Proceedings of the International Conference on Innovations for Sustainable and Responsible Mining: ISRM 2020 - Volume 1 [1st ed.] 9783030608385, 9783030608392

Citation preview

Lecture Notes in Civil Engineering

Xuan-Nam Bui Changwoo Lee Carsten Drebenstedt   Editors

Proceedings of the International Conference on Innovations for Sustainable and Responsible Mining ISRM 2020 - Volume 1

Lecture Notes in Civil Engineering Volume 109

Series Editors Marco di Prisco, Politecnico di Milano, Milano, Italy Sheng-Hong Chen, School of Water Resources and Hydropower Engineering, Wuhan University, Wuhan, China Ioannis Vayas, Institute of Steel Structures, National Technical University of Athens, Athens, Greece Sanjay Kumar Shukla, School of Engineering, Edith Cowan University, Joondalup, WA, Australia Anuj Sharma, Iowa State University, Ames, IA, USA Nagesh Kumar, Department of Civil Engineering, Indian Institute of Science Bangalore, Bengaluru, Karnataka, India Chien Ming Wang, School of Civil Engineering, The University of Queensland, Brisbane, QLD, Australia

Lecture Notes in Civil Engineering (LNCE) publishes the latest developments in Civil Engineering - quickly, informally and in top quality. Though original research reported in proceedings and post-proceedings represents the core of LNCE, edited volumes of exceptionally high quality and interest may also be considered for publication. Volumes published in LNCE embrace all aspects and subfields of, as well as new challenges in, Civil Engineering. Topics in the series include: • • • • • • • • • • • • • • •

Construction and Structural Mechanics Building Materials Concrete, Steel and Timber Structures Geotechnical Engineering Earthquake Engineering Coastal Engineering Ocean and Offshore Engineering; Ships and Floating Structures Hydraulics, Hydrology and Water Resources Engineering Environmental Engineering and Sustainability Structural Health and Monitoring Surveying and Geographical Information Systems Indoor Environments Transportation and Traffic Risk Analysis Safety and Security

To submit a proposal or request further information, please contact the appropriate Springer Editor: – Mr. Pierpaolo Riva at [email protected] (Europe and Americas); – Ms. Swati Meherishi at [email protected] (Asia - except China, and Australia, New Zealand); – Dr. Mengchu Huang at [email protected] (China). All books in the series now indexed by Scopus and EI Compendex database!

More information about this series at http://www.springer.com/series/15087

Xuan-Nam Bui Changwoo Lee Carsten Drebenstedt •

•

Editors

Proceedings of the International Conference on Innovations for Sustainable and Responsible Mining ISRM 2020 - Volume 1

123

Editors Xuan-Nam Bui Hanoi University of Mining and Geology Hanoi, Vietnam Carsten Drebenstedt Institute of Mining and Special Engineering TU Bergakademie Freiberg Freiberg, Germany

Changwoo Lee Department of Energy and Mineral Resources Engineering Dong-A University Busan, Korea (Republic of)

ISSN 2366-2557 ISSN 2366-2565 (electronic) Lecture Notes in Civil Engineering ISBN 978-3-030-60838-5 ISBN 978-3-030-60839-2 (eBook) https://doi.org/10.1007/978-3-030-60839-2 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

We would like to welcome you to the International Conference on Innovations for Sustainable and Responsible Mining-ISRM 2020, which will be held during October 15–17, 2020, at Hanoi University of Mining and Geology (HUMG), Hanoi, Vietnam. ISRM 2020 is organized by HUMG to celebrate the 55th anniversary of the Department of Surface Mining, Faculty of Mining, HUMG. The conference is the effective cooperation of HUMG and both TU Bergakademie Freiberg, Germany, and Hanoi University of Public Health, Vietnam. Especially, the event is financially supported by Vietnam National Coal-Mineral Industries Holding Corporation Limited (VINACOMIN), Dong Bac Corporation (NECO), and other organizations. The main aim of the ISRM 2020 is to provide a platform for researchers, academicians, and engineers in the field of mining, earth resources, civil engineering, and geospatial technologies to present their recent research results. Besides, the conference provides a setting for them to exchange new ideas, innovative thinking, and application experiences face to face, to establish research or business relations, and to find partners for future collaboration. The conference program was organized into four sessions covering topics of Sustainable Development and Responsible Mining, Earth Sciences and Geospatial Technologies, Occupational Safety & Health, and Industry 4.0 in Mining. The ISRM 2020 has received 344 submissions, and among them, 80 high-quality manuscripts were recommended to submit for the section Sustainable Development and Responsible Mining of this Springer proceedings book for a double-blind peer reviewing. Herein, each manuscript has been reviewed for its merit and novelty by at least two reviewers by matching the content areas. As a result, a total of 29 papers have been finally selected for this book. We believe that this proceedings book provides a broad overview of recent advances in above-mentioned topics for readers. Finally, we would like to express our sincere thanks to the university council, the rector and vice-rectors, and the International Office of HUMG for their help in administrative works and other supports. Special thanks to Dr. Nguyen Quoc Long and Dr. Nguyen Hoang, HUMG, the secretary of the ISRM 2020, and Pierpaolo v

vi

Preface

Riva at Springer for help and always responding promptly. We would like to thank all the reviewers for their timely and rigorous reviews of the papers and thank all the authors for their submissions. October 2020

Xuan-Nam Bui Changwoo Lee Carsten Drebenstedt

Organization

List of Reviewers Bui Van Duc Carsten Drebenstedt Chairoj Rattanakawin Changwoo Lee Danial Jahed Armaghani Do Ngoc Anh Hesam Dehghani Hoang Anh Le Hoang-Bac Bui Hoang Huu Duong Hoang Nguyen Jian Zhou Le Tien Dung Le Trung Tuyen Mahdi Hasanipanah Mahdi Shariati Mai Ngoc Luan Manoj Khandelwal Marek Borowski Mohammadreza Koopialipoor Mohsen Hajihassani

Hanoi University of Mining and Geology, Vietnam TU Bergakademie Freiberg, Germany Chiangmai University, Thailand Dong-A University, South Korea University of Malaya, Malaysia Hanoi University of Mining and Geology, Vietnam Hamedan University of Technology, Iran Vietnam National University, Vietnam, Hanoi University of Mining and Geology, Vietnam TU Bergakademie Freiberg, Germany Hanoi University of Mining and Geology, Vietnam Central South University, China Hanoi University of Mining and Geology, Vietnam Institute of Mining Science and Technology, Vietnam University of Tehran, Iran University of Malaya, Malaysia Curtin University of Technology, Australia Federation University Australia, Australia AGH University of Science and Technology, Poland Amirkabir University of Technology, Iran Urmia University, Iran

vii

viii

Nga Nguyen Nguyen Hai An Nguyen Ngoc Bich Nguyen Ngoc Phu Nguyen Quoc Long Nguyen Thi Hong Nguyen-Trang Thao Nguyen Trong Dung Nguyen Van Duc Nguyen Van Thinh Nguyen Viet Nghia Panagiotis Asteris Pham Duc Tho Phung Quoc Huy Pirat Jaroonpattanapong Prashanth Ragam Ramesh Bhatawdekar Romulus-Dumitru Costache Ropesh Goyal Toshifumi Igarashi Tran Thi Thu Huong Trinh Le Hung Trong Vu Xuan-Nam Bui Yosoon Choi

Organization

Hanoi University of Mining and Geology, Vietnam PetroVietnam Exploration Production Corporation (PVEP), Vietnam Hanoi University of Public Health, Vietnam Hanoi University of Mining and Geology, Vietnam Hanoi University of Mining and Geology, Vietnam Can Tho University, Vietnam Ton Duc Thang University, Vietnam Hanoi University of Mining and Geology, Vietnam Dong-A University, South Korea Hanoi University of Mining and Geology, Vietnam Hanoi University of Mining and Geology, Vietnam School of Pedagogical & Technological Education, Greece Hanoi University of Mining and Geology, Vietnam Institute of Mining Science and Technology, Vietnam Chiangmai University, Thailand KL University Hyderabad, India University Teknologi Malaysia, Malaysia Research Institute of the University of Bucharest, Romania Indian Institute of Technology Kanpur, India Hokkaido University, Japan Hanoi University of Mining and Geology, Vietnam Le Quy Don Technical University, Vietnam TU Bergakademie Freiberg, Germany Hanoi University of Mining and Geology, Vietnam Pukyong National University, South Korea

Contents

Assessment of Global Digital Height Models over Quang Ninh Province, Vietnam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nguyen Quoc Long, Ropesh Goyal, Luyen K. Bui, and Xuan-Nam Bui Rock Mass Classification for the Assessment of Blastability in Tropically Weathered Limestones . . . . . . . . . . . . . . . . . . . . . . . . . . . Ramesh Murlidhar Bhatawdekar, Edy Tonnizam Mohamad, Trilok Nath Singh, Pranjal Pathak, and Danial Jahed Armaghani Study on the Reasonable Parameters of the Concentric Hemisphere-Style Shaped Charge for Destroying Rock . . . . . . . . . . . . . Trong Thang Dam, Xuan-Nam Bui, Tri Ta Nguyen, and Duc Tho To Effect of Carbon Nanotubes on the Chloride Penetration in Ultra-High-Performance Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . Pham Manh Hao, Nguyen Van Tuan, Nguyen Cong Thang, Nguyen Van Thao, Luong Nhu Hai, Pham Sy Dong, Nguyen Xuan Man, and Ngo Ngoc Thuy Estimating the Radial Displacement on the Tunnel Boundary Within Efficient Working Area of Rock Tunneling Quality Index (Q-System) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Van Diep Dinh, Ngoc Anh Do, Amund Bruland, and Daniel Dias Two Novel Predictive Networks for Slope Stability Analysis Using a Combination of Genetic Programming and Artificial Neural Network Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hossein Moayedi

