Handbook of Smart Materials, Technologies, and Devices: Applications of Industry 4.0 3030842045, 9783030842048

Citation preview

Chaudhery Mustansar Hussain Paolo Di Sia Editors

Handbook of Smart Materials, Technologies, and Devices Applications of Industry 4.0

Handbook of Smart Materials, Technologies, and Devices

Chaudhery Mustansar Hussain • Paolo Di Sia Editors

Handbook of Smart Materials, Technologies, and Devices Applications of Industry 4.0 With 999 Figures and 216 Tables

Editors Chaudhery Mustansar Hussain Department of Chemistry and Environmental Science New Jersey Institute of Technology Newark, NJ, USA

Paolo Di Sia School of Science University of Padova Padova, Italy School of Medicine Department of Neurosciences University of Padova Padova, Italy

ISBN 978-3-030-84204-8 ISBN 978-3-030-84205-5 (eBook) https://doi.org/10.1007/978-3-030-84205-5 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

Industry 4.0 is a digital revolution, concentrated on all digital technologies able to increase the general interconnection and cooperation of resources (people and IT systems), with changes affecting the industrial sector and the society in all aspects. Data assume a primary role in this process, because they are the basis of any operation. The pillars of this revolution are data, analytics, human-machine interaction and manufacturing. The common factor is the interconnection between multiple elements of a system, high levels of communication and the optimal exploitation of all connected services. The related enabling technologies concern the set of technologies and services closest to IT and the operational level. Industry 4.0 is a revolution that gradually affected a growing number of sectors (medicine, industry, education, etc.), increasing their digitization level through the use of modern technologies, and creating an environment in which the processes will be completely automated. Global revolutions like this one tend to a general improvement of the conditions of man and the environment. The ultimate goal should always remain the good of people and the improvement of everyone’s daily life. Technology in its essence is neutral, is the way in which we use it that makes the difference, protecting people by possible general bad uses and by possible further deterioration of the environment and discrimination of the social conditions. This book provides and discusses relevant topics and details related to the Industry 4.0 revolution. It is both an excellent general introduction, and focuses on important aspects related to smart technologies, robotics, the world of nanotechnologies, defense, environment, IoT, medicine, smart devices, green and smart materials, smart farming, sustainability, circular economy. It is a great effort, one of the most complete works on the subject dealt with, that can be also seen as the starting point for discussions and insights and can be used for further developments. The book is dedicated to and helpful for everyone, both experts in the sector and curious and interested people. Verona, Italy September 2022

Paolo Di Sia

v

Contents

Volume 1 Part I Industry 4.0: Concepts, Themes, and Perspectives . . . . . . . . . . .

1

1

Industry 4.0 Revolution: Introduction . . . . . . . . . . . . . . . . . . . . . . . . Paolo Di Sia

3

2

Industry 4.0 Perspectives: Global Trends and Future Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Antonella Petrillo and Fabio De Felice

3

Changing Manufacturing Landscape: From a Factory to a Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Karl-Erik Michelsen, Mikael Collan, Jyrki Savolainen, and Paavo Ritala

4

Karakuri Solutions and Industry 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . Mariusz Kostrzewski and Wojciech Jerzy Nowak

5

Use of Smart Technologies on Textile Industry Workers to Evaluate the Effect of Work Posture on Lower Extremity Distress in Southern Region of India . . . . . . . . . . . . . . . . . . . . . . . . . S. Shankar, R. Naveenkumar, J. Karthick, P. Mohan Kumar, and R. Nithyaprakash

6

7

8

Monitoring and Modeling of Cylindricity Error Using Vibration Signals in Drilling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . J. Susai Mary, D. Dinakaran, M. A. Sai Balaji, S. Satishkumar, and Arockia Selvakumar Arockia Doss

23

47

69

99

117

Logic Based Path Planning (LBPP) Algorithm for Robotic Library System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sagar Ajanalkar and Harshadeep Joshi

127

Design, Control, and Data Management for Cleaning-in-Place (CIP) Test Rig Used in Process Industries . . . . . . . . . . . . . . . . . . . . . A. S. Patil, M. N. Dhavalikar, and S. A. Chavan

161 vii

viii

9

Contents

Control and Informatics for Demand Response and Renewables Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael Short

Part II

Industry 4.0: Mode of Materials, Technology, and Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

193

213

10

From Industry 4.0 to Pharma 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reza Ebrahimi Hariry, Reza Vatankhah Barenji, and Anant Paradkar

215

11

OHS-Related Risks in an Industry 4.0 Manufacturing Plant . . . . . Mohamed Naceur Ben Aziza, Adel Badri, and Foued Chihi

237

12

Open-Source Framework Based on LoRaWAN IoT Technology for Building Monitoring and Its Integration into BIM Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Martín-Garín, J. A. Millán-García, R. J. HernándezMinguillón, M. M. Prieto, N. Alilat, and A. Baïri

257

Metal Additive Manufacturing Technology Applications in Defense Organizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Panagiotis Stavropoulos

285

Mechanical Properties of Additive Manufactured Part Using Fused Deposition Modeling: Influence of Process Parameters . . . . Ramu Murugan, T. Mohanraj, and Lovin K. John

321

Product Lifecycle Management (PLM): A Key Enabler in Implementation of Industry 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vinayak N. Kulkarni, V. N. Gaitonde, and B. B. Kotturshettar

349

Zero Waste as an Approach to Develop a Clean and Sustainable Society . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nazia Tarannum, Nikhil Kumar, and Km Pooja

381

Role of Industry 4.0 in Maintaining Sustainable Production and Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Muhammad Usman Tariq

425

13

14

15

16

17

18

Biomass as Sustainable Material for Bioethanol Production . . . . . Rozina, Mushtaq Ahmad, and Muhammad Zafar

453

19

Machine Learning–Enhanced Decision-Making . . . . . . . . . . . . . . . . Nikodem Rybak and Maureen Hassall

477

Contents

20

21

Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Atslands Rego da Rocha, Igor Leão dos Santos, Letícia Ali Figueiredo Ferreira, and Augusto da Cunha Reis Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kakali Bhadra

22

Mode of Materials, Technology, and Devices . . . . . . . . . . . . . . . . . . . Shivani Jakhar, Surender Duhan, Supriya Sehrawat, Atul Kumar, Sunita Devi, and Sonia Nain

23

IoT-Based Medication Reminder Devices: Design and Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alivelu Manga N. and Sathish P.

24

Ionic Liquids: The Smart Materials in Process Industry . . . . . . . . Kailas L. Wasewar

25

Microstructure Analysis and Multi-objective Optimization of Pulsed TIG Welding of 316/316L Austenite Stainless Steel . . . . . . . Asif Ahmad

ix

519

559 591

613 647

675

26

Role of IoT in Universal Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ravindra Singh, Sumedha Seniaray, and Partha Pratim Das

27

Nanobiomaterials Administration in Modernization of Biological Science: Current Status and Future Potential . . . . . . . . Ashish Singla and Sreedevi Upadhyayula

729

Adoption of Dematerialization Practices in Knowledge Societies in Order to Achieve Sustainable Outcomes . . . . . . . . . . . . Fernanda E. D. Palandi and Jamile Sabatini-Marques

779

A Perspective on the Frictional Properties of Soft Materials as Smart Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vinit Gupta, Arun K. Singh, Nitish Sinha, and Kailas L. Wasewar

795

Butterflies: A New Source of Inspiration for Futuristic Aerial Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chakravarthi Jada, U. Ashok, B. Pavan, and P. Vinod Babu

813

Significance of Bracing Accessories for Improved Workability: An EMG Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . S. Shankar, J. Karthick, R. NaveenKumar, and R. Nithyaprakash

831

28

29

30

31

709

x

32

33

34

35

36

Contents

Design, Implementation, and Experimental Study on 3-RPS Parallel Manipulator-Based Cervical Collar Therapy Device for Elderly Patients Suffering from Cervical Spine Injuries . . . . . Pavan Kalyan Lingampally and Arockia Selvakumar Arockia Doss

847

Performance Comparison of Two-Stage LED Driver for Tube Light Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vandavasi Harikrishna and Ramachandiran Gunabalan

871

Design and Development of Automated Vertical Farming Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Karthik Warrier, Mukundhan Rajendiran, Shrawan Kumaar Kannan, and R. Ranjith Pillai Research Methodology for Augmenting a Gait Cycle of Lower-Body Exoskeleton, by Using a Data of Mathematical Modeling and Motion Study of a Specific User While Obtaining a Customized Gait for Joint Actuation of Exoskeleton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. S. Ohol, K. D. Kalantri, Y. M. Pirjade, A. U. Kotkar, N. M. Patwardhan, D. R. Londhe, and T. P. Shelke Intelligent, Automated, and Web Application-Based Cradle Monitoring System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Priyanka J. Nair and V. Ravi

37

Review on Deep Learning Algorithms in Medical Devices . . . . . . . G. Ananthi and Arockia Selvakumar Arockia Doss

38

Kinematic Modeling and Analysis of Wheeled In-Pipe Inspection Mobile Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rajendran Sugin Elankavi, D. Dinakaran, R. M. Kuppan Chetty, M. M. Ramya, and Arockia Selvakumar Arockia Doss

891

927

953 967

995

Volume 2 Part III

Industry 4.0: Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1011

39

Cloud Hadoop for Enterprise Collaboration System . . . . . . . . . . . . 1013 Hsiao Kang Lin and Tzu-Jou Liao

40

Smart Devices in Healthcare Sector: Applications . . . . . . . . . . . . . . 1023 Kanika Sharma, Payal Kesharwani, Shiv Kumar Prajapati, Ankit Jain, Neha Mittal, Rahul Kaushik, and Nishi Mody

Contents

xi

41

Oil and Gas Upstream Sector: The use of IEC-61499 and OPC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1051 Carlos A. Garcia, Gustavo Caiza, and Marcelo V. Garcia

42

Lower Extremity Exoskeleton Device for Motion Assistance and Gait Rehabilitation: Design Considerations . . . . . . . . . . . . . . . 1083 Jyotindra Narayan, Aditya Kalani, and Santosha K. Dwivedy

43

Fuzzy Membership Functions in ANFIS for Kinematic Modeling of 3R Manipulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1101 Jyotindra Narayan, Sashwata Banerjee, Durgarao Kamireddy, and Santosha K. Dwivedy

44

Mobile Robot for Gas Leakage Detection System in Pipelines . . . . 1121 Gnana K. Sheela

45

Improved Security Models in Mobile Wireless Vehicle Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1137 Gnana K. Sheela

46

Machine Learning: An Expert Thinking System . . . . . . . . . . . . . . . 1165 T. Mohanraj, Jayanthi Yerchuru, R. S. Nithin Aravind, and R. Yameni

47

The Adoption of Industry 4.0 Technologies Through the Implementation of Continuous Improvement Tools . . . . . . . . . . . . . 1185 Maria Rosaria Sessa, Ornella Malandrino, Giuseppe Fenza, Gianfranco Caminale, and Claudio Risso

48

IIOT Applications for Sustainable Manufacturing . . . . . . . . . . . . . 1221 S. Kamalakkannan and A. K. Kulatunga

49

Nano-biomaterials as a Potential Tool for Futuristic Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1243 Anuron Deka, Pritam Bardhan, Manabendra Mandal, and Rupam Kataki

50

Industry 4.0: Applications and Future Perspectives . . . . . . . . . . . . 1277 Rafael Kunst, Gabriel Ramos, Rodrigo Righi, Cristiano André da Costa, Edison Pignaton, Alecio Binotto, Jose Favilla, Ricardo Ohta, and Rob High

51

4.0 Technology for Port Digitalization and Automation . . . . . . . . . 1307 Chalermpong Senarak and Orawan Mokkhavas

52

Textile and Apparel Industry: Industry 4.0 Applications . . . . . . . . 1321 Sanjeev Swami, Debabrata Ghosh, Charu Swami, and Sonali Upadhyaya

xii

Contents

53

Construction Sector: IR 4.0 Applications . . . . . . . . . . . . . . . . . . . . . 1341 Wesam Salah Alaloul, Syed Saad, and Abdul Hannan Qureshi

54

Design of Low-Cost Soft Ankle Exoskeleton Using Soft Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1391 Kathan Rajesh Sonar, S. Sai Sudeep Reddy, Daniel Schilberg, and Arockia Selvakumar Arockia Doss

55

Robotic Arm for Biomedical Applications . . . . . . . . . . . . . . . . . . . . . 1415 Arockia Selvakumar Arockia Doss, Birupakshya Mishra, Safal Mohammed, Pavan Kalyan Lingampally, and Michael Short

56

Progresses on Green and Smart Materials for Multifaceted Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1439 S. O. Oyedepo, Joseph O. Dirisu, N. E. Udoye, and O. S. I. Fayomi

57

Determining Technologies Trends and Evolution of Smart Building Technologies by Bibliometric Analysis from 1984 to 2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1467 Nadia Karina Gamboa-Rosales, Luis Daniel López-Robles, Leonardo B. Furstenau, Michele Kremer Sott, Manuel Jesús Cobo, and José Ricardo López-Robles

58

Managing Process Safety and Operational Risks with Industry 4.0 Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1501 John Lee, Ian Cameron, and Maureen Hassall

59

Lignin: A Renewable Chemical Feedstock . . . . . . . . . . . . . . . . . . . . 1529 Uroosa Ejaz and Muhammad Sohail

60

Emerging Technologies in Diagnostic Virology and Antiviral Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1545 Goutam Patra and Sumi Mukhopadhyay

61

GTAW Application for Additive Manufacturing and Cladding of Steel Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1559 Vishvesh J. Badheka, Vijay S. Gadakh, V. B. Shinde, and G. Bhati

62

Using Smart Mesoporous Silica in Designing Drug Delivery Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1581 Kayambu Kannan

63

IoT-Based Smart Farming System Using MQTT Protocol and ML Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1613 Sathish P. and Alivelu Manga N.

64

Sustainability of Fusion and Solid-State Welding Process in the Era of Industry 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1637 Vijay S. Gadakh and Vishvesh J. Badheka

Contents

xiii

65

Smart Farming: Applications of IoT in Agriculture . . . . . . . . . . . . 1655 Anil Kumar Singh

66

Smart Materials in Oil and Gas Industry: Application . . . . . . . . . . 1689 Alimorad Rashidi and Soheila Sharafinia

67

Food Industry: Applications of Digitalization . . . . . . . . . . . . . . . . . . 1731 Vahid Mohammadpour Karizaki

68

Detection of Tuberculosis and Lung Cancer Using CNN . . . . . . . . 1751 S. N. Hankare and S. S. Shirguppikar

69

Covid-19 or Viral Pneumonia Detection Using AI Tools . . . . . . . . . 1763 M. V. Pachore and S. S. Shirguppikar

Volume 3 Part IV Industry 4.0: Concept of Smart, Intelligent, and Sustainable Society . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1775 70

Iranian Small and Medium-Sized Industries . . . . . . . . . . . . . . . . . . 1777 S. Jithender Kumar Naik, Malek Hassanpour, and Dragan Pamucar

71

Decentralized Privacy: A Distributed Ledger Approach . . . . . . . . . 1805 Pavlos Papadopoulos, Nikolaos Pitropakis, and William J. Buchanan

72

Toward a Circular Economy in the Copper Mining Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1831 Ingrid Jamett, Ernesto D. R. Santibanez Gonzalez, Yecid Jiménez, and Paulina Carrasco

73

Sustainability Index of Metalworking Fluids in the Manufacturing Industry for Sustainable Manufacturing . . . . . . . . 1853 Muralidhar Vardhanapu and Phaneendra Kiran Chaganti

74

Wind Energy System: Data Analysis and Operational Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1881 Vikas Khare and Cheshta J. Khare

75

Biomedical Data Retrieval Using Enhanced Query Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1921 Muhammad Qadeer, Chuadhery Ghazanfar Hussain, and Chaudhery Mustansar Hussain

76

Machine Learning Applications for The Tensile Property Evaluation of Steel: An Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1957 Hridayjit Kalita, Kaushik Kumar, and J. Paulo Davim

xiv

Contents

77

Business Ecosystem Approach to Industry 4.0 . . . . . . . . . . . . . . . . . 1975 Daniel Alejandro Rossit, Marisa Analía Sánchez, Fernando Tohmé, and Mariano Frutos

78

Modeling the Dynamics of a Smart Factory . . . . . . . . . . . . . . . . . . . 1997 Marisa Analía Sánchez, Daniel Alejandro Rossit, and Fernando Tohmé

79

Robotics in Industry 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2021 Ashwin Misra, Anuj Agrawal, and Vihaan Misra

80

Fossil Fuel Combustion, Conversion to Near-Zero Waste Through Organic Rankine Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2057 A. B. Fakeye, S. O. Oyedepo, O. S. I. Fayomi, Joseph O. Dirisu, and N. E. Udoye

81

Flow Shop Scheduling Problems in Industry 4.0 Production Environments: Missing Operation Case . . . . . . . . . . . . . . . . . . . . . . 2077 Daniel Alejandro Rossit, Adrián Toncovich, Diego Gabriel Rossit, and Sergio Nesmachnow

82

Smart Packaging: O2 Scavenger for Improving Quality of Fish . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2101 C. O. Mohan, S. Remya, K. R. Sreelakshmi, Anuj Kumar, and C. N. Ravishankar

83

Social Responsibility Diagnostics as the Sustainable Development Basis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2123 Iryna Moiseienko, Ivanna Dronyuk, and Igor Moyseyenko

84

Green Nanoparticles: Synthesis and Catalytic Applications . . . . . . 2139 Aniruddha B. Patil, Sharwari K. Mengane, and Bhalchandra M. Bhanage

85

From Agriculture to Sustainable Agriculture: Prospects for Improving Pest Management in Industrial Revolution 4.0 . . . . . . . 2171 Farhan Mahmood Shah and Muhammad Razaq

86

Satellite-Based Environmental Impact Assessment of MSW Dumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2189 Fiza Faizi, Khalid Mahmood, and Wajiha Iftikhar

87

Examining the Impact of Industry 4.0 on Labor Market in Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2207 Syed Jawad Ali Kazmi and Jawad Abbas

88

Recent Advances in the MXenes for Photocatalytic and Hydrogen Production Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 2219 Ikhtiar Gul, Murtaza Sayed, Maleeha Bushra, Faryal Gohar, and Qaiser Khan

Contents

xv

89

Big Data Analytics and Advanced Technologies for Sustainable Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2261 Rubab Zahra Naqvi, Muhammad Farooq, Syed Ali Asad Naqvi, Hamid Anees Siddiqui, Imran Amin, Muhammad Asif, and Shahid Mansoor

90

Geographical Information Systems (GIS) in Industry 4.0: Revolution for Sustainable Development . . . . . . . . . . . . . . . . . . . . . . 2289 Syed Ali Asad Naqvi and Rubab Zahra Naqvi

91

Green Nanomaterials: Design, Synthesis, Properties, and Industrial Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2317 Paulraj Mosae Selvakumar, Samiha Nuzhat, Mahia Mohiuddin Quadrey, Sherin Monichan, Rex Jeya Rajkumar Samdavid Thanapaul, and M. S. A. Muthukumar Nadar

92

Green and Sustainable Battery Materials . . . . . . . . . . . . . . . . . . . . . 2337 Andrew Ng Kay Lup

93

Circular Economy in Brazil Coupled with Industry 4.0 . . . . . . . . . 2367 Camila Callegari, Régis Rathmann, Alexandre Skzlo, Sonia Regina Mudrovitsch de Bittencourt, Antônio Marcos Mendonça, and Márcio Rojas da Cruz

94

Centrality Measures in Finding Influential Nodes for the Big-Data Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2393 Sathyanarayanan Gopalakrishnan, Supriya Sridharan, and Swaminathan Venkatraman

95

A Novel Cluster-Based Routing Technique for Reliable Path Selection in VANET V2V Communication in 5G Using Upper Triangular Matrix Lie Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2411 Supriya Sridharan, Sathyanarayanan Gopalakrishnan, and Swaminathan Venkatraman

96

Multiwall Carbon Nanotubes-Based Micro-fibrillar Polymer Composite Fiber: A Sturctural Biomimetic . . . . . . . . . . . . . . . . . . . . 2427 Neha Agrawal, Manu Aggarwal, Kingsuk Mukhopadhyay, and Arup R. Bhattacharyya

97

Industry 4.0: Mode of Materials, Technology, and Devices . . . . . . . 2461 Sardul Singh Dhayal, Atul Kumar, and Surender Duhan

98

Intrinsic Insights of Nanoparticles via Anaerobic Digestion for Enhanced Biogas Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2481 Puneet Kumar Singh, Slipa Kanungo, Snehasish Mishra, and Ritesh Pattnaik

xvi

Contents

99

Green Nanomaterials: Synthesis, Characterization, and Their Industrial Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2507 Atul Kumar, Surender Duhan, Sushma Kumari, Sunita Devi, and Sardul Singh Dhayal

100

Advanced Materials in the Detection of Arsenic from Aquatic Environment: Advancements in Electrochemical Sensors . . . . . . . 2527 Jongte Lalmalsawmi and Diwakar Tiwari

101

Green Bioenergy for Zero Waste: A Road Toward Clean and Sustainable Society . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2559 P. Ujwal, K. Sandesh, and Vinayaka B. Shet

102

Industrial Innovation Through Sustainable Materials . . . . . . . . . . 2577 Gajanan B. Kunde and B. Sehgal

103

Low Energy/Low Carbon Eco-cementitious Binders as an Alternative to Ordinary Portland Cement . . . . . . . . . . . . . . . . . . . . . 2619 Abhishek Srivastava, Rajesh Kumar, and Rajni Lakhani

104

Industry 4.0: Applications in Oil and Gas Industry . . . . . . . . . . . . . 2641 Anil K. Saroha and Abhijit Bikas Pal

105

Performance of Blended Mortars Containing Industrial and Agricultural By-Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2667 Navdeep Singh, Nitin Ankur, P. Ashik Yashi, and Sunny Gupta

106

Green Nanomaterials: Sustainable Approach for Environmental Remediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2699 Satyajeet Arya, Alka Rathor, Rakhi Tyagi, and Vikas Chaudhary

Part V Industry 4.0: Dangers/Warning Points . . . . . . . . . . . . . . . . . . . . 2717 107

Environmental Side of Fourth Industrial Revolution: The Positive and Negative Effects of I4.0 Technologies . . . . . . . . . . 2719 Adele Parmentola, Ilaria Tutore, and Michele Costagliola Di Fiore

108

Single-Atom Photocatalysts for Energy and Environmental Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2751 Akshat Khandelwal, Dileep Maarisetty, and Saroj Sundar Baral

109

Sustainable Development and Industry 4.0 . . . . . . . . . . . . . . . . . . . . 2789 Muhammad Waqar Akram, Khalil Rehman, Syed Mohsin bukhari, Nida Akram, and Shahla Andleeb

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2813

About the Editors

Chaudhery Mustansar Hussain, PhD, is an adjunct professor and director of laboratories in the Department of Chemistry & Environmental Sciences at the New Jersey Institute of Technology (NJIT), Newark, New Jersey, USA. His research is focused on the applications of nanotechnology and advanced materials, environmental management, analytical chemistry, smart materials and technologies, and other various industries. Dr. Hussain is the author of numerous papers in peerreviewed journals and is a prolific author and editor of around hundred (100) books, including scientific monographs and handbooks, in his research areas. He has published with Elsevier, the American Chemical Society, the Royal Society of Chemistry, Springer, John Wiley & Sons, and CRC Press. Paolo Di Sia is currently adjunct professor/senior lecturer at the University of Padova (Italy). He holds a bachelor’s degree along with a master’s degree and three PhDs. His scientific interests span transdisciplinary physics, classical and quantum-relativistic nanophysics, nano-biotechnology, nano-neuroscience, theories of everything, foundations of physics, history, and philosophy of science. He is author of more than 300 works to date (papers in national and international journals, book chapters, books, internal academic works, works on scientific web pages, and popular papers), is reviewer of some academic books, editor of some international academic books, and reviewer of many international journals.

xvii

xviii

About the Editors

He has obtained several international awards and is member of many scientific societies as well as international advisory/editorial boards. Personal web page: www.paolodisia.com Email address: [email protected]

Contributors

Jawad Abbas Department of Business Administration, Iqra University Islamabad, Islamabad, Pakistan Manu Aggarwal Directorate of Nanomaterials and Sciences, DMSRDE, DRDO, Kanpur, India Department of Textile Technology, Dr. B.R. Ambedkar NIT, Jalandhar, India Anuj Agrawal Delhi Technological University, New Delhi, India Neha Agrawal Department of Neurobiology, DIPAS, DRDO, New Delhi, India Directorate of Nanomaterials and Sciences, DMSRDE, DRDO, Kanpur, India Centre for Research in Nanotechnology and Sciences, IIT-Bombay, Mumbai, India Asif Ahmad Mechanical Engineering Department, PSIT-Kanpur, Kanpur, India Mushtaq Ahmad Biofuel Lab, Department of Plant Sciences, Quaid-i-Azam University, Islamabad, Pakistan Sagar Ajanalkar Dr. Babasaheb Ambedkar Technological University, Lonere, India Muhammad Waqar Akram Ilama University, Sindh, Pakistan Nida Akram Department of Management Science, Govt. College for Women University, Faisalabad, Punjab, Pakistan Wesam Salah Alaloul Department of Civil and Environmental Engineering, Universiti Teknologi PETRONAS, Seri Iskandar, Perak Darul Ridzuan, Malaysia N. Alilat Département Génie Thermique et Énergie (GTE), Université de Paris, Laboratoire Thermique Interfaces Environnement (LTIE), Ville d’Avray, France Imran Amin Agricultural Biotechnology Division, National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan G. Ananthi Department of ECE, Thiagarajar College of Engineering, Madurai, Tamil Nadu, India xix

xx

Contributors

Shahla Andleeb Department of Environmental Science, Government College Women University, Sialkot, Pakistan Nitin Ankur Department of Civil Engineering, Dr B R Ambedkar National Institute of Technology, Jalandhar, India R. S. Nithin Aravind Department of Mechanical Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, India Arockia Selvakumar Arockia Doss Design and Automation Research Group, School of Mechanical Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India Satyajeet Arya Faculty of Management Studies, Sri Sri University, Cuttack, India P. Ashik Yashi Department of Civil Engineering, Dr B R Ambedkar National Institute of Technology, Jalandhar, India U. Ashok Rajiv Gandhi University of Knowledge Technologies, Nuzvid, India Muhammad Asif Agricultural Biotechnology Division, National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan Vishvesh J. Badheka Department of Mechanical Engineering, School of Technology, Pandit Deendayal Energy University, Gandhinagar, Gujarat, India Adel Badri Industrial Engineering Department, School of Engineering, UQTR, Trois-Rivières, QC, Canada A. Baïri Département Génie Thermique et Énergie (GTE), Université de Paris, Laboratoire Thermique Interfaces Environnement (LTIE), Ville d’Avray, France Sashwata Banerjee Department of Electrical Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India Saroj Sundar Baral Department of Chemical Engineering, BITS Pilani K K Birla Goa Campus, Goa, India Pritam Bardhan Department of Molecular Biology and Biotechnology, Tezpur University, Tezpur, Assam, India Reza Vatankhah Barenji Department of Industrial Engineering, Hacettepe University, Ankara, Turkey Mohamed Naceur Ben Aziza School of Management, Université du Québec à Trois-Rivières (UQTR), Trois-Rivières, QC, Canada Kakali Bhadra Department of Zoology, University of Kalyani, Nadia, West Bengal, India Bhalchandra M. Bhanage Department of Chemistry, Institute of Chemical Technology, Matunga, Mumbai, India

Contributors

xxi

G. Bhati Department of Mechanical Engineering, School of Technology, Pandit Deendayal Energy University, Gandhinagar, Gujarat, India Arup R. Bhattacharyya Department of Metallurgical Engineering and Material Science, IIT-Bombay, Mumbai, India Alecio Binotto IBM Consulting, München, Germany William J. Buchanan Blockpass ID Lab, Edinburgh Napier University, Edinburgh, UK Maleeha Bushra Radiation and Environmental Chemistry Laboratory, National Centre of Excellence in Physical Chemistry, University of Peshawar, Peshawar, Pakistan Gustavo Caiza Universidad Politecnica Salesiana, UPS, Quito, Ecuador Camila Callegari Centre for Energy and Environmental Economics, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil Ian Cameron School of Chemical Engineering, The University of Queensland, Brisbane, QLD, Australia Gianfranco Caminale CTO Cyber Security Division, LEONARDO Company, Genoa, Italy Paulina Carrasco Departamento de Ingeniería Industrial, Facultad de Ingeniería, Universidad de Antofagasta, Antofagasta, Chile Phaneendra Kiran Chaganti Department of Mechanical Engineering, Birla Institute of Technology & Science, Pilani Hyderabad Campus, Hyderabad, India Vikas Chaudhary CSIR-Central Scientific Instruments Organisation (CSIO), Chandigarh, India Academy of Scientific and Innovative Research (AcSIR), Ghaziabad, India S. A. Chavan Analogic Automation Pvt. Ltd, Pune, India Foued Chihi Department of Finance and Economics, School of Management, UQTR, Trois-Rivières, QC, Canada Manuel Jesús Cobo Department of Computer Science and Engineering, University of Cadiz, Cadiz, Spain Mikael Collan School of Business and Management, Lappeenranta-Lahti University of Technology, Lappeenranta, Finland VATT Institute for Economic Research, Helsinki, Finland Cristiano André da Costa University of Vale do Rio dos Sinos (UNISINOS), São Leopoldo, Brazil

xxii

Contributors

Márcio Rojas da Cruz Ministry of Science, Technology, and Innovations (MCTI), Brasília, Brazil Michele Costagliola Di Fiore Department of Management and Quantitative Studies, University of Naples “Parthenope”, Naples, Italy Partha Pratim Das Delhi Technological University, Delhi, India J. Paulo Davim Department of Mechanical Engineering, University of Aveiro, Aveiro, Portugal Sonia Regina Mudrovitsch de Bittencourt Ministry of Science, Technology, and Innovations (MCTI), Brasília, Brazil Fabio De Felice Department of Civil and Mechanical Engineering, Università degli Studi di Cassino e del Lazio Meridionale, Cassino, Italy Anuron Deka Department of Energy, Tezpur University,Tezpur, Assam, India Sunita Devi Organic Chemistry Research Laboratory, Department of Chemistry, M.K.J.K. Collage, Rohtak, India M. N. Dhavalikar MIT School of Engineering, MIT ADT University, Pune, India Sardul Singh Dhayal Department of ECE, Guru Jambheshwar University of Science and Technology, Hisar, Haryana, India D. Dinakaran Centre for Automation and Robotics (ANRO), School of Mechanical Sciences, Hindustan Institute of Technology and Science, Chennai, TN, India Joseph O. Dirisu Mechanical Engineering Department, Covenant University, Ota, Ogun State, Nigeria Paolo Di Sia School of Science, University of Padova, Padova, Italy School of Medicine, Department of Neurosciences, University of Padova, Padova, Italy Ivanna Dronyuk ACS Department, LPNU, Lviv, Ukraine Surender Duhan Nanomaterials Research Laboratory, Department of Physics, Deen Bandhu Chhoturam University of Science and Technology, Murthal Sonepat, Haryana, India Santosha K. Dwivedy Department of Mechanical Engineering, Indian Institute of Technology, Guwahati, Assam, India Uroosa Ejaz Department of Microbiology, University of Karachi, Karachi, Pakistan Rajendran Sugin Elankavi Centre for Automation and Robotics (ANRO), School of Mechanical Sciences, Hindustan Institute of Technology and Science, Chennai, TN, India

Contributors

xxiii

Fiza Faizi Remote Sensing, GIS and Climatic Research Lab (National Centre for GIS and Space Application), Centre for Remote Sensing, University of the Punjab, Lahore, Pakistan A. B. Fakeye Mechanical Engineering Department, Federal Polytechnic, Ilaro, Nigeria Muhammad Farooq Agricultural Biotechnology Division, National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan Jose Favilla IBM Global Markets, Coppell, TX, USA O. S. I. Fayomi Department of Mechanical and Biomedical Engineering, Bells University of Technology, Ota, Ogun State, Nigeria Giuseppe Fenza Department of Management & Innovation Systems, University of Salerno, Fisciano, SA, Italy Letícia Ali Figueiredo Ferreira Centro Federal de Educação Tecnológica Celso Suckow da Fonseca, Rio de Janeiro, Brasil Mariano Frutos Departamento de Ingeniería, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina Leonardo B. Furstenau Department of Industrial Engineering, Federal University of Rio Grande do Sul, Rio Grande do Sul, Brazil Vijay S. Gadakh Department of Mechanical Engineering, Amrutvahini College of Engineering, Ahmednagar, Savitribai Phule Pune University, Pune, India V. N. Gaitonde School of Mechanical Engineering, KLE Technological University, Hubballi, India Nadia Karina Gamboa-Rosales CONACYT – Autonomous University of Zacatecas, Zacatecas, Mexico Carlos A. Garcia Universidad Tecnica de Ambato, UTA,Ambato, Ecuador Marcelo V. Garcia University of Basque Country, UPV/EHU,Bilbao, Spain Debabrata Ghosh Operations and Supply Chain Management Area, MIT Global Scale Network – Malaysia Institute for Supply Chain Innovation, Shah Alam, Malaysia Faryal Gohar Radiation and Environmental Chemistry Laboratory, National Centre of Excellence in Physical Chemistry, University of Peshawar, Peshawar, Pakistan Sathyanarayanan Gopalakrishnan Department of Mathematics, School of Arts, Science, Humanities and Education, SASTRA Deemed to be University, Thanjavur, India Ikhtiar Gul Radiation and Environmental Chemistry Laboratory, National Centre of Excellence in Physical Chemistry, University of Peshawar, Peshawar, Pakistan

xxiv

Contributors

Ramachandiran Gunabalan School of Electrical Engineering, VIT Chennai, Chennai, India Sunny Gupta Department of Civil Engineering, Dr B R Ambedkar National Institute of Technology, Jalandhar, India Vinit Gupta Department of Mechanical Engineering, S.B. Jain Institute of Technology, Management and Research, Nagpur, India S. N. Hankare Department of Mechanical Engineering, Rajararambapu Institute of Technology, Rajaramnagar, India Vandavasi Harikrishna IgrenEnergi Services Pvt Ltd, Bengaluru, India Reza Ebrahimi Hariry Department of Pharmacology and Toxicology, Faculty of Veterinary Medicine, Ankara University, Ankara, Turkey Maureen Hassall School of Chemical Engineering, The University of Queensland, Brisbane, QLD, Australia Malek Hassanpour Department of Environmental science, UCS, Osmania University, Hyderabad, Telangana State, India R. J. Hernández-Minguillón CAVIAR Research Group, Department of Architecture, Higher Technical School of Architecture, University of the Basque Country UPV/EHU, Donostia-San Sebastián, Spain Rob High IBM Cloud and Cognitive Software, Durham, NC, USA Chuadhery Ghazanfar Hussain Department of Education, Computer science and Technology, Punjab, Pakistan Chaudhery Mustansar Hussain Department of Chemistry and Environmental Science, New Jersey Institute of Technology, Newark, NJ, USA Wajiha Iftikhar Remote Sensing, GIS and Climatic Research Lab (National Centre for GIS and Space Application), Centre for Remote Sensing and Department of Space Science, University of the Punjab, Lahore, Pakistan Chakravarthi Jada Rajiv Gandhi University of Knowledge Technologies, Nuzvid, India Ankit Jain Department of Materials Engineering, Indian Institute of Science, Bangalore, Karnataka, India Shivani Jakhar Nanomaterials Research Laboratory, Department of Physics, D.C.R. University of Science and Technology, Murthal, India Inorganic Chemistry Research Laboratory, Department of Chemistry, D. C. R. University of Science and Technology, Murthal, India Ingrid Jamett Departamento de Ingeniería Industrial, Facultad de Ingeniería, Universidad de Antofagasta, Antofagasta, Chile

Contributors

xxv

Yecid Jiménez Departamento de Ingeniería Química y Procesos de Minerales, Facultad de Ingeniería, Universidad de Antofagasta, Antofagasta, Chile S. Jithender Kumar Naik University College of Science, Osmania University, Hyderabad, Telangana State, India Lovin K. John Department of Mechanical Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, India Harshadeep Joshi Dr. Babasaheb Ambedkar Technological University, Lonere, India Aditya Kalani Department of Mechanical Engineering, Indian Institute of Technology, Guwahati, Assam, India K. D. Kalantri Department of Mechanical Engineering, College of Engineering Pune (COEP), Pune, India Hridayjit Kalita Department of Mechanical Engineering, Birla Institute of Technology, Ranchi, India S. Kamalakkannan Department of Manufacturing & Industrial Engineering, Faculty of Engineering, University of Peradeniya, Peradeniya, Sri Lanka Durgarao Kamireddy Department of Mechanical Engineering, Indian Institute of Technology,Guwahati, Assam, India Kayambu Kannan PG and Research Department of Chemistry, Raja Doraisingam Government Arts College, Sivagangai, Tamil Nadu, India Shrawan Kumaar Kannan Department of Mechatronics Engineering, College of Engineering and Technology, SRM Institute of Science and Technology, Chennai, Tamil Nadu, India Slipa Kanungo School of Biotechnology, Kalinga Institute of Industrial Technology (Deemed to be University), Bhubaneswar, Odisha, India Vahid Mohammadpour Karizaki Chemical Engineering Department, Quchan University of Technology, Quchan, Iran J. Karthick Department of Mechatronics Engineering, Kongu Engineering College, Perundurai, Erode, Tamil Nadu, India Rupam Kataki Department of Energy, Tezpur University,Tezpur, Assam, India Rahul Kaushik Department of Pharmacy, Ram-Eesh Institute of Vocational and Technical Education, Greater Noida, Uttar Pradesh, India Andrew Ng Kay Lup School of Energy and Chemical Engineering, Xiamen University Malaysia, Selangor, Malaysia College of Chemistry and Chemical Engineering, Xiamen University, Fujian, China

xxvi

Contributors

Syed Jawad Ali Kazmi Department of Business Administration, Iqra University Islamabad, Islamabad, Pakistan Payal Kesharwani Department of Pharmacy, Ram-Eesh Institute of Vocational and Technical Education, Greater Noida, Uttar Pradesh, India Khalil Rehman Department of Environmental Science, Government College Women University, Sialkot, Pakistan Qaiser Khan Radiation and Environmental Chemistry Laboratory, National Centre of Excellence in Physical Chemistry, University of Peshawar, Peshawar, Pakistan Akshat Khandelwal Department of Chemical Engineering, BITS Pilani K K Birla Goa Campus, Goa, India Cheshta J. Khare SGSITS, Indore, India Vikas Khare STME, NMIMS, Indore, India Mariusz Kostrzewski Faculty of Transport, Division of Construction Fundamentals of Transport Equipment, Warsaw University of Technology, Warsaw, Poland A. U. Kotkar Department of Mechanical Engineering, College of Engineering Pune (COEP), Pune, India B. B. Kotturshettar School of Mechanical Engineering, KLE Technological University, Hubballi, India A. K. Kulatunga Department of Manufacturing & Industrial Engineering, Faculty of Engineering, University of Peradeniya, Peradeniya, Sri Lanka Vinayak N. Kulkarni School of Mechanical Engineering, KLE Technological University, Hubballi, India Anuj Kumar ICAR-Central Institute of Fisheries Technology, Kochi, India Atul Kumar Nanomaterials Research Laboratory, Department of Physics, Deen Bandhu Chhoturam University of Science and Technology, Murthal Sonepat, Haryana, India Kaushik Kumar Department of Mechanical Engineering, Birla Institute of Technology, Ranchi, India Nikhil Kumar Department of Chemistry, Chaudhary Charan Singh University, Meerut, India P. Mohan Kumar Department of Mechatronics Engineering, Kongu Engineering College, Perundurai, Erode, Tamil Nadu, India Rajesh Kumar CSIR-Central Building Research Institute, Roorkee, Uttarakhand, India Sushma Kumari Department of Physics, Deenbandhu Chhotu Ram University of Science and Technology, Murthal, India

Contributors

xxvii

Gajanan B. Kunde Department of Chemistry, Indian Institute of Technology Delhi, Hauz Khas, New Delhi, India Rafael Kunst University of Vale do Rio dos Sinos (UNISINOS), São Leopoldo, Brazil R. M. Kuppan Chetty Centre for Automation and Robotics (ANRO), School of Mechanical Sciences, Hindustan Institute of Technology and Science, Chennai, TN, India Rajni Lakhani CSIR-Central Building Research Institute, Roorkee, Uttarakhand, India Jongte Lalmalsawmi Department of Chemistry, School of Physical Sciences, Mizoram University, Aizawl, India John Lee School of Chemical Engineering, The University of Queensland, Brisbane, QLD, Australia Tzu-Jou Liao Department of Industrial Management, I-Shou University, Kaohsiung City, Taiwan Hsiao Kang Lin Department of Industrial Management, I-Shou University, Kaohsiung City, Taiwan Pavan Kalyan Lingampally School of Mechanical Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India D. R. Londhe Department of Mechanical Engineering, College of Engineering Pune (COEP), Pune, India José Ricardo López-Robles Academic Unit of Accounting and Management, Autonomous University of Zacatecas, Zacatecas, Mexico Department of Computer Science and Engineering, University of Cadiz, Cadiz, Spain Luis Daniel López-Robles Academic Unit of Accounting and Management, Autonomous University of Zacatecas, Zacatecas, Mexico Dileep Maarisetty Department of Chemical Engineering, BITS Pilani K K Birla Goa Campus, Goa, India Khalid Mahmood Remote Sensing, GIS and Climatic Research Lab (National Centre for GIS and Space Application), Centre for Remote Sensing and Department of Space Science, University of the Punjab, Lahore, Pakistan Ornella Malandrino Department of Management & Innovation Systems, University of Salerno, Fisciano, SA, Italy Manabendra Mandal Department of Molecular Biology and Biotechnology, Tezpur University, Tezpur, Assam, India

xxviii

Contributors

Shahid Mansoor Agricultural Biotechnology Division, National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan A. Martín-Garín ENEDI Research Group, Department of Thermal Engineering, Faculty of Engineering of Gipuzkoa, University of the Basque Country UPV/EHU, Donostia-San Sebastián, Spain Antônio Marcos Mendonça Ministry of Science, Technology, and Innovations (MCTI), Brasília, Brazil Sharwari K. Mengane Department of Botany, M. H. Shinde Mahavidyalaya, Tisangi, Kolhapur, India Karl-Erik Michelsen School of Business and Management, Lappeenranta-Lahti University of Technology, Lappeenranta, Finland J. A. Millán-García ENEDI Research Group, Department of Thermal Engineering, Faculty of Engineering of Gipuzkoa, University of the Basque Country UPV/EHU, Donostia-San Sebastián, Spain Birupakshya Mishra School of Mechanical Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India Snehasish Mishra Bioenergy Lab and BDTC, Kalinga Institute of Industrial Technology (Deemed University), Bhubaneswar, Odisha, India Ashwin Misra Robotics Institute, Carnegie Mellon University, Pittsburgh, USA Vihaan Misra Netaji Subhas University of Technology, New Delhi, India Neha Mittal Department of Pharmacy, Ram-Eesh Institute of Vocational and Technical Education, Greater Noida, Uttar Pradesh, India Nishi Mody Department of Pharmaceutical Sciences, Dr. Hari Singh Gour Central University, Sagar, Madhya Pradesh, India Safal Mohammed School of Mechanical Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India C. O. Mohan ICAR-Central Institute of Fisheries Technology, Kochi, India T. Mohanraj Department of Mechanical Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, India Syed Mohsin bukhari Department of Wildlife and Ecology, Faculty of Fisheries & Wildlife, University of Veterinary and Animal Sciences, Lahore, Pakistan Iryna Moiseienko Financial Department, Lviv State University of Internal Affairs, Lviv, Ukraine Orawan Mokkhavas Maritime Transportation Program, Department of Nautical Science and Maritime Logistics, Faculty of International Maritime Studies, Kasetsart University, Sri Racha, Thailand

Contributors

xxix

Sherin Monichan Panaiyaanmai (Palmyraculture), The Centre for Self-Reliance and Sustainable Development, Munnetram Green Industry, Tenkasi, Tamil Nadu, India Igor Moyseyenko Department of Theoretical Economics, Lviv Trade and Economic University, Lviv, Ukraine Kingsuk Mukhopadhyay Directorate of Nanomaterials and Technology, DMSRDE, DRDO, Kanpur, India Sumi Mukhopadhyay Department of Laboratory Medicine, School of Tropical Medicine, Kolkata, West Bengal, India Ramu Murugan Department of Mechanical Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, India M. S. A. Muthukumar Nadar Department of Biotechnology, Karunya Institute of Technology and Sciences (Deemed to be University), Coimbatore, Tamilnadu, India Alivelu Manga N. Chaitanya Bharathi Institute of Technology, Osmania University, Hyderabad, India Sonia Nain Inorganic Chemistry Research Laboratory, Department of Chemistry, D. C. R. University of Science and Technology, Murthal, India Priyanka J. Nair Mercedes-Benz Research and Development, Bengaluru, India Rubab Zahra Naqvi Agricultural Biotechnology Division, National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan Syed Ali Asad Naqvi Department of Geography, Government College University, Faisalabad, Pakistan Jyotindra Narayan Department of Mechanical Engineering, Indian Institute of Technology, Guwahati, Assam, India R. NaveenKumar Department of Mechanical Engineering, Kongu Engineering College, Perundurai, Erode, Tamil Nadu, India Sergio Nesmachnow Facultad de Ingeniería, Universidad de la República, Montevideo, Uruguay R. Nithyaprakash Department of Mechatronics Engineering, Kongu Engineering College, Perundurai, Erode, Tamil Nadu, India Wojciech Jerzy Nowak Faculty of Humanities, Institute of Literary Studies, Department of Oriental Studies, Nicolaus Copernicus University, Toru´n, Poland Samiha Nuzhat Science and Math Program, Asian University for Women, Chittagong, Bangladesh S. S. Ohol Department of Mechanical Engineering, College of Engineering Pune (COEP), Pune, India

xxx

Contributors

Ricardo Ohta IBM Research, São Paulo, Brazil S. O. Oyedepo Mechanical Engineering Department, Covenant University, Ota, Ogun State, Nigeria Sathish P. Chaitanya Bharathi Institute of Technology, Osmania University, Hyderabad, India M. V. Pachore Department of Mechanical Engineering, Rajararambapu Institute of Technology, Rajaramnagar, India Abhijit Bikas Pal Honeywell UOP, Unitech Trade Centre, Gurugram, India Fernanda E. D. Palandi Engineering and Knowledge Management, Federal University of Santa Catarina, Trindade, Florianópolis, Brazil Dragan Pamucar Department of logistics, Military Academy, University of Defence, Belgrade, Serbia Pavlos Papadopoulos Blockpass ID Lab, Edinburgh Napier University, Edinburgh, UK Anant Paradkar Centre for Pharmaceutical Engineering Science, University of Bradford, Bradford, UK Adele Parmentola Department of Management and Quantitative Studies, University of Naples “Parthenope”, Naples, Italy A. S. Patil MIT School of Engineering, MIT ADT University, Pune, India Aniruddha B. Patil Department of Chemistry, Maharshi Dayanand College, Parel, Mumbai, India Goutam Patra Department of Laboratory Medicine, School of Tropical Medicine, Kolkata, West Bengal, India Ritesh Pattnaik School of Biotechnology, Kalinga Institute of Industrial Technology (Deemed to be University), Bhubaneswar, Odisha, India N. M. Patwardhan Department of Mechanical Engineering, College of Engineering Pune (COEP), Pune, India B. Pavan Rajiv Gandhi University of Knowledge Technologies, Nuzvid, India Antonella Petrillo Department of Engineering, University of Naples “Parthenope”, Naples, Italy Edison Pignaton Federal University of Rio Grande do Sul (UFRGS), Porto Alegre, Brazil Y. M. Pirjade Department of Mechanical Engineering, College of Engineering Pune (COEP), Pune, India

Contributors

xxxi

Nikolaos Pitropakis Blockpass ID Lab, Edinburgh Napier University, Edinburgh, UK Km Pooja Department of Chemistry, Chaudhary Charan Singh University, Meerut, India Shiv Kumar Prajapati Department of Pharmacy, Ram-Eesh Institute of Vocational and Technical Education, Greater Noida, Uttar Pradesh, India M. M. Prieto Energy Department, Campus de Viesques, University of Oviedo, Gijón, Spain Muhammad Qadeer Department of Education, Computer science and Technology, Punjab, Pakistan Mahia Mohiuddin Quadrey Science and Math Program, Asian University for Women, Chittagong, Bangladesh Abdul Hannan Qureshi Department of Civil and Environmental Engineering, Universiti Teknologi PETRONAS, Seri Iskandar, Perak Darul Ridzuan, Malaysia Mukundhan Rajendiran Department of Mechatronics Engineering, College of Engineering and Technology, SRM Institute of Science and Technology, Chennai, Tamil Nadu, India Gabriel Ramos University of Vale do Rio dos Sinos (UNISINOS), São Leopoldo, Brazil M. M. Ramya Centre for Automation and Robotics (ANRO), School of Mechanical Sciences, Hindustan Institute of Technology and Science, Chennai, TN, India R. Ranjith Pillai Department of Mechatronics Engineering, College of Engineering and Technology, SRM Institute of Science and Technology, Chennai, Tamil Nadu, India Alimorad Rashidi Nanotechnology Research Center, Research Institute of Petroleum Industry, Tehran, Iran Régis Rathmann Centre for Energy and Environmental Economics, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil Alka Rathor Institute of Environmental Studies, Kurukshetra University, Kurukshetra, India V. Ravi School of Electronics Engineering, Vellore Institute of Technology, Chennai, India C. N. Ravishankar ICAR-Central Institute of Fisheries Technology, Kochi, India Muhammad Razaq Department of Entomology, Faculty of Agricultural Sciences & Technology, Bahauddin Zakariya University, Multan, Pakistan

xxxii

Contributors

S. Sai Sudeep Reddy School of Mechanical Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India Augusto da Cunha Reis Centro Federal de Educação Tecnológica Celso Suckow da Fonseca, Rio de Janeiro, Brasil S. Remya ICAR-Central Institute of Fisheries Technology, Kochi, India Rodrigo Righi University of Vale do Rio dos Sinos (UNISINOS), São Leopoldo, Brazil Claudio Risso Critical Infrastructures, EPC & Large Enterprise – Cyber Security Division, LEONARDO Company, Genoa, Italy Paavo Ritala School of Business and Management, Lappeenranta-Lahti University of Technology, Lappeenranta, Finland Atslands Rego da Rocha Universidade Federal do Ceará, Fortaleza, Brasil Daniel Alejandro Rossit Department of Engineering, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina Diego Gabriel Rossit Departamento de Ingeniería, Universidad Nacional del Sur, Buenos Aires, Argentina INMABB UNS CONICET, Departamento de Matemática, Buenos Aires, Argentina Rozina Biofuel Lab, Department of Plant Sciences, Quaid-i-Azam University, Islamabad, Pakistan Nikodem Rybak School of Chemical Engineering, The University of Queensland, St Lucia, QLD, Australia Syed Saad Department of Civil and Environmental Engineering, Universiti Teknologi PETRONAS, Seri Iskandar, Perak Darul Ridzuan, Malaysia Jamile Sabatini-Marques Engineering and Knowledge Management, Federal University of Santa Catarina, Trindade, Florianópolis, Brazil M. A. Sai Balaji Department of Mechanical Engineering, B.S.A Crescent Institute of Science and Technology, Chennai, Tamil Nadu, India Rex Jeya Rajkumar Samdavid Thanapaul Department of Surgery, Boston University School of Medicine, Boston, MA, USA Marisa Analía Sánchez Departamento de Ciencias de la Administración, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina K. Sandesh Department of Biotechnology Engineering, N.M.A.M. Institute of Technology (Visvesvaraya Technological University, Belagavi), Udupi District, Karnataka, India Ernesto D. R. Santibanez Gonzalez CES 4.0, Department of Industrial Engineering, Faculty of Engineering, University of Talca, Curicó, Chile

Contributors

xxxiii

Igor Leão dos Santos Centro Federal de Educação Tecnológica Celso Suckow da Fonseca, Rio de Janeiro, Brasil Anil K. Saroha Chemical Engineering Department, Indian Institute of Technology, Delhi, India S. Satishkumar Department of Mechanical Engineering, Velammal Engineering College, Chennai, Tamil Nadu, India Jyrki Savolainen School of Business and Management, Lappeenranta-Lahti University of Technology, Lappeenranta, Finland Murtaza Sayed Radiation and Environmental Chemistry Laboratory, National Centre of Excellence in Physical Chemistry, University of Peshawar, Peshawar, Pakistan Daniel Schilberg University of Applied Sciences, Bochum, Germany B. Sehgal Department of Applied Chemistry, Faculty of Technology and Engineering, The Maharaja Sayajirao University of Baroda, Vadodara, India Supriya Sehrawat Nanomaterials Research Laboratory, Department of Physics, D.C.R. University of Science and Technology, Murthal, India Paulraj Mosae Selvakumar Science and Math Program, Asian University for Women, Chittagong, Bangladesh Panaiyaanmai (Palmyraculture), The Centre for Self-Reliance and Sustainable Development, Munnetram Green Industry, Tenkasi, Tamil Nadu, India Chalermpong Senarak Maritime Transportation Program, Department of Nautical Science and Maritime Logistics, Faculty of International Maritime Studies, Kasetsart University, Sri Racha, Thailand Sumedha Seniaray Delhi Technological University, Delhi, India Maria Rosaria Sessa Department of Management & Innovation Systems, University of Salerno, Fisciano, SA, Italy Farhan Mahmood Shah Department of Entomology, Faculty of Agricultural Sciences & Technology, Bahauddin Zakariya University, Multan, Pakistan S. Shankar Department of Mechatronics Engineering, Kongu Engineering College, Perundurai, Erode, Tamil Nadu, India Soheila Sharafinia Department of Chemistry, Faculty of Science, Shahid Chamran University of Ahvaz, Ahvaz, Iran Kanika Sharma Department of Pharmacy, Ram-Eesh Institute of Vocational and Technical Education, Greater Noida, Uttar Pradesh, India Gnana K. Sheela APJ Abdul Technological University, Trivandrum, India

xxxiv

Contributors

T. P. Shelke Department of Mechanical Engineering, College of Engineering Pune (COEP), Pune, India Vinayaka B. Shet Department of Biotechnology Engineering, N.M.A.M. Institute of Technology (Visvesvaraya Technological University, Belagavi), Udupi District, Karnataka, India V. B. Shinde Department of Production Engineering, Amrutvahini College of Engineering, Ahmednagar, Savitribai Phule Pune University, Pune, India S. S. Shirguppikar Department of Mechanical Engineering, Rajararambapu Institute of Technology, Rajaramnagar, India Michael Short School of Computing, Engineering and Digital Technologies, Teesside University, Middlesbrough, UK Hamid Anees Siddiqui Agricultural Biotechnology Division, National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan Anil Kumar Singh Department of Life Sciences, Sant Baba Bhag Singh University, Jalandhar, Punjab, India Arun K. Singh Department of Mechanical Engineering, Visvesvaraya National Institute of Technology, Nagpur, India Navdeep Singh Department of Civil Engineering, Dr B R Ambedkar National Institute of Technology, Jalandhar, India Puneet Kumar Singh Bioenergy Lab and BDTC, Kalinga Institute of Industrial Technology (Deemed University), Bhubaneswar, Odisha, India Ravindra Singh Delhi Technological University, Delhi, India Ashish Singla Department of Chemical Engineering, Indian Institute of Technology Delhi, New Delhi, India Nitish Sinha Department of Mechanical Engineering, G.H. Raisoni Institute of Business Management, Jalgaon, India Alexandre Skzlo Centre for Energy and Environmental Economics, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil Muhammad Sohail Department of Microbiology, University of Karachi, Karachi, Pakistan Kathan Rajesh Sonar School of Mechanical Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India Michele Kremer Sott Business School, Unisinos University, Porto Alegre, Brazil K. R. Sreelakshmi ICAR-Central Institute of Fisheries Technology, Kochi, India Supriya Sridharan Department of Mathematics, School of Arts, Science, Humanities and Education, SASTRA Deemed to be University, Thanjavur, India

Contributors

xxxv

Abhishek Srivastava CSIR-Central Building Research Institute, Roorkee, Uttarakhand, India Panagiotis Stavropoulos Laboratory for Manufacturing Systems and Automation (LMS), Department of Mechanical Engineering and Aeronautics, University of Patras, Patras, Greece J. Susai Mary Department of Electronics and Instrumentation Engineering, Chennai, Tamil Nadu, India Charu Swami Department of Home Science, Faculty of Arts, Dayalbagh Educational Institute, Agra, India Sanjeev Swami Department of Management, Dayalbagh Educational Institute, Agra, India Nazia Tarannum Department of Chemistry, Chaudhary Charan Singh University, Meerut, India Muhammad Usman Tariq Abu Dhabi School of Management, Abu Dhabi, UAE Diwakar Tiwari Department of Chemistry, School of Physical Sciences, Mizoram University, Aizawl, India Fernando Tohmé Departamento de Economía, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina Adrián Toncovich Departamento de Ingeniería, Universidad Nacional del Sur, Buenos Aires, Argentina Ilaria Tutore Department of Management and Quantitative Studies, University of Naples “Parthenope”, Naples, Italy Rakhi Tyagi Institute of Environmental Studies, Kurukshetra University, Kurukshetra, India N. E. Udoye Mechanical Engineering Department, Covenant University, Ota, Ogun State, Nigeria P. Ujwal Department of Biotechnology Engineering, N.M.A.M. Institute of Technology (Visvesvaraya Technological University, Belagavi), Udupi District, Karnataka, India Sonali Upadhyaya Department of Management, Dayalbagh Educational Institute, Agra, India Sreedevi Upadhyayula Department of Chemical Engineering, Indian Institute of Technology Delhi, New Delhi, India Muralidhar Vardhanapu Department of Mechanical Engineering, Birla Institute of Technology & Science, Pilani Hyderabad Campus, Hyderabad, India Swaminathan Venkatraman Department of Mathematics, School of Arts, Science, Humanities and Education, SASTRA Deemed to be University, Thanjavur, India

xxxvi

Contributors

P. Vinod Babu Rajiv Gandhi University of Knowledge Technologies, Nuzvid, India Karthik Warrier Department of Mechatronics Engineering, College of Engineering and Technology, SRM Institute of Science and Technology, Chennai, Tamil Nadu, India Kailas L. Wasewar Advance Separation and Analytical Laboratory (ASAL), Department of Chemical Engineering, Visvesvaraya National Institute of Technology (VNIT), Nagpur, India R. Yameni Department of Mechanical Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, India Jayanthi Yerchuru Department of Mechanical Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, India Muhammad Zafar Biofuel Lab, Department of Plant Sciences, Quaid-i-Azam University, Islamabad, Pakistan

Part I Industry 4.0: Concepts, Themes, and Perspectives

1

Industry 4.0 Revolution: Introduction Paolo Di Sia

Contents The Fourth Industrial Revolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industry 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Expected Benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Costs, Security, and Privacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Enabling Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Advanced Human-Machine Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Additive Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cyber Physical System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Internet of Things . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cloud . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Big Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wearable Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Virtual and Augmented Reality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applicative Sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Home and Building Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Logistic Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Automotive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Urban Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . School . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Healthcare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Professional Figures of Tomorrow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4 4 5 6 7 7 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 16 18 19

P. Di Sia () School of Science, University of Padova, Padova, Italy School of Medicine, Department of Neurosciences, University of Padova, Padova, Italy © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_88

3

4

P. Di Sia

Abstract

Until today, three industrial revolutions have occurred over time; starting from that of the late 1700s (note for the use of the “machine steam”) to the advent of the first computers in factories (1960–1970), they have brought a radical change not only in the production in the industry, but also in the society. Technology has never stopped innovating. This increasing evolution has been and is still sustained by man, who seeks progressively advanced tools that improve their existence. This chapter is devoted to the theme of Industry 4.0 and all aspects related to it; starting by the birth and the meaning of the term, the author illustrates all those technologies that have significantly determined the beginning of the fourth revolution, describing their characteristics, their potential, and some methods of use in the industrial field and social sphere. Furthermore, important topics such as ethics, privacy, and security will be considered, in a reality where all data are shared and saved in remote. Keywords

Industry 4.0 · Evolution · Technology · Science · Informatics · Robotics · Automation · Reality

The Fourth Industrial Revolution The term “Industry 4.0” derives from the European Industry 4.0 initiative, inspired by a project thought and done by the German government. The authorship of the German term “Industrie 4.0” is attributed to Henning Kagermann, Wolf-Dieter Lukas, and Wolfgang Wahlster, who used it for the first time in a communication, held at the 2011 Hannover Fair, in which they announced the “Zukunftsprojekt Industrie 4.0.” Realized at the end of 2013, the project for the industry of future “Industry 4.0” involved investing in infrastructure, school, energy systems, research institutions, and companies for modernizing the German production system and bringing German manufacturing back to world level, making it globally competitive. So, when we talk about Industry 4.0, we refer to the fourth industrial revolution, the one characterizing our days. Unlike the revolutions of XVIII, XIX, and XX centuries, it is not possible to attribute to it a precise period or starting date, but, as a revolution, it is leading to a change on global scale (https://www.hannovermesse.de/ home; https://www.vdi-nachrichten.com/Technik-Gesellschaft/Industrie-40-MitInternet-Dinge-Weg-4-industriellen-Revolution; http://reports.weforum.org/futureof-jobs-2016/).

Industry 4.0 The revolution of the early 1970s is known for the entry of electronics and information technology in the industrial sector, leading to an increase of automation

1 Industry 4.0 Revolution: Introduction

5

levels with consequent raising of production. Industry 4.0, called also the “digital revolution,” is concentrated on all those digital technologies able to increase the general interconnection and cooperation of resources (people and IT systems), with changes affecting the industrial sector and the society in all aspects (Schwab 2017). Data assume a primary role in this process, because they are the basis of any operation, a tool that creates value. It is through data that the calculation power of machines and the motion of today and future economy are determined. The four pillars of this revolution are the following: 1) Data. 2) Analytics, that is, all those analysis operations performed after data collection. 3) Human-Machine Interaction (HMI), that is, those modalities of man-machine interface. 4) Manufacturing, also defined as the “bridge” between the digital and the real. Once data have been collected, processed, and made a usable tool, the last step is to find the tools for producing goods. The common factor is the “interconnection” between multiple elements of a system, high levels of communication, and the optimal exploitation of all connected services. The related “enabling technologies” are divided in two large groups. The first one concerns the set of technologies and services closest to IT (Information Technology): 1a) Cloud: management of large amounts of data on external servers, making the information available to anyone who has the authorization. 1b) Big Data: analysis of a large amount of data in order to optimize products and production processes. 1c) Cyber-Security: security during operations on the network and on open systems. The technologies of the second group are closer to the operational level: 2a) Augmented Reality: It can be applied in any sector; in the industrial sector, for example, it acts as support for production and maintenance processes. 2b) Advanced HMI: They allow the man to interact with the machine or with the system in general. 2c) Additive Manufacturing (AM): It covers all the production of manufactured goods carried out using the 3D printer (Kumar Pabbathi 2018).

Expected Benefits Industry 4.0 is a revolution that gradually affected a growing number of sectors (medicine, industry, education, etc.), slowly increasing their digitization level through the use of modern technologies, and creating an environment in which the processes will be completely automated. Supported by a specific communication

6

P. Di Sia

system, they are able to exchange data with other systems, monitor themselves, and act accordingly. Thus, the presence of intelligent machines and equipments is spreading, bringing more efficiency in their respective application areas. In the industrial sector, means will be made available contributing improvements throughout the production line. New technologies are being introduced in every single step that goes from the processing of raw materials upon the delivery of finished product. Production will be performed at reduced costs, with greater speed and avoiding economic problems caused by machine downtime or errors, everything without negatively affecting the product quality. Robots are programmed to work in close contact with the man; augmented reality, following step by step the operator, will make the assembly, maintenance, and testing phases simpler and safer. In a company, however, to affect productivity and position in market is not just machinery; especially for big companies, management choices can produce positive but also negative effects. For this, certain data analysis algorithms (“Data Mining”) go into help, producing useful results that simplify the activity of the decisionmaker. On social level, the evolution will not be less, with effects changing many aspects of our daily life. Many leisure or business activities have already seen the progress using modern and functional tools. It is sufficient to think of the boom in sales of “wearable devices,” about 19 million in 2014 (https://www.wearable-technologies. com/); they allow to monitor the heart rate, to receive calls and messages simply by connecting to a smartphone. An important role is also assumed by the “embedded systems” that take care of repeatedly performing specific operations defined by a software, respecting real time if necessary. They are increasingly present in every kind of object we use. It is easy also to think of a “smart home,” in which such level of interconnection is reachable that all electronic devices are visible on the network and can be managed remotely by the user via smartphone or tablet (Yáñez 2017; O’Driscoll 2017). Expanding this concept further, we talk about “smart city,” a city managed efficiently in every aspect, in order to ensure a sustainable development and a high quality of life, with development factors like economic activities, environmental resources, relationships between people, mobility, and method of administration (Barlow and Levy-Bencheton 2019; Di Sia 2019a).

Costs, Security, and Privacy The effects produced by an event can have a twofold appearance: favorable and unfavorable. The introduction of robots in factories brings not only a more efficient production, but also an inevitable job reduction. The work of a robot often equates that of 10 men, with the difference that the product will be more precise, less subject to imperfections. This problem is common to the entire industrialized population of the world and is caused not only by an increase in technology but also by the request of the necessary knowledge to take advantage of these technologies.

1 Industry 4.0 Revolution: Introduction

7

The possible loss of job is not the only inconvenience of this revolution. Companies and public bodies must invest increasingly on the protection of cyber physical and IoT systems, since these open the door to security problems. The risks in having entire systems connected to each other through the network are an attack on a weak link to break the whole chain; for this, the protection will become necessary not only for data and infrastructure that contain them, but also for their network. The attention to cyber security must be therefore much higher (Di Sia 2017a; Wells et al. 2014). Studies have shown that most violations are committed by a company supplier or an employee who has access to privileged accounts; in 75% of cases, it is sensitive information taken out of the workplace. There are also employees that for disinterest, habit, or negligence do not care for the data they work on, thus exposing the company to possible intrusions or data loss. To solve the problems related to cyber security in business environment, the following are necessary: Increased knowledge of the company’s processed data, knowing how to distinguish important data to those less important, knowing where they are stored, and knowing methods or software that guarantee a minimum of security. Awareness of the risks coming from outside, from e-mail with viruses, to possible access to own server or archive cloud (Vaidya et al. 2018). Privacy is the tool that has been provided to man to protect their confidentiality. But with the progress, lots of personal data like health status, political ideals, and interests become profitable by every operation that one carries out every day making therefore complex and onerous the control on such data.

Enabling Technologies Industry 4.0, although global, has not been uniformly implemented among the various countries both in the timing and in the choice of investments to be carried out. In any case, it is possible to recognize common elements that initiated the revolution: Internet of Things, Big Data, Robotics, and additive manufacturing, through which companies have the possibility of radically innovating their business model.

Advanced Human-Machine Interface The human-machine interface (HMI) is the device or software that allows man to interface with one or more machines. It is therefore possible to supervise and manage, via a single display or multitouch, the correct operation of any simple or complex system. The software used plays the important role of “translator” that shows, in the most user-friendly way possible, all information it receives, making

8

P. Di Sia

them easier for the user to manage machinery (Gorecky et al. 2014; Bannon 2011). The new solutions provide for remote locations that give the possibility to manage also a complex system at distance. If before the man-machine interfaces were mostly used to manage industrial processes, today they are used practically in any social context. With the Internet of Things all objects can be interfaced with the network, and therefore remotely manageable.

Additive Manufacturing The AM is a new technology used for production of 3D objects; one of the principal reasons of its use is accuracy. In some sectors, such as “prototyping” (for example, in the medical field) or “hobby,” it has attained a very high level of precision, at least for now, by milling machines (Gebhardt 2011). The characteristics of the additive printing that decreed its success are the countless varieties of materials and techniques usable: Stereolithography Apparatus (SLA): Stereolithography was one of the first techniques used in AM. It consists of a laser that solidifies portions of liquid resin contained in a tank. The materials used are mostly epoxy resins photosensitive, for the creation of transparent prototypes, and high temperature-resistant ceramics (300 ◦ C) (Bártolo 2014). Multi Jet Modeling (MJM): As in molten deposition, a wax filament is heated and deposited on the platform of construction. Before depositing the next layer, the structure is solidified with the use of UV rays (https://3d-labs.de/ mjm/; https://www.3dprintingmaterialsconference.com/3d-printing-materials/acomparison-of-transparent-am-materials/). Fused Deposition Modeling (FDM): It is the most known and used technique; it consists in depositing material using a nozzle. The extrusion nozzle is heated to the point of get the material malleable but still allows it to cool down in a short time, thus allowing a horizontal and vertical development of the artifact (Attoye et al. 2019). Selective Laser Sintering (SLS): a very popular technique, with which it is possible to create 3D objects of plastic, glass, ceramic, nylon, and metals. It uses a laser that welds the particles together of the chosen material (https://laseroflove.files. wordpress.com/2009/10/dmls_history.pdf). Color Jet Printing (CJP): This is a professional 3D metal-printing technique in which the objects are formed by gluing the metal particles and then sintering them (or melting them). The negative aspects are the need for a separate cooking phase in a special furnace and high costs; it is therefore indicated for artifacts of consistent size (https://www.additive-3d.com/Materials-finishes-Colorjet-3dPrinting.html). Directed Energy Deposition (DED) uses a laser to dissolve the dust that is slowly released and deposited by a robotic arm to form the layers of an object (https:// www.sciencedirect.com/topics/materials-science/directed-energy-deposition).

1 Industry 4.0 Revolution: Introduction

9

The potential of additive printing is endless. In the medical field, this technology does not only limit itself to prostheses, but also reproduces rigid body structures such as bones and teeth. The research aims to achieve artificially the soft tissues of the body making them functional (Zhou and Bhaduri 2019). In the engineering field, the aviation sector has already started projects for the creation of engine parts (nozzles, motor blades, and turbines) for aircrafts using materials such as titanium, aluminum, and nickel-chromium (Jiménez et al. 2019).

Cyber Physical System The Cyber Physical System is perhaps the computer system that has determined the beginning of Industry 4.0 because it has been one of the first systems able to exchange information, in a continuous way, between the physical and the virtual world. CPSs are composed of different parts working together to carry out certain operations. These systems must comply control, communication, and computational capacity; they are software, communication technology, sensors, and actuators for detection and measurement of phenomena in the real world. Many are already used in most varied fields: in medical devices, in environmental and critical infrastructure control systems, in conservation and energy distribution systems, and in smart facilities in general. The potential and usefulness of CPSs increase if they are interfaced with the Internet (https://www.cybersecurityintelligence.com/cpse-labs3225.html; Kim et al. 2018).

Internet of Things Internet of Things (IoT) defines the set of smart objects that, in addition to perform actions of geolocation, processing, acquisition, and identification, are able to interface with the world of the web and consequently access to all services they can offer. Every object, like a CPS, is identified by an IP address or by special labels RFID (Radio-Frequency IDentification), QRcode, or NFC (Near Field Communication) (Di Sia 2016; Di Sia 2019b). With this technology, the network is used as a means to transmit data to remote servers or other devices, also interfaced with the Internet. An object classifiable as IoT can communicate with one or more devices as appropriate. Let us talk then of a connection: One to One: It is a more basic and simple functionality than a device IoT can perform, that is, direct communication with a second device. This is the case of a car equipped with instruments of self-diagnostics; it is able to send information to mechanic’s computer. One to Many: We have larger data transmission; there is a center that instructs many receiving sensors, based on information which in turn had been received and processed by them, as some car manufacturers who improve the efficiency of

10

P. Di Sia

cars, already in circulation, based on data collected by them (https://blogmitcnc. org/2014/08/21/the-tesla-iot-car-case-study/). Many to Many: It is the most complete form of IoT, in which millions of sensors communicate with millions of devices creating a dense network of information capable of managing independently whole activities (Srinidhi et al. 2019). IoT is therefore an evolution that extends the Internet to real objects and places. More and more sectors (home automation, robotics, avionics, automotive and biomedical industry, and telemetry) use an increasing number of devices, connected to the Internet, for monitoring and then carry out consequent actions. For example, in urban planning, street lamps can be equipped with sensors that indicate if the lamp works and detects air quality by informing in an appropriate way the reference body (Rathore et al. 2016).

Cloud The cloud is a service provided by a supplier, such as companies or external providers, which allows any authorized customer to share, store, or process data. A cloud service must respect certain characteristics, among which global accessibility is “in primis.” It constitutes the distinctive character of clouding, that is, the access to information from any terminal and from anywhere in the world (Siebel 2019). Its use, with the integration of the IoT in society, is further expanding. Many devices can request a too large memory to be incorporated into some object, or they need information processed by other microcontrollers; a storage cloud is then used that can contain hundreds of TeraBytes of data and render them accessible to anyone with the authorization (Orban et al. 2017).

Big Data Nowadays, any object connected to the network and any service, online or not, produces data. The flow of information is then enormous. The term “Big Data” refers not only to the actual quantity of data, but also to its analysis. A definition of Big Data is based on three concepts: Volume: the need to determine the relevant data within a huge amount of data. Speed: With a data flow that now travels at speed without precedents, it is necessary to process data fast enough. Variety: Organizations must confront each other with the huge variety of existing data (numeric files, files of text, audio, video, data from quotations on the stock exchange, etc.). Any type of information can be extracted from the data analysis, useful to individuals or companies, and it is performed through “data mining,” a

1 Industry 4.0 Revolution: Introduction

11

knowledge extraction process from large databases. Data mining uses appropriate algorithms and techniques such as “grid computing,” “in-database processing,” and “in-memory analytics,” making information available and immediately usable in the context of decision-making (Marz and Warren 2015; https:// www.britannica.com/technology/data-mining; http://www.gridforum.org/; https:// searchbusinessanalytics.techtarget.com/definition/in-database-analytics; https:// www.intel.in/content/www/in/en/analytics/in-memory-data-and-analytics.html).

Machine Learning With Machine Learning, we go beyond simple automation and enter in the field of artificial intelligence. Machine Learning is similar to data mining; both analyze large amounts of data from which to extract useful information. What differentiates these specific analysis techniques are the users: The information obtained with data mining is exploited in order to make improvements to some operations, while those obtained through Machine Learning are used by machines. So an object becomes intelligent not only because it sends and receives data, but also because it is able to learn from them without the intervention of man, producing more precise and reliable results. The use of an intelligent system allows to reach, in every sector, goals otherwise difficult to achieve (Burkov 2019; Géron 2017).

Wearable Devices Wearable devices are devices or gadgets to be wearable by people and animals. They have the ability to interact with other devices and to interface with the Internet. They may be able to save information on the web by connecting, via Bluetooth or Wi-Fi, to a smartphone. Wearable devices have immediately caught the attention of man. The best known and most popular are fitness trackers with the main function of pedometer and sports watches that, in addition to the pedometer, can detect heart rate, position, watch, and much more depending on the model. There are “smart clothing,” apparently normal clothing but with special technologies that make them more functional, for example, HIGH-TECH clothing for monitoring the vitality parameters of newborns, head-mounted displays for virtual reality, smart jewels, implantable devices, and many others (https://www.lifewire.com/what-are-smartclothes-4176103). These types of devices have been grouped into some macrocategories defining their characteristics (https://www.iop.org/explore-physics; Tong 2018): Complex accessories: To be fully operational, they require the connection to another device. Smart accessories: Slightly more autonomous, they connect to the network and can perform some functions without the support of other devices.

12

P. Di Sia

Smart wearables: They work in complete autonomy, able not only to connect but also to perform complex actions, such as browsing or downloading.

Robotics Robotics is an interdisciplinary science requiring the involvement of several disciplines: informatics, physics, mathematics, psychology, linguistics, automation, mechanics, and biology. It is devoted to the design and construction of automated systems for helping or fully replacing man in their duties. As far as automation is concerned, it has become a reality that involves us in our daily life; by the primary sector of the industry with robotized arms, we arrive to: Humanoid robotics: one of the most fascinating research, creating robots with human features, equipped with artificial intelligence and able to act autonomously with the help of servomotors and video cameras. Service robotics: It deals not only with robots that perform useful services for human beings, for example, carer robots for the elderly, rescue robots, and domestic robots that clean and cook, but also exoskeletons and robotic prostheses for rehabilitation postillness or accidents. Robotics for surgery and medical telepresence: Robots with thin remote-controlled arms allow minimally invasive interventions to heart, prostate, uterus, and lungs. As for medical telepresence, possible solutions allow the specialist to visit remotely the patients. Educational robotics: It introduces children and young people in the world of robotics. Many schools in the world have already integrated this discipline in their course of study (Lynch and Park 2017; https://www.britannica.com/ technology/robotics; Mckinnon 2016).

Virtual and Augmented Reality The market offers today tools that allow users to see virtual objects while they maintain a link with the real world. This operation is carried out by the virtual reality and the augmented reality. “Virtual reality” is a simulation or reconstruction, generated by a computer, of life, environments, or real situations, capable of immersing the user in a nonreal world, involving them totally by stimulating the sense of sight and hearing. At first, virtual reality was used to simulate some jobs, giving the possibility to train and practice. Then, wearable instruments were widespread, using them for all kinds of entertainment such as video games, videos, and navigation on the web. Virtual reality involves uses that can affect all sectors (Rubin 2018; https://www. marxentlabs.com/what-is-virtual-reality/).

1 Industry 4.0 Revolution: Introduction

13

HTC Vive is one of the most advanced tool (for sale to the public) regarding virtual reality; using laser environmental sensors, it allows the user to “enter” the virtual world even with their own body. They will be able to sit, stand up, or run; it is equipped with camera and proximity sensors for detecting the presence of obstacles and does not completely detach the person from the real world as it is able to receive notifications, calls, or messages (https://www.realmore.net/en/htc-vive/). “Augmented reality” is a technology that adds to already existing objects digital levels generated by a computer and with which one can interact, images, information, or instructions that the user can use for their own purposes. In the web, there are applications supported on this concept; some apps, for example, are provided by retailers of furniture items and give the possibility to the buyer to see in real time the arrangement and the style of the future purchase. There are applications dedicated to the consultation of catalogs, maps with navigator, web portals for travels, and restaurants and shops reviews (Pangilinan et al. 2019). As for virtual reality, the tools allowing a person to take advantage of augmented reality are many and differ in features (image quality, accessory tools, and memory) and for portability. The most common is the smartphone, as it needs only a camera (now present in all models) and an application, and smart glasses (https://www. augment.com/blog/virtual-reality-vs-augmented-reality/; https://www.st.com/en/ applications/virtual-augmented-reality.html).

Applicative Sectors Smart Home and Building Automation “Smart Home” is one of the most known sectors because it arouses great interest among people. It is a field that deals with the introduction of modern technologies supporting the objects already existing in the home environment. As “objects,” we refer to those electronic devices falling into the home appliance category such as oven, washing machine, computer, printers, vacuum cleaner, and many others of daily use. All these devices, otherwise independent of each other, are interfaced with the Internet, so becoming visible to all users and systems that can access the server. The concept of “visibility” brings a series of benefits that go from wireless control of entire systems or individual objects to management automated of lighting systems, security systems, TVs, and other appliances that turn on and off according to the presence of persons in the room or if they are not used for a specific period. The resources needed to transform a home into an “intelligent home” are directly proportional to the size of the building as well as the number of devices to interface to the network (Di Sia 2020a). If the idea is to take advantage of new technologies (also of nanotechnologies) on a greater number of environments or on companies, we are in the field of “building automation”; this sector has the same objectives: better livability, greater safety, and

14

P. Di Sia

energy savings. The difference is that they are applied on a larger scale as large industries, offices, or skyscrapers (Di Sia 2014a, 2015a).

Logistic Service Logistics is a very broad sector in terms of application fields, covering many areas that also belong to different realities. Logistics is present in telecommunications and management of the overhead lines (support logistics), but the best known is the logistics sector tied to the productive one and to all operations connected to it (Curley and Salmelin 2017). Operations such as the physical management of materials flow from manufacturer to consumer or the coordination of development activities of projects and systems are being supported and reinforced by intense lines of communication. Any object will be able to transmit and receive real-time information. Cloud systems will allow this, capable of supporting large amounts of data and algorithms able to process them quickly. Many processes will be automated, and companies will be able to see in real time any kind of information concerning own vehicles and own goods (Batalla et al. 2016).

Automotive The Automotive sector is constantly evolving and regroups all integrated systems (ABS, cruise control, and distance sensors) in a transport means that facilitate and make safer displacements. The “parking sensors” are an example: starting with the use of simple cameras or proximity sensors to detect obstacles, up to “Park Assist” (ultrasonic-sensors technology that detects the suitable parking for vehicle, executing then the maneuver) and “Trained Parking” (the car is able to get in a database the favorite parking areas of the driver, selecting the nearest parking to the current position, in automated way and also without the presence of the driver on board, who can follow everything from their smartphone) (https://www2. deloitte.com/us/en/insights/industry/automotive/industry-4-0-future-of-automotiveindustry.html; Di Sia 2017b; https://www.continental-automotive.com/en-gl/Landing-Pages/CAD/Automated-Driving/Driving-Functions/Trained-Parking; https:// www.bosch-mobility-solutions.com/en/highlights/connected-mobility/connectedand-automated-parking/). Trained parking is the first step that will lead to the production of completely autonomous cars, also called “driverless cars.” Moreover, new systems have been introduced that improve existing ones such as ABS (assisted braking) and ESP (antiheeling) or simply improve aspects that so far had not been subject of research, such as new versions of high-beam headlights, real-time communication between two or more cars and emergency vehicles (https:// www.techradar.com/news/self-driving-cars; https://www.technologyreview.com/s/ 520431/driverless-cars-are-further-away-than-you-think/; Taiebat et al. 2018).

1 Industry 4.0 Revolution: Introduction

15

Urban Planning The field of urban planning mainly involves an increase in the grade of livability in small and large cities as well as energy savings, through the use of alternative energy sources and less polluting or toxic materials (Di Sia 2014b, 2018). Also in this sector, embedded technologies and nanotechnology help, carrying out measurements and data processing: street lamps that, if not working, notify the reference entity, or that contain systems for the detection of air quality, equipped also with sensors to assess weather and Wi-Fi repeaters (“multifunction” street lamps) (https://multipole.com.au/). They can collect sensitive data for managing the city more efficiently and will save around 60% in terms of energy consumption. As they are connected to the Internet, they own an IP address that identifies them, giving the possibility to an operator of remotely adjusting the light intensity according to the need. Another area in which the urban sector will intervene is the traffic control: The systems dedicated to the roads management in the coming years will undergo improvements making them capable of actively communicating with automotive products. Installations will be arranged on the streets of the urban center and through sensors will produce a set of data describing volume and speed of traffic in a given place (Elhoseny and Hassanien 2019).

School The new school is an environment completely immersed in the future, in a reality largely dependent on technology, to which students will have to approach by a young age. School innovation means “digitalization” and building redevelopment. In this regard, many schools are already moved, introducing electronic register, new subjects and modernizing teaching methods, the use of “Oculus Rift” for virtual reality, and robots (Di Sia 2020b; https://www.oculus.com/?locale=it_IT; Gleasure and Feller 2016). Other goals are security and modernization level of environments (classrooms, laboratories, canteens, etc.), everything respecting the environment. There will be more focus on new technologies: Technology courses and laboratories, bringing young people closer to Industry 4.0. Computational thinking: Already from school primary, children will have the opportunity to test themselves in theories and programming techniques, the socalled “coding.” Technical institutes will see the entry of programs to teach to students using increasingly sophisticated machines and what characteristics the worker of the future must have. New specialized courses and research doctorates on Industry 4.0, including a percentage with a particular focus on big data (Lehmann and Chase 2015).

16

P. Di Sia

Healthcare Important changes are taking place in the health sector with the birth of the new concept of hospital. HealthCare devices will allow to diagnose pathologies, to treat and prevent diseases, increasing the level of precision, decreasing the bulk, and having the ability to interface with other devices for an increase in features such as: To show clear and specific information. To improve the patient’s state of well-being during a treatment, reducing the level of stress (Zaleski 2015). The “Hospital 4.0” is a technological environment, in which man and machine will work at close contact, for offering an optimal and active service 24/24 hours. “Medical devices” perform some functions: administration of drugs, blood pressure detection, heart rate and sleep quality detection, with the use of a new set of devices, “Health Watches,” working through apps and relying on digital platforms, and using also the “telemedicine” that offers assistance to patients from remote. Many instruments are already in use in some hospitals: operating room robots, driverless trolleys for transporting goods or meals, and automated pharmaceutical cabinets (https://searchhealthit.techtarget.com/definition/telemedicine).

Agriculture Agriculture is a strongly growing sector having as objective more production, to be able to guarantee food for everyone, without consuming more land and using less water, seeds, treatments, and energy. All this would be possible thanks to the use of drones, satellites, agricultural driveless vehicles (HIGH-TECH tractors), rearrangement of water system to minimize waste, and georeferencing work of land for determining levels of humidity, pH, phosphorus, nitrogen, and potassium (https://www.cnbc.com/2016/09/16/future-of-farming-driverless-tractors-agrobots.html). All collected data will be then crossed with surveys and satellite photos, and processed by software that distribute them to tractors. This can lead to an increase of more than 20%, with a consequent reduction in the use of fertilizers by more than 15% (http://theconversation.com/farmers-of-the-future-will-utilizedrones-robots-and-gps-37739).

Professional Figures of Tomorrow In Industry 4.0, data and related systems have great importance; with all these produced data, man will be helped in making decisions, will be recommended in their purchases, and will benefit of optimal health and education system. Industrial

1 Industry 4.0 Revolution: Introduction

17

machinery will be safer and more efficient. To make all this true, it is necessary to analyze data by creating algorithms able to process increasing amount of information and to extract always more precise and reliable results. The working figures of the future will be therefore largely based on “analysis” and “processing.” Important figures will be the following: Regulatory affairs : She/he deals mainly with the supervision of authorization procedures for the products marketing of some companies, as pharmaceutical ones (responsible for research and production of drugs, veterinary products, pesticides, parapharmaceuticals, cosmetics, etc.), that want to know if their products are online with current safety regulations. Business analyst : It is an important figure, because it has the responsibility to contribute to the implementation of companies strategies through design and monitoring of the execution and implementation of tools and technologies already available within the company or within the new ones available on the market. The main tasks are as follows: i) Preparation of the business plan. ii) Identification of the improvement areas in the processes of business, providing ICT solutions with realization of requirements, specifications, and processes related to the proposed solutions. iii) Analysis of information and available documents. HSE specialist – Health and Safety Executive: This is a professional figure much in demand as it takes care of maintenance of quality certifications; she/he provides coordination of all activities related to prevention, protection, and safety at work within the company, in accordance with current legislation, defining the plans of ordinary and extraordinary maintenance. Designer engineer : This is the union of two existing figures: the designer, endowed with creativity, and the engineer, specialized in disciplines such as electronics, mechanics, chemistry, or architecture. This figure is dedicated to the production of prototypes (also with the aid of CAD) and products that can supply solutions in the social field and that meet the needs of the customer. Connectivity and cyber security specialist: With the Internet of Things, more and more objects can communicate and interact through the network; therefore, a higher level of security is required. This specialist is not dedicated only to the introduction of sophisticated techniques that prevent data stealing, but is also interested in the data ensuring integrity and checking for possible losses. Significant data losses can generate an incorrect analysis. Business intelligent analyst: Business intelligence is the process that transforms all kinds of data into “knowledge” then used in decision-making processes. The analyst applies methods, models, and analysis processes to a large amount of data, obtaining precise and detailed information, in companies that must everyday compare numerous data collected from their machinery, products,

18

P. Di Sia

means, and employees, and allows the company to obtain a high return of investment, lower costs/risks, and timely decisions. Data scientist and data specialist: The data scientist, expert in statistics, programming and data mining, and data visualization, produces models and performs analysis and research on the data, in search of notions that can influence in the matter of business. The data specialist processes the data for the purpose to produce useful tools in decision-making. It is not the same job; there are differences between these two tasks: i) The data scientist asks questions that would help to make everything related to business more efficient and then proposing possible solutions, while a data analyst gives solutions to questions asked by a business team. ii) Both roles are dedicated to write queries, working with a team of engineers to search for correct data, for data munging, and to obtain information from the data. A data analyst does not construct statistical models neither works with Machine Learning or advanced programming techniques, but instead on business intelligence with appropriate tool/packages. iii) The data scientist has a good ability about data visualization and the ability to convert data in a business story (https://www2.deloitte.com/insights/us/ en/focus/industry-4-0/overview.html; http://www.globalskillsummit.com/ whitepaper-summary.pdf; Botthof and Hartmann 2014).

Conclusions Revolutions like this one are global and tend, at least from a theoretical point of view, toward a general improvement of the conditions of man and the environment. This involves seriousness and fairness in the work, for the sake of science and technology, thus trying to avoid the entry into play of political interactions that could decrease the goodness of the work in favor of clientelism and power games. The ultimate goal is and should always remain the good of people and the improvement of everyone’s daily life. It is therefore necessary to reflect on the advent of digitization and on the effects, both not only positive but also potentially negative, that it brings into company and in the productive sector, as in the nanotechnology sector, well connected to it (Di Sia 2014c, 2017c). Very important is in every case the contribution of theoretical disciplines (Di Sia 2015b, 2019c, 2021a, b). Questions involving ethics emerge that should not to be set aside, as the advent of nanotechnology. An example among many others concerns the unmanned car: Who will be responsible for damages that the car could do? What about the insurance of such a car? Various solutions have been studied and are under study. A line says that consumer protection is not linked to robotics, but to whom to sell the robot product? Therefore, all sectors should be secured, as well as factories involved in the construction and sale of the particular product; in the case of a driverless car,

1 Industry 4.0 Revolution: Introduction

19

therefore, it is necessary not only to insure passengers, but also builders, software vendors, connectivity providers, and all persons participating in the circulation of the vehicle. We must start from the assumption that technology in its essence is neutral: It is the way in which we use it that makes the difference, for good or for bad. There is therefore the need to have a governance that takes into consideration all problems resulting from the use of advanced technologies and that protects us all from possible general bad uses and from possible further deterioration of the environment and discrimination of the social conditions.

References Attoye S, Malekipour E, El-Mounayri H (2019) Correlation between process parameters and mechanical properties in parts printed by the fused deposition modeling process. In: Kramer S, Jordan J, Jin H, Carroll J, Beese A (eds) Mechanics of additive and advanced manufacturing (vol 8), Conference proceedings of the society for experimental mechanics series. Springer, Cham Bannon L (2011) Reimagining HCI: toward a more human-centered perspective. ACM Interact 18(4):50–57 Barlow M, Levy-Bencheton C (2019) Smart cities, smart future: showcasing tomorrow, 1st edn. Wiley, Hoboken Bártolo PJ (ed) (2014) Stereolithography: materials, processes and applications. Springer, New York Batalla JM, Mastorakis G, Mavromoustakis CX, Pallis E (eds) (2016) Beyond the Internet of Things: everything interconnected, 1st edn. Springer, Berlin, 2017 edition Botthof A, Hartmann EA (eds) (2014) Zukunft der Arbeit in Industrie 4.0. Springer Vieweg, Berlin. 2015 edition Burkov A (2019) The hundred-page machine learning book. Andriy Burkov Curley M, Salmelin B (2017) Open innovation 2.0: the new mode of digital innovation for prosperity and sustainability (innovation, technology, and knowledge management), 1st edn. Springer, Berlin, 2018 edition. https://www.amazon.it/Hundred-Page-Machine-Learning-Book/ dp/199957950X Di Sia P (2014a) Nanobiomaterials for environmental protection: state of the art, applications and modelling. Int J Eng Innov Res (IJEIR) 3(5):688–693 Di Sia P (2014b) Characteristics in diffusion for high-efficiency photovoltaics nanomaterials: an interesting analysis. J Green Sci Technol (ASP) 1:1–4 Di Sia P (2014c) Present and future of nanotechnologies: peculiarities, phenomenology, theoretical modelling, perspectives. Rev Theor Sci (ASP) 2(2):146–180. https://www.ingentaconnect.com/ content/asp/rits/2014/00000002/00000002/art00004;jsessionid=3t67e6ndkrsa0.x-ic-live-01 Di Sia P (2015a) Present and future of Nano-Bio-Technology: innovation, evolution of science, social impact. Online J Educ Technol (TOJET), Special Issue 2 for INTE 2015, 442–449 Di Sia P (2015b) A new analytical transport model for (nano)physics. Int Res J Eng Technol (IRJET) 2(7):1–4 Di Sia P (2016) Internet and new educational perspectives. Int Aced J E-Methodol 3:18–24. https:/ /jecs.pl/index.php/e-methodology/article/view/10.15503.emet2016.18.24 Di Sia P (2017a) Looking at the quantum Internet. Int Acad J e-Methodol 4:31–35. https://jecs.pl/ index.php/e-methodology/article/view/10.15503emet2017.31.35 Di Sia P (2017b) On the use of hydrogen in the automotive sector, nano energy systems. One Central Press 1(2):50–58. https://doi.org/10.24274/nes.2016

20

P. Di Sia

Di Sia P (2017c) Nanotechnologies among innovation, health and risks. Procedia – Soc Behav Sci J 237:1076–1080. http://www.onecentralpress.com/on-the-use-of-hydrogen-in-the-automotivesector-2/ Di Sia P (2018) Nanotoxicology and human health. World Sci News (WSN) 100:86–98 Di Sia P (2019a) Green composites for construction. In: Inamuddin (ed) Green composites 2: industrial applications. Elsevier, in press Di Sia P (2019b) Agri-food sector, biological systems and nanomaterials. In: Molina G, Inamuddin, Pelissari FM, Asiri AM (eds) Food applications of nanotechnology, 1st edn. CRC Press. https://www.routledge.com/Food-Applications-of-Nanotechnology/Molinanamuddin-Pelissari-Asiri/p/book/9780815383819# Di Sia P (2019c) Mathematics and physics for nanotechnology – technical tools and modelling. Pan Stanford Publishing. https://www.routledge.com/Mathematics-and-Physics-forNanotechnology-Technical-Tools-and-Modelling/Sia/p/book/9789814800020 Di Sia P (2020a) Nanotechnologies and advanced smart materials: the case of architecture and civil engineering. In: Hussain CM (ed) The ELSI handbook of nanotechnology: risk, safety, ESLI and commercialization. Wiley-Scrivener Publishers. https://doi.org/10.1002/9781119592990.ch4. https://onlinelibrary.wiley.com/doi/abs/10.1002/9781119592990.ch4. Chapter 4, pp 67–87 Di Sia P (2020b) Looking at the new digital school: didactics 2.0 and school 2.0. Int Acad J E-Methodol (PSP – Pro Scientia Publica) 6:9–20. https://e-methodology.eu/index.php/emethodology/article/view/1189 Di Sia P (2021a) Fourth industrial revolution (4IR) and functionalized MNPs. In: Hussain CM (ed) Analytical applications of functionalized magnetic nanoparticles. Elsevier Publishing (Royal Society of Chemistry) Chapter 19, pp 489–504. https://pubs.rsc.org/en/content/chapter/ bk9781839162107-00489/978-1-83916-210-7 Di Sia P (2021b) On risks and benefits of nanotechnology: the case of medical applications. In: Hussain CM, Kumar V (eds) Handbook of functionalized nanomaterials: environmental health and safety. Elsevier Publishing Chapter 9, pp 235–250. https://www.elsevier.com/books/ handbook-of-functionalized-nanomaterials/hussain/978-0-12-822415-1 Elhoseny M, Hassanien AE (eds) (2019) Emerging technologies for connected internet of vehicles and intelligent transportation system networks, 1st edn. Springer, Berlin, 2020 edition Gebhardt A (2011) Understanding additive manufacturing. Hanser Publishers, Munich. https:// doi.org/10.3139/9783446431621.fm Géron A (2017) Hands-on machine learning with Scikit-Learn and Tensor Flow: concepts, tools, and techniques to build intelligent systems, 1st edn. O’Reilly Media, Sebastopol Gleasure R, Feller J (2016) A rift in the ground: theorizing the evolution of anchor values in crowdfunding communities through the oculus rift case study. J Assoc Inf Syst 17(10):708–736 Gorecky D, Schmitt M, Loskyll M, Zühlke D (2014) Human-machine-interaction in the industry 4.0 era. In: 12th IEEE International Conference on Industrial Informatics (INDIN), Porto Alegre, Brazil. https://doi.org/10.1109/INDIN.2014.6945523 http://reports.weforum.org/future-of-jobs-2016/. Retrieved 23 Sept 2020 http://theconversation.com/farmers-of-the-future-will-utilize-drones-robots-and-gps-37739 . Retrieved 23 Sept 2020 http://www.globalskillsummit.com/whitepaper-summary.pdf. Retrieved 23 Sept 2020 http://www.gridforum.org/. Retrieved 23 Sept 2020 https://3d-labs.de/mjm/. Retrieved 23 Sept 2020 https://blogmitcnc.org/2014/08/21/the-tesla-iot-car-case-study/. Retrieved 23 Sept 2020 https://laseroflove.files.wordpress.com/2009/10/dmls_history.pdf. Retrieved 23 Sept 2020 https://multipole.com.au/. Retrieved 23 Jul 2019 https://searchbusinessanalytics.techtarget.com/definition/in-database-analytics. Retrieved 23 Sept 2020 https://searchhealthit.techtarget.com/definition/telemedicine. Retrieved 23 Sept 2020 https://www.3dprintingmaterialsconference.com/3d-printing-materials/a-comparison-of-transparentam-materials/. Retrieved 23 Sept 2020

1 Industry 4.0 Revolution: Introduction

21

https://www.additive-3d.com/Materials-finishes-Colorjet-3d-Printing.html. Retrieved 23 Sept 2020 https://www.augment.com/blog/virtual-reality-vs-augmented-reality/. Retrieved 23 Sept 2020 https://www.bosch-mobility-solutions.com/en/highlights/connected-mobility/connected-and-automated-parking/. Retrieved 23 Sept 2020 https://www.britannica.com/technology/data-mining. Retrieved 23 Sept 2020 https://www.britannica.com/technology/robotics. Retrieved 23 Sept 2020 https://www.cnbc.com/2016/09/16/future-of-farming-driverless-tractors-ag-robots.html. Retrieved 23 Sept 2020 https://www.continental-automotive.com/en-gl/Landing-Pages/CAD/Automated-Driving/DrivingFunctions/Trained-Parking. Retrieved 23 Sept 2020 https://www.cybersecurityintelligence.com/cpse-labs-3225.html. Retrieved 23 Sept 2020 https://www.hannovermesse.de/home. Retrieved 23 Sept 2020 https://www.intel.in/content/www/in/en/analytics/in-memory-data-and-analytics.html. Retrieved 23 Sept 2020 https://www.iop.org/explore-physics. Retrieved 23 Sept 2020 https://www.lifewire.com/what-are-smart-clothes-4176103. Retrieved 23 Sept 2020 https://www.marxentlabs.com/what-is-virtual-reality/. Retrieved 23 Sept 2020 https://www.oculus.com/?locale=it_IT. Retrieved 23 Sept 2020 https://www.realmore.net/en/htc-vive/. Retrieved 23 Sept 2020 https://www.sciencedirect.com/topics/materials-science/directed-energy-deposition. Retrieved 23 Sept 2020 https://www.st.com/en/applications/virtual-augmented-reality.html. Retrieved 23 Sept 2020 https://www.technologyreview.com/s/520431/driverless-cars-are-further-away-than-you-think/. Retrieved 23 Sept 2020 https://www.techradar.com/news/self-driving-cars. Retrieved 23 Sept 2020 https://www.vdi-nachrichten.com/Technik-Gesellschaft/Industrie-40-Mit-Internet-Dinge-Weg-4industriellen-Revolution. Retrieved 23 Sept 2020 https://www.wearable-technologies.com/. Retrieved 23 Sept 2020 https://www2.deloitte.com/insights/us/en/focus/industry-4-0/overview.html. Retrieved 23 Sept 2020 https://www2.deloitte.com/us/en/insights/industry/automotive/industry-4-0-future-of-automotiveindustry.html. Retrieved 23 Sept 2020 Jiménez M, Romero L, Domínguez IA, del Mar Espinosa M, Domínguez M (2019) Additive manufacturing technologies: an overview about 3D printing methods and future prospects. Complexity:9656938. https://doi.org/10.1155/2019/9656938 Kim NY, Rathore S, Ryu JH, Park JH, Park JH (2018) A survey on cyber physical system security for IoT: issues, challenges, threats, solutions. J Inf Process Syst 14(6):1361–1384. https:// doi.org/10.3745/JIPS.03.0105 Kumar Pabbathi K (2018) Quick start guide to industry 4.0: one-stop reference guide for industry 4.0. CreateSpace Independent Publishing Platform. https://www.amazon.it/Quick-Start-GuideIndustry-4-0/dp/1718978618 Lehmann C, Chase Z (2015) Building school 2.0: how to create the schools we need, 1st edn. Jossey-Bass, San Francisco Lynch KM, Park FC (2017) Modern robotics: mechanics, planning, and control, 1st edn. Cambridge University Press, Cambridge Marz N, Warren J (2015) Big Data: principles and best practices of scalable realtime data systems, 1st edn. Manning Publications, New York Mckinnon P (2016) Robotics: everything you need to know about robotics from beginner to expert. CreateSpace Independent Publishing Platform. https://www.amazon.com/Robotics-EverythingAbout-Beginner-Expert/dp/1523731516 O’Driscoll G (2017) Smart home automation essential guides – the complete series. Independently Published

22

P. Di Sia

Orban S, Jassy A, Cockcroft A, Schwartz M (2017) Ahead in the cloud: best practices for navigating the future of enterprise IT, AWS. CreateSpace Independent Publishing Platform, North Charleston Pangilinan E, Lukas S, Mohan V (2019) Creating augmented and virtual realities: theory and practice for next-generation spatial computing, 1st edn. O’Reilly Media, Sebastopol Rathore MM, Ahmad A, Paul A, Rho S (2016) Urban planning and building smart cities based on the Internet of Things using Big Data analytics. Comput Netw 101(4):63–80. https://doi.org/ 10.1016/j.comnet.2015.12.023 Rubin P (2018) Future presence: how virtual reality is changing human connection, intimacy, and the limits of ordinary life. HarperOne, New York Schwab K (2017) The fourth industrial revolution. Currency, New York Siebel TM (2019) Digital transformation: survive and thrive in an era of mass extinction. RosettaBooks, New York Srinidhi NN, Dilip Kumar SM, Venugopal KR (2019) Network optimizations in the Internet of Things: a review. Eng Sci Technol Int J 22(1):1–21. https://doi.org/10.1016/j.jestch.2018.09.003 Taiebat M, Brown AL, Safford HR, Qu S, Xu M (2018) A review on energy, environmental, and sustainability implications of connected and automated vehicles. Environ Sci Technol 52(20):11449–11465. https://doi.org/10.1021/acs.est.8b00127 Tong R (2018) Wearable technology in medicine and health care, 1st edn. Academic Press, London Vaidya S, Ambad P, Bhosle S (2018) Industry 4.0 – a glimpse. Procedia Manuf 20:233–238. https:/ /doi.org/10.1016/j.promfg.2018.02.034 Wells LJ, Camelio JA, Williams CB, White J (2014) Cyber-physical security challenges in manufacturing systems. Manuf Lett 2(2):74–77. https://doi.org/10.1016/j.mfglet.2014.01.005 Yáñez F (2017) The goal is industry 4.0: technologies and trends of the fourth industrial revolution. Independently Published Zaleski J (2015) Connected medical devices: integrating patient care data in healthcare systems, 1st edn. HIMSS Publishing, Boca Raton Zhou H, Bhaduri SB (2019) 3D printing in the research and development of medical devices In: Biomaterials in translational medicine – a biomaterials approach, Woodhead Publishing Series in Biomaterials, pp 269–289. https://doi.org/10.1016/B978-0-12-813477-1.00012-8

2

Industry 4.0 Perspectives: Global Trends and Future Developments Antonella Petrillo and Fabio De Felice

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Evolution of Industry: The Digital Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Digital Economy and Society Index (DESI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Regional Innovation Scoreboard (RIS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Report Digital 2020 (We Are Social) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . State of the Art on Industry 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Future of Manufacturing: Survey Results and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . Survey Objectives and Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Survey Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges, Directions, and Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Research Direction #1: Technical Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Research Direction #2: Social Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Websites References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

24 26 28 29 30 33 37 37 37 40 41 42 42 43 45

Abstract

All the countries of the world are experiencing a process of business transformation and innovation through the use of digital technologies. COVID-19 has accelerated the digital transformation, but to fully grasp the possibility offered by this crisis, concrete actions are needed to achieve the digital transformation.

A. Petrillo () Department of Engineering, University of Naples “Parthenope”, Naples, Italy e-mail: [email protected] F. De Felice Department of Civil and Mechanical Engineering, Università degli Studi di Cassino e del Lazio Meridionale, Cassino, Italy e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_1

23

24

A. Petrillo and F. De Felice

The effect of new technologies, in the digital first perspective, will lead not only to a more efficient system but above all to relaunch the economy, in particular of some strategic production sectors. The intent of this chapter is to outline global trends and future developments of digitalization in the perspective of Industry 4.0. Thus, a survey and a literature analysis based on structural and conceptual frameworks are developed. The result is the definition of understanding current state of knowledge and to propose future research opportunities in the field of manufacturing digitalization. Keywords

Industry 4.0 · Smart manufacturing · Trends · Perspectives · Challenges Abbreviations COVID-19 CPS DESI I4.0 IoT IR RIS SMLC

Coronavirus Cyber-physical systems Digital Economy and Society Index Industry 4.0 Internet of things Industrial revolution Regional Innovation Scoreboard Smart Manufacturing Leadership Coalition

Introduction An economic system aims to carry out activities to procure the means necessary for the satisfaction of the needs of all stakeholders (Hernandez-de-Menendez et al. 2020; Li et al. 2021). Many systems have undergone and are undergoing profound changes in recent decades due to the advent of automation and digitization (Jeon et al. 2020). This transformation is called Fourth Industrial Revolution or Industry 4.0 or again smart manufacturing. However, we are already beginning to speak, albeit with different meanings, of Industry 5.0. From what we can imagine, the Fifth Industrial Revolution will be characterized by the presence of new technologies which, combining the physical, digital, and biological spheres, will generate consequences in all disciplines and in all economic and productive sectors (Helmann et al. 2020). We will see the integration of technologies already used in the past (big data, cloud, robots, 3D printing, simulation, etc.), further enhanced because they are connected to an intelligent network capable of transmitting digital data to high speed (Hincapié et al. 2020). We are witnessing the birth of a new industrial paradigm that is generating significant changes in the way of conceiving work and man. In this new scenario, empowering people represents a strategic factor both for the quality of goods and services produced and for the efficiency of production systems. The implementation of a new production system will obviously represent a huge change for companies, which already have to face large investments

2 Industry 4.0 Perspectives: Global Trends and Future Developments

25

Fig. 1 Document by years on digitalization. (Source: Scopus, Dec 2020)

today. To remain competitive, one cannot remain indifferent to this epochal change (Culot et al. 2020; Trstenjak et al. 2020). We cannot miss this opportunity. Today, the challenges that companies are facing in an increasingly global market concern the integration of assets and processes; they also require a strategic, flexible, innovative, and future-oriented vision (Ma et al. 2020; Sasiain et al. 2020). The interest on the subject of digitalization in the manufacturing context is demonstrated by the growing number of scientific publications. In fact, a survey pointed out that in 5 years there has been an increase of over 60% in the number of publications as shown in Fig. 1. Understanding the state of the art on innovation means analyzing productive, organizational, and cultural factors from the point of view of the digital revolution (Farooqui et al. 2020; Bougdira et al. 2020). The opportunity for a sudden acceleration was undoubtedly provided by the COVID-19 emergency (Bragazzi 2020). Covid19 is a tragic event, also on an economic and social level, which nevertheless represented a significant change of pace, forcing a march in forced stages, starting from smart working but not only in the direction of a more decisive digitization of activities and processes. The current health emergency COVID19 represents and will undoubtedly represent a strong technological accelerator. In fact, it will be necessary in our opinion: identify new business models, create new and different technological solutions, create a flexible and transparent corporate image, and provide new services and products in a globalized way but characterized by shorter supply chains. These observations motivated us to examine the current state of research in I4.0. Specifically, our intent was to investigate the following questions: What is state of the art in research in I4.0? What are the topics and opportunities for I4.0? Thus, the goal of the present research was to analyze the role

26

A. Petrillo and F. De Felice

that digital technologies play within the production system in the awareness that digital innovation also corresponds to a new theme linked to sustainability. The rest of the chapter is organized as follows: Section “The Evolution of Industry: The Digital Transformation” tries to outline the evolution of industry until digital transformation globally by analyzing European and world data on the degree of technological innovation and the degree of diffusion and use of the Internet and beyond. Section “State of the Art on Industry 4.0” analyzes the state of the art of I4.0 both from a global perception point of view using some data extracted from keyword research tools and using Scopus to provide a more academic view. Section “The Future of Manufacturing: Survey Results and Analysis” presents a proposed survey analysis and results. Section “Challenges, Directions, and Trends” summarizes the main challenges, direction, and trends of I4.0. Finally, Sect. “Conclusion” outlines the main conclusions of the research.

The Evolution of Industry: The Digital Transformation When we talk about digital technologies applied to the manufacturing sector and Industry 4.0 in general, we often also come across the expression smart manufacturing (Shi et al. 2020; Lu et al. 2020). But what is smart manufacturing and how does it differ from I4.0? To answer this question, it is necessary to mention the great foreign digital innovation programs of manufacturing: two of them all – the German one (Industry 4.0) and the American one (Smart Manufacturing Leadership Coalition). History tells us that the German government was the first in the world to define a national strategy to support the digitization of its manufacturing sector, to the point that it is recognized as the paternity of the term Industry 4.0, which appeared for the first time in 2011 (Oláh et al. 2020; Sommer 2015). What is the German smart factory based on? At the basis of this program is the concept of cyber-physical systems (CPS): these systems are composed of thousands of sensors installed directly on the machinery that allow them to be interconnected and lay the foundations for the self-regulation of production systems (Elhoone et al. 2020). Companies that want to follow the principles of Industry 4.0 must move toward the introduction of CPS in factories, of robots that guarantee high production flexibility, also based on the analysis of Manufacturing Big Data collected by themselves. Nearly the same time that the Industry 4.0 program was defined, i.e., in mid2012, the Smart Manufacturing Leadership Coalition (SMLC) was formed in the United States (Bonaccorsi et al. 2020; Corrocher et al. 2020). Unlike the German government program, the American focus is more focused on these principles: cost reduction, sharing of practices and technologies, collective definition of R&D areas, and innovation through collaborative processes (Cioffi et al. 2020). At this point, the question is more legitimate than ever. Are smart manufacturing and Industry 4.0 the same thing? Conceptually, yes because they refer to a common vision: that according to which digital technologies are able to enable the interconnection and cooperation of all the resources used in the factory and along the value chain (O’Donovan et al. 2015). In other words, both terms mean revolutionizing the times and methods of production, bringing technologies and strategies to the

2 Industry 4.0 Perspectives: Global Trends and Future Developments

27

factory (Pinzone et al. 2017). Definitely, the German government was the first in 2011 to define a national strategy (Industry 4.0) to support the digitization of the manufacturing sector, creating an epidemic effect on all other countries (De Felice et al. 2018; Xu 2020). This program, the result of the collaboration of the federal government, technology suppliers, and industry associations with universities and national research centers, has promoted long-term policies for the digitization and innovation of the manufacturing sector (Haipeter 2020; Szabo et al. 2020). The objective is to strengthen German competitiveness. Afterward, all countries developed a national strategy. Denmark (2012) with the Made plan, Belgium (2013) with the Made Different program, England with High Value Manufacturing Catapult, and Holland with Smart Industry (2014) have implemented plans and actions aimed to stimulate innovation and transformation (Denmark and England), to demonstrate the feasibility and sustainability of the factories of the future (Belgium), and to sensitize government, industry, and research centers on the relevance of the issue (Holland). In 2012, the overseas plan (Manufacturing USA) comes to life, thanks to the allocation of 500 million dollars, with the aim of returning to the US production centers of US companies (re-shoring strategy). The plans and Various innovation programs have been promoted around the world, including China (Made in China), Japan (Industrial Value Chain Initiative – IVI), India (Make in India), Canada and Korea (Innovation in Manufacturing 3.0) and the various research and industrial innovation and the forum to which these gave rise (El-kaime and Elhaq 2021; Lopes et al. 2019). Figure 2 summarizes the main national strategies developed in different countries. There is a general aspect that unites all the strategies analyzed: the use of increasingly large amounts of data and information and the increasingly pervasive use of digital technologies to interconnect and cooperate the resources operating

Fig. 2 National strategies on Industry 4.0

28

A. Petrillo and F. De Felice

within the factory and, broadly, throughout the entire value chain. The goal is to optimize the 3Ps: products, processes, and people. To do this in the best possible way, it is necessary to change perspective, exploiting digital to work on new levels of integration and relationship, keeping the customer at the center of development. Innovation in the manufacturing industry is characterized by the centrality of information (Kiraz et al. 2020). From the wishes of customers to the definition of the offer to get to the production and distribution of the product and the corollary services, the virtuous circle of information is enabled by technologies. In related factory, in fact, are not only people to be connected and communicating: as are all assets. From the materials to the media that carry them, from the plants to the products, a new generation of sensors enables improved decision-making very datadriven (Arnold and Voigt 2019).

Digital Economy and Society Index (DESI) Currently, one of the most discussed issues is that of the digitalization of companies or the possibility of giving life to absolutely innovative business models from a smart manufacturing perspective. In this regard, it is worth mentioning the Digital Economy and Society Index – DESI (https://ec.europa.eu/digital-singlemarket/en/digital-economy-and-society-index-desi). The DESI monitors a series of parameters to measure the level of digitization of European countries in five macro-areas: connectivity (worth 25% of the index), digital skills (worth 25% of the index), use of Internet by individuals (worth 15% of the index), integration of digital technologies by companies (worth 20% of the index), and digital public services (worth 15% of the index) (Kletskova et al. 2020; Poor and Basl 2018). The 2019 data on which the DESI 2020 is based show connectivity goes from 12 to 12.5, human capital from 8.01 to 8.11, the use of the Internet by individuals from 6.25 to 6.67, the integration of digital technologies by companies from 6 to 6.25, and digital public services from 9.29 to 10.1 (Fig. 3). In 2020, all member states showed progress in all the main sectors measured by the index. But the EU Commission notes that all states should “intensify their efforts to improve coverage of very high-capacity networks, assign the 5G spectrum to allow the commercial launch of 5G services, improve citizens’ digital skills and further digitize businesses and public sector.” Executive Vice President Margrethe Vestager said, “the COVID-19 crisis has demonstrated how critical it is for citizens and businesses to be connected and able to interact with each other online.” The main consideration that can be drawn is that it is necessary to increase investments in digital services and innovations. As part of the plan for the recovery of Europe, adopted on May 27, 2020, the DESI will have to guide the specific analysis by country in support of the recommendations on digital. This will help member states to target their reform and investment needs and prioritize them. According to the DESI 2020, Finland, Sweden, Denmark, and the Netherlands are the top four countries for global digital performance within the EU, followed by Malta, Ireland, and Estonia. In the last 5 years, however, the most significant progress

2 Industry 4.0 Perspectives: Global Trends and Future Developments

29

Digital Economy and Society Index, by Main Dimensions of the DESI 15

Score (0 to 100)

12.5 10 7.5 5 2.5 0

2015

Legend 1 Connectivity

2016

2 Human Capital

2017

2018

3 Use of Internet

2019

4 Integration of Digital Te...

2020 European Commission, Digital Scoreboard 5 Digital Public Services

Fig. 3 DESI by dimension (year 2020)

has been made by Ireland, followed by the Netherlands, Malta, and Spain. These countries also performed well above the EU average based on the DESI 2020 score. As the pandemic has had a major impact on each of the five aspects examined by DESI, the Commission underlines that the 2020 results should be read in the light of the many measures taken by the Commission and member states to manage the crisis and support the recovery. The report, on the other hand, highlights serious shortcomings as regards the human capital dimension, as shown in Fig. 4. In this specific dimension, which includes basic skills for the use of the Internet by the population and the most advanced skills, Italy is last in Europe. Going into detail, Italy ranks last in terms of the number of graduates in ICT disciplines, but also in all other subdimensions it is far below the EU average.

Regional Innovation Scoreboard (RIS) The Regional Innovation Scoreboard (RIS) is a regional extension of the European innovation scoreboard, assessing the innovation performance of European regions on a limited number of indicators (EU, Regional Innovation Scoreboard Report 2019). The RIS 2019 covers 238 regions across 23 EU countries, Norway, Serbia, and Switzerland. In addition, Cyprus, Estonia, Latvia, Luxembourg, and Malta are included at country level. RIS summarizes the performance based on 17 indicators (Lilles et al. 2020; Arbolino et al. 2019). These indicators are grouped into four main types: framework conditions, investments, innovation activities, impacts, and ten innovation dimensions. Europe’s regions are grouped into four innovation performance groups according to their performance on the Regional Innovation Index relative to that of the EU: innovation leaders, strong innovators, moderate innovators, and modest innovators. The report shows that for most of the 159

30

A. Petrillo and F. De Felice 2 Human Capital, by Sub-dimensions of 2 Human Capital 90 80 70

Score (0 to 100)

60 50 40 30 20 10

Fi nl Sw and ed e N Es n U eth ton ni te erl ia d an Ki d ng s do M m D alt Lu en a xe ma m rk bo Au urg G stri er a m a Ire ny Eu l ro B and pe el an giu U m ni o C n ro C atia ze Sl chi ov a en Sp ia a Fr in Li anc th e u H ania un g Sl ary ov Po aki rtu a g Po al la C nd yp ru La s t G via re Bu ec lg e R ari om a an ia Ita ly

0

Legend 2a Internet User Skills 2b Advanced Skills and D...

European Commission, Digital Scoreboard

Fig. 4 DESI by subdimension “Human Capital” (year 2020)

regions, the performance of innovation has improved in the course of monitoring, which began 9 years ago. Figure 5 shows the Regional Innovation Scoreboard 2019 – relative performance to EU from 2011 to 2019. It emerges that the most innovative region in Europe is Zurich ¨ in Switzerland. Helsinki-Uusimaa (Finland) is the most innovative region in the EU, followed by Stockholm (Sweden). Figure 6 shows an overview for 2019 where countries are classified into four innovation performance groups. The most innovative regions, on average, perform best on most indicators as shown in the radar graph below (Fig. 7). Considering Fig. 5, it is meaningful to note that the strong innovators perform close to average (between 20% below or 20% above the EU average) on almost all indicators, except for lifelong learning (127%).

Report Digital 2020 (We Are Social) A comprehensive look at the state of the Internet, mobile devices, social media, and digitization is provided by Hootsuite with its annual report “We Are Social” (https:// wearesocial.com/digital-2020). Thanks to this report, it is possible to have an overview of the digital world, moving from the use of the Internet and social media at a national and global level to the habits and concerns of online users. The “Digital 2020” report highlights interesting reflections both on the changes taking place in

2 Industry 4.0 Perspectives: Global Trends and Future Developments

31

Fig. 5 Regional Innovation Scoreboard 2019 – relative performance to EU from 2011 to 2019

Fig. 6 Regional Innovation Scoreboard 2019 – overview

the digital scenario and on possible future trends that will also affect social changes with a view to industrialization. For example, as far as eCommerce is concerned, the most interesting data is that for the first time mobile purchases exceed those

32

A. Petrillo and F. De Felice

Fig. 7 Regional Innovation Scoreboard 2019 – average indicator scores by regional performance group

Fig. 8 Digital around the world in 2020 – We Are Social Report 2020. (Sources: population: United Nations; local government bodies; mobile: GSMA intelligence; internet: ITU; GlobalWebIndex; GSMA intelligence; local telecoms regulatory authorities and government bodies; APJII; Kepios analysis; social media: platforms’ self-service advertising tools; company announcements and earnings reports)

from desktops and laptops. Furthermore, the distinction between online and offline purchases is becoming increasingly subtle, giving life to omnichannel shopping experiences, between the real and the virtual world. Figure 8 shows the essential headline data you need to understand mobile, Internet, and social media use. Among the trends to keep an eye on are gaming and voice technology. It is precisely voice-based technologies that have registered an increase of 9% compared to 2019, thanks to the massive spread of voice assistants such as Amazon Echo and Google Home, increasingly used regularly by users of all ages.

2 Industry 4.0 Perspectives: Global Trends and Future Developments

33

State of the Art on Industry 4.0 This section presents the descriptive findings first for the nonacademic literature, thereafter for the academic one. In detail, to analyze the degree of perception related to the terms “Industry 4.0” and “smart manufacturing,” Google Trends has been used as a source of valid knowledge for emerging technologies. In fact, it allows you to know the search frequency on the web search engines for a specific word or phrase. It is a great source for evaluating the seasonality and changes over time of specific keyword volumes on search engines. Figure 9 shows the result of investigation from January 1, 2020 to November 2020 (the period of investigation). The investigation pointed out that Industry 4.0 is more used and more popular than the term “smart manufacturing” even if they are generally used indifferently. It was also interesting to discover the questions people are asking on Industry 4.0 and smart manufacturing. Using AlsoAsked.com, we tried to examine the results. Figure 8 shows that the most frequently asked question is: “How does Industry 4.0 affect industry?” and “Is Industry 4.0 only for manufacturing?” (Fig. 10). In our opinion, the results are an interesting starting point to investigate these aspects in detail. Thus, we conducted on SCOPUS, the largest abstract and citation database of peer-reviewed literature, a search by using the keywords Industry 4.0 and smart manufacturing restricting to TITLE-ABS-KEY (title, abstract, and keywords). We used titles, keywords, and abstracts in order to extract very specific expression describing a technology. Database query highlighted 1763 document results. Figure 11 shows the trend of publications from 2013 (the first year the first publication appeared) until now. It is also interesting to note that 52.1% of the papers are conference papers, 37.4% are articles, and 3.3% are book chapters, as shown in Fig. 12.

Fig. 9 Google Trends results

Fig. 10 Investigation on AlsoAsked.com

34

A. Petrillo and F. De Felice

Fig. 11 Documents by year

Fig. 12 Documents by type

It is no surprise to point out that most of the publications belong to Italian authors (about 14%) and German authors (about 12%) since these two countries are very active in the development of national strategies and beyond (as shown in Fig. 13). In fact, starting from March 2017, France, Germany, and Italy have launched a trilateral cooperation to promote the digitization of the manufacturing sector and to support the European Union’s efforts in this area. The cooperation involves all the stakeholders of the national strategies for Industry 4.0 – Plattform Industrie 4.0 for Germany, Alliance Industrie du Futur for France, and Piano Impresa 4.0 for Italy, who signed a common roadmap in Turin on June 20, 2017.

2 Industry 4.0 Perspectives: Global Trends and Future Developments

35

Fig. 13 Documents by country

Fig. 14 Documents by subject area

Regarding the subject area, it is interesting to note that most of the publications belong to the engineering area (31.8%), followed by computer science (27.0%) and decision sciences (7.2%), as shown in Fig. 14. Of course the most used words are Industry 4.0 and smart manufacturing/factory, but by stratifying the data, it emerges that Internet of things and cyber-physical

36

A. Petrillo and F. De Felice

system are the most used keywords, respectively, with about 27.0% and 18.0% of 1.793 publications (see Fig. 15). Furthermore, it emerged that important trends for the future of manufacturing are represented by the following topics: embedded systems, flow control, engineering education, supply chains, decision-making, life cycle, and predictive maintenance, as shown in Fig. 16.

Fig. 15 Documents by keywords

Fig. 16 Documents by topics

2 Industry 4.0 Perspectives: Global Trends and Future Developments

37

Definitely, the investigation pointed out that the issue on Industry 4.0 is a very potential topic at least from an academic and public perception point of view. However, it is clear that in the case of such a vital topic, a country’s strategies must be based on the needs of the industrial sector, as detailed in the following section.

The Future of Manufacturing: Survey Results and Analysis The aim of this section is to outline a survey that we conducted to identify the drivers of development of manufacturing after the pandemic in the perspective of I4.0. In particular, the goal of the survey was to analyze the role that digital technologies play within the national and international production system. We have tried to answer the following questions: What could be the national challenges for the near future to reduce the digital divide?

Survey Objectives and Setup The survey was developed from April to October 2020. Of course, the survey was conducted online (considering the COVID-19 pandemic) using social networks (such as LinkedIn, Facebook, WhatsApp, etc.). In this way, it was possible to increase the level of diffusion both nationally and internationally. At present, about 2500 companies have responded to the survey. We asked to interview three different company areas within the manufacturing companies with the goal of discovering the needs and innovation priorities but also the challenges these different functions are facing. Figure 17 shows interviewees by company area.

Survey Results The survey consists of 30 questions, as follows: section #1, analysis of the companies under investigation (such as size, number of employees, etc.); section #2, analysis of “digital” skills; and section #3, needs analysis of the digital innovation.

Fig. 17 Interviewees by company area (% of total)

38

A. Petrillo and F. De Felice

Fig. 18 Investment in digitalization perspective

Here below is a summary of the main aspects analyzed and that emerged from the survey. It emerged that 95% of respondents believe that digital technologies are useful for their company. In fact, the issues related to digital transformation are regularly discussed within the company: quarterly in 63% of cases, annually in 15% of cases, and every 6 months in 15% of cases. The topic is never dealt with in 7% of cases. Generally within the company, leadership for digital transformation is assigned as follows: 45% IT manager, 23% R&D manager, 17% chief digital officer, and 15% none. Regarding the IT priorities of manufacturing companies for 2020, it emerged that IT departments continue to be looking for new infrastructural models and architectural to support workloads increasingly data-intensive. The renewal and infrastructural innovation will move towards hybrid models which include the simultaneous use of on-premise services, on private clouds and on public clouds. Security continues to cap a key role, for data protection and systems. The investment made concerned the following sectors, as shown in Fig. 18. More specifically, the survey pointed out that specific investments were made, as shown in Fig. 19. About 30% of the investments concern IT infrastructure, and about 20% concerns web applications, followed by other types of investments. It is remarkable to note that the most requested figures in work teams on digital issues are system administrator/engineer (38%) and Big Data analyst (29%), as shown in Fig. 20. The enabling technologies used most internally in the company are cybersecurity (18%), industrial Internet of things (16%), Big Data and analytics (15%), and so on, as detailed in Fig. 21. However, some obstacles emerge to digitization such as IT infrastructure obsolescence (29%), difficulty in implementing new solutions effectively (23%), and inadequate legacy (18%), as shown in Fig. 22. It is very significant to note that according to the interviewees (93%) there is a close relationship between digital transformation, energy efficiency, and sustainability. In fact, interviewees believe that it is necessary to make investments in digital technologies that will have (direct or indirect) benefits in efficiency and

2 Industry 4.0 Perspectives: Global Trends and Future Developments

39

Fig. 19 Detail of investments in digitalization perspective

Fig. 20 Work teams on digital issues

sustainability. In particular, 65% of them believe that it is essential to investe to improve production processes, 20% to improve internal logistics processes, and 15% to improve the supply chain. The survey shows that companies perceive the importance of investing in digital innovation with a view to sustainability as well, since this aspect will represent a key factor in achieving the competitiveness of the industrial world and beyond. In other words, it emerges that technology is influencing and will increasingly influence structural changes in the economy and will promote sustainable efficiency in production systems. The survey highlighted

40

A. Petrillo and F. De Felice

Fig. 21 Enabling technologies most used

Fig. 22 Obstacles to digitization

some salient aspects for the development of digital transformation. Ultimately, the criteria of economic, social, and environmental sustainability defined by the United Nations and consolidated in the 2030 Agenda must become a beacon in the choices that will determine the development of technologies as tools for building a sustainable future.

Challenges, Directions, and Trends The potential economic benefits of Industry 4.0 are undoubted. The industry is in the midst of a revolution destined to radically change not only the way we work and the business model but also the lives of all of us for new, possible applications. The drivers of development after the pandemic can be summarized as follows: provides digital tool technology and innovative applications, can help

2 Industry 4.0 Perspectives: Global Trends and Future Developments

41

Fig. 23 Driver to digitization

improve productivity, creates a more attractive working environment, and enhances the employee satisfaction and customers (see Fig. 23). According to the previous analysis, some specific key challenges have been defined to do with technological and societal issues, as explained below.

Research Direction #1: Technical Challenges The great transformation is digitalization, which pervades the manufacturing sector, with machines becoming able to “talk” to each other, thus offering development scenarios that were unimaginable only until recently. The first critical issue is integration in the awareness that the challenge is not played only on the terrain of the stable connection between machines, objects, cloud applications, and people but also in the ability and availability to be “contaminated” by new technologies and new digital services, through a rethinking in the sense of all areas of value creation (Gowripeddi et al. 2021; Danielsson et al. 2020). The second critical issue is flexibility to compete on world markets, that is, the ability to adapt as quickly as possible to the volatility of market demands. The third critical issue is skills for the future of manufacturing, i.e., the key role of people in facing the challenges of growing economic, social, and environmental complexities. Skills for the manufacturing of the future are a global challenge (López Ríos et al. 2020). The most important effect of the transformation of work is therefore represented by the

42

A. Petrillo and F. De Felice

growth of skills to carry out jobs of greater added value, more “human,” and this represents the most important industrial revolution that has ever occurred (CantúOrtiz et al. 2020). Among the top skills, there are digital literacy, intercultural and disciplinary mentality, ability to manage increased complexity, and constant openness to change. The fourth critical issue is standardization, legal systems, and infrastructure configuration (e.g., the expansion of advanced power grids and communication networks). Industry 4.0 has the potential of becoming the global language of production. In this regard, the introduction of a uniform industry standard is especially important (Larrañaga et al. 2020; Müller and Voigt 2020). Some standardization has already been developed. For example, the reference model for the intelligent manufacturing architecture of tomorrow’s Industry 4.0 (RAMI 4.0) was presented at the 2015 Hannover Fair.

Research Direction #2: Social Challenges Digitization and sustainability will be levers on which to focus today and in the coming years for the development of manufacturing (Riedelsheimer et al. 2021), a sector that, in the delicate phase of the relaunch after the epidemic, is crucial for the recovery and the creation of economic and social wealth (Romero et al. 2020; Bag et al. 2020). The analysis on the importance of sustainability and digital is also well connected to the pillars identified by the European Union including the “green deal” and the enhancement of innovation as a fundamental economic driver (Acerbi and Taisch 2020; Fokaides et al. 2020). Moreover, in this particular moment, in which the whole world is working on the recovery of the economy after the hard months characterized by the coronavirus epidemic, technology transfer, research, and a green look on production are no longer postponed even in the sector. (Leong et al. 2020; Fatimah et al. 2020).

Conclusion Nowadays, the size or strength of companies is not enough to be successful: it is essential to possess the ability to change, adapt quickly, seize opportunities, and be agile. Production processes are rapidly evolving: new technologies are enabling machinery and production systems to communicate through IP protocols, without the need for human interaction, providing the ability to collect and analyze data and information in real time. Manufacturing companies do not stand addressing only the challenges related to productivity and satisfaction of customers, but they are confronting each other also with innovations related to work environments and skills digital. The rapid and continuous changes in the market, accompanied by a growing competition, are imposing on manufacturing companies a rethinking of products and processes, to exploit and enhance all data available inside and outside the company itself. Data is today at the center of interactions between people, processes, and technologies and is able to make internal activities more efficient, to enable better

2 Industry 4.0 Perspectives: Global Trends and Future Developments

43

decision-making processes, and to improve the customer experience. Innovation will increasingly require collaboration and sharing of information between different business functions. Companies will have to then adopt systems and solutions capable of enabling this collaboration so as not to remain on the margins of transformation digital.

References Acerbi F, Taisch M (2020) A literature review on circular economy adoption in the manufacturing sector. J Clean Prod 273:123086 Arbolino R, Boffardi R, De Simone L (2019) Which are the factors influencing innovation performances? Evidence from Italian Cohesion Policy. Soc Indic Res 146(1–2):221–247 Arnold C, Voigt K-I (2019) Determinants of industrial Internet of Things adoption in German manufacturing companies. Int J Innov Technol Manag 16(6):1950038 Bag S, Yadav G, Wood LC, Dhamija P, Joshi S (2020) Industry 4.0 and the circular economy: resource melioration in logistics. Res Policy 68:101776 Bonaccorsi A, Chiarello F, Fantoni G, Kammering H (2020) Emerging technologies and industrial leadership. A Wikipedia-based strategic analysis of Industry 4.0. Expert Syst Appl 160:113645 Bougdira A, Akharraz I, Ahaitouf A (2020) A traceability proposal for Industry 4.0. J Ambient Intell Humaniz Comput 11(8):3355–3369 Bragazzi NL (2020) Digital technologies-enabled smart manufacturing and Industry 4.0 in the post-COVID-19 era: lessons learnt from a pandemic. Int J Environ Res Public Health 17(13):4785 Cantú-Ortiz FJ, Galeano Sánchez N, Garrido L, Terashima-Marin H, Brena RF (2020) An artificial intelligence educational strategy for the digital transformation. Int J Interact Des Manuf 14(4):1195–1209 Cioffi R, Travaglioni M, Piscitelli G, Petrillo A, De Felice F (2020) Artificial intelligence and machine learning applications in smart production: progress, trends, and directions. Sustainability 12(2):492 Corrocher N, Mavilia R, Giorgio M (2020) The Sino-German alliance for the fourth industrial revolution: dynamics and policy implications. J Econ Policy Reform 23(4):426–446 Culot G, Orzes G, Sartor M, Nassimbeni G (2020) The future of manufacturing: a Delphi-based scenario analysis on Industry 4.0. Technol Forecast Soc Chang 157:120092 Danielsson O, Holm M, Syberfeldt A (2020) Augmented reality smart glasses in industrial assembly: current status and future challenges. J Ind Inf Integr 20:100175 De Felice F, Petrillo A, Zomparelli F (2018) Prospective design of smart manufacturing: an Italian pilot case study. Manuf Lett 15:81–85 Elhoone H, Zhang T, Anwar M, Desai S (2020) Cyber-based design for additive manufacturing using artificial neural networks for Industry 4.0. Int J Prod Res 58(9):2841–2861 El-kaime H, Elhaq SL (2021) Methodology for implementation of Industry 4.0 technologies in supply chain for SMEs. Adv Intell Syst Comput 1193:59–76 Farooqui A, Bengtsson K, Falkman P, Fabian M (2020) Towards data-driven approaches in manufacturing: an architecture to collect sequences of operations. Int J Prod Res 58(16):4947– 4963 Fatimah YA, Govindan K, Murniningsih R, Setiawan A (2020) Industry 4.0 based sustainable circular economy approach for smart waste management system to achieve sustainable development goals: a case study of Indonesia. J Clean Prod 269:122263 ˇ Fokaides PA, Apanaviciene R, Cerneckiene J, . . . Ždankus T (2020) Research challenges and advancements in the field of sustainable energy technologies in the built environment. Sustainability 12(20):8417, 1–20

44

A. Petrillo and F. De Felice

Gowripeddi VV, Ganesan N, Shailendra S, Thangaraju B, Bapat J (2021) Enabling software defined networking for Industry 4.0 using OpenStack. Adv Intell Syst Comput 1231:588–602 Haipeter T (2020) Digitalisation, unions and participation: the German case of “Industry 4.0”. Ind Relat J 51(3):242–260 Helmann A, Deschamps F, Loures EDR (2020) Reference architectures for Industry 4.0: literature review. Adv Transdiscip Eng 12:171–180 Hernandez-de-Menendez M, Morales-Menendez R, Escobar CA, McGovern M (2020) Competencies for Industry 4.0. Int J Interact Des Manuf 14(4):1511–1524 Hincapié M, Valdez A, Güemes-Castorena D, Ramírez M (2020) Use of laboratory scenarios as a strategy to develop smart factories for Industry 4.0. Int J Interact Des Manuf 14(4):1285–1304 Jeon B, Yoon J-S, Um J, Suh S-H (2020) The architecture development of Industry 4.0 compliant smart machine tool system (SMTS). J Intell Manuf 31(8):1837–1859 Kiraz A, Canpolat O, Özkurt C, Ta¸skın H (2020) Analysis of the factors affecting the Industry 4.0 tendency with the structural equation model and an application. Comput Ind Eng 150:106911 Kletskova EV, Titova OV, Vorobyova VV (2020) The fourth technological paradigm as a vector of sustainable development of the modern digital economy: implications for society, government and entrepreneurship. Lect Notes Netw Syst 129:1411–1419 Larrañaga A, Lucas-Estañ MC, Martinez I, Val I, Gozalvez J (2020) Analysis of 5G-TSN integration to support Industry 4.0. In: IEEE international conference on Emerging Technologies and Factory Automation (ETFA) 2020, September, 9212141, pp 1111–1114 Leong WD, Teng SY, How BS, . . . Lam HL (2020) Enhancing the adaptability: lean and green strategy towards the Industry revolution 4.0. J Clean Prod 273:122870 Li Y, Goga K, Tadei R, Terzo O (2021) Production scheduling in Industry 4.0. Adv Intell Syst Comput 1194:355–364 Lilles A, Rõigas K, Varblane U (2020) Comparative view of the EU regions by their potential of university-industry cooperation. J Knowl Econ 11(1):174–192 Lopes J, Farinha L, Ferreira JJM (2019) Reflecting on the innovative performances of European regions in the age of smart specialisation. Glob Bus Econ Rev 21(5):605–623 López Ríos O, Lechuga López LJ, Lechuga López G (2020) A comprehensive statistical assessment framework to measure the impact of immersive environments on skills of higher education students: a case study. Int J Interact Des Manuf 14(4):1395–1410 Lu Y, Xu X, Wang L (2020) Smart manufacturing process and system automation – a critical review of the standards and envisioned scenarios. J Manuf Syst 56:312–325 Ma, S., Zhang, Y., Liu, Y., . . . Ren S (2020) Data-driven sustainable intelligent manufacturing based on demand response for energy-intensive industries. J Clean Prod 274:123155 Müller JM, Voigt K-I (2020) Industry 4.0 – integration strategies for small and medium-sized enterprises. In: 26th International Association for Management of Technology Conference (IAMOT) 2017, pp 564–578 O’Donovan P, Leahy K, Bruton K, O’Sullivan DTJ (2015) An industrial big data pipeline for datadriven analytics maintenance applications in large-scale smart manufacturing facilities. J Big Data 2(1):25 Oláh J, Aburumman N, Popp J, . . . Kitukutha N (2020) Impact of Industry 4.0 on environmental sustainability. Sustainability 12(11):4674 Pinzone M, Fantini P, Perini S, . . . Miragliotta G (2017) Jobs and skills in Industry 4.0: an exploratory research. IFIP Adv Inf Commun Technol 513:282–288 Poor P, Basl J (2018) Czech Republic and processes of Industry 4.0 implementation. In: Annals of DAAAM and proceedings of the international DAAAM symposium, vol 29(1), pp 0454–0459 Riedelsheimer T, Neugebauer S, Lindow K (2021) Progress for life cycle sustainability assessment by means of digital lifecycle twins – a taxonomy. In: Kishita Y, Matsumoto M, Inoue M, Fukushige S (eds) EcoDesign and Sustainability II. Sustainable Production, Life Cycle Engineering and Management. Springer, Singapore. https://doi.org/10.1007/978-981-15-67759_21 Romero D, Stahre J, Taisch M (2020) The Operator 4.0: towards socially sustainable factories of the future. Comput Ind Eng 139:106128

2 Industry 4.0 Perspectives: Global Trends and Future Developments

45

Sasiain J, Sanz A, Astorga J, Jacob E (2020) Towards flexible integration of 5G and IIoT technologies in Industry 4.0: a practical use case. Appl Sci 10(21):7670 Shi Z, Xie Y, Xue W, . . . Xu X (2020). Smart factory in Industry 4.0. Syst Res Behav Sci 37(4):607–617 Sommer L (2015) Industrial revolution – Industry 4.0: are German manufacturing SMEs the first victims of this revolution? J Ind Eng Manag 8(5):1512–1532 Szabo RZ, Herceg IV, Hanák R, . . . Djuriˇcin D (2020) Industry 4.0 implementation in b2b companies: cross-country empirical evidence on digital transformation in the cee region. Sustainability 12(22):9538 Trstenjak M, Opetuk T, Cajner H, Tosanovic N (2020) Process planning in Industry 4.0 – current state, potential and management of transformation. Sustainability 12(15):5878 Xu LD (2020) The contribution of systems science to Industry 4.0. Syst Res Behav Sci 37(4):618– 631

Websites References Digital Economy and Society Index (DESI). https://ec.europa.eu/digital-single-market/en/digitaleconomy-and-society-index-desi Elsevier’s Scopus www.scopus.com Regional Innovation Scoreboard Report 2019 (2019). Publications Office of the European Union, Luxembourg. https://doi.org/10.2873/85586. ISBN 978-92-76-08724-3. https://ec.europa.eu/ growth/industry/innovation/facts-figures/regional_en We are social report 2020. https://wearesocial.com/digital-2020

3

Changing Manufacturing Landscape: From a Factory to a Network Karl-Erik Michelsen, Mikael Collan, Jyrki Savolainen, and Paavo Ritala

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . History of Manufacturing: Change Within Factory Walls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Regime of Accumulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . From Economies of Scale to Economies of Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The End of Division of Labor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Technological and Business Drivers that Underlie Industrial Landscape Change . . . . . . . . . . . Digitalization of Business and Manufacturing Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . Automation and Manufacturing Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Additive Manufacturing Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Logistics Optimization and Supply Chain Risk Management . . . . . . . . . . . . . . . . . . . . . . . . Proactive Versus Reactive Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion: Toward Manufacturing as a Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

48 49 49 52 54 55 56 58 59 60 60 61 64 67

Abstract

This chapter focuses on the technological, the managerial, and the societal transformation from the old manufacturing system into the new, discussing the drivers, the challenges, and the opportunities connected to the transformation.

K.-E. Michelsen · J. Savolainen · P. Ritala School of Business and Management, Lappeenranta-Lahti University of Technology, Lappeenranta, Finland e-mail: [email protected] M. Collan () School of Business and Management, Lappeenranta-Lahti University of Technology, Lappeenranta, Finland VATT Institute for Economic Research, Helsinki, Finland © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_2

47

48

K. E. Michelsen et al.

The suggestion is that we are moving from a Taylorist-Fordian factory model towards “Manufacturing-as-a-Network,” which indicates new types of busines possibilities, risks, and transformative implications to the society at large. Traditionally, manufacturing happens within factory walls, where a factory is understood as a place for mass production of goods. It is an assembly of machines and workers who are organized and managed to maximize efficiency and productivity. “Manufacturing-as-a-Network” is unlike the factory as we know it and answers to the needs of the postindustrial society. It is a network structured to perform specific and tailored products in collaboration with customers, for customers, and sometimes by customers. Keywords

Industry 4.0 · Manufacturing · Factory · Business model · Network

Introduction The industrial and manufacturing landcape is undergoing a major change. We are moving from Taylorist and Fordian model of mass production and scale economies toward a networked and autonomous model of production. In this chapter, the attempt is to outline the history, the present, and the future of this development, by focusing on the changes that take place with regards to how manufacturing is organized. The main argument in this chapter is that the concept of “factory” is radically changing. Activities that take place in the traditional “factory floor” or “within factory walls” are being revolutionized by digital technologies and digitalization and for some industries transform into a digitally controlled manufacturing network, see Fig. 1. The change in the industrial landscape which we are witnessing is not only technical, but socio-technical (Geels 2004). This means that not only digitalization

TRADITIONAL FACTORY Ca. 1850 – Ongoing

HYBRID Ca. 2010 – Future

Manual and semiautomatic mass production-based facilities controlled manually operating as separate entities mostly based on bilateral contracts, supported by small scale customization.

Semi-automated mass production manufacturing facilities operating as a part of a supply chain, supported by advanced flexible manufacturing.

MANUFACTURING AS A NETWORK Future Digitally controlled network of flexible autonomous manufacturing facilities operating near customers and based on demand supported by high-volume mass production tied to the network.

Fig. 1 Transformation from a traditional factory to manufacturing as a network

3 Changing Manufacturing Landscape: From a Factory to a Network

49

changes how industrial work is done, but also who (or what) does the work, how this work is organized, and how it affects the broader economy and the society. In Sect. “History of Manufacturing: Change Within Factory Walls”, the history of production is briefly visited in the context of a factory, informed by Taylorian and Fordist models of industrial production. The change toward post-Fordian production is discussed, where the role of a worker is that of a highly skilled operator of and among increasingly autonomous machines. In Sect. “Technological and Business Drivers that Underlie Industrial Landscape Change”, the new management models and technologies that are driving the change toward a networked manufacturing model are discussed. In Sect. “Conclusion: Toward Manufacturing as a Network”, the new perspective – which is here called “manufacturing-as-a-network” is outlined.

History of Manufacturing: Change Within Factory Walls Regime of Accumulation In his book The Principles of Scientific Management, published in 1907, Frederick Winslow Taylor defined the role of the factory as follows: “In the case of a more complicated manufacturing establishment, it should also be perfectly clear that the greatest permanent prosperity for the workman, coupled with the greatest prosperity for the employer, can be brought about only when the work of the establishment is done with the smallest combined expenditure of human effort, plus nature’s resources, plus the cost for the use of capital in the shape of machines, buildings, etc. Or, to state the same thing in a different way: that the greatest prosperity can exist only as the result of the greatest possible productivity of the men and machines of the establishment” (Taylor 1911). Taylor applied scientific method to the study of work. He conducted time-andmotion studies and based on data, he reorganized the factory work. Taylor’s method eliminated useless tasks and unnecessary movements and created a comprehensive management system that reflected scientific ideals of accuracy and precision. According to Taylor, labor disputes could be eliminated if corporate leaders and managers adapted a new attitude toward industrial work. Maximizing output would automatically maximize profits and income for workers and this goal would make political and ideological disputes obsolete (Warring 1991). Meanwhile, Henry Ford was building a new manufacturing system at the Highland Park automobile factory in Detroit. He combined special-purpose machines with semiskilled labor and organized them along the continuously moving belt. The assembly line combined several manufacturing concepts into the system that revolutionarized the manufacturing of automobiles. Premanufactured and interchangeable parts were placed on the moving belt where semiskilled workers assembled them into the final product. The complex automobile could be manufactured without delays and optimally the assembly line cranked out a new T-Model Ford every 20 s (Hounshell 1984).

50

K. E. Michelsen et al.

Mass production system became the dominant industrial paradigm in the USA during the early part of the twentieth century. It replaced traditional manufacturing systems, which allowed skilled workers to contribute independently to the final product. Mass production system subjected workers to the manufacturing process that utilized economies of scale. As Taylor had promised, the standardized manufacturing system brought higher profits for owners and higher wages for the workers. This in turn improved living standards in industrial societies. New manufacturing systems required massive capital investments, centralized managerial control, and industrial hierarchy that copied modes and methods from the bureaucratic state. Hence, it was no surprise that the industrial landscape was taken over by large, vertically and horizontally integrated corporations (Chandler 1990). Henry Ford and Fredrick W. Taylor were idealistic innovators who operated within a highly competitive American industrial landscape. They tackled the fundamental dilemma that had troubled factory owners and managers since the dawn of the industrial revolution. Factories combined human labor and machines and created, therefore, a series of qualitative and quantitative changes in production systems. How to organize this combination to maximize the output and productivity, but without creating labor disputes? Technological developments improved machine efficiency, but improvements in technology did not bring positive improvements in human labor (Warring 1991). Taylor’s system of scientific management tried to find generalized rules of conduct, based on “laws of nature” that would bring maximum output, high productivity, low cost, high wages, equitable distribution, reduction of unemployment, and rapid economic growth (Thompson 1916). Ford, on the other hand, tried to solve the dilemma by reconceptualizing the production process. He borrowed the idea of assembly line from other industrial sectors and coupled it with the innovations in interchangeability, new special-purpose machine tools, standardization, singlemodel policy, simplification of design, radical de-skilling of work force, and centralized control of the flow of work (Hounshell 1984). Neither Taylor nor Ford was able to solve the problems of division of labor within the factory walls. Taylor’s ambitious search for one best way found too many human variables to be generalized as a law of nature. Ford’s ambitious system believed that if workers would perform only one simple task, they would become parts of the machine system. However, human labor could not match the tempo of the machines and frequent speedups by managers created labor disputes. Therefore, the fundamental dilemma of the division of labor was embedded in the mass production system. As Wang and Siau (2019) point out, the idea of the factory without human labor has been long ago elaborated in science fiction literature, but so far it has been socially and politically too hot a topic to be seriously discussed. Factories provided salaried work to millions of people and industrial societies depend on high employment, mass consumption, and tax revenues. Hence, mass production systems have cautiously moved toward more streamlined production systems that require more technology and less human labor. In the early days, technological innovations took over tasks that required extensive

3 Changing Manufacturing Landscape: From a Factory to a Network

51

physical strength and speed. The next steps were machines that performed high precision tasks. More recently, computers with massive calculation capacity are installed to control complex, vertically and horizontally integrated production systems. This has changed the division of labor within factory walls. Tasks that were previously performed by de-skilled workers are handed over to machines. Human labor is still needed in tasks that require cognitive skills, planning, and refining industrial processes, this was already observed by Pine 30 years ago (Pine 1980). It is often believed that manufacturing systems are shaped by technological developments only. This narrative is challenged by the historical record: Since the birth of the factory, manufacturing systems have interacted with political, economic, and social systems on several different levels. They have reorganized and reconfigured the human-machine relationship, which affects lives of millions of people. In addition, they have established regimes of accumulation that determined economic developments in industrial societies. Finally, manufacturing of inexpensive industrial goods affected the lifestyles and consumptions habits of people living in industrial societies. The effects of the mass production system shaped the industrial landscape during the twentieth century not only in North America, but also in Europe, Soviet Union, and Japan. When the global political map changed at the end of the millennium, Fordism spread rapidly to China, India, and other Asian countries, but also to South America and to Africa. Wherever the mass production system was applied, it shaped social, economic, and political structures. Large factories, connected to the distribution, communication, and energy networks shaped the industrial landscapes and rearranged geographic and demographic structures. As Bob Jessop (1992) concludes, “given that economic activity (mass production system) is always socially embedded, socially regularized, and socially regulated, the state must be involved not only in securing the narrow techno-economic conditions for valorization and labor supply, but also in the broader, socio-economic embeddedness, regularization, and regulation of economic activities. It is this broader context that provides the link between economic and social reproduction, between accumulation and societalization.” Did Henry Ford and Frederik W. Taylor invent a perfect manufacturing system that was applicable to all political and ideological systems? Fordism has demonstrated resilience and adaptability as it has taken over country after country during the twentieth century. Where Fordism was established, societies have undergone fundamental structural changes. However, the interaction between the mass production system and modern industrial society is not a one-way street. Industrial societies control, govern, and regulate industrial developments to enhanced social and economic planning, maintain economic growth, and control social mobility. Fordism has proved to be one of the most efficient ways to organize mass production of industrial goods in modern societies. It has also proved to be an efficient way to ensure macroeconomic growth, full employment, and improvement in social welfare. Fordism has also successfully promoted urban-industrial development, which is dominated by middle-class and wage-earning families (Macdonald 1991).

52

K. E. Michelsen et al.

From Economies of Scale to Economies of Scope What does it take to change the dominant production system? As discussed before, Fordism has penetrated deeply the structures of modern industrial society. It has created links to industrial, educational, political, social, and cultural institutions and shaped the ways of life of millions of people. Industrial societies have become dependent on Fordism and vice versa. This wedlock between society and industry is difficult to break without altering social systems, organizations, and modes of behavior (Boyer and Durand 1993). What happened to Fordism after the mid 1970s is still an unanswered question. What is known is that advanced industrial countries went through fundamental economic changes, because of rapid increases in the oil price. This unexpected exogenous impulse drove Western economies into a stubborn recession. The tidal wave flushed over societies, which had become accustomed to a stable economic growth, increasing wages and profits, and highly standardized, but comfortable lifestyles. The oil crisis changed the long-term relationship between the mass production system and the modern industrial nation. Fordism had reduced relative prices and maintained mass consumption of industrial goods by rationalizing and standardizing industrial production. This, in turn, was based on the assumption that low-priced industrial goods, produced in bulk, would always find a consumer (Boyer and Durand 1993). Although the oil crisis sparked the economic downfall in the 1970s, it was just one factor among many that affected the future of the mass production system. Newly industrial nations in Asia, South America, and Africa had entered the global industrial landscape and they could offer Fordism abundant resources of low-wage semiskilled workers. However, as Lipietz (1985) and others have demonstrated, Fordism adapted a wide variety of shapes when it took over the manufacturing processes in newly industrial countries. Brazil and other vastly populated countries in Asia and South America developed a coexistence of relatively modern and dynamic sectors of production in dynamic urban growth centers and large rural regions, which supplied factories with low-wage and semiskilled industrial labor. In China, Fordism entered the era of development that was shaped by major internal structural changes. Because of its political, social, and cultural background, the industrial trajectory in China has become more complex than just a regime of accumulation (Walker and Buck 2007). China provides textbook circumstances for traditional Fordism with a low-cost labor resource, massive raw material and energy sources, and a vast population hungry for inexpensive industrial goods. As Lüthje (2013) points out, the industrial trajectory in China has been characterized by a coexistence of industrial sectors at various levels of development. There is also a wide gap separating the urban and the rural populations. According to Lüthje (2013), “this co-existence of industries, clusters, and regions with predominantly low-cost and labor intensive production based on rural labor on the one hand, and those with higher levels of capital intensity and social reproduction with mostly urban workforce, on the other, can be regarded as a key feature of China’s emerging

3 Changing Manufacturing Landscape: From a Factory to a Network

53

capitalism. It is closely linked to a governance of one-party state with both, a quasifederal and a highly centralized governance at the same time.” Hence, after the economic crises in the 1970s and the changes in the global political landscape during the final decade of the millennium, there was not one, but many mass production systems coexisting in the world. While traditional Fordism was adapting itself to the politically, socially, and culturally complex societies in Asia and South America, the advanced industrial nations in Europe and in North America, and Japan and South-Korea tried to find ways to get out from traditional mass production paradigm. Fordism was no longer a popular term and it was replaced by several new concepts. Structural crises in advanced industrial nations were identified as an “era of transformation,” “transition,” “post-modern,” “fifth Kondratiev,” “postcollective,” and “post-Fordist.” Although none of them could exactly describe the nature of the change, they all pointed toward the same direction. The era of mass production and mass consumption had come to an end, and advanced industrial nations had to find a new industrial paradigm that could satisfy customer needs, ensure economic growth, and improve the social and the economic standards. Although there were no inventors like Henry Ford or Frederic W. Taylor available, advanced industrial nations had a massive technological and intellectual capacity that could be redirected toward building the new industrial regime (Amin 1994). Flexible response to the global crises mirrored tensions that had built up within Fordism for a long period of time. Centralized control, monotonous and standardized work, and “one-size-fit-all” attitude toward consumer needs overlooked the social, economic, and cultural developments in modern industrial societies. As living standards improved, people became more aware of individual and private needs. Mass production of industrial goods was still needed, but there was a growing demand of tailored, high-quality products. There was also less and less semiskilled workers available in the Western world, because of demographic changes and improvements in education. Large-scale corporation, with the help of national and transnational innovation systems, developed technologies that lift them up in the technological hierarchies. Specialized new machines streamlined manufacturing systems and improved quality of products. These improvements were coupled with the managerial innovations that emphasized flexible specialization, decentralized management, and tailored solutions for identified customer segments. New manufacturing systems were operated by trained professionals, who replaced the semiskilled labor force. Monotonous manual work along the assembly line was taken over by robots and automated machines, whose operations were managed and controlled by skilled workers and managers (Tomaney 1994). As Schumann (1998) defines, post-Fordist system supports a worker, who is technically autonomous and intervenes in the manufacturing process, if it does not operate optimally. Hence, post-Fordism introduced a new division of labor within the factory walls. Human labor was given more independence and freedom, but the governance of the manufacturing process was handed over to the machines.

54

K. E. Michelsen et al.

As Schumann defines it: “if the technical system should work perfectly, the main responsibility of a system’s controller is to check and to service the machine. He himself does a perfect job if succeeds to anticipate deviation and breakdowns in the technical system and proceeds to initiate prevention.” Post-Fordism took a radical step away from orthodox Fordism and proceeded toward a new regime of accumulation. It was dominated by machines and skilled workers. Mass production systems utilized economies of scale, but the focus was shifting from quantity to quality. This transition toward economies of scope was perfected in Japan, where novel mass production systems were developed in the latter parts of the twentieth century. Lean manufacturing applied principles from Fordism, but placed them in the new manufacturing concept. As Jürgens and others (Jürgens 1989) observe, the Japanese manufacturing system builds on flexibility, in utilization of facilities and minimization of quality problems as they arise. Although the Japanese manufacturing system depended heavily on technological innovations, the real significance was not placed in machines themselves, but on how they were used in the manufacturing process (Sayer 1986). In summary, we have witnessed that post-Fordism has replaced Fordism as the dominant manufacturing system in advanced industrial countries during the first two decades of this millennium. Many corporations have struggled to overcome the transition from orthodox Fordism to post-Fordism. The dilemma of division of labor remains as so far there is no consensus on how the problem will eventually be solved. In fact, post-Fordism has created a permanent unemployment issue for the advanced industrial countries and potentially doomed them into slow economic growth. In the meanwhile, orthodox Fordism and its applications are driving economies on high gear in developing countries.

The End of Division of Labor Post-Fordism promised flexibility in production and individuality in consumption, new division of labor within factory walls, and new prosperity to industrial countries. This promise was fueled by the collapse of the Cold War in early 1990s and the end of the ideologically hostile world. Liberal market economy and corporate capitalism pushed aside the socialist planned economy. Global markets were unified by radical innovation in information technologies that made possible the borderless and continuous flows of goods and capital. Post-Fordism used the concepts from old mass production systems, but modified them to satisfy the needs of globalization. It was no longer necessary to centralize production and establish hierarchical management structures. It was equally unnecessary to collect massive stockpiles of raw materials in one location and hire thousands of workers. With the help of information systems, streamlined logistics, and automated manufacturing systems, mass production of inexpensive goods could be decentralized, outsourced, and reorganized to meet the demand of global markets. As Gambao (1988) has argued, the demand of inexpensive consumer goods became homogeneous across national borders and continents. This changed

3 Changing Manufacturing Landscape: From a Factory to a Network

55

the industrial landscape in the Western world, but also in Asia and other parts of the world. Domestic markets were no longer safe and protected against the invasion of inexpensive goods that were manufactured anywhere in the world. In order to survive, companies had to implement new strategies that emphasized mobility and flexibility. Factories that were too far from customers, or inefficient, were either closed or moved to another location, where they could utilize cheap materials, labor, and energy and where environmental and labor regulation were less restrictive. At the same time, global companies invested in high-quality production in advanced industrial countries, which offered skilled labor, research and development resources, and high-income customers. As Kern and Schaumann (1987) point out, a new conscious of qualitative significance of human work performance emerged from the aim to design flexible forms of automation. As automation became the holistic principle in manufacturing systems, it still demanded a highly skilled and specialized labor force that could collaborate with sophisticated machines. Hence, the automation did diminish the role of human labor, but it did not shut out workers from factories. On the contrary, the relationship between machine and worker became more complex as the intelligent machines challenged the creative capabilities of skilled workers. Hence, after a century-long era of mass production, the fundamental question is still with us. Fully automated factories without workers are no longer a narrative of science fiction, but a reality in some industries. On the other hand, manufacturing systems that utilize traditional Fordism are still producing massive amounts of inexpensive industrial goods. As a result, the division of labor that first existed within the factory walls is now existing between the advanced industrial and the developed nations. The remainder of this chapter is about the further changes expected to take place in the industrial landscape, as the trajectory of post-Fordism continues along with the rapid rise of digital technologies and automation.

Technological and Business Drivers that Underlie Industrial Landscape Change This section describes important technological and business model change drivers that propel the migration of several industries toward a networked manufacturing model. The networked manufacturing model is and must be accompanied with a networked model of business to support it, where each independent company in the network must have at least break-even profitability for the whole system to work. The backbone of the networked model of manufacturing (and business) is the ability to control and manage the network – this is a more complex endeavor, than controlling and managing of a typical factory, where the activities take place in a highly concentrated setting. Importantly, from the management and decisionmaking point of view, the management of a typical factory is most often optimizing the activities of separate entities, while the management of a network of a digitalized networked company is optimization of a dynamic system that composes of several digitally interconnected subsystems. This digital connectedness carries a potential

56

K. E. Michelsen et al.

for a higher-level system (or supersystem) optimum, while reaching it means solving several difficult problems. Reaching a state, where a manufacturing company can be said to be digitally networked, means mastering and being able to orchestrate many technologies simultaneously that allow the company to manage the network. These technologies are digital components of the digital system that is the company and that can to a large extent all be controlled remotely. In the following subsections important components of a digital network–based manufacturing business model are presented.

Digitalization of Business and Manufacturing Processes Digitalization, that is the implementation of digital technologies to business processes and models (see, e.g., Bharadwaj et al. 2013; Setia et al. 2013), refers to the manual flow of information, whereas automation refers to manual work. On a small scale, digitalization may only mean the storing of data in a digital form, but the true potential of digitalization is reachable through the digital flow of information and knowledge. Here it is important to note that information and knowledge are not the same as data, as information can be understood as a “recipe” to do something and knowledge as the wisdom of understanding when and under which circumstances the recipe should be put to use. Knowledge, in other words, is information put in a context that typically translates to action. In other words, digital transfer of knowledge in the sense that is referred to here is the transfer of “orders” or “requests” from the managing entity to a node, or nodes, in the networked model. What then underlies the ability to relay orders to nodes in a network and what are the nodes? A key revelation is to understand that an autonomously functioning machine, or a robot-operated warehouse, can be a node. Strictly speaking “a human in the loop” is not a necessity. The ability to relay orders is based on the interconnectedness of the nodes and the ability to “give orders” in the format that is understandable to the nodes. These observations may seem trivial when the ability to communicate requests to a single machine, or node, is discussed, but the situation becomes much more complex when the ability to communicate requests to multiple nodes that operate with different underlying technologies (e.g., machines for different purposes, machines from multiple manufacturers) and in different fields (e.g., logistics, warehouses, and manufacturing) is discussed. What becomes important is the ability to create a direct digital communication interface with the nodes in the network, while the traditional way to interface the nodes has been communication between humans through a digital network. The other side of the coin is the ability to receive information from the nodes, to be able to understand what the status of the network is, in real time or close to real time. This means that the nodes and, e.g., the machinery within a node must necessarily be properly instrumented. This instrumentation falls under the Internet of Things (IoT) paradigm that is a necessary component of networked digitalized manufacturing business models. Digitalization, from the point of view of business and manufacturing processess, opens the possibility for systemic control and it is within the potential for efficiency gains through this control that possible

3 Changing Manufacturing Landscape: From a Factory to a Network

57

breakthroughs can be reached. One can say that digitalization (instrumentation and digital communication) are a necessary baseline for a fully digital business model in manufacturing and elsewhere. Novel management technologies have to be adopted to efficiently control the digitalized factory operating in an industrial network. One of the potential technologies that can be adopted also to management use in the context of a digitalized networked (manufacturing) company is the “digital twin,” a concept originally coined in the aviation industry to enhance efficient use of individual airplanes (see, e.g., Ríos et al. 2015). In a broad sense, as Tuegel, Ingraffea, Eason, and Spottswood (2011) define the concept, a digital twin is a set of high-fidelity, multidisciplinary, computer models of unique physical products with their operational history. In essence, this means that a digital counterpart of a physical entity exists in virtual space, which can be applied for the purposes of product design, maintenance optimization, and flexible collaboration between different stakeholders such as industrial customers and solution providers (see, e.g., Negri et al. 2017; Kostis and Ritala 2020). The recent literature (Rosen et al. 2015) has extended the idea of digital twins to cover entire production lines, where the mutual coordination of equipment is needed. For a review of digital twin technology in manufacturing, the interested reader can refer to Holler, Uebernickel, and Brenner (2016). In the context of a networked factory (or manufacturing company), a repository of digital twins that consists of smaller individually modeled manufacturing processes or even single pieces of equipment is of essence. This same fact is already acknowledged by Li (2018) who names digital twins as one of the cornerstones of the national digitalization projects of “Industrie 4.0” in Germany and “Made in China 2025.” Examples of plant-level implementations of a digital twin include the cases presented in Liu, Zhang, Leng, and Chen (2019) and Zhang, Liu, Chen, Zhang, and Leng (2017), who propose a two-level simulation of a manufacturing plant, where a high-fidelity virtual model is constructed and subjected to a dynamic, random simulation tests to study the robustness of the system. Running plantlevel (or smaller) digital twins for the purposes of asset management today faces challenges such as: (i) communication between physical and the virtual systems; (ii) communication between virtual models, especially in cases where more than one model is in place (see Schroeder et al. 2016); and (iii) general availability of data issues (see Kunath and Winkler 2018). These issues, together with the confidentiality of data, are further highlighted, if and when the digital twin networks consist of nodes that are not under the same owner. In fact, it may very well be that the ideas that underlie using digital twins to predictively analyze single machines and to study designs (in the design rather than in the asset management space) may not be fully compatible with what the management or “control” aspect of a networked factory needs. This is due to the “too high fidelity” for the purpose of management. Digital twins are commonly understood as very hi-fi virtual models that typically exist for single machines as envisioned originally (see, e.g., Tuegel et al. 2011). Creating a network of hifi models for management purposes is “overkill” from the point of view of the management needs. A leap in management efficiency can already be reached with much less detail, and only with information that satisfies management needs. This

58

K. E. Michelsen et al.

infers that for network management purposes a more robust meta-level information layer is needed that is used to manage the network of digital twins, rather than using the full information available in the models. First virtual control systems of production facilities with smart interfaces are already in place and they may be the direction into which future control systems for networked manufacturing will go. While the technical solutions for control develop, the discussion about who will be calling the shots and controlling the network as a whole via the tools is entirely another issue and depends on how the network is composed. If the whole network is owned by a single actor the issue may be clear, but when parts of the network are separately owned the relaxation of control decisions are based on trust and a common understanding of profit sharing between the network (node) owners. Further discussion about these issues is left outside the scope of this chapter.

Automation and Manufacturing Robotics Automation and robotics have been around in manufacturing already for some time. In fact, automation can be said to be the beginning of modern industry, as it was the machine automation of cotton mills that is widely perceived as the pivotal turning point that started the first industrial revolution. Those days of automation are, however, very far away from what is here envisioned the future will bring, because modern automation is “full automation” in the sense that manufacturing machines operate autonomously. Rosen et al. (2015) talk about automatic operation and autonomous operation. Modern manufacturing robots can be used to reach a higher level of automation by using autonomous robots that assist and co-work with digitally operated (or autonomous) machines and with humans and they can be used to bridge the gaps between automated parts of manufaturing processes. Such robots are referred to as cobots and their first generation already exists and is operational in the manufacturing industry (Li et al. 2020). Typically cobots are used in warehouses, where these are used to complement and to replace humans working in many sorting and picking tasks. The ability to create autonomously operating manufacturing equipment, including cobots, requires the ability to fuse information from multiple sources (sensors) at any given point in time. Gabor et al. (2016) point to the ability of digital models to simulate data and thus produce planned reactions based on the data at hand, instead of resorting to the use of fixed rule sets. Using machine learning (ML) that does not require user intervention to tailor action rules may provide good answers for autonomous operation and simple examples of using ML in manufacturing already exist, see, e.g., Priore et al. (2006, 2018). Previously using ML has been inhibited by lack of computing power, but today many of the previous restrictions have been lifted due to faster communication ability and due to cloud-based fast computing. Advanced autonomous robots can be given a task (through a digital management system) that they will then fulfill – from the management system point of view autonomous parts in the network can be treated as black boxes, only the completion, or the lack thereof, is relevant information at the high level.

3 Changing Manufacturing Landscape: From a Factory to a Network

59

Additive Manufacturing Technologies Additive manufacturing, often referred to as 3D printing, is a set of manufacturing technologies that are based on the seemingly simple idea of manufacturing solid objects by adding materials layer by layer (Ngo et al. 2018). The material addition technologies are typically based on using liquifiable substances that can be precisely applied, such as resins and plastics, or on using powders (in connection with metals) and melting to attach the powder-based material. Melting metal powders can be done by, e.g., using precision lasers. What makes additive manufacturing very interesting from the point of view of digitalization is that the machinery is typically digital and computer operated. In essence, an additive manufacturing device is a universal manufacturing device that can manufacture any form or shape within the universe of possible shapes (size and complexity) to meet the customer demands, within the limits of the machin – this includes highly complex shapes that are otherwise impossible to manufacture by conventional manufacturing methods. Importantly, additive manufacturing also allows the use of less raw materials than traditional manufacturing methods (see, e.g., Ford and Despeisse 2016; Gebler et al. 2014) – this may be an important issue, especially in cases where the raw material used is expensive. Additive manufacturing opens many avenues for developing and making manufacturing more efficient, more tailored (Chiu and Lin 2016), and more flexible (Achillas et al. 2015). While this is the case, the variety of new business models around additive manufacturing has been so far quite limited (Savolainen and Collan 2020). Being in possession of a fleet of additive manufacturing machines, or “stations,” allows a manufacturing network to flexibly produce a large assortment of components, which allows for a higher level of optimization of the manufacturing capacity use. Additive manufacturing has been shown to be able to exhibit economies of scale (Baumers et al. 2016) and the ability to reach high capacities is based on a combination of skills that include abilities related to manufacturing technology, abilities related to the management of the potentially produced components, and the ability to sell the capacity, or otherwise create a business that is able to utilize the capacity efficiently. Businesses that may be supported and that may rely on additive manufacturing of components on an ondemand basis include the service and maintenance business that relies on the availabilty of spare parts – there is a natural fit with additive manufacturing and maintenance, although several unresolved issues exist (Holmström et al. 2016; Urbani and Collan 2020). Materials research is a field that is closely connected to additive manufacturing and will in the future widen the range of different materials that can be utilized in additive manufacturing. This development allows widening the range of components and architectures manufactured. Advanced materials such as “advanced” metal alloys may allow for additive manufacturing of simple machines, or more precisely, components that have machine-like characteristics. Memory metal alloys, e.g., may be used to manufacture shapes that can be operated with electric current or

60

K. E. Michelsen et al.

magnetic fields and that allow the creation of unknown types of active structures and capabilities in components.

Logistics Optimization and Supply Chain Risk Management Interestingly, additive manufacturing is not only a revolution of manufacturing, but most importantly a revolution of logistics (Bogers et al. 2016). This is due to the fact that the raw materials used by additive manufacturing machinery are standard powders and liquefiable materials in standardized forms, the logistics and storing of which are most often cheaper than those of ready parts – especially from the point of view of the tied-up capital. This also means that a smaller variety of more standardized items in a larger quantity per item will be needed when manufacturing happens by way of additive manufacturing technologies. While additive manufacturing will affect what is being shipped and where, there will most likely not be an abrupt fundamental change in the way global logistics work – rather there will be a gradual change that will start from the change in various niche areas, where additive manufacturing wins ground. For the great majority of products mass production will remain the most efficient and cost-effective way of production (even when logistics costs are added) for a relatively long time (Mellor et al. 2014; Holmström et al. 2016; Savolainen and Collan 2020). This means that what is known about optimizing logistics will be important also in the future, better yet, the importance will hardly be erased by additive manufacturing or any other manufacturing paradigm, as it is not foreseeable that manufacturing could happen without any raw materials. Supply chain management is an important piece of the digital manufacturing puzzle, where just-in-time (JIT) delivery of goods and raw materials is a key issue in the context of inventory management for manufacturing and businesses also beyond manufacturing. The drive toward optimizing (minimizing) inventories via efficient supply chains is also a source of risks, some of which may materialize in situations of sudden shocks caused by, e.g., pandemics, and which may hamper the ability of businesses to operate normally. In such cases, the flexibility offered by additive manufacturing can be used to remedy some of the problems, if the competences and readiness to do so exist. If the logistic chain cannot provide critical components that can be additively manufactured, a fleet of additive manufacturing systems can be turned into producing the critical components. In this way additive manufacturing technologies also work toward more resilient supply chains (Laplume et al. 2016). In a digitalized environment, where supply chain disruptions can be identified early on, firms with additive manufacturing capacity and the skills to use it to dampen the effects of disruptions may be able to gain competitive advantage.

Proactive Versus Reactive Management There is increasing potential for competitive advantage creation in adopting predictive management in manufacturing. The digitalization of equipment in terms

3 Changing Manufacturing Landscape: From a Factory to a Network

61

of higher numbers of sensors and real-time data collection form the basis of implementing the statistical algorithms for predictive management of industrial equipment. In the realm of supply chain management, the use of predictive risk management models in networks is also a topic of growing interest (Hallikas et al. 2002; Seyedan and Mafakheri 2020). Having a predictive management ability means that management becomes proactive in terms of acting to affect an outcome, rather than acting as a reaction to something that has already taken place. A field within manufacturing where proactive management has already shown to create competitive advantage is predictive maintenance (Urbani et al. 2020). Predictive analytics–based proactive management will most likely spread also to other areas of manufacturing. Traditionally, the data-based management of industrial plants has concentrated on the existing sets of cumulative history data that serve as material for tuning predictive analytics. In the case of networked manufacturing, there is no single fixed factory “entity” that can be used for the systematic and systemic analysis of deviations. As there are many possible constellations of a network with multiple complementing nodes that may be used to reach a desired end result it may be impossible to possess complete data for predictive purposes. This calls for a different type of approach for system-level predictive management and finding system-level optimal management actions – one possible direction is the use of simulation in creating ex ante possibility spaces that allow the identification of good system (nework) configurations for various situations. If it is not possible to obtain real-world historical data from the network, a precise-enough model of the network can be constructed and simulation is used to create the needed data. In large networks, the problem complexity may become an issue and further simplification may be required. Techniques such as meta-modeling (Yang et al. 2018) offer simplification opportunities that may be used to simplify complex manufacturing model parts to simpler input-output systems from the practical point of view. In the long run the development toward more automated manufacturing management will change the role of the post-Fordian human worker increasingly toward those of “a maintenance” person who is responsible for carrying out supporting tasks that keep the automated machine running. This may be viewed as a dystopic development, where the value of the human input decreases. On the other hand the construction and planning needed in making the development possible requires the ingenuity and work of humans for the foreseeable future. With the transfer of manufacturing to a more networked mode, also a transfer of the required competences of human resources is taking place.

Conclusion: Toward Manufacturing as a Network What is argued is that the new networked manufacturing model – fueled with new digital technologies and management mechanisms as discussed thus far – calls for viewing modern factory not as isolated nodes connected merely by logistics and supply relationships, but rather as a developing and dynamic network. In such a model, there is an increasing demand for organizational and individual abilities in

62

K. E. Michelsen et al.

collaboration and coordination, in addition to the traditional demands of efficiency and substance skills. Even if the progression takes place at a different pace across the globe, the argument is that we are witnessing a gradual movement toward the manufacturing-as-a-network model. What does manufacturing as a network mean? First, it refers to undertanding the networked factory as a complex adaptive system (e.g., Phillips and Ritala 2019). This calls for viewing networks as “ecosystems”: networks of interdependent actors, technologies, and institutions that together create value (Aarikka-Stenroos and Ritala 2017). Ecosystems often span multiple industries and industrial logics, and they include co-specialization of assets across different complementary actors (Jacobides et al. 2018). From the perspective of manufacturing as a network, this means that independent factory units of different sizes are interdependent via digital interfaces, creating potential for moving toward a seamless web of fully automated industrial landscape that is operating “24/7.” Such loosely coupled system is thus both distinctive and responsive (Orton and Weick 1990), allowing for global, scalable, flexible, and generative model of manufacturing. Second, manufacturing as a network results in the emergence of new types of risks. Interdependence among different actors does not only provide flexibility and scalability to production, but it creates interdependence risks (Adner 2006). For instance, when one “node” in a manufacturing ecosystem fails to operate, it might create bottlenecks to other parts of the system. Such risks can be mitigated by involving slack in the system, e.g., in terms of alternative producers, but on the other hand, modern production is aiming to cut out slack via, e.g., lean and justin-time methodologies. Furthermore, in a large and interconnected ecosystem there are major risks of incentive misalignment. The higher the number of interconnected actors in the system, the higher the chance that the incentives to contribute to that system are not fully aligned as actors adopt different roles. Therefore, the “factory of the future” needs to balance system-related risks against the benefits of flexibility provided by those systems. Third, manufacturing as a network requires new types of management. Unlike traditional coordination of a firm viewed as a nexus of contracts, modern manufacturing networks require “orchestration” – a model of coordination that pursues to accommodate the incentive structure and division of labor among loosely coupled actors (see, e.g., Dhanaraj and Parhke 2006). Orchestration of a complex networks of autonomous machines, technologies, and production elements requires major orchestration capabilities and skills, often possessed by industrial leading actors (Ritala et al. 2009; Hurmelinna-Laukkanen and Nätti 2018). However, as an alternative to a hub-and-spoke model, where leading actors aim to orchestrate the rest of the network, more collective forms of governance could also be adopted (Fjeldstad et al. 2012). Finally, there are major social and institutional implications. As already mentioned in the introduction, the manual work is increasingly moving to the automated processes and robots in the networked manufacturing model. This means that the industrial work moves from supervising machine operations and production lines toward overseeing the automated systems and training machine learning algorithms,

3 Changing Manufacturing Landscape: From a Factory to a Network

63

for instance. Furthermore, there is also continuing demand for managerial skills in orchestrating (see, e.g., Ritala et al. 2009) networks of actors, and developing overarching undertanding of how the factories’ and production units roles are best positioned in a global production network. These progressions mean that the decreasing demand for manual labor is going to continue, including a “race-to-thebottom” in wages, while the remaining positions might increasingly go to highly paid experts. To conclude, in this chapter, it is noted that it has been demonstrated how the changing industrial landcape has moved from Taylorist and Fordist effectiveness and standardization-based models toward a post-Fordist networked industry model. This change has been fueled by the surge of new digital and manufacturing technologies and by the related business models. There are major implications of adoption of these technologies and resulting ways of industrial production and work. Table 1 summarizes the arguments about how manufacturing as a network differs from the Taylorist-Fordian view of a factory (manufacturing). Going forward, one important question is, whether we will witness a move from loosely coupled industrial governance, where different producers compete against

Table 1 Shift of paradigm from Taylorist-Fordian factory to post-Fordian networked manufacturing

Manufacturing takes place

Underlying business logic Ideal management model

Taylorist-Fordian view of factory In a highly specialized, efficient, and optimized production unit as a part of global value chain(s). The operation is constrained by demand Simultaneously scale production volume and improve quality of the outputs Coordination and control

Role of machinery and robotics The main type of manual work

Automatic, fixed, and manual mobile Supervising fixed machine operations

Payment for manual work

High wages ensured by national labor unions

Risks

Demand-related risks, declining competitivity

Manufacturing as a network In an industrial network, which acts a flexibility resource within the constraints of economic profitability

Increase production scope and output quality to gain valuable roles in the network Orchestration of a loosely coupled system; collective autonomous governance Autonomous machinery and mobile cobots Overseeing and developing the automated production systems; supporting continuous operation of machines Global competition and race to the bottom in low-tech wages; highly skilled workers benefit Interdependence risks, replaceability risks

64

K. E. Michelsen et al.

each other with their networks and ecosystems toward a tightly coupled production model. In principle, a tightly coupled system with fully automated supply and demand, logistics, and manufacturing could be an ideal way for the global networked manufacturing. This type of a system would minimize slack, waste, and risks, while maximizing output. However, if such system is highly centralized among one or few global technology companies, there is risk that the efficiency gains are reaped as monopoly rents. This would leave less room for competition, innovation, and diversity. Therefore, it is expected that manufacturing of the future is run by global “manufacturing networks of the future” – networked production systems that compete head to head with automation, flexibility, and quality. To conclude, this chapter helps to track the progression in the industrial landscape and the changing role of a “factory” from Taylorist and Fordian history toward the networked, automated, and interconnected future of manufacturing. The arguments made are based on the existing evidence and progression, but of course the future development can take alternative paths that have not been taken into account. Therefore, there is a place for more research that would study the changing aspects in the industrial landscape from different perspectives.

References Aarikka-Stenroos L, Ritala P (2017) Network management in the era of ecosystems: systematic review and management framework. Ind Mark Manag 67:23–36 Achillas C, Aidonis D, Iakovou E, Thymianidis M, Tzetzis D (2015) A methodological framework for the inclusion of modern additive manufacturing into the production portfolio of a focused factory. J Manuf Syst 37:328–339. https://doi.org/10.1016/j.jmsy.2014.07.014 Adner R (2006) Match your innovation strategy to your innovation ecosystem. Harv Bus Rev 84(4):98–107 Amin A (1994) Post-Fordism: models, fantasies and phantoms of transition. In: Amin A (ed) PostFordism. A Reader. Blackwell, Oxford, UK, Cambridge, MA, pp 1–40 Baumers M, Dickens P, Tuck C, Hague R (2016) The cost of additive manufacturing: machine productivity, economies of scale and technology-push. Technol Forecast Soc Chang 102:193– 201. https://doi.org/10.1016/j.techfore.2015.02.015 Bharadwaj A, El Sawy OA, Pavlou PA, Venkatraman N (2013) Digital business strategy: toward a next generation of insights. MIS Q 37:471–482 Bogers M, Hadar R, Bilberg A (2016) Additive manufacturing for consumer-centric business models: implications for supply chains in consumer goods manufacturing. Technol Forecast Soc Chang 102:225–239. https://doi.org/10.1016/j.techfore.2015.07.024 Boyer R, Durand J-P (1993) After Fordism. Macmillan Business, London Chandler AD (1990) Scale and scope. The dynamics of industrial capitalism. The Belknap Press of Harvard University Press, Cambridge, MA Chiu M, Lin Y (2016) Simulation based method considering design for additive manufacturing and supply chain: an empirical study of lamp industry. Ind Manag Data Syst 116(2):322–348. https://doi.org/10.1108/IMDS-07-2015-0266 Dhanaraj C, Parkhe A (2006) Orchestrating innovation networks. Acad Manag Rev 31(3):659–669 Fjeldstad ØD, Snow CC, Miles RE, Lettl C (2012) The architecture of collaboration. Strateg Manag J 33(6):734–750 Ford S, Despeisse M (2016) Additive manufacturing and sustainability: an exploratory study of the advantages and challenges. J Clean Prod 137:1573–1587. https://doi.org/10.1016/ j.jclepro.2016.04.150

3 Changing Manufacturing Landscape: From a Factory to a Network

65

Gabor T, Belzner L, Kiermeier M, Beck MT, Neitz A (2016) A simulation-based architecture for smart cyber-physical systems. In 2016 IEEE international conference on autonomic computing (ICAC), pp 374–379 Gambao EC (1988) Globalization of industry through production sharing. In Muroyama JH, Guyford H (eds) Globalization of technology. International Perspective. Council of Academies of Engineering and Technological Sciences. The National Academy Press, pp 86–105. https:// doi.org/10.17226/1101 Gebler M, Schoot Uiterkamp AJM, Visser C (2014) A global sustainability perspective on 3D printing technologies. Energy Policy 74(C):158–167. https://doi.org/10.1016/ j.enpol.2014.08.033 Geels FW (2004) From sectoral systems of innovation to socio-technical systems: insights about dynamics and change from sociology and institutional theory. Res Policy 33(6–7):897–920 Hallikas J, Virolainen V-M, Tuominen M (2002) Risk analysis and assessment in network environments: a dyadic case study. Int J Prod Econ 78(1):45–55 Holler M, Uebernickel F, Brenner W (2016) Digital twin concepts in manufacturing industries – a literature review and avenues for further research. In Proceedings of the 18 international conference on industrial engineering (IJIE), p 9 Holmström J, Holweg M, Khajavi SH, Partanen J (2016) The direct digital manufacturing (r)evolution: definition of a research agenda. Oper Manag Res 9(1–2):1–10. https://doi.org/ 10.1007/s12063-016-0106-z Hounshell DA (1984) From American system to mass production, 1800–1932. The development of manufacturing Technology in the United States. Johns Hopkins University Press, Baltimore Hurmelinna-Laukkanen P, Nätti S (2018) Orchestrator types, roles and capabilities–a framework for innovation networks. Ind Mark Manag 74:65–78 Jacobides MG, Cennamo C, Gawer A (2018) Towards a theory of ecosystems. Strateg Manag J 39(8):2255–2276 Jessop B (1992) Fordism and Post-Fordism: a critical reformulation. In: Storper M, Scott AJ (eds) Pathways to industrialization and regional develoment. Routledge, London, New York, pp 42–61 Jürgens U (1989) The transfer of Japanese management concepts in the international automobile industry. In: Wood S (ed) The transformation of work. Unwin and Hyman, London Kern and Schaumann (1987), Limits of the Division of Labour. New Production and Employment Concepts in West German Industry, First Published May 1, 1987in Economic and Industrial Democracy. 8(2):151–170. Kostis A, Ritala P (2020) Digital artifacts in industrial co-creation: how to use VR technology to bridge the provider-customer boundary. Calif Manag Rev 62(4):125–147 Kunath M, Winkler H (2018) Integrating the digital twin of the manufacturing system into a decision support system for improving the order management process. Proc CIRP 72:225–231. https://doi.org/10.1016/j.procir.2018.03.192 Laplume A, Petersen B, Pearce JM (2016) Global value chains from a 3D printing perspective. J Int Bus Stud 47(5):595–609. https://doi.org/10.1057/jibs.2015.47 Li L (2018) China’s manufacturing locus in 2025: with a comparison of “Made-in-China 2025” and “Industry 4.0”. Technol Forecast Soc Chang 135:66–74. https://doi.org/10.1016/ j.techfore.2017.05.028 Li M, Milojevic A, Handroos H (2020) Robotics in manufacturing – the past and the present. In: Collan M, Michelsen K-E (eds) Technical, economic and societal effects of manufacturing 4.0. Palgrave Macmillan, Cham, pp 97–111 Lipietz A (1985) The world crises: the globalization of the general crises of Fordism. IDS Bull 16(2):6–11. Institute of Development Studies. Sussex Liu Q, Zhang H, Leng J, Chen X (2019) Digital twin-driven rapid individualised designing of automated flow-shop manufacturing system. Int J Prod Res 57. https://doi.org/10.1080/ 00207543.2018.1471243. Taylor & Francis Lüthje B (2013) Why no Fordism in China? Regimes of accumulation and regimes of product ion in Chinese manufacturing industries. IFS, Working Paper #3. Novermber 2013. Institute of Socialforschnung, Frankfurt am Main

66

K. E. Michelsen et al.

Macdonald M (1991) Post-Fordism and the flexibility debate, Post-Fordism and the flexibility debate. Stud Polit Econ 36(1):177–201. https://doi.org/10.1080/19187033.1991.11675447 Mellor S, Hao L, Zhang D (2014) Additive manufacturing: a framework for implementation. Int J Prod Econ 149:194–201. https://doi.org/10.1016/j.ijpe.2013.07.008 Negri E, Fumagalli L, Macchi M (2017) A review of the roles of digital twin in CPS-based production systems. Proc Manuf 11:939–948. https://doi.org/10.1016/j.promfg.2017.07.198 Ngo TD, Kashani A, Imbalzano G, Nguyen KTQ, Hui D (2018) Additive manufacturing (3D printing): a review of materials, methods, applications and challenges. Compos Part B 143:172– 196. https://doi.org/10.1016/j.compositesb.2018.02.012 Orton JD, Weick KE (1990) Loosely coupled systems: a reconceptualization. Acad Manag Rev 15(2):203–223 Phillips MA, Ritala P (2019) A complex adaptive systems agenda for ecosystem research methodology. Technol Forecast Soc Chang 148:119739 Pine JB (1980) Paridgm sift: from mass production to mass customization. M.A. Thesis. Massachusetts Institute of Technology, June 1991. dspace.mit.edu Priore P, de la Fuente D, Puente J, Parreño J (2006) A comparison of machine-learning algorithms for dynamic scheduling of flexible manufacturing systems. Eng Appl Artif Intell 19(3):247–255. https://doi.org/10.1016/j.engappai.2005.09.009 Priore P, Ponte B, Puente J, Gómez A (2018) Learning-based scheduling of flexible manufacturing systems using ensemble methods. Comput Ind Eng 126:282–291. https://doi.org/10.1016/ j.cie.2018.09.034 Ríos J, Hernández JC, Oliva M, Mas F (2015) Product avatar as digital counterpart of a physical individual product: literature review and implications in an aircraft. In Advances in transdisciplinary engineering, pp 657–666. https://doi.org/10.3233/978-1-61499-544-9-657 Ritala P, Armila L, Blomqvist K (2009) Innovation orchestration capability—defining the organizational and individual level determinants. Int J Innov Manag 13(04):569–591 Rosen R, von Wichert G, Lo G, Bettenhausen KD (2015) About the importance of autonomy and digital twins for the future of manufacturing. IFAC-PapersOnLine 48(3):567–572. https:// doi.org/10.1016/j.ifacol.2015.06.141 Savolainen J, Collan M (2020) How additive manufacturing technology changes business models? – review of literature. Addit Manuf 32:101070. https://doi.org/10.1016/ j.addma.2020.101070 Sayer A (1986) New developments in manufacturing: the just in time system. Capital Class 30:43–72 Schroeder GN, Steinmetz C, Pereira CE, Espindola DB (2016) Digital twin data modeling with AutomationML and a communication methodology for data exchange. IFAC-PapersOnLine 49(30):12–17. https://doi.org/10.1016/j.ifacol.2016.11.115 Schumann M (1998) New concept of production and productivity. Econ Ind Democr 19:17–32 Setia P, Setia P, Venkatesh V, Joglekar S (2013) Leveraging digital technologies: how information quality leads to localized capabilities and customer service performance. MIS Q 37:565–590 Seyedan M, Mafakheri F (2020) Predictive big data analytics for supply chain demand forecasting: methods, applications, and research opportunities. J Big Data 7:53 Taylor FW (1911) Principles of scientific management. Harper & Brothers Publishers, New York, London Thompson CB (1916) Relation of scientific management to labor. Q J Econ 30(2):311–351. JSTOR, www.jstor.org/stable/1884997 Tomaney J (1994) A new paradigm of work. Organization and technology? In: Amin A (ed) PostFordism. A Reader. Blackwell, Oxford, Cambridge, MA, pp 156–194 Tuegel EJ, Ingraffea AR, Eason TG, Spottswood MS (2011) Reengineering aircraft structural life prediction using a digital twin. Int J Aerospace Eng 2011:1–14. https://doi.org/10.1155/2011/ 154798 Urbani M, Collan M (2020) Additive manufacturing cases and a vision for predictive analytics and additive manufacturing based maintenance business model. In: Collan M, Michelsen K-E (eds) Technical, economic and societal effects of manufacturing 4.0. Palgrave Macmillan, Cham, pp 131–148

3 Changing Manufacturing Landscape: From a Factory to a Network

67

Urbani M, Petri D, Brunelli M, Collan M (2020) Maintenance management in light of manufacturing 4.0. In: Collan M, Michelsen K-E (eds) Technical, economic and societal effects of manufacturing 4.0. Palgrave Macmillan, Cham, pp 97–111 Walker R, Buck D (2007) The Chinese road -cities in the transition to capitalism. New Left Rev 46:39–66 Wang W, Siau K (2019) Artificial intelligence, machine learning, automation, robotics, future of work and future of humanity: a review and research agenda. J Data Manage 30(1):61–79 Warring SP (1991) Taylorism transformed: scientific management theory since 1945. University of North Carolina Press, Chapel Hill Yang Z, Eddy D, Krisnamurty S, Grosse I, Denno P, Witherell P, Lopez F (2018) Dynamic metamodeling for predictive analytics in advanced manufacturing. Smart Sustain Manufact Syst 2(1):18–39 Zhang H, Liu Q, Chen X, Zhang D, Leng J (2017) A digital twin-based approach for designing and multi-objective optimization of hollow glass production line. IEEE Access 5:26901–26911. https://doi.org/10.1109/ACCESS.2017.2766453

Further Reading https://new.siemens.com/global/en/company/stories/research-technologies/digitaltwin/digitaltwin.html#main-content https://www.ge.com/digital/applications/digital-twin https://www.konecranes.com/resources/digital-twins-new-standard-in-industrial-production https://www.bcg.com/publications/2019/emerging-art-ecosystem-management https://www.siliconrepublic.com/machines/automated-factories-video https://www.visualcapitalist.com/supply-chains-automation-future/ https://www.3yourmind.com/news/5-keys-to-success-in-a-distributed-manufacturing-model https://www.youtube.com/watch?v=nMdjexWNyg4 https://www.youtube.com/watch?v=caTCRVt77s8 https://www.youtube.com/watch?v=jYby_HczyDA https://www.youtube.com/watch?v=x8H1x_OoZBI

4

Karakuri Solutions and Industry 4.0 Mariusz Kostrzewski and Wojciech Jerzy Nowak

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Where to Find Karakuri? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Karakuri Kaizen in Industrial Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selected Reflections on Karakuri . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

70 80 82 90 91 93 96

Abstract

In Japanese, karakuri means “a mechanism,” “a contraption,” or “a trick.” The word karakuri often refers to karakuri ningy¯o (automata), constructed in Japan between seventeenth and nineteenth century, inspired by mechanical devices such as time-measuring mechanisms (Yamamoto et al (2018a) Karakuri IoT – the concept and the result of pre-study. In: Thorvald P, Case K (eds) Advances in manufacturing technology, vol 8. IOS Press, Amsterdam, pp 311– 316). Currently, examples of karakuri, from the Edo period (1603–1867), are an inspiration to various solutions known under collective term karakuri kaizen. These low-cost automations are improvement projects that concentrate on the development of mechanical devices that employ gravity to improve variM. Kostrzewski () Faculty of Transport, Division of Construction Fundamentals of Transport Equipment, Warsaw University of Technology, Warsaw, Poland e-mail: [email protected] W. J. Nowak Faculty of Humanities, Institute of Literary Studies, Department of Oriental Studies, Nicolaus Copernicus University, Toru´n, Poland e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_124

69

70

M. Kostrzewski and W. J. Nowak

ous industrial systems, such as technological transport systems or production systems. These low-cost automations use no electrical, pneumatic, or hydraulic equipment and are not connected to computers, IT networks or any other elements of the Industry 4.0 paradigm. Therefore, karakuri kaizen is sometimes treated as development and implementation phase prior to introduction of Industry 4.0. However, selected solutions, based on karakuri principles, work more effectively than complex and expensive, digital ones. Karakuri happens to be used in industry counter to solutions of Industry 4.0, because of their analogue nature. The first section of the chapter consists of introduction of selected definitions (e.g., karakuri, Industry 4.0) and a brief overview of premodern Japanese clockwork and automata history with description of certain specimens relevant to the authors’ study. In the second section, the authors present analysis of interdisciplinary literature search based on renown scientific databases. The third section focuses on selected solutions of karakuri kaizen in industrial applications. In the chapter the authors take a challenge to present the less known, early pages of the story of last-decade Industry 4.0. Therefore, the fourth section is connected to some reflection on the subject matter. The chapter is finalized with conclusion and proposals for future research. Keywords

Karakuri · Karakuri ningy¯o · Kaizen · Japan · Roller conveyor · Natural force · Industry 4.0 · Logistics 4.0

Introduction Robotics and automation stipulated as crucial elements of Industry 4.0 have become widespread and associated with digital technologies. In the initial section of the chapter, the authors would like to attempt to present the analog beginnings of robotics, on example of karakuri ningy¯o, Japanese premodern automata, focusing on how one could benefit from this technology and what solution can those primal robots provide nowadays, in new context of Industry 4.0 and challenges of twentyfirst century. Since the time immemorial, the technology of contraptions served various purposes. It was an attempt to understand the mechanisms of the world, by mimicking its principles in mechanisms representing it. One of the prime examples of such approach are clocks, on the other hand the technology was an output of the human drive to create. In terms of inventions, actual or mythical, one of the most frequently recurring topics is creating a life-like contraption, an artificial human. The mythical living dolls found their way into fairy tales like Vasiilissa the Beautiful helper-doll, Pinocchio, golem of Prague (Arsénio et al. 2011). In the dawn of twentieth century, during the time of second industrial revolution, the divine

4 Karakuri Solutions and Industry 4.0

71

magic powering those man-made entities was replaced by science. In 1920 Karol ˇ Capek brought to life a name for those machines: a robot, from Slavic robota – labor ˇ (Capek 2008), the name still being used today. The idea of human-like machine was followed by Fritz Lang, director of the motion picture “Metropolis” with Maschinenmensch (Machine-Person) Maria (Lang 1927). Although these robots existed only as a fiction, where did the Western creators, unbound by technical limitations of technology, draw inspiration from? “The Nightingale,” a story by Hans Christian Andersen, gives the reader a hint in a passage describing a parcel delivered to the Emperor of China: “Made of metal with a key to wind it up, this bird was clockwork. But what a magnificent clockwork bird it was, covered in diamonds, rubies and sapphires with a tail of silver and gold. Round its neck was a piece of ribbon with a note. ‘A nightingale from the Emperor of Japan – though it cannot compare with your nightingale,’ the note read” (BBC 2010). Although the clockwork nightingale from the Andersen’s story was fictitious, there is an actual contraption which could inspire the writer, called the Peacock clock, created by English clock master James Cox in 1777 on commission by Prince Grigory Aleksandrovich Potemkin-Tavrichesky as a gift to Russian Empress Katherine II the Great (Pyatnitsky 2017). Now the Peacock clock is a part of Hermitage Museum collection. According to Pyatnitsky (2017), the Peacock clock combines four independent mechanisms: the owl, the peacock, the cockerel, and the mushroom clock. Before considering the main topic of this chapter, it is worthwhile to investigate the progenitors of robots: the automata. In the chapter, the authors decided to narrow their scope of research mainly to Japanese unassisted automata from seventeenth century onwards. The Cartesian idea of animals as mechanisms was literally embodied in the complex mechanical devices, mimicking nature and humans called automata. As the name suggests, they were constructed, as they would act of their own will, facilitating intricate clockwork techniques. The most sophisticated, still extant examples of Western automata portraying humans are Maillardet’s automaton built in London by a Swiss mechanician Henri Maillardet around 1800 (Arsénio et al. 2011), currently in The Franklin Institute of Philadelphia collection and Jaquet-Droz automata in Musée d’Art et d’Histoire of Neuchâtel, in Switzerland, constructed by Swiss clockmakers and mechanics: Pierre Jaquet-Droz, his son Henri Louis and Jean-Frédéric Leschot (Voskuhl 2007). These three automata, namely, the writer (l’écrivain), the draughtsman (le dessinateur), and the woman musician (la musicienne), were introduced to the audience in 1774. Each automaton represents a human figure, performing complex task of either writing, sketching, or playing an instrument. The choice of these tasks will be relevant to the following parts of the chapter. These automata development can be treated as one of the milestones of the humankind history of technology and industrialization. However, it was not the European automata themselves, but the clockworks, which directly influenced the construction of Japanese karakuri from seventeenth century on. The authors

72

M. Kostrzewski and W. J. Nowak

discuss this matter in greater detail later into the chapter. Meanwhile, the following paragraph focuses on the modern times. Most of human life, focused on agriculture, commerce, education, health care, warfare, and industry, depends on relations between humans, society, mechanisms, and increasingly to machines. As a consequence, the technological development is causing a growing appetite for all kinds of energy used to drive machines, digital mechanisms, and devices. As it might be observed in Fig. 1, the energy growth increases with technological development understood here as growth in number of transistors per microprocessors (such a measure for technological development was taken into consideration due to the argument that solutions of Industry 4.0 are highly depended on the computing power). It is no longer a foregone conclusion whether or not the reserves of energy will run out, but when they will. Therefore, it seems necessary to reach more intensively for renewable energy sources, which is a trivial statement nowadays, and it is much more important to reach for solutions that do not require an enormous energy input. The statement, despite being of trivial nature, needs hansei. This Japanese term stands for “self-reflection,” while statement connected to low energy input may be connected to another Japanese term, namely karakuri. The word karakuri has multiple notations in kanji characters (logographic characters of Chinese origin used in the writing system of Japanese language), being: 絡繰, 唐繰, 機巧, 機関, 機, 械, 関 (read the same in every case, as karakuri). The Great Dictionary of Japanese, “Kokugo daijiten” states that the word derives from a verb karakuru which means “to make something move by pulling a string.” Such use of the term has been confirmed around the second part of sixteenth century.

Fig. 1 Comparison of number of transistors per microprocessor (data given in billions on the right ordinate; columns) and global energy consumption (given along left ordinate; data given by solid line include energy acquired from the following sources: coal, solar, oil, gas, traditional biomass, other renewables, hydropower, nuclear, wind, biofuels). (Sources: Roser and Ritchie (2013) for technological development (and additionally number of transistors per microprocessor for 2019 year based on Broekhuijsen 2019; Mujtaba 2019), Ritchie (2014) for global energy consumption)

4 Karakuri Solutions and Industry 4.0

73

In the chapter, the authors focus on karakuri ningy¯o, mechanical dolls (as a reference to modern karakuri-based solutions), which hereafter will be referred to as karakuri. Karakuri are animated, mechanical dolls of humans and animals, which use the principles of clockwork-based mechanisms in performing designated tasks, described in greater detail further in the chapter. As opposed to abovementioned Western automata, Japanese karakuri mechanisms are often composed of wood instead of metal. They do not rely on external power source, but are driven by mechanical forces, provided by springs, or movement of sand, or water inside them. In his chapter, Suzuki (1988, p. 42) proposed the following division of karakuri into four groups: • Karakuri used in religious worship • Karakuri designed to be shown to general audience • Karakuri designed for personal use, which are further divided into those operated manually, or by means of mechanism concealed in them. The examples of karakuri for each of those groups will be discussed in detail further in the chapter. The karakuri mechanisms were accompanied with and inspired by clockworks devices (Yamamoto et al. 2018a). Karakuri ningy¯o has been present in Japan before seventeenth century (Yokota 2009), yet their heyday was Edo period (1603–1867), when newly acquired knowledge on technology, including clock making, imported to Japan from the West provided new possibilities for their further development and widespread recognition (Yokota 2009). The necessary know-how was acquired based on experience gained by Japanese during thorough examinations and repairs conducted on imported clocks. It is worth mentioning after Koch (2020) that unlike in areas of European influence, the mechanical clocks cannot be characterized by the same “revolutionary” effect on time-reckoning in Japan. Yet, the mechanical clocks contributed to development of other sophisticated devices. The oldest extant clock of European origin on Japanese land is a gift of king of Spain, Philip III, given to Tokugawa Ieyasu (the founder of the Tokugawa shogunate of Japan) in 1611 as a token of gratitude for rescuing the stranded Spanish ship 2 years before (Koch 2020). This so-called Western clock, in Japanese y¯odokei, is being preserved at Kun¯ozan T¯osh¯og¯u shrine in Shizuoka City, Shizuoka Prefecture (Koch 2020). This spring-driven clock was made in Madrid in 1581 (Koch 2020). It is worth noticing the evident division of Western clocks from indigenous clocks, called wadokei. The father of Japanese clocks is considered to be Tsuda Sukezaemon (Yokota 2009). The mentioned division between Western and Japanese clockworks is based on the necessity of adjusting the existing mechanism to Japanese approach on time division at the time. Japanese have shown many times their skill in adapting foreign technologies and ideas, as well as their sophisticated refinement for their own domestic use. Wadokei clockwork mechanisms are examples of such adaptations and refinements. Although Tsuda is considered to be the father of Japanese clock, the title of the greatest wadokei designer belongs to Tanaka Hisashige (Yokota 2009). Tanaka

74

M. Kostrzewski and W. J. Nowak

was renown inventor, famous for creating both convenient solutions and numerous automata, some of which the authors describe in detail further in the chapter. This area of his work gained him a nickname Karakuri Giemon, which means the Master of Automata. In Japan his works are widely recognized as groundbreaking contribution to the country’s technical development. His life and achievements are presented in Toshiba Mirai Kagakukan in Tokyo (Toshiba Science Museum 2021). However, the impact of Tanaka’s talent on Japan and the country’s development is undisputed, and the connections of the Master of Automata to foreign countries require further investigation. Andersen’s Nightingale mentioned in the introduction to the chapter presents an interesting lead, when one takes into consideration, that Tanaka, as great creators of automatons in Europe (Mayson 2000, p. 14), also constructed a bird automaton, Hototogisu (the Cuckoo) in 1856 (Toshiba Science Museum 2021). The prime achievement of Tanaka is the so-called Man-nen jimeish¯o (10,000 years self-ringing bell) or simply Man-nen dokei (10,000 years clock), which is described as perpetual chronometer by Hato et al. (2007). He constructed this masterpiece in 1851. The apparatus is said to run for almost a year (225 days to be precise) on a single winding. In 2004, a restoration project aiming at bringing the clock back to the working condition and creating a replica was launched. The project proved to be a success. The results were presented in 2005 during Aichi Expo. Year after Man-nen jimeish¯o was designated as an important cultural property (Yokota 2009, p. 180). Man-nen dokei is of particular importance, because constructing it over the period of 3 years, Tanaka has drawn from his previous experiences of creating karakuri, as well as benefited from western technology, albeit his work definitely was not a copy, but a genuine invention, tailored to local demands. Such approach is reflected in the design of the clock, which has been described in detail by Hato et al. (2007), Yokota (2009), and others. The authors would like to narrow the description to the following features of Man-nen dokei: the six clock faces and the dome. Each of the faces were displaying different time measurement: • First face was of a wadokei, explained in greater detail below. • Second face: solar term with plates for inscriptions of the East Asian lunisolar calendar. • Third face: 7 days of the week. • Fourth face: jikkan j¯unishi which is sexagenary cycle of Chinese Astrological Calendar. • Fifth face: lunisolar calendar and indicator of phases of the Moon. • Sixth face: western style clock face. • Dome with map of Japan showing the position of Sun and Moon during the year. Such construction resembles the mechanism of European astronomical clocks, but on a smaller scale. Miniaturization proved to be a skill very profitable for Japanese industry in the second part of twentieth century onwards.

4 Karakuri Solutions and Industry 4.0

75

The first, wadokei face is the most complex and intricate mechanism. The challenge in this case was to design mechanism that would reflect the irregular measurements of the time, based on the varying length of the day from sunrise to sunset, both indicated on the face of the clock, during the year. It was achieved by fixing the hand indicating hourly units (the hand pointing to lunisolar calendar units remained mobile), and employing the mechanism of a rotating warigoma – 12 motile plates, serving as hour indicators, in 24-h cycle combined with system of changing warigoma positions against each other (placing them closer or further apart) in the yearly cycle, reflecting the alternations in duration of daytime. In the authors opinion this is an example of an early, groundbreaking design merging technology with the world Nature, industrial and environmental, organic approaches. Man-nen dokei, Tanaka’s opus magnum, is a result of its makers own practice and experience. However, even those without such extensive background had the possibility to take a chance on at least re-creating karakuri designs. It is worth noticing that due to high level of literacy in eighteenth century Japan (Kanaya 1979), the indirect transfer of knowledge by means of literature has been possible. From among available primary resources the authors would like to bring to readers’ attention the following: • “Karakuri Kinm¯o Kagami Gusa” (“Notes on the review of automata for beginners”) by Kanch¯usen Tagaya, published in 1730 (Tagaya 1730) • “Kik¯o zui” (“Illustrated compendium of Automata”) by Hosokawa Yarinao, published in 1798 (Hosokawa 1796a, b) ¯ • “Okarakuri e-zukushi” (“Almanach of great automata in pictures”) by Nishimura Shigenaga, published in 1758 (Nishimura 1758) • “Karakuri chigusa no mibae” (“Fruition of thousand types of automata”); author and date of publication unknown (Anonymous) The latter two resources present the variety of karakuri descriptions, the former two serve as open-source manuals on how to construct chosen karakuri and describe how they work. The Japanese approach to automata was more egalitarian than the Western one. Not all automata were concealed in chambers of the noblemen but presented to the wider audience during festivals and theatrical performances. The most notable ¯ theater group utilizing karakuri ningy¯o was Takeda Omi puppet theatre, falling under the term of karakuri shown to the general audience, also described as shibai karakuri, performance automata (Tajima 1983; Yokota 2009). Takeda started his theatre company in 1662 (Tajima 1983; Yokota 2009), and gained enormous popularity in several regions of the country. The One of the proofs for his popularity is highlighted and mentioned in “Settsu meisho zue” (“Depictions of famous places in Settsu province”), a series of books published between 1796 and 1798, a depiction of the performance with Dutch in the audience, accompanied by a comment, which reads “The Dutch were in awe seeing Takeda’s puppet theater” (printed in volume 4 of “Settsu meisho zue”; three Dutch gentlemen presented on the right page and

76

M. Kostrzewski and W. J. Nowak

shibai karakuri on the left page: Akisato and Takehara 1796–1798). Also, whole series of books depicting, and advertising, the performances (see Nishimura 1758, mentioned above) were published on Takeda’s karakuri 1 year after the memorable performance in Edo (present Tokyo). Even after the craze of Takeda’s theater has ceased, its influence of the karakuri technology and techniques on traditional puppet theatre bunraku and kabuki theatre are undebatable. These can be observed in the construction of bunraku puppets as well as in stage contraptions of kabuki theatre, including revolving stage and variety of trap doors, which contribute to creating a momentum in the story presented on stage. It is worth mentioning that shibai karakuri performance were supported with fusuma karakuri (d¯ogugaeshi) – a scenography background or scene-changing elements in performances, equipped with mechanical devices that ensured three-dimensional depth of a scene. In later years a phenomenon of nozoki karakuri peek-in automata has emerged, yet in this case interchangeable plates observed through a lens were used, creating a set similar to European Kaiser-Panorama. The nozoki karakuri device was called also as Oranda megane (Dutch glasses/spectacles) or nozoki megane (peeping glasses), whereas the plates were also known as karakuri-e (tricky picture(s)), Toshiyuki (1998). Another type of karakuri designed for public display are kairaishi ningy¯o (kairaishi means a puppeteer), which were part of the festival floats. The most renown festival employing karakuri ningy¯o is Inuyama Festival. The use of karakuri dates back to 1641, and has been recognized as a part of world intangible cultural heritage by UNESCO (Inuyama Festival 2016). These types of karakuri were also animated by group of people operating sets of strings, connected to certain actions performed by the doll, such as blinking and moving limbs. Placing the automata atop the floats, resembling high-towered cards, enabled the puppeteers to remain unseen by the public, Lancashire (2011, p. 195). The last group is called zashiki karakuri, which means chamber automata (Rani et al. 2015; Bhanu and Kumar 2018; Paraponiaris and Rodríguez 2019). These types of karakuri, powered by wind-up spring, were the most intricate ones, despite their small size, and that is why the authors would like to present detailed descriptions of a few examples: yumihiki d¯oji (young archer), moji kaki ningy¯o (writing doll), and ocha hakobi ningy¯o (tea serving doll) (Yokota 2009). Yumihiki d¯oji consists of a puppet of an archer boy dressed in historical costume sitting atop a box concealing a wind-up mechanism, and a shooting target, placed at a distance from the doll, at which the archer aims. Equipped with three arrows, the boy collects them, one by one, from the stand and shoots them toward the shooting target. Due to the use of strings, the movements of the doll are much smoother than its European counterparts, creating more life-like illusion. Also, the design of the archer face, adopted from traditional Japanese dolls, as well as referring to the art of n¯o theater masks, enabled creating a sense of different facial expressions. An interesting idea is the design making the archer miss the mark on one the attempts. From the point of view of user experience, it is interesting to see the machine fail, and thus familiarize with it more. Two specimens of historical yumihiki d¯oji exist to this day, both created by Tanaka Hisashige, mentioned earlier in the chapter,

4 Karakuri Solutions and Industry 4.0

77

clock-master and inventor, whose expertise in engineering contributed to the rapid modernization of Japan (Yokota 2009). The second mechanism worth mentioning is moji kaki ningy¯o. This karakuri uses a brush dipped in ink to write Chinese characters on a piece of paper. The idea behind this automaton is similar to the European writers, yet, the features of Chinese scripture, including differences in applied pressure and speed of placing the strokes of brush on the paper, make the writing mechanism more intricate (comparing for example to writing boy automaton constructed by Pierre Jaquet-Droz, Henri Louis and Jean-Frédéric Leschot). To make the movement of limbs and the head less rigid, the mechanism utilizes strings with weights, which ensure smooth, rounded motion. In total, four specimens of moji kaki ningy¯o constructed in nineteenth century exist. One of them, retrieved to Japan from the USA, is attributed to Tanaka Hisashige. Another one, holding two brushes (one in hand and one in mouth), was discovered in 2009 in an old estate in present Nagoya (Suematsu 2012). It is attributed to Tamaya Sh¯obei, a karakuri master whose family still produces karakuri automata (Paraponiaris and Rodríguez 2019). The next worth mentioning karakuri is ocha hakobi ningy¯o. This variety of karakuri is of the highest interest for the authors of the chapter, because, as opposed to previous examples, this karakuri is not stationary, but it moves in space and requires human interaction. Ocha hakobi ningy¯o means tea serving doll, and that is what this automaton is designed to realize both in terms of form and function. All of four specimens described in Suzuki (1988) are designed to look like drink-fetching helpers of the era: boys and girls in traditional hairstyles and clothing, holding a tray with a cup in front of them. After placing a cup filled with tea on a tray being held by the doll, it drives toward a “customer” and stops, when the person takes the cup. When an empty cup is placed back onto the tray, the automaton spins around and returns the way it came from. Particular feature of this type of automata is deliberate portrayal of feet, clad in traditional tabi socks, which give the impression of the doll walking, although it moves on a set of wheels, cleverly concealed by long robes. The frame and most of the mechanism parts, except wind-up spring is constructed out of wood. A few specimens are extant, yet none matches exactly the drawings from “Kik¯o Zui” (“Illustrated Compendium of Automata”) (the page that presents such kind of mechanism is given in Hosokawa 1796b). Other type of less complicated, moving only in one direction, carrying automata worth mentioning is kani sakazuki-dai, crab goblet stand. Kyokutei Bakin, in his work “Chosakud¯o Issekiwa” (“Stories told a certain night at the author’s pavilion”), published in 1802, gives a detailed description of such automaton, once a beloved possession of high ranked Kyoto courtesan, Yoshino-tayu the second (Kyokutei 1802). The description is supplemented with detailed drawings, showing the karakuri from the above and from the front. The automaton itself is a life-sized and life-like representation of a crab with raised pincers. It carries the sake-cup on its body using the same principle as the ocha hakobi ningy¯o: the weight of a cup set on the plate in center of automaton’s body triggers the mechanism, and the crab starts mowing on cogwheels protruding from the belly, sideways, until the cup is lifted. The cogs in the mechanism assure the movement of the legs resembles the one of

78

M. Kostrzewski and W. J. Nowak

actual, living animal. Interestingly, kani sakazuki-dai is composed entirely of metal, an exception among Japanese automata, and two existing specimens (both in private collections: one in the family of Yoshino-tayu descendants, second belongs to Hine Yukikazu (Niigata City History Museum 2012)) to the authors’ best knowledge both are in working condition, a rare case, as doll-shaped automata, however extant, are usually immobile. The idea of carrying things to the spot of pick-up and returning to the point of origin is used in modern application of karakuri techniques, as described further in the chapter. Prior to describing the possible industrial application of karakuri principles, the authors would like to consider the presence of historical karakuri in modern Japan. One of the most important and still valid purposes of karakuri is to foster curiosity. The sense of marvel merged with playfulness and awe, lead to questions and explorations in search for solutions to solve a seemingly simple assignment: How does it work? The pursuit for the answer fosters creativity and stresses the importance of thorough understanding of the concept and its meticulous execution. Karakuri are not only valuable sources of know-how, but also a promising vehicle for education and tuition. Numerous institutions and people engaged in karakurirelated work, including transfer of knowledge through exhibitions and presentations, preservation and reconstruction of historical contraptions are taking advantage of the platform Japanese automata have created to educate in fields, of art, history, and technology on different level of education (see Tominaga et al. 2019) and inspire people in various ways, by contributing to enhancement of reliable, historical solutions, while providing joyful human factor. The abundance of publications guiding through the process of re/construction of historical karakuri designs with readily available materials (Hara 2020; Sakano 2001) or pre-produced sets for assembly (Otona no Kagaku 2021), prove that in the digital age, analog automata are still interesting and beneficial, in the traditional, or adapted form. Concerning the latter, modern logic puzzles (Karakuri 2021) based on idea of Akazu-no suzuribako, in-openable calligraphy box made by no one else but Tanaka Hisashige in 1807 (miraikan) are prized in Japan and abroad, Toshiba Mirai Kagakukan (2021). They allow exercise of the mind, a vital task in times when human tend to outsource their abilities onto technologies. The modern application of karakuri is connected to the term karakuri kaizen which in short means constant improvement with use of modernized karakuri applications and the modern application has aroused, albeit still little, interest among researchers as it can be observed based on Fig. 2. Contrary to digitalized Industry 4.0, karakuri kaizen is equivalent of analogue automation. It is not connected to all variety of technologies which have been developed recently, for example, Radio-Frequency IDentification (RFID), Narrow Band Internet of Things (NB-IoT), Wireless Fidelity (Wi-Fi), Near-Field Communication (NFC), Fifth-Generation mobile networks (5G), Global Positioning System (GPS), Wireless Sensor Network (WSN), robotics as Industry 4.0 does (Liu et al. 2018). It is subjected to simplicity of physical phenomena and mechanisms from the era before the invention of electricity, easy to develop and maintain yet complex in the same time. Karakuri

4 Karakuri Solutions and Industry 4.0

79

Fig. 2 The last 5 years of developments on the topic of this chapter; number of publications (i) per consecutive year (y). (Sources: Mašín and Riegr (2016), Sawaguchi (2017), Matsumura et al. (2017), Mašín and Riegr (2017), Katayama (2017), Yamamoto et al. (2018b), Riegr and Mašín (2018), Chen et al. (2018), Kit et al. (2018), Arai et al. (2019), Nascimento et al. (2019), Madisa et al. (2019), Tominaga et al. (2019) Thiptarajan and Lertrusdachakul (2019), Nakagawa et al. (2020), Prasetyawan (2020), Anggrahini et al. (2020), Riegr and Mašín (2020), Tan Gar Heng et al. (2020), Malkova and Zadiranov (2020), Chen (2020))

kaizen is applied in various manufacturing subsectors in form of relatively complex mechanical systems that use the elementary phenomena. Mechanisms for karakuri implementation are as follows: gravity force, magnetic force, friction force, lever mechanisms, weight-shift mechanisms, gears, spring mechanisms, pulley mechanisms, roller mechanisms, water jets, seesaws, counterweights, dampers for energy dissipation, reflection and refraction of light, gas or liquid flow, etc. (Rani et al. 2015) to name only a few, all applied to transform the initial input into a different kind of movement (Besiroglu 2017 recalled them all to the following forces’ terms: force of gravity, clamping force, radial force, tensile force, spring force, hydropower). Arslankaya and Yonar (2015) described karakuri as energyfree motion which is not entirely true. Karakuri technology utilizes natural energy with elemental physical contrivances as Katayama et al. (2019) and Tortorella et al. (2020) mentioned. Gravity, magnetism, and other forces occurring in nature without being caused by human activities, work continuously, although vary in values as these forces’ values depend on latitude, which is related to the shape of the Earth and the influence of centrifugal force, the nonuniformity of the density of the Earth’s

80

M. Kostrzewski and W. J. Nowak

crust and many other parameters. Therefore, it is not surprising that it can be used in nontrivial solutions introduced to the industry. This chapter’s aim is to present such low-cost, analogue yet sophisticated automation concept of karakuri, in contrary to digital and alike sophisticated Industry 4.0 concept, which is a leading area of interest in the published book (karakuri mechanisms are the so-called low-cost automation, Mašín and Riegr 2017; Paraponiaris and Rodríguez 2019). In the second section, the authors present analysis of interdisciplinary literature search based on renown scientific databases. The third section focuses on selected solutions of karakuri kaizen in industrial applications (karakuri kaizen is understood as “a [constant – added by the authors] improvement by creating simple and low-cost mechanical devices that automates some parts of the shop floor operations,” Japan Institute of Plant Maintenance 2009; Yokota 2009). Karakuri kaizen is considered as the basis for a sustainable and efficient manufacturing process in modern Japanese plants (Bock et al. 2019, p. 8). In the chapter, the authors take a challenge to present the less known, early pages of the story of last-decade Industry 4.0, and especially karakuri as progenitors of robotic mechanisms, as described in previous section. In the fourth section, the authors focus on brief description of reflection on the subject matter. The chapter is finalized with conclusion and proposals for future research.

Where to Find Karakuri? As Katayama et al. (2014) deliberated, only a few papers have been published on the subject matter so far. When the authors of this paper have verified this comment based on the data gathered in the Scopus database, it became evident that only seven publications were noted before 2014 (Fig. 3). Katayama et al. (2014) in their work added that despite the fact of a low quantity in number of research papers published, international conferences were considered as quite important aspect of the improvement, sharing, and transfer of knowledge on karakuri. That was one of the reasons to launch Japan Institute of Plant Maintenance in 1993. The organization aim is to promote hypotheses, knowledge, and solutions (equipment) connected to karakuri used in production, logistics, and others technological processes as well as their design (Katayama et al. 2014). Deep and substantial experience accumulated among leading role industries such as car assemblers and parts suppliers (conference/exhibition of Japan Institute of Plant Maintenance: Karakuri Kaizen Kuf¯u Ten translated as Karakuri Improvement Exhibition organized in 2009, METI 2012; Nakano 2005; Saka 2007 – references directly taken after Katayama et al. 2014) drove Japan Institute of Plant Maintenance to prepare Karakuri Exhibition each year in Tokyo and Nagoya alternatively (references to this information was taken originally after Katayama et al. 2014, namely conferences/exhibitions events of Japan Institute of Plant Maintenance: Karakuri Kaizen Kuf¯u Ten translated as Karakuri Improvement Exhibition organized in 2012 and 2013). As Bhanu and Kumar (2018) admitted in their elaboration, industrial employees apply ideas presented during events organized by Japan Institute of Plant Maintenance.

4 Karakuri Solutions and Industry 4.0

81

Fig. 3 Number of publications (i) characterized by karakuri keyword per consecutive year (y). (Source: Scopus database (the database access date: 16 December 2020))

Six years after observations given in Katayama et al. (2014), the authors of this chapter examined the quantity of publications collected in the Scopus database. The results are given in Fig. 3. The researchers’ interest is growing each consecutive year after 2016, and nevertheless currently it seems that actual industrial application raises more interest within companies than among academic audience. When the authors checked the keyword of karakuri for all the searching fields, 127 documents results were returned. Since the interest of the authors was foremost in engineering application, the scope of search was narrowed to Engineering and the language of publications was limited to English. These two limitations resulted in return of 42 documents (notably, before exerting language limitation, only 2 additional publications were returned, both in Japanese). Among the resulted publication 19 items were indexed as conference proceeding papers, 14 items as journal research papers, 5 items as books, 3 items as book series, and 1 item only as trade journal paper. An attempt to promote solutions outside the Japanese market was also made by developing diploma projects. As Bhanu and Kumar (2018) mentioned, their elaboration’s main purpose was to introduce karakuri concept to Europe by means of various solutions’ presentation which are currently applied in Sweden and throughout Asia. Bhanu and Kumar (2018) also considered a method to design these solutions. Also, work by Paraponiaris and Rodriguez (2019) can be mentioned as such attempt. At this point, it should be mentioned that although this quantitative examination presented in Fig. 3 is limited to Scopus database only, it does not mean that during elaboration of this chapter no other publications were included and considered. On

82

M. Kostrzewski and W. J. Nowak

the contrary, the authors additionally searched for publications in various databases via internet browser. For example, in Web of Science only 10 publications were found from 2015 to 2020 (the database access date: 30 December 2020). Whereas, in Google Scholar database 117 records (66 publications in English) were returned in the case of “karakuri kaizen” phrase (the database access date: 30 December 2020). Each of these publications has been verified by the authors in terms of the actual consideration of the karakuri issue in engineering terms. It should be borne in mind as well that some of the presented publications are during the process of indexation in recognized scientific databases, therefore the graph given in Fig. 3 may change not only for 2020, but for the previous years as well. It is worth mentioning that the methodology of this chapter was developed in order to gather as comprehensive information on karakuri kaizen solutions as possible within the limited range of time. A library query and a desk research methodology were conducted. The research was based on both printed matter and resources available in the World Wide Web, in particular in scientific databases as Web of Science, Scopus, and Google Scholar. All of the analyzed documents were verified accordingly.

Karakuri Kaizen in Industrial Applications In this section, the authors describe selected karakuri-based solutions indexed in the scientific databases as described in methodology description above. At the end of this section, selected remarkable design methods connected to karakuri improvement are described. Rani et al. (2015) faced the problem of frequent power cuts which results in production downtimes in the case of power-intensive machines. The authors decided to rearrange tilting unit in the working station by means of employing gravity for tilting process in such a way that a unit does not require electricity to operate. Additional value-added aspect of a newly developed equipment design was reduction of cycle time by one-fourth as no unnecessary downtimes in a tilting process have been observed. After the use of trial and error method, the authors introduced air pump coupled with the counter weight fixed on the beam of a tilting equipment to the design – such solution supported both speed reduction during tilting down and consequently allowed tilting up (application of pressure) and returned the machine to its initial position. Yashvant Khire and Madnaik (2001) developed a karakuri-based mechanism for forming a corrugated paper carton packages for grapes transportation from flat card boards. Typically, in such cases either numerous employees or a robot for cartons folding is essential as Lu and Akella (1999) reported (they developed research on a SCARA robot for carton folding). The authors of described karakuri-based paper folding mechanism excluded the need of robotics implementation and reduced number of employees involved in the process of carton folding. In the authors’ device certain manual operations connected to folding had to be included; however, a lever (accordingly equipped with spring mechanisms) has performed the most

4 Karakuri Solutions and Industry 4.0

83

labor-intensive task. The authors specified all the guidelines for this carton folding process. Paraponiaris and Rodríguez (2019) presented in their report all the design steps for karakuri concept application for redeveloped oil pan assembly station, to be precise, for transporting heavy parts (oil pans) between two cooperating assembly lines (oil pan was chosen in the project as a heavy piece to transport in automotive assembly processes). The simplified visualization of the “self-propelled cart” designated to transport oil pans between two cooperating parallel assembly lines is given in Fig. 4. At first, an oil pan is put into position noted as (1) (Fig. 4; in this solution oil pan works as a trigger, similar to the tea cup in the case of ocha hakobi ningy¯o mentioned in the chapter’s introductory section), which causes lifting of the weight noted as (2) and at the same time lowering a rack (a linear gear). Cooperation of a rack coupled with a pinion (a circular gear) powers a flat belt which triggers wheel(s) rotation. The springs noted as (4) protect the shelf on which an oil pan is placed against excessive decrease. The authors presented essential equations of motion, FEM analyses, aspects of ergonomics for employees supporting origin and destination points between cooperating assembly lines (as a result of ergonomic aspects, a concept sketch of stepping platform for employee fixing small parts onto oil pan was included).

Fig. 4 Simplified visualization of the “self-propelled cart” concept designated to transport oil pans between two cooperating parallel assembly lines (the reader is asked to check reference documents for detailed design of particular solution). (Source: authors elaboration based on Paraponiaris and Rodriguez (2019), p. 22, 52)

84

M. Kostrzewski and W. J. Nowak

Kit et al. (2018) redesigned the lamps manufacturing assembly line. The authors initially gathered data and developed the simulation model of assembly line. Then, after rigorous analysis, three karakuri-based flow racks were incorporated into the line design (each one for other type of production items, namely base, lens and reflector loading bay) and at the same time this solution was recommended as an equipment appropriate to the eliminate unnecessary motion, transportation and waiting. The operation in this solution starts by pressing down a pedal what allows a transfer of production items of the same type in continuous manner and controlled in such a way that the production cycles are aligned. The solution allowed shortening of production cycle time by several seconds due to elimination of non-value addition of elementary activities within the described operation. Presented solution leads to more complex flow racks concept, which is implemented, for example, in Swedish company as Bhanu and Kumar (2018) reported. This type of warehouse rack serves as a buffer between operations (e.g., between storing and order picking processes) which is involved with carrying small containers holding items to be delivered to an employee at an ergonomic height. If one box (e.g., a tray with materials as given in Fig. 5, noted as (2)) is conveyed from the rack, the preceding boxes slide down under the influence of gravitation force to make items available to the employee at the same workstation (Fig. 5, noted as (1)). In the solution, the operation starts by pressing down a pedal included in the mechanism noted as (3), what allows a transfer of production items in the same manner as in the solution presented in Kit et al. (2018). Similarly, as in the case of both solutions presented in Paraponiaris and Rodriguez (2019) and ocha hakobi ningy¯o tray mechanism, a pedal works as a trigger. Karakuri-based solution for handling operation with use of gravitation force can provide efficient storage and easy unloading process, especially in the case of heavy items being moved in logistics, production, and technological processes. In the case of various types of items, boxes can slide in many parallel “aisles” each of them equipped with independent placon rollers. Such solution ensures continuous delivery of materials/items, if such rack is supplied with semi-finished products at its starting point. The alternative version of the flow rack enriched with karakuri-based solution is so-called zig zag flow rack presented, for example, in Abdul Halim et al. (2012). In the case of this type of rack, it is possible for trays with items to slide from a considerable height as a result of reducing tray speed by using the inclination of aisles with placon rollers alternately in relation to the following storage levels (tiers, storeys), namely, once left and once right. The use of placon roller conveyor is relatively often applied equipment in karakuri-like solutions. Shamsudin et al. (2019) applied such an equipment in automotive assembly process in order to decrease the duration of cycle time of process by reducing the need to walk imposed on employees while performing their tasks. Obligation to walk long distances interfered with the full cycle time of the production process and also interfered with a tact time of the whole process. The researchers suggested introduction of direct items supply (automotive parts) to particular workstations located along the assembly line. After the solution’s implementation, the cycle time was reduced by more than 20%. This application is

4 Karakuri Solutions and Industry 4.0

85

Fig. 5 Simplified visualization of the karakuri-based flow rack presented in two steps, namely, (a) and (b) (the reader is asked to check reference documents for detailed design of particular solution); notations: (1) – workstation, (2) – tray with materials, (3) – mechanism with pedal which after pushing executes further material flow. (Source: the authors elaboration based on Kit et al. (2018) (similar solution presented in a loop can be observed in SUS BKK 2016a))

closed to one of the definitions of karakuri as resource saving mechanism (Miyake 2006; Shishir 2010). In solutions focused on the use of karakuri-based concept, energy applied may also be reused. An example of such energy reuse is the cooperation between automated guided vehicles and flow racks. When an automated guided vehicle moves close to such a rack, the vehicle can release a locking mechanism that blocks a tray with items from sliding down along placon rollers in a flow rack. An automated guided vehicle grips the sprocket which, when properly positioned, releases the locking mechanism of a flow rack and a tray with items slides along placon rollers installed in a rack, and, under the influence of gravity, is loaded onto an automated guided vehicle’s storage space. Some companies operate material handling from handling processes to order picking ones with karakuri-based solutions applied. Bhanu and Kumar (2018)

86

M. Kostrzewski and W. J. Nowak

reported use of such solutions by several leading Japanese companies. These solutions refer to the incorporation of low cost racks and shooters (p. 25), conveyors, work benches, wagons, carts and trolleys (p. 25), corner handling and warehouse handling solutions (p. 26), products dedicated to production and logistics such as internal logistics, assembly workstations, transportation, storage and modular systems (p. 26). As Bhanu and Kumar (2018) mentioned, the representatives of selected leading Japanese companies disseminated karakuri-based solutions by offering guided factory tours and courses on logistics systems (p. 25). Bhanu and Kumar (2018) reported as well that selected companies provided special kits for karakuri improvement and implementation and even DIY (do it yourself) type software that allows to design equipment and processes which may facilitate karakuribased principles (such offers included blueprints, tutorials, and components; p. 26), what reminds karakuri construction manuals both ancient as Hosokawa (1796a) and currently published as Hara (2020) and Sakano (2001). Among the suppliers of such solutions, the authors reported representatives of European producers as well, including, for example, German and Swedish manufacturers. As far as promotion materials are considered, it is worth noting that according to the conference paper of Katayama (2017), even karakuri improvement DVD was released to share the knowledge connected to subject matter (Katayama and JIPM 2012). The concept of karakuri can be implemented both in order to support highly developed technological processes of heavy industry factories as it was abovementioned; however, it also applies to in-farm logistics as in Mašín and Riegr (2016), Riegr and Mašín (2018), Prasetyawan (2020), Allan et al. (2020), etc. Mašín and Riegr (2016) focused on karakuri trolley’s design which the authors planned to apply in area of agriculture in order to transport small containers with vegetables or bags filled with compost or animal fodder from initial to final point of transport within the plant. The researchers have analyzed characteristics of speedtime and distance-time based on mechanics’ equations. Acquired solution was based on well-known physics rule of potential energy of moving trolley with containers being changed to kinematic energy. The potential energy accumulated into the press springs coming from a container’s weights (gravitation force) allocated on a trolley was transmitted through a set of sprockets which caused forward movement. This proposal did not consider the operations of loading and unloading a trolley. Brief information on modeling of the mechanical behavior of this karakuri-based trolley was given in as well, whereas Riegr and Mašín (2018) continued their analyses and consequently they developed equation of mechanical spring’s oscillations dampened by Coulomb sliding friction. Prasetyawan (2020) also suggested using karakuri solution in rural production, nevertheless the author focused on a particular technological process namely production of pineapple fruitcake. By using a conveyor, the dough from a stirrer is transferred through the baking process to the cooling process. At the end part of the conveyor, the dough is poured into the cooling compartment underneath the conveyor by utilizing forces of gravity. According to the author, such solution allows to reduce consumption of electric energy by 45%, nevertheless it is difficult

4 Karakuri Solutions and Industry 4.0

87

to include the solution in group of karakuri technology. However, as Weber (2014) mentioned, gravitational flows are used in order to apply energy-saving concept that utilizes the flow of gravity and such flow can be operated in a manner to move bins loaded with parts and components forward, as well as increasing efficiency and ergonomics of the process. As mentioned above, the term karakuri in technological processes is defined as sophisticated mechanical structure that applies energy of natural origin (Katayama 2017; Muñoz et al. 2020). Combining both karakuri and kaizen terms, improvements in manufacturing settings can be introduced by means of implementation of mechanical devices that provide automation with limited use of electrical energy or other external source of power. Such devices rely on mechanical design which takes advantage of natural forces such as gravitation combined with pulley systems to automate or semi-automate different activities in the manufacturing processes. Therefore, the concept of Prasetyawan (2020) is closer to semi-automated karakuri term than full-scale karakuri itself. Using less of energy equipment in food industry raise interest of other researchers as well. Karakuri-based wingko baking equipment was presented in the paper by Anggrahini et al. (2020) (wingko is a traditional snack from East Java region, which ingredients are coconuts, glutinous rice flour and sugar). The karakuri-based baking equipment consists of two placon roller conveyors that utilize gravitational force to bring a baking tray into and out of a baking machine (the term “baking machine” refers to Wingking 2.0 which uses karakuri kaizen concept and full automation). After baking operation is finished and baked wingko removed, an empty tray automatically comes out of baking machine, and it is filled with batter, which weights it down. A spring goes down, a metal platform bends and allows filled tray to move forward and slide onto second conveyor and then it is moved by force of gravity. The solution supports shortening the duration of wingko baking process. Allan et al. (2020) also presented solution based on lean management principles of karakuri kaizen. The researchers aim was to redesign a typical plant cultivating factory layout in order to reduce the necessity to move trays with growing plants from one place to another (places connected to various cultivation processes such as lighting and watering) and enable the implementation of low-cost, simple automation. According to the authors, the prototype enabled reduction of vertical and horizontal moves of trays with growing plants as well as optimized the space used for the growing area in the plant factory. In the initial part of the paper, the authors presented several reasonable evidences on why (urban) farming should be focused on a greener and self-sustainable approach to agriculture. The authors developed a new structural system that maintains operation abilities of plant factories and at the same time creates a safer working environment by means of reduction of reliance on human labor (e.g., tray handling). In order to enable automated mechanical motion based on gravitational forces, classic inclined plane was used. The experiment determined mass (noted as (3) in Fig. 6) needed to move a tray with growing plants and a static coefficient of friction. The result of the experiment concludes that an angle of minimum 2.25◦ is required to enable automated mechanical motion. The tilt angle was increased since trays with plants

88

M. Kostrzewski and W. J. Nowak

Fig. 6 Simplified visualization of the karakuri-based conveyor (the reader is asked to check reference documents for detailed design of particular solution); notations: (1) – a tray with growing plants, (2) – placon roller, (3) – an additional mass. (Source: the authors elaboration based on Allan et al. (2020))

are getting heavier while plants are growing and being watered. Certainly, this was the only part of their solution directly connected to karakuri concept. Karakuri is implemented by small-scale players to transport or move rather light objects (parts, items). However, foremostly it is being introduced by bigger players, for example, by several automotive companies (Paraponiaris and Rodríguez 2019, p. 4), machines and automotive elements manufacturers, as well as providers of components and systems for the automotive industry (Bhanu and Kumar 2018), wire rope industry (Madisa et al. 2019), in order to transport much heavier elements. This choice of karakuri-based applications is not surprising since material handling is one of the most important processes in these industries, especially if it involves movement of heavy components. Bhanu and Kumar (2018) mentioned several karakuri-based systems operating in actual applications, for example: load carrying cart with a set of dears and pulleys as power transmission mechanism, automatic tilting drum with spring and balance lever as power transmission mechanism, shooter machine equipped with set of gears and pulley, a trolley/lift which serves as a dolly and as an assembly jig conveyor, balance lever constructed to support a work space unit facilitating and allowing task performance to elderly and physically disabled people, thread cutting machine that assist the manufacture of threads for screws supported by power gear and cam and follower coupled mechanism as power transmission mechanism. Productivity and ergonomic aspects of karakuri-based solution were also of researchers’ interest. Madisa et al. (2019) validated the standards of work accordingly with the tasks’ productivity for karakuri-based improvements of workstation in a company which manufactures wire ropes. Survey questionnaire among employees of a company were distributed and based on their answers the authors presented the level of employees’ satisfaction. Production rate was also improved slightly after implementation of the solution presented in Madisa et al. (2019). Development of one of the strongest pillar of Industry 4.0, namely, Internet of Things (IoT) solutions, is complex, time-consuming, and challenging activity since software, hardware, and communication modules have to be appropriately composed (Udoh and Kotonya 2018; Muñoz et al. 2020). IoT solutions’ design was

4 Karakuri Solutions and Industry 4.0

89

also an interest of researchers who focused on karakuri concept. Raising popularity of karakuri concept in technological processes evoked interest of designers in developing Karakuri IoT design cards (Muñoz et al. 2020). Muñoz et al. (2020) mentioned after Lucero et al. (2016) that a design process can be supported by technique connected to design cards used during almost every design phase, from idea creating through conceptualization up to evolution of a solution to a particular problem. The goal of Muñoz et al. (2020) was to consider whether shop floor employees, not directly connected to information technology departments’ processes, are able to envision conceptual solutions to particular problems connected to IoT technologies with support of the karakuri cards developed by the researchers. Muñoz et al. (2020) described both the way of Karakuri IoT cards development and selected results of their implementation during a workshop session. Development of cards started by defining problems and opportunities (Yamamoto et al. 2018b) and collecting information on types of sensors and technologies to be used on cards’ descriptions and attributes. The original deck was composed of 42 cards until it was eventually reduced to 27 cards with additional 5 cards (so-called empty jokers; Muñoz et al. 2020). To show the usefulness of the Karakuri IoT cards serious game, the authors engaged various methods and tools such as role-play, storyboard, and gamification. What is worth mentioning, the cards were developed in such form to provide common language between participants of serious game coming from different backgrounds (common language is understood here as solutions’ design). It is worth stressing that, since it is in hand kind of tool (the card deck), it does not require external sources of energy (the authors of current chapter daresay it is possible to consider the case of Karakuri IoT cards as a fun fact connected to the term karakuri). Karakuri IoT cards were not the first attempt to support the karakuri-based technologies design. Katayama (2013) (referenced here after Katayama et al. 2014) presented a template procedure for karakuri analysis, which was designed to help experts in developing essential mechanisms of karakuri cases. Specification of this template includes: (1) title of improvement project, (2) company name, (3) short name of the considered case, (4) objective operation, (5) feature of the expected solution (e.g., variety of parts’ shape and weight, automatic transportation of heavy parts), (6) overall structure of karakuri contrivance (given as appropriate scheme/figure), (7) state definition of karakuri contrivance/device (definitions of the elements) in and their states of karakuri mechanism, (8) elementary operations, (9) state and transition of state (given, e.g., as a binary system notation), (10) state description after transition (presented as concept visualization), (11) sequence of karakuri operations, (12) elementary mechanisms, (13) functions, (14) provided values, (15) other merits (Katayama 2013). All the details of each particular template’s notations are given in the mentioned reference. Moreover, a way of solutions assessment was given there as well. This template may be of interests for those researchers and investigators who plan to construct or reconstruct their solutions into karakuri-based mechanisms.

90

M. Kostrzewski and W. J. Nowak

Selected Reflections on Karakuri As it was presented in the introduction to the chapter, the modern applications of karakuri-based mechanisms/solutions invented in Japan had originated from European-based clockworks and Japanese mechanical dolls known as karakuri ningy¯o. Technology has been intertwined with entertainment to finally emerge in modern times as one of the powerful solutions focused on industrial development. The analysis of contemporary karakuri-based solutions has allowed the authors of this chapter to draw several conclusions, presented below. Reflections on the karakuri-based application allowed to develop several findings: • In order to design karakuri-based equipment, the researchers and engineers used analytical methods (e.g., Kit et al. 2018; Mašín and Riegr 2016; 2018; Madisa et al. 2019; Dewi and Pramono 2019), numerical methods (e.g., Paraponiaris and Rodríguez 2019; Mašín and Riegr 2016; Riegr and Mašín 2018; Prasetyawan 2020; Anggrahini et al. 2020), simulation methods (for example: Kit et al. 2018) and methods based on experiments (e.g., Rani et al. 2015; Yashvant Khire and Madnaik 2001; Katayama et al. 2014; Yamamoto et al. 2018b; Muñoz et al. 2020). • Karakuri-based solutions as significant part of lean manufacturing serve as one of the ideas to minimize waste (muda, meaning “futility,” “uselessness,” “wastefulness”), which according to Anggrahini et al. (2020) and current chapter’s authors observations of technological, production, and logistics processes include muda in areas of transportation, inventory, motion, waiting, overprocessing, overproduction, as well as defects. This waste minimization in karakuri-based solution is indirect and occurs, for example, through saving energy consumption, employees’ vitality, used space, and shortening of cycle time for material handling and accompanying processes. • Karakuri-based concepts applied in European companies mostly consider material handling and flow racks implementation (Paraponiaris and Rodríguez 2019; Bhanu and Kumar 2018). • Karakuri-based solutions support reduction of transport, production cycle times, and operational costs (including initial expenditures), as well as improvement of material handling, ergonomics at workstations, and reduction employees. Most of material handling devices, unless these are, for example, platform pallet trucks or flatbed pushcarts, usually consume substantial amounts of fuel (e.g., fuel-driven forklift trucks) or electricity (electric forklift trucks, high-bay stacker cranes, automated storage and retrieval systems and other similar devices) that increase operational cost of logistics or production company due to daily increase in costs of energy resources (Kit et al. 2018). However, this is only a part of the bigger problem. One of the most distinctive features of karakuri-based solutions’ implementation is the reduction of electrical energy consumption in logistics or production facilities which allows them to achieve minimal environmental impact (minimization of the contribution to environmental poisoning). Lower

4 Karakuri Solutions and Industry 4.0

91

energy consumption serves as contribution to reduction of carbon footprint and greenhouse gas emissions. • Karakuri-based solutions do not require high expenditures and operational and maintenance costs which are certainly lower than in the case of electrified or digitalized solutions – among the reasons for this type of application is the implementation of low-demanding components as lever mechanisms, gears, springs, pulleys, rollers, cams, seesaws, counterweights, dampers and exploitation of gravity force, magnetic force, friction force, etc. • Karakuri-based solutions do not require highly qualified personnel, especially when operations in a particular process are trivial, repetitive, and require quite a lot of energy and power. • Currently, a powerful know-how base on the possibilities of using this type of solutions is accessible in the form of short videos available on the one of the Internet services: post step (SUS BKK 2016b), eject box (SUS BKK 2016c), combination of selected karakuri-based operations presented in a loop (SUS BKK 2016a), karakuri chuter (SUS BKK 2016d) often used at the control airport. Finally, it is worth noting that a concept corresponding to karakuri has also been noticed. Sawaguchi (2017) mentioned jugaad (a colloquial Hindi term which means an innovative solution) as innovative and creative solutions developed in results of local communities’ necessities in India. The author commented on (karakuri) kaizen and jugaad in the context of their mutual entanglements and connections.

Conclusion The authors state that the case of karakuri is a valuable example of how an inspiration for entertainment, a folly, can translate into smart industrial solution. In times, when advanced, digital technology is readily available, when the concept of Industry 4.0 is becoming more and more widespread, examining the ideas of karakuri can bring valuable, up-to-date insights, especially in terms of user experience and electrical energy consumption. Depending on the scope of implementation, one might focus on the outer appearance of the apparatus, and the way to create it in a friendly and ergonomic way, an aspect worth considering within the context of prognosed forthcoming demands for life-assisting or merchandising robots. On the other hand, it is not only the shape itself, but the principles, that could be used for different applications. The karakuri-based solutions can be beneficial both for researchers and business representatives. The karakuri-based solutions give researchers the following research agendas: • Analyses of karakuri-based solution mechanics and operating principles • Analyses of complex technological process with use of karakuri-based equipment in a particular facility

92

M. Kostrzewski and W. J. Nowak

• Continuous improvement, kaizen, of innovative systems, as karakuri-based ones, aimed at sustainable approach in the contexts of environmental pollution’s reduction, operational cost and cycle times optimization, working conditions improvement (e.g., workload reduction, ergonomics improvements), etc. The karakuri-based solutions give business representatives the following advantages: • • • •

Reduction of expenditures. Reduction of operational costs. Reduction of electrical energy and fuel consumption. Reduction of unnecessary activities in technological, logistics, manufacturing, and transport processes. • A sense of being innovative. • High level of automation and Industry 4.0 solutions especially within the scope of Internet of Things technologies are adapted by an increasing number of companies – low cost automation is a good equivalent especially for those of companies which represent small to medium scale industries (Bhanu and Kumar 2018). • Cooperation with the researchers in continuous improvement, kaizen, of abovementioned innovative systems. In the era of Industry 4.0 all seems to be heading toward sole reliance on digitization, full automation – however, it might be worth imagining what would happen if vibrating electrons suddenly stopped “emitting” electricity? Or, what would be more likely, if satellites were to become defunct? and the Internet of Things stopped being reliable? Considering present, unprecedented conditions, in terms of preparation for potential energy shortages, would it be worthwhile to, yet again, take interest in the hundred-years old automation, namely karakuri? It is worth remembering that behind the simplicity of these solutions lie centuries of experience of engineers, mechanics and craftsmen. It can be stated that it is worth going back to the “beginning.” Last, but not least, the interdisciplinary research presented in this chapter is meant as informative, not evaluative, and the authors do not intend to discredit any of mentioned or any other author, publication or journal. Above all, the aim of this publication is to promote the idea of karakuri as one of the possibilities to reduce energy consumption in industry, production processes, logistics, and transport. Acknowledgments The study was inspired by the research visit of the first author in Tokai University, Japan (host department: Department of Navigation and Ocean Engineering, School of Marine Science and Technology, host professor: Dr. Koichi Shintani, term of the visit: between December 2019 and February 2020, title of the research grant: “Maritime Economics and Logistics in the Era of Industry 4.0 and especially Logistics 4.0”), what was a significant opportunity to explore Japanese culture and study several aspects of engineering and technology.

4 Karakuri Solutions and Industry 4.0

93

The authors would like to gratefully acknowledge the editors for invitation to take participation in the book project and to the reviewers, who provided helpful comments and insightful suggestions on the draft of the manuscript.

References Abdul Halim NH, Jaffar A, Yusoff N, Adnan AN (2012) Gravity flow rack’s material handling system for just-in-time (JIT) production. Proc Eng 41:1714–1720. https://doi.org/10.1016/ j.proeng.2012.07.373 Allan TGH, Mohamed H, Mohamed Rafaai ZF (2020) Implementation of lean manufacturing principles in a vertical farming system to reduce dependency on human labour. Int J Adv Trends Comput Sci Eng 9(1):512–520. https://doi.org/10.30534/ijatcse/2020/70912020 Anggrahini D, Prasetyawan Y, Diartiwi SI (2020) Increasing production efficiency using karakuri principle (a case study in small and medium enterprise). IOP Conf Ser Mater Sci Eng 852:012117. https://doi.org/10.1088/1757-899X/852/1/012117 Arai H, Kimoto M, Iio T, Shimohara K, Matsumura R, Shiomi M (2019) How can Robot’s gaze ratio and body direction show an awareness of priority to the people with whom it is interacting? IEEE Robot Automat Lett 4(4):3798–3805. https://doi.org/10.1109/LRA.2019.2929992 Arsénio A, Oliveira M, Caldas L (2011) Social interaction and the development of artificial consciousness. In: Chugo D, Yokota S (eds) Introduction to modern robotics. iConcept Press, New York, pp 93–118 Arslankaya S, Yonar N (2015) Implementation of the lean production approach to production problems in a process at Toyota motor manufacturing Turkey. In: Br˘atianu C, Zbuchea A, Pînzaru F, V˘at˘am˘anescu E-M, Leon R-D (eds) Strategica – international academic conference: local versus global, Bucharest, Romania: Tritonic, 3rd edn. pp 211–219 Besiroglu J (2017) Development of an automation concept for complex process machines under the aspects of low cost automation and lean management. Master’s thesis in Business Administration, Technischen Universität Wien Bhanu MV, Kumar PBS (2018) Global study and implementation of karakuri investigation into the applications of karakuri and its integration with emerging technologies. Master’s thesis in Production Engineering, Chalmers University of Technology Bock T, Linner T, Güttler J, Iturralde K (2019) Ambient integrated robotics automation and robotic technologies for maintenance, assistance, and service. Cambridge University Press, Cambridge, UK ˇ Capek K (2008) R.U.R. – Rossums universal robots (trans. Koreis V). Booksplendour Publishing Chen Y-H (2020) A historical study of rope applications in mechanical devices in three eastern countries. Mech Mach Sci 78:502–509. https://doi.org/10.1007/978-3-030-30036-4_46 Chen Y-H, Ceccarelli M, Yan H-S (2018) A historical study and mechanical classification of ancient music-playing automata. Mech Mach Theory 121:273–285. https://doi.org/10.1016/ j.mechmachtheory.2017.10.015 Dewi D, Pramono AE (2019) Rancang Bangun Karakuri Mixing dan Material Handling pada Proses Pembuatan Cetakan Inti Water Jacket. Prosiding Seminar Nasional Teknik Mesin Politeknik Negeri Jakarta 1:751–760 Hato T, Suzuki K, Tomii Y, Yoshida M, Yokota Y, Kubota Y (2007) Mechanism of “man-nen dokei”, a historic perpetual chronometer (2nd report, power supply). Trans Jpn Soc Mech Eng C 73(729):1537–1544. https://doi.org/10.1299/kikaic.73.1537 Japan Institute of Plant Maintenance (2009) Karakuri improvement case collection. Nikkan Kogyo Shimbun, Tokyo. (in Japanese) Kanaya T (1979) Literacy in Japan. Community Coll Rev 7(1):13–15 Katayama H (2013) Karakuri no kaiseki to sekkei ni kansuru kenkyu-kaisekiyou template kaihatsu to tekiyourei ni tsuite (On design and analysis of “Karakuri” contrivance-template form

94

M. Kostrzewski and W. J. Nowak

development for analysis and its application; in Japanese). In: Proceedings of 16th annual national meeting of Japan Society of Logistics Systems (JSLS), Narashino City, Chiba, Japan, 11–12 May 2013, pp 25–28 Katayama H (2017) Legend and future horizon of lean concept and technology. Proc Manufact 11:1093–1101. https://doi.org/10.1016/j.promfg.2017.07.227 Katayama H, Japan Institute of Plant Maintenance (JIPM) (eds) (2012) Karakuri kaizen jirei shu-jirei yousetsu (Karakuri improvement cases-their summary descriptions; in Japanese) Maruwakari karakuri kaizen DVD (Perfect understanding of karakuri improvement DVD; in Japanese), vol 1 Katayama H, Sawa K, Hwang R, Ishiwatari N, Hayashi N (2014) Analysis and classification of karakuri technologies for reinforcement of their visibility, improvement and transferability: an attempt for enhancing lean management. In: Dundar FK, Anderson TR, Daim T, Niwa K, Perman G (eds) Proceedings of PICMET ‘14: infrastructure and service integration, Kanazawa, Japan, 27–31 July 2014, pp 1895–1906 Katayama H, Murata K, Lee D-J (2019) On advanced topics for reinforcing Leanized management. Proc Manufact 39:599–608. https://doi.org/10.1016/j.promfg.2020.01.427 Kit BW, Olugu EU, Zulkoffli Z (2018) Redesigning of lamp production assembly line. In: Proceedings of the international conference on industrial engineering and operations management, Bandung Koch A (2020) Diplomatic devices: the social lives of foreign timepieces in late sixteenth- and early seventeenth-century Japan. KronoScope 20:64–101. https://doi.org/10.1163/1568524112341454 Lancashire TA (2011) An introduction to Japanese folk performing arts. Routledge, London Lang F (ed) (1927) Metropolis. UFA GmbH, Balelberg Liu X, Cao J, Yang Y, Jiang S (2018) CPS-based smart warehouse for industry 4.0: a survey of the underlying technologies. Computers 7(13):1–17. https://doi.org/10.3390/computers7010013 Lu L, Akella S (1999) Folding cartons with fixtures: a motion planning approach. In: Proceedings 1999 IEEE international conference on robotics and automation, vol 2, Detroit, 10–15 May 1999, pp 1570–1576. https://doi.org/10.1109/ROBOT.1999.772583 Lucero A, Dalsgaard P, Halskov K, Buur J (2016) Designing with cards. In: Markopoulos P, Martens J-B, Malins J, Coninx K, Liapis A (eds) Collaboration in creative design. Springer International Publishing, Cham, pp 75–95. https://doi.org/10.1007/978-3-319-29155-0_5 Madisa IZ, Taib MFM, Reza NA (2019) Implementation of karakuri kaizen to improve productivity and ergonomics in wire rope industry. In: Proceedings of the international conference on industrial engineering and operations management, Bangkok, pp 2765–2775 Malkova MY, Zadiranov AN (2020) Shape optimization of nickel anodes used in the production of galvanic coatings of rocket engine. Adv Astronaut Sci 170:859–865 Mašín I, Riegr T (2016) Dynamic characteristics of the karakuri transport trolley. In: Rostislav Chotˇeborský R, Kováˇr S, Kˇrepˇcík V (eds) Proceeding of 6th international conference on trends in agricultural engineering 2016, Prague, pp 398–403 Mašín I, Riegr T (2017) Advanced modelling of the karakuri mechanism. In: Silva Gomes JF, Meguid SA (eds) Proceedings of the 7th international conference on mechanics and materials in design, Albufeira, paper ref: 6679, pp 747–748 Matsumura R, Shiomi M, Hagita N (2017) Does an animation character robot increase sales? In: Proceedings of the 5th international conference on human agent interaction (HAI’17) Oct 17– 20, Bielefeld, pp 479–482 Mayson GT (2000) Mechanical singing-bird Tabatières. Robert Hale, London Miyake DI (2006) The shift from belt conveyor line to work-cell based assembly systems to cope with increasing demand variation in Japanese industries. Int J Automot Technol Manag 6(4):419–439 Muñoz A, Florin U, Eriksson Y, Yamamoto Y, Sandström K (2020) The karakuri card deck: codesigning industrial IoT conceptual solutions. In: 16th international design conference – design 2020, Cavtat – Dubrovnik, Croatia, 26–29 October 2020, pp 807–816. https://doi.org/10.1017/ dsd.2020.127

4 Karakuri Solutions and Industry 4.0

95

Nakagawa K, Matsumura R, Shiomi M (2020) Effect of robot’s play-biting in non-verbal communication. J Robot Mechatron 32(1):86–96. https://doi.org/10.20965/jrm.2020.p0086 Nakano K (ed) (2005) Tokoton Yasashii TPM no Hon (Easiest TPM; in Japanese), Nikkan Kogyo Shimbun Sha (The Business & Technology Daily News; in Japanese). B&T Books Nascimento SD, Alves MB, Morais JO, Carvalho LC, Lima RF, Alves RR, Calado RD (2019) Karakuri: the application of lean thinking in low-cost automation. In: Silva FJG, Pinto Ferreira LC (eds) Lean manufacturing: implementation, opportunities and challenges. Nova Science Publishers, Hauppauge, pp 335–354 Paraponiaris YP, Rodríguez AM (2019) Product development of material supply - implementation of Karakuri kaizen. Bachelor’s degree project in Product Design Engineering, University of Skövde Prasetyawan Y (2020) Simple automation for pineapple processing combining with karakuri design. IOP Conf Ser Mater Sci Eng 852:012102. https://doi.org/10.1088/1757-899X/852/1/ 012102 Pyatnitsky Y (2017) Love, politics, and fi ne arts. The mechanical automaton “Golden peacock” of Catherine the great. Tyragetia XI[XXVI](2):345–370 Rani D, Saravanan A, Agrewale M, Ashok B (2015) Implementation of karakuri kaizen in material handling unit. SAE Technical Paper, 2015-260074, pp 1–5. https://doi.org/10.4271/2015-260074 Riegr T, Mašín I (2018) New approach to solving mathematical equation for damped oscillations by sliding (Coulomb) friction at the karakuri mechanism. In: Silva Gomes JF, Meguid SA (eds) Proceedings IRF2018: 6th international conference integrity-reliability-failure, Lisbon, paper ref: 7232, pp 177–178 Riegr T, Mašín I (2020) Solution of damped oscillations by coulomb friction at the Karakuri mechanism using MAPLE software. In: Medvecký Š, Hrˇcek S, Kohár R, Brumerˇcík F, Konstantová V (eds) Current methods of construction design. Lecture notes in mechanical engineering. Springer, Cham, pp 383–390. https://doi.org/10.1007/978-3-030-33146-7_44 Saka K (2007) Karakuri no moto: paper craft book (essence of Karakuri: paper craft book; in Japanese). Shubunsha, Shinjuku Sawaguchi M (2017) How does Japanese “Kaizen activities” collaborate with “Jugaad innovation”? In: 2016 Portland international conference on Management of Engineering and Technology (PICMET), Honolulu, 4–8 September 2016, pp 1074–1085. https://doi.org/10.1109/ PICMET.2016.7806683 Shamsudin E, Al Zahra Mat Darus SA, Raja Zainal Abidin RMH (2019) Implement karakuri as a material handling in production sealer line. In: Green Technology & Engineering Seminar 2019 (Greentech’19), Paris, 24–26 July 2019, pp 66–75 Shishir BN (2010) Changing from conveyor to work-cell-based systems to deal with fluctuations in demand. Proc Inst Mech Eng B J Eng Manuf 224(1):169–176. https://doi.org/10.1243/ 09544054JEM1374 Suzuki K (1988) A positive study on the technique of mechanism in the Edo era. Bull Natn Sci Mus E 11:41–61 Tajima M (1983) The fate of Nagoya’s mechanical festival floats. Asian Folklore Stud 42(2):181– 203+205–208 Tan Gar Heng A, Bin Mohamed H, Bin Mohamed Rafaai ZF (2020) Implementation of lean manufacturing principles in a vertical farming system to reduce dependency on human labour. Int J Adv Trends Comput Sci Eng 9(1):512–520. https://doi.org/10.30534/ijatcse/2020/ 70912020 Thiptarajan K, Lertrusdachakul T (2019) Integrating Monozukuri-based learning for undergraduate program in information technology. In: TIMES-iCON 2018 – 3rd technology innovation management and engineering science international conference. IEEE, Bangkok Tominaga H, Mori M, Iida N, Hirose K (2019) A study on Monozukuri education in Karakuri doll production. In: Cocchiarella L (ed) ICGG 2018—proceedings of the 18th international conference on geometry and graphics, advances in intelligent systems and computing, vol 809.

96

M. Kostrzewski and W. J. Nowak

Springer International Publishing AG, New York City, pp 2262–2265. https://doi.org/10.1007/ 978-3-319-95588-9_216 Tortorella G, Sawhney R, Jurburg D, de Paula IC, Tlapa D, Thurer M (2020) Towards the proposition of a lean automation framework: integrating industry 4.0 into lean production. J Manufact Technol Manag. https://doi.org/10.1108/JMTM-01-2019-0032 Toshiyuki I (1998) Uki-e – perspectives of a new kind. J Jpn Trade Industry 11(12):50–51 Udoh IS, Kotonya G (2018) Developing IoT applications: challenges and frameworks. IET CyberPhys Syst Theory Appl 3(2):65–72. https://doi.org/10.1049/iet-cps.2017.0068 Voskuhl A (2007) Producing objects, producing texts: accounts of android automata in late eighteenth-century Europe. Stud Hist Philos Sci A 38(2):422–444. https://doi.org/10.1016/ j.shpsa.2007.03.004 Weber A (2014) Modular framing systems let engineers get creative. Assembly, pp 57–11 Yamamoto Y, Sandstroöm K, Muñoz AA (2018a) Karakuri IoT – the concept and the result of prestudy. In: Thorvald P, Case K (eds) Advances in manufacturing technology, vol 8. IOS Press, Amsterdam, pp 311–316. https://doi.org/10.3233/978-1-61499-902-7-311 Yamamoto Y, Sandström K, Muñoz AA (2018b) Development of methods that support exploration of simple and low-cost IoT-aided improvement solutions at manufacturing plants. In: The 2018 annual autumn meeting of Japan industrial management association JIMA, Kashiwa-shi, Chiba, 26-27 November 2018 Yashvant Khire M, Madnaik SD (2001) Folding cartons using low cost automation – a case study. Assem Autom 21(3):210–212. https://doi.org/10.1108/EUM0000000005618 Yokota Y (2009) A historical overview of Japanese clocks and karakuri. In: Yan HS, Ceccarelli M (eds) International symposium on history of machines and mechanisms. Springer, Dordrecht, pp 175–188. https://doi.org/10.1007/978-1-4020-9485-9_13

Websites Akisato R, Takehara S (1796–1798) Settsu meisho zue. In: National Diet Library Digital Collections. Available via: https://www.ndl.go.jp/nichiran/e/data/R/039/039-003r.html. Accessed 31 Dec 2020 Anonymous (unknown) Karakuri chigusa no mibae. In: Waseda University Library. Available via: https://www.wul.waseda.ac.jp/kotenseki/search.php?cndbn=機関千種の実生/. Accessed 31 Dec 2020 BBC (2010) Tales of Hans Christian Andersen – the nightingale. In: School radio adapted by Kate Stonham. Available via: http://downloads.bbc.co.uk/schoolradio/pdfs/transcripts/ the_nightingale.pdf. Accessed 31 Dec 2020 Broekhuijsen N (2019) AMD’s 64-Core EPYC and Ryzen CPUs stripped: a detailed inside look. In: Tom’sHardware. Available via https://www.tomshardware.com/news/amd-64-coreepyc-cpu-die-design-architecture-ryzen-3000. Accessed 25 Dec 2020 Hara K (2020) Edo Karakuri volume 5. Bow-towing samurai doll restoration. V2-solutions, Nagoya, Japan. Available via: http://www.v2-solution.com/booklist/978-4-434-28205-8.html. Accessed 7 Jan 2021 Hosokawa Y (1798a) Kik¯o zui. In: Waseda University Library. Available via: https:// www.wul.waseda.ac.jp/kotenseki/html/bunko31/bunko31_e0706/index.html. Accessed 31 Dec 2020 Hosokawa Y (1796b) Kik¯o zui, vol 3. In: Waseda University Library. Available via: https://ukiyoe.org/image/bm/AN00036382_001_l. Accessed 31 Dec 2020 Inuyama Festival (2016) Available via: http://www.inuyama-matsuri.com/history/index.html. Accessed 7 Jan 2021 Karakuri (2021) Available via: https://karakuri.gr.jp/ Accessed 7 Jan 2021 Kyokutei B (1802) Chosakudo Issekiwa (Stories told a certain night at the author’s pavilion). Available via: http://codh.rois.ac.jp/pmjt/book/200020269/. Accessed 7 Jan 2021

4 Karakuri Solutions and Industry 4.0

97

Ministry of Economy, Trade and Industry (METI) (2012) Available via: http://www.meti.go.jp/ press. Accessed 16 Dec 2012 Mujtaba H (2019) AMD 2nd Gen EPYC ROME processors feature a gargantuan 39.54 billion transistors, IO Die Pictured in Detail. In: US online retailers of electronics, technology and office supplies. Available via https://tech-battery.blogspot.com/2019/10/amd-2nd-gen-epycrome-processors.html. Accessed 25 Dec 2020 Niigata City History Museum (2012) Available via: http://www.nchm.jp/ozawake/index2012.html. Accessed 7 Jan 2021 Nishimura S (1758) Okarakuri e-zukushi. In: National Diet Library Digital Collections. Available via: https://dl.ndl.go.jp/info:ndljp/pid/1188162?itemId=info%3Andljp%2Fpid %2F1188162&__lang=en . Accessed 31 Dec 2020 Otona no Kagaku (2021) Available via: https://otonanokagaku.net/products/karakuri/edo/ detail.html. Accessed 7 Jan 2021 Ritchie H (2014) Energy. In: Published online at OurWorldInData.org. Available via https:// ourworldindata.org/energy. Accessed 25 Dec 2020 Roser M, Ritchie H (2013) Technological progress. In: Published online at OurWorldInData.org. Available via https://ourworldindata.org/technological-progress. Accessed 25 Dec 2000 Sakano S (2001) Handmade tea-carrying dolls. We serve both tea and coffee (Science Series). Power-sha, Tokyo. Available via: https://honto.jp/netstore/pd-book_02066337.html. Accessed 7 Jan 2021 Suematsu R (2012) Karakuri doll of two pen writer in Anj¯o city. In: Karakuri Frontier. Available via http://karafro.com/anjomojikaki/mecha_paper.pdf. Accessed 31 Dec 2020 SUS BKK (2016a) Available via: https://www.youtube.com/watch?v=2o8PdOA9t4E. Accessed 27 Dec 2000 SUS BKK (2016b) Available via: https://www.youtube.com/watch?v=P9Hzob_L9_U. Accessed 27 Dec 2020 SUS BKK (2016c) Available via: https://www.youtube.com/watch?v=IQkW0-79A1k. Accessed 27 Dec 2020 SUS BKK (2016d) Available via: https://www.youtube.com/watch?v=4Z9WJyzzx3Q. Accessed 27 Dec 2020 Tagaya K (1730) Karakuri Kinmou Kagami Gusa. In: Karakuri tamaya. Available via: http:// karakuri-tamaya.jp/en/pdf/kinmou.pdf. Accessed 31 Dec 2020 Toshiba Mirai Kagakukan (2021) Toshiba Science Museum. Available via: https://toshiba-miraikagakukan.jp/learn/history/toshiba_history/roots/hisashige/index_j.htm and https://toshibamirai-kagakukan.jp/learn/history/toshiba_history/roots/hisashige/hisashige02_j.htm. Accessed 7 Jan 2021

5

Use of Smart Technologies on Textile Industry Workers to Evaluate the Effect of Work Posture on Lower Extremity Distress in Southern Region of India S. Shankar, R. Naveenkumar, J. Karthick, P. Mohan Kumar, and R. Nithyaprakash

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hand Screen Printing Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evaluation of Distress on Lower Extremity Among Hand Screen Printing Workers . . . . . . . Study Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Collection Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Identification of Critical Work Postures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Instruments Used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Analysis Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Impact of WMSD on Various Anatomical Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Demographics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prevalence of Lower Extremity Disorder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Factors Influencing the Occurrence of Lower Extremity Disorder . . . . . . . . . . . . . . . . . . . . Inference from Postural Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rationale Behind the Occurrence of WMSD on Lower Body Extremity Among Textile Workers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

100 100 102 102 102 103 103 103 104 104 105 105 106 111 113 113

Abstract

The main aim of this study is to investigate the various factors influencing the occurrence of work-related musculoskeletal disorders (WMSDs) among the hand screen printing (HSP) industry workers. A cross-sectional study has

S. Shankar () · J. Karthick · P. M. Kumar · R. Nithyaprakash Department of Mechatronics Engineering, Kongu Engineering College, Perundurai, Erode, Tamil Nadu, India R. Naveenkumar Department of Mechanical Engineering, Kongu Engineering College, Perundurai, Erode, Tamil Nadu, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_148

99

100

S. Shankar et al.

been conducted with 385 workers from 41 hand screen printing units using modified Nordic questionnaire. Rapid Entire Body Assessment (REBA) method is adapted to analyze the work posture and twin axial goniometer is employed to find the frequency of trunk bending. Seventy-five-point-one percent of workers participated in the cross-sectional study reported low back pain, 58.7% of workers reported knee pain, and also 55.6% of workers reported that they were suffering from severe ankle/feet pain. REBA analysis also indicated that posture adapted in the HSP process has high risk to health of workers. The study discusses the effect of working condition such as frequent bending and prolonged standing on the occurrence of WMSDs on lower extremity disorder. Keywords

WMSDs · REBA · Hand screen printing · Goniometer · Low back pain

Introduction Textile industries are one of the traditional occupations in India which generate annual revenue of 100 billion US $ (Shahi et al. 2020). Indian textile industry comprises of many unorganized small processing units like dyeing, printing, knitting, etc. which not only generate revenue but also provide job opportunity to the people in rural and urban areas (Gautam and Lal 2020). Textile industry demands high quality of work from its workers within the time frame to stay ahead in the rising global competition. Workers of these industries are mostly from low economic profile with low literacy rate; the high job demand and lack of ergonomic knowledge in the textile industries increase the workers risk of getting affected with workrelated musculoskeletal disorders (WMSDs) (Krishnamoorthy et al. 2019). WMSDs considerably influence the productivity of the workers by increasing absenteeism and also has great impact on a country’s economy because nearly 250 billion US $ are spent directly or indirectly for its treatment (Bathrinath et al. 2020).

Hand Screen Printing Process Hand screen printing (HSP) involves in manual printing of colored design on the plain cloth using chemical pigments. In HSP industry, workers initially place the cloth on the table of length 35 meter and fix the cloth on the table by using pins as shown in Fig. 1a. After that steel frame embedded with silk cloth containing design template is placed over the cloth. And then chemical pigments are applied on the cloth using wooden squeegee; this process is done by pair of workers during which worker from one end of the table wipes down the chemical pigments using squeegee and pass it to the worker in the other end in order to complete the printing process. After completing wiping process, they place the steel frame to next position on the table and repeat the process. During HSP process, workers bend their body more

5 Use of Smart Technologies on Textile Industry Workers to Evaluate. . .

101

Fig. 1 Process of hand screen printing. (a) Placing cloth on the table. (b) Stabling the cloth. (c) Placing the steel frame. (d) Pouring dye. (e) Applying chemical pigments using squeegee. (f) Drying the printed cloth

frequently and they repeat this process throughout the working hours without sitting. Earlier studies about HSP reports that performing repetitive work at awkward position places the workers of HSP industry at high ergonomic risk (Subramaniam et al. 2018; Shankar et al. 2019). In HSP industry workers averagely work more than eight hours a day by maintaining awkward posture and also they work in standing position throughout the day without enough breaks (Shankar et al. 2017, 2018). Meena et al. (2014) reported that factors like working in poor ergonomic environment, job demand, and relationship with coworkers affect quality of life among HSP workers. Lower back, shoulder, and knee of the body were extremely affected by WMSDs compared with other body region of HSP workers by doing repetitive work (Meena and Dangayach 2015). Ergonomic assessment is necessary in screen printing industries in order to reduce the WMSDs. Several studies conducted with various occupations confirmed that the work involving high degree of repetitive motion for prolonged period of time influences the occurrence of WMSDs mostly on the lower body region (Tomita et al. 2010;

102

S. Shankar et al.

Akodu and Ashalejo 2019) (low back, knee, and ankle). A study conducted with Iranian carpet mending workers also stated that performing their work at awkward posture for prolonged period of time is the main reason for the occurrence of WMSDs on lower body extremity (Candan et al. 2019). Study conducted with male kitchen workers revealed that working in awkward posture and standing for prolonged period of time is the main reason for occurrence of WMSDs on lower body region (Subramaniam and Murugesan 2015). The aim of our study is to investigate the effect of lower extremity disorder among the hand screen printing industry workers from southern region of India, Tamil Nadu. Up to our knowledge based on the literature very negligible amount of research studies are only conducted with HSP industry to report the prevalence of WMSD experienced by HSP industry workers. But there are no enough studies carried out to find the root causes for occurrence of WMSDs on lower body extremity, especially low back, knee, and ankle/feet which is highly affected (Meena and Dangayach 2015; Shankar et al. 2017). In this study, the detail information about the influence of adopted work posture, working experience, and age of the workers on lower extremity distress is reported. Since large number of workers depend on this industry for their livelihood this kind of study to make their workplace ergonomically safe will increase their productivity and health.

Evaluation of Distress on Lower Extremity Among Hand Screen Printing Workers Study Participants Totally 385 workers with work experience not less than one year and age between 18 and 60 years were randomly selected based on Cochran’s sampling method (Kotrlik and Higgins 2001; Kothari 2004) from 41 hand screen printing industries located in various districts of Tamil Nadu, India. Among this population 211 (54.8%) were male and 174 were female (45.2%). Workers with good health and without any history of bone dislocation and surgery were only allowed to participate in this survey. Before conducting questionnaire, aim and scope of the study was well explained to both employers and workers participated in the study. The oral consent was obtained from both workers and their employers. This study was conducted in accordance with ethical guidelines for biomedical research on human participants provided by Indian council for medical research (Shah 2005).

Data Collection Method A self-administrative survey was conducted by modifying English version of Standard Nordic Musculoskeletal Questionnaire (NMQ) into four parts (Kuorinka et al. 1987). First part of the questionnaire had questions related to their personal information, second and third section dealt with psychological and medical history,

5 Use of Smart Technologies on Textile Industry Workers to Evaluate. . .

103

respectively. Final section had questions related to WMSDs. The survey was done through face-to-face interaction with the hand screen printing workers in their local language (Tamil). The procedure of the study was briefed to the workers before interrogation. Interview was conducted during the break hours in the morning and noon.

Identification of Critical Work Postures REBA (Rapid Entire Body Assessment) method is used to identify the awkward/non-neutral posture maintained by workers during hand screen printing process (Hignett and McAtamney 2000). For this purpose workers movement and postures during the printing process were videographed using digital camera (Sony Cyber-shot). Then REBA scores indicating risk in posture during HSP process were calculated by considering parameters such as common posture adapted for trunk (four postures), legs (two postures), neck (two postures), upper arm (four postures), and lower arm (two postures), coupling factors and load/force transfer (Kong et al. 2018). REBA score indicated the level of risk present in the present working posture.

Instruments Used To find the frequency of bending and trunk bending angle in order identify the degree of repetitiveness during HSP process twin axial Biometric SG110 goniometer on the lateral side of abdomen and gluteus muscles as shown in Fig. 2 was used. Biometrics DataLog EMG system was used to acquire signal from twin axial goniometer at 50 samples per second. Eight subjects with mean age of 37 ± 13.83 years, average height of 162 ± 5.27 cm, and weight 61.25 ± 4.36 kg, who reported low back pain during the questionnaire survey, participated in the above-said analysis to find the degree of repetitiveness and maximum trunk bending angle during continuous printing process.

Data Analysis Technique Data collected through the questionnaire survey was analyzed using the statistical package for social science (SPSS v. 21). Chi square test, odds ratio, and descriptive analysis were performed on the given data. Work experience of HSP workers were categorized into four groups, namely 15 years. One-way ANOVA test was performed to check the significance difference between the demographic factors reported in Table 1. Post hoc test was conducted to find the significant difference among the subgroups.

104

S. Shankar et al.

Fig. 2 Worker attached with universal goniometer

Table 1 Demographic characteristics of HSP workers

Age (years) Weight (kg) Height (cm) Experience (years) Average working hours/day

≤ %5 years (n = 72) 28.4 ± 7.3

6–10 years (n = 94) 34.5 ± 7.21

11–15 years (n = 83) 38.69 ± 9.2

≥ 15 years (n = 136) 46.4 ± 7.12

Significance (p < 0.05) F 61.31 SIG

57.59 ± 8.28

56.56 ± 6.3

57.04 ± 6.7

57.48 ± 6.5

0.4

NS

158.08 ± 7.7

158.5 ± 7.1

156.75 ± 7.6

159.32 ± 8.4

0.1

NS

13.59 ± 8.09

NA

NA

10.8 ± 1.8

NA

NA

SIG Significant, NS Not significant, NA Not applicable

Impact of WMSD on Various Anatomical Site Demographics Demographic characteristics of the sample population participated in the crosssectional study is shown in the Table 1. Here the analysis of variance test (ANOVA) was conducted between work experience and other demographic by dividing the

5 Use of Smart Technologies on Textile Industry Workers to Evaluate. . .

105

Fig. 3 Reported WMSDs on different body parts in cross-sectional survey

work experience into four groups in order to find the significant difference existing between the two groups. Result shows that significant difference exists between work experience and age.

Prevalence of Lower Extremity Disorder Figure 3, showing the effect of work-related musculoskeletal disorder on different body sites, indicated that 75.1% of workers (n = 289) were suffering from low back pain, 66.2% (n = 255) had shoulder pain, and 58.7% (n = 226) had knee pain followed by ankle and feet pain 55.6% (n = 214), 29.4% reported elbow pain, and 24.2% reported wrist pain. It is clear from Fig. 3 that the low back pain is more prevalent which is followed by shoulder, knees, and ankle/feet. Compared to upper extremity region like shoulder, elbow, and wrist the lower body parts experience relatively more WMSDs. So, we decided to discuss about various factors influencing the occurrence of WMSDs on lower extremity (low back, knee, and ankle/feet).

Factors Influencing the Occurrence of Lower Extremity Disorder Low Back Pain In our statistical analysis, we categorized the workers data into three subgroups based on their age as: 40 years of age. Table 2 shows that 18.2% of workers within the age of 40 years were affected with

106

S. Shankar et al.

low back pain. Here, workers belonging to the age group >40 years had nearly three times (OR-2.898, CI 1.630-5.151) more chance of getting low back pain compared to the workers of other age group. Based on the years of experience workers of HSP industries participated in the survey are grouped into four categories as shown in Table 1. In that, 28.3% of workers with more than 15 years of experience had reported severe low back pain (OR-1.548, CI 0.934-2.563) compared to other three categories. From the results, it was also found that the workers with height < 150 cm (OR-2.103, CI 1.145-3.862) were highly affected with the low back pain compared to the workers with height more than 150 cm.

Knee The questionnaire survey results found that 9.4% of workers belonging to age group less than 30 years (OR = 0.195, CI 0.122–0.312), 23.9% workers belonging to the age group of 31–40 years (OR = 2.215, CI 1.337–3.282), 23.1% of workers from age group more than 40 years experience knee pain (OR = 2.095, CI 1.337–3.282) (Table 2). It is also evident that the workers belonging to the age group of 31– 40 years and more than 40 years had two times more chance of getting affected by knee pain compared to the other age groups. Also, the workers with 15 years of experience had nearly two times more chances of getting affected by knee pain (OR = 1.980, CI 1.277–3.070). Ankle/Feet Increasing age and work experience had great influence on the occurrence of WMSDs on ankle and feet. Such workers belonging to age group of more than 40 years had two times (OR = 2.448, CI 1.562–3.835) more chances for the occurrence of WMSDs on ankle and feet, whereas the workers with more than 15 years of experience had nearly two times (OR = 1.909, CI 1.239–2.940) more chances of getting affected by WMSDs on ankle and feet when compared to other groups (Table 2). Chi square result (Table 2) shows that there is significance difference between age and experience. Whereas other factors like height, jobrelated stress, educational qualification, and smoking habit did not have significance with occurrence of WMSDs on lower body extremity.

Inference from Postural Analysis Table 3 shows the most common work posture adapted during entire HSP process and their relative REBA score. REBA scores for the aligning of cloth (REBA score 9) and drying of cloth (REBA score 10) processes indicated that posture adapted has high postural risk and almost requires revision in their work posture. Obtained REBA (REBA score 12) score for applying the chemical pigments indicates that posture adapted during this process involves very high risk and needs immediate revision in the work posture. Whereas stabling of cloth process had medium-level risk.

Experience < 5 years 6–10 years 11–15 years > 15 years

46 70 64 109

11.9 18.2 16.6 28.3

0.510 0.959 1.153 1.548

Risk factor. Low back pain (75.1%) Odds ratio n (385) % Age 40 111 28.8 2.898

0.294– 0.883 0.562– 1.637 0.649– 2.046 0.934– 2.563

0.05

p value

0.4392.840 >0.05 0.326– 1.216 0.569– 2.139 0.655– 1.617 1.667– 2.002

0.529– 1.335 0.767– 1.715 0.543– 1.710 0.537– 2.375

Ankle and feet (55.6%) Odds ratio 95% CI 95% CI p value n (385) %

108 S. Shankar et al.

95 132 64 11

50 238

significant

EducationIlliterate Elementary High school Degree

SmokingSmoker Nonsmokers

a Statistically

13.0 61.8

24.7 34.3 16.6 2.9

1.050 0.993

1.861 1.283 1.153 1.860

0.567– 1.947 0.504– 1.728

1.073– 3.226 0.802– 2.053 0.649– 2.046 0.405– 8.543 >0.05

>0.05

34 188

83 100 49 4

8.9 48.8

21.6 26 12.7 1

0.735 1.322

2.408 1.069 1.060 0.311

0.432– 1.251a 0.779– 2.224a

1.501– 3.863 0.711– 1.607 0.647– 1.736 0.094– 1.027 >0.05

>0.05

28 184

81 87 55 5

7.3 47.8

21.0 22.6 14.3 1.30

0.537 1.799

2.491 0.740 1.790 0.493

0.314– 0.918a 1.057– 3.063a

1.563– 3.970a 0.490– 1.110a 1.077– 2.975a 0.158– 1.535a >0.05

>0.05

5 Use of Smart Technologies on Textile Industry Workers to Evaluate. . . 109

110

S. Shankar et al.

Table 3 REBA score calculated during HSP process Working posture

Process Aligning the cloth on the table

REBA score 9

Remarks High risk level-needs soon reconsideration of work posture

Stabling of cloth

7

Medium risk level – Needs reconsideration of work

Applying chemical pigments on cloth

12

Very high risk – Requires immediate reconsideration of work posture

Drying of cloth fabric

10

High risk level-needs soon reconsideration of work posture

5 Use of Smart Technologies on Textile Industry Workers to Evaluate. . .

111

Fig. 4 Twin axial goniometer result showing trunk bending angle during hand screen printing process

Rationale Behind the Occurrence of WMSD on Lower Body Extremity Among Textile Workers It was found from questionnaire survey that lower body region was significantly affected with WMSDs compared to upper body extremity as shown in Fig. 2 and also we found that low back pain is most commonly reported WMSDs in lower body region which was followed by knee and ankle/feet. Twin axial universal goniometer recording (Fig. 4) shows that HSP workers averagely bend 36◦ for applying chemical pigments on cloth using squeegee; it is also found that time taken for wiping of squeegee operation is 3–4 s and time for moving the steel frame to next position for performing consecutive wiping process takes 2 s. During the passing of squeegee across the table for applying chemical pigments on cloth, the workers of HSP have to bend their trunk forward twice, namely primary bending and secondary bending within a short span of 6 s for the passing and receiving the squeegee. Thus, this frequent bending of their trunk might be the reason for the occurrence of low back pain. Figure 5 shows the magnified view of the goniometer recording showing primary and secondary bending. Previous studies also confirmed that bending the trunk more frequently during the work is the main cause for the occurrence of low back pain (Osborne et al. 2012; Durlov et al. 2014). Next to low back pain most prevalent WMSD region is knee and ankle/feet region. Because of these sustained standing and perennial movement during their work quadriceps femor muscle in the front thigh gets tightened and becomes inflexible (Lewek et al. 2004; Murray et al. 2015). This injures muscles which ultimately cause knee pain and the same thing happens to anterior muscle group which affects ankle and feet. Table 2 shows the results from the present study which went well with the survey conducted with professionals, whose nature of job involved continuous movement and repeated motion, reporting pain in knee and ankle (Reed et al. 2014; Shankar et al. 2015). Important factor greatly contributing to the occurrence of WMSDs on knee might be the quadriceps angle (Q angle), since workers of hand screen printing industries have to stand and move repeatedly

112

S. Shankar et al.

Fig. 5 Magnified view of primary bending and secondary bending during squeegee wiping operation

for performing activities as shown in Fig. 1. In the tasks like placing the cloth on the table, moving the squeegee, and placing the cloth on the overhead hanger, there is a chance for an increase in quadriceps angle on a longtime perspective and also it might impinge a bilateral asymmetry in quadriceps angle (Hvid and Andersen 1982; Livingston and Mandigo 1999). This increase in the quadriceps angle increases the risk of knee injury as excess quadriceps angle eventually increases the tension of muscles in the front thigh that pulls out knee cap which causes the knee to wear, causing knee injury (Emami et al. 2007). Apart from the above-said nature of the job in HSP industry another significant finding from the study is that increasing age has an association with WMSDs on lower body extremity because aging changes the neurological condition of body and affects its function when it is overloaded and strained by the prolonged muscle activity (Lauretani et al. 2003). Several studies also proved that aging has an influence on occurrence of WMSDs (Heiden et al. 2013; Tonelli et al. 2014). It was found from the present study that the workers who worked in HSP industry for many years (>15 years) were at high risk of developing the WMSDs compared to the worker with lesser experience due to the fatigue loading undergone for many years which might reduce the flexibility of their muscles (Häkkinen and Komi 1983). Statistical studies conducted by Coury et al. (2002) and Nag et al. (2010) confirmed that the longer job tenure had an effect on the occurrence of WMSDs. Findings from work posture analysis using REBA revealed that commonly adapted postures in HSP industries involve high risk to the health of the workers and so it recommends reconsideration. This might be due to the awkward posture involved in the HSP process, such as repetitive motion and frequent bending. Postural studies conducted with different occupation found that working at awkward posture causes threat to the health especially in the low back, knee, and ankle/feet (Le Manac’h et al. 2012; Sarkar et al. 2016). Through direct observation of HSP

5 Use of Smart Technologies on Textile Industry Workers to Evaluate. . .

113

process and interaction with HSP workers during the questionnaire survey, we found that repeated movement, frequent bending of the trunk, and standing throughout the day were major factors influencing the occurrence of WMSDs among the HSP workers. Generally, workers of HSP industry print colored design on 600–1000 m of cloth based on orders received for about 10–12 h a day. Thus, our primary suggestion is that workers of HSP should be given awareness on proper ergonomic posture and also they must be trained to perform stretching exercise between the works in order to increase their muscle flexibility to avoid WMSDs injuries. Another significant risk factor involved in the work is frequent bending of the trunk. It is not possible to completely eliminate the bending movement but the angle and frequency could be reduced through proper ergonomic intervention. Regulating the working hours by providing micro- and macrolevel breaks and allowing the workers to sit on regular intervals between the works would also help in reducing the intensity of the knee and ankle/feet pain.

Conclusion The present study was carried out on the hand screen printing industry workers to find the effect of work posture on the work-related musculoskeletal distress on body lower extremity. Statistical survey carried out using questionnaire survey reported that low back region is high prevalence to WMSD followed by ankle/foot. Increasing age, higher work experience, frequent bending movement, and prolonged standing posture are found to be the risk factors contributing to the occurrence of WMSD. Work posture adopted using REBA method found that present work posture adopted during printing process has very high risk and needs immediate change in the posture in order to prevent the occurrence of WMSD. It is suggested to provide proper ergonomic awareness and intervention to prevent the occurrence of WMSD. Acknowledgments The authors would like to acknowledge the printing industry managers and workers participating in this study for valuable assistance and patiently reporting during their working hours. Funding The author(s) disclosed receipt of the following financial support for the research, and authorship for this Chapter: This study was funded by Science Engineering and Research Board (SERB) Department of Science and Technology, India, through Fast Track Young Scientist Scheme Research Project.

References Akodu AK, Ashalejo ZO (2019) Work-related musculoskeletal disorders and work ability among hospital nurses. J Taibah Univ Medical Sci 14:252–261 Bathrinath S, Bhalaji R, Saravanasankar S (2020) Risk analysis in textile industries using AHPTOPSIS. Mater Today: Proc

114

S. Shankar et al.

Candan SA, Sahin UK, Ako˘glu S (2019) The investigation of work-related musculoskeletal disorders among female workers in a hazelnut factory: prevalence, working posture, workrelated and psychosocial factors. Int J Ind Ergon 74:102838 Coury HJCG, Porcatti IA, Alem ME, Oishi J (2002) Influence of gender on work-related musculoskeletal disorders in repetitive tasks. Int J Ind Ergon 29:33–39 Durlov S, Chakrabarty S, Chatterjee A, Das T, Dev S, Gangopadhyay S, Haldar P, Maity SG, Sarkar K, Sahu S (2014) Prevalence of low back pain among handloom weavers in West Bengal, India. Int J Occup Environ Health 20:333–339 Emami M-J, Ghahramani M-H, Abdinejad F, Namazi H (2007) Q-angle: an invaluable parameter for evaluation of anterior knee pain. Arch Iran Med 10:24–26 Gautam S, Lal M (2020) Analysing competitiveness and trade performance: evidence from Indian textile industry and its select competitors. Transnatl Corp Rev 12(4):1–19 Häkkinen K, Komi PV (1983) Electromyographic and mechanical characteristics of human skeletal muscle during fatigue under voluntary and reflex conditions. Electroencephalogr Clin Neurophysiol 55:436–444 Heiden B, Weigl M, Angerer P, Müller A (2013) Association of age and physical job demands with musculoskeletal disorders in nurses. Appl Ergon 44:652–658 Hignett S, McAtamney L (2000) Rapid entire body assessment (REBA). Appl Ergon 31:201–205 Hvid I, Andersen LI (1982) The quadriceps angle and its relation to femoral torsion. Acta Orthop Scand 53:577–579 Kong Y-K, Lee S-Y, Lee K-S, Kim D-M (2018) Comparisons of ergonomic evaluation tools (ALLA, RULA, REBA and OWAS) for farm work. Int J Occup Saf Ergon 24:218–223 Kothari CR (2004) Research methodology: methods and techniques. New Age International, Uttar Pradesh Kotrlik JWKJW, Higgins C (2001) Organizational research: determining appropriate sample size in survey research appropriate sample size in survey research. Inf Technol Learn Perform J 19:43 Krishnamoorthy VH, Kuberan D, Gopichandran V (2019) Prevalence, patterns and disability due to musculoskeletal disorders among cotton textile industry workers in Tamil Nadu – a cross sectional study. Int J Community Med Public Health 9:66–70 Kuorinka I, Jonsson B, Kilbom A, Vinterberg H, Biering-Sørensen F, Andersson G, Jørgensen K (1987) Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms. Appl Ergon 18:233–237 Lauretani F, Russo CR, Bandinelli S, Bartali B, Cavazzini C, Di Iorio A, Corsi AM, Rantanen T, Guralnik JM, Ferrucci L (2003) Age-associated changes in skeletal muscles and their effect on mobility: an operational diagnosis of sarcopenia. J Appl Physiol 95:1851–1860 Le Manac’h A, Ha C, Descatha A, Imbernon E, Roquelaure Y (2012) Prevalence of knee bursitis in the workforce. Occupational medicine 62:658–660 Lewek MD, Rudolph KS, Snyder-Mackler L (2004) Quadriceps femoris muscle weakness and activation failure in patients with symptomatic knee osteoarthritis. J Orthop Res 22:110–115 Livingston LA, Mandigo JL (1999) Bilateral Q angle asymmetry and anterior knee pain syndrome. Clin Biomech 14:7–13 Meena M, Dangayach G (2015) An ergonomic approach to design hand tool for screen textile printing. Int J Rec Adv Mech Eng 4:59–67 Meena M, Dangayach G, Bhardwaj A (2014) Measuring quality of work life among workers in handicraft industries of Jaipur. Int J Ind Syst Eng 17:376–390 Murray AM, Thomas AC, Armstrong CW, Pietrosimone BG, Tevald MA (2015) The associations between quadriceps muscle strength, power, and knee joint mechanics in knee osteoarthritis: a cross-sectional study. Clin Biomech 30:1140–1145 Nag A, Vyas H, Nag P (2010) Gender differences, work stressors and musculoskeletal disorders in weaving industries. Ind Health 48:339–348 Osborne A, Finnegan G, Blake C, Meredith D, McNamara J, Phelan J, Cunningham C (2012) An evaluation of low back pain among farmers in Ireland. Occup Med 63:53–59 Reed LF, Battistutta D, Young J, Newman B (2014) Prevalence and risk factors for foot and ankle musculoskeletal disorders experienced by nurses. BMC Musculoskelet Disord 15:196

5 Use of Smart Technologies on Textile Industry Workers to Evaluate. . .

115

Sarkar K, Dev S, Das T, Chakrabarty S, Gangopadhyay S (2016) Examination of postures and frequency of musculoskeletal disorders among manual workers in Calcutta, India. Int J Occup Environ Health 22:151–158 Shah N (2005) Ethical guidelines for biomedical research on human subjects. Trends Biomater Artif Organs 18:174–177 Shahi SK, Shiva A, Dia M (2020) Integrated sustainable supply chain management and firm performance in the Indian textile industry. Qual Res Organ Manag Shankar S, Kumar RN, Mohankumar P, Karthick J, Pradeep S (2018) Ergonomics in caring for people. Springer, Uttar Pradesh, pp 131–137 Shankar S, Naveen Kumar R, Mohankumar P, Jayaraman S (2017) Prevalence of work-related musculoskeletal injuries among south Indian hand screen-printing workers. Work 58:163–172 Shankar S, Naveen Kumar R, Mohankumar P, Srinivasan J (2017) Association of Job and Demographical Risk Factor with Occurrence of Neck Pain Among Hand Screen Printing Worker. Work 58:163–172 Shankar S, Naveenkumar R, Karthick J (2019) Management of musculoskeletal shoulder and neck pain through ergonomic intervention: a pre-post design analysis in hand screen printing industry. Int J Bus Innov Res 18:392–409 Shankar S, Shanmugam M, Srinivasan J (2015) Workplace factors and prevalence of low back pain among male commercial kitchen workers. J Back Musculoskelet Rehabil 28:481–488 Subramaniam S, Murugesan S (2015) Investigation of work-related musculoskeletal disorders among male kitchen workers in South India. Int J Occup Saf Ergon 21:524–531 Subramaniam S, Raju N, Jeganathan K, Periyasamy M (2018) Evaluation of vibrant muscles over the shoulder region among workers of the hand screen printing industry. Int J Occup Saf Ergon 24:278–285 Tomita S, Arphorn S, Muto T, Koetkhlai K, Naing SS, Chaikittiporn C (2010) Prevalence and risk factors of low back pain among Thai and Myanmar migrant seafood processing factory workers in Samut Sakorn Province, Thailand. Ind Health 48:283–291 Tonelli S, Culp K, Donham K (2014) Work-related musculoskeletal disorders in senior farmers safety and health considerations. Workplace health Saf 62:333–341

6

Monitoring and Modeling of Cylindricity Error Using Vibration Signals in Drilling J. Susai Mary, D. Dinakaran, M. A. Sai Balaji, S. Satishkumar, and Arockia Selvakumar Arockia Doss

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Materials and Methods for Cylindricity Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cylindricity Measurements Using Vibration Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modeling of Cylindricity Error Using ANFIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

118 120 120 122 124 124

J. Susai Mary () Department of Electronics and Instrumentation Engineering, Chennai, Tamil Nadu, India e-mail: [email protected] D. Dinakaran Centre for Automation and Robotics (ANRO), School of Mechanical Sciences, Hindustan Institute of Technology and Science, Chennai, TN, India e-mail: [email protected] M. A. Sai Balaji Department of Mechanical Engineering, B.S.A Crescent Institute of Science and Technology, Chennai, Tamil Nadu, India e-mail: [email protected] S. Satishkumar Department of Mechanical Engineering, Velammal Engineering College, Chennai, Tamil Nadu, India A. S. Arockia Doss Design and Automation Research Group, School of Mechanical Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_149

117

118

J. Susai Mary et al.

Abstract

Cylindricity is a key factor that affects the assembly and dynamic performance characteristics of mechanical components. Real-time monitoring of cylindricity error during machining reduces the wastages and ensures the quality of the drilled holes. The possible causes of cylindricity error are the run out, mechanical looseness, misalignment, etc. A novel vibration-based monitoring method to predict the cylindricity error during drilling process is developed and presented in this chapter. An adaptive neuro fuzzy system models the cylindricity error with the peak amplitude of vibration at 1xRPM, process parameters, speed, and feed as inputs. The experimental study showed that the ANFIS model is capable of predicting the cylindricity error with 92.4% of accuracy. Keywords

Cylindricity error · ANFIS · Vibration · Drilling

Introduction Drilling is a primary machining process for making holes in the component. Drilling includes 40% of the metal-cutting operations and plays a key role in the assembly of parts. Drilling finds application in many industries such as aerospace, manufacturing, and automobile. In all these sectors, thousands of holes are required for the assembly of different parts using screws, bolts, rivets, etc. (Liu et al. 2012). Deep hole drilling is a highly complicated process. The stiffness and stability of the tool makes the tool shaft prone to deflection causing form errors and affecting the hole quality (Arunkumar et al. 2018). Bad quality holes result in improper assembly which leads to high cost and waste of time (Aamir et al. 2019). Cylindricity is a major form error which results in improper assembly of mechanical parts. Cylindricity describes how the hole surface deviates from a perfect cylinder with respect to geometrical tolerance in 3 Dimension (https://www.taylorhobson.com.de/media/ametektaylorhobson/files/learning-zone/ training-material/cylindricity). Some of the direct methods of cylindricity measurements involve the Radial methods (Adamczak et al. 2011) and Error Separation Techniques (EST) (Chen et al. 2016). Indirect measurement systems include vibration, force, current, acoustic emission, ultrasonics, etc. The methods and advancements in the field of cylindricity measurements are represented in Fig. 1. The effects of the cutting force, the feed rate, and the coolant pressure on the quality of the holes in deep-hole drilling were studied by Aamir et al. The optimal process parameters were identified through a gray relational optimization technique (Aamir et al. 2019). A study on the influence of drilling parameters on the tool wear and hole quality was performed, and it was found that lower cutting speed and feed rate minimize the cylindricity errors (Sultan et al. 2015). Miguel et al. studied the optimum combination of drilling parameters for dry drilling of carbon

6 Monitoring and Modeling of Cylindricity Error Using Vibration Signals in Drilling

119

Fig. 1 Methods of cylindricity measurements

fiber-reinforced plastic (CFRP) composite materials and observed that at a speed of 105 m/min and feed of 300 mm/min, the cylindricity was considerably increased by 5–91% (Alvarez-Alcon et al. 2020). A new geometric method for real-time measuring of deep hole’s axis using ultrasonic thickness measurement was developed by Liu et al., and it was found that this method can measure the deviation and direction accurately (Liu et al. 2018). The effects of spindle speed, feed rate, and vibration amplitude on the circularity error and cylindricity error were studied and simulated. The correlations between the measured and simulated values for the measurement of circularity error were quite encouraging (Reddy and Prasad 2016). From the literatures cited above, it is clear that the cylindricity error in drilling was measured in off-line, and the effect of various parameters such as speed, feed, cutting force, and vibration was studied by many researchers. But, an online monitoring system to predict the cylindricity error is not yet attempted. This paves the way for this research work on developing an online monitoring system using Adaptive Neuro Fuzzy Inference System (ANFIS) models to predict the cylindricity error in drilling in real time. Vibration measurements are widely used in the literature for the analysis of unbalance in the system due to improper holding of tool, overextension of tool, accumulation of debris, etc., which are also the reasons for tool runout, breakage of tools, etc. (Dinakaran et al. 2009; Susai Mary et al. 2017; Uddin et al. 2018). Various machine learning techniques such as Neural Network, Fuzzy Logic, ANFIS, and SVM (Support Vector Machines) were used to model and predict the tool wear, run out, surface roughness, etc., (Susai Mary et al. 2015; Susai Mary et al. 2019; Susai Mary et al. 2019). A neural network model to predict the circularity of drilled holes with different sizes of data set was presented by Umesh et al., and the model shows a regression (R) of 0.99718 (Umesh Gowda et al. 2014). An ANFIS model to predict the hole quality in microdrilling with different features extracted from vibration signals in wavelength domain was developed and found to

120

J. Susai Mary et al.

exert a good agreement with the experimental results (Ranjan et al. 2020). A realtime monitoring to improve the machining efficiency and tool life using ANFIS and vibration and communication particle swarm optimization (VCPSO) algorithm was proposed by Xu et al. and was found to be efficient in predicting tool wear (Xu et al. 2020). This chapter presents an online monitoring method to predict the cylindricity error of the drilled holes. Vibration signals analysis at the unbalance frequency 1× RPM is used for detecting the presence of cylindricity error. The feed and machining speed affect the quality of the drilled holes which are also taken as inputs for modeling. The developed model was able to predict the cylindricity error with an accuracy of 92.4%. The prediction of cylindricity error in online helps in taking corrective actions thereby saving cost, wastages, and improving the quality of drilled holes.

Materials and Methods for Cylindricity Monitoring Drilling operations were performed on a 120 mm circular workpiece of thickness 10 mm made of EN24 steel. Circular components with Pitch Circle Diameter (PCD) drilling are used in many automotive applications as in flange couplings, gears, etc. The tool chosen is a 6 mm HSS twist drill bit. The machining parameters are chosen based on ASTM standards. For a 6 mm HSS tool to drill an EN24 steel workpiece, the recommended range of speed varies from 724 to 1273 RPM and feed range is 59–191 mm/min. The drilling experiments were performed based on an L25 orthogonal array with speed ranges of 800, 950, 1100, 1300, and 1450 RPM. The feed levels were 75, 90,100,130, and 140 mm/min. The machining was done on an LMW (LV-45) 3 axis milling machine. An accelerometer sensor KISTLER (Model: 8636C50) was used to measure the vibration signals during drilling. A data acquisition card (DAC) model (NI9233) was used to acquire the vibration signals using the sound and vibration module of LabVIEW software. The machining setup with the positioning of accelerometer sensor and the DAC is shown in Fig. 2. Based on L25 settings, drilling experiments were carried out and another ten experiments were conducted for validation purposes. The vibration signals were measured after each drilling operation, and the cylindricity error of the drilled holes was measured using a Co-ordinate Measuring Machine (CMM) (Model: Tesa Micro-hite 3D).

Cylindricity Measurements Using Vibration Signals The measured cylindricity error values for all the 35 holes varied from 0.011 mm to 0.045 mm. A Fast Fourier Transforms (FFT) analysis was performed to analyze the effects of cylindricity error on the acquired vibration signals. The frequency analysis of the measured vibration signals showed a sharp rise in peak amplitude at 1×RPM which is also the unbalance frequency. This shows the presence of unbalance in the tool which causes the tool to rotate out of axis causing the cylindricity error

6 Monitoring and Modeling of Cylindricity Error Using Vibration Signals in Drilling

121

Fig. 2 Experimental details

Fig. 3 Sample vibration and frequency signal for circularity error 0.027 mm

in the drilled holes (Uddin et al. 2018). A sample of the vibration signal with the frequency graphs for minimum and maximum cylindricity errors is given in Figs. 3 and 4. It was observed that at places where the cylindricity error is high, the vibration amplitude at 1×RPM is also high. The variations of vibration amplitude at 1×RPM with the cylindricity error are shown in Fig. 5. It is found that the cylindricity error is high even at other places where the vibration peak is less, which is due to the influence of speed and feed of machining (Liu et al. 2012; https://www.taylorhobson.com.de/media/ametektaylorhobson/ files/learning-zone/training-material/cylindricity; Adamczak et al. 2011). Hence,

122

J. Susai Mary et al.

Fig. 4 Sample vibration and frequency signal for circularity error 0.045 mm

Fig. 5 Variations of vibration peak at 1xRPM with cylindricity error

for better prediction of cylindricity error, the machining parameters, speed, and feed are also taken as inputs during the modeling of ANFIS. The cylindricity error may also be caused due the workpiece geometry and tool wear. Hence, in future work, the properties of the workpiece material and tool wear can also be included as inputs for modeling the cylindricity error.

Modeling of Cylindricity Error Using ANFIS ANFIS is an algorithm which finds widespread application and involves a hybrid modeling technique that includes fuzzy logic and neural network (Susai Mary et al. 2019). Drilling operations are nonlinear, and the vibration signals also change significantly with changes in circularity of the drilled holes. Thus, ANFIS is employed in this study to adapt to the fluctuations during machining. The inputs to the ANFIS model are the peak vibration frequency at 1×RPM and the speed and

6 Monitoring and Modeling of Cylindricity Error Using Vibration Signals in Drilling

123

Fig. 6 (a) Training output of ANFIS model; (b) validation of ANFIS model

feed of machining. The output of the model is the cylindricity error. The parameters of the ANFIS model were selected based on the literature and also in a trial and error basis to select the parameter which gives the best fit for the data. A Grid Partition Fuzzy Inference System with 2 Gaussian Membership Functions (GMF’s) for speed, 3GMFs for feed, and 3GMFs for peak vibration at 1×RPM were selected. A hybrid neural network was used for training the network with a maximum of 500 epochs. Out of the 35 experimental data, 25 data were randomly selected for training the network and 9 were used for validating the trained network. One data was left out as it was an outlier. The training and validation error were very minimum of 0.18% and 7.6%. It was found that the model is capable of predicting the cylindricity error of the drilled holes with an accuracy of 92.4%. The training and validation data outputs of the ANFIS model are given in Fig. 6. The model is suitable for mass production where the machining parameters are set as constant throughout the production. However, the changes in the process

124

J. Susai Mary et al.

parameters can be accommodated through a simple data-capturing methodology and train them for a new set of conditions. The peak frequency considered for ANFIS input can be changed with respect to the set speed.

Conclusion This chapter gives a novel online prediction method using the vibration measurements at 1×RPM and machining parameters to detect the cylindricity error of the drilled holes. The vibration peak amplitude at the unbalance frequency increases up to 3 times with increase in cylindricity error. The vibration amplitude and the machining parameters are used as input for the ANFIS model to predict the cylindricity error with an accuracy of 92.4%. The real-time monitoring is necessary for accurate machining operations and also to reduce the form errors by taking remedial actions. A multisensory approach can be adopted in future for the prediction of cylindricity error to increase the accuracy of prediction.

References Aamir M, Tolouei-Rad M, Giasin K, Nosrati A (2019) Recent advances in drilling of carbon fiber– reinforced polymers for aerospace applications: a review. Int J Adv Manuf Technol 105:2289– 2308 Adamczak S, Janecki D, Stepien K (2011) Cylindricity measurement by the V-block method – theoretical and practical problems. Measurement 44(1):164–173 Liu D, Tang Y, Cong W (2012) A review of mechanical drilling for composite laminates. Compos Struct 94:1265–1279 Arunkumar N, Thanikasichalam A, Sankaranarayanan V, Senthilkumar E (2018) Parametric optimization of deep-hole drilling on AISI 1045 steel and online tool condition monitoring using an accelerometer. Mater Manuf Process 33(6):1751–1764 https://www.taylorhobson.com.de/media/ametektaylorhobson/files/learning-zone/trainingmaterial/cylindricity Chen Q, Tao X, Jinshi L, Wang X (2016) Cylindricity error measuring and evaluating for engine cylinder bore in manufacturing procedure. Adv Mater Sci Eng 4212905:1–7 Sultan Z, Sharif S, Kurniawan D (2015) Effect of machining parameters on tool Wear and hole quality of AISI 316L stainless steel in conventional drilling. Procedia Manufacturing 2:202– 207 Alvarez-Alcon M, Lopez LN, de Lacalle F, Fernandez-Zacarias (2020) Multiple sensor monitoring of CFRP drilling to define cutting parameters sensitivity on surface roughness. Cylindricity and Diameter Materials 13(12):2796 Liu J, Wu F, Gao X, Li R (2018) A real-time method to measure the deviation of deep hole. Advances in intelligent systems research (AISR), international conference on computer modeling. Simulation and Algorithm (CMSA) 151:216–219 Reddy YRM, Prasad BS (2016) Simulation of form tolerances using CMM data for drilled holesan experimental approach. Journal of Production Engineering 19(2):77–83 Susai Mary J, Sabura Banu U, Dinakaran D, Nakandhrakumar RS (2017) Adaptive control by multi-objective optimisation for drilling process with fuzzy inference system and neural predictive controller. Insight – Non-Destructive Testing & Condition Monitoring, The Journal of The British Institute of Non-Destructive Testing 59(1):38–44

6 Monitoring and Modeling of Cylindricity Error Using Vibration Signals in Drilling

125

Dinakaran D, Sampathkumar S, Sivashanmugam N (2009) An experimental investigation on monitoring of crater wear in turning using ultrasonic technique. Int J Mach Tools Manuf 49(15):1234–1237 Uddin M, Basak A, Pramanik A, Singh S, Krolczyk GM, Prakash C (2018) Evaluating hole quality in Drilling of Al 6061 alloys. Materials 11:2443 Susai Mary J, Sabura Banu U, Dinakaran D (2015) Adaptive optimization using grey relational analysis and PID control of CNC drilling process. 2015 International conference on Robotics, Automation, Control and Embedded Systems (RACE), Chennai, pp. 1–5 Susai Mary J, Sai Balaji MA, Dinakaran D (2019) Prediction and geometric adaptive control of surface roughness in drilling process. FME Transactions 47(3):424–429 Susai Mary J, Sai Balaji MA, Krishnakumari A, Nakandhrakumar RS, Dinakaran D (2019) Real time monitoring of drill runout using least square support vector machine classifier. Measurement – Journal of the International Measurement Confederation Elsevier 146:24–34 Umesh Gowda BM, Ravindra HV, Ullasa M, Naveen Prakash GV, Ugrasenc G (2014) Estimation of circularity, cylindricity and surface roughness in drilling Al-Si3N4 metal matrix composites using artificial neural network, 3rd international conference on materials processing and characterization (ICMPC 2014). Procedia Mater Sci 6:1780–1787 Ranjan J, Patra K, Szalay T, Mia M, Gupta M. K, Song Q, Krolczyk G, Chudy R, Pashnyov V. A, Pimenov D. Y (2020) Artificial intelligence-based hole quality prediction in micro-drilling using multiple sensors. Sensors (Basel) 20(3): 885 Xu L, Huang C, Li C (2020) Estimation of tool wear and optimization of cutting parameters based on novel ANFIS-PSO method toward intelligent machining. J Intell Manuf. https://doi.org/ 10.1007/s10845-020-01559-0

7

Logic Based Path Planning (LBPP) Algorithm for Robotic Library System Sagar Ajanalkar and Harshadeep Joshi

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Manipulator Arm and Gripper for Manipulation of Book . . . . . . . . . . . . . . . . . . . . . . . . . . Movable Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Robot Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Graphical User Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Book Navigation and Path Planning for Library System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Location Guidance Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Web Application Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Path Planning Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Routing Methods for Path Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives and Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Web Application for Path Planning in a Library System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Add Books . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Book List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Search Books . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Navigate Book Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Logic Based Path Planning (LBPP) Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Web Application Based Path Planning Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Development of Mobile Platform Prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion/Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

128 129 130 130 131 134 135 136 137 138 141 142 142 144 144 144 144 145 152 155 157 158 158

S. Ajanalkar () · H. Joshi Dr. Babasaheb Ambedkar Technological University, Lonere, India e-mail: [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_150

127

128

S. Ajanalkar and H. Joshi

Abstract

Searching a book in a library is one of the difficult and time-consuming tasks. This is a typical example of a warehouse where frequent storage and retrieval is required. With the developments in the field of Information Technology, the graphical representation of the storage of library can help the users to locate the book faster and provides a digital canvas with digital search options. This can further help in identifying and suggesting a faster route to the user especially in the situation when the user is trying for multiple books to be picked up during a check out. In the present work, the importance of graphical interface is discussed, which is developed using open-source languages such as HTML and Java Script. Also the system is discussed in detail in which, MySQL is used as database system for storing the details in the Library and in which the optimum path is identified using Logic Based Path Planning (LBPP) algorithm. The present aim of the work is to discuss the development of an easy tool for the library user for locating the books in the library. However, the extension of the work is intended toward encouragement of the development of a system that can be used for a robot to be used for library automation system. The work can also find applications at various shopping malls, warehouses of medical shops and E-commerce industries, etc. During the practical implementation of the proposed system, it is proposed to use Internet of Things (IoT).

Keywords

Library automation · ASRS · LBPP · Robotic storage retrieval

Introduction In the recent past, the serving robots are finding many applications in industrial, business, and domestic sectors. Library robots, robotic tour guide, robotic nurses, service robots in E-commerce industries, etc. are few of them. Also Integrated Shopping Robot (Tomizawa et al. 2007), Restaurant service robot (Yu et al. 2012), FURO Banquet Hall Setting Robot (F. Vega et al. 2020) are the few examples of service robots other than library service robots. In nearer future, robots can also be seen serving the old age people for their routine needs (Behan and O’Keeffe 2008). Library system is one such area where potential applications of robotic system can be found. Mainly the large variety of books makes it a complex database system and thus makes it difficult for the user to locate the required book easily. To overcome these drawbacks in library system, it is becoming necessary to automate the process of finding books from the library (Animireddy et al. 2018). More technological developments are necessary for getting the complete autonomous library system. Few of the developments are enlisted in this chapter (Fig. 1).

Localization and recognition of tags applied on books based on morphology, local contrast and optical character recognition (Chen et al. 2016) Use of Internet of Things for Library Management System (LMS) (Dong-Ying Li et al. 2016)

Self –navigating library robots using ultrasonic sensors (Rashid et al. 2017) Development of robot using myRIO for library management (Angal and Gade 2017) Microcontroller based library assistant robot for path guidance (Animireddy et al. 2018) Robot assistance for locating books in children’s library (Lin and Yueh 2018)

2019-20

Inventory management in library using autonomous RFID scanning robot (Li et al. 2015) Robot navigation using Infrared optical projectors (Kuriya et al. 2015) Library robot interaction using natural language (Fei et al. 2015) & hand postures (Nguyen et al.2015)

2017-18

2015-16

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

129

Autonomous library assistance robot using Raspberry Pi & advancements in natural speech recognition (Di Veroli et al. 2019) Development of LMS with android mobile usage (Juanatas and Juanatas 2019) Vision based book recognition and retrieval, autonomous navigation and obstacle avoidance (Yu et al. 2019) Development of LMS using Graphical User Interface (Sawant and Patil 2020) Use of modified algorithms for path planning and autonomous robot based inventory management (Mehta and Sahu 2020)

Fig. 1 Developments in the field of robotic library system

Introducing automation of library using library service robot is better solution. In general, library service robot systems may involve main components such as manipulator arm and gripper, movable platform, robot controller, and graphical user interface (Comsa et al. 2014) (Fig. 2). The main components of library service robot are discussed in detail as follows:

Manipulator Arm and Gripper for Manipulation of Book Manipulator arm of a library service robot is having construction similar to the human arm. It is used to support the gripper system and to move the gripper system up to the target book location, from which the book is to be retrieved or restored. Various type of mechanisms are used for designing manipulator arm and the desired motion is achieved. One end of the arm is connected to the main body of the service robot and another end is connected to the gripper system. The manipulator arms has to carry the payload of the book to be retrieved and the weight of gripper system along with self-weight. Also it is designed in such a way that it can also resist the vibrations due to inertia effect. Grippers are also called as end effectors. This system is positioned at the end location of the service robot, which directly comes in contact with the object to be picked up or placed. Grippers are of the types active and passive. Active grippers are used where closed loop control system is needed and passive grippers are used in case where open loop control system is needed. In case of active grippers, various sensors such as Infrared (IR) sensors, load cells, tactile sensors, and/or vision sensors are fitted on the gripper surface for getting the feedback for smooth operation of the whole system. In case of passive grippers, the sensors are not used.

130

S. Ajanalkar and H. Joshi

Fig. 2 Main components of Library Service Robot

Movable Platform It is the base platform on which the manipulator arm and gripper systems are mounted and whole support for arm and gripper system is provided by the base platform. As in case of service robots, the robots has to move from one location to another location. For allowing the motion to service robots, the base platform is enabled with a mechanism which allows the robot system to move. Generally the base platform is provided with the wheels connected to the electric motors and fitted on to the base platform for enabling its motion. The mobile platform is designed in such a way that it can move with maximum possible speed and also with the maximum payload carrying capacity.

Robot Controller Controller is the system which controls the operation of end effector, manipulator arm, and motion of mobile platform. The controller may be a mini computer or a

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

131

microcontroller such as arduino or rasberry pi or a special purpose electronic control unit. The controller unit consisting of number of input-output pins of analog and digital type, using which it is possible to connect various sensors and also it is possible to operate the output pins connected to output devices by processing the input data collected from the sensors attached. Also there are various electronic components such as analog to digital converters or digital to analog converters, operational amplifiers, microcontrollers, operational light emitting diodes, transistor-transistor logic (TTL) converters, etc. are fitted as per the requirement of operation and functionality. Few of the controllers facilitate the users to operate the output devices from remote locations with the help of Wi-Fi module using the trending technology Internet of Things.

Graphical User Interface User interface like proposed simulation web application, capable of high level interaction with the system can improve the performance of the whole system. Desirable characteristics of best user interface are simple to understand, user friendly, less response time, compatible to users, free from future maintenance and easy for modifications, etc. In case of complicated work of record keeping and in case where it is required to keep track of all the associated activities of service robot, it is always preferable to use an internet of things based graphical user interface capable of recording the data, processing the data collected, and presenting the processed data in easily understandable format. User interface can be developed using various available languages and application development software available, but it is preferred to use open source application development software. User interface can be developed as an application to be installed on the computer or smartphones, otherwise it can be developed as a web application which can be operated simply from any web browser available in the computer or smartphone without its installation. Out of these four main components of the system, user interface plays a vital role in understanding the system structure to the end user or customer and also plays a vital role in understanding the customer requirements by the service provider through detailed analysis of huge data sets collected from the search results and using various web application based techniques of literature survey. Few of the useful open source languages and web application development platforms available are discussed in detail as follows:

Hypertext Markup Language (HTML) HTML is an open source language contained in the web browser for presenting the information to the end user, also it is called as standard or core markup language in World Wide Web. This language is developed by Web Hypertext Application Technology Working Group (WHATWG) organization. HTML5 is the latest released version of the HTML language and can be embedded along with

132

S. Ajanalkar and H. Joshi

other languages such as CSS, PHP, SQL, JS, Python languages for betterment of functional operation of the user interface. Also it is widely used for programming the web applications with reduced complications in day today life for improving the effectiveness of the systems. HTML Canvas is an element in HTML5 which is used for presenting and drawing graphics. For showing graphical representation using HTML Canvas, JavaScript can be used along with HTML (WHATWG 2020).

Cascading Style Sheet (CSS) CSS is a Style Sheet Language contained in web browser for setting the style attributes of the font, layout, color of text and background, etc. CSS is developed by an organization entitled as World Wide Web Consortium (W3C) and initially released in the year 1996. Also CSS 2.1 is the latest revised version of the CSS. The CSS style attributes can be applied to any web application in three different possible and optional ways. First one is the internal positioning of the style attributes in the initial part of the programming of the web application using or using style {} element syntax. Second one is the external positioning of the style attributes in the separate file and calling the same in the main HTML file. Third one is the inline positioning of the style attributes in the program by mentioning the required set value of style attribute in the program lines separately as per requirement. Hypertext Preprocessor (PHP) PHP is a widely used and free scripting language used for creating time dependent dynamic web applications and mainly used for interacting the server database. It plays a vital role for establishing a strong connection in between html based user interface and server database. It is developed by Zend Technologies. Java Script (JS) JS is the programming language used in the html based web applications for performing the specific operations once or multiple times repeatedly. The specific operations such as mathematical calculation operations can be programmed using the JS syntax as a specific titled function, further the function is called in the html program using various loops for repeated number of operations to be performed. For performing the operations, it is possible to give performance parameter values in the function as a variable. JavaScript Object Notation (JSON) JSON is a format used for saving the data such as text, numbers, values or array in the form of string and it is also used for interchanging the data stored from servers. The format is also called as data interchange format. JSON is also used to store the data for the single session on to the browser for quick and temporary operation of the program. The JSON can be effectively used along with JS for performing any specific task of the requirement.

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

133

Structured Query Language (SQL) It is very important to create, operate and manage databases for storing huge data generated from the inputs collected from service provider and end users. For the same reason a query language is required to handle all the operations related to database such as creating the database, adding or removing rows and columns in the databases, storing or deleting or retrieving the data in the database, etc. SQL is one of the effective and efficient query language to use for the same reason. It is considered as standard language by ISO as well. This language is useful to operate the various database servers such as MySQL, Oracle, SQL Server, etc.

My Structured Query Language (MySQL) MySQL is an open source database management system using SQL language for interchanging the data stored in the database created in MySQL storage library. End users can access the data stored in the databases using the graphical user interface, which is using the SQL or other supporting languages at the back end for data retrieval. Other database management systems are also available other than MySQL such as SQLite, PostgreSQL, etc.

XAMPP (X-cross Platform, A-Apache, M-MySQL, P-PHP, P-Perl) XAMPP control panel is the open source and freely available software application compatible for Windows, Linux and Mac operating system. It is used for web application development using Apache http server, MariaDB (earlier MySQL) database management system, and PHP and Perl languages in a single packaged application. It is possible to check and modify the web application developed using local host server and also it will be possible then after successful completion of the web application project on the live server. XAMPP control panel is developed by Apache Friends Organization and it was initially released in 2002. In library system, storage and retrieval machine (SRM) is an automated robotic system which can be used as an automated vehicle for storage and retrieval of books in the storage rack. Library is one of the warehouse, where one can store the books instead of goods or products to be delivered. Almost all the systems in library are similar in context with the warehouses in e-commerce industries, logistic industries and medical drug storage warehouses, etc. The chapter is organized as follows: first introduction to the robot library system using web application development is discussed in detail, then discussed the topics on existing and proposed location guidance systems, web application development, path planning algorithm and routing methods for path planning. Further topics such as objectives, assumptions and use of web application for path planning are discussed, then the web application developed using LBPP algorithm and its functioning is discussed in detail.

134

S. Ajanalkar and H. Joshi

Book Navigation and Path Planning for Library System In case of traditional libraries, the printed materials such as reference books, text books, periodicals, magazines, journals and other reference materials are stored on the multi-tier shelves (Ozeer et al. 2019). The procedure of issuing and returning the book is to be carried out manually. Once the user come physically for issuing a book, the librarian or the user will search the book and arrive at the delivery counter, along with the book to be issued for keeping the record of the transaction. After arriving at the counter, barcode on the book is getting scanned by using barcode scanner and the details of the book are getting displayed on the computer user interface. The librarian enter the details either in excel sheet or on the library user borrower card for keeping a track of the book. The similar procedure of keeping record is getting followed when a user come for returning a book. During this manual process of keeping the record, the possibilities of errors may be more due to various factors such as lack of time available, rush of users, pressure of simultaneous work, etc. (Ozeer et al. 2019; Sawant and Patil 2020). Dong-Ying Li et al. and A. Ozeer et al. discussed some issues in the traditional libraries are enlisted and are as follows: (a) (b) (c) (d)

Inventory management of library Searching printed materials available in the library Finding the printed materials placed somewhere around from the actual location Storing the details of newly arrived printed materials and rearranging the already available details and places (if required) of the books (e) Security of the printed materials available, in case where user access is allowed To overcome these issues in the library management, it is advised by Dong-Ying Li et al. and A. Ozeer et al. to use the trending technology known as Internet of Things (Dong-Ying Li et al. 2016; Ozeer et al. 2019). Various users are having different requirements from the libraries. Sastry et al. enlisted the different user groups of library and their most prioritized requirement of library. Few of the user groups from them are students including undergraduate and post-graduate students, Research scholars, Research fellow members, teachers, and professors. Research scholars, Research fellow members, teachers, and professors are using library facilities such as scientific journals, reference books, technical papers, and general reading books, whereas the students use the study materials for notes preparation and general reading books most preferably (Sastry et al. 2011). Digital libraries are one of the effective resources of digital materials and a good alternative for traditional libraries. Da Rosa et al.; Abbas and Faiz compared the two types of library functionalities based on few performance parameters and concluded that the digital library is a good alternative for traditional library (da Rosa et al. 2012; Abbas and Faiz 2013). The results showed that the digital library accessed multiple times more, number of users increased and rate of accessing per item also increased, whereas few users out of total users of the traditional library visited during the

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

135

period of time taken for observation and only few of the printed materials were viewed out of total available materials (da Rosa et al. 2012). User Interface (UI) plays a vital role in increasing the usage of the digital library and for designing the effective UI, it is necessary to understand the challenges in designing the system. Sastry et al. enlisted different challenges according to various orientations such as library user orientation, search orientation, digital content orientation, and network connectivity orientation (Sastry et al. 2011). Although it is advantageous to use digital libraries rather than using traditional libraries in terms of many parameters, but there is always a requirement of the traditional library to store printed materials for the users who are located near to the library (Abbas and Faiz 2013). Use of the concept of digitization and automation in the traditional or conventional libraries is the optimum solution. The topics enlisted below such as location guidance systems, web application development, path planning algorithm and routing methods for path planning will give a brief information about book navigation and path planning for a library system.

Location Guidance Systems Library is having wide range of books placed on different locations. Finding a book in a library is really a tricky problem and it becomes a tedious, time consuming job for a new user to find a book if it is misplaced by someone. In such case, for completing the task easily it is necessary to take the help of the librarian or picker and movers in case of conventional library system. Mikawa et al. proposed a new form of book guidance system which can make use of laser pointer and voice assistance system, by which a library user can quickly understand where a search book results in a library. Since the librarian robot uses the laser pointer and voice assistance system attached to point a target or direction to the target and assist user through voice commands, a library user can easily understand the guidance. The system developed is capable of understanding the language of the user in natural way, also it start greeting the user for doing welcome of the user. This system identifies the book location after getting request of the book location search from the library user and after identification of the book location from available database and geometrical map of library area, it points out the shelf in which the book is located. During this pointing out, it helps users to identify the location, so authors have also worked on gestures of robot (Mikawa et al. 2010). To assist the user who wants to find and collect the book in the library, a prototype robot was created to help by Animireddy et al. In the computer-based system, the user’s request is processed to make the robot notified about the specific shelf or position. The processed data inform the robot to move along the path which is predefined and during this movement, robot will scan the books available on the shelf using barcode scanner to identify the required book. The user is notified after the requested book is identified (Animireddy et al. 2018). Another service robot specially developed for assisting

136

S. Ajanalkar and H. Joshi

library users named “LibRob,” provides users with instructions to locate the book they are looking at for an interactive way to have a more satisfactory experience. LibRob can process a search request either by speech or by text and return an author, topic or title list of related books. LibRob directs them to the shelf containing the book after the user selects the book of interest, and then returns to its base station upon completion. Experimental studies show that the robot reduces the time taken to locate a book by 47.4% and pending 80% of users are pleased (Di Veroli et al. 2019).

Web Application Development Currently web application development becoming more popular due to its various advantages such as easy data retrieval, improved effectiveness of system, simplification of checking the product availability, etc. Mandava and Antony reviewed and analyzed three technologies getting used for the development of web application such as ASP.NET, Open source technologies such as Linux/MySQL/Apache/PHP and Java 2 Platform, Enterprise Edition (J2EE). After comparison, it is concluded by researcher that the developer may go for Linux/MySQL/Apache/PHP technology if less amount is available for project, if project is small, if web application security is needed, if compatible mode of operation required, and if less time is available for project completion with moderate performance rate (Mandava and Antony 2012). Navaraja et al. developed a web application based simulation tool for virtual learning of the kinematics of robotics and real time online control of robots such as Puma 560 and Movemaster RM-501. Web application simulation is developed by MATLAB and J2EE technology. Also for client end web application development, Javascript, Ajax and PHP integrated with HTML languages are used along with Python and VAL for server end programming and command transfer for robot functioning respectively (Navaraja et al. 2016). Development of web application is possible using various languages such as HTML5, Flash, C/C++, etc. The comparison among these three language types is discussed by Wang et al. The parameters used for comparison were openness, visualization effect, production cycle, and enablement on mobile, screen touch enablement, compatibility, necessity of plug-in, and other. In most of the cases, the web application developed using HTML5 gives good performance results In terms of simplicity of programming, least cost due to open source language, convenience of updating, accessibility through web browser from devices such as mobile phones, tablet, and computer (Wang et al. 2019). Hongjiu highlighted the benefits of using a HTML based web application developed for an enterprise by considering the data mining technology. In case of E-commerce industries, the huge data getting collected from user end can give fruitful results for better understanding the need of customers in advance and also used for taking beneficial decisions for the enterprise (Hongjiu 2013). Harvey et al. introduced a web application based assistant system for helping the students virtually 24/7, which is developed using HTML, XAMPP, Program-O, and other such as AIML (Artificial Intelligence Markup Language). In the system developed,

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

137

the avatar animation of female assistant will guide students in the university for various important aspects such as placement related guidance using text to speech conversion (Harvey et al. 2016). Such kind of virtual avatar based support systems will plays a vital role in upcoming days for guiding the users of any system.

Path Planning Algorithm Path planning is very important for optimization of various performance parameters such as time required for completing the task, distance travelled by robot, energy consumption by robot, etc. Also it is useful during obstacle avoidance in case of multi-robot workspace. According to Raja and Pugazhenthi, the path planning is divided into two major types depending on the environment information available before starting the robot motion or during the robot motion: (1) On-line path planning, (2) Offline path planning. The algorithm plays a vital role in applications such as surveillance, identification of landmines, atomic power plants, and in planet investigation (Raja and Pugazhenthi 2012). For developing the LBPP algorithm in the web application developed, the offline type of path planning is used and also the obstacle free environment is considered. Sun et al. proposed algorithm based simulation for path planning of Unmanned Aircraft Vehicle (UAV) flight. The algorithm developed using knowledge of basic mathematics and basic logic of path planning for optimizing the travelling distance and time required to complete the flight (Sun et al. 2013). The similar approach is used for the developed LBPP algorithm in the web application developed for the library system. Naumov et al. proposed an improved Dijkstra algorithm, which uses an image captured of workspace from top. The image was then converted to a twodimensional matrix form for analyzing the obstacles present in the labyrinth and to find the optimum path for robot motion planning. Further simulation is created in a computer for better visualization of the path planning. After simulation, the complete time required for processing the 1600 × 1013 pixels sized labyrinth and for presenting the simulation observed was less than 500 milliseconds (Naumov et al. 2015). M. Korkmaz and A. Durdu used Simultaneous Localization and Mapping (SLAM) algorithm, GMapping for building real time workspace map, and a statistical technique one way Analysis of Variance (ANOVA) for comparing the performance of various path planning algorithms. For comparison, five path planning algorithms are considered such as A* search algorithm, Genetic Algorithm (GA), Rapidly exploring Random Tree (RRT) algorithm, Bidirectional RRT, and Probabilistic Road Map (PRM) algorithm. After comparison it is concluded by M. Korkmaz and A. Durdu that for finding shortest path one can go for A* search algorithm, but it requires large computational time and for the requirement of better results in terms of time optimization, one can go for the PRM algorithm (Korkmaz and Durdu 2018). Han-ye, Wei-ming, et al. reviewed the literature on path planning of robots. From reviewed literature, it was concluded in the paper that the Genetic Algorithm (GA) based approach is most probably getting used for path planning and optimization. Also other approaches such as Particle Swarm, Ant

138

S. Ajanalkar and H. Joshi

Colony Optimization Algorithms, and Artificial Potential Field algorithm approach (PSO, ACO and APF respectively) are used widely after GA approach (Han-ye et al. 2018). Further modification in the ACO using a concept of Killer Nodes is proposed by Guarnizo and Pineda along with the simulation in unity application for better visualization and for reducing negative feedback in case of long path planning problems (Guarnizo and Pineda 2019).

Routing Methods for Path Planning Various optimization techniques for routing are available, out of which any one best suited technique can be applied for finding the optimum path. Routing in a library is similar to the Travelling Salesman Problem (TSP), in which the optimum path can be identified using numerical treatment. In TSP, the optimum path means a finding a path with minimum travelling distance, with less time consumption or with less cost involved (Pankratova and Tarashnina 2017; Zia et al. 2018). TSP is applicable where shortest route is to be identified, where each destination location is to be visited only once and in the case where the salesman (or salesperson) is to be returned back to the original location from where salesman started travelling. Generally, the locations are considered as cities and cost between locations is considered as the distance (or time) between cities (Hahsler and Hornik 2007). TSP is categorized into three different types according to various conditions and complexities of the path planning problems. The types are: (a) Symmetric TSP (sTSP), (b) Asymmetric TSP (aTSP), and (c) Multi TSP (mTSP). The path planning problem in which the optimum travelling path is to be identified between ‘n’ cities to be visited only once during a whole trip, if the to and fro Euclidian distance (i.e., cost) between all two cities is equal then the TSP is known as Symmetric TSP. On the contrary, for the path planning problem in which the optimum travelling path is to be identified between many number of cities to be visited only once during a whole trip, if the to and fro Euclidian distance (i.e., cost) between any two cities is not equal then the TSP is known as Asymmetric TSP. Further, if number of salespersons are more than one and are available at initial city location, from which the travelling needs to be started. Also in case where all the salespersons are to be utilized effectively and each city is to be visited only once during the whole trip of salesperson, it is recommended to use the third type of TSP known as Multiple TSP. The Multiple TSP may be further utilized for solving the TSP-based path planning problems in which variable conditions are applied. Varying or fixed count of number of salespersons, number of cities, number of salespersons available at different cities or at initial city, number of salespersons finishing their trip at specified city location, minimum or maximum number of cities to which each salesperson must visit, time limits for visiting the specified cities, limiting values of distances to be travelled by each salesperson, etc. are the few conditions to be applied as per the requirement for the problems of mTSP. TSP has its wide range of applications in day-today life as

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

139

well as in industries such as service industries and manufacturing industries. Few of the applications enlisted are as follows: (a) Path planning in automated drilling operation (b) Operation sequence finalization in case of x-ray crystallography and gas turbine engine overhauling (c) Path planning of pickers and movers in warehouses (d) Problems where vehicle routing is required such as school bus (e) Scheduling problems (f) Navigation system design (Matai et al. 2010) (g) Courier delivery services (h) Receiving and broadcasting television or radio signals, etc. (Hira and Gupta 2010) If there are ‘n’ number of cities to which a salesman (or salesmen) has to visit during a complete trip only once and return back to home city, then for deciding the optimal path in terms of less time requirement, less cost involved, and in terms of reduced distance to be travelled; it is necessary to find all the possible number of routes. The possible number of routes for travelling salesman starting from home city to ‘nth’ city and returning back to home city can be given mathematically (Eq. 1). N = (n − 1)!

(1)

Where, N = Maximum number of possible routes for travelling ‘n’ cities. Also mathematically the TSP can be represented as a linear programming problem (Eq. 2). Z=

n n  

cij xij

i=1 j =1

Subject to n  j =1 n 

xij = 1 xij = 1

i=1

And xij = 0 or 1; i = 1, 2, . . . , n; j = 1, 2, . . . , n

(2)

140

S. Ajanalkar and H. Joshi

where, cij = Cost or distance or time involved for travelling a salesman from city ‘i’ to city ‘j’ xij = Constant value either zero (0) or one (1) xij = 1 in case salesman directly travels from city ‘i’ to city ‘j’ xij = 0 in case salesman does not travels directly from city ‘i’ to city ‘j’ (Hira and Gupta 2010). For improving the routing of robots or Automated Guided Vehicles (AGV) in warehouses similar to library systems, many researchers proposed various routing methods to improve the efficiency of functioning in terms of reduction in material handing travel time. Roodbergen and Koster developed a new approach for solving the issues in the warehouses where order pickers and movers are continuously changing the aisles. In this chapter, the various approaches are collectively summarized from literature review and an optimized approach is developed and proposed. The various approaches discussed in the paper are as follows: (a) S-Shape approach In this approach, the SRM will start to move from left aisle and initially it will pick-up or place the object from the nearest shelf placed in left side. Further SRM will find the location of nearly placed next object, if the location is in the same aisle then it will move through the same aisle else it will move to the next aisle in S-shape path. Once all the pick-up locations are over in the first cross-aisle, then SRM is instructed to shift the cross-aisle and will again start the same cycle from left aisle in S-shaped path. The same procedure is allowed to be followed up to the second last cross-aisle, then for the last cross-aisle SRM is instructed to follow the cycle in reverse direction, i.e., from right side aisle to the left most aisle till the initial depot location is not arrived. (b) Largest Gap Approach In this approach, the gap is nothing but the distance between two different object locations from same aisles or it is the distance between cross-aisle and the object location. Largest gap criteria is used for segregating number of object locations to be picked in sequence, one is object locations nearest to the back cross-aisles and next is nearest to the front cross-aisles. Initially the SRM is instructed to start the pickup or place operation from the nearest location in the left most aisle, then it will continue picking or placing objects in the same aisles by changing sub-aisles and then it will change the aisles through cross-aisles. After that SRM is instructed to pick-up or place the objects nearest to the back cross-aisles and then instructed to pick-up or place the objects nearest to the front cross-aisles. (c) Aisle-by-Aisle Approach Title of this approach gives a clear idea about the process of pick-up or placing an object or number of objects. In this approach, the SRM is instructed to complete the picking or placing task of multiple objects in a first aisle, then after finishing all the pick-up locations it is instructed to SRM to go for next

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

141

aisle till the last aisle. At last after completing all the aisles, it is instructed SRM to move at the initial depot location. (d) Optimal Approach In all the above approaches, it is planned for routing the SRM or also known as picker or mover during completion of the pick-up or place task. The optimal approach proposed in an earlier literature was modified and proposed by Roodbergen and Koster for finding shortest routes using branch and bound procedure proposed in a literature. Also it is concluded by Roodbergen and Koster from the results of algorithm used, that the material handling rate can be optimized by incorporating optimum number of cross aisles in between shelves. The proposed routing methods can be used for improving the efficiency of manual material handling (Roodbergen and Koster 2001). Further, Roodbergen et al. established a design based approach for improving the layout of warehouse and for applying control strategies in the warehouse. Also the design-based approach considering optimum routing method successfully implemented in the company as a case study, where the performance results improved in terms of reduction in travel distance of manual pickers per unit order handling (Roodbergen et al. 2015).

Summary Looking at the current scenario, complete digitization of the library is not the perfect solution, as traditional libraries have their own importance due to various conditions such as environmental conditions for study, conditional suitability of printed materials, etc. Conversion of conventional library system into smart library system is the perfect solution among all the related categories of library systems. For converting the conventional library into a smart library, A. Ozeer et al. proposed a technique of developing a web-based application for locating the printed materials available in the library (Ozeer et al. 2019). Web application based simulation is a best technique among other techniques to guide the location of book or product stored in a library or warehouse. Also it is possible to keep a track of location of book, availability or non-availability of book and to optimize the path planning for functioning in the library. Using manual operation of keeping the record of all the books getting placed in a dynamic library system is very difficult, where locations of books is continuously changing as per the categorization of books according to its size, shape, type of book (e.g., hardcover or softcover) and retrieval frequency, etc. Also the useful record or information has to be accessible to all its stakeholders from different locations for increasing the utilization of the system. This is possible by using the concept of Internet of Things. So it is highly advisable to use the technology of Internet of Things in the library system which is similar to warehousing system in E-commerce industries. In all the above library assistance systems and other location guidance systems discussed, a huge amount of investment is required for guiding the user to identify

142

S. Ajanalkar and H. Joshi

the books located in the library system due to building the robot and using paid software or languages. To build an open source software, which assists a user by showing simulation is a cost effective solution.

Objectives and Assumptions The following objectives are considered for developing the LBPP algorithm and web application: (a) To search the book with various filters such as book id, title, author, publisher, or location of book. (b) To identify the location of book stored with ease. (c) To simulate the optimum path using web application. The following assumptions are made for the case study: (a) Only five books are placed in each storage rack for simplicity, further one can modify using html code. (b) Sample library layout containing total (4 × 6) racks is considered. (c) Lines of different color are used to indicate the path travelled and numbers indicate the end of path or book location for picking the respective book. (d) The home location for SRM is considered at location ‘0’. (e) Black color filled rectangles are the books and while color filled rectangles are storage racks.

Web Application for Path Planning in a Library System For guiding the users of library various location guidance methods are proposed by researchers, those are discussed in earlier topics. Mikawa et al. proposed a laser pointer and voice assistance-based guidance method for location guidance system, in which it is stated that the feasibility of the proposed system is to be checked with two major practical aspects in real life library system. First point was about to maintain silence in the library system, the point arrived due to the voice assistancebased guidance system developed and second is the decrement in library users due to easy availability of Internet and e-books (Mikawa et al. 2010). Both these points can be answered by using the proposed web-based application for using the library facility from remote location as well. Also the further modified version of this application along with scanning mobile robot platform can help the library users to access the printed form of materials from remote location using Internet of Things, which is similar to the technology getting used in Industry 4.0. The simulation-based web application technology is simple way to guide the users of library for identifying the books. Further it is proposed as future scope

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

143

to design and develop the mobile platform, which follow the simulation in web application and guide the users to navigate book required in the library system. For achieving this target the following methodology can be adopted. (a) Select an open source web application framework development tool such as XAMPP control panel having freeware servers like Apache and MySQL. (b) Develop an interactive web application for adding booklist with its corresponding locations using open source languages such as HTML, CSS, PHP, Java Script, JSON, and SQL. (c) Develop a simulation page using HTML5 Canvas and develop an efficient Logic Based Path Planning (LBPP) algorithm for optimizing the path planning problem along with better visualization of simulation. Web application can be designed and developed as shown in Fig. 3 using open source languages such as HTML, CSS, Java Script, JSON, MySQL, and PHP. For running it on local host server, open source local host servers such as Apache and MySQL can be used which are available in open source application known as XAMPP control panel. Also the simulation part in user interface can be developed using html canvas for representing animation for path planning in library system. The following facilities can also be provided in the ribbon at the left side of the web application in the form of tabs as shown in Fig. 3.

Fig. 3 Layout of web application for path planning simulation

144

S. Ajanalkar and H. Joshi

Add Books “ADD BOOKS” tab on left ribbon can be provided to add book details such as: Title of Book, Name of Author, Publisher, and Location of Book. One can add book details in the html form provided to add details in SQL database table after clicking on “Add Book” button provided.

Book List “BOOK LIST” tab on left ribbon can be provided to check the list of added books in database of books added such as: Book ID, Title of Book, Name of Author, Publisher and Location of Book. After clicking on the tab, the PHP code can interact with the table stored in the SQL database to check the details available in the form of html table.

Search Books “SEARCH BOOKS” tab in the left ribbon can be provided to search book details from database using filters. Also one can provide an additional feature so as to get a list of books after clicking on search button with search keywords and appropriate filter. After getting a list of number of books, the facility for selecting multiple books (up to maximum 4) using checkbox can be provided. After selection using checkbox and after clicking on “Get Selected,” the location of books can be stored in local storage. For further simulation, it is required to convert the location details in the form of string and store the details of location in LOCAL STORAGE using JSON program.

Navigate Book Location “NAVIGATE BOOK LOCATION” tab can be provided in left ribbon to visualize the simulation of path planning for retrieving the number of books in the library system. For which, it is required to convert the string details in the form of an array from local storage and give as an input to Logic Based Path Planning (LBPP) algorithm which can be developed using Java Script, HTML, and CSS program. Further, the algorithm will estimate the path for retrieving number of books as shown in the upcoming sections (Figs. 9a, 10a, 11a, and 12a).

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

145

Logic Based Path Planning (LBPP) Algorithm The complete algorithm can be developed in basic two stages: Stage (I): Initial Setup The flow chart (Fig. 4), is representing the stepwise functioning of the program to define and create a baseline for LBPP algorithm. In this algorithm, initially the X-coordinate and Y-coordinate of the black rectangle representing book in the html canvas is defined for all the books placed in vertical rectangular shelves (Fig. 5). The location coordinates of each book and the location ID is stored as separate array. After selecting the books required from SEARCH tab in web application, the location ID of selected books getting stored in the local storage in the form of string. The string stored in local storage is converted into an array after opening the “NAVIGATE BOOK LOCATION” tab and the array compared with initially stored array of all the book location ID and coordinates. Finally, the HTML canvas is defined of size 615 × 450 pixels and stored as a variable. Further, this defined canvas can be helpful for representing the simulation area

Fig. 4 Flow chart for Stage (I): Initial Setup

146

S. Ajanalkar and H. Joshi

Fig. 5 Coordinates in a HTML canvas for understanding of LBPP algorithm

along with initially represented shelves and books in it. After comparison of arrays and after opening the “NAVIGATE BOOK LOCATION” tab, the color of specific selected books changed from black to blue for better understanding of the book location in simulation. Stage (II): LBPP Algorithm The logical sequence of instructions is structured and developed in this stage for representing path planning simulation using html canvas. The canvas of 615 × 450 pixels is considered for better visualization of simulation (Fig. 5). Again in stage (II) following three sub-stages can be formed for simplicity of programming and for developing a simulation using html canvas: A – Initial Move, B – Intermediate Moves, and C – Retrieval Move A – Initial Move: The flow chart (Fig. 6),is representing the functioning of algorithm developed for showing the movement of SRM in simulation from location ‘0’ to location ‘1’, i.e., from home position to the location of first book placed on storage rack. The initial move is generated in simulation using the following steps: (a) Define style attributes for text and SRM at (0, 210), i.e., at location 0

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

147

Fig. 6 Flow chart for Stage (II): LBPP Algorithm [A – Initial Move]

In HTML Canvas, it is necessary to define the color and other style attributes such as font, font size of text written in the SRM as 0 (Fig. 9a), outline color and outline thickness, etc. Red color is used for representing SRM at location (0, 210) (Fig. 5). (b) Draw rectangle at location 0 to show initial position of SRM. The shape of the SRM is considered as square for drawing the structure of SRM. Also the dimensions of rectangle are taken as(30 × 30). (c) Define style attributes for line going at (x1 ,y1 ), i.e., at book location 1 Line color and line thickness are the style attributes defined for drawing a line from (0, 210) to (x1 ,y1 ). The colour defined here is pink same as

148

S. Ajanalkar and H. Joshi

(d)

(e)

(f)

(g)

that of the colour used to define style attributes of SRM at (x1 ,y1 ), i.e., at location 1, which is helpful for identification of different paths during overlapping of path over one another. Draw a horizontal line till (x1 , 210) After defining style attributes, it is important to use line command to draw the line used to represent the path of SRM. According to the initial location, i.e., (0, 210), always there will be forward movement of the SRM first. So the commands used to draw horizontal line first up to point (x1 , 210) from initial location (0, 210). Draw a vertical line till (x1 ,y1 ) After the forward movement of the SRM, it will take a turn either in the upward direction or in the downward direction depending on the location of book. For showing this movement of the SRM, it is necessary to draw the vertical line either moving upward or in downward direction up to point (x1 ,y1 ) from point (x1 , 210). Define style attributes same as line for text and SRM at (x1 ,y1 ), i.e., at location 1 For showing the location of SRM at first book location, it is necessary in HTML to define the style attributes of the SRM at location 1. The pink color is defined for representing the SRM same as that of the line drawn from location (0, 210) to (x1 ,y1 ). Draw rectangle at location 1 to show final position of SRM. After defining style attributes, it is necessary to use the draw rectangle command in HTML to draw the rectangle of same size (30 × 30) using the style attributes defined in the earlier step at location 1, i.e., at (x1 ,y1 ).

In parallel, after defining style attributes for line and rectangle the following steps followed for calculating and storing the distance travelled by SRM during path planning for initial move. (a) Define a blank array ‘L’ For storing the data in HTML, it is necessary to use the array command. The stored values in an array can be further used for plotting the graph and for calculating the total distance travelled by the SRM. (b) Store length of line as ‘l1 ’ Using mathematical formulation, length of horizontal line is calculated from the coordinates of SRM end locations and this length is stored in newly defined variable ‘l1 ’. (c) Store length of line as ‘l2 ’ Using mathematical formulation, length of vertical line drawn after horizontal line is calculated from the coordinates of SRM end locations and this length is newly defined variable ‘l2 ’ (d) Calculate distance travelled by SRM using formula, L1 = l1 + l2 and Push value of L1 in an array ‘L’.

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

149

Total distance travelled from the initial location of SRM to the first book location is calculated and the value is stored in the defined blank array ‘L’ using PUSH command. B – Intermediate Moves: The flow chart (Fig. 7),is representing the functioning of program developed for showing the movement of SRM in simulation from location ‘1’ to location ‘2’, from location ‘2’ to location ‘3’ and/or from location ‘3’ to location ‘4,’ i.e., for intermediate moves as per number of books to be retrieved. The intermediate move/moves is/are generated in simulation using the following steps: (a) Define style attributes for line going at (xn ,yn ) where, n = 2, 3, 4 and (xn ,yn ) are the x and y-coordinates of the nth book location in canvas. (b) Depending on number of books selected for retrieval and the location of book, the path of SRM is decided. For showing this path in simulation after initial move, i.e., for intermediate move, the style attributes are initially defined. (c) If (x(n − 1) = = xn ) is true, draw a vertical line till (xn ,yn In case the next book selected is placed in a shelf located in the same aisle, the SRM has to travel along the same aisle. So the x-coordinate will not change for the next book retrieval. In that case for showing the path to be travelled for next book, the vertical line is to be drawn from earlier location to (xn ,yn ). (d) If (x(n − 1) = = xn ) is false, check for next condition. If [(10 < y2 < 210 & & (y1 − 10 + y2 − 10) ≤ (210 − y1 + 210 − y2 )] is true, then first draw a vertical line till (x(n − 1) , 10), next to that draw a horizontal line till (xn , 10) and at last draw a vertical line till (xn ,yn ) (Fig. 5). (e) If [(10 < y2 < 210 & & (y1 − 10 + y2 − 10) ≤ (210 − y1 + 210 − y2 )] is false, then go for next condition. (f) If [(210 < y2 < 410 & & (y1 − 210 + y2 − 210) ≥ (410 − y1 + 410 − y2 )] is true, then first draw a vertical line till (x(n − 1) , 410), next to that draw a horizontal line till (xn , 410) and at last draw a vertical line till (xn ,yn ). (g) If [(210 < y2 < 410 & & (y1 − 210 + y2 − 210) ≥ (410 − y1 + 410 − y2 )] is false, then go for the last condition. (h) If (Number of books retrieved ≥ n) is true, then first draw a vertical line till (x(n − 1) , 210), next to that draw a horizontal line till (xn , 210) and at last draw a horizontal line till (xn ,yn ). (i) Define style attributes same as line for text and SRM at (xn ,yn ), i.e., at location ‘n’. (j) Draw rectangle at location n to show final position of SRM. C – Retrieval Move: The flow chart (Fig. 8), is representing the functioning of program developed for showing the movement of SRM in simulation from location ‘n’ to location ‘0’, i.e., for retrieval movement of SRM from location of last book stored in storage rack to home position of SRM.

150

S. Ajanalkar and H. Joshi

Fig. 7 Flow chart for Stage (II): LBPP Algorithm [B – Intermediate Moves]

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

Fig. 8 Flow chart for Stage (II): LBPP Algorithm [C – Retrieval Move]

151

152

S. Ajanalkar and H. Joshi

The final move is generated in simulation using the following steps: (a) If (Number of books retrieved = = 1) is true then u = x1 and v = y1 . Using this web application, user can select up to four number of books. In case of only one book selection, the variables defined (u, v) will store values (x1 ,y1 ), i.e., coordinates of location 1. (b) If (Number of books retrieved = = 1) is false then go for next condition. (c) If (Number of books retrieved = = 2) is true then u = x2 and v = y2 . In case of two number of books selection, the variables defined (u, v) will store values (x2 ,y2 ), i.e., coordinates of location 2. (d) If (Number of books retrieved = = 2) is false then go for next condition. (e) If (Number of books retrieved = = 3) is true then u = x3 and v = y3 . In case of three number of books selection, the variables defined (u, v) will store values (x3 ,y3 ), i.e., coordinates of location 3. (f) If (Number of books retrieved = = 3) is false then go for next condition. (g) If (Number of books retrieved = = 4) is true then u = x4 and v = y4 . In case of four number of books selection, the variables defined (u, v) will store values (x4 ,y4 ), i.e., coordinates of location 4. (h) Draw a vertical line from (u + 30, v) till (u + 30, 240). For showing the return path of SRM to location 0, the vertical line is drawn starting from end location of SRM to (u + 30, 240)In this case, to avoid the overlapping of vertical lines in few cases and for simplicity of understanding the simulation, the offset distance of 30 pixels is taken in addition to ‘u’. (i) Draw a horizontal line till (0, 240). At last, the command used to draw a horizontal line from (u + 30, 240) till (0, 240) (Fig. 8).

Web Application Based Path Planning Simulation The web application based path planning simulation for single to multiple book retrieval processes are shown (Figs. 9a, 10a, 11a, and 12a) which are generated using LBPP algorithm and html canvas. The graph of relative distance travelled versus SRM movement to the specified location is shown (Fig. 9a and b) which is generated after simulation using html chart function. For the same on X-axis, the location of book to be retrieved is taken, e.g., A102 (where as A1 represents as name of storage rack and 02 represents the book number from top in the same rack) and final retrieval movement as Retrieve. For understanding of the movement of SRM, the following color codes are used: (a) Pink Color – Movement of SRM from Location ‘0’ to Location ‘1’, (b) Red Color – Movement of SRM from Location ‘1’ to Location ‘2’, (c) Green Color – Movement of SRM from Location ‘2’ to Location ‘3’

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

153

Fig. 9 Experimental prototype model

a

b Relative distance Travelled in px 266.0 1

265.8 265.6 265.4 265.2

0

265.0 264.8 264.6 264.4 264.2 264.0 A102

Retrieve

Fig. 10 (a) Simulation results for single book retrieval. (b) Graph of relative distance (in pixels) travelled by SRM

(d) Brown Color – Movement of SRM from Location ‘3’ to Location ‘4’, (e) Blue Color – Movement of SRM from Location ‘n’ to Location ‘0,’ i.e., Retrieval Motion. There are several routing methods available in the literature. Among those methods, selection of the proper routing method will decide the effectiveness of the pick-up and retrieval rate in library system where racks are placed in parallel orientation. In this chapter, the algorithm developed for one of the combinational and logical routing method from available literature is discussed for generating simulation of path planning. The results for S-shaped routing method are calculated

154

S. Ajanalkar and H. Joshi

a

b Relative distance Travelled in px 500

450 2 400 0 350 1

300

250 B302

E203

Retrieve

Fig. 11 (a) Simulation results for two books retrieval sequentially. (b) Graph of relative distance (in pixels) travelled by SRM

a

b Relative distance Travelled in px 650 1

600 2 550 500 450

0 400 3 350 300 250 200 A103

C105

F301

Retrieve

Fig. 12 (a) Simulation results for three books retrieval sequentially. (b) Graph of relative distance (in pixels) travelled by SRM

for one of the assumed problem for four cases by using numerical calculation method, whereas the results for routing method used in LBPP algorithm are simulated using web application developed. Further, the results obtained from routing methods used for developing LBPP algorithm and results obtained from S-shaped routing method are compared (Fig. 13). For routing method used in LBPP algorithm, the distance travelled by storage and retrieval machine (SRM) is 530 pixels, 1070 pixels, 1575 pixels, and 1940 pixels for one, two, three, and four books retrieval process, respectively, and for S-shaped routing method, the distance travelled by storage and retrieval machine (SRM) is 530 pixels, 1370 pixels, 1575 pixels, and 2030 pixels for one, two, three, and four books retrieval process, respectively. From this graphical comparison (Fig. 13) of both the routing methods, it is observed that the percentage optimization of the proposed routing method in this chapter for the case considered is 7.09%. The percentage optimization in this case is calculated as follows:

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

a

155

b Relative distance Travelled in px 650 1 600 2 550 500 450

0 4

2

400 350 300

3

250 A102

B302

D405

F301

Retrieve

Fig. 13 (a) Simulation results for four books retrieval sequentially. (b) Graph of relative distance (in pixels) travelled by SRM



(X − Y ) × 100 Percentage Optimization = X



Where, X = Total Distance travelled by SRM using S-shaped routing method Y = Total Distance travelled by SRM using routing method used for LBPP algorithm

Development of Mobile Platform Prototype The task of identifying a book located in a huge library is really tedious for the users of a library and sometimes it’s been a complicated task for librarian to find the book placed on a changed location. For making it easier and less time consuming, the simulation based web application can be further embedded along with the mobile platform (similarly as shown in Fig. 14) to perform the task of locating the books in case of library system. The similar kind of system can be implemented in other industries such as e-commerce industries or in warehouses for identifying the location of product placed somewhere in the workspace easily. Development of such experimental mobile platform is planned for performing the motion task similar to the simulation in the web application. For experimental testing of path planning as per the simulation in web application, the similar model can be developed as shown in prototype model of mobile platform (Fig. 14). The Internet of Things (IoT)-based path planning prototype model is mainly consisting of the following components: (a) NodeMCU ESP 8266 kit Internet of things (IOT) is the trending and effective technology used to implement the projects using internet or wifi network connection. NodeMCU

156

S. Ajanalkar and H. Joshi

Fig. 14 Comparison of routing methods using simulation results

is a microcontroller kit which gives sufficient facilities for producing an open source prototype of IOT based solution. NodeMCU is a low cost development kit and a firmware used for connecting sensors and other hardwares using Wi-Fi enabled microcontroller named as ESP8266. This Wi-Fi System on Chip (SOC) on the NodeMCU development kit is an open source firmware developed by Espressif Systems. This kit is a good cost effective solution to operate the hardwares connected to it using General Purpose Input/Output (GPIO) pins provided on NodeMCU. It consists of 128 kiloBytes Random Access Memory (RAM) and 4 MegaBytes storage memory for storing the program. Also it is enabled with 2.4 GHz Antenna for better connectivity with other Wi-Fi enabled devices. Multiple pins are provided as input or output pins, which are used to interact with the hardwares connected. The Arduino Integrated Development Environment (IDE) can be used for programming the NodeMCU ESP8266, whereas Lua script language is used for internal interaction of firmware. (b) L298N Motor driver circuit module To control the DC motors and stepper motors, it is necessary to use the effective motor driver module. L298N motor driver module enables to control the direction of rotation of motors and to control the speed of motors using pins provided on the module. The module uses 12 volt supply for its operation. It is provided with total four output pins for giving input to the motors named as OUT1 , OUT2 , OUT3 , and OUT4 . OUT1 and OUT2 pins are used for operating left side motors, whereas OUT3 and OUT4 pins are used for operating right side motors. Also motor driver module is provided with four input pins named as IN1 , IN2 , IN3 , and IN4 for controlling the direction of rotation of motors. IN1 and IN2 pins are used to control direction of rotation of left side motors, whereas IN3 and IN4 pins

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

157

are used to control direction of rotation of right side motors. Alongwith these input and output pins, additional Pulse Width Modulation (PWM) pins named as ENA and ENB are provided for controlling the speed of left side motors and right side motors respectively. (c) Battery unit as power supply Power supply is needed for operating NodeMCU ESP8266 kit, L298N Motor driver module and dc geared motors. Selection of proper battery supply is of prime importance for getting the required performance of any module. It is planned to use the 9 volt battery for operating the experimental mobile platform. (d) Other accessories such as dc geared motors, wheels, electric cables, base plate, switch, sensors etc. Four DC geared motors are used and are connected to output pins of L298N driver module. The wheels are fitted on the shaft of four dc motors. For switching on or off the power supply ON-OFF switch is used and the malefemale jumping wires are used whenever necessary for making connections. Further, it is planned to use various sensors for tracking the location and/or motion of the mobile platform. The modifications of the web application and embedding the prototype model, i.e., experimental mobile platform with the web application is in process.

Conclusion/Summary From the simulation results, it can be concluded that the simulation will help the users of library system to navigate the book location with ease. This web application-based system enables users to estimate the path along with distance to be travelled for retrieval of multiple books. LBPP algorithm developed will estimate and simulate the path accurately as it was developed by using trial and error approach. Further, the algorithm can be easily modified for implementing it in different applications. The percentage optimization results by 7.09% will give a clear conclusion that the routing method used for developing LBPP algorithm gives improved results for reducing the travelling time and cost by reducing the travelling distance for retrieval process of books in library system. As the web application is designed and developed using open source languages, it is cost effective solution for managing and organizing the books in the library. The systems can be further enhanced or modified to use it in the e-commerce industries, automated warehouses, shopping malls, medical stores, and in many applications where storage and retrieval is involved. As a future scope, one can further implement this sort of technique in various industries where many number of parts are getting manufactured and keeping record of storage and retrieval tasks is tedious. In such industries, one can implement such system for keeping a track of stored and retrieved products.

158

S. Ajanalkar and H. Joshi

Important Websites 1. K. J. Roodbergen (2016) Warehouse Layout Optimizer Tool, Interactive Warehouse Tool and Literature on Warehouse by Prof. Kees Jan Roodbergen http:// www.roodbergen.com/ Accessed 25 Jan 2021. 2. Dilip Kumar Pratihar (2018) National Program on Technology Enhanced Learning (NPTEL) Online Course on “Robotics” https://nptel.ac.in/courses/112/105/ 112105249/ Accessed 26 Jan 2021. 3. Gaurav Raina and Tanmai Gopal (2016) National Program on Technology Enhanced Learning (NPTEL) Online Course on “Introduction to Modern Application Development” https://nptel.ac.in/courses/106/106/106106156/ Accessed 25 Jan 2021. 4. Refsnes Data (1998) Online programming tutorials, reference material and examples for web application development training https://www.w3schools.com/ Accessed 26 Jan 2021.

References Abbas A, Faiz A (2013) Usefulness of digital and traditional libraries in higher education. Int J Serv Technol Manag 19:149–161. https://doi.org/10.1504/IJSTM.2013.054203 Angal Y, Gade A (2017) Development of library management robotic system. In: 2017 international conference on data management, analytics and innovation (ICDMAI), pp 254–258 Animireddy SP, Singh KP, Neha, NV (2018) Robotic library assistant. In: 2018 second international conference on inventive communication and computational technologies (ICICCT), pp 1443–1447 Behan J, O’Keeffe DT (2008) The development of an autonomous service robot. Implementation: “Lucas” – the library assistant robot. Intell Serv Robot 1:73–89. https://doi.org/10.1007/s11370007-0005-0 Chen Z, Zhang Y, Wang H, Chen W (2016) Real-time tag recognition based on morphology and local contrast. In: 2016 IEEE international conference on real-time computing and robotics (RCAR), pp 614–619 Comsa A, Maniu I, Modler N et al (2014) Automated book manipulator in libraries. In: Pisla D, Bleuler H, Rodic A et al (eds) New trends in medical and service robots: theory and integrated applications. Springer International Publishing, Cham, pp 75–85 da Rosa IB, Ribeiro Lamas D, Peña Saavedra V (2012) Using digital and traditional libraries in developing countries. In: Chen H-H, Chowdhury G (eds) The outreach of digital libraries: a globalized resource network. Springer, Berlin, Heidelberg, pp 363–364 Di Veroli C, Le CA, Lemaire T et al (2019) LibRob: an autonomous assistive librarian. In: Althoefer K, Konstantinova J, Zhang K (eds) Towards autonomous robotic systems. Springer International Publishing, Cham, pp 15–26 Fei Y, Chengyu Z, Wu C (2015) Smart talking robot Xiaotu: participatory library service based on artificial intelligence. Library Hi Tech 33:245–260. https://doi.org/10.1108/LHT-02-2015-0010 Guarnizo O, Pineda I (2019) Path planning simulation in controlled environments using the ants colony optimization algorithm. In: Undergraduate Research Symposium; May 2019; Yachay Tech University, Ecuador; https://doi.org/10.13140/RG.2.2.20544.48649 Hahsler M, Hornik K (2007) TSP – Infrastructure for the Traveling Salesperson Problem. J Stat Softw 23(2):1–21. https://doi.org/10.18637/jss.v023.i02 Han-ye Z, Wei-ming L, Ai-xia C (2018) Path Planning for the mobile robot: a review Symmetry 10:450. https://doi.org/10.3390/sym10100450

7 Logic Based Path Planning (LBPP) Algorithm for Robotic Library System

159

Harvey PH, Currie E, Daryanani P, Augusto JC (2016) Enhancing student support with a virtual assistant. In: Vincenti G, Bucciero A, Vaz de Carvalho C (eds) E-learning, E-education, and online training. Springer International Publishing, Cham, pp 101–109 Gupta PK, Hira DS (2010) Operations Research. S. Chand and Company Limited, New Delhi. Hongjiu G (2013) Data mining in the application of E-commerce website. In: Du Z (ed) Intelligence computation and evolutionary computation. Springer, Berlin, Heidelberg, pp 493– 497 Juanatas RA, Juanatas IC (2019) RFID-based library management system with android mobile access application. In: 2019 International conference on computational intelligence and knowledge economy (ICCIKE), pp 270–274 Korkmaz M, Durdu A (2018) Comparison of optimal path planning algorithms. In: 2018 14th international conference on advanced trends in radioelecrtronics, telecommunications and computer engineering (TCSET), pp 255–258 Kuriya R, Tsujimura T, Izumi K (2015) Augmented reality robot navigation using infrared marker. In: 2015 24th IEEE international symposium on robot and human interactive communication (RO-MAN), pp 450–455 Li R, Huang Z, Kurniawan E, Ho CK (2015) AuRoSS: an autonomous robotic shelf scanning system. In: 2015 IEEE/RSJ international conference on intelligent robots and systems (IROS), pp 6100–6105 Li D-Y, Xie S-D, Chen R, Tan H (2016) Design of internet of things system for library materials management using UHF RFID. In: 2016 IEEE international conference on RFID technology and applications (RFID-TA), pp 44–48 Lin W, Yueh H (2018) Evaluating children’s performance and perception of robotic assistance in library book locating. In: 2018 27th IEEE international symposium on robot and human interactive communication (RO-MAN), pp 1185–1189 Mandava A, Antony S (2012) A review and analysis of technologies for developing web 975 applications. In: ITERA 2012 Matai R, Singh SP, Mittal ML (2010) Chap 1. Traveling salesman problem: an overview of applications, formulations, and solution approaches. In: Traveling salesman problem. IntechOpen, Rijeka Mehta R, Sahu A (2020) Autonomous robot for inventory management in libraries. In: 2020 IEEE international students’ conference on Electrical, electronics and computer science (SCEECS), pp 1–6 Mikawa M, Morimoto Y, Tanaka K (2010) Guidance method using laser pointer and gestures for librarian robot. In: 19th international symposium in robot and human interactive communication, pp 373–378 Naumov V, Karova M, Zhelyazkov D et al (2015) Robot path planning algorithm. Int J Comput Commun 9:96–99 Navaraja AT, Jain N, Sengupta D, Kumar CS (2016) Web based simulation and remote triggered laboratory for robots. In: Mandal DK, Syan CS (eds) CAD/CAM, robotics and factories of the future. Springer India, New Delhi, pp 665–677 Nguyen V, Tran T, Le T, et al (2015) Using hand postures for interacting with assistant robot in library. In: 2015 seventh international conference on knowledge and systems engineering (KSE), pp 354–359 Ozeer A, Sungkur Y, Nagowah SD (2019) Turning a traditional library into a smart library. In: 2019 international conference on computational intelligence and knowledge economy (ICCIKE), pp 352–358 Raja P, Pugazhenthi S (2012) Optimal path planning of mobile robots: a review. International Journal of the Physical Sciences 7. https://doi.org/10.5897/IJPS11.1745 Rashid MR, Uzzaman N, Hossain S, Shuvra NKD (2017) Development of a self-navigating algorithm for library book finder robot. In: 2017 3rd International conference on electrical information and communication technology (EICT), pp 1–6 Roodbergen KJ, Koster R (2001) Routing methods for warehouses with multiple cross aisles. Null 39:1865–1883. https://doi.org/10.1080/00207540110028128

160

S. Ajanalkar and H. Joshi

Roodbergen KJ, Vis IFA, Taylor GD (2015) Simultaneous determination of warehouse layout and control policies. Null 53:3306–3326. https://doi.org/10.1080/00207543.2014.978029 Sastry H, Manjunath G, Reddy LC (2011) User Interface design challenges for digital libraries. Int J Comput Appl 15. https://doi.org/10.5120/1955-2616 Sawant D, Patil R (2020) AD-LIB: automated library system. In: Ranganathan G, Chen J, Rocha Á (eds) Inventive communication and computational technologies. Springer Singapore, Singapore, pp 635–646 Sun Y, Meng H, Li WW et al (2013) Study on UAV path planning simulation. Adv Mater Res 765–767:452–455. https://doi.org/10.4028/www.scientific.net/AMR.765-767.452 Tarashnina SI, Pankratova YB, Purtyan AS et al (2017) On a dynamic traveling salesman problem. Contributions to Game Theory and Management, 10(0):326–338 Tomizawa T, Ohba K, Ohya A, Yuta S (2007) Remote food shopping robot system in a supermarket -realization of the shopping task from remote places. In: 2007 international conference on mechatronics and automation, pp 1771–1776 Vega F, Hwang E, Park L, et al (2020) Service robot navigation and computer vision application in a Banquet Hall setting. In: 2020 10th annual computing and communication workshop and conference (CCWC), pp 0918–0923 Wang T, Zhou Y, Liu Z (2019) 4-peg Hanoi towers algorithm animation demonstration system based on Html5. J Phys Conf Ser 1288:012059. https://doi.org/10.1088/1742-6596/1288/1/ 012059 WHATWG (2020) HTML: the living standard developer’s edition. https://html.spec.whatwg.org/ dev/. Accessed 3 Jan 2021 Yu Q, Yuan C, Fu Z, Zhao Y (2012) An autonomous restaurant service robot with high positioning accuracy. Industrial Robot 39:271–281. https://doi.org/10.1108/01439911211217107 Yu H, Li L, Chen J, et al (2019) Mobile robot capable of crossing floors for library management. In: 2019 IEEE international conference on mechatronics and automation (ICMA), pp 2540–2545 Zia M, Çakir Z, Seker D (2018) Spatial transformation of equality – generalized travelling salesman problem to travelling salesman problem. ISPRS Int J Geo Inf 7:115. https://doi.org/10.3390/ ijgi7030115

8

Design, Control, and Data Management for Cleaning-in-Place (CIP) Test Rig Used in Process Industries A. S. Patil, M. N. Dhavalikar, and S. A. Chavan

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classification of CIP Based on Number of Stages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Three-Stage CIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Five-Stage CIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seven-Stage CIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CIP System Design: A Case Study in Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vessel Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pump Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection of Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PLC Programming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SCADA Screen Development and Programming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HMI Screens Development and Programming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Real-Time Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Program Validation on Hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

162 164 164 165 166 167 168 170 171 171 180 185 188 188 190 191 191 192

Abstract

Cleaning-in-place (CIP) is a very important aspect of every part of the food, beverage, and pharmaceutical industry. Sanitation could greatly affect production

A. S. Patil () · M. N. Dhavalikar MIT School of Engineering, MIT ADT University, Pune, India e-mail: [email protected] S. A. Chavan Analogic Automation Pvt. Ltd, Pune, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_151

161

162

A. S. Patil et al.

because safety and quality are at stake. CIP refers to the method of cleaning a piece of equipment or its parts without disassembling them. As some equipment and machines may be bulky or too complicated to disassemble, it is not advisable to even remove the parts. Manual cleaning is also not possible for most of the parts. The proposed work includes the mechanical system design, process flow design, automation, and testing of the programs of PLC, SCADA, and human– machine interface systems. This CIP trainer kit would demonstrate the various PLC logic testing of the respective CIP processes. It also facilitates the training of newly recruited engineers in aspects of process industries such as food, brewery, pharmaceutical, chemical. Keywords

Cleaning in place · Sanitation · PLC logic · SCADA

Introduction Cleaning is the process of removing the dirt, marks, or any unwanted matter from a piece of equipment. Nowadays, hygiene and sanitation concerns are becoming one of the most important for the food, beverage, dairy, pharmaceutical, and chemical industries. The production system of all these industries regularly contains a number of processing vessels, storage tanks, piping arrangements, heat exchangers, pumps, and valves that have to be cleaned regularly. There are various methods of cleaning this equipment, such as foam cleaning, high-pressure cleaning, mechanical cleaning, Cleaning-Out-of-Place (COP), and Cleaning-In-Place (CIP). All these cleaning processes have their own pros and cons, but CIP is the most suitable, most widely used, safest, and most accurate, efficient, and automated process with minimal production down time and cleaning costs. India is the largest milk producer in the world, with 10 million dairy farms, 96,000 local and 15 state dairy cooperatives, along with the 170 milk producers’ cooperative unions. Milk is the largest crop in India, with a market value of 6.5 lakh crore/annum, which contributes around 26% in total agricultural GDP. Similarly, the food processing and beverage industry is the fifth largest industry of India, which has 1.77 million employees working daily in 39,319 registered units, which have $29.2 billion fixed capital and $144.6 billion aggregate output per year. This industry has a 32% share of India’s total food market, 13% of Indian exports, and 6% of India’s total industrial investment. On the other hand, the Indian pharmaceutical industry is the world’s largest generic medicine producer with third rank for production by volume and tenth by value. India has 3000 pharmaceutical companies with a $20.02 billion domestic market turnover and around $17.27 billion exports. India also supplies 62% of the global demand for vaccinations. The Indian chemical industry is the sixth largest producer of chemicals in the world and the third in Asia. It has 14th rank in the export of chemicals. India supplies 16% of the overall

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

163

world dye demand. The Indian chemical industry has around 80,000 manufacturing products, 2 million employees, and a market value of $35 billion. In 2016, Nighot et al. proposed a CIP system for food and beverage industries with the use of PLC and SCADA. The paper gives basic information about the automated CIP process, its components, and assembly. Automated CIP is developed in RSLogix500 and the testing and monitoring is done on the SCADA software Factory Talkview. The proposed system improves the efficiency, accuracy, and guarantee of cleaning with respect to the conventional cleaning processes (Nighot et al. 2015). In 2015, Vaishnavi et al. proposed a CIP system for the pharmaceutical industry. The paper gives information about the use of CIP in the pharmaceutical industry, its advantages, a basic flowchart of CIP process, whose program logic is developed in RSLogix500 and SCADA screen is developed in Factory Talkview software (Vaishnavi et al. 2015). In 2017, Swapnil et al. proposed use of the CIP process in the dairy industry. The paper gives the basic principles of the CIP process to overcome the disadvantages of existing human oriented processes by using automated methodology. It also provides information about the various CIP processes, cleaning solutions, cleaning procedures, the basic design, and the CIP program description used in the dairy industry (Swapnil et al. 2017). In 2014, Thomas and Sathian reviewed the CIP system used in the dairy industry. It gives thorough information about the various types of cleaning strategies and cleaning solutions used in the food industry, their cleaning procedures, and cleaning efficiency. It concluded that reuse of cleaning solutions a certain number of times improves the cleaning efficiency with the expected accuracy of the cleaning process (Amitha and Sathian 2014). In 2018, Raj et al. reviewed cleaning validation in the pharmaceutical industry. Their article focusses on the validation of the CIP process performed in the pharmaceutical industry to avoid severe hazards formed owing to contaminations and cross-contaminations occurred in improper CIP processes. It gives the various validation protocols, their selection criteria, and basics about the validation report (Raj et al. 2018). In 2010, Alvarez et al. carried out a case study of an ultra-high temperature heat exchanger and recommended the CIP process rationalization in the dairy industry to reduce the operating time and effluents of the CIP process. This CIP process includes a first rinse phase, a chemical cleaning phase, a detergent rinsing phase, which reduces the effluent by more than half. The text of CIP is taken several times and a graph of turbidity vs time is shown (Alvarez et al. 2010). In 2015, Nurgin et al. proposed the CIP process to be used in the dairy industry. The paper elaborates the basic CIP process and the system with material selection guidelines, the design of centralized CIP system, and its program strategy. The overview of the CIP processes is shown on the SCADA screen and the functioning of it is tested using a logic program (Nurgin et al. 2015). Figure 1 shows an example of a CIP process used in a yoghurt production station in the dairy industry. Here, a simple water wash is carried out, along with a caustic and acid wash. All this previous work has been done for particular industrial equipment cleaning. As CIP is a widely used process in different kinds of industries, the generalized

164

A. S. Patil et al. Soft water

Drain

Product pushing

Drain

Drain

Drain

Drain

Pre-rinsing

Caustic circulation

Rinsing

Acid circulation

Sterile water

Rinsing

Drain

Raw water

Caustic circulation

Final rinsing

Drain

Acid circulation

Fig. 1 Process flowchart of cleaning-in-place in a yoghurt production station (Swapnil et al. 2017)

process automation helps newly joined employees, trainees, and interns in their training. Cleaning-in-place is one of the important processes that is used in so many industries. Also, it is one of the best examples of automated processes. Therefore, the training on the CIP test rig helps the operator to upgrade their knowledge of industrial process automation. For this purpose, the design and demonstration of a CIP test rig can perform the desired task with the required output.

