International Conference on Reliable Systems Engineering (ICoRSE) - 2021 [305] 9783030833671, 9783030833688

Citation preview

Lecture Notes in Networks and Systems 305

Daniela Doina Cioboată   Editor

International Conference on Reliable Systems Engineering (ICoRSE) - 2021

Lecture Notes in Networks and Systems Volume 305

Series Editor Janusz Kacprzyk, Systems Research Institute, Polish Academy of Sciences, Warsaw, Poland Advisory Editors Fernando Gomide, Department of Computer Engineering and Automation—DCA, School of Electrical and Computer Engineering—FEEC, University of Campinas— UNICAMP, São Paulo, Brazil Okyay Kaynak, Department of Electrical and Electronic Engineering, Bogazici University, Istanbul, Turkey Derong Liu, Department of Electrical and Computer Engineering, University of Illinois at Chicago, Chicago, USA, Institute of Automation, Chinese Academy of Sciences, Beijing, China Witold Pedrycz, Department of Electrical and Computer Engineering, University of Alberta, Alberta, Canada, Systems Research Institute, Polish Academy of Sciences, Warsaw, Poland Marios M. Polycarpou, Department of Electrical and Computer Engineering, KIOS Research Center for Intelligent Systems and Networks, University of Cyprus, Nicosia, Cyprus Imre J. Rudas, Óbuda University, Budapest, Hungary Jun Wang, Department of Computer Science, City University of Hong Kong, Kowloon, Hong Kong

The series “Lecture Notes in Networks and Systems” publishes the latest developments in Networks and Systems—quickly, informally and with high quality. Original research reported in proceedings and post-proceedings represents the core of LNNS. Volumes published in LNNS embrace all aspects and subfields of, as well as new challenges in, Networks and Systems. The series contains proceedings and edited volumes in systems and networks, spanning the areas of Cyber-Physical Systems, Autonomous Systems, Sensor Networks, Control Systems, Energy Systems, Automotive Systems, Biological Systems, Vehicular Networking and Connected Vehicles, Aerospace Systems, Automation, Manufacturing, Smart Grids, Nonlinear Systems, Power Systems, Robotics, Social Systems, Economic Systems and other. Of particular value to both the contributors and the readership are the short publication timeframe and the world-wide distribution and exposure which enable both a wide and rapid dissemination of research output. The series covers the theory, applications, and perspectives on the state of the art and future developments relevant to systems and networks, decision making, control, complex processes and related areas, as embedded in the fields of interdisciplinary and applied sciences, engineering, computer science, physics, economics, social, and life sciences, as well as the paradigms and methodologies behind them. Indexed by SCOPUS, INSPEC, WTI Frankfurt eG, zbMATH, SCImago. All books published in the series are submitted for consideration in Web of Science.

More information about this series at http://www.springer.com/series/15179

Daniela Doina Cioboată Editor

International Conference on Reliable Systems Engineering (ICoRSE) - 2021

123

Editor Daniela Doina Cioboată Bucharest, Romania

ISSN 2367-3370 ISSN 2367-3389 (electronic) Lecture Notes in Networks and Systems ISBN 978-3-030-83367-1 ISBN 978-3-030-83368-8 (eBook) https://doi.org/10.1007/978-3-030-83368-8 © 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

Foreword

During the last two years, we have witnessed unprecedented challenges. But human beings’ ingenuity is endless, and we were able to find solutions to all the threats we deal with. In this context, organizing a scientific event is not an easy task. This is why I am very proud to be able to coordinate a conference during the pandemic that seems–hopefully–to be coming to an end. And it is due to the degree of excellence that science has attained–in terms of medicine–that we can now say that we stand firm against the threat of COVID-19. And, more than ever, the impact that science has–on the long run–on the well-being of society is easier than ever to spot. The evolution of science and technology, so fast paced, is obvious especially in the fields of electronics and microelectronics, automation and digitization. These have resulted in a shift in focus, from precision mechanics to mechatronics, integronics, adaptronics and currently claytronics and cyber-mix-mechatronics. This is why the National Institute of Research and Development in Mechatronics and Measurement Technique (INCDMTM), which is the main organizer of our conference, and that is an organization that has been writing the history of Mechatronics in Romania for 40 years, has strived to keep up with the trends and developments in the field. Consequently, we have changed the concept and the structure of former International Conference of Mechatronics and Cyber-Mix-Mechatronics (ICOMECYME), and we have established ICoRSE—The International Conference on Reliable Systems Engineering. As of this year, ICOMECYME, which used to have a rich tradition behind (it derives from MECAHITECH International Symposium too), aims to adapt to the challenges of the field of mechatronics and that of related engineering fields, thus becoming ICoRSE. The conference that my team and I are coordinating aims to promote scientific research results and technological developments in reliable systems engineering, cyber-physical systems, mechatronics, applied mechanics and complementary fields, by facilitating the interaction and exchange of experience and good practice between experts in universities, research institutes and private companies.

v

vi

Foreword

I would like to take the chance to thank all the authors for their valuable contributions. Given the reasons mentioned in the first lines, I appreciate their effort more than ever. This is the first year that I am coordinating an international conference, and it is a great honour to benefit from such valuable works. I am also very proud to say that the papers we have reunited in the form of these proceedings come from all major parts of the world. The collected papers are from Europe, Asia, North and South America, and the fact that the country of origin of the authors of papers is so diverse, along with the rich scientific contents debated in the papers, makes me able to strongly stress out the importance of this event and to express my belief that ICoRSE 2021 is–indeed–a conference that can render the current state of the art of mechatronics. To conclude, I would like to thank the organizing team members for their support in all areas. An international event is a matter of team work above anything, and it is important to acknowledge the efforts of my colleagues too. Also, the excellent collaboration with Springer is to be pointed out, and I am highly thankful for the support offered by Mr. Holger Schaepe and Mr. Thomas Ditzinger in the development of the work herein. I genuinely hope that the readers of the proceedings will find the current book interesting and useful. Doina-Daniela Cioboată

Contents

Study of a Test Stand for Determining the Oil Density in Hydraulic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mihai Avram, Valerian-Emanuel Sârbu, and Mariana-Florentina Ştefănescu The Absorbents Nanoporous Structures Regeneration for Industrial Dryers by Microwave Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sergey Dobrotvorskiy, Aleksenko Borys, Vitalii Yepifanov, Yevheniia Basova, Ludmila Dobrovolska, and Viktor Popov

1

8

Lifetime of Optical Fibers Submitted to ThermoMechanical Stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R. El Abdi, R. Leite Pinto, G. Guérard, and C. Capena

23

Design of a Mobile Robot to Work in Hospitals and Trajectory Planning Using Proposed Neural Networks Predictors . . . . . . . . . . . . . . Şahin Yıldırım and Sertaç Savaş

32

Design and Measurement of the Peltier Cell Thermal Actuator for Fine Adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jan Hošek

46

S-Shape Feedrate Scheduling Method with Smoothly-Limited Jerk in Cyber-Physical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Volodymyr Kombarov, Volodymyr Sorokin, Yevgen Tsegelnyk, Sergiy Plankovskyy, Yevhen Aksonov, and Olena Fojtů Selecting the Method for Pre-tightening Threaded Connections of Heavy Engineering Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anatoliy Gaydamaka, Yuriy Muzikin, Volodymyr Klitnoi, Yevheniia Basova, and Sergey Dobrotvorskiy

54

69

vii

viii

Contents

Production Scheduling of Semiconductor Wafer Fabrication Facilities Using Real-Time Combinatorial Dispatching Rule . . . . . . . . . . . . . . . . . Suraj Panigrahi, Srijeta Agrahari, José Machado, and V. K. Manupati Microindentation Hardness Testing of D – gun Sprayed Coatings . . . . . Oleksiy Romanchenko, Oleksandr Lohunov, Yuriy Kharlamov, Volodymyr Sokolov, and Oleg Krol

78 91

Application of Magnetic Field on Lubricating Cooling Technological Condition in Metal Cutting Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Umarov Erkin, Mardonov Umidjon, and Shaozimova Umida Rapid Prototyping of a Lower-Body Exoskeleton for Paraplegia Patients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Filip-Alexandru Harmon and Patricia-Isabela Brăileanu Methods for Testing the Strength of Layers for Different Optical Coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Ciprian Ion Rizescu, Ionela Ghita, and Dana Rizescu Force Simulation of Bird Strike Issues of Aircraft Turbojet Engine Fan Blades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Vyacheslav Merculov, Mykola Kostin, Gennadii Martynenko, Natalia Smetankina, and Volodymyr Martynenko ROBO-PVAFM Proper Software Platform . . . . . . . . . . . . . . . . . . . . . . 142 Adrian Olaru, Tiberiu Dobrescu, and Serban Olaru Hardening, High-Speed Steel R6M5, Using a Combined Heat Treatment Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Umarov Tolibjon Increasing the Abrasive Wear Resistance of Steels by Heat Treatment with Preliminary Preparation of the Structure . . . . . . . . . . . 162 Darob Berdiev, Botir Saydumarov, and Nargiza Makhmudova Biodegradable Starch-Based Polyvinyl Alcohol Films with Zinc-Oxide Particles for Wound Dressing Applications . . . . . . . . . . . . . 171 Mohammad Mohsen Delavari and Ion Stiharu Compliant Positioning System with 6 DOF for High Precision Medical Standing Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 Mihai Tutoveanu, Nichita Larisa Milodin, Nicoleta Mirela Popa, Flavia-Petruța-Georgiana Artimon, and Constantin-Daniel Comeagă Formation Structure of Cement Systems Under the Influence of Chemical Additives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 Makhmudova Naima

Contents

ix

Increasing the Accuracy of Calibration Device for Measuring the Moisture of Bulk Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 Erkin Uljaev, Shohrukh Narzullayev, Ubaydullayev Utkir, and Sulaymonova Shoira Device for Processing Micro-bores by Electrical Discharge Machining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 Aurel Mihail Țîțu and Alina Bianca Pop Communication and Control Algorithms for a Heterogenous Multi-agent System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 Andrei Cristian Dinu, Paul-Nicolae Ancuța, and Victor-Marin Zafiu Development of Measurement Scales for Measuring Performance Value in the Market of Research, Development, and Innovation in Technical Science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 Nina Antičić Mathematical Model for Calculating Heat Exchange . . . . . . . . . . . . . . . 243 Turakhodjaev Nodir, Tursunbaev Sarvar, Jeltukhin Andrey, and Meliboyev Yahyojon Technologies for Thin Layers on Ceramics Substrate . . . . . . . . . . . . . . 250 Georgeta Ionascu, Elena Manea, Raluca Gavrila, and Edgar Moraru Flexural Test of 3D Printed Mecanum Rollers . . . . . . . . . . . . . . . . . . . . 266 Victor-Marin Zafiu, Diana-Maria Cotorobai, Ana Maria Eulampia Rolea, and Andrei Cristian Dinu Predictive Motor Speed Control for an Industrial Robot. A Dead-Beat Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 Iulia Clitan, Cristina Stancioi, and Vlad Muresan Mathematical Model of the Stress State of the Antenna Radome Joint with the Load-Bearing Edging of the Skin Cutout . . . . . . . . . . . . 287 Sergei Kurennov, Natalia Smetankina, Vladimir Pavlikov, Darya Dvoretskaya, and Vladyslava Radchenko Intelligent Network for Measuring Natural Environment Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 Paul-Nicolae Ancuţa and Sorin Sorea Compression Testing of PA2200 Additive Manufactured Lattice Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304 Nichita-Larisa Milodin, Nicoleta-Mirela Popa, Mihai Tutoveanu, and Flavia-Petruta-Georgiana Artimon The Use of CPS for Assistive Technologies . . . . . . . . . . . . . . . . . . . . . . . 316 Pierluigi Rea, Erika Ottaviano, and Maurizio Ruggiu

x

Contents

Mathematical Modeling and Simulation in Sheet Hydroforming Process for the Parts of Space Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 Manh Tien Nguyen and Truong An Nguyen Approach to Product Quality Requirements in the Context of Aeronautical Domain Process Modeling . . . . . . . . . . . . . . . . . . . . . . . 337 Aurel Mihail Țîțu and Gheorghe Ioan Pop Pulsed Fiber Laser Surfaces Micro-processing - Optimization and Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349 Daniela Doina Cioboata, Mircea Udrea, Mihai Selagea, Danut Iulian Stanciu, Silvia Savencu, Radu Mihail Udrea, and Cristian Logofatu Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363

Study of a Test Stand for Determining the Oil Density in Hydraulic Systems Mihai Avram1 , Valerian-Emanuel Sârbu2(B) , and Mariana-Florentina Stef˘ ¸ anescu2 1 Faculty of Mechatronics and Mechanical Engineering, Mechatronics and Precision

Engineering Department, University Politehnica of Bucharest, Bucharest, Romania 2 University POLITEHNICA of Bucharest, Bucharest, Romania

Abstract. The purpose of this paper is to document the development of a test stand that enables experimentally determining the viscosity of the fluid used in a hydraulic system. This test stand also doubles as a method for determining the influence of the temperature and pressure on this parameter. The resulting setup allows full control of the system while also measuring and processing the measured data. For the core of this type of system a microcontroller was selected. High performance transducers are used for ensuring accurate measurements, they provide a proportional electrical output. The microcontroller firmware mostly runs as a command processor, a Lab View program is responsible for communicating with it in order to acquire data and control the system variables in a predefined sequence. Keywords: Hydraulic oil · Density · Hydraulic drive system · The theoretical and experimental analysis

1 Introduction Hydraulic systems use pressurized fluid to transfer the energy from the pump to the desired system. At first, hydraulics appeared and rapidly evolved in applications that require the control of high power, high inertia and forces with relatively high accuracy; furthermore, they also facilitate full control of the position and speed. The method by which the delivery of hydraulic power is adjusted gives hydraulic transmissions advantages compared to electrical and mechanical transmission systems [1, 2]. In order to determine the performances of this type of system, static or dynamic, there are two methods available depending on the circumstances: theoretical and experimental. In practice we generally encounter two scenarios: When the real system is available. When it needs to be designed and then manufactured. In both cases the system needs to be mathematically defined [6]. Mathematical models of systems that use a liquid transport medium include several equations that describe its movement if the following parameters are known: fluid density - ρ, Elasticity modulus - E, Dynamic viscosity - η and cinematic viscosity - μ. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 1–7, 2022. https://doi.org/10.1007/978-3-030-83368-8_1

2

M. Avram et al.

Mathematical models are used for describing existing machines and for designing new ones. The first case is used in attempts to determine optimum functional parameters of the system in order to improve its overall performance. Thus, it requires knowing these parameters and the way they vary in time. This can be accomplished with an experimental test stand that is designed specifically for this purpose. In order to ensure that valid values are measured the system must be designed with accuracy in mind, the transducers need to be of high quality along with the acquisition system. Temperature, pressure and flow sensors are necessary in this setup. The control and data acquisition are done with a control unit built around a microcontroller [3–5]. Designing of this system was done with some targets in mind, some are considered general practice while others are application specific: • • • • • • •

Rapid deployment of the equipment and its components Being able to measure both static and dynamic variables Lightweight, robust and compact design The usage of standard and readily available components Can adapt to different sensors Needs to be designed with current easily available technology Be able to easily acquire the components in due time

Following these guidelines results in: high cost for designing and measurements, complex test setups, expensive equipment, and poor time usage. Up next, the experimental test stand that enables the measurement of the fluid density and the test procedures will be described.

2 Methodology and Experimental Details To be able to experimentally determine the density of a mineral oil, used as work medium in a hydraulic drive system, a stand has been designed and built with the corresponding functional scheme presented in Fig. 1A. The proposed stand contains the following equipment: – the hydraulic power generation group GGE, group containing the fixed flow pump P and safety valve Ssig ; – proportional valve SP; – classic hydraulic valve 3/2, with a preferential position and electrical control DHC; – proportionate butterfly valve DrP; – flow transducer Tq ; – pressure transducers TPav s, i TPam ; – temperature transducer TT ; – electronic acquisition and control block BEAC. Observation: The proportionate butterfly valve DrP used is a proportional hydraulic distributor, model 4 WREE 6 EA08-2X/G24K31/A1V, manufactured by Rexroth Bosch, to which the holes A and T are plugged (Fig. 1B).

Study of a Test Stand for Determining the Oil Density

3

Fig. 1. Functional scheme of the experimental stand (A); Proportional hydraulic valve (B)

Proportional equipment used, proportional valve SP and the proportional butterfly valve DrP are served by proportional amplifiers AESP and AEDrP . Figure 2 presents two images of the experimental stand designed and realized by us, where A represents the electronic control block and B represents the hydraulic power generation group.

Fig. 2. Main components of experimental setup, with the electronic control block (A) and hydraulic power generation group

The presentation will be clear and concise, and the symbols used therein will be specified in a symbol list (if necessary). In the paper the International System measurement units will be used. There will however be no apparatus or installation descriptions. The initial condition of the system is defined as follows: u0 = 0 V, uc1 = 0 V and uc2 = 0 V

4

M. Avram et al.

In this situation the distributor DHC assumes the position (0), the flow section through the proportional hydraulic butterfly valve DrP is zero, and the proportional valve SP is open (no flow resistance). From this moment experimental determinations can start, which involve the following steps: Start the hydraulic power generator unit and apply a control voltage (u0 = 24 V) to the GGE, the distributor electromagnet DHP will engage and the distributor will materialize the position (1)); At the input of the amplifier AESPis a voltage is installed uc2 ∈ [0,10] V which provides the desired hydraulic pressure to the hydraulic system; this pressure will be constantly maintained during the experiment, as in the memory of the microcontroller that is integrated in the electronic acquisition and control block BEAC there is an algorithm specifically designed for this purpose; First the electronic amplifier AEDrP is supplied with a voltage of uc1 = 1 V; Measurement is done using the system transducers: Pam , Pav , q, x and t; Finally, there is the increasing of the voltage uc1 by uc1 ; previous steps are repeated until target voltage uc1 = 10 V is reached. Same pattern but in reverse (from 10 V to 1 V) is then applied;

3 Experimental Determination of Mineral Oil Density

Fig. 3. Experimental results

Figure 3 shows the results obtained for two set feed pressure values Pam , i.e.25 bar and 30 bar.

Study of a Test Stand for Determining the Oil Density

5

For a set of experimentally determined values, the density of the oil can be calculated with the relationship [6]: ρ=

2 · S 2 (uc ) · (Pam − Pav ) q2 c

(1)

The law of variation of the flow section is not known, the only certain thing is the maximum stroke of the drawer xmax = 1,25 mm as this value is specified by the manufacturer in the catalog tab of this equipment. For this reason, we have attempted at the theoretical determination of the variation law of the flow section. For the drawer geometry indicated by the manufacturer in the documentation of this equipment, the following expression of the flow section has been determined:          b 2 b 2 − (x − la ) · x − la · b − (x − la ) S(x) = n · · arccos 1 − · (x − la ) − 2 b 2

(2) where: n – the number of channels processed on the drawer in the control area; b – the width of the channel; la – cover length. This expression is valid for x ∈ [la , l a + b/2]. In the range [0, la ) the flow section is null. Figure 4 shows the variation of the flow section according to the spool position for different situations. Below we consider that the flow section through the butterfly valve corresponds to the variant (*) on the graph (see 4). In this case, the maximum value of the flow section is obtained for xmax = 1.25 mm and is Sc,max = 5,8 mm2 . Following the measurements, for a maximum control voltage were obtained the following values: p = Pam − Pav = 20, 469 bar q = 23, 61 l/ min Oil density can be now calculated with the relation (1) as follows: ρ = 360 ·

2 · 5, 82 · 20, 469 = 889, 39 kg/m3 23, 612

The stand allows determination of density variance with pressure and temperature. Because the connections between the stand equipment are made with metal inserts rubber hoses, the maximum system pressure must be limited to 80 bar. Within this scope, density variance with pressure is insignificant and most of the times the density is assumed to be constant. Keeping the same stand structure and making couplings of the equipment with metal ducts, the working pressure may be increased to 320 bar a situation where the determination of density variation with temperature makes sense.

6

M. Avram et al.

Fig. 4. Variation of flow section depending on the position of the drawer

4 Conclusions Determination of the density of a mineral oil used as a working medium in a hydraulic drive system under certain working conditions, as well as the dependence of the density with temperature and pressure is an important issue for which a solution is still being sought. The experimental stand designed and made by the authors is an effective solution to solve the above-mentioned issues. This stand represents, through its hardware and software structure, a complex system of experimentation and testing, where all activities are computer assisted. One of the advantages of the system consist in that the experimental results obtained can be both directly viewed on the computing system monitor, during the measurement, which is a great advantage in experimental research, and stored in memory in the form of csv files, for further data processing and their printing as test /experimental bulletins, without human operator intervention.

References 1. Avram, M.: Actionari Hidraulice si Pneumatice. Editura Universitara, Bucharest (2005) 2. Wu, W., Yu, C.: Simulation and experimental analysis of hydraulic directional control for displacement controlled system. IEEE Access 99, 1–1 (2007) 3. Valer, D.: Proiectarea Sistemelor Mecatronice. Editura Politehnica, Timisoara (2007)

Study of a Test Stand for Determining the Oil Density

7

4. Ilie, I., Blejan, M.: Distributed Hardware and Software Architecture for Monitoring and Control of Hydraulic Drives, “HIDRAULICA” (No. 4/2017) Magazine of Hydraulics, Pneumatics, Tribology, Ecology, Sensorics, Mechatronics 5. Hristea, A., Radoi, R., Tudor, B., Balan, I.: Electronic Module for Data Acquisition and Transmitting for Portable Power Steering Test Equipment, “HIDRAULICA” (No. 4/2017) Magazine of Hydraulics, Pneumatics, Tribology, Ecology, Sensorics, Mechatronics 6. Avram, M., S, tef˘anescu, M.F., Sârbu, V.E., N˘astase, V.: Theoretical analysis of a hydraulic drive system. The influence of the work environment on the performance of the system. Int. J. Mechatron. Appl. Mech. 7, 152–157 (2020)

The Absorbents Nanoporous Structures Regeneration for Industrial Dryers by Microwave Energy Sergey Dobrotvorskiy1(B) , Aleksenko Borys1 , Vitalii Yepifanov1 Yevheniia Basova1 , Ludmila Dobrovolska1 , and Viktor Popov2

,

1 National Technical University “Kharkiv Polytechnic Institute”,

2, Kyrpychova street, Kharkiv 61002, Ukraine [email protected] 2 Joint Stock Company “FED”, 132 Sumska Street, Kharkiv 61023, Ukraine

Abstract. The presented work is devoted to solving the problem of energy conservation in industrial production. Modern industrial equipment and technological processes use compressed air energy. Wherein deep drying of compressed air is often required. The air preparation process is an expensive and energy-intensive process. Reducing useless energy losses, in this case, is possible through the use of innovative technologies that are technically feasible and economically justified. The adsorption air dryer’s efficiency feasibly increased by using microwave energy to regenerate the adsorbent by deporising water molecules from its nanoporous structure. This paper explores the application of this regeneration technology in air dryers wich have high throughput. The scientific novelty of this submitted research lies in this study of the regularities of the spatial distribution of thermal and electromagnetic energy and their uniform distribution in the adsorbent volume if the proposed innovative technology is applying. Also, it was establishing the dependence of the intensity of exposure to microwave energy depending on the frequency, voltage level, and design features of the adsorption tower. The practical significance lies in the proposition of a new technological process for the regeneration of the adsorbent in adsorption dryers with high performance. Keywords: Nanopores · Air dryer · Regeneration · Microwave energy

1 Introduction Nanotechnology is used intensively in modern industrial [1–4], agricultural [5], and food [6, 7] production. In particular, the use of nanostructures has become widespread in the technological process of drying compressed air using adsorption technologies. Today, adsorption dryers are the main type of equipment that can provide modern technology processes with high-quality compressed air. The problem of increasing the efficiency of adsorption dryers is to reduce energy costs and increase the intensity of the process of periodic desorption of accumulated moisture from the nanoporous structure of the adsorbent. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 8–22, 2022. https://doi.org/10.1007/978-3-030-83368-8_2

The Absorbents Nanoporous Structures Regeneration

9

Today the conventional adsorbent regeneration technology uses heated purge air for drying the adsorbent [8–11]. Air dryers with hot regeneration of the adsorbent are currently the most economical means of producing dried air and the development of the design of this type of equipment is the most important direction to improve the energy efficiency of industrial production. In this case, energy is transferred from the electric heater to the purge air, then from the heated air to the adsorbent material, and only then to the water molecules in the pores. At the same time, at the stages of energy transfer, its non-productive losses occur. It is possible to qualitatively increase the efficiency of adsorption dehumidifiers by applying new technology for removing moisture from the nanoporous structure of the adsorbent, which will make it possible to act directly on water molecules. We have proposed a regeneration technology using the energy of microwave radiation. The electromagnetic field created in the cavity of the adsorption column causes accelerated movement of water molecules in micro-pores, which leads to deporization of moisture and its subsequent evaporation from the surface of the adsorbent granules. The impact of microwave energy on an adsorbent saturated with water leads to an increase in its temperature. This makes it possible to purge the adsorbent volume with a stream of unheated air. Studies show that the use of microwave radiation reduces the energy loss, which is spent during the regeneration of the adsorbent [12, 13]. Microwave energy acts directly on water molecules. This avoids unproductive convective and conductive heat losses, which is inevitable in the case of using classical regeneration technology [14, 15]. Under the influence of microwave radiation, the drying of the adsorbent occurs more intensively, faster [16–29], and at a lower temperature [30]. This makes it possible to reduce the energy consumption for the regeneration of the adsorbent and the consumption of dried air for the subsequent cooling of the adsorbent. The presented article examines the use of high-frequency exposure to adsorbent adsorption dryers with high throughput.

2 Research Problem The process of regenerating serves to remove moisture desorbed from the porous structure of the adsorbent. The completeness of regeneration depends on the uniformity of energy distribution in the volume of the adsorbent [31–37]. All areas of the adsorbent volume must be energized. This eliminates the occurrence of residual zones with high moisture content in the porous structure. Also, the design of the adsorption tower must ensure a uniform laminar flow of purge air through the entire volume of the adsorbent. Stagnant zones with a low flow rate of purge air prevent complete moisture removal, which reduces the quality of regeneration. Our earlier studies [38–42] have shown the possibility of using a magnetron with a wave-guide system for the action of microwave energy on a small volume of the adsorbent. The low price of this design makes it advisable to use microwave regeneration in dryers with a productivity of 1.5 m3 /min and less [43]. The problem of using one or

10

S. Dobrotvorskiy et al.

a group of magnetrons to regenerate a large volume of adsorbent is the limited depth of wave energy penetration into the bulk of the material. It should be taken into account that silica gel loses its useful properties when it overheats above 200 °C. From this, it follows that an increase in temperature above this limit to intensify the process of regeneration of the adsorbent is unacceptable. Thus, an unlimited increase in the parameter is not possible to increase the intensity of the regeneration process. In a column with a diameter of 100 mm, it is possible to achieve a uniform distribution of energy without the appearance of peripheral overheating zones due to heat exchange in the horizontal section of the column. In this case, the microwave energy does not intensively affect the inner layers of the adsorbent (Fig. 1, pos.5).

Fig. 1. Design of adsorption columns with pole plates and waveguide system and energy distribution in the adsorbent volume.

An increase in the number and power of magnetrons leads to the appearance of local overheating zones on the outer layers of the adsorbent volume without increasing the total intensity of regeneration. Dehumidifiers with a productivity of 5 m3 /min and more have an ad-sorption tower diameter of more than 250 mm and cannot use microwave

The Absorbents Nanoporous Structures Regeneration

11

radiation effectively enough due to the limited penetration depth. Meanwhile, highcapacity dehumidifiers are widely used in production. Also, in the chemical and gas industry, adsorption plants are used in which the diameter of the adsorption columns is several meters. In this case, exposure to radiation from the outside does not affect even if the energy is uniformly distributed over the outer surface of the adsorbent volume. To solve this problem, it is proposed to use concentric pole plates, which are located in the volume of the adsorbent (Fig. 1, pos.1). Pole plates are made in some form: 1) several nested cylinders (Fig. 2, pos. 1), the axes of which coincide with the axis of the column; 2) two nested spirals (Fig. 2, pos. 2); 3) a set of parallel planes (Fig. 2, pos. 3) inscribed in the cylindrical volume of the adsorption column.

Fig. 2. Variants of the pole plates placement in the adsorption column cavity (column horizontal section, top view).

The presented study is devoted to the study of the effect of microwave energy on the adsorbent in a dryer with columns having cylindrical pole plates.

3 Materials and Methods The research is carried out using several mathematical multiphysics computer models of adsorption column. The models are constructed to study the movement of the purge airflow through the adsorbent, the intensity of the high-frequency effect on the adsorbent, and to evaluate the uniformity of temperature distribution in its volume. Simulation is performed using the Komsol Multiphysicstm software environment.

12

S. Dobrotvorskiy et al.

The calculation is carried out using the mathematical model of a desiccant adsorption tower with a capacity of 50 m3 /min. The adsorption tower model is a hollow cylindrical body. The cylinder is closed from two ends with elliptical bottoms. The body material is steel. Concentric pole plates are located inside the column body. An adsorbent is placed in the spaces between the pole plates (Fig. 1, pos. 2,4). Thus, the adsorbent volume is divided by plates into separate concentric cylindrical volumes. The purge air blowing is carried out in the vertical direction, with the blowing air moving through the adsorbent bed endlong the pole plates. The purge air moves from the top-down direction, (Fig. 4) along the axis of the adsorption column. Adsorbents have a different porosity at a scale of a separate granule. This defines the adsorption capacity and affects the intensity of the regeneration process. The gas-dynamic process, considered at a scale of the adsorption column, will be determined by the size and shape of the adsorbent granules. This defines the bulk density of the adsorbent volume. The adsorbents, used in dryers, have the same characteristics: adsorbent granules have a spherical shape and a diameter of 3 to 7 mm. The normal range of porosities in granular systems is 5% to 35%. A hexagonal or cubic close pattern is assumed with the calculated packing density is 0.7045, gives a porosity void fraction of 0.2955, provided that all the balls are the same size. The difference in grain size will create a lower porosity. But adsorbent retaining grids of modern dryers can remove small granules and products of their shredding. Due to this the spread in the diameters of the granules is small. This makes it possible to use in the model the value of porosity of the adsorbent volume to be 20%. Granules will not give closed pores, so the effective porosity is considered equal to 100%. The physical properties of silica gel are shown in the table (Table 1). The properties of silica gel [44] are corrected taking into account the moisture content and the bulk density of wet silica gel. Table 1. Silica gel physical properties. Parameter

Dimension Value

Density

kg/m3

2900

Bulk density

kg/m3

780

Heat capacity

j/(kg*K)

730

Relative permeability 1

16

Relative permittivity

1

4,1

Electric conduction

S/m

0

Heat conduction

W/m

0,2

Linear dimensions of the model (Fig. 3) are given in the table (Table 2). The model uses steel material for the column and walls. Since the flow rate of the purge air in the column throughout the entire path does not exceed 1.2 m/s, the flow is calculated as laminar flow using equation [45, 46].

The Absorbents Nanoporous Structures Regeneration

13

Fig. 3. Linear dimensions of the adsorbent volume and the distance between the pole plates.

Table 2. Linear dimensions of the adsorption column. Index Size

Dimension, mm

H

Column height

1000

D

Column diameter

760

a

Distance between the pole plates 32/40/64

The impact on the adsorbent of high-frequency energy is carried out by applying an alternating voltage to the pole plates. The process is modeled as inductive heating. An alternating voltage across the pole plates is generated using a sinusoidal function with a given angular frequency. The voltage is determined by the amplitude of the function. The stationary problem of the air medium’s motion was calculated by the equation:    2  T (1) ρ(u · ∇)u = ∇ −pl + μ ∇u + (∇u) − μ(∇ · u)l + F 3 ∇ · (ρu) = 0 where: u - is the velocity, p - is the pressure, and μ - is the dynamic viscosity.

(2)

14

S. Dobrotvorskiy et al.

Fig. 4. The movement of the purge air in the adsorbtion column cavity and also through the adsorbent layer.

The movement of air in the volume of the adsorbent was modeled as a motion in a porous medium and was calculated by the equation:    2 1 1 1 1  ∇u + (∇u)T − μ (∇ · u)l ρ(u · ∇)u = ∇ −pl + μ ∈p ∈p ∈p 3 ∈p   θm (3) − μk −1 + βF |u| 2 u + F ∈p ∇ · (ρu) = Qm

(4)

where: θm - the value of porosity of the material. The dynamic heat transfer problem was calculated by the equation: ρCp u · ∇T + ∇ · q = Q + Qvd

(5)

q = −keff ∇T

(6)

where: Cp - heat capacity value, T - temperature, Q - quantity of heat. Heat transfer in a porous medium was calculated as:  keff = θp kp + 1 − θp k + kdisp where: θρ - is the volume fraction of the material, kρ - is the thermal conductivity.

(7)

The Absorbents Nanoporous Structures Regeneration

15

The influence of purge air flow on thermal processes in the column is calculated as dynamic heat transfer model through the following equations: ρCp

∂T + ρCp vtrans · ∇T + ∇ · q = Q + Qted ∂t

(8)

where ρ is the density, Cp – heat capacity, q – heat flux, T – temperature, Qe – heat sources and sinks. The initial temperature of the adsorbent is the same at all points in the volume of the adsorbent and is 20 °C. The purge air also has an initial temperature of 20 °C. The same temperature is conventionally considered the adsorption temperature. The electric field generated in adsorbent volume is calculated with equations:

 jσ 2 ε E=0 (9) ∇ × E) − k − ∇ × (μ−1 γ γ 0 ωε0 ∇ · J = Qj,v

(10)

J = σ E + jωD + Je

(11)

ω = 2π f

(12)

where μγ is the complex relative permeability; E – complex amplitude representing an oscillating electric field; εγ – complex relative permittivity; ω – circular (angular) frequency, k0 – phase constant of free space, σ – electrical conductivity, j – is an imaginary unit, J – externally generated current density, D – electric displacement, Qj,v – charge, f – frequency (Fig. 5).