1

13

45

69

81

91

A Review of Artificial Intelligence Applications in Mining and Geological Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Xuan-Nam Bui, Hoang-Bac Bui, and Hoang Nguyen

ix

x

Contents

River Sand Mining Vis a Vis Manufactured Sand for Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Ramesh Murlidhar Bhatawdekar, Trilok Nath Singh, Edy Tonnizam Mohamad, Danial Jahed Armaghani, and Dayang Zulaika Binti Abang Hasbollah Evaluating the Effect of Meteorological Conditions on Blast-Induced Air Over-Pressure in Open Pit Coal Mines . . . . . . . . 170 Quang-Hieu Tran, Hoang Nguyen, Xuan-Nam Bui, Carsten Drebenstedt, Belin Vladimir Arnoldovich, Victor Atrushkevich, and Van-Duc Nguyen Development of a Blasting Vibration Monitoring System Based on Tri-axial Acceleration Sensor for Wireless Mesh Network Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Won-Ho Heo, Jung-hun Kim, Van-Duc Nguyen, Quang-Hieu Tran, Hoang Nguyen, Xuan-Nam Bui, and Chang-Woo Lee Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm to Predict the Dust Concentration in Deo Nai Open-Pit Coal Mine (Vietnam) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 Xuan-Nam Bui, Hoang Nguyen, Carsten Drebenstedt, Hai-Van Thi Tran, Ngoc-Bich Nguyen, Xuan-Cuong Cao, and Qui-Thao Le Evaluating the Air Flow and Gas Dispersion Behavior in a Deep Open-Pit Mine Based on Monitoring and CFD Analysis: A Case Study at the Coc Sau Open-Pit Coal Mine (Vietnam) . . . . . . . . . . . . . . . . . . . 224 Van-Duc Nguyen, Chang-Woo Lee, Xuan-Nam Bui, Hoang Nguyen, Quang-Hieu Tran, Nguyen Quoc Long, Qui-Thao Le, Xuan-Cuong Cao, Ngoc-Tuoc Do, Won-Ho Heo, and Ngoc-Bich Nguyen Numerical Investigation of Characteristics of Mine Ventilation Using One or Two Ducts in Underground Mining Faces . . . . . . . . . . . . 245 Minsik Kim, Jongmyung Park, Youngdo Jo, Dongkil Lee, and Huiuk Yi An Experimental Study on the Turbulent Diffusion Coefficients in Large-Opening Multi-level Limestone Mines . . . . . . . . . . . . . . . . . . . 263 Chang-Woo Lee and Van-Duc Nguyen Increasing Productivity and Safety in Mining as a Chance for Sustainable Development of Vietnam’s Mining Industry . . . . . . . . . 283 Duong Duc Hai, Le Duc Nguyen, Nguyen Duc Trung, Marian Turek, and Aleksandra Koteras

Contents

xi

Analytical Study on the Stability of Longwall Top Coal Caving Face . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 Tien Dung Le Recycling Ash and Slag of the Thermal Power Plant to Replace Protective Pillars in Mao Khe Coal Mine, Vietnam . . . . . . . . . . . . . . . . 320 Phi-Hung Nguyen, Vladimir Ivanovich Golik, Manh-Tung Bui, Thai-Tien-Dung Vu, and Van-Chi Dao Development of the Global 21st Century Mining Technical Services Professional: The WMI-SAGE Collaborative Model . . . . . . . . . . . . . . . 344 Sarfraz Ali, Frederick Cawood, Tariq Feroze, and Hamid Ashraf Assessment of Feasible and Effective Technologies for the Chemical Utilization of Domestic Coal for Value-Added Production in Vietnam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364 Michaela Scheithauer, Patricio E. Mamani Soliz, Roh Pin Lee, Florian Keller, Bernd Meyer, Xuan-Nam Bui, and Tong Thi Thanh Huong Shrinkage Cracking During Filtration Experiments - Influence of Suspension Concentration on Crack Formation . . . . . . . . . . . . . . . . . 385 Thanh Hai Pham and Urs A. Peuker Modern Methods of Dry Mineral Separation - Polish Experience . . . . . 407 Waldemar Mijał, Ireneusz Baic, and Wiesław Blaschke Becher Method Application for Ilmenite Concentrates of Vietnam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426 Tien Thuat Phung and Ngoc Phu Nguyen Upgrading of Vang Danh Coal Fines Using Reflux Flotation Cell . . . . . 436 Nhu Thi Kim Dung, Nguyen Hoang Son, Vu Thi Chinh, and Tran Van Duoc Silicosis Prevalence and Associated Factors Among High-Risk Population Group in Vietnam in 2018–2019 . . . . . . . . . . . . . . . . . . . . . . 453 Huyen Thi Thu Nguyen, Huong Thi Le, Huong Thi Lien Nguyen, Quan Thi Pham, Duy Van Khuong, Anh Ngoc Nguyen, Nguyen Nhu Tran, Thao Thanh Nguyen, Doanh Quoc Nguyen, Huong Thi Mai Phan, Nhung Thi Kim Ta, Anh Mai Luong, and Xuan Thi Thanh Le Knowledge, Attitude, and Practices (KAP) on Silicosis Among High-Risk Worker Population in Five Provinces in Vietnam . . . . . . . . . 469 Viet Nguyen, Huyen Nguyen Thi Thu, Huong Le Thi, Anh Nguyen Ngoc, Duy Khuong Van, Quan Pham Thi, Nguyen Tran Nhu, Thao Nguyen Thanh, Doanh Nguyen Quoc, Huong Phan Thi Mai, Nhung Ta Thi Kim, Anh Luong Mai, Huong Nguyen Thi Lien, and Xuan Le Thi Thanh

xii

Contents

The Impact of Afforestation on Seepage Water Formation on Post-mining Spoil Heaps and Dumps - Results of Water Balance Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485 Christian Hildmann, Lydia Rösel, Beate Zimmermann, Dirk Knoche, and Michael Haubold-Rosar A Conceptual Digital Framework for Near Real-Time Monitoring and Management of Mine Tailing Storage Facilities . . . . . . . . . . . . . . . 498 Iqra Atif, Hamid Ashraf, Frederick Thomas Cawood, and Muhammad Ahsan Mahboob Research on the Use of Fly Ash for Underground Mine Supports in Quang Ninh Coal Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 531 Trong Hung Vo, Van Kien Dang, Ngoc Anh Do, and Van Ha Truong The Potential Use of Waste Rock from Coal Mining for the Application as Recycled Aggregate in Concrete . . . . . . . . . . . . . 550 Nguyen Cong Thang, Nguyen Van Tuan, Dao Ngoc Hiep, and Vu Manh Thang Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563

About the Editors

Xuan-Nam Bui is currently a full professor at the Surface Mining Department, Faculty of Mining, Hanoi University of Mining and Geology (HUMG), Vietnam. He received the B.Eng. and M.Eng. degrees in Mining Engineering from HUMG in 1996 and 2001, and the Dr.Eng. degree in Mining Engineering from TU Bergakademie Freiberg, Germany, in 2005. He has been working at HUMG since 1996. His research interests are environment-friendly mining technology and engineering, occupational safety and health in mining industry, and applications of artificial intelligence and machine learning in geoengineering, mining and environmental issues such as ground vibration, over-pressure, fly rock, air pollution, slope stability. He is the editor-in-chief of the Journal of Mining and Earth Sciences, HUMG, and an editorial board member of several international scientific journals. He has more than 60 articles published in ISI and Scopus indexed journals, and Springer book chapters. Currently, he is the member of the Society of Mining Professors (SOMP) and Vice President of Vietnam Association of Mining Science and Technology. Changwoo Lee received BS in Mineral Engineering at Seoul National University, Korea, in 1978, MS and Ph.D in Mining and Operations Research at Pennsylvania State University, USA, in 1983 and 1986. Currently, he is a full professor at the Department of Energy and Mineral Resources Engineering, Dong-A University, Korea. He served as Dean of the engineering college and Dean of the graduate school of industry and information, and also as Chairperson of Korea Energy and Mineral Resources Engineering Program. He has served on a variety of advisory and affairs committees for the Korean government and state-owned organizations.

Carsten Drebenstedt is since 1999 Professor of Surface Mining and was from 2000-2006 Vice-Rector of research and from 2013 to 2016 Dean of the faculty of geoscience, geoengineering, and mining at Technische Universität Bergakademie Freiberg. Before he starts his career at university, he worked 17 years in different

xiii

xiv

About the Editors

positions in production and planning departments in German lignite industry, including reclamation and land planning department, and was member of the board of an International Consulting Company. Fields of teaching and research activities are mine planning, mining technologies, ecology in mining, reclamation, mine water management, and raw material awareness. Carsten Drebenstedt is invited lecturer in Kasakhstan, Kenia, Laos, Namibia, Serbia, Slovakia, and Russia, and involved in accreditation missions for geology/mining-related study courses in Germany, Romania, and Kasakhstan. He is Secretary General of the World Forum of Resource Universities on Sustainability (WFURS), and member of the Society of Mining Professors (SOMP) where he represented Europe in the Council from 2006 to 2015. Up to now, he has organized more than 50 national and international conferences and published like co-editor/ co-author 16 books in German, English, Mongolian, and Russian languages, further he edited more than 60 proceedings, and more than 400 papers in scientific journals, conference proceedings, and university publications. He is member of the Saxonian academy of sciences and other international scientific academies. He is awarded with Honor PhD in Russia, Ukraine, Romania, and Bulgaria; Honor professor in Mongolia; Honor geologist of Mongolia; and Honor miner of Vietnam.