Classification of CIP Based on Number of Stages As per the requirement and sensitivity of the production process various types of CIP processes are used. There are three main types of CIP processes according to the number of cleaning stages used: three-stage CIP, five-stage CIP, and seven-stage CIP.

Three-Stage CIP The storage tanks of the same product or material are cleaned periodically . This cleaning is simple washing; thus, it is done using the three-stage CIP process. Also, the intermediate cleaning of the simple process vessels is carried out with the help of a three-stage CIP process. The process flow chart of the three-stage CIP process is shown in Fig. 2, where only three sub-stages or cycles are used for overall cleaning, which are Pre-Cold Water Rinsing, Hot Water Rinsing, and Post-

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

165

Fig. 2 Process flowchart of three-stage CIP

Fig. 3 System layout of a three-stage CIP

Cold Water Rinsing. Each cycle contains three subprocesses: Filling, Rinsing, and Draining. Only the hot water is restored in a hot water vessel and reused for some cycles. Figure 3 shows the SCADA system layout of the three-stage CIP system. All the tags are mentioned in the legends table with their respective description on the upper right-hand side of the screen. Below the legends table all the stages of the CIP are listed. Three-stage CIP is one of the simplest CIP process and requires a Hot Water Vessel and a Cold Water Storage Tank along with two inlet valves, two outlet valves, one drain valve, four level switches, two pumps, one water heater, and one temperature sensor.

Five-Stage CIP A five-stage CIP process has two additional stages or cycles compared with the three-stage CIP process mentioned in Fig. 1, which are the Detergent Cycle and the Post-Hot Water Rinsing Cycle. All other subprocesses of the CIP are the same as for three-stage CIP. In five-stage CIP, the recovery of hot water as well as the detergent solution is carried out; thus, after the rinsing of hot water and detergent

166

A. S. Patil et al.

Fig. 4 Process flowchart of five-stage CIP

solution, the restoration process is carried out. As the additional vessel for the detergent solution is required in five-stage CIP, the starting phase of CIP contains two subprocesses: Pre-Filling of Hot Water Vessel and Pre-Filling of Detergent Vessel. Figure 4 shows the detailed process flow chart of the five-stage CIP process along with all the subprocesses of each cycle or stage. This five-stage CIP process is used in intermediate sensitive product vessel cleaning. Figure 5 shows the SCADA system layout of the five-stage CIP system. All the tags are mentioned in the legends table with their respective descriptions at the upper right-hand side of the screen. Below the legends table all the stages of the CIP are listed. Five-stage CIP requires some additional equipment and instruments compared with the three-stage CIP owing to the two additional stages. This system requires a Hot Water Vessel, a Cold Water Storage Tank and the Detergent Vessel along with the four inlet valves, three outlet valves, one drainage valve, eight level switches , two pumps, one water heater, and one temperature sensor.

Seven-Stage CIP The seven-stage CIP process has two more additional stages or cycles compared with the five-stage CIP process mentioned in Fig. 4: which are the Chemical Cycle and the Intermediate Cold Water Rinsing Cycle. All other subprocesses of the CIP are the same as for the five-stage CIP process. In the seven-stage CIP, the recovery of hot water, detergent solution, along with the chemical solution, is carried out; thus, after the rinsing of the hot water cycle, the detergent solution cycle, and the chemical cycle, the restoration process is carried out. As an additional vessel for the chemical solution is required in seven-stage CIP, the starting phase of

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

167

Fig. 5 System layout of five-stage CIP

CIP contains three subprocesses: Pre-Filling of the Hot Water Vessel, Pre-Filling of the Detergent Vessel, and Pre-Filling of the Chemical Vessel. Figure 6 shows the detailed process flow chart of the seven-stage CIP process along with all the subprocesses of each cycle or stage. Figure 7 shows the SCADA system layout of the seven-stage CIP system. All the tags are mentioned in the legends table with their respective description at the upper right-hand side of the screen. In the legends table below all the stages of the CIP are listed. Seven-stage CIP requires some additional equipment and instruments compared with the five-stage CIP owing to the addition of two stages. This system requires a Hot Water Vessel, a Cold Water Storage Tank, a Detergent Vessel, and the Chemical Vessel along with the six inlet valves, four outlet valves , one drain valve, 12 level switches, two pumps, one water heater, and one temperature sensor. Here, the detergent solution and the chemical solutions are used in diluted forms in cold water. The concentrated solutions of the same are stored in the respective tanks, which are poured into diluted tanks in the correct proportions through the additional inlet valves for both vessels. The seven-stage CIP is one of the most complicated CIP processes and it is used when product recipes are changed or in the cleaning of the most sensitive product vessels.

CIP System Design: A Case Study in Industry The proposed system contains two cylindrical water tanks with identical dimensions, one cubical water storage tank, five identical on–off valves, two nonreturn valves, two pumps, and the piping arrangement.

168

A. S. Patil et al.

Fig. 6 Process flowchart of seven-stage CIP

Fig. 7 System layout of seven-stage CIP

Vessel Calculations The Hot Water Vessel and the Main Vessel are of the same size and shape. Thus, all the calculations are the same for the both. The volume of each vessel is 250 l. All the design calculations are made according to the India Standard IS 2825-1969, a code for an unfired pressure vessel. According to these input data the diameter and

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

169

height of the vessel are calculated as follows. Volume of Cylinder = π r 2 h

(1)

Considering, h = 2 (Internal diameter of cylinder) = 2 (di) = 2 (2r) h = 4r

(2)

where r = Radius of cylinder h = Height of cylinder

Putting this value of h in Eq. (1), Volume of Cylinder = π r 2 (4r) 0.25 = 4π r 3 Inner radius of cylinder = r = 0.27 m

Inner diameter of cylinder = di = 2r = 2 ∗ 0.27 = 0.54 m Height of cylinder = h = 2(di) = 2 ∗ 0.54 = 1.08 m. From Eq. (2) Working Pressure = Water Head Pressure P = ρgh = 1000 ∗ 9.81 ∗ 1.08 P = 10, 594.8 N/m2 Design Pressure = Pd = 1.05 P = 1.05 ∗ 10, 594.8 = 11, 124.54 N/m2 Pd = 0.0111 Mpa

(3)

(Bhandari 2017)

Allowable Design Stress = Syt/1.5 (Bhandari 2017) σall = 290/1.5 (SS 316) σcall = 193.33 MPa

(4)

(Ductile Material)

Corrosion Allowance = C C = 1.5 mm (Indian Standard 1998) . (For normal operating condition) Efficiency of Weld Joint = η

(5)

170

A. S. Patil et al.

Class III weld is used with double-welded butt joint with full penetration and nonradiographed η = 0.7 (Indian Standard 1998) Thickness of Pressure Vessel = tCylinder =

P d∗Di (2∗σ all∗η−P d) + C 0.0111∗540 2∗193.33∗0.7−0.0111 + 1.5

(Bhandari 2017)

tCylinder = 1.522 mm (6) • End Closure Calculations Plain formed head is selected as it requires a minimal amount of forming so it is most economical and used for inner diameter (Di) < 7 m. Head Thickness = 0.4 ∗ Di tHead =



Pd +C σ all  0.011 0.4 ∗ 540 193.33

(Bhandari 2017) + 1.5

(7)

tHead = 3.137 mm

Pump Calculations The suction head and delivery head of the pump are considered as, Suction Head = Hs = 1.2 m Delivery Head = Hd = 2.5 m Therefore, Total Head = Suction Head + delivery Head = Hs + Hd = 1.2 + 2.5 H = 3.7 m The pump has to fill the vessel in 5 min. The volume of the vessel is calculated, which is 250 l and is filled with water. Therefore, Discharge of Pump = Q = Mass Flow Rate/Time = (Volume of Vessel ∗ Density of water) /Time = (0.25 ∗ 1000) /300 Q = 0.833 l/s = 2998.8 lph

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

171

Selection of Components The various components such as sensors, controller, and actuators are used in the proposed CIP system. The selection of these components is given in Table 1.

PLC Programming A programmable logic controller (PLC) is basically a controller that can be reprogrammed and is mostly used for heavy automation systems with harsh operating environments such as industrial automation. A PLC is programmed in a graphical programming language called Ladder Diagram or LD or LAD. This programming can be carried out using various software programs such as Logix 500, Logix 5000, Studio 5000, with various versions such as version 24, version 28, and version 32. The proposed system is programmed using Studio 5000 software v. 24, which is the high-end factory automation programming software with the upgraded new features and is owned by Rockwell Automation (which was previously known as “Allen Bradley”). The Input/Output list, which is commonly known as the input/output (I/O) list of the system, as given in Table 2. According to this, the various sensor outputs are considered as inputs along with the various switches such as the start switch, restart switch, and authorization switch. The various actuators mentioned below are activated through the logic of various normally closed (NC) and normally open (NO) switches and are mapped at the output side of the logic. The complete programming of the proposed system is as shown in Fig. 8. There are 29 rungs with the one end rung. Each rung starts with the start switch (SS_01) so that the whole program or the system starts whenever the start switch is pressed. The Start Switch (SS_01) is an NO type of switch, which lights up the attached output only after the pressing action of the assigned hardware switch or the virtual human–machine interface (HMI) window switch. The first two rungs, such as the zeroth rung and first rung, show the programming of two inlet valves of the hot water vessel (IV_02) and the inlet valve of the main vessel (IV_01). The water filling in the hot water vessel will start when the zeroth rung gives the output and this happens when all the NC NO switches attached in the rung are in the ON state. Thus, after the start switch, the Outlet Valve 02 (OV_02) is used in the rung, which is the output but is here used as the input in the format of a normally closed type switch so that it gives output whenever the outlet valve is closed. This condition is used to avoid the opening of the inlet and outlet valve of the same vessel at the same time. At the third position of the zeroth rung the Level Switch High 02 (LSH_02_HV) is used, which indicates the high level of the hot water vessel. This sensor output is used as the normally closed input switch so that it turns OFF the inlet valve when the water level reaches the high level mark. After this the equal logic box is used from the comparison section, which compares the values of Source A and Source B and gives the output only when both values are equal. Source A is the process counter reading in real time and Source B is set at the constant value zero. This logic opens the inlet valve at the time of the filling process. In parallel to this, another equal logic box is used where current process counter reading (Source A) is compared with the

172

A. S. Patil et al.

Table 1 Selection of components Sr. No. 1

Block name Sensor

Component name Level switch

Temperature sensor

2

Controller

Software

Hardware

3

Actuator selection

Valve

Pump

4

SCADA system

Software

Hardware

5

HMI system

Software

Hardware

6

Other mechanical components

Non return valve

Criteria Type Application material Make Mounting Type

Description Float type Water Baumer Side mount RTD

Make Range Make Name Version Make Model

Baumer −200 ◦ C to 450 ◦ C Allen Bradley Studio 5000 24.11 Allen Bradley Compact Logix (L18ER) 16 16 Auto ON–OFF Ball Valve Delval Centrifugal type Kirloskar 300 to 3300 LPH Allen Bradley

Digital input Digital output Type Make Type Make Flow rate Make Name Type Version Make Display size Processor Processor speed Make Name Type Version Make Type Display size Resolution Line size

FT view View site edition 10.00 Dell 18.5 inches Core i3 (10th generation) 3.6 GHz Allen Bradley FT view Machine edition 10.00 Allen Bradley LCD 4 inches 480 * 272 0.5

Make Temperature range Body material

Alfa Laval −10 ◦ C to 140 ◦ C SS 316 L

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

173

Table 2 Input/output (I/O) list Sr. No. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

Instrument description Limit switch high for main vessel Limit switch low for main vessel Limit switch high for hot water vessel Limit switch low for hot water vessel Temperature transmitter ON–OFF inlet valve for main vessel ON–OFF outlet valve for main vessel ON–OFF drain valve ON–OFF inlet valve for hot water vessel ON–OFF outlet valve for hot water vessel Cold water pump Hot water pump

Tag no. LSH_01_MV LSL_01_MV LSH_02_HV LSL_02_HV TT_01 IV_01 OV_01 DV_01 IV_02 OV_02 P_01 P_02

I/O type Digital input Digital input Digital input Digital input Digital input Digital output Digital output Digital output Digital output Digital output Digital output Digital output

constant value two (Source B). Also, the Outlet Valve 01 (OV_01) is connected in series to it to achieve the restoration of the hot water again in the hot water vessel after the hot water rinsing cycle. Last, the authorization of the completion of the CIP process switch is used, which is the NC input switch. This switch keeps the inlet valve closed until authorization is complete. The next rung, which is the first numbered in the logic, shows the logic for the inlet valve opening of the main vessel. This rung is again started with the Start Switch (SS_01); after that, the two parallel branches go towards the output. The first branch indicates the opening and closing conditions of the main vessel inlet valve at the time of Pre Cold Water Rinsing Cycle. This branch contains the Outlet Valve 01 (OV_01), which is normally a closed input switch and gives the output only when the outlet valve of the main vessel is closed to allow the filling only when the outlet is OFF. The Level Switch High 01 (LSH_01_MV) allows the filling of water in the vessel up to the high mark to avoid the overflow condition. Level Switch High 02 (LSH_02_HV) allows the filling of water in the main vessel only when the hot water vessel is filled up to the high level mark, which also indicates the completion of the filling process. The second branch indicates the opening and closing conditions of the main vessel inlet valve at the time of the Hot Water Rinsing Cycle. This branch contains the Outlet Valve 01 (OV_01) as above, along with the Outlet Valve 02 (OV_02), which is normally an open-type input switch, and allows the opening of inlet valve 01 only when the opening of the outlet valve of hot water vessel to avoid the dry running of the second pump. Level Switch High 01 (LSH_01_MV) closed the inlet valve when the high level reached the main vessel. The third branch indicates the opening and closing conditions of the main vessel inlet valve during the Post Cold Water Rinsing Cycle. Here, the equal logic box is used with current reading of the process counter as a Source A and the constant value two as the Source B, which indicates the Post Cold Water Rinsing Cycle. This branch closes the inlet valve when the high level is reached in the vessel

174

A. S. Patil et al.

Fig. 8 Ladder logic for the inlet valve of the main vessel

with the help of a Level Switch High 01 (LSH_01_MV) sensor reading. Finally, authorization of the completion of the CIP process switch is used, which is the NC input switch. This switch keeps the inlet valve closed until authorization is complete. The next rung, which is second numbered in the logic, is used to program the rinsing timing of the all cycles of the main vessel. Here, the retentive type of timer is used, which counts the time when the Start Switch (SS_01) is pressed and the Level Switch High 01 (LSH_01_MV) is giving the high level output of the main vessel. The tag name of this timer is TIMER_01, the timer is set for 5 min with 10-ms time resolution, which counts the 3,00,000 counts that is each count for 10 ms. The Accumulator gives the gradually increasing timer reading when the timer

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

175

is in the ON condition. When the preset valve and the accumulator value are exactly equal the Done bit of timer gives the output. As the retentive type of timer is used here, its reset condition must be given in the next rung; thus, it is given in the third rung of the logic. Level Switch Low 01 (LSL_01_MV) is a sensor output of a low-level mark of the main vessel, which is used here as a NO type of input switch to reset the timer. Also, the hardware restart CIP switch or the onscreen HMI button of the restart CIP resets the timer, as shown in the second branch of the third rung. The Count Up counter is used in the fourth rung of ladder logic to count the subprocesses of the CIP. The tag name of the counter is COUNT1 with preset 3 and accumulator gives gradually increasing readings whenever the Done bit of the Rinsing Timer (TIMER_01) gives the output. When the accumulator reading is equal to the preset value, then the Done bit of the counter gives the output. In the fifth rung, the counter resets when application of hardware or HMI screen Restart CIP button. Also, after counting all three sub-processes of CIP, the counter resets by using the condition in the second branch of the fifth rung, which contains an equal logic box to the process counter (COUNT1) accumulator reading as a Source A and constant value three as Source B. This equal logic box is followed by the Level Switch Low 01 (LSL_01_MV), which resets the counter after draining all the water in the Post Cold Water Rinsing Cycle. The sixth rung of ladder logic deals with the ON–OFF conditions of the Outlet Valve 01 (OV_01) of the Hot Water Vessel. As usual, the rung starts with the Start Switch (SS_01). After that, the Done bit of the Rinsing Timer (TIMER_01) is used, which opens the outlet valve when the Rinsing cycle is done. After archiving the lowermost water level in the tank the Level Switch Low sensor gives the output, which is used here as a NC input switch that finally turns the outlet valve OFF. Last, the three Equal Logic Boxes are used on the input side in the three parallel branches, which give information about the current ongoing cycle. This arrangement also allows the opening of the outlet valve after three cycles only. The seventh rung, provides the operating conditions for the Drain Valve 01 (DV_01) of the Main Vessel. As usual, the rung starts with the Start Switch (SS_01) to allow the opening of the valve only when the CIP process is functional. After the Start Switch, the seventh rung divides into two branches. The first branch of the seventh rung allows the draining of water from the Main Vessel. For this purpose, the Done bit of the Rinsing Timer (TIMER_01) is used, which allows the drain of water only after the completion of the Rinsing Cycle. The Level Switch Low 01 (LSL_01_MV) is used after the Rinsing Timer Done bit, which keeps the Drain Valve open until the lowest water level in the Main Vessel is achieved. This automatically turns OFF the Drain Valve when the Low Level Sensor established at the lower side of the Main Vessel gives the output. After this Level Switch the two Equal Logic Boxes are used in the parallel branches, which allows the Drain Valve to open only after the Pre and Post Cold Water Rinsing Cycle. This arrangement helps in Restoration of the Hot Water by keeping the Drain Valve Closed at the time of Hot Water Rinsing Cycle. So to achieve this, the Equal Logic Boxes are used along with the accumulator reading of the Process Counter as a Source A and

176

A. S. Patil et al.

the constant values one and three as the Source B in the respective branches. The second branch of the seventh rung, specifies the Drain Valve opening and closing condition at the time of draining of hot water from the Hot Water Vessel. After the use of the same hot water for the certain period of times, the quality of the water is reduced below the desired acceptance limit of the reusable water. Thus, after this, the restored hot water is drained out by using this facility. To achieve this, the Drain Hot Water Vessel Switch (Drain_HV) is provided on the hardware switch kit, as well as on the HMI screen to give the command. This command switch is used here in logic as the NO input switch (Drain_HV), which is then followed by the Level Switch Low 02 (LSL_02_HV), allowing the closure of the Drain Valve after achieving the lowermost level in the Hot Water Vessel. The Level Sensor established at the lower side of the Hot Water Vessel gives the low-level signal, which is used here as the NC input switch. The eighth rung of the ladder logic diagram, provides the operating conditions of the Cold Water Pump (P_01). The rung starts with the Start Switch (SS_01) followed by the two parallel branches. The first branch of the eighth rung defines the Cold Water Filling Process in the Hot Water Vessel. Here, the Inlet Valve of the Hot Water Vessel (IV_02) is used as an NO inlet switch, which allows the pump to start only when the inlet valve is open. This avoids excess pressure in the fully closed pipes. After this, the Outlet Valve of the Hot Water Vessel (OV_02) is used as the NC input switch, which allows the vessel to be filled by blocking the outflow. The Level Switch High of the Hot Water Vessel (LSH_02_HV) sensor is used as the NC input switch, which stops the pump when the high level in the Hot Water Vessel is reached. The second parallel branch of the eighth rung defines the Cold Water Filling Process. Here, the Inlet Valve of the Main Vessel (IV_01) is used as an NO inlet switch, which allows the pump to start only when the inlet valve is open. This avoids excess pressure in the fully closed pipes. After this, the Outlet Valve of the Main Vessel (OV_01) is used as the NC) input switch, which allows the vessel to be filled by blocking the outflow. The Level Switch High of the Main Vessel (LSH_01_MV) sensor is used as the NC input switch, which stops the pump when the high level in the Main Vessel is reached. Further, in two parallel branches, the Pre and the Post Cold Water Filling Process is provided by using the Equal Logic Boxes with Source A as the current accumulator reading of the process counter (COUNT1) and Source B as the constants zero and two. The ninth rung of the ladder logic diagram provides the operating conditions of the Hot Water Pump (P_02). The rung starts with the Start Switch (SS_01) followed by the two parallel branches. The first branch of the ninth rung defines the Hot Water Restoration Process. Here, the Outlet Valve of the Main Vessel (OV_01) is used as an NO inlet switch, which allows the pump to start only when the outlet valve is open. This avoids the dry running of the pump. After this, the Drain Valve (DV_01) is used as an NC input switch for blocking the draining of the hot water. Last, the inlet Valve of the Hot Water Vessel (IV_02) is used as the NO input switch, which allows the vessel to be filled by avoiding the excess pressure in the fully closed pipes. The second branch of the ninth rung, defines the Hot Water Filling Process during the Hot Water Rinsing Cycle. Here, the Outlet Valve of the Hot Water Vessel

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

177

(OV_02) is used as an NO inlet switch, which allows the pump to start only when the outlet valve is open. This avoids the dry running of the pump. After this, the Drain Valve (DV_01) is used as an NC input switch for blocking the drainage of the hot water. Last, the Inlet Valve of the Main Vessel (IV_01) is used as the NO input switch, which allows the vessel to be filled by avoiding excess pressure in the fully closed pipes. The tenth rung of the ladder logic diagram provides the operating conditions of the Heater (H_01). The rung starts with the Start Switch (SS_01) followed by the Level Switch High of the Hot Water Vessel (LSH_02_HV), which allows the ON condition of the heater only when the Hot Water Vessel is full of water. After this, the Less Than Logic Box is used, where Source A is the analog input from the RTD Temperature Sensor (TT_01) implemented in the Hot Water Vessel and Source B is the Temperature Set Point (Temp_Set), which can be changed by the HMI Screen. This arrangement keeps the heater ON only when the Source B reading is less than the Source A reading, i.e., when the RTD reading is less than the Set Point. The eleventh rung of the ladder logic diagram provides the operating conditions of the Outlet Valve of the Hot Water Vessel (OV_02). The rung starts with the Start Switch (SS_01) followed by the Outlet Valve of the Main Vessel (OV_01) as an NC input switch, which allows the Main Vessel to be filled. After that the heater is used as the NC input switch, which opens the outlet valve only when the temperature of the hot water reaches the set point. The Level Switch Low of the Hot Water Vessel (LSL_02_HV) turns OFF the outlet valve when the lowest water level in the hot water vessel is reached. Finally, the two branches provide the two conditions for the opening of the outlet valve. The first one allows the restoration of hot water after the Hot Water Rinsing Cycle and the second one allows the drainage of hot water from the Hot Water Vessel after several uses of the hot water. From the 12th rung onwards each CIP process and its Sub-Processes are used as the outputs on the output side to light up the indicating lights in the SCADA as well as in the HMI screens. The 12th rung of the ladder logic diagram shows the start of the Filling Process of the Hot Water Vessel. After pressing of the Start Switch (SS_01), the Process Counter accumulator is at zero, as shown in the Equal Logic Box, with Source A as a current reading of counter accumulator (COUNT1) and Source B as the constant value zero, followed by any one of the inlet valves such as the Inlet Valve of the Main Vessel (IV_01) or the Inlet Valve of the Hot Water Vessel (IV_02) being open. The output is latched so that it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true. The 13th rung of the ladder logic diagram shows the start of the Filling Process of the Main Vessel under the Pre-Cold Rinsing Cycle. After pressing the Start Switch (SS_01), the Process Counter accumulator is at zero or at two, as shown in the Equal Logic Boxes, with Source A as a current reading of the counter accumulator (COUNT1) and Source B as the constant values zero and two, followed by the Inlet Valve of the Main Vessel (IV_01), which is used as an NO input switch. The output is latched so it stays in the ON condition whenever the unlatching condition mentioned in rungs 27 and 28 is true.

178

A. S. Patil et al.

The fourteenth rung of the ladder logic diagram shows the start of the Rinsing Process of the Main Vessel under the Pre-Cold Water Rinsing Cycle. After pressing of the Start Switch (SS_01), the Process Counter accumulator is at zero as shown in Equal Logic Box with Source A as a current reading of counter accumulator (COUNT1) and Source B as the constant value zero, followed by the counting bit of the Rinsing Timer (TIMER_01), which is used as the NO input switch to illuminate the light. The output is latched so it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true. The fifteenth rung of the ladder logic diagram, shows the start of the Draining Process of the Main Vessel under the Pre-Cold Water Rinsing Cycle. After pressing the Start Switch (SS_01), the Process Counter accumulator is at one as shown in the Equal Logic Box, with Source A as a current reading of the counter accumulator (COUNT1) and Source B as the constant value one, are equal, as well as the Drain Valve (DV_01), which is used as the NO input switch is open, then the drainage process indicator of the Pre-Cold Water Rinsing Cycle lights up. The output is latched, so it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true. The sixteenth rung of the ladder logic diagram shows the completion of the PreCold Water Rinsing Cycle. After pressing the Start Switch (SS_01) when Filling (Filling_01), Rinsing (Rinsing_01), Draining (Drain_01) of the Pre-Cold Water Rinsing Cycle are complete and when the Filling process (Filling_02) of the Hot Water Rinsing Cycle starts, then the Pre-Cold Water Rinsing Cycle is considered complete. All the above-mentioned processes are used as the NO input switches so that they light up the output whenever they are true. The output is latched so that it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true. The seventeenth rung of the ladder logic diagram shows the start of the Filling Process of the Main Vessel under the Hot Water Rinsing Cycle. After pressing the Start Switch (SS_01), the Process Counter accumulator is at one, as shown in the Equal Logic Box, with Source A as a current reading of the counter accumulator (COUNT1) and Source B as the constant value one are equal, as well as the Inlet Valve of the Main Vessel (IV_01) along with the Outlet Valve of the Hot Water Vessel (OV_02), which are used as the NO input switches, are open, then the filling process indicator of the hot water rinsing cycle will light up. The output is latched so that it stays in the ON condition whenever the unlatching condition mentioned in rungs 27 and 29 is true. The eighteenth rung of the ladder logic diagram shows the start of the Rinsing Process of the Main Vessel under the Hot Water Rinsing Cycle. After pressing the Start Switch (SS_01), the Process Counter accumulator is at one, as shown in the Equal Logic Box, with Source A as a current reading of the counter accumulator (COUNT1) and Source B as the constant value one, followed by the counting bit of the Rinsing Timer (TIMER_01), which is used as the NO input switch to light up the light. The output is latched so that it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true.

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

179

The nineteenth rung of the ladder logic diagram shows the start of the Recovery Process of hot water from the Main Vessel under the Hot Water Rinsing Cycle. After pressing the Start Switch (SS_01), the Process Counter accumulator is at two as shown in Equal Logic Box, with Source A as a current reading of the counter accumulator (COUNT1) and Source B as the constant value two, are equal, as well as the Outlet Valve of the Main Vessel (OV_01), which is used as the NO input switch, is open, then the Recovery Process indicator of the hot water rinsing cycle will light up. The output is latched so that it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true. The twentieth rung of the ladder logic diagram shows the completion of the Hot Water Rinsing Cycle. After pressing the Start Switch (SS_01) when Filling (Filling_02), Rinsing (Rinsing_02), Recovery (Recovery) of the Hot Water Rinsing Cycle are complete and when the Filling process (Filling_03) of the Post-Cold Water Rinsing Cycle starts, then the Hot Water Rinsing Cycle is considered complete. All the above-mentioned processes are used as the NO input switches so that they light up the output whenever they are true. The output is latched so that it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true. The 21st rung of the ladder logic diagram shows the start of the Filling Process of the Main Vessel under the Post-Cold Water Rinsing Cycle. After pressing the Start Switch (SS_01), if the Process Counter accumulator is at two, as shown in Equal Logic Box, with Source A as a current reading of the counter accumulator (COUNT1) and Source B as the constant value two, are equal, as well as the Inlet Valve of the Main Vessel (IV_01), which is used as the NO input switch, is open, then the filling process indicator of the post-cold water rinsing cycle lights up. The output is latched so that it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true. The 22nd rung of the ladder logic diagram shows the start of the Rinsing Process of the Main Vessel under the Post-Cold Water Rinsing Cycle. After pressing the Start Switch (SS_01), if the Process Counter accumulator is at two as shown in the Equal Logic Box, with Source A as a current reading of the counter accumulator (COUNT1) and Source B as the constant value two, followed by the counting bit of the Rinsing Timer (TIMER_01), which is used as the NO input switch to illuminate the light. The output is latched so that it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true. The 23rd rung of the ladder logic diagram shows the start of the Draining Process of the Main Vessel under the Post-Cold Water Rinsing Cycle. After pressing the Start Switch (SS_01), if the Process Counter accumulator is at three, as shown in the Equal Logic Box, with Source A as a current reading of the counter accumulator (COUNT1) and Source B as the constant value three, are equal, as well as the Drain Valve (DV_01), which is used as the NO input switch, being open, then the drain process indicator of the Pre-Cold Water Rinsing cycle will light up. The output is latched so that it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true. The 24th rung of the ladder logic diagram shows completion of the PostCold Water Rinsing Cycle. After pressing the Start Switch (SS_01) when Filling

180

A. S. Patil et al.

(Filling_03), Rinsing (Rinsing_03), Draining (Drain_03) of the Post-Cold Water Rinsing Cycle are complete with the closing of the Drain Valve (Drain_01), this indicates completion of the Post-Cold Water Rinsing Cycle. All the abovementioned processes are used as the NO input switches and the Drain Valve as the NC input switch, so that they light up the output whenever the processes are true and the Valve is in the OFF state. The output is latched so that it stays in the ON condition whenever the unlatching condition mentioned in rung 27 is true. The 25th rung of the ladder logic diagram shows the completion of CIP process after completion of all the subprocesses. The rung starts with the start switch (SS_01), which is an NO-type input switch. After that, the completion switch of the last subprocess of CIP is used with the NO input switch. Finally, the NC input switch is used for the Drain Valve (DV_01), which gives the output at the closing of the drainage valve. This (CIP_Done) output is a latched output. The 26th rung of the ladder logic diagram unlatches the output of the 25th rung (CIP_Done). The authorization switch provided in SCADA Screen and the hardware panel is used here as an NO input switch to unlatch the CIP_Done output. All the latched outputs of the process indicators are unlatched in the three branches of the 27th, 28th, and 29th rungs. The restart button is used here as the NO input switch. The filling processes of the Pre-Cold Water Rinsing Cycle has another special condition of the opening of the inlet valve of the Hot Water Vessel (IV_02). This special condition is implemented in the 28th rung by using the NO input switch in the second parallel branch, whereas the first branch contains the start button as an NO input switch followed by the restart button as an NO input switch. Similarly, another special condition is required for the filling process of the Hot Water Vessel, which is provided in the 29th rung of the ladder logic diagram. The NO input switch is used for the working of a heater placed inside the Hot Water Vessel at the second parallel rung, whereas the first branch contains the start button as an NO input switch followed by the restart button as an NO input switch. The ladder logic program always ends with the final rung.

SCADA Screen Development and Programming The SCADA Screen is developed in Factory Talk View (FTView) Local Site Edition Studio, version 10. There are three subprocesses and one pre-process of the proposed CIP system, which are shown on the SCADA screen along with the red and green colored animation facility provided in SCADA development software. The first window of Fig. 9 shows the status of the CIP system before the Start Switch is pressed, which is provided on the HMI Screen as well as on the hardware control panel. Currently, all the processes of CIP are in nonworking state; thus, all the flow lines, the subprocess indicators, pumps, valves, and level indicators show red lights. The second window of Fig. 9 shows the status of the CIP system just after starting the process. The Filling process of the Hot Water Vessel is started by starting the cold water pump (P_01). The pump takes the cold water from the Storage Tank

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

181

Fig. 9 SCADA screen status at the various stages of the proposed cleaning-in-place system

and passes it to the Hot Water Vessel through the green-colored line by opening the Inlet Valve of the Hot Water Vessel (IV_02). All other valves are currently red as they are in the closed condition and block water flow through them. The second Non Return Valve (NRV_02) placed at the outlet of the hot water pump (P_02) blocks the flow in the hot water pump and the outlet water line. The low-level switches of the hot water vessel (LSL_02_HV) and the main vessel (LSL_01_MV) are both green, which shows a low water level in both vessels. Currently, the hot water temperature reading given by the RTD sensor is zero as there is no water in the vessel and this is shown by the Temperature Transmitter tag box (TT_02). The middle right-hand side section shows that the Filling Process is currently going on. The third window of Fig. 9 shows the status of the CIP system just after starting the Pre-Cold Water Rinsing Cycle. The Filling process of the Main Vessel is initiated

182

A. S. Patil et al.

by starting the cold water pump (P_01). The pump takes the cold water from the Storage Tank and passes it to the Main Vessel through the green-colored pipe line by opening the Inlet Valve of the Main Vessel (IV_01). All other valves are currently red as they are in closed condition by blocking the water flow through them. The second Non Return Valve (NRV_02) placed at the outlet of the hot water pump (P_02) blocks the flow in the hot water pump and the outlet water line. Now the lowlevel switches of the Hot Water Vessel (LSL_02_HV) is red whereas the high-level switch (LSH_02_HV) is green, which shows that the Hot Water Vessel is fully filled with the water up to the high level where the high level sensor is placed. Similarly, the low-level switch (LSL_01_MV), as well as the high level switch (LSH_01_MV), are both red, which shows that the water level is in between the two level sensors. Currently, the hot water temperature reading given by the RTD sensor is 28 ◦ C, which is the current water temperature present inside the hot water vessel and is shown by the Temperature Transmitter tag box (TT_02). The middle right-hand side section shows that the Filling Process is complete and that currently the Pre-Cold Water Rinsing Cycle is going on. The fourth window of Fig. 9 shows the status of the CIP system just after starting the Rinsing Process of the Pre-Cold Water Rinsing Cycle. After completely filling the main vessel up to the upper-level sensor (LSH_01_MV), which is now showing the green light, the rinsing timer starts counting the time. The timer is set for 5 min, with 10-ms time resolution, which counts the 300,000 counts that is each count for 10 milliseconds. For this 5-min rinsing time, all the valves, pumps, and flow lines are red, as there is no any water flow through the system. The Hot Water Vessel is also completely filled up to the upper level sensor (LSH_02_HV), which is showing the green light. The heater, which is placed inside the Hot Water Vessel, is now working and increasing the water temperature gradually from room temperature to the set point temperature, whereas the water temperature is continuously read by the RTD sensor placed inside the Hot Water Vessel. The current hot water temperature is 40◦ C and is shown in the Temperature Transmitter tag box (TT_02). The middle right-hand side section shows that the Filling Process is complete and that currently the Pre-Cold Water Rinsing Cycle is going on. The fifth window of Fig. 9 shows the status of the CIP system just after starting the Draining Process of Pre-Cold Water Rinsing Cycle. This draining process of the Main Vessel is started when the 5-min rinsing process is completed. The outlet valve of the main vessel (OV_01) and the drain valve (DV_01) are currently open and so they are showing green. All the other valves are currently red, as they are in the closed condition and blocking the water flow through them. Currently, the hot water temperature reading given by the RTD sensor is 60 C, which is the current water temperature present inside the Hot Water Vessel and is shown by the Temperature Transmitter tag box (TT_02). The middle right-hand side section shows that the Filling Process is complete and that currently the Pre-Cold Water Rinsing Cycle is going on. The sixth window of Fig. 9 shows the status of the CIP system just after starting the Hot Water Rinsing Cycle. The Filling process of the hot water in the Main

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

183

Vessel is started by starting the hot water pump (P_02). The pump takes the hot water from the Hot Water Vessel and passes it to the Main Vessel through the greencolored line by opening the Inlet Valve of the Main Vessel (IV_01) and the Outlet Valve of the Hot Water Vessel (OV_02). All other valves are currently red, as they are in the closed condition and blocking the water flow through them. The first Non Return Valve (NRV_01) placed at the outlet of the cold water pump (P_01) blocks the flow in the cold water pump. The low-level switches of the Main Vessel (LSL_01_MV) are green, which shows the low water level of the vessels. Currently, the hot water temperature reading given by the RTD sensor is 80◦ C, which is equal to the Set Point and is shown by the Temperature Transmitter tag box (TT_02). As the Set Point Temperature and the RTD reading are the same, the heater is currently OFF, which prevents overheating or burning of the electric heater. The middle right-hand side section shows that the Filling Process and Pre-Cold Water Rinsing Cycle are complete and that currently the Hot Water Rinsing Cycle is going on. The seventh window of Fig. 9 shows the status of the CIP system just after starting the Rinsing Process of the Hot Water Rinsing Cycle. After completely filling the Main Vessel up to the upper-level sensor (LSH_01_MV), which is now showing the green light, the rinsing timer starts counting the time. The timer is set for 5 min with 10-ms time resolution, which counts the 300,000 counts that is each count for 10 milliseconds. For this 5-min rinsing time, all the valves, pumps, and flow lines are red, as there is no water flow through the system. The Hot Water Vessel is now empty; thus, the upper-level sensor (LSH_02_HV) is showing the red light and the lower-level sensor (LSL_02_HV) is showing the green one. The heater, which is placed inside the Hot Water Vessel is now in the OFF condition. Thus, the current hot water temperature shown by the Temperature Transmitter tag box (TT_02) is 0 ◦ C. The middle right-hand side section shows that the Filling Process and the PreCold Water Rinsing Cycle are complete and that currently the Hot Water Rinsing Cycle is going on. The eighth window of Fig. 9 shows the status of the CIP system just after starting the Recovery Process of the hot water under the Pre-Cold Water Rinsing Cycle. This Recovery Process of hot water is started when the 5-min. rinsing process is complete. The outlet valve of the main vessel (OV_01) and the inlet valve of the hot water vessel (IV_02) are currently open; thus, they are showing green color. The hot water pump (P_02) takes the water from the main vessel through the outlet line and refills the hot water tank, so it is also showing the green indicator light. All other valves, pumps, and flow lines are currently red, as they are in the closed condition and blocking the water flow through them. Currently, the hot water temperature reading given by the RTD sensor is 80 ◦ C, which is the current water temperature present inside the Hot Water Vessel and is shown by the Temperature Transmitter tag box (TT_02). The middle right-hand side section shows that the Refilling Process of the Hot Water Vessel is going on. As it is the refilling process, it is shown by a red indicator light for the normal filling process. The Pre-Cold Water Rinsing Cycle is already complete and the normal Hot Water Rinsing Cycle is going on.

184

A. S. Patil et al.

The ninth window of Fig. 9 shows the status of the CIP system just after starting the Post-Cold Water Rinsing Cycle. The Filling process of the Main Vessel is started by starting the cold water pump (P_01). The pump takes the cold water from the Storage Tank and passes it to the Main Vessel through the green-colored pipe line by opening the Inlet Valve of the Main Vessel (IV_01). All other valves, pumps, and flow lines are currently red, as they are in the closed condition and blocking the water flow through them. The second Non Return Valve (NRV_02) placed at the outlet of the hot water pump (P_02) blocks the flow in the hot water pump and the outlet water line. Now, the low-level switches of the Hot Water Vessel (LSL_02_HV) are red whereas the high-level switch (LSH_02_HV) is green, which shows that the Hot Water Vessel is completely filled with water up to the high level, where the high-level sensor is placed. Similarly, the low-level switch of the Main Vessel (LSL_01_MV) is green, which shows the low level of water in the main vessel. Currently, the hot water temperature reading given by the RTD sensor is 80 ◦ C, which is the current water temperature present inside the Hot Water Vessel and is shown by the Temperature Transmitter tag box (TT_02). The middle righthand side section shows that all the other processes are complete and that currently the Post-Cold Water Rinsing Cycle is going on. The tenth window of Fig. 9 shows the status of the CIP system just after starting the Rinsing Process of the Post-Cold Water Rinsing Cycle. After completely filling the main vessel up to the upper-level sensor (LSH_01_MV), which is now showing the green light, the rinsing timer starts counting the time. The timer is set for 5-min with 10-ms time resolution which counts the 3,00,000 counts that is each count for 10 ms. For this 5-min rinsing time, all the valves, pumps, and flow lines are red, as there is no water flow through the system. The middle right-hand side section shows that all the other processes are complete and that currently the Post-Cold Water Rinsing Cycle is going on. The eleventh window of Fig. 9 shows the status of the CIP system just after starting the Draining Process of the Post-Cold Water Rinsing Cycle. This draining process of the Main Vessel is started when the 5-min rinsing process is complete. The outlet valve of the main vessel (OV_01) and the drain valve (DV_01) are currently open and so they are showing green. All the other valves, pumps, and flow lines are currently red, as they are in the closed condition and blocking the water flow through them. The middle right-hand side section shows that all the other processes are complete and that currently the Post-Cold Water Rinsing Cycle is going on. The twelfth window of Fig. 9 shows the status of the CIP process after completion of the draining process of the Post-Cold Water Rinsing Cycle. The CIP Process Completion Pop-Up message comes up, which shows that the current CIP process and all its subprocesses are complete. This pop-up window further asks about restarting the process or closing the CIP process by authorizing the completion of the current CIP process. The recovered hot water is stored in the Hot Water Vessel, which shows the full-water level, as the high level (LSH_02_HV) indicator is green, whereas the main vessel is empty, with the green light shown by the lowlevel (LSL_01_MV) indicator. At this stage, all the valves, pumps, and flow lines are

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

185

red, as there is no water flow through the system. By authorizing the completion of the current CIP, the main vessel can be further used for the desired application. The middle right-hand side section shows that all the subprocesses of the CIP are done.

HMI Screens Development and Programming The HMI Screens are developed in Factory Talk View (FTView) Machine Edition Studio of 10th version. The current statuses of various vessels and subprocesses of CIP are shown on various HMI screens. The various animation facilities as well as the buttons are used to show the current status of the process. The secure log-in facility is also provided at the start of the HMI screen. The first window of Fig. 10 shows the very first window of the HMI. The upper side of the screen welcomes the operator of the HMI by showing the message “Welcome to the CIP Test Rig.” The bottom left-hand corner shows the current date and time, whereas the bottom right-hand corner shows the two buttons. The first button is for logging into the further control windows of the CIP process. As security is provided for the HMI controls, the login with the desired log-in ID and password is compulsory. After clicking on the Log-In button, the middle section of the HMI screen shows a Pop-Up window named “Log-In As . . . ,” showing the name of the person who has successfully entered his log-in ID and password. Currently, no user ID or password is entered so the Pop-Up screen shows the default name with the message “You are Logging in as DEFAULT.” Below this massage there are another two buttons, the first one, the Continue button, opens the next window screen if the log-in ID and password are correct and authorized. The second button is provided for logging in with another ID, which brings the user back to the LogIn ID and password entering window again. The Shutdown button provided at the bottom rightmost corner is for shutting down the HMI screen. The second window of Fig. 10 shows the Welcome screen of the HMI along with the Log-In Pop-Up screen, which appears when the operator clicks the Log-In button on the Welcome screen. The new Pop-Up window opens to enter the LogIn ID or User Name and the Password. After successfully entering both details, the operator presses the Log-In or Enter button to check the authorization of the entry. If the entries are incorrect, the operator can clear the data entry by clicking on Cancel or the Esc button provided just below the Log-In button on the Log-In Pop-Up. The third window of Fig. 10 shows the Welcome screen of the HMI along with the Log-In Pop-Up screen, which appears when the operator clicks the Log-In button on the Welcome screen. The Log-In ID or User Name is provided here as Ashish and the valid Password provided in the inbox below. After successfully entering both details, the operator presses the Log-In or Enter button to check the authorization of the entry. If the entries are incorrect, the operator can clear the data entry by clicking on Cancel or the Esc button provided just below the Log-In button on the Log-In Pop-Up. The fourth window of Fig. 10 shows the Welcome screen of the HMI. The “LogIn As . . . ” Pop-Up present at the middle section of the Welcome screen shows the

186

A. S. Patil et al.

Fig. 10 HMI screen showing the status of all subprocesses of cleaning-in-place

Log-In ID or the User Name, “Ashish” provided at the previous Log-In screen. This screen validates the entered Log-In details. Now the operator clicks on the Continue button to go to the next screen of the HMI. After the entry of valid log-in details when the operator clicks the Continue button of the “Log-In As” Pop-Up, the actual working screens of the CIP process open, as shown in the fifth window of Fig. 10. There are multiple screens that show the working of the CIP process in various aspects. These aspects are shown on the left-hand side of the screen as buttons. These buttons bring the operator into the screen of that aspect . There are four CIP process aspects: Overall System Layout, Hot Water Vessel, Main Vessel, and the Process Screen, and the last one is the LogOut button. The upper side of the screen shows the user name is Ashish along with the Welcome note. The lower side of the screen shows the four buttons: Restart CIP,

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

187

Go Back, Start CIP, and Stop CIP. The Restart CIP button resets the CIP process and starts it from the beginning by making contact with the restart button (Restart_CIP) in the logic. The Go Back button brings the user to the previous screen. The Start CIP button starts the CIP process by latching the start switch of the logic (SS_01). The Stop CIP button unlatches the Start button (SS_01) of the logic. The middle part of the screen shows the overall working of CIP along with the layout of the system. Like SCADA, the green light shows the work in progress, whereas the red light shows the nonworking state of the line or the equipment. The next screen of the HMI shows the current status of the Hot Water Vessel, as shown in the sixth window of Fig. 10. This screen opens when the operator clicks the Hot Water Vessel button on the left-hand side of the screen. As explained above, the lower part of the screen shows the four buttons and the upper part shows the user name of the logged-in operator along with the welcome note. The middle section shows the status of the Hot Water Vessel in tabular format with the instrument name, tag name, and current status of the respective instrument with the help of green and red lights. The temperature sensor (TT_02) also shows the current temperature of the hot water inside the Hot Water Vessel. Below the table two extra buttons are provided for this particular screen, which are Change Temperature Set Point and the Drain Hot Water Vessel. The Change Temperature Set Point button allows the user to set the new point for the hot water temperature, which is shown in further figures. The Drain Hot Water Vessel button drains the water inside the Hot Water Vessel by opening the outlet valve (OV_02) and the drainage valve (Drain_01). When the operator clicks the Change Temperature Set Point button in the above screen, the next screen opens along with the Pop-Up window of Change Temperature Set Point along with the message “Enter New Temperature Set Point” with a numeric input box and the two buttons at the lower part “Ok” and “Reset,” as shown in the seventh window of Fig. 10. All other buttons such as the left-hand side aspect buttons (Overall System Layout button, Hot Water Vessel button, Main Vessel button, and the Process Screen button, as well as the Log-Out button) whereas the four buttons in the lower part (Reset CIP button, Go Back button, Start CIP button, and Stop CIP button) are all the same as mentioned above , as well as the Welcome note, along with the logged-in user name, are also the same as per the previous screens. When the user clicks on the numeric input box the numeric keypad opens, as shown in the eighth window of Fig. 10. The user enters the new temperature set point number and clicks the enter button, which is shown at the lower rightmost corner with the enter arrow. The user can also clear the entered number by using the lowermost middle button shown by the back arrow. The Esc button is provided at the lowermost left corner of the keypad to escape from the keypad. After clicking on the enter button of the keypad, the entered number of the temperature set point is displayed in the numeric input box, as shown in the ninth window of Fig. 10. The second confirmation of the set point is provided in the Change Temperature Set Point Pop-Up window with the Ok and Reset buttons . The Ok button saves the entered number as a set point and brings the operator to the Hot Water Vessel screen. The Reset button provides a keypad to enter the new numeric

188

A. S. Patil et al.

valve of the set point. All five aspect buttons on the left-hand side, the processing buttons at the bottom, and the user name with Welcome note are all same as per the other screens. The next screen of the HMI shows the current status of the Main Vessel, as shown in the tenth window of Fig. 10. This screen opens when the operator clicks the Main Vessel button on the left-hand side of the screen. As explained above the lower part of the screen the four buttons are shown and the upper part shows the user name of the logged-in operator along with the Welcome note. The middle section shows the status of the Main vessel in tabular format, with instrument name, tag name, and current status of the respective instrument with the help of green and red lights. The next screen, as shown in the eleventh window of Fig. 10, opens when the operator clicks on the Process button present on the left-hand side of the screen. The middle section of the screen shows all the main and subprocesses of the CIP, along with their working status shown by using green and red lights. The rinsing time is also shown in the table and increases gradually up to the set time of the timer. All other buttons, such as the left-hand side aspect buttons (Overall System Layout button, Hot Water Vessel button , Main Vessel button, and the Process Screen button as well as the Log-Out button), as well as the four buttons in the lower part (Reset CIP button, Go Back button, Start CIP button and Stop CIP button), are all the same as mentioned above. The Welcome note and the logged-in user name are also the same as per the previous screens.

Real-Time Data Collection The real-time data of the developed CIP Test rig is stored on an Microsoft Excel sheet. All the inputs, outputs, and internal relays are mentioned in the table with their data type, name, description, tag names such as the ladder logic and SCADA systems and their real-time values. Table 3 shows the sheet of a sample of real-time data gathered on 23 August 2020 at 10:35 AM.

Program Validation on Hardware The proposed system is tested and validated on the hardware training kit. The sensor inputs and switches are mapped to the on/off switches on the hardware. AllenBradley’s Compact Logix PLC is used as the controller where the ladder logic of the proposed CIP system is downloaded. The outputs are mapped with the hardware relays and are shown by the LEDs. The ferrule tags on the input–output wires are provided to map the wires easily. Also, the printed tags are provided on the switches and relays for easy mapping of the inputs and outputs. One main switch is also provided on the hardware panel to cut off the main supply in an emergency for safety purposes. One AC to DC power supply is used to convert the 240 V AC supply to the 24 V DC supply. This 24 V DC is provided to the PLC and the other

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

189

Table 3 Sample of real-time data collection Date Sr. no. 1 2 3

23-08-20 Data type Input

Name Switches

4 5

Sensors

6 7 8 9 10 11 12

Output

Pumps Valves

13 14 15 16 17 18

Internal Relay

Timer Counter

Description Start Switch Restart CIP Process Authorize CIP Process Set Temperature Set-Point Limit Switch High for Main Vessel Limit Switch High for Hot Water Vessel Limit Switch Low for Main Vessel Limit Switch Low for Hot Water Vessel RTD Temperature Sensor Reading Inlet Pump Recovery Pump Inlet Valve for Main Vessel Inlet Valve for Hot Water Vessel Outlet Valve for Main Vessel Outlet Valve for Hot Water Vessel Drain Valve Rinsing Timer Process Counter

Time Tag name SS_01 Restart_CIP Authorize

10:35:00 AM Real-time data 1 0 0

Temp_Set

0

LSH_01_MV

1

LSH_02_HV

1

LSL_01_MV

0

LSL_02_HV

0

TT_01

0

P_01 P_02 IV_01

0 0 0

IV_02

0

OV_01

1

OV_02

0

DV_01 TIMER_01 COUNT_01

1 301274 1

hardware components of the system. The overall validation system is as shown in Fig. 11. The tag data are collected with respect to the time of program simulation on the software, as shown in Fig. 12. The various tag data are mapped on the graph with various colored lines. The x-axis of the graph shows the time whereas y-axis shows the tag values. As all the tag values are in digital form, the tag line moves from the zero and one value lines as shown on the graph. The yellow line shown on the graphs is a tag data line of the process counter accumulator; thus, its output varies from zero to three. Like Fig. 12, the same timing diagram is developed for the hardware simulation as shown in Fig. 13. This graph also shows the various tag data, which are mapped on the graph with the various colored lines. The x-axis of the graph shows the time

190

A. S. Patil et al.

Fig. 11 Overall validation system Fig. 12 Timing diagram of software simulation data

whereas the y-axis shows the tag values. As all the tags values are in digital form, the tag line moves from the zero and one value lines as shown on the graph. The yellow line shown on the graphs is a tag data line of the process counter accumulator; thus, its output varies from zero to three.

Inference • Both timing diagrams, showing the simulation data of the software (Fig. 12) and hardware (Fig. 13), has a time lag of overall 3 s in the hardware system in comparison with the software. Therefore, the overall error in the hardware system is 5.56%. • The proposed automated CIP system has multiple advantages over the various manual cleaning methods in respect of an increase in safety, cleaning quality,

8 Design, Control, and Data Management for Cleaning-in-Place (CIP). . .

191

Fig. 13 Timing diagram of hardware testing data

efficiency, reduction in production down time, cleaning costs, water consumption, and labor requirements. The proposed system also helps in the automatic logging of real-time data, which is not possible in manual systems. • Hence, the proposed system will help to train newly joined employees, college students, and interns in the industrial automation software and hardware field. It also helps in various program validations on the hardware, so that the various errors and data logging can be easily mapped.

Conclusion The generalized three-stage Cleaning-In-Place process test rig has been successfully designed, developed, and automated using PLC programming, data acquisition is performed using the SCADA system and is controlled by an HMI as well as by the hardware switch panel. Also, the data management is carried out in an MS Excel sheet, whereas the system is tested and demonstrated by using the software simulation in Factory TalkView (FTView) Studio. The test rig would be useful for the three-stage Cleaning-in-Place mechanism of the dairy, food processing, beverage, pharmaceutical, and chemical industries.

Future Scope The Internet of Things can be applied in the proposed CIP system in further work. This system will help to gather real-time and historical data in the desired format, which may be in the form of tables or graphs. These data can also be gathered irrespective of user location via the internet facility on any mobile or desktop device by providing a secure authorization log-in to ensure the safety of the data.

192

A. S. Patil et al.

References Alvarez N, Daufin G, Guiziou G (2010) Recommendations for rationalizing cleaning-in-place in the dairy industry: Case study of an ultra-high temperature heat exchanger. American Dairy Science Association. J Dairy Scie 93:808–821 Amitha T, Sathian CT (2014) Cleaning-In-Place (CIP) system in Dairy Plant - Review. IOSR J Environ Sci Toxicol Food Technol (IOSR-JESTFT) 8(6):41–44 Bhandari VB (2017) Design of machine elements, 4th edn. McGraw Hill Education, pp 758–784 http://vt.baumer.com/fileadmin/user_upload/intemational/Services/Download/Datenblaetter/PI/B1_ Electronic_Temperature/BTIN/Baumer_R1_DS_1208.pdf http://vt.baumer.com/fileadmin/user_upload/international/Services/Download/Datenblaetter/PI/E1E2_Level_Switches/BTIN/Baumer_LA_DS_1208.pdf http://www.kirloskarpumps.com/product-pump-monobloc-domestic-monobloc-pumps-jalraaj.aspx http://www1.ap.dell.com/content/products/productdetails.aspx/dell-s2218hmonitor?c=in&l=en&cs =indhs1 https://ab.rockwellautomation.com/Programmable-Controllers/CompactLogix-1200#documenta tion https://www.alfalaval.in/products/fluid-handling/tank-cleaning-equipment/rotary-spray-heads/sani/ https://www.alfalaval.in/products/fluid-handling/valves/control-check-valves/lkc-non-return-valve/ https://www.delvalflow.com/media/uploads/brochures/1580222485-DelVal-Series-65-68-&-69-72Industrial-Process-Floating-Ball-Valves-2020.pdf https://www.ibef.org/exports/chemical-industry-india.aspx https://www.indianmirror.com/indian-industries/dairy.html https://www.indianmirror.com/indian-industries/food.html https://www.indianmirror.com/indian-industries/pharmaceutical.html https://www.rockwellautomation.com/en-us/products/hardware/allenbradley/human-machine-inter face/graphic-terminals.html https://www.rockwellautomation.com/global/products/factorytalk/overview.page?pageti9le=Factor yTalk-View-Site-Edition&docid=fbfaf09e608b1c3a74b3d7d0f95bd25e https://www.rockwellautomation.com/global/products/factorytalk/overview.page?pagetitle=Factory Talk-View-MachineEdition&docid=e8ecadac26887be557f0d580213673f3 https://www.rockwellautomation.com/global/products/factorytalk/overview.page?pagetitle=Studio5000-Logix-Designer&docid=3803f72409175761e586f6107b93c3d9 Indian Standard (1998) Code for unfired pressure vessels (IS: 2825–1969), 8th edn. Bureau of Indian Standards, pp 5–62 Nighot KB, Kunjir RA, Wagh NP, Barure ON (2015) Automation of Cleaning-In-Place for Food and Beverage Industry Using PLC and SCADA. Int J Adv Res Electr Electron Instrum Eng 5(4):3140–3146 Nurgin M, Slevica V M, Milan M, Jelena B, Dragutin D (2015) CIP cleaning process in Dairy indusry. International 58th Meat Industry Conference, Elsevier, vol 5, p 184–186 Raj PG, Arya RKK, Joshi T, Bishat D (2018) A Review on Cleaning Validation in Pharmaceutical Industry. Int J Drug Deliv Ther 8(3):138–146 Swapnil K, Sangitrao RS, Shubham K, Gajanan I, Shubham M (2017) Automated Cleaning in Dairy Industry using CIP Method. Int J Innov Sci Res Technol 2(4):274–277 Vaishnavi D, Amruta D, Dalavi DV (2015) Cleaning-In-Place in Pharmaceutical Industry using PLC and SCADA Software. Int J Adv Res Sci Eng 4(3):257–265

9

Control and Informatics for Demand Response and Renewables Integration Michael Short

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Grid Communications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Networked Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Models and Technical Underpinning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fault-Tolerant Playback Buffer Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Supervisory Controller Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Implementation, Validation, and Illustrative Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Illustrative Example: Model Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Illustrative Example: Statistical Delay Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Illustrative Example: HIL Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

194 197 197 199 200 200 202 204 205 206 208 208 210 211

Abstract

Following the lead of the UK, many governments worldwide have now made commitments to achieve net-zero carbon emissions by 2050. This drive toward a zero-carbon economy requires many technological and social innovations, including advancement and accelerated deployment of industrial digitalization and smart grid concepts. Within the framework of the second digital revolution known as “Industry 4.0,” the smart grid is an augmented energy distribution network that enables real-time communications for monitoring, control, and

M. Short () School of Computing, Engineering and Digital Technologies, Teesside University, Middlesbrough, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_152

193

194

M. Short

protection of energy assets as well as facilitating the physical transfer of energy. Smart grid and energy storage provide a solution for the integration of renewable energy assets, which due to their inherent volatility when compared to traditional forms of generation require careful planning and operation; the coordinated, end-to-end economic dispatch and optimization of generation, storage, and demand assets is required along the entire energy value chain. In this chapter, fundamental and applied research work related to control and informatics for supervisory control of energy-related assets in IoT-based energy management systems (EMSs) are described. Key concepts, design procedures, results, and analysis are presented, along with a summary of implementation activities in the context of funded research and innovation projects. Keywords

Industry 4.0 · Smart energy · Net-zero · Energy management systems · Renewables integration

Introduction In June 2019, the UK government amended the Climate Change Act to commit to 100% greenhouse gas (GHG) emissions reductions – or full net-zero – by 2050 (Energy Catapult 2020). Following the lead of the UK (https://www.gov.uk/government/news/uk-becomes-first-major-economy-to-pass-net-zero-emissions-law), many governments worldwide have now made similar commitments to achieve net-zero emissions by 2050. This commitment by the UK marks an increasing commitment to climate change, starting with agreement to the first Kyoto Protocol in 1997. A timeline of major milestones from 1997 until the current time are shown chronologically in Fig. 1. This drive toward a zero-carbon economy requires

Fig. 1 Timeline of major milestones leading to UK net-zero commitment

9 Control and Informatics for Demand Response and Renewables Integration

195

many technological and social innovations, including advancement of digitalization and accelerated deployment of smart grid concepts. With improved connectivity combined with recent advances in data science, many industries are currently on the verge of a second digital revolution known as “Industry 4.0” (Isaksson et al. 2018). The smart grid lies within this framework as an augmented energy distribution network which enables real-time communications for monitoring, control, and protection of energy equipment as well as the physical transfer of energy (Masters 2004; Ekanayake et al. 2012). Principally for economic and safety reasons, electricity was historically generated by large centralized fossilfueled generators and subsequently transported to consumers via unidirectional transmission and distribution networks. More recent times have seen the emergence of small- and medium-scaled decentralized generation equipment (typically driven by renewable energy sources such as solar PV, wind turbines, and biomass combined heat and power plants (CHP) embedded within the transmission and distribution networks themselves (Energy Catapult 2020; Masters 2004; Ekanayake et al. 2012). Renewable energy sources require careful integration onto the electricity grid, and the network requires careful planning and operation. This is due to their inherent volatility when compared to traditional forms of generation, and the coordinated, end-to-end dispatch of generation, storage, and demand assets is required in order to balance the power system (Energy Catapult 2020; Masters 2004; Ekanayake et al. 2012). In this chapter, previous research work related to control and informatics for demand response and renewables integration which has been carried out at Teesside University in the UK is described. Specifically, the focus of this chapter lies in the application of Industry 4.0 concepts for the integration of semi-controllable (e.g., wind turbines) and fully controllable (e.g., biomass CHP) renewable generation assets, coupled with the use of semi-controllable and fully controllable consumption assets (e.g., HVAC and/or refrigeration systems within buildings) to provide aggregated energy shifting services and demand response (DR) purposes. The rationale for this is twofold. Firstly, wind, solar, and biomass energy have been identified as a key source of renewable energy to be deployed in the UK drive toward net-zero (Energy Catapult 2020; Masters 2004; Ekanayake et al. 2012). Secondly, building stock is estimated to account for around 40% of global energy consumption, exceeding that of other major sectors like industry and transportation (Perez-Lombard et al. 2008; Short et al. 2019). As such, buildings represent the sector with both the largest energy saving and energy shifting potential for DR applications; the decarbonization of heat is critical in the UK context, and will rely heavily on the use of residential electric heat pumps (Energy Catapult 2020) in additional to commercial HVAC. Figure 2 shows a flowchart of the electricity pathways in the calendar year 2019 in the UK, indicating the mix of supply and demand (Department of Business, Energy and Industrial Strategy 2020). A large portion of usage by domestic and commercial users is attributed to HVAC, and generates between 0.0 kg and 0.4 kg of CO2 emissions per kWh depending upon time of use (TOU). This dependency upon TOU is due to the mix of

196

M. Short

Fig. 2 Flowchart of UK electricity pathways in the calendar year 2019

supply sources and the availability of renewable or zero emission generation. The objectives are clearly twofold: firstly, to increase the proportion of the generation mix which comes from renewable energy in order to reduce reliance upon traditional forms of generation, which produce GHGs; and secondly, to shift controllable demand onto time periods in which renewable energy is available. Clearly, the first objective creates volatility due to reliance upon external factors such as weather conditions, creating the need for the second, to balance the need for this increased volatility. Figure 3 shows an example situation of how implicit general demand management, using building HVAC as an example, can be used to help “flatten” a demand curve and reduce the need for GHG-emitting peaking units and support renewable integration. In addition, when adverse events occur (e.g., supply shortfall due to grid issue or under-forecast of wind), this can be extended to explicit specific demand response, when one or more assets are surgically targeted to shift demand to help overcome the adverse event with wider problems such as line trips occurring. In the chapter, tools and techniques which have found to be useful from a control and informatics framework to implement such asset management are described. Specifically, an overview is given of design principles for supervisory controllers in IoT-based energy management systems. The remainder of the chapter is organized as follows. Section “Related Work” presents a review of previous and related work around smart grid informatics and control aspects. Section “Models and Technical Underpinning” presents the models and techniques for integration and management of energy assets. Section “Implementation, Validation, and Illustrative Example” describes and summarizes hardware-in-the-loop (HIL)-based experimental results to illustrate key concepts in a design example, and also discusses wider work on industrial implementation in the UK. The chapter is concluded in Sect. “Summary and Conclusions”.

9 Control and Informatics for Demand Response and Renewables Integration

197

Fig. 3 Employing building stock HVAC for supply following, demand response, and renewables integration

Related Work Smart Grid Communications Historically electricity generation was typically achieved using several large generating stations operating in parallel, and with transmission interconnections possibly spanning several countries for electrical grids; these generators and interconnects were principally under the supervisory control of only a small number of public and private bodies (Masters 2004; Bhowmik et al. 2004). Dispatch instructions for individual or groups of generators (and emergency curtailment commands for large industrial consumers) were normally relayed over basic communications media, and would traditionally consist of dedicated duplex channels between a generation utility and the independent system operator (ISO) for telemetry and telecontrol, with additional voice/data channels over the public switched telephone network (PSTN) as a backup (Elgerd 1981). ICT interfaces enabling real-time information exchange related to the scheduling, monitoring, control, and protection of the interconnected energy assets is critical to the concept of the smart grid. It is generally agreed that Internet Protocol (IP)based connectivity and Industrial Internet-of-Things (IIoT) will facilitate smart grid applications, although alternate proposals have been made (Deshpande et al. 2011; Hopkinson et al. 2009). An emerging concept is that a “utilities intranet” – a data network which is common to the utilities but isolated and protected from the general Internet – provides the backbone for such an infrastructure. The utilities intranet is envisioned to provide for the eventual connection of all regional substations, equipment, assets, and control centers throughout the smart grid (Deshpande et al. 2011; Hopkinson et al. 2009). The utilities intranet, and indeed smart gird traffic in general, has a varying range of quality of service (QoS) and routing requirements, with stricter dependability requirements when compared to regular Internet traffic (Khan and Khan 2013). To handle such dependability requirements,

198

M. Short

differentiated services (DIFFSERV), traffic prioritization, and enhanced multicast UDP/IP concepts (which can be offered by IPv6) need to be leveraged (Deshpande et al. 2011; Hopkinson et al. 2009; Falk et al. 2013; Short and Dawood 2014; Dawood and Short 2014). Unfortunately, end-to-end QoS management for widearea critical streams spanning multiple control areas is still someway off becoming a practical reality. Nevertheless, experimental prototyping work has shown that the IEEE standards on time-sensitive networking (TSN) which are currently under development may provide a suitable framework for carrying critical traffic over wide area networks, including those related to the IEC 61850 standards to be described below (Short et al. 2016). However, the required standards are still in development and TSN technology is not yet fully functional, and there is not yet widespread implementation. Communications with less demanding QoS and dependability requirements are typically supervisory control and data acquisition (SCADA) and/or machine-tomachine (M2M) services. In such cases remote data exchanges using standard UDP/IPor TCP/IP sockets may be employed. Messages may also have to be passed across network elements which are not directly part of the utility Internet, for example, at the end most points of the grid may traverse one or more digital subscriber links (DSL). Care must be taken when such IP-based remote data exchanges have been employed to implement SCADA applications, for example, in load/frequency control and bilateral load following, as packet latency (delay), variability in packet delay (jitter), and packet losses can heavily influence both performance and stability of the closed loop (Bhowmik et al. 2004; Dawood and Short 2014). The impact of such degraded control for an energy management utility is principally an inability to meet contractual power requirements and associated economic penalties. For the grid and gird operator as a whole, in the worst case it may result not only in economic issues, but with security issues caused by the tripping of frequency protection devices and unintentional islanding. The most critical aspects of the smart grid communications infrastructure are related to tele-protection and telecontrol applications and feature very high availability requirements combined with ultralow latency requirements (Deshpande et al. 2011; Khan and Khan 2013; Falk et al. 2013; Short and Dawood 2014). For applications such as situational awareness using phase measurement units (PMUs), allowable worst-case message latencies for PMU data packets are around 40 milliseconds in a 50 Hz electrical system. In tele-protection applications, message latencies less than 20 milliseconds are required (and in some cases less than 3 milliseconds) can be mandated (Deshpande et al. 2011; Khan and Khan 2013). In such cases, the IEC 61850 standard has been proposed to provide methods to help carry and route this more time-critical traffic. The IEC 61850 standard family was originally created to support modular, low-latency communications services in substations and runs directly on switched Ethernet networks (International Electrotechnical Commission (IEC) 2002; Brand et al. 2003). Traffic classes in IEC 61850 include events (GOOSE messages) and sampled data (SMV messages), which are directly mapped into raw Ethernet frames using ASN.1:BER (basic encoding rules). IEEE 802.1Q (priority tagging/VLAN) support at layer 2 is

9 Control and Informatics for Demand Response and Renewables Integration

199

employed to achieve the required response times and multicast routing. Typically, industrial switches with support for all protocol aspects and rugged design for the substation environment are required. Additional traffic classes are also defined, typically for SCADA and M2M services, and generally feature lower availability and latency requirements. Implementations in practice favor mapping to protocols such as the manufacturing messaging specification (MMS) using standard UDP/IP or TCP/IP directly within the IEC61850 framework, or independently using a dedicated IP-based SCADA protocol such as IEC 60870–5-104.

Networked Control Systems Networked control systems do not just appear in energy management and smart grid applications, but also in areas such as process control and automotive. Research investigating the links between network performance and its impact on distributed control systems is several decades old, and research trends mainly include reviews on latency, delays, and security of networks. The effects of the network delay (in particular modeling and analysis, and also delay compensation) and the key roles of bandwidth allocation and packet scheduling for fault-tolerant control applications are often discussed, as well as cybersecurity issues and network component integration and/or reconfiguration (Gupta and Chow 2010). All digital control implementations are susceptible to timing jitter of some sort or another (e.g., clock inaccuracy or drift and task scheduling effects). However, networked and/or distributed control implementations are especially susceptible due to queuing and routing impacts of packet transfers. With the advent and increasing adoption of (I)IoT concepts, modern networked control applications mostly utilize packet switching communication mechanisms and protocols since they have many wellknown advantages and economic benefits (for example, scalability, stability for data transfers over short and long distances, reconfigurable routing, etc.) in comparison with circuit-switching equivalents. However, the former faces problems related to packet-dropout, network-induced delays, and packet disordering that encourages further investigations (Singh et al. 2014). Performance analysis for smart grids has generally followed this direction, with recent focus upon system stability (Zhang et al. 2001). In this work, a networked control scheme and a stability analysis framework to control systems with inherent oscillatory dynamics such as those present in power systems is presented. The allowable data-dropout limit to guarantee a particular closed-loop control performance is explicitly quantified, and requirements for the communication infrastructure and protocol for smart grids are suggested. Research focus has also been put upon playback buffering mechanisms for use in networked control environments, as described in (Liberatore 2006; Short et al. 2015a). Playback buffers can be used to ameliorate (or even in some cases almost totally eliminate) network jitter effects, at the expense of increased latency. This is advantageous for most types of process plant under networked closed-loop control, since jitter is comparably more problematic than an equivalent fixed delay (Cervin et al. 2004). Playback buffer

200

M. Short

parameters must be carefully selected for a given application, and taken into account in any closed-loop analysis, as delays introduce control-theoretic effects such as unbounded negative phase shift at high frequencies (non-minimum phase frequency response) into the plant. Such effects can be effectively engineered out of the control loop using a well-designed predictive controller, as shown in (Short et al. 2015b). As can be deduced from the previous discussions, until such time that there is widespread adoption of IPv6 and TSN for smart grid applications, the use of raw UDP/IP or TCP/IP sockets seems unavoidable for networking applications outside of substations at the current time (and also short- to medium-term future). Therefore, methods which both enhance their basic operation and can be compensated for in the associated control design seem mandated. Such methods are employed within the presented EMS framework in the next section.

Models and Technical Underpinning General Configuration Consider the general arrangement for a cloud-based IoT energy management system (EMS) as shown in Fig. 4 below. A number of energy-related assets (controllable supply and demand units, e.g., wind turbines and HVAC) are to be teleoperated/telecontrolled by an aggregator (EMS) which also interacts with the wider electricity market. The EMS carries out numerous functions including baseline supply/demand predictions, spot/balancing market contracts, bilateral contracts, billing, and ancillary services such as demand response. To implement contracted

Fig. 4 Cloud-based energy management system (EMS) under consideration

9 Control and Informatics for Demand Response and Renewables Integration

201

Fig. 5 Networked control system (NCS) architecture under consideration

power delivery or consumption requirements, the operator uses telecontrol/teleoperation via IP links to edge devices located in the vicinity of the energy asset, interacting with (and in some cases implementing) a local control system. For example, a local control for a wind turbine principally acts to control blade pitch angles to regulate output power, while a HVAC control unit principally controls power delivery to heat exchangers to regulate zone temperature. In this chapter, the interest is in the design of the supervisory controllers and delay/jitter management on the IP links. Next, consider a single-loop control system constructed as a part of such an EMS as depicted in Fig. 5, with the main EMS components being the trajectory generator and feedback controller D(z). Let the supply/demand asset under supervisory control, and its local regulatory/tracking controller be well approximated around the target operating regions by the discrete-time linear transfer function G(z). The EMS computes a value to apply to the local asset set point (the main manipulated variable u(k) in the outer loop), using the controller D(z) acting on the loop error e(k), and transmits a sampled data representation of this variable to the asset every T seconds via the network connection. The asset and its local feedback controls are driven by this command signal and may also experience disturbances. Either in the same (or another) asset, a measurement is taken every T seconds to produce a sampled data representation of the controlled variable y(k) every T seconds, which is then transmitted back to the EMS via the network connection. The controlled variable is subtracted from the reference signal r(k) emitted from the trajectory generator to produce the error. The situation described above reflects many possible configurations for supply and demand asset control including regulation of contracted power delivery or consumption, bilateral contracts for supply or demand following, load/frequency control, and demand response and demand turnup. In each case, sampled data information is transmitted over the IP network in the forward path and also the feedback path, in most cases over logically separate communication pathways. Key to the successful achievement of some of the objectives (e.g., demand response) is optimization of the set point trajectory, taking into account (for example) weather forecasts to exploit preheating or precooling effects (e.g., see (Isaksson et al. 2018; Ekanayake et al. 2012; Short et al. 2019) and the references therein). In some cases, for example, bilateral following contracts, it will be beneficial to couple two or more

202

M. Short

supervisory control loops at a higher level. Such configurations and optimizations are secondary functions of the EMS and are not considered explicitly in this chapter.

Fault-Tolerant Playback Buffer Design Buffer Operation A simple but effective fault-tolerant playback buffering scheme for transmission and reception of a sampled data sequence denoted x(k) via a UDP/IP link has previously been developed (Short et al. 2015a). Under the control of a synchronized clock, at discrete-time step k in the transmitter node, packet k containing the kth sample of the signal x(k) is packed into the data payload of a UDP/IP frame, along with the sequence count k, and appended with a short time history of previous signal values x(k-1),x(k-2), . . . , x(k-N + 1), where N is the (fixed) size of the receiver buffer. An encoding format such as ASN.1: BER is applied to pack the data in the UDP/IP frame, and encryption such as AES 256 applied for security; specific details not discussed further in this chapter. The frame is then transmitted via Internet to the receiver. Let the clocks in both transmitter and receiver be synchronized to a level of accuracy +/− ε. For the types of supervisory control systems this chapter is concerned with typically featuring sample times of the order of seconds, the standard Network Time Protocol (NTP) can achieve an accuracy level of a few milliseconds, which should be sufficient in most cases. Let the receiver node have a buffer of size N samples, and due to this clock synchronization be running in approximate time step with the transmitter (within accuracy +/− ε < k-N must have been received, and hence the sample x(k-N) must exist due to the information redundancy employed by the packetizing scheme in the transmitter. The motivation for the above buffering scheme, which is displayed schematically in Fig. 6, is that latency of transmission is fixed at N sample times, with jitter fixed to accuracy 2ε( qσ ) ≤

q2 + k − 1 √ q 2 k2 + k

(1)

This inequality is effectively a finite-sample version of Chebyshev’s inequality and, although conservative, gives meaningful results for test data obtained over practical timescales (e.g., measured in hours or days) and for confidence probabilities that are not ultrahigh (e.g., in the region of 0.9999). It may be used to determine the probability that packet latency μ + qσ will not be exceeded at run time, and with knowledge of the sample time can be used to specify the required size of a playback buffer using the relationship N = (μ + qσ)/T for value of k satisfying the desired probability level.

204

M. Short

Supervisory Controller Design Suppose that the identified process model is given by the discrete-time pulse transfer function G(z) = B(z)/A(z), where z = e(α + jω)Ts is the z-transform shift operator. Assume that the PTF G(z) is formed from the continuous model of the process G(s) and controller C(s) (if present) via z-transforms with sample time T, using standard means to incorporate any DAC and ADC. Feedforward and feedback delays due to buffering are assumed incorporated into the numerator polynomial B(z), and appear as shifts z-N for a buffer of size N. A method proposed in (Short et al. 2015b) allows the synthesis of a predictive digital controller D(z) directly from the desired closedloop pole specifications encoded in a design polynomial P(z). Assume that B(z), the numerator of the PTF (including the time delay and any zeros), is to be scaled and embedded in the closed-loop transfer function as follows: kp B(z) R(z) = Y (z) P (z)

(2)

where the scalar kp = P(1)/B(1) to ensure unit gain in steady state. Numerous advantages arise from embedding the open-loop zeros of the process into the closed-loop response, including the ability to predictively handle time delays embedded within polynomial B(z), and robustness against inverse response (unstable zeros). Then by rearranging the closed-loop characteristic equation for the required controller D(z), the following controller design achieves the specification: D(z) =

kp A(z) P (z) − kp B(z)

(3)

The relation above was obtained by rearranging the characteristic equation of a process under unity negative feedback, with the scaling gain kp ensuring that the resulting controller contains an integrator (Short et al. 2015b). In order to select the sample time T and design specification P(z), some simple guidelines are as follows. Concerning first P(z), should the (continuous) response be wished to approximate a first-order lag type response 1 + τ s, with time constant τ , the specification 1 + pz−1 can be used, with p = −exp(−T/τ ). For desired setting time ts , τ should be chosen as ts /4.6. For a second-order polynomial of the form s2 + 2ζ ωn s + ωn 2 , with natural frequency ωn and damping ratio ζ , then the specification 1 + p1 z−1 + p2 z−2 can be used with mapping:      p1 = −2 exp −T ζ ωn cos T ωn 1 − ζ 2

(4)

  p2 = exp −2T ζ ωn

(5)

For desired setting time ts and approximate phase margin φ (in radians), ωn and ζ should be chosen as:

9 Control and Informatics for Demand Response and Renewables Integration

sin (ϕ) ζ =√ 4 cos (ϕ) ωn =

205

(6)

4.6 ζ ts

(7)

Selection of sample time T can then proceed by choosing a value satisfying inequality (8), with tro the open-loop risetime of process G(s) and trc the approximate risetime of the closed-loop system, obtained from the design specification as shown in (9):  T ≤ min trc =

tro trc , 10 10

2.23τ 2.23ζ 2 −0.078ζ +1.12 ωn

 (8)

:

First Order

:

Second Order

(9)

Thus, from knowledge of the gross statistical properties of the IP links and dynamics of the local process model – both of which may be determined experimentally – combined with basic specifications for the desired closed-loop response, a full predictive supervisory digital controller may be easily designed using the procedure and relations presented in this section. A simple software library (employing only UDP/IP and NTP elements of the IP stack) provides the informatics framework to help ameliorate the effects of IP packet loss and give jitter compensation, fixing the delay to a known, stable value which may be compensated for directly in the predictive controller synthesis equation. Although the method is principally aimed at cloud-based energy management systems, it is clear that it may be extended to other applications involving tele-operation and telecontrol as required.

Implementation, Validation, and Illustrative Example Research staff at Teesside University in the UK have been developing ICT tools for smart energy management systems for smart grid applications since 2008, and the techniques presented and described in this chapter have been implemented as components in a number of industrial innovation projects both in the UK and abroad. This research has also underpinned the development of new services and products by two companies, a formal technology transfer and educational partnership which is the first of kind in UK, and it has also informed policy design at the EU level (referred to in two policy documents). Specifically, energy management system tools developed in the IDEAS (https://cordis.europa.eu/project/ id/600071) EU project have enabled increased economic profits and reduced GHG emissions from more efficient use of a typical combined heat and power (CHP) plants, of between A C53,000 and A C87,000 per annum for a typical 40 MW plant in the EU. The tools for supply-side energy management developed in that period

206

M. Short

have directly contributed to new commercial product developments, increased economic profits, and contributed to enabling over 4.2 homes in the UK to be powered by offshore wind at the time of writing (https://teesbusiness.co.uk/2020/ 10/30/pioneering-project-could-help-government-meet-wind-power-pledge/). The application of the tools for energy management in the DR-BoB (https://cordis. europa.eu/project/id/696114) EU project have resulted in total savings of 1.67 GWh of total primary energy across four project demonstration sites in 2018, which is equivalent to a reduction of 5.87 tons of CO2 emissions. In terms of economic costs savings, a total benefit of approximately A C13,000 was recorded at these sites (with corresponding large GHG emission reductions) by shifting energy use onto times of renewable availability. An illustrative example using HIL simulation is presented below (reproduced in part from (Short et al. 2015a)), to highlight the design and implementation of the specific techniques described in Sect. “Models and Technical Underpinning” of this chapter.

Illustrative Example: Model Configuration In order to test the elements of the proposed control and communications mechanism, a realistic simulation model of a Biomass CHP plant with steam turbo generator in a microgrid environment has been developed within the framework of a flexible HIL simulator for smart grid systems (Short and Dawood 2014; Dawood and Short 2014). The simulation model incorporates communication link components with configurable fixed packet latency, variable packet latency, and packet error rate. The importance of such simulation-based testing for smart grid applications, including the testing and evaluation of dependable communications, has previously been highlighted (Podmore and Robinson 2010). The model has been developed using Matlab©/Simulink©. Although dedicated simulation tools (such as OMNET) are also useful from the perspective of fine-grained communication network simulation, performance and stability analyses can oftentimes make use of the higher-level statistical impacts of the network when determining its effects on control loops. Therefore, the use of course-grained statistical models of the networkinduced delay and jitter are justified and employed in this example (Bhowmik et al. 2004). The Simulink block diagram of the configuration is shown in Fig. 7. The model features a per unit (p.u.) single equivalent lumped generator incorporating a coordinated boiler/turbine control system as the electrical generation side of the CHP asset (Ordys et al. 1994). The input to the asset is the reference mechanical power output; the main output is the actual mechanical power output. Although a real generator has multiple inputs/outputs and is nonlinear over much of its operating range, the local coordinated control system ensures an offset-free, nearlinear input/output relationship by coordinating and trimming the main inputs (turbine valve setting, furnace biomass fuel/air input) (Ordys et al. 1994). As such the main observable dynamics are that of the turbine itself on the electrical side,

9 Control and Informatics for Demand Response and Renewables Integration

207

Fig. 7 Load/frequency and power exchange microgrid simulation model

and in the simulation, a representative dynamic model for a two-stage turbine with reheater is employed (Ordys et al. 1994). The CHP asset is assumed located within a microgrid (non-islanded mode) and is used – along with a scheduled power exchange (import or export) via a tie-line to a larger grid – to supply power demand in the microgrid local area. The selected units were 1 p.u. power = 1 GW and 1 p.u. frequency = 50 Hz. The area configuration was set using typical representative parameters: The load was assumed to have a power dependency on frequency of 200 MW per Hz, with inertia of the equivalent generator of 6 seconds. The tie-line synchronizing coefficient (dependent upon the reactance of the line) was taken to be 2π/15. When an increase in the local area load occurs (or equivalently a generation shortfall), there is an instantaneous mismatch between power supply and demand and kinetic energy is borrowed from the turbine. This causes the rotational speed of the shaft (and hence the area supply frequency) to dip. Extra power therefore needs to be injected to restore the power balance to a steady state, and eliminate the undesired frequency transient and offset. A proportional speed governor on the connected CHP generator is assumed present (such as is mandated to be employed in UK power systems). The governor proportionally increases/decreases the output power generation in response to frequency decreases/increases; this loop may be seen in the simulation model (speed governor). Although this is sufficient in most cases to keep the frequency within working limits, some frequency offset (known as the “droop”) remains due to the lack of integral action (which cannot be directly employed in parallel connected power systems). A 5% droop between no load and full load, which is typically used in industry, was used in our model by setting the governor gain appropriately. Any drift in frequency away from the nominal value due to droop leads to phase shifts between the local frequency and that of the main grid connected by the tie-line

208

M. Short

occurring. Negative/positive phase shifts induce a net power flow into/out from the microgrid area to restore the power balance and frequency. It is clearly desirable to keep tie-line power flows within safe limits and as close as possible to prescheduled (contracted) levels in most instances, and hence the control objective of the EMS is to monitor and regulate the net power flow over the tie-line despite local changes in demand (load frequency and power exchange control). As shown in Fig. 7, the filtered tie-line power measurement is transmitted via the IP link to the EMS. Real and reactive power flows in the line are assumed instrumented using regular substation equipment (e.g., that used in protection relays). This signal is compared to the contracted power flow, and the error is passed to the regulatory controller D(z). The output of this controller is a bias/trim signal u(k) which is transmitted via the IP link to the speed set point of the local asset controller on the biomass CHP plant. In our experiments, we employ the digital control methodology described in Sect. “Supervisory Controller Design”, to produce a closed loop with critical damping and ∼ =200 seconds settling time. A 1 second sampling time for the control loop was employed. Two experiments were considered, one without buffering and one featuring buffering, as described in the next section.

Illustrative Example: Statistical Delay Model In order to develop representative delay, jitter, and packet loss statistics for the IP links in the model, an experiment was carried out to measure the recorded latency and packet loss statistics experienced by probe packets sent over the public Internet. A public server in Newcastle upon Tyne, UK, was pinged by a client PC in Middlesbrough, UK, once every second for over 10 days. The physical distance between the 2 computers was ∼ = 50 KM. The round-trip delays were measured in each case. The total number of samples was 849,486 with 3,310 packets lost. The distribution of the latencies obtained was as shown in Table 1. From the raw data we obtained the mean latency μ = 28.82 milliseconds and the standard deviation σ = 56.05 milliseconds. From these data we may observe that the majority of the probability mass is concentrated around the mean with a delay between 20 and 30 milliseconds, but there exists a long tail with delays up to around 2,500 milliseconds occasionally experienced. Latencies above this threshold were assumed to be losses, with packet loss probability of 0.0039. The information related to the range and distribution of delays and packet loss probability was employed to calibrate the channel model employed in the simulations.

Illustrative Example: HIL Results Initially, a regulatory controller was designed for the system using the methodology of Sect. “Models and Technical Underpinning” under the assumption that the average delay of 28.82 ms in the communication links will result in a single sample

9 Control and Informatics for Demand Response and Renewables Integration Table 1 Latency statistics for IP link under test

Latency L (ms) 0 ≤ L < 10 10 ≤ L < 20 20 ≤ L < 30 30 ≤ L < 40 40 ≤ L < 50 50 ≤ L < 60 60 ≤ L < 70 70 ≤ L < 80 80 ≤ L < 90 90 ≤ L < 70 100 ≤ L < 2500

209 Probability 0.00000 0.00000 0.95703 0.00884 0.00530 0.00474 0.00365 0.00268 0.00168 0.00142 0.01466

Fig. 8 Power balance error (supervisory control with best-effort IP connection)

delay in both the forward and backward transmission links. A screenshot of the p.u. error following a 0.2 p.u. change in the local microgrid load is as shown in Fig. 8 below. Next, the buffering scheme proposed in Sect. “Models and Technical Underpinning” was also implemented. Given the mean latency μ = 28.82 ms and the standard deviation σ = 56.05 ms, buffer sizes of six samples (6,000 ms) in both forward and backward IP links were employed; from Eq. (1), it can be verified that the probability of exceeding this buffer size in any sample is 90% and xylose >80% both with and without addition of an acid catalyst (Kang et al. 2013).

Fermentation After obtaining simple sugar by hydrolysis conversion of these into ethanol take place with the help of microorganism during fermentation stage. Fermentation is a type of biological process in which hexoses and pentoses of carbohydrates are converted into ethanol by a variety of microorganisms like yeast and bacteria (Wu et al. 2010). Saccharomyces cerevisiae and Zymomonas mobilis are the most commonly used microorganisms in fermentation. However, Saccharomyces cerevisiae cannot metabolize xylose into ethanol. Other types of yeast and bacteria are under investigation for their conversion of xylose into ethanol. Hybrid species of fungi are being produced using genetic engineering that can produce large amount of hydrolysis enzymes, that is, cellulase, hemicellulase, and

18 Biomass as Sustainable Material for Bioethanol Production

469

xylanase, which could convert agricultural (corn and wheat stover, rice straw, and sugarcane bagasse) and forest residue into fermentable sugar. Moreover, genetically engineered microorganisms can efficiently ferment hexoses and pentoses into ethanol (Liu et al. 2013). There are various methods of integration of hydrolysis and fermentation: • In separate hydrolysis and fermentation (SHF), both steps are carried out in different reactors. Liberated cellulose is treated in different reactors for hydrolysis into simple sugar and then subsequent fermentation. It has advantages like facilitating optimization of separate rectors and selection of appropriate microorganisms for fermentation. However, it needs high cost investment for separate reactors; secondly inhibition of high concentrate glues fermenting microorganisms becomes difficult. • Other alternative methods include separate hydrolysis and co-fermentation (SHCF) and simultaneous saccharification and co-fermentation (SSCF) in which both hexoses and pentoses are fermented in same reactor by same strain of microorganism. These can significantly reduce cost investment being economical and increase commercial production of lignocellulose ethanol in short time .

Purification Bioethanol produced at fermentation stage is needed to be purified to reduce water content. Commonly ethanol produced from second-generation feedstock is in the range of 3–6% only which is very low as compared to 12–15% from first generation feedstock (Zhu et al. 2014). Additional distillation efforts are required to improve purity of bioethanol and remove access water content.

Microorganisms Microorganisms are required to hydrolyze lignocellulose hydrolysate to ethanol in industrial fermentation at broad substrate range which produces ethanol with high yield and productivity. There is need of microorganisms which are high tolerant to ethanol and inhibitors formed during pretreatment process. Scientific efforts are continued to produce such hybrid microorganisms that could efficiently ferment xylose into ethanol (Hu et al. 2014). Some groups of bacteria and yeast are able to convert pentose sugar into ethanol but with low productivity. Moreover, xylose fermenting yeast like Pachysolen tannophilus, Candida shehatae, and Pichia stipites usage at large scale is hindered by their sensitivity to higher concentration of ethanol. Careful monitoring of microaerophilic condition, sensitivity to inhibitors, and inability to ferment xylose at low pH is required. Yeast and bacteria have different metabolism pathways for xylose fermentation. In case of bacteria, xylose isomerase enzyme converts xylose into xylulose which is phosphorylated through pentose phosphate pathway (PPP) (Hendriks and Zeeman 2009) while in yeast xylose is first converted into xylitol and then to xylulose by xylose reductase and xylitol dehydrogenase enzyme where NADPH and NAD+ act as cofactors.

470

Rozina et al.

Certain filamentous fungi have been observed with the ability to ferment simple sugar such as glucose, galactose, mannose, xylose, and arabinose found in biomass hydrolysate. Some species of fungi like Monilia, Fusarium, Rhizopus, Aspergillus, Neocallimastix, and Trichoderma have been observed to possess ability to convert cellulose into ethanol (Hahn-Hagerdal et al. 2006). Currently, research is carried out on genetic transformation of Saccharomyces cerevisiae and Zymomonas mobilis to improve its fermentative performance on pentose sugar. Efforts are made to produce recombinant strains of yeast and bacteria to add to the need of industrial fermentation. In this regard first strain of Saccharomyces cerevisiae was obtained through the introduction of for xylose metabolism from Pichia stipitis (van Zyl et al. 2007). Another strain of Saccharomyces cerevisiae was generated by introduction of genes encoding xylose isomerase from Thermus thermophilus bacterium (Wu et al. 2010). Thermoanaerobaterium saccharolyticum (obligatory anaerobic bacterium) has been modified genetically to improve ethanolic fermentation (Kang et al. 2014a). Emphasis is given to generation of strains that would efficiently utilize pentoses and are resistant to high ethanol concentration. Saccharomyces cerevisiae is the most commonly used microorganism for industrial ethanol production. It has exhibited great efficiency to convert lignocellulosic hydrosylate biomass to ethanol (Karimi et al. 2013). However, Saccharomyces cerevisiae cannot ferment xylose to ethanol. Strains of Saccharomyces cerevisiae have been acquired with improved capacity of xylose utilization through mutagenesis, breeding, or adaptation (FAO/IFAD 2005) [77]. It was done by the introduction of Pichia stipitis genes XYL1 and XYL2 encoding xylose reductase and xylitol dehydrogenase respectively in combination with XKS1 encoding xylulokinase (ZK) in Saccharomyces cerevisiae. These transformed industrial strains of Saccharomyces cerevisiae are capable to convert non-detoxified lignocellulose hydrosylate into bioethanol (Kawaroe et al. 2015).

Biocatalyst for Ethanol Production Investigation showed that carbon-based solid catalyst can efficiently convert lignocellulosic biomass into bioethanol. Such catalysts carry functional groups like – SO3 , –COOH, and –OH which result in higher activity than acid catalysts (Table 3). Carbon-based sulfonated catalyst can replace the conventional corrosive H2 SO4 catalyst for fermentable sugar production than commercial catalysts [89]. Various feedstocks like bamboo, cotton, nutshell, and starch are used to derive catalysts for ethanol production. Wu et al. (2010) utilized biocatalysts derived from bamboo, cotton, and starch for ethanol production. They observed a comparatively lesser yield of fermentable sugar compare to other waste biomass-derived catalysts (Wu et al. 2010). Kang et al. (2013) found out from their investigation that lignocellulosic-derived catalysts are more promising than in terms of catalytic stability than acid catalysts. Waste biomass-derived catalysts have higher reusability and are cheaper in cost.

18 Biomass as Sustainable Material for Bioethanol Production

471

Table 3 Biomass-derived solid acid catalyst for bioethanol production Biomass type Lignin

Catalyst Application: name Bioethanol Hydrochar From inulin

Catalyst reusability (cycle) Five

Fermentable sugar yield (%) 65

Bamboo

# BCS

From cellulose

Three

16.71

Cotton

# BCS

From cellulose

Three

19.8

Starch

# BCS

From cellulose

Three

4.6

Nut shell

# CBS

From cellulose

More than nine

59

Sucralose

# CBS

From cellulose

Nine

55

Reference Nigam and Singh (2011) Wu et al. (2010) Wu et al. (2010) Wu et al. (2010) Liu et al. (2013) Hu et al. (2014)

Lignin-derived hydrochar catalyst gave highest bioethanol yield of 65%. It showed moderate surface area and acid strength resulting in maximum ethanol yield. It has the ability to be used up to five successive cycles without further purification (Kang et al. 2013). Liu et al. (2019) concluded from their research that glucose-carbon and sucrose carbon hold greater catalytic activity than nut shell derived sulfonated carbon for crystalline cellulose hydrolysis (Liu et al. 2019). It is because of its lower content –SO3 H functional group than –OH group in sulfonated nut shell catalyst. Sucralose-derived catalysts are proven to be highly effective for hydrolysis of biomass due to the presence of –Cl and SO3 H group, making it more accessible to biomass cellulose to act upon (Chakraborty et al. 2016). These biocatalysts have potential to replace commercially synthesized catalysts which account for high cost productivity. Moreover, it will make bioethanol production a completely renewable process. However, there is great need of scientific investigation and production of such biocatalysts and then their practical implication in ethanol industry (Saeed et al. 2014).

Problems Related with Production of Bioethanol There are some problems related with production of bioethanol from lignocellulosic biomass which are encountered by first-generation feedstocks. During pretreatment and hydrolysis stage certain chemical inhibitors, that is, furans such as furfural and 5-hydroxymthylfufural (5-HMF), and phenol such as vaniline, 4-hydroxybenzaldehyde, and syringaldehyde, are produced which need to be removed to get maximum conversion during fermentation. Saccharomyces cerevisiae and Zymomonas mobilis are the most promising microorganisms that succeed to survive in high ethanol concentration. These are also moderately tolerant to acid, sugar, and other inhibitors making it commercially popular for industrial

472

Rozina et al.

fermentation. Genetically engineered acid resistant strains of these organisms should be produced to further deal with inhibitors and enhance ethanol yield under stress conditions.

Conclusion This review identifies biomass resources for ethanol production and their potential to deal with the current energy demand. A wide variety of resources like agricultural crop residue, forest residue, municipal, and other waste are obtainable resources of biomass distributed throughout the country in different ecological zones. Bioethanol is an incentive source of energy which can add to reduction of release in greenhouse gases to the environment and can substitute fossil fuels. Fuel energy demand is increasing each year. Direct use of sugarcane for ethanol production yields 70 l of ethanol/ton of sugarcane. Development of bioethanol technology will bring a considerable foreign exchange saving. Currently, researchers have focused on biomass resources and their potential for bioethanol production using both first- and second-generation feedstock. Recently, limited information on biomass technologies and biomass conversion in to bioethanol is available. Bioethanol is a promising and proximate carbon neutral renewable energy resource. It is used as an engine fuel and fuel additive in internal combustion engines. It is environment friendly (having low impact on climate change) and reduces emission of CO2 gas. It is used as octane enhancer and oxygenated fuel mix in unleaded gasoline for cleaner combustion. Bioethanol has properties that are suitable for spark ignition in IC engines. First-generation feedstock that has potential for bioethanol production includes the following: sugarcane, maize, sweet sorghum, wheat, cassava, and sweet fruits. Second-generation feedstock are nonedible residue of food crop for ethanol production like wheat and rice straw or nonedible plant or animal biomass like forest residue, grasses, agricultural residue, wood pulp residue, industrial waste, animal waste, and municipal trash and garbage. The valuable third generation feedstock resources can be utilized for ethanol production. Saccharomyces cerevisiae and Zymomonas mobilis are the most promising and commonly used microorganisms that succeed to survive in high ethanol concentration. First-generation feedstocks are directly subjected to fermentation while pretreatment is an essential delignification step in synthesis of ethanol from lignocellulosic as a result of which cellulose becomes more accessible for hydrolysis step and lignin content is reduced. Various techniques used for pretreatment, that is, physical, chemical, and biological pretreatment, are discussed. Ethanol production from these feedstocks has potential to be valuable substitute to gasoline. Ethanol production can contribute to economy by providing job opportunities especially in rural communities. There is great potential to capitalize on ethanol fuel production due to presence of forest and crop land resources. Development of ethanol industry would create construction and operational jobs and strengthen rural economy of the world.

18 Biomass as Sustainable Material for Bioethanol Production

473

References ANFAVEA (2016) Brazilian automotive industry yearbook. Brazilian Automotive Industry Association. Available from: http://www.anfavea.com.br/anuario2016/anuario2016.zip Annamalai K (2013) Respiratory quotient (Rq), exhaust gas analyses, CO2 emission and applications in automobile engineering. Adv Automob Eng 2:2–4 Aradhey A (2016) India biofuel annuals. Global Agricultural Information Network. GAIN report number: IN6088 Bertrand E, Vandenberghe LPS, Soccol CR, Sigoillot JC, Faulds C (2016) First generation bioethanol. In: Soccol CR, Brar SK, Faulds C, Ramos LP (eds) Green fuels technology. Springer, Cham, pp 175–212 Bhatia SC (2014) Ethanol. In: Bhatia SC (ed) Advanced renewable energy systems. Part I. Woodhead Publishing India, New Delhi, pp 523–572 Bracht A, Morais GR, Baesso ML, Correa RCG, Peralta RM (2014) Biological pretreatment of Eucalyptus grandis sawdust with white-rot fungi: study of degradation patterns and saccharification kinetics. Chem Eng J 258:240–246 Cannon JS, Sperling D (eds) (2009) Reducing climate impacts in the transportation sector. Springer, Dordrecht. https://doi.org/10.1007/978-1-4020-6979-6 Chakraborty R, Chatterjee S, Mukhopadhyay P, Barman S (2016) Progresses in waste biomass derived catalyst for production of biodiesel and bioethanol: a review. Procedia Environ Sci 35:546–554 Cui XH, Zhao XB, Zeng J, Loh SK, Choo YM, Liu DH (2014) Robust enzymatic hydrolysis of formiline-pretreated oil palm empty fruit bunches (EFB) for efficient conversion of polysaccharide to sugars and ethanol. Bioresour Technol 166:584–591 Du C, Campbell G, Misailidis N, Mateos-Salvador F, Sadhukhan J, Mustafa M, Weightman R (2009) Evaluating the feasibility of commercial arabinoxylan production in the context of a wheat biorefinery principally producing ethanol, part 1. Experimental studies of arabinoxylan extraction from wheat bran. Chem Eng Res Des 87:1232–1238 Duku MH, Gu S, Hagan EB (2011) A comprehensive review of biomass resources and biofuels potential in Ghana. Renew Sust Energ Rev 15:404–415 FAO (2012) Countries: geography – Nigeria 2012. http://www.fao.org/forestry/country/18310/en/ nga/. Accessed 18 Jan 2016 FAO/IFAD (2005) A review of cassava in Africa with country case studies on Nigeria, Ghana, the United Republic of Tanzania, Uganda and Benin. In: Proceedings of the valid forum on the global cassava development strategy. International Fund for Agricultural Development/Food and Agriculture Organization of the United Nations, Rome Ghosh SK (2016) Biomass & bio-waste supply chain sustainability for bio-energy and bio-fuel production. Procedia Environ Sci 31:31–39 Goldemberg J (2013) Sugarcane ethanol: strategies to a successful program in Brazil. In: Lee JW (ed) Advanced biofuels and bioproducts. Springer Science + Business Media, New York, pp 13–20 Hahn-Hagerdal B, Galbe M, Gorwa-Grauslund MF, Liden G, Zacchi G (2006) Bio-ethanol – the fuel of tomorrow from the residues of today. Trends Biotechnol 24:549–558 Hahn-Hägerdal B, Karhumaa HBK, Fonseca C, Spencer-Martins I, Gorwa-Grauslund MF (2007) Towards industrial pentose-fermenting yeast strains. Appl Microbiol Biotechnol 74(5):937–953 Hendriks ATWM, Zeeman G (2009) Pretreatments to enhance the digestibility of lignocellulosic biomass. Bioresour Technol 100:10–18 Hu S, Smith TJ, Lou W, Zong M (2014) Efficient hydrolysis of cellulose over a novel sucralosederived solid acid with cellulose-binding and catalytic sites. J Agric Food Chem 62(8):1905– 1911 Irfan M, Nadeem M, Syed M (2014) Ethanol production from agricultural wastes using Saccharomyces cerevisiae. Braz J Microbiol 45:457–465 Ishola MM, Brandberg T, Sanni SA, Taherzadeh MJ (2013) Biofuels in Nigeria: a critical and strategic evaluation. Renew Energy 55:554–560

474

Rozina et al.

Jagessar CF (2016) The fermentation of banana (Musa acuminata), Mongo (Mangifera indica L.) and pineapple (Ananas comosus) mash in the absence and presence of additives. Am J Res Commun 4(3):131–140 Kang S, Ye J, Zhang Y, Chang J (2013) Preparation of biomass hydrochar derived sulfonated catalysts and their catalytic effects for 5-hydroxymethylfurfural production. RSC Adv 3:7360 Kang Q, Appels L, Baeyens J, Dewil R, Tan T (2014a) Energy-efficient production of cassavabased bio-ethanol. Adv Biosci Biotechnol 5(12):925–939 Kang Q, Appels L, Tan T, Dewil R (2014b) Bioethanol from lignocellulosic biomass: current findings determine research priorities. Sci World J 13:298153 Karimi K, Chisti Y (2015) Future of bioethanol. Biofuel Res J 5:147 Karimi K, Pandey A (2014) Current and future ABE processes. Biofuel Res J 3:77 Karimi K, Shafiei M, Kumar R (2013) Progress on physical and chemical pretreatment of lignocellulosic biomass in biofuels and biorefineries: recent developments. Springer, Berlin/Heidelberg Kawaroe M, Sari DW, Hwangbo J, Santoso J (2015) Optimum fermentation process for red macroalgae Gelidium latifolium and Gracillaria verrucosa. J Eng Technol Sci 47(6):650–663 Lam MK, Lee KT (2011) Mixed methanol–ethanol technology to produce greener biodiesel from waste cooking oil: a breakthrough for SO4 2− /SnO2 –SiO2 catalyst. Fuel Process Technol 92:1639–1645 Li S, Stuart JD, Li Y, Parnas RS (2010) The feasibility of converting Cannabis sativa L. oil into biodiesel. Bioresour Technol 10:8457–8460 Liska AJ, Yang HS, Bremer VR, Klopfenstein TJ, Walters DT, Erickson GE et al (2009) Improvements in life cycle energy efficiency and greenhouse gas emissions of corn-ethanol. J Ind Ecol 13:58–74 Liu M, Jia S, Gong Y, Song C, Guo X (2013) Effective hydrolysis of cellulose into glucose over sulfonated sugar-derived carbon in an ionic liquid. Ind Eng Chem Res 52(24):8167–8173 Liu M, Li X, Du Y, Han R (2019) Adsorption of methyl blue from solution using walnut shell and reuse in a secondary adsorption for Congo red. Bioresour Technol Rep 1(5):238–242 Maity JP, Chen CY, Bhattachary P (2014) Microalgae for third generation biofuel production, mitigation of greenhouse gas emissions and wastewater treatment: present and future perspectives. Energy 78:104–113 Martinez-Hernandez E, Martinez-Herrera J, Campbell GM, Sadhukhan J (2014) Process integration, energy and GHG emission analyses of Jatropha-based biorefinery systems. Biomass Convers Biorefin 4:105–124 Monceaux DA, Ingledew WM, Kelsall DR, Austin GD, Kluhspies C (2009) Alternative feedstocks for fuel ethanol production. In: The alcohol text book. A reference for the beverage, fuel and industrial alcohol industries, vol 5. Nottingham University Press, Nottingham, pp 47–57 Mussato SI, Machado EMS, Carneiro LM, Teixeira JA (2012) Sugar metabolism and ethanol production by different yeast strains from coffee industry wastes hydrolysates. Appl Energy 92:763–768 Nigam PS, Singh A (2011) Production of liquid biofuels from renewable resources. Prog Energy Combust Sci 37:52–68 Numjuncharoen T, Papong S, Malakul T, Mungcharoen T (2015) Life-cycle GHG emissions of cassava-based bioethanol production. Energy Procedia 79:265–271 Ohgren K, Galbe M, Zacchi G (2005) Optimization of steam pretreatment of SO2 -impregnated corn stover for fuel ethanol production. Appl Biochem Biotechnol 121:1055–1067 Olofsson K, Bertilsson M, Lid G (2008) A short review on SSF – an interesting process option for ethanol production from lignocellulosic feedstocks. Biotechnol Biofuels 1:1–14 Pensria B, Aggarangsi P, Chaiyaso T, Chandeta N (2016) Potential of fermentable sugar production from Napier cv. Pakchong grass residue as a substrate to produce bioethanol. Energy Procedia 89:428–436 Petrova P, Ivanova V (2010) Perspectives for the production of bioethanol from lignocellulosic materials. Biotechnol Biotechnol Equip 24:529–546 Regassa TH, Wortmann CS (2014) Sweet sorghum as a bioenergy crop: literature review. Biomass Bioenergy 64:348–355

18 Biomass as Sustainable Material for Bioethanol Production

475

Renewable Fuel Association (RFA) (2019) http://www.ethanolrfa.org/resources/industry/statistics/ #1454099103927-61e598f7-7643 Romaní A, Garrote G, Parajó JC (2012) Bioethanol production from autohydrolyzed Eucalyptus globulus by simultaneous saccharification and fermentation operating at high solids loading. Fuel 94:305–312 Rozina, Asif S, Ahmad M, Zafar M, Ali N (2017a) Prospects and potential of fatty acid methyl esters of some non-edible seed oils for use as biodiesel in Pakistan. Renew Sust Energ Rev 74:687–702 Rozina AM, Zafar M, Ali N, Lu H (2017b) Biodiesel synthesis from Saussurea heteromalla (D. Don) Hand-Mazz integrating ethanol production using biorefinery approach. Energy 141:1810– 1818 Rozina, Ahmad M, Zafar M, Ali N (2019) Biodiesel synthesis and physiochemical analysis of Taraxacum officinale F. H. Wigg seed oil. Int J Environ Sci Technol 107:1735–1472 Saboori B, Sapri M, Bin BM (2014) Economic growth, energy consumption and CO2 emissions in OECD (Organization for Economic Co-operation and Development)’s transport sector: a fully modified bi-directional relationship approach. Energy 66:150–161 Saeed MA, Medina CH, Andrews GE, Phylaktou HN, Slatter D, Gibbs BM (2014) Agricultural waste pulverised biomass: MEC and flame speeds. J Loss Prev Process Ind 36:308–317 Sassner P, Galbe M, Zacchi G (2006) Fuel ethanol production from steam-pretreated corn stover using SSF at higher dry matter content. Enzym Microb Technol 39:756–762 Sathesh-Prabu C, Murugesan AG (2011) Potential utilization of sorghum field waste for fuel ethanol production employing Pachysolen tannophilus and Saccharomyces cerevisiae. Bioresour Technol 102:2788–2792 Shaaban M, Petinrin JO (2014) Renewable energy potentials in Nigeria: meeting rural energy needs. Renew Sust Energ Rev 29:72–84 Shoemaker CE, Bransby DI (2010) Chapter 9. The role of sorghum as a bioenergy feedstock. Sustain Altern Fuel Feed Oppor challenges road maps six USR eg, Proceeedings of the Sustainable Feedstocks for Advance Biofuels Workshop, Atlanta, GA 28–30 Tao L, Andy A, Elander RT, Pallapolu VR, Lee YY, Garlock RJ et al (2011) Process and technoeconomic analysis of leading pretreatment technologies for lignocellulosic ethanol production using switchgrass. Bioresour Technol 102:11105–11114 Tomas-Pejo E, Oliva JM, Ballesteros M (2008) Realistic approach for full-scale bioethanol production from lignocellulose: a review. Sci Ind Res 67(11):874–884 van Zyl WH, Lynd LR, den Haan R, McBride JE (2007) Consolidated bioprocessing for bioethanol production using Saccharomyces cerevisiae. Adv Biochem Eng Biotechnol 108:205–235 Walia NK, Bedi SS, Kundu K, Karmakar R (2013) Production of bioethanol from mango peel. Int J Eng Res Technol 2(1):32 Wan C, Zhou Y, Li Y (2011) Liquid hot water and alkaline pretreatment of soybean straw for improving cellulose digestibility. Bioresour Technol 102:6254–6259 Wang M, Han J, Dunn JB, Cai H, Elgowainy A (2012) Well-to wheels energy use and greenhouse gas emissions of ethanol from corn, sugarcane and cellulosic biomass for US use. Environ Res Lett 7:045905 Wu Y, Fu Z, Yin D, Xu O, Liu F, Lu C, Mao L (2010) Microwave-assisted hydrolysis of crystalline cellulose catalyzed by biomass char sulfonic acids. Green Chem 12:696–700 Zadeh AKA, Almassi M, Meighani HM, Borghei AM, Azizian J (2011) Suitability of Carthamus oxyacantha plant as biodiesel feedstock. Aust J Crop Sci 5(12):1639–1643 Zhang YJ, Li Q, Su JM (2015) A green and efficient technology for the degradation of cellulosic materials: structure changes and enhanced enzymatic hydrolysis of natural cellulose pretreated by synergistic interaction of mechanical activation and metal salt. Bioresour Technol 177: 176–181 Zhu JY, Zhuang XS (2012) Conceptual net energy output for biodiesel production from lignocellulosic biomass through biorefining. Prog Energy Combust Sci 38:583–598 Zhu QL, Shao R, Dong R, Yun Z (2014) An integrated approach for obtaining biodiesel, sterols, gossypol, and raffinose from cotton seed on a biorefinery concept. Energy 70:149–158

Machine Learning–Enhanced Decision-Making

19

Nikodem Rybak and Maureen Hassall

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Defining Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Artificial Neural Networks and Deep Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Pre-processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Source Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pre-processing of Variable Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pre-processing of Categorical Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Missing Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Visualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Representativeness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Section Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Type of Machine Learning Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Supervised and Unsupervised Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classification Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Regression Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Clustering Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anomaly Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Example Application: Machine Learning-Based System for Real-Time Inattention Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Deployment Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application Outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Section Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

478 479 481 482 484 484 485 486 488 488 489 490 491 491 492 494 497 501 503 504 504 504 508 511 511

N. Rybak () School of Chemical Engineering, The University of Queensland, St Lucia, QLD, Australia e-mail: [email protected] M. Hassall School of Chemical Engineering, The University of Queensland, Brisbane, QLD, Australia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_20

477

478

N. Rybak and M. Hassall

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

512 514 514

Abstract

Machine learning comprises a broad range of analysis methods that can be applied to artefacts such as data, images, and sound recordings to produce insights or findings. This chapter describes what machine learning is and how it works. It then provides an example of real-world application of machine learning to address industry problem. The chapter concludes by discussing important considerations when using machine learning methods to ensure that quality results are produced, and these results are validly interpreted. Keywords

Computer vision · Deep learning · Machine learning · Neural network · Supervised learning · Unsupervised learning

Introduction Just as the twentieth century was the age of electricity, the twenty-first century is the time of artificial intelligence with its applications regularly occupying leading positions in the rankings of the fastest developing, most important or most groundbreaking technologies of the last decade, e.g., speech to text systems, AI-based search engines, predictive and recommendation software engines, and autonomous vehicles to name a few. Thanks to the development of information technologies, in particular technologies of data generation, storage and processing, the amount of digital data is almost doubling every 2 years (Bibri 2019). This means that although the main source of data are the current activities of companies and institutions, more and more of this data is generated and analyzed automatically, by computer systems. This data contains not only a detailed description of the activities of companies or institutions, but also, to a large extent, a description of the world that surrounds us. The information hidden in this data is of great value, and only the methods of extracting it are needed. The importance of the value of data is significant and should not be ignored. To obtain and sustain operational excellence in this rapidly changing, increasingly competitive, and more complex world, industry practitioners need to leverage Industry 4.0 technologies. One set of Industry 4.0 technologies that can be deployed to extract maximum values from the data is machine learning (as shown in Fig. 1) which is explored in this chapter. The chapter begins with a brief overview and concise definitions of machine learning as a subdomain of computer science. Then important issues related to

19 Machine Learning–Enhanced Decision-Making

479

INDUSTRY 4.0 TECHNOLOGY

Advanced automation and robotics

Advanced Data Analytics

Big data and predictive analytics

Industrial Internet of Things

Cyber Physical Systems

Artificial Intelligence

Machine Learning

Deep Learning Fig. 1 Outline of Industry 4.0 technology

data exploratory analysis as a first stage of any practical machine learning implementation are outlined. Next, the chapter discusses important machine learning algorithms and data transformation methods. Each method receives its theoretical background as well as examples of real-world industry applications. The third part of the chapter offers a detailed case study of machine learning-based system that has been developed to detect industry practitioners’ situational attention thereby aiding decision making related to safety and operational efficiency. To conclude, thoughts are offered regarding the use of these technologies in future studies and implications for industry. The objectives for this chapter are twofold, firstly to offer a thorough introduction to machine learning tools for readers who do not have a computer science background, and secondly to present a set of useful machine learning-based technologies for supporting efficient and effective decision-making in real-world implementations.

Defining Machine Learning Machine learning (ML) is a subdomain of artificial intelligence (AI), which is part of a broader field of computer science. Whereas AI encompasses wide-ranging and very diverse concepts like knowledge representation, reasoning, planning and

480

N. Rybak and M. Hassall

Fig. 2 Timeline of Machine Learning

manipulation, the machine learning subdomain typically focuses on learning from data and pattern recognition. There has been numerous theoretical developments and discoveries in other fields like neurosciences that were essential to get to current stage (see Fig. 2). The main objective of machine learning is to deliver mathematical models and algorithmic solutions that can be utilized to solve problems like those that are too demanding to be represented as standard mathematical problems. An example of such problems is recognizing the patterns of facial expressions within an image containing a human face. This function can be performed by a mature human brain with almost no effort, but remains a daunting task to be represented as a mathematical problem due to its vast and mostly unconstrained complexity (Rybak et al. 2017a).

19 Machine Learning–Enhanced Decision-Making

481

To understand how machine learning has revolutionized the field of computer science requires briefly looking into previous solutions. In the past, pattern recognition methods used mainly hand-engineered features to quantify the data. For example, in image processing tasks previous algorithms executed feature extraction separately on each image in a dataset according to predefined image descriptors, returning a vector to quantify the content of an image. The next step in the process was to extract significant structures from a dataset, by performing “flat” statistical inference on received vectors. Traditional hand-engineered image descriptors endeavored to encode features such as color, texture, or shape (Szeliski 2010). Other broadly utilized methods were key point detectors (FAST, Harris, DoG to name a few) and local invariant descriptors (e.g. SIFT, SURF, BRIEF, ORB, etc.) (Nixon and Aguado 2019). Some traditional image processing methods like Histogram of Oriented Gradients (HOG) proved to be efficient at detecting objects in images when the rotation of an object within an image did not significantly vary from what the classifier was trained on (Dalal and Triggs 2005). In all abovementioned approaches, a process was predefined to encode the exact characteristics of an image (i.e., color, shape, texture, etc.) and quantify it. Given an input matrix of pixels, one would apply this hand-defined algorithm to the pixels, and in return receive a feature vector quantifying the image contents – the image pixels themselves did not serve a purpose other than being inputs to the feature extraction process. The feature vectors that resulted from feature extraction were what was truly important as they served as inputs to traditional pattern recognition models. The machine learning methods use a different approach for locating patterns within an image. In place of hand-engineered instructions to extract specific features, these features are repeatedly learned from the training process completed on examples for each instance. There is a no one source with curated resources summarizing and presenting most recent developments in machine learning. Nevertheless, important information and latest updates can be found on industry’s dedicated research and development websites – Google, IBM, and Microsoft.

Artificial Neural Networks and Deep Learning Artificial Neural Networks (ANNs) are a specific class of machine learning procedures conceptually based on the manner in which biological neural structures process signals. ANNs are sets of mathematical tools recently utilized with great success to tackle various classification problems, including speech, images, medical diagnostics, and other signal data (Al-Shayea et al. 2013; Alanis et al. 2019; Amato et al. 2013; Bekey and Goldberg 2012; Cichocki and Unbehauen 1993; Damper 1997; Haykin 1994; Hertz et al. 1991; Jain and Vemuri 1998; Rezzoug and Gorce 2003; Wasserman 1993).

482

N. Rybak and M. Hassall

Deep learning is the latest development in machine learning. Deep learning methods and techniques scale up the size and complexity of traditional ANNsstatistical methods to create increasingly richer functionality. Deep learning models contain dozens or hundreds of hidden sampling layers, whereas previously ANNs only had a few, meaning that deep learning models are capable of hierarchical learning where simple concepts are learnt in the lower layers (e.g., lines, shapes) and more abstract patterns in the higher layers of the network (e.g., faces, objects). A typical instance of the application of deep neural networks to computer vison are tasks like handwritten character recognition which automatically learns discriminating patterns from images by sequentially assembling layers vertically. Lower level layers of the trained network represent configurations of corners and edges, whereas high level layers learn more abstract relations between shapes to discriminate between characters. Therefore, the key aspect of deep learning is that patterns extracted on a level of each layer of neural network. They are automatically learned from data itself utilizing a general-purpose learning procedure. In many applications, deep learning algorithms are now considered the most powerful classifiers and are currently responsible for pushing the state-of-the-art forward in areas such as computer vision, natural language processing, or signal processing – subfields that traditionally leverage machine learning (Goodfellow et al. 2016). In summary, machine learning algorithms learn from data, create connections, search relations, discover patterns, create new examples, etc. on the basis of welldefined and well-curated data sets. The selection of machine learning algorithms, their parameterization and learning process with the use of curated training data is only one stage of this process, which additionally is most often not the most significant one when it comes to the impact on the end results. Figure 2 provides an overview of the steps required to develop a machine learning solution. Typically, the most significant step of this process is performing data analysis and pre-processing to obtain the well-defined and well-curated data. Therefore, before moving on to discussing categories of specific machine learning algorithms and examples of their implementations, firstly a closer look will be taken at the methodology of the first stage of any machine learning project, namely, data analysis and pre-processing; and in the subsequent section practical issues related to models training and models evaluation will be discussed (see Fig. 3).

Data Pre-processing This chapter section intends to introduce the Reader with the first stage of implementing solutions from the scope of machine learning – data analysis and pre-processing. This is a crucial step in deploying valid models because “its output

19 Machine Learning–Enhanced Decision-Making

483

Fig. 3 Diagram overviewing stages of machine learning solution

can only be as accurate as the information entered into it” as per the garbage in – garbage out or rubbish in – rubbish out principle (Branch 2016). The first stage of data pre-processing is Exploratory Data Analysis (EDA). EDA is an approach created within a field of computer science as a result of combining statistics, database technologies, algorithmics, and data visualization. The purpose of EDA is to assist in formulating hypotheses about the sources and causes of observed phenomena. The algorithmics is a study of design and efficiency of algorithms. The most important role under algorithmics is currently played by deep learning. The sources of data for the data science experiments can be constituted by transaction databases, so called data warehouses, big data systems, analytical systems, documents in various formats, image, audio and multimedia files, websites, or network services – virtually every source of data can be used provided that it is possible to transform data originating from that source into tabular form. The tabular form is exceptionally universal. In tables consisting of vertical columns and horizontal rows, you can write not only usual numerical data, but also images (numerically coded values of subsequent pixels will be stored in specific table columns), spectrograms acquired on the basis of voice signal transformation, or text (relevance phrases are divided to be saved in separate columns or rows). The tabular interpretation of source data by the data exploration algorithms, with few exceptions, is the same as its basic statistical interpretation. Rows (records) represent observations (analyzed cases) while columns represent their attributes (variables).

484

N. Rybak and M. Hassall

Source Data Analysis Data exploration algorithms are exceptionally susceptible to the quality of source data. The universal GIGO (Garbage In, Garbage Out) rule, according to which the processing of incorrect data will be incorrect regardless of the correctness of data processing procedure, is exceptionally important with reference to data exploration (Liu et al. 2017). Because it is impossible to create a universal tool that will detect incorrect data, the quality evaluation and appropriate preparation of source data constitute as a necessary condition of successful implementation of machine learning solution. Firstly, the usefulness of data in solving stated problem should be evaluated – the data should not be modified at this stage. When evaluating the data quality, it is relevant to find answers to three basic questions: 1. What information are included in source data? 2. Can those data be used to acquire a solution to stated questions? 3. What problems are to be expected during the creation of exploration models for such data? Next, the errors found in data must be removed – the purpose of this stage is to adjust the training data to technical requirements of its exploration models. At the end, the data should be adjusted to the stated problem – this stage consists of implementing new information into the training data by transforming them.

Data Attributes Object attributes are called variables. Variables are divided into numerical (quantitative) and categorical (qualitative). Only digits can be values of numerical variables and additionally: 1. The collection of available values can be countable and has exceptionally great number of values (like, e.g., age in days) or can be uncountable (e.g., length): numerical variables, the value collection of which is uncountable, are called continuous while such numerical variables, the value collection of which is countable – number of values is definite despite the big amount – are called discrete. 2. These values can be compared with each other (e.g., 2 is greater than 1). 3. It is possible to determine the distance between values (e.g., by calculating the difference between them). 4. It is possible to conduct arithmetic operations, such as adding, multiplying, or averaging, on them.

19 Machine Learning–Enhanced Decision-Making

485

The values of categorical variables can be digits or data of a different type (e.g. text data), but: 1. The collection of available data is always limited (e.g., up to 12 in case of months); for this reason, the categorical variable values are called states. 2. If these values can be compared with each other in a reasonable way (e.g., danger level), then the categorical variable is called sequence. Otherwise, it is a regular categorical variable (as in case of, e.g., genders). 3. The determination of a distance between values is possible only under an adopted model (e.g., distance between colors in an adopted color palette). 4. It is impossible to conduct direct arithmetic operations on them. In the next subsections pre-processing of both variable and categorical data is discussed.

Pre-processing of Variable Data Variables can adopt a single or multiple value. Single-value variables are otherwise known as constant variables. Because they carry no information with them, they should not be used in the data exploration process. Before the constant variables from training data collection is removed, one should check whether or not a given variable is a single-value variable in a given test or in the whole collection of source data – if certain variable values occur exceptionally rare (e.g., once per 1,000,000 cases) then one most likely will not find it in a test counting 10,000 rows. A completely different class is constituted by variables with nonrepeatable values. Although such variables (such as employee identification number) clearly identify objects, but this dependency has no practical value for prediction models due to the nonrepeatability of their values. The third type of variables unusable by prediction models are monotonous variables. The values of such variables are constantly increasing or decreasing. Such type of variables occurs exceptionally often, for example: 1. Values of all variables related to time (such as time of event, report issuance date) are increasing. 2. Values of many variables not directly related to time are also increasing or decreasing – to such variables belong, e.g., invoice numbers. Because values of such variables are continuously increasing or decreasing, they must be transformed to use them as output variables for data exploration models. If this step is ignored, then the training data (data that served as a basis for teaching the model) will differ from data used for prediction. The distribution of numerical variable values is analyzed with the help of descriptive statistics and graphics (Fernández-Delgado et al. 2014).

486

N. Rybak and M. Hassall

Fig. 4 An example of descriptive statistics used to visualize accuracy of regression model and present detected data points. X-axis presents observed (actual) numerical values and Y-axis presents values predicted by the regression model. Colors indicate model accuracy for each data point. Darker colors indicate more accurate prediction

Descriptive statistics can be used to determine quality and representativeness of source data. They determine a central tendency, spread (dispersion), and symmetry of numerical variable values distribution. These parameters can then be discussed with an expert from a given domain (domain expert), to determine quality and representativeness of source data. The simplest way to learn the most important descriptive statistics is to visualize the data set (see example of data visualization in Fig. 4). Visualizations of the data help prompt discussions with the domain expert around the validity of the data especially the outliers, the completeness of the data (e.g., is there missing data) and the representativeness of the data as a measure of real-life scenarios. Aside from descriptive statistics, the following are also useful in the evaluation of data (both numerical and categorical): 1. Number of all existing variable values. 2. Number of unique variable values. 3. Number of missing variable values.

Pre-processing of Categorical Variables The distribution of categorical variables is evaluated with the help of frequency tables or histograms. They contain information about frequency of occurrence for specific states and allow to evaluate the distribution of categorical variables. The objective information quantity measurement methods were developed by Claude Elwood Shannon (Shannon 1948). Shannon’s methods quantify information

19 Machine Learning–Enhanced Decision-Making

487

and use the concept of entropy to quantify the uncertainty involved in categorical variables. The evaluation of usefulness of data is significantly influenced by the amount of information contained in variables. This general definition will be explained further on examples. The smallest information unit is bit (Binary digIT). Because bit can adopt one of two values (0 or 1), one bit communication carries at most one piece of information, two bits allow to convey four pieces of information, etc. Although information can be measures using different methods, the most often it is expressed with the help of bit quantity. The maximum quantity of information is calculated as log2 (number of possible states). For example, four pieces of information can be saved using 2 bits, 16 pieces – using 4 bits, and saving 100 pieces of information requires 7 bits. In practice, the number of information contained in data is smaller than it would result from the number of bits, which in turn results from the redundancy of numerous information channels. For example, the sentences written in English contain approximately half of redundant (repeated) information. The measure of uncertainty in data sets is entropy – average amount of information attributable to a single message from information source. Let’s assume that someone plays a game consisting of drawing a number from 1 to 10 and the task is to guess whether the selected number is even or odd. If the drawing is honest, then the player has a 50% chance to win. The number of possible states for this system is 2 (winning or losing), and the number of information saved in it is equal to –log2 2, that is, 1 bit. Both states are equally probable in this lottery. In order to calculate the amount of information conveyed in the message about winning, one must calculate how many bits of information are contained in each single message and then multiply the acquired number of bits by the probability of its occurrence. In this case the single message contained –log2 0.5 (= 1 bit) of information and the chance to win was 50%. multiplying −1 by 50% gives −0.5 – the message about winning (as well as message about losing) carried 0.5 bit of information. Let’s change rules of the game – the player has to guess the number instead of guessing whether it is even or odd. In this game the chance to win is equal to 10% (1 out of 10 numbers) and the risk to lose is 90%. Now the message about winning will be a greater surprise than previously and therefore it should carry more information. In order to check it, let’s count the quantity of information conveyed in win and loss messages: • Number of possible systems states is still 2. • The probability of winning is 10% and probability of loss – 90%. • The total quantity of information about winning is –log2 0.1, that is, −3.32 and quantity of information about losing is –log2 0.9, that is, −0.15. • Multiplying the total quantity of information in both states by the probability of their occurrence gives the quantity of information conveyed in win and loss messages: the win message carries −3.32 · 10% = 0.32 bit of information; the loss message carries −0.15 · 90% = 0.136 bit of information.

488

N. Rybak and M. Hassall

Let’s notice that this time the total quantity of information in the system amounts to 0.32 + 0.13 = 0.47 bit, which is 0.03 bit less than the previous system contained. The majority of this information is hidden in the win message. The first system (game, in which chance to win was 50%) was characterized by maximum entropy – the probability of occurrence of each state was the same. The entropy of second system (game, in which the chance to win was 10%) was smaller and therefore that system contained less information.

Missing Data The evaluation of missing data should be started from determining whether such data are unknown or nonexistent. Typically, in databases both cases are present and presented using the NULL value, even though the reason they are missing is different as per the following: 1. Unknown value occurs in reality, but has not been measured – for example, if one does not know where exactly an accident happened, then the value of Location attribute will be NULL. 2. The nonexistent value does not occur in reality – for example, if a given person did not participate in an accident, then the value will not exist and will also be saved as NULL. The missing data could be also represented using substitute values, for example, unknown age can be saved as 0. Such substitute NULL values are most often the default column values or empty series of symbols. If a certain value occurs significantly more often than other values and it would result from domain knowledge, then it probably represents missing values. The result of evaluating missing data should be the acquisition of responses to the following questions: 1. Do substitute NULL values occur in source data? 2. If yes, then were all substitute NULL values identified and replaced with other values (e.g., Missing value)? 3. Is it possible to fill the unknown values? 4. Does the fact that a given value is missing carries with it a valuable information? The answers to these questions will help to improve knowledge of attribute values distributions, correlation analysis, and domain expert’s knowledge not only defining missing values, but also substituting them with different values.

Data Visualization Descriptive statistics are often more difficult to interpret when compared to graphic summaries. Moreover, the variables may have the same range, mean, and standard

19 Machine Learning–Enhanced Decision-Making

489

deviation and yet possess different distribution of values. Generally speaking, descriptive statistics are a significant simplification and do not contain as set of information on data (Chatfield and Xing 2019). Therefore, the correct data evaluation is possible in numerous cases only with the use of data visualization. Data visualizations are the more helpful, the more data distribution differs from normal distribution. In particular, distributions other than uni-modal distributions (i.e., such as normal distribution with single dominant) are difficult to analyze using descriptive statistics. For example, if a distribution is bi-modal (distribution with two dominants) and the most often occurring values are extreme (e.g., 5 and 50), then the mean value and median will not have a special meaning because they will not represent an average case – in such set one deals with two different average cases. Let’s imagine a data set of potential customers. If their age distribution is bimodal, then it means that one deals with two different groups, for example, teens and elderly people, who probably differ from each other, and thus building a single model, especially linear or logistic regression model, on the basis of such data will be a bad solution. Graphic presentation of variables also allows to analyze dependencies between variables.

Data Representativeness An important factor, on which the success of implementing a machine learning solution depends, is the representativeness of source data. The source data more often is a trial rather than collection of information on a given subject (population). At first glance it may seem that the greater the trial is, the greater chance there is that it covers all types of modelled cases and therefore it represents the whole population better. However, there are two exceptions to this rule. Firstly, if data are collected in specific conditions, then the size of a trial should not exceed 10% of population. Why? Because the greater trial will reflect a local characteristic of data collection place or time more accurately. The generalization to a whole population will be unsubstantiated, e.g., when the studies were performed only in one company. Such trial is often called “burdened trial,” that is, that cases selected to it are not usual for a broader population. Secondly, the size of a trial influences the level of trust to study results. Therefore, it happens that in order to achieved a set trust level (e.g., 95%), observations are added to the experiment until the results achieve this level. In correctly conducted machine learning solution implementations, the size of trial is determined at the data evaluation and preparation stage. The assumption that source data constitute as trial means that aside from measurement error one also must take into account the method, in which this trial has been selected. There are two main types of measurement errors: 1. Regular error (bias), which over- or underestimates all measurements – for example the lengths of sections measured with inaccurate measure will be either under- or overestimated.

490

N. Rybak and M. Hassall

2. Accidental error (noise), which results in accidental change of measurements despite the fact that the measured size had the same value. If source data are burdened by such error, then models can contain dependencies not occurring in reality and returned predictions will be inaccurate and unreliable. Data evaluation methods presented in this chapter should be applied a. o. in order to detect a measurement error. In addition, the evaluation of accuracy and reliability for machine learning models conducted post factum by separating test data also allows to detect measurement errors that were overlooked earlier. The errors most often made during selection of trial include: preferring more easily accessed observations (e.g., cases from one country) and not taking missing observations into account (e.g., fact that certain devices did not send data or portion of people selected to participate in a survey did not answer questions). In order to protect ourselves from such errors, the data should be selected at random. There are three drawing methods: 1. Random selection of trial from population. 2. Division of data into subsets containing similar cases followed by random selection of trials from all subsets (layer selection method). 3. Division of data into homogenous clusters followed by random selection of trials from certain clusters. The problem of unfamiliarity with attribute values distribution in the whole population is solved thanks to the phenomenon of similarity to the real distribution. The phenomenon consists of increasing the size of variable values distribution trial until it will reach the same as the distribution of whole population. Increasing the trial size after crossing this border will not cause the change of distribution.

Section Conclusions Source data used in machine learning implementations must have a tabular form and they should be transformed to this form. The results of final model entirely depend on source data, in particular on their representativeness, thus their usefulness must be examined – models built on the basis of those data can have a great impact not only on companies’ financial decisions, but also on human health and life (Raghupathi and Raghupathi 2014). The evaluation of source data aims at better understanding them and in consequence building better machine learning models. Since the interpretation of machine learning models results can be difficult (e.g., complex models of reinforced decision-making trees, support vector machines, or deep neural networks), a detailed statistical evaluation of source data and exploratory data analysis are necessary and most important stages in implementing ML solutions.

19 Machine Learning–Enhanced Decision-Making

491

Type of Machine Learning Models The following section of the chapter describes the subsequent stage of implementing a machine learning solution – selection of model category and specific algorithms under a given category. It also discusses real world industry applications in the context of decision making for each presented main categories. All specific machine learning solutions fall into two main groups – supervised and unsupervised learning which will be discussed in the next subsections. The specific categories of machine learning models can be further divided into five categories: 1. 2. 3. 4. 5.

Classification Regression Clustering Prediction Anomaly detection

These 5 categories of machine learning will be discussed in subsections that follow the discussion of supervised and unsupervised learning.

Supervised and Unsupervised Learning As discussed above, the term machine learning is used in a very general form and concerns mostly general techniques of extrapolating trends in big data sets or ability to predict new data on the basis of what had been learned during the analysis of already known data. It is an exceptionally broad definition that covers several different techniques. Machine learning techniques can be roughly divided into two big groups: supervised learning and unsupervised learning. The first machine algorithm class is named supervised learning (Caruana and Niculescu-Mizil 2006; Reed and Marks 1999). It utilizes a set of labelled data in order to classify similar data without labels. Supervised learning assumes the presence of human supervision over the creation of function rendering the system input to its output. The supervision consists of creating learning data set, i.e., pairs: • Input learning object (e.g., vector). • Answer desired by the supervisor (e.g., specific numeric value). The purpose of such model is to learn predicting the correct answer to set stimulation and generalize cases learned as cases with which the system had no prior contact. Supervised learning is frequently used in industry to model technical processes. Insert example.

492

N. Rybak and M. Hassall

Labelled data are data that have already been classified while non-labelled data are such data that have not been classified yet. The second machine learning algorithm class is defined with name unsupervised learning (Baldi 2012; Barlow 1989; Ghahramani 2003). In such cases, data is not labelled in advance, but instead the algorithm is allowed to draw conclusions. One of the most often used and possibly the simplest examples of unsupervised learning is grouping also known as clustering. It is a technique, in which an attempt to divide data into sub-clusters is made. The algorithm divides data into sub-clusters, which constitute as various classes in the context of the data set. In order for the clustering technique to work, each element in each group (also called a cluster) should be characterized by high similarity degree within a class and low similarity degree in relation to other classes. Grouping can work for any number of classes and the idea of grouping methods such as k-means consists of finding k-source data subsets, the elements of which are closer (more similar) to each other than to any other element outside of their class . Obviously, in order to execute this type of task it is necessary to define what closer or more similar means, that is, it is necessary to define certain factor, which defines a distance between points. Unsupervised learning works thanks to the discovery of hidden internal data structures. Those structures, and not labels attached to them, allow a correct classification.

Classification Algorithms Classification in machine learning refers to the problem of predictive modelling, in which the class label is predicted, using mathematical algorithms, for a given example of input data. From the modelling point of view, classification requires a training data set with numerous examples of input and output data, which can be used for learning. Therefore, it is the most prominent group of supervised learning algorithms. The model will use the training data set and calculate the best way to render input data examples to specified class labels. Training data set must be sufficiently representative for a problem and contain multiple examples of each class label. Classification predictive modelling algorithms are evaluated on the basis of their results. The accuracy of classification is a popular measure used to evaluate a model performance on the basis of predicted class labels. There are the following distinct subgroups of classification algorithms: • • • •

Binary classification. Multi-class classification. Multi-symbol classification. Imbalanced classification.

19 Machine Learning–Enhanced Decision-Making

493

Binary Classification Binary classification refers to classification tasks which have two class labels. The examples of binary classification tasks are: • • • •

Clients’ credit capacity testing. Detection of spam. Product quality control. Process quality control.

Usually, the binary classification tasks cover one class that constitutes as a normal state and another class, which constitutes as abnormal state. Normal state classes are assigned to class 0 label, while abnormal state classes are assigned to class 1 label. A common practice is to model the binary classification tasks using a model, which predicts Bernoulli’s probability distribution for each example. Certain algorithms are specifically designed for a binary class and do not natively service more than two classes; the examples cover variation of logistical regression and support vector machines (SVM) algorithms.

Multi-class Classification Multi-class classification refers to classification tasks which have more than two class labels. The examples cover, e.g.: • Identification of people on the base of facial recognition. • Classification of emotions based on facial expression. • Classification of subject’s mental states based facial expression (e.g., detection of multiple stages of fatigue). • Classification of people activity type (aggressive person, calm person, etc.) • Optical sign recognition. • Optical character recognition. Unlike binary classification, multi-class classification does not include definitions of normal and abnormal results. Instead, the examples are classified to one of array of known classes. The number of class labels can be substantial in case of some problems. For example, the model can predict a photo as belonging to one of thousands or tens of thousands of faces in the facial recognition system. Problems related to the prediction of word sequence, such as text translation models, can also be recognized as a special type of multi-class classification. Each word in the sequence of predicted words covers a multi-class classification, in which the size of dictionary defines the number of possible classes that can be predicted and which can have the size of tens or hundreds of thousands of words. Part of the algorithms designed for binary classification can be adjusted for use in case of multi-class problems. It is related to the use of a strategy that focuses

494

N. Rybak and M. Hassall

on adjusting numerous binary classification models for each class in comparison to each other class or a single model for each class pair. In section “Example Application: Machine Learning-Based System for Real-Time Inattention Detection” (Case study), a closer look will be taken at instances of using advanced, deep learning-based fatigue detection system.

Multi-Symbol Classification Classification of multiple labels refers to those classification tasks that have two or more class labels, in which one or more class label can be predicted for each example. It differs from binary and multi-class classification, which predict only one class label for each example. A common practice is to model multi-symbol classification using a model that predicts multiple results, where each output is adopted as Bernoulli’s probability distribution (Weisstein 2002). Basically, it is a model that creates multiple binary classification predictions for each example. Classification algorithms used in both binary and multi-class classification cannot be used directly in multi-symbol classification. Another approach is the use of a separate classification algorithms to predict labels for each class. Imbalanced Classification Imbalanced classification refers to those classification tasks, in which the number of examples in each class is distributed unevenly. Usually the imbalanced classification tasks are binary classification tasks, in which the majority of examples in training data set belongs to a normal class while a small portion of examples belongs to the abnormal class. The examples cover: • Detection of frauds. • Prediction of rare events. • Medical diagnostic tests. These problems are often modelled as binary classification tasks although they may require special techniques. These special techniques can be used to change the composition of samples in a training data set through partial sampling of major class or excessive sampling of major class (Huang et al. 2016; Tang et al. 2008). Classification and clustering are the main ideas standing behind numerous other machine learning techniques and subjects. The ability to classify and recognize certain types of data allows to implement the techniques described above in other computer science areas, such as computer vision, natural language processing, signal processing, building predictive economic or market models, and other areas.

Regression Algorithms Regression is the second, next to classification, most commonly used in industry implementations machine learning method. Similarly, to classification, it is a

19 Machine Learning–Enhanced Decision-Making

495

supervised learning method. The purpose of regression is to find a model, which on the basis of known data will be able to calculate missing values sufficiently accurate. Input variables can be numerical or categorical. If output variable is numerical, then its will be calculated. If it is a categorical variable, then the prediction result will be the probability of given state’s occurrence. The simplest type of regression is linear regression while the simplest linear regression model is a model with one x input variable and one y output variable. In such case, the purpose is to find a function, which will return the y value on the basis of x value. Thus, the purpose of regression in machine learning is to find a function which will calculate output variable values as accurate as possible on the basis of input variables, or in other words, it will minimize the differences between calculated and real values: ei = yi – f(xi ) should be as small as possible. The θ (theta) cost function received using this method allows to calculate a difference between model and training data, that is, a difference between real output variable values and those calculated by the data exploration model (see Fig. 5). The regression error can be calculated as a sum of their absolute values or sum of their squares. In practice, the most often applied method is the sum of squared errors (SSE) because the cost of making an error shows a nonlinear (square) growth along with error value in numerous applications. For example, losses resulting from incorrect estimation of product delivery delay by 30 min are few times greater than those for 10-min error. The SSE value can be calculated from equation from J.H. Ward’s article (Ward 1963):

Fig. 5 An example of on output of regression model. X-axis presents observed (actual) numerical values and Y-axis presents values predicted by the regression model. Colors indicate model accuracy for each data point. Darker colors indicate more accurate prediction

496

N. Rybak and M. Hassall

SSE =

n i=1

xi2 −

2 1 n xi i=1 n

(1)

where n is the number of observations xi is the value of the i-th observation and 0 is the mean of all the observations. The lower the SSE value will be, the more accurate the model will be. Therefore, one should select the initial parameters in a way that minimizes the SSE value – this task can be automatically solved by the dedicated data exploration algorithms (Ward 1963). At this point it is worth to highlight a difference between prediction and extrapolation in regression models. Prediction consists of calculating the value of output variable for input variable values falling within training data values. Extrapolation is the use of the same model to calculate output variable value for yet input variable values falling outside training data values. The standard data exploration model evaluation methods consist of using test data to calculate an error, so therefore on their basis the accuracy of extrapolation results cannot evaluated – this is another reason why each implementation of solutions from the scope of machine learning should start from evaluation and data preparation. Linear regression should be used only when: • There is a linear dependency between input variables and output variable. • The distribution of errors is close to normal. • The error variability is fixed (homoscedasticity of errors). The applications of regression cover solving such problems as: • Problem of commercial agent, who wants to know the profit of partnership with a given customer. • Problem of service supervisor, who is interested in time necessary to repair a given device. • Problem of marketing department manager, who would like to know how many people will visit the company’s website within the next month. • Problem of pharmaceutical company, which would like to evaluate the influence of developed medication on the state of health of patients. • Problem of insurance company, which would like to define the risk of client’s death. • Problem of vehicle lending company, which would like to determine the number of damages on the basis of customer’s historical data. Regression models are also used for classification. Instead of assigning cases to specific classes (e.g., classifying potential borrowers as risky or trustworthy), the degree of risk related to granting a loan is evaluated (in a 0% to 100% scale). Such solution has a significant advantage, namely, it allows to organize cases in terms of prediction results, for example, sort the list of potential borrowers in such a way that the most trustworthy people will be shown first and the least trustworthy

19 Machine Learning–Enhanced Decision-Making

497

people will be shown last. Thus, regression not only allows to classify cases (e.g., people, for whom the risk of granting a loan exceeds 65%, will be described as risky), but also the choice of cases on the basis of their evaluation (e.g., loan will be granted to 55% of customers, who have the lowest risk). The output variable in regression models does not have to be numerical, e.g., Poisson’s regression models and sequence regression were created with a view of categorical sequence variables (Land et al. 1996). Neuron networks can be used for regression and their domain are problems, which require detection of numerous concurrent dependencies between various input variables in order to solve them. Specific layers of deep neuron networks can carry out specific tasks, such as filtering, convolution, or selection of data originating from previous network layer (Qiu et al. 2014).

Clustering Algorithms This section will focus on discussing clustering algorithms (cluster analysis). Unlike classification and regression algorithms presented heretofore, clustering algorithms are unsupervised learning algorithms. Unsupervised learning focuses on finding patterns in a data set without already existing labels and at minimal human intervention. Let us take a look at the idea of clustering as machine learning method shown on an example – organizing a scientific conference. Several applications from potential speakers have been already received when we realized that we forgot to divide speakers according to the subject of their presentations. Luckily, speakers provided in their applications information about themselves, including the name of institution where they work, titles of published scientific articles, and research interests. If that information could be used to create clusters, that is, groups of people with similar interests, then these clusters could be used to create specific conference topic streams. The advantage of clustering is the fact that such distribution takes all input variable into account at the same time. It is important because if one would try to manually divide speakers into groups, then the selection of some attribute is required, e.g., scientific discipline under which the speakers published the most papers, in contrast clustering algorithms will assign speakers to groups while taking any and all available information into account and then it will independently evaluate their impact on the clustering output result. Similarly, to other machine learning techniques, it must be however based on the Ockham’s razor principle. One should always bet on the simplicity of solution when the effectiveness of an algorithm meets the requirements. However, in case of clustering, the empirical data may be unavailable. When the clustering algorithms acts as a search tool, one can only assume that the data set reflects a precise process of generating data samples. If this assumption is true, then the best solution turns out to be the determination of number of clusters that maximizes the internal cohesion (density) and external separation. This means that occurrence of isles should be

498

N. Rybak and M. Hassall

expected, the key components of which contain certain mutual and partially unique properties. The two main groups of methods used for clustering tasks are: principal component analysis and cluster analysis. The principal component analysis is used to reduce the dimensionality of data through discovery and rejection of properties that carry the least amount of information. Cluster analysis is used to group and segment data sets with common attributes in order to extrapolate dependencies occurring in them. Cluster analysis identifies similarities in data and allows to group data that were not labelled, classified, or categorized. Because cluster analysis is based on the presence or lack of such similarities in new data, it can be used to detect anomalies – data, which do not belong to any group. The most important clustering algorithms are: • centroid algorithms also known as k-means, • k-nearest neighbors algorithm based on the KD trees and ball trees, • fuzzy c-means algorithm.

Centroid and k-Means Algorithms In order to properly understand the rules of using these algorithms, a new term has to be introduced – centroid. Centroid method has its roots in geometric decomposition; it is a representative of a given cluster or, in other terms, the center of a given group. If all variances of  i → 0, then the distribution will be reduced to the form of Dirac deltas that symbolize the perfect unit impulse centered in a specific point (Raju 1982). In such situation, one of the reasons to determine the correct cluster is to find the shortest distance between an example and centers of all clusters, that is, all centroids. This technique is also based on a so-called double rule that should be taken into account in case of every clustering algorithm. Clusters must be determined in a way that maximizes: • Cohesion within cluster core. • Separation of specific clusters. This means that the aim is to mark areas with high density that are clearly separated from other areas. If it is impossible to fulfil that condition, then one must move towards an attempt to minimize the average distance between examples and centroid within each cluster. This property is known as inertia . High inertia values suggest small cohesion because it is probable that too many points belong to clusters, the centroids of which are located too far. This problem can be solved by minimizing inertia. However, the computational complexity required to find this minimum is exponential (centroid algorithm is classified as NP-difficult problem). An alternative solution used in the centroids model is Lloyd’s algorithm (Lloyd 1982). It is iterative and starts from the selection of k random centroids followed by their correction until a stable configuration is achieved. Therefore, the centroid algorithm is intuitive and also turns out to be useful in numerous applications, however, two significant factors should be taken into

19 Machine Learning–Enhanced Decision-Making

499

account. The first factor is the convergence acquisition speed. It is easy to prove that each initial designation of centroids leads to a convergence point, but the number of progresses depends to a great degree on their selection and there is no guarantee that the global minimum will be found. If initial centroids are located close to end centroids, then the algorithm will require only few progresses to implement adjustments, but if their choice is completely random, then it will often result in a very large number of iterations. In case of N examples and k centroids, a Nk distance should be calculated in each progress, which in turn leads to small accuracy. The second relevant aspect is the fact that unlike the next algorithm analyzed in this section (KNN), in the centroid method one must immediately set the number of expected clusters. In certain situations, it is a secondary problem as the most appropriate value of the k parameter is already known. However, if the data set is multi-dimensional and our knowledge on possible number of groups is too limited, then the determination of clusters number might be risky. A good solution is to analyze the end inertia for various numbers of clusters. If the maximization of internal coherence is expected, then the small number of clusters will lead to increased inertia. The goal is to try to set the greatest value that is below the maximum tolerance threshold. In theory, one should also designate k = N. In such case we deal with zero inertia because each point becomes a centroid of its own cluster, however, large k values transform the cluster analysis scenario into a fragmented system that is not suitable for catching properties of a uniform group. It is not possible to define the rule for kmax upper limit, but it can be assumed that this value is always smaller than N. The best choice is done through the designation of k value, which minimizes inertia, for example from a scope between 2 and kmax .

K-Nearest Neighbors Algorithm(KNN) This algorithm belongs to a family of instance-based algorithms while the methodology is frequently called learning from examples. This category differs from remaining categories in the fact that it does not constitute as a real mathematical model. The inference process progresses through direct comparing of new examples with already existing ones (known as instances). The k-nearest neighbors algorithm is easy to adjust to the cluster analysis tasks. According to the KNN algorithm philosophy, similar examples should have common properties. For example, the recommendation system can group users with the help of this algorithm and for a new user it can search users, who are more like that user (e.g. on the basis of previously made purchases) in order to recommend the same product category. In general, the similarity function is defined as inversed distance or as cosine measure. Unfortunately, the classic KNN algorithms (also known as brute-force algorithms) can significantly slow down with great number of examples, because it is necessary to calculate all distances between point pairs in order to be able to generate the complete model. The number of those points for M number of operations is equal to M2 , which often constitutes as unacceptable value (e.g., if M = 1000, then each query requires calculating a million of distances). To be more precise, if calculating a distance in N-dimensional space requires N number of

500

N. Rybak and M. Hassall

operations, then the total computational complexity is equal to O(M2 N), which is allowed only for small values of both parameters (M and N). Thus, other methods, such as KD tress or ball trees, are used in the case of large number of instances in order to minimize the computational complexity. All queries in the KNN algorithm can be treated as search problems, thus one of the most effective ways to reduce the total complexity is to transform the data set into hierarchical structure. In the KD tree (single-dimension data), the mean computational complexity of a query is equal to O(logM) because one can assume that there are almost identical number of elements in each branch (if the tree is completely imbalanced, then all elements are implemented in a sequence and the final structure contains only one branching, which in turn results in the computational complexity adopting mentionedO(logM) form). To be precise, the actual complexity is slightly greater than O(logM), but this operation is always much faster comparing to the classic KNN algorithm, the complexity of which is O(M2 ). An alternative solution is the ball trees algorithm. The basis here is constituted by the concept of reorganizing a data set in such a way that this set will become almost impervious to multi-dimensional examples. Here, such ball can be defined as a collection of points, for which the distance from central instance is smaller or equal to a set radius. Starting from the first main ball, one can create in it socketed derivative balls and stop the process after achieving a set height. The basic condition is that the point must always belong to a single ball. This way allows the computational complexity to adopt the form of O(N logM) while taking into account the cost of n-dimensional distance and the algorithm itself does not deal with the curse of multi-dimensionality. The structure here is based on hyperspheres and therefore, the only operation required to search the appropriate ball is the measurement of distance between the instance and centers starting from the smallest balls. Both KD and ball trees can constitute as efficient algorithmic structures, which reduce the complexity of KNN algorithm queries. However, during model learning it is important to take into account both the k parameter (that usually defines mean and standard number of neighbors calculated in a given query) and maximum tree height. The discussed structures are not applied in common tasks (e.g., sorting) and their efficiency is maximized when all required neighbors are located in the same subordinate structure (with size K < < M in order to avoid a hidden use of the classic KNN algorithm). In other words, the role of a tree is to reduce the dimensionality of search space by dividing it into appropriately small areas.

Fuzzy c-Means Algorithm The differences between K-means and KNN clustering algorithms have been discussed. Therefore, the issue of clustering from the perspective of fuzzy logic can be explored (Pedrycz and Gomide 1998). The classic logic sets base on the law of excluded middle, which in terms of clustering can be formulated as follows: instance xi can belong only to one cj cluster.

19 Machine Learning–Enhanced Decision-Making

501

Contrary to other methods, techniques using the fuzzy logic allow to define asymmetric sets, which are not described using continuous functions (e.g., trapezoidal). This way one receives a greater geometrical flexibility. The fuzzy C-means algorithm constitutes as a generalization of classic centroids algorithm and is characterized by so-called soft allocation and flexible clusters. Unlike the centroids algorithm, the sum is not limited to points belonging to a defined cluster because the weight factor causes the contribution of furthest points (wij ≈ 0) will be close to zero. Contrary to the centroids method, no exclusions are imposed. Owning to this a given example can belong to any number of clusters with various affiliation degrees.

Prediction Regardless of utilized algorithm, the task of prediction models is to fill out missing data, that is, responding to predictive queries. For this purpose, models are created using historical data while counting on the fact that the data exploration algorithms used in them will learn to fill out missing data about new cases. For example, model built with the use of decision-making trees can classify new transactions as fraud attempts or legal transactions on the basis of their properties characterizing previous fraud attempts and properties characteristic for legal transactions. All predictive models concern the future in this meaning. Let us review the case of company selling insurance from accidents caused by engaging in winter sports, as an example of using a predictive model. It is known from experience that the greatest number of sold insurances is attributable to winter season, so one can prepare for increased demand during this season with appropriate advance. However, exactly how many insurance policies will be sold this year? It is also known that the company has the greatest turnover in June – however, to what degree an increased sale in first quarter will translate into sales results in other seasons of the year? What impact will have an increased sale on the sale of other insurances and worse sales results in one region of a country will influence the sales of identical insurance policies in other regions? Let us also assume that there is a plan to release a new type of insurance on the market and the aim is to evaluate the influence of that decision on the sale of remaining products. Problems of these types can be approached by using the predictive models. Machine learning algorithms presented so far did not take the order of observation into account in any way. Forecasting on the basis of time series consists exactly of predicting future values on the basis of previous variable values, or to be precise, on the way those values changed. The time series has a systematic and accidental component. The systematic components are described by the following parameters: • Trend, that is, a long-term tendency to single-direction changes (increase or decrease) of values or a fixed level, which occurs when the series has no trend and values fluctuate around a fixed level.

502

N. Rybak and M. Hassall

• Period, under which it is possible to differentiate cyclical fluctuations (fluctuations around a trend or fixed level that repeat in a set time period) and recurring (long-term fluctuations around a trend or fixed level). In the decomposition of time series, it is also helpful to calculate the autocorrelation factor (ACF)(Bence 1995; Hong and Satchell 2015): • If subsequent ACF values are getting closer to zero and then fluctuate around it, then the time series contains a fixed component. • If subsequent ACF values are getting closer to zero and then adopt negative values, then the series has a trend. • If subsequent ACF values are fluctuate around zero and then are significantly greater than zero after certain amount of time, then the time series contains seasonal fluctuations. Forecasting consists of calculating future variable values on the basis of its previous values. The simplest forecast model is linear auto-regression. It was already presented in the previous section that regression consists of estimating a variable value on the basis of values of other variables. The general regression formula is as follows: y = f (w, x, y, z) + ε

(2)

where y is the explained variable (dependent), w, x,y, z are explaining variables (independent), and ε represents the random component. A more advanced forecast method is the ARIMA (Autoregressive Integrated Moving Averages) model, which assumes a non-zero auto-correlation for random component, which quite often leads to more precise forecasts (Contreras et al. 2003; Hillmer and Tiao 1982). As the model’s name suggests that it is a combination of auto-regression (AR) model with moving average (MA) model. The ARIMA model has a chance to be precise if strong or at least partial auto-correlations occur in time series. Both auto-regression and ARIMA model forecast future values by summing their historical values. The use of artificial neural networks for forecast purposes allows to replace a nonlinear sum with activation function, such as, e.g., tanh. Nonlinear models received in such way are characterized by greater stability and accuracy of long-term forecasts, but short-term forecasts can be less accurate than in the case of using auto-regression or the ARIMA method (Zhang 2003). Forecasting requires an exceptionally thorough analysis and preparation of data. Otherwise, the acquired results can be unusable. It should also be taken into account that the accuracy of forecast models often is worse than accuracy of regression models. Therefore, if one has additional data at disposal, it is worth to consider the modelling of a problem using the nonlinear model.

19 Machine Learning–Enhanced Decision-Making

503

Anomaly Detection Anomaly detection , often related to detection of standing out elements, is a task focusing on identification of objects, events, or observations, which significantly stand out from the expected model occurring in uniform data sets. Anomaly detection concerns the prediction of unknown information. Each time invalid information inside data is found, it is usually called an anomaly or outlier element. Nevertheless, those two terms are often used alternatively, in fact they mean two different concepts. The outlier is a correct data point, which significantly differs from average or median within a distribution. It can be something very unusual, like, e.g., a modern passenger car using 30 liters of fuel during one-hundred-kilometer drive, but this result still falls within borders of reality. Anomaly is a datapoint or measurement generated by a process other than the process generating remaining data. The upkeep of a defined pattern of outlier is a signal that something has changed in the system that is monitored. Actual anomaly detection occurs in the case of observing regular deviations in the source data generation process. This also has implications during initial data processing. Contrary to what is done during solving of numerous machine learning problems, one cannot simply filter all standing out elements in case of anomaly detection. Nonetheless, caution should be exercised during effort to determine the nature of such elements. An effort to filter incorrect data elements, remove noises, and normalize the remaining data should be made. The final goal should be an attempt to discover novelties in processed data. Anomalies can occur in every system. From the technical point of view, it is always possible to find an event, which was not seen before and which does not occur in historical data of a given system. Detection of such event in certain contexts may have great consequences (both positive and negative). Anomaly detection can be used to recognize activities of potentially criminal nature. Anomaly detection in a network detection can help in search for external breakin or suspicious user activity – e.g., activity of employee, who accidentally or intentionally introduces great amounts of data outside of the company’s intranet. Anomaly detection mechanisms can also be used to detect actions of a hacker opening connections on nonstandard ports and (or) protocols. Anomaly detection can be used in a specific Internet safety case to stop the spread of new malware applications. For this purpose, all that has to be done is to take a look at increases in movement at untrusted domains. Another similar example are verification systems implemented in online social platforms or by financial institutions. Dedicated software security teams can develop machine learning solutions to measure each action or action sequences and evaluate how far they are in comparison to the behavior median of other users. The system will ask for additional verification every time the algorithms marks a specified action as suspicious. The usage of these techniques may significantly reduce the phenomenon of personal data theft and provide better privacy protection.

504

N. Rybak and M. Hassall

The discovery of inconsistencies generated by human behavior is one of the most popular applications of anomaly detection mechanisms, but at the same time it is one of the most challenging. On one hand there are domain experts, data engineers, and machine learning engineers designing advanced anomaly detection systems. On the other hand, there are hackers, who understand the rules of the game and examine opponent’s moves. Therefore, systems of such type require a great deal of knowledge about given field and should be designed in a proactive and dynamic way. Anomaly detection systems operated manually (e.g., by analysts) do not scale well and usually are not general enough. Deviations from normal behavior are not always obvious. Sometimes the analyst might have difficulties with memorizing the entire historical data to provide effective assessment, which is the base of anomaly detection. The situation gets even complex if the anomaly pattern is hidden within abstract and nonlinear data relations. The demand for intelligent and fully automated systems capable of learning on complex interactions and ensuring a precise realtime monitoring is the next area for innovations in the domain of machine learning.

Example Application: Machine Learning-Based System for Real-Time Inattention Detection In this final part of the chapter, an in-depth examination of a case study is presented. The case study discusses step by step implementation of machine learning model to solve real world industry problem – automated detection of workers’ fatigue. The application is based on authors’ previously published research on deep learningbased face analysis (Rybak 2020; Rybak and Angus 2021; Rybak et al. 2017a).

Aims This project pioneered the development of a prototype, deep learning-based technology for measuring and reporting an individual’s situational attention, with the aim of using this technology to enhance real time risk assessment to prevent inattention and inattention-related accidents and injuries in high hazard industries. Measuring workers’ situational attention should deliver improved decision making, safety, and production related outcomes for various industries.

Background The importance of managing inattention, distraction and fatigue has been recognized by industry. For example, a mining industry report published in 2007 identified that over 65% of vehicle accidents and injuries in mining processes are directly related to drivers’ fatigue and drowsiness (Mining 2007). Similarly, driver distraction has been reported as a contributing factor in both fatal and serious injury crashes on public roads (Lee et al. 2008).

19 Machine Learning–Enhanced Decision-Making

505

Significant amounts of work have gone into driver and heavy equipment operator fatigue detection and management. However, research into fatigue and distraction detection for other safety critical decision-making is yet to be explored to the same degree in other safety critical industries such as the aviation industry (Morris et al. 2018), healthcare industry (Gander et al. 2007), and maritime industry (Jepsen et al. 2015). Safety and production critical roles such as control room operator, inspector, dispatcher, maintainer, and lone worker often require personnel to move and divert attention between competing cues across different spatial distances, timeframes, and levels of details. Measuring the attention/inattention for these non-driver, safety critical roles has additional challenges to measuring driver fatigue and drowsiness. Firstly, the head and eye positions of drivers are relatively fixed in space and over time whereas other safety critical roles may be required to move eyes, heads and bodies to monitor processes and obtain the information required to make decisions. In addition, drivers are often in clean air cabins where the light levels are quite predictable whereas other safety critical roles may incur conditions with variable levels of light. Mental and physical exhaustion of workers, which is a category of general fatigue, constitutes a significant risk in many high hazard industries. Such exhaustion may be a result of insufficient amount of sleep or the necessity to remain concentrated for long periods of time while under increased stress, which is either caused by having to perform highly complex operations in difficult conditions or the monotony that sets in while performing laborious, repetitive tasks. This state manifests itself as fatigue, the typical symptoms of which are sleepiness, lowered concentration and alertness, slower reactions, and impaired perception (Krueger 1989), (Caldwell et al. 2018). Using well-known, efficient bio-signal analysis methods (such as EEG, ECG, and GSR) for the assessment of the aforementioned symptoms, e.g., an EEG electrode cap that would show statistically significant changes in alpha and theta brain rhythms during a state of decreased alertness (Eoh et al. 2005). However, it is impractical to use such technologies in the typical work conditions of high hazard industries due to the difficulty wearing body sensors and in measuring signals in many work contexts. A much more practical approach in terms of simplicity of obtaining diagnostic information involves using inattention and fatigue symptoms visible on the face of the examined person, which can be obtained with computer vision methods and – on the classification level – machine learning techniques. This form of process for fatigue recognition has been a subject of interest in the field of computer vision for over 20 years. The first studies from 1990s describe a methodology that relies on eye activity and uses a measure called PERCLOS (PERcentage of eyelid CLOSure over time) to detect fatigue (Grace et al. 1998; Tock and Craw 1996; Veeraraghavan and Papanikolopoulos 2001). Despite the many innovations that have appeared in subsequent years, the methodology was always based on standard processing methods, a major element of which was “feature engineering” that required an expert description of the image with a vector of parameters, which was then used to create a detector.

506

N. Rybak and M. Hassall

The most common feature extraction methods in fatigue monitoring use a geometric approach, where the information on the state of the face is obtained by locating its specific shapes (e.g., corners of the mouth during a yawn) and characteristic points (Alioua et al. 2014; Lu and Wang 2007; Tang et al. 2016). Geometric features may include distances between such points, angles created by lines linking these points or distances between the points with a specific facial expression and the points on the reference image with a neutral face expression. Such features are popular due to their easy interpretation and calculation speed (given that they are detected in the image). An alternative approach involves using features that determine appearance of the face area, which contain pixel intensity values. An example of such is the holistic approach that treats the entire image as a pixel matrix, which is used to create numerical feature vectors by means of data reduction methods known from biometry, e.g., eigenfaces, which relies on Principal Component Analysis (PCA) (Bergasa et al. 2008). In yet another approach, feature calculation is based on pixels within extracted areas that contain characteristic elements of facial expressions. Here, LBP (Local Binary Pattern) texture detector should be mentioned, where the feature vector is a histogram of the image created by means of thresholded description of a single pixel’s neighborhood in the form of 0 s and 1 s (Y. Zhang and Hua 2015). For feature generation, classic image filtering method with predefined masks with various scales and orientations is also frequently used, e.g. Gabor filter masks (Fan et al. 2008). Due to the fact that potential users of a fatigue detection system are highly different from one another, e.g., in terms of age, facial features, or facial hair, an arbitrary expert description resulting from the aforementioned “feature engineering” will provide relatively unreliable outputs. The innovative approach proposed in this research project suggest using an entirely different image processing technique for the above-mentioned purpose, namely, automatic search of distinct information using a convolutional neural network. Such an approach falls within the subject of deep learning, which is currently the most advanced area of research in the field of modern machine learning (Schmidhuber 2015). A typical instance of applied deep learning is the Convolutional Neural Network (CNN) applied to computer vison tasks like facial expression recognition which automatically learns discriminating patterns from the regions of images by sequentially assembling layers vertically. Lower levels layers of the trained network represent configurations of corners and edges, whereas high level layers learn more abstract relations between shapes to discriminate between patterns. In many applications, CNNs are now considered the most powerful image classifiers (Goodfellow et al. 2016).

Machine Learning-Based Methods The Convolutional Neural Network method uses a different approach for locating patterns within an image. In place of hand-engineered instructions to extract specific features, these features are repeatedly learned from the training process completed

19 Machine Learning–Enhanced Decision-Making

507

on thousands of examples for each instance. This approach is a type of multi-class classification discussed in section “Multi-class Classification.” More importantly, this approach is based on multi-level representation learning, achieved by combining relatively simple nonlinear segments that each transform the representation at one step (starting with the raw input of pixels) into a representation at a higher, more abstract level. This work extended upon previous research which found that highly efficient retrieval of information regarding the operator’s psycho-emotional states in the control room is made technically feasible by continuous visual and acoustic data acquisition and real-time machine learning-based analysis and pattern recognition (Rybak et al. 2017b). The development of CNNS for detailed classification of facial expressions changes are currently key challenges in the fields of computer vision and machine learning. This project focuses on the development and verification of machine learning techniques, and it is anticipated that results from the project would form the basis for future development of technologies targeted to specific application areas (e.g. aviation, maritime, media, health industries).

Previous Machine Learning-Based Approaches There is a number of studies devoted to the use of convolutional networks in vision systems designed for monitoring subject’s facial expressions. Such networks are used both for automatic detection of face in the image and subsequent identification of fatigue, its particular symptoms or driver’s inattention (Yuen et al. 2016a, 2016b; Zhang et al. 2015). A popular application of convolutional networks mentioned in the literature is the detection of face rotation angles in different planes and thus recognition of the direction of sight, which allows for detecting distraction or sleep state (Rangesh et al. 2018; Ribarie et al. 2010; Vora et al. 2017, 2018). Using this approach also allows for detecting yawning (Ribarie et al. 2010) and closed eyes (Yan et al. 2015) based on images of face areas that represent mouth or eyes, respectively. Furthermore, one study (Yan et al. 2015) proposed using convolutional networks for detecting such activities as eating or talking on the cell phone, which may distract the driver. A slightly different approach is based on the systems in which the convolutional network was supposed to differentiate between the state of fatigue and nonfatigue (binary classification) instead of detecting specific symptoms of fatigue to perform multiclass classification (Dwivedi et al. 2014; Huynh et al. 2016; Park et al. 2016). In the study presented by Dwivedi et al., a relatively shallow convolutional network is implemented, and the input data consists of a single image instead of time series representations. Park et al. proposed a method of fatigue assessment based on three simultaneously operating convolutional neural networks with different structures that are used to process the same image. The result of classification is determined by integrating output data of individual networks. Huynh et al. introduced a three-dimensional convolutional neural network to process a sequence of multiple images at the same time instead of a single image. This approach allows

508

N. Rybak and M. Hassall

for assessing fatigue based on features established for the entire sequence of camera image frames, which definitely seems to be a reliable solution. It is, however, likely to be significantly more computationally expensive when compared to methods using two-dimensional networks. Its other disadvantage seems to be the requirement of using high-quality and well-curated video sequences.

Deployment Approach The presented study developed a CNN-based method of processing face images based on the symptoms of fatigue recognized in the literature, which are signaled by specific facial expressions in certain regions of face. It was first established which of the symptoms described in the literature are the most effective with proposed automatic classification technique, secondly authors identified the degree of occurrence of these established symptoms and then adjusted the CNN attention to selected facial regions. Therefore, it should be noted that presented case study also encompasses issues related to physiology and psychology, e.g., establishment of a measure of actual fatigue based on the frequency of symptoms while considering the possibility that these symptoms may not appear in some people. The research material consists of images captured using popular inexpensive cameras fitted with a lens designed to capture image in visible light range (Ghoddoosian et al. 2019). The dataset was subsequently pre-processed utilizing a narrowband filter adjusted dynamically to the wavelength of the illumination. The selection of the dataset including this spectral range and image capturing setup is dictated by the necessity to monitor faces in different light conditions (day and night lighting conditions) in possible future system industry implementations. The preliminary results obtained in the task of recognizing images representing neutral face expressions suggest that a properly designed convolutional network is able to automatically – only based on the provided images – extract facial regions and features that are further used for decision making. Up to this point, the learning process in such research projects used entire face images obtained from a subset of materials provided by project participants, which were then cut out using the Viola-Jones method. The structure of a convolutional network is based on layers, the amount and size of which depend on the complexity of the recognition task (Krizhevsky et al. 2012; Schmidhuber 2015). Due to the nature of the input data (image data) and currently used training methods, it is common to use models with dozens of layers, the selection of which in the described task of recognizing three classes (normal state, maximum fatigue state, and intermediate state) is defined dynamically by the size of the processed images and receives raw values of their pixels, which are then exposed in the learning process along with information on their attribution to certain classes. Layer size is the height and width of input images, which must be made uniform at the preliminary preparation stage.

19 Machine Learning–Enhanced Decision-Making

509

The inputs of each neuron are connected to a changing number of pixels of the input image that create local, rectangular receptive fields of the defined dimensions or to a specific number of elements in the preceding layer. At the training stage, each set of neurons is assigned a specific weight, which is identical for every neuron belonging to that set. Therefore, image processing performed by a trained convolutional layer can be equivalently replaced with image filtration by means of convolution, which uses neurons (filters) with defined weights (coefficients). The set of all images created through the process of convolution will constitute a feature map. The result of the operation performed a convolutional layer may include negative values, which is eliminated by a nonlinear activation function, ReLU (Rectified Linear Unit). This function leaves out all positive pixel values, only changing negative values into zeros. Its presence in deep learning networks is justified by the necessity to minimize the calculation of values in classic activation functions, which is the case, e.g., for the sigmoid function. Networks that use a ReLU-type activation function are trained several times quicker than networks based on classic functions (Glorot et al. 2011). The operation of combining multiple results from a neighborhood in feature maps into a single feature is performed by a reducing layer. This layer performs statistical filtration within a mask with the defined dimensions, providing a specific statistic such as maximum value (so-called MaxPooling) and mean value. All masks that process image in this way do not overlap. Although no training process occurs in this layer, the possibility to select an appropriately large step for mask shifting allows for reducing the dimensions of feature maps (minimized feature maps), which helped to improve training speed. Furthermore, if another convolutional layer is applied after it, the reducing layer gives the neurons a possibility to cover a larger area of the image and thus form additional feature representations. The above proposed network architecture also demonstrates high resistance to shifts of reference images, which was already verified in authors’ previous study (Rybak et al. 2017a). The last pooling layer will be an input layer for several neuron layers, with each neuron being connected to all neurons (outputs) of the preceding layer (a so-called Fully Connected layer). The last layer of this type aggregates information from all preceding layers and closes the structure of the automatic feature generator. The number of its output nodes is equal to the number of recognized classes. In order to provide a classification decision, the output fully connected layer is subjected to activation with a function that is formally identified as a separate layer called SoftMax. The SoftMax function calculates the probability of belonging to a given category for every neuron from the last fully connected layer, which in this case means belonging to one of three classes describing the fatigue level observed on the face. In order to include the possibility of occurrence of over a dozen of identified fatigue symptoms at the same time, the solution simultaneously uses multiple independent detectors, with each detector making a decision for a symptom it is

510

N. Rybak and M. Hassall

assigned to react to, i.e. detecting the degree of intensity of a given symptom or diagnosing the normal state. The reduction of the number of symptoms in the training process occurs only if it does not result in decreased precision/sensitivity of the system and at the same time it speeds up training and detection. The project includes an attempt to train and verify the adopted network structure in a sequential, 10-cross validation manner. In this way, it was possible to obtain an independent result of detection for all subjects. The assessment of the degree of network training and its generalization ability with regard to processing the testing data was performed by means of an error matrix, also known as confusion matrix. The error matrix is a square matrix, where the rows represent the established classes while the columns represent the decisions made by the classifier. Elements of this matrix can be expressed as absolute values or percentages. Such an approach, unlike global measures, reliably includes the instances of confusing the established symptoms with the normal state. To obtain the fastest operation speed at the stage of recognizing new cases – real-time operation – each detector adopts the simplest network structure that gives the best possible precision of recognition. This is achieved by means of a method known as pruning a trained network. It is feasible to implement such a pruned model on mobile device (see Fig. 6). Since it is common to see the phenomenon of overtraining in networks programmed to operate with a large number of iterations, the training process was organized so that it stopped automatically when the average rate of successful recognitions for 20 iterations reaches at least 95%.

Fig. 6 The CNN model for fatigue detection can be implemented on mobile devices

19 Machine Learning–Enhanced Decision-Making

511

Authors hope that developed system will be implemented for real world operation, where it will detect and identify the established symptoms in online mode. The observed short-term false alarms appearing due to the natural facial expressions appearing during normal work, conversation, etc. will be used for further development of the system and improvement of classification effectiveness in the field conditions. Based on previous studies, it is expected that the aforementioned facial expression may temporarily produce an appearance similar to the detected fatigue symptoms, but is it possible to eliminate its influence by filtering a sequence of probabilities established for the subsequent frames by the SoftMax layer.

Application Outcomes There are three classes of states provided as labels in utilized dataset (Ghoddoosian et al. 2019): • Alert: Subjects were advised that being alert meant they were experiencing no signs of fatigue or drowsiness (label 0). • Low Vigilant: This class encompasses subtle instances when some symptoms of fatigue appear, or fatigue is present but no effort to keep alert is required (label 1). • Drowsy: This state means that the subject needs to actively try to not fall asleep (label 2). Owing to the fact that there are no well-defined, gold standard evaluation procedures for fatigue detection systems, the model is compared with the results presented in the original article introducing utilized samples. The system achieved an averaged accuracy of over 86%. This is an improvement over the highest accuracy presented in baseline study which was 82% accuracy reached on label 2 samples (Ghoddoosian et al. 2019).

Section Conclusions This case study presented machine learning pipeline to build an industry specific prototype software, scientifically tested to demonstrate its ability to detect, assess, and provide feedback to operators on their level of situational attention and, as a result, enhance decision-making process. The solution is significant in that it has the potential to improve situational awareness which should lead to fewer accidents and improved production outcome and therefore assist in maximizing the safety in the high hazard industries. The proposed system will operate in a non-interruptive, non-intrusive, and noncontact manner, and it will not require changes in the operators’ environments as it can utilize already installed equipment (cameras and mobile devices).

512

N. Rybak and M. Hassall

Summary Machine learning involves using computers to automatically process and analyze large data sets. This approach can be applied not only by research centers, but also by companies operating in a wide variety of sectors from private to public and manufacturing sectors to trade. As long as relevant data are at disposal, properly conducted exploration allows to gain knowledge that would otherwise be not available. Furthermore, the accuracy and reliability of the rules obtained by means of the machine learning-based systems can be measured and objectively assessed before they are applied. Determining the validity and understanding the uncertainty of the outputs is also important information for decision makers to know. In summary, a typical process of building machine learning model consists of the following stages that were described in the sections “Data Pre-processing” and “Type of Machine Learning Models” of this chapter: • Defining the problem to be solved in the context of the data available. • Collecting the necessary source data – in addition to data warehouses, dedicated big data systems or social networking platforms are becoming increasingly popular sources of data. • Assessing the collected data in terms of its usefulness for solving the problem raised. • Preparing (removing excessive and redundant information) and augmenting the source data for specific data exploration models. • Designing and building machine learning models. • Evaluating the models created and selecting the best of them in relation to the established criteria. • Deploying the chosen model to solve the problem. There are some problems that cannot be solved in an algorithmic manner. For example, the image recognition – one still cannot describe precisely how this process occurs in the human brain, and without knowing the algorithm, it is not possible to implement it in any programming language. However, this does not mean that it is not possible to teach a computer to recognize images, which is evidenced by the case study described in this work. Attempts to algorithmically solve many other real-world problems would not only be very complicated but also highly ineffective in practice. A good example of such problems is the issue of detecting possible financial fraud. Even after successfully encoding all the rules under which a transaction should be classified as an attempted fraud, one would get an extremely complicated software system (in addition to the general rules, account rules specific to the person concerned should be taken into account). What is worse, rules of this kind often change (users do not behave in the same way at all times and fraudsters adjust their actions to circumvent the implemented safeguards), so if one would want the software to be effective, it would require to update it in real-time.

19 Machine Learning–Enhanced Decision-Making

513

Another type of problem that cannot be solved in a purely algorithmic manner are the systems that are unscalable, either because of the amount of data processed or the complexity of the algorithm. An example can be the recommendation systems commonly used by all online stores. The solution to all these problems is to use the machine learning approach. Instead of writing algorithms that would perform specific tasks in a specific way, companies can collect data that describes the expected results (images with a description of their content, sets of transactions with information on whether or not they were an attempt to fraud or the history of purchases made by customers) and use it as training data for an appropriate machine learning model. The model thus created will respond to such predictive queries as “identify the objects present in the image,” “assess the risk of a transaction being an attempt to fraud,” or “recommend to a given user the goods that are most likely to be of interest to the client.” In this chapter authors have presented the methods of creating such systems, with particular emphasis on data exploration methods as the most important stage of any machine learning solution application. It seems that after overcoming multiple technical limitations, a small number of machine learning specialists – expert programmers with appropriate background in the field of advanced statistics and geometry, able to build from scratch the architecture of machine learning models, train models, evaluate their learning progress, and implement them to solve practical problems in industry – have become the main inhibitor of machine learning adoption in industry. This shortage of machine learning specialists frequently results in machine learning development tasks being assigned to software engineers without appropriate background in mathematics as well as information and knowledge theories. With the advent of machine learning solutions industry implementation, a change in the role of domain experts in the machine learning engineering process has been witnessed. Despite not having formal training, domain experts may find themselves collaborating on machine learning models’ development. An example of challenges in collaboration between machine learning engineers and domain experts is a large discrepancy in terminology used by both groups. This problem can be in part tackled by starting any interaction with the introduction outlining terminology originating in machine learning that is commonly used in the development of ML-based systems. Secondly, domain experts employ various software artefacts such as cloudbased micro-services, frequently without a clear picture of their purpose and without adequate knowledge on the internal workings of such solutions. This may cause unexpected and unwanted outcomes when it comes to decision making processes. The adoption of understanding of machine learning-based software development as well as sound knowledge on statistical methods by domain experts is thus crucial for the efficiency and safety of operations. Therefore, industry-based education strategies aiming at strengthening the cooperation between machine learning engineers and domain experts, by offering domain experts a high-level understanding

514

N. Rybak and M. Hassall

of machine learning practices and artifacts, would enable more effective and safer implementations of ML models to enhance decision making.

Glossary ARIMA

Algorithmics Centroid

Entropy Extrapolation

Labelled data Prediction

Probability Distribution

Autoregressive Integrated Moving Average – class of algorithms for forecasting a time series. As the model’s name suggests that it is a combination of auto-regression (AR) model with moving average (MA) model. Subfield of computer science. Study of design and efficiency of algorithms. Machine learning method that has its origins in geometric decomposition; it is a representative of a given cluster or, in other terms, the center of a given group. In information theory, average amount of information attributable to a single message from information source. It is the use of the same machine learning model to calculate output variable value for input variable values falling outside training data values. In machine learning, data that have already been categorized. In machine learning, calculating the value of output variable for input variable values falling within training data values. Function used to compute the likelihoods of occurrence of various observation outcomes.

References Alanis AY, Arana-Daniel N, Lopez-Franco C (2019) Artificial neural networks for engineering applications. Academic Press, Cambridge, Massachusetts Alioua N, Amine A, Rziza M (2014) Driver’s fatigue detection based on yawning extraction. Int J Vehicular Technol, vol. 2014, Article ID 678786, pp. 7 Al-Shayea Q, El-Refae G, Yaseen S (2013) Artificial neural networks for medical diagnosis using biomedical dataset. Int J Behav Healthcare Res 4(1):45–63 Amato F, Lopez A, Pena-Mendez EM, Vanhara P, Hampl A, Havel J (2013) Artificial neural networks in medical diagnosis. Journal of Applied Biomedicine, Elsevier. 11(2):47–58 Baldi P (2012) Autoencoders, unsupervised learning, and deep architectures. Paper presented at the Proceedings of ICML workshop on unsupervised and transfer learning Barlow HBJN c. (1989) Unsupervised learning. 1(3):295–311 Bekey GA, Goldberg KY (2012) Neural networks in robotics, vol 202. Springer Science & Business Media. New York Bence JRJE (1995) Analysis of short time series: correcting for autocorrelation. Journal of Ecological Society of America, Washington, 76(2):628–639

19 Machine Learning–Enhanced Decision-Making

515

Bergasa LM, Buenaposada JM, Nuevo J, Jimenez P, Baumela L (2008) Analysing driver’s attention level using computer vision. Paper presented at the Intelligent Transportation Systems, 2008. 11th international IEEE conference on ITSC 2008 Bibri SEJJ (2019) On the sustainability of smart and smarter cities in the era of big data: an interdisciplinary and transdisciplinary literature review. 6(1):25 Branch MAI (2016) Report on the investigation into the listing, flooding and grounding of Hoegh Osaka–Bramble Bank, the solent, UK on 3 January 2015. Marine Accident Investigation Branch, Southampton Caldwell JA, Caldwell JL, Thompson LA, Lieberman HR (2018) Fatigue and its management in the workplace. Neurosci Biobehav Rev. 96:272–289 Caruana R, Niculescu-Mizil A (2006) An empirical comparison of supervised learning algorithms. Paper presented at the Proceedings of the 23rd international conference on machine learning Chatfield C, Xing H (2019) The analysis of time series: an introduction with R. Chapman and Hall/CRC Press, Boca Raton, Florida Cichocki A, Unbehauen R (1993) Neural networks for optimization and signal processing. Wiley, Chichester. Google Scholar Digital Library Contreras J, Espinola R, Nogales FJ, Conejo AJJI (2003) ARIMA models to predict next-day electricity prices. 18(3):1014–1020 Dalal N, Triggs B (2005) Histograms of oriented gradients for human detection. Paper presented at the 2005 IEEE computer society conference on computer vision and pattern recognition (CVPR’05) Damper R (1997) A biocybernetic simulation of speech perception by humans and animals. Paper presented at the 1997 IEEE international conference on systems, man, and cybernetics. Computational cybernetics and simulation Dwivedi K, Biswaranjan K, Sethi A (2014) Drowsy driver detection using representation learning. Paper presented at the 2014 IEEE international advance computing conference (IACC) Eoh HJ, Chung MK, Kim S-H (2005) Electroencephalographic study of drowsiness in simulated driving with sleep deprivation. Int J Ind Ergon 35(4):307–320 Fan X, Yin B-C, Sun Y-F (2008) Dynamic human fatigue detection using feature-level fusion. Paper presented at the international conference on image and signal processing Fernández-Delgado M, Cernadas E, Barro S, Amorim DJT (2014) Do we need hundreds of classifiers to solve real world classification problems? 15(1):3133–3181 Gander PH, Purnell HM, Garden A, Woodward A (2007) Work patterns and fatigue-related risk among junior doctors. Occup Environ Med. 64(11):733–738 Ghahramani Z (2003) Unsupervised learning. Paper presented at the summer school on machine learning Ghoddoosian R, Galib M, Athitsos V (2019) A realistic dataset and baseline temporal model for early drowsiness detection. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition workshops Glorot X, Bordes A, Bengio Y (2011) Deep sparse rectifier neural networks. Paper presented at the Proceedings of the fourteenth international conference on artificial intelligence and statistics Goodfellow I, Bengio Y, Courville A, Bengio Y (2016) Deep learning, vol 1. MIT Press, Cambridge Grace R, Byrne VE, Bierman DM, Legrand J-M, Gricourt D, Davis B, ... Carnahan B (1998) A drowsy driver detection system for heavy vehicles. Paper presented at the Digital avionics systems conference, 1998. Proceedings, 17th DASC. The AIAA/IEEE/SAE Haykin S (1994) Neural networks: a comprehensive foundation. Prentice Hall PTR, Hamilton, Ontario, Canada Hertz J, Krogh A, Palmer RG, Horner H (1991) Introduction to the theory of neural computation. Phys Today 44:70 Hillmer SC, Tiao GC (1982). An ARIMA-model-based approach to seasonal adjustment. 77(377):63–70 Hong K, Satchell SJQF (2015) Time series momentum trading strategy and autocorrelation amplification. 15(9):1471–1487

516

N. Rybak and M. Hassall

Huang C, Li Y, Loy CC, Tang X (2016) Learning deep representation for imbalanced classification. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition Huynh X-P, Park S-M, Kim Y-G (2016) Detection of driver drowsiness using 3D deep neural network and semi-supervised gradient boosting machine. Paper presented at the Asian Conference on Computer Vision Jain LC, Vemuri VR (1998) Industrial applications of neural networks, vol 3. CRC Press, Boca Raton, Florida Jepsen JR, Zhao Z, van Leeuwen WM (2015) Seafarer fatigue: a review of risk factors, consequences for seafarers’ health and safety and options for mitigation. Int Marit Health 66(2):106–117 Krizhevsky A, Sutskever I, Hinton GE (2012) Imagenet classification with deep convolutional neural networks. Paper presented at the advances in neural information processing systems Krueger GP (1989) Sustained work, fatigue, sleep loss and performance: a review of the issues. Work Stress 3(2):129–141 Land KC, McCall PL, Nagin DSJSM (1996) A comparison of Poisson, negative binomial, and semiparametric mixed Poisson regression models: with empirical applications to criminal careers data. 24(4):387–442, Research Lee JD, Young KL, Regan MAJD (2008) Defining driver distraction. Theory, effects, & mitigation. 13(4):31–40 Liu Y, Zhao T, Ju W, Shi SJJ (2017) Materials discovery and design using machine learning. 3(3):159–177 Lloyd SP (1982) Least squares quantization in PCM 28(2):129–137 Lu Y, Wang Z (2007) Detecting driver yawning in successive images. Paper presented at the Bioinformatics and Biomedical Engineering, 2007. The 1st international conference on ICBBE 2007 Mining, CGJCGMReJ (2007) Viewpoint perspectives on modern mining 21. Mining industry by the Caterpillar Global Mining. http://www.viewpointmining.com/st/mine-features Morris MB, Wiedbusch MD, Gunzelmann G (2018) Fatigue incident antecedents, consequences, and aviation operational risk management resources. Aerosp Med Human Perform 89(8):708– 716 Nixon M, Aguado A (2019) Feature extraction and image processing for computer vision. Academic Press, Cambridge, Massachusetts Park S, Pan F, Kang S, Yoo CD (2016) Driver drowsiness detection system based on feature representation learning using various deep networks. Paper presented at the Asian conference on computer vision Pedrycz W, Gomide F (1998) An introduction to fuzzy sets: analysis and design. MIT Press, Cambridge, Massachusetts Qiu X, Zhang L, Ren Y, Suganthan PN, Amaratunga G (2014) Ensemble deep learning for regression and time series forecasting. Paper presented at the 2014 IEEE symposium on computational intelligence in ensemble learning (CIEL) Raghupathi W, Raghupathi VJH (2014) Big data analytics in healthcare: promise and potential. Health Inf Sci Syst 2(1):3 Raju CJJ (1982) Products and compositions with the Dirac delta function. J Phys A Math Gen 15(2):381 Rangesh A, Deo N, Yuen K, Pirozhenko K, Gunaratne P, Toyoda H, Trivedi MM (2018) Exploring the situational awareness of humans inside autonomous vehicles. Paper presented at the 2018 21st international conference on intelligent transportation systems (ITSC) Reed R, Marks RJ II (1999) Neural smithing: supervised learning in feedforward artificial neural networks. MIT Press, Cambridge, Massachusetts Rezzoug N, Gorce P (2003) A biocybernetic method to learn hand grasping posture. Kybernetes, 32(4):478–490

19 Machine Learning–Enhanced Decision-Making

517

Ribarie S, Lovrenete J, Paveene N (2010) A neural-network-based system for monitoring driver fatigue. Paper presented at the MELECON 2010-2010 15th IEEE Mediterranean electrotechnical conference Rybak N (2020) Technical considerations for the application of deep learning methods for multimodal emotion recognition. School of Information Technology and Electrical Engineering, The University of Queensland, Australia, pp. 223. https://doi.org/10.14264/bf3427d Rybak N, Angus DJ (2021) Tracking conflict and emotions with a computational qualitative discourse analytic support approach. PLoS One. https://doi.org/10.1371/journal.pone.0251186 Rybak N, Hassall M, Parsa K, Angus DJ (2017a) New real-time methods for operator situational awareness retrieval and higher process safety in the control room. Paper presented at the 2017 IEEE international systems engineering symposium (ISSE) Rybak N, Hassall M, Parsa K, Angus DJ (2017b) New real-time methods for operator situational awareness retrieval and higher process safety in the control room. Paper presented at the 2017 IEEE international systems engineering symposium (ISSE) Schmidhuber J (2015) Deep learning in neural networks: an overview. Neural Netw 61:85–117 Shannon CEJT (1948) A mathematical theory of communication. 27(3):379–423 Szeliski R (2010) Computer vision: algorithms and applications. Springer Science & Business Media, New York Tang Y, Zhang Y-Q, Chawla NV, Krasser SJIT (2008) SVMs modeling for highly imbalanced classification. IEEE Trans Syst Man Cybern 39(1):281–288 Tang X, Zhou P, Wang P (2016) Real-time image-based driver fatigue detection and monitoring system for monitoring driver vigilance. Paper presented at the 2016 35th Chinese Control Conference (CCC) Tock D, Craw I (1996) Tracking and measuring drivers’ eyes. Image Vis Comput 14(8):541–547 Veeraraghavan H, Papanikolopoulos NP (2001) Detecting driver fatigue through the use of advanced face monitoring techniques. ITS Institute Center for Transportation Studies, University of Minnesota, Minneapolis. pp. 31 Vora S, Rangesh A, Trivedi MM (2017) On generalizing driver gaze zone estimation using convolutional neural networks. Paper presented at the 2017 IEEE Intelligent Vehicles Symposium (IV) Vora S, Rangesh A, Trivedi MM (2018) Driver gaze zone estimation using convolutional neural networks: a general framework and ablative analysis. arXiv preprint arXiv:1802.02690 Ward JHJJ Jr (1963) Hierarchical grouping to optimize an objective function. 58(301):236–244 Wasserman PD (1993) Advanced methods in neural computing. Wiley, Hoboken, New Jersey Weisstein EWJ (2002) Bernoulli distribution. From MathWorld–A Wolfram Web Resource. https:// mathworld.wolfram.com/BernoulliDistribution.html Yan C, Jiang H, Zhang B, Coenen F (2015) Recognizing driver inattention by convolutional neural networks. Paper presented at the 2015 8th international congress on image and signal processing (CISP) Yuen K, Martin S, Trivedi MM (2016a) Looking at faces in a vehicle: a deep CNN based approach and evaluation. Paper presented at the 2016 IEEE 19th international conference on intelligent transportation systems (ITSC) Yuen K, Martin S, Trivedi MM (2016b) On looking at faces in an automobile: issues, algorithms and evaluation on naturalistic driving dataset. Paper presented at the 2016 23rd international conference on pattern recognition (ICPR) Zhang GPJN (2003) Time series forecasting using a hybrid ARIMA and neural network model. 50:159–175 Zhang Y, Hua C (2015) Driver fatigue recognition based on facial expression analysis using local binary patterns. Optik-Int J Light Electron Opt 126(23):4501–4505 Zhang W, Murphey YL, Wang T, Xu Q (2015) Driver yawning detection based on deep convolutional neural learning and robust nose tracking. Paper presented at the 2015 international joint conference on neural networks (IJCNN)

Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

20

Atslands Rego da Rocha, Igor Leão dos Santos, Letícia Ali Figueiredo Ferreira, and Augusto da Cunha Reis

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Cloud-Assisted IoT Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CIoT Infrastructure Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CIoT Virtualization Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CIoT Application Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cloud-Assisted IoT Application Domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cloud Manufacturing Application Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Healthcare 4.0 Application Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Agriculture 4.0 Application Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges and Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

520 523 524 527 529 531 531 539 547 551 553

Abstract

Combining the Industry 4.0 paradigms of the Internet of Things (IoT), edge computing, and cloud computing in the form of a three-tier architecture, the paradigm cloud-assisted IoT (CIoT) recently emerged. The CIoT paradigm is strongly based on the virtualization of physical sensing devices and on the instantiation of virtual nodes hosted in one of the three tiers of its infrastructure to provide data to applications. Several types of applications can use the CIoT to obtain its data. Examples of such applications are the ones that pertain to the manufacturing, healthcare, and Agriculture 4.0 application domains. In the

A. R. da Rocha () Universidade Federal do Ceará, Fortaleza, Brasil e-mail: [email protected] I. Leão dos Santos · L. Ali Figueiredo Ferreira · A. da Cunha Reis Centro Federal de Educação Tecnológica Celso Suckow da Fonseca, Rio de Janeiro, Brasil e-mail: [email protected]; [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_92

519

520

A. R. da Rocha et al.

manufacturing domain, the virtualization of physical manufacturing resources (e.g., machines, tools, and facilities) and, thus, the instantiation of virtual manufacturing resources, organized into virtual factories, can provide the ondemand use and efficient sharing of capabilities from the physical manufacturing resources among multiple users. The healthcare domain works as a network with numerous instances, processes, departments, routines, and services integrated and connected. This domain shares various resources to achieve a unique goal: assure the well-being of the population by providing proper care. In the Agriculture 4.0 domain, the combined use of the IoT, virtualization of resources (sensors, robots, and machines), and big data are trends to create digital farming and support decision-making in several essential agriculture activities. In this chapter, the CIoT paradigm will be discussed as an infrastructure to support those three application domains under the paradigm of Industry 4.0. Challenges that emerge from CIoT applications will be discussed, and finally, it is pointed out to future research directions. Keywords

Internet of things · Cloud computing · Healthcare · Agriculture 4.0 · Cloud manufacturing

Introduction Since the Hannover Fair occurred in 2011, there has been an increasing interest in the discussion of the topic Industry 4.0 by the academic community. Industry 4.0 is a term that refers to the fourth industrial revolution. This revolution is characterized by the emergence of a novel industrial paradigm, where manufacturers aim for decentralized, integrated, automated, and waste-averse production systems. This paradigm supports simultaneously, through intensive use of information and communication technologies (ICTs), high levels of product personalization and highly flexible mass production of the same customized products. This phenomenon is called mass personalization and is an advance in relation to the mass customization typical from the third industrial revolution (Tsai et al. 2020; Ghobakhloo and Fathi 2019; Wang et al. 2017). Among the main novel ICTs used in Industry 4.0, several authors (Tsai et al. 2020; Ghobakhloo and Fathi 2019; Wang et al. 2017) cite the Internet of Things (IoT) and cloud computing (Atzori et al. 2010; Vaquero et al. 2008), which are key technologies for the discussion in this chapter. Regarding the IoT, its first definition, according to Atzori et al. (2010), comes from the vision that “things” are objects identified by RFID tags, merely. In recent years, such definition evolved, now considering a global-scale networked infrastructure, that links a diverse set of devices, the smart things. Smart things are, for instance, smartphones, laptops, smart sensors, and household appliances. They can provide capabilities necessary to support the development of novel cooperative services in IoT (Atzori et al.

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

521

2010). Smart things in the IoT infrastructure are highly heterogeneous, in terms of functionalities and network protocols. On average, more than 30 billion IoT connections are expected by 2025 (Lueth 2020). Since it achieved a global scale, the IoT is a key technology to support new solutions to mitigate traditional global problems. Problems may comprise energy efficiency, safety, and health of the global population. Thus, the IoT copes with the goal of building the smarter (greener) planet (Hopper and Rice 2008). However, this global scale imposes a challenge to the IoT. How to ensure the manageability of the IoT network infrastructure? And how to handle the high volume of data generated by the smart things connected to this global infrastructure (Rao et al. 2012)? The answer to these challenges may reside in the cloud computing paradigm (Vaquero et al. 2008). Based on Vaquero et al. (2008), it is possible to define clouds as a large (virtually unlimited) pool of user-friendly (and virtual) computing resources, which can be reconfigured dynamically to serve a variable load from users. In the cloud computing paradigm, clouds are exploited in a model of pay-per-use that is mediated by service-level agreements (SLAs). More extensively, Liu and Wassell (2011) highlight six key features over which this paradigm works: (i) virtualization (hides from clients the heterogeneity of infrastructures, platforms, and data), (ii) scalability (dynamically reconfiguring clouds to serve a variable load), (iii) usability (userfriendly interfaces), (iv) reliability (response to failures), (v) security (of data that is sensitive), and (vi) a business model (based on pay-per-use). Moreover, Liu and Wassell (2011) define the services provided by clouds into three levels: infrastructure as a service (IaaS), platform as a service (PaaS), and software as a service (SaaS). In the lowest level (IaaS), virtual units built on physical computing resources (such as processing and storage) are provided to users. In the middle level (PaaS), cloud platforms for building cloud applications are provided to users. Finally, at the highest level (SaaS), cloud applications are exposed. Thus, clouds have what is necessary to help to surpass the challenges of IoT. Thus, it became interesting to combine clouds and the IoT into a two-tier architecture (Madria et al. 2014; Santos et al. 2015b). However, the IoT applications still faced another challenge, regarding their response time in the two-tier architecture. The response time of an IoT application comprises the time for communicating the data to the cloud, the time of processing by the cloud, and the time for the cloud to transmit the feedback to the IoT application. Although the cloud can process huge amounts of data within a reasonable time for IoT applications, such applications had to communicate to the centralized cloud data centers through multiple hops to have their data processed and to receive the feedback. Thus, the communication times involved were not satisfactory to IoT application requirements, having a crucial impact on their response time. To surpass this challenge, the fog computing paradigm emerged (Yi et al. 2015; Banijamali et al. 2020). Fog computing refers to enabling computing at the network’s edge. Thus, it is possible to enable new applications and services for the IoT with low response time. Fog nodes may be smart edge routers that provide its processing speed, through

522

A. R. da Rocha et al.

multiple cores and built-in storage. In this way, the traditional routers become smart and, thus, potential servers. In fog computing, fog nodes exist, which are the either physical or virtual devices positioned at the network edge that provide services like centralized clouds, but with reduced delay. Fog nodes can be either resource-poor, such as a Raspberry Pi, or resource-rich, i.e., actual high-scale servers (Yi et al. 2015; Banijamali et al. 2020). Fog nodes cooperate to produce the services required by the IoT in a distributed fashion and are strategically positioned to provide low response time in communications. However, the processing and storage capabilities of the fog are lower than the capabilities of the cloud. Thus, IoT applications must be designed to have this tradeoff in mind. IoT applications with strict response time requirements (more important than computation requirements) can be met by the fog. Meanwhile, applications with intense computation requirements (more important than the response time requirement) can be met by the cloud (Santos et al. 2018; Li et al. 2017). Therefore, the two-tier architecture was extended to comprise the fog computing tier, resulting in a three-tier architecture. This three-tier architecture thus combines the Industry 4.0 paradigms of IoT, fog computing, and cloud computing. This architecture received different names, behind which figures the same concept. The Cloud of Things (Aazam et al. 2014; Li et al. 2017), the Cloud of Sensors (Madria et al. 2014; Santos et al. 2015b, 2018), and the cloud-assisted IoT, more recently (Sajid et al. 2016; Wang et al. 2020). Figure 1 highlights the recent developments in the literature regarding the cloud-assisted IoT. This chapter refers to this threetier architecture by the name of cloud-assisted IoT (CIoT). The cloud-assisted IoT

Three-tier proposals:

Three-tier Cloudassisted IoT: • Fog/edge computing

• • • •

Aazam et al. (2014) Li et al. (2017) Santos et al. (2018) Wang et al. (2020)

Two-tier proposals: • Madria et al. (2014) • Santos et al. (2015b) • Sajid et al. (2016)

Two-tier Cloudassisted IoT: • Sensor Networks • Cloud Computing • Internet of Things

Fig. 1 The last 5 years of developments on the topic cloud-assisted IoT (CIoT)

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

523

paradigm is strongly based on the virtualization of physical sensing devices. Thus, it foresees the instantiation of virtual nodes (VNs) that can be hosted in one tier of the three-tier architecture (sensors/IoT, edge, and cloud tiers) to provide data to applications. This instantiation occurs according to the rules of a cloud-assisted IoT virtualization model and according to the requirements of the applications. For instance, applications may have strict response time requirements, requiring that VNs are instantiated on either sensors or edge tiers to speed up communications by shortening distances (in the number of hops) among communication devices and applications. As another example, applications may require the execution of high amounts of computation to obtain their data, and the response time may not be a strict requirement. In this last case, the instantiation of VNs to provide data to these applications can occur in the cloud tier (Santos et al. 2015b, 2018). This chapter has the objective of reviewing and discussing application domains of the cloud-assisted IoT (CIoT), raising challenges in these applications. As in every industrial revolution, the fourth industrial revolution introduces changes in every sector of the economy (and society), not only in manufacturing (the secondary sector of the economy). Thus, it is relevant to discuss applications of the Industry 4.0 technology of cloud-assisted IoT beyond manufacturing. In this sense, an application domain from the primary sector (Agriculture 4.0) and an application domain from the tertiary sector (healthcare) were also selected to be discussed in this chapter (Ravina-Ripoll et al. 2019). The possibility of selecting applications from the three different sectors of the economy to discuss is a clue of the extent of the impact of the cloud-assisted IoT in the economy, but do not limit it, since there may be other applications of the cloud-assisted IoT in society. This chapter thus contributes by discussing how several applications, from three different application domains, can use the cloud-assisted IoT and by raising associated challenges. This chapter is organized as follows. Besides this introductory section, Sect. “The Cloud-Assisted IoT Models” discusses the CIoT models. Section “Cloud-Assisted IoT Application Domains” discusses the manufacturing, Agriculture 4.0, and healthcare CIoT application domains. Finally, Sect. “Challenges and Future Directions” discusses challenges in the three CIoT application domains and points out future research directions, concluding this chapter.

The Cloud-Assisted IoT Models The discussion in this chapter is centered on the concept of the cloud-assisted IoT (CIoT), as one of the concepts that enable Industry 4.0 and support its applications. Therefore, it is first necessary to define the concept of CIoT formally. This section was designed to fulfill this objective, based on the literature. Following a didactic strategy to define the concept of CIoT, a division of this concept into three parts was considered: (i) infrastructure, (ii) virtualization, and (iii) applications. Where virtualization is performed over the infrastructure, applications use the virtualized infrastructure. Theoretical models for each of these parts are described in this section, following a bottom-up strategy.

524

A. R. da Rocha et al.

Therefore, Sect. “CIoT Infrastructure Model” describes the CIoT infrastructure model, Sect. “CIoT Virtualization Model” describes the CIoT virtualization model, and Sect. “CIoT Application Model” describes the CIoT application model.

CIoT Infrastructure Model The CIoT infrastructure comprises three tiers, things, fog, and cloud tiers (respectively referred by the acronyms TTier, FTier, and CTier), that form a communication backbone (network) that links data sources and applications. Each tier comprises computational entities (called nodes) that differ among tiers regarding resource constraints and communication delay to the TTier (Banijamali et al. 2020; Santos et al. 2018; Li et al. 2017). The computational entities of the TTier, called thing nodes (TN), are typical things that make up the IoT. For instance, TNs can be the physical sensor and actuator nodes (either wireless or not), laptops, smart devices such as smartphones and smart watches, smart vehicles, and wearable devices. In addition, TNs can be virtual entities (software instances) running on top of physical TNs. For instance, consider a physical smart fridge that is equipped with physical wireless smart sensors, those last endowed with local sensing and processing capabilities. A virtual counterpart of this smart fridge can be created using the available data and services provided by its sensors. This virtual counterpart (a virtual) can be stored logically within the own physical wireless smart sensors to provide smart fridge’s services. An example of such services is to respond to queries by applications (a data provisioning service) about the contents of the smart fridge and its energy efficiency (Fensel et al. 2017; Li et al. 2017; Dinh and Kim 2016; Santos et al. 2015b; Verdouw et al. 2013). The computational entities of the FTier are called fog nodes (FNs). Physical FNs can be, for instance, from a low-cost small-sized computer (closer to TNs) to smart network devices endowed with computational capacities, such as smart routers and switches, edge servers, and clusters of desktop computers (closer to CNs). Virtual FNs are software instances, such as virtual machines, provided by a fog computing virtualization model and running on physical FNs (Bellendorf and Mann 2020; Santos et al. 2018; Li et al. 2017; Pham and Huh 2016; Bonomi et al. 2012). The computational entities of the CTier are called cloud nodes (CNs). Physical CNs are high-performance computing servers and data centers that form large clusters within specific locations of the network. Virtual CNs are software instances, such as virtual machines, provided by a cloud computing virtualization model, and running on physical CNs (Bülbül et al. 2021; Liu et al. 2020; Santos et al. 2018). Given the definitions of TNs, FNs, and CNs, it is important to mention one important issue. The issue relates to the resource constraints of TNs, in comparison to the more resource-rich FNs and the virtually unlimited resources of CNs, what defines the tier each one fits. This classification is reflected in Fig. 2. The TNs have constrained computation, communication, and energy resources, since they are frequently miniaturized, mobile, and battery operated. FNs and CNs are, in

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

Cloud Tier (CTier)

525

Network Infrastructure 3

Network Infrastructure 5

VCN VCN

VCN

VCN

Fog Tier (FTier) Network Infrastructure 4

VFN

VFN

VFN

Network Infrastructure 3

VFN VFN

VFN

VFN

VFN

VFN

VFN

VFN

VFN

Things Tier (TTier) VTN

VTN VTN

VTN VTN

VTN

VCN

Virtual Cloud Node

VFN

Virtual Fog Node

VTN

Virtual Thing Node

VTN

VTN

Network Infrastructure 1

VTN

VTN VTN

Network Infrastructure 2

Network Infrastructure 3

Fig. 2 Three-tier CIoT infrastructure model. Communication links among physical FNs and physical CNs are represented, while links among physical TNs, and physical TNs and physical FNs, had their representations suppressed for the sake of simplicity. However, consider that physical TNs are connected among themselves, and at least one physical TN in an NI serves as a sink node and gateway to one of the bottom-most physical FN, so that the other physical TNs in the NI can access the edge of the fog network

turn, more resource-rich devices than TNs. They are connected to a stable power network and to a stable communication network infrastructure and have no physical dimension limitations. What differs the tiers of FNs and CNs is the scale of computational resources (such as processing power and memory), which is greater for the CNs, and the communication delay to reach the TTier, which is also greater to CNs, imposing a trade-off. Supporting the TTier by the CTier allows processing data faster than by the FTier. However, the results will be returned with a greater delay (Bellendorf and Mann 2020; Santos et al. 2018; Li et al. 2017; Pham and Huh 2016; Bonomi et al. 2012). Figure 2 shows the possible logical classification of each node within the tiers of the CIoT. However, it hides the complexity of the physical topology of the CIoT infrastructure. Meanwhile, Fig. 3 shows details of the physical topology of the CIoT infrastructure that are hidden in Fig. 2. By observing Fig. 3, one is able to understand how the CIoT infrastructure can be dynamic and organic, in both the physical and virtual realms simultaneously. The CIoT infrastructure can grow by adding, respectively, new physical and virtual TNs, FNs, and CNs to the network. Moreover, physical TNs are described in terms of processing speed, total memory, bandwidth, list of sensing and actuation units, remaining energy, network

526

A. R. da Rocha et al.

TTier and FTier

CTier

VTN

VCN

VFN VTN VTN

CTier

VFN

VCN

VTN

VFN

VCN

VTN VTN VFN

VFN

VFN

VFN

VFN VFN

VFN

VTN

VTN

CTier

VTN

VCN

VTN

VFN

VCN

Virtual Cloud Node

VFN

Virtual Fog Node

VTN

Virtual Thing Node

VFN

VTN

Fig. 3 CIoT infrastructure physical topology

identification, and location. Meanwhile, virtual TNs, physical and virtual FNs, and physical and virtual CNs are described in terms of processing speed, total memory, bandwidth, network identification, and location (Santos et al. 2018). Physical and virtual TNs, FNs, and CNs are owned and administered by CIoT infrastructure providers (InP). They are part of network infrastructures (NIs) that sum up to compose the CIoT infrastructure. Physical TNs, FNs, and CNs are deployed over an area and connected by either wireless or cable links (supported by the Internet), so that every node pertains to one or more NI. The InPs define the physical and logical (administrative) boundaries of their NIs (Santos et al. 2018). The deployment of virtual TNs, FNs, and CNs on their respective physical hosts is transparent to the three-tier CIoT infrastructure and is handled by typical virtualization models of IoT, fog, and cloud. As in Dinh and Kim (2016), the CIoT infrastructure is agnostic to the IoT, cloud, and edge computing virtualization models adopted to provision of the virtual TNs, CNs, and FNs. The CIoT follows the principle of overlay virtualization, seen in the work of Khan et al. (2016). Thus, the CIoT virtual nodes (VNs) are built in an overlay layer, over already virtual entities (the virtual TNs, FNs, and CNs). In such model, the whole CIoT infrastructure is offered through overlay virtualization. According to Khan et al. (2016), overlay networks are advantageous: they are distributed, lack a centralized control, and

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

527

allow sharing of resources, thus being CIoT virtualization candidates. Thus, CIoT VNs are built on the available virtual TNs, CNs, and ENs. Therefore, the virtual entities of each tier of the CIoT infrastructure are the ones over which the CIoT virtualization model, described in next subsection, runs.

CIoT Virtualization Model The concept of CIoT virtual nodes (VNs) considers that such entities are abstract representations of capabilities (including data, computation, and communication capabilities) of the CIoT infrastructure. The VNs are allocated to provide data to applications’ requests, fulfilling their requirements. In this sense, VNs link the CIoT infrastructure and CIoT applications. The VNs are computational entities, which are implemented as instances of software, analogous to typical IoT resources (Zhang and Chen 2015), that run on top of the TTier, FTier, and CTier. Examples of attributes of CIoT VNs are processing power, memory, bandwidth, geographic location coordinates and identification of its host, and the list of underlying nodes (virtual TNs, FNs, and CNs that compose the VN). These are all functions of the properties of respective underlying nodes. To manage the VNs, since their instantiation to their operation, a virtualization middleware is necessary. The most important functions of this middleware are described in this section, along with the main properties of VNs (Banijamali et al. 2020; dos Santos et al. 2019; Santos et al. 2018; Khalid et al. 2014; Madria et al. 2014). In the virtualization middleware of the CIoT, there is a resource provisioning function that performs the instantiation of VNs and calculates each of their properties. Resource provisioning can be either proactive, i.e., InPs instantiate VNs and configure their services apart from CIoT operation and independently on application requests (prior to their arrival in the CIoT), or reactive, i.e., the instantiation of VNs occurs dynamically and automatically, in response to the arrivals of requests during the CIoT operation. In addition, a resource allocation function exists in the middleware, with the objective of allocating resources (VNs) to provide data to the applications during the operation of the CIoT and its VNs. Moreover, a task scheduling function is responsible for coordinating the underlying CIoT infrastructure of each VN, during CIoT operation, to obtain the necessary data to be further allocated to an application. The resource provisioning, resource allocation, and task scheduling functions can be either centralized, running on a central virtual TN, FN, or CN, within the respective middleware component and solving a globally defined problem, or decentralized, running partly within the respective middleware components of each VN and solving a localized/distributed problem through collaboration of VNs (dos Santos et al. 2019; Santos et al. 2018). It is important to mention that the VNs can publish and update, proactively, all their information in the TTier, FTier, and CTier, upon changes in their information, by communicating new information to the virtualization middleware. However, the middleware, which instantiated VNs, has the information regarding all VNs, allowing it to also sharing VN information among TNs, FNs, and CNs reactively

528

A. R. da Rocha et al.

to either an application request or a detected change in a VN. A standard, such as the sensor modeling language (Yuriyama and Kushida 2010), for describing VNs’ information can be used, which abstracts the heterogeneity of underlying devices and protocols (dos Santos et al. 2019). The CIoT virtualization model can be information fusion based (Fig. 4), such as Olympus (Santos et al. 2015b). In this case, each VN represents an information fusion technique (Farias et al. 2016; Nakamura et al. 2007), is able to provide data as the output of this technique, and is capable of performing sensing, processing, or actuation on the CIoT infrastructure as required to produce the output data of the VN, i.e., performing a data update through task scheduling (Santos et al. 2018). A data update activates and establishes communication routes among underlying virtual TNs, FNs, and CNs composing the VN, transparently to users and applications. Task scheduling approaches that fit the characteristics and specificities of the CIoT can be found in literature (Li et al. 2014; de Farias et al. 2013; Chengjie Wu et al. 2012). The output data provided by CIoT VNs comprises raw sensing data and/or processed data (by the information fusion technique of the VN) and/or control actions data (feedback data or acknowledgment data in the case of actuation

VTN

M D VCN

M

VFN

VTN F F

VCN

M VTN F VFN

D

VFN

D

M VTN

M M

VTN

VTN

F VFN

M

VFN D VFN F VFN

M

VFN

D

M VTN

VTN

F

Feature level CIoT Virtual Node

D

Decision level CIoT Virtual Node

M F

VTN

F

VFN

M VTN

VCN M

VFN

M

Measurement level CIoT Virtual Node

VFN

VTN

VCN

Virtual Cloud Node

VFN

Virtual Fog Node

VTN

Virtual Thing Node

Fig. 4 Examples of compositions of Olympus virtualization model’s virtual nodes (VNs), in an overlay network in the CIoT. Lines represent that a VN is composed either by another VN or by virtual TNs. Thus, in this figure, communication links are not represented

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

529

tasks). The output data of a VN is, thus, in a given information fusion level, i.e., measurement, feature, or decision (Nakamura et al. 2007). Moreover, the output data of a CIoT VN is of a single data type. Each data type existing in the CIoT is, ideally, unique, thus defined by its unique description. A data type may be, for instance, a damage detection indicator, represented in a data structure (for instance, a multidimensional matrix, vector, or a single integer of a given number of bytes), collected for a specific civil structure, such as a bridge or wind turbine, for instance (Alves et al. 2017; Santos et al. 2016). A data type also defines the information fusion technique and the input data types necessary used to calculate it (dependences among data types). Several VNs can provide data in a same given data type. It is the responsibility of InPs to define and describe to users, and their applications, the data types existing in the CIoT. Thus, a request, of an application, to a given data type can be met by the provisioning of data of this data type, where such provisioning is performed by a VN. To perform this provisioning, each VN has the implementation of a service of data provisioning, with such implementation corresponding to functions necessary to acquire and process the provided data (according to the information fusion technique of the VN) in response to application requests (dos Santos et al. 2019; Santos et al. 2018). The model that governs the CIoT applications is described in Sect. “CIoT Application Model”.

CIoT Application Model The CIoT application model is based on a workflow approach, inspired by the research field on web services (Chirichiello 2008; Ganesarajah and Lupu 2002). CIoT applications are modeled as time-based (Li et al. 2017), i.e., applications define the moment in time for demanding data from the VN. CIoT users model applications as directed acyclic graphs (DAGs) of workflow, where each node of an application graph is a request to a given data type (Li et al. 2017, 2013). A request is, thus, a logical interface that links to, and demands data from, the data provisioning service implemented by VNs. Thus, requests are abstract, i.e., they have no implementation of information fusion techniques or services. Such implementation is accomplished by the VNs and is under responsibility of the InPs that instantiate (provision) the VN, either reactively or proactively, through the implementation of the resource provisioning process of the CIoT virtualization middleware (dos Santos et al. 2019; Santos et al. 2018). Figure 5 shows an example of an application with six requests modeled as a DAG. A request has (i) a list of identifiers of predecessor and successor requests (the list of predecessors is possibly null, when the request is the first in the DAG of the application), (ii) a given data type requested (called the output data of the request), and (iii) a non-negotiable requirements list and levels of negotiable requirements that must be fulfilled by VNs in their data provisioning. Precedences among requests are defined by the data dependence between two or more requests, i.e., a request’s input data may depend on the output data of one or more requests. Non-negotiable requirements (e.g., the data type) must be completely fulfilled by

530

A. R. da Rocha et al.

R4 R1

O

I

M

R1

O

R2

O

F

I F

R1 M

O

I

R4 D

R3 F

I

O

Request

I

Input data

O

Output data

M

Measurement level CIoT Virtual Node

USER

O F

Feature level CIoT Virtual Node

D

Decision level CIoT Virtual Node

Fig. 5 Example of application, modeled as a DAG, and comprising six requests

VNs, while negotiable requirements (e.g., data freshness) allow a given level of partial fulfillment (Perera et al. 2014). Data updates calculate data of a given data type and improve its data freshness (dos Santos et al. 2019; Santos et al. 2018). In turn, an application, as a whole set of its requests, has (i) the moment in time it arrives in an application entry point (AEP) in the CIoT, (ii) a given level of priority in relation to other applications, and (iii) its list of requests. An AEP may be positioned (and run) in a given virtual TN, FN, and CN, being a part (a software component) of the virtualization middleware. In addition, multiple AEPs can exist in the CIoT, while an application arrives with its set of requests in the CIoT through only a single AEP. Applications can arrive, for instance, from user devices (such as smartphones or PCs) connected directly (or through the Internet) to a FN. Upon the arrival of an application in the CIoT, the virtualization middleware can proactively search (according to an implementation of the resource allocation process by the InPs) by the data types and VNs that provide this data type, which can be alternatives to meet the application requests. However, there is also the possibility of VNs proactively disseminate their information periodically in the CIoT, to the resource allocation component (dos Santos et al. 2019; Santos et al. 2018). An application manager subsystem (AMS), a component of the virtualization middleware, has the main functionality of allowing users to develop applications. The AMS allows users to select the relevant data types available in the CIoT and to express the requests to such data types. The AMS may use service discovery methods in literature (Perera et al. 2014) to discover data types and data provisioning services. In addition, the AMS may use domain specific languages (DSL) or semantic queries (Fuentes and Jiménez 2005) to express requests (Santos et al. 2018). The application workflow defines the flows of data among VNs in the CIoT and, consequently, the work performed by VNs to obtain this data. One possible implementation of procedures to handle applications in the CIoT is described as follows. After its development in the AMS, the application joins the CIoT through an AEP that runs a resource allocation algorithm to assign the requests to

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

531

VNs. When assigned to VNs, the requests are transmitted and join the respective VNs’ queues of requests. Now consider a single request X in a VN Y queue. When selected in the queue by the VN Y to run, the request X activates the data provisioning service of the VN. Then, the data provisioning service of VN Y waits for the input data to arrive in it, from other VNs to which the predecessor requests of request X were allocated. With the input data available in VN Y, the VN Y calculates its output data locally through complementary local data collection (if necessary) and execution of its information fusion technique. Finally, the data provided by the VN Y is assigned as the output data of the request X, and thus the request X is finally met. Suppose request X is a final request in the application graph. In that case, the VN transmits the output data of the request X to the application users, located in the original AEP where the application arrived. Otherwise, the output data is transmitted to the VNs that hold successor requests, and the process is repeated in these VNs. Moreover, data reuse, i.e., reuse of data (either input or output data) of a given data freshness level (negotiable requirement), can be implemented in the data provisioning service of a VN. This can avoid unnecessary data updates (reexecutions of the information fusion technique) and thus improve the response time and energy consumption of the CIoT (dos Santos et al. 2019; Santos et al. 2018). In Sect. “Cloud-Assisted IoT Application Domains” application domains of the CIoT, whose models were described in this section, are described.

Cloud-Assisted IoT Application Domains This section describes possible application domains of the cloud-assisted IoT, which are cloud manufacturing, in Sect. “Cloud Manufacturing Application Domain”; Healthcare 4.0, in Sect. “Healthcare 4.0 Application Domain”; and Agriculture 4.0, in Sect. “Agriculture 4.0 Application Domain”.

Cloud Manufacturing Application Domain Under the paradigm of Industry 4.0, the manufacturing application domain has recently realized the emergence of the paradigm of cloud manufacturing (CM). As the cloud-assisted IoT paradigm, this paradigm is also based on virtualization, however, in terms of virtualization of physical manufacturing resources (machines, tools, humans, and facilities, for instance). The instantiation of virtual manufacturing resources (VMRs), organized into virtual factories (VFs), can provide the on-demand use and efficient sharing of capabilities from the physical manufacturing resources among multiple users (each one emitting its production orders) (Zhang et al. 2014). The CM paradigm architecture provides components that implement methods for ensuring, for instance, the man-to-man, man-to-machine, and machine-to-machine communications and allocation of the capabilities of the physical resources to the production orders emitted by users. Therefore, the cloud-assisted IoT paradigm can

532

A. R. da Rocha et al.

play an essential role in supporting the CM paradigm by providing data required by the VMRs and CM architecture components to perform their attributions (Zhang et al. 2014). In this section, the joint use of CM and CIoT paradigms will be referred to as the paradigm of cloud-assisted IoT manufacturing (CIoTM). A possible architecture for the CIoTM will be discussed, based on the original CM architecture from the work of Zhang et al. (2014). Benefits of joining both models are, for instance: (i) benefits of the inclusion of fog computing, missing in the original CM architecture, bringing reduced response time to applications; (ii) better energy efficiency of CIoTM devices in general and lifetime extension for battery-operated devices, through data reduction achieved through information fusion (Santos et al. 2016, 2018; Li et al. 2017; dos Santos et al. 2014); and (iii) integration of CM with other application domains that can be served by the same CIoT infrastructure, because VNs providing data to the CM applications can also provide the same data to other applications from different domains, thus improving data reuse in a same standard format across multiple application domains, fostering the integration of applications, cooperation among applications, and reduced energy consumption (Santos et al. 2018; Li et al. 2017). Moreover, the CIoTM paradigm can be applied to several types of production processes. This section also briefly discusses the application of CIoTM in these types of production processes. Specifically, this section discusses more deeply the application of CMCIoT on the continuous process of wind energy production in wind farms. This section describes the cloud manufacturing paradigm (CM), in Sect. “The Cloud Manufacturing (CM) Paradigm”; the cloud-assisted IoT manufacturing (CIoTM) paradigm, in Sect. “The Cloud-Assisted IoT Manufacturing (CIoTM) Paradigm”; the cloud-assisted IoT manufacturing (CIoTM) applications, in Sect. “Cloud-Assisted IoT Manufacturing (CIoTM) Applications”; and the specific application in a wind farm, in Sect. “Cloud-Assisted IoT Manufacturing (CIoTM) in a Wind Farm”.

The Cloud Manufacturing (CM) Paradigm The CM is a new paradigm of manufacturing that is strongly based on communication networks and virtualization. Cloud computing is used by this paradigm to expose to users the capabilities of physical manufacturing plants and their physical manufacturing resources as services (called CM services) to users. As such, the capabilities can then be managed and operated in an intelligent manner. Under this concept of CM, sharing of physical manufacturing resources and manufacturing capabilities among multiple users becomes possible. Besides that, CM services show characteristics such as (i) safety, (ii) reliability, (iii) high-quality, (iv) low-cost, and (v) on-demand provisioning (Zhang et al. 2014). The CM architecture comprises five layers, as described by Zhang et al. (2014). These are, from top-down, the application, middleware, service, perception, and resource layers. The application layer of the CM architecture is responsible for abstracting a variety of specific application interfaces, for various manufacturing industries. Different users (classified as cloud operators, manufacturing resource

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

533

users, and manufacturing resource providers) access on-demand cloud services through the application layer. CM applications may support the whole life cycle of products and provide cloud-end customized interfaces (Zhang et al. 2014). The middleware layer of the CM architecture has the main attribution of supporting core components and their functions for the operation of manufacturing capability services. Some core components are responsible for the management of functions of cloud service, knowledge, cooperation, platform running, transaction, failure, and energy, for instance (Zhang et al. 2014). The service layer of the CM architecture comprises virtual CM resources and CM services. The primary attribution of this layer is to promote virtualization of the manufacturing resources and capabilities, providing the respective cloud services and thus forming the pool of CM services (Zhang et al. 2014). The perception layer of the CM architecture has the primary goal of sensing physical resources of manufacturing and capabilities of manufacturing, connecting them to the edge/cloud network and the Internet. The use of technologies such as the IoT is proposed in this layer. These technologies enable the connection of a wide variety of manufacturing resources and capabilities to the above layers (Zhang et al. 2014). The resource layer of the CM architecture refers to the physical resources of manufacturing and their capabilities. Physical manufacturing resources can be, for instance, machining centers, machine tools, knowledge, and data. In turn, the formation of capabilities occurs with resources, people, and knowledge. The capability is necessary to complete a manufacturing task. Capabilities can be, for instance, design, simulation, and product capabilities, as well as other life cyclerelated capabilities of the manufacturing process (Zhang et al. 2014). Section “The Cloud-Assisted IoT Manufacturing (CIoTM) Paradigm” describes the integration of CIoT and CM paradigms, considering the architecture of CM described in this section.

The Cloud-Assisted IoT Manufacturing (CIoTM) Paradigm Figure 6 shows the integration of CM and CIoT paradigms, in terms of their concepts. From top-down, it is first necessary to distinguish CM applications and CIoT applications. The application layer of the CM architecture, as described in previous section, remains intact. Its main goal is to provide access to the virtualized manufacturing resources and capabilities by users. CIoT applications, in turn, are described by their workflows (DAGs) that orchestrate VNs in the CIoT. The CIoT applications emerge in the middleware layer of the CM architecture. It is suggested that CM middleware functions are performed through data requests and analysis from the service layer. Thus, CM middleware functions drive the CIoT VNs, defining application requests to them. In turn, VNs provide data, through their data provisioning services, to CM middleware functions. CM application and middleware layers are typically performed by the CTier, according to Zhang et al. (2014), with no direct involvement of the TTier. However, it is suggested that the FTier can also perform CM application and middleware layers, according to the requirements of CM applications regarding response time.

534

A. R. da Rocha et al.

CM Application Layer FTier CTier

CM Middleware Layer

CM Service Layer

M

D F

D D

CM Perception Layer VCN

M

VTN

VTN

TTier FTier CTier

CIoT Application workflow M

Measurement level CIoT Virtual Node

F

Feature level CIoT Virtual Node

D

Decision level CIoT Virtual Node

VCN

Virtual Cloud Node

VFN

Virtual Fog Node

VTN

Virtual Thing Node

VFN

CM Resource Layer TTier

Fig. 6 Integration of the concepts of CIoT and CM

In the CM service layer, the virtualization of the manufacturing resources and capabilities is performed through the instantiation of CIoT VNs. This layer is the ideal candidate to hold the CIoT VNs, because CIoT VNs can, through their information fusion techniques implemented, perform sensing, data processing, and control actions over the manufacturing resources and capabilities, allowing some level of interaction with them, while returning data through their data provisioning services. Thus, it is suggested a data-driven virtualization model of manufacturing resources and capabilities in the CM services layer. So that this model allows the integration of the CIoT models described in Sect. “The Cloud-Assisted IoT Models” and the CM architecture from Sect. “The Cloud Manufacturing (CM) Paradigm”. It is important to mention that VNs in this layer can be built on top of any virtual TN, FN, or CN. Thus, elements of the TTier, FTier, and CTier are present at this layer of CM. In the CM perception layer, figures all the communication infrastructure of the TTier, FTier, and CTier of the CIoT. Thus, this level supports all the necessary technologies to connect the manufacturing resources and capabilities to CIoT VNs in the service layer. Physical and virtual TNs, FNs, and CNs are present in this layer. Most of the CIoT infrastructure is part of the CM perception layer. The part that does not fit the perception layer composes the resource layer. In the resource layer, specific smart things (physical and virtual TNs) from the CM domain exist, representing manufacturing resources and capabilities. Thus, the TTier is divided; part of it is in the perception and service layers, while the other part (specific of the CM application domain) is in the resource layer. It is important to mention that it is being considered that manufacturing resources and capabilities are physical and virtual smart things, respectively, because capabilities may be soft in their

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

535

conception, while manufacturing resources may be real physical machines, for instance. Next section raises potential applications of CIoTM.

Cloud-Assisted IoT Manufacturing (CIoTM) Applications Manufacturing resources and capabilities are provided to users through CM (Zhang et al. 2014). Since the CIoT paradigm can integrate to the CM paradigm, as discussed in Sect. “The Cloud-Assisted IoT Manufacturing (CIoTM) Paradigm”, it is possible to suggest the same to the CIoTM paradigm. It is also possible to consider that different manufacturing processes, with specific manufacturing resources and capabilities, besides specific characteristics of production volume and variety, can consequently benefit from the CIoTM (Chambers 2015). Thus, CIoTM applications in the manufacturing application domain can be classified into the type of production process to which they are designed. The types of production processes are project, jobbing, batch, mass, and continuous. Project production processes are characterized by the lowest production volume and highest production variety among all the types of production processes (in general, projects result in unique products). Examples of project production process are the ones from naval industry (e.g., shipyards), aerospace industry (e.g., aircraft construction), and civil construction (e.g., bridge building). In such processes, manufacturing resources (equipped with sensors) have strong mobility requirements that must be considered by the CIoTM, since they frequently move from storehouses to where the final product (project) is being assembled. Multiple types of data can be extracted from sensors in the product, which generally has large dimensions, thus with large sensing area requirements and low mobility requirements, to be provided through VNs to CIoT applications and, thus, to CM middleware functions. In addition, manufacturing resources (for instance, tools, machine tools, and human resource teams) can also have sensors deployed, collecting data that is also highly variable due to the flexible nature of resources. Jobbing production processes are characterized by not so variable products as in the project production process and a higher production volume. The main difference of this process to other production process is the need to share resources among products. In each product, the transforming resource operates similar transformation operations, however, individualized by product. Examples of jobbing production processes are the ones from tailoring, sawmill, furniture, and specialized toolmaker industries. It is usual, in this production process, that resources have fixed positions within functional departments, which group resources according to their capabilities. Therefore, sensors installed in manufacturing resources usually have no mobility requirements, and information from the departments can be more easily obtained by fusing information from the multiple similar resources in them. Products run through long and complex routes among departments, according to the sequence of operations (performed by the departments) they must go through. Mobility of sensors in products is, thus, a strong requirement in this kind of production process. Batch production processes are characterized by an average volume of goods standardized in batches. Each batch follows a predefined sequence of operations,

536

A. R. da Rocha et al.

and manufacturing resources are more specialized (in terms of the operations they can perform) in comparison to jobbing and project processes. Examples of batch production processes are the manufacture of clothing and footwear on a small scale (consider differentiated products, such as a shirt factory, which produces the same model, in different colors and sizes) and the production of components that feed an assembly line. In batch production process, it is possible to monitor and take decisions to whole batches, not requiring that each product has sensors installed on it. However, fusing information from multiple products within batches is still a possibility. Moreover, manufacturing resources generate data with lower variability in relation to jobbing production process. Mass production processes are characterized by large-scale production of highly standardized products, with repetitive activities and stable demand. Examples of mass production processes are the ones from large-scale and standardized textile products, automotive products, and ceramic products. The mass production process considers a manufacturing resource structure that is highly specialized and not very flexible (in terms of the operations they can perform on products), in comparison to previous production processes. Thus, resources equipped with sensors can be monitored more easily, generating data with lower variability than batch production. In this kind of production process, the steady product flow is more important to be monitored, thus manufacturing resources and the assembly line are more important to be equipped with sensors and monitored than the products individually. Although advantages may still exist in monitoring products individually, this monitoring is more costly than in previous production processes, since the production occurs in large scale. Finally, continuous production processes are characterized by the impossibility of identifying products individually (discretely). The production flow sometimes cannot be interrupted. Products and operations (processes) are interdependent. And high volumes of products must be allocated to production, to justify the fixed cost related to the operation. It is the most inflexible production process. Examples of continuous production processes are the ones from industries of oil and gas extraction, water treatment and distribution, chemicals, and electricity production. In next section, an application of CIoTM in a continuous production process of wind power is described.

Cloud-Assisted IoT Manufacturing (CIoTM) in a Wind Farm In this section, the scenario of application of CIoTM in a wind farm will be described, with the entities of the CIoT infrastructure described, the VNs’ and data types’ description, and a CIoT application workflow description, designed to support the CM middleware function of failure management. This section is concluded with a discussion on the possible CM application to expose to users the manufacturing resources (wind turbines) from this scenario. In continuous production processes, such as wind power production in wind farms, the product flow’s properties can be monitored (and not discrete units of products) regarding power quality, for instance, along with manufacturing resources’ (such as power lines and wind turbines) properties, regarding structural

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

537

health (Alves et al. 2017; Santos et al. 2015a). The application scenario described here will focus on physical TNs that are the wind turbines equipped with sensors, those last capable of measuring acceleration in their positions on the wind turbine. Virtual TNs can be instantiated and run over the physical TNs, representing the wind turbine virtually. In turn, physical FNs, such as smart routers, can be installed in random positions within the large area covered by the wind farm, forming a communication backbone that links to the Internet at a point in the frontier of the wind farm. From this point on, the fog network may comprise several heterogeneous physical FNs, toward the physical CNs. Physical FNs host, each, a virtual FN with its representation. Physical CNs are data centers located far away from the wind farm, somewhere on the Internet, linking several wind farms geographically dispersed in a hypothetical country of large dimensions. One first type of VN (at measurement level) that can be instantiated on virtual TNs is the VN that provides a data type named “A” which consists basically of raw acceleration data. This data type A is defined by a data structure with the fields: (i) given fixed sampling rate used in acquisition, (ii) fixed number of samples, (ii) values of acceleration for each sample, (iii) identification of the wind turbine, and (iv) relative position in the turbine. The information fusion technique that implements the obtention of this data type A consists of reading the accelerometer samples and, possibly, applying filters for suppression of noise on this acquired data (a measurement level information fusion technique). A second type of VN (at feature level), also instantiated on virtual TNs, provides data type “B,” which is a local damage detection indicator, calculated through an information fusion technique seen in dos Santos et al. (2014). A third type of VN (at decision level), also instantiated on virtual TNs, provides data type “C,” which is a warning, issued in case of local damage detection on the wind turbine. A fourth type of VN (at decision level), also instantiated on virtual TNs, provides an actuation feedback, which may shut down the wind turbine in case of damage detection. Data type D depends on C, which depends on B, which depends on A. An application workflow, linking these four data types as specified, can be periodically issued to the CIoT VNs mentioned, to periodically assess the turbines and try to detect damage. This verification can occur several times a day, imposing low communication overhead and requiring little processing within the own virtual TNs, thus allowing local decisions of VNs on their virtual TNs. If damage is detected, the actuation foreseen in data type D can shut down the turbine immediately and issue an actuation feedback to users (the user in this case is the failure management component of the CM middleware, which may be in a virtual FN or CN), avoiding disasters of larger dimensions (Alves et al. 2017; dos Santos et al. 2014). A fifth type of VN can be instantiated on virtual FNs and provides the data type “E,” which is a damage localization coefficient (dos Santos et al. 2014). And a sixth type of VN can also be instantiated on virtual FNs, providing a data type “F,” which is a damage extent determination coefficient (dos Santos et al. 2014). Data type F depends on data types E1, E2, . . . En, which are data type E variations, calculated for several different positions in a wind turbine. As well, each data type E depends on data types B1, B2, . . . Bn. The point is that damage localization and extent

538

A. R. da Rocha et al.

determination is done for a cluster of sensors positioned in different locations on the structure, according to dos Santos et al. (2014). However, localization and extent determination of damage must be determined fast enough to allow fast response in case of damage occurrence. Thus, the FTier is the ideal candidate to handle this greater amount of computation, with faster response than the cloud. The VNs at the virtual FNs that calculate data types E and F perform tasks of cluster heads in the localized algorithm presented by dos Santos et al. (2014). Application workflows can be defined to data types E and F and their variations, with also demanding periods of execution, since data flows through the TTier and FTier. Finally, a data type “G” that depends on data type “A” can be defined, which is the result of a detailed structural properties’ analysis of each wind turbine in the wind farm. A VN providing data type G can be in a virtual CN. And an application workflow can request data types A (with variations A1, A2. . . An) and data type G. Processing and providing data type G requires huge amounts of data and transmission of such data to the cloud. However, the structural analysis is not focused on deciding on damage detection fast. But it is interested on studying the structural behavior of the wind turbine during long periods of time. Data type G must be interpreted by a structural engineer, which can study the evolution of damage that occurs slowly over time. This application runs periodically, but can occur once a day, week, or month, and thus does not require fast response. Moreover, the CN can perform analysis across multiple wind farms, drawing the state of damage of the whole wind power system of a country (dos Santos et al. 2014). Finally, the failure management component of the CM can process the data from all the data types mentioned previously and help managing which physical wind turbines will compose the virtual manufacturing resources (wind turbines) that will be exposed to users. Based on the analysis (information fusion techniques) of each data type, and the states of damage of each physical wind turbine, new sound physical wind turbines can be selected to compose an existing virtual wind turbine dynamically, allowing the virtual wind turbine to operate continuously. Prediction of damage can also be performed, allowing the change of physical wind turbines to occur predictively. Thus, the continuous production flow of wind power is never cut due to failures (damages) in wind turbines of the wind farm. A CM application can use the virtual wind turbine to provide electricity to a given user, and thus, the continuous flow of electricity is ensured by the failure management component of the CM middleware. Concluding this section, it is emphasized that several examples of applications (regarding CIoT and CM applications) were raised in this section, suggesting clearly the possibilities of data reuse and use of each tier of the CIoT for meeting each kind of application. However, the applications in the scenario described in this section are not limited. Several other applications in manufacturing can be raised, either by reusing the data types and infrastructure described or by expanding them, since the CIoT infrastructure and virtualization model are scalable (Santos et al. 2018). Other applications in different production processes can also be raised, using similar concepts described in this section. Finally, other types of CM services can be also be designed, fitting the service categories in Zhang et al. (2014).

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

539

Healthcare 4.0 Application Domain The healthcare sector (HCS) works as a network with numerous components (companies, instances, processes, departments, routines, and services) integrated, connected, and sharing various resources to achieve a unique goal: assure the wellbeing of a population efficiently delivering proper care. These components and their interaction produce a scenario of unexpected and dynamic outputs that may or may not behave differently to its initial conditions and can create new and emergent behaviors and unintended consequences. Therefore, the multiplicity of its configuration and the complex behavior of this sector require not only efficient resource and asset management and process and quality control solutions but also reliable and capable ways to deal with huge amounts of new data and information. What is also important to reinforce is that these needs are closely related. The high volume and diversity of data and data sources that can be received from the system by sensors or other connected devices in a to-be robust analytical HCS environment will make available the information needed for not only monitoring said environment but also making decisions around staffing, resourcing, financial considerations, health programs, operational planning, and as a basis for public health policies. These monitoring and decisioning can be extended to learning and, possibly, automation as we go from a state of changing analog to digital data (digitization) to digitalizing the workflow and using technology to produce new capabilities. Moreover, the nature of the processes involved in the sector call for quick answers and, sometimes, real-time solutions. In this scenario, and with the development of information technology, smart healthcare technology emerges as the employment of technology to connect people, resources, and institutions related to the HCS to actively manage and respond to a medical ecosystem’s needs intelligently. But before getting into that, it is important to understand the HCS. As the key characteristic of the HCS is to provide (and produce) health-related services, it is possible to classify such services according to their types using a “service structures classification” (regarding flexibilization and labor intensity) that is analogous to the “manufacture structures classification” previously used for the manufacturing application. The broadness of the HCS allows it to fit throughout all types of services. Services provided by hospitals and emergency departments (ED), and surgeries in general, are often classified as professional services as there is a high degree of labor intensity and a high degree of interaction with the patient. In contrast, laboratory analysis and mass vaccination campaigns, for instance, can be categorized as mass services as, in the first case, its processes are previously designed, without the direct participation of the patient at the time of providing, and in the second case the patient is handled through the stages of the linear process as in a mass production logic. Thus, it is possible to see that all types of services can be found within this particular HCS. These distinct classifications corroborate with the diversity of the HCS, especially when considered that there are cases in which more than one classification can be perceived within the same installation or configuration (labs are running inside hospitals, for instance).

540

A. R. da Rocha et al.

Similarly to what happened with the service classification, it is possible to observe the healthcare sector through the lens of production systems configuration and perceive its wholesome as being project-oriented, where each output (patient) is unique in its requirements and resources used. However, the HCS can also inherit characteristics from other classifications, such as job shop-oriented processes for its high variety of clinical cases, low volume of patients being attended simultaneously, use of general-purpose machines and facilities for distinct procedures and exams, employment of highly skilled human resources, and a large inventory of materials, medications, machinery, etc. availability. For both cases and types of classification, these different “heritages” can extend throughout its chain of processes, installations, and configurations (chains refer to the existent interconnections among them) according to the particularities of each one of them. Given these parallels and classifications, it is easier to perceive the heterogeneousness and complexity of applications of the HCS as well as its several and distinct components. Moreover, this heterogeneousness also shows how a singular place, such as a hospital or an ED, have diverse needs regarding other two main domains: patient-related (based on the patients themselves and the care provided to them) and process-related (based on the processes and routines running to make this care possible and efficient). These two domains are intertwined and all decisionmaking in the HCS must consider both. Thus, the objective of this chapter is to take a close look at the operations, routines, and processes that support the care. However, it is important to acknowledge that the high-touch and patient-centered aspect of the HCS is still considered, but understanding that operational efficiency is a key factor for the quality of care as well. Regarding decision-making, another way of looking at the HCS is looking at its managerial and hierarchical levels. Despite how a HCS-related domain operates in means of customization/flexibility or volume, service, or production, they all depend on structured decision-making, planning, and control to operate properly and efficiently. Frameworks decompose this hierarchy into three administrative levels: strategic (SL), tactical (TL), and operational (OL) (Bittencourt et al. 2017). These levels guide the decision by not only assigning responsibility to distinct players but also by determining the period of impact (long, medium, and short term), time of response, detailment of information and data frequency, contact with the operation’s edges, and degree of impact to the patients. Considering the smart healthcare domain, the administrative hierarchical levels share similar characteristics with the three tiers/layers of the CIoT: CTier, FTier, and TTier. Analogous to the SL, the CTier is on the top of the hierarchy and is more centralized. SL requires huge amounts of data and information and, therefore, higher computational resources. The long-term decisioning that happens in the SL is better suited to the CTier, which has a greater delay than the other tiers but a more efficient data processing capacity. The FTier is also the intermediate layer responsible for data transformation and distributed storage and network management and control. The FTiers is closer to the operations as it is more local and spread in distinct environments, which is comparable to the TL. Its lower response when compared to the CTier makes it more adequate for quick responses and even on-time decisions,

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

541

which can be suited for middle-complex decisions that require faster responses, such as the one happening in the TL administration. Finally, the TTier devices are not suitable for making big autonomous decisions as their computational capacity enables quick, narrowed resolutions that can be helpful in specific situations. Therefore, it can be well suited to local and operational decisions from the OL. As both the hierarchical levels and the CIoT can be classified and understood based on response time, data aggregation, and decision capacity and responsibility, we hypothesize that there might be a correlation between strategic, tactical, and operational decisions and how they are positioned in the CIoT. Also, we understand that the dynamism and complexity of the HCS create some intersection and overlay between these hierarchies organizations and that the flexibility and dynamism of the data processing layers enable a fluctuation of application among the SL, TL, and OL levels of a HCS organization. The IT interconnection and the presence of devices and technologies based on CIoT is also being recognized as a solution for underlying governance policymaking. The integration and exchange of health information and data helps predict patterns, finding optimized solutions and establishing healthcare as a service and an infrastructure working in a cloud-based system. Regarding policies, technology is not only influencing decision-making and operational performance but how the system itself is going to behave in terms of data integration and security. Therefore, understanding how a government administered public healthcare system currently works becomes a necessary step to understand how it can be inserted and placed in a smart environment.

Healthcare Applications The way a healthcare system is organized dictates how its multiple healthcare applications are organized and their decisions are spread out on the hierarchical strategic framework. Therefore, before untangling these applications, it is important to highlight which type of system will guide this work. Worldwide, there are only 15 countries that implement universal public healthcare with full state support. Among them, Brazil is a country of continental dimensions with the fifth great world population characterized by its multi-ethnic aspect and a recent democratization process. Brazil has a structured political system composed of multiple parties and three levels of independent governments that make decisions at federal, state, and municipal scales. As the country started to establish its political dynamic, it has developed a dynamic, complex public health system (the Unified Health System or SUS), which is based on the principles of health as a citizen’s right and the state’s duty. This system allows coordination, integration, and resource transfers among its three decision levels as it is based on a decentralized policy that grants autonomy of the decision to each one of them (federal, state, and municipality) committed and guided by a set of universal health policies, goals, and responsibilities. Moreover, the delivery of care set by this HCS is threefold into primary, secondary, and tertiary levels of care integrating public and private health organizations.

542

A. R. da Rocha et al.

To manage the decentralized policy, frameworks for government decisionmaking were expanded. The extension of these levels (federal, state, and municipal) is similar to the hierarchical strategic framework (strategic, tactical, and operational) that is of interest in this discussion. Also, the political structures of the Brazilian health system were groundbreaking innovations as they enabled a greater number and variety of stakeholders to take part in the decision-making process and defined areas of institutional responsibility more clearly, making the SUS a role model health system worldwide. Like any other HCS, the Brazilian healthcare system (or SUS) is defined by its complex behavior, multiple nature, and heterogeneous applications. As its premise is integration and strategic coordination, there is an urge for efficient management that requires the support of informatics, data, and information technology systems and infrastructure. However, in the process of digitization versus digitalization, most of the SUS is still facing the digitization phase (Gava et al. 2016). This lack of support and capabilities is still one of the biggest challenges facing hospital and healthcare management personnel, healthcare professionals, and informatics experts alike in Brazil, both in terms of procedural design and accessibility. Therefore, it is important to position the SUS as a complex but still most analogic health system that, to the best of our knowledge, could benefit from the exploration of its variety of data points and sources in a cloud-based system. Given this context, in this section, we will attempt to draw the correlations between the levels of cloud technology and strategic decisions to propose an architecture framework that distributes the decisions of the HCS around the different layers of a cloud-assisted IoT-based system. To this end, the Brazilian HCS will be represented and described by means of some of its applications in which each performs a decision-making process in healthcare management. The applications were sorted out based on their decision level and described according to their current performance policies on the Brazilian healthcare system as stated by the Healthcare Law (Law 8080/90) (BRASIL 2011), Ministry of Health (MOH) (BRASIL 2020), Family Health Programme, Brazilian Mobile Emergency Care Service (Ministério da Saúde 2003), Pact for Health (Ministério da Saúde 2006a, b, c, d, e, f, g, h, i), DATASUS (Ministério da Saúde 2006a), and Localiza SUS (Ministério da Saúde 2020a). Application A: Healthcare system budget planning: the HCS responsibility is decentralized and divided among the three governmental spheres and their counsels and secretariats. In the union sphere, the SUS is administered by the Ministry of Health, responsible for levying taxes and social contributions and repassing them to state and municipal funds. It is also responsible for defining the Global Financial Ceiling (GFC), which value, for each state and each municipality, is defined based on the Granted and Integrated Programming (GIP). In the states, the SUS is administered by the Regional Health Departments and State Secretariat of Health, which is responsible for controlling the health budget for the entire state and repasses resources to the municipalities as well as levying taxes and social contributions for the state. Also, the states manage the States’ Financial Ceiling (SFC), which is composed of the sum of the Financial Ceilings of Assistance

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

543

(FCA), Health Surveillance (HS), and Epidemiology and Disease Control (EDC). In the municipalities, the SUS is administered by the Municipal Secretariat of Health, which is responsible for managing the resources received by the federal and state spheres and repass a fraction of the municipal budget to health programs. All decisions and budget plannings are independent but yet integrated with the superior spheres. Application B: Capacity planning of ambulance services: the Mobile Emergency Medical Service (SAMU) aims at guaranteeing care delivery, appropriate transport, and patient referral to the SUS and can be free requested by telephone nationwide. The MOH is held accountable for acquiring the ambulances and equipment for basic and advanced life support so they could be later transferred to states and municipalities. Application C: Ambulance bases location decision: the base location decision is made at the municipal level and takes into consideration response time, demand points, distance from points of interest (hospitals and 24-hour Emergency Care Units, for example), and relocation costs. Removing ambulances from their old base to a new base would be an operational level decision (low degree of aggregation: for each ambulance). Application D: Ambulance orientation and diversion: ambulance diversion (AD) is required due to patient preference, lack of appropriate facilities, equipment or specialty, or (most common) crowding and overcapacity. In periods of the surge, be it predictable or not, AD is used by EDs and hospitals to reorientate patients and avoid tardiness and waiting. Application E: National Immunization Program (NIP): the NIP runs at the national level and is responsible for creating and managing vaccination campaigns, eradicating diseases in the country and elaborating strategic measures to acquire, distribute, and standardize the use of special immunobiological agents. Application F: Allocation (distribution) of medicaments and drugs or materials to each link (hospital) in the healthcare supply chain: the National Medicine Policy, as any other Brazilian healthcare policy, contains the responsibility of each government hierarchical level. It is the duty of the municipality and their secretariats, through integration with state and federal government, to supply and provide required medicaments to the hospitals under their administration. These medications must follow the National List of Essential Medicines (RENAME). Application G: Scheduling of elective surgeries in a surgical center: Scheduling surgeries for a surgery center can be a tricky task of assigning start times to all surgery-related activities for each individual patient besides the surgery time itself, and it depends on capacity, resource availability, and urgency. It can be a strategic, tactical, or operational activity based on the approach taken. In this case, the focus is on elective surgeries planned within 1 week (or two). Application H: Hospital procurement planning: Brazilian hospitals spend over 50% of their budget on the acquisition of supplies and goods. According to Medeiros e Ferreira (2018), procurement and purchasing are a core activity in healthcare as it requires special care regarding storage and the risk of deterioration. Also, as the SUS purchasing process is decomposed throughout its hierarchical

544

A. R. da Rocha et al.

chain with decentralized responsibilities, this specific application is taking into consideration only the procurement planning of public hospitals under the SUS management. Application I: Hospital supply acquisition: This application refers to placing the order request of the required supplies at the end of the procurement planning. As for public hospitals, there is a public purchasing policy in which all the acquisitions of supplies and medications are done through public bidding. In this case, this application is the publication of the public bidding for the required supplies and goods. Application J: Organizational Human Resource Planning: the human resource for health management, regulation, and planning is the role of the Secretariat of Labor and Education Management in Health (SGTES), created and run under the MOH. Among several actions, the SGTES develop a comprehensive HR information system, plan and execute the Work Management National Policy for SUS, perform the work regulation, and oversee the HRH management at state and municipal level. All SGTES actions are at the federal and strategic level. Application K: Hospital planning of on-call teams and shifts: Hospital administration (a.k.a. nurse chief, clinic chief, resident chief, surgeon chief, etc.) is usually responsible for organizing on-call teams and work shifts of their staff. Because of the unpredictability of the demand, the dynamics of the work, and the public health employment flexibility, this is an operational and recurrent task in the hospitals. Shift and on-call schedules are done weekly. Based on these descriptions, Table 1 presents the complexity of the decision established according to the scope (federal, state, municipal, local), decision planning time frame (short, medium, and long term), level of data aggregation (low, medium, high), and hierarchical level of the decision for each application. As stated, the SUS is a decentralized system that distributes and shares its decisions throughout its hierarchical levels (federal, state, and municipal). Therefore, the same application can be taken on the strategic, tactical, or operational decisioning level depending on the impact of the decision, the decision-maker, and its coverage.

Applications Distributed into the CIoT Tiers Keeping in mind how the current state of the healthcare system works, this section will explore how the CIoT could be inserted in this scenario in the future and, therefore, how the applications might work then. According to how the applications are placed throughout the hierarchical levels, a parallel was established between them and the type or level of responsibility of each tier of the CIoT. This parallel is based on which level the decision would be made. To this end, a system was drawn in which the filled circle indicates where a final decision was made, the empty circle represents where a partial decision is being made, and the dash represents that there is no decision being made at this level. Table 2 presents how each application might distribute their decisions on each level. Decisions of applications running at the strategic level may be the responsibility of the CTier since they need more information and usually aggregates huge amounts

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

545

Table 1 Healthcare Management Applications classification HSC Applications A

Time frame 1 year

Level of data aggregation High

1 year

High

B

Scope Federal, state, and municipal Federal

C

Municipal 1 month

Low

D

Local

Low

E F

Federal 1 year Municipal 6 months

High High

G

Local

1/2 weeks

Low

H

Local

1 month

Low

I

Local

1 day

Low

J

Federal

1 year

High

K

Local

1 week

Low

Online

Hierarchical level of the decision References Strategic (Ministério da Saúde 1996; BRASIL 2011; Vargas et al. 2015; Segatto and Béland 2019) Strategic Ministério da Saúde (2003) and O’Dwyer (2013, 2016) Tactical Ministério da Saúde (1996, 2003) and Nogueira and Castilho (2016) Operational Pham et al. (2006), Ramirez-Nafarrate et al. (2014), and Baek et al. (2020) Strategic Domingues et al. (2012) Tactical Clavel et al. (2018) and Anjomshoa et al. (2018) Operational Riise et al. (2016), Clavel et al. (2018), and Anjomshoa et al. (2018) Tactical Longaray et al. (2018) and Medeiros and Ferreira (2018) Operational Longaray et al. (2018) and Medeiros and Ferreira (2018) Strategic Buchan et al. (2011) and Oliveira (2015) Operational Oliveira (2015), Alves et al. (2015), and McElroy et al. (2020)

of data that can be supported by the computational capacity of this level. Also, as the response time is longer, there is more time for processing and collecting data from different sources along the way. However, there are some exceptions. For instance, applications requiring tactical decisions can be placed in the CTier if they have a high degree of data aggregation, such as Application F. As the physical nodes of the FTier are closer to the applications, the FTier is often responsible for aggregating data for tactical decisions. Also, since the fog structures are not linearly arranged and they can have several levels that intermediate the application and the CTier, they are flexible to support decisions that need different response times. However, they may only be responsible for lower amounts of data than the amount required by strategic decisions.

546

A. R. da Rocha et al.

Table 2 Decision matrix throughout cloud-assisted IoT levels HCS applications A B C D E F G H I J K

CTier responsibility • • ◦ • • • • -

FTier responsibility ◦ ◦ • ◦ ◦ ◦ • • ◦ •

TTier responsibility • ◦ ◦ • -

TTier devices have low computational capacity and quick response time and can usually take over operational decisions. In most applications, they will be replacing mechanical tasks and releasing manpower and human resources that can put more effort into tactical and strategic activities. Another exception for this logical representation is Application G. Even though it refers to an operational decision and has a low aggregation of data, it needs information from higher levels, such as patient demand from the whole system (as SUS cover all the country); at the same time it needs data and information about assets availability and capacity, which is usually stored by the FTier as they are more dynamic. Therefore, the final decision is shared by the CTier and the FTIer. TTier devices do not have the computational capacity for this type of decision. These special cases demonstrate that, even though there is a similarity between the hierarchical levels and the CIoT structures, some flexibility can support the complexity and dynamism of the healthcare sector. This dynamism is closely related to the overlay virtualization and the sense that the healthcare applications can either run on a specific tier or be between layers. First of all, it is important to highlight that all CIoT levels are responsible for collecting and transmitting data, merging and aggregating data, and providing communication infrastructure to enable the decision-makers to collaborate in the decision-making process. Therefore, for instance, a purchasing decision made in Application I would be communicated to Application H, a level above on the hierarchical chain of decision.

CIoT VNs for the Healthcare Applications CIoT VNs can represent the data required from healthcare applications previously described. Each application VN contains a set of data types, and the VNs must be placed in the CIoT tier responsible for the decision. The responsible tier for each application was defined in Table 2. For Application A (strategic budget planning), for instance, the VN is composed of several types of data that come from the strategic steps for elaborating and

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

547

modeling the integrated regional planning. First, there is the re-evaluation and redefinition of the health regions that, based on cultural, financial, and social data, besides communication and transportation infrastructure data, group municipalities to elaborate similar health plannings and actions and define the minimum services of care present in each region. Then, there is the collection of data regarding the health situation of each region and the elaboration of the HealthMap, which aims to gather and systematize the region’s health information to serve as a basis for situational analysis. This step uses the data from current assistancialist parameters. Next, there is the definition of interventions based on guidelines, objectives, goals, and indicators that portray the commitments of federated entities in the organization, shared health actions, and services within the scope of the health region, with a focus on health outcomes. After, it is elaborated on the General Programming of Health Actions and Services, which allows harmonizing the physical and financial quantities of health actions and services to be developed within each region. With the definition of priority interventions, which includes the definition of the responsibilities and goals of each federated entity in the execution of actions and services in the health region, the definition of budgetary and financial responsibilities begins. The inputs that help define these responsibilities are the following: the Annual Health Program of each entity, the Action Plans of the Strategic Networks (Rede Cegonha, Urgency and Emergency Network, Psychosocial Care Network, among others) and the resources of the Surveillance in Health and Primary Care, among others. For the federal government, financial data depends on GFC, taxes, and social contributions. As the federal government is responsible for repassing the funds for the state and municipal spheres, it has to define and explicit the SUS financial automatic flows for each governmental sphere. Also, it has to communicate with the internal and external institutions and the health counselings. Due to the complexity of the system and the number of entries, the financial data is extensive. However, the decentralization of the HCS allows a better share of resources and data between each administration hierarchical level, which helps the information flow between the CIoT layers. Because of the huge amounts of data that compose the financial matrix, the VN of Application A is better suited in the CTier, where its decisions are made. It takes heavy computational resources not only to aggregate but to integrate the diverse amounts of information and data required. The inputs needed to define the financial budget can also be allocated in sub VNs in the FTier, supporting the CTier in this application. Each application has to define its own VN based not only on the data type but the processing methods for the data. The well definition of the VNs allows a functional node allocated in the required CIoT tier.

Agriculture 4.0 Application Domain In this section, the cloud-assisted IoT paradigm is presented as an infrastructure to support data management for Agriculture 4.0 domain applications. The combined use of the Internet of Things, virtualization of resources (sensors, robots, machines, aerial images), and machine learning are trends to create a digital representation of

548

A. R. da Rocha et al.

the farm and support decision-making in several essential agricultural activities. In Agriculture 4.0, one of the main goals is to improve productivity and reduce costs. A digital farm representation supports choosing an ideal supply chain management plan to apply minimum quantities of resources required and target particular areas, such as water, fertilizers, and pesticides. Several agriculture systems are used to choose an ideal supply chain management plan. Specifically, in precision irrigation, it is one of the most water-intensive agricultural activities globally, which have been increasing over time. The new trend blockchain integrated into the cloud-assisted IoT paradigm is applied in recent agriculture activities such as transportation and logistics, creating a unique interaction mode in the supply chain.

Definitions and Technologies Agriculture 4.0 is the fourth evolution in farming technology. Smart agriculture, smart farming, digital farm, and digital agriculture are terms frequently used interchangeably in this context. They refer to the digital representation of the farm, in which data are monitored and analyzed to support an accurate decision about the cultivation issues related to several agricultural activities: seeding, irrigation, harvesting, crop maturity rotation, storing, transport, and logistics. One common characteristic between these terms is the virtualization of sensors, robots, unmanned aerial vehicles, and machines to achieve a digital representation of the farm. Cloudassisted Internet of Things is frequently used as infrastructure to the virtualization approaches since it refers to the thing nodes connected to the Internet, which is able to monitor and share data. More thing nodes have been connected to digital farming in recent years, such as tensiometers and temperature sensors, weather stations, plant health sensors, drones, robots, and machines. For instantiation, various virtual thing nodes enrich the digital representation of the farms providing a large portion of data related to weather (and meteorological data), crop (type, cycle), irrigation system (water demand, irrigation period), plants health (diseases and plagues), and soil (moisture, temperature, nutrients). Thus, data volume, variety, and velocity are changing over the year on the fields. Analyzing and interpreting data and its intrinsic variability, both spatial-wise and time-wise (particular characteristic to the IoT devices), are more complex and accurate. Recently, the CIoT and machine learning have been integrated to support accurate decision-making. The data analysis module can interpret all different kinds of data related to the whole supply chain in the past and the present and perform predictions to make more accurate decisions. According to Zhai et al. (2020), there are four relevant areas to apply prediction techniques aiming to prepare farmers for the inherent agriculture uncertainty: (i) operations to be performed to improve product quality; (ii) climate change to manage crop and avoid unnecessary risks; (iii) potential symptoms and early signs about possible occurrences of pests and diseases to avoid crop losses; and (iv) market fluctuations (consumers demands and the price trend of agricultural products). In Agriculture 4.0, one of the main goals is to improve productivity and reduce costs. Digital farming supports choosing an ideal supply chain management plan to apply minimum quantities of resources required and target particular areas, such as

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

549

water, fertilizers, and pesticides. Several agriculture applications are used to choose an ideal supply chain management plan. Therefore, specific and well-known variables help estimate resource consumption or prepare farmers to handle uncertainty. Forecast of the specific variables can be crucial to enhance the management plan. Specifically, smart irrigation is one of the most water-intensive agricultural activities globally, increasing over time. Reference evapotranspiration is a crucial determinant for predicting water demand for smart irrigation applications. According to Braga et al. (2019), evapotranspiration is the simultaneous occurrence of evaporation and transpiration processes in a crop. This coefficient is usually an input parameter to machine learning techniques for making decisions related to irrigation systems. Several works (Braga et al. 2019; Torres et al. 2020; Nagappan et al. 2020) propose machine learning-based models to estimate the reference evapotranspiration using a reduced-feature model (fewer sensors) in comparison with the conventional methods used in the agriculture domain in order reduce systems costs. In this context, the evapotranspiration parameter plays a vital role in system agriculture to decisionmaking related to suitable irrigation to increase water-saving and reduce system costs. Leveraging machine learning to data analytics in CIoT applications is a potential approach to support accurate decision-making, forecasting several parameters, and handling uncertainties in digital farming (or similar terms frequently used interchangeably). However, machine learning techniques adopt historical data as a training set – historical data containing valuable information to improve decision accuracy. Therefore, in general, it is necessary for a previous step of farm monitoring and data set creation to enable learning successful experiences. How more extensive the historical database is, more machine learning approaches are enabled, and the greater the chance of learning successful experiences.

Applications Several CIoT applications for digital farming have the challenge of optimizing and automating agricultural processes such as irrigation, transportation, and logistics. In this context, some developed applications and its CIoT infrastructure are discussed as follows. Smart Irrigation Application Campos et al. (2020) propose the SmartGreen, a cloud-assisted Internet of Things framework, which provides soil and meteorological data monitoring to support decision irrigation management. The framework performs data analytics to obtain precision irrigation information. The solution also provides a prediction model based on machine learning to estimate the matric potential using meteorological, crop, and irrigation data. The prediction model can be used in fields without (partially or totally) tensiometers to determine the moisture used in an irrigation management plan. Regarding the CIoT infrastructure, data collecting and monitoring are performed on the thing nodes equipped with tensiometers on the lowest layer. Thing nodes are powered with small solar panels. Virtual nodes enrich the framework with

550

A. R. da Rocha et al.

moisture data in three depths to analyze the moisture on different soil levels and, consequently, irrigation needs. Climatic data (solar radiation, evapotranspiration, rain precipitation) from virtual nodes are also taken into account to support irrigation decisions. Concerning actuators, thing nodes can be installed in the fields such as the water pumping system, in which virtual thing node can control the surfaces to be irrigated (period and start time), the pressures to be expected, and the rates of drippers. Processing, IoT analytics, and decisions are performed on the edge or cloud nodes. These activities of the SmartGreen are distributed on these two upper layers. Based on Torres et al. (2020), the framework identifies and removes outliers and uses a fusion data mechanism distributed in levels to optimize data quality and improve decision-making. Specifically, in Agriculture 4.0, fog nodes are not always a resource-rich device compared to cloud nodes. Fog nodes can have constrained computation and mainly constrained communication in comparison to a cloud node. Farms are far from cities in general; therefore, there are issues related to Internet connectivity or cellular networks to deliver a reliable service. Moreover, in general, several digital farming solutions include low-cost hardware (e.g., Raspberry devices) as geo-distributed fog nodes managed in a distributed network. The farm can be a large region, and thus, it can be monitored, controlled, and managed divided into regions. Fusion data and partially distributed decisions performed at the SmartGreen fog level are also beneficial since they can decide and actuate in irrigation systems or punctual decisions and day-to-day terms decisions. It is essential to mention that fog nodes are geographically internal to the farms, significantly reducing the traffic to the cloud node, processing time, and latency. In general, it occurs a holistic vision of the context and sophisticated decisions in the cloud nodes. The prediction model aforementioned is a module implemented on cloud nodes. Logistic and Transport Application Torky and Hassanein (2020) track as one of the top predictions the integration of Blockchain with IoT in global supply chain management by 2030. The new trend blockchain integrated into the cloud-assisted IoT paradigm is applied in recent agriculture activities, creating a unique interaction mode in the supply chain. In this context, several emerging approaches for logistics and transportation have been discussed. Danco Davcev and Mitreski (2018) present an approach that leads to trust, transparency, and information flow to the applications and services within the food supply chains. Blockchain technology addresses security to the supply chain system in the entire process cycle of production, transport, and food distribution on the market. Blockchain is a distributed ledger for recording transactions with a timestamp between several devices in a network. Blockchain identifies the unique device entities, and the identity information is immutable. Thing nodes are responsible for monitoring the crops, farms, and weather. Data processing occurs at the fog nodes in the CIoT infrastructure. A trusted packaging identifying the product is available. The identifier is associated with the food transport data, which must be available to the consumer, about reliability and safety. Thus, the consumer can order, monitor, and control the food products specifying the source, quality,

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

551

delivery place, and other requirements. Consequently, blockchain enhances the management capabilities allowing better interactions of farmers/producers with the consumers. So, this technology enables every transaction to be recorded, replicated, and available for each part of the food supply chain and communicate with consumers via CIoT infrastructure. Regarding the payment, blockchain can also help the customer enabling the crypto payment for agro trading or logistics, as presented in Dasaklis et al. (2019). Despite the benefits of leveraging the blockchain in Agriculture 4.0 applications, it is essential to mention that scalability is a well-known challenge of blockchain technology. Moreover, the decentralized algorithm of blockchain often requires extensive resource consumption (processing and energy), while thing nodes are resource-constrained. For digital farming, the integration of CIoT infrastructure in three tiers (things, fog, and cloud nodes) and blockchain is a potential design pattern, and it is applied in several other works. In this context, blockchain can work as a database from different nodes (things, fog, and cloud) and monitor the transactions. According to Torky and Hassanein (2020), blockchain can be leveraged for providing trust and security as missed links in this kind of integration approach. On the other hand, fog and cloud nodes are acceptable to overcome the resource-constrained issues of the thing nodes.

Challenges and Future Directions This section aims to review and discuss the general challenges that emerge from applications based on CIoT infrastructure. Finally, it points out future research directions and specific challenges for cloud-assisted Internet of Things applications in the Industry 4.0 paradigm discussed throughout the chapter: manufacturing, healthcare, and agriculture. The virtual thing node concept presented in the CIoT system provides a digital representation of a physical thing concept. In a holistic view, a set of VNs provides the digital representation of a specific application domain: manufacturing, healthcare, or agriculture. A common challenge between the Industry 4.0 application domains is to specify and develop a significant number of VNs with the actual states of the corresponding physical things representing the desired domain adequately. This aspect involves interoperability, complexity, cooperation in a standardized way, efficient modeling, data formats, and data analytics. Besides, suitable models still are lacking in many domains. Moreover, the intrinsic constraint resources (computing, storage, and energy) of thing nodes result in some challenges, such as developing security and interoperability requirements. Heterogeneous types, massive volume, and real-time velocity of data produced by VNs result in data analytics challenges. Besides, video and image sensors are prevalent in Industry 4.0 (e.g., diseases and plagues for image detection). Local processing of this kind of data is challenging. Offloading to fog or cloud nodes can be a solution. However, the transmission of the tremendous volumes of data is a challenge since high bandwidth consumption is a bottleneck of wireless communication networks and

552

A. R. da Rocha et al.

resource constraint devices (mainly energy). A hybrid approach of local processing and offloading is often successful. Other issues that are set under consideration are inherent to the Industry 4.0 application domain discussed in this chapter: manufacturing, healthcare, and agriculture. Concerning the manufacturing domain, a well-known challenge is the highly fragmented data generated in several product life cycle processes across industrial sectors, Dai et al. (2019). Therefore, further sharing of data is required for improving data fusion and data analytics techniques. Upgrading of factories is another challenge since thing nodes firmware might be upgraded massively and periodically due to security patches. Firmware updates are usually downloaded from a central server, and it is a bottleneck limiting according to Dai et al. (2019). The main concern and challenges in the healthcare domain are ensuring data security (confidentiality, authentication, and integrity) and preserving patient privacy. The vulnerability of CIoT infrastructure poses security challenges not even in the thing nodes but also in the edge network (fog nodes) and cloud environments. Moreover, the requirements of healthcare applications are defiant because they often need highly reliable transmissions and real-time delays to deliver medical data. Moreover, the requirements of healthcare applications are defiant because they often need highly reliable transmissions and real-time delays to deliver medical data. Regarding Agriculture 4.0 domain, well-known challenges in farms are discussed several times in the last years, such as intermittent communication issues, energy constraints devices, distance from cities, and markets. Other challenges that emerge from this application domain are increased food productivity due to the population growth that directly impacts water scarcity and soil degradation. Moreover, according to Zhai et al. (2020), other upcoming challenges need to be also handled correctly: 1. User-friendly graphical user interface for farmers because they are not familiar with computer systems and algorithms; 2. Improving decision-making support for the life cycle of Agriculture 4.0 according to the duration of agricultural activities (short-term, mid-term, long-term) in an integrated way for day-to-day, seasonal, or yearly decisions to be more enriched. Respective examples are scheduling daily irrigation activities, a reasonable period for fertilizer application based on their experiences, and machinery pieces for replacement based on monitoring the status. Improving decisionmaking support can suggest, for example, integrating two of these activities into a new activity: fertigation, a process that combines fertilization, and irrigation. 3. Handling uncertainty mainly meteorological conditions that have a huge influence on crop and resource consumption (water and fertilizers); One future research direction is integrating blockchain and CIoT platforms for improving solutions related to the challenges discussed in this section. Blockchain can improve security on CIoT applications, specifically data integrity and privacy. Moreover, blockchain and CIoT in three-tiers is the apparent potential in distributed environments for Industry 4.0 applications. One motivator CIoT application is

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

553

product tracking since it is partially overlapped between manufacturing, healthcare, and agricultural domains. Product tracking applications based on blockchain and CIoT concepts can monitor the entire manufacturing chain, as can also monitor the logistics and transport of pharmaceutical products and the track food chain.

References Aazam M, Hung PP, Huh EN (2014) Smart gateway based communication for cloud of things. In: 2014 IEEE ninth international conference on intelligent sensors, sensor networks and information processing (ISSNIP). IEEE, Apr, pp 1–6. https://doi.org/10.1109/ISSNIP.2014. 6827673, http://ieeexplore.ieee.org/document/6827673/ Alves SMP, Coelho MC de R, Borges LH, et al. (2015) The flexibilization of employment relationships in the health sector: the reality in a Federal University Hospital in Brazil. Ciênc saúde coletiva 20:3043–3050. https://doi.org/10.1590/1413-812320152010.11592014 Alves MM, Pirmez L, Rossetto S, Delicato FC, de Farias CM, Pires PF, dos Santos IL, Zomaya AY (2017) Damage prediction for wind turbines using wireless sensor and actuator networks. J Netw Comput Appl 80:123–140. https://doi.org/10.1016/j.jnca.2016.12.027, http://linkinghub. elsevier.com/retrieve/pii/S1084804516303435, https://www.sciencedirect.com/science/article/ abs/pii/S1084804516303435 Anjomshoa H, Dumitrescu I, Lustig I, Smith OJ (2018) An exact approach for tactical planning and patient selection for elective surgeries. European Journal of Operational Research 268:728–739. https://doi.org/10.1016/j.ejor.2018.01.048 Atzori L, Iera A, Morabito G (2010) The internet of things: a survey. Comput Netw 54(15):2787–2805. https://doi.org/10.1016/j.comnet.2010.05.010, http://linkinghub.elsevier. com/retrieve/pii/S1389128610001568, https://www.sciencedirect.com/science/article/abs/pii/ S1389128610001568 Baek S, Lee YH, Park SH (2020) Centralized Ambulance Diversion Policy Using Rolling-Horizon Optimization Framework to Minimize Patient Tardiness. Healthcare 8:266. https://doi.org/10. 3390/healthcare8030266 Banijamali A, Pakanen OP, Kuvaja P, Oivo M (2020) Software architectures of the convergence of cloud computing and the Internet of Things: a systematic literature review. Inf Softw Technol 122(2018):106271. https://doi.org/10.1016/j.infsof.2020.106271, https://linkinghub. elsevier.com/retrieve/pii/S0950584920300215 Bellendorf J, Mann ZÁ (2020) Classification of optimization problems in fog computing. Futur Gener Comput Syst 107:158–176. https://doi.org/10.1016/j.future.2020.01.036, https:// linkinghub.elsevier.com/retrieve/pii/S0167739X19323568 Bonomi F, Milito R, Zhu J, Addepalli S (2012) Fog computing and its role in the internet of things. In: Proceedings of the first edition of the MCC workshop on mobile cloud computing – MCC ’12. ACM Press, New York, p 13. https://doi.org/10.1145/2342509.2342513, http://dl.acm.org/ citation.cfm?doid=2342509.2342513 Braga DJF, da Silva TLC, Rocha A, Coutinho G, Magalhães RP, Guerra PT, de Macêdo JA, Barbosa SD (2019) Time series forecasting to support irrigation management. J Inf Data Manag 10(2):66–80 BRASIL (2011) Dispõe sobre as condições para a promoção, proteção e recuperação da saúde, a organização e o funcionamento dos serviços correspondentes e dá outras providências. http:// www.planalto.gov.br/ccivil_03/leis/l8080.htm Buchan J, Fronteira I, Dussault G (2011) Continuity and change in human resources policies for health: lessons from Brazil. Hum Resour Health 9:17. https://doi.org/10.1186/1478-4491-9-17 Bülbül K, Noyan N, Erol H (2021) Multi-stage stochastic programming models for provisioning cloud computing resources. Eur J Oper Res 288(3):886–901. https://doi.org/10.1016/j.ejor. 2020.06.027, https://linkinghub.elsevier.com/retrieve/pii/S0377221720305701

554

A. R. da Rocha et al.

Campos NGS, Rocha AR, Gondim R, Coelho da Silva TL, Gomes DG (2020) Smart & green: An internet-of-things framework for smart irrigation. Sensors 20(1). https://doi.org/10.3390/ s20010190, https://www.mdpi.com/1424-8220/20/1/190 Clavel D, Xie X, Mahulea C, Silva M (2018) A Three Steps Approach for Surgery Planning of Elective and Urgent Patients. IFAC-PapersOnLine 51:243–250. https://doi.org/10.1016/j.ifacol. 2018.06.308 Chambers S (2015) Process types. Wiley, Ltd, Chichester. https://doi.org/10.1002/9781118785317. weom100221, http://doi.wiley.com/10.1002/9781118785317.weom100221 Chengjie Wu, Xu Y, Chen Y, Lu C (2012) Submodular game for distributed application allocation in shared sensor networks. In: 2012 Proceedings IEEE INFOCOM. IEEE, pp 127–135. https:// doi.org/10.1109/INFCOM.2012.6195490, http://ieeexplore.ieee.org/document/6195490/ Chirichiello A (2008) Two formal approaches for web services: process algebras & action languages. Ph.D. thesis, Sapienza Universita di Roma Dai H, Zheng Z, Zhang Y (2019) Blockchain for internet of things: a survey. IEEE Internet Things J 6(5):8076–8094. https://doi.org/10.1109/JIOT.2019.2920987 Danco Davcev ACVS Ljupco Kocarev, Mitreski K (2018) Blockchain-based distributed cloud/fog platform for IoT supply chain management. In: Proceedings of the eighth international conference on advances in computing, Electronics and Electrical Technology, pp 51–58 Dasaklis TK, Casino F, Patsakis C (2019) Defining granularity levels for supply chain traceability based on IoT and blockchain. In: Proceedings of the international conference on omni-layer intelligent systems. Association for Computing Machinery, New York, COINS ’19, pp 184– 190. https://doi.org/10.1145/3312614.3312652 Dinh T, Kim Y (2016) An efficient interactive model for on-demand sensing-as-a-services of sensor-cloud. Sensors 16(7):992. https://doi.org/10.3390/s16070992, http://www.mdpi.com/ 1424-8220/16/7/992 Domingues CMAS, Teixeira AM da S, Carvalho SMD (2012) National immunization program: vaccination, compliance and pharmacovigilance. Revista do Instituto de Medicina Tropical de São Paulo 54:22–27. https://doi.org/10.1590/S0036-46652012000700009 de Farias CM, Pirmez L, Delicato FC, Wei Li, Zomaya AY, de Souza JN (2013) A scheduling algorithm for shared sensor and actuator networks. In: The international conference on information networking 2013 (ICOIN). IEEE, pp 648–653. https://doi.org/10.1109/ICOIN.2013. 6496703, http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber%=6496703, http:// ieeexplore.ieee.org/document/6496703/ Farias CMD, Li W, Delicato FC, Pirmez L, Zomaya AY, Pires PF, Souza JND (2016) A systematic review of shared sensor networks. ACM Comput Surv 48(4):1–50. https://doi.org/10.1145/ 2851510, https://dl.acm.org/doi/10.1145/2851510 Fensel A, Tomic DK, Koller A (2017) Contributing to appliances’ energy efficiency with Internet of Things, smart data and user engagement. Futur Gener Comput Syst 76:329– 338. https://doi.org/10.1016/j.future.2016.11.026, https://linkinghub.elsevier.com/retrieve/pii/ S0167739X16306653 Fuentes L, Jiménez D (2005) An aspect-oriented ambient intelligence middleware platform. In: Proceedings of the 3rd international workshop on Middleware for pervasive and ad-hoc computing – MPAC ’05. ACM Press, New York, 2, pp 1–8. https://doi.org/10.1145/1101480. 1101482, http://portal.acm.org/citation.cfm?doid=1101480.1101482 Ganesarajah D, Lupu E (2002) Workflow-based composition of Web-services: a business model or a programming paradigm? In: Proceedings sixth international enterprise distributed object computing. IEEE, vol 2002-Janua, pp 273–284. https://doi.org/10.1109/EDOC.2002.1137716, http://ieeexplore.ieee.org/document/1137716/ Gava M, Ferreira LS, Palhares D, Mota ELA (2016) Incorporação da tecnologia da informação na Atenção Básica do SUS no Nordeste do Brasil: expectativas e experiências. Ciência & Saúde Coletiva 21:891–902. https://doi.org/10.1590/1413-81232015213.01062015, https:// www.scielo.br/j/csc/a/7H3q3wDnQcGNv9p7HdJH5LB/abstract/?lang=pt Ghobakhloo M, Fathi M (2019) Corporate survival in Industry 4.0 era: the enabling role of leandigitized manufacturing. J Manuf Technol Manag 31(1):1–30. https://doi.org/10.1108/JMTM-

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

555

11-2018-0417, https://www.emerald.com/insight/content/doi/10.1108/JMTM-11-2018-0417/ full/html Hopper A, Rice A (2008) Computing for the future of the planet. Philos Trans R Soc A Math Phys Eng Sci 366(1881):3685–3697. https://doi.org/10.1098/rsta.2008.0124, http://www.ncbi.nlm.nih.gov/pubmed/18672457, http://rsta.royalsocietypublishing.org/cgi/doi/ 10.1098/rsta.2008.0124, https://royalsocietypublishing.org/doi/10.1098/rsta.2008.0124 Khalid Z, Fisal N, Rozaini M (2014) A survey of middleware for sensor and network virtualization. Sensors 14(12):24046–24097. https://doi.org/10.3390/s141224046, http://www.mdpi.com/ 1424-8220/14/12/24046/ Khan I, Belqasmi F, Glitho R, Crespi N, Morrow M, Polakos P (2016) Wireless sensor network virtualization: a survey. IEEE Commun Surv Tutorials 18(1):553–576. https://doi.org/10. 1109/COMST.2015.2412971, http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber %=7060643, https://ieeexplore.ieee.org/document/7060643/ Longaray A, Ensslin L, Ensslin S, et al. (2018) Using MCDA to evaluate the performance of the logistics process in public hospitals: the case of a Brazilian teaching hospital. Intl Trans in Op Res 25:133–156. https://doi.org/10.1111/itor.12387 Li W, Delicato FC, Zomaya AY (2013) Adaptive energy-efficient scheduling for hierarchical wireless sensor networks. ACM Trans Sens Netw 9(3):1–34. https://doi.org/10.1145/2480730. 2480736 Li W, Delicato FC, Pires PF, Lee YC, Zomaya AY, Miceli C, Pirmez L (2014) Efficient allocation of resources in multiple heterogeneous wireless sensor networks. J Parallel Distrib Comput 74(1):1775–1788. https://doi.org/10.1016/j.jpdc.2013.09.012, http://linkinghub. elsevier.com/retrieve/pii/S0743731513002104, https://www.sciencedirect.com/science/article/ abs/pii/S0743731513002104 Li W, Santos I, Delicato FC, Pires PF, Pirmez L, Wei W, Song H, Zomaya A, Khan S (2017) System modelling and performance evaluation of a three-tier Cloud of Things. Futur Gener Comput Syst 70:104–125. https://doi.org/10.1016/j.future.2016.06.019, http://linkinghub.elsevier.com/ retrieve/pii/S0167739X16302047 Liu R, Wassell IJ (2011) Opportunities and challenges of wireless sensor networks using cloud services. In: Proceedings of the workshop on Internet of Things and Service Platforms – IoTSP ’11. ACM Press, New York, pp 1–7. https://doi.org/10.1145/2079353.2079357, http://dl.acm. org/citation.cfm?id=2079357 http://dl.acm.org/citation.cfm?doid=2079353.2079357 Liu Y, Dong S, Wei J, Tong Y (2020) Assessing cloud computing value in firms through sociotechnical determinants. Inf Manag 57(8):103369. https://doi.org/10.1016/j.im.2020.103369 Lueth KL (2020) State of the IoT 2020: 12 billion IoT connections, surpassing non-IoT for the first time. https://iot-analytics.com/, [Online; accessed 12 Dec 2020] Madria S, Kumar V, Dalvi R (2014) Sensor cloud: a cloud of virtual sensors. IEEE Softw 31(2):70–77. https://doi.org/10.1109/MS.2013.141, http://ieeexplore.ieee.org/xpls/abs_ all.jsp?arnumber=66623%51, http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber= 6662351, http://ieeexplore.ieee.org/document/6662351/ McElroy SF, Olney A, Hunt C, Glennon C (2020) Shift work and hospital employees: A descriptive multi-site study. International Journal of Nursing Studies 103746. https://doi.org/10.1016/j. ijnurstu.2020.103746 Medeiros M, Ferreira L (2018) Development of a purchasing portfolio model: an empirical study in a Brazilian hospital. Production Planning & Control 29:571–585. https://doi.org/10.1080/ 09537287.2018.1434912 Ministério da Saúde (1996) Norma Operacional Básica do Sistema Único de Saúde, https://bvsms. saude.gov.br/bvs/saudelegis/gm/1996/prt2203_05_11_1996.html Ministério da Saúde (2003) Institui o componente pré-hospitalar móvel da Política Nacional de Atenção às Urgências, por intermédio da implantação de Serviços de Atendimento Móvel de Urgência em municípios e regiões de todo o território brasileiro: SAMU- 192, https://bvsms. saude.gov.br/bvs/saudelegis/gm/2003/prt1864_29_09_2003.html Ministério da Saúde (2006a) Diretrizes Operacionais: Pactos pela Vida, em Defesa do SUS e de Gestão. Departamento de Apoio à Descentralização do Ministério da Saúde, Brasil, http:// bvsms.saude.gov.br/bvs/publicacoes/PactosPelaVida_Vol1DiretOperDefesaSUSeGestao.pdf

556

A. R. da Rocha et al.

Ministério da Saúde (2006b) Regulamento: Pactos pela Vida e de Gestão. Departamento de Apoio à Descentralização do Ministério da Saúde, Brasil, http://conselho.saude.gov.br/webpacto/ regulacao.pdf Ministério da Saúde (2006c) Regionalização solidária e cooperativa: Orientações para sua implementação no SUS. Departamento de Apoio à Descentralização do Ministério da Saúde, Brasil, http://bvsms.saude.gov.br/bvs/publicacoes/regionalizacao2006.pdf Ministério da Saúde (2006d) Política Nacional de Atenção Básica. Departamento de Apoio à Descentralização do Ministério da Saúde, Brasil, http://bvsms.saude.gov.br/bvs/publicacoes/ politica_nacional_atencao_basica_2006.pdf Ministério da Saúde (2006e) Diretrizes para a Programação Pactuada e Integrada da Assistência à Saúde. Departamento de Apoio à Descentralização do Ministério da Saúde, Brasil, http://bvsms. saude.gov.br/bvs/publicacoes/DiretrizesProgPactuadaIntegAssistSaude.pdf Ministério da Saúde (2006f) Diretrizes para a Implantação de Complexos Reguladores. Departamento de Apoio à Descentralização do Ministério da Saúde, Brasil, http://bvsms.saude.gov.br/ bvs/publicacoes/DiretrizesImplantComplexosReg2811.pdf Ministério da Saúde (2006g) Política Nacional de Promoção da Saúde. Departamento de Apoio à Descentralização do Ministério da Saúde, Brasil, https://bvsms.saude.gov.br/bvs/publicacoes/ politica_nacional_promocao_saude_3ed.pdf Ministério da Saúde (2006h) Diretrizes e Recomendações para o Cuidado Integral de Doenças Crônicas Não-Transmissíveis. Departamento de Apoio à Descentralização do Ministério da Saúde, Brasil, https://bvsms.saude.gov.br/bvs/publicacoes/diretrizes_recomendacoes_cuidado_ doencas_cronicas.pdf Ministério da Saúde (2006i) Política Nacional de Educação Permanente em Saúde. Departamento de Apoio à Descentralização do Ministério da Saúde, Brasil, https://bvsms.saude.gov.br/bvs/ publicacoes/politica_nacional_educacao_permanente_saude.pdf Ministério da Saúde (2020a) Portal da Saúde - SUS. In: DATASUS. http://www2.datasus.gov.br/ DATASUS/index.php. Accessed 10 Oct 2020 Ministério da Saúde (2020b) Localiza SUS. https://localizasus.saude.gov.br/. Accessed 10 Oct 2020 Nagappan M, Gopalakrishnan V, Alagappan M (2020) Prediction of reference evapotranspiration for irrigation scheduling using machine learning. Hydrol Sci J 65:2669– 2677 Nakamura EF, Loureiro AaF, Frery AC (2007) Information fusion for wireless sensor networks. ACM Comput Surv 39(3):9. https://doi.org/10.1145/1267070.1267073, http://portal.acm.org/ citation.cfm?doid=1267070.1267073 Nogueira DNG, Castilho V (2016) Health services waste: process mapping and cost management as strategies for sustainability in a surgical center. Revista de Gestao USP 23:362–375 O’Dwyer G, Konder MT, Machado CV, et al. (2013) The current scenario of emergency care policies in Brazil. BMC Health Services Research 13:70. https://doi.org/10.1186/1472-696313-70 O’Dwyer G, Machado CV, Alves RP, et al. (2016) Mobile prehospital emergency care: an analysis of implementation in the State of Rio de Janeiro, Brazil. Ciência & Saúde Coletiva 21:2189– 2200. https://doi.org/10.1590/1413-81232015217.15902014 Oliveira ALM (2015) Brazil: Case study on working time organization and its effects in the health services sector. Working Paper. ISBN: 978-92-2-130083-0, https://www.ilo.org/sector/ Resources/publications/WCMS_391010/lang--en/index.htm Perera C, Zaslavsky A, Liu CH, Compton M, Christen P, Georgakopoulos D (2014) Sensor search techniques for sensing as a service architecture for the Internet of Things. IEEE Sensors J 14(2):406–420. https://doi.org/10.1109/JSEN.2013.2282292, http://ieeexplore.ieee.org/ xpls/abs_all.jsp?arnumber=66055%18, http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm? arnumber=6605518, http://ieeexplore.ieee.org/document/6605518/ Pham XQ, Huh EN (2016) Towards task scheduling in a cloud-fog computing system. In: 18th Asia-Pacific network operations and management symposium, APNOMS 2016: management of softwarized infrastructure – proceedings. https://doi.org/10.1109/APNOMS.2016.7737240

20 Industry 4.0: Cloud–Assisted Internet of Things Applications and Challenges

557

Pham JC, Patel R, Millin MG, et al. (2006) The Effects of Ambulance Diversion: A Comprehensive Review. Academic Emergency Medicine 13:1220–1227. https://doi.org/10.1197/j.aem.2006.05. 024 Ramirez-Nafarrate A, Baykal Hafizoglu A, Gel ES, Fowler JW (2014) Optimal control policies for ambulance diversion. European Journal of Operational Research 236:298–312. https://doi.org/ 10.1016/j.ejor.2013.11.018 Rao BBP, Saluia P, Sharma N, Mittal A, Sharma SV (2012) Cloud computing for Internet of Things & sensing based applications. In: 2012 Sixth international conference on sensing technology (ICST). IEEE, pp 374–380. https://doi.org/10.1109/ICSensT.2012.6461705, http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber%=6461705, http://ieeexplore. ieee.org/document/6461705/ Ravina-Ripoll R, Nunez-Barriopedro E, Evans RD, Ahumada-Tello E (2019) Employee happiness in the industry 4.0 Era: insights from the Spanish industrial sector. In: 2019 IEEE technology & engineering management conference (TEMSCON). IEEE, pp 1–5. https://doi.org/10.1109/ TEMSCON.2019.8813623, https://ieeexplore.ieee.org/document/8813623/ Riise A, Mannino C, Burke EK (2016) Modelling and solving generalised operational surgery scheduling problems. Computers & Operations Research 66:1–11. https://doi.org/10.1016/j.cor. 2015.07.003 Sajid A, Abbas H, Saleem K (2016) Cloud-assisted IoT-based SCADA systems security: a review of the state of the art and future challenges. IEEE Access 4:1375–1384. https://doi.org/10.1109/ ACCESS.2016.2549047, http://ieeexplore.ieee.org/document/7445139/ dos Santos IL, Pirmez L, Lemos ÉT, Delicato FC, Vaz Pinto LA, de Souza JN, Zomaya AY (2014) A localized algorithm for Structural Health Monitoring using wireless sensor networks. Inf Fusion 15:114–129. https://doi.org/10.1016/j.inffus.2012.02.002, http://linkinghub.elsevier. com/retrieve/pii/S1566253512000152 Santos IL, Pirmez L, Delicato FC, Carmo LFRdC (2015a) Ensuring energy efficiency of power quality applications in smart grids through a framework based on Wireless Sensor and Actuator Networks. In: 2015 international wireless communications and mobile computing conference (IWCMC), pp 763–768. https://doi.org/10.1109/IWCMC.2015.7289179, http://ieeexplore.ieee. org/lpdocs/epic03/wrapper.htm?arnumber%=7289179 Santos IL, Pirmez L, Delicato FC, Khan SU, Zomaya AY (2015b) Olympus: the cloud of sensors. IEEE Cloud Comput 2(2):48–56. https://doi.org/10.1109/MCC.2015.43, http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber%=7116465, http://ieeexplore. ieee.org/document/7116465/ Santos IL, Pirmez L, Carmo LR, Pires PF, Delicato FC, Khan SU, Zomaya AY (2016) A decentralized damage detection system for wireless sensor and actuator networks. IEEE Trans Comput 65(5):1363–1376. https://doi.org/10.1109/TC.2015.2479608, http://ieeexplore. ieee.org/document/7274680/ Santos IL, Pirmez L, Delicato FC, Oliveira GM, Farias CM, Khan SU, Zomaya AY (2018) Zeus: a resource allocation algorithm for the cloud of sensors. Futur Gener Comput Syst. https://doi. org/10.1016/j.future.2018.03.026 dos Santos IL, Delicato FC, Pirmez L, Pires PF, Zomaya AY (2019) Resource allocation and task scheduling in the cloud of sensors. In: Ammari H (ed) Mission-oriented sensor networks and systems: art and science. Studies in systems, decision and control. Springer, Cham, pp 265–305. https://doi.org/10.1007/978-3-319-91146-5_8, http://link.springer.com/10.1007/9783-319-91146-5_8 Segatto CI, Béland D (2019) Federalism and decision making in health care: the influence of subnational governments in Brazil. Policy Studies 1–19. https://doi.org/10. 1080/01442872.2019.1634187, https://www.tandfonline.com/doi/abs/10.1080/01442872.2019. 1634187?journalCode=cpos20 Torky M, Hassanein AE (2020) Integrating blockchain and the internet of things in precision agriculture: analysis, opportunities, and challenges. Comput Electron Agric 105476. https:// doi.org/10.1016/j.compag.2020.105476, https://www.sciencedirect.com/science/article/abs/pii/ S0168169919324329

558

A. R. da Rocha et al.

Torres AB, da Rocha AR, Coelho da Silva TL, de Souza JN, Gondim RS (2020) Multilevel data fusion for the internet of things in smart agriculture. Comput Electron Agric 171:105309. https:// doi.org/10.1016/j.compag.2020.105309, https://www.sciencedirect.com/science/article/abs/pii/ S0168169919322999 Tsai WH, Lan SH, Lee HL (2020) Applying ERP and MES to Implement the IFRS 8 operating segments: a steel group’s activity-based standard costing production decision model. Sustainability 12(10):4303. https://doi.org/10.3390/su12104303, https://www.mdpi.com/20711050/12/10/4303 Vaquero LM, Rodero-Merino L, Caceres J, Lindner M (2008) A break in the clouds. ACM SIGCOMM Comput Commun Rev 39(1):50–55. https://doi.org/10.1145/1496091.1496100 Vargas I, Mogollón-Pérez AS, Unger J-P, et al. (2015) Regional-based Integrated Healthcare Network policy in Brazil: from formulation to practice KEY MESSAGES. Health Policy and Planning 30:705–717. https://doi.org/10.1093/heapol/czu048, https://pubmed.ncbi.nlm.nih.gov/ 24963157/ Verdouw C, Beulens A, van der Vorst J (2013) Virtualisation of floricultural supply chains: a review from an Internet of Things perspective. Comput Electron Agric 99:160–175. https://doi.org/10. 1016/j.compag.2013.09.006, https://linkinghub.elsevier.com/retrieve/pii/S0168169913002135 Wang B, Fan Ty, Nie Xt (2020) Advanced delay assured numerical heuristic modelling for peer to peer project management in cloud assisted internet of things platform. Peer-to-Peer Netw Appl 13(6):2166–2176. https://doi.org/10.1007/s12083-020-00883-9, http://link.springer.com/ 10.1007/s12083-020-00883-9 Wang Y, Ma HS, Yang JH, Wang KS (2017) Industry 4.0: a way from mass customization to mass personalization production. Adv Manuf 5(4):311–320. https://doi.org/10.1007/s40436017-0204-7, http://link.springer.com/10.1007/s40436-017-0204-7 Yi S, Li C, Li Q (2015) A survey of fog computing. In: Proceedings of the 2015 workshop on mobile big data – mobidata ’15. ACM Press, New York, pp 37–42. https://doi.org/10.1145/ 2757384.2757397, http://dl.acm.org/citation.cfm?doid=2757384.2757397 Yuriyama M, Kushida T (2010) Sensor-cloud infrastructure-physical sensor management with virtualized sensors on cloud computing. In: BT – 2010 13th international conference on network-based information systems. Network-Based Information Systems, pp 1–8. http:// ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=56356%88 Zhai Z, Martínez JF, Beltran V, Martínez NL (2020) Decision support systems for agriculture 4.0: survey and challenges. Comput Electron Agric 170:105256. https://doi.org/10.1016/j.compag. 2020.105256, https://www.sciencedirect.com/science/article/pii/S0168169919316497 Zhang L, Luo Y, Tao F, Li BH, Ren L, Zhang X, Guo H, Cheng Y, Hu A, Liu Y (2014) Cloud manufacturing: a new manufacturing paradigm. Enterp Inf Syst 8(2):167–187. https://doi. org/10.1080/17517575.2012.683812, https://www.tandfonline.com/doi/abs/10.1080/17517575. 2012.683812 Zhang Y, Chen Jl (2015) Constructing scalable Internet of Things services based on their eventdriven models. Concurrency Comput Pract Exper 27(17):4819–4851. https://doi.org/10.1002/ cpe.3469, http://doi.wiley.com/10.1002/cpe.3469, arXiv:1302.5679v1

Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

21

Kakali Bhadra

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Natural Alkaloids with Reference to Anticancer Effect: Its Sources and Types . . . . . . . . . . . Targeted Biomolecules by Anticancer Natural Alkaloids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anticancer Natural Alkaloids Targeting Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anticancer Natural Alkaloids Targeting Cellular Proteins, Enzymes, and Growth Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carbohydrate Target by Small Molecules as Anticancer Alkaloids . . . . . . . . . . . . . . . . . . . Anticancer Natural Alkaloids Targeting Lipid Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

560 562 563 564 576 582 583 585 585

Abstract

Alkaloids are low-molecular-weight nitrogen-containing natural products mostly found in plant and some microorganisms and restricted to few animals also. Historical evidence over time suggests the use of alkaloid both as a medicine and toxicant. The widespread distribution and their wide exhibit of structures make the classification of the alkaloids difficult. Several natural alkaloids display antiproliferation and antimetastasis consequences for different kinds of malignant growths both in vitro and in vivo. This section highlights on the naturally derived alkaloids with potential anticancer properties, highlighting the targeted biomolecules, viz., targeting nucleic acids of different conformation and its associated enzymes, cellular proteins, growth factors, carbohydrates, and lipids. The same natural alkaloids have been accounted to have different target biomolecules for the inhibition of progression and metastasis of specific cancer

K. Bhadra () Department of Zoology, University of Kalyani, Nadia, West Bengal, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_94

559

560

K. Bhadra

cells. Since cancer is a disease process driven by many aberrant oncoproteins related to multiple pathways of signal transduction, single-target therapeutic approach is not going to lead to a successful translational application. Thus, development of multitarget agents is an urgent quest for the treatment. Based on these informations, this study may be of prospective use in a framework to design novel anticancer drug molecules for improved therapeutic applications in the future. But further research and clinical trials are necessary before final acceptance. Keywords

Natural alkaloids · Anticancer · Targeted biomolecules · Nucleic acids · Cellular proteins · Lipid metabolism

Introduction Through adaptive evolution of hundred million years, several toxic plants, poisonous microorganisms, and venomous animal species have developed an enormous number of strong affinity and stable natural-based compounds, which have been isolated and identified with time. Some of these natural compounds possess the capacity to explicitly bind to human receptors/target and have an extraordinary potential to be used as novel model drug candidates for many varieties of diseases, including cancer (Habli et al. 2017). Designing drug molecules to ameliorate human disease is an overwhelming yet thrilling undertaking. In the last two and half decades, many biomolecular researches have been centered around certain types of distinctive molecular targets on cells, which can be perceived by natural-based biomolecules, to uncover their utilization in various parts of clinical forecast and remedial applications. Progress in biological sciences has led to the improvement of new targeted agents, including drugs targeting defective DNA repair, such as PARP inhibitors; drugs targeting angiogenesis; and immune checkpoint inhibitors such as anti-PD-1/PD-L1 antibodies, receptor proteins, and many more. There are four significant classes of natural macromolecules that make up most of a cell’s mass, viz., proteins 10–15%, lipids ∼2%, carbohydrates ∼3%, and nucleic acids 5–7%, and each one a significant part of the cell performing wide exhibit of capacities. Hence, these biomolecules have an expansive scope of applications in targeted drug delivery and have gone to the forefront in recent years. Figure 1 shows the action of few recently studied anticancer natural alkaloids suppressing cancer by modulating multiple signaling pathways targeting different biomolecules. Drug development describes the process of developing a new molecule that effectively targets a specific weak point in a cell. This process involves specific preclinical development and testing, followed by trials in as many person/humans as possible to determine the efficacy of the drug in terms of pharmacokinetic, pharmacodynamic, and pharmacogenomic status of the molecule to determine its ultimate therapeutic status. This chapter highlights the significance of natural alkaloids as

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

561

Fig. 1 Anticancer natural alkaloids restraining cancer by modulating multiple signaling pathways targeting different biomolecules

anticancer drug molecule targeting its specific biomolecule. We intend to reveal insight into promising developing characteristic alkaloids that could be conceivably converted into the clinical anticancer drug molecule. With time, there is rising enthusiasm, looking for better natural origin anticancer drug molecules with high effectiveness and low toxicity. The vast majority of the chemotherapeutic agents, due to their nontargeted actions and toxicity on normal healthy cells, are frequently confined, requiring search for newer drugs having more noteworthy possibility. Drug discovery from natural sources requires a multidisciplinary approach in which the achievement is to a great extent reliant on an all-round picked set of varied assays. Recently, in search for the development of new drugs, the pharmaceutical industry is progressively depending on bioinformatics. To upgrade and optimize the study time and being less expensive, the computational investigations have the bit of leeway over other techniques. There are two approaches in natural product-based drug discovery, viz., forward pharmacological approaches and reverse pharmacological approaches (Viana J de O et al. 2018). The forward pharmacological approach, which is a more traditional way of drug discovery, first determines functional activity by detecting phenotypic changes in complex biological systems based on in vitro and in vivo bioassays and then characterizes the molecular target of the active compounds, while the reverse pharmacology starts by identifying a promising pharmacological target against which compounds are screened, and then the obtained promising compounds are validated by in vivo and in vitro assay. The choice of the assays determined by the study objectives should optimally combine simplicity with good sensitivity and reproducibility.

562

K. Bhadra

Natural Alkaloids with Reference to Anticancer Effect: Its Sources and Types Historically, natural alkaloids have been a proven source of drugs with therapeutic potency, and even today, they are considered to be an important pool of novel drug leads for future. Next to terpenoids (∼28%), alkaloids (∼23%) are probably the largest groups of secondary plant metabolites, being present in several economically important plant families. Plant alkaloids, due to their toxicity, act as defense compounds against pathogens, predators, and herbivores. In general, toxic effects of the alkaloids are dependent on dosage of application, duration of exposure, and individual characteristics, such as affectability, site of activity, and formative stages. It is also suggested that alkaloids may serve as nitrogen storage and act as protective agents against harmful effects of ultraviolet light (Matsuura and FettNeto 2017). Plant alkaloids constitute both taxonomically and chemically extremely varied groups of compounds. As a result, many different types of classifications are possible. They are basic and contain one or more nitrogen atoms with a marked physiological action on different animals. Generally, in view of structures, alkaloids can be classified into quinolines, isoquinolines, indoles, tropanes, steroids, pyrrolidines, pyridines, and pyrrolizidines. Other grouping system is connected with a family of plant species that they occur. Broadly, it is divided into two groups: heterocyclic or typical alkaloid like hygrines, symphitine, nicotine, atropine, cocaine, quinine, berberine, sanguinarine, vinblastine, boldine, beta-carboline, etc. and non-heterocyclic or atypical alkaloid such as hordenine, colchicine, mescaline, etc. True alkaloids are of rare occurrence in lower plants like pteridophytes and gymnosperms. Among them, the lycopodium, ephedra (ephedrine), and Taxol alkaloids have medicinal interest (Eduardo Sobarzo-Sanchez 2015). Induced cytotoxicity of some alkaloids such as vinblastine, vincristine, vinflunine, topotecan, and Taxol have been clinically used in cancer therapy (Habli et al. 2017). FDA has already approved Taxol (paclitaxel), isolated from the bark of Taxus brevifolia, and vincristine, which is isolated from the leaves of Catharanthus roseus, as anticancer drug. Taxol and vincristine are used widely against pancreatic, breast, and nonsmall cell lung cancers (Seca and Pinto 2018; Zhang et al. 2014). These plant-based natural alkaloids prevent cancer mainly by regulating multiple signaling pathways targeting different biomolecules resulting in the restraint of the induction of carcinogenesis, initiation of cell cycle arrest, apoptosis, autophagy or differentiation, and inhibition of metastasis and angiogenesis (Lu et al. 2012). In most cases, amino acids such as lysine, arginine, tryptophan, tyrosine, and phenylalanine act as precursors for the plant-based alkaloids. A couple of changes can transform the above amino acid precursors to the secondary metabolites (O’Connor 2008). Furthermore, many cyanobacteria, fungi, algae, marine sponges, and tunicates have been found to be a source of alkaloids, many of which can possibly compete with potential anticancer plant alkaloids and are in clinical trials or in clinical use. They possess different anticancer activities including induction of apoptosis and cytotoxicity, antiangiogenic, antiproliferative, tubulin polymerization, and suppression of topoisomerase activities (Tohme et al. 2011). Steroidal alkaloid,

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

563

cephalostatin 1, isolated from marine tube worm Cephalodiscus gilchristi is one of the most potent anticancer small molecules affecting the mitochondrial membrane potential inducing apoptosis (Moser 2008). It also induces DNA fragmentation and inhibits protein synthesis (Dey et al. 2019). Two alkaloids derived from tunicates, aplidin and trabectedin, are already in phase II cancer clinical trials. Aplidin initiates ROS-mediated cytotoxicity with G1 arrest and PARP (poly ADPribose polymerase)-mediated apoptosis, causing cleavage of caspases 3 and 9 with the release of cytochrome c (Cuadrado et al. 2003; D’Incalci and Galmarini 2010), while trabectedin causes DNA alkylation, S-phase cell cycle arrest, RNA polymerase II breakdown, upregulation of CCL2 (a potent chemokine for monocytes), and downregulation of VEGF (vascular endothelial growth factor) and IL-6 (interleukin 6) (Tohme et al. 2011). Apart from these, marine tunicates are blessed with wide range of other anticancer alkaloids like ascididemin, granulatimide, lamellarin D, lissoclinidine B, and polycarpines (Anderson et al. 1997; Dassonneville et al. 2000; Fedorov et al. 2004; Pla et al. 2008). The molecules are reported to show various mode and mechanisms of anticancer effects including inhibition of topoisomerases 1 and II, cell cycle arrest, and generation of different apoptosisinducing factors like upregulation of BAX and P53. Marine sponges are equally rich sources of these biologically active secondary metabolites with anticancer effect of unique and diverse chemical structures. Among them, hemiasterlin, kuanoniamine A, neoamphimedine, renieramycin M, spermatinamine, and variolin B are few most important anticancer alkaloids (Tohme et al. 2011). Alkaloids such as halichondrins B and E are isolated from marine sponges like Halichondria okadai. Halichondrin E-7389 (eribulin mesylate) has been used against breast cancer (Fang 2004). Few marine algae like Lophocladia sp. produces lophocladines A and B that were tested for their cytotoxic effects on various cancer cell line like NCI-H460 lung cancer, neuro-2a neuroblastoma, MDA-MB-435 breast cancer cell line, and NCIH460 lung cancer cells (Gross et al. 2006). Marine cyanobacteria also possess number of bioactive secondary metabolites with anticancer potency. Apratoxin, hectochlorin, and lyngbyabellin are few alkaloids isolated from different genera of cyanobacteria species (Russo and Cesario 2012; Tohme et al. 2011). Some endophytic bacteria have been isolated to produce the anticancer alkaloid camptothecin and 9-methoxy camptothecin from Miquelia dentata Bedd. Camptothecin is a quinoline alkaloid extracted from ornamental tree Camptotheca acuminata, which is a classical inhibitor of eukaryotic topoisomerase I (Heinrich et al. 2012; Shweta et al. 2013).

Targeted Biomolecules by Anticancer Natural Alkaloids Cancer is a broad group of diseases, involving uncontrolled cell growth. Through the formation of novel oncogenes, the improper overexpression of normal oncogenes, or by the underexpression of tumor suppressor genes, malignant transformation occurs. It is through the lymphatic or bloodstream that the cancer spread to other distant parts of the body (Hanahan and Weinberg 2000; Wu et al. 2016). As a

564

K. Bhadra

result, different cell, tissue, and organ systems are affected. Hence, cancer therapy varies by patient, tumor type, and location. Among its different treatment options, chemotherapy is one of the treatment with one or more cytotoxic antineoplastic drugs or chemotherapeutic agents as part of a standardized regimen. Chemotherapy in the form of targeted therapy is the highlighted part of this study that targets special molecular differences between cancer and normal cells. Figures 2, 3, and 4 represent few anticancer natural alkaloids isolated from plants, marine invertebrates, and microorganisms, and Table 1 shows the collated list of anticancer natural alkaloids and their targeted biomolecules as described in the text in detail.

Anticancer Natural Alkaloids Targeting Nucleic Acids DNA/RNA → mRNA → protein, this represents the general gross outline of the transfer of information within the cell. The processes of differentiation and carcinogenesis may both be viewed as alterations of this information transferred (Goldthwait 1960; Maiti and Suresh 2010; Brahmachari 2015). The cancer cell possesses a hereditary defect in the normal mechanism of control of cell division. Since nucleic acids are chemical basis of heredity in the cell, it is sensible to presume a modification in nucleic acids in the process of carcinogenesis. Deoxyribonucleic acid (DNA) functionally serves as the long-term storage of genetic information of the cell which leads to the cellular target of many therapeutic molecules (Maiti and Suresh 2010; Brahmachari 2015). Equally significant are ribonucleic acids (RNA). The emerging information of the essential role of cellular RNA in widespread biological processes and as an integral part of many viral genome has prompted a developing enthusiasm in the molecule as a possible target for therapeutic involvement (Bhadra and Suresh Kumar 2011a). The interest in understanding the basics of DNA and RNA interaction with small molecules results from the need to develop new compounds, particularly natural product, which have the ability to target the genome artificially at specified sequences or structures for curing many genetic diseases including cancer through efficient and less toxic chemotherapy. Even multifunctional nucleic acid, Apta-miRNA, is a modified noncoding oligonucleotides, base-pairing with individual microRNA molecules, and is applied for tumor cell treatment (Pofahl et al. 2014). The efficiency and therapeutic potential of biologically active small molecules very much depend on to identify their bio-target especially the active sites, its strength, and specificity of binding. Alkaloids represent such class of interesting small molecules abundantly available in nature showing remarkable medicinal applications as anticancer agents.

Some of the Biologically Significant and Important Nucleic Acid-Binding Anticancer Natural Alkaloids DNA interaction with small molecule can be classified into irreversible covalent interactions and reversible non-covalent interactions. Covalent interaction essentially involves base modifications, alkylation, and cross-linking of strands, strand breakage, etc., resulting in mutations. Non-covalent associations can be further

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

565

Table 1 List of anticancer natural alkaloids and their targeted biomolecules Compounds 2-D-glucose paclitaxel

Aaptamine

Apratoxin A and B

Sources Synthetic analogue of Taxol (isolated from the bark of Taxus brevifolia) A benzonaphthyridine alkaloid extracted from the marine sponge Aaptos suberitoides Marine cyanobacteria

Ascididemin, Marine natural granulatimide, alkaloids lamellarin D, lissoclinidine B, and polycarpines

Berberine

Targeted biomolecules/mode of action Target glucose transporters, GLUT-1

References Hossain and Andreana 2019

Upregulate p21 transcription in p53-mutated human osteosarcoma MG63, inducing G2/M cell cycle arrest

Tohme et al. 2011

Block the fibroblast growth factor receptor (FGFR) pathway via the inhibition of the phosphorylation of the signal transducer and activator of transcription 3 (STAT3) in a time- and dose-dependent manner, thus inhibiting the upregulation of antiapoptotic genes like BclxL and the activation of the cyclin-dependent kinase CDK2. This lead to antiangiogenic effects due to its ability to antagonize the FGFR pathway. Inhibit topoisomerases I and II

Russo and Cesario 2012; Tohme et al. 2011

Fan et al. 2019; Anderson et al. 1997; Dassonneville et al. 2000; Fedorov et al. 2004; Pla et al. 2008 Peumus boldus DNA/RNA damage and p53-dependent Maiti and and Lindera apoptosis; targeting G-quadruples; Suresh, 2010; aggregata. telomerase inhibition; induces MDM2 Brahmachari Distributed in self-ubiquitination and degradation by 2015; Bhadra several botanical inhibiting MDM2-DAXX-HAUSP and Suresh families like interactions; modulation of histone Kumar 2011a; Berberidaceae, deacetylase to induce growth arrest and Bhadra and Papaveraceae, apoptosis; inhibitor of Hsp90; reported to Kumar 2011b; Fumariaceae, suppress the growth of colon cancer cell Ganesan and Xu Menispermaceae, lines, HCT116 and KM12C, by inhibiting 2018; Zhang and many others the glucose uptake and transcription of et al. 2010; Song glucose metabolic genes, GLUT1 (glucose et al. 2018; transporter 1), LDHA (lactate Arunachalam dehydrogenase A), and HK2 (hexokinase et al. 2016; Siva 2); improves hyperlipidemia and fatty liver and Babu 2014; by the activation of AMPK (AMPMao et al. 2018; dependent protein kinase) in liver and Gao and Chen muscle tissues. It also regulates the lipid 2015; Kim et al. metabolism through inhibition of fatty acid 2009 synthetase and 5-tetradecyloxy-2-furoic acid (TOFA), inducing apoptosis in cancer cell. (continued)

566

K. Bhadra

Table 1 (continued) Compounds Betaine

Sources Is found in many foods, including spinach, beets, and whole grains

Boldine

Peumus boldus and Lindera aggregata Steroidal alkaloid Increases Bax protein expression; suppresses Bcl-2, caspase-3 initiation, and segmentation of poly (ADP-ribose) polymerase-1 (PARP-1) in HeLa cells Extracted from Classical inhibitor of eukaryotic ornamental tree topoisomerase I Camptotheca acuminata Isolated from Affecting the mitochondrial membrane marine tube potential inducing apoptosis. It also worm induces DNA fragmentation, inhibits Cephalodiscus protein synthesis, and uses gilchristi Smac/DIABLO for inducing apoptosis. Chelidonium Inhibiting the telomerase activity and majus L. promoting cancer cell death via binding with human telomeric G-quadruplex structure Isolated from Topoisomerase inhibition Cryptolepis sanguinolenta Ocotea Topoisomerase inhibition leucoxylon A quinolone Reported to inhibit STAT3 alkaloid, isolated phosphorylation and regulate from the Chinese downstream molecules to induce herb Evodia apoptosis in breast cancer cells; induces rutaecarpa both caspase-dependent and caspase-independent apoptosis, downregulates Bcl-2 expression, and upregulates Bax expression in some cancer cells; reported to induce cell death by the Fas ligand (Fas-L)/NF-κB signaling pathways Isolated from Is a selective inhibitor of CDK4 with marine sponge IC50 value of 0.35 μM. Based on molecular modelling, it is reported to inhibit CDK4 by binding to the ATP pocket of the kinase.

Briofilin

Camptothecin

Cephalostatin 1

Chelerythrine

Cryptolepine

Dicentrinone Evodiamine

Fascaplysin

Targeted biomolecules/mode of action Improve lipid homeostasis in obesity by participation in mitochondrial oxidative demethylation; reported to alleviate hepatic lipid metabolism via regulating the expression profile of microRNA-182 in nonalcoholic fatty liver disease Telomerase inhibition

References Sivanesan et al. 2018; Zhang et al. 2020

Ganesan and Xu 2018 Dey et al. 2019

Heinrich et al. 2012; Shweta et al. 2013 Moser 2008; Dey et al. 2019

Liu et al. 2017

Fan et al. 2019

Fan et al. 2019 Lu et al. 2012

Song et al. 2018

(continued)

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

567

Table 1 (continued) Compounds Halichondrin B and E

β-Carboline alkaloids: Harmalol, harmaline, and harmine

Hemiasterlins

Hymenialdisine and debromohymenialdisine

Largazole

Lunacridine Lycorine

Manzamine

Sources Marine sponge-derived alkaloid isolated from Halichondria okadai Isolated from the Banisteriopsis caapi (family: Malpighiaceae) and Peganum harmala (family: Zygophyllaceae)

Targeted biomolecules/mode of action References Topoisomerase inhibition; shows tubulin Tohme et al. polymerization action 2011

Induce nucleic acid fragmentation, cellular ultra morphological changes, decreased mitochondrial membrane potential, upregulation of p53 and caspase-3, generation of ROS (reactive oxygen species), a significant increase in G2 /M population, and topoisomerase inhibition Derived from the Exhibit antiproliferative effects against marine sponge murine leukemia p388 cells, human Hemiasterella glioblastoma U373, human mammary minor carcinoma MCF-7, and human ovarian carcinoma HEY by binding to tubulin at nanomolar concentrations and inhibit their polymerization, causing mitotic catastrophe and ultimately apoptosis Isolated from the Inhibit cyclin-dependent kinases marine sponge through competitive inhibition at the Stylotella ATP-binding site. These two aurantium compounds were known to be active in a wide range of CDKs, particularly CDK1, CDK2, and CDK5. Macrocyclic Inhibition of histone deacetylases depsipeptide (HDAC) alkaloid from cyanobacteria Lunasia Amara Topoisomerase inhibition Found in various Targeting histone deacetylases (HDAC) Amaryllidaceae species, such as the cultivated bush lily (Clivia miniata), Lycoris, and daffodils (narcissus) Isolated from Noncompetitive inhibitor of ATP with sponge binding to GSK-3β at IC50 value of Acanthostrongy- 10.2 μM lophora

Sarkar and Bhadra 2018; Bhattacharjee et al. 2018; Sarkar et al. 2020; Ahmad et al. 2020 Tohme et al. 2011

Song et al. 2018

Tohme et al. 2011

Fan et al. 2019 Li et al. 2012

Song et al. 2018

(continued)

568

K. Bhadra

Table 1 (continued) Compounds Matrine

Sources Found in Sophora flavescens Ait and S. japonica Meridianins A–G A group of marine indole alkaloids isolated from marine tunicate Aplidium meridianum Naamidine A sponge-derived alkaloid identified from Leucetta chagosensis

Targeted biomolecules/mode of action It affects proteins involved in cell proliferation or apoptosis, such as E2F-1, Bax, Bcl-2, Fas, and Fas-L

References Lu et al. 2012

Selectively inhibit CDK1, CDK5, and Song et al. 2018 other various protein kinases involved in cancer and other disease

An apoptosis-targeting natural Tohme et al. compound; provokes cell cycle arrest at 2011 G1 phase through stabilization of the p21 cyclin-dependent kinase inhibitor (CDK2), activation of the initiating caspases 8 and 9, and PARP cleavage Neoamphimedine Xestospongia sp. Topoisomerase inhibition Tohme et al. 2011 Nuciferine Found in Improves dyslipidemia in vivo through Zhang et al. Nelumbo PPARα/PPARγ coactivator-1α pathway 2018 nucifera leaves Oxymatrine Found in sophora Reported to alleviate hepatic lipid Zhang et al. roots primarily metabolism via regulating the 2020 from Sophora expression profile of microRNA-182 in japonica nonalcoholic fatty liver disease (kushen) Palmatine Distributed in Photooxidation of DNA, targeting Maiti and several botanical G-quadruples Suresh, 2010; families like Brahmachari Berberidaceae, 2015; Bhadra Papaveraceae, and Suresh Fumariaceae, Kumar 2011a; Menispermaceae, Bhadra and and many others Kumar 2011b Piperine and Extracted from Reported to inhibit STAT3 Chen et al. 2020; piperlongumine Piper nigrum L. phosphorylation and regulate Tinoush et al. and Piper downstream molecules to induce 2020; Matsuda longum L., apoptosis in breast cancer cells; et al. 2008 respectively targeting markers related to cancer drug resistance; inhibit lipid droplet accumulation in mouse macrophages by inhibiting cholesteryl ester synthesis Alkaloids with From the marine Novel HIF-1α/p300, transcription factor, Song et al. 2018 pyrroloiminosponge inhibitors quinone ring Latrunculia sp. structure Saframycin A Isolated from Forms a nuclear ternary complex with Gao and Chen marine tunicates GAPDH (glyceraldehyde 3-phosphate 2015; Azam dehydrogenase) and DNA et al. 2008 (continued)

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

569

Table 1 (continued) Compounds Sanguinarine

Sources Isolated from the Papaveraceae family, which includes Sanguinaria canadensis L. and Chelidonium majus L.

Targeted biomolecules/mode of action DNA/poly(A)/poly(U).Poly(A)* poly(U) binding, targeting G-quadruples, perturb microtubule assembly dynamics through tubulin binding, and induces intrinsic apoptosis in the presence of Bax. The weakened interaction of MELK and STRAP is necessary for the transactivation of Bax from cytosol to mitochondria. Induces apoptosis in the multiple myeloma cells via inhibition of JAK2/STATs signaling; potent suppressor of NF-κB activation induced by TNF, interleukin-1, phorbol ester, and okadaic acid Inhibit EGF (endothelial growth factor) in HL-60

Shearinine A, D, Isolated from and E Penicillium janthinellum Solamargine Steroidal alkaloid Reported to upregulate TNFR-1 (tumor necrotic factor receptor-1) to induce apoptosis Spongiacidin C From the marine Targeting ubiquitin-proteasome system sponge Stylissa massa Staurosporine Produced by Is reported to be a broad-spectrum Streptomyces protein kinase inhibitor bacteria Tetrandrine Isolated from the Activation of glycogen synthase kinase root of Stephania 3β (GSK-3β), generation of ROS, tetrandra activation of p38 mitogen-activated protein kinase (p38 MAPK), and inhibition of Wnt/beta-catenin signaling might contribute to the anticancer effects. It holds a great promise as an MDR (multidrug resistance modulator) for the treatment of P-gp-mediated MDR cancers. Trabectedin Derived from Causes DNA alkylation, RNA tunicates polymerase II breakdown, upregulation of CCL2 (a potent chemokine for monocytes), and downregulation of VEGF (vascular endothelial growth factor) and IL-6 (interleukin 6) Variolin B Marine sponges, Targeting markers related to cancer drug a pyridopyresistance rrolopyrimidine extracted from Kirkpatrickia variolosa Yohimbinic Rauwolfia Topoisomerase inhibition serpentine

References Maiti and Suresh, 2010; Brahmachari 2015; Bhadra and Suresh Kumar 2011a; Bhadra and Kumar 2011b; Gong et al. 2018; Akhtar et al. 2019; Lu et al. 2012

Tohme et al. 2011 Dey et al. 2019

Song et al. 2018

Tamaoki et al. 1986 Lu et al. 2012; Kumar and Jaitak 2019

Tohme et al. 2011

Tohme et al. 2011

Fan et al. 2019

570

K. Bhadra

Fig. 2 Chemical structures of alkaloids extracted from plants

categorized in terms of intercalation, groove binding, and outside binding. RNAs, on the other hand, are highly versatile molecules that can fold into secondary structures and conformations like simple single- and double-stranded structures with A-form to complex structural motifs like hairpin loop or cloverlip structure that could be potential drug-binding pockets. Isoquinoline alkaloids, berberine, is broadly distributed in various botanical families like Berberidaceae, Papaveraceae, Fumariaceae, Menispermaceae, and many others. Several reports on the anticancer property of berberine against different cell

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

571

Fig. 3 Chemical structures of alkaloids extracted from marine invertebrates

lines are available (Maiti and Suresh 2010; Brahmachari 2015). The dose-dependent effect of berberine on cell cycle inhibition and apoptosis in human leukemia HL-60, Balb/c 3 T3 cells, and human osteosarcoma cells by inducing DNA damage and apoptosis of p53-dependent prostate cancer cell death have been reported. Since 1950s, a large number of research group has initiated the berberine-DNA interaction (Bhadra and Suresh Kumar 2011a). Subsequently, Maiti and coworkers, Bhadra et al., studied the interaction of berberine with different natural and synthetic DNAs using multi-spectoscopic and calorimetric techniques and proposed that the alkaloid prefers AT sequences by partial intercalative mode of binding. A strong interaction of berberine to DNA oligonucleotides was also reported by NMR and modeling studies of Mazzini et al. to show groove binding mode. Bhadra et al. explored the binding characteristic of berberine-DNA complexation using UV spectroscopy and characterized that at low-bound alkaloid concentrations, the molecules bind cooperatively with nucleic acid having almost equal percentage of AT and GC

572

K. Bhadra

Fig. 4 Chemical structures of alkaloids extracted from microorganisms

sequences while the binding was noncooperative to DNAs with inherently high AT or GC sequences. With polynucleotides, on the other hand, the binding was observed to be cooperative with poly(purine)poly(pyrimidine) sequences, while with alternating purine-pyrimidine sequences, a noncooperative binding phenomenon was observed (Bhadra and Suresh Kumar 2011a). Complexation of berberine was shown to perturb the circular dichroic spectra of DNA and polynucleotides, and the bound alkaloid molecules acquired induced optical activity to varying extents depending on the base pair composition and sequence. Various binding parameters of berberine DNA complexation obtained from fluorescence spectroscopy, thermal melting, and viscometric analysis were also presented. Energetics of interaction of berberine to various DNAs using isothermal titration calorimetry (ITC) and differential scanning calorimetry (DSC) was studied. Interaction of berberine with poly(A) has been investigated by Maiti and colleagues and Yadav et al. Surprisingly, a stronger affinity of berberine to single-stranded poly(A) over B-DNA and tRNA structures was observed. Later, Giri and Kumar from a careful analysis of the berberine-poly(A) complexation confirmed the noncooperative binding model without any self-structure formation. However, berberine has been found to show a weak interaction with yeast t-RNA molecules by Islam and Kumar from a variety of photophysical techniques. The binding affinity, conformational aspects, and energetics of the interaction of berberine with four single-stranded RNA polymers, viz., poly(G), poly(I), poly(C), and poly(U), were studied. Berberine binds strongly with poly(G) and poly(I) with an order of 105 M−1 , while its binding was practically nil with poly(C) and poly(U). Further, detailed spectroscopic and calorimetric report regarding binding of berberine with double-stranded RNA polynucleotides, viz., poly(A).poly(U), poly(I).poly(C), and poly(C).poly(G), was performed by Islam et al. Among the three double-stranded RNA polynucleotides, the AU sequences showed highest binding with berberine. Thermodynamics of the interaction was

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

573

favored by both positive entropy and negative enthalpy changes. Negative heat capacity, Cpo , changes in all the systems are related to the hydrophobic forces in the complexation. Sinha et al. have reported a weak interaction of the alkaloid with the double-stranded RNA poly(C).poly(G). Berberine was shown to interact strongly (K ∼ 105 M−1 ) with the poly(U).poly(A)* poly(U) RNA triplex using various biophysical techniques (Bhadra and Suresh Kumar 2011a). Unlike berberine, palmatine attracted the attention of researchers only recently probably due to its low level of occurrence in plants. Studies revealed that palmatine like berberine easily binds to DNA and induces DNA photooxidation via singlet oxygen (1 O2 ) generation. Palmatine has potent antitumor activity against HL60 leukemic cells. The first report on the DNA interaction of palmatine was that of Bailly and coworkers who identified DNA as a potential bioreceptor for this alkaloid. This study based on absorption and fluorescence on natural DNA suggested both intercalation and external stacking of the alkaloid on complexation. Palmatine via singlet oxygen formation causes photooxidation of DNA. It was Bhadra et al. (2007) who gave the first detailed investigation on the interaction of palmatine with natural DNAs and synthetic polynucleotides. The multifaceted study employed competition dialysis, spectrophotometry, spectrofluorimetry, thermal melting, circular dichroism, viscometry, and isothermal titration calorimetry techniques (Bhadra and Suresh Kumar 2011a). The study suggested the base and sequence selectivity of the alkaloid to be toward AT sequences. The DNA binding affinity of palmatine was revealed to be higher than that of berberine. Subsequently, the thermodynamics of palmatine-DNA complexation was also investigated in details. Bhadra et al. supported the intercalative mode of binding by hydrodynamic studies where change in specific viscosity was higher for AT polymers. In addition to this, palmatine showed an unwinding angle of 15◦ with covalently closed superhelical Col E1 DNA. Dong and coworkers have also investigated the interaction between calf thymus DNA and palmatine and proposed a mixed mode, viz., intercalative and groove binding for palmatine-DNA complexation. Like berberine, palmatine also showed cooperative and noncooperative mode of binding. Energetics of the complexation suggested that the complex formation in most cases was exothermic and favored by both entropy and enthalpy changes, while in case of the AT polynucleotides, the same was predominantly entropy driven (Bhadra and Suresh Kumar 2011a). Further, study of palmatine and berberine with some oligonucleotides by electrospray ionization mass spectroscopic technique indicated that palmatine exhibited a stronger binding compared to berberine, and this was in concordance with Bhadra et al. Palmatine binds specifically and strongly to poly(A) in a noncooperative manner. A binding affinity of the order of 105 M−1 was observed by spectroscopic and ITC techniques. The interaction was found to be exothermic and entropy driven with a moderate free energy change. No self-structure formation was reported in ss poly(A). Among other singlestranded RNAs, palmatine showed strong binding to poly(G) and poly(I), while among double-stranded RNAs, highest binding was reported with poly(A).poly(U) that was of the order of 106 M−1 , exothermic and entropy driven. The binding of the alkaloid to double-stranded poly(A) was very weak. Detailed studies on

574

K. Bhadra

palmatine-t-RNA complexation was performed by Islam et al. Cooperative binding, thermal stabilization, and perturbation of the RNA conformation and induction of optical activity in bound palmatine molecules were reported. Palmatine has also been reported to interact with triplex RNA, poly(U).poly(A)* poly(U), with a binding order of 105 M−1 using various spectroscopic and calorimetric techniques. Benzophenanthridine isoquinoline alkaloids, sanguinarine (Maiti and Suresh 2010; Bhadra and Suresh Kumar 2011a), a natural alkaloid isolated from Sanguinaria canadensis L. and Chelidonium majus L., belonging to family Papaveraceae, generate wide varieties of biochemical and pharmacological effects including its use in cancer prevention. The alkaloid is known to induce apoptosis in various cell lines. Several mechanisms have been proposed to explain its antiproliferative activities. Results clearly correspond the cytotoxicity of sanguinarine to DNAintercalating ability to induce DNA strand breaks and other DNA-damaging effects. It was Maiti and colleagues who first established the DNA-binding ability of this alkaloid by intercalative mode by spectroscopic and viscometric experiments. The binding affinity of sanguinarine to DNA was reported to be of 105 M−1 (Maiti and Suresh 2010). Out of the two forms of sanguinarine, the charged iminium form is the DNA binding moiety. The alkanolamine form does not bind to DNA, but in presence of large amount of DNA, significant population of iminium form is generated from the alkanolamine form. Results from different biophysical techniques including NMR also suggested a strong GC base pair preference of this alkaloid. Sanguinarine converts the left-handed Z-form DNA cooperatively to the bound right-handed B-form. Energetics of the iminium form was evaluated under various salt and pH conditions by isothermal titration calorimetric technique with natural and synthetic DNAs. Unlike its DNA interaction, little information is available regarding its binding with RNA. Sen and Maiti studied the interaction of sanguinarine with double-stranded RNA structures of AU and IC sequences using various biophysical tools. The interaction of sanguinarine with single- and doublestranded poly(A) came from the contribution of Giri and Kumar. Cooperative CD optical and DSC melting profiles of the complex were provided as clear evidence for self-structured induction of poly(A) by the alkaloid. Triplex RNA binding of sanguinarine was studied by Das et al. who suggested intercalation of the alkaloid to poly(U).poly(A)* poly(U). Another class of natural alkaloids, beta-carbolines, are equally wide group of natural and synthetic indole alkaloids that hold a common tricyclic pyrido indole ring structure, investigated by Shi et al. (2001), Farouk et al. (2008); and many more, possessing a different degree of aromaticity. Beta-carboline alkaloids were originally isolated from the Banisteriopsis caapi (family: Malpighiaceae) and Peganum harmala (family: Zygophyllaceae) that are being utilized traditionally as an herbal drug in the North Africa and the Middle East. Moreover, numerous beta-carboline alkaloids have been isolated from the marine invertebrates including tunicates, soft corals, hydroid, and different sponges. The beta-carboline alkaloids can be classified on the basis of the saturation of their nitrogen (N)-carrying, sixmembered ring. The partially or completely saturated ones are known as dihydrobeta-carbolines like harmalan, harmalol, harmaline, and tetrahydro-beta-carbolines

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

575

or THBCs, respectively, whereas the unsaturated ones are named as fully aromatic beta-carbolines like harmine, harmane, norharman, harman, and harmol. Betacarboline alkaloids are well known for their several biological activities including anticancer and antitumor activities (Sarkar and Bhadra 2018; Bhattacharjee et al. 2018; Sarkar et al. 2020). Interaction of beta-carboline alkaloids with DNA (Sarkar and Bhadra 2018; Bhattacharjee et al. 2018; Sarkar et al. 2020) has been explored by many researchers. Sarkar and Bhadra studied the two structural forms of harmalol and their interaction with CT DNA, where the protonated form showed strong, intercalative mode of binding. They further investigated the sequence-specific binding of harmalol highlighting hetero GC followed by hetero AT-specific binding. It showed both cooperative and noncooperative binding with DNA sequences, and ITC profiles showed both enthalpy- and entropy-driven binding with DNA sequences. The alkaloid was further reported to show maximum cytotoxicity on HepG2 with IC50 of 14 μM (Sarkar and Bhadra 2018; Bhattacharjee et al. 2018; Sarkar et al. 2020). G-quadruplexes are an unusual conformation of DNA with runs of guanines and that are purine rich having four-stranded structures. G-quadruplex structures can inhibit the activity of the enzyme telomerase (vide infra), which adds telomeric repeats to the ends of chromosomes and maintains the proliferation of cancer cells. Thus, proliferated tumor cells can be made inactive by telomerase inhibition. Small molecules that can stabilize the quadruplex structure by binding may act as potential anticancer agents (Bhadra and Kumar 2011b). The idea of investigating natural alkaloids as ligands, interacting and stabilizing quadruplex structures in telomeres, is a current approach for therapeutic intervention. Bhadra and Kumar presented a complete structural and thermodynamic profiles of the binding of isoquinoline alkaloids like berberine, palmatine, coralyne (a synthetic isoquinoline alkaloid), and sanguinarine with G-quadruplex. The results highlight the importance of specific structural elements, particularly the planarity, in the small molecules in stabilizing quadruplex DNA structure for developing better quadruplex targeted therapeutic agents (Bhadra and Kumar 2011b).

Targeting Enzymes, Viz., Topoisomerases I and II, Telomerase, DNA Polymerase, and RNA Polymerase Associated with Nucleic Acids DNA topoisomerases are the key enzymes for DNA replication and repair and have been regarded as target enzymes for many anticancer drugs in clinical use. Ascididemin, granulatimide, lamellarin D, lissoclinidine B, and polycarpines are some of the marine natural alkaloids reported to inhibit topoisomerases I and II (Fan et al. 2019). Topotecan and irinotecan are synthetic analogues of camptothecin, a natural alkaloid, and are FDA-approved anticancer drugs against ovary, colon, and lung as topoisomerase inhibitors. Furthermore, indoquinoline alkaloid like cryptolepine isolated from Cryptolepis sanguinolenta; quinoline alkaloid lunacridine from Lunasia amara; beta-carbolines alkaloids like harmaline and harmine; aporphine alkaloid dicentrinone from Ocotea leucoxylon; indole alkaloid yohimbinic isolated from Rauwolfia serpentine, and many more (Fan et al. 2019) are some of the well-accepted topoisomerase inhibitors. Eribulin mesylate, an

576

K. Bhadra

analogue of the macrocyclic polyether halichondrin B, a marine sponge-derived alkaloid, has reached phases I and II cancer clinical trials for breast cancer through targeting topoisomerases (Tohme et al. 2011). Another natural compound, neoamphimedine, isolated from Xestospongia sp., also has a unique mechanism of action as a topoisomerase inhibitor both in vitro and in vivo. Apart from these, targeting telomere and inhibition of telomerase are an equally viable, effective, and alternative route to achieve effective arrest of cell proliferation, hence enabling the researchers to develop potent ligand derivatives to initiate better diagnostic studies for biomedical applications. Telomerase is a protein involved in cancer where it is present in 85 to 90% of all cases. Taking advantage of this fact, inhibition of telomerase activity by stabilization of the telomeric DNA quadruplex structure as a new treatment in human cancer therapy is very significant. Aporphine alkaloid boldine from Peumus boldus and Lindera aggregata and isoquinoline alkaloid berberine from Berberis vulgaris and Coptis chinensis are some of the natural products as telomerase inhibitors (Ganesan and Xu 2018). It was further found that chelerythrine, a benzene alkaloid isolated from the herb Chelidonium majus L., also inhibits telomerase activity and induces cancer cell death by binding with human telomeric G-quadruplex structure (Liu et al. 2017). Though direct target of DNA polymerase by anticancer natural alkaloids has not yet been reported, major strategy in chemotherapy is to use DNA-damaging agents (vide supra) to inhibit processive DNA polymerases, though trabectedin, a natural alkaloid, reported to cause RNA polymerase II breakdown.

Anticancer Natural Alkaloids Targeting Cellular Proteins, Enzymes, and Growth Factors Targeting Kinases Related to Tumor Proliferation, Progression, and Metastasis Targeting specific oncogene kinases of signaling pathways are expected to be the promising strategy related to the formation and progression of tumor. Protein kinase C (PKC) are family of kinase enzymes that by the phosphorylation of hydroxyl groups of threonine and serine amino acid residues regulate the function of other proteins. Investigation with natural and synthetic compounds that can inactivate PKC is in high need for therapeutic applications. For example, chelerythrine has been evidenced to possess effective antitumor effect on different cancers, like breast, colon, and prostate (Chmura et al. 2000). The mechanism of action involves several pathways in a concentration and time-dependent manners to induce apoptosis. Chelerythrine activates JNK/p38-MAPK (c-Jun N-terminal kinase-mitogen-activated protein kinase) pathways in HeLa cells (Yu et al. 2000), MEK signaling pathway, and mitogen-activated protein kinases and upregulates downstream kinases (p90RSK) in human osteosarcoma cells (Yang et al. 2008). It also causes G1 phase cell cycle arrest and bimodal cell death in human leukemia HL-60 cells (Vrba et al. 2008). Moreover, chelerythrine acts as a specific inhibitor of protein kinase C (PKC) and causes apoptotic cell death in tumor (Liu et al.

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

577

2017). Another natural alkaloid, staurosporine, produced by Streptomyces bacteria, is reported to be a wide-spectrum protein kinase inhibitor (Tamaoki et al. 1986). The cyclin-dependent kinases (CDKs) are family of serine-threonine protein kinases whose activities are required for the cell cycle. These are misregulated in almost 60–70% of human cancers. Intensive screening for the identification of natural alkaloids as inhibitors of CDKs, which display high selectivity and proficiency, is in high demand. These inhibitors are antiproliferative agents that arrest cells in G1 and/or G2/M phases of the cell cycle such as berberine (Lu et al. 2012); evodiamine, a quinolone alkaloid and a bioactive compound isolated from the Chinese herb Evodia rutaecarpa (Lu et al. 2012); piperine (Lu et al. 2012); harmalol; harmaline (Ahmad et al. 2020), and many more. Hymenialdisine and debromohymenialdisine are the alkaloids isolated from the marine sponge Stylotella aurantium. They have been reported to show competitive inhibition at the ATP-binding site and thereby inhibit cyclin-dependent kinases (Song et al. 2018). Another marine natural product, fascaplysin, isolated from marine sponge, is also a specific CDK4 inhibitor by binding to the ATP site of the kinase and has a IC50 value of 0.35 μM. Fascaplysin inhibit the proliferation of endothelial cells and prevent angiogenesis (Song et al. 2018). Another group of marine indole alkaloids, meridianins A–G, connected to an aminopyrimidine ring. They have been isolated from marine tunicate Aplidium meridianum, consisting and found to selectively inhibit CDK1, CDK5, and other protein kinases involved in cancer and other disease. Like the abovementioned alkaloids, meridianins A–G also bind to the ATP pocket of protein kinases and act as ATP competitive inhibitors (Song et al. 2018). Sanguinarine is also reported to induce apoptosis of human lens epithelial cells by increasing ROS (reactive oxygen species) via the MAPK (mitogen-activated protein kinase) signaling pathway (Zhang and Huang 2019). Apratoxin A and B from cyanobacteria showed promising in vitro anticancer activities against many human tumor cell lines including LoVo colon cancer cells and nasopharynx human carcinoma KB cells. Apratoxin causes inhibition of cell cycle at G1 phase leading to apoptosis. It further induces apoptosis by blocking the FGFR (fibroblast growth factor receptor) pathway via the inhibition of phosphorylation of the signal transducer and activator of transcription 3 (STAT3) in a dose- and time-dependent manner and induces apoptosis, thus inhibiting the upregulation of antiapoptotic genes like BclxL and the activation of the cyclin-dependent kinase CDK2. These effects may explain antiangiogenic effects of apratoxin due to its ability to antagonize the FGFR pathway (Tohme et al. 2011). The benzonaphthyridine alkaloid aaptamine extracted from the marine sponge Aaptos suberitoides upregulates p21 transcription in p53mutated MG63 human osteosarcoma cell, inducing G2/M cell cycle arrest, making p21 an interesting target for chemotherapy (Tohme et al. 2011). The insulin-like growth factor-1 receptor (IGF-1R) is also a potential therapeutic target for cancer. They are involved in tumor cell proliferation, survival, and invasion. Inhibitors of IGF-1R are useful for the treatment of solid tumors, including non-small cell lung cancer, small cell lung cancer, and ovarian carcinoma. IGF-1R signaling is transduced through two main pathways: (a) the RAS/RAF/MAP kinase pathway and (b) the phosphoinositide-3 kinase (PI3K)/Akt pathway. Evodiamine

578

K. Bhadra

possesses anticancer activities by causing the cell cycle arrest or apoptosis and suppressing the angiogenesis, invasion, and metastasis in a variety of cancer cell lines in micro- and nanomolar level and in Balb-c/nude mice model (Lu et al. 2012). Evodiamine has effect on tubulin polymerization (Lu et al. 2012) and has been reported to cause cell death by the phosphatidylinositol 3-kinase/Akt/caspase (Lu et al. 2012). Isoquinoline alkaloid, berberine, causes dose-dependent, inhibiting proliferation and induces apoptosis of human osteosarcoma cell line, U20S cell. Berberine inhibits PI3K/Akt activation which, in turn, results in upregulation of Bax and PARP expression and downregulation of Bcl-2 and caspase 3 initiation (Chen 2016). Compared to other anticancer drugs, the cytotoxic effectiveness of berberine is much lower, with an IC50 of 10 μM to 100 μM depending on the cell type and duration of the treatment (Lu et al. 2012). Indole alkaloids, shearinine A, D, and E, sourced from Penicillium janthinellum, are thought to inhibit EGF (endothelial growth factor) in HL-60 (Tohme et al. 2011). Glycogen synthase kinase-3 beta (GSK-3β), a serine/threonine protein kinase that has been broadly involved in critical cell biology processes, is a promising multipurpose kinase for cancer therapeutic target. A complex alkaloid, manzamine, isolated from sponge Acanthostrongylophora, was shown to be a noncompetitivespecific inhibitor of ATP with binding to GSK-3β at IC50 value of 10.2 μM (Song et al. 2018). Based on molecular modeling, it was suggested that phenylmethylene hydantoin (PMH-1) isolated from the Red Sea sponge Hemimycale arabica could be successfully docked into the binding pocket of GSK-3β and control the invasive breast malignant growth by suppressing Ki-67, CD31, p-Brk, and p-FAK expression in tumor samples. Another natural bisbenzylisoquinoline plant alkaloid, tetrandrine, extracted from the root of Stephania tetrandra, was reported to have wide array of medicinal importance including its anticancer effect in micromolar amount. Activation of glycogen synthase kinase 3β (GSK-3β), generation of ROS, activation of p38 mitogen-activated protein kinase (p38 MAPK), and inhibition of Wnt/beta-catenin signaling might bestowed to the anticancer effects of tetrandrine. Tetrandrine also actively upregulates p53, p21, p27, and Fas; downregulates Akt phosphorylation, CDKs, and cyclins; modulates the members of the Bcl-2 family; and initiates caspases (Lu et al. 2012).

Targeting Tubulin Polymerization Hemiasterlin, originally derived from the marine sponge Hemiasterella minor, exhibits antiproliferative effects against human glioblastoma U373, mammary carcinoma MCF-7, murine leukemia p388 cells, and human ovarian carcinoma HEY by binding to tubulin at nanomolar concentrations and inhibits their polymerization, causing mitotic catastrophe and ultimately apoptosis (Tohme et al. 2011). But the in vivo effect of hemiasterlins shows high level of toxicity; hence, synthesis of new analogues with enhanced drug efficacy is in demand. Sanguinarine is also reported to bind with tubulin and perturb microtubular assembly dynamics. Halichondrin B, a marine sponge alkaloid, is also reported to show tubulin polymerization action (Tohme et al. 2011).

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

579

Targeting Apoptosis-Related Biomarkers Apoptosis is a highly synchronized, complicated set of multibranched, multistep pathways with many checks and balances, induced by various stimuli and defined by a series of distinct biochemical and morphological changes. Inhibition of cell proliferation and increasing apoptosis in tumors are the effective ways to eliminate tumor growth and prevent cancers. Drug with a good anticancer potential should have the ability to induce apoptosis in cancer cells and is considered as one of the essential features of the anticancer drug. Mitochondrion-controlled intrinsic pathways or receptor-controlled extrinsic pathways can trigger apoptosis (Dey et al. 2019). Various natural alkaloids activate apoptosis through different pathways in specific cancer cell line. Beta-carboline alkaloids have been reported to induce fragmentation of nucleic acid, ultra morphological changes of cellular organelles, decreased mitochondrial membrane potential, p53 and caspase-3 upregulation, generation of ROS (reactive oxygen species), and a significant increase in G2 /M population that has made hepatic carcinoma HepG2 cells more prone to apoptosis than uterine carcinoma HeLa cells. Steroidal alkaloid, cephalostatin 1, uses Smac/DIABLO for inducing apoptosis (Dey et al. 2019). Smac/DIABLO is a mitochondrial protein that neutralizes one or more members of apoptosis inhibitory proteins. Smac enters the cytosol from mitochondria during apoptosis triggered by UV- or γ-irradiation (Adraine et al. 2001). Another steroidal alkaloid, briofilin, suppresses Bcl-2 and caspase-3 initiation and upregulates Bax protein expression and segmentation of poly (ADP-ribose) polymerase-1 (PARP-1) in HeLa cells (Dey et al. 2019). Solamargine has been reported to upregulate TNFR1 (tumor necrotic factor receptor-1) to induce apoptosis (Dey et al. 2019). The tumor necrosis factor (TNF) represents a multifunctional group of pro-inflammatory cytokines which trigger signaling pathways for cell survival, apoptosis, inflammatory reaction, and cellular differentiation. Isoquinoline alkaloid, sanguinarine, induces intrinsic apoptosis in the presence of Bax. The weakened interaction of MELK (maternal embryonic leucine zipper kinase) and STRAP (serine-threonine kinase receptor-associated protein) is necessary for the transactivation of Bax from cytosol to mitochondria. The accumulation of Bax in the mitochondria induces MOMP (mitochrondrial outer membrane permeabilization), which causes the release of cytochrome c into the cytosol from mitochondria (Gong et al. 2018). Sanguinarine treatment of multiple myeloma cells leads to downregulation of the antiapoptotic proteins including Bclxl, XIAP, Bcl-2, and cyclin D. In addition, it also upregulates proapoptotic protein, Bax. It basically induces apoptosis in the multiple myeloma cells via inhibition of JAK2/STATs signaling (Janus kinases, JAK, signal transducer and activator of transcription proteins, STAT) (Akhtar et al. 2019). It remarkably sensitizes breast cancer cells to tumor necrosis factor (TNF)related apoptosis inducing ligand-mediated apoptosis (Lu et al. 2012). Piperine and piperlongumine extracted from Piper nigrum L. and Piper longum L., respectively, in combination, were reported to suppress STAT3 phosphorylation and control downstream molecules to cause apoptosis in breast cancer cells (Chen et al. 2020). Quinolone alkaloid, evodiamine, has been found to rapidly enhance ROS followed by mitochondrial depolarization that induces apoptosis (Lu et al. 2012). Evodiamine

580

K. Bhadra

induces both caspase-dependent and caspase-independent apoptosis, upregulates Bax expression, and downregulates Bcl-2 expression in some cancer cells. Another natural alkaloid, matrine, a major active compound found in Sophora flavescens Ait and S. japonica, has been evaluated in in vitro 4 T1, H22, MNNG/HOS, and BxPC3 cells and in in vivo BALB/c mice model. It affects proteins such as E2F-1, Bax, Bcl-2, Fas, and Fas-L involved in cell proliferation or apoptosis (Lu et al. 2012). Matrine at 50–100 mg/kg inhibits MNNG/HOS xenograft growth and brings down the pancreatic tumor volumes. In addition to this, it also inhibits the proliferation of various types of cancer cells such as hepatoma G2 cells, K567, and SGC-7901 cells through apoptosis (Lu et al. 2012). Naamidine, a sponge-derived imidazole alkaloid identified in Leucetta chagosensis, is another example of an apoptosistargeting compound (Tohme et al. 2011). Naamidine stimulates caspase-dependent apoptosis in tumors cells through both extrinsic and intrinsic cell death pathways by interruption of mitochondrial potential. It causes cell cycle arrest at G1 phase through stabilization of the p21 cyclin-dependent kinase inhibitor (CDK2) and stimulation of the initiating caspases 8 and 9 and PARP cleavage. Furthermore, naamidine cleaved the effector caspase 3 through p53-independent mechanism both in in vitro and in vivo model (Lu et al. 2012).

Targeting Transcription Factors Controlling Expression of Cancer Gene Transcription factor is a protein that binds to selective DNA sequence and regulates gene expression by promoting or suppressing transcription which plays important role in the development and metastasis of cancer. Hypoxia-inducible factor 1 (HIF1) is one such tumor prospective factor related to tumor cell proliferation, apoptosis, metabolism, and angiogenesis and is one of the most potent targets for treating various cancers. Alkaloids with pyrroloiminoquinone ring structure from the marine sponge Latrunculia sp., which were identified as novel HIF-1α/p300 inhibitors, interrupted the protein-protein interaction between HIF-1α and p300 and actively inhibit the growth of HCT 116 and prostatic carcinoma cell lines in vitro models. It has been reported that the alkaloid berberine induces MDM2 self-ubiquitination and degradation by inhibiting MDM2-DAXX-HAUSP interactions (Zhang et al. 2010). MDM2 (mouse double minute 2) homolog is encoded by the MDM2 gene in human. It is moreover a negative regulator of the p53 tumor inhibitor. Disruption of the regulatory functions by MDM2 is a viable strategy to reactivate p53 (Song et al. 2018). NF-κB, a dimer protein, is another inducible transcription factor; plays an important role in the regulation of immunological, inflammatory, and carcinogenic responses; and has become a major molecular target in drug designing. NF-κB belongs to the Rel family, which includes RelA (p65), RelB, c-Rel, p50 (NFκB1), and p52. Evodiamine has been reported to activate Fas ligand (Fas-L)/NF-κB signaling pathways and cause cancer cell death (Lu et al. 2012). Sanguinarine is also a potent suppressor of NF-κB activation induced by phorbol ester, okadaic acid, TNF, and interleukin-1 (Lu et al. 2012). Moreover, it efficiently inhibits the signal transducer and activator of transcription 3 activation (STAT-3); modulates the members of the Bcl-2 family; activates caspases and upregulates death receptor 5

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

581

(DR-5); downregulates CDKs, cyclins, MMP-2, and MMP-9; and upregulates p21, p27, and phosphorylation of p53.

Targeting Histone Deacetylases to Cause Genomic Instability Through the repressive influence on histone deacetylation transcription, histone deacetylases (HDACs), a group of epigenetic enzymes, control gene expression. They remove acetyl groups from an N-acetyl lysine amino acid on histone and allow the histones to cover the nucleic acid more tightly. The dysregulation of DNA methylation and acetylation of the lysine residues on histone tails result in genomic instability of tumor cells (Song et al. 2018). Hence, HDACs are probable therapeutic targets for cancer treatment. Plant alkaloid berberine has been reported to induce apoptosis and growth arrest in the A549 cell line by causing modulation of histone deacetylase (Arunachalam et al. 2016). Another natural plant alkaloid, lycorine, which is a toxic crystalline alkaloid belonging to family Amaryllidaceae, causes HDAC inhibition in K562 cells and causes cell cycle arrest in the G0/G1 phase (Li et al. 2012). Macrocyclic depsipeptide alkaloid, largazole from cyanobacteria, also suppresses histone deacetylases (HDAC) and exhibits anticancer activity (Tohme et al. 2011). Targeting Ubiquitin-Proteasome System for Cancer Therapy The ubiquitin-proteasome system (UPS) is a highly regulated mechanism of intracellular protein degradation and turnover. Through the combined actions of a series of enzymes, ubiquitin-activating enzyme (E1), ubiquitin-conjugating enzyme (E2), and ubiquitin-protein ligase (E3), proteins are marked for proteasomal degradation (Song et al. 2018). Ubiquitin-proteasome system is involved in the regulation of important processes of carcinogenesis; hence, targeting the ubiquitin-proteasome system has been a therapeutic strategy in clinical treatment of cancer. Spongiacidin C, a pyrrole alkaloid, derived from the marine sponge Stylissa massa has been identified as the first natural USP7 inhibitor with a IC50 of 3.8 μM (Song et al. 2018). Ubiquitin-specific peptidase 7 (USP7) is a human deubiquitylating enzymes (DUBs), which affects the stability and degradation of cellular proteins. USP7 can deubiquitylate Hdm2 and consequently degrade p53. Hence, inhibiting USP7 stabilizes p53 in cells through degradation of Hdm2 and results in the restraint of cancer, and thus, USP7 is an emerging oncology target. The ubiquitylation of protein is reversed by deubiquitylating enzymes (DUBs) and leads to deconjugation of the ubiquitin chain. Targeting Heat Shock Protein (Hsp90) Related to Oncoprotein Maturity Heat shock protein 90 (Hsp90) is a chaperone protein that helps other proteins to fold properly, stabilizes heat stressed proteins, and assists in degradation of protein. It acts as an evolutionarily conserved molecule that plays essential role in apoptosis, cell survival, proliferation, and cellular homeostasis. It also stabilizes

582

K. Bhadra

a number of proteins required for tumor growth. Evidence shows that Hsp90 is often uncontrolled in many solid tumors, like lung cancer, colorectal cancer, breast cancer, and hematological malignancy. Hence, Hsp90 has been recognized as an important target in cancer therapy. Based on computer modelling and docking studies, berberine has been assigned as a natural inhibitor of Hsp90. The ATP sites in Hsp90 such as Gly121, Asn92, Asp79, Asn37, and Leuc34 were selected as targeted amino acids and reported to be inhibited by the alkaloid, thereby suppressing the growth of tumor cells which might contribute to the chemopreventive potential of the alkaloid (Siva and Babu 2014).

Targeting the Markers Related to the Cancer Drug Resistance Cancers often develop resistance to traditional therapies. The increasing number of these drug-resistant cancers demands further research and development. It is influenced by drug target modification, drug efflux, drug inactivation, cell death inhibition, DNA damage repair, epigenetic effects, inherent cell heterogeneity, or any coalition of these mechanisms. P-glycoprotein (P-gp) is one such known marker as multidrug resistance 1 (MDR1) or ATP-binding cassette subfamily B member 1 (ABCB1) and belongs to ABC transporter family. Therefore, exploration of alkaloids as anticancer compounds, which could suppress these ABC transporter proteins, is an effective approach to reverse resistance and further ameliorate the therapeutic efficacy. In mice with MDR, MCF-7/adr, or KBv200 cell xenografts, coadministration of tetrandrine has been found to increase the anticancer activity of doxorubicin, paclitaxel, docetaxel, and vincristine without a significant increase in toxicity. Hence, tetrandrine holds a great promise as a MDR (multidrug resistance) modulator for the treatment of P-gp-mediated MDR cancers (Kumar and Jaitak 2019). Variolin B, a pyridopyrrolopyrimidine extracted from Kirkpatrickia variolosa, is cytotoxic both in micro- and nanomolar concentration against several human cancer cells, including human leukemia cells and ovarian, intestinal, and colon carcinomas which are usually resistant to conventional chemotherapeutic agents due to high levels of Pgp expression (Tohme et al. 2011). Furthermore, harmine, piperine, and piperidine are some of the anticancer alkaloids reported to show MDR reversal effects (Tinoush et al. 2020).

Carbohydrate Target by Small Molecules as Anticancer Alkaloids Carbohydrates are known to play significant roles in cancer metastasis in addition to a large number of other biological and pathological processes. Accumulated glucose and increased metabolism encourage several characteristic features of cancer. In cancer cells, regardless of the availability of the oxygen, a large amount of glucose can be converted to lactic acid, and this phenomenon is known as Warburg’s effect (Hossain and Andreana 2019). Hyperactivated glucose uptake and aberrant glycolytic metabolism are considered as a hallmark of cancer cells. On the surface of tumor cells, overexpressed extracellular tumor-specific carbohydrate antigens

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

583

(TACAs) are considered biomarkers for cancer detection and have been prioritized for the development of novel anticancer drugs. However, detail regulation of the glucose metabolism of cancer cells is still poorly understood. Overexpression of glucose transporters (GLUTs) in cancer cells ameliorates higher glucose consumption than that of normal cells, and this fact has opened new windows for targeted chemotherapy in many cases, and accordingly, many natural alkaloids have been selected as anticancer agent (Gao and Chen 2015). Berberine, a natural alkaloid with tumor-selective anticancer effects, as described earlier, has been shown to upgrade glucose uptake in metabolic tissues and cells. Berberine has been reported to control the growth of colon cancer cell lines, HCT116 and KM12C, by inhibiting the glucose uptake and transcription of glucose metabolic genes, GLUT1 (glucose transporter 1), LDHA (lactate dehydrogenase A), and HK2 (hexokinase 2). Study further indicated that the protein expression of HIF1α (hypoxia-inducible factor 1-alpha), a well-known transcription factor critical for cancer cell with dysregulated glucose metabolism, is inhibited by berberine in HCT116 cell line. Furthermore, berberine treats colon cancer cells via suppressing mTOR (mammalian target of rapamycin)-dependent HIF-1α protein synthesis, which provides not only a novel mechanism for the development of berberine as anticancer drug molecule (Mao et al. 2018). Some groups target GLUT-1 and have conjugated sugars at different positions like glycoconjugated prodrug, 2-D-glucose paclitaxel. After cellular uptake, these prodrugs are cleaved by the hydrolytic enzymes to release the active drugs for controlling tumor cells (Hossain and Andreana 2019). Another natural alkaloid, saframycin A, isolated from marine tunicates has been reported to have potent anticancer effect by forming a nuclear ternary complex with GAPDH (glyceraldehyde 3-phosphate dehydrogenase) and DNA (Gao and Chen 2015). GAPDH, as a glycolytic enzyme, catalyzes the conversion of glyceraldehyde 3-phosphate to 1, 3-diphosphoglycerate. Moreover, GAPDH physically associates with APE1 (human apurinic and apyrimidinic endonuclease), an essential enzyme involved in the repair of abasic sites in damaged DNA, as well as in the redox regulation of several transcription factors. This in turn converts the oxidized APE1 to reduced form and thereby reactivates its endonuclease activity to cleave abasic DNA sites (Azam et al. 2008).

Anticancer Natural Alkaloids Targeting Lipid Metabolism Recent studies disclosed that disorders related to lipid metabolism play significant role in carcinogenesis by exploiting various abnormal gene and protein expression and dysregulation of cytokines and signaling pathways. Once the homeostasis between lipid anabolism and catabolism is disturbed, organisms will develop metabolic disorders that leads to diseases like hyperlipidemia, obesity, and cancer. Natural alkaloids such as piperine, betaine, oxymatrine, berberine, nuciferine, and various analogues of nigellamines and dolabellane have been found to be promising candidates for metabolic diseases to improve lipid metabolism-related

584

K. Bhadra

abnormalities. Lipid intake includes high-fat diet intake or endogenous intake of fatty acids which in turn upregulates VEGF (vascular endothelial growth factor) and promotes angiogenesis via recruitment of M2 macrophages, thereby causing cancer cell proliferation. High-fat diet increases levels of IGF (insulinlike growth factor), microRNA-130a, and various pro-inflammatory cytokines, such as interleukins IL-1α, IL-1β, IL-6, TNF-α (tumor necrosis factor-α), EMR1 (macrophage markers such as EGF-like module-containing mucin-like hormone receptor-like1), CR4 (complement receptor 4), and TLR-4 (toll-like receptor), promoting cancer cell proliferation. High-fat diet also accelerates cancer cell proliferation through activation of MCP-1/CCR2 signaling. Endogenous intake includes LPA (lysophosphatidic acid), PUFAs (ω-3 polyunsaturated fatty acids), and fatty acid. Decreased expression of mRNA encoding LPA5, one of G proteincoupled receptors, accelerates proliferation and infiltration of mammalian cancerous cells. Levels of EDPs (epoxydocosapentaenoic acids), a metabolite of ω-3PUFA produced by cytochrome P450 enzymes in tumor cells, are significantly increased by PUFAs, and as a consequence, expressions of pro-oncogenic genes like C-myc, Axin2, and C-jun are reduced in cancer tissues that leads to the inhibition of cancer growth. Intake of fatty acid also increases the expression of B-scavenger receptor, CD36 (cluster of differentiation 36), a member of cell surface fatty acid receptors that affect cancer metastasis. Berberine improves hyperlipidemia and fatty liver by the activation of AMPK (AMP-dependent protein kinase) in liver and muscle tissues. AMPK activation is required for increasing fatty acid oxidation. It also regulates the lipid metabolism through inhibition of fatty acid synthetase and 5-tetradecyloxy-2-furoic acid (TOFA), inducing apoptosis in cancer cell. Berberine-dependent decrease of dyslipidemia in obesity causes alteration in hepatic and muscular gene expression programs that enhance fatty acid oxidation and reduce lipogenesis or hyperlipidemia (Kim et al. 2009). Piperine has been found to suppress lipid droplet accumulation in mouse macrophages by inhibiting cholesteryl ester synthesis (Matsuda et al. 2008). It significantly reduces the concentration of serum lipids in obese rats to near normal levels and elevates high-density lipoprotein (HDL) in the serum. Betaine is another compound found in many foods, including spinach, beets, and several whole grains. Betaine and choline ameliorate lipid homeostasis in obese condition by participation in mitochondrial oxidative demethylation (Sivanesan et al. 2018). Nuciferine, found in Nelumbo nucifera leaves improves dyslipidemia in vivo through PPARα/PPARγ coactivator-1α pathway (Zhang et al. 2018). Oxymatrine found in sophora roots primarily from Sophora japonica (kushen), both in vivo and in vitro, was reported to alleviate hepatic lipid metabolism via regulating the expression profile of microRNA-182 in nonalcoholic fatty liver disease (Zhang et al. 2020). Lipid-lowering drugs and anti-lipid peroxidation treatment may be the starting point for improving or curing lipid metabolism-related diseases. Further clarification with molecular mechanism between lipid metabolism and malignancy is essential for developing novel therapeutic agents and diagnostic biomarkers for human cancers.

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

585

Concluding Remarks Alkaloids have gained ascending popularity as emerging herbal medicines due to their incontestable physiological properties. They possess great potential as effective bioactive compounds for drug discovery with remarkable antitumoral and anticancer properties. The advantages of these natural compounds are their low toxicity, high efficacy, and background knowledge that made us to understand and exploit them as effective therapeutic agents against the disease process. This is possible by understanding their target-specific binding and structure-activity relationships. In this chapter, up-to-date knowledge on the biological evaluation with promising IC50 concentration and target-specific binding aspects of some of the most important anticancer natural alkaloids are presented. These natural drug candidates act through different methods of activity; some cause damage of DNA molecule, regulate the PI3k/Akt signal transduction course, and raise the levels of ROS, while others initiate a caspase or an MAPK pathway reaction in cells. Nonetheless, many act by cell cycle arrest at different checkpoints and downregulating the NF-κB endurance pathway. Even few drugs lead to the stabilization of G-quadruplexes while others go about as inhibitors of the p-glycoprotein ABCB1 and acting as causative agents against drug resistance in cancer cells. These molecular approaches speak to the reason for further developed advanced research on these alkaloid molecules. Significantly, many of these natural alkaloids could be toxic to human beings. Henceforth, the future challenge would be to make use of the chemical diversity of these alkaloids in order to investigate their cytotoxic selectivity to a large number of cancer cells as well as healthy normal cells and to identify possible targets for improved treatment plan in cancer patients. Such fascinating outcomes warrant more in vivo testing of these compounds as this could yield all the more encouraging and promising results. The elucidation of these recognition mechanism of target-specific binding with the biomacromolecule and the availability of a large volume of experimental results of these prodigious molecules may provide useful guidance for drug designing for their development as potential therapeutic agents in the near future.

Acknowledgments KB is indebted to the Council of Scientific and Industrial Research (CSIR) Government of India 37 (1538)/12/EMR-II; U.G.C., Government of India 41-1434/2012(SR) and DST-RFBR, 2017-19, DST/INT/RUS/RFBR/P-254 for the financial support. The author is also grateful to DST-PURSE, DST-FIST SR/FST/LSI-467/2010C) and PRG, University of Kalyani, 2021–22 for the partial financial support.

References Adraine C, Creagh EM, Martin SJ (2001) Apoptosis associated release of Smac/DIOBLO from mitochondria requires active caspases and is blocked by Bcl-2. EMBO J 20:6627–6636

586

K. Bhadra

Ahmad I, Fakhri S, Khan H, Jeandet P, Aschner M, Yue Z-L (2020) Targeting cell cycle by βcarboline alkaloids in vitro: novel therapeutic prospects for the treatment of cancer. Chem Biol Interact 330(2020):109229. https://doi.org/10.1016/j.cbi.2020.109229 Akhtar S, Achkar IW, Siveen KS, Kuttikrishnan S, Prabhu KS, Khan AQ, Ahmed EI, Sahir F, Jerobin J, Raza A, Merhi M, Elsabah HM, Taha R, El OH, Zayed H et al (2019) Sanguinarine induces apoptosis pathway in multiple myeloma cell lines via inhibition of the JaK2/STAT3 signaling. Front Oncol 9(Article 285):1–15. https://doi.org/10.3389/fonc.2019.00285 Anderson HJ, Coleman JE, Andersen RJ, Roberge M (1997) Cytotoxic peptides hemiasterlin, hemiasterlin a and hemiasterlin b induce mitotic arrest and abnormal spindle formation. Cancer Chemother Pharmacol 39:223–226 Arunachalam K, Chidambaram A, Sankar R, Rajasekaran S, Marshal JJ, Muthusamy K, Ravikumar V (2016) Berberine exhibits chromatin remodeling by modulation of histone deacetylase to induce growth arrest and apoptosis in the A549 cell line. J Agric Food Chem. https://doi.org/ 10.1021/acs.jafc.6b04453 Azam S, Jouvet N, Jilani A, Vongsamphanh R, Yang X, Yang S, Ramotar D (2008) Human glyceraldehyde-3-phosphate dehydrogenase plays a direct role in reactivating oxidized forms of the DNA repair enzyme APE1. J Biol Chem 283(45):30632–30641 Bhadra K, Maiti M, Suresh Kumar G (2007) Molecular recognition of DNA by small molecules: AT base pair specific intercalative binding of cytotoxic plant alkaloid palmatine. Biochim Biophys Acta 1770:1071–1080 Bhadra K, Suresh Kumar G (2011a) Therapeutic potential of isoquinoline alkaloids: an update on their nucleic acids binding aspects and implications for drug design. Med Res Rev 31(6):821– 862 Bhadra K, Suresh Kumar G (2011b) Interaction of berberine, palmatine, coralyne and sanguinarine to quadruplex DNA: a comparative spectroscopic and calorimetric study. Biochim Biophys Acta 1810:485–496 Bhattacharjee P, Sarkar S, Ghosh T, Bhadra K (2018) Therapeutic potential of harmaline, a novel alkaloid, against cervical cancer cells in vitro: apoptotic induction and DNA interaction study. J Appl Biol Biotech 6(04):1–8 Brahmachari G (ed) Bioactive natural products chemistry and biology, Wiley-VCH, Weinheim 2015, Isoquinoline alkaloids and their analogs: nucleic acid and protein binding aspects, and therapeutic potential for drug design, pp 241–277, Suresh Kumar G. Chen Z-Z (2016) Berberine induced apoptosis of human osteosarcoma cells by inhibiting phosphoinositide 3 kinase/protein kinase B (PI3K/Akt) signal pathway activation. Iran J Public Health 45(5):578–585 Chen D, Ma Y, Guo Z, Liu L, Yang Y, Wang Y, Pan B, Wu L, Hui Y, Yang W (2020) Two natural alkaloids synergistically induce apoptosis in breast cancer cells by inhibiting STAT3 activation. Molecules 25:216. https://doi.org/10.3390/molecules25010216 Chmura SJ, Dolan ME, Cha A, Mauceri HJ, Kufe DW, Weichselbaum RR (2000) In vitro and in vivo activity of protein kinase C inhibitor Chelerythrine Chloride induces tumor cell toxicity and growth delay. Clin Cancer Res 6:737–742 Cuadrado A, Garcia-Fernandez LF, Gonzalez L, Suarez Y, Losada A, Alcaide V, Martinez T, Fernandez-Sousa JM, Sanchez-Puelles JM, Munoz A (2003) Aplidin induces apoptosis in human cancer cells via glutathione depletion and sustained activation of the epidermal growth factor receptor, src, jnk, and p38 mapk. J Biol Chem 278:241–250 D’Incalci M, Galmarini CM (2010) A review of trabectedin (et-743): a unique mechanism of action. Mol Cancer Ther 9:2157–2163 Dassonneville L, Wattez N, Baldeyrou B, Mahieu C, Lansiaux A, Banaigs B, Bonnard I, Bailly C (2000) Inhibition of topoisomerase ii by the marine alkaloid ascididemin and induction of apoptosis in leukemia cells. Biochem Pharmacol 60:527–537 Dey P, Kundu A, Chakraborty HJ, Kar B, Choi WS, Lee BM, Bhakta T, Atanasov AG, Kim HS (2019) Therapeutic value of steroidal alkaloids in cancer: current trends and future perspectives. Int J Cancer 145:1731–1744

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

587

Fan M, Chen G, Sun B, Wu J, Li N, Sarker SD, Nahar L, Guo M (2019) Screening for natural inhibitors of human topoisomerases from medicinal plants with bio-affinity ultrafiltration and LC–MS. Phytochem Rev. https://doi.org/10.1007/s11101-019-09635-x Fang FG. Synthetic studies on the halichondrin b analog, E7389. Abstract presented at Frontiers in chemistry and medicine I, 11 November 2004 Farouk L, Laroubi A, Aboufatima R, Benharref A, Chait A (2008) Evaluation of the analgesic effect of alkaloid extract of Peganum harmala L.: possible mechanisms involved. J Ethnopharmacol 115(3):449–54. https://doi.org/10.1016/j.jep.2007.10.014 Fedorov SN, Bode AM, Stonik VA, Gorshkova IA, Schmid PC, Radchenko OS, Berdyshev EV, Dong Z (2004) Marine alkaloid polycarpine and its synthetic derivative dimethylpolycarpine induce apoptosis in jb6 cells through p53- and caspase 3-dependent pathways. Pharm Res 21:2307–2319 Ganesan K, Xu B (2018) Telomerase inhibitors from natural products and their anticancer potential. Int J Mol Sci 19:13. https://doi.org/10.3390/ijms19010013 Gao J-L, Chen Y-G (2015) Natural compounds regulate glycolysis in hypoxic tumor microenvironment. Hindawi publishing corporation. BioMed Res Int 2015, Article ID 354143, 8 p. https:/ /doi.org/10.1155/2015/354143 Goldthwait DA (1960) Nucleic acids and Cancer. Am J Med 29(6):P1034–P1059. https://doi.org/ 10.1016/0002-9343(60)90083-8 Gong X, Chen Z, Han Q, Chen C, Jing L, Liu Y, Zhao L, Yao X, Sun X (2018) Sanguinarine triggers intrinsic apoptosis to suppress colorectal cancer growth through disassociation between STRAP and MELK. BMC Cancer 18:578–593 Gross H, Goeger DE, Hills P, Mooberry SL, Ballantine DL, Murray TF, Valeriote FA, Gerwick WH (2006) Lophocladines, bioactive alkaloids from the red alga lophocladia sp. J Nat Prod 69:640–644 Habli Z, Toumieh G, Fatfat M, Rahal ON, Gali-Muhtasib H (2017) Emerging cytotoxic alkaloids in the battle against cancer: overview of molecular mechanisms. Molecules 22:250. https://doi.org/ 10.3390/molecules22020250 Hanahan D, Weinberg RA (2000) The hallmarks of cancer. Cell 100(1):57–70. https://doi.org/ 10.1016/S0092-8674(00)81683-9 Heinrich M, Barnes J, Gibbons S, Williamson EM (2012) Fundamentals of pharmacognosy and phytotherapy. Elsevier Health Sciences, p 130. 978-0702052316 Hossain F, Andreana PR (2019) Developments in carbohydrate-based cancer therapeutics. Pharmaceuticals 12:84. https://doi.org/10.3390/ph12020084 Kim WS, Lee YS, Cha SH, Jeong HW, Choe SS, Lee M-R, Oh GT et al (2009) Controlling central and peripheral AMPK activity Berberine improves lipid dysregulation in obesity. Am J Physiol Endocrinol Metab 296:E812–E819. https://doi.org/10.1152/ajpendo.90710.2008 Kumar A, Jaitak V (2019) Natural products as multidrug resistance modulators in cancer. Eur J Med Chem 176:268–291. https://doi.org/10.1016/j.ejmech.2019.05.027 Li L, Dai HJ, Ye M, Wang SL, Xiao XJ, Zheng J, Chen HY, Luo YH, Liu J (2012) Lycorine induces cell-cycle arrest in the G0/G1 phase in K562 cells via HDAC inhibition. Cancer Cell Int 12:49 Liu M, Zhao G, Cao S, Zhang Y, Li X, Lin X (2017) Development of certain protein kinase inhibitors with the components from traditional chinese medicine. Front Pharmacol. https:// doi.org/10.3389/fphar.2016.00523 Lu J-J, Bao J-L, Chen X-P, Huang M, Wang Y-T (2012) Alkaloids isolated from natural herbs as the anticancer agents, Evid Based Complement Alternat Med Article ID 485042, pg 1–12. doi:https://doi.org/10.1155/2012/485042. Hindawi Publishing Corporation Maiti M, Suresh Kumar G (2010, 2010) Polymorphic nucleic acid binding of bioactive isoquinoline alkaloids and their role in cancer. J Nucleic Acids:593408, 23 pages. https://doi.org/10.4061/ 2010/593408. SAGE-Hindawi Access to Research Mao L, Chen Q, Gong K, Xu X, Xie Y, Zhang W, Cao H, Hu T, Hong X, Zhan Y-Y (2018) Berberine decelerates glucose metabolism via suppression of mTOR-dependent HIF-1α protein synthesis in colon cancer cells. Oncol Rep. https://doi.org/10.3892/or.2018.6318

588

K. Bhadra

Matsuda D, Satoshi Ohte S, Ohshiro T, Jiang W, Rudel LL, Hong B, Si S, Tomoda H (2008) Molecular target of piperine in the inhibition of lipid droplet accumulation in macrophages. Biol Pharm Bull 31(6):1063–1066 Matsuura HN, Fett-Neto AG (2017) Plant alkaloids: main features, toxicity and mechanisms of action; plant toxins. Springer, Dordrecht, pp 243–261 Moser BR (2008) Review of cytotoxic cephalostatins and ritterazines: isolation and synthesis. J Nat Prod 1:487–489 O’Connor SE (2008) Alkaloid biosynthesis. In: Wiley encyclopedia of chemical biology. https:// doi.org/10.1002/9780470048672.wecb004 Pla D, Albericio F, Alvarez M (2008) Recent advances in lamellarin alkaloids: isolation, synthesis and activity. Anti Cancer Agents Med Chem 8:746–760 Pofahl M, Wengel J, Gunter Mayer G (2014) Multifunctional nucleic acids for tumor cell treatment. Nucleic Acid Ther 24(2). https://doi.org/10.1089/nat.2013.0472 Russo P, Cesario A (2012) New anticancer drugs from marine cyanobacteria. Curr Drug Targets 13:1048–1053 Sanchez ES (ed) (2015) Alkaloids biosynthesis, biological roles and health benefits, biochemistry research trend. Nova Science Publishers, Inc., New York Sarkar S, Bhadra K (2018) Therapeutic role of harmalol targeting nucleic acids: biophysical perspective and in vitro cytotoxicity. Mini-Rev Med Chem. https://doi.org/10.2174/ 1389557518666171211164830 Sarkar S, Bhattacharjee P, Ghosh T, Bhadra K (2020) Pharmaceutical efficacy of harmalol in inhibiting hepatocellular carcinoma. Futur J Pharm Sci 6(29):1–18. https://doi.org/10.1186/ s43094-020-00045-x Seca AM, Pinto DC (2018) Plant secondary metabolites as anticancer agents: successes in clinical trials and therapeutic application. Int J Mol Sci 19:263 Shi C-C, Liao J-F, Chen C-F (2001) Comparative study on the vasorelaxant effects of three harmala alkaloids in vitro. Jpn. J Pharmacol 85:299–305 Shweta S, Hima Bindu J, Raghu J, Suma HK, Manjunatha BL, Mohana Kumara P, Ravikanth G, Nataraja KN, Ganeshaiah KN, Uma Shaanker R (2013) Isolation of endophytic bacteria producing the anti-cancer alkaloid camptothecine from Miquelia dentata Bedd. (Icacinaceae). Phytomedicine 20:913–917 Siva GV, Babu CS (2014) Targeting study of plant alkaloid Berberine act as a natural drug inhibitor of Hsp90(heat shock proteins). Int J ChemTech Res 7(5):2130–2132 Sivanesan S, Taylor A, Zhang J, Bakovic M (2018) Betaine and choline improve lipid homeostasis in obesity by participation in mitochondrial oxidative demethylation. Front Nutr. https://doi.org/ 10.3389/fnut.2018.00061 Song X, Xiong Y, Qi X, Tang W, Dai J, Gu Q, Li J (2018) Molecular targets of active anticancer compounds derived from marine sources. Mar Drugs 16:175. https://doi.org/10.3390/ md16050175 Tamaoki T, Nomoto H, Takahashi I et al (1986) Staurosporine, a potent inhibitor of phospholipid/Ca++ dependent protein kinase. Biochem Biophys Res Commun 135:397–402 Tinoush B, Shirdel I, Wink M (2020) Phytochemicals: potential lead molecules for MDR reversal. Front Pharmacol. https://doi.org/10.3389/fphar.2020.00832 Tohme R, Darwiche N, Gali-Muhtasib H (2011) A journey under the sea: the quest for marine anti-cancer alkaloids. Molecules 16:9665–9696. https://doi.org/10.3390/molecules16119665 Viana J de O, Félix MB, Maia M dos S, Serafim V de L, Scotti L, Scotti MT (2018) Drug discovery and computational strategies in the multitarget drugs era. Braz J Pharm Sci 54(Special):e01010. https://doi.org/10.1590/s2175-97902018000001010 Vrba J, Doležel P, Vièar J, Modrianský M, Ulrichová J (2008) Chelerythrine and dihydrochelerythrine induce G1 phase arrest and bimodal cell death in human leukemia HL-60 cells. Toxicology in Vitro 22:1008–1017 Wu S, Powers S, Zhu W, Hannun YA (2016) Substantial contribution of extrinsic risk factors to cancer development. Nature 529(7584):43–47

21 Anticancer Natural Alkaloids as Drug Bank Targeting Biomolecules

589

Yang R, Piperdi S, Gorlick R (2008) Activation of the RAF/Mitogen-Activated Protein/Extracellular Signal-Regulated Kinase/Extracellular pathway mediates apoptosis induced by Chelerythrine in osteosarcoma. Clin Cancer Res 14:6396–6404 Yu CR, Mandlekar S, Tan TH, Kong AN (2000) Activation of p38 and c-Jun N-terminal kinase pathways and induction of apoptosis by chelerythrine do not require inhibition of protein kinase. J Biol Chem 275:9612–9619. https://doi.org/10.1074/jbc.275.13.9612 Zhang Y, Huang W-R (2019) Sanguinarine induces apoptosis of human lens epithelial cells by increasing reactive oxygen species via the MAPK signaling pathway. Mol Med Rep. https:// doi.org/10.3892/mmr.2019.10087 Zhang X, Gu L, Li J et al (2010) Degradation of MDM2 by the interaction between berberine and DAXX leads to potent apoptosis in MDM2-overexpressing cancer cells. Cancer Res 70(23):9895–9904 Zhang D, Yang R, Wang S et al (2014) Paclitaxel: new uses for an old drug. Drug Des Devel Ther 8:279 Zhang C, Deng J, Liu D, Tuo X, Xiao L, Lai B, Yao Q, Liu J, Yang H, Wang N (2018) Nuciferine ameliorates hepatic steatosis in high-fat diet/streptozocin-induced diabetic mice through a PPARα/PPARγ coactivator-1α pathway. Br J Pharmacol 175:4218–4228 Zhang H, Yang L, Wang Y, Huang W, Li Y, Chen S, Song G, Ren L (2020) Oxymatrine alleviated hepatic lipid metabolism via regulating miR-182 in non-alcoholic fatty liver disease. Life Sci 257:118090. https://doi.org/10.1016/j.lfs.2020.118090

Mode of Materials, Technology, and Devices

22

Shivani Jakhar, Surender Duhan, Supriya Sehrawat, Atul Kumar, Sunita Devi, and Sonia Nain

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nanomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods of Synthesis of Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Immobilization Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Properties of Nanostructured Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Physical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Characterization Techniques of Nanostructured Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Surface Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

593 595 598 598 602 604 604 604 605 605 605 605

S. Jakhar Nanomaterials Research Laboratory, Department of Physics, D.C.R. University of Science and Technology, Murthal, India Inorganic Chemistry Research Laboratory, Department of Chemistry, D. C. R. University of Science and Technology, Murthal, India S. Duhan () · A. Kumar Nanomaterials Research Laboratory, Department of Physics, Deen Bandhu Chhoturam University of Science and Technology, Murthal Sonepat, Haryana, India S. Sehrawat Nanomaterials Research Laboratory, Department of Physics, D.C.R. University of Science and Technology, Murthal, India S. Devi Organic Chemistry Research Laboratory, Department of Chemistry, M.K.J.K. Collage, Rohtak, India S. Nain Inorganic Chemistry Research Laboratory, Department of Chemistry, D. C. R. University of Science and Technology, Murthal, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_105

591

592

S. Jakhar et al.

Surface Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Surface Charge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Crystallography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cosmetics and Sunscreens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nanocomposites as Sensing Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

606 606 606 606 606 607 607 607 608 608 608 609 609

Abstract

Sensors are the fabricated devices that easily respond to any input due to the significant physical alteration in the surrounding atmosphere. The smart performance of a sensor is in correspondence to the sensing material used in its manufacturing. Nanomaterials are the targeted approach and preferable choice of scientists in recent times due to their tremendous desirable characteristics properties in sensing industry. According to ISO and ASTM standard guidelines, we can define nanomaterials as the materials having the particle size within the range of 1 to 100 nm covering its one, two, or all three dimensions. Nanoparticles can be categorized as organic, inorganic, and carbon-based nanostructured particles with their nanoscale particle size dimensions, which have much advanced and improved characteristics properties as compared to when equal mass of same substance is produced with larger size dimensions. Because of their nanoscale sized particles, nanomaterials as a sensing material show many improved desired properties like sensitivity, stability, durability, swift response/recovery time with respect to input from the change in surrounding atmosphere, low hysteresis, significantly increased surface area, etc. There are various literature reported practices adopted for the synthesis of nanostructured particles, which can be broadly classified into three major categories, namely, physical method, chemical method, and mechanical method of synthesis; each one is intended to produce much amended sensing material. These methods of synthesis are still evolving over the years to produce advanced sensing materials to fabricate smart sensors. Another broad classification of nanoparticles includes various methods into bottom-up and top-down approach of synthesis. Sol-gel, spinning, chemical vapor deposition (CVD), pyrolysis, and biosynthesis methods of synthesis are the part of bottom-up approach, whereas mechanical milling, nanolithography, laser ablation, sputtering, and thermal decomposition methods belong to topdown approach of synthesis. In recent years, due to greater surface area of the synthesized material with hydrothermal and nanocasting, chemical methods of synthesis of nanoparticles are extensively used by researchers and scientists.