Fig. 5. Variable electrical potential in the space between the pole plates.

16

S. Dobrotvorskiy et al.

The heating of the adsorbent due to the effect of alternating voltage is calculated as: ρCp u · ∇T = ∇ · (k∇T ) + Qe

(13)

Qe = J · E

(14)

The force, which is effects on the water molecule is calculated as: F =ρ·E F = i E=

(15)

∂E ρi ∂xi

(16)

U d

(17)

E = −∇V

(18)

where F is the electric displacement, ρ – dipole moment, U – potential difference, d – distance between the pole plates, V – electric potential. The calculations were carried out in the time interval from 0 to 180 min, with step 5 min.

4 Results and Discussion The calculation of the computer model shows that the high-frequency action can be used to increase the temperature in the adsorbent volume for regeneration (Fig. 6). Neglecting the change in the absorption coefficient of water as the frequency changes and also taking into account that the dipole moment of water (1.84 Debye) is a constant value, we can influence the efficiency of exposure to microwave energy by changing the frequency of exposure (9), the voltage between the pole plates and the distance between them (17). The calculation of the comparative model shows the positive effect of increasing the frequency on the dynamics of the growth of the average temperature of the adsorbent volume (Fig. 7). It should be noted that obtaining a high polarization frequency is a technically difficult task and requires specialized equipment. This is because most general-purpose semiconducting elements operate effectively in the frequency range up to 20 MHz. In this case, the problem of high voltages commutation is solved relatively easier and cheaper. The performed calculation shows a positive effect of an increase in the voltage which is supplied to the pole plates on the dynamics of an increase in the average temperature of the adsorbent volume (Fig. 8).

The Absorbents Nanoporous Structures Regeneration

17

Fig. 6. Dynamics of temperature rise in the volume of the adsorption column under the influence of microwave energy.

The disadvantage of this method is the danger of high voltage electric shock. Therefore, increasing the voltage on the pole plates to the level of hazardous ones (over 60 V) will require complicating the design of the adsorption column. This is due to the need to ensure the safety of the operating personnel and will entail an increase in the cost of the dryer design (Fig. 9). Also, the dynamics of the growth of the average temperature of the volume of the adsorbent (Fig. 8) are positively affected by a decrease in the distance a between the pole plates (Fig. 3).

18

S. Dobrotvorskiy et al.

Fig. 7. A figure Dynamics of growth of the average temperature in the volume of the adsorbent depending on the frequency. Voltage (peak) = 4 V, distance between pole plates = 4 mm.

Fig. 8. The dynamics of the growth of the average temperature in the volume of the adsorbent depending on the (peak) voltage on the pole plates. Frequency = 2.45 GHz, distance between pole plates = 4 mm.

The Absorbents Nanoporous Structures Regeneration

19

Fig. 9. Dynamics of the average temperature growth in the adsorbent volume depending on the distance between the pole plates. Frequency = 2.45 GHz, Voltage = 4 V.

It should be noted that the specified distance cannot be reduced to a size less than 10 mm, which corresponds to the minimum diameter of the adsorbent granule used in dryers. Also, reducing the distance between the pole plates will require an increase in their quantity. This, in turn, will complicate the design of the equipment and entail an increase in its cost.

5 Conclusions The presented work shows the possibility of using the energy of microwave radiation during the regeneration of adsorbents applied in compressed air dryers with high productivity, more than 50 m3 /min. It is shown that the impact on the adsorbent using pole plates makes it possible to uniformly distribute energy in a large volume of the adsorbent. The influence on the intensity of exposure to microwave energy of such factors as frequency and voltage on the pole plates, as well as the distance between them is considered. The economic effect of the use of microwave energy in the process of regeneration of nanoporous structures is achieved due to the intensification of drying in comparison with blowing with preheated air. Research shows the applicability of the described technology in high throughput adsorption dryers. This opens up opportunities for increasing the efficiency of this type of industrial equipment.

20

S. Dobrotvorskiy et al.

References 1. Mishra, R., Militky, J., Venkataraman, M.: Nanoporous materials (2019). https://doi.org/10. 1016/B978-0-08-102609-0.00007-9 2. Mishra, R. Militky, J.: Future outlook in the context of nanoscale textiles as a technology for the twenty-first century (2019). https://doi.org/10.1016/B978-0-08-102609-0.00009-2 3. Logar, N.Z., Tušar, N.N., Risti´c, A., Mali, G., Mazaj, M., Kauˇciˇc, V.: Chapter 5 - Functionalisation and structure characterisation of porous silicates and aluminophosphates. In: Ordered Porous Solids, pp. 101–126. Elsevier (2009). ISBN 9780444531896, https://doi.org/10.1016/ B978-0-444-53189-6.00005-6. 4. Jibowu, T.: A Review on nanoporous metals. Front Nanosci Nanotech 2 (2016), https://doi. org/10.15761/FNN.1000129 5. Vakal, S., Yanovska, A., Vakal, V., Yarova, T., Artyukhov, A., Shkola, V.: Investigation of morphology and composition of the mineral fertilizer granules with nanostructured areas. In: 2020 IEEE 10th International Conference Nanomaterials: Applications & Properties (NAP), Sumy, Ukraine, 2020, pp. 02IT03–1–02IT03–6 (2020). https://doi.org/10.1109/NAP51477. 2020.9309704 6. Arabmofrad, S., Bagheri, M., Rajabi, H., Jafari, S.M.: 4 - Nanoadsorbents and nanoporous materials for the food industry, Handbook of Food Nanotechnology, pp. 107–159. Academic Press (2020). https://doi.org/10.1016/B978-0-12-815866-1.00004-2, ISBN 9780128158661 7. Sekhon, B.S.: Nanotechnology in agri-food production: an overview. Nanotechnol. Sci. Appl. 2014(7), 31–53 (2014). https://doi.org/10.2147/NSA.S39406 8. Zhao, G., Liu, J., Xu, L., Guo, S.: Comparative study of conventional and microwave heating of polyacrylonitrile-based fibres. J. Pol. Eng. 41(3), 175–183 (2021). https://doi.org/10.1515/ polyeng-2020-0167 9. Cherbanski, R., Komorowska-Durka, M., Stefanidis, G.D., Stankiewicz, A.I.: Microwave swing regeneration vs temperature swing regeneration—comparison of desorption kinetics. Ind. Eng. Chem. Res. 50 (14), 8632–8644 (2011). https://doi.org/10.1021/ie102490v 10. Cavalcante, T.A.B.B., dos Santos Funcia, E., Gut, J.A.W.: Inactivation of polyphenol oxidase by microwave and conventional heating: Investigation of thermal and non-thermal effects of focused microwaves. Food Chem. 340, 127911 (2021). https://doi.org/10.1016/j.foodchem. 2020.127911, ISSN 0308-8146 11. Zhang, H.S., Chen, M.Q., Wei, S.H., Zhong, X.B.: Evaluation on drying performance of typical technical solid waste during microwave-combined hot air. Exp. Heat Transf. 34(2), 121–139 (2021). https://doi.org/10.1080/08916152.2020.1723743 12. Witkiewicz, K., Nastaj, J.: Modeling of microwave-assisted regeneration of selected adsorbents loaded with water or toluene. Drying Technol. 32(11), 1369–1385 (2014). https://doi. org/10.1080/07373937.2014.900506 13. Price, D.W., Schmidt, P.S.: Microwave regeneration of adsorbents at low pressure: experimental kinetics studies. J. Microw. Power Electromagn. Energy. 32(3), 145–154 (1997). https:// doi.org/10.1080/08327823.1997.11688334 14. Datta, A.: Porous media approaches to studying simultaneous heat and mass transfer in food processes. I: problem formulations. J. Food Eng. 80(1), 80–95 (2007). https://doi.org/10.1016/ j.jfoodeng.2006.05.013 15. Datta, A.: Porous media approaches to studying simultaneous heat and mass transfer in food processes. II: property data and representative results. J. Food Eng. 80(1), 96–110 (2007). https://doi.org/10.1016/j.jfoodeng.2006.05.012 16. Sun, Y., Zheng, T., Zhang, G., Liu, B., Wang, P.: Ce-Mn modify Al2O3 adsorbent and the effect on adsorption and regeneration properties. Environ. Sci. Pollut. Res. 25(23), 22818–22828 (2018). https://doi.org/10.1007/s11356-018-2076-1

The Absorbents Nanoporous Structures Regeneration

21

17. Tanaka, M., Takayama, S., Sano, S.: Microwave-assisted regeneration of synthetic zeolite used in tritium removal systems. Fusion Sci. Technol. 67(3), 604–607 (2015). https://doi.org/ 10.13182/FST14-T90 18. Yi, Q., et al.: A simulation method of coupled model for a microwave heating process with multiple moving elements. Chem. Eng. Sci. 231, 116339 (2021). https://doi.org/10.1016/j. ces.2020.116339, ISSN 0009-2509 19. Shi, H., Huang, K., Zhang, Y., Liu, Y., Gou, D.: Incoherent microwaves heating of water: a combined experimental and simulated investigation. Chem. Phys. Lett. 771, 138528 (2021). https://doi.org/10.1016/j.cplett.2021.138528, ISSN 0009-2614 20. Tian, S., et al.: Study on heat transfer process of microwave flash evaporation using water as medium. Int. J. Heat Mass Transf. 166, 120795 (2021). https://doi.org/10.1016/j.ijheatmasstr ansfer.2020.120795, ISSN 0017-9310 21. Demir, H.: Development of microwave assisted zeolite–water adsorption heat pump, Int. J. Refrigerat. 36(8), 2289–2296 (2013). https://doi.org/10.1016/j.ijrefrig.2013.07.005, ISSN 0140–7007 22. Cherba´nski, R., Molga, E.: Intensification of desorption processes by use of microwaves—an overview of possible applications and industrial perspectives. Chem. Eng. Process. Process Intensif. 48(1), 48–58 (2009). https://doi.org/10.1016/j.cep.2008.01.004, ISSN 0255–2701 23. Meloni, E., Martino, M., Pullumbi, P., Brandani, F., Palma, V.: Intensification of TSA processes using a microwave-assisted regeneration step. Chem. Eng. Process. - Process Intensif. 160, 108291 (2021). https://doi.org/10.1016/j.cep.2020.108291, ISSN 0255-2701 24. Legras, B., Polaert, I., Thomas, M., Estel, L.: About using microwave irradiation in competitive adsorption processes. Appl. Thermal Eng. 57(1–2), 164–171 (2013). https://doi.org/10. 1016/j.applthermaleng.2012.03.034, ISSN 1359–4311 25. Li, H., Zhao, Z., Xiouras, C., Stefanidis, G.D., Li, X., Gao, X.: Fundamentals and applications of microwave heating to chemicals separation processes. Renew. Sustain. Energy Rev. 114, 109316 (2019). https://doi.org/10.1016/j.rser.2019.109316, ISSN 1364-0321 26. Bathen, D.: Physical waves in adsorption technology—an overview. Separation Purification Technol. 33(2), 163–177 (2003). https://doi.org/10.1016/S1383-5866(03)00004-2, ISSN 1383–5866 27. Sun, Y., Zhang, B., Zheng, T., Wang, P.: Regeneration of activated carbon saturated with chloramphenicol by microwave and ultraviolet irradiation. Chem. Eng. J. 320, 264–270 (2017). ISSN 1385–8947, https://doi.org/10.1016/j.cej.2017.03.007 28. Briest, L., Wagner, R., Tretau, A., Tsotsas, E., Vorhauer-Huget, N.: Microwave-assisted drying of clay roof tiles. Drying Technol. (2021). https://doi.org/10.1080/07373937.2021.1878369/ 29. Yang, R., Chen, Q., Chen, J.: Comparison of heating performance between inverter and cycled microwave heating of foods using a coupled multiphysics-kinetic model. J. Microw. Power Electromagn. Energy. 55(1), 45–65 (2021). https://doi.org/10.1080/08327823.2021.1877244 30. Cherbanski, R.: Calculation of critical efficiency factors of microwave energy conversion into heat. Chem. Eng. Technol. 34, 2083–2090 (2011). https://doi.org/10.1002/ceat.201100405 31. Lou, B., Sha, A., Barbieri, D.M., Liu, Z., Zhang, F., Jiang, W.: Improved microwave heating uniformity and self-healing properties of steel slag asphalt containing ferrite filler. Mater. Struct. 54(1), 1–14 (2021). https://doi.org/10.1617/s11527-020-01577-7 32. Ayupov, T.A., Vorobiyov, N.G., Vorobiyova, E.G.: Allocation of electromagnetic field in a dehumidifier column for adsorbent regeneration. In: 2010 20th International Crimean Conference “Microwave & Telecommunication Technology”, Sevastopol, Ukraine, pp. 1103–1104 (2010). https://doi.org/10.1109/CRMICO.2010.5632843 33. Xiong, G., et al.: Non-thermal radiation heating synthesis of nanomaterials. Sci. Bull. 66(4), 386–406 (2021). https://doi.org/10.1016/j.scib.2020.08.037, ISSN 2095–9273

22

S. Dobrotvorskiy et al.

34. Ye,J., et al.: A hybrid ALE/implicit function method for simulating microwave heating with rotating objects of arbitrary shape. J. Food Eng. 302, 110551 (2021). https://doi.org/10.1016/ j.jfoodeng.2021.110551, ISSN 0260-8774 35. Zhu, J., Yi, L., Yang, Z., Duan, M.: Three-dimensional numerical simulation on the thermal response of oil shale subjected to microwave heating. Chem. Eng. J. 407, 127197 (2021). https://doi.org/10.1016/j.cej.2020.127197, ISSN 1385-8947 36. Singh, P., Goyal, D.K., Bansal, A.: Microwave heating: Fundasmentals and application in surface modification of metallic materials – a review. Materials Today: Proc. 43(Part 1), 564–571 (2021). ISSN 2214–7853, https://doi.org/10.1016/j.matpr.2020.12.049 37. Malhotra, A., et al.: Temperature homogeneity under selective and localized microwave heating in structured flow reactors. Ind. Eng. Chem. Res. Article ASAP. https://doi.org/10.1021/ acs.iecr.0c05580 38. Dobrotvorskiy, S., Dobrovolska, L., Basova, Y., Aleksenko, B.: Diagnostics of uneven heating of the adsorbent by microwave radiation and measures to increase the uniformity of its heating. Periodica Polytech. Chem. Eng. 63(4), 640–649 (2019). https://doi.org/10.3311/PPch.13234 39. Dobrotvorskiy, S.S., Dobrovolska, L.G., Aleksenko, B.A.: Computer simulation of the process of regenerating the adsorbent using microwave radiation in compressed air dryers. In: Hamrol, A., Ciszak, O., Legutko, S., Jurczyk, M. (eds.) Advances in Manufacturing. LNME, pp. 511– 519. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-68619-6_49 40. Dobrotvorskiy, S., Dobrovolska, L., Aleksenko, B., Basova, Y.: The use of waveguides with internal dissectors in the process of regeneration of industrial adsorbents by means of the energy of ultrahigh-frequency radiation. In: Ivanov, V., et al. (eds.) DSMIE 2018. LNME, pp. 433–442. Springer, Cham (2019). https://doi.org/10.1007/978-3-319-93587-4_45 41. Dobrotvorskiy, S., Aleksenko, B., Dobrovolska, L., Basova, Ye.: Effect of the application of microwave energy on the regeneration of the adsorbent. Acta Polytechnica 58(4), 217–225 (2018). https://doi.org/10.14311/AP.2018.58.0217 42. Dobrotvorskiy, S., Dobrovolska, L., Basova, Ye., Aleksenko, B.: Particulars of adsorbent regeneration with the use of microwave energy. Acta Polytechnica 59(1), 12–23 (2019). https://doi.org/10.14311/AP.2019.59.0012 43. Meier, M., Turner, M., Vallee, S., Conner, W.C., Lee, K.H., Yngvesson, K.S.: Microwave regeneration of zeolites in a 1 meter column. AIChE J. 55, 1906–1913 (2009). https://doi.org/ 10.1002/aic.11793 44. Silica - Silicon Dioxide (SiO2), AZoM. The leading online publication for the Materials Science community. http://www.azom.com/properties.aspx?ArticleID=1114. Accessed 20 Mar 2017 45. RF Module User’s Guide. COMSOL© (2018). https://doc.comsol.com/5.4/doc/com.comsol. help.rf/RFModuleUsersGuide.pdf. Accessed 29 Feb 2018 46. COMSOL User’s Guide. COMSOL© (2018). https://extras.csc.fi/math/comsol/3.4/doc/mul tiphysics/wwhelp/wwhimpl/js/html/wwhelp.htm?context=multiphysics&file=html_guidein tro.5.1.htm

Lifetime of Optical Fibers Submitted to Thermo-Mechanical Stresses R. El Abdi1(B) , R. Leite Pinto1 , G. Guérard2 , and C. Capena2 1 Univ. Rennes-CNRS, Institut de Physique de Rennes, UMR 6251, 35000 Rennes, France

[email protected] 2 Entreprise Acome-Usines de Mortain, 50140 Mortain, France

Abstract. The reliability and the expected lifetime of optical fibers used in telecommunication technologies are closely related to the thermo-mechanical and chemical environment actions on silica network. To ensure the long-term mechanical strength of the optical fibers, a polymer coating is applied onto fibre surface during fiber fabrication. The coating protective action includes several functions, such as: to protect glass fiber from any external damage, to limit chemical attack, in particular those of water and temperature. This paper presents the results of new silica optical fibers aged in hot water between 20 °C and 70 °C and subjected to mechanical static bending stresses from 2.5 GPa to 3.2 GPa. The mechanical strength of optical fibers aged in distilled water at different temperatures was studied. Optical fibers were then wound around ceramic mandrels with different diameters in order to evaluate the influence of the static bending stresses on the lifetime of the fiber. Dependence of the time to failure on temperature was observed and different fatigue parameters such as activation energy, fiber lifetime and stress corrosion parameter can be analyzed and a behavior law can be proposed. Keywords: Optical fibers · Thermo-mechanical stresses · Lifetime · Bending tests · Stress corrosion parameter

1 Introduction Optical fibers are key components in telecommunication technologies. Apart from optical specifications, optical fibers are expected to keep most of their physical properties for 10 to 20 years in current operating conditions. Fiber failure makes an irreversible accident which may occur when an external stress is applied on a defect located on fiber surface. While intrinsic defects, referred to as Griffith flaws, lead to failure only for large applied stresses, other heterogeneous defects make a far more serious concern as they lead to the fiber failure of the aged fibers under moderate stress. The fabrication process includes a proof test that eliminates the largest defects. However, the chemical action of water is likely to induce stress corrosion in the long term and to promote fiber aging.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 23–31, 2022. https://doi.org/10.1007/978-3-030-83368-8_3

24

R. El Abdi et al.

Polymer coatings are currently applied to optical fibers to prevent the formation of surface defects through scratches and abrasion and to minimize the influence of the preexisting defects. They also act as a diffusion barrier against the surrounding humidity reaching the glass surface. Water is known to be one major factor of the propagation of cracks at fiber glass surface because it makes much easier the breaking of the Si-O bonds which build the vitreous network [1]. Accordingly, fiber strength is closely related to the water concentration at the glass surface [2]. It is well known that flaws in glass subject to stress in humid conditions grow subcritically. Crack velocity is related to applied stress and also to relative humidity [3]. It has also been reported that the kinetics of the reaction between silica and water changes at very low water concentration [4, 5]. On the other hand, the temperature effects accentuate the damage when fibers were submitted to combined thermo-mechanical stresses. In this work, optical fibers were wound around ceramic mandrels and aged in distilled water at different temperatures to evaluate the influence of the static bending stresses and the temperature on the fiber lifetime. Change of different physical fiber parameters were analyzed with temperature variations.

2 Used Test Bench The used single mode silica optical fibers with thick dual acrylate coating have 242 μm (±5 μm) in diameter, a cladding with 125 μm (±0.7 μm) in diameter and a zero dispersion wavelength from 1300 nm to 1324 nm (zero dispersion slope ≤ 0.092 ps/nm2 .km). Optical fiber was wound onto a ceramic mandrel of 2.8, 3, 3.2 and 3.4 mm in diameter (Fig. 1). The winding of each fiber was automated and the same applied stress was obtained for each fiber. The real deformation for each fiber depends on the mandrel diameter and each fiber has taken on the exact shape of the mandrel (Fig. 1). Once the fiber was wound around the mandrel, it was placed between a transmitter T and a light receiver R (Fig. 1). The light beam cannot reach the receiver and from then on the time of fiber loading is triggered. The mean time to failure is recorded, and this corresponds to the time required for the fiber strength to degrade until it equals the stress applied through winding round the mandrel. The time to failure is measured by optical detection when the ceramic mandrel moves out of the special holder. When fiber breaks, the mandrel rocks from its vertical static position, the light beam can reach the receiver and the time to failure is directly recorded with an accuracy of ±1 s. The testing setup consists of a large number of vats containing 16 holders each. The fibers (wound around a mandrel) are aged in water at different temperatures.

Lifetime of Optical Fibers Submitted to Thermo-Mechanical Stresses

25

Fig. 1. Ceramic mandrel with wound fibers

3 Theoretical Background The Weibull theory is one of the most used methods to characterize the failure of optical fibers [6]. The statistical Weibull law gives a relationship between the probability F of fiber rupture with a length L and the applied stress σ:     1 1 Ln = m[Ln(σ ) − Ln(σo )] Ln (1) L 1−F where m is a size parameter a scale parameter.

σ o is   and 1 1 according to Ln(σ ) is called Weibull diagram. The evolution of Ln L Ln 1−F This diagram enabled us to calculate m and σ o which respectively correspond to the curve slope and to the curve intersection with the stress axis. The m parameter characterizes the defect size dispersion [7]. A high m value indicates that along the fiber the defects sizes are averaged. A low m value reveals that the defects found at the fiber surface have varying sizes, which results in different values of the failure stress. As to the σ o parameter, it represents the stress for which the cumulated rupture probability of the fiber F is equal to 50%. To plot the Weibull diagram, which corresponds to a fiber thus characterizing a defect population, a series of mechanical tests were carried out on a great number of samples of this fibre (in practice about thirty samples were tested). All the samples were the same length L. Once the tests were carried out, by ascending order: σ1 ≤ σ2 ≤ . . . ≤ σι ≤ σι+1 ≤ . . . ≤ σN

(2)

(N is the total number of tested samples). The obtained breaking stresses were classified. For each breaking stress σ i, a rupture probability Fi is affected using the following estimator. Fi =

i − 0.5 N

(3)

26

R. El Abdi et al.

If the rupture is characterized by a single defect category, Ln [(1/L) {Ln (1/(1−F))}] varies in a linear way with Ln (σ ) and Weibull modulus is given by the slope of this curve. But when the tests are carried out with large fibers lengths, one observes a slope break indicating the passage of a defect distribution to another. On the other hand, the stress-strain relation of optical fibers was first examined by Mallinder and Proctor [8]. They found that the applied stress σ (GPa) for an optical fiber is related nonlinearly to the strain ε (where both σ and ε are positive in the tensile region), according to:    α .ε (4) σ = E0 .ε 1 + 2 where E 0 is the Young modulus (= 72 GPa for the silica); α’ = 0,75 α. A value for the non-linearity constant α of 6 can be used [9]. One can note that a wound fiber was subjected to compressive stresses on its internal part (fiber surface close to the curvature center) and a tensile stress on its external part (fiber surface furthest away from the curvature center). A tensile stress is experienced by the outer surface of the fiber. The bending stress must first be calculated from the mandrel diameter φ. This can be performed with (4) and the simple expression for the maximum strain ε [10]: ε =

dglass φ + dfiber

(5)

φ is the mandrel diameter (in μm); d glace is the glass fiber diameter (125 μm); d fiber is the fiber diameter (242 μm), including the layer polymer coating. This leads, in the case of a usual fiber, to the corresponding stresses of 3.13 GPa for the calibrated diameter mandrel of 2.8 mm.

4 Results and Discussion 4.1 Environment Influence For different mandrel diameters, fibers were aged in air and distilled water at 20 °C. Figure 2 shows that the water weakens the fibers and a significant number of water molecules diffused on the cladding leading to reaction between silica network and water molecules. The ratio between times to failure for aging in the air and in the water was about 4.

Lifetime of Optical Fibers Submitted to Thermo-Mechanical Stresses

27

Fig. 2. Time to failure for fibers aged in air and in distilled water at 20 °C

4.2 Temperature Dependence of Failure Time Fibers rolled around different mandrel diameters (2.8 mm; 3.0 mm; 3.2 mm and 3.4 mm) were aged in hot water for different temperatures (20 °C, 30 °C, 40 °C, 50 °C, 60 °C and 70 °C). Weibull plot curves for water temperature of 20 °C and 70 °C are given in Figs. 3 and 4.

Fig. 3. Weibull’s diagram for fibers aged in distilled water at 20 °C

Temperature plays an essential role in fiber strength weakening. One can note for example for the mandrel diameter of 3.0 mm, the median time to failure was equal to 30 h when water temperature was 20 °C and 0.5 h when water temperature was 70 °C. Matthewson and Kurkjian [11] found that the time to failure is therefore inverse Arrhenius:   Q (6) tf = C. exp RT where Q is the apparent activation energy and R is the perfect gas constant (R = 8.314 J mol−1 K−1 ). Equation (6) shows the high sensitivity of the silica fiber strength resistance as a function of the temperature change.

28

R. El Abdi et al.

Fig. 4. Weibull’s diagram for fibers aged in distilled water at 70 °C

On the other hand, using experimental results, we can obtain the time to failure change according to the temperature (Fig. 5). For each value, the standard deviation was about 8%. Whatever the mandrel diameter, the logarithm of time to failure increases linearly according to the inverse of temperature as following:   1000 + b (7) ln(tf ) = a . T   1000. a tf = C . exp (8) T Using Eq. (6), Q values can be obtained (Q = 1000 a. R). Table 1 summarizes C and Q values. The strain slowly decreases but remains around 3.7%. The activation energy increases when stress increases from 61 kJ/mol to 82 kJ/mol.

Fig. 5. Time to failure in inverse proportion to temperature for different diameters

4.3 Change of Failure Time According to Applied Stress The fatigue curves according to the temperature were given in Fig. 6. Two remarks can be made:

Lifetime of Optical Fibers Submitted to Thermo-Mechanical Stresses

29

Table 1. Q and C values versus applied stresses. Mandrel diameter (mm)

Applied stress (GPa)

Deformation ε (%)

Q (kJ/mol)

C (h)

2.8

3.13

4.1

61.16

2.09.10–11

3

2.93

3.8

63.19

2.28 10–11

3.2

2.75

3.6

67.71

1.96 10–11

3.4

2.59

3.4

82.62

3.15 10–13

1/ The fiber lifetime roughly decreases when temperature increases: for an applied stress σ of 3.13 GPa (Ln (σ ) = 1.14 GPa), the lifetime decreases from 103 min to 2 min when the aging temperature increases from 20 °C to 70 °C. 2/ All the fatigue curves can be described using straight lines. This linear behaviour was not obtained when the temperature was greater than 80 °C. Cuellar et al. [12] observed a change in the curve slope for fibers aged at 85 °C under a stress of 2.7 GPa. But for a temperature lower than 70 °C and during all the other aging tests, the coating doesn’t undergo significant damage and the optical behaviour remains linear.

Fig. 6. Time to failure according to applied stress for different temperatures

4.4 Stress Corrosion Parameter Versus Temperature Stress corrosion parameter n is one of important values which characterize the fiber reliability. The generally accepted corrosion parameter, n, is ~20 for high strength fiber [13]. Equation (9) gives the stress corrosion parameter [6] which is the opposite of the curve slope:   (9) ln tf = −n. ln σ + a

30

R. El Abdi et al.

The n values were determined through experimental results. Figure 7 shows a decreasing behaviour according to temperature. This parameter (n) can be expressed as a linear function of the inverse of temperature (1/T ): n(T ) =

ao + bo T

(10)

where ao = 9371.5 K and bo = −8.03. A high stress parameter n involves a great sensibility to applied stresses.

Fig. 7. Stress corrosion parameter n versus inverse temperature

5 Conclusion A linear dependence of time to failure was observed for the fatigue of silica optical fibers as predicted by the exponential law. The apparent activation energy and the others fiber parameters as the stress corrosion parameter were summarized for all temperaturedependent fatigue data. This survey illustrates the temperature-dependent fatigue data for optical fibers aged in hot water and under bending stresses; these data are critical because they are a standard for all other studies.

References 1. Adams, R., McMillan, P.W.: Review static fatigue in glass. J. Mat. Sci. 12, 643–657 (1977) 2. Gurney, C.: Delayed fracture in glass, Proc. Phys. Soc., London, 59, 169–185 (1947). 3. Wiederhorn, S.M.: Influence of water vapor on crack propagation in soda-lime glass. J. Am. Ceram. Soc. 50, 407–414 (1967) 4. Duncan, W.J., France, P.W., Craig, S.P.: Strength of inorganic glass. In: Kurkjian, C.R. (ed.) The Effect of Environment on the Strength of Optical Fiber, p. 351. Plenum Press, New-York (1985) 5. Mrotek, J.L., Matthewson, M.J., Kurkjian, C.R.: Diffusion of moisture through optical fiber coatings. J. Lightwave Tech. 19(7), 988–993 (2001) 6. Pinto Leite, R.: Caractérisation physique et thermo-mécanique et analyse de la fiabilité des fibres optiques à base de silice, Thèse de doctorat de l’Université de Rennes1, Rennes, Juin (2019)

Lifetime of Optical Fibers Submitted to Thermo-Mechanical Stresses

31

7. Zhao, F.M., Okabe, T., Takeda, N.: The estimation of the statistical fiber strength by fragmentation tests of single-fiber composites. Comput. Sci. Tech. 60, 1965–1974 (2000) 8. Mallinder, F.P., Proctor, B.A.: Elastic constants of fused silica as a function of large tensile strain. Phys. Chem. Glass 5, 91–103 (1964) 9. Glaesemann, G.S., Gulati, S.T., Helfinstine, J.D.: Effect of strain and surface composition on Young’s modulus of optical fibers. Techn. Digest, 11th OFC, TUG5, 26 (1988) 10. Griffioen, W.: Optical fiber reliability. Thesis edited by Royal PTT Netherlands NV, PTT Research, Leidschendam (1994) 11. Matthewson, M.J., Kurkjian, C.R.: Environmental effects on the static fatigue of silica optical fiber. J. Am. Ceram. Soc. 71(33), 177–183 (1988) 12. Cuellar, E., Kennedy, M.T., Roberts, D.R., Ritter, J.E.: Zero stress aging and the static fatigue transition in optical fibers. In: Proceeding of the Society of Photo-Optical Instrumentation Engineers, vol. 1791, pp. 7–17 (1992) 13. Matthewson, M.J.: Optical fiber reliability models, Fiber optics reliability and testing. SPIE Critical Reviews of Optical Science and Technology, CR50, 3–31 (1993)

Design of a Mobile Robot to Work in Hospitals and Trajectory Planning Using Proposed Neural Networks Predictors Sahin ¸ Yıldırım(B)

and Sertaç Sava¸s

Mechatronics Engineering Department, Engineering Faculty, Erciyes University, 38039 Kayseri, Turkey {sahiny,sertacsavas}@erciyes.edu.tr

Abstract. Considering the intense and tiring working conditions in hospitals, healthcare personnel’s performance decreases during prolonged working times, and patients are directly affected by this decrease in performance. This study aims to design and implement a mobile robot that can help healthcare professionals improve the healthcare industry conditions. In this context, the focus is on the mobile robot performing two main tasks. The first task is dispensing medication to patients with an eight-chamber mechanical feeding unit. Thus, patients can take only their medicines from the defined reservoir by selecting their names or photos on the touch screen. The second task is to interact with patients to give moral support with phrases such as “good morning”, “you look great today”. Also, drug delivery activity is recorded in a database, and the health status of the patients can be kept under surveillance with the camera on the mobile robot. The designed mobile robot goes to the patient rooms with magnetic strip tracking. For this purpose, a controller is designed for the omni-drive robot using MATLAB, and its performance is simulated. Also, the control velocities that enable tracking the trajectories are taught to artificial neural networks (ANN), and the requirement magnetic strip for trajectory tracking is eliminated. In this direction, two artificial neural networks are compared in terms of their learning performance. Keywords: Mobile hospital robot · Omni-drive · Controller design · Trajectory tracking · Neural networks

1 Introduction Developing technology has brought robotic systems from factories to our daily lives. Mobile robots, which find many application areas, especially with their mobility features, are among the most important research topics of recent years. Until today, many types of mobile robots have been used to solve different problems. Also, robots used in mobile applications have very different motion systems [1], and the movement patterns of animals generally inspire them in their designs. Although the existence of a wheel or similar system in the animal kingdom is not complete, the wheel design is inspired by the rolling motion of some insects [2]. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 32–45, 2022. https://doi.org/10.1007/978-3-030-83368-8_4

Design of a Mobile Robot to Work in Hospitals and Trajectory Planning

33

Omni-directional wheels are used in many studies on mobile robots due to their driving flexibility. Kosuge et al. [3] developed a mobile robot called Ms. Dancer. This mobile robot was designed for dance partnership, using omni-directional wheels in the motion system. Human-robot physical interaction and coordination were investigated based on ball dances with the developed robot. This study is an excellent example of mobile robot applications’ breadth and the flexibility of omni-directional capability. Many studies have also been done on the control and trajectory tracking of omni-directional mobile robots. Wang et al. [4] used a model predictive control strategy to control an omni-directional mobile robot’s trajectory tracking. Liu et al. [5] developed a non-linear controller consisting of a two-stage structure. While the kinematic control approach is applied in the controller’s outer loop, the Trajectory Linearization Control (TLC) method is used in the inner loop, which is based on the robot’s dynamic model. Karras and Fourlas [6] designed a fault-tolerant control scheme to control an omni-directional mobile robot and accordingly developed a non-linear model predictive controller. Studies have also been carried out for different configurations of omni-directional mobile robots. Wang et al. [7] performed trajectory tracking control with the robust model predictive control (MPC) strategy for a four-mecanum-wheeled omni-directional mobile robot (FM-OMR). Cooney et al. [8] also designed a four-mecanum-wheeled omni-directional mobile robot and performed path tracking. They used an optical mouse to navigate the mobile robot with a dead-reckoning approach. Hashemi et al. [9] also designed a controller for a four-wheeled omni-directional mobile robot. In their work, they designed a model-based PI-fuzzy controller. They used the mobile robot’s linear discrete dynamic model to determine the optimum inputs. Mishra et al. [10] used the behavioral fault-tolerant-control approach for a four-mecanum-wheeled omni-directional mobile robot. Mobile robots are frequently used in the healthcare field. Tan et al. [11] developed a robotic system for transporting surgical tools in hospitals. They proposed a robust integrated system for enabling mobile robots to autonomously perform manipulation of assets. Wang et al. [12] designed a smart robotic hospital bed to transport critical patients in crowded hospital corridors safely. Again, due to their driving ability, omnidirectional wheels are frequently preferred in mobile robot designs to be used in a hospital environment. Takahashi et al. [13] designed an omni-directional mobile robot to be used in transport applications in the hospital domain. They used a four-wheel configuration in the driving system of the robot they designed. Saekow et al. [14] designed an autonomous rehabilitation robot and implemented external force/velocity control. This paper is organized into six sections. In Sect. 2, mechanical, electronic, and software designs of the mobile robot are explained. In Sect. 3, kinematics of the mobile robot is given. In Sect. 4, controller design for trajectory tracking, and in Sect. 5, ANN training of control velocities that will track the determined trajectories is mentioned, study results and findings are given. In the last section, the 6th section, the results, and suggestions are emphasized.