Assessment of Global Digital Height Models over Quang Ninh Province, Vietnam Nguyen Quoc Long1(B) , Ropesh Goyal2 and Xuan-Nam Bui4

, Luyen K. Bui3

,

1 Department of Mine Surveying, Hanoi University of Mining and Geology, Hanoi, Vietnam

[email protected] 2 Department of Civil Engineering, Indian Institute of Technology Kanpur, Kanpur, India 3 Department of Geodesy, Hanoi University of Mining and Geology, Hanoi, Vietnam 4 Department Surface Mining, Hanoi University of Mining and Geology, Hanoi, Vietnam

Abstract. Understanding the surface topography is very essential for numerous Earth science applications and national infrastructure projects. With the advent of the satellite era, there have been numerous Digital Height Models (DHMs) available in the public domain. We use here the term DHM that includes discussions on both Digital Elevation Model (DEM) and Digital Surface Model (DSM). In this study, we analyse three freely-available global one-arc second DSMs of SRTM, ASTER, and ALOS World3D, and a three-arc second DEM of MERIT through validating with precise ground control points over Quang Ninh Province. The DSMs are subsequently inter-compared pixel-wise. Analysing the DHMs over Quang Ninh is of importance because of the 1) topography consisting of mountains, hills and coastal areas, 2) large mining activities, and 3) post-mining recreational activities. The analysis shows that none of the global DHM is suitable for Quang Ninh province because of the old datasets used in generating the global DHMs and horizontal shifts among the DHMs. A few options are then suggested to generate a precise local DHM over Quang Ninh Province. Keywords: Accuracy assessment · Digital surface models · Digital elevation models · Quang Ninh · Vietnam

1 Introduction Height information is one of the most important inputs in various studies and infrastructure projects. However, getting precise heights is challenging. It is a known fact that precise levelling cannot be replaced by any technique to get accurate heights. For the past two decades, there is a substantial increase in planning and infrastructure developments, which includes applications and studies in various fields such as mining, post-mining, smart cities, inter-linking of rivers, flood, and other natural hazards monitoring and mitigation, all of which require height information. As such, there is a need for Digital Height Models (DHMs) at very large scales and bigger areas. Hence, running precise © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 1–12, 2021. https://doi.org/10.1007/978-3-030-60839-2_1

2

N. Q. Long et al.

levelling lines is no longer an option considering the cost-time-accuracy aspects. Therefore, researchers and managers are inclined towards using freely available DHMs. Most of these global height models are Digital Surface Models (DSM), and the only Digital Elevation Model (DEM) available in the public domain is the MERIT [1]. Therefore, in this study, we use the term DHM to discuss collectively DSM and DEM, unless specified particularly. These open-source DHMs have been used for several applications in Vietnam, such as reservoir inundation mapping [2], flood risk analysis and mapping [3–5], rice crops monitoring [6], studying tidal flats [7], small-scale landforms classification [8], environmental modelling [9], land evaluations for different land use types [10], coastal classification [11], among others. In the mining field, DHMs have been employed worldwide in several applications, e.g., establishing a 3D geological solid model in Poland [12], analysing the change of soil moisture content in China [13], monitoring mining-induced subsidence in Australia [14]. Vietnam as a whole has three major issues for which a precise DEM is required: 1) floods, 2) mining and 3) post-mining recreational activities and planning. Though the need for precise DEM is acknowledged from the published studies, a thorough investigation of the precision of freely-available DHMs in Vietnam has not been addressed adequately to date. In this study, we, therefore, carry out a rigorous vertical accuracy assessment of the three DSMs and one DEM (cf. Table 1) over Quang Ninh province. We have chosen this study area because of its varied topography, a large number of islands, active mining and post-mining activities being in practice. Table 1. Digital height models investigated in this study. SRTM V3.0 (SRTM)

ASTER GDEMV2 (ASTER)

ALOS World 3D (AW3D)

MERIT

Model type

DSM

DSM

DSM

DEM

Satellite Mission

Shuttle Radar Topography Mission (SRTM)

Terra

Advanced Land Observing Satellite (ALOS)

SRTM and ALOS

Resolution

1-arc second

1-arc second

1-arc second

3-arc seconds

URL

https://gdex.cr. usgs.gov/gdex/

https://search.ear thdata.nasa.gov/

https://www. eorc.jaxa.jp/ ALOS/en/aw3 d30/index.htm

http://hydro.iis.utokyo.ac.jp/~yam adai/MER IT_DEM

Reference

[15]

[16]

[17]

[1]

The remainder of the study is organised as follows. An overview of the study area and datasets is provided in Sect. 2 followed by methods utilised in Sect. 3. The results and discussions is provided in Sect. 4 and Sect. 5 concludes the study.

Assessment of Global Digital Height Models Over Quang Ninh Province, Vietnam

3

2 Study Area and Datasets 2.1 Study Area Quang Ninh is a province located in the north-eastern part of Vietnam (see Fig. 1), which is around 150 km away from Hanoi–the capital city of Vietnam. It is the home of Ha Long bay, one of the world heritage sites. With its area of more than six thousand square kilometres, Quang Ninh is the home of 1,110,000 inhabitants as of 2020 [18]. Quang Ninh’s topography is dominated by mountains and hills, occupying approximately 80% of the area. It is the principal coal basin of Vietnam with around 90% of Vietnamese coal exploited in this province. There are nearly 2000 islands belonging to Quang Ninh. It is bordered by the East Sea to the East and China to the North.

Fig. 1. (left) Vietnam map with Quang Ninh province occupying in the north-eastern part; (right) one-arc second SRTM digital surface model with ground control points indicated by black triangles used for ground validation.

2.2 Datasets We choose the most used Shuttle Radar Topography Mission (SRTM) and ASTER DSMs, relatively newer ALOS3D DSM, and multi error removed MERIT DEM (cf. Table 1). The spatial resolution of all the DSMs is one-arc second (1-s), while that of MERIT DEM is three-arc second (3-s). All the DSMs and DEM consist of physical heights with vertical datum being the Earth Gravitational Model 1996 (EGM96). For t validation of these height models, we use a total of 210 precise ground control points, which comprise latitude, longitude, and first-order levelling height. The topography of the study area (using SRTM 1-s DSM) along with the locations of GCPs are shown in Fig. 1. The minimum, maximum and mean values of the height information from all the datasets (DSM, DEM, and GCPs) are depicted in Table 2. The large differences in the minimum values among the datasets (cf. Table 2) are likely due to 1) the differences in image acquisition time from different satellite missions,

4

N. Q. Long et al. Table 2. Statistics of the height information of the used data. Units in meters. Min

Max

GCP

−21.2 904.1

SRTM1”

−88

63.3

1495

164.1

1497

163.0

−145 1494

163.5

ASTER1” −6 AW3D 1”

Mean

MERIT 3” −81.6 1473.5 159.7

2) the differences in their processing strategies using different approaches and tools, and 3) the change in the topography of the parts of the study area mainly due to mining activities. The largest minimum heights reaching up to −88 m (SRTM), −145 m (AW3D), and −81.6 m (MERIT) compared to ASTER is possibly because they are newer height models. The areas having these large negative values have been identified as the mining sites (see Fig. 2). We have also provided a ‘.kmz’ file for the readers to visualise the image time series of these sites. The maximum negative values of −21.2 m of the GCPs because the GCPs are not in the lower zones of the mining sites.

Fig. 2. Large negative DHM’s heights at identified two mining sites (a) Site 1 and (b) Site 2.

3 Methods The DHMs are first analysed point-wise by comparison with the GCPs. The DHMs’ heights at the GCP locations are extracted using the Geographic Information System

Assessment of Global Digital Height Models Over Quang Ninh Province, Vietnam

5

(GIS) platform. Further, descriptive statistics (minimum, maximum, mean, and standard deviation) are computed and histograms are plotted to analyse the distribution of errors. Graphical analysis is also done to check if there exists any relationship between the DHM point errors and the levelled heights, latitude, or longitude, separately. The three DSMs (SRTM, ASTER, and AW3D) are also analysed pixel-wise. We do not include MERIT in the pixel-wise analysis because of 1) different spatial resolution compared to those of the three DSMs and 2) different height model types (DSM vs DEM). This analysis is important because of the availability of several open-source DHMs, and to check if one can use any of the DHMs, or there must be some assessment before using the same for their respective application. In Quang Ninh province, flooding is one of the major problems along with another major challenge of post-mining land use planning. For both of these, a precise DEM is needed, which in general is substituted with the SRTM or ASTER DSMs. To assess the applicability of the open-source DSMs for practical planning, this pixel-wise comparison is important. It can reveal or highlight the limitations of using freely available DHMs. The flowchart of processing steps adopted in this study is illustrated in Fig. 3.