22 Mode of Materials, Technology, and Devices

593

Further the synthesized nanoparticle is investigated by various characterization techniques like powder X-ray diffraction technique including both low and wideangle patterns, scanning and transmission electron microscopy, analysis of N2 adsorption and desorption isotherms from BET for enhanced surface area, and energy dispersive X-ray analysis. Finally, this chapter deals with the great utility and applications of the nanostructured particles in various fields. Keywords

Nanocomposite · KIT-6 · Cubic mesoporous · Hydrothermal · Hysteresis · Sensor

Introduction According to the definition defined by Gopel and Schierbaum (Zhao and Ning 2001) “sensors are devices that receive a signal or stimulus and responds with an electrical signal.” Presently the advanced research in sensing industry is mainly focused on practical increase in the reliability factor, stability factor, and effective lowering of production and manufacturing cost of sensing material and sensors that will further transform the chemical as well as physical change into electrical signals. For the same, a variety of sensing materials have been produced and characterized for further investigation for the fabrication of smart sensors (Zhao and Ning 2001). With a significant comparison in characteristics and desirable properties for an effective and practical sensor, metal/metal oxide-doped nanomaterials have emerged as a major sensing material comprising of great mechanical endurance and chemical resistivity in variable atmospheric conditions showing stable and repetitive electrical changes. They have great applicability as humidity and gas sensors as their sensing material (Zhao and Ning 2001; Faraday 1857). A sensor mainly has three components for the sensing mechanism to function, namely, receptor, transducer part, and lastly operation mode. The porous surface of the sensing material acts as the receptor in the sensor fabrication that quickly responds to any significant physical any chemical change in the variable atmospheric conditions. Therefore, all the sorption processes comprising of adsorption and desorption part or any chemical or physical reaction will take place on the surface of metal/metal oxidedoped nanocomposite. Physical adsorption, that is, physisorption, and chemical adsorption, that is, chemisorption, both processes define the adsorption mechanism. In case of chemisorption courses, there is electronic transfer of water molecule and the solid material for humidity sensing. In contrast, there is no transfer for physisorption process. Both processes also depend on the reaction temperature. As a consequence, this adsorption of water molecule in humidity sensing will results in significant alteration in the diminution layer of sensing material which are further converted into electrical signals by the transducer part of the sensor having great impact of mesoporous framework of the sensing film. The porosity of the sensing material and particle size of synthesized nanocomposite mainly govern the output

594

S. Jakhar et al.

signals which are mainly in the form of electrical signals. Other than humidity and gas sensing, the mesoporous metal/metal oxide-doped nanomaterials have huge practical applications like cosmetics and sunscreens, electronics, catalysis, health and medicine, agriculture and food, renewable energy and environmental remediation, construction, and most importantly in sensing industry in the designing of various smart sensors like humidity and gas sensors. This chapter involves detailed description about the target achievements of the present reported research work comprising of detailed synthesis and characterization techniques of the various nanocomposite materials and various diversified humidity sensors describing need of humidity measurement. The precise humidity level calculation, using modern technology developed advanced humidity sensors, is a very important phenomenon for the control of various humidity-dependent industrial activities and mechanized systems. The usage of microelectronics and computers can be significantly counted in the design and fabrication of advanced humidity devices. In the understanding levels, nanoscience can be explained as the science involving the study of transformation of a substance present at atomic, molecular, and macromolecular size to a nanoscale size having physical and chemical properties largely different than its manufacturing materials and further involves investigation, characterization, and applications of the synthesized material in designing humidity devices with controlled shape and size at the nanometer. Due to great applications in research and medical, the synthesis and investigation of nanocomposite is a huge emerging research area for scientists with great funding from governments of all countries. In the prehistoric times, nanoparticles have great application as a dye substance in ceramic materials (Zhao and Ning 2001; Faraday 1857; Postma et al. 2001; Tans et al. 1997); and for the treatment of disease like dipsomania, arthritis, etc., gold nanoparticles were used intensively as a therapy material. After studying all these applications, near about in nineteenth centuries, various series of research experiments were done for synthetization and investigation of nanoparticles (Faraday 1857; Postma et al. 2001; Tans et al. 1997). Many experiments and research work were done to improve the properties and desired characteristics of the synthesized nanoparticles materials. With the help of modern-day progressed scientific technology, scientists now can easily alter any substance present at atomic level to upward scale to attain a material of desired properties. In parallel to demand of the civilization, there were many scientific technologies being developed by scientists to synthesize a material with superior properties that are very much specified in nature for the manufacture and fabrication of humiditysensitive devices. This development in scientific technologies is very important to attain high level human comfort. This all development is achieved at a great pace in past few years. Still there is scope of great improvisation and development in the humidity sensitive devices manufacturing technology. With the advancement of technology in nanoparticles, there is great scope of advancement in the discipline area of physics, chemistry, biology, engineering, and mathematics. The field of nanoscience and nanotechnology is a vast area encompassing synthetization,

22 Mode of Materials, Technology, and Devices

595

characterization, investigation, analysis and further applicability in the field of sensing, catalysis, medicine, and target drug delivery using nanocomposite substances. As stated earlier, a vast utility and significance of nanoparticles can be seen from ancient times mainly in the field of catalysis and target drug delivery mechanism in medical area. For the advancement and development of nanostructured particle devices, the major phenomenon is the construction and design of atomic particles to a miniature scale. Nanoparticle technology is the path to the production of next age of microelectronics. For instance, single-walled carbon nanotubes (SWCNTs) are the base materials for the manufacturing of devices like single electron transistor (Postma et al. 2001; Tans et al. 1997) and field effect transistor (Keren et al. 2003; Tans et al. 1998; Martel et al. 1998) that were developed in past few years and demonstrated as simple logical circuits in the form of proto type (Bachtold et al. 2001; Collins et al. 2001; Rueckes et al. 2000; Derycke et al. 2001). Further many interdisciplinary investigations on nanostructured materials are done on the basis of material nature to enhance the involvement of various scientific technologies comprising of physics, biochemistry, surface chemistry, clusters, catalysis, computer technique including various modeling systems, etc.

Nanomaterials A science-oriented approach based on material is the great findings of nanomaterial fields. This field of research area includes the analysis of morphological features of the substances minimized to nanoscale size. It studies the materials with morphological features belonging to nanoscale having properties specifically derived with respect to their nanosized magnitudes. Nanoscale-sized dimensions can be defined as a scale equal to one thousandth times of a micrometer scale. Nanomaterials research field area is also about the learning of various characterized properties that are derived after reducing the scale size to nanoscale. The vast research area in the nanoscience is very important to make significant changes in the sensing world technology of the sensors. Nanomaterial term is comprising of an enormous variability of nanostructured substances with containment of nanometer range dimension. The structural nanomaterials characteristics of nanomaterials are in between those of atoms and bulk materials. In the manufacture and fabrication of novel advanced devices, semiconducting nanostructured materials show improved derived properties. As electrical, optical and chemical properties can be tuned with nanosized particle in a wide range. Metal oxide semiconductors have attracted much attention on the size-dependent phenomenon. In recent years, metal oxide semiconductor nanoparticles received considerable attention as active components in a wide variety of basic research and technological applications due to their improved electric, optical, and magnetic properties compared to their bulk counter-parts. The infinite possibilities that nanotechnology has on the production of nanomaterials are going to significantly alter the material world. Due to their nanometer size, nanomaterials are already

596

S. Jakhar et al.

known to have many novel properties. Due to their small dimensions, nanomaterials have extremely large surface area to volume ratio, which makes a large fraction of atoms of the materials to be on the surface or interfacial atoms, resulting in more surface dependent material properties. After producing the nanomaterials comparable to Debye length, there is significant notice in the change in desired properties of the nanostructured materials due to change in the surface properties of the materials (Ogawa et al. 1982; Luth 1995). Further we can say that as with many new technologies, there seems to be a lack of complete understanding when it comes to how nanoparticles change the properties of materials. Over the past decade, nanomaterials have been the subject of enormous interest (Cao 2004; Knauth and Schoonman 2002; Edelstein and Cammarata 1998; Rao et al. 2004). Nanomaterials have a great application in the field of industries, biomedical industry and electronic industry, because of their extremely small structured size. As a result of huge funding and investments from government as well as private universities, there is formation of many cooperative associations at individual as well as community stages. There are many structured formations of nanomaterials like composite materials, ceramics, and polymeric materials having notable difference in their structure because of their size and variable dimensions. Because of variable characteristics structured properties and applications of nanomaterials in the biomedical industries, in electrical and mechanical industries like for the formation of batteries, in the packaging industries, and in automobiles industries. Figure 1 illustrates the progress chart due to growth in nanotechnology science. Therefore, now we can clearly define nanomaterials as the materials with very minute size preferably in the range of 1–100 nanometers in all the dimensions. Generally, the stability of highly stable substances can be altered and reduced to a great level by increasing the surface area to volume ratio by making the size infinitesimal small to include in the category of nanomaterials. Therefore, after reducing the size significantly to a level of 100 nm, there is noticeable change in the physical as well as chemical properties and the group of substances is characterized as nanomaterials. There are many methods that are used for the production of nanomaterials like sol-gel route, hydrothermal method, nanocasting practice, etc. Because of variable characteristics, structured properties and applications of nanomaterials in the biomedical industries, in electrical and mechanical industries like for the formation of batteries, in the packaging industries and in automobiles industries. The first usage of nanomaterial was claimed by General Motors for the application in exterior automotive for the running boards in mid-sized vans. Further manufacturing minute electronic circuits in the year 2001 was regarded as the most important scientific development in the year of 2001 in the molecular special edition of Journal Science. It is clear that researchers are merely on the threshold of understanding as well as development and that a great deal of fundamental work remains to be done. Further there is variation in the characteristic properties for nanostructured materials mainly because of two main important phenomena. First reason is identified as the comparatively much larger surface area as compared

597

Fig. 1 Graphical abstract

22 Mode of Materials, Technology, and Devices

598

S. Jakhar et al.

to the equal mass produced the same material in the larger particle size. This will further result in the enhanced stability and significant increase in chemical reactivity affecting the electrical as well as mechanical strength with enhanced properties. Secondly, the quantum effect shows a much greater impact on the electrical, mechanical, optical, as well as magnetic behavior of the material present at nanoscale level. The nanoscale dimension of the nanostructured materials can extend up to one, two, or all the three dimensions. Therefore, in simple words, we can say that the main goal line of nanotechnology and nanoscience is to produce new nanostructured materials with enhanced characteristic properties. Therefore, the various areas covered under the nanoscience field of study can be identified as in sensor industry, drug development, and information and communication technologies.

Methods of Synthesis of Nanoparticles The various literatures reported applied practices, in the synthesis of nanostructured materials, are the said determinant factor for the particle size and specific surface area. The main classification of the adopted synthesis techniques as reported in the literature can be divided into two main broad fields: (a) Chemical Methods Proceeding via Bottom-up Approach (b) Physical Methods Proceeding via Top-down Approach

Chemical Methods The nanostructured product obtained via chemical practice is very much specific and versatile substances in controlling the specific surface area and size of the nanoscale particles. The great homogeneity level can be achieved by chemical methods by mixing the base substances at the molecular level. The whole chemical approach includes various methods as mentioned below: (a) Precipitation methods: (i) Hydrothermal method of synthesis (ii) Emulsion precipitation method (iii) Co-precipitation method (iv) Sol-Gel method The end product of all the above methods is in the form of precipitate. (b) Immobilization methods: (i) Citrate complexation gelation method (ii) Penchini chemical method of synthesis

22 Mode of Materials, Technology, and Devices

599

Precipitation Methods (i) Hydrothermal synthesis method or solvothermal synthesis: The prominent usage of hydrothermal procedure is its relatively very low processing temperature of hydrothermal treatment, which plays determinant role in controlling the pore size and surface area of synthesized nanoparticles. The hydrothermal method can be defined as the derivation of single crystals from the aqueous solutions at very high pressure because of their enhanced characteristics mineral solubility. This adopted hydrothermal treatment for the synthesis of nanostructured materials has major desirable characteristics like proceeding of the crystal growth without usage of any catalyst to activate the reaction, comparatively low reaction proceeding temperature, eco-friendly reaction having product and by product production, greater enhanced specific surface area of the synthesized and very low production cost with very less hazardous properties. The proceeding of the reaction at very low reaction temperature is very specific for the production and functioning of micro as well as plastic electronics. The dilution of water and aqueous solutions is done at relatively high pressure, that is, approximately 80 MPa to 300 MPa, mainly to dissolve metal oxides, metal silicates, and sulfides, which have very minimal solubility at normal temperature and pressure conditions. The determinant factors that play a major role in impartation of desirable characteristics properties of the end product and enhance the rate of reaction in terms of reaction kinetics are the starting pH level of the reaction medium, the time interval, and value of constant maintained temperature and pressure of the proceeding reaction. The steel walled autoclaves, with ability to maintain high pressure and temperature conditions inside the container, are used for the synthesis of nanostructure materials. To achieve the targeted high specific surface area and other desirable characteristic properties, firstly all the reactants will be under hydrothermal treatment at constant room temperature with the noticeable enhancement in the crystallization of available amorphous phases of the reactants during hydrothermal treatment (Komarneni et al. 1993; Meskin et al. 2006). Then the achieved suspension will be shifted to steel walled autoclave to achieve and maintain the high-pressure conditions for the proceeding reaction without maintaining any temperature rise. Same approach is used in various chemical synthesis methods, for example, hydrothermal ultrasonic, electrochemical, microwave, having major rewards as compared to other conventional methods. Another major advantage of hydrothermal treatment is the stability and formation of various metastable phases due to very low reaction proceeding temperature. Also, the time attained in the completion of a hydrothermal process is much less as compared to any other conventional method.

Emulsion Precipitation Method In the emulsion precipitation method , as the name of the process suggests, there is formation of an emulsion phase comprising of having high thermal stability, by the addition required and calculated amounts of surfactants into a phase of water and oil. This emulsion system has many droplets which further have atoms in it. For the formation of thermally stable precipitate, the active exchange of reactant species

600

S. Jakhar et al.

through the droplets is mandatory. Also, according to Einstein-Smoluchowski equation, “The normal rate of the particle growth is faster than the equivalent rate of exchange between droplets.” This further results in the decreased emulsion growth as well as retards the formation and agglomeration of larger molecules due to very less nucleation process as compared to other homogeneous systems. The surfactants also play a major role in deciding the droplet dimensions and enhance the thermal stability of the formed emulsion. After the water extraction, to attain particle stability, many other multi additives are used. In the nanostructured materials synthesis process, firstly the oil phase like containing cyclohexane, heptane, etc., is mixed with multisurfactants like octanol, tergitol, etc., and following the removal of organic phase, there is proceeding of dispersal of particles. Further calculated addition of water is done to obtain a turbid emulsion by dynamic shaking. And the stirring was done for a couple of hours after the dropwise addition of emulsion into the alcoholic alkoxides solutions . Then to remove all the formed micelles, the filtrate was passed into the acetone after solvent extraction. Thus, the obtained solid product was dried completely after removal of organic phase and calcination was done to further transform into nanostructures spinals. The major advantage of this process is the resistance in formation of larger molecules and comparatively great retardation in agglomeration of particles during emulsion formation in its bubbles and that too at very low proceeding reaction temperature (Wouundenberg 2001; Tai et al. 2001). Further to achieve all the desired characteristics properties of the end product, firstly there is formation of the mixture by atomic level mingling resulting into precipitation formation and then the chemical homogeneous mixture is obtained during the enhancement of the reaction for the formation of oxide precipitation having many components. There is also a drawback to overcome during emulsion precipitation process which is the segregation of phase partially due to possibility of precipitation at various levels and with different kinetics during subsequent reaction processing which can be termed as co-precipitation of different precursors.

Co-precipitation Method The formation of precipitate, from the homogeneous solvent during the synthesis process of nanostructured materials, totally depends on the size of particles having nanoscale dimensions. In the proceeding of the reaction, firstly the dissolution of salt like chloride, fluoride, nitride, of inorganic metal is done in water medium. These inorganic metal cations find their stability with water of crystallization molecules in their complex structure which can also be regarded as metal hydrate compounds like Al (H2 O)3+ or Fe (H2 O)6 3+ . These inorganic metal hydrates will then be mixed with the solution having basic medium, for example, the solution of sodium hydroxide (NaOH). After the condensation of the hydrolyzed particles, the uniform repetitive washing was done which further followed by drying and calcination to get the end desired product. Briefly there are two levels of the co-precipitation process. It starts with the agglomeration of impurities because of particle growth during precipitate separation process and then redeployment of the same happens between the solvent medium and the formed particle precipitates. During the very first level, the trapping of impurities can take place at two possible spots: the first

22 Mode of Materials, Technology, and Devices

601

one is on the surface which can be termed as surface co-precipitation and the second one is inside the forming particles which can be termed as volume coprecipitation. In case of growth of nanostructured particles comprising of a fixed crystal structure, there are two very possible locations of localization of impurity particles in the case of volume co-precipitation, that is, in the locality of crystal defects arena or on the surface of formed crystal phase in its crystalline structure. Further there is noticeable transference of impurity particles to the medium having precipitates. After the supersaturation condition is achieved or when the medium having substance particles is super cooled, there observed co-precipitation takes place at the end of the reaction. The kinetics of the reaction, the reaction proceeding and precipitate separation temperature, duration of the reaction (can extend from seconds to hours), the chemical purity of the reactant species, etc., determine the characteristic properties of the desired end product (Schleicher and Friedlander 1996).

Sol-Gel Method The sol-gel synthesis method is very popular in ceramic designing and manufacturing industry as well as in arena of material science. The sol-gel process proceeds through a wet-chemical route. This method of synthesis of nanostructured materials and metal oxides exhibits major advantages in comparison to other literature-reported synthesis methodologies. This practice provides better homogeneous phase throughout the solution, very low reaction proceeding temperature, continuous maintained stoichiometry throughout the reaction medium, more flexible atmosphere for the formation of multilayer or monolayer films, and differently structured nanoparticles. Further the next major step lies in the decantation of the solvent. The end product gel can be obtained in various forms with different chemical and physical properties like aerogel, cryogel, etc. The noticeable point is that throughout the sol-gel synthesis process, the specific nanodimensions of the nanostructured materials remain preserved with specifically high specific surface area. Sol-gel method is based on the principle of suspension of solid reactant atoms approximately in the size dimensions of 0.1–1 μm in a solvent medium existing in liquid state. This dispersion mechanism lies in the explanation given by the Brownian motions. The dispersion of both solid and liquid mediums results in the formation of gel which gives a network of solid-state having liquid particles also resulting in a colloidal solution. Major returns of the usage of sol-gel method of nanoparticle formation are as under: (a) There is formation of a thin film acting as bond coating for the protection of metallic nanostructure. (b) This coating is also corrosion resistant and provides the same protection throughout the reaction process. (c) The desired shape of nanostructured product in the end state of gel can be obtained with varying geometries like films, nanorods, etc.

602

S. Jakhar et al.

(d) The end desired product is in highly pure state because of the property dissolution of the precursor material into the solvent and to form a gel like medium by hydrolysis whose composition is highly specific and can be controlled to a great extent. (e) The sintering process happens at relatively very low temperature, mainly at 200– 600 ◦ C. (f) Further this method is very economical, eco-friendly, and with very simple approach for the production of high effective coatings. Every method has its advantages and disadvantages to exhibit. Due to many limitations, the sol-gel method has never resembled its potential at industrial level because of its very low porosity control mechanism, and stability of nanostructured end material is very less because of weak bonding, low exhibiting resistance, and high permeability tendency. In the sol-gel process firstly, the precursor is present in the form of colloidal solution, termed as sol, for the network of polymers which are in the form of gel. Typical precursors exist in the form of chlorides, nitrates, oxides, etc., of metals, which further proceeds with the mechanism of hydrolysis and polycondensation reactions. In this chemical method of synthesis, there is gradual transformation of “sol” precursor into the gel form having two phase system of both solid and liquid which have their morphologies values ranging from atomic level to particles to further derived polymers (Gusev 2007; Philippou 2000). To recognize the gel-like properties of the end product, a given interval of time was provided for the sedimentation process to occur in the colloidal solution, followed by decantation of extra fluid (Brinker and Scherer 1990; Wright and Sommerdijk 2001). For the sedimentation process and phase separation, centrifugation is also a better option. Further solvent removal is done by drying leading to noticeable shrinkage. The kinetics of the solvent removal and separation process is totally obtained by porous character of the gel. Many structural changes can be executed in the phase formation mechanism to exhibit the final geometry and porosity of the end product. Further this will be used as precursor for the formation of many geometry varying nanostructured materials like ceramics, nanorods, nanopowders, films, etc.

Immobilization Methods Citrate Complexation Gelation Method Here in this method, the stability of metal ions is achieved by using the organic networked precursor medium, to obtain oxide powder of very fine texture through the heating process (Yu et al. 2003). Compositions with more than one component are prepared through this method whose homogeneity is good along with stoichiometric control. Poly-chelates between carbon-oxygen legends of citric acid and metal ions are used here in this method. When we heat these chelates with an alcohol which is polyfunctional then they show polyesterification. Here in this method, evaporation is the stage when chelating process occurs and the precursor medium of metallic salts and citric acid undergoes evaporation. If it is heated even

22 Mode of Materials, Technology, and Devices

603

more, then a glassy gel is produced which is a viscous resin and it is rigid and transparent in nature (Grandqvist and Buhrman 1976). Immobilization of various metal ion mixtures is seen when the making of this rigid system is in its initial stages. Hence, during consequent calcinations, formation of different oxide compositions through segregation is highly minimized.

Penchini Method This is a method through which fine dispersed complexes are formed which are very homogenous. The polymer gel is also an intermediatory result as a product. The basis of this method is: (A) Extreme blending of the positive ions. (B) Formation of polymer gel through conversion of previous solution which is a controlled process. (C) Polymer matrix is removed. (D) Oxide precursor is formed which is highly homogenous. The solution of citric acid along with ethylene glycol is introduced with alkoxides or metal salts and it is asynthetic process (Tai and Lessing 1992a, b; Kakihana and Yoshimura 1999). The balancing of the ions which have different behavior is believed to be done by citric complexes’ formation and its results are comparatively good distribution of ions. The separation of components during the later stages of the process is also prevented. 100 ◦ ´ is the temperature above which citric acid polycondensation starts along with polycondensation of ethylene glycol and this results in the formation of citrate gel which is a polymer. Pyrolysis and oxidation of polymer matrix start when the temperature is more than 400 ◦ ´, which results is production of carbonate precursor (X-ray amorphous oxide). For the desired product which is more homogenous and with more dispersion, then the precursor should be heated more. Nowadays this method is used more for the synthesis of following materials: (A) (B) (C) (D) (E)

Dielectric materials Fluorescent and magnetic materials High temperature superconductors Catalysts For the deposition of oxide films and coatings

Major advantages of this process are comparatively more simplicity, the conditions of the process that are not affected the with the chemical nature of positive ions formed in the product, comparatively lower temperature required for the processing of the precursor, and this leads to the completion of the process with absence of sintering and this results in formation powders of refractory oxides which are nanocrystals in structure.

604

S. Jakhar et al.

Low Temperature Combustion Methods (LCS) This is one of those techniques which saves a lot of time and is very effortless and energy vise very efficient way for the production of powders which are ultrafine (Li et al. 2003). The basis of this method is combustion after gelling. The combustion is of an aqueous medium which contain salts of the metal we desire and some fuels which are organic in nature. The product so formed is very fluffy and has large volume and the surface area is also very large. For example, we can use salts like metal nitrates and fuels like urea, citrate acid, or polyacrylic acid. The fuel more commonly used is citrate acid because of its double functions, as a fuel or reductant and can also be used as a chelating agent. The thing that majorly affects the condition of calcination and the crystallites thus produced is the molar ratio of fuel to nitrate. To prepare the crystalline powders which have homogeneity and have a nanoscale primary particle size, we have to control this ratio and temperature for calcination (Zhang et al. 2010).

Properties of Nanostructured Materials There are many physical and chemical properties specific to nanostructured materials.

Physical Properties The observed physical properties for nanostructured materials are imparted by its varying color, different ultraviolet light absorption, transmission in comparison to intensity of light, reflection and refraction tendencies. Nanostructured materials display high thermal stability, greater electrical and mechanical strength, enhanced elasticity, flexibility, malleability, and ductility. Further, nanostructured materials exhibit intensive properties as a semiconductor, conductor, and superconductor with significant conductivity values. Many carbon-based nanomaterials are totally resistant to ultraviolet light degradation and have high frequency shielding with noticeable sensitivity towards any change on atmospheric humidity, sunlight, air, and any other volatile substances in the atmosphere (Tenne 2002; Huang et al. 2010; De Volder et al. 2013; Fawole et al. 2016).

Chemical Properties The various chemical reactions of nanostructured materials, on the basis of chemical reactivity and chemical nature towards the target molecule, reveal the chemical properties of the nanocomposites. The sensitive behavior towards any variation in moisture content of the atmosphere, heat or any other volatile substance, corrosiveresistant chemical behavior, various oxidative and reductive intermediate reactions, antibacterial and antifungal properties are some of the observed chemical properties

22 Mode of Materials, Technology, and Devices

605

of nanodimensioned materials (Bogutska et al. 2013; Ruales-lonfat et al. 2015; Carlos et al. 2013)

Characterization Techniques of Nanostructured Materials Size The determination of size of the particle of nanostructured materials is very important parameter to characterize for the category revelation of whether nano, micro-, or mesoporous materials. The most commonly used characterization techniques to determine the particle size in the synthesized nanocomposite materials are scanning electron microscopy (SEM) and transmission electron microscopy (TEM). These two microscopic techniques are specific to revelation of size in case of particles only, whereas diffraction methods comprising of laser are specific to that same of bulk solid nanomaterials (Marsalek 2014; Sharma and Rao 2014). The samples present in liquid state have their size determination using centrifugation and photon correlation spectroscopy. For the samples that of in gaseous state, normally scanning mobility particle sizer (SMPS) is used for particulate matter size determination.

Surface Area The specific surface area as well as the ratio of specific area to volume determination of the synthesized nanoparticles is very important parameter to enhance the desired chemical reactivity and characteristics properties of the produced samples. The method used for the determination of specific surface area is known as BrunnerEmmett-Teller method (BET) analysis.

Composition The analysis of chemical as well as elemental composition of the synthesized nanoparticles is a very important parameter to characterize for the determination of how pure the formed sample is and what will be the efficiency of the formed substance as a highly performed sensing material. The removal of foreign elements as impurities is highly important to enhance the efficiency and performance of the sensing material as the foreign elements can result in further reaction to contaminate the solution. The method used for compositional determination normally is X-ray photoelectron spectroscopy (XPS) (Sharma and Rao 2014).

606

S. Jakhar et al.

Surface Morphology The surface morphology plays a very crucial role in the determination of structural features and geometric shape of the sample and of pores present in it, for example, conical, hexagonal, octahedral, cylindrical, spherical. Also, the crystallinity and amorphous layout of the sample surface is also discussed. The common techniques used for the revelation of surface morphology are SEM and TEM (Marsalek 2014; Sharma and Rao 2014; Bzdek et al. 2016; Hodoroaba et al. 2014). In the case of samples being present in liquid state, there is tendency to get accumulated on the surface, whereas when the sample is present in gaseous state, the arresting of the particles is done in an electronic manner or by doing filtration to obtain SEM and TEM characterization photographs.

Surface Charge The charge present on the surface of the synthesized materials is very important aspect to analyze as it decides the quantum of interaction with the targeted molecule. The instrument used for the determination of surface charge is zeta potentiometer. The stability of the produced materials after dispersion in the solution is also determined (Marsalek 2014; Sharma and Rao 2014; Bzdek et al. 2016; Hodoroaba et al. 2014). For the sample in gaseous phase, differential mobility analyzer (DMA) instrument is put in a functional manner to reveal the surface charge.

Crystallography The arrangement and shape of unit cell present in the crystalline solid predicts its crystallinity and the analysis of the came is termed as crystallography. This is determined using powder X-ray method which is comprised of two subcategories, that is, small angle X-ray diffraction (SXRD) and wide-angle X-ray diffraction method (WXRD) (Yano et al. 1996).

Applications Nanoparticles have a very vast range of practical usage applications in the present scenario. Here are some important applications of nanostructured materials discussed briefly towards the establishment of smart sensor.

Cosmetics and Sunscreens Normally sunscreens are used to protect the skin from harmful ultraviolet rays from the sun which have cancerogenic properties and very harmful for the skin. But the

22 Mode of Materials, Technology, and Devices

607

old conventional sunscreens lack in providing the long-term stability and protection towards UV rays. But many nanostructured materials have vast usage in this arena, for example, titanium and zinc oxide special property to reflect the harmful UV rays in the atmosphere residing totally transparent to light passing other than UV rays. This property makes these nanoparticles extra suitable for their usage as pigments in advanced sunscreens. There are many seen usages of iron oxide as pigment in various branded lipsticks (Wiechers and Musee 2010).

Electronics Many nanoparticles in the crystalline state are used vastly in electronics items to improve the display features in terms of larger size and brightness in modern advanced electronic items. Because the display features of electronic items demand lager display size with enhanced brightness, which further intensifies the usage of nanoparticles to fulfill all the technological demands. Many developed advanced light emitting diodes (LED) contain zinc selenide, lead telluride, cadmium sulfide, etc., in the form of nanoparticles to enhance display features (Teng et al. 2008). Whereas the electronic items like phone, laptops, etc., which can be port easily from one place to other have different desired features to fulfill like extremely light in weight, compact size with long lasting batteries. Many nanoparticles in crystalline state like hydrides of nickel and other alkali and alkaline metals can be used in the fabrication of separator plates in batteries in the form of foam named aerogel which is specific to enhance the battery capacity to store much greater energy. Further these nanostructured crystalline materials are larger in specific surface area also which reduces the charging time significantly (Published et al. 2016).

Nanocomposites as Sensing Material Various nanoparticles and metal/metal oxide-doped nanocomposites reveal a great enhancement in the electrical conductivity values because the greater surface area of the nanostructured materials and mesoporous structure of various nanocomposites provide larger number of pores sites for better adsorption of water molecules in case of humidity sensing and other volatile gaseous particles in case of gas sensing. The porous structure of the nanocomposites is responsible for charge transfer from the host nanoparticles to the adsorbed species like any gas particles like of NO2 and NH3 or to water molecules for better binding and enhanced stability to reflect the sensing mechanism (Liu et al. 2011).

Catalysis The nanostructured materials possess much greater surface area, which results in the exhibition of catalytic activity. During many chemical reactions, nanoparticles

608

S. Jakhar et al.

behave as great catalyst towards the fast progression of the chemical reaction because of their high surface to volume ratio (Crooks et al. 2001; Ganesh and Archana 2013). Some of the very famous known applications of nanoparticles as catalyst are in the reduction of NiO to metallic form of nickel using Pt nanoparticles, usage of Pt nanoparticles in the self-propelled catalytic converters.

Medicine The delivery of the particular drug to the specific cell is known as targeted drug delivery and nanoparticles play a very crucial role in the medical field as a targeted carrier for the drug using improved nanotechnology (Mudshinge et al. 2011). The overall side effects observed with the usage of nanoparticles as drug carrier are very negligible in count and can be significantly controlled by varying the drug dosage at the targeted place only which also making the drug very cost efficient. Nanotechnology is also a very efficient method in the repairing action mechanism of the dented tissues leading to their reproduction. Further for the artificial organ grafting, tissue engineering is a very effective technology as proved in the case of evolution of bones in the framework of carbon nanotubes (Shinde et al. 2012). The usage of right amount of gold nanoparticles in mental fitness boosting medicines is very common from the ancient Ayurveda times (Laad and Jatti 2016; Nazari and Riahi 2011).

Food In the food industry, there is a great demand for advanced techniques with improved production, processing, and packaging methods. Nanotechnology plays a crucial role in achievement of all the required goals in food industry. The improvement in production, processing, protection, and packaging of food is attained by implementing nanotechnology as the coating of some nanocomposites has great antimicrobial properties to reflect (Xu et al. 2007). Also, the prominent canola oil production factories use nanoparticles in the form of nanodrops, an additive for incorporating vitamins as well as minerals in the food.

Construction The durability and mechanical strength of the construction concrete material can be improved to a great extent by using nanosilica (SiO2 ) nanoparticles which also makes the construction process way more cost and time efficient with improved safety (Machado et al. 2015). The mechanical strength can also be enhanced by adding hematite (Fe2 O3 ) nanoparticles. New modified steel can be obtained by producing nanosized steel with enhanced tensile as well as mechanical strength in constructing cables (Machado et al. 2015). Next highly used material in the con-

22 Mode of Materials, Technology, and Devices

609

struction industry is glass. Titanium dioxide has great disinfecting and antifouling properties, also acting as a catalyst in the form of coating for chemical reaction involving the breakdown of organic pollutants and other harmful VOCs in the atmosphere (Machado et al. 2015). Also, the corrosion resistivity and self-healing phenomenon of paints can be achieved by the addition of calculated quantity of nanoparticles. The nanoparticles-added paints have water repellent hydrophobic nature which further makes these paints very efficient to make coating layers on metal pipes to make resistant from salt water action (Machado et al. 2015). These are also very cost efficient, economical, and very environment friendly.

Conclusion With the great applicability of nanotechnology in our day to life and in the manufacturing and fabrication of many smart devices, there is great increase in research and modification by scientists and funding industries perspective. Further there is great scope of investigation and research in the terms of nanocomposite synthesis methods for improvement in specific surface area, enhanced adsorption capacity, high thermal and mechanical stability, durability and highly performed smart sensing devices with negligible hysteresis, swift response, recovery time, and high sensitivity.

References Bachtold A, Hadley P, Nakanishi T, Dekker C (2001) Logic circuits with carbon nanotube transistors. Science 294:1317–1320 Bogutska KI, Sklyarov YP, Prylutskyy YI (2013) Zinc and zinc nanoparticles: biological role and application in biomedicine. Ukrainica Bioorganica Acta 1:9–16 Brinker CJ, Scherer GW (1990) Sol-gel science: the physics and chemistry of sol-gel processing. Academic, San Diego Bzdek BR, Zordan CA, Iii GWL, Johnston MV (2016) Nanoparticle chemical composition during new particle formation. Aerosol Sci Technol 45:1041–1048 Cao G (2004) Nanostructures & nanomaterials: synthesis, properties & applications, 1st edn. Imperial College Press, London Carlos L, Einschlag FSG, González MC, Mártire DO (2013) Applications of magnetite nanoparticles for heavy metal removal from wastewater. In: Einschlag FSG, Carlos L (eds) Waste water: treatment technologies and recent analytical developments. InTech, pp 63–77 Collins PG, Arnold MS, Avouris P (2001) Engineering carbon nanotubes and nanotube circuits using electrical breakdown. Science 292:706–709 Crooks RM, Zhao M, Sun LI, Chechik V, Yeung LEEK (2001) Dendrimer-encapsulated metal nanoparticles: synthesis, characterization, and applications to catalysis. Acc Chem Res 34:181–190 De Volder MFL, Tawfick SH, Baughman RH, Hart AJ (2013) Carbon nanotubes: present and future commercial applications. Science 339:535–539 Derycke V, Martel R, Appenzeller J, Avouris P (2001) Carbon nanotube inter-and intramolecular logic gates. Nano Lett 1:453–456 Edelstein S, Cammarata RC (1998) Nanomaterials: synthesis, properties and applications, 2nd edn. Institute of Physics Publishing, London

610

S. Jakhar et al.

Faraday M (1857) The Bakerian lecture: experimental relations of gold (and other metals) to light. Philos Trans 147:145–181 Fawole OG, Cai X, Mackenzie AR (2016) Gas flaring and resultant air pollution: a review focusing on black. Environ Pollut 216:182–197 Ganesh K, Archana D (2013) Review article on targeted polymeric nanoparticles: an overview. Am J Adv Drug Deliv 3:196–215 Grandqvist CG, Buhrman RA (1976) Ultrafine metal particles. J Appl Phys 47:2200–2219 Gusev I (2007) Nanomaterials, nanostructures, and nanotechnologies (in Russian). Fizmatlit Hodoroaba V, Rades S, Unger WES (2014) Inspection of morphology and elemental imaging of single nanoparticles by high-resolution SEM/EDX in transmission mode. Microsc Mircoanal 20:602–612 Huang X, Boey F, Zhang HUA (2010) Graphene-nanoparticle composites. Cosmos 6:159–166 Kakihana M, Yoshimura M (1999) Synthesis and characterization of complex multicomponent oxides prepared by polymer complex method. Bull Chem Soc Jpn 72:1427–1443 Keren K, Berman RS, Buchstab E, Sivan U, Braun E (2003) DNA-templated carbon nanotube field-effect transistor. Science 302:1380–1382 Knauth P, Schoonman J (2002) Nanostructured materials: selected synthesis methods, properties and applications. Springer/Kluwer, Boston Komarneni S, Li Q, Stefansson KM, Roy R (1993) Microwave-hydrothermal processing for synthesis of electroceramic powders. J Mater Res 8:3176–3183 Laad M, Jatti VKS (2016) Titanium oxide nanoparticles as additives in engine oil. J King Saud Univ Eng Sci 30:116–122 Li W, Li J, Guo J (2003) Synthesis and characterization of nanocrystalline CoAl2 O4 spinal power by low temperature combustion. J Eur Cream Soc 23:2289–2295 Liu X, Zhang J, Wang L, Yang T, Guo X, Wu S, Wang S (2011) 3D hierarchically porous ZnO structures and their functionalization by Au nanoparticles for gas sensors. J Mater Chem 21:349–356 Luth H (1995) Surfaces and interfaces of solid materials, 3rd edn. Springer, Heidelberg Machado S, Pacheco JG, Nouws HPA, Albergaria JT (2015) Characterization of green zero-valent iron nanoparticles produced with tree leaf extracts. Sci Total Environ 533:76–81 Marsalek R (2014) Particle size and zeta potential of ZnO. APCBEE Proc 9:13–17 Martel R, Schmidt T, Shea HR, Hertel T, Ph A (1998) Single- and multi-wall carbon nanotube field-effect transistors. Appl Phys Lett 73:2447–2449 Meskin PE, Ivanov VK, Baranchikov AE, Churagulov BR, Tretyakov YD (2006) Ultrasonically assisted hydrothermal synthesis of nanocrystalline ZrO2 , TiO2 , NiFe2 O4 and Ni0.5 Zn0.5 Fe2 O4 powders. Ultrasonics Sonochem 13:47–53 Mudshinge SR, Deore AB, Patil S, Bhalgat CM (2011) Nanoparticles: emerging carriers for drug delivery. Saudi Pharm J 19:129–141 Nazari A, Riahi S (2011) The effects of SiO2 nanoparticles on physical and mechanical properties of high strength compacting concrete. Compos B 42:570–578 Ogawa H, Nishikawa M, Abe A (1982) Hall measurement studies and an electrical conduction model of tin oxide ultrafine particle films. J Appl Phys 53:4448–4455 Philippou J (2000) Sol-gel: a low temperature process for the materials of the new millennium. Sol Gel pub. Postma HWC, Teepen T, Yao Z, Grifoni M, Dekker MC (2001) Carbon nanotube single-electron transistors at room temperature. Science 293:76–79 Published A, Link C, Terms D (2016) Platinum-gold nanoparticles: a highly active bifunctional electrocatalyst for rechargeable lithium-air batteries, J Am Chem Soc 132(2010):12170–12171. https://doi.org/10.1021/ja1036572 Rao NR, Müller A, Cheetham AK (2004) The chemistry of nanomaterials: synthesis, properties and applications. Wiley

22 Mode of Materials, Technology, and Devices

611

Ruales-lonfat C, Barona JF, Sienkiewicz A, Bensimon M, Vélez-colmenares J (2015) Applied catalysis B: environmental Iron oxides semiconductors are efficients for solar water disinfection: a comparison with photo-Fenton processes at neutral pH. Appl Catal B Environ 166–167: 497–508 Rueckes T, Kim K, Joselevich E, Tseng GY, Cheung CL, Lieber CM (2000) Carbon nanotubebased nonvolatile random access memory for molecular computing. Science 289:94–97 Schleicher B, Friedlander SK (1996) Fabrication of aerogel-like structures by agglomeration of aerosol particles in an electric field. J Colloid Interface Sci 180:15–21 Sharma V, Rao LJM (2014) An overview on chemical composition, bioactivity and processing of leaves of Cinnamomum tamala. Crit Rev Food Sci Nutr 54:433–448 Shinde NC, Keskar NJ, Argade PD (2012) Nanoparticles: advances in drug delivery systems. Res J Pharm Biol Chem Sci 3:922–929 Tai W, Lessing PA (1992a) Modified resin-intermediate processing of perovskite powders, part I optimization of polymeric precursors. J Mater Res 7:502–510 Tai W, Lessing PA (1992b) Modified resin-intermediate processing of perovskite powders, Part II Processing for fine, nonagglomerated Sr-doped lanthanum chromite powders. J Mater Res 7:511–519 Tai CY, Lee M, Wu Y (2001) Control of zirconia particle size by using two emulsion precipitation technique. Chem Eng 56:2389–2398 Tans SJ, Devoret MH, Dai H, Thess A, Smalley RE, Geerligs LJ, Dekker C (1997) Individual single-wall carbon nanotubes as quantum wires. Nature 386:474–477 Tans SJ, Verschueren ARM, Dekker C (1998) Room-temperature transistor based on a single carbon nanotube. Nature 393:49–52 Teng W, Jeng S, Kuo C, Lin Y, Liao C, Chin W (2008) Nanoparticles-doped guest-host liquid crystal displays 33(15):1663–1665. https://doi.org/10.1364/OL.33.001663 Tenne R (2002) Fullerene-like materials and nanotubes from inorganic compounds with a layered (2-D) structure. Coll Surf A Physicochem Eng Aspects 208:83–92 Wiechers JW, Musee N (2010) Engineered inorganic nanoparticles and cosmetics: facts, issues, knowledge gaps and challenges. J Biomed Nanotechnol 6:408 Wouundenberg FCM (2001) Nanostructured oxide coating via emulsion precipitation. Thesis, University of Twente, Enschede, Netherlands Wright D, Sommerdijk NAJM (2001) Sol-gel materials-chemistry and applications. Gordon & Breach, Amsterdam Xu X, Stevens M, Cortie MB (2007) In situ precipitation of gold nanoparticles onto glass for potential architectural applications, Chem. Mater. 16(2004):2259–2266. https://doi.org/ 10.1021/CM034744Y Yano F, Hiraoka A, Itoga T, Kojima H, Kanehori K, Mitsui Y (1996) Influence of ionimplantation on native oxidation of Si in a clean-room atmosphere. Appl Surf Sci 100–101:138–142 Yu X, He X, Yang S, Yang X, Xu X (2003) Synthesis and luminescence of Sr2 CeO4 superfine particles by citrate-gel method. Mater Lett 58:48–50 Zhang Z, Wang W, Shang M, Yin W (2010) Low temperature combustion synthesis of Bi2 WO6 nanoparticles as a visible-light-driven photocatalyst. J Hazard Mater 177:1013–1018 Zhao H, Ning Y (2001) China’s ancient gold drugs. Gold Bull 34:24–29

IoT-Based Medication Reminder Devices: Design and Implementation

23

Alivelu Manga N. and Sathish P.

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Block Diagram of the System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . NodeMCU ESP8266 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . System Peripherals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction to IoT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Architecture of IoT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Block Diagram of IoT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Implementation of IoT-Based Medicine Reminder System . . . . . . . . . . . . . . . . . . . . . . . . . . . Arduino IDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Installation Steps to Arduino IDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Salient Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Blynk App . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Libraries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Configuring SMTP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . System Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Project Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interfacing Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

614 616 617 619 623 625 626 627 627 627 631 634 634 634 637 638 638 643 644

A. M. N. () · S. P. Chaitanya Bharathi Institute of Technology, Osmania University, Hyderabad, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_106

613

614

A. M. N. and S. P.

Abstract

IoT (Internet of Things) refers to a system of Internet-connected devices which are capable of sending and receiving the data without human intervention. This technology enabled the remote monitoring in healthcare sector which leads to keep the patients safe and heathy, and ensuring to deliver superlative care. Some people apparently should be taken care by the caretakers and other family members. This is not provided by everyone in today’s life. So, they may forget to take medicines at the right time and may also forget what medicine has to be taken. This project aims to develop a device which alerts the patients to take medicine at right time in an efficient manner. The proposed system has a facility which alerts the patients to take medicine on time and also acknowledges the medicine intake through a Gmail notification. In the proposed system, the alarm is set up with the help of Blynk app indicating the patient to take the medicine. The IR sensor interfaced with the microcontroller unit determines the patient medicine intake. The sensor detects the hand movement of the patient when he opens and takes the medicine from the slot provided. The corresponding message is sent to the caretaker via Gmail based on the IR sensor status. Keywords

IoT · IR sensor · Blynk app · Gmail

Introduction Nowadays healthcare has become a burden for systems that are dealing with an ageing population, the incidence of chronic diseases, and rising expenditures. The rising incidence of chronic diseases and the ageing population require a disproportionate amount of healthcare resources. In the world, about 75% of healthcare dollars go to chronic disease care and two out of every three Medicare recipients suffer from at least two chronic diseases. More than 40% of world population suffers from chronic conditions, often with no cure in sight, which can get hugely expensive. When a patient has a chronic illness, continual health monitoring is more vital than prevention and wellness. Several device manufacturers and other parties are attempting to address this problem by integrating the various streams of data required to accurately monitor the health of a patient with a certain illness. The pressure for relief will grow as the population ages with approximately 10,000 new patients estimated to enroll in medical care every day for the next 15 years. Using the concepts of IoT, Android Application, and cloud computing in patients or their caretaker’s smart phone, the patient can view his/her medication status. Healthcare sector is continuously updated with extensive technologies. The growing healthcare challenges include the rise in chronic diseases and limitations on the capability of the hospitals, doctors, and service providers to deliver quality healthcare services in order to improve patient health. Internet of Things (IoT) technology is one of technologies along with artificial intelligence and machine learning and many

23 IoT-Based Medication Reminder Devices: Design and Implementation

615

Fig. 1 Statistics of Global IoT Health Market. (Source: https:/978-3-030-84205-5/appinventiv. com/blog/iot-in-healthcare/)

more. IoT with its own benefits include real-time monitoring, providing accuracy in collection of data and tracking patients activities, meeting the requirements of the users. IoT helps in the healthcare sector at different phases, from making an appointment with the doctor, analyzing the patient condition, and communicating the condition of the patient using mobile application. These changes focus on the individual needs of people, thereby improving the efficiency and quality of medical care, and represent the future development direction of modern medicine science. The statistics of Global IoT Health Market and forecast is shown in Fig. 1. Smart healthcare arose from IBM’s (Armonk, NY, USA) concept of “Smart Planet” proposed in 2009. Simply described, Smart Planet is an intelligent infrastructure that detects information using sensors, transmits it over the Internet of Things (IoT), and processes it using supercomputers and cloud computing. It has the ability to coordinate and integrate social processes in order to achieve the dynamic and sophisticated management of human civilization. Smart healthcare is a healthcare delivery system that leverages wearable devices, the Internet of Things, and mobile Internet to dynamically access information, connect people, materials, and institutions in the healthcare ecosystem, and then manages and responds to those needs. Smart healthcare can foster interaction among all stakeholders in the healthcare industry, ensuring that participants receive the services they require, assisting parties in making informed decisions, and facilitating resource allocation. In a nutshell, smart healthcare is a higher level of medical information architecture. Chronic diseases have risen to the top of the human disease spectrum and have become a new epidemic since the beginning of the twenty-first century. Chronic diseases have a long course of illness, are incurable, and expensive to treat, thus disease health management is crucial. The traditional hospital and doctor-centered healthcare management approach, on the other hand, appears to be failing. Patient self-management is emphasized more in the new health management approach under smart healthcare. It stresses patient self-monitoring in

616

A. M. N. and S. P.

real time, real-time health data feedback, and timely medical behavior modification. The rise of implantable/wearable smart devices, smart homes, and smart health information platforms linked by IoT technology offers a solution to this problem. Advanced sensors, microprocessors, and wireless modules in third-generation wearable/implantable devices can intelligently sense and monitor numerous physiological indicators of patients while reducing power consumption, enhancing comfort, and allowing the data to be coupled with health information from other channels. It decreases the disease’s related dangers while also making it easier for medical institutions to track the disease’s prognosis. The advent of smart phones, smart watches, and other similar devices has provided a new vehicle for this type of surveillance. Biosensors have been tried to be integrated into smartphones. Users can use a high-performance smartphone to improve portability even more. In order to make a really useful smart pillbox it had to be easily integrated with the recent sweeping smart technologies. While at the same time it had be fit for the elders and their limited knowledge and experience to implement the ease of use. Size and portability were also an important fact that one need to keep in mind (Fasahate 2018).

Block Diagram of the System The block diagram of the IoT-based medicine reminder system is given in Fig. 2. The number of IR sensors used depends upon the number of slots in the box. Each IR sensor is placed near the slot such that it detects the hand movement and also each LED is fixed on each slot indicating that the pill has to be taken from that particular slot. The buzzer which is given as the output to the Blynk app rings at the time of the event created. The LED on one of the slots will turn on. The LEDs connected to each slot are also controlled by the Blynk app. The patient takes the pill from that slot. The IR sensor detects the hand movement as soon as the patient takes the pill from the slot. After few minutes, i.e., according to the input time given in the program, a mail is sent to the caretaker. Depending on the IR sensor output the message in the mail changes. If the IR sensor detects the hand movement the IR value reads “1” and the message says that patient has taken the medicine, else IR value reads “0” and the message says that patient has not taken the medicine. The healthcare industry is always evolving with new technologies. The rise in chronic diseases, as well as limits on hospitals, doctors, and service providers’ ability to give good healthcare services in order to enhance patient health, is among the major healthcare concerns. Internet of Things (IoT) technology is one of technologies along with artificial intelligence and machine learning and many more. IoT, with its own set of advantages such as real-time monitoring, data collecting accuracy, and patient activity tracking, matches the needs of users. IoT assists in the healthcare industry at several stages, including scheduling a doctor’s appointment, assessing the patient’s status, and transmitting the patient’s condition via mobile devices. These changes focus on the individual needs of people, thereby improving the efficiency and quality of medical care, and represent the future development

23 IoT-Based Medication Reminder Devices: Design and Implementation

617

Fig. 2 Block diagram of the system

direction of modern medicine science. Elderly people need to take so many tablets in a day. This leads to forget them to take medicine at sometimes. It will be difficult for them to remember the medication schedule if the consumption is multiple times (https://www.ns-healthcare.com/analysis/iot-in-healthcare/). There are many people who require constant assistance, whether they be our old, family members, or those with special needs (Hayes et al. 2006). Elders are more impacted by the timing of taking a medicine. In order to avoid any dysfunction or sickness, timing is critical. Yet, as people age, their vision and memory deteriorate. What if the patient has dementia such as Alzheimer’s. Some people may forget to take their medications on time or may forget what medications they need to take. It is required to discover a simple, portable, and efficient solution to eliminate the causes of constant need for surveillance, such as nurses, or the chance of a missed dose. Pill boxes do exist; however, the majority of them have limited uses, are unsuitable for older people, or are too large to carry around. To be considered smart, it must be connected to the Internet via a wireless network, which allows it to connect to the Internet for future applications and integration. It is also distinguished by the vast range of Wi-Fi rather than Bluetooth or any other field communication, which eliminates the need for any cables or wired connections, which is what makes it portable in the first place (Kumar et al. 2018). It is connected to the mobile phone over the same network, which allows you to select the dosage interval and notify you in a variety of ways when the dose is due. A buzzer with an LED as a tactile reminder is included.

NodeMCU ESP8266 NodeMCU is an open-source Lua-based firmware and development board designed specifically for IoT applications. Firmware is included in which it runs on Espressif

618

A. M. N. and S. P.

Fig. 3 NodeMCU pin description. (Source: https://i2.wp.com/randomnerdtutorials.com/wpcontent/uploads/2019/05/ESP8266-NodeMCU-kit-12-E-pinout-gpio-pin.png?quality=100&strip= all&ssl=1 )

Systems. ESP8266 Wi-Fi SoC and hardware are based on the ESP-12 board. The ESP-12E module on the NodeMCU ESP8266 development board contains an ESP8266 chip with a Tensilica Xtensa 32-bit LX106 RISC microprocessor. This microprocessor runs on an adjustable clock frequency of 80–160 MHz and supports RTOS. To store data and programs, the NodeMCU consists of 128 KB RAM and 4 MB flash memory. NodeMCU can be powered using micro-USB jack and VIN pin (external supply pin). It has interfaces for UART, SPI, and I2C interface (Fig. 3).

Features It consists of Tensilica 32-bit RISC CPU Xtensa LX106Microcontroller. Its operating voltage is 3.3 V. Input voltage ranges from 7 V to 12 V. It has 16 digital input/output pins (DIO). One analog input pin. One UART and one SPI. One I2C and flash memory of 4 MB. 64 KB of SRAM and clock speed is 80 MHz USB-TTL based on CP2102 is included onboard, enabling plug n play and PCB antenna.

• • • • • • • • •

23 IoT-Based Medication Reminder Devices: Design and Implementation

619

Working Principle of ESP8266 The ESP8266 WIFI Module is a self-contained system on chip with an integrated TCP/IP protocol stack that can connect any microcontroller to Wi-Fi. The ESP8266 may run a program or offload all Wi-Fi networking activities to a separate CPU. This module provides sufficient onboard processing and storage to be used with sensors and other applications. Its excellent on-chip integration means that external circuitry is kept to a minimum, and the front-end module is designed to take up as little PCB space as possible. The ESP8266 offers APSD support for VoIP applications and Bluetooth coexistence interfaces, as well as a self-calibrated RF that allows it to work in any environment and no extra RF parts.

System Peripherals Infrared (IR) Sensor An IR sensor is an electronic system which emits light in order to detect objects in the environment. An infrared sensor can detect motion as well as measure the heat of an object. An IR LED (light emitting diode) serves as the emitter, and an IR photodiode serves as the detector. The photodiode is sensitive to infrared light of the same wavelength as the IR LED. As infrared light strikes the photodiode, the resistances and output voltages change in relation to the intensity of the IR light. A typical infrared detection system consists of five fundamental components: an infrared source, a transmission channel, an optical component, infrared detectors or receivers, and signal processing. Infrared sources include infrared lasers and infrared LEDs with specified wavelengths. There are two types of IR sensors: (i) Active IR sensor. (ii) Passive IR sensor.

Active IR Sensor Active infrared sensors are made up of two parts: an infrared source and a detector. LEDs and infrared laser diodes are examples of infrared sources. Photodiodes and phototransistors are examples of infrared detectors. The infrared source’s energy is absorbed by an object and lands on the infrared detector.

Passive IR Sensor Passive infrared sensors are basically infrared detectors. Infrared sensors that are passive do not have an infrared source or detector. They are divided into two categories: quantum infrared sensor and thermal infrared sensor. Infrared energy is used as a heat source in thermal infrared sensors. Thermal infrared detectors include thermocouples, pyroelectric detectors, and bolometers. Infrared sensors of the quantum kind have a better detection efficiency. It is faster than infrared detectors

620

A. M. N. and S. P.

of the thermal kind. Quantum style detector’s photo sensitivity is wavelength dependent.

IR Sensor Working Principle Infrared transmitters are classified according to their wavelengths, output power, and response time. An IR sensor consists of an infrared LED and an infrared photodiode, which are referred to as a photocoupler or optocoupler when used together. IR Transmitter or IR LED A light emitting diode (LED) that produces infrared radiation is known as an infrared transmitter. Although an IR LED appears to be a regular LED, the radiation it emits is invisible to the naked eye. An IR LED serves as the emitter, while an IR photodiode serves as the detector. An IR LED emits infrared light, which is detected by the IR photodiode (Fig. 4). The resistance and output voltage of the photodiode alter as the temperature rises. When the IR transmitter emits radiation, some of it reaches the object and is reflected back to the IR receiver. The sensor’s output is determined by the intensity of the IR receiver’s reception. Light Emitting Diode (LED) An LED is a type of PN junction diode that emits light. It is made up of semiconductor and is specially doped. Two of the most widely used semiconductors for LEDs are aluminium indium gallium phosphide (Al, In, GaP). Minority carrier electrons are doped into the p-region and corresponding majority carrier electrons are doped into the n-region when the forward biased current IF is applied through the p-n junction of the diode. In the p-region, electron-hole recombination causes photon emission (Fig. 5). Photons (i.e., light) are produced by electron energy transitions through the energy gap, known as radiative recombinations, while phonons are produced by shunt energy transitions, known as nonradiative recombinations (i.e., heat). The dominant wavelength emitted by an LED system is used to describe its color (in nm). Red:626–630 nm, red-orange: 615–621 nm, orange: 605 nm, and amber: 605 nm are the colors produced by AlInGaP LEDs (590–592 nm). Green (525 nm), bluegreen (498–505 nm), and blue are the colors produced by InGaN LEDs (470 nm). Fig. 4 IR sensor working. (Source:https://robu.in/irsensor-working/)

23 IoT-Based Medication Reminder Devices: Design and Implementation

621

Fig. 5 Light emitting diode. (Source:https://www.pintere st.com/pin/44705270049166 1611/ )

The temperature of the LED p-n junction affects the color and forward voltage of AlInGaP LEDs. The luminous intensity decreases as the temperature of the LED p-n junction rises, the dominant wavelength changes to longer wavelengths, and the forward voltage drops. From 20◦ –80 ◦ C, the difference in luminous strength of InGaN LEDs with operating ambient temperature is minimal (about 10%). The dominant wavelength of InGaN LEDs, on the other hand, varies with the LED drive current; as the LED drive current rises, the dominant wavelength shifts toward shorter wavelengths.

Buzzer The buzzer is a small but effective component for adding sound to our project/system. Its small and compact two-pin configuration allows it to be used on a breadboard, or even a PCB, making it a popular component in most of the electronic applications. There are two forms of buzzers available in the market. The one shown here is a simple buzzer that produces a continuous beep sound when powered; the other form is a readymade buzzer, which is bulkier than this and produces a beep. Beep, beep. Beep, beep. It produces sound due to an internal oscillating circuit. But it is the most widely used type of buzzer. This buzzer can be used by simply connecting it to a DC power supply between 4 and 9 volts. A simple 9 V battery can also be used, but a controlled +5 V or +6 V DC supply is preferred. The buzzer is usually connected to a switching circuit that turns it on and off at predetermined times and intervals. Liquid Crystal Display (LCD) LCD stands for liquid crystal display and is commonly used to display various values in various electronic projects and devices. Liquid crystal displays (LCDs) use liquid crystals to produce visible images. A simple LCD module used in DIY electronic projects and circuits is a 16 × 2 liquid crystal display. There are two rows in this LCD module, each of which contains 16 numbers. There are 16 columns

622

A. M. N. and S. P.

Fig. 6 16×2 liquid crystal display. (Source: https:// www.theengineeringprojects. com/wp-content/uploads/ 2019/11/Introduction-to16x2-LCD-Module.jpg)

in this module due to the two rows. These modules have a VA dimension of 66 × 16 mm and a thickness of 13.2 millimeters. It runs on a voltage of plus five or three volts (Fig. 6).

Features Functioning voltages range from 4.7 V to 5.3 V. It uses one milli Amp current for operation. LCD can display alpha numeric values. There are two rows each of 16 characters. Every character is of 40 pixels. It supports on four- and eight-bits mode. It displays two colors in screen backlight (i.e., green and blue).

• • • • • • •

23 IoT-Based Medication Reminder Devices: Design and Implementation

623

Fig. 7 DS3231 RTC module interface. (Source: https://components101.com/asset/sites/default/ files/inline-images/DS3231%20RTC-Module-Interfacing-Circuit.png)

Real-Time Clock (RTC) Module RTC modules are essentially TIME and DATE remembrance systems with a battery setup that keeps the module operating in the absence of external power. The TIME and DATE are kept up to date as a result of this. As a result, one can get correct TIME and DATE from the RTC module at any time. DS3231 RTC Module Features The DS3231 RTC module interface is shown in Fig. 7. • DS3231 RTC counts seconds, minutes, hours, and year. • Accuracy: −2 ppm to +2 ppm for 0 ◦ C to +40 ◦ C, −3.5 ppm to +3.5 ppm for −40 ◦ C to +85 ◦ C. • ±3 ◦ C accuracy in digital temperature sensor • Two time-of-day alarms. • Programmable square wave output. • Register for aging trim. • I2C interface with 400KHz frequency. • Consumes low power. • It has automatic power failure battery switch circuitry. • CR2032 battery backup with two- to three-year life. • Portable size.

Introduction to IoT Nowadays, the Internet has spread to almost every corner of the globe and is shaping human life in unprecedented ways. Entire world entered into an age of far more ubiquitous connectivity, with a wide range of appliances connecting to the Internet and living in the “Internet of Things” (IoT) period. This word has been described in

624

A. M. N. and S. P.

a variety of ways by various writers. Let’s take a look at two of the most commonly used meanings. The Internet of Things, according to Vermes and colleagues, is essentially an interface between the physical and digital worlds. Using a variety of sensors and actuators, the digital world communicates with real world. The Internet of Things is also described as a paradigm in which computing and networking capabilities are embedded in any kind of imaginable object, according to another definition. These features are used to query the object’s state and, if possible, to modify it. IoT is a term used to describe a digital environment in which almost all of the devices and appliances in use are linked to a network. One should work together to complete complex tasks that require a high level of intelligence. The Internet of Things (IoT) is the network of physical devices, vehicles, home appliances, and other items embedded with electronics, software, sensors, actuators, and network connectivity which enable these objects to connect and exchange data (McCall et al. 2013). Sensors and actuators are the devices that enable people communicate with their surroundings. To generate usable insights from the data acquired by the sensors, it must be intelligently stored and analyzed. Note that the term sensor is used in a broad sense; a mobile phone or even a microwave oven can be considered as a sensor if it delivers information about its current state. An actuator, such as the temperature controller of an air conditioner, is a device that is used to effect a change in the environment. Data storage and further processing can take place either at the network’s edge or on the remote server. If data preprocessing is possible, it is usually done at the sensor or another nearby device. The data is usually transferred to a distant server after it has been processed. The data storage and processing capabilities of an IoT item are also limited by the available resources, which are frequently limited due to power, computational capability constraints size, and energy. As a result, the primary research challenge is to ensure having the appropriate data at the required level of precision. Along with the challenges of data collecting and handling, communication is also a problem. Because IoT devices are typically located in geographically separated places, communication between them is primarily wireless. Communication methods are essential to the research of IoT devices in this case because reliably transferring data without too many retransmissions is a major issue. Following the processing of the obtained data, some action must be taken based on the derived inferences. Actions may take a variety of forms (Pawar et al. 2014). Actuators enable us to change the physical environment directly. Alternatively, one might do this virtually. For example, one can send some information to other smart devices. The process of effecting a change within the physical world is often dependent on its state at any point of time. This is called context awareness. Since an application may behave differently in various contexts, each operation is considered in light of the context. The central infrastructure of an IoT architecture consists of sensors, actuators, compute servers, and communication networks. However, there are several software considerations to be made. To begin, a middleware that can bind and manage all of these disparate components is essential. To link so many different devices, a lot of standardization is required (Kavya et al. 2018).

23 IoT-Based Medication Reminder Devices: Design and Implementation

625

Healthcare, wellness, education, culture, social life, energy efficiency, environmental monitoring, home automation, and transportation are all areas where the Internet of Things has applications. IoT systems have greatly reduced human effort while also improving overall quality of life (Minaam et al. 2018).

Architecture of IoT IoT is made up of devices that collect data. They are all Internet-connected gadgets, which means each one of them have an IP address. They range in sophistication from self-driving cars that transport 16 goods around production floors to simple sensors that monitor building temperatures (Fig. 8) (Salgia et al. 2015). They also include personal devices such as fitness trackers that track how many steps people walk each day. To make that data valuable, it must be collected, processed, filtered, and analyzed, all of which can be done in various ways. The data is collected from the devices and sent to a central location. Data can be moved wirelessly or across wired networks using a variety of ways. The data can be delivered via the Internet to a data center or cloud with data storage and computing power, or it can be staged, with intermediary devices gathering the data before sending it on. Data processing can take place in data centers or the cloud, but this is not always possible. The delay in delivering data from the device to a remote data center is too significant in the case of important devices such as shutoffs in industrial settings. Sending data, processing it, evaluating it, and returning instructions can take an excessive amount of time. Edge computing can be used in these situations, where a smart edge device can accumulate data, analyze it, and construct solutions

Fig. 8 Architecture of IoT. (Source: https://www.zibtek.com/blog/iot-architecture/)

626

A. M. N. and S. P.

as needed, all within a short physical distance, reducing time. Upstream connectivity is available on edge devices, allowing data to be sent to be processed and stored.

Block Diagram of IoT IoT Gateways An IoT gateway is a solution for allowing IoT communication, which is usually device to device or device to cloud. Usually, a gateway is a hardware computer that houses application software that performs critical functions. The gateway, at its most basic level, enables the links between various data sources and destinations (Fig. 9). Comparing an IoT gateway to your home or office network router or gateway is an easy way to visualize it. This type of gateway allows you to communicate with your devices while still maintaining security and providing an admin interface where you can perform simple tasks. This is what an IoT gateway does, and much more. IoT gateways have developed to perform a wide range of functions, from data filtering to visualization and complex analytics. These intelligent devices are assisting in the current wave of IoT development. An IoT gateway serves as a network router, connecting IoT devices to the cloud. Most gateway devices delivered traffic in only one direction at first: from IoT devices to the cloud. In today’s world, a gateway device is often used to handle both inbound and outbound traffic. Inbound traffic is utilized for device management tasks including changing device firmware, while outbound traffic is used to deliver IoT data to the cloud. Some IoT gateways are more than just traffic routers. Before transmitting data to the cloud, a gateway device can be used to preprocess it locally at the edge. As a result, the device may deduplicate, summarize, or aggregate data in order to reduce the amount of data

Cloud Server Wireless communication Temperature Humidity Pressure

IoT Gateway/Frame work

Motion Lux

Fig. 9 Block diagram of IoT. (Source: https://iotdunia.com/iot-architecture/)

Mobile app

23 IoT-Based Medication Reminder Devices: Design and Implementation

627

that needs to be sent to the cloud. This can have a significant impact on network transmission costs and response times (Kliem et al. 2012).

Sensors and Controllers Sensors are crucial in the development of IoT solutions. Sensors are the devices which detect and replace external data with a signal which can be understand by both humans and machines. Sensors have enabled data collection in almost any environment and are currently employed in a variety of industries, including nursing care, industrial, logistics, medical care, transportation, agricultural, disaster protection, tourism, regional companies, and many others (Abbey et al. 2012). With the expansion of fields in which sensors play a key role, the market for a variety of sensors is still growing (Zanjala and Talmaleb 2015). IoT sensors are used to detect and quantify a variety of physical events, including heat and pressure, and also the five human senses such as sight, hearing, touch, taste, and smell.

Implementation of IoT-Based Medicine Reminder System Arduino IDE Writing code and uploading it to the board is simple with the open-source Arduino Software (IDE). Any Arduino board can be used with this software. The Arduino integrated development environment is used to create Arduino programs (IDE). The Arduino IDE is a piece of software that runs on your computer and allows you to create sketches (the Arduino equivalent of a programmer) for various Arduino boards. The Arduino programming language is built on processing, a very simple hardware programming language that is akin to C. The sketch should be uploaded to the Arduino board for execution after it has been written in the Arduino IDE.

Installation Steps to Arduino IDE Ensure that the board is ready to accept the software via USB cable. Step 1 − To begin, you will need an Arduino board (any board will do) and a USB cable. You will need a typical USB cable (A plug to B connector) to connect an Arduino Uno, Nano, Diecimila, Duemilanove, or Mega 2560 Diecimila, similar to what you would need to connect a USB printer, as illustrated in Fig. 10 If you are using an Arduino Nano, you will need an A to Mini-B cable instead, as indicated in Fig. 11.

Step 2: Download Arduino IDE Software From the Arduino official website’s download page, you can download several versions of the Arduino IDE. You must choose software that is appropriate for your operating system (Windows, IOS, or Linux). Unzip the file once it has finished downloading.

628

A. M. N. and S. P.

Fig. 10 Universal serial bus cable. (Source: //rarecom ponents.com/store/image/ cache/data/0-1197-22-500x 500.jpg )

Fig. 11 Mini-B cable. (Source: //rarecomponents. com/store/image/cache/data/ 0-1197-22-500x500.jpg)

Step 3: Power up your Board The Arduino Uno, Mega, Duemilanove, and Nano all use the computer’s USB port or an external power supply to get their power. Make sure your Arduino Diecimila is set to draw power from the USB port if you are using one. To select the power source, a jumper, a little piece of plastic that slides onto two of the three pins between the two terminals, is employed. Using the USB cable, connect the Arduino board to your computer. The green power LED (labeled PWR) should light up when turned on.

23 IoT-Based Medication Reminder Devices: Design and Implementation

629

Step 4: Launch Arduino IDE You must unzip the folder after downloading the Arduino IDE program. The application icon with an infinity label can be found inside the folder (application.exe). To launch the IDE, double-click the icon.

630

A. M. N. and S. P.

Step 5: Open your First Project • Create a new project. • Open an existing project example.

Select File → Example → Basics → Blynk to open an existing project example. For this example, one of the Blink instances is chosen. It uses a timer to switch the LED on and off. Any other example from the list can be chosen.

Step 6: Select your Arduino Board Make sure you choose the correct Arduino board name that matches to the board connected to your computer to avoid any issues while uploading your software to the board. Go to Tools → Board and select your board. Step 7: Select your Serial Port Select the serial device on the Arduino board. From the Tools menu, choose Serial Port. COM3 or above is most likely (COM1 and COM2 are hardware serial ports). Disconnecting your Arduino board will reveal the entry that vanishes, which should be the Arduino board. Connect the board to the computer and select the serial port you want to use. Reconnect the board and choose the serial port you want to use.

23 IoT-Based Medication Reminder Devices: Design and Implementation

631

Step 8: Upload the Program to your Board Simply click the “Upload” button in the environment now. The RX and TX LEDs on the board will begin to flash after a few seconds. If the upload is successful, the message “Done uploading” will show in the status bar.

Salient Features (i) The Arduino IDE has a text editor that allows users to create sketches. The procedure is straightforward and uncomplicated. Furthermore, the text editor contains additional capabilities that make the experience more dynamic. (ii) Users can get their projects documented using the program. (iii) Users of the Arduino IDE can share their sketches with other programmers. (iv) Hundreds of libraries are included into the software. The Arduino community created and freely shared these libraries. (v) While the tool is designed particularly for Arduino boards, it also supports third-party hardware via native connections. (vi) A board administration module is included in the program, allowing users to select which board they want to utilize. They can effortlessly select another choice from the drop-down menu if another board is required. PORT data is automatically updated anytime the board is modified or a new board is selected.

632

A. M. N. and S. P.

NodeMCU Libraries #Include The Wi-Fi library for ESP8266 was created utilizing the ESP8266 SDK and the Arduino Wi-Fi library’s naming conventions and overall functionality philosophy. As the number of Wi-Fi functions translated from ESP8266 SDK to esp8266/Arduino expanded, it became clear that it is required to give separate documentation on what is new and extra. #Include The Arduino module and editor are unable to communicate with the LCD’s I2C interface. The Liquid Crystal I2C library, which can be downloaded separately, contains the parameter that allows the Arduino to transmit commands to the LCD. Use the Arduino IDE’s Sketch > Include Library > Add.ZIP Library... to install the

23 IoT-Based Medication Reminder Devices: Design and Implementation

633

Liquid Crystal I2C library (see example). Point to the Liquid Crystal I2C-master.zip file that you downloaded earlier.

#Include The RTC records information such as seconds, minutes, hours, days, dates, months, and years. It also adjusts for months with fewer than 31 days and leap years automatically. The clock can be set to either 24 h or 12 h (with AM/PM). There are additionally two programmable time-of-day alarms and a programmable squarewave output. The RTC is communicated with using an I2C interface with a fixed address.

#Include -Library for I2C CommunicationYou can use this library to communicate with I2C/TWI devices. The SDA (data line) and SCL (clock line) are on the pin headers near the AREF pin on Arduino boards using the R3 configuration (1.0 pinout). Two I2C/TWI ports are available on the Arduino Due. SDA1 and SCL1 are close to the AREF pin, while pins 20 and 21 have an additional one.

634

A. M. N. and S. P.

#Include In this library, the Arduino serves as the master device, allowing you to communicate with SPI devices. Microcontrollers interface with one or more peripheral devices across short distances using the synchronous serial data standard serial peripheral interface (SPI). It can also be used to communicate between two microcontrollers. When using an SPI connection, there is always one master device.

Blynk App Blynk is an iOS and Android app that allows you to control Arduino, Raspberry Pi, and other web-connected devices. It is a virtual dashboard where you may drag and drop widgets to construct a project’s graphic interface. Everything is quite easy to set up, and you will be fiddling in just 5 min. Blynk is not tethered to any certain board or shield. Instead, it is about giving you the freedom to utilize whatever equipment you want. When your Arduino or Raspberry pi is connected to Wi-Fi. Blynk will get online and helps to get ready for Internet of Things. The Internet of Things was a driving force behind the development of Blynk. It can remotely manage devices, display sensor data, save and visualize data, and do a range of other things (Kliem 2012).

Components of Blynk App The three major components are: Blynk app: By combining our numerous widgets one can create many amazing features. Blynk server: This is an interface between the smartphone and hardware and communication is carried out. You have the option of using our Blynk Cloud or setting up your own Blynk server on your own computer. It is free and open source, with the ability to support tens of thousands of devices. It can even run on a Raspberry Pi (Fig. 12).

Libraries Blynk libraries: These libraries enable communication with the server and handle all incoming and outgoing commands for all popular hardware platforms.

Configuring SMTP SMTP (simple mail transfer protocol) is a framework for automatically sending and receiving massive amounts of email from remote places. Developers and marketers typically use it to save time when sending emails in a secure manner due to its

23 IoT-Based Medication Reminder Devices: Design and Implementation

635

Fig. 12 Blynk app components. (source-https://docs.blynk.cc/images/architecture.png)

speedy and dependable service. Its servers and data centers are located all over the world, allowing it to select the closest server and so deliver the fastest service. It can be used in IoT projects to automatically send emails when a specific activity is completed. By using SMTP2GO, email warnings when the flame sensor detects fire are send.

Setting up SMTP2Go Go to https://www.smtp2go.com/ and click on Try SMTP2GO Free. Fill in your name, email address, and password, then click Submit. Then you will be taken to a screen where you will be asked to activate SMTP2GO. Go to your inbox and select the SMTP2GO-received message. Select Activate Account from the drop-down menu. Now type in your username, which is your email address, and your password. Your username will appear on a new page. These login and password should be saved in a notepad file because they will be used later in the Arduino IDE code. Now select Users from the drop-down menu under Settings. Your user name and SMTP server information will appear on a new page. The SMTP server and SMTP

636

A. M. N. and S. P.

port should be saved because they will be used in the Arduino code to connect to the SMTP server (Majumder et al. 2017).

Encoding to Base64 Value One must encrypt login and password into Base64 before sending them to the SMTPTOGO server. Use https://www.base64encode.org/ to encode them. To generate the encoded values, simply input the login and password you want to encode into Base64 and click Encode. The encoded values should be copied and saved. If your user name is “[email protected],” for example, type “[email protected]” in the provided text box and then click “encode.” “cGFzc3dvcmQ=” is the Base64 encoded value. Do the same thing with password. By following the above steps, username and password in Base64 format can be obtained to use them as Wi-Fi credentials in program source code.

23 IoT-Based Medication Reminder Devices: Design and Implementation

637

System Flowchart The libraries and header file required for RTC module, LCD, Wi-Fi, and Blynk app are included in the program. The baud rate for the controller should be set up as 9600. The required credentials should be given for Wi-Fi to send the notification via Gmail. After giving the credentials the Wi-Fi gets connected. The LCD interfaced with RTC module displays the date and time continuously. The buzzer rings according to the time given in the Blynk app. The IR sensor interfaced with the microcontroller unit reads the digital value of IR pin. The RTC module compares the time with the given input time (Polities 2016). Based on the status of the IR pin the mail is sent to the caretaker displaying the corresponding message (Fig. 13).

638

A. M. N. and S. P.

A smart medicine reminder box is included, as well as sensors for health monitoring and diagnosis. A smart home-based medicine box with wireless connectivity allows patients and caregivers to communicate more effectively. The suggested system includes a feature that reminds patients to take their medicines on time. The box is remotely connected to the Wi-Fi and delivers notifications to the caregiver via Gmail in order to make timely modifications (Tsai et al. 2017).

Project Setup The LCD interfaced with RTC module displays the date and time continuously. The buzzer rings at the given input time in the Blynk app which alerts the patient to take medicine. After opening the medicine box, the patient takes the medicine from the respective slot. This is indicated with the help of LED connected to each slot. The IR sensor detects the hand movement and after few minutes, i.e., according to the time setup in the program, the mail is sent to the caretaker displaying the corresponding message (Fig. 14).

Interfacing Modules LCD and RTC Module with NodeMCU LCD and RTC modules are connected to NodeMCU for timing information. The liquid crystal display(16×2) interfaced with the real-time clock module displays the time and date (Fig. 15).

23 IoT-Based Medication Reminder Devices: Design and Implementation

639

Start

sensor status is given by the function irvalue=digital Read(irpin)

Include libraries and header files required for RTC module ,WiFi,Blynk app,LCD display

Set up WiFi and Baud rate

RTC time comparison now. Hour () ==given hour && now. Minute==given_minute

False

Declare the corresponding GPIO pins for LEDs in each slot and variable irpin=0 to read irsensor value

True

irvalue==0 WiFi connection False

True

True

LCD displays time on the screen obtained from RTC module using the function (Date Time.now==RTC.now())

Mail sent saying patient has not taken the medicine

False

Mail sent saying that patient has taken the medicine Buzzer rings according to the time specified in the Blynk app Stop

Fig. 13 Flowchart of the proposed system

Buzzer with NodeMCU The hour is obtained to the controller with the help of Blynk app. Events (Alarm) are created in the Blynk app and the buzzer is connected to one of the GPIOs of the controller. The buzzer rings at that particular instant of time (Fig. 16). IR Sensor with NodeMCU IR sensor detects the object in the surroundings by emitting the light. IR sensor can detect the object as well as the motion of the object. Figure 17 shows the photodiode of the IR sensor detecting the hand movement by turning on the in-built light emitting diodes (LED). Figure 18 shows that there is no motion detection, i.e., LED is off.

640 Fig. 14 Overall project setup

Fig. 15 LCD and RTC module interfaced with NodeMCU

Fig. 16 Buzzer interfaced with NodeMCU

A. M. N. and S. P.

23 IoT-Based Medication Reminder Devices: Design and Implementation

641

Fig. 17 IR sensor detection of hand movement

Fig. 18 IR sensor status is 0

Medicine Reminder through Blynk App One of the GPIO pin is given as the output for the Blynk app. This can be switch ON and switch OFF according to the time in the event created. In this way the buzzer rings at a given time alerting the patient (Fig. 19). Acknowledgment Via Gmail The sensor detection helps us to know the patient’s consumption of pill by sending the corresponding message via Gmail. If the sensor detects the hand movement it is considered that the patient has taken the medicine and the corresponding message is sent. If the hand movement is not detected, the message is sent via Gmail saying that the patient has not taken the medicine (Fig. 20) (Slagle et al. 2011). Deployment of the Proposed System The LCD interfaced with the RTC module displays the date and time (Fig. 21).

642 Fig. 19 Set up reminder using Blynk app

Fig. 20 Mail received to the caretaker

A. M. N. and S. P.

23 IoT-Based Medication Reminder Devices: Design and Implementation

643

Fig. 21 LCD displaying timing information

Fig. 22 Distribution of pills in the slots

The box is divided into three slots (Fig. 22). Each slot is provided with an IR sensor and a LED. The LEDs are controlled by the Blynk app and they turn on each slot indicating that the particular slot pill has to be consumed. The IR sensor in each slot determines the hand movement and sends the notification via Gmail.

Summary The components and characteristics of an embedded system are discussed including the importance of the processor in a system. The debugging tools in an embedded system and also the criteria for choosing a microcontroller are explained. The domain of the proposed system’s Internet of Things is also explained with its evolution over the years. Different ways of communication in IoT are also explained.

644

A. M. N. and S. P.

NodeMCU which is the microcontroller unit for the proposed system is described with its specifications and working. The system peripherals LED, LCD, RTC module, and IR sensor are also discussed along with their features and working principles. The working principle of the proposed system is explained with the help of block diagram. Arduino IDE software required for the system design is explained with installation steps. The libraries required for the system to run are also discussed. Blynk app which helps in reminding the patient to take the medicine is discussed briefly. Installation steps for Blynk app and how to create a reminder is also explained. The results of the proposed chapter are also heighted in the last chapter. The deployment of the proposed system is shown and the working is explained. The developed medicine reminder box assures the medicine intake of the patient according to the specified duration. It also reduces the effort of remembering the medicine which has to be taken as the LCD displays the timing schedule. An efficient algorithm is developed which helps the patient’s medicine intake at right time. With the help of this system there is no need of continuously monitoring the patient so that in the absence of caretaker, the patient can take pill. The LED connected to each slot indicates the patient which pill should be consumed. The IR sensor interfaced with the microcontroller ensures the consumption of pill. The developed medicine box is cost efficient. Further, this work can be extended, not only in health sector, but also in industrial and automotive applications where time management is crucial. In order to develop efficient patient monitoring system, the designed system can be connected to other sensors to sense temperature, pulse, etc., thereby connecting to the system interface so that it may be placed in a nursing station and the pill details of various patients are recorded in the system.

References Abbey B, Alipore A, Camp C, Hofer C (2012) The smart pill box. In: Proceedings of the rehabilitation engineering and assistive Technology Society of North America Fasahate MA (2018) Smart medicine box using IOT. Int J Sci Eng Res 9(2) Hayes TL, Hunt JM, Adami A, Kaye JA (2006) An electronic pillbox for continuous monitoring of medication adherence. In: Proceedings of the 28th IEEE EMBS annual international conference, Aug. 30–Sept. 3 Hiba Zeidan, Khalil Karam, Roy Abi Zeid Daou, Ali Hayek, Josef Boercsoek, MART Learning, Education and Research Center “Smart Medicine Box System” 2018 IEEE International Multidisciplinary Conference on Engineering Technology (IMCET) https://appinventiv.com/blog/iot-in-healthcare/ https://www.businessinsider.in/science/news/iot-healthcare-in-2020-companies-devices-usecases-and-market-stats/articleshow/74126142.cms https://www.hindawi.com/journals/jhe/2021/6632599/ https://www.ns-healthcare.com/analysis/iot-in-healthcare/ Huang S-C, Chang H-Y, Jhu Y-C, Chen G-Y (2014) The intelligent pill box – design and implementation. IEEE Pawar S, Kulkarni PW (2014) Home based health monitoring system using android smart phone. Int J Electr Electron Data Commun 2(2) Feb

23 IoT-Based Medication Reminder Devices: Design and Implementation

645

Kavya M, Reddy GRK (2018) Intelligentpillbox: automatic and programmable assistive technology device. Int J Eng Res Manag Technol Eng 8(X) Oct Kliem A, Hovestadt M, Kao O (2012) Security and communication architecture for networked medical devices in mobility-aware eHealth environments. In: 2012 IEEE international conference on Mobile service, pp 112–114 Kumar SB, Goh W, Balakrishnan S (2018) Smart medicine reminder device for the elderly. In: Fourth international conference on advances in computing, communication & automation (ICACCA), Subang Jaya, Malaysia 26–28 Oct Majumder S, Aghayi E, Noferesti M, Memarzadeh-Tehran H, Mondal T, Zhibo P, Deen M (2017) Smart homes for elderly healthcare – recent advances and research challenges. Sensors 17(11):1–32 McCall C, Maynes B, Zou CC, Zhang NJ (2013) An automatic medication self-management and monitoring system for independently living patients. Medical Engineering and Physics Minaam DSA, Abd-ELfattah M (2018) Smart drugs: improving healthcare using smart pill box for medicine reminder and monitoring system. Future Computing Inform J 3:443e456 Polities C (2016) A new generation of e-health systems powered by 5G. WWRF Ray P (2014) Home health hub internet of things (H 3 IoT): an architectural framework for monitoring health of elderly people. Sci Eng Manag Res 1:3–5 Salgia AS, Ganesan K, Raghunath A (2015) Smart Pill Box. Indian J Sci Technol 8(S2):189–194 Sawand A, Djahel S, Zhang Z, Na F (2014) Multidisciplinary approaches to achieving efficient and trustworthy eHealth monitoring systems. In: 2014 IEEE/CIC international conference on Communications in China (ICCC), pp 187–192 Slagle JM, Gordon JS, Harris CE, Davison CL, Culpepper DK, Scott P, Johnson KB (2011) MyMediHealth – designing a next generation system for child-centered medication management. J Biomed Inform 43(5):27–31 Tsai H-L, Tseng C, Wang L, Juang F-S (2017) Bidirectional smart pill box monitored through internet and receiving reminding message from remote relatives. In: IEEE international conference on consumer electronics – Taiwan (ICCE-TW). Taipei, Taiwan Zanjala SV, Talmaleb GR (2015) Medicine reminder and monitoring system for secure health using IOT. In: International conference on Information Security & Privacy (ICISP2015), Nagpur, India, 11–12 December

Ionic Liquids: The Smart Materials in Process Industry

24

Kailas L. Wasewar

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ionic Liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . What Is Ionic Liquid? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Origin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classification of Ionic Liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Advantages over Organic Solvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications of Ionic Liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analytical and Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electro Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Few Important Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

648 651 651 652 653 653 655 655 659 661 663 664 669 670 671

Abstract

It is mainly considered that the chemical process and allied industries are responsible for environmental pollution. It is true to some extent; hence it is the role of scientists and engineers to fulfill the social expectations by providing sustainable design and development of chemical processes so that there must be less generation of hazardous materials and hence less negative impact on environment. Ionic Liquids (ILs) and IL-based materials are the promising alternative to conventional and traditional materials. ILs and IL-based materials

K. L. Wasewar () Advance Separation and Analytical Laboratory (ASAL), Department of Chemical Engineering, Visvesvaraya National Institute of Technology (VNIT), Nagpur, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_126

647

648

K. L. Wasewar

have been emerged as smart materials in process industry because of many unique properties of ionic liquids such as almost zero vapor pressure, nonflammability, and tunable physiochemical properties for a particular application. The basic properties which can be tuned for specific applications are thermal phase behavior, thermal stability, viscosity, conductivity, etc. Keywords

Ionic liquids · Smart material · Tunable properties · Applications · Process industry Ionic liquids (ILs) comprise the combination of various cations and anions, and these liquids are different from ionic solutions (a molecular solvent salt solution having melting point less than or equal to 100 ◦ C). As per available literature, alternatively ionic liquids are termed as room-temperature molten salts, ionic melts, fused salts, organic ionic liquids, nonaqueous ionic liquids, liquid organic salts, molten salts, and ionic fluids. Ionic liquids are a kind of green and novel compound having wide application in the area of chemistry, biotechnology, chemical engineering, pharmaceutical areas due to their unique properties such as tunable polarity, non-volatility, and high stability. ILs are majorly classified based on cations as imidazolium, pyrazolium, pyrrolidinium, pyridinium, phosphonium, ammonium, morpholinium, piperidinium. Further ILs may classified as conventional, functionalized (mono, dual), renewable, superbase, polymerized, biomaterial, supported IL membranes and materials, reversible, etc. Since last few decades, ILs have been widely employed into many industrial applications which includes biomass processing, electrolytes, electropolishing and electroplating, lubricants, solar cells, biological systems, chromatography, fuel cells, in catalyzed reaction, solvent, in membrane, in microwave-assisted reactions, esterification reactions, synthesis of many chemicals, separation, pharmaceutical (therapeutic, drug delivery, anticancer, antiviral), electrochemistry (electrolyte, fuel cells, metal plating, solar panels, battery), environment (pollution treatment, greenhouse gas, green solvent), materials (nanomaterials, liquid crystals), engineering (coating, polymer, textile, nuclear, additives, engineering fluids), energy (biomass, biofuel, bioenergy), chemistry (analytics, physical, reaction), biotechnology (protein purification, bioreaction, bioprocess), electronic, heat storage, surface and colloid, propellants, solvent, separation, catalyst, and many more. In the present chapter various aspects of ILs, its origin and characteristics, classification and types, synthesis, and various applications in process industry as a smart material have been discussed.

Introduction The world population has been increasing day by day and hence the globalization for the development of the society. There have been a wide range of products used in day-to-day life including utilities, medical, food, and other important products

24 Ionic Liquids: The Smart Materials in Process Industry

649

Fig. 1 Number of publication for 2009–2018 based on Web of Science search (permission taken for Singh S. K., Savoy A. W., Ionic liquids synthesis and applications: An overview, Journal of Molecular Liquids 297 (2020) 112038, 1–23 from Elsevier)

and materials. The increase in the consumption of these materials forced to develop new materials to fulfill global demands. It can be observed that in our day-to-day life various products, materials, and services are used which includes food products, paints, drugs, agrochemicals, dyes, biomolecules, and high-tech substances. These are mainly associated with process industry, where chemicals are used for the production of materials and specific products. These chemicals are an integral part of the day-to-day life in today’s world. Scarcely, no industry can be found without the use of chemicals and also no economic sector without chemicals as it plays a very significant role in industrialization. Ionic liquids (ILs) are mainly fluids which comprise totally ions (cation and anion). Ionic liquids are adaptable fluids having wide potential and alternative applications in the fields of separation processes, catalytic reactions, and electrochemistry, etc. In the last decade, there has been drastic improvement in publications based on various topics and category of ionic liquids and their applications (Figs. 1 and 2) (Singh and Savoy 2020). A lot of research work has been performed on ionic liquids since more than last ten decades. The Google Scholar search on “Research papers on Ionic Liquid” have been performed year wise, which include books, book chapter, review article, research papers, short note, editorial, concept notes, patents, technology, and other papers. Figure 3 shows the growth in research in ionic liquids since 1960, and it can be observed that there have been significant exponential increases in research on ionic liquids since 2000. Ionic liquids mainly comprise cation (organic) and anion (inorganic or organic). These are more attractive due to their tunability of properties by the almost infinite cation-anion combinations and make them as designer solvents (EURARE 2020).

650

K. L. Wasewar

Fig. 2 Publications on ionic liquids in various areas for 2009–2018 based on Web of Science search (permission taken for Singh S. K., Savoy A. W., Ionic liquids synthesis and applications: An overview, Journal of Molecular Liquids 297 (2020) 112038, 1–23 from Elsevier)

Ionic liquids have distinctive properties observed from a complex interplay of Coulombic, hydrogen bonding, and van der Waals interactions of their ions. Ionic liquids are often considered as green solvents due to their exceptional properties as compared to other solvents. These ionic liquids comply with the principles of green chemistry. There are more than 108 combinations of cations and anions, which can be possibly made to have target-specific ionic liquids. Worldwide, the society has been looking for clean, efficient, and ecofriendly ways of sustainable developments; the ionic liquids may be promising alternative to conventional solvents (mainly organic) having many advantages because of inimitable thermal, physical, chemical, and biological features (Singh and Savoy 2020). Ionic liquids have wide applications in the field of process and allied industries. In general, the sustainable development or green chemistry directs the avoidance or reduction of environmental pollution and waste at various laboratory, pilot, and industrial scales. It also explores the development and design of environmentfriendly and economical approaches which may improve the required yield and reduce the waste generation in the process. The synthesis of ionic liquids is not straightforward; it requires sophisticated equipments and controlled conditions to get desired ionic liquids. Various synthesis methods for different ionic liquids have been discussed in literature (Singh and Savoy 2020).

24 Ionic Liquids: The Smart Materials in Process Industry

651

Fig. 3 Number of research documents since 1960 based on Google Scholar. (Source: Google scholar searched on “Research papers on Ionic Liquid” referred on 18th May 2021)

In the present chapter various aspects of ILs, its origin and characteristics, classification and types, synthesis, and various applications in process industry as a smart material have been discussed.

Ionic Liquids What Is Ionic Liquid? Ionic liquids can be defined as salts that exist as liquids below 100 ◦ C or even at room temperatures. Ionic liquids have basically organic cations and inorganic or

652

K. L. Wasewar

Fig. 4 Few cations for various ionic liquids

organic anions linked through various kinds of bonds. The common cations and anions used for ionic liquids for various applications are listed in Figs. 4 and 5. Ionic liquids are also named as room or ambient or room or low-temperature ionic fluid or molten salts or neoteric solvent or liquid organic salts (Singh and Savoy 2020).

Origin It was mentioned in the literature that ionic liquids were first time introduced in 1914 as ethyl ammonium nitrate [EtNH3 ][NO3 ] having a melting point of 12 ◦ C (Sugden and Wilkins 1929; Singh and Savoy 2020). The powder of alkyl pyridinium chloride at little hot condition was added in aluminum chloride, and clear colorless liquid was formed; this was invented by Frank Hurley and Tom Weir (Rice Institute of Texas), which was later considered as ionic liquid (Zhang and Shreeve 2014, Singh and Savoy 2020). The development of ILs during that period was still inquisitive and remained for four to five decades. A few review papers on various standpoints have been presented in the literature (Welton 2018; Angell et al. 2012; Wilkes 2002).

24 Ionic Liquids: The Smart Materials in Process Industry

653

Fig. 5 Few anions for various ionic liquids

Classification of Ionic Liquids Ionic liquids can be classified based on cation, applications, and mainly structure. The ionic liquids based on structural evaluation are classified as conventional ionic liquids, basic ionic liquids, task-specific ionic liquids, polymeric ionic liquids, protic ionic liquids, energetic ionic liquids, bio-ionic liquids, neutral ionic liquids, switchable polarity solvent ionic liquids, metallic ionic liquids, chiral ionic liquids, and supported ionic liquids (Singh and Savoy 2020). The structural classification of ionic liquids has been summarized in Fig. 6.

Advantages over Organic Solvents Ionic liquids have many superior characteristics and properties over conventional organic solvents, which have been widely used for various applications in process

654

K. L. Wasewar

Fig. 6 Typical classification of ionic liquids

industry. These properties are mostly superior (Fig. 7) as compared with most of the conventional organic solvents. The typical superior features of ionic liquids are (Singh and Savoy, 2020): • • • • • • • • • • • • • • • • •

Mostly liquid at a wide range of temperatures Solvating ability Negligible vapor pressure Non-flammable Non-volatile Liquids for a wide range of temperature as compared to water Intrinsic ionic conductors High electrochemical range (maximum >4 V) Tunable hydrophobic and hydrophilic behavior Exceptional properties for lubrication and hydraulic applications Tunable basicity and acidity High thermal stability Mostly colorless and polar Mostly poses low viscosity Good solvent for a wide range of inorganic and organic materials May be used in high vacuum system Easily recycle

24 Ionic Liquids: The Smart Materials in Process Industry

655

Fig. 7 Few features of ionic liquids

• • • •

Good shelf life Highly polar Non-coordinating Conductivity

Applications of Ionic Liquids Ionic liquids have various exceptional properties making them a promising and potential material for numerous applications including process industries and other allied industries. In process industries, it has been employed for separations, reactions, electro processes, analytical, chromatography, and many other applications.

Separations Separations have a significant role in process industry and contributing almost more than 50% production cost. Ionic liquids are a promising alternative to conventional solvents due to their few superior properties. Ionic liquids have been

656

K. L. Wasewar

successfully used for separation of biologically relevant compounds, organics, proteins, hormones, alkaloids, vitamins, antibiotics, aromatic hydrocarbons, sulfur compounds, CO2 , various metals, etc.

Separation of Biologically Relevant Compounds There are many compounds which can be obtained by fermentation such as organic acids or directly from natural methods, and these compounds may be building block chemicals or starting material in chemical, cosmetic, food, and pharmaceutical industries. Organic acids are important building blocks for a range of other chemicals in industry and are produced in considerable volumes (Sprakel and Schuur 2019). These compounds have been separated using many ionic liquids (Marta’k and Schlosser 2008; Marques et al. 2013). The lactic acid extraction was investigated using trihexyltetradecylphosphonium cation with bis(2,4,4-trimethylpentyl)phosphinate (Cyphos IL-104) or chloride anions (Cyphos IL-101), and performance was better than organic solvents. Cyphos IL-104 was employed for the separation of butyric acid, and many fold higher efficiency was observed as compared to lactic acid using organic solvents (Marta’k and Schlosser 2008). Tetracycline was removed efficiently from synthetic fermentation solution using tetrabutylphosphonium chloride, which can extract tetracycline from synthetic solution (Marques et al. 2013). Separation of Organics Separation of various organic compounds from aqueous and organic phases is mostly performed by solvent extraction due to its operation at ambient conditions. Organic solvent possesses many drawbacks such as expensive, toxic, and flammable nature. Ionic liquids have been considered as an alternative and potential solvent for the extraction of organics and other separation applications of metals, alkaloids, and proteins (Khazalpour et al. 2020). Ionic liquids have been employed for the removal and separation of various organic compounds by extractive distillation, aqueous biphasic systems, liquid–liquid extraction, liquid phase micro-extraction, supported liquid membranes, etc. (Poole and Poole 2010). Various ionic liquids have been studied by extraction for the separation of organic compounds. Few of these ionic liquids are trihexyltetradecylphosphonium dicyanamide(P666,14 [N(CN)2 ]), trihexyl tetradecyl phosphonium chloride, trihexyl(tetradecyl)phosphonium bis(trifluoromethylsulfonyl)imide P666,14 [NTf2], trihexyl(tetradecyl)phosphonium bis-2,4,4-(trimethylpentyl)phosphinate P666,14 [Phos], for the separation of aromatics, guaiacol, microalgae lipids, L-tryptophan, b-carotene, rhodamine 6G, caffeine, astaxanthin, terpenes, and terpenoids. The exhaustive and critical review on the separation of organic compounds using ionic liquids has been available in the literature (Poole and Poole 2010). Separation of Proteins Proteins have major application in biopolymers having more than 20 amino acids (Schindl et al. 2019). Ionic liquids were successfully employed as solvent for

24 Ionic Liquids: The Smart Materials in Process Industry

657

the processing of keratin protein, and few of these ionic liquids are [AMIM][Cl], [AMIM][N(CN)2 ], [C4 C1 Im][N(CN)2 ], [C4 C1 Im][Bu2 PO4 ], [C4 C1 Im][Br], etc. Ammonium-based commercial ionic liquid Ammoeng 110 with phosphate buffer was used for the aqueous biphasic separation of various proteins myoglobin, lysozyme, trypsin, albumin, and myoglobin, and 83–100% extraction was observed (Dreyer et al. 2009). Ionic liquid, 1,3-dialkylimidazolium bromide [C4,6,8 MIM]Br was employed with phosphate buffer for the extraction of cytochrome c, trypsin, albumin, and γ-globulin proteins by aqueous biphasic separation with 75–100% extraction efficiency (Pei et al. 2009). The removal of human serum, myoglobin, and immunoglobulin G was investigated using ionic liquids of alkyl-substituted phosphonium-[NTf2 ] (Pei et al. 2007). The exhaustive review describing various ionic liquids and their application for the processing of proteins has been presented in literature (Schindl et al. 2019).

Separation of Hormones, Alkaloids, and Vitamins The content of human urine was determined in terms of testosterone and epitestosterone with 80–90% extraction using 1-methyl-3-butylimidazolium chloride and K2 HPO4 (He et al. 2005). Caffeine and nicotine alkaloids were isolated 100% from human urine using 1,3-dialkylimidazolium chloride with K3 PO4 (Freire et al., 2010). Vitamin B12 present in human urine was measured using 1-hexyl3-methylimidazolium chloride with K2 HPO4 having 97% extraction efficiency (Berton et al. 2012). Ionic liquid butyl-methyl-imidazolium chloride with K3 PO4 or KH2 PO4 was employed for the selective extraction of quinine from human plasma (Flieger and Czajkowska-Zelazko 2015). Separation of Antibiotics The development and use of suitable downstream separation processes for the separation of antibiotics from fermentation broths is one of the challenging technological issues. The aqueous biphasic separation of penicillin G from fermentation broth with more than 90% efficiency was observed using BMIM BF4 and NaH2 PO4 (Lui et al. 2005). The extractive separation of mydecamycin, azithromycin, and roxithromycin were performed using aqueous biphasic separation system containing 1-butyl,3-methylimidazolium tetrafluoroborate (BMIM BF4 ) and salts as NaH2 PO4 , Na2 HPO4 , NaCl, Na2 SO4 , Na2 CO3 , and NaOH with a recovery of 89–96% (Han et al. 2010). Separation of Aromatic Hydrocarbons The separation of various aromatics such as BTX (benzene, toluene, xylene) from other hydrocarbons is challenging because of close boiling points and also formation of azeotropes (Werner et al. 2010). Liquid–liquid extraction is useful for less than 75% aromatic content, and extractive distillation is most suitable for 65–90% aromatic content. For higher concentration of aromatics, azeotropic distillation is more favorable (Werner et al. 2010). These approaches have the drawbacks of mainly high energy requirements and back extraction. Ionic liquids

658

K. L. Wasewar

may overcome these issues and may be the potential solvent for the separation of aromatic hydrocarbons. The liquid–liquid extraction of a naphtha cracker feed having various aromatics were investigated using 1-butyl-4-methylpyridinium [PF6 ], [EMIM][EtSO4 ], and 1-butyl-4-methylpyridinium dicyanamide. The results were compared with commercial sulfolane. Almost 50% saving in cost were observed for this application (Meindersma et al. 2006). Better performance of ionic liquid-based separation for ternary hexane/benzene/[EMIM][NTf2 ] system was observed as compared to current process (Arce et al. 2007). Ionic liquids have high technical potential to extraction aromatics from hydrocarbon streams. Further modification in structure of ionic liquids is needed to achieve higher capacities and selectivities.

Fuel Desulfurization Sulfur compounds have significant environmental and health effects. Sulfur contents in various forms in different organic streams are an issue of environmental concern, and it is necessary to reduce to certain permissible level. Bo¨smann et al. (2001) employed the ionic liquid first time for the removal of sulfur compounds from liquid fuel. Later on different types of ionic liquids have been used for the removal of sulfur compounds from various streams (Dharaskar et al. 2013, 2014, 2015, 2019; Wasewar 2012, 2013). A critical review on the contribution of ionic liquids for desulfurization of liquid fuels has been presented in literature (Ibrahim et al. 2017). Metal Separations Globally, water pollution can affect ground water or surface water by industrial or domestic effluents which have been one of the most alerting and disturbing issues of environmental concern. Toxic and hazardous metals, metal ions, and their compounds can be separated or recovered by extraction using various solvents. In this method, metal-contaminated aqueous phase is allowed to mix with extractive phase, which is mostly immiscible with aqueous phase. Metals as pollutants are distributed in both the phases, and extractive or solvent phase is enriched selectively with the desired metal. These solvents are suffering with toxicity, flammability, and volatility; hence alternative ecofriendly solvent needs to be found due to environmental issues related to conventional organic solvents. Ionic liquids (ILs) are one of the possible and potential alternatives due to non-volatility and non-flammability to conventional toxic organic extractant in separation process (Khazalpour et al. 2020). Ionic liquids have been considered widely for the separation of metal ions. Trihexyl(tetradecyl)phosphonium cation-based ionic liquids have been employed for the separation of various metal ions. Cyphos IL-101 (tetradecyl(trihexyl)phosphonium chloride) in hexane and toluene was employed for the efficient removal of Au(III) and also Pd(II) from acidic solution (Cieszynska and Wisniewski 2010). The separation of Zn(II) from Fe(III) was performed by membrane having metal ion carrier as trihexyl(tetradecyl)phosphonium chloride (Cyphos IL101) or bis(2,4,4-trimethylphenyl)phosphinate (Cyphos IL104) (Baczyn’ska et al. 2018). The aqueous biphasic system having ionic liquid

24 Ionic Liquids: The Smart Materials in Process Industry

659

tetrabutylammonium bromide with (NH4 )2 SO4 was investigated for the removal of Cr (VI), and 97% recovery was obtained (Akama and Sali 2002).

CO2 Capture and Gas Separation Carbon dioxide is one of the key contributors to greenhouse gases (GHG), and its emission is increasing exponentially. Amine technology for the capture of carbon dioxide has certain drawbacks including cost, energy, etc. Other methods such as membrane, cryogenic, and biological are also of interest but not technically or economically feasible at large or industrial scales. In view of this, ionic liquids (ILs) are the one of the best alternatives for the conventional and other technologies. In recent years a lot of studies revealed that ionic liquids can be effectively used for the capture of carbon dioxide. The first-generation or conventional ionic liquid, 1-butyl-3-methylimidazolium hexafluorophosphate ([C4 mim][PF6 ]), was employed first time for the 0.6 mole fraction capture of CO2 at 40 ◦ C and 8 MPa (Blanchard et al. 1999). [hmpy][Tf2 N] was studied to understand the solubility of various gases, and the trend of SO2 > CO2 > C2 H4 > C2 H6 > CH4 > O2 > N2 was observed at 298 K, where carbon dioxide has higher solubility except sulfur dioxide (Anderson et al. 2007). Generally, conventional ILs are not able to capture CO2 efficiently; hence efficiency can be improved by functionalization of ionic liquids (Ramdin et al. 2012; Cui et al. 2016). Functionalized ionic liquids were obtained by performing the reaction of tetrabutylphosphonium hydroxide [P(C4 )4 ][OH] with amino acids (glycine, l-alanine, l-b-alanine, l-serine, and l-lysine). The absorption was observed in less than 100 min, and capacity was retained till four cycles (Yang et al. 2011). The conventional and functionalized ionic liquids have the drawbacks of technical complexity, high amount of energy demand, and loss of solvents in large quantity, and these can be overcome by supported ionic liquids. The ionic liquids as membrane materials having [Tf2 N], [CF3 SO3 ], chloride, and dicyanamide as anions, and 1-ethyl-3-methylimidazolium [Emim] and trihexyltetradecylphosphonium [THTDP] as cations were employed for the separation of CO2 from N2 and (CH4 ) (Scovazzo et al. 2004). The highest permeability of carbon dioxide was observed for [Emim][Tf2 N]. Carbon dioxide was selectively separated from helium by using [Hmim][Tf2 N]-based supported ionic liquid with higher permeability and selectivity as compared to most of the membranes (Ilconich et al. 2007). The exhaustive critical review on CO2 capture using ionic liquids has been presented in the literature (Zhang et al. 2013). Carbon dioxide was captured and removed using various conventional, task-specific, supported in membranes, and polymerized ionic liquids.

Reaction Reaction engineering is the heart of chemical process industries, where the performance of reaction and reactor decides the downstream processing and hence overall

660

K. L. Wasewar

production cost. Ionic liquids have promising and potential application in reaction as catalyst, solvents, and additives.

Catalyst and Additives Ionic liquids are capable to dissolve many transition metal complexes which are used as catalyst. Interaction of transition metal complexes with ILs provides reactivity, and it is essential to understand it for the selection of suitable ionic liquids (Werner et al. 2010). Wilkinson’s catalysts RhCl(TPP)3 (where TPP = triphenylphosphine) in [BMIM][BF4 ], [BMIM][PF6 ], and [BMIM][AlCl4 ] were used for hydrogenation of cyclohexene. Hydrogenation of cumene, chlorobenzene, benzene, ethylbenzene, and toluene were investigated using [H4 Ru4 (η6-C6 H6 )4 ][BF4 ]2 (Dyson et al. 1999, 2003; Werner et al. 2010). Ionic liquids can be a promising replacement for conventional catalysts in many reactions (Khazalpour et al. 2020). The exhaustive review of the application of ILs in reaction as catalysts and additives was presented (Rádai et al. 2018; Amarasekara 2016). Ionic liquids can be used as acid catalyst, basic catalyst, organocatalyst, and soluble support for various reactions. Ionic liquids have been extensively used in various kinds of reactions such as substitutions, alkylations, acylations, esterifications, oxidations, reductions, additions and cycloadditions, eliminations, condensations, halogenations, cross-couplings, Saucy–Marbet reactions, protections and deprotections, and others (Rádai et al. 2018). Triphenyl(propyl-3-sulfonyl)phosphonium cation and trifluoromethane sulfonate anion based ionic liquid [TPPSP][OTf] has been employed to carry Hantzsch synthesis of polyhydroquinolines and acridines by the condensation reaction of b-ketoesters, 1,3-diones, ammonium acetate, aldehydes, and aniline derivatives where recycle of ionic liquid is easily possible (Vahdat et al. 2016). Tributyl(3sulfopropyl)phosphonium hydrogen sulfate (TBSPHS) ionic liquid was used as catalyst to synthesize 1,5-dihydro-2H-pyrrol-2-one derivatives to perform the reaction among aromatic aldehydes, pyruvic acid, and anilines (Yarie et al. 2018). In the presence of ionic liquid catalyst/solvent (4-sulfobutyl)tris(4-sulfophenyl) phosphonium hydrogen sulfate, the multicomponent synthesis was performed for 12-aryl-8,9,10,12-tetrahydrobenzo[a]xanthen-11-one derivatives (Janardhan et al. 2012). There are numerous examples ionic liquids used as catalysts (Khazalpour et al. 2020). Various supported ionic liquids as catalysts have been reviewed in literature (Romanovsky and Tarkhanova 2017). Supported ionic liquids were employed as catalyst for different reactions such as oxidation of sulfur-containing compounds, phenol oxidation, synthesis of metal nanoparticles on the surface, reactions of haloalkanes, Heck reactions, Suzuki reaction, hydroamination of unsaturated compounds, Sandmeyer reaction, isomerization of hydrocarbons, and gas phase alkene hydrogenation reactions. Solvents for Synthesis and Catalyst Applications The ionic liquids are promising alternatives to organic solvents for synthesis and catalyst applications. The critical review on room-temperature ionic liquids has been presented for synthesis and catalysis (Hallett and Welton 2011). ILs

24 Ionic Liquids: The Smart Materials in Process Industry

661

can be used as aedium for various kinds of reactions including electron transfer reactions, transition metal catalyzed reactions, microwave assistant, acid-catalyzed reactions, substitutions, addition, base or nucleophilic reactions, elimination, acid base reactions, and more.

Analytical and Chromatography Analytical and chromatographic applications are important in many industries including pharmaceutical, drug, chemical, food, and other allied industries. Ionic liquids can be a promising alternative for conventional solvents, support, mobile phase, additives, etc., for analytical and chromatography.

Analytical Chemistry Ionic liquids can be used as the stationary phase for the separation of various organic compounds. The type of anions and cations used and their characteristics have a significant effect on selectivity and solubilization of compounds. Due to a few specific and tunable properties of ionic liquids, these can be employed in separation and analytical science and chemistry (Khazalpour et al. 2020). Due to some environmental concerns, ionic liquids can be alternative and have the potential environment friendly in the area of analytical chemistry which includes chromatography, spectrometry, sensing, isolation, extraction, electro-analysis, micro-extraction, analysis of metal ions, bioanalytical, environmental, gas chromatography, high-performance liquid chromatography, capillary electrophoresis, etc. (Singh and Savoy 2020). Tributylethylphosphonium diethylphosphate [P2,4,4,4 ][DEP] was used as gas chromatography stationary phase for various organic solutes and water (Kro’likowska and Orawiec 2016). In reversed-phase high-performance liquid chromatography, quaternary ammonium- and phosphonium-based ionic liquids were considered for analysis. Trioctyl(3/4-vinylbenzyl)phosphonium chloride was employed as modifier in capillary columns of silica-based monolithic (Moravcova´ et al. 2018). Liquid Chromatography Liquid chromatography is mostly used for analytical separations. Liquid chromatography can be operated for separation in various modes: reversed phase, normal phase, hydrophilic interaction, ion pair, ion exchange, etc. (Huang et al. 2013). Mostly, reversed-phase liquid chromatography (RPLC) separation has been used (more than 90%). For the analytical separation and analysis of ionizable or polar compounds, it has the problem of band broadening, band tailing, asymmetric peak, irreproducible retention time, and low efficiency. These problems can be overcome by using more adaptive adsorbents as stationary phase and also by addition of solutes to mobile phase. ILs can be used to resolve the above issues, and it can be the promising and potential substances for liquid chromatography (Huang et al. 2013). The addition of IL in mobile phase as additives shows improvement in separation efficiency, and also it does not have any influence on the pH of the mobile phase.

662

K. L. Wasewar

The few different ionic liquids used as additives in mobile phase are [C4 mim][BF4 ], [C6 mim][BF4 ], [C8 mim][BF4 ], for the targeted analysis of Ephedrines, b-Lactam antibiotics, heterocyclic aromatic amines, Catecholamines, Cytosine, thymine, adenine, 6-chlorouracil, L-his, L-tyr, L-phe, and DL-try, N-carbobenzyloxy-D-phe and D-try, etc. Usage of ILs as additives in mobile phase resulted in decreased band tailing and broadening, improved resolution and retention, and good separation with sharp peaks, more favorable than a common additive. The ILs employed to modify silica stationary phase are [Cnhim][BF4 ], [C4 mim][Br], [C1 pim][Br], [C4 pim][Br], etc., for the analysis of ephedrines, tanshinone, peptide aromatic carboxylic acids, etc., which resulted in the reduction of retention time and fast separation (Chitta et al. 2010).

Gas Chromatography Ionic liquids have been considered for gas chromatography stationary phase due to their low volatility and high thermal stability (Berthod et al. 2008). At early stage, ILs, ethylpyridinium bromide, and ethylammonium nitrate were investigated for characteristics as a stationary phase. Better selectivity of columns with these stationary phases was obtained for H-bond-capable and polar solute (Pacholec and Poole 1983). Enhanced efficiency and selectivity were observed for non-polar and polar solutes using imidazolium-based ILs. The ILs, methoxyphenyl[MIM][TfO], and [BMIM][TfO] were investigated as a stationary phase in gas chromatography, and better performance was observed with increased thermal stability and also better film formation. The gas chromatography capillary columns were prepared by using solvent properties of ionic liquids ([BMIM][Cl]) having cyclodextrins (di- and per-methylated) as chiral selectors. Capillary Electrophoresis Salts as support electrolytes are used in capillary electrophoresis. Ionic liquids (viscous and conductive salts) cannot be directly used as a solvent in capillaries because of high current, low voltage, and electric field. Hence ILs can be applied on walls of capillary as coatings. Ionic liquids are used in capillary electrophoresis as wall covalent coating, as background electrolyte and dynamic wall coating in aqueous capillary electrophoresis, as background electrolytes in nonaqueous capillary electrophoresis, micellar electrokinetic chromatography with ionic liquids, and use of chiral ionic liquids (Berthod et al. 2008). In capillary electrophoresis, ionic liquids are used to reduce the interaction between analytes and the capillary wall. Propyl methyl imidazolium chloride ([PMIM][Cl]) was coated by covalent bond on silica capillary for the separation of sildenafil from its metabolites (Qin and Li 2002). Also, the separation of DNA fragments, and alkyl phosphonic acids and esters was performed by using modified similar kind of capillaries (Qin and Li 2003). Ionic liquids are also used in capillary electrophoresis as background electrolytes which are adsorbed on capillary walls due to hydrophobic nature of cation of ionic liquids. [EMIMBF4 ] and [BMIMBF4 ] were employed for the separation and determination of monohalogenated phenols, anthraquinones, flavonoids, nicotinic acids, and some bioactive flavone derivatives.

24 Ionic Liquids: The Smart Materials in Process Industry

663

For nonaqueous capillary electrophoresis, the liquid phase has very less polarity than water. Ionic liquids were also used for the separation of phenols, polyphenols, aromatic acids, and water-insoluble dyes.

Electro Processes Various processes and operations can be enhanced by using different aids including ultrasound, magnetic field, and electric field. Ionic liquids have application in various electro-driven processes such as gas sensor and analytical chemistry.

Electrochemical Gas Sensor Ionic liquids have wide electrochemical window and good stability and conductivity, which are essential for electrochemical sensor application. Various electrochemical techniques like voltammetry, chronoamperometry, impedometry, and condcutometry can be used to make electrochemical sensors such as chemosensors, actuators, biosensors, and gas sensors (Pau et al. 2020). The electrochemical sensors for various gases mainly CO2 and other gases have been studied using different ionic liquids such as 1-butyl-3-methyl imidazolium acetate [C4 mim][Ac], 1-ethyl-3-methylimidazolium bis (trifluoromethylsulfonyl) imide [C2 mim][NTf2 ], methylpyridinium bis(trifluoromethylsulfonyl)imide [C4 mpyrr][NTf2 ], 1-butyl-3methylimidazolium bis(trifluoromethylsulfonyl)imide [C4 mim][NTf2 ], 1-butyl-3methylimidazolium tetrafluoroborate [C4 mim][BF4 ], 1-butyl-3-methylimidazolium trifluoromethanesulfonate [C4 mim][OTf], 1-butyl-3-methylimidazolium hexafluorophosphate [C4 mim][PF6 ], and 1-hexyl-3-methylimidazolium chloride [C6 mim]Cl. Electrochemistry ILs have some specific properties: high thermal stability, good conducting electrolytes, a wide range of electrochemical potential, high viscosity, a wide range of solids to liquids, tunable solubility, etc., which make them suitable for various applications of electrochemistry (Singh and Savoy 2020). Properties of IL and electrode play a significant role on the performance of electrochemical devices where basically three associations are considered: conductivity, viscosity, and electrochemical potential of ILs. The phosphonium-based ionic liquids have been considered more for electrochemical applications because of high thermal and chemical stabilities, having low viscosities relative to other solvents, and high electrical and thermal conductivities in contrast to other ionic liquids (Khazalpour et al. 2020). Electrophoresis Ionic liquids with lower UV (ultraviolet) absorption at a short wavelength having higher thermal stability are considered for mobile and stationary phases in electrophoresis which significantly improve the resolution and separation. In capillary electrophoresis, ionic liquids may be considered as exceptional additives

664

K. L. Wasewar

in electrolytes which can change electro-osmotic flow and hence improvement in charged ion separation (Khazalpour et al. 2020). Various ionic liquids, tributyl(tetradecyl)phosphonium chloride [P14444 ]Cl, octyltributylphosphonium chloride [P8444 ][Cl], tributyl(tetradecyl)phosphonium chloride and acetate, tetrabutylphosphonium chloride ([P4444 ]Cl, tributyl(hexadecyl)phosphonium bromide ([P16444 ]Br, and triisobutylmethylphosphonium tosylate ([P1444 ][OTs] were used in capillary electrophoresis, micellar electrokinetic chromatography, capillary electrokinetic chromatography, and electrokinetic capillary chromatography.

Other Applications Biodiesel Production The oil extracted from various animal and vegetable sources are one of the nonpolluting alternative energy sources. These oils cannot be used directly in existing combustion engines; hence oils must be converted to biofuels. These oils mainly contain triglycerides and can be converted to biofuels by pyrolysis (cracking), microemulsification, and transesterification. Transesterification is the most popular process for the conversion of oils into biodiesels. The catalyst plays an important role in transesterification reaction and has certain irregularities (Andreani and Rocha 2012). Ionic liquids are the new generation catalyst with potential applications in various segments of chemical industry, and also it can resolve the issues of catalyst in transesterification. Ionic liquid with 1-n-butyl-3-methylimidazolium cation was employed for biodiesel production by lipase-catalyzed transesterification with methyl acetate and [BMIM][PF6 ] as co-solvent. The use of ionic liquid improved the lipase activity and stability, and also it avoids enzyme deactivation. Twenty three different ionic liquids were demonstrated for the production of biodiesel by methanolysis of soybean by lipase, and the maximum yield of 80% at 50 ◦ C in 12 hrs was obtained with 1-ethyl3-methylimidazolium trifluoromethanosulfonate ([Emim][CF3 SO3 ]) and methanol (Ha et al. 2007). The Brønsted acidic ionic liquids having 1-n-butyl-3-methylimidazolium or 1-butylpyridinium cations and the ionic liquid 1-(4-sulfonic acid) butylpyridinium hydrogensulfate ([BSPy][HSO4 ] were investigated for transesterification of cottonseed oil with methanol for biodiesel production with 92% yield in 5 hrs at 170 ◦ C (Wu et al. 2007). Soybean oil was used for the production of biodiesel in the presence of triethylammonium chloroaluminate ([Et3 NH]Cl/AlCl3 ) with 98.5% conversion in 9 hrs 70 ◦ C (Liang et al. 2009). The ionic liquid has several advantages in biodiesel production such as easy in operation, low catalyst (IL) cost, high yields of desired product, and reusability (Andreani and Rocha 2012). Space Technology The space age was begun in the late 1950s, and hence there has been growth of space industry with innovations in various space technologies. The major areas for

24 Ionic Liquids: The Smart Materials in Process Industry

665

innovations are rocketry, spacecraft design, and satellite technology having various engineering and science domains which include computer technology, material science, robotics, analytical technology, and process intensification (Nancarrow and Mohammed 2017). The major drawback of ionic liquids is the high cost, which restricts its application in bulk volume; hence high value application to be considered and that is space technology (Nancarrow and Mohammed 2017). In the last few years, ionic liquids have been used in space technology, and it is the novel development for the application of ionic liquids. Ionic liquids have been employed in space technology in electrospray thrusters, lubricants, hypergolic fluids, and highperformance composite materials. Electrospray/colloid thrusters (electric propulsion engines) provide electrostatic acceleration of liquid propellants. Electrospray has high efficiency and specific impulses but very low thrust as compared to chemical system (Carroll 2015). Various ionic liquids were studied as electrospray thruster propellants in combination of various chemical compounds or alone. The investigated ionic liquids are [C4 mim][GaCl4 ], [C2 mim][NTf2 ], [C2 mim][C(CN)3 ], [C2 mim][BF4 ], and [C2 mim][N(CN)2 ]. Ionic liquids can be used as hypergolic fluids which are highenergy propellants. These are used in spacecraft maneuvering systems and rocket engines having a fuel and an oxidant. Ionic liquids investigated for hypergolic fluids are C4 mim][N(CN)2 ], 1-methyl-3-propargyl-imidazolium dicyanamide, and [PC1im][N(CN)2 ]. In aerospace technology, lubricants have very specific properties including thermal stability for a wide range of parameters, good heat transfer properties, high resistance to radiation, and high resistance to oxidation. Ionic liquids studied for this application are [C6 mim][NTf2 ] and [C6 mim][BF4 ]. Many ionic liquids are stable for the wide range of temperature, and hence C8 mim][PF6 ] and [C4 mim][BF4 ] were investigated as a promising fluid for heat transfer fluid in space applications. The phase change materials suitable for space thermal control should have high phase transition latent heat, high specific heat, high thermal conductivity, low vapor pressure, small change in volume in the phase transition, high density, long-term stability, non-explosive, non-toxic, nonflammable, non-corrosive, chemically stable, and relatively low cost. Ionic liquids with suitable above-said criteria for phase change material in space application, [C16mmim ]Br and [C16 mim]Br were studied (Zhua et al. 2009).

Starch Chemistry Starch is found mainly in plants as energy source having wide application in chemistry, paper, material, pharmaceutical, fermentation, and food industries. Starch is one of the promising renewable materials for various processing with lowprice and biodegradability. Due to biodegradability of starch, it can be one of the best replacements to many synthetic polymers. Organic solvents are required for dissolution and other processing of starch. Ionic liquids can be alternative to organic solvents and found it as an excellent solvent for the derivatization, dissolution, and plasticization of starch (Ren et al. 2020). Different ionic liquids [Bmim][Cl], [Amim][Cl], [Emim][Cl], [Bmim][Br], [Hexmim][HCOO], [Emim][Me2 PO4 ], [Mmim][(MeO)HPO2 ] have been used for

666

K. L. Wasewar

the dissolution, plasticization, and derivatization of starch of maize, potato, rice, wheat, barley, potato, etc.

Surfactant Chemistry Ionic liquids have also been considered in surfactant industry as modifiers (Khazalpour et al. 2020). Ionic liquids based on imidazolium and phosphonium were employed as surfactant modifier agents in cationic exchange reaction for making organic modification in synthetic clays, which have better thermal and mechanical strength and stability as compared to conventional quaternary ammonium salts. Nanoparticle Stabilizers Ionic liquids have been used in the synthesis of metal nanoparticles as stabilizers. Various nanoparticles (titanium oxide, AgI, Pd, copper oxide) have been synthesized using different ionic liquids ([P666,14 ][NTf2 ], [P666,14 ]Cl, tri-t-butyl(decyl)phosphonium tetrafluoroborate). Most of these nanoparticles stabilized in ionic liquids have been used as catalysts in various reactions (Khazalpour et al. 2020). Polymer Science It was observed first time that the imidazolium, ammonium, and phosphonium cation-based ionic liquids can be considered as plasticizers for preparing the polyvinyl chloride to obtain flexible-grade polyvinyl chloride (PVC). Lowering of glass transition temperature was observed for 20% plasticization using ionic liquids with good resistance for leaching and migration and better thermodynamic compatibility (Khazalpour et al. 2020). Various ionic liquids based on imidazolium and phosphonium cations have been investigated as catalyst additives for the photopolymerization of polyethylene glycol-400 dimethacrylate (PEGDM) and triethylene glycol dimethacrylate (TEGDM). The polymerization of epoxynetworked polymeric materials was performed using reactive additive as tributyl (ethyl) phosphonium diethyl phosphate and trihexyl (tetradecyl) phosphonium bis 2,4,4-(trimethyl pentyl)-phosphinate (Nguyen et al. 2014). Drug Delivery Ionic liquids have potential to solve many critical problems in pharmaceutical industry. Ionic liquids can be employed in pharmaceutical applications as solvents, alternative media, to solve polymorphism problems, novel ingredient, and in drug delivery. There are many challenges in the pharmaceutical sector especially in drug delivery systems where skin permeation, stability, drug solubility, and their administration are of major concern (Caparica et al. 2018). Ionic liquids have certain properties which enable them to be suitable to increase the solubility and loading of drug, and permeation of relevant drug through suitable delivery practice. The exhaustive review on the application of ionic liquids in drug delivery system is available in literature (Omar 2016). Two ionic liquids, [Hmim][Cl] and [Bmim][PF6 ], were used to prepare oil– water phase by incorporating in stable emulsion which observed the antimicrobial

24 Ionic Liquids: The Smart Materials in Process Industry

667

activities at more than 5% ionic liquids with higher penetration in skin (Dobler et al. 2012). Further, antimicrobial properties and penetration of drug in the skin of emulsion gel were investigated for [HPyr][Cl], [CDHP], and [Emim][EtSO4 ] in 4-hydroxybenzoic acid propyl ester, caffeine, and testosterone (Santos et al. 2017). Various microemulsion gels were prepared using [Bmim][Br], [Bmim][PF6 ], [HOEIM] [Cl], and [Bmim][C12 SO3 ] in different drugs including 5-fluorouracil, etodolac, tween 80, span 20, and dencichine to investigate the ex vivo permeation (Goindi et al. 2015; Wang et al. 2018). Ionic liquids may improve the penetration by many folds, and hence it has promising application in drug delivery systems.

Lubricants Synthetic lubricants are high-molecular-weight complex chemical compounds having specific application in engine, gear, and other machineries as lubricants. The performance of the lubricant depends on the base oil and other additives having specific designed and targeted properties, and hence lubricants have significant role in machineries. Additives are mainly responsible for enhancing the specific performance including fire resistance, antioxidant, extreme pressure, dispersant, etc. Numerous additives are considered to increase the performance of lubricants, and ionic liquids may be new and promising additives. Ionic liquid tetraalkylphosphonium tetrafluoroborate was investigated as lubricants on still/Al (Kajdas 1994). During the slide of surfaces, the complex tribochemical reactions taken place (Khazalpour et al. 2020). The double bilayer type of the structure of the ionic liquids presented good lubricity as similar to graphite and molybdenum disulfide (Liu et al. 2006). The tribological properties of [N12 ,H,H,H][Cl], 1-methyl-3-hexylimidazolium tetrafluoroborate and 1-ethyl-3hexylimidazolium tetrafluoroborate, and other ionic liquids, were investigated using SRV ball on disk configuration. Batteries Due to the diminishing nature of fossil fuel, it is essential to look for the cleaner alternative energy sources, and rechargeable batteries specially lithium ion batteries (LIBs) are considered as one of the most promising candidates since 1991 of their appearance in the market (Yang et al. 2020). It has wide application in various electrical and electronic devices including aerospace and so on. The key component in such type of batteries is electrolyte. The electrolyte is responsible for fast ion transport and satisfactory electrochemical and chemical stability. There are few safety issues due to thermal instability, flammability, and leakage possibility (Yang et al. 2020). The conventional organic liquid electrolytes are flammable and corrosive having serious safety issues, and also they are not able to stop dendrite growth, which may cause short circuiting. Conventional flammable liquid electrolytes are not suitable for next-generation lithium metal batteries with stable performance, high energy density, and excellent safety (Yang et al. 2020). The solid electrolytes instead of liquid electrolyte are expected to resolve safety issues by reducing side reaction interface and hence increasing the life of the battery. Solid

668

K. L. Wasewar

polymer electrolytes have advantages of processability, flexibility, and adhesion with electrodes. The hybrid electrolytes (HEs) are electrolytes with over two components, and it is expected to improve room-temperature ionic conductivity, mechanical strength, and electrochemical properties. Ionic liquids are the combination of cation and ions with tunable properties having ion conductivity, non-flammability, good thermal stability, wide electrochemical window (up to 5–6 V vs. Li+/Li), and non-volatility. ILs have been used in many electrochemical application includes batteries, supercapacitors, dye-sensitized solar cells, and fuel cells to increase the performance. ILs along with polymers and polymeric ionic liquids used as solid electrolytes in lithium ion batteries have been reported (Yang et al. 2020). ILs used are 1-ethyl-3-methylimidazolium bis(fluorosulfonyl) imide (EMimFSI), 1-ethyl-3methylimidazoliumtriluoromethanesufonate (EMITFSI), tetrabutylphosphonium 2hydroxypyridine (TBPHP), 1-methyl-1-propylpiperidinium bis(trifluoromethanesul fonyl)imide (SiO2PPTFSI), 1-butyl-3-methylimidazolium bis(trifluoromethylsulfon yl)imide, N-methyl-N-butyl pyrrolidinium bis(trifluoromethansulfonyl)imide, N,Ndiethyl-N-(2-methoxyethyl)-N-methylammonium bis(trifluoromethylsulfonyl)imide, N-butyl imidazole, etc., with PEO, PVDF, HFE, etc. (Yang et al. 2020).

Biological Aid Solvents are used at many stages in preparing the drugs. Ionic liquids can be one of the best suited alternatives to conventional organic solvents, and ionic liquids have been getting more and more attention as reaction media in the last few years. Ionic liquids have exceptional and tunable chemical and physical properties and also have high biological activity. Because of such kind of outstanding and attracting properties medicinal scientists, ecologists, and biochemists are significantly more interested to explore the applications of ionic liquids as biological aid (Singh and Savoy 2020). There have been antimicrobial and cytotoxic biological activities of ionic liquids having applications in drug delivery and drug synthesis. Phosphonium- and ammonium-based ionic liquids were explored to investigate the cytotoxicity and anti-tumor activity using human tumor cell of NCI 60 lines. The therapeutic applications can be controlled by tuning the cation and anion combination and also with different substitutes. The ampicillin anions combined with cations (ammonium, imidazolium, phosphonium, and pyridinium) were investigated for anti-tumor activity. The used ionic liquids shows antiproliferative effects against various tumor cell lines. The mono-, di-, and trisubstituted triphenylamine-based phosphonium ionic liquids were found to have antibacterial activity for Grampositive and Gram-negative bacteria (Brunel et al. 2018). Engineering The application of ionic liquids in engineering field was reported almost three decades ago mainly on extraction (Singh and Savoy 2020). Ionic liquids can be employed in many fields of engineering due to their unique properties, which may be resulted in molecular hydrogen bonding, Coulombic, and van der Waals interactions.

24 Ionic Liquids: The Smart Materials in Process Industry

669

Rare Earth Elements Extraction Ionic liquids can also be employed for the extraction and processing of minerals and metals. This is called as ion metallurgy. Ionic liquids as designer solvents can be applied in numerous processing steps for extractive metallurgy of rare earth elements. Ionic liquids are alternative to conventional solvents for efficient and ecofriendly processes in metallurgical industry for dissolution-separation and production of rare earth elements. Ionic liquid (HbetTf2 N) was used for Bauxite residue leaching from Rödberg and Greece ore from Norway to extract rare earth elements (EURARE 2020).

Conclusion Ionic liquids have many promising and tunable properties which make them alternative in numerous applications. Ionic liquids are one of the potential substitutes to various conventional volatile organic solvents, mineral bases, mineral acids, solid acids, and many other applications. Ionic liquids may be a popular option because of non-corrosive, ecofriendly nature as green compound used in many organic transformations and processes in process industry. The typical applications of ionic liquids are mentioned in Fig. 8. Ionic liquids have flexibility to modify chemical, thermal, physical, and biological probable properties, which could be obtained by tuning various combinations of cations and anions. The interest of academicians, scientists, engineers, and industrialists has been toward ionic liquids in the last few decades. There have been many suppliers for ionic liquids for commercial as well as research purposes, which make them more prospective green solvent for further applications. The few suppliers for ionic liquids are Merck KGaA/EMD Chem., Cytec, DuPont, Solvent Innovation, Accelergy, Sigma-Aldrich, Nippon Gohsei, Chemada, Solchemar, Kanto Chemical Co., ACROS, IoLiTec, Scionix, SACHEM, BASF, etc. (Plechkova and Seddon 2008). Due to tunable and designer nature of ionic liquids, these chemicals have been used in a wide range of applications including solvents, catalyst, medium, etc. Furthermore, ionic liquids have been explored to know the scope and potential for wide applications in process industries in sustainable ways in various fields such as solvents, energy storage, analytical, catalysts, electrodes, polymers, and medical aids. Due to negligible vapor pressure of ionic liquids, they are not contributing to air pollution and hence it is a green alternative solvent to conventional organic solvents, which are generally flammable, volatile, and toxic. As a result of this, numerous investigations have been performed on the application of ionic liquids as novel green solvents to replace well-established conventional solvents for specific applications. There is completely no question that the exceptional properties of ionic liquids present great potential to make improvement in various existing engineering applications or to develop attractive novel and new ones. Based on the wide application of ionic liquids in process industry, ionic liquid offers a vast prospective for more safer, more efficient, more better, or completely

670

K. L. Wasewar

Fig. 8 Few applications of ionic liquids

new approaches for process industry and also for other engineering and industrial applications. But much more is needed to be achieved and developed to realize all promising opportunities of ionic liquid applications. Designed ionic liquids for specific application having optimized structures, better and useful physicochemical and engineering data, superior theoretical and experimental prediction tools, and the development of dedicated process units and devices are needed.

Few Important Websites 1. Ionic Liquids – Organic Chemistry Portal https://www.organic-chemistry.org/topics/ionic-liquids.shtm 2. Ionic liquids: a brief history | SpringerLink https://link.springer.com/article/10.1007/s12551-018-0419-2 3. Uses of ionic liquids | Koei Chemical Co., Ltd https://www.koeichem.com/en/en_product/ion/use.html 4. IoLiTec · Ionic Liquids & more

24 Ionic Liquids: The Smart Materials in Process Industry

671

https://iolitec.de/en/products/ionic_liquids?gclid=CjwKCAjwy42FBhB2EiwAJ Y0yQiwymSs1_xGZiltYxaF4Vk28MD4d8BoCcizVg6N0vUMb2oZNLKRGM RoCKdsQAvD_BwE 5. Ionic Liquids Market Analysis | Recent Market Developments ... https://www.marketsandmarkets.com/Market-Reports/ionic-liquid-market163716481.html

References Akama Y, Sali A (2002) Extraction mechanism of Cr (VI) on the aqueous two-phase system of tetrabutylammonium bromide and (NH(4))(2)SO(4) mixture. Talanta 57:681–686 Amarasekara AS (2016) Acidic ionic liquids. Chem Rev 116:6133–6183 Anderson JL, Dixon JK, Brennecke JF (2007) Solubility of CO2 , CH4 , C2 H6 , C2 H4 , O2 and N2 in hexyl-3-methylpyridinium bis(trifluoromethylsulfonyl)imide comparison to other ionic liquids. Acc Chem Res 40(11):1208–1216 Andreani L, Rocha JD (2012) Use of ionic liquids in biodiesel production: a review. Braz J Chem Eng 29(1):1–13 Angell CA, Ansari Y, Zhao Z (2012) Ionic liquids: past, present and future. Faraday Discuss 154: 9–27 Arce A, Earle M, Rodriguez H, Seddon K (2007) Separation of aromatic hydrocarbons from alkanes using the ionic liquid 1-ethyl-3-methylimidazolium bis{(trifluoromethyl) sulfonyl}amide. Green Chem 1:70–74 Baczyn’ska M, Słomka Z, Rzelewska M, Waszak M, Nowicki M, Regel-Rosocka M (2018) Characterization of polymer inclusion membranes (PIM) containing phosphonium ionic liquids and their application for separation of Zn(II) from Fe(III). J Chem Technol Biotechnol 93(6):1767–1777 Berthod A, Ruiz-A’ngel MJ, Carda-Broch S (2008) Ionic liquids in separation techniques. J Chromatogr A 1184:6–18 Berton P, Monasterio RP, Wuilloud RG (2012) Selective extraction and determination of vitamin B12 in urine by ionic liquid-based aqueous two-phase system prior to high-performance liquid chromatography. Talanta 97:521–526 Blanchard LA, Hancu D, Beckman EJ, Brennecke JF (1999) Green processing using ionic liquids and CO2 . Nature 399:28–29 Bo¨smann A, Datsevich L, Jess A, Lauter A, Schmitz C, Wasserscheid P (2001) Chem Commun 23:2494 Brunel F, Lautard C, di Giorgio C, Garzino F, Raimundo J-M, Bolla J-M, Camplo M (2018) Antibacterial activities of mono-, di- and tri-substituted triphenylamine-based phosphonium ionic liquids. Bioorg Med Chem Lett 28(5):926–929 Caparica R, Júlio A, Mota JP, Rosado C, Almeida de TS (2018) Applicability of ionic liquids in topical drug delivery systems: a mini review. J Pharm Clin Res 4(5) JPCR.MS.ID. 555649 Carroll M (2015) Living among giants: exploring and settling the outer solar system, 1st edn. Springer, Littleton Chitta KR, Van Meter DS, Stalcup AM (2010) Separation of peptides by HPLC using a surfaceconfined ionic liquid stationary phase. Anal Bioanal Chem 396:775–781 Cieszynska A, Wisniewski M (2010) Extraction of palladium(II) from chloride solutions with Cyphos® IL 101/toluene mixtures as novel extractant. Sep Purif Technol 73(2):202–207 Cui G, Wang J, Zhang S (2016) Active chemisorption sites in functionalized ionic liquids for carbon capture. Chem Soc Rev 45:4307–4339 Dharaskar S, Sillanpaa M, Wasewar K, Walvekar R (2019) Feasibility study of phosphonium ionic liquids as efficient solvent for sulfur extraction from liquid fuels. Am Inst Phys (AIP) Conf Proc 2137(1)

672

K. L. Wasewar

Dharaskar SA, Wasewar KL, Varma MN, Shende DZ (2013) Extractive deep desulfurization of liquid fuels using Lewis-based ionic liquids. J Energy:1–4 Dharaskar SA, Wasewar KL, Varma MN, Shende DZ, Yoo CK (2014) Extractive desulfurization of liquid fuels by energy efficient green Thiazolium based ionic liquids. Ind Eng Chem Res 53(51):19845–19854 Dharaskar SA, Wasewar KL, Varma MN, Shende DZ (2015) Imidazolium ionic liquid as energy efficient solvent for desulfurization of liquid fuel. Sep Purif Technol 155:101–109 Dreyer S, Salim P, Kragl U (2009) Driving forces of protein partitioning in an ionic liquid-based aqueous two-phase system. Biochem Eng J 46:176–185 Dyson PJ, Ellis DJ, Henderson W, Laurenczy G (2003) A comparison of ruthenium-catalysed arene hydrogenation reactions in water and 1-alkyl-3-methylimidazolium tetrafluoroborate ionic liquids. Adv Synth Catal 345:216–221 Dyson PJ, Ellis DJ, Welton T, Parker DG (1999) Arene hydrogenation in a room-temperature ionic liquid using a ruthenium cluster catalyst. Chem Commun 1999:25–26 EURARE. www.eurare.eu/technologies/ionicLiquidExtraction.html. Referred on 27 October 2020 Flieger J, Czajkowska-Zelazko A (2015) Aqueous two phase system based on ionic liquid for isolation of quinine from human plasma sample. Food Chem 166:150–157 Freire MG, Neves CMSS, Marrucho IM, Lopes JNC, Rebelob LPN, Coutinho JAP (2010) Highperformance extraction of alkaloids using aqueous two- phase systems with ionic liquids. Green Chem 12:1715–1718 Goindi S, Kaur R, Kaur R (2015) An ionic liquid-in-water microemulsion as a potential carrier for topical delivery of poorly water soluble drug: Development, ex-vivo and in-vivo evaluation. Int J Pharm 495. Elsevier B.V Ha SH, Lan MN, Lee SH, Hwang SM, Koo Y-M (2007) Lipase-catalyzed biodiesel production from soybean oil in ionic liquids. Enzym Microb Technol 41(4):480–483 Hallett JP, Welton T (2011) Room-temperature ionic liquids: solvents for synthesis and catalysis. Chem Rev 111(2):3508–3576 Han J, Wang Y, Kang WB, Li CX, Yan XS, Pan JM et al (2010) Phase equilibrium and macrolide antibiotics partitioning in real water samples using a two-phase system composed of the ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate and an aqueous solution of an inorganic salt. Microchim Acta 169:15–22 He C, Li S, Liu H, Li K, Liu F (2005) Extraction of testosterone and epitestosterone in human urine using aqueous two-phase systems of ionic liquid and salt. J Chromatogr A 1082:143–149 Huang YI, Yao S, Song H (2013) Application of ionic liquids in liquid chromatography and electrodriven separation. J Chromatogr Sci 51:739–752 Ibrahim MH, Hayyan M, Hashim MA, Hayyan A (2017) The role of ionic liquids in desulfurization of fuels: a review. 76:1534–1549 Ilconich J, Myers C, Pennline H (2007) Experimental investigation of the permeability and selectivity of supported ionic liquid membranes for CO2/He separation at temperatures up to 125OC. J Membr Sci 298(1–2):41–47. https://doi.org/10.1016/j.memsci.2007.03.056 Janardhan B, Vijaya Laxmi S, Rajitha B (2012) J Chem Pharm Res 4:519 Kajdas C (1994) Importance of anionic reactive intermediates for lubricant component reactions with friction surfaces. Lubr Sci 6(3):203–228 Khazalpour S, Yarie M, Kianpour E, Amani A, Asadabadi S, Seyf JY, Rezaeivala M, Azizian S, Zolfigol MA (2020) Applications of phosphonium-based ionic liquids in chemical processes. J Iran Chem Soc 17:1775–1917 Kro’likowska M, Orawiec M (2016) Activity Coefficients at Infinite Dilution of Organic Solutes and Water in Tributylethylphosphonium Diethylphosphate Using Gas-Liquid Chromatography: Thermodynamic Properties of Mixtures Containing Ionic Liquids. J Chem Eng Data 61(5):1793–1802 Liang X, Gong G, Wu H, Yang J (2009) Highly efficient procedure for the synthesis of biodiesel from soybean oil using chloroaluminate ionic liquid as catalyst. Fuel 88(4):613–616 Liu X, Zhou F, Liang Y, Liu W (2006) Tribological performance of phosphonium based ionic liquids for an aluminum-on-steel system and opinions on lubrication mechanism. Wear 261(110):1174–1179

24 Ionic Liquids: The Smart Materials in Process Industry

673

Lui Q, Xuesheng H, Wang Y, Yang P, Xia H, Yu J et al (2005) Extraction of penicillin G by aqueous two-phase system of [Bmim]BF4/NaH2PO4. Chin Sci Bull 50:1582–1585 Marques CFC, Mouraõ T, Neves CMSS, Lima A´S, Boal-Palheiros I, Coutinho JAP, Freire MG (2013) Aqueous biphasic systems composed of ionic liquids and sodium carbonate as enhanced routes for the extraction of tetracycline. Biotechnol Prog 29(3):645–654 Marta´k JSˇ, Schlosser S Vlcˇkova´ (2008) J Membr Sci 318:298 Meindersma G, Podt A, de Haan A (2006) Ternary liquid- liquid equilibria for mixtures of an aromatic + an aliphatic hydrocarbon + 4-methyl-N-butylpyridinium tetrafluoroborate. J Chem Eng Data 5:1814–1819 Moravcova’ D Planeta J, King AWT, Wiedmer SK (2018) Immobilization of a phosphonium ionic liquid on a silica monolith forhydrophilic interaction chromatography. J Chromatogr A 1552:53–59 Nancarrow P, Mohammed H (2017) Ionic liquids in space technology – current and future trends. ChemBioEng Rev 4(2):2. 1–15 Nguyen TKL, Livi S, Pruvost S, Soares BG, Duchet-Rumeau J (2014) Ionic liquids as reactive additives for the preparation and modification of epoxy networks. J Polym Sci Part A Polym Chem 52(24):3463–3471 Omar M (2016) A review of ionic liquids for advance in drug delivery: theory and pharmaceutical implementation. UK J Pharm Biosci 4(1):41–44 Pacholec F, Poole CF (1983) Stationary phase properties of the organic molten salt ethylpyridinium bromide in gas chromatograph. Chromatographia 17:370–374 Paul A, Muthukumar S, Prasad S (2020) Review—room-temperature ionic liquids for electrochemical application with special focus on gas sensors. J Electrochem Soc 167:037511 Pei Y, Wang J, Liu L, Wu K, Zhao Y (2007) Liquid-liquid equilibria of aqueous biphasic systems containing selected imidazolium ionic liquids and salts. J Chem Eng Data 5:2026–2031 Pei Y, Wang J, Wu K, Xuan X, Lu X (2009) Ionic liquid-based aqueous two-phase extraction of selected proteins. Sep Purif Technol 64:288–295 Plechkova NV, Seddon KR (2008) Applications of ionic liquids in the chemical industry. Chem Soc Rev 37:123–150 Poole CF, Poole SK (2010) Extraction of organic compounds with room temperature ionic liquids. J Chromatogr A 1217:2268–2286 Qin W, Li SFY (2002) An ionic liquid coating for determination of sildenafil and UK-103,320 in human serum by capillary zone electrophoresis-ion trap mass spectrometry. Electrophoresis 23(24):4110–4116 Qin W, Li SFY (2003) Electrophoresis of DNA in ionic liquid coated capillary. Analyst 128:37–41 Rádai Z, Kiss NZ, Keglevich G (2018) An overview of the applications of ionic liquids as catalysts and additives in organic chemical reactions. Curr Org Chem 22:533–556 Ramdin M, de Loos TW, Vlugt TJH (2012) State-of-the-art of CO2 capture with ionic liquids. Ind Eng Chem Res 51:8149–8177 Ren F, Wang J, Xie F, Zan K, Wang S, Wang S (2020) Applications of ionic liquids in starch chemistry: a review. Green Chem. https://doi.org/10.1039/C9GC03738A Romanovsky BV, Tarkhanova IG (2017) Supported ionic liquids in catalysis. Russ Chem Rev 86(5):444–458 Santos De Almeida T, Júlio A, Saraiva N, Fernandes AS, Araújo MEM et al (2017) Choline- versus imidazole-based ionic liquids as functional ingredients in topical delivery systems: cytotoxicity, solubility, and skin permeation studies. Drug Dev Ind Pharm 43(11):1858–1865 Schindl A, Hagen ML, Muzammal S, Gunasekera HAD, Croft AK (2019) Proteins in ionic liquids: reactions, applications, and futures. Front Chem 7:347 Scovazzo P, Kieft J, Finan DA (2004) Gas Separations Using Non-Hexafluorophosphate [PF6]Anion Supported Ionic Liquid Membranes. J Membr Sci 238(1–2):57–63. https://doi.org/ 10.1016/j.memsci.2004.02.033 Singh SK, Savoy AW (2020) Ionic liquids synthesis and applications: an overview. J Mol Liq 297(112038):1–23 Sprakel LMJ, Schuur B (2019) Solvent developments for liquid-liquid extraction of carboxylic acids in perspective. Sep Purif Technol 211:935–957

674

K. L. Wasewar

Sugden S, Wilkins H (1929) CLXVII.-the parachor and chemical constitution. Part XII. Fused metals and salts. J Chem Soc:1291–1298. https://doi.org/10.1039/JR9290001291 Vahdat SM, Zolfigol MA, Baghery S (2016) Straightforward Hantzsch four- and threecomponent condensation in the presence of triphenyl(propyl-3-sulfonyl)phosphoniumtrifluoromethanesulfonate {[TPPSP]OTf} as a reusable and green mild ionic liquid catalyst. Appl Organomet Chem 30(5):311–317 Wang P, Yang L, Wu H, Cao Y, Zhang J, Xu N, Chen S, Decoppet J-D, Zakeeruddin SM, Grätzel M (2018) Stable and efficient organic dye-sensitized solar cell based on ionic liquid electrolyte. Joule 2(10):2145–2153 Wasewar KL (2012) Low sulfur liquid fuel by deep desulfurization using ionic liquids. J Future Eng Tech 8(1):1–5 Wasewar KL (2013) Multistage extractive desulfurization of liquid fuel by ionic liquids. J Future Eng Tech 3(2):10–15 Welton T (2018) Ionic liquids: a brief history. Biophys Rev 10:691–706 Werner S, Haumann M, Wasserscheid W (2010) Ionic liquids in chemical engineering. Annu Rev Chem Biomol Eng 1:203–230 Wilkes JS (2002) A short history of ionic liquids-from molten salts to neoteric solvents. Green Chem 4:73–80 Wu Q, Chen H, Han M, Wang D, Wang J (2007) Transesterification of cottonseed oil catalyzed by Bronsted acidic ionic liquids. Ind Eng Chem Res 46(24):7955–7960 Yang G, Song Y, Wang Q, Zhang L, Deng L (2020) Review of ionic liquids containing, polymer/inorganic hybrid electrolytes for lithium metal batteries. Mater Des 190(108563):1–9 Yang Z-Z, Zhao Y-N, He L-N (2011) CO2 chemistry: task-specific ionic liquids for CO2 capture/activation and subsequent conversion. RSC Adv 1:545–567 Yarie M, Zolfigol MA, Saeidi-Rad M (2018) Tributyl(3-sulfopropyl)phosphonium hydrogen sulfate (TBSPHS) as a novel task-specific phosphonium ionic liquid: A robust catalyst for the synthesis of 1,5-dihydro-2H-pyrrol-2-ones. J Mol Liq 249:144–152 Zhang L, Chen J, Lv JX, Wang SF, Cui Y (2013) Progress and development of capture for CO2 by ionic liquids. Asian J Chem 25(5):2355–2358 Zhang Q, Shreeve JNM (2014) Energetic ionic liquids as explosives and propellant fuels: a new journey of ionic liquid chemistry. Chem Rev 114:10527–10574 Zhua J, Bai L, Chena B, Fei W (2009) Thermodynamical properties of phase change materials based on ionic liquids. Chem Eng J 147(1):58–62. https://doi.org/10.1016/j.cej.2008.11.016

Microstructure Analysis and Multi-objective Optimization of Pulsed TIG Welding of 316/316L Austenite Stainless Steel

25

Asif Ahmad

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Steps in RSM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection of Base Material and Their Mechanical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . Input Parameters with Their Working Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Design of Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Statistical Analysis for Bead Width Using RSM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Development and Evaluation of Regression Equation: Bead Width . . . . . . . . . . . . . . . . . . Adequacy Check of the Mathematical Model for Bead Width . . . . . . . . . . . . . . . . . . . . . . . Perturbation Plot (Bead Width) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Response Surface Plot: Bead Width . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion: Bead Width . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Statistical Analysis for the Depth of Penetration Using RSM . . . . . . . . . . . . . . . . . . . . . . . . . Development and Evaluation of Regression Equation: Depth of Penetration . . . . . . . . . . . Adequacy Check of the Mathematical Model for the Depth of Penetration . . . . . . . . . . . . Perturbation Plot: Depth of Penetration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Response Surface Plot: Depth of Penetration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microstructure Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanical Property Before PWHT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanical Property After PWHT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Optical Micrograph of Base Metal and Weld Metal Before PWHT . . . . . . . . . . . . . . . . . . . Optical Micrograph Weld Metal After PWHT at 800 ◦ C . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

676 678 679 680 680 680 683 684 684 685 689 690 690 692 694 695 699 701 702 704 704 704 705

A. Ahmad () Mechanical Engineering Department, PSIT-Kanpur, Kanpur, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_127

675

676

A. Ahmad

Abstract

A good-quality weld should have enough penetration, desired microstructure, and bright welding profile without any spatter. Modern welding technology started just before the end of the nineteenth century with the development of methods for generating high temperatures in localized zones. In this study, we have used pulsed TIG (tungsten inert gas) welding. In this work, the weld quality comprises of BW (bead width), DOP (depth of penetration), and its microstructure, which influence the output parameter, i.e., mechanical properties like ultimate tensile strength (UTS) and % elongation. To obtain a good-quality weld, it is, therefore, essential to control the input welding parameters. Traditional one factor at a time method of analysis is time-consuming and does not take into consideration the interaction effects between the input parameters; hence, optimization method with a total of 30 experiments was conducted using CCD of response surface methodology (RSM) to determine the optimum combination of each output process. Experimental data were analyzed by RSM using DesignExpert statistical software version 18. The statistical and analytical steps used in RSM are ANOVA; the second-order polynomial regression equation is used to develop mathematical model and response surface plots of the interaction effects of the factors to evaluate optimum conditions for bead geometry and mechanical properties. The linear, quadratic, and linear interactive effects of the input process variables on the output response were calculated, and their respective significance evaluated by ANOVA test. The p-value was used as the basis for measuring the significance of the regression coefficients, and values of p less than 0.05 signified that the coefficient is significant, otherwise insignificant. The adequacy of the model was tested by the coefficient of determination (R2 ) value as compared to the adjusted R2 value. The optimal parameter was obtained for BW (170 A, 90 A, 125 Hz, 50%) and DOP (160 A, 80 A, 100 Hz, 45%). After optimization, microstructure characterization has been done on 316 austenite stainless steel weld specimen before and after PWHT (post-weld heat treatment) to see the change in microstructure and to determine the effect of PWHT on tensile strength and on percentage elongation. Keywords

Tungsten inert gas welding · Response surface methodology · Analysis of Variance (ANOVA) · Post Weld Heat Treatment (PWHT)

Introduction The traditional method of selecting one parameter is a time-taking process and therefore not considered nowadays in the manufacturing industry; hence, an optimization technique concerns the design of experiment (DOE) such as CCD of response surface methodology (RSM) to establish an optimum condition for tensile

25 Microstructure Analysis and Multi-objective Optimization of Pulsed. . .

677

strength (Myers Raymond and Montgomery 2002). In this study, the surface plot is used to explain the main and interaction effect of the process parameter to identify the optimum parameter with their values. RSM is a widely used statistical technique in process optimization (Box and Draper 1987). RSM is a collection of mathematical and statistical techniques useful for analyzing problems in which several independent variables influence a dependent variable or response and the goal is to optimize the response. RSM helps the study of interaction among experiment variables within the range studied, allowing a better understanding of the process while reducing the experiment time and cost (Myers Raymond and Montgomery 2002). A novel RSW system of dissimilar materials of 6008-T66 aluminum alloy and H220YD galvanized high-strength steel was developed (Zhang et al. 2015), and the morphology of welding electrodes was optimally planned. The macrostructure, microstructure, and mechanical properties of welded joints with optimized electrode parameters have also been studied. In contrast with the hybrid spot-welded and bonded joints, Campilho et al. (2012) tested the experimental and numerical welding criteria to test hybrid spot-welded and bonded single lap joints. The interaction between the process parameters and the geometry of the welding bead is nonlinear and complex. It is particularly difficult to determine the extent of the contribution of the individual process parameters to the performance. It is still a puzzling issue that allows operators to only consider the state of the welding process according to practical manufacturing knowledge, maps, and handbooks (Fukuda et al. 1990; Qin et al. 2015; Dhas and Kumanan 2011). The Taguchi philosophy was used (Datta et al. 2008) to investigate the best parametric configurations to obtain the desired weld bead geometry and HAZ-related measurements. In two major fields, Taguchi tackles consistency: off-line and digital quality management. To predict the process parameters for gas metal arc welding, multiple regression analysis may be used (Tarng and Yang 1998; Lee and Rhee 2000; Lee and Um 2000). The use of the quasi-oppositional Jaya algorithm for the optimization of welding processes was implemented (Rao and Rai 2017), while Kim and Rhee (2001) and Dey et al. (2009) recommended the use of generic algorithm (GA). Furthermore, Moradpour et al. (2015) used a non-dominated genetic sorting algorithm (NSGA) to investigate weld penetration, bead width, and bead height in the submerged arc welding operation, while Xu et al. (2015) used the RSM to correlate weld bead geometry in the narrow gap of oscillating arc all location during gas tungsten arc welding (GTAW). The effect of the process parameters’ interaction was also analyzed. The minimization of the weld region was perceived (Kanigalpula et al. 2015) as a result of bead width and bead penetration. Using a generic algorithm, the constraint optimization problem was solved. On the other hand, Parida and Pal (2015) concentrated on the use of fuzzy logic and Taguchi technique in friction stir welding to optimize multiple weld output properties. Experimental design and regression analysis is the most widely employed tool for evaluating process models (Sharma et al. 2018; Sun et al. 2017; Eller et al. 2016; Banerjee et al. 2016), while fractional factorial techniques have been used to measure the dimensions of the weld bead in automated submerged arc welding (Gupta and Parmar 1989; Nielsen et al. 2015). If not regulated, welding processes can create major distortions in the final welded geom-

678

A. Ahmad

etry that cause dimensional control loss, expensive rework, and production delays (Michaleris 2011). Dobrota (2014) and Raghavendra and Kumar (2015) identified the variables that can influence the welding process into two groups, namely, variables that can be regulated (welding current, welding voltage, welding speed, base material surface cleanliness, electrical arc length, preheating temperature) and variables that cannot be controlled (size and temperature of the heat-affected zone, mechanical tensions from welded joints, leakage of molten material). In order to get the required welding hardness without distortion reaching allowable limits, this work implements a constructive means of regulating welding parameters. The goal of the work is to use Taguchi modeling and RSM simulation techniques without costly laboratory trials to model and refine welding processes. The simulated welded components exhibited outstanding strength characteristics and light weight with allowable distortion after effective design. This would decrease the cost of production and improve the efficiency, consistency, and reliability of welding processes for welding. In addition, the evolved model offers the most feasible combination of welding process parameters that, at optimum speed and expense, will deliver the best welding. The application of these research results will not only fulfill the quest of manufacturing industries to strike the right balance between cost and efficiency but will also include design data and predictive model for the manufacturing industries of rail cars and other manufacturing industries that use welding as a collaborative tool to decide the optimum solutions during welding. This improves the integrity of the weld, avoids re-welding, or is a scrap for welding operations. During the assembly of rail car parts, the implementation of the built model would improve the efficiency of welding operations.

Steps in RSM • • • • • • • • •

Identifying the important process control parameter Finding the upper and lower limits of the control variables Developing the design matrix Conducting the experiments as per the design matrix Recording the responses The development of mathematical models Calculating the coefficients of the polynomials Checking the adequacy of the models developed Testing the significance of the regression coefficients, recalculating the value of the significant coefficients, and arriving at the final mathematical models • Presenting the main effects and the significant interaction effects of the process parameters on the responses in two- and three-dimensional (contour) graphical form • Analysis of results

25 Microstructure Analysis and Multi-objective Optimization of Pulsed. . .

679

Selection of Base Material and Their Mechanical Properties “316 sheets of 100 X 75 X 4 mm stainless steel are autogenously welded with a butt joint without edge preparation” (Ahmad and Alam 2018), as seen in Fig. 1. Tables 1 and 2 provide the chemical composition and mechanical properties of the 316 stainless steel sheets.

Fig. 1 Base metal AISI 316 stainless steel

Table 1 Chemical composition of the base material (wt %) Grade 316 Min. Max.

C – 0.08

Mn – 2.0

Si – 0.75

P – 0.045

S – 0.030

Cr 16.0 18.0

Mo 2.0 3.0

Ni 10.0 14.0

N 0.10

Table 2 Mechanical properties of 316 stainless steel Tensile strength

564 MPA

Tensile strength (MPa), min

515

Yield strength, 0.2% Proof (MPa)

205

Elongation (% in 50 mm), min

40

Hardness

Rockwell, HR B max 95

Brinell, HB max 217

680

A. Ahmad

Table 3 Independent parameters with their levels for CCD Input parameter Peak current (I) Base current (I) Pulse frequency (Hz) Pulse-on time (%)

Factor symbol P B F

Level 1 −α 140 60 50

Level 2 −1 150 70 75

Level 3 0 160 80 100

Level 4 1 170 90 125

Level 5 α 180 100 150

T

35

40

45

50

55

Input Parameters with Their Working Range From the literature survey (Peasura 2015) and research work done in the past, the most important process parameters are having a greater influence on bead geometry and mechanical properties. AISI 316 stainless steel sheets of dimension 100 × 75 × 4 mm were used for the butt joint. Peak current, base current, pulse frequency, and pulse-on time are input parameters used for this experiment (Myers Raymond and Montgomery 2002). Input parameters with their levels are given in Table 3. The experiment was carried out at an optimum in the laboratory (Ahmad and Alam 2019).

Design of Experiment The experimental design for this investigation is CCD and the response measured by RSM (Kim et al. 2005). To optimize the process parameter of pulse TIG welding, examine the combined effect of four different input parameters on bead geometry and mechanical properties and drive a mathematical model. Five-level, four-parameter CCD which include 24 = 16 factorial point plus 6 central points and 2 × 4 star point (24 + 2*4 + 6), with a total of 30 experiments, were made in this investigation as shown in Table 4. The framework for the four factors was ranged between five levels (−α, −1, 0, +1, and +α) (Kim et al. 2005; Ahmad and Alam 2019).

Statistical Analysis for Bead Width Using RSM CCD was used to experiment by varying the input process parameter. The experiment was performed by varying input parameters using experimental design CCD. The experiment has been conducted according to different combinations of parameters as shown in Table 5. The experiment results obtained from CCD were fitted to the polynomial regression equation developed by Design-Expert software 18.0 (Ezekannagha et al. 2017; Moi et al. 2019). The statistical steps followed are ANOVA, regression analysis, and response surface plots of the interaction effects of the parameters to evaluate optimum

25 Microstructure Analysis and Multi-objective Optimization of Pulsed. . .

681

Table 4 Design of experiment or central composite design arrangement Run 9 17 12 28 16 1 20 11 8 24 5 18 14 6 27 23 3 15 7 26 19 4 29 22 10 2 13 25 21 30

Std 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Factor symbol A B −1 −1 1 −1 −1 1 1 1 −1 −1 1 −1 −1 1 1 1 −1 −1 1 −1 −1 1 1 1 −1 −1 1 −1 −1 1 1 1 −α 0 α 0 0 −α 0 α 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

C −1 −1 −1 −1 1 1 1 1 −1 −1 −1 −1 1 1 1 1 0 0 0 0 −α α 0 0 0 0 0 0 0 0

D −1 −1 −1 −1 −1 −1 −1 −1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 −α α 0 0 0 0 0 0

Actual factor P B 150 70 170 70 150 90 170 90 150 70 170 70 150 90 170 90 150 70 170 70 150 90 170 90 150 70 170 70 150 90 170 90 140 80 180 80 160 60 160 100 160 80 160 80 160 80 160 80 160 80 160 80 160 80 160 80 160 80 160 80

F 75 75 75 75 125 125 125 125 75 75 75 75 125 125 125 125 100 100 100 100 50 150 100 100 100 100 100 100 100 100

T 40 40 40 40 40 40 40 40 50 50 50 50 50 50 50 50 45 45 45 45 45 45 35 55 45 45 45 45 45 45

conditions for the bead geometry and mechanical properties. The linear, quadratic, and linear interactive effects of the input parameter on the bead geometry and mechanical properties were calculated, and their respective significance was evaluated by ANOVA test (Ezekannagha et al. 2017; Moi et al. 2019). The p-value was used as the basis for measuring the significance of the regression coefficients; values of p less than 0.05 signified that the coefficient is significant, otherwise insignificant. Each response variable of the experimental planning was fitted to a second-order polynomial equation generated by Design-Expert and presented in Eq. (1):

682

A. Ahmad

Table 5 CCD, experimental value bead width

Run. 9 17 12 28 16 1 20 11 8 24 5 18 14 6 27 23 3 15 7 26 19 4 29 22 10 2 13 25 21 30

Std 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

y = b0 +

Factor symbol. A B −1 −1 1 −1 −1 1 1 1 −1 −1 1 −1 −1 1 1 1 −1 −1 1 −1 −1 1 1 1 −1 −1 1 −1 −1 1 1 1 −α 0 α 0 0 −α 0 α 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

C −1 −1 −1 −1 1 1 1 1 −1 −1 −1 −1 1 1 1 1 0 0 0 0 −α α 0 0 0 0 0 0 0 0

n=4

n=4

i=1

bi xi +

D −1 −1 −1 −1 −1 −1 −1 −1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 −α α 0 0 0 0 0 0

i=1

Actual factor. P B 150 70 170 70 150 90 170 90 150 70 170 70 150 90 170 90 150 70 170 70 150 90 170 90 150 70 170 70 150 90 170 90 140 80 180 80 160 60 160 100 160 80 160 80 160 80 160 80 160 80 160 80 160 80 160 80 160 80 160 80

bii xi2 +

Exp. value bead width F 75 75 75 75 125 125 125 125 75 75 75 75 125 125 125 125 100 100 100 100 50 150 100 100 100 100 100 100 100 100

T 40 40 40 40 40 40 40 40 50 50 50 50 50 50 50 50 45 45 45 45 45 45 35 55 45 45 45 45 45 45

2.79 2.41 2.14 2.94 2.52 2.29 3.10 2.57 2.33 3.12 2.62 2.35 3.35 2.69 2.48 3.59 2.78 2.64 3.38 2.75 2.43 3.65 3.14 2.66 3.88 3.21 2.61 3.23 1.85 3.42

n=3 n=4 . bij xi xj + ε i=1

j =i+1

where y is the estimated response (bead geometry and mechanical property) n = no. of input parameter b0 = Constant bi = Coefficient of linear (A, B, C, and D)

(1)

25 Microstructure Analysis and Multi-objective Optimization of Pulsed. . .

683

bii = Coefficient of quadratic (A2 , B2 , C2 , and D2 ) bij = Coefficient of cross-product (A*B, B*C, C*D, and A*D) ε = Random error x = ith term of independent variable

Development and Evaluation of Regression Equation: Bead Width The correlation between process parameters and output response was obtained by using CCD (Ezekannagha et al. 2017; Moi et al. 2019). The second-order polynomial regression equation was fitted between the output response and input process parameter (Ezekannagha et al. 2017; Moi et al. 2019). From the ANOVA result shown in Table 6, it has been found adequacy of the model is suitable to analyze the experimental value. The regression equation based on the regression coefficient of ANOVA results is shown in Eq. (2): Table 6 ANOVA: bead width Source Model A B C D A×B A×C A×D B×C B×D C×D A2 B2 C2 D2 Residual Lack of fit Pure error Cor total R2

Coefficient 3.03 0.0140 −0.0402 0.1806 0.0340 0.1003 −0.0791 0.0809 0.0934 −0.0741 0.0928 −0.1179 −0.0291 −0.0347 −0.0710

= 0.997409, adjusted

Sum of squares 2.05 0.0047 0.0388 0.7830 0.0277 0.1610 0.1000 0.1048 0.1397 0.0878 0.1378 0.3810 0.0233 0.0331 0.1382 4.80 2.30

df 14 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15 10

Mean square 0.1466 0.0047 0.0388 0.7830 0.0277 0.1610 0.1000 0.1048 0.1397 0.0878 0.1378 0.3810 0.0233 0.0331 0.1382 0.3201 0.2300

2.50

5

0.5003

6.85

29

R2

= 0.994991

F-values 3.02 14.45 1.61 3.61 2.51 1.34 0.15 5.44 0.15 6.684 E−15 0.6020 0.5805 5.22 5.22 0.58

p-value 0.0208 0.0468 0.2245 0.0017 0.1341 0.26 0.70 0.03 0.70 1.00 0.45 0.46 0.04 0.04 0.46

0.4598

0.8616

Significant

Not significant

684

A. Ahmad

Bead width = 3.03 + 0.0140A − 0.0402B + 0.1806C + 0.0340D + 0.1003AB − 0.0791AC + 0.0809AD + 0.0934BC − 0.0741BD + 0.0928CD − 0.1179A2 − 0.0291B 2

(2)

− 0.0347C 2 − 0.0710D 2 To obtain a statistically significant regression model p-value, if p-value fr ): Each 50% cycle comprises a part of power delivered to the resonant frequency operation, but it varies as the resonant half cycle does not complete as switching cycle interrupts. Therefore, the turnoff losses on the

Fig. 7 LLC resonant converter

880

V. Harikrishna and R. Gunabalan

Fig. 8 Equivalent resonant circuit

primary side of MOSFETs are increased and hard commutation occurs at secondary rectifiers. Hence, the converter operates at higher input voltage in order to get buck operation. Below resonant frequency (fs < fr ): Each 50% cycle comprises a part of power delivered to the resonant frequency, and the current through the resonant inductor ILr reaches the magnetizing current. The freewheeling operation is carried at the end of the switching 50% cycle. Therefore, conduction losses on the primary side are increased. The converter operates at low input voltage where a boost operation is required. The various components of LLC resonant converter is designed using the following steps: The equivalent circuit at resonant condition is shown in Figure 8. The gain is represented in transfer function.    V oac (s)  f x 2 (m−1)  =  K (q, m, f x) =    2 2  V inac (s)  m f x 2 −1 +f x 2 f x 2 −1 × (m−1)2 × Q2 (15)

where: 

Q = Quality factor = q =

Lr Cr

R.

8Np 2 R π 2N s2 = ff rs

R = Reflected load resistance = Rac = Fx = Normalized switching frequency Fr = Resonant frequency = M=

√1 2π Lr.Cr

Lm+Lr Lr

The high frequency transformer design steps are as follows: Step 1. Calculate the core area and window area of the core

33 Performance Comparison of Two-Stage LED Driver for Tube Light Applications

Ac.Aw =

VI 2 × K × Bm × Fs × J

881

(16)

From the product of Ac and Aw, the core suitable for transformer is selected Step 2. Calculate number of primary and secondary turns Np =

V1 4 × Bm × Ac × Fs

(17)

Ns =

V2 4 × Bm × Ac × Fs

(18)

Step 3. Calculate the primary and secondary conductor size a1 =

I1 J

(19)

a2 =

I2 J

(20)

Step 4. Calculate the primary inductance and secondary inductance Inductance =

No.of turns Reluctance

(21)

le μo μr Ac

(22)

Reluctance (S) = where: le = effective length of the core μo = absolute permittivity μr = relative permeability Ac = effective core area Primary inductance (Lp ) = Np S2 Secondary inductance (Ls ) = NS 2s  Mutual inductance (Lm ) = K Lp Ls

The driver circuit for powering an LED module is depicted in Figure 9. It is a combination of a modified PFC single stage ac–dc converter with a half bridge type LLC converter into a mono-stage power circuit.

882

V. Harikrishna and R. Gunabalan

Fig. 9 A solitary stage driver for street lights for LED street-lighting

Fig. 10 Block diagram

Boost Buck LED Driver Figure 10 illustrates the general block diagram of the dual stage LED driver . The circuit consists of a bridge rectifier with LC filter in input side, a boost converter stage in the middle for power factor improvement, and a buck converter stage at the load side. The first stage of the driver circuit is provided for input PFC (boost converter) and the second stage (buck converter) is designed for current regulation of the LED lamp. Figure 11 describes the LED driver circuit diagram. The operation of the LED driver circuit is divided into three different modes in a working period. In mode 1, both switches are ON. When switch S1 is ON, the inductor L1 is energized through the diode rectifier. The capacitor C1 charges the inductor L2 when switch S2 is ON and supplies power to the load. In mode 2, the switch S1 is in operation and the switch S2 is OFF. The inductor L1 is energized continuously through the switch S1 and the inductor L2 supplies power to the load through diode D6 . In mode 3, both switches are turned OFF. The inductor L1 provides energy to the capacitor C1 and inductor L2 supplies energy to the load through diode D6 in continuous conduction mode. The capacitor C2 supplies energy to the load in discontinuous conduction mode. Once the voltage across the capacitor C1 is less than the supply voltage, diode D5 becomes forward biased. The capacitor C1 charges to a maximum voltage and mode 1 begins again. The dual stage LED driver is designed for 18 W at a voltage rating of 60 V and a current rating of 0.3 A, respectively. The switching frequency of the converter is 50 kHz and the utility supply is 230 V rms, 50 Hz.

33 Performance Comparison of Two-Stage LED Driver for Tube Light Applications

883

Fig. 11 Dual stage LED driver

Boost Converter Design The output voltage of a full wave diode bridge rectifier is obtained as follows: Vdc

√ √ 2 2 2 2 Vac = 230 = 207 V = π π

(23)

With a duty ratio of 0.4, the boost converter output voltage Vo is calculated by 1 1 Vo Vo ; = = ; V0 = 345 V 1−D Vi 1 − 0.4 207

(24)

The design equations for the boost inductor and the capacitor are as follows: Inductor current rippleIL =

DVi fs L1

(25)

With D = 0.4, fs = 50 kHz, a ripple current of 20% of the output current, the boost inductor L1 is calculated by 0.2 ∗ 0.3 =

0.4 ∗ 207 ; L1 = 27.6 mH 50000 ∗ L1

The voltage rippleVc =

DIo fs C1

(26)

884

V. Harikrishna and R. Gunabalan

With D = 0.4, fs = 50 kHz, a ripple voltage of 2% of the output voltage, the boost capacitor C1 is calculated by 0.02 ∗ 345 =

0.4 ∗ 0.3 ; C1 = 34.78 μF 50000 ∗ C1

Buck Converter Design The output of the boost converter is connected to the buck converter which is equal to 345 V. The required output of the buck converter is 60 V. The duty ratio of the buck converter is calculated as D=

60 V0 = = 0.174 Vi 345

(27)

For the smooth functioning of LED, continuous conduction is preferred for the buck converter. The output inductor L2 is designed by the following equation: Inductor L2 =

R (1 − D) 2fs

(28)

The voltage rating of each LED is 3 V and series of 20 LEDs are connected to meet the load demand. The cut in voltage of each LED is 2.5 V as per its V-I characteristics. The current rating of the LED is 0.3 A and remains same in a series path. The voltage equation of LED is VD = ID RD + Vγ

(29)

60 = 0.3RD + 50; RD = R = 33.33 

Inductor L2 =

33.33 (1 − 0.174) = 27.53mH 2 ∗ 50000

For capacitor design, the following equation is used: Capacitor C2 =

1−D ; C2 = 0.75 nF 16fs2 L

(30)

33 Performance Comparison of Two-Stage LED Driver for Tube Light Applications

885

Input LC Filter Design The cut-off frequency is assumed as one-fourth of the switching frequency. For a filter capacitance of C = 0.47 μF, the filter inductor is determined by using the following formula: fr =

1 √

(31)

2π LC

L = 0.345 mH

Performance Analysis The converter design specifications are given in Table 1. The circuit is simulated in powersim (PSIM) environment. The LED driver circuit is connected to the input supply of 230 V, 50 Hz. For closed loop operation, the PI controllers for input power factor correction and output current regulation are designed individually. The supply side voltage and current waveforms are illustrated in Figure 12a. The input current is continuous and the input power factor is 0.98 at rated input voltage. The output voltage and current signals of the LED driver circuit under closed loop operation are presented in Figure 12b. The output dc voltage is 60 V and the current is 0.3 A which is same as the designed values. The output power is 18 W. The input power factor, efficiency and THD are obtained for input voltages between 150 and 230 V. At a rated input voltage of 230 V, high power factor (0.98), low input current THD (11%), and high efficiency (94%) are obtained. The power factor is maintained between 0.89 and 0.98. Table 2 shows the performance parameters of the driver circuit for various input voltages. The input current THD is less than 20% for a supply voltage of 150 V and is 11% at rated voltage. The overall efficiency of the converter is 94% at rated voltage. Table 3 shows the performance comparison among the LED driver circuits for similar power rating. It is observed that the power factor remains same in all the LED driver converter topologies. The projected driver circuit has higher Table 1 Specifications

Parameters Input voltage Output voltage Output current Switching frequency Filter inductance Filter capacitance Boost capacitor Boost inductor Buck capacitor Buck inductor

Ratings 230 V 60 V 0.3 A 50 kHz 0.34 mH 0.47 μF 34.78 μF 27.6 mH 0.75 nF 27.53 mH

886

V. Harikrishna and R. Gunabalan

Current (A)

Voltage (V)

a 500 0 -500

0.2 0 -0.2 -0.4 0.55

0.5

80

Current (A)

Voltage (V)

b 40 0 0

2

0.6 Time (s)

0.7

0.6 Time (s)

0.65

0.4 0.2 0

4 Time (s)

0

2

4 Time (s)

Fig. 12 PSIM simulation results: (a) input voltage and current, (b) output voltage and current Table 2 Performance parameters with respect to input voltages

Input voltage (V) 150 180 200 230

% THD 20 16 13 11

PF 0.89 0.92 0.96 0.98

%η 79 81 90 94

efficiency than the topologies discussed in literatures. The major factors considered for the fabrication is the design of inductors. The ferrite core inductors are generally preferred for medium and high frequency to minimize the eddy current losses. The LED driver circuit directly connected to the mains can be damaged by switching transients. The voltage stresses across switches are shown in Figure. 13. The voltage stress across switch S1 is 325 V. For the same power rating, the voltage stress across the switch S1 is 950 V in a single stage converter with a coupled inductor (Cheng et al. 2016). The switching voltage stress is less in Figure 13a as compared with the single stage driver circuit employing a coupled inductor in Figure 13b. The major significance of the simulation results are: good input power factor with reduced current THD, higher efficiency, and less switch voltage stress. Table 4 shows the cost analysis of the driver circuit with a two stage (Athalye et al. 2012), one-and-ahalf stage (Valipour et al. 2016), and a single stage driver (19W 2011; Cheng et al. 2016) for the nearest power rating. The dual stage circuit presented in this work has a lesser component count than the other driver circuits. The brightness control is required for street lighting and automotive head lighting applications. By adjusting the width of the low frequency signal, the brightness can be controlled.

Conclusion A dual stage LED driver circuit for tube light applications with a boost PFC converter and a buck current regulator has been presented. Using the boost PFC

33 Performance Comparison of Two-Stage LED Driver for Tube Light Applications

887

Table 3 Performance comparison (Harikrishna et al. 2020) Circuit configuration Dual stage LED driver projected in this work One-and-a-half stage LED driver (Valipour et al. 2016) Single stage LED driver (Cheng et al. 2016) Single stage LED driver (19W 2011)

% THD 11 15

Input PF 0.98 0.98

%η 94 84

18 19

7.22 13.5

0.98 0.956

92.42 87.7

b 600 400 200 0 -200 0.23008

Voltage (V)

Voltage (V)

a

Power (W) 18 20

0.23012

1000 500 0 -500 0.27754

0.23016

0.27756

0.27758

0.2776

Time (s)

Time (s)

Fig. 13 Voltage stresses across switches: (a) switch S1 , (b) switch S1 (Cheng et al. 2016) Table 4 Cost analysis

Components Power rating (W) No. of switches No. of line frequency diodes No. of fast recovery diodes No. of inductors Fly-back transformer No. of capacitors Input filter Manufacturing complexity

Dual stage LED driver projected in this chapter 18 2 4

Dual stage LED driver (Athalye et al. 2012) 12 2 4

One-and-ahalf stage LED driver (Valipour et al. 2016) 20 2 4

Single stage LED driver (19W 2011) 19 1 4

Single stage LED driver (Cheng et al. 2016) 18 2 2

2

2

3

2

4

2 (buck and boost) –

3 (coupled and buck) –

1

–

1

1

2 (coupled and resonant) –

2 LC type Simple

3 LC type Difficult

2 LC type Difficult

2 LC type Simple

3 LC type Difficult

circuit, high input power factor is achieved. PSIM simulation software tool was used to build the driver circuit, and simulation results are presented to prove the effectiveness of the converter. The various types of single stage power converter circuits are explained in detail in order to understand the concept of two stage power converter. The performance characteristics of the LED driver have been evaluated for 18 W at a line voltage of 230 V, 50 Hz. At rated voltage, the circuit efficiency is 94% with an input current THD of 11%. The input line power factor is 0.98 at rated supply voltage. The circuit is also tested for wide range of input

888

V. Harikrishna and R. Gunabalan

voltages between 150 V and 230 V. The performance parameters are compared with the presented double stage and single stage driver circuits in literatures. The driver circuit projected in this work is very simple and requires independent control compared with the single stage drive circuits. The circuit could be designed for 50 W to 150 W by varying the duty cycle for street lighting applications. It is also preferred for solar street lighting applications in remote areas with battery backup and automotive lighting applications with dimming control.

Websites for References https://www.globallightingassociation.org/ https://www.cree.com/led-components/ https://www.ti.com/power-management/led-drivers/overview.html https://www.ledsmagazine.com/

References 19W, Single-stage AC/DC LED driver for T8/T10 fluorescent lamp replacement, Texas Instruments. 1–21 (2011) Athalye P, Harris M, Negley G (2012) A two-stage LED driver for high-performance high-voltage LED fixtures. In: Proceedings of applied power electronics conference and exposition, IEEE, Orlando, pp 2385–239 Bailey D (2011) An idea to simply LED lighting purchase decisions Bodo’s Power Systems – Electronics in Motion and Conversion 18 Broeck HVD, Sauerlander G, Wendt M (2007) Power driver topologies and control schemes for LEDs In Proceedings of applied power electronics conference, IEEE, Anaheim, pp.1319–1325 Chen X, Huang D, Li Q, Lee FC (2015) Multichannel LED driver with CLL resonant converter. IEEE J Emerg Select Topics Power Electron 3(2):589–598 Cheng CA, Chang CH, Cheng HL, Chung TY (2014a) A single-stage high-PF driver for supplying a T8-type LED lamp In Proceedings of international power electronics conference, IEEE Hiroshima, Japan, pp 2523–2528 Cheng CA, Cheng HL, Chung TY (2014b) A novel single-stage high-power-factor LED streetlighting driver with coupled inductors. IEEE Trans Ind Appl 50(5):3037–3045 Cheng CA, Chang CH, Chung TY, Yang FL (2015) Design and implementation of a single-stage driver for supplying an LED street-lighting module with power factor corrections. IEEE Trans Power Electron 30(2):956–966 Cheng CA, Chang CH, Cheng HL, Tseng KC (2016) A single-stage LED tube lamp driver with input-current shaping for energy-efficient indoor lighting applications. J Power Electron 16(4):1288–1297 Cheng CA, Chang EC, Tseng CH, Chung TY (2017) A single-stage LED tube lamp driver with power factor corrections and soft switching for energy saving indoor lighting applications. Appl Sci 7(2):115 Harikrishna V, Gunabalan R, Senthil Kumar S (2020) Pulse width modulation converter for lightemitting diode tube light applications. Int Trans Electr Energ Syst 30(4):1–10 Hui SY (2010) A novel passive offline LED driver with long lifetime. IEEE Trans Power Electron 25(10):2665–2672 Khatib M (2009) Ballast resistor calculation – current matching in parallel LEDs. Texas Instruments, Dallas, TX, Application Rep. SLVA325

33 Performance Comparison of Two-Stage LED Driver for Tube Light Applications

889

Kim HC, Choi MC, Kim S, Jeong DK (2017) An ac–dc LED driver with a two-parallel inverted buck topology for reducing the light flicker in lighting applications to low-risk levels. IEEE Trans Power Electron 32(5):3879–3891 Liu W, Chau KT, Lee HT, Jiang C, Wei H, Lam WH (2020) A wireless dimmable lighting system using variable-power variable-frequency control. IEEE Trans Ind Electron 67(10):8392–8404 User Guide for FEBFL7701 L34U018A evaluation board universal input 18.3W LED driver, Fairchild semiconductor. 1–24 (2012) Valipour H, Rezazadeh G, Zolghadri MR (2016) Flicker-free electrolytic capacitor-less universal input offline LED driver with PFC. IEEE Trans Power Electron 31(9):6553–6561 Zhang Y, Rong G, Qu S, Song Q, Tang X, Zhang Y (2020) A high-power LED driver based on single inductor-multiple output DC–DC converter with high dimming frequency and wide dimming range. IEEE Trans Power Electron 35(8):8501–8511 Zhao C, Xie X, Liu S (2013) Multi-output LED drivers with precise passive current balancing. IEEE Trans Power Electron 28(3):1438–1448

Design and Development of Automated Vertical Farming Setup

34

Karthik Warrier, Mukundhan Rajendiran, Shrawan Kumaar Kannan, and R. Ranjith Pillai

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Related Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Systems Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Vertical Stack System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Stacker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Mobile Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Plant Tray Sensor Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Communication Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Arena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sequence of Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Mobile Robot Navigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Robot Alignment Paradigm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Observation and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

892 895 899 900 901 902 906 907 908 909 911 913 920 921 922 923 924

Abstract

Imparting new methodologies into agriculture has become a common order nowadays. The ever-increasing necessity for food production and availability to meet the requirements of the growing population forms a strong backbone to these sustainable new methods. These also propose certain advantages over traditional agriculture such as requirement for less areas of land, immunity to

K. Warrier · M. Rajendiran · S. K. Kannan · R. Ranjith Pillai () Department of Mechatronics Engineering, College of Engineering and Technology, SRM Institute of Science and Technology, Chennai, Tamil Nadu, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_164

891

892

K. Warrier et al.

climate changes, and easy access to urban landscapes altogether leading to an increase in quality. An apt example of such a method would be vertical farming, practiced in an indoor environment. The efficiency of these farms is augmented by automation and the use of robots. The work focuses on developing a modular automated setup prototype, with intended use in a vertical farm. It discusses the design and development of key systems, namely, a storage system with vertical levels for housing the plant modules, known as the Vertical Stack System (VSS). It also features certain mechanical elements that aid in a smooth transfer of these modules in and out. A mobile robot is also discussed in the work, developed to navigate the farm environment and transport the plant modules to and from the VSS. The robot’s behavior in a virtual and real world is presented through the use of a simulation model, developed using the kinematics of the mobile robot, and a computational model, developed using sensor feedback, respectively. The development of an ingenious system called the Stacker is also presented. The Stacker is present on-board the mobile robot that is principally responsible for the transfer of plant modules. Additional work is presented for incorporating automation into the setup, through a sensor setup for monitoring and reporting certain environment variables surrounding the plant and through an interconnected operation. The engineering design of each of these systems is discussed, and the control strategies, operational setup, and results are presented later on. Keywords

Automated vertical farming · Mobile Robot · Internet of things · Automated storage and retrieval system · Agriculture robots

Introduction The rapid increase in urbanization and population has led to an amplified demand in food production, shortage of available land, and food security, as discussed by the UN (United Nations 2017). (By 2050, roughly 80% of the populace is expected to live in cities, directly impacting food demand). Traditional agriculture practices require massive amounts of land for production; they are usually located far away from an urban landscape and affected by the weather conditions (Smit and Nasr 1992). Vertical farming attempts to solve the limitations of traditional farming. It is practiced in an indoor environment, based on the technique of controlled environment agriculture (CEA) (Charron et al. 1996). This methodology preserves the crops against climate changes, infestation, pesticides, and pollution (Al-Kodmany 2018). These farms can be developed right in the middle of a city or an urban landscape, thus reducing the travel costs (AlKodmany 2018). Plants are grown in modules or trays and then stacked vertically in layers, thus eliminating the requirement of massive lands. These modules are continuously monitored, through cameras and manual inspection (Benke and

34 Design and Development of Automated Vertical Farming Setup

893

Tomkins 2017). The environment control variables are preserved or maintained at the desired set points (Pawlowski et al. 2009). Vertical farms support the cultivation of different varieties of crops such as lettuce, kale, chard, collard, chives, mint, basil, and small woody herbs (Beacham et al. 2019). Vertical farms allow for more organic crops to be cultivated, throughout the year production, less use of chemicals and plant diseases. The increased demand in productivity from agriculture and decrease in inputs from labor over time have led to the surge of developments in agricultural technology. The widespread availability of technology has led to advancements in automation and robotics in indoor farming. Additionally, due to a reduction in costs and technology, for example, lower prices of LEDs, capital investments on indoor farming have seen a steady growth. As can be seen in Fig. 1, there is a clear rise in the investments made in farm robotics and automation, which has grown four times in 4 years from 2014 to 2018. The continuation of this trend could lead to added support to the food industry, contributing to a sustainable future. A number of notable companies active in the domain are listed in (https://roboticsandautomationnews. com/2019/05/03/top-25-vertical-farming-companies/22181/). It is evident from Fig. 2 that the vertical farming market in Canada has and continues to witness significant developments with the rise of methodologies

Fig. 1 Trend in capital investments in farm robotics and automation (https://www.ft.com/content/ 0b394693-137b-40a4-992b-0b742202e4e1)

894

K. Warrier et al.

Fig. 2 Impact of automation on vertical farm market in Canada (©Grand View Research, grandviewresearch.com) (https://www.grandviewresearch.com/industry-analysis/vertical-farmingmarket )

such as hydroponics and electronic devices such as LEDs. Hydroponics is the method of growing crops without soil, thus eradicating various plant diseases and permitting easy installations and maintenance. Additionally, it also augments the yield compared to soil-grown methods due to precise control over nutrient deliveries. The deployment of controlled environment agriculture (CEA) has also been predominantly on a surge since 2018, denoted by climate control in the figure. The fundamental purpose of CEA is to shield and preserve ideal development conditions during the plant growth period. Various attributes, such as proper lighting, carbon dioxide, water, humidity, pH levels, and nutrients, are maintained and monitored. This task is carried out through the use of sensors. For example, the use of sensors for pH, water, humidity, and temperature reports back data and also alerts the operator in case of any eccentricities, extremities, or, in some cases, acts by itself. Sunlight has been a prime factor for the process of photosynthesis. However, in a controlled environment where access to sunlight can be denied in certain situations, LEDs are employed as an alternate light source. Availability of LEDs and its key factors such as the absence of harmful chemicals and lesser expulsion of heat energy has surged the use of LEDs in indoor farming and is expected to grow steadily in years ahead. The design aspect of a typical vertical farm can be thought of as a warehouse environment. A general warehouse scenario features rack-based storage of loads in bidirectional fashion (vertical and horizontal). Most often, AS/RS systems are employed in order to automate the task of load pickup and drop-off. Such systems significantly are able to reduce labor constraints, increase the use of the available floor, offer modularity, and increase storage density. AS/RS systems operate on demand, i.e., a load is stored or picked up only when a command is put forth by

34 Design and Development of Automated Vertical Farming Setup

895

the operator. Autonomous mobile robots or AMRs are also used in warehouses that put forward traversing the storage racks at varied heights and conducting operations throughout the warehouse. Loads can be replaced with plant modules, and thus vertical farms can take up a similar fashion as warehouses. The modules are equidistant from each other sufficient enough to aid plant growth. Each module is provided with PAR lights, i.e., photosynthetically active radiation lights. These are artificial sources of light commonly known as grow lights. It is a band of radiation on which plant can carry out photosynthesis, having a wavelength between 400 nm and 700 nm. As per a report by Agrilyst in 2017, manual labor still accounts for more than 50% of the production costs in vertical farms present across the globe. Automation can significantly improve a vertical farm’s overall efficiency and has turned into a requirement for optimization (David et al. 2017). Effective communication between various sensors and actuators will allow smart control over several processes involved in farming, such as delivery and inspection and continuous monitoring of parameters (Monteiro et al. 2018). Robotic systems also play an important role in automating several tasks in modern farms. Such systems provide precise control, guidance, and operations (Cheein and Careli 2013; Lauguico et al. (2019)). Figure 3 shows the current levels of automation trends in the vertical farm sector. Car automation principles have been used to come up with the framework, as discussed in (https://www.agritecture.com/blog/2019/5/10/automation-thefinal-frontier-of-vertical-farming#:~:text=The%20majority%20of%20commercial %20vertical,Level%203%20(conveyor%20automation)). It is to be noted that, at present, most if not all systems have been focused onto incorporating the third level of automation, and the present research has been carried out on implementing the fourth level of automation. In the third level of automation, processes such as seeding, harvesting, and packaging are automated by machines. The fourth level of automation would introduce the aspect of cognition wherein the machines have the knowledge to respond to the plants without human interference.

Related Works Robotics and automation find themselves surrounded by active research when it comes to agriculture. Usage of machinery to handle tasks in a farm environment is long postulated to positively impact efficiency. Due to their design, vertical farms are implemented in high buildings and can be intuitively thought of a replica of rack-based warehouse design. Automation has significantly impacted the operation of warehouses with the use of AS/RS systems, as was noted (Gagliardi et al. 2014). However, it is important to note that warehouses contain more than one rack unit; in single shuttle AS/RS systems, the lifting crane can only traverse along a single passageway at a time, making the retrieval or storage in inner racks difficult (Fu et al. 2015). However, multi-shuttle AS/RS systems can traverse the inner aisles effectively but lead to the requirement of additional space for the entire setup as noted in

896

K. Warrier et al.

Fig. 3 Levels of automation in vertical farms (© OnePointOne, Inc. 2019)

Gagliardi et al. (2014) and can be seen in Mostofi and Erfanian (2018). In order to construct an efficient vertical farm, compactness and farm output must go hand in hand. Vertical farms look to raise yield by cultivating crops in a vertically stretched manner, unlike traditional agriculture which expands the horizontal area; hence, it is a standard note that these farms will lead to high-rise structures. Harvesting is an essential operation in agriculture, and the same goes for vertical farms also. Henten et al. (2002) propose a methodology for automation in harvesting cucumbers using image-processing. However, this method can be expensive to implement in lesser developed nations, as was observed in David et al. (2017). Robot manipulators are also used in a vertical farm environment for pick and place operations. As seen in (MIT Technology Review 2018), the company IronOx (https://www.ironox.com) employs manipulators in the farm to pick plant modules. The use of such manipulators is also seen in Lauguico et al. (2019). These manipulators are only applicable to a certain level in the farm, owing to their fixed base. Henten et al. (2002) proposed the use of an autonomous robot for operation during a cucumber harvesting task. The robot consists of a manipulator arm resting on a mobile base. The arm has 7 DOF, and its end effector handles the load without losing quality. A thermal device is also mounted on the end effector that disables any spread of viruses when picked. A computer vision system is also included that detects the cucumber accurately up to 95%. The manipulator is affixed onto an autonomous vehicle that uses rail guidance for maneuvering between the plant stacks. However, the authors mention that the maximum possible height of harvest is 1.5 m above the ground in a high-rise cultivation scenario for cucumber, but this would not be suitable for a vertical farm as the maximum height could go more than the specified. Vertically lifting the autonomous vehicle with the arm also cannot be considered owing to its overall weight.

34 Design and Development of Automated Vertical Farming Setup

897

The feasibility of vertical farms, in general, has also come under scrutiny, particularly in articles published in Alter (2010) and Cox (2016). The authors collectively note that the practice would lead to a commotion in the rural farming sector, supported by the need to train a large workforce. It is also indicated in Cox (2016) that plants growing only on the topmost level would benefit from solar radiation. However, it is postulated that vertical farms are not developed in order to disrupt the normal farming practices; rather, it is practiced to add support to it. Not all crops can be grown in a vertical farm as was similarly noted in GonzalezDe-Santos et al. (2020). Hence, the need for rural farming would exist. Automated lighting and watering systems also eradicate the issue regarding solar radiation, as advances in technology have now availed lighting systems for individual plant modules. Additionally, rotatable collections of plant modules can also help in increased sunlight, should there be a need, noted in Morrow (2008) and Massa et al. (2008). It is also believed that the economic costs concerning vertical farm implementation could be compensated considering the losses incurred by rural farms owing to climate changes, locusts, and transportation costs, to name a few, also discussed in Fao (2019). Chin and Audah (2017) review the effects of minor changes in the surrounding of the vertical farms that have a huge impact on the plant growth and hence propose a monitoring system to keep track of the physical conditions of the crops. The system utilizes BeagleBone Black wireless board connected with ThingSpeak IoT cloud to display the real-time health conditions. It also provides a remote-control function to the end users that can facilitate in turning on/off the system. However, for their entire setup, the authors have used a single BBB controller and varieties of sensors to measure the external surroundings. This can detect the surrounding conditions in general, but it cannot be used for crop-specific parameters should different crops be grown in the setup. Ziedler et al. (2017) astutely present a complete design of a vertical farming setup for large applications. The report sheds light on the idea of individual plant modules with health monitoring systems enabled in each. The report spans over 80 pages and is a comprehensive solution to an introduction to various vertical farming operations. It depicts the vertical farming scenario when implemented on a large scale. In such a scale, however, the report underlines that human labor is still required to handle several processes. This contends the fact that labor is still the highest contributor to production costs. This could be scaled down with the help of robots and IoT. The authors (Rashid et al. 2019) implement RFID-based warehouse management and object detection and accessing of data remotely with cloud storage in their paper titled, “Smart Warehouse Management System with RFID and Cloud Database.” R.F. tags are used on every container and item being stored in the warehouse, and each container can thus be accessed smoothly. The efficiency increases with cloud storing facility. NODEMCU is used to process and send the data to the cloud. The positions of the container are stored in the cloud, which makes accessing it easier. The authors have used ID-20LA as RFID reader and designated R.F. tags for each container. However, with vertical farms, identification of containers or

898

K. Warrier et al.

plant modules in the farm could be handled using more straightforward methods such as barcodes and Q.R. codes. These codes are easier to fabricate and can store information pertaining to the module, which can then be easily read by a scanner or the most commonly available smartphones. Haris et al. (2019) focus on the infrastructural requirements for indoor vertical farming for the necessary modularity where the proposed system is a serviceoriented platform for three operations, i.e., cloud, fog, and sensors. The authors use a BLE connection between the sensor board and the fog node and have proposed a distributed secure network. The modularity allows them to control several sensors at a time, on each vertical stack. Khasasi et al. (2015) focus on the importance of automated storage and retrieval system. Due to rapid industrialization, warehouses are looking for an integrated automated system that can operate without any interference and improve the efficiency for better performance. The paper investigates the efficiency of the devices like Arduino microcontrollers, Bluetooth technology, and servo motors on working mechanisms of ASRS. Guptha et al. (2018) discuss about the design of an IoT architecture for the order picking process. However, other outcomes like real-time tracking and reduction of cost of operation with safety are also discussed. Considering the difficulty in handling control practices and competitive business environments implementing IoT in warehouses is essential. Berg and Gademann 2000 describe the operation of an automatic storage/retrieval system and examine its control policies. He evaluates the trade-off between storage space requirements and travel times and studies a new storage location policy which combines low storage space requirements with short mean travel times. This report by the author plays a vital role in describing new storage design aspects that would optimize space and also reduce the time taken to transport. Smit and Nasr (1992) propose a greenhouse weather regulation system, depicted as an event-based control. Various wireless sensor networks are used, and event triggers are generated due to external disturbances, which helps prevent actuator wear losses. They discuss a greenhouse system with individual wireless sensor networks, each collecting data over time. Such a system with separate sensor networks can cater to each plant individually. Moreover, event-based systems are advantageous because it is not the passing of time that triggers an action; instead, it is the external event that acts as the trigger. This promotes a good degree of optimization. This methodology can also have a good impact on costs as it can cut down energy charges. On the basis of the abovementioned drawbacks and issues faced and the observations, the primary focus of the work is to design an automated setup for operation in a vertical farm. The work undertaken here is to come up with a setup for automating vertical farming. The work does not directly focus on farming and cultivating crops; rather, it proposes a mechatronic setup, taking into account the available setups and the issues they face. In short, we look to provide a design for the housing of plant modules. The design proposed considers the necessity to retrieve or store a module during operations easily. A mobile robot model is proposed for

34 Design and Development of Automated Vertical Farming Setup

899

logistics purpose, to operate in the farm without the need of human interference, and to cater to its tasks. The robot’s design is carefully undertaken to suit the operation in a vertical farm. The various uses of AS/RS systems in the warehouse have been studied and have been innovated to fit a vertical farm operation. The work also directs the ideology that the farm operator need not be manually handling inspection tasks or in-farm movement in general instead; they are provided with a space outside the farm where they can monitor plant growth effectively through a computer. The work also contemplates the various advantages of using a vertical farm. However, it is to be clearly noted that in this work, vertical farming is considered an aid to rural farming techniques and not as a direct replacement to them. The types of crops grown in a vertical farm are currently limited. A degree of dependence on traditional farming is still very much needed. Overall mechanical and electronic design, along with control system algorithms, is presented, and the secondary goal is also undertaken to develop an onboard sensor network which can be implemented onto the plant module, in order to bolster the practice in general. It consists of sensors that can map the plant’s environment and transfer the data over Wi-Fi to visualize it into the computer. Through this work, a point is asserted that one of the main advantages of deploying an automated vertical farm setup is the nonnecessity of being available at the farm to monitor it throughout the year; rather, through the use of wireless connectivity, a modular IoT model can be developed to interconnect the systems and work at the ease of the operator from anywhere. It is also described that the use of robotics can increase the overall efficiency of the farm. In the situation prevailing across the globe currently, such a provision will only vouch for enhanced safety. Thus, this work discusses about the three main systems, namely, a Vertical Stack System (VSS), for automatic storing and retrieval of plant modules, a mobile robot for navigation in the farm and for logistic purposes along with an auxiliary setup for plant tray transfer (hereon referred to as the Stacker), and a plant tray sensor setup (PTSS) for capturing and visualizing data about plant surroundings. The contents of the chapter are split further into four sections. Section “Systems Design” elaborates on the mechanical and the electronic design of each system. Section “Observations” explains the operational setup and discusses control methods and various observations. Section “Conclusion” provides a conclusion to the work and a future scope outlining certain other developments that can be inducted into the existing work.

Systems Design The mechanical and electronic design features of the VSS, the mobile robot, the Stacker, the PTSS are discussed in this section, split further as “The Vertical Stack System,” “The Stacker,” “The Mobile Robot,” and “The Plant Tray Sensor Setup,” respectively. Additionally, subsection “Communication Interface” describes a communication interface for communication between the robot, the farm operator, and the VSS.

900

K. Warrier et al.

The Vertical Stack System The CAD model of the VSS is shown in Fig. 4. The VSS is designed to stack plant modules vertically. The outer frame is a cuboidal structure of dimensions 340 × 400 × 600 mm made out of joining aluminum extrusions. The need for avoiding complex machining tasks in the scenario where the VSS needs to be scaled up led to the use of aluminum T-profile extrusions of cross section 30 × 30 mm that provide modularity to the design and make it possible to vary the height and extend the dimensions (Fiorino et al. 2014). The yield strength, ease of fabrication, and stiffness are some mechanical properties considered for selecting aluminum extrusion to construct the VSS. The T-slot profile has a yield strength of 25,000 psi and a tensile strength of 36,000 psi. Square cross sections of these extrusions deliver identical strength in the vertical and the horizontal directions. These extrusions are joined to each other using L clamps and special bolt channel joints inserted into these channels. The frame is divided horizontally into two columns of width 170 mm each. Each column consists of passive roller channels that are used further to partition the VSS into two vertical levels. These channels consist of an array of plastic rollers, connected through their axis to an aluminum C-channel to undergo free motion. The space between two opposite channels is 170 mm. Each level has roller channels which are fixed across and between six vertical columns of extrusions on either side of the structure facing inward. These channels are only connected to the extrusions and leave a hollow gap between the columns for the tray’s insertion into the stack by the mobile robot. The roller channels help in effective load distribution and smooth transfer of plant modules in and out of the VSS. The partitioning done on the VSS allows for four modules to be placed

Fig. 4 VSS CAD model

34 Design and Development of Automated Vertical Farming Setup

901

on each level. In comparison, a total of eight modules can be housed in the entire system. The roller channels are also the sole contact areas for the plant modules to rest.

The Stacker The Stacker is an ingenious subsystem on board the mobile robot to principally handle the transfer of plant modules to and from the VSS. The CAD model is shown in Fig. 5. It has a dimension of 170 × 200 × 60 mm. The Stacker features a pair of roller channels on either side, on which the plant module is placed. The module is kept steady on the robot and is then pushed out using a Tray Transfer Module (TTM) on the Stacker. The Tray Transfer Module consists of two plates arranged in between the roller channels. The height of the plates is slightly more than the overall height of the Stacker. This provision is provided in order to hold the plant modules between the Tray Transfer Module plates and push/pull the plant module during operation. The Tray Transfer Module’s rear plate features a brass nut, and a threaded rod of pitch 2 mm, connected at one end to a D.C. motor, passes through the brass nut. This setup converts the motor’s rotary motion to linear motion and extends the Tray Transfer Module horizontally. The Tray Transfer Module thus has the ability to move throughout the length of the lead screw. A roller bearing is kept at the end of the rod to enable smooth motion. A flexible shaft coupling of 8 mm × 5 mm diameters on either end connects the shafts of the D.C. geared motor, and the threaded rod eliminates misalignments and vibrations caused during the transfer. A set of smooth rods and linear bearings prevent toppling and misalignments while moving. Once the mobile robot orients itself with the stack, the Tray Transfer Module moves into the slot between two roller channels of the stack system and stores or retrieves the trays from the stack. The whole Stacker setup traverses vertically on the mobile robot to reach the VSS’s required level where the plants need to be housed. This is achieved via a lift mechanism consisting of a pair of power screw of each 400 mm in length (Gopinath 2014). The components resulting in the motion include two lead screws in the robot frame. A roller bearing holds each screw on the frame on one end, and the other end is held using a flexible bearing conjoint with a stepper motor. The stepper motor has a holding torque of 4.2 kg-cm with a 1.8◦ step angle, i.e., it takes 200 steps per revolution. The shaft of the motors has a diameter of 5 mm. Brass nuts are present at the Stacker’s ends, through which the lead screws have to traverse. The nuts help in converting rotary motion offered by the screw to linear motion. The initial design consisted of only one lead screw mechanism along with a smooth rod with a linear bearing on the sides of the mobile robot for the Stacker’s motion in vertical direction. However, this design failed to provide a smooth motion to the Stacker. An immense load is exerted onto the single lead screw mechanism, leading to misalignments and causing immobility of the Stacker at various lengths of the vertical path.

902

K. Warrier et al.

Fig. 5 Stacker CAD model (isometric view)

An alternate design, consisting of a dual lead screw mechanism, as a replacement to the smooth rod can be used to resolve this drawback. The design called for a synchronous motion of the lead screw setup; this proved to have the better motion to the Stacker. However, this setup now resulted in a staggered motion of the Stacker due to the errors in the copper nut and threaded rod alignment. To overcome the aforementioned problem, a smooth rod was placed at the rear end of the mobile robot to support the Stacker throughout its vertical motion, as shown in Fig. 6. The setup enabled the smooth motion of the Stacker.

The Mobile Robot The CAD model of the mobile robot is shown in Fig. 6. It is 230 × 200 × 400 mm in dimension and resembles a hollow cuboidal structure with a base. The frame is made out of aluminum T − profile extrusions of cross section 30 × 30 mm. Corner brackets and bolt channel joints are used to connect the extrusions together. The base of the robot is made out of an acrylic plastic sheet of dimensions 230 × 200 × 10 mm. The baseplate is connected to the frame using bolt channel joints, and the top surface provides space for housing the electronic components, while the bottom surface is used to connect robot motion elements.

34 Design and Development of Automated Vertical Farming Setup

903

Fig. 6 Mobile robot CAD model

The term “robot motion elements” refers to a collection of components that perform robot motion in the vertical farm. A pair of planetary-gear-type D.C. motors are used on the rear end for locomotion. The motors have a torque of 8.1 kg-cm and a gear ratio of 126:1 with a speed of 40 RPM. The shaft diameter is 6 mm to which standard wheels of diameter 100 mm are coupled. Considering efficiency and smooth operation at maximum load, the motor driver is selected to be the Cytron MDD-10 A. It is a dual-channel motor driver capable of driving two motors at 10 A constant output. At maximum load, the motor consumes 3.7 A. The direction control and speed control are achieved through single dedicated pins on the driver, which makes it convenient for a differential drive mechanism. Swivel Castor wheels of diameter 60 mm are used in the front to provide omnidirectional steering (GonzalezDe-Santos et al. (2020); Siegwart and Nourbakhsh 2004). The robot maintains a ground clearance of 95 mm. The distance is carefully fixed after determining the robot’s total weight, speed of the robot, and safe pivoting requirements. The robot motion elements are shown in Fig. 7. A total of five motors are used in this work. The motor torques are calculated prior to selecting the suitable motors. Criteria such as the total weight and time of operation are considered. Given the radius (r) of the driven wheel to be 0.05 m, the coefficient of static friction (μ) to be 0.6, and the number of wheels (s) to be 4, the torque (τ) is calculated as shown in Eq. 2 (Joseph 2015). The total weight (W) on the robot includes the chassis weight and the load weight, which equals to 160 N:

904

K. Warrier et al.

Fig. 7 Robot motion elements

τ=μ×s×r

(1)

τ = 0.6 × 4 × 5 = 12 kg/cm

(2)

The consideration taken into account while calculating the required motor torque is that the robot would require the maximum torque during the initial stages of motion, i.e., overcoming friction (Joseph 2015). The planetary gear arrangement provides a shared load distribution, which in turn significantly improves torque capabilities (Gawande and Shaikh 2014). The overall weight to be lifted on the Stacker is given by Eq. 3: Mass of the Stacker : 3 kg Mass of the plant module ∼ 0.2 − 0.5 kg Mass of the plant module ∼ 0.2 − 0.5 kg total weight : 3.5 × 9.8 = 34.3 N (3) Since the Stacker needs to hold its position at the required level during the operation, holding torque is preferred over speed. As the vertical motion of the

34 Design and Development of Automated Vertical Farming Setup

905

Stacker is handled by two power screws, actuated by two stepper motors, and the motion of both the motors at the same time is crucial for a balanced lift of the Stacker, it is vital to provide synchronous signals and sufficient input power to both the motors at the same time. The total load on the Stacker is estimated to be approximately 4 kgs. The motor provides a holding torque of 4.2 kg-cm. The NEMA-17, being a current-controlled device, consumes 1.5 A at normal load operation per phase, and it consists of four phases. The current consumption goes up to 2A at maximum load. It provides a step angle of 1.8◦ . Thus, in this work, these motors are together controlled by a single driver module. Considering optimum efficiency at full load, the TB6560 stepper motor driver is chosen. The driver provides a constant 3A output and offers a large heat sink for heat dissipation. Both the motors are connected in parallel connection to drive the driver which maintains a constant current output to both motors. The total weight to be lifted is 3.5 kg, as can be noted from Eq. 3. Nine switches are available on board in the TB6560. The DIP switches SW1, SW2, SW3, and S1 are used for providing varying current values. S2 is used for stop current which is the current required to hold the motor shaft at a specified position. Switches S3, S4 are used for microstep settings. Smaller microstep setting provides smooth and quiet operation of the stepper motor. Switches S5, S6 are used to control the decay settings. The decay setting determines the driver chip handle for the back EMF from the motor. For the proposed work, SW1, SW3, and S1 were switched on, while SW2 was switched off for obtaining a current of 1.5A. S2 was switched off for 50% stop current. S3, S4, S5, and S6 were switched off for obtaining full step microstep setting and 0% decay setting for the motor. The horizontal motion undergone by the Tray Transfer Module to push/pull the plant module is via a D.C. motor coupled with a lead screw using a flexible bearing. Only two modes of operation are considered, full extension of Tray Transfer Module and complete retraction of Tray Transfer Module. Thus, the selected motors as per Eqs. 1 and 2 are summarized in Table 1. The planetary-gear-type D.C. motors are also coupled with two-channel incremental encoders. The encoders provide the primary feedback for the navigation and localization of the robot during the operation. The dual-channel encoders give a resolution of 7 PPR and output a digital pulse. A gap of 30 mm is maintained between the two encoders. This is done to avoid interference between the two magnetic encoder disks (Miyashita et al. 1987).

Table 1 Motor specifications Motor Planetary-gear D.C. motor Hybrid stepper motor D.C. geared motor

Manufacturer Operation Orange Robot motion Electronics NEMA Stacker lift/ lower Johnson

Tray Transfer Module horizontal motion

Specifications 8.1 kg/cm, 40 RPM 4.4 kg/cm, 200 mm/revolution 150 RPM

Nos. 2 2 1

906

K. Warrier et al.

Table 2 Driver specifications Driver Cytron MDD-10A TB6560

For the motor Planetary-gear-type D.C. motors Stepper motors

L298N

D.C. geared motor

Specifications Motor voltage: 5–25 V, bidirectional control, NMOS H-Bridge, Peak: 30 A Operating voltage: 10–35 V, output current – 3A Supply voltage: 5–35 V, peak: 2A/bridge

Nos. 1 1 1

Fig. 8 Mobile robot electronic circuit

The Arduino Mega is used as the primary controller for the robot. Arduino’s external interrupts make it possible to capture each tick from the encoder accurately. The interrupt disables all other communications when triggered. Arduino also provides a sufficient number of digital I/O pins for the entire operation. Motors are driven by using suitable motor drivers, as described in Table 2. Speed control of the robot is also achieved to provide a smooth operation and precise positioning through pulse width modulation. Out of a total of 54 digital I/O pins on the Arduino, 15 can be used for pulse width modulation. The electronic schematic is shown in Fig. 8.

The Plant Tray Sensor Setup The tray used to grow the plant will have a sensor module that will house all the sensors required for the plant monitoring setup and the battery required to power the sensors and control module. The tray would also have a separate compartment to house the extra water leaving the soil. The PTSS is an integrated

34 Design and Development of Automated Vertical Farming Setup

907

Fig. 9 Electronic schematic of PTSS

unit developed to capture certain environment parameters surrounding the plant module. The setup consists of a soil hygrometer module and a digital temperature and humidity (DHT11) sensor, which measure the soil water content and the surrounding temperature and humidity, respectively. The digital temperature and humidity (DHT11) is a low-cost, small-sized sensor that uses a capacitive-type humidity sensor to measure the air nearby and outputs a digital signal on the data pin. New data is received every 2 s. It can measure humidity at a range from 20% to 80% and can measure temperatures from 0 ◦ C to 50 ◦ C. The moisture of the soil plays a crucial role in the sprinkler system. The soil hygrometer uses capacitance to gauge the soil’s water content and provides both analog and a digital output. A NodeMCU ESP8266 Wi-Fi development board is used to acquire data from the sensors and send the data over to the cloud storage and analytics module (Ubidots STEM). The DHT11 detects the humidity in the range of 20–80% and temperature from 0 ◦ C to 50 ◦ C. The sensors, microcontroller, and the battery pack for powering the setup are housed in a container of dimensions 120 × 50 × 15 mm, which is then embedded to the plant module/tray. The sensor probe of hygrometer protrudes outward to immerse into the soil and measure the humidity. In contrast, the digital temperature and humidity (DHT11) is made to stick outward the container. The schematic of the PTSS is shown in Fig. 9.

Communication Interface A communication interface is developed in order to interconnect the systems (VSS, mobile robot, PTSS) and also to have a single command control over the robot.

908

K. Warrier et al.

Fig. 10 Ubidots STEM graphical user interface (©Ubidots 2020)

For these purposes, a central computer is set up, accessed by the farm operator. The computer runs an analytics software online (Ubidots STEM) and also asserts the single command control. Commands relating to the distance to be travelled and the turns to be taken are not directly supplied to the robot during each operation. Rather the entire operation is embedded into a command consisting of alphanumeric codes, which is passed on to the robot wirelessly. Ubidots, under the STEM license, is a free-to-use cloud service and analytics tool which is commonly used for IoT applications, especially the ones surrounding agriculture, thus making it a suitable option for the work. The analytic tools include graphs, plots, set point creation, and many more. The data sent via the ESP8266 is stored on the cloud provided by Ubidots STEM and is henceforth plotted in real time for analysis. The tool allows limited free SMS service as part of alerts sent to the user once a defined set point is crossed. Thus, graphs are plotted separately for data received from digital temperature and humidity (DHT11) and from hygrometer. Set points for each data type are also created, and the SMS service is enabled for alerts. Ubidots STEM is used to acquire sensor data from the PTSS, as shown in Fig. 10.

Observations The mobile robot, VSS, and the Stacker are shown in Figs. 11, 12, and 13, respectively. The section covers the proposed arena and control algorithms, along with certain designs.

34 Design and Development of Automated Vertical Farming Setup

909

Fig. 11 Mobile robot

Arena A typical warehouse environment may consist of multiple storage racks or other elements in a staggered or a uniform distribution. Its purpose is to form the necessary action plan for robot movements, taking into account the available information about the environment. A robust control is required to manage traffic in a busy warehouse environment. The illustration of a sample proposed arena is shown in Fig. 14. The arena was set up in a laboratory, and it consists of a loading area for housing the plant modules and the central computer along with the operator. The operator’s location is at the loading area, shown in Fig. 14, where the tray is prepared initially by the operator and is then placed on the mobile robot. The robot’s home position is marked by “Mobile Robot” in Fig. 14. The loading area is situated exterior to the farm, and the operator is provided with a space for monitoring the farm at the same, such that manual interference during the farm operation can be avoided. The VSS is placed inside the farm. For a simplified sequence, the VSS is placed at a nominal distance, and the dashed lines mark the intended robot motion. A11, A12,

910 Fig. 12 VSS

Fig. 13 Stacker

K. Warrier et al.

34 Design and Development of Automated Vertical Farming Setup

911

Fig. 14 Proposed arena of operation

A21, A22, B11, B12, B21, and B22 represent the stop locations near the VSS that must be arrived at, in order to begin the transfer.

Sequence of Operation Interconnection of various systems in an indoor environment is the primary necessity for automation. Active exchange of data between these systems and the resulting decisions improve upon the efficiency. Cyber-physical systems are the main ordeal in an IoT-enabled workspace (Haris et al. 2019). A connection is established using Wi-Fi. Whereas a Wi-Fi module can be integrated with the microcontroller on the robot, the plant trays require a stand-alone and compact controller board providing wireless communication. The ESP8266 MCU is one such board that provides I/O pins to receive information from sensors and also can transmit them wirelessly. In order to visualize the operation that is to take place in the proposed arena, it can be categorized into two types: “Store” a plant module into a designated level on the VSS and “Retrieve” a plant module from a level on the VSS, as required by the operator. The operator has only a single control over the robot. In this way, it provides a more degree of automation. The Wi-Fi module on the robot receives information in terms of the stop location and the type of operation from the central computer, controlled by the operator. The stop locations are as specified in the arena, in Fig. 14. The overall sequence of operation is presented in Fig. 15. The robot is programmed to execute a sequence of motion to reach the desired level. The Stacker is also programmed to undergo specific motion vertically and horizontally to store or retrieve the module, as shown in Fig. 15.

912

K. Warrier et al.

Fig. 15 Sequence of operation

The sequence of operation is embedded onto barcodes. The barcode is a representation of typescripts and digits, explicitly readable by a machine, and for this work, an android app is additionally used, consisting of a provision to scan the barcodes and transfer the information onto the robot. Codes, each of length three, are developed containing specific information, one example shown in Fig. 16.

34 Design and Development of Automated Vertical Farming Setup

913

Fig. 16 Barcode containing the information for store operation

The barcode example shown depicts one of the operations, namely, “Store.” The first character and the second numeric represent the stop location intended, and the third character depicts the operation (S – store). All other commands pertaining to the operation are not manually handled by the operator, and these are pre-defined on the microcontroller. A specific strategy is developed for the control of the horizontal motion of the Stacker. This is one of the predominant criteria for the design of the Stacker and the roller channels on the VSS. As shown in Fig. 4, the arrangement of the roller channels on each column is made by providing a hollow space between them. Similarly, a hollow space is provided between the roller channels on the Stacker, which is later occupied by the Tray Transfer Module setup, as shown in Fig. 5. The width of this space is 80 mm. The plant module rests on these roller channels. As per the operation intended, the Tray Transfer Module is made to extend or retract between the VSS and the Stacker. The space between the roller channels enables the Tray Transfer Module to have full extension or retraction, along with the desired vertical motion. For the “Store” operation, the robot reaches the specified stop location on the arena. The Stacker is programmed to actuate vertically to the desired height corresponding to the desired level on the VSS. The Tray Transfer Module then undergoes complete extension horizontally to push the tray onto the VSS. Roller channels present on both the Stacker and VSS aid in the smooth transfer. Once transferred, the Stacker is set to actuate a few steps vertically downward. This step is crucial for the Tray Transfer Module to lose its grip on the plant module. The Tray Transfer Module then retracts, and the Stacker goes to its home position on the robot. For the “Retrieve” operation, once the robot reaches the desired stop location, the Stacker then undergoes vertical motion to reach the specified height level on the VSS and then actuates a few steps vertically downward. The Tray Transfer Module then extends its full length. The Stacker then actuates vertically upward, the same number of steps it actuated downward priorly. This sequence of motion confirms the grip of the Tray Transfer Module plates on the plant module, which is now ready to undergo transfer to the robot. The Tray Transfer Module retracts. With the aid of roller channels on either side, the plant module is comfortably transferred onto the robot.

3.3 Mobile Robot Navigation The kinematic equations of the mobile robot form the primary basis for its navigation. The equations help in developing a mathematical model of the robot that closely resembles the motion in the real world. Hence, the general model

914

K. Warrier et al.

and the computational model are described here. The general model is used for understanding and developing a simulation in MATLAB. The simulation is studied and analyzed before deployment in the real world. This helps capture the model behavior, given the set of inputs and obtaining its matched outputs. The computational model consists of a set of equations embedded into the controller that processes the captured data from the various sensors on board and makes the decision to plan the motion of the robot. The general model is relevant to a typical two-wheel differentially driven vehicle; however, the computational model is pertaining to this mobile robot, not explicitly.

Kinematic General Model A point P is selected on the robot as a reference to specify the position of the mobile robot. The coordinates of the point P are {x,y} with respect to the world frame {xw, yw }. The local frame is given by {xl ,yl }. The angle made by the local frame with respect to the world frame is given by θ, as shown in Fig. 17. The pose of the mobile robot is given by ⎡ ⎤ x ξl = ⎣ y ⎦ θ

Fig. 17 General representation for kinematic model

(4)

34 Design and Development of Automated Vertical Farming Setup

915

The mapping of the motion of the robot along the local frame with respect to the world frame is done using orthogonal rotation matrix given by ⎡

⎤ cos (θ ) sin (θ ) 0 R (θ ) = ⎣ sin (θ ) cos (θ ) 0 ⎦ 0 0 1

(5)

The angular velocities of the left wheel ωL and the right wheel ωR of the mobile robot are used to compute the overall linear and angular velocity of the mobile robot. The linear velocities Vr and Vl are given by   l (6) VR = ωR R + 2   l VL = ωL R − (7) 2 where l is the center distance between the wheels and R is the distance from the instantaneous center of curvature (ICC) Linear velocity υ =

ωR + ωL 2

Angular velocity ω =

VR − VL l

(8) (9)

The integration of these velocities can be used to compute the pose of the mobile robot:  Px = v (cos θ ) (10)  Py =

v (sin θ )

(11)

 Qz =

ω(dt)

(12)

where Px , Py are the x, y coordinates and Qz is the heading of robot. The designed CAD model from SOLIDWORKS is imported to MATLAB using Simscape Multibody. The software recognizes the revolute and linear joints automatically and will create a block model of the entire system, as shown in Fig. 19. Simulation of the kinematic model is done to simulate the robot behavior and response in a virtual world before doing so in real time. Inputs are provided in terms of the left and the right wheel angular velocities which are then converted to linear wheel velocities through multiplication with a pre-defined gain. The simulations were performed to understand and visualize the kinematic model of a mobile robot by applying different conditions. The mobile robot’s kinematic model is implemented using the MATLAB and SIMULINK software as shown in Figs. 18 and 19, respectively.

Fig. 18 Block representation of kinematic model in SIMULINK

916 K. Warrier et al.

34 Design and Development of Automated Vertical Farming Setup

917

Fig. 19 Simulated models of the mobile robot

Kinematic Computational Model The robot is very subtle to the individual wheel velocities – small changes in velocity given to individual wheel result in varied trajectories. The robot developed in the work operates based on proprioception, i.e., it uses its internal parameters as reference for calculating the pose. This methodology is known as dead reckoning. The robot rotates at its instantaneous center of curvature or ICC which is located at a distance of D from the center of its two wheels. As Vl , Vr , R, and ω are all time-dependent, the ICC can be computed using Eq. 13 (Dudek and Jenkin 2001):

ICC = (x − D sin θ, y + D cos θ)

(13)

918

K. Warrier et al.

At a forward time instant t + ∂t. The pose of the robot will be given by ⎡

⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ x cos (ω∂t) − sin (ω∂t) 0 x − ICCx I.C.Cx ⎣ y ⎦ = ⎣ sin (ω∂t) cos (ω∂t) 0 ⎦ ⎣ y − ICCy ⎦ + ⎣ I.C.Cy ⎦ 0 0 1 θ θ ω∂t

(14)

The mobile robot uses the encoder feedback as the primary source of measurement to compute the distance travelled, the turning angle, and navigation guidance (Siegwart and Nourbakhsh 2004). The incremental encoder consists of two outputs denoted as channel A and channel B. These are referred to as quadrature signals due to the fact that they are displaced by 90 electrical degrees. This allows for a position as well as direction calculation. The distance travelled is calculated for each wheel (wheelright dist, wheelleft dist ) by considering the wheel radius (r), right and left encoder ticks (ticksr , ticksl ), and total ticks per rotation (tickstotal ), shown in Eq. 17 (Cho et al. 2011):

wheelright dist =

ticksr × 2 × π × r tickstotal

(15)

wheellef t dist =

ticksl × 2 × π × r tickstotal

(16)

wheelright dist + wheel lef t dist 2

(17)

Total distance =

The angle (θ , heading) of the robot is calculated by

considering the distance travelled by each wheel per tick dist , dist , previously stored ticks l r ticks ticks

lef tprevtick , rightprevtick , current ticks (ticksr , ticks l ), and the distance between wheels (dwheels ), shown in Eq. 20 (Cho et al. 2011):

Sr = distrticks × ticksr − rightprevtick

(18)

Sl = distlticks × ticksl − lef tprevtick

(19)

 θ=

Sr − Sl dwheels

 ×

180 π

(20)

And the position of the robot is calculated by x = Sr × cos (θ )

(21)

34 Design and Development of Automated Vertical Farming Setup

919

Fig. 20 Overall control framework for mobile robot

y = Sl × sin (θ )

(22)

The motion control framework for the robot is shown in Fig. 20.

Quadrature Decoding Encoders help in measuring the position of the mobile robot. They count the number of pulses while the wheel is rotating. These types of encoders include the hall effect encoders, optical encoders, etc. In this work, the hall effect-type encoders are used as the primary source of information for the robot’s position and speed control. A magnetic disk is connected to the motor’s rotation shaft, which rotates with the motor and passes near the two hall effect sensors. When the motor turns the magnetic disk, the encoder provides an output as a digital pulse. The encoders have two channels (A and B). Each hall effect sensor produces one output. Hence, each encoder provides two outputs in the form of pulses per rotation (PPR). The hall effect sensors have a separation of 90◦ . The output of the sensors is hence 90◦ out of phase. This output is referred to as quadrature output. This provision of output being out of phase enables for determination of both magnitudes and direction, both of which are crucial for the robot. Quadrature decoding technique is employed to count the pulses and to determine the position and heading. The rising edges and falling edges count the pulses. There are three types of decoding methods X1, X2, and X4 decoding. X4 decoding is used in the work that counts the rising and falling edges of both the channels. This method is preferred as it provides better resolution for position estimation. Figure 21 shows the typical methodology used for X4 decoding (Negrea et al. 2012). For rotary encoders, the position is obtained by dividing the number of edges by the product of PPR and type of decoding:

920

K. Warrier et al.

Fig. 21 Typical X4 decoding procedure

Angular position =

Number of edges P ulses per rev. ∗ x

(23)

where x is the type of decoding Arduino’s ISR routine is called upon during the task of quadrature decoding. Arduino provides a set of specific pins (2, 3, 18, 19, 20, 21 on Mega) for interrupt routines. The channel A of both encoders is connected to Pin 2 and Pin 3. The ISR blocks all other processes happening in the controller when called so that Arduino never misses a tick or a count.

Robot Alignment Paradigm Apt orientation of the robot with the VSS is crucial to transfer the trays without the risk of it falling down. It was observed that a maximum distance of 20 mm could be allowed between the VSS and the robot to initiate a safe transfer; however, a distance in the range 0 − 10 mm is ideal. Also, the heading of the robot has to be in line with The VSS. It was observed that the robot, using only encoder feedback, maintained a distance more than the threshold, and also its heading was misaligned. A different methodology can be implemented to correct these errors. A secondary source of measurement for distance and orientation is introduced using a range sensor and a four-channel I.R. tracing module. Thin black strips, four in number, are attached at equal distance of 50 mm from each, at the front bottom end of each column in the VSS. Correspondingly, four I.R. transceiver modules are attached at the front bottom end of the robot, on its chassis, with each module maintaining a distance of 50 mm from each other, thus imitating the positions of the black strips on the VSS. The width of the column in robot is 170 mm, equal to the width of a single column of the VSS, as described in section “Systems Design.” The sensors are programmed to have a control over the robot motion once it approaches the VSS. At this point, the values from the encoder are not considered as feedback. The range sensor attempts to reduce the distance between the systems gradually, while the I.R. module acquires the reflected rays. The control algorithm is adopted and modified in deployment from line follower robots’ principle of operation (Maniha et al. 2011). The developed algorithm is shown in Fig. 23. An overall control framework for orientation is shown in Fig. 22.

34 Design and Development of Automated Vertical Farming Setup

921

Fig. 22 Robot orientation using I.R. modules

Fig. 23 Robot orientation algorithm using I.R. modules

Observation and Discussion In a typical indoor environment, it is not feasible to use GPS for localization. Several factors such as the block on the signal by the building and interference by concrete structures deliver poor results for localization using GPS. Thus, internal sensors are extensively relied upon for the purpose. These include encoders, gyroscopes, and accelerometers. Other sources may also include external beacons or landmarks or markers that can be accessed by onboard cameras. However, in situations where the abovementioned external sources are not used, the method of dead reckoning is deployed for localization. Dead reckoning estimates the robot’s pose. The primary sources of information for dead reckoning are wheel odometer sensors. Wheel odometry is an inexpensive and a direct method for computation; however, it makes a supposition that rotations of the wheel can be rendered to displacement

922

K. Warrier et al.

(Borenstein and Feng 1996). This supposition fails with only a limited validity majorly due to wheel slippage. Although there may be an availability for smooth shop floors in warehouses, yet a simple unevenness on the surface can result in deteriorations. Thus, though odometry sensors can provide both position and heading values, the error accumulated grows with time. Other contributing factors may include misalignment of wheels and futile wheelbase. Encoder errors such as inadequate sampling frequency and resolution could also be a reason (Borenstein and Feng 1996). The error in robot heading can be corrected through the use of gyroscopes. Gyroscopes are used for measuring angular velocity, essentially rotation, whereas accelerometers are used for measuring linear displacement. The common issue with accelerometers is their output must be integrated twice to calculate linear displacement. Thus, even a small drift in measurement rapidly raises into large position errors (Barshan and Durrant Whyte 1995).

Conclusion The work focuses on the need of incorporating a degree of automation onto the current practice methods employed in a vertical farm, through the use of robotics. Focal attention is given to the engineering design of the systems and not directly onto farming. Firstly, a VSS is explicitly designed that features roller channels that act as a base for housing plant modules and also suffice for smooth transfer of the plant modules in and out of VSS. Secondly, a mobile robot is designed such that it transports the plant modules within the farm through an ingenious platform setup on board named the Stacker. The Stacker is designed featuring roller channels to house the plant module on the robot and to guide while transferring. It also encompasses a separate setup called the Tray Transfer Module (TTM) which is actuated to push or pull the plant module. Apart from the aforementioned systems, additional work is done on the plant module that incorporated various sensors as a single box unit, embedded into the soil present in the module. These report certain environment and soil parameters to the operator wirelessly. The operator is provided with a central computer with an IoT analytics software (Ubidots STEM). The systems are connected via Wi-Fi with the computer. In the work, the operator decided upon the time to call the robot for storing or retrieving the plant module, based on the alerts received from the central computer. This is facilitated using a set of barcodes embedding information regarding various stop locations on the VSS. These barcodes are scanned using an android application, connected over Wi-Fi with the mobile robot that then transfers this information onto the robot to begin the operation. Simulated models of the robot are priorly developed and studied in order to capture its behavior in a virtual world. The general kinematic model was also used for developing the simulation. The computational model is also presented that helps capture the motion of the robot in the real world. Operation algorithm for robot operation in a proposed arena is discussed. The robot’s alignment with the VSS played a crucial role throughout the work, and it was understood that primary odometer sensors were not sufficient enough to properly align the robot

34 Design and Development of Automated Vertical Farming Setup

923

with the VSS, and hence a secondary algorithm involving the use of infrared sensors, which are commonly used in line follower robots, is proposed and presented. This algorithm successfully aided in orientation.

Future Scope With the increase in the contribution made by robots in every aspect of an industry, it is anticipated that these robots will also significantly improve the efficiency of agricultural output, in this case, the vertical farms. The work is aimed at handling various tasks involved in a vertical farm through the use of a mobile robot, automated storage system, and a plant monitoring sensor setup. The system is scalable to industrial standards as it is highly modular. It can be extended to fit a large-scale production unit with further capabilities such as increased storage space and growing various types of crops using innovative indoor farming methods. Automated watering and lighting systems can be implemented, receiving feedback from the plant sensor setup. Additionally, a control over these systems can be implemented, considering the type of soil being utilized for cultivations. Methodologies such as hydroponics or aeroponics can be implemented as a replacement for soil-based cultivation, which can then be stacked in vertical layers. The arena can be divided into an isolated, controlled environment and an exterior chamber for the operator. The isolated controlled environment consists of the VSS and the workspace for robot navigation, while the exterior chamber is a room with a central computer for the operator to monitor the farm. The operator stationed in this room can monitor the data provided by the plant sensor setup and can control the parameters (like temperature, humidity, soil moisture, etc.) inside the controlled environment by regulating the air-conditioning, lighting, and watering systems. The operator can trigger and send specific commands based on priority tasks, to the mobile robot wirelessly which is placed inside the controlled environment using human-machine interface (HMI) which is placed in an exterior room. The control over the plant growth and the surroundings can be amplified using various other sensors for monitoring, which can also be utilized to achieve a further autonomy over the robot. When certain conditions which are set by the operator are satisfied, the additional plant monitoring sensors can send a signal to the central computer, which can be used to communicate with the mobile robot, and the mobile robot autonomously navigates itself in the dynamic environment without any interference of the operator. The mobile robot can also be further developed to equip a vision system for smoother navigation. This vision system can include cameras that are placed on the mobile robot, which can monitor the plants from a close range. These visuals can be processed through complex machine learning and deep learning algorithms developed for recognition of the features of the plant such as color, shape, size, and infections to assess the growth and health of the plant and to detect diseases in plants. They can also help in decision-making by the mobile robot to retrieve a plant tray once the plants are ready. Every soil type has its own characteristics which might not be sufficient enough for plant growth, i.e., Tarai soil is rich in nitrogen and

924

K. Warrier et al.

organic matter but is deficient in phosphate. These need to be manually enhanced in a traditional farm. However, automated systems can be made use of in a vertical farm to balance the soil according to the crop’s necessity. This modularity can help grow a variety of crops in different soil types via autonomously enhancing the soil properties.

References Al-Kodmany K (2018) The vertical farm: a review of developments and implications for the Vertical City. MDPI J Buildings, USA Alter L (2010) The vertical farm: does it make sense? Tree Hugger, 22 November. http:// www.treehugger.com/culture/the-vertical-farm-does-it-make-sense-bookreview.html Auat Cheein FA, Carelli R (2013) Agricultural robotics: unmanned robotic service units in agricultural tasks. IEEE Ind Electron Mag7(3):48–58, Sept 2013. https://doi.org/10.1109/ MIE.2013.2252957 Barshan B, Durrant-Whyte HF (1995) Inertial navigation systems for mobile robots. In: IEEE Transactions on Robotics and Automation, 11(3):328–342. https://doi.org/10.1109/70.388775 Beacham AM, Vickers LH, Monaghan JM (2019) Vertical farming: a summary of approaches to growing skywards. J Hortic Sci Biotechnol:1–7. https://doi.org/10.1080/ 14620316.2019.1574214 Benke K, Tomkins B (2017) Future food-production systems: vertical farming and controlledenvironment agriculture. Sustain Sci Pract Policy 13(1):13–26. https://doi.org/10.1080/ 15487733.2017.1394054 Berg J, Gademann N (2000) Simulation study of an automated storage/retrieval system. Int J Prod Res 38:1339–1356. https://doi.org/10.1080/002075400188889 Borenstein J, Feng L (1996) Measurement and correction of systematic odometry errors in mobile robots. IEEE Trans Robot Autom 12(6):869–880. https://doi.org/10.1109/70.544770 Charron et al (1996) A system and methodology for measuring volatile organic compounds produced by hydroponic lettuce in a controlled environment. J Am Soc Hortic Sci 121(3):483– 487. Accessed Sep 7, 2020. https://doi.org/10.21273/JASHS.121.3.483 Chin Y, Audah L (2017) Vertical farming monitoring system using the internet of things (IoT). AIP Conference Proceedings. 1883. 020021. https://doi.org/10.1063/1.5002039 Cho B, Moon W, Seo W et al (2011) A dead reckoning localization system for mobile robots using inertial sensors and wheel revolution encoding. J Mech Sci Technol 25:2907–2917. https://doi.org/10.1007/s12206-011-0805-1 Cox S (2016) Enough with the vertical farming fantasies: there are still too many unanswered questions about the trendy practice. Alternet, 4 October. http://www.salon.com/2016/02/17/ enough_with_the_vertical_farming_partner David AV, Mohan HT, Bhavani RR (2017) Design and modelling of a cable driven cartrail robot for farm automation. In: 2017 international conference on intelligent computing, instrumentation and control technologies (ICICICT), Kannur, pp 1248–1253. https://doi.org/ 10.1109/ICICICT1.2017.8342748 Dudek G, Jenkin M (2001) Computational principles of mobile robotics. Ind Robot 28(1):80–81. https://doi.org/10.1108/ir.2001.28.1.80.3 FAO (2019) Agriculture and climate change – challenges and opportunities at the global and local level – collaboration on climate-smart agriculture. Rome, 52 pp. Licence: CC BY-NC-SA 3.0 IGO Fiorino L, Macillo V, Mazzolani FM (2014) Mechanical behaviour of bolt-channel joining technology for aluminium structures. Constr Build Mater 73:76–88. https://doi.org/10.1016/ j.conbuildmat.2014.09.086

34 Design and Development of Automated Vertical Farming Setup

925

Fu X, Zhang B, Yu H (2015) Rack design for an autonomous-steerable-vehicle storage and retrieval system of a warehouse. In: 2015 7th international conference on intelligent human-machine systems and cybernetics, Hangzhou, pp 112–116. https://doi.org/10.1109/IHMSC.2015.67 Gagliardi J, Renaud J, Ruiz A (2014) A simulation modeling framework for multiple-aisle automated storage and retrieval systems. J Intell Manuf 25:193–207. https://doi.org/10.1007/ s10845-012-0686-x Gawande SH, Shaikh SN (2014) Experimental investigations of noise control in planetary gear set by phasing. Hindawi Publishing Corporation, J Eng, Article ID 857462 Gonzalez-De-Santos P, Fernández R, Sepúlveda D, Navas E, Armada M (2020) Unmanned ground vehicles for smart farms. IntechOpen. https://www.doi.org/10.5772/intechopen.90683 Gopinath R (2014) Design of a power screw. Middle-East J Sci Res 20:630. https://doi.org/ 10.5829/idosi.mejsr.2014.20.05.77 Haris I, Fasching A, Punzenberger L, Grosu R (2019) CPS/IoT ecosystem: indoor vertical farming system. In: 2019 IEEE 23rd international symposium on consumer technologies (ISCT), Ancona, Italy, pp 47–52. https://doi.org/10.1109/ISCE.2019.8900974 Joseph L (2015) Learn robotics with Python. Packt Publishing, Birmingham Lauguico SC, Concepcion RS, Macasaet DD, Alejandrino JD, Bandala AA, Dadios EP (2019) Implementation of inverse kinematics for crop-harvesting robotic arm in vertical farming. In: 2019 IEEE international conference on cybernetics and intelligent systems (CIS) and IEEE conference on robotics, automation and mechatronics (RAM), Bangkok, Thailand, pp 298–303. https://doi.org/10.1109/CIS-RAM47153.2019.9095774 Massa G, Kim H, Wheeler R, Mitchell C (2008) Plant productivity in response to LED lighting. HortScience 43:1951–1956 MIT Technology Review (2018). https://www.technologyreview.com/2018/10/03/139937/newautonomous-farm-wants-to-produce-food-without-human-workers/ Miyashita K, Takahashi T, Yamanaka M (1987) Features of a magnetic rotary encoder. in IEEE Trans Magn 23(5):2182–2184. https://doi.org/10.1109/TMAG.1987.1065634 Mohammad Khasasi FH, Ali AMM, Mohd Yusof Z (2015) Development of an automated storage and retrieval system in dynamic industrial environment. In: 2015 international conference on BioSignal analysis, processing and systems (ICBAPS), Kuala Lumpur, pp 57–60. https:// doi.org/10.1109/ICBAPS.2015.7292218 Monteiro J, Barata J, Veloso M, Veloso L, Nunes J (2018) Towards Sustainable Digital Twins for Vertical Farming. In: 2018 Thirteenth international conference on digital information management (ICDIM), Berlin, Germany, pp 234–239. https://doi.org/10.1109/ICDIM.2018.8847169 Morrow R (2008) LED lighting in horticulture. HortScience 43:1947–1950 Mostofi A, Erfanian HR (2018) Multi-shuttle automated storage and retrieval system. Rev Ind Eng Lett 4:12–20. https://doi.org/10.18488/journal.71/2018.41.12.20 Nagendra Guptha CK, Bhaskar MG, Meghasree V (2018) Design of IoT architecture for order picking in a typical warehouse. In: 2018 3rd international conference on computational systems and information technology for sustainable solutions (CSITSS), Bengaluru, India, pp 50–53. https://doi.org/10.1109/CSITSS.2018.8768752 Negrea AC, Imecs M, Incze II, Pop A, Szabo C (2012) Error compensation methods in speed identification using incremental encoder. In: 2012 international conference and exposition on electrical and power engineering, pp 441–445. https://doi.org/10.1109/ICEPE.2012.6463857 Nor Maniha Abdul Ghani, Naim F, Yon TP (2011) Two wheels balancing robot with line following capability. Int J Mech Mechatronics Eng. World Academy of Science, Engineering and Technology, vol. 79, pp 634–638. Pawlowski A, Guzman JL, Rodríguez F, Berenguel M, Sánchez J, Dormido S (2009) Simulation of greenhouse climate monitoring and control with wireless sensor network and event-based control. Sensors 9:232–252 Rashid M, Siddique S, Motahar T (2019) Smart warehouse management system with RFID and cloud database. 218–222. https://doi.org/10.1109/ICIEV.2019.8858546 Siegwart R, Nourbakhsh IR (2004) Introduction to autonomous mobile robots. MIT Press, Cambridge, MA

926

K. Warrier et al.

Smit J, Nasr J (1992) Urban agriculture for sustainable cities: using wastes and idle land and water bodies as resources. Environ Urban 4(2):141–152. https://doi.org/10.1177/ 095624789200400214 United Nations, Department of Economic and Social Affairs, Population Division (2017). World Population Prospects: The 2017 Revision, Key Findings and Advance Tables. Working paper no. ESA/P/WP/248 Van Henten EJ, Hemming J, Tuijl B, Kornet J, Meuleman J, Bontsema J, van Os E (2002) An autonomous robot for harvesting cucumbers in greenhouses. Auton Robots 13:241–258. https://doi.org/10.1023/A:1020568125418 Zeidler C, Schubert D, Vrakking, V (2017) Vertical farm 2.0: designing an economically feasible vertical farm – a combined European Endeavor for Sustainable Urban Agriculture

Research Methodology for Augmenting a Gait Cycle of Lower-Body Exoskeleton, by Using a Data of Mathematical Modeling and Motion Study of a Specific User While Obtaining a Customized Gait for Joint Actuation of Exoskeleton

35

S. S. Ohol, K. D. Kalantri, Y. M. Pirjade, A. U. Kotkar, N. M. Patwardhan, D. R. Londhe, and T. P. Shelke Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gait Cycle Analysis and Torque Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Step I: Gait Cycle Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Step II: Torque Estimations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simulation of Two-Legged Robotic Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

928 929 935 935 937 943 944 947 949 950 950

Abstract

Exoskeleton technology has shown its importance in various fields of application such as military, medical, industrial, and commercial. But wide use of this technology is limited due to high cost and customized application according to user. If the exoskeleton has a different gait cycle than the unique natural gait cycle of user, it will lead to injuries. Customization of exoskeleton gait cycle can overcome this challenge. Customized gait cycle according to user means the control unit should provide output in form of position and torque in accordance to natural gait cycle of unique user. In this chapter, a methodology is proposed

S. S. Ohol () · K. D. Kalantri · Y. M. Pirjade · A. U. Kotkar · N. M. Patwardhan · D. R. Londhe · T. P. Shelke Department of Mechanical Engineering, College of Engineering Pune (COEP), Pune, India e-mail: [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_165-1

927

928

S. S Ohol et al.

and validated to obtain customized gait cycle for exoskeleton using motion assessment and MATLAB simulation. Validation of methodology is performed on normal human walking gait cycle. Motion assessment performed in HALEX (Human Assistive Lower-Limb Exoskeleton) is considered as basic method to conduct motion assessment of gait cycle for unique user and data obtained from their experiment considered as base data for position versus time analysis. Latter data obtained from motion assessment is used as input data to obtain results from MATLAB Simscape-Simulink simulations using genetic algorithm. The above method is applied on biped robot to obtain results for normal human gait analysis. Simulation gives results in both position and torque wrt time. Torque values are compared with values obtained from mathematical model, and position values are compared with values obtained from video gait analysis. Validation of methodology gave satisfactory results, and the latter suggested method can be used to obtain resulted torque and gait cycle for exoskeleton. Keywords

Biped robot · MATLAB-Simscape · Genetic algorithm · Exoskeleton · Normal human gait

Introduction If we observe the trend in every industry, we moved from humans to robots and now we are again moving toward humans. Yes, this is the future; that is right; and many limitations of humans and robot can be overcome with the combination of human and robot. This is the basic need of development of exoskeleton technology in field of robotics. A robot price rises due to necessary improvements in control flexibility, whereas repeated/monotonous work, dangerous work, and lesser strength for heavy duty task are not suitable for human workers. The combination of effective and precise control by human and strength of machines is required to get the solution for such challenge. Innovative developments of various exoskeletons using basic technologies of robotics were started in different sectors of applications while solving such type of challenges. Exoskeletons are operated by an electric motor, pneumatic, hydraulic, or a blend of technologies which permit for movement of limbs with enhanced power and perseverance. The exoskeletons are commonly known as powered armors, power armors, exo-frames, powered suits, and exo-suit, which are wearable mobile technology. Humanlike gait-motion can be obtained if links of the exoskeleton are applied with predefined position wrt time. Also, the pre-estimated position of various links wrt actuators is achieved by providing the actuators with preidentified torque related to specific time. During the research, a methodology/approach is proposed for establishing a required motion. Then it is authenticated to achieve personalized gait cycle motion for the exoskeleton using a software of video/motion assessment. The

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . .

929

methodology is validated using MATLAB simulation, performed on a sample of common human’s gait cycle for walking. Motion assessment performed on HumanAssistive Lower-Body Exoskeleton shall be studied as a basic technique to perform motion assessment of gait cycle used for any individual person. Then information achieved from such experiments can be considered as a reference data for posture and motion analysis study. Then this information obtained from motion assessment is employed as an input data to achieve required outcomes from simulation study using Simulink of MATLAB-Simscape by applying a genetic algorithm. Personalized gait pattern obtained accordingly by using these two steps methodology is proposed for further motion assistance. Video/motion analysis using motion assessment software can be a first phase of the procedure and analysis of data, which is obtained from motion assessment, and can be a second phase. This data submitted to X can execute simulation to find torque for real-time gait pattern required for precise motions of exoskeleton. A four-step experiment was performed to validate this methodology. Step-I is motion assessment for obtaining a required gait cycle. Step-II is a formulation of mathematical model for a human physique during plain walking for achieving a torque anticipated at different stages of normal human walking. A Step-III is a simulation for two-legged exoskeleton having a gait cycle feedback obtained from step-I, to acquire precise estimation of position and torque of the joint at relevant stages of time. A Step-IV is an assessment of outcomes of simulation obtained from motion assessment to authenticate the mathematical model and gait cycle which authenticate torque estimation required at several stages of walking.

Literature Review Research into the design and development of the exoskeleton is ongoing since the 1960s (Ref Fig. 1). The idea of human motion and strength enhancement through use of exoskeleton is something that has been frequently depicted in various sci-fi movies and novels. Applications of exoskeleton technology can be seen in various sectors such as military, industrial, domestic, medical, and commercial. Customized solutions/applications according to user and increased cost are the major limitations for current exoskeleton technology. A theoretical model of automated exoskeleton was presented by N. Yagn’s in 1890, and he was awarded a patent for his concept design of lower extremity enhancer (Yagn 1890). This concept model using a long bow could help people in improvising ability of the user for walking, jumping, and running. Different physical factors of users like gender, ages, height, weight, lifestyle, etc. affect human gait cycle. Therefore, customized gait cycle for exoskeleton is essentially required which can be obtained by a user-specific gait cycle and not a general one. While developing exoskeletons, the major concern was achieving the estimated position quickly, inside the permissible error to follow the trajectory of motion during normal walking. For use of improper control, a steady-state error because

930

S. S Ohol et al.

Fig. 1 Review of exoskeletons

of the gravity and additional considerations, such as a control unit, may contribute in opposition to any type of disagreements from the users/patients. Muscle spasticity of various paraplegic patients causes a struggle to match the movements of the actuators pressing an integrator term for reacting quickly. Therefore, a patient may get seriously harmed by overriding of the spasticity while moving the joint (DeLeva 1996). Such type of injury may adversely affect the common patient who is using active exoskeleton as exclusive gait of the person, which is due to overpowered exoskeleton. A method is needed to develop to get customized gait cycle from a specific user. This research studied a recorded motion to obtain gait cycle of specific user and a torque associated to a gait cycle for that specific user using results of MATLAB simulation. An inexpensive exoskeleton, Human Assistive (Active) Lower-Limb (/body) Exoskeleton (HALEX), is discussed by Y. M. Pirjade et al. to provide human assistance during gait locomotion (Pirjade et al. 2019). HALEX is a hybrid lowerlimb exoskeleton which provides enough DOFs to users while locomotion. HALEX design model as shown in Fig. 2 is a basic CAD design for simulation. Motion assessment conducted to study normal human gait for designing control for HALEX is considered as basic human gait for comparison with result obtained from simulation of biped robot. P. Naik and S. S. Ohol et al. have published this paper on full-body exoskeleton that they have designed and developed (Naik et al. 2019). This exoskeleton is full-body exoskeleton with combination of active-type upper body powered by pneumatic artificial muscle (PAM) while passive-type lower body provides support to wearer and exoskeleton to carry out tasks. This exoskeleton has the same DOF in lower body as the model exoskeleton considered for simulation. In this paper, they have used ANSYS for simulation of exoskeleton. Various load cases that were simulated in this paper are joint angle, height of rear plate from ground, and load and load center. Result of this paper also shows importance critical design of knee joint in lower body of exoskeleton, and one of the future scopes suggested that powered joints should be developed using advanced technologies like harmonic drives and planetary gearboxes to overcome mobility issue while carrying payload. Thus, we have considered HALEX exoskeleton model for simulation as its design has advance technologies like planetary gearbox for joint actuation. These results were useful to develop mathematical model for exoskeleton torque requirement.

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . .

931

Fig. 2 Human Assistive Lower-Limb Exoskeleton (HALEX) (Pirjade et al. 2019)

A big number of patients are restricted to wheelchairs owing to spinal cord injuries, which imposes a limited mobility. It may lead to the risk of secondary damages and injuries for those patients. Katherine A. Strausser et al. have presented a paper which states direct use of robotic exoskeleton on patients having spinal cord injury for gait restoration (Strausser et al. 2010). Exoskeleton, which is powered externally offers a perfect method to assist paraplegics to go on foot. A batterypowered mobile exoskeleton is introduced which has hydraulically activated hip and knee joint at the sagittal plane to support a motion of a related body part of the user. A standard human walking mimicked by control strategy stimulates position control of joints to understand average human walking trajectories as per the clinical gait analysis data. This paper shows the possibility of risk of secondary injury due to improper gait training which highlights the importance of actuation of exoskeleton gait with accordance of human gait motion and reliable input to control unit of actuators at joints. Results obtained are shown in Fig. 3. In a Strausser et al.’s paper, use of exoskeleton is mentioned for medical purpose for spinal cord injuries while a method is proposed which will be beneficial for all fields of exoskeleton as it talks about customization of gait cycle for all users. To apply the method suggested in this paper may not be appropriate, since the actual patient may not be able to undergo through motion assessment step. ( mainly due to physical limitations). Customized gait cycle for patients can be acquired from motion assessment data of a volunteer having the same physical parameters as the patient. Exoskeleton presented by Strausser et al. is a mobile, battery-powered device while the method suggested in this paper is a robust method for all types of exoskeletons irrespective of power source, actuation, control unit, and applications.

932

S. S Ohol et al. CGA Data 60

Joint Angle (deg)

50 40 30 20

Knee Hip

10 0 -10 -20 -30

0.0

0.2

0.6 0.4 Time (s)

0.8

1.0

Fig. 3 The range of a hip and a knee joint position required to attain normal human walking is being exhibited by the CGA trajectories joint angle (Strausser et al. 2010)

The main objective of Strausser et al. in their paper is to reduce secondary injuries to patients with spinal cord injuries whereas in the present paper main objective is to customize the gait cycle of exoskeleton, which will be useful for industrial applications also. Gait cycle analysis considered by Strausser et al. is a model based on four-step walking patterns of human walking. The four-step walking model can be understood as (i) stance, (ii) toe-off or late posture, (iii) swing, and (iv) heel-strike. The model used in this paper for gait analysis of human normal gait has eight states in walking pattern, i.e., (i) early contact, (ii) loading reply, (iii) mid stance, (iv) terminal stance, (v) preswing, (vi) starting swing, (vii) mid swing, and (viii) terminal swing. Strausser et al. have used clinical gait analysis data to generate trajectories for walking with some modification in angle at knee and step length to obtain improved clearance in walking. Whereas in this research, a motion analysis by using sport video/motion assessment software has been used to generate trajectories data for walking. Exoskeletons are useful in various sections of medical applications, e.g., a gait cycle rehabilitation and helping for strengthening of paraplegic patients. For a patient, who is using a treadmill, a gait-driven orthosis (DGO) was established that moves the legs in a physiological way. The orthosis can adjust as per size so dissimilar patients can practice it. For training the spasticity, the legs of patients can be trained with DGO with different degrees of paresis as shown in Fig. 4. Use of a driven gait orthosis (DGO) with treadmill training on patient of incomplete spinal cord injury to obtain physiological gait patterns was studied by G. Colombo et al. (2000). Similar to active lower-body exoskeleton, an actuator for knee and hip joints in this DGO was operated through a position controller. Results give three averaged trajectories of joint angle for hip and knee joints.

joint angle (Deg)

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . .

933

normal DGO DGO w. FC

25 15 5 -5 -15 0

10

30

40

50

60

70

80

90 100 % of stride

normal DGO DGO w. FC

70 joint angle (Deg)

20

60 50 40 30 20 10 0 0

10

20

30

40

50

60

70

80

90 100 % of stride

Fig. 4 Results of treadmill training for gait rehabilitation (Colombo et al. 2000)

Generally, all three trajectories are similar. The variations between these three forms do not surpass the inter-patient/subject deviations. The above gaits were used as reference to results obtained from simulation biped robot. The stance and swing stages of regular gait cycle of human being along with other substages can be understood in Fig. 5. A special video recorder along with a personal computer is used by James C Wall and Jack Crosbie to measure the sequential components of human gait for a healthy person’s walking as per self-chosen leisurely mode, intermediate mode, and faster walking mode. Their experiment provided evidence for establishing this method of time-based gait cycle examination, by means of a quality video motion recorder having a slow speed with playback facility which is reliable together and adequately precise for motion study of a gait cycle in a hospital (Walsh et al. 2007). A quasi-passive exoskeleton for lower limb can carry a weight while walking, which is presented by C. J. Walsh, K. Endow, and H. Herr (Berkeley robotics & human engineering laboratory). This exoskeleton is without actuators. It has only ankle joint springs and hip joint springs and knee joint variable damper. This exoskeleton has a total weight of 11.7 kg (Without a payload) and needs electrical power of only 02 Watts while walking with a load. They have proven it that the quasi-passive type of exoskeleton transmits a typical 80% of a total load in the

934

S. S Ohol et al.

RIGHT LEFT INITIAL PRE-SWING CONTACT

RIGHT LEFT INITIAL PRE-SWING CONTACT

RIGHT LEFT INITIAL PRE-SWING CONTACT

Time, percent of cycle

Double support

0%

R. Single support

15%

Double support

100%

60%

45%

R. Swing phase

R. Stance phase L. Swing phase 0%

Double support

L. Single support

L. Stance phase 40%

55%

85%

100%

Fig. 6 Ankle and hip moment for quasi-passive exoskeleton (Berkeley robotics & human engineering laboratory)

Ankle moment (Nm)

Fig. 5 Time percentage and stages of normal human gait cycle (Wall and Crosbie 1997)

0 -50 -100

Hip moment (Nm)

50

0

-50 0

50 % of gait cycle

100

direction of the floor while completing the sole support stage of walking, for a 36 kg payload. Gait analysis obtained is shown in Fig. 6. The BLEEX is an abbreviation of a Berkeley’s Lower Extremity Exoskeleton, which is a first actively autonomous/self-governing lower body exoskeleton having a capacity of transporting a payload, and was established at U.C. Berkeley (Zoss et al. 2005). Its mechanical design having anthropomorphic aspects has been presented by A. Zoss, H. Kazerooni, and A. Chu (Riener et al. 2002). A BLEEX possesses seven DOFs for each leg; four from these are operated by hydraulicly driven linear actuators. It has discussions on the finalizing of the ranges of motion and the degrees

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . .

935

Fig. 7 Human walking anatomy (Zoss et al. 2005) Hip Rotation (compliant)

Hip Abduction/ Adduction (actuated) Toe Flexion/Extension (compliant) Ankle Rotation (un-actuated)

Hip Flexion/Extension (actuated) Knee Flexion (actuated) Ankle Abduction/Adduction (compliant) Ankle Plantarflexion/ Dorsiflexion (actuated)

of freedom. Moreover, it has details of the important design aspects of the major BLEEX components. Three basic concepts of human body design architecture (Ref Fig. 7) were discussed in the research: (i) anthropomorphic architecture, (ii) nonanthropomorphic architecture, and (iii) pseudoanthropomorphic architecture. R. Riener et al. have mentioned that descending and ascending motions were the next stage of floor/ground walking; it means, the motion patterns of floor, i.e., level walking, are associated to the patterns of descent and ascent in a specific way (Riener et al. 2002). Along with the other important kinematic observations, it was also observed that power required is considerable at knee during ascending stairs; while descending stairs, power is absorbed at knee (Ref Fig. 8).

Gait Cycle Analysis and Torque Estimation Step I: Gait Cycle Analysis Analysis of gait cycle for normal human by video/motion assessment to derive angle position versus time plot during a gait cycle. Dartfish company provides video solutions (Ref Fig. 9) (Dartfish). It is based in Fribourg, Switzerland. This company makes online and offline video software to facilitate users to see, edit, and evaluate videos for personal and business usage. The changes in joint angles were recorded for a 21-year-old person using a treadmill for walking by means of a video camera which is fixed in a sagittal plane. These captured videos were analyzed for obtaining data of joint angle measurements using Dartfish Pro, motion assessment software. The video/motion examination technique is to be used for gait analysis because it must offer a specific gait cycle for a sole user which will be distinctly separate from normal gait cycle. Obtaining a tailored output gait for machine/exoskeleton corresponding to patients is an important stage in this total procedure. A unique gait cycle for each person takes care of several bodily factors such as height, weight,

936

S. S Ohol et al.

Fig. 8 Joint powers while descent and ascent are at minimum, normal, and maximum inclinations and while walking are of level walking type almost same for all average humans (Dartfish)

body postures, gender, age, etc. The motion analysis completed with a Dartfish sports analysis computer software is shown in Fig. 10 and 12. It shows readings of the angle of the knee and hip in relation to the time on the setup of the treadmill. The outcome with emphasized single cycle in total gait is shown in Figs. 11 and 13. A video setting of 25 fps, i.e., frames per second, is used, to analyze hip angle and knee angle. Figure 11 indicates a single gait cycle as shown in a rectangle. This measurement happens imitating the similar form over a specific period. It can be observed that angular measurements at the knee joint angle vary starting with 115◦ up to 180◦ , i.e., a range of 65◦ (Fig. 12). It can be observed from Fig. 13 that a single cycle shown in a rectangle is replicating the same pattern over a specific time. Angular measurements at the hip joint angle vary from 160◦ to 200◦ , i.e., for a range of 40◦ .

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . .

937

Fig. 9 Screen of an edit – Dartfish 10 (Dartfish)

As discussed in Fig. 6, a reference gait cycle for average size person is known to us. This gait cycle for typical common person is a source to confirm motion analysis process to find distinctive gait of that unique person. Knee and hip angle measurement positions are shown in a graph (refer Fig. 8). A form of angular position trends displays inverted shape and begins in the vicinity of zero degree on account of variations in reference angle for measurement. Thus, the hip joint ranges from −22◦ to 19◦ , i.e., 41◦ and similarly angular measurements at the knee joint are from 0◦ to 55◦ , i.e., for a range of 55◦ .

Step II: Torque Estimations A torque needed for normal human walking with several phases of standard human gait needs to be calculated using mathematical model. To obtain the torque involved at different stages from generating a mathematical model of human body, the following statistics of human height and weight is utilized as per Tables 1 and 2. The above data is used to obtain a mathematical model for an average human with height of 180 cm and weight of 80 kg. It is observed from a motion analysis that for normal human walking an average gait cycle time is 1 s and length of stride is 60 cm. The torque required for critical stages in a gait cycle is analyzed. The

938

S. S Ohol et al.

Fig. 10 Motion analysis for getting data of various angles of knee wrt time (Pirjade et al. 2019)

required torque at the ankle for a complete gait cycle is insignificant, which can be observed in modeling. The following formulas are derived to achieve different torque values at these secondary levels.

Hip and Knee Torque Calculations for Stance Phase Maximum torque capacity is required at human joint compared to human walking torque. Adequate friction between floor and human foot is necessary to make human

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . .

939

Fig. 11 Knee angle measurements – a single gait cycle is shown in a rectangle (Pirjade et al. 2019) Fig. 12 Motion analysis for getting a data of various angles of hip wrt time (Pirjade et al. 2019)

walk with comfort. In calculation of torque phase, friction has an important role. During stance stage, vertical and horizontal forces are mutually applied upon joints. Ave. step distance = 60 cm = 0.6 m Ave. period of single normal gait = 1 s

940

S. S Ohol et al.

Fig. 13 Hip angle measurements – a single gait cycle form is shown in a rectangle (Yagn 1890) Table 1 Weight distribution in human body (DeLeva 1996) Mass division of different body parts in human body Sr. No. Parts Term Average portion (%) Total weight M 100.00 1 Skull Mh 06.81 2 Trunk Mt 43.02 3 Complete arm Ma 4.715 4 Upper arm Mua 2.630 5 Forearm Mfa 1.500 6 Hand Mh 0.585 7 Total leg Ml 20.37 8 Thigh Mth 14.47 9 Shank Msh 4.570 10 Foot Mf 1.330

  s = u × t + a × t2 /2 0.6 = 0 + (a × 1 × 1) /2 a = 1.2 m/s2 Complete human body mass is M = 80 kg As we know, Force = F = M × a = 80 × 1.20 = 96.00 N

Average mass (kg) 80.00 05.45 34.40 03.77 02.10 01.20 0.47 11.58 11.58 03.66 01.06

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . .

941

Table 2 Height distributions in human body (DeLeva 1996) Height division of different body parts in human body Sr. No. Parts Terms Average portion (%) Total H 100.00 1 Whole leg L 56.20 2 Thigh L1 28.10 3 Shank L2 28.10 4 Foot L3 14.05 5 Full arm La 39.30 6 Head Lh 15.70 7 Torso Lt 28.10

Average height (cms) 180.00 101.16 50.580 50.580 25.290 70.740 28.260 50.580

Fig. 14 A subphase “mid-stance” of typical human gait cycle

Inertial force = I = F = 96.00 N μ = Coeff.of friction = 0.50 N = Standard regular force required = M × g = 80.00 × 9.81 = 784.80 N Fmax = Required maximum force of friction = N × u = 0.50×784.80 = 392.00 N Fc = required force of friction Friction force is a force to be applied on the ground by human, during walking phase. Therefore, maximum available friction force (Fmax) should be more than required friction force (Fc). A mid stance is the important subphase of stance stage in a complete gait cycle with maximum torque requirement (Fig. 14). As per Fig. 13, the following can be the moment acting on the body

942

S. S Ohol et al.

Resultant friction acting on human body = Frc = Fc Resultant moment at knee at hip due to torque = Mrk Resultant moment of Fc at a knee = Mrc Resultant moment of Fc at a knee = Mrc Force necessary for walking at definite speed = F Torque necessary at knee by actuator = Mk Torque necessary at knee b = Mk

Mh = F × 0.30 Nm = 96 × 0.30 = 28.80Nm Mk = Mh + F × 0.80 = 28.80 + 96 × 0.80 = 105.60 Nm Mk = Mrc = Frc × 0.50 = 105.60 Nm Frc = 212 N Frc = Fc = 212 is less than Fmc = 392 N We got, Mh = 28.80 Nm for a hip joint and Mk = 105.60 Nm for a knee joint

Hip and Knee Torque Calculations at Swing Stage A pendulum free-type swing motion is observed for the swing phase of legs. Calculating a torque is required for swing phase which is different torque calculation wrt stance phase. Various torque calculations support that terminal swing is an important stage through maximum torque necessary during a swing stage (Fig. 15). For hip – an equation of moment is

Fig. 15 Terminal swing substages for a common human gait cycle

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . .

943

Mo1 = L1 × 0.50 × sin30 × Mth × g + Msh × g × [L1 + L2 × 0.50] × sin30 + Mfo × g × [(L1 + L2 ) × sin30 + L3 × cos30 × 0.50] Mo1 = 50.58 × 0.50 × 0.50 × 11.58 × 9.81 + 3.66 × 9.81 × [50.58 + 50.58 × 0.50] × 0.50 + 1.06 × 9.81 × [(50.58 + 50.58) × 0.50 + 25.29 × 0.86 × 0.50] Mo1 = 1436.50 + 1362 + 639 Ncm Mo1 = 34.40 Nm As we know, frictional forces are neglected in the calculation of moment and a foot cannot be in contact with a normal force acting on the ground. For knee – a moment equation is Mo2 = Msh ×g×L2 ×0.50×sin30 + Mfo × g × [L2 × sin30 + L3 × 0.5 × cos30] Mo2 = 454 + 376 Ncm Mo2 = 8.30 Nm We get Mh = 34.40 Nm for hip joint and Mk = 8.30 Nm for knee joint

Simulation of Two-Legged Robotic Structure A modest stick and brick form of two-legged robot-walking body structure is available in MATLAB. Simulation utilizing Simscape can be used to evaluate results with motion assessment and torque calculated for this system (The MathWorks 2020; The Global Optimization 2020). Overall, 10,000 trials are performed on the two-legged robot structure to understand human walking using a genetic-algorithm with 100 generation and 100 population size. Results can be analyzed according to primary condition provided at the start of the trials, for genetic-algorithm (Nobile et al. 2011) (Fig. 16). Fig. 16 A simple stick and brick model for two-legged robot system (Nobile et al. 2011)

Hip joint

Knee joint

Ankle joint

944

S. S Ohol et al.

For two-legged robot system, the following are the parameters: Length of thigh = 400 mm. Length of shank = 400 mm. System density = 1000 kg/m3 ; it is known that human body density = 980 kg/m3 . Total weight = 5.588 kg. The above two-legged model has six DOFs to each leg: (i) hip roll, (ii) hip yaw, (iii) hip pitch, (iv) knee, (v) ankle pitch, and (vi) ankle roll, same as that of normal human anatomy required for walking.

Results and Discussion MATLAB simulation of biped robot results obtained is as below (A) The hip roll

hiproll

Joint Angle [rad]

0

-0.05

-0.1

Joint Torque [Nˆm]

-0.15 0

1

2

3

4

0

1

2

3

4

5 Time [s]

6

7

8

9

10

5

6

7

8

9

10

50

0

-50

Time [s]

Fig. 17 The MATLAB scores of joint angle(rad) and joint torque(Nm) wrt time(s) of hip-roll

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . . anklepitch

0.8 Joint Angle [rad]

945

0.6 0.4 0.2 0

0

1

2

3

4

5

6

7

8

9

10

6

7

8

9

10

Joint Torque [Nˆm]

Time [s]

50

0

-50 0

1

2

3

4

5 Time [s]

Fig. 18 The MATLAB results of the joint angle(rad) and joint torque(Nm) wrt time(s) of hip pitch

The torque required at hip roll and the angular position can be obtained from Fig. 17. A single cycle in a gait cycle is shown in the frame. It is observed that a difference of overall angle all over the noted cycle is −0.15 rad to 0 rad, i.e., a range is 8.6◦ . Also, it is observed that a range of torque and maximum torque at the joint is 0 Nm to 20 Nm and 20 Nm, respectively. (B) The hip pitch The torque required at hip pitch and the angular position can be obtained from Fig. 18. A single cycle in a gait is shown in a frame. A difference of entire angle during the cycle is observed as 0.30 rad to 0.80 rad, i.e., a range is 28.60◦ . Also, it is observed that a range of torque with a maximum joint torque is −15 Nm to 20 Nm and 20 Nm, respectively. (C) The hip yaw The torque required at hip yaw and the angular position can be obtained from Fig. 19. A single cycle in a gait is shown in box. It is noticed that the range of joint

946

S. S Ohol et al.

Joint Angle [rad]

0

x 10

-20

hipyaw

-0.5 -1 -1.5 -2 0

1

2

3

4

5

6

7

8

9

10

6

7

8

9

10

Joint Torque [Nˆm]

Time [s]

50

0

-50 0

1

2

3

4

5 Time [s]

Fig. 19 The MATLAB results of joint angle(rad) and joint torque(Nm) wrt time(s) of hip yaw

angle and torque needed is negligibly small. Thus, a role of hip yaw for deciding a gait cycle for the torque required to a standard human walking is insignificant. (D) The knee The torque required at knee joint and the angular position can be obtained from Fig. 20. A difference of overall angle during a cycle is −1.00 rad to −1.50 rad, i.e., a range is 28.60◦ . Also, it is observed that a range and maximum joint torque is 0 Nm to 45 Nm and 45 Nm, respectively. (E) The ankle pitch The torque required at ankle pitch and the angular position can be obtained from Fig. 21. A single series in a gait is shown in a frame. The difference of complete angle during the cycle is 0.40 rad to 0.80 rad; means a range is 22.90◦ . Complete series and maximum joint torque obtained are −10 Nm to 20 Nm and 20 Nm individually.

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . . knee

0 Joint Angle [rad]

947

Single cycle -0.5

-1

Joint Torque [Nˆm]

-1.5 0

1

2

3

4

5 Time [s]

6

7

8

9

10

0

1

2

3

4

5 Time [s]

6

7

8

9

10

50

0

-50

Fig. 20 The MATLAB results of joint angle(rad) and joint torque(Nm) wrt time(s) of the knee

(F) The ankle roll The torque required at ankle roll and the angular position can be obtained from Fig. 22. A particular cycle in a gait is shown in a rectangle. The difference of complete angle all through the gait cycle is 0 rad to 0.15 rad, i.e., it is 8.60◦ . Total range and maximum joint torque are −10 Nm to 20 Nm and 20 Nm, respectively.

Summary of Results 1. The result obtained from simulations using MATLAB Simscape together with genetic-algorithm can offer precise estimations of humanlike gait cycle for working of exoskeletons or two-legged robots. 2. Motion assessment results show that a human gait obtained with an experiment using a treadmill setup (refer Fig. 9) found similarity corresponding to the form of reference gait cycle (refer Fig. 6), for a hip angle of a gait utilizing motion assessment corresponds precisely to the source gait in terms of its form and range. A knee angle of gait cycle for motion assessment is similar to a form of reference gait but in inverted manner. This is due to various reference angles

948

S. S Ohol et al. ankle pitch

Joint Angle [rad]

0.8 0.6 0.4 0.2 0 0

1

2

3

4

5

6

7

8

9

10

6

7

8

9

10

Joint Torque [Nˆm]

Time [s]

50

0

-50

0

1

2

3

4

5 Time [s]

Fig. 21 The MATLAB results for joint angle(rad) and joint torque(Nm) wrt time(s) of ankle pitch

3.

4. 5.

6.

7.

for the amount of several motion assessment and reference gaits. The range is 10◦ less than reference gait for knee angle. Different persons have different gait cycles according to their natural human parameters like gender, age, height, weight, etc. so it is observed that gait cycle obtained from motion assessment is not matching to a whole average reference gait cycle. Hip pitch angle of two-legged robotic system’s gait cycle matches with motion assessment of gait cycle although the range of angle is 12◦ less. Knee angle of gait sequence of two-legged robot system is identical to reference gait cycle at starting and terminal stages of motion assessment gait cycle, but gait cycle of two-legged robot is flattened at mid swing stage. So, the range of the angle varies for 30◦ . Now since mathematical modeling of human average walking is known, a joint torque necessary at knee at a midstance stage is 105.60 Nm. Also, a max torque needed for two-legged robot system is 45 Nm at mid stance stage. It is also observed that the torque for knee joint is more than that of the total joint torque required at the hip joint during the gait cycle for mathematical modeling and for two-legged robot simulations.

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . . ankleroll

0.15 Joint Angle [rad]

949

0.1

0.05

Joint Torque [Nˆm]

0 0

1

2

3

4

5 Time [s]

6

7

8

9

10

0

1

2

3

4

5 Time [s]

6

7

8

9

10

50

0

-50

Fig. 22 The MATLAB results of joint angle(rad) and joint torque(Nm) related to time(s) of ankle roll

8. It was observed with mathematical modeling that an estimated torque required at ankle is insignificant; nevertheless, simulation results show that considerable torque is required at ankle.

Future Scope 1. Using the suggested method and performing simulation on exoskeleton, this method can be further checked. 2. Results of simulation can be improved by improving material and joint properties of biped robot system, i.e., modeling them more humanlike material and soft joints. 3. Simulations results can be improved by using exactly human-like two-legged robot as a replacement for simple stick and brick illustration, so that the weight allocation can be obtained similarly like an actual human body. 4. Improvement in the results could be achieved by properly designing a more detailed cost function and permitting more numbers of iterations to analysis.

950

S. S Ohol et al.

Conclusion 1. All the results obtained from motion/video analysis are highly satisfactory for different test subjects and can be used to achieve customized gait cycle to be used for exoskeleton for a specific user. 2. MATLAB Simscape simulations with genetic-algorithm shall be employed to establish gait walking similar to a normal human being, to model human body motions while the user is wearing an exoskeleton. 3. Simulation results obtained for Torque match to the results obtained for torque from a mathematical modeling of a normal human walking. 4. From the above approach, generating a personalized gait cycle for an exoskeleton developed using motion assessment of specific person and MATLAB simulation can be applied to establish and generate a data of torque values required at each of the joint actuator. Acknowledgments Author would like to thank Dr. B.B. Ahuja, Director, College of Engineering Pune (COEP), for his guidance and support. Also, sincerely thankful to members of Robot Study Circle, COEP and FABLab of MEIM Department, COEP.

References Berkeley robotics & human engineering laboratory, “BLEEX”, [Online]. Available: https:// bleex.me.berkeley.edu/research/exoskeleton/bleex/ Colombo G et al (2000) Treadmill training of paraplegic patients using a robotic orthosis. J Rehabil Res Dev 37(6):693 Dartfish. https://support.dartfish.tv/en/support/solutions/articles/27000049040-dartfish-10-releasenotes DeLeva P (1996) Adjustments to Zatsiorsky-Seluyanov’s segment inertia parameters. J Biomech 29(9):1223–1230 Naik P, Unde J, Darekar B, Ohol SS (2019) Pneumatic artificial muscle powered exoskeleton. In: International conference on advances in robotics, AIR-2019, IIT Madras, Article No.: 33, pp 1–7 Nobile M et al (2011) Further evidence of complex motor dysfunction in drug naïve children with autism using automatic motion analysis of gait. Autism Int J Res Pract 15:263–283 Pirjade YM, Kotkar AU, Patwardhan NM, Londhe DR, Shelke TP, Ohol SS (2019) Human assistive lower limb exoskeleton. Asian J Converg Technol 5(2). ISSN No: 2350-1146, I.F-5.11 Riener R, Rabuffetti M, Frigo C (2002) Stair ascent and descent at different inclinations. Gait Posture 15:32–44 Strausser KA et al (2010) Prototype medical exoskeleton for paraplegic mobility: first experimental results. In: Proceedings of the ASME 2010 dynamic systems and control conference, Cambridge, MA The Global Optimization toolbox from MathWorks website (2020) online https:// www.mathworks.com/products/global-optimization.html The MathWorks website (2020) [online]. https://www.mathworks.com/help/reinforcementlearning/ug/train-biped-robot-to-walk-using-reinforcement-learning-agents.html;jsessionid=fc0 f8c0aa4292db7f004f094bba6 Wall JC, Crosbie J (1997) Temporal gait analysis using slow motion video and a personal computer. Physiotherapy 83(3):109–115

35 Research Methodology for Augmenting a Gait Cycle of Lower-Body. . .

951

Walsh CJ, Endow K, Herr H (2007) A quasi-passive leg exoskeleton for load-carrying augmentation. Int J Humanoid Robot 4(3):487–506 Yagn N (1890) Apparatus for facilitating walking, running, and jumping. US Patents 420179. [Online]. Available: https://patents.google.com/patent/US420179A/en Zoss A, Kazerooni H, Chu A (2005) The mechanical design of the Berkeley lower extremity exoskeleton (BLEEX). In: IEEE/RSJ international conference on intelligent robots and systems, Edmonton

Intelligent, Automated, and Web Application-Based Cradle Monitoring System

36

Priyanka J. Nair and V. Ravi

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Literature Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Working of the Cradle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Block Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Algorithm and Working of Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Intelligent Cradle Monitoring System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Web Portal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Live Streaming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

954 955 956 956 958 962 962 963 964 964 965 965

Abstract

Studies prove that infant cries are acoustically unique, which provides a wide area and opportunity for automation in the field of baby care. Infants require attention all through the day and night, which is practically impossible for parents to provide. The Intelligent Cradle Monitoring System is a step towards efficient baby care. Our idea is to harness this to create a system to soothe the baby upon hearing the baby cry. With features like automatic cradle movement, wet mattress detection, and the possibility to control and monitor the cradle through Internet, while viewing the livestream of the cradle with the baby, we believe that the

P. J. Nair Mercedes-Benz Research and Development, Bengaluru, India V. Ravi () School of Electronics Engineering, Vellore Institute of Technology, Chennai, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_166

953

954

P. J. Nair and V. Ravi

Intelligent Cradle Monitoring System is a step towards efficient baby care. The system alerts the parents through text or call via the Internet when the baby has wet the mattress or has been crying for a long period of time. Keywords

General-purpose input/output · Microcontroller unit · Internet protocol · Integrated development

Introduction With an increase in the number of working parents, the need and demand for child care is also increasing. India is host to a fair amount of child care centers and this proves that there is a market for a system providing better and efficient child care with less human intervention. The Intelligent Cradle Monitoring System allows the parent to be aware of the baby and its activities without being physically present near them. This system also finds itself useful in hospitals, helping nurses take care of multiple babies in maternity wards. Figure 1 presents the recent development in the baby monitoring and automated cradle system.

Fig. 1 Recent development in baby monitoring and automated cradle system

36 Intelligent, Automated, and Web Application-Based Cradle Monitoring System

955

The proposed system has a microphone that detects the baby’s cry that triggers the cradle to sway for an appropriate amount of time to let the baby stop crying. If the baby does not stop crying despite the automatic cradle movement, an alert is sent to the parents or guardian in the form of a call. The parents can then control the cradle through the web portal while logging on to the live streaming session of the baby and the cradle. The parents are also alerted when the baby wets the mattress, in which case human intervention is required. The objectives of this proposed model are automatic cradle movement upon detection of baby cry, live streaming of the cradle via IP Webcam, alert parents via call if the baby cries for a long time or wets the mattress, and allow parents to access the cradle at any given point of time through a web portal. Thus, the Intelligent Cradle Monitoring System is a step towards helping parents take care of their babies in a more efficient manner.

Literature Survey Anjikar (Anjikar et al. 2017) designed a smart bassinet that used motion and surface body detection to detect the baby’s cry and automatically swing the cradle till the baby stops crying. However, an alarm is used to alert the parents, which produces a sound that further disturbs the baby. This is not the case in our research, where a call is being sent to the parents through IoT, thus allowing the baby to sleep peacefully. Kadu (Kadu et al. 2014) presented the mechanical view of swinging the cradle without human intervention upon bed wetting by the baby. This chapter served as a reference for the mechanical connections of the cradle. However, not many sensors are used for monitoring (Jabbar et al. 2019) the baby and there is no provision to alert the parents. The Intelligent Cradle Management System alerts the parents via call when the baby cries or wets the mattress. Rachana Palaskar (Palaskar et al. 2015) introduced the use of Internet to help parents/guardians constantly monitor the baby’s sleep schedule and surrounding parameters. Though the status of the cradle is shown in the web portal, the parents do not have the access to control the cradle or view the cradle, both of which are features of the Intelligent Cradle Monitoring System. Nitin Bhatnagar (Bhatnagar et al. 2016) worked on swinging the cradle automatically upon detection of baby’s cry or wetting of the mattress. The system uses an alarm for alerting the parents, which limits the range of connectivity and also disturbs the baby. Our research work uses the TI CC3200 microcontroller, which has an inbuilt Internet module, allowing for easy access to the Internet, thus extending the range of connectivity. Misha Goyal (Goyal and Kumar 2013) focused on an automatic E–Baby Cradle that automatically sways the cradle based on baby’s cry patterns. The Intelligent Cradle Monitoring System provides parents with the ability to view the cradle and control the same through Internet, in addition to the automatic swinging of the cradle.

956

P. J. Nair and V. Ravi

Working of the Cradle Block Diagram In Fig. 2, the inputs – microphone, camera, and wet sensor and the output – motor driver connected to the motor, which in turn sways the cradle are shown in Fig. 2. The microcontroller used is TI CC3200 LaunchXL. The connections of the inputs and outputs to the controller CC3200 are shown in the block diagram. The microphone’s audio pin is connected to the GPIO Pin 7, wet sensor to GPIO Pin 6, Motor Driver to Pins 1 and 2, and the camera module through cloud. The motor driver is then connected to the motor, which further sways the cradle, upon being triggered by the baby’s cry which is recorded through the microphone. In Fig. 3, the flow chart of the entire system is demonstrated. The system once switched on, checks for Internet access and once connected, shares the port address for access to the videos and controls (Mitra et al. 2015). If the client logs on to the web site, then the system waits for an input from the parents and then performs the

Fig. 2 Block diagram of the system

36 Intelligent, Automated, and Web Application-Based Cradle Monitoring System

957

Fig. 3 Flow chart of the system

respective actions. The step by step process is shown in Fig. 4. Else, if the client has not logged on, then the system acts on the sensor inputs. Initially, the timer is reset to allow for a timely way of monitoring (Kumar and Ravi 2019) the baby. Once a cry is heard, the input is compared with the threshold and if greater, then the time is compared to the threshold time. If both conditions are satisfied, the cradle is rocked. The automatic movement (Joshi and Mehetre 2017; Nejkar et al. 2018; Kavitha et al. 2019) of the cradle is described in detail in Fig. 5. If a sensor input is greater than the threshold, then the alarm module is triggered and the status of the cradle is updated in the web portal. The process of sending an alert to the parents is explained in Fig. 6. The response action was observed within range of 1–2 min of the trigger being observed. The sensor failure is detected through plausibility. The input from the two sensors detecting the same value is verified and compared, and sensor failure or mismatch is detected. The microcontroller used is the TI CC3200 Simple link Wi-Fi LaunchPad – single-chip microcontroller unit (MCU) with built-in Wi-Fi connectivity.

958

P. J. Nair and V. Ravi

Fig. 4 Flow chart of Module 1

For ease of design, the entire system has been split into three modules as explained below. 1. Automatic cradle movement. 2. Live streaming and parental controls. 3. Call alert through the cloud.

Algorithm and Working of Modules Automatic Cradle Movement Figure 4 represents Module 1 which focuses on cradle movement upon detecting baby cry. A LM393, electret microphone, is used to continuously sense the surroundings till a voice is heard. Once an input is sensed, the timer is being set

36 Intelligent, Automated, and Web Application-Based Cradle Monitoring System

959

Fig. 5 Flow chart of Module 2

and the input is compared to the threshold value. If the input is not greater than the threshold, then the microphone continues to sense till another input is heard. Else, if the input is found to be greater than the threshold and the duration has not yet reached the fixed limit, the cradle is rocked and a warning alert is sent to the parents via cloud. If the duration has exceeded the fixed threshold, then the cradle is stopped and the alarm module is triggered. If the baby is not crying, the system will continue to monitor the inputs every 5 minutes (which can be configured by the parents). Similarly, the baby cry is differentiated from other sounds by using appropriate audio filters. A 12 V DC gear motor with 60, connected to the motor driver, L293D is used to sway the cradle. The threshold value is calculated thus:

threshold = 20 log (Vin/Vo) dB

960

P. J. Nair and V. Ravi

Fig. 6 Flow chart of Module 3

where Vin = voltage value when the baby is crying Vo = average reference voltage value when the baby is not crying.

Live Streaming and Parental Controls A feature to allow parents to be able to view their baby and access the cradle accordingly through a web portal (Rajesh et al. 2014; Mishra 2018; Hussain et al. 2019; Choi et al. 2020). An IP webcam module, which is basically the camera of any smart phone or a webcam, is set up at the cradle which records the video and

36 Intelligent, Automated, and Web Application-Based Cradle Monitoring System

961

streams it online. IP Webcam turns your phone into a network camera with multiple viewing options (Lu et al. 2018). The parent can view the baby through the web page mentioned in addition to selecting the mode of operation for the cradle – manual or automated. The processor used for the prototype is of CC3200 LaunchXL which has both application and network stack on one processor which allows easy accessibility over cloud (Mnati et al. 2017). The CC3200 uses files specific to TLS/SSL that may be defined by the user at the application level. The automated mode basically switches off the system except for the live webcam, so that the parent can quit the mode when they wish to do so. The manual mode further has two submodes: one which works with the sensor readings and another which works based on the instructions provided by the parent. Figure 5 shows the flow of actions for Module 2, which focuses on live streaming and parental controls aspect for the cradle. The program is coded in such a way that the port address for access to the live streaming of the cradle and the controls of the same are shared with the parents, every time the program is run. The services of Energia IDE, a rapid prototyping platform for the Texas Instruments MCU Launchpad, are used to code for the system (Patil et al. 2018). Once the user logs on to the site, the board acknowledges that the client is connected and allows the parents to access the cradle. The parents or guardian can select either automated mode, where the cradle sleeps, or they can select the manual mode where the cradle functions based on the parent’s inputs or the sensor-based readings. The cradle is said to be in a particular mode until an exit message is invoked.

Call Alert Through the Cloud If the baby cries for a long time or wets the parents, a call is immediately sent to the parent alerting the parent about the event that triggered the alarm. The cradle is swayed and the status is updated on the web portal. The wet sensor also senses for an input continuously. If the sensed input is found to be greater than the set threshold value, then the cradle is swayed while the alarm module is triggered and an alert message appears on the web portal. The wet sensor used has an operating Voltage between 3.3 V and 5 V and has an on-board LM393 comparator and an on-board power indicator LED, along with an on-board digital switching indicator LED. The services of Temboo Cloud, which generates editable software code in standardized, production-ready blocks for easy-to-implement but powerfully persistent connections to cloud services, APIs, and devices, are used to alert the parents via call or text. Nexmo Server, the Vonage API Platform, provides tools for voice, messaging, and phone verification, allowing developers to embed programmable communications into mobile apps, websites, and business systems and used in our research to convert the text generated by Temboo to voice message that is to be sent to the parents via call.

962

P. J. Nair and V. Ravi

Fig. 7 Intelligent Cradle Monitoring System

Intelligent Cradle Monitoring System This research work was demonstrated through a prototype using the components and services mentioned in the sections. The prototype was proof to the practicality of the idea proposed and can serve to customer needs, if scaled accordingly. Figure 7 shows the snapshot of the Intelligent Cradle Monitoring System which shows the implementation of the concept discussed. This demonstrates the idea of a cradle that helps take care of the baby in an efficient manner with less human intervention. The microphone, wet sensor, motor, camera, motor driver, and controller can be seen attached to the cradle. The controller sends signal to the motor, once the input is received from the microphone or the wet sensor, thus facilitating the movement of the cradle. Meanwhile, the status of the cradle is being constantly updated on the web portal and alerts are sent to the parents upon being triggered by inputs. The camera module constantly records the video of the cradle and streams it through the Internet, unless switched off by the parents through the web portal.

Web Portal The web portal has been designed so as to provide parents the access to control the cradle through the Internet. Once the system is switched on, an IP address is generated for the access of the web portal. The status of the cradle is also displayed to help parents control the cradle. Figure 8 is a screenshot of the web portal that provides access to the parent. In Fig. 8, the parent has logged on to the control mode that allows the parent to stop or sway the cradle. This mode is a submode under manual mode. The status of the mattress and cradle are also being displayed. The mattress is currently displayed

36 Intelligent, Automated, and Web Application-Based Cradle Monitoring System

963

Fig. 8 Web portal screenshot

as being wet and will remain the same till the mattress is changed. The option to stay or quit is also available.

Live Streaming An URL is shared to the parents where the video of the cradle is being streamed. This enables the parents to view the status of the cradle and take necessary actions. A screenshot of the web page through which a video of the cradle is being streamed is shown in Fig. 9. This shows the activity of the cradle, whether it is being swayed or not. Here are the possible scenarios for the system failure and the respective mechanism to detect it. (i) Video not streamed due to camera or network failure: This can be detected at the backend of the web portal (buffer time = 10 min). (ii) The cradle is not being rocked despite observing trigger conditions due to motor failure: This can be detected by the plausibility of two accelerometers placed on the surface of the cradle. (iii) Baby cry or baby wetting the mattress is not detected: This is again detected by the plausibility of the inputs provided by the two microphones and/or wet sensors. (iv) Overall system shutdown – The sensor outputs are monitored at the back end of the web portal every five minutes. If the sensors do not provide plausible outputs, then the inputs are monitored every minute and warning is sent to parents. In all the above-mentioned cases, a warning or an alert is sent to the parents via call.

964

P. J. Nair and V. Ravi

Fig. 9 Live streaming snapshot

Future Scope This system would prove to be extremely useful in day care centers where details of multiple cradles could be uploaded on the cloud for the parents to be informed. The control would be with the monitor providing for an easy way of baby care. Maternity wards in hospitals have a number of cradles that require constant monitoring. This system would help take care of the babies by constantly monitoring the multiple cradles and alerting the nurse in charge, thus reducing the amount of attention required from them. This system has the advantage of being dependent on the details of the cradle and not on that of the baby, which makes the updating of details a one-time process and a fairly easy one.

Conclusion In this research, a method to monitor the cradle with less human intervention was demonstrated. Parameters such as the baby cry and the moisture level of the mattress were taken as inputs to decide whether to rock the cradle or to change the mattress.

36 Intelligent, Automated, and Web Application-Based Cradle Monitoring System

965

The video of the cradle and the status was shared to the parents/guardian. The parents have the provision to control the cradle through the web portal whose IP address was shared to them. Overall, the research work proves to be efficient at a small-scale level.

Important Websites 1. https://core.ac.uk/download/pdf/232204217.pdf 2. https://www.geeksforgeeks.org/project-idea-baby-monitoring-smart-cradle/ 3. https://projects.kluniversity.in/Nirmalapasala/iot-based-smart-cradle-with-babymonitoring-system-embedded-with-s.odi-iot-devices/-/issues?milestone_title= Completion+of+project&sort=priority&state=closed 4. http://umpir.ump.edu.my/id/eprint/28858/ 5. https://www.ijcaonline.org/archives/volume160/number5/borkar-2017-ijca913053.pdf

References Anjikar AD, Vaishnow AR, Warade AI, Nishane SB (2017) Analysis and synthesis of smart BASSINETS for infants. Analysis 4(3):3533–3536 Bhatnagar N, Shinghal K, Saxena A, Tiwari N, Bhatnagar S, Kumar S (2016) Design of automatic & indigenous Ecradle. Imp J Interdiscip Res (IJIR) 2(6):328–333 Choi S, Yun S, Ahn B (2020) Implementation of automated baby monitoring: CCBeBe. Sustainability 12(6):2513 Goyal M, Kumar D (2013) Automatic E-baby cradle swing based on baby cry. Int J Comput Appl (0975–8887) 71(21):39–43 Hussain T, Muhammad K, Khan S, Ullah A, Lee MY, Baik SW (2019) Intelligent baby behavior monitoring using embedded vision in IoT for smart healthcare centers. J Artif Intell Syst 1(15):110–124 Jabbar WA, Shang HK, Hamid SN, Almohammedi AA, Ramli RM, Ali MA (2019) IoT-BBMS: internet of things-based baby monitoring system for smart cradle. IEEE Access 7:93791–93805 Joshi MP, Mehetre DC (2017) IoT based smart cradle system with an android app for baby monitoring. In: 2017 international conference on computing, communication, control and automation (ICCUBEA). IEEE, pp 1–4 Kadu AB, Dhoble PC, Ghate JA, Bhure NB, Jhunankar VA, Sirsat PM (2014) Design, fabrication and analysis of automated cradle. Int J Mech Eng Robot Res 3(2):380 Kavitha S, Neela RR, Sowndarya M, Harshitha K (2019) Analysis on IoT based smart cradle system with an android application for baby monitoring. In: 2019 1st international conference on advanced Technologies in Intelligent Control, environment, Computing & Communication Engineering (ICATIECE). IEEE, pp 136–139 Kumar KA, Ravi V (2019) Design and implementation of real-time data acquisition, monitoring and smart tracking system for solar modules. International Journal of Innovative Technology and Exploring Engineering, 8(6):540–543 Lu CC, Wu CH, Su HK (2018) Intelligent infant monitoring system involving a Wi-fi wireless sensor network. In: International conference on intelligent information hiding and multimedia signal processing. Springer, Cham, pp 269–276 Mishra S (2018) Development of RTOs based internet connected baby monitoring system. Indian J Public Health Res Dev 9(2):345–348

966

P. J. Nair and V. Ravi

Mitra S, Ranjitha MS, Ravi V (2015) Video headend video quality monitoring solution. Indian J Sci Technol 8(19):1–12 Mnati MJ, Van den Bossche A, Chisab RF (2017) A smart voltage and current monitoring system for three phase inverters using an android smartphone application. Sensors 17(4):872 Nejkar VA, Nimbhorkar SR, Paliwal JK, Shrivastav AA (2018) Smart nanny an IoT based baby monitoring system. i-Manager’s J Comp Sci 6(1):28 Palaskar R, Pandey S, Telang A, Wagh A, Kagalkar R (2015) An automatic monitoring and swing the baby cradle for infant care. Int J Adv Res Comp Commun Eng 4(12):187–189 Patil AR, Patil NJ, Mishra AD, Mane YD (2018, January) Smart baby cradle. In: 2018 international conference on Smart City and emerging technology (ICSCET). IEEE, pp 1–5 Rajesh G, Arun Lakshman R, Hari Prasad L, Chandira Mouli R (2014) Baby monitoring system using wireless sensor networks. ICTACT J Commun Technol 5(3):963–969

Review on Deep Learning Algorithms in Medical Devices

37

G. Ananthi and Arockia Selvakumar Arockia Doss

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of Medical Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Medical Device: Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Medical Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Food and Drug Administration (FDA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Deep Learning in Health Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . K Means Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . K Means Clustering Subgroups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hierarchical Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Clustering in Hierarchical Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . K Nearest Neighbors Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PCA Algorithm (Principal Components Analysis) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rules in Association . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dimensionality Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Regression Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Linear Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Logistic Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Decision Tree Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Naïve Bayes Classifier Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Support Vector Machine Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Positive Reinforcement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Negative Reinforcement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Artificial Intelligence in Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

968 971 972 973 975 976 976 977 977 977 977 978 978 978 979 979 979 981 981 981 985 986 986 989

G. Ananthi () Department of ECE, Thiagarajar College of Engineering, Madurai, Tamil Nadu, India e-mail: [email protected] A. S. Arockia Doss Design and Automation Research Group, School of Mechanical Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_167

967

968

G. Ananthi and A. S. Arockia Doss

Autonomous Surgical Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Artificial Intelligence in Surgical Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Real-Time Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Artificial Intelligence in Robotic Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

989 989 990 990 991 992

Abstract

The devices used in medical Industry depend on the senor data and medical images. Using sensor data, the medical data is processed to understand the patients’ health condition. The medical images can be useful to process the human health. The Novel Deep learning algorithms are required to process the data and improve the performance. The Deep learning algorithm is used to detect disease symptoms at the earlier stage. Deep learning offers considerable promise for medical diagnostics. Currently, the medical industry requires innovative ideas to process the large amount of data and improve the quality of service. Hence, intelligent system is needed to detect the symptoms of the disease. Deep learning algorithm is proposed for handling large amount of medical data to classify the disease accurately. The medical data includes patients’ records, insurance records, and medical records for treatment. Deep learning algorithm is helpful for diagnosis in medical industry. In this chapter, the various deep learning algorithms that are used to classify the diseases are explained in detail. Keywords

Deep learning · Medical devices · Health data

Introduction The deep learning algorithms can be useful to improve the health care by contributing earlier diagnosis of diseases. It is a scientific approach that mainly concentrates on computers learning from data. It is a combination of intersection of statistics and in turn to learn relationships from data for efficient computational algorithms. The massive data sets consist of many data points that provide many challenges in mathematics and computer science fields. The deep learning algorithms have been employed to classify diseases using medical images with comparable accuracy. This chapter addresses the various deep learning algorithms for prognosis of diseases using medical devices. The medical devices applications are automation and semiautomation of tasks such as medical image segmentation. A large number of datasets are useful for data mining and are used to uncover the patterns, discovery like drug target biomarker disease.

37 Review on Deep Learning Algorithms in Medical Devices

969

Artificial Intelligence is an area that allows the computers to resemble like a human being. Machine learning is accurate at predicting outcomes without being explicitly programmed. Deep learning used to extract patterns using existing data. These techniques are very efficient and useful to transform health care from health data. The advantages are early disease detection, accurate diagnosis, finding of observations, and development of diagnostics. These algorithms can be useful to learn the healthcare experience and improve their performance. Software as a Medical Device is a Software to learn the Patients health data and develop time to improve care (Unique Device Identification System 2013). Intuitively, the action plan has been released the medical devices machine learning software recommended by the Food and Drug Administration (Artificial Intelligence and Machine Learning in Software as a Medical Device, FDA). The machine learning algorithms in medical devices are studied in Digital health technology (Machine Learning Applications in Medical Devices (invetechgroup.com)). Artificial Intelligence in Medical devices Industries can be explained in the fields of Management of chronic diseases, medical imaging, AI and Internet of things and Robotic surgery applications and machine learning in robotic surgery (AI in Medical Devices – Three Emerging Industry Applications, Emerj). Hence, deep learning algorithms in robotics medical devices are more futuristic research work. Several research works concentrate on the motor ability of Artificial Intelligence and human guidelines for human balance control (Yuan et al. 2020). Several human push revival schemes from Deep learning algorithm given in Figs. 1 and 2. Data labeling does not require data in reinforcement learning due to the important iterations in robots (Bonsignorio et al. 2020). The deep learning algorithms are used to approximate object from optical tactile sensor input. So, deep learning in health care has been widely used in medical applications for surgical planning (Chen and Jain 2020). This book deals with the advances and future of deep learning in medicine and health care. Cough signal processing is a new major research area to diagnose covid-19 disease (Wang and Wong 2020). The development of such medical devices is under major research area. The cough recognition technique is used to detect the position of cough sounds in real time (Fig. 3). Based on a cell phone-recorded cough, machine learning models accurately detect corona virus even in people with no symptoms (Wang and Wong 2020) (Figs. 4 and 5). Internet of Medical things is the network of smart devices operated by the devices for human health care and wellness (Baker et al. 2020) (Fig. 6). Internet of Medical things • Wearables: wristbands, smart watches, etc. • Continuous monitoring: activity, heart rate, sleep, location, etc. • Connectivity: Bluetooth, Wi-Fi, cellular, etc.

970

Fig. 1 Push recovery strategies (Yuan et al. 2020)

• • • • •

Real-time monitoring and just-in-time intervention Cloud: computing and storage Scalable to large cohorts Analytics: machine learning and signal processing Interpret data and predict outcomes

G. Ananthi and A. S. Arockia Doss

37 Review on Deep Learning Algorithms in Medical Devices

971

Fig. 2 Classification of different robots (Yuan et al. 2020)

Types of Medical Software The Food and Drug Administration requires the label of medical device software to contain unique device identification regulation (U.S. Food and Drug Administration 2019). The FDA has also published additional guidance documents about specific types of software that are used in health care and that might be regulated. The significance of software and hardware medical data systems products is to transfer the data formats for display in medical devices. These data systems do not modify the data but display the data and control parameters of the medical data. The systems are intended for monitoring the patients (Software as a Medical Device (SaMD) 2013). These applications might or might not be regulated, depending on other functions of the application. The software is applicable for transfer, store, convert, and display the functions based on regulations. Software helps to store data converts into a format and is useful to display the electrocardiogram of a patient. This software is used to control the functions and device parameters. This software is used to generate alarms related with patients’ health information on a display in medical devices. Similarly, software that detects and highlights abnormalities (computer-assisted detection (CADe)) or software that assesses associated disease severity (computerassisted diagnosis (CADx)) is considered a device by the FDA and is subject to regulatory focus.

972

G. Ananthi and A. S. Arockia Doss

Fig. 3 Cough signal processing

Clinical decision support system software facilitates clinicians and patients to identify patient related information to improve the healthcare monitoring system (Collins and Yao 2018). It consists of several tools that help to improve the decisionmaking skills in clinic. The tools contain patient’s data reports, templates for documents, diagnostics materials, and clinical guidelines. Because of the variety of CDS applications, as well as an evolving regulatory landscape, some CDS software might be regulated by the FDA, others might be regulated by the FDA but under “enforcement discretion,” and some might not be regulated as medical devices.

Medical Device: Definition The use of medical device prevents diseases based on diagnosis of the patients and also affects the body functions. It excludes certain software functions such as data storage, administrative support, and electronic patient records (Medical devices (who.int)).

37 Review on Deep Learning Algorithms in Medical Devices

973

Fig. 4 Different coughs

Medical Devices Examples of Medical Devices and Technologies are Analytical Instruments, Anesthesia, Assay Kits, Cardiology, Cell Biology, Cell Counting, Cleaning and Sterilization, Clinical Emergency Medicine, Cosmetic/Plastic Surgery, Cryogenic, Dental, Dermatology, Ear, Nose and Throat (ENT)/Otolaryngology, Endoscopy, General Hospital, General Practice, General Surgery, Hematology, Imaging, Immunoassays,

974

G. Ananthi and A. S. Arockia Doss

Fig. 5 Medical error analysis

Problem

Diagnosis

Treatment

Traditional Model

Prevention Walking wellness

Running

Swimming

Diagnosis

Treatment

New Model

Happy Person

Wellness and Preventive Care

Fig. 6 Model analysis

Lab Equipment, Microbiology, Microscopy, Microscopy and Laboratory Microscopes, Molecular Biology, Nephrology, Neurology, Obstetrics, Gynecology, Ophthalmology, Patient handling, Pediatrics, Physiotherapy, Prosthetics Radiology, Refrigeration, Spinal and Orthopedic, Support, Software and Services, Ultrasound Urology (Blog (emergobyul.com)). The Details of the medical devices and technology can be explained (50 Best Medical Device and MedTech News Sites,

37 Review on Deep Learning Algorithms in Medical Devices

975

Blogs – Pannam). In (Applications of deep learning in medical device manufacturing – ONdrugDelivery), medical device manufacturing data is used to define a component and measurement.

Food and Drug Administration (FDA) Figure 7 shows the market size for medical devices. The Food and Drug Administration helps in regulating medical devices that diagnose and prevent diseases. Medical devices are used by healthcare providers to diagnose patients (Medical Devices, FDA). Drawbacks in Medical Devices: The obstacles in medical devices are as follows (PT PHC Indonesia (phchd.com)): • • • • • • • • • •

Lack of Information Fascination with technology Deference to personal preference High costs Lack of a single nomenclature Marketing practices Counterfeiting Deficiencies of clinical guidelines Inappropriate design Misinformation of the device

The advantages of using deep learning algorithms in medical devices is that the algorithm runs based on patient’s health data and is useful to predict the disease effectively within a short span of time. No human intervention is required and cost as well.

612.7

425.5

2015

2016

2017

2018

2019

2020

Fig. 7 Market size for medical devices (in dollars)

2021

2022

2023

2024

2025

976

G. Ananthi and A. S. Arockia Doss

Deep Learning in Medical Devices: The supervised algorithm is useful to annotate images with normal state or disease state. This algorithm is trained to recognize the states. After training, if a new image is presented, the algorithm matches the image to the most similar trained state. The training data sets cover a large number of disease data collected from patients and different medical devices (Software as a Medical Device (SaMD) 2013).

Deep Learning in Health Care Unsupervised Learning Unsupervised learning offers an important set of tools in Artificial Intelligence because: • It allows us to better understand and visualize complex, multidimensional datasets. • It succeeds in a situation whether the data is not labeled. • It paves the road to extending beyond the bounds of human performance by detecting nuances that are imperceptible to us. In unsupervised learning, the learning is based on the probabilistic data model. Based on the past inputs, a machine can be used to estimate the probability distribution model for an input. The unsupervised learning techniques are clustering and dimensionality reduction. Clustering means that similar metric data points are grouped together. Clustering is one of the classifications of unsupervised learning algorithm that finds the structure of data in uncategorized manner. It then processes the data and finds the suitable cluster whether data is available inside the cluster. The clustering types are hierarchical clustering, principal component analysis, singular value decomposition, k nearest neighbors, independent component analysis, and K means clustering.

K Means Algorithm It is a partitioning algorithm in a clustering technique. Data can be combined together, and one data belongs to a particular cluster. This algorithm is used to find the maximum value in each iteration. The clusters can be selected based on the maximum values. The cluster data points are arranged into k number of groups. If k is larger, it means that there are smaller groups with maximum granularity. A minimum value of k represents maximum groups with minimum granularity. This algorithm output is named as labels. This algorithm assigns the data point to k groups. Each group creates a centroid point. This centroid captures the points and adds them together to form the cluster.

37 Review on Deep Learning Algorithms in Medical Devices

977

K Means Clustering Subgroups • • • • • •

Agglomerative clustering. Let us assume that each data is a cluster. Based on the distance, it minimizes the clusters in every iteration. Large cluster contains the objects finally. Dendrogram. The height of dendrogram represents the similarity level between the clusters.

Hierarchical Clustering It is an agglomerative clustering system. Here, each data is treated as a cluster. The two nearest clusters form an iterative union that can reduce the number of clusters.

Fuzzy Algorithm Fuzzy sets are a type of overlapping technique for clustering. Each point belongs to clusters that will have separate membership degree. The data is associated with membership value. The algorithm is named as Fuzzy C means algorithm. Probabilistic Distribution It is the type of unsupervised learning algorithm that uses probability distribution to form clusters. Examples are represented here. • • • •

“men’s shoe.” “women’s shoe.” “women’s glove.” “men’s glove.” The algorithm can be clustered into “shoe” and “glove” or “men” and “women.”

Clustering in Hierarchical Scheme It is an algorithm that consists of cluster hierarchy. All data is assigned to each cluster. Two near clusters are represented in the similar cluster. This algorithm ends in one cluster leftward.

K Nearest Neighbors Algorithm It stores and classifies the new examples based on measures similarity. The performance of this algorithm works well based on the distance between the

978

G. Ananthi and A. S. Arockia Doss

examples. If the distance calculation is nontrivial and training set is large, the speed of the learning is slower.

PCA Algorithm (Principal Components Analysis) The principal component is called as base. In higher dimensional space, select a foundation for the space and choose the score on that basis. Select the subset in a space consumes minor in size when compared to the unique space. The data complexity is to be maintained as possible.

Rules in Association The rules in association allows to connect associations between the data objects in the large database. It discovers the relationships between variables in the database. Examples include: People buy home and in turn purchase furniture, and another group of patients affected by cancer and their grouping based on gene measurements, shopper’s purchasing histories making groups, movie reviews making groups based on the rating.

Dimensionality Reduction It is the process of large dimensional data into a lower dimensional data called as feature selection (Sutton and Barto 1998). It is helpful to predict diseases. The dimensionality deals with the number of feature inputs and variables and columns in the dataset. Dimensionality reduction is nothing but it reduces the features from the dataset. The dataset contains large number of feature inputs that makes predictive modeling tough. It seems difficult to make predictions for the dataset training with a large number of features. Dimensionality reduction is defined as, “It is a way of converting the higher dimensions dataset into lesser dimensions dataset ensuring that it provides similar information.” The machine learning methods are considered as a better predictive models for classification and regression problems. The applications of machine learning algorithms are speech recognition, signal processing, and bioinformatics. It is used for noise reduction, data visualization, and cluster analysis (Fig. 8). Supervised learning: The machine learning with labeled data is known as supervised learning. Prediction of discrete values is called classification. Prediction of continuous values is called regression. The supervised learning examples are lung disease findings and findings of different body organs from medical images. It is the process of learning an algorithm to map an input for a particular output. Labeled datasets are collected. Once the mapping is correct, the algorithm is learned successfully. Regression and classification are the types of supervised learning algorithms.

37 Review on Deep Learning Algorithms in Medical Devices

979

Dimensionality reduction Techniques

Dimensionality Reduction

Feature Selection

Missing Value Ratio Low Variance Filter High Correlation Filter Random Forest Backward Feature Extraction Forward Feature Selection

Components/Factors based

Factor Analysis Principal Component Analysis Independent Compone Analysis

Projection Based

ISOMAP t-SNE UMAP

Fig. 8 Dimensionality reduction techniques

Regression Algorithm It is a supervised learning that learns labeled datasets and predicts continuous output for a current data based on an algorithm. Normally, the output in supervised learning is represented in terms of money, height, etc.

Linear Regression It is assumed that there is a linear relationship between the input and output of the learned data. Independent variable is represented as an input variable. Dependent variable is termed as output variable. If any unknown data is given to the algorithm, we can use functions, calculate the input, and map the input for the output in a continuous manner (Fig. 9).

Logistic Regression A set of independent variables are given to this algorithm and predicts discrete values. The prediction is implemented using mapped unknown data based on the

980

G. Ananthi and A. S. Arockia Doss

Fig. 9 Linear Regression

Fig. 10 Logistic regression

logit function. The output lies between the values zero and one. It predicts the new data based on probabilistic approach. The classifier classifies one or zero of two classes (Fig. 10). Classification is a type of learning algorithm maps, and the current data is obtained in any one of the two classes in the dataset. The output is represented in any one of the classes and does not represent in a number. Types: • Naïve Bayes classifier • Decision tree • Support vector machine

37 Review on Deep Learning Algorithms in Medical Devices

981

Root

A

C

B

D

Fig. 11 Decision tree

Decision Tree Algorithm It classifies based on the feature input values. This algorithm uses the information, finds the feature from the dataset, and provides the best information assigned as the root node. It assigns each branch in the decision tree denoted as feature in the dataset (Fig. 11).

Naïve Bayes Classifier Algorithm In this algorithm the features in the dataset are assumed to be independent to each other. This algorithm works very well even in large datasets. Directed acyclic graphs are used for classification too.

Support Vector Machine Algorithm These algorithms use statistical learning theory. The algorithm uses kernel based on learning task that maps input and output. It creates a hyperplane that is used to classify two different classes (Fig. 12).

Semi-Supervised Learning Semi-supervised learning is useful to recognize the activity using sensor data in medical image segmentation. This algorithm helps to learn the human behavior

982

G. Ananthi and A. S. Arockia Doss Margin Y

Hyperplane

X Support Vectors

Fig. 12 Support Vector Machines

Fig. 13 Machine learning algorithms

using labeled and unlabeled data. The use of this algorithm is to know if the data is labeled or unlabeled to change the behavior of learning. If the labeled data is expensive in supervised learning, semi-supervised learning can be used in machine learning and data mining. It includes mixture models, co-training, self-training, graph-based learning, multiview learning, and semisupervised support vector machines. Some specific algorithms are generative models, self-training, S3VMs, graph-based algorithms, and multiview algorithms (Fig. 13). 1. Self-training Step 1: The entire data is split into training and test datas. Train the algorithm based on the labeled data. Step 2: Pseudo labels are named as highest probability of the predicted class in unlabeled data.

37 Review on Deep Learning Algorithms in Medical Devices

983

5 4 3 2 1 0 -1 -2 -3 -4 -5 -5

-4

-3

-2

-1

0

1

2

3

4

5

Fig. 14 Generative Models

Step 3: Combine the pseudo labeled data and labeled data. Step 4: Perform the classification metric based on performance measures. 2. Generative Models Assume labeled data Xl , Yl , each class has a Gaussian distribution (Fig. 14). Model Parameters:     (1) , θ = w1 , w2 , μ1 , μ2 , 1

2

The GMM: p (x, y| θ ) = p( y| θ ) p ( x|y, θ )  = wy N x; μy , y

(2)

Classification: p ( y| x, θ ) = 

p ( x, y| θ )  y  p ( x, y | θ )

(3)

984

G. Ananthi and A. S. Arockia Doss

Fig. 15 S3VMs

3. Semi-supervised Support Vector Machines (S3VMs) It maximizes unlabeled data margins (Fig. 15). Assumption: The data in unlabeled one from different classes can be separated with maximum margin. Implementation steps: • Represent 2u possible values of the set Xu. • Represent Singular Value Machine for each labeled data. • Represent the Singular Value Machine with the maximum margin value. 4. Graph-Based Algorithms The graph is represented using labeled and unlabeled data. The heavy edges are connected using same labeling via instances. Nodes: Xl ∪ Xu Edges: The weights are calculated by means of features using k nearest neighbor graph, and unweighted fully connected graph weight decays using distance (Fig. 16). Graph-Based Algorithms. Harmonic, Local and global consistency, Manifold regularization and Mincut algorithms are called as graph-based algorithms. 5. Multiview Representations Algorithms

37 Review on Deep Learning Algorithms in Medical Devices

985

Fig. 16 Edges

The features are represented as image features and web page text called as multiple views. In data co-training, train an image classifier and text classifier. Assume that the classifiers are trained and taught to each other. Reinforcement Learning: The reinforcement learning algorithm is used to learn the observations, rewards, and actions at a specified time (Anwar et al. 2018). This algorithm is used for context aware symptoms to find the diseases. This learning algorithm is the type of machine learning algorithm. It is used to maximize the reward based on actions in a situation. The machines and software are to find the best path for a particular situation. This algorithm differs from the supervised learning algorithm because the reinforcement agent decides the task. Here, the training dataset is absent and is learned based on the experience. Algorithm: • The initial state is considered as an input and the model has to be started. • The possible output has been represented for a specific problem. • The training has been implemented based on the input. The implemented model returns the state and hence the user decides to give reward option or punishing option based on the output. • This model will continue to learn based on the given parameters. • Maximum reward has been given for the best solution.

Positive Reinforcement Whenever an event occurs for behavior, it increases the power and occurrence of the behavior. The advantages of this approach are that it maximizes the performance and sustains change for longer period of time. The drawback of this method is to provide overload of states that can decreases the consequences.

986

G. Ananthi and A. S. Arockia Doss

Negative Reinforcement It is defined as strengthening a behavior if negative condition is ended. It increases the behavior and provides defiance for the minimum standard performance. The drawback of this negative reinforcement is it provides minimum behavior.

Applications 1. 2. 3. 4. 5. 6.

Robotics for industrial automation Machine learning and data processing Training systems and materials Applicable in known environment model Applicable in simulation model Information about the environment

Deep Learning in Genomics The genomic technology collects a measurement from individual DNA sequences to a set of proteins in blood. Deep learning in Genomics takes raw gene data converts into input data tensors given to the neural network for biomedical applications. The research challenges include stochastic optimization algorithms that are useful to predict the protein structure. Deep learning helps prediction from genetic data which includes traits and disease risk. The prediction phenotype is used to estimate the intermediate molecular phenotypes in terms of gene expression and gene splicing to find the downstream disease predictors. The molecules states consist of proximal signals that are important to predict larger human traits, and hence more extensive training data needed. The predicted features are useful to predict splicing and transcription factor binding. Algorithm: 1. From blood, the DNA is released from the cells. 2. The organ injection fragments are affected by means of graft cells immune system through bacterial infection and cancer. 3. Biomarker data is affected from noise and needs analysis to find the disease. 4. Deep learning algorithms are useful to improve the quality of DNA sequences, gene expression, methylation, chromatin profiles and measurements. Deep Learning in Health Care: Figure 17 represents the different application areas in deep learning methods. The categories of application and examples are represented. The relation between application category and application example is defined.

37 Review on Deep Learning Algorithms in Medical Devices

987

Fig. 17 Biological system

Biological systems: Human biological genomic sequences such as DNA and RNA, and bacterial and viral multiplication and mutation are used to create a predictive model in deep learning algorithms. The prediction is in the form of gene identification, protein interaction, biological data, drug composition reaction profiling, and binding between DNA and proteins. Challenges in Medical Devices and Deep Learning: • • • • • • • • • • •

Medical data representation and transformation Handling biomedical data stream Medical big data analysis Medical big data hardware requirements Guarantee the medical devices to maintain high class product quality Confirmation of product safety standards and regulatory standards Medical devices localization Complexity of computation in deep learning Multitasking deep learning Medical Internet of things and application Semi-supervised learning for biomedical big data

988

G. Ananthi and A. S. Arockia Doss

Table 1 Sample data and Deep learning Sample Data Images of the eye fundus Histological section images Radiology, CT, MRI images Speech, movement patterns Diagnoses, gene data, etc. ECG or EEG signals Internet searches Laboratory values, environmental factors, etc. Patients health records

Function Finding the diabetic retinopathy Recognition of cells Cancer, heart problems Detecting depression Selection and dosage of medicines Heart diseases, brain diseases Detecting epidemics Disease prognoses Time-of-death prognosis for intensive care patients

Sample data and deep learning tasks (Table 1):

Machine learning applications in health care: • Diagnosis and disease identification: Disease identification is a major research issue in initial stage. The advantage of deep learning is to diagnose the disease in initial stages. • Health records improvement: Healthcare record maintenance is a major research issue. • The diabetes prediction: Diabetes is the common and serious disease that causes illnesses. It mostly damages the heart, nerves, and kidneys. The classification algorithms such as KNN, Decision tree, and Naive Bayes are used to predict diabetes. Naive Bayes algorithm is used to improve the performance in terms of computation time. • Liver disease prediction: The liver disease is vulnerable to liver conditions and is similar to hepatitis, cancer and cirrhosis. Using medical big data and machine learning algorithm, it is a very toughest research area to predict the liver disease. Liver disorder data set is available in the Indian liver patient data set. • Artificial Intelligence for finding the best heal • Making diagnoses via image analysis • Personalizing treatment

37 Review on Deep Learning Algorithms in Medical Devices

• • • • •

989

Machine Learning in medicine is making great progress Adjusting behavior Medical research and clinical trial improvement Leveraging crowd sourced medical data Epidemic control

Artificial Intelligence in Surgery The significance of Artificial Intelligence in surgery is the decision-making ability, surgical techniques applications in the surgical procedure, complex surgical procedures, instantaneous complications, and the solutions accessible to patients.

Autonomous Surgical Devices Human surgery is explained by physical and technical variables used to maintain consistency in surgery. The factors are used to provide functional outcomes, barrier rates and survival practice in the Medical field.

Artificial Intelligence in Surgical Robotics The Artificial Intelligence is transformed from existing surgery by autonomous involvement for treating both acute and chronic symptoms. By incorporating such methods, the progress has been achieved in preoperative planning, intraoperative guidance, and surgical robotics. There will be a transition from laparoscopic surgery into robotic surgery. Artificial Intelligence is used to implement the capabilities of surgical robots in terms of autonomy, identify the problems based on the environment and implement the exact actions without the need for human intervention. The use of robotic arms permit the surgeons to perform surgery actions with advantages such as 3-D imaging, ergonomics for the surgeon, the ability to control the arm and 7 degrees of articulation, at the minimum cost of haptics and the inability of the surgeon to be in constant contact with the actual patient. Robotic surgery is a form of invasive surgery, the combination of mechanics and electronics; however, this is a generalization of robotic surgical systems because the power of robotic surgery exists to create to create autonomous actions. Robots are used for surgery, but it is not fully integrated with Artificial Intelligence. Recently three-dimensional representation of high quality image with robotics is popular; enhanced lack of restrictions of movement with articulating hand tools, vibration elimination, and the option of safe suturing in the thin spaces as in open surgery are the advantages of robotic surgery. Artificial Intelligence is used in surgical robotics. The deep learning data is used to automate the behavior programming. Artificial Intelligence decides the

990

G. Ananthi and A. S. Arockia Doss

patterns of surgical procedures and enhances practices and robot accuracies for submillimeter precision. Artificial Intelligence is incorporated with machine vision to analyze scans. Artificial Intelligence implemented in robots empowers to take the decisions on their own. The different types of Artificial Intelligence are as follows.

Weak Artificial Intelligence This Artificial Intelligence is used to create a simulation of human minds and interactions. These robots provide predefined commands and responses. But, the robots cannot understand the commands, but they retrieve the appropriate response if the suitable command is given. Strong Artificial Intelligence This type of Artificial Intelligence is used in robots to perform its task by its own. It does not need any supervision. Once it is programmed, it will perform the task correctly. This Artificial Intelligence can be used in humanoid robots that sense the environment and interact with their surroundings. Specialized Artificial intelligence This Artificial Intelligence is used when the robot needs to perform only the specified special tasks. This Artificial Intelligence is used in Industrial Robots which can perform tasks like painting and tightening.

Real-Time Examples The Children National Medical Centre in Washington demonstrated a robot conduct automated soft tissue surgery. The surgery results are shown to be better than a human surgeon. The future of robotic surgery provides improvement in the technology used in the operating room, with the robots abilities to communicate haptic feedback to the surgeon. This will help for unparalleled sensation and eliminate unintentional tissue contact and injury. The user interface will connect to the surgeon and provide access to the patient, sterile in a procedure, employ head mounted three-dimensional visualization system, and allow the master manipulation of the robot.

Artificial Intelligence in Robotic Applications Assembly It is combined with vision systems, Artificial Intelligence helped with correction, which is used in sectors in terms of aerospace, etc. Artificial Intelligence is used to help a robot learn on its own such as which paths are best for certain processes while it is in operation.

37 Review on Deep Learning Algorithms in Medical Devices

991

Packaging Robotic packaging uses the Artificial Intelligence for quicker, minimum cost, and accurate packaging. Robotic Packaging helps to track the robotic system motions constantly to refine the Robotic System. Customer Service Robots use leverage Artificial Intelligence natural language processing abilities to interact with customers in a human technique. Open Source Robotics The users can teach their robots to do custom responsibilities based on their specific application. Deep Learning Algorithm Challenges • Datasets collection in clinics • Laboratory data collections • Maintain confidence score to find the disease Collar Design: This design consists of a chest, chin support in the front section, and on the other side occipital, upper back (trapezius muscle) support. The chin, occipital supports or platforms that form a top (or) movable platform and the chest, upper back supports that form a base (or) fixed platform are connected by adjustable belts. The platforms connected by extensible and retractable linear electric actuators serve as prismatic links around the neck. Based on the finalized dimensions, the cervical collar is fabricated, as shown in Fig. 3. As the patient is supposed to be wearing the device for the therapy, the weight of the collar is crucial to be very light in weight. To address this issue, fiber material is used in fabricating the chin and chest platforms, and a foam cushion is provided under the platform so that while wearing the device for a long time, it should not cause any kind of discomfort. This foam cushion can absorb sweat if produced by the patient. The overall weight of the cervical collar including the joints and actuators is approximately around 400 g (Figs. 18 and 19). The CROM comprises three-fluid dampened inclinometers. These inclinometers (Doss et al. 2021; Lingampally and Selvakumar 2019) are attached to a lightweight acrylic frame, where this is shaped to fit the head and is fastened using the hookand-loop straps, as shown in Fig. 4a and b.

Conclusion This chapter addresses the deep learning algorithms used in medical devices. The commercial medical devices are addressed and drawbacks of existing devices are also addressed. The significance of deep learning algorithms used in medical devices is also addressed. This chapter is useful for predicting the diseases like covid-19 using deep learning algorithms for medical covid kits.

992

G. Ananthi and A. S. Arockia Doss

Fig. 18 Prototype of collar

CROM

Spherical joint Top platform Electric linear actuator (Prismatic joint) Revolute joint

Base platform

Inclinometers

Magnet

(a)

(b)

Fig. 19 (a) Lateral flexion (left), (b) lateral flexion (Right)

References Anwar SM, Majid M, Qayyum A, Awais M, Alnowami M, Khan MK (2018) Medical image analysis using convolutional neural networks: a review. J Med Syst 42(11):226 Baker SB, Xiang W, Atkinson I (2020) Internet of things for smart healthcare: technologies, challenges, and opportunities. IEEE Access 5:26522–26525 Bonsignorio F, Hsu D, Johnson-Roberson M, Kober J (2020) Deep learning and machine learning in robotics. IEEE Robot Autom Mag 27(2):20–21

37 Review on Deep Learning Algorithms in Medical Devices

993

Chen Y-W, Jain LC (2020) Deep learning in healthcare, paradigms and applications. Springer International Publishers-Book Collins A, Yao Y (2018) Machine learning approaches: data integration for disease prediction and prognosis. Applied Computational Genomics. Springer, pp 137–141 Doss ASA, Lingampally PK, Nurahmi L (2021) Synthesis of a parallel manipulator based rehabilitation cervical collar for c-spine injured patients. Int J Robot Autom 36(1) Lingampally PK, Selvakumar AA (2019) A kinematic and workspace analysis of a parallel rehabilitation device for head-neck injured patients. FME Trans 47(3):405–411 Software as a Medical Device (SaMD): Key definitions, international medical device regulators forum. December 2013 Sutton RS, Barto AG (1998) Introduction to reinforcement learning, vol 2, no. 4. MIT Press, Cambridge U.S. Food & Drug Administration (2019) Medical device data systems. Available at: https://www. fda.gov/medical-devices/general-hospital-devices-and-supplies/medical-device-data-systems Unique Device Identification System: A final rule by FDA. Published 9/24/2013. Docket FDA2011-N-0090, pp 58785–58828 Wang L, Wong A (2020) Covid-19-Net: a tailored deep convolutional neural network design for detection of Covid-19 cases from chest radiography images. Scientific Reports, No. 19549 Yuan K, McGreavy C, Yang C, Wolfslag W, Li Z (2020) Decoding motor skills of AI and human policies: a study on humanoid and human balance control. IEEE Robot Autom Mag 27(2): 87–101

Kinematic Modeling and Analysis of Wheeled In-Pipe Inspection Mobile Robot

38

Rajendran Sugin Elankavi, D. Dinakaran, R. M. Kuppan Chetty, M. M. Ramya, and Arockia Selvakumar Arockia Doss

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kinematics of In-Pipe Inspection Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modeling and Analysis of the Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modeling: A Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis: A Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison of Motion Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

996 997 999 999 1001 1005 1007 1008

Abstract

Research on in-pipe inspection robots (IPIRs) are gaining attention over a few decades as their applications are widening in various fields. Developments in in-pipe inspection robots (IPIRs) are carried out using various types of locomotion. Each type has its advantages and limitations. In this research, an inpipe inspection robot (IPIR) was designed and developed based on wheeled wall press–type locomotion. A kinematic form of the robotic system is developed to determine the robot trajectory and angular velocity. Motion analysis was carried out to find the motion of the robot when it passes through the elbow

R. S. Elankavi () · D. Dinakaran () · R. M. Kuppan Chetty · M. M. Ramya Centre for Automation and Robotics (ANRO), School of Mechanical Sciences, Hindustan Institute of Technology and Science, Chennai, TN, India e-mail: [email protected]; [email protected]; [email protected]; [email protected] A. S. Arockia Doss Design and Automation Research Group, School of Mechanical Engineering, Vellore Institute of Technology, Chennai, Tamil Nadu, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_168

995

996

R. S. Elankavi et al.

and straight pipes. The results from a motion analysis are taken from the case study to compare it with existing research. The findings show that the robot can be employed for in-pipe inspection. Keywords

Pipeline inspection robot · Kinematics · Motion analysis · Mobile robot · Design issues

Introduction Pipelines are used for transporting fluids and gases which are checked for efficient functioning to avoid issues like cracks, corrosion, aging, and mechanical damages. Humans cannot get inside the pipelines, so robots have become the best way to inspect. Developments in in-pipe inspection robots have seen a lot of improvement over the last decades. They are classified based on their motion type. The most used types include pipeline inspection gauge (PIG) (Zhang et al. 2020; Liu et al. 2020), caterpillar-type motion (Kwon and Yi 2012; Zhao et al. 2020), wheel-type motion (Hadi et al. 2020; Li et al. 2020), screw-type motion (Ren et al. 2019; Nayak and Pradhan 2014), inchworm-type motion (Kusunose et al. 2020), wall press–type motion (Feng et al. 2020; Aras et al. 2020), and walking-type motion (Savin et al. 2018; Zagler and Pfeiffer 2003) as shown in Fig. 1. Each type has its advantages and limitations based on its motion type (Elankavi et al. 2020). The PIG type can move through the fluid over a long distance, and the time taken for the process is shorter compared to other types (Zhang et al. 2020; Liu et al. 2020). The caterpillar type has a tracked wheel that keeps contact with the pipe surface in all kinds of pipeline environments and it is best suited for passing through T-branch (Kwon and Yi 2012; Zhao et al. 2020). The wheel type is simple in the mechanism that has less friction between the robot and the inside surface of the pipe and therefore has high mobility (Hadi et al. 2020; Li et al. 2020). The screw-type uses the helical motion to move so that it reduces the damage caused to the wall of the pipe as it does not drag the robot (Ren et al. 2019; Nayak and Pradhan 2014). The inchworm type has a high gripping force compared to others (Kusunose et al. 2020). The wall press type is lighter and smaller in size and has a large contact area (Feng et al. 2020; Aras et al. 2020). The walking type has minimum slippage and climbs easily on vertical pipes (Savin et al. 2018; Zagler and Pfeiffer 2003). The wheeled type has high mobility compared to the other type and it is one of the commonly used types in the in-pipe inspection. The limitation of the wheeled type is its slippage and compared to all other types of in-pipe inspection robot, the wheeled-type in-pipe inspection robot has minimum limitations (Elankavi et al. 2020). The wheeled type is further divided into different types and among them the hybrid type of locomotion is currently used by many researchers. This locomotion type combines with the other types of locomotion to form the hybrid type and this is more efficient in crawling inside vertical pipelines (Roslin et al. 2012).

38 Kinematic Modeling and Analysis of Wheeled In-Pipe Inspection Mobile Robot

997

Fig. 1 Different in-pipe inspection robots (Elankavi et al. 2020)

In this case study, a wheeled wall press–type in-pipe inspection robot (IPIR) is designed and a prototype was built to study the motion of the robot inside the pipeline. It involves the kinematics analysis of the wheeled wall press–type robot. The derived kinematics equation is used to determine the robot trajectory and angular velocity of the wheels. It also solves the issue of irregular motion by locking the rotation of the robot body about the circumferential direction when wheels are mounted 120◦ (Kwon and Yi 2012; Zhao et al. 2020; Li et al. 2020; Aras et al. 2020; Zhang and Yan 2007) apart from each other, and this was observed through motion study. It was found that when wheels are mounted at an angle of 120◦ , 104.88◦ , and 135.12◦ apart from each other, the rotation of the robot body gets locked in the circumferential direction. The purpose of this pipeline inspection robot is additionally to eliminate human factors from labor-intensive and dangerous work, which reduces the number of mishaps that occur because of the lack of routine scrutiny.

Kinematics of In-Pipe Inspection Robot To derive the kinematic equation for a straight pipe the kinematic modeling of the pipeline inspection robot from (Kwon and Yi 2012) is used. The coordinate system and the parameters used for the designed in-pipe inspection robot are shown in Fig. 2. X, Y, and Z show the global reference frame while the x, y, and z represent the local coordinate frame that is present at the center of the robot. The unit vector ˆ jˆ, and k. ˆ for the local coordinate is i, The linear velocity at each wheel V1, V2, V3 is of varying magnitude at the elbows and branches. Thus, the robot linear velocity at the center (Vcz ) along the z-axis is taken, which is accomplished by fetching the wheels (V1, V2, V3 ) average linear velocity.

998

R. S. Elankavi et al.

Fig. 2 The motion of the robot inside the pipeline

Vcz =

r (1 + 2 + 3 ) 3

(1)

The three wheels’ angular velocities are defined as 1 , 2 , and 3. The robot body angular velocity about the x- and y-axis is represented as ωx and ωy. Assumptions are taken to drive the robot: (i) The normal force “P” between the pipe wall and the robot wheel will always point to the axis of the local coordinate frame. That is, P1 will always point to the x-axis. (ii) All wheels are in touch with the inner surface of the pipe. (iii) The body of the robot does not rotate around the local z-axis, but moving along the z-axis is permitted. (iv) The robot wheel keeps line contact inside the pipeline. (v) There is no wheel slippage along the horizontal direction. (vi) The velocities of the three wheels V1, V2, and V3 are the same. The linear velocity of all the wheels exists at the same time because the wheels which are mounted on the legs are connected to the same central hub. So, the linear velocity of the body is given as: Vcz = V1 = V2 = V3

(2)

The robot total angular velocity vector is given as: ω = ωx iˆ + ωy jˆ

(3)

38 Kinematic Modeling and Analysis of Wheeled In-Pipe Inspection Mobile Robot

999

where, ωx =

0.9 r 0.83 r 2 − 3 1.6 a 1.4 a

(4)

and ωy = −

2r 0.44 r 0.56 r 1 + 2 + 3 3a 1.6 a 1.4 a

(5)

Here, “r” denotes the wheel radius and “a” is the distance from the middle of the wheel to the middle of the robot. In the end, the kinematic connection amid the ˙ a ) = (1 2 3 )T are input velocity (u) ˙ = (ωx ωy Vcz )T and the output velocity ( calculated as,   ˙a u˙ = Gua  where,

⎡

⎡ ⎤ 0 ωx  u 2r u˙ = ⎣ ωy ⎦ , Ga = ⎣ − 3a r Vcz 3

0.9 r 1.6 a 0.44 r 1.6 a r 3

r − 0.83 1.4 a 0.56 r 1.4 a r 3

⎤

(6) ⎡

⎤ 1 ⎦ , a ˙ = ⎣ 2 ⎦ 3

˙ a ) and output velocity (u) The kinematic connection amid the input ( ˙ are calculated as:   ˙ a = Gua −1 u˙ 

(7)

From the above equation, for the given linear velocity (Vcz ) and angular velocity of the body of the robot ωx and ωy. , the angular velocity of the wheels (1 , 2 , and 3 ) of the robot is obtained by solving Eq. (7) .

Modeling and Analysis of the Robot Modeling: A Case Study Young-Sik Kwon et al. (2010) designed a three-wheel drive robot with a mechanical clutch that can pass through the pipeline with an inner diameter of 100 mm. On the robot, there are three motorized wheel chains, each with a mechanical clutch. The mechanical clutch is designed using a parallel linkage system. It steers inside the pipeline using a differential drive mechanism. The robot’s foldable mechanism allows it to fit into pipes of many sizes. The robot has two types of modules: one for motion and the other one for retrieval. If in case the robot faces any issues and it gets stuck inside the pipeline, the robot can use the retrieval mechanism to loosen up the contact it provides to always stay in touch with the inner surface of the pipeline. When it loses contact with the pipeline while using the retrieval mechanism the

1000

R. S. Elankavi et al.

robot can be pulled out of the pipeline very easily. Ho Moon Kim et al. (2013) designed a robot that can pass through 150 mm inner diameter gas pipelines. The mechanism is made up of a single motor and a multiaxial differential gear system that presses against the wall. It is made to adapt to varied inner geometries of pipelines, such as elbows, by mechanically regulating the velocities of active wheels. They compress wheels individually so that when it passes over obstacles, the robot always has contact with the pipeline surface. Junghu Min et al. (2014) designed a wheeled-type pipe inspection robot that can pass through 300–500 mm diameter pipe with many elbows. The robot has two modules: an active module and a passive module. Each module features three alternative wheel configurations, each with its mechanism for expanding the wheels. In pipelines, the robot only needs one active mechanism to pass through the pipelines. So, this robot has two modules one active and one passive module. The active module uses motors and the passive module uses springs for compressing the legs. M O T˘atar et al. (T˘atar and Pop 2016) present a robot that can adapt to pipelines having an inner diameter of 220–380 mm. The three wheels are placed 120 degrees apart from each based on the central axis. A passive approach employing elastic components is employed to adapt to the pipe’s internal surface. Wen Zhao et al. (2021) designed a robot that has four wheels touching the ground and it is used for inspecting underground facilities. Except for the steering, it looks just like a car. It has a leader robot and a follower robot. The follower robot follows the leader robot and the connection is set up wirelessly. Atsushi Kakogawa et al. (Kakogawa and Ma 2018) developed an articulated robot made of spherical wheels that can easily adapt to varying diameter pipelines. It bends its body with the help of a torsion spring and uses its force to make the wheel touch the inner surface of the pipeline. It uses spherical and omnidirectional wheels to rotate its body and come out of the pipeline if it gets stuck in a cramped space. H. Tourajizadeh et al. (2021) designed a robot with a manipulator for repairing the pipelines. He also designed an impedance control for the robot’s manipulator. The design comes under wheeled type and combines the wall press mechanism for effective functioning. Three wheels are used to keep the stability of the robot inside the pipeline. Elizabeth Islas-García et al. (2021) designed a robot with wheels that imitates a spider. This robot uses the wheels to apply pressure on the walls of the pipeline for stability. It is designed in such a way that it can pass through horizontal and vertical pipelines. The parts of the robot are 3D printed and they use springs to compress and expand. It uses omnidirectional wheels, so when it meets any kind of obstacle it can easily rotate about the circumferential direction to overcome the problem. All robots mentioned comes under the wheel type and to crawl through vertical pipes they use the wall press–type mechanism which enables them to climb easily (Kwon et al. 2010; Kim et al. 2013; Min et al. 2014; T˘atar and Pop 2016; Zhao et al. 2021; Kakogawa and Ma 2018; Tourajizadeh et al. 2021; Islas-García et al. 2021). The authors of this paper modeled a robot using SolidWorks software to compare it with the existing work. It is made up of polyvinyl chloride as the central shaft. The two fixed joints, two prismatic joints, six small links, and six legs were 3D printed

38 Kinematic Modeling and Analysis of Wheeled In-Pipe Inspection Mobile Robot

1001

Fig. 3 Solid model of the pipeline inspection robot

using polylactic acid (PLA) filament. Six 100 rpm DC geared motors are used to send power directly to the wheels. Four stainless steel pipe clamps were used to lock the fixed joint, prismatic joint, and spring in place. Two springs with a stiffness of 2.2 N/mm are used to supply the required force to the prismatic joint so that the wheels touch the internal surface of the pipe which helps the robot to move through the inclined, curved, varying diameter, horizontal, and vertical pipes. Six straight wheels are given for forwarding motion as shown in Fig. 3. The three legs at the front and back give the robot better stability to move inside the pipeline. The wheels and motors are mounted on the robot legs. Rubber wheels are used to ensure high friction and traction between the wheels and the internal surface of the pipe.

Analysis: A Case Study Young-Sik Kwon et al. (2010) used Matlab to study the behavior of the wheeled mechanism when it passes through the pipeline. It is seen from the simulation that when the clutch wheels touch the pipeline the active and idle wheel lose contact with the pipe surface. Thus, the clutch mechanism is verified through simulation. Junghu Min et al. (2014) simulated his proposed robot and found its angular velocity, linear velocity, wheel velocity, motor voltage, and motor current. The environment used for his simulation is that the robot should follow through the sequence of pipelines. The sequence are as follows: straight (950 mm) – curved (703.63 mm) – straight (700 mm) – curved (703.63 mm) – straight (700 mm) pipeline. The linear velocity of 90 mm/s and angular velocity of 0.2 rad/s are used as boundary conditions for this simulation. It is found that the angular velocity of the robot is constant when it passes through the straight pipeline and is not the same when it passes through the curved pipelines. The speed of the wheels followed the same trend as the angular velocity.

1002

R. S. Elankavi et al.

M O T˘atar et al. (T˘atar and Pop 2016) did a simulation using SolidWorks software with a pipeline having an inner diameter of 300 mm. The working of the robot mechanism is also verified along with the simulation. Atsushi Kakogawa et al. (Kakogawa and Ma 2018) simulated his articulated wheeled-type robot and found that the spring stiffness along with the wheel’s friction plays a vital role in making the robot move through vertical pipelines. It is also found that the maximum static friction force acts on wheel 2 and it shows that wheel 2 should be driven using a motor. H. Tourajizadeh et al. (2021) did a simulation using Matlab software to verify the kinematics and dynamics of the robot. The same is verified again using Adams software. It is seen that the results from both the simulation correlate with each other. Elizabeth Islas-García et al. (2021) carried out a static analysis to find the pressure applied by the legs on the pipe wall with the help of springs. The elasticity constant of the element is found along with the analysis. The authors of this paper carried out motion analysis using SolidWorks by making the robot pass through a pipeline having an internal pipe diameter of 300 mm and a 90-degree elbow. The parameters needed for this simulation are the spring stiffness, gravity, and speed of the motor. In both, the examples discussed below only wheels 2 and 3 are given power while the other wheels are free. Only two wheels were selected so that we could find the cause of the irregular motion and rectify the issue. Example 1 In the first instance, the wheels were mounted 120 degrees apart from each other as shown in Fig. 4. The angle of the wheels is taken from the literature (Kwon and Yi 2012; Zhao et al. 2020; Li et al. 2020; Aras et al. 2020). Case 1: The robot enters the straight pipe and reaches the elbow as shown in Fig. 4a. Case 2: The robot enters the elbow and the body tries to rotate about the z-axis. This is because the three legs of the robot are fixed to the sliding joint. Hence when pressure is applied to wheels all three legs try to compress at the same angle, thus one of the wheels is unable to maintain contact with the internal surface of the pipe as shown in Fig. 4b. Case 3: The robot fails to pass through the elbow because wheel 2 losses contact with the internal surface of the pipe as the robot tries to enter the elbow which is circled in red as shown in Fig. 4c. The angular velocity of wheel 2 vs time based on magnitude when wheels are placed at 120 degrees apart from each other is shown in Fig. 5. It shows that the angular velocity of the robot is constant in the straight pipeline and the red line in the plot shows the entry of the robot in the elbow and straight pipe. It is seen that there is no change in the magnitude of angular velocity (wheel 2) because the wheel losses contact with the internal surface of the pipe and hence the wheel starts to rotate by itself.

38 Kinematic Modeling and Analysis of Wheeled In-Pipe Inspection Mobile Robot

a

1003

b

c

Angular Velocity1 (deg/sec)

Fig. 4 The motion of the robot inside a pipeline when wheels are mounted 120◦ degrees apart

6 09

6 07

STRAIGHT

ELBOW PIPELINE

STRAIGHT PIPELINE

6 04

6 02

6 00 0.00

0.44

0.89

1.33

1.77

2.21

2.66

3.10

3.54

3.98

4.43

Time (sec)

Fig. 5 Angular velocity of wheel 2 vs time when wheels are placed at 120 degrees apart from each other

Example 2 In the second instance, the robot wheels are mounted at an angle of 120◦ , 104.88◦ , and 135.12◦ apart from each other, as shown in Fig. 6. Case 1: The robot enters the straight pipe and reaches the elbow as shown in Fig. 6a.

1004

a

R. S. Elankavi et al.

b

c

Fig. 6 The motion of the robot inside a pipeline when the wheels are mounted at an angle of 120◦ , 104.88◦ , and 135.12◦ apart from each other

Case 2: The robot enters the elbow and maintains contact between the wheel and the internal surface of the pipe, allowing it to move through the elbow by arresting the rotational motion of the body about the z-axis as shown in Fig. 6b. Case 3: The robot exits the elbow and enters the straight pipe without the body rotating about the z-axis as shown in Fig. 6c. The angular velocity of wheel 2 vs time based on magnitude when wheels are placed at an angle of 120◦ , 104.88◦ , and 135.12◦ apart from each other is shown in Fig. 7. It shows that the angular velocity of the robot is constant in the straight pipeline and the red line in the plot shows the entry of the robot in the elbow and straight pipe. The magnitude of angular velocity (wheel 2) increases when it passes through the elbow and comes back to the constant angular velocity when it again enters the straight pipeline. The application of this robot is to pass through straight and elbow pipelines for inspection without any irregular movement about the circumferential direction of the pipe which helps us to find the exact location and distance traveled by the robot inside the pipeline.

Angular Velocity1 (deg/sec)

38 Kinematic Modeling and Analysis of Wheeled In-Pipe Inspection Mobile Robot

1005

6 13

ELBOW

6 09

STRAIGHT PIPELINE

STRAIGHT PIPELINE 6 05

6 00

5 96 0.00

0.65

1.30

1.95

2.60

3.25

3.90

4.54

5.19

5.84

6.49

Time (sec)

Fig. 7 Angular velocity of wheel 2 vs time when wheels are placed at an angle of 120◦ , 104.88◦ , and 135.12◦ apart from each other

Comparison of Motion Study The motion study in general is the study of motion. In our case, the motion of the robot inside the pipeline is studied. Each robot follows different types of motion based on its design. Some use a four-bar mechanism and others use a telescopic mechanism. There are other mechanisms applied apart from these and each gives different results during motion study. So, motion study is important for studying the behavior of robots inside pipelines. Young-Sik Kwon et al. (2010) tested the robot by making it pass through pipelines having six elbows. First, the robot’s driving mode is checked by making it pass through the pipeline in a given sequence, that is, curved-straight-curvedstraight-curved-straight-curved-straight pipeline. Then the robot retrieval mode is checked after the robot passes through all pipelines and reaches its destination. This is done by activation the clutch wheel and pulling the wire that is tied at the back end of the robot. It is seen that the retrieval of the robot is very much easy when activating the clutch wheel when compared to the retrieval of the robot using the active wheel and sometimes, we can’t even retrieve the robot. Thus, the driving and retrieval model of the robot is verified experimentally. Ho Moon Kim et al. (2013) verified the performance of the robot by making it pass through horizontal, vertical, and curved pipes. The curved pipes consist of both horizontal and vertical curved pipes with a curvature of all curved pipes being 225 mm. It is seen from this research that each wheel’s velocity differs when passing through a curved pipeline. The velocity of wheels when passing through the straight pipeline is constant. Thus, there is a difference seen in theoretical velocity compared to the actual velocity. Junghu Min et al. (2014) used the same environment that is used in the simulation. The performance of the robot and the proposed PID controller is verified experimentally. The robot legs compress individually when it crosses the curved pipeline and this helps the robot maintain its stability by always maintaining the central axis of the robot with the central axis of the pipe in a straight line. When it passes through the straight pipeline the angle at which each leg of the

1006

R. S. Elankavi et al.

robot is bent is equal on all three sides. When the robot crosses the curved pipeline one leg compresses at a different angle to the other because of the curvature of the curved pipe. This is where the spring and the design of the robot play the main role in making the robot pass through varying diameter pipelines. M O T˘atar et al. (T˘atar and Pop 2016) verified the forward and backward motion of the robot experimentally. This robot can be deployed on horizontal and vertical pipes. Wen Zhao et al. (2021) tested the performance of the robot chain control system (RCCS) based upon the visible light communication (VLC) method. It is seen that the robot provides inspection with high security and decent quality of communication. Atsushi Kakogawa et al. (Kakogawa and Ma 2018) tested the robot by making it pass through a pipeline having branched pipes. When the robot needs to pass through branched pipes it needs to roll inside the pipeline and change the direction of the bent body to the direction it needs to turn. Thus, the steering of the robot is easily achieved using the help of omnidirectional and spherical wheels. Elizabeth Islas-García et al. (2021) developed a prototype using 3D printing technology and omnidirectional wheels were attached to each of the motors which helps the robot to roll inside the pipeline. It is seen that the robot occupies most of the space inside the pipeline. After observing, it is decided that the robot parts, motors, and camera size should be reduced. A case study is done to observe the motion and function of the robot inside the pipeline and a prototype is developed by the authors of this paper. The robot when passing through the straight and curved pipeline compresses the legs equally unlike the robot developed by Junghu Min et al. (2014) where the legs of the robot compress unequally. This robot leg compresses equally even when passing through the curve because of its prismatic joint design where all the three legs of the robot are fixed to a single joint. The simulation result is verified by the motion study and it is observed that the change in the angle of the mounted wheels allowed it to maintain contact with the internal surface of the pipe and thus during the experiment, the body did not rotate around the z-axis. The navigation of robots inside the pipeline is shown in Fig. 8. Although various motion studies were conducted on wheeled-type in-pipe inspection robot. The study done by Junghu Min et al. (2014) is very similar to the case study done here. The comparison between the two motion studies is given below: (vii) In our motion study, the robot passes through one curved pipe and in the study conducted by Junghu Min et al. (2014) the robot passes through two curved pipes and three straight pipes. (viii) The angular velocity at the elbow with respect to the x-axis increases (Min et al. 2014). (ix) The linear velocity of the robot took about 18 s to catch up to the reference linear velocity (Min et al. 2014). (x) The velocity of each wheels changes when it passes the elbow and remains constant when passing through a straight pipeline (Min et al. 2014). (xi) The results of motor current and motor voltage were considered in their study (Min et al. 2014) and in our case study, it is not considered.

38 Kinematic Modeling and Analysis of Wheeled In-Pipe Inspection Mobile Robot

1007

Fig. 8 Navigation of robots inside pipeline (Min et al. 2014)

(xii) The legs of the robot compress individually (Min et al. 2014) unlike the robot done for the case study. (xiii) The robot (Min et al. 2014) has a PID controller and the case study robot does not have one. (xiv) The legs of the robot (Min et al. 2014) compress up to 200 mm and the robot in the case study can compress up to 100 mm.

Conclusion Various wheeled wall-pressed-type in-pipe inspection robots were designed, developed, and researched during the past decade. In this chapter, a case study is done to study their ability to accommodate pipelines of varying diameters using different mechanisms. The kinematics for the wheeled wall-pressed-type robot is studied from previous literature and derived, which for a given linear velocity (Vcz ) and angular velocity of the body of the robot ωx and ωy. the angular velocity of the wheels (1, 2, and 3 ) of the robot is obtained. A case study is done by designing a wheeled wall-pressed-type in-pipe inspection robot that can compress and expand its leg up to 100 mm. That solid model of the robot is developed using SolidWorks and motion analysis is done. That when the wheels are placed 120 degrees apart the robot body tends to rotate at the elbow and this is rectified by changing the angles in which the wheels are mounted. The design is finalized and a prototype is developed

1008

R. S. Elankavi et al.

to observe the behavior of the robot inside the elbow and straight pipelines. It is observed that the robot locks the rotating motion of the body at the elbow caused when the wheels are placed 120 degrees apart. The steps used by researchers is studied and applied in the case study. The results from the case study correlated with the results from the motion study done by other researchers.

References Aras MSM, et al (2020) Design and development of remotely operated pipeline inspection robot. In: Proceedings of the 11th national technical seminar on unmanned system technology 2019. pp 15–23 Elankavi RS et al (2020) Developments in inpipe inspection robot: a review. J Mech Continua Math Sci 15(5):238–248. https://doi.org/10.26782/jmcms.2020.05.00022 Feng G et al (2020) Development of a wheeled and wall-pressing type in-pipe robot for water pipelines cleaning and its traveling capability. Mechanika 26(2):134–145 Hadi A et al (2020) Developing an adaptable pipe inspection robot using shape memory alloy actuators. J Intell Mater Syst Struct 31(4):632–647 Islas-García E et al (2021) Pipeline inspection tests using a biomimetic robot. Biomimetics 6(1):17 Kakogawa A, Ma S (2018) Design of a multilink-articulated wheeled pipeline inspection robot using only passive elastic joints. Adv Robot 32(1):37–50 Kim HM, et al (2013) An in-pipe robot with multi-axial differential gear mechanism. In: 2013 IEEE/RSJ international conference on intelligent robots and systems. IEEE, pp 252–257 Kusunose K et al (2020) Development of inchworm type pipe inspection robot using extension type flexible pneumatic actuators. Int J Autom Mech Eng 17(2):8019–8028 Kwon YS, Yi BJ (2012) Design and motion planning of a two-module collaborative indoor pipeline inspection robot. IEEE Trans Robot 28(3):681–696 Kwon YS, et al (2010) A pipeline inspection robot with a linkage type mechanical clutch. In: 2010 IEEE/RSJ international conference on intelligent robots and systems. IEEE, pp 2850–2855 Li H et al (2020) Development of a pipeline inspection robot for the standard oil pipeline of China National Petroleum Corporation. Appl Sci 10(8):2853 Liu C et al (2020) Travelling ability of pipeline inspection gauge (PIG) in elbow under different friction coefficients by 3D FEM. J Nat Gas Sci Eng 75:103134 Min J, et al (2014) Development and controller design of wheeled-type pipe inspection robot. In: 2014 International conference on advances in computing, communications and informatics (ICACCI). IEEE, pp 789–795 Nayak A, Pradhan SK (2014) Design of a new in-pipe inspection robot. Procedia Eng 97:2081– 2091 Ren T et al (2019) Driving mechanisms, motion, and mechanics of screw drive in-pipe robots: a review. Appl Sci 9(12):2514 Roslin NS, Anuar A, Jalal MFA, Sahari KSM (2012) A review: hybrid locomotion of in-pipe inspection robot. Procedia Eng 41:1456–1462 Savin S, et al (2018) State observer design for a walking in-pipe robot. In: MATEC web of conferences, vol 161, p 03012 T˘atar MO, Pop A (2016) Development of an in pipe inspection minirobot. In: IOP conference series: materials science and engineering. IOP Publishing, 147(1):012088 Tourajizadeh H et al (2021) Design, modeling, and impedance control of a new in-pipe inspection robot equipped by a manipulator. Sci Iran Trans B Mech Eng 28(1):355–370 Zagler A, Pfeiffer F (2003) MORITZ” a pipe crawler for tube junctions. In: 2003 IEEE international conference on robotics and automation, vol 3, pp 2954–2959 Zhang Y, Yan G (2007) In-pipe inspection robot with active pipe-diameter adaptability and automatic tractive force adjusting. Mech Mach Theory 42(12):1618–1631

38 Kinematic Modeling and Analysis of Wheeled In-Pipe Inspection Mobile Robot

1009

Zhang H et al (2020) Stress and strain analysis of spherical sealing cups of fluid-driven pipeline robot in dented oil and gas pipeline. Eng Fail Anal 108:104294 Zhao W et al (2020) Design and analysis of independently adjustable large in-pipe robot for longdistance pipeline. Appl Sci 10(10):3637 Zhao W et al (2021) A wheeled robot chain control system for underground facilities inspection using visible light communication and solar panel receivers. IEEE/ASME Trans Mechatron

Part III Industry 4.0: Applications

Cloud Hadoop for Enterprise Collaboration System

39

Hsiao Kang Lin and Tzu-Jou Liao

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hadoop for the Syntax and Semantics Cloud Data Integration . . . . . . . . . . . . . . . . . . . . . . . Extract Data from the Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Load the Extracted Data into the HBase Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transform/Off-Load Data Using Hadoop Hive’s Schema-on-Read . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1014 1015 1016 1016 1018 1020 1021

Abstract

Today, a vigorous growth in the creation of digital data communications in synergistic operating networks. Enterprise collaboration systems face the challenges of extracting, processing, and analyzing data from multiple cloud services. Distributed cloud data storage and retrieval via the predefined schema of structured databases has become increasingly inefficient with the advent of Big Data (varies in representation from structured to semi-structured and unstructured formats). The Hadoop Ecosystem System is being adopted to reduce the complexity of moving data to and from the cloud computing service. In the Hadoop approach, the solution supports Hadoop Database (HBase) external tables, thereby enabling users to access data residing on the Hadoop Distributed File System (HDFS). In addition, the agile code-free creation of Hadoop Extract-Load-Transform (ELT) data process to begin analyzing data in minutes, not days or weeks. A case study is conducted to illustrate that the Hadoop ELT approach for the

H. K. Lin () · T.-J. Liao Department of Industrial Management, I-Shou University, Kaohsiung City, Taiwan e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_22

1013

1014

H. K. Lin and T.-J. Liao

syntax and semantics cloud data integration and support the provision of flexible interoperation of global enterprise collaboration systems. Keywords

Enterprise Collaboration System · Extract-Load-Transform (ELT) · Hadoop Database(HBase) · Hadoop Hive

Introduction Enterprise collaboration systems have undergone accelerated changes, driven by the emerging hyperconnected operating systems that are increasingly relying on semantic knowledge services (Popplewell et al. 2008; Choudhary et al. 2019). The EU projects, in particular (FLEXINET: Intelligent Systems Configuration Services for Flexible Dynamic Global Production Networks 2013–2016), provide the software services for supporting enterprise collaboration. They apply advanced solution techniques to the provision of flexible interoperable networks of enterprise systems. For example, a reference ontology for Product-Service-Production and Business Models for Production Network of FLEXINET (Palmer et al. 2017, 2018) integrates different views to a common unified vocabulary and provision of flexible interoperable networks of global production systems. However, the participating enterprises within the Virtual Organizations (VOs) need to transform or map their origin database schema into the common schema/reference ontology/unified vocabulary to facilitate syntax and semantic data interoperability. The Hadoop Ecosystem has been adopted for the syntax and semantics cloud data integration (Nkenyereye and Jang 2017; Schmatz et al. 2018; Zhu and Xu 2020). It is an open source distributed software platform and consists of various components. There are two primary components at the core of Apache Hadoop for storing and processing data – the Hadoop Distributed File System (HDFS) and the MapReduce/Spark parallel processing framework, and related projects such as HBase, Pig, Sqoop, and Hive in a simpler, more scalable, and cost-efficient environment. The Hadoop also provides an alternative approach in which data is “Extracted” from the web sources, “Loaded” into the HBase database,” and then “Transformed” and integrated into the desired format in Hive (ELT). This chapter proposes a novel ELT approach in the form of a set of the Hadoop tools for information integration across a global enterprise collaboration system, without the need for mapping into the common/mediated model. A study was conducted to illustrate that the Hadoop could be adopted across the global production network of notebook and PCs.

39 Cloud Hadoop for Enterprise Collaboration System

1015

Hadoop for the Syntax and Semantics Cloud Data Integration Lin and Harding (2007) and Lin et al. (2012) observed that VOs project team members in different parts of the world each worked using their own preferred terminology. However, when people are brought together from different groups or companies, two common types of problem can occur in communications that share and exchange information. Firstly, that the same term is being applied to different concepts (semantic problem), and secondly, that different terms may be used to denote the same entity (syntax problem). In this chapter, we built a test case involving development of a cloud-based infrastructure for creating, operating, evolution, and eventually dissolving of flexible interoperable within global notebook and PCs enterprises. Figure 1, for example, shows that the column heading “ID” used by Enterprise A has the same meaning as “LineItem” used by Enterprise B: These two different identifiers exist within different models, but mean the same thing. For the Common Ontology Model they propose the single term “Component_ID” that is stored on pre-agreed schema within the collaborative VOs, while simultaneously sharing information through the transforming/mapping mechanisms of the common ontology model. The operation of the Hadoop ELT process is now demonstrated through the example of new VO projects/new collaborative production in a manufacturing

Fig. 1 Predefined schema transformed into the common ontology model for data integration

1016

H. K. Lin and T.-J. Liao

resource e-planning task, using Microsoft Azure HDInsight Service. It is a cloud PaaS (platform as a service) that deploys and provisions Hadoop clusters for the HDFS/MapReduce/Spark software framework and related projects such as HBase, Pig, Sqoop, and Hive. Three main steps of the Hadoop ELT process are: extract data from the web, load the extracted data into the HBase tables, and transform/off-load data by Hadoop Hive’s schema-on-read.

Extract Data from the Web The growing popularity of enterprise collaboration systems on the World Wide Web has resulted in a huge amount of information sources via the Internet. It is necessary to analyze this big volume of data and extract useful information from the web. Chang et al. (2006) pointed out that the Internet presents large amounts of potentially useful information, and it is usually heterogeneous and lacking in structure. To automate the translation of input pages into structured data, considerable effort has been directed toward the area of information extraction (IE). IE produces structured data ready for post-processing, which is crucial to many applications of searching and web mining tools. Moreover, various approaches (Ferrara et al. 2014) have been proposed to efficiently collect the data with limited human effort. Recently, several tools have been developed that allow for Web Data Extraction (WDE) using a simple point-and-click interface to automatically enable the process of extracting specific text and images from any website into text files, spreadsheet, or a wide variety of formats, for example, Mozenda (http://www.mozenda.com) and Microsoft Excel and Win automation (http://www.winautomation.com). In this chapter, Microsoft Excel’s get external data feature (choose Data > From Web) is used to focus on WDE from semi-structured documents. Figure 2 shows a semi-structured web page containing table data to be extracted into the format of individual partners, and the extracted data is then loaded into our designed HBase table – VO Production_Resource (VOPR) HBase table (see Table 1).

Load the Extracted Data into the HBase Tables It is well known that the key characteristic of HBase is called “no schema – on-write,” which means such systems do not require to the data schema to be predefined before loading data into HBase (Hou et al. 2015; Chen and Lee 2017). HBase may offer a better solution for web data extraction and web mining analysis. According to features proposed by the application, a VOPR table was created with multiple Column Families (CFs) – enterprise A and enterprise B, etc. Each partner has been designed to have his own CF that is physically stored together on the same HDFSs. Moreover, CF can be incrementally created while new numbers of partners join the project. In HBase, each CF has one or more non-predefined Column Qualifiers (CQs). Therefore, the individual project partners can use their own preferred terminology in the newly created CQs.

39 Cloud Hadoop for Enterprise Collaboration System

1017

Fig. 2 A semi-structured page containing table data to be extracted from web page

For example, enterprise A can load its own model and original terminology into the VOPR by adding four new CQs: EName, ID, Desc, and Quantity with value of “Micron,” “A9008XUIQ,” “Micron 1TB SATAIII,” and “116,” respectively. Similarly, enterprise B can load its original terminology into the VOPR by adding four new CQs: CName, LineItem, P_Desc and Units with value of “USIO,” “A9009I2F9,” “9 Micron 500GB PCIe,” and “213,” respectively. Table 1 illustrates how the individual partners can each load their original model into the HBase table. Records in HBase are stored in the HDFS as: RowKey-value pairs (RowKey: value) → (NB SSD: Micron), (NB SSD: A9008XUIQ), (NB SSD: Micron 1TB SATAIII), (NB SSD: 116), etc.

The HDFS itself is a binary file and is not human-readable. HDInsight Service uses Azure Blob storage as the big data store for HDFS. In this case, the HDFS file for the VOPR HBase table store resides in a default file system with Azure storage account for the cluster looking something like Table 2: vopr.txt.

1018

H. K. Lin and T.-J. Liao

Table 1 The logical VO Production_Resource (VOPR) HBase table

Table 2 The vopr.txt HDFS file

Transform/Off-Load Data Using Hadoop Hive’s Schema-on-Read Many traditional data systems use the so-called schema-on-write model where users need to decide on the schema of their data before loading data into their system. This model would start out by understanding how the data needed to be used, then design appropriate schemas, and manipulate the data to fit those schemas. In contrast, “schema-on-read” models do not need to know how the data will be stored. This means that “schema-on-read” models can load/write data first and then apply the structure of a schema to the data “on read.” One of the most-cited advantages of the Hadoop Ecosystem is that it enables a “schema-on-read” data analysis strategy. As shown in Table 2, HDFS files allow for the storing all the data without understanding what the data elements are (No schema has been defined yet). A schema to fit the data needs is built later (see Fig. 3). The Apache Hive™ facilitates querying, managing and analysis of large datasets and provides an SQL-like query language called HiveQL (https://cwiki.apache.org/ confluence/display/Hive/GettingStarted). In addition, the Apache Hive™ enables users to build a custom schema and directly query self describing data. This approach enables data to flow into the system in its original form/native format,

39 Cloud Hadoop for Enterprise Collaboration System

1019

Fig. 3 Create voprHive table with new custom schema

and then the schema is parsed at read time, allowing for extreme agility while dealing with complex evolving data structures. In this chapter, Hive’s schema-onread capability is demonstrated to show how data is loaded into HBase tables, stored in HDFS, and managed in the Hive environment, using the Microsoft Azure cloud platform. In particular, Azure HDInsight deploys Hadoop clusters, which include HBase and Hive as well as other technologies under the Hadoop Ecosystem. We created an HDInsight cluster on Azure and ran HiveQL jobs using the HDInsight query console. This case study describes the creation of:

Loading Data from HDFS to External Hive Table The first task is to load data from HDFS to Hive. An external temp_vopr table is created and loaded with the native format data from vopr.txt HDFS files (see Table 2) into the temp_vopr Hive table. The completed query is shown in Fig. 3, which shows the Hive’s command in the HDInsight Hive query console. Creating a New Internal Hive Table with a New Custom Schema The second task is to create a new internal Hive table called “voprHive” and add “Custom Schema” to fit the data in the temp_vopr table. The new “voprHive” table has four columns, namely, Enterprise Name (EName), (Component_ID), Resource Description (rDesc), and the currently available volume (Volume), shown in Fig. 3. Extracting the Data from an External Hive Table and Copying It into the New Schema in an Internal Hive Table The third task is to extract the data from temp_vopr and copy it into the new schema in voprHive table. To do this we build up a multi-line query. The four regexp_extract calls are to extract the EName, Component_ID, rDesc, and Volume fields from temp_vopr, as shown in Fig. 4. As mentioned earlier and shown in Table 2, records in the HDFS are: key-value pairs (rowkey: value) → (NB SSD: Micron), (NB SSD: A9008XUIQ), (NB SSD: Micron 1TB SATAIII), (NB SSD: 116), etc. The regexp_extract function in Hive is used to extract the rowkey-value pair and store the value into the newly created custom columns. The Hive built-in regexp_extract function:

1020

H. K. Lin and T.-J. Liao

Fig. 4 Add data into voprHive from temp_vopr

regexp_extract (string subject, string pattern, int index)

It is used to read data and returns the string extracted using the pattern. The “col_value” key is NB SSD and returns first pattern value: Micron, and then stores it into the EName field. And then, the “col_value” key is NB SSD and returns second pattern value: A9008XUIQ, and then stores it into the Component_ID field, etc. (see Fig. 4). The final task is to execute the query and look at the new “voprHive” table with the new custom schema in Hive: Select EName, Component_ID, rDesc, Volume from temp_vopr,

Once the status shows that the job is completed, click the query name on the screen to see, the voprHive table with the new custom schema, as shown in Fig. 5.