2 Mobile Robot Design The mobile robot will dispense medicine to patients with an eight-chamber mechanical feeding unit controlled through the interface of a touch screen. Thus, patients will be

34

S. ¸ Yıldırım and S. Sava¸s

enabled to take only their medication from the defined reservoir by selecting their names or photos on the screen. Also, the mobile robot will interact with the patients with words such as “good morning”, “you look great today” to give moral support to the patients. In addition to these basic tasks, drug delivery activity will be recorded in a database, and the general health status of the patients will be recorded with a camera on the mobile robot to ensure that the healthcare personnel keep the patients under surveillance. In this direction, the design process of the study consists of three main parts: mechanic, electronic, and software. The mechanical parts of the mobile robot consist of; • Drive system that enables navigation • A mechanism for drug delivery • The general construction of the robot The mobile robot is driven by three omni-directional wheels to be placed at 120° angles and three different servo motors for each wheel. These wheels provide the robot omni-drive. Also, these wheels are specially designed and manufactured to increase the ground grip and increase the carrying capacity of the mobile robot. Omni-directional wheels and servo motors manufactured to be used in the mobile robot’s driving system are shown in Fig. 1.

Fig. 1. Omni-directional wheels and servo motors used in mobile robot driving system.

Also, a rotating mechanism with eight chambers that can deliver drugs to eight different patients is designed. After the patient selection is made on the robot screen, the

Design of a Mobile Robot to Work in Hospitals and Trajectory Planning

35

relevant chamber is positioned in front, and its cover is opened. The CAD (computeraided design) model of the drug delivery mechanism is given in Fig. 2. Figure 3 shows the drug delivery slot on the prototype mobile robot. The outer shell of the mobile robot is made of fiberglass. Figure 3 also shows the CAD model of the mobile robot.

Fig. 2. Mobile robot drug delivery mechanism CAD model.

Fig. 3. CAD model of the mobile robot and the produced prototype model.

Mobile robot electronic equipment; It consists of a control card, communication hardware, sensors placed under the mobile robot to detect magnetic lines, motor drivers,

36

S. ¸ Yıldırım and S. Sava¸s

two batteries, an LCD touch screen, two speakers for audio notifications, a camera placed between the speakers and a charging station where it will be charged in standby mode. The third stage of the design is developing the software; navigation software to go to patient rooms, software to provide medication distribution, database software to keep records for transactions, and communication software to communicate with a computer when necessary.

3 Kinematics of Three-Wheel Omnidirectional Drive Three omni-directional wheel configuration used in the mobile robot drive system is given in Fig. 4. In  order to make inverse kinematic calculations [1] in global coordinates, ˙ of the mobile robot must be defined. the linear (v = x˙ 2 + y˙ 2 ) and angular velocity (ϕ) The velocity of each wheel consists of tangential and radial components. First wheel velocity components and total velocity are given in Eq. 1–4. x˙ and y˙ are the components of the linear velocity in the x and y axes, and R is the distance between the center of the mobile robot and the wheels. v1 = v1t + v1r

(1)

v1t = −˙x sin(ϕ) + y˙ cos(ϕ)

(2)

v1r = Rϕ˙

(3)

v1 = −˙x sin(ϕ) + y˙ cos(ϕ) + Rϕ˙

(4)

Fig. 4. Schematic model of three-wheel omnidirectional drive.

Since the wheels are placed at 120°, θ2 is 120° and θ3 is 240°. Similarly, the second and third wheel velocities are given in Eq. 5–6. Inverse kinematics (v = J q˙ ) in global coordinates is expressed as in Eq. 7. v2 = −˙x sin(ϕ + θ2 ) + y˙ cos(ϕ + θ2 ) + Rϕ˙

(5)

Design of a Mobile Robot to Work in Hospitals and Trajectory Planning

37

v3 = −˙x sin(ϕ + θ3 ) + y˙ cos(ϕ + θ3 ) + Rϕ˙

(6)

⎤ ⎡ ⎤⎡ ⎤ − sin(ϕ) cos(ϕ) R x˙ v1 ⎣ v2 ⎦ = ⎣ − sin(ϕ + θ2 ) cos(ϕ + θ2 ) R ⎦⎣ y˙ ⎦ v3 − sin(ϕ + θ3 ) cos(ϕ + θ3 ) R ϕ˙

(7)

⎡

For mobile robot control, the velocity components (˙qm ) in local coordinates must be defined. In this direction, forward kinematic equations (˙q = Sv) are found by taking the inverse of the Jacobian matrix (˙q = Sv). Finally, the mobile robot omni-drive forward kinematic equation [1] is given in Eq. 8. ⎤⎡ ⎤ ⎡ ⎡ ⎤ v1 − sin(θ1 ) − sin(θ1 + θ2 ) − sin(θ1 + θ3 ) x˙ ⎣ y˙ ⎦ = 2 ⎣ cos(θ1 ) cos(θ1 + θ2 ) cos(θ1 + θ3 ) ⎦⎣ v2 ⎦ 3 1 1 1 v3 ϕ˙ 2R 2R 2R

(8)

4 Controller Design The mobile robot is modeled with MATLAB Simulink, and a controller is designed for line tracking. MathWorks Mobile Robotics Simulation Toolbox and Mobile Robotics Training Toolbox are used in the simulation studies. A PD controller is used as the controller, and the mathematical expression of the controller is given in Eq. 9. Simulink block diagram of the mobile robot control loop is also shown in Fig. 5. PD(s) = P + D

N 1 1 + N · Ts z−1

Fig. 5. Simulink block diagram of the mobile robot control.

(9)

38

S. ¸ Yıldırım and S. Sava¸s

For the hospital scenario to be applied, the map in Fig. 6 is designed, and the tracking performances are examined for two trajectories between “Starting Point - Patient Room 1” and “Patient Room 1 - Patient Room 2”. Mobile robot and controller parameters used in the study are given in Table 1.

Fig. 6. The map designed for the hospital scenario.

Table 1. Mobil robot and control parameters. Parameter

Values

Wheel Radius [m]

0.125

Robot Radius [m]

0.4

Wheel Angles [rd] Sample Time = Ts [sec]

0.01

PD Controller Parameters Filter Coefficient (N)

100

Start Point (SP) (charging station)

X Position [m] = 1

Patient Room 1 (R1)

X Position [m] = 7

Y Position [m] = 5 Y Position [m] = 1

Patient Room 2 (R2)

X Position [m] = 11

Y Position [m] = 1

The trajectory tracked by the PD controller between SP-R1 is given in Fig. 7, the reference heading angle and mobile robot heading angle on this trajectory are given in Fig. 8, and the mobile robot posture error values are given in Fig. 9. The trajectory tracked between R1–R2 is given in Fig. 10, the reference heading angle and mobile robot heading angle on this trajectory are given in Fig. 11, and the mobile robot posture error values are given in Fig. 12. In addition to these studies, the mobile robot’s scenario tests installed in Erciyes University, Mechatronics Engineering Department Laboratory are also performed. In this scenario, room navigation is carried out with magnetic strips placed on the floor and sensors that detect these magnetic strips.

Design of a Mobile Robot to Work in Hospitals and Trajectory Planning

Fig. 7. Reference and PD controlled trajectory between SP-R1.

Fig. 8. Reference and PD controlled heading angle between SP-R1.

Fig. 9. PD controlled mobile robot posture errors for trajectory between SP-R1.

39

40

S. ¸ Yıldırım and S. Sava¸s

Fig. 10. Reference and PD controlled trajectory between R1–R2.

Fig. 11. Reference and PD controlled heading angle between R1–R2.

Fig. 12. PD controlled mobile robot posture errors for trajectory between R1–R2.

Design of a Mobile Robot to Work in Hospitals and Trajectory Planning

41

5 Neural Network Training The system input is the linear and angular control velocity values “vc and wc ” and the output of the designed control loop is the mobile robot posture “x, y, and θ”. The position values of two different trajectories tracked by the mobile robot and the control velocity values required to go to these positions are taught to the artificial neural networks. In this way, the mobile robot’s trajectories will be taught once with line tracking, and then the robot will be able to operate without the need for magnetic strips. Multilayer Perceptron (MLP) and Radial Basis Neural Network (RBNN) types are used for this purpose, frequently used in system modeling and fitting processes in artificial neural network training. Both network models are feed-forward neural networks with 3 inputs and 2 outputs and consist of a single hidden layer. The training process with artificial neural network and system inputs and outputs are given schematically in Fig. 13.

Fig. 13. Training process block diagram with artificial neural network and system inputs and outputs.

Training processes made for MLP are carried out for 3 different training algorithms (Levenberg-Marquardt, Bayesian Regularization, Scaled Conjugate Gradient) and 3 different activation functions (logsig, purelin, and tansig). Also, the number of neurons (H) in the hidden layer is changed between 1–100. Training parameters for MLP are given in Table 2. The lowest error value of each combination is given in Table 3. Network and training parameters in the 6th row, which has the lowest MSE value, are selected. Table 2. MLP training parameters. Parameter

Value

Performance function

MSE

Maximum number of iterations (epochs) 1000 Target error value

1 × 10–10

Learning rate (mu)

0.1

Number of iterations to show results

25

42

S. ¸ Yıldırım and S. Sava¸s Table 3. MSE values for different MLP parameters.

Training Algorithm

Activation Functions

Levenberg-Marquardt Levenberg-Marquardt Levenberg-Marquardt Bayesian Regularization Bayesian Regularization Bayesian Regularization Scaled Conjugate Gradient Scaled Conjugate Gradient Scaled Conjugate Gradient

logsig purelin tansig logsig purelin tansig logsig purelin tansig

Number of Neurons I H O 3 59 2 3 75 2 3 36 2 3 80 2 3 30 2 3 62 2 3 73 2 3 46 2 3 75 2

MSE 3,0671 x 10-3 5,7089 x 10-3 0,1105 x 10-3 3,0698x 10-3 5,71x 10-3 0.0809 x 10-3 3,1519 x 10-3 5,7091 x 10-3 0,2358 x 10-3

Training operations for RBNN are performed by applying different spread values. The MSE values are given in Table 4. The spread value in the 3rd row, which gives the lowest training error in the table, is selected for comparison. Table 4. MSE values for different spreads of RBNN.

Spread 0.0001 0.001 0.01 0.1 1 10 100

MSE 1.9408 x 10-31 1.8743 x 10-31 0.4235 x 10-31 1.3658 x 10-8 1.2552 x 10-4 5.3398 x 10-4 2.1745 x 10-3

The control velocity outputs of MLP and RBNN networks trained with network parameters giving the lowest MSE values are compared with the actual velocity outputs. Actual linear control velocity graph for trajectory tracking between SP-R1 and linear control velocity graphs generated with MLP and RBNN are given in Fig. 14, and graphs for angular control velocity are given in Fig. 15. Likewise, the actual linear control velocity graph for trajectory tracking between R1–R2 and the linear control velocity graphs generated with MLP and RBNN are given in Fig. 16, and the graphs for angular control velocity are given in Fig. 17. As shown from the figures, RBNN has learned the actual control velocity values much more successfully.

Design of a Mobile Robot to Work in Hospitals and Trajectory Planning

43

Fig. 14. Linear control velocities generated with MLP and RBNN for trajectory tracking between SP-R1.

Fig. 15. Angular control velocities generated with MLP and RBNN for trajectory tracking between SP-R1.

Fig. 16. Linear control velocities generated with MLP and RBNN for trajectory tracking between R1–R2.

Fig. 17. Angular control velocities generated with MLP and RBNN for trajectory tracking between R1–R2.

44

S. ¸ Yıldırım and S. Sava¸s

6 Conclusion and Future Research In this study, a mobile robot that can go to the patient room and automatically do some simple tasks performed by nurses or caregivers in a hospital domain is designed, and a prototype is produced. The main purpose of performing these tasks with the mobile robot is to eliminate the confusion that may occur during drug delivery to the patients, make this process automatic and recordable, and keep the patient’s morale at the highest level by the robot’s voice interaction. Also, a controller is designed to track the trajectories that will provide navigation to the patient rooms on the developed scenario map, and simulations are carried out with MATLAB. The position values of the trajectories tracked by the mobile robot and the control velocity values required to go these positions are taught to two different artificial neural networks. Thus, trajectories are taught once with line tracking, and then the robot is enabled to navigate without the need for lines. Artificial neural network studies are also revealed that the RBNN type network is more successful. In future studies, the controller design can be improved, and different control structures can be tested. Besides, mass production can be started for the mobile robot. Acknowledgments. This research results consisted of a part of project number FOA-2013–4150. The authors would like to express their thanks to Erciyes University-Scientific Research Projects Unit (BAP) for supporting this project.

References 1. Klancar, G., Zdesar, A., Blazic, S., Skrjanc, I.: Wheeled Mobile Robotics: From Fundamentals Towards Autonomous Systems. Butterworth-Heinemann, Oxford (2017) 2. Vepa, R.: Biomimetic Robotics Mechanism and Control. Cambridge University Press, Cambridge, p. 343 (2009) 3. Kosuge, K., Hayashi, T., Hirata, Y., Tobiyama, R.: Dance partner robot - Ms DanceR. In: IEEE International Conference on Intelligent Robots and Systems, vol. 4, pp. 3459–3464 (2003) 4. Wang, C., Liu, X., Yang, X., Hu, F., Jiang, A., Yang, C.: Trajectory tracking of an omnidirectional wheeled mobile robot using a model predictive control strategy. Appl. Sci. (Switzerland) 8(2), 231 (2018) 5. Liu, Y., Zhu, J.J., Williams, R.L., Wu, J.: Omni-directional mobile robot controller based on trajectory linearization. Robot. Auton. Syst. 56(5), 461–479 (2008) 6. Karras, G.C., Fourlas, G.K.: Model predictive fault tolerant control for omni-directional mobile robots. J. Intell. Robot. Syst.: Theory Appl. 97(3–4), 635–655 (2020) 7. Wang, D., Wei, W., Yeboah, Y., Li, Y., Gao, Y.: A robust model predictive control strategy for trajectory tracking of omni-directional mobile robots. J. Intell. Robot. Syst.: Theory Appl. 98(2), 439–453 (2020) 8. Cooney, J.A., Xu, W.L., Bright, G.: Visual dead-reckoning for motion control of a Mecanumwheeled mobile robot. Mechatronics 14(6), 623–637 (2004) 9. Hashemi, E., Ghaffari Jadidi, M., Ghaffari Jadidi, N.: Model-based PI-fuzzy control of fourwheeled omni-directional mobile robots. Robot. Auton. Syst. 59(11), 930–942 (2011) 10. Mishra, S., Sharma, M., Santhakumar, M.: Behavioural Fault-tolerant-control of an omnidirectional mobile robot with four-mecanum wheels. Def. Sci. J. 69(4), 353–360 (2019) 11. Tan, H., Mao, Y., Xu, Y., Kannan, B., Griffin, W.B., Derose, L.: An integrated robotic system for transporting surgical tools in hospitals. In: 10th Annual International Systems Conference, SysCon 2016 - Proceedings (2016)

Design of a Mobile Robot to Work in Hospitals and Trajectory Planning

45

12. Wang, C., Savkin, A.V., Clout, R., Nguyen, H.T.: An intelligent robotic hospital bed for safe transportation of critical neurosurgery patients along crowded hospital corridors. IEEE Trans. Neural Syst. Rehabil. Eng. 23(5), 744–754 (2015) 13. Takahashi, M., Suzuki, T., Shitamoto, H., Moriguchi, T., Yoshida, K.: Developing a mobile robot for transport applications in the hospital domain. Robot. Auton. Syst. 58(7), 889–899 (2010) 14. Saekow, P., Neranon, P., Smithmaitrie, P.: External force/velocity control for an autonomous rehabilitation robot. In: IOP Conference Series: Materials Science and Engineering, vol. 297, no. 1 (2018)

Design and Measurement of the Peltier Cell Thermal Actuator for Fine Adjustment Jan Hošek(B) Faculty of Mechanical Engineering, Czech Technical University in Prague, Technická 4, Prague 6, Czech Republic [email protected]

Abstract. The article presents design and experimental validation of a thermal actuator based on peltier cell control. The experiments showed the possibility of fine adjustment of large-scale objects in the working range of 24 µm with a submicrometer resolution capability. We have proved the main body of the actuator is not affected with heat effects during the actuator control. Keywords: Thermal actuator · Peltier cell · Fine adjustment

1 Introduction Fine resolution mechanical actuation is important for a variety of applications in instrumentation. The standard solution for a precise actuation employs flexure systems actuated with piezo actuators. The drawback of using piezo chips or stack actuators is the limited range of motion which can be extended with the appropriate design of lever-based flexure mechanisms [1, 2]. Besides piezo actuators, voice coil actuators can be used for a longer-range motion of actuation [3]. Those and other sophisticated actuators are relatively complicated with expensive power and control systems needed for their actuation. There can be used much simpler physical principles for the desired fine actuation. A very simple principle for a precise actuation provides thermal effects applied to the part material. Thermal effects usually affect the part dimensions, which can be used for the part positioning or actuation. Thermal actuation has been limited in use for a long time due to the fact that the motion provides a long time constant in case of massive objects. This drawback is negligible in small scale micro applications, where thermal actuators can be found for a long time [4]. The application of thermal actuators in a small scale provides variety of motion realization geometries [5], employing variable materials [6] or actuator heating method [7]. Thermal actuation in larger scales has been developed simultaneously where the most common applications can be found in e.g. thermostatic valves [8] where longer time constant is not crucial. There can be found even more effective ways how to increases the motion range or motion direction possibilities with the help of employment of the material phase transitions [9–12]. Our aim was to design a simple linear actuator with the travel range of 20 µm. It is beyond the displacement possibilities of simple piezoelectric chips, but it can be © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 46–53, 2022. https://doi.org/10.1007/978-3-030-83368-8_5

Design and Measurement of the Peltier Cell Thermal Actuator

47

realized with the help of piezoelectric stacks. Due to the piezoelectric actuators’ strong hysteresis and non-linearity in a long-range application we have been looking for other actuation possibilities. For our application, an interesting alternative to piezoelectric actuators seems to be a peltier effect based thermoelectric actuator [13]. We proposed and designed a linear actuator based on flexure hinges where the actuator top platform position has to be controlled. We modified the actuator structure to place the peltier cell for a thermal change actuation. Next sections describe the actuator geometry, performance of the peltier heating experiment, and measured actuator positioning capabilities. We also checked the influence of dissipated heat energy of the peltier cell on the actuator base structure.

2 The Cell Actuator Geometry The aim of the proposed actuator is an approximate linear motion in Z axis with the travel range of 20 µm. The approximate linear motion is achieved with a parallelogram flexure structure of appropriate stiffness to allow the whole range motion without reaching the yield stress limit within the structure. This was achieved by the design of two lever arms parallelogram of the length 12.5 mm and the minimum flexure thickness of 0.3 mm. This allows to reach maximum deformation up to 70 µm without reaching the yield stress limit for the assumed aluminum alloy 6061(T4). 2.1 The Actuator Thermal Properties The motion function of the actuator is due to the dilatation of the heated wall in Z axis direction. The heated wall is 1 mm thick and lo = 26 mm long. Under the assumption of a uniform temperature distribution over the heated wall it can be determined the temperature difference needed for the given actuation range according to the thermal expansion relation: l = lo .α.T ,

(1)

where l is the dilatation of the heated wall, α = 24.10–6 K−1 is the thermal expansion coefficient of the 6061(T4) aluminum alloy, and T is the temperature difference needed for reaching the desired heated wall dilatation. The dynamics of extension can be expressed as a solution of Newton’s cooling equation: τ

T (τ ) = Te + (To − Te )e− k ,

(2)

where T (τ ) is the heated wall temperature value in time τ, T o is the body temperature in time τ = 0, T e is the environment temperature, and k is the time constant of the system. The time constant of the system can be estimated with the relation: k=

c.m , S.α¯

(3)

where c = 896 J.kg−1 .K−1 is the 6061(T4) aluminum alloy thermal capacity, m = 1.12 g is the mass of the heated wall, S = 884 mm2 is the heat exchange area, and a¯ is the average

48

J. Hošek

heat transfer coefficient. The uncertainty of the heat transfer coefficient is high compared to the other parameters. If the wall heating is achieved with an evenly winded resistance wire we can assume the heat exchange surface is exposed to free air. In such a case we can assume the average heat transfer coefficient a¯ = 6, 4 W.m−2 .K−1 . Corresponding steady state temperature power loss can be expressed according to the relation P = a¯ .S,

(4)

what gives the value of 5.6 mW/K. It allows us to determine the necessary power to resistance wire to reach the desired temperature change. Steady state is achieved in 3–5 time constants, what can be evaluated according relation (3) to give k = 177 s. That is inconveniently long time for an effective action of the actuator, especially in case of its cooling, what cannot be accelerated with increases of heating power as in case of heating up sequence. This long time constant drawback is ordinarily solved by additional cooling system, which would increase the complexity of the proposed actuator. We tried to replace the simple resistive wire with a peltier cell as a heat source, because the peltier cell could shorten the cooling time constant due to the presence of cold side of the peltier cell close to the heated wall. It could lead to its fast cooling when the power to the peltier cell is switched off. 2.2 The Actuator Geometry Modification for the Peltier Cell We modified the actuator geometry to fit an appropriate peltier cell. The actuator geometry model and the peltier cell is shown in Fig. 1

Fig. 1. A 3D model of the peltier cell thermal actuator geometry.

We used the STONECOLD TEC1-01708 peltier cell of the size of 15 × 15 × 3.7 mm3 , providing the maximum power of 9.5 W. The peltier cell is placed to contact with a 1 mm thick wall in Z axis direction intended for temperature change with help

Design and Measurement of the Peltier Cell Thermal Actuator

49

of the heat conductive paste Keratherm KP 98. To achieve a temperature gradient over the peltier cell the opposite cell side is provided with an aluminum cooler of the size of 20 × 20 × 10 mm3 . To prevent the heat dissipation to the actuator’s structure the heated wall is provided with thermal baffles increasing the wall thermal-exchange surface. The opposite side of the heater wall was provided with 3D printed part serving as a heat insulator and thermocouple holder.

3 The Experiment Description and Performance The aim of the experiment was to prove the properties and positioning resolution of the peltier cell thermal actuator. It consisted of the measurement of the actuator top platform position and the heated wall temperature as a function of the input power to the peltier cell. For the temperature measurement we used a miniature thermistor NTC 10 kW - AVX NC20K00103JBA and the thermocouple SMD0603 Pt 1000 - Heraeus Nexensos 32207638 placed in contact with the heated wall under the 3D printed holder and to the center of the peltier cell cold side cooler, respectively. The temperature was measured with a digital multimeter Truevolt 34461 A Keysight Technologies. The power to the peltier cell can be controlled with a digital thermostat REES W1209 or directly with a lab power source Diametral P230R51D. The actuator top platform position was measured with a triangulation sensor optoNCDT-1402 achieving the resolution of 0.3 µm within working range 5 mm under sampling frequency 50 Hz. The overall view to the experimental setup is shown in Fig. 2.

Fig. 2. The experimental setup.

50

J. Hošek

3.1 The Time Constant Determination The first aim of the experiment was to determine the time constant of the peltier cell thermal actuator. The peltier cell can be powered by 2 V with max 4 A current. The time constant measurement was performed by heating up the peltier cell with a constant current until the steady temperature compensating the system heat loses is reached. The time constant was determined as a time corresponding to the 63.2% of the final temperature change. An example of measured data during the time constant measurement is shown in Fig. 3.

Fig. 3. A time constant measurement data with two current steps 2.88 A and 3.44 A.

The presented data shows current step reaction of the system. The top platform position corresponds well to the temperature data. In this particular case the electrical current 2.88 A causes the temperature change 11 °C within 100 s the top platform’s position change of 12 µm. Data assessment showed the average heating time constant is 15.3 s and the corresponding system cooling time constant is 14.2 s. It proved the decrease of the system time constant when a peltier cell is employed compared to a conventional approach. 3.2 The Positioning Characteristics Determination The next aim was to prove the set position’s stability and the repeatability of the position setting within the whole range of the peltier cell allowed by power electrical currents. We varied the current level after the steady temperature/position was reached. Position data showed submictometer long term stability of the set position within time of hundreds of seconds where data vary just between two output data points of the positioning sensor. Uncertainty region related to ±3 standard deviations corresponds to 1.2 µm for single measurement cycle. An example of measured data of the top platform position as a function of the peltier cell input current is shown in Fig. 4.

Design and Measurement of the Peltier Cell Thermal Actuator

51

Fig. 4. The actuator top platform’s position measured data as a function of input current.

It is clearly visible that the position characteristics correlate with the second power of input current what corresponds to dissipated power on peltier cell. The graph also shows that the position may deviate from the ideal characteristics especially for higher currents. Even higher deviations up to 3 µm position difference can be found among experiments performed in different experimental days. We have found out that those deviations are closely related to the change of the thermal conductivity within the contact of thermal exchange surfaces. But if there is good contact between the peltier cell’s and the heated wall’s surface the actuator positioning is well repeatable within micrometer uncertainty level. 3.3 The Heat Influence to the Actuator Structure We have measured the temperatures at both sides of the peltier cell. The temperature data of the heated wall shows a nice linear relation to the peltier cell input power. It allows us to evaluate the steady state temperature power loss of the actuator to be 0.37 W/K. This knowledge can be used for the control of the actuator. We have also measured the temperature in the middle of the cooler at the cold side of the peltier cell. It follows an exponential decay where the original ambient temperature steeply decreased for low peltier cell input power. This exponential cooling trend was stopped for the input power above 4.5 W. The temperature of the cold side peltier cell cooler stabilized at temperature 1.5 °C what provides temperature difference to the surrounding air temperature −20.5 °C. We were interested whether those temperature gradients affect the ground structure of the thermal actuator. We checked the whole actuator temperature distribution with a Flir i50 IR camera. Examples of the IR camera measurement is shown in Fig. 5.

52

J. Hošek

Fig. 5. IR camera snapshots for ambient temperature - left and for the maximum reached temperature - right.

The figure shows that even for the highest actuation temperature the temperature of the actuator structure keeps the same level. Just reflections at the surfaces of the actuator body are visible and the hot heated wall is almost masked from the actuator structure.

4 Conclusions We have proposed and tested a new way of a thermal actuator design using the peltier cell heating. The presence of cold side of the peltier cell to the heated wall shortens the cooling time constant to a fraction of the expected value in case of free air cooling, when the power to the peltier cell is switched off. The tested actuator’s geometry showed the average heating time constant is 15.3 s and the corresponding system cooling time constant is 14.2 s. The actuator system showed a capability to reach long enough motion range up to 24 µm and a good position stability at sub-micrometer level. A long term position repeatability of the actuator is mainly affected by keeping good thermo-mechanical contact among the heat exchange surfaces. We have proved the actuator position can be effectively controlled as a function of the second power of peltier cell’s input electrical current or as linearly proportional to the heated wall temperature measurement. We also tested the influence of the temperature gradient over the peltier cell to the holding structure of the actuator flexure with an IR camera. The measurement showed no influence of the heated wall even for the highest temperature change corresponding to 21 °C and holding this temperature for time 300 s.

References 1. Ling, M., Cao, J., Jiang, Z., Lin, J.: Theoretical modeling of attenuated displacement amplification for multistage compliant mechanism and its application. Sens. Actuators A Phys. 249, 15–22 (2016)

Design and Measurement of the Peltier Cell Thermal Actuator

53

2. Lai, L.J., Zhu, Z.N.S.: Design, modeling and testing of a novel flexure-based displacement amplification mechanism. Sens. Actuators A: Phys. 266, 122–129 (2017) 3. Wang, P., Xu, Q.: Design of a flexure-based constant-force XY precision positioning stage. Mech. Mach. Theory 108, 1–13 (2017) 4. Van de Pool, F.C.M., Wonnink, D.G.J., Elwenspoek, M.: A thermo-pneumatic actuation principle for a microminiature pump and other micromechanical devices. Sens. Actuators 17, 139–143 (1989) 5. Potenkina, A., Wang, C.: Review of electrothermal actuators and applications. Actuators 8, 69 (2019) 6. Deng, K., Rohn, M., Guenther, M., Gerlach, G.: Thermal microactuator based on temperature sensitive hydrogels. Procedia Eng. 12, 57–62 (2015) 7. Zheng, D., Zhang, H., Liu, C., Jiang, J.: Theoretical and experimental study of optothermal expansion and optothermal microactuator. Opt. Express 16(17), 13476–13485 (2008) 8. Melone, S.J.: Valves having a thermostatic actuator controlled by a Peltier device. Patent US20050056799A1 (2005) 9. Kim, T.H., et al.: Biomimetic thermal-sensitive multi-transform actuator. Sci. Rep. 9, 1–7 (2019). Article number: 7905 10. Mertmann, M., Vergani, G.: Design and application of shape memory actuators. Eur. Phys. J. Spec. Top. 158, 221–230 (2008) 11. Benafan, O., et al.: Shape memory alloy actuator design: CASMART collaborative best practices and case studies. Int. J. Mech. Mater. Des. 10(1), 1–42 (2013). https://doi.org/10.1007/ s10999-013-9227-9 12. Mann, A., Germann, T., Ruiter, M., Groche, P.: The challenge of upscaling paraffin wax actuators. Mater. Des. 190, 108580 (2020) 13. Neto, J.V.F., de Sá, D.F.S., Lopes, L.R.: State space modeling of thermal actuators based on peltier cells for indirect measurements and optimal control. In: Al-Dabass, D., Orsoni, A., Cant, R., Abraham, A. (eds.) UKSim 12th International Conference on Computer Modelling and Simulation, pp. 392–397. IEEE (2010)

S-Shape Feedrate Scheduling Method with Smoothly-Limited Jerk in Cyber-Physical Systems Volodymyr Kombarov1 , Volodymyr Sorokin1 , Yevgen Tsegelnyk2(B) Sergiy Plankovskyy2 , Yevhen Aksonov2 , and Olena Fojt˚u3

,

1 National Aerospace University “Kharkiv Aviation Institute”,

17 Chkalova Street, Kharkiv 61070, Ukraine 2 O. M. Beketov National University of Urban Economy in Kharkiv,

17 Marshala Bazhanova Street, Kharkiv 61002, Ukraine [email protected] 3 Research Center of Manufacturing Technology, Czech Technical University in Prague, 3 Horska Street, 128 00 Prague 2, Czech Republic

Abstract. For the successful implementation of innovative automated technological processes, such as high speed machining, 3D laser welding, 3D laser cutting and laser cladding it is necessary to provide highly effective control of the working tool movement speed in multi-axis space of cyber-physical system. A simplified three-interval model S-shaped feedrate profile with smoothly-limited jerk using the sin2 function are proposed. An algorithm for pre-scheduling of the feedrate before the start of motion control, which has no constraints on the number of analyzed blocks and the scheduling execution time is proposed. Modeling the scheduling process on the example of a test trajectory has shown the high efficiency of the proposed model. Reduction of the processing time of the control section of the trajectory by 7.4% to 10.7% in comparison with the Heidenhain iTNC530 system with the same acceleration and jerk limitations are obtained. Keywords: CNC · S-shaped feedrate profile · Smoothly-limited jerk · Snap

1 Introduction One of the most challenges in modern industry is improving the accuracy, reliability, and effectiveness of cyber-physical systems such as high-speed machining [1, 2], laser cutting [3, 4], welding [5, 6], etc. A modern machine tool with a numerical control system (CNC) is a technological system that can have various functions for implementing various high-speed processes. When moving with high feeds, due to the influence of inertia forces, the machine bodies cannot immediately respond to variations in the tool path in sections with bends in the path [7]. To avoid inertial casts during the variation of motion direction, modern CNC systems use intelligent feed scheduling methods implemented by Look-ahead algorithms [8, 9]. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 54–68, 2022. https://doi.org/10.1007/978-3-030-83368-8_6

S-Shape Feedrate Scheduling Method

55

To ensure high efficiency of processing, studies are being carried out to simulate various processes and determine their optimal parameters [10–14], while the real geometry of the parts being manufactured is the reason for the formation of complex curved processing paths along which it is not possible to move with a constant feed. To form the optimal treatment conditions need to be considered constraints that arise when moving on a curved path. When scheduling feedrates in real time, there are severe processing time constraints due to control cycle time constraints. Real-time scheduling can take up most of the control cycle time. In addition, it is not necessary to analyze the number of blocks, but a section of the trajectory of a certain length. An alternative to real-time feedrates scheduling methods is to pre-scheduling feedrates before motion control begins. Feedrate scheduling is carried out for the tool path specified in the NC program and unchanged during its execution. The absence of time constraints when performing feedrate scheduling allows using of the most complete set of constraints when calculating the allowable motion feedrates in each block, to apply S-shaped acceleration/deceleration laws that ensure smooth changes in the differential kinematic characteristics. In this paper for simplicity of calculations in the CNC system, a three-interval model S-shaped feedrate profile with smoothly-limited jerk is considered.