Fig. 3. A flowchart of processing steps adopted in this study

4 Results and Discussion The descriptive statistics of the point-wise analysis of the four DHMs by comparing with the 210 GCPs is provided in Table 3. Figure 4 depicts the distribution of the point errors, in which the difference shown on the x-axis indicates the point height error obtained by subtracting the DHMs-extracted point heights from the levelled heights of GCPs. From Tables 3 and 4, the mean values for all the DHMs height error are negative and also most of the height error points are congregated towards the left of zero on the histograms. This indicates that there is a substantial number of GCPs of which heights are comparatively lower than the corresponding DHM extracted heights. This trend might be reasonable because Quang Ninh province is an active mining site and the global opensource DHMs are not updated regularly. It is possible that there is a significant time difference between the DHM creation and the GCPs collection. The statistics shown in Table 3 do not give any idea on the paramountcy of one DHM over the others. However,

6

N. Q. Long et al. Table 3. Statistics of the DEM validation with respect to the GCPs. Units in meters. SRTM ASTER AW3D MERIT Min

−28.5

Max

45.8

45.6

61.0

64.5

Mean −16.1

−14.1

−16.7

−11.7

13.5

9.8

12.8

Std. dev.

8.8

−40.2

−26.7

−31.0

Fig. 4. Histograms of DEM (a. SRTM, b. ASTER, c. AW3D, and d. MERIT) validation w.r.t to the 210 GCPs

with the available ground, it does signify that the ASTER DSM is the least accurate among other DHMs in consideration. Because Quang Ninh province consists of mountains, hills and coastal areas distributed in different geographic locations, we try to find any possible relationship between the DHM errors with respect to increasing levelled heights (Fig. 5), latitude (Fig. 6) and longitude (Fig. 7). It is observed that the errors in the ASTER DSM fluctuates the most. This strengthens our observation from Table 3 regarding the ASTER DSM to be the least accurate among others. From Fig. 5, we suspect that Quang Ninh province is under subsidence, but we acknowledge that a detailed deformation analysis must be done to claim this interpretation. However, to some extent it is validated with the available Vietnam news on mining in Quang Ninh Province and a few recent published studies [19, 20]. Figure 6 shows that the height error varies significantly as the latitude decreases. This is possibly because of the presence of many islands towards lower latitudes in the study area. The descriptive statistics for the pixel-wise DSM analysis are listed in Table 4 with the pixel-wise difference map for the pairs of SRTM, ASTER, and AW3D DSMs provided

Assessment of Global Digital Height Models Over Quang Ninh Province, Vietnam

7

Fig. 5. DEM height errors with respect to increasing observed height

Fig. 6. DEM height errors with respect to increasing latitude

Fig. 7. DEM height errors with respect to increasing longitude

in Figs. 8, 9 and 10. There are large differences among DSMs with a range of −249 m to 235 m in Quang Ninh province. This might be possibly because of the spatial shifts

8

N. Q. Long et al. Table 4. Statistics of the inter-DEM comparison SRTM-ASTER

SRTM-AW3D

AW3D-ASTER

Min

−148.0

−249.0

−159.0

Max

200.0

183.0

235.0

1.1

0.5

0.6

10.8

9.6

11.3

Mean Std. dev.

among the DSMs. A non-negligible factor for these large differences can be the variation in the time of capturing and processing the images from different satellite missions. Furthermore, from Figs. 8, 9 and 10, it can be seen that most of the large differences lie in mining zones followed by the hilly areas. Also, Figs. 9 and 11 indicate bias and noise in ASTER DSM (see within 21°N–21.25°N and 106.75°E–107°E), possibly the stripe errors.

Fig. 8. A map of DSM pixel-wise difference for SRTM-ASTER

These observations suggest that one should not use any freely-available global DHMs, as-is, for practical applications unless a validation at local scales for horizontal and vertical accuracies is conducted. We acknowledge that horizontal accuracy assessment will make the study more reliable. However, based on vertical accuracy assessment only, we have shown that the available DHMs should not be used for planning purposes in Quang Ninh province. Hence, estimating the horizontal errors will not change our

Assessment of Global Digital Height Models Over Quang Ninh Province, Vietnam

9

Fig. 9. A map of DSM pixel-wise difference for SRTM-AW3D

Fig. 10. DSM pixel-wise difference maps for AW3D-ASTER

interpretation. We, suggest that a precise DEM should be generated using dense highprecise levelling [21], UAV [22], or LiDAR [23], in which the UAV method has been proven to have advantages in precise DEM generation and high spatial resolution in mining applications in Vietnam [24–27].

10

N. Q. Long et al.

Fig. 11. Scatter plots for DEM inter-comparison: (a) SRTM-ASTER, (b) SRTM-AW3D, (c) AW3D-ASTER (note the difference in y-axis)

The scatter plots of the errors from an inter-comparison of the three DSMs are shown in Fig. 11. The SRTM and AW3D DSMs make the pair having the largest number of small-difference pixels, which is also confirmed by the descriptive statistics in Table 4. This also indicates that there are significant biases in ASTER DSM.

5 Conclusions In this study, three DSMs and one DEM have been validated with the ground control points. The assessment was conducted only on the vertical accuracy and not on the horizontal accuracy. We could have done the horizontal accuracy assessment if we would have found any of the available DHMs accurate enough to be used ‘as-is’. The study revealed that Quang Ninh province cannot rely on global freely available DHMs for the main reason being mining activities and topography consisting of mountains, hills, and islands. Furthermore, we also suspect the subsidence in the province, which should be studied in detail. Even for post-mining recreational activities, a precise DHM, preferably DEM is needed which must be generated using 1) UAV, 2) LiDAR, 3) dense precise levelling, 4) dense GNSS survey complemented with a precise geoid model, or 5) a combination of these methods. Since these DHMs will be much reliable, given their updated height information, applications at hand should be pursued after vertical accuracy assessment followed by horizontal accuracy assessment. Acknowledgements. This work was financially supported by the Ministry of Education and Training (MOET) in Viet Nam under grant number B2020-MDA-14

Assessment of Global Digital Height Models Over Quang Ninh Province, Vietnam

11

References 1. Yamazaki, D., Ikeshima, D., Tawatari, R., Yamaguchi, T., O’Loughlin, F., Neal, J.C., Sampson, C.C., Kanae, S., Bates, P.D.: A high-accuracy map of global terrain elevations. Geophys. Res. Lett. 44, 5844–5853 (2017) 2. Aekakkararungroj, A., Chishtie, F., Poortinga, A., Mehmood, H., Anderson, E., Munroe, T., Cutter, P., Loketkawee, N., Tondapu, G., Towashiraporn, P.: A publicly available GIS-based web platform for reservoir inundation mapping in the lower Mekong region. Environ. Model Softw. 123, 104552 (2020) 3. Ho, L., Umitsu, M., Yamaguchi, Y.: Flood hazard mapping by satellite images and SRTM DEM in the Vu Gia-Thu Bon alluvial plain, Central Vietnam. Int. Arch. Photogram. Remote Sens. Spat. Inf. Sci. 38, 275–280 (2010) 4. Dang, K.B., Bui, Q.T., Pham, T.P.N.: A convolutional neural network for coastal classification based on ALOS and NOAA satellite data. IEEE Access 8, 11824–11839 (2020) 5. Huong, D.T.V., Nagasawa, R.: Potential flood hazard assessment by integration of ALOS PALSAR and ASTER GDEM: a case study for the Hoa Chau commune, Hoa Vang district, in central Vietnam. J. Appl. Remote Sens. 8 (2014) 6. Phan, V.H., Nguyen, A.T.: Applying water footprint model to estimate water requirements of rice crops. In: ICSCEA 2019, pp. 849–856. Springer (2020) 7. Tong, S.S., Deroin, J.P., Pham, T.L.: An optimal waterline approach for studying tidal flat morphological changes using remote sensing data: a case of the northern coast of Vietnam. Estuar. Coast. Shelf Sci. 236, 106613 (2020) 8. Ho, L.T.K., Yamaguchi, Y., Umitsu, M.: Delineation of small-scale landforms relative to flood inundation in the western Red River delta, northern Vietnam using remotely sensed data. Nat. Hazards 69, 905–917 (2013) 9. Hirose, K., Sanga, T., Arakawa, Y., Maruyama, Y., Tran, T.B., Tuan, H.T., Anh, N.H., Hang, H.T.M.: A case study of spatial analysis for environmental monitoring. In: Asian Association on Remote Sensing - 26th Asian Conference on Remote Sensing and 2nd Asian Space Conference, ACRS 2005, pp. 1626–1630 (2005) 10. Herzberg, R., Pham, T.G., Kappas, M., Wyss, D., Tran, C.T.M.: Multi-criteria decision analysis for the land evaluation of potential agricultural land use types in a hilly area of Central Vietnam. Land 8, 90 (2019) 11. Dang, A.T.N., Kumar, L.: Application of remote sensing and GIS-based hydrological modelling for flood risk analysis: a case study of District 8, Ho Chi Minh city Vietnam. Geom. Nat. Hazards Risk 8, 1792–1811 (2017) 12. Jaskulski, M., Nowak, T.: Transformations of landscape topography of the Bełchatów Coal Mine (central Poland) and the surrounding area based on DEM analysis. ISPRS Int. J. Geo-Inf. 8, 403 (2019) 13. Zhengfu, C.L.B.: DEM-based analysis of the change of soil moisture content in mine surface layer: with shendong mining area as the instance. Metal Mine 11 (2008) 14. Ge, L., Chang, H.C., Rizos, C.: Mine subsidence monitoring using multi-source satellite SAR images. Photogram. Eng. Remote Sens. 73, 259–266 (2007) 15. Farr, T.G., Rosen, P.A., Caro, E., Crippen, R., Duren, R., Hensley, S., Kobrick, M., Paller, M., Rodriguez, E., Roth, L.: The shuttle radar topography mission. Rev. Geophys. 45 (2007) 16. Tachikawa, T., Kaku, M., Iwasaki, A., Gesch, D.B., Oimoen, M.J., Zhang, Z., Danielson, J.J., Krieger, T., Curtis, B., Haase, J.: ASTER global digital elevation model version 2-summary of validation results. NASA (2011) 17. Tadono, T., Ishida, H., Oda, F., Naito, S., Minakawa, K., Iwamoto, H.: Precise global DEM generation by ALOS PRISM. ISPRS Ann. Photogram. Remote Sens. Spat. Inf. Sci. 2, 71 (2014)