Conclusion The Hadoop Ecosystem running in the cloud platform is generating new opportunities and presenting new challenges for businesses across every industry sector. The challenges of data integration incorporating data from the web and other unstructured data from multiple sources are the most urgent issues facing the successful implementation of enterprise collaboration system. Based on our evaluation, using the Microsoft Azure cloud platform and Hadoop tools for enterprise ELT processes to achieve collaborative production in an example of a manufacturing

39 Cloud Hadoop for Enterprise Collaboration System

1021

Fig. 5 The voprHive table with new custom schema

resource e-planning task, the greatest benefits derived are in the areas of improved performance and functionality. There is still however a time and cost overhead in this method as individual partners need to extract the rowkey-value pair and store the value into the newly created custom columns. Using the current manual methods this can be slow. Hence, a limitation in the research reported in this chapter is the current manual mapping process which is very ineffective and may cause major barriers to the large-scale use in information integration for global supply chain’s network. However, likely future advances in this area should reduce this overhead. These include features for formal mapping representation, such as algorithms and heuristics to identify similarities between the two value, machine learning to value match, and knowledge discovery. These topics are therefore recommended for future investigation.

References Chang CH, Kayed M, Girgis MR, Shaalan KF (2006) A survey of web information extraction systems. IEEE Trans Knowl Data Eng 18(10):1411–1428 Chen J, Lee W (2017) Data conversion from RDB to HBase. In: 2017 IEEE 8th International Conference on Awareness Science and Technology (iCAST) Choudhary A, Harding J, Tiwari MK, Shankar R (2019) Knowledge management based collaboration moderator services to support SMEs in virtual organisations. Prod Plan Control 30(10–12):951–970 Ferrara E, De Meo P, Fiumara G, Baumgartner R (2014) Web data extraction, applications and techniques: a survey. Knowl-Based Syst 70:301–323 FLEXINET (2013–2016) FLEXINET: intelligent systems configuration services for flexible dynamic global production networks. EU grant agreement no. 608627 from home (flexinetfof.eu) Hou Y, Yuan S, Xu W, Wei D (2015) Transformation of an E-R model into HBase tables: a data store design for IHE-XDS document registry. In: 2015 IEEE 12th international conference on ubiquitous intelligence and computing and 2015 IEEE 12th international conference on autonomic and trusted computing and 2015 IEEE 15th international conference on scalable computing and communications and its associated workshops (UIC-ATC-ScalCom)

1022

H. K. Lin and T.-J. Liao

Lin HK, Harding JA (2007) A manufacturing system engineering ontology model on the semantic web for inter-enterprise collaboration. Comput Ind 58(5):428–437 Lin HK, Harding JA, Tsai WC (2012) A rule-based knowledge system on semantic web for collaboration moderator services. Int J Prod Res 50(3):805–816 Nkenyereye L, Jang J (2017) Integration of big data for querying CAN bus data from connected car. In: 2017 ninth International Conference on Ubiquitous and Future Networks (ICUFN) Palmer C, Urwin EN, Young RIM, Marilungo E (2017) A reference ontology approach to support global product-service production. Int J Prod Lifecycle Manag 10(1):86 Palmer C, Usman Z, Canciglieri Junior O, Malucelli A, Young RIM (2018) Interoperable manufacturing knowledge systems. Int J Prod Res 56:1–20 Popplewell K, Stojanovic N, Abecker A, Apostolou D, Harding J (2008) Supporting adaptive enterprise collaboration through semantic knowledge services. In: Proceedings of the 4th international conference on interoperability for enterprise software and applications, IESA 2008, Berlin, 26–28 Mar 2008 Schmatz K, Berwind K, Engel F, Hemmje ML (2018) An interface to heterogeneous data sources based on the mediator/wrapper architecture in the Hadoop ecosystem. In: 2018 IEEE international conference on bioinformatics and biomedicine (BIBM) Zhu H, Xu Y (2020) Sports performance prediction model based on integrated learning algorithm and cloud computing Hadoop platform. Microprocess Microsyst 79:103322

Smart Devices in Healthcare Sector: Applications

40

Kanika Sharma, Payal Kesharwani, Shiv Kumar Prajapati, Ankit Jain, Neha Mittal, Rahul Kaushik, and Nishi Mody

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Orthopedic Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wearable Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Inhalers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pulse Sensor Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Temperature Sensing Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of Smart Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Sensor for Gastrointestinal Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Sensors for Liver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Sensors for Lungs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Sensor for Eyes Surgery and Its Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Sensors for Brain Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cardiac and Motor Activity Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Sensors for Parkinson’s Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Devices in Management of Metabolic Disorder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion and Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1024 1027 1027 1029 1029 1031 1032 1032 1032 1034 1034 1035 1036 1038 1040 1042 1044 1045 1046 1046

K. Sharma · P. Kesharwani · S. K. Prajapati · N. Mittal · R. Kaushik Department of Pharmacy, Ram-Eesh Institute of Vocational and Technical Education, Greater Noida, Uttar Pradesh, India A. Jain () Department of Materials Engineering, Indian Institute of Science, Bangalore, Karnataka, India e-mail: [email protected] N. Mody Department of Pharmaceutical Sciences, Dr. Hari Singh Gour Central University, Sagar, Madhya Pradesh, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_23

1023

1024

K. Sharma et al.

Abstract

Today the world is surrounded by an embedded system to build an easier, convenient, and effective life. The expansion of the Internet of things (IoT) and the industrial cyber-physical system established by Industry 4.0 concept has flourished man’s life to bloom up stage by providing benefits to the healthcare and biopharmaceutical sector with new research and development. Devices with smart sensor act as an auxiliary therapy in diversified cases with modern tools and technology which have refined mankind. With the usage of smart devices, one can now monitor real-time data and get the accurate biological functioning of the body. Nowadays, wristband equipped with an accelerometer, optical sensor, and electrodes is most trending among people equipped for predicting the body functions by sensing temperature, blood pressure, heart rate, sleep pattern, etc. This chapter highlights the potential applications of smart devices in sundry clinical scenarios. This chapter focuses on various smart sensors such as smart orthopedic devices, smart inhalers, insulin devices, pulse sensors, temperature sensors, cardiac and motor activity monitors and device sensing mechanisms. Along with this, the employment of these sensors in the healthcare sector for different body organs like lungs, liver, gastrointestinal tract, brain, and heart has been discussed. Moreover, various challenges faced by the user, in conjunction with its future perspective, are emphasized. Keywords

Industry 4.0 · Smart sensors · Wearable devices · Robots · Neurodegenerative disease · Metabolic syndrome

Introduction The embedded system technology has the world in its grip owing to variegated benefits along with a range of development and paradigms shift. It utilizes advanced computing technology at a cost-effective range and a better network which connects people all around the world (Silvestre-Blanes et al. 2020). The healthcare industry is one of the fastest growing sectors all over the world which not only provides critical services to patients but also contributes huge revenue to a country’s economy. The integration of Internet of things (IoT) into embedded technology equipped with smart healthcare sensor devices helps the patients to obtain an easy, low cost, fast treatment with modern facilities. The various IoTs are being now interconnected to cloud technology to solve the issue faced due to low storage and limited processing capacity. The cloud technology provides huge storage capacity for data and its processing and analysis (Amin et al. 2019). The unfolding of various new sensing systems has stepped its significant footprint over technology ranging from the industrial society to business, and academics to diagnosis; management and treatment of various illnesses has been simplified because of the commencement

40 Smart Devices in Healthcare Sector: Applications

1025

Fig. 1 Number of publications relating to smart sensors in the last 5 years

of Industry 4.0 (Silvestre-Blanes et al. 2020). This advancement is evident from the rising number of publications in the last few years (Fig. 1). This has upgraded human life by increasing the average life span of a person gifting many more valuable years to lead life happily. The healthcare monitoring devices with sensors have enhanced the diagnostic capabilities of diseases, real-time monitoring during treatment, and generalized monitoring of patient’s health. The number of people who take advantage of wearable devices has widely increased from 300 million since 2016 to 1,000 million until 2022. The conventional devices used for diagnosis purposes are costly, inconvenient for the patients, external expertise is required, daily monitoring is not possible, the wired system makes it messy, it has network connectivity issues, etc. (Jin et al. 2017). The smart grid achieved by sensing devices is increasingly becoming very popular among the people as they possess powerful sensing capabilities to analyze the real-time performance of a device in detecting the potential failure (Morales-Velazquez et al. 2017). Such an approach ameliorates the health monitoring structure together with a data analyzer that possesses the skills to identify high fidelity and detect damage using hardware and software elements. The admirable properties such as self-diagnosis, better connectivity, and convenient to use and carry with accurate and reliable sensing signals have captured the market. The sensing devices used can be simple sensors (passive infrared sensor), wireless sensing devices (wrist communicators and motion detectors), and sensors fitted to camera (machine interpretation of an image). The smart sensing device consists of an embedded microcontroller, micro-processor, algorithmic analyzer to convert the signals, proliferation network, and diagnostic software. The sensors fitted in the

1026

K. Sharma et al.

devices should be highly flexible and ultra-thin, have lower modulus, be lighter in weight and stretchable, and consist of electrode and substrate embedded inside. The electrode is made of hybrid material having carbon nanofibers and nanomaterialbased multi-dimension (An et al. 2017). The microprocessor filters the multiplexed signals and helps in driving useful information. The useful signals are passed to a digital converter and monitored by data acquisition computer. The analyzed data is stored in a digital signal processor having a memory storage capability. Various connecting devices of sensor and database system such as Bluetooth, maglev porous nanogenerator (MPNG) (Jin et al. 2017), and many other wireless sensor networking protocols help in transmitting data to a mobile or computer database depicting the IoT. It has the likelihood of obtaining bio-signals from sensor nodes and transferring it to gateway channels through wireless communication portals. The real-time information received by a database undergoes real-time processing, visualization, and diagnosis. This sensing device is used medically in various disorders like diabetes, Parkinson’s disease, cardiac heart failure, lung and liver diseases, kidney disorders, etc. The main utilization of this sensing device is its utilization in diagnosing the dysfunction of a particular organ by analyzing its pattern of response and helping in early treatment if required. A wearable electronic device can monitor the patient’s health condition, for instance, heart rate, wrist pulse, motion, blood pressure, intraocular pressure, concentration of glucose in blood, patient strain, exhaled gas composition, different ions, and biomolecules that exist in bloodstream, etc. (Kim et al. 2015). The optical fiber is used for preparation of sensors. These are light in weight, chemically stable, have multiplexing capability, and are compactly embedded in a soft structure in a wearable device. Various physical parameters such as angle, refractive index, temperature, humidity, acceleration, pressure, and oxygen level can be used for diagnosis. Different materials such as are polylactic acid (PLA) or acrylonitrile butadiene styrene (ABS), thermoplastic polyurethane (TPU), etc. can be used along with optical fibers for the D printing technology used in sensors. Various applications can be employed using optical fiber sensor such as bone decalcification study, stress distribution, intervertebral disc evaluation, dental splint, cardiac monitoring, pathogen detection, pressure analysis, and prosthesis socket, joint angle, and plantar pressure monitoring (Leal-Junior et al. 2019). Figure 2 depicts the applications of sensors helping in the diagnosis of dysfunctions in various body organs. This chapter will discuss the tenet of smart sensors in various medical fields. Various devices and utilization of smart sensors in variegated fields have also been discussed. The potential applications of smart devices in sundry clinical scenarios have been discussed. Various smart sensors like smart orthopedic devices, Fitbit watches, smart inhalers, insulin devices, and cardiac and motor activity monitors which enhance quality of life of users are discussed in this chapter. Moreover, this chapter also covers perfected robotic surgeries in variegated fields. Wearable devices in management of neurodegenerative and metabolic disorders are also emphasized. Various challenges that are encountered during the usage of sensors by consumers have been thoroughly debated.

40 Smart Devices in Healthcare Sector: Applications

1027

Fig. 2 Application of sensor helping in diagnosis of dysfunctions in various body organs

Devices Orthopedic Devices Orthopedic disorders are assessed to be a minimum of 4.4% of total diseases and is serious detriment to major population worldwide especially the older ones (Barrett and Liebman 2020). The dysfunction of bone as well as ligaments includes injuries or musculoskeletal diseases affecting bones, joints, muscles, connective tissues, tendons, ligaments, and also the nervous system, with a maximum percentage of patients having osteoarthritis (OA) and osteoporosis. Injuries can be marked by a bone fracture or a ligament tear known as repetitive strain injury or syndromes such as cubital tunnel syndrome and lateral/medial epicondylitis. The convention method used for treating such conditions may be temporal resolution and nonintrusive giving pain to the sufferer and they are also time consuming. Orthopedic sensors are employed to monitor the recovery of bone as well as tissue regeneration, the real-time healing capacity of bone and ligaments, and also to check physical and chemical conditions of bones (Karipott et al. 2018). Tribocorrosion is frequently encountered by patients who have a surgically implanted metal device inside their body. Tribocorrosion is an amalgamation of mechanical process (wear) and chemical process (corrosion) (Mathew M.T. et al, Wear 2011). About 2.5 million people in America have implanted with a metal device, and 10% of them suffer due to implant failure. The implant devices often release metal ions in body leading to tribological activity in a corrosive internal environment. The discharge of metal ions is undetected until the threshold of pain felt by the individual is reached due to adverse tissue reaction. The unsuccess of implant devices needs an urgent flourish of a fast-reliable technique which monitors

1028

K. Sharma et al.

Fig. 3 The two modules shown work together to obtain real-time drilling angle information, to locate the distal hole in real-time (Choi et al. 2017)

the leakage of the metal ions in orthopedic patient (Pichetsurnthorn et al. 2012). The sensor aids in metal ion detection and monitors the clinical functioning of the device (Bolotsky et al. 2019). The orthopedic patient can routinely check their condition at home. The concept of electrochemical impedance along with open circuit potential test helps in detecting the changes while dissolving metal ions in phosphate-buffered saline and bovine calf serum as the reference solution. The eddy current and the associated magnetic field provide a valid prediction of the concentration of metal ions (O’Hara et al. 2013). Thus, a smart biosensor is an advantageous tool to identify the metal ion concentrations in orthopedic patients. The navigation system together with sensor also helps to give a real-time data of a drilling angle to locate distal hole in a damaged area through the skin for intramedullary nailing in orthopedic surgery (Choi et al. 2017). Older radiography in accordance to 2D fluoroscopic image limits its use as a consequence of the exposure of unwanted radiation for extended time. In new intervention, the module is involved in handling the integrated laser guidance guided by nine axes inertial sensor with a Bluetooth connector for targeting the point of insertion in the skin. This technique minimizes the rotation and dislocation of the patient’s fractured bone and has a great potential toward orthopedic surgery (Choi et al. 2017). Figure 3 demonstrates the working of sensor in finding the real-time drilling angle by locating the distal hole. Practically, the physician cannot monitor the healing process in a fractured bone. The sensor devices become a gateway that allows close monitoring of fractured bone healing process. The optical fibers having strain and shear sensor can measure the pressure felt at the knee amputee socket and help in relief of pain. Wireless sensors

40 Smart Devices in Healthcare Sector: Applications

1029

are also inserted along with implant to examine the real-time chemical and physical context of bones, tendons, muscles, and nearby tissues to study the disease/injury to the area. Currently, this technique is under investigation by scientists but can practically soon be available with the effort (Karipott et al. 2018).

Wearable Devices Wearable devices are electronics equipped with sensors which can be worn on the body to collect different kinds of health data by sensing the movement, pulse rate, heart rate, etc. The different wearable devices are available for their respective purpose such as wearable sensor robotic device, sensor walker device, plantar pressure device, accelerometer, pedometer, etc. The wearable robotic device is also used for gait rehabilitation where the sensor device is fit to lower limb (de Fátima Domingues et al. 2020). The sensor walker can be classified into four categories: sensor for upper limb, sensor for kinetic assessment of gait, odometry sensor, and sensor used for environment testing. The foot plantar pressure is used to detect the pressure generated due to ulceration in foot caused by increase in sugar level. An accelerometer is fitted inside a sensor that examines the movement or a heart rate monitor that measures heartbeat. A pedometer is a trending wearable device used by many health-conscious people, which counts the number of footsteps, distance traveled, and speed covered by a person. Its significant effect for managing neurodegenerative diseases like Parkinson’s Disease and metabolic disorders like obesity and diabetes may also be obtained (Lu et al. 2020). The wristband sensor can be used to monitor daily activities of an epileptic patient and help in clinical assessment required during emergency. Due to less monitoring of body functions, there always remains a gap between the initiation of symptoms and administration of pharmacological therapy. This hiatus can be overpowered through a wearable device as it regularly monitors the various bodily data. The wearable sensor transmits the change recorded for real-time clinical assessment. This helps in monitoring intervention response, providing customized care, and recommending medication (Fig. 4) (Godoi et al. 2019).

Smart Inhalers Smart inhaler is a sensor-based device for measuring the airflow of inhaled gas and it controls the drug release. Targeting the drug to the affected area of lungs is always a tough task in respiratory ailments such as asthma and bronchial inflammation. Traditional gadgets like pressurized metered-dose inhaler (pMDI), an aerosol inhaler, are commonly used to deliver medicament to the lungs. The efficiency to deliver a drug is reduced as turbulent mixing produces a large deposition of particles near the larynx area, making the drug unable to reach the lungs (Pausley and Seelecke 2008). To surmount such an issue and for smooth distribution of drug in the lungs, the development of smart sensor devices is emerging as a muse visible area in

1030

K. Sharma et al.

Fig. 4 Working of wearable devices in providing wholesome medical care based on remote monitoring of clinical manifestations in patients

pharmaceutical digital technology. The sensor-based technology results in accurate dose delivery to the target site and adherence and also stores data received through IoT in a mobile or a laptop database. Smart inhaler is a wave of improvement for respiratory care technology for improving patient convenience and outcomes. The sensor and microprocessor are equipped in devices for inhalation such as meterdosed inhaler, inhalation sensor, smart mist device, MDILog etc. (van der Kamp et al. 2020).

Metered Dose Inhaler A metered dose inhaler (MDI) is used to deliver the prescribed drug in diseases like bronchospastic conditions including asthma. The dosage form like solution, emulsion, or suspension including active ingredient can be administered along with propellant through a pressurized canister. The asthma attack which involves bronchodilation can be life-threatening, if an accurate, adequate, consistent dose with reproducibility is not delivered at each actuation as indicated on the label. Therefore, a sensor-based inhaler device holds a promise to support and optimize self-assessment of asthma by counting the doses delivered by MDI. The MDI is also introduced with a miniaturized pressure sensor with a microprocessor. The pressure sensor measures the pulse generated at the transfer channel fitted in a mouthpiece guiding the discharge of formulation in a metered dose. The microprocessor keeps the record of the dynamics of pressure pulse along with the count of doses administered through a digital display or audible signal present in a device. It also avoids the error generated due to false and missing counts, thus recording the reliable dose delivery (Rocci Jr et al. 2000).

40 Smart Devices in Healthcare Sector: Applications

1031

Inhalation Sensor The traditional cigarettes or alternative chewing gums and nicotine patches are having some limitations such as cigarettes produce discomfort (carbon monoxide, coughing, heat, etc.) through the smoke produced while nicotine chewing gums and patches may deliver increased amount as they produce sensationless effect due to inhalation of vaporized nicotine solution which increases the urge of consumption. The inhalation sensor is now invented to conquer the shortcomings of traditional use and helps to deliver the required amount of diluted nicotine or (Tetrahydocannabinol) THC solution to the consumer. The sensor is fitted with a processor, and a memory storage is in-built in it which keeps the record of the amount of nicotine solution or THC delivered to the consumer. This helps to diminish the unwanted consumption of nicotine in a safe and efficient manner (Spinka and Spinka 2015). Smart Mist Device The smart mist device possesses the sensing potential of monitoring the inhalation pattern and records the inspiratory flow grade and inspiratory firing volume. The device reasons the actuation mechanism of releasing medication from the inhaler and its accurate administration in lungs. It also logs down the date and time of administration in its memory and displays it in small LCD screen, thus helping the healthcare professional to track the patient deeply and help in the effective therapy of asthma (Lavorini et al. 2010). MDILog It is similar device like the smart mist, which is approved by the FDA and provides accuracy and reliability in delivering a drug through inhalation technique. The sensing component fitted to a device is a different form of smart mist. The device additionally includes an accelerometer and sensor to evaluate temperature and inhaler actuation. The accelerator and temperature sensor also records the date and time of inhaler shaking and actual inhalation, respectively. MDILog is a wireless system that is connected to computer easily, and it also possesses a small LCD screen and auditory tone output for interface (Penza-Clyve et al. 2004).

Pulse Sensor Device A pulse sensing element is a simple device that offers idea of functioning condition of a heart. Heart arrhythmias are monitored by a pulse sensor, which measures the pulse rate sensing by the flow of blood in veins, and the flow rate gives an idea of the heart’s working capacity. The basic sensor possesses an amalgamation of three basic pins, namely, ground, Vcc, and an input signal (which is also known as A0 signal). The pin indicates the heart pace and is present in a breadboard or in the printed circuit board (PCB) attached to a sensor. When it is attached to an Arduino or with the ESP8266 Wi-Fi module, the LED is in an active state which works either in 3v or 5v mode. Asada et al. (2002) introduced a ring sensor device fabricated

1032

K. Sharma et al.

to assess heart rate along with blood oxygen saturation concentration which is selfsufficient for its use. The device is worn in fingers, and a unified tactic used is artifact reduction in motion which obtains the accurate measurements (Asada et al. 2002). Yan and Zhang (2008) introduced a gadget that helped to record the arterial oxygen saturation. The application of this device is in accordance with an optical model that utilizes photon diffusion analysis. The results obtained prove that the novel algorithm is more robust and shows better execution comparing with the motion resistant algorithm distinct saturation transform (DST), hence efficient to use (Yan and Zhang 2008). The pulse sensor is also used to monitor the heart pulse through the neck arteries. Arduino nano R3, a microcontroller pulse sensor is fit into the helmet which detects the sleepy state and causes buzzing vibration in ear.

Temperature Sensing Device The temperature sensing devices measure the body temperature by utilizing sensors. The developed technology allows the measurement of temperature from any part of body. The nasal skin temperature reflects autonomic nerve activity and their measurements help to estimate diverse human psychological and physiological conditions. Wei et al. (2019) discussed a smart device that records the body temperature using the backside of body. A 3D printer is utilized for fabricating the framework which helps in the execution of different tests and in vivo experiments (Fig. 5). The error values in these devices were found to be less than 0.1 ◦ C from 25 to 40 ◦ C. Young males (10) were chosen to record their body temperatures using devices fitted to their back which helped establishing relationship with ear temperature. To assess the fitting of the device, information about the following parameters were evaluated: the sum of squares due to error (33.0874), R-square (0.0212), adjusted R-square (0.0117), and root mean square error (0.3998). The experimental results showed the difference of mean value of error obtained using the value of ear temperature and assessed core body temperature which was found to be ±0.19 ◦ C. The mean bias was also obtained to be 0.05 ± 0.14 ◦ C when the subjects were in a steady state (Wei et al. 2019). The advance wearable strain sensor in combination with thermal management is used for monitoring the motion and its management.

Application of Smart Sensor Smart Sensor for Gastrointestinal Tract Carcinoma in digestive tract such as stomach, duodenum, jejunum, ileum, cecum, colon, rectum, etc. is a complex diagnosis in patients who are around 50 years of age. Till now no such significant method for carcinoma identification at an early stage has been introduced and it is detected only when a person shows

40 Smart Devices in Healthcare Sector: Applications

1033

Fig. 5 (a) 3D model of the shell organization. (b) In vivo experiment for measuring body temperature from the back of the body (Wei et al. 2019)

some symptoms. Cancer grows and may become metastasized affecting diverse organs, which leads to low survival statistic of a person. The sensors help to measure and analyze the peristalsis movement which helps in diagnosing the abnormality in a user system. A wireless capsule arrangement embedded with a sensor analyzes real-time physiological parameters. They capture the image and communicate to processor which computes abnormal motility by comparing the images. The traditional system of endoscopy is time-consuming and inconvenient for the diseased. The capsule system helps in surmounting these disadvantages and also helps in providing a thorough examination of gastrointestinal tract (GIT). This gadget is provided with capacitive pH, capacitive pressure, resistive temperature with frequency and pulse width modulated IC circuit, microcontroller, and receiver that are implanted in a capsule arrangement. Integrated circuit (IC) helps in the conversion of physiological signals detected in GIT to electrical signals (Arefin et al. 2017). The flexible piezoelectric device is possibly employed for the study of GIT

1034

K. Sharma et al.

motility which measures the mechanical variations and energy and thus serves as a guide for optimal therapeutic intervention (Dagdeviren et al. 2019). For example, in an emergency where the upper GIT is turned out to be bleeding and needs immediate attention and treatment, an intracorporeal bleeding detecting sensor will provide real-time data and quick identification of acute bleeding at the upper part of GIT. The sensor that is placed within a capsule and is swallowed through GIT diagnoses the bleeding in the suspected patient. Telemetric communication helps to detect blood and tracks its activity inside the body. The sensor possesses a phototransistor which utilizes the optical property (light intensity) of blood to detect its presence and thus helps in diagnosing the bleeding in upper GIT on a timely basis. The sensor can also be used to detect the nonmalignant tissue in human intestine. The refractive index analogue through sensor helps to identify the tissues in the intestine (Sinha et al. 2020).

Smart Sensors for Liver Chronic liver ailment is a global issue that impairs the health as well as quality of life, identified as cirrhosis leading to ascites, gastroesophageal varices, jaundice, hepatic encephalopathy (HE) , etc. The dysfunction of the liver leads to augmentation levels of biomolecules (Bale et al. 2018). The smart sensor devices assist in quick identification of increased traces of biomolecules and enhance the life of a patient. The HE leads to a progressive increase in ammonia due to dysfunctioning of liver where it is transformed to urea or uric acid through the urea cycle. The increased level of ammonia reaches the blood and is exhaled out through the breath while blood reaches lungs for purification. The high-level of ammonia gas can be diagnosed by a sensor device equipped with CuBr thin film with high selectivity and sensitivity when exhaled out. This device tracks the liver’s operating status and helps to support a healthy life (Ishida et al. 2020). The identification of hepatocellular carcinoma (HCC) at the initial phases is beneficial for patient health. For this to be achieved, the procedure employed for diagnosis should be efficient enough. The HCC can be identified by existence of a lone polypeptide chain glycoprotein named α-fetoprotein. Nano optical sensors could be utilized as a detecting device for liver carcinoma which can assess alpha-fetoprotein in a serum sample of a liver patient. The nano binucleated Pt(abi)(bpy) emits a strong spectrum in water at a wavelength of 380 nm. End of the emission spectrum helps in the analysis of abnormality. The device provides satisfactory results in valuation of nano binuclear Pt(abi)(bpy) for the detection of liver carcinoma (Attia et al. 2018).

Smart Sensors for Lungs Congestive heart failure happens due to weak heart muscles and faulty valves that fail to pump blood to different organs. Such conditions result in pulmonary edema where fluid accumulates in the lung’s cavity. The doctors diagnose such

40 Smart Devices in Healthcare Sector: Applications

1035

conditions through the breathing pattern analyzed by various sounds produced during breathing. Magnetic resonance imaging and computerized tomography scan are employed in diagnosing such conditions. It is an arduous task visiting hospitals, and moreover the present equipment is costly and needs expertise to operate. Therefore, a sensor-based device was introduced which allows the patient to assess the breathing patterns themselves. It is a sound sensor with a good quality signal processed by 3D printers (Jain et al. 2018a). The sound signals received with the help of the sensor are recorded by a voice memo application linked to a smartphone whose data after analysis by using algorithms can be saved in the computer for further investigations (Jain et al. 2018b; Wong 2020). The smart sensor also diagnoses the initial lung condition by observing the body temperature and fingernail color which helps in quick detection of lung condition and helps in the immediate treatment if required. The illnesses like pneumonia and tuberculosis can be diagnosed by using temperature and color sensors designed utilizing the Naïve Bayes method (Maulana et al. 2020). The patients suffering from asthma, pneumonia, chronic obstructive pulmonary disease (COPD), and sleep apnea need constant surveillance of respiration rate. The most commonly used method for this purpose is ECG signals, Impulse Radio (IR) UWB (Ultra-Wide Band) radar systems, etc. But these equipment are costly and need a string signal production and identification system operated by expertise. A new signaling sensor-based device is introduced which acquires the signals wirelessly by an indicator through Bluetooth technology employed in a hybrid-spiral antenna in a multi-material fiber of a cotton T-shirt. The variation in volume of air in lungs changes the respiratory pattern which could be identified mechanically by the antenna’s configuration as the chest changes the dielectric property of the antenna. Smartphones may be coupled with a Bluetooth device that aids in receiving the respiratory signals and helps in catching any dysfunction of the lungs. Figure 6 depicts a T-shirt fitted with a sensor device implanted in the fabric (Gorgutsa et al. 2017).

Smart Sensor for Eyes Surgery and Its Management The unstable intraocular pressure leads to injury of optic nerves causing glaucoma detected at the later stage leading to blindness. The advancement in technology has introduced a micro- machine capacitive pressure sensor for assessing the intraocular pressure. The sensor is operatively planted on various areas of the eyes like cornea and sclera but the anterior region present in the eyes has additional advantages. There are various techniques introduced for quantifying the external pressure in the eyes such as wired sensing technique, inductively coupled telemetry, and intraocular pressure (IOP) sensor. The intraocular pressure sensor continuously monitors alteration in pressure in eyes by passive LC resonance technique, where resonance frequency measured by sensor circuit finds a way to sensor capacitance which relates to external pressure in eyes (Katuri et al. 2010). Lloyd et al. described a pressure detecting gadget implanted into the aqueous chamber of the eyes for

1036

K. Sharma et al.

Fig. 6 T-shirt fitted with a sensor device implanted in the fabric (Gorgutsa et al. 2017)

measuring IOP. The signal is produced via a device which is attached to a wireless system. The device acquires, processes, and provisionally stocks the information in the database which is received through sensor and keeps recorded IOP (Lloyd et al. 2005). The wave sensor is gaining significant interest and advances in assessing and checking the aberrations in eyes. It is very useful in surgeries like cataract, conductive keratoplasty, lasik, and corneal correction. An integrated amalgamation of a wavefront sensor and microscope allows the doctor to assess the wavefront of the patient. This device supports and maintains the balance of integrated device during surgery by repositioning it and assessing the optical property of the eyes. A soft contact lens smart sensor is another device used for checking the tear glucose. The sensor comprises a microelectromechanical system where sensor features a film electrode on the contact lens and in the sensing region of the electrode glucose oxidase is immobilized. The sensor can be held for 50 days and shows efficient results (Kudo et al. 2012).

Smart Sensors for Brain Disease Smart sensors are outstanding diagnostic tools that have uses in diagnosing brain diseases. By now, several technologies have been introduced like smart sensors, digital devices, and intelligent applications that assist people of each generation to monitor their health. An approach for an integrated technological solution would fix issues faced by elderly patients and would lead to reinventing their quality of life, health, and safety (Uddin et al. 2018). The sensor is used for different types of brain disorders like brain tumors, Alzheimer’s disease (AD), Parkinson’s disease, dementia, etc. which develop gradually, and thus, it is arduous to recognize pathological processes in reference to clinical phenotype alone. In a study, AlZubi

40 Smart Devices in Healthcare Sector: Applications

1037

Fig. 7 (a) Wearable device-based information gathering procedure. (b) Parkinson disease diagnosis functioning procedure framework. [Reproduced with permission from (AlZubi et al. 2020), 4911870189010]

et al. (2020) discussed the heuristic tube optimized sequence modular neural network (HTSMNN)-based Parkinson disease recognition process. Herein, a deep brain stimulation smart IoT sensor gadget is positioned on the patient to collect the brain features (Fig. 7a and b). The data gained is then processed by HTSMNN. The method inquired the information constantly and freely to anticipate alteration going on in the brain. The method identified the alterations in brain functioning swiftly,

1038

K. Sharma et al.

lowering the prediction delay and in turn enhancing Parkinson’s disease seriousness (AlZubi et al. 2020). EEG is a commonly used device for recording the brain activity and diseases such as Alzheimer’s, stroke, epilepsy, depression, and brain injuries. But it requires expert opinion for diagnosis which is time consuming. EEG-based diagnosis integrated with smart sensor device interconnected with cloud technology helps in smart diagnosis of any brain activity. EEG sensors help to acquire the signals and transmit them to cloud server for processing. These multimodal sensors help in sensing the difference in gesture, movement, EEG, facial expression, and voice which determines patient health. The signals are transmitted by sensors to processing machine where a pathological detection is performed and immediate response is decided. The experts analyze the response and take the immediate treatment required for the patient (Amin et al. 2019). In older adults, mobility metrics are obtained from sensor recordings which are related to parkinsonism. von et al. (2019) inspected the productivity of theses metrics to envisage incident parkinsonism. The patient fastens the sensor on their back and is allowed to walk 32 foot, standing posture, and Timed Up and Go (TUG) tasks. Six out of 12 sensorbased measures are related to incident parkinsonism (von et al. 2019). To evaluate AD-related physiological alterations Saif et al. (2019) reported the feasibility of wearable biosensor (WHOOP). Forty patients with AD and no or nominal cognitive grievances were selected and out of them 34 convinced to wear a WHOOP. One patient lost his device. Twenty-four were grouped as usual cognition and were asymptomatic, six were grouped as subjective cognitive downward, and other three were amyloid-positive. Sleep-cycle, heart cardiac rhythm alterations, and activism measures were gathered via WHOOP. Twenty-seven attendees finished the surveys. Six months later, 24 attendees were contented with WHOOP, and 23 intended to continue WHOOP. The outcome showed that physiological data obtained via the device may reveal itself as a reliable technique to monitor cognitive alterations associated with preclinical AD (Saif et al. 2019).

Cardiac and Motor Activity Monitor Almost 17.7 million (31%) deaths have been because of cardiovascular disorders (CVDs) in 2015, as reported. Therefrom 7.4 million have been a consequence of coronary heart disease, and almost 7 million because of stroke. Early and quick diagnosis is important for the successful prognosis of these diseases. Many devices are in use for monitoring continuous cardiac and motor activity. Some examples include OLAM chest band, chest and headband, Doppler radar sensor lodged in a manual wheelchair, EPIC electric field sensors, Smart Helmet, Joule Earing, and cosinuss◦ (Schmidt 2020). Oregon State University Life and Activity Monitor (OLAM) is a gadget that offers the potentiality of combining motion figures with heart-rate records, thereby permitting assessment of tangible physical activity. It involves the employment of a

40 Smart Devices in Healthcare Sector: Applications

1039

heart sensor with a 5-axial inertial valuation section which allows concurrent heart, respiration, and movement surveillance and does not necessitate direct placement on skin. It consists of a simple, repeatable, robust capacitive electrode. Amplifier circuits to buffer the weak signal acquired electrode are housed on the top. In the center, a snap connector is employed to provide mechanical stability. The plate that is present at the bottom is a solid copper fill, which in turn represents a parallel plate capacitor along with the body to harness biopotential signals. However, in this device, the copper fill is not segregated by the solder mask which enables it to alternatively work as a dry contact electrode. This sensor permits the gadget to be fully enveloped inside a cloth material without altering the sensing capabilities. This gadget is fastened on body with a thin chest band giving it a smooth connection along with the freedom to stretch the user’s movements (Albright et al. 2011). The same gadget could also be accustomed as a headband, and would continuously monitor ECG along with EEG. A cardiac respiratory monitor for patients dependent on wheelchair was introduced by Postolache and coworkers in 2011. This involved frequency modulated continuous wave (FMCW). Doppler radar sensors incorporated in a wheelchair are employed for the assessment of cardiac, respiratory, and physical activity of the user. Another radar sensor used to assess motor action via distance travelled by wheelchair, while the user carries out the manual workings of wheelchair. The conditioning circuit comprises of active filters together with a microcontroller supported primary processing component that is configured and put into practice to convey the information through Bluetooth transmission procedure to an Android OS tablet computer. Signal treating of Doppler radar assessment channel signals, graphical user link, data storage, and Wi-Fi data coordination with remote physiological and physical activity database are the major abilities of the computer program which is created utilizing Android SDK and Java (Postolache et al. 2011). EPIC (electric potential integrated circuit ) electric sensors are demonstrated to be helpful in detecting of physio-behavioral parameters in rodents (Noble et al. 2017). These gadgets provide a cost-effective, noncontact detection by placing the sensor outside the rodent cages. Plessey semiconductors are an order of ultra-high resistance and dry-contact capacitive coupling electric field sensors. These high sensitivity sensors are advertised as capable of use for contact or noncontact-based identification of proximity to the sensor, movement or specific gestures, and ECG activity in humans. Each EPIC Plessey PS25251 sensor is 1 cm2 in size with four pins: Vdd (the positive power supply = +5V), Vss (the negative power supply = – 5V), Gnd (ground), and output. Smart Helmet is a helmet that has to be fastened on the head and is incorporated with sensors that record/monitor EEG (Von Rosenberg et al. 2016). It also constitutes a belt that is supposed to be fastened on the chest which helps in recording ECG. The sensor for recording ECG is the comparable dry contact electrode. Similarly, Joule Earing is the device that records cardiac pulse, calorie loss, and activity level. This has to be employed as a piece of jewelry – earring. However, the drawback that might be faced is that some people probably do not find it comfortable wearing as an earring (Joule Earring Backings 2020).

1040

K. Sharma et al.

Smart Sensors for Parkinson’s Diseases Millions of neuronal deaths in particular portions of brain cause Parkinson’s disease. Symptoms include memory loss, forgetfulness, apathy, anxiety, agitation, akathisia, mood changes, and other symptoms. Consequently, pharmacological therapy merely provides symptomatic relief and fails to prevent the advancement of the disease. Therapeutic supervision of cases has always remained a difficult task majorly due to dearth of instruments to properly measure therapeutic responses of the pharmacological intervention and also because of the insufficiency of data of motor responses displayed by the individual in their daily routine. Patients are often asked to keep a diary/journal of their motor symptoms when in-home, and this, however, is often tiring, not accurate, and the patient tends to forget to note down the symptoms (Ossig et al. 2016). To overcome this issue, implementation of these devices has been introduced. Various wearable devices like the Ambulsono sensor system, MercuryLive, Wearable multi-sensor motor unit, finger-based sensor system, and eGaIT are employed in the control of PD (Bhidayasiri et al. 2014). Supervision of specific symptoms by various devices is discussed in Table 1. These sensors most commonly utilize either accelerometer, gyroscope, or both. Gyroscopes are somewhat advantageous over accelerometer when it comes to measuring rotational motion because its reading is not affected by gravity, and therefore an amalgamation is used (Wong et al. 2007). Wearable engineering has offered a way to diagnose the disposition and intensity of motor and cognitive deterioration, which acts as assistive devices for surmounting these impairments that produce hindrance in bodily activities and enhances the capabilities of exercise by keeping a track of training progression, the optimized exercise “dosing,” and by consolidation of exercise with varied mental and mood regulation such as through motivational music stimulation (Jain et al. 2019). The Ambulsono sensor system (Chomiak et al. 2015) is the engineering that is harnessed in assessing various parameters in PD patients. Ambulsono is used on the wrist, thigh, and knee. This system utilizes Apple Operating System (iOS) and runs off on Gait Reminder App that produces auditory instructions and records step size using iOS gyro and accelerometer. This device is specifically utilized in the 6-minute walking test (6MWKT) wherein the patients are instructed to walk for 6 minutes in a marked space. Although this test has high acceptance, the outcome, without using this wearable device, was affected by various parameters, not to mention the time consumed in performing the test. The employment of the Ambulsono sensor has helped in standardizing and automating this 6MWKT. The same sensor-based device is involved in another diagnostic test in PD sufferers, namely, the stepping dual-task test. In this test, the individual is asked to step in place while performing another cognitive activity. Height of step is the outcome measure. Ambulsono sensors placed in the ears dispense the instructions for the cognitive activity, whereas the second sensor which is fastened on the knees records height of steps. Another sensor used is called MercuryLive. Longitudinal tracking of motor manifestations in a private setting, which earlier was not possible, can now be done easily by using of MercuryLive. This system works at 3 ranks – a sensor established information

40 Smart Devices in Healthcare Sector: Applications

1041

Table 1 Smart devices used based on symptoms of PD Symptoms Tremor

Bradykinesia

Gait Dyskinesia Motor function

Devices Wearable accelerometers/gyrostats, DigiTrac (Doudet et al. 2004), ActiTrac (Ruiz and Bernardos 2008),SomnoMedics GmbH, Randersacker (Ossig et al. 2016) MercuryLive (Patel et al. 2011), Wearable multi-sensor motor unit (WMSMU) (Tzallas et al. 2014), a finger-based sensor (Bhidayasiri et al. 2014) Several applications are present for iOS and Android mobile phones, in accordance to accelerometry, and some apps are however not regulated and are therefore not approved. The same accelerometer/gyrostat sensors are employed in monitoring/assessing Gait. eGAIT, GmbH, Erlangen, Germany (Klucken et al. 2013), MercuryLive (Patel et al. 2011), Wearable multi-sensor motor unit (WMSMU) (Tzallas et al. 2014) eGAIT, Astrum IT GmbH, Erlangen, Germany (Klucken et al. 2013), Ambulsono sensor system (Chomiak et al. 2015) Watch type accelerometers, like PKG, Global Kinetics Corporation(Griffiths et al. 2012) The accelerometer fastened to variegated body parts (Patel et al. 2011) Watch-like accelerometers, for example, PKG (Griffiths et al. 2012), a finger-based sensor (Bhidayasiri et al. 2014)

assembling system, an Internet gateway for live streaming and storage of data, and a user interface for two-way connection among the user and doctor. Another research involved similar model. A wearable multi-sensor motor unit (WMSMU) is usually physically fastened on body to monitor daily motor activity. It comprises of four triaxial accelerometers for every extremity along with one accelerometer/gyroscope to be fastened on waistline, which would record body movement acceleration along with angular velocity while turning. All this recorded signal would be received by a Parkinson Daily Data Set Logger (data acquisition unit), which may therefore be shared with clinical organization for the betterment of pharmacotherapy (Tzallas et al. 2014). Finger-based sensor used by Bhidayasiri et al. (2014) involved a sensor section attached to a thin wire onto a transmitter which has to be fastened on the wrist. This sensor is utilized in detection of tremors suffered by PD patients. The sensor involves the use of a 3-axis accelerometer that can evaluate acceleration in three translational planes and a 3-axis gyroscope for measuring three angular oscillations about axes in these planes. For this reason, the whole unit has the capability of capturing motion with 6 points of freedom. Both sensors are automated to collect information at the gap of 4 milliseconds and deliver 10-bit resolution digital output for the accelerometer and 16-bit for the gyroscope, transmitting data wirelessly as far as 10 meters to a data processor in real-time exploiting a 2.4 GHz Bluetooth TM radio, or saving the information in a microSD card. This information is then

1042

K. Sharma et al.

accessed by the physician to modulate the pharmacotherapy as per starting of the symptoms (Bhidayasiri et al. 2014). Embedded gait analysis using intelligent technology (eGaIT) is the sensor-based wearable device that has been employed to review motor-related gait impairment as per the Hoehn&Yahr (H&Y) staging. It is identified by shuffling gait, decreased length of step, inflicted gait commencement, and reduction in gait speed. This gadget is not directly fastened to the user’s body but on their shoes. It employs harnessing the similar accelerometer and gyroscope for assessment of sensor signal with a Bluetooth device for transmittance of data (Klucken et al. 2013). Although the employment of smart gadgets has been quickly adopted in clinical setting, the acceptability of physicians or the primary healthcare givers is substantially more as observed with PD patients.

Devices in Management of Metabolic Disorder A metabolic disorder is an involved condition noted by insulin resistance, obesity, dyslipidemia, hypertension, and hyperglycemia. It is classified among the risk elements for other conditions like cardiovascular disorders, noninsulin-dependent diabetes mellitus, stroke, chronic kidney disease, and cancers. Managing glucose levels is very important in preventing many micro- and macrovascular diseases like cardiovascular disease, nephropathy, and retinopathy along with the slow recovery of the wound. The crucial endeavor of continuous supervision of patients specifically those patients who are at higher exposure of being afflicted with metabolic syndrome, also gathering real-time information and indication from patients and other individuals, and preventive education to manage the disease, is still hard and expensive to tackle specifically with the traditional technologies (Moradi et al. 2018). Earlier, the method for checking the concentration of glucose in blood was by usually utilizing a needle to prick the finger, which is generally known as lancet, to capture capillary blood in glucose meter, this method however provided accurate results but users were often reluctant by virtue of pain experienced, disposal of the lancet, unable to detect sudden dip or spike of glucose amount and blood wastage (Gao et al. 2018). Nonetheless, this traditional method was unable to deliver uninterrupted glucose level supervision. Advancement in the scope of engineering along with the inception of Internet of Things (IoT) paved the way for better management of diabetic blood glucose levels. These advanced sensors are employed as a continuous glucose monitoring system (CGMS). Variegated gadgets are obtainable for the supervision of metabolic disorders, ranging from simple sensor devices to collaborative technologies that promise comprehensive stewardship of metabolic disorders. Devices like GlucoWatch, Wearable Graphene-Gold Sweat Patch, Metabolic Syndrome Control and caring system (MetaSyCar), SugarBEAT® (SugarBEAT, Daily Non-Invasive CGM 2020), 3-μ W CMOS Glucose Sensor, glucose detecting contact lense, optical smart sensors, automated microneedle drug delivery system, and fluorescent polymeric nanosensors/smart tattoos, etc. are used (Jain et al. 2013a, b; Jain and Jain 2015;

40 Smart Devices in Healthcare Sector: Applications

1043

Jain and Jain 2016; Jain and Jain 2018; Moradi et al. 2018; Prajapati et al. 2019). Lee et al. (2016) amalgamated gold mesh with graphene to enhance its electrochemical sensitivity and used it as a wearable sweat patch to monitor sweatbased diabetes. This device shows stretchable property and consists of serpentine dual layer of gold mesh and graphene that gives a competent electrochemical link for firm transmittance of electrical signal. The patch perspiration check part, sensing part, and therapeutic part are employed for delivering drugs transcutaneously. By thermally activating this patch, we can deliver metformin in animals suffering from diabetes (to lower glucose concentration in blood) (Lee et al. 2016). The electrochemical glucose sensor uses amperometric sensors that are linked with an electrochemical enzymatic approach to constantly quantify the amount of glucose which is present in interstitial fluid of subcutaneous tissue. Oxidation of glucose, which is done through glucose oxidase, quantifies the creation of electrons; the amount of current generated commensurate to the concentration of glucose present in interstitial fluid. These systems are in reference with the volume of glucose in ISF and relate to the abundance of glucose present in plasma. In this arrangement, a very thin needle is introduced in the cellular tissue of the abdomen to quantify the measure of glucose in ISF or via employment of an external sensor. The signal that is generated is gathered and treated in an extraneous screen (Torres et al. 2010). There are many devices that are supposed to be introduced in the body, whereas GlucoDay® has a sensor that can be positioned extraneously. Generally, these devices consist of: (a) A glucose sensor, which is used for diagnosis purpose (b) An insulin delivery system (c) A feedback mechanism that bridges between glucose detection and medication delivery units Figure 8 depicts the various components of an archetype microneedle-supported device for the supervision of diabetes, and it is important as it decreases the chances of hypo- or hyperglycemia in patients suffering from diabetes. In this system, glucose concentration is monitored at frequent intermissions by glucose sensor and also the required dose of insulin is injected through an augmented pump (Khanna et al. 2008). The measure of glucose present can be sensed topically via the skin through electro-osmotic flow (reverse iontophoresis), by applying a low degree of electrical current. This process of iontophoretic extraction has been amalgamated within a gadget that involves in situ glucose sensor called GlucoWatch (Potts et al. 2002). The sensing electrode is working in a layer of Pt/C ink, while the sensor and electrodes are composed of Ag and Ag/AgCl coating. The counter electrode also serves as the iontophoretic one. The hydrogel discs placed over each sensor are in connection with the skin and serve as the smart sensor electrolyte and also as the repositories wherein glucose is gathered. The biosensor chemistry utilizes direct detection of H2 O2 produced by the reaction of glucose with glucose oxidase (Potts et al. 2002). It also needs calibration via a single fingerstick test. SugarBEAT® ,

1044

K. Sharma et al.

Fig. 8 Elements of a prototypical microneedle-based diabetes therapy system

(https://sugarbeat.com/) like other devices, utilizes the reverse iontophoresis process for gaining glucose from interstitial fluid. It has an adhesive patch sort of design which gives the potentiality of daily disposal. Liao et al. (2011) introduced a noninvasive wireless sensor manifesto for continuous health monitoring. This device unifies a loop antenna, wireless sensor link, with a glucose sensor. This device has a reach of 0.18 μ A mm–2 mM. The arrangement consumes current of 3 μW from 1.2 V supply as well as quantifies scope of 0.05–1 mM. The patent has been granted to Google for a similar, disposable noninvasive sensor-based contact lens that can discern bioanalytes present in tears (Liao et al. 2011). The system consists of: (i) Contact lens, which will function as a measurement unit (ii) Reader tool, which will transmit the power to the contact lens (iii) A display device that would transmit the data This patent explicitly discusses the probability of spectacles, jewelry, or clothing operating as the reader. The principle requisite is that gadget must be sufficiently close to either of the lenses, thereby guaranteeing transmission to the reader. Likewise, mobile phones or computers can function as a display tool.

Challenges Various psychometric parameters of assessments are crucial in clinical scenarios of smart sensor estimation, comprising of precision and robustness, dependability and uniformity, clinical utilization, sensitization to alterations, quality of being generalizable, and economically feasibility. Several challenges are faced while using

40 Smart Devices in Healthcare Sector: Applications

1045

sensors. Calibration and validation is an issue that is frequently encountered. To calibrate the instrument, there occurs a requirement for the skilled individuals that should be fulfilled at each occurrence, as a non-calibrated instrument might give wrong readings. For instance, the problem with the majority of systems that measure glucose is that they require frequent calibration utilizing traditional method of finger pricking/capillary testing. Because of this, these sensor-based technologies act as only supplementary facilities along with capillary quantification of glycemia. Another challenge that is observed is commercial prices of the device. Some devices may weigh a higher monetary burden on individual patients. However, in clinical organizations or alternatively in a collaborative healthcare activity this cost might be distributed. Nonetheless, these costs would be still lesser on comparing with the amount the patient would have to spend for the hospital visit, getting himself tested with traditional and heavy machinery. Sensors with adhesive electrodes are irritating and uncomfortable to be home used, ultimately causing low compliance. To assess the cardiac health-related problems, variegated clinical practices like nuclear myocardial perfusion scan Echocardiogram (Echo), Electrocardiogram (ECG), computerized tomography scan (CT), etc are present. However, most clinical practices are either highly expensive and require special devices are not practicable to be designed using body sensors. ECG can be detected through body sensors; however, it can only detect the cardiac electrical activities, which offer little knowledge on various cardiac mechanical operations like the activity of heart valves, blood circulation into ventricles, and suppression-relaxation of ventricle walls, etc. Sensors for specifically monitoring cardiac activity consist of capacitive type noncontact electrodes. Several issues are encountered in the engineering of these sensors, for instance, the capacitive electrode shows very high circuit noise floor, it shows much greater sensitivity to 50/ 60Hz line noise, and even slight amounts of motion/friction saturates the signal (Koydemir and Ozcan 2018). It has been recorded that information gathered by most wearable devices that are available for commercial use does not fit the performance scrutiny because of deficiency in dependability and hence their employment are circumscribed to the introductory detection techniques like surveilling heart rate as well as blood pressure. Due to insufficiency of acquiring location-specific information limited accuracy is observed in wearable devices. Additionally, some sensors that deliver unrefined data become too complicated to elucidate. Wearable devices are either supposed to be fastened on wrists or chest functions by utilizing motion detectors but it depends upon the users’ capability and willingness to wear them.

Conclusion and Future Perspectives The advent of Industry 4.0 and progressions in the field of IoT has ignited the ambits of medical fields to a new level. These gadgets offer fine possibilities in diagnosis, management, and therapy of variegated diseases and conditions by providing continuous monitoring and real-time data. There are various marketing tools that are accessible in the stores for commercial use and some are yet in

1046

K. Sharma et al.

the labs perfecting upon the designs. Studies that will be planned in future need to focus on larger sample sizes, different smart sensor devices, and more robust analytical methods to expand upon these preliminary findings and can assist in explaining the competencies of the wearable biometric device. Therefore, it is warranted from future studies to work on sensor positions and angle adjustments to improve lightness and accuracy. Utilization of infrared thermography for risk assessment while comparing experiments will be necessary for the employment of the structure that does not impede the view.

Important websites • Joule Earring Backings. (2020). http://shopjoule.com/. Accessed 23-09-2020 2020 • https://www.abilifymycite.com/about • https://www.cognitionkit.com/ • https://www.maximintegrated.com/en/products/sensors/healthcare-sensor-ics. html • https://www.peerbridgehealth.com/ • https://www.prescouter.com/2019/12/future-healthcare-smart-sensors-digitalbiomarkers/ • https://www.apple.com/healthcare/apple-watch/ • https://www.wipro.com/en-IN/business-process/what-can-iot-do-for-healthcare-/ Acknowledgment Dr. AJ gratefully acknowledges the financial support as C.V. Raman postdoctoral fellowship under the Institution of Eminence (IoE) scheme at the Indian Institute of Science, Bengaluru (Karnataka), India.

References Albright RK, Goska BJ, Hagen TM, Chi MY, Cauwenberghs G, Chiang PYOLAM (2011) A wearable, non-contact sensor for continuous heart-rate and activity monitoring. In: 2011 annual international conference of the IEEE engineering in medicine and biology society. IEEE, pp 5625–5628 AlZubi AA, Alarifi A, Al-Maitah M (2020) Deep brain simulation wearable IoT sensor device based Parkinson Brain disorder detection using heuristic tubu optimized sequence modular neural network. Measurement 61:107887 Amin SU, Hossain MS, Muhammad G, Alhussein M, Rahman MA (2019) Cognitive smart healthcare for pathology detection and monitoring. IEEE Access 7:10745–10753 An BW et al (2017) Smart sensor systems for wearable electronic devices. Polymers 9:303 Arefin MS, Redouté J-M, Yuce MR (2017) Integration of low-power ASIC and MEMS sensors for monitoring gastrointestinal tract using a wireless capsule system. IEEE J Biomed Health Inf 22:87–97 Asada H et al (2002) Significant antitumor effects obtained by autologous tumor cell vaccine engineered to secrete interleukin (IL)-12 and IL-18 by means of the EBV/lipoplex. Mol Ther 5:609–616. https://doi.org/10.1006/mthe.2002.0587

40 Smart Devices in Healthcare Sector: Applications

1047

Attia M, Youssef A, Khan ZA, Abou-Omar M (2018) Alpha fetoprotein assessment by using a nano optical sensor thin film binuclear Pt-2-aminobenzimidazole-Bipyridine for early diagnosis of liver cancer. Talanta 186:36–43 Bale A, Pai CG, Shetty S, Balaraju G, Shetty A (2018) Prevalence of and factors associated with minimal hepatic encephalopathy in patients with cirrhosis of liver. J Clin Exp Hepatol 8:156–161 Barrett J, Liebman C (2020) Gaining competence in musculoskeletal care as a primary care provider. J Nurse Practitioners 16:44–47 Bhidayasiri R, Petchrutchatachart S, Pongthornseri R, Anan C, Dumnin S, Thanawattano C (2014) Low-cost, 3-dimension, office-based inertial sensors for automated tremor assessment: technical development and experimental verification. J Parkinsons Dis 4:273–282 Bolotsky A et al (2019) Two-dimensional materials in biosensing and healthcare: from in vitro diagnostics to optogenetics and beyond. ACS Nano 13:9781–9810 Choi J, Kim J, Hwang JY, Je M, Kim J-Y, Kim S-Y (2017) A novel smart navigation system for intramedullary nailing in orthopedic surgery. PLoS One 12:e0174407 Chomiak T et al (2015) A new quantitative method for evaluating freezing of gait and dualattention task deficits in Parkinson’s disease. J Neural Transm 122:1523–1531 Dagdeviren C, Traverso CG, Langer RS (2019) Flexible piezoelectric devices for gastrointestinal motility sensing. Google Patents de Fátima Domingues M et al (2020) Wearable devices for remote physical rehabilitation using a fabry-perot optical fiber sensor: ankle joint kinematic. IEEE Access 8:109866–109875 Doudet D, Cornfeldt M, Honey C, Schweikert A, Allen R (2004) PET imaging of implanted human retinal pigment epithelial cells in the MPTP-induced primate model of Parkinson’s disease. Exp Neurol 189:361–368 Gao W, Brooks GA, Klonoff DC (2018) Wearable physiological systems and technologies for metabolic monitoring. J Appl Physiol 124:548–556 Godoi BB, Amorim GD, Quiroga DG, Holanda VM, Júlio T, Tournier MB (2019) Parkinson’s disease and wearable devices, new perspectives for a public health issue: an integrative literature review. Rev Assoc Med Bras 65:1413–1420 Gorgutsa S, Bellemare-Rousseau S, Guay P, Miled A, Messaddeq Y (2017) Smart T-shirt with wireless respiration sensor. In: 2017 IEEE SENSORS. IEEE, pp 1–3 Griffiths RI et al (2012) Automated assessment of bradykinesia and dyskinesia in Parkinson’s disease. J Parkinsons Dis 2:47–55 Ishida J et al (2020) Real-time breath ammonia measurement using a novel cuprous bromide sensor device in patients with chronic liver disease: a feasibility and pilot study. J Breath Res Jain A, Jain SK (2015) Environmentally rzesponsive Chitosan-based Nanocarriers (CBNs) handbook of polymers for pharmaceutical technologies. Biodegradable Polymers 3:105 Jain A, Jain SK (2016) In vitro release kinetics model fitting of Liposomes: an insight. Chem Phys Lipids 201:28–40. https://doi.org/10.1016/j.chemphyslip.2016.10.005 Jain A, Jain SK (2018) Stimuli-responsive Smart Liposomes in cancer targeting. Curr Drug Targets 19:259–270. https://doi.org/10.2174/1389450117666160208144143 Jain A, Gulbake A, Jain A, Shilpi S, Hurkat P, Jain SK (2013a) Dual drug delivery using “smart” liposomes for triggered release of anticancer agents. J Nanopart Res 15:1772 Jain A, Gulbake A, Shilpi S, Jain A, Hurkat P, Jain SK (2013b) A new horizon in modifications of Chitosan: syntheses and applications. Crit Rev Ther Drug Carrier Syst 30:91–181. https:// doi.org/10.1615/critrevtherdrugcarriersyst.2013005678 Jain A, Bansal K, Tiwari A, Rosling A, Rosenholm J (2018a) Role of polymers in 3D printing technology for drug delivery - an Overview. Curr Pharm Des 25 Jain A, Tiwari A, Verma A, Jain SK (2018b) Ultrasound-based triggered drug delivery to tumors. Drug Deliv Trans Res 8:150–164. https://doi.org/10.1007/s13346-017-0448-6 Jain A, Hurkat P, Jain SK (2019) Development of liposomes using formulation by design: Basics to recent advances. Chem Phys Lipids 224:104764. https://doi.org/10.1016/ j.chemphyslip.2019.03.017

1048

K. Sharma et al.

Jin L et al (2017) Self-powered wireless smart sensor based on maglev porous nanogenerator for train monitoring system. Nano Energy 38:185–192 Joule Earring Backings (2020). http://shopjoule.com/. Accessed 23 Sept 2020 Karipott SS, Nelson BD, Guldberg RE, Ong KG (2018) Clinical potential of implantable wireless sensors for orthopedic treatments. Expert Rev Med Devices 15:255–264 Katuri KC, Ramasubramanian MK, Asrani SA (2010) Surface micromachined capacitive pressure sensor for intraocular pressure measurement. In: Proceedings of 2010 IEEE/ASME international conference on mechatronic and embedded systems and applications. IEEE, pp 149–154 Khanna P, Strom JA, Malone JI, Bhansali S (2008) Microneedle-based automated therapy for diabetes mellitus. J Diabetes Sci Technol 2:1122–1129 Kim J et al (2015) Highly transparent and stretchable field-effect transistor sensors using graphene– nanowire hybrid nanostructures. Adv Mater 27:3292–3297 Klucken J et al (2013) Unbiased and mobile gait analysis detects motor impairment in Parkinson’s disease. PLoS One 8:e56956 Koydemir HC, Ozcan A (2018) Wearable and implantable sensors for biomedical applications. Ann Rev Anal Chem 11:127–146 Kudo H, Chu M, Miyajima K, Arakawa T, Mitsubayashi K, Hiranuma Y, Hiramatsu H (2012) Soft contact-lens biosensor for real-time tear sugar monitoring at the eye. In: 2012 IEEE international conference on Systems, Man, and Cybernetics (SMC). IEEE, pp 2048–2051 Lavorini F, Levy ML, Corrigan C, Crompton G (2010) The ADMIT series-issues in inhalation therapy. 6) Training tools for inhalation devices. Primary Care Respir J 19:335–341 Leal-Junior AG, Diaz CA, Avellar LM, Pontes MJ, Marques C, Frizera A (2019) Polymer optical fiber sensors in healthcare applications: a comprehensive review. Sensors 19:3156 Lee H et al (2016) A graphene-based electrochemical device with thermoresponsive microneedles for diabetes monitoring and therapy. Nat Nanotechnol 11:566–572 Liao Y-T, Yao H, Lingley A, Parviz B, Otis BP (2011) A 3-$\mu\hbox {W} $ CMOS glucose sensor for wireless contact-lens tear glucose monitoring. IEEE J Solid-State Circ 47:335–344 Lloyd JR, Grotjohn TA, Weber AJ, Rosenbaum FR, Goodall GA (2005) Implantable microscale pressure sensor system for pressure monitoring and management. Google Patents Lu R et al (2020) Evaluation of wearable sensor devices in Parkinson’s disease: a review of current status and future prospects. Parkinson’s Dis:2020 Maulana R, Kurniawan D, Ichsan MHH (2020) Smart devices for self-diagnosing of lung condition based on body temperature and fingernail color. In: 2020 2nd World Symposium on Artificial Intelligence (WSAI). IEEE, pp 113–117 Moradi M, Moradi M, Anvari-Moghaddam A (2018) MetaSyCar: a system for metabolic syndrome control and caring. In: 2018 IEEE international conference on internet of things (iThings) and IEEE green computing and communications (GreenCom) and IEEE Cyber, Physical and Social Computing (CPSCom) and IEEE Smart Data (SmartData). IEEE, pp 183–188 Morales-Velazquez L, de Jesus R-TR, Herrera-Ruiz G, Morinigo-Sotelo D, Osornio-Rios RA (2017) Smart sensor network for power quality monitoring in electrical installations. Measurement 103:133–142 Noble DJ, MacDowell CJ, McKinnon ML, Neblett TI, Goolsby WN, Hochman S (2017) Use of electric field sensors for recording respiration, heart rate, and stereotyped motor behaviors in the rodent home cage. J Neurosci Methods 277:88–100 O’Hara SP, Tabibian JH, Splinter PL, LaRusso NF (2013) The dynamic biliary epithelia: molecules, pathways, and disease. J Hepatol 58:575–582 Ossig C et al (2016) Wearable sensor-based objective assessment of motor symptoms in Parkinson’s disease. J Neural Transm 123:57–64 Patel S et al (2011) Longitudinal monitoring of patients with Parkinson’s disease via wearable sensor technology in the home setting. In: 2011 annual international conference of the IEEE engineering in medicine and biology society. IEEE, pp 1552–1555 Pausley ME, Seelecke S (2008) Multifunctional SMA-based smart inhaler system for improved aerosol drug delivery: design and fabrication. In: Active and passive smart structures and integrated systems 2008. International Society for Optics and Photonics, p 69280T

40 Smart Devices in Healthcare Sector: Applications

1049

Penza-Clyve SM, Mansell C, McQuaid EL (2004) Why don’t children take their asthma medications? A qualitative analysis of children’s perspectives on adherence. J Asthma 41:189–197 Pichetsurnthorn P, Vattipalli K, Prasad S (2012) Nanoporous impedemetric biosensor for detection of trace atrazine from water samples. Biosens Bioelectron 32:155–162 Postolache O, Girão PS, Postolache G, Gabriel J (2011) Cardio-respiratory and daily activity monitor based on FMCW Doppler radar embedded in a wheelchair. In: 2011 annual international conference of the IEEE engineering in medicine and biology society. IEEE, pp 1917–1920 Potts RO, Tamada JP, Tierney MJ (2002) Glucose monitoring by reverse iontophoresis. Diabetes Metab Res Rev 18:S49–S53 Prajapati SK, Jain A, Jain A, Jain S (2019) Biodegradable polymers and constructs: a novel approach in drug delivery. Eur Polym J 120:109191. https://doi.org/10.1016/ j.eurpolymj.2019.08.018 Rocci ML Jr, Rossiter DJ, Isensee RK, Ferrara M, Salerno AH (2000) Dosage counter for metered dose inhaler (MDI) systems using a miniature pressure sensor. Google Patents Ruiz PJG, Bernardos VS (2008) Evaluation of ActiTrac® (ambulatory activity monitor) in Parkinson’s Disease. J Neurol Sci 270:67–69 Saif N et al (2019) Feasibility of using a wearable biosensor device in patients at risk for alzheimer’s disease dementia. J Prev Alzheimer’s Dis:1–8 Schmidt G (2020) INVADE – Home monitoring of cardiovascular risk factors. https://www. cosinuss.com/en/portfolio-items/study-invade-home-monitoring-of-cardiovascular-risk-factors/. Accessed 23 Sept 2020 Silvestre-Blanes J, Sempere-Payá V, Albero-Albero T (2020) Smart sensor architectures for multimedia sensing in IoMT. Sensors 20:1400 Sinha S, Singh PK, Lohia P, Dwivedi D (2020) Evanescent wave fiber optic biosensor: a theoretical study for detection of human intestine tissues. Sensor Lett 18:150–156 Spinka K, Spinka X (2015) Inhalation sensor for alternative nicotine/thc delivery device. Google Patents SugarBEAT, Daily Non-Invasive CGM (2020) SugarBEAT. https://sugarbeat.com/. Accessed 23 Sept 2020 Torres I, Baena MG, Cayon M, Ortego-Rojo J, Aguilar-Diosdado M (2010) Use of sensors in the treatment and follow-up of patients with diabetes mellitus. Sensors 10:7404–7420 Tzallas AT et al (2014) PERFORM: a system for monitoring, assessment and management of patients with Parkinson’s disease. Sensors 14:21329–21357 Uddin M, Khaksar W, Torresen J (2018) Ambient sensors for elderly care and independent living: a survey. Sensors 18:2027 van der Kamp MR et al (2020) WEARCON: wearable home monitoring in children with asthma reveals a strong association with hospital based assessment of asthma control von Coelln R et al (2019) Quantitative mobility metrics from a wearable sensor predict incident parkinsonism in older adults. Parkinsonism Relat Disord 65:190–196 Von Rosenberg W, Chanwimalueang T, Goverdovsky V, Looney D, Sharp D, Mandic DP (2016) Smart helmet: wearable multichannel ECG and EEG. IEEE J Transl Eng Health Med:4 Wei Q, Park H-J, Lee JH (2019) Development of a wireless health monitoring system for measuring core body temperature from the back of the body. J Healthcare Eng:2019 Wong SQ (2020) Sensor system design for respiratory sound classification Wong WY, Wong MS, Lo KH (2007) Clinical applications of sensors for human posture and movement analysis: a review. Prosthet Orthotics Int 31:62–75 Yan Y-S, Zhang Y-T (2008) An efficient motion-resistant method for wearable pulse oximeter. IEEE Trans Inf Technol Biomed 12:399–405

Oil and Gas Upstream Sector: The Use of IEC-61499 and OPC

41

Carlos A. Garcia, Gustavo Caiza, and Marcelo V. Garcia

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results from Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . State of Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OPC-UA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IEC 61499 standard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OPC-UA And IEC 61499 SIFBs for Integration of Communications within Industry 4.0 Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Set of SIFBs for OPC-UA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions and Ongoing Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1052 1056 1056 1057 1059 1059 1064 1067 1067 1075 1078 1079

Abstract

Oil production is currently decreasing, prices maintain their downward trend and extraction and operating costs have not changed as production has decreased, consequently, it is relevant to carry out initiatives aiming to reduce expenses

C. A. Garcia Universidad Tecnica de Ambato, UTA, Ambato, Ecuador e-mail: [email protected] G. Caiza Universidad Politecnica Salesiana, UPS, Quito, Ecuador e-mail: [email protected] M. V. Garcia () University of Basque Country, UPV/EHU, Bilbao, Spain e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_24

1051

1052

C. A. Garcia et al.

in the upstream sector of the oil & gas industry. This is the reason why oil production and exploration companies worldwide are immersed in what the 4th industrial revolution has called digital transformation as a step towards the increase of efficiency in processes. Taking this definition into account, it can be inferred that the digital transformation of this industrial sector is closely related to what is called Industry 4.0, which will allow the oil industry to have flexible architectures to process different types of raw oil, low automation cost and optimization of resources generating benefits both at the process, product and business model level. The aim of this chapter is to show an implementation of Industry 4.0 architecture in the oil & gas industry using the standard OPCUA and IEC-61499 in order to integrate distributed information from wells and transmit in a secure way. Keywords

OPC-UA · Industry 4.0 · IEC-61499 · Industry communication

Introduction The oil and gas industry are currently energy sources that are used worldwide, according to studies it is estimated that the world demand for oil will remain present during the next 20 years, despite the initiatives to use renewable energy as shows in Wanasinghe et al. (2020a, b). The high demand for production, difficult extraction and operating costs are a challenge that industries face on a daily basis, in addition to the oil price falls, the oil and gas industries tend to implement smart and innovative technologies with the purpose of maximizing production and reducing costs as found in previous studies (Lu et al. 2019; Shinkevich et al. 2020; Wanasinghe et al. 2020a). When talking about Industry 4.0, one of the concepts that has acquired the most relevance is the Internet of Things (IoT). The concept of IoT is important for the digital transformation of the industry. As a general concept it is not new, at the beginning the Internet was a network of networks that allowed connecting computers to share data. The term IoT was used for the first time in 1999, referring to identifiable objects (things) and their virtual representation in a digital infrastructure (Raza et al. 2018; Khan et al. 2017; Anderson 2017a; Aalsalem et al. 2017a). In its synthetic definition, IoT is a network that connects the physical (devices) and virtual (systems) worlds, where millions of devices and systems collaborate with each other and with others to provide intelligent services (smart) to users. Although the concept is not new, it is gaining more importance in the industry and in this field it has been conceptualized as the Industrial Internet of Things (IIoT), where the things to interconnect are machines, people and objects within the industrial environment (Lu 2017). Within Industry 4.0 another enabling technology that plays an important role is machine-to-machine communication (M2M). M2M is considered to be any

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1053

technology that allows devices that are in the same network to exchange information and perform actions in a totally autonomous way. This form of communication is mainly used for remote supervision of the machine itself and the environment in which it is located, turning it into a great help in the upstream sector of the oil industry because the extraction sites are distributed throughout a huge area in the jungle, desert or sea region (Biral et al. 2015). In this chapter, the authors must consider M2M as the base level of communication, while IIoT is using that communication to provide a solution for either the industrial or the personal field. In the short term, even those aggressively pursuing Industry 4.0 will have a mix of smart devices and equipment with traditional products and machines that need M2M communication to guide both the vertical and horizontal integration process. A third enabling technology on which Industry 4.0 is based is Cyber Physical Systems (CPS). CPS is the result of providing the components or physical objects that we regularly find in our work environment with computing and communication capabilities, turning them into intelligent objects. This update allows them to surpass the current simple embedded systems in terms of capacity, security, scalability, adaptability, resilience and usability, now being able to work together to form distributed and totally autonomous ecosystems (Monostori et al. 2016; Hajizadeh 2019; Sohraby et al. 2018). CPS go beyond the concept of IoT as it refers to more complex systems made up of other systems and that are capable of learning from the interactions they have with the physical world, in a way that makes environments intelligent. This concept applied to the industry has been called Cyber Physical Production Systems (CPPS). CPPS are made up of industrial control devices with extensive computing and communication capabilities, both local and remote (Kusiak 2018). Furthermore, the IEC 61499 automation standard promotes a framework for model-based development for distributed control systems. CPPS devices in conjunction with IEC 61499 will allow the modeling and development of software and hardware components for distributed control systems, making easier the collection and processing of data at the level of sensors and actuators, optimizing predictive maintenance techniques for equipment and allowing the incorporation of new control techniques in the upstream sector of the oil and gas industry (Patil et al. 2018). To develop distributed software components, IEC 61499 provides the Service Interface Function Block (SIFB) model, that allows the encapsulation and abstracts the user from accessing hardware, communications or resources of the Application Programming Interfaces (API). The IEC 61449 standard provides the following characteristics (Dai et al. 2018): • Portability: support and correctly interpret software components and configurations created by different software tools. • Interoperability: the different embedded devices can work together to carry out the functions of distributed applications.

1054

C. A. Garcia et al.

• Configurability: any device and its software components can be configured by multi-vendor software tools. • Reconfigurability: involves the ability to modify control hardware and software during process operation. • Distribution: allows the distribution of software components in different hardware devices regardless of the provider, which is a necessary requirement given by the automation industry. The use of standards such as IEC 61499 and CPPS architectures would imply considerable savings in the automation hardware used by the oil industry since this type of technology can be implemented in low-cost devices as found by Castellanos et al. (2017). Low-cost automation enables new technologies to be incorporated quickly and with a low impact on company budgets, its main advantages are the use of hardware from various manufacturers, optimization of production and incorporation of greater flexibility and individualization of the automation of oil extraction and production processes. Another problem that the oil production industry is having in the upstream, mindstream and downstream cases, is its characteristic of being formed by a structure distributed in multiple geographical locations very distant from each other, as mentioned in previous paragraphs. Therefore, there is a clear need to improve the management of the distributed process, adapting the control system to the characteristics of multiple locations for the process components, while offering the possibility of monitoring from different locations and responding to production demands in real time. To achieve this, it is necessary to have the possibility of remotely monitoring the status and evolution of the crude dehydration, transportation and refining process. In this sense, the oil production industry could additionally benefit from proposals promoted in the Industry 4.0 as found by de Melo and Godoy (2019). This initiative, as explained in previous paragraphs, promotes the introduction of factory concepts such as IIoT which defines a set of technologies for remote access to the production system and the Internet of Services (IoS), which provide mechanisms for the cloud-based manufacturing and process virtualization. These initiatives will have a collective impact on all aspects of manufacturing companies and it is convenient to introduce these concepts in real industries such as the oil and gas sector (Bassi 2017). The developments during the last years about the use of IIoT into Oil & Gas industry is shown in Fig. 1. Industry 4.0 insists on the importance of using consolidated standards to formally represent the information that is captured, stored and transmitted, and to specify the communication technologies for its transmission. In this context, the OPC-UA standard (IEC 62541) is a promising alternative for the implementation of M2M communications as it has a service-oriented architecture that offers data security and reliable information models at the same time. Unfortunately, and perhaps due to the lack of availability of suitable platforms, this technology is not yet widely used industrially as You and Feng (2020) show.

Fig. 1 Developments about the use of IIoT into Oil & Gas industry

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC 1055

1056

C. A. Garcia et al.

OPC-UA uses configuration files for servers to perform data integration at the plant level. These configuration files must necessarily follow the models of the plant and the devices that supply the data. Within the standards widely used at the industrial level we have the ISA95 standard (IEC 62264) and the physical model of ISA88 (IEC 61512) which propose a hierarchical model where the role that equipment plays in production processes is characterized. The use of these standards will make it simpler for the initial configuration files of OPC-UA servers to follow an information model accepted at an industrial level, allowing a rapid integration of this communication technology in the Ecuadorian oil industry sector. Considering the needs identified in the oil production sector, the motivation for this work is to build bridges between the state of the art and the design of industrial control applications. This can be achieved through the establishment of methodologies and the integration of a CPPS platform for their use in distributed control under the IEC 61499 standard in low-cost embedded devices, and an easily configurable OPC-UA architecture with servers and clients managed by IEC 61499 applications with configuration files that follow meta-models based on ISA 88/95. Therefore, the presented proposal allows an easy deployable software platform destined to remotely collect process data and that enables vertical integration from the company or business areas to the plant production level. The design of the document is as follows, section “Literature Review” shows some related works that have been used as a starting point for this research. Section “State of Technology” describes the state of the art that introduce the elements used in the methodology of this research The proposed solution where the IEC61499 SIFBs and OPC-UA integration is presented in section “OPC-UA And IEC 61499 SIFBs for Integration of Communications within Industry 4.0 Architectures.” Section “Case Study” illustrates a case study where an Industry 4.0 architecture is implemented in an Oil & Gas industry. Finally, some conclusions and future work are established in section “Conclusions and Ongoing Work.”

Literature Review This section was based on the review of recent developments and practices in this area, to address works and methodologies used by other authors and in this way know the progress at an industrial level, considering the benefits and the impact that certain technologies provide for the deployment of Industry 4.0 in the oil and gas industry, where they are leaning towards using cutting-edge technology that allows them to be more competitive and efficient within the market

Methodology Based on the specified topic, search equations were performed in different databases, such as Scopus, IEEE, Science Direct, Springer, ACM, and Google Scholar. In addition, the selection of the documents to be studied was made

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1057

fundamentally based on the content that each of the authors’ ideas contributes to the development, use, and vision of how Industry 4.0 is being implemented within O&G companies (OIL and GAS). The scope of the literature review described in this chapter in the O&G sector. The initial search based on keywords and later we identify and eliminate articles that do not address the required topic. Identifying articles started with extensive inquiries about databases of the scholarly literature of a series of topically essential keywords. For instance, the following keywords were used, typically in sets of two and groups of three or more as well: [“industry 4.0” and “oil,” “industry 4.0”], [“gas,” “industry 4.0,” and oil], [“gas,” “intelligent oil fields,” “intelligent oil fields,” and “pipeline digitalization”], and [“IIoT in oil and gas”].

Results from Literature Review The information extracted from each document is based on answering if the use of Industry 4.0 helps Oil & gas real industries in all the world. Various aspects such as the fields in which Industry 4.0 is developed for communication integration, process control, process optimization, advantages, disadvantages, and the gaps of this digital tool have been considered. On the other hand, a detailed description of the studies included in the review is presented in the next paragraphs. Through technological development, the optimization of oil and gas wells is sought with monitoring in real-time, decision-making through data analysis and remote control of devices as found in Huiyun et al. (2020), to increase the speed of exploration and detection, optimization of reserves, acceleration of exploitation, increase in production, reduction of risks to health and the environment as found in previous studies Roshani et al. (2017a); Alguliyev et al. (2018). As Qing and Heripracoyo (2019) shows Industry 4.0 is at the forefront of digital transformation since it allows the collection, configuration, processing, and analysis of data in real-time with better visibility and administration capabilities, dealing with processes such as production, purchases, sales, and logistics. Due to the advantages that this technology presents, it is being implemented in the oil and gas industry to take advantage of all its advantages and characteristics, in recent years digital technologies such as Industrial IoT (IIoT), Big Data, artificial intelligence, chain of blocks, etc (TOMA and POPA 2018; Nguyen et al. 2020). IIoT is the massive connection of smart devices and networks with CPS that comprise cloud computing platforms and has the potential to contribute to economic growth and global competitiveness Allahloh and Mohammad (2018), in terms of improving productivity, efficiency, maintenance prediction, etc. as presented by other studies as Choo et al. (2018); Tyagi (2019). In addition, it aims to take advantage of open high-bandwidth protocols and low-cost smart grids (Paez and Tobitsch 2017), and one of the main challenges when applying this technology is data security in these systems, for which cryptographic techniques can be used considering that IIoT devices have limited resources.

1058

C. A. Garcia et al.

In the area of oil and gas, the connectivity of the pipelines and processing plants with refineries is increasingly full of intelligent sensors that is like the nervous system of the industry (Anderson 2017b; Mangayarkarasi et al. 2019) and all these data must be analyzed and processed using intelligent control techniques in order to optimize processes. The transformative trend of IIoT is the link between connectivity, cloud, and analytical technologies to simplify process automation, which is essential to improve remote production and monitoring and thus be more competitive in these difficult times in the oil and gas industry. The oil and gas industry can use these advancements for 3D viewing on a PC and this, in turn, can revolutionize traditional oil and gas workflows. The use of PC technologies will generate greater opportunities for collaboration and faster response times. Furthermore, web services and technologies will continue to grow, and the oil industry will substantially increase the number of web applications developed and used (Evans et al. 2002). There are various applications and proposals for IIoT within the oil and gas field. The research works of Allahloh and Mohammad (2018) and Anderson (2017b) demonstrated that have developed IIoT systems for management and control applying intelligent control techniques such as Neuro-Fuzzy controllers and integration with Enterprise Resource Planning (ERP) and Systems, Applications, Products in Data Processing systems (SAP), and predictive and prescriptive analytics techniques to maximize production. In addition as found in previous studies Alguliyev et al. (2018); TOMA and POPA (2018); Zhao et al. (2008) highlight the capabilities of the cloud, big data, and data mining to provide security solutions and real-time monitoring for IoT that could be applied in the oil and gas industry. the integration of sensor data, communication channels for data analysis and management. As presented in the papers of Tyagi (2019); Wang et al. (2019); Aalsalem et al. (2017b) shows an intelligent architecture design for management and decision making with a resource exchange center that has an IIoT system that is designed for the management of exploration and development of the gas field, they also propose an intelligent monitoring system based on IoT. Smart IoT objects are capable of detecting important parameters and thus reporting failure notification and predictive good maintenance in order to reduce production downtime. The new smart control techniques for machines have been implemented in several works to develop intelligent and reliable models, in studies conducted by Roshani et al. (2017b); Yang et al. (2019); Fath et al. (2020); Eyoh and Kalawsky (2018) implement neural networks to identify flow regimes and predict volume fractions, estimate The gas-oil ratio and perform predictive maintenance based on performance can also be used to give an early warning of third-party threat and even damage to oil and gas pipelines. The results have shown that the models proposed in these works exceed the empirical correlations considered. Ling (2021) presents a gray box and neural network model is presented to predict oil production, the results provide a meaningful guide and provide a reliable basis for the oil field development plan. Another method used by Wu et al. (2017) is the knowledge approach driven by thresholds where

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1059

an intelligent oil well production architecture is built and a management system is developed where the results show that it improves the effectiveness and precision of early warning and the fault alarm. García et al. (2018) present an architecture for vertical integration based on cyber-physical systems, under the IEC 61499 standard and using OPC UA, in which the results showed that it is suitable for use in flexible manufacturing in the oil and gas industry. The transport of oil and gas is of great importance worldwide because it allows the distribution of this from its point of extraction to the processing plants, in order to avoid leaks or ruptures in the pipes due to the fact that this directly impacts the ecosystem and economy. There are several jobs where they use different methods and techniques shown below. Sasloglou et al. (2013) propose an empirical channel model to optimize the placement of wireless sensors in oil transmission pipelines, where it is sought to have a minimum number of sensors and the best possible performance. Hausamann et al. (2005); McConnell (2007) propose to use unmanned aerial vehicles (UAV) for the inspection of pipelines and thus reduce the costs and risk of inspections in addition to analyzing the design of the information operation and maintenance system of oil and gas pipelines of high availability and long-distance. Another method used is the one proposed in the studies conducted by Pajany and Hemalatha (2019); Meléndez Pertuz et al. (2017) whose objective is to detect the gas leak in the pipeline, monitoring at regular intervals using gas detection sensors and will pass the information and location to the operator. Digitization is also used in the management part, as shown by Galiakhmetov et al. (2020) where it considers the digitalization instruments of logistics business processes in the selection of the most efficient oil transportation option, which will allow to considerably reduce the workforce. Furthermore, Shinkevich et al. (2020); Alguliyev et al. (2016) have carried out a study of the industries and how digital transformation in conjunction with big data has been used for the sustainable development of the oil and gas sector, obtaining an architectural model for integration and big data convergence, business analytics, providing guidelines to enable companies to digitally transform and implement IoT technologies to optimize and improve business processes demonstrated by Buhulaiga et al. (2019).

State of Technology This section will explain all the concepts that will be applied for the development of the research proposal.

OPC-UA The latest tools for supervision and management of manufacturing processes need to collect in real-time information from production means whose operation

1060

C. A. Garcia et al.

is directed (in many cases) by controllers from different manufacturers. Typical vertical integration investigations usually separate automation systems into different layers. The resulting structure is usually known as the automation pyramid. Although this is a reliable approach, this means that all requests from systems such as: invocation of functions, data acquisition, data transfer, etc, must be transferred from one layer to another. When performing this exchange they generally go through proprietary communication interfaces where they must perform the associated translations, but sometimes these translations are prone to errors (Kim and Sung 2018). Despite the existence of communication standards such as OPC-DA (acronym for Ole for Process Control-Data Access) and programming standards such as the IEC 61131-3 standard, automatic data capture in the plant tends to be an extremely complicated and problematic section. The causes of this problem lie in the different structures, contents and views of the data that reside in each controller and for which there is still no general rule to follow. The Unified Architecture (UA), proposed by the OPC Foundation, is capable of giving a solution to the problems that the integration of data at the plant level currently has on: how to locate the interlocutor of a communication, how the data, metadata and metadata of the control logic are identified, and how the information that must be made visible to other devices is defined. However, all points of view of the data integration problem at the plant level must be taken into account if the different manufacturers are to be able to exchange information between them in an appropriate way (Grner et al. 2016). OPC-UA provides an industrial interoperability framework of the future. OPCUA defines two main pillars that support interoperability: the communications infrastructure and the meta-model, as seen in Fig. 2. The communications infrastructure defines how information is exchanged, and the meta-model defines what information must be exchanged. The Web Server (WS*) communications infrastructure can facilitate its integration with business applications and Internet access. On the other hand, the use of a binary transmission protocol based on Transmission Control Protocol (TCP) allows efficient data exchange in terms of performance (Vimos and Sacoto Cabrera 2018). In addition to all this, a series of special mechanisms defined in the API and in the protocols are integrated to guarantee reliable communication in the case of working with distributed systems. The OPC-UA standard consists of 13 different parts of which some have already been adopted by the International Electrotechnical Commission as the IEC 62541 standard. This standard has different chapters which we summarize below: the first chapter presents a overview of this standard, a broad security concept is presented in Chap. 2, the Address Space model is given in Chap. 3, Chap. 4 gives its concepts and abstract services functions, Chap. 5 indicates the model information, technology mappings in Chap. 6, profiles in Chap. 7, access to data is given in Chap. 8, data visualization in Chap. 9, method of calls and historical access data in Chaps. 10 and 11, respectively and functionalities for OPC-UA servers in Chaps. 12 and 13 (see Fig. 3).

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1061

Fig. 2 The foundation of OPC UA

Fig. 3 OPC UA Specifications

Information Model from OPC-UA While the classical OPC standard has a very elementary meta-model because it is made up of labels in a simple and straightforward hierarchy, OPC-UA offers a

1062

C. A. Garcia et al.

rich information meta-model with object-oriented techniques using more elaborate hierarchies and with the possibility of creating inheritance instances between them. OPC-UA provides a standard meta-model, known as Address Space, which can be used to modify it to define specific Information Models. It is hierarchically designed and all the higher levels of the OPC-UA servers are designed in the same way, thereby achieving interoperability between them. The meta-model in OPC-UA defines the objects to be used: types, variables, data types, etc. (Mahnke et al. 2009). Also, it has information on the set of instances (or objects) existing in the system and the type of each one of them. In this way, any system, no matter how complex, can be fully described employing object-oriented mechanisms. Typically, very few embedded systems will implement all of the modeling options described above. By using meta-models, information with known semantics can be exchanged efficiently instead of simple data. The information model is always present in an OPC-UA server, in this way the client does not need to have this model integrated. The idea is that OPC-UA specifies how data is exchanged, while standard information models specify what information is exchanged. The great interest in information modeling has generated the impetus to standardize information models based on OPC-UA. This standard defines a series of web services, which allow the OPC-UA client to search and edit objects in the address space of a server. OPC-UA additionally defines a set of abstract services that can be executed in different communication infrastructures and use the meta-model as a basis to define the appropriate parameters for the services. The basic information model of OPC-UA provides the primary object types and access points to the server address space. Standard or provider-specific information models can be established on the base information model. OPC-UA already defines various standardized information models for accessing data, alarms and conditions, programs, historical data, and aggregate functions. As can be seen in Fig. 4. It also provides the mechanisms to support multiple information models on one server. This allows to provide a high level of interoperability because the data can not only be changed in a simple way between devices, but also with clearly defined semantics.

Security from OPC-UA Security is a fundamental requirement for OPC-UA and therefore it is integrated within the architecture. The mechanisms are based on a detailed analysis of the threats. OPC-UA manages security with the authentication of clients and servers, the integrity and confidentiality of the exchanged messages, and the verifiability of the functional profiles of equipment. It is based on the architecture shown in Fig. 5. The three levels of security are user security, application security, and data transport security. The OPC-UA user-level security mechanisms run once when a session is configured. The client transmits an encrypted security token, which identifies the user, to the server. The server authenticates the user based on the token and then authorizes access to the objects on the server. The OPC-UA standard does not define authorization mechanisms such as access control lists.

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1063

Fig. 4 Comparison between OPC and OPC-UA models

Fig. 5 OPC UA Security Architecture

OPC-UA has security at the application level and is also part of the session configuration and includes the exchange of digitally signed certificates. Two types are the instance certificates that identify the specific installation of the communication software and the software certificates that identify the client and the software server and the implemented OPC-UA profiles. They describe server capabilities, such as support for a specific information model as Schleipen et al. (2016) shows.

1064

C. A. Garcia et al.

Transport-level security certificates can be used to provide integrity by signing messages and confidentiality by encrypting messages. This prevents the disclosure of the information exchanged and ensures that the messages have not been tampered with. The security mechanisms are realized as part of the OPC-UA stacks. They are included in a software package provided by the OPC Foundation and are ready to be used by the client and the server. OPC-UA offers a complete solution that meets the requirements of all vertical layers for access to remote devices. It provides an efficient and secure infrastructure for communications between sensors and controllers in the automation and control systems of manufacturing and the processes of the companies. Provides secure communications using computer industry standards. This allows a multitude of devices to receive and convert remote data into useful information to make intelligent decisions. They provide a multi-vendor infrastructure, cross-platform interoperability, and security for industrial automation environments.

IEC 61499 standard This standard was created for distributed control systems, including their architecture and software tool requirements. It was developed as a consequence of the growing interest in new technologies and architectures to create the next generation of industrial systems and based on the IEC 61131 standard. IEC 61499 was designed by the technical committee TC-65 for measurement, control and automation of industrial processes (TC, Technical Committee), which belongs to the IEC, the first version being approved in August 2005 (Vyatkin 2011a). This standard defines a generic architecture and a guide for the use of Functional Blocks (FBs) in Control Systems and Measurement of Distributed Industrial Processes (IPMCSs). One of the main objectives of IEC 61499 is to promote the development of heterogeneous systems composed of control devices from different manufacturers allowing dynamic reconfiguration, resulting in the modification of the configuration of a system while the control application continues to run. IEC 61499 is seen as the next generation of standards in automation systems and is designed to cover interoperability, portability, and reconfigurability, which are not covered by IEC 61131-3. At the moment, in industrial practice there are few systems based on IEC 61499, but nowadays, a large number of research works accept and use the basic concepts of the standard.

IEC 61499 Specifications The IEC 61499 standard is divided into the following four sections (Strasser et al. (2008):

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1065

– Architecture: IEC 61499-1 contains the general requirements, definitions and reference models. In addition to this, it has the rules for declaring the types of FBs and rules for their behavior. – Software tool requirements: IEC 61499-2 defines the requirements of the software tools used for this standard, which support the execution of engineering tasks in distributed control systems and the specification of the types of FBs. – Informative Manual: IEC 61499-3 contains the information for the understanding, acceptance and applicability of both the IPMCS architecture and software tools that comply with the specifications of the standard. – Rules and Compliance Profiles: IEC 61499-4 contains the definition of the rules for the development of profiles in accordance with the standard, which specify the characteristics to implement sections 1 and 2.

Architecture IEC-61499 defines a generic and hierarchical architecture of models, allowing to understand the organization of the system and its components. It develops a new framework for distributed control applications. The models are generic, independent of the domain and extensible with the definition and use of FBs. The models are: Functional Block Model (FB), Resource Model, Device Model, System Model, Application Model, Distribution Model and Management Model. Functional Block Model (FB) is the smallest element in a distributed control system. The FB is made up of two parts: head and body. The first consists of a head that is connected to the flow of events. It accepts input events and generates output events, as shown in Fig. 6. The second part is the body, which is connected to the data flow, accepts the input data and generates output data. The dynamic behavior of the FB is defined by the Execution Control Chart (ECC, Execution Control Chart) that processes the event inputs and generates event outputs (Vyatkin 2011b).

Fig. 6 IEC-61499 Functional Block Model (FB)

1066

C. A. Garcia et al.

An FB in the IEC-61499 standard remains passive until it is triggered by an input event, that is, all input events are used to activate a functional block. The FB executes and produces events and output data as shown in Fig. 6. The ECC describes the internal behavior of the basic FB instances. It helps the programmer break complex behavior down into small parts called states. Each state is valid under a certain set of conditions. The states are associated with one or more algorithms and / or with output events. The activation of the state implies the execution of the attached algorithms. The FB functionality is provided by algorithms. An algorithm can be written in any of the 5 languages mentioned in IEC 61131-3: IL, ST, LD, FBD and SFC. Also in other high-level languages such as: C, C ++, Java and Delphi. The algorithm processes internal and/or external inputs and data, generating output data. Internal variables or status information are not accessible by the data flow.

Framework platform for IEC-61499 4DIAC™ Framework for Distributed Industrial Automation and Control is an open source engineering tool based on the Eclipse platform for distributed, reconfigurable automation and control software. 4DIAC™ was created in 2000 by PROFACTOR GmbH & Vienna University of Technology.

Development environment: 4DIAC-IDE The editor is an integrated development environment called 4DIAC-IDE. The objective of the 4DIAC initiative is to provide tools in accordance with the standard that allow the establishment of an automation and control environment, based on the objectives of portability, configurability and interoperability, which are mentioned in IEC 61499. 4DIAC pursues the following goals (Zoitl and Vyatkin 2009): – Provide a common basis for development, industrial device and research of IEC 61499. – Supply a package containing a runtime environment for different embedded control platforms and the engineering environment. – Provide real examples at the prototype level to increase the acceptance of IEC 61499 in the industry. – Provide an incentive for the use of IEC 61499 with industry. The most relevant characteristics of 4DIAC are: (i) 4DIAC-IDE is an IEC 61499 tool based on Eclipse. It has elementary data types according to IEC 61131-3. (ii) Has event and data connections. (iii) This software has command configuration (create, write, start according to IEC 61499). (iv) Has communication FBs (Client/Server, Publish/Subscribe for Ethernet). (v) It can execute basic FBs, Composite FBs, FBs, SIFBs service interfaces, and adapters. It runs under Windows, Linux and Solaris platforms.

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1067

Execution Environment: FORTE The 4DIAC runtime environment according to IEC 61499 is 4DIAC-RTE (FORTE). FORTE is a small portable implementation of a runtime environment according to IEC 61499 focused on small embedded control devices (16/32 Bit), implemented in C++ and is portable for multiple platforms (Seilonen et al. 2019). The execution mechanisms in FORTE allow limited real-time execution of IEC 61499 control configurations triggered by external events, where different parts of the configuration can meet different real-time constraints and the execution of low-priority processes does not disturb the execution of the highest priority processes. In the runtime environment FORTE makes use of an event trigger for scheduling FBs. The scheduler introduces a queuing approach where all incoming events are delivered to the target FBs in FIFO order (first in-first out). The event trigger decouples the FB event sending execution from the receiving block, thereby creating the blocking period of an FB independent of the network topology (Vyatkin and ISA 2007). The runtime environment continues to change when it comes to optimization of execution and implementation of communications interface. FORTE has been written to work independently of the platform to be used, this allows to make it easier to use it on various types of hardware and operating system platforms. The current version of FORTE is supported by the following systems: Windows (Win32), eCos, POSIX, Lego Mindstorms NXT controller, etc.

OPC-UA And IEC 61499 SIFBs for Integration of Communications within Industry 4.0 Architectures Typically, a profile for OPC-UA is designed for vertical integration of industrial control and automation systems. The use of an OPC-UA architecture allows a complete description of any automation system data regardless of its complexity. In this sense, the information described in the Address Space of the OPC-UA server may be sufficient for most industrial systems. Next, vertical integration at the plant level using IEC 61499 as a tool to implement and integrate process data using OPC UA as a communication protocol within the Industry 4.0 architecture is explained.

Set of SIFBs for OPC-UA This section presents the development of a set of SIFBs that encapsulate the operations offered by OPC-UA services, following the previously explained methodology.

Service Interface Function Block for implement an OPC-UA server By using this SIFB it is possible to manage the configuration and operation of a static OPC-UA server. The OPC-UA server is configured using a file in XML format and allows integrating the devices at plant level using native protocols of industrial networks such as Modbus TCP, S7 communications, etc.

1068

C. A. Garcia et al.

Fig. 7 OPC-UA server deployment model

This file follows the CPPS architecture and includes all the essential parameters for the operation and configuration of the OPC-UA server such as: the URL address, the URI identifier, name of the provider, the name of the server, the version, etc. The schema of the XML configuration file is presented in Fig. 7. This XML configuration file declares the following: • Server ID (Identification, Server information): This section gives basic information about the OPC-UA server such as the domain and its version. • Access to the Server (Server Access): This element provides access to the OPCUA server, since it allows entering the server name, URL address, and URI identifier. The configuration file depends on the communication stack used for the server implementation. • Address Space (NodeTypes, NodeInstances): This section declares the OPC-UA nodes (types and instances) that constitute the information model

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1069

corresponding to the Logical Process Nodes (LPNs), as well as the logical process data (tags) available in each of the nodes. In turn, this section is made up of two groups of elements: (i) Types of Nodes (NodeType), here the Types of Objects (ObjectTypes) and the Types of Data Variables (DataVariableTypes) associated with them are declared. (ii) The elements of the Node Instances (NodeInstance) are included in the Object Instances (ObjectInstances) - these are instances that have already been declared in the Object Types (ObjectTypes) - and the Data Variables (DataVariables) the which are instances of the Variable Types (DataVariableTypes). The Object Instances (ObjectInstances) constitute the LPNs, while the Variable Instances (DataVariables) constitute the tags associated with the LPNs. The explained can be seen in Fig. 8. • Field Devices (FieldDevices): This section groups together the definitions of the field devices to which the server has access. The process data (FieldData) supplied by each device are also defined. Field devices are characterized by the communication protocol or the process data access mechanism, and the information related to it. See Fig. 9. • Data Mapping (DataMapping): In this section the existing relationships between the data variables (DataVariables) already declared in the Address Space section and their corresponding process data (FieldData) of the field devices (FieldDevice) are defined. An example of the configuration file for the OPC-UA server in XML format can be seen in Fig. 10. Once the OPC-UA server configuration file has been explained, we proceed to explain the SIFB that allows communication at the field level by integrating them with the OPC-UA protocol. In addition to common events of SIFBs in the IEC 61499 standard such as INIT, REQ, INITO and CNF explained previously (Fig. 11), the SIFB OPCUA_SERVER also presents the following input and output parameters: • QI (BOOL): This input data works together with the INIT event to connect or disconnect the OPC-UA server. If the INIT event is requested, and if QI is TRUE, the OPC-UA server tries to connect; if QI is FALSE (FALSE), the OPC-UA server ends its execution. • CONFIGFILE (WSTRING): Contains the full name of the configuration XML file. • QO (BOOL): Reports on the result of the last process executed. • STATUS (STRING): Provides information on the status of the server.

Service Interface Function Block for implement an OPC-UA server dynamically This SIFB allows OPC-UA communication using the native IEC 61499 communication layer called FORTE CommLayer. This network communication interface has been designed to be as flexible as possible. The basic steps to design this interface start from implementing the OPC-UA protocol with libraries developed in C++ that allow interacting with the standard design given by the norm. The used communication stack allows to integrate OPC-UA in FORTE (4DIAC RTE), which

Fig. 8 OPC-UA Address Space Meta-Model

1070 C. A. Garcia et al.

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1071

Fig. 9 Field Device Model

Fig. 10 OPC-UA Server Configuration File in XML format

is fully scalable, supports multi-threaded architectures, and where each connection or session is operated by separate threads. This communication stack can be run on various devices, either in industrial computers or in low-cost embedded devices such as Raspberry Pi™ or BeagleBone Black™ cards. The architecture of this implementation is shown in Fig. 12. Here you can see the different types of FORTE classes and the basic interaction between them. The class of the Function Block called CCommFB is the only class that interacts

1072

C. A. Garcia et al.

Fig. 11 SIFB for implement an OPC-UA server

Fig. 12 OPC UA server SIFB implementation architecture

with IEC 61499 applications and allows sending and receiving events based on the IEC 61499 standard. The function block and the communication layer implemented are related using the following data inputs and outputs implemented in the SIFB (Fig. 13): • ID (WSTRING): It is an entry in the SIFB that is used to implement all the parameters that the OPC UA protocol needs for its correct configuration. Parameters are entered as a character string, there are mandatory [m] and optional [o] parameters. These parameters are described below: Server name (Server name) [o], its input mode is svr: < servername> the name allows identifying the implemented OPC UA service; Configuration file address [m] this parameter is entered using the cnf code line: < configfile> reads a memory file in xml format which contains the URL, URI, author of the OPC UA server, etc. ; Field device name [o] use the fd code: < fdname, driver, params, ...> enter the name of the field device, the industrial communication protocol driver that we use to

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1073

Fig. 13 SIFB SERVER OPC-UA dinamically

integrate the variables the process and additional parameters for its configuration; Process Variable (Process Tag) [o] the input format is tg [num]: < TagName, Type, AccessLevel, FDName ...> with this line of code we enter process variable followed by a number and include its name, type variable, its access level, etc. • SD (ANY): It allows writing the value in the address space of the OPC-UA server. It is an input memory parameter (parameter - memory tag) [o] it is entered using the following format sd [num]: < TagName, Type, AccessLevel> as in the previous case the variable is entered but only found in the memory of the embedded device to be created. This variable must be accompanied by a number and include its name, type and access level. • RD (ANY): Data provided by the server to which it is connected. As in the previous entry, this parameter is memory tag [o], the input code line is rd [num]: < TagName, Type, AccessLevel>.

Service Interface Function Block for implement an OPCUA Client Read This SIFB makes it possible to implement an OPC-UA client capable of synchronously querying variables from an OPC UA server. In addition to the common events, this SIFB has the following input and output parameters (Fig. 14a): • URLSERVER (WSTRING): Indicates the URL of the OPC-UA server. • DATANAME (WSTRING): Name of the variable to be accessed on the OPCUA server.

1074

C. A. Garcia et al.

Fig. 14 Different SIFB for OPC-UA Clients. a) CLIENT_READ SIFB b) CLIENT_WRITE SIFB c) CLIENT_SUBSCRIBE

• RD (ANY): Value of the data returned by the OPC-UA server to which the client is connected. The type of this parameter is ANY to improve its reuse in any type of application. • SOURCETIMESTAMP (DATE AND TIME): Indicates the time information associated with the element in the source resource. • SERVERTIMESTAMP (DATE AND TIME): Indicates the time information associated with the element on the OPC-UA server.

Service Interface Function Block for implement an OPCUA Client Write This SIFB (Fig. 14b) creates a client for synchronous writing of variables on your OPC-UA server. It presents the following parameters: • TYPE (WSTRING): It expresses the data type for the value to be written to the OPC-UA server. • SD (ANY): Indicates the value of the data to be written. As in the previous SIFB, the type of this parameter is ANY in order to increase generality.

Service Interface Function Block for implement an OPCUA Client Subscribe Besides the synchronous read and write services, OPC-UA enables the subscription mechanism. The subscription maintains a local copy of the parameters of the item to be monitored. These local copies can be altered by updating their properties without affecting the state on the server. To implement this mechanism, a SIFB has been

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1075

created (Fig. 14c) that allows creating subscribing clients to monitor variables from OPC-UA servers. The input and output parameters of this SIFB are: • MODE (WSTRING): Two modes are allowed to monitor the OPC-UA server variables: “REPORTING” and “SAMPLING.” • PERIOD (ULINT): Sets the sampling period in milliseconds for items in sampling mode. • RD (ANY): Provides the value of the monitored elements supplied by the OPCUA server. • SOURCETIMESTAMP (DATE AND TIME): Indicates the Timestamp associated with the element of the source resource. • SERVERTIMESTAMP (DATE AND TIME): Shows the Timestamp associated with the OPC-UA server element .

Case Study The case study focuses on Petroamazonas EP, an Ecuadorian public company dedicated to the exploration and production of hydrocarbons. It is in charge of the operation of 21 Blocks, located in the Oriental basin and in the Litoral area of Ecuador. Due to the infeasibility of implementing the system in all the blocks, the design has been simplified towards the set of wellpads of a specific Block. In our case, Block 18 has been selected. In this Block there are currently 4 wellpads and each of them groups approximately between 28 to 30 oil wells. The example aims to obtain the variables of a production well, specifically, it refers to well PAA-001 whose scheme is shown in Fig. 15. The head of the well called x-mas wellhead, labeled PAA-01-01, has two indicator transmitters: TIT-WPAA01-01 that indicates the temperature and PITWPAA01-01 the pressure of the extracted oil. It also has a production emergency valve, the SDV-WPAA01-01. Because the natural pressures of the reservoirs in Ecuador do not provide enough pressure for the oil to reach the surface, each well uses a BESPA1 electricsubmersible pumping system, which corresponds to element M of Fig. 15. The equipment BES has two pressure transmitters: PT-BESPA1-01 for the inlet pressure of the pump (intake) and the PT-BESPA1-02 for the discharge pressure of the pump (exhaust); three temperature transmitters: TT-BESPA1-01 for fluid temperature, TTBESPA1-02 for engine temperature and TT-BESPA1-03 for outlet fluid; a sensor that measures the motor current, CT-BESPA1-01; The motor voltage is measured by VT-BESPA1-01 and the FT-BESPA1-01 indicates the value of the frequency of the frequency inverter, VSD. Furthermore, the core of the ESP system is composed of a motor (M-WPA01-1), multi-stage pump (P-WPA01) and a VSD variable speed drive (VSD-WPA01). Figure 16 shows the current communication system in which the OPC-UA server is located.

1076

C. A. Garcia et al.

Fig. 15 Electric-submersible pump system and xmas wellhead

The design and development of an Information Components (ICs) using the IEC61499 architecture and OPC UA communication is a structured, but simple process, if the following steps are followed: Step i) Decompose the acquisition process of the field variables in a detailed and hierarchical way to be able to be represented with the FBs developed for OPC-UA and the existing ones of the IEC 61499 standard. Step 2) Select the corresponding FBs from the library of OPC-UA or the standard FBs for all the sub-functions designed in the previous step, then you must create the instances of the selected FBs, configure their parameters and connect these FBs with data flow and event flow. Step 3) Validate the entire IC model to avoid errors. Step 4) Integrate the physical components to the ICs and perform complete communication tests of the different tags. As an example of a case study, the design of an IC using the standard FBs and FBs for OPC-UA communication developed in this work is shown. Figure 17 shows the architecture of a network of IC FBs. The IC consists of three main modules: i) the static OPC-UA server, which reads its configuration in XML format; ii) the dynamic creation of new labels to process variables using the server FB that uses the FORTE communication layer and, if required, use Atomic Services FBs (Data Storage, Data Reporting, Data Reading, etc. ) and iii) finally, once the ICs have been created, the integration of communication services with the field devices must be carried out using the client FBs (Write, Read, Subscribe) of OPC-UA.

1077

Fig. 16 Electric-submersible pump system and xmas wellhead

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1078

C. A. Garcia et al.

Fig. 17 ICs implementation into IEC-61499 Architecture

Conclusions and Ongoing Work This chapter presents an approach to access to field data in process control systems using OPC-UA servers in low-cost CPPS architectures while using applications under the IEC 61499 standard. In this case, the application of this architecture has been directed towards the oil and gas industry. The use of this type of architecture helps to introduce new elements about CPPS within the Industry 4.0 paradigm. The proposed architecture provides an M2M infrastructure for plant-level communications and the integration of higher-level devices into the production process. The use of a set of SIFBs is proposed to implement OPC-UA clients and servers, including subscription mechanisms. These function blocks allow, in a quick and easy way, the construction of new distributed applications based on components through programming environments under the IEC 61499 standard. Using function blocks to design information components provides obvious and potential benefits. Perhaps the simplest, but most important, is that the same information component, which, in addition to being reusable, is applicable in different industrial processes through software reconfiguration. It is also feasible to

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1079

add, remove and replace some functions of an information component in production when requested by the application requirements. As future works, the extension of the SIFBs library is necessary to provide the designer of distributed control systems with a set of networks sufficient to be able to solve any type of industrial communication, including protocols such as MQTT or AMQP. The methodologies and tools developed in this chapter open the field for the possible integration of new tools and functionalities compatible with the IEC 61499 standard.

References Aalsalem M, Khan W, Gharibi W, Armi N (2017a) An intelligent oil and gas well monitoring system based on internet of things. vol 2018-January, pp 124–127, https://doi.org/10.1109/ ICRAMET.2017.8253159 Aalsalem MY, Khan WZ, Gharibi W, Armi N (2017b) An intelligent oil and gas well monitoring system based on internet of things. In: 2017 International Conference on Radar, Antenna, Microwave, Electronics, and Telecommunications (ICRAMET), IEEE, https://doi.org/10.1109/ icramet.2017.8253159 Alguliyev R, Fataliyev T, Mehdiyev S (2018) Some issues of application of internet of things in the oil and gas complex. In: 6th International Conference on Control and Optimization with Industrial Applications, vol 1, pp 46–48 Alguliyev RM, Aliguliyev RM, Hajirahimova MS (2016) Big data integration architectural concepts for oil and gas industry. In: 2016 IEEE 10th International Conference on Application of Information and Communication Technologies (AICT), IEEE, https://doi.org/10.1109/ icaict.2016.7991832 Allahloh AS, Mohammad S (2018) Development of the intelligent oil field with management and control using IIoT (industrial internet of things). In: 2018 2nd IEEE International Conference on Power Electronics, Intelligent Control and Energy Systems (ICPEICES), IEEE, https://doi.org/ 10.1109/icpeices.2018.8897439 Anderson R (2017a) ‘petroleum analytics learning machine’ for optimizing the internet of things of today’s digital oil field-to-refinery petroleum system. vol 2018-January, pp 4542–4545, https:// doi.org/10.1109/BigData.2017.8258496 Anderson RN (2017b) ’petroleum analytics learning machine’ for optimizing the internet of things of today’s digital oil field-to-refinery petroleum system. In: 2017 IEEE International Conference on Big Data (Big Data), IEEE, https://doi.org/10.1109/bigdata.2017.8258496 Bassi L (2017) Industry 4.0: Hope, hype or revolution? In: 2017 IEEE 3rd International Forum on Research and Technologies for Society and Industry (RTSI), pp 1–6, https://doi.org/10.1109/ RTSI.2017.8065927 Biral A, Centenaro M, Zanella A, Vangelista L, Zorzi M (2015) The challenges of m2m massive access in wireless cellular networks. Digital Communications and Networks 1(1):1 – 19, https:/ /doi.org/https://doi.org/10.1016/j.dcan.2015.02.001 Buhulaiga EA, Telukdarie A, Ramsangar S (2019) Delivering on industry 4.0 in a multinational petrochemical company: Design and execution. In: 2019 International Conference on Fourth Industrial Revolution (ICFIR), IEEE, https://doi.org/10.1109/icfir.2019.8894790 Castellanos EX, Garcia CA, Rosero C, Sanchez C, Garcia MV (2017) Enabling an automation architecture of cpps based on uml combined with iec-61499. In: 2017 17th International Conference on Control, Automation and Systems (ICCAS), pp 471–476, https://doi.org/ 10.1109/CONISOFT.2017.00015 Choo KKR, Gritzalis S, Park JH (2018) Cryptographic solutions for industrial internet-ofthings: Research challenges and opportunities. IEEE Transactions on Industrial Informatics 14(8):3567–3569, https://doi.org/10.1109/tii.2018.2841049

1080

C. A. Garcia et al.

Dai W, Song Y, Zhang Z, Wang P, Pang C, Vyatkin V (2018) Modelling industrial cyberphysical systems using iec 61499 and opc ua. pp 772–777, https://doi.org/10.1109/INDIN.2018. 8472099 de Melo PFS, Godoy EP (2019) Controller interface for industry 4.0 based on rami 4.0 and opc ua. In: 2019 II Workshop on Metrology for Industry 4.0 and IoT (MetroInd4.0 IoT), pp 229–234, https://doi.org/10.1109/METROI4.2019.8792837 Evans F, Volz W, Dorn G, Fröhlich B, Roberts DM (2002) Future trends in oil and gas visualization. In: Proceedings of the Conference on Visualization ’02, IEEE Computer Society, USA, VIS ’02, p 567570 Eyoh J, Kalawsky RS (2018) Reduction of impacts of oil and gas operations through intelligent maintenance solution. In: Proceedings of the International Conference on Intelligent Science and Technology - ICIST ’18, ACM Press, https://doi.org/10.1145/3233740.3233747 Fath AH, Madanifar F, Abbasi M (2020) Implementation of multilayer perceptron (MLP) and radial basis function (RBF) neural networks to predict solution gas-oil ratio of crude oil systems. Petroleum 6(1):80–91, https://doi.org/10.1016/j.petlm.2018.12.002 Galiakhmetov R, Pushina N, Sokolova N (2020) Instrumentation for digitalization of logistic business processes of oil production organizations. In: Proceedings of the 2nd International Scientific and Practical Conference “Modern Management Trends and the Digital Economy: from Regional Development to Global Economic Growth” (MTDE 2020), Atlantis Press, https:/ /doi.org/10.2991/aebmr.k.200502.101 García MV, Irisarri E, Pérez F, Estévez E, Marcos M (2018) Arquitectura de automatización basada en sistemas ciberfísicos para la fabricación flexible en la industria de petróleo y gas. Revista Iberoamericana de Automática e Informática industrial 15(2):156, https://doi.org/ 10.4995/riai.2017.8823 Grner S, Pfrommer J, Palm F (2016) Restful industrial communication with opc ua. IEEE Transactions on Industrial Informatics 12(5):1832–1841, https://doi.org/10.1109/TII.2016.2530404 Hajizadeh Y (2019) Machine learning in oil and gas; a swot analysis approach. Journal of Petroleum Science and Engineering 176:661–663, https://doi.org/10.1016/j.petrol.2019. 01.113 Hausamann D, Zirnig W, Schreier G, Strobl P (2005) Monitoring of gas pipelines – a civil UAV application. Aircraft Engineering and Aerospace Technology 77(5):352–360, https://doi.org/ 10.1108/00022660510617077 Huiyun M, Chenggang Y, Liangliang D, Yukun F, Chungang S, Hanwen S, Xiaohua Z (2020) Review of intelligent well technology. Petroleum 6(3):226–233, https://doi.org/10.1016/ j.petlm.2019.11.003 Khan W, Aalsalem M, Khan M, Hossain M, Atiquzzaman M (2017) A reliable internet of things based architecture for oil and gas industry. pp 705–710, https://doi.org/10.23919/ ICACT.2017.7890184 Kim W, Sung M (2018) Standalone opc ua wrapper for industrial monitoring and control systems. IEEE Access 6:36,557–36,570, https://doi.org/10.1109/ACCESS.2018.2852792 Kusiak A (2018) Smart manufacturing. International Journal of Production Research 56(1-2):508– 517, https://doi.org/10.1080/00207543.2017.1351644 Ling M (2021) Prediction on petroleum demand under the epidemic crisis based on computerimplemented grey bp neural network algorithm. Advances in Intelligent Systems and Computing 1244 AISC:274–281, https://doi.org/10.1007/978-3-030-53980-1_41 Lu H, Guo L, Azimi M, Huang K (2019) Oil and gas 4.0 era: A systematic review and outlook. Computers in Industry 111:68 – 90, https://doi.org/https://doi.org/10.1016/ j.compind.2019.06.007 Lu Y (2017) Industry 4.0: A survey on technologies, applications and open research issues. Journal of Industrial Information Integration 6:1–10, https://doi.org/10.1016/j.jii.2017.04.005 Mahnke W, Leitner SH, Damm M (2009) OPC unified architecture. https://doi.org/10.1007/978-3540-68899-0 Mangayarkarasi N, Raghuraman G, Kavitha S (2019) Influence of computer vision and IoT for pipeline inspection-a review. In: 2019 International Conference on Computational Intelligence in Data Science (ICCIDS), IEEE, https://doi.org/10.1109/iccids.2019.8862109

41 Oil and Gas Upstream Sector: The use of IEC-61499 and OPC

1081

McConnell VP (2007) Military uavs claiming the skies with fuel cell power. Fuel Cells Bulletin 2007(12):12–15 Meléndez Pertuz FA, Gonzalez Coneo JE, Comas Gonzalez ZV, Nuñez Perez BE, Viloria Molinares P (2017) Integridad estructural de tuberías de transporte de hidrocarburos: Panorama actual 38(17) Monostori L, Kdr B, Bauernhansl T, Kondoh S, Kumara S, Reinhart G, Sauer O, Schuh G, Sihn W, Ueda K (2016) Cyber-physical systems in manufacturing. CIRP Annals 65(2):621–641, https:/ /doi.org/10.1016/j.cirp.2016.06.005 Nguyen T, Gosine RG, Warrian P (2020) A systematic review of big data analytics for oil and gas industry 4.0. IEEE Access 8:61,183–61,201, https://doi.org/10.1109/access.2020.2979678 Paez M, Tobitsch K (2017) The industrial internet of things: Risks, liabilities, and emerging legal issues. NYL Sch L Rev 62:217 Pajany M, Hemalatha A (2019) Pipeline gas leakage detection and location identification system. In: 2019 IEEE International Conference on System, Computation, Automation and Networking (ICSCAN), IEEE, https://doi.org/10.1109/icscan.2019.8878721 Patil S, Drozdov D, Vyatkin V (2018) Adapting software design patterns to develop reusable iec 61499 function block applications. pp 725–732, https://doi.org/10.1109/INDIN.2018.8472071 Qing CHG, Heripracoyo S (2019) Oil & gas industry mapping analysis with GIS approach: A literature review. In: 2019 7th International Conference on Cyber and IT Service Management (CITSM), IEEE, https://doi.org/10.1109/citsm47753.2019.8965340 Raza M, Aslam N, Le-Minh H, Hussain S, Cao Y, Khan N (2018) A critical analysis of research potential, challenges, and future directives in industrial wireless sensor networks. IEEE Communications Surveys and Tutorials 20(1):39–95, https://doi.org/10.1109/ COMST.2017.2759725 Roshani G, Nazemi E, Roshani M (2017a) Intelligent recognition of gas-oil-water three-phase flow regime and determination of volume fraction using radial basis function. Flow Measurement and Instrumentation 54:39–45, https://doi.org/10.1016/j.flowmeasinst.2016.10.001 Roshani G, Nazemi E, Roshani M (2017b) Intelligent recognition of gas-oil-water three-phase flow regime and determination of volume fraction using radial basis function. Flow Measurement and Instrumentation 54:39–45, https://doi.org/10.1016/j.flowmeasinst.2016.10.001 Sasloglou K, Tsao G, Glover IA, Gazis V, Frangiadakis N, Kikiras P (2013) Empirical channel model for placement optimisation of sensors deployed on oil & gas transmission pipelines. In: Proceedings of the 17th Panhellenic Conference on Informatics - PCI ’13, ACM Press, https:/ /doi.org/10.1145/2491845.2491886 Schleipen M, Gilani SS, Bischoff T, Pfrommer J (2016) Opc ua & industrie 4.0 - enabling technology with high diversity and variability. vol 57, pp 315–320, https://doi.org/10.1016/ j.procir.2016.11.055 Seilonen I, Vyatkin V, Atmojo U (2019) Opc ua information model and a wrapper for iec 61499 runtimes. vol 2019-July, pp 1008–1013, https://doi.org/10.1109/INDIN41052.2019.8971964 Shinkevich AI, Baygildin DR, Vodolazhskaya EL (2020) Management of a sustainable development of the oil and gas sector in the context of digitalization. Journal of Environmental Treatment Techniques 8(2):639–645 Sohraby K, Minoli D, Occhiogrosso B, Wang W (2018) A review of wireless and satellite-based m2m/iot services in support of smart grids. Mobile Networks and Applications 23(4):881–895, https://doi.org/10.1007/s11036-017-0955-1 Strasser T, Rooker M, Ebenhofer G, Zoitl A, Snder C, Valentini A, Martel A (2008) Framework for distributed industrial automation and control (4diac). pp 283–288, https://doi.org/10.1109/ INDIN.2008.4618110 TOMA C, POPA M (2018) IoT security approaches in oil & gas solution industry 4.0. Informatica Economica 22(3/2018):46–61, https://doi.org/10.12948/issn14531305/22.3.2018.05 Tyagi AK (2019) Building a smart and sustainable environment using internet of things. SSRN Electronic Journal https://doi.org/10.2139/ssrn.3356500 Vimos V, Sacoto Cabrera EJ (2018) Results of the implementation of a sensor network based on arduino devices and multiplatform applications using the standard opc ua. IEEE Latin America Transactions 16(9):2496–2502, https://doi.org/10.1109/TLA.2018.8789574

1082

C. A. Garcia et al.

Vyatkin V (2011a) Iec 61499 as enabler of distributed and intelligent automation: State-of-the-art review. IEEE Transactions on Industrial Informatics 7(4):768–781, https://doi.org/10.1109/ TII.2011.2166785 Vyatkin V (2011b) Iec 61499 as enabler of distributed and intelligent automation: State-of-the-art review. IEEE Transactions on Industrial Informatics 7(4):768–781, https://doi.org/10.1109/ TII.2011.2166785 Vyatkin V, ISA (2007) Iec 61499 function blocks for embedded and distributed control systems design Wanasinghe TR, Gosine RG, James LA, Mann GKI, de Silva O, Warrian PJ (2020a) The internet of things in the oil and gas industry: A systematic review. IEEE Internet of Things Journal 7(9):8654–8673, https://doi.org/10.1109/jiot.2020.2995617 Wanasinghe TR, Wroblewski L, Petersen BK, Gosine RG, James LA, Silva OD, Mann GKI, Warrian PJ (2020b) Digital twin for the oil and gas industry: Overview, research trends, opportunities, and challenges. IEEE Access 8:104,175–104,197, https://doi.org/10.1109/ access.2020.2998723 Wang S, Peng X, Lü Q, Long F, Jiang Y, Meng Y (2019) Design and implementation of the overall architecture of the puguang intelligent gas-field project. Natural Gas Industry B 6(3):262–271, https://doi.org/10.1016/j.ngib.2018.10.007 Wu C, Wei Y, Chu X, Tan Y (2017) Research and application of knowledge. In: Proceedings of the 2017 International Conference on Artificial Intelligence, Automation and Control Technologies - AIACT ’17, ACM Press, https://doi.org/10.1145/3080845.3080849 Yang Y, Li J, Tian M, Zhou Y, Dong L, xin He J (2019) Signal analysis of distributed optic-fiber sensing used for oil and gas pipeline monitoring. In: Proceedings of the 2019 International Symposium on Signal Processing Systems - SSPS 2019, ACM Press, https://doi.org/10.1145/ 3364908.3364918 You Z, Feng L (2020) Integration of industry 4.0 related technologies in construction industry: A framework of cyber-physical system. IEEE Access 8:122,908–122,922, https://doi.org/ 10.1109/ACCESS.2020.3007206 Zhao Q, Gao X, Wang H, Yang Y (2008) The application of oil and gas wells intelligent wireless monitoring system in oil field system. In: 2008 IEEE International Symposium on IT in Medicine and Education, IEEE, https://doi.org/10.1109/itme.2008.4743998 Zoitl A, Vyatkin V (2009) Iec 61499 architecture for distributed automation: The ‘glass half full’view. IEEE Industrial Electronics Magazine 3(4):7–23

Lower Extremity Exoskeleton Device for Motion Assistance and Gait Rehabilitation: Design Considerations

42

Jyotindra Narayan, Aditya Kalani, and Santosha K. Dwivedy

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Market Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preliminary Design of the Lower Extremity Exoskeleton Device . . . . . . . . . . . . . . . . . . . . . Modifications in the Preliminary Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Proof-of-Concept Design of Lower Extremity Exoskeleton Device . . . . . . . . . . . . . . . . . . . Comparative Cost Analysis: Preliminary Design Versus Proof-of-Concept Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1084 1088 1091 1095 1096 1096 1098 1099

Abstract

As population increases, there is a significant rise in cases of weak muscles and impaired nerves for elderly age group. In literature, several lower extremity exoskeleton devices (LEEDs) have been developed to assist the elderly people in the activities of daily living (ADL). However, due to design and cost constraints, users are unable to access these devices in many developing countries. Therefore, there is an emergent need to develop an affordable lower extremity exoskeleton device with cost-effective design features. This work presents the design factors associated with the development of the lower extremity exoskeleton device for motion assistance. At first, a market survey regarding awareness, type, functionality, safety factors, and cost of the exoskeleton device is carried out. A list of design aspects, required for the development of a LEED, is prepared according to the survey responses. Thereafter, an 8-DOF lower extremity exoskeleton is

J. Narayan () · A. Kalani · S. K. Dwivedy Department of Mechanical Engineering, Indian Institute of Technology, Guwahati, Assam, India e-mail: [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_25

1083

1084

J. Narayan et al.

primarily designed in SOLIDWORKS software for motion assistance and gait rehabilitation. However, after consulting with institute physiotherapy and manufacturing staffs, few design aspects, like material selection of links, reduction in transverse DOF, placement of hip, as well as knee actuator and support on wheeler stand, are modified for cost-effectiveness of the exoskeleton device and physiological safety of the user. At last, a final design of the 6-DOF exoskeleton device as a proof of concept is presented and compared with the preliminary design based on cost involved in various components. Keywords

Lower extremity exoskeleton devices · Design factors · Market survey · Motion assistance · Cost-effectiveness

Introduction Mobility is a critical requirement to perform the activities of daily living (ADL) in the life of an individual. However, impairment to the nervous system, stroke effects, and Parkinson’s disease can lead to mobility disorder in the upper extremities and/or lower extremities of the human body. Moreover, stroke survivors are always at risk to lose the ability to move the body limbs partially or completely. According to the Global Health Estimates, one of the primary sources of disability with the effect of impaired reasoning is stroke (World Health Organization 2019), and over 13 million cases of strokes per year have been reported in 2019 globally (Lindsay et al. 2019). Estimates of 2015 show that India has a total of approximately 1.8 million cases of stroke annually (Patnaik et al. 2015), with the number rising each year. Moreover, the cardiovascular diseases, leading to stroke, are also greatly induced by environmental factors like air pollution (Hussain 2019). When the loss of sensory functions gradually increases in the lower extremities of the body, majority of paraplegic-stroke patients are forced to use a wheelchair to restore their mobility. A wheelchair device provides limited mobility assistance where individuals sit and steer themselves to the desired destination. Many of the wheelchair variants comprise of two motors that actuate the rear wheels to provide forward motion and direction change. Front wheels provide support and move as required. However, wheelchair has a major functionality limitation where users have to maintain a sitting position for a substantial period of time. This advances the problem of muscle spasticity, chronic pain, and gradual muscle degeneration (Physiopedia 2019). Therefore, an urgent need has emerged to develop a cost-effective lower extremity device, which can help patients to regain their mobility and retrieve a healthy gait. Recent technological advancements in powered robotics exoskeletons can create powerful adjunctive tools for rehabilitation and potentially accelerate functional recovery. An exoskeleton, referred as a powered wearable mechanical device, has an anthropomorphic design architecture that enhances the functioning of the motor operations in the human body (Herr 2009). Lower limb exoskeleton

42 Lower Extremity Exoskeleton Device for Motion Assistance and Gait. . .

1085

Fig. 1 Significant developments in lower limb exoskeletons in the last 5 years

provides augmented support to the patients, assists them in walking, and helps them in their rehabilitation. These devices are augmented by making all the joints of the exoskeleton move in a trajectory similar to healthy human gait patterns. Over the last two decades, various lower extremity exoskeleton devices have been proposed by the researchers based on ergonomics design (Moreno et al. 2009; Chen et al. 2013; Han et al. 2018; Wang et al. 2020), compliant degrees of freedom (DOFs) (Kim et al. 2013; Lu et al. 2013; Bartenbach et al. 2016; Ouyang et al. 2016), and compact actuation (Zoss et al. 2006; Banala et al. 2007; Hyon et al. 2011; Asbeck et al. 2015; Ouyang et al. 2016; Chen et al. 2018) for the physiological safety of the user. Within the last 5 years, the significant developments in the field of lower limb exoskeletons are illustrated in Fig. 1. An ergonomic design offers the flexibility to the user by allowing unrestricted motion of the lower limb without any discomfort. Employing ergonomics features, the interactions between human and elements of exoskeleton device can be understood to optimize the performance of the overall system. The kinematics used in the design of a lower extremity exoskeleton or orthosis depends on the biomechanics of the human gait pattern. If the kinematics of the exoskeleton and the human body is not compliant with each other, then a non-ergonomics internal interaction force may arise due to misalignment. Moreno et al. (2009) presented a dual-stage analysis to estimate the interaction between the user and a lower limb exoskeleton. From the

1086

J. Narayan et al.

study, it was observed that there is a need for reduced mediolateral reaction forces for a post-polio patient to acquire a physiological gait. Chen et al. (2013) designed a wearable exoskeleton for the rehabilitation of the lower extremity derived from bionic design approaches, which consider the human anatomy and bone surgery. Carrying out a strategic design for crutches to support lower extremity exoskeleton devices, Han et al. (2018) performed an experiment with 30 subjects to identify the significant design improvements based on human-machine interaction (HMI). In a recent study by Wang et al. (2020), an exoskeleton device, AIDER, is utilized to confirm the ergonomic design for the user based on the interaction forces involved. The results proposed that there is a need for improvement in mechanical structure and control strategy for better ergonomics of the exoskeleton device. In another recent work on exoskeleton devices, Gupta et al. (2020) introduced a wheelchairbased sit-to-stand device for paraplegic people. Furthermore, following similar design principles, a child exoskeleton is designed by Narayan et al. (2020), and joint actuator torques are estimated using backpropagation neural networks for different heights of the children. Another crucial design objective of the exoskeleton device is to achieve the necessary degrees of freedom for varied applications. In addition to the motion assistance for paraplegic people, they are also utilized for the industry labor workers and soldiers to carry the additional loads. Considering lower limb joints of the human body (trunk, hip, knee, and ankle-foot), the lower extremity exoskeleton devices are primarily classified into multi-joint and single-joint exoskeletons (Kalita et al. 2020). The multi-joint exoskeletons are utilized to assist two or more joints simultaneously, whereas the single-joint exoskeletons are employed to provide assistance for one joint at a time. Kim et al. (2013) developed a 14-DOF lower limb exoskeleton to provide gait assistance, having a 3-DOF hip joint, a 1-DOF knee joint, and a 3-DOF ankle joint for each limb. A force sensor-based hardware system was installed to measure the center of pressure (CoP) and validated for a healthy person wearing the exoskeleton device. Lu et al. (2013) developed a 4-DOF novel lower limb exoskeleton, actuated at hip, knee, and ankle joints in the sagittal plane, for physical assistance and rehabilitation. The hip joint consists of 2-DOFs for flexion/extension (f/e) and abduction-adduction movement (b/d), while knee and ankle joint poses a 1-DOF movement each for flexion/extension and plantar/dorsiflexion, respectively. Bartenbach et al. (2016) introduced modular and reconfigurable design of the lower limb exoskeletons, which can be extended up to 4-DOFs per leg: hip (b/d, f/e), knee (f/e), and ankle (f/e). The device functionality was improved due to the adaptation of different users and a wide range of applications. In other works on multi-joint exoskeleton devices, Ouyang et al. (2016) proposed a hydraulic power unit for the rehabilitation of trunk-hip-kneeankle-foot (THKAF) while having the advantage of the high power-to-weight ratio. In the powered exoskeletons, three actuation methods, i.e., hydraulic, electric, and pneumatic, are conventionally used to move the lower limb joints. A compact hydraulic actuation-based exoskeleton unit, driven by 2.4 kW two-stroke internal combustion engine at 13000 rpm, was established to deliver 1.45 kW power

42 Lower Extremity Exoskeleton Device for Motion Assistance and Gait. . .

1087

(Ouyang et al. 2016). Another hydraulic-powered device, HUALEX (Chen et al. 2018), is recently designed with load-carrying features and zero load control approach for military services. However, hydraulic actuators are used seldom due to its heavy weight and high impedance features. Most of the prominent exoskeleton devices, viz., BLEEX (Zoss et al. 2006), ALEX (Banala et al. 2007), and Soft Exosuits (Asbeck et al. 2015), exploited the electric actuators due to their ease of availability and low maintenance cost. To address the problems of high metabolic costs associated with electric and hydraulic actuation system, the researchers introduced pneumatic artificial muscle actuators (PAMs) to run the exoskeleton devices (Hyon et al. 2011). Employing a hybrid actitation of PAMs with electric motors, XoR assistive device is developed for mobility assistance to the elderly people (Hyon et al. 2011). The electric actuator offers the dynamic compensation, while PAM works as a gravity balancer. Moreover, with the advent of nanoengineered products and components in electronic industry (Hussain 2018) as well as their commercialization in different sectors (Hussain 2020), the miniaturized form of printed circuits can be further used for more compact design of actuation modules in exoskeleton systems. Although several functionality aspects of lower extremity exoskeleton devices have been addressed in the past, there are still few design issues that need to be extensively enlisted and worked upon. These issues are as follows: adaptability for different age groups, selection as well as placement of joint actuators, material selection of device links, and affordability of the device. Moreover, in developing countries, awareness regarding availability and effectiveness of these devices should be widely spread to avoid implications during clinical procedures. Therefore, in this work, a market survey, having questionnaires related to awareness, type, functionality, safety factors, and cost of the exoskeleton devices, is conducted to enlist the design requirements of the lower extremity exoskeleton device. A preliminary design of 8-DOF lower extremity exoskeleton device is proposed based on the survey responses. To maintain the cost-effectiveness and physiological safety to the wearer, the device is modified for few design aspects, like reduction in transverse DOF, material selection of links, placement of hip as well as knee actuator, and support on wheeler stand. The modifications are carried out after the consultation with the institute therapists and manufacturing staffs. Finally, a proofof-concept design for the lower extremity exoskeleton device is presented along with a comparative analysis of the involved expenditures in both models. The organization of the rest of the chapter is as follows. Section “Market Survey” presents the market survey and responses followed by the necessary design specifications. In section “Preliminary Design of the Lower Extremity Exoskeleton Device,” a preliminary design of the lower extremity exoskeleton is discussed. Section “Modifications in the Preliminary Design” elaborates the possible modifications in the existing design. A proof-of-concept design for the exoskeleton is presented in section “Proof-of-Concept Design of Lower Extremity Exoskeleton Device” along with the expenditure details. Section “Conclusions” concludes the work of the chapter.

1088

J. Narayan et al.

Market Survey In developing countries, the healthcare industry is lagging behind because of the less adaptability to the requirements of diverse population. These countries rely heavily on the imports of medical devices, eventually making such devices expensive and accessible to select few. Therefore, to understand the market dynamics of a potential exoskeleton usage, a survey was conducted in the physiotherapy section of an institute hospital. A total of ten questions were prepared based on age group, mobility dysfunction, affected extremity, affected extremity segment(s), gait correction, external assistance, experience with wheelchair, awareness about exoskeletons, design preferences in exoskeleton devices, and expected cost involved. In the hospital, 10–16 incomings/day (Mean, 13; SD: ±3) were recorded having the problem of nerve impairment anywhere in the body. The survey lasted 10 days, leading to 130 responses from the patients. However, around 30 incomings were not interested to answer all the questions and left the survey mid-way. With due consent to the patients, the questions (Fig. 2) were asked to explore the distinct opinions of nerve-impaired patients for better understanding of design characteristics of exoskeleton. The technical keywords, like “mobility dysfunction,” “extremity,” “gait,” “robotic exoskeletons,” “rehabilitation,” “ADLs,” “motion assistance,” etc., were explained while asking the questions. The first question, as shown in Fig. 2a, is asked to understand the age group of different individuals. It is found from the response that 96% patients were above 50 years of age and 4% of patients were from 30 years to 50 years of age group, who require assistance from the physiotherapy section. However, out of 100 patients, 88 patients were suffering from mobility dysfunction (Fig. 2b), in which 67 patients were facing movement imbalance in the lower extremity of the body (Fig. 2c). As the age increases, mobility impairments are quite prevalent in the lower limb due to decline in muscle strength and functional strength. Out of 67 patients, 45.8% patients were having mobility issues in both legs, while 37.5% patients were facing disorder in one leg only, as shown in Fig. 2d. Around 12.5% patients either were not aware about the source of gait disorder or did not wish to disclose the same. After excluding the 12.5% responses, it is evident from Fig. 2e that 76% individuals have not retrieved the original gait by means of manual therapy though recovered from mobility impairment. Moreover, the responses shown in Fig. 2f indicate that 59% of patients (out of 59) received some form of assistance from another person even after their impairment recovery. Regarding the comfort level with conventional methods, 73.9% patients were disappointed when using wheelchairs, sticks, crutches, etc., as shown in Fig. 2g. Moving on to the next question, shown in Fig. 2h, about patient awareness of robotic exoskeletons, it was worth noting that 96% of 59 patients had no idea of such devices for motion assistance and gait rehabilitation. Thereafter, with the brief introduction of exoskeleton, 59 patients were asked for their preferences in the design of such robotic devices to provide complete assistance in walking and to offer rehabilitation. Nearly all patients had marked for affordability, ease of maintenance, and lightweight features for the design of exoskeleton devices as

42 Lower Extremity Exoskeleton Device for Motion Assistance and Gait. . .

Fig. 2 (continued)

1089

1090

J. Narayan et al.

Fig. 2 Questions asked and responses received in the market survey

shown in Fig. 2i. Another crucial design feature responded was adaptability of the device to different user sizes. Most of the patients had pointed out that the human body exhibits a wide range of size differences in both skeletal bone lengths and limb and torso girth; therefore, a single exoskeleton unit must be adaptable to different users of the same family. As shown in Fig. 2j, the last question reflects the upper bound of the purchasing capability of the patients. Almost 57% patients were not ready to purchase the product worth 195,000 INR, and 13% patients were not sure whether the device will be cost-effective or not. However, the positive responses from 30% patients reflected that there is still a considerable market scope for exoskeleton devices. Although several Indian start-ups have emerged with robotic solutions, there is still far less awareness about exoskeleton devices and their benefits for motion assistance and rehabilitation. There is a dire need for effective marketing or advertising campaigns so that more and more people get to know about this. Based on the survey responses, the following design specifications should be considered while designing a lower extremity exoskeleton device for motion assistance and rehabilitation: • Design should be ergonomic so that normal motion could not be restricted and provide comfort to the user. • The exoskeleton device must be anthropometric, affordable, and lightweight for the user. The weight of the device can be reduced by employing rated actuators less than the required and generating certain portion of power via sticks or crutches or wheeler stands. • The design should be customizable according to the specific size and weight requirements of different age groups. As adaptability of the device will be

42 Lower Extremity Exoskeleton Device for Motion Assistance and Gait. . .

1091

increased, the number of potential customers could be increased, especially in developing countries. • The main focus of exoskeleton device is to achieve the freedom of movement for the impaired by providing additional locomotion strength. The purpose of the rehabilitation trainings should be implicitly fulfilled while performing ADLs. • The design should reflect the cost-effectiveness of the lower extremity exoskeleton device by keeping the purchasing and maintenance costs within 195,000 INR.

Preliminary Design of the Lower Extremity Exoskeleton Device Designing a lower extremity exoskeleton device, the central approach should be focused to ensure the strength and stability during transmission of the required force vectors through chassis, frames, and joints. However, maintaining the strength of the device as well as keeping the mechanical and electric components lightweight is always a matter of trade-off between these two design features. For a specified set of motions, selection and positioning of the joint actuators is another crucial design aspect to be considered in accordance with physiological comfort to the user. Furthermore, selection of degrees of freedom for each lower extremity joint also plays a pivotal role in the design of the lower extremity exoskeleton devices. Although the user might experience more comfort with increased DOFs, providing all possible DOFs at the initial stage of assistance and rehabilitation may lead to uncontrolled movements. In general, as shown in Fig. 3, each human leg comprises of a 7-DOF structure, with three rotational DOFs at the hip (abduction/adduction (a/a), flexion/extension (f/e), and hip intra/extra rotation (i/e)), one at the knee (flexion/extension (f/e)), and three at the ankle (abduction/adduction (a/a), dorsi/planter flexion (d/p), and ankle inversion/eversion (i/e)). However, to avoid the involuntary motions, 4-DOF movements are considered in this work, i.e., hip abduction/adduction (a/a), hip flexion/extension (f/e), knee flexion/extension (f/e), and ankle dorsi/planter flexion (d/p). Employing the above design aspects in SOLIDWORKS software, the 8-DOF (six active, two passive) LEED is modeled with waist link, thigh links, calf links, and foot links as shown in Fig. 4. The proposed device is designed for the average weight and average height ranging between 55 kg and 90 kg and 152 cm and 182 cm, respectively. The main objective of the device is to assist the individuals suffering from weak muscles and impaired nerves. The allowable range of motion (ROM) for each lower extremity joint in the device is shown in Table 1. In the CAD modeling of lower extremity exoskeleton device, telescopic links of cast alloy steel are designed to accommodate people of different heights and body sizes as shown in Fig. 5a, b. The waist link varies between 180 mm and 220 mm, the thigh link varies between 330 mm and 470 mm, the calf link varies between 345 mm and 480 mm, and the foot link varies between 180 mm and 260 mm based on the body ratios of a healthy human (Gupta et al. 2020). The actuation modules are considered after fixing the weight and height of the human body at the upper bound. To maintain a fine trade-off among weight, rated torque, and cost involved

1092

J. Narayan et al.

Fig. 3 Different joint movements of the lower limb (a) hip abduction/adduction (a/a), intra/extra rotation (i/e), and flexion/extension (f/e), (b) knee flexion/extension (f/e), (c) ankle dorsi/planter flexion (d/p), abduction/adduction (a/a), and ankle inversion/eversion (i/e) (Kalita et al. 2020)

Fig. 4 SOLIDWORKS model of preliminary LEED in (a) isometric view and (b) front view

in the hip joint actuator, a DC stepper motor could be a better option. A comparison analysis for the performance aspects of stepper motor, servomotor, and harmonic drive is presented in Table 2. The hip joint’s flexion and extension motion is driven by belt and pulley mechanism as shown in Fig. 6a. There are three pulleys, i.e., driver, driven, and idler, connected by a belt. By varying the sizes of the driver and the driven pulley, the

42 Lower Extremity Exoskeleton Device for Motion Assistance and Gait. . .

1093

Table 1 Range of motion (ROM) for different lower extremity joints Lower limb joint Hip joint Knee joint Ankle joint

Motion Flexion/extension Abduction/adduction Flexion/extension Dorsiflexion/plantar flexion

Range of motion (ROM, degrees) 44/−22 12/−12 80/−5 15/−25

Fig. 5 Design of telescopic links for (a) thigh and calf segments and (b) waist segment

Table 2 Performance aspects of stepper motor, servomotor, and harmonic drive Criteria Accuracy Cost Weight

Torque

DC stepper motor Accurate Less costly than harmonic drive Less bulky with geared setup. Not compact in shape Easily available for high torques

Servomotor Less accurate More costly than stepper motor More bulky than stepper for high torque requirement Easily available for low toques

Harmonic drive Most accurate Very costly (2 lakhs INR/joint) Lightweight and very compact in shape Readily available for high torques

input torque of the DC stepper motor can be reduced. The motor drives the smaller pulley, which eventually rotates the driven pulley (larger) connected to the hip joint by keeping the transmission ratio 1.85. The design parameters of timing belt-pulley drive module, shown in Table 3, are selected or calculated from the catalogue of Cross and Morse Power Transmission Solutions (2020). The material of the pulley is stainless steel, while the timing belt is considered to be made of high hardness synthetic rubber with glass core wire and nylon cloth. The third pulley is exploited as an idler to maintain the tension in the belt. The size of the idler pulley is kept equal to the size of the driver pulley connected to the motor drive. Moreover, the idler is positioned in such a way that the minimum arc of contact would be maintained. The weight of complete actuation module for driving a hip joint, which includes motor drive, driver pulley, idler pulley, driven pulley, and timing belt, is estimated to be 16 kg.

1094

J. Narayan et al.

Fig. 6 Design of actuation cum transmission mechanism (a) hip f/e, (b) hip b/d, and (c) knee f/e Table 3 Design parameters of timing belt drive for hip f/e Design parameters Transmission ratio Maximum output torque Maximum input torque Output angular velocity Input angular velocity Driver pulley pitch diameter Driven pulley pitch diameter Driver pulley teeth Driven pulley teeth Centre distance between the driver and the driven pulley Timing belt width Belt pitch

Design values 1:1.85 120.25 nm 65 nm 10 rev/min 40 rev/min 60 mm 122.1 mm 26 48 225 mm 50 mm 8 mm

During the gait initiation, a mediolateral instability arises due to the difference between the center of mass (CoM) and center of pressure (CoP) for the human body. To minimize this instability when one lifts the leg, CoM and CoP should try to move closely with each other for the complete human-device arrangement. Therefore, abduction/adduction movement is required to achieve the balanced state in transverse plane while walking. In the exoskeleton design, this transverse motion of hip joint of both legs is driven and altered with the help of a single motor drive. The mechanism involves three pinions and a rack, having a total weight of 8 kg, as shown in Fig. 6b. The pinion at the center is driven by a motor which drives the rack. The remaining two pinions, connected to the both hip joint, are driven by the rack to provide b/d motion. The specifications of actuation design for the hip b/d is presented in Table 4. For the f/e motion of the knee joint, a lead screw-based linear actuator with 300–350 mm stroke length for 40 Nm rotational torque is utilized as shown in Fig. 6c. One end of the actuator is placed at the thigh link and other end at the calf link by maintaining a ratio of 2:1 with respect to the knee joint. For d/p movement of ankle, a passive joint is designed, having a torsional spring of stiffness of 6.5 Nm/0 for the upper limit of foot length. Combining the weights of links and actuation mechanisms, the total weight of the lower extremity exoskeleton device is nearly 44 kg.

42 Lower Extremity Exoskeleton Device for Motion Assistance and Gait. . . Table 4 Design parameters of timing belt drive for hip b/d

Design parameters Transmission ratio Maximum torque Angular velocity Pinion pitch diameter Pinion teeth Rack width Pitch

1095 Design values 1:1 72 nm 5 rev/min 66 mm 32 70 mm 12 mm

Fig. 7 Design of (a) aluminum slider rails, (b) adjacent placement of hip joint motor, and (c) linear actuator as an inside arrangement

Modifications in the Preliminary Design After completion of the preliminary design, suggestions are asked from the therapists of institute hospitals and skilled staffs from the manufacturing workshops. Following their overall recommendations, few issues are found and need to be modified in the final design. The number of control variables can be reduced by eliminating hip abduction/adduction movement. This movement in the exoskeleton devices, in general, is not necessary for partial motion assistance and gait rehabilitation. To establish a balanced relation among less deformation, minimal weight, and high strength, aluminum-based slider rails are found to be a better alternative for the cast alloy steel-based telescopic links. In the modified design, three slider rails, as shown in Fig. 7a, are considered for every thigh and calf link. As shown in Fig. 7b, the adjacent positioning of hip joint actuator for f/e motion instead of back side is found to be a more cost-effective alternative. This kind of actuator placement eliminates the need for transmission mechanism, which overall reduces the weight of the device by 18 kg. Similarly, the lead screw linear actuator at the outside of the thigh and calf links could be replaced with ball screw linear actuator appended at the cuffs from inside for smooth linear motion. The modified arrangement of linear actuators at knee joint is shown in Fig. 7c.

1096

J. Narayan et al.

Although a direct drive for hip joint actuators reduces the overall weight of the device as compared to timing belt-pulley-based driven mechanism, the load applied at the hip joint is still high. Therefore, there is a need for wheeler-based stand module to place the hip actuators on both sides. This provision also fulfills the requirement of sticks/crutches support responded by the patients in the market survey. Moreover, considering the responses received in Fig. 3d, the usability of the device can be more improved by removing the waist link and designing two separate arrangements for two lower limbs, each having 3-DOFs in the hip, knee, and ankle joint.

Proof-of-Concept Design of Lower Extremity Exoskeleton Device After carrying out the modifications in the preliminary design, following costeffective design features of a lower extremity exoskeleton device are confirmed. The proof-of-concept design for the 6-DOF LEED is shown in Fig. 8a. • Three degrees of freedom are considered in one extremity of the exoskeleton device for hip f/e, knee f/e, and ankle d/p. • Aluminum-based slider rails, readily available in the market, are employed for changing thigh and calf links in case of different heights of the users. A locking plate is provided to allow the user to adjust the leg length. • Stepper motors are utilized to directly drive the hip joints without any provision of transmission mechanism. • Ball screw linear actuators are placed via cuffs from inside of the extremities to avoid any obstruction while altering the slider rails for different heights. • Ankle movements are kept passive by using two torsional springs at both joints. • A wheeler stand has been added to support the weight of the stepper motors, which drive the hip joint. Moreover, the stand module satisfies the requirement of sticks/crutches. • Another variant of the modified design consists of two separate legs, which are connected with the stand. This design variant, as shown in Fig. 8b, is consistent with the survey fact to assist both legs of individuals simultaneously as well as separately.

Comparative Cost Analysis: Preliminary Design Versus Proof-of-Concept Design For any robotic device, increasing the affordability feature often helps to achieve greater market outreach and longevity in the industry. Furthermore, emerging instances of muscle weakness and nerve impairments are expected to increase the application of exoskeleton technology in the near future. However, there is limited

42 Lower Extremity Exoskeleton Device for Motion Assistance and Gait. . .

1097

Fig. 8 Proof-of-concept design of LEED with (a) wheeler stand and (b) separate limb arrangements

adoption of this technology in many developing countries due to low per capita income and availability of cheaper short-term alternatives. Therefore, to understand the manufacturing and assembly cost involved, a comparative analysis for both design models of lower extremity exoskeleton device is presented briefly as shown in Table 5. In the preliminary design, four alloy steel telescopic links (MISUMI India Private Limited 2020) are chosen for the development of thigh and calf segments, which costs a total of 24,225 INR (6056 INR each). In the modified design, the strength of three aluminum slider rails from the same company is equivalent to one alloy steel telescopic link, which costs 9318 INR. Therefore, the total amount for four lower segments is 37,272 INR, with taxes excluded. Moving further, as per the motor torque requirement of 65 Nm for hip joint f/e in the preliminary design, a worm-geared stepper motor (Bholanath Precision Engineering Private Limited 2020) is selected, which costs 32,254 INR including motor driver. The total cost for two such motors is 64,508 INR, inclusive of motor drivers. Moreover, the belt-pulley transmission mechanism (Cross and Morse Power Transmission Solutions 2020) costs 42,888 INR for both hip joints. In modified design, a worm-geared stepper motor of capacity 120 Nm costs 40,926 INR, including motor driver, and the total cost for two such motors along with drivers is 81,852 INR. However, as there is no transmission mechanism in the proof-ofconcept design, the respective expenditure is saved. For hip b/d motion in the preliminary design, a motor drive and rack-pinion arrangement costs of 22,114 INR and 16,442 INR, respectively, lead to a total of 38,556 INR. However, this expenditure is saved in the final design due to the elimination of transverse degrees

1098

J. Narayan et al.

Table 5 Comparative cost analysis for preliminary and proof-of-concept design Component Telescopic links/slider rails (thigh and calf segments) Stepper motor (hip joints f/e) Transmission module (hip joints f/e) Stepper motor (hip joints b/d) Transmission module (hip joints b/d) Linear actuator (knee joints f/e) Miscellaneous costs

Total cost

Company name MISUMI India Private Limited

Bholanath Precision Eng. Pvt. Limited Cross and Morse Power Transmission Solutions Bholanath Precision Eng. Pvt. Limited Cross and Morse Power Transmission Solutions Bholanath Precision Eng. Pvt. Limited Assam Steel and Bholanath Precision Eng. Pvt. Limited

Preliminary design cost (INR) 24,225

Proof-of-concept design cost (INR) 37,272

64,508

81,852

42,888

1000

22,114

0

16,442

0

32,820

36,910

10,000

15,000

212,997

171,034

of freedom in the hip joint. The above amount of motor, motor drivers, belt-pulley mechanism, and rack-pinion arrangements is presented without including the taxes. The costs involved in lead screw and ball screw linear actuator in respective design models (Bholanath Precision Engineering Private Limited 2020) are 32,820 INR and 36,910 INR, excluding taxes. In the preliminary design, the miscellaneous cost for bearings, torsional springs, hubs, nut-bolts, cuffs, lock plates, etc. is roughly 10,000 INR. However, due to the presence of wheeler stand in the final design, the miscellaneous costs approximately 15,000 INR. The tubes, plates, and connectors in both design models are fabricated from steel material (Assam Steel 2020) or chosen (Bholanath Precision Engineering Private Limited 2020). After estimating all costs incurred in both design models, the total cost for the final design is observed to be less than 195,000 INR, being preferable for the users as responded in the last question of the market survey.

Conclusions Several lower extremity exoskeleton devices have been designed in the literature for motion assistance and gait rehabilitation in case of weak muscles and impaired nerves. However, users from many developing countries neither are aware about the exoskeleton technology nor able to afford it even if they are aware of it. Therefore, in this work, a market survey has been conducted in the therapy section of an institute hospital to create awareness about exoskeleton and to ask users of their preferences

42 Lower Extremity Exoskeleton Device for Motion Assistance and Gait. . .

1099

for cost-effective design features of the exoskeleton devices. A set of ten questions have been asked to understand the design features, functionality, and affordability of such devices. Based on the responses received, an 8-DOF lower extremity device has been designed with complete actuation modules. Thereafter, few design modifications, viz., reduction in transverse DOF of hip joint, placement of hip as well as knee actuator, link materials, and addition of wheeler stand, have been carried out after taking the suggestions from institute therapists and manufacturing staffs. At last, a 6-DOF proof-of-concept device has been presented to assist the patients in pathological gaits. A comparative cost analysis between preliminary design and final design has also been presented to realize the cost-effectiveness of the device. This work will serve enthusiastic researchers and therapists to understand the cost-effective features of the LEEDs during the design process. Future work could be explored for dynamics and robust control required for gait rehabilitation. Acknowledgments The authors acknowledge the Department of Scientific and Industrial Research, India, for starting the initiative PRISM (Promoting Innovations in Individuals, Startups and MSMEs), under which this project is carried out. The authors are grateful to the amiable support of medical therapist Mr. Kandarpa Jyoti Das, IIT Guwahati, in performing the research experiments.

References Asbeck AT, De Rossi SM, Holt KG, Walsh CJ (2015) A biologically inspired soft exosuit for walking assistance. Int J Robotics Res 34(6):744–762 Assam Steel (2020) Products and services. https:/978-3-030-84205-5/www.tradeindia.com/Seller2629320-ASSAM-STEEL/product-services.html. Accessed 17 May 2020 Banala SK, Agrawal SK, Scholz JP (2007) Active Leg Exoskeleton (ALEX) for gait rehabilitation of motor-impaired patients. In: 2007 IEEE 10th international conference on rehabilitation robotics. IEEE, pp 401–407 Bartenbach V, Gort M, Riener R (2016) Concept and design of a modular lower limb exoskeleton. In: 2016 6th IEEE international conference on biomedical robotics and biomechatronics (BioRob). IEEE, pp 649–654 Bholanath Precision Engineering Private Limited (2020) Stepper motors. https:// steppermotors.bholanath.in/. Accessed 17 May 2020 Chen C, Zheng D, Peng A, Wang C, Wu X (2013) Flexible design of a wearable lower limb exoskeleton robot. In: 2013 IEEE international conference on robotics and biomimetics (ROBIO). IEEE, pp 209–214 Chen Q, Cheng H, Shen W, Huang R, Chen X (2018) Hybrid control for human-powered augmentation exoskeleton. In: 2018 IEEE 8th annual international conference on CYBER technology in automation, control, and intelligent systems (CYBER). IEEE, pp 682–687 Cross and Morse Transmission Power Solutions (2020) Timing belt drives. https:// www.crossmorse.com/en/our-products/timing-belt-drives. Accessed 17 May 2020 Gupta M, Narayan J, Dwivedy SK (2020) Modeling of a novel lower limb exoskeleton system for paraplegic patients. In: Maity D, Siddheshwar P, Saha S (eds) Advances in fluid mechanics and solid mechanics. Lecture notes in mechanical engineering. Springer, Singapore Han J, Jin Hyun D, Jung K, Yoon Kim K, Youn S (2018) Ergonomic design strategy for crutches of a lower-limb exoskeleton for paraplegic individuals: an experimental study. In: Proceedings of the human factors and ergonomics society annual meeting, vol 62, no 1. Sage, Los Angeles, pp 1012–1016

1100

J. Narayan et al.

Herr H (2009) Exoskeletons and orthoses: classification, design challenges and future directions. J Neuroeng Rehabil 6(1):21 Hussain CM (ed) (2018) Handbook of nanomaterials for industrial applications. Elsevier Hussain CM (ed) (2019) Handbook of environmental materials management. Springer, Cham Hussain CM (ed) (2020) The ELSI handbook of nanotechnology: risk, safety, ELSI and commercialization. Wiley Hyon SH, Morimoto J, Matsubara T, Noda T, Kawato M (2011, September) XoR: hybrid drive exoskeleton robot that can balance. In: 2011 IEEE/RSJ international conference on intelligent robots and systems. IEEE, pp 3975–3981 Kalita B, Narayan J, Dwivedy SK (2020) Development of active lower limb robotic-based orthosis and exoskeleton devices: a systematic review. Int J Soc Robot 13:775–793 Kim JH, Han JW, Kim DY, Baek YS (2013) Design of a walking assistance lower limb exoskeleton for paraplegic patients and hardware validation using CoP. Int J Adv Robot Syst 10(2):113 Lindsay MP, Norrving B, Sacco RL, Brainin M, Hacke W, Martins S, Pandian J, Feigin V (2019) World Stroke Organization (WSO): global stroke fact sheet 2019. Int J Stroke 14(8):806–817 Lu R, Li Z, Su CY, Xue A (2013) Development and learning control of a human limb with a rehabilitation exoskeleton. IEEE Trans Ind Electron 61(7):3776–3785 MiSUMi India Private Limited (2020) Automation components. https://in.misumi-ec.com/vona2/ mech/. Accessed 17 May 2020 Moreno JC, Brunetti F, Navarro E, Forner-Cordero A, Pons JL (2009) Analysis of the human interaction with a wearable lower-limb exoskeleton. Appl Bionics Biomech 6(2):245–256 Narayan J, Jhunjhunwala S, Gupta M, Dwivedy SK (2020) Backpropagation neural network based design of a novel sit-to-stand exoskeleton at seat-off position for paraplegic children. In: 2020 6th international conference on control, automation and robotics (ICCAR). IEEE, pp 546–552 Ouyang X, Ding S, Fan B, Li PY, Yang H (2016) Development of a novel compact hydraulic power unit for the exoskeleton robot. Mechatronics 38:68–75 Patnaik L, Sahoo HS, Sahu T (2015) Awareness of the warning symptoms and risk factors of stroke among adults seeking health care from a rural hospital of India. Ann Indian Acad Neurol 18(4):487 Physiopedia (2019) Wheelchair users. https://www.physio-pedia.com/Wheelchair_Users. Accessed 13 July 2020 Wang Y, Qiu J, Cheng H, Zheng X (2020) Analysis of human–exoskeleton system interaction for ergonomic design. Hum Factors. https://doi.org/10.1177/0018720820913789 World Health Organization (2019) Global Health Estimates: Life expectancy and leading causes of death and disability. https://www.who.int/data/gho/data/themes/mortality-and-global-healthestimates. Accessed 13 August 2021 Zoss AB, Kazerooni H, Chu A (2006) Biomechanical design of the Berkeley lower extremity exoskeleton (BLEEX). IEEE/ASME Trans Mechatron 11(2):128–138

Fuzzy Membership Functions in ANFIS for Kinematic Modeling of 3R Manipulator

43

Jyotindra Narayan, Sashwata Banerjee, Durgarao Kamireddy, and Santosha K. Dwivedy

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ANFIS Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Triangular Membership Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Trapezoidal Membership Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gaussian Membership Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Generalized Bell Membership Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sigmoidal Membership Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kinematic Modeling of 3-DOFs Robotic Manipulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of ANFIS for 3-DOFs Robotic Manipulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparative Study of Fuzzy Membership Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1102 1106 1107 1107 1108 1108 1108 1109 1112 1114 1117 1118

Abstract

In this work, a soft computing technique named adaptive neuro-fuzzy inference system (ANFIS) is employed for kinematic modeling of 3R robotic manipulator. Solving the forward kinematic problem using Denavit-Hartenberg (DH) parameters is straightforward; however, obtaining inverse kinematic results for higher degrees of freedom manipulator is computationally expensive. Therefore, a combined form of artificial neural networks (ANN) and fuzzy logic (FL),

J. Narayan () · D. Kamireddy · S. K. Dwivedy Department of Mechanical Engineering, Indian Institute of Technology, Guwahati, Assam, India e-mail: [email protected]; [email protected] S. Banerjee Department of Electrical Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_26

1101

1102

J. Narayan et al.

ANFIS technique, is exploited to show its worth for the inverse kinematic problem by considering a 3-DOFs robotic manipulator having three revolute joints. Moreover, ANFIS model is considered with five membership functions (MFs), viz., triangular, trapezoidal, Gaussian, generalized bell, and sigmoidal. Thereafter, manipulator’s joint variables are estimated for a semicircular trajectory for every membership function. The joint variables are compared with analytical inverse kinematic solutions to analyze the response of different MFs. From the simulation study, it is realized that the most promising order of selecting MFs for kinematic analysis of the robotic manipulator is Gaussian, generalized bell, sigmoidal, trapezoidal, and triangular in ANFIS training model. Finally, a qualitative justification of selecting the MFs in different literature works is extensively reported. Keywords

ANFIS · Kinematic analysis · Robotic manipulator · Membership functions

Introduction From the last three decades, soft computing has significantly evolved to address the complex mathematical models using different approximation approaches. Soft computing deals with the limited truth and uncertainty of the system and provides cost-efficient solutions. Unlike conventional methods, it is found to be more promising to solve complex problems using intelligent control, optimization, and decision-based support strategies (Jang et al. 1997). In a study by Zadeh (1994), soft computing is considered as amalgamation of various methods like fuzzy logic, artificial neural networks (ANNs), genetic algorithms, etc. In the literature till date, soft computing technique is utilized in numerous software- and hardwarebased applications like signal processing (Cichocki et al. 1993), automated voice identification (Komori et al. 1993), manufacturing and automation (Roy et al. 2012), diseases control (Ulieru et al. 2006), image segmentation (Senthilkumaran and Rajesh 2009), environmental engineering (Yetilmezsoy 2019), and robot trajectory tracking (Narayan et al. 2018; Abbas et al. 2019). Takagi and Sugeno (1993) explained the interpretation of system’s stability using dynamic model-based defuzzification process. The system identification is performed for two industrial applications, that is, water cleaning and steel-making procedures under defined constraints. The classical model of fuzzy logic (Zadeh 1994) was proposed to characterize and simplify the approximate human knowledge. Fuzzy logic is a well-known approach to represent the human understanding levels using IF and THEN rules based linguistic designs. Bai et al. (2007) presented the fundamentals of fuzzy logic in image processing, medical engineering, and control strategies for different industrial processes. Few years later, Nachtegael et al. (2013) edited a book on mathematical concepts of fuzzy filters for noise reduction, edge detection, and image enhancement. They have also presented few applications

43 Fuzzy Membership Functions in ANFIS for Kinematic Modeling of. . .

1103

of fuzzy filters in digital image processing (DIP) and lossy image reconstruction. In the field of industrial image processing, the concept of fuzzy logic has been exploited to detect the defects of industrial products using X-ray images (Amza and Cicic 2015). In a different work by Goyal et al. (2016), the formation of fuzzy models was discussed using subtractive and fuzzy C-Means clustering methods. Recently, Ali et al. (2021) proposed a fuzzy-based classifier to estimate the bread quality where fuzzy membership vectors are used to select optimal weights in relevance vector machine (RVM). Artificial Neural Network, another knowledge-based model, is driven in accordance with the biological nervous system and the brain. ANN constitutes single or multilayer neurons in an integrated way to solve the nonlinear relationship between input and output datasets. This technique has several advantages like adaptability, better management, and fault tolerance in real-time scenarios. However, the accuracy of the predicted results depends on the available input datasets to a great extent. The prediction of output is done using either supervised or unsupervised learning algorithms (Hoffmann 2003). The most common learning algorithm is the Backpropagation algorithm to train the input dataset in neural network architecture (Goh 1995). ANNs are generally exploited for prediction- and classification-based applications like error prediction and minimization (Boné and Crucianu 2002), pattern or data recognition (Ripley 2007), image processing (Zhao et al. 2016), and solar radiation estimation (Singh and Mittal 2020). The implementation of ANN model is reported by Mouss et al. (2020) to predict the accretion of crack density and corresponding length in cancellous bone. Thereafter, they conducted a parametric study to investigate the effects of load level on crack accretion using the ANN model. Recently, Narayan and Dwivedy (2021) proposed a Bayesian regularized BPNN model to estimate the biomechanical movements of lower-limb during healthy walking. The designed model is built upon the kinematic, spatiotemporal, and biological parameters of the healthy subjects. Furthermore, Jang (1993) proposed a significant combination of Artificial Neural Networks and Fuzzy Logic technique, that is, adaptive-network-based fuzzy inference system (ANFIS). In the study, ANFIS technique is presented using hybrid learning approach of IF-THEN rules and specified input-output datasets. This technique has been utilized in various fields of engineering to predict the desired output. The recent and significant developments in the application of ANFIS from last 5 years are shown in Fig. 1. In a study by Lei et al. (2008), condition monitoring of rotating machinery is done using ANFIS and GA techniques. Maher et al. (2015) exploited the ANFIS model to predict the surface properties in wireEDM process. Over last few years, ANFIS is combined with more nature-inspired optimization algorithms such as cultural, bees, invasive weed, particle swarm, and firefly algorithms (Tien Bui et al. 2018; Hasanipanah et al. 2018; Yaseen et al. 2018). Tien Bui et al. (2018) proposed ANFIS models for flood susceptibility mapping where the parameters of membership function are optimized using cultural-, bees-, and invasive weed algorithms. The ANFIS along with particle swarm optimization (PSO) approach is presented to estimate the fragmentation of rock during mine blasting (Hasanipanah et al. 2018). A similar kind of model is used to predict

1104

J. Narayan et al.

Fig. 1 Recent developments in the application of ANFIS

the electricity prices using hybrid form of mutual information (Gahrooei and Hooshmand 2019). Yaseen et al. (2018) presented a novel freefly algorithm (FFA)optimized ANFIS model to predict the rainfall pattern from the Pahang river. They further compared the proposed model with the standard ANFIS model and found the proposed to be effective. In a work on predicting leakage locations throughout the complicated water distribution facility, Yalçın et al. (2018) proposed an ANFIS approach as a reliable and affordable solution where the training dataset is formed using sensorized information of acceleration, pressure, and flow rate at specific points in the distribution facility. Moreover, literature shows many applications of ANFIS technique in the domain of robotics (Alavandar and Nigam 2008; Manjaree et al. 2015; Narayan and Singla 2017; Narayan et al. 2018; Chawla and Singla 2019). Robots play a significant role for many medical- and industry-based applications (Kim et al. 1987). The position of robot’s end effector in medical surgery is quite important for better accuracy and precision (Narayan et al. 2018). The identification and study for a fragment of a protein in the metabolism of human being is carried out using robot kinematics (Chen et al. 2020). Although, finding forward kinematics of any robotic system using D-H parameters is a simple task; however, estimating the inverse kinematic solutions is computationally expensive. Moreover, the most common available techniques (geometric, algebraic, and numerical) are not enough efficient due to the incapability of forming closed form solutions, undesirable solution convergence based on starting point and nonfunctionality

43 Fuzzy Membership Functions in ANFIS for Kinematic Modeling of. . .

1105

within the singularity zone (Schilling 1996). Therefore, researchers have started to explore the ANFIS technique to solve the inverse kinematics problem (Duka 2015; Manjaree et al. 2015; Narayan and Singla 2017; Narayan et al. 2018; Vu et al. 2018; Abbas et al. 2019). For a 3-DOF planar manipulator, Duka (2015) proposed an ANFIS-based inverse kinematic solution with bell membership function. The simulation results were verified by planning of circular, square, and triangle-shaped trajectory. However, the quantitative discussion on the deviation of robot’s end effector from the desired trajectory was not carried out. In a work by Manjaree et al. (2015), an inverse kinematic problem is solved for 5-DOFs robotic arm using ANFIS and compared with analytical approach. Moreover, the solutions are validated with experimental setup for circular trajectory. They have exploited Gaussian and bell MFs in the ANFIS architecture for their study. In Narayan and Singla (2017), ANFIS approach is used to solve the inverse kinematic problem for 4-DOFs selective compliance articulated robot arm (SCARA). The results are validated with the analytical solutions and found to be promising. One year later, Narayan et al. (2018) presented the ANFIS model with generalized bell MF to estimate the inverse kinematic solutions of a patient side medical manipulator. However, the effectiveness of different fuzzy membership functions in ANFIS architecture is yet to be explored for the kinematic analysis of robotic manipulators. In other work on path planning for 3-DOFs excavator arm, Vu et al. (2018) proposed an ANFIS model to generate the via-points in joint space while following the desired trajectory. They have considered two cases where number of fuzzy rules is kept constant and varying for each joint, respectively. The results with different rules are found to be more effective. In a recent work by Abbas et al. (2019), inverse kinematic solution of 5R robotic manipulator is carried out and further utilized in the tracking control as desired trajectory. The proposed approach reduces the computational time and complexity while tracking the desired trajectory. In this chapter, ANFIS architecture is explained with different fuzzy membership functions. Thereafter, a 3-R robotic manipulator is considered for the direct and inverse kinematic modeling. For a desired semicircular path, the ANFIS-based inverse kinematic solutions for five different MFs are compared with analytical solutions. The maximum relative deviation is estimated to analyze the effectiveness of different membership function. Selecting different MFs in literature for path planning of the robotic arm is supported by this study. Therefore, this study presents the clarity on selection of particular membership function in the path planning of industrial robots. The rest of the chapter is structured as follows: ANFIS architecture is represented with explanation of five membership functions in section “ANFIS Architecture.” The forward and inverse kinematic problems are addressed for 3-R robotic manipulator in section “KinematicModeling of 3-DOFs RoboticManipulator.” Section “Application of ANFIS for 3-DOFs RoboticManipulator” presents the details of training dataset and implementation of ANFIS model for considered robotic arm. Section “Comparative Study of FuzzyMembership Functions” illustrates the inverse kinematic solutions and generation of desired path for each membership function, showing a state of comparison. Moreover, the effect of selection of MFs in different

1106

J. Narayan et al.

literature works is also presented. Section “Conclusions” underlines the conclusions and future aspects of the work.

ANFIS Architecture The augmented neuro-fuzzy strategy, named ANFIS (Jang 1993), is the hybrid form of fuzzy logics and artificial neural networks. This strategy abridges the inference system of fuzzy logic with learning capability of neural network for better level of inference. In ANFIS, backpropagation or hybrid method is exploited to alter the inferences of the membership functions. A neuro-fuzzy strategy utilizes the fuzzy-based input variables and non-fuzzy input-based output variables (Takagi and Sugeno 1993). The ANFIS architecture is demonstrated with five prominent layers: node layer, membership layer, rule layer, defuzzification layer, and output layer, as shown in Fig. 2. The first layer, that is, node layer delivers the input parameters to the next layer by forming a fuzzy set. The corresponding set acts as an input for the next layer. The second layer, named as membership layer, evaluates the firing strength based on multiplication of every membership function with each other. Thereafter, the respective degree of the input bounds for the third layer is regulated using a firing strength. In the third layer, that is, rule layer, activation level for each

Fig. 2 Schematic representation of ANFIS structure

43 Fuzzy Membership Functions in ANFIS for Kinematic Modeling of. . .

1107

Fig. 3 Membership functions

rule is estimated by relating each firing strength with summation of all firing strengths. In the fourth layer, that is, defuzzification layer, a set of subsequent design parameters form a linear relationship with premise parameters using IF-THEN rules and provide resulting output. In last layer, that is, output layer, the ANFIS model is trained automatically for required MF using least square and back propagation algorithm. There are many types of membership function available in the literature (Narayan and Singla 2017; Sambariya and Prasad 2017). In this work, the considered membership functions, as shown in Fig. 3, are triangular (trimf ), trapezoidal (trapmf ), Gaussian (gaussmf ), generalized bell curve (gbellmf ), and sigmoidal (dsigmf ). The effectiveness of using these membership functions in ANFIS architecture is compared while solving the inverse kinematic problem of a robotic arm. A systematic explanation of all the MFs is given below.

Triangular Membership Function A triangular MF, defined by (1), depends on three parameters ã, b, and c´ . These parameters specify the three corner point coordinates on the x-axis and form the triangular MF.     x − a˜ c´ − x ,0 , F (x; a , b, c´ ) = max min b − a˜ c´ − b

(1)

Trapezoidal Membership Function A trapezoidal membership function is illustrated in (2). It relies on four parameters ã, b, c´ , and d. ¯ These parameters denote the x-coordinates at four corner points which

1108

J. Narayan et al.

form the MF. It is to be noted that trapezoidal MF reduces to triangular one for b equal to c´ .     a˜ d-−x F (x; a, ˜ b, c, ´ d-) = max min x− b−a˜ , 1, d-−´c , 0

(2)

Since triangular and trapezoidal MFs are made up of straight lines; therefore, not smooth at corner points. For smooth corners, Gaussian, generalized bell, and sigmoidal MFs are introduced.

Gaussian Membership Function As defined by (3), an axisymmetric Gaussian membership function is characterized by two parameters σ and ã which signify the curve width and distance from the starting point (origin). F (x; σ, a˜ ) = e

−(x−˜a)2 2σ 2

(3)

Generalized Bell Membership Function The generalized bell MF is defined by (4), where ã, b, and c´ denote the width, shape, and center of the MF, respectively. In the expression, b is usually positive; however, it could be negative for inverted bell-shaped MF. The bell-shaped MF comprises of one extra parameter than Gaussian MF, thereby increasing one DOF to tune the steepness at crossover positions. F (x; a˜ , b, c´ ) =

1+



1 |x−´c| |˜a|

2b

(4)

The Gaussian and generalized bell MFs can easily attain the smoothness behavior; however, unable to deliver the asymmetric curve. Therefore, the sigmoidal MF is utilized for the same.

Sigmoidal Membership Function A sigmoidal MF , specified by (5), depends on two parameters ã and c´ . In this MF, the parameter ã adjusts the slope at the crossover point (x = c´ ). It could be of open right or left in shape, based on the sign of ã. For a closed MF, two sigmoidal MFs can be used. F (x; a , c´ ) =

1 1 + e−˜a(x−´c)

(5)

43 Fuzzy Membership Functions in ANFIS for Kinematic Modeling of. . .

1109

Kinematic Modeling of 3-DOFs Robotic Manipulator For any robotic manipulator or arm with multiple degrees of freedom, the kinematic modeling can be categorized into direct kinematic and inverse kinematic modeling (Schilling 1996). In forward kinematic analysis, the position as well as the angular orientation of the robot’s end effector can be evaluated by employing the joint variables. On the other hand, inverse kinematic analysis describes the procedure to obtain the joint angles and offsets for the desired location and orientation. In this section, following a similar approach given in Narayan and Singla (2017), a 3-DOFs robotic manipulator with RRR configuration (three revolute joints) is considered for kinematic analysis. The 3-DOFs robotic arm is designed in the SolidWorks software, as shown in Fig. 4. The coordinate frames are properly assigned using D-H conventions, as shown in Fig. 5. In Table 1, the respective joint and end-effector movements are specified in a detailed manner. The D-H parameters in terms of joint offset, link length, joint angle, and twist angles are tabulated in Table 2 for the formulation of direct kinematic. The analytical approach to estimate the inverse kinematic solutions for the robotic arm is formulated using algebraic method. In general, the conversion matrix Kii−1 from frame Fi-1 to the adjacent frame Fi can be defined as (6): Fig. 4 CAD model of the 3R manipulator

1110

J. Narayan et al.

Fig. 5 DH frame representation of 3R manipulator

Table 1 Workspace for 3-DOFs manipulator S. No. 1. 2. 3.

Link type First link Second link Wrist

Robot segment Waist Shoulder Wrist roll

Movement direction Right/left Right/left Clockwise/anticlockwise

⎤ Cθi −Sθi Cαi Sθi Sαi ai Cθi ⎢ Sθi Cθi Cαi −Cθi Sαi ai Sθi ⎥ ⎥ =⎢ ⎣ 0 Sαi Cαi di ⎦ 0 0 0 1

Defined workspace (− π/4): π/4 (rad) (− π/4): π/4 (rad) (− π/4): π/4 (rad)

⎡

Kii−1

(6)

43 Fuzzy Membership Functions in ANFIS for Kinematic Modeling of. . .

1111

Table 2 DH parameters of 3R manipulator S.No. 1. 2. 3.

Joint angle Θ i (rad) Θ1 Θ2 Θ3

Joint offset di (mm) d1 = 400 0 d3 = 150

Link length ai (mm) a1 = 250 a2 = 150 0

Twist angle α i (rad) π 0 0

From Table 2, the D-H parameters are placed in (3) to formulate the direct kinematic relations for 3-DOFs robotic arm with the product of K10 , K21 , and K32 . ⎡

⎤ ⎡ .. Cθ1−2−3 −Sθ1−2−3 . p R 3×3 1×3 ⎥ ⎢ ⎢ ⎥ ⎢ Sθ1−2−3 Cθ1−2−3 K30 = ⎢ ⎣ ... ... ... ⎦ = ⎣ 0 0 . 0 0 0 .. 1 0 0

⎤ 0 a1 Cθ1 + a2 Cθ1−2 0 a1 Sθ1 + a2 Sθ1−2 ⎥ ⎥ ⎦ 1 d1 − d3 0 1 (7)

where, Cθ i = cos(θ i ), Sθ i = sin (θ i ), Cα i = cos(α i ), and Sα i = sin(α i ), Cθ i – j = cos(θ i – θ j ), Sθ i – j = sin(θ i – θ j ), Cθ i – j – k = cos(θ i – θ j – θ k ), Sθ i – j – k = sin(θ i – θ j – θ k ), Using Eq. (4), px = a1 Cθ1 + a2 Cθ1−2

(8)

py = a1 Sθ1 + a2 Sθ1−2

(9)

pz = d1 − d3

(10)

where px , py , pz indicate translation in x, y, and z direction, respectively. After simplifying the above relations, the inverse kinematic solutions, that is, the joint variables (joint angles in this case) can be obtained in the form of position variables.

px 2 + py 2 − a1 2 − a2 2 θ2 = ±arccos 2a1 a2 



a2 Sθ2 px + (a1 + a2 Cθ2 ) py θ1 = arctan (a1 + a2 Cθ2 ) px − a2 Sθ2 py  θ3 = θ1 − θ2 − arctan

Sθ1−2−3 Cθ1−2−3

(11)  (12)

 (13)

1112

J. Narayan et al.

As it can be clearly observed from Eqs. (11)–(13), obtaining multiple solutions is a challenging task with analytical approach. Furthermore, if there will be an increase in degrees-of-freedom, several computational complexities will rise in estimating the inverse kinematic solutions due to the non-closure form of equations (Narayan and Singla 2017; Narayan et al. 2018). Therefore, to address these problems, the intelligent techniques like ANFIS can be used in a promising manner to evaluate the inverse kinematics solution. This approach guarantees the reduction of computational cost with the consumption of minimal time.

Application of ANFIS for 3-DOFs Robotic Manipulator ANFIS technique executes in two stages: training and testing (Narayan et al. 2018). In first one, the tip coordinates (x, y and z) and joint variables (θ 1 , θ 2 , and θ 3 ) of the 3-R manipulator work as training input and output parameters. Here, three training datasets are formed as (x, y, z, θ 1 ), (x, y, z, θ 2 ), and (x, y, z, Sθ 1–2–4 , Cθ 1–2–4 , θ 3 ), respectively. In first two datasets, five MFs are selected for each input parameter and comprises of total 125 rules. The effect of wrist roll is considered in the third dataset leading to the five input parameters. The third dataset also contains five MFs for each input parameters and having a total of 3125 rules. The fuzzy rules are formed to institute the knowledge base. After carrying out the numerical experiment, 10 epochs are found to be efficient for whole training process. As soon as required tolerance is attained by the architecture, the training procedure turns out to be stopped. In the testing phase, modifiable limits are utilized due to their adaptive behavior. Thereafter, the validation with the independent data is carried out in the second phase. Conventionally, ANFIS exploits two different learning techniques to correlate the input and output sets; backpropagation- and hybrid-based. In case of backpropagation-based learning method, a gradient descent algorithm is used to compute the node error during ANFIS training. On the other hand, hybrid form of learning method employs a least square method with the gradient descent algorithm to adjust the errors while training. In this work, the hybrid form of learning is utilized with the input-output dataset while training the ANFIS architecture. A detailed implementation of ANFIS algorithm for the robotic manipulator is shown in Fig. 6. Initially, the input parameters are defined in terms of joint variables using DH frame assignment. These parameters form an input dataset required during training of ANFIS architecture. In case of robot path planning, the input dataset is also known as training workspace which could be less than or equal to the actual workspace of the robotic arm manipulator. Based on the input dataset, the Cartesian coordinates (x, y, and z) are computed using forward kinematic equations. Moreover, if the considered robotic manipulator is planar one, the value in z-direction is kept constant; otherwise, keep varying with respect to the time. The coordinate vectors constitute an output dataset in designed ANFIS model. An input-output training dataset is altogether formed using vectors of joint variables and Cartesian coordinates. After forming training dataset, fuzzy inference system (fis) models are generated using specified type and number of membership functions. A separate

43 Fuzzy Membership Functions in ANFIS for Kinematic Modeling of. . .

1113

Fig. 6 Implementation of ANFIS algorithm (Narayan et al. 2018)

membership function is allocated for every input parameter, that is, joint variable. The type and number of MF pertaining to the input parameters could be similar or different at a time. However, in the current study, a similar specification of MF is kept for all three input parameters at once. Moving further with fis models, the ANFIS architecture is trained using anfis syntax for certain number of epochs. The training is carried out with display settings labeled as 0 or 1 where 0 and 1 suppress and display the error results at each epoch, respectively. Thereafter, desired trajectory is defined within the trained input workspace. Invoking the evalfis syntax, the joint variables are evaluated using desired Cartesian vectors and trained ANFIS architectures. Moreover, the desired Cartesian coordinates are also fed into the analytical-based inverse kinematic solutions to compute the joint variables. Finally, the analytical and ANFIS-based inverse kinematic solutions are compared for permissible error tolerance. If the deviation of ANFIS-based solutions is within the limits, the results are considered as

1114

J. Narayan et al.

satisfactory; otherwise, the tuning of training parameters like dataset range, type and number of MFs, and number of epochs is carried out. This process repeats itself till the deviations are found to be satisfactory and attains the permissible tolerance level. In general, though the allowable tolerance level is considered as approaching to zero, it varies one application to other. For example, in case of surgical manipulation, the permissible error could be approximately 10−3 mm as mentioned by Narayan et al. (2018). In contrast, this error could be in the range of 0.1–10 mm for path planning of mobile manipulators and humanoid robots.

Comparative Study of Fuzzy Membership Functions Following the Manjaree et al. (2015), Narayan and Singla (2017), and Narayan et al. (2018), a desired semicircular path/trajectory for 3-DOFs robotic manipulator is considered as x0 = 0; y0 = 0; r = 400; x = x0 + rcosθ ; y = y0 + rsinθ ; z = 250 where, −45 ≤ θ ≤ 45 (degrees) where x0 and y0 denote the manipulator’s base in Cartesian coordinates. The radius of the semicircular trajectory is represented by ‘r’. x, y, and z denote the position coordinates in the respective directions. Considering the 3-R robotic manipulator as a planar one, z coordinate is kept constant with varying x and y positions for the desired trajectory. The responses for joint variables (θ 1 , θ 2 , θ 3 ) for the desired path with every membership function are plotted using MATLAB, as shown in Fig. 7a–c.The ANFIS-based inverse kinematic solutions for five membership functions along with the analytical one are shown with six different color based marker type. Total number of samples (N) used are 90 for all simulation runs. To clearly differentiate the deviations of ANFIS-based inverse kinematic solutions from analytical one, magnifier boxes are drawn at 69th, 49th, and 38th sample for θ 1 , θ 2 , and θ 3 , respectively. Moreover, percentage relative error for each MF is estimated at the maximum deviation of every joint angle from the desired one. Table 3 presents the detailed analysis of the same, where the sample number (at x-axis) for maximum deviation is shown in the brackets. It can be noted that the maximum deviations are observed at the starting or near the end of the solutions. This is due to the upper and lower bounds of training workspace near which predicted solutions are not

43 Fuzzy Membership Functions in ANFIS for Kinematic Modeling of. . .

1115

Fig. 7 (a) Response of “θ1 ” with “Number of samples.” (b) Response of “θ2 ” with “Number of samples.” (c) Response of “θ3 ” with “Number of samples”

1116

J. Narayan et al.

Table 3 Relative percentage error at maximum deviation Membership Function (MF) Gaussian gbell Sigmoidal Trapezoidal Triangular

Relative Percentage Error (θ1 ) (θ2 ) 6.05 (n = 82) 12.23 (n = 82) 7.28 (n = 83) 16.40 (n = 8) 9.34 (n = 81) 15.20 (n = 81) 9.55 (n = 82) 21.80 (n = 9) 16.81 (n = 90) 39.10 (n = 89)

(θ3 ) 11.47 (n = 9) 13.95 (n = 9) 18.36 (n = 80) 22.00 (n = 70) 40.97 (n = 89)

accurate. The accuracy of predicting the inverse kinematic solution degrades further and behaves erratically, especially if desired trajectory is completely out of training workspace. It can be observed from Table 3, the relative percentage error at maximum deviation is found to be lowest with Gaussian MF (6.05, 12.23, and 11.47) and highest with triangular MF (16.81, 39.10, and 40.97) for all joint variables. Moreover, the relative percentage errors at maximum deviation with Gaussian and generalized bell MFs are close to each other as compared to other MFs. The typical order of MFs with increasing relative percentage error is Gaussian, gbell, sigmoidal, trapezoidal, and triangular, respectively. However, for second joint variable (θ 2 ), the relative percentage error at maximum deviation is found to be less with sigmoidal MF (15.20) in comparison to the bell one (16.40). Furthermore, for Gaussian and generalized bell MFs, the order of joint variables with increasing relative percentage error is θ 1 , θ 2 , and θ 3 . The order changes and turns out to be θ 1 , θ 3 , and θ 2 for remaining three MFs. From the simulation runs, effect of Gaussian and generalized bell MFs are found to be more promising for kinematic analysis of 3R manipulator. Following the effectiveness of MFs, a qualitative state of justification for selecting them can be explored in case of different literature works. For instance, in Manjaree et al. (2015), total eight combinations of dataset are considered in view of multiple solutions for 5-DOF robotic manipulator. After computing average of percentage error, the least values with Gaussian and bell MF are observed for seventh and fourth dataset. Moreover, in case of tracking a circular trajectory, the Gaussian MF-based ANFIS model is found to be more effective by 64.12%, 84.16%, and 98.13% in x-, y-, and z-direction, respectively. In other work by Narayan and Singla (2017), Gaussian MF is exploited in ANFIS training architecture to predict the desired and analytical trajectory. The maximum deviation of end-effector from the desired trajectory is 0.3774 mm, 0.4135 mm, and 0.0027 mm in x-, y-, and z-direction, respectively. The analytical one is also found to be in close agreement with the desired trajectory. However, in another work by Narayan et al. (2018), the generalized bell MF is selected in ANFIS architecture during path planning of a 5-DOF patient-side surgical manipulator. The maximum absolute deviations between desired an ANFIS predicted trajectory are 1.07%, 1.32%, and 0.19% in x-, y-, and z-direction, respectively. All the errors are found within the tolerance

43 Fuzzy Membership Functions in ANFIS for Kinematic Modeling of. . .

1117

level of 10−3 as acceptable in case of surgical procedures. Moreover, for robot’s home position, the absolute deviation of joint variables with PRRR configuration is 2.5 mm, 4.22◦ , 0.41◦ , 1.22◦ , and 3.01◦ . It is worth to be mentioned that authors have applied subtractive clustering method to generate the fis models instead of grid partitioning due to the number of inputs greater than five. Therefore, it can be understood that bell MF is more effective than Gaussian one for subtractive clustering method. At last, it is worth mentioning that the accuracy of ANFIS-predicted solutions is not regulated only by selection of MFs, but also depends on other training parameters such as number of MFs, number of epochs, shape of desired trajectory, and training workspace.

Conclusions In this chapter, the significance of MFs’ selection in the ANFIS approach, for kinematic analysis of 3R manipulator, is studied as Gaussian, generalized bell, sigmoidal, trapezoidal, and triangular. Initially, the inverse kinematic solutions for 3R manipulator have been estimated for semicircular trajectory using analytical method. Thereafter, ANFIS technique has been utilized along with five MFs to prove its worth by performing inverse kinematic analysis for same trajectory. The avoidance of finding inverse kinematic equations and solving their nonclosure forms are the noteworthy benefits of ANFIS technique. The effect of five MFs has been compared with each other, for three joint variables, by evaluating relative percentage error at maximum deviation. The Gaussian and generalized bell MFs have shown least relative percentage error at maximum deviation in comparison to other ones. This comparative analysis has been carried out at lower computational cost than the analytical approach. However, proper ANFIS training and training parameters adjustments (epochs, dataset, and number of MF) are required to realize the true potential of each membership function. At last, the selection of particular MF in the literature works has been discussed for ANFIS-based path planning of robotic manipulator. In future, authors will try to analyze the effect of different MFs for obstacle avoidance and complex trajectories. It would also be interesting to investigate the bounds of permissible deviations from the desired path with varied nature of applications. Moreover, as application of ANFIS is mostly limited to the path planning of nonredundant industrial manipulator; therefore, investigating the potential of ANFIS approach for redundant ones is another concern which could be explored in the future. Acknowledgments The authors would like to acknowledge the support received from Mechatronics and Robotics Laboratory, Indian Institute of Technology Guwahati.

1118

J. Narayan et al.

References Abbas M, Narayan J, Dwivedy SK (2019) Simulation analysis for trajectory tracking control of 5-DOFs robotic arm using ANFIS approach. In: 2019 5th international conference on computing, communication, control and automation (ICCUBEA). IEEE, Pune, India, pp 1–6 Alavandar S, Nigam MJ (2008) Inverse kinematics solution of 3DOF planar robot using ANFIS. Int J Comput Commun Control 3:150–155 Ali ZN, Askerzade I, Abdulwahab S (2021) Estimation model for bread quality proficiency usig fuzzy weighted relevance vector machine classifier. Appl Bionics Biomech 2021:1–9 Amza CG, Cicic DT (2015) Industrial image processing using fuzzy-logic. Procedia Eng 100: 492–498 Bai Y, Zhuang H, Wang D (eds) (2007) Advanced fuzzy logic technologies in industrial applications. Springer Science & Business Media, London Boné R, Crucianu M (2002) Multi-step-ahead prediction with neural networks: a review. 9emes rencontres internationales: Approches Connexionnistes en Sciences 2:97–106 Chawla I, Singla A (2019) System identification of an inverted pendulum using adaptive neural fuzzy inference system. In: Yadav N, Yadav A, Bansal J, Deep K, Kim J (eds) Harmony search and nature inspired optimization algorithms. Advances in intelligent systems and computing, vol 741. Springer, Singapore Chen Y, Liu G, Yang Y (2020) Structure of the singular manifold of protein backbone and its applications. IEEE Access 8:48438–48450 Cichocki A, Unbehauen R, Swiniarski RW (1993) Neural networks for optimization and signal processing, vol 253. Wiley, New York Duka AV (2015) ANFIS based solution to the inverse kinematics of a 3DOF planar manipulator. Procedia Technol 19:526–533 Gahrooei YR, Hooshmand RA (2019) Short term electricity price forecasting by hybrid mutual information ANFIS-PSO approach. Sci J Comput Intell Electr Eng 10(1):63–72 Goh AT (1995) Back-propagation neural networks for modeling complex systems. Artif Intell Eng 9(3):143–151 Goyal LM, Mittal M, Sethi JK (2016) Fuzzy model generation using subtractive and fuzzy c-means clustering. CSI Trans ICT 4(2–4):129–133 Hasanipanah M, Amnieh HB, Arab H, Zamzam MS (2018) Feasibility of PSO–ANFIS model to estimate rock fragmentation produced by mine blasting. Neural Comput & Applic 30(4):1015– 1024 Hoffmann F (2003) An overview on soft computing in behavior based robotics. In: International fuzzy systems association world congress. Springer, Berlin/Heidelberg, pp 544–551 Jang JS (1993) ANFIS: adaptive-network-based fuzzy inference system. IEEE Trans Syst Man Cybern 23(3):665–685 Jang JSR, Sun CT, Mizutani E (1997) Neuro-fuzzy and soft computing-a computational approach to learning and machine intelligence [Book Review]. IEEE Trans Autom Control 42(10):1482– 1484 Kim W, Tendick F, Stark L (1987) Visual enhancements in pick-and-place tasks: Human operators controlling a simulated cylindrical manipulator. IEEE J Robot Autom 3(5):418–425 Komori Y, Sagayama S, Waibel AH (1993) A neural fuzzy training approach for improving speech recognition. Syst Comput Japan 24(8):82–94 Lei Y, He Z, Zi Y (2008) A new approach to intelligent fault diagnosis of rotating machinery. Expert Syst Appl 35(4):1593–1600 Maher I, Ling LH, Sarhan AA, Hamdi M (2015) Improve wire EDM performance at different machining parameters-ANFIS modeling. IFAC-PapersOnLine 48(1):105–110 Manjaree S, Nakra BC, Agarwal V (2015) Comparative analysis for kinematics of 5-DOF industrial robotic manipulator. Acta Mech Autom 9(4):229–240 Mouss ME, Zellagui S, Nasraoui M, Hambli R (2020) Parametric investigation of the effects of load level on fatigue crack growth in trabecular bone based on artificial neural network computation. Proc Inst Mech Eng H J Eng Med 234(8):784–793

43 Fuzzy Membership Functions in ANFIS for Kinematic Modeling of. . .

1119

Nachtegael M, Van der Weken D, Van De Ville D, Kerre EE (eds) (2013) Fuzzy filters for image processing, vol 122. Springer, New York Narayan J, Dwivedy SK (2021) Biomechanical study and prediction of lower extremity joint movements using Bayesian regularisation based backpropagation neural network. J Comput Inf Sci Eng 22(1):014503 Narayan J, Singla A (2017) ANFIS based kinematic analysis of a 4-DOFs SCARA robot. In: 2017 4th international conference on signal processing, computing and control (ISPCC). IEEE, Solan, India, pp 205–211 Narayan J, Singla E, Soni S, Singla A (2018) Adaptive neuro-fuzzy inference system–based path planning of 5-degrees-of-freedom spatial manipulator for medical applications. Proc Inst Mech Eng H J Eng Med 232(7):726–732 Ripley BD (2007) Pattern recognition and neural networks. Cambridge university press, Cambridge, UK Roy R, Furuhashi T, Chawdhry PK (eds) (2012) Advances in soft computing: engineering design and manufacturing. Springer Science & Business Media, London Sambariya DK, Prasad R (2017) Selection of membership functions based on fuzzy rules to design an efficient power system stabilizer. Int J Fuzzy Syst 19(3):813–828 Schilling RJ (1996) Fundamentals of robotics: analysis and control. Simon & Schuster Trade, Upper Saddle River, NJ Senthilkumaran N, Rajesh R (2009) Image segmentation-a survey of soft computing approaches. In: 2009 international conference on advances in recent technologies in communication and computing. IEEE, Kottayam, India, pp 844–846 Singh A, Mittal M (2020) Prediction of solar radiation using hybrid discriminant-neural network. In: 2020 sixth international conference on parallel, distributed and grid computing (PDGC). IEEE, Waknaghat, India, pp 150–153 Takagi T, Sugeno M (1993) Fuzzy identification of systems and its applications to modelling and control. In: Readings in fuzzy sets for intelligent systems, vol 15, no 1, pp 387–403 Tien Bui D, Khosravi K, Li S, Shahabi H, Panahi M, Singh VP, Chapi K, Shirzadi A, Panahi S, Chen W, Bin Ahmad B (2018) New hybrids of ANFIS with several optimization algorithms for flood susceptibility modeling. Water 10(9):1210 Ulieru M, Hadzic M, Chang E (2006) Soft computing agents for e-Health in application to the research and control of unknown diseases. Inf Sci 176(9):1190–1214 Vu NTT, Tran NP, Nguyen NH (2018) Adaptive neuro-fuzzy inference system based path planning for excavator arm. J Robot 2018:1–7 Yalçın BC, Demir C, Gökçe M, Koyun A (2018) Water leakage detection for complex pipe systems using hybrid learning algorithm based on ANFIS method. J Comput Inf Sci Eng 18(4):041004 Yaseen ZM, Ghareb MI, Ebtehaj I, Bonakdari H, Siddique R, Heddam S, Yusif AA, Deo R (2018) Rainfall pattern forecasting using novel hybrid intelligent model based ANFIS-FFA. Water Resour Manag 32(1):105–122 Yetilmezsoy K (2019) Applications of soft computing methods in environmental engineering. In: Hussain C (ed) Handbook of environmental materials management. Springer, Cham Zadeh LA (1994) Fuzzy logic, neural networks, and soft computing. Commun ACM 37(3):77–85 Zhao H, Gallo O, Frosio I, Kautz J (2016) Loss functions for image restoration with neural networks. IEEE Trans Comput Imaging 3(1):47–57

Mobile Robot for Gas Leakage Detection System in Pipelines

44

Gnana K. Sheela

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Related Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Image Enhancement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Histogram Equalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Image Thresholding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Binarization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Feature Extraction and Decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . LCD Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermal Camera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SET UP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mobile Platform Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sensing Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inspection Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1122 1123 1126 1128 1128 1128 1129 1129 1129 1129 1131 1131 1131 1131 1131 1131 1134 1135

Abstract

Gas is a matter that is freely moving. Gas can be either in solid state or liquid state. On compressing under high pressure and transported through pipelines. Transportation of gas through pipeline can create some difficulties. If the pipelines undergo any corrosion or any leakage at flanges or joints, the gas will be leaked. So in order to prevent the leakage, much faster and safer method should be important. An automated system is required to detect the gas leakage

G. K. Sheela () APJ Abdul Technological University, Trivandrum, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_27

1121

1122

G. K. Sheela

properly. The gas releases a small amount of energy that can only be detected by infrared radiation. Thermal imaging is a technique is used as a method to detect the gas leakage because it is used for improving visibility of objects in a dark environment by detecting the objects infrared radiation and creating an image based on that information. After obtaining the image using thermal camera it is converted to thermal image using image processing. The thermal camera captures the thermal image. This thermal image can only show the area of gas leakage. The position where the leakage happened can be obtained by using distance calculation. The distance will be sent to emergency number through GSM. Thus it creates an alert and makes necessary precaution to prevent further leakage and to protect from huge disaster to both environment and human life. Keywords

Robot · Gas Leakage · Thermal Image · Image processing

Introduction Most of the chemi compounds and gasses are invisible to the human eye. Several companies work with these substances before, during, and after their production processes. Most of the companies have strict regulations to trace the gas leakage to rectify the problem and report any leaks of gaseous compounds. Also there are regulations regarding the procedures used to carry out. While searching for potential gas leaks, all systems at the points have to be checked that may or may not have been identified. These inspections are carried out regularly for identifying the leakage. Technologies currently in existance to identify the gas leakage may expose human operators to invisible and harmful chemicals. It will produce inaccurate measurements. The gases can be toxic or nontoxic. The continuous inhaling of nontoxic gas can also eventually cause health hazards. The gas leakage cannot be detected by naked eyes because the energy released from the object is very small. But can be read by using thermal images. Thermal imaging technique is used for improving the visibility of objects in a dark environment. It is done by detecting the infrared radiation emitted by object and creating an image based on that information obtained. Infrared technology uses the principle of infrared radiation to measure the temperature and the radiant energy of the object. Cameras can be used in a number of different ways to inspect industrial field. There are several benefits for using thermal imaging techniques. Measurements can be carried out rapidly and at relatively low cost and most importantly problems can be identified at an early stage. The research work aims to develop a system that is able to perform inspections in industrial areas without having to access hazardous areas directly and without requiring any human presence. The proposed mobile robot can be used for routine inspections for targeted inspections of specific system parts. The independent mobility of the system was implemented with various navigation sensors and

44 Mobile Robot for Gas Leakage Detection System in Pipelines

1123

inspection module. This system is also equipped with video and gas sensing system, which enables to inspect system parts that were difficult to inspect by a human operator. The development of innovative monitoring system helps to improve the reliability, efficiency, and it will reduce the cost required for the inspections. At the same time, it explains technical side, time-consuming, and labor intensive tasks. Important reason for using the system is that, it will minimize discharges of gas and other volatile organic substances. The role of leak detection can be evaluated either in the determination of spill size and dispersion or as a stand-alone element that is then used to adjust previous consequence estimates. The former approach is logical and consistent with the real-world scenarios. The benefit of leak detection is indeed its potential impact on spill size and dispersion (Lukonge and Cao 2020). Currently existing technologies uses a network of sensors which are held stationary at particular locations. The main drawback is that gas leakage cannot be identified correctly. After sensor giving an indication that leakage has happened, a human operator has to identify the area were leakage occurred. But it will be with less accuracy. So there occurs a need for robotic system which equipped with sensors and camera to identify the leakage. A new technology called thermal imaging is used here. By thermal image processing the area of leakage can be identified. Figure 1 shows the Flow chart of different gas pipeline leakage detection approaches.

Related Works Kamal et al. (2006) introduced Air Quality Index (AQI) system which plays an important role in conveying to both decision-makers and the general public. The objective of this method is to investigate the effectiveness of Artificial Neural Network. BPNN is a method is used to predict the AAQ (Kamal et al. 2006). So et al. (2007) introduced laser spectroscopic trace gas sensor method for gas detection. Laser SPECKs integrates miniaturized quartz enhanced photo acoustic gas sensing technology. This uses an infrared laser spectroscopy technique. Traditional gas sensing devices has an advantage such that same sensitivity and specificity. This reduces cost and power consumption. The basic principles behind laser based trace gas detection methods, design issues, and outlines are the implementation of a miniaturized trace gas sensor from commercial-off-the-shelf (COTS) components. Gas sensor implements quartz enhanced photo acoustic spectroscopy (QEPAS), a new technology. The merits are high specificity (So et al. 2007). Wang et al. (2010) developed an ambient air quality monitoring technique with wireless sensor networks. One of the common poisonous gas is carbon monoxide produced from the incomplete oxidation of carbon during the combustion process. Necessary equipment needs to be housed and operated inside a room, and protected from rain, dust, and sunlight. Thus these preventive issues make this method complicated, cumbersome, and are expensive (Wang et al. 2010). Wang et al. (2010) put forward an urban air quality monitoring system based on the wireless sensor network technology and integrated with the global system

1124

G. K. Sheela

Pipeline Leakage Detection Methods

Exterior methods

Visual/Biological Methods

Interior/computatonal based methods

Acoustic sensing AUV/Drone

Mass/Volume

Trained Dog/Human

Negative pressure

Accelerometer

Fibre optic sensing

Vapour sampling

Visual-based bolted joints monitoring

Pressure point analysis

Infrared thermography Digital signal Ground penetration Dynamic modelling Flurescence Electromechanical impedance

State estimators

Capacitive sensing

Other methods

Fig. 1 Flow chart of different gas leakage detection approaches (Adegboye et al. 2019)

for mobile communications. The system consists of sensor node, a gateway, and a back end platform. The advantages are small, easy to set up, and inexpensive. The performances of WSNs are subject to unit computing speed, memory capacity and stability of communication, etc. In order to overcome the limitations of hardware, many issues on WSNs software have been considered, such as routing protocols, media access control, coverage and power management. Here it proposes an automatic micro scaled air quality monitoring system for areas with high density of population and vehicles (Thamrin et al. 2012).

44 Mobile Robot for Gas Leakage Detection System in Pipelines

1125

Thamrin et al. (2012) put an idea about Simultaneous Localization and Mapping Based Real-Time Inter-Row Tree Tracking Technique for Unmanned Aerial Vehicle (UAV). This is used for small-scaled unmanned aerial vehicle where substantial devices or equipment that has enormous weight and size. There is several other row detection technique, they are vision based, laser-based, and stereo vision-based techniques which are found to be an accurate row-detection technique. In order to monitor the autonomic vehicles and robots in correct way, the GPS receiver allows for precise navigation and localization in the previous implementation (Rossi and Brunelli 2013). Rossi and Brunelli (2013) developed a method to reveal and measure natural gas presence in air. The advantage is that it extends the autonomy of battery powered system. The most important characterization of the sensing device has been conducted using a range of humidity conditions to demonstrate the effectiveness of the proposed approach. There are two different duty cycle rates that are characterized and simulated to demonstrate it. Thus this demonstrates an innovative strategy for chemo resistive MOX gas sensors. These gas sensors can achieve ultra-low power performance in environmental monitoring applications. Thus it achieved a remarkable energy saving in real-time applications where sensors are always switched on (Rossi et al. 2014). Rossi et al. (2014) developed a portable Gas Sensing System on UAVs for Gas Leakage localization. For environmental monitoring and measuring of volatile chemical concentration gas leakage recognition is important. UAVs are used to measure spatially distributed gas concentration. This is quite recent and few efforts have been dedicated to the design of integrated sensing instruments. That aims at the optimization of crucial features as weight, dimension, and energy autonomy. The Gas Sensing System (GSS) is a fully autonomous board. That is based on a 32bit MCU with 30 min autonomy (on its own battery), data storing, wireless connectivity for real-time feedback, and embeds a custom micro-machined MOX (Metal Oxide) sensor (Manekiya and Arulmozhivarman 2016). Adnan et al. (2015) proposed a new technique for gas leakage detection. The pipelines are operated for a number of years and it will tend to corrode. If the gas filled pipelines leakage happened, it can lead to explosion. The detection of leak in gas pipeline is more difficult compared to water because of the poor signal noise to ratio (SNR). The noise coming from gas flow itself is larger. The geometrical features can also be one of the noises and make the leak detection in gas pipeline more complicated (Jadin and Ghazali 2014). Manekiya and Arulmozhivarman (2016) developed a new technique for leakage detection using infrared thermography. It is used in many fields such as environment, military, industry, etc. IR thermography can detect several problems before they occurs and especially helpful to detect the overloads, worn, and circuit brakes. A new technique was introduced for the flow detection in pipe lines and extracts the effected leakage areas in the pipe. There are three ways of transmission for thermal energy. They are conduction, convection, and radiation. This uses basic concepts of image processing using MATLAB. IR thermography camera works based on its spatial resolution. This camera works from a particular distance. Thus the object should be kept at a particular distance and camera should focus that object. Camera

1126

G. K. Sheela

can detect the infrared rays emitted by the object. By using thermal equipment, the object can be detected. IR camera can distinguish the object by cooler area and warm area showing in the thermal image with a different color of that particular area (Adnan et al. 2015). Adegboye et al. (2019) reviewed different leakage detection and localization in pipeline systems and their strengths and weaknesses are highlighted (Adegboye et al. 2019). An acoustic emission technology for pipeline (2020) leak detection has been explained. The recent developments are discussed and compared with conventional techniques (Lukonge and Cao 2020).

Theory The leakage of gas in the pipelines is one of the major problem that present economy faces. The gas leaked can easily spread in atmosphere. This will later cause pollution and health problems. Figure 2 describes the flow chart of gas leakage detection system. The system function starts with the detection of gas leakage. So there should present an automated system with more reliable and safer methods to detect the leakage. The gas is invisible to naked eye. It can only be identified by using thermal imaging technology. Because the energy radiated by the gas will be smaller, it can only be detected using thermal camera. The leakage detection is identified using sensors. After the gas leakage is identified, the camera captures the image. Then transmit the image to image processing for identifying the leaked area. If leakage is identified, a dialog box appears showing the results as gas is detected or not. When

Fig. 2 Flow chart of gas leakage detection system

Start Leakage is identified Capture Image and transmit

Gas detected No Yes

Calculate Distance

Send location and alert Stop

44 Mobile Robot for Gas Leakage Detection System in Pipelines

1127

leakage is displayed, the distance from the starting location and current position will be send as a message to control room. The thermal imaging technique is used to locate or obtain the image of an object by using the heat given off by that. It was developed for military purposes firstly, later this technique was started using for firefighting, medical, security, transportation, and industries. It works at ambient temperature. Grey scale is a common nature. The black color is for cold objects. And white color for hot objects. And the depth of variation indicates the temperature change. Thermal imaging adds color to the objects. The thermal image is formed based on heat signature. Thermal imaging technique uses a thermal camera. There is a phased array to scan the focused light. The phased array is a part of infrared-detector elements. The detector creates a very detailed temperature pattern called a thermogram. Infrared thermography is equipment or a method, which detects infrared energy, emitted from the object and converts it to temperature then displays image of temperature distribution. Gaussian filtering is a technique that is used for removing noise in image processing. Gaussian noise is a white noise. This requires mean and variance as the additional inputs. It converts the images to blur images and remove noise from original images. They are also known as linear smoothing filters. The Gaussian mask comprises certain elements which are determined by a Gaussian function. Gaussian blur is also known as Gaussian smoothing which is the result of blurring an image by a Gaussian function. It is a widely used effect typically to reduce image noise and reduce detail. There are five steps in thermal image processing. They are data acquisition, image pre-processing, image processing, feature extraction, and classification decision. Data acquisition is the process whereby how the data were acquired. In this step, the tools are used in order to get the data. Image pre-processing can be defined as the process that we want to enhance and improve the original image to be used in next process which is called image processing. Image processing is an important stage in methodology because at this stage the image will be analyzed and digitized in order to get the desired image. Generally, feature extraction is the process of creating features to be used in the classification. Classification process is the process to make decisions based on test and analysis done on the image [18]. The Fig. 3 describes about the block diagram of image processing.

Data Acquisition

Image Enhancement

Apply Gaussian filter

Histogram Equalization

Image Thresholding

Binarization

Erosion

Fig. 3 Block diagram for thermal image processing

Classification

1128

G. K. Sheela

Data Acquisition In this step, data were acquired by using an thermal camera and save as AVI (Audio Video Interleaved) file (Jadin and Ghazali 2014). Thermal camera is used for collecting data.After recording the video of the gas situation, that video is converted into image in JPEG image format by using MATLAB coding. There are three states for the leakages are described in figure. They are before leakage, during leakage and after leakage. The gas leakage is clearly appeared. The color of the surrounding changes as the gas released. The problem on using digital camera is that, the changes of the surrounding color cannot be seen. The hole of the leakage is invisible through digital camera, so this is the main reason why thermal camera is more preferred in thermal image processing compared to the normal digital camera. Tracks could be extracted by video. Video is converted to frame by frame. The pixels with same thermal signature are grouped. These pixels are later grouped to objects. The parameters such as the amount of thermal signature similarity, proximity of objects, etc. are considered for grouping same pixels. Efficiency of track identification using observed-based counts was tested the tracks within segments of sample video. The difference in count was a part of parameterization. Original track attributes including measures of thermal intensity, object size, and rate of travel concerns the view.

Image Enhancement Image enhancement means the process that lies on the image pre-processing steps. This is important in order to remove noise from the original image because most of the original images are unclear and blurred. Image will be clear in appearance compared to the original one. RGB image consists of three color arrays which are red, green, and blue. Thus these color arrays needs to be converted into grayscale. After this step, the image is enhanced by image filtering and noise removing.

Histogram Equalization Histogram equalization is a technique in which the dynamic range of the histogram of an image is increased. Thus the output image contains a uniform distribution of intensities. These techniques can be used on a whole image or just on a part of an image. It redistributes intensity distributions. In this case, if the histogram of any image has many peaks and valleys, it will still have peaks and valley after equalization, but peaks and valley will be shifted. Due to this, spreading is a better term than flattening to describe histogram equalization. Each pixel is assigned a new intensity value. This value is based on its previous intensity level.

44 Mobile Robot for Gas Leakage Detection System in Pipelines

1129

Image Thresholding The coolest region is identified by using image thresholding. After the binarization process, it is easy to analyze the image. Images are classified based on the pixel value in the eroded image. The arrow in the figure shows the area of thresholding process. After the detection of coolest region, the image is once again converted into a grayscale image in order to enhance the coolest region. By using thermal camera the video is recorded and then abstracted to be an image. The original image is converted to a grayscale image in the first step. After that the filtering process takes part. Then thresholding process, binarization, process using im2bw also are applied, but there has no change to the image. The output is not visible. So we can conclude that the gas leakage does not exist for this image (Jadin and Ghazali 2014).

Binarization The Binarization Method is that it converts the gray scale image (0 up to 256 gray levels) into a black and white image (0 or 1). The binarized image can give more accuracy in character recognition as compared to original image because noise is present in the original image. The main problem is that which binarization algorithm is appropriate for all images. The selection of the binarization algorithm is difficult. This is true in the case of historical documents images with variation in contrast and illumination. Two types of algorithms are: (a) Global Binarization (b) Local Binarization. The global binarization methods use single threshold value for the whole image and the local binarization method use the threshold value calculated locally pixel by pixel or region by region (Puneet and Garg 2013).

Feature Extraction and Decision Feature extraction is the process is the value of the eroded image which is in binary value. The image will tell that the leakage happened or not based on the feature extracted. After getting coolest region of the image, the image needs to convert to binary to easy analysis. Finally, the morphology technique was used in order to achieve the desired output. The closing and erosion operations are used. In this project, the image was classified based on the pixel value that we extract in the eroded image (Jadin and Ghazali 2014).

Description This section consists of ARM microcontroller (LPC2148), LCD display, motor driver, GSM, camera, and sensor. The ARM microcontroller is used in this project because the power consumption is less in this controller. The voltage required is

1130

G. K. Sheela

SENSOR

LCD DRIVER

LCD DISPLAY

DUAL H BRIDGE

MOTOR

USART

GSM

ARM Micro controller THERMAL CAMERA

(LPC2148)

RESET

Fig. 4 Hardware block diagram

3.3 V. LCD display is used to display the value of gas sensed by the sensors. The sensors are interfaced to ADC pin of ARM microcontroller. After sensing the ADC value, it is converted in to digital values. These digital values are displayed on LCD display. A camera is interfaced to this controller. This is used to identify the gas leakage if happened. For the robotic mechanism a motor is interfaced. The motor is interfaced using dual H bridge circuit. The need for dual H Bridge is that for proper working of motor; the current should flow in two directions. Only a H bridge circuit can make current flow in two directions. The message showing the leakage has happened is send to the control room using a GSM. The GSM is interfaced using USART (Fig. 4). ARM Microcontroller LPC 2148 is used. The features are 128-bit wide interface/accelerator that enables high-speed 60 MHz operation, Single flash sector or full chip erase in 400 ms and Full-speed compliant device controller with 2 kB of endpoint RAM [20]. Keil software is used for Dual H bridge. L293D is a Motor driver IC which allows DC motor to drive on either direction. It consists of two DC motors simultaneously in any direction. It means that it is possible to control two DC motor with a single L293D IC. Dual H-bridge is a motor driver integrated circuit (IC). It works based on the concept of H-bridge. Voltage is needed to change its direction. Then only the motor can rotate in clockwise or anticlockwise direction; therefore, these ICs are ideal for driving a DC motor. In a single chip there are two h-Bridge circuit inside the IC. It can rotate two dc motor independently. Its size is important factor. Based on its size, it is widely used in robotic application for controlling DC motors.

44 Mobile Robot for Gas Leakage Detection System in Pipelines

1131

LCD Display LCD means Liquid Crystal Display. It is the technology used for displaying in notebook and other smaller computers. LCD allows displays to be much thinner. One of the advantages is that it consumes less power than LED and gas-display displays because it works on the principle of blocking light rather than emitting it.

Thermal Camera Thermal camera is used to take an image using infrared radiation. The operating wavelength of a thermal camera is 14,000 nm. The main features are to visualize gas leaks in real time, it can be used for temperature measurement applications; inspect without interruption of process and it can reduce inspection time.

SET UP Mobile Platform Robot A mobile robot means it is an automatic machine that is capable of locomotion. They have the capability to move around in their environment and are not fixed to one physical location. Mobile robots can be autonomous. They can navigate through an uncontrolled environment. This robot is having a chain driven platform. This chain driven mobile platform helps to move through any type of surfaces. Thus it can easily monitor the area of leakage.

Sensing Module Sensing module consist of a sensor. It is an electronic component. A sensor is always used with other electronics. Sensors capture the analog value and convert it in to digital values. Sensors can be of different types for measuring pressure, temperature, flow parameters, etc. Analog sensors such as potentiometers and force-sensing resistors are widely used. Sensitivity of a sensor means how much the sensor’s output changes when the input quantity. Small sensor will improve sensitivity. Several sensors on combining can form a micro mems systems. The speed of micro sensors is high.

Inspection Module The inspection module mainly consists of an thermal camera. This camera detects the temperature, released by the body, and identify the area where the leakage occurs. That processed to produce a thermal image on a video monitor. The heat

1132

G. K. Sheela

Fig. 5 Hardware set up

sensed by a thermal camera can be very precisely quantified. An infrared image without any measurements can be misleading. Thermal image without an accurate measurement says very little about the condition of an electrical connection or mechanical part. The hardware set up for the gas leakage detection system is shown in Fig. 5. This system moves in a predefined path around the leakage area. It continuously monitors the pipelines and transmits the information to control room wirelessly. This system consists of chained mobile platform. Two gas sensors for detecting the leakage. It consists of GSM module. GSM is used for transmitting the distance where the leakage has happened. The image is captured using thermal camera. Then it is transmitted to control room for MATLAB processing in order to get the thermal image from that. The image is obtained after using a Gaussian filter. The original image is in the form of RGB color format. It is converted to grey scale color. Thus the original image always contains some noises. It will be blurred and unclear. So a Gaussian filter is used to remove the noise. The Fig. 6 is the result of thermal imaging technique after the thresholding process. After the image thresholding the coolest region is detected. This coolest region is considered as the region of gas leakage. Figure 7 is the thermal image after binarization process. In this process again the image is converted to binary image for more easy analysis. Thus the region of leakage is clearly identified.

44 Mobile Robot for Gas Leakage Detection System in Pipelines

1133

Fig. 6 Image obtained after image thresholding

Fig. 7 Binary image

Figure 8 gives the final result of image processing. That is, it tells that whether the gas leakage is detected or not. If leakage is identified then dialog box will appear on the screen showing that “gas detected.” The result showing that the gas detected is displayed on a dialogue box. The distance calculated will be send as massage to the control room by GSM. Thus, an alert is provided. Also necessary precaution can be taken to avoid further leakage.

1134

G. K. Sheela

Fig. 8 Gas detection results

Conclusion Gas leakage detection has grown in importance because of safety concerns, a performance of systems, and health conditions. Leaks can happen because of poor seals and connections, as well as from inadequate welds. So for the detection of these dangerous gases, a well-developed detection system should be there. By replacing a human operator a robotic system can detect the leakage of gas. The sensors in the robot identify the gas leakage. An alarm will provide the alert. These interfacing are done using Keil software. The thermal camera can capture the image. This captured video is then transmitted to image processing in MATLAB. The thermal images help to identify the gas leaked area more easily. Because the temperature changes at the point of gas leakage can be identified easily from thermal image. In MATLAB after image processing, the thermal image will be clearly visible after the noise removal and binarization process. The leaked area is seen as white color and rest of the surrounding in black color. The distance from the starting position to the location of the vehicle will be obtained by distance calculation method and finally, it is sent as a message to the control room GSM. The simulation results show that the image obtained after each stage in image processing. Finally, for easy analysis, the binary image is obtained. Then a dialogue will be displayed as the final result showing the gas is detected. As instead of sending the message through GSM it can be obtained in the MATLAB software itself as a result of a modification.

44 Mobile Robot for Gas Leakage Detection System in Pipelines

1135

References Adegboye MA, Fung W-K, Karnik A (2019) Recent advances in pipeline monitoring and OILLEAKAGE detection technologies: principles and approaches. Sensors 19:1–36 Adnan NF, Ghazali MF, Amin MM, Hamat AMA (2015) Leak detection in gas pipeline by acoustic and signal processing – a review. 3rd international conference of mechanical engineering research (ICMER 2015), pp 1–9 Jadin MS, Ghazali KH (2014) Gas leakage detection using thermal imaging technique. 16th International conference on computer modelling and simulation (UKSIM), pp 301–305 Kamal MM, Jailani R, Shauri RLA (2006) Prediction of ambient air quality based on neural network technique. 4th student conference on research and development (SCOReD 2006), Shah Alam, Selangor, Malaysia, pp 27–28 Lukonge AB, Cao X (2020) Leak detection system for long-distance onshore and offshore gas pipeline using acoustic emission technology. A review. Trans Indian Inst Metals 1–13 Manekiya MH, Arulmozhivarman P (2016) Leakage detection and estimation using IR thermography. International conference on communication and signal processing (ICCSP), Melmaruvathur, India, pp 1516–1519 Puneet P, Garg NK (2013) Binarization techniques used for grey scale images. Int J Comput Appl 1:1–11 Rossi M, Brunelli D (2013) Ultra low power CH4 monitoring with wireless sensors. IEEE conference on SENSORS, Baltimore, pp 1–4 Rossi M, Brunelli D, Adami A, Lorenzelli L, Menna F, Remondino F (2014) Gas-drone: portable gas sensing system on UAVs for gas leakage localization. In Proceedings of IEEE SENSORS, Valencia, Spain, pp 1–4 So S, Koushanfar F, Kosterev Anatoliy (2007) LaserSPECks: laser SPECtroscopic trace-gas sensor networks sensor integration and applications. 6th international symposium on Information processing in sensor networks, Cambridge, MA, pp 226–235 Thamrin NM, Arshad NHM, Adnan R, Sam R, Razak NA, Misnan MF and Mahmud SF (2012) Simultaneous localization and mapping based real time inter-row tree tracking technique for unmanned aerial vehicle. 2012 IEEE international conference on control system, computing and engineering, Penang, Malaysia, pp 322–327 Wang D, Agrawal DP, Toruksa W, Chaiwatpongsakorn C, Lu M, Keener TC (2010) Monitoring ambientair quality with carbon monoxide sensor-based wireless network, Commun of ACM 53(5):138–141

Improved Security Models in Mobile Wireless Vehicle Networks

45

Gnana K. Sheela

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L-DRAND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CarTALK 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zone Flooding Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zone Diffusion Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dynamic Time-Stable Geocast Routing in Vehicular Ad-Hoc Networks . . . . . . . . . . . . . Chain Collisions of Vehicles Equipped with Vehicular Communications . . . . . . . . . . . . . A Vertical Handoff Method via Self-Selection Decision Tree . . . . . . . . . . . . . . . . . . . . . . Mobile Location Estimator Using Extended Kalman-Based IMM and Data Fusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scalability of Vehicle-to-Vehicle Communication with Prediction- based STDMA . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1138 1141 1145 1146 1146 1147 1148 1152 1152 1153 1153 1154 1156 1158 1161 1162

Abstract

The recent developments in technology are growing at a faster pace in numerous areas. The world’s first driverless public transport system in vehicles is on the schedule to be implemented in Dubai before EXPO 2020. The majority of research on data collection using wireless mobile vehicle network emphasizes the reliable delivery of information. So far various algorithms have been developed and implanted and different case studies have been conducted till date in the area of VANET. Various algorithms are analyzed such as TDMA, STDMA, and

G. K. Sheela () APJ Abdul Technological University, Trivandrum, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_28

1137

1138

G. K. Sheela

self-adaptive sensing model. The main purpose of the project is to enhance the self-adaptive sensing model. The main areas of concern in self-adaptive sensing model are the security issues. Performance requirements such as packet delivery ratio and delay are not given prime importance, thus making data collection ability of vehicular nodes in real application environment inferior. The security feature was added to self-adaptive sensing model and the performance analysis was carried out. A comparative study of the algorithms was carried and the results were analyzed and evaluated. The comparative analysis on throughput, delay, packet delivery ratio, and routing overhead shows the enhanced capability of secure data collection algorithm. Keywords

Hand off · Protocol · Wireless · Vehicle networks

Introduction The mobile wireless vehicle network attracts more attention in recent years in real applications. Mobility has brought latest challenges to intelligent vehicle networks. Since vehicular nodes always are dynamic and since they move faster, routing for WSN data transmissions may easily loose security and stability. Also, manage with rapid change of dynamic topology and improving efficiency of data dissemination are also some urgent issues to be addressed in a wireless vehicle network. Vehicular networks are the combination of transport systems and the Internet systems formed with the main motive to increase the safety of passengers, although nonsafety applications are also provided by vehicular networks. Internet of Things (IoT) has a subsection called Mobile Ad hoc Network (MANET), which in turn has a subsection called Vehicular Ad hoc Network (VANET). Internet of Energy (IoE) is a new domain that is formed of electric vehicles connected with VANETs. As a large number of transport systems are coming into operation and various pervasive applications are designed to handle such networks, the increasing number of attacks in this domain is also creating threats. As IoE is connected to VANETs extension with electric cars, the future of VANETs can be a question if security measures are not significant. The present survey is an attempt to cover various attack types on vehicular networks with existing security solutions available to handle these attacks. This study will help researchers in getting in-depth information about the taxonomy of vehicular network security issues which can be explored further to design innovative solutions. This knowledge will also be helpful for new research directions, which in turn will help in the formulation of new strategies to handle attacks in a much better way. VANETs support a wide range of applications – from simple one hop information dissemination of, for example, cooperative awareness messages (CAMs) to multihop dissemination of messages over vast distances. Most of the concerns of interest to mobile ad hoc networks (MANETs) are of interest in VANETs. Rather than

45 Improved Security Models in Mobile Wireless Vehicle Networks

1139

moving at random, vehicles tend to move in an organized fashion. The interactions with roadside equipment can likewise be characterized fairly accurately. And finally, most vehicles are restricted in their range of motion, for example by being constrained to follow a paved highway. Some of the examples of VANET include electronic brake lights, which allow a driver (or an autonomous car or truck) to react to vehicles braking even though they might be obscured (e.g., by other vehicles). Platooning allows vehicles to closely (down to a few inches) follow a leading vehicle by wirelessly receiving acceleration and steering information, and thus forms electronically coupled “road trains.” VANETs can use any wireless networking technology as their basis. The most prominent are short range radio technologies like WLAN (either standard Wi-Fi or ZigBee). In addition, cellular technologies or LTE can be used for VANETs. Major standardization of VANET protocol stacks is taking place in the United States, Europe, and Japan, corresponding to their dominance in the automotive industry. In the United States, the IEEE 1609 WAVE (Wireless Access in Vehicular Environments) protocol stack builds on IEEE 802.11p WLAN operating on seven reserved channels in the 5.9 GHz frequency band. The WAVE protocol stack is designed to provide multi-channel operation (even for vehicles equipped with only a single radio), security, and lightweight application layer protocols (Fig. 1).

Fig. 1 VANET – an overview

1140

G. K. Sheela

Within the IEEE Communications Society, there is a Technical Subcommittee on Vehicular Networks and Telematics Applications (VNTA). The main aim is to enable road and vehicle safety, real-time traffic monitoring, intersection management technologies, future telematics applications, and ITS-based services. In Europe, ETSI ITS G5 builds on a variant of the same radio technology with some adaptations operating on up to five reserved channels in the 5.9 GHz frequency band. The ETSI ITS G5 protocol stack is designed to provide multi-radio multi-channel operation, security, and a complex hierarchy of higher layer protocols integrating a broad range of basic services. In Japan, ARIB STD-T109 builds on a variant of the same radio technology operating on a single frequency in the 700 MHz band. The protocol stack provides TDMA operation to split use between road side services and pure vehicle-to-vehicle communication. Major issue of TDMA is too much of interference. As the number of vehicles increases the interference becomes more so that the ability to be implemented lacks in areas where the number of vehicles is more. Another algorithm that is commonly used is STDMA, but the main factor of concern of this algorithm is its reliability. So after a thorough research of the existing algorithms that is being used in this area the self-adaptive sensing model (SASM) was chosen for the proposed research work. When bulk information is being transferred, the amount of information the receiver node gets is also a main factor of concern. Thus adding a security feature to existing self-adaptive sensing model will thoroughly improve its capability for a wide range of applications. An efficient data collection algorithm is proposed for mobile vehicle network environment by taking into account the features of nodes in wireless IoVs, such as large scales of deployment, volatility, and low time delay. A self-adaptive sensing model is designed to establish vehicular data collection protocol, which adopts group management in model communication. The most favorable aim is to build the more useful, efficient, and safer roads through vehicular networks by informing basic authorities and drivers in time in the future. Another target is to implement the advancement of vehicular ad hoc networking (VANET) wireless technologies. The aim is to secure and to make possible commercial requests through range of communication systems and/or other networks which go short to medium. These technologies would support main concern for critical time secure communication and fulfill the QoS needs of other multimedia software or e-commerce mobile. The work is to add a security feature in self-adaptive sensing model and to show a better performance capability in packet delivery ratio, delay, throughput, and routing overhead. The packet delivery ratio is improved by the addition of security feature, thus ensuring secure data transmission. Since, in this the data collected are passed to a road side unit and then sent to the corresponding node, the stability will not be affected. Another main objective is data security. Data transmitted in wireless environment is vulnerable, and intruders can perform attacks like eavesdropping and active attacks like tampering, spamming, etc. VANET shall satisfy requirements like authentication of message and integrity, non-repudiation of message, authentication of entity, access control, confidentiality of message and availability, privacy, and susceptibility identification.

45 Improved Security Models in Mobile Wireless Vehicle Networks

1141

The scope of the research work is to compare algorithms and to prove that enhanced self-adaptive sensing model is the best algorithm in throughput, packet delivery ratio, delay, and routing overhead in VANET. The enhanced self-adaptive sensing model is the addition of SASM with more security through powerful encryption and decryption schemes. This approach is used to manipulate the input data and encode it by using data encoding methodology. Once encoded the approach encrypts the data and forwards it to the receiver or destination end. The data is received at the receiver end undergoes the decryption process. A proper analysis of SASM and enhanced SASM in throughput, delay, packet delivery ratio, and routing overhead also needs to be studied.

Comparative Analysis Based on the survey, various algorithms are discussed. The techniques that are used are explained in Table 1. The major challenge of WSNs is balancing of loads and lifetime of the network. A major solution for the above challenges is clustering the networks, clustering has proven to improve network lifetime. Clustering technology is widely used in different operational areas but to use it in WSNs it need to be more specific in terms of WSNs. In (He and He 2010) the authors explain the major challenges and issues on the deployment of clustering techniques in WSNs. Existing research on data collection using wireless mobile vehicle network emphasizes the reliable delivery of information. However, other performance requirements such as life cycle of nodes, stability, and security are not set as primary design objectives. This makes data collection ability of vehicular nodes in real application environment inferior. By considering the features of nodes in wireless IoV, such as large scales of deployment, volatility, and low time delay, an efficient data collection algorithm is proposed for mobile vehicle network environment. An adaptive sensing model is designed to establish vehicular data collection protocol. The protocol adopts group management in model communication. The vehicular sensing node in group can adjust network sensing chain according to sensing distance threshold with surrounding nodes. It will dynamically choose a combination of network sensing chains on the basis of remaining energy and location characteristics of surrounding nodes. In addition, secure data collection between sensing nodes is undertaken as well. It can be represented that vehicular node can realize secure and real-time data collection. Moreover, the proposed algorithm is superior in vehicular network life cycle, power consumption, and reliability of data collection compared to other algorithms. In this ever-booming technological advancement, the day to day need for IVC protocols and its applications has become inevitable. IVC protocols are widely used in different areas such as (Samara et al. 2010): 1. Peer to peer network for web surfing 2. Co-ordinated braking

1142

G. K. Sheela

Table 1 Comparative analysis of various existing techniques Author Younis et al. (2006)

Method Node clustering

Merits It aids data aggregation. It improves network lifetime

Shangguang et al. (2007)

A vertical handoff method via self-selection decision tree Type 3 application in intervehicular communication A vehicle-tovehicle communication protocol An adaptive peer-to-peer collision warning system

It supports the VHO among WAVE, WiMAX, and 3G cellular High reliability

Very high reliability

Lacks deterministic communication

Pijus Kumar Pal et al. (2014)

TDMA protocol

Share the same frequency channel by dividing the signal into different time slots

Daniel Verenzuela et al. (2015) Leung et al. (2019)

STDMA protocol

Muhammad Sameer Sheikh et al. (2020)

VANET

Based on the higher priority it is organizing Based on secured protocol reliable

Create interference at a frequency which is directly connected to the time slot length Lack of reliability

Cassell et al. (2009)

Yang et al. (2011)

Miller et al. (2013)

VANET

High delivery ratio for two lane road

Demerits This method cannot be directly implemented to WSN due to its typical operation characteristics

Delay due to deterministic protocol. Very low latency Medium latency

Remarks To implement node clustering into WSNs need specific clustering algorithm

The method can avoid the negative impact of service changes and movement changes Highly-probable information delivery to an intended group Services for which delayed information may result in compromised safety Vehicle motion planning involving global optimizations or negotiations and that may or may not involve group motion regulation Based on time slots it is allocated to each user. So when mobility speed is high it is difficult

As speed increases reliability is a major issue secure

High speed

45 Improved Security Models in Mobile Wireless Vehicle Networks

1143

3. Runway incursion prevention 4. Adaptive traffic control 5. Vehicle formations Here depending on the IVC applications, a systematic approach has been adopted to classify these protocols into different categories. As a first step to these approach with the sufficient background study. In this system, using Dynamic Source Routing Protocols to perform the successful communication between source and destination ends. The protocol adopts group management in model communication, which indicates that multiple groups are present in the network and form the grouping nodes with proper request and response methodology. The security constraint concentration falls into account via encryption and decryption schemes, with the help of these security manipulations can perform the successful data transmission between both the ends. Along with the concept proposed earlier, the attack detection strategies such as Malicious Node detection scheme are also included. Once the source node starts transmitting data, the start and end point time is mentioned and the route request is passed to the next neighbor node. If the neighbor node replies to the request within the particular point of time, then that node is considered as the next node for further precedence otherwise it is marked as a malicious node; this kind of node sensing is called selfadaptive sensing. Today technology is changing at a pacing rate. Thus it is mandatory for automobiles to become intelligent as time progress. A self-driving car is a vehicle that is capable of sensing its environment and navigating without human input. Self-driving cars can detect surroundings using a variety of techniques such as radar, lidar, GPS, odometry, and computer vision. Advanced control systems interpret sensory information to identify appropriate navigation paths, as well as obstacles and relevant signs. Autonomous cars have control systems that are capable of analyzing sensory data to distinguish between different cars on the road, which is very useful in planning a path to the desired destination. Some demonstrative systems, precursory to autonomous cars, date back to the 1920s and 1930s. The first self-sufficient (truly autonomous) cars appeared in the 1980s, with Carnegie Mellon University’s Navlab and ALV projects in 1984 and Benz and Bundeswehr University Munich’s Eureka Prometheus Project in 1987. Since then, major companies and research organizations have developed working prototype autonomous vehicles. But still this has many disadvantages in the time delay, security issues, etc. In order to overcome this we described an algorithm that overcomes the disadvantages of the previous algorithms. The algorithm proposed is self-adaptive sensing model algorithm. In truck platooning application, the lead vehicle sets the pace and communicates maneuvers to the followers. The followers attempt to maintain a constant time or distance headway to the vehicle directly in front of them. This can be done manually by a driver using, for example, a two-second following rule or remaining 50 meters behind the preceding vehicle. However, if the goal is to take advantage of draft aerodynamics for fuel economy and further increase traffic capacity, it may be desirable to tighten the vehicle spacing so much that human reaction times are

1144

G. K. Sheela

challenged. To deal with this, each vehicle collects its position coordinates and other motion data using local sensor inputs and periodically broadcasts this information to its follower. Each follower computes the relative distance, velocity, and acceleration to its predecessor and uses these in closed loop control of acceleration and braking to reduce spacing errors, etc. In addition, the lead vehicle broadcasts its motion data to all, which use it as a common reference to keep spacing errors from propagating and amplifying down the platoon’s length. The leader may also broadcast control messages that reorganize the group or announce path plan changes. An organizational approach permitted to identify the communication requirements unique to each type and focus on the most important protocol design issues facing developers Multi-channel MAC protocols has recently obtained considerable attention in wireless networking research because they promise to increase capacity of wireless networks significantly by exploiting multiple frequency bands (Wang et al. 2014). However, most of these protocols remain as pure academic interest since they only exist on paper and in simulation code but have no practical implementation. We analyzed the implementation of three representative multi-channel MAC protocols: asynchronous multi-channel coordination protocol (AMCP), multi-channel MAC (MMAC), and slotted seeded channel hopping (SSCH) on off-the-shelf IEEE 802.11 hardware. The major findings of our performance evaluation are: 1. All multi-channel MAC protocols underperform the original 802.11 MAC at low load. 2. All multi-channel MAC protocols give better performance than the original 802.11 MAC at medium and high load. 3. AMCP performs worst among all multi-channel MACs in one-hop and multihop 802.11b scenario but delivers the best performance in multi-hop 802.11a scenario. 4. SSCH attains the best results in one-hop scenarios or at low loads but loses its effectiveness at high loads in multi-hop scenarios. Implementing AMCP, MMAC, and SSCH on off-the-shelf IEEE 802.11 hardware is an ambitious endeavor. The implementation of each protocol takes an experienced kernel developer several months to complete. The major findings include: first, common techniques for clock synchronization rely on the exchange of timestamps or broadcast references. However, because broadcast frames are not protected by RTS/CTS and are susceptible to collisions, the transmission of timestamps or broadcast references is unreliable under heavy load. Further, it is very difficult to achieve synchronization accuracy finer than 1 ms using standard hardware and software implementations. We also note that time synchronization incurs additional implementation complexity and prolongs the implementation process. Thus, unless time synchronization is available via an external signal like GPS, a multi-channel MAC that does not rely on clock synchronization appears attractive.

45 Improved Security Models in Mobile Wireless Vehicle Networks

1145

Second, since channel switching delay on off-the-shelf IEEE 802.11 hardware is rather large (several milliseconds), it is important to design a multi-channel MAC protocol that avoids frequent channel switching. Otherwise, the performance gain realized by multi-channel MAC can be offset by the performance penalty incurred by frequent channel switching. On the other hand, it is desirable if hardware vendors can provide IEEE 802.11 hardware with low switching delay. Third, due to imperfect synchronization and possibly different channel switching delay, it is possible that a sender arrives at a channel earlier than a receiver. For this reason, it is recommended that the sender establishes a handshake with the receiver before starting packet transmission. Similarly, when the receiver decides to leave a channel, it is recommended that the sender is informed. Fourth, when devising a multi-channel MAC, a protocol designer should consider that a node can be simultaneously involved in multiple data exchanges with other nodes. Special attention is required to avoid lockout and to provide multiplexing between different flows at a node. This mindset is crucial in a multi-hop topology. Fifth, a multi-channel MAC generally requires that a node can overhear its neighbors channel selection. For this reason, control frames are usually transmitted as broadcast frames. However, since broadcast frames are not acknowledged and also not protected by RTS/CTS, a protocol design should be as simple as possible to avoid undesired consequences due to possible loss of control frames. Further, additional reliable mechanisms need to be provided for the exchange of control frames. Z-MAC is a hybrid MAC protocol which combines the advantages of CSMA and TDMA protocols and has enhanced in terms of bandwidth utilization compared with other WSN protocols. Z-MAC protocol switches TDMA and CSMA depending on the contention situation to use the bandwidth effectively. Z-MAC slot assignment algorithm, DRAND, was designed to adopt a node conflict resolution procedure based on randomized ODP.

L-DRAND Lamports bakery algorithm is one of mutual exclusion algorithms that is designed to prevent concurrent threads entering critical sections to eliminate the risk of data corruption. The algorithm solves the following conditions assuming N asynchronous threads: 1. At any time, at most one thread may be in its critical section 2. Each thread must eventually be able to enter its critical section (unless it halts). 3. Any thread may halt in its noncritical section The problem of the above two is slot allocations did not converged and lots of packet exchanges by consecutive request retrying in the traffic occurred during the sessions. By using the proposed scheme, priority control for nodes in the network can be performed in the MAC layer according to the collected distance measurement

1146

G. K. Sheela

information, and it can also increase efficiency for slot allocation by reducing the processing time for it, and reduce the system energy consumption, drastically.

CarTALK 2000 CarTALK 2000 is a European Project focusing on new driver assistance systems which are based upon intervehicle communication. The main objectives are the development of cooperative driver assistance systems on the one hand and the development of a self-organizing ad-hoc radio network as a communication basis with the aim of preparing a future standard (Willke et al., 2009). The main issues are: 1. Assessment of today’s and future applications for co-operative driver assistance systems 2. Development of software structures and algorithms, i.e. new fusion techniques 3. Testing and demonstrating assistance functions in probe vehicles in real or reconstructed traffic scenarios

Theory The world is moving rapidly and mobility has become an inevitable part of our life as it is used in the day to day life. Many developments and researches are going in the area of VANET. Numerous algorithms have been developed and selfautomated vehicles have also come into reality. Several algorithms are being used in VANET; after a comparison study and analysis, SASM algorithm is taken for the research study. An improvement in the existing method which improves the overall effectiveness of self-adaptive sensing model has been done. The new enhanced self-adaptive sensing model is developed with additional features of security in it. Comparative analysis of self-adaptive sensing model without security and with security on the delay, throughput, packet delivery ratio, and routing overhead was done. Signcryption is a public key cryptography that simultaneously fulfills both the functions of digital signature and public key encryption logically in one step, and with a cost significantly lower than that required by the traditional “signature and encryption” approach. Discovery of public key cryptography has made communication between people who have never met before over an open and insecure network in a secure and authenticated way possible. It enables one person to send a digitally signed message to another person and the receiver can verify the authenticity of this message (Sheikh et al. 2020). This scheme uses the private key of the sender to sign the message and the receiver uses the sender’s public key to verify the signature. Signcryption provides the properties of both digital signature and encryption schemes in a way that is more efficient than signing and encrypting separately. This approach is knows as signature then encryption. The steps that have been taken into account are as follows:

45 Improved Security Models in Mobile Wireless Vehicle Networks

1147

(i) Creation of VANET Topology: Create a Vehicular Network topology with more number of vehicles to transmit the message from sender vehicle to receiver vehicle. (ii) Implementation of self-adapting sensing model: Implementation of selfadapting sensing model by using roadside units to transmit the message from one vehicle to another vehicle. The mobile vehicles are divided into clusters with each group head as the roadside unit. The message sent from the sender vehicle and through different vehicles which are chosen based on the energy and distance of each road-side unit finally reaches the receiver vehicle. (iii) Performance Analysis of Self Adapting Sensing Model: Packet delivery ratio, throughput, delay, and routing overhead are measured for self-adapting sensing model and outputs are analyzed. (iv) Implementation of enhanced self adapting sensing model: Implementation of enhanced self-adapting sensing model by using cryptographic technique to transmit the message from one vehicle to another vehicle is done. It provides user authentication and identity-based signcryption. Only if the authentication user and password matches the entire transmission will take place. (v) Performance analysis of enhanced self-adapting sensing model: Packet delivery ratio, throughput, delay, and routing overhead are measured for enhanced self adapting sensing model is measured and outputs are analyzed. (vi) Comparison: Self-adapting sensing model and enhanced self-adapting sensing model are compared using parameters such as packet delivery ratio, throughput, delay, and routing overhead and outputs are analyzed.

System Architecture The basic algorithm of the proposed enhanced self-adaptive sensing model is shown in Fig. 2. There are 25 nodes in the network. Assume a source point and destination point. The main goal is to send packets to the destination. The shortest path is measured after the information about the path for source and destination. After the transmission ranges are identified the packets are sent to the neighbor nodes. The packets are sent to destination taking into account maximum strength of neighbor nodes. Then, check whether the neighbor node strength is poor or good. Check for an alternate, if it is poor, repeat the steps again. If it is good, encrypt the data packets and then forward the packets. Finally, packets reach the destination and the packets are decrypted. Hence the process has been completed with successful transmission. Implementation of enhanced self-adaptive sensing model is performed by using cryptographic technique. The cryptographic technique used is signcryption. In order for the transmission to take place a username and password is inserted. If the username and password matches with source file username and password when logging in, then the transaction will take place if not a command called unauthorized user will be executed. The system has been designed through various error and trial methods. Initially the performance was tried through cygwin software. But it faced many issues such as it could be performed only through various theoretical equations and the

1148

G. K. Sheela

Check for Alternate Node and Path and try to retransmit

Found the data transmission range to send packets

Source and Destination pairs are fixed

Nodes {1-25}

Poor

Check

Verifying Node Strength

Neighbor Nodes Fixed

Good

Encrypt data packets

Forward the packets

Destination Received the Packets and Start Decryption

Data Received in Destination with Decrypted Structure

Process Completed with successful transmission of packets with Decryption procedures

Fig. 2 Flowchart of proposed enhanced SASM

graph comparison through trace file was also not possible. Then ns-2.35 Linux version was installed. There trace file could be done. The encryption technique studied was Caesar Cipher. It was a shifting algorithm technique, which was a basic technique with not much security. This was because the cipher text used was a shifting alphabet. But the alphabet has only 26 letters. And so with some different arrangements the attacker can easily identify the encrypted file and take the message. Finally signcryption method was tried.

Analysis The simulation of the research has been performed through ns-2.35. The VANET topology has been developed and the model realized is self-adaptive sensing model. The performance analysis based on self-adaptive sensing model and proposed enhanced self-adaptive sensing model is done using throughput, packet delivery ratio, delay, and routing overhead. For simple VANET topology the AODV pro-

45 Improved Security Models in Mobile Wireless Vehicle Networks

1149

tocol is used. This chapter mainly describes about the simulation tool, boundary parameters, node movements, and performance analysis. A sequence diagram is an interaction diagram that shows how objects operate with one another and in what order. It is a construct of a message sequence chart. A sequence diagram shows object interactions arranged in time sequence. It depicts the objects and classes involved in the scenario and the sequence of messages exchanged between the objects to carry out the functionality of the scenario. Sequence diagrams are typically associated with use case realizations in the Logical View of the system under development. Sequence diagrams are sometimes called event diagrams or event scenarios. A sequence diagram shows, as parallel vertical lines (lifelines), different processes or objects that live simultaneously, and, as horizontal arrows, the messages exchanged between them, in the order in which they occur. This allows the specification of simple runtime scenarios in a graphical manner. Source check the Neighbor Node and start communication with that node until it transmits Route Response to Source Node. If Route Response is established, it indicates that node is not a Malicious Node. It finds the next shortest path. This process is repeated until it reaches the destination. Finally packets are sent to the destination. After the response is received, finally the acknowledgment is sent to the source. Since it is not a malicious node, the encrypted data from the source is decrypted at the destination. The packets are moving from the source to destination through the self-adaptive sensing model routing protocol. According to the protocol, the fixed vehicles nodes send message Staticnode_Msg to surrounding nodes. Mobile vehicle node receives message Staticnode_Msg and calculates the distance. Mobile vehicles send message Mobilenode_Msg to the closest sensing node. All mobile vehicle nodes can find their sensing nodes. After selecting a fixed vehicle node as the main reference point, other common nodes are classified into one group. Each fixed vehicle node will compare to mobile sensing node and then broadcast message Cluster_Msg. After some time, fixed vehicle nodes receive message Mobilenode_Msg from neighboring mobile sensing nodes. Now the IDs of mobile vehicle nodes within the group, the remaining energy and sensing distance can be known by the fixed node. Mobile nodes can store the information into local neighbor table. Fixed vehicle node selects a high remaining energy from the neighbor table and finds the closest sensing node. Sensing vehicle node is selected based on remaining energy and position. However, common mobile vehicle node will select a closest fixed node as main reference point. After the group is classified, the node can send data to specific group head. Boundary parameters are shown in Table 2. The MAC_802.11 protocol is used. This is used for the short range of communication, and this protocol is mainly used for the VANET topology. The simulation time is 30 s. Adjustment of simulation time is possible according to requirement. Assume a simulation area of 1000*1000 m2 . Within this area the static nodes and mobile nodes are provided. The constant bit rate (CBR) means that the rate at which a codec’s output data should be consumed is constant. The traffic is transmission control protocol (TCP) which is a dynamic reliable congestion protocol that is used to provide reliable transport of packets from one host to another host by sending

1150 Table 2 Boundary parameters

G. K. Sheela Type Simulator Channel type Antenna type Link layer type Mac type Protocols studied Simulation time Simulation area Traffic type Data packet size (byte) Mobile, Init _ Energy (J) Graph

Parameter Ns-2.35 Wireless channel Omni antenna LL Mac/802_11 AODV, SASM 30 s 1000*1000 m2 CBR (TCP) 812 100 (max), 20 (min) Gnuplot

acknowledgments on proper transfer or loss of packets. TCP requires bidirectional links in order for acknowledgments to return to the source. The data packet size is 812 bytes. The maximum energy for the mobile nodes is 100 Joules. The energy decreases as the node is used for transmission. The plotting software used for analysis is Gnuplot. In proposed enhanced self-adaptive sensing model, the entire packet transmission takes place only after the proper authentication. This means that there is a public key that is known to the nodes where transmission takes place. A comparison takes place based on the private key and public key. When the comparison matches, the authentication activates and the encrypted message will be sent and finally gets decrypted at the receiver side. Taking the different parameters such as delay, packet delivery ratio, average throughput, and routing overhead into consideration, the effect of enhancement to self-adaptive sensing model can be clearly studied. Transmission delay plays a vital role in effective data communication; if the delay is more, it means that efficient data transmission is not taking place. Delay is calculated using awk script which processes the trace file and produces the result. Delay is the difference between the time at which the sender generated the packet and the time at which the receiver received the packet. Now moving on to analyze the delay time in self-adaptive sensing model with and without security, for a smaller simulation time there is a difference of about 15,000 ms. As the simulation time increases, the difference in delay is about 4000 ms. This is because as the simulation time increases, it is becoming more dynamic and so the delay reduces. The conclusion is that adding the security feature in self-adaptive sensing model reduces the delay time for data communication. In Fig the graph compares the delay time with the simulation time for self-adaptive sensing model (SASM) and enhanced self-adaptive sensing model (ESASM). The performance of routing protocol in mobile vehicle network is simulated on the basis of urban street model. Life Cycle in data collection of mobile vehicle nodes is compared with TDMA, STDMA, and ACMAP.

45 Improved Security Models in Mobile Wireless Vehicle Networks

1151

Comparison with different algorithms is performed and is represented by using a graphical representation. Comparison of lifecycle in data collection for mobile vehicle nodes is plotted with TDMA, STDMA, and ACMAP algorithms with x-axis as number of rounds and y-axis as number of vehicle nodes. In ACMAP algorithm, all nodes communicate within the range of d0. In this case, energy consumption problem in data transmission can be addressed. Data forwarding in vehicle network are realized by awareness, which saves much energy by comparing to one-way communication. However, since STDMA protocol does not consider relationship between energy attenuation and communication distance in data transmission, energy of node is easy to use up in short time. In STDMA, sensing nodes require more energy to form sensing group and record information of each member node. However, in ACMAP algorithm, mobile vehicle nodes are responsible for collecting sensing information and sending the fused information to fixed node. Other tasks such as selection of sensing nodes, TDMA time slots partition, are finished by fixed nodes which have more energy. Therefore, energy consumption can be reduced, prolonging network lifetime. Recent news illustrate the frequent occurrence of pileup crashes on highways. A predominant reason for the occurrence of such crashes is that current vehicles (including those equipped with an automatic cruise control system) do not provide drivers with advance information of events occurring far ahead of them (Sato and Sakata 2013). The use of intervehicular communication to provide advance warnings to enhance automotive safety is therefore being actively discussed in the research community. Since wireless communication delays are on the order of milliseconds, they can be considered small, compared with human reaction time delays; therefore, information of a slowdown is near, which is simultaneously propagated to all the cars within the communication range, addresses the specific issues of car pileup crashes and their alleviation. The goal of this chapter is, therefore, to describe how the slowdown warning system enables pileup crashes to be averted, even in a mixedsensing environment, wherein only a limited number of vehicles are equipped with the warning system. Thus in general the concept of a slowdown warning system can be briefed as follows, while driving on a highway, if a car in a platoon abruptly decelerates, then such a system near simultaneously provides advance information to all the drivers behind that car. This advance information gives these drivers more time to react in preparation for the impending slowdown and, accordingly, increases their headway to the car ahead. This increased headway can alleviate collisions, particularly pileup crashes. Furthermore, it has theoretically been shown that there exist conditions wherein even a partial equipage with such a system can be sufficient to alleviate crashes, even in the unequipped cars. This is because each equipped car, with its increase in headway, acts as an attenuator that arrests the amplification of the velocity perturbation of the lead car as it propagates through the line of cars; therefore, even with a few (appropriately distributed) equipped cars, it is possible to keep the level of amplification below the threshold that leads to car pileups. Sufficient conditions governing the requisite number and distribution of

1152

G. K. Sheela

the equipped vehicles that are required to avert pileup crashes are discussed. A prototype of such a system has been developed and installed in a few cars, and experimental road tests have been conducted. These road tests have confirmed the satisfactory work of the slowdown warning equipment. Vehicular ad-hoc network is an emerging research area focusing on communication infrastructures that support vehicles and road-signs in distributing road state data such as information about hazardous road conditions ahead, approaching emergency vehicles, and traffic delays (Hubaux et al. 2004). Vehicular ad-hoc networks combine the areas of sensor networks (data acquisition) with mobile ad-hoc networks (highly dynamic topology and lack of preexisting infrastructure). One of the main challenges of vehicular ad-hoc networks is the data dissemination protocols capable of distributing road-state information among vehicles. This chapter presents two candidates for dissemination protocols: a zone flooding protocol and a zone diffusion protocol.

Zone Flooding Protocol The zone flooding protocol is a variant of basic flooding with three modifications to limit the dissemination of packets. It can be seen as a special case of floodingbased geocasting in the sense that the source is located inside the geocast zone. Traditionally the problem with flooding-based protocols is that they congest the network with hordes of packets.

Zone Diffusion Protocol The zone diffusion protocol is based on data aggregation which is a commonly used technique in sensor networks. Each node maintains an environment representation (ER) representing the surrounding environment. The ER is updated every time data arrives from the sensors. To disseminate data the ER is periodically broadcasted. When an ER is received from another node it is aggregated with the local ER by merging the information in the received ER that intersects with the area covered by the local ER. Contrary to the zone flooding protocol, packets are never forwarded. However, data about the local environment is indirectly forwarded to other nodes since nodes periodically broadcast their ER. The protocol is thus data-centric as opposed to node-centric. The protocols only rely on the assumption that the relevance of information about a particular phenomenon decreases with the distance to that phenomenon and can therefore be used in typical vehicular ad-hoc network applications. The conclusion is that flooding protocol generally achieves better awareness percentage and information distance than the zone diffusion protocol, but zone diffusion protocol achieves reasonable performance at a much lower network utilization. By a crude form of congestion containment, the zone flooding protocol adaptively decreases the size of the area in which data is disseminated when the networks get

45 Improved Security Models in Mobile Wireless Vehicle Networks

1153

congested. The zone diffusion protocol could be improved by adding congestion control – either by limiting the dissemination area or by decreasing the number of broadcasts per second.

Dynamic Time-Stable Geocast Routing in Vehicular Ad-Hoc Networks This is a novel time-stable geocast protocol that works well even in too sparse networks. Moreover, since commercial applications sometimes make it necessary to change the duration of the stable message within the region, the dynamic nature of a geocast protocol should allow this time to be extended, reduced, or canceled without any additional cost. Therefore, we call it a dynamic time-stable geocast, DTSG, protocol (Liu et al. 2006). It works in two phases (the pre-stable period and the stable period), and the simulation results show that it works well in its performance metrics (delivery ratio and network cost). DTSG is a dynamic time-stable geocast protocol that guarantees delivery of the message to the intended vehicles entering the region for a certain amount of time. This time can be expanded or can be canceled by the dynamic characteristic of this protocol. This is done by the supplementary message without any new effect on the network cost.

Chain Collisions of Vehicles Equipped with Vehicular Communications Improvement of traffic safety by cooperative vehicular applications is one of the most promising benefits of vehicular ad hoc networks (VANETs). However, to properly develop such applications, the influence of different driving parameters on the event of vehicle collision must be assessed at an early design stage. The model enables the computation of the average number of collisions that occur in the platoon, the probabilities of the different ways in which the collisions may take place, as well as other statistics of interest (Liang et al. 2016). Although an exponential distribution has been used for the traffic density, it is also valid for different probability distributions for traffic densities, as well as for other significant parameters of the model. Moreover, the actual communication system employed is independent of the model since it is abstracted by a message delay variable, which allows it to be used to evaluate different communication technologies. The goal is to describe and analyze the risk of colliding for a set of moving vehicles forming a platoon (or chain) and equipped with a warning collision system when the leading vehicle stops suddenly. The main practical utility of this model lays in its ability to quickly evaluate numerically the influence of different parameters on the collision process without the need to resort to complex simulations in a first stage. Such an evaluation provides relevant guidelines for the

1154

G. K. Sheela

design of vehicular communication systems as well as chain collision avoidance (CCA) applications. This approach proposed and derived a stochastic model for the probability of collisions in a chain of vehicles where a warning collision system is in operation. The fact that a warning notification system is used allows us to overcome the difficulties for obtaining stochastic models for such vehicular scenarios, since we can assume that all the drivers/vehicles react to the warning message independently, and therefore, the motion equations can be simplified (Ylianttila et al. 2001). We also propose a good matching approximation to the exact model to further reduce the required computations to calculate the vehicle collision probabilities. In both cases, its validity has been confirmed by Monte Carlo simulations. The model is independent of the particular communication system employed as long as its operation can be abstracted and characterized by an appropriate message notification delay, including communication latency and driver reaction times. Therefore, it also enables the performance evaluation of different technologies. Indeed, a future line of this work is to assess the performance of current VANET technology based on contention (CSMA) MAC protocols for those cases where delay is actually relevant for the collision process outcome. Similarly, different probability distributions for the intervehicular spacing can seamlessly be incorporated into the model due to the fact that the distribution of the initial intervehicle spacing is independent of the actions that drivers make after receiving the warning messages. Here, we have used an exponential distribution, which is considered appropriate for low vehicle traffic densities. As a future work, we plan to employ a lognormal distribution that describes well high vehicle traffic densities. Finally, we compute the probability that collisions occur in different forms (both vehicles in motion, one stopped and one in motion, etc.), which opens a promising way to define detailed accident severity functions, that is, by assigning different grades of severity to each collision possibility. This is an interesting approach that we leave as future work as well. Although we have shown some examples of the application of the model, a quantitative evaluation requires a careful definition of the scenarios of interest. Therefore, we leave as future but imminent direction to pursue a systematic characterization and evaluation of the different scenarios for a wider and more accurate extent of the model parameters.

A Vertical Handoff Method via Self-Selection Decision Tree This method establishes the respective handoff probability distribution of vehicles according to network attributes and movement trend. Then, based on handoff probability distributions and defined user preferences, we propose a novel handoff method by the self-selection decision tree for IoVs (Chen et al. 2006). Finally, we also present a feedback decision method according to the feedback of vehicle handoff, to improve next handoff quality when vehicle movement trend and vehicle service status change.

45 Improved Security Models in Mobile Wireless Vehicle Networks

1155

Fig. 3 Vehicle motion state

IoV allows the vehicle to share Internet access to other devices both inside and outside the vehicle. Often, in IoVs, the vehicle is outfitted with special technologies that tap into the Internet access or the WLAN and prov. Figure 3 shows the diagram of vehicle motion state. Existing unstable network statuses and different user preferences result in vehicle frequent VHOs. In order to solve the issue, a VHO method based on a self-selection decision tree for IoVs can be used. Main features include: 1. A terminal states of motion in the handoff decision 2. A handoff method based on a self-selection decision tree for VHO among WAVE, WiMAX, and 3G cellular 3. A feedback decision method, which makes the next handoff timely and accurate for IoVs The decision tree makes decision according to user preferences, and the feedback decision method in line with the feedback of services and movements on vehicles can avoid the negative impact of service changes and movement changes (Younis et al. 2006). The method reflects the specific needs of vehicle to the network. Moreover, there may be some other network attributes and user preferences that were not taken into account. In addition, the proposed decision tree method can be further optimized. A topological analysis of urban transit system gives a functional view on nodes named a transit line. Statistical measures are computed and introduced in complex network analysis. It shows that the urban transit system forms small-world networks and exhibits properties different from random networks and regular networks. The urban transit network is a complex network, in which the nodes can be seen as transit sites and links corresponding to the routers linking O-D7. A dual approach is adopted based on the functional view. Nodes are defined as named transport lines and links represent the convenience of transfer. The convenience

1156

G. K. Sheela

Fig. 4 Illustration of dual approach

of transfer here means which nodes are named transit line and links represent the convenience of transfer (Willke et al. 2009). A range of statistical measures are computed for structural analysis introduced in complex network analysis. It is shown that the urban transit system forms small-world networks and exhibits properties different from random networks and regular networks. Furthermore, the topological properties of the transit-line network from the points of view of public transportation engineering are investigated to get some useful conclusions (Figs. 4 and 5). The aim of the functional representation of urban transit network is to investigate the inherent topological properties of the relations between transit lines, and it indicates the importance of loop lines in the whole network.

Mobile Location Estimator Using Extended Kalman-Based IMM and Data Fusion An extended Kalman-based interacting multiple model (EK-IMM) smoother is proposed for mobile location estimation with the data fusion of the time of arrival (TOA) and the received signal strength (RSS) measurements in a rough wireless environment. The extended Kalman filter is used for nonlinear estimation (Pal and Chatterjee 2014). The IMM is employed as a switch between the line-of-sight (LOS) and non-LOS (NLOS) states, which are considered to be a Markov process with two interactive modes. Combining extended Kalman filtering with the IMM scheme for accurately smooth range estimation between the corresponding base station (BS) and mobile station (MS) in the rough wireless environment, the proposed robust mobile location estimator, in association with data fusion, can efficiently mitigate the NLOS effects on the measurement range error.

45 Improved Security Models in Mobile Wireless Vehicle Networks

1157

COMPARISON 20000

SASM ESASM

18000 16000

Delay(ns)

14000 12000 10000 8000 6000 4000 2000

10

15

20 Simulation time (seconds)

25

30

Fig. 5 Simulation time verses average end-to-end delay

However, the accuracy of mobile location estimation is still a very difficult problem for any type of estimation method used alone in the urban region because the wireless environment is rough. To solve this problem, mobile location estimation in association with the data-fusion technique has recently been proposed to improve accuracy. Special antennas are needed if the AOA technology is used. It is expensive and not easy for real implementation. The data fusion of RSS and TOA estimation can simultaneously be made without requiring any additional hardware components. A Kalman filter is used for mobile location estimation. The identification of LOS/NLOS condition can be attained by a simple hypothesis that the standard deviation of range measurement in the NLOS case is significantly larger than that in the LOS case. However, the LOS/NLOS transition will cause a serious measurement error for the range estimation because the estimated covariance matrices of the measurement noise by the corresponding Kalman filter are not adaptively adjusted to match the true covariance variation in the LOS and NLOS cases. The IMM estimator is one of the most effective methods for estimation in a hybrid dynamic system under uncertain environmental conditions. It can accurately estimate the state of a dynamic system with several available switching modes. In particular, it has a self-adjusting variable bandwidth filter that makes it suitable for switching-mode systems. The major challenge is the computation complexity which is relative to the total number of multiplications for the algorithm. Based on the aforementioned statement, the total number of multiplications for the proposed algorithm is about

1158

G. K. Sheela

98 in each iteration. The complexity of the conventional extended Kalman filter algorithm is about 86 in each iteration. The computation complexity of the proposed algorithm is increased by about 14%.

Scalability of Vehicle-to-Vehicle Communication with Prediction- based STDMA Self-organized TDMA (STDMA) has been proposed as a medium access control (MAC) protocol for vehicle-to-vehicle (V2V) communication. Although it avoids the unbounded channel access delay faced by other CSMA-based protocols, the reliability in high traffic density is still unsatisfactory for critical road safety applications. An enhanced prediction-based STDMA protocol utilizes the spatial information available in the messages exchanged for road safety applications. Thus it allocates the radio resources more efficiently and minimizes the mutual interference among different vehicles (Jiun et al. 2011). The prediction-based STDMA offers a significant improvement in the scalability and coverage of the V2V communication system for road safety applications with strict reliability requirement. With CSMA, all vehicles must listen to the wireless channel before transmitting, and if the channel is busy, the nodes must defer their access. To overcome the channel access delay problem, we proposed to utilize self-organizing time division multiple access (STDMA) as a replacement of the CSMA-based MAC.STDMA has a synchronized time-frame structure and allows multiple vehicles to transmit in the same time slot thus limiting the channel access delay. When a vehicle needs to select a time slot for reuse, it selects the one that is being used by the vehicle currently located furthest away. This selection is done to minimize the mutual interference caused by simultaneous transmissions. Assume all vehicles are equipped with transceivers for real-time communications. The data transmissions consist of periodical broadcasting of cooperative awareness messages (CAM) to nearby vehicles. Each CAM contains status information such as position and velocity. Due to the broadcast nature of transmissions, these messages are received by the all vehicles within the coverage area. In order to cement on effective data transmission the packet delivery ratio gives a clear idea so during simulation carried out the comparative analysis of Packet delivery ratio verses simulation time. Packet delivery ratio is the ratio of packets that are successfully delivered to a destination compared to the number of packets that have been sent by sender. It is calculated using awk script which processes the trace file and produces the result. In the graph it clearly indicates that red line is the self-adaptive sensing model and green line indicates the enhanced self-adaptive sensing model. Now moving on to analyze the packet delivery ratio in self-adaptive sensing model with and without security, for a smaller simulation time there is a difference of about 0.2582. And as the simulation time increases the difference in packet delivery ratio is about 0.2051. In this also 67% of packet could be sufficiently delivered compared to the 38% packet delivery of SASM. There are enough packet losses; because the simulation time increases it is becoming more dynamic and so

45 Improved Security Models in Mobile Wireless Vehicle Networks

1159

COMPARISON 0.75

SASM ESASM

0.7 0.65

PDR

0.6 0.55 0.5 0.45 0.4 0.35 10

15

20 Simulation time (seconds)

25

30

Fig. 6 Comparison of PDR versus simulation time

the packet delivery ratio also reduces. The conclusion is that adding the security feature in self-adaptive sensing model improves the packet delivery ratio for data communication. The comparison graph is shown in Fig. 6. Figure 7 shows the comparison of the throughput in self-adaptive sensing model and enhanced self-adaptive sensing model. Throughput is the number of successfully received packets in a unit time and is represented in bps. Throughput is calculated using awk script which processes the trace file and produces the result. Throughput can be calculated by dividing the file size by the difference between the stop time and start time. When throughput is more it means that much file is going at that particular time which shows an improvement. Routing overhead is the number of routing packets required for network communication. How much loads are transmitted, that is requested message, reply, finding the channel, how many signals, sensing data. All the above measures can be known through the routing overhead. When the simulation time is less, there is a routing head difference of 1.552, whereas as the simulation time increases, there is a routing difference of 0.385 between the two models. As simulation time increases the dynamic will be more and efficiency reduces in a small way. But still the enhanced self-adaptive sensing model improves the routing overhead by an average of 1.187 whereas the average routing overhead of self-adaptive sensing model is 2.420 (Fig. 8).

1160

G. K. Sheela COMPARISON

400

SASM ESASM

Throughput (kbps)

350

300

250

200

150

100

10

15

20 Simulation time (seconds)

30

25

Fig. 7 Comparison of throughput versus simulation time

COMPARISON 2.6

SASM ESASM

2.4 2.2

Routing overhead

2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 10

15

20 Simulation time (seconds)

Fig. 8 Comparison of routing overhead verses simulation time

25

30

45 Improved Security Models in Mobile Wireless Vehicle Networks

1161

Table 3 Analysis of SASM and proposed enhanced SASM Parameters Average end-to-end delay (ms) Average throughput (kbps)

Packet delivery ratio (PDR) Routing overhead (RO)

SASM 19348.2

Enhanced SASM 3549.19

Start time = 0 s Stop time = 30 s Received size = 632,320 bytes Throughput = 168.62 kbps

Start time = 0 s Stop time = 30 s Received size = 757,952 bytes Throughput = 202.12 kbps Received packets = 911 bits Sending packets = 1359 bits PDR = 0.6703 Routing packets received = 3947 bits Data packets received = 3326 bits RO = 1.187

Received packets = 760 bits Sending packets = 2038 bits PDR = 0.3729 Routing packets received = 1839 bits Data packets received = 760 bits RO = 2.420

A comparative analysis table based on self-adaptive sensing model and enhanced self-adaptive sensing model with parameters such as average end-to-end delay, packet delivery ratio, throughput, and routing overhead is shown in Table 3. From the table, it can be easily understood that proposed enhanced self-adaptive sensing model is better than self-adaptive sensing model.

Conclusion In the recent years VANET has become an active area in the field of communication. Day by day new technologies are coming into the mainstream. Numerous algorithms have been developed for VANET and it is on the advancement to find the best protocol which will be very useful for real-time applications and will be of great help to the civilization. Thus the scope of work included the performance analysis and evaluation of self-adaptive sensing model and enhanced self-adaptive sensing model using ns-2.35. The proposed enhanced SASM is done based on security, that is, signcryption. The simulation model was created and the algorithms code for selfadaptive sensing and the proposed enhanced self-adaptive sensing algorithm was implanted. The results demonstrate the capability of self-adaptive sensing model. The parameter that has been used for analysis and comparative studies include throughput, average end-to-end delay, packet delivery ratio, and routing overhead. The enhanced self-adaptive sensing model has got an additional security feature. The security feature has been implemented using signcryption technology. The comparative study and analysis in the parameters such as throughput, delay, packet delivery ratio, and routing overhead has been done. It has been proved that the delay and routing overhead is less and the packet delivery ratio and throughput is better for enhanced self-adaptive sensing model. The analysis and simulation

1162

G. K. Sheela

results conceived that enhanced self-adaptive sensing model is better than selfadaptive sensing model. As a future scope, synchronization into data collection is a possibility in mobile vehicle network. The requirements of synchronization for network nodes need to be reduced.

References Chen LX, Zeng X, Cao Y (2006) Mobile Ad Hoc networking (MANET) – technology of MANET, pp 205–220 Fengzhong Q, Wu Z, Wang F-Y, Cho W (2015) A security and privacy review of VANETs, IEEE Trans Intell Transp Syst 1–9 He Z, He Y (2010) Study on key technologies of MANET. In: International conference of information science and management engineering, pp 112–115 He Y, Yu FR, Wei Z, Leung V (2019) Trust management for secure cognitive radio vehicular ad hoc networks. AdHoc Netw 86:154–165 Horng S, Tzeng S-F, Pan Y (2013) B-specs+: batch verification for secure pseudonymous authentication in Vanet. IEEE Tran Inf Forensics Secur 8(11):1860–1875 Hubaux J, Capkun S, Epfl J (2004) The security and privacy of smart vehicles. IEEE Secur Priv 2(3):49–55 Jiun H, Lo Y, Hung C (2011) ABAKA: an anonymous batch authenticated and key agreement scheme for value-added services in vehicular Ad Hoc networks. IEEE Trans Veh Technol 60(1):248–262 Johnson D, Menezes A, Vanstone S (2001) The elliptic curve digital signature algorithm (ECDSA). Int J Inf Secur 5:36–63 Liang W, Zhiqiang RUAN, Mingdong TANG, Peng LI (2016) A secure efficient data collection algorithm based on self-adaptive sensing model in mobile internet of vehicles. 52:121–129 Lin X, Sun X, Ho P, Shen X (2007) GSIS: a secure and privacy preserving protocol for vehicular communications. IEEE Trans Veh Technol 56(6):3442–3456 Liu M, Li ZC, Guo XB, Dutkiewicz E, Zhang DK (2006) Performance evaluation of vertical handoff decision algorithms in heterogeneous wireless networks. IEEE Syst J 5:1–10 Ming W, Dan L, Lie Z, Yong X, Fei W (2014) LESPP: lightweight and efficient strong privacy preserving authentication scheme for secure VANET communication. Computing 1:133–139 Pal PK, Chatterjee P (2014) A survey on TDMA-based MAC protocols for wireless sensor network. Int J Emerg Technol Adv Eng 4(6):43–50 Papadimitratos P, La Fortelle A, Evensen K, Brignolo R, Cosenza S (2009) Vehicular communication systems: enabling technologies, applications, and future outlook on intelligent transportation. IEEE Commun Magazine 47:553–559 Raya M, Hubaux J (2005) The security of vehicular ad hoc networks. Laboratory for computer Communications and Applications (LCA), pp. 11–21 Samara G, Salihy W, Sures R(2010) Security analysis of vehicular ad hoc networks. In: Second international conference on network applications, protocols and services, pp 55–60 Sato K, Sakata S (2013) A power-efficient distributed TDMA scheduling algorithm with distancemeasurement for wireless sensor networks. Wireless Pers Commun 75:1012–1017 Sheikh MS, Liang J, Wang W (2020) Security and privacy in vehicular Ad Hoc network and vehicle cloud computing: a survey. Wirel Commun Mob Comput 2020 Studer A, Bai F, Bellur B, Perrig A (2009) Flexible, extensible, and efficient VANET authentication. J Commun Netw 11(6):574–588 Verenzuela D, Liu C, Lu W, Shi L (2014) Improving scalability of vehicle-to- vehicle communication with prediction-based STDMA. Vehic Technol Conf IEEE 58:1–5 Verma A, Saha R, Kumar G (2021) The security perspectives of vehicular networks: a taxonomical analysis of attacks and solutions. Appl Sci 11:1–25

45 Improved Security Models in Mobile Wireless Vehicle Networks

1163

Wang S, Fan C, Hsu C-H, Sun Q, Yang F (2014) A vertical handoff method via self-selection decision tree for internet of vehicles. IEEE Syst J 2:99–105 Willke T, Tientrakool P, Maxemchukv N (2009) A survey of inter-vehicle communication protocols and their applications. IEEE Commun Surv Tutorial 11:1–6 Ylianttila M, Pande M, Makela J, Mahonen P (2001) Optimization scheme for mobile users performing vertical handoffs between IEEE 802.11 and GPRS/EDGE networks. Glob Telecommun Conf 50:120–125 Younis O, Krunz M, Ramasubramanian S (2006) Node clustering in wireless sensor networks: recent developments and deployment challenges. IEEE Netw 20:20–25

Machine Learning: An Expert Thinking System

46

T. Mohanraj, Jayanthi Yerchuru, R. S. Nithin Aravind, and R. Yameni

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Factory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cyber-Physical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Features of CPS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Components of CPS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Supervised Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Unsupervised Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reinforcement Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Support Vector Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Naive Bayes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . K-Nearest Neighbors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Logistic Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Decision Tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Random Forest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . K-Means Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hierarchical Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mean Shift Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Artificial Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hidden Markov Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications of ML in Condition Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1166 1168 1170 1171 1173 1173 1175 1175 1176 1177 1178 1178 1178 1178 1179 1179 1179 1179 1179 1180 1180 1182 1182

T. Mohanraj () · J. Yerchuru · R. S. N. Aravind · R. Yameni Department of Mechanical Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_29

1165

1166

T. Mohanraj et al.

Abstract

Industry 4.0 makes it believable to collect and investigate data across machines, aiding more efficient and flexible processes to manufacture the parts with high quality at a low cost. The technologies which enable this are digital twin, big data analytics, autonomous robots, Internet of things, cybersecurity, cloud computing, augmented reality, and additive manufacturing. Thus, interconnected intelligent machines allow autonomous manufacturing using decentralized decision-making systems that cooperate with each other, making the manufacturing process more efficient. Machine maintenance can be categorized into three types, namely, predictive maintenance (supervised), run to failure (semi-supervised), and preventive maintenance (unsupervised). Self-diagnostic machines are an integral part of smart factories. Predictive maintenance is a proactive maintenance strategy that predicts failure. These predictions are based on data gathered through condition monitoring sensors using IoT, analyzed using big data, and predicted using machine learning algorithms. This can lead to major cost savings and increased availability of the systems, thus optimizing performance. Keywords

Industry 4.0 · Supervised and unsupervised maintenance · Preventive maintenance · Predictive maintenance · Run to failure · Cyber-physical system · IoT · Big data

Introduction A system is said to learn if it advances its future performance after diligently observing and analyzing the present conditions. Why should the system learn? This is because the designer cannot anticipate and program solutions for all the future situations the system might face. So, it has to adapt to the present and future conditions based on its past experiences. Industry 4.0 makes it believable to acquire and examine the data across machines to enable faster, efficient, and flexible processes, eventually producing parts with high quality at a low cost. Within this scenario, industrial production management envisages the use of big data analytics (automatized processing of large data arrays), cloud technologies (to store information in the virtual environment, reducing the risk of loss during hardware and software failures), the Internet of things, cybersecurity, digital twin, 3D printing, and automatized production (autonomous manufacturing using decentralized decision-making systems, making the manufacturing efficient), thus reducing human intervention to simply monitor and control the work. The future outline of Industry 4.0 is shown in Fig. 1. The typical illustration of an expert system is shown in Fig. 2. It consists of an editing module, learning module, inference engine, knowledge base, memory storage, and interface for input and output devices. The expert system has a

46 Machine Learning: An Expert Thinking System

1167

Fig. 1 Future outline of Industry 4.0

Fig. 2 Simple expert system

significant role in Industry 4.0. True to these words, the current world leaders have realized the significance of Industry 4.0 in development. The leading example is that of Germany, the first nation to adopt and implement it in 2012. The UK has followed Germany’s footsteps and has adopted Industry 4.0 as the leading sphere of growth for the industry and envisages the “eight great technologies” (Sergi et al. 2019), which is shown in Fig. 3.

1168

T. Mohanraj et al.

Fig. 3 Eight great technologies

The typical applications of the expert system are shown in Fig. 4 (Expert system 2020). The applications include prediction (Woolery and Grzymala-Busse 1994), condition monitoring (Zhu et al. 2020), interpretation (Zalazinsky et al. 2020), design analysis (Purnomo and Hidayatuloh 2020), process planning (Zalazinsky et al. 2020), monitoring and control, enterprise resource planning (Mondal et al. 2020), material resource planning (Perera 2020), inventory control (Harifi et al. 2020), etc.

Smart Factory The revolution from conventional manufacturing to smart manufacturing conspiracies plans the long-term benefits of the manufacturing sector at the global level. The concept of Industry 4.0 introduced in the manufacturing industry is to advance the

46 Machine Learning: An Expert Thinking System

1169

Fig. 4 Applications of expert system

process and reduce the processing time. As a core of Industry 4.0, smart factory assimilates the physical as well as cyber technologies besides producing novel techniques to enhance the performance, quality, and controlling ability and to reduce the manufacturing lead time (Shi et al. 2020). The term “smart factory” has been introduced by academic and industrial experts from several viewpoints (Bicocchi et al. 2019). Many of the definitions relate to the intelligent manufacturing and cyber-physical system (CPS). It is an integrated manufacturing system connected by a network and collects the entire data about the manufacturing process in real time over the Internet and automatically fine-tunes the machining parameters/methods or changes the raw materials to enhance the features of the production system dynamically (Park 2016). A smart factory is a manufacturing CPS that assimilates the physical systems like machinery, belts and conveyors, and processed parts with information systems, such as material resource planning (MRP), planning control (PC), and enterprise resource planning (ERP) to device flexible, just-in-time, as well as agile manufacturing system (Wang et al. 2016). Conferring to (Chen et al. 2017), it is an intelligent manufacturing system that integrates the data-sharing process, computation process,

1170

T. Mohanraj et al.

Fig. 5 Typical smart factory

and process control in the industry to face the current needs of the industry. A typical layout of the smart factory is shown in Fig. 5. Machine maintenance can be categorized into three types, namely, predictive maintenance (supervised) (Naren and Subhashini 2020), run to failure (semisupervised) (Vercruyssen et al. 2020), and preventive maintenance (unsupervised) (Cerquitelli et al. 2019). Self-diagnostic machines are an integral part of smart factories. Predictive maintenance is a proactive maintenance strategy that predicts failure. These predictions are based on data gathered through condition monitoring sensors using IoT, analyzed using big data, and predicted using machine learning algorithms. This leads to reducing the cost and improving the availability of the systems, thus optimizing performance. “Intelligent agent” is a device that observes its environment and takes corrective action to complete the task successfully (Russell and Norvig 2010). An energetic portion of agent-based methods in the industrial sector comprises the observation to regulate the settings of their surroundings. The utilization of transducers and additional devices to acquire the statistics from the shop floor can be used for ML algorithms. Owing to the observation of agents, several learning tactics, followed by the agents’ emphasis on supervised learning algorithms as the sensor signal stipulates an accurate response, are employed for training the agents. Predictive maintenance aims to determine the condition of the machine to forecast the timing of the occurrence of the failure.

Cyber-Physical Systems Cyber-physical systems (CPS) are the outcome of the incorporation of computers connected with systems in the physical world. CPS extensively employs sensors and actuators to monitor and control the physical elements of the CPS. The lifetime of

46 Machine Learning: An Expert Thinking System

1171

the sensors and actuators employed in CPS plays a vital role in the trustworthiness and accessibility of the entire system. CPS is generally applied across domains like engineering and technology to empower the process optimization and earlier unattainable functionality. CPS becomes the crucial infrastructure that supports the progress of automated manufacturing (Li 2018). In an automated manufacturing system, there is no human involvement preferably, or to a certain level, there is reduced human involvement. So, the complete production line in a manufacturing plant would be made highly independent without any human involvement. The entire automation can be feasible with the aid of CPS through various elements like smart sensors and actuators. Smart sensors are going to be facilitating technologies to fulfill the needs of Industry 4.0. The subsequent epoch of smart computing will be entirely dependent on CPS (Nayyar and Puri 2016). CPS is a co-engineered interrelating network of physical and computational elements. These systems will afford the foundation of essential infrastructure, form the fundamentals of budding and upcoming smart services, and advance the quality in various domains (Griffor et al. 2017). Recently, IoT plays an essential role in renovating the “conventional technology” to “next-generation everywhere computing.” IoT is attaining a significant place in research across the globe, particularly in sensor technologies and automation. The most important aspect of Industry 4.0 is automation. CPS is an embedded system. Conceptually, CPS is the integration of an embedded system along with the physical system together in the cyber-physical system. Nowadays, embedded systems are the enabling technologies for making systems smarter. They possess a certain level of computation, communication, and control capabilities. So, the interaction with the physical world through different sensors and actuators is called cyber-physical systems. The typical example of CPS is shown in Fig. 6. The comparison of an embedded system with CPS is presented in Table 1.

Features of CPS The prominent features of CPS are described below (Alur 2015): • Reactive Computation: Reactive computation is a mandatory one, and it means the interaction of the system with the physical environment in a continuous manner. Accordingly, there is a classification of observed inputs and outputs in the process and deals with the systems. • Concurrency: The simultaneous execution of multiple processes is called concurrency . So, concurrent processes would exchange the information to accomplish certainly anticipated output, and these processes could be synchronous or asynchronous in terms of their operation.

1172

T. Mohanraj et al.

Fig. 6 Typical cyber-physical system

Table 1 Comparison of an embedded system with CPS Embedded system Devices embedded with data processing systems Typically narrowed to a single device Limited resources to perform a limited number of tasks Real-time response and reliability are the major issues

CPS System have physical components and software Networked with a set of embedded systems Resources constrain is not here Timing and concurrency are the major issues

• Feedback Control: Feedback control of the physical systems has a control system with a certain form of the control element. So, the sensors sense from the physical environment, and the actuators produce the corrective action on the environment. To perform the complex task, a hybrid control system is applied. • Real-Time Computation: Real-time computation is employed in a time-sensitive operation such as coordination, resource allocation, etc.

46 Machine Learning: An Expert Thinking System

1173

Components of CPS The major components of CPS are listed below: • Sensors: – Active sensor – Passive sensor • Data acquisition system: – Data logger • Connectivity: – IEEE standards – Zigbee – 6LoWPAN – HART – Z-wave – Bluetooth – RFID – LoRa (Long Range) • Advanced manufacturing system: – Additive manufacturing process – Flexible manufacturing system – SCADA/CAD/CAM – Unmanned industries • Data analytics: – Machine learning algorithms – Cloud computing – Edge computing • Human-machine interaction: – Autonomous robots – Man-machine interface – Swarm robots – Interconnected robots – Cobots (collaborative robots)

Machine Learning Human expertise should have a better understanding of the functionality and behavior of the system before the users of expert systems. Machine learning (ML) specifies the ability of the software to investigate large quantity of data and to study how to resolve the difficulties spontaneously. The various ML algorithms, like support vector machine (SVM), hidden Markov model (HMM), convolutional neural networks (CNN), decision tree (DT), etc., (Mahdavinejad et al. 2018) are

1174

T. Mohanraj et al.

Fig. 7 Classification of ML based on learning prospect

used for expert thinking. Big data is associated with the usage of consumer data for optimizing the product design and processes. ML is the science of accomplishing computers to perform exclusive programming (Blog 2017). In manufacturing industries, the ML algorithm is useful to process and analyze the data sensibly. The signals acquired from field devices of the shop floor industry cannot be processed manually. ML can play a major role in the transformation process of the product life cycle and product variety to enhance customer satisfaction. The classification of ML based on learning prospect is shown in Fig. 7. Figure 6 reveals a set of algorithms suitable for different types of learning process. In general, 70% of the data are used for training, and 30% of the data are used for testing and validation. The research of ML is classified into three main classes conferring to the learning approach (Hadfield-Menell et al. 2016; Monostori 2003; Szarvas et al. 2007): • Supervised learning: – Classification – Regression • Unsupervised learning: – Clustering • Reinforcement learning

46 Machine Learning: An Expert Thinking System

1175

Fig. 8 Process of supervised learning

Supervised Learning Supervised learning concentrates on cataloging (labeling) of training data, and the algorithm first learns from the given training data. The data employed for training purposes includes different patterns, in which the algorithm will learn from the patterns. It seems the output is readily available. But, from the training data, it has different outputs. Here, supervising assists the model to forecast the correct output. Supervised learning algorithms are typically employed in the machining industry, particularly owing to the data-rich but knowledge-scarce kind of problems. Figure 8 shows the learning process involved in supervised learning. Supervised ML is used in various field of manufacturing, metal cutting, monitoring, and control, which is the most significant one (Wuest et al. 2016). The application of supervised learning ranges from the machine level to process planning and design.

Unsupervised Learning Unsupervised learning is a learning methodology in ML. Unlike supervised learning, unsupervised learning does not label the data with which the user wants to train the model. Labeling the data means to classify the data into different categories. This labeling mainly takes place in supervised learning. But, in unsupervised learning, there is no labeling. Figure 9 shows the learning process involved in supervised learning. The model learns through training itself from the data. Unsupervised

1176

T. Mohanraj et al.

Fig. 9 Process of unsupervised learning

learning draws inferences from datasets without labels, and the assessment of the action is not offered since in this learning, the teacher is not involved. But in supervised learning, the right answer is provided by the supervisor. A detailed survey on deep learning was presented by (Hatcher and Yu 2018).

Reinforcement Learning Reinforcement learning is used along with supervised and unsupervised learning. The thought of offering rewards for every positive outcome is the basis of this algorithm. Reinforcement learning is considered by limited feedback; then relatively proper action is not initiated, and the evaluation of the action is provided by the teacher. This learning contains a hybrid method of unsupervised learning and is applied for a laser welding system to enhance the quality of welding (Günther et al. 2014) and is also used in the manufacturing industry to enhance production (Monostori et al. 1996). The process of reinforcement learning is shown in Fig. 10. The framework of the expert system is shown in Fig. 11. After acquiring the data from the industry, it has to be preprocessed in terms of signal conditioning, feature extraction, and feature selection process. Feature selection is performed to identify the most significant features that have vital information about the process. The feature selection ensures the recommended volume of data, which reduces the significant computational time while training the model with ML. After training and testing the ML, it has to be validated with real-time as well as benchmarking datasets to avoid overfitting and underfitting problems. This chapter focuses exclusively on the situations involving machine/tool condition monitoring, fault diagnosis, and early detection. Various works discuss the sensor signals and features used for condition monitoring applications (Mohanraj

46 Machine Learning: An Expert Thinking System

1177

Fig. 10 Process of reinforcement learning

Fig. 11 The framework of an expert system using ML

et al. 2020; Shankar et al. 2019b; Thangarasu et al. 2020). The subsequent section discusses the ML algorithms and their applications in machine tool condition monitoring.

Support Vector Machine Support vector machine (SVM) is predominant, but flexible SVM algorithms are used both for problems like classification and for regression types. In general, SVM is mainly used for classification problems. During the 1960s, SVM was introduced and advanced in the year 1990. SVM has its exclusive way of execution as related

1178

T. Mohanraj et al.

to other ML algorithms. Recently, SVM is enormously popular for its capability to handle various continuous and definite variables.

Naive Bayes Naive Bayes (NB) is a classification algorithm based on the Bayes’ theorem with a robust hypothesis that all the forecasters are self-regulating to each other. The hypothesis is that the incidence of a feature in a class is independent on the incidence of any other feature in the same class.

K-Nearest Neighbors K-nearest neighbors (KNN) is one of the supervised ML algorithms and employed to solve the classification as well as the regression nature of problems, though KNN is primarily applied to solve classification problems. The properties of KNN are described below: • Lazy learning: It is an exhausting learning algorithm since the dedicated training phase is not available and utilizes the entire data for training. • Nonparametric learning: It is a nonparametric learning algorithm since it does not guess anything about the original data. It uses “feature similarity” to forecast the values of new datasets, which means that the new dataset will be allotted a value based on how exactly it matches the data in the training set.

Logistic Regression Logistic regression is one of the supervised learning algorithms applied to forecast the possibility of a goal. The goal is dichotomous in nature, which means there would be only two types. The dependent variable is binary type, and the data is coded as either 1 (Yes) or 0 (No). Scientifically, it forecasts P(Y = 1) as a function of X. It is one of the simple ML algorithms applied for several classification problems like detection of spam, prediction of diabetes and cancer, etc.

Decision Tree Decision tree (DT) is a prognostic modeling tool used in various fields of applications. DT can be built by an algorithmic approach that can split the dataset in dissimilar ways based on the dissimilar conditions. DT is one of the most powerful algorithms that cataracts under the group of supervised algorithms. DT is applied to solve classification and regression problems.

46 Machine Learning: An Expert Thinking System

1179

Random Forest Random forest (RF) is a subset of a supervised learning algorithm applied for solving classification problems largely and suitable for the regression problem also. RF generates DT on data points and gets the estimation from each of them and lastly chooses the finest solution through voting. It is a collective approach that is better than a single DT, as it diminishes the overfitting problem by using the average of the predicted results.

K-Means Clustering K-means clustering calculates the centroids and repeats until the optimum value for centroid is found. It assumes that it has a known number of clusters. The number of clusters recognized from data points by the algorithm is denoted by “K.” The dataset is allotted to a cluster in such a way that the sum of the squared distance among the datasets and centroid would be least as possible. The least difference inside the clusters will lead to more similar datasets within the same cluster.

Hierarchical Clustering Hierarchical clustering is an unsupervised learning algorithm used for clustering the unlabeled datasets that have the same characteristics. Hierarchical clustering algorithms have two classes: • Agglomerative hierarchical algorithms: Every data is considered as an individual cluster and consecutively combines the pair of clusters. • Divisive hierarchical algorithms: The entire data is considered as a single large cluster, and the clustering process involves dividing the single large cluster into numerous minor clusters.

Mean Shift Algorithm The mean shift algorithm is one of the unsupervised powerful nonparametric learning algorithms. The mean shift algorithm essentially allows the datasets to the clusters iteratively by shifting points toward the highest density of datasets (cluster centroid). The number of clusters can be identified by the algorithm for the data.

Artificial Neural Network Artificial neural network (ANN) is based on how the human brain works by making the correct networks. The basic element of ANN is a neuron, and each neuron is

1180

T. Mohanraj et al.

connected with other neurons through axons. A neuron can transfer the message to other neurons to handle the issue or does not send it forward. ANN consists of multiple layers (input, hidden, output), which replicate the biological neurons of the human brain. The nodes can take input data and make simple actions on the data. All the links are associated with their weight. ANN is proficient in learning, through changing the values of weight.

Hidden Markov Model Hidden Markov model (HMM) is a graphical model that consents to forecast an arrangement of unknown (hidden) variables from a set of observed variables. An HMM can be observed as a Bayes net unfolded through time with remarks made at a sequence of time steps being applied to forecast the best arrangement of hidden states.

Applications of ML in Condition Monitoring ML algorithms play a vital role in the automation and condition monitoring process. An online condition monitoring system can be built through the implementation of ML. In the turning process, tool wear was monitored by analyzing the machined surface with SVM classification through Gaussian and polynomial kernels and found that polynomial kernel with degree 7 or 9 exhibits the higher classification accuracy (Bhat et al. 2016). In the milling process, tool wear was monitored with cutting force and vibration signals with CNN, SVM, Bayesian network (BN), and KNN. They reported that the performance of CNN was better than other algorithms and validated the accuracy of the CNN model with NASA_AMES benchmarking data (Aghazadeh et al. 2018). During the milling of Ti6Al4V, tool wear was classified with vibration signals using v-SVM and s-SVM. They applied locality preserving projection (LPP) to identify the most significant features of vibration signals and found that v-SVM with LPP reveals better classification accuracy (Wang et al. 2014). The accuracy of the results depends on the kernel function of the SVM (Elangovan et al. 2011). During the face milling of 42CrMo4 steel alloy, the tool condition was monitored with sound signals using SVM. The performance of SVM was compared with ANN, DT, and NB. SVM proved its higher classification accuracy of 83% compared to other classifiers (Madhusudana et al. 2017). A detailed review of ML in milling process monitoring was briefed (Zhou and Xue 2018). In the turning process, tool condition was monitored with cutting force signatures using v-SVR, BRNN, and BPNN models. The results were evident that v-SVR exhibits the classification accuracy of 96.76% than other algorithms (Li et al. 2017). In the grinding process, wear of the grinding wheel was monitored with acoustic (AE) signals using DT, ANN, and SVM. They identified that SVM trained with cubic kernel performed better than DT and ANN in terms of classification accuracy

46 Machine Learning: An Expert Thinking System

1181

(Arun et al. 2018). In the milling process, tool wear was monitored with vibration signals using a long short-term memory network (LSTM) and found a considerable improvement in classification accuracy. The accuracy of the proposed model was validated with a benchmarking dataset (Cai et al. 2020). In the milling process, tool wear was monitored with AE and vibration signatures using various ML algorithms. The feature level and decision fusion were made to predict the tool condition accurately. They used DT, SVM, NB, and ANN with time, frequency, and time-frequency domain features. They reported that SVM and ANN had better accuracy in the time domain and frequency domain, respectively (Krishnakumar et al. 2018). During the milling process, the tool condition was monitored with the sound signal, cutting force, and vibration signals. The statistical features like root means square, mean, kurtosis, and skewness were considered as features, and ANN was used as a decision-making algorithm. They found that ANN with selected features was used to predict the tool condition properly (Shankar et al. 2019a). They found that the performance of ANN with cutting force and sound signals was better than ANFIS (Shankar et al. 2019b). Table 2 shows the classification accuracy of various ML in tool condition monitoring applications. Table 2 Prediction accuracy of various ML in condition monitoring applications References Zhou et al. (2020) Zhou et al. (2019)

Zhou et al. (2020)

Sakthivel et al. (2010a) Sakthivel et al. (2012)

Sakthivel et al. (2014)

Sakthivel et al. (2010b)

Machine learning algorithm SVM SVM with transition point identification method KNN DT SVM KNN DT ANN SVM KNN DT DT SVM GEP (gene expression programming) P-SVM Wavelet -GEP Kernel PCA Manifold chart MVU PCA DT Fuzzy DT-PCA

Classification accuracy (%) 81.9 90.8 81.3 79.3 85.7 85.7 84.9 82.1 98.7 93.7 96.3 99.66 99.93 99.93 96.66 99.83 96.53 75.06 77.82 99.3 99.33 97.50 96.6

1182

T. Mohanraj et al.

Conclusion The future industry depends on Industry 4.0 and the entire cyber-physical system. The cyber-physical system, along with the innovation of the Internet of things (IoT), made the system become fully automated, and the system can be monitored and controlled anywhere in the world. The most important thing in Industry 4.0 is the thread of cyber-security and attacks. The entire control of the industry can be monitored and controlled through machine learning algorithms. With the advent of ML, an expert system can be designed and implemented in the unmanned industry. The expert system with ML can be used for many real-world applications.

References Aghazadeh F, Tahan A, Thomas M (2018) Tool condition monitoring using spectral subtraction and convolutional neural networks in milling process. Int J Adv Manufact Technol 98:3217–3227. https://doi.org/10.1007/s00170-018-2420-0 Alur R (2015) Principles of cyber-physical systems. MIT Press Arun A, Rameshkumar K, Unnikrishnan D, Sumesh A (2018) Tool condition monitoring of cylindrical grinding process using acoustic emission sensor. Mater Today Proc 5:11888–11899. https://doi.org/10.1016/j.matpr.2018.02.162 Bhat NN, Dutta S, Vashisth T, Pal S, Pal SK, Sen R (2016) Tool condition monitoring by SVM classification of machined surface images in turning. Int J Adv Manufact Technol 83:1487–1502 Bicocchi N, Cabri G, Mandreoli F, Mecella M (2019) Dynamic digital factories for agile supply chains: an architectural approach. J Ind Inform Integr 15:111–121 Blog (2017) Machine learning. https://expertsystem.com/machine-learning-definition/. Accessed 22 Oct 2020 Cai W, Zhang W, Hu X, Liu Y (2020) A hybrid information model based on long short-term memory network for tool condition monitoring. J Intell Manufact 31:1–14 Cerquitelli T, Proto S, Ventura F, Apiletti D, Baralis E (2019) Towards a real-time unsupervised estimation of predictive model degradation. In: Proceedings of real-time business intelligence and analytics. United States Association for Computing Machinery, New York, pp 1–6 Chen B, Wan J, Shu L, Li P, Mukherjee M, Yin B (2017) Smart factory of industry 4.0: key technologies, application case, and challenges. IEEE Access 6:6505–6519 Elangovan M, Sugumaran V, Ramachandran KI, Ravikumar S (2011) Effect of SVM kernel functions on classification of vibration signals of a single point cutting tool. Expert Syst Appl 38:15202–15207. https://doi.org/10.1016/j.eswa.2011.05.081 Expert system (2020). https://www.expert.ai/ Griffor ER, Greer C, Wollman DA, Burns MJ (2017) Framework for cyber-physical systems: Special Publication (NIST SP), National Institute of Standards and Technology, Gaithersburg, volume 1, overview. https://doi.org/10.6028/NIST.SP.1500-201. Accessed 9 Aug 2021 Günther J, Pilarski PM, Helfrich G, Shen H, Diepold K (2014) First steps towards an intelligent laser welding architecture using deep neural networks and reinforcement learning. Proc Technol 15:474–483 Hadfield-Menell D, Russell SJ, Abbeel P, Dragan A (2016) Cooperative inverse reinforcement learning. In: Advances in neural information processing systems. 29:3909–3917 Harifi S, Khalilian M, Mohammadzadeh J, Ebrahimnejad S (2020) Optimization in solving inventory control problem using nature inspired Emperor Penguins Colony algorithm. J Intell Manuf 32:1–15 Hatcher WG, Yu W (2018) A survey of deep learning: platforms, applications and emerging research trends. IEEE Access 6:24411–24432

46 Machine Learning: An Expert Thinking System

1183

Krishnakumar P, Rameshkumar K, Ramachandran K (2018) Feature level fusion of vibration and acoustic emission signals in tool condition monitoring using machine learning classifiers. Int J Progn Health Manag 9:1–15 Li D (2018) Study and application of the computing architecture of petrochemical cyberphysical system (PCPS). In: Eden MR, Ierapetritou MG, Towler GP (eds) Computer aided chemical engineering, vol 44. Elsevier, pp 2023–2028. https://doi.org/10.1016/B978-0-44464241-7.50332-3 Li N, Chen Y, Kong D, Tan S (2017) Force-based tool condition monitoring for turning process using v-support vector regression. Int J Adv Manuf Technol 91:351–361. https://doi.org/ 10.1007/s00170-016-9735-5 Madhusudana CK, Kumar H, Narendranath S (2017) Face milling tool condition monitoring using sound signal. Int J Syst Assur Eng Manag 8:1643–1653. https://doi.org/10.1007/s13198-0170637-1 Mahdavinejad MS, Rezvan M, Barekatain M, Adibi P, Barnaghi P, Sheth AP (2018) Machine learning for Internet of Things data analysis: a survey. Digit Commun Netw 4:161–175 Mohanraj T, Shankar S, Rajasekar R, Sakthivel N, Pramanik A (2020) Tool condition monitoring techniques in milling process – a review. J Mater Res Technol 9:1032–1042 Mondal KC, Nandy BD, Baidya A (2020) Fact-based expert system for supplier selection with ERP data. In: Algorithms in machine learning paradigms. Springer, pp 43–55 Monostori L (2003) AI and machine learning techniques for managing complexity, changes and uncertainties in manufacturing. Eng Appl Artif Intell 16:277–291 Monostori L, Márkus A, Van Brussel H, Westkämpfer E (1996) Machine learning approaches to manufacturing. CIRP Ann 45:675–712 Naren R, Subhashini J (2020) Comparison of deep learning models for predictive maintenance. In: IOP conference series: materials science and engineering, vol 2. IOP Publishing, p 022029 Nayyar A, Puri V (2016) Smart farming: IoT based smart sensors agriculture stick for live temperature and moisture monitoring using Arduino, cloud computing & solar technology. In: Proceedings of the international conference on communication and computing systems (ICCCS2016). pp 9781315364094–9781315364121 Park S (2016) Development of innovative strategies for the Korean manufacturing industry by use of the Connected Smart Factory (CSF). Proc Comput Sci 91:744–750 Perera JCP (2020) Sustainability-based expert system for additive manufacturing and CNC machining Purnomo M, Hidayatuloh S (2020) Development of rule-based expert system for conceptualisation of poster design. In: IOP conference series: materials science and engineering, vol 1. IOP Publishing, p 012019 Russell SJ, Norvig P (2010) Artificial intelligence-a modern approach, third international edition. Pearson Education, London Sakthivel N, Sugumaran V, Babudevasenapati S (2010a) Vibration based fault diagnosis of monoblock centrifugal pump using decision tree. Expert Syst Appl 37:4040–4049 Sakthivel N, Sugumaran V, Nair BB (2010b) Comparison of decision tree-fuzzy and rough setfuzzy methods for fault categorization of mono-block centrifugal pump. Mech Syst Signal Process 24:1887–1906 Sakthivel N, Nair BB, Sugumaran V (2012) Soft computing approach to fault diagnosis of centrifugal pump. Appl Soft Comput 12:1574–1581 Sakthivel N, Nair BB, Elangovan M, Sugumaran V, Saravanmurugan S (2014) Comparison of dimensionality reduction techniques for the fault diagnosis of mono block centrifugal pump using vibration signals. Eng Sci Technol Int J 17:30–38 Sergi BS, Popkova EG, Bogoviz AV, Litvinova TN (2019) Outlines of the context for industry 4.0 understanding industry 40: AI, the Internet of Things, and the future of work. Emerald Group Publishing 3–10 Shankar S, Mohanraj T, Pramanik A (2019a) Tool condition monitoring while using vegetable based cutting fluids during milling of inconel 625. J Adv Manuf Syst 18:563–581

1184

T. Mohanraj et al.

Shankar S, Mohanraj T, Rajasekar R (2019b) Prediction of cutting tool wear during milling process using artificial intelligence techniques. Int J Comput Integr Manuf 32:174–182 Shi Z, Xie Y, Xue W, Chen Y, Fu L, Xu X (2020) Smart factory in Industry 4.0. Syst Res Behav Sci 37:607–617 Szarvas G, Farkas R, Busa-Fekete R (2007) State-of-the-art anonymization of medical records using an iterative machine learning framework. J Am Med Inform Assoc 14:574–580 Thangarasu S, Shankar S, Mohanraj T, Devandran K (2020) Tool wear prediction in hard turning of EN8 steel using cutting force and surface roughness with artificial neural network. Proc Inst Mech Eng C J Mech Eng Sci 234:329–342 Vercruyssen V, Meert W, Davis J (2020) “Now you see it, now you don’t!” Detecting suspicious pattern absences in continuous time seriesc. In: Proceedings of the 2020 SIAM international conference on data mining. SIAM, pp 127–135 Wang GF, Yang YW, Zhang YC, Xie QL (2014) Vibration sensor based tool condition monitoring using ν support vector machine and locality preserving projection. Sensors Actuators A Phys 209:24–32. https://doi.org/10.1016/j.sna.2014.01.004 Wang S, Wan J, Zhang D, Li D, Zhang C (2016) Towards smart factory for industry 4.0: a selforganized multi-agent system with big data based feedback and coordination. Comput Netw 101:158–168 Woolery LK, Grzymala-Busse J (1994) Machine learning for an expert system to predict preterm birth risk. J Am Med Inform Assoc 1:439–446 Wuest T, Weimer D, Irgens C, Thoben K-D (2016) Machine learning in manufacturing: advantages, challenges, and applications. Prod Manuf Res 4:23–45 Zalazinsky A, Shveykin V, Titov V (2020) The expert system for improving the technological processes of composite manufacturing. In: IOP conference series: materials science and engineering, vol 4. IOP Publishing, p 044115 Zhou Y, Xue W (2018) Review of tool condition monitoring methods in milling processes. Int J Adv Manuf Technol 96:2509–2523. https://doi.org/10.1007/s00170-018-1768-5 Zhou C, Guo K, Yang B, Wang H, Sun J, Lu L (2019) Singularity analysis of cutting force and vibration for tool condition monitoring in milling. IEEE Access 7:134113–134124. https:// doi.org/10.1109/ACCESS.2019.2941287 Zhou CA et al (2020) Vibration singularity analysis for milling tool condition monitoring. Int J Mech Sci 105254:166. https://doi.org/10.1016/j.ijmecsci.2019.105254 Zhu N, Song B, Lu Y, Li Y, Zhu Z (2020) Research on equipment fault prediction expert system based on big data dimension reduction. In: Journal of Physics: conference series, vol 3. IOP Publishing, p 032045

47

The Adoption of Industry 4.0 Technologies Through the Implementation of Continuous Improvement Tools Maria Rosaria Sessa, Ornella Malandrino, Giuseppe Fenza, Gianfranco Caminale, and Claudio Risso

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Main Characteristics of Quality 4.0 Paradigm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Implementation of WCM Through Quality 4.0 Principles . . . . . . . . . . . . . . . . . . . . . . . Introduction of RAMI 4.0 for Intelligent Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . WCM for the Intelligent Production in LEONARDO of Tomorrow . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1186 1188 1192 1198 1204 1209 1216 1218

Abstract

In recent years, the term Industry 4.0 refers to a multiplicity of changes which are revolutionizing the production methods of different industrial sectors. The purpose of this chapter is to identify different tools to support continuous improvement of performance in order to adopt Industry 4.0 model (see figure below). To reach this aim, the authors carried out an analysis on possible implementation and integration of different Quality Management approaches, as World Class Manufacturing (WCM) and RAMI 4.0 in an Italian aerospace,

M. R. Sessa () · O. Malandrino · G. Fenza Department of Management & Innovation Systems, University of Salerno, Fisciano, SA, Italy e-mail: [email protected]; [email protected]; [email protected] G. Caminale CTO Cyber Security Division, LEONARDO Company, Genoa, Italy e-mail: [email protected] C. Risso Critical Infrastructures, EPC & Large Enterprise – Cyber Security Division, LEONARDO Company, Genoa, Italy e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_30

1185

1186

M. R. Sessa et al.

defense and security company, within the scope of national program PON 2014– 2020, Leonardo 4.0, and the program ECSEL-Innovation Actions-2018. This research provides information about possible integration between managerial approaches and 4.0 technologies. This integration will allow the real implementation Industry 4.0 model. This can serve to academicians and practitioners in the field as incipit for development of actionable strategies and practices for successful transition from traditional manufacturing into the Industry 4.0. The present chapter contributes to the debate surrounding Industry 4.0 by stressing that willing firms undertake the transition from actually manufacturing into the Industry 4.0, can’t build this phase on only application of enabling technologies but must be aware that is necessary have to implement precise organizational and managerial approaches. All stakeholders need to be aware that to adopt the Industry 4.0 model requires not only the application of the new technologies but also the development and implementation of a series of Quality Management tools and practices that become an essential key to face the fourth Industrial Revolution. In fact, better managerial practices can support the correct adoption of the Industry 4.0 model. Keywords

Industry 4.0 · New technologies · Managerial approaches · Integration

Introduction In recent decades, companies have felt the need to redefine their business model for the purpose of facing multiple challenges produced by the emergence of new trends, such as the expansion of markets in a global dimension, which involves an increase in the intensity of international competition and a reduction in the life cycle of the product; a change in the average consumer behavior, who is more and more careful to principles of economic, environmental, and social sustainability (Dossou 2018); the advent of the fourth industrial revolution or Industry 4.0 which significantly reduces the positive effects of economies of scale and allows the configuration of products, processes, and, more generally, of value chains through the use of enabling technologies (Schuh et al. 2014). There is no Industry 4.0 without Big Data Management and Business Intelligence supported by increasingly customized Analytics for each single company. In fact, the data, put into the system, offer companies a new predictive analysis capacity, which ensures a wide information base to improve products and services, supporting decisions in the best possible way. On this point, new digital technologies will have a deep impact in the context of four development guidelines: the first concerns the use of data, computing power, and connectivity and is divided into Big Data, Open Data, Internet of Things, Machine-to-Machine, and Cloud Computing for the centralization of information and its conservation; the second one is that of Analytics: after the data

47 The Adoption of Industry 4.0 Technologies Through the. . .

1187

has been collected, it is necessary to obtain a value. Today only 1% of the data collected is used by companies, which could instead obtain advantages starting with Machine Learning, by machines that improve their performance by learning from the data gradually collected and analyzed; the third development line is the interaction between man and machine, which involves both touch interfaces, which are increasingly popular, and intensified reality; at last, the fourth guideline concerns the whole sector that deals with the transition from digital to real world and which includes additive manufacturing, 3D printing, robotics, communications, Machineto-Machine interactions, and new technologies for storing and using energy in a focused way, rationalizing costs and optimizing performance. Therefore, Industry 4.0 represents a highly innovative management model, but without adequate methodological support – which in this report will be called Quality 4.0 – there is a risk that technology will prevail over the methodological approach, without generating the expected benefits compared to the investments made. This factory configuration leads to different challenges and approaches compared to the Industry 3.0 model, so it is necessary to expand the field of analysis: customer and provider must be considered as crucial points in the value chain and quality must be a feature present starting from the top to the bottom of the supply chain. Thanks to Industry 4.0 we therefore move from product design to factory design: the physical object is no longer separated from its digital value, on the contrary it is possible to start from the virtual project and then overturn it on the physical world. Nevertheless, the 4.0 model can be effectively implemented only if the tools intended for continuous improvement of process and product performance, typical of Industry 3.0 but implemented through the digital technologies of the fourth industrial revolution, continue to be taken into consideration. Therefore, an organization will be able to choose this model only if it is able to combine the Quality 4.0 methodological approach with the digital innovations of Industry 4.0, in order to make processes and products more efficient and sustainable and enhance the skills of the human capital. In light of these premises, the objective of this contribution is to examine the World Class Manufacturing (WCM) and RAMI 4.0 tools for the management of the digital business system, in order to suggest the adoption of the Industry 4.0 model through the Quality 4.0 methodological approach in LEONARDO, a leader company in the Italian aerospace, defense, and security sector. In particular, we will proceed with the analysis of the possible integration between some of technical pillars of WCM with Key Enabling Technologies (KET) for the effective implementation of Industry 4.0 model. In this regard, the main approaches at the basis of WCM will be analyzed, namely, Total Quality Management, Lean Production, and Just in Time. The recognition of business processes and tools designed to manage them, according to the principles of quality, will come to a standardization of procedures and, therefore, the development of an integrated system of effective tools and methodologies for entire production reality in which the final solution will be implemented.

1188

M. R. Sessa et al.

Background The term “Industry 4.0” was used for the first time in 2011, during the Hannover Messe on the initiative of the working group led by Siegfried Dais (Robert Bosh GmbH) and Henning Kagermann (Acatech) as part of the High-tech 2020 plan, in order to create production systems capable of monitoring physical processes through Enabling Technologies – creating a so-called digital twin of the physical world – and to make intelligent decisions through real-time communication and cooperation between humans and machines (Wang et al. 2016). Indeed, industry 4.0 combines production systems with intelligent production processes by exploiting advanced information and production technologies to create flexible, intelligent, and reconfigurable projects capable of facing an increasingly dynamic and global market (Shen and Norrie 1999). It also allows all physical processes and information flows to be available when and where they are needed through holistic production chains, where the use of supporting technologies allows devices or machines to vary their behaviors in response to different situations and requirements based on past experiences and learning skills (McFarlane et al. 2003) and to communicate directly with production systems in order to solve different problems and make adaptive decisions in a timely manner. Although this conception of Industry 4.0 is widely shared, actually, to date considering the embryonic nature of the concept, we have not arrived yet at a univocal definition of the same concept, which is why very often are used different terminological expressions such as Cyber Physical Systems, Internet of Things, Smart Manufacturing, or Advanced Manufacturing to represent and describe the fourth industrial revolution. Therefore, there are many definitions of Industry 4.0 present in the literature. According to Kang et al. (2016) Industry 4.0 or Intelligent Manufacturing is identifiable in the fourth industrial revolution, which represents a new paradigm and consists in the convergence of ICT technologies and cutting-edge manufacturing, providing a solid basis for making effective and optimized decisions through smart and accurate decision-making systems. While Ivanov et al. (2016) represent Industry 4.0 through the concept of smart manufacturing based on the network, in which machines and products interact with each other without the need for human intervention. Industry 4.0, according to Kagermann et al. (2013), concerns the technical integration of cyber-physical systems in manufacturing and logistics and the use of the Internet of Things and Internet of Services in industrial processes. This will have consequences on the methods of generating value, on business models, on the services provided downstream, and, in general, on the whole organization of work. But, Zhong et al. (2017) argue that understanding the Industry 4.0 context is only possible if another concept is taken into consideration, such as intelligent manufacturing or Smart Manufacturing, so through a systematic review of the literature, it will identify more than 350 articles (in Scopus and Google Scholar), from 2005 to 2015, in which is made reference to intelligent production when trying to define the concept of Industry 4.0. Intelligent manufacturing is aimed

47 The Adoption of Industry 4.0 Technologies Through the. . .

1189

at optimizing production processes through the use of advanced technologies, in order to generate a new business performance management model that significantly enhances the design, production, management, and integration of the whole product life cycle. This implies that the management and evaluation, in terms of quality, of the entire product life cycle can be facilitated using various smart devices as well as adaptive decision models (Li et al. 2017). However, it is necessary to underline that the embodiment of this concept can be found in the Intelligence Manufacturing System, a new generation system that uses the network architecture via the Internet to provide collaborative, customizable, flexible, and reconfigurable services to end users and start a highly integrated man-machine production system (Feeney et al. 2015). This integration aims to establish an ecosystem of the various manufacturing activities involved in the smart factory system, so that the organizational, managerial, and technical levels are perfectly connected (Chen et al 2018). So, probably, the concept behind Industry 4.0 is that of collaboration-cooperation that can take place in three ways: between people, between machines, or between machines and men. All this is possible precisely through the use of technologies that facilitate communication practices and that allow machines to share the information collected in real time as well as to improve the quality of processes and products and to reduce time to market from product conception to marketing (Brettel et al. 2014). The main characteristics required for a correct conversion of traditional production into intelligent production are the following: – Horizontal integration between the actors of the value chain through the sharing of information regarding production processes and data – End-to-end integration of engineering processes through advanced communication and virtualization tools to enable the customization of products and services – Vertical integration of business processes to establish cross-functional collaboration and cooperation between departments In this context, technology is one of the driving forces of change which is capable of automating business processes, making them flexible and efficient and facilitating internal and external communication. The enabling technologies of Industry 4.0 are varied and heterogeneous. For example, technologies commonly classified under the name of Cyber-Physical Systems (CPS) or the Internet of Things (IoT) are taken into consideration for communication and collaboration between objects, systems, environments, and people. In this dense network of information exchanges, huge amounts of data are also generated and collected. The scientific branch that follows this trend is Big Data Analysis. Furthermore, other technologies that are redefining the industry are augmented reality and virtual reality, cloud computing, additive printing or 3D printing, and newly developed robotics. More precisely, the cyber-physical system can be understood as the set of technologies capable of “dialogue” with the physical and real world, constituting a network that incorporates all the business activities of a company (Faller and

1190

M. R. Sessa et al.

Höftmann 2018). It is based on a set of devices, used to communicate and collaborate with each other and with the surrounding external environment, controlling activities, collecting or providing data in real time, and making services available via the Internet (Lu 2017). According to Lee et al. (2015) the architecture of CPS systems can be represented through five levels: – Smart connection, an area in which there is the connection of data, a network of sensors, the connection to development data, the management system, and databases. – Data-to-information conversion, in this phase the collected data must be converted into processable information, for which algorithms are used to understand what type of data has been collected. – Cyber level, in which the cyber system acts by coordinating the devices through communication. – Cognition level refers to the cognitive capacity of the system once the global situation has been understood, in order to understand how to act and interact with users and make adequate decisions. – Configuration level, in this phase the configuration of the real world starts on the basis of the choices coming from the cyber-physical world. The good functioning of a CPS system requires an adequate IT division, capable of withstanding a greater use of services and the load of the network infrastructure. The latter must be able to support a large number of connected devices with greater information processing, without slowing down or interrupting the connected services. All this implies larger physical environments, technical quality, systems reliability, and qualified personnel (Kagermann et al. 2013). While, the term Internet of Things (IoT) is used for the first time by a researcher from MIT (Massachusetts Institute of Technology) – Kevin Ashton – to give a name to the interconnection of physical objects via the Internet. Biography of Kevin Ashton. (http://ethw.org/Kevin_Ashton). According to Schwab (2016), an organization’s use of IoT could lead to a number of positive, negative, or yet unknown consequences on business performance. The main positive consequences are increased productivity; increased efficiency in the use of resources; reduction of environmental impacts; increased demand for storage space and broadband; labor market transformation; request for new professional skills; possibility of having rigid devices that provide information in real time through traditional communication networks; creation of products that can be digitally connected; new knowledge through the connection of intelligent objects. While the negative consequences are less privacy; precariousness for lowskilled individuals on these aspects; risk of cyber attacks; lower security; higher levels of complexity; and loss of control. Finally, the aspects with consequences not yet known could be impact of the value of the contents on the business model; possibility for each organization to implement an application; creation of new businesses in the field of data marketing; changes in the area of privacy; dislocation of infrastructures to take advantage of information technologies; work automation.

47 The Adoption of Industry 4.0 Technologies Through the. . .

1191

In Italy, the knowledge of the IoT is not, to date, particularly thorough but the data of the IoT Observatory of the School of Management of the Politecnico di Milano leave no room for negative impressions, reporting a significant growth in the use of this technology. Observatory of the School of Management of the Politecnico di Milano (https://www.osservatori.net/it_it/osservatori/osservatori/ internet-of-things). In particular, Bellini (2016) believes that the use of IoT in Italy can be identified in three different application areas: consolidated, experimental, and embryonic. The consolidated applications are the most successful and widespread, since they are immediately applied and are very simple, such as video surveillance and security in smart homes with the aim of controlling or avoiding illicit intrusions. While experimental and embryonic applications are classified under the Radio Frequency Identification (RFID) technologies used within the Supply Chain. The most relevant problem with respect to the spread of the IoT could be represented by the protection of privacy, as the ability to process data remotely and transfer them using remote connections if not carried out with transparency can create damage to personal security. Therefore, for the realization of the IoT, characteristics such as compatibility, modularity, and scalability of the solutions are necessary (Wan et al. 2015). This means providing an integration on the production equipment, a network, and a supervision system in order to carry out a remote control of the vehicles, systems, and also of the personnel and to collect data and information in real time, allowing to make choices that increase productivity and efficiency and lead to the overall optimization of the system, regardless of how complex it may be (Chui et al. 2013). A further enabling technology can be identified in the use of Big Data, that is, the ability to collect, share, select, aggregate all the information useful to process, analyze, and find new solutions to old and new issues through algorithms capable of dealing with many variables in a short time and with computational resources however limited in relation to the amount of input. It is a technology suitable for Industry 4.0 as it adapts to different use scenarios (Golzer et al. 2015). The growing need to obtain more memory, greater flexibility, and speed of calculation finds its solution in Cloud Computing, a term which identifies a set of technologies capable of using the software and hardware available in the network to process, store, and archive data. More precisely, Cloud Computing guarantees organizations a reduction in fixed capital, personnel costs, and the management and maintenance of the IT system. Furthermore, it allows you to manage any problems quickly and efficiently: this means greater timeliness and reliability. At the same time, the data centers installed inside structures equipped with access control and redundant hardware reduce the risk of data loss compared to normal company servers. Finally, the data servers located within the server farm guarantee a high speed of execution when requesting data. A further technology that is already unique to the Industry 4.0 is represented by traditional robots and robot Collaborative who take the name of Cobot (Beltrametti et al. 2017). These are advanced machines able to collaborate with each other and with the staff working inside the factory, in order to perform complex tasks in synergy. This means that the operator is supported in the production line by smart robots able to perform teamwork, creating a hybrid team,

1192

M. R. Sessa et al.

in which the characteristics of the staff and the robot are evaluated as a single group that combines the flexibility of the human being, to the accuracy of the machine (Richert et al. 2016). Therefore, the requirements sought are flexibility, reliability, and security rather than high speed of execution of the activities. Finally, one of the symbols of production 4.0 is 3D printing which uses additive techniques to create any type of object designed using special software. Usually the process that allows the creation of solid objects through the use of 3D printing is divided into three phases (Bongio and Distefano 2015): – Modeling, which is the information acquisition phase and consists in reporting the details of the object on a CAD software, in which any changes are reported. – Slicing, in which the previously created virtual model is divided into layers, which will be compatible with the model in use of the 3D printer and can be reproduced in succession by the latter. – Print, in which the virtual object is reproduced. There are several printing methodologies, including extrusion, photopolymerization, granular technologies, and lamination. The main advantages of this technology can be summarized as follows: – Realization of systems and parts of systems as well as single pieces, without the need for subsequent assembly and which contain electrical parts, sensors, and batteries. – Reduction of production costs. Products are made with characteristics equal to or superior to those observed in products made using traditional technologies, obtaining, moreover, unique quality standards. – Possibility to change the design of the product by improving the basic characteristics of the same object. – Possibility of storing virtual models rather than finished products. This allows a reduction in fixed capital and all those costs related to warehouse management. – Allows to cope with the sudden and unexpected lack of components for line production. – Facilitates the prototyping phases in all those sectors in which economies of scale are not decisive but critical success factors, such as agility and speed of response.

The Main Characteristics of Quality 4.0 Paradigm The reference framework, described in the previous paragraph, allows to affirm that Industry 4.0 represents a highly innovative management model, but without adequate methodological support – which can be identified in the Quality 4.0 paradigm – there is a risk that technology will prevail over the “methodological approach,” without generating the expected benefits with respect to the investments made.

47 The Adoption of Industry 4.0 Technologies Through the. . .

1193

The intelligent production model in Industry 4.0 can only be effectively implemented if the tools aimed at continuous improvement of process and product performance, typical of Industry 3.0 but implemented through the use of enabling technologies, continue to be taken into consideration typical of the fourth industrial revolution. Quality 4.0 combines new technologies with traditional methods to achieve new results in terms of Operational Excellence, performance, and innovation. New technologies and advances in terms of data, analysis, connectivity, scalability, and collaboration will have an impact on all company performance and on internal and external stakeholders within the same organization. It was in the 1890s that discussions began with Taylor about the scientific organization of work. This concept has been expanded, refuted but also criticized by subsequent contributions that have followed over time and to which have been added further concepts such as Lean Production, Just in Time, and Total Quality Management (Taylor 1919). The concept of Lean Production refers to the management of company performance according to the logic of eliminating waste in order to maximize the value-cost ratio of production. This means that a production is Lean when it uses the least amount of man hours, materials, machines, and economic amount obtaining the best results within the foreseen time frame. Started to talk about Lean production in the second half of the last century, when in 1988 John Krafcik (Engineer at New United Motor Manufacturing Inc., Joint venture of Toyota and General Motors), with the article Triumph of the Lean Production System (1988), presented the concept of Lean Manufacturing or Lean Production as a set of formalized approaches. The key principles that must be pursued by an organization to adapt and implement the Lean Production methodology are the following: – Value. The starting point for Lean Production is the concept of value that must be rethought from the customer’s point of view. Only a small fraction of the actions and total time it takes to produce a specific product add real value to the end customer. It is of fundamental importance to define the value of the product according to the customer’s perspective, so that all activities that do not add value can be removed step by step. – Map. You need to focus on analyzing the activities that create value. The analysis involves all activities ranging from design, order management to product production. – Flow. After defining the value and after the value stream has been completely mapped and all kinds of waste have been eliminated, we focus on the activities that create value. The goal is to ensure that these value-creating activities flow steadily and continuously. To do this, it is necessary to review how to organize the work, what type of equipment to use to facilitate production in order to avoid backward flows, rejects and stops, what structure to create to facilitate the flow and what kind of professionals to look for.

1194

M. R. Sessa et al.

– Pull. This term indicates that upstream goods are not produced until the moment in which the downstream customer requests them. This makes it possible to avoid raising the level of stocks by the producer of the good, its suppliers and the companies producing the raw materials. – Perfection. Once the value has been accurately defined, the value stream identified, and all the previous steps carried out, it is necessary to pursue perfection through continuous improvements according to the Kaizen philosophy. Therefore, one of the basic approaches of Lean Production is Value Stream Mapping, according to which it is necessary to map and analyze the flow of value generation with the aim of distinguishing the activities that produce added value from those that do not create it in order to reduce costs, process complexity, and process lead time as well as increasing production capacity by making better use of available resources. Therefore, once the “AS-IS” status has been analyzed, that is the current state of business processes, we move on to the identification of the criticalities, in order to create a map of the “TO-BE” scenarios, or a map of the processes in which critical issues will be resolved in order to make production leaner. In this regard, the hunt for waste, or all those activities or ways of using resources that do not provide the product with an added value, is the first step in building “TOBE” scenarios. In the context of lean manufacturing, seven different types of waste are identified: – Overproduction. It consists in producing a quantity of components or finished products that exceeds the demand. – Wait. It is the most easily identifiable waste. It occurs whenever an operator does not carry out any work waiting for material (from the supplier or warehouse) or production means. – Transportation. Every time a product is handled it risks being damaged, lost, etc. It turns out to be a non-value-added activity. – Process. This type of waste occurs when the production process does not have adequate means (equipment, machinery, operators) and procedures. – Stocks. The materials produced in excess of the real needs wherever they are found, on production lines, in warehouses, in order from suppliers, are considered a waste of both space and financial resources. – Useless movements. Useful work is that particular type of movement that produces value. Unproductive movements are all those types of movements that involve unnecessary movements due to poorly designed layouts or oversized structures and unproductive actions attributable to workplaces that have not been ergonomically designed. – Rework. Whenever you perform an operation that produces a defective part, the defect must be corrected. A noncompliant product entails large financial and image burdens for the company. Defects slow down production and increase lead time. If even the defects are detected by the customer, the costs increase further,

47 The Adoption of Industry 4.0 Technologies Through the. . .

1195

as it becomes necessary to set up a structure able to manage complaints, bear the costs deriving from repairs, disassembly, and reassembly and delivery. The main causes of such waste can be identified in the badly organized layout, in the lead times that are too long, in the inadequacy of the production process, in inadequate maintenance, in poor work procedures (Visual Management), in the lack of personnel training (Skill Matrix), poor supervision skills, incorrect product or production process design, lack of performance indicators, inefficient planning and programming of production, inadequate equipment, lack of organization of the workspace (methodology 5S), and in the quality of suppliers. However, it is necessary to underline that the loss of value does not depend only on waste but also on other negative elements, such as “Muda,” “Muri,” and “Mura.” The Japanese term Muda refers to the different types of waste, as described above, while Muri is the term that indicates the overload of people or resources. Overloading for people can lead, in the long term, to the possibility of occupational injuries or illnesses due to the overexertion to which workers are subjected. The effect is the absence from work for shorter or longer periods by the workers and general dissatisfaction of the staff. Similarly, the excessive exploitation of machinery can lead, in the long term, to accelerated wear, to breakages with consequent production stop for maintenance and repair, or even the need to change machinery may arise. Finally, Mura indicates fluctuations, variations, irregularities in the workload (of demand). These fluctuations lead to phases in which there is an overload of work (Muri) and to other phases in which the workforce and machinery are oversized (pauses are created – Muda) and, therefore, the production flow is disturbed. Instead, approaches to evaluate the efficiency of Lean Production have mainly three: the Kaizen, which refers to the quality management; SMED Techniques for what concerns setup times; the Total Productive Maintenance. Managing company activities according to the Kaizen philosophy means seeking continuous improvement, starting from the assumption that every action once improved can become a standard that can be further improved. In order to pursue the continuous improvement of performance, the involvement of all interested parties is necessary. While the SMED is a tool that is totally integrated within Lean Manufacturing and arises from the need to minimize the internal and external setup times of a machine. The purpose of this process is to be able to quickly switch from one production to another in the same plant. Reducing times means eliminating the adjustment actions on equipment, tools, machines and systems that do not bring added value to the finished product. Finally, Total Productive Maintenance (TPM) is an efficiency approach to infrastructure maintenance. It too is a philosophy of continuous improvement and teamwork, aimed at involving all operators, maintenance technicians, and supervisors so that they themselves can exercise direct control over the correct functioning of their machinery. One of the aims is to obtain an attitude of greater responsibility and attention toward the plants on which we work every day (Galgano 2002).

1196

M. R. Sessa et al.

It is an approach developed in the Japanese industrial concern in the early seventies, and today is widely used in a variety of organizations, because of considerable support in the total efficiency management plant them and, in general, the whole company. The spread of this approach is linked to the demand for greater productivity and flexibility and the containment of operating costs and inventory levels of products and materials. Therefore, the main objective is to reduce the number of failures and defects and at the same time to maximize the efficiency of the plants, through the involvement of the staff. The application of the TPM provides for: maximizing the overall efficiency of the entire production system; the promotion of an adequate infrastructure maintenance system; the empowerment of all business functions relating to maintenance; the active involvement of all internal stakeholders; the promotion of TPM through a motivating management, that is, for small groups. Furthermore, the TPM is based on some fundamental pillars (Fig. 1): Therefore, the objective of the TPM is to have plants and machinery that are able to guarantee the maximum reliability of the process in order to avoid the need for safety stocks and have more streamlined flows, a necessary condition for Just in Time production. Just in Time (JiT) is a logistical-production method aimed at eliminating stocks and inventories of material in the factory. It is based on the concept of producing only when needed. This way of organizing the production process, together with the adoption of ever smaller production batches allowed

Fig. 1 Total Productive Maintenance pillars. (Source: Lean Production, lean made easy by Vorne https://www.leanproduction.com/tpm.html)

47 The Adoption of Industry 4.0 Technologies Through the. . .

1197

by the introduction of quick setup techniques (SMED), eliminates or drastically reduces the stationary material waiting to be processed, thus reducing the total crossing time of the same material on the production line, which goes from days to hours. This approach, which is of fundamental importance in the broader Lean Production methodology, marks the transition from a push type logic, for which finished products are made to be used in stock, to a pull type, according to which production is started only on the basis of a real need and at the exact moment in which this need arises. In fact, the JIT consists of three elements: Pull system; OnePiece Flow system; Takt Time. With the Pull system , the progress of the production flow is guided by the customers: upstream goods are not produced until the downstream customer, both internal and external, requests them. The instrument that governs this system is the Kaban, a visual system that transmits a series of instructions communicating information on the materials to be procured or the components to be produced. While the second element that makes up the Just in Time approach is the OnePiece Flow system, which is a way to organize the advancement of materials, with the possibility of changing the product model at each step, in a continuous flow. In this way, the individual pieces pass from one production phase to another without accumulation between the machines, contributing to the reduction of the Time Line (the material passes through the departments in the fastest way), to obtaining maximum flexibility, important measure of intermediate stocks, to the recovery of physical space inside the line, thanks to the use of smaller machines, which are brought closer together due to the presence of small lots. Finally, the Takt Time is the parameter that usually links production to the market, that is, it is a number that expresses a time within which a unit of product must be obtained. It is essentially the pace of production. The last approach underlying the Quality 4.0 paradigm is Total Quality Management (TQM). Adapted at the end of the 1990s in Japan, the TQM is the fruit of the evolutionary history of the concept of Quality. With the advent of ISO standards since 1987, the concept of Quality takes root as a strategic investment for the organization, capable of generating profits. This was the fundamental step for a systemic vision of quality from which he finally came to the most recent programs based on dynamic processes aimed at continuous improvement of the effectiveness and efficiency of performance of an organization. In fact, contents relating to the conceptual framework of Quality have undergone a radical transformation over time. From a sectorial content (Product Quality) it has gradually expanded to take on global connotations (Total Quality), passing from a closed and static system, mainly aimed at analyzing the past, to an open and dynamic system, oriented toward the future. In short, we have moved from an orientation toward production to one toward the market and, in a broader vision such as the one currently affirmed, to a perspective that is also attentive to the balance and protection of the natural environment, to solidarity and social cohesion, in the awareness of the interdependence and complementarity between the management of aspects of quality, the environment, safety, and corporate social responsibility (Proto et al. 2010).

1198

M. R. Sessa et al.

And this is how a truly broad vision of Quality has gradually established itself, which goes beyond the traditional economic approach, as it is aimed at satisfying the expectations of all interested parties. The implementation of this methodological approach requires the adoption of a series of operational tools, including the Deming Cycle and the Seven Tools, designed to manage the activities connected with the improvement programs, whose main objective is the removal of any causes of problems. The area where these fall methodological approaches is the “Operational Excellence,” a term which indicates the set of methods and tools by which an organization can set their own goals in light of the culture of continuous performance improvement. Typical examples of Operational Excellence models are the X-Production Systems, which is the model that a generic X organization can adopt to improve and maintain the efficiency and competitiveness of the performance of its operations. The most accredited Operational Excellence model is certainly World Class Manufacturing (WCM), which will be discussed more fully in the next paragraph.

The Implementation of WCM Through Quality 4.0 Principles World Class Manufacturing can be understood as the most significant and typical model of the Lean methodological approach. Schonberger (1987), in the 1980s, defined WCM as a continuous rapid improvement in the area of quality, investments, and flexibility of an organization, so that continuous improvement is possible by removing all obstacles to production, in order to achieve maximum simplification. In this context, the participation of personnel is of fundamental importance, who, through their cognitive abilities and, in particular, their soft skills, will be able to positively influence the production process. But the most accredited definition is certainly the one theorized by Yamashina (2000), in the early 2000s, according to which the WCM has undergone a rapid evolution on a global level, as a tool capable of developing the competitive advantage of an organization. Over the years these first interventions on the definition of the WCM have resulted in others who have tried to deepen and refine the initial definition, as well as to develop principles and technical tools aimed at achieving excellence in manufacturing. The spread of this model in Italy began in 2005, when FCA decided to adopt the WCM methodology to manage its production in the various Fiat factories, in order to respond adequately to the continuous changes in the production system in the national and international context. This means focusing on a greater speed of response, an increase in operational efficiency and the logic of flexibility, through production systems that are highly synchronized with market demand and characterized by a strong focus on reducing waste in the process. Therefore, WCM can be understood as a structured production system that promotes long-term systemic improvements, aimed at evaluating and reducing all

47 The Adoption of Industry 4.0 Technologies Through the. . .

1199

Fig. 2 World Class Manufacturing Structure. (Source: Evoluzione della fabbrica. Dalla fabbrica tradizionale a quella in ottica WCM (http://www.provincia.torino.gov.it/fidati/orientarsi/archivio_ orientarsi/dorientarsi/evoluzione_fabbrica.pdf))

types of losses and wastes that can be detected, applying methods and standards with rigor and through the involvement of all the arts concerned (Fig. 2). The WCM is inspired by some basic concepts, defined in this work as Quality 4.0 methodological approaches and schematically shown in Fig. 3. This model refers to the activities and organization of the entire factory: from the management of workers’ health and safety, to the management system for Quality in the broadest sense of the term, to maintenance systems and workplace organization, logistics, and the environment. Therefore, WCM is mainly based on the concept of continuous performance improvement which tends to achieve recognized and certified excellent results. The main feature of the WCM is to seek the “Zero,” that is to eliminate all waste and losses through a program of continuous improvement of all company performance and the involvement of all internal stakeholders. This orientation can be summarized in the following factors: 1. 2. 3. 4. 5.

Zero customer dissatisfaction Zero misalignments Zero bureaucracy Zero shareholder dissatisfaction Zero waste

1200

M. R. Sessa et al.

World Class Manufacturing

TQM

TIE

TQC

TPM

TIE

Total Industrial Engineering

Total Quality Control

Total Productive Maintenance

Just in Time

Productivity

Quality Improvement

Zero Waste

Zero Defects

Focus

Goals VALUES

Cost Deployment

Method

Technical Efficiency

Zero Failures

Service Level

Zero Stocks

People Involvement, Value Creation, Customer Satisfaction

Fig. 3 The WCM structure according to the logic of Total Quality Management. (Source: own processing)

6. 7. 8. 9.

Zero work that does not create added value Zero stops Zero missed opportunities Zero information lost

The pursuit of these factors will allow the organization to obtain the excellence of the entire production cycle in terms of quality and efficiency that will determine the company’s success. This means that it is essential to follow certain procedures, such as punctuality in deliveries to customers; knowledge of key customers and the strengths of the products they make; a team that is attentive to customer needs; the elimination of final storage; the reduction of setup times (SMED techniques); continuous training of its employees; the elimination of any action that does not bring added value to the customer; the search for total quality through the use of statistical control tools; the engineering of critical operations; preventive and proactive maintenance of machinery (Total Productive Maintenance); maintaining clean and tidy work areas (5S methodology); the continuous interface between production and design; the reduction of paper reporting and the Supply Chain controlled through the Kanban system. Therefore, certainly WCM is a methodology that recalls other typical approaches of Lean Production but a particular characteristic that makes it unique is its structure in pillars and steps of a multiplicity of technical and managerial activities. In fact, the WCM is compared to a temple (Fig. 4), whose columns represent the technical pillars while the managerial aspects are placed at the base of the pillars.

47 The Adoption of Industry 4.0 Technologies Through the. . .

1201

Fig. 4 The temple of World Class Manufacturing. (Source: own processing)

Each technical pillar refers to a specific area of the production system, while the managerial factors are reflected in the management methods of the entire production system of the company. This is a path that must be followed at every stage, so it is not possible to imagine developing one technical pillar rather than another. On the contrary, all the technical pillars should be developed, although they are independent of each other and that there is close collaboration and communication between the different areas of the system, in order to add the objectives of the WCM. In addition, through the Cost Deployment matrix, the areas on which each pillar must actively intervene are identified and, at the same time, once a model area has been identified, staff are trained on aspects related to the design, guidance, support, and monitoring of the development of their own area in the company. In fact, the WCM approach begins with learning in the model area, which will then be extended to other areas of the system, until it covers the entire organization. The development structure of a pillar is to establish the basic conditions, define the objectives and the improvement plan of the model area, implement the improvement in it, review the objectives and plan the extension, and develop the extension improvement in the entire production system. These actions, aimed at pursuing continuous improvement, require the use of a methodological approach to measure the different company performances. In this case, the most used indicators are KPI (Key Performance Indicator) and KAI (Key Activities Indicator). With the former it is possible to represent the degree of improvement of traditional variables of a production system, such as profit, sales, and product quality rate. While the latter measure the actions and effort required to achieve an improvement goal. It is essential to choose the most appropriate indicator to be able to quantify the progress, in terms of continuous improvement, of the actions to change a specific performance. Therefore, each step of each single pillar will be considered concluded only after the actual achievement of the set objectives

1202

M. R. Sessa et al.

is established during the audit. Once this has been verified for a certain step and, therefore, considered that it has been completed, you can move on to the next one. The ten technical pillars of the WCM are the following: – Safety. It refers to the safety and health of employees in the workplace, for which it is necessary to adopt tools aimed at the management and prevention of risks for humans. – Cost Deployment. This pillar allows you to analyze costs and benefits deriving from your production system, as well as any losses of the organization. The Cost Deployment analysis is foreseen for each single pillar, as it is necessary to identify for each one which are the activities that must be improved as a priority in order to reduce costs. – Focused Improvement. Its main objective is to eliminate the main loss items identified through the Cost Deployment analysis, avoiding to allocate effort and resources toward nonpriority issues. – Autonomous Activities. It is divided into activities, such as Autonomous Maintenance and Workplace Organization. The first has the objective of eliminating the tastes due to the lack of maintenance of the basic conditions that generate deterioration of machines and equipment. While the second has as its objective the improvement of the work areas, trying to eliminate waste and losses present in the production process, often due to deteriorated equipment. – Professional Maintenance. This pillar refers to all those activities aimed at creating a maintenance system capable of minimizing the defects of the equipment, trying to extend the useful life of the machines and their components. – Quality Control. The main objective of this pillar is to satisfy the customer’s needs through the continuous improvement of performance. It is a systemic, logical, and detailed approach to reduce quality defects deriving from the machine, production method, manpower, and material, which represent the main causes of variability, that is, the shift from optimal production conditions. – Logistics and Customer Service. This pillar pursues the objective of managing the internal flow of the production process, also improving it through the involvement of external stakeholders, such as logistics. In order to optimize customer service and minimize logistics costs, the production and distribution processes must be integrated with each other to allow the fastest and most effective flow of products from upstream to downstream. This means operating from a Just in Time perspective, or offering the right product, at the right time, in the right quantities and conditions. – Early Equipment Management. This pillar aims to optimize costs and installation times for a new machine or production line. – People Development. In this case, reference is made to the continuous improvement of employee skills, for which an adequate training system is required as well as a system capable of classifying the formal, nonformal, and informal skills required for each specific task and reducing the gaps and/or skills gap. Furthermore, this pillar can support the reduction of human errors, the ability of individuals to carry out autonomous maintenance, the development of highly

47 The Adoption of Industry 4.0 Technologies Through the. . .

1203

qualified technical personnel, and the assumption of responsibility for continuous improvement by personnel. – Environment. The last pillar refers to the satisfaction of environmental management requirements, in compliance with mandatory and/or voluntary environmental regulations and standards, aimed at the continuous improvement of environmental performance. The standards commonly used to implement management systems for environmental quality are ISO 14001 and the EMAS regulation. While managerial pillars, which apply within each pillar technical and reflect on the involvement of the entire factories, are as follows: – Management Commitment. Managers must know how to use the key tools of this methodology and must be able to identify and transmit the objectives that the company has set itself, with the related actions to be taken, even at the lowest levels. Furthermore, management must be able to delegate responsibilities to employees to facilitate the creation of an attitude that is inclined to initiative and autonomy. The entire workforce must collaborate in identifying waste and losses with the identification of their causes. – Clarity of Objectives (KPI). The goals set by the company must be Smart, that is, specific, measurable, achievable, stimulating, and time-based. Everyone inside the plant must be aware of the path taken and the result to be achieved. To make the communication of objectives more effective and direct it is necessary to set up specific areas in the plant in which the various projects undertaken to eliminate losses and waste and the results obtained are indicated. – Route Map of WCM. It corresponds to a time map in which the results that each single area must achieve are specified. The Route Map of each pillar must be directed, specific, feasible, and desirable. – Allocation of Highly Qualified People of Model Areas. It is necessary, especially in the early stages of WCM implementation, to know how to choose and place the most qualified people. These resources must learn the concepts and tools of the methodology in the field and then transfer this knowledge throughout the plant. – Commitment of the Organization. An active involvement of the entire organizational system must be created, starting from a change of mentality. The organization must be aware of existing problems and have an open mind, willing to collaborate in the activities of World Class Manufacturing and predisposed to improvement. – Competence of Organization toward Improvement. The company must know how to spread the tools necessary for the implementation of the methodology, starting from the basic ones and then moving on to the more advanced ones. – Time and Budget. It is very important to quantify the improvement projects economically and temporally in order to provide a clear indication of the direction to take, the means to use, and the resources to reach the goal. – Level of Detail. It is necessary to have an excellent level of detail in the collection of data for the identification of waste and losses, thus allowing to focus attention

1204

M. R. Sessa et al.

and resources on the costs that have a more negative impact on the plant. The possession of a high level of detail is required every time it is necessary to solve a problem, in order to eliminate the real root cause. – Level of Expansion. The knowledge and notions learned in the model areas must be implemented and adapted to the entire production system. The implementation will end only when all wastes and losses are eliminated. – Motivation of Operators. Operators must be the owners of machinery and equipment, actively participating in continuous improvement. The tools and knowledge necessary to recognize anomalies and resolve them in a timely manner are provided. In this way a motivation is created in the operators that push them to always give their best. Therefore, in light of this, it is possible to state that the strengths of the World Class Manufacturing methodology are manifold. One of the main benefits deriving from the adoption of this methodology lies in the area of safety, for which the priority objective of the WCM is not the reduction or elimination of losses and waste, but the achievement of zero serious accidents and injuries. A further benefit can be identified in the clear and clear division of the tasks and objectives to be achieved required of each worker or work group. In fact, each of the WCM pillars focuses its attention on a specific area of expertise. The path that each working group will have to follow is known from the beginning, thanks to the characteristic steps of each pillar. Furthermore, through the WCM there is a precise identification, for each single pillar, of the losses to be resolved. The priorities are identified based on an economic evaluation obtained thanks to the work of the Cost Deployment. This allows us to focus attention and efforts on the cause which, for each pillar, has the highest financial loss. This avoids the investment of time, resources, and means on factors that do not have a significant impact on the loss of the single pillar. Furthermore, World Class Manufacturing allows to acquire the concepts and tools of the methodology directly in the field. Once the model area of each pillar has been identified, the concepts and means learned in the theory are put into practice. Therefore, a learning site is created which will allow to extend what has been learned in other areas of the company, thus allowing to reduce implementation times. Finally, each pillar is made up of a team of trained and competent people. In each work group the roles and tasks of all individuals are clear, coordinated by a Pillar Leader who turns out to be the one who possesses the greatest knowledge and ability in using the characteristic tools of that pillar. Each team collaborates, in the different steps, with other pillars, thus allowing to create multi-skills working groups.

Introduction of RAMI 4.0 for Intelligent Production As already stressed several times above, Industry 4.0 can be understood as the implementation of the smart factory to provide smart products and services that meet the individual needs of consumers.

47 The Adoption of Industry 4.0 Technologies Through the. . .

1205

As was the case for the first three industrial revolutions, also in the context of Industry 4.0, technical innovation integrates itself, vertically and horizontally, into production systems, so that continuous digital engineering is developed for the entire cycle of life of the product and there is a decentralization of IT resources. This means that to enable such an intelligent network, new technologies such as modern enabling technologies and flexible hardware and software interfaces are needed. It is only by equipping products and production environments with these new technologies that Industry 4.0 will offer multiple opportunities to organizations that decide to adopt this paradigm, in order to improve efficiency and make production processes more flexible. This will allow for an increase in the value of products and the development of new business models. Therefore, the skills of Industry 4.0 in the field of communication technologies, digitalization, and IT infrastructure in companies, considered key enabling technologies, must be promoted to strengthen the implementation of the Industry 4.0 paradigm, and finally, to fully realize the potential benefits (Anderl et al. 2015). In this regard, to support research, standardization processes, and operators in this area, Platform Industry 4.0 in 2015 developed the Reference Architectural Model Industry 4.0 (RAMI 4.0). Plattform Industrie 4.0. RAMI4.0 - a reference framework for digitalisation (https://www.plattform-i40.de/PI40/Redaktion/EN/ Downloads/Publikation/rami40-anintroduction.pdf?__blob=publicationFile&v=7). The model allows to classify and identify the production areas to be allocated to Industry 4.0, creating a solid basis for the further development of technologies in the production system. However, to date, the model itself is quite abstract and its application in practice still too complicated. At the moment the use of RAMI 4.0 is mainly limited to research institutions and first individual use cases. Due to the abstract design of the model and its embryonic nature, in this contribution only the generic applicability of RAMI 4.0 will be considered, in order to understand the real benefits of its hypothetical implementation in an innovative business reality, such as that of LEONARDO. The RAMI 4.0 model, (Alignment Report for Reference Architectural Model for Industrie 4.0/Intelligent Manufacturing System Architecture. (https://www.platt form-i40.de/PI40/Redaktion/EN/Downloads/Publikation/hm-2018-manufacturing.p df?__blob=publicationFile&v=5)) whose application ranges from production to process technologies, is articulated on a three-dimensional space in which the hierarchical levels of a production plant connected to the Internet are represented, the life cycle of plants and products as well as the IT representation of an Industry 4.0 component. The hierarchical levels basically correspond to the levels of the automation pyramid (Fig. 5). L’asse 1 – Hierarchy levels can be represented as follows (Fig. 6). These levels represent the functional characteristics of the factory components and are defined according to the international standards IEC 62264 and IEC 61512. They can be understood as the first axis of the RAMI 4.0 model. The lowest level, called Product, includes products that, thanks to their ability to communicate, are active elements within the production system. They provide information on their individual properties and the necessary production steps. The

1206

M. R. Sessa et al.

LAYERS HIERARCHY LEVELS (IEC 62264 // IEC 61512) Business Functional

Co En nnec t t Wo erpri Wor ld se r k Sta Un Co tion its Fie ntrol D Pro ld De evic du vice e ct

Information Communication Integration Asset Development

Maintenance Usage

Type

Production

Maintenance Usage

Instance

LIFE CYCLE & VALUE STREAM (IEC 62890)

Fig. 5 The structure of the RAMI 4.0 model. (Source: own processing on https://www. plattform-i40.de/PI40/Redaktion/EN/Downloads/Publikation/rami40-an-introduction.pdf?__blob= publicationFile&v=7) Entreprise

Connected World

Work Centers Station

Smart Factory

Control Device Field Device

Smart Products

Product Industry 3.0

Industry 4.0

Fig. 6 First axis of RAMI 4.0 model. (Source: own processing on https://www.plattform-i40.de/ PI40/Redaktion/EN/Downloads/Publikation/rami40-an-introduction.pdf?__blob=publicationFile& v=7)

Field Device level includes intelligent field devices such as sensors and actuators. While the Controller level refers to the controllers in turn, the built-in controllers and controllers. Instead, intelligent production machines, robots, or logistics vehicles are located on the Station level. Additionally, both manufacturing facilities and entire departments within a company are assigned to the Machining Centers level. The Enterprise level considers the business organization as a whole and the Connected World level represents its external networks, that is, collaboration with business

47 The Adoption of Industry 4.0 Technologies Through the. . .

1207

Business

Organisation and business processes What is the customer willing

Functions

Functions of the asset

What is my product supposed to do?

Necessary data

What data does my product have to provide?

Access to information

How do I or my customer access the data?

Information

Communication

to pay for?

Which parts of my product

Integration

Asset

Integration of assets into the world of are digitally available in the network?

“Things” in the physical world

How do I integrate my product with the process to move it in the real world?

Fig. 7 Vertical axis of the RAMI 4.0 model. (Source: https://www.plattform-i40.de/ PI40/Redaktion/EN/Downloads/Publikation/rami40-an-introduction.pdf?__blob=publicationFile& v=7.)

partners, customers, and other interested parties as well as services based on the use of the Internet. The vertical axis 2 (Fig. 7) of the model represents the structural properties of an asset (component of the factory) or of a set of assets, for which each component is made up of six layers. The lowest layer Asset refers to the resources, or the representation of physical reality. It contains all physical objects such as machines, sensors, and documents as well as human resources. As well as intangible objects, that is, models, ideas, or patents are similarly attributed to this layer. The Integration layer supports the provision of computer-usable information about physical resources, hardware and software, of the layers above. It contains all the elements associated with IT and generates events based on the acquired information. The integration layer performs the final check of the technical processes. The purpose of the Communication layer is to enable communication between the different elements of the network based on uniform communication protocols and data formats. It also provides services to control the level of integration. Within the Information layer, the data used, generated, or modified by the technical functions of this layer is processed. To do this, the data is checked for integrity, summarized into new, higher quality data, and made available to higher levels via interfaces. Events are received by the communication layer, transformed, and forwarded accordingly. While the Functionality layer represents the runtime environment for services and applications. It is the platform for the horizontal integration of the various functions and generates rules and application logic. Remote access and integration

1208

M. R. Sessa et al.

Fig. 8 Horizontal (third) axis of the RAMI 4.0 model. (Source: https://www.plattformi40.de/PI40/Redaktion/EN/Downloads/Publikation/rami40-an-introduction.pdf?__blob=publication File&v=7.)

of applications and functions occurs only in this layer, without interfering with the underlying layers and ensuring the integrity of the information. Finally, the Business layer refers to the management of the commercial aspects of an organization. It also provides legal and regulatory frameworks to follow and ensures the integrity of functions throughout the value chain. The third axis (Fig. 8), defined in accordance with the IEC 622890 standard, describes the life cycle and the value chain of an Industry 4.0 component. This axis presupposes a basic distinction between two concepts of fundamental importance: yype of product and Instance. While the term Type refers to something that exists, starting from the basic idea of the product, the taking of the order, and the development of the product up to the production of prototypes. After all testing and validation, the type is prepared for mass production: in fact, the type of any component creates a basis for serial production. Each product manufactured represents an instance of that type, for example, it has a unique serial number. An Instance involves the transition from design to production after successfully running a series of tests. The manufactured product thus represents the instantiation of the type. The change from Type to Instance can be repeated several times. The horizontal axis structure shows a division of the Type in development, maintenance, and use, while Instances consist of production, maintenance, and use. The function of the layers in the left horizontal axis can be explained in the following example: the development of a new electric drive represents the creation of a new type of motor. The drive (controlled motor) is developed, initial samples are installed and tested, and a first series of prototypes are produced and validated. After successfully passing the tests, the new unit type is released for sale (product designation in the manufacturer’s sales catalogue). A first series production can be started at this time. Each unit in serial production has its own serial number (a unique identification) and is an instance of the previously developed electrical unit. Customer feedback to type instances can lead to the implementation of corrective actions in the production

47 The Adoption of Industry 4.0 Technologies Through the. . .

1209

process. These are changes in the type, that is, they are applied as changes to the type documentation and new instances of the changed type are produced. The left side of the RAMI 4.0 model also represents the value chain, so the connection between digitalization and the work chain (in the idea and practice of Industry 4.0) represents a great potential for continuous improvement of product types. The value chain in fully digitalized production will allow the connection between the different industrial areas: purchasing, order planning, assembly and assembly, logistics, maintenance, customer and suppliers, and so on (Zezulka et al. 2016). The model thus described does not yet allow to obtain information about the importance or the level of use of the same in practice, but could be populated with the technologies of Industry 4.0 and to observe the related interrelations, in order to innovate production processes, within the 4.0 paradigm. Therefore, the RAMI 4.0 reference architecture model could support the implementation of the main aspects of the Industry 4.0 context, providing organizations – determined to orient their production systems toward these principles – with an overview of the technologies enabling the fourth industrial revolution. Indeed, the model integrates different perspectives for users and offers a common way of seeing the technologies of the fourth industrial revolution. With RAMI 4.0, industry requirements – from manufacturing automation and mechanical engineering to process engineering – can be addressed in industry associations and standardization committees, providing a common understanding of standards and use cases (which to date still do not find a wide diffusion). The RAMI 4.0 model can be considered a map of Industry 4.0 solutions together with national and international standards, or rather an orientation to trace the needs of the sectors to define and further develop the new paradigm. The model ensures that all participants involved in Industry 4.0 procedures, processes, and activities have a common structure and terminology. Therefore, this architecture can represent, in line with the methodological approaches described above (in particular to the WCM), a useful support to the objective of continuous improvement of company performance through the use of the “enabling” technologies of Industry 4.0. Even if the model is in these years taking its first steps and, in particular, it can affirm that LEONARDO might represents king one of the first cases of use of such an architecture, helping to define a model of management Industry 4.0 holistic and adaptive for industrial automation.

WCM for the Intelligent Production in LEONARDO of Tomorrow In light of what has been presented up to now, in particular, taking into consideration the main characteristics of WCM and RAMI 4.0, we will proceed, in this paragraph, to show which are the most appropriate approaches and managerial tools – Quality 4.0 – to implement process innovations and address the concrete needs of the future production needs of the LEONARDO industrial sites. Following the analysis of the LEONARDO context, a series of macro-objectives have been identified that can be

1210

M. R. Sessa et al.

divided into “horizontal” or “transversal” and “vertical.” The “horizontal” macroobjectives are the basic ones that the LEONARDO solution will have to address regardless of the peculiarities of the plant and on which all the services and vertical applications that will be created will rest. These include: – Cyber Security, for the protection of acquired information. – Cloud Services, relating to the ways in which LEONARDO will be provided. By “vertical” macro objectives, on the other hand, we mean those functional areas that characterize the actual functional offer provided by the final platform, which have a direct impact on those who work in the new generation manufacturing industry and which are an integral part of the Industry 4.0 paradigm. Among these, the following main vertical macro-objectives are highlighted: – Monitoring of the production process: control of the operation and verify the efficiency of the machines (OEE = Overall Equipment Effectiveness) – Energetic efficiency of machines and installations: measurement of absorption parameters of the machines, identification of optimization of energy consumption for the purpose of energy saving solutions and the sustainability of the factory – Predictive maintenance and cleaning of machines and plants: acquisition of historical parameters of operation of the machines, machine application techniques learning, introduction of predictive maintenance policies, aimed at anticipating possible failures and malfunctions – Tracking materials: application of “smart tag” to the materials/semi-finished products for tracking by the phase incoming material in the plant, to productive process or assembly and of the systems, including the operating phase of the system – Control quality of production: solutions for the control of the production process and its realignment during the execution of the production itself, in order to avoid waste – A predictive analysis undergoing testing, certification, and after sales systems: solution similar to that for the predictive maintenance of machines and installations, applied to the LEONARDO products (complex systems in the process of testing and certification or in operation) – Solutions for mobile working: technological solutions support all efficiency operative operator (actually increased, virtual, mixed reality, instant messaging) Therefore, taking into consideration the characteristics of Industry 4.0 and Lean Management, as a useful theoretical-conceptual background of reference, we want to highlight the relationship that could arise from their relationship, taking into consideration the staff, products, suppliers, infrastructure, and operation of business operations, and then try to transfer the implementation of such a combination at industrial sites in LEONARDO. With respect to the personnel sphere, belonging to an organization oriented to accept the principles of Industry 4.0, some innovations can be detected, such as

47 The Adoption of Industry 4.0 Technologies Through the. . .

1211

an autonomous detection of problems using the Jidoka tool, greater integration with production systems and personal devices and training continues on the characteristic aspects of the Industry 4.0 paradigm, for which greater skills, competences, and flexibility are assumed in relation to the role one covers and the task to which one is called Enke et al. (2018). In this case, tools such as electronic Kanban for communication and the CPS system are used in the workstation to integrate into the entire production system and respect the Lean elements of production on demand. While, as far as the product is concerned, in Industry 4.0 the focus is on a personal and individual production, through the detection of the needs of individual customers. Also in this area, the principles of waste reduction are followed according to Lean logic but, with the introduction of new technologies, the aim is to increase the value of the product for the customer, improving response times and adapting to an increasingly competitive market. The relationship of mutual benefit with suppliers is fundamental, with which a flow of information and data exchange must be established for the proper functioning of the company system and the development of the product itself. Therefore, it is important to introduce Just in Time into the production system, in order to improve the synergy between the different work sectors, or between the production line and logistics. Another important sphere to be taken into consideration in the business system is that relating to infrastructures, that is, machinery, equipment, and various physical supports. These means in Industry 4.0 operate in Pull, placing emphasis on the ability to execute production in a flexible and lean manner. This is possible by means of a real-time communication, using for example the RFID technology, which allows to continuously adapt the production to the downstream request. Furthermore, in this context, great emphasis is placed on the maintenance of machinery and equipment with methods such as TPM and SMED techniques, so it is possible to carry out more targeted and punctual interventions also through the so-called autonomous maintenance performed directly by the machinery operator. In general, it can be said that the transition from Industry 3.0 to Industry 4.0 and, from the more traditional management tools to those of Lean and Total Quality Management, is marked by the objective of continuous improvement of company performance. This improvement activity in Industry 4.0 can be supported by means of the availability of a large amount of data and information: Big Data. Big Data analysis, through complex simulations on the collected data and virtual models, allows to find solutions to various problems that can be detected in the company system and, therefore, to improve the expected results according to the logic of Kaizen. A further enabling technology, the IoT, always in accordance with the principles of lean production, will allow the creation of a decentralized control system, adopting devices that communicate locally with each other. In fact, the Internet of Things guarantees greater and faster communication between machinery and operators and between machinery and machinery and/or operators and operators belonging to different work areas. This implies greater coordination in carrying out the production process and communicating with operators who, among other things,

1212

M. R. Sessa et al.

Table 1 Summary features of the combination Industry 4-0 – Lean Management Lean Production Tools and Principles Produced in value added

Solutions from Industry 4.0 to Support Availability of customer feedback through the Cloud Unique identification of the product, advanced customization Concurrent Engineering, virtual prototyping and Big Data Reduction of Mapping of value streams through continuous data collection costs Identification of activities that bring value with Big Data processing Full integration, horizontal and vertical in the company Elimination Virtual model and physical consideration in the CPS of waste Process and inventory optimization through Big Data Real-time assessment of the state of the establishment Reduction of inaccuracies and errors in communication and data recording Transparency Real-time control and display of the status of the plant and Decentralized and robust system to problems in the network reliability of operations Pull and Electronic versions of the Kanban between workstations, machines, and Kanban operators production Continuous control of workstation consumption and autonomous maintenance Cloud of requests from customers and suppliers, from downstream to upstream Unique addressing of material and data between production resources Just in Time Status of the Supply Chain in real time, transport optimization Coordination with suppliers and between workstations via Cloud and IoT Maintaining flows of materials and “queue” control in stations One-piece RFID integration in components, readers on machinery flow Autonomous reconfiguration of machinery for each product Handling with dynamic paths for components or products SMED Visual support in the execution of manual work and automatisms Systemic review of setup procedures, proactive maintenance Machines, workstations, and plug-n-play components Jidoka Continuous control of systems status to detect anomalies Increase of machines controlled by a single operator Notices addressed directly to competent personnel Poka-yoke Correctness control of manual and automatic activities Visual Computing to instruct and guide the activities to be carried out Removal of errors and unreliability of manual data collection Andon Monitors or mobile devices with contextualized information Provide information, fulfill requests and changes for each subject/object in the corporate social network TPM IoT devices and sensors for data collection, virtual models in the CPS Big Data for the definition of intervention strategies Remote service and control by suppliers Heijunka Sequences to optimize the integration between market and production Automatic setup and use of mixed model productions

47 The Adoption of Industry 4.0 Technologies Through the. . .

1213

Table 1 (continued) Lean Production Tools and Principles Standard operations Kaizen

Solutions from Industry 4.0 to Support Augmented Operator to carry out more diversified activities Dynamic timing Virtual learning and augmented reality Decentralized IoT system to communicate and optimize operations Make reliable databases available for improvement activities Data mining to support business decisions

Fig. 9 Mapping between identified macro-objectives and LEONARDO industrial sites. (Source: own processing)

could supervise a greater number of machinery, in order to keep the production flow constant. Furthermore, from the adoption of the Industry 4.0 paradigm, the benefits deriving from opening up to Cloud Computing and the service market should not be overlooked. The following table (Table 1) presents a summary relating to the binomial Industry 4.0 – Lean Management, in which the possibility of using the typical tools of lean production in symbiosis with the enabling technologies has been identified, in order to obtain greater benefits and respond more adequately to the challenges of the future. Therefore, with respect to the macro-objectives identified and taking into account the peculiarities of the various industrial sites, it is possible to hypothesize the implementation of the technical pillars underlying at the basis of WCM, as an Operational Excellence tool of fundamental importance for the pursuit of the continuous improvement of company’s performance (Fig. 9 and Table 2). In order to create a technical solution to the macro-objectives identified, only the technical pillars of the WCM are taken into consideration. For the managerial

1214

M. R. Sessa et al.

Table 2 Mapping between the macro-objectives of LEONARDO and its industrial sites Macro-Objectives Production process monitoring (OEEE)

Industrial Site Site 1 Site 2 Site 1 Site 1 Site 1 Site 2 Site 3 Site 1

Energy efficiency of machines and plants Predictive maintenance of machines and plants Materials tracking Production quality control Predictive analysis, testing, certifications, and after sales Solutions for mobile worker

Management

Clarity of

Route Map of

Allocation of Highly

Commitment

Objectives

WCH

Qualified People

Production process monitoring (OEEE)

Energy efficiency of machines and plants

Commitment of Organization

Preditive maintenance of machines and plants

Competence of Organization

Materials tracking

Time and Budget

Productive quality control

Level of Detail

Level of Expansion

Predictive analysis, testing, certifications and

Environment

People Development

Early Product/Equipment Management

Logistics

Quality control

Professional Maintenancee

Autonomous Maintenance & Workspace Organization

Focused Improvement

Cost Deployment

Safety / Hygiene & working environment

World Class Manufacturing

Motivation of Operators

Solutions for mobile worker

Fig. 10 The support of the WCM model for the development of macro-objectives in LEONARDO. (Source: own processing)

pillars there is no figure of the Leader Pillar who only deals with the management of the area for which he is responsible, but is the plant manager responsible for all the pillars. Furthermore, the managerial pillars make more reference to elements that could be defined as intangible, since it involves managing attitudes such as awareness, responsibility, and motivation that are part of that wealth of skills, not always observable and evaluable, of the human being. Therefore, the figure (Fig. 10) shows what are the specific technical pillars that you need to consider and implement using Lean tools, responding adequately to the macro-objectives identified, the understand the importance of introducing standardized procedures in the production system and increase one’s competitive advantage at national and international level. The technical pillars thus identified involve following a series of steps to respond to the macro-objectives of the LEONARDO.

47 The Adoption of Industry 4.0 Technologies Through the. . .

1215

The Autonomous activities pillar is divided into the Autonomous Maintenance and Workplace Organization sub-pillars. For the first sub-pillar, the characteristic steps to follow, for its correct implementation in the company reality, are the following: 1. 2. 3. 4. 5. 6. 7.

Initial cleaning and inspection Measures against sources of contamination Initial standards General inspection Autonomous inspection Improvements of standards Fully applied self-management system

In this case, the managerial tools useful for the actual application of these steps in the macro-objectives identified are the Key Performance Indicators and Key Activity Indicators. Compared to the second sub-pillar, the fundamental steps to follow are: 1. 2. 3. 4. 5. 6. 7.

Initial cleaning Reordering of the process Initial standards Product characteristics training Supply of materials in Just in Time Improvement of the standard Sequence of work standards

For the proper functioning of this pillar it is possible to adopt the Muda, Muri, and Mura analysis, mainly aimed at eliminating waste. The activities of this pillar are in close synergy with the activities of the Logistics pillar, in order to reduce waste deriving from the movement of materials within a process and between the various processes and to create, therefore, a Golden zone. With reference to the pillar of logistics and service satisfaction, the steps to follow for its correct implementation are as follows: 1. 2. 3. 4. 5. 6. 7.

Re-engineer the lines to satisfy the customer Refit the logistics internal Refit the logistics external Leveling the production Refine internal and external logistics Integrate sales, production, and purchasing networks Adopt a sequence programming – fixed time

The tools to be used for managing the activities of this pillar are, mainly, the Value Stream Mapping, the Milk Run, the Kanban, and the classification of materials.

1216

M. R. Sessa et al.

While, compared to the Professional Maintenance pillar, the steps to follow are: 1. 2. 3. 4.

Elimination and prevention of accelerated degradation Failure analysis, recovery and reversal of degradation Definition of maintenance standards Countermeasures on the weak points of the machines and extension of the average life of the components 5. Construction of a periodic maintenance system 6. Construction of a predictive maintenance system (trend management) 7. Management of maintenance costs, construction of a planned maintenance system The Lean tools that can be used in this area are OEE (Overall Line Effectiveness), MTBF (Mean Time Between Failure), and MTTR (Mean Time to Repair). Finally, for the implementation of the Quality Control pillar, it is necessary to follow the following steps: 1. 2. 3. 4. 5. 6. 7.

Select the topic Understanding the situation and goals Plan your activities Analyze the causes Define and implement countermeasures Check the results Standardize and institute control

For the management of the activities of this pillar it is possible to use different quality control tools belonging to the Total Quality Management philosophy. In particular, statistical tools (Seven Tools), monitoring tools, SMED techniques and managerial tools, or the Quality Function Deployment, Six Sigma, and Benchmarking, can be taken as a reference.

Conclusion Lean Management, in this contribution, represents the theoretical-conceptual framework for the implementation of the Industry 4.0 model. In fact, the Lean Management tools, implemented through Enabling Technologies, can support the management of business processes considering the operators, products, suppliers, production means, and the functioning of business operations. Therefore, based on the analysis of the LEONARDO context and taking into consideration the main production activities, the main Lean tools that can be used and the main Enabling Technologies can be applied are shown below, in order to respond to the macro-objectives identified with respect to the management of company processes (Table 3).

47 The Adoption of Industry 4.0 Technologies Through the. . .

1217

Table 3 Lean Management and Industry 4.0 in LEONARDO Industrial Sites Main Activities (Objectives) Site 1 Saving, displaying, and analyzing data relating to key parameters (e.g., % of occupancy of the individual machines, completion times of the single process, production waste on a single machine); AR solution with Smart Glasses to support work activities; preventive maintenance of machinery; bottleneck identification of the line Site 2 Study and analysis of AR and MR technologies to support testing, validation, and training in the production area

Automating the functional testing of assemblies and subassemblies by radio frequency via ATE (Automatic Test Equipment); CS (Compressive Sensing) in the reduction of antennas testing times; design of a remote test equipment command and control system; assembly and training with Mixed Reality Realization of a CAD/CAM environment for optical constructions; interface between measurement tools and simulation software; tracking management of materials in the production, transport and processing phases Automatic analysis of care cycles through Machine Learning; implementation of a tool life cycle monitoring system

Site 3

Tracking of materials and digitalization of logistics flows; predictive analysis to support the product design, test and certification phases.

Lean Tools and Principles Value added product; cost reduction; elimination of waste; Just in Time; Pull and Kanban production; TPM; Heijunka; Standard operations; Kaizen

Main Ket Big Data Analytics; Augmented e Mixed Reality; Horizontal & Vertical Integration; Industrial IoT.

Value added product; cost reduction elimination of waste; Just in Time; TPM; Heijunka; Standard operations; Kaizen Just in Time; TPM; Heijunka; Standard operations; Kaizen

Big Data Analytics; Augmented e Mixed Reality; Industrial IoT

Value added product; cost reduction; elimination of waste; TPM; Just in Time; Kaizen; Pull and Kanban production; One - piece –flow. Value added product; cost reduction; elimination of waste; SMED; Jidoka; One-piece-flow; Poka –yoke; Just in Time;TPM; Kaizen. Value added product; cost reduction; elimination of waste; TPM; Just in Time; Kaizen; One-piece-flow; Pull and Kanban production

Big Data Analytics; Industrial IoT.

Big Data Analytics; Augmented e Mixed Reality; Industrial IoT

Industrial IoT; Big Data Analytics.

Big Data Analytics.