2 The Concept of Feedrate Pre-scheduling and Control in a Two-Level CNC System One of the most common CNC architectures is a two-level system structure. In Fig. 1 shown the main stages of data exchange in such a CNC system. At the application level, the architecture of the CNC system is determined both by the composition of application sections, which are called control tasks, and by the need to execute them in real time [8, 9, 15].

Represent as a spline Approximation Vmax, Amax, Jmax Parameters limitation

XS

VXS

%F

US XF VXF

High-speed equipment

Real-time PC

Position regulator

Linearcircular

Non-real-time PC

Trajectory interpolation

Spline

Spline-interpolator in CNC

Feed generation F(t)

CAD/CAM Trajectory setting

Feedrate correction

Fig. 1. Generalized structure of a two-level CNC system.

Such tasks include: geometrical (servo drives control); logical (electro-automatic control); technological (maintenance or optimization of the technological process parameters); dispatching (control of other tasks at the application level); terminal (dialogue

56

V. Kombarov et al.

with the operator, system states displaying, NC programs editing and verification); communication (information exchange with the enterprise management system). The hardware of the complex is a two-level CNC system, in which tasks control is distributed between two PCs (Fig. 1). Communication between PCs builds on industrial IoT technology. Operator panel unit contains a PC, which operates by Windows operation system and solves tasks that do not require real-time mode. The real-time unit contains a CPU module and FPGA-based specialized modules connected to the internal bus: discrete signal input-output modules, motor controllers, feedback modules, and solves tasks for which real-time mode is crucial. Real-time block comprises a CPU module and specialized FPGA-based modules, connected to the internal bus: the input/output modules, motor controllers, and feedback modules. It is known that the CAD/CAM system calculates the tool paths and create the program in the form of the G-code, which does not allow to describe the federate variation along the length of the block. This problem is solved by the CNC system using look ahead algorithms. Note that all the parameters which determine the necessity for federate variations along the toolpath are uniquely defined and known to the CNC system before starting the execution of the NC program. The toolpath is defined by the NC program. For each section of the toolpath is known a priori geometrical parameters of the tool. The dynamic characteristics of the servo drive for a specific machine are defined and set in the CNC system settings. The mathematical parameters of the S-shaped feedrate profile are also predefined in the CNC system settings. In this regard, it can be argued that the task of constructing the federate variation scheduling along toolpath can be solved in advance to prevent the real-time CPU module utilization by calculating the allowable feedrates of the tool movement. The concept of pre-scheduling provides the separation by the stages on a schedule of the federate variation along toolpath (F C (L), Fig. 1) and the implementation of real-time feed control (F(t), Fig. 1) [15]. The implementation of the pre-scheduling feedrate concept in contrast to the look ahead concept, which is implemented in real-time, allows to perform feed analysis for an unlimited sections of the toolpath up to the entire NC program. Execution of scheduling before the control beginning allows avoiding restrictions on time of data processing and the size of the preview buffer. The absence of constraints on the time and amount of data processing allows analyzing any complexity of the toolpath. The calculation of the feedrate F(t) in real-time is carried out using tabular or analytically represented Sshaped feedrate profile. The input data for such calculations is a schedule of the feedrate variation along toolpath length F C (L).

3 Methods of the Feedrate Schedule Variation Set in Blocks of NC Program Feedrate is programmed in the text of the NC program at address F for the sequence of blocks or each block separately. The program does not contain information about the method and behavior of feedrate during the transition from block to block or federate variation during the block. In the CNC system during the process of compiling the source

S-Shape Feedrate Scheduling Method

57

text of the NC program implemented text translation software, represented as G-code into the internal format. At the stage of the internal representation of the NC program formation based on the analysis of the trajectory shape subsystem scheduling feedrate has the ability to calculate for each block schedule to feedrate variation F C (L). This takes into account the constraints of the kinematic and dynamic characteristics of a machine tool. It is possible to use a parametric or functional model of setting the feedrate variation schedule in the block of the NC program [8, 9, 16, 17]. Blocks of a linear-circular trajectory are characterized by the constancy of geometric characteristics along the block length. This allows us to set the movement in the block with a constant feedrate, and the diagram of the feedrate variation in the block sets two parameters: the feed start (F BSi ) and end (F BEi ) of the i block. Path sections of feedrate variations implemented within the block using smooth S-shaped laws of acceleration/deceleration. The blocks of the toolpath, given by involutes, splines, and other curves, are characterized by a continuous variation of geometric characteristics along the block length. When setting the curved blocks, it is possible to form a section of the trajectory with several extremes in curvature, which requires the formation of a schedule feedrate variation with several sections of acceleration/deceleration within one block. For such a schedule of the feedrate variation can also be formed by applying a parametric model. In this case, the parameters of the feedrate start F BS (L i ) and the end F BE (L i ) for the path section are set for individual sections of the block along the toolpath length (L i ) inside the block, while for linear-circular blocks the parameters are set for block boundaries. In Fig. 2a an example of setting a feedrate variation schedule for a curvilinear block using a parametric model is shown.

F

F(L) FBS(L1)

FBS(L2)

F(u)

FBS(LE)

F F(u1)

FBE(L1)

F(u2)

FBE(L2) FBE(LE)

F0 L0

L1

L2

(a)

FBE(uE)

F0 LE L

u0

u1

u2

uE u

(b)

Fig. 2. Variations in the feedrate on the length of the curved block: (a) parametric model of the feedrate; (b) functional model of the feedrate.

The formation of the feedrate variation schedule via continuous function F(u) at full-length curved block (Fig. 2b) is implemented using a functional model of schedule feedrate variation. In contrast to the parametric model of the feedrate variation, which forms a symmetrical law of acceleration/deceleration, the use of a functional model takes into account the uneven variation of curvature of the trajectory. The functional model provides the

58

V. Kombarov et al.

ability to calculate the feed rate for any point of the trajectory, taking into account the local curvature of the trajectory and the constraints of the kinematic and dynamic characteristics of the machine tool. However, the use of a parametric model of the feedrate variation schedule requires a much smaller amount of data and the cost of computation in real-time.

4 Mathematical Model of the High-Speed Machining Feedrate Variation Law in Path Sections of Acceleration/Deceleration A large number of the models were proposed for the S-shape feedrate. For example, [17] used the trapezoidal acceleration profiles. Cubic polynomials for a feedrate profile with a discontinuous jerk were applied in the work [18]. In the work [19] a polynomial function to jerk continuous profile was applied. Some authors suggested application of trigonometric functions for the S-shape feedrate profile. The model proposed in the work [20] provides a continuous acceleration profile with discontinuous jerk. In the work [21] a continuous jerk profile model based on the use of sin and cos functions was proposed. Seven-interval model of S-shaped feedrate profile with smoothly-limited jerk using sin2 function is proposed in [16]. In this paper, the authors propose a simplified three-interval model S-shaped feedrate profile with smoothly-limited jerk using the sin2 function (Fig. 3). T1

DJmax 0 -DJmax Jmax JP

T2

T3

DJ(t)

(a)

J(t)

0

(b)

-JP -Jmax Amax AP

(c)

A(t)

0 F (d)

∆F

V(t) F0 t0

t1

t2

t3=tP

Fig. 3. Diagram of kinematic motion parameters: (a) is the jerk derivative (b) is the jerk; (c) is the acceleration; (d) is the velocity.

The following symbols for kinematic processing parameters are proposed: F 0 is a feedrate at the moment of frame beginning, m/s; F is a specified feedrate, m/s; AP is a largest achieved acceleration, m/s2 ; Amax is a maximum acceleration limit, m/s2 ; J P is a largest achieved jerk m/s3 ; J max is a maximum jerk limit, m/s3 ; DJ max is a maximum jerk derivative (snap) limit, m/s4 ; t P is a duration of velocity change from F 0 to F, s;

S-Shape Feedrate Scheduling Method

59

t P is an end time for velocity change, s; S(t) is a travel function, m; V (t) is a velocity along the trajectory (feedrate), m/s; A(t) is a acceleration function, m/s2 ; J(t) is a jerk function, m/s3 ; DJ(t) is a jerk derivative (snap) function, m/s4 . The following boundary conditions should be observed to achieve the smoothness of tool motion during acceleration: V (t 0 ) = F 0 ; V (t P ) = F; A(t 0 ) = A(t P ) = 0; J(t 0 ) = J(t P ) = 0; DJ(t 0 ) = DJ(t P ) = 0. The S-shape feedrate profile (Fig. 3), which usually includes three intervals, is proposed:   (1) Ti = ti−1 , ti , i ∈ [1, 3]. Acceleration starts at time t 0 and finishes at t 3 = t P , complete process duration is Δt P . The model covers the following intervals: – T 1 is an acceleration from A(t 0 ) = 0 to A(t 1 ) = AP , interval duration is t 1 ; – T 2 is a motion with constant acceleration A(t 1 …t 2 ) = const, interval duration Δt 2 ; – T 3 is an acceleration variation from A(t 2 ) = AP to A(t 3 ) = 0, interval duration is t 3 . In the particular case, intervals of constant acceleration T 2 may be not used. At each i interval time from it start is defined as τ(t) = (t − ti−1 ), t ∈ Ti .

(2)

tP = t3 − t0 .

(3)

Complete process time is

The jerk variation law is based on the sin2 function with period defined by angular frequency ω (s–1 ). Value of ω is computed from the condition sin2 (ω · t1 ) = 0:  ω = π t1 . The jerk variation function is defined as: ⎧ ⎨ Jp · FJ (t), t ∈ T1 , J (t) = 0, t ∈ T2 , ⎩ −Jp · FJ (t), t ∈ T3 ,

(4)

where FJ (t) = sin2 (ω · τ(t)). The jerk derivative function DJ(t) is found by differentiation of the system (4): ⎧ J · F (t), t ∈ T1 , dJ (t) ⎨ p DJ DJ (t) = (5) 0, t ∈ T2 , dt ⎩ −Jp · FDJ (t), t ∈ T3 , where FDJ (t) = ω · sin(2 · ω · τ(t)).

60

V. Kombarov et al.

Acceleration, velocity and traveled distance functions can be obtained by sequential integration of jerk, acceleration and velocity functions (Fig. 3)  ti  ti  ti J (t)dt, V (t) = A(t)dt, S(t) = V (t)dt, A(t) = ti−1

ti−1

⎧ t ∈ T1 , ⎨ JP · FA(t), A(t) = A(t1 ), t ∈ T2 , ⎩ A(t1 ) − JP · FA(t), t ∈ T3 ,

where: FA(t) =

τ (t) 2

−

ti−1

(6)

sin(2·ω·τ (t)) ; 4·ω

⎧ t ∈ T1 , ⎨ JP · FV (t) + F0 , V (t) = A(t1 ) · τ (t) + V (t1 ), t ∈ T2 , ⎩ A(t1 ) · τ (t) − JP · FV (t) + V (t2 ), t ∈ T3 ,

2 2 sin(ω·τ (t)) − . where: FV (t) = τ(t) 2 2·ω Equation to determine the distance traveled during acceleration: ⎧ ⎪ t ∈ T1 , ⎨ JP · FS(t) 2+ F0 · t, S(t) = A(t1 ) · τ(t) + V (t ) · τ(t) + S(t ), t ∈ T2 , 1 1 2 ⎪ ⎩ τ(t)2 A(t1 ) · 2 − JP · FS(t) + V (t2 ) · τ(t) + S(t2 ), t ∈ T3 ,

(7)

(8)

τ(t) sin(2·ω·τ(t)) . where: FS(t) = τ(t) 12 − 8·ω2 + 16·ω3 The acceleration, velocity and traveled distance, which will be achieved at the end of T 1 interval (at the t 1 ), are found from (4)–(8):

t2 1 t1 1 + F0 · t1 . − A(t1 ) = JP · , V (t1 ) = JP · 1 + F0 , S(t1 ) = JP · t13 · 2 4 12 8 · π2 3

The velocity and traveled distance, which will be achieved at the end of T 2 interval (at the t 2 ), are found from (4)–(8): V (t2 ) = F − JP ·

t12 F + F0 , S(t2 ) = · (t2 − t1 ) + S(t1 ). 4 2

The distance traveled in the acceleration section is determined by the equation S(tP ) =

F + F0 · (t2 + t1 ). 2

The proposed mathematical model simplified three-interval S-shaped feedrate profile provides smoothly-limited jerk and simplicity of calculations in the CNC system. To apply the proposed mathematical model in the system of scheduling and control of feedrate, it is necessary to be able to calculate the time intervals of the acceleration cycle, taking into account the constraints on the kinematic parameters [8, 16]. The time parameter t 2 is present in the equations of velocity and traveled distance, which will be

S-Shape Feedrate Scheduling Method

61

achieved at the end of interval T 2 depends on the parameter t 1 . This dependence can be determined from the equation A(t1 ) · (t2 − t1 ) = F − F0 − 2 · [V (t1 ) − F0 ],

(9)

From which, taking into account the ratios for acceleration, velocity, which will be achieved at the end of T 1 interval we obtain t2 =

2 · (F − F0 ) . JP · t1

(10)

The total acceleration time t P , taking into account that the time intervals T 1 and T 3 are the same, will be determined by the equation tp = t 2 + t 1 .

(11)

At t = t 1 and A(t) = Amax find the initial acceleration interval t 1A , where the acceleration reaches the maximum allowable value t1A = 2Amax /JP .

(12)

From Eq. (5) at t = t 1 /4 and DJ(t) = DJ max we find the time of the initial acceleration interval t 1J , where the derivative jerk reaches the maximum allowable value t1J =

π · JP . DJmax

(13)

It is obvious that the parameter t 1 should satisfy the condition 4 · π · JP 2 · Amax ≤ t1 ≤ , DJmax JP hence the second constraint on the value of J P .  Amax · DJmax . JP ≤ 2·π

(14)

(15)

It was previously noted that for certain combinations of initial data, the acceleration interval T 2 may be absent at the acceleration section. In this case, taking into account the velocity, which will be achieved at the end of the T 1 interval, we can write the equation F = F0 + 2 · JP ·

2 t1F , 4

(16)

from which we find the time of the initial acceleration interval t 1F for the case when the system reaches a given operating feed F without a constant acceleration interval T 2 :  2 · (F − F0 ) . (17) t1F = JP

62

V. Kombarov et al.

If t 1A < t 1F , then the acceleration section will contain a non-zero interval of constant acceleration (t 2 > t 1 ), and if t 1F < t 1A , the interval of constant acceleration will be absent (t 2 = t 1 ), but the acceleration at the acceleration section will not exceed the maximum admissible (A(t) < Amax ). It is easy to verify that at t 1 = t 1F , when there is no constant acceleration interval in the acceleration section, the value of t 2 calculated by Eq. (10) will be equal to the value of t 1 calculated by Eq. (17). With a small interval F – F 0 , the value of the parameter t 1 , calculated by Eq. (17), may decrease so that the constraint (14) on the maximum allowable rate of the jerk variation will be violated. To exclude this possibility, we introduce the third constraint on the value “jerk acceleration” which is defined by inequality  π · JP 2 · (F − F0 ) ≤ , (18) DJmax JP solving which according to J P , we obtain  2 3 2 · (F − F0 ) · DJmax JP ≤ . π2

(19)

Thus, the optimal value of the parameter J P , which ensures the implementation of constraint on velocity, acceleration, jerk, and jerk velocity, is determined by the equation ⎞ ⎛   2 2 · A · DJ 2 · (F − F ) · DJ 3 max max 0 max ⎠ , , (20) JP = min⎝Jmax , π π2 and the optimal time t 1 of the initial acceleration interval by the equation t1 = min(t1A , t1F ).

(21)

Thus, the proposed mathematical model of the feedrate variation law in path sections of acceleration/deceleration solved the problem of determining acceleration cycle intervals time subject to the constraints of kinematic parameters.

5 Pre-scheduling Feedrate Using Smooth S-Shaped Laws of Acceleration/Deceleration Consider the procedure for feedrate scheduling in the CNC system when applying the parametric model of setting the profile for feedrate variation. With this model, the diagram of the feedrate variation in the block or in the feed control section can be set by two parameters – feedrate at the beginning (F BSi ) and at the end (F BEi ) of the section or block (Fig. 4).

S-Shape Feedrate Scheduling Method

Sacc i

F(L) FBS i

63

Sdec i

FBE i LB i

FBE i-1

L

Li-1

Li

Fig. 4. Diagram of the federate variation at the feed control section or block length

The movement in the block always starts with the feedrate achieved in the previous block (F BEi-1 ). At the beginning of the block, feedrate is accelerated to feed the block (F BSi ). The movement during the main part of the block is carried out on the feedrate of the current block (F BSi ). At the end of the block is deceleration before the end of the current block (F BEi ). The feedrate scheduling algorithm consists of two stages, which correspond to the formation of the diagrams F I and F II (Fig. 5). In the first scheduling stage, the problem of determining the feed values of the beginning F BSi and the end F BEi of each block is solved from the condition of the possibility of movement in the current block, taking into account the following constraints. F(L) FT

m-4 FII BS m-4 FII BS m-3

m-3

m-2 FI

m-1 FI BE m-1

m

FII BS m-2 FII

FII BS m-1 FII BS m

FBE m = 0 L

Fig. 5. Stages of construction feedrate variation scheduling along the trajectory.

The least of the following is selected as the block feedrate beginning (F BSi ): feedrate, defined in the text for the G-code stream block; feedrate limited by the maximum allowable centrifugal acceleration when moving along a curvilinear trajectory; feedrate, admissible according to the condition of limiting axial feedrates; feedrate achieved by constant acceleration from the feed of the previous block F BEi −1 end on the block length. The least of the following is selected as the block federate end (F BEi ): feedrate represented as the block beginning feedrate (F BSi ); feedrate, allowed for the transition from the current to the next block when feedrate direction vector variation subject to the limitations of new instant acceleration and jerks to each control axis. The result of the first scheduling stage is the F I diagram (Fig. 5), which specifies the possible motion feeds in the block. In the second stage, the problem of providing the possibility of slowing down the feedrate at the end of the motion section or in any arbitrary block of the section, in accordance with the algorithm (Fig. 6).

64

V. Kombarov et al. Begin Calculate Sacc(FB i-1,FBS i) and Sdec(FBS i,FBE i) Sdec + Sacc > LB i

False

True Calculate transition path from FBE i-1 to FBE i: SchngF = S(FB i-1,FBS i) SchngF = LB i False SchngF < LB i False FBE i-1 > FBE i

True

True

Do not change feedrate. Save FBS i and FBE i Change feedrate for current block start: FBS i = FBE i-1. Save FBE i-1; FBE i without changes

Find normalized feedrate X for block start from equation: Sacc(FBE i-1,X) + Sdec(X,FBE i) = LB i

Change feedrate for current block start: FBS i = X. Save FBE i-1; FBE i without changes

Find start feedrate X for decceleration section from equation: SchngF(X,FBE i) = LB i

Change feedrate for previous block end and current block start: FBE i-1 = X; FBS E i = X; Save FBE i without changes

True

False First stage algorithm error End

Fig. 6. Algorithm of the second stage of the feedrate scheduling.

To enable deceleration is decreasing values and set the differential feedrate in the block. The result of the second scheduling step is an adjusted schedule of feedrate variation F C (L), which corresponds to the diagram F II (Fig. 5). Adjustments reduce the length of the acceleration/deceleration sections. This automatically ensures that the previously imposed constraints on the feedrate schedule variation diagram are provided and no additional checks are required. It is not known in advance the number of blocks in which it is necessary to adjust the feed to perform a feed deceleration from the specified in the program to the final feed of the m-th block. Sequential sorting of blocks from the end to the beginning of the program is carried out. In each block, the feeds X m , X m−1 are calculated, from which the length of the block can be slowed down to the feed specified at the end of the corresponding block, and taken as the feed of the beginning of the current frame (F BSi ) and the feed of the end of the previous frame (F BEi -1 ). The developed pre-scheduling algorithm is implemented in a two-level CNC system [15] has no constraints on the number of analyzed blocks and the execution time of the analysis.

6 Testing the Method of the Feedrate Control Comparison of the results of feedrate scheduling for the test trajectory obtained by applying the proposed method and control of the MCV1000 machine tool with the CNC Heidenhain iTNC530 system is carried out (Fig. 7a). The test trajectory is a set of linear blocks (Fig. 7b). The trajectory presents the main types of conjugation of adjacent blocks and various combinations of a sequence of small and large length blocks, which provides the ability to check the constraints inherent in the methods of the feedrate control. The analysis of the results of the feedrate scheduling was carried out according to the data

S-Shape Feedrate Scheduling Method

65

of the motion velocity interpolation, recorded using the technical means of a two-level CNC system [15] and the Scope subsystem iTNC530.

Fig. 7. 5-axis MCV1000 with Heidenhain iTNC530 CNC system (a) and test trajectory (b).

The diagram of the interpolation velocity variation along the toolpath is obtained with the following constraints: – – – –

the feedrate F = 2000 mm/min; constraint of the axial movements feedrate F max = 10000 mm/min; the maximum allowable acceleration when moving on trajectory a = 4 m/s2 ; the maximum allowable jerk when moving on trajectory J = 20 m/s3 .

In Fig. 8 shown the diagram of the federate variation, which are formed using the proposed method in comparison with the CNC Heidenhain iTNC530. With the general similarity of the diagrams of the feedrate variation along the toolpath are obvious differences in the feed values at the time of passage of the conjugation point of blocks with a large fracture angle (curve 6, Fig. 8) and feed values achieved at short block length (curve 7, Fig. 8), although the magnitude of the feed at the points of conjugation with a small fracture angle is almost the same for both Heidenhain and the proposed method (curve 5, Fig. 8). F, m/min

2

Y, mm

1 2

3

7

8 9

1.5 4 50

1 0.5 5

6

t, s

0 0

0.2 0.4 0.6 0.8

1

1.2 1.4 1.6 1.8

X, mm

0 0

50

100

Fig. 8. Comparison of the feedrate diagrams formed by the proposed method with feedrte in the Heidenhain iTNC530 CNC: 1 – feed at a = 4 m/s2 , J = 20 m/s3 , dJ/dt = 6600 m/s4 ; 2 – feed at a = 4 m/s2 , J = 20 m/s3 , dJ/dt = 1600 m/s4 ; 3 – feed at a = 4 m/s2 , J = 10 m/s3 , dJ/dt = 1700 m/s4 ; 4 – feed Heidenhain iTNC530 at a = 4 m/s2 , J = 20 m/s3 ; 5 – feed at the junction of blocks with a small fracture angle; 6 – feed at the junction of blocks with a large angle of fracture; 7 – feed achieved at the length of the short block: 8 – the exact trajectory specified in the NC program; 9 – trajectory approximated to the Heidenhain iTNC530 CNC.

66

V. Kombarov et al.

For the proposed control method, the diagram of the feedrate variation at constraints of feed, acceleration, and jerk of similar, used in Heidenhain iTNC530 are obtained. In addition to the velocity, acceleration, and jerk constraints, the value of the allowable jerk variation rate dJ/dt = 1600, 6600 m/s4 varied (curve 1, 2, Fig. 8). The coincidence of the feedrate at the points of coupling with a small fracture angle indicates the coincidence of the models taking into account the limitation of the kinematic characteristics of the trajectory used in Heidenhain iTNC530 and the proposed method (curve 5, Fig. 8). A significant difference in feeds in the sections of joining blocks with a large bend angle (curve 6, Fig. 8) is explained by the presence of the mode of approximation smoothing of acute angles in the iTNC530 system (curve 9, Fig. 8), while in the experimental two-level CNC system sharp corners were interpolated without smoothing (curve 8, Fig. 8). In Table 1 shown the test results. With similar constraints on the amount of acceleration and jerk, the proposed S-shape feedrate scheduling method with smoothly-limited jerk provides a reduction in processing time from 7.4 to 10.7%. Table 1. Comparison of test results. Parameter

Heidenhain Experimental two-level CNC system

a, m/s2

4

4

4

4

J, m/s3

20

20

20

10

dJ/dt, m/s4 –

6600

1600 1700

t, s

1.87

1.67

1.73 1.8

%

–

10.7

7.4

3.7

7 Conclusion A pre-scheduling feedrate method, which has no constraints on the number of analyzed of CNC blocks and the planning execution time is proposed and experimentally tested. Using of a simplified three-interval model S-shaped feedrate profile with smoothlylimited jerk using the sin2 function when scheduling the feedrate is proposed. An experimental investigation using the example of the control section of the trajectory confirmed a reduction in processing time by 7%…10% compared to the Heidenhain iTNC530.

References 1. Permyakov, A., Dobrotvorskiy, S., Dobrovolska, L., Basova, Y., Ivanova, M.: Computer modeling application for predicting of the passing of the high-speed milling machining hardened steel. In: Ivanov, V., et al. (eds.) DSMIE 2018. LNME, pp. 135–145. Springer, Cham (2019). https://doi.org/10.1007/978-3-319-93587-4_15

S-Shape Feedrate Scheduling Method

67

ˇ 2. Dodok, T., Cuboˇ nová, N., Císar, M., et al.: Influence of CNC milling strategies on complex surface machining. IOP Conf. Ser. Mater. Sci. Eng. 776(1), 012002 (2020). https://doi.org/ 10.1088/1757-899X/776/1/012002 3. Plankovskyy, S., Tsegelnyk, Y., Shypul, O., Pankratov, A., Romanova, T.: Cutting irregular objects from the rectangular metal sheet. In: Nechyporuk, M., Pavlikov, V., Kritskiy, D. (eds.) Integrated Computer Technologies in Mechanical Engineering. AISC, vol. 1113, pp. 150–157. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-37618-5_14 4. Anghel, C., Gupta, K., Jen, T.C.: Analysis and optimization of surface quality of stainless steel miniature gears manufactured by CO2 laser cutting. Optik 203, 164049 (2020). https:// doi.org/10.1016/j.ijleo.2019.164049 5. Bernatskyi, A.V., Berdnikova, O.M., Klochkov, I.M., et al.: Laser welding in different spatial positions of T-joints of austenitic steel. IOP Conf. Ser. Mater. Sci. Eng. 582(1), 012048 (2019). https://doi.org/10.1088/1757-899X/582/1/012048 6. Plankovskyy, S., Shypul, O., Tsegelnyk, Y., et al.: Simulation of surface heating for arbitrary shape’s moving bodies/sources by using R-functions. Acta Polytechnica 56(6), 472–477 (2016). https://doi.org/10.14311/AP.2016.56.0472 7. Krol, O., Sokolov, V.: Modeling carrier system dynamics for metal-cutting machines. In: 2018 International Russian Automation Conference (RusAutoCon), pp. 1–5. IEEE, Sochi (2018). https://doi.org/10.1109/RUSAUTOCON.2018.8501799 8. Chen, Y., Ji, X., Tao, Y., Wei, H.: Look-ahead algorithm with whole S-curve acceleration and deceleration. Adv. Mech. Eng. 5, 974152 (2013). https://doi.org/10.1155/2013/974152 9. Qin, Q., Huang, J., Yao, J.: A real-time adaptive look-ahead speed control algorithm for FDMbased additive manufacturing technology with Hbot kinematic system. Rapid Prototyping J. 25(6), 1095–1107 (2019). https://doi.org/10.1108/RPJ-11-2018-0291 10. Plankovskyy, S., Myntiuk, V., Tsegelnyk, Y., Zadorozhniy, S., Kombarov, V.: Analytical methods for determining the static and dynamic behavior of thin-walled structures during machining. In: Shkarlet, S., Morozov, A., Palagin, A. (eds.) MODS 2020. AISC, vol. 1265, pp. 82–91. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-58124-4_8 11. Petrakov, Y., Shuplietsov, D.: Contour milling programming technology for virtual basing on a CNC machine. Eastern-Eur. J. Enterprise Technol. 98(2/1), 54–60 (2019). https://doi.org/ 10.15587/1729-4061.2019.162673 12. Kolar, P., Sulitka, M., Matyska, V., Fojtu, P.: Optimization of five-axis finish milling using a virtual machine tool. MM Sci. J. 2019(05), 3534–3543 (2019). https://doi.org/10.17973/ MMSJ.2019_12_2019037 13. Vovk, A., Sölter, J., Karpuschewski, B.: Finite element simulations of the material loads and residual stresses in milling utilizing the CEL method. Procedia CIRP 87, 539–544 (2020). https://doi.org/10.1016/j.procir.2020.03.005 14. Jacso, A., Szalay, T., Jauregui, J.C., Resendiz, J.R.: A discrete simulation-based algorithm for the technological investigation of 2.5D milling operations. Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 233(1), 78–90 (2019). https://doi.org/10.1177/0954406218757267 15. Aksonov, Y., Kombarov, V., Fojt˚u, O., et al.: Investigation of processes in high-speed equipment using CNC capabilities. MM Sci. J. 2019(04), 3271–3276 (2019). https://doi.org/10. 17973/MMSJ.2019_11_2019081 16. Kombarov, V., Sorokin, V., Fojt˚u, O., et al.: S-curve algorithm of acceleration/deceleration with smoothly-limited jerk in high-speed equipment control tasks. MM Sci. J. 2019(04), 3264–3270 (2019). https://doi.org/10.17973/MMSJ.2019_11_2019080 17. Erkorkmaz, K., Altintas, Y.: High speed CNC system design. Part I: jerk limited trajectory generation and quintic spline interpolation. Int. J. Mach. Tools Manuf. 41(9), 1323–1345 (2001). https://doi.org/10.1016/S0890-6955(01)00002-5

68

V. Kombarov et al.

18. Liu, Q., Liu, H., Yuan, S.: High accurate interpolation of NURBS tool path for CNC machine tools. Chin. J. Mech. Eng. 29(5), 911–920 (2016). https://doi.org/10.3901/CJME.2016.040 7.047 19. Fan, W., Gao, X.S., Yan, W., Yuan, C.M.: Interpolation of parametric CNC machining path under confined jounce. Int. J. Adv. Manuf. Technol. 62(5), 719–739 (2012). https://doi.org/ 10.1007/s00170-011-3842-0 20. Lee, A.C., Lin, M.T., Pan, Y.R., Lin, W.Y.: The feedrate scheduling of NURBS interpolator for CNC machine tools. Comput. Aided Des. 43(6), 612–628 (2011). https://doi.org/10.1016/ j.cad.2011.02.014 21. Ni, H., Yuan, J., Ji, S., et al.: Feedrate scheduling of NURBS interpolation based on a novel jerk-continuous ACC/DEC algorithm. IEEE Access 6, 66403–66417 (2018). https://doi.org/ 10.1109/ACCESS.2018.2813334

Selecting the Method for Pre-tightening Threaded Connections of Heavy Engineering Objects Anatoliy Gaydamaka , Yuriy Muzikin , Volodymyr Klitnoi(B) Yevheniia Basova , and Sergey Dobrotvorskiy

,

National Technical University “Kharkiv Polytechnic Institute”, 2, Kyrpychova str, Kharkiv 61002, Ukraine [email protected]

Abstract. The choice and application pre-tightening method of the threaded connection for each specific technical object is determined both by the specifics of production and by established traditions that have passed many years of experience in operational tests. At the same time, there is no formalized scientific justification for the decision on the choice of the tightening method for the threaded connection. In this work, based on the results of a review of the known methods of pretightening of threaded connections, the choice of tightening the impeller with the shaft of axial hydraulic turbines is justified. An assessment of the probability of failure-free operation of threaded connections is proposed, taking into account the applied tightening method and its weakening during operation. As a result of the research, it was found that for the known methods of implementing and controlling the tightening of threaded connections, namely with a torque wrench, according to the rotation angle, according to the deformation of the elastic washer, according to the elongation of the bolt, 20% weakening corresponds to the probability of failure-free operation equal to 0.5. The method of tightening with a torque wrench has the softest effect on the probability of failurefree operation according to the criterion of relaxation of tightening forces. The elongation measurement tightening method up to 15% loosening guarantees high reliable operation of the threaded connection with a failure rate of 0.99. The results obtained will make it possible to scientifically select a method for the implementation and control of tightening of threaded connections of heavy engineering objects, depending on their purpose and design features. Keywords: Threaded connection · Pre-tightening · Uptime probability · Tightening weakening · Methods of implementation and control of tightening

1 Introduction Heavy engineering is referred to as material-intensive industries such as metallurgical, mining, lifting and transport, energy, heavy machine-tool building, sea and river shipbuilding, railway rolling stock, where a significant amount of metal and energy is consumed with a relatively low labor intensity. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 69–77, 2022. https://doi.org/10.1007/978-3-030-83368-8_7

70

A. Gaydamaka et al.

The specifics of the work of heavy engineering objects threaded connections (TC) with large fasteners (d = 30… 200 mm) are often determined by high variable loads, significant rotation speeds of rotor systems, and adverse environmental influences [1]. In full measure, the noted features of the operation of heavy engineering facilities are also true for the TC of axial hydraulic turbines impellers. To reduce the external negative impact on the TC, special vibration protection supports can be used [2], the quality characteristics of bearings can be improved [3, 4]. However, first of all, to ensure the reliable operation of such machines, a high-quality tightening of the TC is required, which is characterized by the size and uniformity of the preliminary tightening of the group connection, the accuracy of the normalized tightening of each single connection, and the stability of the connection tightening during operation [5]. Known cases of heavy engineering machines and equipment TC destruction with catastrophic consequences [6–8] require an increase in their reliability. This, among other things, forms the problems of small businesses, which hinder the pace of their development in the context of the current trend in the development of digital technologies [9]. Therefore, the first step in solving this problem is the scientific substantiation of the choice of the method of implementation and control of the TC preliminary tightening.