12

N. Q. Long et al.

18. https://www.gso.gov.vn/Default_en.aspx?tabid=491 19. Quang, N.H., Takewaka, S.: Land subsidence and its effects on coastal erosion in the Nam Dinh Coast (Vietnam). Continental Shelf Research 104227 (2020) 20. Nguyen, L.Q.: A novel approach of determining the parameters of Asadi profiling function for predictiong ground subsidence due to inclied coal seam mining at Quang Ninh coal basin. J. Min. Earth Sci. 61, 86–95 (2020) 21. Nelson, A., Reuter, H., Gessler, P.: DEM production methods and sources. Dev. Soil Sci. 33, 65–85 (2009) 22. Akturk, E., Altunel, A.O.: Accuracy assessment of a low-cost UAV derived digital elevation model (DEM) in a highly broken and vegetated terrain. Measurement 136, 382–386 (2019) 23. Dou, J., Yunus, A.P., Tien Bui, D., Sahana, M., Chen, C.-W., Zhu, Z., Wang, W., Thai Pham, B.: Evaluating GIS-based multiple statistical models and data mining for earthquake and rainfall-induced landslide susceptibility using the LiDAR DEM. Remote Sens. 11, 638 (2019) 24. Bui, D.T., Long, N.Q., Bui, X.-N., Nguyen, V.-N., Van Pham, C., Van Le, C., Ngo, P.-T.T., Bui, D.T., Kristoffersen, B.: Lightweight unmanned aerial vehicle and structure-from-motion photogrammetry for generating digital surface model for open-pit coal mine area and its accuracy assessment. In: International Conference on Geo-Spatial Technologies and Earth Resources, pp. 17–33. Springer (2017) 25. Bui, X.N., Lee, C., Nguyen, Q.L., Adeel, A., Cao, X.C., Nguyen, V.N., Le, V.C., Nguyen, H., Le, Q.T., Duong, T.H.: Use of unmanned aerial vehicles for 3D topographic mapping and monitoring the air quality of open-pit mines. In˙zynieria Mineralna 21 (2019) 26. Nghia, N.V.: Building DEM for deep open-pit coal mines using DJI Inspire 2. J. Min. Earth Sci. 61, 1–10 (2020) 27. Van Le, C., Cao, C.X., Le, V.H., Dinh, T.: Volume computation of quarries in Vietnam based on Unmanned Aerial Vehicle (UAV) data. J. Min. Earth Sci. 61, 21–30 (2020)

Rock Mass Classification for the Assessment of Blastability in Tropically Weathered Limestones Ramesh Murlidhar Bhatawdekar1 , Edy Tonnizam Mohamad1 , Trilok Nath Singh2 , Pranjal Pathak3 , and Danial Jahed Armaghani4(B) 1 Geotropik-Centre of Tropical Geoengineering, Department of Civil Engineering,

Universiti Teknologi Malaysia, 81310 Skudai, Johor Bahru, Malaysia 2 Earth Science Department, Indian Institute of Technology Bombay, Powai,

Mumbai 400076, India 3 Department of Mining Engineering, Indian Institute of Technology,

Kharagpur, Kharagpur 721302, India 4 Department of Civil Engineering, Faculty of Engineering, University of Malaya,

50603 Kuala Lumpur, Malaysia [email protected]

Abstract. Rock mass classification systems have been developed in various applications of rock engineering design such as tunnelling, foundation, slopes, rippability, and excavatability. These systems are reviewed and gaps with respect to blastability are identified. Geomechanical properties and geological features of the rock mass are known to have a great and direct influence on the effects of blasting operations. The tropical climate which is characterized by heavy rainfall of about 2000 mm/year and wide temperature variation resulted in thick weathering profile consist of heterogeneous zones. In addition, tropical climate and post tectonic impacts on the rock mass often cause severe and deep weathering in complex rock formations. The uniqueness of tropical influences on the geoengineering properties of a rock mass leads to significant effects on blast performance, especially in the developmental blasting stage. This study aims to determine the significant factors of tropically weathered rock mass properties that influence blast performance and to develop a rock mass classification system for blastability in tropically weathered limestone. The studied limestones are classified into four classes, namely W1, W2, W3 and W4 depending on the extent of weathering W1 type limestone is absent in Sri Lanka. (Marine) sedimentary, sedimentary and metamorphosed limestones are found in Sri Lanka, Cambodia and Thailand respectively. Uniaxial compressive strength (UCS) of limestones tested is 55 MPa and 77 MPa. Rock mass properties (Range of rock quality designation (RQD), range of cavity (%) for limestones at the studied sites are (15–50%, 0%), (30– 85%, 0–14%) and (50–90%, 0–12%) respectively. Range of geological strength index (GSI) varies from 40–70, 30–60, 20–50 and 15–45. Blastability index (BI) is a better parameter as it considers joint orientation and other rock mass properties as compared to GSI. After reviewing various blastability systems, W1 and W2 limestone are comparable with intermediate spaced joints of blast quality system (BQS) for joint spacing 0.1–1 m as rock mass description - ‘blocky’ matches. On © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 13–44, 2021. https://doi.org/10.1007/978-3-030-60839-2_2

14

R. M. Bhatawdekar et al. the other hand, W3 and W4 type limestone are comparable with closely spaced BQS for joint spacing less than 0.1 m as rock mass description - ‘powdery/friable’ matches. Karst classification systems for engineering or cone and tower are not suitable for blastability as these systems compare the change in karst topography over a couple of centuries and wider area. Rock mass classification system for blastability is proposed for tropically weathered limestone based on limestone density, degree of weathering, RQD, cavities, porosity, water absorption, point load strength index and BI. Keywords: Blastability index · Cone and tower karst classification · Degree of weathering · Engineering karst classification · Geological strength index (GSI) · Rock mass classification (RMC) · Rock quality designation (RQD)

1 Introduction Limestone is commonly found near the surface and is mined for use in the manufacturing of cement and as dimension stone in construction sites. The researchers over the years have presented a significant amount of work explaining the behaviour of these rocks [1–6]. Weathering is defined as the transformation and disintegration of rock at or close to the Earth’s surface by physical, mechanical and/or chemical processes [7]. In tropical regions, due to high temperature, humidity, and rainfall, the rate of weathering is higher as compared to that in other areas. Rock mass classification is a system of classifying a rock mass into groups or classes based on defined ranges of physical properties and assigning a specific description (or number) to it so as to estimate the behaviour of the rock mass can be estimated. Rock mass classification systems are used in defining rock mass behaviour in pre-feasibility studies, preliminary design and final design of excavations in rock. Tropical climates result in different degrees of weathering of limestone, which affect the physico-mechanical properties of the rock mass and has a direct impact on its blastability. Weathering is the physical and chemical process of disintegrating of rock and minerals. Weathering of limestone occurs below the surface resulting in karst, dolines or sinkholes and the formation of caves. During the weathering process, geomechanical properties of limestone rocks resulting in lower compressive strength or weaker limestone formation. Weathering classification as per the British standard is fresh, slightly weathered, moderately weathered, highly weathered, and completely weathered. Tectonic activities such as faulting and folding in tropical rock masses result in changes in the physical and mechanical properties of the rock. In addition, these rock masses are further exposed to weathering which affect their geomechanical properties. In civil and mining projects, blasting may be necessary for breaking rocks prior to excavation. The engineering properties of these rocks have a significant influence on the environmental effects of blasting (such as flyrock and ground vibration) and the efficiency of blasting operations (such as fragmentation and back break). Thus, the effective assessment of the different geomechanical properties is necessary prior to blasting to avoid unnecessary adverse effects. Backbreak can cause future flyrock. A new uncertain rule-based fuzzy approach for prediction of backbreak was developed by [8].