1218

M. R. Sessa et al.

Therefore, the implementation of some of the technical pillars of WCM through Lean Management approaches and the operational support of the Enabling Technologies of Industry 4.0 will allow to improve and innovate the management of its business processes and respond adequately to the principles of the Industry 4.0 model. In this regard, the company should orient its activities and production area toward the Total Quality Management and Kaizen philosophy; extend this methodological approach to all production areas of the organization, in order to find improvement solutions that are not an end in themselves but useful for all areas of the plant; conduct a third-part external audit that will lead, in the coming years, the company to be considered a best in class according to the assessments of the World Class Manufacturing Association. The results of the analysis conducted so far constitute a preliminary evaluation of the research activity aimed at assessing the contribution of the Quality 4.0 approach to the Industry 4.0 model in LEONARDO. Possible research evolutions could foresee the adoption of the RAMI 4.0 model – still in a theoretical-conceptual definition phase – to integrate the WCM, in order to consider LEONARDO as one of the first cases of use of this architecture and to contribute to the creation of a standardized Industry 4.0 management model for the entire Italian industrial automation sector. Moreover, a further in-depth study in the future could be addressed the management of human capital according to the principles of the Semantic Web. It is, in fact, appropriate not to overlook the transversal skills of human resources considered a source of added value and creator of competitive advantages for organizations. For this reason, it is essential to create skill profiles appropriate to the tasks and roles assigned, in light of a greater need for specific knowledge not only of a declarative and procedural nature. Therefore, in the same way – by exploiting Enabling Technologies and Total Quality Management methodological approaches, in order to foster the recognition, classification, assessment, and certification of human capital skills and smart mobility of workers. Acknowledgments This research was supported by the MIUR (Ministero dell’Istruzione dell’Università e della Ricerca) under the national program PON 2014–2020, Leonardo 4.0 (ID ARS01 00945), and the ECSEL-JU under the program ECSEL-Innovation Actions-2018 (ECSELIA) for research project CPS4EU (ID-826276) (The chapter reflects only the author’s view. JU is not responsible for any use that may be made of the information it contains_. This project has received funding from the ECSEL Joint Undertaking (JU) under grant agreement No 826276. The JU receives support from the European Union’s Horizon 2020 research and innovation programme and France, Spain, Hungary, Italy, Germany

References Alignment Report for Reference Architectural Model for Industrie 4.0/Intelligent Manufacturing System Architecture. Aviable: https://www.plattform-i40.de/PI40/Redaktion/EN/Downloads/ Publikation/hm-2018-manufacturing.pdf?__blob=publicationFile&v=5. Last accessed: 15 Sept 2020

47 The Adoption of Industry 4.0 Technologies Through the. . .

1219

Anderl R, Picard A, Wang Y, Fleischer J, Dosch S, Klee B, Bauer J (2015) Guideline Industrie 4.0 – guiding principles for the implementation of Industrie 4.0 in small and medium-sized businesses. VDMA Forum Industrie 4:1–31 Bellini M (2016) Internet of things: gli ambiti applicativi in Italia, industry 4.0. Available: https:/ /www.internet4things.it/iot-library/internet-of-things-gli-ambiti-applicativi-in-italia/. Last accessed: 10 Sept 2020 Beltrametti L, Guarnacci N, Intini N, La Forgia C (2017) La fabbrica connessa. La manifattura italiana (attra) verso Industria 4.0. Guerini, Milano Biography of Kevin Ashton. Available: http://ethw.org/Kevin_Ashton. Last accessed: 10 Sept 2020 Bongio A, Distefano A (2015) Digital manufacturing: Cogliere l’opportunità del rinascimento digitale. Available: https://www.pwc.com/it/it/publications/assets/docs/digital-manufacturing.pdf. Last accessed: 10 Sept 2020 Brettel M, Friederichsen N, Keller M, Rosenberg M (2014) How virtualization, decentralization and network building change the manufacturing landscape: an industry 4.0 perspective. Int J Mech Aerosp Ind Mechatron Manuf Eng 8(1):37–44 Chen B, Wan J, Shu L, Li P, Mukherjee M, Yin B (2018) Smart factory of industry 4.0: key technologies, application case, and challenges. IEEE Access 6:6505–6519 Chui M, Manyika J, Bughin J, Brad B, Roberts R, Danielson J, Gupta S (2013) Ten it-enabled business trends for the decade ahead. Available: http://www.mckinsey.com/industries/high-tech/ our-insghts/ten-it-enabled-business-trends-for-the-decade-ahead. Last accessed: 10 Sept 2020 Dossou PE (2018) Impact of sustainability on the supply chain 4.0 performance. Procedia Manuf 17:452–459 Enke J, Glass R, Kreß A, Hambach J, Tisch M, Metternich J (2018) Industrie 4.0–competencies for a modern production system: a curriculum for learning factories. Procedia Manuf 23:267–272 Evoluzione della fabbrica. Dalla fabbrica tradizionale a quella in ottica WCM. Available: http://www.provincia.torino.gov.it/fidati/orientarsi/archivio_orientarsi/dorientarsi/ evoluzione_fabbrica.pdf Last accessed: 20 Sept 2020 Faller C, Höftmann M (2018) Service-oriented communication model for cyber-physicalproduction-systems. Procedia CIRP 67:156–161 Feeney AB, Frechette SP, Srinivasan V (2015) A portrait of an ISO STEP tolerancing standard as an enabler of smart manufacturing systems. J Comput Inf Sci Eng 15(2) Galgano A (2002) Le tre rivoluzioni in Le principali tecniche della produzione snella. Gruppo Galgano e associati 333–377 Golzer P, Cato P, Amberg M (2015) Data processing requirements of industry 4.0 – use cases for big data. ECIS 2015 research-in-progress papers, 61 Ivanov D, Dolgui A, Sokolov B, Werner F, Ivanova M (2016) A dynamic model and an algorithm for short-term supply chain scheduling in the smart factory industry 4.0. Int J Prod Res 54(2):386–402 Kagermann H, Helbig J, Hellinger A, Wahlster W (2013) Recommendations for implementing the strategic initiative INDUSTRIE 4.0. Final report of the Industrie 4.0 Working. Geschäftsstelle der Plattform Industrie 4.0, Berlin Kang H, Lee J, Choi S, Kim H, Park J, Son J, Kim B, Noh S (2016) Smart manufacturing: past research, present findings, and future directions. Int J Precis Eng Manuf 3(1):111–128 Krafcik JF (1988) Triumph of the lean production system. Sloan Manag Rev 30(1):41 Lean Production, lean made easy by Vorne. Available: https://www.leanproduction.com/tpm.html. Last accessed: 20 Sept 2020 Lee J, Bagheri B, Kao HA (2015) A cyber-physical systems architecture for industry 4.0-based manufacturing systems. Manuf Letters 3:18–23 Li B, Hou B, Yu W, Lu X, Yang C (2017) Applications of artificial intelligence in intelligent manufacturing: a review. Front Inform Tech El 18(1):86–96 Lu Y (2017) Industry 4.0: a survey on technologies, applications and open research issues. J Ind Inf Integr 6:1–10 McFarlane D, Sarma S, Chirn JL, Wong CY, Ashton K (2003) Auto ID systems and intelligent manufacturing control. Eng Appl Artif Intell 16(4):365–376

1220

M. R. Sessa et al.

Observatory of the School of Management of the Politecnico di Milano- Available: https://www.osservatori.net/it_it/osservatori/osservatori/internet-of-things Last accessed: 15 Sept 2020 Plattform Industrie 4.0. RAMI4.0 – a reference framework for digitalisation. Available: https://www.plattform-i40.de/PI40/Redaktion/EN/Downloads/Publikation/rami40-anintroduction.pdf?__blob=publicationFile&v=7. Last accessed: 15 Sept 2020 Proto M, Supino S, Malandrino O (2010) Sistemi e strumenti per la qualità. Percorsi evolutivi e approcci manageriali. Giappichelli Editore, Torino Richert AS, Shehadeh M, Muller SL, Schroder S, Jeschke S (2016) Robotic workmates – hybrid human robot teams in the industry 4.0. 11th international conference on e-learning (ICEL). Academic Conferences and Publishing Limited 127-131 Schonberger RJ (1987) World-class manufacturing: Le nuove regole per una produzione di classe mondiale. FrancoAngeli, Milano Schuh G, Potente T, Wesch-Potente C, Weber A, Prote JP (2014) Collaboration mechanisms to increase productivity in the context of Industrie 4.0. Procedia CIRP 19:51–56 Schwab K (2016) La quarta rivoluzione industriale. FrancoAngeli, Milano Shen WM, Norrie DH (1999) Agent-based systems for intelligent manufacturing: a state-of-the-art survey. Knowl Inf Syst 1(2):129–156 Taylor FW (1919) The principles of scientific management. Harper and Brothers Publishers, New York Wan J, Cai H, Zhou K (2015) Industry 4.0: enabling technologies. In Proceedings of 2015 international conference on intelligent computering and internet of things, ICIT pp 135–140 Wang S, Wan J, Zhang D, Li D, Zhang C (2016) Towards smart factory for industry 4.0: a selforganized multi-agent system with big data based feedback and coordination. Comput Netw 101:158–168 Yamashina H (2000) Challenge to world class manufacturing, in International Journal of Quality & Reliability Management 17(2):132–143 Zezulka F, Marcon P, Vesely I, Sajdl O (2016) Industry 4.0 – an introduction in the phenomenon. IFAC Paper Online 49(25):8–12 Zhong RY, Xu X, Klotz E, Newman ST (2017) Intelligent manufacturing in the context of industry 4.0: a review. Engineering 3:616–630

IIOT Applications for Sustainable Manufacturing

48

S. Kamalakkannan and A. K. Kulatunga

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Importance of Sustainable Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Role of Information and Communication Technology in Industry 4.0 . . . . . . . . . . . . Sustainable Value Creation Through Life Cycle Management . . . . . . . . . . . . . . . . . . . . . . Challenges in Life Cycle Management Toward Sustainable Development . . . . . . . . . . . . Importance of Facilitating IIOT for Life Cycle Management . . . . . . . . . . . . . . . . . . . . . . . IIOT- Based Life Cycle Management Toward Sustainable Manufacturing . . . . . . . . . . . . Implementation of IIOT System for Life Cycle Management . . . . . . . . . . . . . . . . . . . . . . . . System Architecture of IIOT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Identification of Parameters and Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Design and Development of IIOT Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IIOT-Based Real-Time Data Acquisition and Manipulation . . . . . . . . . . . . . . . . . . . . . . . . Example of an IIOT-Based Life Cycle Management in Sri Lankan Tea Industry . . . . . . . . . IIOT-Based Real-Time Data Acquisition for LCI and LCM . . . . . . . . . . . . . . . . . . . . . . . IoT- based Real-Time Data Manipulation for LCI and LCM . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1222 1222 1223 1224 1225 1226 1227 1228 1231 1233 1233 1234 1235 1236 1238 1240 1240

Abstract

Industry 4.0 technologies have significant potential for sustainable value creation in the economic, environmental, and social dimension of sustainability by improving resource and energy efficiency. The industrial value creation is causing a fundamental change in manufacturing with the increasing global resource constraints and long-term goals of sustainable development, with the

S. Kamalakkannan · A. K. Kulatunga () Department of Manufacturing & Industrial Engineering, Faculty of Engineering, University of Peradeniya, Peradeniya, Sri Lanka e-mail: [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_31

1221

1222

S. Kamalakkannan and A. K. Kulatunga

support of emerging information and communication technologies (ICT) such as Internet, cyber-physical system, Internet of Things (IoT), cloud computing, and big data. However, the role played by IoT in the industry in the sustainable value creation is remarkable. Notably, the industrial Internet of Things (IIoT) enabled availability of live data on environmental indicators such as energy and resource consumption could make it easier for all manufacturers to access such information and thus effectively improve the sustainability of industries through life cycle management (LCM) approach. Undoubtedly, the application of IIoT toward sustainable development will be very opportune for the industries to become competitive in the global market. Although handful scientific researches and projects deal with the implication of the IIoT for sustainable development, the linkages between digitalization and sustainability are attracting increasing attention nowadays. This chapter explores the concept and application of IIoTbased LCM toward sustainable value creation in the manufacturing industries. It also offers an example of an experimental IIoT-based LCM in the manufacturing industry. Keywords

Industrial internet of things · IIoT · Internet of things · IoT · Industry 4.0 · Life cycle management · LCM · Life cycle assessment · LCA

Introduction Importance of Sustainable Manufacturing Over the last few decades, environmental issues such as global warming, pollution, and depletion of resources have attracted much attention around the world. In this regard, sustainable development plays a crucial role in striking a balance between the demands of social productivity and the reserves of natural resources (Chang et al. 2014). This has been aggravated due to rapid globalization which exponentially increased the consultative lifestyles. Since the manufacturing sector is one of the basic and primary sectors which directly link with societal needs when demand increases, manufacturing sector damages environment through direct and indirect means in the way of excessive consumption and polluting the environment through waste generation (Cai et al. 2019). Escalation of climate change–related consequences and rapid environmental pollution have compelled the global manufacturing sector and other utility sectors such as power generation and logistics and transportation sectors, etc. to concern about environmental sustainability. Consequently, there is a greater tendency for the manufacturing sector to move toward “Green concepts” to have a higher level of appreciation by the society and even to secure global markets such as EU so that products manufactured through green concepts will have a better demand than their competitors (Kamalakkannan et al. 2020).

48 IIOT Applications for Sustainable Manufacturing

1223

In particular, the consequences of two major negotiations, the green economy concept and the sustainable development goals, influence global sustainable development policies for the coming decade (Beier et al. 2018). The policies supporting a green economy, shortly defined as “a low-carbon, resource efficient, and socially inclusive” economy, were adopted at the 2012 United Nations Conference on Sustainable Development. Moreover, the United Nations (UN) has stated that sustainability can be considered as the core of business strategy in the sustainability 2030 agenda and Industry 4.0 technologies help to achieve sustainability in business practice (Jamwal et al. 2021). On the other hand, sustainability and Industry 4.0 are two prominent research areas in engineering. Industry 4.0 technologies have significant potential for sustainable value creation in the economic, environmental, and social dimension of sustainability by improving resource efficiency (Jamwal et al. 2021). Sustainable designing, life cycle management (LCM), circular economy, environmental management, lean and green management, and remanufacturing are the main disciplines of sustainability and Industry 4.0. In the sustainable value creation of Industry 4.0, sustainable manufacturing will be realized by using the ubiquitous information and communication technologies (ICT) infrastructure (Li et al. 2020a). The industrial revolutions aim to not only to enhance and directly respond the needs of the industry in a productive side fact, but also aim for the sustainable development in terms of triple bottom line (Tabaa et al. 2020). In general, sustainability benefits of Industry 4.0 are expected on improving resource efficiency, productivity, and flexibility, and reduction of energy, waste, consumption, and overproduction. Notably, when considering the current situation, COVID-19 pandemic, which is currently threatening the entire world, is causing significant impacts such as loss of businesses and unemployment on the world economy. Besides, the pandemic lockdown has caused severe threats to sustainability in both positive and negative ways due to raising of human disease, a contemporary pause of business, stock market and industries, low production, climate changes, labor shortage, internal migration of workforce, etc. On the other hand, COVID-19 pandemic significantly enhances environmental sustainability primarily due to slowing down of global economies and reduction of movements between the territories, regions, etc.

The Role of Information and Communication Technology in Industry 4.0 Globalization is a multidimensional as well as a systematic technological advancement. The ICT revolution is one of the most important factors in globalization. Through ICT, the industries have started to become more intelligent and smart. The Industry 4.0 is all about including modern technologies from the ICT platform for processes of automation and real-time data exchange in manufacturing organizations. In addition, the ICT allows for product and process innovation through generating new factors of production that help with economic restructuring and

1224

S. Kamalakkannan and A. K. Kulatunga

transition. In addition, with the purpose of industrial revolution with sustainable value creations, the industrial activities from research and development to manufacturing processes has all been made possible by ICT. The industrial value creation is causing a fundamental change in manufacturing with the increasing global resource constraints and long-term goals of sustainable development, with the support of emerging ICT such as Internet, cyber-physical system, Internet of Things (IoT), cloud computing, and big data. Accordingly, the manufacturing has begun to move toward intelligence, real-time, interconnection, globalization, and personalization. With the emergence of these features, there is a transformation in manufacturing paradigm to cyber-physical manufacturing systems, industrial Internet of Things (IIoT), intelligent manufacturing, cloud manufacturing, sustainable manufacturing, global manufacturing, and mass customization (Li et al. 2020a). However, the role played by IoT in the effective utilization of resources and economic value creation is remarkable. The extension of IoT in the industry is known as IIoT and it is the integration of numerous collection sensors or control sensors and actuators, ubiquitous technologies, communication technology, artificial intelligence, big data analytics, augmented reality, virtual reality, and security mechanisms. In particular, there are currently three major applications in the adaptation of IoT for industries, namely (1) implementation of monitor and control by collecting data, (2) providing a reference for business decisionmaking by the analyzation of big data generating from the objects of IoT, and (3) accomplishment of information sharing and collaboration between people and things (Li et al. 2020a).

Sustainable Value Creation Through Life Cycle Management Environmental threats and pollution has increased due to the industrial revolution and escalation of societal needs over the centuries. A complete sustainable life cycle and continued sustainable development are achieved by following a sustainable triple bottom line and helical economy. Since helical economy, which is the extension of circular economy, promises to simultaneously create sustainable value creation and encourage continued innovation and economic growth (Bradley and Jawahir 2019). Also, the triple bottom line states that, in addition to its economic performance, in order to be truly sustainable, a company must concentrate on its environmental and social performances as well (Fauzi et al. 2010). In that respect, the sustainable development and value creations are the most concerning substances over the significance of industrial revelations. To improve the sustainable value of the manufacturing sector, it is essential to consider the entire life cycle of products and processes to identify environmental hotspots mitigated through LCM by subsequent product designs as eco-design or eco-innovation (Kamalakkannan et al. 2020). Hence, managing the life cycle of a product and process is vital for their green initiative and to enhance the environmental performance

48 IIOT Applications for Sustainable Manufacturing

1225

of the industries. In addition to enhancing the environmental performances due to multiple requirements and pressure from different groups, environmental hotspots of all the phases from cradle to grave of the product have to be identified and mitigated through eco-design or eco-innovation as well. On the other hand, around 70% of the product sustainability decisions are taken at the design stage (Fu et al. 2020). Therefore, establishing life cycle thinking, green product concepts, and environmentally sound products are necessary at this stage (Belucio et al. 2021; Brundage et al. 2018). The decisions made at the early stages set the general conditions for the following stages of the design process. Hence, the decisions made at the early stages of the design process have a robust influence on the product. Thus, the substantial potential for optimization and reduction of emissions remain in the early stages of the design process (Basic et al. 2019). Accordingly, in order to initiate continuous sustainable value creation through ecodesign and eco-innovations, this stage provides significant steering opportunities. The actions taken in eco-design and eco-innovation design process are the crucial, in order to achieve a product, process, and system with a minimal environmental impact. In this respect, the National Institute of Standards and Technology (NIST) has recommended that industry needs to include sustainability in the product design phase (Eddy et al. 2013). In addition, elements such as efficiency, speeding up productivity, quality, eco-innovation, eco-friendliness, and power to distinguish oneself from other market players have all become extremely important (Zbicinski and Stavenuiter 2006). As a result, these have simulated the designers to develop eco-friendly processes and products. Even though all manufacturers are inclined to sustainability, very few are prepared to pay for products with an outspoken environmental profile (Luttropp and Lagerstedt 2006).

Challenges in Life Cycle Management Toward Sustainable Development The manufacturing domain is integrated with a huge amount of heterogeneous information including operational, supply chain, materials, energy, resource, waste, and preliminary data from the whole life cycle of the products or services. However, the lack of information and high degree of uncertainty during the early design phases hinder the use of tools such as life cycle assessment (LCA) (Bouyarmane and Sallaou 2019; Ng and Chuah 2017). Due to the wide range of required data, evaluating environmental impact through the implementation of a complete LCA is very complex, time consuming, and expensive (Ameli et al. 2017). Hence, in order, the acquisition of a product or process-related life cycle data is of vital importance to manage the life cycle of the product, process, and systems through the LCA approach. However, with the growing complexity of the analyzed product, the calculation of an LCA requires a large effort. Due to the wide range of required data, evaluating environmental impacts through the implementation of a complete LCA is very complex, time consuming, and expensive. An LCA is not sufficient to

1226

S. Kamalakkannan and A. K. Kulatunga

make a final decision depending only on environmental assessment results. In order to maintain and improve the environmental performances of a product or process, an LCM is of paramount importance for manufacturing sectors. However, LCM is not a gradual one-way process because, in the life cycle, certain changes can be taken place over the time period of product lifespan. Therefore, continuous reevaluation of product life cycle needs to be considered from time to time when the product is being available in the market, due to various reasons such as changes in raw material (RM) or RM supply, alteration of technology, the variation of energy source, etc. Generally, the data are collected manually from the field visits. The collection of life cycle inventory (LCI) data requires the support of process engineers, supervisors, technicians, operators, production record books, forms, bills, open literature, LCA libraries, and others from the manufacturing domain. Hence, manual data collection is the most difficult, complex, and resource- and time-consuming process. Besides, there are some other practical difficulties during manual data collection such as resource and energy allocation, data accuracy, data reliability and uncertainty, data availability, acquisition hardness, and technical supports. Therefore, to overcome these issues, it is essential to incorporate IoT-based realtime data acquisition system with the LCM.

Importance of Facilitating IIOT for Life Cycle Management Integrative approaches for managing life cycle of a product usually confront designers and manufacturers with a huge amount of data, numbers, and facts. Therefore, it is required to have adequate and timely information about the entire product life cycle of the product and processes it manufacture and typically it is a time-consuming and costly operation (Bhander et al. 2003). Also, environmental sustainability information in the manufacturing industry is not easily shared between stages in the product life cycle (Brundage et al. 2018). Besides, LCA requires detailed data on product development that is not available in the early stage of conceptual design (Prastawa and Hartini 2019). This issue can be eased some extent through the IoT which is tracking data real time (Bhander et al. 2003). IoT is one of extremely high expected technologies (Goto et al. 2016). Many benefits are expected to be enabled by implementing the IoT technologies through the product life cycle management process, such as remote monitoring of field service and predictive quality reliability engineering design in research and development (Goto et al. 2016). Notably, the IIoT enabled availability of live data on environmental indicators such as energy and resource consumption could make it easier for all manufacturers to access such information and thus effectively improve the sustainability of industries through LCM approach. Although handful scientific projects and researches deal with the implication of the IIoT for sustainable development, the linkages between digitalization and sustainability are attracting increasing attention nowadays (Beier et al. 2018).

48 IIOT Applications for Sustainable Manufacturing

1227

IIOT- Based Life Cycle Management Toward Sustainable Manufacturing The manufacturing industry consumes an excessive amount of materials, mainly from virgin sources and energy streams, which contributes significantly to global environmental problems such as climate change. Therefore, managing the life cycle of a product is of paramount importance to the sustainability of the product. Hence, there is a requirement to integrate proper data acquisition mechanism to assist LCM, and in that way it is possible to increase the sustainable value of a product or a process throughout its lifespan in the market (Liu and Zhou 2012). In recent years, the rapid development of the Internet provides a global platform for machines and smart objects in terms of communication, dialogue, computation, and coordination. Correspondingly, the wide applications of IIoT have been witnessed in some fields including smart industry, smart home, smart energy, smart transport, and smart health (Li et al. 2020b). Although IoT is an emerging domain, it has already helped to enable or enhance countless applications, which has conveyed number of changes in our day to day lives (The Internet of Things (IoT)). The IoT application has been widely started to facilitate in many sectors such as transportation, smart home, smart city, lifestyle, retail, agriculture, smart factory, supply chain, emergency, health care, user interaction, culture and tourism, and environment and energy (Elsonbaty 2019). Since IoT refers to a variety of equipment and systems, tracking the information through IoT has gained a drastic momentum over the years. Equipment and systems, such as sensor networks, radio frequency identification (RFID) reading device, bar code, and two-dimensional code equipment, own many global industries influencing various industrial development (Liu and Zhou 2012). Moreover, the IoT data can be collected from multiple different sources and it consists of various structured and unstructured data where data storage components are expected to have the ability to deal with heterogeneous data resources (Kamalakkannan et al. 2020). Many product development and life cycle management processes and overall decisionmaking processes in the manufacturing sector have been influenced by the approach of IoT (Papakostas et al. 2016). IoT technologies play a very important role toward managing the product life cycle, improving the performance, and more efficient process controlling and monitoring (Papakostas et al. 2016). Meantime, the IoT will help manufacturers to gain a better understanding of the entire life cycle information that can be delivered in real time. IoT provides end-to-end transparency almost in real time, allows the optimization across factory sites in the area of production, and then improves the factory efficiency (Shrouf et al. 2014). Therefore, the application of IIoT toward sustainable development will be very opportune for the industries to become competitive in the global market. Figure 1 shows the annual scientific production of IIOT applications toward sustainable manufacturing from 2005 to the present. This analysis was based on systematic literature survey related to IIOT applications in sustainable manufacturing. The search term ([“IoT” or “IOT” or “internet of things*” or “Industrial internet of things*” or “IIOT” or “IIOT”] and [“sustainable manufacturing*” or

1228

S. Kamalakkannan and A. K. Kulatunga

400

Articles

300

200

100

0 2005

2007

2009

2011

2013

2015

2017

2019

2021

Year Fig. 1 Growth of IIOT application toward sustainable manufacturing over the past decades

“manufacturing industry*” or “manufacturing*” or “life cycle management*” or “LCM”]) was used to collect “2000” research articles including articles, books, book chapters, and conferences from Scopus database. The data from Scopus was imported to the biblioshiny web interface for bibliometrix analysis. Analysis shows that the annual growth rate is 32.29% which highlights the rapid growth and importance of IIoT in sustainable manufacturing.

Implementation of IIOT System for Life Cycle Management LCM is a single tool or methodology, which is a flexible integrated framework of concepts, techniques, and procedures. LCM inherently takes a life cycle approach in considering environmental, economic, and social aspects of products and organizations. A multiple-step methodology has been built based on LCM concepts and a special focus is given to the environmental aspect. To develop the IoT-based real-time data acquisition system, the expected techniques and tools are listed under inputs, and they include the sensors, devices, IoT platform, and the cloud computing. The key steps that need to be followed to achieve the real-time data are listed under the design and development section (which is in the second column of Fig. 2). The key steps are defined based on the requirement for a real-time data acquisition. Moreover, when considering the correlation between the inputs and the design and development phase, the sensors and devices incorporate the development of sensor networks and IoT platform. Here,

48 IIOT Applications for Sustainable Manufacturing

Input

Sensors

1229

Design & Development

Output

Life Cycle Stage

Monitoring & Controlling

(Product, Process, System)

LCI Devices

IoT Platform

Cloud Computing

IoT Based Data Acquisition System

PDD Parametric LCA

Life Cycle Inventory

Finest Eco-Design

NO

(Product, Process, System) Eco-Design

Fig. 2 IIoT-based real-time data acquisition system for LCM [PDD – product design and development; DSS – decision support system]

cloud computing is used to transfer and store the LCI data to cloud storage. To this end, the collected and stored data will be analyzed to prepare LCI summary. Finally, the output of this framework, which describes process monitoring and controlling and LCI summary development, is facilitated based on the outcome resulting from an LCI data. Eventually, the LCI will be used for parametric LCA modeling and decision-making. The basic concept of this parametric LCA approach is combining the principles of parameterization techniques with a simplified process of the LCA method. A parametric LCA will help designers and manufacturers to evaluate and compare the environmental performance of the product, process, and system with various scenarios of alternative approaches during the design process (Kamalakkannan and Kulatunga 2021). In sum, the effects of this IoT-based real-time data acquisition system indicate how the approach contributes to the ecodesign. Implementing eco-design for the product life cycle during the product design and development stage will lead to LCM through the creation of sustainable value and improvement of environmental performance. This framework includes an IIoT-based real-time data acquisition and manipulation system to facilitate the LCM (Fig. 3). The IIoT system is facilitated to gather LCI information associated with the product manufacturing. The bigdata traced from IIoT system will be stored in the cloud database such as ThingSpeak, 000WebHost, etc. The stored data will be manipulated and filtered to smart data that will be used for life cycle analysis. After that, the hotspot will be identified, which will help identify the place where the necessary action needs to be taken in terms of

Data Logger

2

Web Server

3

DSS Feeder DSS User (Plugging)

16(a)

4

6

10

Setup

Raw data 5 11(b)

IIOT Data Manipulation

User Interface

Environmental Impact

8

LCI

Redesign

LCIA

11(a)

17

Fig. 3 Schematic diagram of component network of IIoT system. (Source: (Kamalakkannan et al. 2020))

Factory

1

7

IIOT Based LCM

12

9

Redesign Reports

16(b)

LCIA Reports

LCI Reports

Product or Process Design & Development

Eco-Design

15

Hotspot Identification

Redesign Scenario & Parameters

13(b)

13(a)

Background Surveys

14

1230 S. Kamalakkannan and A. K. Kulatunga

48 IIOT Applications for Sustainable Manufacturing

1231

operations and environmental aspects to improve production performance and sustainability. Afterward, consequences from life cycle analysis of the manufacturing product, processes, and system-related improvement can be incorporated to manage the life cycle of product, process, and system. Importantly, through manipulating this method, it is easy to manage the life cycle of the product, process, and system on the environmental aspect because LCM is concerned on life cycle thinking and product sustainability operations to attain continuous improvement (Life Cycle Management).

System Architecture of IIOT When considering the different IoT systems in various applications, the fundamentals for the IoT architecture as well as its general data process flow is almost the same. IoT architecture is a system of numerous elements which include sensors, Wi-Fi module, circuits, protocols, actuators, and cloud services. In the IoT architecture system, there are three major divisions such as data source, data collector, and data gateway. The data collector part is a fundamental part in the data collection architecture. The mobility of the data collector allows gathering a huge amount of the data from the IoT environment. The data collectors which are sensors and actuators are able to sense the required information and then pass the accumulated information on to IoT gateways. However, when pursuing the big data through IoT, the data collector should be well equipped with a significant number of communication interfaces, large storage area, and long lifetime. The function of gateway that collects unprocessed data converts it into digital streams and stores it in the cloud server which can be used for further analysis. The system architecture of IoT is constructed and displayed in two levels such as unit process level and entire life cycle level.

Unit Process–Based System Architecture of IIOT The IoT-based data collection process could be facilitated in the unit process level which is the smallest element in the product life cycle. In the unit process level, the LCI data which is used for impact assessment associated with the unit process such as inputs and outputs tends to be collected through sensors and devices. The schematic diagram of the unit process–based system architecture is given in Fig. 4. Figure 4 shows how the unit process level LCI data are transmitted to a cloud environment. The appropriate sensors and devices which are installed with the exchange points of the unit process will gather entire data and transmit them to the data logger. Another indispensable element of this system is the data logger or actuators. The data logger consists of the Wi-Fi module which is programmed to gather data and circuits which connect Wi-Fi module with the sensors. The system obtains an input from the sensors, and then the system analyzes the situation in real time. Subsequently, the system commands the actuators to trace the data. Data collection, filtering, and transfer to edge infrastructure and cloudbased platforms could be identified as the functions of gateway (The Internet of

1232

S. Kamalakkannan and A. K. Kulatunga

Fig. 4 Unit process–based system architecture of IIoT

Things (IoT)). As intermediaries between the connected things and the cloudbased platforms, gateways provide the necessary connection point that ties the remaining stages together. The cloud-based system is facilitated to store, process, and analyze massive volumes of data which is called big data. The cloud computing contributes to higher production rates, reduction of energy consumption, and many other business benefits. Finally, the cloud-based big data provides the necessary smart data that contains all exchange data associated with the unit process to conduct the comprehensive analysis.

Product Life Cycle–Based System Architecture of IIOT The comprehensive life cycle analysis of a product or process or system requires the LCI data from cradle to grave that includes premanufacturing, manufacturing, use, and post-use. Therefore, the collection of entire life cycle exchange data is essential. In that respect, to collect data from cradle to grave using IoT techniques, the life cycle–based system architecture is constructed. The schematic diagram of product life cycle–based system architecture is given in Fig. 5. Figure 5 shows how life cycle level LCI data are transmitted to a cloud environment. The appropriate sensors and devices which are installed with the exchange points from each life cycle phases will gather entire data and transmit it to the data logger. The process from data logger to cloud-based platform is similar to the unit process–based IIoT system as mentioned in previous section. The big data from the cloud platform will be converted as smart data to generate essential LCI reports to conduct the comprehensive environmental assessments. In addition to that, the cloud platform will allow users to monitor and control the processes or system and make informed decisions on the basis of reports and data viewed in real time.

48 IIOT Applications for Sustainable Manufacturing

1233

Fig. 5 Product life cycle–based system architecture of IIoT

Identification of Parameters and Sensors The purpose of IoT-based data acquisition system is to collect the LCI data which is used for LCA in real time. Therefore, the parameters will be the inputs and outputs that make an impact on the environment. The LCI parameter could be product quality based or product design based or process based. The most common and essential LCI parameters are raw materials, electricity, LP gas, water, petroleum fuel, quality factors, weather related parameters, etc. The selection of LCI parameter should accomplish the requirement of conducting LCA of a product or process or system. Besides, the selection of sensors will be based on the LCI parameters and the sensor should be precise and reliable. The advanced and intricate devices that are used frequently to detect and respond to electrical and optical signals are known as sensors. Sensors convert the physical parameters such as resource consumptions, temperature, humidity, speed, etc. into a signal which can be measured electrically. In the current world, many types of sensors are used for numerous applications in several areas (Rao et al. 2012). Sensors and sensor networks are being used by various applications such as smart home, smart city, and smart industries.

Design and Development of IIOT Platform Design and Development of the Device The sensor network devices are integrated by many distributed and interacting components that are usually heterogeneous in terms of hardware devices, communication protocols, software interfaces, and data. Initially, the general and specific requirements need to be defined in order to develop the sensor networks. The sensors, devices, and software components need to communicate with each other.

1234

S. Kamalakkannan and A. K. Kulatunga

Since to define the way of data sensing, connect sensors and Wi-Fi module, connect IP and transfer data to cloud, and amplitude the signals, it is essential to develop the middleware device which consists Wi-Fi module, sensors, and circuits. The Wi-Fi module is a transmission layer of IoT that can connect traditional serial device and controlled device with Wi-Fi network to realize control and management. According to the data requirements, the sensor networks will be designed and developed.

Device and Sensor Calibrations Proper calibration is a key factor in maintaining accuracy in real-time data analytics. It is safe to say that the sensor calibration is inevitable for an IoT system because the data analytics process could be catastrophically influenced by providing inaccurate data that renders itself nonactionable. The poor-quality sensors and the sensors without calibration which provide inaccurate data will convert decision-makers to make wrong decisions, so certain amount of calibration is necessary to ensure accuracy. In addition, when using the sensors for monitoring and control the process in dangerous situation is very dangerous. Ordinarily, structural errors are shown in the sensors that creates differences between the expected output and the measured output (Lin et al. 2019). However, calibration removes structure errors of the sensor outputs as well as improves the accuracy of the sensor. Sensor’s response to an input is defined by the characteristic curve in calibration, while the process of calibration succeeds to map the sensor’s response to an ideal linear response. When considering the ideal scenario, a straight line would be the characteristic curve (So, How Do We Calibrate?). The greatest deviation of the characteristic curve from a reference line is described as nonlinearity. However, the calibration is achieved by the adjustment of the characteristic curve. Besides, the calibration can be examined with the similar sensor which is likely ideal and accurate. Thereafter, the calibrated data is analyzed through an analytical software like Minitab® .

IIOT-Based Real-Time Data Acquisition and Manipulation The IoT-based data acquisition system is installed in the factory process premises which are required to collect the data for conducting LCA after the sensors and device calibrations. The collected data transmit to the cloud database in real time through Wi-Fi modules. In that respect, there are many cloud databases for individual users or organizations to store their data on the cloud platform. Effortless collection, access, process, visualization, archive, share, and search of huge bundles of data from different processes and places is enabled by these cloud platforms (Casola et al. 2013). Besides, it will support users as a decision support system to monitor and control the life cycle process in real time. In fact, this IoT-based data acquisition system is beneficial in two aspects such as process controlling and environmental impact assessment. The process monitoring and controlling of a product life cycle is mandatory unless it will affect not only the product quality but also the consumption volume of energy, resource, and raw material. Since the real-time process monitoring helps to maintain the quality of the product as well

48 IIOT Applications for Sustainable Manufacturing

1235

as helps to optimize resource and energy consumption. The data storage framework should have the ability to deal with various types of data, which are collected from various devices and in that way it is feasible to fulfill the requirements to manage massive IoT data in cloud platform (Elsonbaty 2019; Dodda et al. 2016). These data are different in data structures, volume, units, accessing methods, and in some other aspects. However, cloud platform like ThingSpeak provides certain access to analyze, filter, and visualize the collected data. Finally, the required data for the impact assessment which is called smart data receives from the collected data stored in the cloud database.

Example of an IIOT-Based Life Cycle Management in Sri Lankan Tea Industry Integrating the IoT-based real-time data acquisition system into LCA toward LCM can resolve the problems and challenges that occur during the LCA and LCM practices. The presented example of an IIoT-based LCM in Sri Lankan tea industry i ntended to investigate the performance, effectiveness, and feasibility of the IIoTbased data acquisition system for LCM. Sri Lankan tea industry, acclaimed as the best tea in the world, has its inherent unique characteristics and reputation running through more than 150 years. Globally, Sri Lanka is the second largest tea producing country, and the production share is around 10% in the international sphere. Sri Lanka is the world’s largest producer of orthodox tea and is one of the world’s leading exporters with a share of around 23% of the global demand (Sri Lanka Export Development Board (EDB)). Sri Lankan tea has a marvelous reputation with the branding of Ceylon Tea. However, the recent incidents such as Khapra beetle scare in Russia (Daily Mirror) and contamination issues in Japan (News First) have startled the tea industry in Sri Lanka. The tea industry directly contributes to the national GDP and it gives more than 900,000 employment opportunities (4.5% of Sri Lanka’s population) (Global Press Journal). The main environmental issues identified in the tea production are excessive energy consumption in the form of electricity to run the machinery and thermal energy for withering and drying. Munasinghe uncovered many issues including energy efficiency of the industry, GHG emissions, and occupational health hazards in Sri Lankan tea industry (Munasinghe et al. 2017). Countries like India, Sri Lanka, and Vietnam use abundant amount of fertilizers and pesticides for tea cultivation, which give a negative effect to the local and wider environment by increasing water pollution and reducing soil biodiversity (Wal 2008). Therefore, the sustainability of tea industry is crucial to the Sri Lankan economy. Hence, IIoT application experimented to realize the potential and effectiveness in the Sri Lankan tea industry. Real-time data acquisition system is facilitated in the manufacturing process using the IoT sensors and devices. Out of the many processes of producing black tea, withering and drying are the two main operations, and these operations consume high electrical and thermal energy. The final quality of the black tea often critically depends on the withering and fermentation operations.

1236

S. Kamalakkannan and A. K. Kulatunga

Hence, LCI data of these processes are crucial to conduct a proper LCA of black tea processing. Therefore, this experiment was facilitated to explore the variations of relative temperature, humidity, moisture content, and electricity consumption. Conclusively, this presented example of an IIoT-based LCM which clearly shows how IoT-based data acquisition can be facilitated to obtain LCI data and how it facilitates LCM.

IIOT-Based Real-Time Data Acquisition for LCI and LCM Development of IIOT-Based Real-Time Data Acquisition System With the aim of acquiring real-time data, monitoring, and controlling black tea processing, the essential environmental parameters were identified. Out of the many processes of producing black tea, withering and drying are the two key operations, and these operations consume high electrical and thermal energy. The final quality of the black tea often critically depends on the withering and fermentation operations. Hence, LCI data of these processes are crucial to conduct a proper LCA of black tea processing. Therefore, this study was focused to explore the variations of relative temperature, humidity, moisture content, and electricity consumption through IoT. Thereafter, the sensors were precisely calibrated and mathematical equations were expressed using analytics software called “Minitab.” The used and developed sensors and devices are shown in Table 1 and Fig. 6. Then, the IoT devices were installed on the factory process premises, and the monitored and traced processrelated data were stored in the cloud database and manipulated by ThingSpeak software. Fig. 7 shows the factory environment in which IoT devices are installed. Figure 8 shows the real-time visualization of ThingSpeak interface. Figure 9 shows the observed real-time electricity consumption for (a) withering and (b) rolling processes.

Table 1 Sensors and specifications used for the IoT system Parameter Temperature and humidity

Sensor type DHT22

Moisture content

Arduino Moisture module

Temperature

Sensor probe

Current

AC current sensor

Specification Power supply 3.3–5.5 V DC Humidity.0–100%RH Temperature − 40–80 ◦ C Accuracy + − 2%RH; + − 0.5 ◦ C Power supply 3.3–5 V DC Output signal 0 ∼ 4.2 V Current 35 mA Power supply 3.0–5.5 V DC Temperature range − 55 to125 ◦ C Accuracy over range − 10 to 85 ◦ C is + − 0.5 ◦ C Rated input 0–30A Rated output 0–1 V Accuracy 1% and turns ratio 1:1800

48 IIOT Applications for Sustainable Manufacturing

Fig. 6 Fabricated devices for data collection

Fig. 7 IoT-based data collection from various processes

1237

1238

S. Kamalakkannan and A. K. Kulatunga

Fig. 8 Real-time data visualization of ThingSpeak interface b 5 Current(A)

4 3 2 1 0

3:02:57 PM 3:40:49 PM 4:19:51 PM 4:57:47 PM 5:35:38 PM 6:17:18 PM 7:08:24 PM 8:01:24 PM 8:54:24 PM 9:43:37 PM 10:34:43 PM 11:32:40 PM 12:42:43 AM 1:54:42 AM 4:25:44 AM 6:40:16 AM 7:47:25 AM

Phase current(A)

6

Time

20 18 16 14 12 10 8 6 4 2 0

1:15:12 AM 1:26:33 AM 1:36:01 AM 1:47:25 AM 1:58:47 AM 2:43:49 AM 9:38:08 AM 9:54:40 AM 10:06:02 AM 10:15:30 AM 10:23:56 AM 10:33:54 AM 10:45:52 AM 10:56:35 AM 11:13:19 AM

a

Time

Fig. 9 Monitored LCI electricity consumption data for (a) withering process and (b) rolling process

IoT- based Real-Time Data Manipulation for LCI and LCM As an outcome of this real-time process data monitoring, the unit process based LCI data was saved in the cloud environment. However, the monitored data cannot be used to assess impact directly because the measurement units, time frame, and LCI allocations are different for every data acquisition system and devices. Hence, the entire possible and available data were tracked and saved through ThingSpeak software. The interval of data tracking can be defined based on LCA requirements and device capability. Primarily, the quality of black tea depends on temperature, moisture content, and humidity. Therefore, maintaining these key factors of black tea processing is mandatory unless it will affect not only the quality of tea but also the energy, resource, and material wastages. Hence, real-time data monitoring helps to maintain the quality of the product as well as helps to optimize resource and energy consumption.

48 IIOT Applications for Sustainable Manufacturing

1239

Table 2 Summarized IIoT-based real-time monitored data from black tea processing (Source: (Jamwal et al. 2021)) Process Withering Rolling Rotorvane

Phase current /A (Mean) 05.053 11.828 16.197

Standard deviation 0.058 1.469 4.545

Working hours/h 14 8 8

No of machine 6 3 2

Electrical energy consumption/kWh 448.218 299.767 273.715

In addition, this data monitoring was also used for LCA. The stored LCI data which is related to energy and resource consumption was extracted through ThingSpeak and categorized based on a unit process as shown in Table 2. Thereafter, the data were analyzed and converted into a functional unit. At the same time, by using the entire data population, the mean and standard deviation of the input and output were calculated. Power consumption

W = 3Vp Ip Cosθ

(1)

W: Watts Vp : Line voltage Ip : Phase current Cosθ : Power factor A summary of the energy between the most important processes in the tea industry is shown in Table 2. It is evident from the table that the electricity consumption of the withering process is high due to high processing time and large capacity motor usages, where the rolling and rotorvane processes are approximately consuming the same level of energy. Further, the monitored temperature of fermentation result was used to control the exact fermenting process time unless it will reduce the tea quality due to an excessive chemical reaction. Moreover, the monitored temperature and moisture of the withering process was used as a control parameter to control the air flow rate and the pressure by changing the RPM of the motor. The results of an example study reveal that unless sustainability-related key performance indicators are not traced frequently it is difficult to evaluate, realize, and implement the environmental performances and eco-design opportunities through LCM approach. Since IoT techniques are economical, easily adaptable, and could cover wide spectrum of activities, it is easy to step into the operational level information tracing, process monitoring and controlling, and sustainable value creation. Hence, from this experimental study, it can be seen that an IIoT application toward LCM and sustainable manufacturing is paramount and effective.

1240

S. Kamalakkannan and A. K. Kulatunga

Conclusion Currently world is moving toward the significant environmental problems such as climate change and natural resource depletion. Manufacturing sector plays a pivotal role in these global issues. Thereby, all manufacturing industries have been pressured to contribute toward sustainable development by transforming their supply chains toward sustainability. As the most essential step, they need to adapt sustainable manufacturing practices. However, this needs significant amount of data gathering to visualize the current levels and to evaluate level of sustainability very frequently. Unless sustainability-related key performance indicators are not tracked frequently it is difficult to align their processes and supply chains toward sustainable manufacturing. Therefore, it is paramount important to trace supply chain and process data real time. This issue could be handled significantly with the development of IIoT techniques. Since IIoT techniques are cheaper and easily adaptable and could cover wide spectrum of activities it is easy to step into the operational level information tracing and develop sustainability-related KPIs. This chapter presented simple approach of adapting IIoT technologies to track the process sustainability through product life cycle perspective for one of the oldest manufacturing industries in Sri Lanka. Therefore, IIoT imposes sustainable value creation opportunities and techniques in manufacturing industries, which direct to enhance sustainability. Further, it is caused to guide industry professionals through hotspot without worsening time and money. Hence, the combined technique of LCA, which is linked to IIoT-based LCM, will be used as a stranded ladder for the green future.

References Ameli M, Mansour S, Ahmadi-Javid A (2017) A sustainable method for optimizing product design with trade-off between life cycle cost and environmental impact. Environ Dev Sustain 19(6):2443–2456 Basic S, Hollberg A, Galimshina A, Habert G (2019) A design integrated parametric tool for real-time Life Cycle Assessment–Bombyx project. In IOP conference series: earth and environmental science (vol 323, No. 1, p. 012112). IOP Publishing Beier G, Niehoff S, Xue B (2018) More sustainability in industry through industrial internet of things? Appl Sci 8(2):219 Belucio M, Rodrigues C, Antunes CH, Freire F, Dias LC (2021) Eco-efficiency in early design decisions: a multimethodology approach. J Clean Prod 283:124630 Bhander GS, Hauschild M, McAloone T (2003) Implementing life cycle assessment in product development. Environ Prog 22(4):255–267 Bouyarmane H, Sallaou M (2019) Environmental assessment in the early stages of product design. In: 5th international conference on optimization and applications (ICOA) (pp 1–5). IEEE Bradley R, Jawahir IS (2019) Designing and redesigning products, processes, and Systems for a Helical Economy. Procedia Manuf 33:168–175 Brundage MP, Bernstein WZ, Hoffenson S, Chang Q, Nishi H, Kliks T, Morris KC (2018) Analyzing environmental sustainability methods for use earlier in the product lifecycle. J Clean Prod 187:877–892

48 IIOT Applications for Sustainable Manufacturing

1241

Cai W, Lai KH, Liu C, Wei F, Ma M, Jia S, Jiang Z, Lv L (2019) Promoting sustainability of manufacturing industry through the lean energy-saving and emission-reduction strategy. Sci Total Environ 665:23–32 Casola V, De Benedictis A, Rak M, Aversano G, Villano U (2013) An SLA-based approach to manage sensor networks as-a-service. In: IEEE 5th international conference on cloud computing technology and science (vol 1, pp 191–197). IEEE Chang D, Lee CKM, Chen CH (2014) Review of life cycle assessment towards sustainable product development. J Clean Prod 83:48–60 Daily Mirror [online] http://www.dailymirror.lk/article/SL-plays-down-Russian-tea-bansaysinsect-detection-an-isolated-incident-142319.html/ (Accessed 15 Dec 2017) Dodda R, Gaddam V, Prasad JR, Rao BS (2016) The evolution of internet of things (IoT) and its impact on existing technology. IJSTE – Int J Sci Technol Eng 2(7):96–103 Eddy DC, Krishnamurty S, Grosse IR, Wileden JC, Lewis KE (2013) A normative decision analysis method for the sustainability-based design of products. J Eng Des 24(5):342–362 Elsonbaty AA (2019) Smart college using the internet of things. GSJ: Global Sci J 7(9):347–356 Fauzi H, Svensson G, Rahman AA (2010) “Triple bottom line” as “Sustainable corporate performance”: a proposition for the future. Sustainability 2(5):1345–1360 Fu Y, Wang L, Li L (2020) Conceptual design scheme automatic generation and decision-making considering green demand. Procedia Manuf 43:407–414 Global Press Journal [online] https://globalpressjournal.com/asia/sri_lanka/fighting-povertysrilankas-tea-estate-workers-demand-pay-increase/ (Accessed 11 Jan 2017) Goto S, Trolio E, Yoshie O Tamaki KY (2016) Multi-party interactive visioneering workshop for smart connected products in global manufacturing industry considering PLM. In: IFIP International Conference on Product Lifecycle Management (pp 501–511). Springer, Cham Jamwal A, Agrawal R, Sharma M, Giallanza A (2021) Industry 4.0 Technologies for Manufacturing Sustainability: a systematic review and future research directions. Appl Sci 11(12):5725 Kamalakkannan S, Kulatunga AK (2021) Optimization of eco-design decisions using parametric life cycle assessment. J Sustain Prod Consum 27:1297–1316 Kamalakkannan S, Kulatunga AK, Bandara LADAD (2020) The conceptual framework of IoT based decision support system for life cycle management. Procedia Manuf 43:423–430 Li K, Zhou T, Liu BH (2020a) Internet-based intelligent and sustainable manufacturing: developments and challenges. Int J Adv Manuf Technol 108:1767–1791 Li L, Zheng Y, Yang M, Leng J, Cheng Z, Xie Y, Jiang P, Ma Y (2020b) A survey of feature modeling methods: historical evolution and new development. Robot Comput Integr Manuf 61:101851 Life Cycle Management, [online], (Accessed 21 Dec 2019) Lin YB, Lin YW, Lin JY, Hung HN (2019) SensorTalk: an iot device failure detection and calibration mechanism for smart farming. Sensors 19(21):1–19. https://doi.org/10.3390/s19214788 Liu Y, Zhou G (2012) Key technologies and applications of internet of things. In: Fifth international conference on intelligent computation technology and automation (pp 197–200). IEEE Luttropp C, Lagerstedt J (2006) Eco-design and the ten Golden rules: generic advice for merging environmental aspects into product development. J Clean Prod 14(15–16):1396–1408 Munasinghe M, Yvani D, Nisitha D, Harshani K (2017) Economic, social and environmental impacts and overall sustainability of the tea sector in Sri Lanka. Sustain Prod Consum 12(1):155–169. Elsevier BV News First [online] https://www.newsfirst.lk/2018/01/trouble-brewing-tea-export-sector/ (Accessed 27 Jan 2018) Ng CY, Chuah KB (2017) Environmental impact evaluations on product design alternatives using the combined evidential reasoning with fuzzy set. JEAPM 19(03):1750014 Papakostas N, O’Connor J, Byrne G (2016) Internet of things technologies in manufacturing: application areas, challenges and outlook. In: International conference on information society (i-Society) (pp 126–131). IEEE

1242

S. Kamalakkannan and A. K. Kulatunga

Prastawa H Hartini S (2019) The influence of product design on environmental impacts using life cycle assessment. In: AIP conference proceedings (vol. 2114, No. 1, p. 030015). AIP Publishing LLC Rao BP, Saluia P, Sharma N, Mittal A, Sharma SV (2012) Cloud computing for internet of things & sensing based applications. Sixth International Conference on Sensing Technology (ICST) (pp 374–380). IEEE Shrouf F, Ordieres J, Miragliotta G (2014) Smart factories in Industry 4.0: a review of the concept and of energy management approached in production based on the Internet of Things paradigm. In: IEEE international conference on industrial engineering and engineering management, (pp 697–701). IEEE So, How Do We Calibrate? [online], https://learn.adafruit.com/calibrating-sensors/so-how-do-wecalibrate, (Accessed 28 Dec 2020) Sri Lanka Export Development Board (EDB) (2016) Industry capability report tea [online] http:// www.srilankabusiness.com/pdf/industry_capability_reports/tea.pdf Tabaa M, Monteiro F, Bensag H, Dandache A (2020) Green industrial internet of things from a smart industry perspectives. Energy Rep 6:430–446 The Internet of Things (IoT), [online] (Accessed 21 Dec 2019) Wal, S.V.D (2008) Sustainability issues in the tea sector: a comparative analysis of six leading producing countries, The Centre for Research on Multinational Corporations (SOMO) Amsterdam, ISBN: 978-90-71284-23-6 Zbicinski I, Stavenuiter J (2006) Product design and life cycle assessment. The Baltic University Press, Uppsala

Nano-biomaterials as a Potential Tool for Futuristic Applications∗

49

Anuron Deka, Pritam Bardhan, Manabendra Mandal, and Rupam Kataki

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nanotechnology: The Beginning of a New Era . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nano-biomaterials: Making Our Lives Better . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of Nano-biomaterials (Fig. 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Polymeric Nano-biomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Metallic Nano-biomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ceramic Nano-biomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carbon-Based Nano-biomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Silica-Based Nano-biomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Semiconductor-Based Nano-biomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications of Nano-biomaterials (Fig. 5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nano-biomaterials in Therapeutics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nano-biomaterials in Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nano-biomaterials in Environmental Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nano-biomaterials in Bioenergy Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fabrication Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cytotoxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biocompatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion and Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Websites/Links Related to the Topic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

∗ Anuron

1244 1244 1244 1246 1246 1247 1249 1250 1251 1253 1253 1253 1256 1258 1258 1260 1262 1264 1265 1265 1266 1275

Deka and Pritam Bardhan contributed equally with all other contributors.

A. Deka · R. Kataki () Department of Energy, Tezpur University, Tezpur, Assam, India e-mail: [email protected] P. Bardhan · M. Mandal Department of Molecular Biology and Biotechnology, Tezpur University, Tezpur, Assam, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_32

1243

1244

A. Deka et al.

Abstract

Nanotechnology has been able to carve out a special place for itself in the scientific community. Nano-materials such as graphene have enabled mankind to build supercapacitors that can store as much energy as an average-sized Lithiumion battery. Nano-biomaterials are the nano-materials that can be introduced inside a human body as a part of a drug, a medical device, or an organ and have to be biocompatible, nontoxic, and noncarcinogenic. They are widely used in drug delivery, cell tracking, bio-marking, cancer treatment, tissue engineering, gene therapy, and manufacturing of artificial limbs. Besides, nanomaterials also have profound application in food technology. Recently, they have also made their mark in the field of bioenergy and controlling environmental pollution. In this chapter, we will discuss different types of nano-biomaterials, their applications, and challenges for the future. Keywords

Nano-biomaterials · Nanomaterials · Therapeutics · Bioenergy · Food applications

Introduction Nanotechnology: The Beginning of a New Era The day was December 29, 1959, and the event was the annual meeting of the American Physical Society where Richard Feynman delivered his famous speech titled “There’s Plenty of Room at the Bottom.” This was perhaps the first time when the scientific community put their heads together around the concept of nanotechnology (Feynman 1959). The term “nano” originates from the Greek word “nano” and means “dwarf” (Leon et al. 2020). Nanotechnology is a relatively new field and the term was first used by Norio Taniguchi in his 1974 paper on “Production technology that creates objects and features on the order of a nanometer” (www.trynano.org). Cambridge dictionary defines nanotechnology as “an area of science that deals with developing and producing extremely small tools and machines by controlling the arrangement of separate atoms” (dictionary.cambridge.org). Nanotechnology encapsulates a very vast area of science under its wings. It is the research and development of technology that allows us to tinker with properties of elements and compounds by altering its structure in the atomic level (Corbett et al. 2000).

Nano-biomaterials: Making Our Lives Better Biomaterials are the materials that are biocompatible and can be introduced into an animal body without compromising the natural functioning of the body or the

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1245

cells and tissues that surround it. They can be both naturally occurring as well as synthetically engineered to meet certain requirements (Park and Lakes 2007). Metals, ceramics, polymers, and other carbon- and silica-based materials are the most common biomaterials. Biomaterials have played a significant role in making our lives easier and more comfortable and people have used them since ancient times. In ancient Egypt, doctors used a special type of linen to stitch wounds, and elephant tusks, walrus teeth, and some special kind of wood as a replacement for bone (Williams and Cunningham 1979). Even in ancient India, there are written records of Sushruta using waxes, glues, and regenerative tissues to repair injured noses (Bhat 2002). Nano-biomaterials are essentially biomaterials with surface size not greater than 10 nm. These materials have unique properties which vary from regular biomaterials, such as enhanced mechanical properties and optical properties, and tunable electrical and magnetic properties (Hasirci et al. 2006). They are also widely used in novel therapeutic and diagnostic biomedical techniques, dentistry, gene therapy as well as fabrication of biosensors for pathogen detection including COVID-19, packaging, and food delivery systems (Abd Elkodous et al. 2019; Jandt and Watts 2020; Sampathkumar et al. 2020; Srivastava et al. 2020b; Reddy et al. 2008). Nano-biomaterials have given a huge boost to the global nano-medicine market which is estimated to reach $350.8 billion by 2025 (www.grandviewresearch.com) (Fig. 1). Similarly, nanomaterials find wide applications in food technology particularly in development of smart food packaging materials (Kaur et al. 2020). Besides, semiconductor-based nanomaterials have garnered recent research attention for photocatalytic environmental and energy applications (Tahir et al. 2019;

Fig. 1 Global nanomedicine market (by application, 2020)

1246

A. Deka et al.

Sarkar et al. 2020). In this chapter, we will discuss different types of nanobiomaterials, their applications, and challenges for the future.

Types of Nano-biomaterials (Fig. 2) Polymeric Nano-biomaterials Polymers are long-chained molecules formed by repetitions of a single unit called monomers. These repeating units are held together by covalent bonds (Percec et al. 2006). They are one of the most versatile classes of materials and have played a huge role in the recent developments in the scientific world. Based on the source of the monomer, a polymer can be either naturally occurring like natural rubber or synthetic like polyethylene (Lendlein 2010). However, there are various other ways a polymer can be classified based on the type and properties of the monomer. Polymeric nano-biomaterials are one of the most versatile classes of nanobiomaterials. They can have biological or chemical origins (Lendlein et al. 2010). However, all of them are biocompatible and some of them are even bioactive. Bioactive materials are those materials that are nontoxic and are able to form a chemical bond with the host tissue by inciting a biological reaction (Kohane and Langer 2008). They are mostly used for therapy or diagnosis; however, we also see their use in prosthetics.

Types and Properties A polymeric nano-biomaterial consists of biocompatible nanoparticles as fillers. The matrix itself may or may not be biocompatible (Smith et al. 2009). They may be a derivative of a synthetic polymer which has been altered to make it

Fig. 2 Types of nano-biomaterials

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1247

biocompatible or may be a naturally occurring biopolymer which is biocompatible by default. Biopolymers are polymers that are derived from living organisms. They are usually polysaccharides, polypeptides, polynucleotides, or some other long chains of biomolecules (Leena et al. 2016). Collagen is one of the best examples of biopolymers. Skin and cartilages are the major tissues that contain collagen in our body. It also helps in maintaining skeletal integrity (Rhee et al. 1992). On the other hand, synthetic biocompatible polymers are engineered specifically to mimic their naturally occurring counterparts. Poly (lactic acid) (PLLA), poly (glycolic acid) (PGA), and poly (hydroxyl butyrate) (PHB) are some of the most widely used synthetic biocompatible polymers (Liao et al. 2011). Biopolymers are known to be more nurturing toward the growth of attached cells and tissues. However, most of the biopolymers lack good mechanical properties. Synthetic biocompatible polymers have excellent mechanical properties and their biocompatibility can also be engineered to meet specific needs. Nanomaterials that are incorporated in the polymeric matrix can be in the form of particles, platelets, fibers, whiskers, sheets, or tubes. Incorporation of nanomaterials significantly alters the properties of the polymeric matrix (Koo 2006). Polymeric nano-biomaterials are widely used in bone tissue engineering, nerve tissue engineering, vascular tissue engineering, and dental implants. They are often used as hydrogels and scaffolds that support an injured tissue. However, they can also be implanted as artificial organs or vectors for drug delivery (Rana et al. 2014).

Metallic Nano-biomaterials Metallic biomaterials such as stainless steel, cobalt-chrome alloys, and titanium alloys have been in use for a long time now. They have excellent mechanical properties. However, many of them tend to corrode which not only jeopardize its integrity but can also be toxic for surrounding cells and tissues (Niinomi et al. 2015). They can also be subjected to failure related to fatigue because of their size (Parida et al. 2012). Metallic nano-biomaterials negate most of the shortcoming of metallic biomaterials. Like all other nanomaterials, they can be synthesized as nanoparticles (0D), nanotubes, nanowires or nanorods (1D), nanosheets or nanoplatelets (2D), and nanoshells and other nanoporous structures (3D) (Edmundson et al. 2014). They can be easily synthesized and functionalized to suit various needs. Because of their extremely small size, the surface area of the metallic nanoparticles also increases which makes them extremely efficient drug delivery vectors (Vicky et al., 2010).

Types and Properties Most widely used metallic nano-biomaterials are gold (Au), silver (Ag), copper (Cu), iron oxide (Fe2 O3 ), zinc oxide (ZnO), titanium dioxide (TiO2 ), platinum (Pt), Selenium (Se), Gadolinium (Gd), and Palladium (Pd) (Fig. 3).

1248

A. Deka et al.

Fig. 3 Types of metallic nano-biomaterials

Gold Nanoparticles They are probably the most commonly used metallic nano-biomaterials. They have enhanced and tunable optical properties such as surface plasmon absorption and near-infrared (NIR) fluorescence and good biocompatibility (Das et al. 2011). They are widely used for photo targeted drug delivery, anticancer therapy, cell tracking, as contrasting agent in medical imaging, antiviral treatments, and photothermal therapy (Boisselier and Astruc 2009). Silver Nanoparticles Silver nanoparticles are synthesized in different forms; however, spherical silver nanoparticles are the most commonly used ones (Sarkar et al. 2007). Owing to its huge surface area, AgNPs are widely used as a vector for drug delivery and allows for a huge scope of functionalization with various ligands. AgNPs also display excellent optical, electrical, and thermal properties because of which they are widely used from photovoltaic equipment to biological and chemical sensors (Yuan-Tao and Huai-Zhi 2003). They also have antibacterial properties because of which they are widely used in medicines and dressing of wounds. However, the toxicity of AgNPs is not yet fully known and is subjected to extensive research worldwide (Dananjaya et al. 2016). Iron Oxide Nanoparticles Iron oxide nanoparticles (IONP) can be broadly divided into two categories: (a) superparamagnetic iron oxide nanoparticles (SPION), which are larger than 50 nm, and (b) ultra-small superparamagnetic iron oxide nanoparticles (USPION), which are smaller than 50 nm in size (Wu et al. 2008). Owing to its magnetism, IONPs hold a very special place among nanoparticles. SPIONs are biocompatible and widely used as MRI contrast agents (Gupta and Gupta 2005). They can also be used as efficient drug delivery vehicles with surface modifications. USPIONs are used as

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1249

blood pool agents for MR angiography as well as targeted drug delivery (Teja and Koh 2009).

Ceramic Nano-biomaterials Ceramic is the class of materials that are derived from nonmetallic inorganic elements. They are both amorphous and crystalline in nature and include inorganic oxides, non-oxides, and composites (Richardson et al. 2000). Ceramics are known to have outstanding hardness, excellent thermal properties, great chemical resistance, and biocompatibility. However, they are extremely brittle and have extremely poor thermal shock resistance (Naplocha et al. 2000). The properties of ceramic materials are dependent on their structure and the type of bond between the atoms. Ceramics generally have either ionic bond or covalent bond. The ionic bond is seen between two elements with different electronegativity, generally between a metal and a nonmetal. This is an extremely strong bond and transfer of electrons takes place between the elements. On the other hand, covalent bonds occur between elements with similar electronegativity and electrons are shared between the elements (Dalgleish et al. 1988).

Types and Properties Calcium Phosphates Calcium phosphates are the most widely used ceramic nano-biomaterials. Various tissues of the human body such as bone and dental tissues contain a high concentration of calcium. Thus, ceramic nanomaterials are widely used in those applications (Kalita et al. 2007). Calcium phosphate nanomaterials exhibit excellent biocompatibility, which is because of the similarity of composition with human bones. Different forms of calcium phosphate nanomaterials that are widely used in the biomedical industry are hydroxyapatite, ß-tricalcium phosphate, α-tricalcium phosphate, biphasic calcium phosphate, monocalcium phosphate monohydrate, and unsintered apatite (Kalita and Bhatt 2007). Calcium phosphate bio-nanomaterials, apart from being biocompatible, are also found to be osteoconductive (Chae et al. 2003). This means that these materials promote the growth of bone tissues over it when used as implants or scaffolds. Calcium phosphate nanomaterials are produced by wet chemical synthesis from calcium nitrate, ammonium hydrogen phosphate, sodium carbonate, calcium acetate, and phosphoric acid (Ma and Zhu 2010). Aluminum Oxide Aluminum oxide (Al2 O3 , alumina) is another prominent ceramic nano-biomaterial. It is modified as α-Al2 O3, ß-Al2 O3 , and γ-Al2 O3 or as alumoxanes (Yang et al. 2009). All the different forms or Al2 O3 are bioactive, have high thermal resistance, and are chemically inert. They are thus used in dentistry, arthroplasty, and in the treatment of fractures (Deville et al. 2003). Alumoxanes are used as nanofillers in the production of nanocomposites, mostly polymer nanocomposites and known

1250

A. Deka et al.

to improve biocompatibility and hydrophobicity of the nanocomposite (Landry et al. 1995). Alumina nanopowder can be synthesized by plasma spraying of liquid precursors or flame aerosol technology (Renuka et al. 2012; Park et al. 2005). Zirconium Dioxide Zirconium dioxide, zirconium (IV), oxide or zirconia (ZrO2 ) is the most commonly found oxide of zirconium, which is a transitional metal (Trunec and Maca 2007). Zirconia can be synthesized and modified into different geometries such as monoclinic, tetragonal, and cubic (Mueller et al. 2004). All the geometries of zirconia are temperature dependent. At room temperature, it is found in the monoclinic form. When heated to 2370 ◦ C, monoclinic geometry is converted to tetragonal and at 2690 ◦ C tetragonal geometry is converted to cubic form. Zirconia can also be doped with other metallic and nonmetallic oxides to get improved properties such as higher strength and fracture toughness (Tsunekawa et al. 2003). Zirconia nanobiomaterials are widely used in orthopedic implants, prosthetic knee implants, and dental restorations (Bartolome et al. 2007).

Carbon-Based Nano-biomaterials Carbon is one of the most important elements on the planet. Owing to its tetravalent structure, carbon can form single, double, or triple bonded structures with other elements such as hydrogen and nitrogen that makes it an extremely versatile element. Carbon nanomaterials also come in various shapes and forms. There are quantum dots, graphene and graphite sheets, diamond nanoplatelets, carbon nanotubes (CNT), and carbon nanofibers (CNF) just to name a few (Yanhong et al. 2006). These carbon nanomaterials, when infused with other materials, form extremely useful nano-biomaterials that have proved extremely useful to humans (Lin et al. 2016).

Types and Properties Graphene was isolated by Professors Andre Geim and Kostya Novoselov at the University of Manchester in 2004 for which both of them won the Nobel Prize in Physics. It is a two-dimensional (2D) material which forms the building blocks for a host of other nanomaterials like CNT. Graphene has excellent chemical, mechanical, and optical properties that make it useful for a host of applications. Moreover, it can also be made biocompatible by surface functionalization and because of this it is widely used in carbon-based nano-biomaterials (Feng et al. 2013). Graphene, CNTs, and CNFs are widely used as scaffolds to support the growth of injured bone and nerve tissues because of their excellent mechanical and electrical properties. Functionalization of CNTs with 4-hydroxynonenal has also shown to induce neural growth and branching (Mattson et al. 2000). Graphene oxide and SWNT can also be used as vectors for drug delivery because of their pH-sensitive nature (Xu and Wang 2006). Carbon nanomaterials such as fullerenes and CNTs also

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1251

display near-infrared (NIR) fluorescence and excellent electrochemical properties and are used as biosensors and bio-imaging agents (Chung et al. 2013).

Surface Modifications Graphene and CNTs are the most commonly used carbon nanomaterials. Thus, we will discuss some of the surface modifications of these nanomaterials. Graphene can be functionalized with nucleic acids, aptamers, and carbohydrates which make it more receptive to electrical signals (Bitounis et al. 2013). It can also be complexed with therapeutic molecules such as DOX for more efficient drug delivery systems. When coated with positively charged polyethyleneimine, graphene can be used as an extremely efficient drug delivery system (Zhang et al. 2012). Similarly, when CNTs are oxidized with carbonyl-based couplings, its solubility increases along with the increased possibility of conjugating with bioactive and therapeutic agents (Liu et al. 1998). When an amino group is attached with a CNT, it can link with other amino acids in the body and can be used for fluorescent probes (Bianco et al. 2005).

Silica-Based Nano-biomaterials Silica nanoparticles (SNP) are the most widely used nanoparticles for the production of polymer nano-composites. This is because SNPs are easy to produce and do not have any toxic by-products as is the case with some other nanoparticles. Also they possess many desirable properties such as high thermostability, biocompatibility, and also improve mechanical, thermal, and electrical properties of the NCs along with imparting biodegradability (Rahman and Padavettan 2012).

Production Sol-gel and micro emulsion are the most commonly used production techniques for SNPs. Hydrolysis of a precursor solution can also be done to produce SNPs. It was first reported by Stöber and Fink in 1968 (Stöber et al. 1968). SNPs have a 3D network structure as a result of which silanol and siloxane groups are created on the surface. The surface of SNPs are usually terminated by three silanol types; free or isolated silanol, H–bonded or vicinal silanols, and geminal silanols. These form H bonds with each other which lead to the formation of aggregates of SNPs (Wu et al. 2005). Thus, they have to be modified physically or chemically to counter the formation of agglomerates (Fig. 4). Modification by chemical interaction results in much stronger bonds between the modifiers and silica nanoparticles. It is by using modifying agents. The modifying agent or coupling agent has a hydrolysable group which reacts with the hydroxyl group present on the silica surface. The agent also contains an alkyl group which interacts with the polymer chains. Some common modifiers are TDI, epichlorohydrin, glycidyl phenyl ether, octadecylamine, etc. (Islam et al. 2013). Physical modification is mostly done on the surface of the SNP using surfactants or macromolecules. These molecules are adsorbed on the surface of the SNP. The

1252

A. Deka et al.

Fig. 4 Common silanes and agglomeration of SNP

principle of surfactant treatment works on the preferential adsorption of polar groups onto the surface of SNP by electrostatic interactions (Tang et al. 2007b). A surfactant can reduce agglomeration among SNPs by reducing physical interaction among each other which leads to uniform dispersion in the polymer matrix. It can also lead to increased hydrophobicity of the SNP (Tang et al. 2007a).

Properties SNPs when incorporated to a composite increases the tensile strength of the resultant nano-composite (Rong et al. 2001). However, the tensile strength increases to a particular loading point after it decreases again due to agglomeration of the SNPs. Other mechanical properties such as impact strength, hardness, and scratch resistance also improve when SNPs are incorporated to a composite system (Shang et al. 2002). SNPs are highly thermostable. They enhance the thermal properties of the NCs as they act as superior insulators and mass transport barriers to the volatile products generated during the decomposition process (Ray and Okamoto 2003). SNPs also form char when burned and act as a protective layer that stops oxygen penetration thus making the nano-composite flame retardant (Brancatelli et al. 2011). Transparency is the most important optical property for any material. However, introduction of SNPs, even at very low concentration often leads to opaque NCs. This is due to the scattering of light caused by the SNPs. For an NC containing SNP to be transparent, the NPs must be very finely dispersed in the matrix which can be obtained using a coupling agent such as 3-glycidyloxypropyltrimethoxysilane (GOTMS). GOTMS gets hydrolyzed to form silanol groups which can polycondensate with hydrolysis products of SNPs. GOTMS also get hydrolyzed to form hydroxyl groups that can form hydrogen bonds with carbonyl or hydroxyl groups

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1253

present in the polymer. These two factors together reduce the size and result in good dispersion of the SNPs (Zou et al. 2008).

Semiconductor-Based Nano-biomaterials Semiconductors are the class of materials with electrical conductivity between conductors and insulators. They can be elements such as silicon and germanium or compounds such as gallium arsenide and cadmium selenide (Peck 1991). Semiconductors have played an integral role in the ballistic pace of scientific development of the past five decades. They are an indispensible unit of the electronics industry. Starting from the one-dollar radio to the multi-million dollar satellite orbiting the earth, nothing would have been possible without the semiconductors (Weisbuch and Vinter 2014). Semiconductor nanomaterials are also used in biomedical applications such as bio-sensing, drug delivery, gene therapy, and in vitro and in vivo animal cell imaging (Zhou et al. 2015).

Types and Properties Cadmium-based nanoparticles are the most common semiconductor-based nanobiomaterials. They are mostly synthesized in the form of CdE quantum dots, where E = sulfide, selenide, and telluride (Mo et al. 2017). These quantum dots (QD) can be synthesized from cadmium containing precursors by organometallic synthesis methods (Bhattacharya et al. 2004), colloidal synthesis (Smith et al. 2008), and aqueous systems at low temperature (Esteve-Turrillas and Abad-Fuentes 2013). Properties of the cadmium quantum dots largely depend on their structure. QDs with a wide bandgap, such as CdS, display photoluminescence (Hines and GuyotSionnest 1996). Another very important property of the QDs is that they can be excited by a wide range of wavelengths because they have broad absorption spectra and can be used as an extremely potent bio-marker as a dye in bio-imaging (Han et al. 2001). Cd QDs can also be encased within other magnetic nanoparticles such as Fe2 O3 that imparts magnetic property to the entire core-shell structure (Schwartz et al. 2003). Cd QDs are inherently non-biocompatible. However, they can be modified using biodegradable moieties such as polymers, chitosan, cellulose, or hydroxyapatite that makes them biodegradable without hindering the intrinsic properties (Huang and Lee 2006). These nanoparticle systems can be used for drug delivery systems and gene therapy (Sharma et al. 2006).

Applications of Nano-biomaterials (Fig. 5) Nano-biomaterials in Therapeutics Nanostructured materials are widely employed for biological and biomedical applications due to their attractive physicochemical properties such as crystal structure (shape), size (nanoscale), and biocompatibility. Several of these materials

1254

A. Deka et al.

Fig. 5 Different applications of nano-biomaterials

are derived from natural resources which can be processed and transformed into structures (scaffolds or matrices) that find applicability in prevention, diagnosis, and treatment of a range of diseases. Also, considering the widespread prevalence of antimicrobial resistance (AMR) as reported recently by the World Health Organization’s (WHO) Global Antimicrobial Resistance Surveillance System (GLASS), nano-biomaterials are innovative tools to combat drug resistance by providing alternative medicine and prophylaxis approaches (Torres-Sangiao et al. 2016). Moreover, nano-biomaterials-based electrochemical sensors offer a sensitive and efficient early-stage diagnosis of cancer and other infectious diseases. Furthermore, the piezoelectric properties (charge accumulation in response to applied mechanical stress) of nano-biomaterials have been extensively exploited in developing regenerative medicine, implants, tissue engineering, and drug delivery systems (Sharma and Hussain 2020; Kapat et al. 2020). In this section, we highlight some of the recent applications of smart nano-biomaterials in: (i) antimicrobial therapy; (ii) regenerative medicine/artificial organs; (iii) sensors and diagnostics.

Nano-biomaterials in Antimicrobial Therapy In recent times the use of magnetic nanoparticles (NPs) particularly iron oxide NPs for biomedical applications (in vivo drug delivery) has been authorized by WHO due to its intrinsic antimicrobial activity, cost-effective synthesis, controlled release of the drug, and low toxicity (Rodrigues et al. 2019). Lipid-based polymeric nanoparticles are used for topical applications of antimicrobial agents. In

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1255

this regard, biodegradable and biocompatible polymers (nano-biomaterials) are attractive raw materials for the formulation of topical therapeutics (Severino et al. 2016). For instance, a promising antibacterial wound healing agent was developed by absorbing silver nanoparticles (AgNPs) on eggshell membranes (Li et al. 2019a). Similarly, a range of synthetic and natural polymers including polyacrylonitrile and chitosan were used to develop antibacterial mats or scaffolds (nanofibers) for prolonged delivery of natural bioactive drugs (Yadav and Balasubramanian 2016). Furthermore, several novel nanocomposites with antimicrobial properties have been developed using Nano-biomaterials like chitosan decorated nanotubes/hydrogels and octadecylamine capped Cu/reduced graphene oxide nanohybrids (Bellingeri et al. 2018; Ghosh et al. 2019). Antimicrobial hybrids with broad-spectrum antibiotic activity and controlled drug release properties such as silver phosphate-pectin microparticles loaded with levofloxacin advantageous as microdevice acting as biocide matrix (Bayón et al. 2016). Similar hybrids were developed by synthesizing silver phosphate microparticles on bacterial cellulose or silver nanoparticles loaded on poly-vinyl alcohol-lignin nanofibers (Bayón et al. 2018; Aadil et al. 2018).

Nano-biomaterials in Regenerative Medicine Nano-biomaterials that exhibit piezoelectric properties are attractive tools for therapeutic applications such as tissue regeneration (implants), drug delivery, and theranostics. The wide range of piezoelectric biomaterials includes naturally occurring piezocrystals (quartz, Rochelle salt), natural polymers (chitin, starch, fibrin, keratin, cellulose, and collagen), synthetic polymers (such as nylon-11, polyuria polyvinylidine fluoride (PVDF) and its copolymer trifluoroethylene-PVDF) and ceramics having large piezoelectric charge constant like BaTiO3 , PZT (lead zirconate titanate), and PMN-PT (lead magnesium niobate-lead titanate) (Kapat et al. 2020). Carbon nanomaterials like graphene oxide, carbon dots, carbon nanotubes, fullerenes, and nanodiamond offer good properties as scaffolds for tissue (bone, cartilage, tendon, and ligament) engineering by allowing adequate cell adhesion, growth, and differentiation. In addition, such carbon-based scaffolds or matrices offer biocompatibility, low toxicity, and mechanical strength (Eivazzadeh-Keihan et al. 2019). Several novel drug delivery nanosystems such as poly (L-lactic acid) or polycaprolactone NPs loaded with antithrombotic drug Dipyridamole are used extensively in cardiovascular treatments (Bakola et al. 2017, 2019). Collagen-based nano-biomaterials are used intensively for soft tissue engineering of cardiovascular, corneal, skin, muscle, and nerve regeneration (Purcel et al. 2016). The major advantages of collagen are its low antigenicity and biodegradability as a natural polymer. For example, collagen-inspired mineral-hydrogel nanocomposites and gelatin/PVA scaffold coated with nanofibrillated collagen has been used recently for hard tissue regeneration (Kim et al. 2019; Patel et al. 2020). Moreover, nano-biomaterials are also used extensively in theranostics (both therapeutics and diagnostics). The applications of such materials in diagnosis and sensors are described in the next section.

1256

A. Deka et al.

Nano-biomaterials as Sensors and Diagnostics Piezoelectric nano-biomaterials such as those based on Barium titanate-doxorubicin (BT-DOX), gadolinium doped zinc oxide-DOX (ZnO-GD), and bismuth ferrite (BFO) are widely used in radiotherapy, fluorescence, magnetic resonance (MR), and computed tomography (CT) imaging for real-time, cost-effective cancer diagnostics (Kapat et al. 2020; Rajaee et al. 2018). Several different nanomaterials such as quantum dots, gold NPs, magnetic NPs, and carbon nanotubes have been reported to be used for detection of malignant tumors (Zhang et al. 2019). 2DNanomaterials (graphene oxide, reduced graphene oxide) have garnered vast impact as electrochemical biosensors in cancer diagnosis (biosensing of cancer biomarkers like proteins and nucleic acids) (Wang et al. 2016). Furthermore, nanostructured ZnO-based materials have multifunctional biomedical applications such as imaging, drug delivery systems, and tissue regeneration and also as biosensors (Zhu et al. 2016). Over the past few years, aptamer-conjugated nano-biomaterials as biosensors (aptasensors) have revolutionized the detection of pathogenic bacteria in infectious diseases (Sharifi et al. 2020). These aptamers (single-stranded nucleic acids) integrated with a variety of nanoparticles such as gold NPs and supermagnetic iron oxide NPs having unique physicochemical properties are widely applied in calorimetric, fluorescent, and MR imaging for medical diagnostics (Lee et al. 2010). Functionalized nanomaterials also find wide applications as sensors for dope test in sports, detection of illicit drugs, biological fluids, gunshots, and other explosives during forensic investigation (Rawtani et al. 2019; Anil et al. 2020).

Nano-biomaterials in Food Nano-biomaterials have attractive applications in food and nutraceutical industry as platforms for immobilization of enzymes and other bioactive agents, smart packaging materials, nano-delivery systems in agri-food, direct food additives for preservation and fortification (Sampathkumar et al. 2020; Khan et al. 2018). Particularly, the use of nanomaterials like carbon nanotubes (CNTs), NPs, quantum dots in food safety applications (biosensing pollutants, pathogenic microbes, pesticides, and pH) is of immense importance in recent times (Socas-Rodríguez et al. 2017). Furthermore, the nano-biomaterials are also used as sorbents in extraction of different analytes from complex food materials and as emulsion stabilization agents.

Nano-biomaterials in Food Packaging Bio-nanocomposite materials are considered as an alternative to conventional nonbiodegradable plastic-based packaging as it is eco-friendly, lightweight, nontoxic and offers better performance (antioxidant, antimicrobial, and detection of exposure to oxygen level and temperature). Such materials are derived from renewable sources like natural polymers (cellulose, chitosan, starch, cellophane, and poly lactic acid) (Youssef and El-Sayed 2018; Khan et al. 2018). Nanomaterials that are applied onto the packaging matrix are categorized as inorganic (ZnO, AgNPs, TiO2 , Fe2 O3 , nanoclays) and organic (phenols, quaternary ammonium salts, polysaccharides, or proteins like chitin, chitosan, and whey protein isolates). The inorganic

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1257

nanomaterials offer strong bioactivity against pathogenic microbes. In addition, materials like nanoclays are used to improve mechanical strength of packaging biopolymers and also offer resistance to moisture and oxygen (Huang et al. 2018). In recent times, research has been directed toward utilization of lignocellulosic agro-wastes. In this regard, cellulosic nanomaterials like cellulose nanocrystals and nanofibrils provide multifunctional food applications particularly as packaging materials due to high mechanical strength and biocompatibility (Grishkewich et al. 2017; Zimmermann et al. 2010). Micro/nanofibrils derived from bacterial cellulose (gelatinous polysaccharide membrane) produced by bacteria G. xylinus also find potential applications in the packaging industry as a durable material with antimicrobial activity (Ullah et al. 2016).

Nano-biomaterials in Enzyme Immobilization Enzymes such as cellulase, lipase, β-glucosidase, α-amylase, and pectinase find wide applications in agro-food industry for bioconversion of cellulosic wastes into monomeric sugars, extraction of edible oil, as an animal feed additive to improve digestibility and clarification of fruit juices and wines (Husain 2017). However, the high cost of enzyme production coupled with limited stability renders the process expensive (Naqash et al. 2019). Development of nanoscale biocatalyst by nanoencapsulation (entrapment of bioactive enzymes onto carrier materials in nanoscale range) technique offers several advantages like increased stability to heat, pH and storage, resistant to inhibitory products and retention of catalytic activity on repeated use (Prakash and Khare 2019). In this context, nanomaterials such as gold and AgNPs and magnetic NPs provide biocompatible surfaces for immobilization and easy enzyme recovery. In addition, enzyme immobilizations on carbonaceous nanomaterials like nanowires and nanotubes, graphene oxide (GO), and reduced graphene oxide (rGO) preserves the native structure of the enzyme and its biological function (Liu and Dong 2020). Furthermore, in recent times, the applications of enzymatic nanobiosensors (enzyme immobilized onto transducer surface) in food safety assessment is imperative due to the widespread contamination of food by pollutants, pesticides, and food-borne pathogens (Verma 2017). Such nanobiosensors are also employed in the nutraceutical industry for detection of metabolites (flavonoids, phenolics, pigments such as carotenoids and anthocyanins) (Della Pelle and Compagnone 2018). Nano-biomaterials in Nano Food Delivery Systems Nano bio-based delivery system allows the incorporation of metabolites, bioactive ingredients, and nutraceuticals into food and beverages to produce “functional foods” with performance benefits like better digestibility, bioavailability, increased stability, and bioactivity of the nutrient (Jafari and McClements 2017). Several novel nano-food delivery systems have been developed in recent years using a wide range of nano-biomaterials like whey and soy protein-based hydrogels, chitosan-coated insect protein, lipid-based carriers (nanoliposomes), oligo-hyaluronic acid-curcumin polymer, etc. (Abaee et al. 2017; Okagu et al. 2020; Akhavan et al. 2018; Guo et al. 2018). Nanoencapsulation of major micronutrients such as phenolics, flavonoids, vitamins using biopolymer-based NPs, and natural

1258

A. Deka et al.

nano carriers (chitosan, sodium alginate, gelatin, caseins, and cyclodextrins) are widely used in the food industry as delivery systems (Esfanjani and Jafari 2016; Aditya et al. 2017). In addition, starch, chitosan-based nanocarriers (nanofibers, nanocomposites, nanohydrogels and NPs), dendrimers, and gums (Arabic gum, xanthan) are desired for efficient encapsulation of a broad range of phytochemicals and other bioactive compounds including carotenoids, essential oils, and alkaloids (Rostamabadi et al. 2019; Akbari-Alavijeh et al. 2020; Yousefi et al. 2020; Taheri and Jafari 2019).

Nano-biomaterials in Environmental Applications Environmental pollution (air, water, and soil) caused by the release of toxic chemicals as a result of industrial or human activities have received much attention by the scientific community all around the globe. Nanotechnology provides novel mitigation approaches for the removal of pollutants from the air, water, wastewater, and soil. Carbon-based nano-adsorbents (carbon nanotubes) are superior materials for adsorption of gases, dissolved organic substances (pollutants) as compared to powder activated carbon due to its large surface area, specific sorption sites, variable pore size, and surface chemistry (Panahi et al. 2018). Excessive use of agrochemicals, especially fertilizers and pesticides, has degraded the soil quality. However, nanomaterials nanoclays, zeolites, and metal oxide NPs can adsorb some of these pollutants as well as act as nutrient carriers at the same time (Pulimi and Subramanian 2016; Rani and Shanker 2018). Such nanomaterials have also been applied for the plant growth-promoting activities along with concomitant removal of pollutants from the soil (Mohamed et al. 2018). The release of industrial effluents in water bodies or oils spillage raises water pollution to alarming levels and resource usage concerns. Cellulose nanofibrils and nanocrystals incorporated nanocomposites provide wide applications in wastewater treatment processes such as filtration, sorption, and flocculation (Mohammed et al. 2018; Palit and Hussain 2020). Besides, nanocomposite materials are favored for scale-up applications in water-treatment by integrating the advantages of metal NPs and different solid host matrices of large size (Zhang et al. 2016). Magnetic NPs such as super-paramagnetic iron oxide NPs, magnetic nanocomposites, and carbon nanotubes are valuable tools for cleaning up of oil spills or leaks in water (Singh et al. 2020; Cardona et al. 2019). Some of the nano-biomaterials as smart devices or technologies developed in recent times for environmental applications have been discussed in (Table 1).

Nano-biomaterials in Bioenergy Applications Depletion of natural fossil fuel reserves and increase in emission of greenhouse gases contribute significantly to energy security and global climate change issues. Bioenergy refers to the production of gaseous and liquid fuels derived from renewable sources like biodiesel, bioethanol, biobutanol, biogas, bio-hydrogen, microbial fuel cells (MFCs), and other speciality biofuels such as higher alcohols,

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1259

Table 1 Environmental applications of nano-biomaterials and its advantages Environmental application Pollution control of air, water, plastic degradation

Nano-biomaterials ZnO-based nanostructures

Micronutrient availability and improved utilization efficiency for better crop yields Removal of heavy metals (Cu2+ , Pb2+ )/wastewater treatment

Microcrystalline cellulose and attapulgite-based bio-composite

Phosphate (PO4 3− ) removal from contaminated water

Mixed metal oxides (cerium/manganese) embedded on cellulose pine wood shaving (nanocomposites) Fe2 O3 NPs incorporated into permeable concrete

Microbiological and physicochemical pollutant removal from urban run-off/water quality monitoring

Zoledronate functionalized hydroxyapatite (nano-hybrid)

Water disinfection and microbial control

Carbon nanotubes

Adsorption of organic dye (methylene blue) from textile industrial wastewater

Apatite/attapulgite/ alginate composite hydrogels

Bioremediation of marine environment by (biostimulation and/or bio augmentation)

Metal NPs, carbon nanomaterials, metal oxide NPs, polymer-based nanocomposites

Advantages Semiconductor photocatalysis (under UV, solar light), degrades organic pollutants to CO2 and H2 O, generation of clean fuels pH-controlled release of foliar fertilizer, high adsorption capacity, micronutrient immobilization Enhanced adsorption capacity, pH-dependent removal of metals ions, facile, low-cost, eco-friendly Better compared to the individual metal oxide, economically feasible, reusability

References Chakrabarti et al. (2020)

Photo catalytic, wide range of pollutant in the surface runoff water could be degraded by oxidation, environmental friendly Antimicrobial properties, easily incorporated with polymers or NPs to form nanocomposite membrane, surface coating applications High adsorption capacity, improved mechanical strength of the gel, biodegradable and nontoxic Directly catalyze degradation of pollutants, promote the growth of microorganism able to degrade toxic materials

Ortega-Villar et al. (2019)

Wang et al. (2016)

Fang et al. (2020)

Nakarmi et al. (2020)

Liu et al. (2018)

Li et al. (2019b)

Cappello and Mancini (2019)

(continued)

1260

A. Deka et al.

Table 1 (continued) Environmental application Plant protection and growth

Removal of NOX gases from polluted air

Nano-biomaterials Chitosan-based nanomaterials (NPs, nanogels and nanocomposites)

Polymeric nanocomposites (poly vinylidene fluoride and poly dimethyl siloxane with TiO2 NPs)

Advantages Bioactivity toward broad-spectrum pathogenic microbes, biodegradable, induce seedling growth, photosynthesis, and nutrient uptake Photocatalytic, electrospun fibers as membrane filters, high surface area, porosity, and enhanced adsorption

References Kumaraswamy et al. (2018)

Majidi et al. (2018)

alkanes, and alkenes and energy-dense isoprenoids compounds (Bardhan et al. 2019). Nano-biomaterials find wide applications in bioenergy as novel supporting matrix for enzyme immobilization (cellulase, xylanase, and β-glucosidase for hydrolysis of complex lignocellulosic substrates, lipases for biodiesel production), bio-electrodes or conducting polymers of biofuel cells, green catalyst for biomass conversion, direct additives in biogas production by microorganisms (Verma et al. 2016; Sharma et al. 2020; Bhanja and Bhaumik 2016; Dehhaghi et al. 2019). Furthermore, metal nanoparticles such as iron and nickel NPs significantly improve bio-hydrogen yield as they act as (1) co-factors of hydrogenase and nitrogenase enzymes, (2) oxygen scavengers thereby creating anaerobic condition required for the activity of hydrogenase enzyme (Srivastava et al. 2020a). Porous nanomaterials (micro/mesoporous carbon, resin, silica, microporous zeolites, mesoporous metal oxides, and organic polymers) find wide application as a solid acid catalyst for the conversion of platform chemicals (5-hydroxymethyl furfural, levulenic acid, 2, 5-dimethylfuran) into liquid biofuels (Bhanja and Bhaumik 2016). Nanomaterials such as Fe3 O4 NPs impregnated eggshell, Cu impregnated TiO2 , K/Fe2 O3/γAl2 O3 -based nanocatalyst also play an active role as catalyst for transesterification of oil into biodiesel (Chingakham et al. 2019; Yazdani et al. 2019; De and Boxi 2020). A detailed list of nano-biomaterials with their advantages in different aspects of bioenergy applications have been discussed in Table 2.

Challenges In the past couple of decades, nano-biomaterials have made their way into the various aspects of our lives and made it more comfortable. However, there are still challenges associated with these materials which need to be overcome. Some of them are easy and efficient fabrication techniques and a true assessment of the level of toxicity of nano-biomaterials.

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1261

Table 2 Application of nano-biomaterials in different aspects of bioenergy and its advantages Bioenergy applications Lignocellulosic biomass pretreatment Delignification of sorghum stover/biohydorgen production

Nano-biomaterials

Advantages

References

Chitosan coated Fe3 O4 @SiO2 NPs.

Shanmugam et al. (2020)

Pretreatment of sugarcane bagasse/bioethanol production Enzymatic hydrolysis of rice straw in combination with ionic liquid pretreatment

Acid-functionalized magnetic NPs

Pretreatment of sugarcane bagasse/biohydorgen and biogas Cellulolytic hydrolysis of sugar beet pulp

NanoTiO2

Immobilization of Trichoderma asperellum laccase, higher catalytic efficiency, and reusability Easy recovery using a magnetic field, reusability, improved sugar yield Immobilization of Aspergillus niger cellulase, improved enzyme activity, reusability, and easy recovery Enhanced sugar, biohydorgen yield

Graphene oxidehydrogel bioconjugates (nanocarriers)

Improved saccharification efficiency, thermotolerant enzyme activity, enhanced specific activity, and storage stability

Ariaeenejad et al. (2020)

Porous nanomaterials (resin, mesoporous metal oxides, organic polymers) Ru/TiO2 nanocatalysts

Green catalysts, high surface acidity, and porous nanostructures (high surface area) Improved catalytic performance and stability, Ru dispersion and distribution on TiO2 surface High conversion yield, energy saving, eco-friendly

Bhanja and Bhaumik (2016)

Catalytic conversion of biomass to platform chemicals Production of HMF, furfural and 2,5-furandicarboxylic acid for liquid fuels Production of γ-valerolactone from methyl levulinate biowaste Photo catalytic conversion of glucose to gluconic acid, xylitol, arabinose, and formic acid

β-cyclodextrinconjugated magnetic NPs

TiO2 photocatalysts

Ingle et al. (2020)

Huang et al. (2015)

Jafari and Zilouei (2016)

Xu et al. (2018)

Payormhorm et al. (2017)

(continued)

1262

A. Deka et al.

Table 2 (continued) Bioenergy applications Furfural from biomass-derived xylose

Nano-biomaterials Zn doped CuO NP (nanoscale catalyst)

Oligomerization of itaconic anhydride (biomass-derived added value product)

Chitosan-cross linked magnetic Fe3 O4

Microbial fuel cells (MFCs) and carbon capture storage Electrode materials in single and double chamber MFCs

Carbon/graphene-based nanomaterials

Proton exchange membrane (PEM) in MFCs

Sulfonated graphene oxide (SGO)@SiO2 (nanocomposite)

Cathode catalyst

Rod shaped metallic NPs (Fe, Ni, Fe/Ni)

CO2 capture

CO2 derived-nonporous carbon (crystalline nanosheets)

CO2 capture and bioconversion to succinate

Magnetic (Fe3 O4 ) NPs coated A. succinogenes cell wall

Advantages Large surface area, enhanced catalytic activity, complete conversion at high temperature without any side products Immobilization of lipase from yeast Candida antarctica, good storage stability, reusability

References Mishra et al. (2019)

Large surface area, high electro catalytic activity, conductivity, and mechanical stability Improved proton conductivity, anti-fouling of PEM, increased electricity generation Enhanced power output, improved bacterial growth for biosurfactant production from waste vegetable oil in anodic chamber High CO2 adsorption capacity due to large surface area and volume of narrow microspores Improved CO2 capture efficiency, carbon resource utilization

Valipour et al. (2016), Ci et al. (2015)

Hosseini et al. (2019)

Xu et al. (2019)

Liu and Vipulanandan (2017)

Liu et al. (2020)

Li et al. (2016)

Fabrication Techniques Molecular Self-Assembly Molecular self-assembly is a fabrication process where the molecular components organize themselves into ordered structures with definite patterns without any external intervention (Whitesides and Grzybowski 2002). Various forms of nano-biomaterials such as nanoparticles, nanofibers, nanotubes, nanowires, and

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1263

nanopatterned surfaces can be fabricated using this technique (Whitesides and Boncheva 2002). The process of self-assembly can be either dynamic or static depending on the equilibrium of the systems involved (Philip and Stoddart 1996). This process is a result of the balance between the repulsive and attractive interactions such as Van der Waals forces and hydrogen bond, which is developed between different entities. By fine-tuning the different components and environment of the system, these forces come together to initiate self-assembly processes (Sun et al. 2002). Electrostatic interactions lead to the fabrication of nanofibers from ironic selfcomplementary peptides (Altman et al. 2000), whereas hydrophobic interactions lead to the formation of micelles which becomes a site for the fabrication of peptide fibers as well as supramolecular polymers (Bishop et al. 2009). Nano-biomaterials fabricated by this technique are biocompatible, biodegradable, nontoxic, and nonimmunogenic thus making them extremely useful for biomedical applications (Koutsopoulos 2012).

Electrospinning Electrospinning (ES) is the process of fabricating extremely thin layers of nanofibers from solutions or melt using an extremely high electric field. It is mostly used to fabricate polymer nanostructures (Greiner and Wendorff 2007). However, it is not a new process and the first instance of ES can be traced back to more than 70 years ago (Huang et al. 2003). In the process of ES, the solution that contains the precursor is forced through a syringe that is pushed automatically at a predetermined rate. An extremely high electric field (100–500 kV/m) is applied to the tip of the spinneret which also acts as an electrode. The collector is usually placed at a distance of 15–30 cm, depending on the setup. Against the surface tension, the applied voltage causes a cone-shaped deformation of the drop of solution at the spinneret tip, in the direction of the collector. When the two opposite forces are equal, the droplet of solution forms the famous “Taylor cone,” and the cone angle is 49.3◦ (Taylor 1964). At the critical value of the electrical field, a charged jet is ejected from the tip toward the collector. On its way to the collector, the solvent evaporates and a thin film of dry solute is deposited on the collector. Scaffolds that are used in tissue engineering applications can also be fabricated using ES technique (Lee et al. 2008). Electrospun nanofiber mats can also be used for drug delivery systems (Nikkola et al. 2006) and self-degrading wound dressings (Khanam et al. 2007). Nanopatterning Nanopatterning is a surface modification technique for nano-biomaterials. Various factors such as surface roughness, topography, and attached functional groups can alter properties of nano-biomaterials (Ostrovidov et al. 2015). Surface modification techniques for nano-biomaterials can be classified into two categories: (i) physical and chemical modification and (ii) coating the surface of a nano-biomaterial with biocompatible and/or bioactive agents to favor cell growth and the subsequent

1264

A. Deka et al.

Fig. 6 Types of nanopatterning

functions (Manickam et al. 2017). Nanopatterning can be of five types as seen in the diagram below (Fig. 6). Nano-biopatterning is making certain patterns at nanometer scale on biological samples such as cells and tissues (Zheng et al. 2005). It can alter cell properties such as adhesion and regeneration (Padmanabhan et al. 2014). Topographical patterning refers to the use of substrates with textured patterns by modulating its shape and size in the nanometer range. It is a physical modification and can be used to alter cellular responses in cells (Dalby et al. 2002). 3-D patterning is a relatively new approach and uses 3-D bioprinting. It creates an artificial 3-D environment for cell cultures and the cells will have different properties compared to regularly cultured cells (Sun et al. 2006). Chemical patterning refers to the use of various chemical agents to alter the surface of a biomaterial at nanoscale (Franco et al. 2000). Combinatioral patterning, as the name suggests, combines both chemical and topographical techniques to alter the surface of biomaterials to promote cell growth and cell orientations. It can also be used to fabricate biosensors (Charest et al. 2005).

Cytotoxicity Although nanomaterials (NMs) are being increasingly used for commercial and industrial applications, toxicological assessment for potential harmful effects on human cells is imperative. The cytotoxic effect like DNA-damaging properties of nanoparticles is primarily mediated through intrinsic generation of reactive oxygen species (ROS) and associated oxidative stress (Unfried et al. 2007). A comparative study of 10 engineered NMs found that Ag and ZnO NM in the concentration range of 0.3–80 μg/cm2 were cytotoxic in three tested human epithelial cell lines, whereas the cytotoxicity was absent in case of TiO2 and multi-walled carbon

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1265

nanotubes (Thongkam et al. 2017). Similarly, mesoporous silica NMs including MCM-41, two functionalized analogues of MCM-41, and SiO2 were found to be cytotoxic toward human neuroblastoma cells and the cytotoxicity was found to be associated with the absorptive surface area of the NM and the nature of the functional groups (Di Pasqua et al. 2008). The carbon-based NMs including single, multiwalled carbon nanotubes and fullerene were evaluated for cytotoxic effects like impaired phagocytosis, necrosis, and degeneration of alveolar macrophages and were reported to be dependent on the mass and concentration of the exposed NM (Jia et al. 2005). Furthermore, semiconductor quantum dots that find wide applications in biomedical devices are mostly made of heavy metal ions like Cd2+ which is often associated with in vitro cytotoxicity and therefore hamper their practical applications (Chen et al. 2012).

Biocompatibility Nanomaterials find wide biological and biomedical applications, such as tissue engineering, drug delivery, bio imaging, and biosensing. However, toxicity assessment before using NMs as implant devices or regenerative medicine is of great importance. It is well known that the embedded foreign material could elicit an immunologic response of the body and thereby hamper the functionality of such devices (Trindade et al. 2016). Therefore, it is imperative to design NMs keeping in mind the immune system responses to the implanted device. NMs with a high degree of biocompatibility will interact with the body without inducing immunogenic, thrombogenic, carcinogenic, and unwanted toxicity (Yoshioka et al. 2016). In this context, the physicochemical characteristics such as size, surface topography, electrostatic interactions, and mechanical properties of NMs are critical considering their role in immunogenicity, complement activation, coagulation, and biocompatibility (Rahmati and Mozafari 2019). Moreover, the cellular uptake, degradation, and bio-distribution properties of the NMs also influence the biological responses (Kunzmann et al. 2011).

Conclusion and Future Perspectives Humans have made great strides in fabrication of nano-biomaterials. They have considerably improved the quality of life. An injury or an ailment that could have seriously hindered an individual’s movement and subjected them to a wheelchair is no longer an issue. Moreover, these materials have also significantly helped in developing treatment for life-threatening diseases such as cancer. Not only biomedical applications, nano-biomaterials have also made an impact in the field of energy and environment. Because of its surface area, they are actively used in removal of pollutants from the environment. However, toxicity still remains one of the major issues associated with nanobiomaterials. As most of these materials are synthetically fabricated, they have

1266

A. Deka et al.

varying degrees of toxicity, cytotoxicity to be more specific. Researchers all over the world are working together to counter this aspect. Fabrication techniques of these materials may hold a key solution with regards to reducing the toxicity of nanobiomaterials. While conventional methods rely on extensive use of chemical agents, newer techniques such as eletrospinning and molecular self-assembly do not have any by-products. Thus, as these materials are used more and more in the products of mainstream markets, quality control and standardization is going to be ever so important. Therefore, the role of the scientific community will be more important than ever before to ensure the safety of the masses by ensuring strict quality control and streamlining the fabrication techniques of nano-biomaterials.

References Aadil KR, Mussatto SI, Jha H (2018) Synthesis and characterization of silver nanoparticles loaded poly (vinyl alcohol)-lignin electrospun nanofibers and their antimicrobial activity. Int J Biol Macromol 120:763–767 Abaee A, Mohammadian M, Jafari SM (2017) Whey and soy protein-based hydrogels and nanohydrogels as bioactive delivery systems. Trends Food Sci Technol 70:69–81 Abd Elkodous M, El-Sayyad GS, Abdelrahman IY, El-Bastawisy HS, Mosallam FM, Nasser HA et al (2019) Therapeutic and diagnostic potential of nanomaterials for enhanced biomedical applications. Colloids Surf B Biointerfaces 180:411–428 Aditya NP, Espinosa YG, Norton IT (2017) Encapsulation systems for the delivery of hydrophilic nutraceuticals: food application. Biotechnol Adv 35(4):450–457 Akbari-Alavijeh S, Shaddel R, Jafari SM (2020) Encapsulation of food bioactives and nutraceuticals by various chitosan-based nanocarriers. Food Hydrocoll 105:105774 Akhavan S, Assadpour E, Katouzian I, Jafari SM (2018) Lipid nano scale cargos for the protection and delivery of food bioactive ingredients and nutraceuticals. Trends Food Sci Technol 74: 132–146 Altman M, Lee P, Rich A, Zhang S (2000) Conformational behavior of ionic self-complementary peptides. Protein Sci 9(6):1095–1105 Anil D, Pandey G, Tharmavaram M, Rawtani D, Mustansar Hussain C (2020) Sensors for dope tests in sports. In: Technology in Forensic Science: sampling, analysis, data and regulations. Wiley-VCH, Weinheim, pp 259–278 Ariaeenejad S, Motamedi E, Salekdeh GH (2020) Stable cellulase immobilized on graphene oxide@ CMC-g-poly (AMPS-co-AAm) hydrogel for enhanced enzymatic hydrolysis of lignocellulosic biomass. Carbohydr Polym 230:115661 Bakola V, Karagkiozaki V, Pappa F, Tsiapla AR, Pavlidou E, Moutsios I, Logothetidis S (2017) Drug delivery nanosystems for cardiovascular stents. Mater Today 4(7):6869–6879 Bakola V, Tsiapla AR, Karagkiozaki V, Pappa F, Pavlidou E, Moutsios I, Logothetidis S (2019) Electrospray encapsulation of antithrombotic drug into poly (L-lactic acid) nanoparticles for cardiovascular applications. Mater Today 19:102–109 Bardhan P, Gupta K, Mandal M (2019) Microbes as bio-resource for sustainable production of biofuels and other bioenergy products. In: New and future developments in microbial biotechnology and bioengineering. Elsevier, pp 205–222 Bartolome JF, De Aza AH, Martín A, Pastor JY, Llorca J, Torrecillas R, Bruno G (2007) Alumina/zirconia micro/nanocomposites: a new material for biomedical applications with superior sliding wear resistance. J Am Ceram Soc 90(10):3177–3184 Bayón B, Bucalá V, Castro GR (2016) Development of antimicrobial hybrid mesoporous silver phosphate–pectin microspheres for control release of levofloxacin. Microporous Mesoporous Mater 226:71–78

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1267

Bayón B, Cacicedo ML, Álvarez VA, Castro GR (2018) Self-assembly stereo-specific synthesis of silver phosphate microparticles on bacterial cellulose membrane surface for antimicrobial applications. Colloid Interface Sci Commun 26:7–13 Bellingeri R, Mulko L, Molina M, Picco N, Alustiza F, Grosso C, Barbero CA (2018) Nanocomposites based on pH-sensitive hydrogels and chitosan decorated carbon nanotubes with antibacterial properties. Mater Sci Eng C 90:461–467 Bhanja P, Bhaumik A (2016) Porous nanomaterials as green catalyst for the conversion of biomass to bioenergy. Fuel 185:432–441 Bhat SV (2002) Overview of biomaterials. In: Biomaterials. Springer, Dordrecht, pp 1–11 Bhattacharya P, Ghosh S, Stiff-Roberts AD (2004) Quantum dot opto-electronic devices. Annu Rev Mater Res 34:83–122 Bianco A, Kostarelos K, Partidos CD, Prato M (2005) Biomedical applications of functionalised carbon nanotubes. Chem Commun 5:571–577 Bishop KJ, Wilmer CE, Soh S, Grzybowski BA (2009) Nanoscale forces and their uses in self assembly. Small 5(14):1600–1630 Bitounis D, Ali-Boucetta H, Hong BH, Min DH, Kostarelos K (2013) Prospects and challenges of graphene in biomedical applications. Adv Mater 25:2258–2268 Boisselier E, Astruc D (2009) Gold nanoparticles in nanomedicine: preparations, imaging, diagnostics, therapies and toxicity. Chem Soc Rev 38(6):1759–1782 Brancatelli G, Colleoni C, Massafra MR, Rosace G (2011) Effect of hybrid phosphorus-doped silica thin films produced by sol-gel method on the thermal behavior of cotton fabrics. Polym Degrad Stab 96(4):483–490 Cappello S, Mancini G (2019) Use of nanomaterials for marine bioremediation: a perspective. Ann Mater Sci Eng 4(1):1035 Cardona DS, Debs KB, Lemos SG, Vitale G, Nassar NN, Carrilho ENVM, Labuto G (2019) A comparison study of cleanup techniques for oil spill treatment using magnetic nanomaterials. J Environ Manag 242:362–371 Chae SK, Kim HY, Lee HY, Lee CH, Seo KM (2003) U.S. Patent No. 6,537,589. U.S. Patent and Trademark Office, Washington, DC Chakrabarti S, Banerjee P, Mitra P, Roy A (2020) Zinc oxide–based nanomaterials for environmental applications. In: Handbook of smart photocatalytic materials. Elsevier, pp 73–107 Charest JL, Eliason MT, García AJ, King WP, Talin AA, Simmons BA (2005) Polymer cell culture substrates with combined nanotopographical patterns and micropatterned chemical domains. J Vac Sci Technol B 23(6):3011–3014 Chen N, He Y, Su Y, Li X, Huang Q, Wang H, Fan C (2012) The cytotoxicity of cadmium-based quantum dots. Biomaterials 33(5):1238–1244 Chingakham C, David A, Sajith V (2019) Fe3O4 nanoparticles impregnated eggshell as a novel catalyst for enhanced biodiesel production. Chinese J Chem Eng 27(11):2835–2843 Chung C, Kim Y-K, Shin D, Ryoo S-R, Hong BH, Min D-H (2013) Biomedical applications of graphene and graphene oxide. Acc Chem Res 46:2211–2224 Ci S, Cai P, Wen Z, Li J (2015) Graphene-based electrode materials for microbial fuel cells. Sci China Mater 58(6):496–509 Corbett J, McKeown PA, Peggs GN, Whatmore R (2000) Nanotechnology: international developments and emerging products. CIRP Ann 49(2):523–545 Dalby MJ, Riehle MO, Johnstone HJH, Affrossman S, Curtis ASG (2002) Polymer-demixed nanotopography: control of fibroblast spreading and proliferation. Tissue Eng 8(6):1099–1108 Dalgleish BJ, Lu MC, Evans AG (1988) The strength of ceramics bonded with metals. Acta Metall 36(8):2029–2035 Dananjaya SHS, Godahewa GI, Jayasooriya RGPT, Lee J, De Zoysa M (2016) Antimicrobial effects of chitosan silver nano composites (CAgNCs) on fish pathogenic Aliivibrio (Vibrio) salmonicida. Aquaculture 450:422–430 Das M, Shim KH, An SSA, Yi DK (2011) Review on gold nanoparticles and their applications. Toxicol Environ Heal Sci 3(4):193–205

1268

A. Deka et al.

De A, Boxi SS (2020) Application of Cu impregnated TiO2 as a heterogeneous nanocatalyst for the production of biodiesel from palm oil. Fuel 265:117019 Dehhaghi M, Tabatabaei M, Aghbashlo M, Panahi HKS, Nizami AS (2019) A state-of-the-art review on the application of nanomaterials for enhancing biogas production. J Environ Manag 251:109597 Della Pelle F, Compagnone D (2018) Nanomaterial-based sensing and biosensing of phenolic compounds and related antioxidant capacity in food. Sensors 18(2):462 Deville S, Chevalier J, Fantozzi G, Bartolomé JF, Requena J, Moya JS, Díaz, L. A. (2003) Lowtemperature ageing of zirconia-toughened alumina ceramics and its implication in biomedical implants. J Eur Ceram Soc 23(15):2975–2982 Di Pasqua AJ, Sharma KK, Shi YL, Toms BB, Ouellette W, Dabrowiak JC, Asefa T (2008) Cytotoxicity of mesoporous silica nanomaterials. J Inorg Biochem 102(7):1416–1423 Edmundson MC, Capeness M, Horsfal L (2014) Exploring the potential of metallic nanoparticles within synthetic biology. New Biotechnol 31(6):572–578 Eivazzadeh-Keihan R, Maleki A, De La Guardia M, Bani MS, Chenab KK, Pashazadeh-Panahi P, Hamblin MR (2019) Carbon based nanomaterials for tissue engineering of bone: building new bone on small black scaffolds: a review. J Adv Res 18:185–201 Esfanjani AF, Jafari SM (2016) Biopolymer nano-particles and natural nano-carriers for nanoencapsulation of phenolic compounds. Colloids Surf B Biointerfaces 146:532–543 Esteve-Turrillas FA, Abad-Fuentes A (2013) Applications of quantum dots as probes in immunosensing of small-sized analytes. Biosens Bioelectron 41:12–29 Fang X, Zhu S, Ma J, Wang F, Xu H, Xia M (2020) The facile synthesis of zoledronate functionalized hydroxyapatite amorphous hybrid nanobiomaterial and its excellent removal performance on Pb2+ and Cu2+. J Hazard Mater 392:122291 Feng L, Wu L, Qu X (2013) New horizons for diagnostics and therapeutic applications of graphene and graphene oxide. Adv Mater 25(2):168–186 Feynman RP (1959) There’s plenty of room at the bottom. Eng Sci 23 Franco M, Nealey PF, Campbell S, Teixeira AI, Murphy CJ (2000) Adhesion and proliferation of corneal epithelial cells on self-assembled monolayers. J Biomed Mater Res 52(2): 261–269 Ghosh T, Bardhan P, Mandal M, Karak N (2019) Interpenetrating polymer network-based nanocomposites reinforced with octadecylamine capped cu/reduced graphene oxide nanohybrid with hydrophobic, antimicrobial and antistatic attributes. Mater Sci Eng C 105:110055 Greiner A, Wendorff JH (2007) Electrospinning: a fascinating method for the preparation of ultrathin fibers. Angew Chem Int Ed 46(30):5670–5703 Grishkewich N, Mohammed N, Tang J, Tam KC (2017) Recent advances in the application of cellulose nanocrystals. Curr Opin Colloid Interface Sci 29:32–45 Guo C, Yin J, Chen D (2018) Co-encapsulation of curcumin and resveratrol into novel nutraceutical hyalurosomes nano-food delivery system based on oligo-hyaluronic acid-curcumin polymer. Carbohydr Polym 181:1033–1037 Gupta AK, Gupta M (2005) Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications. Biomaterials 26(18):3995–4021 Han M, Gao X, Su JZ, Nie S (2001) Quantum-dot-tagged microbeads for multiplexed optical coding of biomolecules. Nat Biotechnol 19(7):631–635 Hasirci V, Vrana E, Zorlutuna P, Ndreu A, Yilgor P, Basmanav FB, Aydin E (2006) Nanobiomaterials: a review of the existing science and technology, and new approaches. J Biomater Sci Polym Ed 17(11):1241–1268 Hines MA, Guyot-Sionnest P (1996) Synthesis and characterization of strongly luminescing ZnScapped CdSe nanocrystals. J Phys Chem 100(2):468–471 Hosseini SM, Kim SM, Sayed M, Younesi H, Bahramifar N, Park JH, Pyo SH (2019) Lipaseimmobilized chitosan-crosslinked magnetic nanoparticle as a biocatalyst for ring opening esterification of itaconic anhydride. Biochem Eng J 143:141–150 Huang CY, Lee YD (2006) Core-shell type of nanoparticles composed of poly ((n-butyl cyanoacrylate)-co-(2-octyl cyanoacrylate)) copolymers for drug delivery application: synthesis, characterization and in vitro degradation. Int J Pharm 325(1–2):132–139

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1269

Huang ZM, Zhang YZ, Kotaki M, Ramakrishna S (2003) A review on polymer nanofibers by electrospinning and their applications in nanocomposites. Compos Sci Technol 63(15):2223– 2253 Huang PJ, Chang KL, Hsieh JF, Chen ST (2015) Catalysis of rice straw hydrolysis by the combination of immobilized cellulase from aspergillus Niger on β-cyclodextrin-Fe3O4 nanoparticles and ionic liquid. Biomed Res Int 2015. https://doi.org/10.1155/2015/409103 Huang Y, Mei L, Chen X, Wang Q (2018) Recent developments in food packaging based on nanomaterials. Nano 8(10):830 Husain Q (2017) Nanomaterials immobilized cellulolytic enzymes and their industrial applications: a literature review. JSM Biochem Mol Biol 4(3):1029 Ingle AP, Philippini RR, da Silva SS (2020) Pretreatment of sugarcane bagasse using two different acid-functionalized magnetic nanoparticles: a novel approach for high sugar recovery. Renew Energy 150:957–964 Islam MS, Masoodi R, Rostami H (2013) The effect of nanoparticles percentage on mechanical behavior of silica-epoxy nanocomposites. J Nanosci 2013 Jafari SM, McClements DJ (2017) Nanotechnology approaches for increasing nutrient bioavailability. In: Advances in food and nutrition research, vol 81. Academic Press, pp 1–30 Jafari O, Zilouei H (2016) Enhanced biohydrogen and subsequent biomethane production from sugarcane bagasse using nano-titanium dioxide pretreatment. Bioresour Technol 214:670–678 Jandt KD, Watts DC (2020) Nanotechnology in dentistry: present and future perspectives on dental nanomaterials. Dent Mater Jia G, Wang H, Yan L, Wang X, Pei R, Yan T, Guo X (2005) Cytotoxicity of carbon nanomaterials: single-wall nanotube, multi-wall nanotube, and fullerene. Environ Sci Technol 39(5):1378– 1383 Kalita SJ, Bhatt HA (2007) Nanocrystalline hydroxyapatite doped with magnesium and zinc: synthesis and characterization. Mater Sci Eng C 27(4):837–848 Kalita SJ, Bhardwaj A, Bhatt HA (2007) Nanocrystalline calcium phosphate ceramics in biomedical engineering. Mater Sci Eng C 27(3):441–449 Kapat K, Shubhra QT, Zhou M, Leeuwenburgh S (2020) Piezoelectric Nano-biomaterials for biomedicine and tissue regeneration. Adv Funct Mater:1909045 Kaur J, Sood K, Bhardwaj N, Arya SK, Khatri M (2020) Nanomaterial loaded chitosan nanocomposite films for antimicrobial food packaging. Mater Today 28:1904–1909 Khan A, Wen Y, Huq T, Ni Y (2018) Cellulosic nanomaterials in food and nutraceutical applications: a review. J Agric Food Chem 66(1):8–19 Khanam N, Mikoryak C, Draper RK, Balkus KJ Jr (2007) Electrospun linear polyethyleneimine scaffolds for cell growth. Acta Biomater 3(6):1050–1059 Kim H, Yang GH, Kim G (2019) Three-dimensional gelatin/PVA scaffold with nanofibrillated collagen surface for applications in hard-tissue regeneration. Int J Biol Macromol 135:21–28 Kohane DS, Langer R (2008) Polymeric biomaterials in tissue engineering. Pediatr Res 63(5): 487–491 Koo JH (2006) Polymer nanocomposites. McGraw-Hill Professional Pub Koutsopoulos S (2012) Molecular fabrications of smart nanobiomaterials and applications in personalized medicine. Adv Drug Deliv Rev 64(13):1459–1476 Kumaraswamy RV, Kumari S, Choudhary RC, Pal A, Raliya R, Biswas P, Saharan V (2018) Engineered chitosan based nanomaterials: bioactivities, mechanisms and perspectives in plant protection and growth. Int J Biol Macromol 113:494–506 Kunzmann A, Andersson B, Thurnherr T, Krug H, Scheynius A, Fadeel B (2011) Toxicology of engineered nanomaterials: focus on biocompatibility, biodistribution and biodegradation. Biochim Biophys Acta Gen Subj 1810(3):361–373 Landry CC, Harlan CJ, Bott SG, Barron AR (1995) Galloxane and Alumoxane hydroxides: (Ga12tBu12 (μ3-O) 8 (μ-O) 2 (μ-OH) 4) and (Al6tBu6 (μ3-O) 4 (μ-OH) 4). Angew Chem Int Ed Engl 34(11):1201–1202 Lee SJ, Liu J, Oh SH, Soker S, Atala A, Yoo JJ (2008) Development of a composite vascular scaffolding system that withstands physiological vascular conditions. Biomaterials 29(19):2891–2898

1270

A. Deka et al.

Lee JH, Yigit MV, Mazumdar D, Lu Y (2010) Molecular diagnostic and drug delivery agents based on aptamer-nanomaterial conjugates. Adv Drug Deliv Rev 62(6):592–605 Leena M, Rana D, Webster TJ, Ramalingam M (2016) Accelerated synthesis of biomimetic nano hydroxyapatite using simulated body fluid. Mater Chem Phys 180:166–172 Lendlein A (2010) Polymers in biomedicine. Macromol Biosci 10(9):993–997 Lendlein A, Rehahn M, Buchmeiser MR, Haag R (2010) Polymers in biomedicine and electronics. Macromol Rapid Commun 31(17):1487–1491 Leon L, Chung EJ, Rinaldi C (2020) A brief history of nanotechnology and introduction to nanoparticles for biomedical applications. In: Nanoparticles for biomedical applications. Elsevier, pp 1–4 Li Q, Zhang R, Wu D, Huang Y, Zhao L, Wang D, Ma G (2016) Cell-nanoparticle assembly fabricated for CO2 capture and in situ carbon conversion. J CO2 Util 13:17–23 Li X, Cai Z, Ahn DU, Huang X (2019a) Development of an antibacterial nanobiomaterial for wound-care based on the absorption of AgNPs on the eggshell membrane. Colloids Surf B Biointerfaces 183:110449 Li Y, Liu SJ, Chen FM, Zuo JE (2019b) High-strength apatite/attapulgite/alginate composite hydrogel for effective adsorption of methylene blue from aqueous solution. J Chem Eng Data 64(12):5469–5477 Liao S, Ramakrishna S, Ramalingam M (2011) Development of nanofiber biomaterials and stem cells in tissue engineering. J Biomater Tissue Eng 1(2):111–128 Lin X, Clasky A, Lai K, Yang L (2016) Carbon-based nano biomaterials: design, fabrication and application. Biomed Nanomater 49 Liu DM, Dong C (2020) Recent advances in nano-carrier immobilized enzymes and their applications. Process Biochem 92:464–475 Liu J, Vipulanandan C (2017) Effects of Fe, Ni, and Fe/Ni metallic nanoparticles on power production and biosurfactant production from used vegetable oil in the anode chamber of a microbial fuel cell. Waste Manag 66:169–177 Liu J, Rinzler AG, Dai H (1998) Fullerene pipes. Science 280:1253–1256 Liu D, Mao Y, Ding L (2018) Carbon nanotubes as antimicrobial agents for water disinfection and pathogen control. J Water Health 16(2):171–180 Liu S, Jin Y, Bae JS, Chen Z, Dong P, Zhao S, Li R (2020) CO2 derived nanoporous carbons for carbon capture. Microporous Mesoporous Mater:110356 Ma MG, Zhu JF (2010) Recent progress on fabrication of calcium-based inorganic biodegradable nanomaterials. Recent Pat Nanotechnol 4(3):164–170 Majidi R, Parhizkar J, Karamian E (2018) Photocatalytic removal of NOx gas from air by TiO2/polymer composite nanofibers. Nanochem Res 3(2):212–218 Manickam LP, Bhatt A, Rana D, Ostrovidov S, Pasricha R, Wang X, Ramalingam M (2017) Nanopatterning techniques. In: Nanobiomaterials: classification, fabrication and biomedical applications. Wiley, pp 189–204 Mattson MP, Haddon RC, Rao AM (2000) Molecular functionalization of carbon nanotubes and use as substrates for neuronal growth. J Mol Neurosci 14(3):175–182 Mishra RK, Kumar VB, Victor A, Pulidindi IN, Gedanken A (2019) Selective production of furfural from the dehydration of xylose using Zn doped CuO catalyst. Ultrason Sonochem 56:55–62 Mo D, Hu L, Zeng G, Chen G, Wan J, Yu Z, Cheng M (2017) Cadmium-containing quantum dots: properties, applications, and toxicity. Appl Microbiol Biotechnol 101(7):2713–2733 Mohamed MA, Hashim AF, Alghuthaymi MA, Abd-Elsalam KA (2018) Nano-carbon: plant growth promotion and protection. In: Nanobiotechnology applications in plant protection. Springer, Cham, pp 155–188 Mohammed N, Grishkewich N, Tam KC (2018) Cellulose nanomaterials: promising sustainable nanomaterials for application in water/wastewater treatment processes. Environ Sci Nano 5(3):623–658 Mueller R, Jossen R, Pratsinis SE, Watson M, Akhtar MK (2004) Zirconia nanoparticles made in spray flames at high production rates. J Am Ceram Soc 87(2):197–202

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1271

Nakarmi A, Chandrasekhar K, Bourdo SE, Watanabe F, Guisbiers G, Viswanathan T (2020) Phosphate removal from wastewater using novel renewable resource-based, cerium/manganese oxide-based nanocomposites. Environ Sci Pollut Res:1–16 Naplocha K, Janus A, Kaczmar JW, Samsonowicz Z (2000) Technology and mechanical properties of ceramic preforms for composite materials. J Mater Process Technol 106(1–3): 119–122 Naqash F, Masoodi FA, Rather SA (2019) Food enzymes and nanotechnology. In: Enzymes in food biotechnology. Academic, pp 769–784 Niinomi M, Narushima T, Nakai M (2015) Advances in metallic biomaterials. Springer, Heidelberg Nikkola L, Seppälä J, Harlin A, Ndreu A, Ashammakhi N (2006) Electrospun multifunctional diclofenac sodium releasing nanoscaffold. J Nanosci Nanotechnol 6(9–10):3290–3295 Okagu OD, Verma O, McClements DJ, Udenigwe CC (2020) Utilization of insect proteins to formulate nutraceutical delivery systems: encapsulation and release of curcumin using mealworm protein-chitosan nano-complexes. Int J Biol Macromol 151:333–343 Ortega-Villar R, Lizárraga-Mendiola L, Coronel-Olivares C, López-León LD, Bigurra-Alzati CA, Vázquez-Rodríguez GA (2019) Effect of photocatalytic Fe2O3 nanoparticles on urban runoff pollutant removal by permeable concrete. J Environ Manag 242:487–495 Ostrovidov S, Shi X, Sadeghian RB, Salehi S, Fujie T, Bae H, Khademhosseini A (2015) Stem cell differentiation toward the myogenic lineage for muscle tissue regeneration: a focus on muscular dystrophy. Stem Cell Rev Rep 11(6):866–884 Padmanabhan J, Kinser ER, Stalter MA, Duncan-Lewis C, Balestrini JL, Sawyer AJ, Kyriakides TR (2014) Engineering cellular response using nanopatterned bulk metallic glass. ACS Nano 8(5):4366–4375 Palit S, Hussain CM (2020) Carbon-based polymer nanocomposite and environmental perspective. Emerging Carbon-Based Nanocomposites for Environmental Applications:121 Panahi Y, Mellatyar H, Farshbaf M, Sabet Z, Fattahi T, Akbarzadehe A (2018) Biotechnological applications of nanomaterials for air pollution and water/wastewater treatment. Mater Today 5(7):15550–15558 Parida P, Behera A, Mishra SC (2012) Classification of biomaterials used in medicine. Inter J Adv Appl Sci 1(3):31–35 Park J, Lakes RS (2007) Biomaterials: an introduction. Springer Science & Business Media Park YK, Tadd EH, Zubris M, Tannenbaum R (2005) Size-controlled synthesis of alumina nanoparticles from aluminum alkoxides. Mater Res Bull 40(9):1506–1512 Patel A, Zaky SH, Schoedel K, Li H, Sant V, Beniash E, Sant S (2020) Design and evaluation of collagen-inspired mineral-hydrogel nanocomposites for bone regeneration. Acta Biomater 08(112):262–273 Payormhorm J, Chuangchote S, Kiatkittipong K, Chiarakorn S, Laosiripojana N (2017) Xylitol and gluconic acid productions via photocatalytic-glucose conversion using TiO2 fabricated by surfactant-assisted techniques: effects of structural and textural properties. Mater Chem Phys 196:29–36 Peck RL (1991) U.S. Patent No. 5,055,171. U.S. Patent and Trademark Office, Washington, DC Percec V, Guliashvili T, Ladislaw JS, Wistrand A, Stjerndahl A, Sienkowska MJ, Sahoo S (2006) Ultrafast synthesis of ultrahigh molar mass polymers by metal-catalyzed living radical polymerization of acrylates, methacrylates, and vinyl chloride mediated by SET at 25 C. J Am Chem Soc 128(43):14156–14165 Philp D, Stoddart JF (1996) Self assembly in natural and unnatural systems. Angew Chem Int Ed Engl 35(11):1154–1196 Prakash O, Khare S (2019) Immobilization of α-amylases and their analytical applications. In: Biocatalysis. Springer, Cham, pp 113–138 Pulimi M, Subramanian S (2016) Nanomaterials for soil fertilisation and contaminant removal. In: Nanoscience in food and agriculture, vol 1. Springer, Cham, pp 229–246 Purcel G, Meli¸ta˘ D, Andronescu E, Grumezescu AM (2016) Collagen-based nanobiomaterials: challenges in soft tissue engineering. In: Nanobiomaterials in soft tissue engineering. William Andrew Publishing, pp 173–200

1272

A. Deka et al.

Rahman IA, Padavettan V (2012) Synthesis of silica nanoparticles by sol-gel: size-dependent properties, surface modification, and applications in silica-polymer nanocomposites – a review. J Nanomater 2012. https://doi.org/10.1155/2012/132424 Rahmati M, Mozafari M (2019) Nano-immunoengineering: opportunities and challenges. Curr Opin Biomed Eng 10:51–59 Rajaee A, Wensheng X, Zhao L, Wang S, Liu Y, Wu Z, Si-Shen F (2018) Multifunctional bismuth ferrite nanoparticles as magnetic localized dose enhancement in radiotherapy and imaging. J Biomed Nanotechnol 14(6):1159–1168 Rana D, Kumar TS, Ramalingam M (2014) Cell-laden hydrogels for tissue engineering. J Biomater Tissue Eng 4(7):507–535 Rani M, Shanker U (2018) Remediation of polycyclic aromatic hydrocarbons using nanomaterials. In: Green adsorbents for pollutant removal. Springer, Cham, pp 343–387 Rawtani D, Tharmavaram M, Pandey G, Hussain CM (2019) Functionalized nanomaterial for forensic sample analysis. TrAC Trends Anal Chem 120:115661 Ray SS, Okamoto M (2003) Prog Polym Sci 28:1539 Reddy LRM, Reddy MO, Thaidun M, Reddy PG (2008) Biomaterial and its Applications Renuka NK, Shijina AV, Praveen AK (2012) Mesoporous γ-alumina nanoparticles: synthesis, characterization and dye removal efficiency. Mater Lett 82:42–44 Rhee W, Wallace DG, Michaels AS, Burns RA Jr, Fries L, DeLustro F, Bentz H (1992) U.S. Patent No. 5,162,430. U.S. Patent and Trademark Office, Washington, DC Richardson JT, Peng Y, Remue D (2000) Properties of ceramic foam catalyst supports: pressure drop. Appl Catal A Gen 204(1):19–32 Rodrigues GR, López-Abarrategui C, de la Serna Gómez I, Dias SC, Otero-González AJ, Franco OL (2019) Antimicrobial magnetic nanoparticles based-therapies for controlling infectious diseases. Int J Pharm 555:356–367 Rong MZ, Zhang MQ, Zheng YX, Zeng HM, Friedrich K (2001) Improvement of tensile properties of nano-SiO2/PP composites in relation to percolation mechanism. Polymer 42(7):3301–3304 Rostamabadi H, Falsafi SR, Jafari SM (2019) Starch-based nanocarriers as cutting-edge natural cargos for nutraceutical delivery. Trends Food Sci Technol 88:397–415 Sampathkumar K, Tan KX, Loo SCJ (2020) Developing Nano-delivery Systems for Agriculture and Food Applications with nature-derived polymers. Iscience:101055 Sarkar S, Jana AD, Samanta SK, Mostafa G (2007) Facile synthesis of silver nano particles with highly efficient anti-microbial property. Polyhedron 26(15):4419–4426 Sarkar M, Prakash RM, Hussain CM, Shankar MV (2020) Ferroelectric semiconductors for photocatalytic energy and environmental applications. In: Handbook of smart photocatalytic materials. Elsevier, pp 3–19 Schwartz DA, Norberg NS, Nguyen QP, Parker JM, Gamelin DR (2003) Magnetic quantum dots: synthesis, spectroscopy, and magnetism of Co2+−and Ni2+−doped ZnO nanocrystals. J Am Chem Soc 125(43):13205–13218 Severino P, Fangueiro JF, Chaud MV, Cordeiro J, Silva AM, Souto EB (2016) Advances in nanobiomaterials for topical administrations: new galenic and cosmetic formulations. In: Nanobiomaterials in Galenic formulations and cosmetics. William Andrew Publishing, pp 1–23 Shang XY, Zhu ZK, Yin J, Ma XD (2002) Compatibility of soluble polyimide/silica hybrids induced by a coupling agent. Chem Mater 14(1):71–77 Shanmugam S, Krishnaswamy S, Chandrababu R, Veerabagu U, Pugazhendhi A, Mathimani T (2020) Optimal immobilization of Trichoderma asperellum laccase on polymer coated Fe3O4@ SiO2 nanoparticles for enhanced biohydrogen production from delignified lignocellulosic biomass. Fuel 273:117777 Sharifi S, Vahed SZ, Ahmadian E, Dizaj SM, Eftekhari A, Khalilov R, Labib M (2020) Detection of pathogenic bacteria via nanomaterials-modified aptasensors. Biosens Bioelectron 150:111933 Sharma D, Hussain CM (2020) Smart nanomaterials in pharmaceutical analysis. Arab J Chem 13(1):3319–3343

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1273

Sharma P, Brown S, Walter G, Santra S, Moudgil B (2006) Nanoparticles for bioimaging. Adv Colloid Interf Sci 123:471–485 Sharma A, Singh G, Arya SK (2020) Biofuel cell nanodevices. Int J Hydrog Energy. https://doi.org/ 10.1016/j.ijhydene.2020.02.164 Singh H, Bhardwaj N, Arya SK, Khatri M (2020) Environmental impacts of oil spills and their remediation by magnetic nanomaterials. Environ Nanotechnol Monit Manag 100305 Smith AM, Duan H, Mohs AM, Nie S (2008) Bioconjugated quantum dots for in vivo molecular and cellular imaging. Adv Drug Deliv Rev 60(11):1226–1240 Smith IO, Liu XH, Smith LA, Ma PX (2009) Nanostructured polymer scaffolds for tissue engineering and regenerative medicine. Wiley Interdiscip Rev Nanomed Nanobiotechnol 1(2): 226–236 Socas-Rodríguez B, González-Sálamo J, Hernandez-Borges J, Rodríguez-Delgado MÁ (2017) Recent applications of nanomaterials in food safety. TrAC Trends Anal Chem 96:172–200 Srivastava N, Srivastava M, Mishra PK, Kausar MA, Saeed M, Gupta VK, Ramteke PW (2020a) Advances in nanomaterials induced biohydrogen production using waste biomass. Bioresour Technol:123094 Srivastava M, Srivastava N, Mishra PK, Malhotra BD (2020b) Prospects of nanomaterials-enabled biosensors for COVID-19 detection. Sci Total Environ:142363 Stöber W, Fink A, Bohn E (1968) Controlled growth of monodisperse silica spheres in the micron size range. J Colloid Interface Sci 26(1):62–69 Sun S, Anders S, Hamann HF, Thiele JU, Baglin JEE, Thomson T, Terris BD (2002) Polymer mediated self-assembly of magnetic nanoparticles. J Am Chem Soc 124(12):2884–2885 Sun T, Jackson S, Haycock JW, MacNeil S (2006) Culture of skin cells in 3D rather than 2D improves their ability to survive exposure to cytotoxic agents. J Biotechnol 122(3):372–381 Taheri A, Jafari SM (2019) Gum-based nanocarriers for the protection and delivery of food bioactive compounds. Adv Colloid Interf Sci 269:277–295 Tahir MB, Rafique M, Sagir M, Ullah S, Kiran H (2019) Nanomaterials for photocatalytic applications Tang JC, Lin GL, Yang HC, Jiang GJ, Chen-Yang YW (2007a) Polyimide-silica nanocomposites exhibiting low thermal expansion coefficient and water absorption from surface-modified silica. J Appl Polym Sci 104(6):409–4105 Tang JC, Yang HC, Chen SY, Chen-Yang YW (2007b) Preparation and properties of polyimide/silica hybrid nanocomposites. Polym Compos 28(5):575–581 Taylor GI (1964) Disintegration of water drops in an electric field. Proc R Soc Lond A Math Phys Sci 280(1382):383–397 Teja AS, Koh PY (2009) Synthesis, properties, and applications of magnetic iron oxide nanoparticles. Prog Cryst Growth Charact Mater 55(1–2):22–45 Thongkam W, Gerloff K, van Berlo D, Albrecht C, Schins RP (2017) Oxidant generation, DNA damage and cytotoxicity by a panel of engineered nanomaterials in three different human epithelial cell lines. Mutagenesis 32(1):105–115 Torres-Sangiao E, Holban AM, Gestal MC (2016) Advanced nanobiomaterials: vaccines, diagnosis and treatment of infectious diseases. Molecules 21(7):867 Trindade R, Albrektsson T, Tengvall P, Wennerberg A (2016) Foreign body reaction to biomaterials: on mechanisms for buildup and breakdown of osseointegration. Clin Implant Dent Relat Res 18(1):192–203 Trunec M, Maca K (2007) Compaction and pressureless sintering of zirconia nanoparticles. J Am Ceram Soc 90(9):2735–2740 Tsunekawa S, Ito S, Kawazoe Y, Wang JT (2003) Critical size of the phase transition from cubic to tetragonal in pure zirconia nanoparticles. Nano Lett 3(7):871–875 Ullah H, Santos HA, Khan T (2016) Applications of bacterial cellulose in food, cosmetics and drug delivery. Cellulose 23(4):2291–2314 Unfried K, Albrecht C, Klotz LO, Von Mikecz A, Grether-Beck S, Schins RP (2007) Cellular responses to nanoparticles: target structures and mechanisms. Nanotoxicology 1(1):52–71

1274

A. Deka et al.

Valipour A, Ayyaru S, Ahn Y (2016) Application of graphene-based nanomaterials as novel cathode catalysts for improving power generation in single chamber microbial fuel cells. J Power Sources 327:548–556 Verma ML (2017) Enzymatic nanobiosensors in the agricultural and food industry. In: Nanoscience in food and agriculture, vol 4. Springer, Cham, pp 229–245 Verma ML, Puri M, Barrow CJ (2016) Recent trends in nanomaterials immobilised enzymes for biofuel production. Crit Rev Biotechnol 36(1):108–119 Wang M, Zhang G, Zhou L, Wang D, Zhong N, Cai D, Wu Z (2016) Fabrication of pH-controlledrelease ferrous foliar fertilizer with high adhesion capacity based on nanobiomaterial. ACS Sustain Chem Eng 4(12):6800–6808 Weisbuch C, Vinter B (2014) Quantum semiconductor structures: fundamentals and applications. Elsevier Whitesides GM, Boncheva M (2002) Beyond molecules: self-assembly of mesoscopic and macroscopic components. Proc Natl Acad Sci 99(8):4769–4774 Whitesides GM, Grzybowski B (2002) Self-assembly at all scales. Science 295(5564):2418–2421 Williams DF, Cunningham J (1979) Materials in clinical dentistry. Oxford University Press, Oxford Wu CL, Zhang MQ, Rong MZ, Friedrich K (2005) Silica nanoparticles filled polypropylene: effects of particle surface treatment, matrix ductility and particle species on mechanical performance of the composites. Compos Sci Technol 65(3–4):635–645 Wu W, He Q, Jiang C (2008) Magnetic iron oxide nanoparticles: synthesis and surface functionalization strategies. Nanoscale Res Lett 3(11):397 Xu M, Wang LV (2006) Photoacoustic imaging in biomedicine. Rev Sci Instrum 77(4):041101 Xu C, Ouyang W, Muñoz-Batista MJ, Fernández-García M, Luque R (2018) Highly active catalytic Ru/TiO2 nanomaterials for continuous production of γ-valerolactone. Chem SusChem 11(15):2604–2611 Xu Q, Wang L, Li C, Wang X, Li C, Geng Y (2019) Study on improvement of the proton conductivity and anti-fouling of proton exchange membrane by doping SGO@ SiO2 in microbial fuel cell applications. Int J Hydrog Energy 44(29):15322–15332 Yadav R, Balasubramanian K (2016) Bioabsorbable engineered nanobiomaterials for antibacterial therapy. In: Engineering of Nanobiomaterials. William Andrew Publishing, pp 77–117 Yang L, Hu C, Nie Y, Qu J (2009) Catalytic ozonation of selected pharmaceuticals over mesoporous alumina-supported manganese oxide. Environ Sci Technol 43(7):2525–2529 Yan-Hong Y, Run-Cai M, Jin-Tao B, Xun H (2006) Photoluminescence of multiwalled carbon nanotubes excited at different wavelengths. Chinese Phys 15(11):2761 Yazdani F, Akia M, Khanbolouk F, Hamze H, Arandiyan H (2019) Novel heterogeneous base nanocatalysts supported on a spray dried gamma alumina applying optimized production of biodiesel from waste cooking oil. Biofuels 1–7 Yoshioka Y, Higashisaka K, Tsutsumi Y (2016) Biocompatibility of nanomaterials. In: Nanomaterials in pharmacology. Humana Press, New York, pp 185–199 Yousefi M, Narmani A, Jafari SM (2020) Dendrimers as efficient nanocarriers for the protection and delivery of bioactive phytochemicals. Adv Colloid Interf Sci 102125 Youssef AM, El-Sayed SM (2018) Bionanocomposites materials for food packaging applications: concepts and future outlook. Carbohydr Polym 193:19–27 Yuan-Tao N, Huai-Zhi Z (2003) Silver in Nano-materials. Precious Metals Yunnan 24(3):54–60 Zhang Y, Nayak TR, Hong H, Cai W (2012) Graphene: a versatile nanoplatform for biomedical applications. Nanoscale 4(13):3833–3842 Zhang Y, Wu B, Xu H, Liu H, Wang M, He Y, Pan B (2016) Nanomaterials-enabled water and wastewater treatment. NanoImpact 3:22–39 Zhang Y, Li M, Gao X, Chen Y, Liu T (2019) Nanotechnology in cancer diagnosis: progress, challenges and opportunities. J Hematol Oncol 12(1):137 Zheng GF, Patolsky F, Lieber CM (2005) Multiplexed electrical detection of cancer marker proteins and single viruses by nanowire fet arrays. In: Abstracts of papers of the American chemical society, vol 229, pp u782–u782. 1155 16th st, nw, Washington, dc 20036 USA: Amer Chemical Soc

49 Nano-biomaterials as a Potential Tool for Futuristic Applications

1275

Zhou J, Yang Y, Zhang CY (2015) Toward biocompatible semiconductor quantum dots: from biosynthesis and bioconjugation to biomedical application. Chem Rev 115(21):11669–11717 Zhu P, Weng Z, Li X, Liu X, Wu S, Yeung KWK, Chu PK (2016) Biomedical applications of functionalized ZnO nanomaterials: from biosensors to bioimaging. Adv Mater Interfaces 3(1):1500494 Zimmermann T, Bordeanu N, Strub E (2010) Properties of nanofibrillated cellulose from different raw materials and its reinforcement potential. Carbohydr Polym 79(4):1086–1093 Zou H, Wu S, Shen J (2008) Polymer/silica nanocomposites: preparation, characterization, properties, and applications. Chem Rev 108(9):3893–3957

Important Websites/Links Related to the Topic https://www.elsevier.com/books/nanobiomaterials/narayan/978-0-08-100716-7 https://www.icevirtuallibrary.com/toc/jbibn/current https://www.longdom.org/proceedings/application-of-nanobiomaterials-in-healthcare-32734.html https://www.researchgate.net/publication/236123036_Nanobiomaterials_Handbook_Chapter_1_ Nanobiomaterials_Current_and_Future_Prospects

Industry 4.0: Applications and Future Perspectives

50

Rafael Kunst, Gabriel Ramos, Rodrigo Righi, Cristiano André da Costa, Edison Pignaton, Alecio Binotto, Jose Favilla, Ricardo Ohta, and Rob High

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Internet of Things . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Real-Time Location Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Computer Vision Techniques to Analyze the Behavior of Moving Entities . . . . . . . . . . . Sensors, Integration Board, and Computing Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . . Data Compressing to Optimize IoT-Data Representation . . . . . . . . . . . . . . . . . . . . . . . . . . Edge Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Products and Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Cities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Healthcare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Military Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1278 1280 1280 1283 1284 1286 1288 1292 1292 1294 1296

R. Kunst () · G. Ramos · R. Righi · C. A. da Costa University of Vale do Rio dos Sinos (UNISINOS), São Leopoldo, Brazil e-mail: [email protected]; [email protected]; [email protected]; [email protected] E. Pignaton Federal University of Rio Grande do Sul (UFRGS), Porto Alegre, Brazil e-mail: [email protected] A. Binotto IBM Consulting, München, Germany e-mail: [email protected] J. Favilla IBM Global Markets, Coppell, TX, USA e-mail: [email protected] R. Ohta IBM Research, São Paulo, Brazil e-mail: [email protected] R. High IBM Cloud and Cognitive Software, Durham, NC, USA e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_33

1277

1278

R. Kunst et al.

Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1298 1302 1302

Abstract

Industry 4.0 introduces several changes to the original approach of industrial automation. Internet of Things (IoT) and cyber-physical system (CPS) technologies play huge roles in this context introducing cognitive automation and consequently implementing the concept of intelligent production, leading to smart products and services. This approach leads companies to face challenges of a much more dynamic environment. Many of these companies are not ready to deal with this new scenario where the existence of a large amount does not always collaborate to increase productivity. This book chapter explores the concepts of IoT and machine learning along with the consequent applications in the context of Industry 4.0. Applications covered throughout the chapter include solutions in the areas of IoT, edge computing, agriculture, smart cities, healthcare, and military operations. Keywords

Industry 4.0 · IoT · Machine learning · Smart cities · Smart agriculture · Healthcare · Cyver-physical system

Introduction The possibility of having billions of connected devices with specific purpose processing units and users connected by mobile devices brings the availability of low-cost processing power, storage capacities, and Internet connectivity on those devices. When shifting part of the processing to such smart devices, the concept of smart automation in different industries, like a smart factory, becomes a reality in a way that actions can be performed in real time where the data was created. Data and, more importantly, analytics are changing how we see our machines, processes, products, and operations composed of several smart devices. Artificial intelligence, combined with big data approaches, can identify patterns in the data, uncovering model behaviors. This is promoting emerging technology breakthroughs, covering wide-ranging fields such as robotics, cyber-physical systems (CPS), the Internet of Things (IoT), autonomous vehicles, nanotechnology, smart energy networks, energy storages (Abraham et al. 2020), currently COVID-19 solutions for back to work, etc. Figure 1 shows one of the results of a systematic literature review (Dalzochio et al. 2020) that covered the last 5 years of research involving machine learning and reasoning in the context of industry 4.0. Although this is only part of the

50 Industry 4.0: Applications and Future Perspectives

ACM

IEEE

1279

Science Direct

Springer

Google Scholar Initial Search

4

56

123

63

316

Database Joint 562

Impurity Removal 288 155

Exclusion Criteria Filter by Title

124 60

Filter by Abstract Filter by Full Text

44 38

Filter by Quality Assessment

Fig. 1 Research on Industry 4.0 in the last 5 years (Dalzochio et al. 2020)

possible applications of Industry 4.0, one can see that the topic gained a lot of attention in recent years. The initial search in the leading scientific databases that focus on Industry 4.0 resulted in 562 papers. After removing duplicates found in more than one database, surveys, reviews, book chapters, or nonscientific papers like magazine articles, 288 out of 562 papers remained in the authors’ analysis. These are the papers that are more relevant for industry 4.0. The literature review’s remaining phases excluded documents that were not focused on more specific Industry 4.0 topics, like using machine learning models or ontologies to solve particular problems. The previous industrial revolution focused mainly on improving the physical manufacturing processes, expanding human power with additional power sources (machinery, steam power), establishing a mass production process through the introduction of assembly lines, and introducing electronics and automation. The 4th industrial revolution, also known as Industry 4.0, focuses primarily on creating a digital representation of the physical processes to get better insights into the physical processes. For example, production equipment may have some early signs that something is going wrong and that a breakdown may happen soon. These signs may be detected by predictive models that indicate the deviation from normal operating conditions. The digital model can provide early insights about the equipment’s status, allowing the maintenance personnel to determine the best time to repair it, moving from a reactive to planned repair (Dalzochio et al. 2020). Industry 4.0 has a very ambitious scope, aiming to create digital factories, i.e., a digital representation of the physical operations, sometimes called cyber-physical models or digital twins (Grieves 2015). It aims at integrating processes from the

1280

R. Kunst et al.

top floor to the shop floor and from suppliers to the end clients, creating vertical and horizontal integration across the value chain. Another goal is to reduce the product design life cycle by creating a digital thread that integrates key processes to design, build, operate, and maintain the equipment. It is also relevant to establish a feedback loop from operation to product engineering to create a piece of fully connected equipment. The connected equipment, regardless of its location in the factory, provides the basis for predictive maintenance. The main idea is to collect various online and off-line signals from the equipment to feed models that can detect an early indication of an anomaly or fault. In the path toward cognitive enterprises, a platform assumes that three logical steps should be covered through a timely, sequenced road map to ensure value: data gathering (structured and unstructured), analytics (predictive analytics, prescriptive analytics, and industry-specific AI models), and pattern visualization (for decisionmaking if actuation is not done automatically and dynamically). In this sense, this chapter is focused on discussing the current applications of Industry 4.0 and the perspectives of the field. Toward the focus of the chapter, section “Internet of Things” discusses IoT as an enabling technology to implement the concepts of Industry 4.0. Section “Edge Computing” discusses the importance of edge computing in the context and presents challenges regarding this approach. Section “Smart Products and Services” focuses on the presentation and discussion of four typical applications of Industry 4.0. Conclusions and future perspectives are presented in section “Conclusion”.

Internet of Things First, this section presents the technologies that one can use in an industrial plant to collect data about people’s position and their behavior in a specific moment. In particular, the section addresses real-time location systems (RTLSs) and computer vision techniques. Also, considering the monitoring of machines, it also explores sensors that used to monitor devices’ health, presenting a viable infrastructure to collect data on Industry 4.0 environments. Finally, the section describes an optimization technique for data visualization data that uses data compression to receive better and plot Industry 4.0-related information.

Real-Time Location Systems RTLS are solutions for indoor identification and location tracking of people and assets. Such systems consist of a set of fixed readers or anchors reading wireless signals from tags (Boulos and Berry 2012). The system applies position estimation methods to these signals, outputting the tag position in its coordinate system. Thus, by assigning a tag to a specific target, it is possible to monitor its location within a building facility (indoor spaces). Nowadays, specifically in the healthcare

50 Industry 4.0: Applications and Future Perspectives

1281

landscape, solutions are employing wireless aensor network (WSN) technologies in RTLS, such as Wi-Fi, Radio Frequency Identification (RFID), Bluetooth, and Ultra-wideband (UWB) (Tan et al. 2015; Adame et al. 2018). Particularly, UWB technologies surpass low-frequency technologies when it comes to interference since it implements adaptive frequency hopping (Lee et al. 2007). WSNs can provide location information, which is a crucial factor in understanding the application context (Liu et al. 2012). The low cost of sensor technology has eased the proliferation of WSNs in many areas, such as healthcare and smart buildings (Mainetti et al. 2011). A WSN is a network of tiny devices that cooperate using wireless protocols to collect information about a target physical environment. WSNs comply with a wide range of solutions, thus being characterized by their high heterogeneity (Mainetti et al. 2011). Data gathered by different devices can be stored and combined locally or sent to other networks, such as the Internet. A typical WSN usage scenario is to employ technologies such as Wi-Fi or Bluetooth for indoor location purposes. WSNs applied to the healthcare field aim to improve monitoring systems and services. For example, Wheeler (2007) demonstrates the value of a WSN that can report the location of patients, medical staff, and critical equipment. Another example is presented in Fig. 2, which depicts the Logitrak (https://logi-tag. com/real-time-location-system/) solution from Logi-Tag Systems. These technologies generally employ sensor devices embedded with memory, processor, and wireless communication protocols to transmit data. A WSN uses a collection of devices to produce information about things and the context. However, a WSN itself is not able to identify a target object independently. Conversely, RFID is a modern technology aiming to provide object identification in a short range. Therefore, combining WSN and RFID technologies is an attractive solution for better information monitoring. Differently from printed code technologies, RFID (ISOs 15693 and 14443) permits short-range identification without requiring visibility between readers and tags. RFID tags have exclusive IDs and can store and transmit information about the manufacturer, environment, and technical parameters. They segment into two principal standards: active, which require a power source like batteries, and passive, which do not require a power source. Proposals extensively apply RFID solutions aiming at activity recognition since RFID is a mature and low-cost technology. In turn, the creation of the Near Field Communication (NFC) Forum in 2004 is a direct result of attempts to spread RFID applications further. This forum’s goal is to bring together existing mobile RFID standardization efforts and introduce short-range communication capabilities into RFID. It also aims to standardize mechanisms in which sensor devices can exchange information in very short distances. NFC technologies that operate in the high-frequency band at 13.56 MHz (ISO 14443, ISO 18092) support tag readings from distances of 10 cm (Want 2011). Regarding wireless communication protocols, UWB, Wi-Fi, and Bluetooth are part of the short-range wireless field (Lee et al. 2007). UWB is a radio frequency technology that provides information exchange by transmitting data through continuous short radio pulses. Wi-Fi is a well-known protocol that allows data

1282

R. Kunst et al.

Fig. 2 An RFID-based RTLS called LogiTrack, which provides hospitals, nursing homes, and clinics a method to accurately track, locate, and monitor assets and people, and trigger events in real time based on location and status

transmission in higher ranges with larger data throughput. However, these factors result in higher values of energy consumption. Lastly, Bluetooth aims to cover wireless communication in short ranges focusing on low-cost devices. This protocol works in the 2.45 GHz band employing frequency hopping strategies to increase performance. The IEEE standard 802.15.4 (IEEE Computer Society 2016) aims at wireless communication devices with low-energy consumption, cost, and data rate (Lee et al. 2007). Many communication technologies apply this standard, for example, Bluetooth Low Energy (BLE) (Bluetooth 2017), 6LoWPAN (RFC 4944) (Montenegro et al. 2007), and ZigBee (Alliance 2017; Farahani 2008). BLE is an attractive choice for WSN applications that require high data transmission in short distances between devices. Similarly to BLE, ZigBee is a wireless communication technology for applications that focus on low-energy consumption and cost (Lee et al. 2007). Likewise, the 6LoWPAN standard adapts the IPv6 over IEEE 802.15.4 networks focusing on low-energy approaches. Recent work promotes the adoption of this standard instead of proprietary, rigid ones (Mainetti et al. 2011). Table 1 summarizes technical aspects from the aforementioned technologies.

50 Industry 4.0: Applications and Future Perspectives

1283

Table 1 Technical details of the communication standards and protocols Technology RFID NFC Wi-Fi Bluetooth UWB ZigBee 6LoWPAN Specification ISO ISO IEEE IEEE IEEE IEEE IEEE 15693, 14443, 802.11 802.15.1 802.15.3a 802.15.4 802.15.4, ISO ISO RFC 4944 14443, 18092 ISO 18000 Frequency 1 tending to a maximum value of 15, as can be seen that in Fig. 23, the area which has FOS < 1 is the same region from Fig. 22 which experiences greater amount stress but that does not prove that there are chances of failure at that point. It cannot be proved without any proper physical testing of the part. Hence, instead of reinforcing that region, we are willing to take the risk as weight reduction had been one of our objectives (Figs. 24 and 25). The von Mises stress is used to predict yielding of materials under complex loading from the results of uniaxial tensile tests. In Fig. 26, the same region from

55 Robotic Arm for Biomedical Applications

1431

Fig. 22 Principle stress analysis of back arm

Fig. 23 FOS analysis of back arm

Fig. 23 and 24 shows that the rest of the region has a higher amount of equivalent (von Mises) stress. Force and load analysis of the design are conducted. In Fig. 27, the side hinges of the shoulder experience a lateral force of 500 N or -500 N in +X or –X direction, respectively. In Fig. 28, the clamp for the smaller driving gear experiences a bending moment of “100 N-mm.” In Fig. 29, the extruded controller attached to the palm was applied with a moment of 100 N-mm.

1432

A. S. Arockia Doss et al.

Fig. 24 FOS analysis of shoulder

Fig. 25 FOS analysis of palm

In Fig. 30, the teeth of larger gear attached to the forearm end would experience an 800 N reaction force in “–Y” direction from the teeth of the smaller gear attached to the shoulder. In Fig. 31, the hinge plates of the forearm experience a lateral reaction force of 500 N in either +X or –X direction by the hinges of the shoulder.

Summary and Future Scope A prosthetic arm model comprising of 20 degrees of freedom was designed and analyzed. Based on the design and analyses, the following points are summarized.

55 Robotic Arm for Biomedical Applications

1433

Fig. 26 Von Mises stress analysis of back arm

Fig. 27 Force analysis of shoulder

The conceptual model was developed with 18 revolute joints and 2 twisting joints, each providing 1 degree of freedom. Three joint linkages in the fingers help in attaining multiple grip positions compared to a single joint linkage mechanism found in other major designs. The cosmetic appeal of this design is better than other prosthetic arms available. These joints provide the model with a superior range of motion and grip system. Starting from meshing, various structural analyses, including the stress flow analysis, von Mises stress analysis, and point load analysis of individual parts, were carried out. The model was applied with 800 N force to access its ability to withstand activities of daily living. The data collected was used to make design changes and assess the amount of deformation of the model. The factor of safety

1434

A. S. Arockia Doss et al.

Fig. 28 Load analysis of shoulder

Fig. 29 Load analysis of palm

(FOS) of each part was improved and as a result, the endurability of the overall design was increased. The designed model is lighter than a human arm and most of the robotic prosthetic arm available today. A human arm on average weight is 3.6 kg, while the average weight of a prosthetic arm is 1.2 kg. The prosthetic arm’s designed model weighs at 950 g, which is 70% lighter in weight than a human arm and 20% lighter than most of the prosthetic arms available. This research would pave a path for future research on robotic prosthetic arms and their development. The derived design can be manufactured using 3D Printing, aiding efficiency. As the 3D printing technology advancements are made,

55 Robotic Arm for Biomedical Applications

1435

Fig. 30 Force analysis on shoulder gear

Fig. 31 Force analysis on shoulder hinges

the manufacturing cost would lower considerably and make the prosthesis more affordable and popular. A further strengthening of the design could be carried out and materials with a higher strength to weight ratios can be used for manufacturing. Though this would lead to a further increase in costs. The control system would remain a challenge for the researchers as it is necessary to maintain ease of use and simplicity from the perspective of the amputee.

1436

A. S. Arockia Doss et al.

References Abayasiri A, Madusanka K, Arachchige N, Silva A, Gopura R (2017) MoBio: a 5 DOF transhumeral robotic prosthesis. In: IEEE international conference on rehabilitation robotics. pp 1627–1632 Ahmed S et al (2020) Fabrication of cost-effective prosthetic arm using electroencephalography signal. In: International conference on industrial & mechanical engineering and operations management, Dhaka, Bangladesh, 26–27 December 2020 Ali AMM, Ambar R, Kadir KA, Mohd MNH, Khan MRB, Halim SA, Sharon WAW, Aznizam NN (2018) Preliminary design of a robotic exoskeleton for arm rehabilitation. In: International postgraduate conference on applied science & physics Anh MNTD, Tang VT, Trung LC, Hieu NH, Tu HL, Minh LH Quoc (2018) Cost-Effective Design and Development of a Prosthetic Hand. In International Conference on the Development of Biomedical Engineering in Vietnam. 225–229. Springer, Singapore Arjunan SP, Kumar DK, Bueno L, Villarejo Mayor J, Bastos Filho TF (2014) Upper limb prosthesis devices. Devices for Mobility and Manipulation for People with Reduced Abilities (1st ed.), CRC Press, 1–16. https://doi.org/10.1201/b16870 Bajaj NM, Spiers AJ, Dollar AM (2019) State of the art in artificial wrists: a review of prosthetic and robotic wrist design. IEEE Trans Robot 35(1):261–277 Bandara DSV, Gopura RARC, Hemapala KTMU, Kiguchi K (2017) Development of a multi-DoF transhumeral robotic arm prosthesis. Med Eng Phys 48:131–141 Bridgwater LB et al (2012) The robonaut 2 hand-designed to do work with tools. In: Proceedings of the IEEE international conference on Robotics and Automation, pp 3425–3230 Carignan C, Liszka M (2005) Design of an arm exoskeleton with scapula motion for shoulder rehabilitation. In: Proceedings of the 12th international conference on advanced robotics, 2005. ICAR’05. IEEE, pp 524–531 Christopher C, Evans CO, Smith SB, Muller AH, Kerwin JM, Schnellinger TS, Michael Guay G, van der Merwe DA (2019) Arm prosthetic device. United States Patent Application Publication Clement R, Bugler K, Oliver C (2011) Bionic prosthetic hands: a review of present technology and future aspirations. J R Coll Surg Edinburgh 9:336–340. https://doi.org/10.1016/ j.surge.2011.06.001 Degirmenci A, Hammond FL, Gafford JB, Walsh CJ, Wood RJ, Howe RD (2015) Design and control of a parallel linkage wrist for robotic microsurgery. In: Proceedings of the IEEE international conference on intelligent robotics and systems, pp 222–228 Dholariya K, Arockia Selvakumar A (2016) Conceptual design and kinematic analysis of humanoid robot. Indian J Sci Technol 9:34 Dovat L, Lambercy O, Gassert R, Maeder T, Milner T, Leong TC, Burdet E (2008) HandCARE: a cable-actuated rehabilitation system to train hand function after stroke. IEEE Trans Neural Syst Rehabil Eng 16(6):582–591 Duan X, Yang Y, Cheng B (2016) Modeling and analysis of a 2-DOF spherical parallel manipulator. Sensors (Switzerland) 16(9):1–15 Fan S, Jiang L, Liu H (2016) A design of a miniaturized prosthetic wrist based on repetition rate of human wrist daily tasks. In: Proceedings of the international conference on robotics and biomimetics, pp 1643–1648 Flesher SN, Downey JE, Weiss JM, Hughes CL, Herrera AJ, Tyler-Kabara EC, Boninger ML, Collinger JL, Gaunt RA (2021) A brain-computer interface that evokes tactile sensations improves robotic arm control. Science 372(6544):831–836 Gopura RARC, Kiguchi K (2009) Mechanical designs of active upper-limb exoskeleton robots: State-of-the-art and design difficulties. In: 2009 IEEE international conference on rehabilitation robotics, pp 178–187 Gopura R, Kiguchi K, Mann GKI, Torricelli D (2018) Robotic prosthetic limbs. J Robot 2018:1–2. https://doi.org/10.1155/2018/1085980

55 Robotic Arm for Biomedical Applications

1437

Gutiérreza SC, Zotovicb R, Navarroa MD, Meseguera MD (2017) Design and manufacturing of a prototype of a lightweight robot arm. In: Manufacturing engineering society international conference 2017, MESIC 2017, 28–30 June 2017, Vigo (Pontevedra), Spain Hussin Q, Sadun AS (2021) User-driven prosthetic robotic arm for patients with lower arm amputation. Prog Eng Appl Technol 2(1):753–768 Kamil A, Arockia Selvakumar A (2018) Simulation and analysis of 10-DoF biped robot. Int J Mech Prod Eng Res Dev 8(8):113–119 Kim N, Yun S, Shin D (2019) A bioinspired lightweight wrist for high-DoF robotic prosthetic arms. IEEE/ASME Trans Mechatron 24(6):2674–2683 Klein J, Spencer S, Allington J, Bobrow JE, Reinkensmeyer DJ (2010) Optimization of a parallel shoulder mechanism to achieve a high-force, low-mass, roboticarm exoskeleton. IEEE Trans Robot 26(4):710–715 Krebs HI, Volpe BT, Williams D, Celestino J, Charles SK, Lynch D, Hogan N (2007) Robotaided neurorehabilitation: a robot for wrist rehabilitation. IEEE Trans Neural Syst Rehabil Eng 15(3):327–335 Kulkarni T, Uddanwadiker R (2016) Overview: mechanism and control of a prosthetic arm. Mol Cell Biomech 12:1–46 Kumar VS, Aswath S, Shashidhar TS, Choudhary RK (2017) A novel design of a full length prosthetic robotic arm for the disabled. In: Kim JH, Karray F, Jo J, Sincak P, Myung H (eds) Robot intelligence technology and applications 4. Advances in intelligent systems and computing, vol 447. Springer, Cham Kundu SK, Kiguchi K, Horikawa E (2008) Design and control strategy for a 5 DOF above-elbow prosthetic arm. Int J Assist Robot Mechatron 9:79–93 Maat B, Smit G, Plettenburg D, Breedveld P (2017) Passive prosthetic hands and tools: a literature review. Prosthetics Orthot Int 42:030936461769162. https://doi.org/10.1177/ 0309364617691622 Manna SK, Bhaumik S (2013) A bioinspired 10 DOF wearable powered arm exoskeleton for rehabilitation. J Robot 2013:1–15. https://doi.org/10.1155/2013/741359 Mao Y, Agrawal SK (2011) A cable driven upper arm exoskeleton for upper extremity rehabilitation. In: 2011 IEEE international conference on robotics and automation (ICRA). IEEE, Shanghai, China, pp 4163–4168 Mendoza-Vázquez JR, Escudero-Uribe AZ, Tlelo-Cuautle E (2008) Modeling and simulation of a parallel mechanical elbow with 3 DOF. In: 2008 electronics, robotics and automotive mechanics conference (CERMA’08), pp 455–460 Nef T, Mihelj M, Riener R (2007) ARMin: a robot for patient-cooperative arm therapy. Med Biol Eng Comput 45(9):887–900 O’Malley MK, Sledd A, Gupta A, Patoglu V, Huegel J, Burgar C (2006) The RiceWrist: a distal upper extremity rehabilitation robot for stroke therapy. In: ASME 2006 international mechanical engineering congress and exposition. American Society of Mechanical Engineers, Chicago, Illinois, pp 1437–1446 Pehlivan AU, Celik O, O’Malley MK (2011) Mechanical design of a distal arm exoskeleton for stroke and spinal cord injury rehabilitation. In: 2011 IEEE international conference on rehabilitation robotics (ICOR. IEEE, Zurich, Switzerland, pp 1–5 Plagenhoef S, Evans FG, Abdelnour T (1983) Anatomical data for analyzing human motion. Res Q Exerc Sport 54:169–178 Reza Kashefa S, Amini S, Akbarzadeh A (2019) Robotic hand: a review on linkage-driven finger mechanisms of prosthetic hands and evaluation of the performance criteria. Department of Mechanical Engineering, Ferdowsi University of Mashhad, Mashhad Rosen J, Perry JC, Manning N, Burns S, Hannaford B (2005) The human arm kinematics and dynamics during daily activities – toward a 7 DOF upper limb powered exoskeleton. In: ICAR’05. Proceedings of the 12th international conference on advanced robotics, 2005, pp 532–539

1438

A. S. Arockia Doss et al.

Saliba MA, Camilleri D, Farrugia MJ (2005) Development of an anthropomorphic robot hand and wrist for teleoperation applications. In: International conference on information and automation. pp 203–208 Sameer A (2016) Human hand prosthesis. https://doi.org/10.13140/RG.2.1.4460.8889 Sarah Mattes MS (2014) Is lighter better? Thoughts on the relationship between device weight and function. The O&P EDGE, 5:30–38. https://www.oandp.com/articles/2014-05_07.asp Scarcia U, Melchiorri C, Palli G (2015) Towards simplicity: on the design of a 2-DOFs wrist mechanism for tendon-driven robotic hands. In: Proceedings of the IEEE international conference on Robotics and Biomimetics, pp 1317–1322 Ueki S, Kawasaki H, Ito S, Nishimoto Y, Abe M, Aoki T, Mouri T (2012) Development of a hand-assist robot with multi-degrees-of-freedom for rehabilitation therapy. IEEE/ASME Trans Mechatron 17(1):136–146 Vatsal V, Hoffman G (2017) Wearing your arm on your sleeve: studying usage contexts for a wearable robotic forearm. 974–980. https://doi.org/10.1109/ROMAN.2017.8172421 Vujaklija I, Farina D, Aszmann O (2016) New developments in prosthetic arm systems. Orthop Res Rev 20168:31–39. https://doi.org/10.2147/ORR.S71468 Wolf EJ, Cruz TH, Emondi AA et al (2020) Advanced technologies for intuitive control and sensation of prosthetics. Biomed Eng Lett 10:119–128 Zaidan AH, Wail MK, Yaseen AA (2021) Design and implementation of upper prosthetic controlled remotely by flexible sensor glove. In: IOP conference series: materials science and engineering, vol 1105. IOP Publishing, Baghdad, Iraq

Progresses on Green and Smart Materials for Multifaceted Applications

56

S. O. Oyedepo, Joseph O. Dirisu, N. E. Udoye, and O. S. I. Fayomi

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multifunctional Applications in Energy, Construction, Infrastructure, Electronics, and Buildings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Concept of Smart Materials, Traditional Materials, Modern Materials, Nanomaterials, and Composite Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Material and Waste Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Materials for the Industry and Smart Industrial Equipment . . . . . . . . . . . . . . . . . . . . . Concept of Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Cooking Appliances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of Smart Materials in Energy and Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of Smart Materials in Building and Construction . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1440 1443 1443 1449 1449 1450 1452 1461 1462 1463 1463

Abstract

Sustainable development is widely known as a critical issue for the future growth and well-being of our society. Technological advancement has provided smart material development opportunities for multifunctional applications in energy, construction, infrastructure, electronics, and building. Green materials

S. O. Oyedepo () · J. O. Dirisu · N. E. Udoye Mechanical Engineering Department, Covenant University, Ota, Ogun State, Nigeria e-mail: [email protected] O. S. I. Fayomi () Department of Mechanical and Biomedical Engineering, Bells University of Technology, Ota, Ogun State, Nigeria e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_41

1439

1440

S. O. Oyedepo et al.

are proficient, safe, economically viable, and reliable because of their ecofriendliness and renewable prowess. This chapter looks into the progress of natural and resource-efficient materials for smart manufacturing processes that could be used for intelligent application and address contemporary materials’ challenges. Keywords

Sustainable development · Smart materials · Technological advancement · Manufacturing processes · Material development

Introduction The subject of sustainability depicts the sustenance of living things, especially humans in the ecosystem, as a function of using resources over the generations. A product from a resource could be recommendable but may become a threat to the next generation. Sustainable development has been a universal topic of interest to the international academic communities, building stakeholders, international conferences, and the theme of developers and green environment crusaders (Mensah and Casadevall 2019). Smart materials have distinct properties that are developed in controlled factors such as stress, moisture, and temperature control, and electric and magnetic fields control. Intelligent material selection influences the composite makeup of modern-day products without compromising the product’s integrity and quality during the performance. The future generation benefits from this product in meeting their target as there is satisfaction in the material selection (Boström et al. 2018). Smart technology employs materials that can considerably alter their thermomechanical, electromagnetic, and optical properties in a regulated environment (Konarzewska 2017a; Li et al. 2017). Materials from the earth’s crust, such as clay, sand, and stone, are green materials as their location is beneath the ground. It does have numerous utilizations in diverse disciplines of medical science and technology. The increase for the advanced materials is a pointer to the opportunities in stock innovative materials for the present and future generation (Li et al. 2017). Active smart materials and passive intelligent materials are the two major classes of smart materials. Several green and smart material applications include piezoelectric, magnetostrictive, magnetic shape memory alloys, and artificial intelligence, amongst others (Müller and Schmid 2019). Piezoelectric materials such as bone, crystals, certain ceramics, DNA, enamel, silk, dentin, etc. are materials that produce interior electrical charge from applied mechanical stress (Mayeen and Kalarikkal 2018). A magnetostrictive material is materials that have tiny ferromagnets. These ferromagnets act like tiny permanent bar magnets consisting of iron, nickel, or cobalt, having small magnetic moments due to their “3d” shells partly filled with electrons (Zverev et al. 2018). The magnetic shape memory is used to design actuators

56 Progresses on Green and Smart Materials for Multifaceted Applications

Thermoelectric Material

Shape Memory Alloy

Magnetostrictive Material

Piezoelectric Material

1441

Rheological Fluid

Smart Materials

Others

Fig. 1 Different types of smart materials (Applications of Smart Materials 2020)

where the element protrudes based on a magnetic field’s existence. Temperatureresponsive polymers are polymers that show a severe and disjointed change of their physical properties with temperature. It is mainly used when the property is soluble in a given solvent, and other properties are affected. Halochromic material is a material that changes color when pH alteration occurs (Rosace et al. 2017; Zakirullin and Odenbakh 2020; Goldan and Nistor 2019; Zhu et al. 2016). Different types of innovative materials are shown in Fig. 1, and the applications of smart material to transportation are shown in Fig. 2. The term “green materials” refers to the materials that enhance the environment during processing, consumption, or disposal by conserving resources and minimizing the use of toxic agents, pollution, and waste. In other words, green materials offer potential benefits to the environment and human health. Hence, both researchers and environmentally conscious consumers have shown an enhanced interest in green materials. Available literature has shown that relevant research has been carried out on reviews for green product innovation, development and process innovation, sustainable product innovation, eco-innovation, and sustainabilityoriented innovation (Dangelico 2016; Tariq et al. 2017; Pereira and Vence 2012; Adams et al. 2016). The trends of research and development in green materials– related areas have spread since 1964. It started with two publications in 1964, 27 publications in 2005, 23 publications in 2006, and 185 publications in 2019. Figure 3 shows the trends of research publications and citations in the field of green materials (Bhardwaj et al. 2020).

1442

S. O. Oyedepo et al.

Fig. 2 Smart material and application in transport industry (Applications of Smart Materials 2020)

Fig. 3 Trends of research publications and citations in the field of green materials (Bhardwaj et al. 2020)

56 Progresses on Green and Smart Materials for Multifaceted Applications

1443

Multifunctional Applications in Energy, Construction, Infrastructure, Electronics, and Buildings There is a need to use new construction systems and green materials to increase construction, improve service life, and protect buildings against earthquakes. Mahmoudian (Mahmoudian and Sharifikheirabadi 2019) concluded that the best approach to minimize energy consumption in buildings involves using green materials and techniques. The green materials help to reduce energy waste in construction, optimize the present materials and products, minimize the application of raw materials and energy, reduce manufacturing steps and maintenance cost, improve construction speed, in effective use of materials, conservation and preservation of natural resources, and recycling and plasticity (Mahmoudian and Sharifikheirabadi 2019). Al-Baldawi (Mohamed 2017) researched developing and using smart materials to improve the product’s economic, functional, technical, and aesthetic view. The author stated that renewable materials, such as bamboo and wool, could substitute within a decade of the growing forestry cycle. Products made from rapidly renewable sources gain broader interior applications, giving the environmental importance of reducing nonrenewable materials’ demands (Mohamed 2017). The higher chances of bringing nanotechnology’s development caused an increased number of new and arrived innovative architectural results. Engineered cementitious composites (ECCs) are a class of smart materials in a block of concrete with high-performance properties and smart multifunctionalities designed to retain the potential to satisfy the assumed civil infrastructure needs of the twenty-first century (Li 2019). Smart materials are green and are required for numerous high-performance modern concrete materials. High-performance concrete helps in infrastructure sustainability and obstructs hazardous failure triggered by a brittle failure of the material (Li 2019). The principle of green architecture and smart materials involve conservation of energy, working with the climate to achieve eco-friendliness, and reducing the use of new resources. Structures are now embedded with smart technologies to ensure safety, ease of lives, and property. Figures 4 and 5 present buildings that employ smart technology.

The Concept of Smart Materials, Traditional Materials, Modern Materials, Nanomaterials, and Composite Materials Advancements in science and technology lead to changes in material selection. There are ranges of modern materials with remarkable properties, also traditional materials such as wood and metal (Wood 2008). Traditional materials exist from primitive age, such as wood, stone, paper, and metals. On the other hand, modern materials are traditional improvised materials with better properties due to artificial improvization such as reinforced concrete, aluminium, and steel (Mostafaei et al.

1444

S. O. Oyedepo et al.

Fig. 4 Smart material in structural engineering (Applications of Smart Materials 2020)

2021). More discoveries of materials have modified the mindset of researchers and manufacturers of products. For instance, graphene, a modern material, is a single carbon layer material, which is supposedly stronger than steel by multiples of 100. It is invisible and quasi-fluffy and is acclaimed to protect the body from bullets (Rudrapati 2020). Titanium is another modern material used in the sporting and medical fields. It has successfully been used for the replacement of hip joints and advanced bicycles as it possesses high strength to weight ratio and is corrosion resistant. Metal foams are lightweight and robust materials produced by injecting a gas or foaming agent into molten metal. They are used in planes and cars as they absorb shock efficiently during crashes (Hu and Yoon 2018). Smart materials tend to have an additional “smart” function embedded within the material. An example would be a conductive component that allowed the product to be heated, or perhaps allowed an input, or could include an integrated antenna. Smart materials can be appropriately called “reactive materials” due to their response to external influence, such as electric and magnetic fields, stress, moisture and temperature, light, pressure, voltage, pH, or chemical compounds (Esther et al. 2014). This change is reversible and can be repeated many times. Smart materials will help reduce weight, component size, and complexity while improving design flexibility, functionality, reproducibility, and reliability – the benefit spans aerospace, medical, textile, construction, and electronics industries (Wiklund et al. 2021).

56 Progresses on Green and Smart Materials for Multifaceted Applications

1445

Fig. 5 (a) Smart building (b) Smart building with smart tools (Applications of Smart Materials 2020)

Smart materials involve composite materials embedded with fiber optics, actuators, sensors, microelectromechanical systems (MEMSs), vibration control, sound control, shape control, product health or lifetime monitoring, cure monitoring, and intelligent processing. For the future, for autonomous systems, one can segment the market into several areas such as implant material, sensors, and actuators,

1446

S. O. Oyedepo et al.

structural materials, shape memory, smart fluid, alloys, etc. (Di Rito et al. 2020). Using a smart material instead of conventional methods to sense and respond can simplify devices, reducing weight and the chance of failure. Smart materials respond predictably to changes in their environment (Addington and Schodek 2012). Smart materials have several functions, such as embedded sensors or actuators, and support the structural weight. A typical example is Nitinol. After mechanical deformation, it can be heated up and return to the initial structural shape. Leadzirconate-titanate (PZT) is a ceramic that reacts to mechanical deformation by causing an electrical potential. On the other hand, an applied electrical potential result in an expansion of the material. Many categories of smart materials exist, each displaying unique properties that can be utilized in various high-tech and everyday applications (Akhras 2000). These include shape memory alloys, piezoelectric materials, magnetostrictive materials, and chromic materials. Piezoelectric materials generate voltage when stress is applied. Deformation occurs in shape memory alloys by varying the temperature or stress. Magnetostrictive materials also exhibit shape changes at the function of magnetic field and mechanical stress (Dineva et al. 2014). Composite materials require at least two different materials to be used in a composite material, which allows the engineering of desired properties such as strength, thermal conductivity, stiffness, etc. An example is glass-reinforced plastic (GRP). Generally, this is done where the combination of materials has a synergistic benefit (e.g., more significant benefit than the sum of the parts, e.g., glass cloth will flex, resins are not particularly strong in tension, but together the composite makes a strong, stiff part) (Haruna et al. 2014). Nanomaterials imply a material that has been modeled at the nanoscale, which includes developing a nanosized particle to produce composites, building surfaces, thin films coatings such as the oleophobic coatings on smartphone screens that repel greasy fingerprints, or hydrophobic materials that repel water, etc. (Cunha and Gandini 2010). Nanomaterials also refer to a part or item of nano-dimensions, this is less than micro(meter) size or < 0.001 mm, such as putting carbon nanotubes or platelets into a resin system to improve its mechanical strength. Also, such a resin system could be used in a composite structure, and since it is conductive, a smart feature can be added such as a capacitance touch input. The product, therefore, has multiple functions. An example might be a tennis racket that allows you to control a smart device such as the volume or rejection of calls (Momeni 2018). Four-dimensional printing technology applies smart materials such as single shape memory polymers, liquid crystal elastomers, composite hydrogel, composites, multimaterial, and other multifaceted material due to their physical, mechanical, thermal, and microstructural properties. They can adapt their shapes over external stimuli. Any 4D printed object will react and change size or shape when external stimuli such as temperature, water, etc. come in contact with it. The benefit of 4D printing is that objects that are bigger than the size of the printer can be designed efficiently, can enhance printed products, and have better manufacturing efficiency, lower production cost, and reduced climate footprint (Momeni 2018). Smart material for 4D printing is one of the trending research fields where different materials are blended and deformed in their reactions to different external stimuli. Presently, researchers are using 4D selective

56 Progresses on Green and Smart Materials for Multifaceted Applications

1447

Fig. 6 The flow chart of the smart materials in achieving 4D printing (Bajpai et al. 2020)

laser melting methods, direct inkjet cure, fused deposition modeling, stereography, and laser-assisted bioprinting achievable with the knowledge of mathematical modeling. Types of smart materials are electroresponsive polymers, shape memory polymers, smart inorganic polymers, temperature-responsive polymers, magnetic shape memory alloys, electroactive polymers, memory alloys, and photoresponsive polymers (Haleem et al. 2021). Smart and programmable materials are employed in 4D printing technologies to bring up different features when heat is applied. Thermal composite hydrogel and bilayer SMP composites are combined with polymeric photo materials to achieve 4D printing. The process of achieving 4D printing is first applying smart materials and dimensions, which allows for complex creations and additive processing of such time-dependent materials with an explicit reaction to external stimuli for a calculated time. The flow chart in achieving 4D printing is shown in Fig. 6 (Bajpai et al. 2020). Smart materials in combination with multimaterial composites have been employed to produce new materials with a desirable performance by imbibing nanoparticles such as graphene, carbon nanotubes, and biomaterials and combining them with 3D systems leading to 4D printing (Ahmed et al. 2020). Smart materials are increasingly used for 4D bioprinting leading to the emergence of medicine 4.0 aside from Industry 4.0. At the moment, smart materials such as shape memory polymers can be applied for 3D printing tailored organs that change shapes appropriately when added into a human body called smart implantations which are fit when there is a dearth of medical implants more especially during emergency surgeries. (Ashima et al. 2021) This multifaceted path is demonstrated in the stateof-the-art use of smart materials, which combine physical and applied sciences, as well as electrical and mechanical engineering. The synergy of knowledge across disciplines enables the development of devices and products that are smart, user-friendly, and eco-friendly. Smart materials are responsible for the evolution of devices, operations, and cities that are wired with smart Internet and sustainable. Therefore systems, products, and even cities are circuited with communication devices. Thus, smart materials endorse the Internet of a Thing (IoT) as the whole globe and people can be connected, and communications or informed decisions can be made easily (Gandhi and Thompson 1992). The smart materials research field seeks to advance and apply new and distinctive materials with better performance, such as self-cleaning and self-healing characteristics,

1448

S. O. Oyedepo et al.

photomechanical materials, shape-memory materials, and electroactive magnetoactive materials. These materials are applied to smart structures to improve our health and habits solving socioeconomic and ecological challenges. The production of smart materials provides eco-friendly products. It has advantages over mass production approaches such as low cost, material utilization, excellent pattern, adaptability to ambient temperatures, and difficult products. Smart materials can also be called smart behavior of a material that responds to external stimuli in a precise, consistent, and suitable way. Smart technology has been used to improve electrical products such as actuators and sensors and a considerable collection of products, spanning from domestic goods to innovative automotive parts and medical devices, sports, entertainment, security, consumer electronics, banking, and food. Examples of smart materials are not limited to piezoelectric materials, electrostrictive materials, electrostrictive polymers, piezoresistive materials, pHsensitive or chromic systems materials, and lignocellulosic biomass (Oliveira et al. 2018). Smart materials find their application in power harvesting that combines the operation of harvesters, the structure of piezoelectric materials, processing, and technological application of smart materials to achieve a power harvester (Batra and Alomari 2017). Gels are examples of smart materials which are effective for their prompt response to chemical change and physical activations. It is a novel material that can be altered to actual fabrication of different structures for applications in medicine, electronics, pattern making, solar cells, automobiles, etc. Peptides are further examples of smart materials with massive capacities (Fairman and Åkerfeldt 2005). Smart materials technologies are important to the twenty-first century where competition is high, and the better technology overrides the lesser since it offers a better solution to the challenges of humans. The introduction of smart technology has enabled numerous materials to increase in operation and durability. Smart materials thus performs a major function in building technology; as the materials are adaptable to the environment due to high sensibility becoming sensitive structures, so smart materials in this sense can be called living buildings because as humans are sensitive to stimuli, these materials are capable of performing such roles due to embedded elements. This technology phases out the traditional methods and combines both modern and composite materials with the inclusion of sensitive elements to achieve a smart composite material. This becomes the future direction as it solves challenges faced with the climate and humans (Addington and Schodek 2005). Smart material is applicable in everyday life as every section wants a smarter way of solving challenges. One example is the development of smart technology for communities around fishery ports to achieve a quasi-null carbon footprint. Modern researches endeavor to solve the challenges of material selection, carbon footprint, noxious emission, climate change, pollution, improved air quality, and energy demand. The world organizations aim at reducing these challenges by the application of renewable energy and smart materials. Efforts are continuously being made by standard organizations to adopt renewable energy and eco-friendly materials. Smart technologies come into play when communities are interconnected, and devices can sense the environment, such as light emission. Smart technology has the advantage of conserving energy by reducing redundant or indiscriminate

56 Progresses on Green and Smart Materials for Multifaceted Applications

1449

energy use. The outcome is also that energy loss will be minimized, operation cost lowered, and the power grid efficiency can be improved (Alzahrani et al. 2020).

Smart Material and Waste Management Smart material also solves the challenge of scarcity of resources as it supports agile manufacturing. The geometry and dimensions of materials are determined from the start to avoid wastage of material. The material is prudently 3D/4D printed, and all necessary components are embedded in the product to make it an intelligent material, thus saving energy, time, and minimizing waste. Another advantage is achieving zero waste and optimizing the environment for other valuable purposes than having landfills. According to María-Laura “Zero Waste has a goal that is ethical, economical, efficient and visionary, to guide people in changing their lifestyles and practices to emulate sustainable natural cycles, where all discarded materials are designed to become resources for others to use.” According to (Abi-Akar et al. 2017), a perceived nonrecyclable material is seen as a valuable material to make a sustainable product that supports the goal of sustainable material management (Abi-Akar et al. 2017). Population upsurge and economic activities have led to increased tonnage of waste and, if not utilized, will result in pollution of the environment and water bodies. A further challenge to urban settlement is that the waste occupies the scarce land meant for human consumption in comfort, agriculture, and housing. The economic impact is a health threat due to compacted buildings and other social vices. It is been projected that if landfills in the urban areas are gathered, their area will be equivalent to a city like Paris. To achieve a zero-landfill goal is to recycle waste by reusing materials such as plastics, metals, glass, etc., imbibing waste to wealth initiative and transforming the waste into viable products. Some of the materials can be utilized as smart materials when dutifully sorted out. Energy can as well be recovered from this waste which is sustainable and powers a smart city or intelligent device (Bourtsalas 2019).

Smart Materials for the Industry and Smart Industrial Equipment Sustainability ensures continuity of a process and not just a transient goal. The next generation can benefit maximally from the previous generation due to the viable product bequeathed to the latter. Animals’ extinction, depletion of the ozone layer, pollution upsurge, deforestation, squandering of the earth’s resources, and other human activities that had adversely affected the ecosystem and eco-chain are all consequences of not factoring sustainability in manufacturing activities. While countries are keen on boosting the economic returns on investment, nature suffers due to indiscriminate exploration (Kuhlman and Farrington 2010). When nature’s patience is exceeded, there will be consequences such as global warming; climate change; rise in sea level due to melting caused by global temperature rise;

1450

S. O. Oyedepo et al.

increment in greenhouse gases (GHG) such as methane, chlorofluorocarbons, and carbon (IV) oxide; extinction of plant and animal species as a result of deforestation and pollution; baldness of the ozone layers due to untreated industrial effluent chemicals released to the atmosphere or hydrosphere; and food chain pollution with noxious metals such as mercury, lead, soil acidity, soil erosion, etc. These escalating threats to nature are also challenges that the next generation of solutions that are not proffered can also be inherited (Zhang et al. 2020). An insightful solution to these menaces is for humanity to embrace recycling waste and minimize the exploration of resources from the earth. The development of quasi-natural materials will also reduce waste and pollution of the atmosphere as materials are almost plantlike in composition. Waste minimization is very crucial to replenishing nature and sustaining the ecosystem. Sustainability is linked with the attitude posed by the society on the environment in boosting the economy. The advancement of the course for sustainability hinges on the community, manufacturer, and all stakeholders to utilize resources to maintain or improve nature which in turn will increase public health and better working conditions (The World Bank 2018). The concept of waste management or its utilization can be further explored. Waste was defined as a byproduct or a useless product from the main product that can disintegrate the strength of nature if not reused.

Concept of Industry The industry is the aspect of an economy that showcases the level of development of a nation. Technology, no doubt, serves as a benchmark to assess the outlook of industrialization. The first industrial revolution was dominated by mechanization, while the second industrial revolution ushered in the intensive use of electrical energy. The introduction of computer technology and communications birthed the third industrial revolution. The current shift from the previous paradigm allows smart technology in products to achieve a remarkable material product property. The concept of Industry 4.0 is imbibed by multidiscipline to accomplish a creative output more closely in building technology to achieve a desirable building product property (Bajpai et al. 2020). Sustainable manufacturing blends with the fourth stage of industrialization to solve the growing quest for a durable and eco-friendly world. Value infuses into products, considering personal satisfaction (Bajpai et al. 2020). The evolution of Industry 4.0 becomes a culture in manufacturers to consider ergonomics and sustainability by factoring these when developing products for end users (Sima et al. 2020). Figures 7 and 8 present how industry has displaced traditional applications with smart technology. Three-dimensional printing replaces conventional mold, mobile devices, cloud computing, and Internet of a Thing, among others. This introduction of 3D printing helps a product be viewed in software, modified as requested by the client, and satisfies consumer’s demand before the mass production of the prototype (Rayna and Striukova 2021).

56 Progresses on Green and Smart Materials for Multifaceted Applications

1451

Fig. 7 Impact of technology on industrial production (Bajpai et al. 2020)

Fig. 8 Relationship between Industry 4.0 and conventional manufacturing (Bajpai et al. 2020)

1452

S. O. Oyedepo et al.

Fig. 9 Two-slice high-speed smart toaster

Smart Cooking Appliances We are living in the era of smart technology that now influences everything we do. We are done with traditional appliances and now engage in energy-efficient electronic appliances that sense the environment and give useful feedback to human users. Much of the applications can be seen in our home appliances. (Mattern et al. 2010). Smart cooking appliances are programmed devices that replace most human efforts in the kitchen. Figures 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, and 22 present smart cooking appliances that have reduced human effort drastically. Examples of these appliances abound, such as:

Two-Slice High-Speed Smart Toaster A two-slice high-speed smart toaster is shown in Fig. 9. It is a high-profile touch screen electronic device that delivers the toast assorted bread without any burnt bread. The muncher can program the numbers of sliced bread and the texture of the food. The food is ready within 15 seconds, which replaces much time wasted by the traditional method. Instant Pot Smart Wi-Fi 8-in-1 Pressure Cooker Fig. 10 shows an instatnt pot smart Wi-Fi 8-in-1 pressure cooker . It is a fantastic smart device that can cook your recipe when you are away from home by clicking the app on your phone, and while at home, you can be engaged in other activities while the device does your job. So there is more time for gym, office, and relaxation at home. It gives flexibility, ease of stress, safety, and comfort to the user (Elias 2009).

56 Progresses on Green and Smart Materials for Multifaceted Applications

1453

Fig. 10 Instant pot smart Wi-Fi 8-in-1 pressure cooker

Fig. 11 Sensate touchless kitchen faucet with voice-activated technology

Sensate Touchless Kitchen Faucet with Voice-Activated Technology This device shown in Fig. 11 and is another evidence of breakthrough of technology such that after a successful installation, programming, and calibration of the sensate faucet with the phone device, water can be dispensed without you touching the device as it can recognize the voice command and fill the number of cups of your

1454

S. O. Oyedepo et al.

Fig. 12 Automatic pan stirrer with timer

choice. The device’s design corresponds to the COVID 19 goal of “no touch” on objects which become safe for users.

Automatic Pan Stirrer with Timer Figure 12 shows an automatic pan stirre with timer . This device fits every pan with three legs with a battery-enabled electric motor. The speed can be regulated, and the device will give your hand rest from stirring activity on the pot. Works Smart Coffee Maker with Alexa Works smart coffee maker with Alexa is a 12-cup capacity programmable device built with corrosion-resistant stainless steel material. It is built with an intelligent speaker and Wi-Fi sensor to enable remote operation by installing the Alexa app on a smartphone device. With this, your coffee will be ready at your predetermined time. Many advantages can be garnered, such as stress relief, time saving, reliability, functionality, no struggle to prepare the coffee, and simple operation. Figure 13 shows works smart coffee maker with Alexa. Smart Wi-Fi Air Fryer Smart Wi-Fi air fryer (Fig. 14) is a twenty-first-century smart kitchen device that can cook chips or recipes up to 85% more than traditional fried food with highly reduced calories. The recipe can be downloaded and cooking scheduled using the smartphone. Then the user can relax and pick the already prepared food. You can snap the food to check the cooking history. This is a fantastic contribution of intelligent technology through artificial intelligence (AI). Food can be 3D printed

56 Progresses on Green and Smart Materials for Multifaceted Applications

1455

Fig. 13 Works smart coffee maker with Alexa

with the user setting the desired food product on the smartphone without much interaction with the cooking device (St˛epie´n 2020).

Smart Wireless Meat Thermometer Figure 15 shows a smart wireless meat thermometer . This is a simple design with a probe for checking the temperature of meat during steaming. This can be called a “meater” with an inbuilt rechargeable battery which can last long during operation. The device can be connected to a smartphone to monitor the temperature during the cooking operation. Touchless Automatic Soap Dispenser Figure 16 shows a touchless automatic soap dispener . The device is a simple battery-enabled device with a stand that can release soap, hand sanitizer, and liquid products without touch. It is calibrated to control the volume of soap per unit time. It can be used everywhere to minimize germs spread through touch. Further work here is to make it fully automatic through interaction with a smartphone.

1456

S. O. Oyedepo et al.

Fig. 14 Smart Wi-Fi air fryer

Fig. 15 Smart wireless meat thermometer

Balance Smart Blender Balance smart blender (Fig. 17) is a Bluetooth-enabled blender for smoothies and vegetables. All your vegetable recipes can be selected online and then programmed through the enabled application for the blender. The blender can be operated to start

56 Progresses on Green and Smart Materials for Multifaceted Applications Fig. 16 Touchless automatic soap dispenser

Fig. 17 Balance smart blender

1457

1458

S. O. Oyedepo et al.

Fig. 18 Brava mini oven

Fig. 19 Cordless electric kettle

remotely using the phone device, and the desired smoothie is ready for consumption or vegetables ready to prepare the stew. The application can predict the nutritional status of the vegetables, thus improving the health status of consumers and reducing medical emergencies. This is made possible as the nutritional value of all added vegetables is calculated by applying the nutribullet balance app.

56 Progresses on Green and Smart Materials for Multifaceted Applications

1459

Fig. 20 Anova precision cooker

Other innovative housewares are: Brava mini oven that serves as an alternative to the toaster oven because it can heat up to 500 ◦ C within seconds and with the advantage of Wi-Fi access to the heated recipe.

Cordless Electric Kettle Figure 19 shows a cordless electric kettle. This device is an electric kettle made of corrosion-resistant stainless steel, accommodating up to 1.7 liters of water. It has a programmable memory setting to regulate the temperature and time for the tea to be ready. Anova Precision Cooker Anova precision cooker (Fig. 20) is designed to operate when the user is either at home and home away preparing the food favorites to perfection. It is Bluetooth enabled to monitor the cooking progress to control the temperature at a steady or

1460

S. O. Oyedepo et al.

Fig. 21 Vacuum sealer

Fig. 22 Temperature control smart mug

unsteady state. It solves the challenge of being new to a recipe as it can be browsed on the Internet and prepared through setting the app.

Vacuum Sealer Vacuum sealer (Fig. 21) is another homemade device to package food by sealing it by evacuating air. The seal can be refrigerated after the operation, thereby improving the shelf life of the food product. Temperature Control Smart Mug Temperature control smart mug (Fig. 22) is a specially designed ceramic mug with extraordinary heating power for liquid poured in it. The success of the operation

56 Progresses on Green and Smart Materials for Multifaceted Applications

1461

lies in the saucer charging base using a rechargeable battery. The mug has a WiFi-enabled sensor that detects when liquid enters it, and the charging base regulates the liquid to preset desired temperature. The temperature can be regulated using an app-enabled smartphone.

Application of Smart Materials in Energy and Electronics Smart technology influences our digital and communication world. Technology determines the advancement outlook of society. Tools engaged in solving societal needs decades ago are no longer employed due to a better approach. A society that employs traditional tools shows its level of crudeness, unlike a society with advanced technology. Our electronic products were designed to precision as a function of smart material selection such as sensors in automobiles, security ˇ cameras, and public services (Cukuši´ c et al. 2019). Better materials to prevent building collapse in infrastructure and enable low energy consumption are now developed and employed in our buildings. Advanced energy-saving bulbs are now seen in our homes (Farag 2019; Konarzewska 2017b). Figures 23 and 24 show the application of smart technology to our energy system. Lightweight efficient utensils, home apparatus, and low-energy appliances are replaced with traditional devices. Our refrigeration systems now employ nanorefrigerants that are well researched to curb environmental pollution and improve the exergy of the refrigeration cycle with improved refrigerating units. Hybrid automobiles are born to ensure a green environment devoid of carbon emissions and pollutants (Krishna and Thirumal 2015; Kochovski and Stankovski 2018). Smart material has also made achievements in the biomedical community. Shape memory polymer (SMP) is a major breakthrough in medicine in solving challenges

Fig. 23 Smart energy system (Krishna and Thirumal 2015)

1462

S. O. Oyedepo et al.

Fig. 24 (a) 3D filament (Hu and Yoon 2018) (b) Structured electronic car (Hu and Yoon 2018)

such as wounds, cancer, and treating infection according to (Mohamed 2017). SMP changes its shape by adapting to the environment, the altering of the shape can be indefinite based on environmental factors, and then returns to the original shape when triggered to do so. SMP has functioned successfully in surgical operation and in stitching of wounds. The technology is also applied to medical implants and ease of urination for urinal patients. The shape change is triggered by the body temperature of the patient (Roy et al. 2016; Li and Su 2018; Nguyen et al. 2018; Fraser et al. 2018; Cruz et al. 2020).

Application of Smart Materials in Building and Construction The technological advancement that encourages the sustainability and application of renewable resources in the building and construction industry is a significant breakthrough in the emerging population for sustainable buildings (Salvarli and Salvarli 2020). The arithmetic progression of the teeming population draws a solid resolution for climatic alterations, and requesting for gradual changes is the structuring path of a sustainable future for the next generations. Porotherm bricks are produced from a blend of clay and other green products such as rice husk, ash, and sawdust, giving a long-lasting renewable material that can be applied in construction. Construction companies used these bricks as construction materials due to their lightweight properties. Other properties include small water intake, fire retardant, and thermal insulation, which improve their usage in the construction sector. The use of waste products to manufacture fly ash brick obstructs the release of a harmful substance into the environment, thereby helping to generate renewable construction materials. The bricks are preferred over clay as a green building material that can absorb little heat and water, attracting future investors (Karslio˘glu et al. 2021). The energy performance of these bricks is gaining more recognition in the market apart from being light in weight. Construction companies utilized waste materials with recycled products such as mining wastes, glass wastes, and burnt clay to produce renewable materials for construction. The produced green concrete reduced the maintenance rate and duration of renewable construction material (Safiuddin et al. 2010). Manufacturers developed these concretes by taking

56 Progresses on Green and Smart Materials for Multifaceted Applications

1463

notes on the construction life cycle, comprising structural design parameters, as well as manufacturing methods and maintenance patterns. The advantage of using this concrete cannot be overemphasized since it discharges no harmful emissions and reduces CO2 into the environment (Yüksek and Karadayi 2017).

Conclusions This chapter demonstrates the importance of using the high-execution properties of the material with smart features to improve infrastructure’s long-lasting resilience and viability. The application of smart material has improved all facets of the industry by providing users comfort and ease of use. Smart materials in construction help preserve the energy and optimize the use of amenities and reverse initial capital to reduce energy consumption. The chapter observed that smart material in green buildings reduces costs more significantly than construction costs. It also minimizes the negative environmental impacts and the attitude of society to sustainable performance. The generation present and next depend on smart technologies to reduce human efforts to the barest minimum and save redundant energy for alternative utilization. Smart equipment will now build efficient structures and products, thereby saving materials from wasting and eliminating downtime. Smart materials have proven a dependable technology that applies to all fields of human endeavors. A devoted exploration by researchers and developers will produce a geometric efficiency of a product more than the traditional methods.

References Abi-Akar R, Jones G, Tang Y (2017) Opportunities for sustainable materials management and zero waste in detroit Adams R, Jeanrenaud S, Bessant J, Denyer D, Overy P (2016) Sustainability-oriented innovation: a systematic review. Int J Manag Rev 18:180–205 Addington DM, Schodek DL (2005) Smart materials and new technologies: for the architecture and design professions. Routledge Addington M, Schodek D (2012) Smart materials and technologies in architecture. Routledge Ahmed K, Shiblee MNI, Khosla A, Nagahara L, Thundat T, Furukawa H (2020) Recent progresses in 4D printing of gel materials. J Electrochem Soc 167(3):037563 Akhras G (2000) Smart materials and smart systems for the future. Can Military J 1(3):25–31 Alzahrani A, Petri I, Rezgui Y, Ghoroghi A (2020) Developing smart energy communities around fishery ports: toward zero-carbon fishery ports. Energies 13(11):2779 Applications of Smart Materials. Date Accessed October, 21, 2020 Ashima R, Haleem A, Bahl S, Javaid M, Mahla SK, Singh S (2021) Automation and manufacturing of smart materials in additive manufacturing technologies using internet of things towards the adoption of industry 4.0. Mater Today: Proc 45:5081–5088 Bajpai A, Baigent A, Raghav S, Brádaigh CÓ, Koutsos V, Radacsi N (2020) 4D printing: materials, technologies, and future applications in the biomedical field. Sustainability 12(24):10628 Basheer AA (2020) Advances in the smart materials applications in the aerospace industries. Aircr Eng Aerosp Technol 92(7):1027–1035 Batra AK, Alomari A (2017) Power harvesting via smart materials. SPIE Press, pp 1–15

1464

S. O. Oyedepo et al.

Bhardwaj AK, Garg A, Ram S, Gajpal Y, Zheng C (2020) Research trends in green product for environment: a bibliometric perspective. Int J Environ Res Public Health 17(8469):1–21 Boström M, Andersson E, Berg M, Gustafsson K, Gustavsson E, Hysing E et al (2018) Conditions for transformative learning for sustainable development: a theoretical review and approach. Sustainability 10(12):4479 Bourtsalas A (2019) In: Themelis NJ (ed) Recovery of materials and energy from urban wastes: a volume in the encyclopedia of sustainability science and technology. Springer, New York Cruz DM, Mostafavi E, Vernet-Crua A, Barabadi H, Shah V, Cholula-Díaz JL et al (2020) Green nanotechnology-based zinc oxide (ZnO) nanomaterials for biomedical applications: a review. J Phys: Mater 3(3):034005 ˇ Cukuši´ c M, Jadri´c M, Mijaˇc T (2019) Identifying challenges and priorities for developing smart city initiatives and applications. Croat Oper Res Rev 10(1):117–129 Cunha AG, Gandini A (2010) Turning polysaccharides into hydrophobic materials: a critical review. Part 1. Cellulose. Cellulose 17(5):875–889 Dangelico RM (2016) Green product innovation: where we are and where we are going. Bus Strategy Environ 25:560–576 Di Rito G, Chiarelli MR, Luciano B (2020) Dynamic modelling and experimental characterization of a self-powered structural health-monitoring system with M.F.C. piezoelectric patches. Sensors 20(4):950 Dineva P, Gross D, Müller R, Rangelov T (2014) Piezoelectric materials. In: Dynamic fracture of piezoelectric materials (pp. 7–32). Springer, Cham Elias MJ (2009) Food in the United States, 1890–1945. ABC-CLIO Esther L, Piselli A, Faucheu J, Delafosse D, Del Curto B (2014) Smart materials: development of new sensory experiences through stimuli responsive materials. In: 5th S.T.S. Italia conference a matter of design: making society through science and technology (pp 367–382). S.T.S. Italia Fairman R, Åkerfeldt KS (2005) Peptides as novel smart materials. Curr Opin Struct Biol 15(4):453–463 Farag SG (2019) Application of smart structural systems for smart sustainable cities. In: 2019 4th M.E.C. International Conference on Big Data and Smart City (ICBDSC) (pp 1–5) Fraser AG, Butchart EG, Szyma´nski P, Caiani EG, Crosby S, Kearney P, Van de Werf F (2018) The need for transparency of clinical evidence for medical devices in Europe. Lancet 392(10146):521–530 Gandhi MV, Thompson BD (1992) Smart materials and structures. Springer Science & Business Media Goldan T, Nistor MC (2019) Reducing self-heating coal stockpile for prevention of environmental hazard. Int Multidiscip Sci GeoConference: SGEM 19(1.3):513–519 Haleem A, Javaid M, Singh RP, Suman R (2021) Significant roles of 4D printing using smart materials in the field of manufacturing. Adv Ind Eng Polymer Res Haruna VN, Abdulrahman AS, Zubairu PT, Isezuo LO, Abdulrahman MA, Onuoha DC (2014) Prospects and challenges of composites in a developing country. ARPN J Eng Appl Sci 9(7):1069–1075 Hu CY, Yoon TR (2018) Recent updates for biomaterials used in total hip arthroplasty. Biomaterials Res 22(1):1–12 Karslio˘glu A, Balaban E, Onur MI˙ (2021) Insulation properties of bricks with waste rubber and plastic: a review. J Nat 1:20–27 Kochovski P, Stankovski V (2018) Supporting smart construction with dependable edge computing infrastructures and applications. Autom Constr 85:182–192 Konarzewska B (2017a) Smart materials in architecture: useful tools with practical applications or fascinating inventions for experimental design. In: I.O.P. Conf series: materials science and engineering, 245, 052098 Konarzewska B (2017b) Smart materials in architecture: useful tools with practical applications or fascinating inventions for experimental design. In: I.O.P. Conf series: materials science and engineering (Vol. 245, p. 052098) Krishna JG, Thirumal JR (2015) Application of smart materials in smart structures. Int J Innov Res Sci Eng Technol 4(7)

56 Progresses on Green and Smart Materials for Multifaceted Applications

1465

Kuhlman T, Farrington J (2010) What is sustainability? Sustainability 2(11):3436–3448 Li VC (2019) High-performance and multifunctional cement-based composite material. Engineering 5:250–260 Li X, Su X (2018) Multifunctional smart hydrogels: potential in tissue engineering and cancer therapy. J Mater Chem B 6(29):4714–4730 Li X, Shang J, Wang Z (2017) Intelligent materials: a review of applications in 4D printing. Assem Autom Mahmoudian M, Sharifikheirabadi P (2019) Uses of new/smart materials in the green building with sustainability concerns. Int Trans J Eng Manag Appl Sci Technol:1–9 Mattern F, Staake T, Weiss M (2010) I.C.T. for green: how computers can help us to conserve energy. In: Proceedings of the 1st international conference on energy-efficient computing and networking (pp 1–10) Mayeen A, Kalarikkal N (2018) Development of ceramic-controlled piezoelectric devices for biomedical. Fundamental Biomaterials: Ceram 47 Mensah J, Casadevall SR (2019) Sustainable development: meaning, history, principles, pillars, and implications for human action: literature review. Cogent Soc Sci 5(1):1653531 Mohamed ASY (2017) Smart materials innovative technologies in architecture; towards innovative design paradigm. Energy Procedia 115:139–154 Momeni F (2018) 4D Printing as a new paradigm for advanced manufacturing (Doctoral dissertation) Mostafaei A, Elliott AM, Barnes JE, Li F, Tan W, Cramer CL et al (2021) Binder jet 3D printing— process parameters, materials, properties, modeling, and challenges. Prog Mater Sci 119:100707 Müller P, Schmid M (2019) Intelligent packaging in the food sector: a brief overview. Foods 8(1):16 Nguyen PQ, Courchesne NMD, Duraj-Thatte A, Praveschotinunt P, Joshi NS (2018) Engineered living materials: prospects and challenges for using biological systems to direct the assembly of smart materials. Adv Mater 30(19):1704847 Oliveira J, Correia V, Castro H, Martins P, Lanceros-Mendez S (2018) Polymer-based smart materials by printing technologies: improving application and integration. Addit Manuf 21: 269–283 Pereira Á, Vence X (2012) Key business factors for eco-innovation: an overview of recent firmlevel empirical studies. Cuad Gest 12:73–103 Rayna T, Striukova L (2021) Assessing the effect of 3D printing technologies on entrepreneurship: an exploratory study. Technol Forecast Soc Chang 164:120483 Rosace G, Guido E, Colleoni C, Brucale M, Piperopoulos E, Milone C, Plutino MR (2017) Halochromic resorufin-GPTMS hybrid sol-gel: chemical-physical properties and use as pH sensor fabric coating. Sensors Actuators B Chem 241:85–95 Roy S, Mishra H, Mohapatro BG (2016) Creating sustainable environment using smart materials in smart structures. Indian J Sci Technol 9(30) Rudrapati R (2020) Graphene: fabrication methods, properties, and applications in modern industries. In: Graphene production and application. IntechOpen, Rijeka, pp 9–22 Safiuddin M, Jumaat MZ, Salam MA, Islam MS, Hashim R (2010) Utilization of solid wastes in construction materials. Int J Phys Sci 5(13):1952–1963 Salvarli MS, Salvarli H (2020) For sustainable development: future trends in renewable energy and enabling technologies. In: Renewable energy-resources, challenges and applications. IntechOpen Sima V, Gheorghe IG, Subi´c J, Nancu D (2020) Influences of the industry 4.0 revolution on the human capital development and consumer behavior: a systematic review. Sustainability 12(10):4035 St˛epie´n GJ (2020) Internet of things (I.O.T.): smart kitchen appliences for the U.S. market (Doctoral dissertation) Tariq A, Badir YF, Tariq W, Bhutta US (2017) Drivers and consequences of green product and process innovation: a systematic review, conceptual framework, and future outlook. Technol Soc 51:8–23 The World Bank (2018) What a waste: an updated look into the future of solid waste management. World Bank

1466

S. O. Oyedepo et al.

Wiklund J, Karakoç A, Palko T, Yi˘gitler H, Ruttik K, Jäntti R, Paltakari J (2021) A review on printed electronics: fabrication methods, inks, substrates, applications and environmental impacts. J Manuf Mater Process 5(3):89 Wood J (2008) The top ten advances in materials science. Mater Today 11(1–2):40–45 Yüksek I, Karadayi TT (2017) Energy-efficient building design in the context of building life cycle. Energy Efficient Build:93–123 Zakirullin R, Odenbakh I (2020) Chromogenic materials in optical filters for smart windows. In: Novel optical materials and applications (pp JTu4C-19). Optical Society of America Zhang Y, Beggs PJ, McGushin A, Bambrick H, Trueck S, Hanigan IC et al (2020) The 2020 special report of the M.J.A.–lancet countdown on health and climate change: lessons learnt from Australia’s “black summer”. Med J Aust 213(11):490–492 Zhu FB, Zhang CL, Qian J, Chen WQ (2016) Mechanics of dielectric elastomers: materials, structures, and devices. J Zhejiang Univ-Sci A 17(1):1–21 Zverev VI, Pyatakov AP, Shtil AA, Tishin AM (2018) Novel applications of magnetic materials and technologies for medicine. J Magn Magn Mater 459:182–186

57

Determining Technologies Trends and Evolution of Smart Building Technologies by Bibliometric Analysis from 1984 to 2020 Trends, research opportunities and future perspectives

Nadia Karina Gamboa-Rosales, Luis Daniel López-Robles, Leonardo B. Furstenau, Michele Kremer Sott, Manuel Jesús Cobo, and José Ricardo López-Robles Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Performance Bibliometric Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Publications, Citations, and Impact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1468 1469 1470

N. K. Gamboa-Rosales CONACYT – Autonomous University of Zacatecas, Zacatecas, Mexico e-mail: [email protected] L. D. López-Robles Academic Unit of Accounting and Management, Autonomous University of Zacatecas, Zacatecas, Mexico e-mail: [email protected] L. B. Furstenau Department of Industrial Engineering, Federal University of Rio Grande do Sul, Rio Grande do Sul, Brazil e-mail: [email protected] M. K. Sott Business School, Unisinos University, Porto Alegre, Brazil e-mail: [email protected] M. J. Cobo Department of Computer Science and Engineering, University of Cadiz, Cadiz, Spain e-mail: [email protected] J. R. López-Robles () Academic Unit of Accounting and Management, Autonomous University of Zacatecas, Zacatecas, Mexico Department of Computer Science and Engineering, University of Cadiz, Cadiz, Spain © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_42

1467

1468

N. K. Gamboa-Rosales et al.

Most Productive and Cited Authors, Geographic Distribution of Publications, Organizations, Research Areas, Sources, and Funding Sponsors . . . . . . Most Relevant Publications, Authors, Organizations, and Countries Related to SBT Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conceptual Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis of Content of the Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conceptual Evolution Map of Smart Building Technologies . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1471 1474 1480 1482 1493 1495 1497

Abstract

Today, all industries – especially the building industry – are faced to find environmentally friendly and smart technologies which help us to reduce the human impact on the environment and promote a global sustainable development. These solutions have to consider aspects such as economic feasibility, sustainability, and social equitability. As a result of this need, there is a wide green technologies portfolio that seeks to become the core technology, one of the most representatives being Smart Building Technologies. In this way, the aim of this contribution is to develop a bibliometric analysis that evaluates the evolution of the intellectual and cognitive structure of Smart Building Technologies that supports communities to identify, improve, and reach the research about these technologies. To do that, more than 2,398 publications from 1984 to 2020 related to Smart Building Technologies were retrieved from Web of Science Core Collection and analyzed using advanced bibliometric tools, techniques, and methodologies. Keywords

Smart Building technologies · Industry 4.0 · Sustainable development · Strategic intelligence · SciMAT · Competitive intelligence

Introduction Nowadays, building companies, research centers, and universities are focusing their efforts on the Smart Building Technologies (SBT). Nowadays, building companies, research centers, and universities are focusing their efforts on the Smart Building Technologies (SBT). Bearing in mind the fast development of the IT industry and the technologies that enable it, would be beneficial to use an advanced tools and techniques that allow analyze and visualize in which topics the researchers are focusing on are necessary. Furthermore, to detect and explore the gaps identified in order to develop and advance in this knowledge area. The aim of the present research is to analyze and illustrate the evolution of the Smart Building Technologies research field and its research themes using bibliometric techniques and tools (Glenisson et al. 2005; Moral-Muñoz et al. 2020). To do that, the evolution of the SBT and its performance are evaluated using a

57 Determining Technologies Trends and Evolution of Smart Building. . .

1469

bibliometric analysis software based on a bibliographic network (López-Robles et al. 2018; Gamboa-Rosales et al. 2019; Herrera-Viedma et al. 2020; Furstenau et al. 2020b; Ghansah et al. 2020). In order to recognize the main research themes related to SBT and its evolution, a bibliometric approach is suggested based on the analysis of the scientific map between 1984 and 2019 and its main performance indicators. To this purpose, the documents and their bibliographic information available in Web of Science Core Collection have been compiled – considered as one of the main scientific and academic databases – and subsequently processed with SciMAT (Cobo et al. 2012). The information collected allow the analysis of the main performance indicators: productivity of authors in terms of publications and citations, countries, and journals. This first stage of the analysis has also been evaluated using h-index and H-Classics, thus homogenizing the indicators presented by each of the authors and the main documents (Martínez et al. 2014). In this respect, this manuscript is structured as follows: section “Performance Bibliometric Analysis” describes the performance bibliometric analysis of SBT field. In section “Conceptual Analysis,” conceptual analysis and science mapping are presented and discussed. Finally, section “Conclusions” covers the conclusions, limitations, and future actions.

Performance Bibliometric Analysis This section analyzes the bibliometric performance of Smart Building Technologies (SBT) literature in terms of productivity, quality, and impact. The bibliometric performance analysis is developed into four stages: (1) whole production and impact of published documents; (2) production of main authors, organizations, countries, sources, and research lines; (3) h-index and H-classic analysis; and (4) analysis of content of the documents retrieved. Taking into account the above, the performance and science mapping analysis is carried out using the publications related to Smart Building Technologies (SBT) from 1984 to 2020. The publications and their citations included in this analysis have been collected on October 4, 2020. The publications were retrieved from Web of Science Core Collection – for many researchers, the most important bibliographic database – using the following advance query: TS=(“smart build*” OR “smart building technol*”). Refined by: DOCUMENT TYPES=(PROCEEDINGS PAPER OR ARTICLE OR REVIEW) AND LANGUAGES=(ENGLISH). Timespan=All years. Indexes=SCIEXPANDED, SSCI, A&HCI, CPCI-S, CPCI-SSH, BKCI-S, BKCI-SSH, ESCI, CCREXPANDED, IC. This process retrieved a total of 2,398 publications from 1984 to 2020, distributed into three publications type: proceedings paper (1,378 publications), articles (986 publications), and reviews (61 publications). It is important to highlight that some publications can have multiple classifications.

1470

N. K. Gamboa-Rosales et al.

As a next step, the SciMAT period manager was used to avoid data flatness. The best approach to analyzing the evolution of the data is to evaluate them year by year, but on occasion it is advisable to group them in periods in order to generate a sufficient corpus of publications for analysis. In this case, the best option was to divide the time span into three comparable periods. Therefore, the analysis period (1984–2020) was split into 1984–2010, 2011–2015, and 2016–2020, with 120, 577, and 1,701 publications, respectively.

Publications, Citations, and Impact Figures 1 and 2 present the distribution of publications and cites per year related to Smart Building Technologies (SBT) from 1984 to 2020. In terms of productivity, since the first year, two milestones in the development of the knowledge area can be observed. The first milestone corresponds to the beginning of the growth trend (2002), and the second milestone is related to first historical maximum in terms of annual publications after a period of growth fourteen consecutive years (2019). It is important to highlight that the results of last year (2020) are not definitive, because the data of retrieval and indexing process could modify the final data during 2021. Then, it is expected that 2020 could be a similar or better year as 2019 in terms of production and citations. Moreover, Fig. 2 shows the distribution of citations count achieved by those publications from 1984 to 2020 according to Web of Science Core Collection. From 1997 the patter of citations received by year is positive, achieving in the last year around 3,800 cites. From 2012, where around 220 citations are reached, the citations

Fig. 1 Distribution of publications by year from 1984 to 2020

57 Determining Technologies Trends and Evolution of Smart Building. . .

1471

Fig. 2 Distribution of citations by year from 1984 to 2020

received have increased constantly until reaching around 3,800 citations in 2020. Therefore, it could be predicted that the citations will continue improving in the next years. The publication process and citation distribution showed a positive developmental trend from 1984 to 2020. During this period, a total of 18,125 citations (including self-citations) were recorded, the average citation per publication of 7.56 and the hindex 55 publications.

Most Productive and Cited Authors, Geographic Distribution of Publications, Organizations, Research Areas, Sources, and Funding Sponsors To understand the evolution of Smart Building Technologies, it is also important to know which are the most productive and cited authors, their geographic distribution, affiliation, funding sponsors, sources, and research areas. Table 1 shows the most productive and most cited authors during the period 1984–2020. There was a tie in some positions between various authors, thus all of them have been included in an alphabetical order, and this rule applies for all the research. It is important to mention that only two of the most productive authors are among the most cited authors: Skarmeta, AF (21 publications and 376 cites) and Wang, LF (11 publications and 324 cites). This situation reflects two scenarios, the first one related to most productive authors and the second to most cited authors, but both scenarios promote the growth and development of the knowledge field.

1472

N. K. Gamboa-Rosales et al.

Table 1 Most productive and most cited authors (1984–2020) Publications 22

Author(s) (Cites) Javaid, N (118);

21 18

Skarmeta, AF (376); Lazarova-Molnar, S (49); Spanos, CJ (143) Benhaddou, D (87); Agarwal, Y (222); Kayal, M (48) Stamatescu, G (14); Zou, H (142) Lilis, G (42); Rodriguez, IB (25); Zhou, YX (114) Jin, M (69); Kjaergaard, MB (96); Spanos, C (64); Wang, LF (324) Gupta, R (61); Koshizuka, N (15); Moreno, MV (120); Pacheco, J (67); Tarazaga, PA (71) Balaji, B (98); Koh, J (80); Ramesh, MV (16); Schmeck, H (34); Spachos, P (108); Stamatescu, I (13); Tan, SXD (29)

17 15 13 12 11 10

9

Citations Author(s) (Publications) 1,243 Hall, DL (2); Llinas, J (2) 376 Skarmeta, AF (21); 367 Lee, EK (4); 366 357 356 324

Gerla, M (4); Pau, G (3); Lee, U (2); Wang, LF (11);

321

Visser, HJ (2);

319

Vullers, RJM (1);

298

Huang, P (1); Mutka, MW (1); Soltani, S (1); Xi, N (1); Xiao, L (1)

Table 2 Most productive countries and organizations (1984–2020) Publications 579 237 210 149 125 123 93 84 81 75

Country(ies) United States of America China Italy France Spain; United Kingdom India Germany; South Korea Canada Australia Singapore

Publications 55 53 37 34 31 28 26 25 23 22

Organization(s) University of California Berkeley Nanyang Technology University Carnegie Mellon University Politecnico di Milano University of Southern Denmark Universidad de Murcia Politehnica University of Bucharest Tsinghua University University of Houston Politecnico di Torino

During 1984–2020, the United States appeared to be the most productive country with 579 publications, followed by China and Italy with 237 and 210 publications, respectively. In this way, the most productive organizations were University of California Berkeley, Nanyang Technology University, Carnegie Mellon University, Politecnico di Milano, University of Southern Denmark, Universidad de Murcia, Politehnica University of Bucharest, Tsinghua University, University of Houston, and Politecnico di Torino (Table 2). In terms of geographical distribution, the Smart Building Technologies have to been developed by 88 countries from the 5 continents. Among the most productive countries, a balance can be observed between American, European, and Asian countries. In terms of geographical distribution, the Smart Building Technologies

57 Determining Technologies Trends and Evolution of Smart Building. . .

1473

Table 3 Most productive research areas and WoS categories (1984–2020) Publications 1358 994

Research areas Engineering Computer Science

Publications 998 420

406 371 232

Telecommunications Energy Fuels Construction Building Technology Science Technology other Topics Automation Control Systems Instruments Instrumentation Materials Science Chemistry

406 371 365

213 184 116 93 89

232 205 193 184 169

WoS categories Engineering Electrical Electronic Computer Science Theory Methods Telecommunications Energy Fuels Computer Science Information Systems Construction Building Technology Computer Science Artificial Intelligence Computer Science Interdisciplinary Applications Automation Control Systems Engineering Civil

have to been developed by 88 countries from the 5 continents. Among the most productive countries, a balance can be observed between American, European, and Asian countries. It is important to add that the most productive countries stand out for being part of the most influential countries in the world. These results also coincide with the performance presented by the most productive organizations. Table 3 shows the most productive research areas and WoS categories related to SBT publications from 1984 to 2020. These categories are mainly related to Engineering, Computer Science, Telecommunications, Energy, and Automation. In this context one can also mention the emergent research areas such as Chemistry, Material Science, and Interdisciplinary Applications. Table 4 shows the journals with the largest number of publications. It highlights the concentration of journals related to energy and electronics. This is aligned with the rest of results, which means that the SBT research has a multiples approaches but it is driven by electronics and energy themes. Finally, the main funding sponsors are European Union EU (129 publications), National Natural Science Foundation of China NSFC (105 publications), National Science Foundation NSF (99 publications), United States Department of Energy DOE (42 publications), United States Department of Defense (22 publications), Fundamental Research Funds for the Central Universities (19 publications), Engineering Physical Sciences Research Council EPSRC (18 publications), Ministry of Science and Technology Taiwan (17 publications), Innovation Fund Denmark (16 publications), Natural Sciences and Engineering Research Council of Canada (15 publications), Republic of Singapore National Research Foundation (14 publications), Defense Advanced Research Projects Agency DARPA (13 publications), and Singapore National Research Foundation (13 publications).

1474

N. K. Gamboa-Rosales et al.

Table 4 Journals with the highest number of publications (1984–2020)

Name Energy and Buildings Sensors IEEE Access Energies Applied Energy IEEE Internet of Things Journal IEEE Transactions on Smart Grid Sustainability Sustainable Cities and Society IEEE Industrial Electronics Magazine

Publications related to SBT 57

Total Cites (Citable items 2019) 40,757

Journal Impact Factor (IF-2019) 4.867

5-Year Impact Factor 5.055

Impact Factor without Journal Immediacy Self Cites Index 1.060 4.153

55 51 45 34 24

63,306 51,038 29,605 98,570 12,832

3.275 3.745 2.702 8.848 9.936

3.427 4.076 2.822 9.086 11.705

0.744 0.615 0.738 2.100 2.303

2.570 2.429 1.903 7.069 7.881

22

26,670

8.267

9.758

2.545

7.269

20 20

35,095 7,140

2.576 5.268

2.798 5.143

0.738 1.660

1.711 4.115

19

1,801

13.593

14.708

0.300

13.037

Based on these results, the most relevant authors and publications are analyzed below in terms of citations according to the h-index and H-Classics.

Most Relevant Publications, Authors, Organizations, and Countries Related to SBT Field The H-classics method (Martínez et al. 2014), based on the well-known h-index (Hirsch 2005), assists as an neutral criterion to categorize the identification of the core publications of any research field. This method is used to discover the most relevant paper in the SBT field, and therefore, to identify the authors, countries, organizations, research areas, and funding sponsors that have contributed more to the development of the SBT field. The advanced search query used in the database Web of Science Core Collection has an h-index of 55 (46 Articles, 5 Proceedings Papers and 4 Reviews). Using this h-index score as a reference, the relevant publications were identified and are listed in Table 5. In terms of citations, its distribution showed a positive developmental trend. Thus, a total of 7,150 citations (including self-citations) were achieved, and the average citation per cited article is 130.00. Thus, the most relevant publications

57 Determining Technologies Trends and Evolution of Smart Building. . .

1475

Fig. 3 Distribution of most cited publications by year from 1984 to 2020

distribution and the total number of citations for the SBT field from 1984 to 2020 are shown in Fig. 3. Taking into account the most productive and most cited authors within H-Classic, it can be observed that the most productive authors are not included in the most cited, and vice versa (Table 6). Nevertheless, this situation is consistent with the results of the bibliometric analysis performance. It confirms two frameworks, the first one includes the most productive and cited authors and the second one considers the authors within the H-Classics. This means that the knowledge area is developed for all the stakeholders. With regard to production by countries, the United States is the most productive country within H-Classic with 28 publications, followed by China with 7 publications. Additionally, the rest of the countries considered as relevant for their contribution to the most cited publications are Canada (5 publications), the United Kingdom (5 publications), Greece (4 publications), Switzerland (4 publications), Belgium (3 publications), India (3 publications), Italy (3 publications), the Netherlands (3 publications), Saudi Arabia (3 publications), Australia (2 publications), Denmark (2 publications), and France (2 publications). In this way, the most productive organizations identified in this section are Purdue University (4 publications), Universidad de Toledo (3 publications), Zurich from Polyterrace ETH (2 publications), Huazhong University of Science and Technology (2 publications), Technological Educational Institute of Piraeus (2 publications), University of California, Berkeley (2 publications), London’s Global University UCL (2 publications), and University of California, Los Angeles (2 publications).

1476

N. K. Gamboa-Rosales et al.

Table 5 H-Classics of Smart Building Technologies field Rank 1 2

3

4 5 6 7 8

9 10

11

12

13 14

15

16 17 18

Title (Year, First author, Source) An introduction to multisensor data fusion (1997, Hall, DL, Proceedings of the IEEE) RF Energy Harvesting and Transport for Wireless Sensor Network Applications: Principles and Requirements (2013, Visser, HJ, Proceedings of the IEEE) Internet of Vehicles: From Intelligent Grid to Autonomous Cars and Vehicular Clouds (2014, Gerla, M, 2014 IEEE World Forum on Internet of Things) The Evolution of MAC Protocols in Wireless Sensor Networks: A Survey (2013, Huang, P, IEEE Communications Surveys and Tutorials) Ultrabroadband Elastic Cloaking in Thin Plates (2009, Farhat, M, Physical Review Letters) PDA: Privacy-preserving data aggregation in wireless sensor networks (2007, He, W, INFOCOM 2007, vols. 1-5) Applications of peptide and protein-based materials in bionanotechnology (2010, de la Rica, R, Chemical Society Reviews) Efficient energy consumption and operation management in a smart building with microgrid (2013, Zhang, D, Energy Conversion and Management) An inverse gray-box model for transient building load prediction (2002, Braun, JE, HVAC&R Research) A Survey on Electric Power Demand Forecasting: Future Trends in Smart Grids, Microgrids and Smart Buildings (2014, Hernandez, L, IEEE Communications Surveys and Tutorials) IoT Considerations, Requirements, and Architectures for Smart Buildings-Energy Optimization and Next-Generation Building Management Systems (2017, Minoli, D, IEEE Internet of Things Journal) Fog computing: A cloud to the ground support for smart things and machine-to-machine networks (2014, Stojmenovic, I, 2014 Australasian Telecommunication Networks and Applications Conference) The Smart Grid-State-of-the-art and Future Trends (2014, El-Hawary, ME, Electric Power Components and Systems) A System Architecture for Autonomous Demand Side Load Management in Smart Buildings (2012, Costanzo, GT, IEEE Transactions on Smart Grid) Efficient IoT-based sensor BIG Data collection-processing and analysis in smart buildings (2018, Plageras, AP, Future Generation Computer Systems-the International Journal of escience) Narrow Band Internet of Things (2017, Chen, M, IEEE Access) WSN- and IOT-Based Smart Homes and Their Extension to Smart Buildings (2015, Ghayvat, H, Sensors) A Self-Powered and Flexible Organometallic Halide Perovskite Photodetector with Very High Detectivity (2018, Leung, SF, Advanced Materials)

Citations (average) 1,243 (51.79) 319 (39.88)

311 (44.43)

298 (37.25) 265 (22.08) 207 (14.79) 195 (17.73) 177 (22.13)

173 (9.11) 156 (22.29)

154 (38.50)

145 (20.71)

139 (19.86) 135 (15.00)

126 (42.00)

123 (30.75) 118 (19.67) 116 (38.67)

(continued)

57 Determining Technologies Trends and Evolution of Smart Building. . .

1477

Table 5 (continued) Rank 19

20

21

22 23

24

25

26

27 28

29

30 31

32

33

34

Title (Year, First author, Source) Performance Analysis and Comparison on Energy Storage Devices for Smart Building Energy Management (2012, Xu, Z, IEEE Transactions on Smart Grid) Enhanced Fingerprinting and Trajectory Prediction for IoT Localization in Smart Buildings (2016, Lin, K, IEEE Transactions on Automation Science and Engineering) Recent advances in solar photovoltaic systems for emerging trends and advanced applications (2016, Pandey, AK, Renewable & Sustainable Energy Reviews) Smart buildings (2003, Snoonian, D, IEEE Spectrum) Personal comfort models - A new paradigm in thermal comfort for occupant-centric environmental control (2018, Kim, J, Building and Environment) Multi-objective optimization for decision-making of energy and comfort management in building automation and control (2012, Yang, R, Sustainable Cities and Society) Optimal integration of a hybrid solar-battery power source into smart home nanogrid with plug-in electric vehicle (2017, Wu, X, Journal of Power Sources) Supercooling suppression of microencapsulated phase change materials by optimizing shell composition and structure (2014, Cao, F, Applied Energy) Environmental effects of information and communications technologies (2011, Williams, E, Nature) Vortex shedding induced energy harvesting from piezoelectric materials in heating, ventilation and air conditioning flows (2012, Weinstein, LA, Smart Materials and Structures) Semisupervised Deep Reinforcement Learning in Support of IoT and Smart City Services (2018, Mohammadi, M, IEEE Internet of Things Journal) Optimal design and operation of building services using mixed-integer linear programming techniques (2013, Ashouri, A, Energy) Privacy-Preserving and Efficient Aggregation Based on Blockchain for Power Grid Communications in Smart Communities (2018, Guan, Z, IEEE Communications Magazine) Optimal operation of an energy management system for a grid-connected smart building considering photovoltaics’ uncertainty and stochastic electric vehicles’ driving schedule (2018, Thomas, D, Applied Energy) Fault detection analysis using data mining techniques for a cluster of smart office buildings (2015, Capozzoli, A, Expert Systems with Applications) Multi-agent control system with information fusion based comfort model for smart buildings (2012, Wang, Z, Applied Energy)

Citations (average) 114 (12.67)

109 (21.80)

102 (20.40)

93 (5.17) 89 (29.67)

89 (9.89)

87 (21.75)

87 (12.43)

85 (8.50) 81 (9.00)

79 (26.33)

79 (9.88) 77 (26.00)

77 (25.67)

77 (12.83)

76 (8.44) (continued)

1478

N. K. Gamboa-Rosales et al.

Table 5 (continued) Rank 35

36

37 38 39 40 41 42

43 44

45 46 47

48 49

50

51 52

Title (Year, First author, Source) Automated Demand Response for Smart Buildings and Microgrids: The State of the Practice and Research Challenges (2016, Samad, T, Proceedings of the IEEE) Economic and environmental scheduling of smart homes with microgrid: DER operation and electrical tasks (2016, Zhang, D, Energy Conversion and Management) Patchy Polymer Colloids with Tunable Anisotropy Dimensions (2011, Kraft, DJ, Journal of Physical Chemistry B) Occupant centered lighting control for comfort and energy efficient building operation (2015, Nagy, Z, Energy and Buildings) Synergistic self-assembly of RNA and DNA molecules (2010, Ko, SH, Nature Chemistry) RSSI-Based Indoor Localization With the Internet of Things (2018, Sadowski, S, IEEE Access) Microlocation for Internet-of-Things-Equipped Smart Buildings (2016, Zafari, F, IEEE Internet of Things Journal) Phase distribution and microstructural changes of self-compacting cement paste at elevated temperature (2007, Ye, G, Cement and Concrete Research) Last-Meter Smart Grid Embedded in an Internet-of-Things Platform (2015, Spano, E, IEEE Transactions on Smart Grid) A Design of Greenhouse Monitoring & Control System Based on ZigBee Wireless Sensor Network (2007, Zhou, Y, 2007 International Conference on Wireless Communications, Networking and Mobile Computing, vols. 1-15) The role of smart grids in the building sector (2016, Kolokotsa, D, Energy and Buildings) From Buildings to Smart Buildings-Sensing and Actuation to Improve Energy Efficiency (2012, Weng, T, IEEE Design & Test of Computers) Improving energy efficiency via smart building energy management systems: A comparison with policy measures (2015, Rocha, P, Energy and Buildings) Characterizing the energy flexibility of buildings and districts (2018, Junker, RG, Applied Energy) Energy management for a commercial building microgrid with stationary and mobile battery storage (2016, Wang, Y, Energy and Buildings) Smart Personal Sensor Network Control for Energy Saving in DC Grid Powered LED Lighting System (2013, Tan, YK, IEEE Transactions on Smart Grid) Integration of plug-in hybrid electric vehicles into energy and comfort management for smart building (2012, Wang, Z, Energy and Buildings) A Routing Protocol Based on Energy and Link Quality for Internet of Things Applications (2013, Machado, K, Sensors)

Citations (average) 75 (15.00)

74 (14.80)

73 (7.30) 71 (11.83) 71 (6.45) 69 (23.00) 68 (13.60) 68 (4.86)

67 (11.17) 67 (4.79)

65 (13.00) 64 (7.11) 62 (10.33)

61 (20.33) 58 (11.60)

57 (7.13)

57 (6.33) 56 (7.00) (continued)

57 Determining Technologies Trends and Evolution of Smart Building. . .

1479

Table 5 (continued) Rank 53

54

55

Title (Year, First author, Source) Indoor Air Quality Monitoring using Wireless Sensor Network (2012, Bhattacharya, S, 2012 Sixth International Conference on Sensing Technology) A Survey of Cyber-Physical Advances and Challenges of Wind Energy Conversion Systems: Prospects for Internet of Energy (2016, Moness, M, IEEE Internet of Things Journal) Using machine learning techniques for occupancy-prediction-based cooling control in office buildings (2018, Peng, Y, Applied Energy)

Citations (average) 56 (6.22)

55 (11.20)

55 (18.33)

Table 6 Most productive and most cited authors (1984–2020) Publications 3 2

1

Author(s) (Cites) Wang, LF (222); Yang, R (222) Chen, M (232); Dounis, AI (133); Nagy, Z (126); Papageorgiou, LG (251); Wang, Z (133); Zhang, D (251)

Citations 1,243 319 311

187 authors

265

298

Author(s) (Publications) Hall, DL (1); Llinas, J (1) Visser, HJ (1); Vullers, RJM (1) Gerla, M (1); Lee, EK (1); Lee, U (1); Pau, G (1) Huang, P (1); Mutka, MW (1); Soltani, S (1); Xi, N (1); Xiao, L (1) Enoch, S (1); Farhat, M (1); Guenneau, S (1)

Table 7 Most productive research areas and WoS categories (1984–2020) Publications 36 16 13 12

Research areas Engineering Energy Fuels Computer Science Telecommunications

Publications 24 16 12 9

9

Construction Building Technology Chemistry

6

WoS categories Engineering Electrical Electronic Energy Fuels Telecommunications Computer Science Information Systems; Construction Building Technology Engineering Civil

5

Engineering Chemical

7

Table 7 shows the most productive research areas and WoS categories related to SBT publications from 1984 to 2020. These categories are mainly related to previous results. Table 8 shows the journals with the largest number of publications. It highlights the concentration of journals related to energy and electronics. This context is aligned with the rest of the results, which means that the SBT research has a multiples approach but it is driven by electronics and energy themes, mainly.

1480

N. K. Gamboa-Rosales et al.

Table 8 Most productive journals within H-Classics (1984–2020)

Name Applied Energy Energy and Buildings IEEE Internet of Things Journal IEEE Transactions on Smart Grid Energy Conversion and Management IEEE Access IEEE Communications Surveys and Tutorials Sensors

Total Cites Publications (Citable items related to 2019) SBT 5 98,570 5 40,757

Journal Impact Factor (IF-2019) 8.848 4.867

5-Year Impact Factor 9.086 5.055

Immediacy Index 2.100 1.060

Impact Factor without Journal Self Cites 7.069 4.153

4

12,832

9.936

11.705

2.303

7.881

4

26,670

8.267

9.758

2.545

7.269

2

62,270

8.208

7.447

2.119

6.082

2 2

51,038 18,995

3.745 23.700

4.076 25.928

0.615 4.856

2.429 22.522

2

63,306

3.275

3.427

0.744

2.570

Finally, considering all of the results obtained in the bibliometric performance it can be observed that the Smart Building Technologies are a growing field and area of interest for scientific, academic, and business communities. In this respect, the following section presents the cognitive and intellectual structure of SBT research field.

Conceptual Analysis An overview is provided of the science mapping and the hidden relationships between core themes related to Smart Building Technologies (SBT). This overview is structured into two complementary sections: (i) analysis of the content of the publications and (ii) a conceptual evolution map. The bibliometric mapping or science mapping is one of the most accepted techniques to understand how scientific, academic, or business concepts, knowledge areas, disciplines, and research fields evolve and are related to each other. Additionally, such methods are increasingly valued as a tool for measuring professional, scientific, or academic quality, productivity, and evolution and uncovering the hidden key elements in different research fields (Martinez et al. 2015; Garfield 1986; Callon et al. 1983; López-Robles et al. 2019a; López-Robles et al. 2019b; Guallar et al. 2020).

57 Determining Technologies Trends and Evolution of Smart Building. . .

1481

Although there are various software tools for analyzing science mapping (Cobo et al. 2011a; Börner et al. 2015), SciMAT was used in the present research (Cobo et al. 2012). In this way, the bibliometric methodology applied in this research identifies four stages of analysis in this knowledge area within a specified period: Detection of research themes. For the periods evaluated, the research themes are identified by applying a clustering algorithm (Juliani and de Oliveira 2016) over a normalized co-words network (Callon et al. 1983). Lastly, the similarity between the themes is assessed using the equivalence index (Callon et al. 1991). Visualizing research themes and the thematic network. The research themes defined are determined based on their density and centrality rank values using two tools: the thematic network and strategic diagram (Herrera-Viedma et al. 2020; Callon et al. 1991; He 1999; López-Robles et al. 2020; Gamboa-Rosales et al. 2020). Centrality (c) measures the degree of interaction of a network with other networks, and density (d) measures the internal strength of the network. By considering both types of measures, a knowledge area can be visualized as a set of research themes and plotted on a two-dimensional strategic diagram (Fig. 4a). Therefore, four research themes can be classified (Cobo et al. 2011b): • Motor themes (upper-right quadrant): The themes located in this quadrant are relevant for developing and structuring the sources, disciplines, specialties, knowledge areas, and research fields. These themes are known as the motor themes of the knowledge area, given that they present high density and strong centrality. • Basic and transversal themes (lower-right quadrant): These themes are not well developed but these are relevant for the knowledge area. This quadrant covers general basic themes and transverse. • Highly developed and isolated themes (upper-left quadrant): These themes are strongly related, highly specialized, and peripheral, but these do not have the appropriate importance or background for the area. • Emerging or declining themes (lower-left quadrant): These themes have a low centrality and density and are relatively weak. These themes mainly characterize either disappearing or emerging themes. Discovery of thematic areas. The research themes are analyzed using an evolution map (Fig. 4c), which links the themes of a consecutive period that retains a conceptual nexus (keywords in common). Performance analysis. The contribution of thematic areas and research themes to the entire knowledge area of research are measured quantitatively and qualitatively. It is used to establish the most productive and relevant areas within the field. In this case, the bibliometric indicators used are published documents, number of citations, average citation, and h-index. In this way, the theme’s performance was computed taking into consideration the documents linked with it and its h-index. Moreover, the present research tries to identify the citation classics. For this purpose, the concept of H-Classics proposed by Martínez et al. (Martínez et al.

1482

N. K. Gamboa-Rosales et al.

Fig. 4 (a) Strategic diagram. (b) Thematic network. (c) Thematic and evolution map

2014) was used. H-Classics is well-defined as follows: “H-Classics of a research area ‘A’ could be defined as the H-core of ‘A’ that is composed of the ‘H’ highly cited papers with more than ‘H’ citations received” (Martínez et al. 2014).

Analysis of Content of the Publications Smart Building Technologies (SBT) research field performance is presented below, identifying and visualizing the core themes from 1984 to 2020 using SciMAT, in three periods: (i) 1984–2010, (ii) 2011–2015, and (iii) 2016–2020. Each period presents the strategic diagram, including the number of documents that each theme concentrates, number of citations, h-index, and average citations achieved by each one according to the results obtained from the advanced search query applied in Web of Science Core Collection and the relevant thematic networks for the development of SBT field. During the first period (1984–2010) (Fig. 5), 19 research themes related to Smart Building Technologies were identified and displayed in the strategic diagram: PERVASIVE-TECHNOLOGIES, SUSTAINABLE-SMART-CITY, BLUETOOTHLOW-ENERGY, ARCHITECTURAL-DESIGN, SMART-CAMERA-NETWORK, LIGHTING-CONTROL-SYSTEM, PUBLIC-DATA-SERVICES, NEURAL-NETWORK, SMART-BUILDING-SYSTEM, VIBRATION-SENSING-DETECTIONAND-CONTROL, WIRELESS-SENSOR-NETWORK, SUPPORT-VECTOR-MA CHINE, SENSOR-NETWORKS, SMART-BUILDING, STRUCTURAL-HEALT H-MONITORING, BUILDING-AUTOMATION-AND-CONTROL-SYSTEM, TEMPERATURE-MONITORING-AND-CONTROL, ENERGY-MANAGEMENTSYSTEM, and POSITIONING-AND-NAVIGATION-SYSTEM. The three most productive themes are included in the Motor themes (SUSTAINABLE-SMART-CITY (5 publications) and Basic and transversal themes (SMART-BUILDING (14 publications) and SENSOR-NETWORKS (7 publications)). In this regard, the three most cited themes are included in the Emerging or declining themes (WIRELESS-SENSOR-NETWORK (72

57 Determining Technologies Trends and Evolution of Smart Building. . .

1483

Fig. 5 Strategic diagram and performance of the themes in the period 1984–2010

cites) and Basic and transversal themes (SMART-BUILDING (113 cites) and TEMPERATURE-MONITORING-AND-CONTROL (68 cites)). In terms of productivity, the three most productive themes related to SBT are SMART-BUILDING, SENSOR-NETWORKS, and SUSTAINABLE-SMART CITY. The first one is related mainly with SUSTAINABILITY-STRATEGY, SECURITY-AND-SURVEILLANCE-SYSTEM, CONTEXT-AWARENESS,

1484

N. K. Gamboa-Rosales et al.

ZERO-NET-ENERGY-BUILDING, WIRELESS-SENSOR-NODE, UBIQUITOU S-COMPUTING, MICRO-GRIDS-AND-SMART-GRID, ENERGY-METER, DES IGN-METHODOLOGY, SMART-STRUCTURES, and VIDEO-IMAGERY. Furthermore, SENSOR-NETWORKS is related to SACCADIC-MOTION, MOBILESINK, INSTINCTIVE-COMPUTING, VISUALIZATION-TECHNOLOGIES, ENERGY-EFFICIENCY-MANAGEMENT, CLUSTERING-ALGORITHM, EMBEDDED-SYSTEMS, SMART-ENVIRONMENT, VIDEO-SURVEILLANCE, SMART-MOBILITY, and SMART-WEB-TECHNOLOGIES. Lastly, SUSTAINAB LE-SMART-CITY is related to ECOLOGICAL-HOUSE, ECOLOGICALDESIGN, AUTOCAD-DESIGN, SMART-SENSORS-AND-ACTUATORS, BIOMI METICS, BUILDING-TECHNOLOGIES, SUSTAINABLE-ARCHITECTURE, SMART-MATERIALS, SMART-ARCHITECTURE, INTELLIGENT-DESIGN, and EFFICIENT-DESIGN. Taking into account that this theme is allocated in the Quadrant 1 (Q1) and Quadrant 4 (Q4), its role are motor and transversal for the rest of research themes. Figure 6 presents the research themes that are considered core to the growth of the SBT research field (Motor themes and Basic and transversal themes): SUSTAINABLE-SMART-CITY, PERVASIVE-TECHNOLOGIES, BLUETOOTH-LOW-ENERGY, SMART-BUILDING, SENSOR-NETWORKS, TEMPERATURE-MONITORING-AND-CONTROL, ENERGY-MANAGEMEN T-SYSTEM, POSITIONING-AND-NAVIGATION-SYSTEM, STRUCTURALHEALTH-MONITORING, and BUILDING-AUTOMATION-AND-CONTROLSYSTEM. Finally, this period was focused on the information and communication technologies and its application to develop smart capabilities in the building sector. Additionally, the artificial intelligence was other core research theme during 1984– 2010 period. During the second period (2011–2015) (Fig. 7), 24 research themes related to Smart Building Technologies were identified and displayed in the strategic diagram: DATA-FUSION, VIBRATION-SENSING-DETECTION-AND-CONTROL, PARTICLE-SWARM-OPTIMIZATION, HYBRID-AND-ELECTRIC-VEHICLES, WIRELESS-SENSOR-NETWORK, OPTIMIZATION-ALGORITHMS, SMARTENVIRONMENT, GREENHOUSE-GAS-EMISSIONS, NEURAL-NETWORK, WIRELESS-SENSOR-NODE, SMART-OFFICE-BUILDING, SENSOR-AGENTROBOT, REGULATION-SERVICE-PROVISION, ENERGY-SAVINGS-MODEL, SUSTAINABLE-SMART-CITY, USER-INTERACTION, SERVICE-ORIENTEDARCHITECTURE, BIG-DATA, SMART-CITIES, MULTI-AGENT-CONTROLSYSTEM, SMART-HOME, CYBER-PHYSICAL-SYSTEMS, BUILDING-AU TOMATION-AND-CONTROL-SYSTEM, and INTERNET-OF-THINGS. The three most productive themes are included in the Motor themes (WIRELESS-SENSOR-NETWORK (241 publications)) and Basic and transversal themes (MULTI-AGENT-CONTROL-SYSTEM (75 publications) and SMARTHOME (71 publications)). In this connection, the most cited theme is the most productive, and it achieved 2,949 cites (WIRELESS-SENSOR-NETWORK).

57 Determining Technologies Trends and Evolution of Smart Building. . .

1485

Fig. 6 Thematic network for core themes in the period 1984–2010

In terms of productivity, the most productive theme related to SBT is WIRELESS-SENSOR-NETWORK, which is related mainly with SMARTBUILDING, ZIGBEE-TECHNOLOGY, SMART-SENSORS-AND-ACTUATORS, ENERGY-EFFICIENCY-MANAGEMENT, CLUSTERING-ALGORITHM, MACPROTOCOLS, SMART-METERING, GAME-THEORY, APPLICATION-SCENA RIO-PLANNING, 6LOWPAN, and 3D-VISUALIZATION. Furthermore, the MULTI-AGENT-CONTROL-SYSTEM is related to themes such as THERMALCOMFORT-CONTROL, SUPPORT-VECTOR-REGRESSION, DEMAND-SIDEMANAGEMENT, WIRELESS-SENSOR-AND-ACTUATOR-NETWORK, SHOR T-TERM-LOAD-FORECASTING, ENERGY-MANAGEMENT-SYSTEM, GENETI

1486

N. K. Gamboa-Rosales et al.

C-ALGORITHM, BUILDING-MANAGEMENT-SYSTEM, ARTIFICIAL-NEURA L-NETWORK, ENERGY-EFFICIENT-BUILDINGS, and BUILDING-ENERGYMANAGEMENT-SYSTEM. Lastly, SMART-HOME is related to REAL-TIMEPRICING, DESIGNING-SMART-BUILDINGS, BUILDING-INTELLIGENCE-Q

Fig. 7 Strategic diagram and performance of the themes in the period 2011–2015

57 Determining Technologies Trends and Evolution of Smart Building. . .

1487

UOTIENT, WASTEWATER, NZEB-BUILDING, NONINTRUSIVE-LOADMONITORING, MIXED-INTEGER-PROGRAMMING, MICRO-GRIDS-ANDSMART-GRID, SMART-BUILDING-SYSTEM, WEB-SERVICES, and THERMA L-ENERGY. Taking into account that this theme is allocated in the Quadrant 1 (Q1) and Quadrant 4 (Q4), its role is motor and transversal for the rest of the research themes. Figure 8 presents the research themes that are considered core to the growth of the SBT research field (Motor themes and Basic and transversal themes): WIRELESS-SENSOR-NETWORK, PARTICLE-SWARM-OPTIMIZATION, OPTIMIZATION-ALGORITHMS, HYBRID-AND-ELECTRIC-VEHICLES, DATA-FUSION, VIBRATION-SENSING-DETECTION-AND-CONTROL, SMA RT-HOME, MULTI-AGENT-CONTROL-SYSTEM, INTERNET-OF-THINGS, BUILDING-AUTOMATION-AND-CONTROL-SYSTEM, CYBER-PHYSICALSYSTEMS, and SMART-CITIES. Finally, this period was focused on the technologies 4.0, new electronic devices, mobility, and its integration in cities and building industry. Additionally, the big data and its utilization was other core research theme during 2011–2015 period. During the third period (2016–2020) (Fig. 9), 22 research themes related to Smart Building Technologies were identified and displayed in the strategic diagram: NEURAL-NETWORK, MACHINE-LEARNING, SMART-SENSORSAND-ACTUATORS, BUILDING-MANAGEMENT-SYSTEM, BLUETOOTHLOW-ENERGY, INDOOR-TEMPERATURE, LOW-POWER-WIDE-AREANETWORKS, SERVICE-ORIENTED-ARCHITECTURE, SMART-INFRAST RUCTURE, UBIQUITOUS-COMPUTING, ENERGY-FLEXIBILITY, CYBERPHYSICAL-SYSTEMS, MULTI-OBJECTIVE-OPTIMIZATION, THERMALAND-ENERGY-MANAGEMENT, THERMAL-COMFORT-CONTROL, INTERNET-OF-THINGS, MICRO-GRIDS-AND-SMART-GRID, SMART-HEATINGVENTILATION-AIR-CONDITIONING-SYSTEM, ENERGY-CONSUMPTION, ENERGY-STORAGE-TECHNOLOGIES, TECHNOLOGY-4.0, and BUILDINGENERGY-MANAGEMENT-SYSTEM. The three most productive themes are included in the Basic and transversal themes (THERMAL-COMFORT-CONTROL (806 publications), INTERNETOF-THINGS (607 publications), and MICRO-GRIDS-AND-SMART-GRIDS (285 publications)). In this regard, the most cited themes are included in the Basic and transversal themes (THERMAL-COMFORT-CONTROL (4,409 cites), INTERNET-OF-THINGS (3,409 cites), and MICRO-GRID-AND-SMART-GRID (1,917 cites)). In terms of productivity, the most productive theme related to SBT is THERMAL-COMFORT-CONTROL, which is related mainly with SMARTBUILDING, OPTIMIZATION-ALGORITHMS, OCCUPANT-BEHAVIOR, OCCUPANT-CENTERED-CONTROL, LIGHTING-CONTROL-SYSTEM, BU ILDING-OPERATION, INDOOR-ENVIRONMENT, MONITORING-SYSTEM, PERSONAL-COMFORT-MODELS, INDOOR-AIR-QUALITY, and FUZZYLOGIC. In this regard, INTERNET-OF-THINGS is mainly related to BIG-DATA, SMART-CITIES, SMART-HOME, WIRELESS-SENSOR-NETWORK, FOG-

1488

N. K. Gamboa-Rosales et al.

Fig. 8 Thematic network for core themes in the period 2011–2015

COMPUTING, ENERGY-EFFICIENCY-MANAGEMENT, CLOUD-SERVICES, SECURITY-AND-SURVEILLANCE-SYSTEM, MONITORING-AND-CONTRO L-SYSTEM, LONG-RANGE-COMMUNICATION-TECHNOLOGIES, and EDGE-COMPUTING. Lastly, MICRO-GRID-AND-SMART-GRID is related to DISTRIB-UTED-ENERGY-RESOURCES, DEMAND-SIDE-MANAGEMENT, MODEL-PREDICTIVE-CONTROL, ENERGY-MANAGEMENT-SYSTEM, HYBRID-AND-ELECTRIC-VEHICLES, COMBINED-HEAT-POWER, OVERL AY-NETWORKS, SIDE-MANAGEMENT, LOAD-SHIFTING, and LOADCONTROL. Taking into account that this theme is allocated in the Quadrant 4 (Q4), its role is transversal for the rest of the research themes.

57 Determining Technologies Trends and Evolution of Smart Building. . .

1489

Fig. 9 Strategic diagram and performance of the themes in the period 2016-2020

Figure 10 presents the research themes that are considered core to the growth of the SBT research field (Motor themes and Basic and transversal themes): SMART-SENSORS-AND-ACTUATORS, NEURAL-NETWORK, MACHINELEARNING, BUILDING-MANAGEMENT-SYSTEM, THERMAL-COMFORT-

1490

N. K. Gamboa-Rosales et al.

Fig. 10 Thematic network for core themes in the period 2011–2015

CONTROL, INTERNET-OF-THINGS, MICRO-GRIDS-AND-SMART-GRID, ENERGY-STORAGE-TECHNOLOGIES, ENERGY-CONSUMPTION, BUILDIN G-ENERGY-MANAGEMENT-SYSTEM, SMART-HEATING-VENTILATIONAIR-CONDITIONING-SYSTEM, and TECHNOLOGY-4.0. Finally, this period was focused on the artificial intelligence, smart devices, and technologies 4.0, and its integration in cities and building industry. Additionally, the smart infrastructures were other core research theme during 2016–2020 period. On the other hand, taking into account the strategic diagrams from Figs. 5, 7, and 9, Table 9 presents the main research themes developed and their performance according to the number of documents from 1984 to 2020.

57 Determining Technologies Trends and Evolution of Smart Building. . .

1491

Table 9 SBT intellectual structure from 1984 to 2020 Theme NEURAL-NETWORK BLUETOOTH-LOW-ENERGY BUILDING-AUTOMATIONAND-CONTROL-SYSTEM CYBER-PHYSICAL-SYSTEMS INTERNET-OF-THINGS SERVICE-ORIENTEDARCHITECTURE SUSTAINABLE-SMART-CITY VIBRATION-SENSINGDETECTION-AND-CONTROL WIRELESS-SENSORNETWORK ARCHITECTURAL-DESIGN BIG-DATA BUILDING-ENERGYMANAGEMENT-SYSTEM BUILDING-MANAGEMENTSYSTEM DATA-FUSION ENERGY-CONSUMPTION ENERGY-FLEXIBILITY ENERGY-MANAGEMENTSYSTEM ENERGY-SAVINGS-MODEL ENERGY-STORAGETECHNOLOGIES GREENHOUSE-GASEMISSIONS HYBRID-AND-ELECTRICVEHICLES INDOOR-TEMPERATURE LIGHTING-CONTROLSYSTEM LOW-POWER-WIDE-AREANETWORKS MACHINE-LEARNING MICRO-GRIDS-AND-SMARTGRID MULTI-AGENT-CONTROLSYSTEM MULTI-OBJECTIVEOPTIMIZATION OPTIMIZATIONALGORITHMS

P1: 1984–2010 Q2 (2|0|0) Q1 (2|1|1) Q4 (2|1|1)

P2: 2011–2015 Q2 (9|109|7)

P3: 2016–2020 Q1 (78|620|14) Q2 (55|417|10)

Q4 (45|673|13) Q4 (48|560|13) Q4 (59|838|15) Q3 (5|36|3)

Q1 (5|26|1) Q2 (2|4|1)

Q3 (12|103|5) Q1 (15|185|9)

Q3 (4|72|2)

Q1 (241|2,949|28)

Q3 (60|232|8) Q4 (607|3409|28) Q2 (12|100|4)

Q2 (2|1|1) Q3 (9|87|4) Q4 (163|890|15) Q1 (57|512|11) Q1 (16|334|8) Q4 (131|986|17) Q2 (16|128|7) Q4 (2|52|1) Q3 (17|410|6) Q4 (115|987|16) Q2 (9|64|4) Q1 (26|463|10) Q2 (21|97|5) Q2 (2|5|2) Q2 (6|59|3) Q1 (103|544|11) Q4 (285|1,917|23) Q4 (75|1,106|16) Q3 (39|361|10) Q1 (37|518|10) (continued)

1492

N. K. Gamboa-Rosales et al.

Table 9 (continued) Theme PARTICLE-SWARMOPTIMIZATION PERVASIVE-TECHNOLOGIES POSITIONING-ANDNAVIGATION-SYSTEM PUBLIC-DATA-SERVICES REGULATION-SERVICEPROVISION SENSOR-AGENT-ROBOT SENSOR-NETWORKS SMART-BUILDING SMART-BUILDING-SYSTEM SMART-CAMERA-NETWORK SMART-CITIES SMART-ENVIRONMENT SMART-HEATINGVENTILATION-AIRCONDITIONING-SYSTEM SMART-HOME SMART-INFRASTRUCTURE SMART-OFFICE-BUILDING SMART-SENSORS-ANDACTUATORS STRUCTURAL-HEALTHMONITORING SUPPORT-VECTOR-MACHINE TECHNOLOGY-4.0 TEMPERATUREMONITORING-ANDCONTROL THERMAL-AND-ENERGYMANAGEMENT THERMAL-COMFORTCONTROL UBIQUITOUS-COMPUTING USER-INTERACTION WIRELESS-SENSOR-NODE

P1: 1984–2010

P2: 2011–2015 Q1 (27|642|12)

P3: 2016–2020

Q1 (3|35|2) Q4 (2|4|1) Q2 (2|0|0) Q2 (3|18|2) Q2 (2|0|0) Q4 (7|58|4) Q4 (14|113|6) Q2 (2|0|0) Q2 (2|10|1) Q4 (32|166|6) Q2 (9|94|4) Q4 (141|835|14)

Q4 (71|1,307|17) Q2 (13|53|3) Q2 (2|10|1) Q1 (125|955|15) Q4 (2|2|1) Q3 (2|16|1) Q4 (79|421|10) Q4 (2|68|1)

Q3 (34|188|7) Q4 (806|4,409|28) Q2 (14|73|3) Q3 (4|78|2) Q2 (6|29|4)

Note: Quadrant (Publications|Citations|h-index)

Finally, according to the results obtained and taking into account the main research themes related to the most cited themes, it could be possible to state that the publications and research themes related to Smart Building Technologies (SBT) are robust, linked, and synergic between them, and these will be growing up in the following years.

57 Determining Technologies Trends and Evolution of Smart Building. . .

1493

Conceptual Evolution Map of Smart Building Technologies Considering the results of analyzing the content of the publications for each period, a second analysis focusing on the conceptual evolution of the main themes was carried out. Hence, four thematic areas were recognized: (i) Smart Building Technologies (green area), (ii) Smart Hardware (pink area), (iii) Smart Software (blue area), and (iv) Smart Building Strategies (purple area). These thematic areas consolidate the main themes and research areas within SBT field (Fig. 11). Therefore, it is important to remember that the size of the spheres is proportional to the number of publications identified in each theme and the colored areas represent the clusters of themes pertaining to the same thematic area. Moreover, the dotted line indicates that related thematic areas share some keywords while the solid line represents the thematic link between the different areas. Lastly, the thickness of the lines is proportional to the rate of inclusion (Cobo et al. 2011b). Smart Hardware (pink area) is the most representative thematic area within the conceptual evolution map. It accounts for 1,258 publications, 9,247 cites of the total citations, and 43 documents highly cited according to the hindex. In terms of structure and thematic composition, it remains Basic and transversal themes and Highly developed and isolated themes, but it covers all the quadrants during the three periods. This thematic area covers themes related to BLUETOOTH-LOW-ENERGY, SMART-CAMERA-NETWORK, SENSORNETWORKS, WIRELESS-SENSOR-NETWORK, INTERNET-OF-THINGS, CYBER-PHYSICAL-SYSTEMS, USER-INTERACTION, SENSOR-AGENTROBOT, WIRELESS-SENSOR-NETWORK, WIRELESS-SENSOR-NODE, BLUETOOTH-LOW-ENERGY, CYBER-PHYSICAL-SYSTEMS, INTERNETOF-THINGS, LOW-POWER-WIDE-AREA-NETWORKS, MICRO-GRIDSAND-SMART-GRID, TECHNOLOGY-4.0, SMART-SENSORS-ANDACTUATORS, and UBIQUITOUS-COMPUTING. Smart Building Technologies (green area) is the second thematic area within the map in terms of the number of publications. It has 1,144 publications, 7,137 citations, and 39 highly cited publications according to the h-index. This thematic area remains the Motor themes and Basic and transversal themes in all periods, even though it was Basic and transversal themes. Regarding its thematic composition, it covers themes related to BUILDING-AUTOMATIONAND-CONTROL-SYSTEM, ENERGY-MANAGEMENT-SYSTEM, LIGHTIN G-CONTROL-SYSTEM, PERVASIVE-TECHNOLOGIES, POSITIONINGAND-NAVIGATION-SYSTEM, SMART-BUILDING-SYSTEM, STRUCTURALHEALTH-MONITORING, TEMPERATURE-MONITORING-AND-CONTROL, VIBRATION-SENSING-DETECTION-AND-CONTROL, BUILDING-AUTOMATION-AND-CONTROL-SYSTEM, ENERGY-SAVINGS-MODEL, GREENHOUSE-GAS-EMISSIONS, HYBRID-AND-ELECTRIC-VEHICLES, REGUL ATION-SERVICE-PROVISION, VIBRATION-SENSING-DETECTION-ANDCONTROL, BUILDING-ENERGY-MANAGEMENT-SYSTEM, BUILDING-

1494

Fig. 11 SBT thematic evolution from 1984 to 2020

N. K. Gamboa-Rosales et al.

57 Determining Technologies Trends and Evolution of Smart Building. . .

1495

MANAGEMENT-SYSTEM, ENERGY-CONSUMPTION, ENERGY-FLEXIBILIT Y, ENERGY-STORAGE-TECHNOLOGIES, INDOOR-TEMPERATURE, SMAR T-HEATING-VENTILATION-AIR-CONDITIONING-SYSTEM, THERMALCOMFORT-CONTROL, and THERMAL-AND-ENERGY-MANAGEMENT. Smart software (blue area) is the third thematic area within the conceptual evolution map. It accounts for 361 publications, 3,589 cites of the total citations, and 31 documents highly cited according to the h-index. In terms of structure and thematic composition, it remains Motor themes, Highly developed and isolated themes, and Emerging or declining themes, but it covers all the quadrants during the three periods. This thematic area covers themes related to NEURAL-NETWORK, PUBLIC-DATA-SERVICES, SUPPORT-VECTORMACHINE, BIG-DATA, DATA-FUSION, MULTI-AGENT-CONTROL-SYSTEM, NEURAL-NETWORK, OPTIMIZATION-ALGORITHMS, PARTICLE-SWARMOPTIMIZATION, SERVICE-ORIENTED-ARCHITECTURE, MACHINE-LEARNING, MULTI-OBJECTIVE-OPTIMIZATION, NEURAL-NETWORK, and SERVICE-ORIENTED-ARCHITECTURE. Smart Building Strategies (purple area) is the fourth thematic area within the map in terms of the number of publications. It has 150 publications, 1,824 citations, and 22 highly cited publications according to the h-index. This thematic area remains the Highly developed and isolated themes, mainly. Regarding its thematic composition, it covers themes related to ARCHITECTURAL-DESIGN, SMART-BUILDING, SUSTAINABLE-SMART-CITY, SMART-CITIES, SMART-ENVIRONMENT, SMART-HOME, SMART-OFFICE-BUILDING, SUSTAINABLE-SMART-CITY, and SMART-INFRASTRUCTURE. Finally, it could be mentioned that the Smart Building Technologies intellectual structure is focused on new, original, and innovative research and technological developments such as energy efficiency, comfort, surveillance, mobility, and connectivity.

Conclusions The growth and evolution of Smart Building Technologies research field is positive since its beginning to nowadays. Moreover, given the large volume of publications and citations received, as well as the research themes identified and their evolution in the main sources, it is expected that the scientific, academic, and business communities’ interests will continue over the coming years. In the bibliometric performance analysis, SBT research field is composed by 2,398 publications, distributed into three publications type: proceedings paper (1,378 publications), articles (986 publications), and reviews (61 publications). It is important to remember that some publications can have multiple classifications. In this connection, the SBT research field integrates around 6,863 researches from 1,997 organizations and 25 different countries during the period 1984–2020. In terms of productivity and impact, the balance between the most productive and most cited authors reflects the novelty, interest, and quality of the publications

1496

N. K. Gamboa-Rosales et al.

related to SBT. Based on this, the United States appears to be the most productive country with 579 publications, followed by China and Italy with 237 and 210 publications, respectively. Thus, the most productive organizations were University of California Berkeley, Nanyang Technology University, Carnegie Mellon University, Politecnico di Milano, University of Southern Denmark, Universidad de Murcia, Politehnica University of Bucharest, Tsinghua University, University of Houston, and Politecnico di Torino. Additionally, the most productive journals related to SBT research field are Energy and Buildings, Sensors, IEEE Access, Energies, Applied Energy, IEEE Internet of Things Journal, IEEE Transactions on Smart Grid, Sustainability, Sustainable Cities and Society, and IEEE Industrial Electronics Magazine. According to the conceptual analysis developed using SciMAT tool, four main research themes groups are identified: (i) Smart Building Technologies (green area), (ii) Smart Hardware (pink area), (iii) Smart Software (blue area), and (iv) Smart Building Strategies (purple area). The first thematic area could be considered as core for their contribution to the growth of the SBT research field and is related to themes such as BLUETOOTHLOW-ENERGY, SENSOR-NETWORKS, SMART-CAMERA-NETWORK, WIRELESS-SENSOR-NETWORK, CYBER-PHYSICAL-SYSTEMS, INTERNE T-OF-THINGS, SENSOR-AGENT-ROBOT, USER-INTERACTION, WIRELESSSENSOR-NETWORK, WIRELESS-SENSOR-NODE, BLUETOOTH-LOWENERGY, CYBER-PHYSICAL-SYSTEMS, INTERNET-OF-THINGS, LOWPOWER-WIDE-AREA-NETWORKS, MICRO-GRIDS-AND-SMART-GRID, SMART-SENSORS-AND-ACTUATORS, TECHNOLOGY-4.0, and UBIQUITOU S-COMPUTING. In this way, the second thematic area is core for their contribution to the growth of the field too, and it is related to BUILDING-AUTOMATIONAND-CONTROL-SYSTEM, ENERGY-MANAGEMENT-SYSTEM, LIGHTINGCONTROL-SYSTEM, PERVASIVE-TECHNOLOGIES, POSITIONING-ANDNAVIGATION-SYSTEM, SMART-BUILDING-SYSTEM, STRUCTURALHEALTH-MONITORING, TEMPERATURE-MONITORING-AND-CONTROL, VIBRATION-SENSING-DETECTION-AND-CONTROL, BUILDING-AUTOMATION-AND-CONTROL-SYSTEM, ENERGY-SAVINGS-MODEL, GREENHOUSE-GAS-EMISSIONS, HYBRID-AND-ELECTRIC-VEHICLES, REGULATION-SERVICE-PROVISION, VIBRATION-SENSING-DETECTIONAND-CONTROL, BUILDING-ENERGY-MANAGEMENT-SYSTEM, BUILDINGMANAGEMENT-SYSTEM, ENERGY-CONSUMPTION, ENERGY-FLEXIBILITY, ENERGY-STORAGE-TECHNOLOGIES, INDOOR-TEMPERATURE, SMARTHEATING-VENTILATION-AIR-CONDITIONING-SYSTEM, THERMALAND-ENERGY-MANAGEMENT, and THERMAL-COMFORT-CONTROL. Moreover, the third and fourth thematic areas are oriented to support the rest ones, and these are related to NEURAL-NETWORK, PUBLIC-DATA-SERVICES, SUPPORT-VECTOR-MACHINE, BIG-DATA, DATA-FUSION, MULTI-AGENTCONTROL-SYSTEM, NEURAL-NETWORK, OPTIMIZATION-ALGORITHMS, PARTICLE-SWARM-OPTIMIZATION, SERVICE-ORIENTED-ARCHITECTUR E, MACHINE-LEARNING, MULTI-OBJECTIVE-OPTIMIZATION, NEURAL-

57 Determining Technologies Trends and Evolution of Smart Building. . .

1497

NETWORK, SERVICE-ORIENTED-ARCHITECTURE, ARCHITECTURALDESIGN, SMART-BUILDING, SUSTAINABLE-SMART-CITY, SMART-CITIES, SMART-ENVIRONMENT, SMART-HOME, SMART-OFFICE-BUILDING, SUSTAINABLE-SMART-CITY, and SMART-INFRASTRUCTURE. On analyzing these themes and their relationship, the research finds that the development of Smart Building Technologies field will mainly support the following areas: artificial intelligence, intelligent decision support systems, neural networks, big data, Internet of things, predictive maintenance intelligence, modeling, optimization, planning and scheduling, recommender systems, and computer vision, among other capabilities. Finally, it is important to note that this analysis allows the identification of common themes that can be used to achieve the research lines related to Smart Building Technologies. Moreover, as future research works, a global analysis could be carried out taking into account a yearly time span to understand the evolution of each theme as its thematic network and establish a specific roadmap of each Smart Building Technology. Acknowledgments The authors acknowledge the support by the CONACYT-Consejo Nacional de Ciencia y Tecnología (Mexico), COZCyT Consejo Zacatecano de Ciencia Tecnología e innovación (Mexico) and Research Coordination of the Brazilian Ministry of Education to carry out this study. Additionally, this work has been supported by the Spanish State Research Agency through the project PID2019-105381GA-I00/AEI/10.13039/501100011033 (iScience).

References Börner K, Theriault TN, Boyack KW (2015) Mapping science introduction: past, present and future. Bull Assoc Inf Sci Technol 41(2):12–16. https://doi.org/10.1002/bult.2015.1720410205 Cobo MJ, López-Herrera AG, Herrera-Viedma E, Herrera F (2012) SciMAT: a new science mapping analysis software tool. J Am Soc Inf Sci Technol 63(8):1609–1630. https://doi.org/ 10.1002/asi.22688 Callon M, Courtial JP, Turner WA, Bauin S (1983) From translations to problematic networks: an introduction to co-word analysis. Information (International Social Science Council) 22(2):191– 235. https://doi.org/10.1177/2F053901883022002003 Callon M, Courtial JP, Laville F (1991) Co-word analysis as a tool for describing the network of interactions between basic and technological research: The case of polymer chemsitry. Scientometrics 22(1):155–205. https://doi.org/10.1007/BF02019280 Cobo MJ, López-Herrera AG, Herrera-Viedma E, Herrera F (2011a) Science mapping software tools: review, analysis, and cooperative study among tools. J Am Soc Inf Sci Technol 62(7):1382–1402. https://doi.org/10.1002/asi.21525 Cobo MJ, López-Herrera AG, Herrera-Viedma E, Herrera F (2011b) An approach for detecting, quantifying, and visualizing the evolution of a research field: A practical application to the Fuzzy Sets Theory field. J Inf 5(1):146–166. https://doi.org/10.1016/j.joi.2010.10.002 Furstenau LB, Sott MK, Kipper LM, Machado ÊL, López-Robles JR, Dohan MS, Cobo MJ, Zahid A, Abbasi QH, Imran MA (2020a) Link between sustainability and industry 4.0: trends, challenges and new perspectives. IEEE Access 8:140079–140096. https://doi.org/10.1109/ ACCESS.2020.3012812 Furstenau LB, Sott MK, Homrich AJO, Kipper LM, Al Abri AA, Cardoso TF, López-Robles JR, Cobo MJ (2020b) 20 years of scientific evolution of cyber security: a science mapping. In:

1498

N. K. Gamboa-Rosales et al.

International conference on industrial engineering and operations management. IEOM Society International, Dubai, pp 314–325 Gamboa-Rosales NK, Ayastuy-Arizti JL, Gamboa-Rosales H, Galván-Tejada CE, Luna-García H, Galván-Tejada JI, Celaya-Padilla JM, Otegi-Olaso JR, López-Robles JR (2019) Past, current and future of fuel cell technologies: a bibliometric analysis of the research published during the 1968-2018 period. Paper presented at the VII Symposium on Hydrogen, Fuel Cells and Advanced Batteries (HYCELTEC 2019), Barcelona (Spain) Gamboa-Rosales NK, Celaya-Padilla JM, Hernandez-Gutierrez AL, Moreno-Baez A, GalvánTejada CE, Galván-Tejada JI, González-Fernández E, Gamboa-Rosales H, López-Robles JR (2020) Visualizing the intellectual structure and evolution of intelligent transportation systems: a systematic analysis of research themes and trends. Sustainability 12(21):8759. https://doi.org/ 10.3390/su12218759 Garfield E (1986) Towards scientography. Current Contents 43:3–14 Ghansah FA, Owusu-Manu D-G, Ayarkwa J (2020) Project management processes in the adoption of smart building technologies: a systematic review of constraints. Smart Sustain Built Environ. https://doi.org/10.1108/SASBE-12-2019-0161 Glenisson P, Glänzel W, Janssens F, De Moor B (2005) Combining full text and bibliometric information in mapping scientific disciplines. Inf Process Manag 41(6):1548–1572. https:// doi.org/10.1016/j.ipm.2005.03.021 Guallar J, López-Robles J-R, Abadal E, Gamboa-Rosales N-K, Cobo M-J (2020) Revistas españolas de Documentación en Web of Science: análisis bibliométrico y evolución temática de 2015 a 2019. Prof Inf 29(6). https://doi.org/10.3145/epi.2020.nov.06 Herrera-Viedma E, López-Robles J-R, Guallar J, Cobo M-J (2020) Global trends in coronavirus research at the time of Covid-19: a general bibliometric approach and content analysis using SciMAT. El Profesional de la Información, 2020, vol 29, num 3. https://doi.org/10.3145/ epi.2020.may.22 He Q (1999) Knowledge discovery through co-word analysis. Libr Trends 48(1):133–159 Hirsch JE (2005) An index to quantify an individual’s scientific research output. Proc Natl Acad Sci 102(46):16569–16572. https://doi.org/10.1073/pnas.0507655102 Hoy MB (2016) Smart buildings: an introduction to the library of the future. Med Ref Services Q 35(3):326–331. https://doi.org/10.1080/02763869.2016.1189787 Juliani F, de Oliveira OJ (2016) State of research on public service management: Identifying scientific gaps from a bibliometric study. Int J Inf Manag 36(6):1033–1041. https://doi.org/ 10.1016/j.ijinfomgt.2016.07.003 Lawrence TM, Boudreau M-C, Helsen L, Henze G, Mohammadpour J, Noonan D, Patteeuw D, Pless S, Watson RT (2016) Ten questions concerning integrating smart buildings into the smart grid. Build Environ 108:273–283. https://doi.org/10.1016/j.buildenv.2016.08.022 López-Robles JR, Otegi-Olaso JR, Gamboa-Rosales NK, Gamboa-Rosales H, Cobo MJ (2018) 60 years of business intelligence: a bibliometric review from 1958 to 2017. In: New trends in intelligent software methodologies, tools and techniques: proceedings of the 17th international conference SoMeT_18, 2018. IOS Press, p 395. https://doi.org/10.3233/978-1-61499-900-3395 López-Robles JR, Otegi-Olaso JR, Porto-Gómez I, Cobo MJ (2019a) 30 years of intelligence models in management and business: a bibliometric review. Int J Inf Manag 48:22–38. https:// doi.org/10.1016/j.ijinfomgt.2019.01.013 López-Robles JR, Guallar J, Otegi-Olaso JR, Gamboa-Rosales NK (2019b) El profesional de la información (EPI): bibliometric and thematic analysis (2006-2017). El profesional de la inf 28(4):e280417. https://doi.org/10.3145/epi.2019.jul.17 López-Robles JR, Otegi-Olaso JR, Porto-Gomez I, Gamboa-Rosales H, Gamboa-Rosales NK (2020) Understanding the intellectual structure and evolution of Competitive Intelligence: A bibliometric analysis from 1984 to 2017. Tech Anal Strat Manag 32(5):604–619. https://doi.org/ 10.1080/09537325.2019.1686136

57 Determining Technologies Trends and Evolution of Smart Building. . .

1499

Martínez MA, Herrera M, López-Gijón J, Herrera-Viedma E (2014) H-Classics: characterizing the concept of citation classics through H-index. Scientometrics 98(3):1971–1983. https://doi.org/ 10.1007/s11192-013-1155-9 Martinez MA, Cobo MJ, Herrera M, Herrera-Viedma E (2015) Analyzing the scientific evolution of social work using science mapping. Res Soc Work Pract 25(2):257–277. https://doi.org/ 10.1177/1049731514522101 Minoli D, Sohraby K, Occhiogrosso B (2017) IoT considerations, requirements, and architectures for smart buildings—energy optimization and next-generation building management systems. IEEE Internet Things J 4(1):269–283. https://doi.org/10.1109/JIOT.2017.2647881 Moral-Muñoz JA, Herrera-Viedma E, Santisteban-Espejo A, Cobo MJ (2020) Software tools for conducting bibliometric analysis in science: An up-to-date review. https://doi.org/10.3145/ epi.2020.ene.03 O¨zadowicz A (2017) A new concept of active demand side management for energy efficient prosumer microgrids with smart building technologies. Energies 10(11):1771. https://doi.org/ 10.3390/en10111771 Sott MK, Furstenau LB, Kipper LM, Rodrigues YPR, López-Robles JR, Giraldo FD, Cobo MJ (2021) Process modeling for smart factories: using science mapping to understand the strategic themes, main challenges and future trends. Bus Process Manag J. https://doi.org/10.1108/ BPMJ-05-2020-0181

Managing Process Safety and Operational Risks with Industry 4.0 Technologies

58

John Lee, Ian Cameron, and Maureen Hassall

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Process System and Process Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Process Safety Management Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modelling of Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industry 4.0 Impacts on Process Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . What Are the Opportunities and Threats from Industry 4.0? . . . . . . . . . . . . . . . . . . . . . . . What Are the Next Steps and Recommendations? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion/Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1502 1507 1509 1510 1513 1518 1521 1524 1525 1525

Abstract

Major process safety incidents are still frequently occurring despite many process safety and risk management improvement initiatives. Although digital systems have a long history of application throughout the life cycle of process plants, Industry 4.0 has the transformative potential to reduce process safety incidents and optimize the management of risk. A core technology of Industry 4.0 is the digital twin, a term used for various forms of models connected to plant data that can be utilized for improved analysis, predictions, and decision-making. To realize the transformative potential of Industry 4.0 in the process industries, a fundamental systems thinking approach needs to be applied to the implementation of the digital twin and should be applied for each intended use case. In addition, having a standardized language and ontology, such as ISO15926, is important as it enables the use of reasoning engines and the ability J. Lee () · I. Cameron · M. Hassall School of Chemical Engineering, The University of Queensland, Brisbane, QLD, Australia e-mail: [email protected]; [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_54

1501

1502

J. Lee et al.

to interconnect models and systems across the process and product lifecycle. At present, components of Industry 4.0 are currently being applied to improve process safety, which should extend in the future to more use cases and forms of the digital twin to improve safety and optimize the management of operational risks within the process industries. Successful applications of Industry 4.0 initiatives require an overall strategy aligned with business priorities to deliver transformation in business processes and the management of process safety. Although there are immense opportunities available through the implementation of Industry 4.0, poor implementation can also create threats to managing process safety that need to be considered and mitigated. Keywords

Process safety · Digital twin · Digitalization · Industry 4.0 · Models · Risk management

Introduction Across the world, the process industries are responsible for producing a plethora of products including fuels, metals, chemicals, plastics, processed foods, beverages, chemicals, rubber, textiles, composites, ceramics, wood and paper products, personal care products, and pharmaceuticals. The sustainability, productivity, competitiveness, and reputation of this industry sector depends on its ability to ensure safe operations regardless of the hazardous nature of the materials being processed (e.g., flammable, explosive, toxic, etc.) and despite the hazards inherent in the processing operations (e.g., high temperature, pressures, concentrations, etc.). Activities associated with ensuring safe operations are typically referred to as process safety and/or operational risk management. Managing and improving process safety is a high priority for the process and other high hazard industries. Catastrophic and major accidents continue to occur across industries, causing enormous numbers of fatalities, injuries, economic losses, and environmental damage. Examples over the last 10 years include the following: • 2020 Beirut ammonium nitrate explosion causing over 200 fatalities, more than 6500 injuries, and at least US$15B in losses (Allahoum and Linah 2020); 2020 Vizag gas leak at the LG Polymers chemical plant in India causing 11 fatalities and over 1000 injuries (News India 2020) • 2019 Tlahuelilpan pipeline explosion in Mexico causing 137 fatalities (Rincón 2019); 2019 Yancheng chemical plant explosion in China causing 78 fatalities and 617 injuries (Coote 2019) • 2018 chemical plant in Zhangjiakou city in China’s northern Hebei province causing 22 fatalities and more than 22 injuries (ABC News 2018); 2018 chemical

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

•

• •

•

•

•

•

1503

plant in Sichuan, China, causing 19 fatalities and 12 injuries (China Labour Bulletin 2018) 2017 Tangerang fireworks disaster in Indonesia causing 49 fatalities and 46 injuries (ABC News 2017); 2017 NTPC power plant explosion in India causing 43 fatalities and more than 100 injuries (Proctor 2018) 2016 Dangyang power plant explosion in China causing 22 fatalities and 4 injuries (Wong 2016) 2015 Port of Tianjin explosion in China causing 165 fatalities and 565 injuries (Huang and Zhang 2015); 2015 Bosley wood flour mill explosion in the UK that caused 4 fatalities and devastated the facility (BBC News 2015) 2014 Kunshan metal dust explosion in China causing 146 fatalities, 114 injuries, and US$53M in losses (Li et al. 2016); 2014 Kaohsiung gas explosions in Taiwan causing 32 fatalities and 321 injuries (Central News Agency 2014) 2013 West Texas ammonium nitrate explosion in the USA causing 15 fatalities, over 300 injuries, and more than US$100M in losses (Insurance Journal 2018); 2013 Dhaka garment factory building collapse in Bangladesh causing 1132 fatalities and over 2500 injuries (International Labour Organization 2018) 2012 Paraguana Refinery Complex explosion in Venezuela causing 48 fatalities and over 100 injuries (De Abreu 2012); 2012 Tamil Nadu’s Sivakasi fireworks factory blaze causing 38 fatalities and 75 injuries (India Today Online 2012); 2012 Pakistan garment factory fire in Karachi causing 289 fatalities and 249 injuries (China Daily 2012) 2010 BP Deepwater Horizon blowout causing 11 fatalities, 17 injuries, the worst oil spill in history, and over US$60B in losses (CSB 2016); 2010 San Bruno gas pipeline explosion in the USA causing 8 fatalities and over US$1.6B in losses (NTSB 2011)

These accidents continue to occur at high rates, despite decades of attention and improvements made to process safety management systems as illustrated by (Jarvis 2015) with Fig. 1. As this figure illustrates, more should be done to reduce major losses in the process industries. Further analysis of the major events has found that the primary causes include loss of mechanical integrity, inadequate operating procedures and practices, inadequate control of maintenance work, and inadequate hazard identification (Jarvis 2016). Although inadequate hazard identification was the primary cause in only 10% of these major events, it was a contributing cause to 40%. This research by Jarvis suggests that a priority focus area for Industry 4.0 applications should include improving hazard identification. Good hazard identification requires a thorough understanding of the proposed and current performance of processing systems. The process industry has improved its measuring, reporting, and modeling of processes. Process systems engineering (PSE) focuses on developing models to help personnel design, simulate, control, troubleshoot, and optimize processes. Modelling is an important concept that underpins good decision-making associated with hazard identification, risk management, and operational performance optimization.

1504

J. Lee et al.

Fig. 1 Major losses in the onshore oil, gas, and petrochemical industry 1996–2014 (Jarvis 2015)

To ensure the models used inform, not misinform, decision-makers, the attributes of a model need to align with a fundamental systems conceptualization when being created for process safety applications. Because of its importance, the fundamental systems conceptualization will be discussed in detail in the next section of the chapter. Digitalization features in many of the tools, models, and processes that have been introduced and evolved over the last 50 years, and many of the core technologies of Industry 4.0 already feature to varying degrees within modern processing plants. However, these implementations could be considered more as an evolution of computerization and digitization, which are focused on specific use cases to deliver on business priorities rather than true digital transformation that has fundamentally changed the process industries. Many of these applications are disparate with scattered information that over the years have created a behemoth to manage and maintain. A lot of resources have been invested into these applications, and the disparities that have been created are not necessarily a sufficient driver for change. The process industries also have many other pressing priorities alongside process safety management (PSM), and this can create challenges to maintain the focus and dedicate resources on PSM. So how can Industry 4.0 or digital transformation be leveraged to further improve process safety? For Industry 4.0 to successfully transform and improve PSM, not only should the change demonstrably reduce risk, but it should also make it easier to manage and reduce the already high demand on resources for PSM. New digitalization applications obviously need to be aligned to business priorities to warrant change.

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1505

When it comes to PSM, without an obvious business case and just implementing change for change sake can increase rather than reduce risk. Depending on what has been previously implemented or what has fallen into a state of disrepair, there may be “low-hanging fruit” within the Industry 4.0 landscape that can be easily or quickly implemented. Examples include high hazard plants that don’t have functioning, updated or accurate steady-state flowsheet models, or operator training simulators. Also, simply improving the visibility of data and information across an organization may enable the identification of new areas for PSM improvement or avert a major accident. Or there may be other opportunities such as adding lowcost wireless fidelity (WiFi) devices to measure and transmit corrosion rate data in high-risk areas or to improve monitoring of critical equipment to avoid running to failure. Although specific tools that fall under the Industry 4.0 umbrella can be applied in isolation to address business priorities, the ensuing benefits are incremental. Process safety involves all parts of the lifecycle from concept to decommissioning and across most enterprise activities, such as design, operations, procurement, and maintenance. Industry 4.0 may offer a lot of opportunities for improving process safety; however, its transformative potential derives from being applied as a universal concept across an enterprise. Such opportunities can be shown as a maturity or journey model. Several companies such as Bentley Systems, Inc., and ARC Advisory Group have developed journey or maturity models for Industry 4.0 implementation. The German National Academy of Science and Engineering, acatech, conducted a project to develop a comprehensive leading example – “Industrie 4.0 Maturity Index.” The maturity index was developed systematically and collaboratively with academia and industry to help provide guidance for the manufacturing industries (Schuh et al. 2017). These maturity models help enterprises to holistically consider gaps and to plan the journey to Industry 4.0 while aligning activities with business priorities. Figure 2 is an example of maturity levels as specifically applied to the use of flowsheet models for improving process safety. In recent years, there has been an exponential increase in services and offerings for the implementation of digital transformation and digital twins, with goals or side benefits of improving process operation’s hazard and risk management. However, if new applications are implemented without the fundamental prerequisites (as per Fig. 2 example for the use of flowsheet models) or without an overall strategy that aims to simplify the management and optimization of operation risks and reduce resource demands over time, then it will not generate step-change improvements in process safety. Such an overall strategy should consider how to bring together disparate systems to make them easier to use and maintain and to integrate them into a “single source of the truth” that considers not just the plant but also the people and procedural interactions in a manner that enables processes and process risks to be optimized. Thus, the adoption of the concepts of Industry 4.0 needs to consider the human perspectives of change as well as the implementation of new technology (Gallagher 2017). Stakeholders need see significant improvements, whether that is new and improved ways to further reduce risks or whether it makes their lives easier

1506

J. Lee et al.

OPTIMISED: ADVANCED: DYNAMIC: DEFINED: BASIC: Basic process measurement but poor/outdated or no system models

Good quality and comprehensive measurement of processes and models that represents steady state operations

Dynamic models available and provide up to date representations of plant. Dynamic modelling of normal and abnormal operations used for design and operator training Good in-house understanding of dynamic model design and operation

Use of dynamic modelling to assess process safety risks implications of all decisions from project concept to decommissioning In-house capability available to update and use dynamic model Steady state and dynamic models use compatible software and have common representation of plant (no duplication)

Incorporation of other business risks and IA into models to allow risk-based optimisation of processing operations over entire lifecycle. Changes to operations automatically update models. In-house dynamic model building and maintenance capabilities Steady state and dynamic models connected to plant data. Automated health monitoring, detection and assessment of degradations of capability and faults Dynamic model used to develop advanced abnormal situation management capabilities. Models capable of self-calibration

Fig. 2 Maturity model for the use of flowsheet models for risk management

and less resource demanding for true Industry 4.0 transformation and to sustain on the journey. In the remainder of this chapter, we describe in more detail how the transformative potential of Industry 4.0 might be leveraged to reduce process safety incidents and to optimize a process operation’s risks. We start with explaining processes and process safety in fundamental systems terms and why this is important. Next, we discuss process safety management requirements before describing the role that modelling plays in the process industries, across a project’s lifecycle perspectives, before outlining contemporary uses of models in the process industries. Then we provide an overview of Industry 4.0 technological changes relevant to process safety and then some of the potential opportunities and threats that could emerge from this technology. We conclude by outlining the next steps and recommendations for using Industry 4.0 advancements to optimize the management of process safety and operations risks.

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1507

Constraints {k} Boundary

S System

Inputs

{û}

functions, capabilities states, parameters

Operating modes

{o}

S {f,c} {x,p}

Methods

Tasks

{m}

{t}

Outputs

{y}

Fig. 3 General systems conceptualization. (Adapted from Cameron and Gani 2011)

Process System and Process Safety In order to develop, use and interpret Industry 4.0 technologies, it is prudent to understand the system they are being introduced to and/or are modelling. A general representation of a dynamic process system adapted from Cameron and Gani (2011) is depicted in Fig. 3. The features of this conceptualization are the specification of the boundary that defines the domain of a dynamic system being modelled, which has inputs, states, and outputs determined by the nature of the system. These inputs include manipulated variables as well as measured and unmeasured disturbances. The system and its subsystems have intended functions that are delivered through its components’ capabilities and are defined by a set of parameters. There are certain constraints on states and outputs within which the system should operate. The system is impacted by changes in inputs and disturbances. Tasks may be carried out within the system to deliver certain goals, and a series of tasks may be regularly completed in a defined succession to serve as a method of doing something. An operating mode relates to the several ways that something operates, such as normal operation, start-up, and shutdown. Representing a system in this way helps to explore the impacts of introducing Industry 4.0 technologies and the features and variations in parameters that have been included and excluded from Industry 4.0 models. These variations could include changes in the system boundary, inputs, outputs, system functions and capabilities, system behavior, operating modes, constraints, tasks, and methods. To explore the interactions among the system “actors” as well as the environment, the P3 schema has been developed as shown in Fig. 4. This schema helps analysts consider not only the roles of plant, people, and procedures but also their interactions, which are crucial for ensuring safe processing systems (Seligmann et al. 2012, 2019). This perspective highlights that the system is dynamic with many interfaces and interactions between and among plant, people, and procedures,

1508 Fig. 4 High-level systems view capturing key actors and interactions: a “P3 schema.” (Adapted from Lee et al. 2019)

J. Lee et al.

Environment System

Plant

People

Procedures

Fig. 5 Formative functional systems framework (Hassall 2013)

which can be shaped or influenced by environmental conditions and changes. These are important aspects and need to be considered in the design and maintenance of system integrity within a dynamic environment and can present many challenges in the management of process safety. Major process safety events are usually a result of a combination of factors which can be derived, at least in part, from the conceptualizations shown in Figs. 3 and 4. These can include changes in inputs, disturbances from the environment, failures of components or changes in component capabilities, and internal deviations, all of which can lead to divergences in intended functions or the execution of other action options, which can affect the system’s operating state and outputs as shown in Fig. 5. These system frameworks have been developed to help people identify, describe, and analyze a processing system to identify vulnerabilities to accidents and opportunities to enhance process safety. It helps analysts identify hazards, analyze process safety risks, and determine the management interventions required to assure system

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1509

safety. In addition, and importantly, they can be used to design and evaluate the process models that are crucial for any Industry 4.0 transformational change in the process industries. The next subsection discusses the determination of management interventions, and the following subsections describe process models.

Process Safety Management Systems Frameworks have also been developed to highlight important interventions that should be considered when managing hazardous operations. Such interventions include policies, standards, procedures, and practices as well as details on roles and responsibilities, all of which need to be in place and be effective to prevent major accidents and to mitigate their potential impact should they occur. Many regulators and industry bodies publish guidance on safety management systems. An example of a process safety management system framework from the CCPS is shown in Fig. 6. The World Steel Association has the following process safety management fundamental principles (Worldsteel 2019): 1. 2. 3. 4.

Ensure there is a commitment to process safety management. Establish a hazard and risk analysis program. Implement and maintain a process safety risk management and control system. Strive to excellence in learning from experience.

Fig. 6 Process safety management system (CCPS 2007)

1510

J. Lee et al.

5. Utilize continuous improvement to ensure process safety management effectiveness. 6. Maintain a sense of vulnerability in process safety management. Another example framework from the Asia Industrial Gases Association (AIGA 2017) includes the following: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

Leadership, commitment, and responsibility Compliance with legislation and industry standards Employee selection, training, and competency Workforce involvement Communication with stakeholders Hazard identification and risk assessment Documentation, records, and knowledge management Process and operational status monitoring and handover Operating procedures Management of operational interfaces Standards and practices Management of change Operational readiness and process start-up Emergency management Inspection and maintenance Management of safety critical devices Work control, permit to work, and task risk management Contractors and suppliers – selection and management Incident investigation Audit, management review, and intervention Measures and metrics

The introduction of industry 4.0 should be done with consideration to all the elements of a process safety management system to ensure it enhances and integrates the elements and not weaken or silo them.

Modelling of Processes In the early 1960s, process systems engineering (PSE) was introduced to enable integrated design and optimization of industrial processes (Sargent 1963). Digital computer systems had been recently introduced and were to play a critical role in PSE supported by appropriate modelling over the life cycle. PSE has continuously evolved over the last 60 years, in parallel with the exponential improvements in computing and programming as shown in Fig. 7. The result is an ecosystem of models for use across the life cycle as shown in Fig. 8. The growth in computing power and software over the last seven decades has made it possible to develop high-fidelity dynamic simulations of the entire process

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1511

Commercial Process Simulation Tools 1990 - 1995

Digitization of plant measurements and control 1950 - 1960 Introduction of digital computers 1940 - 1950 Development of analogue computers

1981 - 1990

Internet

GUI and

emerges

2011 Industry 4.0

dynamic simulation capability 1961

2004 - 2019 1994

Concept of Process Systems

ISO15926

WWW Consortium created

Engineering introduced

1940

2020

1992 WiFi 1965 - 1975

1985 - 1995

Commercial CAD tools become available

Distributed Control Systems implemented 1975 - 1985

Digital systems for maintenance

2007 5G invented 2004 - 2009

Ontology Web Language Developed

, finance , optimization implemented

Fig. 7 The history of digitalization within the process industries. (Adapted from Lee et al. 2019)

Fig. 8 A model-centric view of engineered systems. (Adapted from Cameron and Gani 2011)

plants. The simulations are a form of digital twin that can represent the interactions between people, plant, and procedures, at least in an operating perspective. Such models present new opportunities to improve risk management practices by being better able to consider human factors and operational policies. Over the last decade, the process industries have been applying components of Industry 4.0 to improve in critical areas that are the primary causes of major process safety incidents. As previously discussed in the Introduction, these critical areas are mechanical integrity, control of maintenance, operational procedures, design, and hazard identification.

1512

J. Lee et al.

With a strong focus on improving mechanical integrity, the process industries have been applying digital systems to improve corrosion management systems, including risk-based inspection, repairs, corrosion models, and connections to live plant data. WiFi technology has also enabled an increase in the implementation of online corrosion rate sensors. Greater connectivity with remote monitoring systems has helped to reduce rotating equipment failures. Machine learning and predictive analytics are being applied to data to alert to issues with equipment that may lead to failure. Maintenance activities have resulted in many process safety incidents, and digital systems have been applied to improve the control of work and energy isolation processes, improve risk assessment, and even provide real-time risk assessment for when critical barriers are nonfunctional or out of service for maintenance. To reduce process safety incidents related to operator error, the process industries have improved control systems interfaces, implemented operator training simulators (OTS), and advanced alarm management systems. Getting the design of a process plant right is the most impactful stage of the life cycle to improve process safety. Dynamic models similar to those used for OTS are also being utilized to assist with design by assessing specific abnormal situations and the performance of control and safety instrumented system design. Computer-aided design (CAD) tools have been developed and evolved over the years, and they connect engineering design, project documentation, equipment data, process and instrumentation drawings (P&IDs), process flow diagrams (PFDs), and 3D visualization via a common language and ontology. These tools also assist with project management and implementation including scheduling, approvals, version control, management of change, procurement, and construction to increase efficiency, reduce error, and improve the quality of the physical asset. These tools are increasingly being utilized after commissioning to assist with ongoing operation and maintenance of the asset. With the aim of improving risk assessment processes, qualitative process hazard analysis software has been developed such as ExpHAZOP+ (Rahman et al. 2009), PHASuite (Zhao et al. 2005), and BLHAZID (Németh et al. 2011), although these have not seen widespread industrial application. Models have been developed to assess the consequences of loss of primary containment, such as for assessing dispersion, fire, and explosion. Some critical scenarios may utilize complex computational fluid dynamics to assess consequences. In contemporary process operations, the dynamic models used within operator training simulators are typically one of the most advanced forms of digital models. They are often designed and used to improve operator training, assess competency, and develop and test procedures. OTS are being increasingly implemented for complex and integrated plants, such as those in oil and gas, chemical, and mineral processing. These simulators are also finding application early in the life cycle of plants for assessing specific abnormal events to improve design, assessing potential design changes, designing control systems and safety instrumented systems (SIS), reducing commissioning and start-up issues and duration, and developing emergency procedures.

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1513

There are challenges to implementing and maintaining OTS associated with resources and costs. Often the creation and maintenance of the dynamic models and simulators requires specialty third-party expertise in the software with limited inhouse capability. The creation of an OTS for an operating plant can cost over US$1 million, not to mention the cost of ongoing training costs and funding to support improvements and changes for plant modifications. Often the flowsheeting software in the OTS is not the same as what the process engineers use for design, health monitoring, and optimization and often not available for use outside of the OTS environment. Being familiar with the dynamic model and the simulation software is important for developing awareness of simplifications and ability to represent complex behaviors that may limit the models’ accuracy. Having such awareness is vital to being able to correctly use and interpret the outcomes from the model. In addition, there has been research undertaken to investigate options for re-tasking these models for other use cases, such as process hazard analysis. Preliminary findings indicate that process safety and domain expertise is required to recognize and handle model deficiencies as well as to interpret the outputs. Industry 4.0 offers further opportunities to leverage and improve dynamic models and OTSs.

Industry 4.0 Impacts on Process Operations Although Industry 4.0 refers to the fourth industrial revolution, it represents the evolution of digital and computer systems, applied in an integrative way, enabling a step change in performance and decision-making capability. Industry 4.0 entails several key technological advancements as shown in Fig. 7. Although these advancements are shown as discrete categories, they are highly integrated with some enabling and informing others. At the core of the depiction in Fig. 9 is the high-level P3 schema perspective essential to process safety considerations within the overall space of the Internet of things (IoT) technologies that apply to Industry 4.0. Surrounding the process safety perspective is a box depicting the hierarchical components of data providing the foundation for information to build knowledge that creates wisdom to make good decisions. IoT technologies, such as models for predictive analytics, artificial intelligence, virtual and augmented reality, automation, autonomous systems, and digital twins, rely on reliable and cleansed data as a foundation in order to derive useful information and knowledge and to make proper decisions. Although this depiction is shown as a set of concentric boxes, IoT technologies interact with each of the P3 schema actors or impact the interactions between them. The interconnectivity provided by the Internet is a key component of Industry 4.0 and has fundamentally changed how businesses operate in sectors such as banking, retail, communications, and travel. It has also fundamentally changed social interactions and education. The Internet has enabled large volumes of data to be brought together from a variety of data sources and accessible widely to other systems and people. The Internet has also provided the ability to apply virtually

1514

J. Lee et al.

Internet of things (IoT) Artificial intelligence Predictive analytics

5G

Information

Decision making

Digital twins

WiFi

VR-AR-MR systems

Data

Knowledge

Autonomous/automated systems Internet Fig. 9 Landscape of Industry 4.0 technologies that could impact process operation. (Adapted from Lee et al. 2019)

unlimited computing power at local problems. WiFi and 5G have enabled access to remote real-time transfer of data to provide additional inputs from people and plant and new paths for outputs to actuate plant. Virtual, augmented, and mixed reality are categories of cyber-physical systems that allow people to interact with the plant and its digital representation often within a 3D representation and can improve the contextual interactions between people and plant. A digital twin, which is a term used for the cyber representation of something physical, can be used to improve analysis, make predictions, and improve decisionmaking. Some of the first digital twins were created in the 1970s for the aerospace industry. The digital twin can be connected to live information of the actual asset to enhance its capability and use cases. The digital twin may also consist of an ecosystem of models, each developed for a specific use case and lifecycle phase. These models may be based on fundamental physics/chemistry/mathematics, while others may be empirically based or both. Some models may be a “black box” approach with machine learning providing algorithms to predict outputs.

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1515

OREDA database

CAD module

Dynamic Simulation Model

ISO15926 ontology in OWL/RDF

Maintenance management systems

LP models

Blended HAZID Module

Other engineering models

Production scheduling models

Inspection management systems Reasoning Agent Financial management systems

Data Validation

Historian

Live Plant

Use Cases: Process Hazard Analysis, causal diagrams, optimization, production scheduling, health monitoring, self calibration, fault detection, abnormal situation management, Process safety management, reliability assessments, RBM, RBI, management of change.

A conceptual digital twin for the process industries

Fig. 10 Conceptual example of dynamic model within the digital twin concept. (Adapted from Lee et al. 2019)

As discussed in the previous section, a lot of Industry 4.0 components are being applied within the process industries to improve process safety, not to mention other applications delivering on other business priorities. However, they are often on disparate systems with disparate data sources, making them resource-intensive to keep up to date and increasing the risk of human error in accessing and transferring information between systems. Industry 4.0 gives digital twins the opportunity to integrate the disparate data sources into one easy-to-use, single source of truth. Figure 10 depicts an overall digital twin concept that interconnects an ecosystem of models and systems connected via a common language and ontology, which can help manage models, data, and information flows and simplify change management. Connecting the digital twin with live data can assist with health monitoring, fault detection, process and risk optimization, and abnormal situation management, thereby helping to improve human decision-making, process safety, plant reliability, and production outcomes. Having a common language and ontology is crucial to developing an integrated interconnected digital twin. ISO15926, the international standard for “Industrial automation systems and integration – Integration of life-cycle data for process plants including oil and gas production facilities,” provides a framework for integrating, sharing, and exchanging data. It provides a common language and ontology for the process and manufacturing industries suitable across the life cycle of plant. This ISO standard has been adopted widely in the process industries within CAD

1516

J. Lee et al.

tools typically used in the design, construction, and commissioning of new plant. There are some existing tools developed to utilize this information, such as stress analysis software, standards comparison software (HAZID 2019), and qualitative risk assessment software (Németh et al. 2011).There are potential opportunities to maintain the information of the plant in this format over the entire life cycle of the plant following construction. There are also opportunities to connect this representation of the plant with other models, such as steady-state and dynamic flowsheeting models to drive further business value through new use cases. For many processing plants built before the use of CAD and ISO15926, there is quite a challenge to “back-cast” this ontology. There are commercial tools available that purport to be able to scan P&IDs and utilize artificial intelligence (AI) to recognize equipment and text (Bilfinger 2018; HAZID 2017) and also photometry/laser scanning tools also utilizing AI to recognize equipment types. Research work is being conducted by Lee et al. (2019) investigating parsing dynamic flowsheeting schemas into the ISO15926 schema to help lay an initial skeleton ontology for the process plant. Such a parsing tool may also be useful in projects as typically a flowsheeting model (or at least a very high-level view) is available very early in a project. There are opportunities to increasingly use robotics such as robots and drones and AI visual interpretation to remove personnel from high-risk tasks, such as for inspection, leak repairs, confined space entry, catalyst removal, and cleaning. These applications address personal safety more than process safety. Automation is already applied extensively within process plants via instruments, analyzers, control systems, and the remote operation of valves and switches. These directly impact the operation of the plant and process safety. Industry 4.0 initiatives may assist the management of operational risks by improving the availability or representation of data and information. They may also provide advice to operators by the assessment of operational risk and the provision of timely advice on imminent failures and response actions. Automation and robotics applied to process safety may take the form of additional automated start-up, controlled shutdowns, emergency shutdowns, and control system overrides. The ability to apply these forms of automation or control can be associated with the degree of complexity of the plant. For example, it is not uncommon for steam boilers, combustors, and furnaces to have automated start-up while much less common on large, complex processing plants such as fluid catalytic crackers, crude distillation, hydrocrackers, and reformers. Start-up procedures for complex plant often have thousands of steps. It is also not uncommon to have issues with instruments and equipment at start-up, especially if they are dedicated for startup and rarely utilized. Such issues can thwart an automated start-up, requiring human intervention. For automation to handle such issues would require additional redundancy such as two out of three instrument systems, and even then this does not necessarily remove start-up issues. Field operator activities during start-ups involve the movement of manual valves, the movement of blinds, the use of vents and drains, the connection and disconnection of utility hoses, starting and stopping rotating equipment, sampling, and sometimes sample analysis. Much of the associated

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1517

equipment is not utilized during normal operation, and making changes to allow automation would be extensive and costly. Automated shutdown of process plant is being increasingly implemented across the process industries to reduce risk, especially for high hazard processes. These automated shutdown systems aim to take a process plant to a safer state; however, they can also increase process risk if they unnecessarily shutdown a process. Even emergency automated shutdown systems typically require operator intervention to purge, clean, and isolate equipment for maintenance. Additional automation and control implementation needs to be aligned with critical priorities and potential adverse risks considered and mitigated. Additional automation to reduce operator resources needs to be considered holistically as the minimum operator manning of a refinery may be set by many factors, including abnormal situations, emergency response, or even turnaround isolation and permitting requirements. Some process plants have remote detection of the release of hazardous materials or fire and have automated water deluge systems to reduce emergency response demands. Many complex processing plants have advanced control such as dynamic matrix control (DMC) operating as an “autopilot” for the optimal operation of the plant within constraints. These advanced controls can handle changes in feed quality, feed rates, and product targets; however, during abnormal situations such as equipment failure, these may hand back control to an operator. Because the advanced control may operate the plant most of the time, the operator may lose competency and capability to operate the process and is one of the drivers behind the requirement for an OTS. Advanced data analytics is another category of Industry 4.0 technologies, which includes big data, predictive analytics, and artificial intelligence. Big data is a term used to represent the extremely large datasets of information that can be analyzed to identify patterns, trends, correlations, and associations among the data. When related to a processing plant, big data could include high-fidelity engineering design data that defines the entire processing plant as well as historical and live operational data collected from sensors, testing, analysis, and observations. Advanced analytics, predictive analytics, and machine learning are automated or semi-automated methods of examining large amounts of existing or historical data to generate insights or make predictions. Artificial intelligence (AI) is also a core Industry 4.0 technology. Artificial intelligence refers to technology that seeks to mimic human cognition, particularly the ability to perceive the environment, assess the situation, and then select and execute a response. Some more advanced forms of AI also possess the ability to learn and adapt based on that learning. Example process safety applications of AI include the meta-analysis of natural language data (e.g., incident reports, shift reports, etc.) to derive insights into accident precursors and control vulnerabilities, CCTV analysis to alert operators and/or emergency response systems of leaks, and the use of deep learning to prioritize and rationalize alarms to prevent alarm flood scenarios. However, the introduction and use of AI, and any other Industry 4.0 technology in high hazard and safety critical processing environment, has the

1518

J. Lee et al.

potential to improve and worsen safety management endeavors. The opportunities and threats associated with Industry 4.0 technologies are discussed next.

What Are the Opportunities and Threats from Industry 4.0? A survey of oil refinery professionals conducted by Accenture (2018) indicated that they believed significant value may be delivered by further implementing new technologies. In the survey, although improving reliability was considered a significant driver for the implementation of Industry 4.0 and has the potential to reduce some process safety events, improving process safety was not specifically identified as a driver. Industry 4.0 can offer new opportunities to reduce risks but there are also threats to process safety. As discussed in Sect. “Conclusion/Summary”, there are already Industry 4.0 technologies being applied throughout a plant’s life cycle that improve design, mechanical integrity, operating procedures, operator competencies, maintenance of plant, and hazard identification. There are opportunities to improve operations and further reduce process safety risks by better interconnection of these systems throughout the life cycle and reduced duplication of data. Reducing the time and resources to identify risks and make decisions and implement actions, such as the modification of plant, reduces the duration of exposure to risks and increases the ability to reduce more risks. The transformative potential of Industry 4.0 is derived when applied across an enterprise within a common framework and strategy, which may fundamentally change business processes. One could envision a future state for the operation of a process plant akin to flying and controlling the “Starship Enterprise” from Star Trek: The Next Generation with artificial intelligence, augmented reality, and autonomous operations. Abnormal situations are almost always associated with process safety incidents, and SIS have been implemented to override the operator and bring the process to a safe state, often to a shutdown requiring restart, which in itself introduces risk. There are opportunities for artificial intelligence to provide earlier identification and optimal corrective advice or intervention to further reduce risk and improve reliability. It may be as simple as having more open and accurate data or information could help to avert an accident. Knowledge management and the retention of experience are challenges facing many process plants with aging workforces. A well thought-out and implemented Industry 4.0 strategy has the ability to greatly improve existing knowledge management practices and provide opportunities to capture and embed the experience or wisdom of experienced personnel within computer systems. Industry 4.0 can also cause the deterioration of knowledge and competency of personnel, which will be discussed later. There is a degree of hype and urgency associated with Industry 4.0; however, exemplar implementation within the process industries is not yet clear. There are significant threats to regret spend and poor outcomes where an overall strategy is

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1519

missing or there is a lack of willingness or competencies to lead, implement, and sustain change. Implementing an Industry 4.0 agenda without a well thought-out strategy has the potential to direct resources away from ongoing process safety management, which can lead to accidents. A poorly thought-out strategy may also lead to adding more systems and complexity for little improvement, compounding demands on resources, and increasing the risk of human error and process safety events. An Industry 4.0 strategy should consider what digital systems and tools have already been implemented and not only how new initiatives can provide new opportunities but also how heritage digital systems can be incorporated into the strategy to make managing change easier and reduce risk and resource demands. Outside of the threats to PSM due to resource demands and regret spend on Industry 4.0, there are also direct process safety threats associated with the implementation of Industry 4.0. From data quality, acquisition, interpretation, and decision-making to taking action, errors at any of these stages can lead to or fail to avoid process safety incidents. Similarly, badly designed or incorrectly interpreted outcomes from digital twins, artificial intelligence, or predictive analytics can lead to incidents. These risks must be considered and mitigated whether Industry 4.0 technologies are providing information or advice or taking autonomous actions. Even where autonomous intervention is designed to improve safety, it is possible that new risks are introduced, as experienced with crashes of the Airbus A400M in 2015 (Flightsafety 2015), the Tesla autonomous vehicle in 2018 (NTSB 2020), and the Boeing 737 MAX in 2018/2019 (The House Committee on Transportation and Infrastructure 2020). Autonomous operation must be designed for and tested against every possible scenario and operating mode to ensure robust safety, and even then human override over autonomous operation should always be an option. George Box once said that “all models are wrong, but some are useful.” This is an important tenet for Industry 4.0 and the digital twin. At the core of a digital twin is a digital model (or ecosystem of digital models) designed for a specific use case that will likely be a good replica of the physical asset, but it will be “wrong.” To correctly use and interpret output from digital twins and to develop reasoning agents to interface with the digital twins requires a combination of skills, including process-specific design and operational experience and capability, as well as indepth model design knowledge and IT skills. Such combination of skills is required to understand or identify a model’s weaknesses, which is critical to avoid potential deleterious outcomes from blindly accepting outcomes or directions from a digital twin. As discussed previously, Industry 4.0 technologies impact the interactions between people, plant, and procedures. There are other threats associated with the implementation of enhanced automation, not only through bad data, system design errors, and programming faults but also due to the removal of the human element that can often deal with previously unknown situations and events better than computers. Increasing automation has the potential to result in reduced operator manning levels both at the DCS and in the field. The human senses of sight, sound, smell, and touch (heat/vibration) are in play during routine operator rounds and have a

1520

J. Lee et al.

significant role in managing operational risk. Removing such means of detection of abnormal situations would require significant additional engineered systems such as cameras, gas detectors, vibration sensors, and temperature probes, and then there is the requirement to analyze this information, draw conclusions, and recommend or take action. Auto-starting equipment like pumps and compressors without human involvement can introduce risks such as undetected seal failures and vibration, thermal shock, pump cavitation, and compressor liquid impingement. These risks need to be considered and typically require additional engineered systems to mitigate them. Greater implementation of Industry 4.0 can reduce the human interactions with plant and create a dependency on the technologies. The loss of human experience and knowledge over time can lead to a degradation of capability to control a process when human intervention is required. Implementation of new technologies can also lead to new failure modes or loss of ability to control a process. All of these risks need consideration and mitigation. Simulators play an important role in such situations where operators can practice both normal and abnormal operation routinely in a safe environment, while normal operation of the plant is by automated systems such as DMCs. Taking the human out of the “loop” introduces risks as machines and computers are not necessarily good at dealing with new, previously unconsidered scenarios. An example of the beneficial human response to unexpected scenarios includes the pilot-controlled landing of US Airways Flight 1549, which landed without power on the Hudson River after multiple bird strikes at takeoff. Human experience can build wisdom and improve decision-making. Industry 4.0 implementations also need to be designed with the ability to continuously improve and capture new learnings and knowledge. There are risks if the right competencies and capabilities are not applied to implementing Industry 4.0 initiatives. Although IT and computer science competencies are critical resources, the involvement and direction from those with domain expertise will be critical to successful outcomes. This domain expertise should entail a collaboration of engineering and operational expertise. In terms of engineering expertise, this involves experience applying the professional competencies described by professional organizations for chemical and process engineers, such as the IChemE and AIChE within. It should also entail the ability to apply fundamental systems thinking as previously discussed. From an operational perspective, domain expertise should span field and control room operations as well as electrical, mechanical, maintenance, instrumentation, and control system personnel. The inputs of these domain experts should span both technical and human factors to ensure Industry 4.0 initiatives improve both human and overall system performance and well-being. Another important aspect to consider when thinking about Industry 4.0 initiatives is the massive amounts of data that are routinely collected from operating facilities. This data arises from each of the major components of plant, people, and procedures as seen in Fig. 4. Typically, billions of data points from control systems connected to the plant are sampled each day, manipulated, and held in memory or archived in

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1521

plant historians. The advent of improved, diverse, and cheaper sensing technologies has driven this data tsunami. This can lead to a situation of being data rich but information poor (DRIP). Optimal location of sensors to maximize information content in data is crucial, as is the recognition that the underlying physics, chemistry, and/or biology within the system makes system data highly correlated. This means information redundancy and the need to apply intelligent approaches to univariate and particularly multivariate data analysis tools and the models built from them. Additionally, those models need to be “fit for purpose” in the sense that their use is directed to aiding specific decision-making activities. Those decisions are often constrained by time, risk, and insights, often leading to a set of models for a range of end-goals. This means that interpretation by data scientists or engineers alone without the consideration of the fundamentals of the system will provide limited insights and poorer-quality predictions. Predictive analytics that use historical data require consideration of any data sampling rates and subsequent averaging or filtering that has occurred prior to use in building predictive tools. In the case of dynamic systems, normal operational variations are usually inadequate to characterize the underlying system dynamics, thus requiring specific system perturbations to build effective models. As well, there is an important use of multivariate methods such as principal component analysis and related factor methods that can effectively build predictive tools that consider highly correlated datasets. Care must be exercised to ensure that data is appropriate and properly treated and that data correlation is adequately dealt with in many predictive methods that include neural networks and machine learning techniques (Kresta et al. 1991). Many data-driven models and phenomenological models are used in real-time fault diagnosis. The quality of those models and the adaptation to changing process circumstances are vital for high-quality detection and diagnostic applications (Dash and Venkatasubramanian 2000).The aphorism that “all models are wrong, but some are useful” can be extended to advanced data analytics and artificial intelligence applications. A significant threat to these technologies is the quality of the data underpinning them as previously mentioned. Another threat is the degree of bias built into the algorithms and the interpretation of results. This requires deep domain knowledge and understanding of the models being used. Obtaining and retaining the combination of skills required to build, maintain, and correctly utilize such systems is crucial for effectively managing operational risks and enhancing the safety of process systems.

What Are the Next Steps and Recommendations? As previously discussed, there are many use cases for digitalization and digital twins to improve the various aspects that contribute to process safety incidents. Dynamic models, based on a combination of fundamental- and empirical-based calculations, can offer accurate predictions of plant behavior under most circumstances and represent the interactions between people, plant, and procedures, at least from

1522

J. Lee et al.

an operational perspective. Such digital models offer further opportunities for the application of reasoning agents for assessing process safety risks. The development of reasoning agents requires the “encoding” knowledge and experience of design, operations, and process safety as well as the knowledge of model deficiencies. Even better still would be if the ability to identify and manage deficiencies was encoded within a reasoning agent; however, such a higher-order reasoning agent may currently be beyond reach. The application of qualitative risk assessment, in parallel with dynamic model quantitative risk assessment, may help the reasoning agent to recognize questionable outcomes. There is no well-publicized exemplar implementation of Industry 4.0 within the process industries suitable as a blueprint for others. Although there may be many exemplar digitalized applications and forms of digital twins, there is no exemplar implementation of an overall configuration framework for enterprisewide application over the life cycle of the asset which can promote coordination of change, simplify processes, and deliver further value through enabling synergies across applications with the ability to apply the range of the Industry 4.0 tools. It is postulated that an exemplar application would involve the concept of a centralized platform using a common standard language and ontology that is able to communicate with other existing digital models and applications. Reasoning engines could be developed that leverage off ontological web language tools and existing applications to deliver synergies and improved business outcomes. Within the process industries, ISO15926 is the standard language and ontology applicable across the entire life cycle and is commonly used with CAD tools for the design and construction of greenfield plants. Typically, this information provides all equipment details, process flow diagram, and P&ID details as well as a 3D representation of the plant. There are opportunities for the owner to ensure that the ISO15926 information is handed over with a project as it contains one of the most detailed and comprehensive representations of the physical asset in an internationally recognized language and ontology. This information alone, however, does not address the ability to represent the function of the plant and has the ability to make predictions like flowsheeting models. Certainly, steady-state flowsheeting models would have been created as part of the design process and/or to support the operation. Also many recent greenfield plants would also have detailed high-fidelity dynamic models associated with OTS for training, competency assessment, control system design/tuning, and commissioning. The ability to accurately replicate the function of a process plant and its components and thus provide predictive capabilities is an essential property of a digital twin to deliver high value in the operating phase of the life cycle. Significant resources are required to develop and maintain OTS and where these have been created can be exploited to improve procedures, improve human centric design, analyze abnormal situations, and rationalize alarms. The representation of the human-machine interfaces and operational procedures within the OTS facilitates consideration of all of the system actors within a process hazard analysis. In the future, there is potential to connect the OTS with reasoning agents, live data, and

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

BLHAZID Module

ISO15926 ONTOLOGY in OWL/RDF Process Units/ Streams

Goals and Functions Substances

Procedures

Topology

Parameters

Plant item state

Compare outcomes

Dynamic Model

Analysis Module

Unmitigated consequences

Sort scenarios

FMEA

OREDA

Significant Consequences

IPLs

Discrepancies Causal graphs/ diagrams

Additional Inputs

1523

HAZOP style Left hand side output of bow tie

Risk rank

Alarm signatures

Process safety time

Fig. 11 Conceptual example of PHA using an OTS. (Adapted from Lee et al. 2019)

other models or applications to facilitate additional use cases such as higher levels of automation or enhanced operational assistance. The development of such tools can improve productivity, reduce error, and enhance knowledge management. Although many models and applications may be used within a digital twin, dynamic flowsheeting models, such as those used in operating training simulators, have a role to play. Figure 11 below is a conceptual example from research conducted by Lee et al. (2019) of how a dynamic model associated with an OTS may be used to enhance process hazard analysis. It is evident from this research that the development of such an application requires the involvement of personnel with a high level of knowledge and experience in process design, process safety, and operating the physical plant. As well as the indepth engineering capability, a high level of capability in computer science and IT is required for the development of such an application. Currently this combination of high-level skills is rarely available in personnel. Even upskilling the computer science and IT capabilities of the experienced engineer may not be sufficient as it is a challenge to dedicate sufficient time for an experienced engineer to be a competent computer programmer. It is more likely that collaboration of highly experienced computer scientists and engineers is required to successfully develop these sorts of applications. A significant amount of time and resources will be required to develop new and novel applications. Since such tools would be applicable across the process industries, they should be advanced through collaboration across industry and with academia rather than developed by individual companies, especially in the field of improving process safety.

1524

J. Lee et al.

Conclusion/Summary Process safety remains a priority for the process industries, and many technologies associated with Industry 4.0 are already in use to reduce process safety risks. Computer modelling for various use cases has been utilized for many decades; however, it is often on disparate platforms with many versions of the same data. The lack of the use of a common language and ontology has restricted the ability to share information among applications in an automated way, making the use of these tools complex and resource-intensive. The use of ontologies can be an enabler for the development of reasoning engines to be applied to the models and applications. The limited connectivity and integration has restricted the development of a digital twin suitable for use across the life cycle of a process. To fully exploit the opportunities of Industry 4.0, systems thinking is also needed. Systems engineering provides a fundamental systematic approach for laying out the requirements of the model to deliver on a specific use case. This approach allows one to lay the framework for model building, including determining boundaries, components, and constraints as well as modelling assumptions. Harnessing the full potential of Industry 4.0 to improve process safety requires highly experienced multidisciplinary teams who are knowledgeable in process engineering, operational aspects, process safety, and computer science. These skillsets will be important to identify opportunities and manage threats associated with digitalization initiatives and will be important determinants of success. In the future, engineers will increasingly need IT skills as they will be required to interact with digital twins of the processes that they design and operate. The use case for any digitalization initiative needs to align with the priorities of the organization to ensure appropriate resources are allocated. Mapping these priorities against existing and potential applications can help an organization focus on what digitalization initiatives should be progressed first. If reducing incidents due to operator error is a high priority, implementing proven OTS technology may represent the best application of resources. Or if reducing process safety incidents due to mechanical integrity failure is a high priority, implementing corrosion models, online corrosion sensors, or enhanced digital systems for inspection and predictive maintenance should be targeted. Organizations with good process safety performance may choose to prioritize digitalization opportunities to improve optimization of production or reduce costs rather than improve process safety management. The OTS is a form or component of a digital twin that has demonstrated the ability to improve operator competency and procedures and help to reduce process safety incidents. The dynamic model within the OTS is based on fundamental physics, mathematics, and chemistry and has the ability to accurately represent the interactions between plant, people, and procedures. These OTS could be utilized as a systems safety model for process hazard analysis. There are opportunities to leverage off existing OTS investments to improve process hazard analysis, provided that its limitations are recognized and the output is correctly used and interpreted. There may be additional opportunities to connect the dynamic model

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1525

to live data for other applications such as abnormal situation management, fault detection, and health monitoring as well as non-process safety-related applications like optimization. A key attribute for the success in the implementation of Industry 4.0 is the use of a standard language and ontology such as ISO15926 to support interconnectivity, replication, and information sharing and to provide a framework that can be built on by others in the future. Successful implementation of Industry 4.0 concepts not only should deliver improved business outcomes but also needs to make things easier for stakeholders and reduce resource demands by reducing manual data handling, reducing heuristic analysis, and reducing complexity to make the maintenance of systems easier.

Important Websites Website https:/978-3-030-84205-5/www.15926.org https://www.aiche.org/ccps https://dexpi.net https://en.acatech.de/publication/industrie-40-maturity-index-update-2020/ https://itwinjs.org https://www.mimosa.org

Description The Semantic Web for Engineering Center for Chemical Process Safety Data Exchange in the Process Industry Industry 4.0 Maturity Index. Managing the Digital Transformation of Companies Open source infrastructure digital twins Open Standards for Physical Asset Management

References ABC News (2017) Fireworks factory explosion kills at least 47 outside Jakarta. Retrieved from http://www.abc.net.au/news/2017-10-26/indonesia-fireworks-factory-explosion-kills-dozens/90 90478 ABC News (2018) At least 22 dead in blast outside northern China chemical plant. Retrieved from https://www.abc.net.au/news/2018-11-28/blast-in-china-hebei-province-kills22-injures-at-least-22/10562194 Accenture (2018) The intelligent refinery – distilling digital, driving innovation. Retrieved from https://www.accenture.com/_acnmedia/pdf-79/accenture-2018-refining-research.pdf#zoom=50 AIGA (2017) Process safety management framework guidance document. Retrieved from http:// asiaiga.org/uploaded_docs/en_AIGA_099_17_Process_Safety_Management_Framework-Guid ance_Document.pdf Allahoum TR, Linah AR (2020) Endemic corruption’ caused Beirut blast, says Diab: Live updates. Retrieved from https://www.aljazeera.com/news/2020/08/beirut-police-fire-tear-gas-protestersregroup-live-updates-200810010528285.html BBC News (2015) Bosley mill explosion: HSE finds ‘dust issues’ at blast site. Retrieved from https://www.bbc.com/news/uk-england-33887081 Bilfinger (2018) Bilfinger introduces new technology for the digitalization of plant documents. Retrieved from https://www.bilfinger.com/fileadmin/corporate_webseite/Presse/pressenotizen/ 2018/achema/2018-06-11-FP-Bilfinger-PIDGraph-EN.pdf Cameron IT, Gani R (2011) Product and process modelling: a case study approach. Elsevier Science

1526

J. Lee et al.

CCPS (2007) Guidelines for risk based process safety. American Institute of Chemical Engineers/Wiley-Blackwell, Hoboken Central News Agency (2014) 12 indicted for Kaohsiung gas explosions. Focus Taiwan. Retrieved from https://focustaiwan.tw/society/201412180010 China Daily (2012) 289 killed in Karachi factory fire in Pakistan. Retrieved from http:// www.chinadaily.com.cn/world/2012-09/13/content_15754743.htm China Labour Bulletin (2018) Sichuan chemical plant explosion: investigation reveals familiar problems. Retrieved from https://clb.org.hk/content/sichuan-chemical-plant-explosioninvestigation-reveals-familiar-problems Coote D (2019) Death toll rises to 78 in Chinese chemical plant explosion. Retrieved from https://www.upi.com/Top_News/World-News/2019/03/26/Death-toll-rises-to-78-in-Chinese-ch emical-plant-explosion/1601553571988/ CSB (2016) Macondo blowout and explosion. Retrieved from https://www.csb.gov/macondoblowout-and-explosion/ Dash S, Venkatasubramanian V (2000) Challenges in the industrial applications of fault diagnostic systems. Comput Chem Eng 24(2–7):785–791 De Abreu L (2012) A devastating fire at Venezuela’s main oil refinery, spread Monday. Retrieved from http://www.petroleumworld.com/storyt12082802.htm Flightsafety (2015) Aviation safety network database. Retrieved from https://aviation-safety.net/ database/record.php?id=20150509-0 Gallagher S (2017) Industry 4.0 Testlabs in Australia -preparing for the future, a report of the prime Minister’s industry 4.0 taskforce. Retrieved from http://www.industry.gov.au/sites/g/files/ net3906/f/July%202018/document/pdf/industry-4.0-testlabs-report.pdf Hassall ME (2013) Methods and tools to help industry personnel identify and manage hazardous situations. Doctor of Philosophy, The University of Queensland, Queensland, Australia HAZID (2017) HAZID conversion integrity. Retrieved from https://www.hazid.com/conversionintegrity HAZID (2019) HAZID technology. Retrieved from https://www.hazid.com/ Huang P, Zhang J (2015) Facts related to August 12, 2015 explosion accident in Tianjin, China. Process Safety Progr 34(4):313–314. https://doi.org/10.1002/prs.11789 India Today Online (2012) Toll in Sivakasi cracker unit fire rises to 38, 6 people arrested, search on for factory owner. Retrieved from https://www.indiatoday.in/india/south/story/owners-ofsivakasi-cracker-factory-arrested-115319-2012-09-06 Insurance Journal (2018) Settlements reached in West Texas fertilizer plant explosion. Retrieved from https://www.insurancejournal.com/magazines/mag-features/2018/01/22/477523.htm International Labour Organization (2018) The Rana Plaza accident and its aftermath. Retrieved from https://www.ilo.org/global/topics/geip/WCMS_614394/lang%2D%2Den/index.htm#:∼:te xt=On%2024%20April%202013%2C%20the%20collapse%20of%20the,Tazreen%20Fashions %20factory%20on%20the%20outskirts%20of%20Dhaka Jarvis R (2015) An analysis of common causes of major losses in the onshore oil, gas and petrochemical industries. Retrieved from https://www.icheme.org/membership/communities/ special-interest-groups/safety-and-loss-prevention/resources/hazards-archive/hazards-27/ Jarvis R (2016) An analysis of common causes of major losses in the onshore oil, gas and petrochemical industries. Retrieved from http://www.lmalloyds.com/LMA/Underwriting/ Non-Marine/Onshore_Energy/Onshore_Energy_Wordings/Common_Causes_of_Losses_in_th e_Oil%20_Petro_Industry.aspx Kresta JV, MacGregor JF, Marlin TE (1991) Multivariate statistical monitoring of process operating performance. Can J Chem Eng 69:35–47 Lee J, Cameron I, Hassall M (2019) Improving process safety: what roles for digitalization and industry 4.0. Process Saf Environ Prot 132:325–339. https://doi.org/10.1016/j.psep.2019.10.021 Li G, Yang HX, Yuan CM, Eckhoff RK (2016) A catastrophic aluminium-alloy dust explosion in China. J Loss Prev Process Ind 39:121–130. https://doi.org/10.1016/j.jlp.2015.11.013 Németh E, Seligmann BJ, Hockings K, Oakley J, O’Brien C, Hangos KM, Cameron IT (2011) Generating cause-implication graphs for process systems via blended hazard identification

58 Managing Process Safety and Operational Risks with Industry 4.0 Technologies

1527

methods. In: Computer aided chemical engineering, vol 29. Elsevier, Amsterdam, pp 1070– 1074 News India (2020) Vizag gas leak LIVE updates: 11 dead, over 5,000 sick after leak at LG polymers plant; CM announces Rs 1 crore for Kin of deceased. Retrieved from https://www.news18.com/news/india/visakhapatnam-gas-leak-live-updates-lg-polymersvizag-andhra-pradesh-2608889.html NTSB (2011) Pacific Gas and Electric Company natural gas transmission pipeline rupture and fire. Retrieved from https://www.ntsb.gov/Investigations/AccidentReports/Pages/PAR1101.aspx NTSB (2020) Highway accident report 20/01 Proctor D (2018) Report: human error to blame in fatal India plant accident. Retrieved from https:/ /www.powermag.com/report-human-error-to-blame-in-fatal-india-plant-accident/ Rahman S, Khan F, Veitch B, Amyotte P (2009) ExpHAZOP+: knowledge-based expert system to conduct automated HAZOP analysis. J Loss Prev Process Ind 22(4):373–380. https://doi.org/ 10.1016/j.jlp.2009.01.008 Rincón E (2019) Sobreviviente de la tragedia en Tlahuelilpan regresa a Hidalgo. Retrieved from https://www.excelsior.com.mx/nacional/sobreviviente-de-la-tragedia-en-tlahuelilpan-regresa-ahidalgo/1319364 Sargent RWH (1963) Chemical engineering and engineering science. Inaugural lecture as Professor of Chemical Engineering, Imperial College, March 12. Chem Eng 168:CE151–CE155 Schuh G, Anderl R, Gausemeier J, ten Hompel M, Wahlster WE (2017) Industrie 4.0 maturity index. Managing the digital transformation of companies (acatech study). Herbert Utz Verlag, Munich. Retrieved from https://www.acatech.de/wp-content/uploads/2018/03/ acatech_STUDIE_Maturity_Index_eng_WEB.pdf Seligmann BJ, Németh E, Hangos KM, Cameron IT (2012) A blended hazard identification methodology to support process diagnosis. J Loss Prev Process Ind 25(4):746–759. https:// doi.org/10.1016/j.jlp.2012.04.012 Seligmann BJ, Zhao J, Marmara SG, Corbett TC, Small M, Hassall M, Boadle JT (2019) Comparing capability of scenario hazard identification methods by the PIC (Plant-PeopleProcedure Interaction Contribution) network metric. Saf Sci 112:116–129. https://doi.org/ 10.1016/j.ssci.2018.10.019 The House Committee on Transportation and Infrastructure (2020) Final committee report on the Boeing 737 MAX. Retrieved from https://transportation.house.gov/imo/media/doc/ 2020.09.15%20FINAL%20737%20MAX%20Report%20for%20Public%20Release.pdf Wong E (2016) Explosion at coal-fired plant in central China kills at least 21. Retrieved from https://www.nytimes.com/2016/08/12/world/asia/china-danyang-power-plant-blast.html Worldsteel (2019) Process safety management fundamentals. Retrieved from https://www. worldsteel.org/en/dam/jcr:eb3f7e7a-4f7a-4eef-b2bb-85ce7b2825ea/Process%2520Safety%252 0Management.pdf Zhao C, Bhushan M, Venkatasubramanian V (2005) PhaSuite: an automated HAZOP analysis tool for chemical processes part II: implementation and case study. Process Saf Environ Prot 83(6B):533–548. https://doi.org/10.1205/psep.04056

Lignin: A Renewable Chemical Feedstock

59

Uroosa Ejaz and Muhammad Sohail

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Properties of Lignin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classification of Lignin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lignin as a Renewable Feedstock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Traditional Use of Lignin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Advanced Use of Lignin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Use of Lignin Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Use of Lignosulfonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Use as Copolymer Component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Extraction of Lignin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kraft Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lignosulfonate Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Organosolv Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Use of Steam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alkaline Pulping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Role of Ionic Liquids in Lignin Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Relevant Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1530 1531 1532 1533 1534 1534 1534 1534 1535 1535 1536 1536 1537 1538 1538 1539 1540 1540 1543

Abstract

Increased consumption of fossil fuels has resulted in environmental pollution and global warming. Therefore, chemical industry is making efforts to find sustainable and alternative renewable resources. Plant biomass, especially lignin, is one of the most abundant natural renewable raw materials. Currently, most of

U. Ejaz · M. Sohail () Department of Microbiology, University of Karachi, Karachi, Pakistan e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_55

1529

1530

U. Ejaz and M. Sohail

this resource (∼98%) is burnt; therefore, material potential of lignin remains largely untamed. Lignin can be used either chemically modified or in native form, as a component for copolymers and composites, heavy metal sequestrant, emulsifier, dispersant agent for pesticides, or adhesives, yet it possesses more potential for developing green products. Depending on the methods adopted for extraction, several lignin preparations have been reported including Kraft lignin, lignosulfonates, organosolv lignin, and steam-exploded lignin. Preparation of lignin using ionic liquids (ILs) is an emerging technology, and keen interest in this area is increasing exponentially. Along with the comparison of different extraction methods, utilization of currently available and prospecting future lignin preparations has been summarized in this chapter. Keywords

Feedstock · Extraction · Ionic liquid · Lignin

Introduction Lignocellulose is considered as an important and renewable feedstock of chemicals and energy as it does not compete with the food crops (Korányi et al. 2020). Valorization of lignin, a component of lignocellulose that resists to biodegradation, is a beneficial approach which can contribute to biorefineries in sustainable and economical way (Zhao et al. 2020). The word lignin was firstly used by the Swiss botanist A. P. Candolle (1778–1841) derived from the Latin word lignum which means “wood.” It is the second most sustainable and abundant component in wood and carbon source in nature after cellulose (Pin et al. 2020). It accounts for 30% of all terrestrial nonfossil carbon (Arapova et al. 2020). Hemicellulose and cellulose form a framework in plant cell wall in which lignin is embedded as a connector which strengthen the cell wall by solidifying it like a tar (Khan et al. 2020; Ejaz and Sohail 2020a). Cell wall lignification makes plants resistant to pests and wind (Irmer 2017), reduces the biochemical stress by suppressing the enzymatic degradation of structural polysaccharides, and also controls the flow of fluids (Boerjan et al. 2003). Worldwide, Kraft lignin is produced 100 million tons annually with about 70 million tons of that coming from pulp and paper industry (Mandlekar et al. 2018; Luo and Abu-Omar 2017); this production witnesses an ever-increasing trend (Bajwa et al. 2019). Only 2% of this lignin is commercially utilized, while the remaining is burnt as low-grade fuel (Luo and Abu-Omar 2017). This represents a huge opportunity for deriving value-added products from lignin feedstock (Fig. 1). Due to its aromatic and high-functionality property, lignin can be utilized for the production of bulk and fine chemical compounds (Moreno and Sipponen 2020) such as vanillin, muconic acid, and polyhydroxyalkanoate (Kline et al. 2010). Recently, biological conversion of lignin to useful products has opened a new horizon which represents relatively green process. Lignin properties have a major impact on bioconversion of lignin by microorganisms. Briefly, bioconversion of

59 Lignin: A Renewable Chemical Feedstock

1531

Fig. 1 An overview of utilization of lignocellulosic biomass

lignin consists of three steps: fractionation/pretreatment of lignocellulosic biomass, depolymerization of lignin, and catabolism of aromatic compound to obtain valuable products (Zhao et al. 2020). Substantial efforts have been done to improve the efficiency of these processes for lignin valorization (Li et al. 2019). Lignin has low thermal conductivity, and it is capable of absorbing UV radiation (Toh et al. 2005). Lignin is also more resistant to chemical and biological actions as compared to hemicellulose/cellulose (Thakur and Thakur 2015) and also possesses antimicrobial and antioxidant activity like other bioactive polymers (Cruz et al. 2001). The antimicrobial and antiradical activity of lignin is of great interest for applications such as in food packaging additives (Alzagameem et al. 2019). Other beneficial properties of lignin are hydrophobicity or hydrophilicity; good rheological characteristics, such as film forming ability, viscosity, and elasticity; and compatibility with other chemicals (Doherty et al. 2011).

Properties of Lignin Adler in 1977 proposed the structural model of lignin (Adler 1977). The model showed that oxidation of p-hydroxycinnamyl alcohols releases all the types of structural elements detected in lignin. Later, Zhang and LeBoeuf (2009) reported the thermodynamic properties of lignin including solubility parameter (δ), density (ρ), thermal expansion coefficient (α), and glass transition temperature (Tg). With a molecular weight of 1000–15,000 Da (Arapova et al. 2020), lignin is heavier than other abundant components of plant cell wall. In most of the cases, lignin is composed (Fig. 2) of syringyl (S), guaiacyl (G), and/or hydroxyphenyl (H) units (Pin et al. 2020). Its structural features render it highly resistant to chemical and physical

1532

U. Ejaz and M. Sohail

Fig. 2 Structure of lignin

actions (Strassberger et al. 2015). Fragmentation of various chemical bonds, i.e., 5–O–4 , 5–5 , β–β , β–5 , and β–O–4 , present in lignin (Río et al. 2015; Menezes et al. 2017) results in the formation of phenolic groups (Chakar and Ragauskas 2004). Monolignols are mainly bound by β-O-4 ether bonds, which accounts for more than 50% fraction in lignin (Chakar and Ragauskas 2004). Polymerization occurs due to the radical polycondensation of several monolignols (sinapyl alcohols, coniferyl, and p-coumaryl) (Liu 2012; Brien et al. 2012). Extraction condition and source of biomass highly influence the structure of lignin (Menezes et al. 2017). The ratio and composition of lignin in plant depends on the type of plant species. In general, lignin in plant feedstock decreases in the order: coniferous (softwood) > deciduous (hardwood) > grass. In hardwood, softwood, and cereal straw, lignin is about 20–25%, 30%, and 12–20 wt %, respectively (Arapova et al. 2020). Along with monolignols, ferulic acids (FA) and p-coumaric acids (pCA) are present in the lignin extracted from sugarcane bagasse (Río et al. 2015).

Classification of Lignin As noted above, the source of lignin, i.e., plant species, has an influence over different lignin structural properties. General classification of lignin depends on plant taxonomy. Angiosperm lignin consists of both syringyl and guaiacyl residues, gymnosperm lignin has more guaiacyl residues, and grass lignin contains all the three aromatic residues (Calvo-Flores and Dobado 2010). This classification of lignin greatly differs; therefore, chemical approach is used to classify lignin. There

59 Lignin: A Renewable Chemical Feedstock

1533

are mainly four types of lignin based on abundance of the basic phenol units p-hydroxyphenyl (H), syringyl (S), and guaiacyl (G) in the polymer, namely, H-G type, H-G-S type, G-S type, and G type (Calvo-Flores and Dobado 2010).

Lignin as a Renewable Feedstock Lignin is the most promising abundant carbon source after cellulose. The physicochemical properties and composition of lignin make it a promising biomaterial to be utilized for energy production. It can serve as a raw material for producing not only arenes but also synthesize gas such as hydrogen, which are valuable petrochemical products (Edwards et al. 2008). Grafted polymers can also be synthesized by modifying the lignin molecules using target fragments, such as antibodies or peptides. Lignin can be utilized as a heavy metal sequestrant, emulsifier, and dispersant agent. Significant scope for diverse applications (Fig. 3) is principally segmented as aromatics, macromolecules, and energy/power. There is immense scope and opportunities to turn lignin into biofuel as companies are readily investing in the research and development for related applications (Mandlekar et al. 2018).

Fig. 3 Uses of lignin

1534

U. Ejaz and M. Sohail

Traditional Use of Lignin Lignin plays a key role in the manufacturing of low-value chemicals. Currently, lignin is largely used for fuel generation, although there are other diversified applications of lignin which include the use of Kraft lignin or lignosulfonates for oil well drilling, binder and resin composition, dispersants, food and feed additives, dust control, and concrete admixture. Lignin is also used as rubber additives, battery expanders, and emulsifiers and used for leather tanning, cleaning chemicals, and water treatment (Agrawal et al. 2014).

Advanced Use of Lignin Lignin-based carbon fiber is one of the brightest examples of successful lignin upgrading technology which has been developed to pilot scale (Baker and Rials 2013). The use of upgraded technical lignins in polymer composites is another example of nontraditional lignin application. Furthermore, production of valuable oxygenated aromatic compounds by depolymerization of lignin, such as olefin in replacement of petrochemicals, is also an example of emerging nontraditional lignin applications (Pandey and Kim 2011). Utilization of lignin’s functionality in polymer composites is perhaps one with the large market potential in future lignin demand.

Use of Lignin Derivatives Lignin is known as a free-radical scavenger. Previously, water-soluble lignin derivatives showed antiviral activity in vitro (Calvo-Flores and Dobado 2010), whereas steam-exploded or Kraft lignin exhibited antioxidant activity in human red blood cells (Vinardell et al. 2008). Lignin-based biopolymers also possess antioxidant capacity. Anticarcinogenic and antibiotic activities of other lignin derivatives have also been reported. Alcell, a sulfur-free lignin, functioned as prebiotic in monogastric animals. Moreover, this type of lignin improved the morphological structures in the intestines and improved the growth of beneficial bacteria. Lignin dietary fiber can also absorb bile acid and hence affect the lipid metabolism (Calvo-Flores and Dobado 2010).

Use of Lignosulfonates Lignosulfonates are considered as nonhazardous for environment and human body. There are many applications of lignosulfonates. They can be used as emulsifiers, dispersant agents for pesticides, and binders (Azadi et al. 2013). Lignosulfonates possess chelating property and, therefore, have a remarkable affinity for some metal ions and can be used as heavy metal sequestrants. In addition, lignosulfonates

59 Lignin: A Renewable Chemical Feedstock

1535

stabilize the emulsions of immiscible liquids including dyes, pigments, wax emulsions, pesticide formulation, and asphalt emulsion and, hence, prevent the breakdown of these materials. Moreover, these can also be used as a binder in soil stabilizers or animal feed due to their nonhazardous nature and low toxicity. Furthermore, lignin binders are used in particle boards or plywood, briquetting of mineral dust (turnings, shavings, fines), and ceramics or coal briquettes (Zakzeski et al. 2010). It also prevents settling and lumping of unsolved particles in mixture of concrete; is also used in gypsum board, ceramics, and leather tanning; and has other applications. Lignosulfonate can also serve as dust suppressor and moistureretention agent and has pronounce application to decrease dust particles in racing circuits, sports facilities, airports, unpaved roads, or public works (Calvo-Flores and Dobado 2010).

Use as Copolymer Component Addition of lignin to improve the mechanical properties of copolymers has been investigated for decades; however, practical results of this research are still modest. There are mainly two ways to use lignin as a copolymer component: First is the direct incorporation of lignin to polymer without any chemical modification, and, second, lignin is chemically modified prior to its use with polymer. Lignin can be transformed by nitroxide formation, phosphorylation, silylation, sulfonation, sulfomethylation, oxidation and reduction, methylolation, hydrogenolysis, nitration, halogenation, acylation, carboxylation, amination, oxyalkylation, dealkylation, or alkylation due to its aromatic nature (Thielemans and Wool 2005).

Extraction of Lignin Many lignin preparations are used for different purposes as they possess different physical properties and chemical functionalization. There are different lignin extraction methods such as Kraft, lignosulfonate, alkaline, organosolv, steam, and alkaline processes (Fig. 4). Significant changes occur in the structure of lignin during its separation from other components. Moreover, sulfur becomes part of lignin during lignosulfonate pulping and Kraft processes (Table 1) which are most frequently used (Kumar et al. 2020). Table 1 Types of lignin (Kumar et al. 2020) Type Organosolv Lignosulfonate Kraft

Pretreatment method Acid Acid Alkaline

Purity High Low Moderate

Sulfur amount Free High Moderate

Production scale Pilot Industrial Industrial

1536

U. Ejaz and M. Sohail

Fig. 4 Process of lignin extraction

Kraft Process Kraft process is used worldwide to convert wood into wood pulp. Kraft (sulfate) pulping accounts for about 85% of lignin production (Tejado et al. 2007). There are two types of reaction occurred to lignin during Kraft pulping: (1) conversion to smaller lignin fragments during decomposition reaction and (2) formation of bonds resistant to cleavage in alkalis that decreases the efficiency of processing during condensation reaction. Cellulose gets separated, and hemicellulose and lignin come to a solution of NaOH/Na2 S (Kraft white liquor) (Zakzeski et al. 2010; Galkin and Samec 2016). Due to release of organic acids from lignin and hemicellulose, pH gradually decreases. After preliminary exposure of solution at 170 ◦ C, dissolved inorganic and organic compounds are separated as a basic aqueous layer containing 15% solids (weak black liquor). Degradation and condensation reactions resulted in various structural and chemical changes in lignin (Azadi et al. 2013) such as percentage of strong C–C bonds between propylphenol monomers becomes much higher (Chakar and Ragauskas 2004). Acidifying black liquor resulted in high yield of Kraft lignin, whereas only ∼2% of wood processing enterprises produce commercial Kraft lignin (Galkin and Samec 2016). In contrast, most of the wood enterprises produce steam and electricity by burning the organic component of black liquor in waste heat boilers (Arapova et al. 2020). Kraft lignin has low molecular mass (1000–3000 Da) as compared to original lignin, but it may reach to 15,000 Da (Galkin and Samec 2016; Glasser et al. 1993). It contains 1.5–3.0 wt % sulfur in its structure which is much lower than the lignin obtained from lignosulfonate cellulose (4–8%) (Arapova et al. 2020), but it is soluble only at pH > 10 and is also hydrophobic in nature (Calvo-Flores and Dobado 2010).

Lignosulfonate Process Lignosulfonate process is commonly employed in paper and pulp industry (Zakzeski et al. 2010). Waste liquid from softwood is used to produce lignosulfonates. Various salts of sulfurous acid (bisulfites or sulfites) are used for lignin extraction from

59 Lignin: A Renewable Chemical Feedstock

1537

Table 2 Sulfite-pulping procedures for lignin extraction (Pye 2008; Bykov 2008; Calvo-Flores and Dobado 2010) Process Anthraquinone/alkaline sulfite Neutral sulfite Bisulfite Acid sulfite

Temperature (◦ C) 50–175 150–175 150–175 125–145

pH 9–13 6–7 3–5 1–2

Reactive agent(s) Na2 SO3 HSO3 − /SO3 2HSO3 SO2 /HSO3 -

wood during sulfite-pulping procedure (Calvo-Flores and Dobado 2010). Sulfitepulping procedure can be performed in different conditions (Table 2). It includes a reaction between sulfur dioxide, metal sulfite, and lignin. The frequently used sodium, magnesium, or calcium ions act as counterions and make the pH of the solution neutral (Azadi et al. 2013). Depending on the concentration and chemical composition of compound, pH can vary from 2 to 13 (Table 2). The extraction process is usually carried out for 1–5 h at 120–180 ◦ C (Arapova et al. 2020). Lignosulfonates have both hydrophilic and hydrophobic characteristics. The sulfur content of the lignosulfonates can vary from 4% to 8% (Arapova et al. 2020). Lignosulfonate has higher molecular mass than Kraft lignin due to addition of sulfonate groups into the lignin polymer, and the extraction process is also less aggressive as compared to Kraft pulping (Calvo-Flores and Dobado 2010). Extracted lignin from hardwood or coniferous species has a molecular weight of 1000–50,000 Da, respectively (Zakzeski et al. 2010; Galkin and Samec 2016). However, high concentrations of sulfur and other pollutants, increased ash content, and formation of new ´–´ bonds are the disadvantages associated with lignosulfonate process (Galkin and Samec 2016).

Organosolv Process Lignin is extracted by organic-solvent-based procedure known as organosolv lignin (Calvo-Flores and Dobado 2010). Peels remaining after fruit juice extraction or fibrous wood residues are mainly used to obtain organosolv lignin (Arapova et al. 2020). High pressure/temperature is used for lignin extraction by using water/organic solvent solution (Calvo-Flores and Dobado 2010). Different organic solvents used for organosolv process are given in Table 3. Alcohols are the most widely used solvent, together with other mixtures of reagents and solvent. Particularly, ethanol, methanol, acetic acid, and formic acid are used, and the treatment is given at 170–190 ◦ C (Tejado et al. 2007). Destruction of α-O-4 lignin– carbohydrate bonds has occurred in organosolv process, whereas β-O-4 bonds are less prone to cleavage (Galkin and Samec 2016). The ability to organize the separate streams of lignin, hemicellulose, and cellulose is the main advantage of using organosolv process (Arapova et al. 2020). Organosolv lignin has a molecular mass of 500–5000 Da and has a high degree of purity due to low carbohydrate and ash content (Galkin and Samec 2016; El Hage et al. 2010). Organosolv lignin is

1538

U. Ejaz and M. Sohail

Table 3 Solvent used for organosolv process (Calvo-Flores and Dobado 2010; Kline et al. 2010; Glasser et al. 1993) Solvent Methanol pulping, followed by methanol, NaOH, and anthraquinone pulping Hydrogen peroxide/formic acid Water/formic acid/acetic acid Water/acid/phenol Alkaline sulfite/anthraquinone/methanol HCl pulping/acetic acid Water/acetic acid Water/ethanol

Lignin type Organocell Milox Formacell Geneva phenol/Battelle ASAM Acetosolv Alcetocell Alcell

mostly sulfur-free and has less modified structure than Kraft lignin. Lignin structure remained preserve in this process; therefore, this procedure can be opted as the most effective method of lignin extraction (Galkin and Samec 2016; Azadi et al. 2013). Moreover, severe conditions and sulfur-containing compounds are not used in organosolv process which make it more environmentally friendly as compared to Kraft lignin (Arapova et al. 2020). Although organosolv process has many advantages, only Organocell and Alcell lignins are commercially available (CalvoFlores and Dobado 2010). High cost of solvent, extensive corrosion of equipment, and lower quality of the resulting pulp are the main disadvantages of organosolv process (Doherty et al. 2011). Furthermore, random condensation may occur with increasing acidity instead of depolymerization (Galkin and Samec 2016).

Use of Steam Lignin is partially hydrolyzed when biomass is treated at high temperature (180– 200 ◦ C) and pressure with steam in the presence of some chemicals. These conditions resulted in a water-insoluble and low molecular weight lignin with less impurities and carbohydrate. This process is often followed by enzymatic hydrolysis (Arapova et al. 2020).

Alkaline Pulping This process is used for the processing of non-wood materials such as sugarcane, grass, and straw (Galkin and Samec 2016). Alkaline pulping is analogous to Kraft process. Sodium hydroxide is used to pretreat plant biomass at 140–170 ◦ C. Oxidation of aliphatic hydroxyl groups resulted in carboxyl group formation which makes it difficult to separate alkaline lignin by centrifugation or filtration due to high concentration of carboxyl groups (Doherty et al. 2011). Alkaline lignin has a molecular weight of 1000–3000 Da depending on the type of plant origin (Azadi

59 Lignin: A Renewable Chemical Feedstock

1539

et al. 2013; Prinsen et al. 2013). The use of an alkali/anthraquinone mixture or alkaline medium seems preferable to other lignin extraction method as sulfur is not present in the precipitated lignin. It is noteworthy that the presence of sulfur in lignosulfonate and Kraft lignins has always been an obstacle for further catalysis of these lignins (Doherty et al. 2011).

Role of Ionic Liquids in Lignin Chemistry Other lignin-target pretreatments such as ionic liquid pretreatment have been thoroughly studied in recent years (Ejaz and Sohail 2020b; Zhao et al. 2020; Ejaz et al. 2019). Ionic liquids (ILs) are organic salts and have many advantages such as recyclability, low flammability, low vapor pressure, and high thermal stability (Pin et al. 2020). Selective extraction of lignin from plant biomass by using ILs has also been achieved in previous years. Previously, 1-ethyl-3-methylimidazolium chloride-based ILs were used to extract lignin from sugarcane bagasse at 190 ◦ C, and 93% lignin and 46% cellulose pulp were recovered. Moreover, small particles (0.1250.250 mm) of both hardwood and softwood could be completely dissolved in 1-ethyl-3-methylimidazolium acetate (Lan et al. 2011). For conversion and surface modification of lignin to other products, ILs can be used as a solvent or catalyst. IL which can be used for lignin extraction must possess the following properties: (1) simple and low-cost process, (2) easy lignin regeneration, (3) no lignin decomposition, (4) chemically stable, (5) nontoxic, (6) nonvolatile, (7) good thermal stability, (8) low melting point, and (9) high dissolution capacity for lignin (Olivier-Bourbigou et al. 2010). The structure of lignin can be modified through hydroxyalkylation and amination reactions by using protic ILs (Pin et al. 2020) which resulted in targeted modified lignin and opens new pathways of lignin valorization (Wang et al. 2018). The nature of anions present in IL affects the solubility of lignin. Pu et al. (2007) used the imidazolium-based ILs for dissolution of residual softwood lignin and reported that ILs containing non-coordinating, large anions like [BF4 ]− and [PF6 ]− are not effective for dissolving lignin. Lee et al. (2009) reported that 1-ethyl-3-methylimidazolium acetate can be used for lignin extraction without much disruption of the hemicellulose and cellulose structures. Moreover, Fu and Mazza (2011) also reported that in 1-ethyl-3-methylimidazolium, acetate is effective for lignin extraction from triticale straw, flax shives, and wheat straw. Pinkert et al. (2011) used imidazoliumacesulfamate ILs, for extraction of lignin from E. nitens wood and P. radiata flour without disrupting crystallinity of cellulose. Lignin can also be extracted after complete dissolution of biomass. Lateef et al. (2009) isolated 69% and 49% lignin by using 1-propyl-3-methylimidazolium bromide and 1-ethyl3-methylimidazolium diethylphosphate, respectively. However, Sun et al. (2009) dissolved both hardwood and softwood in 1-ethyl-3-methylimidazolium acetate, and lignin was extracted using acetone/water (1:1 v/v). Ejaz et al. (2020b) used methyltrioctylammonium chloride for lignin removal from sugarcane bagasse and

1540

U. Ejaz and M. Sohail

argued that ammonium-based ILs are less toxic than imidazolium-based IL and are also water insoluble. Although IL-based lignin extraction offers several advantages, it also has some issues which need to be solved. ILs are more expensive compared to traditional and common solvents. Hence, recyclability of ILs should be more investigated. Recyclability of methyltrioctylammonium chloride for six times retaining its 80% pretreatment efficiency by evaporating the solvent was reported (Ejaz et al. 2020b). In addition, the π–π interaction between lignin and ILs causes problems in separation of lignin from IL (Zakzeski et al. 2010). Therefore, this must also be addressed in future investigations in this area.

Concluding Remarks Lignin which is one of the three components of lignocellulose accounts for 30% of the world carbon source. It is the largest naturally occurring reservoir of aromatic cross-linked polymer. Lignin can be chemically modified to form many valueadded products. It can be used as a heavy metal sequestrant, emulsifier, binder, dispersing agent for pesticides, or component for copolymers. Different processes of lignin extraction such as Kraft, lignosulfonate, organosolv, steam, and alkaline are available. However, each method has its advantages and disadvantages. The research on the usage of ILs for lignin extraction is still at its budding stage, but keen interest in developing this area is having an increasing trend.

References Adler E (1977) Lignin chemistry-past, present and future. Wood Sci Technol 11:169–218. https:// doi.org/10.1007/BF00365615 Agrawal A, Kaushik N, Biswas S (2014) Derivatives and applications of lignin – an insight. SciTech J 1:30–36 Alzagameem A, Klein SE, Bergs M et al (2019) Antimicrobial activity of lignin and lignin-derived cellulose and chitosan composites against selected pathogenic and spoilage microorganisms. Polymers (Basel) 11. https://doi.org/10.3390/polym11040670 Arapova OV, Chistyakov AV, Tsodikov MV, Moiseev II (2020) Lignin as a renewable resource of hydrocarbon products and energy carriers (a review). Pet Chem 60:227–243. https://doi.org/ 10.1134/S0965544120030044 Azadi P, Inderwildi OR, Farnood R, King DA (2013) Liquid fuels, hydrogen and chemicals from lignin: a critical review. Renew Sust Energ Rev 21:506–523. https://doi.org/10.1016/ j.rser.2012.12.022 Bajwa DS, Pourhashem G, Ullah AH, Bajwa SG (2019) A concise review of current lignin production, applications, products and their environment impact. Ind Crop Prod 139:111526. https://doi.org/10.1016/j.indcrop.2019.111526 Baker DA, Rials TG (2013) Recent advances in low-cost carbon fiber manufacture from lignin. J Appl Polym Sci 130:713–728. https://doi.org/10.1002/app.39273 Boerjan W, Ralph J, Baucher M (2003) Lignin Biosynthesis. Annu Rev Plant Biol 54:519–546. https://doi.org/10.1146/annurev.arplant.54.031902.134938

59 Lignin: A Renewable Chemical Feedstock

1541

Brien JAO, Daudi A, Butt VS, Bolwell GP (2012) Reactive oxygen species and their role in plant defence and cell wall metabolism. Planta 236:765–779. https://doi.org/10.1007/s00425012-1696-9 Bykov I (2008) Characterization of natural and technical lignins using FTIR spectroscopy. Master’s Thesis submissted to Lulea University of Technology, Sweden 1–43. Calvo-Flores FG, Dobado JA (2010) Lignin as renewable raw material. ChemSusChem 3:1227– 1235. https://doi.org/10.1002/cssc.201000157 Chakar FS, Ragauskas AJ (2004) Review of current and future softwood kraft lignin process chemistry. Ind Crop Prod 20:131–141. https://doi.org/10.1016/j.indcrop.2004.04.016 Cruz JM, Domínguez JM, Domínguez H, Parajó JC (2001) Antioxidant and antimicrobial effects of extracts from hydrolysates of lignocellulosic materials. J Agric Food Chem 49:2459–2464. https://doi.org/10.1021/jf001237h Doherty WOS, Mousavioun P, Fellows CM (2011) Value-adding to cellulosic ethanol: Lignin polymers. Ind Crop Prod 33:259–276. https://doi.org/10.1016/j.indcrop.2010.10.022 Edwards PP, Kuznetsov VL, David WIF, Brandon NP (2008) Hydrogen and fuel cells: towards a sustainable energy future. Energy Policy 36:4356–4362. https://doi.org/10.1016/ j.enpol.2008.09.036 Ejaz U, Sohail M (2020a) Supporting role of lignin in immobilization of yeast on sugarcane bagasse for continuous pectinase production. J Sci Food Agric. https://doi.org/10.1002/ jsfa.10764 Ejaz U, Sohail M (2020b) Ionic liquids: green solvent for biomass pretreatment. In: Nanotechnology-based industrial applications of ionic liquids. Springer Nature, Switzerland. pp 27–36 Ejaz U, Muhammad S, Ali FI et al (2019) Methyltrioctylammonium chloride mediated removal of lignin from sugarcane bagasse for thermostable cellulase production. Int J Biol Macromol 140:1064–1072. https://doi.org/10.1016/j.ijbiomac.2019.08.206 Ejaz U, Muhammad S, Hashmi IA et al (2020a) Utilization of methyltrioctylammonium chloride as new ionic liquid in pretreatment of sugarcane bagasse for production of cellulase by novel thermophilic bacteria. J Biotechnol 317:34–38. https://doi.org/10.1016/j.jbiotec.2020.04.013 Ejaz U, Muhammad S, Imran F et al (2020b) Cellulose extraction from methyltrioctylammonium chloride pretreated sugarcane bagasse and its application. Int J Biol Macromol 165:11–17. https://doi.org/10.1016/j.ijbiomac.2020.09.151 El Hage R, Brosse N, Sannigrahi P, Ragauskas A (2010) Effects of process severity on the chemical structure of Miscanthus ethanol organosolv lignin. Polym Degrad Stab 95:997–1003. https:// doi.org/10.1016/j.polymdegradstab.2010.03.012 Fu D, Mazza G (2011) Aqueous ionic liquid pretreatment of straw. Bioresour Technol 102:7008– 7011. https://doi.org/10.1016/j.biortech.2011.04.049 Galkin MV, Samec JSM (2016) Lignin valorization through catalytic lignocellulose fractionation: a fundamental platform for the future biorefinery. ChemSusChem 9:1544–1558. https://doi.org/ 10.1002/cssc.201600237 Glasser WG, Davé V, Frazier CE (1993) Molecular weight distribution of (semi-) commercial lignin derivatives. J Wood Chem Technol 13:545–559. https://doi.org/10.1080/ 02773819308020533 Irmer J (2017) Lignin – a natural resource with huge potential. BIOPRO Baden-württemb GmbH. pp 1–6 Khan MT, Ejaz U, Sohail M (2020) Evaluation of factors affecting saccharification of sugarcane bagasse using cellulase preparation from a thermophilic strain of Brevibacillus. Curr Microbiol. https://doi.org/10.1007/s00284-020-02059-3 Kline LM, Hayes DG, Womac AR, Labbé N (2010) Simplified determination of lignin content in hard and soft woods via UV-spectrophotometric analysis of biomass dissolved in ionic liquids. BioResources 5:1366–1383. https://doi.org/10.15376/biores.5.3.1366-1383 Korányi TI, Fridrich B, Pineda A, Barta K (2020) Development of ‘lignin-first’ approaches for the valorization of lignocellulosic biomass. Molecules 25:2815

1542

U. Ejaz and M. Sohail

Kumar A, Anushree KJ, Bhaskar T (2020) Utilization of lignin: a sustainable and eco-friendly approach. J Energy Inst 93:235–271. https://doi.org/10.1016/j.joei.2019.03.005 Lan W, Liu CF, Sun RC (2011) Fractionation of bagasse into cellulose, hemicelluloses, and lignin with ionic liquid treatment followed by alkaline extraction. J Agric Food Chem 59:8691–8701. https://doi.org/10.1021/jf201508g Lateef H, Grimes S, Kewcharoenwong P, Feinberg B (2009) Separation and recovery of cellulose and lignin using ionic liquids: a process for recovery from paper-based waste. J Chem Technol Biotechnol 84:1818–1827. https://doi.org/10.1002/jctb.2251 Lee SH, Doherty TV, Linhardt RJ, Dordick JS (2009) Ionic liquid-mediated selective extraction of lignin from wood leading to enhanced enzymatic cellulose hydrolysis. Biotechnol Bioeng 102:1368–1376. https://doi.org/10.1002/bit.22179 Li X, He Y, Zhang L et al (2019) Discovery of potential pathways for biological conversion of poplar wood into lipids by co-fermentation of Rhodococci strains. Biotechnol Biofuels 12: 1–16. https://doi.org/10.1186/s13068-019-1395-x Liu C (2012) Deciphering the enigma of lignification: precursor transport, oxidation, and the topochemistry of lignin assembly. Mol Plant 5:304–317. https://doi.org/10.1093/mp/ssr121 Luo H, Abu-Omar MM (2017) Chemicals from lignin. Elsevier, Netherland Mandlekar N, Cayla A, Rault F et al (2018) An overview on the use of lignin and its derivatives in fire retardant polymer systems. In: Lignin – trends and applications. IntechOpen, London, UK Menezes FFD, Rencoret J, Nakanishi SC et al (2017) Alkaline pretreatment severity leads to different lignin applications in sugarcane biorefineries. ACS Sustain Chem Eng 5:5702–5712 Moreno A, Sipponen MH (2020) Lignin-based smart materials: a roadmap to processing and synthesis for current and future applications. Mater Horiz 7:2237–2257. https://doi.org/10.1039/ d0mh00798f Olivier-Bourbigou H, Magna L, Morvan D (2010) Ionic liquids and catalysis: recent progress from knowledge to applications. Appl Catal A Gen 373:1–56. https://doi.org/10.1016/ j.apcata.2009.10.008 Pandey MP, Kim CS (2011) Lignin depolymerization and conversion: a review of thermochemical methods. Chem Eng Technol 34:29–41. Pin TC, Nascimento VM, Costa AC et al (2020) Structural characterization of sugarcane lignins extracted from different protic ionic liquid pretreatments. Renew Energy 161:579–592. https:// doi.org/10.1016/j.renene.2020.07.078 Pinkert A, Goeke DF, Marsh KN, Pang S (2011) Extracting wood lignin without dissolving or degrading cellulose: investigations on the use of food additive-derived ionic liquids. Green Chem 13:3124–3136. https://doi.org/10.1039/c1gc15671c Prinsen P, Rencoret J, Gutiérrez A et al (2013) Modification of the lignin structure during alkaline delignification of eucalyptus wood by kraft, soda-AQ, and soda-O2 cooking. Ind Eng Chem Res 52:15702–15712. https://doi.org/10.1021/ie401364d Pu Y, Jiang N, Ragauskas AJ (2007) Ionic liquid as a green solvent for lignin. J Wood Chem Technol 27:23–33. https://doi.org/10.1080/02773810701282330 Pye EK (2008) Industrial lignin production and applications. In: Biorefineries-industrial processes and Products: Status quo and future directions. WILEY-VCH Verlag GmBH & Co Río C, Lino AG, Colodette JL et al (2015) Differences in the chemical structure of the lignins from sugarcane bagasse and straw. Biomass Bioenergy 81:322–338. https://doi.org/10.1016/ j.biombioe.2015.07.006 Strassberger Z, Prinsen P, Van Der Klis F et al (2015) Lignin solubilisation and gentle fractionation in liquid ammonia. Green Chem 17:325–334. https://doi.org/10.1039/c4gc01143k Sun N, Rahman M, Qin Y et al (2009) Complete dissolution and partial delignification of wood in the ionic liquid 1-ethyl-3-methylimidazolium acetate. Green Chem 11:646–665. https://doi.org/ 10.1039/b822702k Tejado A, Peña C, Labidi J et al (2007) Physico-chemical characterization of lignins from different sources for use in phenol-formaldehyde resin synthesis. Bioresour Technol 98:1655–1663. https://doi.org/10.1016/j.biortech.2006.05.042

59 Lignin: A Renewable Chemical Feedstock

1543

Thakur VK, Thakur MK (2015) Recent advances in green hydrogels from lignin: a review. Int J Biol Macromol 72:834–847. https://doi.org/10.1016/j.ijbiomac.2014.09.044 Thielemans W, Wool RP (2005) Lignin esters for use in unsaturated thermosets: lignin modification and solubility modeling. Biomacromolecules 6:1895–1905. https://doi.org/10.1021/bm0500345 Toh K, Nakano S, Yokoyama H et al (2005) Anti-deterioration effect of lignin as an ultraviolet absorbent in polypropylene and polyethylene. Polym J 37:633–635. https://doi.org/10.1295/ polymj.37.633 Vinardell MP, Ugartondo V, Mitjans M (2008) Potential applications of antioxidant lignins from different sources. Ind Crop Prod 27:220–223. https://doi.org/10.1016/j.indcrop.2007.07.011 Wang B, Chen TY, Wang HM et al (2018) Amination of biorefinery technical lignins using Mannich reaction synergy with subcritical ethanol depolymerization. Int J Biol Macromol 107:426–435. https://doi.org/10.1016/j.ijbiomac.2017.09.012 Zakzeski J, Bruijnincx PCA, Jongerius AL, Weckhuysen BM (2010) The catalytic valorization of lignin for the production of renewable chemicals. Chem Rev 110:3552–3599. https://doi.org/ 10.1021/cr900354u Zhang L, LeBoeuf EJ (2009) A molecular dynamics study of natural organic matter: 1. Lignin, kerogen and soot. Org Geochem 40:1132–1142. https://doi.org/10.1016/j.orggeochem. 2009.08.002 Zhao Z-M, Liu Z-H, Pu Y et al (2020) Emerging strategies for modifying lignin chemistry to enhance biological lignin valorization. ChemSusChem. https://doi.org/10.1002/ cssc.202001401

Relevant Websites Department of Energy [DOE] (2007) Alternative, renewable and novel feedstocks for producing chemicals. Department of Energy [DOE], Washington, DC. Available at: http://www1. eere.energy.gov/manufacturing/pdfs/v2020_alternate_feedstock_report.pdf Grand View Research (GVR) (2015) Lining market analysis. Grand View Research [GVR], San Francisco. Available at: http://www.grandviewresearch.com/industry-analysis/lignin-market NNFCC (National Non Food Crops Centre (UK)) (2011) NNFCC renewable chemicals factsheet: lignin. NNFCC, New York. Available at: http://www.nnfcc.co.uk/publications/nnfccrenewablechemicals-factsheet-lignin US Energy Information Administration [EIA] (2016) Monthly energy review. Available at: http:// www.eia.gov/totalenergy/data/monthly/archive/00351604.pdf

Emerging Technologies in Diagnostic Virology and Antiviral Strategies

60

Goutam Patra and Sumi Mukhopadhyay

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CRISPR-Cas System-Based Detection Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Paper-Based Lateral Flow Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fluid-Phase Immunoassay and Luciferase Immunoprecipitation System Followed by Paramagnetic Technology (LIPSTICKS) Assay . . . . . . . . . . . . . . . . . . . . . . . . Multiplex Microsphere Immunoassay (MIA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sandwich-Type Electrochemiluminescence Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nanoparticle-Based Immunodetection Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Solid-Phase Immunoelectron Microscopy (SPIEM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Loop-Mediated Isothermal Amplification (LAMP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biosensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aptamers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microarray Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transmission Electron Microscope Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MALDI (Matrix-Assisted Laser Desorption Ionization) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Antiviral Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1546 1547 1548 1548 1549 1549 1550 1550 1550 1551 1551 1551 1552 1552 1553 1555 1556

Abstract

Timely and accurate diagnosis of viral infections has an important role in public healthcare management. Accordingly, there is a growing need to update on emerging diagnostics, which would be helpful in identifying newer uncharacterized viruses. With increasing viral pandemics in the past century, development of novel antiviral strategies also poses an important challenge to the current healthcare system. Herein, we describe several recent advancements on rapid

G. Patra · S. Mukhopadhyay () Department of Laboratory Medicine, School of Tropical Medicine, Kolkata, West Bengal, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_97

1545

1546

G. Patra and S. Mukhopadhyay

viral detection and antiviral methods. Among the various diagnostic approaches, CRISPR-Cas system-based detection of viral nucleic acid on a paper-based lateral flow assay, fluid-phase immunoassay technology, and luciferase immunoprecipitation system followed by paramagnetic technology (LIPSTICKS) appear promising. Recently developed multiplex microsphere immunoassay (MIA), sandwich-type electrochemiluminescence, nanoparticle-based immunodetection, solid-phase immunoelectron microscopy (SPIEM), loop-mediated isothermal amplification (LAMP), biosensors, aptamers, and microarray technique also open a new era in viral diagnostics. Apart from these, transmission electron microscope study and MALDI (matrix-assisted laser desorption/ionization) methods are also some sophisticated technologies for detection of uncharacterized viruses. The review will further highlight on the emerging antiviral strategies. MHC class I and class II associated peptide proteomics (MAPPs) is a powerful tool for directly profiling neoantigen. The sequencing data of neoantigen can help in the synthesis of effective peptide vaccines that can stimulate humoral and cell-mediated immune responses. Type I IFN-based antiviral therapy, followed by the usage of stimulator of IFN gene (STINGs) as vaccine adjuvants to boost up vaccine efficacy against viral infection, also appears to be a novel method. Taken together, this review will highlight on our current understanding of these high-throughput viral detection techniques as well as novel antiviral approaches.

Keywords

Diagnostic virology · Vaccine · Virus · CRISPR-Cas system · LIPSTICKS · SPIEM · LAMP

Introduction Viruses have enormous repertoire diversity such that if all the 1 × 1031 viruses on earth were laid end to end, they would stretch for 100 million light years (Mourya et al. 2019). Viruses are evolving at an alarming rate and rapidly spreading. Statistical estimations reveal that there could be over 320,000 mammalian viruses in existence. Till date, 219 virus species are known to be able to infect humans (Kiselev et al. 2020). In recent years, there has been marked increase in the incidence of several viral infections, like SARS, Zika, Coronavirus 2019, etc. However, the variety of available viral diagnostic assays seems unequal to the growing number of diseases, as they are capable of detecting only a limited number of pathogens, approximately 0.07% of viral entities (Kiselev et al. 2020). Human virus infections cause significant morbidity and mortality throughout the world and manifest as acute, chronic, or lifelong infections. The immune status of the patient and their ability to clear virus infection, as well as the characteristics of the pathogen, play a vital role in the establishment of the infection. The finding

60 Emerging Technologies in Diagnostic Virology and Antiviral Strategies

1547

Fig. 1 Evolution of Viral Diagnostics

of viral nucleic acids and antiviral antibodies has greatly helped in the development of improved technologies for the accurate diagnoses of viral disease(s) (Fig. 1). Significant developments in the field of viral diagnostics have improved the precision and timely diagnosis of a viral infection, which has greatly benefited patient management and disease control. As the repertoire of antiviral drugs and treatment options has greatly increased in the past few decades, clinicians now rely greatly on the diagnosis of viral infections, which are useful in clinical decisionmaking and management of patients. To aid the clinicians in providing better and faster diagnostics, point-of-care testing has been developed. Though molecular diagnostics has greatly shaped viral diagnosis, significant treads have also been made in the field of serological assays for the development of rapid tests. Thus, viral molecular and serological tests have been immensely useful for differential diagnosis. The aim of this chapter is to provide an outline of the advances made in the field of viral diagnostics as well as the latest development in the field of antiviral therapy.

CRISPR-Cas System-Based Detection Method The CRISPR-Cas system is the most advanced RNA-based gene therapeutics and currently has diverse potential remedial applications. The clustered regularly interspaced short palindromic repeats (CRISPR) and CRISPR-associated (Cas) nuclease system is a simple yet powerful tool for editing genomes. The technology has allowed altering DNA sequences, thereby modifying gene function. The CRISPR system targets the DNA strands and is guided by a single-guide RNA (sgRNA). The sgRNA, which is bound to the Cas protein, has a custom-designed nucleotide spacer, which targets the genome to be modified. The spacer next hybridizes with the target genomic sequence which lies next to a protospacer adjacent motif (PAM), thereby cleaving the target DNA, causing a double-strand break (DSB). The Cas-mediated DSB is then repaired by the cellular DNA repair machinery either via homologydirected repair (HDR) or the non-homologous end joining (NHEJ) pathway. Thus,

1548

G. Patra and S. Mukhopadhyay

the NHEJ is usually used to insert or delete or inactivate the target gene, while HDR is used for precise nucleotide sequence modifications; all processes are mediated by the Cas endonuclease. There are 29 CRISPER/Cas systems recognized till now (Makarova et al. 2015). Out of these, the Cas12a is created specifically for the viral DNA. The reporter and quencher are present on both ends of the ssDNA. After the action of Cas12a, the fluorophore reporter releases the fluorescence, which is then identified by the quencher molecule; the DNA is then further amplified by polymerase chain reaction. However, different kinds of the CRISPER system have been developed like CRISPR screen based on the plasmid library, lentiviral-based CRISPR screen-based sgRNA plasmid pool, etc. These techniques are being used for the diagnosis of several viral infections like human papillomavirus infection.