2 Literature Review For TC with large fasteners, there are two technologically possible methods of obtaining a preliminary tightening: heating and cold. These methods differ both in the physical principles of their implementation and in the peculiarities of the technological equipment, and the choice of the method is significantly influenced by the traditions of production and the experience of long-term operation of the equipment. The heating method of tightening TC is provided due to the temperature change in the length of the bolt (stud) without mechanical impact on the connection elements [5]. The heating method depends on both the bolt size and the design features of the connection, and the heating source used is determined by both the assembly conditions and the technical capabilities of production. Therefore, both the open flame of gas burners and various electric heaters are used as heating sources. The relative simplicity of the thermal method of tightening in the absence of a force effect on the bolts (studs) at the time of its installation is undoubtedly its advantage. However, the method has significant drawbacks, the main of which are: the need to use special fasteners adapted to the conditions of the thermal tightening method; availability of special heaters with adjustable performance; creation of conditions regulating the rate of heating and cooling of the TC; availability of certified methods of control of temperature and linear parameters of fasteners; ensuring increased requirements for safety, fire safety and sanitary conditions of work. Heating tightening of large threaded fasteners is used for hydraulic and steam turbines, generators, crimping and rolling mills, cone and jaw crushers, screens [10]. Difficulty of meeting these requirements, and often the impossibility of implementing some of them, significantly limit the cases of using the thermal tightening method, and therefore, in each specific case, a technical and economic analysis of the possibilities of its use is necessary.

Selecting the Method for Pre-tightening Threaded Connections

71

The cold method of tightening a large TC can be considered as an alternative and, depending on the implementation methodology, is based on the measurement of the following parameters: • torque applied to the nut when tightening; • the angle of rotation of the nut relative to the initial position; • elongation of the bolt (studs), generated by bolt (studs) tightening. The most widespread in practice is the torque control method using calibrated wrenches. The moment T applied to the nut is related to the axial force Fax and is balanced by the moments from the friction forces in the thread and at the end of the nut under the assumption of uniform loading of the mating surfaces   Dm (1) T = Fax · d2 · · f + tan(β ± α) , d2 where d2 - average thread diameter; Dm - average diameter of the bearing surface of the nut; f - coefficient of friction between the nut end and the supporting surface; β angle of friction in the thread; α- angle of rise of the thread; the «+» sign corresponds to tightening, and the «–» sign to unscrewing. Actually, the stresses arising in various elements of the threaded mating depend on many design and operational factors, the consideration of which is a difficult technical problem. The main of these parameters are: the quality of processing of mating surfaces, types of coating and the presence of lubricants on them, heat treatment, specific pressure, repeatability of assembly, screwing speed and a number of other factors. In addition, the relationship between the torque and the tightening force does not remain constant and therefore the error of this method is ±30% [11]. In the rotation angle method, the value of the angle is determined by the expression ϕ = 360◦ ·

n Fax  · λi , S

(2)

i=1

where

n 

λi - total flexibility of the bolt and parts to be tightened; n- the number of parts

i=1

in the connection; S- thread pitch. The rotation angle method is simple, does not depend on the individual characteristics of the threaded connection and does not require special equipment during the implementation process. However, the complexity of calculating the flexibility of the parts being pulled together, as well as determining the initial value of the angle ϕ0 at which the gaps in the joint are completely selected, make this method ineffective, and the error is on the order of ±20% [10]. For this reason, the use of such a tightening for short bolts is practically excluded, since the required angle of rotation is commensurate with the error of the method. In the most critical cases, when a violation of the preliminary tightening of large fasteners can lead to serious accidents, the elongation measurement method is used with a tensor jack (tensioner stud) [12]. Tensioning jacks are used to apply a preliminary calibrated tension force on the bolt, followed by loose tightening of the nut (up

72

A. Gaydamaka et al.

to the stop). Tensor jacks are used to assemble heavy-duty TC with high-strength fasteners for compressors, generators, presses, diesel and marine engines, transformers, gearboxes, couplings, aircraft gas turbines, rolling mills, presses, finishing machines, heavy machines, extruders, pumps, tower cranes, excavators and mining equipment, pipelines, pressure vessels, heat exchangers, wind generators, turbines of thermal and nuclear power plants with a range of bolt diameters (studs) from 5 to 500 mm [13]. The principle of operation of a tensor jack (Fig. 1) is as follows [14]. First, the twist chuck is installed over the nut to be tightened. At the same time, the hydraulic device grips the bolt body. The tightening (working) nut is screwed onto the protruding part of the bolt. The applied hydraulic pressure stretches the bolt axially. When the required pressure (load) level is reached, the nut is tightened without effort using a nutrunner. Upon completion of the operation, the pressure in the system is released and the chuck returns to its original position. Thus, tightening is done through axial tension of the bolt without friction or twisting. The TC is dismantled in the same way.

Fig. 1. General view (a) and scheme (b) of the tensor jack (tensioner stud).

The hydraulic principle of the TC tightening has the following advantages: high accuracy of the specified load and repeatability of results ±2.0%; maximum approximation of the load value to the elastic limit of the material; the possibility of simultaneous tightening of several bolts with the same force; simplicity and safety in operation; the possibility of implementing automation systems for tightening bolts; the ability to tighten nuts and bolts from any type of material. The elongation measurement method is especially effective for TC with long stepped bolts and for TC of heavy loaded with high strength bolts. The disadvantages of the elongation measurement method are: loss of load, the need for expensive equipment and special tools, and less efficiency.

Selecting the Method for Pre-tightening Threaded Connections

73

The load loss is taken into account in the calculation in advance and added to the preload value to determine the initial load value. Therefore, for greater accuracy, it is recommended to apply force to the bolt and tighten the nut twice. At the first loading, the joint is upset with a force exceeding the calculated one, which causes plastic deformation of the metal in the point contact areas due to the creation of stresses 0.7… 0.8 of the yield point with a holding time of up to 3 h. This provides compensation for gaps and deformation of microroughnesses of surfaces, which ensures an even distribution of the load. The second loading is necessary to achieve the required accuracy of the final tightening force TC. Uniformity of loading when tightening a group of bolts (studs) is ensured by simultaneous connection to one hydraulic station of several tensor jacks up to 100% fasteners [15] to synchronize their work (Fig. 2).

Fig. 2. Tensioner jacks on the assembly of bolted connections of a wind turbine.

With a smaller coverage of simultaneously tightened bolts or studs (50, 25, 10, 5%), special studies and calculations are required to carry out preliminary tightening of a group TC. Until now, there is no formalized procedure for the quantitative comparative evaluation of the decision on the choice of the method for implementing the TC preliminary tightening. Therefore, a study aimed at developing a mathematical model for evaluating a decision on the choice of a method for implementing a preliminary tightening of an TC based on a probabilistic approach should be considered relevant.

74

A. Gaydamaka et al.

3 Research Methodology The aim of the work is to develop a mathematical model for a comparative assessment of the methods of preliminary TC tightening of machines and equipment for heavy engineering. To achieve this goal, it is necessary to solve the following tasks: • develop a probabilistic model for comparative evaluation of the ways of TC preliminary tightening; • calculate the probability of failure-free operation of the TC with preliminary tightening in various ways; • formulate recommendations for the use of the most effective method of preliminary TC tightening of the impeller with the shaft of axial turbine of hydroelectric power plants. 3.1 Probabilistic Model for Evaluation the Methods of TC Pre-tightening Known methods for calculating the TC based on the probabilistic approach are presented in [16–18]. Mathematical models for calculating the probability of no-failure operation (PNFO) of the TC based on the criteria of non-opening and non-displacement of the joint are most acceptable for a comparative assessment of the methods of performing preliminary tightening. However, the known PNFO models contain eight independent variables. Therefore, it is proposed that the probability of failure-free operation Ptight of the TC by the criterion of relaxation of tightening forces for various implementation methods is represented by a mathematical model   Ptight = probability σtight  σT , (3) and its value is determined by the calculated quantile of the normalized normal distribution for three random variables from the expression uPtight = − 

ntight − 1 n2tight

· νσ2T + νσ2tghit

,

(4)

T ; νσT , νσtghit - respectively, where ntight - safety factor for tightening value, ntight = σσtghit the coefficients of variation of random mean values for the yield point and tightening stress σT , σtight .

4 Results The results of calculating the probability of failure-free operation Ptight according to the criterion of relaxation of tightening forces for various methods of implementation on the example of TC with high-strength bolts (σtghit = 0.8 · σT ) are shown in the form of graphs in Fig. 3.

Selecting the Method for Pre-tightening Threaded Connections

75

For known methods of tightening TC with a torque wrench control, according to the nut rotation angle, according to the deformation of the elastic washer, according to the elongation of the bolt, its 20% weakening corresponds to PNFO = 0.5. The softest effect on the PNFO according to the criterion of relaxation of tightening forces from the action of variable loads is exerted by the method of tightening with a torque wrench. The method of tightening the bolt axially up to 15% loosening guarantees its reliable assessment (PNFO = 0.999).

Fig. 3. Graphs of the functions of the probability of failure-free operation Ptight of a threaded connection according to the criterion of relaxation of tightening forces for various methods of its implementation (curve 1 with a torque wrench control; 2 - according to the nut rotation angle; 3 according to the deformation of the elastic washer; 4 - according to the bolt elongation).

The reliability of the TC tightening by the elongation measurement method under conditions of 15% loosening exceeds the reliability when tightening by the methods with a torque wrench, rotation angle, and the deformation of the elastic washer by 31, 11, 7%, respectively. Developed probabilistic assessment model makes it possible to comprehensively, without calculating the PNFO of the TC according to individual criteria (non-opening of the joint and non-displacement of the joint), to solve the question of the advantage of one or another method of performing preliminary tightening. In addition, the proposed probabilistic model saves time and money for determining the random variables required for calculations. Taking into account the above, the use of the proposed model will also be promising for assessing the effectiveness of the newly proposed technical solutions for tightening the projected TC. Proposed probabilistic model for evaluation the methods for performing preliminary TC tightening, as well as the known models [17, 18], has limitations related to experiments. However, this requires almost three times less experimental data. In addition, the proposed model does not take into account the influence of the variable nature of

76

A. Gaydamaka et al.

external loads on the tightening value, material defects of fasteners, assembly errors, and possible degradation of the connection under the influence of the external environment. The disadvantages of the proposed probabilistic model open up prospects for further research in the direction of a formalized assessment of the methods for carrying out preliminary tightening of the TC and, first of all, in taking into account the influence of the variable nature of external loads.

5 Conclusions Cold pre-tightening methods are the most common and can be implemented in any assembly area with a wide range of fasteners. A probabilistic model has been developed for evaluation the methods of preliminary tightening of the TC, with the help of which it is possible to solve the issue of choosing the most acceptable option in a complex way without calculating according to individual criteria and effectively with a smaller amount of experimental data. The reliability of the TC when tightening by the elongation measurement method under conditions of 15% loosening exceeds the reliability when tightening by the methods with a torque wrench, rotation angle, and the deformation of the elastic washer by 31, 11, 7%, respectively. The most effective way to ensure the preliminary tightening of the impeller with the shaft of axial hydraulic turbines TC is the method of axial drawing of the studs using tensor jacks, which provide high reliability of the tightening force during automated synchronous group assembly.

References 1. Iosilevich, G.B., Stroganov, G.B., Sharlovskij, J.: Tightening and locking threaded connections: Handbook. Mashinostroenie, Moscow (1985) 2. Klitnoi, V., Gaydamaka, A.: On the problem of vibration protection of rotor systems with elastic adaptive elements of quasi-zero stiffness. Diagnostyka. 21(2), 69–75 (2020) 3. Gaydamaka, A., Klitnoy, V., Muzikin, Y., Tat’kov, V., Hrechka, I.: Construction of a model for the distribution of radial load among the bearing’s rolling bodies. Eastern-Euro. J. Enterp. Technol. 6, 7(96), 39–44 (2018) 4. Gaydamaka, A., et al.: Devising an engineering procedure for calculating the ductility of a roller bearing under a no-central radial load. Eastern-Euro. J. Enterp. Technol. 3, 7(99). 6–10 (2019) 5. Bickford, J.H.: Introduction to the Design and Behavior of Bolted Joints. CRC Press, New York (2007) 6. Berlin, V.V., Murav’ev, O.A.: Technical aspects of the failure of the second generating set at the Sayano-Shushenskaya HPP. Power Technol. Eng. 44, 263–268 (2010) 7. Jones, D.R.H.: Failure analysis case studies II: A sourcebook of case studies selected from the pages of engineering failure analysis 1997–1999. Elsevier, Pergamon (2001) 8. Krol, O., Sokolov, V.: Research of modified gear drive for multioperational machine with increased load capacity. Diagnostyka 21(3), 87–93 (2020) 9. Dobrotvorskiy, S., Basova, Y., Dobrovolska, L., Sokol, Y., Kazantsev, N.: Big challenges of small manufacturing enterprises in industry 4.0. In: Ivanov, V., Trojanowska, J., Pavlenko, I., Zajac, J., Perakovi´c, D. (eds.) DSMIE 2020. LNME, pp. 118–127. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-50794-7_12

Selecting the Method for Pre-tightening Threaded Connections

77

10. Gaponov, V.S., Muzikin, Y.D., Tat’kov, V.V., Vyshnyvetskyi, S.M., Voitovych, O.I., Moskalenko, P.P.: Increasing the bearing capacity and reliability of threaded connections in rolling equipment. Bull. NTU KhPI, 47(953). 106–110 (2012) 11. Birger, I.A., Iosilevich, G.B.: Threaded and flange joints. Mashinostroenie, Moscow (1990) 12. Miao, R., Shen, R., Zhang, S., Xue, S.: A review of bolt tightening force measurement and loosening detection. Sensors 20, 3165 (2020) 13. Monville, J.M.: Optimal tightening process of bolted joints. Int. J. Simul. Multisci. Des. Optim. 7(A4), 1–22 (2016) 14. SKF.: Bolt Tightening Handbook. SKF Montigny le Bretonneux. France (2001) 15. Müller, R., Vette, M., Hörauf, L.: An adaptive and automated bolt tensioning system for the pitch bearing assembly of wind turbines. Robot. Comput. Integr. Manufact. 36, 119–126 (2015) 16. Yin, T., Wang, X.Y., Zhu, H.P.: A probabilistic approach for the detection of bolt loosening in periodically supported structures endowed with bolted flange joints. Mech. Syst. Sign. Process. 1–54 (2019) 17. Antsupov, A.V., Antsupov, V.P.: Methodology of machine elements reliability prediction by means of various criteria. Dependability. 3(46), 15–23 (2013) 18. Novoselac, S., Kozak, D., Ergi´c, T., Damjanovi´c, D.: Fatigue damage assessment of bolted joint under different preload forces and variable amplitude eccentric forces for high reliability. In: Pluvinage, G., Milovic, L. (eds) Fracture at all Scales. Lecture Notes in Mechanical Engineering. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-32634-4_13

Production Scheduling of Semiconductor Wafer Fabrication Facilities Using Real-Time Combinatorial Dispatching Rule Suraj Panigrahi1 , Srijeta Agrahari1 , José Machado2 , and V. K. Manupati1(B) 1 Department of Mechanical Engineering, NIT Warangal, Warangal, Telangana, India

[email protected]

2 Department of Mechanical Engineering, University of Minho, 4800-058 Guimarães, Portugal

[email protected]

Abstract. Fabrication of semiconductor wafers is a complicated and challenging process with huge system complexity and stochasticity in operations. To improve the above challenges and to enhance the system’s efficiency an effective scheduling technique is indispensable. In recent times Real-time scheduling with dynamic dispatching rules has been widely discussed. However, choosing the appropriate combination of dispatching rules in a dynamic environment is a challenge. In this study, we proposed a multi-objective non-dominated sorting genetic algorithm (MO-NSGA-II) approach for optimizing the dispatching rules by considering the multiple objectives as minimization of work in process, minimization of delay time, and minimization of makespan. To implement the proposed approach, the optimal parameters for the multi-objective evolutionary algorithm (MOEA) are selected based on the hierarchical combination method and by deploying the response surface methodology (RSM) the best combination of rules is generated. Further, a real-time simulated environment is created using Flexsim to check the significance of the proposed approach and the robustness of the generated combinatorial rules. Results stated that the proposed approach can improve the performance of a system to a greater extent. Keywords: Wafer fabrication · Response surface methodology · Combinatorial dispatching rules · Multi-objective evolutionary algorithm · Discrete event simulation

1 Introduction Production scheduling in a semiconductor manufacturing system is extremely essential as it is one of the most intricate manufacturing environments with multiple machines following several processing steps, unstable production conditions, re-entrant flows, and batch processing capabilities. Generally, it involves four phases, namely: wafer fabrication, wafer probe, assembly, and final inspection, among which this paper address the wafer fabrication process as it is the most time-consuming and capital intensive [1]. Due to the dynamic nature and modeling complexity of the scheduling problem, the © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 78–90, 2022. https://doi.org/10.1007/978-3-030-83368-8_8

Production Scheduling of Semiconductor Wafer Fabrication Facilities

79

semiconductor production system is interpreted as NP-hard in nature [2], for which obtaining a real-time optimal solution within a reasonable time is not possible [3]. As a result, heuristic methods are widely utilized in place of analytical models to find feasible solutions in a real-time dynamic scheduling environment, where the arrival of urgent jobs, breakdown of machines, and change in due dates can occur [4]. Due to the wide-scale and complexity of semiconductor production systems, dispatching rules with an uncertain nature are mostly applied for solving the scheduling issues [5, 6]. Appropriate selection of dispatching rules can minimize inventory, waiting time, and tardiness cost of the system. Although there are several effective dispatching rules available, it is observed that using a single dispatching rule for all the operations in a system is not possible [7] particularly in a dynamic scheduling environment. In this context, [8] developed rules by categorizing machines into batches, dedicated steppers, and non-dedicated steppers. In recent years, many researchers tried solving dispatching decision problems i.e., material handling and processing tools separately, an integrated approach can improve the system’s efficiency. In unstable production scheduling conditions of a wafer manufacturing system, incorporation of heuristic approaches and evolutionary algorithms is important, when it comes to machine-dependent rule selection. [9] divided the entire scheduling process into two sets of rules namely release and dispatch. Chen et al. [10] proposed a genetic algorithm for choosing the batch rules, dispatching rules, and set up rules to improve the system performance. To make the scheduling process more optimal, Song et al. [11] developed an evolutionary strategy by minimizing the overall cost. Batch processing steps in a wafer fabrication system may tend to 10 h. while compared to other processing steps which take hardly 2 h. of processing. For solving such batch machine scheduling problems, meta-heuristics was utilized by Mathirajan and Sivakumar [12]. However, several research-work was done in the past to find the best rules but the studies were only focused on the single-stage problem. This study deals with mapping the dispatching rules for the real-time semiconductor wafer in Fabs using a hierarchal combination-based method. Through a survey of the literature, very few studies have analyzed the application of evolutionary algorithms for the selection of rule combinations. The contribution of this paper is to propose an effective evolutionary algorithm approach to handle the aforementioned limitations and with that, it will address the following questions mentioned below: i.

How the combination of effective dispatching rules can be obtained to optimize the entire fabrication process? ii. How the design of the technique of the experiment can be integrated to get the optimum parameters for the MOEA? iii. How to evaluate the obtained combination of rules by utilizing a simulation analysis? For this, we propose a multi-objective NSGA-II based evolutionary algorithm approach which is combined with RSM to automatically generate combinatorial dispatching rules. Thereafter, a Flexsim based discrete simulation environment is created to test the feasibility of the developed rules. By utilizing this approach, the user will be benefited to select the appropriate solutions according to their choice.

80

S. Panigrahi et al.

This paper is organized as follows. In Sect. 2, the framework of the proposed methodology is presented and traced. Section 3 details the evolutionary algorithm used to generate various combinatorial dispatching rules. In Sect. 4, we map the problem with discrete-event simulated Flexsim simulation software. Section 5 discusses the performance results under varied conditions. Finally, Sect. 6 concludes and summarizes the future applications and research work.

2 Proposed Framework and Rules Construction This section exhibits an approach to integrate AMHS with the processing tools, then applied to real-time dynamic scheduling-based problems for wafer fabrication in a semiconductor manufacturing system. The methodology is divided into two parts namely cloud scheduler and self-contained scheduler. The cloud scheduler operates as an information transacting medium among the users, semiconductor manufacturing system, and the self-contained scheduler, along with that it monitors the entire system. The cloud scheduler coordinates with the self-contained scheduler to initialize it and for authorizing the combinatorial rules generated by the self-contained scheduler. Finally, it applies those rules to the real-time system. The cloud scheduler can also be utilized in times when new combinatorial rules are reselected (see Fig. 1). A dynamic wafer fabrication system majorly deals with two types of stochastic problems i.e., major uncertainties and minor uncertainties [13]. Unknown machine failures for which the repair time is not fixed, urgent job arrivals come under major uncertainties, and breakdowns for which the required repair times are known comes under minor uncertainties. Based on this, Kim and Jeong [14] proposed three scenarios for reselecting rules, namely, at the beginning of the process, when a major uncertain problem occurs and when the performance of the system is not up to the estimated performance measures. These rules are utilized in this study to determine the time for reselecting the dispatching rules. The self-contained scheduler is integrated with 4 modules: MOEA, a DOE module for optimizing the parameters, a simulation module consists of Flexsim, and the dispatching rules modules. At first, the MOEA and the DOE module assist in selecting the optimal parameters that are crossover rate, stopping criterion, mutation rate, and the population size for the Algorithm and number of iterations and simulation time for the simulation related parameters. These parameters are automatically generated based on the trigger signals by a user or cloud scheduler. Normally, such trigger signals are passed when there is a processing tool failure or change in material plan or any user request. MOEA and the Flexsim module are the most important part of the self-contained scheduler. Iterations are performed by applying MOEA, to generate rules with a specific code number, and then these generated rules are sent to the simulation module where the various combination of the rules are generated and validated to obtain the optimal set of combinatorial rules. For validating the obtained results, in this paper, we considered three key performance indicators i.e., work in progress (WIP) level, cycle time (CT), and delay time.

Production Scheduling of Semiconductor Wafer Fabrication Facilities

81

Fig. 1. Framework of the proposed methodology

2.1 Selecting Dispatching Rules for Lot, Batch and AGV In a complex and dynamic manufacturing environment, assigning jobs to work station is a very tedious process as it involves multiple jobs and machines, scheduling such systems requires various sequencing techniques to specify what particular jobs to be processed and at which work station. These sequencing techniques are stated as dispatching rules based on which the entire scheduling process of wafer fabrication takes place. Researchers in the past, utilized single dispatching rules for scheduling, while increasing the performance and reducing the overall completion time of wafer fabrication, in here, we tried to combine several dispatching rules. In general, the dispatching rules used for Lot dispatching are basically of single attribute type or multi-attribute type. First Come First Serve (FIFO), EDD, Shortest Remaining Processing time (SRPT), SPT, Time in System (TIS) FALL under single attribute category. Whereas, Slack (SL), processing time/ time in the system (PT/TIS), Critical Ratio (CR) are included in the multi-attribute type. In this paper, to achieve better performance, rules consisting of FIFO, SPT, PRI-SPT, and PRI-FIFO are considered for lot dispatching. In PRI-FIFO and PRI-SPT, “PRI” represents a prioritized rule for hot lots. For batch dispatching, four rules are considered namely, Minimum Batch

82

S. Panigrahi et al.

Size rule (MBS), Modified Dynamic Batching Heuristic (MDBH), Modified Minimum Batch Rule (MMBS) and Process Urgency Classification heuristic (PUCH) and for AGV dispatching, First Come First Serve (FCFS), Critical ratio (CR) and First Encounter First Serve (FEFS) are utilized. 2.2 Response Surface Methodology Based Parameters Optimization The generated parameters such as crossover rate, population size, mutation rate, and the number of generations have an impact on the results obtained from the MOEA. As a result, achieving a proper set of these parameters is necessary when there is a time constraint. For this, we utilized the Design of Experiments (DOE) approach performed by the RSM. RSM utilizes both first-order as well as second-order models widely to obtain an optimal response. Even when only a little information is available for the response variables and the explanatory variables, this method can be adapted to obtain results. As the first-order model indicates linear functionality for the explanatory variables and might be a lack of fit for the MOEA, in this study, we used a second-order model. p/g (population/generation), C (crossover rate), and M (mutation rate) are utilized for constructing the response surface, as these factors affect the performance most. The steps performed for the RSM are shown below: STEP 1: Systematic design of a series of experiments and generation of the response data. For each of the three factors conveniently selecting three levels while sustaining the constraint P1 × G1 = P2 × G2 = P3 × G3 to make sure the computational time is fixed. STEP 2: Selection of a suitable RSM model capable to represent the data. Conduct 23 + 2(3) + 3 = 17 (No of runs = 2n + 2n + n, where n is the number of input parameters) experimental treatments from the combination of Three-factors three-levels central composite design in his study. m ¯ = δ0 + δ1 n1 + δ2 n2 + δ3 n3 + δ4 n21 + δ5 n22 + δ6 n23 + δ7 n1 n2 + δ8 n2 n3 + δ9 n1 n3 + δ10 n1 n2 n3 + ω

(1)

STEP 3: Incorporating the measured data into the mathematical model. Based on the regression model articulated in Eq. (1), the regression analysis is implemented as per the results from step 2. Factors having P-value ≤ 0.05 are adequate with a 95% confidence level. It is observed that C, C 2 , and M 2 are the weighty terms. Based on this the objective function is written as: m ¯ = δ0 +

k 

δi ni +

i=1

k  i=1

δii n2i +

 i

j

δij ni nj + ω

(2)

0.016998635054776p + 0.561973239636741 C g p p + 14.496563340281465 M + 0.007856021039124 ∗ g g − 1.104295374647964 C ∗ C − 446.872239436614 M ∗ M

RAVG = 2.079276917057858 −

Production Scheduling of Semiconductor Wafer Fabrication Facilities

83

p p ∗ C − 3.140120030321786e − 14 ∗ M g g − 2.178805221351669e − 14 C ∗ M

− 2.915103954243903e − 15

p − 0.575223923661956 C g p p − 1.633671261501979 M + 0.034397196087636 ∗ g g p + 0.029576464788722 C ∗ C + 47.323942661960364 M ∗ M + 0.1000 *C g p + 0.333333333333335 ∗ M − 0.999999999999954 C ∗ M g

σ = 0.797638294309693 − 0.227856494512691

where “m” is the estimated response, “ δ0 ” is constant,“ δ1 ”, “ δ12 ”, “ δij ” represent the coefficients of linear, quadratic, and cross-product terms, “ω” is error respectively. n reveals the coded variables. STEP 4: Authentication of the model by optimizing the objective with new optimal MOEA parameters thereby checking different accuracy or error parameters. The response surface methodology is conducted by Minitab software to predict the responses for the given input parameters. Figure 2 and 3 show the responses of RAVG and σ for the experimented data through residual and surface plots and the developed parameters are depicted in Table 1.

Fig. 2. Residual plot of RAVG .

84

S. Panigrahi et al.

Fig. 3. Residual plot of sigma.

Table 1. Parameters from RSM. Levels Factors p/g

C

M

1

0.4

0.7 0.015

2

0.6

0.8 0.018

3

1.0

0.9 0.02

3 Hierarchical Combination Based Evolutionary Algorithm for Generating Affective Hybrid Dispatching Rules For a manufacturing scheduling problem, Evolutionary Algorithms (EAs) have always been a favorable approach. Generally, it is a meta-heuristic algorithm based on a genetic population that utilizes biological evolution mechanisms i.e., mutation, selection, reproduction, and recombination. The process begins with a set (population) and operators of solutions (individuals), followed by iterations, and in each iteration, out of the population, good solutions are selected (survival of the fittest), and then new solutions are generated by combining two old solutions (crossover) and later modifying it (mutation). All these steps are followed until an optimal solution is achieved or a maximum number of iterations are performed. Also, to run a particular operation at a station with say ’n’ rules, the number of permutations would be ‘n!’. If even ten rules were to be considered, the number of permutations would result up to 36,28,800.

Production Scheduling of Semiconductor Wafer Fabrication Facilities

85

As illustrated in the figures below, one-point or two-point crossover can be utilized to exchange combined rules (see Fig. 4). Two-point crossover interchanges the two hierarchal combinations of rules of machines 2 and 3 between two individuals as shown in Fig. 5. Operators can be used for permutation-based encoding such as linear order crossover to exchange the order of rules within a combined rule. This has been demonstrated in Fig. 6, wherein the arrangement of rules is exchanged in the combined rules of machines 2 and 4.

Fig. 4. Implementing dispatching rules on 2 machines.

Fig. 5. Crossover mutation.

86

S. Panigrahi et al.

Fig. 6. Weight function for machines

4 Flexsim Simulation Analysis This section describes a simulation study conducted using the Flexsim simulation platform to evaluate the efficacy of the developed combinatorial rules. In literature, for analyzing the dispatching rules, simulation is considered an impactful method. Representing fabrication facilities with considering all the constraints in a real-time simulation environment is a time-consuming and complicated task [15]. To establish a simulation model of the considered problem the generated combinatorial dispatching rules from MOEA as illustrated in Table 2 served as an input. Each lot of wafers in the simulation is specified as an entity and the processing time for each entity includes the setup time, in specific case rework and operator unavailability. The simulation model evaluates each combinatorial dispatching rule separately to calculate the performance of each element i.e., material handling, tools, process, etc. The advantage of this model is, it keeps a track of record that consists of various key performance parameters such as delay tie, throughput, and work in progress. Table 2. Combinatorial set of dispatching rules No. Dispatching rules Lot

Batch

AGV

1

FIFO

MBS

FCFS

2

SPT

MDBH CR

3

PRE-FIFO

MMBS FEFS

4

PRI-SPT

PUCH

EDD

For running the simulation model certain real-time constraints were considered as parameters settings in flexsim are stated as follows: i. The arrival time for each entity to a specific station is considered the same. ii. The lot size is considered as 24 wafers per lot and it’s a constant in the model. iii. Layers for each wafer are fixed as 10.

Production Scheduling of Semiconductor Wafer Fabrication Facilities

87

For batch operations settings, when a batch of entities is being processed on a processing station, another set keeps waiting in a queue until the processing is being completed. When the queue gets empty a new batch is generated and gets picked by the AGV to the corresponding processing station conducting the following operation. The time period a batch waits in the queue is stated as delay time. For each set of experiments, the combination of rules is changed and the corresponding results are stored. Each experiment is repeated for n replications for obtaining an optimal set of rules.

5 Experimentation and Results In this section, we considered a semiconductor wafer fabrication system, that fabricates 20,000 wafers per month. The manufacturing plant has 16 processing tools, 8 upper and lower processers out of which two of them are batch machines used to dispatch batches. In total, there are 3 AGVs present in this layout, 1 busy OHT, and the rest are idle OHTs. A set of 25 wafers forms a lot and a set of 10 such lots form a batch. In every 72 h, production of 2 NPW lots, 1 hot lot and 25 regular lots take place. In every 24 h, 2 NPW lots having a processing time 1/25 times that of a regular lot is produced. A snapshot of the real-time digital twin of the shop floor is illustrated in Fig. 7.

Fig. 7. FlexSim based simulation model

The defined R function can be utilized to optimize the parameters using the RSM based on Eq. (1). The following results from the analysis are depicted in Table 3.

88

S. Panigrahi et al. Table 3. Regression analysis based on two factors Term value Constant

1.93088 −0.00180

Co-efficient SE Coef T- value

p/g

0.00399 484.27

P-value

0.000

C

−0.12049 −0.00286

0.00295 −0.67

M 0.00295

−40.89

0.561

0.00295 −0.97

0.000

0.364

The results obtained from the study are articulated in Table 4 along with the results of change [16]. From the results, it can be concluded that sequence 1 i.e., FIFO-MBS-FCFS indicates similar values when compared to results from the literature. Table 4. A comparison study of results Dispatching combinatorial rules (Lot- Batch- AGV)

CT Minutes

WIP Minutes

Time Delay in minutes

No.

Rules proposed in [16]

1

FIFO-MBS-FCFS

66.324

3.269

26.42

2

SPT-MDBH-CR

44.599

1.6904

19.23

3

PRI-FIFO- MMBS -FEFS

149.29

8.8457

44.95

4

PRI-SPT-PUCH-EDD

41.8199

1.8594

19.67

No.