Rock Mass Classification for the Assessment

15

The geological structure has an impact on flyrock [9–12]. Various artificial intelligence techniques have been developed for prediction of flyrock [13–17]. Prediction of air overpressure is done through artificial intelligence techniques [12, 18]. Prediction of ground vibration due to blasting is developed using artificial intelligence techniques [18, 19]. The aim of this research is to investigate the relevant geomechanical properties related to the blastability of tropical limestone deposits in Thailand, Cambodia and Sri Lanka for developing appropriate rock mass classification systems which in turn would be used to assess blastability relevant to these areas. Building an information model of tropically weathered rock incorporating various multiparameter features of the deposit is useful for drilling and blasting throughout the life of mine [20]. Mining of limestone has an adverse impact on the surrounding environment [21–24]. Several countries have come out with frameworks of corporate social responsibility and sustainability, which are reported by organizations across the world [25–28]. For example, economic classification of gold considering long term data on gold mining and reported sustainability issues based on the degradation of gold mineral quality, degradation of the environment surrounding the mining areas and increased cost of production was carried out by [29]. 78 Chinese big size cities which are dependent upon the economics of mining minerals were studied by [30]. The key factors depending on Sustainable Development of Mineral Resources (DSDMR) were zanalyzed and classified into six clusters of mining, suggesting appropriate government policy. The system of ‘social license to operate’ in Canada was zanalyzed by [31] and suggested the JEL classification for further research in the governance and sustainability of minerals. Thus, every classification system related to limestone or mineral commodity has been useful for policyholders at government level or mine management level for maintaining the sustainability of mining operations. 1.1 Statement of the Problem Geomechanical properties and geological features have a huge influence on blasting operation. The complexity of these parameters is more crucial while dealing with tropical rock mass as the extreme climate change results in a different degree of weathering. Also, geomechanical properties of rocks are significantly influenced by tectonic related features such as fold, faults and joints. The objective of this research is to investigate the effect of geomechanical parameters of tropically weathered limestone on blastability. 1.2 Objectives of the Study This study aims to investigate the uniqueness of the tropical rock mass and its material properties that influence blastability. The subsequent objectives are designed as under 1. To determine significant factors of tropically weathered limestone properties that influence blastability 2. To develop a unique rock mass assessment system for blastability of the tropically weathered limestone.

16

R. M. Bhatawdekar et al.

1.3 Scope of the Study It includes the following: 1. Literature review on existing various systems of rock mass classification and assess gaps for blastability, 2. Reviewing rock mass classification of tropically weathered rock and comparing them with other rock mass classification systems in the tropical region. 3. Visit limestone quarries in Thailand, Cambodia and Sri Lanka which are situated in tropical regions to collect geological information and to study different rock types and their rock mass properties. 4. Analysis of common geoengineering parameters of tropical limestone suitable for assessing blastability.

2 Literature Review Rock mass classification is essential as it simplifies the complexity of the actual rock mass which could be understood by many professionals. Initially, all rock mass classification systems developed for tunnel support systems were based on qualitative parameters. The rock mass classification with quantitative parameters is practised globally in many countries [32–35] due to the following reasons: 1. Common understanding between designers, geologists, field engineers and contractors 2. Subjective observation and personal experience can be correlated effectively through quantitative system 3. Overall estimation of rock quality is better 4. Organization knowledge is improved and built up through a classification system. It can be also be used as a powerful tool in the feasibility of the design. Rock mass classification systems for tunnel support and blastability have a common factor of rock mass properties. However, the tunnel support system considers factors against gravity. On the other hand, for blastability, breaking of rock is important. Rock mass classification systems for excavation and rippability are connected with the respective equipment. Any of these equipment do not play any role in blastability. Rock mass classification systems for slope stability and blastability depend upon rock mass properties. However, blastability has little impact on water. Slope failure may occur with natural weak planes, gravitational force and external factors such as ground vibration etc. Blastability depends upon ease of breaking rock. Table 1 shows different types of rock mass classification systems and the purpose of application development with respective factors. A karst classification for engineering is suited for foundation design. While cone and tower karst classification are used mostly in Chinese Region, this classification system is more of regional level. Later the same would be discussed with the status of each limestone deposit with respect to this classification and the level of complexity with blastability. Table 2 shows the development of blastability system during the last six

Rock Mass Classification for the Assessment

17

Table 1. Development of different rock mass classification systems Type of system

Purpose of application development

Tunnel support

Load factor theory [36]

Shear zone, stand Ground condition uptime, safety, rate

Slope stability

Types of slope failures, potential of landslides [37–39]

Shear zone

Landslides

Excavatability

Ease in excavation, selection of equipment, rate of excavation [40]

Difficult to excavate

Size of equipment Rate of excavation

Rippability

Ease in ripping

Difficult to excavate

HP of dozer

Foundation

Dams, depth of socketing, type of foundation needed

Shear zone

Building bearing pressure

Blastability

Fragmentation, environmental issues

Muck profile

Blast design

Rate of tunnelling

Rate of ripping

Powder factor

decades with the main consideration for each system and research gap with blastability in the tropical rock mass. Impact of Rock Mass Properties on Blastability. Sedimentary rocks like limestone have structural discontinuities, variation in strata, cavities which has impacts on blastability of limestone. These various aspects are reviewed in this section. Structural Discontinuities. Discontinuities or joints refer to structural features and textural features that are inherent in situ rock masses, in macro or microforms. These could be joint planes, bedding planes, fractures and fault planes, cleavages and foliation etc. They have a significant influence on the geomechanical properties and dynamic properties of rock and therefore influence the extent to which the rocks deform under dynamic loading. Consequently, joints and discontinuities determine the response of rock masses in blasting. The nature of discontinuities, their location and orientation determine the strain wave propagation and the deformation characteristics of rock masses under different forms of applied loads. When the discontinuities are in a direction favourable to that of the strain waves, the fragmentation due to explosive induced transient stress waves would be enhanced. On the other hand, if the discontinuities are in an adversely oriented direction, it has a detrimental effect on the effect of the fragmentation.

18

R. M. Bhatawdekar et al. Table 2. Development of blastability system

Blastability systems Main consideration for development of blast ability assessment

Gaps with blast ability in tropical rock mass

[41, 42]

(i) Empirical equations developed due to complex nature of rock mass specific to certain sites

(i) Rock masses in tropics is subjected to high weathering. The profile is thick, and the geomechanical properties differ much with various weathering stage

[43]

(i) Blast ability index developed considering geology, rock mass properties and can be quantified in percentage

(i) Does not consider blast design, e.g. hole diameter, bench height (ii) Powder factor with blast ability index to be developed for each site separately

[44]

(i) Blast ability index developed for coal-bearing rock and can be quantified in percentage

(i) The nature of coal-bearing rock mass and tropical rock mass with limestone are different and will have different impact on blast ability

[45]

(i) Blast ability concept for coal-bearing strata (ii) Strength index proportional to powder factor (iii) Introduced heave index and breakage index proportional to Young’s modulus (iv) Modifier index proportional to structure, confinement

(i) Many numbers of indices which may be difficult to be easily understood by field personnel in charge of blasting (ii) Indices for coal-bearing rock mass will not suit for tropical rock mass for blasting due to its nature

[46, 47]

Indexed system for blast ability consisting of 12 factors (strength, deformability, discontinuity plane’s strength, in-situ block size, hardness, fragility, sturdiness and integrity of rock mass, resistance to fracturing and dynamic loading)

(i) Geomechanical properties only considered (ii) Influence of blast design is not considered (iii) For tropical rock mass, all the properties of rock mass may not have relevance and the same can be simplified

[48–50]

Blast ability index calculated for closely spaced, (i) Geomechanical properties only considered intermediate spaced and widely spaced rock (ii) Influence of blast design is not considered (iii) Tropical rock mass can be comparable with closely spaced rock. However, geomechanical properties vary exponentially, and there is scope for simplifying rock mass classification for tropical rock mass

A number of researchers [51–53] have studied the effect of discontinuities on blasting performance using full scale, reduced scale and small-scale model blasting. Their experiments showed that joints and discontinuities have a greater effect on blast performance than explosives properties and blast geometry. By small scale tests on bench shaped granite blocks, [54] concluded that joints and discontinuities even at 5% level have a significant impact on fragmentation as compared to granite blocks without joints and discontinuities. Influence of joints, bedding planes, hard bands and cavities influence blast performance has been discussed in detail by [55]. The relationship between orientations of discontinuities on different aspects of blast performance has been studied by several researchers [54, 56–59]. To study the influence of joints and discontinuities on blast performance, small scale tests were carried out in bench shaped granite blocks [57, 59]. In these tests, joints are discontinuities oriented in known directions as opposed to tests under field blasting conditions, where the orientation of joints and discontinuities are not always known and visible.

Rock Mass Classification for the Assessment

19

Presence of Cavities. Cavities (also called Vughs) are formed in rock bodies due to the dissolution of parts of rock bodies by groundwater. Sulphide ores, limestones and some iron ore deposits are subject to the formation of cavities. Large cavities (in some cases as much as 150 mm) have a detrimental effect on blast performance. Cavities cause bits to jam during drilling operations. When ANFO or other bulk explosives are delivered into boreholes intersecting cavities, it can cause explosives to flow into the cavities and accumulate there. This could result in overcharge of explosives energy in some spots of the borehole and reduced charge in the rest of the borehole. This could be overcome by using additional column charge in the stemming rise of the borehole immediately above the cavity or where it is not possible to obtain the stemming rise, explosives of higher energy in the upper part of the surrounding boreholes could be used. A method to charge a hole with a cavity was suggested by [60] and shown in Fig. 1.