Paper-Based Lateral Flow Assay The lateral flow assay (LFA) is a modern, advanced, and rapid field adaptable technique. The method is based on a one-site immunometric assay using a monoclonal antibody (MAb) against pathogen for the identification of specific antigen or antibody from the infected sample. MAb is bound with colored gold nanoparticles and appears as a test line on a polyvinylidene fluoride (PVDF) or nitrocellulose (NC) membrane. Targeted antigens present in the collected sample combine with the gold nanoparticles and form antibody-antigen-nanoparticle conjugate complexes, which move forward along the NC or PVDF membrane. Then, the complexes are captured by the MAb present in the membrane, appearing as a colored line(s) in the test line. This rapid test permits for on-site determination of infection in the case of a speculated disease epidemic. Further, it is also a useful technique for the detection of suspected infected patients from the endemic region. Previously, several groups of scientists used this technique for the detection of a pathogen like hand, foot, and mouth disease virus (Oem et al. 2009), malaria (Reboud et al. 2019), influenza (Hwang et al. 2018), dengue, chikungunya, human immunodeficiency virus, COVID-19, etc. However, the disadvantages of this method are low sensitivity (Sher et al. 2017) and test variability (Sajid et al. 2015), and also the test output is qualitative instead of quantitative (Quesada and Merkoçi 2015).

Fluid-Phase Immunoassay and Luciferase Immunoprecipitation System Followed by Paramagnetic Technology (LIPSTICKS) Assay Fluid-phase immunoprecipitation (FPIP)-based methods offers an extended range for detection and exhibit excellent sensitivity and specificity level for evaluating the levels of specific antibodies. This technique applicability appears to be ideal for point-of-care treatment (POCT). The uniqueness of the FPIP assay is that all the epitopes of native antigens are kept intact, enabling the proper detection of antibodies. However, the main disadvantage of this method is it utilizes biohazardous radiolabelled isotope for the detection of the pathogen. That is practically impossible

60 Emerging Technologies in Diagnostic Virology and Antiviral Strategies

1549

in the field-stage diagnosis. Thus, a modification of fluid-phase immunoassay technology has been introduced, which is called luciferase immunoprecipitation systems (LIPS). LIPS technique uses light-emitting luciferase-antigen blending proteins instead of the radiolabeled isotope. This technique has also exhibited high specificity and sensitivity for detecting immunoglobulin for specific autoimmune and infectious diseases. Previously, Burbelo et al. have detected HIV and EpsteinBarr virus using this LIPS method (Burbelo et al. 2017). This assay may be conducted in different ways, like in microplate (Burbelo et al. 2009), single tube (Burbelo et al. 2005), and microfluidic platform (Zubair et al. 2011). However, the LIPS technique is impractical for the field as it requires more than 2 h for completion. Thus, a further modification has been made to LIPS technique to make it suitable for POCT by combining luciferase-tagged antigens with the direct capture of immune complexes on the end of a paramagnetic “sticks” or protein A/Gcoated beads. This enables a rapid detection of disease-specific antibodies, thereby reducing the assay time.

Multiplex Microsphere Immunoassay (MIA) This assay consists of internally dyed microspheres, the surface of which have a suspension array of specific capture moieties. Subsequent to binding reactions on the surface of these microspheres, analyte binding is detected using a reporter fluorophore. These microspheres are then analyzed through a Luminex analyzer to quantitate the surface analyte binding, as detected by the fluorophore emission. Microsphere-based immunoassays (MIAs) have been utilized for screening multiple viruses in a single sample, though, previously, CDC has used dual-IgM antibody test for the identification of St. Louis encephalitis (SLE) and West Nile (WN) viruses (Johnson et al. 2005). Subsequently, Alison et al. reported detection of six alphaviruses, six flaviviruses, and one bunyavirus through MIA (Alison and Kalanthe 2013). The advantage of this method is the need of very small amount of sample that could be concurrently used for testing toward various viral antigens. Similarly, the diagnosis of the Zika and dengue virus simultaneously by using flavivirus MIA has many advantages. However, the major drawback of this method is that it cannot discriminate new from past infections.

Sandwich-Type Electrochemiluminescence Assay Among the different assays known for the detection of pathogenic virus in the patient’s body, sandwich-type electrochemiluminescence (ECL) is one of the most promising assays. ECL immunosensor technique has been developed for specific and ultrasensitive determination of different viral surface antigens like the hepatitis virus and HIV-1. This technique is based on the primary immunoglobulin, which is raised against the virus and immobilized on the surface of carboxyl magnetic nanoparticles (MNPs). After that, dendrimer with many amine functional groups

1550

G. Patra and S. Mukhopadhyay

is applied as a vehicle for immobilizing the quantum dots (QDs). Finally, the secondary antibody amplifies the ECL beacon of QDs (Babamiri et al. 2018). This technique is considerably more sensitive and specific for viral disease diagnosis. Similarly, influenza A subtype virus H1N1 has also been detected from the upper respiratory region using this technique.

Nanoparticle-Based Immunodetection Assay Nanoparticles have opened a new era in the field of biomedical and diagnostic applications. Nanoparticles are used in some modern detection approaches based on lateral flow immunoassay (LFIA) as a diagnostic tool. Currently, LFIA is also being used for the detection of infectious agents and diseases. LFIA methods are paper-based and point-of-care assays; it is rapid and a field-adaptable stripbased biosensor designed to identify a particular analyte in virus-affected samples (Banerjee and Jaiswal 2018). Previously, Draz et al. reported that the AuNP-based detection technique could be applied for the detection of several pathogens (Draz and Shafiee 2018).

Solid-Phase Immunoelectron Microscopy (SPIEM) Immunoelectron microscopy (IEM) is one of the most assuring techniques for the discovery of the virus based on antigen and antibody reaction. The structure of various viruses has been differentiated morphologically by using IEM, like rotavirus A, B, and C. Solid-phase IEM (SPIEM) is a current modification of IEM; it is based on viral elements that are captured instantly over the solid covering of a grid. This technique is rapidly being used for the identification of norovirus by using a specific antibody (Lewis et al. 1988). Generally, 50,000–60,000 X magnification is being used for the observation of virus particles. Expensive instrumentation and low sensitivity are the main disadvantages of SPIEM.

Loop-Mediated Isothermal Amplification (LAMP) Loop-mediated isothermal amplification (LMAP) is one of the rapid methods that specifically amplify the target genetic material at 60 ◦ C. However, this method is an inexpensive but also accurate diagnostic method as the method is based on the application of a set of nucleic acid-specific primer. The chance to lose chemical and sample is low as the whole experiment occurs in a separate tube and less than 30 min is needed for the experiment. Reverse transcriptase loop-mediated isothermal amplification (RT-LAMP) is a modified version that helps in the identification of bovine rotavirus, porcine epidemic diarrhea (PED), a novel swine acute diarrhea syndrome-coronavirus (SADS-CoV), etc. (Yu et al. 2015; Wang et al. 2018). After that, the LAMP technique has undergone several modifications for better and

60 Emerging Technologies in Diagnostic Virology and Antiviral Strategies

1551

field adaption like paper-based LAMP (paper LAMP), LAMP lateral flow dipstick (LAMP-LFD), immunocapture LAMP (IC-LAMP), etc. The sensitivity/specificity and simplicity of these techniques were more superior as compared to conventional PCR. Hence, these methods could be adopted for the detection of various viruses.

Biosensors Biosensors help in the rapid detection of viruses from a variety of specimens. The main aspects for the construction of biosensors are dependent on the selection of the membrane, like nitrocellulose membrane, sodium azide, and polyether sulfonate/nylon membrane. The main goal of all biosensors is to identify a complete virus particle or virus that produces any protein component. Previously, biosensors have been developed for several viruses like bovine viral diarrhea (Luo et al. 2010; Park et al. 2014), rotavirus (Pineda et al. 2009), human norovirus (Connelly et al. 2012; Escudero-Abarca et al. 2014), etc. Further, the identification of hepatitis and rotavirus molecular device biosensors has been revealed by Zhang et al. (2015).

Aptamers The aptamer is a small fragment of oligonucleotide or polypeptide that acts as a substitute form of antibodies. For an efficient and reliable diagnosis, aptamerlinked immunosorbent assay (ALISA) has been developed like ELISA. The main advantages of ALISA are high sensitivity and quick detection (Kieboom et al. 2015). Through this assay norovirus strain, Bovine viral diarrhea has also been identified from the reservoir (Escudero-Abarca et al. 2014). Like antibodies, aptamers are thermostable and more economical, so it could easily be used in resource-constricted settings (Song et al. 2012).

Microarray Technique Microarrays are extremely appropriate for bulk investigation of viral samples and should also be utilized toward trailing the reservoirs in the endemic regions for viral disease. These methods simultaneously detect larger than a hundred kinds of viruses at a time (Wang et al. 2002). Interestingly, Martínez et al. developed a nucleic acid microarray, and simultaneously 14 gastrointestinal viral infections have been examined toward the affirmation of his experiment (Martínez et al. 2015). Further, Buss et al. developed a film array technique based on self-nucleic acid purification, reverse transcription, and amplification (Buss et al. 2015). Currently, DNA microarray platforms served in the genotyping of hepatitis A viruses and norovirus (Quinones et al. 2017).

1552

G. Patra and S. Mukhopadhyay

Transmission Electron Microscope Study Electron microscopy (EM) is the pioneer of the identification and characterization of several viruses. The tobacco mosaic virus is the first virus that was detected by EM in 1970. Several gastrointestinal viral infections are difficult to identify only through their etiology. Hence, EM is the only way to detect these viruses. In the year 1976, sapovirus was for the first time detected by EM and differentiated from another rotavirus, parvovirus, etc. Further, an EM study was deployed for regular detection of the gastric enteric virus as EM result could be obtained on the same day of sample collection, though the sample amount requirement was high. Currently, the COVID-19 virus was initially detected and characterized by EM. Transmission electron microscopy (TEM) is one great beginning in the study of viruses for many causes, including the fact that, given tissue, storage is not the main issue, and this method can work uniformly well on current, aged, partly deteriorated, or dried samples (Huger 1967, 1974; Koenig and Lesemann 2001; Gentile and Gelderblom 2014). TEM is a classical technique and its advantage is direct visualization of morphology of virus. The merged use of both TEM and scanning electron microscope (SEM) reveals the characterization of several objects, such as baculovirus (Gencer et al. 2018). TEM allows high visual resolution of the morphology of virus particle, localization, quantity of virus particles, and specific ultrastructural aspects within enclosed occlusion of the prepared sample. However, SEM investigations are perfect for high-output screening of specimens and diameter measurements for the differentiation between various viral isolates (Gencer et al. 2018). Microbiologists have also utilized the electron microscope for the detection of enteric microbes (Leland and Ginocchio 2007), adenovirus, rotavirus (Morinet et al. 1984; Oka et al. 2015), etc.

MALDI (Matrix-Assisted Laser Desorption Ionization) The spectrometry analysis of viral disease diagnosis has produced a notable achievement. Mass spectroscopy techniques have developed through the efforts of Dempster (1918) and Aston (1919). After that, this technique was successfully being used for viral disease diagnosis at the proteomic levels. In these circumstances, the pathogen could be identified by mass spectrometry-based identification, such as matrix-assisted laser desorption ionization time-of-flight mass spectrometrybased systems (MALDI-TOF MS). Further, ions are produced in the specimens and then separated by their size-to-charge ratio identified on a detector. After that, the result is investigated and correlated with a directory database. The mass spectrometry assay is more reliable than conventional methods for the detection of co-infection with many viruses in a particular specimen. Another method that has been used for the rapid and particular detection of several enteric viruses is surfaceenhanced Raman spectroscopy (SERS), which is a Raman spectroscopic (RS) assay. These methods provide considerably enhanced Raman signals from Raman-active

60 Emerging Technologies in Diagnostic Virology and Antiviral Strategies

1553

analyte molecules that have adsorbed onto several specially prepared metal surfaces. Previously, Driskell et al. (2010) identified rotavirus strains by using SERS. The disadvantage of these techniques is that they are more time-consuming and have dependence on the skills of the operator and there is a need for self-regulation and for automation.

Antiviral Strategies Vaccination During the past decade, several viral infections have caused a huge global disease burden; many outbreaks took place, largely caused by the avian influenza virus and the SARS and MERS coronaviruses, Zika, and Ebola viruses. Most of these RNA viruses from different families have been a matter of great concern and can cause very high rates of mortality (>60%). The influenza A viruses in particular have been a global threat to public health, as they are rapid spreaders due to their antigenicity and virulence. The pandemic during the early 1900 is believed to be due to these behaviors of the virus, which caused millions of deaths (Joshi et al. 2013). Vaccines have an important role in the management and prevention of diseases. Every year, millions of children worldwide receive vaccinations to prevent serious diseases that were once prevalent and linked to even death. Some of the wellknown vaccines include those for measles, mumps, rubella, tetanus, polio, hepatitis B, diphtheria, pertussis, and others. Estimations reveal that vaccination usually prevents about two to three million deaths annually (WHO). During the nineteenth and early twentieth century, critically important live pediatric vaccines for the prevention of viral measles, mumps, and rubella became available. The attenuation of the constituent virus of these vaccines is largely by experimental means, including practices like numerous passages in cell culture or embryonated eggs. These procedures, however, produce modest levels of vaccineassociated side effects and thus have been overall useful. The near global elimination of paralytic poliomyelitis under WHO-sponsored programs over the past four decades was one of the biggest success stories in the field of vaccinology. During the past few years, measles, mumps, and rubella vaccine viruses have been substantially improved which fortunately are stable. This includes the substitution with strains that are highly immunogenic and/or are related with fewer adverse reactions. Improvement in the vaccination regimes and the introduction of trivalent measles, mumps, and rubella (MMR) and quadrivalent measles, mumps, rubella, and varicella (MMRV) vaccines have greatly controlled these viral infections. Apart from these traditional vaccines, more vaccines in the 1990s have been registered, which include oral rotavirus vaccines, intranasal live attenuated influenza vaccines, subcutaneous varicella zoster virus (VZV) vaccines, etc. Though vaccines could be successfully developed for many diseases, still for several diseases, the development of a safe and effective vaccine remains elusive.

1554

G. Patra and S. Mukhopadhyay

Many viruses like HIV-1 or influenza have exceptionally broad sequence diversity among their constituent family members and undergo significant antigenic drift. Successful vaccine development in these viral infections has thus suffered major setbacks. Dengue vaccine development has been particularly challenging as the viral components elicit a complex immune reaction, thereby contributing to disease severity (Malonis et al. 2020). Recent developments in the field of proteomics have paved the way for comprehensive understanding of peptides and protein and their use in vaccines. Cytotoxic T cells are one of the most important weapons of our immune system and are effective against virus-infected cells. T cells are activated and triggered by professional antigen-presenting cells, such as dendritic cells (DCs). Though antiviral drugs are available to treat viral diseases, they are usually associated with side effects and resistance through viral escape. Vaccination, on the other hand, is an effective method to fight viral diseases. Most of the vaccination strategies are based on activating humoral response and are often efficient in preventing virus infection. However, due to the rapidly changing virus surface proteins, the virus often escapes and infects host cells, which are then taken care of by the cellular response. Most of the vaccines against viral infections are based on triggering humoral responses and are usually poor inducers of T-cell responses. As T cells are an important component of the immune system, T-cell-inducing vaccines are much needed. Usually, small protein fragments or peptides are included in a vaccine, which upon presentation by the major histocompatibility complexes on the antigenpresenting cells prime the CD4+ and CD8+ T cells, thereby inducing specific T-cell responses. The first peptide-based licensed vaccine is against HPV. This vaccine contains synthetic peptides along with Montanide, which triggers specific CD4+ and CD8+ T-cell responses in all patients. Klade et al., using the IC41 vaccine in HCVinfected patients, demonstrated decreased viral RNA load, which is composed of five synthetic peptides along with a Th1-type adjuvant, poly-I-arginine. On the other hand, many studies are being conducted for the development of a preventive peptide vaccine. In these vaccines, usually conserved sequences are targeted, thereby leading to the development of a universal vaccine. Many such conserved domains have been identified for viruses like HIV, HCV, and influenza. These preventive peptide-based vaccines aim to elicit the memory T-cell repertoire and thus rapidly remove the virus from circulation. Thus, naïve T cells need to be primed efficiently to induce a pool of both memory CD4+ T cells and memory CD8+ T cells. Thus, professional antigen-presenting cells process and present the antigen on MHC molecules to both CD4+ and CD8+ T cells. Therefore, by identifying the conserved domain of the virus, peptide-based vaccines can be designed, which appears as a promising concept. Currently, several peptide-based vaccines, for viruses like Epstein-Barr virus, hepatitis B virus, and influenza virus, are being evaluated in clinical trials (Rosendahl et al. 2014).

60 Emerging Technologies in Diagnostic Virology and Antiviral Strategies

1555

Though live attenuated vaccines are usually very immunogenic, there have been concerns over their stability and safety. Modern vaccines, therefore, focus on other components like recombinant proteins, peptides, or DNA. However, to increase the effectiveness of vaccines, co-administration of adjuvants is required. Since time immemorial, adjuvants have been used in human vaccines. However, their mechanisms of action, till date, remain incompletely understood. More recent advances in the field of immunology have identified new adjuvants that activate specific immunological pathways. Viruses are usually recognized by cellular pattern recognition receptors (PRRs). Similarly, upon activation of the innate response, adaptive immunity is triggered. Antigen-presenting cells, like dendritic cells, play an important role in this process. DCs are well-equipped with several PRRs that recognize viruses and trigger intracellular pathways and produce cytokines and chemokines. Thus, dendritic cells after presenting antigens usually migrate to lymphoid tissues wherein they communicate with T and B lymphocytes to modulate our adaptive immune response. Therefore, PRR stimulation by adjuvants is an important aspect for assessing vaccine performance. Several PRRs have been identified in humans, including TLRs, RIG-I-like receptors (RLRs), cytosolic DNA sensors (CDs), nucleotide-binding oligomerization domain, and (NOD)-like receptors (NLRs). Most of the current innate immunitybased adjuvants target TLR. The stimulator of IFN genes (STING) recognizes the cyclic guanosine monophosphate-adenosine monophosphate dinucleotide (c-di-GAMP), which further activates IRF3 and NF-κB transcription factors, which promotes DC maturation and Th1 response. Though STING agonists as vaccine adjuvants have shown promising results in mice, their potential application in human vaccines is still a matter of concern. Vaccine performance is increased manifold if several innate receptors can be triggered. Some of the most successful adjuvant systems are AS01, AS02, AS03, and AS04. More recently, AS01 has been incorporated in the malaria vaccine Mosquirix. The adjuvant efficacy of AS01 and AS02 in vaccines against HIV, tuberculosis, and HBV is also being studied. AS03 is being used in H1N1 vaccines (Acosta et al. 2020). Taken together, all these abovementioned adjuvant combinations induce rapid activation of the immune system. Virus vaccines are still evolving and have miles to go to achieve a disease-free world.

Conclusions The past several decades have witnessed the development of numerous sophisticated efficient techniques of virus detection. With the advent of the recent pandemics, the need for rapid sensitive, specific, and reliable assay is on the rise. Along with the discovery of newer pathogenic viral strains, the development of novel antiviral strategies is also the current need of the hour. Rapid technological advancement in the field of diagnostics and therapy promises to fulfill these growing needs.

1556

G. Patra and S. Mukhopadhyay

References Acosta PL, Byrne AB, Hijano DR et al (2020) Human type I interferon antiviral effects in respiratory and reemerging viral infections. J Immunol Res 2020:1372494 Alison J, Kalanthe H, Amanda J et al (2013) Multiplex microsphere immunoassays for the detection of IgM and IgG to Arboviral diseases. PLoS One 8(9):e75670 Babamiri B, Hallaj R, Salimi A (2018) Ultrasensitive electrochemiluminescence immunosensor for determination of hepatitis B virus surface antigen using CdTe@ CdS-PAMAM dendrimer as luminescent labels and Fe3O4 nanoparticles as magnetic beads. Sensors Actuators B Chem 254:551–560 Banerjee R, Jaiswal A (2018) Recent advances in nanoparticle-based lateral flow immunoassay as a point-of-care diagnostic tool for infectious agents and diseases. Analyst 143(9):1970–1996 Burbelo PD, Goldman R, Mattson TL (2005) A simplified immunoprecipitation method for quantitatively measuring antibody responses in clinical sera samples by using mammalianproduced Renilla luciferase-antigen fusion proteins. BMC Biotechnol 5:22 Burbelo PD, Ching KH, Klimavicz CM et al (2009) Antibody profiling by Luciferase Immunoprecipitation Systems (LIPS). J Vis Exp 32:1549 Burbelo PD, Gunti S, Keller JM et al (2017) Ultrarapid measurement of diagnostic antibodies by magnetic capture of immune complexes. Sci Rep 7(1):3818 Buss SN, Leber A, Chapin K et al (2015) Multicenter evaluation of the Bio Fire Film Array gastrointestinal panel for etiologic diagnosis of infectious gastroenteritis. J Clin Microbiol 53:915–925 Connelly JT, Kondapalli S, Skoupi M et al (2012) Micro-total analysis system for virus detection: microfluidic preconcentration coupled to liposome-based detection. Anal Bioanal Chem 402:315–323 Draz MS, Shafiee H (2018) Applications of gold nanoparticles in virus detection. Theranostics 8(7):1985 Driskell JD, Zhu Y, Kirkwood CD et al (2010) Rapid and sensitive detection of rotavirus molecular signatures using surface enhanced Raman spectroscopy. PLoS One 5:e10222 Escudero-Abarca BI, Suh SH, Moore MD et al (2014) Selection, characterization and application of nucleic acid aptamers for the capture and detection of human norovirus strains. PLoS One 9:e106805 Gencer D, Bayramoglu Z, Nalcacioglu R et al (2018) Characterisation of three Alphabaculovirus isolates from the gypsy moth, Lymantria dispar dispar (Lepidoptera: Erebidae), in Turkey. Biocontr Sci Technol 28:107–121 Gentile M, Gelderblom HR (2014) Electron microscopy in rapid viral diagnosis: an update. New Microbiol 37:403–422 Huger AM (1967) Elektronenmikroskopie bei der Diagnose von Insektenkrankheiten. In: van der Laan FA (ed) Proc. Intern. Colloq. “Insect pathology and microbial control”. North Holland, Amsterdam, pp 29–53 Huger AM (1974) Methoden und Bedeutung der diagnose von Insektenkrankheiten/methods and importance of diagnosis of insect diseases. J Plant Dis Protect 81:372–388 Hwang SG, Ha K, Guk K et al (2018) Rapid and simple detection of Tamiflu-resistant influenza virus: development of oseltamivir derivative-based lateral flow biosensor for point-of-care (POC) diagnostics. Sci Rep 8:12999 Johnson AJ, Noga AJ, Kosoy O et al (2005) Duplex microsphere-based immunoassay for detection of anti-West Nile virus and anti-St. Louis encephalitis virus immunoglobulin m antibodies. Clin Diagn Lab Immunol 12:566–574 Joshi VG, Dighe VD, Thakuria D et al (2013) Multiple antigenic peptide (MAP): a synthetic peptide dendrimer for diagnostic, antiviral and vaccine strategies for emerging and re-emerging viral diseases. Indian J Virol 24(3):312–320 Kieboom CH, Beek SL, Meszaros T et al (2015) Aptasensors for viral diagnostics. Trends Anal Chem 74:58–67 Kiselev D, Matsvay A, Abramov I et al (2020) Current trends in diagnostics of viral infections of unknown etiology. Viruses 12(2):211

60 Emerging Technologies in Diagnostic Virology and Antiviral Strategies

1557

Koenig R, Lesemann DE (2001) Plant virus identification. In: Encyclopedia of life sciences. Wiley, Chichester Leland DS, Ginocchio CC (2007) Role of cell culture for virus detection in the age of technology. Clin Microbiol Rev 20:49–78 Lewis DC, Lightfoot NF, Pether JV (1988) Solid-phase immune electron microscopy with human immunoglobulin M for serotyping of Norwalk-like viruses. J Clin Microbiol 26:938–942 Luo Y, Nartker S, Miller H et al (2010) Surface functionalization of electrospun nanofifibers for detecting E. coli O157: H7 and BVDV cells in a direct-charge transfer biosensor. Biosens Bioelectron 26:1612–1617 Makarova KS, Wolf YI, Alkhnbashi OS et al (2015) An updated evolutionary classification of CRISPR–Cas systems. Nat Rev Microbiol 13(11):722 Malonis RJ, Lai JR, Vergnolle O (2020) Peptide-based vaccines: current progress and future challenges. Chem Rev 120(6):3210–3229 Martínez MA, Soto-Del Río Mde L, Gutiérrez RM et al (2015) DNA microarray for detection of gastrointestinal viruses. J Clin Microbiol 53:136–145 Morinet F, Ferchal F, Colimon R, Perol Y (1984) Comparison of six methods for detecting human rotavirus in stools. Eur J Clin Microbiol Infect Dis 3:136–140 Mourya DT, Yadav PD, Ullas PT et al (2019) Emerging/re-emerging viral diseases & new viruses on the Indian horizon. Indian J Med Res 149(4):447–467 Oem JK, Ferris NP, Lee K-N et al (2009) Simple and rapid lateral-flow assay for the detection of foot-and-mouth disease virus. Clin Vaccine Immunol 16(11):1660–1664 Oka T, Wang Q, Katayama K, Saif LJ (2015) Comprehensive review of human sapoviruses. Clin Microbiol Rev 28:32–53 Park JW, Jin Lee S, Choi EJ, Kim J et al (2014) An ultra-sensitive detection of a whole virus using dual aptamers developed by immobilization-free screening. Biosens Bioelectron 51:324–329 Pineda MF, Chan LLY, Kuhlenschmidt T et al (2009) Rapid specfic and label-free detection of porcine rotavirus using photonic crystal biosensors. IEEE Sensors J 9:470–477 Quesada GD, Merkoçi A (2015) Nanoparticle-based lateral flow biosensors. Biosens Bioelectron 73:47–63 Quiñones B, Lee BG, Martinsky TJ et al (2017) Sensitive genotyping of foodborne-associated human noroviruses and hepatitis A virus using an array-based platform. Sensors 17:2157 Reboud J, Xu G, Garrett A et al (2019) Paper-based microfluidics for DNA diagnostics of malaria in low resource underserved rural communities. Proc Natl Acad Sci 116(11):4834–4484 Rosendahl HS, Van BJ, de Jonge J et al (2014) T cell responses to viral infections – opportunities for Peptide vaccination. Front Immunol 5:171 Sajid M, Kawde AN, Daud M (2015) Designs, formats and applications of lateral flow assay: a literature review. J Saudi Chem 19:689–705 Sher M, Zhuang R, Demirci U et al (2017) Paper-based analytical devices for clinical diagnosis: recent advances in the fabrication techniques and sensing mechanisms. Expert Rev Mol Diagn 17(4):351–366 Song KK, Lee S, Ban C (2012) Aptamers and their biological applications. Sensors 12:612–631 Wang D, Coscoy L, Zylberberg M et al (2002) Microarray-based detection and genotyping of viral pathogens. Proc Natl Acad Sci U S A 99:15687–15692 Wang H, Cong F, Zeng F et al (2018) Development of a real time reverse transcription loopmediated isothermal amplification method (RT-LAMP) for detection of a novel swine acute diarrhea syndrome coronavirus (SADS-CoV). J Virol Methods 260:45–48 Yu X, Shi L, Lv X et al (2015) Development of a real-time reverse transcription loop-mediated isothermal amplification method for the rapid detection of porcine epidemic diarrhea virus. Virol J 12:76 Zhang J, Zhao Z, Xu M et al (2015) The establishment of biosensor technology based on F0F1ATPase molecular motor for detection of rotavirus and hepatitis A virus. Biosens J 4:121 Zubair A, Burbelo PD, Vincent LG et al (2011) Microfluidic LIPS for serum antibody detection: demonstration of a rapid test for HSV-2 infection. Biomed Microdevices 13(6):1053–1062

GTAW Application for Additive Manufacturing and Cladding of Steel Alloys

61

Vishvesh J. Badheka, Vijay S. Gadakh, V. B. Shinde, and G. Bhati

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Current Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gas Tungsten Arc Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Progress in GTAW Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hotwire GTAW and Narrow Gap Hotwire GTAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Orbital Hotwire GTAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Activated TIG Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Automated Wire Feeding in GTAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inter Pulse TIG Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . K-TIG Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Twin Wire/Tandem GTA Cladding Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Internal GTAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multicathode GTAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Additive Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cladding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Optimization of Welding Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1560 1561 1563 1564 1564 1565 1565 1565 1566 1566 1567 1569 1569 1570 1573 1576 1577 1577 1578

V. J. Badheka () · G. Bhati Department of Mechanical Engineering, School of Technology, Pandit Deendayal Energy University, Gandhinagar, Gujarat, India e-mail: [email protected]; [email protected] V. S. Gadakh Department of Mechanical Engineering, Amrutvahini College of Engineering, Ahmednagar, Savitribai Phule Pune University, Pune, India e-mail: [email protected] V. B. Shinde Department of Production Engineering, Amrutvahini College of Engineering, Ahmednagar, Savitribai Phule Pune University, Pune, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_109

1559

1560

V. J. Badheka et al.

Abstract

Steel alloys are widely used in many industry sectors namely construction, automobile, and aerospace. With additive manufacturing (AM), complex and rare parts of steel alloys can be manufactured efficiently in days in an economical way. On the other hand, conventional manufacturing methods include timeconsuming and cost-incurring die and mold preparation and postmachining of the parts. The other benefits associated with AM are low cost governed by lowenergy input, low-material wastage, and automation with high strength. The gas tungsten arc welding (GTAW) process produces sound weld with high integrity with the parent metal. Therefore, GTAW has found the place as a method for AM. Additionally, the components in these industries are many times subjected to high wear, corrosion, and abrasion. Their durability at the work is increased by using the cladding process. GTAW is widely used in the cladding process for hard facing and corrosion resistance. With its increased deposition rate using advanced hotwire multi-cathode GTAW processes, high applicability to different metals, defect-free weld, and high strength of weld, among other welding processes like shielded metal arc welding (SMAW), gas metal arc welding (GMAW), plasma metal arc welding (PMAW), laser welding, and electron beam welding (EBW), GTAW is a promising process for AM and cladding process. The deposition rate can be further improved by optimization of weld parameters like type of current supply (DC supply, DC pulsed supply, and AC supply), mean voltage, wire feed rate, frequency, and hotwire current. In this chapter, the study of different advanced hotwire GTAW processes and optimization of weld parameters is explained. Keywords

GTAW cladding · Single cathode GTAW · Additive manufacturing · Hotwire welding · Dual cathode GTAW

Introduction In a global competitive environment, different industries are aiming for the least price for manufacturing of their products. This is closely related to lightweight manufacturing. Most of the time, the production cost is increased by the lightweight technologies due to the new processes and equipment needs (Tisza and Czinege 2018). Steel is the most popular among other materials used in different industries like construction, automobile, and aerospace. According to rankings released by the World Steel Association (WSA) in 2019, China was the world’s largest crude steel producer followed by India, Japan, and the USA. As per the WSA report in 2018, 50 percent of the crude steel production was utilized by the construction industry. According to the American Iron and Steel Institute (AISI), as of now, more than 200 steel grades are available which are three to four times stronger than the latest aluminum alloys available in the market. Similarly, these steels perform better with

61 GTAW Application for Additive Manufacturing and Cladding of Steel Alloys

1561

safety benefits using an existing manufacturing infrastructure and eliminating the major manufacturing cost by the introduction of alternative materials. With additive manufacturing (AM), complex and rare parts of steel alloys can be manufactured efficiently in days in an economical way. On the other hand, conventional manufacturing methods include time-consuming and cost-incurring die and mold preparation and postmachining of the parts. The other benefits associated with AM are low cost governed by low-energy input, low-material wastage, and automation with high strength. The gas tungsten arc welding (GTAW) process produces sound weld with high integrity with parent metal (Shah and Agrawal 2019). Therefore, GTAW has found the place as a method for AM. GTAW is widely used in the cladding process where two metals are joined together by welding to the surface and adding a layer for hard facing and corrosion resistance. The mining components such as high-speed rotating components are subjected to high wear and abrasion. Their durability at the work can be increased by using the cladding process. Hence, GTAW cladding has an important application in the mining industry (Wang et al. 2016). GTAW use in the mining industry can be further extended to AM of tools and components. The other benefits associated with it are low cost governed by low-energy input, low-material wastage, and automation with high strength. However, traditional GTAW has a low deposition rate and travel speed. Recently, many improved variants of GTAW have been proposed to increase the efficiency of the welding process such as GTAW with powder, single cathode single hotwire, single cathode dual hotwire, dual cathode single hotwire, and dual cathode dual hotwire (Egerland et al. 2015). In the hotwire-cladding process, a hotwire is fed into the puddle, where hotwire acts as a secondary source of heat into the plasma and hence increases the deposition rate by pouring more weld material into the clad. This process has better weld quality compared to GTAW and GTAW with a powder (Egerland et al. 2015). Many variants of hotwire GTAW namely single cathode single hotwire, single cathode dual hotwire, dual cathode single hotwire, and dual cathode dual hotwire have increased deposition rate as well as low overspray. The low dilution rate achieved with these processes also improves weld quality by increasing corrosion resistance and hard facing. Also, proper chemistry is achieved with fewer layers of the weld. All these benefits are attractive when cladding applications are considered.

Current Trends Recently, Twin-wire / Tandem GTA and other variants of cladding processes have been developed and researched by Fronius International GmbH, Austria. It is found that the arc pressure varies when two arcs coming from two cathodes are single arcs, overlapping arcs, and coupled arcs as shown in Fig. 1. Figure 2 shows the elemental distribution when different combinations of the cathode and hotwire are used. Hence, it can be seen that the twin hotwire single cathode GTAW has better elemental distribution. Further, at the University of Alberta, Canada, the GTAW process has been utilized for wear-resistant overlays.

1562

V. J. Badheka et al.

Fig. 1 Comparison of arc pressure for a different combination of arcs (Egerland et al. 2015)

Fig. 2 Effect of various combinations of cladding on surfacial elemental distribution (Egerland et al. 2015)

When we talk about AM, there is phenomenal growth in engineering. The startups like Relativity and Desktop Metal are 3D printing even at higher speeds than the classical manufacturing methods. This shows the potential of AM in the digitalization of the industry. In India primarily research in the field of GTAW has been done by the Welding Research Institute (WRI), Bharat Heavy Electrical Limited (BHEL). WRI has developed hotwire GTAW, narrow gap hotwire GTAW, orbital hotwire GTAW, activated

61 GTAW Application for Additive Manufacturing and Cladding of Steel Alloys

1563

Fig. 3 Comparison of arc pressure for single cathode and two cathodes (Shah and Agrawal 2019)

TIG welding, automated wire feed GTAW, etc. Besides WRI, BHEL, Institute of Infrastructure Technology Research and Management (IITRAM), Ahmedabad, has researched twin TIG where two tungsten electrodes are used in one torch. Their research shows that there is a significant effect of many electrodes used on arc pressure. Welding Research laboratory of Pandit Deendayal Energy University (PDEU), Gandhinagar working extensively in the area of Activated- TIG welding and its variants like FB-TIG and FZ –TIG for ferrous, non-ferrous and dissimilar metal combinations. Recently researchers have also developed welding procedure for 16 mm thick plate of copper using hot wire GTAW. The arc pressure curve gets flattened when two cathodes are used as shown in Fig. 3.

Gas Tungsten Arc Welding Fusion welding processes are divided into three types as flux-shielded welding processes, gas-shielded welding processes, and beam-welding processes. Each of these processes has unique features and finds a wide range of applications. The gas-shielded welding processes – GTAW, GMAW, and plasma metal arc welding (PMAW) – are extensively employed in almost all fabrication industries. Among these processes, GTAW is suitable to weld almost all metals and for uniform root penetration. Due to the development of newer materials, new variants of GTAW processes like hotwire GTAW activated TIG, continuous wire feed GTAW, etc. have been developed for improving productivity and enhancing its scope for diverse areas. The traditional GTAW process has certain issues like the weld pool dynamics and slower manual wire feed rates. A highly skilled operator is required for the manual GTAW process. To enhance productivity, increase of current increases the arc pressure which results in defects and poor weld bead quality. But these processes can be further optimized based on the other parameters like type of current supply (DC supply, DC pulsed supply, and AC supply), mean voltage, wire feed rate, frequency, and hotwire current.

1564

V. J. Badheka et al.

Progress in GTAW Process Hotwire GTAW and Narrow Gap Hotwire GTAW The hotwire GTAW method is able to produce high-quality welds with improved deposition rates and finds a feasible place in different applications where it was not before. It uses an additional power source for wire heating by which faster weld speed and improved wire feed rates are obtained that reduces the weld defects. A schematic diagram for hotwire GTAW is shown in Fig. 4. The narrow gap hotwire GTAW process is an extension of the hotwire GTAW process aimed for productivity enhancement. Due care has been taken to achieve a constant sidewall fusion in a narrow gap with special accessories and techniques. Different metals and alloys with thick jobs can be welded using this process. This process gives high quality and efficient welds with excellent mechanical properties. A typical setup for narrow gap hotwire GTAW process is shown in Fig. 5.

Fig. 4 Hotwire GTAW (Santhakumari 2018)

Fig. 5 (a) Narrow gap hotwire TIG; (b) welding in progress (Santhakumari 2018)

61 GTAW Application for Additive Manufacturing and Cladding of Steel Alloys

1565

Fig. 6 Orbital hotwire GTAW (Santhakumari 2018)

Orbital Hotwire GTAW Orbital hotwire GTAW is a hybrid process that combines the conventional orbital GTAW process and the hotwire method. The process has been developed to increase the productivity, weld quality, and its wider applications in aircraft, aerospace industry, nuclear, boiler, and high-pressure tubes. A typical setup of the orbital hotwire GTAW process for pipe welding is shown in Fig. 6.

Activated TIG Welding Activated TIG or A-TIG Welding was developed in the middle of the 1960s by Paton Welding Institute (PWI), Ukraine. In this process to improve the weld penetration, a thin flux material coating is applied on top of the joint surface before welding. It is reported that the weld penetration is increased two to three times than the conventional GTAW process. It is presently used for steels and nickel base alloys and found useful in orbital pipe-welding applications. A typical setup for an activated TIG welding process is shown in Fig. 7.

Automated Wire Feeding in GTAW It is a hybrid process that combines manual and automated GTAW wire feed control (Tip TIG / Top TIG) combined with a hotwire method. It is suitable in any position of the weld of any thickness material. A typical setup of Tip TIG which enables to weld root is shown in Fig. 8.

1566

V. J. Badheka et al.

Fig. 7 Activated TIG welding (Santhakumari 2018)

Fig. 8 Root pass welded by Tip TIG process (Santhakumari 2018)

Inter Pulse TIG Welding Inter Pulse TIG patented by VBC Group, UK, is specially designed to give highly constricted and fine welding arc to weld “difficult-to-weld” materials. A comparison of the DC TIG arc and the inter-pulse arc is depicted in Fig. 9. It can be seen that the inter-pulse arc is finer with smaller heat-affected zone (HAZ).

K-TIG Welding K-TIG (Keyhole TIG) welding technique was developed in 1990–1993 by Dr. Laurie Jarvis, Australia, which is a single-pass full-penetration keyhole method. In this process, the weld pool surface is fixed to the top and bottom surface making

61 GTAW Application for Additive Manufacturing and Cladding of Steel Alloys

1567

Fig. 9 (a) DC TIG arc 85 Amps; (b) inter-pulse arc 85 Amps average current (Santhakumari 2018)

Fig. 10 (a) K-TIG welding process; (b) macro of the K-TIG weld (Santhakumari 2018)

a stable arc that is moved along the weld path as a keyhole by increasing the arc pressure. A typical K-TIG welding technique is explained in Fig. 10.

Twin Wire/Tandem GTA Cladding Process Conventional GTAW process is not recommended for cladding due to its slower production caused by lesser dilution which resulted in weld defects. To overcome these defects, in the late 1990s Yamada patented a novel high-efficiency GTAW method known as multicathode GTA for improving process efficiency and weld quality shown in Fig. 11a, where a welding torch contains the electrically insulated electrodes with independently operated power supplies. A plot of weld travel speed

1568

V. J. Badheka et al.

Fig. 11 (a) A schematic of multicathode GTAW process; (b) comparison of weld travel speed and current (Egerland et al. 2015)

Fig. 12 (a) Tandem GTAW process setup; (b) Tandem GTAW torch setup (Santhakumari 2018)

and current consisting of single and multicathode GTAW is shown in Fig. 11b (Egerland et al. 2015). This process is capable of increase in weld speed without weld defects. The Tandem GTA cladding process is an extended version of the twin hotwire GTAW process where the two electrodes are combined with two independently controlled preheated welding wires which permit higher weld speeds with higher deposition rate. A typical Tandem GTAW process and torch setup are shown in Fig. 12.

61 GTAW Application for Additive Manufacturing and Cladding of Steel Alloys

1569

Fig. 13 Internal GTAW (Santhakumari 2018)

Fig. 14 A schematic of single and twin hotwire GTAW (Egerland et al. 2015)

Internal GTAW Joining of stub tube to pipe known as internal GTAW system was developed by WRI and is depicted in Fig. 13. It has a servomotor drive which is controlled by a programmable logic controller (PLC) that handles all the inputs like voltage, current, and gas flow rate and outputs.

Multicathode GTAW There are four variants of hotwire GTAW as single cathode single hotwire GTAW, single cathode double hotwire GTAW, double cathode single hotwire GTAW, and double cathode double hotwire GTAW. A schematic of single and twin hotwire GTAW is shown in Fig. 14. The dissimilar material joining is recognized as a challenge where there are increasing demands for high strength and lightweight alloys in different industries.

1570

V. J. Badheka et al.

In such a case, the twin GTAW process can be beneficial to join by supplying less heat input to highly thermal conductive material and more heat input to low thermal conductive material (Shah and Agrawal 2019).

Additive Manufacturing AM processes have different variants which include vat photopolymerization, binder jetting, material extrusion, material jetting, sheet lamination, powder bed fusion, and directed energy deposition. The benefits of AM processes are less material waste and a minimum processing cycle. It is successfully applied to fabricate components for a variety of metals. In these processes, different sources of heat such as arc, laser beam, or electron beam are used. The wire arc-additive manufacturing (WAAM) process was developed by Baker in the year 1925. The WAAM process belongs to the directed energy deposition method where the wire is heated then melted and transferred to the melt pool, and it continues to form a component by building layer by layer. The process can be employed using GMAW, GTAW, PAW, or cold metal transfer welding (CMT) as a heat source and build a component as depicted in Fig. 15 (Jin et al. 2020). Currently, WAAM is one of the popular fabrication processes applied for different materials like Ti, Al, Ni alloy, and steel. It is reported that directed energy deposition GMAW has deeper penetration, 5–10 times arc power than powder bed fusion with a laser heat source, and directed energy deposition with a laser heat source, and produces stainless steel (SS) 316 components with much higher speeds and least fusion defects (Jin et al. 2020; Mukherjee and DebRoy 2019). Recently, the WAAM process has attracted the attention of many researchers in producing large-scale SS parts with high efficiency and low cost. Eagar (Eager 1995) opinioned that, “a new welding technology is often get commercialized before a fundamental science emphasizing the underlying physics and chemistry can be developed.” It is truly agreed and applicable to the WAAM process as well. The different factors that affect the performance of the WAAM process are depicted in Fig. 16. The dimensional accuracy and surface quality are based on the WAAM process parameters. The thermal history, solidification behaviors, and phase transformations are closely related to each other and have a strong influence on the microstructure. Anisotropy problems are frequently seen in WAAM-processed parts, and they can be overcome by in-process mechanical working like rolling followed by WAAM. Abe and Sasahara (Abe and Sasahara 2016) successfully deposited first YS308L stainless steel weld bead on a SUS304 stainless steel substrate and then finally deposited Ni6082 weld bead onto the previous stainless steel weld bead using WAAM process. They have studied the structure-properties of the dissimilar metal deposition layers using a nickel-based alloy. Comparable bond strength was obtained at YS308L and Ni6082 bond area than the tensile strength of the YS308L and Ni6082 weld metals. It is argued that the bond produced has high heat resistance, and corrosion resistance along with low weight of the object due to

61 GTAW Application for Additive Manufacturing and Cladding of Steel Alloys

1571

Fig. 15 Schematic diagram of the WAAM setup (Xia et al. 2020)

its inside rib structure. Such kind of inside rib structure is difficult to fabricate using other traditional manufacturing processes. Chen et al. (2017) fabricated 316 L austenitic SS using GMAW-AM and studied the mechanical properties and microstructure. They found that the microstructure has σ,δ, and γ phases which are differently orientated and have different morphologies. Also, the mechanical properties are similar to wrought 316 L with ductile mode of fracture. The increase of microhardness and tensile strength is attributed due to σ phase; however, there is reduction in tensile yield strength, and ductility. They have extended their work (Chen et al. 2018), to investigate the mechanical, corrosion properties, and analyzed the effect of heat treatment on the microstructure. The heat treatment causes improvement in corrosion resistance of the steel. However, due to σ phase formation, there is an increase in corrosion attack sensitivity. Rodriguez et al. (2018) have employed CMT and Top TIG processes to deposit the 316 L SS and compared the mechanical properties and microstructures of deposit produced using these processes. Both of these processes are able to deposit thick walls with high accuracy and sound flatness. Higher material deposition rates were obtained for CMT (3.7 kg/h) than Top TIG process (2 kg/h). Li et al. (2019) deposited H08Mn2Si steel filler wire onto Q235B low-carbon steel as a substrate using GMAW-AM process and investigated the molten pool stability at different wire feed speed, weld speed, and inclination angles, i.e., angle between the GMAW torch and substrate. It is found that the bead width increases with an increase in the inclination angle, with decrease in bead height

Fig. 16 Cause and effect diagram for WAAM process

1572 V. J. Badheka et al.

61 GTAW Application for Additive Manufacturing and Cladding of Steel Alloys

1573

and penetration depth. Similarly, the forward position of the GMAW torch gives a stable molten pool with superior quality than other positions. Ge et al. (2019) studied microstructural evolution, defect distribution, and residual stresses developed in additively manufactured 2Cr13 SS thin wall parts created using robotic CMT technology and numerical simulations. The residual stress distribution grows frequently with the addition of the deposited layers and was found highest for the final deposited layer. Rafieazad et al. (2019) deposited ER70S-6 low-carbon low-alloy steel wire onto ASTM A36 mild steel substrate using GMAW-AM process in advanced surface tension transfer mode. They have studied the microstructure and mechanical properties of the deposits. In the microstructure, the main phases observed were fine ferrite and laminar pearlite. The additively manufactured part of microstructure has found equiaxed grains with weak cubic texture. A similar kind of mechanical properties of deposits as that of substrate was found in both deposition and building direction. The mechanical properties of different steels produced using WAAM process are depicted in Table 1.

Cladding In metal cladding, a thin layer of coating material is applied onto the base material or substrate. A variety of methods employed for cladding such as accumulative roll bonding (ARB) (Selvaraj et al. 2020), resistance welding, SMAW, submerged arc welding (SAW), overlay welding, electroslag welding (ESW), GTAW, flux-cored arc welding (FCAW), laser beam cladding, oxyacetylene welding, laser powder welding, pulsed GMAW, tubular core covered electrodes, hotwire plasma process, explosive cladding, plasma transferred arc (PTA) cladding, hybrid methods, etc. (Saha and Das 2016, 2018). Chen et al. (2009) have successfully deposited multicomponent alloy fillers (Ni, Co, Cr, Al, and Mo with Si) on low-carbon steel substrates using GTAW cladding process. They evaluated the microstructure and wear properties of this multicomponent alloy. They observed that the FeMoSi as a principal dendritic and BCC interdendritic phases with both phases have multiple elements. The hardness and wear resistance of cladding layer are function of Si content. Both enhancement in the hardness and wear resistance of cladding layer are affected by FeMoSi dendrites, which have strong covalent bonds. Chen and Lee (2016) have attempted to alter the surface properties like surface hardness, wear resistance, and heat dissipation of AISI 410 martensitic SS with an AlN clad layer by specific addition of Si, W, and Co using GTAW process. It is found that the surface properties of cladding material were better than the parent material. Lv et al. (2008) investigated GTAW cladding of S201 copper (Cu) alloy under different processing conditions on carbon steel. Their work focuses on the influence of clad current on the concentration and morphology of Fe solute in the clad layer and the hardness of the Cu/Fe bond area. The Fe morphology in the Cu layer

Filler metal wire ER70S-6

316L

316L

316L

Heat source GMAW

GMAW

GMAW

GMAW

316L

–

–

Substrate A36

WD: 1.2 mm, SpeedPulse WAAM, mean current (I)-22.1 A; mean voltage (U)- 135 V; arc power P-2984W; Layer thickness (δ)-1.5 mm WD: 1.2 mm, SpeedArc WAAM, I -19.5 A; U- 140 V; P-2730W; δ-1.8 mm

WD: 1.2 mm, solution treated

WD: 1.2 mm, as deposited WD: 1.2 mm, cold worked

Input process condition WD: 0.889 mm, horizontal WD: 0.889 mm, vertical WD: 1.2 mm, as deposited WD: 1.2 mm, 1000◦ C/1 h, WQ WD: 1.2 mm, 1100◦ C/1 h, WQ WD: 1.2 mm, 1200◦ C/1 h, WQ WD: 1.2 mm, 1200◦ C/4 h, WQ WD: 1.2 mm, wrought 316L

Table 1 Mechanical properties of different steels produced using WAAM processes

YS: 417.9 MPa; UTS: 553 ± 2 MPa

Mechanical properties YS: 400 MPa; UTS: 500 MPa YS: 385 MPa; UTS: 500 MPa YS: 235 MPa; UTS: 533 MPa; %El: 48 YS: 255 MPa; UTS: 549 MPa; %El: 41 YS: 323 MPa; UTS: 498 MPa; %El: 56 YS: 215 MPa; UTS: 474 MPa; %El: 57 YS: 204 MPa; UTS: 494 MPa; %El: 70 YS: 222–265 MPa; UTS: 505 MPa; %El: 56–63 YS: 235 MPa; UTS: 533 MPa; %El: 48 YS: 255–310 MPa; UTS: 525–632 MPa; %El: 30 YS: 222–265 MPa; UTS: 505–578 MPa; %El: 56–63 YS: 418.0 MPa; UTS: 550 ± 6 MPa

Wang et al. (2019)

Chen et al. (2017)

Abe and Sasahara (2016)

Ref Rafieazad et al. (2019)

1574 V. J. Badheka et al.

Al and LS422750/4 99.5% black annealed iron 316L

316L

DH36

WD: 1.2 mm, CMT-vertical- continuous mode (CM) WD: 1.2 mm, vertical- pulsed mode (PM) WD: 1.2 mm, horizontal-CM WD: 1.2 mm, horizontal-PM WD: 1.2 mm, base material WD: 1.2 mm, top TIG -vertical WD: 1.2 mm, top TIG - horizontal

WD: 0.9 mm, as deposited

YS: 331 MPa; UTS: 536 MPa; %El: 45.6 YS: 364 MPa; UTS: 577 MPa; %El: 43.4 YS: 374 MPa; UTS: 588 MPa; %El: 45.1 YS: 346 MPa; UTS: 651 MPa; %El: 47 YS: 322 MPa; UTS: 539 MPa; %El: 43.1 YS: 365 MPa; UTS: 590.3 MPa; %El: 42.3

YS: 336 MPa; UTS: 574 MPa; %El: 42

YS: 847 MPa; UTS: 944 MPa; %El: 3.27

WD wire diameter (mm), YS yield strength (MPa), UTS ultimate tensile strength (MPa), and %El percentage elongation

CMT and TopTIG

GTAW

Rodriguez et al. (2018)

Shen et al. (2015)

61 GTAW Application for Additive Manufacturing and Cladding of Steel Alloys 1575

1576

V. J. Badheka et al.

depends on its concentration. The Fe concentration is exponentially related with growth of clad current. The hardness of Cu layer is a function of clad current and its Fe content. Silwal et al. (2019) employed low current GTAW cladding of hot-wire filler of IN625 Inconel alloy on the 347 austenitic SS substrate. It is found that the torch angle has significant effect on the bead formation. The low current resulted in minimum Fe dilution on the cladding whereas the high current (90–100 A) has resulted in liquation cracking formation in the HAZ of 347 SS. Malhotra (2020) employed GTAW process for top clad layer remelting of austenitic stainless steel 316 L which was previously deposited on low-carbon steel substrate using GMAW. It has been found that the pitting corrosion resistance was higher than the GMAW cladding with comparatively smaller degree of sensitization and obtained depth of penetration around 2.34 mm. Knerek et al. (2021) have successfully deposited 625 Inconel alloy on API 5L X65 carbon steel pipe as a substrate using hotwire TIG process. The cladding layer is free from defect or cracks during bend test, and the resulting cladding possesses enhanced mechanical properties. Moradi and Ketabchi (2015) have cladded 625 Inconel alloy on the A516 Grade 70 carbon steel plate using GTAW process. It is found that the shear strength of 232.5 MPa was obtained which is greater than regular standards. The microstructure shows mainly Niobium carbide (NbC) and Fe2 Nb precipitate phases. In general, metal cladding provides high hardness, corrosion and/or erosion resistance, and good bonding where clad material of different thickness is deposited over the substrate surface (Saha and Das 2016). Recently, weld cladding processes find applications in numerous industries as a cost-effective engineering solution method.

Optimization of Welding Parameters Much of the research work reported on the effect of welding parameters in GTAWbased metal deposition process. It is reported that the current, voltage, torch velocity, wire feed speed, and TIG welding torch with the substrate are vital weld parameter that affects the performance of the metal deposition process (Gokhale et al. 2019). The WAAM deposition path, thermal history, and phase transformation control during WAAM can be optimized to reduce the residual stress and distortion of the parts (Jin et al. 2020). Still, other weld parameters like current supply (DC supply, DC pulsed supply, and AC supply), frequency, and hotwire current are not explored so far. It is reported that the heat input in the layer by layer deposition needs to be controlled for getting a high deposition rate and good forming appearance (Dadbakhsh et al. 2012). To date, researchers have employed response surface methodology (Srivastava et al. 2018; Youheng et al. 2017; Balasubramanian et al. 2009; Lakshminarayanan et al. 2008), mathematical modeling (Geng et al. 2017),

61 GTAW Application for Additive Manufacturing and Cladding of Steel Alloys

1577

and developed controllers (Bonaccorso et al. 2011; Xiong et al. 2016) for optimizing the weld parameters, genetic algorithm (GA) for weld bead geometry optimization in PTA hard-faced austenitic SS plates (Siva et al. 2009). Traditional optimization methods give local optimal solutions and are not robust. The meta-heuristics (MH) is a global search technique which deals with all kinds of objective functions and design variables, and these are flexible as there is no need of data training (Saka et al. 2016). There is an ardent need to explore these MH and other machine learning algorithms for AM and cladding to optimize the weld parameter and predict the weld performances. Furthermore, for making state-ofthe-art processes (Tutum and Hattel 2011), a clear understanding of the thermomechanics of these welding processes is essential by which the mathematical models can be developed. With these developed models as an input to the different efficient algorithms in which the process parameters can be optimized efficiently with multiple inputs to get commercially acceptable parts. The AM and cladding of metal parts with advanced properties have attracted great attention in various fields. Still, AM faces substantial techno-commercial logical and scientific issues like difficulty in microstructure control, properties and defects, lack of standards, slower rate of AM, and availability of filler materials for several commercial alloys and economy. Similarly, the effectiveness of digital twins is well recognized; however, it is not fully explored for AM. Efforts are needed to develop or modify the digital twin building blocks and their analysis for variety of alloys and AM processes (DebRoy et al. 2021).

Conclusion This chapter elucidates the application of GTAW for the additive manufacturing and cladding process of steel alloys. It covers different techniques employed for additive manufacturing and cladding processes of steel alloys. Parametric optimization for GTAW-based additive manufacturing and cladding processes with multiple input weld parameters using meta-heuristic algorithms and machine learning algorithms are suggested for improving the weld performance.

Websites 1. https://economictimes.indiatimes.com/small-biz/productline/building-materials/ alloy-steel-everything-you-need-to-know-about-alloy-steels-and-their-role-in-bu ilding-and-construction-industry/articleshow/70344024.cms. Accessed 19 Jan 2021. 2. https://steel.gov.in/overview-steel-sector. Accessed 19 Jan 2021. 3. https://www.steel.org/wp-content/uploads/2020/12/2020-AISI-Profile-Book. pdf. Accessed 19 Jan 2021.

1578

V. J. Badheka et al.

References Abe T, Sasahara H (2016) Dissimilar metal deposition with a stainless steel and nickel-based alloy using wire and arc-based additive manufacturing. Precis Eng 45:387–395 Balasubramanian V, Lakshminarayanan AK, Varahamoorthy R, Babu S (2009) Application of response surface Methodology to prediction of dilution in plasma transferred arc hardfacing of stainless steel on carbon steel. J Iron Steel Res Int 16:44–53 Bonaccorso F, Cantelli L, Muscato G (2011) Arc welding control for shaped metal deposition process. IFAC Proc 44:11636–11641 Chen TM, Lee HM (2016) Surface analysis of AISI 410 stainless steel cladding with AlN, Si, Co, and W powders. Sens Mater 28:1005–1011 Chen JH, Chen PN, Hua PH et al (2009) Deposition of multicomponent alloys on low-carbon steel using gas tungsten arc welding (GTAW) cladding process. Mater Trans 50:689–694 Chen X, Li J, Cheng X et al (2017) Microstructure and mechanical properties of the austenitic stainless steel 316L fabricated by gas metal arc additive manufacturing. Mater Sci Eng A 703:567–577 Chen X, Li J, Cheng X et al (2018) Effect of heat treatment on microstructure, mechanical and corrosion properties of austenitic stainless steel 316L using arc additive manufacturing. Mater Sci Eng A 715:307–314 Dadbakhsh S, Hao L, Sewell N (2012) Effect of selective laser melting layout on the quality of stainless steel parts. Rapid Prototyp J 18:241–249 DebRoy T, Mukherjee T, Wei HL et al (2021) Metallurgy, mechanistic models and machine learning in metal printing. Nat Rev Mater 6:48–68 Eager TW (1995) Welding and joining: moving from art to science. Weld J 74:49–55 Egerland S, Zimmer J, Brunmaier R et al (2015) Advanced gas tungsten arc weld surfacing current status and application. Soldag Inspeção 20:300–314 Ge J, Ma T, Han W et al (2019) Thermal-induced microstructural evolution and defect distribution of wire-arc additive manufacturing 2Cr13 part: numerical simulation and experimental characterization. Appl Therm Eng 163:114335 Geng H, Li J, Xiong J et al (2017) Optimization of wire feed for GTAW based additive manufacturing. J Mater Process Technol 243:40–47 Gokhale NP, Kala P, Sharma V (2019) Thin-walled metal deposition with GTAW welding-based additive manufacturing process. J Braz Soc Mech Sci Eng 41:569–80 Jin W, Zhang C, Jin S et al (2020) Wire arc additive manufacturing of stainless steels: a review. Appl Sci 10:1563 Knerek R, Lemos GVB, Vander VG et al (2021) Investigating an API X65 steel pipe cladded with alloy 625. Tecnol em Metal Mater e Mineração 18:e2465 Lakshminarayanan AK, Balasubramanian V, Varahamoorthy R, Babu S (2008) Predicting the dilution of plasma transferred arc hardfacing of Stellite on carbon steel using response surface methodology. Met Mater Int 14:779–789 Li Y, Xiong J, Yin Z (2019) Molten pool stability of thin-wall parts in robotic GMA-based additive manufacturing with various position depositions. Robot Comput Integr Manuf 56:1–11 Lv SX, Xu ZW, Wang HT, Yang SQ (2008) Investigation on TIG cladding of copper alloy on steel plate. Sci Technol Weld Join 13:10–16 Malhotra D (2020) Effect of GTAW remelting on the corrosion performance of AISI 316L cladding. https://doi.org/10.1002/maco.202011666 Moradi MJ, Ketabchi M (2015) An experimental study of microstructure and mechanical behavior of alloy 625 weld overlay deposited on ASTM A516 grade 70. Indian J Sci Technol 8:1–6 Mukherjee T, DebRoy T (2019) Printability of 316 stainless steel. Sci Technol Weld Join 24: 412–419 Rafieazad M, Ghaffari M, Vahedi Nemani A, Nasiri A (2019) Microstructural evolution and mechanical properties of a low-carbon low-alloy steel produced by wire arc additive manufacturing. Int J Adv Manuf Technol 105:2121–2134

61 GTAW Application for Additive Manufacturing and Cladding of Steel Alloys

1579

Rodriguez N, Vázquez L, Huarte I et al (2018) Wire and arc additive manufacturing: a comparison between CMT and TopTIG processes applied to stainless steel. Weld World 62:1083–1096 Saha MK, Das S (2016) A review on different cladding techniques employed to resist corrosion. J Assoc Eng India 86:51 Saha MK, Das S (2018) Gas metal arc weld cladding and its anti-corrosive performance – a brief review. Athens J Technol Eng 5:155–174 Saka MP, Hasançebi O, Geem ZW (2016) Metaheuristics in structural optimization and discussions on harmony search algorithm. Swarm Evol Comput 28:88–97 Santhakumari A (2018) Advances in gas tungsten arc welding technology. Weld Des Fabr:14–18 Selvaraj SK, Srinivasan K, Deshmukh J et al (2020) Performance comparison of advanced ceramic cladding approaches via solid-state and traditional welding processes: a review. Materials (Basel) 13:1–22 Shah P, Agrawal C (2019) A review on twin tungsten inert gas welding process accompanied by hot wire pulsed power source. J Weld Join 37:41–51 Shen C, Pan Z, Ma Y et al (2015) Fabrication of iron-rich Fe-Al intermetallics using the wire-arc additive manufacturing process. Addit Manuf 7:20–26 Silwal B, Walker J, West D (2019) Hot-wire GTAW cladding: Inconel 625 on 347 stainless steel. Int J Adv Manuf Technol 102:3839–3848 Siva K, Murugan N, Logesh R (2009) Optimization of weld bead geometry in plasma transferred arc hardfaced austenitic stainless steel plates using genetic algorithm. Int J Adv Manuf Technol 41:24–30 Srivastava M, Rathee S, Maheshwari S, Kundra TK (2018) Multi-objective optimisation of fused deposition modelling process parameters using RSM and fuzzy logic for build time and support material. Int J Rapid Manuf 7:25 Tisza M, Czinege I (2018) Comparative study of the application of steels and aluminium in lightweight production of automotive parts. Int J Light Mater Manuf 1:229–238 Tutum CC, Hattel J (2011) State-of-the-art multi-objective optimisation of manufacturing processes based on thermo-mechanical simulations. In: Wang L, Ng AHC, Deb K (eds) Multiobjective evolutionary optimisation for product design and manufacturing. Springer, London, pp 71–133 Wang S, Tadic D, Jiang J, McKinnell J (2016) Weld cladding for mining and mineral processing: a Canadian perspective. In: CWA, pp 52–66 Wang L, Xue J, Wang Q (2019) Correlation between arc mode, microstructure, and mechanical properties during wire arc additive manufacturing of 316L stainless steel. Mater Sci Eng A 751:183–190 Xia C, Pan Z, Polden J et al (2020) A review on wire arc additive manufacturing: monitoring, control and a framework of automated system. J Manuf Syst 57:31–45 Xiong J, Yin Z, Zhang W (2016) Closed-loop control of variable layer width for thin-walled parts in wire and arc additive manufacturing. J Mater Process Technol 233:100–106 Youheng F, Guilan W, Haiou Z, Liye L (2017) Optimization of surface appearance for wire and arc additive manufacturing of Bainite steel. Int J Adv Manuf Technol 91:301–313

Using Smart Mesoporous Silica in Designing Drug Delivery Systems

62

Kayambu Kannan

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Synthesis of Mesoporous Silicas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Synthesis of Biodegradable MPSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Functionalization of MPSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Improvement of Biocompatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Loading of Drug and Biomolecules on MPSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Targeting Delivery Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Passive Targeting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Active Targeting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gate Keeping of Drug Loaded MPSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stimuli-Responsive Smart Delivery System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Internal Stimuli-Responsive Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . External Stimuli-Responsive Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multi-stimuli Responsive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion and Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1582 1583 1586 1586 1588 1588 1589 1589 1590 1592 1593 1595 1600 1603 1606 1608

Abstract

The drug delivery systems are playing an important role in targeting site of action. The delivery systems are broadly classified into organic and inorganic materials. Among the drug delivery systems the inorganic smart mesoporous silicas (SMSs) have greater advantages over the other system. Many synthetic techniques have been reported for tailoring of SMSs by fine tuning of structure, biocompatibility, and functionalization with requisite functional groups. The

K. Kannan () PG and Research Department of Chemistry, Raja Doraisingam Government Arts College, Sivagangai, Tamil Nadu, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_111

1581

1582

K. Kannan

well-designed targeting approaches have been found to be remarkable in reducing the side effects as compared to the intended drug effect as well as protecting the drug molecules from degradation. The critical part of the delivery process is the way by which the drug is released upon exerting a stimulus such as pH, redox reagents, thermal, magnetic, and ultrasound with controlled accumulation within certain region around the targeting tissues. In this chapter, we illustrate the different synthetic strategies for designing the drug delivery system using smart mesoporous silicas with multi-stimuli responsive carriers. Keywords

Mesoporous silica · Functionalization · Stimuli responsive · Therapeutic molecules · Drug delivery system

Introduction Healthcare issues ever have the highest priority in the world and hence many efforts are focused on the improvement of treatment by concerning the cost, duration, and efficacy. Porous materials have shown great promise to offer sustainable solutions to therapeutics such as drug delivery, biomedical imaging, and tumor therapy. In the last few decades, the applications of mesoporous silicas (MPSs) as delivery systems for biomolecules and medicine have attracted much interest among the pharmaceutics (Gupta and Mukherjee 2013; Palit and Hussain 2020; Di Sia 2014). The mesoporous silica materials exhibit remarkable physico-chemical properties such as (i) tunable pore size (2–50 nm) and high pore volumes (ca. 1 cm3 /g), (ii) high specific surface areas (up to 1500 cm2 /g), (iii) high density of silanol groups on the surface for wide functionalization, (iv) robust silica framework that allows harsh reaction conditions, and (v) great biocompatibility. The first work of drug delivery of anti-inflammatory drug ibuprofen using MCM-41 mesoporous silica nanoparticle opened up the opportunity to design silica-based nanoparticle for medical applications. The high delivery efficiency with controlled release profile of vancomycin and adenosine triphosphate was observed using cadmium sulfidefunctionalized MCM-41 material (Huang et al. 2020). This achievement inspired the researchers to develop the MPS-based drug delivery systems for biomedical applications. Nowadays, MPSs are widely used as carriers for the treatment of complex diseases; however, their clinical translation remains great challenges. Ideal drug delivery system is one that should accumulate only in the target area and that could be able to release their cargo only inside the infected area or tumor cell. Practically, the delivery systems have to face several barriers, when they are administered into a patient. These not only reduce the efficacy of the treatment but also prevent their successful translation into the clinical applications (Sábio et al. 2019). To overcome the obstacles, researchers have to focus on the development of MPSs delivery systems along with targeting ligand and suitable gatekeeper for stimuli-responsive delivery of drug molecules in specified area (Fig. 1).

62 Using Smart Mesoporous Silica in Designing Drug Delivery Systems

1583

Fig. 1 Some recent development on designing of drug delivery systems (DDSs) using smart mesoporous silicas (MPSs)

This chapter is initially intended to provide an account on the preparation and the surface functionalization of MPSs toward the biocompatibility with requisite functional moieties. Afterward the influences of drug or biomolecules loaded into the functionalized MPSs with or without suitable gatekeeper and their performance as targeting drug delivery system in therapeutic treatment. Finally, the strategies used to prevent premature drug release and the on-demand stimuli-responsive drug deliveries are discussed. In addition the drug delivery systems with their responsiveness to various stimuli, chemical, physical, and biological factors used for various therapeutic treatments are discussed.

Synthesis of Mesoporous Silicas In the literature, several different methods have been reported for synthesis of mesoporous silica particles to yield a variety of engineered particles with unique properties (Fig. 2). The sol–gel method has unique advantages over other methods to attain the ordered, high surface area, narrow pore size distribution, the geometry and connectivity of their pores in very well defined particles. In biomedical applications of MPSs, a careful selection of synthesis method plays vital role to obtain particles with requisite dimensions for each specific application (Fig. 3). In brief, surfactant is stirred in a mixture of water/alcohol under basic or acidic conditions and

1584

K. Kannan

Fig. 2 Schematic representation of the formation mechanisms of various MPSs

tetraethylorthosilicate (TEOS) or other silicates are added under agitation. Spherical MCM-41 materials with a hexagonal pore array are usually synthesized in basic solution using cetyltrimethyl ammonium bromide (CTAB) as the structure-directing agent. The surfactant with silicon molar ratio also influences the mesostructure and size, and CTAB/alkoxysilane molar ratio of 0.13 or higher is required for obtaining highly dispersed MPSs (Yamada et al. 2014). The pH of the reaction mixture also could affect the MPSs properties, a higher pH resulting in a wormlike morphology for MCM-41 particles. The formation of regular pore geometry and the dimensions are strongly dependent on the supramolecular self-assembling of surfactant micelles with silica molecules. Nonionic triblock copolymers have also been used in the synthesis of Santa Barbara Amorphous materials (SBA). The various ratio of poly ethylene oxide to propylene oxide provided to achieve the desired symmetry of mesoporous materials (SBA-11 (cubic), SBA12 (3-d hexagonal), SBA-15 (hexagonal) and SBA-16 (cubic cage-structured) with larger pore size. Highly ordered hexagonal mesoporous SBA-15 was synthesized using Pluronic P123 triblock copolymer (EO20–PO70–EO20, BASF) as a template and TEOS as a silica source in acidic conditions (Kannan and Jasra 2009).

62 Using Smart Mesoporous Silica in Designing Drug Delivery Systems

1585

Fig. 3 Strategies for designing mesoporous silica’s properties through synthesis

Folded sheets mesoporous material (FSM-16) is another type of mesoporous material that is synthesized using quaternary ammonium salt as a template and layered polysilicate kanemite (Tozuka et al. 2005). However, the modulation of pore volume remains a great challenge, which limits the encapsulation of drug molecules and macromolecules including proteins, enzymes, antibodies, RNA, and DNA for biomedical applications. A different synthetic strategy proposed to synthesis the porous ultra-large spherical mesopores silicas by using swelling agents like 1, 3, 5-triisopropylbenzene, cyclohexane, xylene, ethylbenzene and toluene to get dissolved moderately in the micelle of a specific surfactant in order to produce a clear micell template structure with significantly enlarged pores. We synthesized the siliceous meso cellular foam (MCF) using the Pluronic P123 BASF in acidic condition with trimethyl benzene (TMB) as swelling agent at 35–40 ◦ C (Kannan and Jasra 2011). Further some studies revealed that the correlations between the excessive pore expanding could lead to a mechanically unstable thinner pore wall. Recently, dendritic mesoporous silica nanoparticles (DMSNs) with open 3D dendritic super structures and center-radial pore channels have been prepared in an aqueous solution using TEOS and bis(triethoxysilyl)ethane (BTEE) as precursor and CTAB/sodium salicylate (NaSal) as structure-directing agent. The increased molar ratio of CTAB/NaSal from 0.75:1 to 1:1 could increase the pore size from 8.1 to 17.5 nm. Further the structure-dependent and glutathione-responsive biodegradable dendritic mesoporous organosilica nanoparticle was observed for safe protein delivery (Yang et al. 2016).

1586

K. Kannan

Synthesis of Biodegradable MPSs The biodegradability with consequent clearance of MPSs must be taken into serious consideration for biomedical applications. Under physiological conditions the MPSs degraded into silicic acid including monomeric silicic acid and polysilicic acids through successive hydration, hydrolysis, and ion-exchange steps, which can be excreted through the urine. However, the degradation process of MPSs is relatively slow due to its stable Si-O-Si frameworks and this reluctant biodegradation could lead to unwanted accumulation of MPSs within the body, which possibly causes the severe tissue inflammation and related problems. Therefore, the improvements of MPSs with better biodegradability extensively interests to promote their application in clinical translation. Recently, several strategies have been reported to improve the biodegradability of MPSs. For example, the doping of Mg2+ into the silica framework increased the degradability of the obtained hollow mesoporous silicas (HMSs) due to the much weaker Si-O-Mg network compared to the Si-O-Si network. In brief, the Si-O-Mg bonds were sensitive to mild acidic environment, including tumor tissue, which causes the breakup of Mg-O bonds to generate abundant defects within the framework to accelerate the biodegradation. The Ca, Mn, Zn, and Na elements have also been introduced into the MPSs framework to facilitate the biodegradability, and the rapid degradation is triggered by typical environment such as pH (for Ca and Mn) and specific peptides (for Fe and Mn) (Hadipour Moghaddam et al. 2019). A redox triggered degradable HMSs was fabricated by using phenylene and bis (propyl) tetrasulfide-bridged organoalkoxysilanes, where the bis (propyl) tetrasulfide acted as a self-destructor in reductive environment (Guimarães et al. 2020). Enzymatically degradable MPSs were prepared by using phenylene and oxamidebridged organoalkoxysilanes, the oxamide in the silica framework is responsible for enzymatic degradation, which could be triggered in the presence of trypsin. This research led to work toward the drug delivery to organs containing specific proteins for targeted therapy (Croissant et al. 2016). The disulfide containing silsesquioxanes is also an interesting material for the preparation of degradable MPSs to focus on specific triggered degradable strategies. The application of MPSs in tissue engineering is a relatively interesting field that has gained much research interest. The different roles of MPSs comprise the delivery of bioactive molecules, inherent bioactivity, stem cells labeling, and the impacts on scaffolds (Abdo et al. 2020).

Functionalization of MPSs The functionalization or modification of the surface of MPSs plays crucial roles for designing the drug delivery carrier with requisite physical and chemical properties The mesoporous silica materials with organic functional groups are widely prepared by two routes: (i) in situ functionalization (ii) post-synthesis functionalization. The former route involves the co-condensation of a functionalized alkoxysilane

62 Using Smart Mesoporous Silica in Designing Drug Delivery Systems

1587

Fig. 4 Post-synthesis functionalization of mesoporous silicas using triethoxyorganosilanes

(e.g., 3-aminopropyltriethoxysilane with tetra-ethoxysilane (TEOS)). The surfactant template of these materials could be removed by solvent extraction and its result was typically about ca. 80% successful, which probably caused by block copolymer remains trapped within pores of the silicas. The percentage of functional groups was limited to ca. 20% and also usually causes loss of structural order with high functional group contents. The post-synthesis route involves a reaction of the functionalized alkoxysilane with the calcined mesoporous silicas (Fig. 4). The inclusion of functional groups likes –NH2 , -SH, -COOH, -Cl, -CN, -SO3 H, alkyl, and phenyl on silica surfaces as well on pore walls is a well-established procedure. As a result, the ability of interaction between the MPSs surface to drug molecule by hydrogen bonds or van der Waals force, electrostatic, or covalent bonds can be manipulated. We prepared various functionalized MPSs like SBA-15@NH2 , SBA-15@COOH, MCF@NH2 , and MCF@CHO for covalent binding with alkaline serine endopeptidase and cellulase enzymes (Kannan and Jasra 2009; Kannan and Jasra 2011). Many researchers have focused on obtaining suitable functionalized MPSs as vehicles for different therapeutically important molecules to control their adsorption and release.

1588

K. Kannan

Improvement of Biocompatibility The biocompatibility of MPSs is strongly influenced by the kind of moiety present on surfaces, and hence surface modification plays a critical role in improving the biocompatibility of MPSs. Nonfunctionalized MPSs have the tendency to associate with serum proteins, and then be cleaned by phagocytic cells from circulation. Moreover the MPS silanol groups exposed on the surface can interact with biological molecules and lead to their damage (Slowing et al. 2009). In order to improve the biocompatibility and their circulation time in vivo, the surface of MPSs should be coated with suitable biocompatible substances. PEGylation significantly improves the hemolytic activity and cytotoxicity and prevents the MPSs from being captured by phagocytic cells (Li et al. 2019). Some research groups have demonstrated that lipid coating on the surface of MPSs improves the biocompatibility and better drug delivery performance by combining with liposome to provide a safe environment. The protocells formed by spontaneous fusion of phosphatidylcholine-structured liposomes to the surface of MPSs upon mixing are known as lipid bilayer coated MPSs (Ashley et al. 2011).

Loading of Drug and Biomolecules on MPSs The drug loading process is to bind the drug molecules onto the surface with or without functionalized mesoporous silica. An ideal drug loading system would be quickly able to load a huge amount of drug on the internal surface of MPSs and then unload it with a desired release profile with as little waste as possible to the environmental factors. A variety of cargoes such as drugs molecule along with or without biomolecules, contrast agents, and bio-sensing agents can be loaded into MPSs through two main routes: 1) in situ loading during fabrication, and 2) adsorption of cargoes onto mesopores of MPSs (either as physisorption or chemisorptions). Wan et al. (2016) employed a route for in situ loading of heparin and ibuprofen into SBA-15 by evaporation-induced self-assembly (Wan et al. 2016). The most reliable approach for loading of cargoes into MPSs is adsorption process, a concentrated drug molecule solution is immersed with MPSs, the mesopores adsorb the drug molecules through capillary action, and then the drug loaded silicas are separated from the solution through filtration or centrifugation. Xie et al. (2014) reported a higher doxorubicin (DOX) loading efficiency (21.6%) on COOH@MPSs by regulating electrostatic interaction between silica and protonated drugs for controllable drug release rates (Xie et al. 2014). Moreover, this adsorption method can also be adapted for co-loading of hydrophilic and hydrophobic drugs. Doxorubicin and rapamycin were loaded on to magnetic silicas using consecutive adsorption processes by adjusting the solubility of the drug in solvent (Liu et al. 2012).

62 Using Smart Mesoporous Silica in Designing Drug Delivery Systems

1589

The dimethyl bifunctional silyl ether of antitumor prodrug camptothecin (CPT) was successfully tethered onto the surface of HS@MPSs through Thiol-Ene click chemistry and the silyl ether bonds showed an acid-responsive function. The acidcleavable silyl ether bonds could be degraded at the acidic pH in tumor issues (pH = 6.8) and remained stable in normal plasma conditions (pH = 7.4). In another study, the pH-sensitive releasing profile of DOX was observed by conjugating the DOX with OHC@MPSs through covalent attachment, which provided the platform for versatile and easy assembling of drug with functionalized MPSs (Llinas et al. 2018).

Targeting Delivery Systems Passive Targeting Passive targeting is based on the enhanced permeation and retention (EPR) effect. It is based on the size of the drug loaded particles that are administered into the bloodstream for circulation, while most of the particles tend to escape through abnormal neovasculature and accumulate into the targeting tissues. To accomplish the effective targeting drug delivery of MPSs, it is essential for MPSs to be 10 nm diameter and 100–200 nm size. Meng et al. (2011) developed an effective passive targeting delivery system using polyethyleneimine-polyethylene glycol (PEG) coated MPSs with a particle size of 50 nm. Use of this system not only achieved the EPR effect but also an improvement in DOX delivery through passive diffusion to tumor cells was observed. This preferably induced cellular apoptosis and the reduction of tumor size along with severe DOX cytotoxicity in a murine xenograft tumor model (Meng et al. 2011). In another study, MPSs coated gold nanorods showed an efficient passive targeting effect. Two synergistic therapeutic effects exerted on this unique nanodevice, the chemotherapeutic effect of loaded DOX into the MPSs shell and photothermal effect of the gold core. In vivo study of this nanosystem showed exceptional efficiency compared with pure DOX in targeting tumor tissues and also inducing damage to Ehrlich ascites carcinoma for significant cytotoxicity to a breast cancer cell line (MCF-7) on comparison with normal cells (Monem et al. 2014). Subramaniam et al. (2019) evaluated the distinct size (40 nm and 100 nm) of MPSs loaded with rifampicin (RIF) for enhancing the intracellular bacterial infection treatment. An in vitro study of intracellular infection model was established using small colony variants (SCV) of Staphylococcus aureus in macrophages. MPSs100 exhibited cellular uptake values up to 80% of macrophage internalization and MPSs-40 showed lower uptake values (up to 40%), and thus indicated the exocytosis process was more assigned for smaller size MPSs. The observed higher MPSs-RIF internalization and an antibacterial activity enhancement enable these nanoplatforms to treat bacterial biofilms (Subramaniam et al. 2019).

1590

K. Kannan

Active Targeting Active targeting can be employed to promote the internalization of nanoparticles into tumor cells based on biological recognition as the fundamental tool for the fabrication of targeted therapies. This kind of active targeting delivery requires the surface modification of MPSs nanoparticles with a suitable ligand that tend to interact with membrane receptors overexpressed in diseased organs, tissues, and cells. This ligand-receptor interaction will direct a specific retention to increase the endocytosis of nanoparticles into the target cells. The ligands could be antibodies, peptides, aptamers, proteins, small biomolecules (like folic acid), and saccharides (Fig. 5). Folic acid has been used extensively for selective localized delivery of several anticancer drugs to various tumors including ovarian, breast, lung, endometrial, colon, kidney, and brain cancer; it recognizes the folate receptors that are overexpressed on their surfaces (Feng et al. 2016). A recent development on a smart system comprised MCM-41 silica functionalized with two different molecular weights of hyaluronic acid (HA) (6.4 and 200 kDa) to achieve active targeting. On using this, enhancement stability and dispersity of MPSs in biological fluids was

Fig. 5 Schematic representation of multifunctional MPSs loaded with drug and targeting ligands such as aptamers, antibodies, proteins, peptides, small biomolecules, and saccharides

62 Using Smart Mesoporous Silica in Designing Drug Delivery Systems

1591

observed. It also indicated that high molecular weight HA functionalized MCM41 showed higher biocompatibility, lowered cytotoxicity, and favored the targeted binding to HA receptor (CD44), which are overexpressed in many tumor cells. These results revealed that HA-functionalized MPSs are a promising platform for active targeting of cancer cells (Ricci et al. 2018). Antibodies are the most efficient and specific targeting ligands; Mandal et al. (2018) developed an innovative therapeutic approach for targeting leukemia stem cells (LSCs) using succinic anhydride coated MPSs tagged with an anti-B220 antibody. An in vitro study of the anthracycline daunorubicin loaded system showed efficient internalization into murine B220 positive acute myeloid leukemia (ML) and preference to kill these cells compared to B220 negative AML LSCs. Furthermore, these MPSs significantly delayed the leukemia development in recipient mice. These result revealed that active targeting of AML LSCs could be enhanced by using antibody functionalized MPSs as carriers for anti-leukemic drugs (Mandal et al. 2018). Mao and coworkers designed a theranostic system combining synergistic therapy and real-time imaging. Initially the polyvinylpyrrolidone decorated rodlike Bi2 S3 nanoparticles were encapsulated with MPSs and loaded with DOX. The resultant nanoparticles (NPs) were conjugated with trastuzumab (a monoclonal antibody targeting HER-2 overexpressed breast cancer cells) to obtain Tam-Bi2 S3 @MPSs core-shell NPs. Both in vitro and in vivo studies indicated that the Tam-Bi2 S3 @MPSs bear multiple features such as good biocompatibility, drug loading capacity as well as active tumor targeting. This system can serve as excellent contrast agent for computed tomography deep tissue imaging as well as therapeutic agent in cancer treatment (Li et al. 2018). Sheng and coworkers have designed a novel delivery platform, tumor endothelial marker 1 (TEM1)/endosialin (Ab-/scFv) conjugated with MPSs to target ovarian cancer cells. The resultant nanosized MPSs exhibited a controlled release of bevacizumab (BVC) in pH 7.4 for promising anticancer efficacy profile. Furthermore, it increased the cellular uptake and intracellular distribution of BVC in Ovcar5 cancer cells. Overall, Ab-/scFv-conjugated MPSs showed a superior anticancer effect with profound apoptosis as effective strategy for ovarian cancer treatment (Zhang et al. 2015). Certain proteins are directly involved in the accelerated metabolism of tumor cells and can be employed to active targeting (Gupta and Roy 2021). Transferrin (Tf) is blood-plasma glycoprotein, which involves in the transportation of iron into cells and epidermal growth factor (EGF). It stimulates cell growth and differentiates with high demand of this protein by cancer tissues. Recently, Vallet-Regí group developed a nanoplatform for the cancel cell targeting nucleation using MPSs decorated with Tf followed by the immobilization of Ag NPs ((MSNs-Tf-AgNPs). The receptor-mediated endocytic mechanism facilitated the internalization of the NPs, and the transported Ag NPs dissolved toxic Ag+ ions within the lysosomes during the retention time by following the “lysosome-enhanced Trojan horse effect.” This affected the cell proliferation, and the transcripts involved in cell cycle regulation were validated by quantitative proteomics using qPCR (Castillo et al. 2019).

1592

K. Kannan

Peptides are short amino acid sequences usually less than 50, which have many similarities with their parent proteins. The use of peptides has significant advantages such as easy synthesis, conjugation, lower cost, and reduced immune response. Hu et al. (2016) constructed the drug delivery system using polydopamine (PDA) coated MPSs and functionalized with Asn-Gly-Arg (NGR), a ligand able to target cancer neovasculature throughout the cluster of differentiation 13 (CD13) receptor. In vitro study of this delivery system showed an enhanced intracellular accumulation in primary brain capillary endothelial cells (BCECs) and C6 cells with higher bloodbrain barrier (BBB) permeability. Furthermore, the strong antiangiogenesis and more antitumor efficacy of this system indicated that the dual-targeting vehicles are potentially useful in brain glioma therapy (Hu et al. 2016). Aptamers are a special class of nucleic acids that have higher affinity to specific target molecules with great structural flexibility and thermal stability. All these features make aptamers highly valuable components to active targeting delivery system. Tang et al. (2015) developed a novel photoresponsive aptamer-targeting drug delivery system. The Cy5.5-AS1411 aptamer conjugated on the surface of graphene oxide grafted and doxorubicin (Dox)-loaded MPSs through noncovalent assembly (MPSs-Dox@GO-Apt). The “off–on” switches of the MPSs-Dox@GOApt could be controlled by aptamer targeting and laser irradiation, respectively. Aptamer Cy5.5-AS1411 provided the nucleolin specific targeting with real-time indicator abilities and the GO acted as gatekeeper to prevent the DOX release in the absence of laser irradiation. Interestingly, with an increase in laser power the system showed synergism of chemotherapy and photothermal therapy for effective treatment of cancer cells (Tang et al. 2015). Li et al. (2017) fabricated a promising therapeutic platform for epithelial cell adhesion molecule (EpCAM) positive colorectal cancer (CRC). Maytansine derivative (DM1) was loaded into the pores of MPSs and the surface was decorated with hydrochloride dopamine (PDA), PEG, and EpCAM aptamer (APt) for the active targeted treatment of CRC. In an in vivo study of this system, the PDA was used as a pH-sensitive gatekeeper to release the DM1, and EpCAM APt-guided active targeting increased the delivery of DM1 to CRC as well as reduced toxicity by minimizing the exposure of normal tissues to DM1 (Li et al. 2017).

Gate Keeping of Drug Loaded MPSs The major advantages of MPSs as drug delivery systems is the possibility to design zero premature drug release by blocking the pore openings using suitable gatekeepers. In these systems, the pores of MPSs can be blocked by the blockers such as biomolecules, nanoparticles, linear molecules, cyclic molecules, polymers, and polyelectrolyte layers. These blockers are usually attached with MPSs via electrostatic interactions, hydrogen bond, or covalently, which are being easily detached or cleavable by action of internal or external stimuli. Kim et al. fabricated the glutathione (GSH) stimulus-responsive MPSs, in which the gate keeper β-cyclodextrin (β-CD) was covalently attached on surface of particles via disulfide

62 Using Smart Mesoporous Silica in Designing Drug Delivery Systems

1593

bonds. In vitro study showed that the addition of GSH could remove the gatekeepers via the cleavage of disulfide bond between the moiety and then release the anticancer drug in the pore (Kim et al. 2010). A high loading HMS-based delivery system was developed for both hydrophobic and hydrophilic drugs using noncovalently bound PEG-PDS-DPA copolymer as gatekeeper. At neutral pH, the PEG-PDS-DPA polymer forms a dense layer on the surface of HMSs and blocks the pores to prevent the drug leakage, but in low pH = 5.0–5.5, the disulfide bonds in the polymer are degraded to release the drug from the HMSNs (Palanikumar et al. 2017). PEGylated interior surfaces coated with poly (N-isopropylacrylamide) (PNIPAm) as gatekeepers for MPSs were developed for the thermo-sensitive drug delivery of bovine hemoglobin. Surface treated cadmium quantum dots have been used as photosensitive gatekeepers with the ability to avoid capillary condensation in the MPSs pores for anticancer treatment using camptothecin (Nik et al. 2020). Redox-responsive glycoprotein transferrin gatekeeper was applied as targeting ligand on MPSs and showed the targeted drug delivery of doxorubicin and doxorubicin in hepatocellular carcinoma cells. MPSs end capped with natural chitosan were stimulated to unblock for delivering drug by creating the acidic environments due to lysozymes activity. Lysozymes are inherent immune system enzymes, which are overproduced by myelomonocytic leukemia. MPSs capped with a peptide sequence containing Arg-Gly-Asp acting as a gatekeeper were demonstrated to have significant drug release ability in response to enzyme activity, where the peptide functioned capping as a triple targeting motif via integrin binding in apoptotic cells. The release from nanogold capped to the MSNs through electrostatically or covalently with 3-(propyldisulfanyl)ethylamine was monitored using differential interface contrast microscopy. In uncapping process associated with endocytosis, the reducing reagents such as dithiothreitol and GSH target the disulfide bond and cleave the nanogolds from MPSs channels (Nik et al. 2020) (Table 1).

Stimuli-Responsive Smart Delivery System Nowadays, the designing of smart MPSs drug delivery system that responds to specific stimuli has attained a lot of attention due to its incredible promises in diagnostic level as well as in therapeutic applications. However, these platforms lack the ability to transport the therapeutic molecules without any leakage until they reach the targeted area. Besides that, these systems hardly accomplish the higher release of drugs in restricted area and with temporal control. Taking these into account, the fabrication of responsive MPSs system on demand can regularize the circulating drug concentration in the body to eliminate the side effects against the healthy cells and decomposition and denaturation of the drugs. Stimuli-responsive behavior can be accomplished by blocking pores throughout using linkers, which can be cleaved upon exposure to given stimuli. The stimuli are classified into internal stimuli, that is, class of the treating pathology such as pH, redox potential, and enzymes, and external stimuli such as light, temperature, magnetic field, and ultrasounds. These applied stimuli response lead to alteration in surface structure

1594

K. Kannan

Table 1 Various gatekeepers for stimulus responsive triggered drug release from MPSs (Nik et al. 2020) Gate keeper Chitosan Polyaspartic acid Chitosan oligosaccharide TPGS (d-a-tocopheryl poly-ethylene glycol 1000 succinate) Zeolitic imidazolate framework (ZIF)-8 β-cyclodextrin ((β-CD)- PEG Blue Fluorescent N-GQDs, Hyaluronic acid Graphene quantum dots (GQDs) Polydopamine Carbon Quantum Dots Poly (allylamine hydrochloride)/Sodium Poly (styrene sulfonate) Natural gelatin Poly (poly (ethylene glycol) methylether methacrylate-co-poly-(2methacryloxyethoxy)benzaldehyde) Amphiphilic peptide (P45) Polydopamine Cross-linked chitosan Sodium Alginate Curcumin polymer Fe-BTC and Zn-BTC Modified (PGMA)-Cucurbituril Polydopamine Poly ((N-isopropylacrylamide)-co-methacrylic acid) Bismuth Sulfide 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan2-yl) benzyl acrylate Hyaluronic Acid Hybrid Lipid d-α-tocopherol polyethylene glycol 1000 succinate (vitamin E) Transferrin (Tf) PEGylated organosilica Polyacrylic acid Carboxymethyl β-cyclodextrin (CβCD)

Stimulus pH pH pH pH

Loaded drug DOX DOX DOX DOX

pH pH pH pH pH pH pH

DOX DOX DOX DOX DOX DOX DOX

pH pH

DOX DOX

pH pH-GSH pH-GSH pH-GSH pH-GSH pH-Liposome pH-Host/Guest Interactions pH-Ultrasound pH-Temperature

DOX DOX DOX DOX DOX DOX DOX

NIR Irradiation Temperature-ROS

DOX DOX

DOX DOX

Enzyme DOX Oxidation/Reduction DOX Oxidation/Reduction/NIR DOX GSH GSH pH pH

DOX DOX Etoposide Etoposide (continued)

62 Using Smart Mesoporous Silica in Designing Drug Delivery Systems

1595

Table 1 (continued) Gate keeper Polyethylene glycol poly (ethyleneimine) conjugated folic acid CdS β-cyclodextrin GST (glutathione-S-transferase)-HER2-Afb

Stimulus pH pH UV Irradiation Two-photon excitation Enzyme

Loaded drug Celastrol Erlotinib Camptothecin Camptothecin DOX-Camptothecin

of MPSs and release the drugs in the desired area or tissue. Thus researchers are working toward enhanced targeting delivery and avoid the premature drug by applying a combination of several stimuli-responsive strategies (Abdo et al. 2020; Guimarães et al. 2020).

Internal Stimuli-Responsive Delivery The metabolic and biochemical variations between normal and pathological conditions (e.g., environment around tumors and inflammation sites) have been exploited as the response for drug delivery systems without requirement of external mediation to trigger drug release. In this concern, researchers have focused on the design of MPSs that respond to internal stimuli such as pH, redox potential variations, and deregulations of various proteins (mainly enzymes) or small biomolecules.

pH Responsive From literature, pH is one of the most employable internal stimuli, especially associated with drug release for treatment of cancer and inflammations. Tumors exhibit (environment pH = 6–7) significant variations in pH values compared with normal tissues (pH = 7.4), especially in intracellular lysosomes (pH = 30

No

OFF the Air Exhaust

Yes

No

Analysis of data

OFF the Water Pump

ON the Air Exhaust

Data set given to the machine learning

Send SMS TO Farmer

Training the Algorithm With Real time data set

Prediction of the Values

MQTT Collection of data from the sensors

Interface the data to the cloud using MQTT

Stop

Fig. 4 Algorithm for the proposed system

to his/her personal mobile. Even in case of temperature say for a particular crop if the greenhouse temperature should not be more than 30 ◦ C the automatically the exhaust fan will be activated to reduce the greenhouse atmosphere temperature. Even here the farmer will be getting an offline SMS to his/her personal mobile. The collected data is also further processed and given as the dataset for algorithms for machine learning (decision tree and random forest). Both the algorithms are trained with the real time data set and the parameters are predicted. As well as the accuracy of the algorithms is also generated. The decision tree algorithm is a means of supporting decisions that have a map or decision-making models like a tree with chance results, resource costs, and utility. A decision tree is a structure that measures a particular attribute of each node, each branch is the outcome and each node is a class label as shown in Fig. 5. The algorithm steps are as follows: Step 1: Start the root node tree, says S that contains the entire dataset. Step 2: Use Attribute Selection Measure (ASM) to find the best attribute in the dataset. Step 3: Divide the S into subsets containing the best attributes possible values. Step 4: Build the node for the tree, with the highest attribute. Step 5: A set of latest decision trees are recurrently being made using the data set sub-sets generated in step 3.

63 IoT-Based Smart Farming System Using MQTT Protocol and ML Algorithms

1619

Fig. 5 Decision tree algorithm

Continue until a step has been achieved in which nodes are not further classified and the final node is called a leaf node. Random Forest algorithm is an ensemble tree-based learning algorithm, which combines multiple classifiers in order to overcome a complex issue and improve model efficiency. In two steps, firstly, by combining N decision tree, it will construct the random forest and secondly predict each generated tree during the first step. In the following steps, the work cycle can be clarified as shown in Fig. 6. Step 1: From the training set pick K random points. Step 2: Create decision-making trees for the chosen points. Step 3: Choose the N number to create the N number of decision trees. Step 4: Repeat Step 1 & 2. Step 5: Find out the forecasts of each decision tree with new data points and allocate the new data elements to the majority voting list.

Experimental Analysis of Smart Farming The experimental setup consists of a processing unit with a sensor network that read samples for every 13 s from the smart farm/greenhouse system. For the realtime monitoring, the processing unit also updates the ThingSpeak server. As well as the customized designed Python Flask webpage for the farmers will also be updated. The values are taken from Hyderabad, Almasguda cabbage field in realtime at different time stamps on different dates. The data for the proposed system is collected for over 15 days for 24 h of each day. The entire hardware setup of the smart farm/greenhouse system is shown in Figs. 7 and 8 with laptop.

1620

S. P. and A. M. N.

Fig. 6 Random forest algorithm Fig. 7 Hardware Setup of Smart Farm/Greenhouse System

Data Acquisition Through Different Field Trials ML depends heavily on data, without data, an artificial intelligence can’t learn. This is the most critical part of algorithm preparation. Here the data collection can be called as data set. In simple terms, the data set is consistent with the content of a single database table or the single statistical data matrix in which each column of the table represents the particular variable, and each row corresponds to that

63 IoT-Based Smart Farming System Using MQTT Protocol and ML Algorithms

1621

Fig. 8 Hardware Smart Farm/Green House System in Farm

Fig. 9 Cabbage Farm during the field trial visit

particular data member. As the data collections play an important role in machine learning, for this smart farm system the real-time data is collected from various field trials on different days at different time slots. Initially, for testing purposes, the data was collected from the home garden at the location of 17.4486◦ N, 78.3908◦ E coordinates, i.e., at Madhapur. The field trial visit was to cabbage farm located at 17.3125◦ N, 78.5363◦ E as shown in Fig. 9. The visit was on 26/02/2020 and the time of visit was around 13:30 h. The data was collected for more than an hour from different locations of the farm for a period of interval. The data was directly exported to the ThingSpeak Server from the farm using MQTT protocol by placing the hardware as shown in Fig. 10. Around more than 540 data points are collected at the first field trial of all the parameters that were required for the proposed system. The reason to select a cabbage farm is the amount of fertilizer effect on the crop concerning other crops. Since cabbage grows like a flower, it’s the vegetable that has petals of leaves covered over it. So whenever the farmer sprays the fertilizer to protect the crop from other weeds, pests, and other insects, the fertilizer concertation goes into the crop even and stays as under the leaves.

1622

S. P. and A. M. N.

Fig. 10 Data Acquisition during the field trials

Design of 3D Printing Prototype As when the two field trials were visited a problem statement came into observation, i.e., as the proposed system needs to place in an agricultural field, the MCU and the sensors need to place in different locations. But few locations can be on the wet ground surface, in the middle of the crop where water, fertilizer will be sprayed due to which the system can be damaged. As the technology is growing day by day an MCU canister is designed in the PTC Creo tool according to the proposed system specifications by measuring the dimensions of the components. The entire MCU Canister in different views are shown in Fig. 11. Figure 11a–e shows the side, front, top orientation, and assembly views of the MCU canister, respectively. After the 3D model was designed in the Creo tool, the part file (i.e., prt file) is generated using the tool itself. Later this part file is given as input to the MakerBot 3D printer for printing. The part file is transferred over Wi-Fi to the 3D printer. Before printing the 3D printer is loaded with polylactic acid (PLA) material. It is available in different colors. But for the system design, the opted color is blue. After loading with the filament the 3D printer extruder is allowed to preheat up to 180 ◦ C which is the initial printing temperature. When the print job is transferred the extruder heat up to 220 ◦ C to begin printing since it is the melting point for the PLA. The top and bottom files were 3D printed with different specifications on different build plates. The top part was with 80% infill which took around 8 h of

63 IoT-Based Smart Farming System Using MQTT Protocol and ML Algorithms

1623

Fig. 11 (a) Side view of the MCU Canister. (b) Front view of the MCU Canister. (c) Top view of the MCU Canister. (d) Orientation view of the MCU Canister. (e) Assembly view of the Entire MCU Canister

printing time. The bottom of the MCU canister is 3D printed with 60% infill of the material and the time taken to print was 28 h. The final 3D printed model is shown in Fig. 6.6 with top view in Fig. 12a The model designed to hold the system MCU, sensors, and relay. The system is also

1624

S. P. and A. M. N.

Fig. 12 (a) Top View of the 3D printed model. (b) Components Placement in the Canister

designed with good air ventilation on the top part of the MCU canister since the Raspberry Pi generates heat. The canister can accommodate all the sensors, wires, and relays. The Raspberry Pi is fixed to the base of the canister with M2.5 screws and nuts so that even in any movement the Raspberry Pi will not be damaged. Accordingly, all the other sensors are fixed and wired as shown in Fig. 12b. The top and bottom parts are fixed with an M2.5 head long screw and supported by clamp fit.

63 IoT-Based Smart Farming System Using MQTT Protocol and ML Algorithms

1625

Temperature and Humidity Sensor Initially, the DHT-11 sensor module measures the temperature and humidity of the surrounding environment. The surrounding temperature is measured as a special temperature must be held for a particular farm in greenhouse. Not only that an optimum temperature required in for a crop to give a good yield (Agritech 2020). Higher temperatures negatively impact mineral food, shoot growth and the growth of pollen, resulting in low yields (Murari et al. 2018). Using MQTT protocol the measured values are updated for every 13 s in the ThingSpeak Server. The temperature values (◦ C) and the humidity values (%) are updated in field 1 and field 2 respectively as shown in Fig. 13. The temperature value at the time instance of 18:25 h is 24 ◦ C and from field 2 at the same instance of time is 69% humid. But in the same graph at 18:40 h the temperature raised to 25 ◦ C and at that same interval the humidity of atmosphere is gradually getting reduced to 65%.

Fig. 13 Measured temperature and humidity values with respect to the Time in field 1 and field 2, respectively

1626

S. P. and A. M. N.

Field 3 Chart

Pressure(Pa)

Pressure 100

0 18:25

18:30

18:35

18:40

Time(Hrs) ThingSpeak.com

Fig. 14 Pressure Sensor output in ThingSpeak Server with field 3 Pressure vs. Time

Pressure Sensor As shown in Fig. 14, field 3 in the ThingSpeak Server are updated with air pressure values. The server is getting updated every 13 sec. The time at which 18:25 h air pressure in the farm 50 Pa. Since the values are taken in different locations of the farm the pressure values are varying from 100 Pa and less than 100 Pa. As observed in Fig. 12, as time passes the pressure is toggling between 50 Pa and 10 Pa, i.e., at 18:40 h the pressure is around 40 Pa.

Gas Sensor The fertilizer content sprayed in farm is measured using MQ-2 gas sensor. Farmer sprays the fertilizer in the farm thrice a week to protect from pests and other insects. But spraying fertilizer thrice a week is not good for the farm as it will grow with the chemical content in them. Due to growing with chemical the vegetables will spoil the heath of human beings. So monitoring the fertilizer level in the farm is very important. As shown in Fig. 15, field 4 in the ThingSpeak Server are updated with fertilizer content value. At the time interval of 18:25 h the fertilizer sprayed content is 195 ppm and gradually decreasing to 178 ppm and in various fields of farm it is different.

Soil Moisture Sensor, Status of Motor, and Location Finally, the soil moisture values in percentage, as well as the motor status i.e., the motor is ON or OFF of the farm are updated in the Server, as shown in Fig. 16. It is observed that the value of field 5 at time 18:25 h is 39% moisture in the soil. At the

63 IoT-Based Smart Farming System Using MQTT Protocol and ML Algorithms

1627

Field 4 Chart Fertilizer Content Fertilizer(ppm)

190

180 18:25

18:30

18:35

18:40

Time(Hrs) ThingSpeak.com

Fig. 15 Gas Sensor output in ThingSpeak Server with field 4 Fertilizer vs. Time Fig. 16 Soil Moisture Sensor output in ThingSpeak Server with field 5 Soil Moisture vs. Time and field 6 Motor Status vs. Time

same time since the soil moisture is less than 50% the motor status is 1 i.e., it is high so the water pump is activated. As we take the parameter values at a different point in the farm then soil moisture content is more than 50% so the water motor gets

1628

S. P. and A. M. N.

Fig. 17 Channel location field

cut-off at 18:37 h. The sensor reads the soil moisture value in analog so to convert into percentage the Eq. 1 is used. mp = (sa/1023.00) ∗ 100

(1)

In Eq. 1, mp is moisture value in percentage where this variable is sent into ThingSpeak server and sa is sensor analog input value. The analog value is divided 1023 so the value of soil moisture will be converted into digital and multiplied by 100 so that the percentage of the water content in soil is measured. The location of the farm is located 17.3125◦ N, 78.5363◦ E as shown in Fig. 17.

Messaging Services As shown in Fig. 18, the farm parameters will be sent to farmers personal mobile with the help of twilio account. Whenever the status of water motor or the exhaust fan will change from ON to OFF or OFF to ON the SMS will be triggered from the system to the farmer including all the parameters like temperature, humidity, air pressure, fertilizer content and soil moisture of his/her farm/greenhouse.

ThingSpeak Mobile Application The monitoring of smart farm parameters using ThingSpeak mobile application will be useful for the farmers. After installation, the authorized users can log in and access their account of their smart farm/greenhouse. Different channel ID’s are created to monitor the farm parameters as shown in Fig. 19. After successful creation of the channel ID’s all the parameters are visualized on the graphs of the application as shown in Fig. 20.

63 IoT-Based Smart Farming System Using MQTT Protocol and ML Algorithms

1629

Fig. 18 SMS Alert to Farmer Mobile

Design of Webpage The customized web Page for farmers is designed using Python Flask API and HTML Bootstrap 4. As the customized webpage uses python to run it on the web browsers dynamically Ngrok tunneling is activated on the Raspberry Pi so that on that tunneling web address the customized web page will be published. The customized web page has different web pages as shown in Fig. 21 (a) login page, (b) Home page, (c) status page and (d) control page. The login page, home page, status page and login pages of the customized web pages are created. Nowadays as the internet is available everywhere the farmers can open the webpage on their smartphones also so accordingly the user interface will adjust according to the screen dimensions.

1630

S. P. and A. M. N.

Fig. 19 ThingSpeak App installation

Decision Tree Algorithm The Data collected from different field Trials at different intervals of time is given to Decision Tree Machine Learning Algorithm to predict the parameters, accuracy of prediction, Mean Absolute Error (MAE), and Mean Squared Error (MSE). Tabulation of the predicted parameters during the field trials using the decision tree algorithm is listed in Table 1. The average of the actual parameters and predicted parameters with error is calculated. For Decision Tree Algorithm the accuracy for one day of the parameters is observed as 97.83%.

Random Forest Algorithm The Data collected from different field Trials at different intervals of time is given to the Random Forest Machine Learning Algorithm to predict the parameters, accuracy of prediction, Mean Absolute Error (MAE), and Mean Squared Error (MSE).

63 IoT-Based Smart Farming System Using MQTT Protocol and ML Algorithms

1631

Fig. 20 All the Parameters in ThingSpeak mobile application

Tabulation of the predicted parameters during the field trials using a random forest algorithm is listed in Table 2. The average of the actual parameters and predicted parameters with error is calculated. For the Random Forest Algorithm, the accuracy of the parameters is observed as 98.76%.

Comparison of ML Algorithms Accuracy The data acquisition for different days in field trail visits are giving to both the algorithms and the accuracies of both the algorithms are calculated and compared as shown in Fig. 22. The comparison for 5 days, 10 days, and for 18 days’ data samples accuracies, MAE, MSE are shown in Table 3.

1632 Fig. 21 (a) Login Page. (b) Home Page. (c) Status Page. (d) Control Page

S. P. and A. M. N.

63 IoT-Based Smart Farming System Using MQTT Protocol and ML Algorithms

1633

Table 1 Comparison of Actual and Predicted Parameters for decision tree algorithm Parameters Temperature (◦ C) Humidity (%) Fertilizer Content (ppm) Soil Moisture (%)

Actual 25.9417 66.647 223.3676

Predicted 25.9411 66.4986 224.6176

Error 0.0294 0.1484 1.25

MAE 0.12 1.03 2.04

MSE 0.44 2.12 28.19

59.9664

56.1323

3.8341

4.91

18.5

Table 2 Comparison of Actual and Predicted Parameters for random forest algorithm Parameters Temperature (◦ C) Humidity (%) Fertilizer Content (ppm) Soil Moisture (%)

Actual 25.411 66.3235 224.397

Predicted 25.4201 66.4411 224.9183

Error 0.008 0.1176 0.5213

MAE 0.06 0.81 1.36

MSE 0.03 1.17 4.35

59.3235

58.7508

0.5727

2.32

5.11

Fig. 22 Comparison of Accuracy’s for 18 days

Summary In this Chapter, the proposed system for smart farm/greenhouse environment monitoring and controlling is a secured and robust IoT solution in real-time. The developed system has Raspberry Pi 4 as a target board and interfaced with several sensors and actuators successfully. A 3D prototype model for the proposed system is successfully designed and developed. A web-based application, i.e., ThingSpeak with MQTT protocol is used for sending the parameters like temperature, humidity, the pressure of the surrounding atmosphere, fertilizer content sprayed, and soil

1634

S. P. and A. M. N.

Table 3 Comparison of Algorithm Parameters for several days Algorithm Parameters Decision Tree Accuracies (%) Algorithm MAE MSE Random Accuracies (%) Forest MAE Algorithm MSE

Number of Days 5 Days 97.93 2.025 4.91 98.76 1.137 3.76

10 Days 98.18 1.292 2.535 98.35 1.137 1.59

18 Days 90.86 10.347 21.13 92.03 8.412 19.54

moisture through the web server and also to the customized webpage which is designed for the farmers. Using ThingSpeak smartphone application these measured parameters also be monitored and the required SMS alert is generated when there is a change in the parameters. Various field trials are performed for the cabbage farm to measure the vital parameters. The measured temperature value ranges from 24 ◦ C to 30 ◦ C, humidity ranging from 46% to 71% for location at 17.3125◦ N latitude and 78.5363◦ E longitude. The fertilizer content sprayed 190 ppm to 160 ppm, soil moisture from 25% to 100% in different parts of the farm on 21st February 2020 at 3:30 PM. In order to predict the vital parameters, the supervised machine learning algorithms such as decision tree and random forest are implemented successfully. The obtained accuracy, mean absolute error and mean square error for decision tree algorithm 90.86%, 10.347 and 21.13 respectively and similarly for random forest algorithm the obtained accuracy, mean absolute error and mean square error are 92.03%, 8.412 and 19.54, respectively. The random forest algorithm gives better performance in terms of accuracy, mean absolute error, and mean square error compared with the decision tree algorithm.

References Agritech (2020) Influence of climate on crops. http://agritech.tnau.ac.in/agriculture/ agri_agrometeorology_temp.html Benyezza H, Bouhedda M, Djellout K (2018) Smart irrigation system based on ThingSpeak and Arduino. In: International conference on applied smart systems, IEEE conference, 24–25 Nov 2018 Carlos J, Mendoza-Moreno MA, Faeq A, Arun Kumar N, Ramirez-González G (2018) An IoTbased traceability system for greenhouse seedling crops. In: IEEE access on new trends in brain signal processing and analysis, October 18 Foundation, Raspberry Pi, Raspberry Pi 4 Model B specifications. 2019. https:// static.raspberrypi.org/files/productbriefs/200521+Raspberry+Pi+4+Product+Brief.pdf Gas Sensor (MQ-2) specifications. 2020. https://components101.com/mq2-gas-sensor Kishore R, Samar Sarjerao B (2017) A low-cost smart irrigation system using MQTT protocol. In: IEEE region 10 symposium, 19 Oct 2017 Kishore R, Soratkal SR (2017) MQTT based home automation system using ESP8266. In: IEEE region 10 humanitarian technology conference (R10-HTC), 24 Apr 2017

63 IoT-Based Smart Farming System Using MQTT Protocol and ML Algorithms

1635

Kranthi Kumar V, Sai Sandeep R, Ramanjinailu R (2019) Measuring soil moisture using thingspeak by IoT sensing device. Int Res J Eng Technol 6(2):927–928 Murari KK, Mahato S, Jayaraman T, Swaminathan M (2018) Extreme temperatures and crop yields in Karnataka, India. Rev Agrarian Stud 8(2):92–113 Python Documentation. 2020. https://docs.python.org/3/ Raghav Kumar T, Aiswarya B, Suresh A, Jain D, Balaji N, Sankaran V (2018) Smart management of crop cultivation using IOT and machine learning. Int Res J Eng Technol 5(1):845–850 Shete R, Agrawal S (2016) IoT based urban climate monitoring using Raspberry Pi. In: International conference on communication and signal processing, 6–8 Apr 2016, India Vij A, Vijendra S, Jain A, Bajaj S, Bassi A, Sharma A (2020) IoT and machine learinng approaches for automation of farm irrigation system. Procedia Comput Sci 167:1250–1257 Yokotani T, Sasaki Y (2016) Comparison with HTTP and MQTT on required network resources for IoT. In: ICCEREC

Sustainability of Fusion and Solid-State Welding Process in the Era of Industry 4.0

64

Vijay S. Gadakh and Vishvesh J. Badheka

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industrial Revolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industrial Revolution: A Welding Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industry 4.0 in Fusion and Solid-State Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industry 4.0 in Fusion Welding Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industry 4.0 in Solid-State Welding Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1638 1638 1640 1642 1643 1645 1646 1650 1650 1651

Abstract

Manufacturing is considered as the heart of any industrialized developed nation. Welding and joining, being a part of manufacturing activity, contributes a large for the country’s long-term growth. In order to sustain in the global competitive environments, welding industries nowadays continue to strive hard for finding new ways do the same function either by mimics from existing product or adopting the concept from nature. This includes not only the cost and energy saving, waste minimization, optimal parameters, and eco-friendly but also with the advent newer computer technologies like smart factories or factories of the

V. S. Gadakh () Department of Mechanical Engineering, Amrutvahini College of Engineering, Ahmednagar, Savitribai Phule Pune University, Pune, India e-mail: [email protected] V. J. Badheka Department of Mechanical Engineering, School of Technology, Pandit Deendayal Energy University, Gandhinagar, Gujarat, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_113

1637

1638

V. S. Gadakh and V. J. Badheka

future, industrial internet of things (IIoT), cyber-physical systems, cloud systems, big data, digital manufacturing, and lastly the Industry 4.0. The present work elucidates the different approaches employed for assessing and improvement in the weld quality using meta-heuristic algorithms, in process weld quality inspection; defects detection and its control in fusion and solid-state welding processes. Keywords

Sustainability · Fusion welding · Solid-state welding · Industry 4.0 · Industrial internet of things · Cyber-physical systems · Cloud systems · Big data · Digital manufacturing

Introduction Manufacturing is considered as the heart of engine development of any developed nation which is directly related to its financial well-being. It consists of use of mechanical, physical, and chemical processes to alter the properties and geometry of raw material into the end products (Rao 2007). Mainly, manufacturing process involves casting, forming, machining, and joining processes. Out of these manufacturing processes welding and joining play an important role in major industries like automobile, aviation, construction, defense, oil and gas, naval, space, etc. which contribute a large for the country’s long-term growth. The current scenario of welding process in different sectors in Indian context is explained in Fig. 1 where few newer technologies have been introduced like dynamic oxide control system for aluminum welding, intelligent gas control for large-scale savings in gas consumption, and intelligent arc control for automatic arc adjustment. Similarly, robotics and automation is being deployed by the industries for achieving the high weld quality and productivity improvement. Similarly, welding process related expenses contributed significantly toward the US economy which shows one-third of the total US Gross Domestic Product (Miller et al. 2002).

Industrial Revolution From the industrial perspective, a revolution is nothing but a significant change and growth in technology and thereby standard of living of the people. This change in the technology takes place in following ways. The first one called innovation which means modification of the existing product by either conventional trial-anderror approach or process parameter optimization or Jugaad (Frugal Engineering) (Singh et al. 2012; Prabhu et al. 2012) or TRIZ approach (Kohnen 2004). The second one adopting concepts from nature (Wahl 2004). The details of different industrial revolutions are depicted in Fig. 2. The first phase of industrial revolution caused by invention of steam engines led to an increase in production called as “Industry 1.0.”

64 Sustainability of Fusion and Solid-State Welding Process in the Era. . .

1639

Others Shipbuilding Construction Railways Automotive Heavy Engineering 0%

FY 2020 FY 2016

Heavy Engineering 33% 31%

5%

10%

15%

20%

25%

30%

35%

Automotive

Railways

Construction

Shipbuilding

Others

23% 20%

8% 6%

11% 10%

4% 3%

21% 30%

Fig. 1 Current scenario of welding process in different sectors in Indian context

Fig. 2 Industrial revolutions (Petrillo et al. 2018)

In the second revolution, significant changes were made caused by inventions of electricity and steel making processes and is called as “Industry 2.0.” In the third revolution, a rapid transformation happened caused by the application of electronics in manufacturing processes and is called as “Industry 3.0.” In the current fourth revolution, transformation from machine led manufacturing to digital manufacturing is called as “Industry 4.0” (Gunal 2019) and is based on nine technologies as shown

1640

V. S. Gadakh and V. J. Badheka

Fig. 3 Nine pillars of fourth industrial revolution

in Fig. 3. The future industrial revolution is known as “Industry 5.0” wherein with the significant development in the robotics field humans will be able to accomplish considerably better than the Industry 4.0 (W˛eglowski 2018).

Industrial Revolution: A Welding Perspective In parallel with different industrial revolutions, significant technology also had developed in the welding and Joining. Though the joining of pieces together can be traced back to more than 2000 years, welding emerged as a feasible manufacturing process only in the late 1800s. It is at the heart of many great engineering achievements. It is important for agriculture mechanization, energy generation, clean water supply, and medical devices production (Debroy 2015). The development of fusion and pressure welding processes was caused by eruption of World Wars I and II as shown in Table 1 (W˛eglowski 2018). When we say the technology or industrial revolution gets advanced or upgraded, it means that the things are easy-to-use. This also results in resolving the problems or issues that were found in the past technology. This advancement in the technology is possible when the manufacturing firms involve the customers or the end users or receive feedback from the customers on the concept design phase of design cycle. In the fourth industrial revolution, most of the welding industries are striving hard to achieve the production efficiency as well as quality effectiveness across

64 Sustainability of Fusion and Solid-State Welding Process in the Era. . .

1641

Table 1 Development of welding technologies (W˛eglowski 2018) Year 1802 1885 1895 1901 1902 1904

Country USA Russia Kingdom of Poland Germany France Sweden

Inventor Robert Hare Nikolaj Benardos Stanislaw Olszewski Hans Goldschmidt Charles Picard Edmond Fouche Oscar Kjellberg

1926

USA

1929

USSR

H. M. Hobart P. K. Devers D. A. Dulczewskij

1948

USA

P. K. Devers

1949

USSR

1949

Germany

B. E. Paton G. Z. Woloszkiewicz K. H. Steigerwald

1953 1953

USA USSR

1953

USA

1962 1965

USA USA

1970

Great Britain

M. S. Holander Robert Soloff and Symour Linsley Martin Adams

1977 1981 1991

Germany Canada UK

W. M. Steen John G. Church Wayne Thomas

2004

Austria

Manfred Schorghuber

Robert M. Gage K. V. Lyubavskii N. M. Novoshilov A. A. Bernard

Process Oxy-Hydrogen Welding Carbon Arc Welding Thermit Welding Oxyacetylene Welding Manual Metal Arc Welding (MMAW) Gas Tungsten Arc Welding (GTAW) Submerged Arc Welding (SAW) Gas Metal Arc Welding (GMAW) Electroslag Welding Electron Beam Welding (EBW) Plasma Welding CO2 shielded MAG Welding Self-shielded Flux Cored Arc Welding (FCAW) Friction Welding (FW) Ultrasonic welding Laser Beam Welding (LBW) Hybrid Laser Welding T.I.M.E. Welding Friction Stir Welding (FSW) CMT Welding

the globe. On the contrary, few of the small-scale industries (SSI) are in dilemma whether to go or not with such industrial revolution due to high initial investments in such a welding metamorphism world. It is well agreed that at the start of any of the industrial revolution, initially the cost of the product is comparatively higher than the previous one. This cost is acceptable over a period of time, which in turn, the living standard of people and hence significant development of that country. W˛eglowski (2018) suggested few approaches related with welding engineering of the fourth industrial revolution as: (1) welding technologies should move from

1642

V. S. Gadakh and V. J. Badheka

traditional push forward approach to acquisition approach, (2) intelligent robots and software programs flexibility, (3) welding programs integration into “smart” production processes, and (4) flexibility of welding software provider. Digitization, control systems, and transfer of data are the important measures of the fourth industrial revolution. Selection of weld process parameters are generally based on parent material, thickness, and other related parameters that can be successfully implemented when welding knowledge is digitized accordingly. As explained earlier most of the welding manufactures are already taking inventive steps to store welding data intelligently which will be ready for documentation and data analysis. Still, at which level of detail the data must be available has not yet been precisely evaluated for the creation of appropriate algorithms (Posch et al. 2018). On the contrary, lot of development also has been done in the virtual simulation module, for example, augmented reality and virtual reality (AR/VR) for training purpose. It consists of doing virtual welding where a welding torch is controlled by simulating it in tutorial and training mode, its evaluation, etc. (Torres-Guerrero et al. 2019).

Sustainability More than one million workers across the globe are presently engaged as full-time welders. A number of epidemiologic studies have reported that a higher incidence of respiratory illness in welders takes place (Antonini et al. 1996), and there is an ardent need of innovative welding technologies that not only create elegant products but indeed eliminate these health- and environment-related issues. Hence, sustainability or sustainability development is vital for welding industries in any of the industrial revolutions. As it has a variety of dimensions, important aspects, performance measures and it is dynamic one. Although in manufacturing studies economic, societal, and environmental sustainability have given a prime importance (Bi 2011), still sustainability is considered a complex, unstructured issue and is a vital issue for the current and future perspectives (Garetti and Taisch 2012). As of now, there is no specific definition for sustainability or sustainability development. According to the past studies (Garetti and Taisch 2012; Agustiady and Badiru 2013) sustainability is a way for improving living standards and comfort for the current and future perspectives. In the global competitive environment, in order to be sustainable , manufacturing industries need more efforts in terms of time and cost for their products (Garetti and Taisch 2012; Garbie 2013). Kamble et al. (2018) suggested a framework for sustainable industry 4.0 which consists of Industry 4.0 technologies, process integration, and sustainable outcomes as depicted in Fig. 4. Their proposed framework considers that Industry 4.0 provides the business units integration through the cyber-physical systems which make the manufacturing system more flexible, efficient, and eco-friendly.

64 Sustainability of Fusion and Solid-State Welding Process in the Era. . .

1643

Fig. 4 Framework for sustainable Industry 4.0 (Kamble et al. 2018)

Industry 4.0 in Fusion and Solid-State Welding Globally, in the arena of fourth industrial revolution welding equipment manufacturers are already supplying their consumers with the state-of-the-art welding solutions. In 2013 Fronius launched Trans Process Solution Intelligent (TPS/i) system where the system interfaces are incorporated into system bus architecture and is able to provide data in a real time. This data can be used to monitor, analyze, and document the process (WeldCube). In addition to this, the company also has developed a system of facilitating welder training (Virtual Welducation) and a WeldConnect software program. Byrd et al. (2015) have employed virtual reality simulations VRTEX® 360 as an assessment tool for experienced and trained novice welders. Sirius Electric offers ultrasonic welding machine that has remote software administration and control facility. Similarly, ABICOR BINZEL offers a system that manages the welding gases (W˛eglowski 2018). In recent years, SMW Group, Queensland, have automated their welding repair processes with advanced laser seam tracking, adaptive welding software, a new generation welding system, and a modular configuration robot. They are able to reduce welding time by 70–90%, reduction in overall production costs, improvement in safety, quality, and reporting. The first novel method proposed by Backer et al. (De Backer et al. 2014; De Backer and Bolmsjo 2013; De Backer 2014) for friction stir welding (FSW), process to monitor, and control system architecture where the temperature controller modifies the robotics FSW spindle speed for maintaining a constant weld temperature.

1644

V. S. Gadakh and V. J. Badheka

Fig. 5 FSW smart industry (Mishra et al. 2018)

Later Mishra et al. (2018) reviewed different techniques and methods for sensorbased quality and control monitoring of defects in FSW. They proposed a roadmap to achieve the goal of Industry 4.0 in FSW as can be seen in Fig. 5. They also proposed an FSW monitoring and control architecture similar to Backer et al. (De Backer et al. 2014; De Backer and Bolmsjo 2013; De Backer 2014) with addition of different sensor modules which integrates the existing experimental data that acts as a knowledge base and the different sensory systems which captures, extracts, processes, and monitors the data, and finally using correlation the final decision is made. On the similar platform, different welding equipment manufacturers provide the equipment that meet the expectations of Industry 4.0 either by adopting existing technologies or retrofitting. Another vital component of Industry 4.0 is digital twins. It is a kind of virtual simulator which mimics the physical object/system/process and it can be employed for analysis and to simulate the process in a real-world environment so as to accommodate the changes, optimize the process parameters, and improve the weld quality, reduce time, and costs. SORPAS® created first resistance spot welding process digital twins for its complete lifecycle from designing, optimizing, planning, producing, and evaluating the welds. With the exploitation of machine learning and artificial intelligence (AI) approaches, the weld digital twins work with physical welding process in dynamic environment (Fig. 6).

64 Sustainability of Fusion and Solid-State Welding Process in the Era. . .

1645

Fig. 6 Digital twins for resistance spot welding (Ref. www.swantec.com)

Industry 4.0 in Fusion Welding Process The concepts like Industry 4.0, the smart factory, and digital transformation provide several greater prospects for welders. Due to digitization of welding equipment, there is enhancement in productivity and efficiency of the welding process. Similarly, superior weld quality, repeatability, better conditions for welders, and sustainable competitiveness are obtained. The computers alone provide right decisions similar to experienced welders. To achieve this, the high-performance information and communication technology along with sensors, powerful network infrastructure for data transfer, communication security, and storage are needed. The literatures on fusion welding process in Industry 4.0 are limited. In most of the reported work, machine learning approach has been used which is often termed as “intelligent welding.” Computational modeling and simulation is the heart of intelligent weld manufacturing. Chen et al. (2000) have described the framework for intelligent weld manufacturing. It requires a multidisciplinary integrated computational welding engineering approach (process, automation, control, microstructure, and property) for intelligent weld manufacturing (David et al. 2018). In gas tungsten arc welding (GTAW), gas metal arc welding (GMAW), laser welding, and resistance welding machine learning approach were employed to optimize weld process parameters, enhance weld quality, defect detection, etc. (David et al. 2018). Similarly, with development of bioinformatics model of machine learning, risk and

1646

V. S. Gadakh and V. J. Badheka

danger associated with welding processes can be minimized (Mahadevan et al. 2021). The example shows that how fusion welding process works well in the arena of Industry 4.0. Cold metal transfer (CMT) welding is advanced joining process; a variant of GMAW which includes filler wire feed drive in a close loop control gives extremely stable arc, low heat input, and spatter free process to join different material combinations. Vendan et al. (2020) have adopted data from Kumar et al. (2016) to optimize the CMT welding process parameters of Aluminum 6061 using supervised machine learning. The initial process parameters were considered as welding current (50 A, 60 A, 70 A) and weld speed (400, 500, and 600 mm/min), respectively, whereas the performance measures were penetration depth, width of weld pool, and height of reinforcement. After their initial study, they used machine learning techniques for data analysis and predicting the optimal process parameters. They employed NumPy and Pandas for data analysis and Seaborn and Matplotlib for data visualization. They employed seaborn library for making the co-relation of independent variables. After co-relation matrix the next essential step is data preprocessing which is used for nonlinearity. Generally, it is applied using linear regression technique for training the higher degree model. After data preprocessing the next step is prediction analysis using Scikit-Learn. They employed a typical method of dividing the data into 80% of trained data and 20% of test data from the dataset for predicting the dependent parameters. After training the model, it is again used for data prediction. Recently, ADLINK Technology Inc. provided built in Intel® distribution of OpenVINO™ toolkit for automated arc-welding porosity defects in real time. By means of neural network (NN), the two things are performed by the system, that is, first it identifies the defects in real time, and second is auto-stopping the welding process using robot actuation prior to the defect is extended and the part is damaged beyond repair. Further, the NN model can also retrained for detecting other type of defects in the welding process. EWM AG Xnet 2.0 provides welding management system like a smart factory with digital transformation. The Xnet 2.0 is a kind of intelligent monitoring system which keeps transparency in the welding processes right from process planning, actual production, and final cost estimation of weld seams. Recently, Novarc Technologies launched the NovEye™, a vision-based software that employs machine learning for keeping high accuracy track at the root pass, also measures the root gap and automatically detects the tracks. Recently, Kemppi India Pvt. Ltd. launched the X8 MIG Welder, an intelligent welding system which provides connectivity to the WeldEye welding management software and innovative performance software. The WeldEye welding management software provides simplicity, traceability, and WPS management in welding production.

Industry 4.0 in Solid-State Welding Process Similar to the fusion welding process, the literatures are scarce on Industry 4.0 in solid-state welding process . The following examples show that how solid-state

64 Sustainability of Fusion and Solid-State Welding Process in the Era. . .

1647

welding process well in the arena of Industry 4.0. The FSW is advanced welding process, also called “green” technology owing to its versatile benefits (Mehta and Badheka 2016). In contrast to the traditional weld techniques, it consumes notably less energy. It involves complex material displacement and plastic deformation. Researchers across the globe now understand the merits of using the solid-state welding method as a potential alternative for difficult-to-weld material. Due to its solid-state nature, it is expected to result in components with superior metallurgical and mechanical properties. Most of the available literature has applied ANN, adaptive neuro-fuzzy inference system (ANFIS), support vector machine (SVM) techniques to optimize FSW and FSSW processes. Fleming et al. (2008) and Gibson et al. (2013) employed dimensional reduction techniques for weld quality, and defect formation and tool wear (Gibson et al. 2018). Hartl et al. (2020) used ANN for predicting weld surface quality. Shojaeefard et al. (2015) used ANN to correlate tool parameters (pin and shoulder diameter) and heat-affected zone, thermal, and strain value in the weld zone. Fleming et al. (2007) employed SVM for fault detection like gap occurrence and find out the gap depth in FSW process. Fratini et al. (2009) integrated ANN to finite element model (FEM) and calculated the average grain size values of butt, lap, and T type FSW joints. Boldsaikhan et al. (2006, 2011) used artificial neural network (ANN) for defect detection. They statistically correlated, controlled, and developed a system for weld quality evaluation using tool X, Y forces, and the tool torque feedback signals. The force signal feature extraction, and generation of feature vectors for each of the force signal was evaluated using Discrete Fourier Transform (DFT). The ANN was applied to these resulting feature vectors which categorize the weld into two. The first category has/not has metallurgical defects and the second category gives the weld strength as per requirement. The obtained results were promising; still some issues were suggested like noise, network robustness, prefiltering of machine dependent component of output signals before categorization. Vendan et al. (2020) have adopted data from Suban et al. (2017) to optimize the FSW process parameters of AA6061/TiB2 /SiC metal matrix composite by varying weight percentages of SiC, B4 C, and TiB2 using supervised machine learning. The initial process parameters were considered as tool rotational speed (600, 800, and 1000 rpm) and weld speed (50, 60, and 70 mm/min), respectively, whereas the performance measures were mechanical properties. A similar procedure was employed to predict the optimal process parameters as explained in case of fusion welding. Based on these, it is able to get the predicted values for the performance measures along with other characteristics as torque developed, experimental heat generated, and theoretical heat input. In their another work (Vendan et al. 2020), they have applied supervised machine learning in Magnetically Impelled Arc Butt Welding (MIAB) process. The process was studied by the E. O. Paton Electric Welding Institute and after its development it was commercially applied by Kuka Welding systems. The process is fully automated solid-state welding and able to weld tube diameters approximately 10–220 mm with wall thickness of 0.7–13 mm. Vendan et al. (2020) have optimized the MIAB process parameters of alloy steel tubes (T11) by varying current and time using

1648

V. S. Gadakh and V. J. Badheka

supervised machine learning. They considered following parameters notch strength ratio, ultimate tensile strength, and mean weld interface hardness as performances. A similar procedure was employed to predict the optimal process parameters as explained in case of fusion welding and FSW process. Based on these, it is able get the predicted values for the performance measures. Jayaraman et al. (2011) have employed ANN approach for predicting tensile strength of high strength Aluminum A356 alloy. Tansel et al. (2010) have applied genetically optimized NN system (GONNS) to evaluate the optimal process parameters of FSW process. Satpathy et al. (2018) have developed the regression, ANN, and ANFIS models for simulating and predicting the joint strength of ultrasonic metal welding of dissimilar sheets of Al-Cu. Dewan et al. (2016) developed ANFIS model to predict the tensile strength of FSW joints. Zhu et al. (2002) have applied SVM technique to predict the welding joint quality. Conventional optimization methods give local optimal solution and are not robust. The meta-heuristics (MH) is a global search technique which deals with all kinds of objective functions and design variables and these are flexible as there is no need of data training (Saka et al. 2016). Significant research is reported in the literature (Rao and Pawar 2010) that the conventional techniques such as Powell’s Method, Fletcher-Reeves Method, Dynamic Programming, and the Reduced Gradient Method, etc. have been employed for such kinds of problems. However, these techniques lack to perform well over a wide range of problems. With these issues as an input, efforts are being made to model and optimize the complex problem using MH techniques. Most frequently employed algorithms in the literature are scatter search (SS), simulated annealing (SA), Tabu search (TS), artificial immune system (AIS), Ant Colony Algorithm (ACO), differential evolution (DE), multiobjective GA (MOGA), particle swarm optimization (PSO), imperialist competitive algorithm (ICA), teaching learning-based optimization (TLBO), and artificial bee colony algorithm (ABC). However, there is a need to explore these and other hybrid algorithms in welding applications for optimization of process parameter and predicting the performance measures. Nadeau et al. (2020) applied different machine learning algorithms for predicting the defective welds over various process variables. Out of these machine learning models, K-nearest neighbor (KNN) technique performs best for predicting defective welds. Automatic identification algorithm was developed by Sikora et al. (2013) for detecting and classifying the weld defects by employing X-ray radiograph images. Chen et al. (2003) studied online monitoring of FSW process and defects using acoustic emission technique. A fractal dimension algorithm (Das et al. 2016), Power spectral density (Das et al. 2018) was employed to extract various acquire signals for weld process monitoring and detect the defects. Thermography was employed for online monitoring to measure the weld joint quality using FSW process by monitoring temperature plots during the welding (Serio et al. 2016; De Filippis et al. 2017). A free vibration technique was employed as NDT to find out weld quality which depends on natural frequencies of the weld samples (Crâ¸stiu et al. 2017).

64 Sustainability of Fusion and Solid-State Welding Process in the Era. . . Table 2 Category scores for fusion and solid-state welding processes (Jamal et al. 2020)

Category Physical performance Environmental impact Economic impact Social impact

GTAW 0.23 0.20 0.34 0.56

1649 GMAW 0.21 0.72 0.58 0.56

FSW 0.39 0.92 0.14 0.96

Even though the parameter optimization and corresponding prediction of the characteristics has been done, still one has to see these welding processes from the sustainability viewpoint. Jamal et al. (2020) have done sustainability assessment for fusion and solid-state welding process using statistical approach by data collection and segregation into environmental impact, economic impact, social impact, and physical performance. They have considered weld yield strength and toughness under physical performance; carbon footprint, material wastage auxiliary material usage, and weld emissions under environmental impact; cost of consumables, labor, energy, welded part, and equipment under economic impact; and finally incident rate under social impact. After that normalization has been made for above indicators by assigning weights for each category. Finally they determined overall sustainability score and its comparison with each of the welding processes. It was concluded based on the statistical approach that FSW is the most sustainable process due to the highest physical performance, social impact, and environmental impact scores in comparison with GTAW and GMAW processes. Table 2 shows the summary of category scores for fusion and solidstate welding processes. It can be observed from Table 2 that FSW process did not score more due to initial tool plunging time, and FSW tooling cost. While GMAW achieved maximum score due to less cost of fillers, its storage, transport, and use. Chang et al. (2015) employed life cycle assessment approach to evaluate social and environmental impacts of fusion welding processes. Life cycle assessment (LCA) and social life cycle assessment (SLCA) are the modern techniques used for the evaluation of social and environmental impacts which affects the product, or process. “LCA is an ISO standard technique employed to assess environmental impacts of products through its entire life cycle.” On the contrary, “SLCA is a technique employed to assess social and socio-economic impacts related to human beings affected by products/services throughout the life cycle” (Chang et al. 2015). Chang et al. (2015) considered manual metal arc welding (MMAW), manual GMAW, automatic GMAW, and automatic laserarc hybrid welding (LAHW) processes for LCA and SLCA. It is revealed that the MMAW process gives maximum environmental impacts in terms of nature, and welders’ health issues when compared with other considered welding processes. Nobrega et al. (2019) have reviewed different manufacturing processes including welding from the sustainable point of view. Garbie (2016) has estimated the sustainability based on the three pillars, namely, economic (E), societal (S), and environmental (N). The sustainability/sustainable development (S/SD) is expressed as:

1650

S/SD Manufacturing Enterprise

V. S. Gadakh and V. J. Badheka

⎧ ⎫ ⎨ S/SD E ⎬ = S/SD S = f (S/SD E , S/SD S , S/SD N ) ⎩ ⎭ S/SD N

S/SD Manufacturing Enterprise = wE (S/SD E ) + wS (S/SD S ) + wN (S/SD N ) Where S=SDE is the value of attribute of economic issues (E) with respect to S/SD assessment S=SDS is the value of attribute of social issues (S) with respect to S/SD assessment S=SDN is the value of attribute of environmental issues (N) with respect to S/SD assessment wE , wS, and wN are the relative weights of the economic, societal, and environmental issues, respectively. These are estimated using the analytic hierarchy process (AHP) and determine the S/SD index of the manufacturing enterprise.

Conclusion The present chapter elucidates the sustainability of fusion and solid-state welding process in the era of Industry 4.0. It covers different approaches employed for assessing and improvement in the weld quality using meta-heuristic algorithms, in process weld quality inspection; defects detection and its control in fusion, and solid-state welding processes. However, sustainability of fusion and solidstate welding process in the era of Industry 4.0 is a missing link that needs to be taken into consideration. Furthermore, welding simulator can be considered as mandatory assessment tool for welder qualification. Past studies have applied statistical approaches and AHP-based assessment for sustainability of manufacturing enterprise. Such similar kind of assessment for sustainability of fusion and solidstate welding process gives remarkable insights into the fourth industrial revolution.

Websites 1. http://www.manufacturingchampions.in/Document/Presentation%201%20%20Overview%20of%20the%20Welding%20Industry%20in%20India.pdf. (access date 15.01.2021) 2. https://www.wfmj.com/story/42569322/welding-equipment-market-size-2020industry-recent-developments-progression-status-latest-technology-and-forecast-research-report-2026 (access date 16.01.2021) 3. https://www.canadianmetalworking.com/canadianfabricatingandwelding/article/welding/next-generation-welding-tech (access date 16.01.2021)

64 Sustainability of Fusion and Solid-State Welding Process in the Era. . .

1651

4. https://industrialautomationreview.com/automation-robotics-welding/ (access date 16.01.2021) 5. https://www.millerwelds.com/resources/article-library/the-latest-weldingtechnologies-can-save-time-and-money (access date 16.01.2021) 6. https://www.manmonthly.com.au/features/welding-4-0-future-industry/ (access date 11.08.2021) 7. https://www.ciiblog.in/technology/weld-4-0-industry-4-0-in-welding/ (access date 11.08.2021) 8. https://www.swantec.com/welding-solutions/digital-twins-and-ai/(access date 12.08.2021) 9. https://weldfabtechtimes.com/news/adlink-enables-automated-arc-welding-defect-detection-with-industrial-machine-vision-edge-solution-toward-industry4-0/ (access date 28.08.2021) 10. https://weldfabtechtimes.com/news/the-leading-manufacturer-of-arc-weldingtechnology-expands-ewm-ag-opens-a-location-in-france/ (access date 28.08.2021) 11. https://weldfabtechtimes.com/products-update/novarc-increases-autonomy-ofits-spool-welding-robot-with-breakthrough-noveye-software-product/ (access date 28.08.2021)

References Agustiady T, Badiru AB (2013) Sustainability—Utilizing lean six sigma techniques. CRC Press, Boca Raton Antonini JM, Murthy GGK, Rogers RA et al (1996) Pneumotoxicity and pulmonary clearance of different welding fumes after intratracheal instillation in the rat. Toxicol Appl Pharmacol 140:188–199 Bi Z (2011) Revisiting system paradigms from the viewpoint of manufacturing sustainability. Sustainability 3:1323–1340 Boldsaikhan E, Corwin E, Logar A, Arbegast W (2006) Neural network evaluation of weld quality using FSW feedback data. In: 6th international symposium on friction stir welding. Nr Montreal, pp 1–14 Boldsaikhan E, Corwin EM, Logar AM, Arbegast WJ (2011) The use of neural network and discrete Fourier transform for real-time evaluation of friction stir welding. Appl Soft Comput J 11:4839–4846 Byrd AP, Stone RT, Anderson RG, Woltjer K (2015) The use of virtual welding simulators to evaluate experienced welders. Weld J 94:389s–395s Chang YJ, Sproesser G, Neugebauer S et al (2015) Environmental and social life cycle assessment of welding technologies. Procedia CIRP 26:293–298 Chen SB, Lou YJ, Wu L, Zhao DB (2000) Intelligent methodology for sensing, modeling and control of pulsed GTAW: Part 1 - bead-on-plate welding. Weld J (Miami, Fla) 79:151-s Chen C, Kovacevic R, Jandgric D (2003) Wavelet transform analysis of acoustic emission in monitoring friction stir welding of 6061 aluminum. Int J Mach Tools Manuf 43:1383–1390 Crâ¸stiu I, Nyaguly E, Deac S et al (2017) Defects detection on the welded reinforcing steel with self-shielded wires by vibration tests. MATEC Web Conf 126:1007 Das B, Pal S, Bag S (2016) Defect detection in friction stir welding process using signal information and fractal theory. Procedia Eng 144:172–178

1652

V. S. Gadakh and V. J. Badheka

Das B, Pal S, Bag S (2018) Weld defect identification in friction stir welding using power spectral density. IOP Conf Ser Mater Sci Eng 346:12049 David SA, Chen J, Gibson BT, Feng Z (2018) Intelligent weld manufacturing: role of integrated computational welding engineering. In: Chen S, Zhang Y, Feng Z (eds) Transactions on intelligent welding manufacturing. Springer Singapore, Singapore, pp 3–30 De Backer J (2014) Feedback control of robotic friction stir welding. PhD dissertation, University West, Trollhättan De Backer J, Bolmsjo G (2013) Thermoelectric method for temperature measurement in friction stir welding. Sci Technol Weld Join 18:558–565 De Backer J, Bolmsjö G, Christiansson AK (2014) Feedback Control of Robotic Friction Stir Welding. Temperature control of robotic friction stir welding using the thermoelectric effect. PhD dissertation, University West, Trollhättan. Int J Adv Manuf Technol 70:375–383 De Filippis LAC, Serio LM, Palumbo D et al (2017) Optimization and characterization of the friction stir welded sheets of AA 5754-H111: monitoring of the quality of joints with thermographic techniques. Materials (Basel) 10:1165 Debroy T (2015) Welding in the digital age. Weld J 94:58–63 Dewan MW, Huggett DJ, Warren Liao T et al (2016) Prediction of tensile strength of friction stir weld joints with adaptive neuro-fuzzy inference system (ANFIS) and neural network. Mater Des 92:288–299 Fleming PA, Fleming KA, Lammlein D, et al (2007) Automatic fault detection in Friction Stir Welding. In: Materials science and technology conference and exhibition, MS and T’07 “Exploring structure, processing, and applications across Multiple materials systems.” pp 3309– 3315 Fleming P, Lammlein D, Wilkes D et al (2008) In-process gap detection in friction stir welding. Sens Rev 28:62–67 Fratini L, Buffa G, Palmeri D (2009) Using a neural network for predicting the average grain size in friction stir welding processes. Comput Struct 87:1166–1174 Garbie IH (2013) DFSME: design for sustainable manufacturing enterprises (an economic viewpoint). Int J Prod Res 51:479–503 Garbie I (2016) Sustainability in manufacturing enterprises: concepts, analyses and assessments for industry 4.0. Springer International Publishing, Cham Garetti M, Taisch M (2012) Sustainable manufacturing: trends and research challenges. Prod Plan Control 23:83–104 Gibson BT, Wilkes DM, Cook GE, Strauss AM (2013) In-process detection of faying surface sealant application flaws in friction stir welding. J Aircr 50:567–575 Gibson BT, Tang W, Peterson AG et al (2018) Evaluating the potential for remote in-process monitoring of tool wear in friction stir welding of stainless steel. J Manuf Sci Eng Trans ASME 140:1087–1357 Gunal MM (2019) Simulation and the fourth industrial revolution. In: Gunal MM (ed) Simulation for industry-past, present, and future. Springer Nature, Cham, pp 1–18 Hartl R, Praehofer B, Zaeh MF (2020) Prediction of the surface quality of friction stir welds by the analysis of process data using artificial neural networks. Proc Inst Mech Eng Part L J Mater Des Appl 234:732–751 Jamal J, Darras B, Kishawy H (2020) A study on sustainability assessment of welding processes. Proc Inst Mech Eng Part B J Eng Manuf 234:501–512 Jayaraman M, Sivasubramanian R, Balasubramanian V, Lakshminarayanan AK (2011) Prediction of tensile strength of friction stir welded A356 cast Aluminium alloy using response surface methodology and artificial neural network. J Manuf Sci Prod 9:45–60 Kamble SS, Gunasekaran A, Gawankar SA (2018) Sustainable industry 4.0 framework: a systematic literature review identifying the current trends and future perspectives. Process Saf Environ Prot 117:408–425

64 Sustainability of Fusion and Solid-State Welding Process in the Era. . .

1653

Kohnen JB (2004) Simplified TRIZ: new problem-solving applications for engineers & manufacturing professionals. CRC Press, Boca Raton, Florida Kumar NP, Vendan SA, Shanmugam NS (2016) Investigations on the parametric effects of cold metal transfer process on the microstructuralaspects in AA6061. J Alloy Comp 658: 255–264 Mahadevan RR, Jagan A, Pavithran L et al (2021) Intelligent welding by using machine learning techniques. Mater Today Proc. https://doi.org/10.1016/j.matpr.2020.12.1149 Mehta KP, Badheka VJ (2016) A review on dissimilar friction stir welding of copper to aluminum: process, properties, and variants. Mater Manuf Process 31:233–254 Miller C, French R, Crawford MH, Saenger F (2002) Welding-related expenditures, Investments and productivity measurement in U.S. manufacturing, construction, and mining industries. Technical Report, American Welding Society, Miami Mishra D, Roy RB, Dutta S et al (2018) A review on sensor based monitoring and control of friction stir welding process and a roadmap to industry 4.0. J Manuf Process 36:373–397 Nadeau F, Thériault B, Gagné MO (2020) Machine learning models applied to friction stir welding defect index using multiple joint configurations and alloys. Proc Inst Mech Eng Part L J Mater Des Appl 234:752–765 Nobrega JHC, Pio PGC, Fernandes GL et al (2019) Sustainability in manufacturing processes: practices performed in metal forming, casting, heat treatment, welding and electrostatic painting. Int J Sustain Dev World Ecol 26:684–697 Petrillo A, De FF, Cioffi R, Zomparelli F (2018) Fourth industrial revolution : current practices, challenges, and opportunities. In: Petrillo A, De FF, Cioffi R (eds) Digital transformation in smart manufacturing. IntechOpen, London, pp 1–20 Posch G, Bruckner J, Ennsbrunner H (2018) Industry 4.0 in welding. Biul Inst Spaw 2018:97–105 Prabhu J, Radjou N, Ahuja S (2012) Jugaad Innovation. Penguin Random House, Haryana Rao RV (2007) Introduction to decision making in the manufacturing environment. In: Decision making in the manufacturing environment : using graph theory and fuzzy multiple attribute decision making methods. Springer-Verlag, London, pp 3–5 Rao RV, Pawar PJ (2010) Parameter optimization of a multi-pass milling process using nontraditional optimization algorithms. Appl Soft Comput 10:445–456 Saka MP, Hasançebi O, Geem ZW (2016) Metaheuristics in structural optimization and discussions on harmony search algorithm. Swarm Evol Comput 28:88–97 Satpathy MP, Mishra SB, Sahoo SK (2018) Ultrasonic spot welding of aluminum-copper dissimilar metals: a study on joint strength by experimentation and machine learning techniques. J Manuf Process 33:96–110 Serio LM, Palumbo D, De Filippis LAC et al (2016) Effect of friction stir process parameters on the mechanical and thermal behavior of 5754-H111 aluminum plates. Materials (Basel) 9:122 Shojaeefard MH, Khalkhali A, Akbari M, Asadi P (2015) Investigation of friction stir welding tool parameters using FEM and neural network. Proc Inst Mech Eng Part L J Mater Des Appl 229:209–217 Sikora R, Baniukiewicz P, Chady T et al (2013) Detection and classification of weld defects in industrial radiography with use of advanced AI methods. In: Far East Forum on nondestructive evaluation/testing. New Technology and Application, Jinan Singh R, Gupta V, Mondal A (2012) Jugaad-from “making do” and “quick fix” to an innovative, sustainable and low-cost survival strategy at the bottom of the pyramid. Int J Rural Manag 8:87–105 Suban AA, Perumal M, Ayyanar A, Subbiah AV (2017) Microstructuralanalysis of B4C and SiCreinforced Al alloy metal matrix composite joints. IntJ Adv ManufTechnol 93:515–525 Tansel IN, Demetgul M, Okuyucu H, Yapici A (2010) Optimizations of friction stir welding of aluminum alloy by using genetically optimized neural network. Int J Adv Manuf Technol 48: 95–101

1654

V. S. Gadakh and V. J. Badheka

Torres-Guerrero F, Neira-Tovar L, Torres-Treviño L (2019) An introductive training for welding workshop: a biometric evaluation using virtual reality scenes to improve practice. In: Trojanowska J, Ciszak O, Machado JM, Pavlenko I (eds) Advances in manufacturing II volume 1 – solutions for industry 4.0. Springer Nature, Cham, pp 319–331 Vendan SA, Kamal R, Karan A et al (2020) Welding and cutting case studies with supervised machine learning. Springer Nature, Singapore Wahl DC (2004) Design and nature II: comparing design in nature with science and engineering. In: Brebbia CA (ed) Design and nature. WIT Press, Southampton, pp 289–298 W˛eglowski MS (2018) Industry 4.0 vs welding engineering. Biul Inst Spaw:23–33 Zhu L, Cao C, Wu W, Xu X (2002) A novel approach based on support vector machine to forecasting the quality of friction welding. In: Proceedings of the World Congress on Intelligent Control and Automation (WCICA), pp 335–339

Smart Farming: Applications of IoT in Agriculture

65

Anil Kumar Singh

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evolution of Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SMART Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Internet of Things (IoT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Major Equipment and Technologies Enabling IoT-Based Smart Farming . . . . . . . . . . . . . . . Wireless Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IoT-Based Tractors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Harvesting Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Communication in Smart Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smartphones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cloud Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Unmanned Aerial Vehicles in Smart Farming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of IoT in Smart Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Irrigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fertilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Crop Disease and Pest Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yield Monitoring, Forecasting, and Harvesting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Protected Cultivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Livestock Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Some Important Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1656 1659 1660 1663 1665 1665 1668 1668 1669 1671 1672 1673 1674 1674 1675 1676 1677 1678 1680 1680 1682 1683 1687

A. K. Singh () Department of Life Sciences, Sant Baba Bhag Singh University, Jalandhar, Punjab, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_114

1655

1656

A. K. Singh

Abstract

The fourth industrial revolution also known as Industry 4.0 has significantly transformed traditional manufacturing and industrial practices by use of contemporary smart technologies. In the same line, agriculture sector has also evolved to become data-centered, precise, and smarter to face the future challenges. The employment of recent advance technologies like Internet of Things (IoT), Cloud Computing (CC), Remote Sensing (RS), Machine Learning (ML), Artificial intelligence (AI), Unmanned Aerial Vehicles (UAVs), and Big Data Analytics (BDA) has modernized several traditional agricultural practices. The modern agriculturist aims to increase agriculture efficiency by enhancing food production with reduced cost and minimum environmental impact. The IoT has potential to offer diverse means of modernizing agriculture sector. Recent advances in IoT-based technologies have empowered agriculture on the path of precise and smart farming. Research institutions as well as the industries are racing to develop farmer friendly and efficient IoT-based agriculture technologies. Efforts by scientific groups, research institutions, and industries are likely to make IoT an imperative technology for agriculture in near future. The present chapter discusses basics of smart agriculture and IoT as well as evolution of agriculture from traditional Agriculture 1.0 to modern Agriculture 4.0. The chapter also includes major equipment and technologies enabling IoT-based smart farming and application of IoT in smart agriculture. Current trends and future challenges of IoT in agriculture have also been highlighted in the chapter. Keywords

Agriculture robotics · Sensors · Cloud computing · Agriculture 4.0

Introduction Agriculture is a broad term referring to occupations concerned with the cultivation of soil for producing crop and raising livestock to provide food, wool, and many other products. Human civilization has been significantly influenced by the advancements in agricultural practices. Shift from foraging to farming during Neolithic era improved nutrition, increased life-span, and decreased workload of human. History suggests that most powerful civilizations of world has well-developed and established agriculture sector. Agriculture supports human society not only by providing food and raw material for industries but also by providing employment. In current scenario, agriculture embraces more prominence than earlier times due to ever-increasing population and dwindling natural resources. According to Food and Agricultural Organization (FAO) the global population will reach nearly 9.6 billion people by 2050 (FAO 2009). Feeding such a huge population will require an increase of about 70% in agriculture output worldwide (Zhang et al. 2018). Last few decades have also seen decrease in total agricultural land. In 1991 total

65 Smart Farming: Applications of IoT in Agriculture

1657

cultivated agricultural land was around 19.5 million square miles which accounted for about 39.47% of the earth surface. The cultivated land decreased to 18.6 million square miles (nearly 37.73% of the earth surface) by 2013 (World Bank 2016). Thus, agriculture sector is facing a formidable task of increasing production with less available agricultural land. Additionally, future of agriculture sector is also challenged by climate change. The change in climate is believed to adversely affect crop production and health of livestock (Arora 2019). Change in global temperature, water availability pattern, soil erosion, and frequent outbreak of disease and pest is expected to decrease overall agriculture output in coming decades. Such an extraordinary increase in the production under limited land resource and challenges from climate changes can only be achieved by modernization and intensification of agricultural practices. Intensification of agricultural practices with traditional approaches has significantly increased the environmental footprint of agriculture. Extensive use of water, fertilizers, weedicides, herbicides, pesticides, and changes in land use practice has raised several environment-related concerns (Lampridi et al. 2019). Current trend indicates that agriculture sector can achieve formidable target of feeding rapidly increasing global population in sustainable manner by embarrassing Industry 4.0 technologies, application, and solution. Industry 4.0 or fourth industrial revolution helped in achieving higher level of operational efficiency and productivity by use of contemporary smart technologies. Farms that choose to be technologydriven in some way or other exhibited several advantages, such as saving money and labor, having an increased production or a reduction of costs with minimal effort and producing quality food with more environment friendly practices (Zhang et al. 2018). Díez (2017) advocated that larger use of smart farming services is vital not only for improving a financial condition of farmers but also to meet the food needs of an expanding population. Smart farming encompasses modern technologies like computer mapping, guidance, and variable-rate equipment during cultivation, harvesting, and post-harvest processing. Smart agriculture technologies apply input whenever and wherever required (Saiz-Rubio and Rovira-Más 2020). This enables food production in a sustainable manner. Hence, smart agriculture is seen as promising strategy to accomplish future demand on sustainable pathway. Smart technologies have enabled precision agriculture feasible for progressive farmers. Contrarily, traditional agriculture practice employs inputs throughout the field in predetermined manner, thus resulting in wastage of resources and extra expenditure. The United State Department of Agriculture (USDA) reported that corn farm following precision agriculture incurred 163 dollars per hectare higher profit than for non-adopters (Schimmelpfennig 2016). The success of precision agriculture relies on systems that generate data in farms. The information generated on farms is then transferred and processed in such a way to make proper strategically and profitable operational decisions (Tzounis et al. 2017). Internet of Things (IoT)-based technologies have aided agriculture to generate, transfer, and process such a big amount of valuable information (Zhang et al. 2018). The modern agriculture sector is expected to be highly influenced

1658

A. K. Singh

by the advances in IoT technologies (Saiz-Rubio and Rovira-Más 2020). The new philosophy centered on agricultural data has been referred with several names like Agriculture 4.0, Digital Farming, or Smart Farming. Smart Farming was born when telematics and data management were combined to the already known concept of Precision Agriculture, thus improving the accuracy of operations (Saiz-Rubio and Rovira-Más 2020). Agriculture 4.0 is based on precision agriculture principles with producers using systems that generate data in their farms, which will be processed in such a way to make proper strategically and operational decisions. Internet of Things (IoT) in an agricultural context refers to the use of sensors and other devices to turn every element and action involved in farming into data (Tzounis et al. 2017). According to estimation nearly 10–15% of US farmers are using IoT solutions on the farm across 1200 million hectares and 250,000 farms (Saiz-Rubio and Rovira-Más 2020). According to a research report on smart agriculture market, it is expected to grow from USD 13.8 billion in 2020 to USD 22.0 billion by 2025 at compound annual growth rate of 9.8% (Smart Agriculture Market 2020). As seen in Fig. 1, there is a continuous increase in the number of publications with the term “Smart Agriculture” and “IoT” along with the term “Agriculture” in the scientific literature. These highlight the importance and growing interest of researchers in IoT-based technologies for empowering smart agriculture. In present scenario, the demand of IoT-based agriculture technologies has increased significantly, particularly in developed countries. Research institutes and

Fig. 1 Increase in the number of publications related to “Smart Agriculture,” “IoT,” and “Agriculture” as seen in Science Direct (www.sciencedirect.com)

65 Smart Farming: Applications of IoT in Agriculture

1659

industries are making efforts to satisfy the future needs of farmers through the development of IoT-based agriculture technologies. The main purpose of this chapter is to demonstrate how IoT-based technologies has modernized agriculture sector. The present chapter introduces readers to the concept of IoT applications in agriculture as well as discuses current trends and future challenges of IoT-based agriculture technologies.

Evolution of Agriculture Agriculture has evolved through ages with aim of increasing the agricultural yield and reduces human labor. Beginning of agriculture can be dated back to about 22,000 years ago when man learned to collect wild fruits as food. Archaeological evidences suggest that various crops were cultivated as earlier as 9500 BC (Hillman 1996; Walsh 2009). The demand for more foods to feed ever-increasing population has propelled agriculture evolution from traditional to more sophisticated form. During course of evolution agriculture has moved on from Agriculture 1.0 to Agriculture 4.0 (Fig. 2). Traditional agriculture also referred as Agriculture 1.0 was entirely dependent on manpower and animal forces. Agriculture 1.0 was labor intensive with low output and involved usage of simple tools (Zhai et al. 2020). Invention and improvement of steam engine during nineteenth century not only brought about progressive influence on various industries but also become force for second agriculture revolution. Agriculture 2.0 involved usage of agriculture machineries and synthetics agrochemicals. Agriculture output in terms of efficiency and productivity increased tremendously in Agriculture 2.0 as compared to Agriculture 1.0. However, improved agriculture output was also accompanied by several harmful effects such as environment pollution, destruction of ecological environment, wastage of natural resources, and excessive consumption of powers (Zhai et al. 2020). The twentieth century saw the rapid development in computer technology and robotics resulting in third agriculture revolution. Agricultural machines with robotics and computer programs became more efficient and intelligent. In Agriculture 3.0, precision and intelligence of agricultural machines reduced use of agrochemicals and wastage of natural resources leading to more productive agriculture (Zhai et al. 2020). Previous few years have seen integration of Information and Communication Technology (ICT) with traditional agriculture. This integration has resulted in fourth agricultural revolution designated as Agriculture 4.0. Implementation of recent technologies like Internet of Things (IoT), Cloud Computing (CC), Remote Sensing (RS), Artificial Intelligence (AI), Machine Learning (ML), and Big Data Analytics (BDA) has brought about advances in agriculture through development of Smart Farming practices (Zhai et al. 2020). Agriculture 4.0 is based on principles of monitoring, recording, collecting, and processing on field data for making suitable and strategically operational decisions. Traditional farming required farmers to visit the fields, check the condition of their crops, and make decisions based

1660

A. K. Singh

Fig. 2 Evolution of agriculture from Agriculture 1.0 to Agriculture 4.0

on their experience or seek suggestion from experts. It is estimated that on an average farmers spent 70% of farming time in monitoring and understanding the crop status instead of doing actual field work (Saiz-Rubio and Rovira-Más 2020). Traditional approach of manually monitoring and managing large farm is not feasible, particularly under labor force limitations. Future agriculture must develop and rely on technologies that can impart efficiency and sustainability under limited workforce. IoT empowered smart farming technologies have potential to provide practical solution to future needs.

SMART Agriculture Smart farming or Smart agriculture is promising concept of contemporary agriculture sector that aim to minimize waste and enhance productivity by use of advance supplementary technologies (Saiz-Rubio and Rovira-Más 2020). Present agriculturist and policy makers consider smart farming as a green technology approach as it reduces ecological footprint in comparison to traditional farming practices (Navarro et al. 2020). Implementation of advance information and communication

65 Smart Farming: Applications of IoT in Agriculture

1661

Sensing Technologies

Data analytics Solutions

Software Applications

Smart Farming Hardware & Software systems

Communication System

Telematics, positioning Technologies

Fig. 3 Important components of smart farming

technologies has improved the quantity and quality of agriculture products while optimizing the human labor requirements and using fewer natural resources. Smart farming monitors and analyzes environmental conditions (such as growth status, soil status, irrigation water, pest, fertilizers, and weed management) for optimizing various input processes (Doshi et al. 2019). Smart farming uses hardware and software to record the data and give valuable insights to handle all on-farm activities (Fig. 3). IoT are important hardware for proper functioning of Smart farming. IoT-based smart farming enables real-time monitoring, recording, diagnosis, decision making, and accordingly appropriate activities in farms (Fig. 4). Generally, IoT-based system consist of input and output interface for sensors, interface for connecting to the Internet, interface for memory and storage, as well as interface for audio or video. Demand for technology-driven smart farming has increased in recent past particularly after wide acceptance and recognition by progressive agriculturists. There are several commercially available IoT solutions for smart agriculture (Table 1).

1662

A. K. Singh

Fig. 4 Important events in IoT-based smart farming cycle

Table 1 Some renowned IoT solutions used in agriculture IoT platform Easyfarm Farmx

Services Farm record-keeping Impact of management on health of canopy Cropx Optimize input application, leading to significant saving of resources like water, fertilizer, energy, and labor Farmlogs Farm management system software to automatically record and analyze day-to-day operation MbeguChoice Suggest type of crop seed to farmers of Kenya KAA Connect and manage IoT devices via an open cloud system Determine water demand, growth, and Phytech health of tree Semios Assess and optimize response to insect, disease, and plant health in real time OnFarm Automatically collect, analyze, and store data to provide real-time field situation in user-friendly dashboard Farmtrx Precision yield monitor system that can be attached with any combine

websites http://www.easyfarm.com/ https://www.farmx.co/ https://www.cropx.com/

https://farmlogs.com/

http://www.mbeguchoice.com/ https://www.kaaproject.org/ https://www.phytech.com/ https://semios.com/ https://www.onfarm.com/

https://www.farmtrx.com/

65 Smart Farming: Applications of IoT in Agriculture

1663

Internet of Things (IoT) The Internet of Things (IoT) is considered as one of the big technological revolution of the present world. IoT-based technologies are expected to touch the life of entire human race directly or indirectly (Doshi et al. 2019). The IoT allows things to be controlled from distance via an existing network infrastructure, thus creating possibility for direct amalgamation between the physical world and computer-based systems. Utilization of IoT-based technologies in several sectors (like manufacturing industry, healthcare, transportation, communications, homes, and agriculture) has reduced inefficiencies and improved the performance. The phrase “Internet of Things” (IoT) was first coined by British innovator Kevin Ashtonin in 1999 to describe the network of physical devises that are connected to Internet for exchanging data (Ayaz et al. 2019; Tzounis et al. 2017). Since its conception, IoT has improved a lot by embracing various empowering techniques such as wireless sensor networks, cloud computing, big data embedded systems, security protocols and architectures, communication protocols, and web services. Basically, the IoT is assembly of three layers, namely, the perception layer, the network layer, and the application layer (Fig. 5a). The three-layer architecture defines the basic idea of IoT but it is not always sufficient in achieving desire results, hence several researchers have included finer aspects and added more layers to architecture of IoT (Kumar and Mallick 2018). The five-layer architecture of IoT includes perception layer, transport layer, processing layer, application layer, and business layer (Fig. 5b). The role of perception layer and application layer of fivelayer architecture is same as that of three-layer architecture (Tzounis et al. 2017). The perception layer is physical layer consisting of sensors for sensing and gathering information from environment (Chen and Yang 2019). This layer has technologies such as Radio Frequency Identification (RFID), Wireless Sensor Network (WSN), and Near Field Communications (NFC). RFID technology is considered as the first and the most suitable example of interconnected “Things.” RIFD uses electromagnetic waves to automatically identify and track tags attached to the objects. RFID tags hold data in the form of the Electronic Product Code (EPC). The RFID readers triggers read and manipulate a large number of tags, thus allowing object identification, tracking, and data storage on active or passive tags. Active tags are powered by embedded power supply consisting of battery. Passive tags do not require embedded power supply as it is powered by energy from interrogating radio waves of RIFD readers (Welbourne et al. 2009). A typical WSN is equipped with sensing and computing device, radio transceivers, and power components for monitoring and recoding the physical condition of the environments. Data collected by hundred thousands of sensors are organized at central location. Sensor nodes of WSN communicate among themselves using radio signals. NFC is a set of communication protocols used for communicating between two electronic devices over a very short distance of 4 cm or less. NFC enables a lowspeed connection with simple setup to bootstrap more-capable wireless connections. NFC devices can act as electronic identity documents and keycards (Tzounis et al. 2017).

1664

A. K. Singh

Fig. 5 IoT solution architecture that includes (a) three layers and (b) five layers

The transport layer constitutes the second layer in five-layer IoT architecture. This layer transfers the data collected by sensor from the perception layer to the processing layer and vice versa. This layer uses network protocols to enable communication between the perception layer and the processing layer. Network protocols are employed to construct wireless communication between sensor nodes and application layers. Each protocol has distinctive characters, such as the working range, data exchange rate, and power consumption. Based on distinct features network protocols can be grouped into short range, cellular networks, and long range (Fernández-Ahumada et al. 2019). Short-range network protocols allow communication over short distances and hence are applied for the communication between devices that are located close to each other. Bluetooth, ZigBee, and Wi-Fi are good examples of short-range protocols. These protocols have high data transmission rate and low power requirement. Cellular networks like GPRS and 3G enable long distance communication with high transmission rate. However, licensing cost and high power consumption sometimes creates hurdle for extensive application (Mekki et al. 2019). Long-range network protocols allow communication over very long distance. LoRaWAN and Sigfox are good examples of long-range network protocols (Fernández-Ahumada et al. 2019). These protocols have a low power consumption as well as low data transmission rate. Long-range network protocols are considered suitable when a small amount of data has to be transferred over very long distances (Navarro et al. 2020; Fernández-Ahumada et al. 2019).

65 Smart Farming: Applications of IoT in Agriculture

1665

The network layer constitutes the second layer in three-layer IoT architecture. This layer is concerned with transmitting and processing of sensor data. Data collected by wireless sensor nodes are communicated to neighboring nodes or a gateway and further forwarded to remote infrastructure for storage, analysis, processing, and dissemination through network layer (Gubbi et al. 2013). Communication protocols formulate over wireless standards, such as 802.15.4, help the device networking, and fill the gap between the Internet-enabled gateways and the end-nodes. Such protocols include ZigBee, ONE-NET, Sigfox, Wireless HART, ISA100.11a, and 6LowPan. Recently, Bluetooth Low Energy (BLE), LoRa/LoRaWAN, DASH7, and low-power WiFi have also been reported to perform efficiently in network layer (Suhonen et al. 2012; Tzounis et al. 2017). The processing layer is also known as the middleware layer of five-layer IoT architecture. It stores, analyzes, and processes enormous amount of data coming from the transport layer. This layer can handle and provide a diverse set of services to the lower layers. It utilizes numerous technologies such as databases, cloud computing, and big data processing modules (Tzounis et al. 2017). Modeling information by artificial intelligence (AI) and machine learning for the development of decision support systems and automation of agriculture process has brought about true value of precision farming (Navarro et al. 2020). The application layer is responsible for delivering application-specific services to the user. This layer is regarded as the most important layer as it facilitates the realization of the IoT-based smart farming (Tzounis et al. 2017; Navarro et al. 2020). The business layer handle the whole IoT system, including applications, business and profit models, and users’ privacy.

Major Equipment and Technologies Enabling IoT-Based Smart Farming IoT-based smart farming has been made possible by use of several equipment and technologies. This section discusses some major equipment and technologies that have contributed toward realization of smart farming.

Wireless Sensors Wireless sensors are the most important part of smart farming equipment. Almost every part of advance agricultural tools and heavy machineries contain wireless sensors for gathering on site information. Depending upon the application requirements, there may be following major type of sensors, namely, acoustic sensors, field-programmable gate array (FPGA)-based sensors, optical sensors, ultrasonic ranging sensors, optoelectronic sensors, airflow sensors, electrochemical sensors, electromagnetic sensors, mechanical sensors, mass flow sensors, Eddy covariancebased sensors, soft water level-based (SWLB) sensors, light detection and ranging (LIDAR), telematics sensors, and remote sensing (Ayaz et al. 2019).

1666

A. K. Singh

Acoustic sensors function by quantifying the change in the noise intensity as the tool containing these sensors interact with other materials such as seeds and soil particles. It has been used successfully for monitoring and detection of pest (Ayaz et al. 2019). Gasso-Tortajada et al. (2010) have used novel acoustic sensor for classifying different seeds varieties using sound absorption spectra. Acoustic sensors are preferred especially in portable equipment as it provides low-cost solutions with fast response (Ayaz et al. 2019). Field-programmable gate array (FPGA)-based sensors are being used in smart agriculture due to their flexibility of reconfiguration. It has been employed for measuring real-time plant transpiration, irrigation, and humidity (Millan-Almaraz et al. 2010). FPGA-based sensors utilization for agricultural purposes is in infancy stage primarily due to limitation issues such as size, cost, and power consumption. These sensors are not suitable for continuous monitoring systems as they have high operational power requirement (de la Piedra et al. 2012). FPGA-based sensors have potential to offer satisfactory solutions in smart agriculture once it overcomes aforementioned limitations. Optical sensors use light reflectance phenomena for measuring various soil parameters. These sensors test soil parameters such as organic content, moisture, color, and mineral composition by reflecting light of different electromagnetic spectrum on samples. The changes occurring in reflected wavelength help to quantify various parameters. Molina et al. (2011) have reported use of integrated optical sensors with microwave scattering for characterizing olive grove canopies. Fluorescence-based optical sensors have been used for supervising the fruit maturation (Pajares 2011). Ultrasonic ranging sensors are often considered better than several other sensors. These sensors have low cost, ease to use, and ability to become part of diverse applications. It has been used for monitoring tank, crop canopy, and measuring spray distance (Dvorak et al. 2016). Combination of camera and ultrasonic ranging sensors has been employed for the detecting weeds in the crop field (Pajares et al. 2013). Ultrasonic sensors detect height of plants while the camera determines the weed and crop cover. Optoelectronic sensors are known to distinguish between different plant types. It helps to identify weeds, herbs, and other unwanted plants among crops. Andújar et al. (2009) have used combination of optoelectronic sensor and location information to map the weed distribution in maize crop field. Optoelectronic sensors can also be employed for mapping vegetation area as it can differentiate between plant cover and soil through reflection spectra. Airflow sensors are capable of measuring various soil characteristics by pushing a predetermined quantity of air into the ground at a prescribed depth. It has been used to determine soil air permeability, soil moisture content, and soil types. It can perform measurements at a singular fixed position or over long range in the mobile mode (García-Ramos et al. 2012). Electrochemical sensors are widely used in smart agriculture for soil test automation. Standard chemical analysis of soil is expensive and time-consuming process. However, application of electrochemical sensors can easily substitute these

65 Smart Farming: Applications of IoT in Agriculture

1667

tests by determining various chemical characteristic of soil such as salinity, pH, and macro and micro-nutrients of the soil (Cocovi-Solberg et al. 2014). Electromagnetic sensors record electrical conductivity and transient electromagnetic response to adjust rate of applications in the authentic situation. These sensors determine various properties of soil by measuring the capability of soil particles to conduct or accumulate electrical charges. Yunus and Mukhopadhyay (2011) had demonstrated that electromagnetic sensors can be applied to determine residual nitrates and organic matter in the soil. Mechanical sensors are widely used to determine soil mechanical compaction. These sensors enter through the soil and record the force experienced by strain gauges (Hemmat et al. 2013). Soil compaction knowledge aids farmers to determine the tillage requirement. Mass flow sensors have been used to determine the yield by measuring the amount of grain flowing through combine harvester. Sensing mass flow for determining the crop yield has been used extensively in IoT-based smart farming. Yield monitoring systems apart from mass flow sensors also contain several other modules, like the grain moisture sensor, data storage device, and software to analyze the recorded data (Ayaz et al. 2019). Eddy covariance-based sensors measures exchange of gases, water vapors, and energy between surface of the earth and atmosphere. The sensors based on this technology can quantify exchanges of carbon dioxide, methane, or other gases, and energy between agriculture field and atmosphere (Kumar et al. 2017). These sensors are preferred over other sensors due to its ability to measure continuous flux over large areas with high precision. Soft water level-based (SWLB) sensors are used in agriculture sectors to describe hydrological behaviors of catchment area. These sensors measure rainfall and stream flows by determining water level and flow (Crabit et al. 2011). SWLB has helped agriculture sector by enabling efficient and effective irrigation, thus reducing the wastage of water. Light detection and ranging (LIDAR) is method used to measure distance. It has been used in a range of agriculture applications, such as land mapping and segmentation, determining soil type, monitoring soil erosion and soil loss, and yield forecasting (Ayaz et al. 2019; Montagnoli et al. 2015). These sensors operate by illuminating the target with laser light and then measure the reflection from targets. Difference between laser return times and wavelength is used to make digital threedimensional representation of targeted site (Ayaz et al. 2019). Telematics sensors help in telecommunication between two places. In agriculture-based application telemetry sensors have been employed for communicating between two vehicles. These sensors collect data from remote location and inform farmers on how the components of machine are working. These sensors also record location and travel routes of operating machines. Telematics sensors-based technologies have enabled farmers to record and store all information related to farm operations automatically (Ayaz et al. 2019). Remote sensing-based technologies has enabled researcher and farmers to monitor on field crop, forecast yield dates, model and forecast yield, identify

1668

A. K. Singh

weeds and pests, and map land degradation. These sensors capture and store the geographic information. Information collected by remote sensors is further analyzed, manipulated, managed, and presented as spatial or geographical data (Ayaz et al. 2019). Argos-sensor is a well-known example of remote sensor-based technology. It is a satellite-based sensor system for collecting, processing, and distributing environmental data from fixed and mobile platforms worldwide (Rose and Welsh 2010).

IoT-Based Tractors Tractors are farm vehicles used for imparting power and traction to mechanized agricultural task like tillage. Demand for tractors has increased with decrease in rural labor resource. It has been reported that an average size tractor can work 40 times faster at less expenses than traditional farm labor (Ayaz et al. 2019). Different agriculture equipment manufacturers are continuously adding advance features to tractor for fulfilling the farm requirements. Several manufacturers have added automatic-driven and Cloud-computing capabilities to advance tractors. Manufacturers are working on to commercialize self-driving tractors. The main benefit of self-driving tractors is their ability to make very precise turns without the physical presence of driver. Self-driving tractors offer precision with reduced errors that are mostly unavoidable when a human controls the machinery (Ayaz et al. 2019).

Harvesting Robots Harvesting is the last and most critical phase of crop cultivation. It has significant influence on crop quality and quantity. Some crops are harvested once while some others crops are picked several times after plants have reached a certain stage. Harvesting crop at the correct time in a proper manner is very important for obtaining optimum return from cultivated crops. Early or late harvesting generally reduces quality and quantity of crops. Labor are predominantly required for harvesting crops. However, labor shortage is hampering several agriculture functions significantly. According to Ayaz et al. (2019), there is decline in crop production annually due to labor shortage. United States Department of Agriculture reported that costs of wages and labor may range something between 14% and 39% of total cultivation cost (Ayaz et al. 2019). Considering the importance of this stage and labor issues, farm experts and managers believe that use of agriculture robotics will ease the labor pressure and provide the freedom to harvest as per desire (Ayaz et al. 2019). Recent decades have seen automation of harvesting process by use of sensitive robots that can detect shape, size, color, and localization of fruits (Zujevs et al. 2015). Harvesting robots require sophisticated sensors capable of accumulating accurate information of particular crop and fruit. In real field scenario, the task of

65 Smart Farming: Applications of IoT in Agriculture

1669

detecting the right target is not simple as most of the fruits are partially or fully hidden among leaves and branches (Bac et al. 2017). The successful automation of harvesting process employ computer vision, image processing, and machine learning techniques. As different fruits vary in shapes, sizes, and colors, researchers are designing and developing specific robots for specific crops (Zujevs et al. 2015). SW 6010, Octinion, SWEEPER robot, and FFRobot are some of the leading harvesting robots being used for crop harvesting. FFRobot is harvesting robots employed for harvesting tree-based fruits like apples. It has capability to pick up to 10,000 fruits per hour. SW 6010 and Octinion are able to collect strawberries. SWEEPER robot is employed to harvest peppers. Another automated strawberry harvester Tektu T-100 is an all-electric rechargeable harvester that can run silently with zero emission inside the poly-tunnels (Ayaz et al. 2019; Defterli et al. 2016).

Communication in Smart Agriculture Smart farming is heavily dependent on accurate and timely communication and reporting the information to stakeholders. The real purpose of smart agriculture can only be achieved by firm, reliable, and secure connection among various participating objects. Successful implementation of IoT in agriculture sector on a large scale can be achieved only after providing a suitably large architecture for communication. Deciding upon the means of communication is influenced by several factors like cost, coverage, energy consumption, and reliability. Currently several communication modes and technologies are employed in smart agriculture, depending on the availability, scalability, and application requirements. Cellular communication through 2G, 3G, 4G, or 5G, ZigBee, LoRa, Bluetooth, and Sigfox has been used successfully to communicate in smart agriculture. Cellular communication modes from 2G to 5G have been used for communicating information or data. However, availability of appropriate cellular network in rural region is a major concern worldwide particularly in developing and underdeveloped countries. Data transmission via satellite can overcome aforementioned problem. The cost of communication via satellite is very high thus making it unsuitable for small- and medium-sized farms. The choice of communication mode is also determined by the application requirements. Farms employing sensors that operate with low data rate but work continuously for long periods require long battery life. Such farms often consider relatively a new range of Low Power Wide Area Network (LPWAN). Several research consider LPWAN a better solution for cellular connectivity, due to long battery life and a wider connectivity range at reasonable rates (2 to 15 USD per year) (Beecham Research 2016). Several short-range and medium-level communications are also being used in mesh networks for communicating information (Zulkifli and Noor 2017). In a mesh-network based communication system, sensor nodes collect data and transmit it to the gateway located in the same area. The gateway then sends collected data via WAN network to the farm management system.

1670

A. K. Singh

Fig. 6 Different topologies used for supporting the ZigBee networks: (a) Star, (b) Tree, (c) Mesh

The ZigBee technology has emerged as the new standard for superior wireless technology in agriculture communication. The ZigBee hardware uses physical devices to offer a low cost accomplishment. Depending on the application requirements the devices based on ZigBee protocol uses three logical device types, namely, ZigBee coordinator, ZigBee Router, ZigBee End Devices (Oliveira et al. 2017). The ZigBee network uses three different topologies, namely, Star, Tree, and Mesh (Fig. 6). Star topology is simplest of all topologies. It consists of a ZigBee coordinator and few ZigBee End Devices. All end devices are connected directly to coordinator (Fig. 6a). In the star topology, ZigBee coordinator is accountable for initiating, maintaining, and controlling the entire end devices on the network. Tree topology consists of coordinator, few routers, and end devices (Fig. 6b). Router serves as extension for network coverage. End devices can be connected to coordinator or router. Mesh topology consist of one coordinator and several router and end devices (Fig. 6c). It is also referred as Peer-to-Peer network. This topology has capability to find alternate path to destination, in case node fail to communicate, hence also known as self-healing topology. ZigBee can play vital role especially in the greenhouse environment where usually short-range communications are required. Data collected by sensors enable real-time monitoring of various parameters. Data collected by sensors is transferred through ZigBee network to end server. For the applications like irrigation and

65 Smart Farming: Applications of IoT in Agriculture

1671

fertilization, ZigBee modules are networked for communication. For example, in drip irrigation, soil content like moisture is monitored. Further, SMS is forwarded to the farmer to update about the field data where GSM is required at long distance or Bluetooth module can help at the shorter distances. Bluetooth is a wireless communication standard for connecting small-head devices over shorter distances. It is used in several smart farming applications as it is easy to use, utilize less power, and have low cost. Hong and Hsieh (2016) successfully tested Bluetooth and PLC (programmable logic controller) with ICS (integrated control strategy) for controlling irrigation timing depending upon soil moisture content to carry out smart irrigation. Advanced form of Bluetooth commonly known as Bluetooth Smart or Bluetooth Low Energy (BLE) has been used in sensors devoted to monitor moisture, ambient light, and temperature in open field conditions as well as under protected cultivation area. LoRa wireless technology offers LPWAN connectivity between the wireless sensors and the Cloud as it is a long-range and low-power platform. Sensors based on LoRa can be used in smaller devices for consistent monitoring. LoRa signals have capability to penetrate thick and insulated objects, hence allowing coverage over larger network area. As compared to several network platforms, LoRa-based network have longer lifespan with little maintenance. Jedermann et al. (2018) successfully tested LoRa-based networks platform in apple storage warehouse for monitoring temperature and airflow with packet rate of more than 96%. Sigfox is wireless network connectivity services employed to connect lowpowered objects that emit small amounts of data continuously. This wireless network connectivity service uses narrowband or ultra-narrowband technology. It offers high level performance as several sensors can transmit data at the same time (Lavric et al. 2019).

Smartphones Smartphones are important part of present-day lifestyle. Recent advances in the smartphone industry have resulted in sharp decrease in price thus making it affordable for small farmers. Cellular communication is the key technology in rural areas. Mobile phones are important mode of communication within farming community. It can be easily used to contact or update farmers in the rural setting. With flexibility and functionality (such as the camera, GPS, microphone, accelerometer, and gyroscope), smartphones have tremendous potential to participate in IoT-based smart farming (Table 2). Realizing the tremendous potential of Mobile-phone-based agriculture services, IT professionals are developing mobile apps to fulfill various needs of farmers (Alfian et al. 2017). In recent years researchers have developed several Mobile phone-based agriculture services, particularly in developing countries like Cameroon, China, Turkey, India, Kenya, Ghana, Nigeria, Mali, Uganda, and Zimbabwe to address local farmers issues (Pongnumkul et al. 2015). Table 3 summarizes some important smartphone-based application for farmers.

1672

A. K. Singh

Table 2 Some of the smartphone-based sensors used for various agriculture applications Smartphone Sensor Image sensors (camera)

Function Capture pictures of object

Global positioning system (GPS)

Measure the latitude and longitude

Microphone

Detects usual/unusual sound

Accelerometer

Measures acceleration forces

Gyroscope

Detect the angular velocity of object during rotation

Barometer

Measures air pressure

Inertial sensor

Determine altitude of object in relation to the inertial system

Agriculture uses Detect disease, chlorophyll status, fruit ripeness, leaf area index (LAI), soil erosion Determine location information during machine driving and tracking, land management, crop mapping Help maintaining machine, detect pest, make audio enquiries Detect movement or rotation of camera, detect worker or machine movement Estimate movement of equipment, determine canopy structure Measures the elevation of agriculture farm as an altimeter Help in measuring distance of plant, leaves, or any other object with the help of camera

References Camacho (2018), Chung et al. (2018) Wan et al. (2018), Stiglitz et al. (2017) Kou and Wu (2018) Orlando et al. (2016)

Kou and Wu (2018) Frommberger et al. (2013) Orlando et al. (2016)

Cloud Computing Cloud computing offers on-demand various services (like data storage, server space, database networking, and software) through the Internet. It is a popular choice for people as it offers cost saving, high speed, efficiency, performance, and security. Cloud computing is named so because the information accessed by individual is present remotely in the virtual space. It has supported smart farming applications by two means. Firstly, it offers space for collection and storage of information transmitted from remote sources. Secondly, it allows processing of collected information and displaying the results to the users. Hiring Cloud-based services in smart farming has several opportunities but with few challenges. Smart farming uses vast range of sensors with its own data format and semantics. Also the most of the decision-support systems utilized in smart farming are application-specific. Thus, the Cloud-based decision-support system has to handle the diversity of data and their formats as well as configure these formats for diverse function. AgJunction developed an open Cloud-based system for gathering and disseminating the data from different precise agriculture controllers. Fujitsu

65 Smart Farming: Applications of IoT in Agriculture

1673

Table 3 Some important smartphone applications used for various agricultural purposes Mobile Apps PocketLAI

Application Irrigation

WheatCam

Crop insurance

AMACA

Machinery/tools

Ecofert

Fertilizer management

eFarm

GIS

Weedsmart

Weed management

BioLeaf

Health monitoring

cFertigUAL

Fertigation

Features Determine plant’s water requirement by estimating leaf area index. App uses the mobile camera and accelerometer sensor Offers picture-based insurance (PBI) to simplify the process of crop insurance. Smartphone camera is used to take picture pre and post damaged insured areas Estimate the cost of machinery and its performance in various field operations Calculates the best possible combination of fertilizers and cost of fertilizer based on current market prices Applied for sensing, mapping, and modeling of agriculture farm land This tool is capable of enhancing weed management for a specific paddock. Based on the answers given for nine questions about a paddock‘s farming system, the app assesses herbicide resistance and weed seed bank risk Detects leaf damage and monitor crop foliar status Calculates the amount of fertilizer and water needed for the major crop types in greenhouse farming

References Orlando et al. (2016)

Ceballos et al. (2018)

Sopegno et al. (2016) BuenoDelgado et al. (2016) Yu et al. (2017) Scholz (2018)

Machado et al. (2016) Pérez-Castro et al. (2017)

proposed Akisai-Cloud for increasing the food supply chain in future by focusing on food and agricultural industries.

Unmanned Aerial Vehicles in Smart Farming UAVs are aircraft without human pilot on board. These aircraft are controlled by remote control or autonomous control unit fitted with sensors. UAVs are more commonly known as drones. Massachusetts Institute of Technology considered agricultural UAVs as green-tech tool for smart farming. Agricultural UAVs has been utilized to address several farm-related issues. It provides farmers with luxury of collecting precise data from large areas without much labor. Collected data can be further processed and analyzed to draw valuable information. UAVs fitted with highresolution cameras and precise sensors have been used for monitoring plant growth

1674

A. K. Singh

on an individual plant-level. The application of UAV technology in smart farming has been successfully employed for weed detection and management in crop field of sugar beet, paddy, sunflower, and cotton. UAV technology has been applied to monitor wheat breeding in large trial, monitor key developmental stages of winter wheat, detect plant stress, predict as well as estimate grain yield and pest monitoring. Faiçal et al. (2014) used an algorithm to self-adjust UAV route during spray in crop field to reduce waste of pesticide and fertilizers. UAV-based system has also been used to detect and chase away birds from cultivated fields by producing sounds.

Application of IoT in Smart Agriculture The IoT-assisted agricultural operation mainly has three stages, namely, data collection, cloud-side data analysis, and decision-making. Data are collected with on field sensors or remote sensors or UAVs. Collected raw data are filtered and processed by data analysis tool to draw valuable information. The filtered information is converted to critical information by data mining and machine learning tools. Information analysis with specific factors such as crop state, soil, and environment condition allows decision-making. Appropriate application of IoT in smart agriculture has allowed control of all agricultural operations such as field preparation, seeding, irrigation, fertilization, pest control, and harvesting (Fig. 7).

Irrigation Water is crucial input for good agriculture production and maintaining food security. Irrigation is the process of artificially applying water to the crops for fulfilling water requirements. Farmers generally depend on wells, ponds, lakes, canals, tubewells, and even dam for irrigation. According to an estimate, agriculture sectors use more than 70% of all freshwater globally (FAO 2019). Competition for water resource particularly for agriculture sectors is expected to increase in near future due to population growth, urbanization, and climate change. Future agriculture sector demands judicious and precise use of water in irrigation. The IoT has been more widely used in agriculture for executing precision irrigation. The system uses wireless sensor network (WSN) along with in-field sensors to monitor plant water requirement. Sensors present in the field measures temperature and humidity from air, soil, and canopy. Data recorded by sensors are fed into network gateway. The gateway can be accessed via Internet wirelessly by use of technologies such as 4G LTE mobile communication network at relatively low cost. The data transmitted can be received by subscribed web services on the cloud. The information gathered from farmland and other source (like weather station and satellite imaging) are analyzed to decide on irrigation index value for each site on farmland. These results are transmitted back to network gateway and further forwarded to irrigation controller present on the field. Irrigation controller utilizes results to manage precise irrigation (Zhang et al. 2018). Specifically

65 Smart Farming: Applications of IoT in Agriculture

1675

Fig. 7 Architecture of IoT-based agriculture technology

developed web application can enable farmers to access all data and results, as well as make adjustments in irrigations. Lorvanleuang and Zhao (2018) reported development of IoT-based automated irrigation system that used smartphone to capture and process image of soil near root zone of crop. Rawal (2017) reported development of automated irrigation system capable of monitoring and maintaining the desired soil moisture content through automatic watering system. Control unit was developed with Microcontroller ATMEGA328P on arduino uno platform. Soil sensors present in the field measure and maintain appropriate quantity of water so as to avoid over or under irrigation (Rawal 2017).

Fertilization Fertilizers are natural or synthetic material that can provide essential nutrients to crop plant for proper growth and development. The current agriculture practices depend heavily on fertilizer application for increasing the production of food, feed, fuel, fiber, and other plant products. Continuous cropping exhaust soil from essential nutrients, hence a regular replenishment of nutrients is necessary in order to maintain the fertility of the soil so as to maintain or improve the quality of the harvest (Zhang et al. 2018).

1676

A. K. Singh

Depending upon the requirement nutrients can be grouped into major and minor nutrients. Nitrogen (N), Phosphorus (P), and Potassium (K) are important primary major nutrients. Calcium (Ca), Magnesium (Mg), and Sulfur (S) are secondary major nutrients. Boron (B), Chlorine (Cl), Cobalt (Co), Copper (Cu), Iron (Fe), Manganese (Mn), Molybdenum (Mo), and Zinc (Zn) are minor nutrients. The amount of fertilizer required is influenced by various factors such as crop type, soil type and soil absorption capability, product yield, fertility type, fertilizer utilization rate, weather condition, and climate condition (Ayaz et al. 2019; Zhang et al. 2018). Application of fertilizer in excess amount adversely affects the environment by depleting the soil quality, deteriorating ground water resource and contributing to the climate changes across the globe. Adverse effect of fertilizer can be reduced by precise application of fertilization. Precise fertilization is relatively complicated as compared to precise irrigation. The main hurdle in precise fertilization is determination of fertilizer requirement in soil patches and supplying nutrient accordingly. Remote sensing and IoT approaches have helped in estimating fertilizers requirement at particular site with acceptable accuracy and minimum labor. Normalized Differences Vegetation Index (NDVI) has been used for estimating crops health, vegetation vigor, density, and soil nutrient assessment. NDVI is estimated from satellite images or images from UAV or IoT ground stations. NDVI value is used to generate the site-specific fertilization index map. Fertilizer is applied with automated agriculture machines and vehicles in accordance to the fertilization index map (Zhang et al. 2018).

Crop Disease and Pest Management Yield loss due to disease and pest is one of the serious problems of agriculture sector. According to FAO (2019) nearly 20% to 40% of global crop yield are lost annually due to pest and diseases. Agriculture 3.0 saw application of large amount of agrochemicals like pesticides, fungicides for overcoming losses due to diseases and pest. Globally more than two million tons of agrochemical is used annually. Most of these agrochemicals are harmful to living organisms. Use of IoT-based system allows cultivators with real-time monitoring, modeling, and diseases forecasting. Traditionally, diseases and pest management include application of agrochemicals at particulars stage of crop or pre-prescribed time irrespective of infection. The effectiveness of crop disease and pest management system depends on the sensing, evaluation, and treatment strategies. First stage of IoT-based crop disease and pest management system is to collect real-time crop physiological and pathological data. Image processing is most commonly employed for acquiring physiological and pathological data. Image can be obtained from on field sensors or remote sensing devices installed on aircraft or satellite. Data obtained from remote sensing system has higher efficiency and comparatively less cost as compared to on field sensors. On field sensors enable more function as compared to remote sensing system. On field sensors can be used to monitor and collect data from every corner of field throughout the cultivation period (Zhang

65 Smart Farming: Applications of IoT in Agriculture

1677

et al. 2018). IoT-based system has been employed to develop automated pest traps. IoT-enabled pest traps capture, count, and characterize pest. Information collected by automated traps is then uploaded to cloud for analysis (Ramalingam et al. 2020). Data collected by crop disease and pest management system are sent and stored at central management system. The center processes and analyzes data or image using sophisticated models and algorithms to provide functions like pest identification, diseases identification, behavior of pest, and recommendations from experts. Early warning and suitable interventions can be suggested to the farmer from the central managing system in various ways such as text messages or emails. Information obtained from central system are also utilized in precision farming to precisely apply diseases treatment and pesticide using automated spray or unmanned aerial vehicles or automated variable rate technology. Use of advance robot enabled with multispectral sensing devices, precision–spraying, can locate and deal with the problem. Several IoT technology systems have been successfully used in smart farming for monitoring various aspects of crop production. Zhang et al. (2014) developed IoT system for monitoring diseases, pest, and weeds in field crops of wheat. The system also has capability to diagnose and predict diseases, pest, and weeds of wheat. Depending upon data analysis, the system can provide recommendation to farmers (Zhang et al. 2014). An IoT system developed by Lee et al. (2017) predicted diseases and pest by utilizing model based on correlation information. This IoT system helped farmers to control diseases and pest by judicious usage of pesticides and fungicides.

Yield Monitoring, Forecasting, and Harvesting Yield monitoring is the means of analyzing quality and quantity of harvest. Grain mass flow, moisture content, and harvested grain quantity are some important parameters analyzed during yield monitoring. Accurate yield monitoring helps in estimating crop performance and deciding on the future course of action. Yield monitoring is the important part of smart farming as it not only helps in deciding on the time of harvest but also inform about the quality of yield. Yield quality depends on several internal and external factors. Genetic composition of the crop is an important internal factor influencing yield quality. Nutrient availability, good quality pollination, photoperiod, and sufficient pollination are some external factors influencing yield quality and quantity. With globalization of agriculture sector, the demand for good quality fruits has increased tremendously in recent times. Moreover, good quality fruits must reach right market at right time for harnessing maximum profit. Consumers prefer fruits with attractive external features like fruit size, color, aroma, and texture. All these attractive features of fruit can be achieved only if fruits are harvested at right time and maintained under appropriate postharvest care. Crop forecasting is an art to predicting the yield and production before the harvesting period. Forecasting is done after taking into account the current status of crop plant. Forecasting of crop yield helps farmers to decide on near future

1678

A. K. Singh

plan. Predicting yield quality and maturity time is critical in decision making and deciding on the harvesting time. Yield quality and maturity time prediction take into account the data collected during various development stages of fruit. Fruit color, size, texture, and appearance are some important parameters that are considered for predicting right harvesting time. It must be noted that right harvesting time helps to maximize the crop quality and quantity. It also offers an opportunity to accordingly adjust the post-harvest management strategy. Harvesting is the last but most crucial stage of agricultural process. Proper harvesting at right time can make a great difference. Farmers must know the right time for crop harvesting as it is essential for drawing real benefits from crops. Farm area network (FAN) helps in real-time monitoring of the whole farm and deciding on the harvesting time. A yield monitoring system can be installed on harvester combine and linked with the mobile app FarmRTX. Mobile app displays live harvest data and uploads it automatically to web-based platform of the manufacturer. Mobile app FarmRTX can also generate high-quality yield maps. These maps can be shared with any agriculture experts. Farmer also has option to export information and analyze them on other farm management software. Manfrini et al. (2015) estimated the production and quality of yield by monitoring fruit growth on apple trees. The group considered the fruit growth as the most fundamental and appropriate parameter for determining the progression of the crop. Torbick et al. (2017) used satellite imagining for monitoring rice crop yield in vast areas. Satellite imagining allows monitoring of large area and thus aids in estimating yield of large farms. Such method can also be useful in estimating yield of entire state or country. Sentinel-1A Interferometric images were used to map the rice crop yield and intensity in Myanmar (Torbick et al. 2017). Color (RGB) depth imagining has been used to track the different fruit conditions in mango and papaya farms (Wang et al. 2017). Multiple optical sensors have been employed successfully to monitor the shrinking of papayas particularly during drying conditions.

Protected Cultivation Protected cultivation is a process of growing crops in controlled environments. Green house farming is a good example of protected cultivation. Green house cultivation is more intense than open farm system, hence requires relatively more monitoring and controlled farming operations. There are cloud-based IoT solutions used for monitoring and controlling various operations in green house. High-precision monitoring of green house has been made possible by IoT-based technologies. Sensors present in the green house collect data. These data are uploaded to a cloud infrastructure by Internet facility. Collected data are analyzed with well-evaluated equations and crop and climate models to provide grower with valuable information. Depending upon the information grower may take better decisions or get early warnings.

65 Smart Farming: Applications of IoT in Agriculture

1679

Online precise irrigation scheduling for greenhouses (OpIRIS) was developed on the basis of well-evaluated scientific knowledge. OpIRIS uses web application for communicating with long distance sensors installed in the greenhouses. The system has climate sensors and machinery for regulating fertigation. Sensors robotically sense drainage sample and transfer data to the infrastructure. The data is processed and employed to precisely predict the crop water requirement and notify growers about when to irrigate and volume of nutrient solution to apply. Tong-ke (2013) developed automate irrigation system for green house with agricultural information cloud and a hardware combination of IOT and RFID. The system was able to achieve high efficiency in resource utilization and significant improvement in maintaining water quality. The increasing demand for food requires more farmable lands but soil erosion and pollution has destroyed nearly one-third of farm land. Current intense agricultural practices are also damaging the soil quality at faster rate than it can be replenished. Researchers believe that Vertical Farming (VF) and plant factories are better options to overcome limitation arising due to reduction of arable land and decrease in water resource issues. Plant factories are closed plant growing system facilitated with artificially controlled light, temperature, moisture, and carbon dioxide concentrations. These enable constant production from plant throughout the year, even under unfavorable external environment condition. Artificial growth systems heavily rely on IoT-based technology for imparting suitable condition for plant growth. Vertical Farming is a form of urban agriculture that offers prospect to stack the plants in vertical racks. Plants are allowed to grow in limited space under controlled environment. Vertical Farming is highly efficient in terms of resources consumption due to smart farming approaches. This method enhances production several folds by increasing the number of stacks under fraction of ground surface as compared to conventional agriculture practices. Indoor farm developer Mirai reported that Japanese vertical farm of 25,000 square meters produce 10,000 heads of lettuce per day which is double the amount produced under traditional field condition. Japanese vertical farms used nearly 40% less energy and 99% less water consumption as compared to open field type cultivation practice. Aerofarms, another vertical farm developer, reported 390 times higher yields and nearly utilizing 95% less water as compared to conventional cultivation practice. All abiotic parameters required for plant growth is controlled by sensors. Carbon dioxide concentration in plant growing environment is very critical parameter. Non-dispersive infrared (NDIR) carbon dioxide sensors play important role to detect and control the concentration in artificial vertical farms system. Edinburgh Sensors designed Boxed Gascard for vertical farms environment. Boxed Gascard employs a pseudo dual beam NDIR measurement system to enhance the stability and reduced optical complexity. Mint Controls reported development of IoT-connected vertical farm that do not require human hands to touch the crops at any stage (Chowdhury et al. 2020).

1680

A. K. Singh

Livestock Applications Livestock are important part of agriculture sectors. Livestock or farm animals are domestic animals raised for producing labors and different commodities (such as meat, eggs, milk, etc.). Water buffalo, cattle, sheep, donkey, camel, and goat are some common examples of livestock. Maintaining livestock in clean and optimal environment is necessary for good productivity. Optimum supply of food and water is also essential for maintaining proper health of livestock. IoT-based application has been used successfully for managing livestock. WSN-based technology has been employed for monitoring and controlling the environmental conditions. It has also been used for monitoring feeding practice, animal tracking, and animal behaviors. Murphy et al. (2015) reported use of WSN technology for monitoring a beehive colony. Their group collected important information about activity and environment within a beehive as well as its surrounding area.

Future Challenges IoT-based smart agriculture has several challenges that need to be addressed before it becomes imperative part of agricultural practice. Challenges range from hardware problem to security concerns. The equipment present at the perception layer are directly exposed to harsh environmental conditions capable of destroying the electronics circuits. Environmental condition such as soaring solar radiation, extreme temperatures, rain, humidity, winds, vibrations, and dust storms can damage electronic circuits and render perception layer nonfunctional. The devices of IoT system has to stay active and function perfectly over long period with limited availability of power resources from batteries. Suitable programming tools and low-power utilizing system needs to be developed as frequent battery replacement or reset of the system is not an easy task in the large-scale open fields. Small power harvesting modules like solar panels, wind turbines can be applied to generate power to some extent in open fields but still low power utilizing equipment are prerequisite for successful IoT system. Ziegeldorf et al. (2014) highlighted the importance of small number of interconnected devices produces exceedingly large amount of data. Small-scale server infrastructures usually fail to handle such large amount of data (Atzori et al. 2010). Logistics infrastructure dealing with food and agricultural sector aims to assist the exchange of information as well as transport of goods. Proper logistics can optimize the production process and the supply chain networks globally. Including IoT-based technology in logistics have progressively transformed post-harvest agriculture business processes by providing more precise and real-time information to the movement of materials and products. Inclusion of cloud computing into IoTbased logistics system has increased service quality primarily by offering ample storage and computational resources to store and process the data generated at the edge of the network. Botta et al. (2014) used the “CloudIoT” to highlight the ideal

65 Smart Farming: Applications of IoT in Agriculture

1681

accompaniment of IoT technologies and cloud computing. The enormous amount of data generated at the edge of the network can sustain unpleasantly high cost for transferring it to the cloud. Therefore, the researchers and industries must develop technologies that can draw best possible balancing between the edge storage and processing as well as determine the portion of the workload that is to be done on the cloud. Expanding cloud technologies and tools is must for satisfying the future requirements of IoT-based technologies. Fog computing which is an extension of the Cloud Computing paradigm offers such advancements. It is decentralized computing infrastructure that uses edge devises to carry out sizeable amount of computation, data storage, and application locally and routed over the Internet (Jukan et al. 2019; Bo and Wang 2011). Another challenge for IoT-based system is the network layer. Normally wireless communication systems are preferred over wired base data transfer system in various agricultural deployments. It is well established that environment condition has considerable influence on wireless link quality. Environment condition can decrease wireless link quality through the multi-path propagation effects and its contribution to background noise (Wang et al. 2017). In the real world, performance of popular transceivers has been found to be influenced by temperature, humidity, and presence of human and other obstacles within the space where a wireless node attempts to communicate. Thus, there is need for robust and reliable data transfer technologies that can work accurately according to the requirements and challenges of the rural environment. Security concerns are another challenge commonly faced by IoT-based system. The safe IoT-based system transfer data through interconnected Internet that can ensure the security, legitimacy, confidentiality, and privacy of the stakeholders involved in this group. A well-protected IoT-based system guard data present in perception layer, network layer, and application layer from external attacks. It ensures that only approved entities can access and alter data in the application layer. IoT security requires to fulfill three primary requirements, namely, authentication, confidentiality, and access control. The first layer, that is, perception layer often faces most common security breach. This layer requires security not only for the acquired information but also physical security for the hardware components. In case of agriculture, the physical security of hardware components is quite important, since the devices are deployed in open fields and work continually without proper surveillance for long periods under diverse environmental condition. Very often a single security protocol may not be effective in providing protection to all devices as they are deployed in diverse environments and distributed nature of IoT. RFID commonly used in IoT has security concern related to leakage of information. Information leakage from RFID may reveal the location identity and other similar sensitive data. RFID security has been achieved by several countermeasures like data encryption, use of blocker tags, and tag frequency modification. Jamming of tags and tag destruction policy which enables physical ending of a tag’s life has also been used as countermeasures. Security of sensor can be achieved through several policies like cryptographic algorithms, identity authentication mechanisms,

1682

A. K. Singh

data flow control policies, and data filtering mechanisms (Li et al. 2012). This layer is also threatened by wiretapping, tampering, cheating, and replay attacks. Therefore, data acquisition and key management protocols must be guaranteed with authenticity, confidentiality, and data integrity. Secure routing policies must be adopted and sensor node verification policies must be controlled to prevent data access by unauthorized things. Middleware layer of IoT also has specific security requirements. This layer is positioned between the perception and application layers and is accountable for data processing. Middleware layer enables communication interface between network and application layers. Confidentiality and secure data storage are primary security requirement of middle layer. Data transmissions through wireless medium often offer security challenges to IoT system. IoT systems are easily exposed to threat, such as denial of service attacks, virus attack, man-in-the-middle attacks, and unauthorized access. These security threats often target and affect privacy and data integrity of network layer. Security threats of middleware layer can be overcome by authentication, intrusion detection, key management, and negotiation mechanisms (Demesticha et al. 2020; Farooq et al. 2019). Application layer is the top layer in the IoT system. Enormous amount of data streams end up in this layer, thus requiring improved storage and computational resources. Often the application layer is so closely related with the cloud and thus security threats of application layer are similar to the security issues of the cloud itself. Application layer experiences threats for data security, privacy, backup, and recovery. Secured application layer must have mechanisms to manage the rights and ownership of data. It must also control the access rights to all, or part of the information, both for users and between machines, or even organizations (Farooq et al. 2019).

Conclusions The agriculture sector has undergone several important changes during last few decades. These changes have transformed traditional agriculture sector into a new smart agriculture. The fourth revolution in agriculture has transformed traditional farming approaches by introducing ICT technologies. The use of contemporary software and hardware technologies like WSNs, IoT, UAVs, cloud computing, and machine learning has enabled development of smart agriculture. These technologies have potential to further enhance crop yield, improve harvest quality, reduce cost of production, and lessen the ecological footprint of traditional farming. Smart farming technologies have aided farmers to practice modern process by monitoring crops even at a per plant level. IoT is expected to optimize the agriculture process by many means. Recent advances in IoT-based open-field and protected cultivation has paved pathway to move farming from precision to a micro-precision model of agricultural production. Precise monitoring, pervasive computing, and proper agriculture decision support system are bound to provide the optimal growing or living conditions for both crops

65 Smart Farming: Applications of IoT in Agriculture

1683

and animals. Automation of agriculture systems will allow optimization of service and resource usage as well as control the crop production in harmony to the market condition. This approach is bounded to maximize the profit and minimize the cost of production in every possible way. In a real world, IoT-based agriculture system employ diverse equipment, with different technical specifications and sensor features. Data coming from vast heterogeneous sources can be used to optimize agriculture decision support system or expert system only after making it interpretational, filtration, and the semantic annotation of the data. Food supply chains and agriculture logistics system armed with WSN and RFID equipment can allow monitoring of product at each stage of its life thus imparting feeling of safety for consumer, through a transparent product lifecycle information system. All the abovementioned approaches are the optimistic advances of the IoTbased agriculture system. However, in this perception, several individual players have to participate in accord. Foremost, the local networks systems have to be protected against interference from other unwanted networks. Undoubtedly, IoT-based agricultural technologies have tremendous potential to serve future human demand in sustainable manner. However, still lots of issues need to be addressed before IoT-based agricultural technologies become imperative part of farming and affordable for every farmer. IoT-based agricultural technologies along with robotics and artificial intelligence algorithms are about to bring about fifth revolution in agriculture.

References Alfian G, Syafrudin M, Rhee J (2017) Real-time monitoring system using smartphone-based sensors and NoSQL database for perishable supply chain. Sustainability 9:2073 Andújar D, Ribeiro A, Quintanilla C F, Dorado J (2009). Assessment of a ground-based weed mapping system in maize. In: Book Precision Agriculture. vol 09, pp 355–362 Arora NK (2019) Impact of climate change on agriculture production and its sustainable solutions. Environ Sustain 2:95–96 Atzori L, Iera A, Morabito G (2010) The internet of things: A survey. Computer Networks 54(15):2787–2805 Ayaz M, Ammad-Uddin M, Sharif Z, Mansour A, Aggoune EM (2019) Internet-of-things (IoT)based smart agriculture: toward making the fields talk. IEEE Access 7:129551–129583. https:// doi.org/10.1109/ACCESS.2019.2932609 Bac CW, Hemming J, Tuijl BAJ, Barth R, Wais E, Henten EJ (2017) Performance evaluation of a harvesting robot for sweet pepper. J Field Robotics 34:1123–1139 Beecham Research (2016) An introduction to LPWA public service categories: matching services to IoT applications. https://www.iot-now.com/2016/04/26/46490-introducing-lpwapublic-service-categories-matching-services-to-applications/ Bo Y, Wang H (2011) The application of cloud computing and the internet of things in agriculture and forestry. In: 2011 international joint conference on service sciences, Taipei, pp 168–172 Botta A, de Donato W, Persico V, Pescapé A (2014) On the integration of cloud computing and internet of things. In: International conference on future internet of things and cloud, pp 23–30. https://doi.org/10.1109/FiCloud.2014.14

1684

A. K. Singh

Bueno-Delgado MV, Molina-Martínez JM, Correoso-Campillo R, Pavón-Mariño P (2016) Ecofert: an android application for the optimization of fertilizer cost in fertigation. Comput Electron Agric 121:32–42 Camacho HAA (2018) Smartphone-based application for agricultural remote technical assistance and estimation of visible vegetation index to farmer in Colombia: AgroTIC, Remote Sens. Agric Ecosyst Hydrol SPIE Remote Sens 10783 Ceballos F, Kramer B, Robles LM (2018) The Feasibility of Picture-Based Insurance (PBI): smartphone pictures for affordable crop insurance. IFPRI Discussion Paper 1788. Available at SSRN: https://ssrn.com/abstract=3324536 Chen J, Yang A (2019) Intelligent agriculture and its key technologies based on internet of things architecture. IEEE Access 7:77134–77141 Chowdhury MEH, Khandakar A, Ahmed S, Al-Khuzaei F, Hamdalla J, Haque F, Reaz MBI, Al Shafei A, Al-Emadi N (2020) Design, construction and testing of IoT based automated indoor vertical hydroponics farming test-bed in Qatar. Sensors 20:5637 Chung S, Breshears LE, Yoon JY (2018) Smartphone near infrared monitoring of plant stress. Comput Electron Agric 154(August):93–98 Cocovi-Solberg DJ, Rosende M, Miro M (2014) Automatic kinetic bio accessibility assay of lead in soil environments using flow-through micro dialysis as a front end to electro thermal atomic absorption spec-trometry. Environ Sci Technol 48:6282–6290 Crabit A, Colin F, Bailly JS, Ayroles H, Garnier F (2011) Soft water level sensors for characterizing the hydrological behaviour of agricultural catchments. Sensors 11:4656–4673 de la Piedra A, Braeken A, Touhafi A (2012) Sensor systems based on FPGAs and their applications: a survey. Sensors 12(9):12235–12264 Defterli SG, Shi Y, Xu Y, Ehsani R (2016) Review of robotic technology for strawberry production. Appl Eng Agric 32(3):301–318 Demesticha K, Peppes N, Alexakis T (2020) Survey on security threats in agricultural IoT and smart farming. Sensors 20(22):6458 Díez C (2017) Hacia una agricultura inteligente (Towards and intelligent Agriculture). Cuaderno de Campo 60:4–11 Doshi J, Patel T, Bharti SK (2019) Smart farming using IoT, a solution for optimally monitoring farming conditions. Proc Comp Sci 160:746–751., ISSN 1877-0509. https://doi.org/10.1016/ j.procs.2019.11.016 Dvorak JS, Stone ML, Self KP (2016) Object detection for agricultural and construction environments using an ultrasonic sensor. J Agric Saf Health 22(2):107–119 Faiçal B, Costa FG, Pessin, Ueyama J, Freitas H, Colombo A et al. (2014) The use of unmanned aerial vehicles and wireless sensor networks for spraying pesticides. J Syst Archit 60(4): 393–404 FAO (2009) Global agriculture towards 2050. Retrieved August 10, 2016, from http://www.fao.org/ fileadmin/templates/wsfs/docs/Issues_papers/HLEF2050_Global_Agriculture.pdf FAO Accessed: Apr. 15, 2019. [Online]. Water for Sustainable Food and Agriculture by FAO. www.fao.org/3/a-i7959e.pdf Farooq MS, Riaz S, Abid A, Abid K, Naeem MA (2019) A survey on the role of IoT in agriculture for the implementation of smart farming. IEEE Access 7:156237–156271. https:/ /doi.org/10.1109/ACCESS.2019.2949703 Fernández-Ahumada LM, Ramírez-Faz J, Torres-Romero M, López-Luque R (2019) Proposal for the Design of Monitoring and Operating Irrigation Networks Based on IoT, cloud computing and free hardware technologies. Sensors 19:2318 Frommberger L, Schmid F, Cai C (2013) Micro-mapping with smartphones for monitoring agricultural development. ACM DEV ’13: Proceedings of the 3rd ACM Symposium on Computing for Development. https://doi.org/10.1145/2442882.2442934 García-Ramos FJ, Vidal M, Boné A, Malón H, Aguirre J (2012) Analysis of the air flow generated by an air-assisted sprayer equipped with two axial fans using a 3D sonic anemometer. Sensors (Basel) 12(6):7598–7613

65 Smart Farming: Applications of IoT in Agriculture

1685

Gasso-Tortajada V, Ward AJ, Mansur H, Brøchner T, Sørensen CAG, Green OA (2010) Novel acoustic sensor approach to classify seeds based on sound absorption spectra. Sensors (Basel) 10(11):10027–10039 Gubbi J, Buyya R, Marusic S, Palaniswami M (2013) Internet of things (IoT): a vision, architectural elements, and future directions. Futur Gener Comput Syst 29(7):1645–1660 Hemmat A, Binandeh AR, Ghaisari J, Khorsandi A (2013) Development and field testing of an integrated sensor for on-the-go measurement of soil mechanical resistance. Sensors Actuators A Phys 198:61–68 Hillman GC (1996) Late Pleistocene changes in the wild plant foods available to huntergatherers of the northern Fertile Crescent: possible preludes to cereal cultivation. In: Harris DR (ed) The origins and spread of agriculture and pastoralism in Eurasia, UCL Press, London, pp 159–203 Hong G-Z, Hsieh C-L (2016) Application of integrated control strategy and Bluetooth for irrigating Romaine lettuce in greenhouse. IFAC-PapersOnLine 49(16) Jedermann R, Borysov M, Hartgenbusch N, Jaeger S, Sellwig M, Lang W (2018) Testing Lora for food applications – example application for airflow measurements inside cooled warehouses with apples. Proc Manuf 24:284–289 Jukan A et al (2019) Fog-to-cloud computing for farming: low-cost technologies, data exchange, and animal welfare. Computer 52(10):41–51 Kou Z, Wu C (2018) Smartphone based operating behaviour modelling of agricultural machinery. IFAC-PapersOnLine 51(17):521–525 Kumar NM, Mallick PK (2018) The internet of things: insights into the building blocks, component interactions, and architecture layers. Proc Comput Sci 132:109–117 Kumar A, Bhatia A, Fagodiya RK et al (2017) Eddy covariance flux tower: a promising technique for greenhouse gases measurement. Adv Plants Agric Res 7(4):337–340 Lampridi MG, Sørensen CG, Bochtis D (2019) Agricultural sustainability: a review of concepts and methods. Sustainability 11:5120 Lavric A, Petrariu AI, Popa V (2019) Long range SigFox communication protocol scalability analysis under large-scale, high-density conditions. IEEE 7:35816–35825 Lee H, Moon A, Moon K, Lee Y (2017) Disease and pest prediction IoT system in orchard: A preliminary study. In: 9th international conference on ubiquitous and future networks (ICUFN). pp 525–527. https://doi.org/10.1109/ICUFN.2017.7993840 Li H, Chen Y, He Z (2012) The survey of RFID attacks and defenses. In: 8th International conference on wireless communications, networking and mobile computing. pp 1–4. https:// doi.org/10.1109/WiCOM.2012.6478720 Lorvanleuang S, Zhao Y (2018) Automatic irrigation system using android. Open Access Library Journal 5:1–7 Machado BB, Orue JPM, Arruda MS, Santos CV, Sarath DS, Goncalves WN, Silva GG, Pistori H, Roel AR, Rodrigues-Jr JF (2016) BioLeaf: a professional mobile application to measure foliar damage caused by insect herbivory. Comput Electron Agric 129:44–55 Manfrini LL et al (2015) Monitoring strategies for precise production of high quality fruit and yield in apple in Emilia-Romagna. Chem Eng Trans 44:301–306 Mekki K, Bajic E, Chaxel F, Meyer F (2019) A comparative study of LPWAN technologies for large-scale IoT deployment. ICT Express 5:1–7 Millan-Almaraz JR, Romero-Troncoso RJ, Guevara-Gonzalez RG, Contreras-Medina LM, Carrillo-Serrano RV, Osornio-Rios RA, Duarte-Galvan C, Rios-Alcaraz MA, Torres-Pacheco I (2010) FPGA-based fused smart sensor for real-time plant-transpiration dynamic estimation. Sensors 10:8316–8331 Molina I, Morillo C, García-Meléndez E, Guadalupe R, Roman MI (2011) Characterizing olive grove canopies by means of ground-based hemispherical photography and spaceborne RADAR data. Sensors (Basel) 11(8):7476–7501 Montagnoli A et al (2015) Estimating forest aboveground biomass by low density lidar data in mixed broad-leaved forests in the Italian pre-Alps. Forest Ecosyst 2:10

1686

A. K. Singh

Murphy FE, Popovici E, Whelan P, Magno M (2015) Development of an heterogeneous wireless sensor network for instrumentation and analysis of beehives. In: Instrumentation and measurement technology conference (I2MTC), 2015 IEEE international. pp 346e351 Navarro E, Costa N, Pereira A (2020) A systematic review of IoT solutions for smart farming. Sensors (Basel) 20(15):4231. https://doi.org/10.3390/s20154231 Oliveira KVD, Esgalha Castelli HM, José Montebeller S, Prado Avancini TG (2017) Wireless sensor network for smart agriculture using ZigBee protocol. In: 2017 IEEE First Summer School on Smart Cities (S3C). Natal, pp 61–66, https://doi.org/10.1109/S3C.2017.8501379 Orlando F, Movedi E, Coduto D, Parisi S, Brancadoro L, Pagani V, Guarneri T, Confalonieri R (2016) Estimating leaf area index (LAI) in vineyards using the PocketLAI smart-app. Sensors (Basel) 16(12):1–12 Pajares G (2011) Advances in sensors applied to agriculture and forestry. Sensors 11:8930–8932 Pajares G, Peruzzi A, Gonzalez-de-Santos P (2013) Sensors in agriculture and forestry. Sensors (Basel). 13(9):12132–12139 Pérez-Castro A, Sánchez-Molina JA, Castilla M, Sánchez-Moreno J, Moreno-Úbeda JC, Magán JJ (2017) cFertigUAL: a fertigation management app for greenhouse vegetable crops. Agric Water Manag 183:186–193 Pongnumkul S, Chaovalit P, Surasvadi N (2015) Applications of smartphone-based sensors in agriculture: a systematic review of research. J Sens. Open Access Article ID 195308 Ramalingam B, Mohan RE, Pookkuttath S et al (2020) Remote insects trap monitoring system using deep learning framework and IoT. Sensors (Basel) 20(18):5280 Rawal S (2017) IOT based smart irrigation system. Int J Comput Appl 159(8):7–11 Rose I, Welsh M (2010) Mapping the urban wireless landscape with Argos. In: Proceedings of the 8th ACM conference on embedded networked sensor systems, SenSys ‘10. ACM, New York Saiz-Rubio V, Rovira-Más F (2020) From smart farming towards agriculture 5.0: a review on crop data management. Agronomy 10:207 Schimmelpfennig D (2016) Farm profits and adoption of precision agriculture. USDA 217:1–46 Scholz M (2018) Enhancing adoption of integrated Weed Management – an Australian Farmer’s perspective. Outlooks Pest Manag 29(2):66–69(4) Smart Agriculture Market Published Date: Apr 2020 | Report Code: SE 4043. https:// www.marketsandmarkets.com/PressReleases/smart-agriculture.asp Sopegno A, Calvo A, Berruto R, Busato P, Bocthis D (2016) A web mobile application for agricultural machinery cost analysis. Comput Electron Agric 130:158–168 Stiglitz R, Mikhailova E, Post C, Schlautman M, Sharp J, Pargas R, Glover B, Mooney J (2017) Soil color sensor data collection using a GPS-enabled smartphone application. Geoderma 296: 108–114 Suhonen J, Kohvakka M, Kaseva V, Hamalainen TD, Hannikainen MH (2012) Low-power wireless sensor networks: protocols, services and applications 2012. Springer. https://doi.org/10.1007/ 978-1-4614-6381-8 Torbick N, Chowdhury D, Salas W, Qi J (2017) Monitoring Rice agriculture across Myanmar using time series Sentinel-1 assisted by Landsat-8 and PALSAR-2. Remote Sens 9:119 Tong-ke F (2013) Smart agriculture based on cloud computing and IOT. J Converg Inf Technol 8:210–216 Tzounis A, Katsoulas N, Bartzanas T, Kittas C (2017) Internet of things in agriculture, recent advances and future challenges. Biosyst Eng 164:31–48 Walsh F (2009) Human-animal bonds I: The relational significance of companion animals. Fam Process 48:462–480 Wan X, Cui J, Jiang X, Zhang J, Yang Y, Zheng T (2018) Smartphone based hemispherical photography for canopy structure measurement. In: Proceedings of the international conference on optical instruments and technology, Beijing, China, p 6 Wang Z, Walsh KB, Verma B (2017) On-tree mango fruit size estimation using RGB-D images. Sensors 17(12):2738 Welbourne E, Battle L, Cole G, Gould K, Rector K, Raymer S et al (2009) Building the internet of things using RFID: the RFID ecosystem experience. IEEE Internet Comput 13(3):48–55

65 Smart Farming: Applications of IoT in Agriculture

1687

World Bank (2016). http://data.worldbank.org/indicator/AG.LND.AGRI.ZS Yu QY, Shi Y, Tang HJ, Yang P, Xie AK, Liu B, Wu WB (2017) eFarm: a tool for better observing agricultural land systems. Sensors 17 Yunus MAM, Mukhopadhyay SC (2011) Novel planar electromagnetic sensors for detection of nitrates and contamination in natural water sources in IEEE. Sensors 11(6):1440–1447 Zhai Z, Martínez JF, Beltran V, Martínez NL (2020) Decision support systems for agriculture 4.0: survey and challenges, computers and electronics in agriculture. 170:105256. ISSN 0168-1699 Zhang S, Chen X, Wang S (2014) Research on the monitoring system of wheat diseases, pests and weeds based on IOT. In: 9th international conference on computer science and education. pp 981–985. https://doi.org/10.1109/ICCSE.2014.6926609 Zhang L, Dabipi IK, Brown WL Jr (2018) Internet of things applications for agriculture. In Internet of things A to Z, Q. Hassan (Ed.). https://doi.org/10.1002/9781119456735.ch18 Ziegeldorf JH, García-Morchón Ó, Wehrle K (2014) Privacy in the internet of things: threats and challenges. Secur Commun Networks 7:2728–2742 Zujevs A, Osadcuks V, Ahrendt P (2015) Trends in robotic sensor technologies for fruit harvesting: 2010–2015. Proc Comput Sci 77:227–233 Zulkifli C, Noor N (2017) Wireless sensor network and internet of things (IoT) solution in agriculture. Pertanika J Sci Technol 25(1):91–100

Some Important Websites http://www.easyfarm.com/ http://www.mbeguchoice.com/ https://edinburghsensors.com/industries/agriculture/ https://farmlogs.com/ https://www.eip-water.eu/projects/opiris-online-precise-irrigation-scheduling-algorithm https://www.farmtrx.com/ https://www.farmx.co/ https://www.marketsandmarkets.com/Market-Reports/smart-agriculture-market-239736790.html; https://www.cropx.com/ https://www.onfarm.com/

Smart Materials in Oil and Gas Industry: Application

66

Alimorad Rashidi and Soheila Sharafinia

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart Materials Used in Oil and Gas Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of Smart Materials in Oil and Gas Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of Smart Materials in Enhanced the Oil Recovery (EOR) . . . . . . . . . . . . . . . Application of Smart Materials in Drilling Fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of Smart Materials in Oil Well Cement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of Smart Materials in Oil-in-Water and Water-in-Oil Separation . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1690 1692 1694 1694 1703 1706 1708 1722 1722

Abstract

In recent years, smart materials are used as a different generation of materials in different types of fields due to their environmentally friendly advantages, high efficiency, and low cost. These materials have properties that can be manipulated for a reversible and controllable response to environmental changes such as pressure, heat, humidity, voltage, stress, chemical compounds and mechanical force, pH, electricity, or magnetism. Therefore, these materials are known as smart, stimuli-responsive or reactive materials. Smart materials are used in a variety of fields, for example, the oil and gas industry, personal medical diagnosis, telephones and entertainment centers, construction, wastewater treatment,

A. Rashidi () Nanotechnology Research Center, Research Institute of Petroleum Industry, Tehran, Iran e-mail: [email protected] S. Sharafinia Department of Chemistry, Faculty of Science, Shahid Chamran University of Ahvaz, Ahvaz, Iran © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_115

1689

1690

A. Rashidi and S. Sharafinia

civil engineering applications, and aerospace. In this review, we have focused on investigating the potential application of smart materials in the oil and gas industry. Smart materials are used in the oil and gas industry in different fields. These materials are used as nanotracer and smart water for crude oil exploration and oil recovery, respectively. One of the important applications of this material is nano additives, not only is it low cost, but it has high stability and efficiency. Switchable wettability membranes also showed high performance in the separation of oil-in-water and water-in-oil emulsions. In general, smart materials are economically suitable, and do not require complex equipment, so they improve the speed of the production process, increase the final harvest, reduce adverse environmental effects, significantly improve oil recovery, and also increase water and oil separation, and information of production. Keywords

Smart mud · Nanocement · Nanosmart additive · Smart water · Nanosensor

Introduction Smart materials are called smart because they can reversibly respond to external stimuli, including temperature, mechanical, light, pH, and electrical and chemical energy (Rezaeian et al. 2020). This material has wide applications, for example, controlled drug delivery, robots, switches, artificial muscles, the oil and gas industry, actuators, and chemical sensors (Rezaeian et al. 2020). In this study, we have focused on the investigation of smart material applications in the oil and gas industry. One of the most essential subjects in the oil and gas field is requirements to oil transfer, which in recent years have been more attentive (see Graph 1). Major part of oil reserves is formed of carbonate reservoirs, and these carbonate reservoirs, due to their wettability and tightness of matrix, have little oil recovery efficiency and therefore the oil recovery in carbonate reservoirs is down. Since the permeability of the network has more fractures than the rock matrix, the oil inside the fractures exit when water is transported; however, an enormous value of oil exists in the rock matrix. The automatic ignition mechanism is one of the primary methods for oil production in fractured carbonate reservoirs, but the rock matrix oil-wet nature prevents this method’s effectivity. Recently, the smart water injection is one of the best approaches for oil recovery from fractured carbonate reservoirs (Nowrouzi et al. 2020). Also, with the expansion of the industry, the release of dangerous waste, for example, petroleum, has entered the water and caused severe damage to the environment (Nowrouzi et al. 2020; Harrison 2007). The different materials such as activated carbon, smart materials, (Walcarius and Mercier 2010) metals oxides, metal−organic frameworks (MOF), polymeric and synthetic organic materials, macroalgae, and inorganic as well as natural solid have been extensively

66 Smart Materials in Oil and Gas Industry: Application

1691

50

Count

40 30 20 10 0 2008

2010

2012

2014

2016

2018

2020

2022

Year Graph 1 Developments in the use of smart materials in the oil and gas industry over the past 10 years

used for the removal of pollutants from the wastewater (Kim et al. 2017). But, often used traditional materials used are usually nonrecyclable and nonselective which can come with high costs, and recently, smart materials have been able to challenge conventional (traditional) materials because they have selective adsorption and high desorption efficiency. One other case in the oil and gas industry is the drilling fluids. These fluids in the drilling industry play the blood role, in which hollow drill pipes are pumped down and then flushed back up using drill cuttings. The shear-thinning rheological factor is one of the important parameters, which must be highly adhesive to cause transforming drill cuttings from the down of the good hole (Li et al. 2015a). These theologies using a lot of nanomaterials such as CNTs, graphene oxide, carbon ash, magnesium aluminum silicate, laponite, CuO, ZnO, Al2 O3 , TiO2 , SiO2 , Fe3 O4 , and Fe2 O3 have been modified (Barry et al. 2015). Today, with population growth and subsequent increase in demand and reduction of oil resources, drilling operations are carried out in challenging physicochemical conditions such as salinity, pressure, pH, and temperature, which lead to significant changes in the drilling fluids rheology. One way to confront with this problem is to produce smart drilling fluids that change their rheology in response to stimuli, including magnetic/electric field, mechanical force, salinity, temperature, and pH (Rezaeian et al. 2020). The next case in the oil industry is oil well cement; the empty places between the pipe and rock formation are filled using these materials. Also good types of cement have been used to protect from casing against corrosion, and translocate the drilling fluids, prevent of corrosion. The most important parameters of cements is their resistance to corrosive liquids and high pressures. One of the applications of oil well cement

1692

A. Rashidi and S. Sharafinia

is to increase the strength while having a higher ductility, and currently different additive nanoparticles such as smart materials are used for this purpose.

Smart Materials Used in Oil and Gas Industry As mentioned in the types of literature, smart nano-water flooding is obtained by injecting nanoparticles dispersed in alcohol, brine, and water. The amount of enhancement in oil recovery depends on the intrinsic properties and the large surface area of the injected nanoparticles, for example, substances such as SiO2 NPs greatly reduce interfacial tension (IFT). Nanoparticles such as MgO, NiO, Fe2 O3 , CuO and Al2 O3 can also be used to increase the viscosity of injected materials, reduce the viscosity of the oleic phase, change the moisture content of the rock surface and improve oil recovery. A summary of the materials used in the oil recovery process is presented in Table 1 (Gbadamosi et al. 2019). Smart cements are fabricated by the addition of materials such as nano-SiO2, ZnO, multiwalled CNTs (MWCNTs) and single-walled CNTs (SWCNTs) (Gao et al. 2009), graphene oxide (GO), carbon nanofibers (CNF) (Phrompet et al. 2019), nanocarbon black (NCB), and Fe2 O3 NPs to increase the life and safety of the cement structure. Materials such as nano-Al2 O3 , nickel powder, steel slag, carbon black, steel fiber, and carbon fiber play a filler and electrical conductive role in cement-based nanocomposites. The presence of these materials causes increasing pressure, cracks, stress, strain, and diagnosis (Bogue 2012). The well-known smart materials that are used to oil-in-water and water-in-oil separating are (1,3-dimethylamylamine) (DMAA)grafted Si surface (light and temperature-responsive), poly (N-isopropylacrylamide) (PNIPAM)/Poly[2 (Dimethylamino)ethyl Methacrylate] (PDMAEMA) grafted 1H, 1H, 2H, 2H perfluorooctyltrichlorosilane (PFOTS) modified Al2 O3 (pH and electrolyte-responsive), P(NIPAAm-co-AAc) thin films (PNIPAAm+PAAc), Cellulose-PDMAEMA, and PDMAEMA hydrogel coated mesh (pH and Temperature) (Xia et al. 2007). The next material is smart polymer membranes, which are widely used in the field of oil-in-water and water-in-oil separation. These membranes reversibly switch the amount of wettability by changing external stimuli such as temperature, pH, light, and electricity. Nanoparticles of TiO2 and ZnO have the property of responding to ultraviolet light, which is easily coated on the surface of the materials (Gao et al. 2013). Also, the poly(4-vinylpyridine) and the mix of HS(CH2 )9 CH3 and HS(CH2 )10COOH were grafted with the nylon membrane, preparing pH-responsive membranes (Zhang et al. 2020). The membranes prepared of PDMAEMA and PNIPAM (Zhang et al. 2020) indicated hydrophobicity-hydrophilicity switch ability in different temperatures, therefore they are used to separating oil-in-water and water-in-oil separation.

66 Smart Materials in Oil and Gas Industry: Application

1693

Table 1 Some of oil recovery using smart materials NPs SiO2 Al2 O3 TiO2

Base fluid Deionized water

Salinity (wt%) Brine 1 (2 wt% NaCl, 0.2 wt% KCl, 0.2 wt% CaCl2 , 0.1 wt% MgCl2 ) Brine 2 (3 wt% NaCl) NaCl (3 wt%)

Rock type Limestone

Oil recovery –

LHPN, NWPN, HLPN HLPN SiO2 TiO2 Hydrophobic fumed SiO2 Al2 O3 , TiO2 , SiO2 SiO2 , IIT

Ethanol and water

Sandstone

0.75–36.67%

Ethanol Brine Brine Ethanol

NaCl (3 wt%) NaCl (3 wt%) NaCl (0.5–1.0 wt%) NaCl (5 wt%)

Sandstone Sandstone Sandstone Sandstone

19.31% 0–14.29% ∼31% 25.43% 14.55%

PVP

NaCl (3.0 wt%)

Sandstone

19–25%

DIW

NaCl, KCl, MgCl2 ·6H2 O, CaCl2 , Na2 SO4 , NaHCO3 (2.0 wt%) NaCl (7.0 wt%)

Sandstone

50–55%

Sandstone

SiO2 (5.14–13.88% Al2 O3 (− 8.18–4.65%) 0.9–9.49 wt%

SiO2 , Al2 O3

Brine

Fe2 O3 , Al2 O3 , SiO2

Brine

SiO2 SiO2 SiO2

Brine Brine Brine

SiO2

Distilled water Brine

SiO2

Brine A (3.0 wt%) Brine B (1.5 NaCl,1.0 CaCl2 , 0.5 MgCl2 ) NaCl (3.0 wt%) NaCl (3.0 wt%) NaCl (5.0 wt%)

Sandstone

Sandstone Sandstone Carbonate

–

Micromodel

4.26–5.32% 5.0–15% 9–12% 16–17% (24 h aging) 8.7–26%

NaCl (6.5 wt%)

Sandstone

9.0–19%

1694

A. Rashidi and S. Sharafinia

Application of Smart Materials in Oil and Gas Industry Application of Smart Materials in Enhanced the Oil Recovery (EOR) The effect of the concentration of soluble ions in seawater was studied using smart water samples. Acetone and methanol have also been used as a solvent in different volume ratios. When using acetone as a solvent, a more significant reduction in surface tension (IFT) is observed than in methanol solvent, it can be said this reduction is directly related to the presence of solvents in water and the pressure of CO2 . The value of IFT is decreased with increasing solvent at constant concentration of seawater, which requires further study of the structure and behavior of the solvents. Methanol has composed of hydrogen bonds with two parts, one of part is nonpolar hydrophobic, and the other part is polar hydrophilic. When it was dissolved in water the water-water hydrogen bond was weakened due to the formation of the methanol-water hydrogen bond, and as a result, it decreases IFT (Biscay et al. 2011). On the other hand, the density of binary solutions is effective in reducing IFT. Since water is denser than acetone and methanol solvents when these solvents are dissolved in water, the solution density is much lower than the density of water (i.e., close to the density of crude oil). The density difference calculated from Eq. 1: 

γ=

ρgD 2 H

(1)

In this relation parameters ρ (g/cm3 ), g (cm/s2 ), and D (cm) refer to the difference in the solutions densities, gravitational acceleration of the earth, and the larger diameter of the droplet, respectively. Also, H parameter depends on the shape of the drop. Simultaneously with the adsorption of the solvent on the water-oil interface, a new layer is created, which leads to the enhanced adsorption of ions and thus increases the thickness of the oil solvent-ion layer (Fig. 1). As salinity increases, the amount of IFT increases and the solubility of solvents decreases, thus increasing the transfer of solvent mass from the aqueous phase to the oil phase. Therefore, oil swelling increases through the adsorption of solvent molecules by the oil phase. By increasing the aging of the carbonate surface of the crude oil, oil-wet is obtained. Figure 2a and b shows the contact angles of 31 and 161 for oil dropping on the carbonate surface before and after aging, respectively. This change is due to the absorption of carboxylic acids in crude oil during the aging process. The contact angle is decreased with increasing concentration of Mg2+ and Ca2+ ions in smart water. But because Mg2+ ions have a higher charge density than Ca2+ ions, they have a more significant effect on the change in wettability. As mentioned in different kinds of literature, carboxylic acid adsorption from the crude oil by the rock surface affects the wettability, and the cations adsorbed these acids. In the following, the mechanism of the change of wettability by smart water will be discussed. The concentration of SO4 2− ions affects the divalent cations and the

66 Smart Materials in Oil and Gas Industry: Application

1695

Fig. 1 The performances of ions in the formation of the oil solvent ion (Nowrouzi et al. 2020)

wettability of the carbonate rock surface. Increasing these ions increases the divalent cations and decreases the wettability of the treatment solution, respectively. These divalent cations, in turn, lead to the separation of carboxylic acids on the rock surface. Therefore, the contact angle is significantly reduced in smart water due to the increase in SO4 2− ion concentration. Also, as the concentration of divalent cations in smart water increases, their activity and performance in dissociating the carboxylic acids adsorbed on the rock surface will increase. The divalent ion activity is affected by the concentration of NaCl. As the NaCl concentration increases, the activity of divalent ion decreases sharply, thereby reducing their access to carboxylic acids adsorbed. Therefore, the contact angle is increased while the activity of effective ions is reduced (Strand et al. 2008). As Fig. 2c shows, in the presence of a cationic surfactant (CTAB) and anionic surfactant (SDS), the contact angle is hugely reduced. But the nonionic surfactant (TX-100) did not affect the wettability mode and contact angle. In confirmation of this discussion, we can refer to the work of Karimi et al. (2015), which has succeeded in reducing the contact angle and increasing the oil recovery by adding a (CTAB) to the solution. In Jarrahian et al. (2012), the addition of C12 TAB surfactant to the wettability modifier solution changed the wettability to a significantly compared SDS and TX-100. It also has a mechanism similar to increasing the concentration of divalent cations in a smart water solution (SW + CTAB). It dissociates the carboxylic acids adsorbed on

1696

A. Rashidi and S. Sharafinia

Fig. 2 The contact angles for oil dropping on the carbonate surface (a) before and (b) after aging. The mechanism of wettability alteration using (c) SW + C12TAB and (d) SW + SDS (Ahmadi et al. 2020)

the rock surface using their positive charge, resulting in increasing surface water wetness. For SDS to cause a considerable change in the wettability of carbonate rock, it must be mixed with smart water. The SDS by decreasing the positive charge on the rock surface reduces the access to divalent cations, therefore, reducing the contact angle. The mechanism of changing the wettability of carbonate rock is presented in Fig. 2d. In a discussion of EOR processes, S. Habibi et al. studied the ionic composition effect of SO4 2− and Ca2+ ions available in smart water in the presence of nanofluid.

66 Smart Materials in Oil and Gas Industry: Application

1697

They designed IFT measurements, contact angle, and stability to investigate the effect of this parameter on the ionic composition in EOR processes (Habibi et al. 2020). Also, for the first time, they used amine/organosiloxane@Al2 O3 /SiO2 (AOAS) nanocomposite along with smart water to change carbonate rock wettability, and observed that nanofluid with 50 ppm concentration was very stable in during the flooding cycle. When the SO4 2− and Ca2+ ions concentration increased, the contact angle is decreased and following that the rock wettability shifted toward water wetness. As shown in Fig. 3a, the wettability increases with increasing Ca2+ ions but SO4 2− ions alone cannot change the wettability. Therefore, with lack of Ca2+ ions, the oil recovery factor improvement is not considerable. According to Fig. 3b, when nanofluid enters saltwater, the nanofluid that has high stability to improve the wettability decreases for the contact angle. The reason of these observations is that nanofluids have a negative surface charge and thus reduce the positive charge of carbonates. As it approaches the surface, the concentration of Ca2+ ions is increased due to the reduction of repulsive electrostatic forces. Therefore, crude oil, which has negative carboxyl groups, binds to these ions and leaves the surface (see Fig. 3c) (Habibi et al. 2020). Composite membranes including montmorillonite (MMT), polyphenylsulfone (PPSU), and polyamide (PA) to produce smart water with lower concentrations of monovalent ions and higher concentrations of Mg2+ , SO4 2− , and Ca2+ were used for EOR processes synthesized by M.S. Rezaeian et al., and also have been examined to the change of these parameters: reaction time, trimesoyl chloride (TMC) concentration, and 1, 3-phenylenediamine (MPD) concentration; FESEM results are indicated in Fig. 4 (Rezaeian et al. 2020). As Fig. 4 shows, by the increasing these parameters, the thickness of selective layers of polyamide is increased. The high solubility of MPD in hexane than the TMC monomer’s low solubility in water leads to diffusion of MPD molecules from the aqueous phase into the organic phase and finally the increase in polyamide selective layer thickness. They studied polymerization conditions effect on the membrane’s efficiency, and observed that the polyamide layers’ thickness was increased by enhancement of organic phase monomer, reaction time, and aqueous phase monomer and have investigated the different ions’ rejection percentage, such as Mg2+ , Ca2+ , SO4 2− , Na+ , and Cl− , in the synthesized membranes under various polymerization conditions. The increase in the abovementioned parameters has caused rejection values of these ions to appear in the order SO4 2− > Mg2+ > Ca2+ > Cl− > Na+1a 1a . Sangwai et al., using crude oil from an Indian offshore oilfield, have studied low salinity water flooding efficiency of crude oil with a low acid number (Kakati et al. 2020). These studies were performed by core-flooding tests on sand, including two percentage of bentonite saturated with crude oil. Also, to see the low salinity parameter effects on the crude oil/brine/rock properties, IFT studies were carried out. To understand EOR process, effluent brine has been investigated, and the results obtained represented that seawater salinity was inversely related to the EOR process. To get a clear insight on different low salinity waterflooding effects, they performed experiments and concluded that with increasing salinity seawater flood, the crude oil recovery efficiency enhanced until it reached a flat plateau.

1698

A. Rashidi and S. Sharafinia

Fig. 3 The change of contact angle on carbonate surfaces to the (a) brines and (b) the same brines+ nanofluid. (c) Proposed mechanism of nanoparticle accumulation in the presence of Ca2+ (Habibi et al. 2020)

Haghighi et al. have successfully used smart water and a novel green surfactant for the EOR process from a carbonate oil-wet rock. The surfactant, smart water, and surfactant/ smart water combinations were used because these materials could provide wettability condition and IFT to the conversion of oil-wet to the water-wet from these rocks. They concluded that the oil recovery factor for the surfactant/smart water combination (66%) was higher than water and surfactant, because it has

66 Smart Materials in Oil and Gas Industry: Application

Fig. 4 FESEM results of TFC membranes (Rezaeian et al. 2020)

1699

1700

A. Rashidi and S. Sharafinia

become powerful by combination with smart water. When using smart water, it was observed that increasing concentration of Ca2+ , Mg2+ , and SO4 2− ions decreased the contact angle. This behavior happened at the 4Ca2+ , 4Mg2+ , and 6SO4 2− concentrations, and thus the smart water due to the modification of seawater ions includes 4Ca2+ , 4Mg2+ , and 6SO4 2− . As the surfactant concentration is increased, the pH of the surfactant solution is decreased, but the conductivity of the solution is increased, which is related to the chemical structure of dodecanoyl glucosamine. Thus, when the surfactant concentration reaches the optimal content (i.e., 800 ppm), the value of IFT reaches 19 mN/m. Besides, wettability conditions are also affected by the concentration of surfactant, that is, as the surfactant concentration is increased, the contact angle is decreased, and the wetting condition is changed from wet-oil to wet-water. The IFT and contact angle of the solutions were measured to understand the effect of ions of Ca2+ , Mg2+ , and SO4 2− in smart water combined with surfactant. According to Fig. 5a with increasing concentrations of ions in the surfactant solutions, the IFT amounts are reduced. Therefore, when Smart is combined with a surfactant, the wettability of the rock surface also changes and the lowest IFT and contact angle are obtained. The chemical mechanism was related to the modification of wettability when injecting seawater, as shown in Fig. 5b and c. According to Fig. 5b, when the temperature is low, the carboxylic group adsorbed on the surface highly reacts with Ca2+ ions and is released from the surface. The Ca2+ – carboxylate group is also displaced by Mg2+ ions at high temperatures (Fig. 5c). As a result, modification of wettability occurred only at high temperatures. Mg2+ – carboxylate bond is much weaker than the Ca2+ –carboxylate bond (Haghighi and Firozjaii 2020). In another study conducted by A. H. Saeedi Dehaghani et al. using an injection of a cationic surfactant combination with improved Persian Gulf water as smart water, the EOR process in an oil-wet carbonate reservoir has been investigated. In this study CTAB and SDS played the role of cation and anion, respectively, and the effects of the surfactants’ concentration and smart water salt were investigated. They found that when SW-1NaCl, SW-4Na2 SO4 , and SW-3KCl were used, smart water salt concentration was optimum. Moreover, optimized smart water was studied with 0.1CMC and 0.9CMC to the SDS and CTAB, respectively, and they concluded that the combination of the surfactant with smart water in carbonate reservoirs significantly affects the change in wettability. According to the results, in all salts used, the increases their contact angle with the carbonate rock when increasing salt concentration in smart water. The contact angle was decreased after reaching its maximum. In the neutral pH, the surface of carbonate rocks is positively charged; therefore, SO4 2− active anions bonded with the Ca2+ cations on the surface of the rock, releasing the carboxylic groups of the oil that were previously bonded with Ca2+ . This can be generalized to 2 CaCl2 and MgCl. To describe this phenomenon, it can be stated that with improving the bonding between Ca and phosphate, the electrostatic repultion on the rock surface is reduces, that leads to the cations transfer to the surface of rock. As a result, it cleans the surface of the stone of oil molecules due to the formation of bonds with carboxylate ions (Mofrad and Dehaghani 2020;

66 Smart Materials in Oil and Gas Industry: Application

1701

Fig. 5 (a) The influence of the surfactant and modified sea-water on the reduction of IFT. Proposed mechanism when Ca2+ and SO4 2− are active at (b) low and (c) high temperature (Haghighi and Firozjaii 2020)

Fathi et al. 2011). As soon as CTAB enters to the solution of SW-1NaCl, the contact angle is decreased, and the rock approaches the wet-water condition, but for SW-3KCl and SW-4Na2 SO4 , no uniform decrease or increase is observed. At very high concentrations of CTAB, SW-1NaCl has the most excellent effect on wetting the rock. This indicates that the CTAB optimal concentration is different depending on the salt concentration in the seawater. The solution’s contact angle increases by mixing with CTAB and SDS, and the minimum contact angle is related to 0.1CMC of SDS and 0.9 CMC of CTAB in smart water (Mofrad and Dehaghani 2020). Kharrat et al., by injection method of hybrid smart carbonated water, were able to achieve the highest recovery factor in EOR than the other methods (Soleimani et al.

1702

A. Rashidi and S. Sharafinia

Fig. 6 (a) The results of oil recovery factor and (b) differential pressure for smart brine and hybrid smart carbonated water (Soleimani et al. 2020)

2020). Oil recovery for smart water and hybrid smart carbonated water injection is 55% and 70%, respectively, as oil recovery amount for the hybrid smart water (15%) is higher than smart water due to smart and carbonated water dual impact in hybrid method (Fig. 6a and b). For reasons such as pH variation, wettability alteration, IFT reduction, and ion exchange in pore volumes injected less than 0.3, oil recovery for hybrid brine injection is less than smart water injection method. The surface of the carbonate core becomes negative through the aging of the core, resulting in the surface covered with oil. With smart water injection, we are seeing more change in the wettability because of injected negative ions; for example, SO4 is displaced by negative oil droplets on the surface, and cations are produced according to equations of 2–5 as follows: CaCO3 + CO2 + H2 O → Ca2+ + 2HCO− 3

(2)

H2 O + CO2 → H2 CO3

(3)

H2 CO3 → 2H+ + CO2− 3

(4)

H2 CO3 → H+ + HCO2− 3

(5)

Injection of smart carbonated water leads to the CO2 transfer from saltwater to oil. Because the solubility of CO2 in oil is higher than salt water, the oil is distributed by a negative charge on the rock surface. As a result, the oil is separated from the surface. Therefore, the change in wettability for smart water is lower than for smart carbonated water (Soleimani et al. 2020).

66 Smart Materials in Oil and Gas Industry: Application

1703

Exploration and mapping of subsurface crude oil reservoirs is a major global issue. But the information available from oil reservoirs is based on assumptions, fluid data, well pressure, wireline data, and so on. Currently, information about oil reservoirs is obtained using probes around boreholes. Since this information is just around a guess, they are not reliable (Yu 2012). Other techniques to obtaining information include the use of isotopes and artificial and natural tracers. However, in environments with abundant microbiotic activity, high temperatures and nonneutral pH are unstable (Sabir et al. 2000). The other nanoparticles used in the cementitious matrices are carbon quantum dots (CQDs). D. Wen et al. synthesized a new hydrocarbon sensor based on CQDs using a hydrothermal way. They used xylose precursor to produce the sensor and employed it to detect oil saturation in sandstone cores and reservoirs. They carried out the packed column experiments for investigating the effect of CQDs. Their results showed that CQDs displayed tracer-like breakthrough curves (BTC) characteristics, in other words, during the post-flood process, the transport effect first increases and then decreases rapidly until it approaches a relative concentration (C/C0). Also, CQDs showed a different performance than conventional NPs, meaning that when injected into the pores of the glass column (i.e., where particle agglomeration and retention were impossible with the presence of electrolytes), they quickly fill up and are completely removed from the column by washing process. Moreover, according to the results, the transport behavior of NPs is strongly affected by the presence of salt (Hwang et al. 2012). Varying the ionic strength of the salt has no significant effect on the change in the shape of the BTCs. Therefore, the CQDs displayed similar behavior to tracers. BTC for CQDs still acts like a tracer to about 60 ◦ C but starts to fluctuate when it reaches 80 ◦ C. Since the maximum C/C0 cannot reach 100%, when the temperature increases, the rate of retention also increases. A significant change in the shape of BTCs was observed by replacing calcite with glass beads. The CQD breakthrough for calcite column occurred at about 1 PV, which is the same as the glass bead packed columns (Hu et al. 2019).

Application of Smart Materials in Drilling Fluids The fluids used in drilling operations are called drilling fluids, which can be employed in circulation to balance the fluid formation pressure. Drilling fluid is a colloidal fluid that includes a combination of several phases of liquid and different solids used to facilitate and continue drilling operations of various formations. The reasonable drilling costs (about 80%) are too dependent on the drilling mud cost, so these prices can be controlled by employing suitable nano additives (Zamani et al. 2019). Nanotechnology provides a way to manufacture drilling mud with lower costs, highest efficiency, and stability. Nanofluids are materials that can be synthesized via adding small amounts of the nanoparticles (NPS) to a fluid. Also, these materials are called smart fluids because it has controllable properties. Therefore, their specific heat, thermal conductivity, viscosity, and density can be easily tuned toward the optimum levels (Zamani et al. 2019). The NPS are inorganic

1704

A. Rashidi and S. Sharafinia

and can be suspended in the liquids such as oil, and water. In addition they can be designed to be well-matched with reservoir fluids and are environmentally friendly. The main advantages of smart nanoparticles used in liquids are as follows (Zamani et al. 2019; Singh et al. 2010): 1. Better transmission via micro-channels 2. Decrease in erosion and not a requirement of pumping 3. Perform controllable reactions with other available materials Lee et al. studied a magnetically controllable drilling fluid for viscosity control using NPs. To achieve this purpose, they prepared magnetic NPs with three different sizes and two other bentonites such as water-based bentonite and pure bentonite and then have investigated their role in the complex fluids. They also used four different surfactants for particle stability in either paraffin oil or water (Amanullah et al. 2011). They obtained the following results: 1. The fluid viscosity is enhanced when increasing magnetic NPs and bentonite. 2. The fluid viscosity is decreased by adding a surfactant. 3. The use of microparticles (iron oxide with a size of 3 nm) increases the viscosity more than nanoparticles. It can be related to the microparticle’s ability to disperse more effectively than the nanoparticles. Wu et al. synthesized smart water-based drilling fluids (WDFs) with temperatureresponsive property via graft between bentonite and dual-functionalized cellulose nanocrystals (fCNCs). They observed that the thermocontrollable property of fCNCs was too related to the molar ratio of poly(N-isopropylacrylamide) (PNIPAM) and poly (2-acrylamide-2-methyl-1-propane sulfonic acid) (PAMPS) monomers (Kong and Ohadi 2010). As a result of the presence of PAMPS in the fCNCs structure, a strong bond is formed between BT platelets and fCNCs due to the negatively charged sulfonate and amide groups on the PAMPS. Due to the presence of negatively charged sulfonate and amide groups on PAMPS in the fCNCs structure, a strong bond is formed between BT platelets and fCNCs. These negative groups lead to a bond between the bentonite and fCNCs. The BT/fCNC clusters are formed at high temperatures through thermo responsive PNIPAM grafts and led to thermothickening rheological performance of BT/fCNC-WDFs. The synthesized smart BT/fCNC-WDFs with good cyclability, in situ rheology controllability, and high sustainability have excellent efficiency in the drilling industry. Vipulanandan et al. used the hydrophilic bentonite-based mud to evade the drilling mud fluid loss, raise the sensing electrical properties, and modify the rheological efficiency (Vipulanandan et al. 2018). The particle size range of drilling mud is 12–20 nm and contained 2–8 W% water at a temperature range between 25 and 85 ◦ C. According to the results obtained, electrical resistivity was recognized as an external stimulus to give smart properties to drilling mud, and they also observed that:

66 Smart Materials in Oil and Gas Industry: Application

1705

1. When increasing nanoclay content, temperature, and bentonite content, the electrical resistance was decreased. 2. The yield stress of drilling mud was enhanced (2 Pa to 31 Pa) by enhancing the bentonite content (2% to 8%). It can be said that the yield stress is a controlling parameter for the fluid loss because addition of 1% nanoclay to the drilling mud enhanced the yield stress and the maximum shear stress tolerance is observed with bentonite content and temperature. 3. The maximum shear stress of the drilling muds was decreased by increasing the temperature, but it increased with the bentonite content (Vipulanandan et al. 2018). Vryzas et al. prepared the controllable, magnetic drilling fluid. This fluid is formed of custom-made iron oxide magnetic nanoparticles (CM Fe3 O4 NPs) that, by applying a magnetic field, lead to control of yield stress and viscosity in the drilling fluid (Vryzas et al. 2017). The X-ray diffraction (XRD) pattern and transmission electron microscope (TEM) of the CM Fe3 O4 NPs are shown in Fig. 7a and b, respectively. Figure 7a confirms that the peaks corresponding to Fe3 O4 NPs have no impurities and according to Fig. 7b the diameter of Fe3 O4 NPs is in the range of 6–8 nm. When the CM Fe3 O4 NP range is 0.5–1 wt%, shear rates increased about 10–30% due to the NPs optical dispersion. The Herschel- Bulkley (HB) rheological model is the best model to describe these examples. In another study, macro and micro type fluid additives have been challenged due to inadequate thermal and environmental, chemical, mechanical, and physical characteristics by M. D Amanullah et al. (2011) They used the nanomaterials for the preparation of smart drilling fluids. They studied the mud cake and solids control, filtration and rheological properties, and stability of these materials. The results are as follows: 1. The micro and nanomaterials have a higher surface area and lowest imperfection than macro material and as a result lower thermal, physical, chemical, and mechanical properties.

Fig. 7 (a) X-ray diffraction (XRD) pattern and (b) transmission electron microscope (TEM) of the CM Fe3 O4 NPs (Soleimani et al. 2020)

1706

A. Rashidi and S. Sharafinia

2. The nanofluid’s efficiency has been increased due to the concatenation of functional groups of the nanomaterial. 3. The use of nanomaterials also reduces adhesion to the differential pipe and increases the penetration rate in rock structures. Husein and Hareland (Zakaria et al. 2012) studied the filtration and stable rheology properties of different smart drilling fluids under high-temperature and high-pressure conditions. The novel smart drilling fluid has been expanding in the Li et al. work (Li et al. 2016). The smart drilling fluid was prepared by using materials such as nano-silica, XC- polymer, bentonite, and KCl. The properties of these materials were investigated using filtration and rheological experiments. According to the results obtained the rheology properties were better and the filtration was decreased by the addition of nanomaterials. Dimitrios et al., to increase the hydraulic properties of drilling processes, prepared drilling fluids using Fe3 O4 nanoparticles (Fe3 O4 NPs) (Gerogiorgis et al. 2015). They used the Herschel-Bulkley rheological model to compare the correlation shear rate and shear stress of experimental data. The parameters of Fe3 O4 NPs concentration, shear rate rheological effects, and temperature are described by viscosity correlation and shear stress rheological effects. Increasing the temperature enhances the viscosity error and shear stress. It also affects fluid-structure and rheology.

Application of Smart Materials in Oil Well Cement The important reason for failure in some accidents for example, Macondo operations, is due to a defect in the cementing job. This is confirming the pressure accumulation in places under high stress (Christou and Konstantinidou 2012). One of the essential issues in the oil and gas industry is well cementing because it provides proper conditions for water, oil, and gas release from the wellbore as well as prevents pollution and corrosion in hydrocarbons production. In general, the main purpose of well-cementing consists of the following: 1. The primary cementing includes providing a hydraulic seal, protecting useable water, making of zonal isolation, preparing structural support for the casing, protecting casing for correlation, and isolating coverage seat for later drilling (Christou and Konstantinidou 2012) 2. Remedial cementing such as squeeze cementing and plug cementing (Christou and Konstantinidou 2012) In the oil and gas industry, some mistakes such as formulation errors, lead to explosions on oil rigs. This is usually due to the poor performance of the materials used to make the cement. They are of lower tensile strength than the polymers and metals (Han et al. 2011). One of the new ways to avoid these problems in the oil and gas industry is the use of smart materials. Cement materials with the ability to cause

66 Smart Materials in Oil and Gas Industry: Application

1707

structural and chemical changes in the face of external stimuli such as magnetic and electric fields, light, chemical composition, and temperature, smart cement or cement responsive to external stimulants are called. To prepare smart-cement, microorganisms, inorganic nanoparticles, and natural and synthetic polymers as admixtures are added in the cement formulation (Bogue 2012). Nanoparticles have significantly been used than micro and macro particles because they have a high surface area, thermal and electrical properties, and excellent chemical activity. Therefore, the addition of nanofibers and NPs to cement matrices improves their electrical conductivity. In D. M Abdullah et al.’s work (Gao et al. 2009) oil well cement was modified by replacing these particles with ZnO NPs. The results showed the particle size of the synthesized powder was about 100 nm. Also, it improves the structural properties by the development of the hydration phase of calcium silicate. Reducing free water increases compressive strength and density, thus giving it a smart property. The nano powders lead to reduction of porosity by filling in microstructures (Isaia et al. 2003). Addition of 1% cement dust/ZnO, ZnO, and cement dust at 380 ◦ C increases the compressive strength by 12.6%, 9.96%, and 15.2%, respectively. The reason for these differences in the mortars is related to the pozzolanic reaction between the cement and the nano powder. Since the CD and ZnO are able to fill the space between the types of cement and thus produce smaller pores, they are more affected by pozzolanic reactions because these reactions are related to the amount of surface available (Gao et al. 2009). Ch. Ruttanapun et al. (Phrompet et al. 2019) improved electrochemical, antibacterial, thermal conductivity, dielectric constant, and micro-hardness properties and developed graphene oxide-nanosized C3 AH6 cement nanocomposite via hydration method. To determine the micro-hardness, the samples of rGO-C3AH6 were prepared with different percentages of rGO and determined according to Eq. 6 (Kamalak et al. 2016).   HV = 1854 P/d2

(6)

The parameters of HV, P, and d indicate Vickers micro-hardness value, the applied load in kg, and the average diagonal length of the impression in mm, respectively. They observed that the amount of HV enhances with increase of rGO percent. Also, the dielectric constant, for example, of C3 AH6 is less than the sample of x%rGO-C3 AH6 , and the smaller the particle size, the larger this value. The experiment results of x%rGO-C3 AH6 samples on gram-negative bacteria showed that cement of C3 AH6 is ineffective in killing bacteria, but with the loading of nanoparticles rGO on C3 AH6 , the antibacterial activity increased because the sample of x%rGO-C3 AH6 releases phospholipids from the membrane (Kamalak et al. 2016). The study on smart cement with different percentages of water to oil was performed by A. Mohammed et al. (API, R65 2002), which used a conductive filler (0.1%) to create smart properties at different temperatures. The apparent viscosity

1708

A. Rashidi and S. Sharafinia

was increased by increasing temperature (85 ◦ C) and proportion of water/oil. Also, values of the shear stress and yield stress were increased by increasing temperature, but these were decreased by increasing the ratio of water to oil. One of the properties of cementitious materials is electrical conductivity affected by their micro and nanobehaviors. For improving the electrical conductivity, addition of nanoparticles into the cementitious matrices is the best way. One of the promising materials is carbon nanotubes (CNTs), because of their piezoresistive characteristics changing the electrical resistivity upon strain. The amount of CNTs introduced into cementitious matrices is very important. When the amount of CNTs in the composite matrix reaches a critical fraction, percolation starts and the material becomes a conductor. In order to achieve a conductor, it is essential to add sufficient CNTs into the matrix. It changes the resistance and reactance of strain in matrix, and the information obtained can be exploited for the purpose of structural diagnosis. A cement matrix also leads to easy embedding of sensors in concrete elements and conversion of structures into distributed sensor network systems (Chung and Wang 2003).

Application of Smart Materials in Oil-in-Water and Water-in-Oil Separation With growing population and modern industry, water pollution has created more concerns and adverse effects to public health and the environment. Among the various pollutants, oil contaminants have obtained abundant consideration in recent years and have become a great part of water pollution due to their extensive applications in the industry. Therefore, they must be removed from water sources. In the process of oil removal from waters, a variety of methods have been proposed, including gravity biological oxidation, coagulation, separation, flotation, adsorption, and centrifugal separation (Wang et al. 2017). Among these methods, the adsorption method has been widely studied as an excellent method for removing pollutants of oil from water due to its easy studying, high performance, high capacity, and low cost. Recently, smart materials have been raised as a suitable and effective way for preparing materials with different diameters, surface areas, and high porosity with high efficiency (Cai et al. 2014). In the following, we will discuss these materials in the oil-in-water and water-in-oil separation industry. Smart materials used in oil-in-water and water-in-oil separation significantly are controllable and employ different methods to synthesize materials with a single super-wetting surface that can intelligently separate oil-in-water and water-in-oil. Compared to simple oilwater mixtures, treatment of oil-in-water and oil-in-oil emulsions has become a major challenge (Kwon et al. 2012). Membrane separation is an excellent way to treat these emulsions. These membranes are in a hydrophilic state or hydrophobic state. When they are hydrophilic, water droplets penetrate the pores and they do not allow it to enter the oil by blocking the membrane cavities. Also, the opposite is true of hydrophobic membranes (Cai et al. 2014; Chen and Xu 2013). Therefore, with the development of membranes with the wettability switchable, the high investment

66 Smart Materials in Oil and Gas Industry: Application

1709

costs and consumption of materials to separate this emulsion can be significantly reduced. A membrane that has the ability to change the wettability property in response to external stimuli such as electrical energies, light, pH, and temperature is called an intelligent response membrane. These membranes play a very important role in separating both oil-in-water and water-in-oil emulsions (Xue et al. 2013). In general, those membranes that have ionizable functional groups are responsive to pH, and can by controlling pH affect the wettability. Among the materials with ionizable functional groups can refer to poly (acrylic acid) (PAA) and carboxylic. At acidic pH, these materials cause oleophobic properties on the surface through the formation of hydrogen bonds. Li et al. used poly (4-vinylpyridine) (PMMA-bP4VP) and poly (methyl methacrylate) to fabricate smart membranes and observed that in neutral environments membrane adhesion increases due to PMMA lipophilic behavior (Xue et al. 2013). However, this membrane showed oleophobic behavior because the small amount of water was trapped in its cavities. In the oil-water separation process, due to its superhydrophobic/superoleophilic surface of this membrane, the oil penetrated the membrane structure while retaining water in its structure. As indicated in Fig. 8a, the opposite result was obtained when acidic water (pH = 3) was used to treat this membrane. The important point is that underwater, the conversion between superoleophilicity and superoleophobicity and the conversion between superhydrophobicity and superhydrophilicity is continuously repeated (see Fig. 8c and d). The coating of fluorine-free superwetting was also synthesized by Wang et al., which showed an efficiency of about 99.9% for the separation of oil/ water. The connection to the porous substrate was made through polyethylene (Xue et al. 2013). In solutions where the pH changes, wettability and surface energy are affected by protonated N. Because the protonated N increases in an acidic environment, the membrane is hydrophilic and, conversely, exhibits hydrophobic behavior in the basic medium (Fig. 8b) (Xue et al. 2013). A new surface with temperature-responsive properties is obtained by bonding between the membrane surface and thermo-sensitive material, which leads to a wettability behavior. The low critical solution temperature Poly (Nisopropylacrylamide) (PNIPAAm) is about 33 ◦ C (Fig. 9a) (Xue et al. 2013). Xin et al. prepared a temperature-responsive membrane via bonding of this polymer with the surface of regenerated cellulose. At temperatures below LCST the PNIPAAm is hydrophilic due to the hydrogen bond formation. Still, it indicated hydrophobicity in temperatures higher LCST (Fig. 9a) leading to selectively absorb water or oil at different temperatures, as shown in Fig. 9b and c. According to the Fig. 9d Lu et al. synthesized a smart membrane with thermal- responsive property using polymethyl methacrylate (PMMA) and PNIPAAm, which has superhydrophilicity at temperatures below LCST (Fig. 9e–g), but it shows hydrophobicity and oleophilicity at temperatures above LCST (Fig. 9h and i) (Wang et al. 2015). The wettability of this membrane is also unchanged for several cycles (Fig. 9j). The materials such as TiO2 and ZnO (metal oxides) are photo-sensitive due to their photocatalytic properties and are used in the preparation of smart membranes. These membranes are stimulated by UV and visible light and act reversibly. For this purpose, Bhushan et al. have synthesized a UV-sensitive membrane (Li et al.

1710

A. Rashidi and S. Sharafinia

Fig. 8 (a) Process of the preparation of external stimulus responsive membrane, (b) the mechanism of wettability conversion, (c) the wettability conversion between superhydrophilicity and superhydrophobicity of the membrane, and (d) the wettability conversion between superoleophilicity and underwater superoleophobicity of the membrane (Xue et al. 2013)

2019). The low surface energy of the 1H,1H,2H,2H- Perfluorodecyltrimethoxysilane (PFDMS) used in this coating leads to water repellency (Fig. 10a). As the UV radiation increases, the contact angle of the oil is increased but the contact angle of the water is decreased, respectively. The wettability would change after 20 minutes of the UV radiation and can be recovered by heat treatment (Fig. 10d and e). Also, many other organic compounds, such as the N = N of azo, break down into cistrans isomers under ultraviolet irradiation, as a result of altering their wettability properties (Fig. 10b). Feng et al. studied the formation of a coating of aminoazobenzene and nano-Ag pine needles on a metal network. The coordination bonds caused that silver particles to aminoazobenzene. Aminoazobenzene molecules have improved wettability via the reduction of steric hindrance (Coyle et al. 2004). The hydrophobicity of this mesh is about 150.0 and in the ultraviolet region (365 nm) it showed hydrophilicity equivalent to 10.0 degrees (Fig. 10f). Also, the speed of this mesh for separating water and oil is 4 × 105 L m−2 h−1 (Fig. 10g and h). The next case is related to the electricity-responsive membranes, which separate water and oil in a much shorter time due to their high rate of wettability transformation. The electrical potential leads to chemical change of the surface by an oxidation-reduction reaction and a change in the amount of wettability. Tian et al. prepared a superhydrophobic electrical-response membrane (Xue et al. 2014). The reason for its hydrophobicity is related to hydrostatic pressure (Youngblood and McCarthy 1999). At high voltages, an electric capillary pressure is produced (Prins et al. 2001), and if this pressure exceeds the hydrostatic pressure the membrane exhibits hydrophilic behavior (Fig. 11a–i). Li et al. synthesized a poly3-methylthiophene (P (3-MTH)) membrane for separating oil and water mixtures (Li et al. 2015b). They controlled surface wettability by controlling the voltage to

66 Smart Materials in Oil and Gas Industry: Application

1711

Fig. 9 (a) Membrane wettability conversion process. (b and c) The adsorption of water and oil at different temperatures by membrane. (d) The preparation and wettability change of the membrane. (e and f) Wetting of water and oil droplets at the temperature below the LCST. (g) Spread water drops on the membrane. (h and i) Wetting of water and oil droplets at the temperature above the LCST (Xue et al. 2013)

control P (3-MTH) redox. The presence of P (3- MTH) in the membrane structure causes its hydrophobicity. The higher the voltage, the more ClO4 is produced, which leads to the formation of hydrophilic dipoles due to the presence of impurities (Fig. 11j–l). When the voltage reached above 0.8, the wettability changed from hydrophobic to hydrophilic and was repeated several times (Fig. 11m and n). Xi et al. by co-depositing TA/DETA on the PPMM prepared a wettability switchable membrane. This membrane initially had a hydrophilic behavior but after treatment with ethanol showed excellent hydrophobicity. The efficiency of these membranes to the separation of the water-in-water emulsions in the hydrophobic state and the oil-in-water emulsions in the hydrophilic state was excellent (> 98%) (Zhang et al. 2020). External stimuli lead to ionization of the surface functional groups and thus switch the wettability of the surface. Membranes with carboxylic groups such as nylon membranes are hydrophilic but exhibit hydrophobic behavior

Fig. 10 (a) Mechanism of the wettability conversion of the membrane. (b) The effects of UV irradiation on the breaking, rotation, and regeneration of the bond N = N. (c) The wettability of membrane under ultraviolet irradiation and heat. (d and e) Changes of water contact angle. (d) Underwater oil contact angle and (e) after ultraviolet irradiation and heat treatment. (f) Cycles of wettability conversion. (g and h) Separation of oil-water mixture after ultraviolet irradiation. (g and h) Visible light irradiation (Xue et al. 2013)

1712 A. Rashidi and S. Sharafinia

66 Smart Materials in Oil and Gas Industry: Application

1713

Fig. 11 (a–c) Water contact angle images of the membrane with different voltage. (d–f) Photographs of the underwater oil contact angle of the prepared membrane after different voltage treatments. (g–i) Membrane wettability schematics and the analogous enlarged schematic of the liquid-solid contact at different voltages, respectively. (j) Schematic of wettability alteration between doped and undoped after electrical treatment. (k and l) The water contact angle of membrane in before and after electrical treatment. (m) The water contact angle with changing voltage. (n) Cycles of wettability alteration between hydrophobic and hydrophilic (Xue et al. 2013)

in acidic environments due to the de-protonation and protonation process (Cheng et al. 2017). Hydrophobicity of membrane is related to the increase in free surface energy resulting from de-protonated functional groups (Cheng et al. 2017; Cai et al. 2018). On other hand, the hydrophobic surface is created due to the collapse of structures on the surface by the protonated functional groups (Li et al. 2018b). Deprotonated TA shows a higher surface wettability than protonated TA because it reduces hydrogen bonds between DETA and TA in TA/DETA networks. They used molecular simulation to calculate the Gibbs free energy change (G) in the TA ionization process in the presence of DETA and considered the solvent as ethanol or water. According to Fig. 12, the G for ethanol (G > 0) is higher than water (G < 0), so TA is not ionized in ethanol, but first-order ionization occurs in water. As a result, it can be said that the switchable wettability of the TA/DETAfunctionalized PPMM is changed because TA is protonated in ethanol but becomes de-protonation in water (Zhang et al. 2020). Xu et al. prepared a polypropylene membrane with switchable surface wettability using diethylenetriamine and diethylene triamine co-deposition methods. These membrane surfaces are hydrophilic and hydrophobic before and after treatment by ethanol, respectively, and in other words, they have a translation hydrophilicityhydrophobicity (Zhang et al. 2020). The prepared membrane showed 98% efficiency in separating both oil-in-water and water-in-oil emulsions. Qu et al. synthesized superwettable smart materials that are pH-responsive to oil/water separation. This smart material can by chemical reaction separate oil and water, and it also displays chemical stability in difficult conditions. Less surface energy and roughness are two important parameters to achieve the surface of superhydrophobic. They synthesized a rough surface with superhydrophobic property for water repellency through AIBN polymerization in a mixture of methyl methacrylate

1714

A. Rashidi and S. Sharafinia

Fig. 12 Results of B3 LYP method on ionization TA in ethanol or water solvents (Zhang et al. 2020)

(MMA) and 3-(Trimethoxysilyl)propyl acrylate (KH570) and succeeded in synthesizing a pH-responsive copolymer. In the first stage, the silica surface was modified by aminopropyltriethoxysilane (AMEO). Since AMEO has a large number of amino groups, it can be easily bonded with carboxyl copolymer groups. It also has the ability to switch between superhydrophilicity and superhydrophobicity states at different pH, as shown in Fig. 13 (Li et al. 2007). Deng et al. synthesized superhydrophobic, durable, and magnetic carbon sponges with high BET surface area (309.7 m2 /g), mechanical durability, and chemical stability, excellent chemical resistance, high oil/water separation ability, and high superhydrophobicity/superoleophilicity. Since the carbon sponges are able to selectively control adsorption of oils, due to the presence of magnetic particles on the sponge, they are therefore called smart sorbents. In order to study mechanical, environmental, and chemical stability of superhydrophobic carbon sponges, this sorbent is tested in different conditions, for example, the superhydrophobic carbon chemical stability was examined in basic solutions, acidic solutions, and salt with pH = 7, 1, and 13 respectively. Fig. 14a–e show the hydrophobicity of adsorbent in acidic, alkaline, and salt media has changed slightly than to pure water. thus confirming the excellent stability of this adsorben. Ahmed et al. have reported a switchable smart sorbent manufacture by a very fast, easy method of modifying different substrates. This method involves increasing the interaction of melamine foam with a substrate and subsequently by poly(2-vinylpyridine-b-dimethylsiloxane). In fact, PDA is used for bonding between melamine copolymer poly(2-vinylpyridine-b-dimethylsiloxane) and foam. The stability of poly(2-vinylpyridine-b-dimethylsiloxane) on the PDA-MF sponge surface is provided through the formation of π − π and hydrogen bonding between PDA and MF and subsequently the formation of a π-conjugated bond between PDA-MF sponge and poly(2-vinylpyridine-b-dimethylsiloxane) to fabricate P2VP-b-PDMSPDA-MF. The modified materials used in this method have excellent efficiency in

66 Smart Materials in Oil and Gas Industry: Application

1715

Fig. 13 Schematic image of preparation and application of the pH-responsive smart fabric in the oil/water separation (Qu et al. 2020)

the oil removal from water at various pH and they are also very stable under different conditions including alkaline and acidic (Ong et al. 2019). This method includes dopamine oxidation on the material surface and then functionalizing by poly(2vinylpyridine-b-dimethylsiloxane), and the materials are pH-prepositive. In order to study sorption and desorption of oil by the smart sorbent prepared, toluene and

1716

A. Rashidi and S. Sharafinia

Fig. 14 Photographs of (a) the water droplets (Rhodamine B dye) and (b) the aqueous droplets of HCl (lacmus dye, pH 1), NaCl (0.1 M, pH 7), and NaOH solutions (lacmus dye, pH 13) on the surface of an MSCS. Photographs of the dyed water droplets on the MSCS after (c) heating at 200 ◦ C for 2 h, (d) freezing at −196 ◦ C for 2 h, and (e) UV irradiation (λ = 365 nm, 500 W) for 24 h. Insets: the related measurements of the water contact angle. (f) The water contact angle of the MSCS after abrasion for 50 times by sandpaper. Insets: image of the dyed water droplets on the surface and its water contact angle after abrasion (50 times) (Wang and Deng 2019)

66 Smart Materials in Oil and Gas Industry: Application

1717

Fig. 15 Adsorption and desorption of oil by P2VP-b-PDMS-PDA-MF at pH = 2 and 7 (Ong et al. 2019)

1,2-dichloroethane removal from water were tested at pH = 7, and the desorption at pH = 2 using the formerly oil-loaded oil smart sorbent. In water with pH = 2, all chains of P2VP were protonated, and as a result, exterior surface became hydrophilic and underwater superoleophobic (see Fig. 15) (Zhang et al. 2012). Wang et al. prepared carbon blackcoated membrane adsorbent with suitable chemical composition and a hierarchical surface structure and ideal underoil superhydrophobic and underwater superoleophobic smart properties and tested them in oil-water separation. Results reveal that carbon black-coated membrane enable sufficiently to separate the multiphase emulsion mixture, esatablishing an oil area. By this way, they were able to overcome problems such as the blocked liquid layer that occurs when mixing multi-phase emulsions with traditional membranes (Cao et al. 2019). The fabrication of a smart sorbent with self-cleaning property, pH-responsive, and superwettable properties using a mild and convenient method has been studied by M. Qu et al. Among the important factors for preparing a superwettable surface are low surface energy and roughness (Li et al. 2007). As shown in Fig. 16, a combination of ethylene base [3 (trimethoxycylcyl) propylene], PFOA, and kaolin is used to fabricate a highly pH-responsive and smart surface. Kaolin particles are interconnected together by bis[3-(trimethoxysilyl) propyl]ethylene binders. In the presence of these binders, kaolin particles can be compressed. The bis [3(trimethoxysilyl) propyl] ethylene diamine groups are affected by the amidation reaction with PFOA due to their high activity, and thus lead to the formation of fluorocarbon chains. These chains have low surface tension and easily bond

1718

A. Rashidi and S. Sharafinia

Fig. 16 The pH-responsive wettability of the superwettable material surface (Qu et al. 2019)

with the surface of kaolin particles, resulting in a superamphiphobic surface that has oil and water repellency properties. This smart sorbent is dependent on pH due to having a rough surface with specific chemical composition and therefore can vary reversibly between superoleophobicity-superhydrophilicity and superamphiphobicity for many times The obtained smart sorbent can be used for different types of separation such as oil/water separation due to having an effective and sequential separation. Also, smart materials reported in this work are very stable than the mechanical abrasion and solutions containing salt. These superamphiphobic materials are more useful and better than other materials due to their pH-responsive properties in the oil/water separation process (Qu et al. 2019). To show superhydrophilicity properties in the oil/water separation process, nhexane/water mixture with ratio of 1:1 at pH = 7 was passed over this material and it was observed that water solution passed through these superwettable materials, and therefore, the separation of n-hexane/water mixture was carried out successfully. Also, the pH-responsive properties in the n-hexane/water mixture separation process with ratio of 1:1 at pH = 7 were displayed by adding water solution at pH = 13 into the mixture, and they found that the pH of the passed solution through these materials was equal to 12.7.

66 Smart Materials in Oil and Gas Industry: Application

1719

Fig. 17 Free radical polymerization process to modify cotton fabric by DMA (Liang et al. 2019)

X. Meng et al., investigated on the oil/ water separation by a new cotton fabric with interesting features such as switchable wettability and CO2 responsive. During the manufacturing process, the cotton fabric was polymerizedwith 2 (dimethylamino) ethyl methacrylate and then CO2/ N2 alternation. this investigation has successfully demonstrated that oil/water separation can use for control of CO2 . Also, the modified fabric by N2 has wettability, and therefore they are useful in the smart separation of oil and water. Figure 17 represents the protonation process, and during this process, the hydrophilic and hydrophobic transfer behavior of CO2 - released from the modified fabric using periodic CO2 addition and removal of deprotonate and protonate amines was reversible and repeatable. Because the modified fabric with 2-(dimethylamino) ethyl methacrylate (DMAEMA) polymers can be converted from a hydrophobic to a hydrophilic state by stimulating CO2 , bicarbonate was produced due to the reaction of amine with CO2 in the aqueous medium (Liang et al. 2019). X. Meng et al., investigated on the oil/ water separation by a new cotton fabric with interesting features such as switchable wettability and CO2 responsive. During the manufacturing process, the cotton fabric was polymerizedwith 2 (dimethylamino) ethyl methacrylate and then CO2/ N2 alternation. this investigation has successfully demonstrated that oil/water separation can use for control of CO2. Shami et al. prepared multi-purpose and PH-responsive smart nanofibers. They used styrene−acrylonitrile (SAN) copolymer by electrospinning process and then treatment by NaOH which can well switch between superoleophobic and superhydrophobic states (Shami et al. 2019). Figure 18 shows the fabrication method and production mechanism of these nanofibers. By pouring the neutral water on the NaOH-treated mesh, a hydration layer was formed via COO− groups interaction on the mesh surface with the water. Since these groups are hydrophilic, they easily form a bond. This layer prevents oil contact with the mesh, and results in the fabrication of the superoleophobicity and on the other hand increases the superhydrophilicity of the mesh. Also, the states of

Fig. 18 Fabrication process of the SAN mesh, treatment with by NaOH and switching between oil-removing and water-removing wetting states at different pH (Shami et al. 2019)

1720 A. Rashidi and S. Sharafinia

66 Smart Materials in Oil and Gas Industry: Application

1721

superhydrophobic and superoleophilic have been confirmed by treatment of the acid of mesh with neutral water. This means that due to their contact, the contact angle of the oil decreases (0◦ ) and, conversely, the water contact angle increases to 159.3◦ . Therefore, the carboxylate and carboxylic groups form the intermolecular and the intramolecular hydrogen bonds with the water, respectively, and the water contact angle increases to 159.3◦ . The carboxylic group ionization state of the poly (acrylic acid) leads to a reversible structural transition from the extended coil to the collapsed ball at pH 4.5: RCOOH  H+ + RCOO−

(7)

and the amount of neutral water absorbed on the mesh is gradually increased by the increase in pH to 7. Therefore, the wettability is dependent on pH (Shami et al. 2019). The surface ability to control the material wettability in response to external changes is called the wettability control. The surface wettability of the material undergoes a transition between oleophobic-oleophilic and hydrophobic-hydrophilic, which leads to regulatable oil-water separation (Li et al. 2015b). Wang et al. using the electrochemical deposition method produced a pHresponsive superhydrophobic copper network. This copper network represented various properties such as superhydrophilicity to alkaline water droplets and showed superoleophobic and superhydrophobic properties to the acidic and neutral water droplets (Wang and Guo 2013). In another study by Cheng et al., a pH-responsive reversible wettable biomass was prepared using acrylamide and acrylic acid bond with eucalyptus paste cellulose. By conversion of the pH from 1 to 9, the hydrophilic oleophobic cellulose-g-PAM (cellulose-g-pam) and the hydrophobic lipophilic cellulose-gPAA (ellulose-g-paa) were changed into hydrophobic lipophilicity and oleophobic oleophobicity, respectively (Cheng et al. 2019). Gao et al. discovered that ZnO, TiO2 , and other materials have response to UV light properties. so, using of a responsive to UV light inorganic material can generate a smart compound which can be applied for water/oil separation. Therefore, they used a hydrothermal method to create a bond between TiO2 NPS with the copper mesh surface and finally to produce a superoleophobic- superhydrophobic copper mesh using octadecylphosphoric acid. In addition, electric field and temperature external stimuli can also reason a change in wettability. A controlled surface can be achieved by connection between the surface and a thermo-responsive material such as poly (N-isopropylacrylamide) (PNIPAM) (Gao et al. 2013). Zheng et al. by polymerization process have created a bond between polyaniline and the steel mesh surface to construct polyaniline. When the voltage has changed from 0 to 160 V, the polyaniline keeps on superoleophobic while the water droplets reduce contact angle from 146◦ to 70◦ . This confirms the ability of the smart wettability and superoleophobic property under the influence of an electric field (Xue et al. 2014). Some of the dual superwettability and smart materials are reported in Table 2.

1722

A. Rashidi and S. Sharafinia

Table 2 Dual stimuli-responsive super-wettable materials Materials Cellulose-PDMAEMA P(NIPAAm-co-AAc) thin films (PNIPAAm PAAc) Fe3 O4 @SiO2 /PNIPAM PNIPAM/PDMAEMA grafted PFOTS modified Al2 O3 DMAA-grafted Si surface PDMAEMA hydrogel coated mesh

Stimulus – –

Ref Zhao et al. (2018) Xia et al. (2007)

Magnetic and Thermo (Temperature) pH and Electrolyte

Chen et al. (2014) Liu et al. (2010)

Light and Temperature pH and Temperature

Moreno et al. (2019) Gao et al. (2013)

Conclusion Potential applications of smart materials in the oil and gas industry were studied. According to their properties, they are used in different fields of oil and gas industry. These materials are used as nanotracer and smart water for crude oil exploration, enhanced oil recovery, and nano additives in drilling fluids. According to the observed results, the quantum dots are the best candidate to produce the sensor and detect oil saturation in sandstone cores and reservoirs. Quantum dots can be utilized as a significant sensor for hydrocarbon compounds. As sensors, the CQDs can easily detect the oil saturation in the matrix core and reservoirs. In the contrary of pure brine matrix, the existence of oil would affect the incivility of quantum dots. It can be concluded that nowadays improved smart cements with nanostructures have attracted the attention of researchers due to advantages such as easy measurement of pressure and stress, increased flexural strength and ductility, and cost-effectiveness. The traditional methods used in oil/water separation process are usually non-recyclable and nonselective that can be coming with high costs. Recently, smart materials have been able to challenge traditional sorbents because they have properties such as selective adsorption, efficient desorption, and high sensitivity and can easily respond to external stimuli pH, temperature, light, and electricity. Also, the wettability membranes can separate the water-in-oil and oil-inwater emulsion with very high speed and efficiency, even despite the long distance between the actual industry and laboratory, these membranes have an excellent ability to separate these emulsions.

References Ahmadi S, Hosseini M, Tangestani E, Mousavi SE, Niazi M (2020) Wettability alteration and oil recovery by spontaneous imbibition of smart water and surfactants into carbonates. Pet Sci:17(3):712–721

66 Smart Materials in Oil and Gas Industry: Application

1723

Amanullah M, AlArfaj MK, Al-abdullatif ZA (2011) Preliminary test results of nano-based drilling fluids for oil and gas field application. In: SPE/IADC drilling conference and exhibition. Society of Petroleum Engineers API, R65 (2002) Cementing shallow water flow zones in deepwater wells. API, Washington, DC; (b) Vipulanandan, C.; Mohammed, A., Smart cement rheological and piezoresistive behavior for oil well applications. Journal of petroleum Science and Engineering 2015, 135, 50–58 Barry MM, Jung Y, Lee J-K, Phuoc TX, Chyu MK (2015) Fluid filtration and rheological properties of nanoparticle additive and intercalated clay hybrid bentonite drilling fluids. J Pet Sci Eng 127:338–346; (b) Vryzas, Z.; Nalbandian, L.; Zaspalis, V. T.; Kelessidis, V. C., How different nanoparticles affect the rheological properties of aqueous Wyoming sodium bentonite suspensions. Journal of Petroleum Science and Engineering 2019, 173, 941–954; (c) Cai, J.; Chenevert, M. E.; Sharma, M. M.; Friedheim, J. E., Decreasing water invasion into Atoka shale using nonmodified silica nanoparticles. SPE Drilling & Completion 2012, 27 (1), 103– 112; (d) William, J. K. M.; Ponmani, S.; Samuel, R.; Nagarajan, R.; Sangwai, J. S., Effect of CuO and ZnO nanofluids in xanthan gum on thermal, electrical and high pressure rheology of water-based drilling fluids. Journal of Petroleum Science and Engineering 2014, 117, 15–27; (e) Huang, X.; Lv, K.; Sun, J.; Lu, Z.; Bai, Y.; Shen, H.; Wang, J., Enhancement of thermal stability of drilling fluid using laponite nanoparticles under extreme temperature conditions. Materials Letters 2019, 248, 146–149; (f) Wang, K.; Jiang, G.; Liu, F.; Yang, L.; Ni, X.; Wang, J., Magnesium aluminum silicate nanoparticles as a high-performance rheological modifier in water-based drilling fluids. Applied Clay Science 2018, 161, 427–435; (g) Meng, X.; Zhang, Y.; Zhou, F.; Chu, P. K., Effects of carbon ash on rheological properties of water-based drilling fluids. Journal of Petroleum Science and Engineering 2012, 100, 1–8; (h) Ismail, A.; Aftab, A.; Ibupoto, Z.; Zolkifile, N., The novel approach for the enhancement of rheological properties of water-based drilling fluids by using multi-walled carbon nanotube, nanosilica and glass beads. Journal of Petroleum Science and Engineering 2016, 139, 264–275 Biscay F, Ghoufi A, Malfreyt P (2011) Surface tension of water–alcohol mixtures from Monte Carlo simulations. J Chem Phys 134(4):044709; (b) Russo, N.; Anastassopoulou, J.; Barone, G., Properties and Chemistry of Biomolecular Systems: Proceedings of the Second Joint GreekItalian Meeting on Chemistry and Biological Systems and Molecular Chemical Engineering, Cetraro, Italy, October 1992. Springer Science & Business Media: 2012; Vol. 11 Bogue R (2012) Smart materials: a review of recent developments. Assem Autom; (b) Chung, D., Piezoresistive cement-based materials for strain sensing. Journal of Intelligent Material Systems and Structures 2002, 13 (9), 599–609 Cai Y, Lin L, Xue Z, Liu M, Wang S, Jiang L (2014) Filefish-inspired surface design for anisotropic underwater oleophobicity. Adv Funct Mater 24(6):809–816; (b) Zhang, F.; Zhang, W. B.; Shi, Z.; Wang, D.; Jin, J.; Jiang, L., Nanowire-haired inorganic membranes with superhydrophilicity and underwater ultralow adhesive superoleophobicity for high-efficiency oil/water separation. Advanced Materials 2013, 25 (30), 4192–4198 Cai Y, Chen D, Li N, Xu Q, Li H, He J, Lu J (2018) A smart membrane with antifouling capability and switchable oil wettability for high-efficiency oil/water emulsions separation. J Membr Sci 555:69–77 Cao G, Wang Y, Wang C, Ho S-H (2019) A dually prewetted membrane for continuous filtration of water-in-light oil, oil-in-water, and water-in-heavy oil multiphase emulsion mixtures. J Mater Chem A 7(18):11305–11313 Chen P-C, Xu Z-K (2013) Mineral-coated polymer membranes with superhydrophilicity and underwater superoleophobicity for effective oil/water separation. Sci Rep 3:2776; (b) Zhu, Y.; Zhang, F.; Wang, D.; Pei, X. F.; Zhang, W.; Jin, J., A novel zwitterionic polyelectrolyte grafted PVDF membrane for thoroughly separating oil from water with ultrahigh efficiency. Journal of Materials Chemistry A 2013, 1 (18), 5758–5765; (c) Wang, R.; Zhao, X.; Lan, Y.; Liu, L.; Gao, C., In situ metal-polyphenol interfacial assembly tailored superwetting PES/SPES/MPN membranes for oil-in-water emulsion separation. Journal of Membrane Science 2020, 615, 118566

1724

A. Rashidi and S. Sharafinia

Chen Y, Bai Y, Chen S, Ju J, Li Y, Wang T, Wang Q (2014) Stimuli-responsive composite particles as solid-stabilizers for effective oil harvesting. ACS Appl Mater Interfaces 6(16):13334–13338 Cheng B, Li Z, Li Q, Ju J, Kang W, Naebe M (2017) Development of smart poly (vinylidene fluoride)-graft-poly (acrylic acid) tree-like nanofiber membrane for pH-responsive oil/water separation. J Membr Sci 534:1–8; (b) Li, X.; Zhang, Q.; Zhang, W.; Qu, R.; Wei, Y.; Feng, L., Smart Nylon Membranes with pH-Responsive Wettability: High-Efficiency Separation on Demand for Various Oil/Water Mixtures and Surfactant-Stabilized Emulsions. Advanced Materials Interfaces 2018, 5 (21), 1801179; (c) Dang, Z.; Liu, L.; Li, Y.; Xiang, Y.; Guo, G., In situ and ex situ pH-responsive coatings with switchable wettability for controllable oil/water separation. ACS Applied Materials & Interfaces 2016, 8 (45), 31281–31288 Cheng M, He H, Zhu H, Guo W, Chen W, Xue F, Zhou S, Chen X, Wang S (2019) Preparation and properties of pH-responsive reversible-wettability biomass cellulose-based material for controllable oil/water separation. Carbohydr Polym 203:246–255 Christou M, Konstantinidou M (2012) Safety of offshore oil and gas operations: lessons from past accident analysis. Joint Research Centre of the European Commission, 1–60; (b) Mangadlao, J. D.; Cao, P.; Advincula, R. C., Smart cements and cement additives for oil and gas operations. Journal of Petroleum Science and Engineering 2015, 129, 63–76 Chung D, Wang S (2003) Self-sensing of damage and strain in carbon fiber polymer-matrix structural composites by electrical resistance measurement. Polym Polym Compos 11(7): 515–525 Coyle B, McCann M, Kavanagh K, Devereux M, McKee V, Kayal N, Egan D, Deegan C, Finn GJ (2004) Synthesis, X-ray crystal structure, anti-fungal and anti-cancer activity of [Ag2 (NH3) 2 (salH) 2](salH2= salicylic acid). J Inorg Biochem 98(8):1361–1366; (b) Liu, Q.; Wang, Y.; Li, W.; Wu, L., Structural characterization and chemical response of a Ag-coordinated supramolecular gel. Langmuir 2007, 23 (15), 8217–8223 Fathi SJ, Austad T, Strand S (2011) Water-based enhanced oil recovery (EOR) by “smart water”: Optimal ionic composition for EOR in carbonates. Energy Fuel 25(11):5173–5179 Gao D, Sturm M, Mo Y (2009) In 13. Development of carbon nanofiber self-consolidating concrete. In: 2nd international symposium on design, performance and use of self consolidating concrete. RILEM Publications Sarl, pp 126–134; (b) Coppola, L.; Buoso, A.; Corazza, F. In Electrical properties of carbon nanotubes cement composites for monitoring stress conditions in concrete structures, PROTECT 2011. Performance, Protection & Strenghtening of Structures under Extreme Loading. Third International Workshop., Trans Tech Publications: 2011; pp 118– 123; (c) Abdullah, D. M.; Abdul-Hamead, A. A.; Othman, F. M. In Smart Cement Modified by Adding Multi Nano Powder for Oil Well Applications, Defect and Diffusion Forum, Trans Tech Publ: 2020; pp 8–15 Gao C, Sun Z, Li K, Chen Y, Cao Y, Zhang S, Feng L (2013) Integrated oil separation and water purification by a double-layer TiO2 -based mesh. Energy Environ Sci 6(4):1147–1151; (b) Tian, D.; Zhang, X.; Tian, Y.; Wu, Y.; Wang, X.; Zhai, J.; Jiang, L., Photo-induced water– oil separation based on switchable superhydrophobicity–superhydrophilicity and underwater superoleophobicity of the aligned ZnO nanorod array-coated mesh films. Journal of Materials Chemistry 2012, 22 (37), 19652–19657 Gbadamosi AO, Junin R, Manan MA, Agi A, Yusuff AS (2019) An overview of chemical enhanced oil recovery: recent advances and prospects. Int Nano Lett 9(3): 171–202 Gerogiorgis DI, Clark C, Vryzas Z, Kelessidis VC (2015) Development and parameter estimation for a multivariate Herschel-Bulkley rheological model of a nanoparticle-based smart drilling fluid. In: Computer aided chemical engineering, vol 37. Elsevier, pp 2405–2410 Habibi S, Jafari A, Fakhroueian Z (2020) Wettability alteration analysis of smart water/novel functionalized nanocomposites for enhanced oil recovery. Pet Sci 17(5):1318–1328 Haghighi OM, Firozjaii AM (2020) An experimental investigation into enhancing oil recovery using combination of new green surfactant with smart water in oil-wet carbonate reservoir. J Pet Explor Prod Technol 10(3):893–901

66 Smart Materials in Oil and Gas Industry: Application

1725

Han B, Yu X, Ou J (2011) Multifunctional and smart carbon nanotube reinforced cement-based materials. In: Nanotechnology in civil infrastructure. Springer, pp 1–47 Harrison RM (2007) An introduction to pollution science. J Anal At Spectrom 22:599–607 Hu Z, Gao H, Ramisetti SB, Zhao J, Nourafkan E, Glover PW, Wen D (2019) Carbon quantum dots with tracer-like breakthrough ability for reservoir characterization. Sci Total Environ 669: 579–589 Hwang C-C, Wang L, Lu W, Ruan G, Kini GC, Xiang C, Samuel EL, Shi W, Kan AT, Wong MS (2012) Highly stable carbon nanoparticles designed for downhole hydrocarbon detection. Energy Environ Sci 5(8):8304–8309 Isaia GC, Gastaldini ALG, Moraes R (2003) Physical and pozzolanic action of mineral additions on the mechanical strength of high-performance concrete. Cem Concr Compos 25(1):69–76 Jarrahian K, Seiedi O, Sheykhan M, Sefti MV, Ayatollahi S (2012) Wettability alteration of carbonate rocks by surfactants: a mechanistic study. Colloids Surf A Physicochem Eng Asp 410:1–10 Kakati A, Kumar G, Sangwai JS (2020) Oil recovery efficiency and mechanism of low salinityenhanced oil recovery for light crude oil with a low acid number. ACS Omega 5(3):1506–1518 Kamalak H, Erdem M, Altin E, Demirel S, Oz E, Altin S (2016) Giant dielectric constant and mechanical, heat conducting and structural properties of dental composites by adding Ca ions. J Dent Orofac Surg 1(3):1–8 Karimi M, Al-Maamari RS, Ayatollahi S, Mehranbod N (2015) Mechanistic study of wettability alteration of oil-wet calcite: the effect of magnesium ions in the presence and absence of cationic surfactant. Colloids Surf A Physicochem Eng Asp 482:403–415 Kim H, Watthanaphanit A, Saito N (2017) Simple solution plasma synthesis of hierarchical nanoporous MnO2 for organic dye removal. ACS Sustain Chem Eng 5(7):5842–5851; (b) Sharma, Y., Adsorption characteristics of a low-cost activated carbon for the reclamation of colored effluents containing malachite green. Journal of Chemical & Engineering Data 2011, 56 (3), 478–484; (c) Zhao, X.; Wang, K.; Gao, Z.; Gao, H.; Xie, Z.; Du, X.; Huang, H., Reversing the dye adsorption and separation performance of metal–organic frameworks via introduction of− SO3H groups. Industrial & Engineering Chemistry Research 2017, 56 (15), 4496–4501; (d) Du, X.-D.; Wang, C.-C.; Liu, J.-G.; Zhao, X.-D.; Zhong, J.; Li, Y.-X.; Li, J.; Wang, P., Extensive and selective adsorption of ZIF-67 towards organic dyes: performance and mechanism. Journal of colloid and interface science 2017, 506, 437–441 Kong X, Ohadi M (2010) Applications of micro and nano technologies in the oil and gas industry-overview of the recent progress. In: Abu Dhabi international petroleum exhibition and conference. Society of Petroleum Engineers Kwon G, Kota AK, Li Y, Sohani A, Mabry JM, Tuteja A (2012) On-demand separation of oil-water mixtures. Adv Mater 24(27):3666–3671 Li X-M, Reinhoudt D, Crego-Calama M (2007) What do we need for a superhydrophobic surface? A review on the recent progress in the preparation of superhydrophobic surfaces. Chem Soc Rev 36(8):1350–1368; (b) Wen, L.; Tian, Y.; Jiang, L., Bioinspired super-wettability from fundamental research to practical applications. Angewandte Chemie International Edition 2015, 54 (11), 3387–3399 Li M-C, Wu Q, Song K, Qing Y, Wu Y (2015a) Cellulose nanoparticles as modifiers for rheology and fluid loss in bentonite water-based fluids. ACS Appl Mater Interfaces 7(8):5006–5016; (b) Hamed, S. B.; Belhadri, M., Rheological properties of biopolymers drilling fluids. Journal of Petroleum Science and Engineering 2009, 67 (3–4), 84–90; (c) Li, M.-C.; Wu, Q.; Song, K.; Lee, S.; Jin, C.; Ren, S.; Lei, T., Soy protein isolate as fluid loss additive in bentonite–water-based drilling fluids. ACS applied materials & interfaces 2015, 7 (44), 24799–24809; (d) Kosynkin, D. V.; Ceriotti, G.; Wilson, K. C.; Lomeda, J. R.; Scorsone, J. T.; Patel, A. D.; Friedheim, J. E.; Tour, J. M., Graphene oxide as a high-performance fluid-loss-control additive in water-based drilling fluids. ACS applied materials & interfaces 2012, 4 (1), 222–227 Li J-J, Zhou Y-N, Luo Z-H (2015b) Smart fiber membrane for pH-induced oil/water separation. ACS Appl Mater Interfaces 7(35):19643–19650

1726

A. Rashidi and S. Sharafinia

Li S, Osisanya S, Haroun M (2016) Development of new smart drilling fluids using nano-materials for unconventional reservoirs. In: Abu Dhabi international petroleum exhibition & conference. Society of Petroleum Engineers Li L, Zhang J, Wang A (2018a) Removal of organic pollutants from water using superwetting materials. Chem Rec 18(2):118–136; (b) Cheng, Q.; Li, M.; Yang, F.; Liu, M.; Li, L.; Wang, S.; Jiang, L., An underwater pH-responsive superoleophobic surface with reversibly switchable oil-adhesion. Soft Matter 2012, 8 (25), 6740–6743 Li J-J, Zhou Y-N, Luo Z-H (2018b) Polymeric materials with switchable superwettability for controllable oil/water separation: a comprehensive review. Prog Polym Sci 87:1–33 Li F, Kong W, Bhushan B, Zhao X, Pan Y (2019) Ultraviolet-driven switchable superliquiphobic/superliquiphilic coating for separation of oil-water mixtures and emulsions and water purification. J Colloid Interface Sci 557:395–407 Liang L, Dong Y, Wang H, Meng X (2019) Smart cotton fabric with CO2 -responsive wettability for controlled oil/water separation. Adv Fiber Mater 1(3–4):222–230 Liu X, Ye Q, Yu B, Liang Y, Liu W, Zhou F (2010) Switching water droplet adhesion using responsive polymer brushes. Langmuir 26(14):12377–12382 Mofrad SK, Dehaghani AHS (2020) An experimental investigation into enhancing oil recovery using smart water combined with anionic and cationic surfactants in carbonate reservoir. Energy Rep 6:543–549 Moreno M, Castillo-Corullón S, Pérez-Ruiz E, Luna MC, Antón-Pacheco JL, Mondejar-Lopez P, De-la-Serna O, Villa JR, Osona B, Torres-Borrego J (2019) Spanish multicentre study on morbidity and pathogenicity of tracheal bronchus in children. Pediatr Pulmonol 54(10):1610– 1616 Nowrouzi I, Mohammadi AH, Manshad AK (2020) Utilization of methanol and acetone as mutual solvents to reduce interfacial tension (IFT) in enhanced oil recovery process by carbonated smart water injection. J Mol Liq 304:112733; (b) Nowrouzi, I.; Manshad, A. K.; Mohammadi, A. H., The mutual effects of injected fluid and rock during imbibition in the process of low and high salinity carbonated water injection into carbonate oil reservoirs. Journal of Molecular Liquids 2020, 305, 112432 Ong C, Shi Y, Chang J, Alduraiei F, Ahmed Z, Wang P (2019) Polydopamine as a versatile adhesive layer for robust fabrication of smart surface with switchable wettability for effective oil/water separation. Ind Eng Chem Res 58(12):4838–4843; (b) Lee, H.; Dellatore, S. M.; Miller, W. M.; Messersmith, P. B., Mussel-inspired surface chemistry for multifunctional coatings. science 2007, 318 (5849), 426–430; (c) Yang, H. C.; Waldman, R. Z.; Wu, M. B.; Hou, J.; Chen, L.; Darling, S. B.; Xu, Z. K., Membranes: Dopamine: Just the Right Medicine for Membranes (Adv. Funct. Mater. 8/2018). Advanced Functional Materials 2018, 28 (8), 1870052; (d) Chen, J.-K.; Chang, C.-J., Fabrications and applications of stimulus-responsive polymer films and patterns on surfaces: A review. Materials 2014, 7 (2), 805–875 Phrompet C, Sriwong C, Ruttanapun C (2019) Mechanical, dielectric, thermal and antibacterial properties of reduced graphene oxide (rGO)-nanosized C3AH6 cement nanocomposites for smart cement-based materials. Compos Part B 175:107128; (b) Xu, Z. H.; Liu, Z. T. In Fatigue damage sensing in smart carbon fiber concrete by electrical resistivity measurement, Key Engineering Materials, Trans Tech Publ: 2007; pp 345–348 Prins M, Welters W, Weekamp J (2001) Fluid control in multichannel structures by electrocapillary pressure. Science 291(5502):277–280 Qu M, Ma L, Wang J, Zhang Y, Zhao Y, Zhou Y, Liu X, He J (2019) Multifunctional superwettable material with smart ph responsiveness for efficient and controllable oil/water separation and emulsified wastewater purification. ACS Appl Mater Interfaces 11(27):24668– 24682; (b) Xu, Z.; Zhao, Y.; Wang, H.; Wang, X.; Lin, T., A superamphiphobic coating with an ammonia-triggered transition to superhydrophilic and superoleophobic for oil–water separation. Angewandte Chemie International Edition 2015, 54 (15), 4527–4530 Qu M, Liu Q, He J, Li J, Liu L, Yang C, Yang X, Peng L, Li K (2020) A multifunctional superwettable material with excellent pH-responsive for controllable in situ separation multiphase oil/water mixture and efficient separation organics system. Appl Surf Sci 515:145991

66 Smart Materials in Oil and Gas Industry: Application

1727

Rezaeian MS, Mousavi SM, Saljoughi E, Amiri HAA (2020) Evaluation of thin film composite membrane in production of ionically modified water applied for enhanced oil recovery. Desalination 474:114194; (b) Leo, D. J., Engineering analysis of smart material systems. John Wiley & Sons: 2007; (c) Smith, R. C., Smart material systems: model development. SIAM: 2005 Sabir IH, Haldorsen S, Torgersen J, Alestrom P, Gaut S, Colleuille H, Pedersen TS, Kitterod N-O, Alestrom P (2000) Synthetic DNA tracers: examples of their application in water related studies. IAHS Publication (International Association of Hydrological Sciences) (262), pp 159–165 Shami Z, Gharloghi A, Amininasab SM (2019) Multifunctional pH-switched superwetting copolymer nanotextile: surface engineered toward on-demand light oil–water separation on superhydrophilic–underwater low-adhesive superoleophobic nonwoven mesh. ACS Sustain Chem Eng 7(9):8917–8930 Singh SK, Ahmed RM, Growcock F (2010) Vital role of nanopolymers in drilling and stimulations fluid applications. In: SPE annual technical conference and exhibition. Society of Petroleum Engineers Soleimani P, Shadizadeh SR, Kharrat R (2020) Experimental assessment of hybrid smart carbonated water flooding for carbonate reservoirs. Petroleum, 7(1):80–90 Strand S, Austad T, Puntervold T, Høgnesen EJ, Olsen M, Barstad SMF (2008) “Smart water” for oil recovery from fractured limestone: a preliminary study. Energy Fuel 22(5):3126–3133; (b) Fathi, S. J.; Austad, T.; Strand, S., “Smart water” as a wettability modifier in chalk: the effect of salinity and ionic composition. Energy & fuels 2010, 24 (4), 2514–2519; (c) Fernø, M.; Grønsdal, R.; Åsheim, J.; Nyheim, A.; Berge, M.; Graue, A., Use of sulfate for water based enhanced oil recovery during spontaneous imbibition in chalk. Energy & fuels 2011, 25 (4), 1697–1706; (d) Zahid, A.; Shapiro, A.; Stenby, E. H.; Yan, W., Managing injected water composition to improve oil recovery: A case study of North Sea chalk reservoirs. Energy & fuels 2012, 26 (6), 3407–3415 Vipulanandan C, Mohammed A, Samuel R (2018) Fluid loss control in smart bentonite drilling mud modified with nanoclay and quantified with Vipulanandan fluid loss model. In: Offshore technology conference, Houston Vryzas Z, Kelessidis VC, Bowman MB, Nalbantian L, Zaspalis V, Mahmoud O, Nasr-El-Din HA (2017) Smart magnetic drilling fluid with in-situ rheological controllability using Fe3 O4 nanoparticles. In: SPE middle east oil & gas show and conference. Society of Petroleum Engineers Walcarius A, Mercier L (2010) Mesoporous organosilica adsorbents: nanoengineered materials for removal of organic and inorganic pollutants. J Mater Chem 20(22):4478–4511; (b) Sivarajasekar, N.; Mohanraj, N.; Sivamani, S.; Maran, J. P.; Moorthy, I. G.; Balasubramani, K., Statistical optimization studies on adsorption of ibuprofen onto Albizialebbeck seed pods activated carbon prepared using microwave irradiation. Materials Today: Proceedings 2018, 5 (2), 7264–7274; (c) Sivarajasekar, N.; Mohanraj, N.; Balasubramani, K.; Maran, J. P.; Moorthy, I. G.; Karthik, V.; Karthikeyan, K., Optimization, equilibrium and kinetic studies on ibuprofen removal onto microwave assisted–activated Aegle marmelos correa fruit shell. Desalin Water Treat 2017, 84, 48–58; (d) Roy, E.; Nagar, A.; Chaudhary, S.; Pal, S., A Review on Advanced Properties and Applications of AIEgens Inspired Smart Materials. Industrial & Engineering Chemistry Research 2020; (e) Mahmoodi, N. M.; Hayati, B.; Arami, M., Textile dye removal from single and ternary systems using date stones: Kinetic, isotherm, and thermodynamic studies. Journal of Chemical & Engineering Data 2010, 55 (11), 4638–4649; (f) Zou, W.; Li, K.; Bai, H.; Shi, X.; Han, R., Enhanced cationic dyes removal from aqueous solution by oxalic acid modified rice husk. Journal of Chemical & Engineering Data 2011, 56 (5), 1882–1891; (g) Albadarin, A. B.; Collins, M. N.; Naushad, M.; Shirazian, S.; Walker, G.; Mangwandi, C., Activated lignin-chitosan extruded blends for efficient adsorption of methylene blue. Chemical Engineering Journal 2017, 307, 264–272; (h) Zaini, M. A. A.; Salleh, L. M.; Azizi, M.; Yunus, C.; Naushad, M., Potassium hydroxide-treated palm kernel shell sorbents for the efficient removal of methyl violet dye. Desalin. Water Treat 2017, 84, 262–270; (i) Mirzaie, M.; Rashidi, A.; Tayebi, H.-A.; Yazdanshenas, M. E., Removal of anionic dye from

1728

A. Rashidi and S. Sharafinia

aqueous media by adsorption onto SBA-15/polyamidoamine dendrimer hybrid: adsorption equilibrium and kinetics. Journal of Chemical & Engineering Data 2017, 62 (4), 1365–1376; (j) Sivarajasekar, N.; Baskar, R.; Ragu, T.; Sarika, K.; Preethi, N.; Radhika, T., Biosorption studies on waste cotton seed for cationic dyes sequestration: equilibrium and thermodynamics. Applied Water Science 2017, 7 (4), 1987–1995; (k) Naushad, M.; Abdullah ALOthman, Z.; Rabiul Awual, M.; Alfadul, S. M.; Ahamad, T., Adsorption of rose Bengal dye from aqueous solution by amberlite Ira-938 resin: kinetics, isotherms, and thermodynamic studies. Desalination and water treatment 2016, 57 (29), 13527–13533 Wang N, Deng Z (2019) Synthesis of magnetic, durable and superhydrophobic carbon sponges for oil/water separation. Mater Res Bull 115:19–26 Wang B, Guo Z (2013) pH-responsive bidirectional oil–water separation material. Chem Commun 49(82):9416–9418 Wang Y, Lai C, Hu H, Liu Y, Fei B, Xin JH (2015) Temperature-responsive nanofibers for controllable oil/water separation. RSC Adv 5(63):51078–51085; (b) Li, J.-J.; Zhu, L.-T.; Luo, Z.-H., Electrospun fibrous membrane with enhanced swithchable oil/water wettability for oily water separation. Chemical Engineering Journal 2016, 287, 474–481 Wang Z, Xiao C, Wu Z, Wang Y, Du X, Kong W, Pan D, Guan G, Hao X (2017) A novel 3D porous modified material with cage-like structure: fabrication and its demulsification effect for efficient oil/water separation. J Mater Chem A 5(12):5895–5904; (b) Zhang, J.; Seeger, S., Polyester materials with superwetting silicone nanofilaments for oil/water separation and selective oil absorption. Advanced Functional Materials 2011, 21 (24), 4699–4704; (c) Tang, X.; Si, Y.; Ge, J.; Ding, B.; Liu, L.; Zheng, G.; Luo, W.; Yu, J., In situ polymerized superhydrophobic and superoleophilic nanofibrous membranes for gravity driven oil–water separation. Nanoscale 2013, 5 (23), 11657–11664; (d) Li, J.; Kang, R.; Tang, X.; She, H.; Yang, Y.; Zha, F., Superhydrophobic meshes that can repel hot water and strong corrosive liquids used for efficient gravity-driven oil/water separation. Nanoscale 2016, 8 (14), 7638–7645; (e) Tai, M. H.; Gao, P.; Tan, B. Y. L.; Sun, D. D.; Leckie, J. O., Highly efficient and flexible electrospun carbon– silica nanofibrous membrane for ultrafast gravity-driven oil–water separation. ACS applied materials & interfaces 2014, 6 (12), 9393–9401; (f) Yu, L.; Han, M.; He, F., A review of treating oily wastewater. Arabian journal of chemistry 2017, 10, S1913-S1922; (g) Saththasivam, J.; Loganathan, K.; Sarp, S., An overview of oil–water separation using gas flotation systems. Chemosphere 2016, 144, 671–680 Xia F, Ge H, Hou Y, Sun T, Chen L, Zhang G, Jiang L (2007) Multiresponsive surfaces change between superhydrophilicity and superhydrophobicity. Adv Mater 19(18):2520–2524 Xue B, Gao L, Hou Y, Liu Z, Jiang L (2013) Temperature controlled water/oil wettability of a surface fabricated by a block copolymer: application as a dual water/oil on–off switch. Adv Mater 25(2):273–277; (b) Qu, R.; Liu, Y.; Zhang, W.; Li, X.; Feng, L.; Jiang, L., Aminoazobenzene@ Ag modified meshes with large extent photo-response: towards reversible oil/water removal from oil/water mixtures. Chemical science 2019, 10 (14), 4089–4096; (c) Jung, M. C.; Cho, S.-H.; Kim, S. H.; Kim, H.-Y.; Lee, H. J.; Oh, K. H.; Moon, M.-W., UVresponsive nano-sponge for oil absorption and desorption. Scientific reports 2015, 5, 12908; (d) Qiu, L.; Sun, Y.; Guo, Z., Designing novel superwetting surfaces for high-efficiency oil– water separation: design principles, opportunities, trends and challenges. Journal of Materials Chemistry A 2020, 8 (33), 16831–16853 Xue Z, Cao Y, Liu N, Feng L, Jiang L (2014) Special wettable materials for oil/water separation. J Mater Chem A 2(8):2445–2460; (b) Zheng, X.; Guo, Z.; Tian, D.; Zhang, X.; Jiang, L., Electric field induced switchable wettability to water on the polyaniline membrane and oil/water separation. Advanced Materials Interfaces 2016, 3 (18), 1600461 Youngblood JP, McCarthy TJ (1999) Ultrahydrophobic polymer surfaces prepared by simultaneous ablation of polypropylene and sputtering of poly (tetrafluoroethylene) using radio frequency plasma. Macromolecules 32(20):6800–6806; (b) Tian, D.; Zhang, X.; Wang, X.; Zhai, J.; Jiang, L., Micro/nanoscale hierarchical structured ZnO mesh film for separation of water and oil. Physical Chemistry Chemical Physics 2011, 13 (32), 14606–14610

66 Smart Materials in Oil and Gas Industry: Application

1729

Yu H (2012) Transport and retention of surface-modified nanoparticles in sedimentary rocks. Dissertation, University of Texas at Austin Zakaria M, Husein MM, Harland G (2012) Novel nanoparticle-based drilling fluid with improved characteristics. In: SPE international oilfield nanotechnology conference and exhibition. Society of Petroleum Engineers Zamani A, Bataee M, Hamdi Z, Khazforoush F (2019) Application of smart nano-WBM material for filtrate loss recovery in wellbores with tight spots problem: an empirical study. J Pet Explor Prod Technol 9(1):669–674; (b) Jervis, T. T.; Jardine, S. I., Alarm system for wellbore site. Google Patents: 1999 Zhang L, Zhang Z, Wang P (2012) Smart surfaces with switchable superoleophilicity and superoleophobicity in aqueous media: toward controllable oil/water separation. NPG Asia Mater 4(2):e8–e8; (b) Chang, S.; Singamaneni, S.; Kharlampieva, E.; Young, S. L.; Tsukruk, V. V., Responsive hybrid nanotubes composed of block copolymer and gold nanoparticles. Macromolecules 2009, 42 (15), 5781–5785 Zhang X, Liu C, Yang J, Huang X-J, Xu Z-K (2020) Wettability switchable membranes for separating both oil-in-water and water-in-oil emulsions. J Membr Sci 624:118976 Zhao L, Li L, Wang Y, Wu J, Meng G, Liu Z, Guo X (2018) Preparation and characterization of thermo-and pH dual-responsive 3D cellulose-based aerogel for oil/water separation. Appl Phys A 124(1):9

Food Industry: Applications of Digitalization

67

Vahid Mohammadpour Karizaki

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classifying the Application of Digitalization in Food Industry . . . . . . . . . . . . . . . . . . . . . . . Different Levels of Digitalization’s Relation with a Process in Food Industry . . . . . . . . . . . Legislation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Communication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Useful Websites for Further Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1732 1734 1736 1736 1736 1737 1741 1744 1744 1747 1748 1748

Abstract

The utilization of digitalized technologies, data, and systems in food industries can offer potential solutions to different challenges across personalized nutrition, food processing, production line, and supply chain monitoring and management. This chapter classifies the applications of digitalization in the food industries on the basis of different criteria. The shifting toward digitalization is a strategic and necessary change that can increase the satisfaction of consumers and producers of the food industry throughout the world. It is also shown that how the digitalization can be related to a process in food industries. Different levels of digitalization’s relation with a process in food industry include legislation, data collection, processing, monitoring, economy, and communication. It will

V. M. Karizaki () Chemical Engineering Department, Quchan University of Technology, Quchan, Iran e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_131

1731

1732

V. M. Karizaki

be expected from traditional producers to accept and apply digital changes in the future. The food producers that resist the digital changes will not be able to compete with those have applied digitalized systems for production. Keywords

Digitalization · Food industry · Digitalized technology

Introduction Digitalization is generally defined as utilization of digitalized technologies, data, systems, and their interconnection which creates new or modified activities. The entire systems of control, management, production, and consumption as well as modern economies are being transformed by digitalized products, services, and technologies (FAO 2020). Connecting producers and consumers by digitalization in novel methods is an opportunity for renewing business models. In other words, access of entrepreneurs to global markets is facilitated by digitalization that opens up new and different opportunities (Commission 2019; FAO 2020). Digitalization can increase the connectivity of sectors in food industries, and decrease losses and inefficiencies in different ways. For example, consumers and producers can service and share required information even internationally with access to a stable Internet connection. Digitalized systems can be used by producers for working more efficiently, precisely, and sustainably (Garitta et al. 2013; Mohamed and Shalaby 2019; Shafaei et al. 2019). Also, these systems have a good potential for offering greater transparency to consumers as to how a product is prepared (Commission 2019). The quick development of digital technology makes companies and industries eager for changing and modifying their business models. The five domains of how digital technology can change a business model have been considered by Rogers (2016). The first domain called customer perspective is a critical issue for manufacturers to use social media and digital communication for interacting with the customers. Competition with the other companies is the second domain that often results in development of industry. The third domain is data, which is how to analyze, manage, and apply the available information and statistics. The innovation process is the next domain. The digitalized systems and services make it easier for industries to utilize novel approaches for creation of new products (MaryAnne 2018). Also, receiving continuous feedback from the market for new products can be carried out easily (FAO 2020). Value is the last domain that is introduced by Rogers (2016). New entrant companies and organizations are digitally born. In other words, their business models are digitalized from the beginning. In contrast, the established industries must struggle to overcome difficulties and challenges that are created by new entrants. Consequently, changing and improving their value proposition is very important in this respect. Over the last years, producers and manufacturers have been faced with several challenges that are related to the demand volatility and to the changing requirements

67 Food Industry: Applications of Digitalization

1733

from suppliers as well as consumers. One of these challenges is the introduction and integration of novel techniques and technologies for improving efficiency, quality, and competitiveness. The other challenge in food industry deals with changing supply base to demand-based approaches, in which the customers tell manufacturers what they want to buy. Production of foodstuff with different and individual tastes shows that production will be adapted to consumer demand. According to this view, companies are trying to boost applying of digitalization in different elements such as machines, processes, storage systems, and utilities (Demartini et al. 2018). The sharing of digital information is also one of the advantages of digitalization. It will show a new surface of accountability between primary producers and customers and among other actors at different levels of the food chain. Availability of information with higher transparency enhances trust and confidence among customers, and trading partners alike (WHO 2019). However, the adoption of digitalization also creates questions related to data ownership, privacy, use, and so on that must be determined and addressed. Digitalization can offer potential solutions to different challenges across personalized nutrition, food processing, production line, and supply chain monitoring and management (Pichawadee and Kim 2020). Certainly, food manufacturers that resist the digital changes will not be able to compete with those that have applied digitalized systems for production. In order to find the current level of published data related to digitalization in the food industries, literature review can be conducted. Investigating the scientific body of knowledge shows the subject “application of digitalization in food sector” is increasingly being studied by researchers in the recent years. As an example, the number of publications in Sciencedirect database concerning this topic in the last five years is shown in Fig. 1. The word “digitalization” in the title of article and the

Fig. 1 Development of the topic “digitalization in food industry” in the literature in the last five years

1734

V. M. Karizaki

“food” in the journal name are selected for retrieving the articles of interest. As it is possible to understand from the figure, the numbers of publications started growing from 2017 and reached its maximum in 2020. This chapter presents the different applications of digitalization in food industries.

Classifying the Application of Digitalization in Food Industry The applications of digitalization in the food industries can be classified on the basis of different criteria. One of the important criteria is processing step that shows which step (before, during, or after processing) is related to digitalization. In other words, this criterion determines which digitalized system is involved in which of the following parts: raw materials, process in progress, or products. The shifting toward digitalization is a strategic and necessary change that can increase the satisfaction of manufacturers and consumers of the food industry all around the world. The mentioned change may involve the processing of a food material, just like three-dimensional printing of food ingredients that is a completely digitalized processing in the food industry. Furthermore, digitalization before or after processing is an important part of approaches that are changing the food production and consumption throughout the world. For example, sampling and quality control of raw materials using digital image analysis is a digitalized step that may perform before processing in food factories. Also, digitalized packaging of products is an example for final steps in manufacturing lines. Packaging that is defined as the art and the science of protecting or enclosing products for sale is usually applied after processing of food materials. As is shown in Fig. 2 using of digitalization can be applied before, during, or after processing of food materials. The presented example in this figure can be explained in terms of “intersection of sets” in mathematics. Figure 2 contains three set of data called A (before processing), B (during processing), and C (after processing). The intersection of two sets consists of all data that is in the first set and also belongs to the second set. For example, “food economy” is an element that is common in both A and B. Also, the intersection of two sets B and C includes the term food economy. Therefore, it is concluded that digitalization of food economy is a key factor for all steps of food industry (i.e., before, during, and after processing). The other criterion that can be used for categorizing the application of digitalization in food industry is whether the subject has been commercialized in the last decades, or will be considerably developed in the future years. For example, characterization of products by digital analysis is a common technique that has been applied from last decades. Sapirstein et al. (1987) used the image analysis for determining the morphologic characteristics of whole grains. Guillaume et al. (1996) designed a sensor and analyzed the in-flow images for characterization of mill products.

67 Food Industry: Applications of Digitalization

1735

Fig. 2 Intersection of sets in mathematics for presenting an example of involving digitalization in food industries before, during, and after processing

The subjects of characterization and analysis of products by digitalized systems have been developed very well in the past, although many advances will be expected to happen in the future. As an example of subjects that aren’t commercialized all around the world so far, the three-dimensional printing of food ingredients can be mentioned. The 3D printing systems are in the nascent form currently, and limited business projects have been commercialized up to now (Jayaprakash et al. 2019). Although the subject of 3D printing of food materials is being considered by many researchers in the world, the practical applicability of these printers is still low due to the technical difficulties and market challenges. Also, the type of digitalization’s relation to processes can be considered as an important criterion. Figure 3 shows how digitalization can be related to a process in food industries. This relation should be first defined in legislation level. It is a requirement for governments and international organizations to develop plans for adapting laws on utilizing the digital technologies in different food processes (Sidorenko and van Arx 2020). Digitalization for data collection is a main part that can be dealt with the process. Digital tools offer a faster approach in high quality for collecting the required data. The other level of digitalization’s relation with the process is processing. Digitalization of a unit operation or process means that the process itself is related to digitalized system. In other words, digitalization is applied during processing of food materials. Monitoring, data analysis, and control are also important parts in the food industries. A process can be affected by digitalization of these parts by reducing the cost and increasing the speed of analysis and control. Digital economy is the next part that has close relation with a process. Digital economy is known as an economy that is formed on the base of digital computing technologies. This type of economy is also referred to as Web economy, or Internet economy which implies that the businesses are being conducted and controlled today based on the World Wide Web.

1736

V. M. Karizaki

Fig. 3 Different levels of digitalization’s relation with a process in food industry

Finally, digital communication can be used for increasing the satisfaction of producers and consumers of food products. Digital communication is related to the design and application of a variety of systems such as voice digital telephone, video digital telephone, wireless LANs, satellite devices, and so on (Barry et al. 2004). Sending any image or text-based message through streaming videos, website, audio, digital photography, or graphic design is defined by digital communication.

Different Levels of Digitalization’s Relation with a Process in Food Industry Legislation With the rapid development of digitalization, creation of global, comprehensive, and coherent legal safeguards in order to minimize risks of digital technology and to legitimize new assets is essential. Governments and international organizations are currently developing strategies to adapt different laws on using of novel digital techniques (Sidorenko and van Arx 2020).

Data Collection Data collection is a procedure for measuring and gathering information on variables in a system or process. Determination of fat content of potato during frying process

67 Food Industry: Applications of Digitalization

1737

can be considered as an example of data collection. Digitalized systems or devices employed for data collection differ by unit operation and vary from one type of producer to another (McNamara 2017). For example, a digital thermometer may be used in osmotic dehydration for measuring the osmotic solution temperature, and a digital flow-meter may be utilized in a cabinet dryer for determination of hot air flow rate during drying process. As mentioned earlier, digitalization offers a rapid and accurate approach for achieving better results. For example, the moisture content of a food sample in different processes such as drying and evaporation can be determined by traditional methods. Different methods for moisture content determination all consist of removing moisture from food sample through heating. A common technique is the procedure proposed by Association of official analytical chemists (AOAC 1990). The amount of water that is removed after 24 h in an oven at a temperature of 105 ◦ C can be defined as moisture content value. This traditional technique is time-consuming and relatively accurate. Employing a digital moisture analyzer instrument provides faster and more accurate results.

Processing Several researches related to application of digitalization in process study are published in literature (Li et al. 2014, 2015; Tomita et al. 2019). Process study by digitalized system or device is an effective way for evaluation of the process, and finding the important parameters involved in the process. Some researchers have applied digitalization for process study in food industry. For example, Li et al. (2014) considered the application of digital image analysis for studying the gelatinization process. The authors combined information from differential scanning calorimetry and light microscopy. Then, the digital image analysis method was employed for information analysis. The effect of sodium chloride solution on the gelatinization of potato, corn, and pea starch was studied in this work. Li et al. (2015) investigated the effect of different sugar concentrations on the gelatinization process of corn starch by using digital image analysis. In the other research work, Tomita et al. (2019) studied the absorption process by using digital image analysis technique. They developed a novel approach in which a digital microscope is applied to visualize the process of water absorption in real time. The characteristic of water absorption in rice grain at 10, 25, 40, and 55 ◦ C was consistent with the obtained results of digital image analysis. The digital image analysis method is also employed for studying the baking process of bread. Paquet et al. (2012) used a multilevel oven for conducting the baking process. An imaging and analysis system was able to detect color and volume changes of bread rolls online. Therefore, by presenting color, shape, and size of dough, the state of the bread-baking process was identified reliably. Digitalization during packaging is the other subject that has been considered by investigators. Although packaging is a unit operation that is usually applied at the end of production lines, it may be considered as an independent process

1738

V. M. Karizaki

or step in food industries. Protecting products or raw materials from deteriorative effects is the main goal of packaging. Furthermore, packaging is known as an effective way of marketing to communicate with the customers. Over the past decades, terms such as smart packaging and intelligent packaging have been used in literature. These terms often show the capability of a packaging process in carrying out smart/intelligent features (such as detecting, tracing, sensing, recording, or communicating) to make easier a decision for shelf life extension, safety increment, information and data collection, and quality improvement (Schaefer and Cheung 2018). Digital technology is a powerful tool for creation of more transparent and efficient packaging system. This system consists of different steps such as packaging design, artwork preparation, pre-press operations, package printing, cutting, folding, and distribution. All stakeholders, from designers and suppliers, to machine manufacturers, and company owners need to adapt with digitalization techniques. A considerable amount of published data on application of digitalization in process study is related to 3D printing of food materials. The three-dimensional printing in the food industry is a digitalized process to create layer-by-layer construction of simple or complex 3D food objects. The process is also known as additive food manufacturing, or solid freeform fabrication. The whole operation is automatically controlled by a simulation program with the lowest level of human interaction. The different types of materials such as natural polymers and a wide variety of food ingredients can be utilized in 3D printing systems. Unlike robotic technology that can easily control and automate the homemade cooking procedure, additive food manufacturing allows production of individualized edible components through printing of food materials. Although different studies related to 3D printing have been presented in the literature over the last decades, the first 3D food printer called Fab@Home model 1 capable of manufacturing edible structure was introduced by researchers from Cornell University in 2007 (Godoi et al. 2016; Jayaprakash et al. 2019). However, 3D food printing techniques are at the beginning of a long way, and limited projects have been commercialized so far. In other words, the products of 3D food printing are obtained with higher cost in comparison with the conventional food manufacturing technologies (Jayaprakash et al. 2019). The other drawback of 3D printing is that the general public prefer the traditional product, because the 3D food printed materials are unknown and unrevealed (Sun et al. 2018). The available information in the scope of 3D food printing has to do with the applying of the printable food ingredients, rheology optimization of the food materials, and development of the 3D food printing technology. Also, the researchers are focusing on the broad range of 3D food printing applications from applying in home kitchen and restaurant, to the installing in the airplane and shuttles (Jayaprakash et al. 2019). Due to the change of consumption patterns in different groups such as children, elderly, and athletes, the researchers are working on new approaches for production of flavor, vitamins, and additives with pleasant properties, and longer lifetime duration.

67 Food Industry: Applications of Digitalization

1739

Employing 3D printing for teaching and education is an amazing example related to application of digitalization in food processes. Teaching shape, color, and taste to children is an important step in early education. Therefore, using the 3D food printing for production of edible pieces with special and customized shapes is an efficient method for learning. The researchers showed that this idea is successful in practice. They attracted children in an exhibition station for showing 3D food printing. A large number of children were invited for recognizing the shapes of cookies in an exhibition. The results confirm that the use of 3D printing can improve the education process (Sun et al. 2018). The other application of 3D printing is introduced by the term “personalized nutrition.” It is a unique nutrition recipe that is prepared and served for a special group of consumers. For example, personalized nutrition can be produced by 3D printing for elderly people that have difficulties in chewing or swallowing (Sun et al. 2018). The three-dimensional printing is one of those techniques that is involved in building unique food structures. Depending on the food material supply, the current 3D printing equipments are classified into three groups. The first group is liquid system that has been applied to generate 3D constructs through deposition of liquid-based ingredients. The layerby-layer deposition of liquid materials can be carried out via ink jet printing or extrusion process (Godoi et al. 2016). The second printing group is powder system that produces powder-based structure. This type of structure is formed by edible powder deposition. A heat source such as hot air or laser is applied simultaneously or at the following of deposition process. The last group is the cell culture system, also known as bio-printing technique. This system is mainly applied for production of meat analogue. Ink jet tool as a common printing system is based on the accumulation of food liquid droplets deposited by printer nozzles (Kruth et al. 2007; Godoi et al. 2016). This type of 3D printing system is shown schematically in Fig. 4. Usually, thermal or piezoelectric heads are utilized for operating the ink jet printers. Thermal head in a printer is electrically heated to create pressure pulses for moving liquid droplets from the nozzle. Piezoelectric head includes a piezoelectric crystal that generates sound wave. The created waves are applied to separate the food liquid into droplets. The low viscosity materials such as liquids, paste, or semi-liquid ingredients can be generally deposited by ink printers. As a result, these printers can’t be used for generation of complex food structures. The most common food materials deposited by ink jet printer include liquid dough, meat paste, chocolate, sugar icing, gels, cheese, butter, honey, jams, and so on (Murphy and Atala 2014; Godoi et al. 2016). The extrusion process as a printing method has been primarily utilized in deposition of hot-melt polymers layer by layer. Currently, this technique is adapted for printing of food materials to build 3D edible objects. Depending on the food ingredients entered to the extruder, the different binding mechanisms can be performed for deposition of layers. Several studies showed that the control of rheological properties of food material is an effective way for proper and accurate accommodation of layers (Lipton et al. 2010; Godoi et al. 2016; Periard et al. 2007).

1740

V. M. Karizaki

Fig. 4 Schematic illustration of ink jet printing system (Godoi et al. 2016)

For example, Lipton et al. (2010) investigated the effect of concentration of food ingredients including yolk, butter, and sugar on the 3D printing process. Solidification of a melted material upon cooling is another way for building and fixing the edible layers. Deposition of melted chocolate to create a 3D object was reported by investigators (Hao et al. 2010a, b). The authors produced the solid chocolate upon cooling of melted materials. Gel formation is also used as a common method for creation of layers in 3D food objects. The gel forming mechanism during 3D printing of food ingredients such as protein and starch has been considered by Cohen et al. (2009). The extrusion-based 3D printing has several steps that have been presented in Fig. 5. The first step is designing a three-dimensional model. Using a computer program for converting the 3D designed model into separated and individual layers is the second step. In this step the available software creates the correspondence codes for printing process. In the third step, the final programming code is uploaded and transferred into a 3D food printer. Then, the required food recipe is selected for printing. Three-dimensional printing of food materials is the final step. According to the designed patterns by the machine program, the food materials are extruded and dispensed by extruder nozzle. Depending on the type of products, a post-printing process such as cooking or baking may be applied on the 3D food printed pieces. The extrusion process in three-dimensional printing is digitally controlled. It has a robotic structure that can create complex 3D pieces from food ingredients. At the beginning of the process,

67 Food Industry: Applications of Digitalization

1741

Fig. 5 Schematic illustration of extrusion-based three-dimensional printing steps; 1) designing of 3D model, 2) generation of programming code, 3) uploading the code into printer, 4) printing of food material (Sun et al. 2018)

the raw materials are loaded by the extruder. Then the charged equipment pushes and moves the materials through a predesigned path. Finally, a bonding mechanism is carried out during layer deposition for creation and fixing a solid structure (Sun et al. 2018). The speed of layer deposition in the 3D food printing is one of the most important parameters in the process. This factor determines the process productivity and is related to the resolution of the extrusion process. The characteristics and properties of food materials are also main parameters in 3D printing. For example, deposition of very thin layers of a low viscosity food material leads to increase the cost and time of operation (Godoi et al. 2016). Table 1 summarizes several researches related to application of 3D printing systems in food industries.

Monitoring As mentioned earlier, digital technology can play an important role in monitoring, data analysis, and control of different processes in food industries. Monitoring, data analysis, and control are procedures that focus on determining the defects of products. In other words, these procedures will try to ensure consumers receive the final product without deficiency. Instead of monitoring, data analysis, and control, some researchers have applied other terms such as “quality management”

1742

V. M. Karizaki

Table 1 Application of 3D printing systems in food industries Printed material base Cocoa butter

Printing temperature (◦ C) 18, 20

2

Garden pea, carrot, and bok choy

Room temperature

Threedimensional food printing of different vegetable inks for dysphagic patients

3

Starch gels

[70–80]

4

Cookie dough

Room temperature

Investigation of the relationship between rheological properties, structures, and 3D printability of food material Threedimensional food printing of pre-heated cookie dough

5

Wheat starch, cassava starch

Room temperature

6

Egg yolk

[72–84]

No. 1

Main goal of study Investigation of the print stability of chocolate in different temperatures

Improvement of the 3D printability of wheat and cassava starch by using pulsed electric field (PEF) treatment Investigation of the relationship between protein structure, rheology, and printing behavior of egg yolk

Highlight result The print stability and heat transfer dynamics of the process are strongly changed with temperature changes. Optimized ink formulations for printed materials show minimal water seepage, and excellent 3D printability. Inhomogeneity of foodstuff formed new crystal structure during the process of 3D printing.

Reference (Rando and Ramaioli 2021)

Pre-heating of foodstuff can significantly improve the quality of printed cookie dough structures. Pulsed Electric Fields (PEF) treatment affected the properties of wheat and cassava starch.

(Pulatsu et al. 2021)

The heat treatment of egg yolk at 76 ◦ C for 8 min resulted in obtaining the best printing performance.

(Xu et al. 2020)

(Pant et al. 2021)

(Zeng et al. 2020)

(Maniglia et al. 2021)

67 Food Industry: Applications of Digitalization

1743

that provide a similar concept. The shift toward digitalization leads to a better performance in quality management. The routine methods and techniques of monitoring, data analysis, and control are usually time-consuming and costly. The high consumption of reagents and waste generation in utilization of these techniques are also reported. The other drawbacks of using routine methods include complex sample preparation, inaccurate or incomplete results generation, and so on (Silva and Rocha 2020). Now, using digitalized system is known as a novel technique for quantification and classification of raw materials or final products, quality assessment, online monitoring and control. Digital technology generally provides a simple and inexpensive tool for obtaining rapid and accurate results (Lei et al. 2020; Reile et al. 2020). Several applications of digitalization in food industries in this issue have been considered by investigators. Many studies have reported the application of digital technology in data analysis and quality control in food industries. The published works show that different parameters such as color, texture, moisture content, fat content, and sugar content of a food material can be detected or evaluated by a digital system. As an example, Silva and Rocha (2020) determined protein content of milk by digital image calorimetry. The authors developed a rapid, novel, cost-effective, and practical way for detection of milk frauds that are caused by dilution. Their approach involves the milk proteins precipitation by the salting-out of copper sulfate and determination of the remaining copper concentration. Because the intensity of color measured by digital image calorimetry is inversely proportional to the milk protein content, this technique can be used as an indicator of milk adulteration. The other example is work of ElMasry et al. (2009) related to digital image analysis of peanut. The authors considered the effect of moisture content on morphological properties of peanut by digital imaging technique. For this purpose, moisture-mechanical and moisture-morphological features equations were built. Also, the peanuts depending on their properties were classified at different moisture contents. Similar works on application of digitalization for classifying the grains and cereals, and determining the seed features have been published in recent years. Uniformity determination of soybeans (Shahin et al. 2006), detection of color and shape of flax seeds (Dana and Ivo 2008), determination of geometrical features in rapeseeds (Ta´nska et al. 2005) are some examples in this respect. As an example for application of digital technology in data analysis the work of Yam and Papadakis (2004) is considered. In food industries, it is very important to analyze the color of food sample surfaces both quantitatively and qualitatively. The food samples are visually inspected and compared with together in qualitative analysis. In quantitative analysis, the color distribution and the color averages can be obtained. The authors presented a simple way that combines computer, digital camera, and graphic software for measuring and analyzing the color of microwaved pizza.

1744

V. M. Karizaki

Also, Table 2 shows a brief review of recent attempts for employing digitalization in monitoring, data analysis, and control of different processes in food industries, from freshness detection of packaged chicken breast, to determination of lettuce shelf life, to analysis of crumb softness in bread.

Economy Digital economy is also known as the Internet economy. The basis of digital economy is digital computing technologies. It is a term that captures the effect of digitalization on patterns of consumption and production. In other words, digital economy determines how goods, products, and services are marketed, sold, traded, and paid for. The new wave of economical and technological development has been formed by superior connectivity speed and computing power. Also, the number of physical objects or things connected to the Internet (i.e., the Internet of things) is considerably being increased. The Internet of things definition is the network of physical things that are connected to other systems and devices over the Internet for the purpose of exchanging data and information. The Internet of things is predicted to contribute to future economical growth: about 125 billion devices and systems are estimated to be connected to the Internet by 2030 (CEPS and Barilla 2019). Due to the importance of digital economy in food industries, the manufacturers and company owners are greatly focusing on electronic commerce (e-commerce). E-commerce is being increased and developed every year and traditional retailers are moving to online retailers. An online retailer is not only global but is also available everywhere all the time. Information speed to online retailing determines the degree of maturity and ripeness of a food company. Old manufacturers that started before digitalization age must rethink for rapid joining to the digital economy (FAO 2020).

Communication Connectivity is a factor that shows whether or not a food process is involved with digitalized systems. More than 4.5 billion people had Internet access in 2020, encompassing near to 60% of the world’s population. The Internet access through the world in the future years will be increased. In addition, it will be easier to connect a process with other digitalized systems and devices that are connected to the Internet. Therefore, connectivity of million tools and devices to Internet network and people will change human life in the near future. Furthermore, digitalization will change all areas of food industry: from the way raw materials are processed, to the way final products are packaged, to the way we buy and consume foodstuff (McNamara 2017). Growing population throughout the world is a significant ongoing change that results in increasing of global food

67 Food Industry: Applications of Digitalization

1745

Table 2 Application of digitalization in monitoring, data analysis, and control of food processes

No. 1

Food material Chicken breast

2

Lettuce

3

Bread

4

Milk

Purpose of digitalization Freshness detection and monitoring of packaged foodstuff

description A colorimetric array for indicating freshness was used to monitor spoilage in chicken breast over the storage time. Determination Changes in color of shelf life composition and and acceptbrown area ability percent of foodstuff were determined. Lettuce quality was evaluated by digital image analysis. Analysis of The digital images crumb of bread crumbs softness were compared against real bread samples via strain compression experiments and visual texture.

Highperformance detection of pathogen

The high-performance molecular experiment for detecting mycobacterium bovis in milk samples was performed by digital LAMP (digital loop mediated isothermal amplification).

Main results Digital technology can provide a sensitive, specific, and low-cost pathway for monitoring of freshness in food products. The digital analysis of browning is an effective and reliable research tool for quantitatively determination of shelf life and quality. The bread crumbs in digital form exhibited similarities to real foodstuff in terms of wall thicknesses as obtained from surface appearance. Digital LAMP technique provided higher sensitivity and accuracy for detection of pathogen in milk, in comparison with the common methods.

Reference (Lee et al. 2019)

(Zhou et al. 2004)

(Wang et al. 2013)

(Tao et al. 2020)

(continued)

1746

V. M. Karizaki

Table 2 (continued)

No. 5

Food material Meat

6

Edible oil

7

Edible powder

8

Pear

9

Olive oil-lemon juice

10

Broccoli

Purpose of digitalization Determination of nitrite residue in processed meat

description Residual amounts of nitrite in meat products were determined through a novel analytical approach that applies digital images. Evaluation Peroxide values of of lipid edible oils and oxidation in their emulsion edible oil were evaluated by using digital image colorimetric analysis. Assessment The authors of the utilized the quality of coordinates of edible food CIELAB color powder space for mixes describing the color of edible powders. Determination Digital pear of sample surface enzymatic images have been browning utilized for kinetics of describing pear enzymatic samples browning. Determination Digital image of aging analysis was mechanisms employed for in emulsions measuring of mean droplet size of emulsions. Sensory Evaluations of assessment digital images of broccoli were compared over time with the real broccoli appearance.

Main results The obtained results of using the digital image method were in good agreement with the available data that has been published by official AOAC. The proposed method is nondestructive and can be employed as a rapid monitoring technique.

Reference (Azeem et al. 2019)

Digital color imaging has good potential for assessing the quality of powder mixes.

(Shenoy et al. 2014)

A digital methodology for texture analysis of pear was successfully applied.

(Quevedo et al. 2009)

Rheological behavior of olive oil-lemon juice emulsion was evaluated.

(Silva et al. 2010)

The corresponding images with the real broccoli obtained from digital sensory analysis were in good agreement.

(Garitta et al. 2013)

(Singkhonrat et al. 2019)

67 Food Industry: Applications of Digitalization

1747

demand. A growth of 70% of food demand by 2050 is provided by food and agriculture organization (FAO 2020). As a result, the total number of humans living all around the world are considered as an important piece of digitalization puzzle in food industries (Casper 2019). On the other hand, it is obvious from the interview that the food in the future is not only dependent on science and technology but also on the social conditions (Jayaprakash et al. 2019). Therefore, digitalization in the food industries without paying attention to digitalization of social media is an incorrect decision (Tania and Michelle 2018). Digital technologies such as web and mobile systems have transformed the business method in food industry. The different processes such as food ordering have been changed to a simple and fast form. The signs of oversaturation in food distribution and delivery market resulted in providing opportunities to create novel business models by digital innovation. Online services through Internet are working on these opportunities stronger than ever. Linking between offline and online worlds by mobile applications allows the customers to choose from the list option in sale order, check pieces, compare products, and place orders online (Tanmay 2019). Digitalization also provides information about cancelation, unusual transaction or returns. Thus, a transparent and efficient system is formed for serving customers. In addition to mobile applications, social media platforms such as Facebook, Twitter, Instagram, Telegram, and so on can be employed for communication between producers and consumers (Sima 2017; Tanmay 2019). Social media and food industry share a unique relationship, so that an inseparable whole is created. Understanding the reputation and historical background of a brand, finding the different target markets, asking from manufacturers and sales, and responding to the consumers are a small part of activities that can be rapidly and accurately performed due to the digital technology and social media relationship (Tanmay 2019).

Conclusion Paying attention to digitalization as a novel approach for employing in food industries is very essential and vital. New or modified activities can be created by using digitalized technologies, data, system, and their interconnection. Digitalization can offer potential solutions to different challenges across personalized nutrition, food processing, production line, and supply chain monitoring and management. Additionally, trust and confidence among customers and trading partner will be increased due to the availability of information with higher transparency. It will be expected from traditional producers to accept and apply digital changes in the near future. Otherwise, they will not be able to compete with those that have applied digitalized systems for production.

1748

V. M. Karizaki

Useful Websites for Further Information http://chocedge.com/ Accessed January 27, 2021 https://www.foodjet.com/ Accessed January 27, 2021 https://www.naturalmachines.com/press-kit/ Accessed January 28, 2021 https://www.impactmybiz.com/blog/digitalization-food-and-beverage-industry-outlook/ Accessed January 28, 2021

References AOAC (1990) Official methods of analysis, 15th edn. AOAC Azeem SMA et al (2019) Determination of nitrite in processed meat using digital image method and powdered reagent. J Food Compos Anal 81:28–36 Barry JR et al (2004) Digital communication, 3rd edn. Springer Casper T (2019) What drives digitalization in the food production industry? Feed & Livestock Magazine CEPS, C. f. E. P. S, C. f. F. N. F. Barilla (2019). Chapter 1-the digital technology stack: a primer. Digitising AgriFood: pathways and challenges Cohen DL et al (2009) Hydrocolloid printing: a novel platform for customized food production. 20th annual international solid freeform fabrication symposium, pp 807–818 Commission, E (2019) Digital transformation in agriculture and rural areas citation. Agriresearch factsheet Dana W, Ivo W (2008) Computer image analysis of seed shape and seed color for flax cultivar description. Comput Electron Agric 61(2):126–135 Demartini M et al (2018) Food industry digitalization: from challenges and trends to opportunities and solutions. IFAC-PapersOnLine 51(11):1371–1378 ElMasry G et al (2009) Investigating the effect of moisture content on some properties of peanut by aid of digital image analysis. Food Bioprod Process 87(4):273–281 FAO (2020) Realizing the potential of digitalization to improve the agri-food system: proposing a new International Digital Council for Food and Agriculture. A concept note. Rome Garitta L et al (2013) Sensory analysis of broccoli over time: consumer defined critical attributes and evaluation of digital photographs in comparison to real product appearance. Food Qual Prefer 29(1):48–52 Godoi FC et al (2016) 3d printing technologies applied for food design: status and prospects. J Food Eng 179:44–54 Guillaume S et al (1996) Characterization of mill products by analysis of in-flow digitalized images. J Food Eng 27(3):311–322 Hao L et al (2010a) Material characterisation and process development for chocolate additive layer manufacturing. Virtual Phys 5(2):57–64 Hao L et al (2010b) Extrusion behavior of chocolate for additive layer manufacturing. Innovative developments in design and manufacturing – advanced research in virtual and rapid prototyping, pp. 245–250 Jayaprakash S et al (2019) Chapter 11 – prosumer-driven 3D food printing: role of digital platforms in future 3D food printing systems. In: Godoi FC, Bhandari BR, Prakash S, Zhang M (eds) Fundamentals of 3D food printing and applications. Academic, pp 331–354 Kruth JP et al (2007) Consolidation phenomena in laser and powder-bed based layered manufacturing. CIRP Ann 56(2):730–759 Lee K et al (2019) Colorimetric array freshness indicator and digital color processing for monitoring the freshness of packaged chicken breast. Food Packag Shelf Life 22:100408

67 Food Industry: Applications of Digitalization

1749

Lei S et al (2020) Absolute quantification of Vibrio parahaemolyticus by multiplex droplet digital PCR for simultaneous detection of tlh, tdh and ureR based on single intact cell. Food Control 114:107207 Li Q et al (2014) Application of digital image analysis method to study the gelatinization process of starch/sodium chloride solution systems. Food Hydrocoll 35:392–402 Li Q et al (2015) The influence of different sugars on corn starch gelatinization process with digital image analysis method. Food Hydrocoll 43:803–811 Lipton J et al (2010) Mutlimaterial food printing with complex internal structure suitable for conventional post-processing. 21st annual international solid freeform fabrication symposium – an additive manufacturing conference, pp 809–815 Maniglia BC et al (2021) Pulsed electric fields (PEF) treatment to enhance starch 3D printing application: effect on structure, properties, and functionality of wheat and cassava starches. Innovative Food Sci Emerg Technol 68:102602 MaryAnne MG (2018) Digitalization, Digitization, and Innovation. Research-Technology Management 61(4):56–59 McNamara C (2017) Digitalization: the future of food and beverage. Food Processing Mohamed AA, Shalaby AA (2019) Digital imaging devices as sensors for iron determination. Food Chem 274:360–367 Murphy SV, Atala A (2014) 3D bioprinting of tissues and organs. Biotech 32(8): 773–785 Pant A et al (2021) 3D food printing of fresh vegetables using food hydrocolloids for dysphagic patients. Food Hydrocoll 114:106546 Paquet-Durand O et al (2012) Monitoring baking processes of bread rolls by digital image analysis. J Food Eng 111(2):425–431 Periard D et al (2007) Printing food. 18th solid freeform fabrication symposium, pp 564–574 Pichawadee K, Kim HT (2020) A framework for food supply chain digitalization: lessons from Thailand. Prod Plan Control 31(2-3):158–172 Pulatsu E et al (2021) Effects of ingredients and pre-heating on the printing quality and dimensional stability in 3D printing of cookie dough. J Food Eng 294:110412 Quevedo R et al (2009) Quantification of enzymatic browning kinetics in pear slices using nonhomogenous L∗ color information from digital images. LWT Food Sci Technol 42(8):1367– 1373 Rando P, Ramaioli M (2021) Food 3D printing: effect of heat transfer on print stability of chocolate. J Food Eng 294:110415 Reile CG et al (2020) Qualitative and quantitative analysis based on digital images to determine the adulteration of ketchup samples with Sudan I dye. Food Chem 328:127101 Rogers DL (2016) The digital transformation playbook: rethink your business for the digital age. Columbia University Press Sapirstein HD et al (1987) An instrumental system for cereal grain classification using digital image analysis. J Cereal Sci 6(1):3–14 Schaefer D, Cheung WM (2018) Smart packaging: opportunities and challenges. Procedia CIRP 72:1022–1027 Shafaei SM et al (2019) Development and implementation of a human machine interfaceassisted digital instrumentation system for high precision measurement of tractor performance parameters. Eng Agric Environ Food 12(1):11–23 Shahin MA et al (2006) Determining soya bean seed size uniformity with image analysis. Biosyst Eng 94(2):191–198 Shenoy P et al (2014) Investigation of the application of digital colour imaging to assess the mixture quality of binary food powder mixes. J Food Eng 128:140–145 Sidorenko EL, van Arx P (2020) Transformation of law in the context of digitalization: defining the correct priorities. Digit Law J 1(1):24–38 Silva AFS, Rocha FRP (2020) A novel approach to detect milk adulteration based on the determination of protein content by smartphone-based digital image colorimetry. Food Control 115:107299

1750

V. M. Karizaki

Silva KA et al (2010) Evaluation of aging mechanisms of olive oil–lemon juice emulsion through digital image analysis. J Food Eng 97(3):335–340 Sima H (2017) Digital food marketing: what we know, what we should know? Br J Market Stud 5(8):12–26 Singkhonrat J et al (2019) Digital image colorimetric analysis for evaluating lipid oxidation in oils and its emulsion. Food Chem 286:703–709 Sun J et al (2018) Extrusion-based food printing for digitalized food design and nutrition control. J Food Eng 220:1–11 Tania L, Michelle P (2018) Food/media: eating, cooking, and provisioning in a digital world. Commun Res Pract 4(3):207–211 Tanmay K (2019) Digital innovations in the food industry. Burger King India Ta´nska M et al (2005) Measurement of the geometrical features and surface color of rapeseeds using digital image analysis. Food Res Int 38(7):741–750 Tao Y et al (2020) High-performance detection of Mycobacterium bovis in milk using digital LAMP. Food Chem 327:126945 Tomita H et al (2019) Development of the visualization and quantification method of the rice soaking process by using the digital microscope. J Food Eng 243:33–38 Wang S et al (2013) Digital bread crumb: creation and application. J Food Eng 116(4):852–861 WHO (2019) Digitalization, food safety and trade. WHO Xu L et al (2020) Impact of thermal treatment on the rheological, microstructural, protein structures and extrusion 3D printing characteristics of egg yolk. Food Hydrocoll 100:105399 Yam KL, Papadakis SE (2004) A simple digital imaging method for measuring and analyzing color of food surfaces. J Food Eng 61(1):137–142 Zeng X et al (2020) Insights into the relationship between structure and rheological properties of starch gels in hot-extrusion 3D printing. Food Chem 342:128362 Zhou T et al (2004) Determination of acceptability and shelf life of ready-to-use lettuce by digital image analysis. Food Res Int 37(9):875–881

Detection of Tuberculosis and Lung Cancer Using CNN

68

S. N. Hankare and S. S. Shirguppikar

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Description of Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion and Working of Various Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1752 1753 1754 1755 1759 1760 1761

Abstract

Tuberculosis (TB) is contagious disease spread through air from bacteria called M. tuberculosis. This disorder has high risk factor in humans. It also causes a high risk of lung cancer. Pulmonary tuberculosis and lung cancer share same clinical as well as radiological features. This imitation of symptoms between tuberculosis and lung cancer may lead radiologist to misdiagnose one for the other. This chapter discusses the Machine Learning (ML) tools which is to be used in the detection of tuberculosis and lung cancer. The two main ML tools from deep learning networks in convolutional neural network (CNN) we used are the VGG16 and VGG19 (ILSVRC Runner-up 2014). The datasets of CT

Md Badrul Alam Miah, Mohammad Abu Yousuf, “Detection of Lung cancer from CT image using image processing network”, 2015 International Conference on Electrical Engineering and Information Communication Technology (ICEEICT), May 2015 S. N. Hankare · S. S. Shirguppikar () Department of Mechanical Engineering, Rajararambapu Institute of Technology, Rajaramnagar, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_134

1751

1752

S. N. Hankare and S. S. Shirguppikar

scan and X-rays of tuberculosis and lung cancer has been taken from Kaggle Data Science to train an Algorithm differentiating between a tuberculosis patient and a lung cancer patient. We achieve an accuracy of 71.2% to 74.3% by VGG16 and 89.2% to 97.6% by VGG19. The proposed work illustrates the possibility of the proposed technique in helping the radiologists to differentiate between a tuberculosis patient and a lung cancer patient. Keywords

Convolutional neural network · Deep learning · Tuberculosis detection · Lung cancer detection · Machine learning

Introduction Tuberculosis is an infectious disease. If it is not treated in early stages, the risk of spreading the disease to other body parts like spine and brain increases. Nearly ten million individuals were diagnosed with tuberculosis in 2019 out of which 1.4 people were died (Global Tuberculosis report 2019). The number of cases of tuberculosis are even worse in 2017, 10.4 million patients were reported positive for TB out of which 1.7 million were reported dead (Global Tuberculosis Report 2017). It is registered as top 10 most dangerous diseases having a death rate of 4000 lives each day (Global Tuberculosis Report 2020). While, cancer is the second major disease causing most number of deaths globally, lung cancer alone has 25% deaths of all the cancer deaths. Lung cancer leads to divide the cells in lungs uncontrollably. This phenomenon increases the growth of tumors which can cause severe breathing problems. In 2018, 9.6 million deaths occurred around the globe out of which 2.09 people died due to lung cancer alone (WHO Report 2018). Tuberculosis and lung cancer both share similar prodrome as well as the initial procedure of diagnosis of the disease is also same. As a result, it becomes hard for radiologists to differentiate between these two cases. Chest radiography images (X-Ray) and computerized tomography scans (CT scan) usually being the first test used for diagnosis of tuberculosis as well as for lung cancer (Hussein et al. n.d.). This may lead doctors to misdiagnose the tuberculosis with lung cancer. However, early diagnosis of these vicious diseases will lead to providing necessary treatment and will reduce the death cases around the world. The Deep Learning or Deep Neural Network is an Artificial Neural Networks (ANN) which contains various different layers between input and output layers. In the past few years, it has been considered to be one of the most powerful tools due to its ability to handle a substantial amount of data. The idea of getting hidden layers recently began to become popular over classical methods and has a great future in various fields. In recent years, deep learning methodologies have accomplished remarkable outcomes in the field of machine learning (Schmidhuber 2015). Convolutional Neural Networks (CNN) (LeCun et al. 2010) has made its mark in image classification

68 Detection of Tuberculosis and Lung Cancer Using CNN

1753

applications such as face detection (Li et al. 2015; Farfade et al. 2015), image-driven autonomous cars (Huval et al. 2015), morphology prediction (Dieleman et al. 2015), video classification tasks (Karpathy et al. n.d.), etc. Due to such advancement in Technology especially in the sector of artificial intelligence grant an advantage to researchers to study the automatic diagnosis possibilities of certain diseases. The automatic detection and diagnosis of both the disorders by differentiating them may come really handy for doctors to move further. This speed up the process of diagnostics and lowers its cost. The classification between tuberculosis and lung cancer have been done by very few researchers.

Related Work ANN (Artificial Neural Network) methods can be used for nodule detection (Chen et al. 2015; Akram et al. 2015). Promising results have been shown by CNN in lung nodule detection [35]. Researchers can build their own CNN architecture or there are Ready to be used [36]. A. Teramoto et al. (Teramoto et al. 2016) developed a method for reducing false positive cases. Authors used 104 PET/CT for pulmonary nodule detection with Convolutional Nueral Networks (CNNs). The results show sensitivity of 90.1% eliminating almost half of the False positives of Previous studies with only 4.9 false positives/case. According to S. Antani (Antani 2015), tuberculosis detection can be possible by detecting specific patterns in the chest radiographs. Some of these patterns are opacity in lobes, military patterns, enlargement of lymph nodes, plural effusion and airways enlargement. Presence of these patterns can be helpful for detection of tuberculosis. Also, Patients infected with other pulmonary infections also can have similar patterns that a Tuberculosis infected person (Daley et al. 2003). J Melendez et al. (Melendez et al. 2015) developed a tuberculosis detection method. In his proposal, every radiograph is separated into instances and extracted its features based on pixel intensity distributions. As the standard formulation of SVM was unable to classify radiographs at instance level the researchers proposed a new reformulated SVM (MiSVM) (Andrews et al. 2003). This allows them to classify groups of samples rather than just individual isolated samples. The three datatsets evaluated by authors were Zambia, Tanzania and Gambia. The obtained results showed promising results with an AUC in the range 0.86 to 0.91. B. Van Ginneken et al. (Van Ginneken et al. 2002) was one of the first researchers to invent a CAD (Computer Aided Diagnosis) for Tuberculosis detection. The features of the chest images were extracted with the help of multiscale filter bank. The classification is done with the help of weighted nearest-neighbor scheme. For validation purpose LOOCV is used. The results were taken out for two different small datasets giving AUC of 0.86 and 0.82. L. Hogeweg et al. (Hogeweg et al. 2010) proposed a model for TB detection having combination of pixel-level textural abnormality analysis with other techniques. Results showed an accuracy of AUC from 0.67 to 0.86.

1754

S. N. Hankare and S. S. Shirguppikar

J.H. Tan et al. (Tan et al. 2012) introduced a method for statistical details of pixel distribution is given as input. Test were carried out in a small dataset which gave best accuracy of 0.949. Jaeger et al. (Jaeger et al. 2014) combined different algorithms for extraction of features associated with radiographic images. This experimentation is into various parts – Segmentation (Stirenko et al. n.d.; Candemir et al. 2012), feature extraction (Dalal and Triggs 2005), and classification. Segmentation is done to divide the data into various parts and remove unnecessary data associated with image (Candemir et al. 2012). Feature extraction is done by combining various algorithms like oriented gradient histograms (Dalal and Triggs 2005), local binary patters (Ojala et al. 2002), hu moments (Hu 1962), etc. (Tamura et al. 1978; Howarth and Ruger 2005). This data is fed to binary classifier for classification of image as healthy and unhealthy. Support vector Machine (SVM), Multiplayer perceptron (MLP), logistic regression, and decision trees were used as classifiers of which logistic regression and linear SVM has given best results. For Montgomery dataset the value of AUC was 0.87 with an accuracy of 78.3% and for Shenzhen dataset the AUC value obtained is 0.9 with accuracy of 84%. Chan and Vese et al. (Chan 2001) investigated a curve evolution method for detection and segmentation of lungs. It is used for energy minimization based segmentation. Kumar et al. (Kumar et al. 2015) proposed a CAD system with the help of deep features taken out from autoencoder. It is used for classification of lungs into malignant or benign. J. Tan et al. (Tan et al. 2017) invented a method for lung nodule detection with significant amount of reduction in false positive cases with the help of CNN and DNN. The dataset was comprised of 85 patients. The final results give a sensitivity of 82% and the number of false positives was reduced by 0.329 with DNN. R. Golan et al. (Golan et al. 2016) designed a framework to detect lung nodules in the CT scan data. The framework trains the weights with the help of back propagation. The results showed 78.9% sensitivity with 20 being false positive. (c) shows Lung cancer cases

Description of Dataset In this study, the dataset taken consist of 5932 frontal chest X-ray images provided form Kaggle and non-disclosed local Hospitals or Covid-19 Wards. The images in the dataset are of resolutions 96dpi whereas the varying dimensions such as 728x368 to 2746x2382. There are 1040 normal case, 3808 tuberculosis case 1062 lung cancer images in the dataset. Also, for organizing the dataset optimization techniques were also used (Phan et al. 2021; Shirguppikar and Dabade 2018; Shirguppikar et al. 2020). X-ray samples from the dataset have been shown in Fig. 1 and Table 1 shows the dispensation training data, validation data, and testing data phases of the model introduced in this chapter.

68 Detection of Tuberculosis and Lung Cancer Using CNN

1755

Fig. 1 Data samples from the dataset, (a) shows normal cases and (b) shows tuberculosis cases Table 1 Distribution of dataset

Normal Lung Cancer Tuberculosis

Train 1040 3808 1062

Validation 305 390 92

Test 305 390 92

Discussion and Working of Various Architectures A. Image Processing and Neural Network [35]: The proposed system focuses on detecting lung cancer from a CT image, which is done automatically. The method for detecting lung cancer is dependent on machine vision. This thesis’ key contribution is the development of a rotation, scaling, and

1756

S. N. Hankare and S. S. Shirguppikar

Image Prepossessing

Binarization of CT Image

Y

#B>Thresh1

Abnormal Full Lung

N

Input Lung CT Image

Segmentation

N

Normal or Abnormal left/ right Lung

Neural Network Recognition

Segmented Right Lung

Segmented Left Lung

Binarization of Right Lung

Binarization of Left Lung

#B>Thresh2

#B>Thresh3

Y Abnormal Right Lung

Y Abnormal Left Lung

N

Feature Extraction

Fig. 2 Lung Cancer and Tuberculosis Detection System Architecture [35]

translation invariant feature extraction system for lung cancer identification. Md Badrul Alam Miah et al. [35] established the overall lung cancer identification device architecture seen in Fig. 2. and the same architecture will be used for Detection of Tuberculosis. The whole lung cancer screening procedure is broken down into the steps below. Neural Network Detection, Binarization, Segmentation, Image Preprocessing. I. Image Processing: The image preprocessing steps are applied to the images. The block diagram of image preprocessing steps is seen in Fig. 3. II. Binarization: Image binarization is a method for converting a grayscale image to a black-and-white image. Binary images are also known as two-level or bilevel images. This means that each pixel is stored as a single bit, i.e., a 1 or a 0, as seen in Fig. 4. III. Segmentation: Image segmentation is the method of partitioning a visual image into several segments in a computer vision device. The aim of segmentation is to make an image more coherent and easier to interpret by simplifying and/or changing its representation. Points and borders (lines, circles, etc.) in

68 Detection of Tuberculosis and Lung Cancer Using CNN

1757

Fig. 3 Block Diagram of image preprocessing

photographs are usually located using image segmentation. To be more specific, the method of assigning a label to each pixel in an image such that pixels with the same label share certain characteristics is known as image segmentation. In the proposed system, segmentation processes consist of different steps. At first convert the original grayscale image into edge only image shown in Fig. 5. Then, as seen in Fig. 6, transform the edge-only image to a dilated image and then to a filled image. Finally, the picture in Fig. 7 is segmented into Right Lung and Left Lung. IV. Neural Network Detection: Following the Thresholding process, the majority of the Lung Cancer and Tuberculosis Detection System employs a neural network, which is a highly efficient and effective method. Following the feature extraction process, the features are fed into a neural network, which is used to train the algorithm for classification or identification. Image Acquisition, Image Preprocessing, Segmentation, Feature Extraction, and Neural Network Classification are the measures that make up the proposed training method for lung cancer and tuberculosis detection. The suggested method’s outcome is achieved in two ways: Binarization Technique and Neural Network Binarization Technique. Binarization Technique provides a 99% (approximately) accurate result for this system. Table 2 shows the preliminary effects of the system’s neural network.

1758

Fig. 4 (a) Gray scale image, (b) Binarized image

Fig. 5 (a) Original Grayscale image, (b) Edge Only

Fig. 6 (a) Dilated image, (b) Filled image

S. N. Hankare and S. S. Shirguppikar

68 Detection of Tuberculosis and Lung Cancer Using CNN

a

1759

b

20

20

40

40

60

60

80

80

100

100

120

120

140

20 20

40

60

80

40

60

80

100

Fig. 7 (a) Segmented right Lung, (b) Segmented Left Lung Table 2 Total Detection rate of the system Feature Normal Lung Cancer Tuberculosis

Total Images 305 390 92

Correct detection (%) 98.63 95.33 87.31

Error (%) 1.37 4.67 12.69

Conclusion Lung cancer and tuberculosis are two deadly illnesses that must be detected at an early stage. Lung cancer and tuberculosis, on the other hand, are very difficult to diagnose. According to the literature review, a variety of methods are used to diagnose lung cancer and tuberculosis, but each has its own set of limitations. Our proposed method uses methods in which binary thresholding is done first, followed by feature extraction, and then these features are used to train and evaluate the neural network. From CT scan and X-ray scans, the proposed device successfully detects lung cancer and tuberculosis. By the end of the day, the system can be seen to have met its objectives. The proposed system tested 150 different types of lung CT images and came up with a result of 96.67% total system performance, which meets the system’s expectations. This procedure may be used to identify brain tumors, breast cancer, and other cancers in the future.

1760

S. N. Hankare and S. S. Shirguppikar

References Akram S, Javed MY, Qamar U, Khanum A, Hassan A (2015) Artificial neural network based classification of lungs nodule using hybrid features from computerized tomographic images. Appl Math Inform Sci 9(1):183–195 Andrews S, Tsochantaridis I, Hofmann T (2003) Support vector machines for multipleinstance learning. Adv Neural Inf Process Syst:561–568 Antani S (2015) Automated detection of lung diseases in chest X-rays, Technical Report to the LHNCBC Board of Scientific Candemir S, Jaeger S, Palaniappan K, Antani S, Thoma G (2012) Graph-cut based automatic lung boundary detection in chest radiographs. IEEE Healthcare Technol Conf Transl Eng Health Med:31–34 Chan TF (2001) Vese, ‘Active contours without edges’. IEEE Trans Image Process:266–277 Chen J, Chen J, Ding H-Y, Pan Q-S, Hong W-D, Xu G, Yu F-Y, Wang Y-M (2015) Use of an artificial neural network to construct a model of predicting deep fungal infection in lung cancer patients. Asian Pac J Cancer Prev 16(12):5095–5099 Dalal N, Triggs B (2005) Histograms of oriented gradients for human detection. In: Computer Vision and Pattern Recognition. CVPR 2005, IEEE Computer Society Conference on, vol 1, IEEE, pp 886–893 Daley CL, Gotway M, Jasmer R (2003) Radiographic manifestation of tuberculosis. A Primer Clin 1 Dieleman S, Willett KW, Dambre J (2015) Rotation-invariant convolutional neural networks for galaxy morphology prediction. Mon Notices R Astron Soc 450:1441–1459 Farfade SS, Saberian MJ, Li L-J (2015) Multi-view face detection using deep convolutional neural networks. In: Proceedings of the 5th ACM on international conference on multimedia retrieval. ACM, pp 643–650 Global Tuberculosis Report 2017, 2017 Global Tuberculosis report 2019, 2019 Global Tuberculosis Report 2020, 2020 Golan R, Jacob C, Denzinger J (2016) Lung nodule detection in ct images using deep convolutional neural networks. In 2016 International Joint Conference on Neural Networks (IJCNN), pp 243–250 Hogeweg L, Mol C, de Jong PA, Dawson R, Ayles H, van Ginneken B (2010) Fusion of local and global detection systems to detect tuberculosis in chest radiographs. In: Medical image computing and computer-assisted intervention–miccai 2010. Springer, pp 650–657 Howarth P, Ruger S (2005) Robust texture features for still-image retrieval. IEE Proc Vision Image Signal Process 152:868–874 Hu M-K (1962) Visual pattern recognition by moment invariants. IRE Trans Inf Theory 8:179–187 Hussein S, Cao K, Song Q, Bagci U (n.d.) Risk stratification of lung nodules using 3D CNN-based multi-task learning. Center for Research in computer vision (CRCV) at University of Central Florida, Orlando, FL Huval B, Wang T, Tandon S, Kiske J, Song W, Pazhayampallil J, Andriluka M, Cheng-Yue R, Mujica F, Coates A et al 2015 An empirical evaluation of deep learning on highway driving, arXiv Prepr. arXiv:1504.01716 Jaeger S, Karargyris A, Candemir S, Folio L, Siegelman J, Callaghan F, Xue Z, Palaniappan K, Singh RK, Antani S et al (2014) Automatic tuberculosis screening using chest radiographs, medical imaging. IEEE Trans 33:233–245 Karpathy A, Toderici G, Shetty S, Leung T, Sukthankar R, Fei-Fei L (n.d.) Large-scale video classification with convolutional neural networks. In: Proceedings of the IEEE Kumar D, Wong A, Clausi DA (2015) Lung nodule classification using deep features in ct images. In: 2015 12th conference on computer and robot vision, pp 133–138

68 Detection of Tuberculosis and Lung Cancer Using CNN

1761

LeCun Y, Kavukcuoglu K, Farabet C (2010) Convolutional networks and applications in vision. In: Proceedings of the IEEE international symposium on circuits and systems (ISCAS). IEEE, pp 253–256 Li H, Lin Z, Shen X, Brandt J, Hua G (2015) A convolutional neural network cascade for face detection. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 5325–5334 Melendez J, van Ginneken B, Maduskar P, Philipsen RH, Reither K, Breuninger M, Adetifa IM, Maane R, Ayles H, Sanchez CI (2015) A novel multipleinstance learning-based approach to computer-aided detection of tuberculosis on chest x-rays, medical imaging. IEEE Trans 34: 179–192 Ojala T, Pietikainen M, Maenpaa T (2002) Multiresolution gray-scale and rotation invariant texture classification with local binary patterns. IEEE Trans Pattern Anal Mach Intell 24:971–987 Phan, N.H., Van Dong, P., Dung, H.T. Muthuramalingam T, Shirguppikar S, Tam NC, Ly NT (2021) Multi-object optimization of EDM by Taguchi-DEAR method using AlCrNi coated electrode. Int J Adv Manuf Technol 116(12):1429–1435. https://doi.org/10.1007/s00170-02107032-3 Schmidhuber J (2015) Deep learning in neural networks: an overview. Neural Netw 61:85–117 Shirguppikar S, Dabade U (2018) Experimental investigation of dry electric discharge machining (Dry EDM) process on bright mild steel. Materials Today: Proceedings 5(2):7595–7603 Shirguppikar SS, Patil MS, Vinayak G (2020) Grey Fuzzy multiobjective optimization of process parameters by dry electro discharge machining process. Materials Today: Proceedings 27: 671–676 Stirenko S, Kochura Y, Alienin O, Rokovyi O, Gordienko Y (n.d.) Chest X-ray analysis of tuberculosis by deep learning with segmentation and augmentation Tamura H, Mori S, Yamawaki T (1978) Textural features corresponding to visual perception. IEEE Trans Syst Man Cybern 8:460–473 Tan JH, Acharya UR, Tan C, Abraham KT, Lim CM (2012) Computer-assisted diagnosis of tuberculosis: a first order statistical approach to chest radiograph. J Med Syst 36:2751–2759 Tan J, Huo Y, Liang Z, Li L (2017) A comparison study on the effect of false positive reduction in deep learning based detection for juxtapleural lung nodules: Cnn vs dnn. In: Proceedings of the Symposium on Modeling and Simulation in Medicine, MSM ‘17. Society for Computer Simulation International, San Diego, pp 8:1–8:8 Teramoto A, Fujita H, Yamamuro O, Tamaki T (2016) Automated detection of pulmonary nodules in PET/CT images: ensemble false-positive reduction using a convolutional neural network technique. Med Phys 43(6):2821–2827 Van Ginneken B, Katsuragawa S, ter Haar Romeny BM, Viergever MA et al (2002) Automatic detection of abnormalities in chest radiographs using local texture analysis, medical imaging. IEEE Trans 21:139–149 WHO Report on Cancer, September 2018

Bibliography https://www.cancer.org/cancer/lung-cancer/about/what-is.html https://www.cancer.gov/types/lung https://www.kaggle.com/datasets https://medlineplus.gov/tuberculosis.html https://www.who.int/news-room/fact-sheets/detail/tuberculosis

Covid-19 or Viral Pneumonia Detection Using AI Tools

69

M. V. Pachore and S. S. Shirguppikar

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Description of Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion and Working of Various Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CovXNet (Mahmud et al. 2020) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modified AlexNet Network (Chen et al. 2020) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1764 1766 1767 1767 1769 1771 1772

Abstract

Viral pneumonia is a disease which occurs in lungs due to bacterial infection. Since middle of December 2019, many cases of pneumonia with unknown cause were found in Wuhan City, China; at present, it has been confirmed that it is a new respiratory disorder caused due to coronavirus infection. Lungs abnormality is highly risky condition in humans; the reduction of the risk is done by enabling quick and efficient treatment. The Covid-19 pneumonia is mimicking viral pneumonia, that is, their symptoms are undistinguished. Lung’s abnormality is detected by Computed Tomography (CT) scan images or X-ray images. By viewing the X-rays or CT scan images, even for a well-trained radiologist, it is difficult to detect Covid-19/viral pneumonia. For quick and efficient treatment, it is necessary that proper detection must take place and during this epidemic situation, late detection can lead to doubling of cases; hence, there is a need of proper tool for quick detection of Covid-19/viral pneumonia. This chapter is

M. V. Pachore · S. S. Shirguppikar () Department of Mechanical Engineering, Rajararambapu Institute of Technology, Rajaramnagar, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_136

1763

1764

M. V. Pachore and S. S. Shirguppikar

discussing various AI tools for quick detection as a part of our contribution for quick detection and cure of Covid-19 to front line corona worriers and safety of viral pneumonia patients from Covid-19. The two AI tools are from deep learning (DL), that is, Convolutional Neural Networks (CNN) and Recurrent Neural Network (RNN), which are used for the detection of Covid-19/viral pneumonia. The algorithm is trained using available X-ray images of health lungs, viral pneumonia-affected lungs, and Covid-19-affected lungs available through Kaggle and nondisclosed local hospitals or Covid-19 wards. Also transfer learning method is also used for long-lasting validation of the model. The results give us an accuracy for CNN 83.2 to 94.1% results which are also matched with practically tested positive Covid-19 patients using swab tests by doctors. After testing the various models, we also came through that every model of DL has its own specialty. Keywords

Covid-19 · Viral pneumonia · Computed Tomography (CT) · X-ray image · Deep learning (DL) · Convolutional Neural Network (CNN) · Recurrent Neural Network (RNN) · Transfer learning

Introduction Many cases of pneumonia were suddenly found in Wuhan, China, with unknown cause which has been later considered as an acute respiratory infectious disease caused by coronavirus until the February 12, 2020, after the announcement of official calcifications of new coronavirus, that is, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) by the International Committee on Taxonomy of Viruses (Sohrabi et al. 2020; Lai et al. 2020; Gorbalenya et al. 2020). The alarming increase in mortality rates throughout the world hence made WHO to call out a global pandemic which has crumbled the healthcare organizations throughout the globe leading to high demand in diagnosis and prevention against the increasing spread of the disease (Rothan and Byrareddy 2020). The Covid-19 symptoms detected from patients are so varying of which fever, dry cough, tiredness are being common and sore throat, headache, acute respiratory distress syndrome (ARDS) which are almost similar to the symptoms of viral pneumonia (Bai et al. 2020). Since the finding of new mutation of Covid-19 in UK has created a great risk to mankind (Dawood 2020). Both reverse transcription-polymerase chain reaction (RT-PCR) and SWAB diagnosis test used for Covid-19 have a low sensitivity and large test period. Furthermore, the test is only carried out in specific labs and test samples are transported to these labs, which causes more time consumption and the limited supply of expensive test kit (European Center for Disease prevention and Control 2020) is making the situation even worse.