Proposed combinatorial rules

1

FIFO-MBS-FCFS

66.210

3.417

26.26

2

SPT-MDBH-CR

22.07

0.756

9.7665

3

PRI-FIFO- MMBS -FEFS

70.52

3.5798

26.65

4

PRI-SPT-PUCH-EDD

27.093

0.9043

12.26

6 Conclusion One of the most extensive, high-powered, and complex manufacturing industries continues to exist in today’s world and those are the semiconductor industries. At times it can get very tedious to process large batches and dispatch them. This includes batch processing, multiple machines to process, re-entrant flows, and sequence-dependent setup times. This paper proposes an algorithm-based simulation-optimization approach to generate combinatorial dispatching rules automatically based on NSGA-II. It is necessary to optimize the parameters of the adopted multi-objective evolutionary algorithm and to do so a design of experiments has been conducted with response surface methodology (RSM). For improved combinatorial dispatching rules, these optimized parameters are

Production Scheduling of Semiconductor Wafer Fabrication Facilities

89

very advantageous. We also conducted experimental studies to find if the developed technique would be able to perform better on the marked performance measures i.e., maximum completion time required for the last job, work in process, and time delay. Later, we formulated a real-time simulated environment using those parameters in it to understand the productivity of the generated rules. Furthermore, we compared our rules with the previously developed rules from the literature exhibiting the robustness of our automatically generated rules.

References 1. Shim, J., Kang, S., Cho, S.: Active learning of convolutional neural network for cost-effective wafer map pattern classification. IEEE Trans. Semicond. Manuf. 33(2), 258–266 (2020) 2. Pfund, M.E., Balasubramanian, H., Fowler, J.W., Mason, S.J., Rose, O.: A multi-criteria approach for scheduling semiconductor wafer fabrication facilities. J. Sched. 11(1), 29–47 (2008) 3. Waschneck, B., et al.: Deep reinforcement learning for semiconductor production scheduling. In: 2018 29th Annual SEMI Advanced Semiconductor Manufacturing Conference (ASMC), pp. 301–306. IEEE (2018) 4. Tadumadze, G., Emde, S., Diefenbach, H.: Exact and heuristic algorithms for scheduling jobs with time windows on unrelated parallel machines. OR Spectrum 42(2), 461–497 (2020). https://doi.org/10.1007/s00291-020-00586-w 5. Oukil, A., El-Bouri, A.: Ranking dispatching rules in multi-objective dynamic flow shop scheduling: a multi-faceted perspective. Int. J. Prod. Res. 59(2), 388–411 (2021) 6. Chiang, T.C., Fu, L.C.: Using dispatching rules for job shop scheduling with due date-based objectives. Int. J. Prod. Res. 45(14), 3245–3262 (2007) 7. Barhebwa-Mushamuka, F., Dauzère-Pérès, S., Yugma, C.: Multi-objective optimization for Work-In-Process balancing and throughput maximization in global fab scheduling. In: 2019 IEEE 15th International Conference on Automation Science and Engineering (CASE), pp. 697–702. IEEE (2019) 8. Wu, M.C., Chiou, S.J., Chen, C.F.: Dispatching for make-to-order wafer fabs with machinededication and mask set-up characteristics. Int. J. Prod. Res. 46(14), 3993–4009 (2008) 9. Kim, Y.D., Kim, J.G., Choi, B., Kim, H.U.: Production scheduling in a semiconductor wafer fabrication facility producing multiple product types with distinct due dates. IEEE Trans. Robot. Autom. 17(5), 589–598 (2001) 10. Chen, J.H., Fu, L.C., Lin, M.H., Huang, A.C.: Petri-net and GA-based approach to modeling, scheduling, and performance evaluation for wafer fabrication. IEEE Trans. Robot. Autom. 7(5), 619–636 (2001) 11. Song, D.P., Hicks, C., Earl, C.F.: An ordinal optimization based evolution strategy to schedule complex make-to-order products. Int. J. Prod. Res. 44(22), 4877–4895 (2006) 12. Mathirajan, M., Sivakumar, A.I.: A literature review, classification and simple meta-analysis on scheduling of batch processors in semiconductor. Int. J. Adv. Manuf. Technol. 29(9–10), 990–1001 (2006) 13. Kim, H., Lim, D.E., Lee, S.: Deep learning-based dynamic scheduling for semiconductor manufacturing with high uncertainty of automated material handling system capability. IEEE Trans. Semicond. Manuf. 33(1), 13–22 (2020) 14. Jeong, K.C., Kim, Y.D.: A real-time scheduling mechanism for a flexible manufacturing system: using simulation and dispatching rules. Int. J. Prod. Res. 36(9), 2609–2626 (1998)

90

S. Panigrahi et al.

15. Mönch, L., Fowler, J.W., Mason, S.J.: Production Planning and Control for Semiconductor Wafer Fabrication Facilities: Modeling, Analysis, and Systems. Springer, New York (2012). https://doi.org/10.1007/978-1-4614-4472-5 16. Chang, X., Dong, M., Yang, D.: Multi-objective real-time dispatching for integrated delivery in a Fab using GA based simulation optimization. J. Manuf. Syst. 32(4), 741–751 (2013)

Microindentation Hardness Testing of D – gun Sprayed Coatings Oleksiy Romanchenko(B) , Oleksandr Lohunov , Yuriy Kharlamov , Volodymyr Sokolov , and Oleg Krol Volodymyr Dahl East Ukrainian National University, 59-a, pr. Tsentralnyi, Severodonetsk 93400, Ukraine

Abstract. The prospects of application for the method of non-destructive testing by continuous indentation of an indenter for determination of the mechanical properties for thermal sprayed coatings are considered. Research to assess the microhardness of composite coatings based on the tungsten and chromium carbides obtained by the D – gun spraying have been carried out. The penetration diagrams of coatings made of powders WC – 8% Co (VK8), WC – 15% Co (VK15), WC – 18% Co (VK18S), 3%Cr – 2%C – 15% Ni (KHN15S) are considered. It is shown that under optimal spraying conditions, the microhardness of obtaining coatings is close to the microhardness of sintered hard alloys of the same composition. The values of variation coefficients of microhardness values with an increase of the indentation depth for specimens from sintered hard alloy and sprayed coatings were determined. Under optimal spraying conditions the maximum values of microhardness and more homogeneous structure and mechanical properties of coatings are achieved. Keywords: Indenter · Indentation depth · Variation coefficient · Microhardness · Spraying modes · Phase composition

1 Introduction The application of protective coatings in mechanical engineering provides a decrease of material and energy consumption of production and operation of machines, an increase of their reliability, allows to solve the problems of creating and mastering of new products of modern technology [1–5]. Methods of thermal spraying of coatings from powder materials are the most promising, and among them is D – Gun spraying [6–10]. The main advantages and important advantages of the D – Gun spraying are: low requirements to the sprayed surface quality; possibility of obtaining durable coatings; wide opportunities of thermal cycle regulation of formed coating and product; sufficiently high growth rate of the coating thickness; high density coatings; high reliability and relative simplicity of technological equipment, etc. The development and introduction into production of modern technological processes and equipment [11–15] for obtaining of high-performance protective coatings based on D – Gun spraying requires a comprehensive research of their mechanical properties. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 91–99, 2022. https://doi.org/10.1007/978-3-030-83368-8_9

92

O. Romanchenko et al.

2 Literature Review An important direction of development of research methodology of resource capabilities and service characteristics of structural materials is the localization of the studied volumes of material to the micro- and nanometer scale. The resulting physical and mechanical properties are significantly different from the properties that appear on the macroscopic scale [16–20]. Conventional methods that are usually used for determination of materials properties in macro volumes are unsuitable for studying and diagnosing objects with a structure of the order of 1 µm or less. At present, a methodology for determination of mechanical properties by precision means is being developed within the framework of a promising scientific direction, which is called the mechanics of macro - and nanotesting [20–25]. Indentation is the most accurate, theoretically and experimentally confirmed standardized method for determination of the physical and mechanical properties of the micro- and nanometer scale for materials with a fine structure, a local gradient of properties, materials in a state of severe plastic deformation or after thermal action, as well as superhard, brittle and elastic materials, the mechanical properties of which cannot be determined in other ways [26–30]. The objective of this work is experimental research of mechanical properties of D – Gun spraying coatings based on tungsten and chromium carbides by micro-indentation method with the registration of the tip indentation process in form of «load – unload» diagram.

3 Research Methodology The method of measuring indentation is as follows [11, 21]: an indenter of a known shape is pressed into the sample surface under the action of a load P at a constant speed. When a given load or indentation depth is reached, the movement stops for a certain time to keep the material under load. After that, the indenter is retracted in the opposite direction. In the process of loading and retraction of the indenter (unload), the values of the load and corresponding displacements are recorded. The resulting dependence (Fig. 1) is a curve of «load on the indenter – indentation depth» (or «load – unload»). In fact, the «load – unload» diagram represents the work spent by the indenter to overcome the material resistance (the area under the load branch S1), and the work spent by the material to restore its properties (the area under the unload branch S2). In conventional testing, the microhardness is found as the ratio of the load (applied to the indenter) to the square of indenter’s print after testing. This leads to the loss of information about the mechanical behavior of the material during testing. During these tests, the dependence of the diamond indenter displacement on the load is recorded, both with an increase and decrease of load. The displacement of indenter is measured with high resolution, which allows testing at low loads and indentation depths. As a result, there is no need for difficult unreliable measurement of indenter’s print size obtained at low loads on the indenter. The hardness of measured by a non – restored print with registration of its depth under load («non – restored» hardness) is not equal to hardness measured by restored (unloaded) print using a microscope («restored» hardness). The hardness is calculated according to the first branch of the diagram – the dependence of

Microindentation Hardness Testing of D – gun Sprayed Coatings

93

penetration depth of indenter on the load, and since the indenter unloading process is elastic, the elastic modulus is determined by unload curves. Using the selected theoretical model, the hardness is automatically calculated by experimental curve.

Fig. 1. Diagram «load – unload» during indentation and profile of the indenter’s print in material: h1 – maximum indentation depth, h2 – reconstructed depth.

4 Results In accordance with the described technique, experimental research of the physical and mechanical properties of D – Gun spraying coatings were carried out, the coatings were obtained from a standard powders of alloys WC – 8% Co (VK8). The measurement of kinetic microhardness on each sample were carried out at least six times. Under optimal spraying conditions, the microhardness of coatings is close to the microhardness of sintered hard alloys of the same composition. A change in nature and intensity of phase transformations of powders under various technological conditions also leads to a change of mechanical properties of obtaining coatings. The X – ray studies of coatings phase composition on samples for microhardness measurement were performed. Qualitatively, the phase composition of coatings differed insignificantly; at the same time, the average values of the microhardness of samples varied from 10418 to 22495 MH/m2 through different compositions of coatings phases. From the analysis of load branches of micromechanical test diagrams, important conclusions were made: branches of load for sintered hard alloy are steeper than for sprayed coatings. So, the coatings contain phases with lower hardness; the hardness of the coatings largely depends on the composition of the detonating acetylene – oxygen mixture; influence of powder consumption, which controlled by the thickness of a single spraying spot and the spraying distance are noticeable. With an increase of spraying distance from 120 to 150 mm and the thickness of a single spray spot from 7 to 13,7 µm, the hardness of coatings slightly increases, as evidenced by the steeper slope of the load branch. An increase of spraying distance from 150 to 180 mm and a decrease of a single spray spot thickness from 13 to 6,8 µm led to a decrease of coatings hardness and steepness of the load branches. Therefore, there are optimal values of the powder consumption, at which the intensity of undesirable phase transformations in the original powder is suppressed. At optimal spraying conditions, the following regularities are observed: there is a certain correspondence between the ratio of parameters of the curves of microhardness

94

O. Romanchenko et al.

values from indentation depth and the ratio of parameters of load branches for same samples; with a change of indentation depth for a sample with large values of microhardness, a greater difference of its values on curve is observed; the microhardness of samples from a sintered hard alloy is higher at practically all values of indentation depth than in all sprayed samples; the magnitude and nature of changes of microhardness values with increasing indentation depth depend largely on gas composition, powder consumption and spraying distance; for a number of samples the presence of a minimum on the microhardness curve at indentation depths of 1,75–3,5 µm is observed. The coefficient of variation reaches its highest values at a small indentation depth. This can partly be explained by measurement errors associated with the microgeometry of surface of machined coatings. The most stable measurement results (low values of the coefficient) were obtained for samples, that showed high values of microhardness. This indicates the superior homogeneity of their mechanical properties. The regularity of the decrease of hardness of refractory compounds with increasing load, especially with values 0–3 N, is common [7]. The increased scatter of values of coatings microhardness in the area of low loads can be explained by the mobility of crystals of surface layer. Crystals of the surface layer are only unilaterally bound to neighboring grains, and upon deformation of the sample in elastic area, they can smoothly or abruptly return to angles of tens of minutes and noticeably shift. A significant increase of microhardness with a decrease of indentation size can be explained by high rate and degree of strain hardening due to facilitated conditions for nucleation and motion of dislocations near surfaces. The increase of hardness at small values of indenter penetration depth (and thus at low loads) for coatings based on refractory carbides, depending on indentation depth upon penetration of indenter, is the same as for many brittle refractory compounds and hard alloys. For tungsten carbide and its alloys containing 4–20% Co at a force of less than 3 N, the hardness increases significantly [4]. One of the main models for explaining the fact of an increase of «non-restored» microhardness in area of small forces (deviation from the Kirpichev – Kick mechanical similarity law) is the idea that the general shape change upon indentation is realized mainly as a result of elastic deformation, acts at relatively large distances, and to a lesser extent – as a result of plastic deformation proceeding by dislocation path and localized in a narrow area. The increase of hardness at low forces is explained either by an increase of contribution of work of material elastic deformation, or by influence of scale factor, that determined by the structure [4, 16, 17]. The scale factor is possibly manifested in the fact that in low – cobalt alloys a large number of WC – WC contacts enhances the «skeleton» effect of dislocation network under the indenter. The diagrams of indentation of indenter into the coating of spheroidized powders of chromium carbides with nickel 3%Cr – 2%C – 15%Ni (KHN15S), curve 1, mechanical mixture of powders WC – 15% Co (VK15), curve 2, and spheroidized powder WC – 18% Co (VK18S), curve 3, are presented in Fig. 2. Comparing them with similar data for coatings from a standard powders of alloys WC – 8% Co (VK8) shows that the load branches for these coatings are less steep. Their hardness, especially at low loads, is much lower. An increase of hardness of test material leads to a regular increase of growth rate of microhardness with a decrease of indentation size [30]. This is confirmed

Microindentation Hardness Testing of D – gun Sprayed Coatings

95

by the experimental data. In the latter case, the coatings contain more metal bonds, that reduce the overall hardness of coating, and also increase the probability of initial contact of indenter with metal component of alloy. A similar phenomenon was observed for micro – indentation of hard alloys of the WC – Co system [4]. For a solid solution based on cobalt and viscocobalt alloy (VK50), hardness does not depend on force in the range of 1 … 12,5 N. For tungsten carbide and low – cobalt alloys (4 … 20% Co), with a force of less than 3 N occurs deviation and the hardness increases significantly.

Fig. 2. Diagrams of indentation of a pyramidal indenter into a coating from powders: spheroidized chromium carbide with nickel KHN15S (1); mechanical mixture of tungsten carbide with cobalt VK15 (2); spheroidized tungsten carbide with cobalt VK18S (3).

The dependences of the microhardness of coatings made of VK15 powder in the center and at the periphery of spraying spot-on penetration depth of the indenter are presented in Fig. 3. The hardness of coatings at periphery of the spot is slightly higher, which can be explained by the less intense transformations of WC → W2 C → W in process of single coating layer cooling. In the center of spraying spot, the thickness of a single layer is higher, and at the spot periphery the faster cooling of coating is associated with its smaller thickness, but also with possibility of heat removal into and along surface of substrate. Moreover, apparently, the coating contains a significant amount of W. The heterogeneity of coating structure is much higher than that of coatings made of VK8 powder.

Fig. 3. Dependences of coatings microhardness on penetration depth of indenter in the center of the spray spot (1) and at the periphery of the spot (2), powder – mechanical mixture of tungsten carbide with cobalt VK15.

96

O. Romanchenko et al.

The coefficient of microhardness variation is approximately two times higher than that of coatings made of VK8 powder. If we proceed from the composition of initial powder, then one should expect an increase of structure homogeneity and, thus, microhardness of coatings with an increase of Co content, especially in high loads area. However, one should take into account the differences in coating technology. A coating of VK8 powder was applied on a gun with a short barrel of variable cross – section and a powder supply system, that ensures uniform distribution. The other coatings were applied on a gun with a cylindrical barrel; and uniform distribution of powder over the cross – section of the barrel was not ensured. In addition, the powder was injected continuously through an open tube, so some of it was exposed to hot combustion products before entering the barrel. The dependences of microhardness of coatings with spheroidized powder of tungsten carbide with cobalt VK18S with a particle size of about 50 µm are presented in Fig. 4. In this case, the conversion WC → W2 C → W proceeds less intensively. Differences of hardness of coatings in the center and at the periphery of the spraying spot are observed. In the center of the spraying spot there is a pronounced minimum of microhardness value at an indenter penetration depth of 1,5 µm. A high concentration of powder in the center of D – gun jet leads to a slower cooling of central area of a single layer and to an intensification of its interaction with the surrounding gas medium, and thereby the oxidation and transformation of WC → W2 C → W. Therefore, within the thickness of a single layer, the formation of a weakened surface area is possible, which further determines the structure of the boundary area between single layers. In addition, a high concentration of powder in the center of jet promotes an increase of coating porosity. This is confirmed by higher value of variation coefficient of microhardness at a penetration depth of 1,5 µm.

Fig. 4. Dependences of coatings microhardness on penetration depth of indenter in the center of the spray spot (1) and at the periphery of the spot (2), powder – spheroidized tungsten carbide with cobalt VK18S.

The similar dependences of coatings microhardness from spheroidized KHN15S powder are presented in Fig. 5. In this case, the dependence of hardness on the penetration depth of indenter is less pronounced, especially in the center of the spraying spot. This is consistent with the previously described dependences of hardness of materials in area of low loads. The hardness of chromium carbides is significantly lower than tungsten carbides. A wide deflection of microhardness curve in the range of penetration depths

Microindentation Hardness Testing of D – gun Sprayed Coatings

97

of 1,5–3 µm is also characteristic, which can be explained by the lower hardness of chromium carbides, higher porosity of coatings, and increased fragility. In most cases, cracks are forming around the prints.

Fig. 5. Dependences of coatings microhardness on penetration depth of indenter in the center of the spray spot (1) and at the periphery of the spot (2), powder – spheroidized chromium carbide with nickel KHN15S.

5 Conclusions Mechanical tests by continuous indentation of the indenter are a convenient express method for determining the microhardness and mechanical properties of hard coatings based on refractory carbides obtained by D – gun spraying and other methods of thermal spraying. Under optimal conditions of D – gun spraying of fine powders of WC – Co alloy, the mechanical properties of obtained coatings are close to baked alloys of the same composition. The microhardness of coatings is significantly influenced by composition of the detonating acetylene – oxygen mixture, the spraying distance and powder consumption during single spraying cycle. For coatings obtained at optimal conditions, the higher homogeneity of mechanical properties is characteristic, that is, low values of the variation coefficient.

References 1. Zhigachev, A., et al.: Ceramic Materials Based on Zirconium Dioxide. Technosphere, Moscow (2018) 2. Ivanov, V., Dehtiarov, I., Pavlenko, I., Liaposhchenko, O., Zaloga, V.: Parametric optimization of fixtures for multiaxis machining of parts. In: Hamrol, A., Kujawi´nska, A., Barraza, M.F.S. (eds.) MANUFACTURING 2019. LNME, pp. 335–347. Springer, Cham (2019). https://doi. org/10.1007/978-3-030-18789-7_28 3. Kotliar, A., Basova, Y., Ivanov, V., et al.: Ensuring the economic efficiency of enterprises by multi-criteria selection of the optimal manufacturing process. Manage. Prod. Eng. Rev. 11(1), 52–61 (2020) 4. Tumanov, V., et al.: Deformation and elastic properties of hard alloys of the WC-Co system during micro- and macro-indentation. Probl. Strength 3, 99–103 (1985)

98

O. Romanchenko et al.

5. Rogovyi, A., Khovanskyy, S., Grechka, I., Pitel, J.: The wall erosion in a vortex chamber supercharger due to pumping abrasive mediums. In: Ivanov, V., et al. (eds.) DSMIE 2019. LNME, pp. 682–691. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-22365-6_68 6. Tretyakov, V.: Fundamentals of Metal Science and Technology of Sintered Alloys Production. Metallurgy, Moscow (1976) 7. Andrievsky, R., Lanin, A., Rymashevsky, G.: Strength of Refractory Compounds. Metallurgy, Moscow (1974) 8. Samsonov, G., Vitryanyuk, V., Chaplygin, F.: Tungsten carbides. Nauk. Dumka, Kiev (1974) 9. Kundrák, J., Mitsyk, A., Fedorovich, V., Morgan, M., Markopoulos, A.: The use of the kinetic theory of gases to simulate the physical situations on the surface of autonomously moving parts during multi-energy vibration processing. Materials 12(19), 3054 (2019) 10. Fedorovich, V., Mitsyk, A.: Mathematical simulation of kinematics of vibrating boiling granular medium at treatment in the oscillating reservoir. Key Eng. Mater. 581, 456–461 (2014) 11. Oliver, W.C., Pharr, G.M.: An Improved technique for determining the hardness and elastic modulus using load displacement sensing indentation experiments. J. Mater. Res. 7, 1564– 1583 (1992) 12. Sokolov, V., Krol, O., Baturin, Y.: Dynamics research and automatic control of technological equipment with electrohydraulic drive. In: 2019 International Russian Automation Conference (RusAutoCon). IEEE (2019) 13. Krol, O., Sokolov, V.: Parametric modeling of transverse layout for machine tool gearboxes. In: Gapi´nski, B., Szostak, M., Ivanov, V. (eds.) MANUFACTURING 2019. LNME, pp. 122–130. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-16943-5_11 14. Krol, O., Sokolov, V.: Modeling carrier system dynamics for metal-cutting machines. In: 2018 International Russian Automation Conference (RusAutoCon). IEEE (2018) 15. Sokolov, V., Krol, O., Romanchenko, O., Kharlamov, Y., Baturin, Y.: Mathematical model for dynamic characteristics of automatic electrohydraulic drive for technological equipment. In: Journal of Physics: Conference Series, vol. 1553, p. 012013 (2020) 16. Mott, B.: Micro-Indentation Hardness Test. Metallurgizdat, Moscow (1960) 17. Grigoriev, O., et al.: Resistance of covalent crystals to micro-indentation. Powder Metall. 8, 72–80 (1977) 18. Rogovyi, A., Khovanskyi, S., Hrechka, I., Gaydamaka, A.: Studies of the swirling submerged flow through a confuser. In: Ivanov, V., Pavlenko, I., Liaposhchenko, O., Machado, J., Edl, M. (eds.) DSMIE 2020. LNME, pp. 85–94. Springer, Cham (2020). https://doi.org/10.1007/ 978-3-030-50491-5_9 19. Pavlenko, I., Trojanowska, J., Ivanov, V., Liaposhchenko, O.: Parameter identification of hydro-mechanical processes using artificial intelligence systems. Int. J. Mechatron. Appl. Mech. 5, 19–26 (2019) 20. Ivanov, V., Dehtiarov, I., Pavlenko, I., Kosov, M., Hatala, M.: Technological assurance and features of fork-type parts machining. In: Ivanov, V., et al. (eds.) DSMIE 2019. LNME, pp. 114–125. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-22365-6_12 21. Khokhlova, Y., Ishchenko, D., Khokhlov, M.: Indentation from macro to nanometer level and examples of studying the properties of materials with a special structure. Tech. Diagn. Nondestr. Control 1, 30–36 (2017) 22. Gogolinsky, K., Syasko, V.: Methods and means for monitoring of mechanical properties of micro- and nanometer coatings and modified near-surface layers. In: The World of NonDestructive Control, vol. 3, no. 61, pp. 43–48 (2013) 23. Bulychev, S., Alekhin, V.: Material Testing by Continuous Indentation of the Indenter. Mechanical Engineering, Moscow (1990)

Microindentation Hardness Testing of D – gun Sprayed Coatings

99

24. Kovalevskyy, S., Kovalevska, O., Koshevoy, A., Tasi´c, I.: Using wave signatures for identifying mechanical objects. In: IOP Conference Series: Materials Science and Engineering, vol. 568, p. 012117 (2019) 25. Sokolov, V.: Transfer functions for shearing stress in nonstationary fluid friction. In: Radionov, A.A., Kravchenko, O.A., Guzeev, V.I., Rozhdestvenskiy, Y.V. (eds.) ICIE 2019. LNME, pp. 707–715. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-22041-9_76 26. Babich, M.: Inhomogeneity of Hard Alloys by Carbon Content and Its Elimination. Nauk. Dumka, Kiev (1975) 27. Goldschmidt, H.: Implementation Alloys. Mir, Moscow (1971) 28. Kiparisov, S., Perelman, V.: Some aspects of deformation theory of powder and composite materials as the basis for improvement of the processes of pressing and sintering. In: Progressive Technological Processes in Powder Metallurgy, pp. 61–76 (1982) 29. Krol, O., Sokolov, V.: 3D modelling of angular spindle’s head for machining centre. In: Journal of Physics: Conference Series, vol. 1278, p. 012002 (2019) 30. Ternovsky, A., et al.: The nature of the dependence of microhardness on the size of the indentation and peculiarities of materials deformation in conditions of concentrated loading of surface. In: New in the Field of Microhardness Testing, pp. 71–81 (1974)

Application of Magnetic Field on Lubricating Cooling Technological Condition in Metal Cutting Process Umarov Erkin , Mardonov Umidjon(B)

, and Shaozimova Umida

Tashkent State Technical University, Tashkent 100097, Uzbekistan

Abstract. In this paper, the implementation of static magnetic field on lubrication cooling technological condition in the machining process is discussed. Moreover, the new construction of UMD-1 and SMD-2 magnetizing devices is given in the article. Nowadays, increasing the wear resistance of cutting tools used in metal cutting processes is one of the controversial issues of the manufacturing industry. Using the effect of magnetic field on lubricating cooling technological condition can be a modern solution to this problem. For this purpose, the authors designed a new device for magnetizing flowing lubrication cooling liquids in the machining process. Technical specifications of UMD-1 and SMD-2 magnetizing devices are also given in the paper. Keywords: Cutting process · Magnetic field · Magnetized cutting fluid

1 Introduction Over the last decade, a lot of researches about the effect of magnetic field on the physical and chemical properties of water have been published. When magnetic field effects on liquids they become magnetized liquids and there are some differences between physicalchemical properties of magnetized and non-magnetized liquids. Nikolskiy et al. analytically and experimentally investigated the influence of magnetic and electric fields with tension gradient in the direction of the movement of the contacting gas-liquid phases [1]. Rashid et al. presented an investigation of water evaporation through a magnetic field of 0.5 T, which was located at different locations of tested water height (water-air interface, water mid-height, and bottom) [2]. Mardonov et al. studied the effect of magnetic field on the dynamic viscosity coefficient of flowing water, they found that the dynamic viscosity coefficient of water changed after magnetization. They magnetized flowing water using different magnetic field strengths. After magnetization, they measured the coefficient and found that the dynamic viscosity of water was decreased [8]. The effect of magnetic field on liquids is still a controversial issue, and there is a lack of researches in this field. Almost all of the researches about the magnetic field effect on liquids were conducted on water and the majority of those researches used in the agriculture field effectively. The authors of this paper suggest that use of magnetic field treatment in manufacturing process would have a great advantage. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 100–106, 2022. https://doi.org/10.1007/978-3-030-83368-8_10

Application of Magnetic Field

101

2 Effect of Magnetized Cutting Fluid on the Metal Cutting Process The manifestation of the indicated actions of cutting fluids and other lubricants and a beneficial effect on their cutting process and wear of the cutting tool are possible only if the liquid penetrates the contact surface. At a low cutting speed, the contact zone between the processed material and the tool is not continuous, the entire zone is dotted with the smallest pores (capillaries) ranging in size from fractions to several tens of micrometers. The resulting chips are also penetrated by capillaries throughout their entire thickness. Through these capillaries, the cutting fluid enters the friction zone of the tool with the processed material [4–6]. Periodic braking and seizure of individual volumes of the processed metal on the surface of the tool lead to the formation of vacuum cavities that facilitate the penetration of liquid. Since the cutting process is always accompanied by oscillations of the machine-fixture-tool-part system, and the oscillations of the workpiece (low-frequency) and the tool (high-frequency) do not coincide in phase and amplitude, the contacting surfaces are periodically opened for the penetration of coolant. Although the mechanism of coolant penetration in the cutting zone has not been completely studied, it is indisputable and established that the coolant penetrates the contact surface and significantly affects the cutting process, the wear rate of the cutting tool, and the quality of surface treatment. Since the temperature in the cutting zone is almost always higher than the boiling point of water, the process fluid enters the contact surface not in the usual state, but in the form of vapors and individual particles - molecules, radicals, or ions of the substances that make up the liquid. The efficiency of the process fluid depends not only on its chemical composition but also on the method of feeding it into the cutting zone. Most often, the coolant enters the cutting zone in the form of a freely falling jet. In addition, in some cases, high-pressure jet cooling is used, in which the coolant is supplied under a pressure of 2–3 MPa to the rear surface of the cutting part of the tool with a thin jet through a nozzle with a diameter of 0.3–0.5 mm. The method of cooling and lubricating the cutting zone with sprayed liquids is relatively often used; at the same time, a jet of the air-liquid mixture obtained in a special spray device is supplied to the cutting zone [7–11]. In terms of the mechanism of action on the cutting process, magnetized cutting fluids are apparently similar to those of cutting fluid, but unlike the latter, magnetized cutting fluids can be fed into the cutting zone using a magnetic field. Under its influence, the washing effect is enhanced due to the improvement of wettability, the lubricating effect is intensified since the conditions for the penetration of magnetized cutting fluids on the contact surface are improved. Magnetized cutting fluids have a stronger cooling effect since they surpass the base in terms of heat capacity and thermal conductivity. Magnetized cutting fluids can be used when conventional cutting fluids and methods of their suppliers are not applicable, for example, when using power-operated hand tools, when working at high altitudes, in confined isolated spaces, and other special conditions. In this regard, the use of magnetized cutting fluids in cutting metals can and should be considered as a new way of supplying a technological medium to the cutting zone using a directed magnetic field. Magnetized cutting fluids can be used, for example, as technological environments for collecting non-magnetic materials. Such material can be not only non-magnetic

102

U. Erkin et al.

metal shavings but also other non-magnetic objects such as industrial waste, garbage, and others. Highly dispersed materials such as sludge and abrasive dust from metal abrasives can be collected. By spraying magnetized cutting fluids in a stream of dusty air, magnetized cutting fluids droplets wet particles of nonmagnetic abrasive dust and make them magnetically controlled, subject to the influence of a magnetic field through which a stream of dusty air passes. In a magnetic field, the air is purified, since dust particles moistened with magnetized cutting fluids are attracted to the poles of the magnets under the influence of the MDF of an uneven magnetic field and settle on them. As the effect of magnetic fields on liquids is becoming a controversial issue in modern science, new devices used for magnetizing liquids would be one of the important inventions.

3 Magnetizing Devices 3.1 Specification of UMD-1 Magnetizing Device For the application of magnetic field in the machining process, it is essential to design a new magnetizing device. Universal Magnetizing Device - UMD-1 is developed to magnetize flowing liquids or more accurately, it is very comfortable to apply it to any metal cutting machine (lathe, milling machine, grinding machine). The main purpose of the UMD-1 magnetizing device is to hold the super magnets with different poles and change the distance between these magnets (Fig. 1).

Fig. 1. UMD-1 magnetizing device. 1–4 – Supports, 5–8 – magnet holders, 9–14 – Horizontal hexagon bolts, 15–20 – Vertical hexagon bolts, 21–28 – Screws to fix magnet holders, 29, 30 – Magnets.

Figure 1 shows that, magnets (29, 30) are placed on lower (1, 4) and upper (2, 3) supports with different poles. And long horizontal hexagon bolts (9–14) are fixed according to magnets width, then holders (5–8) are fixed to hold magnets hard by screws (21–28). Holders and supports are developed to change the distance between them. Lower and upper supports have the same construction and the distance between the lower and upper supports can be changed easily by vertical hexagon bolts (15–20). Changing the distance between the lower and upper supports is purposed to change the distance between the magnets placed on lower and upper supports. We can easily change

Application of Magnetic Field

103

the magnetic field strength by changing the distance between lower and upper magnets. Moreover, the distance between the horizontal supports can be changed according to the sizes of the magnets. The design of the UMD-1 magnetizing device is universal because it is possible to place magnets with different sizes (length, width, height). Magnets that have a minimum size of 20 × 10 × 5 (mm) and a maximum size of 300 × 80 × 30 mm can be placed on UMD-1 magnetizing device. The distance can be changed from 0 to 50 mm with respect to the sizes of the magnets placed on the supports. Furthermore, UMD-1 magnetizing device is very portable and it is very easy to change its position on cutting machines. 3.2 Magnetizing Scheme of Flowing Cutting Fluids Magnets (1) were placed on the other magnets with the distance between them. They were separated by long hexagon bolts between them. Different types of magnetic field strength can be selected to magnetize the liquids and magnetic field strength was controlled by changing the distance between magnets, or more clearly, we change the height with the help of hexagon bolts between upper and lower magnets to make the various magnetic field strength. The PVC pipe is placed between the middle of the magnets, and the diameter of it is can be chosen according to the height between the upper and lower magnets. Polyvinyl chloride (PVC) pipe passed through UMD-1 magnetizing equipment and the length of the magnetic field would be a maximum of 350 mm. When liquid flows through the PVC pipe, it turned into magnetized liquid. The scheme of the magnetizing process is given in Fig. 2.