Fig. 1. Charging of hole with a cavity (Hagan and Reid 1983)

In case where the design requires all blast holes to have uniform column length of explosives and uniform stemming columns, there is a tendency for a blast hole with a cavity to get overcharged and result in over breaks or cut-offs or high flyrock. There is a considerable drop in explosives pressure as the gases of explosion flow into cavities through joints and discontinuities. Sometimes, crack formations proceed in unintended directions due to the presence of cavities. Another serious consequence of cavities could be the premature initiation of explosives columns in the neighbourhood of cavities with explosives filled in it, thus disturbing the entire sequence of blasting operations and resulting in flyrock. Variability of Strata. In different sections of a mine, strata conditions vary widely. Most variation is observed in overburden strata, where every layer exhibits different geomechanical properties. For example, in a blast, if it is found that the lower layers of the strata is a hard competent rock mass and the upper layers are weak, fractured weathered rock, then explosives charges could be concentrated at the bottom of the borehole to fragment the

20

R. M. Bhatawdekar et al.

harder lower parts, while the upper part of the strata could get displaced and muck-piled to give easy digging during excavation operations.

3 Blastability Index (BI) This section describes blastability indices developed by [43] and [44] and compares them based on various input parameters. 3.1 Blastability Index by [43] An empirical method of BI based on observation of rock mass was developed by [43]. BI is site-specific and is developed on the basis of rock mass description, specific gravity and hardness, joint density and orientation. BI is calculated for various rock types as in (Eq. 1). BI = 0.5 × (RMD + JPS + JPO + SGI + H )

(1)

where, RMD = Rock mass description JPS = Joint plane spacing JPO = Joint Plane Orientation SGI = Specific gravity influence H = Mho’s scale of hardness. Table 3 shows blastability parameters with classification values. Rock mass description is given value of 10 for powdery or friable rock, 20 for blocky rock and 30 for totally massive rock. Joint plane spacing is specified as 10 for close spacing (1.0 m). Joint plane orientation value is fixed as 10 for horizontal joint planes, 20 for when joint plane dips out of face, 30 for strike normal to face and 40 for dip into face. Specific gravity influence is calculated by multiplying specific gravity of rock (t/m3 ) by 25 and deducting value by 50. Mho scale of hardness is in between 1 to 10 based on the type of rock. BI and powder factor relationship need to be developed for each site separately. Blastability Quality System (BQS). BQS is a beneficial approach involving the most valuable rock mass features, which can be simply estimated and employed in situ. The BQS connects the systems of classifying rock mass, Geological strength index (GSI), Rock Mass Rating (RMR), structural data, the hardness of rock mass, and BI. In the first stage, the discontinuities spacing is distinguished. In the second stage, the orientation of discontinuities in addition to the hardness of rock mass is described. The determination of the BI range will be easy if the above classification is done completely. Looking at a rock mass picture, discontinuities can be easily distinguished along with their spacing and orientation. Also, by using a Schmidt Hammer rock mass hardness can be estimated. At the final stage, the structure and surface conditions can be combined in order to estimate GSI [61] and RMR.

Rock Mass Classification for the Assessment

21

Table 3. Blastability parameters with classification values [43] Parameter

Description

RMD

Type of rock mass 10 = powdery, friable

JPS

20 = blocky

30 = totally massive

20 = intermediate (0.1 to 1.0 m)

30 = wide (>1.0 m)

Type of spacing 10 = close ( 1.04 can be ripped; while rocks with 3.2 < Q > 5.2 can be ripped and/or blasted which is consistent with the investigation was suggested by [71]. The rock excavation assessment as under rock masses excavatability assessment using GSI chart was suggested by [76]. The GSI [77] relates to the overall rock mass quality. It is based on an assessment of the lithology, structure and condition of discontinuous surfaces in the geological foundations and is estimated through visual examination of rock mass exposed in crops, surface excavations such as road cuts, tunnel faces or borehole cores. It utilizes two fundamental parameters of the geological process (block size of the mass and discontinuities characteristics); hence it takes into consideration the main geological constraints that govern a formation. Additionally, this index is simple to assess in the field. The GSI is therefore built on the linkage between descriptive geological terms and measurable earth field parameters such as joint spacing or roughness. So, based on this information, GSI uses the description of rock mass structure – as laminated and sheared, disintegrated, blocky and disturbed, very blocky, blocky and intact of massive – referring to the block size and discontinuity space and the description of surface conditions as very poor, poor, fair, good and very good – referring to the joint surface conditions. Blasting is required for rock masses with GSI > 65 and blocky or very blocky rock structures. They suggested hydraulic breaking necessary for the loosening of rock masses with GSI between 55 and 65. Successful ripping is suggested for rocks with GSI < 55; for very blocky rock masses GSI ~ 25; for disintegrated material and blocky/disturbed/seamy ~35; digging GSI 25 to 35 with increasing difficulty. These guidelines can be used for medium to medium-hard rock. GSI does not provide weightage for joint orientation with respect to face which is important during blasting. The excavation class intervals concerning mechanical efforts are classified into 8 classes from hand spade to blasting was proposed by [69, 78]. This classification is suitable for all types of natural materials. For blasting materials, classes 4 to 8 are comparable. Various parameters identified for excavatability as mass strength number, block size number, and joint strength number can be correlated with blastability of rock mass.

24

R. M. Bhatawdekar et al.

4.2 Assessment of Slope Stability with Rock Mass Classification The adjustment factors with respect to RMR, which is used by many researchers for slope stability assessment was developed by [37–39]. Rock mass classification system has been utilized by various researchers [79, 80]. Slope stability probability classification (SSPC) is developed by [81]. All these methods are used to assess the stability of rock mass while for blasting breakage of rock mass is important. Assessment is required for completed lifetime of the project. Earthquake, groundwater and environmental conditions have an impact on slope stability. There is no impact on blasting due to the earthquake. Environmental conditions such as fog, the weather may affect the management of blasting operation. Blast design needs to be modified by using more explosives for watery conditions. Rock mass classification for tropically weathered rock needs to be developed. Based on SSPC, comparing properties between weathered rock mass and fresh rock mass is an important input for developing the assessment system for blasting.

5 Development Rock Mass Classification Systems for Karst This system discusses the development of engineering classification of karst and development of cone and tower karst system. 5.1 Karst Classification from an Engineering Perspective To classify karst, the engineering field has defined different intricacies of ground conditions and probable threats. Karst in this field is organized into five classes in a progressive mode, which covers the phenomenon from its immature to its extreme mode. In this sense, there are three main parameters: caves (their size and extent), sinkholes (abundance and frequency of collapse), and rockhead (profile and relief). In the classification process, the intact rock strength is not considered. Figure 2 represents the concept of five classes. To be addressed well by the foundation engineer, this classification has characterized karst regarding two parameters, namely difficulty and complexity. In folded limestone, more complicated dissolution characteristics can be observed. Majority of the characteristics of the lower classes can be observed in karsts of higher maturity. During any ground investigation on karst, the most significant challenge is how to detect the underground cavities or how to find hidden voids. Much borehole drilling is needed for ground investigation in mature karst. For the detection of the underground cavities, two effective techniques are geophysical survey (for instance microgravity survey) and ground penetration radar survey. 5.2 Tower and Cone Karst System Chinese literature introduces fenglin and fengcong as two main types of karst terrain, which have a loose correlation with the Western terms of tower and cone karst, respectively [82].

Rock Mass Classification for the Assessment

25

Fig. 2. a Juvinile karst kI, b Youthful karst kII, c Mature karst kIII, d Complex karst kIV, e Extreme karst kV [82]

Fengcong Evolution. A broadly recognized fact is that fengcong karst results from a natural evolution of doline karst, which is formed once dolines expand toward coalescence. In this situation, they leave residual hills that finally tend toward conical in both plan view and profile.

26

R. M. Bhatawdekar et al.

Fig. 3. Stages in landscape evolution. Starting an initial plain, to Doline karst, to Fengcong to Fenglin Karst, and then back to Karst plain (Waltham 2008) a Stage 1: Initial surface, b Stage 2: Doline karst, c Stage 3: Fengcong karst, d Stage 4: Maturity of Fengcong karst, e Stage 5: Base Level is reached, f Stage 6: Fenglin karst is formed, g Stage 7: Slow tectonic uplift, h Stage 8: Lowering of towers, i Stage 9: Formation of karst plain [82]

Fenglin Evolution. In this phenomenon, a significant process is enhancement of the rock dissolution, which happens because of the chemically-aggressive water that is found at the water table. In other words, fenglin is a form of karst terrain with an extreme maturity level, which is able to progress only in the course of surface lowering through a great thickness of limestone. The simple sequence presented in Fig. 3 [82].

6 Findings of Site Study Limestone quarries in Thailand, Cambodia and Sri Lanka, which are in the tropical region are selected for this study (Fig. 4). These countries fall in wet tropical area having SW monsoon season in June- October and NE monsoon in December- March.