69 Covid-19 or Viral Pneumonia Detection Using AI Tools

1765

Hence, chest X-rays and CT scans have been provided to all individuals having positive symptoms of pneumonia to diagnose and isolate patients. As a serious shortage of experts to differ patients as having large similarities of Covid-19 and traditional pneumonia, an artificial intelligence (AI) assists which automatically detects the patient’s infection can be a significant change for filtration of Covid19 patients form traditional pneumonia and separate isolation of individuals. Also, this can also help in deciding the priority order of the testing samples of the patients. Today, much research is being conducted in the area of deep learning, and convolutional neural networks (CNN) have shown considerable promise in image recognition and classification (Krizhevsky et al. 2012). The central idea behind CNNs is to create an artificial replica of the human brain’s visual cortex. It has the potential to remove more important features from the whole picture rather than handcrafted features, the key benefit of CNNs (Krizhevsky et al. 2012; Xie et al. 2020). Various CNN-based deep networks were developed by researchers and these networks achieved state-of-the-art results in computer vision recognition, segmentation, target detection, and localization (Krizhevsky et al. 2012; Ronneberger et al. 2015). In addition to natural machine vision issues, CNNs have obtained very good outcomes in the resolution of medical problems such as identification of breast cancer (Ragab et al. 2019), segmentation of brain tumors (Pereira et al. 2016), diagnosis of Alzheimer’s disease, recognition of skin lesions (Esteva et al. 2017; Ayan and Ünver 2018). Detailed articles on deep learning in medical image processing are presented here (Ker et al. 2018; Litjens et al. 2017). As far as we are concerned, we are aware that there are a few experiments using deep learning to classify pneumonia. Antin et al. used a DenseNet-121 sheet in 2017 using the transition leaning technique and obtained 0.60% area under the curve value (AUC) (Antin et al. 2017). A 121-layer convolutionary neural network based on DenseNet (Huang et al. 2017) was proposed by Rajpurkar et al. in 2017 and named as CheXNet (Rajpurkar et al. 2017). They trained their network of 10,000 frontal view chest X-ray images containing 14 different diseases. They calculated the performance of their network with four expert radiologists using the f1 score metric, which is the harmonic average of the accuracy and recall metrics. CheXNet received an f1 score of 0.435 (95 percent CI 0.387, 0.481), which was higher than the radiologist’s average of 0.387 (95 percent CI 0.330, 0.442). On the other hand, RNN architecture-based methods have also been put forward. LSTM (Rajpurkar et al. 2017), which is capable of learning about long-term dependencies, is the most popular RNN. Usually, Xu et al. (Zhou et al. 2018) suggested a model with SDP knowledge introduced using long short-term memory networks (LSTM) to help explore essential text constructs for classification of relationships. Cai et al. (Xu et al. 2015) also introduced LSTM with SDP encoding into neural networks without taking SDP information into account as a standard function, allowing their model to outperform all previous models on the SemEval2010 task 8 dataset. Zhang and Wang (Zhang et al. 2018), on the other hand,

1766

M. V. Pachore and S. S. Shirguppikar

used a simple RNN-based model and outperformed the KBP37 dataset of CNN models. Although these strategies make use of automatic RNN feature extraction, their efficiency is limited because the information obtained by RNN does not often contain local features. Some of the other works are based on a mixture of CNN and RNN to do the role of relationship classification, aside from the approaches described above. And it was also possible to benefit from certain sentence classification work (Hassan and Mahmood 2018). Rotsztejn et al. (2018), who introduced a method of classification of relationships based on an ensemble of CNNs and RNNs, suggested a relatively representative approach. Moreover, there are works that are also common based on attention (Li et al. 2018). While their solution performs better on three of the four subtasks on SemEval-2018 challenge 7, ensemble learning methods are more nuanced and require much more time and energy to compute than single models. Since the size of a dataset is small, ensemble models can take much longer to adjust parameters, and it may also be easier to overfit. Also, to organize the data some optimization techniques were also referred (Shirguppikar and Dabade 2018; Shirguppikar and Patil 2020; Nguyen Huu et al. 2021). Our ultimate aim is to develop a simple, single deep learning model that can easily and effectively differentiate relationship types without the need to train multiple models.

Description of Dataset To test this approach, the dataset consists of collection of five various datasets of 210 X-ray and 398 CT images of Covid-19. The distribution of collected dataset is shown in Tables 1 and 2. These datasets are unique as they are collected from different sources and from different parts of the world which is important for building an advanced tool to help medical experts diagnosing Covid-19 all over the world. Also, the images from dataset are openly available for researchers and public (Cohen et al. 2020; Covid-19 bsti imaging database 2020; Hacking and Bickle 2020; Mooney 2020). The images used in this will collectively be available in a GitHub repository (Cohen et al. 2020). The given data in Table 1 clearly shows that we do not have enough data of Covid-19 images which is publicly available. This makes hard for research communities to conduct full investigation and the need of more images from radiology is high (Fig. 1). Table 1 Covid-19 X-ray and CT scan images and their source

Covid-19 X-ray CT

GitHub 70 16

BSTI 15 187

Total 85 203

69 Covid-19 or Viral Pneumonia Detection Using AI Tools Table 2 Normal X-ray and CT scan images and their source

Normal X-ray CT

1767 Kaggle 206 42

Radiopedia 89 153

Kaggle

Radiopedia and GitHub

BSTI

Radiopedia

Total 295 195

X-ray

CT

Fig. 1 X-ray and CT scan Images from different Sources

Discussion and Working of Various Architectures CovXNet (Mahmud et al. 2020) The schematic representation of the workflow is given in Fig. 2. The similarities between traditional pneumonia and pneumonia caused by Covid-19 are very high from clinical and physiological views (Bai et al. 2020; Chung et al. 2019); transferring information acquired from chest X-rays of traditional pneumonia patients can be an effective way to utilize smaller Covid-19 X-rays for extracting additional features. Hence, the training phase should have a big dataset containing both X-rays (normal pneumonia and non-Covid pneumonia) and should be used for training CovXNet. Then for optimization of predictions, an algorithm is applied on all the networks through meta-learner. As the convolution layers get optimized to extract

1768

M. V. Pachore and S. S. Shirguppikar

Fig. 2 The workflow of the proposed by Tanvir Mahmud et al. (2020)

Fig. 3 Significant portions of the test X-rays that instigate the decision are localized by imposing the activation heatmap obtained from CovXNet (Mahmud et al. 2020)

69 Covid-19 or Viral Pneumonia Detection Using AI Tools

1769

features from X-rays, weights of layers are directly transferred in the transfer learning phase. After that a dataset having Covid-19 and other pneumonia patients is used to train other layers integrated with the CovXNet. At last, the modified CovXnet is applied to predict X-ray images. A deep neural network is used (CovXNet) for the detection of Covid-19 and different pneumonia types with distinctive localization from chest images as shown in Fig. 3. Depth-wise convolution is used in the CovXNet with changing dilation rates which integrate features from various receptive fields for analyzing abnormalities in the images. For utilization of a few X-rays, large dataset is utilized having x-rays from normal and various traditional pneumonia patients for initially training deep network. Because of the similar characteristics of Covid-19 and pneumonia, we can obtain satisfactory results with the help of smaller datasets containing Covid-19 Xrays by transfer of trained convolutional layers with additional fine-tuning layers. Also, it can be seen that a stacking algorithm provides improved performance by further optimization of prediction obtained from variations of CovXNet. A generated class activation map also provides discriminative localization of abnormal areas which is able to help to diagnose the variations of clinical features of pneumonia on X-ray. Improvement in the performance of these schemes can be done by including more samples of Covid-19 X-rays for training in transfer learning phase. Extensive simulation of experimental results can also be effective to quick diagnostics of Covid-19 and pneumonia patients. Hence, the given CovXNet model has a high scalability with high receptive capacity which can be applied to various computer vision applications.

Modified AlexNet Network (Chen et al. 2020) AlexNet is a pretrained network which referred to CNN family having trained more than million images on ImageNet in a wide range of objects (Maghdid et al. 2004). By modifying the pretrained AlexNet, we are able to transfer the weights, bias, and features for the detection of Covid-19. After this, input datasets (CT scans and X-ray images) are applied on the parameters for training. These transfer learning algorithms lay out perception of how an individual can design new architectures to detect Covid-19. It is due to the fact that training a CNN network from scratch using random weights values would take far longer than tuning a pretrained network. It also requires less computational power especially if the dataset has small number of images. However, in order to use a pretrained network for our project, the model is changed by replacing the last layers with the layers we choose to use. Furthermore, as seen in Fig. 4, our collected datasets were used to train a new network. When the width is equal to 227, the height is equal to 227, and the channel color number is equal to 3, the input images of the dataset are clipped and the sizes of the images are unified according to the AlexNet model. In addition, the updated

Fig. 4 The modified AlexNet network for diagnosing disease Covid-19 (Chen et al. 2020)

1770 M. V. Pachore and S. S. Shirguppikar

69 Covid-19 or Viral Pneumonia Detection Using AI Tools

1771

Fig. 5 Identified all Covid-19 X-ray images correctly, while 23 normal X-rays images out of 25 images (Chen et al. 2020)

network’s minimum batch size has been set to 10, the number of epochs has been set to 20, the learning rate has been set to 3e-4, shuffling has been set at every epoch, and the validation frequency has been set to 3. When checked on CT images of Covid-19 patients, pretrained AlexNet performed well in distinguishing Covid-19 X-ray images from standard X-rays, but not as well as X-ray scans of chest radiographs. Furthermore, unlike traditional CNN, the Pretrained AlexNet is not learned from scratch. The AlexNet transfer learning algorithm, which has been trained, performs exceptionally well in extracting silent features from Covid-19 X-ray images (Fig. 5). As seen in Fig. 5, the enhanced AlexNet accurately diagnoses all Covid-19 X-ray images, thus correctly identifying 23 regular X-ray images out of 25 images, implying a 98% overall accuracy.

Conclusion The latest integrated kernel is demonstrated on a chest X-ray dataset from a multibranch convolution network in this chapter. In a simple network, it is obvious that numerous divisions of convolution are capable of extracting vital data from textural knowledge. It has been discovered that a stacking algorithm improves performance by improving predictions obtained from different versions of CovXNet that are primarily configured for different resolutions. Based on the results of comprehensive simulations, it seems to be a viable option for quicker diagnosis of Covid-19 and other pneumonia patients. Furthermore, the proposed CovXNet is

1772

M. V. Pachore and S. S. Shirguppikar

highly flexible and has a large receptive power, making it suitable for a wide range of other computer vision applications. This research aims to assist radiologists in using a device that can be used on cell phones, particularly because the majority of people nowadays have smartphones with adequate computational capacity. As a result, if a basic CNN model can be built and shown to be accurate, it is simple to evaluate prospective radiography images and does not take a lot of computing resources. It can also be readily tailored to smartphones. Hopefully, the findings of this research would act as a first step toward developing a sophisticated Covid-19 disease detection system from X-ray or CT images as quickly as possible, in order to save as many lives as possible. While the proposed models’ accuracy is insufficient, the effect of the detection may be offset by using other disease signs.

References Antin B, Kravitz J, Martayan E (2017) Detecting pneumonia in chest X-rays with supervised learning. http://cs229.stanford.edu/proj2017/finalreports/5231221.pdf, Corpus ID: 46632050 Ayan E, Ünver HM (2018) Data augmentation importance for classification of skin lesions via deep learning. In: 2018 electric electronics, computer science, biomedical engineering meeting (EBBT). IEEE, pp 1–4 Bai HX, Hsieh B, Xiong Z, Halsey K, Choi JW, Tran TML, Pan I, Shi L-B, Wang D-C, Mei J, Jiang X-L, Zeng Q-H, Egglin TK, Hu P-F, Agarwal S, Xie F-F, Li S, Healey T, Atalay MK, Liao WH (2020) Performance of radiologists in differentiating COVID-19 from non-COVID-19 viral pneumonia at chest CT. Radiology 296(2):E46 Chen J, Wu L, Zhang J, Zhang L, Gong D, Zhao Y, Hu S, Wang Y, Hu X, Zheng B et al (2020) Deep learning-based model for detecting 2019 novel coronavirus pneumonia on high-resolution computed tomography: a prospective study. medRxiv Chung M, Bernheim A, Mei X, Zhang N, Huang M, Zeng X, Cui J, Xu W, Yang Y, Fayad ZA, Jacobi A, Li K, Li S, Shan H (2019) CT imaging features of 2019 novel coronavirus (2019nCoV). Radiology 295(1) Cohen JP, Morrison P, Dao L (2020) Image data collection, vol 6. arXiv 2003.11597 Covid-19 bsti imaging database (2020, March) [Online]. https://www.bsti.org.uk/training-andeducation/covid-19-bsti-imaging-database/ Dawood AA (2020) Mutated COVID-19 may foretell a great risk for mankind in the future. New Microbes New Infect 35 Esteva A et al (2017) Dermatologist-level classification of skin cancer with deep neural networks. Nature 542(7639):115 European Center for Disease prevention and Control (December 2020) Rapid increase of a SARSCoV-2 variant with multiple spike protein mutations observed in the United Kingdom Gorbalenya AE, Baker SC, Baric RS et al (2020) Severe acute respiratory syndrome related coronavirus: the species and its viruses – a statement of coronavirus study group. BioRxiv Hacking C, Bickle I (2020, March) Normal chest imaging examples. [Online]. Available: Radiopaedia, https://radiopaedia.org/articles/normal-chest-imaging-examples?lang=gb Hassan A, Mahmood A (2018) Convolutional recurrent deep learning model for sentence classification. IEEE Access 6:13949–13957 Huang G, Liu Z, Van Der Maaten L, Weinberger KQ (2017) Densely connected convolutional networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition. IEEE, pp 4700–4708 Ker J, Wang L, Rao J, Lim T (2018) Deep learning applications in medical image analysis. IEEE Access 6:9375–9389

69 Covid-19 or Viral Pneumonia Detection Using AI Tools

1773

Krizhevsky A, Sutskever I, Hinton GE (2012) Imagenet classification with deep convolutional neural networks. In: Advances in neural information processing systems. Curran Associates, pp 1097–1105 Lai C-C, Shih T-P, Ko W-C, Tang H-J, Hsueh P-R (2020) Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and corona virus disease-2019 (COVID-19): the epidemic and the challenges. Int J Antimicrob Agents 55:105924 Li N, Zhang H, Chen Y (2018) Convolutional neural network with SDPbased attention for relation classification. In: 2018 IEEE international conference on big data and smart computing (BigComp). IEEE, pp 615–618 Litjens G et al (2017) A survey on deep learning in medical image analysis. Med Image Anal 42:60–88 Maghdid HS, Asaad AT, Ghafoor KZ, Sadiq AS, Khan MK (2004) Diagnosing COVID-19 pneumonia from X-ray and CT images using deep learning and transfer learning algorithms. arXiv.org > eess > arXiv:2004.00038 Mahmud T, Rahman MA, Fattah SA (2020) CovXNet: a multi-dilation convolutional neural network for automatic COVID-19 and other pneumonia detection from chest X-ray images with transferable multi-receptive feature optimization. Comput Biol Med 122:103869 Mooney P (March, 2020) Chest x-ray images (pneumonia). [Online]. https://www.kaggle.com/ paultimothymooney/chest-xray-pneumonia/mtadata Nguyen Huu P, Dong P, Tien D, Van Thien N, Muthuramalingam T, Shirguppikar S, Chi Tam N, Ly N (2021). Multi-object optimization of EDM by Taguchi-DEAR method using AlCrNi coated electrode. In: The International Journal of Advanced Manufacturing Technology. vol 116. Springer, London, pp 1–7 Pereira S, Pinto A, Alves V, Silva CA (2016) Brain tumor segmentation using convolutional neural networks in MRI images. IEEE Trans Med Imaging 35(5):1240–1251 Ragab DA, Sharkas M, Marshall S, Ren J (2019) Breast cancer detection using deep convolutional neural networks and support vector machines. PeerJ 7:e6201 Rajpurkar P et al (2017) Chexnet: radiologist-level pneumonia detection on chest x-rays with deep learning. arXiv preprint arXiv:1711.05225 Ronneberger O, Fischer P, Brox T (2015) U-net: convolutional networks for biomedical image segmentation. In: International conference on medical image computing and computer-assisted intervention. Springer, pp 234–241 Rothan HA, Byrareddy SN (2020) The epidemiology and pathogenesis of coronavirus diseases (COVID-19) outbreak. J Autoimmun 109:102433 Rotsztejn J, Hollenstein N, Zhang C (2018) ETH-DS3Lab at SemEval-2018 task 7: effectively combining recurrent and convolutional neural networks for relation classification and extraction. In: Proceedings of the 12th international workshop on semantic evaluation. Association for Computational Linguistics, pp 689–696 Shirguppikar S, Dabade U (2018). Experimental investigation of dry electric discharge machining (Dry EDM) process on bright mild steel. Materials Today: Proceedings 5:7595–7603. https:// doi.org/10.1016/j.matpr.2017.11.432 Shirguppikar S, Patil MS (2020) Grey fuzzy multiobjective optimization of process parameters for dry electro discharge machining process. Materials Today: Proceedings 27:671–676. https:// doi.org/10.1016/j.matpr.2020.02.234 Sohrabi C, Alsafi Z, O’Neill N, Khan M, Kerwan A, Al-Jabir A, Iosifidis C, Agha R (2020) World health organization declares global emergency: a review of the 2019 novel coronavirus (COVID-19). Int J Surg 76:71 Xie X, Zhao Z, Zheng C, Wang F, Liu J (2020) Chest CT for typical 2019-nCoV pneumonia: relationship to negative RT-PCR testing. Radiology. https://doi.org/10.1148/radiol.2020200343 Xu Y, Mou L, Li G, Chen Y, Peng H, Jin Z (2015) Classifying relations via long short term memory networks along shortest dependency paths. In: Proceedings of the 2015 conference on empirical methods in natural language processing. Association for Computational Linguistics, pp 1785– 1794

1774

M. V. Pachore and S. S. Shirguppikar

Zhang R, Meng F, Zhou Y, Liu B (2018) Relation classification via recurrent neural network with attention and tensor layers. Big Data Mining Anal 1(3):234–244 Zhou JT, Di K, Du J, Peng X, Yang H, Pan SJ, Tsang IW, Qin YLZ, Goh RSM (2018) SC2Net: sparse LSTMs for sparse coding. In: The thirty-second AAAI conference on artificial intelligence (AAAI-18). AAAI Press, pp 4588–4595

Part IV Industry 4.0: Concept of Smart, Intelligent, and Sustainable Society

Iranian Small and Medium-Sized Industries

70

S. Jithender Kumar Naik, Malek Hassanpour, and Dragan Pamucar

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Weighing and Ranking Models Applied in Classification of Iranian Industries . . . . . . Iranian Industries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reviewing Statistical Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classified Iranian Industries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Iranian Future Industries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Websites for Further Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1778 1781 1788 1789 1790 1798 1798 1799 1803

Abstract

The definition of small and medium industries varies from country to country and is governed by economic and industrial circumstances. Some of the criteria used to determine the type of industry (small, medium, and large) are: number of employees, invested budget, total assets, sales, and production quantities, energy consumed, etc. By present review, Iranian industries classified based on five main criteria in various multicriteria decision-making systems according to initial screening of the Iranian evaluator team to confirm and confer installation permissions. Hereby, a list of nine group industries has discussed classification

S. Jithender Kumar Naik University College of Science, Osmania University, Hyderabad, Telangana State, India M. Hassanpour () Department of Environmental science, UCS, Osmania University, Hyderabad, Telangana State, India D. Pamucar Department of logistics, Military Academy, University of Defence, Belgrade, Serbia © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_57

1777

1778

S. Jithender Kumar Naik et al.

stages in weighing and ranking practices along with some statistical analysis. Iranian government encouraged to support the stakeholders to implement modern techniques to generate high value-added industries from waste and raw materials streams. Findings developed a narrow channel to expand the science and will mature a valuable database based on current and future preliminary screening information of evaluator teams prior to the complete appearance of an industrial project on the territory of Iranian country. Therefore, the classification can be deployed to comprise new industries along with executing the energy stream. Keywords

Iranian industries · MCDM · Assessment · Project

Introduction Globally, industries divided into four sections such as micro, small, medium, and large industries. Large industries beset about 2% of Iranian industries proportion with a major role in the industrial revolution in Iran. But below we have another type of Iranian industries classification including mining and construction aggregates, food materials, beverage and tobacco industries, textile, apparel and leather industries, cellulose materials industries, chemical industries and miscellaneous materials, appliance, and basic metal industries, machinery, tools, equipment, and metal products industries and recycling industries (Ghanatabadi 2005; Hassanpour 2020). By current review Iranian industries listed into nine main groups such as Iranian Food Manufacturing and Processing Industries (IFMPI), Iranian Wood and Cellulose Industries (IWCI), Iranian Plastic Industries (IPI), Iranian Mining and Aggregate Industries (IMAI), Iranian Chemical Industries (ICI), Iranian Automotive Industries (IAI), Iranian Textile and Leather Industries (ITLI), Iranian Household Appliance Industries (IHAI), and Iranian Electronic Products Manufacturing Industries (IEPMI). Ghanatabadi (2005) classified Iranian industries as electrical and electronics, information technology, petroleum, oil and gas, construction, communication, healthcare, automobile and aerospace, agriculture, food, and retail, distribution, and warehousing, banking, and manufacturing partitions. Global classification of industries has been done as Cluster 1: food products – crop using, Cluster 2: food products – processed food using, Cluster 3: food products – livestock using, Cluster 4: tobacco products, Cluster 5: apparel and textile products, Cluster 6: miscellaneous textile products, Cluster 7: leather apparel, Cluster 8: manmade fibers and chemical products, Cluster 9: chemical products, Cluster 9.1: chemical products (cont’d.), Cluster 10: paper products, Cluster 11: printing services, Cluster 12: petroleum products, Cluster 13: lumber and wood products, Cluster 14: stone and pottery products, Cluster 15: concrete and glass products, Cluster 15.1: concrete and glass products (cont’d.), Cluster 16: metals and fabricated metal products, Cluster 16.1: metals and fabricated metal products (cont’d.), Cluster 17: machinery, Cluster 17.1: machinery (cont’d.), Cluster 18: nonelectrical machinery, Cluster

70 Iranian Small and Medium-Sized Industries

1779

19: telecommunications equipment, Cluster 20: miscellaneous manufacturing, and Cluster 20.1: miscellaneous manufacturing (cont’d.) and those industries without detailed materials records. To understand the implication of the small and medium-sized industries in Iran, it needs to explain that there is no certain and approved definition for the medium-sized industries according to existing laws and documents. The small industries included up to 50 staff and vice versa. The first proclamation about the classification of Iranian industries was posed in 1955 called the program for Economic Development Plan (EDP). By the way, the industries classified into two groups the artisan firms and the large industries. EDP proclamation devoted up to 500 staff for large group industries. According to the Fourth EDP rules (1968– 1972), artisan industries were defined based on two criteria: the number of staff (less than 10) and invested budget (less than USD 70,000.00). After the revolution of 1979, the Central Bank of Iran changed the criteria, limiting it to the number of employees and adjusting the numbers to less than 50 for small industries and more than 500 for large firms. In 1991, in a new endeavor, the Central Bank of Iran changed the criteria to 10 employees, at the same time. The Ministry in charge of small industries used 50 employees as the criteria for defining the small firms. At present, according to the law, artisan firms with less than 10 employees are defined as small businesses. With regard to the bereavement of existing information about the right classification of small and medium-sized industries and the difficulties of accessibility to the precise references and sources, we need to notice that due to published data by European Union, up to 250 staff defined as the upper limit for medium-sized firms (Ghanatabadi 2005). Iranian industrial sector sought to achieve methods and practices to reduce manufacturing expenses, impede soaring and floating waste materials flow, escalate commodities quality, elevate productivity, and increase customer satisfaction along with business excellence run-up. Universally, industries are prominent resources of production with heavy demand in energy supply aims particularly for water, electricity, and fossil fuel consumption, etc. Iranian government pays a huge budget to compensate for energy consumed in the industrial sector. So, according to the aforementioned demands and plan of the environment protection agency, all industrial projects need to go through the Environmental Impact Assessment (EIA) program once start confirming an industry by in-charge organizations. By the current review, we tried to present the energy demand evaluation of industries including the number of employees, water, fuel, and power consumed in industries individually along with areas devoted to building and landscaping activities. Iranian evaluator team gathered the mentioned information to allow the stakeholders and industries managers to receive the required permissions to dig the well, install power facilities (according to their demands), the stake for fuel subsidiary, and recruit staff to run the plants. Some of the important aspects followed by the current review can be listed as (1) benchmarking the spectrum for future growth rates, (2) comparison of Iranian industries with industries of other nations, and (3) comparison of the levels of energy resources consumption including water, electricity, and fuel and employee recruited, etc.

1780

S. Jithender Kumar Naik et al.

One prominent challenge that came into view in the expansion of industrial ecology encompassed energy stream networks and lack of access to a highly granted database to design and execute energy channels towards industries loops. There is no database about the preliminary screening of Iranian industries prior to constructing industries, ranking, and weighing the industries and their criteria. The current data encompassed the most authoritative database for Iranian industries, which is the first report in English in this regard. This review noticed quantities of energy consumed for industries, a weighted average of factors among whole industries, a new type of classification for industries, significant differences and correlation among industries, and their factors in a certain framework. This review encompassed the secondary data collected from the initial database of both Iranian protection agency and Iranian industries organization in Environmental Impact Assessment (EIA) plan. In a clustering method, around 405 types of Iranian small and mediumsized Industries studied individually before (as case studies). Obtained data came through of decision-making systems supported by SPSS and EXCEL software in different studies and then the results collected in the current review. Recently lots of multicriteria decision-making (MCDM) models have been emerged and introduced to classify, prioritize, and sort out both the uncertain and big data. As mentioned above, we collected full details of references to cover all concepts about MCDM models applied to offer a realistic decision by a list of literature reviews in this regard (Hassanpour 2020). The MCDM models of Fuzzy set theory, Fuzzy Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), TOPSIS, Weighted Aggregated Sum-Product Assessment (WASPAS), Simple Additive Weighting (SAW), COmplex PRoportional ASsessment (COPRAS), Vlse Kriterijumska Optimizacija I Kompromisno Resenje (VIKOR), Additive Ratio ASsessment (ARAS), and COmbinative Distance-based ASsessment (CODAS) were requested to assess Iranian industries (Shirazia et al. 2017; Zagorskas et al. 2014; Kaklauskas et al. 2006). By Table 1, we offered a strong literature review of recent studies inside Iran and all over the world associated with the subject discussed. Kavousi and Salamzadeh (2016), Zagorskas et al. (2014), Hosseini et al. (2013), Onat et al. (2016), Tobiszewski et al. (2017), Indahingwati et al. (2018), Georgiadis et al. (2013), Farzami and Vafaei (2013), Dace et al. (2014), Rostami et al. (2017), Askarifar et al. (2018), Dinmohammadi and Shafiee (2017), and Forghani et al. (2018) used TOPSIS model for Indicators influencing the success of a strategic planning process, historical buildings preservation (five new insulation materials chosen to reclaim), weighing and ranking 50 large industries on Tehran Stock Exchange since 2009–2011, sustainability efficiency of alternative vehicles (both hybrid and plug-in hybrid electric vehicles were the excellent options to supersede), evaluating the environmental distribution of solvents in dyeing industry (both alcohols and esters were much harmless than aromatic hydrocarbons including 1 to 78 chemicals), four kinds of fertilizer selection, finding the overwhelming technique of weapon systems, the best contractor selection, choosing a relevant catalyst for CO2 conversion and CH4 selectivity, financial performance of chemical companies in Tehran stock exchange (Ahvas Petrochemical Company, Persian

70 Iranian Small and Medium-Sized Industries

1781

Gulf, and Iran chemical industries enlaced the highest efficiency), investment opportunities for public demands, wind turbine system selection, and suppliers selection of four pharmaceuticals, respectively. The COPRAS model exploited by Zolfani and Zavadskas (2013), Popovic et al. (2012), Medineckiene et al. (2011), Vujicic et al. (2016), Jamali et al. (2015), and Zarbakhshnia et al. (2018) for evaluating the sustainability of five types of buildings with 13 criteria and 5 options (Iranian traditional systems used in deserts > light gauge steel frames > insulating concrete frameworks > Tronco systems > 3D Sandwich panels), selecting the best investment projects among four cases, evaluating building life-cycle assessment in sustainability, the best option selection in purchase a fluorescent lamp, ranking five Iranian automotive manufacturers containing seven criteria and prioritizing thirdparty reverse logistics creators in the presence of risk factors, respectively. The WASPAS model assigned for selection of the best green suppliers, removing the difficulties of eight industries, selection of the position of a garage considering four alternatives, and six criteria and the best supplier selection including 10 alternatives along with five criteria respectively (Yazdani et al. 2016; Chakraborty and Zavadskas 2014; Bausys and Juodagalyiene 2017; Azadfallah 2016). Hereby, to complete the review, we have done a depth literature review, then used the weighting and ranking systems, statistical analysis results, and managerial application of study to declare the first report of evaluating Iranian industries based on five main criteria in the EIA plan. To the best of our knowledge, this is the first report and review which encompassed whole Iranian industries to pass through the EIA program conducted by the in-charge organizations. There is no report to encompass Iranian small and medium-sized industries for the five main criteria. In Table 1, we tried to point out the recent studies completed using MCDM models and a brief description of their methodology, type of industry, the covering content, and main achievement of the study.

The Weighing and Ranking Models Applied in Classification of Iranian Industries According to Fig. 1, the steps followed in the current review complied sections of project identification and initial screening of the Iranian evaluator team, introducing decision-making systems used, introducing weighing and ranking models applied, introducing obtained ranking, and literature review collection. It needs to explain that the raw data of initial screening have been passed through the decision-making systems, weighing, and ranking stages to approve the projects in previous studies of the authors. The public involvements in projects demand some pollutants discovery practices that are not a big deal in EIA for industrial projects. This review collected the obtained results from previous studies of the authors, which are shown in figures and tables. In Iran, due to the lack of laws, in some cases, no evaluation report was prepared before. But recently under environmental laws, the government has given the authority to require environmental inspections for some projects. These laws

1782

S. Jithender Kumar Naik et al.

Table 1 Literature review of some recent studies

Methodology Fuzzy Shannon entropy and fuzzy TOPSIS

Type of industries or location Heat and power systems

Fuzzy AHP

10 biggest Iranian hotels

Fuzzy TOPSIS

–

Delphi and TOPSIS Fuzzy system

–

Fuzzy TOPSIS Fuzzy network model Fuzzy system

Key energy-saving technologies Investment strategy selection Climate policy

Iranian shipping industries

Investment practices in technology transfer

National Iranian gas Company Iranian automobile industries Iranian power plant industry

Performance analysis Stock exchange

Fuzzy system + SPSS

133 automotive industries

Fuzzy AHP

9 enterprises

Fuzzy AHP

Publishing industry 8 Iranian cement companies in the Tehran stock exchange

Fuzzy system

Subject Combined heat and power systems assessment

Risk identification and management Performance measurement

The importance of 4 criteria and 20 approaches Sustainable development Financial performance

Main achievements Gas turbine > steam turbine > fuel cell > reciprocating engine > microturbine 40 energy factors classified into five groups Prioritized the existing alternatives Weighing and ranking Joint venture and the subsidiary companies got the highest and lowest priorities Offering a score and rank system The bubble growth of stock exchange A new ranking model outlined Classifying options, weighting and ranking Ranking criteria

Ranking factors Sabhan, Sarab, Sedasht, Safar, Sekaroun, Sakarma, Sanir, and Sahrmoz companies prioritized with scores of around 0.55, 0.51, 0.50, 0.49, 0.42, 0.37, 0.36, and 0.33, respectively

References Cavallaro et al. (2016)

Mardani et al. (2016) YazdaniChamzini et al. (2013) Nikas et al. (2018) Radfar and Ebrahimi (2012)

Parsa et al. (2016) Sorayaei et al. (2012) Ebrahimnejad et al. (2012) Behrouzi et al. (2011)

Hsu and Hu (2008) Shaverdi et al. (2013) Moghimi and Anvari (2014)

(continued)

70 Iranian Small and Medium-Sized Industries

1783

Table 1 (continued)

Methodology Fuzzy cognitive map

Type of industries or location Iranian manufacturing enterprises

Entropy Shannon

Urban sprawl

Entropy Shannon

Kish airport

Entropy Shannon Fuzzy + entropy Shannon Fuzzy TOPSIS

Supplier

AHP-TOPSIS

Fuzzy TOPSIS

Basin Detergent powder up to 25 criteria in an Iranian industry Iranian petrochemical industry Offshore boat engine

Subject Agile new product process and their connections

Finding the urban sprawl based on four criteria Classification of the 13 projects containing 7 criteria Supplier selection Eutrophication index Four suppliers assessment of tripolyphosphate Connections among three major criteria Demystifying the main causes of defeat To select a temporary military road Development process

DEMATEL and COPRAS

Serbia

VIKOR + entropy Shannon

Sanitation in Ardabil province

VIKOR

Research projects in Thailand

Offering a strategy to impede expanding pollution and hazards

Fuzzy + TOPSIS + VIKOR

Turkish manufacturing industry

A performance analysis

Main achievements Items were classified and sorted into six main groups and the connections tabulated Weighting

References Fekri et al. (2009)

Danaei (2017)

Weighting

Mkhalet et al. (2018)

Weighting

Zhao et al. (2017) Taheriyoun et al. (2010) Roshandel et al. (2013)

Weighting Weighing and ranking

Weighing and ranking

Rahdari (2016)

Classification of overall scores

OsezuaAikhuele et al. (2017) Pamucar et al. (2018)

Weighting and ranking Ardabil County placed the first rank and Sarein County the last rank The sustainability index calculated along with ranking and weighting values. Obtained results revealed the same classification style

Yazdani et al. (2017)

Thipparat and Thaseepetch (2013)

Yalcin et al. (2012)

(continued)

1784

S. Jithender Kumar Naik et al.

Table 1 (continued)

Methodology VIKOR + AHP VIKOR + fuzzy analytical hierarchy

Type of industries or location Iranian state universities Basic metals generating industries

Subject 22 Iranian state universities Assessing the basic metals generating industries Parameters selection difficulties

VIKOR

A texture industry

VIKOR

Suppliers

Supplier selection with four criteria

Fuzzy VIKOR + AHP

–

VIKOR and TOPSIS techniques

Iranian 24 supplier companies

Difficulties of risk assessment processes Supplier selection including six criteria

Shannon entropy + VIKOR

–

AHP and VIKOR

Coastal zones in Khuzestan, Iran

VIKOR + fuzzy sets theory

Mehrcam pars company of the automobile, Iran

Classifying 25 types of equipment failures with four alternatives Assess conservation priority based on 17 factors and 6 alternatives Difficulties raised from supplier priority in materials chain supply networks

Main achievements Weighting and ranking Weighting and ranking

References Mazdeh et al. (2013) Farrokh et al. (2016)

Weighting and ranking system between 0–0.4 and 0–1, respectively Weighting around 0.713, 0.472, 0, and 1 for four suppliers Ranking system valued from 0 to 1 The Saze Pouyesh company got the highest rank in the TOPSIS method while Kosar Sanat Abzar company reached the highest rank in VIKOR technique. Weighting and ranking

Fallahpour and Moghassem (2012)

Developing a ranking system in the range of 0–1

Pourebrahim et al. (2014)

Weighing and ranking 15 criteria for five supplier companies

Amiri et al. (2011)

Alimardania et al. (2013)

Liu et al. (2015) Azar et al. (2011)

Omidvar and Niromand (2017)

(continued)

70 Iranian Small and Medium-Sized Industries

1785

Table 1 (continued)

Methodology Grey relational analysis

Type of industries or location Iranian dates industries

Delphi-AHP and fuzzy-GIS approaches

Large extractive industrial units (100,000 t/year)

AHP and Expert Choice 2000

Iranian particle board industries

Major criteria intensities

Cochran test

Iranian mega motor company

AHP method and expert choice software

Iranian facial tissue industries

Assessment of technological capability Weighing factors

Markov chain Grey model

Iran’s industrial sector

Energy demand forecast

Integrated VIKOR technique

Iranian mining sector

Evaluating the strategies

Subject Analyze the interaction among the major barriers Site selection of large extractive industrial units in Iran

Main achievements Figure out the main barriers and ranking them Prioritizing and ranking the criteria and identifying the suitable sites in such applications Density of the products and its high intensity got the highest priority Ranking based on sub and main dimensions Softness, time of absorption, appearance quality, basis weight, and price criteria have high priority respectively Scientific basis for the planned development of the energy supply of industrial sector in Iran Improving exploitation ability and production outperforms; other strategies

References Ghane (2014)

Kamali et al. (2015)

Azizi et al. (2009)

Khamseh and Mohagheghi (2014) Azizi (2007)

Kazemi et al. (2013)

Azimi et al. (2011)

1786

S. Jithender Kumar Naik et al.

Fig. 1 (This study)

include items 7 and 6 on Environment Protection and improvement and item 50 of the Constitution. The first EIA resolution was adopted in 1997 by the Supreme Council for Environmental Protection, which was amended in 1997. With this decree, it became clear that some projects are required to submit an EIA report. The Supreme Council for Environmental Protection is composed of 12 members, including the Ministers of Health, Treatment and Medical Education; agrarian; industry; country; housing; urban planning and Jihad Sazandegi Organization; Head of Program and Budget Organization; the head of the Environment Protection Organization; and four qualified persons and officials on the proposal of the head of the Environment Protection Organization and the approval of the President for 3 years. The High Chairman of the Council is the President. The Scientific Committee for Environmental Assessment consists of the head of the Environment Protection Organization and five experts selected by the head of the organization; agents of the Program and Budget Organization; forests and pastures; Institute of Standards and Industrial Research; the ministry or organization is involved in the evaluation plan. EIA first began in the United States in 1970, Germany in 1971, Sweden in 1972, the United Kingdom in 1973, Canada, Australia, and Denmark in 1974, and France

70 Iranian Small and Medium-Sized Industries

1787

in 1976. Environmental concerns as a result of economic growth in the 1950s and 1960s, resource degradation and economic scarcity in the 1970s, the discussion of sustainable development in the 1980s, the development of new perspectives on economic progress with an emphasis on natural resources management, and the ideas presented at the Rio Conference in 1992 contributed to the EIA plan. Rule 17 of the Rio Declaration emphasizes that EIA should, as a national instrument, be required for proposed activities that have significant, potential adverse effects on the environment and require national decision-making. Project management operates in a broader environment in comparison with the project environment itself. The project management team needs to understand this broader concept. Managing the day-to-day activities of the project is necessary for success, but it is not enough. Key aspects of project management include several dimensions such as project stages and project life cycle, project stakeholders, organizational impact, key public management skills, social, economic, and environmental impacts. The sum of the project stages is called the project life cycle. The project life cycle is used to determine the start and end of a project. The project life cycle will determine: Whether the feasibility study is the first stage of the project or is considered as a separate independent project. What transition measures are planned at the beginning and end of the project and what measures are not planned? The project life cycle definition can be used to relate the project to the current operations of the executive organization. Due to the inseparability of the two issues of environment and development, the need for environmental management instruments in development programs to minimize the damage to the environment is identified. The environmental management system has various features, including complementing and continuing sustainable development, combining different development perspectives, etc. It integrates social sciences, policy-making, and planning, making it inevitable to pay attention to the needs of human societies and to expand them globally and regionally. The objectives of the evaluation include the above (1) erasing and repairing damage to the environment, (2) increasing public awareness, (3) using the views and opinions of the public in the decision-making process, (4) recognizing issues and problems of environmental damage that they are likely to occur, (5) predicting the occurrence of significant and sustainable environmental impacts, (6) creating a balance between the long-term development goals and the need for the majority of people to have development resources to protect the environment, (7) increasing the level of cooperation and coordination between public and private organizations, (8) applying and integrating environmental criteria in development planning, (9) defining the duties of each government agency to protect the environment, (10) establishing a balance between population and environmental resources, (11) maintaining the quality of renewable resources to maximize efficiency with proper maintenance of life cycles, (12) providing a healthy and active life for the community, (13) identifying the correct methods of using the environment, and (14) recognizing critical environmental issues and problems that need to be studied, controlled, and cared for. The reasons for the need to evaluation can be mentioned in the following cases. Community professionals and their representatives want to identify the various activities of a proposed project and seek to recognize its effects. This desire is formed based on the following needs:

1788

S. Jithender Kumar Naik et al.

(1) awareness of how to locate and implement the project and reduce the effects for project managers, (2) damages and losses to the project for project investors, (3) the need to know the results of the evaluation to provide a project permit for environmental officials, (4) the impact of the project on other projects for other government officials, and (5) the impact of the project on the quality of life of the local community or their representatives. The benefits of an assessment include (1) resolving public dissatisfaction, (2) eliminating incompatibility between individuals in society and government agencies, (3) increasing the quality of the environment, (4) crediting the government at the international level, (5) creating appropriate planning, (6) increased environmental awareness and knowledge at various levels of society, officials, and decision-makers, (7) and increased executive power referred to laws and regulations. Evaluation by the project employer (public or private sectors), responsible departments (usually government). Planner (usually government) or a combination of the two or more above. Factors involved in an assessment are the sensitivity of the local environment, the degree, and dimensions of development and the proposed project and its potential effects, the social value of the project at national and local levels in terms of improving environmental quality, resources, and scientific expertise of the country and assessment time (Qavamabadi 2013). Weighing and ranking systems: The source for the data gets back to initial screening of engineering and industrial projects in the preliminary steps of study and industries confirmation to acquire the required licenses which have been accomplished by both Iranian industries organization and Iranian protection agency once in the beginning. Both weighing systems of Friedman test and Entropy Shannon have been chosen to conduct the classification of industries based on the five main criteria, that is, the number of employees, water, fuel, energy consumed along with land area occupied by each industry in previous studies of authors. The reason for assigning these weighing systems refers to the fact that the Friedman test is a valuable test to present low, high, and medium weights for the low, high, and medium values. This prominent point was acquired by author research to conduct the thesis in this regard. The Entropy Shannon weighting system provides an excellent platform to judge about industries situations containing positive and negative criteria to sort out. Actually, the Friedman test sets up a matrix of data in the framework of SPSS software to offer the values of weights for configured data in the software (Eisinga et al. 2017; Hassanpour and Pamucar 2019). The ranking systems used were Fuzzy set theory, Fuzzy-TOPSIS, TOPSIS, WASPAS, SAW, COPRAS, VIKOR, ARAS, and CODAS.

Iranian Industries By the current review, nine groups (including 405 industries) of Iranian industries have come through the MCDM models to classify the industries pertain on five main criteria such as IWCI (16 various kinds), ITLI (38 various kinds), IMAI (26 various kinds), IFMPI (57 various kinds), IAI (71 various kinds), IPI (21 various kinds), IEPMI (33 various kinds), ICI (118 various kinds), and IHAI (25 various kinds).

70 Iranian Small and Medium-Sized Industries

1789

The existing information indicates the five main industry-specific factors that are calculated by the team of evaluators of Iranian organizations for industries.

Reviewing Statistical Analysis Results In this step, the statistical outputs of industrial groups were collected based on five main criteria to further assess. There was no significant difference among the criteria of IFMPI via t-test analysis. Also, Friedman’s two-way analysis of variance by ranks emerged to zero significance in this regard. In IWCI were found significant differences around (p-value ≤0.001, 0.002) among the criteria via t-test analysis. An equal probability manifested about 0.982, 0.437 (via one-sample Chi-Square test), 0.299 (via one-sample Kolmogorov Smirnov test), and 0.309 and 0.185 (via onesample Kolmogorov Smirnov test) among criteria of IWCI. The findings proved a normal distribution among the criteria of IWCI. The t-test analysis had shown a significant difference among parameters (p-value ≤. 0.009) along with a high value of Pearson correlation for both factors of land and water approximately 0.835 (in IPI). The related samples Friedman’s two-way analysis of variance by ranks appeared with a significant value of about 0.007 among the values of IPI. The normal distribution trend was proved by one-Sample Kolmogorov-Smirnov test among criteria with a significant difference (p-value ≤0.015) among the criteria using the t-test (in IMAI). The highest correlation distinguished by Pearson correlation sig. (2-tailed) about 0.797 between the values of land and employees (The values of IMAI). According to the t-test, there was no significant difference among the criteria of ICI. The Chi-square test obtained around 388.645 (N = 118). The Null hypothesis completely rejected. No significant difference was acquired using both t-test and related samples Friedman’s two-way analysis of variance by ranks for the IAI. There was a normal distribution among the criteria. The Null hypothesis completely rejected. Pearson correlation sig. (2-tailed) came into view with the maximum correlation between land and water amounts approximately 0.82. The distribution of criteria obtained normally and Pearson correlation analysis had shown the top correlation between land and power values ranged about 0.818 with zero significant differences among criteria of ITLI. Using the Chi-square test provided a value of around 90.321 via Friedman test along with 0.879 for the highest correlation between two criteria of land area used and employee’s number (in IHAI). Also, it has acquired a significant difference (p-value ≤0.025) via t-test analysis among the five main criteria of IHAI. The Pearson correlation sig. (2 tails) appeared with a value of about 0.837 between both criteria of water and power consumed (In IEPMI). It has completed the statistical analysis with emerging zero significant differences among the main criteria of IEPMI (Hassanpour 2018a, b, c; Hassanpour 2020; Hassanpour 2019a, b, c, d). Similar studies covering the same concept are listed as follows. Hosseininia and Ramezani (2016) used SPSS analysis to classify and prioritize with the call out of 130 participants and 12 owner-managers in the food industry considering the fact about the sustainability in small and medium-sized industries. The study of

1790

S. Jithender Kumar Naik et al.

Hourali et al. (2008) weighted indicators for around 45 Iranian small and medium enterprises (SME) using Minitab and SPSS software. Dadashpoor and Allan (2010) used SPSS analysis to find intra-metropolitan supply linkages, industrial clustering, and driving forces in the Iranian automotive sector. By the way, it was achieved with insufficient evidence of industrial clustering. Significant relationships between accounting variables were observed in the study by Salehi and Hematfar (2012) in 10 models SPSS analysis on the data of chemical companies listed on the Tehran Stock Exchange from 2005 to 2010. The relevant quality improvement was reported by Yunus et al. (2016) in a process efficiency assay of Iranian food industries via SPSS analysis. SPSS analysis of Kaab Dogloss carried out on Iranian smallscale industries resulted in development of two mathematical production functions (Shahroudi 2011). The promotion of knowledge management practices has been taken into attention by Valmohammadi (2010) in a study on the Critical Success Factors (CSFs) of Iranian SMEs via SPSS analysis. A significant relationship has been obtained in an investigation based on the connection between new and old industries to survive using SPSS analysis in new firms established in Mazandaran province (Madhoushi and Nasiri 2011). The Esfahan Steel Company got a survey to identify CSFs via SPSS analysis (Moohebat et al. 2011). A model for E-Readiness has been assessed concerning weight the indicators in 45 Iranian SMEs using Minitab and SPSS analysis (Hourali et al. 2008).

Classified Iranian Industries In the following steps, the values of weights (VW) for nine groups of industries are listed in Figs. 2, 3, 4, 5, 6, 7, 8, and 9, and the classification of industrial groups is summarized in Table 2 by the number of industries in each cluster, weight, and rank released for them. To rank, the IHAI, ITLI, IAI, and IMAI have used both the weighing systems of Friedman test and Entropy Shannon regarding the future expansion in industries

Fig. 2 VW developed in the ranking system for IMAI (Hassanpour 2019a)

70 Iranian Small and Medium-Sized Industries

1791

Fig. 3 VW developed in the ranking system for IHAI (Hassanpour and Pamucar 2019)

Fig. 4 VW developed in the ranking systems for IHAI and ITLI (Hassanpour and Pamucar 2019; Hassanpour 2019b)

1792

S. Jithender Kumar Naik et al.

Fig. 5 VW developed in the ranking systems for IAI (Hassanpour 2019c)

Fig. 6 VW developed in the ranking systems for IFMPI and IPI (Hassanpour 2018a, b)

70 Iranian Small and Medium-Sized Industries

1793

Fig. 7 VW developed in the ranking systems for IWCI (Hassanpour 2018c)

Fig. 8 VW developed in the ranking system for ICI (Hassanpour 2020)

including negative and positive criteria among the five main criteria mentioned. The WASPAS model assigned to rank the IMAI in both systems of weighting and revealed no significant difference between them. Also, the IAI have been classified using SAW and COPRAS ranking models in the same weighting systems (p-value ≤0.00). IHAI have been classified based on both systems of CODAS and ARAS. According to the findings, the ARAS and CODAS created very close results in a ranking system with a significant difference around (p-value ≤0.001) for the ranking system of ARAS with the weighting system of Entropy Shannon. Then, ITLI came through the ranking systems of SAW and VIKOR to prioritize the industries with both weighing systems of Friedman test and Entropy Shannon (p-value ≤0.00). According to Fig. 6, IPI and IFMPI were ranked in Delphi fuzzy set (integrated with the SAW model) via normalization, de-fuzzification, and criteria/symbols versus factors based on the Likert Scale. The Friedman test was employed to estimate the VW for five main criteria. Then SAW model integrated with Delphi

1794

S. Jithender Kumar Naik et al.

Fig. 9 VW developed in the ranking systems for IEPMI (Hassanpour 2019d)

fuzzy set requested to release the weights for each industry and finalized the ranking system. The same procedure was applied for the IWCI (Delphi fuzzy set integrated with the SAW model) with IPI and IFMPI along with two other methods of Fuzzy TOPSIS and TOPSIS supported with the weighting system of Friedman test for the criteria. The normalizing and weighing for alternatives (industries) followed the TOPSIS method and finalized with the calculation of Euclidean distances and division of them. The Fuzzy TOPSIS also underwent the same procedure except using fuzzy numbers instead of real numbers in order to normalize the alternatives values. By the way, it observed no significant difference between the three methods of ranking systems. Also, the TOPSIS method was assigned to classify ICI, so the procedure was united with the weighting system of the Friedman test for the criteria. Therefore, 118 various types of ICI were classified in a certain cluster. The ranking systems of ARAS and SAW classified the IEPMI using the weighting system of the Friedman test for the criteria that resulted in the same ranks with different VW (p-value ≤0.00). The sequence number of the VW in various ranking models is drawn in Fig. 10. The sequence diagram usually shows interactions between different classes of models discussed to prove the functionality of the explained models where they graphically undergo development steps. The distribution of the values of the weights

70 Iranian Small and Medium-Sized Industries

1795

Table 2 Classified Iranian industries (Hassanpour 2018a, b, c, 2019a, b, c, d, 2020) Industries/ranking system IFMPI/fuzzy set theory

IWCI/fuzzy set theory

IPI/fuzzy set theory

IMAI/WASPAS

ICI/TOPSS

Classification (42) > (43) > (51) > (14) > (31) > (1) > (41) > (7) > (4) > (52) > (23) = (25) > (15) > (5) > (17) > (19) > (3) > (11) > (35) > (34) > (47) > (48) > (45) > (33) > (54) > (16) > (53) > (20) > (49) > (9) > (29) > (39) > (6) > (12) = (50) > (18) > (57) > (56) > (21) > (38) > (24) > (2) > (36) > (46) > (37) > (8) = (26) > (30) > (32) > (27) > (40) > (55) > (22) > (44) > (10) = (13) = (28) Weighing based on Friedman tests Land (4.98)> power (3.95) > fuel (2.26) > employees (2.17) > water (1.64) Fuzzy set logic 15 > 11 > 16 > 5 > 1 > 3 > 10 > 4 > 8 > 7 > 9 = 12 > 2 > 6 > 13 = 14 Fuzzy TOPSIS 15 > 7 > 11 > 1 > 5 > 16 > 3 > 10>8 > 4 > 12 > 9 > 2 > −; (6 = 13 = 14) TOPSIS 15 > 11 > 1 > 16 > 5 > 3 > 4 > 8 > 10 > 7 > 12>9 > 14 > 2 > 13 > 6 Weighing based on the Friedman test Employees (2.59), power (4), water (1.53), fuel (1.88), and land (5). 1> 2> 7> 8> 13> 10> 3> 20> 18> 9 > 6> 14> 19 > 11 > 16 = 4 > 15> 5> 12 > 17 > 21 Weighing based on the Friedman test Employees (2.88), power (3.86), water (1.74), fuel (1.52), and land (5). 26 > 17 > 9 > 4 > 21 > 10 > 19 > 11 = 12 > 20 > 2 > 5 > 16 > 24 > 3 > 8 > 25 > 22 > 6 > 14 > 1 > 18 > 13 > 15 > 23 > 7. Weighing based on the Friedman test Employees (2.31), power (3.96), water (1.79), fuel (1.94), and land (5). 17 > 26 > 2 > 9 > 21 > 20 > 10 > 19 > 4 > 16 >11 = 12 > 6 > 5 > 24 > 3 > 13 > 8 > 22 > 25 > 1 > 14 > 7 > 23 > 18 > 15. Weighing based on the entropy Shannon Employees (0.07547), power (0.1074), water (0.1767), fuel (0.4554), and land (0.1849) 85 > 10 > 68 > 54 > 108 > 11 > 70 > 110 > 31 > 95 > 1 > 118> 48> 115>116> 33> 55> 96> 72>113 > 24> 60>97 > 103>27> 21>107 > 14 > 50> 32> 53 > 30> 114> 8> 81> 109 > 3> 62> 94> 65> 41> 34> 19> 56 > 76> 84> 100> 18> 5> 73> 63> 80> 47> 42> 101> 43> 83> 16> 87> 38> 15> 99 > 86> 40> 88> 37 > 105> 98> 51> 44> 46 > 17> 29> 58> 34> 112> 23 > 79> 36> 26> 82> 117> 90> 67> 89> 92> 69> 13> 77> 6> 2> 20> 66> 93> 35> 104> 28> 7> 78 > 106> 9 > 52> 75> 111> 102> 64> 59> 91> 93> 35> 12> 4 > 57> 71> 74> 45> 22> 61. Weighing based on the Friedman test Employees (2.52), power (3.94), water (1.6), fuel (1.94), and land (5) (continued)

1796

S. Jithender Kumar Naik et al.

Table 2 (continued) Industries/ranking system IAI/SAW and COPRAS

Classification Saw;52>2>48>13>5>7>70>67>17>56>24>27>65>66>53>12>38> 3>68>30>16>36>19>34>21>4>37>50>41>29>39>43>15>61>35> 47>63>26>10>62>54>42>64>57>23>8>31>51>46>33>25>58>49> 40>11>1>69>6>60>59>45>28>32>14>55>44>22>71>9>18>20 Weighing based on the Friedman test Employees (2.94), power (4), water (1.78), fuel (1.27), and land (5). COPRAS; 70 > 52 > 50 > 2 > 3 > 53 > 48 > 25 > 19 > 38 > 13 > 66 > 17 > 56 > 62 > 27 > 36 > 24 > 5 > 57 > 59 > 15 > 65 > 39 > 10 > 63 > 29 > 47 > 33 > 42 > 44 > 35 > 21 > 55 > 58 > 9 > 64 > 46> 14 > 34 > 40 > 51 > 8 > 43 > 26 > 23 > 68 > 31 > 11 > 49 > 12 > 18 > 71 > 1 > 69 > 20> 54 > 45 > 67 > 6 > 61 > 28 > 4 >60 > 22 > 41 > 37 > 16 > 7 > 30 > 32. Weighing based on the entropy Shannon Employee (0.0801), power (0.1592), water (0.1039), fuel (0.5589), and land (0.0976). VIKOR; 10 > 34 > 3 > 21> 17 > 19 > 15 > 6 > 11 > 18 > 9 > 14 > ITLI/VIKOR 36 > 4 > 31 > 37 > 7 > 2 > 28 > 8 > 27 > 1 > 38 > 35 > 26 > 23 > 24 > 29 > 32 > 13 > 25 > 16 > 33 > 5 > 30 > 20 > 22 > 12. Weighing based on the Friedman test Employees (2.8), power (3.88), water (1.7), fuel (1.64), and land (4.97) SAW; 36 > 10 > 34 > 3> 21 > 4 > 17 > 11 > 15 > 19 > 6 > 18 > 9 > 14 > 31 > 8 > 28 > 7 > 35 > 27 > 38 > 1 > 2 > 37 > 24 > 23 > 26 > 16 > 25 > 32 > 13 > 5 > 29 > 33 > 20 > 30 > 22 > 12. IHAI/ARAS and CODAS ARAS; 5 > 22 > 23 > 19> 25 > 9 > 24 > 17 > 8 > 10 > 16 > 3 > 15 > 4 > 14 > 11 > 20 > 7 > 18 > 21 > 2 >12 > 13 > 6 > 1. Weighing based on the Friedman test Employees (2.84), power (3.98), water (1.72), fuel (1.46), and land (5) ARAS; 22> 23 > 5 > 19> 9 > 24 > 17 > 25 > 10 > 8 > 16 > 15 > 3 > 4 > 11 > 7 > 20 > 18 > 14 > 21 > 2 > 12 > 13 > 1 > 6. Weighing based on the entropy Shannon Employees (0.0945), power (0.1946), water (0.1368), fuel (0.4763), and land (0.0975) CODAS rank; 5 > 22 > 23 > 19> 9 > 25 > 17 > 28 > 8 > 10 > 16 > 3 > 14 > 15 > 4 > 20 > 11 > 7 > 18 > 21 > 2 > 6 > 12 > 13 > 1. Weights based on the Friedman test; Employees (2.84), power (3.98), water (1.72), fuel (1.46), and land (5). CODAS rank; 22 > 23 > 5 > 24> 9 > 17 > 19 > 25 > 16 > 10 > 8 > 15 > 3 > 4 > 11 > 7 > 18 > 20 > 14 > 21 > 2 > 1 > 12 > 13 > 6. Weights based on the entropy ShannonEmployees (0.09454463), power (0.19462021), water (0.136873268), fuel (0.476367041), and land (0.097594852) (continued)

70 Iranian Small and Medium-Sized Industries

1797

Table 2 (continued) Industries/ranking system IEPMI/SAW; ARAS

Classification SAW; 32 > 1 > 25 > 15> 28 > 17 > 10 > 29 > 33 > 30 > 31 > 23 >26 > 16 > 3 > 24 > 4 > 5 > 27 > 14 > 20 > 7 > 19 > 13 > 12 > 6 > 9 > 21 > 11 > 18 > 2 > 22. ARAS; 32 > 1 > 25 > 15> 28 > 17 > 10 > 29 > 33 > 30 > 31 > 23 > 26 > 16 > 3 > 24 > 4 > 5 > 27 > 14 > 20 > 7 > 19 > 13 > 12 > 6 > 9 > 21 > 11 > 18 > 2 > 22. Weighing based on the Friedman test Land (5.00) > power (3.91) > employee (3.08) > water (1.89) > fuel (1.12)

Fig. 10 (Supplementary analysis)

in MCDM models has not appeared in linear consequences approaches. On the other hand, there is no high overlapping for the functions deployed. The VW for IHAI (Entropy Shannon, ARAS) had shown a significant difference (p-value ≤0.001) among 18 groups of the values of weights for nine groups of industries via t-test.

1798

S. Jithender Kumar Naik et al.

The managerial applications of the current review encompassed a valid reference for the energy demand of Iranian industries based on nominal capacity as benchmarking, comparison of industries in the national, local and global levels, set up a national database, possibility in the exploitation of available resources, future studies, and extensions towards demystifying efficient industries, offer to develop a relevant classification of industries, optimization of energy demands for the stockholder’s claims and expansion in financial surveys, etc.

Iranian Future Industries Diamond synthesis and deposition using plasma reactors with raw and waste materials stream: Wastes and lots of raw materials provided suitable feedstocks to create high value-added products for human demands. By the way, recent signs of progress and development moved towards plasma technology applications to rotate the steer for converting and depositing some new products employing Chemical Vapor Deposition (CVD) and make up the value-added commodities, valuable gasses, H2 generation, diamond, and diamond-like materials deposition using wastes as a feedstock. The methods matured for synthesis and generate diamond and analogous materials included many practices such as synthesis of high pressure and high-temperature diamond, flame diamond synthesis, bias-enhanced nucleation diamond synthesis, synthesis of detonation diamonds, ion implantation diamond synthesis, and CVD. Similarly, new MCDM methods also are developing and need to be applied for current issues in this regard (Purushothaman et al. 2011; Yoshimoto et al. 2001; Honga et al. 2002; Khachatryan et al. 2008).

Conclusion Considering the need to expand and develop a useful database to evaluate Iranian industries at the stage of obtaining the necessary permissions, this review has taken a step towards achieving this goal. The need for comprehensive information on energy consumption in industries has recommended and emphasized in many new studies, depending on the nominal capacity of the industry. Awareness of the energy used to compensate for production costs and how to use alternative energy resources are very important points. By this review, classified Iran’s small and medium-sized industries with different ranking systems were described based on the five main factors. The Iranian government has gained a golden opportunity to build new industries in line with the production of valuable materials, especially diamonds, from raw materials and waste streams as new suggestions to expand the industrial scope and area according to sustainable development aims. The growth rates and signs of progress in developing the ecological industry targeted to universal indices, so optimal exploitation can be cited in the future plans of the current review. The future developments demand allocating whole industries in certain clusters depending

70 Iranian Small and Medium-Sized Industries

1799

on nominal capacity. Data envelopment analysis will enhance materials streams integrated with the energy stream and classification of industries more specifically.

References Alimardania M, Zolfani SH, Aghdaiec MH, Tamosaitien J (2013) A novel hybrid SWARA and VIKOR methodology for supplier selection in an agile environment. Technol Econ Dev Econ 19(3):533–548. https://doi.org/10.3846/20294913.2013.814606 Amiri M, Ayazi SA, Olfat L, Moradi JS (2011) Group decision making process for supplier selection with VIKOR under fuzzy circumstance case study: an Iranian car parts supplier. Int Bull Bus Adm 10:62–75 Askarifar K, Motaffef Z, Aazaami S (2018) An investment development framework in Iran’s seashores using TOPSIS and best-worst multi-criteria decision making methods. Decis Sci Lett 7:55–64. https://doi.org/10.5267/j.dsl.2017.4.004 Azadfallah M (2016) A hybrid MCDM approach for ranking suppliers by considering ethical factors. Electron J Bus Ethics Org Stud 21(1):20–25. http://ejbo.jyu.fi Azar A, Olfat L, Khosravani F, Jalali R (2011) A BSC method for supplier selection strategy using TOPSIS and VIKOR: a case study of part maker industry. Manage Sci Lett 1:559–568. https:// doi.org/10.5267/j.msl.2011.05.005 Azimi R, Yazdani-Chamzini A, Fooladgar MM, Basiri MH (2011) Evaluating the strategies of the Iranian mining sector using an integrated model. Int J Manag Sci Eng Manag 6(6):459–466. https://doi.org/10.1080/17509653.2011.10671196 Azizi M (2007) Determination of effective criteria in Iranian facial tissue industries with respect to customer’s perspective. Proceedings of the ISAHP symposium Viña Del Mar. 1–7 Azizi M, Khakifirooz A, Moghimi F (2009) Evaluation of the major criteria intensities for Iranian particleboard products with respect to manufacturer’s aspects. ISAHP 2009 Symposium, July 29-August 1, Pittsburgh Bausys R, Juodagalvien˙e B (2017) Garage location selection for residential house by WASPAS-SVNS method. J Civ Eng Manag 23(3):421–429. https://doi.org/10.3846/ 13923730.2016.1268645 Behrouzi F, Wong KY, Behrouzi F (2011) A study on lean supply chain performance measures of SMEs in the automotive industry. IEEE Int Conf Ind Eng and Eng Manag, Singapore, 6–9 DEC. https://doi.org/10.1109/IEEM.2011.6117914 Cavallaro F, Zavadskas EK, Raslanas S (2016) Evaluation of combined heat and power (CHP) systems using fuzzy Shannon entropy and fuzzy TOPSIS. Sustainability 8(6):2211–2219. https://doi.org/10.3390/su8060556 Chakraborty S, Zavadskas EK (2014) Applications of WASPAS method in manufacturing decision making. Informatica 25(1):1–20 D¯ace E, Rusanova J, Gusca J, Blumberga D (2014) Selecting a catalyst for methanation process: technical and economic performance based TOPSIS analysis. In: Proceedings of the 27th international conference on efficiency, cost, optimization, simulation and environmental impact of Energy Systems (ECOS 2014), Finland, Turku, and 15–19 June. Åbo Akademi University, Turku, pp 1–9 Dadashpoor H, Allan A (2010) Industrial clustering, innovation and competitive advantage in the metropolitan regions: evidence from the auto-parts cluster within the Tehran metropolitan region. Intl J Humanit 17(1):19–46 Danaei J (2017) Prioritization of Kish airport projects using multi-criteria decision-making (weighting: Shannon entropy). Amazonia Investiga 6(11):122–131 Dinmohammadi A, Shafiee M (2017) Determination of the Most suitable technology transfer strategy for wind turbines using an integrated AHP-TOPSIS decision model. Energies 10:642. https://doi.org/10.3390/en10050642

1800

S. Jithender Kumar Naik et al.

Ebrahimnejad S, Mousavi SM, Tavakkoli-Moghaddam R, Heydar M (2012) Evaluating high risks in large-scale projects using an extended VIKOR method under a fuzzy environment. Int J Ind Eng Comput 3:463–476. https://doi.org/10.5267/j.ijiec.2011.12.001 Eisinga R, Heskes T, Pelzer B, Te Grotenhuis M (2017) Exact p-values for pairwise comparison of Friedman rank sums, with application to comparing classifiers. BMC Bioinf 18(68):1–18. https://doi.org/10.1186/s12859-017-1486-2 Fallahpour AR, Moghassem AR (2012) Evaluating applicability of VIKOR method of multicriteria decision making for parameters selection problem in rotor spinning. Fibers Polymers 13(6):802–808. https://doi.org/10.1007/s12221-012-0802-8 Farrokh M, Heydari H, Janani H (2016) Two comparative MCDM approaches for evaluating the financial performance of Iranian basic metals companies. Iran J Manage Stud (IJMS) 9(2): 359–382 Farzami SM, Vafaei F (2013) Evaluation and selection of optimal contractor to execute project using FTOPSIS method (Case study: Kermanshah gas company). Int Res J Appl Basic Sci 6(4):450–459 Fekri R, Aliahmadi A, Fathian M (2009) Predicting a model for agile NPD process with fuzzy cognitive map: the case of Iranian manufacturing enterprises. Int J Adv Manuf Technol 41:1240–1260. https://doi.org/10.1007/s00170-008-1565-7 Forghani A, Sadjadi SJ, Moghadam BF (2018) A supplier selection model in pharmaceutical supply chain using PCA, Z-TOPSIS and MILP: a case study. PLoS One 13(8):e0201604. https:/ /doi.org/10.1371/journal.pone.0201604 Georgiadis DR, Mazzuchi TA, Sarkani S (2013) Using multi criteria decision making in analysis of alternatives for selection of enabling technology. Syst Eng 16(3):287–303. https://doi.org/ 10.1002/sys.21233 Ghanatabadi F (2005) Internationalization of small and medium-sized enterprises in Iran. Lulea University of Technology Department of Business Administration and Social Sciences Division of Industrial Marketing and e-Commerce. PhD Thesis Ghane S (2014) The analysis of barriers to the development of industries Iranian dates. Academia Arena 6(10):8–13 Hassanpour M (2018a) Hierarchical cluster classification of Iranian food manufacturing and processing industries. J Food Safe Hyg 4(1–2):27–40 Hassanpour M (2018b) Evaluation of Iranian plastic industries. J Waste Recycling 3(2): 1–11 Hassanpour M (2018c) Evaluation of Iranian wood and cellulose. Decis Making 2(1):13–34. https:/ /doi.org/10.31181/dmame1901013h Hassanpour M (2019a) Evaluation of Iranian mining and aggregate industries. Ind Eng J 12(3): 1–15. https://doi.org/10.26488/IEJ.12.3.1179 Hassanpour M (2019b) Evaluation of Iranian textile and leather industries. J Appl Res Ind Eng 6(1):33–51. https://doi.org/10.22105/jarie.2019.167432.1072 Hassanpour M (2019c) Evaluation of Iranian automotive industries. Int J Res Ind Eng 8(2):115– 139. https://doi.org/10.22105/riej.2019.183148.1085 Hassanpour M (2019d) Evaluation of Iranian electronic products manufacturing industries using an unsupervised model, ARAS, SAW and DEA models. J Ind Eng Manage Stud 6(2):1–24. https://doi.org/10.22116/JIEMS.2019.92255 Hassanpour M (2020) Evaluation of Iranian small and medium-sized industries. A Ph. D thesis submitted to Osmania University, India Hassanpour M, Pamucar D (2019) Evaluation of Iranian household appliance industries using MCDM models. Oper Res Eng Sci 2(3):2620–1747. https://doi.org/10.31181/oresta1903001h Honga SP, Yoshikawa H, Wazumia K, Koga Y (2002) Synthesis and tribological characteristics of nanocrystalline diamond film using CH4-H2 microwave plasmas. Diam Relat Mater 11 (3–6):877–881. https://doi.org/10.1016/S0040-6090(02)00006-8 Hosseini S-H, Ezazi ME, Heshmati MR, Hosseini Moghadam SMR (2013) Top companies ranking based on financial ratio with AHP-TOPSIS combined approach and indices of Tehran stock exchange, a comparative study. Int J Econ Financ 5(3):126–133. https://doi.org/10.5539/ ijef.v5n3p126

70 Iranian Small and Medium-Sized Industries

1801

Hosseininia G, Ramezani A (2016) Factors influencing sustainable entrepreneurship in small and medium-sized Enterprises in Iran: a case study of food industry. Sustainability 8(10):1–20. https://doi.org/10.3390/su8101010 Hourali M, Fathian M, Montazeri A, Hourali A (2008) A model for E-readiness assessment of Iranian small and medium enterprises. J Fac Eng 41(7):969–985 Hsu CW, Hu AH (2008) Green supply chain management in the electronic industry. Int J Environ Sci Tech 5(2):205–216. https://doi.org/10.1007/BF03326014 Indahingwati A, Wajdi MBN, Susilo DE, Kurniasih N, Rahim R (2018) Comparison analysis of TOPSIS logic method on fertilizer selection. Int J Eng Technol 7(2–3):109–114 Jamali G, Farrokhnejad K, Mohammadi M (2015) Decision making on analyzing advanced manufacturing systems dimensions: SWARA and COPRAS_G integration (Case study: automotive industry).Teknol. Tanaman, Bil.12. Tambaban 2:266–274 Kaklauskas A, Zavadskas EK, Raslanas S, Ginevicius R, Komka A, Malinauskas P (2006) Malinauskas, P. Selection of low-e windows in retrofit of public buildings by applying multiple criteria method COPRAS: a Lithuanian case. Energ Buildings 38(5):454–462. https://doi.org/ 10.1016/j.enbuild.2005.08.005 Kamali M, Alesheikh AA, Khodaparast Z, Hosseinniakani SM, Borazjani SAA (2015) Application of Delphi-AHP and fuzzy-GIS approaches for site selection of large extractive industrial units in Iran. J Settlements Spatial Plann 6(1):9–17 Kavousi S, Salamzadeh Y (2016) Identifying and prioritizing factors influencing success of a strategic planning process: a study on National Iranian Copper Industries Company. Asian Soc Sci 12(8):230–244. https://doi.org/10.5539/ass.v12n8p230 Kazemi A, Modarres M, Mehregan M (2013) Energy demand forecast of Iran’s industrial sector using Markov chain Grey model. Intl J Humanit 20(1):1–12 Khachatryan AK, Aloyan SG, May PW, Sargsyan R, Khachatryan VA, Baghdasaryan VS (2008) Graphite-to-diamond transformation induced by ultrasound cavitation. Diam Relat Mater 17:931–936. https://doi.org/10.1016/j.diamond.2008.01.112 Khamseh A, Mohagheghi M (2014) Assessment of technological capability in Iranian automotive industries through developing the model. Eur Online J Natural Soc Sci 2(3):1730–1736 Liu HC, You XJ, You XY, Shan MM (2015) A novel approach for failure mode and effects analysis using combination weighting and fuzzy VIKOR method. Appl Soft Comput 28:579– 588. https://doi.org/10.1016/j.asoc.2014.11.036 Madhoushi M, Nasiri A (2011) The influence of industry characteristics on new Firms’ survival: Iranian study. Aust J Basic Appl Sci 5(3):653–661 Mardani A, Zavadskas EK, Streimikiene D, Josuh A, Nor MDK, Khoshnoudi M (2016) Using fuzzy multiple criteria decision making approaches for evaluating energy saving technologies and solutions in five star hotels: a new hierarchical framework. J Energy 117:131–148. https:// doi.org/10.1016/j.energy.2016.10.076 Mazdeh MM, Razavi S-M, Hesamamiri R, Zahedi MR, Elahi B (2013) An empirical investigation of entrepreneurship intensity in Iranian state universities. High Educ 65:207–226. https:// doi.org/10.1007/s10734-012-9539-y Medineckiene M, Turskis Z, Zavadskas EK (2011) Life-cycle analysis of a sustainable building, aplying multicriteria decision making method. The 8th international conference May 19–20, Vilnius Mkhalet MEL, Aziz S, Saidi R (2018) The application of entropy methods to formulate global performance for selecting the automotive suppliers in Morocco. J Theor Appl Inf Technol 96(16):5522–5536 Moghimi R, Anvari A (2014) An integrated fuzzy MCDM approach, and analysis, to the evaluation of the financial performance of Iranian cement companies. Int J Adv Manuf Technol 71 (1–4):685–698 Moohebat MR, Jazi MD, Asemi A (2011) Evaluation of the ERP implementation at Esfahan steel company based on five critical success factors: a case study. Int J Bus Manag 6(5):236–250 Nikas A, Doukas H, Opez LM (2018) A group decision making tool for assessing climate policy risks against multiple criteria. J Heliyon 4(3):1–38. https://doi.org/10.1016/ j.heliyon.2018.e00588

1802

S. Jithender Kumar Naik et al.

Omidvar M, Nirumand F (2017) An extended VIKOR method based on entropy measure for the failure modes risk assessment – A case study of the geothermal power plant (GPP). Safety Science 92:160–172. https://doi.org/10.1016/j.ssci.2016.10.006 Onat NC, Gumus S, Kucukvar M, Tatari O (2016) Application of the TOPSIS and intuitionistic fuzzy set approaches for ranking the life cycle sustainability performance of alternative vehicle technologies. Sustain Production Consumption 6:12–25. https://doi.org/10.1016/ j.spc.2015.12.003 Osezua-Aikhuele DO, Sorooshian S, Hannis-Ansah R, Mohd-Turan F (2017) Application of intuitionistic fuzzy topsis model for troubleshooting an offshore patrol boat engine. Polish Marit Res 24(2):68–76. https://doi.org/10.1515/pomr-2017-0051 Pamucar D, Bozanic D, Lukovac V, Komazek N (2018) Normalized weighted geometric bonferroni mean operator of interval rough numbers – application in interval rough DEMATELCOPRAS model. Facta Universitatis Series 16(2): 1–22. DOI: https://doi.org/10.22190/ FUME180503018P Parsa M, Jaferi F, Biranvand M (2016) Performance Evaluation using Balanced Scorecard (BSC) and Fuzzy TOPSIS technique Case Study: National Iranian Gas Company. Proceedings of the 2016 International conference on industrial engineering and operations management Detroit, Michigan, September, 23–25 Popovic G, Stanujkic D, Stojanovic S (2012) Investment project selection by applying copras method and imprecise data. Serbian J Manage 7(2):257–269. https://doi.org/10.5937/sjm7-2268 Pourebrahim S, Hadipour M, Mokhtar MB, Taghavi S (2014) Application of VIKOR and fuzzy AHP for conservation priority assessment in coastal areas: case of Khuzestan district, Iran. Ocean Coast Manag 96:1–7. https://doi.org/10.1016/j.ocecoaman.2014.05.009 Purushothaman J, Sajanlal PR, Ponnavaikko M, Pradeep T (2011) Nano and sub-micro inclusions as probes into the origin and history of natural diamonds. Diam Relat Mater 20(7):1050–1055. https://doi.org/10.1016/j.diamond.2011.06.003 Qavamabadi MHR (2013) Environmental protection in the constitutional rights of the Islamic Republic of Iran. Q J Judicial Law Perspect 63:93–140. [In Persian] Radfar R, Ebrahimi L (2012) Fuzzy multi criteria decision making model for prioritizing the investment methods in technology transfer in shipping industries. Proceedings of the 2012 international conference on industrial engineering and operations management Istanbul, Turkey, July 3–6 Rahdari AH (2016) Developing a fuzzy corporate performance rating system: a petrochemical industry case study. J Clean Prod 131:421–434. https://doi.org/10.1016/j.jclepro.2016.05.007 Roshandel J, Miri-Nargesi SS, Hatami-Shirkouhi L (2013) Evaluating and selecting the supplier in detergent production industry using hierarchical fuzzy TOPSIS. Appl Math Model 37(24):10170–10181. https://doi.org/10.1016/j.apm.2013.05.043 Rostami AAA, Saberi M, Hamidian M, Esfandiar Pour M (2017) Evaluating and ranking the firms in chemical industry listed in Tehran stock exchange with TOPSIS. Adv Math Finan Appl 2(2):73–79 Salehi M, Hematfar M (2012) Comparing linear and non-linear relationships between accounting variables and dividend: evidence of Iranian chemical industries. Afr J Bus Manag 6(6):2143– 2151. https://doi.org/10.5897/AJBM11.2484 Shahroudi SM (2011) Estimation of small-scale industries production functions. Int J Econ Bus Model 2(2):139–141 Shaverdi M, Heshmati MR, Eskandaripour E, Akbaritabar A (2013) Developing sustainable SCM evaluation model using fuzzy AHP in publishing industry. Proc Comput Sci 17:340–349. https:// doi.org/10.1016/j.procs.2013.05.044 Shirazia F, Kazemipoor H, Tavakkoli-Moghaddam R (2017) Fuzzy decision analysis for project scope change management. J Decis Sci Lett 6:395–406. https://doi.org/10.5267/j.dsl.2017.1.003 Sorayaei A, Salimi MR, Divkolaii MS (2012) Probing efficiency scale of fuzzy neural network on forecasting stock exchange of the automobile industries in Iran. Indian J Sci Technol 5(4):2590– 2592

70 Iranian Small and Medium-Sized Industries

1803

Taheriyoun M, Karamouz M, Baghvand A (2010) Development of an entropy- based fuzzy eutrophication index for reservoir water quality evaluation. Iran J Environ Health Sci Eng 7(1):1–14 Thipparat T, Thaseepetch T (2013) An integrated VIKOR and fuzzy AHP method for assessing a sustainable research project. World Appl Sci J 22(12):1729–1738. https://doi.org/10.5829/ idosi.wasj.2013.22.12.2787 Tobiszewski M, Namie´snik J, Pena-Pereira F (2017) Environmental risk-based ranking of solvents using the combination of a multimedia model and multi-criteria decision analysis. Green Chem 19:1034–1042. https://doi.org/10.1039/C6GC03424A Valmohammadi C (2010) Identification and prioritization of critical success factors of knowledge management in Iranian SMEs: an experts’ view. Afr J Bus Manag 4(6):915–924 Vujicic M, Blagojevic M, Papic M (2016) Application of COPRAS MCDM method for choosing the best compact fluorescent lamp. International science conference 18–19 November 2016, GABROVO Yalcin N, Bayrakdaroglu A, Kahraman C (2012) Application of fuzzy multi-criteria decision making methods for financial performance evaluation of Turkish manufacturing industries. Expert Syst Appl 39:350–364. https://doi.org/10.1016/j.eswa.2011.07.024 Yazdani M, Hashemkhani Zolfani S, Zavadskas EK (2016) A new integration of MCDM methods and QFD in selecting green suppliers. J Bus Econ Manag 17(6):1097–1113. https://doi.org/ 10.3846/16111699.2016.1165282 Yazdani MH, Ramazantash DH, Mostafahpour M (2017) Analysis of health indicators status in districts of Ardabil Province. J Health 7(5):687–697 Yazdani-Chamzini A, Fouladgar MM, Zavadskas EK, Moini SHH (2013) Selecting the optimal renewable energy using multi criteria decision making. J Bus Econ Manag 14(5):957–978. https://doi.org/10.3846/16111699.2013.766257 Yoshimoto M, Furusawa M, Nakajima K, Takakura M, Hishitani Y (2001) Diamond film growth in an oxygen atmosphere. Diam Relat Mater 10(3–7):295–299. https://doi.org/10.1016/S09259635(00)00485-4 Yunus MFM, Taib CA, Iteng R (2016) A preliminary study on the application of statistical process control (SPC) towards process efficiency: case study in food industries. Rosman Iteng/Sains Humanika 8(4–2):25–31. https://doi.org/10.11113/sh.v8n4-2.1055 Zagorskas J, Zavadskas EK, Turskis Z, Burinskiene M, Blumberga A, Blumberga D (2014) Thermal insulation alternatives of historic brick buildings in Baltic Sea region. Energ Buildings 78:35–42 Zarbakhshnia N, Soleimani H, Ghaderi H (2018) Sustainable third-party reverse logistics provider evaluation and selection using fuzzy SWARA and developed fuzzy COPRAS in the presence of risk criteria. Appl Soft Comput 65:307–319. https://doi.org/10.1016/j.asoc.2018.01.023 Zhao J, You XY, Liu HC, Wu SM (2017) An extended VIKOR method using intuitionistic fuzzy sets and combination weights for supplier selection. Symmetry 9(9):169: 1–16. https://doi.org/ 10.3390/sym9090169 Zolfani SH, Zavadskas EK (2013) Sustainable development of rural areas’ building structures based on local climate. Proc Eng 57:1295–1301. https://doi.org/10.1016/j.proeng.2013.04.163

Important Websites for Further Information https://en.doe.ir/portal/home/ https://isipo.ir/ https://teh.mimt.gov.ir/ https://www.mimt.gov.ir/ mimt.gov.ir

Decentralized Privacy: A Distributed Ledger Approach

71

Pavlos Papadopoulos, Nikolaos Pitropakis, and William J. Buchanan

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Privacy Paradox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Blockchain Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hyperledger Fabric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Distributed Ledger Technology Use Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Healthcare Oriented Use Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Passive DNS Use Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . How Blockchain Technology Ensures Data Privacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Internet of Things . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Big Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cloud Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Supplementary Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1806 1807 1807 1808 1812 1817 1817 1820 1824 1824 1824 1825 1826 1827 1827 1827

Abstract

Our world due to the technological progress became fast-paced and is constantly evolving, thus changing every single day. Consequently, the most valuable asset on earth is not gold or oil anymore but data. Big data companies try to take advantage of this situation and maximize their profits using people’s data. It is common to offer free of charge services to collect vast amounts of data that

P. Papadopoulos () · N. Pitropakis () · W. J. Buchanan Blockpass ID Lab, Edinburgh Napier University, Edinburgh, UK e-mail: [email protected]; [email protected]; [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_58

1805

1806

P. Papadopoulos et al.

most of the times are considered personal and sensitive. On the other hand, their users, everyday and ordinary people without any technological background, take the opportunity to communicate with their friends and family freely and use their services without hesitation and second thoughts. But, if we take a step back and observe these free of charge services, the first concern that comes out of our mind is how all these companies are profitable since they offer a free product/platform/service. Most of the times, big data companies sell their users’ data to advertisers to offer personalized ads. There is an innovative technology that can securely store the aforementioned data but also enable its monetization to the actual producers of it, the people. This technology is the blockchain and its variants. Keywords

Privacy · Big data · Distributed ledger technology · Blockchain · Hyperledger Fabric

Introduction Privacy concerns attract more popularity over time. Europe had already taken care of the privacy of the citizens to a certain point, from 1995 with the Data Protection Directive (DPD) (Yeh 2018). This legislation aimed to facilitate cross-border data transfer and required an absolute recognition of individual privacy rights. Further harmonization occurred under the General Data Protection Regulation (GDPR). GDPR implies there is always a controller that can be held responsible for complying with a set of regulations if personal data are processed. These regulations are (1) the principle of awfulness, fairness and transparency; (2) the principle of purpose limitation; (3) the principle of data minimization; (4) the principle of accuracy; (5) the principle of storage limitation; (6) the principle of integrity and confidentiality; and (7) the principle of accountability. GDPR specifies that the processing of special, sensitive personal data and data relating to criminal convictions and offences is forbidden outside specific regulated circumstances or without explicit consent. Additionally, data controllers must be transparent about the processing of personal data, including the purposes for which data are processed. Individuals can apply their data subject rights, including the rights to access data, request correction of incorrect data, erasure (the right to be forgotten), data portability, and dispute to and not be subjected to a decision based merely on automated processing, including through profiling. Big data companies that collect citizens’ personal data were acting uncontrollably these years, in the altar of profit (Yeh 2018). GDPR legislation regulates this kind of actions and gives the opportunity on people to control their personal data. This can be further enhanced by utilizing the blockchain technology. Except for the ability it gives to people, to sell their data to companies, it also helps to protect their privacy from devices they use daily. In addition to that, sometimes citizen’s views on privacy

71 Decentralized Privacy: A Distributed Ledger Approach

1807

change if they have some kind of profit that directly impacts their lives (Kokolakis 2017). Blockchain technology provides an innovative concept for information storage that is able to build trust within a non-trusted network, by executing and reserving time-stamped transactions. Cybersecurity and cryptography disciplines could evolve by elaborating blockchain technologies in the future, with multiple case studies ranging from financial applications that utilize cryptocurrency ecosystems such as the Bitcoin to automated electronic health record (EHR) management systems that can perform functions without involving human interaction. Recently, the interest of the blockchain technology has grown exponentially both in academia and industry. However, since it is still immature, the security and privacy challenges of the blockchains are being discussed extensively when deploying blockchain in distinctive applications (Zhang et al. 2019). Blockchain technology allows the participants within a transaction to be sure that the rest can view identical information. This technology does not require a trusted intermediary to approve and present each transaction. Often the “blockchain” term is confused with the “distributed ledger technology” (DLT), which is the underlying framework of blockchains. DLT is similar to a decentralized database that is controlled by various participating entities, where blockchain is a type of DLT that each stored transactions on the ledger that are certified with a unique cryptographic stamp called a hash. A number of transactions compose a blockchain block, and each one carries a hash of the previous block. The very first blockchain block is called genesis block. Blockchains get their name from the fact that their blocks are chained together (Zhang et al. 2019). DLT is one of the most significant advancements of the recent years, with features such as a decentralized architecture, immutability and transparency. These features offer a new approach to data storage and protection. Generally, a reliable method to store data is using a conventional database, and, similarly, a private permissioned DLT does not majorly diverge from it. The actual power of blockchains shines in public permissioned or permissionless DLTs. Public blockchains are compliant with various data protection legislation by enabling powerful cryptographic mechanisms to achieve it. However, particular features may affect the users’ privacy if not precisely set, and vast ecosystems such the Bitcoin ledger may collapse if not created properly. A non-rigorous decentralized infrastructure may harm more than a centralized equivalent (Zhang et al. 2019). Recently, various applications had been developed that utilize DLT and blockchains for several use cases and disciplines (Casino et al. 2019). In Fig. 1, a few crucial case studies are presented.

Background Knowledge Privacy Paradox According to Kokolakis (2017), the definition of privacy is composed of three main aspects:

1808

P. Papadopoulos et al.

Fig. 1 Blockchain development overview

1. Territorial privacy, which describes the physical area surrounding a person 2. Privacy of a person, which refers to the protection of a person against unjustified interference, like physical search 3. Informational privacy, which is how personal information is gathered, stored, processed, and interpreted However, an interesting question that derives is “Do people care about their privacy?” (Kokolakis 2017). More and more people use services like social media, arbitrary, without realizing possible costs for their privacy. The article unveiled that the privacy paradox is common when people who express that they are responsible regarding their privacy and their data tend to disclose their sensitive personal data, when they have an economic incentive, such as discounts. The reason is that the judgment for a decision is being taken at the time of question and not before it, without realizing potential security risks. Privacy has no direct impact on their lives, and they do not have anything to gain in return; for all that, they choose an immediate profit like a discount (Kokolakis 2017).

Blockchain Technology Blockchain is a technology created by Satoshi Nakamoto which originally served as the underlying technology for the cryptographic currency named Bitcoin – both originally described in Bitcoin’s whitepaper (Nakamoto et al. 2008). The identity of Satoshi Nakamoto remains a mystery, with a belief that this name belongs to a group of people (Lemieux 2013). The authors established trust in a distributed system designed for finance transactions. Moreover, the solution is based on digital signatures, timestamps, and distributed storage among peers where not a single entity is allowed to tamper data. Furthermore, blockchain technology can be adopted by more areas of interest in addition to finance. Scientists have developed techniques and mechanisms that use the advantages of blockchain technology such as the

71 Decentralized Privacy: A Distributed Ledger Approach

1809

immutability that leads to a tampered proof ledger and the ease of auditing since stored information can be publicly available (Carlozo 2017). Blockchain technology may be perceived as the development of secure cryptographic algorithm applications on modern decentralized databases. In specific, a distributed ledger is a sequence of blocks containing a complete history of transaction records and a cryptographic hash value of the previous linked block. Additionally, the very first block of the chain is called the genesis block. As mentioned the blockchain network is a distributed append-only timestamped peer-to-peer (P2P) network, where non-trusted entities can securely interact with each while eliminating the requirement of a central trusted jurisdiction (Casino et al. 2019). There are several participating entities, called nodes that verify each transaction. These nodes verify and validate both the participating users and the produced transaction by successfully calculating the corresponding hash value through adopting an authoritative consensus algorithm, in order to generate a new transaction, that would be included in a block of transactions (Zheng et al. 2018). In a blockchain, the decentralized network of peers is capable of storing data in a ledger where no participating party can arbitrarily change the data. All participants and data are digitally signed to create a distributed trust relationship. Each peer keeps a history of all the transactions, and each transaction must be approved by the majority of the network, eradicating circumstances of tampered data (Nakamoto et al. 2008). Some variations of the blockchain technology include a cryptocurrency such as the Bitcoin and Ethereum (Wood et al. 2014), to store transactions on the ledger. Some other technologies, such as the Hyperledger Fabric, do not involve a cryptocurrency at all (Androulaki et al. 2018). A distributed ledger network handles the following cryptographic mechanisms in order for the blockchain structure to operate successfully, while the identity of an associated user is preserved, and the validation of the generated transactions is continuously monitored: – Asymmetric cryptography – a procedure also recognized as public-key cryptography, where a pair of cryptographic keys is operated in order to achieve proper encryption and authentication – Cryptographic hash function – a mathematical algorithm that is being used by the computer software that produces a one-way value known as a hash which protects the integrity of data The hash function such as SHA-256 that is currently deployed by blockchain frameworks like Bitcoin is capable of mapping an arbitrary-length data input to an exclusive fixed-length binary output while eliminating the probability of hijacking the generated output, in order to recover the original input or produce the exact output for two different inputs (Wang et al. 2018). Therefore, hashing algorithms allow blockchain nodes to create a digital digest of the transaction, known as a hash pointer, and append the hash value to the original block, thus generating a digitally certified document (Zhang et al. 2019).

1810

P. Papadopoulos et al.

In addition, an asymmetric cryptographic mechanism such as the Elliptic Curve Digital Signature Algorithm (ECDSA) can be utilized to assure reliable encryption and authentication. Specifically, in the work of Wang et al. (2018), the authors address how the users practice a private key, which remains hidden to the entire blockchain network, as a digital signature function, to provide a fixed-length string for any random-length input, whereas a widely acknowledged key is associated with a verification function in order to validate a signed transaction. It should be mentioned that it is mathematically infeasible to produce an identical public key with a dissimilar private key and the other way round (Zhang et al. 2019). As a result, the aforementioned cryptographic procedures are integrated with regard to protecting the data integrity through hashing while successfully encrypting the related auditable transaction and validating the authenticity of a blockchain node. Moreover, effective hashing procedures are utilized in the context of blockchain transaction size reduction. In specific, in the work of Wang et al. (2018), the authors present that each block includes both a hash pointer within the block header and the hashcodes of the associated blocks, in the form of a cryptographic data structure of a Merkle tree, to preserve the integrity of the generated chain. The related binary tree contains both the limited size hashcode of each transaction in the form of a leaf and the hash values of the sequential child leaves, where the root node of the Merkle tree is known as Merkle root. In addition, a typical block structure contains a block header and a primary block body. Specifically, according to Zheng et al. (2018), a block header is consisting of: – – – – – –

Block version – a representation of the block validation rules Parent block hash – a 256-bit hash value that indicates the previous block Merkle tree root hash – a hash value of all the activities within a block Timestamp – present time as seconds in universal time since 1 January 1970 nBits – ongoing hashing target of the valid block hash Nonce – a 4-byte value, which is initiated with a zero value and then rises as the hash calculation advances

Furthermore, the main block body is composed of a transaction counter and transaction data. Strong validation and protection of data are achieved by utilizing feasible cryptographic mechanisms. The amount and the authenticity of the validators and the end users of the described decentralized platform, which generates the transactions, may be initially specified (permissioned), while public and anonymous approaches could be implemented alternatively (public/permissionless). One such case example is Hyperledger Fabric, a permissioned open-source blockchain platform, which endorses strong security and identity features (Stamatellis et al. 2020). The DLT and the blockchains demonstrate several significant benefits over distributed database management systems (DDBMS). In the work of Kuo et al. (2017), the authors address the five assets of blockchain technology as follows:

71 Decentralized Privacy: A Distributed Ledger Approach

1811

– Decentralization: A peer-to-peer network that provides decentralized database management as all non-trusted entities can individually comply to a regulated set of rules to operate on the distributed ledger network. – Immutability: One of the most fundamental attributes of blockchain is that the authorized participating nodes may only view and create transactions; hence, blockchain is suitable to record and manage critical records in an immutable ledger. – Data provenance: Blockchain data ownership may be altered only by the data owner, while the origin of the information is identifiable and could be examined as a ledger verifying technique. – Robustness and availability: The complete record of the verified transactions is held by each participating node independently; therefore, blockchain achieves a high level of data redundancy and availability. – Security and privacy: Effective implementation of the National Institute of Standards and Technology (NIST)-certified cryptographic procedures such as the SHA-256 and 256-bit ECDSA produces secure identities for the users and preserves the digital asset management.

Blockchain Consensus Algorithms This subsection presents a review of various consensus mechanisms while critically investigates the potential benefits and difficulties that they have, regarding the efficient selection of blockchain consensus algorithm. A distributed P2P network consensus scheme requires a formal agreement between the blockchain validators in order to interact with each other to ensure the successful authentication, integrity of the transactions, non-repudiation, sufficient fault tolerance, decentralized governance, and efficient network performance (Baliga 2017). Furthermore, there are various proposed consensus algorithms implemented in the blockchain frameworks that involve digital currencies referred to as cryptocurrencies. The two leading criteria that define a blockchain architecture and a recommended general agreement process, as proposed by Kravchenko (2016), are as follows: – Level of anonymity of validators: The nodes which approve the blockchain transactions or participate in the distributed network may either be anonymous (public blockchain) or verified to a certain extent (private blockchain) through identity certificates. – Level of trust within the validators: The validators’ authority to interact in a blockchain network and a penalty for misconduct are to be defined in each incident; therefore, permissionless and permissioned blockchain designs refer to the classification of the node and user permissions. Moreover, the fundamental consensus protocols, from which numerous variations are derived and are currently being deployed in several blockchain applications, are the following:

1812

P. Papadopoulos et al.

– Proof of work (PoW): Bitcoin blockchain utilizes the corresponding consensus mechanism as a method of establishing proof that each node has performed an amount of work for a block to be appended to the distributed network chain (Baliga 2017). In specific, in the work of Baliga (2017), the author analyzes how the first node which successfully computes a hash value of the desired block, through a dynamically rising difficulty level process, broadcasts it over the P2P network, and receives a mining reward. Additionally, the author also notices that the Bitcoin PoW consensus algorithm operates satisfactorily in public and permissionless network where every participant engages as a validator with no prior knowledge or authentication. However, in the work of (Chaudhry and Yousaf 2018), the authors examine that the PoW consensus algorithm provides notable scalability regarding the amount of the participating nodes but at the same time produces trivial transaction processing rate and massive energy consumption. – Proof of stake (PoS): The concept of proof of stake consensus design is that each node willing to participate in the mining procedure needs to own an amount of the associated cryptocurrency. That introduces the appropriate blockchain license and is bet as a reward to be received in the event of a lucrative block contribution or as a penalty to be deducted in the event of fraudulent activity (Nguyen and Kim 2018). In the work of Baliga (2017), the author presents that a validator’s amount of stakes provides analogous block creation possibilities through a pseudorandom selection process. The PoS algorithm is designed to overcome the disadvantages of PoW design in terms of energy consumption. – Practical Byzantine Fault Tolerance (PBFT): The Byzantine Fault Tolerancebased consensus has been deployed by the Hyperledger Fabric framework, as a technique of eliminating crashing and corrupted nodes when reaching consensus between the verified validators of a privately distributed network (Nguyen and Kim 2018). Furthermore, three-round message exchanges are utilized between the participating nodes in order to reach to an agreement. Additionally, the PBFT consensus model demonstrates moderate scalability of a peer network and decentralization, due to the amount of exchanging messages. Since the identity of nodes is already distinguished, it supports high transaction rates and no economical cost for the participation (Baliga 2017). – Federated Byzantine Agreement (FBA): A variation of the Byzantine Fault Tolerance consensus model is adopted by blockchain frameworks such as Stellar and Ripple whose task is to achieve open-ended participation of trusted end users (Baliga 2017). Nevertheless, the FBA model does not achieve optimal safety against ill-behaved nodes in contrast to PBFT (Bach et al. 2018).

Hyperledger Fabric Permissioned blockchains originally developed for small to medium enterprise networks where the identity of each participating entity can be validated. Hyperledger Fabric is a distributed ledger technology where each action of the participating

71 Decentralized Privacy: A Distributed Ledger Approach

1813

entities can be specified. The blockchain can be private to one or several organizations that form a consortium. It is also possible that different ledgers can be present and only authorized organizations and entities have access to them. The consensus mechanism can be defined during the time of creation, and complex fault-tolerant algorithms can be used for each transaction’s validation (Androulaki et al. 2018). Each entity is validating itself on each action on the ledger, and a Membership Service Provider (MSP) is used to generate and validate their identities. Hyperledger Fabric allows the use of chaincode to perform actions on the ledger. Chaincode is a blockchain program that runs autonomously performing a set of actions defined by the developer (Androulaki et al. 2018). It shares the same logic as the smart contracts of Ethereum (Wood et al. 2014), though the main difference is that, in Ethereum, the program code is written in a blockchain-specific programming language, named Solidity. Nevertheless, in Hyperledger Fabric, the chaincode is written in generalpurpose programming languages such as Java, JavaScript, or Go. The transactions in Hyperledger Fabric adhere to the following order: – Execution: Each peer executes the chaincode according to the designated policy to interact with the blockchain ledger and signs the transaction with its obtained credentials from a Membership Service Provider (MSP), an entity that is responsible for the identities’ management of all the participants. – Order: Each peer sends the constructed transaction to the ordering service, which is a group of nodes also referred to as orderers. The orderers are able to combine various accepted transactions into a single block that is transmitted to all participating peers. – Validation: Each peer receives the block of transactions, verifies these transactions according to the specified policy, and updates its local ledger state. Hyperledger Fabric employs novel security mechanisms such as the private data collection, which allows specific data to be accessed only from particular authorized participants (Papadopoulos et al. 2020). Additionally, Hyperledger Fabric is able to utilize sophisticated zero-knowledge proof (ZKP) security mechanisms to create authorized identities and ensure the anonymity of its users, such as the Identity Mixer (Idemix) cryptographic protocol (Stamatellis et al. 2020). Each participant is associated with an identity certificate in order to interact with the distributed ledger. The identities issued are X.509 digital certificates signed from the certificate authority (CA) and examined from the corresponding MSP. These identities can be generated with the cryptogen tool for development environments during the creation of the system (Androulaki et al. 2018). The X.509 digital certificates issued by the CA involve cryptographic techniques that use the public keys of the users in combination with the private key of the CA (Chokhani et al. 1999). Additionally, Hyperledger Fabric’s team has developed various projects such as Hyperledger Iroha, Burrow, Cello, Composer, and Explorer combining features from other blockchain technologies, extending its capabilities and offering qualityof-life improvements to its developers. Hyperledger Composer is an extension of the original project that offers the creation and management of a blockchain project

1814

P. Papadopoulos et al.

in a development environment. Hyperledger Explorer provides a visualization of the whole blockchain network to its developers, thus enabling its management via a graphical user interface (Dhillon et al. 2017). In Hyperledger Fabric, the participating entities can be constructed using Docker containers. A Docker container (Bernstein 2014) is a virtualization method, often confused with a virtual machine (VM). Docker containers use the host operating system instead of their own, contrary to VMs, and only the Dockerized applications run isolated. The Dockerized applications include all the related programming code and dependencies to execute effectively. Moreover, a Docker container is a lightweight deployment compared to a typical virtual machine that needs fewer resources while providing the same functionality from the blockchain’s perspective. Another critical feature that extends their capabilities is that multiple Docker containers can exist under the same Kubernetes cluster. Kubernetes (Bernstein 2014) is an open-source platform created by Google that allows the orchestration and management of groups of Docker containers. Coexisting Docker containers form a pod that can be separated from other pods in the same Kubernetes cluster. Kubernetes offers semi-infinite scalability to its applications since new Docker containers can be added automatically when needed, sharing the same features as the rest of them. Kubernetes provides self-recovering capabilities from fails to its applications, alongside with the management of the distribution of hardware units. In Hyperledger Fabric, the scalability is considered from the number of peers, organizations, ordering services, and channels. The chaincode that contains all the blockchain’s logic, security structures, and competences is installed and instantiated in the peers and the orderer nodes (Papadopoulos et al. 2020). The main advantage of chaincode written in Go programming language is that it requires fewer resources to run in each container than in JavaScript that needs a library of modules to be installed. During the creation of the blockchain network, the state database that each entity is going to use can be defined. An example of that is CouchDB (Thakkar et al. 2018). CouchDB is a complete database available in Hyperledger Fabric that stores data in key-value pairs and also offers rich queries to them. Using rich queries from the CouchDB and a set of APIs, data can be available to users in many forms, covering their needs and extending the capabilities of the infrastructure as a whole (Thakkar et al. 2018). Peers possess the most crucial role in the blockchain network since they install the chaincodes and host the blockchain ledgers. Peers can communicate privately with other participants by hosting multiple ledgers and chaincodes. That can be achieved by creating private Channels where groups of peers can interact privately with participants only within the channel. Each peer should join a channel to interact with others and perform actions on the ledger. In addition to that, peers are part of Organizations. A group of peers forms an organization, and the permissions of the whole organization could be defined by the established policies (Papadopoulos et al. 2020). The blockchain network is formed by different organizations that all together form a consortium (Dhillon et al. 2017). In the case that a peer fails, the

71 Decentralized Privacy: A Distributed Ledger Approach

1815

other peers continue to operate normally. When the peer recovers back, it uses the gossip protocol to update its ledger from the other peers (Papadopoulos et al. 2020). Each peer holds its identity certificate which has been composed by the certificate authority. This certificate is a .X509 digital certificate that contains all the required information about its owner (Papadopoulos et al. 2020). For the validation of those certificates, another entity called Membership Service Provider (MSP) is able to authenticate the identity of each blockchain participant. The role of this entity is to manage and examine all the cryptographic mechanisms and certificates that peers are using to perform actions in the ledger (Androulaki et al. 2018). Ordering service is the entity that receives the transactions from the peers’ applications and updates the ledger according to the defined consensus. The administrator of the blockchain defines the consensus that the orderers are using to approve or reject transactions (Androulaki et al. 2018). In test environments, only one orderer is needed for the creation of the blocks. The problem of a single point of failure arises for the writes in the ledger. It can be easily prevented in a production environment where more ordering services are being used under a Kafka cluster (Androulaki et al. 2018). Hyperledger Fabric is able to use Apache Kafka to create a cluster of ordering services to create new blocks on the ledger. Peers of the blockchain network are sending the transactions directly to the Kafka broker, and it specifies the orderer that is going to create the new block. In situations of an orderer failure, the operations of the blockchain can be continued normally, as long as one or more orderers are still available. When the orderer approves a transaction, broadcasts it to the peers in order to update their own ledgers, where each one of them is performing a validation of the transaction. Peer nodes that approve the transaction are updating their local ledgers. A peer node is participating in the agreed consensus protocol for the rest of its life cycle since it has to ensure that the data it possesses is remaining valid (Androulaki et al. 2018). The administrators are configuring the agreed consensus protocol in Hyperledger Fabric. It is not a proof-of-work or proof-of-stake algorithm; instead, it can be configured to be one such as Paxos, RAFT, or even one of the BFT algorithms. Since the agreed consensus protocol is something different than resource-hungry PoW, the performance of the blockchain is better. Hence, with a BFT consensus protocol, it is able for the system to defend in situations where adversaries took control of some peers, continuing its operations normally (Baliga 2017). For each transaction in Hyperledger Fabric, each peer is executing the installed chaincode and signs its result with his identity. The corresponding MSP examines its identity, and if it succeeds, it sends the transaction to the orderer. According to the defined consensus, the orderer rejects a transaction or creates a new block on the ledger, signs it with its own identity and delivers it to the peers. Lastly, each peer that receives the new block checks the orderer’s identity and then saves it to its correlated ledger. Transactions are stored in a state database on each peer with the most common to be GoLevelDB and CouchDB (Thakkar et al. 2018). An overview of the architecture in Hyperledger Fabric can be seen in Fig. 2. Another essential feature of Hyperledger Fabric is the gossip protocol. Peers can initiate the gossip protocol after a crash, to query other peers of the network,

1816

P. Papadopoulos et al.

Fig. 2 Hyperledger Fabric overview (Papadopoulos et al. 2020)

for potential updates to the ledger (Androulaki et al. 2018). Furthermore, peers are using the gossip protocol to update their private data collections that only authorized entities hold a copy of it. Since Hyperledger Fabric v1.2, private data collections can be created and configured to allow access to specific data only to authorized participants in a single channel. The private data is sent peer-to-peer to each authorized participant via the gossip protocol. All the other peers have a hash of the data for proof of evidence in auditing. Private channels are used over private data collection when peers want to keep entire ledgers and transactions private, instead of situations that only a subset of fields must remain private. Consequently, when it comes to private data collection, the data is transferred peer-to-peer, and it remains private from the ordering service and even the administrators of the blockchain (Papadopoulos et al. 2020). Hyperledger Fabric provides the necessary chaincode APIs to extend the functionalities of the peers by utilizing command-line (CLI) tools. These APIs are distinguished in the Init API, Invoke API, and Query API. The Init API is used when initialization or upgrade of the chaincode is executed. The Invoke API and Query API are used when storing or reading transactions to the ledger have been performed (Papadopoulos et al. 2020). Peers of the blockchain can store data on the ledger using Hyperledger Fabric’s Invoke API and CLI tools. First, they have to specify their identity and then use the Invoke API with the corresponding storing function and the arguments in JSON format to send each transaction to the orderer. The orderer receives the data and performs the storing function defined. In case of success, this procedure will create

71 Decentralized Privacy: A Distributed Ledger Approach

1817

a new block on the ledger and will send an update signal to each of the peers to update their ledgers. To receive data from the ledger, peers use the Query API, sending a query transaction to the orderer. Peers should specify their identity and then use the query function with the arguments in JSON format to send the transaction to the orderer. The orderer receives the transaction and, in continuation, displays only the allowed data to the recipients according to the defined query function and the private data collection configuration. The specified identity functionality enables a peer to query only specific blocks. The chaincode is installed to all the peers and instantiated from the orderer. To store data in the blockchain ledger, the CLI tools can be used via the command line of the Docker containers’ interface. Additionally, the blockchain’s administrator is able to configure a Hyperledger Fabric Software Development Kit (SDK) to interact with the blockchain. SDKs are tools that the administrator can use to manage multiple channels, install and instantiate chaincode or simply invoke and query transactions. SDKs are communicating directly with the Hyperledger Fabric’s APIs for each process, and the officially supported SDKs are written in Node.js and Java. There are more SDKs written in Go, Python, and Rest that are available for testing (Papadopoulos et al. 2020; Androulaki et al. 2018). Without an SDK, the administrator can only use the CLI tools for each process. Data stored in the blockchain’s immutable ledger cannot be manipulated by potentially malicious actors. Each transaction is authorized by the policy, thus making unauthorized requests to be rejected automatically. Each participating entity prior to interacting with the blockchain needs to install the associated chaincode that contains all the blockchain logic and security mechanisms. A collection configuration is developed to advise the orderer about the state of the stored data, the time of their availability until they purge, and each corresponding entity that has access to them. Any access attempt by unauthorized entities is denied. Only authorized entities are allowed to store and receive data from the ledger. Each peer is obliged to prove its identity to the orderer before each transaction. According to the configured policy, the store and query transactions are restricted to peers which are not included in the policy. These fundamental principles eliminate the possibility of a malicious actor to store arbitrary data to the ledger without the correct identity. Furthermore, a malicious actor is not able to query data at all. The private data can be queried only by specified entities, and the rest of the data are available only to participants.

Distributed Ledger Technology Use Cases Healthcare Oriented Use Case Healthcare is one of the disciplines that due to its importance and complexity needed considerable more time to adapt to the new digital era. Health records contain highly sensitive patients’ data, and their privacy and security must be ensured. It

1818

P. Papadopoulos et al.

is common in healthcare institutions to maintain patient’s health records physically in papers. These institutions have to follow several regulations, auditing, and compliance regarding these records, and since the arrival of General Data Privacy Regulation (Voigt and Von dem Bussche 2017) in Europe, their sustainability has been challenged. Healthcare institutions that favored to use EHR instead faced other novel challenges. Privacy, at first, was not a concern, particularly when these records are being routinely shared with other healthcare providers, pharmacies, and patients in order to improve the diagnosis and treatment (Stamatellis et al. 2020). EHR are considered highly sensitive, and they should only be shared with other parties only after patients’ approval and consent. However, their management and sharing with other necessary parties implies and is decisive that the EHR would be encrypted in a way that a correlation to the patients’ identity would not be possible (Papadopoulos et al. 2021; Abramson et al. 2020). Blockchain technology promotes the aforementioned privacy-preserving measures that need to be taken. Therefore, many researchers utilized blockchain’s innovative technology in order to provide novel precautionary measures (Hölbl et al. 2018), but also extend its capabilities in other areas of the healthcare industry, such as drug counterfeiting and medical research (Casino et al. 2019). There are a number of researches that combined blockchain technology with eminent cloud computing. Some of them focused on improving and extending familiar centralized infrastructures with decentralized features and methods (Zhang et al. 2019). These decentralized features involve cloud storage and access in order to improve the system’s availability, scalability, and cost-efficiency; thus, appropriate access policy and identity management are crucial (Papadopoulos et al. 2020; Stamatellis et al. 2020). The perseverance of privacy, the availability, and scalability have been identified as the most important features of an efficient blockchain EHR management infrastructure. Cryptography, in general, has a key role for that perseverance of privacy. In their work Dubovitskaya et al. (2015), the authors utilized a public key infrastructure (PKI) to encrypt patients’ medical data in their cloud storage and local databases, in order to provide a scalable privacy-preserving system. Furthermore, the main challenge for a successful blockchain decentralized EHR management infrastructure is the fact that data sit in multiple devices and organizations. There are several authors who presented efficient and scalable systems without a real-world implementation, Roehrs et al. (2017), Guo et al. (2018), and Patel (2019) yet. Those research attempts include practical proofs of concept developed in both permissioned and permissionless blockchain schemes. Most of the permissioned blockchain infrastructures were developed on top of Hyperledger Fabric framework (Stamatellis et al. 2020; Ichikawa et al. 2017; Liang et al. 2017b), while most of the permissionless blockchain systems were developed on top of the Ethereum network (Azaria et al. 2016; Al Omar et al. 2017; Yang and Yang 2017). The research approaches that utilized Hyperledger Fabric performed their operations faster and more efficient and can be easier extended and operated by multiple devices, such as mobile devices (Stamatellis et al. 2020; Ichikawa et al. 2017; Liang et al. 2017b). From its nature, Hyperledger Fabric identities management

71 Decentralized Privacy: A Distributed Ledger Approach

1819

minimizes the risks of malicious participants, since the identity of each participant is known. However, the uncertainty of an insider attack still exists (Stolfo et al. 2008). Another approach with characteristics similar to permissioned blockchains is MedShare (Xia et al. 2017), although their underlying technology is not explicitly specified. On the other hand, notable researches were utilized on top of the Ethereum network such as MedRec (Azaria et al. 2016; Yang and Yang 2017) and MediBchain (Al Omar et al. 2017). MedRec presented a decentralized and easily auditable EHR management system; however, its scalability, alongside users’ privacy, through anonymity and unlinkability, has been questioned. Their infrastructure has been further extended to enhance the protection mechanisms to a certain degree (Yang and Yang 2017). A combination of the permissionless blockchain network and cloud infrastructure has been proposed in MediBchain (Al Omar et al. 2017). One of its advantageous characteristics is its scalability. However, the cost of each transaction and potential information leakages should be resolved in order for this system to be utilized in production environments.

Challenges A centralized collection of private records in either local or cloud-based databases introduces several loss-making complexities, apart from common disadvantages such as the likelihood of a single point of failure or violation of privacy and anonymization as a result of a third-party service provider’s unethical behavior. Certain healthcare providers misinterpret national regulations such as Health Insurance Portability and Accountability Act (HIPAA) by sharing limited medical information, thus restricting patients and proxies from accessing data while creating costly obstacles with regard to effective EHR distribution (Ivan 2016). In the work of Albeyatti (2018), the author demonstrates how a medical error, which could be generated by ill-informed clinical decisions, is the third leading cause of death in the United States in 2016 and at the same time telemedicine market was estimated to be worth 23.8 billion dollars in 2017 and is projected to exceed 55 billion dollars by 2021. This is due to both intentionally or accidentally tampered record and fragmentary medical information distribution. It should also be noted that conventional infrastructures regarding health record management and storage demonstrate particular threats concerning data breaches and cybersecurity attacks. Healthcare information is considerably more valuable than other industry data for exchanging in the black market regarding unethical or illegal actions, while the average cost of a hijacked medical record is 380 dollars, which is twice the average cost across all industry-related data breaches (Coventry and Branley 2018). Furthermore, the economic incentive leads the malicious actors to craft sophisticated malicious software in order to infect as many machines as possible. That malicious software is commonly in the form of a ransomware, which is a software that completely encrypts all the files of an infected machine until the associated ransom is paid. Traditional decryption techniques are often incompetent, and only a complete reconstruction of the file system is able to restore the system to a normal

1820

P. Papadopoulos et al.

operation state (Coventry and Branley 2018). Recently, there was a disastrous attack such as the aforementioned, namely, WannaCry ransomware, that compromised millions of machines worldwide. Victims of that attack were also governmental bodies such as the National Health Service (NHS) computers and servers, with losses reaching to 92 million pounds (Smart 2018). Finally, in the work of Alvarez (2017), the author examines that weak security mechanisms provided by a third-party vendor that usually offers management solutions to healthcare providers led to an extensive compromisation of over a quarter-million healthcare records from multiple organizations located in the United States. Nevertheless, 68% of all security attacks within healthcare institutions are carried out by malicious insider individuals, who recklessly or unwittingly introduce threats such as a 400,000 Protected Health Information (PHI) record loss from an unencrypted password-protected laptop in February 2016 (Alvarez 2017). The following attack vectors target the healthcare industry and EHR handling: – Command injection: The leading attack mechanism involving manipulation of malicious data input to databases, such as Structured Query Language (SQL) database injection, allows unauthorized access to critical data and the compromisation of users and healthcare facilities. – Data structure manipulation: The attacker attempts to gain unlawful access by exploiting common vulnerabilities existing in current database designs. – System resource manipulation: The resources of a distributed network are manipulated in order for a successful denial of service (DOS) or arbitrary code execution to be achieved, thus undermining availability and data privacy. – Probabilistic technique employment: The malicious actor explores and overcomes the security features of a target by profitably calculating system credentials and gaining access to the healthcare server. Consequently, there is a need for a countermeasure against all the aforementioned attacks. This defensive mechanism needs to be adequately flexible to preserve the privacy of the stored records but at the same time robust to guard them effectively against misuses. The adoption of a distributed ledger technology solution can succeed in those terms and assist various healthcare institutes to defend against insider and malicious attacks.

Passive DNS Use Case Domain Name System (DNS) created to translate servers’ IP addresses into easily remembered names, as in Fig. 3. Each DNS query encapsulates crucial information that can be used from security analysts to identify malicious misuses such as phishing domain names (Papadopoulos et al. 2020; Christou et al. 2020). Passive DNS data is a concept introduced by Florian Weimer (2005), who used recursive name servers to log responses received from different name servers and then copied this logged data to a central database. These passive DNS data includes

71 Decentralized Privacy: A Distributed Ledger Approach

1821

Fig. 3 Overview of the Domain Name System (Papadopoulos et al. 2020)

the queries and responses from the authoritative name servers before the recursive name servers. Taking their example, many researchers have used passive DNS data in conjunction with machine learning to build domain name reputation scoring systems to detect abuse on the Internet (Antonakakis et al. 2010). EXPOSURE (Bilge et al. 2011) follows a similar technique but needs less training and can detect a range of malicious services (e.g., fast-flux networks, phishing, botnets). Khalil et al. (2016) proposed a passive DNS analysis through graphs, using public aggregated passive DNS data. Notos and EXPOSURE rely on DNS queries that may contain sensitive data of the end users. Following a similar way of thinking, Lever et al. (Lever et al. 2016) used passive DNS to identify possible domain ownership changes, while Alrawi et al. (2019) evaluated home-based IoT devices using the DNS traffic. Khalil et al. (2016) relied on public passive DNS databases which belong to companies such as Farsight (DNSDB) (Farsight Security DNSDB: https://www.farsightsecurity.com/solutions/dnsdb/) and VirusTotal (VirusTotal Passive DNS replication: https://www.virustotal.com/gui/home/search). Related research in identification of malicious domain names through passive DNS collection and analysis such as Notos (Antonakakis et al. 2010) assigns a reputation score to each website based on DNS queries. Tian et al. (2018) proposed a novel methodology which detects phishing squatted domains based on a classifier that introduces features from visual analysis and

1822

P. Papadopoulos et al.

optical character recognition (OCR) which managed to overcome the heavy content obfuscation from attackers. Kidmose et al. (2018) stated that the detection should take place during the pre-registration time before the first update to the zone, where it is guaranteed that the domain has not been abused on the Internet, as it has not yet been published in the TLD zone. From another point of view, Piredda et al. (2017) achieved the detection of typosquatted domains by creating a similarity measure using n-gram-based representation and DNS traffic analysis. In a similar way of thinking, Selvi et al. (2019) used masked n-grams to detect algorithmically generated malicious domain names because it provides a great combination of training time and accuracy. Shulman (2014) challenged the security of the DNS and summarized that to effectively and efficiently protect DNS, a combination of mechanisms should apply. The authors proposed that defensive mechanisms such as the recursive authoritative name server (RANS) have the ability to reduce the traffic of the infrastructure and perform their operations faster. They also mentioned that public DNS resolvers could solve similar issues, but the privacy of the end users in the DNS caching is crucial and must be ensured. Ranjan (2012) developed a patent for the identification of DNS fast-flux attacks, where a domain name is changing swift IP addresses to forward users to malicious web servers. The users can have no idea since moments before they may have used the same website that responded to a benign web server. Kambourakis et al. (2007) proposed a system to protect local DNS servers from DNS amplification attacks. This type of attacks aims to waste the recursive DNS server’s resources to perform a DOS to legitimate users. The attackers, to achieve their goal, are sending numerous DNS requests even from various sources to flood the infrastructure. To successfully protect end users, the authors proposed solutions such as the acceptance of DNS queries only from trusted sources reducing the size of the sample significantly (Kambourakis et al. 2007). DNS cache poisoning is a kind of attack where a malicious actor is able to forge the cache of a benign name server with potential malicious information. It is one of the most crucial DNS attacks, and defensive mechanisms are difficult to protect users completely. In his work, Stewart (2003) stated that only with the configurations of DNSSEC name servers can be fully protected. The author created a method where a DNS query goes through two different DNS servers instead, to minimise the possibility of both of them to be exposed (Stewart 2003). Despite the success of Weimer’s concept, the impact of the collection of passive DNS to end user privacy was soon questioned. Users could be clueless if a passive DNS collector is placed in their DNS resolver. In situations where the passive DNS collector is placed in the ISP or the TLD, and the dataset is massive, the privacy issue arises. Since each query can be correlated to each user and their DNS behavior can be tracked, the personal data must remain private (Spring and Huth 2012). Consequently, the security community focused on this issue, and several approaches were published. One of the first approaches was using tools that could eliminate confidential information from collected network packets (Zdrnja 2007). Another point of view

71 Decentralized Privacy: A Distributed Ledger Approach

1823

instructed that a cryptography-based prefix preserving anonymization algorithm should be used to address this issue (Govil and Govil 2007) or other encryption techniques that would secure the IP prefix (Xu et al. 2002). Other researchers trying to overcome this conflict came up with a totally different solution, the collection of active DNS data (Kountouras et al. 2016). In specific, they created a system called Thales which can systematically query and collect large volumes of active DNS data using as input an aggregation of publicly accessible sources of a vast amount of domain names and URLs. These include but are not limited to public blacklists, the Alexa ranking, the Common Crawl project, and various top-level domain (TLD) zone files. Liu et al. (2018) proposed a decentralized DNS (DecDNS) system which has a stored database of DNS records and performs the resolution using the nodes of the blockchain. The advantages and security mechanisms of blockchain by default, such as the tampered proof state of the data and the distributed denial of service (DDoS) attack resilience, are essential features of the system. The DNS resolution is performed by each node of the network, so in the case of a DDoS attack that multiple nodes may go offline since one or more continue operating normally, the proposed system is able to resolve DNS queries. Their attempt is a potential solution to the existing DNS issues without significant changes (Liu et al. 2018). Liang et al. (2017a) proposed a system that combines two technologies such as blockchain and cloud computing to effectively and efficiently create a decentralized storage system as a DNS record database. Regarding the privacy of the data stored, they used a hashed version of the sensitive data to work as a proof of identity, and only the administrator of the system has the ability to correlate each identity to the hashed data. From another point of view, there are some systems developed to change the existing DNS infrastructure. These promising systems such as Namecoin (Kalodner et al. 2015) and Blockstack (Ali et al. 2016) were developed to build a more secure, easily audited, transparent domain name organization. They created a substitution for the Internet Corporation for Assigned Names and Numbers (ICANN), where each user is not relying on a third party to buy a domain name. The proposed system is built on a blockchain network, Bitcoin in their case, where users can mine for the domain name cryptocurrency. Then users are able to use this cryptocurrency to buy domain names with new .bit, .id TLDs that were not existing before (Ali et al. 2016). The privacy of the users can be ensured since their identity is protected from Bitcoin’s identity management. The downside of these systems is that users need specific extensions to be able to query blockchain registered domain names. A system that is built on Hyperledger Fabric, utilizing its privacy features, such as the private data collection, and takes into consideration the existing DNS infrastructure is PRESERVE DNS (Papadopoulos et al. 2020). In their work, the authors presented a secure, scalable, and efficient infrastructure that is able to store passive DNS data, by ensuring the privacy of the end users. In their proof of concept, there is a simulation of a real-world scenario where multiple participating entities have access to the blockchain ledger, with some having access to only specific data and others to all data.

1824

P. Papadopoulos et al.

In a similar approach, by utilizing the Hyperledger Fabric framework, the DNS Trusted Sharing Model (DNSTSM) (Yu et al. 2020) is a high-performance and efficient system that can mitigate various DNS attacks. DNSTSM can be utilized in the current global DNS infrastructure without any changes needed. However, due to the older v1.1 version of Hyperledger Fabric, that the DNSTSM is using, the private data collection feature is not possible to be implemented since it has been introduced in the later v1.4 version. Thus, the DNSTSM needs a complete re-design of its architecture to benefit from the enhanced security mechanisms that became available in the newer versions of its backbone technology. In the literature, the majority of the proposed solutions try to improve DNS by reinventing many of its features or even the whole of it. Judging from the adoption of proposed solutions, it might take longer than expected as the whole Internet infrastructure functions are built on top of the existing DNS form. Consequently, priority should be given to the creation of systems that can be implemented on top of the existing functionalities of DNS, securing it, with always taking the privacy into serious consideration.

Challenges Commonly, the human factor is the weakest link in systems that include certificates and identities such as the Hyperledger Fabric. A malicious user could perform arbitrary queries to the blockchain ledger in a potential theft of the identity certificates of a blockchain participant. The system’s security could be completely exploited according to the endorsed policy. However, in Hyperledger Fabric, this scenario is not possible, since no participating entity controls the blockchain ledger, even the administrators of it. A possible risk inherited from traditional code programming lays into the chaincode that is installed and executed by each peer. Since chaincode is an autonomous piece of code that runs without supervision, extensive examination and testing should occur to ensure that it executes as intended (Androulaki et al. 2018). Another challenge of blockchain technology relies on the fact that it is still immature; new bugs and attacks may be introduced in the future. Moreover, all the systems that involve passwords, encryption mechanisms, and hashes may be at risk when quantum computing would be developed. A direct countermeasure to quantum computing is to utilize quantum-robust techniques from now if they are efficient and expedient (Papadopoulos et al. 2020).

How Blockchain Technology Ensures Data Privacy Internet of Things In the work of Dorri et al. (2017), the authors mention that the Internet of things (IoT) devices produce, process, and exchange immense amounts of security and safety-critical data as well as privacy-sensitive information; therefore, they are an

71 Decentralized Privacy: A Distributed Ledger Approach

1825

appealing target of various cyber-attacks. The author analyzes the example of a blockchain-based smart home and the relationship between privacy and information security of the IoT data, ensuring the three core principles of security: confidentiality, integrity, and availability (CIA). Subsequently, the author presents the example of blockchain-based smart home and explains how privacy can be ensured. Existing security mechanisms are not correctly suited for IoT because of the high energy they consume. Low-resource IoT devices can efficiently use the blockchain technology that delivers a platform to interconnect reliably and avoid the threats that plague central server models. The main characteristic of blockchains is that their data is decentralized and available to all of its nodes. The security and privacy of the stored data can be assured since blockchain is utilizing reliable and strong encryption algorithms. This way of protection also complies with the GDPR and adds an extra layer of protection since an unauthorized transaction is immediately perceived and is rejected automatically. Encrypted data is considered pseudonymized, and GDPR defines that the pseudonymization of data should be performed in a way that a correlation back to each individual cannot be achieved (Voigt and Von dem Bussche 2017). However, one of the concerns of the IoT is data security and privacy. There is intense competition between the IoT developers, and, as a result, they release their devices to the market urgently without ensuring about their stability and security. Unsecure devices may lead to exploitation attacks such as recently in which the Mirai botnet spread in IoT devices (Kelly et al. 2020). In terms of privacy, the GDPR requires that if any personal data, like a name or email address, is exposed during a data breach, the affected individual must be notified. Likewise, users of IoT devices have the right to be forgotten from companies’ data centers. Furthermore, the personal data economy promotes the concept that individuals need to handle their data, and to be in complete control of it, allowing its further commercialization to third parties on their terms (Elvy 2017). This model is based on the exchange of data that is threatening the traditional model of big data companies selling their users’ data.

Big Data Everyone leaves a digital data trail on a daily basis, sometimes by buying a cap of coffee using a credit card or a mobile application. Vendors track and collect data about consumers’ use of their services. Consumers may, at first, regard such use as harmless and believe the vendors use the collected information only for promotional reasons. Many consumers believe that it is not concerning since they do not have anything to “hide” (Solove 2011). However, a promotional strategy is not only what is produced from the collected data. What may initially appear naive at an individual transaction level may not be harmless when data is gathered and aggregated on a large scale through big data analytics. Using months’ or years’ worth of transactions, vendors can construct considerably accurate pictures of individuals. These simple insights into private lives are extremely valuable. According to an official report in 2012, the annual revenue of the nine largest data brokers in the United States was approximately US$426 million (Commission et al. 2012).

1826

P. Papadopoulos et al.

This profit raised the last few years from the increased use of social networks. Such a high return is why most data vendors retain the right, through privacy policies and user agreements, to sell customer information to third parties (Yeh 2018). Since profits are enormous, data brokers may try to bypass GDPR in Europe and continue their operations in the United States since there is no regulative framework about personal data. Particularly, the possibility of exceptions, divergent interpretations, legal cultures, and national laws that lack harmonization remains of concern. Another concern of the author is that because GDPR states that the process of anonymized data for statistical and research objectives is permitted. Hence, data brokers may claim that they process anonymous data, denying the fact they are a data controller or structure their operations to avoid European jurisdiction (Yeh 2018) and taking advantage of this loophole since pseudonymization has not yet been standardized. Although consumers generally do not read the privacy policies, and even when they do, they often do not completely understand them, to make sensible decisions. Consumers, in general, are unaware of how data brokers consolidate, aggregate, analyze, and sell their data. Unlike the legislation in the United States, which leaves personal data largely unprotected in the private sector, European data protection legislation covers all private-sector processing of personal data (Yeh 2018).

Cloud Computing In recent years, cloud computing is being used vastly on a personal and business level, offering benefits such as elasticity of resources. Nevertheless, privacy and security remain a gray area. However, it should be noted that even if data is not stored centralized in servers on-premises, it may be stored in a single cloud platform, thus facing similar issues. The security risks of a centralized infrastructure in the cloud may be reduced but not eliminated. In the work of Roman et al. (2018), the authors suggest methods such as fog computing, mobile edge computing, and mobile cloud computing, to protect the three principles of security, confidentiality, integrity, and availability, from various attacks and malicious intents. Furthermore, the authors analyze threats and challenges that appear in edge data centers and provide the security mechanisms that should be present in all edge examples. Furthermore, various threats such as DDoS, man in the middle, privacy leakage, and privilege escalation attacks could be mitigated successfully when proper security mechanisms set within these infrastructures. The security mechanisms consolidate identity and authentication, access control systems, protocol and network security, error resistance, and durability (Pitropakis et al. 2020). Blockchain technology and particularly the Hyperledger Fabric framework provide these related security mechanisms to protect the stored data from exploitation and malicious usage. By utilizing its novel architecture and backbone technologies, a combination with cloud infrastructure such as Kubernetes can efficiently keep the data private and secure, with high availability (Papadopoulos et al. 2020).

71 Decentralized Privacy: A Distributed Ledger Approach

1827

Conclusion Blockchain technology attracted much interest in the last few years. Mostly, this interest was focused on its financial aspects, the cryptocurrencies. However, it has been evolved to benefit a broader spectrum of fields that can help humanity to solve previously non-investigated problems. This growth created similar architectures such as DLTs. This chapter presents the background knowledge related to one of the most important DLTs, the Hyperledger Fabric, alongside significant and novel related use cases. Additionally, this chapter discussed other areas that can benefit from this innovative technology such as the IoT, Big Data, and Cloud computing. The aim of this chapter was to review the literature in order to provide valuable insights into the described topic. Moreover, its intention is to present the future direction of blockchains and DLTs, to create a secure world. However, the question that derives is whether we can build a secure world, on top of an insecure one (Bruce Schneier in Privacy, Trust and the Future at Edinburgh Napier University: https://www.youtube.com/watch?v=eFmsCSIEMlw). As the whole world becomes data-centric, the privacy of the end users is remarkably valuable, thus motivating all future solutions to aim to preserve it.

Supplementary Material Hyperledger Fabric community provides a range of tutorials and supplementary material on their website (https://hyperledger-fabric.readthedocs.io/en/latest/ tutorials.html) and their GitHub repository (https://github.com/hyperledger/fabricsamples). Additionally, there is a range of case studies focused on the adoption and development of Hyperledger Fabric as a whole (https://www.hyperledger.org/learn/ case-studies).

References Abramson W, Hall AJ, Papadopoulos P, Pitropakis N, Buchanan WJ (2020) A distributed trust framework for privacy-preserving machine learning. In: International conference on trust and privacy in digital business. Springer, pp 205–220 Al Omar A, Rahman MS, Basu A, Kiyomoto S (2017) Medibchain: a blockchain based privacy preserving platform for healthcare data. In: International conference on security, privacy and anonymity in computation, communication and storage. Springer, pp 534–543 Albeyatti A (2018) White paper: medicalchain. MedicalChain self-publication Ali M, Nelson J, Shea R, Freedman MJ (2016) Blockstack: a global naming and storage system secured by blockchains. In: 2016 {USENIX} annual technical conference ({USENIX} {ATC} 16), pp 181–194 Alrawi O, Lever C, Antonakakis M, Monrose F (2019) Sok: security evaluation of home-based IoT deployments. In: IEEE S&P, pp 208–226 Alvarez M (2017) Security trends in the healthcare industry. Somers: IBM, pp 2–18 Androulaki E, Barger A, Bortnikov V, Cachin C, Christidis K, De Caro A, Enyeart D, Ferris C, Laventman G, Manevich Y et al (2018) Hyperledger fabric: a distributed operating system for permissioned blockchains. In: Proceedings of the thirteenth EuroSys conference. ACM, p 30

1828

P. Papadopoulos et al.

Antonakakis M, Perdisci R, Dagon D, Lee W, Feamster N (2010) Building a dynamic reputation system for DNS. In: USENIX security symposium, pp 273–290 Azaria A, Ekblaw A, Vieira T, Lippman A (2016) Medrec: using blockchain for medical data access and permission management. In: 2016 2nd international conference on open and big data (OBD). IEEE, pp 25–30 Bach L, Mihaljevic B, Zagar M (2018) Comparative analysis of blockchain consensus algorithms. In: 2018 41st international convention on information and communication technology, electronics and microelectronics (MIPRO). IEEE, pp 1545–1550 Baliga A (2017) Understanding blockchain consensus models. Persistent 4:1–14. Available online at: https://www.persistent.com/wp-content/uploads/2017/04/WP-UnderstandingBlockchain-Consensus-Models.pdf. Accessed on 01 Aug 2021 Bernstein D (2014) Containers and cloud: from lxc to docker to kubernetes. IEEE Cloud Comput 1(3):81–84 Bilge L, Kirda E, Kruegel C, Balduzzi M (2011) Exposure: finding malicious domains using passive DNS analysis. In: NDSS, pp 1–17 Carlozo L (2017) What is blockchain? J Account 224(1):29 Casino F, Dasaklis TK, Patsakis C (2019) A systematic literature review of blockchain-based applications: current status, classification and open issues. Telematics Inform 36:55–81 Chaudhry N, Yousaf MM (2018) Consensus algorithms in blockchain: comparative analysis, challenges and opportunities. In: 2018 12th international conference on open source systems and technologies (ICOSST). IEEE, pp 54–63 Chokhani S, Ford W, Sabett R, Merrill C, Wu S (1999) Rfc 2527: Internet x. 509 public key infrastructure certificate policy and certification practices framework. Internet Engineering Task Force (IETF), RFC Christou O, Pitropakis N, Papadopoulos P, McKeown S, Buchanan WJ (2020) Phishing url detection through top-level domain analysis: a descriptive approach. arXiv preprint arXiv:2005.06599 Commission FT et al (2012) Protecting consumer privacy in an era of rapid change: recommendations for businesses and policymakers. Federal Trade Commission, Washington, DC Coventry L, Branley D (2018) Cybersecurity in healthcare: a narrative review of trends, threats and ways forward. Maturitas 113:48–52 Dhillon V, Metcalf D, Hooper M (2017) Blockchain enabled applications: understand the blockchain ecosystem and how to make it work for you. Springer, Apress Dorri A, Kanhere SS, Jurdak R, Gauravaram P (2017) Blockchain for IoT security and privacy: the case study of a smart home. In: 2017 IEEE international conference on pervasive computing and communications workshops (PerCom workshops). IEEE, pp 618–623 Dubovitskaya A, Urovi V, Vasirani M, Aberer K, Schumacher MI (2015) A cloud-based ehealth architecture for privacy preserving data integration. In: IFIP international information security and privacy conference. Springer, pp 585–598 Elvy S-A (2017) Paying for privacy and the personal data economy. Columbia Law Rev 117:1369 Govil J, Govil J (2007) 4g mobile communication systems: turns, trends and transition. In: 2007 international conference on convergence information technology (ICCIT 2007). IEEE, pp 13–18 Guo R, Shi H, Zhao Q, Zheng D (2018) Secure attribute-based signature scheme with multiple authorities for blockchain in electronic health records systems. IEEE Access 6:11676–11686 Hölbl M, Kompara M, Kamišali´c A, Nemec Zlatolas L (2018) A systematic review of the use of blockchain in healthcare. Symmetry 10(10):470 Ichikawa D, Kashiyama M, Ueno T (2017) Tamper-resistant mobile health using blockchain technology. JMIR mHealth uHealth 5(7):e111 Ivan D (2016) Moving toward a blockchain-based method for the secure storage of patient records. In: ONC/NIST use of blockchain for healthcare and research workshop. ONC/NIST Gaithersburg, pp 1–11 Kalodner HA, Carlsten M, Ellenbogen P, Bonneau J, Narayanan A (2015) An empirical study of namecoin and lessons for decentralized namespace design. In: WEIS. Citeseer

71 Decentralized Privacy: A Distributed Ledger Approach

1829

Kambourakis G, Moschos T, Geneiatakis D, Gritzalis S (2007) Detecting DNS amplification attacks. In: International workshop on critical information infrastructures security. Springer, pp 185–196 Kelly C, Pitropakis N, McKeown S, Lambrinoudakis C (2020) Testing and hardening iot devices against the mirai botnet. In: 2020 international conference on cyber security and protection of digital services (cyber security). IEEE, pp 1–8 Khalil I, Yu T, Guan B (2016) Discovering malicious domains through passive DNS data graph analysis. In: Proceedings of the 11th ACM on Asia conference on computer and communications security. ACM, pp 663–674 Kidmose E, Lansing E, Brandbyge S, Pedersen JM (2018) Detection of malicious and abusive domain names. In: 2018 1st international conference on data intelligence and security (ICDIS). IEEE, pp 49–56 Kokolakis S (2017) Privacy attitudes and privacy behaviour: a review of current research on the privacy paradox phenomenon. Comput Secur 64:122–134 Kountouras A, Kintis P, Lever C, Chen Y, Nadji Y, Dagon D, Antonakakis M, Joffe R (2016) Enabling network security through active DNS datasets. In: International symposium on research in attacks, intrusions, and defenses. Springer, pp 188–208 Kravchenko P (2016) Ok, I need a blockchain, but which one. Medium. Available online at: https:// medium.com/@pavelkravchenko/ca75c1e2100 Kuo T-T, Kim H-E, Ohno-Machado L (2017) Blockchain distributed ledger technologies for biomedical and health care applications. J Am Med Inform Assoc 24(6):1211–1220 Lemieux P (2013) Who is satoshi nakamoto? Regulation 36(3):14 Lever C, Walls R, Nadji Y, Dagon D, McDaniel P, Antonakakis M (2016) Domain-z: 28 registrations later measuring the exploitation of residual trust in domains. In: 2016 IEEE symposium on security and privacy (SP). IEEE, pp 691–706 Liang X, Shetty S, Tosh D, Kamhoua C, Kwiat K, Njilla L (2017a) Provchain: a blockchainbased data provenance architecture in cloud environment with enhanced privacy and availability. In: Proceedings of the 17th IEEE/ACM international symposium on cluster, cloud and grid computing. IEEE Press, pp 468–477 Liang X, Zhao J, Shetty S, Liu J, Li D (2017b) Integrating blockchain for data sharing and collaboration in mobile healthcare applications. In: 2017 IEEE 28th annual international symposium on personal, indoor, and mobile radio communications (PIMRC). IEEE, pp 1–5 Liu J, Li B, Chen L, Hou M, Xiang F, Wang P (2018) A data storage method based on blockchain for decentralization DNS. In: 2018 IEEE third international conference on data science in cyberspace (DSC). IEEE, pp 189–196 Nakamoto S et al (2008) Bitcoin: a peer-to-peer electronic cash system Nguyen G-T, Kim K (2018) A survey about consensus algorithms used in Blockchain. J Inform Process Syst 14(1):101–128 Papadopoulos P, Abramson W, Hall AJ, Pitropakis N, Buchanan WJ (2021) Privacy and trust redefined in federated machine learning. Mach Learn Knowl Extract 3(2):333–356 Papadopoulos P, Pitropakis N, Buchanan WJ, Lo O, Katsikas S (2020) Privacy-preserving passive DNS. Computers 9(3):64 Patel V (2019) A framework for secure and decentralized sharing of medical imaging data via blockchain consensus. Health Inform J 25(4):1398–1411 Piredda P, Ariu D, Biggio B, Corona I, Piras L, Giacinto G, Roli F (2017) Deepsquatting: learningbased typosquatting detection at deeper domain levels. In: Conference of the Italian association for artificial intelligence. Springer, pp 347–358 Pitropakis N, Katsikas S, Lambrinoudakis C (2020) Cloud security, privacy and trust baselines. In: Cloud computing security: foundations and challenges, CRC Press, p 45–64 Ranjan S (2012) Detecting DNS fast-flux anomalies. US Patent 8,260,914 Roehrs A, da Costa CA, da Rosa Righi R (2017) Omniphr: a distributed architecture model to integrate personal health records. J Biomed Inform 71:70–81 Roman R, Lopez J, Mambo M (2018) Mobile edge computing, fog et al.: a survey and analysis of security threats and challenges. Future Gener Comput Syst 78:680–698

1830

P. Papadopoulos et al.

Selvi J, Rodriguez RJ, Soria-Olivas E (2019) Detection of algorithmically generated malicious domain names using masked n-grams. Expert Syst Appl 124:156–163 Shulman H (2014) Pretty bad privacy: pitfalls of DNS encryption. In: WPES, pp 191–200 Smart W (2018) Lessons learned review of the wannacry ransomware cyber attack. Dept Health Soc Care England UK Lond 1(20175):10–1038 Solove DJ (2011) Nothing to hide: the false tradeoff between privacy and security. Yale University Press, New Haven/London Spring JM, Huth CL (2012) The impact of passive DNS collection on end-user privacy. Securing and Trusting Internet Names Stamatellis C, Papadopoulos P, Pitropakis N, Katsikas S, Buchanan WJ (2020) A privacypreserving healthcare framework using hyperledger fabric. Sensors 20(22):6587 Stewart J (2003) DNS cache poisoning–the next generation Stolfo SJ, Bellovin SM, Hershkop S, Keromytis AD, Sinclair S, Smith SW (2008) Insider attack and cyber security: beyond the hacker, vol 39. Springer Science & Business Media, Boston Thakkar P, Nathan S, Viswanathan B (2018) Performance benchmarking and optimizing hyperledger fabric blockchain platform. In: 2018 IEEE 26th international symposium on modeling, analysis, and simulation of computer and telecommunication systems (MASCOTS). IEEE, pp 264–276 Tian K, Jan ST, Hu H, Yao D, Wang G (2018) Needle in a haystack: tracking down elite phishing domains in the wild. In: Proceedings of the Internet measurement conference 2018. ACM, pp 429–442 Voigt P, Von dem Bussche A (2017) The EU general data protection regulation (GDPR). A practical guide, 1st edn. Springer International Publishing, Cham Wang W, Hoang DT, Xiong Z, Niyato D, Wang P, Hu P, Wen Y (2018) A survey on consensus mechanisms and mining management in blockchain networks. arXiv preprint arXiv:1805.02707, pp 1–33 Weimer F (2005) Passive DNS replication. In: FIRST conference on computer security incident, p 98 Wood G et al (2014) Ethereum: a secure decentralised generalised transaction ledger. Ethereum Project Yellow Pap 151(2014):1–32 Xia Q, Sifah EB, Asamoah KO, Gao J, Du X, Guizani M (2017) Medshare: trust-less medical data sharing among cloud service providers via blockchain. IEEE Access 5:14757–14767 Xu J, Fan J, Ammar MH, Moon SB (2002) Prefix-preserving ip address anonymization: measurement-based security evaluation and a new cryptography-based scheme. In: 10th IEEE international conference on network protocols, 2002. Proceedings. IEEE, pp 280–289 Yang H, Yang B (2017) A blockchain-based approach to the secure sharing of healthcare data. In Proceedings of the Norwegian Information Security Conference. Nisk J, Oslo, Norway, 11–15 Yeh C-L (2018) Pursuing consumer empowerment in the age of big data: a comprehensive regulatory framework for data brokers. Telecommun Policy 42(4):282–292 Yu Z, Xue D, Fan J, Guo C (2020) DNSTSM: DNS cache resources trusted sharing model based on consortium blockchain. IEEE Access 8:13640–13650 Zdrnja B, Brownlee N, Wessels D (2007) Passive Monitoring of DNS Anomalies. Detection of Intrusions and Malware, and Vulnerability Assessment Springer Berlin Heidelberg, Berlin, Heidelberg 129–139 Zhang R, Xue R, Liu L (2019) Security and privacy on blockchain. ACM Comput Surv (CSUR) 52(3):1–34 Zheng Z, Xie S, Dai H-N, Chen X, Wang H (2018) Blockchain challenges and opportunities: a survey. Int J Web Grid Serv 14(4):352–375

Toward a Circular Economy in the Copper Mining Industry

72

An Industry 4.0 Approach Ingrid Jamett, Ernesto D. R. Santibanez Gonzalez, Yecid Jiménez, and Paulina Carrasco

Contents Introduction to Circular Economy and Industry 4.0 in the Mining Industry . . . . . . . . . . . . . Adoption of Industry 4.0 and Circular Economy in the Copper Mining Industry in Chile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CODELCO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Escondida Mine– BHP Billinton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Antofagasta Minerals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Perspectives in the Implementation of Industry 4.0 in the Copper Mining in Chile . . . . . . . Use of Seawater in the Copper Mining Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Water Consumption and Supply Problem in Chile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Projections of Water Consumption in Copper Mining Processes . . . . . . . . . . . . . . . . . . . . A Solution to the Intensive Consumption of Continental Water . . . . . . . . . . . . . . . . . . . . . Current Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1832 1835 1836 1836 1837 1838 1840 1840 1841 1842 1848 1849 1850

I. Jamett · P. Carrasco Departamento de Ingeniería Industrial, Facultad de Ingeniería, Universidad de Antofagasta, Antofagasta, Chile e-mail: [email protected]; [email protected] E. D. R. Santibanez Gonzalez () CES 4.0, Department of Industrial Engineering, Faculty of Engineering, University of Talca, Curicó, Chile Y. Jiménez Departamento de Ingeniería Química y Procesos de Minerales, Facultad de Ingeniería, Universidad de Antofagasta, Antofagasta, Chile e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_59

1831

1832

I. Jamett et al.

Abstract

In this chapter we analyze the application of Industry 4.0 concepts to advance the adoption of the circular economy in the mining industry. The mining industry makes intensive use of some essential and currently scarce resources such as water and non-renewable fossil fuels, and at the same time is one of the main generators of CO2e. The adoption of a circular economy framework makes it possible to assess the impact that changes in the production processes from copper mining towards more sustainable mining may have on the management of scarce resources such as water and on cost management. Desalination projects in the mining industry are described in the context of a circular economy practice to reduce the consumption of continental water. We briefly explore different mining projects that integrate industry 4.0 technology and its potential integration in the implementation of circular economy practices. The results of this study will inform managers, professionals, and decision makers about the importance of using Industry 4.0 in a circular economy framework to increasing sustainability in the mining extractive sector. Keywords

Circular Economy · Industry 4.0 · Mining Industry · Supply Chain · Water and Energy

Introduction to Circular Economy and Industry 4.0 in the Mining Industry To cope the demand of 9 billion inhabitants in 2050, production systems will devour about 140 billion tons per year of minerals, ores, fossil fuels, and biomass – three times the current consumption, and global food production alone needs to increase by 50%. In the still prevalent and traditional supply chain management approach where goods or products are made, used and then disposed of, and raw materials and natural resources have been severely depleted. Large amount of greenhouse gases are every day emitted to the atmosphere percussing climate change and aggravating its effects, great volume of finished goods are lost even before to reach the customers, and important quantities of waste is generated alongside the supply chain which end up in landfills without any further use. In addition to losses generated by institutional agents in the supply chain, consumers continue to devour products and services under the same approach of unlimited availability of resources (Gonzalez et al. 2018; Santibañez-Gonzalez et al. 2019). Some products and parts of them end up in landfill without any treatment or recovery effort. For example, globally, only between 10% and 40% of generated e-waste and just 10% of plastics are recovered and recycled. This non-virtuous model of natural resources exploitation, production of goods and services, and their consumption has brought our planet to a critical situation, putting at risk the availability of natural resources and their derivatives as products and services

72 Toward a Circular Economy in the Copper Mining Industry

1833

necessary to satisfy the needs of future generations. In this context, mining industry is not the exception, it consumes large volume of natural resources such as water and produces huge volumes of waste. Mining operations produce several types of waste, e.g., overburden – material that lies above an area that lends itself to economic exploitation, such as the rock and soil, waste rock – bedrock that has been extracted and transported out of the pit with metal concentrations without economic interest, tailings – slurry that remains after the mineral of interest has been extracted from the crushed ore at the processing plant, heap leach spent ore – tail of the heap leaching process, and lateral but not less important waste derived from the use of a large variety of inputs such as vehicle tires. In addition to the huge volume of waste generated by the mining industry and the embedded carbon and water footprint of several inputs, mining operations make directly use of immense volume of water and energy. To tackle with environmental and sustainable issues, a number of efforts undertook by some mining companies around the world have been documented in the scientific and grey literature. However, environmental indicators claim for renewed efforts and the adoption of strategies to decouple grow population and demand from this intensive natural resource consumption. While the actual volume of mine waste worldwide that has to be disposed of in dumps and tailings storage facilities is difficult to assess, according to Blight (2011), over 35 × 109 t/annum of waste were generated in 1995 alone by the world’s iron, copper, gold, lead, and bauxite (aluminum) mines together. A no less important consequence of this traditional take-make-consume approach, known as linear economy, has been the generation of large quantities of greenhouse gases, which synergy has undoubtedly enhanced the scarcity of natural resources. Circular economy (CE) is the paradigm shift, which should allow to reverse this situation transform the linear approach into a virtuous circle, where the use of natural resources and energy consumption are optimized, waste generation is reduced and recovering, and restoring activities are maximized (Gonzalez et al. 2018; Santibañez-Gonzalez et al. 2019). The widespread adoption of CE should help to walk toward a more sustainable planet, seeking an efficient balance between the exploitation of natural resources and the economic objectives of companies, respecting society and the environment. According to The Ellen MacArthur Foundation by 2030, a shift toward a CE could reduce net resource spending in the European Union (EU) by A C600 billion annually, improve resource productivity by up to 3% annually, and generate an annual net profit of A C1.8 trillion (Ellen MacArthur Foundation 2015). For example, taken Sweden as a case in terms of environmental benefits and compared to today, Wijkman and Skanberg (2015) estimate about 25% increase in energy efficiency and 25% increase in overall material efficiency “by organizing manufacturing according to CE principles, minimizing waste, and maximizing the reuse and recycling of materials.” Ecofys and Circular Economy (2016) report that the implementation of CE practices could reduce up to 7.5 billion tonnes of carbon dioxide equivalent globally through materials substitution in construction sectors, nutrient recovery in agriculture, chemical leasing, and shared ownership models in transport systems

1834

I. Jamett et al.

worldwide. Despite these promising economic benefits and a substantial amount of recent literature deals with this subject, only a small amount – 6% – of all materials processed around the world are recycled (Haas et al. 2015). The transition toward a CE model in the mining sector is possible by implementing innovative business models that takes advantages of secondary material marketplaces and collaboration alongside the supply chain to adopt novel material stewardship practices. However, the implementation of CE practices in mining operations has been very scarce (Kinnunen and Kaksonen 2019; Lèbre et al. 2017). In the past few years, research on the application of CE in mining has been dominated by resource (valuable metal) recovery from the mine waste stream, particularly from tailings. While some authors recognize that mining operations consume an enormous volume of water and energy, literature on how CE practices can contribute to improve water and energy management is limited. In addition to those limited efforts so far, the mining industry need to consider the adoption of disruptive technology as a means to make a qualitative leap toward new forms of CE implementation. In this context, Industry 4.0 (I4.0) paradigm can play a key role in integrating technology-based CE solutions making use of blockchain for traceability, Internet of Things (IoT), Big Data Analytics (BDA), Artificial Intelligence, among others. According to Lu (2017) I4.0 aims to provide real time information on production, machines, and flow of components, integrating this information in order to help managers to make decisions, monitor performance, and track parts and products. On spite of the applications of I4.0 in CE is in its infancy, significant potential benefits can be envisioned both in mining operations and throughout the mining supply chain. I4.0 disruptive technology has the power to transform a high volume of data into valuable information to make sound decisions in mining companies. An increasing number of researchers and practitioners have realized that with such a volume of data being generated so quickly in many different contexts, the key challenge to face is how we can help decision makers to improve resource management and make better decisions for business. In this context, big data analytics can enable successful implementation of circular economy strategies by improving the availability of data. But more important than that, it is by proposing decision makers the best alternatives to effectively balance economic objectives with environmental and social impacts of the mining company operations in a broad sense. For example, in a general setting, Gupta et al. (2019) explore the use of BDA as a fundamental pillar for making informed and data-driven decisions in supply chain networks supporting CE. Our propose in this chapter is to advance the knowledge on the applications of Industry 4.0 – based technology in circular economy in the mining industry. We analyze and discuss two real cases on the application of CE based on technology in mining processes and key inputs. First, from the perspective of supply chain, we focus on a single copper mining company and the natural capital as a provider. Second, we select water as a natural resource and one of the main inputs in the copper mining industry. Third, we focus in some process in the copper mining industry that makes intensive use of water and can generate large volume of waste with a large embedded hydric footprint. Figure 1 shows the evolution over time of

72 Toward a Circular Economy in the Copper Mining Industry

Early mining projects in Industry 4.0

1835

Generation of proposals at the national level for water efficiency and industry 4.0

Implementation of projects associated with the water problem Identification of the need to incorporate Industry 4.0 in the mining sector

Fig. 1 Industry 4.0 evolution: Chilean mining water efficiency. (Based on: COCHILCO (2018), Consejo Minero, F. C. and C. A. L. (2020))

the adoption of the concepts of CE and I4.0 to address the water shortage problems in the mining operations of the north region of Chile.

Adoption of Industry 4.0 and Circular Economy in the Copper Mining Industry in Chile Mining 4.0, to refer to the application of Industry 4.0 in the mining industry, has been uneven strategy adopted for large mining companies around the world. While some companies are seriously investing in the adoption of technologies to improve the performance of their core business, other companies still do not consider that technologies such as Industry 4.0 can properly support their business and therefore have stayed away from these advances (CESCO 2020). Example of companies adopting Industry 4.0 technologies to improve the performance of their processes are Anglo American and its FutureSmart Mining project, Rio Tinto with its Mine of the future® project, CODELCO with its Integrated Operations Center Project, Escondida with its Smart Mining Market and its EWS desalination plant, Minera Centinela with its project Integrated Operations Center of Minera Centinela and its desalination plant. Efficient water management is one of the great challenges facing society, where technology is a fundamental tool for developing good water management. Desalination becomes a necessary process so that mining companies can continue with their production processes, applying artificial intelligence technology, data analysis, sensors, among other.

1836

I. Jamett et al.

Table 1 Technologies associated with the fourth industrial revolution Group 1 2

3

4

Technologies Digital technologies, blockchain, or internet of things Technologies with physical connection, artificial intelligence with robotics, advanced materials, or IT manufacturing (3D printing) Technologies with Human connection, biotechnologies, neurotechnologies, and virtual and augmented reality. Sustainable technology associated with the capture, storage, and transmission of energy, geoengineering, and space technologies.

Empresa CODELCO, Escondida Mine CODELCO, Escondida Mine, Antofagasta Minerals

Antofagasta Minerals

Escondida Mine, Antofagasta Minerals

Based on Schwad (2016)

According to Schwab (2016), there are 12 types of technologies associated with the industrial revolution 4.0, which can be classified into four groups as indicated in Table 1. Taking this classification as a reference, we briefly discuss the adoption of I4.0 in different copper mining operations in Chile.

CODELCO The Integrated Operations Center (CIO) and Management project of CODELCO in Chile, consists of the automation of the control room of all the operations of the company in the country (CODELCO 2020). It began in 2010 with the Andina mining operation, where the first CIO room. Three years later a new room was inaugurated in the Ministro Hales mining operation. In 2019, the automation of the underground mine began where the transport and trucks uploading, the crushing system, the conveyor belt system to reach the plant, the concentration processes, the smelter, and refinery processes can be controlled remotely from a single place. Within the Schwab classification (Schwab 2016), CODELCO would be in group 1 and 2, since, with the Integrated Operations and Management Centers, it is adding technologies such as the Internet of Things, technologies with physical connections, digital technologies, and the application of artificial intelligence.

The Escondida Mine– BHP Billinton The Escondida is the largest copper mine (by production) in the world and is located in the Atacama Desert in northern Chile, one of the driest deserts on Earth (BHP 2020). She has launched the “Smart Mining Market” project valued at US $ 6.8 billion in 2019 and it is estimated that it will reach US $ 20.31 billion by 2025.

72 Toward a Circular Economy in the Copper Mining Industry

1837

This project will bring automation, artificial intelligence, Internet of Things, drones, smart businesses, 3D modeling techniques, among other technologies, allowing the company to gain in efficiency, safety, and sustainability. On the other hand, the mining company also has a 2018–2030 Water strategy, which will allow it, through technology adoption, to cease groundwater extraction for operational mining processes. Escondida began planning a transition from the use of groundwater resources to desalinated seawater since the mid-2000s. In 2006, Escondida started up the first large-scale desalination plant for industrial use in Chile. In early January 2020, a second desalination facility started production. The project, named Escondida Water Supply Expansion (EWSE), started in 2017, and is the largest desalination plants in the world. The facility has a maximum design capacity of 3800 l per second, over seven times the capacity of the desalination plant in Jeddah’s international airport, on the coast of the Red Sea, Saudi Arabia. With the operation of these two desalinations facilities and the adoption of I4.0, in 2020 (10 years ahead of target) Escondida ceased groundwater extraction from the aquifers Salar de Punta Negra and Monturaqui.

Antofagasta Minerals Antofagasta Minerals is adopting Industry 4.0 technologies in several ways under the umbrella project Integrated Operations Center of Minera Centinela (Antofagasta Minerals 2020). The operation of a fleet of autonomous trucks in the new pit of this mine, the developing of a digital literacy program for its workers, the use of seawater in its processes, among other initiatives that it has been developing since 2015. Antofagasta Minerals is adopting Industry 4.0 technologies in various ways, one of which is the general project Integrated Operations Center of Minera Centinela. The operation of a fleet of autonomous trucks in the new pit of this mine, the development of a digital literacy program for its workers, the use of seawater in its processes, augmented reality, among other initiatives that it has been developing since 2015. In the industry 4.0 application, Antofagasta Minerals was the first company to incorporate seawater desalination technology in mining operations worldwide, allowing efficient use of water resources, during 2016. Currently, 48% of the water used in its operations it came from the sea. In desalination, sensor technology, big data, artificial intelligence, among others, are applied. Augmented reality (AR) is another technology adopted in some Antofagasta Minerals operations. For instance, Minera Centinela, which is owned by Antofagasta Minerals, in 2019 it produced 276,000 t of copper (cathodes and concentrates) from the treatment of approximately 30,000,000 t of ore, accounting for about 40% of Antofagasta PLC’c group copper production. This operation has a 3D Design of the Semi-autogenous grinding (SAG) Mill Lining. SAG mill is one of the most critical equipment in a concentrator plant which needs to be shut down every year commonly due to a liner maintenance among other devices. With the use of AR, maintenance times are reduced.

1838

I. Jamett et al.

Perspectives in the Implementation of Industry 4.0 in the Copper Mining in Chile Today there are numerous projects under development by several mining companies toward achieve mining 4.0 in Chile, in order to improve production and mitigate environmental impacts. Published literature on the applications of 4.0 in the mining industry is scarce. Readers interested in this field are recommended to see,for example, Bertayeva et al. (2019) and Deloitte Insights (2020). Although the investments are high, the benefit that the projects would bring to the companies is greater. The Roadmap “Digitization for mining 4.0” seeks to identify those areas in the mining industry where the adoption of new technologies can add value to the companies in the next 15 years (Consejo Minero 2020). It is an initiative carried out by the Mining Council, Fundación Chile and Corporación Alta Ley, with the support of CORFO/Chile and the technical advice of the Interop program. This roadmap project is part of the strategy of the Chilean Ministry of Mining and aims to provide guidance at the sector level, to take advantage of the opportunities offered by digital technologies to the mining sector. The structure of the Roadmap is composed of two types of cores (a) the traction cores, which are those that, at the heart of the mining process, constitute the most crucial technological challenges for the evolution of the mining industry and (b) the enabling cores, which group Those dimensions that, without being exclusive to the mining process, condition the industry’s ability to carry out its development plan, Fig. 2 shows the structure of the Roadmap. The driving core NT2, named Green Mining, focuses on the adoption of 4.0 technologies to make mining more sustainable, minimizing its impacts on the environment and communities, reducing emissions, managing waste, increasing the use of renewable energies and the efficient use of energy resources and water, and promoting circular economy (Consejo Minero, Fundación Chile and Corporación Alta Ley 2020). Figure 3 presents the challenges and solutions presented in the traction core NT2.

TRACTION CORES

- NT1: Integrated and Intelligent Mining - NT2: Green Mining - NT3: Safe Mining

ENABLING CORES

-

NH4: Digitization NH5: Cybersecurity NH6: Development human capital NH7: Social and political license to innovate

Fig. 2 Roadmap structure: digitization for mining 4.0. (Based on: Consejo Minero, Fundación Chile and Corporación Alta Ley 2020))

72 Toward a Circular Economy in the Copper Mining Industry

2: SHORT TERM (2020TERM (2025-2030)

CHALLENGE 4: MEDIUM TERM (2025-2030)

CHALLENGE 3: MEDIUM

CHALLENGE

CHALLENGE 1: SHORT TERM (2020-2025)

TRACTION CORE 2: GREEN MINNING Transparent control and monitoring of resources and relevant parameters as we operate (Water, air, pollution, contamination and others) and delivery to the community

Include energy and water efficiency in projects

1839

SOLUTION

− Capture and monitoring technologies of

operational data in real time − Data storage technology (Cloud) − Implementation of management systems -

energy, water, emission - responsible budget – objectives − Open and public platform for manage and share information

− − − − −

Data Science Scenario simulation Include in BIM Planning Predictive modeling Decrease in water consumption

− Real-time consumption sensors − Carbon footprint measurement − Link sustainable / environmental KPIs

Improve traceability in energy consumption throughout the process

with productive − Definition of traceability standard − Blockchain implementation for asset

identification − Platform to identify source / origin of

energy − Generate a platform of clean energy

suppliers

− Resource models. Data capture, storage,

Include energy and water efficiency in projects

analysis and optimization − Management system implementation − Interoperability (autonomous cisterns,

dispatch sensors) − Use of artificial intelligence

Fig. 3 Challenges and solutions driving core 2: Green mining. (Consejo Minero, Fundación Chile and Corporación Alta Ley 2020)

One of the solutions presented in the short-term challenge 2 is the reduction of water consumption in the different mining processes. The main objective is the implementation of a closed circuit with zero water losses. The investigations have

1840

I. Jamett et al.

focused on the study of pool water evaporation, low cost dry tailings, comminution and dry mineral separation. This approach is new for what the mining industry is doing today. It consists in the use of seawater in the different mining processes due to the scarcity of water in the country. In this direction, we can mention the initiatives of BHP Billiton (Escondida), CODELCO and Sierra Gorda, and Antofagasta Minerals. In this last case, in addition to use seawater in their operations, they also incorporated options for water reclamation in tailings disposal, managing to recover between 75% and 80% of the water used. More details on the efficient use of water and the incorporation of seawater in mining companies are presented in the following sections.

Use of Seawater in the Copper Mining Industry While the public available information regarding the several desalination projects in the mining industry is very limited, in this section we describe the importance of the adoption of this technology toward a substantial reduction in the use of groundwater in the mining site. The mining industry contributes significantly to the depletion of some essential natural resources on the planet, such as groundwater. One of the objectives of the implementation of CE practices in the industry is that they can help save essential natural resources. Then, as it is described in this section, the implementation of desalination facilities in the mining industry is helping to save groundwater, and in some cases to cease the extraction of groundwater from aquifers in the most dried dessert of the world.

Water Consumption and Supply Problem in Chile Currently, there is a worldwide problem that is undoubtedly present in Chile related to the climate change impact and the footprint left by industrial activities on the planet. Water scarcity is one of these important consequences, especially in the mining industry, which has a constant need for water supply for the development of its processes. It is estimated that, by 2029, water consumption will increase by 56% on year 2018, moving from 16.25 m3 /sg in 2018 to 25.35 m3 /s in 2029 (Chilean Copper Commission 2018). In the mining industry is necessary to develop plans that promote the adoption of technologies, and practices in line with sustainability principles (Bebbington and Bury 2009; Humphreys et al. 2007; Kates and Parris 2003). The that balance of environmental and social impacts with economic and productive activities should contribute to reducing the environmental footprint and the social impact generated by the mining industry in the planet and, in particular, in Chile as one of the main copper producers in the world. Chile has had a decrease in water resources for several years. The National Water Balance of the General Water Directorate (DGA), indicate that the availability of water has decreased by up to 37% in some sectors of the country and national rainfall

72 Toward a Circular Economy in the Copper Mining Industry

RECIRCULATED WATER

CONTINENTAL WATER

1841

SEAWATER

Fig. 4 Data extracted from the report on water consumption in mining (National Mining Society Chile 2018)

have decreased by 30% since the last measurement (DGA 2017). The 37% decrease in water supply corresponds to the central area of the country, between Aconcagua and Maule, so the increase in the consumption of this resource in mining presents a challenge in its operations and future projects. Of the total exploitation of copper in Chile, 54% is concentrated in the Antofagasta region and its consumption of continental water represents about 40% (Chilean Copper Commission 2018). This region is home to several multinational mining companies such as: BHP Billiton, Antofagasta Minerals, CODELCO, Goldcorp and Teck. The Antofagasta Region, with more than 30,000 Km2 , is an area anchored in the driest desert in the world, with annual rainfall of 1.7 mm. Therefore, the consumption of water in mining is currently centralized in the Antofagasta region, identifying three types of water to consume, see Fig. 4 (Chilean Copper Commission 2018; Escenarios Hídricos 2030 2020). 1. Recirculated Water: They correspond to all those flows that are reinjected into the system, these can be previously treated or not. Its consumption in copper mining corresponds to 57.55 m3 /s, 74.8% of the total water consumption. Year 2018. 2. Water of continental origin: considers all permanent bodies of water found in the interior of the continent, away from coastal areas. Some continental waters are rivers, lakes, flood plains, reserves, and wetlands, among others, their consumption in copper mining corresponds to 11.75 m3 /s, 19.3% of the total water consumption. Year 2018. 3. Water of oceanic origin: They come from the sea and have a high brackish content. Its consumption in copper mining corresponds to 4.55 m3 /s, 5.9% of the total water consumption Year 2018.

Projections of Water Consumption in Copper Mining Processes The Chilean Copper Commission (COCHILCO) has made a projection of the consumption of continental and sea water in mining, applying the projection in

1842

I. Jamett et al. Continental Water

Sea Water

25

m3/seg

20

15

3,28

4,74

6,18

6,81

8,14

8,75

9,43

9,68

9,96

10,33

10,69

10,82

14,21

14,09

13,87

13,82

14,2

14,53

2024

2025

2026

2027

2028

10

5

12,98

13,76

13,81

13,79

13,89

13,82

2018

2019

2020

2021

2022

2023

0 2029

Año

Fig. 5 Water projection in copper mining 2018–2029 (Chilean Copper Commission 2018)

the production of copper, according to unit coefficients of water consumption to obtain the future demand of a certain period. This projection includes consumption by origin based on the different categories of projects, including existing desalination and impulsion projects. The study methodology was based on the unit consumption of each task in each process; the maximum production profiles determined through the registry of mining investments, and the probability of materialization of such portfolio by virtue of the historical background. According to values obtained through a Monte Carlo simulation, (Chilean Copper Commission 2018) the estimated total water consumption of continental origin expected to 2029 reaches 14.53 m3 /s, which represents an increase of 12% regarding the expected consumption for 2018. In the case of sea water, it is expected to reach 43% of the total water required in the copper mining industry, reaching 10.82 m3 /s by 2029, which represents an increase in 230% compared to the expected value for 2018, see Figs. 5 and 6.

A Solution to the Intensive Consumption of Continental Water The projections of water consumption, both continental and sea, for the Antofagasta region are presented in Table 2. As can be observed in Table 2, the consumption of continental water shows a decrease, very contrary to the consumption of seawater, which shows a considerable increase, of approximately 178% more in 2029, compared to 2018. Clearly the decrease in water continental is related to the increase in the consumption of seawater, a resource that is projected to be increasingly used in current operations and in future projects. The copper mining processes where water is consumed are presented in Fig. 7.

Fig. 6 Water consumption, according to stages of the mining process. LX, Leaching; SX, Solvent extraction; EW, Electrowinning (Chilean Copper Commission 2018)

72 Toward a Circular Economy in the Copper Mining Industry 1843

1844

I. Jamett et al.

Table 2 Projection of water consumption in copper mining, Antofagasta region, 2018–2029 period Continental L•s−1 Seawater L•s−1

2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 5.4 5.6 5.4 5.4 5.3 5.1 5.2 5.1 4.9 4.8 4.5 4.4 2.8

4.2

5.1

5.5

6.4

6.5

6.9

7.1

7.3

7.6

7.7

7.8

Based on COCHILCO (2018) Fig. 7 Copper mining processes where water is consumed

The mine section includes the underground or open-pit mine, as well as the transport of the material to the primary crushing. The greatest use of water in this area is associated with the extraction and pumping from underground operations, and the dust abatement on roads. In the case of the concentration area, this includes comminution, classification, flotation, and thickening of minerals. Considering the total volumes, this area represents the highest water consumption, among all the areas mentioned. The hydrometallurgical part covers the heap leaching, solvent extraction and electrowinning processes. In this process, the leaching stage is the one that consumes the most water due to its evaporation from the leaching heaps. Fresh water consumption in hydrometallurgy is of the order of 0.13 m3 /t of mineral and it is considered that leaching uses 1/3 of the fresh water used by flotation. Next, there is the smelter and refinery, where the dry concentrate, product of the concentration stage, is subjected to a pyrometallurgical process to obtain copper anodes. This product can be further refined or commercialized, in the case of refining, this is carried out in electrolytic cells with a sulfuric acid solution, where an electric current is applied, dissolving the copper from the anode and depositing on the cathode. The other area includes services such as the use of water for drinking, cooking, washing, irrigation and bathrooms. This consumption with respect to the aforementioned is much lower. In the Antofagasta region, the operations that currently use seawater are: Centinela, Escondida, Michilla (in the process of reopening), Antucoya, Las Cenizas Taltal, Mantos de la Luna, Sierra Gorda, and J.A. Moreno.

72 Toward a Circular Economy in the Copper Mining Industry

1845

Likewise, there are projects that plan the implementation of desalination facilities, among others, we can mention: CODELCO’s northern district plant to supply the divisions of Chuquicamata Radomiro Tomic and Ministro Hales progressively, Esperanza upgrade and its subsequent network extension to supply the Encuentro project, Spence Growth Project, and El Abra Mill Project. Table 3 shows the companies in the Antofagasta Region that use seawater in their operations, desalinated or not, as well as the projects in the area. (*) Production started during 2020. The use of seawater in the region is long-standing, and in this sense the following operations can be mentioned in chronological order: • Mantos Blancos 1961, heap leaching of the ore atacamite and chrysocolla with belt agglomeration with sodium chloride. 90% extraction of copper is achieved, with an acid consumption of 3 kg H2 SO4 / kg copper. • Tocopilla 1987 plant, leaching with seawater and later solvent extraction. • Lince 1991 Plant. Specifies the production of grade A copper cathodes, overcoming the problems of transferring the chloride ion to the electrowinning stage. • Minera Michilla: Leaches its ore (secondary oxides and sulfides) using seawater and calcium chloride, CaCl2 , in a sulfuric acid medium. • Las Luces, Mantos de la Luna, heap leaching. On the other hand, in recent years the Pampa Norte Division of BHP Billiton has studied at an industrial level the addition of sodium chloride in its mixed minerals and secondary sulfides. Similarly, Sierra Gorda and its Catabela plant mine complex is supplied with sea water supplied from the filtering plant located in the city of mussels through a 142 km long, 36 diameter aqueduct and two pumping stations located in its trajectory and stored in a process water pool, replacing the reservoir that for the same purpose had originally been proposed. The water that is used in its entirety comes from the sea, and then from the cooling system of the thermoelectric plant, located in the industrial district of Mejillones. This is captured in an accumulation pond in the filtering plant and sent to the process water pool of the Catabela mine-plant at an approximate rate of 1305 L/s. In the mine plant sector, water is used in different amounts and types: seawater without desalination for the processing of sulfides, feeding the osmosis plant and the fire network (4737 m3 /h); fresh water obtained from the desalination plant for the processing of oxides, washing trucks and machinery, and sanitary systems (227 m3 /h); and treated drinking water for human consumption (39 m3 /h) (Cisternas and Moreno 2014). Mining only represents 9% of water consumption at the national level (DGA 2013); however, in the case of the Antofagasta region, it is 68% and this must be resolved urgently with the participation of all stakeholders, water consumers, and state bodies with decision-making power. The portfolio of efficiency and strategic use of water, (see Fig. 8) presents four types of measures associated with preserving water resources and adopt actions aimed at reducing the water gap in Chile.

1846

I. Jamett et al.

Table 3 Desalination plants and projects in the Antofagasta Region Production Stage start

–

–

Closed (Expected reoperation 2019) Operating

–

Operating

–

Operating

–

Operating

–

Operating

– –

Operating Operating

–

Operating

2020

Feasibility

2020

Feasibility

2020

Feasibility

2021

No environmental impact assessment Feasibility stop Feasibility

2023 2024 2025

No environmental impact assessment

Owner

Name

Capacity

Capacity

Haldeman

Michilla

desalination l/s 75

seawater l/s 25

ENAMI

Planta J.A. Moreno (Taltal) Las Cenizas Taltal

–

15

9

12

Mantos de la luna

5

20

Escondida – Planta Coloso Distrito Centinela (Esperanza + El Tesoro) Antucoya Sierra Gorda

525

–

50

1500

20 –

280 1315

Escondida EWS

3800(*)

–

–

850

1000

–

630

–

740

–

Santo Domingo

30

400

Mining development Centinela – Stage 2 El Abra Mill Project

–

1650

500

–

Las Cenizas Mantos de la luna BHP Billiton AMSA

AMSA KGHM Int. BHP Billiton AMSA

Mining development Centinela – Stage 1 BHP Spence Growth Billiton Option CODELCO Desalination plant adaptation RT Sulfides – Stage 1 Goldcorp Nueva Unión y Teck

Capstone Mining AMSA Freeport McMoran

Source: (Montes 2019)

The third line being “efficiency and strategic use of water resources” and the fourth line “Migration and incorporation of new water sources,” those destined to

72 Toward a Circular Economy in the Copper Mining Industry Axes of the Water Description

1847 Actions

transition Management

Delivers the enabling conditions to achieve security

institutionality of the Water

water

•

Integrated management of water resources

•

Integrated national water information system

•

Incentives to savings and efficiency in sectors productive

Conservation and

It considers as the basis of management development

•

protection of our

of water policies, programs and plans that guarantee

conservation of rivers,

Recovery and

ecosystems Water

the protection, recovery and conservation of hydric

banks, wetlands, cushion

ecosystems

bogs, estuaries and peat bogs

Efficiency and use

Raises efficient use and responsible for water, the

strategic of the

reduction of withdrawals by users intensive in water

resource Water

consumption, taking care of their quality

•

Systems and techniques optimized irrigation

•

Replacement minor crops request hydric

•

Reservoirs for accumulation of waters

Migration and

Refers to the planning, design and enabling of

incorporation of

systems multipurpose that allow the incorporation of

•

Desalination through osmosis inverse with and

new Water sources. water for address the Water Gap in the territories,

without energies

allowing a decoupling between the development of

renewable

productive sectors and the use of fresh water in the

•

basin.

•

Treatments wastewater Water reuse residual in Emissaries Submarines

Fig. 8 Axes of the hydric transition, based on Scenarios hydric 2030

the use of technology 4.0 and the concepts of circular economy. It can be indicated that the examples of application of these strategies are not yet consolidated, remaining as practices associated with nations and companies in particular. Some of these practices have a high environmental commitment, such as the installation of desalination plants, which is mostly promoted by mining companies. Optimized systems and techniques are mostly used in the agricultural industry, but it is highlighted that they do not present negative environmental impacts. In order to promote the efficient use of water resources in the mining industry, it is necessary to implement some of the technological strategies outlined by Schwab (2016) and taking as a reference the Green Mining proposals contained in the Roadmap digitization for mining 4.0. Although some of these technologies are

1848

I. Jamett et al.

currently being applied in the industry, it is recommended the adoption of industry 4.0 to in conjunction with circular economy to continue to reduce the consumption of continental water in several mining processes. The need for an effective reuse and recycling technology becomes in conjunction with an adequate management of supply sources to reduce consumption are essential circular practices to move toward mining 4.0. The use of seawater, salty or desalinated, already represents 23% of all the water used by mining operations. It should be noted that according to the latest Cochilco report, by 2029 it is expected that the consumption of seawater will represent 43% of that required by the industry. Considering that large-scale mining operates in arid areas, and that there is a country challenge due to the scarcity of water, “seawater is and will continue to be an important source of supply for the sector.” Beyond the benefits that desalination provides as a water source, this type of supply source implies a high cost that many operations cannot assume, especially those far from the coast and at a higher altitude above sea level. This higher cost is mainly due to the huge amount of electricity that is required to carry out desalination and, mainly to drive water to mining sites located, for example, more than 3000 m above sea level, with a desalinated seawater supply costs of up to 8% on mining production values. Due to the costs involved in using seawater, the first concern of the sector is the efficient use of this resource. This is how recirculation has remained above 70%, which means that only 30% or less of the water used in mining is new water. On the other hand, the consumption of water per ton of processed mineral in the concentrator plants has decreased persistently, from 0.57 m3 /t in 2013 to 0.36 m3 /t in 2018, which means a decrease of 36% in the last 5 years.

Current Projects One strategy to address the water shortage is also to increase the rate of recirculation of water in several mining processes so that the resource can be reused within the operation, generating significant savings as a result of optimization. This is achieved by introducing innovations and technological advances that allow less evaporation and use of water in tailings. One such example is Minera Centinela, owned by Antofagasta Minerals, which has incorporated tailings thickened with seawater on an industrial scale, allowing greater efficiency and a lower risk of infiltration in the soil and subsoil layers. Attempts are being made to recover the maximum amount of used water, for example, in obtaining copper concentrate – a process that requires the most water – between 75% and 80% of the water used is recovered from the tailing dams. In this regard, if previously only a certain amount of this element was recovered from the tailing dams, today the aim is to recover the highest percentage of water resources and reuse as much water as possible. That is the great challenge that mining faces today. An objective that is constantly seen in mining companies investing in water management and treatment projects, through projects called “zero effluents.”

72 Toward a Circular Economy in the Copper Mining Industry

1849

Conclusions From the analysis carried out is concluded that the mining industry has not taken a key role in the fourth industrial revolution. Due to large investment involved, the diversity of geographical location of mining projects, and the age of some mining operations, among others, are influencing on the adoption mining 4.0, and it has still a long way to a widespread adoption in the industry. But there are successful projects, where technology has become the protagonist of the mining production process, achieving lower costs, clean productions, and reducing accidents. In Chile, a Roadmap was generated to implement Mining 4.0, which has short-, medium-, and long-term activities, within 15 years planning horizon. One of its shortterm activities (2020–2025) is to include energy and water efficiency in mining projects, through the use of data science, scenario simulation, BIM planning, and predictive modeling. The integration of these Industry 4.0 related technologies with circular economy practices in the mining industry is still in its infancy. To continue advancing in the reduction of inland water consumption and the efficient management of seawater in mining operations, strategies that consider at least the following three areas must be implemented: (1) large-scale use of sensors and IoT applications along the water supply network. This will allow real-time monitoring of water consumption in relevant stages of mining production; (2) the use of cloud computing and big data to cope with the increasing amount of data captured from sensors deployed in the water supply network. The use of IoT allows a greater number of devices to be connected, generating a greater volume of data to be captured. The captured data must be stored, then processed and converted into information, to finally be translated into useful knowledge to make the right decisions; (3) The adoption of decision-making tools based on Artificial Intelligence will allow the use of collected and stored data to transform it into real value-added information so that decision makers can make better decisions, in order to reduce the consumption of continental water, recirculate water, and recycle it at each stage of mining operations. As a result of the water scarcity in the Antofagasta Region and the high demand for the use of water by mining companies, the incorporation of seawater in operations has been a solution not only to water scarcity but has also been beneficial for processes such as the leaching of certain copper sulfides, as well as in the concentration process. This is reflected in the increased consumption of seawater in the region, as well as in the increase in projects for desalination plants for seawater or without desalination. It is concluded that the reduction in water consumption in operations is associated in most companies with the reduction of water losses due to evaporation as well as the increase in recirculation. In this line, although there are advances such as those of Antofagasta Minerals, there are still challenges associated with materials, equipment, control, monitoring, among other, that need to be addressed to continue reducing water consumption in mining operations. Acknowledgments Research of the author Ernesto DR Santibanez Gonzalez was partially funded by FONDECYT No. 1190559 (regular project, ANID-Chile).

1850

I. Jamett et al.

References Antofagasta Minerals (2020). https://web.mineracentinela.cl/ Bebbington AJ, Bury JT (2009) Minería, desarrollo y ciencia de la sostenibilidad. Retrieved from https://hummedia.manchester.ac.uk/schools/seed/andes/es/publications/papers/ Bebbin gtonandBury2010_espanol.pdf Bertayeva K, Panaedova G, Natocheeva N, Kulagovskaya T, Belyanchikova T (2019) Industry 4.0 in the mining industry: global trends and innovative development. E3S Web Conf 135:1–7. https://doi.org/10.1051/e3sconf/201913504026 BHP (2020). https://www.bhp.com/ Blight G (2011) Mine waste: a brief overview of origins, quantities, and methods of storage. In: Waste. pp 77–88. https://doi.org/10.1016/B978-0-12-381475-3.10005-1 CESCO (2020) Hacia una minería 4.0: Recomendaciones para impulsar una industria nacional inteligente. CESCO, Santiago Chilean Copper Commission (2018) Water projection in copper mining 2018–2019. Chilean Copper Commission, Santiago Cisternas L, Moreno L (2014) El agua de mar en la minería: Fundamentos y aplicaciones (R. Editores., Ed.). R. Editores, Santiago CODELCO (2020). https://www.codelco.com/ Consejo Minero, F.C, Corporación Alta Ley (2020) Roadmap: digitalización para una mineria 4.0. Retrieved from https://digitalizacionmineria.fch.cl/ Deloitte Insights (2020) Tracking the trends 2020. Leading from the front. Retrieved from http:// arxiv.org/abs/2007.10148 Dirección de Gestión de Aguas (DGA). (2013) Balance de gestión integral. Ministerio de Obras Públicas, Gobierno de Chile. Santiago, Chile. Ecofys and Circular Economy (2016) Implementing circular economy globally makes Paris targets achievable. Ecofys and Circular Economy. Netherlands Ellen MacArthur Foundation (2015) Delivering the circular economy: a toolkit for policymakers. Ellen MacArthur Foundation, Cowes Escenarios Hídricos 2030 (2020). https://escenarioshidricos.cl/publicaciones/ Gonzalez EDS, Zhu J, Zanoni S, Maculan N (2018) Trends in operational research approaches for sustainability. Eur J Oper Res 269:1–4. https://doi.org/10.1016/j.ejor.2017.12.014 Gupta S, Chen H, Hazen BT, Kaur S, Santibañez Gonzalez EDR (2019) Circular economy and big data analytics: a stakeholder perspective. Technol Forecast Soc Chang 144(2017):466–474. https://doi.org/10.1016/j.techfore.2018.06.030 Haas W, Krausmann F, Wiedenhofer D, Heinz M (2015) How circular is the global economy? An assessment of material flows, waste production, and recycling in the European Union and the world in 2005. J Ind Ecol 19(5):765–777. https://doi.org/10.1111/jiec.12244 Humphreys M, Sachs JD, Stiglitz JE (2007) Escaping the resource curse. Columbia University Press, New York Kates RW, Parris TM (2003) Long-term trends and a sustainability transition. Proc Natl Acad Sci 100(14):8062–8067 Kinnunen H, Kaksonen AH (2019) Towards circular economy in mining: opportunities and bottlenecks for tailings valorization. J Clean Prod 228:153–160. https://doi.org/10.1016/ j.jclepro.2019.04.171 Lèbre É, Corder G, Golev A (2017) The role of the mining industry in a circular economy: a framework for resource management at the mine site level. J Ind Ecol 21(3):662–672. https:// doi.org/10.1111/jiec.12596 Lu Y (2017) Journal of Industrial Information Integration Industry 4.0: A survey on technologies, applications and open research issues. J Ind Inf Integr 6:1–10. https://doi.org/10.1016/ j.iii.2017.04.005 Montes C (2019) Proyecciones de agua y energía eléctrica en la minería del cobre al 2027. Comisión Chilena del Cobre (COCHILCO), Santiago, Chile

72 Toward a Circular Economy in the Copper Mining Industry

1851

National Mining Society Chile (2018) Water consumption report in mining 2018. National Mining Society Chile, Santiago Santibañez-Gonzalez EDR, Koh L, Leung J (2019) Towards a circular economy production system: trends and challenges for operations management. Int J Prod Res 2(1):7209–7218. https:// doi.org/10.1080/1478643YYxxxxxxxx Schwab K (2016) The fourth industrial revolution (World Economic Forum, Ed.). World Economic Forum, Ginebre Wijkman A, Skanberg K (2015) The circular economy and benefits for society Swedish case study shows jobs and climate as clear winners. The Club of Rome, Rome

Sustainability Index of Metalworking Fluids in the Manufacturing Industry for Sustainable Manufacturing

73

Sustainability Index Muralidhar Vardhanapu and Phaneendra Kiran Chaganti

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sustainability Index (SI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sustainable Manufacturing Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Process Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Product Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Four Principles of Sustainability Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Consolidated Framework for Sustainability Index (SI) Assessment . . . . . . . . . . . . . . . . . . . Evaluation of Deterministic Components of SI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Understanding the Evaluation of Nondeterministic Components Using a Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Overall Sustainability Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Material and Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sustainability of Various Elements Used during Various Stages of the Life Cycle of MWFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Categorization of Deterministic Components and Nondeterministic Components . . . . . . Evaluation of SI for Deterministic Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evaluation of Nondeterministic Components of SI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sustainability Index of the System under Consideration . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion/Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1855 1858 1861 1862 1863 1864 1864 1865 1868 1870 1871 1871 1871 1871 1872 1876 1876 1877 1879 1880

M. Vardhanapu · P. K. Chaganti () Department of Mechanical Engineering, Birla Institute of Technology & Science, Pilani Hyderabad Campus, Hyderabad, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_60

1853

1854

M. Vardhanapu and P. K. Chaganti

Abstract

Having said that metalworking fluids (MWFs) are advantageous in many ways during the machining operations in the manufacturing of a product, yet MWFs are always a debatable topic for the researchers because of their adverse effects on the environment and operator health. Better surface finish, enhanced tool life, avoiding the formation of built-up edges (BUEs), avoiding tool tip welding, and reduced thermal deformation of the work-piece, as well as the tool, are included as the advantages of MWFs, while the disadvantages being contamination of the machinery and workplace, weak storage stability, polluting the work zone atmosphere, causing respiratory problems for the operators and disposal problems. Industry 4.0 is not only known for improvising productivity but also paves the way for solving social problems. One of the major contributions to society through Industry 4.0 is the value addition to a process or product by concentrating on sustainability aspects without deteriorating the environment. As a countermeasure to hard challenges faced by our planet, entire mankind is thriving toward establishing sustainable systems. While the whole world is routing its steps toward achieving sustainability, the manufacturing industry is no bar. Sustainable manufacturing is gaining a lot of interest of the renowned OEM’s (original equipment manufacturers) across the globe. As a part of achieving sustainable manufacturing, there is an urging need to assess the sustainability index of the MWFs with respect to socioeconomic concerns. Assessing the sustainability of any system is quite difficult in practice because of the uncertainty of various factors that are under consideration in a study. In practice, the sustainability index is defined first, in order to define any system. The sustainability index is often defined as a decision-making tool or parameter which helps the firm or a retailer to make decisions regarding a certain product or a system keeping its responsibility intact toward social and environmental areas. On the other hand, the sustainability index consists of two components, namely, deterministic and nondeterministic. The deterministic factors are machining cost, waste management, and power consumption, while the nondeterministic factors are safety, health, and environmental issues. In this work, an assessment of the sustainability index of MWFs with respect to various deterministic and nondeterministic components is carried out. A hypothetical case study has been presented in this chapter with the assumed data and in reference to the previous works in this domain. This chapter of the book provides insights of the assessment of the sustainability index of MWFs with MQL setup in the manufacturing industry. Keywords

Sustainability · Sustainability index · Metalworking fluids · MQL · Fuzzy logic

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . .

1855

Introduction Industry 4.0 represents the fourth revolution that occurred in the manufacturing industry, which is often referred to as digital transformation and value creation to the production process. The productivity has claimed to be increased considerably with the advent of Industry 4.0. Sustainability is an integral part of Industry 4.0, as its origin is from Germany which is one of the most competitive manufacturing industry hubs in the world, and Industry 4.0 is the strategic initiative used by the Germans to sustain their position in the market (Rojko 2017). Thus, Industry 4.0 is termed as an enabler of sustainable development. Sustainable development is defined as the integration of various processes and making the system capable of producing higher quality products and services using sustainable resources while keeping the environment and social circles safer during the life cycle of a process (Machado et al. 2020). Industry 4.0 technologies and sustainable operations are having interlinks, while few studies suggest the underdevelopment of convergence between digital transformation and sustainability (Machado et al. 2020). Industry 4.0 and sustainable production systems are built on the nine pillars of technological advancements, namely, autonomous robots, simulation, horizontal and vertical systems integration, Internet of things (IoT), cybersecurity, cloud, additive manufacturing, augmented reality (AR), and big data analysis. Sustainability has become a significant area of interest across the globe with the production systems concentrating on the control of environmental impacts. Societal, ecological, and economic impacts are the three dimensions of sustainable development as illustrated in Fig. 1. These are also called a triple bottom line (TBL). Sustainability index (SI) is used to measure the sustainability of any system, considering various factors. The sustainability index is often used at the industry level to determine a company’s performance and compare it with other companies worldwide. SI even helps the consumer to understand a product’s environmental impact by providing information regarding the process involved in making the product and how the process affects the environment. In this chapter, the sustainability index of metalworking fluids (MWFs) is assessed as a part of achieving sustainable manufacturing. Metalworking fluids (MWFs) are part of the manufacturing industry for various advantageous reasons but are often a debatable topic for researchers because of their adverse effects on human health and the environment. Metalworking fluids are used at the cutting zone between the tooltip and the workpiece for the reduction of temperatures, for providing lubrication effect, for flushing away metal fines, and for many other advantages. Better surface finish, enhanced tool life, avoiding the formation of built-up edges (BUEs), avoiding tool tip welding, and reduced thermal deformation of the workpiece are included as the advantages of MWFs. The disadvantages are contamination of the machinery and workplace, weak storage stability, polluting the work zone atmosphere, causing respiratory problems for the operators, and disposal problems. There are different ways of applying MWFs at the cutting zone like liquid and gaseous lubrication. In general, the MWFs are

1856

M. Vardhanapu and P. K. Chaganti

Fig. 1 Three dimensions of sustainable development

introduced at the cutting zone using the flood coolant system (liquid lubrication), but of late, a lot of industries are implementing the minimum quantity lubrication (MQL) technique for reduced quantity usage of MWFs, which produces a very similar effect to the Flood coolant system. MQL can be a potential solution to the increased quantity usage of MWFs (in flood coolant lubrication type) that are responsible for environmental degradation as well as fossil fuel depletion. Also, MQL offers more required cooling and lubrication effect at the cutting zone where the high temperature is generated due to friction developed during chip formation, as MQL spray can be targeted at the right area while the regular flood coolant type fails to reach the actual cutting zone (Ali et al. 2011). In the MQL type of lubrication, the MWFs are applied on to the cutting zone with a flow rate of 50–500 ml/hr., which is found to be 3 to 4 orders of magnitude lesser than the quantity that is used in the flood coolant lubrication (Ali et al. 2011). In general, there are three types of MQL setups available in the market. The first type are the low-pressure spray systems with a flow rate of 0.5 to 10 l/hr. The second type of system uses dosing pumps to deliver a definite amount of lubricant on to the machining surface without air. These types of systems have a flow rate of 0.1 to 10 ml/cycle with an adjustment up to 260 cycles/min. The third type is the most widely used pressure systems in which the lubricant is mixed with compressed air at a certain pressure to form a mist of the lubricant. The lubricant is pumped to the nozzle, where it is mixed with compressed air to form the mist. The quantities of compressed air and lubricant can be varied separately. The generated mist is targeted or sprayed on the machining surface to enhance the effect of lubrication and heat dissipation. In this type of system, the flow rate generally varies from 50 to 300 ml/hr., making it much viable and operational among the available systems. Earlier, petroleum-based

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . .

1857

mineral oils are regularly used, as they give well-desired outputs. Later vegetable oils came into existence as supplements to those petroleum-based mineral oils to counter the adverse effects of petroleum-based oils (Sen et al. 2019). Currently, the trend has changed; the nanoparticles based MWFs are prevailing (Hegab et al. 2018). Metalworking fluids (MWFs) are harmful to the environment as well as to human health due to their high oil content, biological oxygen demand, and hazardous surfactants. Several research studies found out that the MWFs show carcinogenic effects on the operators’ health (Lu et al. 2012). Also, these MWFs are prone to various instabilities caused by ingredient depletion, evaporation, contamination, and microbial growth, not only making them harmful to the atmosphere but also not economically viable in a broader view. Because of these adversities, avoiding these metalworking fluids and adopting the dry machining can reduce a lot of costs associated with them; however, the lack of cooling and lubrication effect provided by MWFs may result in rapid tool failure, poor surface finish, and dimensional errors (Skerlos et al. 2008). Dry machining may be encouraged to an extent in case of turning and milling operations, but it is very uneconomical to prefer dry machining in case of drilling, tapping, and grinding. Hence, there is a scope for sustainability assessment of MWFs and to develop sustainable metalworking fluid systems. In the recent times, focus has shifted to sustainability irrespective of discipline and gained the interest of researchers. With a massive shift in researcher’s and industries’ ideology to adopt sustainable practices, Fig. 2 shows the recent research trends in sustainability assessment across the domains. In similarity, Fig. 3 demonstrates the trend of sustainability assessment over the past 5 years in manufacturing in terms of the number of journal paper publications. Figure 4 provides insights into the number of publications related to sustainability assessment in context with machining, nano-MWFs, and MQL.

Fig. 2 The percentages of journal papers in sustainability across all the fields for the past 5 years obtained from Scopus and Google scholar

1858

M. Vardhanapu and P. K. Chaganti

Fig. 3 The number of journal papers with the keyword “sustainability AND manufacturing” between 2015 and 2019. (Data was obtained from Scopus in October 2020.)

This shows the significant importance of sustainability assessment of any system. The sustainability can be assessed by evaluating the sustainability index. In this research work, MWFs along with MQL setup while conducting a specific machining operation is the system under consideration for the sustainability assessment. An example has been presented for better understanding of the calculation of SI.

Sustainability Index (SI) Sustainable development is said to be achieved only when all the three elements, societal sustainability, environmental sustainability, and economic sustainability are very well balanced (Palit and Hussain 2018). The sustainability index is used to assess any system’s sustainability under consideration by taking into account various factors. The sustainability index (SI) is a decision-making tool that helps the firm or a retailer make decisions regarding a particular product or a system keeping its responsibility intact toward social and environmental areas (Sandoval-Solis et al.

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . .

1859

Fig. 4 The number of publications in the area of sustainability in context with machining, nanoMWFs, and MQL

2011). The sustainability index is basically divided into two components, namely, deterministic and nondeterministic components. The deterministic components are those factors that can be quantified, for example, the machining cost, which can take a certain value. The nondeterministic are those that cannot be physically quantified, for example, the health of an operator or safety aspects that are nonquantifiable. With metalworking fluids in the manufacturing system as a subject of consideration for this study, the deterministic components of SI are considered to be the manufacturing/machining cost, waste management, and power consumption that are associated with the usage of MWFs. The nondeterministic components are safety, the health of an operator, and environmental issues that may be caused due to usage of MWFs. These six factors constitute the implementation of sustainable practices during the use of MWFs in the manufacturing industry (Najiha et al. 2016). Figure 5 illustrates these six essential elements of a sustainable manufacturing process. The sustainability index has to be assessed first in order to assess the sustainability of any manufacturing system. The critical challenge in assessing the

1860

M. Vardhanapu and P. K. Chaganti

Fig. 5 Six basic elements of the sustainable manufacturing process (Lu et al. 2012)

sustainability of any machining process is that among the six basic elements of sustainability, only three components can be quantified by using analytical and numerical models because they are relatively deterministic in nature (Jawahir and Jayal 2011). The other three components, because of their nondeterministic nature, require logic-based models such as fuzzy logic for their assessment, which is quite difficult. The system under consideration for this study is the machining operations performed with the metalworking fluids used as lubricants and these MWFs are applied onto the cutting zone using the MQL technique. To measure the sustainability index of this system, the six basic elements of sustainability are to be assessed.

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . .

1861

Sustainable Manufacturing Systems Sustainability in manufacturing can be evaluated at various levels, namely, product, process, and systems (Gholami et al. 2020). The sustainable manufacturing system in the MWFs point of view assists the eco-design of MWFs formulation and the development of sustainable metalworking fluid systems in countering the economic, health, environmental, and performance liabilities associated with MWFs (Skerlos et al. 2008). The target objectives of the sustainable MWF systems are improved sustainability, improved microbial stability, improved longevity, reduced eco-inputs, and improved performance, as illustrated in Fig. 6. The sustainable manufacturing systems emphasize the requirement of physiochemical sensors to track the various parameters such as pH, concentrations of corrosion inhibitors, surfactants, biocides, ions, and particulate in MWFs over a period to maximize the performance. These automated technologies can increase the speed of microbial quantification at reduced costs, thereby helping the cause. Along with the sensor technologies, sustainable manufacturing systems focuses on the development and upgrade of MWF actuator technologies. Actuators are employed for various applications, viz., chemical addition, phase separation, magnetic separation, filtration, and microbial inactivation. These actuators play an important role in controlling contamination, reducing health risks, and environmental loadings. Altogether, the development of MWF control systems making use of the physiochemical sensors, biological sensors, and contamination controlling actuators aids the achievement of a sustainable MWF system. Figure 7 pictographically represents the conceptualized sustainable MWF system with various control

Fig. 6 Target objectives of sustainable MWF Systems (Skerlos et al. 2008)

1862

M. Vardhanapu and P. K. Chaganti

Fig. 7 MWF control system (Skerlos et al. 2008)

systems. These control systems are designed with the aim to maximize the lifetime of the sustainable MWF systems. Thus, these sustainable MWF systems are targeted toward achieving minimized system inputs and outputs such as materials, energy, and toxicity; this maximizes the life span of the system by maintaining the physical, chemical, and biological parameters within the functionality limits.

Process Sustainability Process sustainability can be evaluated for manufacturing using the 6R (Reuse, Reduce, Recycle, Recover, Redesign, and Remanufacture) approach (Kishawy et al. 2018). There were a fair amount of papers published in this domain, where metricsbased sustainability evaluation of the various manufacturing processes were carried out (Lu et al. 2012). Qualitative and quantitative metrics are necessary for evaluating the sustainability of the manufacturing process. These metrics are developed with the goal of reducing the decision-making burdens involved in optimizing the

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . .

1863

Table 1 Potential sustainable manufacturing process metrics (focus is on the process) (Haapala et al. 2013) Metric type Environmental impact

Manufacturing cost/economic cost Operational safety/worker safety Energy consumption

Waste management

Personal safety/worker health

Example  Emission (kg CO2 eq./unit)  Ratio of renewable energy used (%)  Total water consumption (kg/unit)  Labor cost ($/unit)  Energy cost ($/unit)  Maintenance cost ($/unit)  Exposure to corrosive/toxic chemicals (incidents/person)  Injury rate (injuries/unit)  Near misses (misses/unit)  In-line energy use (kWh/unit)  Energy use for maintaining the working environment (kWh/unit)  Energy consumption for material handling (kWh/unit)  Mass of disposed consumables (kg/unit)  Consumable re-use ratio (%)  Ratio of recycled chips and scrap (%)  Chemical contamination of the working environment (mg/m3 )  Mist/dust level (mg/m3 )  Physical load index (dimensionless)

economic and environmental concerns in a manufacturing process life cycle. The potential sustainable manufacturing process metrics are listed in Table 1 (Haapala et al. 2013). The metrics mentioned in Table 1 can be used to assess the sustainability of a process. The total life-cycle approach can also be used to develop metrics for sustainable product development (Gupta et al. 2011). The design of sustainable manufacturing, an approach in developing a sustainable manufacturing process by integrating the information with various factors throughout the life-cycle of a process, was also studied by some authors (Gupta et al. 2011; Gao et al. 2016b).

Product Sustainability Having discussed the assessment of sustainability at system and process levels, it is said that there is a major impact of the product on the TBL (triple bottom line) throughout its life cycle during various stages, namely, pre-manufacturing stage, manufacturing stage, and use and post-use stages (Weinert et al. 2004; Jaafar 2006). The product sustainability index method can be used as a tool for assessing the sustainability of a product (Schmidt 2007). Product Sustainability Index (PSI), as an outcome of a rating system that can be used for a wide range of products, was developed as a new assessment methodology to evaluate product sustainability. The product sustainability index can be divided into three subindices belonging to environmental, ecological, and social aspects, which are again classified into

1864

M. Vardhanapu and P. K. Chaganti

subclusters. Once the subclusters are classified for a particular product, the product sustainability index can be calculated by making use of the normalizing, weighting, and aggregating approaches (Zhang et al. 2012).

The Four Principles of Sustainability Measurement There are four fundamental principles of sustainability measurement (Fiksel et al. 1999). The first principle emphasizes that the evaluation of sustainability should address the dual perspectives of resource consumption and value creation. The general list of resources that are associated with product development, namely, energy, material, water, land, waste, cost, human capital, and investment capital. As per the literature, the general categories of performance associated with value creation, namely, functional performance, information content, customer satisfaction, environmental quality, economic value-added, business competency, human health, and social welfare. The second principle accentuates the consideration of the triple bottom line, the social, economic, and environmental aspects while evaluating sustainability. The third principle highlights the systematic consideration of each stage in the product life cycle as an important event while evaluating the sustainability as the resource consumption and value creation takes place throughout the life cycle of a product, that is, during the supply of raw materials, manufacturing, and use and post-use (disposal). In other words, the different stages, namely, the pre-production, production process, product usage, and post usage, are to be considered while evaluating the sustainability of a product/service. The fourth principle points out the need for combining lagging and leading performance indicators while evaluating sustainability. The lagging indicators are the measure for outcomes of improvements that are made in the processes, while the leading indicators are the measure of efforts made to improve future performance. Having given the principles of sustainability measurement, insights of sustainability performance measurement process were also provided along with different steps to be followed during the various phases of the process, namely, plan phase, review phase, and implementation phase (Fiksel et al. 1999).

Consolidated Framework for Sustainability Index (SI) Assessment Sustainability index assessment of metalworking fluids was done by quite a few researchers (Skerlos et al. 2001). Some of the researches in this area focused on the sustainability assessment of the manufacturing system as a whole but very few researchers have focused on the sustainability assessment of MWFs (Jawahir and Jayal 2011). Most of the papers published in this area were either quantifying deterministic or nondeterministic elements of sustainability of MWFs, but a holistic view is missing in the literate. The present study aims to present the assessment of the sustainability index of the MWFs used during the machining process with

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . .

1865

the MQL setup. This study focuses on quantifying the overall sustainability index of the MWFs in the manufacturing system applied at the cutting zone using the MQL setup. The sustainability index can be assessed based on the valuation and aggregation of deterministic and nondeterministic components.

Evaluation of Deterministic Components of SI The deterministic components associated with MWFs used during the machining applied with MQL setup are the costs related to machining, energy consumption, and waste management. The scope of the work is restricted to the sustainability evaluation of MWFs that are applied to the cutting zone with the MQL technique. Figure 8 demonstrates the various elements of the manufacturing system with MWFs and MQL setup. These costs can be numerically evaluated using the techniques available in the literature. The total costs include the various costs that are shown in Fig. 9. Even though all these costs are to be mentioned while discussing product manufacturing, as the main focus of this chapter is on MWFs and hence the Fig. 10 illustrates the cost associated with MWFs and MQL setup that are specifically adopted from the Fig. 9 (Gao et al. 2016a). The deterministic components associated with MWFs and MQL are classified as demonstrated in Fig. 10. To evaluate these deterministic components, process-based cost modeling can be used. This process-based cost modeling also called the bottom-up cost calculation method, has been used to evaluate the costs associated with product manufacturing

Fluids - Fatty alcohols - Synthetic esters Machine tool - MQL supply - Upgradability

Equipment - Int. / Ext. feed - 1 or 2 Channels

MQL

Tools - Internal feed - External feed

Settings - Air flow - Oil flow

Fig. 8 Elements of the manufacturing system with MWFs and MQL (Weinert et al. 2004)

1866

M. Vardhanapu and P. K. Chaganti

Fig. 9 Total costs associated with product manufacturing (Gao et al. 2016b)

Fig. 10 Total costs associated with MWFs and MQL setup

(Gao et al. 2016a), and the equations, as well as the analysis, have been thoroughly resonating the past literature (Lajevardi et al. 2011). The limitations of the previously available cost models are addressed in this process-based cost model (Gao et al. 2016a). The various costs demonstrated in Fig. 10 are evaluated using the

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . .

1867

following equations, and upon evaluating all of them and summing up all of them would give the deterministic components of SI. The methodology used to formulate the equations is based on the literature (Gao et al. 2016a).

Raw Material Cost The raw materials that are used to prepare the MWFs are oils, add-ons like surfactants, nanoparticles (in case of nanoparticles-based MWFs), corrosion inhibitors, etc. Let V be the volume of the MWF and cR is the total raw material cost. If i (i = 1, 2, 3 . . . x) types of raw material are used. Let PR be the unit price of each item and the quantity used is nR . The total cost of raw materials used in the preparation of MWFs can be obtained by using the below formula: cR =

x i=1

PRi × nRi

(1)

MWFs Synthesis Process Equipment Cost To synthesize the MWFs, special equipment like glass wear, magnetic or regular stirrers, vibrating machinery, heat source, etc. are used. The cost of the equipment is to be incorporated into the total cost. Let tE represent the life span of the equipment in years. P is the volume of MWF produced using the equipment over the life span of the equipment. Let c1 , c2 , c3 . . . . cn be the costs of the “n” number of equipment that are used in the process of synthesis of MWFs. Let cE be the total cost of equipment. To calculate the cost of the equipment, the below-mentioned formula can be used. cE =

(c1 + c2 + c3 . . . . . . . . . . + cn ) tE × P

(2)

Similarly, the remaining costs, namely, MQL setup cost (cSO ), maintenance cost (cM ), and storage cost (cST ), associated with the usage of MWFs in combination with MQL that are represented in Fig. 9 can be calculated numerically and all these costs are summed up to obtain the total machining cost component amongst the deterministic elements of SI. Let CM be the total machining cost component and the total machining cost component (CM ) is the f (cR , cE , cSO , cM , cST ). By using a similar method, the remaining two deterministic components, the energy (power consumption) and waste management components, can be numerically determined. Let CP be the total energy consumption component associated with usage of energy during the production of MWFs and during their usage of machining a product. Therefore, the CP is the f (ePR , eU ), where ePR is the cost component of energy consumed during the production of MWFs and eU is the cost component of energy consumed during the usage of MWFs for machining (i.e., for pumping the MWFs from the MWF storage tank to the cutting zone via MQL setup). Let CW is the total waste management components associated with the management of waste that is generated during the production and usage and post-usage of MWFs. Therefore, the CW is the f (wPR , wD , wR ), where wPR , wD are the cost component of waste managed during the production and disposal of MWFs, respectively. The wR

1868

M. Vardhanapu and P. K. Chaganti

is the cost component of the recycled waste, which is a positive input to the system as it again adds value in the form of revenue, while wPR and wD are negative inputs. Among these waste management components, the wD and wR are negligible as the proposed system is using an MQL setup for the application of MWFs at the cutting zone. The whole idea of using MQL is to spray the MWF in the form of a mist which cannot be recycled and the disposal only happens in the case of spoiled or outdated/unused MWFs. Once all these three cost equivalent components are determined numerically, the total deterministic component of SI for the system under consideration can be obtained by multiplying each of three deterministic components with a weighting factor and summing up all of them. The weighting factors for three individual components can be taken according to their respective impact on the triple bottom line and their respective performance levels. Let SOP be the total deterministic component of the sustainability index and the SOP can be obtained by the summation of the three deterministic components of SI, that is, the cost components associated with machining, energy consumption, and waste management. Therefore, SOP can be mathematically written as (Jaafar et al. 2007), SOP = (WM × CM ) + (WP × CP ) + (WW × CW )

(3)

WM , WP , and WW are the weighting factors respective to machining, power consumption, and waste management. These weighting factors take a value depending on the performance levels.

Understanding the Evaluation of Nondeterministic Components Using a Case Study Several analytical tools can be used to calculate the deterministic component of the sustainability index for the system under consideration. In contrast, the nondeterministic components are nonquantifiable, and their fuzziness makes it a difficult task to evaluate them. The three nondeterministic components are social, economic, and environmental components. In a manufacturing system, they are more precisely termed as the operational safety, personal health of the operator, and the ecological impact. To evaluate these nonquantifiable components, Najiha et al. (2016) implemented the fuzzy logic technique in his work, and this approach gives very reliable results in case the uncertainties are involved. Making use of fuzzy logic to evaluate these nondeterministic components is very well echoing the past literature. Granados et al. (2009) also employed the fuzzy logic and analytical models to assess sustainability in manufacturing, more precisely in machining. Applying fuzzy logic, which is a theoretically comprehensive and straightforward methodology to evaluate nondeterministic components of SI, can be well justified as it deals with parameters that are not easy to understand, and it also takes into consideration the values or the opinions. Fuzzy logic happens to play a successful role in the case of human subjectivity.

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . .

1869

A similar attempt was made in the past to quantify the nondeterministic components of SI of machining a part using MWFs for lubrication with MQL setup (Granados et al. 2009). The fuzzy logic was applied to MWFs used in machining operation involving MQL type of lubrication to predict the nondeterministic component of SI using the fuzzy logic toolbox of the MATLAB software (Najiha et al. 2016). The assumptions were made at the three different levels contributing to three different factors of operational safety, operator’s health, and environmental impact, as illustrated in Table 2. Once these three levels are well defined, the three factors were considered as linguistic variables and processed through fuzzy logic rules. As described in this work, the application of fuzzy logic is a threetier process consisting of fuzzification of inputs, application of fuzzy operator, and application of the aggregation method. Once the composition and aggregation is done, the aggregated fuzzy output is then to be defuzzified to get a numeric value. The CoG defuzzification method can be used to defuzzify the aggregated fuzzy output. A three-level fuzzy logic-based model is constructed to evaluate the uncertainties, thereby assessing the nondeterministic components of SI, and after defuzzification of the aggregated result outputs, the numerical values obtained by Najiha et al. (2016) has been tabulated in Table 3. The total nondeterministic component of SI for a particular machining operation with MQL setup obtained through their study is 64.2%. Their results are compared with the results obtained by Granados et al. (2009), who considered the entire conventional turning operation to evaluate the nondeterministic component, while Najiha et al. (2016) considered only MWFs with MQL set up as part of their study. The results did not vary much, though. Let SSHE be the total nondeterministic component of the system considered by them.

Table 2 Three different levels contributing to three different factors as segregated by Najiha et al. (2016) Level 1 Environmental friendliness Process safety Operational safety

Level 2 Safety operator Safety of process Waste generation Workplace contamination Level of recycling

Level 3 Mist generation Coolant losses Chip recycling rate Toxic & dangerous additives Equipment malfunction

Table 3 Numerical interpretation of outputs (sustainability indicators) (Najiha et al. 2016) Output Environmental friendliness (%) Operator’s personal safety (%) Operational safety (%) SSHE

Sustainability Indicators (%) 79 50 63.6 64.2

1870

M. Vardhanapu and P. K. Chaganti

Overall Sustainability Index The scheme for the evaluation of the overall or total sustainability index is demonstrated in Fig. 11. Initially, while evaluating the sustainability of any system, the societal, economic, and ecological factors are to be identified, and then those factors are to be converted into sub-indices. These indices are again classified as deterministic and nondeterministic components depending upon the nature of the component. The total nondeterministic component of the sustainability index of any system or operation under consideration can be evaluated using a fuzzy logicbased methodology. The total nondeterministic component obtained can be summed up with the numerically or analytically calculated total deterministic component to evaluate the total overall sustainability index of any system under consideration. Let “S” be the overall sustainability index of a system under consideration and it can be numerically written as, S = (CSH E × SSH E ) + (COP × SOP )

Fig. 11 Scheme for the evaluation of the overall sustainability index

(4)

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . .

1871

where S is the overall sustainability index, SSHE is the total nondeterministic component of SI, SOP is the deterministic total component of SI {can be obtained from Eq. (3)}, COP is the weighting factor for deterministic component, and CSHE is the weighting factor for the nondeterministic component.

Example Having discussed the various aspects such as elements, metrics, and measurement processes of the sustainability index in the above sections, in this section, the evaluation SI is illustrated with an appropriate example.

Material and Process The system under consideration for evaluating the SI is metalworking fluids (MWFs) used during the machining process with minimum quantity lubrication (MQL). The machining process selected for this example is a simple step turning. Based on the available literature, with more work being focused on steels, mild steel/AISI D2 tool steel/alloy steel are the materials chosen for this example. In the machining process, the independent process parameters such as speed, feed, and depth of cut play a major role in achieving optimum levels of the dependent parameters like surface roughness, tool life, material removal rate (MRR), tool wear rate (TWR), and residual stresses. There are various MWFs that are taken into consideration and this example is confined to petroleum-based mineral oils only.

Sustainability of Various Elements Used during Various Stages of the Life Cycle of MWFs One of the principles of sustainability measurement is the consideration of each and every stage in the life cycle of a product (Fiksel et al. 1999). The various stages in the product life cycle are pre-production, production, usage, and post usage. Figure 12 illustrates the different elements and energies involved in various stages of the life cycle of MWFs [3].

Categorization of Deterministic Components and Nondeterministic Components The categorization of deterministic and nondeterministic components falling under the triple bottom line is shown in Table 4. In each of the TBL, the various factors are again classified as deterministic and nondeterministic components. There may

1872

M. Vardhanapu and P. K. Chaganti

Fig. 12 Various elements and energy involved in various stages of the MWFs life cycle

be other abysmal and many relatable factors apart from the ones that are mentioned in Table 4; however, these are the factors identified from the literature. The deterministic components can be numerically evaluated using analytical methods, while the nondeterministic components are to be quantified by employing the fuzzy logic method. To evaluate the deterministic components, the weighting and aggregating techniques discussed in Sect. “Evaluation of Deterministic Components of SI” can be employed, and the overall deterministic component of the MWFs system in combination with MQL setup while performing a turning operation can be determined. The nondeterministic components of the case study, as mentioned in Table 4, can be quantified and aggregated by using a fuzzy logic-based methodology, as discussed in Sect. “Understanding the Evaluation of Nondeterministic Components Using a Case Study”. Decisions regarding the nondeterministic elements like the toxicity of chemical components and the toxicity data sheets can be employed to find out the permissible limits. The reliability and probabilistic models can be employed individually to quantify the nondeterministic elements like rate of injuries, labor injury costs, machine equipment downtimes, and related costs. Thus, once the individual nondeterministic components are quantified, the weighting and aggregation is done to find out the total nondeterministic component of SI.

Evaluation of SI for Deterministic Components In this example, the concentration is on the MWFs that are applied on to the cutting zone using minimum quantity lubrication (MQL), sometimes referred to as near dry machining or micro-lubrication (Ali et al. 2011). The quantity of MWFs used for a

Economic Aspects Deterministic Raw material cost Recycle revenue Equipment costs Capital Labor cost Storage cost Electricity cost Disposal costs

Nondeterministic Labor injury costs Downtime costs (equipment malfunction)

Environmental Aspects Deterministic Life-cycle energy CO2 emissions The carbon footprint of electricity generation and water treatment Biodegradability of the disposed components Nondeterministic Smog creation in the workplace Depletion of fossil fuels used in MWFs preparation Level of recycling Workplace contamination

Table 4 Categorization of deterministic and nondeterministic components of the hypothetical example Societal Aspects Deterministic Rate of injuries Lost time for injuries

Nondeterministic Employee satisfaction Mist generation Occupational hazards over prolonged usage of MWFs (cohort studies) Toxicity of chemical components

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . . 1873

1874

M. Vardhanapu and P. K. Chaganti

Mixing chamber

Dynamometer

Pressure Gauge

Tool

Head stock Pressure Gauge

Work-piece

Tail stock

MWFs

Compressed air

Fig. 13 Line diagram of the experimental setup for turning operation along with MQL circuit

Fig. 14 Schematic diagram of the workpiece after turning operation

particular MRR over a time period is calculated. A cylindrical bar (rod) with 35 mm diameter and 120 mm length, which is to be step turned to a final diameter of 33 mm as represented in Fig. 14, is considered for the hypothetical example. The MRR of the hypothetical step turning is determined using the following formula (Pant et al. 2017) (Table 5, Fig. 14): MRR = π ∗ f ∗ d ∗ Davg ∗ N

(5)

The experimental setup for the example is represented in the form of a line diagram in Fig. 13. For this case, the feed, cutting speed, depth of cut, length of the rod, initial and final diameter of the rod, and rotational speeds are assumed to be 0.2 mm/rev, 190 m/min, 0.1 mm, 120 mm, 35 mm and 33 mm, 500 rpm, respectively, from the available literature, assuming this optimal setting of the process parameters will give the nominal output. MRR for the cylindrical bar with the abovementioned optimal setting of process parameters is calculated to be 1068.14 mm3 /min. The cutting time (also known as machining time) for this example can be obtained by the L/(f ∗ N) (Pant et al. 2017). The cutting time is calculated to be 1.2 min. The motive of finding the MRR is to calculate the average lubricant quantity consumed for its

Initial Diameter (Di ) 35 mm

Final Diameter (Df ) 33 mm

Average Diameter (Davg ) 34 mm Length (L) 120 mm

Feed (f ) 0.2 mm/rev

Table 5 Assumed values of the various parameters used in the hypothetical example Rotational Speed (N) 500 rpm

Depth of Cut (d) 0.1 mm

Cutting Speed (v) 190 m/min

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . . 1875

1876

M. Vardhanapu and P. K. Chaganti

corresponding machining time. Assuming a flow rate of MQL to be 150 ml/hr., the total quantity of lubricant that is flown at the cutting zone for a cutting time of 1.2 min is 3 ml. Once the cutting time is determined, the amount of MWFs used in that particular time is determined. The cost of the MWF for the determined usage quantity over a period can be calculated, which is one of the deterministic components of SI. Depending upon the cutting time calculated for the hypothetical example, the annual usage of MWFs is approximately calculated to be 252 liters assuming the number of working days to be 20 per month with 7 actual cutting hours per day. The total cost (cMWF ) of MWFs for 252 liters is | 25,200 (assuming cost of MWFs/liter = | 100). The cost (cSO ) of MQL setup is assumed to be | 1000. The average annual storage and maintenance cost (cM & cST ) of MWFs is assumed to be | 2000. Based on the above set of assumptions, the total cost component (CM ) for the annual usage of MWFs during the turning operation can be determined by weighting and aggregating all the costs. The other deterministic component is the cost of the power or the energy consumed during the production and usage of MWFs. Let the total energy consumed for the pumping of 252 liters of MWFs is 504 units or kWh (assuming that the 5 liters of pumping consume 10 units of energy). Therefore, the total cost (eU ) for pumping 252 liters is determined to be | 4032 (| 8 per unit). Let the total units of energy assumed to be consumed during the production of MWFs be 50 units and the cost of energy consumed (ePR ) during the production of MWFs is determined to be | 8. The total deterministic component (CP ) of energy can be determined by weighting and aggregating both the abovementioned costs. Having calculated the two deterministic components, it is assumed that the total annual cost component (CW ) of waste management is approximately | 2000. The overall deterministic component can be calculated by using Eq. (3) for different weights of WM , WP , and WW . The total deterministic component of the system under consideration is calculated to be 82.12% at the weights of WM , WP , and WW taken as 0.2, 0.4, and 0.4, respectively.

Evaluation of Nondeterministic Components of SI Following a similar kind of division of levels of the nondeterministic factors and the same set of assumptions for the present work, the SI index of nondeterministic components for the system under consideration can be obtained from the literature (Najiha et al. 2016). The SI of nondeterministic components (SSHE ) for the MWFs with MQL setup, thus obtained from the previous work, is 64.2%, which was quantified using a fuzzy logic-based methodology.

Sustainability Index of the System under Consideration To calculate the overall sustainability, the weights for the deterministic and nondeterministic components are assumed to be 0.6 and 0.4. The SI for deterministic

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . .

1877

components and nondeterministic components are 82.12% and 64.2%, respectively (from Sects. “Evaluation of Nondeterministic Components of SI” and “Sustainability Index of the System under Consideration”). Therefore, the overall sustainability index aggregated as per the weights for this hypothetical case study is 74.96%. Thus, the SI for any system can be calculated using this methodology, which is a combination of numerical or analytical methods to compute the deterministic components and fuzzy logic-based methodology to determine the nondeterministic components of SI. From the Eq. (4), S = (CSH E × SSH E ) + (COP × SOP ) Therefore, S = (0.4 × 64.2) + (0.6 × 82.12); S = 74.96% The other aspects to be considered while calculating the deterministic components of SI for any system is the quantification of CO2 emissions, the Carbon footprint of electricity generation and water treatment, the biodegradability of the disposed components, rate of injuries, and lost time for injuries which were not accounted in the presented example. The most realistic approach would be to consider these aspects while calculating the deterministic part of the sustainability index for its precise assessment.

Conclusion/Summary Sustainability can be classified into process sustainability, product sustainability, or overall system sustainability. In this chapter, previous literature is thoroughly referred to assess sustainability at the process and the product levels. This chapter focuses on the process implementation to assess the sustainability index with which the overall sustainability of any system under consideration of the study can be determined. The system under consideration in this chapter is the MWFs used for various advantageous purposes during machining operations and in combination with minimum quantity lubrication (MQL). The sustainability index is a combination of three deterministic components and three nondeterministic components, hence to assess the overall sustainability of any system, these deterministic and nondeterministic components are to be evaluated. In this chapter, the numerical or analytical approach of evaluating the deterministic cost components of the SI is illustrated, which can be evaluated using bottom-up cost calculations. These deterministic components are relatively easy to evaluate because of their deterministic nature, while the difficult part remains to be to evaluate the nonquantifiable/nondeterministic components that are operational safety, the health of the operator and environmental impact. To evaluate these value or opinion based components, the Fuzzy logic technique can be used effectively as a fuzzy logic gives reliable justifications to the relatively difficult to understand parameters and for things that involve human subjectivity. The defuzzification methods can be implemented to convert the aggregated outputs of nondeterministic components

1878

M. Vardhanapu and P. K. Chaganti

to achieve a numerical value. To understand the computations of nondeterministic components, an example is presented. Once the deterministic and nondeterministic components are evaluated, they can be summed up as per equations available in the literature to find out the overall sustainability index, taking into account the weighting factors. This chapter paves the way for understanding a structure and scheme of evaluating the sustainability index of any MWFs system with MQL setup. Thus, the sustainability level of any system can be assessed by evaluating the sustainability index as a part of achieving sustainable manufacturing in the era of Industry 4.0.

Nomenclature MWFs SI TBL AHP MQL PR nR cR tE V cE cSO cM cST ePR eU wPR wD wR CM CP Cw SOP SSHE S CSHE COP MRR TWR f d Di Df Davg N

Metal Working Fluids Sustainability Index Triple bottom line Analytical hierarchy process Minimum quantity lubrication Unit price of raw material per Kg (in |) Quantity of raw material in Kg Total cost of the raw materials (in |) Total life span of equipment in years Volume of MWFs Total cost of equipment used for the synthesis of MWFs (in |) Total cost of MQL setup (in |) Total maintenance cost of MWFs (in |) Total Storage cost of MWFs (in |) Total cost for the energy consumed in the production of MWFs (in |) Total cost of energy consumed for pumping the MWFs from the storage to the cutting zone (in |) Total cost associated with waste management during the production of MWFs (in |) Total cost associated with waste management during the disposal of the used MWFs (in |) Total cost associated with recycling of the used MWFs (positive input to the system) Total production cost of MWFs (in |) Total cost of energy consumption during the production and usage of MWFs (in |) Total cost associated with the waste management during production, usage and postusage of MWFs Total value for deterministic components of SI Total value for nondeterministic components of SI Overall sustainability index weighting factor for nondeterministic components (0 to 1) weighting factor for deterministic components (0 to 1) Material removal rate in mm3 /min Tool wear rate in mm3 /min Feed in mm/Rev. Depth of cut in mm Initial diameter of cylindrical work piece in mm Final diameter of cylindrical work piece in mm Average diameter of the cylindrical work piece in mm Rotational speed of the work piece in rpm

73 Sustainability Index of Metalworking Fluids in the Manufacturing. . . v L cMWF

1879

Cutting speed in m/min Length of the work piece Total Purchase cost of MWFs

References Ali SM, Dhar NR, Dey SK (2011) Effect of minimum quantity lubrication (MQL) on cutting performance in turning medium carbon steel by uncoated carbide insert at different speed-feed combinations. Adv Prod Eng Manag 6:185–196 Fiksel J, McDaniel J, Mendenhall C (1999) Measuring progress towards sustainability principles: process and best practices. Green Ind Netw Conf Best Pract Proc 2693:1–25 Gao Q, Lizarazo-Adarme J, Paul BK, Haapala KR (2016a) An economic and environmental assessment model for microchannel device manufacturing: Part 1 – methodology. J Clean Prod 120:135–145. https://doi.org/10.1016/j.jclepro.2015.04.142 Gao Q, Lizarazo-Adarme J, Paul BK, Haapala KR (2016b) An economic and environmental assessment model for microchannel device manufacturing: Part 1 – methodology. J Clean Prod 120:135–145. https://doi.org/10.1016/j.jclepro.2015.04.142 Gholami H, Saman MZM, Sharif S et al (2020) A general framework for sustainability assessment of sheet metalworking processes. Sustain 12. https://doi.org/10.3390/su12124957 Granados S, Jawahir IS, Fernandez J (2009) A Comprehensive Criterion for Sustainability Evaluation of Machining Processes, in: Proceedings of the 7th Global Conference on Sustainable Manufacturing, IIT Madras (Chennai, India), pp.385–391 Gupta A, Jayal AD, Chimienti M, Jawahir IS (2011) Glocalized solutions for sustainability in manufacturing. Glocalized Solut Sustain Manuf 240–245. https://doi.org/10.1007/978-3-64219692-8 Haapala KR, Zhao F, Camelio J et al (2013) A review of engineering research in sustainable manufacturing. J Manuf Sci Eng Trans ASME 135. https://doi.org/10.1115/1.4024040 Hegab H, Darras B, Kishawy HA (2018) Sustainability assessment of machining with Nano-cutting fluids. Proc Manuf 26:245–254. https://doi.org/10.1016/j.promfg.2018.07.033 Jaafar IH (2006) Total life-cycle considerations in product design for sustainability: a framework for comprehensive evaluation 10th international research/expert conference Total life-cycle considerations in product design for sustainability: a framework for Compreh Jaafar IH, Venkatachalam A, Joshi K, et al (2007) Product Design for Sustainability: a new assessment methodology and case studies Jawahir IS, Jayal AD (2011) Advances in sustainable manufacturing. Adv Sustain Manuf:299–305. https://doi.org/10.1007/978-3-642-20183-7 Kishawy HA, Hegab H, Saad E (2018) Design for sustainable manufacturing: approach, implementation, and assessment. Sustain 10:1–15. https://doi.org/10.3390/su10103604 Lajevardi B, Leith SD, King DA, Paul BK (2011) Arrayed microchannel manufacturing costs for an auxiliary power unit heat exchanger. 61st Annu IIE Conf Expo Proc Lu T, Rotella G, Feng SC et al (2012) Sustainable manufacturing. Sustain Manuf:59–64. https:// doi.org/10.1007/978-3-642-27290-5 Machado CG, Winroth MP, Ribeiro da Silva EHD (2020) Sustainable manufacturing in industry 4.0: an emerging research agenda. Int J Prod Res 58:1462–1484. https://doi.org/10.1080/ 00207543.2019.1652777 Najiha MS, Rahman MM, Kadirgama K (2016) Minimum quantity lubrication: quantifying nondeterministic component of sustainability index for machining operations. Int J Automot Mech Eng 13:3190–3200. https://doi.org/10.15282/ijame.13.1.2016.6.0266 Palit S, Hussain CM (2018) Environmental management and sustainable development: a vision for the. Future 1–17 Pant G, Kaushik S, Rao DK, Negi K (2017) Study and analysis of material removal rate on lathe operation with Varing parameters from CNC lathe machine. Int J Emerg Technol (Special Issue NCETST-2017) 8:683–689

1880

M. Vardhanapu and P. K. Chaganti

Rojko A (2017) Industry 4.0 concept: background and overview. Int J Interact Mob Technol 11:77– 90. https://doi.org/10.3991/ijim.v11i5.7072 Sandoval-Solis S, McKinney DC, Loucks DP (2011) Sustainability index for water resources planning and management. J Water Resour Plan Manag 137:381–390. https://doi.org/10.1061/ (ASCE)WR.1943-5452.0000134 Schmidt W-P (2007) Ford of Europe’s product sustainability index cost. OECD Work Sustain Manuf Prod Compet:1–10 Sen B, Mia M, Krolczyk GM et al (2019) Eco-friendly cutting fluids in minimum quantity lubrication assisted machining: a review on the perception of sustainable manufacturing. Korean Society for Precision Engineering Skerlos SJ, Adriaens P, Hayes K et al (2001) Challenges to achieving sustainable aqueous systems: a case study in metalworking fluids. Proc – 2nd Int Symp Environ Conscious Des Inverse Manuf 566–571. https://doi.org/10.1109/ECODIM.2001.992425 Skerlos SJ, Hayes KF, Clarens AF, Zhao F (2008) Current advances in sustainable metalworking fluids research. Int J Sustain Manuf 1:180–202. https://doi.org/10.1504/IJSM.2008.019233 Weinert K, Inasaki I, Sutherland JW, Wakabayashi T (2004) Dry machining and minimum quantity lubrication. CIRP Ann-Manuf Technol 53:511–537. https://doi.org/10.1016/S00078506(07)60027-4 Zhang X, Lu T, Shuaib M et al (2012) A metrics-based methodology for establishing product sustainability index (ProdSI) for manufactured products. Leveraging Technol a Sustain World – Proc 19th CIRP Conf Life Cycle Eng 435–441. https://doi.org/10.1007/978-3-642-29069-5_74

Important Websites https://www.bbva.com/en/what-is-a-sustainability-index-used-for/ https://www.inderscience.com/info/inarticle.php?artid=19233 https://www.lifecycleinitiative.org/starting-life-cycle-thinking/life-cycle-approaches/life-cyclesustainability-assessment/ https://www.oecd.org/innovation/green/toolkit/oecdsustainablemanufacturingindicators.htm https://www.productionmachining.com/articles/sustainable-metalworking-with-minimumquantity-lubrication https://www.tandfonline.com/doi/abs/10.1080/00207543.2014.993773

Wind Energy System: Data Analysis and Operational Management

74

Vikas Khare and Cheshta J. Khare

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Basics of Descriptive Statistics in the Field of Wind Energy System . . . . . . . . . . . . . . . . . . Measurement of Central Tendency of Wind Energy System Data . . . . . . . . . . . . . . . . . . . . . Percentile and Quartile (PQ) Assessment of Wind Energy System Data . . . . . . . . . . . . . . . . Mean Absolute Deviation, Variance, and Standard Deviation (MVS)-Based Assessment of Wind Energy Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Frequency Distribution of Prefeasibility Data of Wind Energy System . . . . . . . . . . . . . . . . . Measurement of Central Tendency and Variability of Wind Energy Grouped Data . . . . . . . Regression Analysis of Wind Energy System Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross Tabulation and Scatterplot of Wind Velocity Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . Operational Management of Wind Energy System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Operational Management of Flexible Generation System . . . . . . . . . . . . . . . . . . . . . . . . . . . Operational Management-Based Maintainability and Availability Function of Wind Energy System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Maintainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Confidence Level of Repairable System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Website for Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1882 1885 1886 1894 1895 1897 1897 1901 1904 1904 1908 1911 1911 1916 1918 1918 1919 1920

V. Khare () STME, NMIMS, Indore, India C. J. Khare SGSITS, Indore, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_62

1881

1882

V. Khare and C. J. Khare

Abstract

In the recent scenario, renewable energy system plays very vital role, and growth of such type of green industry is increased in very splendid way because green energy also decreases the pollution from the atmosphere. Wind energy system is one of the highly recognized renewable energy sources to generate the electricity in all over the world. This chapter introduces data analysis and operational management of wind energy system because in industry 4.0, data analytics and operation and strategic management are the very hot topic and are utilized for prediction analysis and total quality management of wind energy system. In this chapter, we analyzed frequency distribution of wind velocity of certain location and qualitative and quantitative analysis of wind energy system, and further result of given data is also assessed through R-language. After the completion of prefeasibility assessment of wind energy system, we explain optimum design of wind energy system through operational management process. A three-tier model of wind energy system for assessing the competitiveness of a location is presented, and location decision in wind energy distribution system must address the requirement of responsiveness. Wind energy component manufacturing company needs to spend a considerable amount of money to carry inventory. The cost of stores and the administrative cost related to maintaining inventory and accounting for it form a significant part of this cost. So arise such type of issues in this chapter to explain inventory and total quality management of wind energy system for increasing the system efficiency and reducing the loss. Keywords

Wind energy system · Total quality management · Decision science · R-language

Introduction In the recent scenario, it is necessary to develop green and effective environment and reduce greenhouse gases from the atmosphere. Nowadays, lots of country are working toward the renewable energy sources such as solar energy, biomass, small hydro system, tidal energy system, wave energy system, and wind energy system. Wind energy system is one of the most prominent energy sources, and with phenomenal development in the field of electricity generation through wind energy system, wind power data sources have raised sharply. Wind speed or wind flow velocity is a fundamental atmospheric rate which is caused by air moving from high pressure to low pressure usually due to changes in temperature difference. Wind speed data are measured at height above the ground level and is known as anemometer height. The mathematical modeling of wind energy conversion system depends on wind turbine dynamic characteristics and parameter of the wind generator. Nondimensional shown as a motivation behind the tip speed proportion is a fundamental trait of wind turbine. Fundamentally produced power is, to a great extent, relying upon the 3D square of the breeze speed. In the wind energy

74 Wind Energy System: Data Analysis and Operational Management

1883

framework, it is important to successful use of given information and viable working of in general wind energy framework. Comprehensive utilization of wind power information investigation and operational administration can give a compelling method to safe activity of excellent force flexibly of a breeze energy framework. The viable objective of information examination and operational administration research in the field of wind energy framework is to gotten mindful of ignorance, and unbiasedly, there are verifiable truths and obscure realities which influence the working of wind energy framework. Information Analysis causes us, in realizing what we don’t have, the foggiest idea and is done in two different ways: from little too huge and from large excessively little. It is important to build up an appropriate knowledge perspective on information examination in the field of wind energy framework and model out the structure of wind energy framework through the successful operational administration. The logical disobedience in the breeze energy framework area is creating huge volumes of information with significant power in the business and viable cycles of framework administrators, age organization, and the customer. Information investigation strategies can be applied to state assessment, estimating, and control issues, just as to help the cooperation of market specialists in the power market. The information removed from a breeze energy framework and market information has a huge effect in key execution pointers, as operational effectiveness (e.g., working costs), speculation deferral, and nature of gracefully. Besides, plans of action identified with information examination and operational administration handling are developing and boosting new energy administrations. Wind speed information is estimated tallness over the ground and is known as anemometer stature. The numerical plan of wind energy change framework incorporates wind turbine elements and generator demonstrating. Nondimensional introduction as a component of the tip speed proportion is a fundamental attribute of wind turbine. Fundamentally created energy is generally relying upon the shape of the wind speed. The output of mechanical power captured from the wind by a wind turbine can be formulated as Wt =

Cp λρAV 3 2

and torque developed by a wind turbine can be expressed as Tt =

Wt ωm

where Wt is the output power, Tt the torque developed by wind turbine, Cp the power coefficient, λ the tip speed ratio, ρ the air density in kg/m3 , A the frontal area of wind turbine, and V the wind speed. In wind energy system data, Weibull distribution factor which is the distribution of wind speed over a year is 1.961. Autocorrelation factor is 0.86 which presents random behavior of wind speed. Lots of researchers work toward the data analysis of wind energy system, where Azad and Rasul (2015) described assessment of power generation through wind energy system by Weibull distribution. Three different Weibull distribution methods

1884

V. Khare and C. J. Khare

were used to find out Weibull parameters which were verified using different widely acceptable statistical tools. Relative percentage error, chi-square error, analysis of variance, etc., are the efficient statistical tools to rank the method which was used in this study. Khare (2020) explained assessment of solar wind hybrid renewable energy system through big data analysis. The objective of this research is to present a technological framework for the management of large volumes, variety, and velocity of solar system-related information through big data tools such as Hadoop to support the assessment of solar and wind energy system. Zhang (2020) described calculation and analysis of wind turbine health monitoring indicators based on the relationships with SCADA data. Taking the normal operation state model of the wind turbine as the standard reference and based on the Euclidean distance between the state model curve and the standard model curve, the health index of the wind turbine operation state is proposed. Luo (2017) explained wind turbine condition monitoring by physical-based analysis. Physics-based data analysis methods for improving damage feature extractions, especially for that lower-frequency shaftrelated component, are proposed and demonstrated. This chapter developed new insights in the field of wind energy system and is distributed into 12 sections. Sections “Basics of Descriptive Statistics in the Field of Wind Energy System”, “Measurement of Central Tendency of Wind Energy System Data”, “Percentile and Quartile (PQ) Assessment of Wind Energy System Data”, “Mean Absolute Deviation, Variance, and Standard Deviation (MVS)-Based Assessment of Wind Energy Data”, “Frequency Distribution of Prefeasibility Data of Wind Energy System”, “Measurement of Central Tendency and Variability of Wind Energy Grouped Data”, “Regression Analysis of Wind Energy System Data”, and “Cross Tabulation and Scatterplot of Wind Velocity Data“” described assessment of wind energy system through data analysis further. Sections “Operational Management of Wind Energy System”, “Operational Management of Flexible Generation System“”, and “Operational Management-Based Maintainability and Availability Function of Wind Energy System” explained operational management of wind energy system. Section “Basics of Descriptive Statistics in the Field of Wind Energy System” described basics of descriptive statistics in the field of wind energy system. Sections “Measurement of Central Tendency of Wind Energy System Data” and “Percentile and Quartile (PQ) Assessment of Wind Energy System Data” explained measurement of central tendency and percentile and quartile assessment of wind energy system data, respectively. In Sect. “Mean Absolute Deviation, Variance, and Standard Deviation (MVS)-Based Assessment of Wind Energy Data”, assessment of wind energy system data is done through the mean absolute deviation, variance, and standard deviation. Frequency distribution of prefeasibility data of wind energy system is assessed in Sect. “Frequency Distribution of Prefeasibility Data of Wind Energy System”. Section “Measurement of Central Tendency and Variability of Wind Energy Grouped Data” described measurement of central tendency and variability of wind energy grouped data. Section “Regression Analysis of Wind Energy System Data” analyzed regression analysis of wind energy system data. In Sect. “Cross Tabulation and Scatterplot of Wind Velocity Data”, cross tabulation and scatterplot of wind velocity data are

74 Wind Energy System: Data Analysis and Operational Management

1885

presented. Section “Operational Management of Wind Energy System” described operational management of wind energy system. Operational management of flexible generation system is analyzed in Sect. “Operational Management of Flexible Generation System”. Section “Operational Management-Based Maintainability and Availability Function of Wind Energy System” described operational managementbased maintainability and availability function of wind energy system. This chapter ends with the conclusion and reference section. In this chapter, some analysis is also done through the R-language programming.

Basics of Descriptive Statistics in the Field of Wind Energy System Prior to installation and operation, the prefeasibility study of wind energy system has to be done. In any wind energy projects, an initial study is undertaken to determine whether it is worthwhile to continue to the viability study stage. A precise feasibility study should provide a chronological background of the projects. In addition to climate condition of the application site, availability of wind energy sources, potential of wind energy sources, and load demand of application sites are included to find out the best location to develop a wind energy system. Generally, a feasibility study is preceded by technical development and project implementation. It must therefore be conducted with an impartial approach to provide information upon which decisions can be based. Here, you can apply concept of descriptive statistics in primary level assessment of wind energy system. Following are the primary level qualitative factors, which are necessary to assess through descriptive statistics for proper functioning of wind energy system (Tummala 2019): • • • • •

Availability of wind velocity Load capacity Size of the generator Hub height Cost analysis of the overall project

If any wind energy company want to establish 50 kW wind power plant at a particular location, top-level management visits number of sites and decides which location will be more compatible for wind energy system, so in that case, one question keeps in mind, amount of wind velocity at that place. So now, it is necessary to data analysis in terms of descriptive statistics of wind velocity of a number of locations. Here, first, you can assess measures of central tendency of wind velocity in terms of ungrouped data. Measures of central tendency yield information about the center, or middle part, of a group of number. For the analysis of wind energy system, it is necessary to know the level of data measurement (Fig. 1) represented by the numbers being analyzed. The appropriateness of the data analysis depends on the level of measurement of the data gathered. The phenomenon represented by the numbers determines the level of data measurement. The lowest level of data measurement of wind energy system

1886

V. Khare and C. J. Khare

Nominal Level

• Prefeasibility Assessment Data • Optimum Sizing Data • Control System Data • Reliability Analysis Data

Ordinal Level

• Ranking of wind energy system according to the plant capacity

Interval Level

• Day-wise, month-wise and year-wise wind velocity and load demand

Ratio Level

• Efficiency of generator, turbine and overall system

Fig. 1 Data measurement of wind energy system

is nominal data level, and it is only used to classify or categorize the given data sets. In the wind energy system, the terminology such as prefeasibility analysis, modelling, controlling, and reliability analysis comes under the category of nominal data because in this type of data measurement, whole data of wind energy system is categorized into prefeasibility analysis data, modelling data, controlling data, and reliability analysis data. In the ordinal level of wind energy system, data measurement is used to rank or order the different factors, for example, ranking of wind energy system according to the capacity of the plant. Nominal and ordinal data measurement is the nonmetric type of data because it shows only qualitative analysis of the wind energy system. At the next stage, data measurement through interval level and ratio level is called the metric level type and quantitative analysis of wind energy system. In interval level of data measurement of wind energy system, we can assess day-wise charging and discharging rates of electric battery and daywise, month-wise, and year-wise data of wind velocity at the particular location. Ratio-level data measurement is the highest level of data measurement, in which we can identify efficiency of generator, battery, and overall wind energy system.

Measurement of Central Tendency of Wind Energy System Data When you will apply the concept of central tendency to assess the wind velocity, then assessment is done through the “3M,” which means MODE, MEDIAN, and MEAN. The mode is the most frequently occurring value of wind velocity in a given set of data. Median is the middle value of wind velocity in an ordered array data. For an odd number of values of wind velocity, find the middle value of the

74 Wind Energy System: Data Analysis and Operational Management

1887

ordered array as a median, and for an even number of values of wind velocity, find the average of the middle two terms. The arithmetic mean is the average of a group of numbers and is computed by summing all numbers and dividing by the number of numbers (Hussain 2018a). So again, for establishment of 50 kW wind energy system, a company identifies three possible locations and collects monthwise average wind velocity (m/s) of an individual location in Table 1. Table 2 shows data of wind velocity in ordered array. Then according to the given data set of wind velocity, find out the value of mean, mode, and median. Figure shows the mean of wind velocity for locations 1, 2, and 3 through R-language programming (Figs. 2, 3, and 4). Mean of location 1 = (5.1 + 6.4 + 7.4 + 9.4 + 9.4 + 11.1 + 11.2 +11.3 + 12.6 + 13.9 + 14.3 + 15.8) /12 = 127.9/12 = 10.65 Mean of location 2 = (5.8 + 6.9 + 8.3 + 9 + 11 + 12.4 + 12.4 + 13.3 + 15 +16.5 + 17.3 + 18) /12 = 145.9/12 = 12.15

Table 1 Month-wise average wind velocity (m/s)

Table 2 Month-wise average wind velocity in ascending order (m/s)

Month January February March April May June July August September October November December Induction 1 5.1 6.4 7.4 9.4 9.4 11.1 11.2 11.3 12.6 13.9 14.3 15.8

Location 1 5.1 6.4 11.3 9.4 12.6 11.2 15.8 14.3 7.4 11.1 13.9 9.4

Location 2 9 6.9 11 12.4 15 13.3 17.3 18 5.8 12.4 16.5 8.3

Induction 2 5.8 6.9 8.3 9 11 12.4 12.4 13.3 15 16.5 17.3 18

Location 3 4.5 7.3 7.3 8.9 6.2 10.8 11.6 12.9 14.6 7.4 5.1 8.9

Induction 3 4.5 5.1 6.2 7.3 7.3 7.4 8.9 8.9 10.8 11.6 12.9 14.6

1888

Fig. 2 Mean of wind velocity for location 1 through R-Language

Fig. 3 Mean of wind velocity for location 2 through R-Language

Fig. 4 Mean of wind velocity for location 3 through R-Language

V. Khare and C. J. Khare

74 Wind Energy System: Data Analysis and Operational Management

1889

Mean of location 3 = (4.5 + 5.1 + 6.2 + 7.3 + 7.3 + 7.4 + 8.9 + 8.9 + 10.8 +11.6 + 12.9 + 14.6) /12 = 105.5/12 = 8.79 In the next step, you identify mode of wind velocity data of locations 1, 2, and 3. Figure shows mode of wind velocity data of locations 1, 2, and 3 through RLanguage programming (Figs. 5, 6, and 7). Mode of location 1 = most frequent value of wind velocity = 9.4 (number of occurrence 2) Mode of location 2 = most frequent value of wind velocity = 12.4 (number of occurrence 3)

Fig. 5 Mode of wind velocity for location 1 through R-Language

Fig. 6 Mode of wind velocity for location 2 through R-Language

1890

V. Khare and C. J. Khare

Fig. 7 Mode of wind velocity for location 3 through R-Language

Fig. 8 Median of wind velocity for location 1 through R-Language

Mode of location 3 = most frequent value of wind velocity = 7.3, 8.9 (number of occurrence 2) After the mean and mode, we identify the median of locations 1, 2, and 3. Figure shows median of locations 1, 2, and 3 through R-language programming (Figs. 8, 9, and 10). Median of location 1 = odd number of values, so median is the average of 6th and 7th number of the value of wind velocity = (11.1 + 11.2) /2 = 11.15 Median of location 2 = (12.4 + 12.4) /2 = 12.4 Median of location 3 = (7.4 + 8.9) /2 = 8.15

74 Wind Energy System: Data Analysis and Operational Management

1891

Fig. 9 Median of wind velocity for location 2 through R-Language

Fig. 10 Median of wind velocity for location 3 through R-Language

Table 3 Comparative analysis of data

Mean Mode Median

Location 1 10.65 9.4 11.15

Location 2 12.15 12.4 12.4

Location 3 8.79 7.3, 8.9 8.15

Table 3 shows that in descriptive statistics of all the location and result according to the data of wind velocity, location 2 is suitable rather than location 1 and location 3. Because the mean of location 2 is 12.15 m/s, which is greater than the mean of other location, so easily understand that 12.15 is suitable value of wind velocity for electricity generation. The mode of location 2 is 12.4 which is greater than mode of locations 1 and 3. So that central tendency is useful to find out the descriptive statistics of wind velocity and also useful to find out suitable location for establishment of wind power plant. After the descriptive assessment, now you can develop data visualization of wind velocity data through bar chart in Fig. 11. Now, it is necessary to identify the descriptive statistics of load demand of location 2 (Table 4), which is finalized to establish the wind energy system.

1892

V. Khare and C. J. Khare

Month-wise wind velocity for 3 location 20 17.3

18 16

15

14

12.4 12.6 11.3 11

12 10

16.5

15.8 13.3 11.2 10.8

14.3 12.9 11.6

14.6

13.9

12.4 11.1 9.48.9 8.3

9.48.9

9

7.3 6.9 6.4 5.1 6 4.5 4 8

18

7.4 5.8

7.3 6.2

7.4

Location 1 Location2

5.1

Location3

2 0

Fig. 11 Month-wise wind velocity for three locations Table 4 Month-wise average load capacity kW

Month January February March April May June July August September October November December

Load capacity (kw) 32 29 35 37 36 41 32 34 28 26 29 29

Mean of location 2 = (32 + 29 + 35 + 37 + 36 + 41 + 32 + 34 + 28+ 26 + 29 + 29) /12 = 388/12 = 32.34 Mode of location 2 = most frequent value of demand capacity = 29 (number of occurrence 3) Median of location 2 = (41 + 32) /2 = 36.5 Above descriptive statistics shows that mean of load demand is 32.34 KW and median is 36.5 KW. Value of mode for the load demand of location 2 for wind power plant is 29, and number of occurrence is 3, which shows that load demand is

74 Wind Energy System: Data Analysis and Operational Management

1893

Load Capacity (KW) 45 40 35 30 25 20 15 10 5 0

Fig. 12 Month-wise load capacity

45

41

40 35 35

37

Load Capacity (kw)

36

32

32 29

30

Wind Velocity (m/s)

34 28

29

29

26

25 20

17.3 15

15 10

9

11 6.9

12.4

18

16.5

13.3

12.4 8.3 5.8

5 0

Fig. 13 Comparative analysis of load capacity and wind velocity

saturated or steady state (Hussain 2018b). Now, you can develop data visualization of load demand and also comparative data visualization between wind velocity of location 2 and load demand of location 2 (Figs. 12 and 13).

1894

V. Khare and C. J. Khare

Percentile and Quartile (PQ) Assessment of Wind Energy System Data In the previous analysis, assessment of wind energy system data is evaluated through the “3M” concept, which means mean, mode, and median. Now, assessment is done through the “PQ” concept. PQ is the percentile and quartiles measurement of wind energy system data. Percentiles are measures of central tendency that divide a group of data into 100 parts. There are 97 percentiles because it takes 97 dividers to separate a group of data into 100 parts (Hussain 2019a). In the percentiles calculation, first arrange the wind energy system data into an ascending order, and then identify the percentile location (k) by  k=

 P n 100

where P is the percentile of interest, k is the percentile location, and n is the number in the data set. If k is a whole number, the Pth percentile is the average of the value at the kth and the value at the (k + 1)st location, and if k is not a whole number, the Pth percentile value is located at the whole number part of j + 1. Quartiles are measure of central tendency that divide a group of data into four subgroups or parts. The three quartiles are denoted as Q1, Q2, and Q3. The first quartile Q1 separates the first or lowest one fourth of the data from the upper three fourths and is equal to the 25th percentile. The second quartile Q2 is located at the 50th percentile and equals the median of the data. The third quartile is equal to the value of the 75th percentile. In the “PQ” assessment, you can determine 40th percentile of the following 12month average wind velocity data of location 2 (Table 5).

Table 5 Wind velocity data of location 2

Month January February March April May June July August September October November December

Location 2 (wind velocity) 9 6.9 11 12.4 15 13.3 17.3 18 5.8 12.4 16.5 8.3

74 Wind Energy System: Data Analysis and Operational Management

1895

First, organize the data into ascending order: 5.8, 6.9,  9, 11, 12.4, 12.4, 13.3, 15, 16.5, 17.3, 18  8.3, 40 12= 4.8 k = 100 Because k is not a whole number, the value of k + 1 is 4.8 + 1 = 5.8. The whole number of 5.8 is 6. The 40th percentile is located at the sixth value. The sixth value is 12.4, so 12.4 is the 40th percentile of wind velocity data set. In the quartile measurement, you want to determine the values of Q1, Q2, and Q3 for the above number set of 12th month average wind velocity of location 2 (Shtessel et al. 2014). The value of Q1is found  at the 25th percentile, P25 , by 25 For n = 12 k = 100 12 = 3 Because k is a whole number, P25 is found as the average of third and fourth value of wind velocity P25 = (8.3 + 9)/2 = 8.65 The value of Q1 is P25 = 8.65. The value of Q2 is equal to the median, n = 12, so median average of two middle values is Q2 = (12.4 + 12.4) /2 = 12.4 The value of Q3is found  at the 75th percentile, P75 , by 75 For n = 12 k = 100 12 = 9 Because k is a whole number, P25 is found as the average of ninth and tenth value of wind velocity P75 = (15 + 16.5)/2 = 15.75

Mean Absolute Deviation, Variance, and Standard Deviation (MVS)-Based Assessment of Wind Energy Data MVS is another measure of variability in terms of “M = mean absolute deviation,” “V = variance,” and “S = standard deviation.” The mean absolute deviation is the average of the absolute values of the deviations around the mean for a set of numbers. The variance is the average of the squared deviations about the arithmetic mean for a set of numbers. The population variance is denoted by σ 2 . The standard deviation is the square root of the variance (Guediri 2017). The population standard deviation is denoted by σ . Table 6 shows deviation from the mean for wind velocity data of location 2. Table 7 shows variance and standard deviation of wind velocity data. Now, the value of variance and standard deviation is  Variance = σ = 2

183.19 (x − μ)2 = = 15.26 n 12

1896

V. Khare and C. J. Khare

Table 6 Deviation from the mean for wind velocity data of location 2 Month January February March April May June July August September October November December

Wind velocity (x) 9 6.9 11 12.4 15 13.3 17.3 18 5.8 12.4 16.5 8.3  x=145.9 Mean μ = 145.9/12 = 12.15

Deviation from mean (x–μ) 9–12.15 = −3.15 6.9–12.15 = −5.25 11–12.15 = −1.15 12.4–12.15 = −0.25 15–12.15 = 2.85 13.3–12.15 = 1.15 17.3–12.15 = 5.15 18–12.15 = 5.85 5.8–12.15 = −6.35 12.4–12.15 = −0.25 16.5–12.15 = 4.35 8.3–12.15 = −3.85  (x − μ) = 0

Table 7 Variance and standard deviation of wind velocity data Month January February March April May June July August September October November December

Deviation from mean (x–μ) 9–12.15 = −3.15 6.9–12.15 = −5.25 11–12.15 = −1.15 12.4–12.15 = −0.25 15–12.15 = 2.85 13.3–12.15 = 1.15 17.3–12.15 = 5.15 18–12.15 = 5.85 5.8–12.15 = −6.35 12.4–12.15 = −0.25 16.5–12.15 = 4.35 8.3–12.15 = −3.85  (x − μ) = 0

Wind velocity (x) 9 6.9 11 12.4 15 13.3 17.3 18 5.8 12.4 16.5 8.3  x=145.9 Mean μ = 145.9/12 = 12.15

 Standard Deviation = σ =



(x − μ)2 9.9225 27.5625 1.3225 0.0625 8.1225 1.3225 26.5225 34.2225 40.3225 0.0625 18.9225 14.8225  (x − μ)2 = 183.19

√ (x − μ)2 = 15.26 = 3.906 n

The standard deviation is the separate entity and is a part of other analysis of wind energy system.

74 Wind Energy System: Data Analysis and Operational Management

1897

Frequency Distribution of Prefeasibility Data of Wind Energy System A first step in exploring and analyzing data is to reduce important and sometimes expensive data to a graphic picture that is clear, concise, and consistent with the message of the original data. Frequency distribution is a summary of data presented in the form of class interval and frequencies. Frequency distribution is the assessment of grouped and ungrouped data. Raw data, or data that have not been summarized in any way, are sometimes referred to as “ungrouped data.” Data that have been organized into a frequency distribution are called “grouped data.” The distinction between ungrouped and grouped data is important because the calculation of statistics differs between the two types of data (Naidu 2020). Table 8 shows the raw data of wind velocity of 364 days for a particular location. The table shows the frequency distribution of wind velocity data of 364 days. In this analysis, whole data of 364 days is categorized into number of intervals, and minimum and maximum value of class interval is 5 and 19, respectively. Frequency of class interval shows the number of occurrences in between the given class interval. The table shows the class midpoints, relative frequency, and cumulative frequency of given data sets of wind velocity. The midpoint of the class interval is called the class midpoint and is some-times referred to as the class mark. Relative frequency is the proportion of the total frequency that is in any given class interval in a frequency distribution (Mann 2018). The cumulative frequency is a running total of frequencies through the classes of a frequency distribution (Tables 9 and 10). In the next section, we assess and find out central tendency of grouped data to find out the key insights of wind energy data measurement.

Measurement of Central Tendency and Variability of Wind Energy Grouped Data Grouped data do not provide information about individual values. Hence, measure of central tendency and variability for grouped data must be computed differently from those for ungrouped or raw data. Mean: The mean for grouped data is then computed by summing the products of the class midpoint and the class frequency for each class and dividing that sum by the total number of frequencies. The formula for the mean of grouped data follows.  Mean of grouped data : μGrouped =

F i Mi = n



F i Mi Fi.

where Fi is the class frequency, Mi is the class midpoint, and n is the total frequency. Median: The median for ungrouped or raw data is the middle value of an ordered array of numbers. For grouped data, solving for the median is considerably more complicated. The calculation of the median for grouped data is done by using the following formula:

5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 6 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 14.2

7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 14.3

8.9 9 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 10 10.1 10.2 10.3 10.4 10.5 10.6 10.7 14.4

10.8 10.9 11 11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8 11.9 12 12.1 12.2 12.3 12.4 12.5 12.6

12.7 12.8 12.9 13 13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8 13.9 14 14.1 14.2 14.3 14.4 14.5

14.6 14.7 14.8 14.9 15 15.1 15.2 15.3 15.4 15.5 15.6 15.7 15.8 15.9 16 16.1 16.2 16.3 16.4

16.5 16.6 16.7 16.8 16.9 17 17.1 17.2 17.3 17.4 17.5 17.6 17.7 17.8 17.9 18 18.1 18.2 18.3

18.4 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8

8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 9 9.1 9.2 10.5 10.6 10.7 10.8 10.9 11 11.1

Table 8 Wind velocity (m/s) data of 364 days for a particular location 11.2 11.3 11.4 11.5 11.6 11.7 11.8 11.9 12 12.1 12.2 12.3 12.4 12.5 12.6 12.7 12.8 12.9 13

13.1 13.2 14.9 15 15.1 15.2 15.3 15.4 15.5 15.6 15.7 15.8 15.9 16 16.1 16.2 16.3 16.4 16.5

16.6 16.7 16.8 16.9 17 17.1 17.2 17.3 17.4 17.5 17.6 17.7 6.4 6.5 6.6 6.7 6.8 6.9 7

7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9

9 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 10 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8

10.9 11 11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8 11.9 12 12.1 12.2 12.3 12.4 12.5 12.6 12.7

12.8 12.9 13 13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8 13.9 14 14.1 14.2 14.3 14.4 14.5 14.6

14.7 14.8 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 6 6.1 6.2 6.3 6.4 6.5 6.6 6.7

6.8 6.9 7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8 8.1 8.2 11.9 12 12.1 12.2

12.3 12.4 12.5 12.6 12.7 12.8 12.9 13 13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8 13.9 14 14.1

1898 V. Khare and C. J. Khare

74 Wind Energy System: Data Analysis and Operational Management Table 9 Frequency distribution of wind velocity (m/s) data of 364 days for a particular location

Class interval 5 under 7 7 under 9 9 under 11 11 under 13 13 under 15 15 under 17 17 under 19

1899 Frequency 51 73 48 71 58 40 23

Table 10 Class midpoints, relative frequencies, and cumulative frequency Class interval 5 under 7 7 under 9 9 under 11 11 under 13 13 under 15 15 under 17 17 under 19 Total

Frequency 51 73 48 71 58 40 23 364

Class midpoints 6 8 10 12 14 16 18

Relative frequencies 0.14011 0.200549 0.131868 0.195055 0.159341 0.10989 0.063187

 Median of the grouped data = Median =

l+

n 2

Cumulative frequencies 51 124 172 243 301 341 364

− CF P FMED

 D

where l is the lower limit of the median class interval, CFP are a cumulative total of the frequencies up to but not including frequency of the median class, FMED is the frequency of the median class, and D is the width of the median class interval. Mode: The mode for grouped data is the class midpoint of the model class. The modal class is the class interval with the greatest frequency. Measures of Variability: The measures of variability for grouped data are presented here: the variance and standard deviation. Again, the standard deviation is the square root of the variance. “Original Formulas” for Population Variance and Standard Deviation of Grouped Data:  σ = 2

Fi (Mi − μ)2 n  σ = σ2

Formula for “computational version” is  σ2 =

Fi Mi2 − ( n



Fi Mi ) n

2

1900

V. Khare and C. J. Khare

Table 11 Calculation of grouped mean Class interval 5 under 7 7 under 9 9 under 11 11 under 13 13 under 15 15 under 17 17 under 19 Total μGrouped =



Class midpoints (Mi ) 6 8 10 12 14 16 18

Frequency (Fi ) 51 73 48 71 58 40 23 364 Fi Mi n

=

4088 364

Cumulative frequencies 51 124 172 243 301 341 364

Fi Mi 306 584 480 852 812 640 414  Fi Mi = 4088

= 11.23

Mean = 11.23

where Fi is the frequency, Mi is the class midpoint, n is the total frequencies of the population, and μ is the grouped mean for the population. Refer to the table wind velocity (m/s) data of 364 days for a particular location and frequency distribution of wind velocity (m/s) data of 364 days for a particular location. Table 11 shows calculation of grouped mean. The first step in calculating a grouped median is to determine the value of n/2, which is the location of the median term. Since there are 364 values (n), the value of n/2 is 364/2 = 182. The median is 182th term. The question to ask is, Where does the 182th term fall? This can be answered by determining the cumulative frequencies for the data, as shown in Table. An examination of these cumulative frequencies reveals that the 182th term falls in the fourth-class interval because there are only 172 values in first three class intervals. Thus, the median value is in the third-class interval between 11 and 13. The class interval containing the median value is referred to as the median class interval. Since the 182th value is between 11 and 13, the value of median must be at least 11. How much more than 11 is the median? The difference between the location of the median value, n/2 = 182, and the cumulative frequencies up to but not including the median class interval, 172, and the frequency of median class is 71 (Wawrzinek 2020). Width of the class interval is 2. Now, we put the value in the formula of median: Median =

11+

364 2 −172

71



 2 = 11+

   10 182 − 172 2 = 11 + 2 = 11.28 71 71

Median = 11.28 The mode for grouped data is the class midpoint of the modal class. The modal class is the class interval with the greatest frequency. Modal class is 11 under 13,

74 Wind Energy System: Data Analysis and Operational Management

1901

Table 12 Calculating grouped variance and standard deviation with the original formula Class interval 5 under 7 7 under 9 9 under 11 11 under 13 13 under 15 15 under 17 17 under 19 Total

Frequency (Fi ) 51 73 48 71 58 40 23 364 

Class midpoints (Mi ) 6 8 10 12 14 16 18

Fi Mi 306 584 480 852 812 640 414  Fi Mi = 4088

(Mi –μ) μ = 11.23 6–12 = −6 8–12 = −4 10–12 = −2 12–12 = 0 14–12 = 2 16–12 = 4 18–12 = 6

(Mi –μ)2 36 16 04 00 04 16 36  (Mi − μ)2 = 112

Fi (Mi –μ)2 1836 1168 192 0 232 640 828  Fi(Mi – μ)2 = 4896

F (M −μ)2

4896 i i σ 2 = Variance = n √ = √ 364 = 13.45 Standard deviation = σ = σ2 = 13.45 = 3.66

Table 13 Calculating grouped variance and standard deviation with the computational formula Class interval 5 under 7 7 under 9 9 under 11 11 under 13 13 under 15 15 under 17 17 under 19 Total 

F M2−

Frequency (Fi ) 51 73 48 71 58 40 23 364 (



F i Mi ) 2

Class midpoints (Mi ) 6 8 10 12 14 16 18

50592− (4088)

Fi Mi 306 584 480 852 812 640 414  Fi Mi = 4088

Fi Mi 2 1836 4672 4800 10,224 11,368 10,240 7452  F i M 2i = 50, 592

2

i i n 364 σ2 = = = 50592−45911 = 13.45 n 364 √ 364√ 2 Standard deviation = σ = σ = 13.45 = 3.66

and then class midpoint is 12. Mode = 12 Measures of Variability: “Original Formulas” for Population Variance and Standard Deviation of Grouped Data (Tables 12 and 13).

Regression Analysis of Wind Energy System Data Regression analysis is a process which identifies the relationship between the two parameters, and it is categorized into two-part linear regression analysis and multiregression analysis. In linear regression analysis, one is dependent parameter, and

1902

V. Khare and C. J. Khare

Table 14 Data of solar radiation and wind velocity Month 1 2 3 4 5 6 7 8 9 10 11 12

Solar radiation (x) (kWh/m2 /day) 6.3 7.1 8.4 8.6 8.8 9.0 7.6 7.1 6.5 6.6 6.8 6.4

Wind velocity (y) (m/s) 9 6.9 11 12.4 15 13.3 17.3 18 5.8 12.4 16.5 8.3

one is independent parameter. The equation of regression analysis is given by y = b0 + b1 x, where y is the dependent variable, x is the independent variable, b0 is the intercept of regression analysis, and b1 is the slope of the regression analysis (Padhya 2020). In this analysis, find out the relationship between wind velocity and solar radiation, where wind velocity is the dependent parameter, which depends on the amount of solar radiation, or in the other way, solar radiation is the independent parameter of this linear regression analysis (Table 14). Slope of the Regression Analysis: 

 b1 =



x y) xy − ( )( n  2 ( x )2 x − n

Intercept of Regression Line:  b0 =

n

y

 − b1

x

n

Correlation is a measure of the degree of relatedness of variables (Table 15). The term “r” is a measure of linear correlation of the two variables. In this analysis, identify the correlation between wind velocity and solar radiation, and the value of r is 0.3419, so the + value denotes the perfect positive relationship between these two parameters (Dupuis et al. 2020).

74 Wind Energy System: Data Analysis and Operational Management

1903

Table 15 Computation of correlation “R” Solar radiation (x) Month (kWh/m2 /day) 1 6.3 2 7.1 3 8.4 4 8.6 5 8.8 6 9.0 7 7.6 8 7.1 9 6.5 10 6.6 11 6.8 12 6.4  X = 89.2

  x y n    2 ( x )2 x − n



r=

Wind velocity (y) (m/s) 9 6.9 11 12.4 15 13.3 17.3 18 5.8 12.4 16.5 8.3  Y=145.9

xy −

1 × 1957.09 −

b1 =

1100.57 −

×

145.92 12

1 

y2 − (



y) n

2

y2 81 47.61 121 153.76 225 176.89 299.29 324 33.64 153.76 272.25 68.89  2 Y = 1957.09

=

xy 56.7 48.99 92.4 106.64 132 119.7 131.48 127.8 37.7 81.84 112.2 53.12  XY = 1100.57

1100.57 − 89.2×145.9 12 2 674.24 − 89.2 12

= 0.3419

89.2×145.9 12 2 674.24 − (89.2) 12

b0 =

x2 39.69 50.41 70.56 73.96 77.44 81 57.76 50.41 42.25 43.56 46.24 40.96  2 X = 674.24

=

16.05 1100.57 − 1084.52 = = 1.43 674.24 − 663.05 11.19

89.2 145.9 − 1.43 = 12.15 − 10.62 = 1.53 12 12

y = 1.53 + 1.43x

Wind velocity = 1.53 + 1.43 (solar radiation)

1904 Table 16 Wind energy system data by wind velocity and solar radiation

V. Khare and C. J. Khare Days 1 2 3 4 5 6 7 8 9 10 11 12 . . . 365

Solar radiation (kWh/m2 /day) 6.3 7.1 8.4 8.6 8.8 9.0 7.6 7.2 6.5 6.6 6.8 6.4 . . . 0.401

Wind velocity (m/s) 9 6.9 11 12.6 15 13.3 17.3 18 5.8 12.4 16.5 8.3 . . . 6.3

Cross Tabulation and Scatterplot of Wind Velocity Data It is very common in descriptive statistics to want to analyze two variables simultaneously in an effort to gain insight into a possible relationship between them. Two of the more elementary tools for observing the relationship between two variables are cross tabulation and scatterplot. Cross tabulation is a process for developing a two-dimensional matrix that displays the frequency counts for two variables simultaneously (Beltran et al. 2012). As an example of prefeasibility analysis of wind energy system, gather a sample of 365 days’ data of solar radiation (kwh/m2 /day) and wind velocity (Table 16). By tallying the frequencies of responses for each combination of categories between solar radiation and wind velocity, data are cross tabulated according to the two variables (Table 17). A scatterplot is a two-dimensional graph plot of pairs of points from two numerical variables. The scatterplot is a graphical tool that is often used to examine possible relationship between two variables. Figure 14 shows the scatterplot between wind velocity and solar radiation.

Operational Management of Wind Energy System An operational management of wind energy system is defined as the assessment of several activities such as generation, transmission, and distribution of electricity through the wind power plant. In wind energy system, operations management is a systematic approach to understanding the nature of issues and problems related to electricity generation. In systematic approach, we studied the data of

74 Wind Energy System: Data Analysis and Operational Management

1905

Table 17 Cross tabulation table of wind energy data Wind velocity (m/s) Solar radiation 6.3 6.4 6.5 6.6 6.8 7.1 7.2 7.6 8.4 8.6 8.8 9

5.8 4 5 3 2 5 4 4 3 2 3 3 3 41

6.9 3 5 2 4 5 3 4 2 2 3 3 3 39

8.3 6 4 1 3 4 5 4 3 2 4 4 4 44

9 2 3 1 5 3 2 3 1 1 1 2 1 25

11 3 6 2 2 6 3 5 1 1 1 5 1 36

12.4 5 1 4 3 1 2 1 2 2 2 1 2 26

12.6 1 3 3 2 3 1 3 1 1 1 3 1 23

13.3 1 4 5 1 4 2 3 2 2 2 3 2 31

15 3 2 2 2 2 2 2 2 2 2 2 2 25

16.5 3 3 3 1 3 3 3 3 3 3 3 3 34

17.3 3 1 2 1 2 1 1 1 1 1 1 2 17

18 1 3 1 1 3 1 3 3 2 2 3 1 24

35 40 29 27 41 29 36 23 21 26 33 25 365

Solar Radiation V/S Wind Velocity 20 18 16 14 12 10 8 6 4 2 0 0

1

2

3

4

5

6

7

8

9

10

Fig. 14 Scatterplot of wind velocity and solar radiation

the prefeasibility analysis, modelling, and controlling of wind energy system. For successful operations management, the focus should be on developing a set of tools and techniques to analyze the problem faced within an operations system of wind energy system. Operations management involves addressing various issues of wind energy system. Simple problem includes deciding how to provide supply and find alternative when turbine or generator is not capable to convert the energy from one form to another. On the other hand, decisions include such as where to

1906

V. Khare and C. J. Khare

locate the wind power plant, what capacity to build in the system, and in which manner to distribute the energy to the electricity consumer. Operational management provides alternative methodologies to address such wide-ranging issues in an organization. Transformation and conversion processes are central to operations system and ensure that wind velocity as an input is converted into useful output in terms of electricity generation. Therefore, the focus of operations management is to address the design, planning, and operational control of the transformation process. The operations management also involves the development of performance evaluation systems and framework through which the operating system can make improvements to meet targeted performance measures (Evangelista et al. 2017). In wind energy system, location decision pertains to the choice of appropriate geographical sites for locating variations manufacturing and/or service facilities of an organization. Locating facilities in regions that offer attractive cost advantages is one aspect of the trade-off. It may be expensive to set up a distribution system that is both efficient and responsive to an increase in transportation costs and the costs of coordinating and communicating with the supplier about products and availability. Location planning essentially addresses different issues and provides the operations manager the needed tools and techniques to study the location problem. Location issues have been more prominent in recent years due to the increased pace of economic reforms in several countries and the consequent globalization of markets. Figure 15 shows affecting parameters of location decision for an organization. The most significant factor that drives globalization is the ongoing financial and regulatory reforms in several developed and developing countries. Beginning

Financial Reforms

Saturation of Financial Development

High Investment Cost

How Location choice Affect on Organization

Regulatory and Policy Issue

Market Condition

Multi-Location Environment

Fig. 15 Affecting parameters for location decision on organization

74 Wind Energy System: Data Analysis and Operational Management

1907

from 1991, we embarked on a set of regulatory changes that made India much more attractive in terms of locating a manufacturing facility. Two events have been broadly responsible for this. First is the reduction in customs and excise tariffs and a move forward the single-point value-added tax (VAT) regime. The other is the delicensing of several sectors of industry and the progressive removal of the cap on foreign direct investment. Another relevant point for location planning with respect to regulatory issue is the emergence of regional trading blocs (Hussain 2019b). Location Factor Rating: Factor rating is a simple methodology to assess the attractiveness of each potential location. This method involves four steps in which the relevant parameters are identified, their relative importance established, the performance of each location in each factor assessed, and finally, all this information combined to rank the locations. The four-step process of identifying an appropriate location is as follows: (a) Identify and list all the relevant parameters for the location decision for wind energy system. (b) Establish the relative importance of each parameter in the final decision for location of wind energy system. (c) Rate the performance of each candidate location using a rating mechanism. One of the wind energy company is actively considering five alternative locations to install the wind power plant. Based on the survey related to wind power plant, company has arrived at eight factors to be considered for final site selection for wind power plant. The ratings (Table 18) of each factor on a scale of 1–100 provide this information of location of wind energy system. The first step in the solution is to develop the relative importance of each parameters using a normalization factor. The sum of all the parameters related to wind energy system is 325. Therefore, by dividing each parameter rating by 415, one can obtain the relative weight of the parameters (Tables 19 and 20). Since the normalized weight of each parameter, one can compute how each location fares by weighing the rating of the location against each factor with the weight for the factor. The computation for location 1 is as follows:

Table 18 Factor rating Factors Amount of wind velocity Load demand in KW Monsoon circulation Technical potential Government policy toward wind energy at a particular location Subsidy policy toward wind energy at a particular location

Rating (1–100) 80 75 70 60 65 65

1908

V. Khare and C. J. Khare

Table 19 Factor rating Factors Amount of wind velocity Load demand in KW Monsoon circulation Technical potential Government policy toward wind energy at a particular location Subsidy policy toward wind energy at a particular location Sum of all factor rating

Rating (1–100) 80 75 70 60 65

Relative weight 0.192771084 0.180722892 0.168674699 0.144578313 0.156626506

65

0.156626506

415

1

Table 20 Rating of each location of wind energy system against each factor Factors Amount of wind velocity Load demand in KW Monsoon circulation Technical potential Government policy toward wind energy at a particular location Subsidy policy toward wind energy at a particular location Sum of all factor rating Ranking of the locations

Relative Location weight 1 0.192771084 45

Location 2 60

Location 3 75

Location 4 62

Location 5 42

0.180722892 0.168674699 0.144578313 0.156626506

30 35 30 40

58 41 45 40

70 63 55 60

56 45 25 42

55 47 51 50

0.156626506 35

30

55

35

45

36.084

46.433

63.698

45.337

48.216

5

3

1

4

2

(45 × 0.192771084) + (30 × 0.180722892) + (35 × 0.168674699) + (30 × 0.144578313) + (40 × 0.156626506) + (35 × 0.156626506) = 36.084 Similarly, factor rating of locations 2, 3, 4, and 5 is 46.433, 63.698, 45.337, and 48.216, respectively. So according to the above assessment, location 3 is a more feasible location for wind energy system.

Operational Management of Flexible Generation System In the operational management of wind energy system, flexible generation system usually consists of dynamic component control system as a wind farm controller (Fig. 16). The purpose of the control system of a wind turbine is to manage the safe,

74 Wind Energy System: Data Analysis and Operational Management

ELECTRICAL ENERGY

SINGLE

WIND FARM

1909 HYBRID

MODERATE COMPLEXITY

HIGH COMPLEXITY

LOW COMPLEXITY

MODERATE COMPLEXITY

NO OF ENERGY CONVERSION

MECHANICAL ENERGY AC LOAD

LOAD

DC LOAD

Fig. 16 Flexible generation system of wind energy system

automatic operation of the turbine. This reduces operating costs, provides consistent dynamic response and improved product quality, and helps to ensure safety. This operation is usually designed to maximize annual energy capture from the wind while minimizing turbine loads. In flexible generation system, wind turbine control systems are categorized into the three separate forms: 1. A controller that controls numerous wind turbines in a wind farm 2. A supervisory controller for each individual turbine 3. Separate dynamic controllers for the various turbine subsystems in each turbine The flexible wind energy generation systems offering several flexibilities are as follows: • Machine Flexibility: This is related to the ease of making changes required to generate an electricity from the wind generator or store the energy in the number of batteries. Technological process, proper operational management, and the technological capability of blade, hub, wind turbine, and wind generator along with the different technical factors will improve this flexibility. • Process Flexibility: For effective operational management of wind energy system, it is necessary to develop process flexibility in the generation of electricity. In the process flexibility, it should be proper functioning of rotor shaft and generator, and use ideal switchgear system to transfer the energy from generator to the power transmission tower. • Routing Flexibility: This is the ability of wind power plant to handle breakdown of supply and continue the generation of electricity through another number of wind turbine. This ability exists if either a part of electricity generation can be processed by several transmission routes or each operation of wind farm is performed by more than one wind generator. • Expansion Flexibility: It is the capability to build a wind power plant of fixed capacity, and in the future, it can be expanded according to the load demand

1910

V. Khare and C. J. Khare

because load demand is variable according to the consumer load (Hussain 2019c). There is much complexity in operational management of wind energy system due to the number of stages or number of energy conversion, types of load, and types of wind power plant. The complexity of operational management of wind energy system is described through the low, moderate, and high complexity. Complexity is mainly depending on the types of wind farm, which may be single or hybrid wind energy system, number of stages of energy conversion, and types of load, which may be AC or DC load. In the low complexity, the consumers demand an AC load, and during the process, turbine converts kinetic energy into the mechanical energy. During the electrical energy conversion, moderate complexity is occurring because it is necessary to assess which type of generator will be better for the conversion of mechanical to electrical energy. According to the load, in most of the cases, it may be AC load so in that case, induction or synchronous generator is much better compared to the DC generator. If the load is DC, some complexity will be occurring because in the case of induction or synchronous generator, it is necessary to convert AC into DC through the suitable converter. High complexity is occurring when hybrid wind energy system is connected at the particular location because hybrid wind energy system is just like a jumbled process, and the overall process depends on which type of auxiliary system is used with wind energy system. Some conventional system is connected with wind energy system, but in the new era of electricity generation, solar wind hybrid renewable energy system is the most prominent source of electricity generation. From the operational management point of view, solar wind hybrid renewable energy system is a much complex system because setup of solar power plant is very much different from the wind energy system. The operations function and its linkages with wind energy system are shown in different forms such as consumer layer, operational layer, technical layer, and core operational layer (Fig. 17). In every manufacturing or service organization, there are several subfunctions within the core operations functions. For proper operational management of wind energy system, it is necessary to perfect coordination between the different layers. Consumer layer is depending on the load demand, types of load which may be AC or DC, and transmission and distribution of electricity through the wind energy system. The capacity of wind energy system always depends on the proper functioning of consumer layer. Operational layer is related to the operation and maintenance of wind energy system, which is the combination of different layers such as unit sizing and accurate modelling of wind energy system. Unit sizing of wind energy system is the part of optimum sizing of wind turbine, generator, and converter and their working according to the overall capacity of wind power plant. Nowadays, the importance of information technology in the field of wind energy system increases in the tremendous way because lots of software such as supervisory control and data acquisition system (SCADA), RET Screen, and HOMER are used for assessment of prefeasibility, modelling, and controlling of wind energy system (Hussain 2020a, b; Wu et al. 2008).

74 Wind Energy System: Data Analysis and Operational Management

1911

OPERATIONAL LAYER CONSUMER LAYER LOAD DEMAND TRANSMISSION OF ELECTRICITY DISTRIBUTION OF ELECTRICITY

MAINTENANCE OF WIND POWER PLANT QUALITY OF ENERGY SUPPLY OPTIMUM DESIGN OF WIND ENERGY SYSTEM WIND ENERGY SYSTEM WITH INFORMATION TECHNOLOGY

TECHNICAL LAYER CORE OPERATIONAL LAYER INNOVATION STRATEGY RESEARCH AND DEVELOPMENT OF WIND ENERGY SYSTEM

TESTING ASSEMBLY FABRICATION MACHINING & SERVICE DELIVERY SYSTEM OF WIND ENERGY SYSTEM

Fig. 17 Operational layer of wind energy system

For effective functioning of wind energy system throughout the life span, it is necessary for innovation and research and development in the field of wind energy system. It is also necessary to apply artificial intelligence and machine learning in the field of wind energy system to enhance the performance of overall wind power plant. Technical layer of operational management of wind energy system is directly linked with the core operational layer of wind energy system, which consists starting final stage of wind power plant. Manufacturing of turbine blade, hub, generator, and different types of converter and their fabrication and testing are the main component of this layer (Hu et al. 2010; Javadi et al. 2018; Hussain 2020c).

Operational Management-Based Maintainability and Availability Function of Wind Energy System Maintainability In operational management of wind energy system, if any component and elements such as wind turbine generator are repaired in a specific period of time, then that is known as maintainability of wind power plant. Maintainability is a most important

1912

V. Khare and C. J. Khare

part of operational management. For example, in electricity generation through wind power plant, it is necessary to maintain transmission and distribution system in a proper way. Then it transmits large amounts of electrical power. Maintainability is a very specific property of modelling of a wind energy system framework. It is always related to the concept of economy, replacement, and safety measures, which is very helpful to run any physical system. Maintainability of any system is classified into following categories: • • • • •

Wind energy system component activity Wind energy system component repair time Wind energy system repair time Wind energy system downtime Wind energy system verification time

Component Activity: A specific task performed by individual component of wind energy system of any specific physical arrangement is known as component activity. For example, wind turbine is used to convert kinetic energy into mechanical energy, and generator is used to convert mechanical energy into electrical energy. Maintainability of individual component directly affects the maintainability of whole system because overall wind power plant efficiency is depending on the performance of individual component. The component activity is also related to what manner individual components are connected. If wind turbine and generator are connected in series, then performance of the overall system is affected at large level through performance of individual components. If possible, it is necessary to connect a large number of components in parallel to maintain the framework of the complete system in a proper way (Hussain 2020d, e). Component Repair Time: A time required to minimize the failure of different elements of wind energy system is known as component repair time. Wind energy component repair time is directly proportional to the maintainability of the overall system. System Repair Time: A time required to minimize the failure of the overall wind power plant is known as system repair time. System repair time is directly proportional to the maintainability of the overall system. System Downtime: It is a time required for the wind power plant when it reaches from ON condition to OFF condition (Kenne et al. 2015; Lei et al. 2006). Availability: It deals with the duration of uptime for operations of wind energy system and is a measure of how often the system is alive and well. It is often expressed as Availability =

Availability =

Uptime Uptime + Downtime

MT BF MT BF + MT R + MT W S

74 Wind Energy System: Data Analysis and Operational Management Table 21 Comparison between three parameters for wind energy system

Operational management Constant Constant Decreases Decreases

Maintainability Decreases Increases Constant Constant

1913 Availability Decreases Increases Increases Decreases

MTBF = mean time between failure MTR = mean time to repair MTWS = mean time to waiting for spares, reflecting supply In operational management, one is worried about outlining a thing to keep going to the extent that this would be possible without disappointment; in maintainability, the accentuation is on planning a thing so a disappointment can be redressed as fast as would be prudent (Table 21). It is a component of the gear, plan and the establishment, workforce accessibility in the required expertise levels, sufficiency of support techniques and test hardware, and the physical condition under which upkeep is performed. Similarly, as with operational management, practicality parameters are additionally probabilistic and are examined by the utilization of consistent and discrete arbitrary factors, probabilistic parameters, and measurable circulations. We consider performance of any component so that concepts of maintainability in terms of functions are analogous to the terms of operational management. They may be derived in a way identical to that done for operational management by merely substituting t (time-to-restore) for t (time-to-failure), μ (repair rate) for λ (failure rate), and M(t) probability of successfully completing a repair action in time t, or P (T ≤ t) for F(t) probability of failing by age t (Hussain 2020f, 2021). In other words, the following correspondences prevail in maintainability and operational management of wind energy system. 1. The time-to-failure probability density function in operational management corresponds to the time-to-maintain power density function in maintainability. 2. The failure rate function in operational management corresponds to the repair rate function in maintainability. Repair rate is the rate with which a repair action is performed and is expressed in terms of the number of repair actions performed and successfully completed per hour. 3. The probability of system failure, or system operational management, corresponds to the operational management of successful system maintenance or system maintainability. Figure 18 shows probability of repair of wind energy system. These and other analogous functions are summarized in the following: There are three categories of availability:

1914

V. Khare and C. J. Khare

Fig. 18 Probability of repair of wind energy system Pr Probability of Repair

Time to Repair

T

1. Inherent availability: AI =

MT BF MT BF + MT T R

AA =

MT MA MT MA + MMT

AO =

MT MA MT MA + MDT

2. Achieved availability:

3. Operational availability:

where MTBF = mean time between failure, MTTR = mean time to repair Frequency of Failure: This is the number of failure of wind power plant that occurs divided by the total elapsed calendar time during which those events occur or by the total number of demands, as applicable. The frequency of failure is directly related to the wind turbine and wind generator failure and component failure mode. Failure Density: This is the ratio of the number of failures during a given unit interval of time to the total number of components of the primary stage of the test. Repairable System: In order to assess repairable system, we already discussed above the function of availability, where availability is the operational management that the particular system, component, and element are operating and working satisfactorily at time “t.” On the other hand, operational management is the probability that the given system has operated satisfactorily over the time interval of 0 to t. Repairable system always increases the system life span and performance of the given system. By and large, upkeep is characterized as any activity that reestablishes fizzled units to an operational condition or holds a non-fizzled unit in an operational state of wind power plant. For repairable frameworks, support assumes a fundamental part in the life of a framework. It influences the framework’s general dependability, accessibility, downtime, cost of the task, and so forth. A

74 Wind Energy System: Data Analysis and Operational Management

1915

repairable system is a system that can be restored to an operating condition following a failure. Questions of interest in repairable systems analysis include: • How many failures will occur over a fixed time interval in wind energy system? • What is the probability of a failure in the next time interval of wind energy system? • What is the availability of the system of wind energy system? • How many spare parts should be purchased of wind energy system? • What is the cost of maintaining the wind energy system? • What is the optimum overhaul time of wind energy system? Mean Time to Repair : Mean time to repair (MTTR) is a basic measure of the maintainability of repairable items. It represents the average time required to repair a failed component or device of wind energy system. Expressed mathematically, it is the total corrective maintenance time for failures divided by the total number of corrective maintenance actions for failures during a given period of time. It generally does not include lead time for parts not readily available or other Administrative or Logistic Downtime (ALDT). The equation for MTTR will be obtained in two ways. Mean time to repair (MTTR) is the average time required to troubleshoot and repair failed equipment and return it to normal operating conditions. It is a basic technical measure of the maintainability of equipment and repairable parts of wind energy system. Maintenance time is defined as the time between the start of the incident and the moment the system is returned to production. This includes notification time, diagnostic time, fix time, wait time, reassembly, alignment, calibration, test time, back to production, etc. It generally does not take into account lead time for parts. Mean time to repair ultimately reflects how well an organization can respond to a problem and repair it. Expressed mathematically, the MTTR calculation is the total maintenance time divided by the total number of maintenance actions over a specific period. MT T R = (Total maintenance time) / (no.of repairs) Let N be the items undergoing repairs simultaneous under identical conditions. Let t be the time intervals when we note down the number of units that have been repaired. Let n1 be the number of units repaired during the first t interval and nk the number of units repaired during the k-th t interval. Let the last t interval when all the N units have been repaired be the l-th t interval. n1 + n2 + . . . nk + . . . nl =

l i=1

ni = N

Let kt = t. Then the number of units repaired within time t is n1 + n2 + . . . nk =

k i=1

ni

1916

V. Khare and C. J. Khare

 2 +...nk Hence, the maintainability Ma (t = kt) = n1 +nN = N1 ki=1 ni . The number of items repaired during the k-th t interval was nk . Its value is obviously nk = N [Ma (t = kt ) − M (t = kt − t)]

MT T R =

1 l 1 nk (kt) [n1 t + n2 2t + . . . nk kt + · · · + nl lt ] = i=1 N N

Practical Model of Repairable System : When we develop wind energy system and after the failure of those systems, there are two possibilities. They are repaired and may be replaced, but some complex systems such as wind turbine, wind generator, inverter, transmission and distribution systems, etc., are repaired and are not replaced when they fail. The operational management of repairable system is different from a reliability of actual systems. The reliability of repairable system depends on the number of failures during the warranty period, which is mentioned by the shopkeeper, the total lifetime of the product, and the replacement cost of the product. The most powerful theory to determine the reliability of repairable system is a power law model. The power law model is easy to understand, and it is used to find out the frequency of failures. In statistics, a power law is a functional relationship between two quantities, where a relative change in one quantity results in a proportional relative change in the other quantity, independent of the initial size of those quantities: one quantity varies as a power of another. If power law is applied in reliability assessment, then performance of individual component is directly related to the performance of the overall system. The power law model is often used to analyze the reliability for complex repairable systems in the field. A system of interest may be the total system such as wind energy system, or it may be subsystem, such as rotor blade of wind energy system, which is also complex repairable system. The chances of repairing of any system depend on the failure rate, where if the failure rate is increasing, then this is indicative of component wear out. If the failure rate is decreasing, then this is indicative of infant mortality. If the failure rate is constant, then the component failures follow an exponential distribution.

Confidence Level of Repairable System On account of repairable framework, it is important to discover at what level shortcoming is limited, and the framework is fixed at the ideal level. The evaluation of a repairable framework discovers by the idea of certainty level, the certainty level is the likelihood of a framework at standard boundary, and those boundary lies among lower and maximum cutoff points. The certainty level of a repairable framework additionally relies on the ideal yield and reference yield. The maximum furthest reaches of the span are known as certainty cutoff of the general framework.

74 Wind Energy System: Data Analysis and Operational Management

1917

The certainty level of the framework is done in two habits: confidence level of individual component of wind energy system and confidence level of overall system of wind energy system There are three possible cases are developed to a confidence level of individual and overall systems. 1. The confidence level is greater than actual operational management. 2. The confidence level is equal to actual operational management. 3. The confidence level is less than actual operational management. In the event that the certainty level is more prominent than genuine operational administration, at that point, flawed framework is fixed more than the ideal level, and it expands the lifetime of the general framework. On the off chance that the certainty level is not exactly genuine unwavering quality, at that point, it is important to supplant the individual segment or generally arrangement of wind energy framework with another one. The ideal condition is discovered when the certainty level is equivalent to the real operational administration. Figure 19 shows confidence of reliability of wind energy system. The probability of repairable system is a ratio of the number of times successful maintenance (MS ) is to be done to repair the system to total number of maintenance (MT ) is to be done. Probability of repairable system (RP ) =

MS (t) MT (t)

Rewriting the numerator yields: RP =

MF (t) MT (t) − MF (t) =1− = 1 − Q(t) MT (t) MT (t)

where MF (t) = number of missions of “t” duration each than unsuccessful maintenance and Q(t) = probability of unsuccessful maintenance.

Fig. 19 Confidence of reliability

Output Confidence Level is greater than actual reliability Confidence Level = Actual Reliability Confidence Level is less than actual reliability

Input

1918

V. Khare and C. J. Khare

This applies when one unit undertakes MT (t) missions out of which Ms (t) mission succeed, and when the unit fails, it is restored to the same condition it was started on the first mission. If all maintenance of repairable system is failing, then Ms (t) = 0. If Ms (t) missions are completed successfully, the definition of probability of repairable system as (M _ S(t))/(M _ T(t)) is again conditional to the fact that all mission objectives and design for performance objectives are attained. Their design adequacy is 100%. A better way of seeing the complete picture of repairable system or system effectiveness is defined by Effectiveness of system = O × R × D =

M AC MT

where O is the operational management that the actual system is either available at the beginning of the project, R is the operational management that has completed the successful maintenance without failure, and D is the operational management after the maintenance system is performed successfully, and this is called design adequacy.

Future Scope It is at the center of a growing ecosystem of data analysis technologies that are primarily used to support advanced analytics initiatives, including predictive analytics, data mining, and machine learning applications. Following are the possibilities through data analysis in the field of wind energy system: • Create Hadoop Distributed File System (HDFS) for prefeasibility analysis, modelling system, and control system data which is also used for predictive analysis of wind power plants. • Create YARN-based framework for sustainable force plant, which incredibly extended the applications that Hadoop groups can deal with to incorporate stream handling and constant examination applications in wind velocity-based wind energy system. • Create a basket model of technical and managerial parameters of wind power plant.

Conclusion This chapter is the complete assessment of wind power plant through the data analysis and operational management. There are a number of technical parameters in the wind power plant, but it is necessary to take perfect decision from cradle to grave process of wind power plant or wind energy system. From the mean, mode, and median, you can assess ideal value of wind velocity and load demand. All the data of wind energy system-related parameters should not be analyzed in the same

74 Wind Energy System: Data Analysis and Operational Management

1919

way statistically because the entities represented by the numbers are different. For this reason, it is necessary to apply the concept of data analysis and operational management in the field of wind power plant or wind energy system.

References Azad AZ, Rasul MG (2015) Analysis of wind energy prospect for power generation by three Weibull distribution methods. Energy Procedia 75:722–727 Beltran B, Benbouzid MEH, Ahmed-Ali T (2012) Second-order sliding mode control of a doubly fed induction generator driven wind turbine. IEEE Trans Energy Convers 27(2):261–269. https://doi.org/10.1109/TEC.2011.2181515. Dupuis D, du Mouza C, Travers N (2020) Real-time influence maximization in a RTB setting. Data Sci Eng 5:224–239 Evangelista CA, Pisano A, Puleston P, Usai E (2017) Receding horizon adaptive second-order sliding mode control for doubly-fed induction generator based wind turbine. IEEE Trans Control Syst Technol 25(1):73–84. https://doi.org/10.1109/TCST.2016.2540539. Guediri AK (2017) Modeling and fuzzy control of a wind energy system based on double-fed asynchronous machine for supply of power to the electrical network. Int J Syst Assur Eng Manag 8:353–360 Hu J, Nian H, Hu B, He Y, Zhu ZQ (2010) Direct active and reactive power regulation of DFIG using sliding-mode control approach. IEEE Trans Energy Convers 25(4):1028–1039. https:// doi.org/10.1109/TEC.2010.2048754. Hussain CM (2018a) Nanomaterials in chromatography: current trends in chromatographic research technology and techniques. Elsevier Hussain CM (2018b) Handbook of nanomaterials for industrial applications. Elsevier, New York Hussain CM (2019a) Handbook of environmental materials management. Elsevier, New York Hussain CM (2019b) Modern environmental analysis techniques for pollutants, 1st edn. Elsevier, New York Hussain CM (2019c) Handbook of nanomaterials in analytical chemistry: modern trends in analysis. Elsevier, New York Hussain CM (2020a) Handbook of functionalized nanomaterials for industrial applications. Elsevier, New York Hussain CM (2020b) Handbook of manufacturing applications of nanomaterials. Elsevier, New York Hussain CM (2020c) The ELSI handbook of nanotechnology: risk, safety, ELSI and commercialization. Wiley, Hoboken Hussain CM (2020d) Handbook on miniaturization in analytical chemistry: application of nanotechnology. Elsevier, New York Hussain CM (2020e) Handbook of industrial applications of polymer nanocomposites. Elsevier, New York Hussain CM (2020f) Handbook of nanomaterials for sensing applications. Elsevier, New York Hussain CM (2021) Concepts of advanced zero waste tools: present and emerging waste management practices. Elsevier, New York Javadi M, Malyscheff AM, Wu D, Kang C, Jiang JN (2018) An algorithm for practical power curve estimation of wind turbines. CSEE J Power Energy Syst 4(1):93–102. https://doi.org/10.17775/ CSEEJPES.2016.00980 Kenne G, Nguimfack Ndongmo JDD, Fochie Kuate R, Fotsin HB (2015) An online simplified nonlinear controller for transient stabilization enhancement of DFIG in multimachine power systems. IEEE Trans Autom Control 60(9):2464–2469. https://doi.org/10.1109/ TAC.2015.2427591.

1920

V. Khare and C. J. Khare

Khare V (2020) Solar-wind energy assessment by big data analysis. In: Innovation in energy systems-new technologies for changing paradigms. IntechOpen, London Lei Y, Mullane A, Lightbody G, Yacamini R (2006) Modeling of the wind turbine with a doubly fed induction generator for grid integration studies. IEEE Trans Energy Convers 21(1):257–264. https://doi.org/10.1109/TEC.2005.847958 Luo H (2017) Physics-based data analysis for wind turbine condition monitoring. Clean Energy 1(1):4–22. https://doi.org/10.1093/ce/zkx005 Mann HS (2018) Effect of number of blades in ducted turbine system on kinetic energy extraction from chimney flue gases – benchmarking with wind energy system. J Mech Sci Technol 32:5443–5455 Naidu RPK (2020) Performance investigation of grid integrated photovoltaic/wind energy systems using ANFIS based hybrid MPPT controller. J Ambient Intell Humaniz Comput 12:5147–5159 Padhya M (2020) R-OO-KASE: Revocable Online/Offline Key Aggregate Searchable Encryption. Data Sci Eng 5(4):391–418 Shtessel Y, Edwards C, Fridman L, Levant A (2014) Sliding mode control and observation. Springer, New York. https://doi.org/10.1007/9780-8176-4893-0. Tummala ASLV (2019) Modified vector controlled DFIG wind energy system based on barrier function adaptive sliding mode control. Prot Control Mod Power Syst 4:1–8 Wawrzinek J (2020) Exploiting latent semantic subspaces to derive associations for specific pharmaceutical semantics. Data Sci Eng 5(4). https://doi.org/10.1007/s41019-020-00140-2 Wu F, Zhang X-P, Ping J, Sterling MJH (2008) Decentralized nonlinear control of wind turbine with doubly fed induction generator. IEEE Trans Power Syst 23(2):613–621. https://doi.org/ 10.1109/TPWRS.2008.920073. Zhang F (2020) Calculation and analysis of wind turbine health monitoring indicators based on the relationships with SCADA data. Appl Sci 10(1):410

Website for Further Reading https://www.ncbi.nlm.nih.gov/books/NBK481626/ www.tutorialspoint.com www.sas.com www.webopedia.com www.sciencedirect.com

Biomedical Data Retrieval Using Enhanced Query Expansion

75

Muhammad Qadeer, Chuadhery Ghazanfar Hussain, and Chaudhery Mustansar Hussain

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Motivation and Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review of Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biomedical Datasets Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Trec Genomic Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ohsumed Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Query Expansion and Relevance Feedback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Automatic Query Expansion in Biomedical Data Retrieval . . . . . . . . . . . . . . . . . . . . . . . . Biomedical Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ontologies in Biomedical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Retrieval Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biomedical Data Retrieval Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Query Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Documents Ranking Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Probabilistic Retrieval Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Query Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biomedical Data Retrieval Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Query Expansion from Ranked Documents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Term Scoring Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Average Precision (AP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mean Average Precision (MAP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Semantic Filtering for Biomedical Information Retrieval . . . . . . . . . . . . . . . . . . . . . . . . . . Semantic Similarity Measures and Query Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1923 1923 1923 1924 1924 1925 1925 1925 1926 1927 1927 1929 1929 1930 1932 1932 1934 1934 1936 1936 1937 1941 1942 1942 1942

M. Qadeer · C. G. Hussain Department of Education, Computer science and Technology, Punjab, Pakistan C. M. Hussain () Department of Chemistry and Environmental Science, New Jersey Institute of Technology, Newark, NJ, USA e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_63

1921

1922

M. Qadeer et al.

Evaluation of Semantic Similarity Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classification of Semantic Similarity Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methodology Adopted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . We Used Three Different Documents Retrieval Models: LM Jelinek-Mercer, LM Dirichlet, and BM25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection of Documents Ranking Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evaluation of Terms Scoring Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Semantic Similarity Measures Application for Query Expansion . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Biomedical Information Retrieval Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Research Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1943 1943 1946 1947 1947 1948 1948 1949 1950 1951 1953 1953 1954 1954 1954

Abstract

Biomedical data is growing up rapidly, and a better retrieval system is the need for its utilization. A basic problem while retrieving data from a system related to the queries is mismatch of words, which indicates the use of dissimilar words for expressing the identical concepts in given queries and in the stored documents. Two techniques are commonly used to solve this problem, i.e., query paraphrasing and query expansion. Query paraphrasing refers that the query is paraphrased by using synonyms of terms in the query. Query expansion techniques are further categorized as local and global. Local query expansion technique focuses on the analysis of the documents having top ranks retrieved for a query. Different ranking models have been introduced to rank documents in collections based on terms and features. A collection of candidate terms is obtained for expanding the given query from these documents. On feature selection from term pool, final selected candidate expansion terms contain a few terms which cause query drift problem. To overcome this problem, the semantic filtering technique was used. Semantic similarity measures are the basic techniques for successful semantic filtering. However, global query expansion relies on the analysis of the whole collection to find out word relationships. Synonyms of query words are extracted from a dictionary or thesaurus. In this research, we evaluated the famous probability-based ranking models such as LMDirichlet, LM Jelinek-Mercer, and BM25 for biomedical data retrieval process. We performed experimental analysis using diverse preprocessing techniques iteratively on 36 biomedical-related queries for the evaluation. State-of-the-art biomedical dataset Trec Genomic was used as a core for whole experimentation. It was observed that BM25 was the best information retrieval model for biomedical data. We used different terms scoring techniques such as Baseline, BNS, Chi-Square, CoDice, BIM, KLD, LRF, PRF, and RSV to score the terms related to the query. The average of MAP scores of all the queries was compared that exhibited BNS term scoring technique is the best for biomedical data. Different

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1923

semantic similarity measures such as path-based, Wu and Palmer, Leacock, and Chodorow were applied on terms extracted from BNS to get most appropriate terms for query expansion. Finally, queries expanded with the most similar terms each time and documents retrieved through the expanded queries and the MAP results were evaluated for the purpose of final declarations of this research. The results of biomedical data retrieval through query expansion were improved, and the LCH semantic similarity measuring technique found best for query expansion in biomedical data retrieval system. Keywords

Biomedical data · Query paraphrasing · Retrieval system · Automatic query expansion · Information retrieval system · Unified medical language system

Introduction Background Data retrieval is concerned to retrieve related documents from a large collection of the documents according to the user’s desire mentioned in the query. Various relevancy measuring techniques have been developed and applied in open as well as in the biomedical domain. Biomedical information retrieval systems usually find documents (the large text units) to satisfy user needs mentioned in a query. Some keywords or terms exist in the query. These terms are actually indexing units which have been simplified through different processes of normalization such as stemming and tokenization. Most of the techniques are query terms dependent, and systems compare these terms with the term occurrences in an individual document. They require initial retrieval through the ranking of documents on the basis of term similarity. The focus of this research is the betterment in biomedical data retrieval systems which is challenging in this era. There is also a need to consider and improve the performance of techniques and methods used in data retrieval systems. This research is based on Enhanced Biomedical Retrieval Using Discriminative Term Selection for Pseudo Relevance Feedback. The base research is covered in limited approaches and techniques. The better data retrieval techniques for biomedical data are to be proposed through enhancement of research and evaluation (Wasim et al. 2018).

Motivation and Problem Statement The information retrieval system is the core need for most of the networkbased applications. Prominent widespread information retrieval systems are the search engines. Today, it is impossible to find our desired information on the World Wide Web without them. Biomedical information retrieval system has also

1924

M. Qadeer et al.

become a demand with growing research in this field. Intrinsic uncertainty of biomedical information retrieval system distinguishes it from other information systems. Uncertainty about a document whether it has query relevant contents or not is called intrinsic uncertainty. Information needs can be precisely mapped in query formulation at least in standard database applications because properly defined elements of a database can form the exact answer, but it is much difficult in biomedical information retrieval system. Query formulation can never represent user-required information precisely, and there is not a clear procedure which can decide whether an object is the relevant answer or not. As the volume of biomedical information is rising exponentially, new problems and challenges are also appearing constantly for biomedical information retrieval systems. Developments in information retrieval technologies require theories and experimentations side by side to improve these systems. If a theory does not guide the experimentation, trials and errors become the parts of engineering which are not enough to solve newly arising problems and meeting the new challenges in the biomedical domain. Since the last two decades, a lot of new theories have been presented and empirically tested to cope with the problems and meet new challenges. These theories are known as formal models. Each of them guides toward the development of some information retrieval technology tool. It is necessary to learn these models or theories to understand these tools. Probabilistic models are most famous from them which have been deployed specifically to cope with the uncertainty in biomedical information retrieval. There are two approaches used in probabilistic models: The first is a classical approach which is based on the concept of relevance. It means a user gives a judgment about documents through relating to the query given by the user. The second novel approach is generalizing the proof of theoretical models for a particular result which overcomes the multidimensional judgment about an answer of given query to biomedical information retrieval system (Hiemstra 2009).

Review of Literature Biomedical Datasets Structure A dataset is the collection of related data which corresponds to a particular database table contents or a statistical data matrix. Each column in the matrix or table corresponds to a particular variable, and every row represents a member of the dataset. Dataset term is also used for the data collection of closely related tables, which corresponds to a particular event or experiment. Major properties of the dataset are attributes of a variable, have specific types and different statistical measures, and can be applied to them. Research in biomedical is growing which is determined by analysis and acquisition of large digital datasets. These datasets promote reproducibility in different ways, i.e., extraction of the available biomedical data, which is usually expensive in collecting and analyzing. Though the propagation of openly available data exhibits new challenges for researchers, the datasets

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1925

are relevant to their concern and might be scattered across the multiple repositories and indexed conflictingly (Cohen et al. 2017).

Trec Genomic Dataset It is also the collection of biomedical documents containing 162,259 documents from the 49 journals in the format of HTML. It is a total of 59 .zip files, one for each journal. Each document file is named by PMID with .html extension. There are 36 official topics available in this dataset, and available fields are the body, header, and other HTML basic tags. The python script is available to evaluate the results.

Ohsumed Dataset It is the collection of biomedical documents with 348,566 references commencing MEDLINE in nxml format. The presented fields are the title, abstract, the MeSH indexing terms, the author, the source, and the publication type. Fields include the titles or abstracts extracted from 270 medical journals over the five-year period (Xiong and Callan 2015).

Query Expansion and Relevance Feedback In the biomedical domain, the retrieval of query-relevant documents which can fulfill the user requirements is the great challenge for data retrieval systems. One of the most reliable and successful techniques is pseudo relevance feedback (PRF) to meet the challenge. While considering problems in the retrieval of required information, the PRF-QE is a feasible approach when some top-retrieved documents in the first retrieval are used to expand user’s given query. The approach is shown in Fig. 1. Here required an automatic relevance-based method for applying query

Fig. 1 Pseudo relevance feedback architecture

1926

M. Qadeer et al.

expansion technique that can automatically remake the given query (Singh and Sharan 2015).

Automatic Query Expansion in Biomedical Data Retrieval Biomedical data has been increasing dramatically, but the number of searched query terms remained almost constant. In the past, the average number of query terms were two to three. According to recent research, average has become five numbers of terms which are being used in a query which are required to expand with an appropriate automatic method, so that can retrieve most relevant results. Most of the users still use one, two, or three words in their queries to search required biomedical information. In this situation, the need for an automatic query expansion (AQE) has increased in biomedical data retrieval system. A user mentions required information in the query, and after normalization, the terms are given to a particular ranking model such as BM25 which ranks and retrieves corpus documents by comparing query terms with the terms in biomedical documents. Top k documents are considered to extract a pool of terms for further query expansion process. Terms collected in the pool are evaluated through different available methods for the purpose of scoring. Different scoring methods like Kullback-Leibler divergence (KLD), Co-occurrence, probabilistic relevance feedback (PRF), and Robertson Selection Value (RSV) are the available methods for biomedical data retrieval. Then from these scored terms, broad ranking algorithm is applied to rank terms. Finally, these ranked terms pass through semantic similarity measures for semantic filtering, and at last, the filtered terms are supplemented in the original query for query expansion. These expanded queries retrieve a huge collection of relevant biomedical documents for the user. This automatic expansion process may also be repeated multiple times for better retrieval of information for the user. The cyclic approach is mentioned in Fig. 2 (Singh and Sharan 2015). There are still some problems that occur while using automatic query expansion (AQE) in the biomedical information retrieval system. The major problem is that it does not work well when integral sparseness related to the terms is used by the user in query for the corpus with high dimensions. Another difficulty is that all the terms in top-retrieved documents (feedback) are not significant for expansion of the query. Some of them can be irrelevant or redundant and can misguide the retrieval process. Removal of irrelevant terms from the bulk of terms retrieved from the highranked documents is the main target of query expansion. The terms selected through query expansion may contain reliable and sufficient biomedical information about original documents. So query expansion should reduce high dimensionality of a corpus (pool of terms from feedback documents) as well as provide a high level of understanding with biomedical documents for improving the results of automatic query expansion. In AQE, various feedback-based query expansion term selection methods are broadly used. It has been detected that these methods increase accuracy and efficiency of information retrieval methods. Most conventionally used term selection

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1927

Fig. 2 Automatic query expansion model

methods are term association based or corpus statistics based depending upon the algorithms used in biomedical documents retrieval model (Mahdabi and Crestani 2014).

Biomedical Sources Most of the biomedical data sources (datasets) are ontology based. The definition of ontology is categorized into three groups of thought. First is the philosophical group, and it defines, “The ontology is the theory of existence.”

Implications of this definition in artificial intelligence (AI) purposes are very common. The second gives the conceptual view of ontology such as, “Ontology is an explicit specification of the conceptualization.”

Ontology is normally considered for the AI community. Third group splendid definition, which is valid in different knowledge-based fields such biomedical, is, “Ontology is the body of knowledge which comprehensively describes a specific domain, generally the commonsense domain.”

The last definition is of our concern which shows that ontology is a central body of knowledge, not as a way to express the knowledge (Harispe et al. 2014).

Ontologies in Biomedical The interest of researchers in the term ontology has been growing up day by day in the field of biomedical and bioinformatics. The ontology-based knowledge structures lined up biomedical resources like drugs, diseases, genes, clinical records, and scientific articles over clear conceptualization.

1928

M. Qadeer et al.

Unified Medical Language System (UMLS) UMLS contains a big multilingual and multipurpose meta-thesaurus which contains information related to the health and biomedical conceptions. This is developed from the electronic versions of different sets of codes, thesauri, taxonomies, and the list of different controlled medical terms. This system constitutes information over one million conceptions related to biomedical. Five million names of concept are extracted from hundred classifications (in several languages) or more systems and incorporated controlled vocabularies. Each thesaurus concept is allocated to at least a single semantic type and convinced semantic relationships between the semantic-type members. UMLS contains semantic network and meta-thesaurus (a database) which is a collection of terms and relationships between them and controlled vocabularies of biomedical concepts. Both are biomedical knowledge sources, but they have complementary nature. The structures of these have been compared in many types of research and studies, but their alignment has not been done (Slimani 2013). The search engine technologies become participants of different growing domains like lexical and biomedical, etc. These technologies have become a necessary part of biomedical data retrieval systems. The effective retrieval of biomedical data has become the need due to the evolution in biomedical resources on wide area of networks like Internet (Grossman and Frieder 2012). One of the finest examples is the UMLS (Unified Medical Language System) developed by the National Library of Medicine (NLM). UMLS is the framework to represent the knowledge of the biomedical domain. Various controlled medical ontology-based terminologies like Systematized Nomenclature of Medicine–Clinical Terms (SNOMED-CT) and Medical Subject Headings (MeSH) are incorporated semiautomatically in meta-thesaurus of UMLS (Demner et al. 2010). Ontologies are making possible to model biomedical domains through the concepts and semantic relationships between the concepts. Major ontology-based source of our interest is a part of this research (Lv and Zhai 2010). Researchers focused to improve the performance of biomedical retrieval system with individual term selection method, but most of the time, it remained a challenge. However, multiple term selection methods are also available for performance enhancement for biomedical information retrieval system. Past experiences of combining term selection methods did not analyze theoretically, but recently, evaluation is done of two high-performing but unassociated query expansion term selection methods (Harish et al. 2010). Biomedical concepts are classified and arranged in different hierarchies and semantic networks ontologies related to the biomedical domain. The concepts are labeled by Atomic Unique Identifiers (AUI). Relation existence between the concepts is represented by Concept Unique Identifier (CUI). It means AUIs of two related concepts are mapped to a CUI. The meta-thesaurus consists of two tables that provide relational information between CUIs (Hiemstra 2009). The ontology-based semantic similarity measures are classified into single ontology SSMs and multiple ontologies SSMs. Both have relative contribution and

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1929

limitations. The conceptual similarity is suggested by mapping in terminologies (McInnes et al. 2009).

Data Retrieval Application “A search engine is the web-based application of data retrieval that collects and organizes the contents from all over the internet. On a query given by the user the engine provides links to the content that matches query terms.” (The Balance). The content provided by a search engine could comprise of text, images, and different types of documents. Some of them may be used to mine data in different databases or open directories. The WHOIS (pronounced as who is) is a query response protocol used for data querying from Internet resources and was used in 1982. The first suitable search engine was Archie designed to search the content files (FTP files) in 1990 (Manning et al. 2008).

Biomedical Data Retrieval Process Most of the data is stored in the form of documents in biomedical data retrieval system. This system assists the user in getting required biomedical information and notifies the position of existing documents that may contain information related to the user’s requirements. Some of these biomedical documents will expectantly satisfy the information needs of the user. These are relevant documents. A seamless biomedical data retrieval system retrieves only those documents which are relevant and discard irrelevant, but it does not exist and may not be in the future because search query of the user is predictably incomplete, as well as relevance subjectively depends upon the user’s judgment. Practically, it has been observed that in two users given the same query but judged the relevance differently from retrieved documents, one is satisfied with the results, but the other is not. Biomedical data retrieval system has to support three essential processes: content representation in the document, user’s information need representation in the query, and the comparison of these two. These processes can be viewed in Fig. 3. Each square box represents data, and rounded box represents a process. Generally, documents are represented in the indexing process. This process may include storage of the documents but partly such as title, abstract, and actual information-related position of each document. A user expresses the desired information in query formulation process. This query is actually an interactive dialogue between user and biomedical information retrieval system. Matching is the comparing of user’s query with document representation in the system. The results of matching appear normally in the form of ranked documents list. More relevant documents allocated high ranks are put toward the top in a hierarchy. Quality of a retrieval process is dependent on response to the query given by user with high degree of accuracy to the relevant documents (Pedersen et al. 2007).

1930

M. Qadeer et al.

Retrieved Documents

Feedback

Indexing

Query

Indexed Documents

Query Formation

Indexing

Information Need

Documents

Fig. 3 Documents retrieval process

Query Preprocessing A user specifies needs in the query which is normalized for exact representation of requirement. For this purpose, various processing techniques are applied on biomedical data retrieval systems. These techniques may be used due to morphological and grammatical reasons. This segment is to trim down inflectional forms of the words into the common base form (Moffat et al. 2007).

Tokenization The chopping of a query into the parts (called tokens) by throwing away the punctuation and unwanted characters is known as tokenization. The same task can be applied over a document to assure that sequence of characters in the text is matching with query terms. For example, a biomedical query is tokenized as following:

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1931

Which change expressions with relationship through disease movement in lupus?

After tokenization, it will be Which change expressions with

relationship through disease movement in

lupus

Multi Words Grouping They are the words which give proper meaning in the form of group, and they are grouped into a single token. Disease activity is a group of words in the query so they will be in one token. Which change expressions with

relationship through disease movement

in

lupus

Removal of Stop Words The words occurring in maximum documents having little impact are called stop words. They are excluded from vocabulary of the query. These have none semantic value to the relative domain of documents. A list of the stop words or terms in the corpus of documents is organized through their occurrences (frequency). They are discarded by comparing the list. Change

expressions

relationship

disease movement

lupus

Stemming or Lemmatization Stemming refers to the chopping the ends of the words in hope to retrieve the stem of the words which give most accurate meanings. It includes sometimes the removal of the derivational affixes. Change

Express

relation

disease movement

Lupus

Lemmatization is the technique which analyzes the word morphologically and identifies their basic forms, which are called lemma. It is the more complex form of stemming. It implies for identifying synonyms of words in the given query. Both lemmatization and stemming simplify the searching job of an information retrieval system. The process of normalization is shown in Fig. 4.

1932

M. Qadeer et al. change transform

Express Phrase

relation association

Fig. 4 Query normalization process

disease movement soaring

Lupus infection

Input Query

Tokenization

Multi Words Grouping

Removal of stop words

Stemming or Lemmatization

Normalized Query terms

Documents Ranking Models A well-designed and effective ranking algorithm uses frequently occurring terms in a document and other statistical information like the number of pointing links toward the document. Statistics-based ranking algorithms make the search process easy and reduce user’s reading time. Probabilistic models are most effective in recent information retrieval systems (Pérez-Agüera and Araujo 2008).

The Probabilistic Retrieval Model Probability distribution over the terms of query is used effectively in ranking of the documents.

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1933

The analysis of probabilistic ranking models which are being most of the time considered for biomedical information retrieval systems is the task here. A considerable detail of the models is given in the following:

LM Jelinek-Mercer Similarity This algorithm captures important patterns in the text while leaving out the noise. λ is a ranking parameter in similarity. It is the meta parameter, and its value is constant which is independent of document and query. Correct setting of its value is very important. pλ (wi | d, C) =



((1 − λ) .pd (wi | d) + λ.pC (wi | C))

(1)

wi ∈q

Pd (wi | d) = tfd /doclen and PC (wi | C) is probability of the term in corpus. Optimal value of the λ is about 0.1 for short or title queries and about 0.7 for long queries. The default value is considered 0.1. When value approaches to 0, documents matching more terms in the query are ranked higher than those having fewer matching terms (Urbain et al. 2006).

Okapi-BM25 BM (Best Matching) 25 is sound-gaining implementation of the probability ranking model in the biomedical domain. All biomedical documents with binary relevance attribute described which document is relevant and which is not. This model assumes that the measure is a probabilistic itself with binary relevance. BM25 algorithm is used in biomedical information retrieval system to rank the documents rendering to the significance of a given query. It was implemented firstly in Okapi data retrieval system. In Okapi, it assigns ranks to the documents on the basis of appearance of query terms in each document irrespective to the relationship between terms within the document. This performs preliminary experiments in a better way as compared to the other models. It is recognized as a best weighting scheme for Ohsumed collection. Its formula is score(q, d) =

|q|  i=0

tf (qi , d) . (k1 + 1) N − df (q1 ) + 0.5  . log  |d| df (qi ) + 0.5 tf (qi , d) + k1 . 1 − b + b. avgdl (2)

tf (qi ,d) is the term frequency which shows how many times the query term qi appears in document d, where |d| is the length of document d in words or terms, and avgdl is the average document length of overall documents. There are two free parameters, k1 and b, usually considered k1 = 2.0 and b = 0.75, N shows the total documents in corpus, and df(qi ) represents the documents containing qi , the query term (Lu et al. 2005). Sentiment analysis is the part of Natural Language Processing (NLP) that surveys and classifies the effective states and subjective evidence about the topic or entity.

1934

M. Qadeer et al.

The sentiment analysis is correlated with the sentiment as firm by the users. This can be done by looking at the subjectivity, the objectivity, and the magnitude of sentiment in the text at the sentence level while keeping track of contextual.

LM (Language Model) Dirichlet Using of Jelinek-Mercer has a problem that longer documents provide better estimates and could get by with less smoothing. Fractional number occurrences can be added to each term frequency to achieve smoothing. Pμ (wi | d) =

  f req (wi , d) + μp (wi |C)  | d | +μ w ∈q

(3)

i

where freq(wi ,d) represents frequency of term/word in a document, P(wi | C) is the frequency of term in the corpus, |d| is length of the document, and μ is tuning parameter for best results μ = 2000 (Fujita et al. 2004). Documents and queries are modeled on the basis of probability distribution. The distribution is applied on the sequence of the words in biomedical retrieval. Three main approaches are used for applying probability in the domain. First, queries can be generated from documents language model which is called query likelihood. Second, documents are generated from a query language model called document likelihood model. The third is a comparison of query likelihood and document likelihood model which is known as divergence model (Amati et al. 2003). The search facility can include different types of documents having text, images, and sounds. They may be used as data mining in different databases or open directories. The WHOIS protocol is very effective for data querying from Internet resources. Technologies are the necessary part of biomedical data retrieval systems in this era. The effective retrieval of biomedical data has become the need due to the evolution in biomedical resources on Internet (Jerome et al. 2001).

Materials and Methods Query Expansion The process of supplementing the additional phrases or terms to the given query for improving the performance of data retrieval is known as query expansion. In the biomedical data retrieval system, query expansion has been suggested to deal with various problems such as word mismatch. Query expansion is an incredibly effective technique which can boost up the overall performance of an information retrieval system. Automatic query expansion is a manner of supplementing the additional phrases/terms to the novel query without intervention of the user to improve the performance of biomedical information retrieval. A great number of approaches have been used, but recent of those techniques have been focused which can analyze the corpus used as well as discover relationships of words and can also

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1935

analyze documents retrieved through initial query as initial feedback. All available biomedical documents are analyzed to find the relationships between biomedical terms or word in documents with given query terms in global query expansion. Synonyms of the words or terms are extracted from thesaurus or dictionary available globally, and these are also utilized in the analysis of documents.

Local and Global Query Expansion in Biomedical Domain In local query expansion, the whole corpus of biomedical documents is analyzed, and the documents are ranked against biomedical terms given by the user in the query. Moreover, relevant terms extracted from the top-ranked documents are used to expand user’s given query. Local query expansion is focused for the local feedback. This is more operative than global query expansion. Some local query expansion techniques are not strong and damage seriously retrieval process when a few relevant documents exist. Local Context Analysis is the famous and effective local query expansion technique used in initial biomedical retrieval systems. In this technique, biomedical terms are selected from top-ranked documents on co-occurrence basis with the terms original query. Experiments prove that a limited collection of documents in local context analysis show consistent results. Sometimes, local query expansion techniques use general statistics like documents frequency of the biomedical terms. These statistics might be cheap and easily available. The source of expansion terms is the set of biomedical documents which have top ranks. Many top-ranked biomedical documents use relevance feedback for query expansion (Lin and Huang 2017). HARPER and CROFT suggested these techniques where the information from top-ranked biomedical documents estimates the probabilities of the terms in relevant set of the query. Chosen terms through local feedback for biomedical query expansion are the most frequent terms (except stop words). Recently, Trec results in confirmation by local feedback significantly improved effectiveness of biomedical information retrieval. Every time, local feedback cannot be proved as a vigorous technique and can extremely degrade biomedical retrieval performance. Local Concept Analysis can also be considered as a prime tool for the techniques of global query expansion. Biomedical features known as “CONCEPTS” are extracted from top-ranked biomedical documents. The concepts may be simplified into a single term or the pair of terms. Moreover, sophisticated biomedical concepts may be nouns or noun phrases. This analysis gives ranks the concepts in accordance with their co-occurrences. The top-ranked concepts from top-ranked documents are used for query expansion. The experimental results show the local context analysis which produces more operative and stronger query expansion than other surviving techniques in biomedical domain. Global Query Expansion is based on analysis of term relationships and requires some corpus including statistics which take remarkable amount of the computer resources for computing co-occurrence of data nearly all possible term pairs in the collection.

1936

M. Qadeer et al.

Fig. 5 Biomedical information retrieval process

Clustering is a global query expansion technique in which words are grouped into clusters based on the co-occurrence. These clusters are used to expand given query. Some other popular global query expansion techniques are the latent semantic indexing, phrase finder, and similarity thesauri. Usually, the global expansion techniques do not express consistent and reliable results until well-off strategies for the term selection are used. Global techniques usually need the co-occurrence of information for each pair of terms. It is a computationally challenging task for the large collections (Xu and Croft 2017).

Biomedical Data Retrieval Process A user enters query to retrieve desired information to the biomedical data retrieval system. After preprocessing, the query is given to the searching model such as BM25. This model matches the query terms with document terms present in corpus. According to the number of occurrences of terms, documents are ranked in the corpus. This whole process is shown in Fig. 5.

Query Expansion from Ranked Documents A large number of models for ranking biomedical documents have been introduced which are based on terms or features present in the collections. A set of biomedical expansion terms (called candidate terms) is obtained from these documents to expand the user-specified query. Selected candidate terms from a pool of biomedical terms also included some terms which can cause the serious problem of query drift. Semantic filtering is a popular technique to cope up this problem. Different semantic similarity measuring (SSM) techniques are available for successful

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1937

semantic filtering of terms in the biomedical domain. Query expansion in biomedical domain is used to get the most relevant information in the form of documents from huge collections but still requires improvements. It has been widely researched that some expansion techniques provide insufficient results. The documents are ranked according to the given query. Out of these initial set of documents, most appropriate terms are designated and then supplemented to the original query. The expanded query is re-executed on the corpus. Two important parameters are used for the analysis query expansion process. First is number of the documents from which expansion terms can be possibly reclaimed, and the other is the number of expansion terms which are to be supplemented in the original query. Adding new terms commences a risk of changing the query topicality, which leads search in a different direction. Adding such a term, the query might drift. This drift can be handled with the help of semantic filtering techniques (Roy et al. 2016).

Term Scoring Methods It is obvious that terms extracted from ranked biomedical documents may be irrelevant and redundant which may lead to the ambiguous query expansion. To prevent this situation, all terms of the pool should be ranked on information basis through statistical methods. Here, we describe famous term ranking methods for query expansion.

Robertson Selection Value (RSV) RSV term scoring technique assumes to retrieve terms by ranking them by measuring association with query. This principle is proposed by Swets for examination of the distribution of values of match function over the whole document collection. Moreover, two distributions are considered, one for relevant documents and the other for nonrelevant documents. The score of a candidate term is calculated by the following formula: RSV (T , C) =



W (T , D) (P R − PN R )

(4)

R∈D

where PR is the probability of terms in relevant documents, and PNR is the probability in nonrelevant documents in corpus (Singh and Sharan 2018).

Binary Independence Model (BIM) It is a probabilistic technique for data retrieval which is centered on simple assumptions for estimating document and query terms relevancy or similarity. Binary independence assumes Ti of document D is independent statistically in the relevant class R as well as in nonrelevant class NR. An assumption is signified in the following formula:

1938

M. Qadeer et al.

BI M (T , C) = log

P (T i /R) (1 − P (T i /NR)) P (T i /NR) (1 − P (T i /R))

(5)

where T is the term, R is the number of relevant documents, and NR is the number of nonrelevant documents and probabilities P(Ti /R) and (1 − P(Ti /NR)) to be estimated from relevant as well as nonrelevant documents. C represents the corpus having documents. Candidate terms are graded based on the score obtained as of above formula. After ranking, high score terms can be selected for further query expansion (Singh and Sharan 2018).

Lavrenko Relevance Feedback (LRF) It is a language model-based technique and assigns a score to the terms for query expansion through Lavrenko relevance model. It has been formulated as following: LRF (T ) =

 All−R

log

P (T |M R ) P (T |G )

(6)

where P(T|MR ) is calculated from the following formula: P (T |M R ) = λ. 

T F (T , R) + (1 − λ) .P (T |G ) T ∈R T F (T , R)

(7)

 TF(T, R) is frequency of the term T in relevant document R, and T ∈ R TF(T, R) is summation of all the term frequencies in the relevant document. λ is the tuning parameter that is adjusted during the experimentation. The value 0.6 gives the finest results. P(T|G)is the probability related to the occurrence of the term T in corpus (Lavrenko and Croft 2017, August).

Baseline TF-IDF (Term Frequency and Inverse Document Frequency) is a frequently incorporated algorithm for computing similarity between queries with documents. Another way to make more useful, this algorithm is more expanding the formula of TF-IDF as the product of TF weight and IDF weight which is signified as:

Score(T , D) = log 1 + T F T ,D . log (N/DF T )

(8)

where TFT, D is the term frequency in document D, and DFT is the documents frequency having term T. N is the number of documents in corpus. It can be accumulated for whole corpus (Abdulla et al. 2016).

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1939

Binormal Separation (BNS) BNS is the feature selection measuring method which is proposed by Forman. It uses standard normal distribution and cumulative probability. Its formula is the following: 

      −1 P (T , Ci ) −1 P T , C i −F BNS (T , Ci ) = F    P (Ci ) P Ci

(9)

where F is standard normal distribution, whereas F−1 is its inverse cumulative probabilities function. To avoid undefined value, F−1 (0) is considered by 0.0005. The Ci represents category of relevant features cluster, and C i is other than feature category. The larger the BNS value, the larger the difference between the prevalence of term t in the category Ci and C i . BNS performance is better in the evaluation metrics on recall rate (Basu and Murthy 2016).

Kullback-Leibler Divergence (KLD) KLD is an effective term-evaluating technique which is commonly used in language processing application for speech processing-based statistical language modeling. It is a commonly used information theory. The score to query relevant terms is assigned on probability bases in relevant documents and corpus. KLD(T ) = PR (T ). log

PR (T ) P C (T )

(10)

whereas PR (T) is probability of occurrence related to the term in top rankeddocuments R which is calculated by the formula:     tf T D D∈R    PR (T ) =    D∈R T ∈D tf T D 

(11)

PC (T) is probability of occurrence of the term in the corpus and calculated by     tf T D    PC (T ) =    tf T D D∈C T ∈D 

D∈C

(12)

KLD(T) assigns the score to the terms found in the pool of terms. Because it uses a probabilistic approach, score assigned is between 0 and 1. The term with 0 scores is an irrelevant term. On the other hand, score 1 represents that the term is completely relevant and can be an excellent candidate for query expansion (Robertson and Zaragoza 2009).

1940

M. Qadeer et al.

Chi-Square It measures how much the independent term T found in relevant as well as in irrelevant class. This divergence of the two events is called Chi-Square. The lesser independence is given a higher score. The formula of the Chi-Square is Chi − Square =

[P R (T ) − P C (T )]2 P C (T )

(13)

Here, PR (T) is term probability of T in relevant documents. PC (T) is probability of the term T in corpus (Forman et al. 2003).

Co-Occurrence It is the terms association-based technique assigns a score to the terms on the presence in a terms pool. The score is assigned in this technique by evaluating the relationship of the candidate terms to the word of a query. It is the algorithm to find a relationship between terms in query and corpus. The coefficients such as CoDice, Cosine, and CoJaccard are used to find an association of the co-occurrence between two terms. So it is formulated as following:

Co − Dice Ti , Tj =

dfij df i + df j − df ij

(14)

dfi ,dfj are the documents frequencies in which term i and term j exist, respectively. dfij is number of the documents in which the terms i and j collectively occur together. D is the number of top-retrieved documents. Here, a problem occurs. When query expands with greatly similar terms, it causes query drift problem. For avoiding this problem, the IDF (inverse document frequency) concept is introduced. Co-degree can be considered which also caters the IDF. A mathematical expression can be in the following way: 

I DF (cT ) CoDegree (qi , cT ) = log10 (CoDice (qi , cT ) + 1) . log10 (D)

 (15)

where qi is query term, ct is candidate term, and then co-degree may be calculated. D represents the number of top-retrieved documents.  I DF (cT ) = log10

N Nc

 (16)

where Nc is the number of documents in the corpus which holds ct candidate term, and N is the total number of documents in a corpus.

75 Biomedical Data Retrieval Using Enhanced Query Expansion



CoOccurance (Q, ct ) =

(Co deg ree (qi , ct ))

1941

(17)

qi ∈Q

The value related to the candidate term against all the terms of query can be obtained by the above formula (Yang et al. 2003).

Probabilistic Relevance Feedback (PRF) This technique assigns a score to the terms found in the pool of terms by calculating their probability in the relevant and nonrelevant documents. A term having a higher probability in relevant class is considered more suitable term for query expansion. The formula of PRF is the following:

P RF (T ) =

P Rel (T ) P N on−rel (T )

(18)

where PRel (T) is probability of the term in relevant documents, and PNon − rel (T) is probability of the term in nonrelevant documents (Salton and Buckley 1990).

Average Precision (AP) Biomedical information retrieval system returns the ranked sequence of biomedical documents; it is preferred to present the documents in that order in which they returned. For this average precision, values from the ranks are to be calculated for a retrieved relevant document (Rivas et al. 2014). It can be calculated from following formula: N 

Avg _ Pr =

(P (r).Rel(r))

r=1

Cr

(19)

where r indicates the rank of biomedical document. N represents number of retrieved documents. P(r) shows precision which means proportion of a retrieved set is relevant at a specified rank. Rel(r) is a binary function which indicates whether a document is relevant or not, and Cr represents corpus.

1942

M. Qadeer et al.

Mean Average Precision (MAP) It summarizes the results of ranking documents retrieved from numerous queries by averaging the Avg _ Pr. It can be calculated from the following expression:

Q 

Avg _ Pr =

(Avg _ Pr(q))

q=1

|Q|

(20)

Automatic query expansion is the way of enhancing the supplementary phrases/terms to the original query without interference of the user for improving the performance of biomedical data retrieval. The approaches used have been focused which can also analyze the corpus used as well as discover relationships of words in documents retrieved through original query as initial feedback (Imran and Sharan 2009).

Semantic Filtering for Biomedical Information Retrieval Poor performing terms extracted through different documents retrieval models can mislead in query expansion. They must be filtered out to get highly related terms by applying semantic filtering. For the purpose of semantic filtering, the similarity measures of terms/concepts are the foundation of ontology-based information retrieval (Harispe et al. 2014). We are to describe currently available semantic similarity measures and also to compare them for the purpose of resolving the problem under consideration. This study has a great contribution in the biomedical IR research. By recognizing the limitations in the global perspective of existing techniques for query expansion, a novel approach can be proposed to address the misleading in query expansion.

Semantic Similarity Measures and Query Expansion In biomedical domain, ontology-based data storage is preferred in a corpus. Semantic similarity measures (SSMs) are used to discover the similarity among concepts defined in ontologies to make valuable information retrieval. Measuring of concepts is taxonomically compared in the ontology. This is applicable to design algorithms for information retrieval, disambiguation of text, repositioning the drug, and clustering genes according to their molecular functionality (Alipanah et al. 2010).

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1943

Evaluation of Semantic Similarity Measures Semantic similarity measures (SSMs) are used to discover the similarity of biomedical concepts defined in ontologies. A valuable biomedical information retrieval system can be developed on the basis of ontologies and concepts similarities in biomedical ontologies. The concepts are taxonomically compared in these ontologies. This is suitable to design information retrieval algorithms, text disambiguation, drug positioning, and gene clustering according to their molecular functionality for biomedical information retrieval system.

Classification of Semantic Similarity Measures The information processing and retrieval systems require computational knowledge sources to measure the similarity between concepts. Concept-based semantic similarity measures (SSMs) are classified for single ontology and for multiple ontologies. These are computational-based classifications. The similarity is measured by ontological, structural, or informational content. All measures are demonstrated in the form of parameterized functions. Semantic similarities measures between concepts are applied on single ontology based on the quantitative approaches like path, depth, information content, and features (Wei et al. 2007). These are divided into the following categories: • Edge-counting measure • Path-based measure • Depth relative measure • Information content (IC)-based measure • Hybrid measure

Edge-Counting Measure These are the measures in which edges from one node to the other are counted for the purpose to measure similarity. These are the following: Path-Based Measure The path-based measure is used when the structure of biomedical ontology shows is-a-type relationship between two concepts. It is a typical representation of the hierarchical relationship between concepts. It was suggested and employed for studying a particular case of semantic relations semantic distance. Path-based measure considers the distance separating two concepts. The degree of similarity is determined by path length in such a way that the shortest path between two concepts is measured by counting edges (Rada et al. 1989). It works for spatial orientation and fine for networks or maps, but sometimes, is-a hierarchy does not give reasonable approximation because deeper paths tend to travel less semantic distance. The nodes related to biomedical concepts are counted

1944

M. Qadeer et al.

in this similarity. Here, parent-child (is-a) relationship is exploited specifically. Semantic similarity between two biomedical concepts, c1 and c2, is calculated by the following formula: Similarity (c1, c2) =

1 d

(21)

where d is the number of nodes between c1 and c2 following through shortest path between them. Above equation is used where we have to set the minimum distance between ancestor and seed concept in the document. A concept has one value of measure when similarity is to itself (self-similarity). Weighted Links Measure It is an extension of the path-based measure. Weighted links are proposed for computing the two concepts of similarity. There are two factors which can affect weights of links. First is the depth of the specific hierarchy, the density of taxonomy for a given level, and the strength of association between parent and child nodes. Therefore, the distance between the two concepts is obtained by summing up the weights of the traversed links instead of counting them. Criticism on edge-counting approach is that quality of the taxonomy sensitivity can make a problem to apply this approach. Due to attractive features of semantic and mathematical traceability, Rada concluded that if it is operated on better semantic nets, it can capture conceptual similarity. Depth Relative Measure The depths of edges which connect two concepts in the whole structure of ontology are considered with the shortest path approach because path only doesn’t account for the specificity of concepts. Deeper concepts hold more specific paths, and they travel less semantic distance. Depth is calculated from the root to the targeted concept. Different depth relative measures have been proposed, but well-liked measures are the following: Wu and Palmer Let C1 and C2 be two biomedical concepts in taxonomy. Similarity measure between these concepts considers the position of C1 and C2 to the position of C which is the most specific common concept. Numerous parents may be shared by C1 and C2 through various paths, but the most specific common concept is the closest common ancestor C (common parent related to the minimum links with concepts C1 and C2). This closest common ancestor is also called as least common subsumer. The formula is SimW u&p (c1, c2) =

2∗ N N1 + N2 + 2∗ N

(22)

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1945

N1 and N2 are distances (number of links or is-a relationships) which separate C1 and C2 from the specific common concept, whereas N is the distance of the closest common ancestor of C1and C2 from the root node. Leacock and Chodorow Leacock and Chodorow (LC) proposed the relatedness similarity measure. The length of shortest path between two biomedical concepts (taken by counting nodes) and the maximum depth of taxonomy are considered in this measure. The shortest path between the two concepts of ontology is restricted to taxonomic links by dividing the double of maximum depth of the hierarchy.  Simlc (c1, c2) = − log

N 2∗ D

 (23)

where N is the length of the shortest path between c1 and c2 concepts related to biomedical calculated by using the node-counting process, and D is the maximum depth of taxonomy. The shortest path between concepts c1 and c2 in the restricted ontology to the taxonomic links is normalized by dividing N by double of the maximum depth of the hierarchy.

Information Content Measures These measures use information content (IC) of the concepts to evaluate the semantic similarity between two concepts or terms. The IC value of a concept is found out by calculating the frequency of concept term in the collection of documents. Some fine semantic similarity measures are discussed in the following section. Resnik Measure Information content (IC) of shared parents of biomedical terms is used in this measure. If two biomedical concepts present more shared information, they are considered more similar, and shared information by two concepts is specified by the information content of the subsume concept in the taxonomy. We have two input concepts, c1 and c2, for this measure. Then the output will be a numeric similarity measure. These measures rely on least common subsumer (LCS). It is the most unambiguous concept that can be shared ancestor by two concepts. SimResnik (c1, c2) = IC (LCS (c1, c2))

(24)

Information about the size of the corpus is extracted from this measure; a large numerical value indicates the large corpus. The Resnik measure is coarse since several different pairs of the concepts can share the same LCS. Lin Measure This measure is based on an ontology which is restricted to a corpus and hierarchic links. This similarity considers the information shared by two concepts like Resnik,

1946

M. Qadeer et al.

but there is a difference in the definition. The definition has the same components as Resnik measure, but the combination taken is ratio, not a difference. Simlin (c1, c2) =

2∗ IC (LCS (c1, c2)) I C(c1) + I C(c2)

(25)

This measure compares the terms of given ontology and shows a better ranking of similarity than Resnik measure. Jiang and Conrath Measure This measure is in a similar manner as Resnik, but a corpus was used in addition to hierarchic ontology (links of taxonomy). The distance between C1 and C2 is formulated as a difference between the sum of the IC of the two concepts and the IC of their most informative subsume. Simj cn (c1, c2) =

1 I C(c1) + I C(c2) − 2∗ IC (LCS (c1, c2))

(26)

IC(c1) and IC(c2) are information contents of two concepts, respectively. LCS is least common subsume of c1 and c2.

Hybrid Measures Hybrid measures are a combination of structural characteristics (such as path length, depth, and local density) and information content or feature-based measures. However, for some concrete situations, their accuracy appears higher than basic edge-counting measures. It depends upon the empirical alteration of the weights according to input terms and the ontology. Hybrid measures combine the knowledge which is derived from different information sources. The main benefit of these approaches is that if the knowledge related to an information source is insufficient, then it can be derived from alternate sources of information.

Methodology Adopted In this research, we adopted simple but effective methodology to address the problems during the information retrieval process and for better biomedical data retrieval. • A user enters a query to retrieve biomedical information from biomedical retrieval system. • The query is normalized by the removal of stop words, stemming, etc., through normalization system.

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1947

• The normalized query is passed to the document retrieval model BM25. This model ranks “n” biomedical documents of Trec the collection of biomedical documents in HTML format. • A pool of query-related terms from ranked documents is evaluated. • Query-related terms are awarded score through scoring method such as BNS. • Some of the high-scoring terms are further evaluated for similarity to the query terms through semantic similarity measures through UMLS-based thesaurus and assigned score. The terms which are most similar to the query terms gain higher scores. • Higher-scoring terms are used to expand the original query. • This expanded query is passed again to the biomedical retrieval model to give more biomedical documents related to the query. • The results are evaluated through MAP (mean average precision). The complete procedure is shown in Fig. 6 and briefly described below.

Experimental Setup Experimental setup of this research consists of the following: • Trec dataset was locally used to evaluate biomedical information retrieval. This dataset consists of 162,259 biomedical documents, while 10,000 documents were retrieved after indexing. A set of 36 biomedical queries was used for evaluation. • We used an open-source search platform “Solr” for experimentation. We configured Solr and indexed Trec documents through Solr indexer (Smiley and Pugh 2011).

We Used Three Different Documents Retrieval Models: LM Jelinek-Mercer, LM Dirichlet, and BM25 • Then we used different term scoring methods such as Baseline, BNS, ChiSquare, CoDice, BIM, KLD, LRF, PRF, and PRF. • MAP (mean average precision) of each method was found for final documents retrieval evaluation. • For semantic similarity measures of query terms and retrieved document terms, we configured UMLS on our local server (Bodenreider et al. 2004). • We measured similarities between query terms and top-retrieved document terms by using different semantic similarity measures through MeSH thesaurus. • The original query was expanded with five top similarity scoring biomedical terms. • Finally, documents retrieved through expanded query and MAP are calculated.

1948

M. Qadeer et al.

User Query

Query Preprocessing

Retrieval Model

Final Docs Retrieved

Top n Retrieved Documents having Terms Pool

Scoring Terms through BNS method

Semantic Similarity Based Query Expansion

UMLS Thesaurus (MeSh)

Expanded Query

Fig. 6 Methodology adopted

Results and Discussion Selection of Documents Ranking Model Three different document retrieval models were evaluated in this research to trace out the best for biomedical information retrieval using enhanced query expansion. All three models were evaluated through Baseline standard scoring technique. The results extracted from the experimental setup are given in Table 1.

75 Biomedical Data Retrieval Using Enhanced Query Expansion Table 1 Document retrieval models scores

FD 5 15 25 35 45 55 Average

BM25 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703

LM Jelinek-Mercer 0.2438 0.2438 0.2438 0.2438 0.2438 0.2438 0.2438

1949 LM Dirichlet 0.2381 0.2381 0.2381 0.2381 0.2381 0.2381 0.2381

Fig. 7 Comparison between document retrieval models scores

Figure 7 shows the comparison between models for documents retrieval with 36 selected biomedical queries applied on Trec the corpus. It is clearly observed that the BM25 model gives better results than LM Jelinek-Mercer and LM Dirichlet. Furthermore, other term scoring techniques are applied over BM25 for query expansion.

Evaluation of Terms Scoring Method Ten popular term scoring methods were used for the purpose of evaluation which is the best for biomedical information retrieval. The methods are applied on retrieved documents through BM25 data retrieval model. All methods were applied experimentally on different candidate terms such as 5, 15, 25, 35, 45, and 55 from retrieved biomedical documents 5, 15, 25, 35, 45, and 55. The results extracted from the experimental setup are given in following Table 2. The Baseline, BNS, Chi-Square, CoDice, BIM, KLD, LRF, PRF, and RSV techniques are used on retrieved biomedical documents through BM25 model. Fig. 8 shows that BNS technique gave highest average score of terms-related MAP. The top scoring terms through BNS were evaluated through semantic similarity measures.

1950

M. Qadeer et al.

Semantic Similarity Measures Application for Query Expansion The popular semantic similarity measures such as Path base, Wu Palmer, and Leacock Chodorow were applied on top scoring terms extracted from retrieved documents to expand the query. Table 2 Comparison of term selection methods FT FD 5 5 15 25 35 45 55 5 15 15 25 35 45 55 5 25 15 25 35 45 55 5 35 15 25 35 45 55 5 45 15 25 35 45 55 5 55 15 25 35 45 55 Average

Baseline 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703 0.2703

BNS 0.2545 0.2815 0.2738 0.2791 0.2803 0.2850 0.2993 0.2987 0.2974 0.2936 0.2945 0.2934 0.3050 0.3014 0.3030 0.2999 0.2978 0.2899 0.3078 0.2975 0.2921 0.2926 0.2960 0.2934 0.3094 0.3049 0.2927 0.2985 0.2922 0.2944 0.3021 0.3012 0.2866 0.2834 0.2865 0.2860 0.2929

ChiSqr 0.2850 0.2801 0.2754 0.2758 0.2751 0.2781 0.2862 0.2908 0.2852 0.2833 0.2836 0.2831 0.2956 0.2925 0.2872 0.2884 0.2871 0.2870 0.2852 0.2953 0.2944 0.2918 0.2885 0.2868 0.2973 0.3051 0.2931 0.2922 0.2904 0.2905 0.3002 0.2987 0.2925 0.2924 0.2887 0.2867 0.2886

CoDice 0.2710 0.2808 0.2752 0.2774 0.2847 0.2807 0.2802 0.2786 0.2758 0.2743 0.2745 0.2715 0.2829 0.2793 0.2744 0.2663 0.2602 0.2577 0.2764 0.2656 0.2644 0.2704 0.2682 0.2597 0.2690 0.2578 0.2495 0.2529 0.2508 0.2478 0.2483 0.2503 0.2480 0.2389 0.2403 0.2347 0.2650

BIM 0.2755 0.2770 0.2742 0.2739 0.2742 0.2736 0.2592 0.2584 0.2504 0.2478 0.2484 0.2481 0.2559 0.2558 0.2529 0.2482 0.2443 0.2436 0.2515 0.2494 0.2467 0.2414 0.2385 0.2379 0.2519 0.2504 0.2403 0.2342 0.2275 0.2236 0.2512 0.2474 0.2349 0.2279 0.2252 0.2201 0.2489

KLD 0.2418 0.2376 0.2193 0.2270 0.2184 0.2159 0.2530 0.2468 0.2415 0.2354 0.2319 0.2258 0.2642 0.2325 0.2055 0.1995 0.1928 0.2023 0.2649 0.2202 0.1991 0.1900 0.1838 0.1739 0.2455 0.2101 0.1872 0.1805 0.1795 0.1751 0.2498 0.2152 0.1845 0.1797 0.1780 0.1740 0.2134

LRF 0.2857 0.2854 0.2726 0.2687 0.2654 0.2609 0.2897 0.2761 0.2737 0.2685 0.2675 0.2624 0.2912 0.2734 0.2644 0.2564 0.2496 0.2442 0.2816 0.2665 0.2590 0.2513 0.2417 0.2366 0.2609 0.2416 0.2235 0.2112 0.2107 0.1995 0.2540 0.2474 0.2245 0.2205 0.2095 0.2030 0.2527

PRF 0.2823 0.2758 0.2736 0.2746 0.2753 0.2781 0.2657 0.2675 0.2624 0.2647 0.2607 0.2565 0.2590 0.2572 0.2619 0.2566 0.2546 0.2539 0.2599 0.2568 0.2563 0.2538 0.2499 0.2481 0.2562 0.2603 0.2570 0.2504 0.2456 0.2409 0.2564 0.2578 0.2554 0.2515 0.2459 0.2403 0.2590

RSV 0.1524 0.1074 0.0991 0.0901 0.0843 0.0881 0.1432 0.1030 0.0905 0.0880 0.0850 0.0833 0.1525 0.0914 0.0800 0.0747 0.0785 0.0778 0.1437 0.0888 0.0750 0.0724 0.0679 0.0733 0.1420 0.0880 0.0726 0.0649 0.0626 0.0601 0.1781 0.0978 0.0841 0.0709 0.0737 0.0688 0.0932

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1951

Fig. 8 Comparison of term selection methods

Here, selected pool of 5, 15, 25, 35, 45, and 55 terms scored through BNS from which 5 terms having high semantic similarity score are selected for query expansion. The number of documents is changed iteratively 5, 15, 25, 35, 45, and 55 to enhance term pool. This expanded query is again used to retrieve documents through retrieval model BM25. MAP score is calculated of all queries for the purpose of evaluation shown in following table. We analyzed the results of biomedical data retrieval through BNS before and after query expansion. Table 3 and Fig. 9 show that MAP results before query expansion are just named as BNS and after query expansion through semantic similarities of Path based, WuP, and LCH. The MAP score of LCH semantic similarity measure is observed best than other measures. So it is proposed the semantic similarity measuring technique for query expansion in biomedical data retrieval system.

Conclusion Three standard document ranking models BM25, LM Jelinek-Mercer, and LM Dirichlet on Baseline common term scoring technique are evaluated. BM25 found higher-scoring documents ranking model on the biomedical corpus. So this model is selected for further analysis of term scoring and query expansion techniques in the biomedical domain. Here analyzed are nine well-known query-related term scoring techniques, Baseline, BNS, Chi-Square, CoDice, BIM, KLD, LRF, PRF, and RSV, on the biomedical corpus. BNS found the highest term scoring technique to

1952 Table 3 MAP score of expanded query on retrieval

M. Qadeer et al. FD 5

15

25

35

45

55

FT 5 15 25 35 45 55 5 15 25 35 45 55 5 15 25 35 45 55 5 15 25 35 45 55 5 15 25 35 45 55 5 15 25 35 45 55

LCH 0.2599 0.2943 0.2990 0.2743 0.2850 0.2875 0.2946 0.3036 0.2995 0.3124 0.3108 0.3084 0.3202 0.3131 0.3182 0.3150 0.3033 0.2851 0.3130 0.2985 0.3173 0.3002 0.3011 0.3087 0.3238 0.3101 0.3180 0.3237 0.3175 0.2995 0.3272 0.3263 0.3108 0.3009 0.2954 0.3114

Path 0.2436 0.2693 0.2609 0.2672 0.2686 0.2740 0.2864 0.2880 0.2858 0.2815 0.2816 0.2815 0.2945 0.2892 0.2905 0.2870 0.2853 0.2776 0.2975 0.2862 0.2813 0.2805 0.2853 0.2817 0.2975 0.2922 0.2828 0.2883 0.2821 0.2821 0.2896 0.2903 0.2754 0.2721 0.2745 0.2753

WUP 0.2422 0.2715 0.2632 0.2681 0.2691 0.2735 0.2833 0.2849 0.2826 0.2778 0.2815 0.2781 0.2891 0.2861 0.2875 0.2848 0.2844 0.2762 0.2948 0.2833 0.2788 0.2775 0.2812 0.2795 0.2931 0.2893 0.2798 0.2849 0.2772 0.2804 0.2872 0.2873 0.2703 0.2704 0.2708 0.2713

BNS 0.2545 0.2815 0.2738 0.2792 0.2799 0.2850 0.2993 0.2986 0.2973 0.2937 0.2945 0.2934 0.3049 0.3013 0.3030 0.2999 0.2980 0.2900 0.3079 0.2975 0.2921 0.2927 0.2960 0.2934 0.3087 0.3049 0.2928 0.2985 0.2922 0.2944 0.3020 0.3012 0.2857 0.2835 0.2863 0.2861

apply in biomedical domain related to the query terms. Different semantic similarity measures are applied to find most appropriate terms for query expansion. The results of biomedical data retrieval through query expansion with semantic similarity measures are improved particularly through LCH semantic similarity measuring technique.

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1953

Fig. 9 MAP score of expanded query on retrieval

Summary The Biomedical Information Retrieval Process While retrieving information, a user enters a query related to his/her need about required biomedical information. This query is normalized through different methods for exact extraction of related biomedical information. Then a set of queryrelated terms is retrieved by using different retrieval models. These terms can be utilized in query expansion through different techniques, but some of these terms lead toward poor performance or guide toward irrelevant information retrieval. These terms should be filtered during the query expansion process. After a suitable query expansion process, newly formed queries are weighted for further information retrieval. This process is continued till the desired information retrieval of the user. It is showed in the following figure for an understanding of biomedical information retrieval process. The query-relevant documents retrieval with accuracy and performance are a major challenge for almost all information retrieval systems. Terms in some topranked documents which are retrieved in the first iteration can be used to expand the original query. There should be an automatic query expansion technique which is able to reformulate the original query automatically. Since some last decades, the volume of available biomedical data has been increased radically, whereas the rate of biomedical terms of search query remained very low. All the terms extracted

1954

M. Qadeer et al.

from documents may not fit for expanding the query given by the user. Some terms may lead toward incorrect retrieval which is called query drift. Therefore, proper terms selection is very important for improvement in performance of the system. Most of the available methods of query expansion have been broadly investigated. All individual methods have some weaknesses and strengths. We used multiple term selection methods to utilize strengths and overcome the weaknesses of the methods. Different semantic similarity measures are used to select terms which are similar semantically with given query terms. The whole research is done very carefully to get accurate inferences.

Research Contributions This research elaborates the biomedical information retrieval process, explores and evaluates biomedical data sources, analyzes benchmark dataset, and identifies the role of semantic filtering through semantic similarity measures for biomedical query expansion. Probabilistic models are discussed which are based on a network of inferences. Finally, results from evaluations through mean average precisions of some proposed methods are the core part of the contribution for biomedical query expansion. All these are evaluated through a biomedical dataset and showed the results for comparison of models and techniques to reach on conclusions which are best in performance, efficiency, and cost-effectiveness.

Future Work In the future, this research will be enhanced to improve the performance of biomedical information retrieval system by applying new methodologies and improved techniques. There is big margin to enhance the biomedical documents retrieval models and to improve term scoring as well as semantic similarity measuring techniques for query expansion. With the advancement in biomedical field, improved data retrieval will always be the need of the field. This work gives a line-of-sight view to improve biomedical data retrieval systems to meet new challenges.

References Abdulla AAA, Lin H, Xu B, Banbhrani SK (2016) Improving biomedical information retrieval by linear combinations of different query expansion techniques. BMC Bioinf 17(7):238 Alipanah N, Parveen P, Menezes S, Khan L, Seida SB, Thuraisingham B (2010) Ontology-driven query expansion methods to facilitate federated queries. In: 2010 IEEE international conference on service-oriented computing and applications (SOCA). IEEE, pp 1–8 Amati G, Joost C, Rijsbergen V (2003) Probabilistic models for information retrieval based on measuring the divergence from randomness. ACM Transactions on Information Systems (TOIS), 20(4), 357–389

75 Biomedical Data Retrieval Using Enhanced Query Expansion

1955

Basu T, Murthy CA (2016) A supervised term selection technique for effective text categorization. Int J Mach Learn Cybern 7(5):877–892 Bodenreider O (2004) The unified medical language system (UMLS): integrating biomedical terminology. Nucleic Acids Res 32(suppl_1):D267–D270 Cohen T, Roberts K, Gururaj AE, Chen X, Pournejati S, Alter G, Xu H (2017) A publicly available benchmark for biomedical dataset retrieval: the reference standard for the 2016 bioCADDIE dataset retrieval challenge. Database 2017. https://doi.org/10.1093/database/bax061 Demner-Fushman D, Mork JG, Shooshan SE, Aronson AR (2010) UMLS content views appropriate for NLP processing of the biomedical literature vs. clinical text. J Biomed Inform 43(4):587–594 Forman G (2003) An extensive empirical study of feature selection metrics for text classification. J Mach Learn Res 3(Mar):1289–1305 Fujita S (2004) Revisiting again document length hypotheses TREC 2004 genomics track experiments at Patolis. In: TREC Grossman DA, Frieder O (2012) Information retrieval: algorithms and heuristics, vol 15. Springer Science & Business Media, New York Harish BS, Guru DS, Manjunath S (2010) Representation and classification of text documents: a brief review. IJCA, Special Issue on RTIPPR 2:110–119 Harispe S, Sánchez D, Ranwez S, Janaqi S, Montmain J (2014) A framework for unifying ontology-based semantic similarity measures: a study in the biomedical domain. J Biomed Inform 48:38–53 Hiemstra D (2009) Information retrieval models. Information Retrieval: searching in the 21st Century, pp 1–17 Imran H, Sharan A (2009) Thesaurus and query expansion. Int j Comp Sci Infor Technol (IJCSIT) 1(2):89–97 Jerome RN, Giuse NB, Gish KW, Sathe NA, Dietrich MS (2001) Information needs of clinical teams: analysis of questions received by the clinical informatics consult service. Bull Med Libr Assoc 89(2):177 Lavrenko V, Croft WB (2017) Relevance-based language models. In ACM SIGIR Forum, ACM, New York, 51(2):260–267 Lin SM, Huang CM (2017) Personalized optimal search in local query expansion. In ROCLING Lu W, Robertson S, MacFarlane A (2005) Field-weighted XML retrieval based on BM25. In: International workshop of the initiative for the evaluation of XML retrieval. Springer, Berlin/Heidelberg, pp 161–171 Lv Y, Zhai C (2010) Positional relevance model for pseudo-relevance feedback. In: Proceedings of the 33rd international ACM SIGIR conference on research and development in information retrieval. ACM, pp 579–586 Mahdabi P, Crestani F (2014) The effect of citation analysis on query expansion for patent retrieval. Inf Retr 17(5–6):412–429 Manning CD, Raghavan P, Schütze H (2008) Introduction to information retrieval, vol 1, no 1. Cambridge university press, Cambridge, p 496) McInnes BT, Pedersen T, Pakhomov SV (2009) UMLS-Interface and UMLS-similarity: open source software for measuring paths and semantic similarity. In: AMIA annual symposium proceedings, vol 2009. American Medical Informatics Association, p 431 Moffat A, Webber W, Zobel J, Baeza-Yates R (2007) A pipelined architecture for distributed text query evaluation. Inf Retr 10(3):205–231 Pedersen T, Pakhomov SV, Patwardhan S, Chute CG (2007) Measures of semantic similarity and relatedness in the biomedical domain. J Biomed Inform 40(3):288–299 Pérez-Agüera JR, Araujo L (2008) Comparing and combining methods for automatic query expansion. arXiv preprint arXiv:0804.2057 Rada R, Mili H, Bicknell E, Bletner M (1989) Development and application of a metric on semantic nets. IEEE Trans Syst Man Cybern 19(1):17–30 Rivas AR, Iglesias EL, Borrajo L (2014) Study of query expansion techniques and their application in the biomedical information retrieval. Sci World J 2014

1956

M. Qadeer et al.

Robertson S, Zaragoza H (2009) The probabilistic relevance framework: BM25 and beyond. Found Trends Inf Retr 3(4):333–389 Roy D, Paul D, Mitra M, Garain U (2016) Using word embeddings for automatic query expansion. arXiv preprint arXiv:1606.07608 Salton G, Buckley C (1990) Improving retrieval performance by relevance feedback. J Am Soc Info Sci 41:288–297 Singh J, Sharan A (2015) Relevance feedback based query expansion model using borda count and semantic similarity approach. Comput Intell Neurosci 2015:96 Singh J, Sharan A (2018) Rank fusion and semantic genetic notion based automatic query expansion model. Swarm Evol Comput 38:295–308 Slimani T (2013) Description and evaluation of semantic similarity measures approaches. arXiv preprint arXiv:1310.8059 Smiley D, Pugh DE (2011) Apache Solr 3 Enterprise search server. Packt Publishing Ltd., Birmingham Urbain J, Goharian N, Frieder O (2006) IIT TREC 2006: genomics track. In: TREC Wasim M, Khan MUG, Mahmood W (2018) Enhanced biomedical retrieval using discriminative term selection for Pseudo relevance feedback. J Med Imaging Health Inform 8(5):1000–1008 Wei CP, Hu PJH, Tai CH, Huang CN, Yang CS (2007) Managing word mismatch problems in information retrieval: a topic-based query expansion approach. J Manag Inf Syst 24(3):269–295 Xiong C, Callan J (2015) Esdrank: connecting query and documents through external semistructured data. In: Proceedings of the 24th ACM international on conference on information and knowledge management. ACM, pp 951–960 Xu J, Croft WB (2017) Query expansion using local and global document analysis. In ACM SIGIR Forum, ACM, New York, 51(2):168–175 Yang J, Peng W, Ward MO, Rundensteiner EA (2003). Interactive hierarchical dimension ordering, spacing and filtering for exploration of high dimensional datasets. In IEEE Symposium on Information Visualization 2003 (IEEE Cat. No. 03TH8714) (pp 105–112). IEEE.high dimensional datasets. Information Visualization, 2003. INFOVIS 2003. IEEE Symposium on. IEEE, 2003

Machine Learning Applications for The Tensile Property Evaluation of Steel: An Overview

76

Hridayjit Kalita, Kaushik Kumar, and J. Paulo Davim

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemical Composition/Heat Treatment Parameter-Based ML Models for Tensile Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microstructure-Based ML Models for Tensile Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1958 1960 1964 1967 1969 1969 1970 1974

Abstract

With consistent developments and increased application of ML algorithms across different fields, various researches suggested incorporation of the technology to material designing. Evaluation of tensile properties is a significant topic for designers due to its association with complex and large number of processing variables, involvement of variations in constituents, and microstructure volume fractions. ML applications can provide a robust and an interpretable relationship between these complex input variables and output tensile properties. In this chapter, a review on application of ML models based on different input feature space adopted for the evaluation of tensile properties has been provided. In almost all the previous studies concerning tensile properties, ML models have

H. Kalita · K. Kumar () Department of Mechanical Engineering, Birla Institute of Technology, Ranchi, India e-mail: [email protected] J. P. Davim Department of Mechanical Engineering, University of Aveiro, Aveiro, Portugal © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_64

1957

1958

H. Kalita et al.

been integrated with FE, multiscale physical modeling (MPM), and optimization strategies to perform different tasks. These integrations improve the overall performance of the models, making it more reliable, accurate, and robust. Since the variables associated with the production phase are huge and inaccurate, initial dimensionality reduction needs to be carried out employing FE and MPM approach. Instead of eliminating the initial features on the basis of correlation calculation and quality of data as in the case of FE, MPM adopts different approach and has been found to be a more appropriate one. Keywords

Feature engineering · Random forest regression · ANN · Multiscale physical model · Microstructure · Machine learning

Introduction In an era of Industry 4.0, there is a dire need for intelligent steel production that optimizes and reduces the overall cost and time. In order for such accomplishments, intelligent design processes comprising of reliable and real time forecasting of various mechanical properties of special steels, online/remote decision making, and quality assessment need to be adopted (Runde and Bruns 2019; Peters 2017; Hsu et al. 2017). Since artificial intelligence has tremendous contribution in materials science advancements, integration of computing algorithms to materials designing for the desired mechanical properties of steels remained one of the most discussed topics in the field research (Wang et al. 2016; Olson 2013; Wang et al. 2015). Mechanical properties specifically the tensile properties of structural steels are of high priority to be considered in the design and service life evaluation of steels. But the evaluation in terms of their production conditions, chemical constituents, and microstructures is a major challenge due to unreliable, inaccurate, and a large number of variable involvements. During the early stage of computer integration in material design, classical mean field dislocation theory-based traditional physical metallurgical models were used for tensile property predictions which eliminated the repeated manual practice of specimen preparation and testing operations (Wang et al. 2013a, b, Wang et al. 2017a, b). The models basically include the HallPetch model (Hansen 2004; Zhu et al. 2014), Peierls-Nabarro model (Wang et al. 2004), Kocks-Mecking model (Al-Abedy et al. 2018; Hariharan and Barlat 2019), Orowan dislocation looping model (Proville and Bako 2010), and Friedel’s shear cutting model (Friedel 1964) with each having different strengthening mechanisms. These individual strengthening models could also be superimposed for simulation of a collective effect on the yield strength of steels and later came to be known as multimethod strengthening (Wang et al. 2013a, b; Galindo-Nava et al. 2016). Later it was again modified by Wang (Wang et al. 2017a, b) to take into account also the temperature and irradiation effects. Even with all the modifications, strengthening models suffered from complexity in their constitutive equations involving huge number of parameters and inability to accommodate the plasticity behavior of steels (Wang et al. 2017a, b). It was due to which researchers were in constant search

76 Machine Learning Applications for The Tensile Property Evaluation. . .

1959

for models that focused on universality rather than physical complex mechanisms (Tavassoli et al. 2002; Kemp et al. 2006). This resulted in adoption of machine learning (ML) algorithms as tensile property prediction models in Datta et al. (2007), Guo et al. (2019), Datta et al. (2008) with few studies also incorporating optimization algorithms along with ML (Ganguly et al. 2009. Ganguly et al. 2007). ML models when fed to computers can perform different functions and identify physical laws (without any human intervention or manual experiments and complex constitutive equations) using vast amount of available data and rule selection methodologies (Butler et al. 2018). ML models have been used for tensile property predictions of a variety of steels including structural steels, TRIP steels, pipeline steels, etc. (Mohanty et al. 2011; Ganguly et al. 2009; Datta et al. 2008), with almost all the studies focused on prediction based on data obtained either from the production processes/chemical constituents or from the volume fraction (or characteristics) of the microstructures. ANN can be applied to determine the optimized processing parameters and to attain superior mechanical properties by training the complex relationship network between input and output feature space (Bulling et al. 2014; Zhang et al. 2008; Abhyankar et al. 1997; Huang et al. 2019; Meredig and Wolverton 2013). Sidhu (Sidhu et al. 2012) investigated the alloying elements in the low carbon steel for the determination of the bainitic volume fraction using ML techniques. In this chapter, few examples regarding application of ML models in the field of determining tensile properties have been covered considering two aspects of the input feature space: chemical composition/heat treatment and microstructure-based parameters and are described separately in Sects. “Chemical Composition/Heat Treatment Parameter-Based ML Models for Tensile Properties” and “Microstructure-Based ML Models for Tensile Properties”, respectively. Based on the previous work done on this topic, following figures (Figs. 1 and 2) have been given displaying quantities of paper by their types and by their countries.

Documents by type Conference Pape... (22.2%)

Conference Revi... (48.1%)

Article (29.6%)

Fig. 1 Percentage of papers by types. (Source Scopus and web of science)

1960

H. Kalita et al.

Documents by country or territory Compare the document counts for up to 15 countries/territories. China Germany Italy Belgium Brazil Hong Kong Iran Portugal United Kingdom United States 0

0.25 0.5 0.75

1

1.25

1.5 1.75 Documents

2

2.25 2.5 2.75

3 3.25

Fig. 2 No. of papers by countries/territories. (Source Scopus and web of science)

Discussions and inferences on the topic including the suitability of the models in the current scenario of large data availability have been provided in Sect. “Discussions”. Section “Conclusion” gives the future scope of the technology and finally concluding in Sect. “Future Scope”.

Chemical Composition/Heat Treatment Parameter-Based ML Models for Tensile Properties A number of attempts considering the composition and manufacturing conditions for the evaluation of the tensile properties of steels have been performed using analytical and empirical (multiscale) models in previous works (Bok et al. 2015, Cruz-Chávez et al. 2018 , Bohemen 2018). Though these models gave strong correlated predictions of the real situations (with domain knowledge and understanding of the mechanism and numerous variables involved in the production processes), they suffer from increased computational burden (or large iterations) and time. These models are also hindered by missing, incomplete, and unreliable datasets obtained from the various sensors caused due to human negligence and machine failures. All these hurdles make the design of steels a very complicated one that relies on deeper understanding of the metallurgical characteristics at every step of processing.