Fig. 2. Magnetizing scheme of flowing liquid. 1 – UMD-1 magnetizing equipment, 2 – PVC pipe, 3 – container, 4 – pump, 5 – liquid, 6 – tap, 7 – flowing direction of the liquid.

SMD-2 special magnetizing laboratory device based on a solenoid was projected and built to make a pulsating electromagnetic field for magnetizing the liquids (Fig. 3). The device is made of 12 layer solenoid coils. Each layer contains 200 turns off wire thickness 1.06 mm. The pipes of the device have an inner diameter (18.75 mm) and outer diameter (25 mm) with wall thickness (6.25 mm). The length of the solenoid coil is 20 mm. The 220 V voltage from the source (9) is reduced by transformer 7 and passed through a

104

U. Erkin et al.

Fig. 3. Magnetizing scheme of flowing liquid with SMD-2 magnetizing device. 1-Solenoid, 2Ceramic pipe, 3-Container, 4-Pump, 5-Liquid, 6-Rheostat, 7-Transformer, 8-Voltage, 9-Pulse signal.

rheostat 6 in order to turn to a regulated pulsating voltage 9. The pulsating voltage from the diode bridge is transmitted to the solenoid and a pulsating electromagnetic field is created. The voltage value is controlled using a rheostat and is set to the desired voltage. Container 3 is filled with liquid 5, and the liquid in the container is removed from the container using pump 4 through ceramic tubes 2. The liquid that moving through the pipe following the direction that is shown in the diagram passes through the center of solenoid 1 with a diameter of 18.75 mm and is poured back into the container. The liquid coming through the tube passes through the center of the solenoid, which creates a pulsating electromagnetic field, and then turns into a magnetized liquid and is poured back into the container. The pressure of the liquid is controlled by tap 10 and the required pressure is obtained. After a few minutes of the cycle, the liquid in the container becomes fully magnetized.

4 Discussions It can be noted that magnetized cutting fluids expand the arsenal of tools for increasing the efficiency of machine-building production. They should not be used where conventional cutting fluids have traditionally been used quite effectively. At the same technological operations of mechanical processing of materials, where conventional cutting fluid is not used for some reason, magnetized cutting fluids can increase the wear resistance of the cutting tool, the quality of surface treatment, reduce cutting forces and torques, eliminate the breakdown of the threads of the cut thread and breakage of the cutting tool when machining parts made of titanium alloys, corrosion-resistant steels, and other difficult-to-process materials. Developed devices are very essential in the magnetic treatment of different liquids in the manufacturing process because the changes in physical and chemical properties of liquids have an impact on the wear resistance of the cutting tool used in the machining process. The influence of magnetized lubricating cooling liquids on the wear resistance

Application of Magnetic Field

105

of cutting tools helps to increase the efficiency in the machining process and saves energy [3, 16, 17]. However, this issue about magnetic field treatment of liquids remains controversial. Complete understanding of the influence of magnetic fields on fluids has a great impact on agriculture, industry, and other fields [12–14]. Some scientists are working on the issue of magnetic field treatment of water, but there lack of researches on this issue, especially the magnetic field effect on different liquids.

5 Conclusion As magnetic treatment of different liquids are being one of the important fields of the modern world and science, the requirement for magnetic treatment devices is increasing. Because of this, UMD-1 and SMD-2 universal magnetic treatment devices have been developed. These two devices are very easy to use in any field such as agriculture, manufacturing and etc. They are also portable to use in laboratory conditions while conducting scientific experiments. The design of the devices was constructed to control magnetic field strength by changing the distance between the magnets.

References 1. Nikolskiy, V.Y.: Primenenie elektromagnitnyx poley dlya intensifikatsii teplomassoobmena v sovmewennyx gazojidkostnyx protsessax [application of electromagnetic fields to intensify heat and mass transfer in combined gas-liquid processes]. Vostoch. Yevrop. J. Pered. Texnol. 3(8), 1729–3774 (2017) 2. Rashid, F.L., Hassan, N.M.: Increasing water evaporation rate by magnetic field. Int. Sci. Invest. J. 2, 61–68 (2013) 3. Umarov, E.O., Mardonov, U.T., Shoazimova, U.K.: Influence of the magnetic field on the viscosity coefficient of lubricoolant that is used in the cutting process. Int. J. Mech. Appl. Mech. 8(2), 144–149 (2020). https://doi.org/10.17683/ijomam/issue8.50 4. Kaminski, J., Alvelid, B.: Temperature reduction in the cutting zone in water-jet assisted turning. J. Mater. Process. Technol. 106(1–3), 68–73 (2000) 5. Lawal, S.A., Choudhury, I.A., Nukman, Y.: A critical assessment of lubrication techniques in machining processes: a case for minimum quantity lubrication using vegetable oil-based lubricant. J. Clean. Prod. 41, 210–221 (2013) 6. Dhar, N.R., Bhowmic, T.P., Chattopadhyay, A.B.: The influence of cutting fluids on the wear characteristics and life of carbide cutting tools. In: Proceedings of the International Conference on Mechanical Engineering, 26–28 December, Dhaka, Bangladesh, pp. 73–77 (2001) 7. Kurjmóto, T., Barrow G.: The wear of high speed steel cutting tools under the action of several different cutting fluids. In: Proceeding of the Twenty-Second International Machine Tool Design Research Conference, Manchester, UK, pp. 237–246 (1981) 8. Yildiz, Y., Gunay, M., Seker, U.: The effect of the cutting fluid on surface roughness in boring of low carbon steel–technical communication. Mach. Sci. Technol. 11, 553–560 (2007) 9. Yahya, I.: An experimental investigation on effect of cutting fluids in turning with coated carbides tool. J. Mech. Eng. 56, 1–7 (2010) 10. Debnath, S., Reddy, M.M., Yi, Q.S.: Environmental friendly cutting fluids and cooling techniques in machining: a review. J. Clean. Prod. 83, 33–47 (2014)

106

U. Erkin et al.

11. El-Hossainy, T.M.: Tool wear monitoring under dry and wet machining. Mater. Manuf. Processes 16(2), 165–176 (2001) 12. Wang, X.Y., Zhang, B., Gong, Z., et al.: The effect of a static magnetic field on the hydrogen bonding in water using frictional experiments. J. Mol. Struct. 1052(11), 102–104 (2013). https://doi.org/10.1016/j.molstruc.2013.08.021 13. Uljayev, E., Ubaydullayev, U.M., Tadjitdinov, G.T., Narzullayev, S.: Development of criteria for synthesis of the optimal structure of monitoring and control systems. In: Aliev, R.A., Yusupbekov, N.R., Kacprzyk, J., Pedrycz, W., Sadikoglu, F.M. (eds.) WCIS 2020. AISC, vol. 1323, pp. 559–563. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-68004-6_73 14. Cai, R., Yang, H., He, J., et al.: The effects of magnetic fields on water molecular hydrogen bonds. J. Mol. Struct 938(1–3), 9–15 (2009). https://doi.org/10.1016/j.molstruc.2009.08.037 15. Toledo, E.J.L., Ramalho, T.C., Magriotis, Z.M.: Influence of magnetic field on physical– chemical properties of the liquid water: insights from experimental and theoretical models. J. Molecul. Struct. 888(1–3), 409–415 (2008). https://doi.org/10.1016/j.molstruc.2008.01.010 16. Umarov, E.O., Mardonov, U.T.: Analysing the effect of magnetic field on lubricoolant in machining process. Int. J. Appl. Res. 6(5), 347–352 (2020) 17. Turakhodjaev, N., Tashbulatov, S., Zokirov, R., Tursunbaev, S., Baydullaev, A.: Studying the scientific and technological bases for the processing of dumping copper and aluminum slags. J. Crit. Rev. 7(11), 441–444 (2020). https://doi.org/10.31838/jcr.07.11.79

Rapid Prototyping of a Lower-Body Exoskeleton for Paraplegia Patients Filip-Alexandru Harmon1 and Patricia-Isabela Br˘aileanu2(B) 1 University Politehnica of Bucharest, FIM, Bucharest, RO 060042, Romania 2 University Politehnica of Bucharest, FIMM, Bucharest, RO 060042, Romania

[email protected]

Abstract. Since the early 2000s, powered exoskeletons have been used in military and industrial environments to increase the strength, stability, and overall safety of the user. Due to fast and notable improvements shown in the field of robotics and medicine in recent years, various medical exoskeletons have been developed, to aid paraplegia patients regain a certain degree of their lower body’s motor function. However, many of the models available on the market currently still present certain limitations and design issues, mainly due to the complexity of the human body’s construction and the user size variations. This paper aims to design a medical exoskeleton variant to solve some of the limitations and design issues still present in medical exoskeletons sold on the market today, and to optimize the price to quality ratio to the maximum degree possible, making this device a viable option even for people from disadvantaged countries. Keywords: Exoskeleton · Prototyping · Medical · Paraplegia

1 Introduction 1.1 About Paralysis Paralysis is an incurable condition in most of its forms and represents a serious trauma, both physiologically and psychologically. There are multiple variants of paralysis, depending on the affected area. The majority have a neurological root cause, for example a lesion of the spinal cord. The most frequent events that can lead to such a lesion are car accidents, shooting or stabbing, falls and sports injuries [1]. The main symptom of the disease is the loss of motor function in and below the affected area, often paired with loss of sensory function. Depending on the severity of the lesion, the affected area can be very large and in extreme cases, the patient can become paralyzed either from the waist down (paraplegia) or even from the neck down (tetraplegia). Various complications can result, proportional to the severity of the illness. Such complications can include urinary tract infections, loss of muscle mass and bone density, thrombosis, and various pulmonary, cardiovascular, and lymphatic problems [2]. The condition has an irrepairable impact on the life quality of the sufferer. Many individuals that suffer from a paralysis form will remain either immobilized in bed © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 107–118, 2022. https://doi.org/10.1007/978-3-030-83368-8_11

108

F.-A. Harmon and P.-I. Br˘aileanu

or dependent upon a wheelchair for the rest of their lives. Considering this situation, physical work can be excluded. In the case of a complete tetraplegia, most other jobs become impossible because the individuals cannot move their hands. Thus, many of the sufferers must give up their place in the work force. Furthermore, basic activities become a challenge, leaving the patient in constant need of a caretaker. In many cases, the individual is left with the family. Those who do not have family members alive or willing to take care of them are forced to hire a caretaker. The problem which arises is that the patient, in many cases unable to work, will not be able to financially afford a caretaker for a very long time. Unfortunately, humanitarian associations that aid individuals who suffer from paralysis are not available to many people, mostly due to economic difficulties. 1.2 Medical Diagnosis To diagnose the severity of the spinal cord injury, various standard scales are used, especially in the case of paraplegia, the main ones being Frankel, ASIA and Hauser [3]. Evaluation on the Frankel and ASIA scales is alphabetical, from A to E. Grade A represents a complete injury, which implies a total lack of motor and sensory function below the lesion. Grade E represents a minor injury, which can cause minor anomalies in the motor and sensory functions. These anomalies are not dangerous, and a grade E patient can be almost fully recovered. The Hauser scale represents an ambulation index, numbered from 0 to 9, which describes the patient’s symptoms depending on the distance they can travel in a certain time interval, with or without unilateral, bilateral or wheelchair support [4]. The use of such measuring scales is essential in finding the optimal path towards treatment, even in cases where motor or sensory function cannot be restored to its original state. 1.3 Therapies Available and Rehabilitation Exoskeletons Frequently used therapies in the case of paraplegia patients are the administration of certain glucocorticoids, electrostimulation and nerve decompression surgery. Other relatively recent therapies which are still experimental, such as bone marrow and stem cell transplantation can give hope to sufferers [3]. Despite the fact that branches of medicine such as neurology and orthopedics are continuously evolving and various studies regarding concepts such as nerve regrowth methods are ongoing, the therapies available at this moment are unfortunately not sufficient to treat paraplegia sufferers. Taking into consideration all the above mentioned, the only viable method available to aid paraplegia sufferers is the use of a medical exoskeleton unit. Rehabilitation exoskeletons are wearable motorized devices, which can help the individual walk without assistance, for an amount of time, depending on the efficiency of the batteries. Currently, there are a few lower-body exoskeleton models on the market. Their prices are highly influenced by various criteria such as production difficulty, battery life, material properties and functionality. Regarding full-body medical exoskeletons destined for tetraplegia sufferers, there are very few models available, most of them being experimental and not able to provide sustained usability. There is also limited literature on which to base an

Rapid Prototyping of a Lower-Body Exoskeleton

109

accurate research in this field, compared to lower-body exoskeletons. Overall, most medical exoskeletons are extremely expensive, and their availability is very limited, making it difficult for people who are financially disadvantaged to acquire one. The aim of this paper is to present the prototype of a 3D-printable, low-cost, wearable lower-body medical exoskeleton, which can be used as an alternative for people who are financially disadvantaged.

2 Virtual Prototyping 2.1 Disadvantages of Exoskeletons Exoskeleton units are still mostly in development, due to the limitations of current technology. Ever since their emergence in the late 1960s, these devices showed various problems, which have been solved to a certain extent. Some of the aspects that both industrial and military models struggled with are power consumption, restrained area of use and battery overheating. Devices that use power sources such as Li-Ion batteries tend to overheat, due to the nature of these accumulators. Thermal runaway is especially dangerous in the case of exoskeletons or exo–suits because they interface directly with the body. If the temperature exceeds the safe limit, the device can be severely damaged both internally and externally and more importantly, the wearer can suffer burns. In extreme cases of overheating or short circuiting, the accumulators can even explode [5]. Rehabilitation exoskeleton designers managed to solve some of the problems present in other types of exoskeletons, however, other demands arise, such as stair climbing function, better range of motion and even higher energy efficiency [6]. Compared to industrial exoskeletons, which can be tethered to a power generator all day long, rehabilitation exoskeletons require an efficient power supply for prolonged use outdoors. Another disadvantage posed by robotic exoskeletons is the extensive training required by the patient to learn how to use the equipment. Lower extremity exoskeletons designed for paraplegia rehabilitation should fulfill various requirements to be approved, such as biocompatibility and proof of electrical, mechanical, and thermal safety [5]. The interface between the frame and the wearer’s skin should be designed according to the user’s needs, from a dimensional point of view and while taking into consideration the fact that paraplegia patients are prone to skin issues [7]. 2.2 Available Models Currently, there are very few medical exoskeleton models on the market, the most remarkable of them being the ReWalk™ and Indego. Both devices have been extensively tested, and studies have been conducted to determine them safe to use. Two studies concluded that there were no adverse events for the ReWalk™ exoskeleton and minimal adverse events for Indego, with all those enrolled being able to complete their protocol [8, 9]. The devices are similar in functionality and design, and both are approved for sale in both the USA and Europe. ReWalk™ was the first exoskeleton to be FDA-approved. The major disadvantage posed by both these exoskeletons is the restrictive price tag of 80,000 USD for the Indego and 69,000–85,000 USD for ReWalk™ [10, 11].

110

F.-A. Harmon and P.-I. Br˘aileanu

2.3 Advancements in Exoskeleton Construction In comparison with exoskeletons used for military and industrial applications, which in majority are full-body models, medical exoskeletons offer a minimalistic construction, the main reason being the requirement of good balance and weight distribution. The actuators used are more efficient regarding power consumption, because of the lower effort and the battery life is much better. Such devices are destined for paraplegia sufferers, so they do not require the additional upper-body part that other types of exoskeleton feature (Fig. 1).

Fig. 1. Exoskeleton design parts

Due to the rapid advancements in technologies such as 3D-printing, the construction of medical exoskeletons became considerably easier, the structural quality greatly increased, and the price decreased. Medical exoskeletons also benefit from the advancements in the fields of mechanics and electronics, because of their use of electricalpowered motors. Power improvements in the motors and size reduction of both the batteries and motors can help decrease the overall size and weight of the exoskeleton, making it more comfortable and stable. In the last years, due to the factors mentioned above, medical exoskeletons became available to a larger number of paraplegia sufferers. However, the price of such a unit is still considerably high, the lowest being approximately 40,000$ [10]. 2.4 The Prototype The presented device is a rehabilitation exoskeleton prototype intended to be used by paraplegia sufferers. The main objective that was established when designing this prototype was to obtain a model with an attractive price–quality ratio, which is both reliable

Rapid Prototyping of a Lower-Body Exoskeleton

111

and aesthetically pleasing due to the possibility of it being customized according to patient’s needs and preferences. Most of the exoskeleton modules and parts can be 3D-printed, thus facilitating the construction. The device represents an assembly of 40 pieces, with 7 major units, 6 of which are grouped in pairs, each of them corresponding to a segment of the lower limb. The ungrouped unit represents the lower back support. Five of these modules feature interior compartments which house the electronic components. Movement of the articulations is realized by the actuators fitted in 3 joint slots: hip, knee, and ankle. Furthermore, the exoskeleton features 6 leg straps and a belt to maintain the stability of the patient during use. The equipment has a total of 8 mobile articulations (4 on each side), which are operated using a remote controller programmed to work with the Arduino uno microcontroller fitted into each major compartment. Most of the parts that make up the exoskeleton are replaceable, including the accumulators. The part assembly is mainly realized using screws, so that access to the electronic parts on the inside of the exoskeleton is facilitated to decrease maintenance costs and time (Fig. 2).

Fig. 2. Exoskeleton for paraplegia patients overall prototype

Regarding customization, the exoskeleton will be constructed taking into consideration both the physiological requirements of the patient (size) and the aesthetical preferences. The individual can choose from a variety of colors for the whole device. The covers of the major units are also customizable, offering multiple designs to select from.

112

F.-A. Harmon and P.-I. Br˘aileanu

2.5 Material Assignment Exoskeleton frames are commonly fabricated using materials such as carbon fibers, aluminum alloys, steel, or common thermoplastics. The motors used are often made of steel or iron and in some cases, components are even made from bronze. Other materials such as polyethylene or nylon-6 can be used for the belt and leg straps. Material assignment should be realized while taking into consideration production costs, weight, and factors such as mechanical, thermal and electroconductive properties. The material should be solid enough so that the motor can move the wearer’s leg segments without breaking the exoskeleton frame. The users should be able to maintain their balance with or without cane support, meaning that the added weight of both the frame and the internal components cannot be too big, and lastly, to avoid serious accidents, the material should have good thermal stability in both warm and cold environments and not be electroconductive. Considering the requirements and the objective of making this prototype 3D-printable, a suitable material that can be used is ABS plastic. Acrylonitrile butadiene styrene, commonly known as ABS is a copolymer obtained through polymerization of styrene, acrylonitrile and polybutadiene and it is mainly used as a thermoplastic. ABS is one of the most common materials used in 3D-printing applications [12]. Other applications include car components, mobile phones and smartphone accessories, TV components, radios, keyboards, control panels [13]. Some of the most important properties of ABS plastic are its high rigidity, high impact resistance even in low temperature, strain resistance, mechanical and structural stability over a prolonged time, electrical insulation, good weldability, and decent resistance to diluted acids and ionic salts [14]. The latter is especially important in extreme cases such as battery leaks when lead-acid accumulators are used. The mechanical and thermal properties of the acrylonitrile butadiene styrene polymer can be further improved depending on the conditions under which the material is processed. Table 1. Mechanical properties of ABS plastic copolymer [14, 15] Property

Value

Measurement unit [SI]

Elastic modulus

2.28

GPa

Flexural modulus

1.6–2.4 GPa

Flexural strength

77

MPa

Tensile strength(max) 43

MPa

Density

1.05

g/cm3

Poisson’s ratio

0.35

N/A

2.6 Electronics and Functionality The exoskeleton features 8 mobile joints which can help move the wearer’s lower limb segments. Each articulation contains an actuator, and the placement is as follows:

Rapid Prototyping of a Lower-Body Exoskeleton

113

For the hip joint, the Exoskeleton Drive GEN.1 was the chosen actuator. It was released by Maxon Motors in October 2017, at an accessible price and it was specifically made for rehabilitation exoskeleton construction. The motor can generate 54 Nm of continuous torque and 120 Nm on a 20% duty cycle, with a maximum actuation speed of 22 rpm, which represents more than enough for its designed use. It can be operated on a power supply between 10 and 50 V DC, thus facilitating the selection of accumulators [15]. Exoskeleton Drive GEN.1 matches the imposed requirements of this prototype, both dimensionally and functionally, the only disadvantage being that it is the most expensive component of the assembly (Fig. 3).

Fig. 3. Exoskeleton leg section and components

For the knee joint, lower power-generating actuators were used, because the knee joint actuation engages less of the lower limb segments in motion than in the case of hip joint actuation, so a higher power-generating motor is not required. Furthermore, financial, and energetic resources can also be saved. Thus, the actuator of choice in the knee joint slot is a flat BLDC motor, which matches the designed slot, because of its decent power generation: energy consumption rate. Currently, there are many flat BLDC motors available on the market, offering the possibility of constructing this exoskeleton according to the needs of various users. Actuators which can operate on a power supply of 12 and 24 V DC, together with those that can operate on variable power supplies are preferred, to reduce additional costs and to avoid internal compartment modifications that would be necessary if larger batteries were to be installed. The articulations created by the union of the lateral arms and back segment have the purpose of realizing the abduction/adduction of the lower limb. Such as in the case of hip joint actuation, a higher power-generating actuator is required. BLDC motors which can operate on 12/24 V/variable DC power supplies are suitable. For the ankle articulations, BLDC motors of reduced dimension can be used since they are not required to realize such complicated movements as in the case of the other joint actuators.

114

F.-A. Harmon and P.-I. Br˘aileanu

On the interior of the exoskeleton, there are 5 major compartments which house the power supplies (replaceable accumulators in this case), microcontrollers used to program the functionality of the actuators and the corresponding wiring. The batteries installed in the femoral compartment power the motors corresponding to the hip joint, the tibial segment batteries power the knee and ankle joint actuators and the battery fitted in the back compartment, which is also the largest of the system, powers the motors responsible for the abduction/adduction movements of the lower limb. Suitable power supplies for this configuration are 12/24 V DC, with capacities between 10 and 40 Ah, offering a use time between 6 and 8 h, depending on whether the actuators are used continuously or with breaks. The actuators do not consume energy while on standby (Fig. 4).

Fig. 4. Exoskeleton interior section and components

All the components chosen offer both a performance and a financial benefit. The presented exoskeleton prototype offers the possibility of being constantly updated, because it benefits from various technological advancements, such as the emergence of higher capacity accumulators with reduced dimension. Price is especially important to fulfill the desired goal. Taking into consideration the market value of all the components mentioned, the overall price of the exoskeleton would fall under 15,000–20,000 USD, which is less than half of the cheapest exoskeleton price available.

3 FEA Static Simulation To verify the mechanical behavior of the exoskeleton assembly of different applied forces, a static FEA analysis was performed in Inventor 2021. The forces can represent

Rapid Prototyping of a Lower-Body Exoskeleton

115

different loads which can occur to the joints of the exoskeleton, in the event of an incident. The model was simplified and only the upper joint was kept, the one in the hip area, to simplify the FEA analysis, and the connection between the components was defined as a bonded connection. The assigned global material for this static study was ABS plastic copolymer with the mechanical properties indicated in Table 1. In terms of boundary conditions, the constraints were applied to the faces of the exoskeleton back support element, and forces were applied to the side of the joint to observe its behavior in the event of an accident (because most cases of exoskeleton damage have happened in various incidents that can occur in patients’ daily life activity) (Fig. 5).

Fig. 5. Exoskeleton FEA static study boundary conditions

The stress results in case of each force applied are shown in Table 2. We can observe a linear increase, directly proportional to the applied force with a maximum stress of 398.15 MPa on Von Mises Scale, in case of a 2000 N force. Table 2. Static FEA stress results Force Minimum stress Maximum stress [N] [MPa] [MPa] 0.15·10–10

120.30

1200 0.29·10–10

238.86

1500

0.38·10–10

298.52

2000

0.52·10–10

398.15

600

The displacement results in case of each force applied are shown in Table 3. We can observe a linear increase, directly proportional to the applied force with a maximum displacement of 49.59 mm, in case of a 2000 N force. Although they may seem significant displacements at first glance, we must keep in mind that the overall material applied was

116

F.-A. Harmon and P.-I. Br˘aileanu Table 3. Static FEA displacement results Force [N]

Minimum displacement [mm]

Maximum displacement [mm]

600

0

14.89

1200

0

29.75

1500

0

37.19

2000

0

49.59

an ABS, and for an equivalent of about 200 kgf applied to the exoskeleton joint, the displacement is not very large. The strain in case of each force applied are shown in Table 4. We can observe a linear increase, directly proportional to the applied force with a maximum strain result of 0.159 Equivalent Strain, in case of a 2000 N force. Table 4. Static FEA strain results Force Minimum strain Maximum [N] [Equivalent strain] [Equivalent strain] 0.59·10–14

0.048

1200 0.11·10–13

0.095

0.15·10–13

0.119

2000 0.20·10–13

0.159

600 1500

The results of the FEA static simulation and their distribution on the exoskeleton in case of the 600 N force can be visualized in Fig. 6.

Fig. 6. Exoskeleton FEA static study graphic results. a. Stress results; b. Displacement results; c. Strain results

Rapid Prototyping of a Lower-Body Exoskeleton

117

4 Conclusions Due to the advancements in 3D-printing technologies and latest trends in additive manufacturing, medical exoskeleton fabrication becomes much easier and less expensive, offering the possibility for such devices to become a common form of prosthetic for disabled people in the near future. Medical prosthetics construction is greatly facilitated by the use of CAD programs, which offer access to precise virtual design and measurement tools and various types of simulations or environments that can detect errors and allow for easy modification, before a physical model is created. Variables such as the patient’s anatomical landmarks should always be taken into consideration when modeling a medical prosthetic, regardless of type and producers should work together with healthcare providers and humanitarian associations in order to make such equipment available to a greater number of people. In regard to the rehabilitation exoskeleton prototype that was presented, one of the main goals established for the future is the creation of a physical model, through 3D-printing, simple assembly, and programming.

References 1. Nas, K., Yazmalar, L., Sah, ¸ V., Aydın, A., Öne¸s, K.: Rehabilitation of spinal cord injuries. World J. Orthop. 6(1), 8–16 (2015). https://doi.org/10.5312/wjo.v6.i1.8 2. Farris, R.J., Quintero, H.A., Murray, S.A., Ha, K.H., Hartigan, C., Goldfarb, M.: A preliminary assessment of legged mobility provided by a lower limb exoskeleton for persons with paraplegia. IEEE Trans. Neural Syst. Rehabil. Eng. 22(3), 482–490 (2014). https://doi.org/ 10.1109/TNSRE.2013.2268320 3. Vyshka, G., et al.: Recent advances on acute paraplegia. J. Acute Dis. 5(6), 445–449 (2016) 4. Hauser, S.L., et al.: Intensive immunosuppression in progressive multiple sclerosis. A randomized, threearm study of high-dose intravenous cyclophosphamide, plasma exchange, and ACTH. N. Engl. J. Med. 308(4), 173–180 (1983) 5. Misiti, A.: Ensuring the Electrical Safety of Powered Exoskeletons | Bench Ta. Mouser.com (2018). https://www.mouser.com/blog/ensuring-the-electrical-safety-of-pow ered-exoskeletons. Accessed 5 May 2021 6. Gorgey, A.S.: Robotic exoskeletons: the current pros and cons. World J. Orthop. 9(9), 112–119 (2018). https://doi.org/10.5312/wjo.v9.i9.112 7. Chen, B., et al.: Recent developments and challenges of lower extremity exoskeletons. J. Orthop. Transl. 5, 26–37 (2016) 8. Zeilig, G., Weingarden, H., Zwecker, M., Dudkiewicz, I., Bloch, A., Esquenazi, A.: Safety and tolerance of the ReWalkTM exoskeleton suit for ambulation by people with complete spinal cord injury: a pilot study. J. Spinal Cord. Med. 35(2), 96–101 (2012). https://doi.org/ 10.1179/2045772312Y.0000000003 9. Tefertiller, C., Hays, K., Jones, J., et al.: Initial outcomes from a multicenter study utilizing the indego powered exoskeleton in spinal cord injury. Top. Spinal Cord. Inj. Rehabil. 24(1), 78–85 (2018). https://doi.org/10.1310/sci17-00014 10. Hruska, J.: A new budget exoskeleton could help paraplegics walk at a drastically lower price - ExtremeTech. ExtremeTech (2016). https://www.extremetech.com/extreme/222396-a-newbudget-exoskeleton-could-help-paraplegics-walk-at-a-drastically-lower-price. Accessed 5 May 2021 11. FDA approves Indego exoskeleton for clinical and personal use. EurekAlert! (2016). https:// www.eurekalert.org/pub_releases/2016-03/vu-fai031016.php. Accessed 5 May 2021

118

F.-A. Harmon and P.-I. Br˘aileanu

12. Zhan, J., et al.: Metal-plastic hybrid 3D printing using catalyst-loaded filament and electroless plating. Addit. Manuf. 36, 101556 (2020) 13. Plastic Properties of Acrylonitrile Butadiene Styrene (ABS): May 15, 2010, at the Wayback Machine. Accessed 4 May 2021 14. Acrylonitrile Butadiene Styrene (ABS Plastic): Uses, Properties and Structure. Omnexus.specialchem.com (2021). https://omnexus.specialchem.com/selection-guide/acr ylonitrile-butadiene-styrene-abs-plastic. Accessed 4 May 2021 15. Peters, E.N.: Plastics: Thermoplastics, Thermosets, and Elastomers. Wiley-Interscience, New York (2002) 16. Exoskeleton Joint Actuator: Maxongroup.us (2017). https://www.maxongroup.us/maxon/ view/news/Exoskeleton-joint-actuator. Accessed 6 May 2021

Methods for Testing the Strength of Layers for Different Optical Coatings Ciprian Ion Rizescu, Ionela Ghita, and Dana Rizescu(B) University POLITEHNICA of Bucharest, Splaiul Independentei 313, Bucharest, Romania [email protected]

Abstract. The first part of the paper presents the test techniques and the equipment used to make them, as well as the morphological characteristics of the tested parts and the properties of the materials that compose them. Also, there are presented briefly the results obtained after testing the resistance of several types of optical coatings are presented. The full description of the testing operations and acquiring experimental results will be described in a future paper. The authors used techniques found in the literature, but also standardized ones, comparing the mechanical behavior of frequently used materials (for example: MgF2 , TiO2 , Ta2 O5 and SiO2 ), deposited on a glass substrate (BK7). The tested parts are flat surfaces, with a diameter of 25 mm, covered with anti-reflective layers for the visible range. In this paper, various methods for testing the strength of thin films for three types of optical coatings have been studied and developed. The three types of coatings tested had the following morphological characteristics: AR (anti-reflective) single layer, AR 4 layers and AR 6 layers. From the results obtained from the tests, it was concluded that the strength of the optical coatings differs depending on the test to which the sample is subjected. The most resistant types of optical abrasion coatings were single-layer coating and four-layer coating, followed by 6-layer coating. Keywords: Optical coatings · Anti-glare coatings · Layers strength

1 Introduction Optical coatings are used for optical, microelectronic, biomedical, and decorative applications to improve elements on tribology (reduced friction), mechanics (wear / abrasion resistance), chemical composition (barrier to aggressive gases), optical, magnetic, and electrical properties [1]. With the development of optical technologies and for more comfort in wearing glasses, layered lenses have become increasingly popular. Considering the variety of these layers, we can say that they have a common feature, namely their role to replace the deficiencies of the basic lens. Deficiencies can be the following: the gloss of glass lenses, scratches on plastic lenses, shine and the fact that they get dirty more easily. Anti-glare coatings are used in a wide variety of applications where light passes through an optical surface and low reflection is desired. Examples include coatings on © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 119–128, 2022. https://doi.org/10.1007/978-3-030-83368-8_12

120

C. I. Rizescu et al.

spectacle lenses, but also on camera lenses / cameras, as well as anti-glare coatings on solar cells. The simplest form of anti-glare coating was discovered by Lord Rayleigh in 1886. The optical glass available at the time tended to develop a stain on its surface over the years due to chemical reactions with the environment. Rayleigh tested some old, slightly stained pieces of glass and found to his surprise that they emitted more light than the new, clean pieces. The stains replace the air-glass interface with two interfaces: a scratch interface and a glass interface. Because the stain formed on the glass has a refractive index between the two glass and air interfaces, each of these interfaces has a lower degree of reflection than the air-glass interface. In fact, the total of the two reflections is smaller than that of the air-glass interface, as can be calculated from the Fresnel equations [2]. The simplest optical coatings are thin metal layers, such as aluminum, which is deposited on glass supports to make mirror surfaces Fig. 1. The metal used determines the reflection characteristics of the mirror; Aluminum is the cheapest and most common coating and provides a reflection of about 88%–92% of the visible spectrum. More expensive is silver, which has a reflection of 95%–99% even in far infrared but suffers from decreased reflection ( 75

[M] 15 > 16 [M] 40

[M] 200 > 48 [M] 350

[M] 24 > 60 [M] 18

Successive mov.

Type of robot Arm Simultaneously > 43 mov. [M] 2700 Successive > 45 mov. [M] 3000

4 Conclusion The proposed ROBO-PVAFM software platform provides, compared to other similar platforms, the following research possibilities: (1) analysis of the variation of positions (P), linear and angular velocities (V), linear and angular accelerations (A) as well as of forces (F) and moments (M) in the conditions of simultaneous or successive movements; (2) PVAFM analysis in the conditions of changing the direction of the various joints slide-element, or element-slide, rotation joint-element, or element-rotation joint; (3) PVAFM analysis in the conditions of movement with or without the object to be manipulated; (4) complete analysis by modifying the various dimensions of the bodies or their materials; (5) analysis by partial or total modification of the parameters of the velocity characteristics, parameters such as acceleration time, cycle time, time from the common origin of time, or the time after which the deceleration starts, or linear or angular displacement values for various joints; (6) the platform offers a broad framework for analysis by highlighting both 2D and 3D features, as well as the variation on each axis of various kinematic and dynamic parameters; (7) the possibility to interactively calculate the inertia tensors for each body and to ensure the visualization of the way in which the

152

A. Olaru et al.

dimensions of the bodies influence the dynamics of the movement is important for the constructive optimization taking into account the optimization of the overall dynamic behavior of the robotic structure. The proposed software platform ROBO-PVAFM is general and can be applied to any robotic structure or to any complex multirobot structure with up to 5 DOF.

References 1. 2. 3. 4. 5. 6. 7. 8. 9.

https://new.abb.com/products/robotics/robotstudio http://www.workspacelt.com https://roboticsimulationservices.com Rofer, T.: SimRobot Website (2005). http://www.tzi.de/simrobot. https://www.ode.org http://www.roboanalyzer.com https://www.parallemic.org/RoKiSim.html https://www.ros.org/about-ros/ García-Martínez, J.R., Rodríguez-Reséndiz, J., Cruz-Miguel, E.: A new seven-segment profile algorithm for an open source architecture in a hybrid electronic platform. Electron. J. (2019). https://www.researchgate.net/publication/333676286

Hardening, High-Speed Steel R6M5, Using a Combined Heat Treatment Technology Umarov Tolibjon(B) Tashkent State Technical University, Tashkent 100097, Uzbekistan

Abstract. This article provides an increase in the reliability of drills while reducing the processing cycles of high-speed steel tools, namely cycles 3, 4, 5 (Fig. 1) by creating a combined processing technology that allows these cycles to be created in one cycle. Keywords: Drill · General process · Cycle · Thermal · Operation · Tool · Fabrication · Reliability · Carbide dissolution · Tempering · Temperature · Durability · Wear-resistant · Combined · Deposit coating · Microstructure · Hardness · Salt bath · Viscosity

1 Introduction The general technological process of processing a tool, including drills, is usually presented in the form of a diagram (Fig. 1), consisting of various cycles. In the diagram of the general process, mechanical, two thermal and one assembly cycles are provided, in each of which operations are concentrated, included in various private technological processes, which have a single route of movement of tools during their manufacture. As you can see (Fig. 1), the final properties of the tool, its quality, performance and reliability are determined in 3, 4, 5 processing cycles. At present, the industry has received the most widespread method according to which tools made of high-speed steel R6M5 are processed according to the technology, including quenching from temperatures of 1200–1350 °C corresponding to the maximum dissolution of carbides, three-time 1 h tempering at a temperature of 540–560 °C, final machining: grinding-sharpening, then, in order to increase the tool’s durability, apply a wear-resistant Tin coating to the surface at a temperature of 450 °C. When creating a combined technology for hardening a tool made of high-speed steel, the most optimal tempering with the process of obtaining a coating by the CIB method (cathode-ion bombardment in vacuum). In this case, it became possible to deposit the coating at higher temperatures, which is very important from the point of view of increasing the coating product in tools, the edge according to [12] it is possible to use a shortened high-temperature tempering instead of three one-hour tempering at 560 °C.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 153–161, 2022. https://doi.org/10.1007/978-3-030-83368-8_15

154

U. Tolibjon

To determine the temperature-time parameters that ensure the greatest efficiency of the combined processing of tools made of high-speed steel R6M5, it was necessary to solve two problems: 1. To develop a technology for heat treatment of steel R6M5 with high-temperature short-term tempering; 2. Combine the process of high-temperature short-term tempering with obtaining a wear-resistant titanium nitride coating in vacuum on R6M5 steel. To solve the first problem, we studied the effect of temperature during quenching, preliminary and final tempering of high-speed steel R6M5 on its microstructure, hardness, heat resistance, strength and toughness, as well as on cutting properties.

Fig. 1. General process flow diagram.

2 Materials In the study of the microstructure, hardness, and heat resistance, plates with dimensions of 5 × 20 × 20 mm (height × width × length) were used as samples. Strength, the amount of retained austenite was investigated on specimens 6 × 6 × 60 mm, impact toughness - on specimens with dimensions 10 × 10 × 55 mm. Resistance when cutting tool plates and drills. Heating of samples and tools for hardening is carried out in two stages: the first stage - heating to a temperature of 800 °S in a salt bath of NaCl, holding for 5 min per 1 mm, the second stage - heating to temperatures of 1230 °S in a salt bath BaCb, holding

Hardening, High-Speed Steel R6M5

155

0.25 min for 1 mm of the product. Then a preliminary vacation is carried out at 300 °C– 350 °C for 1 h. The final tempering was carried out at 200–400 °C, 450 °C, temperatures from 600 °C to 650 °C (holding time 5–20 min, Fig. 3). Preliminary tempering was carried out after quenching at different temperatures in order to determine the temperature at which reduced values of hardness are provided for steel R6M5. This temperature was found to be 350 °C with a specimen hardness of HRC 58.5–59. (Fig. 2,a), with hardness after hardening HRC 62–63. The austenite grain size after quenching was determined (according to GOST 5639-82) by comparing with the standards on a scale, the grain size did not exceed a point of 10. The final tempering at different temperatures and holding times showed (Fig. 2,b) that the most optimal for short-term tempering is a temperature in the range of 600–620 °C with a holding time of 15–20 min.

Fig. 2. a-influence of tempering temperature on hardness HRC of steel R6M5 after hardening.KS, MJ/cm2 ; b-influence of the final tempering temperature on the hardness of R6M5 steel (intermediate tempering 350 ºC).

3 Results and Discussions The flexural strength, as can be seen from Fig. 3, decreases monotonically with an increase in the tempering temperature at all values of the holding time. Dependences

Fig. 3. Dependence of ultimate strength al in bending of specimens from steel R6M5 on tempering temperature.

156

U. Tolibjon

of impact strength on temperature and holding time are shown in Fig. 2b. The amount of retained austenite for all selected tempering temperatures and all selected intervals ranges from 4–5%. Thus, analyzing the data obtained, it can be seen that the most acceptable modes for the combined processing of tools made of steel R5M5 are the following hardening modes from 1230 °C, preliminary tempering 350 °C, Reduced final tempering 610–620 °C for 10–15 min. The phase composition of samples from steel R6M5 with such heat treatment is as follows: special carbides KS, MJ/cm2 . Exposure time 5 min. Holding time 10 min Exposure time 15 min Holding time 20 min Dashed line - maximum allowable value

Fig. 4. Heat resistance of R6M5 steel.

1 - samples that went through the proposed heat treatment 2 - samples that went through standard heat treatment (undissolved during austenitization and released during tempering) ~ 20–25%; residual austenite ~ 4–5%; the rest is tempered martensite. Heat resistance tests were carried out on specimens prepared according to the new recommended heat treatment in comparison with specimens that underwent standard heat treatment (Fig. 4). Tests for heat resistance consisted in the fact that after the final heat treatment, tempering was carried out at 600, 615, 630.650 °C. The selected modes of new heat treatment were checked by comparative resistance tests during turning and drilling. For turning, a 1K62 screw-cutting lathe was used. Workpieces of 0150 mm and a length of 500 mm from steel 45 were processed. As a cutting tool, we used cutters with mechanical fastening of cutting plates (dimensions 12 × 12 × 5 mm) made of steel R6M5, having the following geometrical parameters: y = 10°; a = 6°; f = 75°; 1 = 0°; i = l mm. Cutting mode for turning: V = 50 m/min; S—0.1 mm/rev; t = 2 mm. Cooling - 5% solution of EGT emulsifier in water. The test results are shown in Fig. 5. The tests for each variant of heat treatment were repeated 3 times. 1. Suggested heat treatment

Hardening, High-Speed Steel R6M5

157

Fig. 5. a-diagram of specific volumetric wear P of plates made of steel R6M5, heat-treated according to two options; b-durability diagram T of plates made of high-speed steel R6M5, heat-treated according to two options.

2. Standard heat treatment Comparative resistance tests were also carried out for twist drills 010 mm with a tapered shank made of R6M5 steel, thermal bath NaCl, exposure 0.5 min per 1 mm of the product, the second stage was heating to 1230 °C in a BaCb salt bath, exposure 0.25 min per 1 mm diameter of the product. Steel quenching is carried out with cooling in oil. After quenching, preliminary tempering is carried out at a temperature of 350 °C for 1 h. Then carry out grinding and sharpening of the tool. The final tempering is carried out in vacuum, combining it with the writing of a wear-resistant coating with titanium nitride, the instrument, previously cleaned by ultrasound, is placed in the vacuum chamber of the spraying unit. Upon reaching a vacuum of 2–10 4Pa, a voltage of −900 V is applied, and an arc discharge is ignited at the cathode with simultaneous heating quenching 1200–1230 °C (Figs. 6 and 7).

Fig. 6. Scheme of heat treatment and additional hardening of a tool made of high-speed steel R6M5 of the tool surface to a final tempering temperature of 610–620 °C for 10–15 min, as a result of which the final cleaning of the tool surface by bombardment with titanium ions is performed.

158

U. Tolibjon

Fig. 7. Scheme of combined processing of tools made of high-speed steel R6M5.

When the surface temperature reaches 620 °C (the temperature is controlled by an infrared pyrometer with an accuracy of ±10 °C), the voltage is reduced to 200 V and the coating is deposited with titanium nitride by the method of electric arc evaporation of titanium in a nitrogen atmosphere and condensation of titanium nitride vapors on the tool surface for 15 min.

Fig. 8. Durability diagram of 10 mm diameter twist drills made of P6M5 steel when cutting 45 steel. a) a drill that has undergone standard heat treatment; b) a drill that has undergone combined heat treatment.

Hardening, High-Speed Steel R6M5

159

To determine the effectiveness of the developed combined hardening, comparative resistance tests were carried out. Drills that have passed the proposed and standard processing. Drills 010 mm were tested when drilling blind holes in steel 45 and steel 12X18H10T to a depth of 3d, on a 2R135F machine. For cooling, a solution of ETT-2 emulsol was used (flow rate 5–7 l/min, watering). Processing mode: n = 500 rpm, S = 0.2 mm/rev. The wear criterion is 0.5 mm flank wear. Tests have established that when processing steel 45, the durability of drills processed using the combined technology increases by 4.8 times compared to standard heat treatment, and when processing steel 12X18H10T by 1.9 times (Fig. 8, 9) (Fig. 10).

Fig. 9. Diagram of the resistance of spiral high-speed drills with a diameter of 10 mm from steel R6M5 when cutting steel 12X18H10T. a) drill, passed standard heat treatment; b) a drill that has undergone combined heat treatment.

160

U. Tolibjon

Fig. 10. Dependence of the durability of spiral high-speed drills made of steel R6M5 from wear on the flank surface when cutting steels.a-standard heat treated drill b-drill, which has passed the combined heat treatment.accordingly, the intensity of wear

4 Conclusions 1. The most promising and effective methods of hardening spiral high-speed drills are methods of hardening the coating by the CIB method. 2. The developed combined method of processing a tool made of high-speed steel makes it possible to combine the process of obtaining a coating by the CIB method with the final tempering of the tool. 3. The temperature of the final tempering (610–620 °C) is set for the combined processing of the tool. 4. The combined technology makes it possible to reduce the processing time of the tool by 3–4 h due to the combination of the processes of obtaining a coating and final tempering. 5. The combined technology provides the required level of ultimate strength in bending, impact toughness, hardness, while increasing the heat resistance of HSS tools. 6. The optimal modes of deposition of the TiN coating have been established for the combined treatment of the tool. (j = l 50 A/m; u = 200 V; p = 6–10 Pa) 7. The combined technology allows to carry out grinding - sharpening of the tool in more favorable conditions (lower values of hardness HRC 58–58.5 in comparison with the standard technology HRC 60–61). 8. The combined technology increases the durability of twist drills made of high-speed steel when processing structural carbon steels by 4.8 times, and in the processing of structural high-alloyed by 1.9 times.

References 1. Umarov, T., Bekturov, D.: Drilling in heat treated steel. Asian J. Multidimen. Res. (AJMR) 10(3), 425–430 (2021) 2. Umarov, T.U., Mardonov, U.T., Khasanov, O.A., Ozodova, S.O., Yusupov, B.D.: Research of the variation of firmness of pointed drills by method of simulation modeling of process of wear. Int. J. Psychosoc. Rehabil. 24(04), 1885–1901 (2020)

Hardening, High-Speed Steel R6M5

161

3. Umarov, E.O., Mardonov U.T., Shoazimova, U.K.: Influence of the magnetic field on the viscosity coefficient of Lubricoolant that is used in the cutting process. Int. J. Mechatron. Appl. Mech. 8(2), 144–149 (2020). https://doi.org/10.17683/ijomam/issue8.50 4. Vinogradov, A.A.: Physical bases of the drilling process of hard-to-process materials with carbide drills. Kiev Nauka dumka 264 (1985) 5. Lawal, S.A., Choudhury, I.A., Nukman, Y.: A critical assessment of lubrication techniques in machining processes: a case for minimum quantity lubrication using vegetable oil-based lubricant. J. Clean. Prod. 41, 210–221 (2013) 6. Turakhodjaev, N., Tursunbaev, S., Tashbulatov, S., Kuchkorova, M.: Analysis of technological solutions for reducing the copper concentration in slags from oxygen-flare smelting of copper sulfide concentrates. J. Crit. Rev. 7(5), 449–452 (2020) 7. Kurjmóto, T., Barrow G.: The wear of high speed steel cutting tools under the action of several different cutting fluids. In: Proceeding of the Twenty-Second International Machine Tool Design Research Conference, Manchester, UK, pp. 237–246 (1981) 8. Turakhodjaev, N., Tashbulatov, Sh., Zokirov, R., Tursunbaev, S., Baydullaev, A.: Studying the scientific and technological bases for the processing of dumping copper and aluminum slags. J. Crit. Rev. 7(11), 441–444 (2020). https://doi.org/10.31838/jcr.07.11.79 9. Uljayev, E., Ubaydullayev, U.M., Tadjitdinov, G.T., Narzullayev, S.: Development of criteria for synthesis of the optimal structure of monitoring and control systems. In: Aliev, R.A., Yusupbekov, N.R., Kacprzyk, J., Pedrycz, W., Sadikoglu, F.M. (eds.) WCIS 2020. AISC, vol. 1323, pp. 559–563. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-68004-6_73 10. Cai, R., Yang, H., He, J., et al.: The effects of magnetic fields on water molecular hydrogen bonds. J. Mol. Struct. 938(1–3), 9–15 (2009). https://doi.org/10.1016/j.molstruc.2009.08.037 11. Toledo, E.J.L., Ramalho, T.C., Magriotis, Z.M.: Influence of magnetic field on physical– chemical properties of the liquid water: insights from experimental and theoretical models. J. Mol. Struct. 888(1–3), 409–415 (2008). https://doi.org/10.1016/j.molstruc.2008.01.010 12. Umarov, E.O., Mardonov, U.T.: Analysing the effect of magnetic field on Lubricoolant in machining process. Int. J. Appl. Res. 6(5), 347–352 (2020)

Increasing the Abrasive Wear Resistance of Steels by Heat Treatment with Preliminary Preparation of the Structure Darob Berdiev(B) , Botir Saydumarov, and Nargiza Makhmudova Tashkent State Technical University, Tashkent 100097, Uzbekistan [email protected]

Abstract. The abrasive wear resistance of samples made of steels St3Gsl, 35, 45, 65G, U8 is investigated. Samples of technical iron are used as a reference material. To increase the abrasive wear resistance of carbon and low-alloy steels, it is proposed to carry out preliminary normalization with an extreme heating temperature of 1100 °C before the final heat treatment. After re-hardening and low tempering, the wear resistance can increase from 20 to 50%. Keywords: Abrasive wear · Heat treatment · Wear resistance

1 Introduction Tillage machines operate under abrasive soil and dust conditions. In the composition of soil and dust, the abrasive effect is mainly caused by quartz particles. Abrasive wear is the most destructive and therefore a large amount of metal is consumed annually for the manufacture of spare parts. For the manufacture of such parts, medium and high-carbon steels 40G, 45, 65G, 70G, L53, L65, St6 are used [1]. In order to increase the wear resistance, attempts were made to use chromium steels such as XF, XG, SHX15, X6F1, X12F1, but did not get the expected results, because the cost of the product itself did not pay off [2]. In the opinion of the author of work [1], there is no sufficient justification for the correct choice of material for parts of tillage machines. The choice of steels and hardening methods is mainly based on many years of experience in the production and operation of tillage machines. This disadvantage is aggravated in repair production due to the lack or absence of the necessary steel grades. Any replacement of steel or hardening method must be justified by sufficiently long tests. Otherwise, the results of such replacements will be unpredictable in terms of durability. In this work, an attempt is made to solve two problems simultaneously: 1. To increase the wear resistance of steels for the working bodies of tillage machines by heat treatment with preliminary preparation of the structure for extreme conditions; © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 D. D. Cioboatˇa (Ed.): ICoRSE 2021, LNNS 305, pp. 162–170, 2022. https://doi.org/10.1007/978-3-030-83368-8_16

Increasing the Abrasive Wear Resistance of Steels by Heat Treatment

163

2. Establish the empirical dependences of the wear resistance of steels with sliding friction on a loose abrasive material on their structural parameters. On the basis of the obtained dependencies, develop a computer program for the selection of materials and methods of hardening to achieve the required level of wear resistance, that is, to replace the steel grade, modes of its thermal hardening already with predictable results in terms of wear resistance.

2 Materials and Methods of the Research The objects of research were samples from steels St3Gsl, 35, 45, 65G, U8. Technical iron was used as a reference material. It is known from previous studies [3] that pre-heat treatment can significantly affect the properties of steel after final heat treatment. Therefore, the heat treatment of the samples was carried out in the following modes: – preliminary heat treatment consisted in carrying out normalization from different heating temperatures, starting from the one usually accepted for each steel grade (Ac3 + 30 ÷ 50 °C or Ac1 + 30 ÷ 50 °C), up to 1000, 1100, 1150 and 1200 °C; – the final heat treatment consisted in reheating to temperatures usually accepted for quenching each steel grade, quenching cooling to a martensitic structure (steels 35, 45, U8 - in water, 65G - in oil, St3Gsl in 10% NaCl solution in water) and subsequent vacation 200, 350, 450 and 600 °C. The final heat treatment provided approximately the same grain size of austenite (within each steel grade), the minimum amount of retained austenite, but a different level of defectiveness of the crystal structure - phase. In order to assess the effect on the wear resistance of the amount of the pearlite component of the structure at a minimum level of defectiveness of the crystal structure - phase, some of the samples were subjected to complete annealing according to the standard mode. Wear tests were carried out with sliding friction on a loose abrasive material on a PV-7 installation in accordance with the procedure [4, 5]. The abrasive material was pulverized quartz sand, which was fed in portions using a dispenser onto the rubbing surface of the test specimen and a polyurethane screw. The choice of installation and method of testing for abrasive wear is explained by the fact that our previous studies [6] found the existence of similarity of wear resistance rows when tested on a PV - 7 machine and field tests of a cotton seeder opener compactor in the fields of a number of regions of Uzbekistan. There was a coincidence of the order of the materials and the quantitative values of the relative wear resistance obtained during wear under laboratory and field tests. The parameters of the structure were estimated by quantitative metallography (grain size, amount of structural components) electron microscopy of carbon replicas (size and number of cementite particles in tempered steel), X-ray diffraction analysis (calculation of dislocation density from the physical width of the X-ray line) according to known techniques [7–9].

164

D. Berdiev et al.

3 Results The steels in the annealed state showed a linear dependence of the increase in wear on the amount of the pearlite component in the structure. A decrease in the amount of wear from the amount of the pearlite component in the structure is described by the ratio QP = QOFe , where QOFe - is the amount of wear of annealed technical iron - 7.3 mg [6]. After normalization, the steels in the structure had some deviation from the equilibrium state, which consists in an increase in the fraction of the pearlite (quasi-keutectoid) component with an increase in the normalization temperature and an increase in the dislocation density. At the same time, a sharp increase in the austenite grain was observed upon heating under normalization to temperatures above 1000 °C. However, the amount of wear did not depend on either the grain size or the amount of the pearlite component (within each grade) and did not even depend on a certain increase in hardness, but was in accordance with the level of dislocation density (Fig. 1). As seen from Fig. 2, an extremum of the dislocation density is observed at a normalization temperature of 1100 °C, in accordance with which the amount of wear changes. The formation of an extremum is explained by the fact that when heated for normalization (or quenching) in the temperature region of 1100–1150 °C, refractory impurity phases dissolve in austenite. The onset of dissolution of these phases is characterized by the chemical micro-heterogeneity of the solid solution. Upon cooling, in the course of γ - α transformation, a structure with an increased level of dislocation density is formed. A further increase in the heating temperature leads to homogenization of austenite, and upon cooling in the α-phase, the dislocation density is lower [10]. Accordingly, minimal wear is observed after normalizing steels from a heating temperature of 1100 °C. The decrease in the amount of wear of steels after normalization from an extreme temperature in comparison with normalization from the usually accepted temperatures turned out to be quite large (Table 1). The steels after quenching and low tempering (final heat treatment - heating Ac3 (or Ac1 ) + 30 ÷ 50 °C, quenching, tempering 200 °C) after preliminary normalization from different heating temperatures inherited the elements of the original sub microstructure, because the position of the extremum remained, although the overall level of the dislocation density increased sharply. Table 1. Reduced wear of steel during normalization from extreme heating temperatures. Steel grade

Technical iron

Steel, St3Gsl

Steel 35

Steel 45

Steel U8

Reduced wear, %

2

11

18

31

36

Similarly, the amount of wear decreased sharply, but the minimum wear is observed at extreme temperatures (Fig. 2). At the same time, the decrease in the amount of wear, in comparison with the state of preliminary normalization from the usually accepted heating temperatures, turned out to be quite large (Table 2). For high-carbon steel U8, this reduction reaches 50%, that is, the wear is halved. Taking into account the existence of a similarity in wear resistance in laboratory and field tests, it can be argued that the presented mode of preliminary preparation of the structure can, without significant additional costs, almost double the wear resistance of parts of the working bodies of soil cultivation machines.

Increasing the Abrasive Wear Resistance of Steels by Heat Treatment

165

Table 2. Reduction of steel wear after preliminary normalization from 1100 °C, repeated phase recrystallization of hardening and low tempering. Steel grade

Technical iron

Reduced wear, % 15

Steel St3Gsl

Steel 35

Steel 45

Steel U8

Steel 65G

19

23

30

50

40

The results of experiments on quenching steel St3Gsl without tempering are of considerable interest. As can be seen from Fig. 2, direct quenching from 1100 °C of St3Gsl steel without tempering gives the same results as the preliminary preparation of the structure and re-quenching from the usually accepted temperature (950 °C). The use of quenching without tempering can significantly increase the wear resistance [11– 13]. This is due to the strain hardening of the rubbing surface of untempered steel. Steel St3Gsl even in the quenched state without tempering has a sufficient margin of toughness, which excludes brittle fracture, and in terms of wear resistance, in this state, it corresponds to low tempered steels with a carbon content of 0.35–0.40% (see Fig. 2).

Fig. 1. Influence of heating temperature during normalization on structure parameters and properties of carbon steels. 1 – technical iron, 2 – steel St3Gsl, 3 – steel 35, 4 – steel 45, 5 – steel U8

Fig. 2. Influence of the temperature of preliminary normalization on the density of dislocations and the amount of wear of steels: 1 – steel St3Gsl, 2 - steel 35, 3 – steel 45, 4 – steel U8, 5 – steel 65G

166

D. Berdiev et al.

The steels after quenching to tempering at 350, 450 and 600 °C after preliminary normalization from different heating temperatures had structures of a ferrite-cementite mixture of varying degrees of dispersion depending on the tempering temperature: but retained the position of the extremum of the dislocation density in the - phase and the minimum of wear.

4 Quantitative Relationships Between Wear Value and Steel Structure Parameters Since the structure parameters and wear values of the studied steels were already determined experimentally, an attempt was made to quantify the contribution of each parameter to the decrease in the wear value. In the low-fluff investigated steels, two parameters varied in wide redistributions: the amount of carbon in the martensite composition (depending on the steel grade) and the level of dislocation density (also depending on the steel grade and the preliminary normalization temperature). In steels after quenching and tempering at 350 °C and above, the number and size of cementite particles and the level of dislocation density change within a fairly wide range. If we plot the graphs of the dependence of the wear value Q on the square root of the dislocation density, that is, Q = f, then we will have Fig. 3 and 4. These are equations of straight lines, the steps between which in the first case (low-tempered steels) are determined by the amount of carbon in martensite (solid solution hardening). In the second case (tempering 350 °C above) there is no carbon in the solid solution (or almost no), the steps between the straight lines are determined by the size and amount of cementite particles in the ferrite matrix (hardening by the second phase). In low-tempered steel, the amount of wear will depend on the parameters of the structure according to the type. √ Q = QOSt3 − α ρ − a(C)b ; where QOSt3c is the amount of wear of steel with structural tempered martensite at a minimum (0.28%) carbon content in steel - 2.2 mg. α, a and b – corresponding ratios and exponent. C - the difference between the carbon content in the studied steel and steel St3Gsl. √  ρ - the difference between the square root of the dislocation density of the steel under study and the St3Gsl steel, which have structures of tempered martensite. In numerical terms, this ratio will be written: √ Q = 2.2 − 0.4 ρ − Q − 0.82(C)0.48

(1)

In steels tempered at 350 °C and above, the wear value can be represented √ Q = QOFe − α ρ − Kλ−1 − ξ %C − η%M

(2)

where QOFe is the amount of wear of the annealed technical iron, equal to 7.5 mg;

Increasing the Abrasive Wear Resistance of Steels by Heat Treatment

167

ρ - dislocation density; λ - average distance between cementite particles in tempered steel. Since in alloy steel during tempering, part of the carbon and alloying element can be in solid solution, the values of %C and %M show their presence in this solution. The coefficients ξ and η were constant, ξ = 6.2; η = 0.046 for manganese steel. Coefficients α and K were variable and depended on the tempering temperature: – at 350 °C α = 0.70; K = 0.35. – at 450 °C α = 0.86; K = 1.0 – at 600 °C α = 1.0; K = 3.8 According to the experimental data, the intervals for varying the parameters of the structure and their possible combinations in heat treated steels were established, which, together with relations (1) and (2), were entered into the computer program. The algorithm and the program of calculations for the selection of the necessary materials and hardening technologies are designed for cases when only one value is set - relative wear resistance:

Fig. 3. Dependence of the amount of wear of low-tempered steels on the density of dislocations: 1 – steel St3Gsl; 2 – steel 35; 3 – St3Gsl no vacation; 4 – steel 45; 5 – steel U8; 6 – steel 65G

E=

QOFe , Q

where QOFe is the wear value of the standard - annealed technical iron; Q – wear value obtained from relations (1) and (2). Since the intercementite distance in tempered above 300 °C in carbon or low alloy steel depends on both the carbon content and the tempering temperature, the tempering

168

D. Berdiev et al.

Fig. 4. Dependence of the amount of wear of steels on the density of dislocations, steels were hardened and highly tempered: 1 – armco - iron; 2 – steel 35; 3 – steel 45; 4 – steel U8.

temperature was found according to the diffusion equations [3].   2 dav q , T = − × ln 2 R A × 0.05C 3 where q – activation energy of diffusion of carbon in ferrite; λ × 21 [11], d av - average diameter of cementite particles, which can be found dav = π/6f if the intercementite distance λ and the volume fraction of cementite f are known; R – gas constant; A – constant, practically independent of temperature.

5 Discussions When choosing materials for the production of parts (as well as when replacing steel under conditions of repair production), it is advisable to focus on the already well-proven

Increasing the Abrasive Wear Resistance of Steels by Heat Treatment

169

units. In particular, studies of the material of the German-made plow blade showed that it was made of 40G steel. On the cutting edge, the blade is hardened to a hardness of HRC 48..50. Tests of the sample cut from the hardened part of the blade on the PV7 installation showed that the relative wear resistance is E = 4.68. Calculations have shown that this level of relative wear resistance can be achieved when quenching steel 45 with low tempering, steel St3Gsl when quenching without tempering, steel 35 when quenching under extreme conditions, steel U8 after quenching under extreme conditions and tempering at 450 and 600 °C. Thus, the list of steels and options for heat hardening can be very wide for a given level of wear resistance. As indicated in [2], there is no need to use expensive wearresistant materials when it is more expedient to comply with the required geometry of the working body of the tillage machine. When assessing wear resistance, it is also impractical to orientate only by hardness, because at one value of hardness, a different combination of structural parameters can give a difference in wear resistance almost twofold (Fig. 4).

6 Conclusions 1. Pre-heat treatment with heating to extreme temperatures after cooling (normalization, quenching) creates a structure with the maximum level of dislocation density. At the same time, when sliding friction against loose abrasive particles of quartz sand, wear resistance increases depending on the carbon content in steel from 11 to 36%. 2. The repeated phase recrystallization of steel with heating to temperatures usually accepted for each steel takes place under conditions of inheritance of submicrostructure elements. Therefore, after re-quenching, the position of the extremum (preliminary normalization at 1100 °C) of the dislocation density is retained. 3. Preliminary preparation of the structure by normalization at a heating temperature of 1100 °C after re-quenching and tempering can increase the wear resistance of medium and high-carbon steels by 25–35%. 4. An algorithm and a program for calculating the choice of materials, modes of their thermal hardening to ensure the required level of relative wear resistance have been created. This achieves the necessary validity when choosing or replacing materials and thermal hardening of the working bodies of tillage machines.

References 1. Tkachev, V.N.: Iznos i povishenie dolgovechnosti detaley selskoxozyaystvennix mashin (Wear and durability of agricultural machinery parts). Mashinostroenie 264 (1971) 2. Brenshteyn, D.B.: Abrazivnoe iznashivanie lemeshnogo lezviya i rabotosposobnost pluga (Abrasive wear of the ploughshare blade and plow performance). Traktori i selskoxozyaystvennie mashina 6, 39–42 (2002) 3. Berdiev, D., Yusupov, A., Umarova, M., Ibodullaev, T.: Improving wear resistance of steel products to unconventional heat treatment methods. Int. J. Sci. Technol. Res. 9(2), 2504–2509 (2020)

170

D. Berdiev et al.

4. Tenenbaum, M.M.: Soprotivlenie abrazivnomu iznashivaniyu (Abrasion resistance). Mashinostroenie 267 (1976) 5. Garkunov, D.N.: Tribotexnika (Tribotechnics). MSXA, Russia (2001) 6. Berdiev, D.M., Toshmatov, R.K.: Otlichitelnie osobennosti fazovix i strukturnix prevraweniy pri netraditsionnix rejimax termicheskoy obrabotki konstruksionnix stali (Distinctive features of phase and structural transformations under unconventional modes of heat treatment of structural steel). Vestnik Mashinostroeniya 10, 63–65 (2020) 7. Gorelik, S.S., Skakov, Yu.A., Rastorguev, L.N.: Rentgenograficheskiy i elektronnoopticheskiy analiz (X-ray and electro-optical analysis). MSXA, Russia (1994) 8. Umanskiy, Ya.S., Skakov, Yu.A., Ivanov, A.N., Rastvorguev, L.N.: Kristallografiya, rentgenografiya i elektronnaya mikroskopiya (Crystallography, X-ray and Electron Microscopy). Metallurgiya 632 (1982) 9. Xramov, A.S., Nazipov, R.A., Lukyanov, I.V.: Rentgenostrukturniy analiz polikristallov (Xray structural analysis of polycrystals). Uchebnoe posobiya, chapters 1,4, 5, KGU, Kazan (2009) 10. Mukhamedov, A.A.: The influence of the thermal history on the structure and properties of steel. Phys. Metals Metallogr. 74(5), 482–487 (1992) 11. Korshunov, L.G., Makarov, A.V., Osinseva, A.L.: Issledovanie iznosostoykosti i strukturni prevraweniy pri abrazivnom iznashivanii stali U8, uprochnennoy lazerom (Investigation of wear resistance and structural transformations during abrasive wear of U8 steel, hardened by a laser). Trenie Iznos 9(1), 52–59 (1998) 12. Goldshteyn, M.I., Grachev, S.V., Veksler, Yu.G. Spesialnie stali (Special steels).: Metallurgiya 484 (1985) 13. Berdiev, D.M., Yusupov, A.A.: Improving the wear resistance of gear teeth by cyclic quenching with inductive heating. Russ. Eng. Res. 40(6), 473–475 (2020). https://doi.org/10.3103/S10 68798X20060064

Biodegradable Starch-Based Polyvinyl Alcohol Films with Zinc-Oxide Particles for Wound Dressing Applications Mohammad Mohsen Delavari(B) and Ion Stiharu Concordia University, Montreal, QC 3G 1M9, Canada [email protected]

Abstract. Starch-based composites as wound dressings feature biodegradability, antibacterial activity, and healing properties around wound areas. Available wound dressings have some significant drawbacks, such as water vapor transmission rate (WVTR) and low absorption of wound exudates. In this work, starchbased polyvinyl alcohol (PVA) blend with embedded zinc oxide (ZnO) particles (