76 Machine Learning Applications for The Tensile Property Evaluation. . .

1961

Machine learning applications (with or without feature selection methods) can overcome these hurdles with relative ease by providing a robust and an interpretable relationship between the input variables (processing conditions and compositions) and the output variables (tensile properties in this case) (Hosseini et al. 2004; Chokshi et al. 2017). It eliminates the requirement of a large number of orthogonal experiments, which relies on huge data collected from the previously conducted experiments in the literature. A few metal composition/processing parameter-based ML approaches adopted in the previous studies have been covered below. Hosseini et al. (2004) adopted ANN model for determining the mechanical properties of Si-Mn TRIP steel considering the 2-stage heat treatment process conditions and its compositions. Tensile tests were first performed considering the standard ASTM E8 (Dyson and Holmes 1970) with test specimens having a gage length and width of 30 mm and 6 mm, respectively, and 0.001 s−1 as strain rate. True stress and strain curve was obtained from the measured load and elongation data. The input layer of the ANN model involved 10 variables including the heat treatment processing parameters such as intercritical annealing temperature (T1 ), time of annealing (t1 ), bainitic holding temperature (T2 ), and time of holding (t2 ), and the material composition, that is, Si, Mn, C, P, S, and Al, while the output layer included the tensile properties (tensile strength and the percent uniform elongation or UEI%). Out of a total database consisting of 85 vectors, partitioning was done with 60 vectors reserved for training and remaining 25 for testing the network. Normalization of the input-output parameters was done by setting their values within the range between 0 and 1 by the following formula (Eq. 1): XN =

X − Xmin Xmax − Xmin

(1)

where X, Xmin , Xmax , and XN represent the variable value, minimum, maximum, and normalized values of that variable dataset, respectively. Feed forward training of the model is carried out in a supervised manner with tan h activation function and threshold value being associated with each neuron (Eq. 2).    1 − exp −α xj − tj    1 + exp −α xj − tj

(2)

where xj and tj are the weighted output and threshold value of the jth neuron, respectively. Gradient search method was applied as the back propagation algorithm which minimizes the error function (by following the steepest descent curve on the 3D surface of the error) and updates weights and biases in the process. The error function is given by Eq. 3. n   2 dj − oj j =1

(3)

1962

H. Kalita et al.

In the first layer, the nodes varied in the range between 5 and 25 in number, while the second layer nodes were selected based on Hecht-Nielson’s (Hecht-Nielsen 1987 ) and Kolmogorav’s (Kolmogorav 1963) theorem. The whole architecture was trained up to 5 times for each individual parameter and the error during training and testing was found to be decreasing until the hidden layer size reached 40, above which the time of computation is compromised and no further improvement was observed. The robustness and reliability of the ANN model were controlled by the parameters such as the learning constant, initial random distribution of the weights (during start of the training), power of the activation functions, and the number of hidden layer nodes. Optimality of the solutions was evaluated by considering several models out of which the best five were selected based on the ranked error values while calculating the averaged value of the output of those five selected models. Wang et al. (2019a, b) established a universal feature engineering (FE)-based ML model for RAFM steels that considered heat treatment processing/composition features as input parameters and yield strength/elongation features as output parameters. The heat treatment processing variables included the tempering temperature and time, while the composition feature space included the carbon and chromium content. Before machine learning selection of features, 11 features had been chosen after comparing an original database of 100 references comprising of 60 samples of RAFM steel (Ma et al. 2010; Klueh et al. 2002; Chen and Rong 2015; Klueh et al. 1995; Lee et al. 2017; Puype et al. 2018; Tan et al. 2016; Wang et al. 2013a, b). A common strategy is to choose the number of features much smaller than the sample size to prevent problems arising from dimensionality (Lu et al. 2018). Feature engineering is an important step to reduce dimensionality and can be mainly accomplished in two steps, that is, by artificial selection and machine learning (ML) selection. Artificial selection was done by eliminating the features that had poor testing accuracy and data quality, narrow range, and lesser amount of data. This resulted in selection of 11 features from a total 19 original features. In the second step for the ML selection, features that had minimum effect on the predicted values of tensile properties were eliminated using random forest algorithm. Thus, from the selection result, it was found that the tempering temperature, time, and the content of C and Cr had the most significant effect on the yield strength and elongation outputs than all other features. Pearson’s correlation coefficient was applied to find how linearly the pairs of features were correlated with each other. After eliminating the features that had the highest linear correlation, the total number of features left with was seven considering both yield strength and plastic elongation outputs. The selected seven features are then fed to the ML algorithms for evaluating yield strength and elongation. Different supervised learning algorithms or regressors had been analyzed such as Back propagation artificial neural network (Lu et al. 2019) (as also discussed in the previous example), support vector regressor (SVR) (Khemchandani et al. 2018), gradient boosting regressors (GBR) (Li et al. 2018), random forest regressor (RFR) (Ding and Joseph 2017), and Kernel ridge regressor (Yang et al. 2014). Both SVR and RFR generally display enhanced performance for high dimensional data but

76 Machine Learning Applications for The Tensile Property Evaluation. . .

1963

SVR in the current database had R2 correlation for training set greater than that of testing set by 10% which was inappropriate. Also associating decision trees in large number to the RFR algorithm, the over-fitting problem could be overcome much better than SVR. Therefore, RFR had been finally selected as the ML model in their work. As partitioning strategy, 80% of the datasets were reserved for training of the model and the other 20% for testing. Higher than 80% generally leads to underfitting with R2 correlation of the testing set greater than the training set while lower than 80% result in overfitting. It was found that higher yield stress was obtained at a C content ranging in 0.06– 0.08 wt% and at a tempering temperature of 670–760 ◦ C. For higher elongations or enhanced plasticity, a tempering temperature in the range of 750–760 ◦ C had to be maintained with Cr content of 8–9 wt % and tempering time duration of 30–120 min. There is a mention of the significance of including microstructural information in this ML model such as phase fraction, phase growth rate, phase driving force though these were not explored. Wang et al. (2019a, b) adopted another strategy incorporating multiobjective optimization algorithm into FE-ML for enhancing accuracy of the solution space and evaluating the desired mechanical properties of RAFM steels. This integration also enhanced comprehensibility of the mechanical behavior of the steel by considering both tensile and impact toughness properties. The datasets for the ML models were established from the already published references (Ma et al. 2010; Chen and Rong 2015; Klueh et al. 2002; Klueh et al. 1995; Lee et al. 2017; Puype et al. 2018; Tan et al. 2016; Wang et al. 2013a, b) which contained 60 data groups with each having 11 critical features including the composition of the steel, that is, content of C, Cr, Si, Ti, Ta, V, B, W, N, and the processing parameters, that is, the tempering temperature and tempering time. The output parameters include the yield strength and the impact toughness. FE integrated ML models enable selecting the useful features from the initial feature space and discarding the ones that are redundant (Tan et al. 2016). The ML model generalization was enhanced by using standard feature engineering (FE) which reduced the feature space to only seven highest correlated features out of 11 (as explained in the previous example) for training and testing of the model. The 11 features were assessed within the ML selection (as defined in the previous example) by comparing the prediction accuracy quantitatively by the random forest algorithm, thereby eliminating the features that are inefficient. Partitioning of the datasets with 80% as training and 20% as testing sets and normalization of the inputs and the outputs (to set the values within 0–1) was performed. Thus, the prediction of the output variables (yield strength/impact toughness) was made robust using FE guided RFR. The predicted values of the impact toughness and yield strength for RAFM steels showed promising results with FE-based RFR algorithms with better generalization and accuracy. The R2 correlation coefficient exhibited a % higher than 90 for both testing and training sets between the predicted and expected values of the output parameters. The mean absolute deviation (MAD) between the predicted and the expected yield strength was found to be lower than in the previous researches (Wang

1964

H. Kalita et al.

et al. 2017a, b) where FE has not been implemented. Similar MAD result was obtained in case of impact toughness. Along with FE-based ML algorithm, a high throughput optimization algorithm (NSGA-II) with elitist strategy was also designed for the 200 solutions from the previous ML approach each containing all the characteristics of the correlated features. A chain of solutions containing all the information on their feature values were taken as a chromosome. Euler distribution was employed for the evaluation of the degree of crowding of all the 200 solutions around the closest solution. Based on this degree along with the objective values, evaluation and sorting were carried out and first generation pareto front obtained. New solutions were generated then by crossovers and mutations of the parent chromosome using genetic operators and pareto front results. Finally strategy of elitism is applied for selection of the optimal solutions. The design optimal parameters selected after the end of the GA final generation exhibited yield strength greater than 700 MPa and impact toughness greater than 200 J, much higher than the traditional RAFM steels, making the model worthy of experimental verification. A reasonable composition of the RAFM steel was predicted as compared to the original steels and the tensile and impact solutions were accepted based on the spearman correlation coefficient indicating improved comprehensive mechanical properties and facilitating innovations in alloys. The optimal solutions exhibited a prediction error less than 10% with specific ranges in the input features though beyond those ranges, and the prediction error was harder to determine. The ML model can be further made robust, explicable, and rational with the introduction of physical metallurgical constraints which could enable new alloy designs for development.

Microstructure-Based ML Models for Tensile Properties Microstructure generally forms a bridge between composition/heat treatment process parameters and mechanical properties in ML models, providing a better understanding of the material mechanism (Jung et al. 2020). Different models for the analysis of the volume fractions of steel microstructures and their effects on mechanical behavior can be found in the previous studies (Yoozbashi and Yazdani 2010; Jo et al. 2017; Ramazani et al. 2012; Ayatollahi et al. 2016) with or without the ML applications. Attempts have been made to employ structure-based strength calculation model in evaluating the alloy yield strength considering hardening parameters such as grain boundary hardening, precipitation hardening, dislocation hardening, and solid solution hardening (Kim et al. 2019; Sun et al. 2018). But these parameters suffer from limitations in terms of overlapping dependence (Park et al. 2019; Zhou et al. 2019), which has been tackled in the literature employing microstructure-based prediction models. Bouquerel et al. (2006) employed a physical micromechanical model to simulate stress-strain behavior of multiphase transformation-induced plasticity (TRIP) steels considering the customizable microstructure phases such as polygonal ferrite,

76 Machine Learning Applications for The Tensile Property Evaluation. . .

1965

martensite/austenite, and bainitic ferrite. Chowdhury et al. (2016) adopted an automated microstructure recognition system using deep neural network for feature extraction, selection, and classification of microgragh images. Convolution neural network and classifiers (such as SVM, nearest neighbors, random forests) were employed for the purpose of extracting and classifying features, respectively. Different possible combinations of these models (extraction, selection, and classifications) were then adopted based on the performance comparisons. Chokshi et al. (2017) attempted an ANN prediction model to tailor the properties of hot stamped 22MnB5 boron steel considering their phase distributions and their results validated experimentally. Two of the relevant studies conducted in the literature considering ML models for the evaluation of the tensile properties from the microstructure behavior of the material are discussed in detail below. In the study conducted by Jung et al. (2020), ML models in the form of linear regression algorithm and artificial neural network have been proposed to evaluate tensile properties of high strength steels considering the volume fraction of microstructures as input features while also making comparison between the two approaches. The volume fraction of the microstructures are usually experimentally obtained giving due consideration to the chemical compositions/heat treatment process parameters such as reheating temperatures, cooling conditions after rolling the heated metal, cooling rate, starting and finishing of the cooling temperature. The microstructure obtained can be characterized into Bainitic ferrite (BF), polygonal ferrite (PF), Granular bainite (GB), martensite (M), and Acicular ferrite (AF). All these microstructure characteristics generally exhibit different mechanical behavior which results in different mechanical properties of the metal structure as a whole. Big data containing the experimental information from the previous literatures (Zhang et al. 2012, Sung et al. 2016, Sung et al. 2012a, b , Sung et al. 2014, Min et al. 2011, Sung et al. 2013, Lee et al. 2018, Sung et al. 2011a, b , Sung et al. 2015, Sung et al. 2012a, b) on the microstructure volume fractions and corresponding mechanical property data were first accumulated. As the first ML model, linear regression model with gradient descent search methods was employed with the volume fractions of (previously abbreviated) M, PF, GB, AF, and BF as input variables. After training and testing the model, it was found that the elongation or yield ratio (YR) could not be predicted effectively. A lower value of the yield strength (YS) was also observed with increased volume fraction of the PF, while a higher value was observed with AF, BF, and GB. M had a negatively proportional effect on the value of YS, the increased volume fraction of which enhances the YS of the metal while the importance of the magnitude was realized (Costin et al. 2016). The effects of GB and AF on YS were found to be significant and insignificant, respectively, while the significance of PF on YS was more than M. As the second ML model, ANN model for the abovementioned input and output features had been constructed with multiple hidden layers. More number of hidden layers and nodes leads to higher accuracy and less error but at the same time reduces interpretability. The initial weights and biases are assumed first in the forward propagation and then updated during back propagation minimizing errors in the process. Hyper-parameter tuning improves the efficiency of the model (Yilmaz

1966

H. Kalita et al.

and Ertunc 2007; Decost et al. 2017), the optimum combinations of which were determined by stochastic-approximate-based random search algorithm with 300 of it. The model can suffer from local minimum problem at low learning rate (Brahme et al. 2009) while can result in overshooting at higher learning rate. Thus, high learning rate was obtained from larger number of iterations or epochs. The performance and stability of the model was improved by batch normalization (Rahaman et al. 2019; Dutta et al. 2019). For training, testing, and validation, the available database was partitioned in the ratio 8:1:1 (Decost and Holm 2015) and cross-validation was adopted for data reliability considering the small number of available input features. Model loss function and predictability (Myttenaere et al. 2016) were determined by employing the root mean square error (RMSE) and mean absolute percentage error (MAPE) which are given as below:   N 1  RMSE = (yi − ti )2 N

(4)

i=0

1 MAPE = N



 N   ti − yi ti

(5)

i=1

where N, ti , and yi are the total observations, expected, and predicted value. The final optimized model had seven hidden layers, 85 nodes, learning rate of 0.01497, 8121 epochs, and 6.591% minimum MAPE percentage. Thus, the experimentally obtained volume fractions of the microstructures can now be fed into the ANN model for the desired yield strength, tensile strength, and yield ratio. Another work by Jiang et al. (2020) transforms the process parameter/composition feature space to the microstructure feature space using multiscale physical metallurgy models unlike the previous example where microstructures were first obtained experimentally and then fed to ML. Here, the author attempted to compare ML algorithms on the basis of their performance metrics, that is, the mean and maximum relative errors. Also the data and domain knowledge-based strategies for the reduction of initial dimensional feature space were compared on the basis of the accuracy between actual and predicted tensile strength values. Based on the accuracy of the tensile strength predictions with the actual values. In case of the data-based strategy, the production parameter/composition datasets of the pearlitic steel were first collected manually from the different sensors and then taken for the dimensional reduction. The missing values in the raw dataset were inputted by inferring from the knowledge of the known part (Little and Rubin 2019) using univariate feature imputation. The 40 features with 150 instances thus obtained were preprocessed where the low variance features removed (García et al. 2015 ), then scaling by normalizing to values between 0 and 1 in the Scikit-learn python framework and then evaluating the Pearson’s correlation coefficients (Schober et al. 2018) between all the pairs of feature variables. The features that were linearly correlated to a high degree (or a correlation coefficient

76 Machine Learning Applications for The Tensile Property Evaluation. . .

1967

value closer to 1) were eliminated from their consideration in the ML models. Thus, the final numbers of features selected were reduced from 40 to 37. After the feature selection, ML models such as the support vector regressor, gradient tree boosting (GTB) (Cortes and Vapnik 1995; Hoerl and Kennard 1970; Friedman 2002), and ridge regression were applied to relate the input features with the output, that is, the tensile strength. Cross validation to prevent overfitting in the training dataset (Pedregosa et al. 2012), Grid Search CV for parameter tuning, and mean and maximum relative errors (RE) for accuracy were then applied. It was found that all the three ML models exhibited the average maximum RE of value higher than 30% which proved the nonapplicability of data-based strategy for high dimensional feature space and small available data in industries. In the domain knowledge-based strategy, the composition/production parameter feature space was first transformed into microstructure space by applying multiscale integrated physical models such as grain growth, cooling phase transition calculation, dynamic recrystallization, and temperature field models. The final microstructural feature space included the proeutectoid ferrite content, pearlite content, pearlite lamellar spacing, and elemental compositions. The ML regression algorithms adopted in this case were the Multilayer Perceptron (Rumelhart et al. 1988; Hecht-Nielsen 1992; Kingma and Ba 2014), K-Neighbors (Goldberger et al. 2005), random forest (Breiman 2001), bagging (Breiman 1996; Louppe and Geurts 2012), partial least squares (Wegelin 2000), extra trees (Geurts et al. 2006), Gaussian process (Williams and Rasmussen 2006), AdaBoost (Freund and Schapire 1997; Hastie et al. 2009), linear regression (Seber and Lee 2012), SVR, RR, and GTB. Cross validation and parameter tuning were performed as described in the previous example and values for the mean and maximum RE evaluated. In contrary to the data-based strategy, the mean and maximum RE values for all the models in this case were found to be lower than 1.0% and 30%, respectively, thus proving the significance and effectiveness of this approach for higher dimension and small dataset size applications. GTB model was selected and validated with 10 separate samples. It was found that the predicted T.S. values and the actual measured values matched well with improved accuracy.

Discussions The above discussion aims at looking into different ML methodologies that have been employed in the previous literature for the evaluation of tensile properties of steels. Basically two different approaches were commonly observed based on the initial database source of model inputs. Some researches attempted to include data in the input feature layers of the ML models directly from the production/processing phase. While some other researches concentrated on first collecting the data about the formation of different types of microstructures from various production process parameters and compositions either manually by experiments or by applying multiscale integrated physical models and then using these microstructure data as input features in the ML model.

1968

H. Kalita et al.

In the section concerning evaluation of tensile properties directly from the processing/ composition parameters, two of the patterns are observed: i. Directly obtaining the solution space (tensile property data) from the large processing/composition datasets. ii. Obtaining the solution space from the processing/composition datasets and then optimizing the input features (processing variables and composition) by applying optimization algorithms for the desired output features (tensile properties). Though the tensile properties can be suitably evaluated using the processing/composition data as input features in ML models, the method is not robust and accurate as a large number of variables are associated and the parameters can easily be affected by sensor malfunction and human errors. Evaluation of tensile properties from the microstructure point of view proves to be a much better approach in this regard which helps in enhancing robustness, predictability, and efficiency in reducing feature dimensionality. In the third section of this chapter, the ML models concerning the evaluation of tensile properties from the microstructure point of view have been discussed which can mainly be subcategorized into two approaches: i. Obtaining the microstructure volume fractions from the corresponding processing parameters and composition by conducting experiments or using previously experimented data and then using the microstructure dataset as the input feature space in the ML models. ii. Transforming the production/composition dataset into microstructure feature space using various multiscale physical models (MPM) and then utilizing the microstructure data as inputs to the ML models. ML models employed in the literature for the evaluation of tensile properties and that has been discussed in the current review work include ANN, FE-RFR, FE-RFR + NSGA-II, linear regression with gradient descent search and MPM integrated GTB. Considering all models, it can be inferred that MPM integrated ML models area better option than feature engineering (FE) integrated models in terms of techniques for reduction in initial dimensionality of the complex production input space. FE-integrated models generally reduce the initial dimension of the vast amount of production data by determining the effects of each one of them on the final tensile outputs, by eliminating the low quality features (or accuracy) and that are highly linearly correlated using Pearson’s correlation coefficient. On the other hand, MPM integrated ML models reduce dimensionality by relying on the actual physical mechanism between the high dimension production inputs and the microstructure characteristic outputs with the involvement of few predefined model parameters that can be systematically fitted. Though the application of only MPM model for determination of the mechanical behavior from complex microstructures of steels is not feasible (Jung et al. 2020), an MPM integrated ML model can perfectly yield a greater accuracy, efficiency, and robustness in comparison to all other ML models. MPM-ML models are always suitable to be applied in large industries where situations regarding high dimension production

76 Machine Learning Applications for The Tensile Property Evaluation. . .

1969

datasets and standardized data values (for consistent product quality and leading to lesser variation range in the parameter values) are usually encountered (Jiang et al. 2020).

Conclusion In the current chapter, two different approaches of implementing ML techniques for the evaluation of tensile properties in steels have been discussed based on the type of input feature space adopted. In the first approach, the high dimensional space and large volume of processing parameters and composition data were used as the input feature space for the ML algorithms. In the second approach, the processing parameters and composition datasets were first reduced to a lower dimensional space comprising of microstructure volume fraction data which were then used as the input feature space for ML algorithms. The conversion from a higher dimensional space to lower dimensional space was basically done by integrating feature engineering and multiscale physical modeling techniques to ML models of which the MPM technique has been found to be the most robust, accurate, and efficient one. This approach of handling complex data eliminated the problems encountered in industrial scenarios for higher dimensions and small size datasets. Few of the ML algorithms that were found suitable and common in these applications were the RFR, ANN, linear regression with gradient descent search, and gradient tree boosting.

Future Scope Tensile testing procedures have always been a costly and a time-consuming affair in industries. With the increase in demand for improved and customizable properties of advanced materials, these procedures also cannot be avoided and remain an integral part of material designing. As is discussed above, a large number of variables are involved which influences the tensile behavior of materials including the processing parameters, compositions, and volume fractions of microstructures. ML applications can simplify such complex conditions by finding patterns automatically between the input features and output features with the available big data from experimental results. ML can also be integrated with multiscale physical models (Jiang et al. 2020) and optimization techniques (Wang et al. 2019a, b) to make the model more robust, globally optimized and predictions accurate. This can facilitate industrial product designers to simulate evaluation of mechanical properties of materials from the production parameter inputs without the need to perform any traditional physical tests and without any domain expertise. The entire evaluation can be performed in real time online-based system (Jiang et al. 2020) from a remote location without the need of any human intervention. Domain knowledge-based (MPM-ML integration) strategy in comparison to data-based strategy as adequately demonstrated in the study (Jiang et al. 2020) proved to be a superior approach for initial dimensionality reduction with decrease in output relative errors and better

1970

H. Kalita et al.

compliance with the real scenarios. Remote virtual mechanical labs implementing this technology can be employed to impart visual education to engineering students in accordance with increased demand of industries in the setup of Industry 4.0 (Grodotzki et al. 2018).

References Abhyankar A, Copeland LS, Wong W (1997) Uncovering nonlinear structure in real-time stockmarket indexes: the S&P 500, the DAX, the Nikkei 225, and the FTSE-100. J Bus Econ Stat 15:1–14 Al-Abedy HK, Jones IA, Sun W (2018) Small punch creep property evaluation by finite element of Kocks-Mecking-Estrin model for P91 at elevated temperature. Theor Appl Fract Mec 98: 244–254 Ayatollahi MR, Darabi AC, Chamani HR, Kadkhodapour J (2016) 3D micro-mechanical modeling of failure and damage evolution in dual phase steel based on a real 2D microstructure. Acta Mech Solida Sin 29:95–110 Bohemen S (2018) Exploring the correlation between the austenite yield strength and the bainite lath thickness. Mater Sci Eng A 731:119–123 Bok HH, Kim SN, Suh DW, Barlat F, Lee MG (2015) Non-isothermal kinetics model to predict accurate phase transformation and hardness of 22MnB5 boron steel. Mater Sci Eng A 626: 67–73 Bouquerel J, Verbeken K, Cooman BCD (2006) Microstructure-based model for the static mechanical behavior of multiphase steels. Acta Mater 54:1443–1456 Brahme A, Winning M, Raabe D (2009) Prediction of cold rolling texture of steels using an artificial neural network. Comput Mater Sci 46:800–804 Breiman L (1996) Bagging predictors. Mach Learn 24(2):123–140 Breiman L (2001) Random forests. Mach Learn 45(1):5–32 Bulling A, Blanke U, Schiele B (2014) A tutorial on human activity recognition using body-worn inertial sensors. ACM Comput Surv 46:1–33 Butler KT, Davies DW, Cartwright H, Isayev O, Walsh A (2018) Machine learning for molecular and materials science. Nature 559(7715):547–555 Chen S, Rong L (2015) Effect of silicon on the microstructure and mechanical properties of reduced activation ferritic/martensitic steel. J Nucl Mater 459:13–19 Chokshi P, Dashwood R, Hughes DJ (2017) Artificial neural network (ANN) based microstructural prediction model for 22MnB5 boron steel during tailored hot stamp- ing. Comput Struct 190:162–172 Chowdhury A, Kautz E, Yener B, Lewis D (2016) Image driven machine learning methods for microstructure recognition. Comput Mater Sci 123:176–187 Cortes C, Vapnik V (1995) Support-vector networks. Mach Learn 20(3):273–297 Costin WL, Lavigne O, Kotousov A (2016) A study on the relationship between microstructure and mechanical properties of acicular ferrite and upper bainite. Mater Sci Eng A 663:193–203 Cruz-Chávez MA, Serna-Barquera SA, Juárez-Chávez J, Romero R, Cruz-Rosales MH, CampilloIllanes B (2018) Optimization-heuristic of mechanical properties of acicular ferrite steel. Mater Sci Eng A 721:65–73 Datta S, Pettersson F, Ganguly S, Saxén H, Chakraborti N (2007) Designing high strength multiphase steel for improved strength–ductility balance using neural networks and multi-objective genetic algorithms. ISIJ Inter 47:1195–1203 Datta S, Pettersson F, Ganguly S, Saxén H, Chakraborti N (2008) Identification of factors governing mechanical properties of TRIP-aided steel using genetic algorithms and neural networks. Mater Manuf Process 23:130–137

76 Machine Learning Applications for The Tensile Property Evaluation. . .

1971

Decost BL, Holm EA (2015) A computer vision approach for automated analysis and classification of microstructural image data. Comput Mater Sci 110:126–133 Decost B, Jain H, Rollett A, Holm EA (2017) Exploring the microstructure manifold: image texture representations applied to ultrahigh carbon steel microstructures. JOM 133:30–40 Ding J, Joseph ZB (2017) MethRaFo: MeDIP seq methylation estimate using a random forest regressor. Bioinformatics 33:3477–3479 Dutta T, Dey S, Datta S, Das D (2019) Designing dual-phase steels with improved performance using ANN and GA in tandem. Comput Mater Sci 157:6–16 Dyson DJ, Holmes B (1970) Effect of alloying additions on the lattice parameter austenite. J Iron Steel Inst 208:469 Freund Y, Schapire RE (1997) A decision theoretic generalization of online learning and an application to boosting. J Comput Syst Sci 55(1):119–139 Friedel J (1964) Dislocations. Pergamon Press, Oxford, pp 223–226 Friedman JH (2002) Stochastic gradient boosting. Comput Stat Data An 38(4):367–378 Galindo-Nava EI, Rainforth WM, Rivera-Díaz-del-Castillo PEJ (2016) Predicting microstructure and strength of maraging steels: elemental optimization. Acta Mater 117:270–285 Ganguly S, Datta S, Chakraborti N (2007) Genetic algorithms in optimization of strength and ductility of low-carbon steels. Mater Manuf Process 22:650–658 Ganguly S, Datta S, Chakraborti N (2009) Genetic algorithm-based search on the role of variables in the work hardening process of multiphase steels. Comput Mater Sci 45:158–166 García S, Luengo J, Herrera F (2015) Data preprocessing in data mining. Springer Geurts P, Ernst D, Wehenkel L (2006) Mach Learn 63(1):3–42 Goldberger J, Hinton GE, Roweis ST, Salakhutdinov RR (2005) Neighbourhood components analysis. Adv Neural Inf Proces Syst:513–520 Grodotzki J, Ortelt TR, Tekkaya AE (2018) Remote and virtual labs for engineering education 4.0: achievements of the ELLI project at the TU Dortmund University. Procedia Manufact 26:1349– 1360. https://doi.org/10.1016/j.promfg.2018.07.126 Guo S, Yu J, Liu X, Wang C, Jiang Q (2019) A predicting model for properties of steel using the industrial big data based on machine learning. Comput Mater Sci 160:95–104 Hansen N (2004) Hall-petch relation and boundary strengthening. Scr Mater 51:801–806 Hariharan K, Barlat F (2019) Modified Kocksmecking estrin model to account non-linear strain hardening. Metall Mater Trans A 50A:513–517 Hastie T, Rosset S, Zhu J, Zou H (2009) Statistics and its interface stat. Interface 2(3):349–360 Hecht-Nielsen R (1987) Kolmogorov’s mapping neural networks existence theorem. Proceeding of the first IEEE international conference on neural networks, vol. 3, San Diego, pp. 112–114 Hecht-Nielsen R (1992) Neural networks for perception. Elsevier, pp 65–93 Hoerl AE, Kennard RW (1970) Ridge regression: biased estimation for non-orthogonal problems. Technometrics 12(1):55–67 Hosseini S, Zarei-Hanzaki A, Panah M, Yue S (2004) ANN model for prediction of the effects of composition and process parameters on tensile strength and percent Elon- gation of Si–Mn TRIP steels. Mater Sci Eng A 374:122–128 Hsu CY, Kang LW, You TY, Jhong WC (2017) Vision-based automatic identification tracking of steel products for intelligent steel manufacturing, 2017 IEEE International Symposium on Multimedia (ISM), IEEE, pp. 376–377 Huang W, Martin P, Zhuang HL (2019) Machine-learning phase prediction of high-en- tropy alloys. Acta Mater 169:225–236 Jiang X, Jia B, Zhang G, Zhang C, Wang X, Zhang R, Yin H, Qu X, Song Y, Su L, Mi Z, Hu L, Ma H (2020) A strategy combining machine learning and multiscale calculation to predict tensile strength for pearlitic steel wires with industrial data. Scr Mater 186:272–277 Jo MC, Lee SG, Sohn SS, Kim KS, Kim WK, Lee CS, Lee S (2017) Effects of coiling temperature and pipe-forming strain on yield strength variation after ERW pipe forming of API X70 and X80 linepipe steels. Mater Sci Eng A 682:304–311

1972

H. Kalita et al.

Jung ID, Shin DS, Kim D, Lee J, Lee MS, Son HJ, Reddy NS, Kim M, Moon SK, Kim KT, Yu JH, Kim S, Park SJ, Sung H (2020) Artificial intelligence for the prediction of tensile properties by using microstructural parameters in high strength steels. Materialia 11:100699 Kemp R, Cottrell GA, Bhadeshia HKDH, Odette GR, Yamamoto T, Kishimoto H (2006) Neural-network analysis of irradiation hardening in low-activation steels. J Nucl Mater 348: 311–328 Khemchandani R, Goyal K, Chandra S (2018) Generalized eigenvalue proximal support vector regressor for the simultaneous learning of a function and its derivatives. Int J Mach Learn Cyb 9:2059–2070 Kim YK, Choe J, Lee KA (2019) Selective laser melted equiatomicCoCrFeMnNi high-entropy alloy: microstructure, anisotropic mechanical response, and multiple strengthening mechanism. J Alloys Compd 805:680–691 Kingma, D.P., Ba, J., (2014) Adam: A method for stochastic optimization. arXiv preprint arXiv (2014) Klueh RL, Kai JJ, Alexander DJ (1995) Microstructure mechanical properties correlation of irradiated conventional and reduced-activation martensitic steel. J Nucl Mater 225:175–186 Klueh RL, Alexander DJ, Sokolov MA (2002) Effect of chromium, tungsten, tantalum and boron on mechanical properties of 5-9 Cr-WVTab steels. J Nucl Mater 304:139–152 Kolmogorav AN (1963) On the representation of continuous functions of many variables by superposition of continuous functions of one variable and addition math. Soc Trans 28:55 Lee CH, Park JY, Seol WK, Moon J, Lee TH, Kang NH, Kim HC (2017) Microstructure and tensile and charpy impact properties of reduced activation ferritic martensitic steel with Ti. Fusion Eng Des. 124:953–957 Lee DH, Sohn SS, Song H, Ro Y, Lee CS, Lee S, Hwang B (2018) Effects of start and finish cooling temperatures on the yield strength and uniform elongation of strain-based API X100 pipeline steels. Metall Mater Trans A 49:4536–4543 Li X, Li W, Xu Y (2018) Human age prediction based on DNA methylation using a gradient boosting regressor. Genes 9(9):424 Little RJ, Rubin DB (2019) Statistical analysis with missing data. John Wiley & Sons Louppe G, Geurts P (2012) Ensembles on random patches, Joint European conference on machine learning and knowledge discovery in databases. Springer, pp 346–361 Lu S, Zhou Q, Ouyang Y, Guo Y, Li Q, Wang J (2018) Accelerated discovery of stable lead free hybrid organic-inorganic perovskites via machine learning. Nat Commun 9:3405 Lu FY, Yin ZQ, Wang C, Cui CH, Teng J, Wang S, Chen W, Huang W, Xu BJ, Guo GC, Han ZF (2019) Parameter optimization and real-time calibration of a measurement-device-independent quantum key distribution network based on a back propagation artificial neural network. J Opt Soc Am B 36:B92–B98 Ma R, Yang Y, Yan Q, Yang Y, Li X, Ge C (2010) Effect of alloying on the property of 9Cr low activation martensitic steels. Acta Metall Sin –Engl 23:451–460 Meredig B, Wolverton C (2013) A hybrid computational-experimental approach for au- tomated crystal structure solution. Nat Mater 12:123–127 Min KZ, Bae JH, Kim K, Lee DJ (2011) Effects of Mo and Nb on tensile and Charpy impact properties of API X80 linepipe steels. Kor J Met Mater 49:766–773 Mohanty I, Bhattacharjee D, Datta S (2011) Designing cold rolled IF steel sheet with optimized tensile properties using ANN and GA. Comput Mater Sci 50:2331–2337 Myttenaere A, Golden B, Grand B, Rossi F (2016) Mean absolute percentage error for regression models. Neurocomputing 192:38–48 Olson GB (2013) Genomic materials design: the ferrous frontier. Acta Mater 61:771–781 Park JM, Choe J, Kim JG, Bae JW, Moon J, Yang JS, Kim KT, Yu JH, Kim HS (2019) Superior tensile properties of 1%C–CoCrFeMnNi high-entropy alloy additively manufactured by selective laser melting. Mater Res Lett 8:1–7 Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V, Vanderplas J, Passos A, Cournapeau D, Brucher M, Perrot M, Duchesnay E, Louppe G (2012) Scikit-learn: machine learning in python. J Mach Learn Res 12

76 Machine Learning Applications for The Tensile Property Evaluation. . .

1973

Peters H (2017) How could Industry 4.0 transform the steel industry. Future Steel Forum, Steel Times International, Warsaw Proville L, Bako B (2010) Dislocation depinning from ordered nanophases in a model fcc crystal: from cutting mechanism to Orowan looping. Acta Mater 58:5565–5571 Puype A, Malerba L, De Wispelaere N, Petrov R, Sietsma J (2018) Effect of W and N on mechanical properties of reduced activation ferritic/martensitic EUROFER based steel grades. J Nucl Mater 502:282–288 Rahaman M, Mu W, Odqvist J, Hedström P (2019) Machine learning to predict the martensite start temperature in steels. Metall Mater Trans A 50:2081–2091 Ramazani A, Mukherjee K, Prahl U, Bleck W (2012) Modelling the effect of microstructural banding on the flow curve behavior of dual-phase (DP) steels. Comput Mater Sci 52:46–54 Rumelhart DE, Hinton GE, Williams RJ (1988) Learning representations by back-propagating errors. Cognitive Modeling 5(3):1 Runde W, Bruns M (2019) Hi-tech steel production planning. Steel Times Int 43(1):35–42 Schober P, Boer C, Schwarte LA (2018) Correlation coefficient: appropriate use and interpretation. Anesth Analg 126(5):1763–1768 Seber GA, Lee AJ (2012) Linear regression analysis. John Wiley & Sons Sidhu G, Bhole SD, Chen DL, Essadiqi E (2012) Development and experimental validation of a neural network model for prediction and analysis of the strength of bainitic steels. Mater Des 41:99–107 Sun S, Tian Y, Lin H, Yang H, Dong X, Wang Y, Zhang Z (2018) Transition of twinning behavior in CoCrFeMnNi high entropy alloy with grain refinement. Mater Sci Eng A 712:603–607 Sung HK, Shin SY, Hwang B, Lee CG, Kim NJ, Lee S (2011a) Effects of carbon equivalent and cooling rate on tensile and Charpy impact properties of high-strength bainitic steels. Mater Sci Eng A 530:530–538 Sung HK, Shin SY, Hwang B, Lee CG, Kim NJ, Lee S (2011b) Effects of rolling and cooling conditions on microstructure and tensile and Charpy impact properties of ultra-low-carbon highstrength bainitic steels. Metall Mater Trans A 42:1827–1835 Sung HK, Shin SY, Hwang B, Lee CG, Lee S (2012a) Effects of B and Cu addition and cooling rate on microstructure and mechanical properties in low-carbon, high-strength bainitic steels. Metall Mater Trans A 43:3703–3714 Sung HK, Sohn SS, Shin SY, Lee S, Kim NJ, Chon SH, Yoo JY (2012b) Effects of finish rolling temperature on inverse fracture occurring during drop weight tear test of API X80 pipeline steels. Mater Sci Eng A 541:181–189 Sung HK, Shin SY, Hwang B, Lee CG, Lee S (2013) Effects of cooling conditions on microstructure, tensile properties, and Charpy impact toughness of low-carbon high-strength bainitic steels. Metall Mater Trans A 44:294–302 Sung HK, Lee S, Shin SY (2014) Effects of start and finish cooling temperatures on microstructure and mechanical properties of low-carbon high-strength and low-yield ratio bainitic steels. Metall Mater Trans A 45:2004–2013 Sung HK, Lee DH, Shin SY, Lee S, Jung YY, Hwang B (2015) Effect of finish cooling temperature on microstructure and mechanical properties of high-strength bainitic steels containing Cr, Mo, and B. Mater Sci Eng A 624:14–22 Sung HK, Lee DH, Lee S, Kim HS, Ro Y, Lee CS, Hwang B, Shin SY (2016) Correlation between microstructures and tensile properties of strain-based API X60 pipeline steels. Metall Mater Trans A 47:2726–2738 Tan L, Snead LL, Katoh Y (2016) Development of new generation reduced activation ferriticmartensitic steels for advanced fusion reactors. J Nucl Mater 478:42–49 Tavassoli AAF, Rensman JW, Schirra M, Shiba K (2002) Material design data for reduced activation martensitic steel type F82H. Fusion Eng Des 61-62:617–628 Wang G, Strachan A, Ça T, in, GoddardIii, W.A (2004) Model. Simul Mater Sc 12:S371–S389 Wang P, Chen J, Fu H, Liu S, Li X, Xu Z (2013a) Effect of N in the precipitation behavior of the reduced activation ferritic/martensitic steel CLF-1 after thermal ageing. J Nucl Mater 442: S9–S12

1974

H. Kalita et al.

Wang JS, Mulholland MD, Olson GB, Seidman DN (2013b) Prediction of the yield strength of a secondary-hardening steel. Acta Mater 61:4939–4952 Wang C, Zhang C, Yang Z, Su J, Weng Y (2015) Multi-scale simulation of hydrogen influenced critical stress intensity in high Co–Ni secondary hardening steel. Mater Des 87:501–506 Wang C, Zhang C, Yang Z, Su J, Weng Y (2016) Microstructure analysis and yield strength simulation in high Co–Ni secondary hardening steel. Mater Sci Eng A 669:312–317 Wang C, Zhang C, Yang Z, Zhao J (2017a) Multiscale simulation of yield strength in reducedactivation ferritic/martensitic steel. Nucl Eng Technol 49:569–575 Wang C, Zhang C, Zhao J, Yang Z, Liu W (2017b) Microstructure evolution and yield strength of CLAM steel in low irradiation condition. Mater Sci Eng A 682:563–568 Wang C, Shen C, Cui Q, Zhang C, Xu W (2019a) Tensile property prediction by feature engineering guided machine learning in reduced activation ferritic/martensitic steels. J Nucl Mater S0022– 3115:30667–30661 Wang C, Shen C, Huo X, Zhang C, Xu W (2019b) Design of comprehensive mechanical properties by machine learning and high-throughput optimization algorithm in RAFM steels. Nucl Eng Technol S1738–5733:30686–30682 Wegelin JA (2000) A survey of partial least square method with emphasis on the two block case. University of Washington, Tech. Rep Williams CK, Rasmussen CE (2006) Gaussian processes for machine learning. MIT Press Cambridge Yang S, Yang L, Feng Z, Wang M, Jiao L (2014) Neurocomputing 134:173–180 Yilmaz M, Ertunc HM (2007) The prediction of mechanical behavior for steel wires and cord materials using neural networks. Mater Des 28:599–608 Yoozbashi MN, Yazdani S (2010) Mechanical properties of nanostructured, low temperature bainitic steel designed using a thermodynamic model. Mater Sci Eng A 527:3200–3205 Zhang YM, Yang S, Evans JRG (2008) Revisiting Hume-Rothery’s rules with artificial neural networks. Acta Mater 56:1094–1105 Zhang X, Gao H, Zhang X, Yang Y (2012) Effect of volume fraction of bainite on microstructure and mechanical properties of X80 pipeline steel with excellent de- formability. Mater Sci Eng A 531:84–90 Zhou R, Liu Y, Liu B, Li J, Fang Q (2019) Precipitation behavior of selective laser melted FeCoCrNiC0.05 high entropy alloys. Intermetallics 106:20–25 Zhu YZ, Wang SZ, Li BL, Yin ZM, Wan Q, Liu P (2014) Grain growth and microstructure evolution based mechanical property predicted by a modified Hall-Petch equation in hot worked Ni76Cr19AlTiCo alloy. Mater Des 55:456–462

Further Reading http://industryofthingsvoice.com/wpcontent/uploads/2017/10/steel_whitepaper.pdf https://www.aist.org/AIST/aist/AIST/Publications/Digital%20Transformations/19-dec-digitaltransformation.pdf

77

Business Ecosystem Approach to Industry 4.0

Daniel Alejandro Rossit, Marisa Analía Sánchez, Fernando Tohmé, and Mariano Frutos

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background Concepts from Management Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Business Ecosystems and Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Platform-Based Business Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background Concepts from the Manufacturing Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . Cyber-Physical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cloud Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Big Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Digital Twins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cloud Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industrial Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Factory-as-a-Platform Business Model Proposal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Manufacturing Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Limitations of Platform Interactions in Smart Factories . . . . . . . . . . . . . . . . . . . . . . . . . . . Illustrative Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Final Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1976 1978 1978 1979 1981 1981 1981 1982 1983 1984 1987 1988 1990 1991 1991 1992 1994 1994 1994

D. A. Rossit () Department of Engineering, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina e-mail: [email protected] M. A. Sánchez Departamento de Ciencias de la Administración, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina F. Tohmé Departamento de Economía, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina M. Frutos Departamento de Ingeniería, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_65

1975

1976

D. A. Rossit et al.

Abstract

Industry 4.0 creates opportunities for defining new ways of creating value. Traditional manufacturing business models should be transformed to embrace its benefits. This transformation has the potential to induce changes not only in the shop-floor operation, but also on the strategies and processes supporting value creation activities. Most of the research focuses on engineering issues but not so much on how manufacturing firms shape their business, redefining their role in the business ecosystem, capturing more value thanks to network effects. The goal of this chapter is to lay the ground lines for a comprehensive business model for manufacturing firms, based on Industry 4.0 technologies. A factoryas-a-platform business model is proposed, illustrating it with a concrete example. The description of the platform-based business model embodies the answers to questions about how to make decisions regarding the core interaction among participants, how to grow up the platform user base, how to monetize network effects, and to what extent should the platform be open.

Keywords

Ecosystem · Platform · Complementary · Industry 4.0 · Manufacturing · Production

Introduction The widespread adoption of emerging technologies has led to a paradigm shift in the business models of firms. While in the industrial era, giant companies relied on supply-side economies of scale, most Internet era giants run demand-side economies of scale. That is, in a world of network effects, the relationships with users constitute the new sources of competitive advantage and market dominance (Van Alstyne et al. 2016). This new competitive environment is described in the management literature as an “ecosystem” comprising a platform owner and complementors that enhance the value of the platform (Ceccagnoli et al. 2012; Gawer and Cusumano 2008). The platform owner enables complementors to join the platform via shared or open-source technologies and/or technical standards (Jacobides et al. 2018). By connecting to loosely affiliated ecosystems, firms are able to create a global network of partners they do not even know beforehand. This makes it possible to generate highly valuable products and services for their customers (Parker et al. 2016a). Smart factories enjoy the benefits of customized manufacturing (Yao et al. 2017; Wang et al. 2017) sometimes enhanced with platform-mediated interactions between the complementors and end users (Adner 2006). The traditional phases of design, fabrication, execution, and service of traditional industrial plants can benefit from these new forms of collaboration. The penetration of Internet technologies into

77 Business Ecosystem Approach to Industry 4.0

1977

Industry 4.0 systems can go far beyond its uses in production flow processes. There exists a large literature on Industry 4.0, mostly focusing on its engineering aspects. Research on how manufacturing firms shape their business, redefine their role in the business ecosystem, and build on network effects to capture more value is, instead, lacking. In this chapter it is shown how factories using Industry 4.0 can shift from a manufacturing to a factory-as-a-platform business model. The challenges posed by such redefinitions require strategies for the creation of value for the entire ecosystem, instead of merely for the firm. Adamson et al. (2017) state that cloudbased manufacturing models (Xu 2012) will require the development of new business models. Traditional hierarchical models will no longer be competitive since massive online exchanges will largely improve the agility and innovativeness of production processes, becoming able to quickly satisfy the demands of customers (Rossit et al. 2019a). New models should facilitate the participation of dynamically structured fabrication entities through online collaborations, as shown in Fig. 1. The purpose of this work is to present a factory-as-a-platform business model intended for factories using Industry 4.0 technologies and based on a platform approach. This business model is consistent with the design principles of the cloudbased design and manufacturing (CBDM) architecture (Wu et al. 2015), which covers a whole spectrum of aspects, ranging from the design of prototypes to the production of final goods. In this way, factory-as-a-platform model extends the potentialities of Industry 4.0 by defining how a smart factory can develop a platform as a means to coordinate value creation.

Fig. 1 Important milestones in business models evolution

1978

D. A. Rossit et al.

In order to approach the problem of defining a business model it is necessary to take into account research on both management and manufacturing. Sect. “Background Concepts from Management Literature” presents the concept of business model and the relevant decisions to be made in a platform-mediated setting. Sect. “Background Concepts from the Manufacturing Literature” briefly describes technologies in Industry 4.0 to understand its potential. In addition, cyberphysical systems and cloud manufacturing are introduced with a focus on the design requirements of a smart factory. The factory-as-a-platform business model proposal is described in Sect. “Factory-as-a-Platform Business Model Proposal”. Before concluding the chapter, Sect. “Final Remarks” discusses the contributions.

Background Concepts from Management Literature Business Ecosystems and Platforms A business ecosystem is an economic community in which a variety of interrelated stakeholders coevolve. It creates value and brings competitive advantages to the participating companies by initiating, identifying, and integrating its stakeholders (Rong et al. 2018). A platform provides the infrastructure and rules for a marketplace, bringing together producers and consumers. Platforms comprise four types of players engaged in value-creating interactions. The owners of the platforms control their intellectual property and governance (for instance, Google owning Android). Providers serve as the platforms’ interface with users (mobile phone companies support devices that run Android). Producers create their products (e.g., Android apps), and consumers use them (Van Alstyne et al. 2016). When a platform is opened to external contributions, the demand for the owner’s products goes up thanks to the complementarities with the demands for the products of other participants. Platforms do thus take advantage of those network effects, harnessing the collective power of the crowd. Opening up platforms provide more benefits to the owner, by creating a greater volume and variety of contributions, motivations, and ideas than the owner alone could have mustered. These contributions increase consumer surplus and push up the demand curve for complementary products. Owners also infer the customers’ preferences up from the data (McAfee and Brynjolfsson 2017). Consider as an example the Apple operating system iOS. Each free app has the effect of shifting the demand curve of iPhones outward, increasing the number of people who are willing to pay a higher price. Companies in a business ecosystem do not only work cooperatively and competitively but also coevolving around new innovations to satisfy the needs of customers (Adner 2006; Rong et al. 2010; Bengtsson and Raza-Ullah 2016). Co-creation refers here to settings in which communities produce marketable value in voluntary activities mediated through platforms, conducted independently of any established organization (Karhu et al. 2011). Companies are figuring out how to take advantage of crowdsourcing their problems to the contributing participants in the value chain (McAfee and Brynjolfsson 2017). Crowdsourcing is defined by Jeff Howe (2006)

77 Business Ecosystem Approach to Industry 4.0

1979

as the act of taking a job traditionally performed by a designated agent (usually an employee) and handing it over to an undefined, generally large, group of people in the form of an open call. There are different types of crowdsourcing: cloud labor (Amazon’s Mechanical Turk), crowd creativity (YouTube), distributed knowledge (Wikipedia), open innovation (Innocentive.com), crowdfunding, etc. Cloud labor services constitute a particularly useful form of crowdsourcing. A coordination platform serves as an interface between requesters who need to get work done and a large crowd of workers willing to carry it out (Kern 2014). Sometimes requesters do not want to bring together an entire crowd but just to find out, as quickly and efficiently as possible, the right person or team that may help with some tasks.

Platform-Based Business Models The notion of business models in the literature and the practice of management is quite relevant since it expresses the business logic of specific firms. Osterwalder et al. (2005) identify the most common building blocks among business models that are structured in a Business Model Canvas. From this synthesis, four pillars emerge that describe the value proposition, the customer interface (target customer, distribution channels, and the kind of links between the company and the customers), the management infrastructure (activities, core competencies, and partners network), and financial aspects (cost structure and revenue model) (Osterwalder et al. 2005). The concept was conceived from a firm-centered perspective (Fehrer et al. 2018). This means that a model definition is based on a value proposition that does not consider the firm as being part of an ecosystem in which the way of creating value does not restrict to the inside of the organization. In this work, it is adopted a platform-based conceptualization of a business model based on a servicedominant logic proposed by Fehrer et al. (2018), and Parker (Parker et al. 2016b) and Cusumano et al. (2019) research on platform creation. Parker et al. (2016b) embodies the platform business model in a set of design decisions. The first decision is about the core interactions among the participants and the value units (information, goods or services, and some form of currency). Information may refer to prices, results of a search on eBay, or the availability of a driver in Uber. Services are provided inside the platform or outside, as is the case of an Airbnb booking. As a result of information exchange, participants may decide to consume goods or services. Another design decision involves ways to ensure that valuable core interactions start to happen, attracting more and more participants to the platform. The core interaction involves three components: the participants, the value unit, and filters (Parker et al. 2016b). Participants are those that create value and those that consume it. The value unit is offered to different users based on filters. For example, Netflix or YouTube define a customized offer for each user. A filter is an algorithm with the aim of increasing successful interactions. A good filter assures that users receive relevant value units. A poor design may make users receive irrelevant value units and leave the platform. Then, there is

1980

D. A. Rossit et al.

the question on how to monetize network effects and decisions regarding manager/sponsor participation, developer participation, and user participation. When platform participants exchange goods or services, a payment occurs. This may take the form of money transfer using a credit card or a payment button. There are other forms of payment such as likes in a post or a good recommendation. Another issue regarding monetization is deciding which side of the market pays. Finally, governance plays a role in engaging platform participants to choose good behaviors and discourage bad interactions. Cusumano et al. (2019) describe how to build a platform in terms of four strategic steps: (a) identifying the various market sides for the platform and how to create value through them; (b) determining how to attract increasing number of users or complementors in order to generate strong network effects; (c) designing how to monetize those network effects; and (d) establishing and enforcing ecosystem rules. These steps resemble the same design decisions as introduced by Parker et al. (2016b). In addition, Cusumano et al. (2019) discuss these four steps with regard to two basic types of platforms depending on their primary function. The “innovation platforms” usually consist of common technological building blocks that the owner and ecosystem partners can share in order to create new complementary products and services. For example, Microsoft Windows and Google Android are commonly used operating systems that serve as innovation platforms for computer and smartphones ecosystems (Cusumano et al. 2019). The “transaction platforms” are intermediaries or online marketplaces that make it possible for people and organizations to share information or to buy, sell, or access a variety of goods and services. eBay, Amazon Marketplace, Uber, Instagram, or Twitter are transaction platforms. The authors use the term “hybrid” to refer to the combination of innovation and transaction platforms within the same company or within the same platform infrastructure. Amazon illustrates this type of platforms since Amazon Marketplace and online store are transaction platforms and Amazon Web Services serve as innovation platform providing cloud computing services. The design issues discussed so far provide a guide to create platform. However, the successful creation of a platform business is more difficult for traditional firms than for digital giants or entrepreneurs. Incumbents are based on value-chain models in which inputs at one end of the chain undergo a series of steps that transform them into an output that has gained added value. The challenge to the incumbents is particularly relevant in the manufacturing sector where there exist many firms established a long time ago and were conceived under a value-chain model. Van Alstyne et al. (2016) notice that the move from value chain to platform involves three key shifts. The first amounts to understand that the chief assets are no longer tangible or intangible ones like mines or intellectual property. The main asset is the user base of the platform (producers and consumers). This leads to the second shift that is from the internal optimization of the value chain to the management of interactions among producers and consumers. The final shift amounts to the transition from the focus on customer value to a focus on the value of the ecosystem in creating an attractive platform for producers, consumers, and complementors.

77 Business Ecosystem Approach to Industry 4.0

1981

Background Concepts from the Manufacturing Literature Industry 4.0 involves the creation of “smart factories” able to adapt their production processes assigning more efficiently their resources. The key technologies are IoT, cyber-physical systems (CPS), cloud computing, and big data.

Cyber-Physical Systems CPS are physical resources with incorporated computational capacities (Lee 2008), integrating physical aspects of production processes with their associated data processing aspects. In particular, they include computers and integrated networks that monitor and control physical processes using computations and communication loops to improve the quality of the production activities (Ghaffarinasab 2020). These systems can be applied to a wide range of areas, from pacemakers to national energy grids (Wang et al. 2015), but their largest impact is on industrial activities (Monostori 2014; Lee et al. 2015). CPSs obtain real-time information of the physical processes of production and submit them to data processing facilities involved in decision-making (Rossit and Tohmé 2018). This, in turn, yields the integration of the different control levels systems in a single system, providing reliable decisions on the fly (Rossit et al. 2018). The production system can thus be flexibly adapted and reconfigured at the different scenarios that a firm can face. It is even possible to create CPS able to “self-configure” (Wang et al. 2015; Rossit et al. 2019b) yielding higher levels of productivity by adapting to changes. The research and innovation funding program for 2007–2013 (FP7) of the EU established that a future networked society had to be grounded on four ft.: Internet by and for the People (IoP) (Lyons 2017), Internet of Contents and Knowledge (IoCK), Internet of Things (IoT), and Internet of Services (IoS) (Yao et al. 2017). IoT links cyber and physical systems making fabrication processes intelligent. IoP connects all the participants, eliminating the barriers between producers and consumers, creating online communities for the design, creation, and sale of products. IoS uses the Internet as a medium for the exchange of services by applying technologies like service-oriented architectures (SOA) or cloud computing. Finally, IoCK transforms data into information/knowledge that can be used in manufacturing systems (Yao and Lin 2016). Since all these applications of Internet are already being used, they will become even faster and cheaper, facilitating closer interactions between customers and production units, connected in platforms.

Cloud Computing The NIST defines cloud computing as “a model for enabling ubiquitous, ondemand access to a shared pool of configurable computing resources (e.g. computer networks, servers, storage, applications and services), which can be rapidly

1982

D. A. Rossit et al.

provisioned and released with minimal management effort or service provider interaction” (Mell and Grance 2011). The NIST also specifies five features that characterize cloud computing systems: (i) being demand-driven (customers place their requests that are automatically satisfied); (ii) wide network access (through mobile phones, laptops, etc.); (iii) the computational resources can simultaneously provide to different users; (iv) high-demand elasticity (increasing requests are immediately satisfied); and (v) services are measurable (servers can be monitored and controlled) (Mell and Grance 2011). For companies this means that lower costs and larger yields can be achieved reducing risks and increasing the accessibility to consumers and providers (Wang and Wang 2018). When cloud computing did not exist, resources and applications were provided in the form of products that were sold or licensed, and then used locally in the on-premises computing infrastructure. With the cloud computing model, you pay for what is used: for the storage required in a temporary window, or for software only when you use it. Cloud computing enables the linking of manufacturing resources and capabilities of companies and thus optimizes internal and external logistics. There are three levels of services, namely: (a) Software as a Service (SaaS) in which software applications are offered and used on the Internet, being an alternative to the acquisition and execution of packages for local self-use. (b) Platform as a Service (PaaS) provides application development facilities, including the design, implementation, testing, operation, and support of web applications and Internet services purchased or created by the user, using vendorprovided tools remotely. (c) Infrastructure as a Service (IaaS) in which processing or storage resources are offered as a service to users. From an enterprise point of view, cloud computing is an alternative with significant benefits, including the reduction of costs in storage infrastructure, the availability of robust services, securing data management, and the ability to select the most appropriate software licensing model. It is important to highlight the potential of scaling that it yields to an organization. If its operations require a larger storage capacity, it can immediately procure what it needs, without having to invest on extra servers.

Big Data The enormous generation of new data has given rise to what has been called big data. The ensuing large databases are characterized by five V’s (Chen et al. 2014): volume (large amounts of data), variety (data comes in different formats and is generated by different sources), velocity (data is generated and renewed by fast processes), veracity (data is used to reduce error levels, inconsistencies, incompleteness, ambiguities, noise, and other kinds of inaccuracies), and value (the

77 Business Ecosystem Approach to Industry 4.0

1983

marginal worth conveyed by data). The volume of data is not defined by a specific size but by the size for which it is not possible to run analyses using traditional tools. In addition, the speed of data creation leads authors like Tomas Davenport to refer to “data streams” (Davenport et al. 2012). Consider, for example, the generation of data in social networks, or by smart and connected products. Big data tools provide means to assess the performance of systems and detect patterns in consumption. In manufacturing, these tools allow analyzing large volumes of heterogeneous and multisource data generated along the life cycle of industrial production (Li et al. 2015). Industry 4.0 environments are ready for the implementation of big data analytics (Rossit et al. 2019c), thanks to the sensors that provide information on events and states. In turn, enterprise systems provide market data (Babiceanu and Seker 2016). The systematic computational analysis of data outputs from these environments help to make “informed” decisions, improving the quality of smart manufacturing processes. For these reasons, data-driven systems are necessary requirements for the implementation of Industry 4.0 environments (Tao et al. 2018).

Digital Twins The concept of a digital twin was first introduced to refer to a rich digital representation of actual devices, being widely used in the aerospace field (Grieves 2014). In 2015, the scope of the original definition was expanded, opening the possibility of using it in other fields (Rios et al. 2015). Virtual factories are characterized as the digitalization of plants integrated with the real systems assisting the production process along the lifecycle of each asset (Sacco et al. 2010). This development led to two different concepts of what a digital twin is: Some researchers believe that it can be identified with its virtual component (i.e., the simulation process) while others emphasize on the connections between the virtual and the physical aspects (Tao et al. 2019). Now consider a more formal definition in terms of a specific architecture. All the aforementioned technologies (CSP, cloud computing, and big data) allow the efficient integration of the different functionalities of a production system. To illustrate the integration of these functionalities, the ISA 95 architecture is considered as a reference architecture, which is officially (ISA-95 2019) defined as: “ISA-95 is the international standard for the integration of enterprise and control system. ISA-95 consists of models and terminology that can be used to determine which information has to be exchanged between systems for sales, finance and logistics and systems for production, maintenance and quality” (Scholten 2007). In terms of a control specification ISA-95 of 5 levels, it can be seen that all the processes can be digitalized, other except for those that require human decisions (as the definition of the goals of the firm). These processes can be incorporated into Industry 4.0 systems (Monostori 2014; Rossit and Tohmé 2018). This integration can be both at horizontal and vertical levels. The vertical integration ranges from the level 0 of ISA-95 to level 3 (manufacturing execution systems) (Rossit et al.

1984

D. A. Rossit et al.

Fig. 2 ISA-95 levels integration

2018; Jalil et al. 2019), as shown in Fig. 2. These levels, handled by CPS, translate the physical events into data, creating a digital twin of the production system. A smart factory integrates the physical factory in the real world and a virtualization of events and physical components in virtual or digital components in the cyber space. That is, an event Xi in the physical space is traduced into an event Xi  in the cyber space. Both components, physical and digital, constitute the Industry 4.0 factory. Digital twins provide useful information about the real workload of the production system, which can be used in planning and business strategizing (Rossit et al. 2019a; Saurikhia et al. 2018). Managers can simulate the inner workings of the plant under different scenarios, providing further information that contributes to better decisions on the entire production process.

Cloud Manufacturing The incorporation of new technologies led to novel production architectures like cloud manufacturing. Cloud manufacturing is a service-oriented architecture that uses cloud computing to relate design and innovation activities with production ones (Xu 2012; Wu et al. 2015). More specifically, starting up from NIST’s definition of cloud computing, Xu (2012) redefined cloud manufacturing as “a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable manufacturing resources (e.g. manufacturing software tools, manufacturing equipment, and manufacturing capabilities) that can be rapidly provisioned and released with minimal management effort or service provider interaction.” Cloud manufacturing, thus, promises elasticity and adaptability through the on-demand provision of resources and services, allocating them through a payas-you-go system. In this way, cloud manufacturing can adequately address the challenges that SME companies face nowadays, as for instance the lack of basic technologies, the restrictions to the access to external resources and capacities,

77 Business Ecosystem Approach to Industry 4.0

1985

the lack of skills to manage complex IT systems, as well as the inability to share efficiently resources and capabilities (Wang and Wang 2018). In the architecture of cloud manufacturing systems (Adamson et al. 2017), it can be distinguished three main agents: the provider, the cloud manufacturing operator and finally the costumer. The provider is the agent that runs the physical production resources, managing the resource layer in the architecture. The cloud manufacturing operator bridges the gap between the ends of the production system. This operator has to handle the main layers of the architecture, starting with the perception layer, managing the data obtained from CPS-like systems, translating them into a format that is friendly toward the rest of the layers. The virtualization layer is in charge of virtualizing the resources and fabrication capacities (which become digital twins), encapsulating them into cloud manufacturing systems. These resources and capacities are easily accessed by other components of the system. The cloud service layer manages systems, services, resources, and tasks being compatible with different activities and service applications as, for instance, those involved in the description, registration, publication, composition, and monitoring of systems. The last layer in the cloud manufacturing operator is the application layer, in which the provider’s services are delivered, allowing customers the possibility of selecting the different properties of pieces, under the constraints of size, material, and tolerances defined by customers. Finally, the customer layers are called interface layer, linking them with the cloud manufacturing operator, facilitating the submission of requirements and the exploration of systems already available. Other designs of cloud manufacturing architectures, developed for instance by Xu (2012), include less layers and subsume the virtualization layer in the domain of the provider. Adamson et al. (2017) review Xu’s (2012) as well as other architectures presented in the literature. In general, these designs have in common a three-agent scheme with finer details varying according to the approach and implementation in each case. Cloud manufacturing intends to solve fabrication problems. It starts with customers requiring fabrication services in order to execute a self-contained task, contacting providers of those services through the architecture of the system. In order to provide manufacturing services in the cloud, their specifications must be clearly stated (Wang and Wang 2018; Halty et al. 2020). To illustrate the relation between services, capabilities, and resources the reader is referred to Fig. 3. In Fig. 3 it is shown that cloud manufacturing services are nested inside manufacturing capability, which in turn reside inside manufacturing resources. The characterization of each of these components is as follows: – Manufacturing resources: provide manufactured and nonmanufactured materials, including equipment, machinery, devices, and smart processes. – Manufacturing capability: has the capacity of transforming things in the domain of fabrication, using tools drawn from manufacturing resources. – Cloud manufacturing service: includes the packages of autonomous manufacturing services, rapidly configurable to satisfy the demands of customers. Their variability is large; CMS services can be randomly activated as well as been active in the long or short term and even strategically enabled.

1986

D. A. Rossit et al.

Fig. 3 Structure of a cloud manufacturing service

Table 1 Manufacturing capability and manufacturing resources Manufacturing capability Design capability Production capability Experimentation capability Management capability Communication capability

Manufacturing resource Hard resource Manufacturing equipment Monitoring device Material Computational resource Storage Transportation

Soft resource Software Knowledge Skill Personnel Experience Business network

A cloud manufacturing system implements manufacturing capabilities in the cloud encapsulating them as manufacturing services packages. Manufacturing capability includes the capacity of designing, producing, experimenting, managing, and communicating. In turn, each manufacturing capability is supported by manufacturing resources, be they hard or soft. In Table 1 each of these concepts are detailed. The language of a manufacturing capability is designed to facilitate its acknowledgment by a cloud manufacturing service providing a formatted way of identifying the supporting resources and the resource/capability relations. Wang and Wang (2018) present an instance of a 5-tuple, relating each capability to the associated resources, in such a way that the cloud manufacturing service can identify those resources and their ability to provide assistance.

Design Requirements of Cloud Manufacturing Architectures Cloud manufacturing provides for digital fabrication and design innovation, as shown by Wu et al. (2015). In an idealized scenario of smart drone deliveries, these authors postulate a cloud manufacturing architecture incorporating design processes in the cloud, as well as integral management of manufacturing services

77 Business Ecosystem Approach to Industry 4.0

1987

and supply chains. They postulate eight requirements that any cloud manufacturing design should satisfy: 1. The system has to provide connections between customers and providers, supporting communications and information and knowledge flows. 2. It has to allow elastic cloud storing of 3D design files, which can be shared and synchronized among the users. 3. It has to provide the capacity to process large amounts of data in a parallel and distributed way, using open code programming languages. 4. It must lend SaaS (Software-as-a-Service) to customers in a multitenancy structure in which a single instance serves several users accessing it through web browsers. 5. It has to assign and control efficiently and effectively the fabrication resources (like production cells and assembly lines). Flows of material and information on the availability and capacity of fabrication resources must be ensured, using IoT tools. 6. The architecture has to provide users with X-as-a-Service applications (e.g., IaaS, PaaS, HaaS, and SaaS). 7. It must have a smart search engine allowing users to find fabrication resources in the cloud. 8. It must provide online price quoting tools to rapidly budget commercial proposals that may arise inside the system. Cloud technologies provide an opportunity to reframe manufacturing businesses, in particular for SME. Combined with SOA, it yields a framework for one of a kind production. Cloud manufacturing shows to be more appropriate for catering specialized and personalized demands, thanks to its flexible and rapid reaction capacities (Wang and Wang 2018). It becomes thus imperative to postulate new models and business strategies able to generate value up from these potentialities.

Industrial Practice An increasing number of companies are developing and implementing cloud manufacturing platforms. General electric is an example of an incumbent that transformed from a product-based to a platform business. In particular, GE Predix (GE 2020) is an industrial platform for the implementation of intelligent systems to monitor and control physical devices or systems through the industrial Internet (Chen 2017). The platform integrates machine sensors to generate a continuous data flow that is allocated in the cloud. Developers can add applications on the platform. The value of Predix is that it generates knowledge in real time. For example, based on data analysis it optimizes the rotation of wind turbines to increase electricity production. The GE strategy consisted in enabling data analysis, remote monitoring of equipment, and facilitating massive machine-to-machine communication. Microsoft Azure IoT Hub (Microsoft 2020) is a fully managed

1988

D. A. Rossit et al.

service integrated into Microsoft Azure’s cloud offering that enables reliable and secure bidirectional communication between millions of IoT devices and an IoT application. Intel IoT platform provides an end-to-end platform for connecting devices, which works with third-party solutions to provide a foundation for seamless and secure connection of devices (Intel 2020).

Factory-as-a-Platform Business Model Proposal A platform-based business model expresses the business logic of cloud manufacturing. Then in order to define a cloud manufacturing business the principles and guidelines already proposed by the platform-based business models’ literature can be followed. As already mentioned in Sect. “Background Concepts from Management Literature” of this chapter, for incumbents (traditional manufacturing plants adopting cloud manufacturing) this means that they have to transition from a value-chain model to platform-based business. The smart factory operates as an orchestrator of the interactions among individual parties and a factory-wide platform facilitates those interactions. The core asset of the platform is the producers and users base and all design decisions have impact on the size of the platform. In what follows, these decisions for the case of a cloud manufacturing platform are discussed. In this section a proposal for a hypothetic platform is presented. The platform business model is described in terms of the design decisions exposed by Parker et al. (2016b), Cusumano et al. (2019), and Fehrer et al. (2018) in their studies of the key issues relevant to create and sustain a platform business. The smart factory is the owner and manager of the platform and controls who and how can participate. Other actors are designers, retailers, and investors. The platform broadcasts a general goal as well as some basic requirements. Designers publish their proposals and retailers vote on them. The digital twin simulates the most voted designs in order to generate data on their costs and manufacturing requirements. One of them is selected and given the information provided by the digital twin a layout for its production is submitted to the factory. Investors provide the funding that allows starting the production of the new good. In the following, some issues that should be defined in order to build a platform are explored. – Participants: This includes the smart factory (owner and platform manager), the designers (producers), retailers, and investors (consumers). – Exchange of information: The platform provides details of design projects enabling users (designers) to know the goal of the project and its basic requirements. Designers upload designs and prospective ones are published. Retailers vote from them. Information on the selected design is published and funding is requested. The platform should facilitate the exchange of information because as a result of these exchanges, designers create proposals and retailers can vote, having the information needed to make a decision. – Exchange of services: The platform provides the mechanisms facilitating the upload of designs and for retailers to vote for their preferred design. A

77 Business Ecosystem Approach to Industry 4.0

1989

crowdsourcing mechanism allows reaching out to investors for funding. The proliferation of prospective designs attracts more investors, and as more investors participate in the platforms, both designers and retailers are attracted to be part of projects. – Exchange of value: Designers enhance their reputation with the votes received by their designs – even if they do not go into production. Reputation plays the role of currency in the platform. The platform also gives designers access to specialized software tools to develop their projects. Retailers have the opportunity to vote for projects that are more promising according to their market and business knowledge. The platform may also support retailers with market data, financial calculators, and other tools helping in the assessment of projects. Investors provide funding for designs that go into production and reduce investment risk, thanks to the assurance given by the votes of retailers who in turn reflect market insights. As it can be noticed, reputation mechanisms, access to relevant tools, and access to knowledge or data should be used to attract and maintain platform participants. – Monetization: Higher value creation by designers on the platform attracts more retailers and investors, who, in turn, attract more designers and further value creation. This powerful positive growth dynamics challenges monetization decisions. Platform business rarely charges all their users since this may discourage participation. A pricing choice is to charge members of one category of users while allowing members of another category of users to participate for free (Parker et al. 2016b). The smart factory platform may offer designers free participation as a way to attract innovative design projects. The platform may charge retailers for access. In this case, retailers looking for promising products to commercialize may get access to design proposals and have the right to vote. This form of monetization benefits both parties: Designers are motivated to publish their best work on the platform, while retailers get access to new proposals and vote on them. When a design is selected for production, investors pay the designer for the rights to produce and commercialize the design. The smart factory generates its profits from all the projects that go into production. – Openness: A platform is open to the extent that no restrictions are placed on participation in its development, commercialization, or use. No restrictions seem reasonable and nondiscriminatory, and thus, if they exist, are applied uniformly to all potential platform participants (Parker et al. 2016b). The decision about the level of openness affects usage, developer participation, monetization, and regulation (Fehrer et al. 2018). The smart factory firm manages and sponsors the platform. The platform manager organizes and controls interactions, and retains legal control over the technology (such as the software code that controls its operation). In order to facilitate extensions of the platform functionality, the firm may open its business to participation to extension developers. However, the platform is not open as to allow all developers to protect the quality of the design proposals and to retain control over the revenue streams of the platform. Extension developers add features and value to the platform and normally are not employed by the platform management firm (Parker et al. 2016b). The smart factory platform would benefit from tools that simplify

1990

D. A. Rossit et al.

transportation arrangements for retailers. In order to facilitate this extension of its platform functionality, the smart factory might open its business to the participation of extension developers. To prevent poor-quality service providers joining the platform, an approval process is required. Another type of developer is the generator of data aggregators that enhance the matching function of the platform by adding data from platform users and the interactions they engage in. Considering the confidential nature of the smart factory platform, it is a wise decision not to open it up to this kind of developers. Unauthorized disclosure of information about design proposals, costs, and manufacturing requirements should be penalized. Another concern is related to an unrestricted participation of designers. Although the smart factory would like to facilitate participation, high-quality content is a major concern, otherwise retailers and investors may move away from the platform. Designers should upload their projects using a platform tool to assure that the required data is provided (facilitating the screening of projects). Curation may also take the form of voting and feedback. A designer’s reputation is based on past projects ratings. At the same time, platform managers need to continuously monitor designers’ participation and suggestions to prevent designers to get discouraged of participating.

Manufacturing Architecture The core of the factory-as-a-platform model is a system based on Industry 4.0 technology that connects all the stakeholders in a dynamical integrated structure, a platform. The factory, which already did the transition to Industry 4.0, provides the tools, in particular the digital twin, to carry out all the actions needed to reach the production phase. For instance, it can evaluate the workload of the factory and coordinate delivery dates, already in the evaluation of designs. This factory can be embedded in the CBDM architecture since it can easily satisfy the requirements of the latter. For instance, the digital twin satisfies requirements 4 and 5 of CBDM, which prescribe the provision of software and applications needed for multitenancy participation. Requirement 8, concerning the ability to quote the resources demanded by the production process, can also be fulfilled by Industry 4.0 factories. Requirement 1, which involves the connection with stakeholders, requirement 2 on the management of information flows, and requirement 3 on the assessment of the data generated by the system are all satisfied thanks to the technologies already incorporated into the factory. Finally, requirements 6 and 7 are easily satisfied through the interactions in the platform, allowing the direct the connection among the interested parties. Therefore, the factory-as-a-platform business model can be implemented in the CBDM architecture, yielding value to all the participants in the ecosystem.

77 Business Ecosystem Approach to Industry 4.0

1991

Limitations of Platform Interactions in Smart Factories A limitation of smart factories is that there still exist tasks that require a substantial human participation. There are, in particular, three instances in which this is relevant for our proposal: – When the solution depends on the participation of large numbers of human users, like in the cases of reCAPTCHA (Law and Von Ahn 2011) or Wikipedia (Kaplan and Haenlein 2014). In the case of production plans, the need of supervised training of expert components through deep learning or similar technologies can be highlighted. – When humans are just better in carrying out tasks. McAfee and Brynjolfsson (2017) analyze which abilities will remain essentially human in the future. Creative tasks are, until now, better handled by human beings but there are some (e.g., the generation of hypotheses (Spangler 2014)) that are starting their automatization. A limit may be found in tasks that require social skills like empathy, ability to work in teams, leadership, etc. (McAfee and Brynjolfsson 2017). In a factory this may be the case when parts of a design have to be modified on the fly in response to unforeseen collateral effects. The providers of inputs can be affected by this and thus may participate in the search of a solution. This is similar to addressing challenges by crowdsourcing the solution process (for instance, the Kaggle Data Challenge (Garcia Martinez 2017)). – When multiple alternatives are conceived and one must be chosen (as for instance when the best transport and distribution option is sought) and the different parties face disparate costs and benefits that have to be pondered and negotiated.

Illustrative Example In this section, it is introduced an idealized manufacturing scenario based on currently existing technologies and the factory-as-a-platform proposal. In this scenario, an established cosmetics manufacturing firm owns and manages a platform that provides solutions responding to other manufacturers of cosmetics and toiletries. The factory also owns a range of filling equipment (bottle, jar, tube, stand-up pouch, and Doypack filling), labeling and wrapping machines (linear and rotary bottle or carton labeling, horizontal and vertical cartoning, and tray packing). By transforming its business model from a traditional cosmetics manufacturer to a platform-based business model, the company shifts from making money based on units produced to monetizing network effects. The platform facilitates other manufacturers to access and use assets to encourage innovation by third parties. The benefit of cosmetics manufactures that participate in the platform by outsourcing their projects is the opportunity to innovate and to scale their production without investing in equipment. Cosmetics manufacturers access the smart factory user base (designers and investors) that is a key asset of the platform. On the other

1992

D. A. Rossit et al.

hand, it may raise a risk for cosmetics manufacturers because of the platform’s ability to appropriate their complementary innovations. Nambisan et al. (2018) discuss regulations and policy-related factors with implications of platforms on entrepreneurship (for example, antitrust policies, or national or regional regulations to promote technology-based development). Based on its own or a customer’s requirements (another cosmetics manufacturer), the platform crowdsources the design process. Online competitors create a new packaging or improve an already existing item. Anyone may join the design community and membership in the community grants access to submit a design. To submit a design, community members follow the requirements for the type of product (e.g., balms, creams, and lotions), package (e.g., bottle, carton, display boxes, and tube), and concept of the product defined by the cosmetics manufacturer. From a conceptual design perspective, crowdsourcing platforms allow the design team to solicit design ideas from more sources, thereby enhancing innovation. Designs remain available for the assessment of cosmetics manufacturers and the highest scoring designs are selected and a digital twin simulates them in order to generate production data. The platform may also enable retailers to vote for designs. The benefit of this option is that the votes of retailers reveal an implicit knowledge about market demands. However, this decision depends on the customer preference about having full control on the design selection or not. After the design phase is finished, the design team builds prototypes of selected proposals and makes a decision about the design that will go into production. The platform may also support crowdfunding and investors may pay the designer for the rights to produce and commercialize the design. Cloud-based manufacturing allows for rapid manufacturing capacity scalability by outsourcing manufacturing tasks to qualified global suppliers. Also, if the design requires a processing that cannot be satisfied by the available equipment in the plant factory, cloud-based manufacturing allows retrieving a list of machines that are capable of producing the design.

Final Remarks The factory-as-a-platform model integrates all the participants in the ecosystem. Designers provide the innovations, which the digital twin allows to evaluate; the retailers, who know the preferences of the market (because they control distribution), evaluate the commercial viability of projects and investors provide funding. The scale of production becomes measured not only in terms of the sheer volume of production but depends also on its variety. The crowdsourcing solution facilitates identifying promising innovations, supporting it from outside the firm. Figure 4 summarizes all these interactions, showing all the relevant players in the ecosystem. The proposal differs substantially from a traditional manufacturing business model. It requires a strategic transition to foster ecosystem value around an incumbent factory that is challenged by the transformation. The CBDM architecture

77 Business Ecosystem Approach to Industry 4.0

1993

Fig. 4 The factory-as-a-platform business model where the smart factory owns a platform that enables interactions between ecosystem actors

underlying the business model supports information flows among ecosystem actors, providing also the means to process the ensuing large volume of data. The different interventions within the platform can be shared synchronically by all of them and the virtualization of the resources and assets allows the simultaneous assessment of plans and the negotiation of delivery conditions of the physical output of the factory. This architecture provides means to solve problems that Industry 4.0 cannot address by itself, like the generation of innovations. The platform enables interactions among all stakeholders, for example, allowing retailers to rate proposed designs. Crowdsourcing facilitates finding funding to carry out production. The entire chain, from the original design to the finished good, is coordinated and implemented on the platform, accelerating the life cycle of product generation. Along the way, producers and complementors participate suggesting alternatives (cost reduction activities or the use of new kinds of inputs, for instance), fostering value for all the participants in the ecosystem. Industry 4.0 becomes the production engine for the entire ecosystem, relating the virtual realm with the physical one. Digital blueprints are transformed into actual products, in such way that the demand is assuredly satisfied. The use of Internet technologies leads to planning the completion of work orders in order to deliver the goods to the customers at the requested dates.

1994

D. A. Rossit et al.

All these features enhance diversity through complementary products and services, incorporating idiosyncratic tastes and demands into the design of products in an efficient way. Customers, in turn, get information about details of the production process of their requested products, making the whole process more transparent for both sides of the market and ensuring that its focus is on the satisfaction of demands.

Conclusions The question of how Industry 4.0 can benefit from technological changes and how smart factories may transform their business models is addressed in this chapter. The proposed factory-as-a-platform model provides the blueprint for the way in which the smart factory can become a business ecosystem, defining its main players, the interactions between producers and consumers as well as how to maximize value for the whole ecosystem, making monetization and openness decisions. The cloud-based design and manufacturing architecture implements it, covering from the design of prototypes to the production of final goods. In this way, Industry 4.0 is enhanced, fulfilling further its promise of creating larger value for the stakeholders of the firms using it.

Important Websites • What is a platform?: https://www.applicoinc.com/blog/what-is-a-platformbusiness-model/ • Why business models fail? Pipes versus Platforms https://www.wired.com/ insights/2013/10/why-business-models-fail-pipes-vs-platforms/ • What is cloud manufacturing? https://www.erpsoftwareblog.com/cloud/2016/06/ what-is-cloud-manufacturing/#: :text=Cloud%20manufacturing%20(CMfg)%20 is%20the,at%20any%20time%20or%20place

References Adamson G, Wang L, Holm M, Moore P (2017) Cloud manufacturing–a critical review of recent development and future trends. Int J Comput Integr Manuf 30(4–5):347–380 Adner R (2006) Match your innovation strategy to your innovation ecosystem. Harv Bus Rev 84(4):98–107 Babiceanu RF, Seker R (2016) Big data and virtualization for manufacturing cyber-physical systems: a survey of the current status and future outlook. Comput Ind 81:128–137 Bengtsson M, Raza-Ullah T (2016) A systematic review of research on coopetition: toward a multilevel understanding. Ind Mark Manag 57:23–39 Ceccagnoli M, Forman C, Huang P, Wu D (2012) Co-creation of value on a platform ecosystem: the case of enterprise software. MIS Q 36(1):263–290 Chen Y (2017) Integrated and intelligent manufacturing: perspectives and enablers. Engineering 3:588–595

77 Business Ecosystem Approach to Industry 4.0

1995

Chen M, Mao S, Liu Y (2014) Big data: a survey. Mobile Netw Appl 19(2):171–209 Cusumano M, Gawer A, Yoffie D (2019) The business of platforms. Strategy in the age of digital competition, innovation, and power. HarperCollins publishers, New York Davenport T, Barth P, Bean R (2012) How “big data” is different. MIT Sloan Manag Rev:22–24 Fehrer J, Woratscheck H, Brodie R (2018) A systemic logic for platform business models. J Serv Manag 29(4):546–568 Garcia Martinez M (2017) Inspiring crowdsourcing communities to create novel solutions: competition design and the mediating role of trust. Technol Forecast Soc Chang 117:296–304 Gawer A, Cusumano M (2008) How companies become platform leaders. MIT Sloan Manag Rev 49(2):28–35 GE (2020) GE digital. [Online]. Available: https://www.ge.com/digital/iiot-platform. Accessed 7 Oct 2020 Ghaffarinasab N (2020) A tabu search heuristic for the bi-objective star hub location problem. Int J Manag Sci Eng Manag:1–13 Grieves M (2014) Digital twin: manufacturing excellence through virtual factory replication, White paper, pp 1–7 Halty A, Sánchez R, Vázquez V, Viana V, Piñeyro P, Rossit DA (2020) Scheduling in cloud manufacturing systems: recent systematic literature review. Math Biosci Eng 17(6):7378–7397 Howe J (2006) The rise of crowdsourcing. Wired Mag 14(6):1–4 Intel (2020) What is the Intel® IoT platform. [Online]. Available: https://www.intel.la/content/ www/xl/es/internet-of-things/iot-platform.html. Accessed 7 Oct 2020 ISA-95 (2019) ISA-95.com. [Online]. Available: http://www.isa-95.com. Accessed 2 Oct 2020 Jacobides M, Cennamo C, Gawer A (2018) Towards a theory of ecosystems. Strateg Manag J 39:2255–2276 Jalil SA, Hashmi N, Asim Z, Javaid S (2019) A de-centralized bi-level multi-objective model for integrated production and transportation problems in closed-loop supply chain networks. Int J Manag Sci Eng Manag 14(3):206–221 Kaplan A, Haenlein M (2014) Collaborative projects (social media application): about Wikipedia, the free encyclopedia. Bus Horiz 57:617–626 Karhu K, Botero A, Vihavainen S, Tang T, Hämäläinen M (2011) A digital ecosystem for cocreating business with people. J Emerg Technol Web Intell 3(3):197–205 Kern R (2014) Dynamic quality management for cloud services, vol 192. Springer International Publishing, Cham Law E, Von Ahn L (2011) Human computation. Syn Lect Artificial Intell Mach Learn 5(3):1–121 Lee EA (2008) Cyber physical systems: design challenges. In Object oriented real-time distributed computing (ISORC), 2008 11th IEEE international symposium, Florida Lee J, Bagheri B, Kao H (2015) A cyber-physical systems architecture for industry 4.0-based manufacturing systems. Manuf Lett 3:18–23 Li J, Tao F, Cheng Y, Zhao L (2015) Big data in product lifecycle management. Int J Adv Manuf Technol 81(4):667–684 Lyons T (2017) The book of Fermat. Fermat Project, New York McAfee A, Brynjolfsson E (2017) Machine, platform, crowd. Harnessing our digital future. W. W. Norton & Company Ltd., New York Mell P, Grance T (2011) The NIST definition of cloud computing. NIST, Washington, DC Microsoft (2020) Azure IoT Hub. [Online]. Available: https://azure.microsoft.com/en-us/services/ iot-hub/. Accessed 7 Oct 2020 Monostori L (2014) Cyber-physical production systems: roots, expectations and R&D challenges. Procedia Cirp 17:9–13 Nambisan S, Siegel D, Kenney M (2018) On open innovation, platforms, and entrepreneurship. Strateg Entrep J 12:354–368 Osterwalder A, Pigneur Y, Tucci C (2005) Clarifying business models: origins, present, and future of the concept. Commun Assoc Inf Syst 15:1–25 Parker G, Van Alstyne M, Xiaoyue J (2016a) Platform ecosystems: how developers invert the firm. Boston University Questrom School of Business, Boston

1996

D. A. Rossit et al.

Parker G, Van Alstyne M, Choudary S (2016b) Platform revolution. How networked markets are transforming the economy and how to make them work for you. W. W. Norton & Company, New York Rios J, Oliva M, Mas F (2015) Product avatar as digital counterpart of a physical individual product: literature review and implications in an aircraft. In 22nd ISPE Inc. International conference on concurrent engineering, Delft Rong K, Hou J, Shi Y, Lu Q (2010) From value chain, supply network, towards business ecosystem (BE): evaluating the BE concept’s implications to emerging industrial demand. In IEEE international conference on industrial engineering and engineering management, Macao Rong K, Lin Y, Li B, Burström T, Butel L, Yu J (2018) Business ecosystem research agenda: more dynamic, more embedded, and more internationalized. Asian Bus Manag 17(3):167–182 Rossit D, Tohmé F (2018) Scheduling research contributions to smart manufacturing. Manuf Lett 15:111–114 Rossit D, Tohmé F, Frutos M (2018) Industry 4.0: smart scheduling. Int J Prod Res 57(12):3802– 3813 Rossit D, Tohmé F, Frutos M (2019a) An industry 4.0 approach to assembly line resequencing. Int J Adv Manuf Technol 105(9):3619–3630 Rossit D, Tohmé F, Frutos M (2019b) Production planning and scheduling in cyber-physical production systems: a review. Int J Comput Integr Manuf 32(4-5):385–395 Rossit D, Tohmé F, Frutos M (2019c) A data-driven scheduling approach to smart manufacturing. J Ind Inf Integr:1–11 Sacco M, Pedrazzoli P, Terkaj W (2010) VFF: virtual factory framework. In 2010 IEEE international technology management conference, Lugano Saurikhia A, Ahmed S, Haleem A, Gangopadhyay S, Khan MI (2018) Evaluating technology management factors for fly-ash utilization in the road sector using an ISM approach. Int J Manag Sci Eng Manag 13(2):108–117 Scholten B (2007) The road to integration: a guide to applying the ISA-95 standard in manufacturing. ISA, Research Triangle Park Spangler S (2014) Automated hypothesis generation based on mining scientific literature. In Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining, New York Tao F, Qi Q, Liu A, Kusiak A (2018) Data-driven smart manufacturing. J Manuf Syst 48:157–169 Tao F, Zhang H, Liu A, Nee A (2019) Digital twin in industry: state-of-the-art. IEEE Trans Indust Inform 15(4):2405–2415 Van Alstyne M, Parker G, Choudary P (2016) Pipelines, platforms, and the new rules of strategy. Harv Bus Rev 94:54–62 Wang L, Wang X (2018) Cloud-based cyber-physical Systems in Manufacturing. Springer, London Wang L, Törngren M, Onori M (2015) Current status and advancement of cyber-physical systems in manufacturing. J Manuf Syst 37:517–527 Wang Y, Ma H-S, Wang K-S, Yang J-H (2017) Industry 4.0: a way from mass customization to mass personalization production. Adv Manuf 5(4):311–320 Wu D, Rossen D, Wang L, Schaefer D (2015) Cloud-based design and manufacturing: a new paradigm in digital manufacturing and design innovation. Comput Aided Des 59:1–14 Xu X (2012) From cloud computing to cloud manufacturing. Robot Comput Integr Manuf 28(1):75–86 Yao X, Lin Y (2016) Emerging manufacturing paradigm shifts for the incoming industrial revolution. Int J Adv Manuf Technol 85(5):1665–1676 Yao X, Zhou J, Lin Y, Yu H, Liu Y (2017) Smart manufacturing based on cyber-physical systems and beyond. J Intell Manuf 30(8):2805–2817

78

Modeling the Dynamics of a Smart Factory Marisa Analía Sánchez, Daniel Alejandro Rossit, and Fernando Tohmé

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Production Technologies: A Chronology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industry 4.0 Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cyber-Physical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Internet of Things . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Digital Twins and Smart Factories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A System Dynamics Representation of Digital Twin CPS . . . . . . . . . . . . . . . . . . . . . . . . . . . Causal Loop Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stocks and Flows Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Model Validation and Result Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scenario Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1998 2000 2002 2002 2003 2004 2006 2007 2008 2013 2014 2016 2017 2017

M. A. Sánchez Departamento de Ciencias de la Administración, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina e-mail: [email protected] D. A. Rossit () Department of Engineering, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina e-mail: [email protected] F. Tohmé Departamento de Economía, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_66

1997

1998

M. A. Sánchez et al.

Abstract

“Digital twining” is one the main ways of establishing data channels in cyberphysical systems combining the outputs of a virtual model with real time data collected by sensors. The purpose to this chapter is to outline the digital twin of a cyber-physical production system. The System Dynamics paradigm to the case of a shop-floor factory devoted to cloud manufacturing is applied. The digital twin uses data from the real production line, providing assistance to maintenance procedures triggered by inconsistencies between the real and the virtual processes. Keywords

Digital Twin · Cyber-physical systems · System dynamics · Simulation

Introduction One of the main components of the Industry 4.0 infrastructure are “smart factories,” able to adapt their production processes to demand, assigning their resources more efficiently than traditional factories. The key technologies enabling smart factories are the Internet of Things (Yao et al. 2019), Cyber-physical Systems (CPS) (Lee 2008; Wang and Wang 2018), Cloud Computing (Wang and Wang 2018), and Big Data (Li et al. 2015; Tao et al. 2018). These technologies allow the efficient integration of the different functionalities of the control specification ISA-95 into a single production system. In terms of its five levels, all the processes that can be digitalized can be thus incorporated into Industry 4.0 systems (Monostori 2014; Rossit and Tohmé 2018). This integration can be both at horizontal and vertical levels. The vertical integration extends from the physical process level (level 0 of ISA-95) to Manufacturing Execution Systems (level 3 of ISA-95) (Rossit et al. 2018). These levels, handled by CPS, translate the physical events into digital data, creating a digital twin (DT) of the production system. The digital twin provides useful information about the real workload of the production system, which can be used in planning and business management processes (Parsanejad and Matsukawa 2016; Rossit et al. 2019a). Managers can simulate the inner workings of the plant under different scenarios, gaining further information, contributing to make better decisions. The ultimate objective of digital twins is to improve the operation and efficiency of manufacturing assets, reducing costs by forecasting future states and supporting advanced decisionmaking throughout the entire manufacturing lifecycle (Damjanovic-Behrendt and Behrendt 2019). A DT of a production system is a type of simulation that allows the real-time control of production. Ding et al. (2019) use the term “digital twining” to refer to the process of building a digital twin in the cyber world of physical objects and systems, establishing data channels for cyber-physical connection and synchronization. The

78 Modeling the Dynamics of a Smart Factory

1999

Fig. 1 Evolution of research on digital twin, IIoT and simulation since 2015. (Source: Scopus database)

implication of this definition is that a DT uses both simulated values and real time data collected by sensors. The simulation has the ability to compress time, running ahead the real world process, by setting the simulation clock at one relevant period ahead of real time. If at time tr a sensor reports a machine failure that was not generated in the simulation when its internal clock was at tr , the latter should be reset at tr updating its state with this new information. This feature enables a DT to keep the simulation running at a par with the real production line (Fig. 1). The aim of this work is to describe a simulation model for a digital twin cyberphysical system. In order to define the guidelines for a smart plant environment, the modelling methodology of System Dynamics can be used to handle the uncertainties and nonlinear relationships among interacting system components (Barlas 1996; Morecroft 2007). The complexity in the definition of the model may ensue from its structural or dynamic aspects. Structural complexity refers to the number of components in a system, or the number of combinations involved in making a decision. Dynamic complexities result from the nonlinear and history-dependent nature of self-organizing and adaptive systems (Sterman 2000). System Dynamics addresses these two kinds of complexity by postulating that the behavior of complex systems results from an underlying structure of flows, delays, and feedback loops (Forrester, Industrial Dynamics 1961). The emphasis in System Dynamics is not on forecasting the future, but on learning how the actions in a period of time can trigger reactions in the future (Senge 1990; Sánchez 2013). This chapter is organized as follows. Sect. “Production Technologies: A Chronology” reviews the history of production technologies up to the inception of Industry

2000

M. A. Sánchez et al.

4.0. Sect. “Industry 4.0 Technologies” discusses the characteristics of Industry 4.0 systems. Sect. “Digital Twins and Smart Factories” introduces the concept of Digital Twin and how it contributes to the design of smart factories. Sect. “A System Dynamics Representation of Digital Twin CPS” presents a System Dynamics representation of a Digital Twin Cyber-Physical System, and Sect. “Causal Loop Diagrams” uses this model to simulate different scenarios that may arise in actual production processes. Finally, Sect. “Stocks and Flows Diagrams” concludes.

Production Technologies: A Chronology The first great process of economic, social, and technological transformation recorded in modern times is the so-called First Industrial Revolution (sometimes also simply called the Industrial Revolution). This process took place in the second half of the eighteenth century in Great Britain, where the largest set of transformations in the history of mankind took place since the Neolithic. Before the Industrial Revolution, the economy was basically rural, based on agriculture and commerce. This process of change lasted approximately from 1760 to 1840, introducing the steam engine as a great technological improvement, making possible the transformation of the thermal energy contained in water steam into mechanical energy. This invention made it possible to drive mechanisms and devices with more energy than that provided by living beings. This, in turn, made it possible to work with materials that either had been very difficult to handle or that had allowed only exploitation at a much smaller scale. That is why metallic components began to be more and more common in machinery and mechanisms. Metals became processed at a much larger scale, creating the conditions of possibility for technologies that used these large pieces of metal (allowing, for instance, the emergence and proliferation of railways). On the other hand, this revolution had a direct impact on society, by modifying and creating new social classes. The economy, then, became urban, industrialized and mechanized. Drastic changes also ensued in the production process, transitioning from artisan workshops to factory systems where machinery powered by steam engines produced large quantities of goods satisfying a growing demand for new products. In the nineteenth century, driven by the increasing availability of the first industrialized products, a large number of scientific and technical developments followed one another, deepening the dominance of the industrial paradigm. Many of these developments amounted to the development of new chemical processes like the Bessemer procedure for the production of steel at large scale. As steel is a major constituent of the vast majority of production machines and tools, the ensuing reduction in costs and the remarkable improvement in quality, allowed the development of new machinery and the creation of an infrastructure that facilitated the transition from steam/mechanical energy to the use of electrical power. The application of electrical mechanisms allowed the adoption of machinery

78 Modeling the Dynamics of a Smart Factory

2001

for industrial operations, making them more flexible. These developments led to what became a milestone in this industrialization process: the adoption of serial production. Design improvements became not only a matter of products but also of production processes. This led to a new paradigm in production, in which the product and its production process are conceived and developed together. In this way, large mass productions arose, benefitting from economies of scale and satisfying the needs of customers in a scale unseen until then. The iconic case of this is the automobile industry of the early twentieth century. In the twentieth century, scientific and technological advances were the great drivers of industrial innovation. The knowledge gained in different branches of science and engineering led to a great number of industrial developments and improvements, ranging from the creation of new materials, such as those with a polymeric structure, to the generation of tools on which small operations could be carried out automatically, based on the first PLC (Programmable Logic Controller) mechanisms. This last class of devices gave rise to what was called the third industrial revolution. Although the development of the first PLCs can be traced back to the 1950s, it can be argued that by the 1970s their industrial-scale implementation was widespread enough to have a global impact. They allowed the great leap from a production controlled purely and exclusively by human operators, to self-controlled systems. This control capacity, incipient in the beginning, jointly with the increasing transition from analog to digital devices, laid the foundations for the design of robotized systems. In terms of productivity, it was possible in many areas to start producing 24 h a day, 7 days a week. Since the beginning of the twenty-first century, a new industrial revolution, the fourth one (Industry 4.0) has been brewing. Its greatest difference with previous production paradigms is centered on the growing autonomy of the production process. On top of the advances in automation and robotics that constituted the third industrial revolution, this new phase incorporates the possibility of interconnecting its components through the Internet of Things (IoT). By connecting the different physical production assets together, it became possible to achieve global control and monitoring of production processes, not only at the local and individual level allowed by traditional control systems, It is now possible to analyze and predict future states of the production system more precisely, since it has access in real time to the information generated by each device and sensor in the factory. In turn, by being able of transmitting information through the Internet, it is no longer necessary to limit the processing and analysis activities to those that can be carried out by the embedded systems in the production machinery. It can now also make use of servers and specialized computer centers accessed through the Internet. All of this results in the possibility of generating a complete digital version of the factory, in which different Artificial Intelligence and Big Data tools can be implemented to improve the decision-making process associated to production (Hermann et al. 2016; Rossit et al. 2019a, 2020; Kumar et al. 2020). Figure 2 shows the main characteristics that stand out and define the four industrial revolutions.

2002

M. A. Sánchez et al.

Fig. 2 The four Industrial Revolutions. (Source: Christoph Roser at AllAboutLean.com)

Industry 4.0 Technologies As discussed in the last section, industrial revolutions are identified with the creation and implementation of new technologies for industrial production activities. Industry 4.0 is not the exception, and in this section, it is made a brief presentation of the technologies that allow us to consider Industry 4.0 as a genuine industrial revolution.

Cyber-Physical Systems Cyber-physical systems are perhaps the starting point of all (or the vast majority) of the advantages of Industry 4.0. They represent the new production assets used in the production processes of Industry 4.0. These CPS make it possible to closely and deeply link information technologies and physical processes, allowing in turn a new approach to physical production processes based on a deep digitization of everything that happens at the physical level. In other words, the CPS consists of a set of sensors linked to the shop floor and the machine tool analyzed. This allows obtaining data in real time about the production process, considerably improving the ability to monitor and control the production process since this data can be sent to a Decision Support System (DSS) or to another linked CPS (Wang et al. 2015). Therefore, CPS allow the passage of information from the physical world to the digital or cybernetic world. Among the features of CPS, their information processing capacity stands out (Babiceanu 2016; Yao and Lin 2016). Traditionally, all the information collected in production plants was sent and analyzed globally by planners and supervisors, allowing them to define an action plan. However, given that CPS have a higher processing capacity than traditional production systems, it is possible to incorporate in them the control capacities

78 Modeling the Dynamics of a Smart Factory

2003

associated with SCADA systems (Supervisory Control And Data Acquisition) as well as some functions of the higher levels of the control structure (for example, Manufacturing Execution Sys-items, or Quality analysis) (Rossit and Tohmé 2020). In this way, what was previously centralized in those higher levels can now be decentralized and controlled autonomously by the CPS (Monostori 2014; Lee et al. 2015), leading to an effective autonomy of the CPS-based production system (Monostori 2016).

Internet of Things All the characteristics that make Industry 4.0 a new production paradigm are based on the ability to communicate and transfer information, basically unavailable in old production systems. This is now possible using the technologies of the Internet of Things (IoT). IoT allows the sensors distributed in the machines and on the shop floor to transmit the data collected during the execution of the production process, as well as to communicate the different systems and production machines with each other. Therefore, the information that in traditional production systems used to be confined to the device that collected it can now be accessed by other systems. This accessibility to information allows the integration of the functionalities of the different systems involved in the production process. By integrating and unifying the information handled by the different systems and decision-making processes of the factory, the ability to autonomously plan operations and activities becomes greatly enhanced (Rossit, Tohmé, & Frutos, 2019). The different systems analyze the same scenario described in the same terms, that is, the same information becomes transparent to all systems. Then, the diagnostics of operations, maintenance, quality, service level, etc., will be consistent with each other. Since these systems can communicate with each other, any conflict arising between them can be resolved directly by the affected parties. On the other hand, IoT opens the possibility of creating such massive flow of information that its analysis becomes really unmanageable (Sanchez 2018). This possibility makes necessary the use of efficient tools for processing this large amount of data (Tao et al. 2018). Therefore, the implementation of IoT in Industry 4.0 systems is coupled with the need to use Big Data tools to exploit this technology to its full potential. The data processing tools for Big Data make it possible to predict future scenarios and develop action plans to anticipate possible unfavorable situations (Rossit and Tohmé, Scheduling research contributions to Smart manufacturing 2018; Rossit et al., Industry 4.0: Smart Scheduling 2018). At the same time, the ability to record this large amount of data facilitates the generation of knowledge out of it, using it to improve future developments and do long-term planning. This knowledge, generated by data on its own production process, refers thus to the concrete production process and its natural evolution. This eliminates barriers to the applicability of developments designed for the specific environment of the factory (Li et al. 2015; Wang and Wang 2018).

2004

M. A. Sánchez et al.

However, all these benefits derived from the use of Big Data tools for the analysis of the production process imply the requirement of commensurate computing capacities. Faced with this need, IoT once again provides the appropriate support, allowing access to Cloud Computing services. Cloud Computing represents those services that provide access to the computing resources (both for storage and processing) without the need to spend on installing an entire support physical system (servers and data centers) (Mell and Grance 2011). These services allow users to take advantage of the availability of huge datasets, as in the case of Industry 4.0 technologies. In turn, Cloud Computing services provide a very flexible scalability, lending resources and services on demand. Therefore, the amount of investment and associated risk becomes much more affordable and manageable for industries. The business model of Cloud Computing services providers, based on the payas-you-go logic, makes it a technology easily adaptable by even relatively small firms (Adamson et al. 2017). On the other hand, since the data and services are available in the cloud, Cloud Computing makes possible for any company with plants in different geographical locations to share and access the same information (Wu et al. 2015). This increases the transparency of the information within the group of industrial plants, and facilitates the synchronization and coordination of different activities. There are different services that Cloud Computing offers: Software as a Service (SaaS), Platform as a Service (PaaS), and Hardware as a Service (HaaS) (Wang and Wang 2018). SaaS offers the possibility of running applications or software from the cloud, without the need of installing software on users’ computers, saving the consumption of hardware resources, as well as costs associated with software maintenance. On the other hand, the execution from the cloud prevents users from owning the software and disallowing the free modification of software, although many SaaS services offer the possibility of making adjustments and configurations according to the user’s preference (Tao and Zhang 2017). In contrast, when Cloud Computing services are of the PaaS type, users have greater access and freedom to develop and install their own software. Therefore, a company could install a set of APIs of its own development, but which, through the cloud platform, can be accessed remotely and individually by different users of the same industry. At the same time, it would allow joint development between users from different geographical locations. And the last type of service, HaaS, is the one that gives the users the greatest freedom, since the service is contracted for the hardware itself. In this way, HaaS users can develop, deploy, and run as many software and applications as they want. HaaS service contracts enable standardized storage and processing capacity to be delivered to the user over the network (Wang et al. 2014; Yao et al. 2019).

Digital Twins and Smart Factories The intensive digitalization process of production systems induced by Industry 4.0 has led naturally to the idea of creating digital replicas of those systems, known as Digital Twins. They create virtual but accurate models of the physical aspects

78 Modeling the Dynamics of a Smart Factory

2005

of production. This allows evaluating and probing questions on a 1:1 map of the production process with the great advantages lend by computing simulations. Digital Twins represent faithfully the data and parameters of the real-world machinery, diagnosing from afar the evolution of the production process. The data gathered in the plant, transmitted through IoT reach the Digital Twin that may be generated, for instance, by a PaaS Cloud Computing service. An expert may, in turn, access the Digital Twin from far away to assess the current state of the machinery. This reduces the costs in time and money of physical inspections. Real time data facilitate emitting alarms from afar to the operators in the plant, preventing potential damages. Another type of alarm can be emitted when the equipment is used inefficiently or violating its warrants. Furthermore, Digital Twins allow programming more precise maintenance tasks, avoiding excessive plant shutdowns. The concept of digital twin as a rich digital representation of actual devices was initially adopted by the aerospace industry (Grieves 2014). In 2015, the scope of the original definition was expanded, opening the possibility of using it in other fields (Rios et al. 2015). Virtual factories were characterized as the digitalization of plants integrated with the real systems assisting the production process along the lifecycle of each asset (Sacco et al. 2010). This development led to two different concepts of what a digital twin is: some researchers believe that it can be identified with its virtual component (i.e., the simulation process) while others emphasize on the connections between the virtual and the physical aspects (Tao et al. 2019). Tao and Zhang (2017) present a design for a smart factory including four main components: a physical shop-floor (PS), a virtual shop-floor (VS), the data generated by the digital twin (D), and a shop-floor service system (SSS). The physical part includes entities such as human, machines, and materials. The virtual part refers to the digital twin of the physical shop-floor which supports the decision-making and control of the physical part (Tao and Zhang 2017). It provides control orders for the PS and optimization strategies for the SSS. The data component integrates physical, virtual, and service data with the aim of providing consistent information. The service component embodies services like the definition of the production and resource allocation plans as well as guidelines for data fusion. These services, in turn, provide information for both the physical and the virtual shop-floor. The authors propose a three stage operation of the digital twin: • Before production, the SSS defines a production plan based on data collected from different sources (customer orders, real-time data read by sensors, outputs from the simulation, as well as business data) and transmits it to the VS for verification. • The VS simulates the production plan and if it works correctly, it sends control orders to the PS to start production. The real-time data generated by the PS is then recorded by the VS. • Once the production process finishes, its recorded history is analyzed to extract knowledge that can be used by the VS to simulate future production plans.

2006

M. A. Sánchez et al.

Fig. 3 Simulation model of the digital twin cyber-physical system. Dashed lines: data flow, solid lines: control flow

In the digital twin shop-floor framework, physical and simulated data are synchronized. The operational implication is that the simulation is fed with realtime sensor data, sending back controls to the PS in a continuous interaction between the physical and virtual components. The framework of Tao and Zhang (2017) is the basis for implementing the simulation model of a digital twin cyber-physical system. We, in turn, add on top of that the process of updating in real-time the production plan in the context of a cloud manufacturing environment. The ensuing digital twin embodies many specialized digital twins: parts DT, machines DT, and a shop-floor DT. The model assumes the use of real-time production data (shop-floor and cloudshop floor) to support the factory operations. While simulation has been traditionally used to identify bottlenecks in production plan, it can now be incorporated into smart machines or tools equipped with devices able to capture, process and transmit data. A simulation fed with such information becomes able to check whether the behavior of the physical factory is consistent with the “simulated” behavior. In particular, when (real) data indicates an unforeseen failure, the simulation synchronizes accordingly (data fusion), updating its behavior, and transmitting new control orders to smart devices. Hence, such simulation of the digital twin cyber-physical system helps to handle failures. Figure 3 depicts our simulation model of a digital twin cyber-physical system.

A System Dynamics Representation of Digital Twin CPS Let us consider a smart factory consisting of two production lines A and B. Orders are scheduled on line A. If A is operating above its capacity or it is being repaired, the orders are sent to line B. If line B is operating above its capacity, then the remaining orders are assigned to cloud manufacturing suppliers. For simplicity, it

78 Modeling the Dynamics of a Smart Factory

2007

Table 1 Probability distributions describing cycle time of production facilities Production facilities A B Cloud manufacturing suppliers

Cycle time (in days) Normal distribution (mean = 2, st. deviation = 2%). Normal distribution (mean = 4, st. deviation = 2%) Normal distribution (mean = 2, st. deviation = 2%)

is possible to assume that only line A suffers failures and the time to repair is of 3 time units. The cycle times of production lines A, B, as well as that of the cloud suppliers are represented by the probability distributions detailed in Table 1. These normal distributions are assumed to have been fitted with historical data of demand.

Causal Loop Diagrams The first step towards creating a dynamic model of the digital twin CPS is to draw its Causal Loop diagram (Sánchez 2013). In a Causal Loop diagram variables are connected by arrows denoting the causal influences among them: x → y means that the input variable x has some causal influence on the output variable y. A positive influence, denoted by +, means “a change in x, being the rest of variables unchanged, causes y to change in the same direction.” In turn, a negative influence means “a change in x, being the rest of variables unchanged, causes y to change in the opposite direction.” An increase in a cause variable does not necessarily mean the effect will actually increase (Sterman 2000). Since a variable may have more than one input, a change depends on the combined effect of all its input variables. The process wherein one component x initiates changes in other components, and those modifications lead to further changes in x itself is said a feedback loop (McGarvey and Hannon 2004). According to the system of interest, the diagram may be decomposed in several sub-diagrams. For the model of the digital twin cyber-physical system, five diagrams are developed. To save space only a partial view of the full causal loop diagram is included. Figure 4 depicts the interaction between demand and the production plan. The diagram shows the causal links among the state of the backlog and the accepted orders rate, the dispatch rate of production lines A, B and cloud manufacturing, and the sales volume. The “New orders” constitutes the central input to the production system. As orders increase, the backlog grows, increasing the forecasted number of orders to produce. An increase in the capacity of the production line A adds up the number of units it produces but decreases the shop floor production of line B as well as the production in cloud manufacturing. In turn, increases in failures of production line A decreases its running time and thus its available capacity. As shop-floor production increases, the work in progress and the required dispatch rate

2008

M. A. Sánchez et al.

Fig. 4 Causal loop diagram of demand and the production plan. This represents a partial view of the causal loop diagram of the complete system. (Source: own elaboration)

grow in time. Finally, increases in the dispatch rate impacts positively on the rate of fulfilled orders and on sales. The accumulation of resources in a system (stocks), their rates of change (flows), and the feedback loops are the main components of the syntax of dynamic systems models. Stocks are represented by rectangles, inflows and outflows represented by pipes pointing into or out of stocks; valves that control the flows indicate the rates of change; and clouds represent the sources and sinks for the flows originating outside the boundary of the model. Converters representing processes that convert inputs into outputs are depicted as circles.

Stocks and Flows Diagrams For structuring the smart shop-floor simulation model, five interacting sectors are introduced. Cloud manufacturing requests are incorporated in addition to the physical shop-floor production in such a way that production is not limited by plant capacity. The overall structure of interactions between customer demand and flows of orders and products is based on the supply chain models in Sterman (2000).

78 Modeling the Dynamics of a Smart Factory

2009

Fig. 5 Stock and flow diagram of Demand. (Source: own elaboration)

Demand sector. The stock Backlog is the accumulation of accepted orders less the satisfied orders. It is assumed that orders cannot be changed or cancelled. There is a limit on the amount of production based on the capacity of the plant and the number of manufacturing orders that can be requested to other manufacturers in the cloud. This limit sets the theoretical maximum production rate that may depend on constraints such as the availability of materials, labor, equipment in the plant, and equipment of partners in the cloud. Hence, the rate of fulfilled orders depends on the dispatch rate of production lines A and B and on partners. Orders that cannot be met due to limited capacity are recorded and costs of lost sales or penalties may be computed. Figure 5 depicts the Stock and flow diagram for the Demand sector. There are converters and flows defined in such a way as to capture the logic of the synchronization when the behavior of the simulation is not consistent with sensor data. The Real time gap is set to 10 meaning that the simulation clock is 10 time units ahead of real time (this value is arbitrarily defined only for illustration). Let the simulation clock be ts . Sync backlog is a variable recording the value of Backlog at period ts − Real time gap − 1. The initial value of converter Reboot simulation is 0 and is set to 1 when simulation synchronization is required. Then, if Reboot simulation is 1 at simulation time ts , then the Backlog stock is emptied (the flow Empty Bck is habilitated) and set to the historical value at ts − Real time gap − 1. Virtual shop-floor sector. Production is captured with a chain of two stocks: Work in progress and Finished products (Fig. 6). For the purpose of this model, all

2010

M. A. Sánchez et al.

Fig. 6 Stock and flow diagram of the Virtual Shop-floor. (Source: own elaboration)

stages of the production process are aggregated together into the Work in progress stock as prescribed in Sterman (2000). There are two manufacturing facilities in the plant (A and B) and if necessary, requests are made to cloud manufacturing partners. As mentioned in the description of the Demand sector, there are stocks and flows defined to support synchronization. Orders are scheduled on production line A. If A is operating above its capacity or it is being repaired, the orders are sent to line B. If also B is operating above its capacity, the required orders are placed to cloud manufacturing providers (Figs. 6, 7, 8, and 9). There exist a maximum of orders that can be requested from the cloud depending on previous agreements with partners. The cycle time is dependent on process-technology and product design (converters Mean cycle time A, Mean cycle time B). Also, plant A is subject to failure events and the rate of production is thus dependent on the availability of the plant. The production rate of A is 0, if it is downtime.

78 Modeling the Dynamics of a Smart Factory

2011

Fig. 7 Stock and flow diagram of Cloud manufacturing. (Source: own elaboration)

Fig. 8 Stock and flow diagram of Service Production Plan. (Source: own elaboration)

Most of the time, the inflow Begin production A assumes the value Shop floor production A. However, when simulation synchronization is required at instant ts , the simulation should reflect machine A failure at time ts , and a downtime interval of length ts + Time to repair, resetting stock Work in Progress A with the

2012

M. A. Sánchez et al.

Fig. 9 Stock and flow diagram of Digital twin data. (Source: own elaboration)

value at period ts − Real time gap − Time to repair (value recorded by converter Sync prod A). Assume that products are dispatched taking into account the dispatch delay goal. For example, if the dispatch delay goal is set to be 3 days and current time is ts , then finished products which orders were placed at time ts − 3 become dispatched. The dispatch delay takes into account the time since an order is included in the backlog until it is ready for dispatch. Service production plan sector. This sector is in charge of recommending orders of production to each production line (Fig. 8). The number of orders to produce is dependent on the dispatch delay goal and on the available capacity of production lines A, B, and of the cloud manufacturing providers. Calculations use data provided by (a) simulations (the available capacity of production lines A and B), and (b) production managers (capacity utilization and the dispatch delay goal). Hence, the forecast of orders to produce are defined as the amount in the Backlog divided by dispatch delay goal. The production volume is based on forecasts, assuming that no more than the available capacity is allowed. The shop floor production for A is minimum between its available capacity and the forecast of orders to produce. If production line is no able to produce all orders, B is scheduled to produce them, and if B has not enough capacity, requests are forwarded to the cloud. Digital twin data sector. The digital twin data sector emulates (Fig. 9) the digital twin data component introduced by Tao and Zhang (2017). This component is in charge of calculating measures used by the Service Production Plan and the Virtual Shop-floor. These measures are ratios between the capacity utilization of plants A and B, the number of orders that can be requested to cloud partners, and the

78 Modeling the Dynamics of a Smart Factory

2013

maximum number of orders that can be accepted considering the current level of utilization of plants A and B and cloud services. The dispatch performance rate compares the current dispatch delay and the dispatch goal defined by the firm’s goals. The time to repair machine A is defined as constant, but it can be also represented with a probability distribution function or can be a simulation input entered by the operator. The most relevant part of the digital twin sector is the comparison between data retrieved by sensors and simulated data. For simplicity, only consider a sensor for machine A (converter Sensor A). The sensor detects machine A failures. If this information does not agree with simulated data (converter Failure A) at period tr , then the behavior of the simulation after tr should be updated to take into account this (the converter Reboot simulation is set to 1). In a more realistic setting, the presence of failures should be checked on a time interval (not a point interval). The current available capacity of machine A is restricted to its maximum (physical) capacity. In addition, the downtime in case of failure and the reboot of the simulation are also taken into account.

Model Validation and Result Accuracy Validation is concerned with determining that a model is an accurate representation of the real system. No model is ever totally representative of the system under study, so in practice validation refers to the problem of how one knows whether a model is satisfactory to make decisions about modeled systems (Sánchez 2013). System Dynamics focus on describing how a system operates; hence, the ultimate objective of validation is to establish the validity of the structure of the model (“right output behavior for the right reasons”) (Barlas 1996). Barlas (1996) proposes to first validate the structure of the model and then start testing the behavior accuracy. Structural tests include direct structure test and structure-oriented behavior tests. Direct structure using direct inspection of structures, stocks and flows, and equations are performed. For dimensional consistency that checks the correct use of units of measure (Forrester and Senge 1980) the validity indicators provided by the software tool are used. In addition, many controls on equations checking for zero or negative results are included. Structure-oriented behavior tests assess the validity of the structure applying tests to the simulation output (Sánchez 2013). Tests for model behavior were implemented using more likely condition scenarios, extreme condition scenarios, and behavior sensitivity tests as suggested by Carson and Flood (Carson and Flood 1990). Finally, results accuracy should be considered (Barlas 1996). In this work, an illustrative example is used, and as a consequence, there is no real data to compare the model input-output transformations and the corresponding input-output transformations for the real system.

2014

M. A. Sánchez et al.

Scenario Modelling By using the System Dynamics diagrams as a modeling tool and its derived equations it is possible to compute numerical solutions. Tools such as Stella™ allow automatizing this computation. There are many techniques of numerical integration of differential equations that can be used to solve the resulting system. The most popular are the Runge-Kutta methods, although their use of finite steps and the approximation to average rates over the interval introduce errors, affecting the results (Table 2). Customers’ orders at the start initiate the simulation execution. In order to test different configurations of the production system, the dispatch delay goal, the initial capacity and mean cycle time of production lines, the capacity utilization of shop floor production lines, are defined as input parameters that are updated from a simulation control panel. Running the system it is found that during the first 5 days, the system is unstable and dominated by the details of the initial conditions. So this period is discarded for analytical purposes. There is a simulated failure of production line A at time 10 and the line is down for 3 days. There is an increase in line B and cloud production (Fig. 10). There is a failure of line A (detected by sensor data) at real time 20 (the simulation clock is set 10 units ahead of real time). Since the simulation behavior did not consider a failure at time instant 20, then production lines A, B and cloud are set to 0 and re-initialized with real values recorded at simulation time 20 (Fig. 10). In addition, line A is down for 3 days. Simulation forecasts are “corrected” by incorporating real time data when available. Hence, a synchronization of the physical and virtual factory takes place. The available cloud capacity is not enough (notice that Desired cloud capacity in Fig. 11 increases after the simulated failure at period 10 and after synchronization due to a real failure of A). This suggests that during downtime plant periods (or an abrupt increase in demand), improving the access to more cloud providers would be recommended. The maximum acquired cloud capacity is not enough to satisfy the desired dispatch delay of 3 days (Fig. 12). In (Hekimoglu and Barlas 2010) the authors use sensitivity analysis to evaluate the way that behavior patterns vary with different parameters values. This enables to assess the impact of different levels of cloud manufacturing capacity. An increase in that variable would reduce the number of rejected orders. At the same time, such increment entails higher costs incurred in partner agreements.

Table 2 Parameters for the scenario simulation Parameter description Demand (number of orders) Dispatch delay goal Initial capacity of line A Initial capacity of line B Available cloud capacity

Value Normal distribution (mean = 500, st. deviation = 10%) 3 days 500 units 400 units 1000 units

78 Modeling the Dynamics of a Smart Factory

2015

Fig. 10 Digital twin interface including graphics depicting the state of the smart factory and slider input devices to set values of capacity utilization, dispatch delay goal, initial capacity of production lines A and B, maximum limit of requests to cloud manufacturers, and mean cycle times

Fig. 11 Production of lines A, B, and cloud partners

Using this model, it is possible to define alternative scenarios to assess recommend settings with respect to different variables. For example, minimize unfilled orders, inventory-holding costs, or cloud manufacturing costs. In addition, the resultant response behavior of the system under different customer order patterns can be analyzed. This basic model may be enhanced by including more factors that have impact in the production system. As mentioned in the description of the model, the theoretical

2016

M. A. Sánchez et al.

Fig. 12 Demand represented by the New orders line; cloud manufacturing capacity utilization (between 0 and 1); and desired extra cloud capacity (defined as the difference between forecasted production orders and available shop-floor and cloud capacity)

maximum production rate that may depend on constraints such as the availability of materials, labor, equipment in the plant and equipment of partners in the cloud. A more comprehensive model may disclose among the different factors and analyze the impact of any change of changes in these factors on production performance indicators. Regarding cloud manufacturing, its main benefit is its contribution to the elimination of unfilled orders. However, cloud costs or the time required to take products from one facility to another should be incorporated in the model to have an understanding of the impact of different production plans configurations.

Conclusions The purpose to this chapter was laying down the guidelines for a digital twin of a cyber-physical production system based on a System Dynamics approach. Although it is not intended to cover all systems that may interact with the digital twin, it was possible to describe how to synchronize the simulation when an inconsistency is detected when data comes from different sources (data retrieved from sensors and simulated data). It is possible to extend the model to integrate the digital twin with different databases of the organization. Due to the anticipatory nature of simulation runs, different production plans can be examined to choose the most convenient one according to criteria like optimizing cost, dispatch rate or rejected orders. This information may be useful to support human decision-making or to control a physical machine. This enhances the value of a digital twin (simulation) by supporting real-time production control.

78 Modeling the Dynamics of a Smart Factory

2017

Other simulation paradigms can be used to model and implement a digital twin. For example, discrete event simulation (Banks et al. 2001) is adequate to model the shop-floor production lines. However, a digital twin of a cyber-physical production system is a complex system (many components or subsystems interact and the knowledge of the impact of a change in a variable over another is incomplete). To model problems with such levels of uncertainty an iterative learning process and modeling tools able to deal with incomplete information are required (Sánchez 2013). The model presented in this work serves as a basis to add interactions with other subsystems belonging to different tiers of the supply chain. The system dynamics paradigm is appropriate to model uncertainty and study the evolution of the system over time (for example, the interactions among demand, logistic disruptions, and machine fatigue). System dynamics shows its potential in forecasting or understanding the far future. Hence, a promising approach is to use a hybrid method based on discrete event simulation to handle short-term phenomena (machine failure) and system dynamics comprising all subsystems and assess the far future.

Important Websites CESMII. The Smart Manufacturing Institute: https://www.cesmii.org/ Digital Twin: https://www.linkedin.com/pulse/beginners-guide-digital-twin-illustra tive-approach-sanjoy-paul/ Digital Twin Consortium: https://www.digitaltwinconsortium.org/ Industrial Internet Consortium: https://www.iiconsortium.org/ Industry 4.0 Consortium: https://www.i40c.com/ Internet of Things Consortium: https://iofthings.org/ ISA-95: https://www.isa.org System Dynamics Society: https://www.systemdynamics.org/

References Adamson G, Wang L, Holm M, Moore P (2017) Cloud manufacturing–a critical review of recent development and future trends. Int J Comput Integr Manuf 30(4–5):347–380 Babiceanu RF (2016) Big Datfia and virtualization for manufacturing cyber-physical systems: a survey of the current status and future outlook. Comput Ind 81:128–137 Banks J, Carson J, Nelson B, Nicol D (2001) Discrete-event system simulation. Prentice Hall, Upper Saddle River Barlas Y (1996) Formal aspects of model validity and validation in system dynamics. Syst Dyn Rev 12(3):183–210 Carson ER, Flood RL (1990) Model validation: philosophy, methodology and examples. Trans Inst MC 12(4):178–185 Damjanovic-Behrendt V, Behrendt W (2019) An open source approach to the design and implementation of Digital Twins for Smart Manufacturing. Int J Comp Integrat Manufact 57(20):1–19

2018

M. A. Sánchez et al.

Ding K, Chan F, Zhang X, Zhou G, Zhang F (2019) Defining a Digital Twin-based Cyber-physical Production system for autonomous manufacturing in smart shop floors. Int J Product Res 6315– 6334. https://doi.org/10.1080/00207543.2019.1566661 Forrester J (1961) Industrial dynamics. Pegasus Communications, Cambridge, MA Forrester J, Senge P (1980) Tests for building confidence in system dynamics models. In: Legasto A, Forrester J, Lyneis M (eds) System dynamics. North-Holland, Amsterdam Grieves M (2014) Digital twin: manufacturing excellence through virtual factory replication. White paper, 1–7. Retrieved from Digital Twin: Manufacturing Excellence through Virtual Factory Replication: http://www.apriso.com/library/Whitepaper_Dr_Grieves_ DigitalTwin_ManufacturingExcellence.php Hekimoglu M, Barlas Y (2010) Sensitivity analysis of system dynamics models by behavior pattern measures. Proceedings of the 28th international conference of the system dynamics society, Seul Hermann M, Pentek T, Otto B (2016) Design principles for industrie 4.0 scenarios. 2016 49th Hawaii international conference on system sciences (HICSS). IEEE, Hawaii, pp 3928–3937 Kumar S, Suhaib M, Asjad M (2020) Industry 4.0: complex, disruptive, but inevitable. Manag Product Eng Rev 11(1):43–51 Lee EA (2008) Cyber physical systems: design challenges. In Object oriented real-time distributed computing (isorc), 2008 11th ieee international symposium. IEEE, Florida, pp 363–369 Lee J, Bagheri B, Kao H (2015) A cyber-physical systems architecture for industry 4.0-based manufacturing systems. Manufact Lett 3:18–23 Li J, Tao F, Cheng Y, Zhao L (2015) Big data in product lifecycle management. Int J Adv Manuf Technol 81(4):667–684 McGarvey B, Hannon B (2004) Modeling dynamic systems. Springer, New York Mell P, Grance T (2011) The NIST definition of cloud computing. NIST, Washington, DC Monostori L (2014) Cyber-physical production systems: roots, expectations and R&D challenges. Procedia CIRP 17:9–13 Monostori LK (2016) Cyber-physical systems in manufacturing. CIRP Ann 65(2):621–641 Morecroft J (2007) Strategic modelling and business dynamics. A feedback systems approach. Wiley, Chichester Parsanejad M, Matsukawa H (2016) Work-in-process analysis in a production system using a control engineering approach. J Jpn Indust Manag Assoc 67(2):106–113 Rios J, Oliva M, Mas F (2015) Product avatar as digital counterpart of a physical individual product: literature review and implications in an aircraft. In: Curran R, Wognum N, Borsato M, Stjepandi´c J (eds) 22nd ISPE Inc. International conference on concurrent engineering. Volume 2: transdisciplinary lifecycle analysis of systems. IOS Press, Delft, pp 657–666 Rossit D, Tohmé F (2018) Scheduling research contributions to smart manufacturing. Manuf Lett 15:111–114 Rossit D, Tohmé F (2020) The tolerance scheduling problem for maximum lateness in Industry 4.0 systems. In Advances in mathematics for industry 4.0. Elsevier, London, pp 95–113 Rossit D, Tohmé F, Frutos M (2018) Industry 4.0: smart scheduling. Int J Product Res 1–12. https:/ /doi.org/10.1080/00207543.2018.1504248 Rossit D, Tohmé F, Frutos M (2019a) A data-driven scheduling approach to smart manufacturing. J Ind Inf Integr:1–11. https://doi.org/10.1016/j.jii.2019.04.003 Rossit D, Tohmé F, Frutos M (2019b) An industry 4.0 approach to assembly line resequencing. Int J Adv Manufact Technol 1–12. https://doi.org/10.1007/s00170-019-03804-0 Rossit D, Tohmé F, Frutos M (2019c) Production planning and scheduling in Cyber-Physical Production Systems: a review. Int J Comp Integrat Manufact 1–11. https://doi.org/10.1080/ 0951192X.2019.1605199 Rossit D, Tohmé F, Mejía Delgadillo G (2020) The tolerance scheduling problem in a single machine case. In: Scheduling in industry 4.0 and cloud manufacturing. Springer, Cham, pp 255–273 Sacco M, Pedrazzoli P, Terkaj W (2010) VFF: virtual factory framework. In: 2010 IEEE international technology management conference. IEEE Press, Lugano, pp 1–8

78 Modeling the Dynamics of a Smart Factory

2019

Sánchez M (2013) Modeling for system’s understanding. In: P. Fonseca i Casas (ed) Formal languages for computer simulation: transdisciplinary models and applications. IGI Global, Hershey, pp 38–61 Sanchez MA (2018) How Internet of things is transforming project management. In: Gontar ZH (ed) In smart grid analytics for sustainability and urbanization. IGI Global, Warsaw, pp 73–102 Senge P (1990) The fifth discipline: the art and practice of the learning organization. Doubleday/Curency, New York Sterman J (2000) Business dynamics: systems thinking and modeling for a complex world. McGraw-Hill, New York Tao F, Zhang M (2017) Digital twin shop-floor: a new shop-floor paradigm towards smart manufacturing. IEEE Access 5:20418–20427 Tao F, Qi Q, Liu A, Kusiak A (2018) Data-driven smart manufacturing. J Manuf Syst 48:157–169 Tao F, Zhang H, Liu A, Nee A (2019) Digital twin in industry: state-of-the-art. IEEE Trans Indust Inform 15(4):2405–2415 Wang L, Wang X (2018) Cloud-based cyber-physical systems in manufacturing. Springer, London Wang W, He Z, Huang D, Zhang X (2014) Research on service platform of Internet of things for smart City. In Jiang J, Zhang H (eds) ISPRS technical commission IV symposium. XL-4. Suzhou, Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci, pp 301–303 Wang L, Törngren M, Onori M (2015) Current status and advancement of cyber-physical systems in manufacturing. J Manuf Syst 37:517–527 Wu D, Rossen D, Wang L, Schaefer D (2015) Cloud-based design and manufacturing: a new paradigm in digital manufacturing and design innovation. Comput Aided Des 59:1–14 Yao X, Lin Y (2016) Emerging manufacturing paradigm shifts for the incoming industrial revolution. Int J Adv Manuf Technol 85:1–12 Yao X, Zhou J, Lin Y, Yu H, Liu Y (2019) Smart manufacturing based on cyber-physical systems and beyond. J Intell Manufact 30(8):1–13

Robotics in Industry 4.0

79

Ashwin Misra, Anuj Agrawal, and Vihaan Misra

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Industrial Revolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industry 4.0 Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Role of Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sensors and Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Path Planning in Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Artificial Potential Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Grid-Based Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cell Decomposition (CD) Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rapidly Exploring Random Tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Probabilistic Road Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fuzzy Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ant Colony Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bacterial Foraging Optimization (BFO) Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AI in Robotics and Industry 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analytics Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Platform Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Operations Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2022 2023 2025 2027 2028 2037 2040 2040 2041 2041 2042 2043 2043 2044 2045 2045 2046 2046 2047 2047 2047 2047

A. Misra () Robotics Institute, Carnegie Mellon University, Pittsburgh, USA A. Agrawal Delhi Technological University, New Delhi, India V. Misra Netaji Subhas University of Technology, New Delhi, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_68

2021

2022

A. Misra et al.

Challenges in Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cybersecurity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Variation in Labor Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . New Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Artificial Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ethics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2051 2051 2051 2052 2052 2052 2052 2053

Abstract

The advent of robotics in industrial automation has been an integral part of the fourth industrial revolution. Industry 4.0 has various components including robotics, artificial intelligence, machine learning, and internet of things, among various other technologies. The utilization of robots in industries has been on an exponential rise since they increase productivity, efficiency, accuracy, and human safety by a substantial factor. This chapter provides a detailed overview on the different types of robotic technologies that are currently being used in various industries. It also gives an insight of the various control structures and algorithms implemented in an industrial environment to execute targeted tasks with maximum accuracy. It discusses various state estimation and localization techniques as well as motion planning algorithms currently used. Alternatively, this has opened up various opportunities for research, which in turn contributes to industrial automation and optimizes current systems continuously. Applications in various industries and the current challenges faced are also discussed to provide a holistic overview to the role of robotics. Keywords

Industry 4.0 · Robotics · Artificial intelligence · Machine learning · Sensors · Path planning · Robotic manipulators · Drones

Introduction The dynamic global markets are the driving forces of technological innovations. Lately, robotics and artificial intelligence have been at the forefront of research and development with groundbreaking contributions from academia and the industry. Automation has led to vast developments, ranging from the first mass-produced product, the T-Ford, to many products ranging from mobile phones to automobiles. This rapid progression has been possible due to the rapid pace of industrial innovation and the adaptive process that it follows. The onset of robotics in the industrial domain has opened a new door to unlimited possibilities. The robots’ high repeatability and accuracy rates allowed factories to mass-produce components

79 Robotics in Industry 4.0

2023

that were never mass-produced before due to specific operational difficulties. The efficiency and productivity of the plants reached unprecedented levels. Industrial robots’ modularity also makes them useful for various computer visions, sensors, and artificial intelligence predictive models. Robotics is a reasonably old industrial automation concept, beginning from the first generation of robots in the 1960s to current intelligent industrial robots. The third industrial revolution propelled significant groundbreaking research in robotics; Joseph Engelberger and George Devol developed the first industrial robot in 1959, which used hydraulic actuators and a magnetic drum to die-cast hot metals. This drum stored sequential commands in the joint coordinate’s axis, which were executed to perform a specific industrial task. This robot was used to make automotive interiors in a production plant in Trenton, New Jersey, USA. However, this was not the first cylindrical robot. The first cylindrical robot was introduced in 1962 and was installed in the Ford Factory, Canton, USA, by the American Machine and Foundry (AMF) (Grau et al. 2017).

The Industrial Revolution Industry 4.0, or the fourth industrial revolution, was coined by the German government in 2017. It refers to developing a new initiative in industrial technology and data processing to complement the industrial automation procedure. With the rapid pace at which industrial innovation and production are progressing, the changing global markets require quick and adaptive strategies to keep up with the global demand. A brief review of the industrial revolution would provide a summary of the developments and the relevant timeline (Fig. 1). The 1800s marked the first industrial revolution; it was possible due to mechanical power generation and mechanization of various processes. Mainly in the textile industry, the first manufacturing processes were executed as opposed to manual work. A significant product of this revolution was steam engines, which ultimately became the backbone of the nineteenth century’s transportation. This step marked the subsequent process of improving the quality of the lives of the general public. The second industrial revolution started with mass production and the introduction of electricity in industrialization. The electrification of processes improved the speed and reduced the load on industrial workers. Henry Ford’s contribution to this revolution was immense by launching the first mass-produced car – Ford T Model. This revolution marked the production of goods in large quantities through a conveyor-belt mechanism, which boosted the plant’s production capacity. Automation, microchips, and digitalization brought the third industrial revolution. This revolution phase brought flexibility in manufacturing, with programmable machines, setups, and adaptable production lines. This phase allowed for various products to be manufactured; however, the quantity of production could still not be controlled.

2024

A. Misra et al.

Fig. 1 The industrial revolution timeline (Rojko 2017)

The research on the internet of things, cyber-physical systems, and smart automation paved the way for the latest and current fourth industrial revolution. It introduces the decentralization of various decision-making processes in a manufacturing plant to enhance decision-making. It allows dynamic production systems that are characterized by flexibility in custom mass production and production quantity. It solves all the prior shortcomings but poses new ones in its own domain. Industry 4.0 aims to encapsulate various cutting-edge technologies such as artificial intelligence, robotics, Industrial IoT, and smart factories into a package to improve industrial automation and explore new possibilities to increase productivity and efficiency. It has profound advantages: decreasing production, quality control, and logistics cost by a significant percentage. All the current practices to improve productivity and drive the resource cost down were exhausted; hence, Industry 4.0 is the result of new technologies and technological advancement and brought various new methods such as just-in-time manufacturing. Various added operational benefits include process transparency, reduced time-to-market, data analytics, custom mass production, increased efficiency, increased productivity, and easy troubleshooting (Fig. 2).

79 Robotics in Industry 4.0

2025

Fig. 2 Industry 4.0 ecosystem

Industry 4.0 Structure The core feature of an Industry 4.0 plant is the reconfigurable manufacturing systems. Through a digital approach, the systems can be modified to execute various specific tasks. The particular component which makes it possible is cyber-physical systems, which combine traditional physical processes and computational power. Recent developments have led to the availability and affordability of data modules, transducers, and networks to inch closer to a highly technical environment. Broadly, a CPS has two main functional components: (1) smart data acquisition through high-speed connectivity and real-time information relay and (2) computational data analytics, warehousing, management, and data-driven decision-making. However, various theories suggest the architecture of a CPS in the light of Industry 4.0. Lee et al. (2015) proposes a 5C architecture as a systematic guideline to the CPS Architecture. The 5Cs include the connection, conversion, cyber, cognition, and configuration.

2026

A. Misra et al.

The connection level denotes the Smart Connection through which multimodal and reliable data is acquired from machines, controllers, and sensors. This is the first stage in the development of a CPS structure. There are various types of data to be acquired, numerical, categorical, image, etc., and transferring this data to the central processing unit through efficient tracks with minimal data loss and noise. The selection of accurate and proper transducers is one of the most important considerations of this level. Conversion is the stage in which the acquired data has to be converted to meaningful information. Several tools and methodologies are developed, according to the facility, to achieve such meaningful conversion. Recently, the medical and health industries have focused on specific algorithms in the conversion domain to achieve specific health data. This level imparts a certain level of self-awareness to the CPS. The cyber level is at the center, the central processing unit of the whole CPS structure. The various information collected from multiple sources in the network must be analyzed to better understand multiple processes and gain important insights. The huge pool of data is analyzed carefully to keep a check on various machines throughout the operation. Multiple machines are benchmarked and are compared to the different machines to improve individual performance. Predictive analytics are also employed to predict future process parameters such as scheduled delivery time, quantity, etc. The various acquired knowledge must be clearly presented to let the subject experts make essential decisions about the facility’s processing. The cognition level refers to maintaining proper infographics, analytics, and comparative study to optimize the decision-making process (Shariatmadar et al. 2019). The configuration level refers to the corrective loop between the cyberspace and physical space and helps the machines adapt and configure according to the decisions taken. This stage acts as a means to apply preventive and corrective decisions made at the other levels (Fig. 3). How an Industry 4.0 environment can use a CPS System is explained (Jazdi 2014). Various components are added to introduce adaptability to the environment. The features include smart networking, mobility, flexibility, and integration of the customer and new business models. These features introduce the CPS as a dynamic structure which could communicate easily with the industrial ecosystem, flexible in various data types, and also handles the customer needs from end to end. It also highlights the fact that the management team has to overhaul their business models to suit an Industry 4.0 facility (Fig. 4). This figure shows the gap in recent manufacturing systems to the level of Industry 4.0. Various verticals among the industrial domain have seen developments over the last 5 years, with various research articles studying the gaps in these domains (Fig. 5). This figure highlights the domains in a contemporary manufacturing plant that have been developed and studied. All these domains are constantly evolving with continuous research especially in the robotics and planning components. The combination of all these components leads to the package we know as Industry 4.0.

79 Robotics in Industry 4.0

2027

Lower Upper

MANUFACTURING CONCEPTS

INDUSTRY 4.0

Fig. 3 The 5C architecture (Lee et al. 2015) Self-Configuration Self-Optimization Early-Aware Decision Making Predictive Maintenance Real-time Response Customization Flexibility Standardization Communication Digitalization

The Gap

Single-Station Automated Cell

Automated Assembly System

Flexible Manufacturing System

ComputerIntegrated Manufacturing System

Reconfigurable Manufacturing System

The State

CURRENT MANUFACTURING SYSTEMS

Fig. 4 Identified gaps from research to application in Industry 4.0 (Qin et al. 2016)

Role of Robotics The advent of robotics in industrial manufacturing was due to their ability to execute accurate and faster production processes, operate in dangerous environments, and reduce operational costs (Ustundag and Cevikcan 2017). Artificial intelligence and

2028

A. Misra et al.

Fig. 5 Recent research topics pertinent to Industry 4.0(Mourtzis et al. 2014)

big data in light of Industry 4.0 have contributed to the significant role of robotics in Industry 4.0. This role is ever-changing as it keeps up with the present research and market demands. Traditional robots used up till now were placed in specified spaces and are designed to repeatedly execute a particular sequence of actions. Hence, they are developed to make reconfiguration difficult for a different or new production line. A very few engineers are adept with programming a robot, reconfigure, and use it to its maximum advantage. Many such problems can be solved by the use of data science, as proposed in Industry 4.0. The collaboration between humans and robots is an essential part of a holistic approach to production; in such aspects, the robot is programmed to ensure the safety of the human worker (Colgate et al. 2008). Thus, various recent technologies lay the basis for the industry’s future: Cloud Technology (Xu 2012), IoT for remote control and monitoring, machine learning, and Big Data Analytics. A framework combining all such technologies is continuously being integrated in the Industry 4.0 architecture to improve data analytics, transparency, and quality. The robotic technology highly depends on various factors such as the cost, technological advancements, quality, and processors.

Types of Robots Increasingly, robots are replacing humans in repetitive tasks in the industrial and manufacturing sectors. With the advent of Industry 4.0, these robots are now being equipped with intelligent decision-making capabilities, reducing humans’ need in the loop and increasing efficiency. Unlike traditional robots that carry out just one predefined task, these modern age robots can sense the environment and robot’s internal state, take decisions, and act upon it. Various robotics platforms are available to carry out industry-specific tasks:

79 Robotics in Industry 4.0

2029

Aerial Robots: More commonly known as drones or unmanned aerial vehicles (UAVs). They are fully/partially automated robotics that use propulsive thrust, aerodynamics, or a combination of both to fly. They consist of an autopilot subsystem that is the UAV’s brain and consists of an array of sensors (such as an inertial measurement unit, barometer, compass, and GPS) and a microcontroller that computes and executes the control command based on the sensory input. Broadly, the mechanical structure of the heavier-than-air UAVs can be divided into two categories: (a) Rotary Wing – This type of drone consists of multiple motors with two to three rotor blades revolving around a fixed mast. The combined thrust generated by all the motors generates the required lift. They are available in different configurations depending on the number of arms: bicopter (2), tricopter (3), quadcopter (4), hexacopter (6), and octocopter (8). The control of such systems is achieved by varying the input thrust and torque to different motors. (b) Fixed Wing – This type of heavier-than-air UAVs design consists of a wing structure with a predefined airfoil. A forward airspeed is generated by the propulsion system when passed over the airfoil of the wings generating the required lift. The control of such systems is achieved by deflection of the control surfaces present on the wings (ailerons) and tail (elevator and rudder) (Fig. 6). With the availability of cheaper electronics and hardware, there has been a tremendous increase in drones for research and industrial purposes (Liew et al. 2017; Kardasz et al. 2016). The interconnection of drone payload to the Internet of things architecture has opened an immense possibility of integrating such systems in the industry.

Fig. 6 A drone used at Audi factory of transportation of parts

2030

A. Misra et al.

Various industries are using drones in indoor environments. These robots with help an array of vision-based sensors keep track of inventory and perform intralogistic tasks such as transporting parts and equipment using grippers, increasing the worker’s efficiency, and minimizing dangerous tasks for the human workforce (Wawrla et al. 2019). In an open environment, these drones are used in inspection and surveillance tasks in areas such as oil rigs and mining sites. An interconnected visual feed from drone surveillance with IOT enables the ground workers to quickly and accurately identify sources of problems and faults. This results in a holistic system to identify, detect, and repair faults in the plant. Swarms of UAVs can be used to perform the same task faster and in an efficient manner such as in humanitarian aid missions. These UAVs communicate with each other and relay attitude and sensor information to intelligently plan individual and swarm path while avoiding obstacles (Agrawal et al. 2020). Manipulators: Industry-specific robots execute multiple functions, such as lifting and positioning items, and activities designed to mimic how a fully functional human arm performs related manual activities. Such kinds of robotic arms are known as robotic manipulators. Every robotic manipulator available in the market consists of a controller and manipulator arm. The manipulator’s performance depends on its payload weight, precision, and speed. However, the manipulator’s structure determines the reach of its end effectors, overall workspace, and work orientation. These robot specifications help in the kinematic and dynamic analysis of the manipulator, which gives vital information on its functioning (Misra and Singh 2019). The manipulators are composed of a series of joints that jointly behave like a human arm. These joints are movable points of the manipulator that enables relative movement between the adjacent links. A robotics manipulator is designed utilizing the rigid links connected by joints with one closed support and one end accomplish a specific function, such as shifting a package from one position to another. The manipulator can be split down into two sections, each with separate functions: 1. Body and arm: The body and arm consist of significant links connected by joints. They are used to transfer items and devices in the workspace and position them in. 2. Wrist: The wrist’s purpose is to provide orientation to the end effector and organize items or instruments in the workspace. A robotic wrist’s structural characteristics are made of two to three small joints. The Robotic Arm Configuration: Robotic architecture usually follows the frame of coordinates with which they are defined. The most common joints that are used in robotic arms are as follows: 1. Prismatic joint (P): In this joint, there is a linear movement along the axis. 2. Revolute joint (R): In this joint, the arm rotates about a given axis. 3. Spherical joint (S): This joint is composed of a spherical ball fitted in a cavity.

79 Robotics in Industry 4.0

2031

The configuration of the robotic arm is specified by the order P, R, or S designation. Based on these joints, the robotic arm can be classified as follows: (a) Cartesian Coordinate System (3P): In this configuration, arms consist of three prismatic joints succeeded by each other. These joints enable the arm to reach in any desired coordinate in a rectangular workspace. These joints enable the robot for linear movement along all the three axes (X, Y, Z) of the coordinate system. The cartesian coordinate robot’s arms have widespread application in the industries for drilling holes with high precision and accuracy. (b) Cylindrical Robot System (RP2): Cylindrical robot system consists of two prismatic joints and one revolute joint. These combinations of joints allow the endpoints of the arm to move in a cylindrical workspace. (c) Spherical Robot System (R2P): Spherical robot follows the polar coordinate system where one prismatic and two revolute joints are used for positioning of items in the workspace. (d) Articulated Robot (3 R): Articulated robots have three revolute joints, and its functioning is similar to a human hand. Due to the similarity between the human hand and articulate robot, it has various applications in industries such as for welding, tooling, and pick and place of objects. (e) Selective Compliance Assembly Robot Arm (SCARA): SCARA has two revolute joints, followed by prismatic joints. Joints used in SCARA are the same as in spherical robots, but the order is different. The two revolute joints in SCARA allow the robots to move in a horizontal plane, and the prismatic joints allow the movement in a vertical direction. SCARA robots play a vital role in assembly operation due to specific compliance characteristics in the x-y plane but stiffer along the z-axis, therefore providing selective compliance. This type of robot is used in pick and drop and assembly operations (Figs. 7, 8, 9, and 10). With advancements in technology, farmers have also started using these robots for fruit harvesting, such as the Kiwi Picker, using two collaborating cartesian arms (Barnett et al. 2020). The optimization and uniform work distribution between the arms minimize the harvesting time. In uncertain times like the Covid-19, there is a demand for contactless services, so convenience stores in Japan have been testing remote control object pick and place on shelves. Coupled with internal sensors to measure the joint positions and external sensors such as RGB and depth cameras, these robots can accurately position themselves in space and perform various tasks. The researchers (Ali et al. 2018) have integrated a camera system with the gripper of the manipulator’s arm and an image processing pipeline that detects objects’ location and information in the camera frame and commands the arm’s controller to grasp and place the objects. GraspNet (Mousavian et al. 2019) framework developed at Nvidia using deep learning on point cloud data detects feasible grasp position on cluttered sets of objects placed in a table. This can be applied to grasp and manipulate diverse sets of objects placed randomly, such as couriers in post offices or cartons in departmental stores. Various methods

2032 Fig. 7 Cartesian coordinate system robot

A. Misra et al.

Z Elevation

X Base Travel

Y Reach

Z Elevation

Fig. 8 Cylindrical system robot

q Base Rotation R Reach (Extend & Retract)

(Misra et al. 2020) have been proposed to develop such manipulators at cheap costs and integrate into manufacturing processes. Continuous research is improving human-robot collaborations for tasks that require human input for operational intelligence, making the robots safe and predictable, enabling humans to work in proximity to the robots in a shared workspace, and extending to industries that are not automated currently (Vysocky and Novak 2016) (Fig. 11).

79 Robotics in Industry 4.0

2033 R Reach (Extend & Retract) f Elevation (Pitch) q Base Rotation

Fig. 9 Spherical system robot Base Rotation

Reach (Elbow Bend) I)

d f Elevation (Shoulder Bend) q Base Rotation

Fig. 10 Articulated system robot

Fig. 11 (a) Car assembly using robots at Tesla factory. (b) Remote-controlled robot stocking shelves at a Japanese convenience store

Legged Robotics: Legged robots have leglike articulated limbs mechanism to provide motion inspired by human locomotion. With the ability to use discrete steps for each foot, these robots can walk over uneven and discontinuous surfaces compared to wheeled robots. This ability enables robots to climb stairs, walk

2034

A. Misra et al.

Fig. 12 (a) Atlas robotdeveloped by Boston Dynamics. (b) ASIMO robot developed by Honda

on irregular terrains, and turn without slipping by selectively adapting the leg configuration according to the ground surface. Studies have shown that these robots are energy-efficient compared to wheeled rovers as they require less energy to come out of surface depressions. The state of such systems’ art development is a long way to go with intricate build design, low speeds, and difficult model stability and control algorithms (Tenreiro Machado and Silva 2006). While there are many types of humanoid robots, the most common robot configurations based on the number of legs are (Fig. 12): 1. Biped/Humanoid Robots: These robots mimic human motion with a two-legged structure. The development of such robots has been restricted to academia due to difficult control and dynamic balancing algorithms. Its market is expected to grow significantly in the coming year with applications in human assistance tasks. These tasks include a helper in daily life, inspection, and maintenance, most importantly in areas where human life is at risk, radioactive nuclear power plants, humanitarian aid, disaster relief, and unexplored caves and mines. ASIMO robot developed by Honda is aimed at daily use. It can recognize its surroundings, objects, and people using a host of sensors such as the stereo camera system embedded in the eyes of the head and laser and infrared sensors in the torso. The robot interacts using hand and head gestures, such as shaking hands and nodding, and can walk and avoid obstacles (Sakagami et al. 2002). One of the most advanced robots developed by Boston Dynamics is the Atlas humanoid robot. Driven by a hydraulic system designed to be used in emergency, search, and rescue tasks, this robot can perform acrobatic movements alongside tasks

79 Robotics in Industry 4.0

2035

such as opening doors and valves, operating tools, driving vehicles, climbing stairs and ladders, and avoiding and removing objects. 2. Quadruped Robot: Quadruped robots have four limbs and mimic animals such as dogs. Due to the presence of four limbs, these robots are more stable than the bi-legged robotics. They can navigate rugged terrains while carrying heavy payloads. These robots’ four-legged design gives them an unparalleled advantage over wheeled robots, such as walking on uneven surfaces, climbing stairs, and even turning in place. With onboard sensors such as RGBD cameras and LiDAR, these robots can efficiently develop environment maps and plan navigation paths. These robots have immense applications in surveying, monitoring, warehouse management, delivery, and disaster response. Commercially available robots such as Spot (Boston Dynamics 2020) by Boston Dynamics are being used in industries to carry various tasks. These tasks include but are not limited to remote monitoring of oil and gas plants, an inspection of constructions by creating digital models, and autonomous routing through radiation dense areas. The userspecific payload can be mounted on the robot with customization options such as a robotics arm or mapping sensors. Another similar robot developed at ETH Zurich is the ANYmal quadruped robot(Hwangbo et al. 2019). This robot has similar structure as well as application but, for the meantime, largely restricted to academia (Fig. 13). Robotic Ground Vehicles: One of the most widely deployed robots, ground vehicles, are a class of autonomous systems that operate in contact with the surface through wheels or continuous tracks. Equipped with various application-specific capabilities, these robots can perform tasks ranging from material handling in warehouses and workshops to uneven exploration and navigation in agricultural fields and mines. Based on the area of use application, these can be classified into two broad categories:

Fig. 13 (a) ANYmal quadruped robot developed by ETH Zurich. (b) Spot robot developed by Boston Dynamics

2036

A. Misra et al.

1. Automated Guided Vehicles: AGVs are extensively used in various stages of manufacturing in indoor industries and factories. They perform tasks such as handling raw materials, transportation of intermediate outputs through the production cycle, storage of finished products, and loading trailers with custom orders. These unmanned autonomous systems operate in a known environment with predictable trajectories. The guidance and navigation use externally mounted sensors and technology based on the industry and localization accuracy level. Some standard methods are: (a) Magnetic Tapes: Magnetic tapes and tags are executed on defined paths that assist the AGV in calculating the position and next action. These vehicles have additional obstacle avoidance mechanics such as a bumper or LiDAR sensor that assist in stopping or rerouting the path (Fedorko et al. 2017). (b) Laser Navigation: A wireless mode of localization and navigation, these methods consist of a rotating sensor header with a transmitter and a receiver. The laser reflects from reflective markers mounted on the walls that help estimate the position based on a stored map. The article (Jung et al. 2014) used an unscented Kalman filter (UKF) over motor encoders, a laser navigation system, and a gyroscope. (c) Natural Feature Navigation: This method uses one or more localization and range finding sensors such as depth camera, RGB cameras, and LiDAR. This method fuses these sensor readings and tries to position the robot based on a predefined map generated using the same sensors. This navigation system is advantageous in dynamic environments as it is fast to set up and flexible toward failures such as the one developed for hospital logistics (Baˇcík et al. 2017). Using ground vehicles drastically increases accountability, automatic workforce, and output while reducing damages and operational costs. 2. These are more robust systems developed for operations on uneven and outdoor areas, especially in hazardous environments for human operations. The sensors mounted on these rovers help the robotic system map, localize, plan, and navigate the path through unknown terrains. Simultaneously sending data for analytics on a central server to the remote operator for intelligent decision-making or overriding, the basic modules of UGV can be classified as: 1. Perception: The sensor and processing modules are responsible for environmental understanding. Sensors such as depth cameras and LiDAR enable the robot to create a rich and precise map of the surrounding and accurate localization. This is executed through a technique known as Simultaneous Localization and Mapping (SLAM). This map is updated continuously with every sensor data and information. 2. Planning: The planning modules compute the trajectory of the rover using the perception module data and the rover’s kinematics model. This calculation calculates the path based on some global and local planning algorithms. 3. Control: This module is a controller for the UGV and converts the generated trajectory to steering angle, wheel speeds, and acceleration.

79 Robotics in Industry 4.0

2037

Fig. 14 AGVs developed by Amazon for warehouse management

UGVs are deployed in numerous use cases such as precious agriculture, mining, inspection, humanitarian aid, and disaster relief. The paper Quaglia (2019) presented the design of a UGV that moves over unstructured terrains while providing a self-stabilizing charging and landing pad for UAVs. It also incorporates a robotic manipulator arm that can be used for crop harvesting or weeding. The article (Pierzchała et al. 2018) implemented a mobile robot capable of mapping forests by generating 3D point cloud maps using a stereo camera, IMU, GPS, and LiDAR sensors. Nine teams participating in the DARPA Subterranean Challenge have demonstrated a multi-agent system of UAVs, bipeds, and UGVs that can effectively map, navigate, and find artifacts in unknown, degraded, and dynamic underground settings (Ackerman 2019) (Fig. 14).

Sensors and Perception Various sensors are used in robotics both for internal feedback control and external interaction with the environment. A sensor is a window to the environment for a robot. They allow the robot to understand and measure the objects’ geometric and physical properties in their surrounding environment. Humans and animals have different sensors. For instance, when humans wake up, they do not have to look down to check their body parts. The neurons in muscles send signals to the brain and determine the state of each muscle.

2038

A. Misra et al.

Similarly, the robots have sensors such as encoders, resolvers, and potentiometers that send signals to the controller to determine each joint’s state. Moreover, humans have a sense of vision, smell, and touch to interact with the outside world. Similarly, robots may possess similar sensors to communicate with the external world, which perform similar functions. A sensor’s performance can be defined by its various characteristics such as economy, performance, ease of application, and applicability of the sensor. Some essential characteristics to be compared are: 1. Sensitivity – The change in the output relative to the unit change in the input or the physical partner’s minimum input will create a detectable output change. 2. Range – Range is the difference between the minimum quantity a sensor can measure and the maximum quantity. It can also be defined as the difference between the sensor’s lower and upper limits. 3. Repeatability – Repeatability refers to the degree of reproducibility of the measurement. If a sensor measures the same value of input for a repeated number of times, the output may differ each time measured by a small value. Repeatability measures how output value varies each time. 4. Resolution – Resolution is the minimum detectable incremental change of measure quantity, which can be detected in the output signal. It is the least count of measuring device that can measure the minimum step size change in the measuring sensor range. 5. Accuracy – Accuracy is the relative difference between the actual value and the measured value. The smaller the difference, the higher the accuracy of the sensor. Various types of sensors are available in the market with different characteristics and measuring value: 1. Proprioceptive Sensor: These sensors measure the internal state of the system. (a) Wheel Encoders: These encoder sensors are attached with the wheel of the robot to measure the distance, speed, acceleration, and position of the wheel. They consist of a revolving disk with alternative window patterns. This is attached to the wheel across a fixed LED and photodetector. When the wheel rotates, it generates an alternating LED light pattern on the photodetector, which provides position information. (b) Torque and Force Sensor: This sensor is used to measure the torque and force generated in rotating joints and motors. Using a combination of six precisely positioned strain gauges, they can measure the component of force and torque in the three axes. (c) Inertial Measurement Unit: This sensor is a combination of accelerometer and gyroscope sensors used to measure the acceleration and angular rate of all three axes (X, Y, and Z). They are essential components in the state estimation of autonomous aerial and underwater robots as they help to compute the linear and angular position. (d) GNSS Sensor: Global Navigation Satellite System (GNSS) is a geospatial positioning sensor that uses various satellite systems such as GPS,

79 Robotics in Industry 4.0

2039

GLONASS, BeiDou, and Galileo to estimate the longitude, latitude, and altitude of the robot. This sensor is used in the outdoor navigation of robots alongside other sensors such as IMU and magnetometer. GNSS sensors work by calculating the distance traveled by signal to reach from satellites’ constellations in medium earth orbit (MEO). These satellites move on predefined trajectories, and by measuring the distance from four or more satellites, they can accurately calculate the system’s position. (e) Magnetometer: Also known as a compass, these sensors measure the robot’s direction and orientation concerning Earth’s magnetic north pole at a particular location. They do this by measuring the change in voltage induced across the semiconductor due to Earth’s field, also known as the Hall Effect. 2. Exteroceptive Sensor: These sensors are used to measure and collect information about the robot’s environment. (a) Cameras: Cameras or image sensors play a role similar to human eyes on robots. They provide a vast amount of information about the surrounding environment that can be exploited for dynamic interactions such as obstacle avoidance, human and activity recognition, object manipulation using traditional computer vision techniques, and, more recently, artificial intelligence techniques. The most common type of robotics sensor is the Charged Coupled Device (CCD). They contain numerous pixels that output analog values corresponding to the photons’ number that reach the sensor region through the lens. When used in a stereo setup, they can be used to estimate depth from images. (b) IR Sensor: Infrared sensors work by measuring the IR radiation in the environment. The sensor consists of an LED that emits infrared radiations and an IR receiver. The IR light from the LED strikes the surface in front of the sensor and reflects. This, when detected by the receiver, helps in obstacle detection. (c) Ultrasonic Sensor: This sensor is used to measure distance/detect obstacles using ultrasonic waves. The sensor design consists of an ultrasonic emitter that emits waves using piezo crystals and a receiver that detects the reflected waves from surfaces and obstacles. By calculating the flight time between transmitting and receiving the wave, these sensors can measure the distance between the sensor and the object’s surface. (d) LiDAR Sensor: Light Detection and Ranging sensor is a class of sensors that uses light pulses to generate high-resolution maps of the surrounding. They are an active source of light; hence, these sensors can work in night and foggy conditions. The sensor consists of a rotating light beam that pulsates the light of a specific wavelength. This light pulse is scattered by the surrounding and finally received by the sensor. By calculating the time of flights, these sensors provide highly dense point cloud information that can be easily integrated with other sensors such as IMU and GNSS to generate a map of the environment. LiDAR is used by drones to generate survey maps of the area or by legged robots and ground vehicles to localize and plan obstacle-free paths.

2040

A. Misra et al.

(e) Microphones: Microphones as a sensor provide audio interactions with robots such as automated voice services. The audio received through the microphone is passed through a speech recognition software that converts audio to machine-understandable commands.

Path Planning in Robots Path planning is an essential challenge for an autonomous robot. There exist numerous paths between the start and the goal location to get the feasible path some parameters are optimized such as shortest distance, smoothness of the path, minimum energy consumption, etc. The path planning can be categorized into twofolds: (i) local path planning and (ii) global path planning. Local path planning is the one in which the robot system does not have complete knowledge of the surrounding environment, while in the global path planning, robot control system has complete information such as a map so that it can use a preplanned path to reach the final destination. Global path planning methods, however, show limited applications due to a less robust terrain uncertainty in comparison to local path planning methods that show more flexibility in partially known/unknown environments and provide an optimized path. Use of a robot path planning is introduced in restorative and surgical uses, individual assistance, safety, storage and circulation applications, and additionally ocean and space investigation, robotic-guided vehicles for moving merchandise in a plant, unmanned bomb transfer robots, and planet investigation robots. The path planning algorithm for a known environment is based on classical approaches. These are traditional algorithms and have limited intelligence, while local navigational approaches are known as reactive approaches as they are more intelligent and able to control and execute a plan autonomously.

Artificial Potential Field APF algorithm (Khatib 1986) presented where the goal and obstacles act as the charged surfaces and total potential in the system creates the reactive force on the robot. This reactive force attracts the robot toward the goal and keeps the robot away from the obstacles. In this approach, the robot follows the negative gradient approach to avoid the obstacles and reach the goal. The paper Garibotto and Masciangelo (1991) presented the application of this algorithm for mobile robots using a vehicle model. Such algorithms are used in scenarios where the environment is partially or fully known such as object transportation in an indoor factory (Fig. 15). In the paper Borenstein and Koren (1989), a solution was presented to the problem of local minimum conditions while considering the dynamic properties of robot navigation through a cluttered environment. The analysis of APF in the dynamic environment for the avoidance of obstacles is brought out in papers Ge

79 Robotics in Industry 4.0

2041

Target

Target

Obstacle

Repulsive force

Resultant force Obstacle Attractive force

Start

Start

Fig. 15 Mobile robot navigation by the APF approach

and Cui (2002) and Montiel (2015). Numerous improvements and variations to the algorithms are developed such as the use of electrostatic law which generates a collision-free path in real-time using electrostatic potential. Other variations such as developed by Huang (2009) took into consideration velocity control mechanism to understand the location of moving obstacles while reaching for the goal.

Grid-Based Method For an n-Dof robot, the number of avoidable configurations is called free space Cfree . The remaining region in the entire C space is termed as the obstacle region. The grid-based method discretized the entire configuration space in points. Robots are allowed to move to an adjacent free grid point using a heuristic-based planner such as A* algorithm that guides the robots toward the goal location following the shortest route. Changing the size of the grid point produces different results. Coarser grid points will make the search faster but might fail in narrow portions. Similarly, finer grid points increase the size of C space to be explored, thus increasing planning time. Moving to higher dimensionality C space exponentially increases the time to plan the path, making this algorithm unsuitable for 6-DOF manipulators. In case of a change in the configuration space of the robot by dynamic obstacles, the algorithm has to repeatedly search each grid point to generate the optimal path. This makes it unfavorable for real-time planning.

Cell Decomposition (CD) Approach This approach divides the area into nonoverlapping grids (cells) and uses connectivity graphs to navigate from one cell to another to accomplish the goal. During the navigating process, perfect cells (cells without obstacles) are selected to achieve path planning from the start point to the target point. Tainted cells (obstacle-

2042

A. Misra et al.

Goal

14

11

2

4 6 9

15

10

5

3

13

1 12

8

7

Robot position

Fig. 16 Cell decomposition

containing cells) present in the trajectory are further divided into two new cells to obtain a perfect cell, and this perfect cell is added to the sequence while examining the optimal route from the start point to the target point. In this approach, the start and end cells denote the starting and ending positions of the path. The required path is shown by the sequence of the pure cells connected (Fig. 16).

Rapidly Exploring Random Tree A rapidly exploring random tree (RRT) is an algorithm developed (Lavalle 1998) and designed to efficiently search non-convex high-dimensional spaces by randomly building a space-filling tree. The tree is developed step-by-step from randomly drawn samples from the search space and has a bias to develop/grow toward the unsearched areas. RRT trees handle complex problems with hurdles and differential constraints and have been widely adopted in motion planning of autonomous robots. These trees can be seen as a technique to develop open-loop trajectories for nonlinear problems with general space constraints. The root of the RRT tree is in the starting configuration using samples drawn randomly from the search space. As each sample is drawn, a connection is attempted between it and the nearest tree state. If the connection is possible, this will result in the addition of a new state to the tree. The length of the connection between the tree and the new state is often limited by a growth factor. If the random sample distance is greater from its nearest tree state than this limit allows, a new state will be used at the maximum distance from the tree along the line to the random sample instead of the random sample itself. Random samples can then be seen as controlling the direction of tree growth while the growth factor determines the rate of tree growth. This regulates the bias of spatial-filling condition of RRT while restricting the size of incremental growth. RRT growth can

79 Robotics in Industry 4.0

2043

be skewed by increasing the probability of sampling states from a specific area. Most of the practical implementations of the RRTs make use of this to guide the search toward the goals of the planning problem. This is achieved by introducing a small sampling probability for the goal of the state sampling procedure. The higher the probability, the more rapaciously the tree grows toward the goal.

Probabilistic Road Map The general concept behind PRM is to take random samples from the robot’s configuration area, test them to see if they are in free space, and use the local planner to try to connect these configurations to other configurations in the immediate area. Starting and target configurations are added, and a graph search algorithm is applied to the resulting graph to determine the path between start and goal configurations. The probabilistic road map planner consists of two phases: the construction phase and the query phase. A road map (graph) is constructed during the construction phase, approximating the movements that can be made in the environment. First, a random setup is created. It is then connected to some neighbors, typically either the nearest k neighbors or all neighbors less than some predetermined distance. Configurations and connections are added to the graph until the map is sufficiently dense. This map can be reused in multiple planning problems for the same region. During the query phase, the Dijkstra algorithm determines the path, and the graph is also connected to the start and end configurations. Given some comparatively low conditions on the shape of free space, PRM is likely probabilistically complete, which means that as the number of points sampled increases without a bound, the probability that the algorithm will not find a path if one exists is zero. The convergence rate is dependent on the free space’s visible parameters, in which the local planner determines the visibility quotient.

Fuzzy Logic It is used in situations where there is a high level of uncertainty, complexity, and nonlinearity. Pattern recognition, automatic control, decision-making, and data classification are just a few of them. The hypothesis of the FL framework is encouraged by the remarkable human ability to process perception-based information. It uses the human-provided rules (if-then) and converts these rules to their mathematical equivalents. This simplifies the work of the system designer and the computer to obtain more accurate information on how systems should operate in the real world and is therefore used to plan the path of a mobile robot (Fig. 17). The navigation based on fuzzy (Sugeno) is presented (Zavlangas and Tzafestas 2003) for the omnidirectional mobile robot. The authors (Castellano et al. 1997) developed an automatic fuzzy rule generation system for the early detection of obstacles for effective navigation. Al-Khatib and Saade (2003) and Lee et al. (2012) in their paper have presented fuzzy logic for path planning in dynamic and changing environments using a data-driven approach.

2044

A. Misra et al.

Fuzzy membership function

Input as a crisp value

Fuzzification

Fuzzy IF-Then rules fuzzy linguistic variables

Fuzzy controller

Output as a crisp value

Defuzzifier

Fig. 17 Fuzzy logic framework X1

X2 Input parameters X3

Wk1

Wk2 Σ

f (.)

Yk Output

Wk3 Summing junction

Xk

Uk

Activation function

Wkk

Fig. 18 Neural network

Neural Network Artificial neural network is a kind of machine learning algorithm in which we have simple and highly interconnected elements. In the network, we have basically three types of layers called the input layer, hidden layers, and output layers. These elements transfer the information by their capability of dynamic state response to external inputs. A basic neural network consists of a fully connected network from input to the output layer. Each node receives weighted data from all the nodes of the previous layer. This node then sums the input and passes through a nonlinear activation function (Fig. 18). In this, we give input to the input layers which then contact with the hidden layer which then after iteration gives output to the output layer. Janglova (2004) presented the application of a neural network for wheeled mobile robot navigation in a partially unknown environment. The paper proposed two neural networks for computing a collision-free path. The two networks work in conjugation with each other. This first network uses the sensor data to classify free space in the environment, and the second network uses this information to find local trajectory by avoiding obstacles near the robot. Fast Simultaneous Localization and Mapping

79 Robotics in Industry 4.0

2045

Technique (Li et al. 2015) uses neural networks to remove the drift accumulated due to sensor noise and inaccurate model. The neural network coupled with the SLAM algorithm generates an accurate map of the environment that helps the mobile robot to navigate through the environment collision-free.

Ant Colony Optimization Ant colony optimization (ACO) is a population-based approach to solve combinatorial optimization problem developed by Dorigo and Gambardella (1997). This swarm intelligence algorithm is inspired from the behavior of ants in colonies. The worker ants find the shortest path between the food source and their nesting site. The ACO algorithm is already applied to various fields of science and engineering such as job-shop scheduling, vehicle routing, quadratic assignment problem, traveling salesman problems, graph coloring, and many more (Fig. 19). Ant colony optimization can be used for planning of obstacle-free path for mobile robots such as presented by Guan-Zheng (2007). In the paper, the author presented the application of ACO for real-time path planning of mobile robots. ACO provides higher convergent speed to be used for real-time applications and is computationally efficient.

Bacterial Foraging Optimization (BFO) Algorithm This nature-inspired algorithm (Passino 2002) aims to optimize the path planning method through mimicking the behavior of E. coli and M. xanthus bacteria. These microorganisms search for food by maximizing the use of energy achieved per unit time. The bacteria communicate with each other using a chemical gradient that transmits specific signals to others. In a given area, the bacteria always travels toward nutrient-rich regions. As the organism reaches sufficient food levels, they split into two equal parts, else they disperse in the surrounding and die. The Food

Food

?

Food

Food

?

Preferred path

Nest

Nest

Nest

Fig. 19 Behavior of ant while searching for food

Nest

2046

A. Misra et al.

bacteria with less nutrient sends out warning signals to others nearby, and those with higher-nutrient surrounding attract others. Overtime, they get accumulated in nutrient-rich regions of the environment for where they get dispersed to newer areas of the environment. Inspired by this behavior, BFO algorithm is used for mobile navigation such as proposed by Coelho (Coelho and Sierakowski 2005) in a static surrounding environment using different probability distributions. BFO can also be used for navigation of unmanned aerial vehicles alongside a proportional, integral, and derivative controller (PID controller) (Oyekan and Hu 2010). This reduces the long process of parameter tuning and gets optimal parameters navigation in the space.

AI in Robotics and Industry 4.0 The fields of machine learning, robotics, computer vision, speech recognition, natural language processing, etc. combine to form artificial intelligence. ML and AI have been up and coming technologies for the past decade, while such methodologies depend on the specific developer and operations used. Industrial AI is a systematic method that includes developing, deploying, and validating models in real time. There is a difference between intelligent manufacturing and smart manufacturing. Intelligent manufacturing has evolved into smart manufacturing with advanced data acquisition, communication, and analytics. However, a quantitative effect on industrial growth is yet to be observed, various processes have been developed, and monitoring has improved drastically. An industrial AI environment has been proposed (Lee et al. 2018) as a method for better understanding and implementing different technologies. The key elements are ABCDE: A, analytics technology; B, big data technology; C, cyber technology; and E, evidence. Analytics is the centerpiece of AI. The information stream is provided by big data technology and cloud. Industrial AI, however, requires domain expertise and knowledge, which include the ability of the physical variables in the process and how they are associated with the software components of the plant. This also includes knowing the difference in different machines, how they vary, and how this change affects the system. It is essential to know the system thoroughly so that the right quality and quantity of data are stored. This is also essential for data mining, i.e., to mine patterns in the data and find meaningful information, identifying the main aim of the problem and implementing artificial intelligence into this system to solve them. There are five co-enablers to industrial artificial intelligence, ABCDE:

Data Technologies Acquiring meaningful data with significant metrics across different modalities is very important. Hence, data technologies enable 5C architecture and are components of the “Smart Connection” step – data communication in smart manufacturing.

79 Robotics in Industry 4.0

2047

Hence for interaction between smart manufacturing machines with continuous relaying of data from the factory to the cloud, the physical components have to be synced with cyberspace and compute subsequent analytics.

Analytics Technologies This part converts the physical data from sensors to vital information for further processing. This includes data-driven modeling of uncertain correlations, unpredicted behaviors, and various other useful information. Analytics ultimately improves productivity and innovation through integration. The information sourced from this can be used further for health asset prediction and other predictive features.

Platform Technologies Data storage, processing, and feedback require hardware architecture. A compatible platform architecture for analyzing data is a major deciding factor for realizing smart manufacturing characteristics such as agility and complex event processing. Three major types of platform configurations are generally found – stand-alone, embedded, and cloud. The cloud platform provides random service deployment, customization, and high scalability.

Operations Technologies The operational decisions and actions that process the industry data are vital to control the whole plant’s execution. The inputs to a decision are extensive and range from machine to machine communication to changing specific operational and process parameters to operator data. This is the last step which makes the plant adept in: 1. 2. 3. 4.

Self-comparison Self-configuration Self-awareness Self-prediction

Application of Robots 1. Aerospace: The aerospace industry relies heavily on robots to perform various tasks, from construction to flight readiness. Construction of aircraft requires exceptionally high precision and accuracy that is repetitive such as drilling holes and painting. Robot manipulators with embedded vision systems can accurately locate and perform this task. These robots are faster and cost-effective

2048

A. Misra et al.

as compared to manual labor with a smaller window of error. Similarly, the task of painting and welding is hazardous to humans when performed for long durations. Robots manipulators can work around the clock to coordinate with other robots to speed up the process and reduce the operational hazard. Airbus has recently demonstrated the use of drones for aircraft maintenance. The use of camera-mounted drones significantly increases the quality of surface inspection for cracks and faults compared to human vision and speeds up the process. These drones provide real-time operator overview and software analysis (Fig. 20). 2. Construction: To date, the construction industry is one of the least technologically enabled sectors with labor-intensive tasks. Robots provide numerous advantages to the construction industry with a faster, cheaper, and consistent workforce. A 3D printing robot is a fast-growing use case; complex, layered structures and assembly of homes, buildings, and bridges can be printed with cement providing standardized design. An entire two-story house was built by a Belgian company Kamp C in 15 days using a 3D printing machine with concrete material. It is also projected that robots will be used for bricklaying where they are fed with bricks and CAD design, and the robot efficiently lays the bricks in a fraction of time. Demolition robots are another type of robot that can be used to break down walls and demolish concrete. They tend to be slower than human crews but are far safer to use in hazardous areas. As companies are looking to make construction projects faster, cheaper, and safer, robots’ demand will rise in the coming years (Fig. 21). 3. Agriculture: There is the immense application of robots in the agriculture industry, such as crop assessment and monitoring, seed plantation, weed control,

Fig. 20 Drone used by Airbus of aeroplane inspection

79 Robotics in Industry 4.0

2049

Fig. 21 3D printing of house by Kamp C

and crop harvesting. There is a steady rise in these robots due to a labor shortage and increased food demand to cater to the growing population. Robotic manipulators mounted on autonomous ground vehicles replace backbreaking manual fieldwork, such as picking up delicate fruit and vegetables. Drones are being used to spray pesticides and insecticides onto the crop, which reduces long exposure of labor to such chemicals. Further, the drones are being used to gather images of the field to gauge the crop’s health and count the fruit and vegetable yield. Another area of work is weed removal that is essential for crop yield. Weed management robots use advanced artificial intelligence software to detect and differentiate between crop and weeds and remove using weeders. 4. Manufacturing: To meet the ever-growing demand for products and supplies, industries need to be dynamic in their manufacturing process. Automating the manufacturing sector with intelligent robots that can sense the environment and make decisions based on user input strives to improve productivity and safety in the manufacturing process. Robots in the manufacturing industry fulfill a variety of roles. They can also manufacture complex parts (Mittal et al. 2020) through tedious processes accurately which are not possible by human counterparts. They are used from repetitive tasks that produce high outputs while maintaining accuracy in human assistive tasks. Work performed by some robots in industries are: (a) Material handling (b) Welding and drilling (c) Part assembly (d) Inventory management (e) Warehouse logistics

2050

A. Misra et al.

5. Mining: The age-old industry of extracting minerals and fossil fuels from the Earth’s crust is moving toward a huge change by introducing robots in almost all aspects of mining stages: (1) prospecting, (2) exploration, (3) development, (4) exploitation, and (5) reclamation (Marshall et al. 2016). The major driving forces behind the need for robots in the mining industry are as follows: (a) Dangerous working environments: Some mines are located in difficultto-reach places such as mountainsides, high-altitude regions, and deep underwater trenches with extreme operating conditions for humans to work. Robots can easily be deployed in such areas, which results in faster operation at minimum cost. (b) Health: The mining sites are, often, a source of huge pollution that over time affects the health of workers and laborers. The extreme conditions make mining activities challenging to execute and pose severe health risks. (c) Operational output: Using robots may consider the labor shortage and work schedules around the clock, thus producing higher efficiency. With advanced sensors and tools, these robots can analyze and work faster than humans. Industrial robots also provide secondary services such as surveying the area for potential sites on exploration, carrying ores from mines to storage facilities, underground drilling for breaking up ores, and human assistive tasks. Various companies, such as Caterpillar, design an autonomous fleet of haul trucks with onboard intelligence to move heavy materials even in congested areas (Caterpillar 2020). CSIRO research group developed automated swing loading technology for electric mining shovels that used laser sensors to create a model of the surrounding shape and form, position, and the position of other transport vehicles such as trucks. This technology helps traffic management around the swing loader by calculating the loading path position (CSIRO 2006) (Fig. 22). 6. Health care: One of the most time-critical industries use robots to help in reducing repetitive tasks for health officials, nursing staff, and doctors. AGVs help in the transportation of medicines, essential supplies, and reports around the facilities. Artificial intelligence is used to detect patterns and anomalies through studying health-care data and assists in accurately diagnosing diseases. The paper Bajwa (2020) presented a deep neural network that assisted in computer-aided diagnosis of twenty-three different skin diseases and achieved a state-of-the art accuracy of around 80%. The Covid-19 pandemic has emphasized the need for robots in the health sector, such as in the contactless disinfection of hospitals and clinics. Mobile ground robots such as one developed by UVD Robotics use powerful short-wavelength ultraviolet-C light that can shred the DNA strand of any microorganism that is within its reach (Ackerman 2020). Another area where robots are emerging is surgical tasks. The da Vinci surgical system developed by Intuitive [Intuitive] gives the surgeon the ability to perform minimally invasive surgery (MIS) with superior precision. These systems extend the capability and range with multiple interactive robotics arms and a magnified view of the surgical area.

79 Robotics in Industry 4.0

2051

Fig. 22 (a) Cat 793F autonomous truck for material handling. (b) Automatic mining machinery by CSIRO

Challenges in Robotics There are various challenges to robotics in Industry 4.0 (Pira 2018; Infopulse 2019; Lewis 2018). These include broad aspects as well as specific targeted problems. These pose major challenges to the widespread adoption of Industry 4.0 and have to be countered. These are explained as follows:

Cybersecurity With increasing interconnectedness and data communication, cybersecurity is a significant issue in Industry 4.0 (Liew et al. 2017). Cyberattacks can execute malicious activities with whole industries, devices, and enterprises. Iran’s uranium enrichment operations were disrupted by a virus developed by Israeli intelligent services, Stuxnet, highlighting how dangerous a computer virus could be. There is a vital need to develop and adopt integrated cyber solutions to safeguard vital systems and equipment. Various research articles (Singh et al. 2020) have suggested methods to identify and block intrusive transactions in large networks. Such methods can be adopted in the Industry 4.0 framework to make it more secure.

Variation in Labor Forces Automation and robotics are leading to the reduction of low-skilled manual labor. As they will be redundant and replaced by their more accurate counterparts, this will improve the necessity for highly skilled workers such as roboticists and software engineers. Currently, industrial plants are located near strategic centers of cheap labor; this will allow them to relocate near customers or raw materials.

2052

A. Misra et al.

New Materials The gears and mechanical and electromechanical actuators are essential to the operation of various robots today. However, the latest research shows the advent of tactile robotics, i.e., artificial muscles, to be induced in robotic operations. Such bioinspired motional muscles are softer and can operate as a human and detail through shrinkage and expansion of the material. However, using soft robotics compromises the robot’s strength, so it has to be a calculated trade-off.

Artificial Intelligence Artificial intelligence in industrial robots is somewhat restricted to object detection and pattern recognition. There is a need to further introduce the vast spectrums of AI in Industry 4.0, such as NLP, to improve the Industry 4.0 framework. It can be also further applied in intelligent tutors (Mian et al. 2019) to teach supervisors about various assisting methods.

Ethics One of the most impending impacts of robotics in Industry 4.0 is regulation and ethics. Firstly, it reduces human operators’ accountability in the industry due to the increased reliance on robotics. A human should always handle sensitive tasks such as manual inspection for quality and supervision. With the current rate, they would soon be AI-based. Increased unemployment is a significant factor in automation and robotics. Increased training programs and industrial education would solve this.

Conclusion Through the course of this chapter, the various components of robotics are explained in the light of Industry 4.0. Industry 4.0 and industrial IoT provide a holistic environment which comprises smart communication, analytics, and self-correction. The sensors enable machines to communicate with the physical world and learn stochastically and achieve a level of intelligence. Recently, various new modelling techniques such as probabilistic modelling of uncertainty modelling and digital twin machine learning models are being integrated to the Industry 4.0 decision-making structures. These modelling techniques reduce the error in predictive analytics and further improve the smart industry framework. Human and robot collaboration is also being continuously tested to optimize the decision frameworks as a human operator imparts a high level of precision to operations. A substantial improvement in customer satisfaction, productivity, efficiency, and delivery speed is observed

79 Robotics in Industry 4.0

2053

due to the introduction of robotics in industrial development. Various applications and challenges faced by robotics have also been discussed in this chapter. The future development in this domain would source from solving these problems and broadening the applications of robotics.

References Ackerman E (2019) DARPA Subterranean challenge: teams of robots compete to explore underground worlds. https://spectrum.ieee.org/automaton/robotics/industrial-robots/subt-thenext-darpa-challenge-for-robotics Ackerman E (2020) Autonomous robots are helping kill coronavirus in hospitals. In: IEEE Spectrum: Technology, Engineering, and Science News 2020 Mar 11. https://spectrum.ieee.org/auto maton/robotics/medical-robots/autonomous-robots-are-helping-kill-coronavirus-in-hospitals Agrawal A, Gupta A, Bhowmick J, Singh A, Nallanthighal R (2020) A novel controller of multiagent system navigation and obstacle avoidance. Procedia Comput Sci 171:1221–1230 Ali MH, Aizat K, Yerkhan K, Zhandos T, Anuar O (2018) Vision-based robot manipulator for industrial applications. Procedia Comput Sci 133:205–212 Al-Khatib M, Saade JJ (2003) An efficient data-driven fuzzy approach to the motion planning problem of a mobile robot. Fuzzy Sets Syst 134(1):65–82 ˇ Baˇcík J, Durovský F, Biroš M, Kyslan K, Perdukova D, Padmanaban S (2017) Pathfinder– development of automated guided vehicle for hospital logistics. IEEE Access 5:26892–26900 Bajwa MN, Muta K, Malik MI, Siddiqui SA, Braun SA, Homey B et al (2020) Computer-aided diagnosis of skin diseases using deep neural networks. Appl Sci 10(7):2488 Barnett J, Duke M, Au CK, Lim SH (2020) Work distribution of multiple Cartesian robot arms for kiwifruit harvesting. Comput Electron Agric 169:105202 Borenstein J, Koren Y (1989) Real-time obstacle avoidance for fast mobile robots. IEEE Trans Syst Man Cybern 19(5):1179–1187 Boston Dynamics (2020) Boston Dynamics Spot. https://www.bostondynamics.com/spot Castellano G, Attolico G, Distante A (1997) Automatic generation of fuzzy rules for reactive robot controllers. Robot Auton Syst 22(2):133–149 Caterpillar (2020) A World leader in autonomous mining. https://www.caterpillar.com/en/news/ caterpillarNews/customer-dealer-product/a-world-leader-in-autonomous-mining.html#: :text=T he%20Caterpillar%20autonomous%20fleet%20is,without%20a%20lost%20Dtime%20injury Coelho LDS, Sierakowski CA (2005) Bacteria colony approaches with variable velocity applied to path optimization of mobile robots Colgate E, Bicchi A, Peshkin MA, Colgate JE (2008) Safety for physical human-robot interaction. In: Springer handbook of robotics. Springer, Cham, pp 1335–1348 CSIRO Robotics and Autonomous Systems Group (2006) Shovel Loading Automation. https:// research.csiro.au/robotics/shovel-loading-automation/ Dorigo M, Gambardella LM (1997) Ant colony system: a cooperative learning approach to the traveling salesman problem. IEEE Trans Evol Comput 1(1):53–66 Fedorko G, Honus S, Salai R (2017) Comparison of the traditional and autonomous agv systems. In: MATEC Web of Conferences, vol 134, p 00013. EDP Sciences Garibotto G, Masciangelo S (1991, June). Path planning using the potential field approach for navigation. In: Fifth International Conference on Advanced Robotics’ Robots in Unstructured Environments. IEEE, pp 1679–1682 Ge SS, Cui YJ (2002) Dynamic motion planning for mobile robots using potential field method. Auton Robot 13(3):207–222 Grau A, Indri M, Bello LL, Sauter T (2017) Industrial robotics in factory automation: from the early stage to the Internet of Things. In: IECON 2017-43rd annual conference of the IEEE Industrial Electronics Society. IEEE, pp 6159–6164

2054

A. Misra et al.

Guan-Zheng TAN, Huan H, Sloman A (2007) Ant colony system algorithm for real-time globally optimal path planning of mobile robots. Acta Automat Sin 33(3):279–285 Huang L (2009) Velocity planning for a mobile robot to track a moving target – a potential field approach. Robot Auton Syst 57(1):55–63 Hwangbo J, Lee J, Dosovitskiy A, Bellicoso D, Tsounis V, Koltun V, Hutter M (2019) Learning agile and dynamic motor skills for legged robots. Sci Robot 4(26):eaau5872 Infopulse (2019) The main benefits and challenges of INDUSTRY 4.0 adoption in manufacturing. https://www.infopulse.com/blog/the-main-benefits-and-challenges-of-industry-4-0adoption-in-manufacturing/ Intutive, da Vinci Systems – About da Vinci Systems. https://www.davincisurgery.com/da-vincisystems/about-da-vinci-systems Janglová D (2004) Neural networks in mobile robot motion. Int J Adv Robot Syst 1(1):2 Jazdi N (2014, May) Cyber physical systems in the context of industry 4.0. In: 2014 IEEE international conference on automation, quality and testing, robotics. IEEE, pp 1–4 Jung K, Kim J, Kim J, Jung E, Kim S (2014) Positioning accuracy improvement of laser navigation using UKF and FIS. Robot Auton Syst 62(9):1241–1247 Kardasz P, Doskocz J, Hejduk M, Wiejkut P, Zarzycki H (2016) Drones and possibilities of their using. J Civil Environ Eng 6(3):1–7 Khatib O (1986) Real-time obstacle avoidance for manipulators and mobile robots. In: Autonomous robot vehicles. Springer, New York, pp 396–404 LaValle SM (1998) Rapidly-exploring random trees: a new tool for path planning. Computer Science Dept., Iowa State University, Tech. Rep. TR 98–11 Lee H, Jung J, Choi K, Park J, Myung H (2012) Fuzzy-logic-assisted interacting multiple model (FLAIMM) for mobile robot localization. Robot Auton Syst 60(12):1592–1606 Lee J, Bagheri B, Kao HA (2015) A cyber-physical systems architecture for industry 4.0-based manufacturing systems. Manuf Lett 3:18–23 Lee J, Davari H, Singh J, Pandhare V (2018) Industrial artificial intelligence for industry 4.0-based manufacturing systems. Manuf Lett 18:20–23 Lewis B (2018) The 10 biggest challenges facing robotics. https://australiascience.tv/the-10biggest-challenges-facing-robotics/ Li QL, Song Y, Hou ZG (2015) Neural network based fast SLAM for automobile robots in unknown environments. Neurocomputing 165:99e110 Liew CF, DeLatte D, Takeishi N, Yairi T (2017) Recent developments in aerial robotics: a survey and prototypes overview. arXiv preprint arXiv:1711.10085 Marshall JA, Bonchis A, Nebot E, Scheding S (2016) Robotics in mining. In: Springer handbook of robotics. Springer, Cham, pp 1549–1576 Mian S, Goswami M, Mostow J (2019, June) What’s most broken? Design and evaluation of a tool to guide improvement of an intelligent tutor. In: International conference on artificial intelligence in education. Springer, Cham, pp 283–295 Misra A., Singh G. (2019) Kinematic and dynamic analysis of an industrial six axis robotic manipulator, international conference on robotics, automation and non-destructive evaluation Misra A, Sharma A, Singh G, Kumar A, Rastogi V (2020) Design and development ofa low-cost CNC alternative SCARA robotic arm. Procedia Comput Sci 171:2459–2468 Mittal A, Misra A, Sharma A, Gupta A, Ansari N (2020) Design and analysis of modified radiator fins to improve overall cooling efficiency (No. 2020-01-2029). SAE Technical paper Montiel O, Orozco-Rosas U, Sepúlveda R (2015) Path planning for mobile robots using bacterial potential field for avoiding static and dynamic obstacles. Expert Syst Appl 42(12):5177–5191 Mourtzis D, Doukas M, Bernidaki D (2014) Simulation in manufacturing: review and challenges. Procedia Cirp 25:213–229 Mousavian A, Eppner C, Fox D (2019) 6-dof graspnet: Variational grasp generation for object manipulation. In: Proceedings of the IEEE international conference on computer vision, Seoul, pp 2901–2910

79 Robotics in Industry 4.0

2055

Oyekan J, Hu H (2010, June) A novel bacterial foraging algorithm for automated tuning of PID controllers of UAVs. In: The 2010 IEEE international conference on information and automation. IEEE, pp 693–698 Passino KM (2002) Biomimicry of bacterial foraging for distributed optimization and control. IEEE Control Syst Mag 22(3):52–67 Pierzchała M, Giguère P, Astrup R (2018) Mapping forests using an unmanned ground vehicle with 3D LiDAR and graph-SLAM. Comput Electron Agric 145:217–225 Pira S (2018) Robotics in Industry 4.0 – five major challenges for the packaging industry. https://www.automationworld.com/factory/robotics/article/13319394/robotics-in-industry40-five-major-challenges-for-the-packaging-industry Qin J, Liu Y, Grosvenor R (2016) A categorical framework of manufacturing for industry 4.0 and beyond. Procedia Cirp 52:173–178 Quaglia G, Visconte C, Scimmi LS, Melchiorre M, Cavallone P, Pastorelli S (2019) Design of the positioning mechanism of an unmanned ground vehicle for precision agriculture. In: IFToMM World Cong Mech Mach Sci. Springer, Cham, pp 3531–3540 Rojko A (2017) Industry 4.0 concept: background and overview. Int J Interactive Mobile Technologies (iJIM) 11(5):77–90 Sakagami Y, Watanabe R, Aoyama C, Matsunaga S, Higaki N, Fujimura K (2002) The intelligent ASIMO: system overview and integration. In: IEEE/RSJ international conference on intelligent robots and systems, Lausanne, vol 3. IEEE, pp 2478–2483 Shariatmadar K, Misra A, Debrouwere F, Versteyhe M (2019, October) Optimal modelling of process variations in industry 4.0 facility under advanced p-box uncertainty. In: 2019 IEEE Student Conference on Research and Development (SCOReD). IEEE, pp 180–185 Singh I, Manuja M, Mathur R, Goswami M (2020) Detecting intrusive transactions in databases using partially-ordered sequential rule mining and fractional-distance based anomaly detection. Int J Intell Eng Inform 8(2):138–171 Tenreiro Machado JA, Silva M (2006) An overview of legged robots. In: Proceedings of the MME 2006 international symposium on mathematical methods in engineering. Ankara, pp 1–40 Ustundag A, Cevikcan E (2017) Industry 4.0: managing the digital transformation. Springer, Cham Vysocky ALES, Novak PETR (2016) Human-robot collaboration in industry. MM Sci J 9(2): 903–906 Wawrla L, Maghazei O, Netland T. (2019) Applications of drones in warehouse operations. Whitepaper. ETH Zurich, D-MTEC Xu X (2012) From cloud computing to cloud manufacturing. Robot Comput Integr Manuf 28(1):75–86 Zavlangas PG, Tzafestas SG (2003) Motion control for mobile robot obstacle avoidance and navigation: a fuzzy logic-based approach. Syst Anal Model Simul 43(12):1625–1637

Fossil Fuel Combustion, Conversion to Near-Zero Waste Through Organic Rankine Cycle

80

A. B. Fakeye, S. O. Oyedepo, O. S. I. Fayomi, Joseph O. Dirisu, and N. E. Udoye

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zero Waste Strategic Plan and Sustainable Development Goals . . . . . . . . . . . . . . . . . . . . . . Indices for Zero Waste Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Waste Heat Recovery and Utilization in Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Drivers for Waste Heat Recovery in Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zero Waste Energy Index (ZWeI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Waste Heat Recovery Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2058 2061 2063 2064 2066 2068 2068 2071 2072

Abstract

Zero waste manufacturing (ZWM) conceptually transforms the economies of nations to a circular economy by employing sustainable technologies in reducing waste to barest minimum possible through the entire value chain. A number of indicators have therefore been proposed by many researchers to assess zero waste management right from producing raw materials to product manufacturing and finally waste disposal. Much attention has been given to waste disposal and recycling in ZWM. However, for better resource efficiency, zero waste index

A. B. Fakeye Mechanical Engineering Department, Federal Polytechnic, Ilaro, Nigeria S. O. Oyedepo () · J. O. Dirisu · N. E. Udoye Mechanical Engineering Department, Covenant University, Ota, Ogun State, Nigeria e-mail: [email protected] O. S. I. Fayomi Department of Mechanical and Biomedical Engineering, Bells University of Technology, Ota, Ogun State, Nigeria © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_69

2057

2058

A. B Fakeye et al.

(ZWI) was proposed to quantify energy, material, and water conservation through recycling efforts rather than simply measuring waste diverted from landfills. The most significant influence on the earth is energy generation and consumption. Hence, to limit the exploitation of the earth within its carrying capacity, the zero waste energy index (ZWeI) is hereby proposed to assess and promote energy efficiency in value chain through low-grade energy utilization and waste heat recovery (WHR). The ZWeI is a measure of the energy efficiency in product manufacturing processes and the potential of energy recovery from product waste. In this study, organic Rankine cycle (ORC) technology is being proposed to achieve ZWEI in energy-intensive industries. Keywords

Zero waste energy index · Organic rankine cycle · Waste heat recovery · Energy conservation · Low-grade energy

Introduction In modern terms, waste depicts inefficiency and overexploitation of resources in a civilized world. The severe degradation of the ecosystem the world is currently experiencing has been attributed to the world economic growth occurring since the mid-twentieth century when fossil fuels were discovered. The 500% progression in world economy was at the cost of 60% degradation of world ecosystem which establishes a positive correlation between economic growth and overexploitation of the earth (Zaman and Lehmann 2014). Recently, there are many initiatives such as circular economy and related concepts that are being promoted toward zero waste in order to uncouple economic well-being from resource depletion and environmental degradation. Nations and cities across the world have different levels of development and area at different stages of industrialization. Wang et al. (2019) investigated decoupling statuses between economic growth and their carbon footprints in 15 Chinese cities in relation to their levels of industrialization. Results showed diversities in decoupling statuses, driving factors, and decoupling efforts among cities, even in cities at the same level of industrialization. It was however established that energy intensity was the homogeneous factor with the highest influence on emission reduction, which increases proportionally to the presence of more heavy manufacturing industries. In another research, Egilegor et al. (2019) resolved to retrieve no less than 40% of the waste heat in the flue gases and reuse them within three selected factories instead of fouling the environment. They were able to achieve above 40% cost-effective waste heat recovery in all the three considered factories across Europe, providing savings on primary energy up to 597, 3020, and 4003 MWh/year and correspondingly cutting down carbon footprints by 135, 600, and 797 t/year and with payback periods of less than 3 years. They however identified the challenges to implementation of sustainable technologies to include high temperatures of waste sources and heat sinks, variations in waste heat

80 Fossil Fuel Combustion, Conversion to Near-Zero Waste Through. . .

2059

stream flow rate and temperature, and also the presence of corrosive contents and other particles in the waste streams harmful to the heat exchangers. The prime objective of the zero waste approach is to advance a design evaluation model that incorporates mass- and energy-efficient techniques into energy technologies to enhance environmental integrity (Khan and Islam 2017). Hence, a concept of designing for the environment (DfE) evolved to preclude or reduce a product’s undesirable effect on the environment by adapting products to a sustainable life cycle at the development phase of products. However, a key factor of DfE among similar concepts that have been identified to curtail the environmental footprint of a product was energy utilization, waste, and emissions during production (Pettersson et al. 2016). Energy efficiency is increasingly becoming a key target with universal adoption of the guidelines to achieve it (World Energy Council 2020). Egilegor et al. (2019) sustain that energy efficiency is a fundamental stake of a sustainable energy policy and economy decarbonization strategy that can improve industrial competitiveness, provide job opportunities, reduce energy bills, restore environmental integrity, and guarantee energy security for the future, among other benefits. Accordingly, for better resource efficiency in the manufacturing sector, zero waste index (ZWI) was proposed to quantify energy, material, and water conservation through recycling processes rather than simply quantifying waste diverted from landfills. According to BP Energy Outlook (2019), the industrial sector currently accounts for half of global energy consumption. According to Papadis and Tsatsaronis (2020), for the objective hazards of global warming to be eradicated, the carbon footprint associated with the energy sector and interdependent enduse energy sectors, most especially manufacturing industry, must be significantly reduced by the deployment of appropriate sustainable technologies and policies consistent with sustainable environment, energy security, economic stability, and social features. Papadis and Tsatsaronis (2020) however affirmed that the most crucial issue was ensuring availability of accessible and affordable technology options. Though the proportion of renewable energy sources in the energy mix is progressively increasing, the annual consumption of fossil fuels and its associated hazards is continuously on significant increase due to growing population and industrialization (Fakeye and Oyedepo 2019). As identified by BP Energy Outlook (2019), the prevailing challenge of our time is the urgency of meeting the increasing energy demand while concurrently reducing the carbon footprint. In the work of Papadis and Tsatsaronis (2020) in investigating emissions from secondary and enduse energy sector, the highest emissions were obtained from generation of heat and electricity, followed by the manufacturing industries in every nation and region. Globally, secondary energy carriers such as electricity, heat, petroleum products, and synthetic fuels are primarily derived by combustion of fossil fuels. Consequent on the global concerns to preserve environmental integrity and energy security, especially arising from the rising trend in energy demand and low deployment of renewable resources, coupled with fast rate of depletion of the limited resources, much attention is being focused on development and commercialization

2060

A. B Fakeye et al.

of various technologies which have been studied toward sustainable exploitation of fossil fuel sources and, in general, more efficient energy utilization. According to Papadis and Tsatsaronis (2020), most manufacturing processes are highly energyintensive producing huge carbon deposits. Heat energy loss in manufacturing industries is substantially high because of high temperatures and several heatintensive processes involved leading to rejection of huge amount of low-grade heat to the environment. Industries in this category are metallurgical (iron and steel making, aluminum, copper), chemical (refineries, plastic, fertilizers), nonmetallic minerals (cement and lime, ceramics, glass), pulp and paper, textiles and leather, food processing, and mining. In agreement with several other similar researches, they proposed process integration as the viable option to improve overall efficiency while reducing resource consumption and waste of manufacturing procedures. Several researches have investigated the effect of energy efficiency on product quality, cost, and environmental impacts. Lee and Okos (2011) appraised consumption of water and energy in processing of some products using plant-scale audit data collection to evaluate energy consumption, laboratory-scale experiments for variations in product quality with decrease in water and energy usage. Their results evidently showed possibility of reduction in water and energy consumption up to 95% and 80%, respectively, as well as reduction in cost, without compromising standards. Ziemele and Vigants (2018) applied WHR for district heating. They examined its economic benefits of district heating from a nearby factory. After accounting for heat losses during transmission, 4–12% savings of energy load was achieved during heating season, reducing CO2 emissions from the heating by 5–10% monthly. As well, the technoeconomic investigation of David, Michel, and Sanchez (2011) has verified that ORC technology is viable to decrease environmental impact from industrial and energy sectors through waste heat-to-power and system integration. For example, the energy consumption of industrial sector in France dropped from 36% in early 1970 to 23% in 2007 through the use of sustainable technologies. Ammar et al. (2012) however concluded from studies that all hindrances to utilization of lowgrade heats in process industry such as organizational, financial, and economic could be subdued and benefits from an all-inclusive perception could be gained through robust government policy and regulation incentives. However, in the last two decades, there have been growing interest and researches on waste heat recovery in industry through application of organic Rankine cycle (ORC) technology. The goal of such research is to improve energy efficiency, minimize the consumption of finite fossil fuel resources, and encourage reuse of waste heat from industrial processes so as to abate their environmental impacts (Tchanche et al. 2014; Imran et al. 2016). The trend of progress in research publication and the patents in the field of ORC technology in the past two decades (from 2000 to 2016) are shown in Fig. 1. The total number of publications amounts to 2120 in the journals and conference proceedings (Imran et al. 2018).

80 Fossil Fuel Combustion, Conversion to Near-Zero Waste Through. . . 600

2061

Papers Patents

Number of papers and patents

500

400

300

200

100

0 2000

2002

2004

2006

2008

2010

2012

2014

2016

Year Fig. 1 Trend of research publication and patents in ORC technology from 2000 to 2016 (Imran et al. 2018)

Zero Waste Strategic Plan and Sustainable Development Goals Zero waste (ZW) focuses on waste preventive measures based on sustainable design and utilization practices, maximum waste recovery, and not waste management by incineration (Zaman 2015). ZW concords with voidance and prevention of waste instead of waste treatment. According to Zero Waste International Alliance (ZWIA), “Zero Waste is a goal that is ethical, economical, efficient and visionary, to guide people in changing their lifestyles and practices to emulate sustainable natural cycles, where all discarded materials are designed to become resources for others to use” (ZWIA 2015). In real sense, practicing zero waste is capable to remove all emissions or discharges to land, water, or air that are a threat to the ecology. Zero waste principle can accelerate sustainable production and consumption, recovery, and recycling and restricts incineration. In other words, zero waste entails waste control through sustainable design of consumption patterns, effective managing, and optimal utilization of resources and products at the expiring of their life span. In view of this, zero waste proffers solution to global challenges of managing and

2062

A. B Fakeye et al.

processing natural resources without adverse effects on people and the quality of the environment. ZW brings about a cleaner and sustainable environment (Zaman 2017). The ZW Strategic Plan concords with Green Vision (GV) goals in leading the world in sustainability initiatives such as renewable energy and adequate managing and increasing recycled of waste usage. To achieve this strategy, the GV specified such goals as (i) create green jobs, (ii) reduce per capita energy use, (iii) generate clean renewable energy resources, (iv) build green buildings, (v) convert waste to energy, (vi) reuse of wastewater, (vii) plan with measurable standard for sustainable development, and (viii) transport sector to use alternative fuels. Meanwhile, the prime objective of the plan is the identification of the strategies to achieve zero waste (Branchini 2015). Majorly, zero waste serves as pathway for achieving and sustaining a greener society. Hence, to attain real sustainability, plan strategies should consider people, planet, and profit as principal factors for achieving social, environmental, and economic sustainability (Wali et al. 2019 ). Contrary to the conventional principle “cradle to grave” process, where a product has no use at the end of its life, rethink in cycles: “cradle to cradle” has come to bear. This implies that, at the expiry of the life span of a material or product, it should commence to be utilized new product – just like our natural ecosystems. Figure 2 presents the Future Cyclic Process Toward Resource Efficient and Zero Waste Society (Franco-García et al. 2019).

Fig. 2 Future cyclic process toward resource efficient and zero waste society (Franco-García et al. 2019)

80 Fossil Fuel Combustion, Conversion to Near-Zero Waste Through. . .

2063

In this present era, global energy supply from fossil fuels is being used in transportation sector, manufacturing, and power industry, and this situation is anticipated to continue for at least the next few decades. For effective and sustainable utilization of this limited fossil fuel resource, implementation of near-zero emissions technologies is paramount.

Indices for Zero Waste Measurement Solid waste is composed of partly renewable constituents and partly fossil-based constituents. The combustion waste is usually implemented to dispose of the portion of waste that cannot be recycled or reused, and the heat liberated is recovered to generate either electric power or domestic or industrial heating (Taherzadeh and Richards 2016). Several authors have suggested establishing a more appropriate model and index for the zero waste concept (Hogland et al. 2017). The zero waste index was introduced by Zaman and Lehmann (2014) to forecast the quantity of raw materials, energy, water, and greenhouse gas emissions that can be offset by recovered resources from waste stream. The zero waste index is basically a performance analysis of waste management systems in metropolitan cities which evaluates the potential energy, greenhouse gas (GHG), and water savings due to resource recovery from municipal solid waste. It hence reflects a measure of the resources that are exploited, consumed, wasted, recycled, recovered, and eventually reused in place of raw materials to reduce resource extraction. Among others, a fundamental objective of the zero waste idea is zero or near-zero depletion of natural resources. One of its fundamental principles is therefore designing processes of products manufacturing to methodically avoid and abate the amount and toxicity of waste and preserve, recover, and reuse all resources (Laura et al. 2019). Here, the zero waste energy index is proposed for energy conservation in product manufacturing to indicate the relation of possible maximum energy efficiency and retrievable energy content from its waste to the actual energy consumed in manufacturing the product. Unlike zero waste index, the zero waste energy index is proposed to be a standard of energy efficiency that measures and compares energy efficiency of a factory in producing a particular product with others producing the same products. Studies conducted by Sundana et al. (1987) concluded that existing indices in zero waste management were one sided, considering only environment factor without factoring other attributes such as economy, social, and financial, in addition to policy and regulation of waste management system. They found out limitations of existing indices were inapplicability to every region and country and hence recommended formulation of an index to suit local or regional conditions. However, environmental protection and minimization of global resource requirements are basic issues of environmental concern. Achieving 100% conversion of solid waste is inadequate and insufficient to represent the core values of zero waste concept. Since energy utilization affects the ecosystem beyond entire diversion rate of waste from landfill and it as well cuts across all the factors argued, it presents a better

2064

A. B Fakeye et al.

evaluation of zero waste strategies. Zero waste strategies according to Rathoure (2020) should not be limited to diversion rate of waste from landfill alone but should also technologically apply to companies, communities, industries, educational institutions, and homes. They hence proposed an all-inclusive assessment of sustainable avoidance and management of waste and resources through resource management alongside product management. The baseline for resource management and product management is energy, and since energy extraction and consumption is the most significant influence on the earth, ZWeI is inclined of being a suitable indication of zero waste concept, surpassing the limitations of the previous indices.

Waste Heat Recovery and Utilization in Industry In industry processes worldwide, large amounts of heat in the temperature range 125–250 ◦ C are rejected to the environment. Utilization of this surplus heat for direct reuse or upgrading is the most efficient and cost-effective way of energy conservation and mitigation of environmental pollution (Nikolaisen et al. 2019). Waste heat is accepted as a thermal energy discharged into the environment at low temperature. This waste heat emanates from process devices or equipment inefficiencies. Often, part of waste heat could potentially be used for some useful purpose. At present, about 20–50% of energy used in industry is rejected as waste heat. It has been recognized that waste heat is experience largely in mechanical and thermal processes. Industry waste heat originates from combustion products being discharged to the atmosphere, heated water released into environment, heated products exiting industrial processes, heat transfer from hot equipment surfaces, etc. (Arzbaecher and Parmenter 2007). Based on the nature of source, waste heat differs regarding the aggregate state, temperature range, and frequency of their occurrence. Table 1 shows exhaust temperature of some on-site industrial application of WHR. Industrial heating and energy production processes accounted for the highest quantity of waste heat being lost into the environment (Oyedepo 2019). The exact amount of industrial waste heat is difficult to quantify; however, researches revealed that up to 20–50% of industrial energy used is ultimately discharged as waste heat and that between 18% and 30% of this waste heat could be reused. Optimal utilization of industrial waste heat integrated with thermally efficient machineries brings about improved industrial energy efficiency (Woolley et al. 2018). Application of the technique of waste heat recovery in industrial processes is a pathway to enhance the effective utilization of limited available fossil fuel and reduce the running costs, the pollutants, and greenhouse emissions to the environment (Daniele et al. 2018). Industrial processes such as drying, heating, and combustion, there are significant waste heat releases in the forms of vapor, fume, exhaust, wastewater, and heat, being released from industrial devices such as furnaces, kiln, heaters, refrigeration systems, boilers, etc., that are not reused. The temperature of waste heat varies with industrial processes, and the range is very broad, from as low as 30 ◦ C to more than 1000 ◦ C. Waste heat recovery has

80 Fossil Fuel Combustion, Conversion to Near-Zero Waste Through. . . Table 1 Exhaust temperature for some on-site industrial applications

2065 Temperature (◦ C)

Manufacture of some basic metals Iron and steel making Steel electric arc furnace Nickel refining furnace Aluminum reverberatory furnace Cooper refining furnace Manufacture of nonmetallic minerals Cement sintering Glass melting furnace Calcining of limestone in the kiln Calcination of magnesia Cement kiln Clinker cooler waste air Kiln system exhaust/preheater Manufacture of nonmetallic minerals Furnace black process Ammonia catalyst reaction Paint and varnish depolymerization Plastic and rubber

1450–1550 1370–1650 1370–1650 1100–1200 760–820 1450 1300–1540 900 600–800 450–620 177–232 382–816 1200–1900 510 288–343 90–200

Source: Miró et al. (2016)

been recognized as a technique of trapping low-grade heat from waste streams of existing industrial process and using this heat directly, upgrading it to a high-grade temperature, and/or converting it to electric power or cooling (Oyedepo and Fakeye 2020). The thermal energy produced from waste heat recovery can be utilized either in industrial site or transported to neighboring facilities or converted to electrical or heat distribution networks. Waste heat recovery techniques provide opportunity for energy savings and substantial greenhouse gas emission reductions (Ling-Chin et al. 2018). The suitability of a waste heat source for useful recovery depends on its characteristics. Waste heat streams are characterized by type (i.e., flue gas, water, oil), temperature, flow rate, profile, and composition. A common approach is to classify waste heat according to temperature ranges. There is currently no unambiguous agreement on waste stream classification, but oftentimes temperature ranges are used. Tchanche et al. (2011) gave a range for low-temperature (up to 230 ◦ C), medium-temperature (230–650 ◦ C), and high-temperature (above 650 ◦ C) waste heat. The thermal and economic value of low-temperature heat is lower than for high-temperature streams, yet its potential is unambiguous due to its large availability (US DoE 2008). Studies have shown that industrial low-grade heat has the prospect to be make-use of in industry in producing (i) another single product such as mechanical and electric power or heating; (ii) cogeneration (two products simultaneously) such as heating and power; and (iii) tri-generation (three products

2066

A. B Fakeye et al.

simultaneously) such as heating, cooling, and electricity simultaneously (Lingfeng et al. 2018), with the deployment of advanced technologies. Utilizing waste heat as input for another process generally enhances energy efficiency because of the better application of the primary energy put into the main process. Also, it offsets (part of) the energy requirements in the secondary process (Oyewunmi et al. 2017). For example, utilizing waste heat for heating purposes is an attractive option because it avoids conversion from high-quality fuel inputs to lower quality heat. In case no useful heat demand exists, the waste heat may generate useful output in terms of high-quality electricity. In the case, there is a heating demand, but at a very low temperature, such low-temperature power generation systems may even serve as cogeneration unit (Peris et al. 2015).

Drivers for Waste Heat Recovery in Industry Significant quantity of waste heat is dissipated from industrial plants and processes. The source of this unutilized heat has economic and environmental impacts, if this waste heat is not recovered nor stored nor transported to other processes or users. In production industry, process heating accounts for largest proportion of the total energy used in manufacturing applications. The proportion of energy used for process heat varies from industries to industries. As energy costs continue to rise, industrial plants need effective ways to reduce the energy used for process heating. Figure 3 shows examples of industry process heating systems such as combustion systems (fossil and biomass), electric systems, and heat recovery and exchange systems. In manufacturing or process industry, over 80% of energy source for process heating is from the combustion of fossil fuels. Hence, the potential methods to reduce the environmental impacts of combustion-related emissions from industry

Fig. 3 Key components of a process heating system (US DoE 2007)

80 Fossil Fuel Combustion, Conversion to Near-Zero Waste Through. . .

2067

include advanced processes, improved designs, thermal efficiency, etc. (US DoE 2007). Various factors such as global climate change and its various effects on human life are among the major drivers of technologies for waste heat recovery in industry for achieving sustainable society. Waste heat from industry is among factors responsible for global greenhouse gas (GHG) emissions. Hence, the concept of “zero waste heat” is capable to address environmental issues such as GHG emissions and the provision of potential specific solutions for emission reduction and sustainable waste management in industry (Zaman and Lehmann 2011). A significant proportion of consumed energy in industry is wasted as heat in the form of exhaust gases, air streams, and liquids leaving industrial facilities. Technically and economically, it may not be feasible to recover all waste heat from industry. However, a gross estimate is that waste heat recovery could substitute for 9% of total energy used by industry. Moreover, an increased use of waste heat recovery technologies by industry has potential to ameliorate greenhouse gas (GHG) emissions. Researches have shown that the major sources of waste heat in industrial facilities include combustible gases from fossil fuel used in fired furnaces/kilns, boilers, and process heating equipment. These types of high-grade waste heat sources can readily be used to preheat compressed air in gas turbine and boiler feedwater, generate heat in generator used in absorption refrigeration system, process loads, etc. According to Reis and Gallo (2018), an indicative target of 27% energy savings by 2030 for European energy consumption including transport, electric power generation, and industry has been endorsed by the European Council, on its energy efficiency directive (ECEED). Success in this aspect will depend largely on industrial energy efficiency, particularly on energy efficiency of energy-intensive industries. Studies have shown that the energy sector, i.e., power generation and energyconsuming sectors, represents by far the principal source of greenhouse gas (GHG) emissions. This really accounts for two-thirds of global GHG and 80% of CO2 emissions (Riccardo et al. 2019). Hence, reliable and effective strategies are essential in the energy sector to alleviate the climate change problem (Shu et al. 2012). Moreover, depleting fossil fuel resources and the horrible environmental impacts due to burning fossil fuels emphasize the importance of waste recovery technologies in energy-intensive industries (Afsaneh et al. 2019). It is important to note that it is practically possible to reuse captured industrial waste heat within the same process (such as in power generation), or it can be transferred to an external process. The captured waste heat can be used to preheat air or water used in gas turbine air preheater and heat recovery steam generator (HRSG), respectively. Such process reduces time and quantity of fuel needed to boil water in HRSG and also eliminate discharge of exhaust gases to the environment. Studies have shown that less than 75% of the energy supplied to industrial process heating equipment is actually used for heating; the rest is often lost to exhaust streams such as flue gases. In view of this, modern heat recovery technologies allow at least part of energy in exhaust streams be reused. Moreover, operating fuel combustion devices at optimal efficiency reduces heat losses to the

2068

A. B Fakeye et al.

environment. Also, by seeking for possible ways to recover and reuse waste heat could be the most beneficial improvements of energy utilization in industry. The waste heat recovery in industry is a paradigm to enhance the efficiency and reduce the costs, the pollutants, and greenhouse emissions. Energy conservation refers to as process for forestalling the wasteful use of energy. It is an essential tool for industrial energy users to cut down their energy use in order to curtail its negative effects on the environment (Huang et al. 2017). Energy conservation from perspective of combustion optimization principle has potential of reducing both energy consumption and environmental pollution. Studies have shown that a 2% reduction in excess oxygen in a combustion process can reduce fuel consumption by 4% and NOx emissions by 40% (Lion et al. 2017; Kheiri et al. 2017). Moreover, combustion optimization technologies allow power-generation facilities to balance their boilers, center the fireball, minimize NOx, and maximize MW per ton of fossil fuel.

Zero Waste Energy Index (ZWeI) The zero waste energy index (ZWeI) is a novel proposition for energy efficiency in industries for manufacturing of products and the potential of energy recovery from product waste through the use of sustainable technologies, particularly the ORC. This entails the manufacturing of products at minimal energy and fuel input so as to reduce pressure on resources. Typically, energy efficiency systems essentially involve both private costs and social gains. The capital investments on such systems are costly, but they bring about significant reduction in a company’s effective cost of energy services per unit of real output. Optimal investment in energy efficiency programs in industry has severally been identified by researchers and practitioners as having the potentials to achieve environmental protection and energy security, hence the call for collaboration between manufacturing industries, government, and policy makers to ease the difficulties to achieving the benefits an average of 3.1% improvement in TFP of the companies enlisted on the program over (Filippini et al. 2020). Filippini et al. (2020) evaluated the effect of Chinese energy efficiency program on short-run total factor productivity growth (TFPG) of metallurgical industry. They established an average growth of 6.4% in total productivity factor (TFP) of the industry with the first 3 years of program implementation. They hence suggested the program for similar industries in developing countries through sufficient awareness in order to improve their energy decisions and productivity.

Waste Heat Recovery Technologies The impacts of low-grade heat are undermined by industry, and this has raised a serious concern on the environment as a result of thermal enrichment. The unutilized waste heat is discharged to the airspace at all stages of an industrial process, through the use of ineffective energy conversion systems. In view of this effect, studies have

80 Fossil Fuel Combustion, Conversion to Near-Zero Waste Through. . .

2069

been carried out on effective technologies not only to reduce the burden on limited conventional sources of energy but also to take steps toward achieving a sustainable environment. Moreover, considering the increasing paucity of fossil fuels, attaining optimal effectiveness of energy conversion processes is one of the key challenges for optimizing primary energy use. On this basis, low-grade waste heat from various industrial processes is becoming interesting as a secondary energy source (Anup 2016). Presently, fossil fuels provide the majority of the world’s energy, and this situation is expected to continue for at least the next few decades. For effective and efficient utilization of fossil fuels in an environmentally sustainable way, near-zero waste heat emission technologies are essential. Waste heat by industry can be recaptured for reuse either directly or, more commonly, indirectly. Direct heat recovery is more contemptible recourse, but its use is limited by locality and foulness considerations. In indirect heat recovery, two fluid streams are separated by a heat transfer surface. Systems that transform low-grade heat to electricity and can be integrated subsist power generation plants to improve their thermal performance and reduce emissions of enthrallment. Such systems that transform low-grade heat to electric power are of great benefits on two bases: firstly, by the enhancement of the performance of contemporary power generation system and, secondly, in adoption to sustainable energy sources that are, to date, dormant. Traditional technologies exist and are under maturing that could enable more cost-effective recovery of waste heat from the industrial processes. There are two clear technology classes for this technology, and these are process heat exchange and thermal energy conversion devices. Based on the complexity of industrial process, hence, waste heat from industry varies in the temperature grades at hand and distribution pattern of the thermal discharges. This complexity in nature demands different techniques be adopted to recuperate waste heat depending on the quality of heat available and the nature of the ancillary processes and the ability to redeploy heat within the processes (Viswanathan et al. 2006). Three main techniques recognized to abating the CO emissions from fossil fuel energy conversion processes are: • Improve the efficiency of energy transformation and advance technology in power plant and power cycles • Use of alternative fuel, through the use of low-carbon fossil fuels such as natural gas or by resorting to renewable or nuclear technology • Carbon management through the development and deployment of clean coal and CO capture and storage technologies. There are four thermodynamic cycles popularly used for waste heat recovery. These include trilateral flash cycle (TFC), transcritical CO2 cycle (T-CO2 ), Kalina cycle (KC), and organic Rankine cycle (ORC). In comparison, the Kalina cycle systems are complex, bulky, corrosion-prone, and hence expensive; transcritical CO2 power cycles are disadvantaged by the inappropriateness of multiphase expansion and the complexity of cooling below ambient temperature, while trilateral

2070

A. B Fakeye et al.

flash cycle’s problematic two-phase expansion limits its practical applications (Shu et al. 2016). Organic Rankine cycles (ORCs) are, however, characteristically simplestructured, highly reliable, and easy to maintain, modify, and effectively adapt to various heat sources and may be operated in subcritical or transcritical cycles (Feng et al. 2015).

Organic Rankine Cycle for Waste Heat Recovery There are enormous potentials for utilizing waste heat sources from engine exhaust gases or industrial processes in industry. Studies have revealed that low-grade waste heat accounts for 50% or more of the total heat produced in industry. Among the suggested techniques for waste heat recovery is the organic Rankine cycle (ORC). ORC system is the most widely used. The main benefits of ORC are the simplicity and the availability of its components. In this system, the working fluid is an organic substance, better adapted than water to lower heat source temperatures (Sylvain et al. 2011). The heat recapture through ORC system can be an appropriate solution for several industrial sectors, especially in fields where the industrial process is characterized by a gigantic need of energy. The industrial processes in which the ORC serves a high potential energy efficiency solution are the energy concerted industries such as iron and steel, glass, and cement (Daniele et al. 2012). ORC is a mature, suitable, and commercially available technology in realizing greenhouse gas emission reductions and conservation of the excessively depleting fossil fuel reserve by improving the efficiency of energy use by utilizing low-power and low-temperature heat source. Waste heat sources obtainable for recapture from various industrial processes are capable of effectively powering ORC systems that can deliver output energy scales ranging from about 10 kW to 10 MW (Önder et al. 2020). However, Reshid et al. (2017) stated that ORC technology is uneconomical for any application when the heat source temperature is lower than 100 ◦ C. ORCs are the most widespread low-grade waste heat recapture systems predominantly because of their lucidity and readily accessible components (Mohammed et al. 2020). The advantages of ORC over a steam Rankine cycle become especially evident for low-grade heat sources when appropriate working fluids and operating conditions are selected (Quoilin 2011). According to Liu et al. (2015), ORC applications are advantageous over the steam Rankine cycle at the lower temperatures because the thermal efficiency of ORCs becomes economically feasible by using low-boiling organic fluids to recover waste heat at temperatures below 300 ◦ C, especially when used as bottoming cycles for low-temperature waste heat recovery in process industries, enhance the efficiency improvement in a power station generating less than 20 MW, and recuperate heat from geothermal sources. Another preferential advantage of ORC systems is utilization of organic working fluids that typically require single stage expanders which are simpler and more cost-effective as regards capital costs and maintenance (Liu et al. 2004).

80 Fossil Fuel Combustion, Conversion to Near-Zero Waste Through. . .

2071

ORCs for Near-Zero Waste Thermal Conversion of Fossil Fuels The organic Rankine cycles have many practical and commercial applications for recovery and reuse of degraded heats across every industry such as power generation, metallurgical, nonmetallic minerals, chemical, food processing, and others. In addition, the ORC has proven applications for power conversion in renewable energy sources and thermal conversion of solid waste to electricity, hence improving the proportion of geothermal, solar thermal biomass, etc., in the energy mix, which in essence inherently reduces dependence on fossil fuels. Researches have established that the ORC has potential to achieve sustainable consumption of fossil fuels via cutting down on greenhouse gas emission to a tolerable limit and elimination of excessive depletion of fossil fuels. Their simplicity, readily available components, and wider flexibility of modifications to cycle architectures in order to adapt them to the peculiarities and variability of thermal regimes of operations have earned ORC’s wider acceptance and practicability over other similar technologies. A lot of work has been done on the use of ORC technology in achieving sustainable consumption of fossil fuels and promoting energy efficiency in energy and manufacturing industries and thermochemical conversion of solid waste, as well as in achieving economy decarbonization through energy transition to renewable sources such as geothermal and solar-thermal. Moreira and Arrieta (2019) examined the benefits of using subcritical basic and regenerative ORCs with superheater for WHR in the operations of Brazilian cement manufacturing plants with clinker productive capacities between 3000 and 6300 ton/day. Results showed that the proposed ORCs were able to generate roughly 80 MW for the cement industry of the state of Minas Gerais, thus cutting down carbon footprint by 221,069 kg CO2 /year, and as typical of ORC systems, the payback period was less than 2 years with internal rate of return above 80%/year. Oyedepo and Fakeye (2020) investigated a 5.75 MW gas turbine power plant (GTPP) and revealed from the preliminary investigation that incorporating a bottoming and ORC for power conversion of the exhaust gases from GTPP can generate in excess of 16.8% of the nominal power of a GTPP at no additional fuel consumption or emission to the atmosphere which effectively reduced carbon footprint per kW of energy generated, compatible for sustainable environment as well as boosting energy security.

Conclusion How much longer can the earth effectively meet the ever-increasing energy demands resulting from population accretion and economic development? The question may not have a ready answer, but efficient use of resources will certainly prolong natural resources and preserve the environment. The term effective must therefore be tangible like efficiency with a measurement tool to assess the environmental, economic, and social impacts. The zero waste energy index (ZWeI) is proposed as a standard for each product to account for minimum energy required for product

2072

A. B Fakeye et al.

manufacturing and the amount of recoverable energy from the product waste. ORC technology is proposed for energy efficiency for system integration and waste heat-to-power conversion for near-zero waste thermal conversion of fossil fuels to promote environmental integrity and prolong the life span of the limited resource to secure future generations.

References Afsaneh N, Abbas N, Mokhtar B (2019) Exergoeconomic comparison and optimization of organic Rankine cycle, trilateral Rankine cycle and transcritical carbon dioxide cycle for heat recovery of low-temperature geothermal water. Proceedings of the Institution of Mechanical Engineers Part A Journal of Power and Energy 233(8):1068–1084 Ammar Y, Joyce S, Norman R, Wang Y, Roskilly AP (2012) Low grade thermal energy sources and uses from the process industry in the UK. Appl Energy 89(1):3–20. https://doi.org/10.1016/ j.apenergy.2011.06.003 Anup S (2016) Energy and exergy analysis of waste heat recovery systems using organic Rankine cycle. MSc thesis, Bangladesh University of Engineering and Technology (BUET), Dhaka, p 98 Arzbaecher C, Parmenter K (2007) Industrial waste-heat recovery: benefits and recent advancements in technology and applications. In: ACEEE summer study on energy efficiency in industry, pp 1–13 BP Energy Outlook (2019), The Energy Outlook explores the forces shaping the global energy transition out to 2040 and the key uncertainties surrounding that transition. BP Energy Economic, Brunel International N.V., The Netherlands, p 73 Branchini L (2015) Waste-to-energy: advanced cycles and new design concepts for efficient power plants. Springer, Cham, p 150 Daniele F, Dario DS, Francesco C (2018) Innovative system for electricity generation from waste heat recovery. In: ECEEE industrial summer study proceedings, pp 393–403 Daniele F, Nicola R, Veronica V, Marco B, Dario D (2012) Heat recovery for electricity generation in industry, Conference: ECEEE industrial summer study - Retool for a competitive and sustainable industry at: Arnhem, 523–534 David GF, Michel L, Sanchez L (2011) Waste heat recovery projects using organic Rankine cycle technology – examples of biogas engines and steel mills applications. World Engineers’ Convention, Geneva Egilegor B, Jouhara H, Zuazua J, Al-Mansour F, Plesnik K, Montorsi LM (2019) ETEKINA: analysis of the potential for waste heat recovery in three sectors: aluminium low pressure die casting, steel sector and ceramic tiles manufacturing sector. Int J Thermofluids 2019:100002. https://doi.org/10.1016/j.ijft.2019.100002 Fakeye AB, Oyedepo SO (2019) Designing optimized organic Rankine cycles systems for waste heat-to-power conversion of gas turbine flue gases. J Phys Conf Ser 1378(2019):032097. https://doi.org/10.1088/1742-6596/1378/3/032097 Feng Y, Hung TC, Zhang Y, Li B, Yang J, Shi Y (2015) Performance comparison of lowgrade ORCs (organic Rankine cycles) using R245fa, pentane and their mixtures based on the thermoeconomic multi-objective optimization and decision makings. Energy 93(2015), 2018– 2029. https://doi.org/10.1016/j.energy.2015.10.065 Filippini M, Geissmann T, Karplus VJ, Zhang D (2020) The productivity impacts of energy efficiency programs in developing countries: evidence from iron and steel firms in China. China Econ Rev 59:101364. https://doi.org/10.1016/j.chieco.2019.101364 Franco-García MG, Carpio-Aguilar JC, Bressers H (2019) Towards zero waste, circular economy boost: waste to resources. Springer Nature Switzerland AG, p 283

80 Fossil Fuel Combustion, Conversion to Near-Zero Waste Through. . .

2073

Hogland W, Kaczala F, Jani Y (2017) Beyond the zero waste concept. Linn Eco-Tech. https:// doi.org/10.15626/Eco-Tech.2014.028 Huang F, Zheng J, Baleynaud JM, Lu J (2017) Heat recovery potentials and technologies in industrial zones. J Energy Inst 90(6):951–961. https://doi.org/10.1016/j.joei.2016.07.012 Imran M, Usman M, Park B-S, Yang Y (2016) Comparative assessment of organic Rankine cycle integration for low temperature geothermal heat source applications. Energy 102:473–490. https://doi.org/10.1016/j.energy.2016.02.119 Imran M, Haglind F, Asim M, Alvi JZ (2018) Recent research trends in organic Rankine cycle technology: a bibliometric approach. Renew Sust Energ Rev 81:552–562 Khan MM, Islam MR (2017) Zero waste engineering, 2nd edn. Wiley, Hoboken. Retrieved from [email protected] Kheiri R, Ghaebi H, Ebadollahi M, Rostamzadeh H (2017) Thermodynamic modeling and performance analysis of four new integrated organic Rankine cycles (a comparative study). Appl Therm Eng 122:103–117. https://doi.org/10.1016/j.applthermaleng.2017.04.150 Laura M, Garcia F, Carlos J, Aguilar C, Bressers H (Ed) (2019) Towards zero waste, circular economy boost: waste to resources. Springer Nature Switzerland, p 283 Lee W, Okos MR (2011) Sustainable food processing systems – path to a zero discharge: reduction of water, waste and energy. Ital Oral Surg 1:1768–1777. https://doi.org/10.1016/ j.profoo.2011.09.260 Ling-Chin J, Bao H, Ma Z, Taylor W, Roskilly AP (2018) State-of-the-art technologies on lowgrade heat recovery and utilization in industry. IntechOpen 55–74. http://doi.org/10.5772/ intechopen.78701 Lingfeng S, Gequn S, Hua T, Shuai D (2018) A review of modified organic Rankine cycles (ORCs) for internal combustion engine waste heat recovery (ICE-WHR). Renew Sust Energ Rev 92: 95–110 Lion S, Michos CN, Vlaskos I, Taccani R (2017) A thermodynamic feasibility study of an organic Rankine cycle (ORC) for heavy duty diesel engine (HDDE) waste heat recovery in off-highway applications. Int J Energy Environ Eng 8:81–98. https://doi.org/10.1007/s40095-017-0234-8 Liu BT, Chien KH, Wang CC (2004) Effect of working fluids on organic Rankine cycle for waste heat recovery. Energy 29(8):1207–1217. https://doi.org/10.1016/j.energy.2004.01.004 Liu C, Gao T, Xu J, Zhu J, Xu X (2015) Analysis of pure fluid and zeotropic mixtures used in lowtemperature reheating organic Rankine cycles. In: Proceedings of the 3rd international seminar on ORC power systems, pp 1–10 Miró L, Gasia J, Cabeza LF (2016) Thermal energy storage (TES) for industrial waste heat (IWH) recovery: a review. Appl Energy 179:284–301. https://doi.org/10.1016/j.apenergy.2016.06.147 Mohammed RH, Qasem NAA, Zubair SM (2020) Enhancing the thermal and economic performance of supercritical CO2 plant by waste heat recovery using an ejector refrigeration cycle. Energy Convers Manag 224:113340 Moreira LF, Arrieta FRP (2019) Thermal and economic assessment of organic Rankine cycles for waste heat recovery in cement plants. Renew Sust Energ Rev 114:109315. https://doi.org/ 10.1016/j.rser.2019.109315 Nikolaisen M, Skjervold V, Andresen T (2019) Evaluation of heat recovery heat exchanger design parameters for heat-to-power conversion from metallurgical off-gas. In: IIR international Rankine 2020 conference – heating, cooling and power generation, Glasgow, 26–29 July 2020, pp 1–8 Önder K, Onur B, Nehir T, Muhammed MA (2020) First and second law evaluation of combined Brayton–organic Rankine power cycle. J Therm Eng 6(4):577–591 Oyedepo SO (2019) Energy use and energy saving potentials in food processing and packaging: case study of Nigerian industries. In: Grumezescu AM, Holban AM (eds) Bottled and packaged water, volume 4: the science of beverages. Woodhead Publishing, Oxford, UK, pp 423–452 Oyedepo SO, Fakeye AB (2020) Electric power conversion of exhaust waste heat recovery from gas turbine power plant using organic Rankine cycle. Int J Energy Water Resour. https://doi.org/ 10.1007/s42108-019-00055-3

2074

A. B Fakeye et al.

Oyewunmi OA, Ferré-Serres S, Lecompte S, Van Den Broek M, De Paepe M, Markides CN (2017) An assessment of subcritical and trans-critical organic Rankine cycles for waste-heat recovery. Energy Procedia 105:1870–1876. https://doi.org/10.1016/j.egypro.2017.03.548 Papadis E, Tsatsaronis G (2020) Challenges in the decarbonization of the energy sector. Energy 205:118025. https://doi.org/10.1016/j.energy.2020.118025 Peris B, Navarro-Esbrí J, Molés F, Mota-Babiloni A (2015) Experimental study of an ORC (organic Rankine cycle) for low grade waste heat recovery in a ceramic industry. Energy 85:534–542. https://doi.org/10.1016/j.energy.2015.03.065 Pettersson A, Niklasson F, Richards T (2016) Combustion of wastes in combined heat and power plants. In: Taherzadeh MJ, Richards T (eds) Resource recovery to approach zero municipal waste. CRC Press/Taylor & Francis Group, Boca Raton/London/New York, pp 142–163 Quoilin S (2011) Sustainable energy conversion through the use of organic Rankine cycles for waste heat recovery and solar applications, October, pp 1–183. Retrieved from http:// orbi.ulg.ac.be/handle/2268/96436 Rathoure AK (Ed) (2020) Zero Waste management practices for environmental sustainability. CRC Press Taylor & Francis Group, London, New York, p 350 Reis MML, Gallo WLR (2018) Study of waste heat recovery potential and optimization of the power production by an organic Rankine cycle in an FPSO unit. Energy Convers Manag 157:409–422. https://doi.org/10.1016/j.enconman.2017.12.015 Reshid MN, Muhamad WMW, Majid MAA (2017) Transient simulation of a waste heat recovery from gas turbine exhaust. J Appl Sci 17(1):22–31. https://doi.org/10.3923/jas.2017.22.31 Riccardo B, Maurizio C, Sonia L, Parisi LM (2019) Life cycle assessment of energy systems and sustainable energy technologies: the Italian experience. Springer, Cham, pp 3–15 Shu G, Zhao J, Tian H, Liang X, Wei H (2012) Parametric and exergetic analysis of waste heat recovery system based on thermoelectric generator and organic Rankine cycle utilizing R123. Energy 45(1):806–816. https://doi.org/10.1016/j.energy.2012.07.010 Shu G, Shi L, Tian H, Li X, Huang G, Chang L (2016) An improved CO2 -based transcritical Rankine cycle (CTRC) used for engine waste heat recovery. Appl Energy 176:171–182. https://doi.org/10.1016/j.apenergy.2016.05.053 Sundana EJ, Sutadian AD, Juwana I, Lingkungan T (1987) Zero Waste Management Index – Sebuah Tinjauan – a review. Creat Res J 6:55–62 Sylvain Q, Sébastien D, Bertrand FT, Vincent L (2011) Thermo-economic optimization of waste heat recovery organic Rankine cycles. Appl Therm Eng 31(14–15):2885 Taherzadeh MJ, Richards T (Ed) (2016) Resource recovery to approach zero municipal waste. CRC Press is an imprint of the Taylor and Francis Group, London, New York, p 360 Tchanche BF, Lambrinos G, Frangoudakis A, Papadakis G (2011) Low-grade heat conversion into power using organic Rankine cycles – a review of various applications. Renew Sust Energ Rev 15(8):3963–3979 Tchanche BF, Petrissans M, Papadakis G (2014) Heat resources and organic Rankine cycle machines. Renew Sust Energ Rev 39:1185–1199. https://doi.org/10.1016/j.rser.2014.07.139 U.S. DoE (2007) Improving process heating system performance: a sourcebook for industry. Lawrence Berkeley National Laboratory, Berkeley, p 114 U.S. DoE (2008) Waste heat recovery: technology and opportunities in the US industry. U.S. Department of Energy, Washington, DC, p 112 Viswanathan VV, Davies RW, Holbery J (2006) Opportunity analysis for recovering energy from industrial waste heat and emissions. U.S. Department of Energy, Washington, DC, p 135 Wali E, Wizor CH, Nwankwoala HO (2019) Waste-to-wealth, towards a sustainable zero-waste in a circular economy: an overview. Int J Emerg Eng Res Technol 7(11):1–11 Wang R, Zheng X, Wang H, Shan Y (2019) Emission drivers of cities at different industrialization phases in China. J Environ Manag 250:109494. https://doi.org/10.1016/j.jenvman.2019.109494 Woolley E, Luo Y, Simeon A (2018) Industrial waste heat recovery: a systematic approach. Sustain Energy Technol Assess 29:50–59 World Energy Council (2020) World energy issues monitor 2020: decoding new signals of change. World Energy Council United Kingdom, London, p 173

80 Fossil Fuel Combustion, Conversion to Near-Zero Waste Through. . .

2075

Zaman AU (2015) A comprehensive review of the development of zero waste management: lessons learned and guidelines. J Clean Prod 91:12–25 Zaman AU (2017) A strategic framework for working towards zero waste societies based on perceptions surveys. Recycling 2(1):1–15. https://doi.org/10.3390/recycling2010001. www. mdpi.com/journal/recycling Zaman AU, Lehmann S (2011) Challenges and opportunities in transforming a city into a “zero waste city”. Challenges 2:73–93. https://doi.org/10.3390/challe2040073 Zaman AU, Lehmann S (2014) The zero waste index: a performance measurement tool for waste management systems in a “zero waste city”. J Clean Prod 50:123–132. https://doi.org/10.1016/ j.jclepro.2012.11.041 Ziemele J, Vigants E (2018) Assessing the feasibility of using the heat demand-outdoor temperature function for a long-term heat demand forecast. Energy Procedia 147:315–321. https:// doi.org/10.1016/j.egypro.2018.07.098 ZWIA (2015) Zero waste international alliance. http://zwia.org/aboutus/. Accessed 1 Sept 2017

Flow Shop Scheduling Problems in Industry 4.0 Production Environments: Missing Operation Case

81

Daniel Alejandro Rossit, Adrián Toncovich, Diego Gabriel Rossit, and Sergio Nesmachnow

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flow Shop Scheduling Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Non-permutation Flow Shop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Non-permutation Flow Shop with Missing Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . Non-permutation Flow Shop with Total Tardiness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Industry 4.0 Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scheduling in Manufacturing Cells: Mathematical Formulations . . . . . . . . . . . . . . . . . . . Numerical Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Optimization Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2078 2080 2081 2083 2084 2085 2085 2088 2090 2090 2092 2094 2095 2096

D. A. Rossit () Department of Engineering, Universidad Nacional del Sur (UNS), Bahía Blanca, Argentina e-mail: [email protected] A. Toncovich Departamento de Ingeniería, Universidad Nacional del Sur, Buenos Aires, Argentina e-mail: [email protected] D. G. Rossit Departamento de Ingeniería, Universidad Nacional del Sur, Buenos Aires, Argentina INMABB UNS CONICET, Departamento de Matemática, Buenos Aires, Argentina e-mail: [email protected] S. Nesmachnow Facultad de Ingeniería, Universidad de la República, Montevideo, Uruguay e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_71

2077

2078

D. A. Rossit et al.

Abstract

The Fourth Industrial Revolution or Industry 4.0 is forcing a complete reorganization of the manufacturing systems in order to implement increasingly automatized processes and customized products. Within this context, advanced computer-aid tools can contribute to give support to decision-makers in increasingly complex conditions. As a contribution to this process, this chapter addresses an optimization problem that has become progressively common within the Industry 4.0 context: the missing operations flow shop scheduling problem. Conversely, to the traditional flow shop, this problem considers the customization of the final products based on the requirements of the clients. Thus, several operations of the manufacturing cell can be skipped. Moreover, the missing operations can vary from one client to another, increasing the difficulty of the decision-making process. In this chapter we revise the missing operations flow shop scheduling problem under two of the main paradigms of the scheduling literature: considering only permutative schedules, that is, the same job sequence is used for all the machines involved, and the more computationally expensive case of allowing the optimization problem to consider non-permutative schedules, that is, different job schedules can be used for different machines in the production line. For these two cases, mathematical formulations that aim at minimizing total tardiness are presented. Furthermore, a two-echelon resolution approach is discussed. This involves firstly a genetic algorithm (GA), which only considers permutative schedules, and secondly, a simulated annealing algorithm, which taking as an input the solution of the GA expands the search space by considering non-permutative schedules. Computer experimentation was performed on a set of instances with different proportions of missing operations in order to represent a large variety of the situations that occur in practice at real-world manufacturing cells. Keywords

Industry 4.0 · Missing operation · Flow shop scheduling problem · Genetic algorithm · Simulated annealing

Introduction The Fourth Industrial Revolution or Industry 4.0 symbolizes the modern tendency to automatization of the manufacturing process by taking in different cuttingedge technologies, such as cyberphysical systems (CPS), Internet of Things (IoT), and cloud computing (Xu et al. 2018). Additionally, these technologies allow to increase the connectivity among different elements of the production process and stakeholders (Hermann et al. 2016; Zhong et al. 2017). Under this paradigm, practitioners can receive real-time data gathered directly from the shop floor (Monostori 2014; Rossit et al. 2019a), which increases the flexibility of the production system

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2079

allowing designs of innovative products and the development of personalization capabilities (Zhong et al. 2017; Rossit and Tohmé 2018). This has led to a mass customization of products in which the production line can be adapted to obtain personalized products that match the individual customer preferences (Zheng et al. 2019). However, in order to produce these personalized products and organize the corresponding variations in the production line, efficient sequencing approaches have to be implemented to fulfill customers’ specifications (Dolgui et al. 2019) and this is the main motivation for this chapter. As aforementioned, customization of products will be reflected in different usage of machines in the manufacturing cell of the production line. In general, the design of manufacturing cells aims at grouping operations that are involved in producing a certain family of similar products. This distribution increases both flexibility and client service quality (Pugazhendhi et al. 2003). However, when the customer selects the options for his/her particular product, defining the specifications, it is highly probable that not all the available operations will be required. Thus, some operations will be skipped. This is the so-called flow shop scheduling with missing operations (Glass et al. 1999; Pugazhendhi et al. 2004a; Dios et al. 2018). In the literature, several authors have addressed this problem (Ribas et al. 2010). The majority of them have considered only permutation sequences (PFS) (Venkataramanaiah 2008), meaning that each machine processes all the jobs in the same order. Another way of solving this problem is considering non-permutative schedules (NPFS), in which the production sequence does not need to be the same for all the machines involved. The expansion of the space search leads to a more complex optimization problem that is harder to solve (Rossit et al. 2020a). Since NPFS solutions include as a particular case the PFS solutions, NPFS solutions are at least as efficient as PFS (Pugazhendhi et al. 2003; Tseng et al. 2008; Benavides and Ritt 2016; Benavides and Ritt 2018; Rossit et al. 2019b). Despite the increasing importance of the skipping operations problem in the context of Industry 4.0, NPFS solutions have not yet been entirely studied in this context (Liu et al. 2018; Rossit et al., 2018, 2019c). Therefore, this work makes a contribution in that regard introducing a two-echelon approach for solving this problem. An initial stage applies a genetic algorithm (GA) to solve the problem considering only PFS solutions. This solution is later improved in the second stage with a simulated annealing (SA) algorithm that considers NPFS solutions. The results reported in the literature for both problem variants indicate that NPFS solutions outperform PFS solutions, for problems with missing operations, considering the makespan as objective function. Fig. 1 shows how different objectives functions have been addressed in the NPFS literature, where due-dates based objective functions represent less than 10% of the literature, meanwhile completion time based objective functions represents 73% (Rossit et al., 2018). This work is organized as follows. Section “Flow Shop Scheduling Problems” reviews the related research works. Section “Problem Statement” describes the skipping operations flow shop scheduling problem. Section “Numerical Examples” presents the resolution approach and Section “Conclusions” describes the computational experimentation and discusses the results.

2080 Fig. 1 Distribution of objective functions considered in the literature (Source: Rossit et al., 2018)

D. A. Rossit et al.

Multi-Obj % Costs %

Mono-Obj Studies 4%

Due-date 8%

Completiontime 73%

Flow Shop Scheduling Problems Flow shop problems are those problems where all the jobs must follow the same production sequence. That is, there is a set of n jobs and m machines, and each job j is made up of oij operations, where the operation oij corresponds to the i-th operation of job j to be performed at machine i. Therefore, given an ordering of the m machines, ordered as i = 1, 2, ..., m, then, all the jobs j perform their operations on the machines following that order (Pinedo 2012). The question to be solved in flow shop scheduling problems is the ordering of the n jobs in such a way that some criterion is optimized (Rajendran 1994). Regarding the complexity of scheduling problems, flow shop is NP-hard for the cases in which m is greater than or equal to three (Garey et al. 1976). One particularity of flow shop problems is that all jobs have the same sequence of operations, but this does not imply that jobs must be processed in the same order by all machines (Pinedo 2012). That is, from the point of view of technological constraints, each machine can process the n jobs regardless of the order in which they were processed in previous machines. The only condition, for a machine i to process a job j, is that before performing operation oij at machine i, job j must have finished operation oi − 1j at machine i − 1. Regarding this point, it is possible to introduce the particular case of when the m machines are forced to process the n jobs in the same sequence. This problem is known as the permutation flow shop (PFS) (Osman and Potts 1989). The PFS is the most widely studied version of flow shop problems (Benavides and Ritt 2016). The main advantage that the PFS environment provides is that it limits the size of the search space for feasible solutions to n!, while if that condition is relaxed, the size of the search space would be n!m (Tseng et al. 2008). Nonetheless, for objective functions such as the makespan, this search space can be reduced a little (Rossit et al. 2019b). Furthermore, in his seminal work Johnson demonstrated that this way of approaching the problem (i.e., considering

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2081

only permutation solutions) allows to obtain the optimal solution when considering the makespan as objective function, for the cases of two and three machines (Johnson 1954). That is, using the notation of Graham et al. (1979), for the problems of F2||Cmax and F3||Cmax the PFS approach provides the optimal solution. Basically, Johnson’s result shows that, for the case of two machines, it is possible to demonstrate that if a non-permutation job ordering is optimal, then the job ordering of the second machine can be replicated on the first machine, without delaying any of the starts of the operations performed on the second machine. Therefore, the makespan would not be increased. And, on the other hand, Johnson showed that a similar analysis makes it possible to ensure that an optimal solution can be obtained if the ordering of jobs for the last and second-to-last machines is kept constant (i.e., i = m, i = m − 1). Both results are for the makespan as an objective measure. This is the reason why if a system has three machines, and the ordering of the first two machines is the same, and that one of the last two machines also coincides, necessarily, the ordering of the jobs must be the same along the three machines. In other words, for Flow Shop problems with m less than or equal to three, it is enough to solve the problem in the permutation format. In turn, in that same work from the 1950s, Johnson proposed one of the most widely used algorithms in scheduling, which allows us to ensure the optimum for PFS problems with two machines. This procedure is known as the Johnson’s rule. We direct the reader to Gharbi et al. (2014) for the complete demonstration of the fact that for problems with three or fewer machines it is enough to solve the PFS variant. However, for the case in which m is greater than or equal to four, the PFS ceases to ensure the optimal solution even for the study of makespan (Conway et al. 1967). Therefore, although the PFS study allows reducing the dimension of the search space, it is possible that the permutation condition excludes the optimal solution of the problem (Benavides and Ritt 2018). That is why for problems where the number of machines is greater than four, the problem must be addressed in its most general version to ensure high quality solutions, that is, allowing the machines to process the jobs in a different order. This problem is often referred to as non-permutation flow shop (NPFS) (Rossit et al., 2018).

Non-permutation Flow Shop The NPFS is the most general case of flow shop problems, where only the technological constraints of the problem are incorporated (Pinedo 2012). In other words, in the NPFS problem the m machines are not required to process the n jobs respecting the same order. In this way, the PFS solutions are included as a subset in the feasible solutions of the NPFS set (note that, non-permutation solutions do not exclude the possibility of the job orderings being the same) which also incorporates other possible solutions. The inclusion of more feasible solutions, as mentioned above, causes the search space to grow in size considerably reaching n!m feasible solutions (Vahedi Nouri et al. 2013). Although for the makespan case, the search space size of feasible solutions is given by n!max{m − 2,1} (Rossit et al. 2020a).

2082

D. A. Rossit et al.

However, solving the scheduling problem using the NPFS format implies a greater degree of difficulty due to the large space of feasible solutions it has. In contrast to this increase in difficulty, NPFS solutions have yielded considerable improvements in terms of the objective function value compared to PFS solutions. In the case of objective functions related to due dates, these improvements have been around 10% (Liao et al. 2006). While for objective functions of the makespan type (i.e., objective functions based on the completion time of the jobs), the improvements have not been so pronounced, hovering around 1% (Benavides and Ritt 2018). However, if it is analyzed that these improvements occur by the simple fact of solving the problem directly without making any cut or modification, it is highly beneficial to study the problems in NPFS format (we have already commented that, in the vast majority of cases, the permutation condition is an aggregate when solving the problem and not an inherent restriction of the problem). Therefore, the possibilities of achieving substantial improvements in the performance of the calculated schedules, and the improvements in the computationally capacity, as well as in the algorithms, has caused the interest of the scientific community in NPFS problems to increase consistently during the last 20 years. Furthermore, if it is considered that at least a quarter of the production systems worldwide can be modeled as flow shop systems (Pan et al. 2011), the improvement due to NPFS may have a great impact (Rossit et al. 2016, 2018b). Nevertheless, a condition that the flow shop production system must have to allow the existence of non-permutation solutions is the availability of intermediate storages. Since, to reverse the precedence between two jobs (consecutive or not) it is necessary that the job that is being processed first must wait somewhere, so that the job that comes from behind can be processed in advance in that machine or in the following machines. In this sense, it is interesting to comment that an NPFS treatment requires the existence of intermediate buffers between the machines that allow these job surpassing. There are several types of intermediate buffers, depending on capacity and configuration. The most common case in the literature is the case of unlimited buffer capacity. On the other hand, the opposite case of unlimited capacity has also been studied, which is when there is no intermediate buffer defined. Naturally, between unlimited capacity and no intermediate stock are the case studies where buffers exist and have limited capacity. There are also two other remaining cases, the first case is when the intermediate buffer comprehends two or more of the introduced features, that is, mixed buffers, and the second case is when the jobs in process cannot wait at the buffer place, that is, exist the buffer, so jobs released the previous machine, but jobs cannot spend time at the buffer. Now, for a buffer to be cataloged as unlimited buffer, it is sufficient that it can hold n − 1 jobs. Therefore, if the storage has that capacity, then for the purposes of the scheduling problem it is no longer a restriction (there is no possibility that any operation will be affected by the buffer being saturated). The opposite case is when storage is always a restriction to consider within the scheduling problem, this case occurs when buffer capacity is null or does not exist such buffer. That is, every time a job j finishes its operation oij at machine i, it can only release machine i if machine i + 1 has finished processing job k (whatever job k). Whereas if machine i + 1 did

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2083

not finish processing job k, job k will not be able to release machine i. Note that the same happens with machine i + 1 and the successive machines, that is, i + 2 . . . . Therefore, in this situation it is impossible for a job j to surpass another job k, so NPFS orderings are not possible in this type of system. On the other hand, limited buffers allow to store | bi | units after the machine i has finished running (where bi is i-th buffer of the system, and | bi | the capacity of buffer bi ), with | bi | being less than n − 1 units. This implies that, if job j ends up on machine i, and bi located between i and i + 1, is not full while machine i + 1 is processing job k, then, job j can release machine i and job j pass to bi . However, if bi is full, then job j must wait (blocking machine i for further processing) until it finds a place in the buffer bi . bi will be able to free up space once machine i + 1 finishes job k and transfers the result to i + 2 or to the bi + 1 buffer (between machines i + 1 and i + 2), and so on. In these cases of limited buffers, NPFS schedules are possible.

Non-permutation Flow Shop with Missing Operations In this section, the most significant contributions related to flow shop scheduling research that consider production environments with missing or skip operations and non-permutation solutions are presented. The objective of this section is to contextualize our work within the framework of the non-permutation flow shop with missing operations literature. In Rajendran and Ziegler (2001), heuristics and dispatch rules are used to minimize the total flow time in a flow shop problem with missing operations considering non-permutation solutions. To address the same problem, Pugazhendhi et al. (2003) resort to insertion heuristics in order to optimize the total flow time. The largest problems covered in the article have a maximum of 40 jobs and 50 machines. In order to deepen the analysis, in Pugazhendhi et al. (2004a) non-permutation heuristic and dispatch rules solutions are contrasted against permutation solutions. In Pugazhendhi et al. (2004b) sequence-dependent setup times are incorporated into the study of the flow workshop problem with missing operations. The authors use a single-objective approach to jointly optimize makepan and total weighted flow time, addressing instances that have up to 30 jobs and 20 machines. The extension of the analysis of missing operations to hybrid flow shops (where there is more than one machine per production stage) is considered in Tseng et al. (2008). The problem is linked to a case from the stainless steel industry and makespan minimization is pursued during the solution process. A heuristic that incorporates permutation solutions as input is used to generate non-permutation solutions. An improved version of the heuristic from Pugazhendhi et al. (2003) is proposed in Henneberg and Neufeld (2016) to address the missing operations flow shop scheduling problem. Furthermore, the authors propose a two-step simulated annealing algorithm to solve the problem. Instances of up to 100 machines and x jobs were considered in the experimental work carried out by the authors.

2084

D. A. Rossit et al.

Non-permutation Flow Shop with Total Tardiness Although NPFS is a topic intensely studied by the scheduling community, more than 70% of the NPFS literature is devoted to the study of performance measures based on the completion times of jobs, such as the makespan, the total completion time, the total weighted completion time, the flow time, among others (Rossit et al., 2018). This is to be expected, since those performance measures dominate the rest of the PFS literature, and scheduling in general, where they are the most widely studied (Pinedo 2012). In this chapter, the emphasis will be placed on papers that consider due-datebased objective functions. The computational complexity of these problems is known to be high, and even “binary NP-hard” in two-machine cases (Błazewicz et al. 2005). Nonetheless, there is extensive research in the field of PFS configurations with due-date-based measures (Bła¨zewicz et al. 2008; Pesch and Sterna 2009). In this section, papers that consider total tardiness, total weighted tardiness, and maximum tardiness as performance criteria will be reviewed. The NPFS problem with total tardiness as performance measure is studied by Liao and Huang (2010). They present and evaluate three different MIP formulations and two Tabu Search algorithms. When contrasting NPFS and PFS approaches, they found that NPFS is better suited to address the problem. Swaminathan et al. (2004) consider minimizing total weighted tardiness criteria in a dynamic flow shop scheduling problem with variable processing times. Initially only permutation sequences are considered and then the permutation schedule requirement is relaxed to assess the benefits that can be obtained. Extending the analysis of the influence of enforcing the permutation condition on the general flow shop (non-permutation) problem, Swaminathan et al. (2007) investigate the problem with the total weighted tardiness objective employing three approaches: pure permutation, shift-based, and pure dispatching. Non-permutation schedules are obtained by using the third approach. They conclude that NPFS solutions produce better results and this approach is more effective to tackle the problem. In Xiao et al. (2015) the authors address a flow shop scheduling problem that considers order acceptance under a weighted tardiness objective using two solution approaches. First, for small instances of the problem, they present a MIP model solved through CPLEX. Second, for medium and large sized instances of the problem, they propose a two-phase genetic algorithm to solve a nonlinear integer programming (NLIP) formulation. A two-phase heuristic is introduced in Ziaee (2013) to solve the problem of minimizing total weighted tardiness in an NPFS environment with sequence-dependent setup times. The solution approach follows a standard twostep procedure, that is, first step generates a promising permutation solution, and the following one tries to improve the solution using a non-permutation local search method.

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2085

Problem Statement In this section, we properly introduce the problem that we address in this chapter. Firstly, we analyze the impact of Industry 4.0 in manufacturing cells, and how this influences the interaction with the customer. Then, we formalize our conception of the problem, relating our approach to the scheduling literature, and, finally, we model the problem considered in this work as mixed-integer programming formulations.

Industry 4.0 Environment Industry 4.0 proposes a digital transformation of traditional production systems (Zhong et al. 2017). This transformation is based on the implementation of cyberphysical systems (CPS), which integrate virtual and physical processes in the same system (Lee 2008). In turn, through IoT the CPS can communicate with each other and with the Decision Support Systems, which in terms of production planning, allows directly linking the shop floor with the support systems for production decision making (Almada-Lobo 2016; Rossit et al. 2019a). On the other hand, CPS allow generating a digital twin of the physical system (Lee et al. 2015), making it possible to analyze situations and making more informed decisions more agilely than in a traditional production system. This results in a much more flexible production system, better suited to different scenarios (Lu et al. 2019; Wang et al. 2019). In this way, the manufacturing system becomes a smart manufacturing system capable of providing an improved service to the customer, satisfying its needs in a personalized way (Yu et al. 2018; Rossit et al. 2019d). This allows to bring the final product closer to the true requirements of the client, who is actively involved in the design of his product (Zheng et al. 2017; Lu and Xu 2019). The customer expresses his/her preferences by specifying different variants of a base product belonging to a given family of products (Vollmann et al. 2005), constructing what is known as a personalized variant of the product (Simpson et al. 2006). In this way, by making the production system more flexible, and making it smart, Industry 4.0 allows to offer the customer a much higher level of personalization than traditional production systems (Yao and Lin 2016; Lu and Xu 2019).

Cyber-Physical Systems Cyber-physical systems (CPS) are systems that allow direct and deep integration of physical production processes with cybernetic or digital systems (Lee 2008). In this way, a much more efficient control and supervision is achieved than in traditional production systems, since the CPS allow access in real time to all the information generated in the production system and the physical process. This information is

2086

D. A. Rossit et al.

collected through sensors that can be associated with the same machine tools, as well as distributed on the shop floor. The sensors collect the data and send it to other CPS or decision support systems (DSS) linked via IoT. Therefore, it is clear that CPS are key technologies in this entire Industry 4.0 revolution since they are the systems that allow us to go from the physical process to the digital processing. To delve into these technologies, we resort to the definition generated from the CyPhERS Project (“Cyber-Physical European Roadmap and Strategy”) (Cengarle et al. 2013): A Cyber-Physical System (CPS) consists of computation, communication and control components tightly combined with physical processes of different nature, e.g., mechanical, electrical, and chemical. Typically, a CPS is defined and understood (evaluated) in a social and organizational context.

From this definition, it is clear that CPS surpass traditional control systems by incorporating greater processing and communication capacity, and another aspect highlighted by this definition is the adoption of context for its conception. In our case of interest, this context is the organizational one, mainly manufacturing type organizations. On the other hand, Wang et al. (2015) propose a series of characteristics that allow us to understand the technological leap represented by CPS compared to previous technologies. As the first characteristic, the authors make the difference between deeply embedded and IT (Information Technology) dominated technologies. Traditional embedded systems are represented by dedicated computer systems with very limited computing capabilities, as well as directly integrated with the physical processes with which they interact. However, currently, the distance between embedded systems and computer technologies themselves has begun to be absorbed by this type of CPS systems, where computing capacity is increasing while proximity to physical is kept. Even, given the increasing connectivity, the computing capacity of the CPS can be associated with a server or cluster, which increases it considerably. In other words, CPSs are not closed systems like embedded systems used to be. On the other hand, embedded technologies used to be associated with single domain applications, while CPS can include applications from multiple domains, which increases their versatility. Hand in hand with the increase in domains and functionalities, the CPS allow the increase of the autonomy of the systems. That is, they can reduce the need for human intervention or supervision, and operate safely and reliably. In this sense, Wang et al. (2015) highlight the CPS ability to govern the system, considering autonomous cars as an example. However, this governing ability will be highly exploited at the factory level enabling the possibility of decentralizing control and decision-making processes. Since, by including greater computing capacities, it is possible to implement tools based on artificial intelligence that allow controlling and solving the problem locally. That is, without the need to centralize all control and operations in hierarchical architecture. The decentralization, at the manufacturing level, greatly enhances the flexibility of the production system. Flexibility in manufacturing systems evokes the ability of the system to adapt and adjust to different production requirements. These production requirements may have their origin in an increasingly demanding market

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2087

for highly customized products, where the same product has an increasing number of non-standardized variants. To cope with these changes in the finished product, it is necessary that the production process can also be changed, and for that changes, flexibility is required. In traditional production systems, one of the main limitations to production flexibility is the need to centralize the entire production planning and decision-making process. Therefore, to face changing or undefined scenarios in the design stage of the production system, it is very difficult for a traditional production system to adapt to change. This situation is accentuated if it is considered that these scenarios can occur unexpectedly. Therefore, CPS-based systems can address in a decentralized way those issues that will be known and precisely defined, once this unfamiliar/unplanned scenario is faced (Yao and Lin 2016; Rossit and Tohmé 2020). That is, for each deviation or unexpected event that occurs, the system itself will be able to address the deviation or event, without the need to centralize all the information (considering that centralization implies data compatibility, filtering, data collation, etc.), which naturally increases the agility or flexibility of the system. In this sense, the ability of CPS to adapt to uncertain conditions is a key factor in their development as technologies. That is why CPS at the architecture and design level is a complex and challenging process (Rossit et al. 2020b). On the other hand, the design of CPS must consider the role that humans will play. That is, CPS may or may not require the supervision of a person. Even in the case in which they are supervised by a person, the control of the process associated with the CPS can be handled in a hybrid way, that is, they alternate CPS-Person and vice versa. Although this last configuration is of special importance at the level of logic and control design, since it must be a design that allows to clearly identify who is in command, and thus prevent both, the CPS and the person, from executing an involuntary control or action. And just as control can be fully integrated within a CPS, the same can happen with the data collection processes, data conversion into information, and then the processing stages (Rossit et al. 2019e, f). Therefore, the level of integration is another defining characteristic of CPS. In some cases, a total integration is proposed, where the CPS collects the information, processes and analyzes it, and generates the necessary configurations to continue with the process in execution (Lee et al. 2015).

IoT The IoT is another of the technologies that supports the production paradigm proposed by Industry 4.0. IoT allows the interconnection between the different systems and devices that make up the production process (Lee et al. 2015). This “Internet of Things” expression is attributed to a 1999 lecture by Kevin Ashton on RFID systems. Since then, the development acquired by IoT has revolutionized information technologies, and through them, all the areas where information technologies are of special importance. This revolution lies in the ease with which IoT allows you to collect data. In this sense, Kevin Ashton himself, in an interview conducted in 2009, explained (Ashton 2009):

2088

D. A. Rossit et al.

Today computers -and, therefore, the Internet- are almost wholly dependent on human beings for information. Nearly all of the roughly 50 petabytes (a petabyte is 1,024 terabytes) of data available on the Internet were first captured and created by human beings -by typing, pressing a record button, taking a digital picture or scanning a bar code. Conventional diagrams of the Internet include servers and routers and so on, but they leave out the most numerous and important routers of all: people. The problem is, people have limited time, attention and accuracy -all of which means they are not very good capturing data about things in the real world. And that’s a big deal. We’re physical, and so is our environment. Our economy, society and survival aren’t based on ideas or information -they’re based on things. You can’t eat bits, burn them to stay warm or put them in your gas tank. Ideas and information are important, but things matter much more. Yet today’s information technology is so dependent on data originated by people that our computers know more about ideas than things. If we had computers that knew everything there was to know about things -using data they gathered without any help of us- we would be able to track and count everything, and greatly reduce waste, loss and cost. We would know when things needed replacing, repairing or recalling, and whether they were fresh or past their best.

This is why IoT allows a qualitative leap in the collection and transmission of data and information of any kind at an industrial level. By complementing IoT with CPS, what is generated is a system with the capacity to collect, transmit, process, and resolve (almost) any type of data and events that arise at the factory level. Since, as Ashton mentions in his explanation, information technologies collect a large amount of data with high fidelity and constantly. Therefore, all control activities that require data will be greatly benefited by this type of technology (Halty et al. 2020).

Scheduling in Manufacturing Cells: Mathematical Formulations In this chapter, the problem of scheduling a set of jobs on a set of machines is considered, taking into account that all jobs can require processing on all the machines and follow the same technological sequence, that is, a flow shop scheduling problem. In addition, it is considered that each job has its own unique requirements, under a mass customization perspective encouraged by Industry 4.0 (Zhong et al. 2017; Wang et al. 2019), which implies a differential use of resources expressed through varying processing times. This differential use of resources in certain cases may directly result in some of the jobs not being processed in some of the machines, which causes the jobs to skip the corresponding operations (flow shop scheduling problem with missing operations). In the production environment under study, the objective is to schedule operations in order to achieve the highest levels of customer service, fulfilling the due dates of the jobs. To this end, the total tardiness related to the due date of jobs is adopted as the performance measure. At the same time, in addressing the problem, only permutation solutions (permutation flow shop (PFS)) could be taken into account, or non-permutation solutions (non-permutation flow shop (NPFS)) could also be incorporated. In the PFS setting, the sequences represented by the possible permutations of the n jobs are considered, that is, the solution search space is made up of a total of n! feasible solutions. The size of the solution space increases to n!m , where m represents the number of machines, when the processing order of the jobs can be modified in each

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2089

production stage of the flow shop, that is, NPFS setting. Using the classic notation of scheduling theory (Graham et al. 1979), the previous problems can be denoted, respectively, as F|prmu/missing|TTard and F|missing|TTard. In the following section, the mixed-integer programming model for PFS is introduced. Then, in Section “NPFS Model”, this model is further expanded in order to consider NPFS solutions.

PFS model Sets Jobs J indexed by {j}, j = 1, . . . ,n Machines I indexed by {i}, i = 1, . . . ,m Parameters pi, j Processing time of product j on machine i  large positive number dj due date of job j Variables Ci, j Completion time of job j on machine i xj  ,j Binary variable: 1 if job j’ is processed before job j Tj Tardiness of job j Minimize



Tj

j

Subject to: Ci,j ≥ Ci−1,j + pi,j , ∀j, i > 1

(1)

  Ci,j ≥ Ci,j  + pi,j − 1 − xj  ,j • , ∀i, j  = j

(2)

Ci,j  ≥ Ci,j + pi,j  − xj  ,j • , ∀i, j  = j

(3)

xj  ,j + xj,j  = 1, j = j 

(4)

  Tj = max Cm,j − dj ; 0 , ∀j

(5)

Ci,j > 0; xj,j  {0, 1}

(6)

The objective function to be minimized is the total tardiness, and it is defined as the sum of tardiness of all jobs. Constraint (1) forces the precedence of operations, that is, a job must be completed on the current machine before it passes to the next one. Constraints (2) and (7) work together indicating the ordering of jobs. If job j’ is processed before job j, then xj’,j becomes 1 and constraint (2) becomes active, while

2090

D. A. Rossit et al.

constraint (3) turns redundant. In expression (4) the logical order is respected: if job j’ is processed before job j, the converse cannot be valid. Constraint (5) defines the tardiness of each job j, which requires the comparison between the completion time of job j in the last machine m (Cm, j ) with its corresponding due date dj . While (6) enforces the non-negativity and binary conditions on the decision variables.

NPFS Model The NPFS model is similar to the PFS model, the main difference being that the job sequence may vary from machine to machine. For this, the variable x becomes now indexed by the set m of machines, as follows: xj  ,j,i Binary variable: 1 if job j’ is processed before job j on machine m Then the equations that are modified are (6), (7), and (8) and we get,   Ci,j ≥ Ci,j  + pi,j − 1 − xj  ,j,i • , ∀i, j  = j

(7)

Ci,j  ≥ Ci,j + pi,j  − xj  ,j,i • , ∀i, j  = j

(8)

xj  ,j,i + xj,j  ,im = 1, j = j  , ∀i

(9)

Here, constraints (6) and (8) are analogous to (2) and (3), but now the sequence of jobs may change at each machine of the system. Constraint (9) is a similar logical condition as (4) but now it is evaluated on every machine.

Numerical Examples In this section the case study analyzed in this work will be presented. The case analyzed consists of the study of a flow shop problem with missing operations. This case allows to analyze the impact that an extremely customized production system will have, where the production facilities have sufficient capacity to adapt to the personalized requirements of the clients. For this, it is started from the problem of traditional scheduling, that is, where the products respond to the installed production system, so there are no jobs that skip operations. And then cases are studied in which customers can order different variants of the product that imply differences in the production process. For this, the optimization methods used for both the problem in its PFS version and in its NPFS version will be presented. And then, the results, where the benefits of working on this problem from an NPFS perspective are observed, are shown.

Optimization Methods To solve the scheduling problem proposed here, approaches based on metaheuristics were used. Since, as mentioned above, meta-heuristics have been shown

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2091

to be one of the most efficient methods to address scheduling problems, and particularly, NPFS-type problems (Rossit et al. 2018). For the case of the PFS problem, a genetic algorithm of the steady type was used, which was programmed in Java using JMETAL as the base library (Durillo et al. 2006). This algorithm works with a permutation of length n, where each element of the array represents one of the n jobs that must be sequenced. That is, for a vector or individual to be feasible, it must not contain repeating elements. In turn, partially mapped crossover (PMX) was used as the genetic crossover operator, which ensures that crosses between individuals give rise to new feasible individuals. Therefore, it is not necessary to incorporate correction or repair methods of individuals. The PMX operator is the most used for those problems where the representation of the solution follows a permutation, as for example in the Traveling Salesman problem. To perform this operator, two chromosomes from different individuals (selected by some selection mechanism) are aligned, and two crossing points are randomly selected. The step-by-step process of PMX is as follows: 1. Each parent will have a subtring defined by the two crossing points. 2. Copy the substring of the first parent into the second child. 3. Do the same step, but with the subtring of the second parent and copy it into the first child. 4. Then, do a matching with the elements of both substrings from both parents, the match must relate value of the element and position of the element. 5. Complete the rest of the elements of the first child with the elements of the first parents, copying the value of each same positions. 6. In the case, that a value of the first parent, it is already present in the first child (that is, this value is already in the substring that was copied from the second parent), then, copy the value that corresponds to the match done in the Step 4. 7. Repeat steps 5 and 6 with the second child. On the other hand, to solve the NPFS problem, another widely used metaheuristic for scheduling problems was used, this is simulated annealing (Kirkpatrick et al. 1983). Simulated annealing (SA) algorithms have proven to be very efficient for non-permutation problems (Liao et al. 2006). SA is a local search method that uses an initial solution and performs a search in the neighborhood of that solution trying to improve the value of the objective function. In case of finding a solution that improves the value of the objective function, the method adopts that solution as the “current solution,” and then continues searching in the neighborhood. Therefore, in general lines it can be said that applying this procedure systematically, solutions that represent local optimum will be found. Nevertheless, this local optimum can be very inefficient regarding the whole problem, then, to avoid being stagnated in this solution and explore the rest of the feasible space, it is required to introduce some diversification method. In SA algorithms this diversification method is implemented by generating a variant in the first neighborhood exploration rule: (move only from “current solution” only if it improves to the current solution) and it is allowed to modify the current solution even when the candidate solution, do not improve

2092

D. A. Rossit et al.

the current objective function value. Naturally, the latter method occurs much less frequently than the initial one during the optimization process.

Results The problems to be studied are flow shop type problems, as described in Section “Flow Shop Scheduling Problems”. The objective of the experiments is to compare the performance of PFS and NPFS solutions to solve problems with missing operations with total tardiness as objective function. A natural difference between the two types of problems is the representation used for each algorithm. In the case of PFS, it is enough to use a permutation type chromosome (i.e., a linear vector of n elements), while for NPFS a permutation (of n elements) is required for each machine. Since the objective function to be analyzed is Total tardiness, the results of (Johnson 1954) are not valid (that results were presented in section “Non-permutation Flow Shop with Missing Operations”) and the entire space of n!m solutions must be explored. Therefore, the representation for the NPFS is an n x m matrix, where row i is associated with the production sequence that machine i will have for the n jobs. On the other hand, given the nature of each of the meta-heuristic methods presented, the genetic algorithm will be used to solve the PFS problem, while the SA will be used to solve the NPFS, since to work with genetic operators for a matrix chromosome type is rather cumbersome, while an SA algorithm operates more naturally in a matrix representation.

Data Generation For the generation of data to develop the experiments, the experimental designs developed by other previous works in the literature was followed, considering the two main features of this study, from one side the missing operation modeling, and from the other side, the objective function modeling, that is, due-dates considerations. For missing operations modeling we followed (Pugazhendhi et al. 2004a; Henneberg and Neufeld 2016) and for the design of due dates (Ruiz and Stützle 2008; Toncovich et al. 2019). To define the different parameters defined in Section “Scheduling in Manufacturing Cells: Mathematical Formulations,” the references mentioned were followed. The main parameter that gives the reason for this problem with missing operations is the processing time (pi,j ). These values are always integers, and are obtained using a pseudo-uniform distribution on the interval [0; 100], where the possibility of 0 is the one that represents the missing operation. In the pseudo-uniform distribution, the value zero (i.e., pi,j = 0), which is the probability of omitting an operation, has a special probability and, the rest of the integer values of the interval [1; 100] all have the same probability. Two groups of instances were constructed, where the proportion of missing operations in the problem varies, a group in which the special probability of missing operation is 5%, and another group, where this probability is 10%. The relatively large probability assigned to the value zero is intended to amplify the impact of missing operations on the scheduling

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2093

problem. For the calculation of the due dates of each job the main input is the processing time of each job (Toncovich et al. 2019). It is because, in Ruiz and Stützle (2008)the delivery date is defined according to the following expression: dj = rj + i∈I pi,j • (1 + random • 3). This representation includes all the operations that work j must go through for its completion plus an additional time considered by the term of the random number, which is a random number uniformly distributed in the interval [0; 1]. So dj is rounded to the nearest integer number. As with different percentages of missing operations, different sizes of problems were also experimented with. For this, different values of n and m were used, for n the values used were 40, 80, 120, and 150, while for m they were 15, 20, and 30. From these values, the instances of 40 × 15, 40 × 20, 80 × 20, 120 × 20, 120 × 30, and 150 × 30 were generated.

Experiments Results Given that the objective of the experiments is to deepen the study on the impact of restricting the study of flow shop problems only to permutation solutions. Therefore, what will be analyzed in the experiments is whether the NPFS approach manages to improve or not the PFS approach. For this, the PFS solution (PFSsol ) is analyzed in terms of objective function with respect to the NPFS solution (NPFSsol ). The comparison is made in percentage terms through the Improve indicator, which is calculated as follows: I mprove =

P F Ssol − NP F S sol • 100% P F Ssol

where a positive value of Improve implies that NPFSsol < PFSsol , that is, NPFSsol achieved a better value in the objective function, since the sense of the optimization problem is minimization . Whereas if Improve is equal to zero, it implies that NPFSsol = PFSsol , so the NPFS solution does not overcomes the result of the PFS problem. It is interesting to note that it is not possible that NPFSsol > PFSsol , since the PFS is only a particular case of the NPFS, so the Non-permutation solution is at least as good as the permutation one. Therefore, Improve cannot be negative. Since we worked with meta-heuristics, 30 runs were made for the same problem for each data set. The values presented in Table 1 represent the average values. Consequently, the standard deviation of Improve is also presented, that is, Dev.Improve. The results in Table 1 are grouped by problem size, as well as percentages of zeros. From the results of Table 1 it is observed that in general lines the NPFS approach is superior to the PFS since all the Improve values are greater than zero. In the last row the average of all the values is observed, and it can be observed that in general terms NPFS obtains an improvement of 1.1% for the case of 5% of missing operations, while for 10% of missing operations that value is of 1.8%. Therefore, as was assumed at the beginning, the higher the percentage of missing operation, the greater the benefits reported by the NPFS. Analyzing the specific cases for each instance, it can be commented that for the cases of smaller problems the difference between NPFS and PFS is greater than for the cases of larger problems. As in the

2094 Table 1 Comparative Results

D. A. Rossit et al. m 15 20

n

40 40 80 120 30 120 150 Average

Improve 5% 1.83% 2.80% 0.50% 0.53% 0.58% 0.55% 1.1%

10% 3.53% 3.97% 0.95% 0.82% 0.90% 0.81% 1.8%

Dev.Improve 5% 10% 1.73% 2.20% 3.36% 3.74% 0.43% 0.50% 0.27% 0.38% 0.27% 0.44% 0.23% 0.33% 1.0% 1.3%

case of 15 machines and 40 jobs, where if the missing operations proportion is of 5%, Improve is 1.83%, while if the missing operation proportion is extended to 10% Improve grows up to 3.53% (near double in Total tardiness reduction regarding PFS solution). In the other extreme, is the case of 30 machines and 150 jobs, where Improve is 0.55% is missing operations proportion is of 5%, and 0.81% is missing operation proportion is of 10%. In both extreme punctual cases, it is observed that the greater the missing operations percentage, the greater the difference between NPFS and PFS solutions. Regarding the deviation of the improvements with respect to the average (columns Dev.Improve from Table 1) it is observed that it is greater for the cases in which Improve is greater. Although the average Dev.Improve values are 1% and 1.3% for 5% and 10% of missing operation (last row of Table 1), it is noticeable that for most cases the Dev.Improve is less than 1%. Which indicates that the NPFS improvements are stable. However, it is observed as a particular case the problem of 20 machines and 40 jobs where deviations exceed 3%, although this peak in the deviation value coincides with the peak value for the improvement of the NPFS over the PFS.

Conclusions This chapter presents the study of the impact on scheduling problems of Industry 4.0 systems and the business models they drive. The business models that Industry 4.0 promotes are those based on deep customization by customers, since Industry 4.0 allows to considerably increase the flexibility of the production system. These customizations allow to get closer to the ideal of being able to produce for each client “the” product they need, reaching a production of the type “one of a kind,” in contrast to the traditional models that seek to standardize production, which are usually associated with the motto “many of the same.” The main technologies driving this fourth industrial revolution are Cyber-physical Systems and the Internet of Things. CPS allow a growing integration in the management of data and information, since CPS can collect, process, and analyze the information. But they can also develop actions based on artificial intelligence methods based on this processing, for example, to increase the autonomy of the system. All this is

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2095

magnified by the communication and transmission capacity that IoT grants to the CPS, so that the CPS are no longer closed in on themselves, but can share the information with other CPS (or with the cloud), and by this, increase the visibility of the processes and status of the shop floor. As a particular case study, scheduling problems in flow shop configurations are considered, as it is one of the main production configurations. As a particularity of the chapter, this problem was addressed in both the permutation and nonpermutation forms. The permutation format is the most used and analyzed when solving flow shop scheduling problems, while the non-permutation one is a more general case than the permutation one, and ensures the same or better result than the permutation one. The trade-off of these improvements is that the space for feasible solutions grows exponentially. On the other hand, given that the motivation for the problem is highly customized production, it was used as a strategy to address these cases: problems where these customizations mean that the production processes of the different products are not all the same. Therefore, given a flow shop production system, there are operations that the products should not carry out. This type of problem, called missing operation, can be studied from the perspective of the PFS and NPFS. In this chapter, two meta-heuristics were used to solve both approaches. One of the meta-heuristics was based on a Genetic algorithm of the Steady type, while the other meta-heuristic used Simulated Annealing algorithm. Both algorithms were able to obtain efficient solutions for the two problems. When comparing both solutions it was observed that the NPFS improved to the PFS solution. Therefore, for the study of this type of problem, it is recommended to use approaches that consider nonpermutation solutions to the problem. On the other hand, the impacts of the presence of missing operations and their proportion were analyzed. As future lines of work, the study of production systems based on other configurations is proposed, such as the job shop or the hybrid flow shop. The case of the Hybrid flow shop is of singular interest since in these environments the type of machines plays a special role, since in the case in which the machines of the same stage of the process are not identical, then the eligibility of machines so that the final result is optimal. In turn, it would also be interesting to study other types of objective functions, where the criterion to evaluate a production sequence is one based on completions times. This last line of research based on objective functions is of special interest for problems based on Industry 4.0 as it is a new production paradigm, which impacts many of the traditional production conditions.

Important Websites International Society of Automation: https://www.isa.org/ Reference Architectural Model Industrie 4.0 (RAMI 4.0) https://ec.europa.eu/ futurium/en/system/files/ged/a2-schweichhart-reference_architectural_model_ industrie_4.0_rami_4.0.pdf The Scheduling Zoo: http://schedulingzoo.lip6.fr/

2096

D. A. Rossit et al.

References Almada-Lobo F (2016) The industry 4.0 revolution and the future of manufacturing execution systems (MES). J Innov Manage 3(4):16–21 Ashton K (2009) That ‘internet of things’ thing. RFID J 22(7):97–114 Benavides AJ, Ritt M (2016) Two simple and effective heuristics for minimizing the makespan in non-permutation flow shops. Comput Oper Res 66:160–169 Benavides AJ, Ritt M (2018) Fast heuristics for minimizing the makespan in non-permutation flow shops. Comput Oper Res 100:230–243 Cengarle V, Törngren M, Bensalem S, McDermid J, Sangiovanni-Vincentelli A, Passerone R (2013) Structuring of CPS domain: characteristics, trends, challenges and opportunities associated with CPS. D2.2 of CyPhERS FP7 Project. http://www.cyphers.eu/sites/default/files/ D2.2.pdf Bła¨zewicz J, Pesch E, Sterna M, Werner F (2005) A comparison of solution procedures for twomachine flow shop scheduling with late work criterion. Computers & Industrial Engineering 49(4):611–624 Bła¨zewicz J, Pesch E, Sterna M, Werner F (2008) Metaheuristic approaches for the two-machine flow-shop problem with weighted late work criterion and common due date. Computers & Operations Research, 35(2):574–599 Cengarle V, Törngren M, Bensalem S, McDermid J, Sangiovanni-Vincentelli A, Passerone R (2013) Structuring of CPS domain: characteristics, trends, challenges and opportunities associated with CPS. D2.2 of CyPhERS FP7 Project. http://www.cyphers.eu/sites/default/files/ D2.2.pdf Conway RW, Maxwell WL, Miller LW (1967) Theory of scheduling. Courier Corporation Dios M, Fernandez-Viagas V, Framinan JM (2018) Efficient heuristics for the hybrid flow shop scheduling problem with missing operations. Comput Ind Eng 115:88–99 Dolgui A, Ivanov D, Sethi SP, Sokolov B (2019) Scheduling in production, supply chain and Industry 4.0 systems by optimal control: fundamentals, state-of-the-art and applications. Int J Prod Res 57(2):411–432 Durillo JJ, Nebro AJ, Luna F, Dorronsoro B, Alba E (2006) jMetal: a java framework for developing multi-objective optimization metaheuristics. Departamento de Lenguajes y Ciencias de la Computación, University of Málaga, ETSI Informática, Campus de Teatinos, Tech Rep ITI-2006-10 Garey MR, Johnson DS, Sethi R (1976) The complexity of flowshop and jobshop scheduling. Math Oper Res 1(2):117–129 Gharbi A, Labidi M, Louly MA (2014) The nonpermutation flowshop scheduling problem: adjustment and bounding procedures. J Appl Math:273567, 14 p. https://doi.org/10.1155/2014/ 273567 Glass CA, Gupta JN, Potts CN (1999) Two-machine no-wait flow shop scheduling with missing operations. Math Oper Res 24(4):911–924 Graham RL, Lawler EL, Lenstra JK, Kan AR (1979) Optimization and approximation in deterministic sequencing and scheduling: a survey. In: Annals of discrete mathematics, vol 5. Elsevier, pp 287–326 Halty A, Sánchez R, Vázquez V, Viana V, Piñeyro P, Rossit DA (2020) Scheduling in cloud manufacturing systems: Recent systematic literature review. Math Biosci Eng 17(6):7378–7739 Henneberg M, Neufeld JS (2016) A constructive algorithm and a simulated annealing approach for solving flowshop problems with missing operations. Int J Prod Res 54(12):3534– 3550 Hermann M, Pentek T, Otto B (2016) Design principles for industrie 4.0 scenarios. In: 2016 49th Hawaii international conference on system sciences (HICSS). IEEE, pp 3928–3937 Johnson SM (1954) Optimal two-and three-stage production schedules with setup times included. Naval Res Logistics Q 1(1):61–68 Kirkpatrick S, Gelatt CD, Vecchi MP (1983) Optimization by simulated annealing. Science 220(4598):671–680

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2097

Lee EA (2008) Cyber physical systems: design challenges. In: 2008 11th IEEE international symposium on Object oriented real-time distributed computing (isorc). IEEE, pp 363–369 Lee J, Bagheri B, Kao HA (2015) A cyber-physical systems architecture for industry 4.0-based manufacturing systems. Manuf Lett 3:18–23 Liao CJ, Liao LM, Tseng CT (2006) A performance evaluation of permutation vs. non-permutation schedules in a flowshop. Int J Prod Res 44(20):4297–4309 Liao LM, Huang CJ (2010) Tabu search for non-permutation flowshop scheduling problem with minimizing total tardiness. Appl Math Comput, 217(2):557–567 Liu Y, Wang L, Wang XV, Xu X, Zhang L (2018) Scheduling in cloud manufacturing: state-of-theart and research challenges. Int J Prod Res 32(8):1–26 Low C (2005) Simulated annealing heuristic for flow shop scheduling problems with unrelated parallel machines. Comput Oper Res 32(8):2013–2025 Lu Y, Xu X (2019) Cloud-based manufacturing equipment and big data analytics to enable ondemand manufacturing services. Robot Comput Integr Manuf 57:92–102 Lu Y, Peng T, Xu X (2019) Energy-efficient cyber-physical production network: architecture and technologies. Comput Ind Eng 129:56–66 Luo H, Fang J, Huang GQ (2015) Real-time scheduling for hybrid flowshop in ubiquitous manufacturing environment. Comput Ind Eng 84:12–23 Metropolis N, Rosenbluth AW, Rosenbluth MN, Teller AH, Teller E (1953) Equation of state calculations by fast computing machines. J Chem Phys 21(6):1087–1092 Monostori L (2014) Cyber-physical production systems: roots, expectations and R&D challenges. Procedia CIRP 17:9–13 Nesmachnow S (2014) An overview of metaheuristics: accurate and efficient methods for optimisation. Int J Met 3(4):320–347 Osman IH, Potts CN (1989) Simulated annealing for permutation flow-shop scheduling. Omega 17(6):551–557 Pan QK, Tasgetiren MF, Suganthan PN, Chua TJ (2011) A discrete artificial bee colony algorithm for the lot-streaming flow shop scheduling problem. Inf Sci 181(12):2455–2468 Pesch E, Sterna M (2009) Late work minimization in flow shops by a genetic algorithm. Comput Ind Eng, 57(4):1202–1209 Pinedo M (2012) Scheduling, New York, 29 Ponnambalam SG, Reddy M (2003) A GA-SA multiobjective hybrid search algorithm for integrating lot sizing and sequencing in flow-line scheduling. Int J Adv Manuf Technol 21(2):126–137 Pugazhendhi S, Thiagarajan S, Rajendran C, Anantharaman N (2003) Performance enhancement by using non-permutation schedules in flowline-based manufacturing systems. Comput Ind Eng 44(1):133–157 Pugazhendhi S, Thiagarajan S, Rajendran C, Anantharaman N (2004a) Relative performance evaluation of permutation and non-permutation schedules in flowline-based manufacturing systems with flowtime objective. Int J Adv Manuf Technol 23(11–12):820–830 Pugazhendhi S, Thiagarajan S, Rajendran C, Anantharaman N (2004b) Generating nonpermutation schedules in flowline-based manufacturing sytems with sequence-dependent setup times of jobs: a heuristic approach. Int J Adv Manuf Technol 23(1–2):64–78 Rajendran C (1994) A heuristic for scheduling in flowshop and flowline-based manufacturing cell with multi-criteria. Int J Prod Res 32(11):2541–2558 Rajendran C, Ziegler H (2001) A performance analysis of dispatching rules and a heuristic in static flowshops with missing operations of jobs. Eur J Oper Res 131(3):622–634 Ribas I, Leisten R, Framiñan JM (2010) Review and classification of hybrid flow shop scheduling problems from a production system and a solutions procedure perspective. Comput Oper Res 37(8):1439–1454 Rossit D, Tohmé F (2018) Scheduling research contributions to smart manufacturing. Manuf Lett 15(B):111–114 Rossit DA, Tohmé F (2020) The tolerance scheduling problem for maximum lateness in Industry 4.0 systems. In: Advances in mathematics for industry 4.0. Academic Press, London, pp 95–113

2098

D. A. Rossit et al.

Rossit D, Tohmé F, Frutos M, Bard J, Broz D (2016) A non-permutation flowshop scheduling problem with lot streaming: a mathematical model. Int J Ind Eng Comput 7(3):507–516 Rossit DA, Tohmé F, Frutos M (2018a) The non-permutation flow-shop scheduling problem: a literature review. Omega 77:143–153 Rossit DA, Vásquez ÓC, Tohmé F, Frutos M, Safe MD (2018b) The dominance flow shop scheduling problem. Electron Notes Discrete Math 69:21–28 Rossit DA, Tohmé F, Frutos M (2019a) Industry 4.0: smart scheduling. Int J Prod Res 57(12):3802– 3813 Rossit DA, Vásquez ÓC, Tohmé F, Frutos M, Safe MD (2019b) A combinatorial analysis of the permutation and non-permutation flow shop scheduling problems. Eur J Oper Res. https:// doi.org/10.1016/j.ejor.2019.07.055 Rossit DA, Tohmé F, Frutos M (2019c) Production planning and scheduling in cyber-physical production systems: a review. Int J Comput Integr Manuf 32(4–5):385–395 Rossit DA, Tohmé F, Frutos M (2019d) An industry 4.0 approach to assembly line resequencing. Int J Adv Manuf Technol 105(9):3619–3630 Rossit DA, Tohmé F, Frutos M (2019e) A data-driven scheduling approach to smart manufacturing. J Ind Inf Integr 15:69–79 Rossit DA, Tohmé F, Frutos M (2019f) Designing a scheduling logic controller for Industry 4.0 environments. IFAC-PapersOnLine 52(13):2164–2169 Rossit D, Tohmé F, Frutos M, Safe M, Vásquez Ó (2020a) Critical paths of non-permutation and permutation flow shop scheduling problems. Int J Ind Eng Comput 11(2):281–298 Rossit DA, Tohmé F, Delgadillo GM (2020b) The tolerance scheduling problem in a single machine case. In: Scheduling in industry 4.0 and cloud manufacturing. Springer, Cham, pp 255–273 Ruiz R, Stützle T (2008) An Iterated Greedy heuristic for the sequence dependent setup times flowshop problem with makespan and weighted tardiness objectives. Eur J Oper Res 187(3):1143–1159 Shim SO, Park K, Choi S (2017) Innovative production scheduling with customer satisfaction based measurement for the sustainability of manufacturing firms. Sustainability 9(12):2249 Simpson TW, Siddique Z, Jiao RJ (eds) (2006) Product platform and product family design: methods and applications. Springer Science & Business Media Swaminathan R, Fowler JW, Pfund ME, Mason SJ (2004) Minimizing total weighted tardiness in a dynamic flowshop with variable processing times. In: IIE Annual Conference. Proceedings (p. 1). Institute of Industrial Engineers-Publisher Swaminathan R, Pfund ME, Fowler JW, Mason SJ, Keha A (2007) Impact of permutation enforcement when minimizing total weighted tardiness in dynamic flowshops with uncertain processing times. Comput Oper Res, 34(10):3055–3068 Toncovich A, Rossit DA, Frutos M, Rossit DG (2019) Solving a multi-objective manufacturing cell scheduling problem with the consideration of warehouses using a simulated annealing based procedure. Int J Ind Eng Comput 10(1):1–16 Tseng CT, Liao CJ, Liao TX (2008) A note on two-stage hybrid flowshop scheduling with missing operations. Comput Ind Eng 54(3):695–704 Uhlmann IR, Frazzon EM (2018) Production rescheduling review: opportunities for industrial integration and practical applications. J Manuf Syst 49:186–193 Vahedi Nouri B, Fattahi P, Ramezanian R (2013) Hybrid firefly-simulated annealing algorithm for the flow shop problem with learning effects and flexible maintenance activities. Int J Prod Res 51(12):3501–3515 Venkataramanaiah S (2008) Scheduling in cellular manufacturing systems: an heuristic approach. Int J Prod Res 46(2):429–449 Vollmann TE, Berry WL, Whybark DC, Jacobs R (2005) Manufacturing planning and control for supply chain management, 5th edn. McGraw-Hill, Irwin Wang L, Törngren M, Onori M (2015) Current status and advancement of cyber-physical systems in manufacturing. J Manuf Syst 37:517–527

81 Flow Shop Scheduling Problems in Industry 4.0 Production. . .

2099

Wang M, Zhong RY, Dai Q, Huang GQ (2016) A MPN-based scheduling model for IoT-enabled hybrid flow shop manufacturing. Adv Eng Inform 30(4):728–736 Wang Y, Zheng P, Xu X, Yang H, Zou J (2019) Production planning for cloud-based additive manufacturing – a computer vision-based approach. Robot Comput Integr Manuf 58:145–157 Xiao Y, Yuan Y, Zhang RQ, Konak A (2015) Non-permutation flow shop scheduling with order acceptance and weighted tardiness. Appl Math Comput, 270:312–333 Xu LD, Xu EL, Li L (2018) Industry 4.0: state of the art and future trends. Int J Prod Res 56(8):2941–2962 Yao X, Lin Y (2016) Emerging manufacturing paradigm shifts for the incoming industrial revolution. Int J Adv Manuf Technol 85(5-8):1665–1676 Yu C, Mou S, Ji Y, Xu X, Gu X (2018) A delayed product differentiation model for cloud manufacturing. Comput Ind Eng 117:60–70 Zheng P, Yu S, Wang Y, Zhong RY, Xu X (2017) User-experience based product development for mass personalization: A case study. Procedia CIRP 63:2–7 Zheng P, Lin Y, Chen CH, Xu X (2019) Smart, connected open architecture product: an IT-driven co-creation paradigm with lifecycle personalization concerns. Int J Prod Res 57(8):2571–2584 Zhong RY, Xu X, Klotz E, Newman ST (2017) Intelligent manufacturing in the context of industry 4.0: a review. Engineering 3(5):616–630 Ziaee M (2013) General flowshop scheduling problem with the sequence dependent setup times: A heuristic approach. Inform Sciences, 251:126–135

Smart Packaging: O2 Scavenger for Improving Quality of Fish

82

C. O. Mohan, S. Remya, K. R. Sreelakshmi, Anuj Kumar, and C. N. Ravishankar

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Status of Smart Packaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reduced O2 Packaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . High Barrier Films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vacuum Packaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modified Atmosphere Packaging (MAP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . O2 Scavenger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Web References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2102 2104 2106 2107 2109 2110 2113 2119 2119 2121

Abstract

Oxygen (O2 ) is the life line of animals and many other organisms including human beings which is essential for their living. But for perishable food commodities, O2 becomes a limiting factor to enhance its shelf life and quality. Removing this O2 inside the food pack is essential to increase the eating quality of perishable food commodities. Fish is regarded as one of the healthy food commodities but at the same time it is also highly perishable. Proper preservation of this nutrient-rich food is essential to ensure nutritional as well as food security. Smart packaging technology, particularly O2 scavenger, is one of the emerging technologies that assumes great importance in enhancing the shelf life of perishable food commodities including fish. This chapter describes different reduced oxygen packaging technologies giving more emphasis on the O2 scavenging technology for fish preservation.

C. O. Mohan () · S. Remya · K. R. Sreelakshmi · A. Kumar · C. N. Ravishankar ICAR-Central Institute of Fisheries Technology, Kochi, India © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_72

2101

2102

C. O Mohan et al.

Keywords

Smart packaging · O2 scavenger · Perishable commodity · Reduced oxygen pack

Introduction Fish has long been an important part of human diets and increasingly a major source of economic value. Globally, fish accounted 6.7% of all protein consumed by humans. Fish is a rich source of omega-3 fatty acids, particularly EPA and DHA, vitamins, calcium, zinc, and iron. There is a great demand for fish and shellfishes in the international market due to its proven health benefits. Fish business not only provides foreign exchange, it also helps in ensuring nutritious food, employment to millions of people, many of whom are below the poverty line. As per FAO (2018), the world fish production in 2016 has reached 171 million tonnes and 88% of this was used for direct human consumption. Of the total fish produced, aquaculture represented 47%. Global sale value of fisheries in 2016 was estimated at US$ 362 billion of which US $ 232 billion was contribution from aquaculture production. The value of global fish exports in 2017 was USD 152 billion, up from $8 billion in 1976, and 54% of this was originating from developing countries, indicating the contribution of seafood export to the building of nations. Nearly 57 million people are engaged in the primary fish production sector, a third of them in aquaculture. Of all the global merchandise trade, fishery products accounted for 1% in terms of value, representing more than 9% of total agricultural exports. Export of fishery product is one of the major foreign exchange earners in developing countries which accounted to over USD 80 billion in 2017, providing higher net trade revenues than meat, tobacco, rice, and sugar combined. The per capita consumption during 2016 reached 20.3 kg and is expected to increase further. On an average, fish provides nearly 6.7% of all protein consumed by human beings. These indicate there is an ever-increasing demand for fish across the countries, which should be met by increasing the production. Due to ever increasing demand for fish, global requirement is increasing steadily. From the available data, it is observed that both fish production and fish consumption increased from 1975. Till 1990s, capture fisheries was the major contributor and later, the importance of aquaculture increased resulting in sizeable contribution to total production. As the years pass, the fish consumption level is also increasing. Per capita consumption was 9 kg in 1961 which increased to 17.2 kg in 2008 and further increased to 20.5 kg/year in 2018 (FAO 2020). Considering a stable consumption at 2008 level, the fish requirement estimated as nearly 140 and 152 million tonnes by 2025 and 2050, respectively. However, as the consumption is not stable and it is increasing steadily, the additional demand considering the rate of increase in consumption between 1975 and 2008 will be 164 and 232 million tonnes by 2025 and 2050, respectively. Meeting this huge demand is a herculean task, which creates huge pressure on both capture production and aquaculture. As it is very

82 Smart Packaging: O2 Scavenger for Improving Quality of Fish

2103

important to increase the production of fish to meet the global demand, it is also equally important for its proper utilization without wasting by adopting responsible handling and advanced processing and packaging to reduce the postharvest loss. It is estimated that nearly 18–20% of global fish production is wasted as postharvest loss resulting in huge loss of valuable nutrient-rich food commodity. This also affects economy of the country negatively. For instance, in India alone, very huge quantity of fish produced is lost annually, which is very huge loss to the economy as well as loss of nutritive food commodity. Apart from proper handling, there is a need to adopt advanced packaging technologies, particularly cost-effective smart packaging to overcome this problem. Although advanced packaging technologies like vacuum and modified atmosphere packaging (MAP) technologies are available, their adoption in middle- and low-income countries is very insignificant due to its high cost and its maintenance. In vacuum packaging, air inside the pack is removed completely so that oxidative changes and growth of aerobic spoilage and pathogenic bacteria can be controlled. In MAP, package atmosphere is replaced with suitable gases to enhance the quality and shelf life of food product. Generally recommended combination of gases used for lean and fatty fishes is CO2 :N2 :O2 in the ratio 40:30:30 and 60:40:0, respectively. However, the gas composition needs to be optimized for fishes depending on their fat content to obtain better results. Apart from reducing postharvest losses, providing quality and safe fish products without use of any chemical preservative is a challenge world is facing today. Advanced, low-cost packaging options to enhance the quality, shelf-life, and safety are the need of the hour. Although advanced technologies like vacuum packaging and modified atmosphere packaging are developed, lack of its adaptation uniformly across the countries arises many questions. These are as follows: Although vacuum and MAP technologies are effective, is the cost of vacuum and MAP technologies are the major deciding factor in adopting technology? Are the vacuum and MAP technologies maintain their effectiveness throughout their storage life to retain the quality of the products? Whether smart packaging technologies will be effective to replace vacuum and MAP? Will the adaptation rate be higher for smart packaging technologies compared to vacuum and MAP? Whether nutrient-rich fishery products can be transported to distant locations maintaining its freshness and quality by adopting smart packaging technologies? Whether postharvest loss can be minimized by effective adaptation of smart packaging? The smart active O2 scavenging packaging system can be a suitable alternative to address these issues effectively. The smart packaging, particularly active O2 scavenging sachets and films, is gaining wide popularity across the globe. As fish is highly perishable, the developed active sachet/film will have immediate application to enhance the quality and shelf life of fishery products thereby helping in reducing the postharvest loss. Smart packaging includes both active and intelligent packaging. Active packaging is a type of packaging that changes the condition of the packaging and “maintains throughout the storage period” to extend the shelflife and safety or to improve safety or sensory properties, while maintaining the quality. Active packaging refers to alternation of package atmosphere/incorporation

2104

C. O Mohan et al.

Market value (Billion US$) 3

Billion US $

2.5 2 1.5 1 0.5 0 2016

2018

2020

2022*

2025*

Year Fig. 1 Market value of O2 scavenger between 2016 and 2025. *Projected value

of suitable compounds to enhance the quality and shelf life, whereas intelligent system monitors the condition of packaged food to give information regarding the quality of food (Rooney 1992; Ahvenainen 2003). The global demand for smart packaging market is expected to grow from USD 40,018.58 million in 2019 to USD 58,791.56 million by the end of 2025 (https://www.reportlinker.com/p05881809/? utm_source=PRN). Smart packaging is very crucial to the fish products to minimize the loss of valuable nutritious commodity and to enhance the quality, shelf life, and safety of fishery products. Among the smart packaging, O2 scavenger has wide spread usage due to its wide applications. The market value of O2 scavenger is given below. Figure 1 indicates that the market value of O2 scavenger was 1.73 billion US $ in 2016 which has increased to 1.9 billion US $ by 2020. It is projected to increase to 2.67 billion US $ by 2025, indicating its increasing demand.

Status of Smart Packaging Traditionally, food packaging is aimed for protection, communication, convenience, and containment (Paine 1991; Robertson 2006). The package is used to protect the product from the deteriorative effects of external environmental conditionals like heat, light, presence or absence of moisture, pressure, microorganisms, gaseous emissions, and so on. It also provides the consumer with the greater ease of use and time-saving convenience and contains product of various size and shapes (Yam et al. 2005; Marsh and Bugusu 2007). The key safety objective for traditional packaging

82 Smart Packaging: O2 Scavenger for Improving Quality of Fish

2105

materials which comes in contact with food is to be inert as possible. While the smart packaging systems like active and intelligent packaging concepts are based on the useful interaction between packaging environment and the food to provide active protection to the food. Packaging may be termed active when it performs some role other than providing an inert barrier to the external environment (Rooney 1995a, b). Active packaging can be defined as a system in which the product, package, and the environment interact in a positive way to extend the shelf life or to achieve some characteristics (Miltz et al. 1997). It has also be defined as a type of packaging that changes the condition of the packaging to extend shelf life or to improve safety or sensory properties while maintaining the quality of the packaged food (Ahvenainen 2003). According to regulation 1935/2004/EC and 450/2009/EC, active materials and articles are intended to extend the shelf life or to maintain or improve the condition of packaged food. They are designed to deliberately incorporate components that would release or absorb substances into or from the packaged food or environment surrounding the food (Sivertsvik 2007; Floros et al. 1997). The goal of active packaging is to enhance the preservation of food in the package and prolonging shelf life involves application of various strategies like temperature control, oxygen removal, moisture control, addition of chemicals such as salt, sugar, carbon dioxide, or natural acids or a combination of these with effective packaging (Robertson 2006; Restuccia et al. 2010). These developments in active packaging have led to advances in many areas including delayed oxidation in muscle foods, controlled respiration rate in horticultural products, microbial growth, and moisture migration in dried products. The concept of internal migration of preservatives to food and the communication function of the package to facilitate decision making are related with intelligent packing. According to EC/450/2009, intelligent materials and articles means materials and articles which monitor the condition of packaged food or the environment surrounding the food. Intelligent packaging systems provide the user with information on the conditions of the food or its environment (temperature, pH). It is an extension of the communication function of traditional packaging and communicates to the consumer based on its ability to detect, sense, and record the changes in the products environment (Restuccia et al. 2010; Realini and Marcos 2014). In contrary to active components, intelligent components do not have the intention to release their constituents into the food. The intelligent packaging can also contribute to improving Hazard Analysis and Critical Control Points’ (HACCP) and Quality Analysis and Critical Control Points’ (QACCP) systems (Heising et al. 2014), which are developed to onsite detection of unsafe food, identify potential health hazards, and establish strategies to reduce or to eliminate their occurrence. It also helps to identify processes that strongly affect the quality attributes and efficiently improve the final food quality (Vanderroost et al. 2014). Basically there are three intelligent systems; sensors, indicators, and radiofrequency identification (RFID) systems (Kerry et al. 2006; Vanderroost et al. 2014). In India, the research on development of smart packaging devices for perishables including fish is very limited. Few organizations in India have initiated preliminary studies in this direction and notable one includes work carried out at Central Institute

2106

C. O Mohan et al.

of Fisheries Technology, Kochi, working under the Indian Council of Agricultural Research, New Delhi of Ministry of Agriculture & Farmers Welfare, Govt. of India, and research work on active packaging is reported (Mohan 2008; Mohan et al. 2008, 2009a, b, 2010, 2012, 2018, 2019a; Remya 2017; Remya et al. 2014, 2015, 2016, 2017, 2018). Similar to active packaging systems, ICAR-CIFT has also developed various intelligent packaging systems. A simple, easy to use, and cheap freshness indicator is developed for indicating quality of fish and shellfishes. The effectiveness of freshness indicator is evaluated with fishes of freshwater, marine, and brackish water. Studies on nanoparticle-based intelligent packaging system to develop temperature history sensor were developed in association with University of Wisconsin-Madison, USA (Mohan et al. 2019b). Nanocomposite of chitosan and gold nanoparticles (AuNPs) was used to develop sensors that can indicate the frozen state and thermal history of foods and other temperature-sensitive products based on the visual color change (Wang et al. 2018). A greener method is used for the synthesis of gold nanoparticle using chitosan to develop temperature history indicator to ensure the quality and safety of frozen stored perishable food and pharma products during shipment and transportation (Mohan et al. 2019). Apart from this, gold, silver and copper nanoparticles are developed and characterized using different chemicals and biological sources of marine origin which finds application as biogenic amine and heavy metal detecting sensor. Developed paperbased colorimetric nano-biosensor strip for the detection of food-borne pathogens including E. coli 0157:H7 and E. coli which reduced the detection time and a detailed reviews on active and intelligent packaging systems is documented (Mohan et al. 2009, 2010, 2018; Biji et al. 2015; Remya et al. 2020). Apart from this, institute is also steer heading the research on developing biodegradable smart packaging materials with improved properties. Chitosan, collagen, and chitosan-collagenbased films are developed for their application as wrap. PLA-based biodegradable packaging material with improved mechanical properties and heat sealability is developed. Seaweed-based functional and edible films developed exhibited good sealing and antioxidant properties and can be used as novel packaging material in food industry as a sachet/pouch/bag for seasoning powder for instant noodles, instant coffee/tea, etc. However, research on developing O2 scavenging film is initiated as it has wider application across food systems.

Reduced O2 Packaging Oxygen is the life line to animal and plants in their living condition. But when the organism dies, the same O2 will become detrimental for its prolonged preservation. Various approaches are in use across the globe to preserve food products including fish. These include low temperature preservation techniques (icing, refrigeration, super chilling, or freezing), high temperature techniques (thermal processing, smoking, sous-vide, pasteurization, etc.), reducing water available for action of enzymes and microorganisms (drying, curing, freeze drying, spray drying, osmotic treatments), using radiation (both ionizing and nonionizing radiations),

82 Smart Packaging: O2 Scavenger for Improving Quality of Fish

2107

chemical-based preservations like use of organic acids or natural extracts to enhance quality and shelf life and microbial-based products such as bacteriosins, useful bacterial cultures as in bio-preservation. In addition, hurdle technology, in which combination of different technologies are used to exert preservation effects (altered atmosphere packaging, high pressure processing, pulsed light preservation, pulsed electric field, ultrasound, ozone and microwave processing technologies). In spite of the existence of these technologies, its adaptation across the globe is not uniform resulting in huge postharvest loss of food across the sector. In addition to these processing and preservation technologies, novel packaging technologies are to be adopted to maintain the quality and extending the shelf life. Reduced oxygen atmosphere is one of the important novel packaging techniques which finds application in food and other systems. This can be achieved by adopting following methods. (a) (b) (c) (d)

Use of high barrier films Vacuum packaging Modified atmosphere packaging Oxygen scavenger

High Barrier Films Selection of appropriate packaging material depending on the product characteristics and its intended use and shelf life is very essential. Packaging material used should contain, protect, and preserve food products apart from other functions like presenting, convenience, communication, how to use, traceability, tamper indication, and providing storage history. Earthen pots, paper, glass, metal, and petroleum-based packaging materials are commonly used food packaging materials. Since ages, earthen pots are in use for short-term storage and preservation of food products mainly grains and other dried food items. In fish, earthen pots find its special place mainly in the preparation of traditional fermented fish products. These are ethnic food products in which fish with appropriate ingredients like salt are packed and sealed and kept for fermentation time ranging from few months to a year. Paper is another natural-based packaging material commonly used for packaging including food products. As paper-based material can easily absorb moisture from food and its surrounding appropriate treatment like wax coating, providing thin liner using flexible packaging is commonly practiced for its food applications. Glass is made from silica, is inert in nature, and is best suited for food applications. It can be prepared in varied shapes and sizes with attractive color. Metal containers like tin, stainless steel, tin-free steel, aluminum are commonly used for food packaging applications after providing an inner lacquer layer to prevent its direct contact with food. Being fragile and heavy, its use in food packaging including fish packaging is very limited. Petroleum-based plastic packaging materials are being commonly used for packaging applications of food. These are made from petroleum, a nonrenewable resource extracted and processed using very high energy-intensive techniques.

2108

C. O Mohan et al.

Because of its unique properties like light weight, adjustable barrier properties using combination of plastic materials, and its cost, these are very commonly used in fish packaging applications. These petroleum-based materials are either polyethylene terephthalate (PET), high-density polyethylene (HDPE), polyvinyl chloride (PVC), low-density polyethylene (LDPE), polypropylene (PP), polystyrene (PS), or others, which are numbered serially from 1 to 7 for easy identification. Other category may include combination of many layers or multilayered packaging materials which contains more than one type of plastic materials including food can liners, compact discs, computers, cell phones, sip-up bottles, etc. Among these, HDPE, LDPE, and PP are considered as somewhat safe compared to other packaging materials and can be used for direct food contact applications. Apart from these, many petroleum-based packaging materials are available in the market for their food application either as direct or indirect contact and these include: (a) polyolefins (LDPE, linear LDPE, HDPE, biaxially oriented polypropylene (BOPP)), (b) copolymers of ethylene (ethylene vinyl acetate (EVA), ethylene-vinyl alcohol (EVOH), ethylene-acrylic acid (EAA)), (c) substituted olefins (polystyrene, highimpact polystyrene, oriented PS, poly-vinyl alcohol (PVOH), poly-vinyl chloride (PVC), poly-vinylidene chloride (PVdC), poly-tetrafluoroethylene (PTFE)), (d) polyesters (PET, polyethylene naphthalate (PEN), relative copolymer PET-PEN), (e) polycarbonates, (f) polyamide (PA) or nylon, (g) acrylonitrile like polyacrylonitrile (PAN) and acrylonitrile/styrene (ANS), (h) regenerated cellulose, and (i) polylactic acid (PLA). The arrangement of molecules in these plastic packaging materials decides its permeability to gases and water vapor. Thickness, temperature, and relative humidity affect the permeability of packaging materials. Of the different physical properties of packaging materials, gas permeability, water vapor permeability, and migration of residues decide its applications. For applications in food products, the packaging material should comply with the requirements for migration of residues and many countries have stipulated the acceptable limit as 60 mg L−1 or 10 mg dm−2 . Selection of simulants and testing conditions depends on the properties of food intended to be packed. Water vapor permeability is another important property which decides the application of packaging materials especially for moisture-sensitive products. The gas permeability of the packaging material is an important attribute for the selection of packaging material. It is well known that the fat oxidation can take place even at low temperature, and the oxygen-sensitive foods need to be packed in appropriate packaging material with low gas permeability. Fat oxidation results in the formation of peroxides and aldehydes. Peroxides result in off-odor and can be toxic to consumers. Aldehydes affect the texture of food. Apart from this, if fish is packed in packaging material with poor gas barrier properties, it may lead to growth of aerobic bacteria, fungi, and molds resulting in inferior quality. This also results in flavor and color changes due to oxidation of fat molecules and pigments. The gas permeability of commonly used single and laminated packaging materials is given in Table 1. Table 1 clearly indicates the OTR and WVTR of packaging materials vary with the thickness and the polymer materials used. Among the packaging materials,

82 Smart Packaging: O2 Scavenger for Improving Quality of Fish

2109

Table 1 Oxygen and water vapor transmission rate of packaging materials Packaging material Polyester-LDPE laminate Polyester-aluminum foil-cast PP Nylon-EVOH-polyethylene PEST-PE laminate High impact Polypropylene tray EVOH PVDC PE PP Nylon PET

Thickness (μm) 87 110 138 88 880 15 25 50 50 25 12

Oxygen transmission rate 65 10.15 3.86 236.38 13.2 0.2 2 300 29,900 77 64

Water vapor transmission rate 3.62 0.22 139.02 80.96 0.89 NA NA NA NA NA NA

NA – not analyzed OTR as cc/m2 /atmosphere/24 h at room temperature (28–32 ◦ C) WVTR as g/m2 /24 h at 37 ◦ C and 90 ± 2% RH

EVOH and PVDC can be used as high barrier packaging materials for reduced oxygen atmosphere packaging.

Vacuum Packaging Important properties by which consumers judge fish and shellfish products are appearance, texture, and flavor. Appearance, specifically color, is an important quality attribute influencing the consumer’s decision to purchase. In fresh red meat fishes, myoglobin can exist in one of three chemical forms. Deoxymyoglobin, which is purple, is rapidly oxygenated to cherry red oxymyoglobin on exposure to air. Over time, oxymyoglobin is oxidized to metmyoglobin, which results in a brown discoloration associated with a lack of freshness. Low oxygen concentrations favor oxidation of oxymyoglobin to metmyoglobin. Therefore, in order to minimize metmyoglobin formation in fresh red meats, oxygen must be excluded from the packaging environment to below 0.05% or present at saturating levels. Lipid oxidation is another major quality deteriorative process in muscle foods resulting in a variety of breakdown products which produce undesirable off-odors and flavors. Hence, O2 may cause off-flavors (e.g., rancidity as a result of lipid oxidation), color changes (e.g., discoloration of pigments such as carotenoids, oxidation), nutrient losses (e.g., oxidation of vitamin E, β-carotene, ascorbic acid) and accelerates microbial spoilage thereby causing significant reduction in the shelf life of foods. Therefore, control of oxygen levels in food package is important to limit the rate of such deteriorative and spoilage reactions in foods. Oxygen level in the package can be controlled by using the vacuum packaging technique in which the air present in

2110

C. O Mohan et al.

the pack is completely evacuated by applying vacuum and then package is sealed. Vacuum packaging, which is also referred as skin packaging, involves removal of air inside the pack completely and maintaining food material under vacuum conditions, so that oxygen available for the growth of microbes and oxidation will be limited. This will help in doubling the shelf life of fish under chilled conditions. This technique is particularly useful in fatty fishes, where the development of undesirable odor due to the oxidation of fat is the major problem. Vacuum packaging for chilled and refrigerated fishes doubles the shelf life compared to normal air packaging (Mohan et al. 2008). Application of this to frozen fishes is also commonly followed as it helps in reducing problem of freezer burn. This technique can be applied to fresh meat and fishes, processed meat and fishes, cheese, coffee, cut vegetables, etc. One of the important aspects in the vacuum packaging is the use of packaging material with good barrier properties. Normally polyester-polyethylene or nylon-polyethylene laminates are used. Polyester and nylon provides good strength and acts as good barrier to oxygen. Polyethylene proves good heat sealing property and is resistant to water transmission. The advantages of vacuum packaging include reduction in fat oxidation, growth of aerobic microorganisms, reduction in evaporation thereby dryness, and freezer burn in frozen products, extends shelf life, and reduces volume for bulk packs containing lighter materials. Disadvantages include difficulty in use for sensitive crispy products and products with sharp edges, requires high barrier packaging material to maintain vacuum, and creates anaerobic condition, which may trigger the growth and toxin production of Clostridium botulinum and the growth of Listeria monocytogenes. Additional barriers/hurdles are needed to control these microorganisms and also it is capital intensive. Alternative to vacuum packaging, reduced oxygen level in the package can be achieved by using active packaging system like oxygen scavenger. Use of oxygen scavenger is very effective in reducing the oxygen level to 80% within 24 h, respectively (Mohan 2008). Dual action sachet, which combines the scavenging of O2 and emits CO2 , is also developed and found to extend the shelf life of fatty fish up to 25 days (Mohan 2008). These developed active packaging systems follow first-order reaction and extend the shelf life of fishery products significantly (Mohan 2008). Studies indicated that O2 scavenger was very efficient in reducing oxygen concentration by 99.58% within 24 h inside the packages and found to extend the catfish steaks shelf-life up to 20 days, compared to 10 days in control air packs (Mohan et al. 2008). Studies on Seer fish indicated a shelf life extension of 20 days under O2 scavenger compared to only 12 days for air packs and inhibited the formation of biogenic amines, especially histamine by inhibiting bacterial enzyme activity (Mohan et al. 2009a). Use of O2 scavenger positively extended the shelf life by inhibiting the formation of volatile bases and inhibiting the nucleotide degradation resulting in delayed formation of hypoxanthine, which is associated with the spoilage of fish (Mohan et al. 2009b). Use of O2 scavenger improved the shelf life of barracuda steaks by 20 days (Remya et al. 2018) and Indian oil sardine by 15 days (Mohan et al. 2019a) under chilled storage. A delay in the growth of microorganisms including specific spoilage flora like Pseudomonas spp. and H2 S forming bacteria was observed in fishes packed with O2 absorber by extending the lag phase which is mainly due to the effect of altered atmosphere (Mohan et al. 2010). A shelf life of 9–10 days was observed for long tail tuna (Thunnus tonggol) packed under O2 scavenger under chilled stored (Mohan et al. 2014). Active antimicrobial packaging films prepared using chitosan incorporating ginger (Zingiber officinale) essential oil (GEO) were effective against food-borne pathogens (Remya et al. 2015). Keeping quality of steaks of barracuda (Sphyraena jello) fish improved significantly in the chitosan films with GEO (Remya et al. 2015). Antimicrobial packaging film incorporating silver nanoparticles synthesized using low and high molecular weight and other chemicals can be used effectively to control the growth of food-borne pathogens (Kishore et al. 2018). Combination of O2 scavenger and antimicrobial film incorporating essential oil resulted in enhanced quality retention and reduced oxidation and extended the shelf life up to 30 days in chilled storage condition (Remya et al. 2017). Combination of curry leaf essential oil and O2 scavenger resulted in increased lag phase and reduced oxidation in Rachycentron canadum and extended shelf life up to 30 days (Remya et al. 2014). Antimicrobial coating with chitosan resulted in reduced microbial growth, volatile formation, oxidation, drip loss, and improved water holding capacity and improved the texture of Indian oil sardine (Mohan et al. 2012; Renuka et al. 2016). The formation of total volatile base nitrogen and trimethylamine nitrogen was less by 14.9–32.7 and 26.1–49% for different concentration of chitosan-treated samples (Mohan et al. 2012). Biodegradable antioxidant packaging film developed using rosemary essential oil resulted in improved DPPH activity and total phenolic content (Mohan et al. 2018).

82 Smart Packaging: O2 Scavenger for Improving Quality of Fish

2119

Summary Smart packaging is gaining increased attention as they find applications in many fields. Among the smart packaging techniques, O2 scavenger has increased role to play as it helps in achieving enhanced shelf life of perishable commodity. These active smart packaging techniques are cost effective compared to vacuum and MAP and are equally or more efficient. The advantage of cost of this technology will pave way to its increased adoption in coming days. Oxygen scavengers are very effective in reducing the residual oxygen level to less than 0.01% within 24 h and maintain it throughout the storage life which cannot be achieved in vacuum packaging and MAP. Simple to use and low cost make O2 scavenger one of the highly sought after smart packaging technology in the coming days.

References Ahvenainen R (2003) Active and intelligent packaging: an introduction. In: Ahvenainen R (ed) Novel food packaging techniques. Woodhead Publishing Ltd., Cambridge, UK, pp 5–21 Biji KB, Ravishankar CN, Mohan CO, Srinivasa Gopal TK (2015) Smart packaging systems for food applications: a review. J Food Sci Technol 52(10):6125–6135 Brody AL, Budny JA (1995) Enzymes as active packaging agents. In: Rooney ML (ed) Active food packaging. Blackie Academic Professional, London, pp 174–192 Daniels JA, Krishnamurthi R, Rizvi SSH (1985) A review of effects of carbon dioxide on microbial growth and food quality. J Food Prot 48:532–537 Dixon NM, Kell DB (1989) The inhibition of CO2 on the growth and metabolism of microorganisms. J Appl Bacteriol 67:109–136 FAO (2018) The State of World Fisheries and Aquaculture 2018 – meeting the sustainable development goals. FAO, Rome, p 227 FAO (2020) The State of World Fisheries and Aquaculture 2020. Sustainability in action. Rome, p 224. https://doi.org/10.4060/ca9229en Farber JM (1991) Microbiological aspects of modified atmosphere packaging technology- a review. J Food Prot 54:58–70 Floros JD, Dock LL, Han JH (1997) Active packaging technologies and applications. Food Cosmet Drug Packag 20:10–17 Heising JK, Dekker M, Bartels PV, Van Boekel MA (2014) Monitoring the quality of perishable foods: opportunities for intelligent packaging. Crit Rev Food Sci Nutr 54(5):645–654 https://www.reportlinker.com/p05881809/?utm_source=PRN Kerry JP, O’Grady MN, Hogan SA (2006) Past, current and potential utilisation of active and intelligent packaging systems for meat and muscle-based products: a review. Meat Sci 74: 113–130 Kishore P, Panda SK, Minimol VA, Mohan CO, Ravishankar CN (2018) Nanotechnology for pathogen detection and food safety. In: Mohan CO, Carvajal-Millan E, Ravishankar CN (eds) Research methodology in food sciences: integrated theory and practice. CRC & Apple Academic Press, Oakville, pp 3–18 Marsh K, Bugusu B (2007) Food packaging: roles, materials, and environmental issues. J Food Sci 72:39–55 Miltz J, Passy N, Mannheim CH (1997) Trends and applications of active packaging systems. In: Ackerman P, Jagerstad M, Ohlsson T (eds) Foods and packaging materials: chemical interactions. Royal Society of Chemistry, Milton Road, Cambridge, pp 201–210

2120

C. O Mohan et al.

Mohan CO (2008) Shelf life extension of seer fish (Scomberomorus commerson) steaks using O2 scavenger and CO2 emitters in chilled condition. Thesis submitted to ICAR-CIFE, Mumbai, India, p 248 Mohan CO, Ravishankar CN, Srinivasa Gopal TK (2008) Effect of O2 – scavenger on the shelf-life of catfish (Pangasius sutchi) steaks during chilled storage. J Sci Food Agric 88:442–448 Mohan CO, Ravishankar CN, Srinivasa Gopal TK, Ashok Kumar K (2009a) Nucleotide breakdown products of seer fish (Scomberomorus commerson) steaks stored in O2 scavenger packs during chilled storage. Innov Food Sci Emerg Technol 10:272–278 Mohan CO, Ravishankar CN, Srinivasa Gopal TK, Ashok Kumar K, Lalitha KV (2009b) Biogenic amines formation in Seer fish (Scomberomorus commerson) steaks packed with O2 -scavenger during chilled storage. Food Res Int 42:411–416 Mohan CO, Ravishankar CN, Srinivasa Gopal TK (2010) Active packaging of fishery products- a review. Fish Technol 47(1):1–18 Mohan CO, Ravishankar CN, Srinivasa Gopal TK, Lalitha KV (2012) Effect of chitosan edible coating on the quality of double filleted Indian oil sardine (Sardinella longiceps) during chilled storage. Food Hydrocoll 26(1):167–174 Mohan CO, Ravishankar CN, Srinivasa Gopal TK (2014) Quality of Longtail tuna (Thunnus tonggol) chunks packed under reduced oxygen atmosphere in chilled storage condition. In: International Symposium “MECOS-2 Marine Ecosystems - Challenges and Opportunities.” The Marine Biological Association of India at Cochin during 2–5 Dec 2014 Mohan CO, Ashitha VA, Kishore P, Panda SK, Ravishankar CN (2018) Influence of mono- and multilayered packaging material on the quality of Seer fish (Scomberomorus commerson) during chilled storage. J Packag Technol Res 2(1):67–76 Mohan CO, Abin J, Kishore P, Panda SK, Ravishankar CN (2019a) Effect of vacuum and active packaging on the biochemical and microbial quality of Indian Oil Sardine (Sardinella longiceps) during iced storage. J Packag Technol Res 1–13 Mohan CO, Gunasekaran S, Ravishankar CN (2019b) Chitosan-capped gold nanoparticles for indicating temperature abuse in frozen stored products. npj Sci Food 3(2):1–6 Moleyar V, Narasimham P (1994) Modified atmosphere packaging of vegetables: an appraisal. J Food Sci Technol 31(4):267–278 Paine FA (1991) The packaging users handbook. AVI, Van Nostrand Reinhold, New York Parkin KL, Brown WD (1982) Preservation of seafood with modified atmospheres. In: Martin RE, Flick GJ, Hebard CE, Ward DR (eds) Chemistry and biochemistry of marine food products. AVI Publishing Co., Westport, pp 453–465 Parry RT (1993) Introduction. In: Parry RT (ed) Principles and applications of MAP of foods. Blackie Academic and Professional, New York, pp 1–18 Phillips CA (1996) Review: modified atmosphere packaging and its effects on the microbiological quality and safety of produce. Int J Food Sci Technol 31:463–479 Realini CE, Marcos B (2014) Active and intelligent packaging systems for a modern society. Meat Sci 98:404–419 Remya S, Basu S, Venkateshwarlu G, Mohan CO (2014) Quality of shrimp analogue product as affected by addition of modified potato starch. J Food Sci Technol 52(7):4432–4440 Remya S, Mohan CO, Revishankar CN, Gopal TKS, Venkateshwarlu G (2015) WANT TO KEEP YOUR FISH FRESH LONGER? Adopt Reduced Oxygen Packaging (ROP) Technologies. Fishing Chimes 34(11):49–52 Remya S, Mohan CO, Bindu J, Sivaraman GK, Venkateshwarlu G, Ravishankar CN (2016) Effect of chitosan based active packaging film on the keeping quality of chilled stored barracuda fish. J Food Sci Technol 53(1):685–693 Remya S, Mohan CO, Venkateshwarlu G, Sivaraman GK, Ravishankar CN (2017) Combined effect of O2 scavenger and antimicrobial film on shelf life of fresh cobia (Rachycentron canadum) fish steaks stored at 2◦ C. Food Control 71:71–78 Remya S, Mohan CO, Ravishankar CN, Sivaraman GK, Jha AK, Venkateshwarlu G (2018) Effect of active packaging atmosphere on the shelf life of chilled stored steaks of barracuda Sphyraena jello. Indian J Fish 65(4):109–115

82 Smart Packaging: O2 Scavenger for Improving Quality of Fish

2121

Remya S, Mohan CO, Ravishankar CN (2020) Oxygen scavenger packaging for seafood preservation. Fish Technol 57:147–155 Renuka V, Mohan CO, Kirplani Y, Sivaraman GK, Ravishankar CN (2016) Effect of chitosan edible coating on the microbial quality of Ribbonfish, Lepturacanthus savala (Cuvier, 1929) steaks. Fish Technol 53(2):146–150 Restuccia D, Spizzirri UG, Parisi O, Cirillo G, CurcioM IF, Puoci F, Vinci G, Picci N (2010) New EU regulation aspects and global market of active and intelligent packaging for food industry applications. Food Control 21:1425–1435 Robertson G (2006) Food packaging principles and practices. Taylor & Francis, Boca Raton Rooney ML (1985) Oxygen scavenging from air in package headspaces by singlet oxygen reactions in polymer media. J Food Sci 47(291–294):298 Rooney ML (1992) Reactive packaging materials for food preservation. In: Proceedings of the First Japan–Australia workshop on Food Processing, Tsukuba, Japan, pp 78–82 Rooney ML (1995a) Active packaging in polymer films. In: Rooney ML (ed) Active food packaging. Blackie Academic and Professional, London, pp 74–110 Rooney ML (1995b) Overview of active food packaging. In: Rooney ML (ed) Active food packaging. Blackie Academic and Professional, London, pp 1–37 Sivertsvik M (1997) Active packaging of seafood – an evaluation of the use of oxygen absorbers in packages of different seafood products. Paper presented at 27th WEFTA Conference, October 19–22, 1997, Instituto del frio (CSIC), Madrid, Spain Sivertsvik M (2007) Lessons from other commodities: fish and meat. In: Wilson CL (ed) Intelligent and active packaging for fruits and vegetables. CRC Press, Boca Raton/London, pp 151–161 Smith JP, Ramaswamy HS, Simpson BK (1990) Developments in food packaging technology. Part 2: Storage aspects. Trends Food Sci Technol 11:111–118 Vanderroost M, Ragaert P, Devlieghere F, Meulenaer BD (2014) Intelligent foodpackaging: the next generation. Trends Food Sci Technol 39:47–62 Wang Y-C, Mohan CO, Jiehao G, Ravishankar CN, Gunasekaran S (2018) Chitosan and gold nanoparticles-based thermal history indicators and frozen indicators for perishable and temperature-sensitive products. Food Control 85:186–193 Yam KL, Takhistov PT, Miltz J (2005) Intelligent packaging: concepts and applications. J Food Sci 70:1–10

Web References https://en.wikipedia.org/wiki/Oxygen_scavenger; https://www.sciencedirect.com/topics/agricultu ral-and-biological-sciences/oxygen-scavenger https://www.intechopen.com/books/structure-and-function-of-food-engineering/oxygenscavengers-an-approach-on-food-preservation https://www.saesgetters.com/oxygen-scavenger https://www.sciencedirect.com/science/article/pii/S0956713516302729 https://krishi.icar.gov.in/jspui/bitstream/123456789/20461/2/Effect%20of%20active%20packaging %20atmosphere%20on%20the%20shelf%20life%20of%20chilled%20stored.pdf https://www.researchgate.net/publication/229474306_Effect_of_O2_scavenger_on_the_shelflife_of_catfish_Pangasius_sutchi_steaks_during_chilled_storage

Social Responsibility Diagnostics as the Sustainable Development Basis

83

Iryna Moiseienko, Ivanna Dronyuk, and Igor Moyseyenko

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . National Model Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Features of National Model Formation for Business Social Responsibility . . . . . . . . . . . System Methodology Analysis of Entrepreneurship Social Responsibility . . . . . . . . . . . . Entrepreneurship Social Responsibility Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods for Diagnosing the Parameters of Entrepreneurship Social Responsibility . . . . Recommended Social Monitoring Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2124 2125 2125 2129 2130 2130 2134 2136 2137

Abstract

The new type of economy formation provides for an increasing role of intellectual and social capital in management structures at the macro and micro levels, social entrepreneurship is becoming a strategic factor in the sustainable development of national economies. The formation of a national model of social responsibility of entrepreneurship as a factor of sustainable development requires a deeper theoretical understanding, methodological formalization, and justification of methods of controlling the level of responsibility. The presence of

I. Moiseienko Financial Department, Lviv State University of Internal Affairs, Lviv, Ukraine e-mail: [email protected] I. Dronyuk () ACS Department, LPNU, Lviv, Ukraine e-mail: [email protected] I. Moyseyenko Department of Theoretical Economics, Lviv Trade and Economic University, Lviv, Ukraine © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_74

2123

2124

I. Moiseienko et al.

a variety of methodological approaches to the analysis of the structure, types, and forms of social responsibility of entrepreneurship necessitated the comparison and streamlining of diagnostic and monitoring problems. The entrepreneurship development problems analysis from the social responsibility point of view is important both from the theoretical point of view – finding out the social and economic nature of social entrepreneurship responsibility, and from the practical point of view – the social entrepreneurship development as a sustainable development factor. Formation of the national entrepreneurship social responsibility model as a sustainable development factor requires deep theoretical reflection, methodological formalization, and justification of responsibility level monitoring methods. The purpose of the study is to substantiate the methodological basis for entrepreneurship social responsibility analysis as a condition for monitoring Ukraine sustainable development. Keywords

Business processes · Social responsibility of entrepreneurship · National model of social responsibility · Diagnostics · Social monitoring · Monitoring methods · Sustainable development

Introduction In the context of a pandemic and the construction of a socially oriented economy of Ukraine, there is a growing need for management decisions adequate to the needs of social and sustainable development to ensure effective control over their implementation. Accordingly, the need for prompt and reliable information about social responsibility and the state and trends of social processes is growing. The solution to this problem is possible through the use of specific scientific tools and tools for analysis and evaluation of corporate social responsibility. Corporate social responsibility is a set of models, mechanisms, and tools of social policy of the enterprise, ethics, and culture of business, which are certainly part of its intellectual resources (Adamczyk 2009; Moyseyenko 2014). Diagnosis involves the formalization of the processes of data collection and processing with the developed methods and procedures of analysis. Information and analytical support for the diagnosis of social policy of sustainable development should be based on an extensive system of social indicators, which would comprehensively reflect the objective socioeconomic processes and the level of use of resource potential. Problems of research and analysis of the parameters of social responsibility of entrepreneurship are as follows: Narrow discipline and fragmentary research in the field of corporate social responsibility. Lack of methodologies for systematic analysis of social policy in terms of the principles of security theory, the theory of sustainable development, and the theory of resource potential.

83 Social Responsibility Diagnostics as the Sustainable Development Basis

2125

Imperfection of methods for assessing the level of social capital and social responsibility, the choice of priorities and parameters for assessing transformational change. Lack of information and analytical support for the transparency of local governments, management decisions, and public participation in the regulation of social responsibility. Methodical problems of the analysis of a condition of social responsibility of business include: (1) the derivation of generalized indexes for each major structural unit of social responsibility of participants in social processes, (2) the identification of representative indicators representing each structural component, and (3) the selection of a set of probable indicators of development, those that are most closely related to solving current social problems and achieving certain social goals of sustainable development (Volkmann et al. 2019; Trunov 2015). The purpose of this study is a theoretical and scientific-applied substantiation of the basics of diagnostics of social parameters as a component of state regulation of social responsibility of entrepreneurship. Achieving this goal involved solving the following tasks: – Describe the importance of social diagnosis of development – Propose to public authorities and local governments – Methodical bases of social diagnostics taking into account international indicators of social development, indicators of social and labor relations and components of efficiency of social policy Analysis of research on the functioning of business entities revealed the absence of a systematic analysis of social policy in terms of the principles of the theory of sustainable development and the theory of use of resource potential; typologies of social responsibility of entrepreneurship by levels of responsibility and levels of its implementation; methodological support for making management decisions by levels of management and levels of responsibility (Volkmann et al. 2019; Godelnik 2012; Porter and Kramer 2006; Gogula and Kudinova 2011; Introduction to CSR 2021; ISO 26000 2010; Moyseyenko 2014; State Fiscal Service of Ukraine 2020; Moiseienko and Moyseyenko 2016). This investigation is an extended version the thesis (Moiseienko et al. 2020).

National Model Formation Features of National Model Formation for Business Social Responsibility In modern economics, researchers mostly pay attention to certain components of social responsibility, such as corporate practices, human capital, intellectual assets, and social capital. At the same time, the factors of formation of social responsibility of entrepreneurship by levels of management and levels of responsibility, their

2126

I. Moiseienko et al.

impact on the processes of the macroeconomic environment, and sustainable development of Ukraine remain insufficiently studied (State Statistics Service of Ukraine 2018; The concept of the business 2020; Moiseenko and Moiseenko 2021). The strategies of socioeconomic development at the state and regional levels do not provide for measures to use socially responsible business activities of all sizes: large, medium, small, and micro. Thus, for the formation of a national model of social responsibility of entrepreneurship, it is necessary to introduce social monitoring into economic practice by levels of government (macro, meso, micro) and levels of responsibility to the state, staff, society (ISO 26000 2010; State Statistics Service of Ukraine 2018). Implementation of the national model of corporate social responsibility involves: achieving a unified understanding of the principles of business social responsibility by participants in social processes and their coordination; formation of mechanisms for managing the level of social activity at different levels of management at the macro, meso, and micro levels; acceptance by the state of responsibility for business activities; and promotion of positive impact in the framework of its activities on the environment, consumers, workers, communities, and all other stakeholders (Kutsyk and Moyseyenko 2019; Leonard 2019). Entrepreneurs should integrate the developed principles of social responsibility into business plans, which will help increase awareness of the values and benefits of such a process, including in the context of best international practices and international documents. Aspects of the Social Responsibility Study are presented in Table 1. Thus in economic practice the high level of differentiation of factors of formation of social responsibility is observed. The forms of their manifestation are presented Table 2. These factors determine the relevance of theoretical and analytical research on sustainable development, taking into account the role and importance of corporate social responsibility, the formation of a national model of social responsibility. Corporate social responsibility is a set of models, mechanisms, and tools of social policy, which has a multilevel structure and consists of three main levels. The basic level provides for the fulfillment of the following obligations: timely payment of taxes, timely payment of wages appropriate for the restoration of the level, and the provision of new jobs. The second level is to provide employees with adequate working and living conditions: staff training, preventive treatment, housing construction, and development of social infrastructure (corporate social responsibility). The third level of responsibility includes the three previous levels and involves the implementation of charitable, sponsorship, and philanthropic activities (Volkmann et al. 2019). Currently, economic practice considers five basic principles of corporate social responsibility: business responsibility to the consumer, which consists in fair pricing, quality of goods and services, care for the health and safety of consumers, fair competition and advertising, compliance with ethical standards doing business; social protection of employees of enterprises: labor rights and decent remuneration for work, labor protection, safety and health at work, staff development and support; attitude to the environment: ecological safety of production, economical

83 Social Responsibility Diagnostics as the Sustainable Development Basis

2127

Table 1 Aspects of the social responsibility study Source The National Strategy Concept of Corporate Social Responsibility in Ukraine

EU Green Paper on Corporate Social Responsibility

ISO 26000 “Guidelines for social responsibility”

Concept Social responsibility is responsibility for the decisions impact and actions on society, the environment through transparent and ethical behavior, which promotes sustainable development, including health and well-being of society; takes into account the expectations of stakeholders; complies with current legislation and international standards of conduct; integrated into the activities of the organization and practiced in its relations. Corporate Social Responsibility is a concept whereby companies integrate social and environmental issues into their business and engage with stakeholders on a voluntary basis. Social responsibility is the responsibility of the organization for the impact of its decisions and activities (namely products and services) on society and the environment, implemented through transparent and ethical behavior, consistent with the sustainable development and well-being of society, taking into account stakeholder expectations. Does not contradict the relevant legislation and international norms of conduct

Developed by authors with data from (ISO 26000 2010; The concept of the business 2020; Commission of the European Communities 2002)

consumption of natural resources and their reuse, waste disposal; partnership with the local community and government is also a prerequisite for the participation of business in the socioeconomic development of the region in areas affecting the core activities of companies, including charitable programs of social and cultural development, support for community initiatives and civil society institutions; attitude to human rights (Godelnik 2012; Gogula and Kudinova 2011; Introduction to CSR 2021). The variety of tools for the formation of the national model of social responsibility is reduced to the following types: information and communication (data collection, information, diagnosis, monitoring, counseling); financial (economic and tax levers of influence); structural and functional (classification of elements and tools of the social responsibility system); organizational (adoption of the concept, development of the strategy, state target programs of social responsibility, social programs of business entities, etc.) (Moyseyenko 2014; Moiseienko and Moyseyenko 2016). To solve the problem of implementing the institutional foundations of the national model of regulation of social responsibility of entrepreneurship, the types and levels of implementation of social responsibility by levels of management and levels of responsibility (Moiseienko and Moyseyenko 2016). It is proposed to regulate the social responsibility of entrepreneurship on the basis of the principles: a combination of different forms and manifestations of social responsibility of

2128

I. Moiseienko et al.

Table 2 Factors in the formation of entrepreneurship social responsibility Directions Exacerbation of competition in national and international markets The growing role of business structures Exacerbation of global, in particular environmental, problems Finding new ways to fight for the consumer The growing role of qualified personnel Increasing the share of intangible assets and intellectual capital Development of consumerism

Legislative activity Growing activity and demanding civil society Activities of international organizations Social policy at the state and regional levels

Forms of manifestation Globalization, development of TNCs, formation of regional integration associations, reduction, and elimination of customs barriers Increase in size, resources, impact on social processes Environmental pollution, climate change, deforestation socio-ethical marketing, relationship marketing “Attractive social face” In particular image and brands in the capital structure The struggle of consumers for their own rights, the growth of product quality standards, the formation of mechanisms of responsibility for violating the rights of the buyer Standards of corporate governance, reporting, environmental and social behavior Nongovernmental organizations, local communities, public opinion on ethical and social aspects of business UN concepts and programs, IBC Quality of life, social standards

Developed by authors with data from (Gogula and Kudinova 2011; ISO 26000 2010)

entrepreneurship; taking into account the interests of the Ukrainian state, business and civil society; the need to develop human, intellectual, and social capital as objects of social responsibility. It is recommended to develop the national model of regulation on a combination of formal and informal institutions, dialogical concept of social responsibility of entrepreneurship (the state forms institutional support and stimulation, entrepreneurship realized social responsibility by levels of management and levels of responsibility), socially oriented approach to implementation of principles of social responsibility of entrepreneurship, viscosity of the implementation of social responsibility and implementation of social reporting (Moiseienko and Moyseyenko 2016).

System Methodology Analysis of Entrepreneurship Social Responsibility In the knowledge economy, the system methodology for research of entrepreneurship social responsibility requires the development of tools for analysis of the use of intellectual and social capital, their potentials and criteria for effective management.

83 Social Responsibility Diagnostics as the Sustainable Development Basis

2129

According to the levels of management and levels of responsibility, the social responsibility of business entities is proposed to be studied as a set of relationships in the triad: “state – business – staff” in order to realize economic, social, environmental effects and use potentials (Moiseienko and Moyseyenko 2016). As a systemic concept of entrepreneurship social responsibility as a set of objects and subjects of social responsibility, the relationship between them can be described as follows: SV P = Fl t (O, S, N, V , Z)

(1)

where O is objects of entrepreneurship social responsibility at the following stages: design, production, consumption, recycling; S is social responsibility subjects (stakeholders); N is directions of social responsibility realization on external and internal environment; V is type of social responsibility economic: legal, moral, philanthropic; t is temporary conditions; l is economic, social, environmental goal; Z is social responsibility level: 1 is with the state, 2 is with the personnel, 3 is with the society. The state regulation of social responsibility of entrepreneurship will be considered by levels of responsibility and levels of management of types of responsibility and in terms of the size of business entities. The formalized record of diagnostics and social monitoring of the parameters of the realized responsibility will look like this   KA = ∪ RUm , RVi , RS kj , KVl , T P p , V F iv ,

(2)

where RUm is control level: m = 1 is macro, m = 2 is meso, m = 3 is micro; RVi is social responsibility level; i = 1 is to the state, i = 2 is to the staff, i = 3 is to the society; RSj k is type of business entity: j is size of enterprises: j = 1 is large, j = 2 is medium, j = 3 is small, j = 4 is micro; k is level of use of the social responsibility concept: k = 1 is not used, k = 2 is partially used, k = 3 is used (corporate or social enterprises); KVl is regulatory and ethical responsibility criteria in the following areas: l = 1 is economic, l = 2 is social, l = 3 is environmental; TPp is types of potentials p = 1 is intellectual, p = 2 is human, p = 3 is market, p = 4 is organizational and information, p = 5 is creative, p = 6 is natural resource; VFv i is types of financial responsibility; v is types of financial responsibility v = 1 is tax responsibility, v = 2 is wage responsibility to staff, v = 3 is consumer responsibility at price (Moyseyenko 2014). Thus, the diagnosis of social responsibility of entrepreneurship can be considered as one of the main parts of state regulation of social processes, which characterizes the mutual responsibility of subjects and objects of responsibility in the use of elements of their intellectual, social, market, human, and emotional potential.

2130

I. Moiseienko et al.

Diagnosis of the parameters of social responsibility of entrepreneurship as a component of state regulation of the social environment performs a number of functions in the system of public administration, in particular: • Study of the situation, identification of existing problems and trends, assessment of social development • Measuring the results of economic and social policy in the field of entrepreneurship • Measuring the effectiveness of business responsibility • Providing information base for the process of formation and implementation of social entrepreneurship policy • The use of a system of social indicators to determine the prospects of dynamics and create scenarios for the development of society (Adamczyk 2009; Godelnik 2012; Moyseyenko 2014, 2017) Thus, diagnostics of the social environment on the basis of monitoring the indicators of corporate social responsibility is a functional component of the mechanism of public administration, which is a tool for measuring policy effectiveness, finding out the causes and possible consequences of existing trends, forecasting important social processes of sustainable development.

Entrepreneurship Social Responsibility Diagnostics Methods for Diagnosing the Parameters of Entrepreneurship Social Responsibility The model of analysis of social responsibility of entrepreneurship consists of a system of macroeconomic, socioeconomic and social indicators. They demonstrate a certain state of social responsibility by levels of government and levels of responsibility, which is due to social policy and resource potential. Systematic analysis of social processes involves the identification of the composition of key indicators and their relationship. The main economic indicators include the growth rate of gross domestic product, production volumes and efficiencies, inflation, etc. Socioeconomic indicators are the unemployment rate, the percentage of people living below the poverty line, the ratio of average wages to the value of the consumer basket, and so on. Social indicators in terms of international rankings include HDI, the Humanitarian Development Index, the Human Poverty Index, and the Poverty Index; integrated indicator of socioeconomic protection of the population; welfare index (Porter and Kramer 2006; Gogula and Kudinova 2011; Moiseenko and Moiseenko 2021; Trunov 2015), etc. Indicators should be built on a single conceptual basis, integrated into a single data bank within an information system, and correlated with international standards. The system of diagnostic indicators should both reflect the degree of achievement

83 Social Responsibility Diagnostics as the Sustainable Development Basis

2131

of the strategic goals of the region and characterize the main factors of positive or negative changes. • The number of indicators should be quite limited, but fully characterize the state of social responsibility by levels of management and levels of responsibility in terms of the size of enterprises. • Indicators should cover the most important characteristics of the sphere of social policy in the country or region; indicators should provide representativeness for international and interregional comparisons. • Statistics on selected indicators should be objective, reliable, accurate, accessible, and collected without unnecessary costs and additional efforts. • Indicators should not duplicate or at least partially overlap the directions of the social sphere, that is, they should be self-sufficient. • Indicators of social diagnostics should belong to the competence of the relevant body of state executive power, which is responsible for a certain area of social policy, as the core of domestic policy (Porter and Kramer 2006; Moiseienko and Moyseyenko 2016). The list of indicators is formed according to the principles: Representativeness (includes the most significant indicators that affect the level of social security of the state) Reliability (adequately reflect the state of the security component) Information accessibility (official data of the State Statistics Service, state bodies, and public expert assessments are used during the calculation) Criteria and indicators are especially important in social diagnostics. In the scientific literature and statistical practice for international and interregional comparisons use different approaches to the construction of integrated and unit indicators. Each indicator (criterion) is represented by a system of indicators that can be defined and presented with varying degrees of detail. Criteria highlight the direction of management, and indicators record the level achieved. Criteria and indicators are differentiated by spheres of life: economic, social, political, etc. Diagnostic criterion – a specific feature on the basis of which the assessment or classification is performed, the significance or insignificance of the state of the object is determined (The concept of the business 2020; Ojo 2016; Moiseenko and Moiseenko 2021; Trunov 2015). When diagnosing the social development of a country or region, it is advisable to use indicators of corporate social responsibility. Indicators of social responsibility of entrepreneurship as part of the system of social indicators are statistical indicators that are based on observations, monitoring and directly reflect the quantitative aspects of social processes and are closely related to their qualitative state. To determine the causal factors and possible policy solutions in the field of entrepreneurship, a realistic view of the assessment of potential, analysis of similarities and differences of regions is needed. The regional index of social responsibility

2132

I. Moiseienko et al.

of entrepreneurship should be not only more detailed, but also adapted to the realities of a particular region, given the existing problems and data (Introduction to CSR 2021; Moyseyenko 2014, 2017). At the conceptual level, the proposed method provides for a systematic analysis of the social development of the region in terms of social responsibility of entrepreneurship. The upper position is an integrated index of social development, the lower generalizing indicators of regional components. The method of analysis of indicators of social responsibility of entrepreneurship in the region involves the implementation of such stages. 1. The choice of a system of indicators of social responsibility of entrepreneurship in the region. 2. Reducing the number of indicators based on priorities and evaluation criteria. The task of this stage is to reduce the number of indicators and obtain an array of data that most informatively describes the state of the social environment of the regions. Such an analysis can include both the evaluation of graphical data and regression-correlation analysis of the impact of selected indicators on the macro parameters of sustainable development. 3. Calculation of the integrated indicator of the level of social responsibility of entrepreneurship of the regions. It is performed using the methods of taxonomic analysis. The taxonomic indicator is a synthetic quantity. This indicator makes it possible to assess the level of development of each of the regions compared to others (Moyseyenko 2017; State Statistics Service of Ukraine 2018). Regression analysis of the social investment impact on macro indicators with linear and parabolic models is presented on Figs. 1 and 2. Here GRR means gross regional product. 4. Analysis of the obtained data. The final one is the formation of general conclusions on the state of social responsibility by levels of government and levels of responsibility of Ukraine in order to develop appropriate measures of state regulation (Moyseyenko 2014, 2017).

Fig. 1 Analysis of the social investment impact on macro indicators, linear model (Moyseyenko 2017; State Statistics Service of Ukraine 2018)

83 Social Responsibility Diagnostics as the Sustainable Development Basis

2133

Fig. 2 Analysis of the social investment impact on macro indicators, parabolic model (Moyseyenko 2017; State Statistics Service of Ukraine 2018)

Table 3 Dependence of GDP on tax liability of entrepreneurship in Ukraine

Factor Corporate income tax, UAH billion Single tax for small businesses, UAH billion Single tax on legal entities, UAH billion

Equation and calculated valuesStudent’s t-test for regression coefficients y˜ = 86.8 + 25.5x 0.19 2.39** y˜ = 638.4 + 128.8x 5.88* 5.80* y˜ = 340.7+ +703.4x 2.14 5.52*

Correlation coefficient 0.624

Determination Fisher’s coefficient F-test 0.389 5.73**

DW-test DurbinWatson’s test 0.45

0.888

0.789

33.67*

0.51

0.879

0.772

30.45*

0.55

Developed by authors with data from (State Fiscal Service of Ukraine 2020; State Statistics Service of Ukraine 2018)

To clarify the list of diagnostic indicators for indicators of social responsibility of entrepreneurship, it is proposed to build macro-indices of social diagnostics, which, thanks to the available information base, will expand the range of indicators for monitoring social policy. Trends in the level of entrepreneurship taxation according to the general and simplified tax system for the period from 2007 to 2017 have been analyzed, where taxes have the greatest impact on economic growth. The pairwise linear regression equations of GDP dependence on income tax burden in Ukraine show that the highest level of influence on economic performance has the small businesses fiscal responsibility. Regression analysis is presented in Table 3. Detailed analysis of social indicators is a prerequisite for the formation, implementation, or adjustment of economic and social policy. The value of information and analytical support of social policy using the individual indicators described above may be as follows: in the event of a crisis, the definition of “emergency” policies; an appropriate incentive policy is used to stabilize the situation; for

2134

I. Moiseienko et al.

dynamic growth becomes the basis for activities, which is defined as regulatory policy. The application of the methodology provides both the validity of comparisons of social development of individual regions of the country and the dynamics of social development over a period of time for a particular region; allows to make correct comparisons both on an integral index, and on its components - indices of separate aspects of development; helps to identify the most problematic regions.

Recommended Social Monitoring Parameters To overcome the problems of diagnosis and monitoring of social parameters of business responsibility, a method of social monitoring has been developed, which provides for: identification of representative indicators of each structural component by levels of social responsibility (unrealized and realized responsibility by regions) and management levels; determining the dynamics of indicators by types of social responsibility (economic, environmental, social), indicators that have the greatest impact on macroeconomic indicators of sustainable development or shadowing of the economy. According to Ukrainian legislation, the structure of Ukrainian entrepreneurship is divided into large, medium, small, and micro enterprises. To analyze the social responsibility of entrepreneurship to the state, we use a set of indicators and coefficients (Moiseienko et al. 2020). For the entrepreneurship social responsibility analysis to the state we use a complex of indicators and coefficients, characterizing sustainable development factors of stabilizers and destabilizers at the macro level: tax responsibility (tax burden) and unrealized tax responsibility (debts to the budget) are implemented (see Fig. 3). Based on statistical analysis, it is established that the highest level of business tax liability to the state by the size of the subjects have small businesses. To determine the integrated indicator for assessing the level of social responsibility of business entities used a list of indicators that reflect the levels and principles of social responsibility, capacity utilization, cost-effectiveness. The parameters envisaged for sustainable development are reflected in the list of social monitoring indicators (see Table 4). The parameters provided for sustainable development are reflected in the list of indicators of social monitoring (see Table 4). The integrated indicator of the assessment of the business entities social responsibility level (IPSR) is calculated by the next formula  I P SR =

6

I P LP RP OI P KP P RP ∗ ∗ ∗ ∗ ∗ , D D D D D D

(3)

where D is profit of enterprises and other values are from Table 4 (Moiseienko et al. 2020).

83 Social Responsibility Diagnostics as the Sustainable Development Basis

2135

Fig. 3 Analysis of the tax debts dynamics in Ukraine for 2016–2018 years, millions UAH (data from (State Fiscal Service of Ukraine 2020))

Table 4 The reasons of increasing process complexity Potentials Intelligent (IP) Human (LP) Market (RP) Organizational Infrastructure (OIP) Creative (KP) Natural Resources (PRP)

Measurement indicators Expenses on intangible assets Staff costs Selling expenses Administrative expenses Culture costs Expenditures on nature protection

Developed by authors with data from (Moiseienko and Moyseyenko 2016; Moyseyenko 2017)

The calculated integrated indicators of social responsibility of entrepreneurship at the macro level allowed to establish that the highest values of realized social responsibility are demonstrated by medium-sized business entities. The method of assessing the level of social responsibility reflects: the definition of internal and external manifestations of social responsibility of the enterprise; taking into account the results of social and financial reporting of the enterprise; targeting a limited number of indicators. The combination of structural analysis on the choice of objects of social responsibility and quantitative analysis allows to achieve a comprehensive analysis and identify areas for change to increase the level of social responsibility. The application of the methodology of social monitoring provides validity and comparability of the data of the consolidated analysis by levels of management; analysis of the dynamics of indicators and determination of the level of social responsibility by factors stabilizers and destabilizers as realized and unrealized social responsibility; implementation of factor analysis of social responsibility of entrepreneurship by their size to determine the direction of changes

2136

I. Moiseienko et al.

in socioeconomic policy; transparency and validity of decisions state policy on fiscal and social responsibility of entrepreneurship (Moiseienko et al. 2020). The essence of entrepreneurship social responsibility regulation is to establish a set of rules and mechanisms aimed at implementing the principles of entrepreneurship social responsibility in order to harmonize the socioeconomic interests of the parties and ensure the economic achievement, social and environmental development priorities. In our opinion, the conceptual basis for the implementation of the national model of management of entrepreneurship social responsibility provides for the development and approval of the national Concept of Entrepreneurship Social Responsibility, determination of methods, tools, and norms of implementation and mechanisms of regulation. In order to generalize the formation factors of the entrepreneurship social responsibility national model, they have been systematized in the directions and forms of manifestation taking into account modern conditions and transformation processes in the country’s economy (State Fiscal Service of Ukraine 2020). A variety of tools for the formation of a national model of social responsibility is reduced to such kinds of information and communication (data collection, information diagnostics, monitoring, consulting financial) and economic and tax levers of influence of structural and functional classification of elements and tools of the system of social responsibility of organizational (adoption of the Concept of development of the Strategy of State Targeted Programs for Social Responsibility of Business Entities). To solve the problem of introducing the institutional foundations of the national model of regulation of social responsibility of entrepreneurship, the types and levels of implementation of social responsibility were systematized by management and responsibility levels.

Conclusion and Future Work At the present stage of transformation of Ukraine’s economy there is a need for further research aimed at finding ways to overcome existing regulatory problems and increase the efficiency of the corporate social responsibility system to ensure sustainable development. At the forefront are the problems of improving the institutional framework for the formation of a system of entrepreneurship social responsibility at the levels of government in Ukraine. The peculiarities of the formation of the national model of social responsibility of economic entities were studied on the basis of the use of the principles of systems theory and the theory of potentials, institutional foundations, and international practice. The typology of specific manifestations of social responsibility of entrepreneurship is determined on the basis of object models. The state of entrepreneurship social responsibility as a function of the listed types of manifestation, reflecting their temporal, spatial, and other states by size of enterprises and levels of responsibility, is analyzed. Diagnosis of the parameters of social responsibility of entrepreneurship and measuring the level of its implementation is carried out using modified models of economic and regression analysis.

83 Social Responsibility Diagnostics as the Sustainable Development Basis

2137

The level of using for entrepreneurship social responsibility before the state and its impact on macro indicators is calculated, which is defined as the difference between the states in space-time and economic dimensions. This allows from a fundamentally new methodological standpoint to consider the concept, structure, and importance of corporate social responsibility for the effective functioning of the national model of social responsibility in an innovative economy. The tool of research of social responsibility of business and diagnostics of social parameters has scientific novelty. Selected methods of formal modeling of economic processes and complex methods that combine quantitative parameters of measuring capacity and sustainable development can reflect the complexity and hierarchical management decisions. The use of quantitative methods of analysis of social responsibility statistics allows to determine the level of relationships between the factors of social responsibility of entrepreneurship. The stated theoretical provisions of systematic research and analysis of the processes of formation of the national model expand and improve knowledge about the essence and importance of social responsibility of entrepreneurship as a factor of sustainable development. The results of the study will serve as a basis for improving approaches to the implementation of the national model of corporate social responsibility in Ukraine.

References Adamczyk J (2009) Social activity of enterprises. Challenges and barriers of the 21st century economy. Cracow University of Economics, Cracow, pp 65–73 Commission of the European Communities (2002) Corporate social responsibility: a business contribution to sustainable development. https://eur-lex.europa.eu/legal-content/EN/LSU/ ?uri=celex:52002DC0347 Godelnik R (2012) Philanthropy, CSR and the social responsibility of business 2012. https:// www.triplepundit.com/2012/08/philanthropy-csr-social-responsibility-of-business Gogula ±P, Kudinova IP (2011) Business social responsibility: monograph. Lysenko, Nizhyn. 175 p. (In Ukrainian) Introduction to CSR. (2021). http://tutor2u.net/business/strategy/corporate-social-responsibilityintroduction.html ISO 26000: Guidance on social responsibility. (2010). http://www.iso.org/iso/iso_catalogue.htm Kutsyk V, Moyseyenko I (2019) Modeling the parameters of entrepreneurship fiscal responsibility. The potential of modern science, vol 2. Science Publishing, London, pp 117–127 Leonard K (2019) Four types of corporate social responsibility. http://smallbusiness.chron.com/ four-types-corporate-social-responsibility-54662.html Moiseenko IP, Moiseenko IV (2021) Social factors of provision economic security of the state.: collective monograph, vol 2. SPOLOM, Lviv, pp 138–156. (In Ukrainian) Moiseienko IP, Moyseyenko IV (2016) Transformation systemic bases for state regulation of Ukraine economy socialization. In: Lviv State University of internal affairs proceedings. Seria “Economics”. Lviv, 2016, vol 2. pp 56–67. (In Ukrainian) Moiseienko I, Dronyuk I, Moyseyenko I (2020) Responsibility social monitoring as the sustainable development basis. Decision Aid Sciences and Application (DASA), Sakheer, Bahrain, pp 122–126. https://doi.org/10.1109/DASA51403.2020.9317288., https://ieeexplore.ieee.org/xpl/ conhome/9316858/proceeding

2138

I. Moiseienko et al.

Moyseyenko IV (2014) A systematic approach to the social responsibility study. In: Kherson State University proceedings. Seria “Economics”. Kherson, Publishing House “Gelvetika”, vol 9–1, Part. 2, pp 44–47. (In Ukrainian) Moyseyenko IV(2017) Social diagnostics as a basis of social policy. Bulletin of the Volyn Institute of Economics and Management, vol 19. ¥I¨¯, Lutsk, pp 170–178. (In Ukrainian) Ojo M (2016) Designing optimal models of financial regulation in a changing financial environment. 234 p Porter ME, Kramer MR (2006) Strategy and society: the link between competitive advantage and corporate social responsibility. Harvard Bus Rev 12:78–92. https://hbr.org/2006/12/strategyand-society-the-link-between-competitive-advantage-and-corporate-social-responsibility State Fiscal Service of Ukraine. (2020). http://sfs.gov.ua/en/ State Statistics Service of Ukraine. (2018). http://www.ukrstat.gov.ua The concept of the business social responsibility national strategy in Ukraine. (2020). On approval of the Concept for the implementation of state policy in the field of promoting the development of socially responsible business in Ukraine for the period up to 2030. https://zakon.rada.gov.ua/ laws/show/66-2020-%D1%80?lang=en#Text Trunov A (2015) An adequacy criterion in evaluating the effectiveness of a model design process. Eastern-Eur J Enterp Technol 1(4):36–41. https://doi.org/10.15587/1729-4061.2015.37204 Volkmann C, Fichter K, Klofsten M, Audretsch DB (2019) Sustainable entrepreneurial ecosystems: an emerging field of research. Small Bus Econ. https://doi.org/10.1007/s11187-019-00253

Green Nanoparticles: Synthesis and Catalytic Applications

84

Aniruddha B. Patil, Sharwari K. Mengane, and Bhalchandra M. Bhanage

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Green Routes/Techniques for Nanoparticles Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biological Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nanoparticles Synthesis by Botanical Extracts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microwave-Assisted Nanomaterial Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ultrasonic-Assisted Nanomaterial Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrochemical-Assisted Nanomaterial Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Solar Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Natural Fibers: Green Nanomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Catalytic Applications of Green Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Catalytic Applications of Metal Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Catalytic Applications of MMNPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Catalytic Applications of Metal Oxide Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Websites for Further Detailed Understanding . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2140 2142 2142 2145 2146 2147 2149 2149 2151 2154 2156 2160 2162 2163 2165 2165

Abstract

The discovery of green nanoparticles (GN) has fascinated scientific community and has significantly improved in recent years. GN exhibited appealing applications in various fields such as catalysis, energy harvesting, and electrocatalysis.

A. B. Patil Department of Chemistry, Maharshi Dayanand College, Parel, Mumbai, India S. K. Mengane Department of Botany, M. H. Shinde Mahavidyalaya, Tisangi, Kolhapur, India B. M. Bhanage () Department of Chemistry, Institute of Chemical Technology, Matunga, Mumbai, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 C. M. Hussain, P. Di Sia (eds.), Handbook of Smart Materials, Technologies, and Devices, https://doi.org/10.1007/978-3-030-84205-5_75

2139

2140

A. B. Patil et al.

The performance of nanoparticles is interrelated with their singular features, which mainly relied on size and shape. In-depth knowledge of synthetic materials engineering and assembly dynamic is highly beneficial for discovering new protocols for the fabrication of nanoparticles with tunable properties. This chapter is focused on the different synthetic techniques reported for the synthesis of GNs, which mainly cover monometallic nanoparticles, bimetallic nanoparticles, and metal oxide nanoparticles. Different synthesis techniques that demonstrated for the desired performance or performance enhancement of nanomaterials have been discussed in a comprehensive manner. As a main area of GNs, naturally occurring biofibers and their applications in different areas of science and technology have been discussed. The latter part will focus on the catalytic applications of GNs. This section will discuss applications of all above discussed GNs for different catalytic reactions. Furthermore, the merits of GNs over conventional catalytic systems will also be covered. We envision that the collected information about the controlled synthesis of GNs and their catalytic applications for organic transformations will provide a strong platform to the researchers and learners working in the field of green nanomaterials. Keywords

Green nanomaterials · Natural fibers · Nanoscience and nanotechnology · Metal nanoparticles · Catalytic applications

Introduction As a leading branch of science and technology, nanoscience and nanotechnology are rapidly evolving fields which are gaining wide range applicability in science and technology (Patil and Bhanage 2014). Nanoparticles exhibit excellent and programmable features, hence they are considered the hottest topic of research activities at laboratory as well as industrial scale. Though the organic nanomaterials demonstrated a broad range of applications, their preparation is normally relied on the use of hazardous chemicals and derivatives, causing serious environmental issues (Fig. 1). In order to give bypass to increased anthropogenic activities in synthesis of nanoparticles, new types of materials and preparative methods with low environmental impact are urgently required. In this concern, the progress of efficient and sustainable chemical methods for the preparation of metal nanoparticles (MNPs) has become an excellent alternative to the current research dimensions. The term GN mainly comprises preparation of nanoparticles by green methods or the applications of nanosized materials in electronic smart devices (Wu et al. 2019a; Ma et al. 2019), or energy generation (Ma et al. 2019, 2020a) and storage accessories (Ma et al. 2020b). The available biological methods based on series of sustainable,

84 Green Nanoparticles: Synthesis and Catalytic Applications

2141

Fig. 1 Structure performance correlation and applications of green nano particles in the various fields

non-polluting and biodegradable materials such as use of varies plants product like leaf extracts, flower extracts, fruit extracts, seeds. In addition to this, use of algae, bacteria, enzyme, and fungi is also popular and well explored under biological methods of nanoparticle preparation. Use of microwave irradiation is also reported for MNPs synthesis as a green route. Ultrasound-assisted sonochemical method is also well explored for such synthesis. The concentrated solar energy (CSE) method developed by Patil et al. for the preparation of MNPs and metal oxides nanoparticles (MONPs) is also found to be useful for the preparation of green nanoparticles (Patil and Bhanage 2013a). There is another class of nanoparticles that can be considered as green nanoparticles, in which different naturally occurring species are used directly for the

2142

A. B. Patil et al.

assembly of nanoparticles. In this, silk extracted from silkworm, silk spun by spiders, cellulose, and other biomasses are used. Nanoparticles made using these materials were found to be more biocompatible and sustainable. Since 1959 the term “Nanotechnology,” which was introduced by the renowned physicist Richard Feynman, the research area has enormously expanded and has been reviewed meticulously at the recent time. Noteworthy, the purpose of this chapter is to create a thorough understanding about the GN and its catalytic performance. Rather than producing list of experimental or theoretical investigations, we preferably provide detailed discussion on the latest understanding in this field. In this regard, this chapter will cover various green aspects of nanoparticles synthesis, that is, green additives (reductant and capping agents) and green solvents. Afterward, the chapter includes green methods of synthesis of MNPs, MONPs, and multi-metallic nanoparticles (MMNPs). Here, green methods such as biological methods, microwave, ultrasonic-sonochemical, electrochemical, and solar energy have been discussed. The third part of this chapter will cover catalytic applications of green nanoparticles.

Green Routes/Techniques for Nanoparticles Synthesis Wet chemical methods of nanomaterial preparations possess economic and ecological problems and therefore a green chemistry approach has been introduced to overcome these limitations. The green approach to nanomaterial preparation focuses on unconventional green practices that address economic and environmental issues by identifying clean pathways for nanomaterial preparation (Fig. 2). In this regard, green methods such as biological, microwave, ultrasonic, electrochemical, sono-electrochemical, and solar have established their application for nanomaterial synthesis.

Biological Method Among the different green methods of nanoparticles preparation, the synthesis of nanoparticles by biological methods is determined as safe, economic, user friendly, sustainable, and eco-friendly. So far, different MNPs, MONPs, and MMNPs have been prepared by using biosynthesis methods, for instance, silver, gold, copper, platinum, palladium, copper oxide (CuO) nanoparticles, silver oxide (Ag2 O) nanoparticles, gold–silver (Au/Ag), and gold-palladium (Au/Pd) (Maheshwaran et al. 2020).

Nanoparticles by Bacteria Two popular approaches are involved in this class, namely, intracellular and extracellular mechanisms of synthesis of inorganic nanomaterials. In principle, the bacterial system demonstrates metal ion detoxification by either reduction or

84 Green Nanoparticles: Synthesis and Catalytic Applications

2143

Fig. 2 Schematic illustration showing primary essential factors for the green nanoparticles’ synthesis

precipitation. MNPs produced by the extracellular mechanism have more viable applications in various fields, whereas the MNPs obtained using the intracellular mechanism resulted in particles with specific dimensions and low polydispersity. Biosynthesized gold nanoparticles (AuNPs) can be synthetized intracellularly and extracellularly, whereas silver nanoparticles (AgNPs) are significantly prepared extracellularly. The AuNPs are created the day after chlorite precursor is added, whereas AgNPs are obtained 7 days later. Intracellular Preparation of Nanoparticles by Bacteria In this class different bacterial systems have been found for the intracellular fabrication of MNPs and MONPs, some of which are discussed as representatives. Bacillus subtilis 168 is the most popular moiety successfully used to reduce Au(III)

2144

A. B. Patil et al.

to Au(0) possessing octahedral morphology of 5–25 nm (Beveridge and Murray 1980). Furthermore, nature based airborne Bacillus SP. has used as reductant for the synthesis of AgNPs of size 5–15 nm from Ag precursor (Velmurugan et al. 2014). In addition, Co, Cr, and Ni may be incorporated into magnetite crystals prepared by thermophilic iron-reducing bacterium Thermoanaerobacter ethanolicus (TOR-39) (Klaus-Joerger et al. 2001). This process led to the development of octahedralshaped magnetite (Fe3 O4 ) NPs (5 nm is also known (Sheny et al. 2012).

Microwave-Assisted Nanomaterial Synthesis Microwaves (MW) are radio waves with frequencies in between 300 MHz (0.3 GHz) and 300 GHz. The use of developing MW-assisted chemistry method in conjunction with mild reaction medium reduces chemical waste intensity and has gained significant consideration in the organic synthesis and inorganic materials preparation. Principally, quick reaction times and high-throughput effects induced by “hotspot” heating of microwave technique helps for the preparation of different inorganic nanoparticles. In this concern, different research groups reported use of microwave technique for the preparation of various metal and metal oxides nanomaterials. The microwave radiation technology was used to prepare nickel nanoparticle (NiNPs) from a nickel acetate tetrahydrate [Ni(CH3 CO2 )2 .4H2 O] salt and sodium hypophosphite monohydrate (NaPH2 O2 .H2 O) (Fig. 3) (Eluri and Paul 2012). The preparation of PdNPs using short time microwave (MW) irradiation technique is known. In irradiation, PdNPs were obtained from ethanol solution of Pd(OAc)2 in presence of PVP under mild reaction conditions (Galletti et al. 2010). Microwave irradiation technique demonstrated its applicability toward preparation of nanosized metal oxides. Cu2 O nanoparticles (Cu2 ONPs) synthesis using benzyl alcohol under microwave irradiation is also reported using copper acetate precursor using benzyl alcohol under MW irradiation (Bhosale et al. 2013). Additive free preparation of nanosized magnesium hydroxide and magnesium oxide nanoparticles (MgONPs) using MW irradiation technique is also known (Bhatte et al. 2012). The synthesis of ZnONPs under MW radiations is also prepared using same method.

Fig. 3 Schematic illustration showing microwave radiation technology to prepare NiNPs with TEM image of prepared NiNPs. (Reprinted from Eluri and Paul (2012) with permission from Elsevier)

84 Green Nanoparticles: Synthesis and Catalytic Applications

2147

Ultrasonic-Assisted Nanomaterial Synthesis The use of ultrasonic way toward preparation of nanoparticles is well-established area of nanomaterial preparation. Sonochemistry involves the creation, development, and collapse of bubbles in a liquid that creates high pressure and temperature trailed by high rate of cooling. Such acoustic cavitation is driving force in variety of materials synthesis and reactions. These properties are responsible for the preparation of morphology controlled selective nanoparticles (Patil and Bhanage 2016). Sonochemical studies are linked with the effects of ultrasonic waves on the chemical system (Fig. 4). Herein, the chemical effects of ultrasound will not have direct effect with reacting materials. In its place, energy obtained from collapse of bubbles, that is, acoustic cavities interact with molecular species. Ultrasonic horn is commonly used for the preparation of various MNPs and MONPs.

MNPs Preparation Using Sonochemical Method The preparation of PdNPs from palladium acetate Pd(CH3 CO2 )2 precursor at room temperature by sonochemical reduction method is reported in presence of myristyltrimethylammonium bromide, [CH3 (CH2 )13 N(CH3 )3 Br] (NR4 X), in THF or methanol. NR4 X demonstrates stabilizing as well as reducing agent toward preparation of PdNPs. These PdNPs find application in C–C coupling reaction, in

Fig. 4 Schematic illustration showing application of ultrasonic horn

2148

A. B. Patil et al.

the absence of phosphine ligands, to a reasonable degree of 30 conversions (Arul Dhas and Gedanken 1998). The use of sonochemical synthesis of PdNPs and PtNPs from salts of K2 PdCl4 and H2 PtCl6 , respectively, is reported. The particle size of 3.6 ± 0.7 nm and 2.0 ± 0.3 nm were recorded, respectively (Fujimoto et al. 2001).

MMNPs by Sonochemical Route In addition to preparation of monometallic nanoparticles using sonochemical method the technique is found to be best for the fabrication of MMNPs preparation. The colloidal aqueous solution of bimetallic Pt/RuNPs have been obtained by sonochemical reduction of ionic Pt and Ru (Vinodgopal et al. 2006). Sonochemical-assisted synthesis of gold and palladium bimetallic nanoparticles, that is, Au@PdNPs, is also known. In aqueous medium, under optimum reaction conditions, Au+3 and Pd+2 from corresponding precursors were reduced to metallic form using sodium dodecyl sulfate (SDS) and ultrasound irradiation technique. TEM results revealed formation of spherical particles of 8 nm size (Mizukoshi et al. 1997). Metal Oxide Nanoparticles by Sonochemical Route With the application of sonochemical methods for the synthesis of monometallic and MMNPs, the technique for the synthesis of MONPs has been well explored. The additives free preparation of ZnONPs using ZnC4 H6 O4 salt and diol solution, that is, HOCH2 CH2 CH2 CH2 OH using ultrasonic irradiations is reported (Fig. 5) (Bhatte et al. 2011). The synthesis of nanomaterials of chromium oxide (Cr2 O3 ) and manganese oxide (Mn2 O3 ) by ultrasonic irradiation of aqueous solutions containing respective

Fig. 5 TEM image of ZnONPs prepared by sonochemical route. (Reprinted from Bhatte et al. (2011) with permission from Elsevier)

84 Green Nanoparticles: Synthesis and Catalytic Applications

2149

salts is reported (Dhas et al. 1997). The amorphous tungsten oxide (WO2 ) from tungsten hexacarbonyl W(CO)6 salt can be prepared. In typical synthesis process nanomaterial of WO2 has been prepared by ultrasound irradiation of a solution W(CO)6 in diphenylmethane (DPhM) at 90 ◦ C temperature condition in the presence of Ar (80%)–O2 (20%) gaseous mixture. Report showed snowflake-like dendritic particles consisting of a mixture of monoclinic and orthorhombic WO2 crystals (Koltypin et al. 2002). The preparation of ZnO hollow nanospheres around 500 nm size made up of ZnO nanoparticles (particle diameter of about 15 nm) has been reported by a simple and rapid sonochemical method (He et al. 2010).

Electrochemical-Assisted Nanomaterial Synthesis Electrochemistry has proven its diversity of applications in the field of physics and chemistry. In recent investigation electrochemical synthesis found an effective method for preparation of nanostructured materials. Using electrochemical method shape and size selective nanoparticles can be easily prepared at low cost and low synthesis temperatures. The synthesis route of nanoparticles using electrochemical method demonstrated several benefits in terms of quickness, clean, simplistic, inexpensive, and environmental benign. Reetz and Helbig introduced electrochemical method for the preparation of PdNPs and NiNPs (Reetz and Helbig 1994). In typical synthetic protocol, the intermediate metal salt was formed by reduction at the cathode, which was further stabilized by tetra alkyl ammonium salts along with co-electrolyte acetonitrile/tetra hydrofuran. PdNPs were also prepared by using electrochemical synthesis in an ionic liquid that plays the role of electrolyte cum stabilizer (Deshmukh et al. 2011). The electrochemical method for preparation of stable and durable AgNPs suspension in aqueous solution as well as silver powders is reported. The as obtained AgNPs is found to be spherical in the size in the range of 2–20 nm (Khaydarov et al. 2009). In this regard, different nanoscale materials, such as nanoparticles, nanowires of Au, Pt, Ni Co, Fe, Ag, etc., are synthesized using electrochemical deposition method.

Solar Energy The application of solar energy has put its remarkable application as a green and inexpensive energy source for nanomaterial preparation. Here metal nanoparticles were rapidly prepared photochemically by treating metal ions. The dendrimerprotected AuNPs were prepared by application of sunlight (Fig. 6) (Luo 2008). Similarly the single step dendrimer-protected AuNPs using natural solar radiations can be prepared (Luo 2009). Though, natural solar energy is found to be suitable for synthesis of AuNPs and AgNPs, considering the scope and area of subject the same protocol fails for the synthesis of several other MNPs and MONPs, which needs high intensity energy.

2150

A. B. Patil et al.

Fig. 6 (a and c) TEM images obtained at molar ratio 1:1 and 16:1. (b and d) Size distribution histogram of AuNPs obtained with molar ratio 1:1 and 16:1. (Reprinted from Luo (2008) with permission from Elsevier)

To overcome this problem the concept of concentrated solar energy (CSE) was developed for the preparation of metal nanoparticles of low standard reduction potential. The idea of concentrated energy satisfied the necessity of high energy flux essential for the reduction of metal ions like Pd. The use of “Fresnel lens” as a solar concentrator was helpful for trapping and concentrating solar radiations by which temperature of the reaction mass can rise up to 95 ◦ C in aqueous medium. As a combined effect from duel energy sources the faster synthesis of nanoparticles has been achieved. Moreover, this protocol can be used for the controlled and faster preparation of metal oxides nanoparticles. This technique was used for the synthesis of PdNPs. In a typical preparation protocol, reaction mixture was irradiated under

84 Green Nanoparticles: Synthesis and Catalytic Applications

2151

Fig. 7 (a) Schematic representation of reaction setup and reaction progress as a result of solar irradiation by color change (inset). (b) TEM images of PdNPs. (c) FEG–SEM image of PdNPs. (Reprinted from Patil et al. (2012a) with permission from Elsevier)

CSE for different time intervals (Fig. 7). The acquired results showed 30–45 nm size of PdNPs particles (Patil et al. 2012a). The concept is found to be applicable for the shape selective nanoparticles preparation as well (Patil et al. 2012b; Patil and Bhanage 2013b). The concept of CSE was also applied for the preparation of MONPs such as ZnONPs and MgONPs. The ZnO nanoparticles of 10–15 nm size were prepared using Zn(CH3 COO)2 and 1,4-butanediol (Patil et al. 2012c). Whereas 5–20 nm size MgO nanoparticles were prepared from magnesium acetate precursor and 1,4butanediol under influence of CSE (Patil and Bhanage 2013c).

Natural Fibers: Green Nanomaterials Naturally occurring fibers such as cocoon silk, spider silk, cellulose, wool, and botanical fibers are considered as carbon-neutral materials that normally contain fibrils of