Fig. 4. Location of limestone quarries in the tropical region - Thailand, Cambodia and Sri Lanka

Rock Mass Classification for the Assessment

27

6.1 Site Study of Limestone Quarry at Thailand Construction aggregates in Thailand consist of limestone, basalt and granite. Potential aggregate resources and working quarries are located in various regions of Thailand. The quarries producing more than 200,000 cubic meters per month are termed as ‘Large aggregate’ quarries, and the quarries otherwise are termed as ‘small size’ quarries [83]. Most of large quarries are located in the Central part of Thailand, 100 km north of Bangkok. The large limestone quarries are mainly supplying limestone for manufacturing Portland cement. Regional Geology. Limestone study under study belongs to Thung Song Limestone Group of Ordovician age (505–438 million years) described as a banded argillaceous limestone. The formation was probably rather Thung Song Limestone, which deposited in Ordovician Age. The limestone can be divided into 3 units: first is a thin bedded argillaceous limestone, second is an argillaceous limestone, and the third one not bounded is a massive dark color limestone. Structure geology main fault is in NE-SW direction and main anticline recumbent folding type was across middle of project area. Limestone in these resources was re-crystallized due to metamorphism rocks. The geological map

Fig. 5. Geological map of limestone quarry

28

R. M. Bhatawdekar et al.

is shown in Fig. 5. Limestone deposit consists of highly weathered limestone, laminated limestone, somewhat weathered limestone and massive limestone. Exploration. 10 drill holes, 100 m/hole, spacing 500 × 500 m. Weekly update pit mapping indicated the geological profile and rock qualities estimation. The rock composition is as follows: 1. 80% Aggregates limestone: Thick bed argillaceous limestone and laminated limestone. 2. 20% Waste and cavity: Sheared rock, reddish rock, very thin bed/weathered limestone and cavity. Based on borehole logs following are observations are made: 1. Cavities in limestone varied from Nil to a maximum of 12% with average of 2%. 2. RQD varied from 50% to 90% with an average of 76%. Uniaxial compressive strength (UCS) of massive limestone of core sample was 77 MPa. The deposit is divided into four layers, given in Table 5. Table 5. Details of the limestone deposit in Thailand Layer name and type of material

Layer name

Thickness

The 1st layer is the overburden

D

2 to 5 m

The 2nd layer is highly weathered limestone

C

2 to 20 m

The 3rd layer is slightly weathered

B

2 to 30 m

The 4th layer is massive limestone

A

1 to 100 m

Figure 6a shows the developed limestone quarry faces. Limestone was classified based on GSI as blocky, very blocky, blocky/seamy and disintegrated as shown in Figs. 6b, c, d and e respectively. Exploration is carried out at quarry with 500 × 500 m holes. GSI, RQD (%) and BI for different rock types (blocky, very blocky, blocky/seamy, disintegrated) are given in Table 6.

Rock Mass Classification for the Assessment

29

Fig. 6. Limestone Quarry faces and classification of limestone based on GSI a Developed limestone faces showing benches, b Blocky, c Very blocky, d Blocky/seamy, e Disintegrated

Table 6. BI and GSI at limestone quarry in Thailand Type of limestone

Blocky

Very blocky

Blocky/seamy

Disintegrated

GSI - Minimum

40

30

20

15

GSI - Maximum

70

60

50

45

BI - Average

55.5

45.5

40.5

33

RQD (%) - Minimum

45

35

20

10

RQD (%) - Maximum

90

80

65

50

6.2 Site Study of Limestone Quarry at Cambodia The selected limestone deposit is located in Kampot province of Cambodia situated in different cone type hills. Figure 7 shows general topography of limestone hills. Limestone has karst, and one cave is found at the bottom of the hills, as shown in Fig. 8. Regional Geology. The limestone deposit consists of carbonate rocks of Permian age, located on the east side of the Pursat-Kampot Indonesian II Fold Belt. In the south of Cambodia, in Kampot province, the Devonian-Carboniferous deposits and Permian Limestone (Oolitic and organic-detrital varieties with intercalations of shales, marls and Argillaceous grits) are emerging as pinnacles from quaternary plain (alluvial sediments of clays, silts, pebbles and silts).

30

R. M. Bhatawdekar et al.

Fig. 7. a Limestone conical hill covered with vegetation, b Exposed limestone hill, c Series of limestone hills in Cambodia

Fig. 8. a Front view of cave, b Inner view of limestone cave in Cambodia

Limestone deposit in Cambodia has Argillaceous Limestone, Upper Micritic and Lower Micritic limestone, Upper Chetty and Lower Chetty Limestone, Semimble. Volcanic intrusion is intersecting different portions of limestone. Cherts are also found in limestone. Fault and fold are located in the deposit. Figure 9 shows several Limestone deposits. Upper Micritic Limestone (MIC1). MIC1 is a high-grade Limestone consisting of light gray compacted Limestone, partially weathered and containing some cracks filled with brownish sandy clay. This unit has a maximum thickness of 27 m and outcrops on a small area in the east side of the Western area. Argillaceous Limestone (ARG1). ARG1 consists of grey bedded Limestone, alternated with black Argillaceous or shaly Limestone. Some short lamina sequences may be seen in the greyish Limestone, but also in the blackish Limestone. This unit has a thickness of about 20 m and outcrops on a small area in the east side of the Western area.

Rock Mass Classification for the Assessment

31

Fig. 9. a Upper micritic limestone, b Outcrop of lower Micritic Limestone (MIC2), c Outcrop of Argillaceous Limestone in the field, d Character of Chert in Limestone massive rock

Lower Micritic Limestone (MIC2). MIC2 is the second horizon of high-grade Limestone consisting of light grey to grey compacted Limestone, containing fossil of crinoid stems. Short lamina sequences of blackish material can also be seen. The last three meters show a strong zone of recrystallization. This unit has a thickness of about 68 m and outcrops on a larger area in the east side of the Western area at the bottom of the hill. Cherty Limestone (CH). CH consists of dark grey compacted silica-rich Limestone containing fossils of fusulinids and crinoid stem. In the field, silicified fossils of Fussilina sp. and discontinuous layers of chert were observed. This unit has a thickness of over 31 m (considered down to 0 m MSL) and outcrops on a small area in the east side of the Western area. Upper Cherty Limestone (CH1). CH1 consists of grey to dark grey compacted Limestone with the fossil of brachiopods, weathered in the upper part with cavities and fillings of brownish clay or sandy clay. Some interlayers (2–7 m thickness) of dark grey to black Limestone can also be found in the drilling cores. Some intrusions of diabase were intercepted in two drillings. The chemical composition of the rocks shows several intervals of dolomitization of the compact Limestone. The thickness of this unit ranges between 28 and 62 m. This variation is due to the diabase intrusions and the weathering process. This unit outcrops on large surfaces on the south side of the hill of limestone depositeast area.

32

R. M. Bhatawdekar et al.

Argillaceous Limestone (ARG2). ARG2 consists of black Argillaceous and shaly Limestone. The lamina or shale sequences often contain organic material (graphite), which is laminitic and very soft. The compact black limestone shows white calcite veins. The strongest intrusions of diabase were intercepted in this unit. Semi-marble (SM). SM is the most important high-grade source. The SM consists of white to very light grey limestone partially metamorphosed (low-grade metamorphism), sometimes showing large white crystals of pure calcite. This unit has a thickness of about 120–140 m and outcrops on a limited surface on the north and northeast sides of the hill of east area. Lower Cherty Limestone (CH2). CH2 is very similar in aspect to the previous Cherty level (CH1). The CH2 is also consisting of grey to dark grey compacted Limestone with the fossil of brachiopods. The thickness exceeds 60 m (the lower contact was not intercepted by drillings) and it outcrops on a small area on the north side of the hill of east area, north of the semi-marble. Volcanic Intrusion. Diabase (DIA) Volcanic intrusions of aphanitic, fine grained diabase were discovered in the core samples. Due to the irregular geometry of the diabase occurrences, resources cannot be estimated reliably. The potential usefulness of the material will be discussed in a qualitative way. The total thickness of diabase for different rocks, as shown in Table 7. Table 7. Diabase in the drill holes Rock code Thickness (m) ARG2 SM ARG2

4.72 1.47 41.61

CH1

6.16

ARG2

6.25

CH1

2.2

Total

62.41

Chert. The chert distribution and its morphological shapes are variable, ranging from no chert or a few cherts (as in semi-marble or Micritic Limestone) up to between 20 and 30% of the rock mass, as in some parts of the Cherty Limestone or of the Argillaceous Limestone. In the field, several types of chert may be found: 1. Cherts Randomly Distributed in the Massive Body of Rocks: Dimension: from 1 cm up to 30 cm; Shape: from angular to round; Quantity: low concentration but locally up to 10% of the rock mass.

Rock Mass Classification for the Assessment

33

2. Cherts Distributed in the Rock Stratification: Dimension: several cms up to continuous strong strata; Shape: elongated or stratiform; Quantity: low concentration but locally up to 30% of the rock mass. Geological Structure. In the southwest corner and in the Western area, the geophysical work carried out for the plant foundations identified two NW-SE faults, probably of the strike-slip type. These two faults are responsible for the karstification, followed by a fast erosion of the Limestone from the Western area. The structure of the Eastern area of the deposit is a monocline with dipping to the southwest as below: 1. Fault (Fig. 10a): The various structures found in the West are strongly tectonized, which develop major faults and high folding. Major fault found fewer criterions in fieldwork, when compared with the Aerial photo to interpret. Two major faults expecting from Aerial photo have trended in NE-SW. The positions of the major faults have been measured and mapped.

Fig. 10. a Fault located in the limestone, b Fold located in limestone deposit of Cambodia

2. Fold (Fig. 10b): Folding structures are important in this area. It is the main structure to control the distribution of rock units. The results of traverse mapping and field measurement rock attitude indicated that, TMW has both of syncline and anticline. The main folding is syncline in the north-south direction. It occurred by compressive force in East-West direction after minor folding. Geological structure illustrates in the geological map and cross-section. Karsts. Karst phenomena are very common in limestone deposit of Cambodia, and the dissolution of the surface of the deposit is very strong. The field observations and also the drilling findings proved that the entire deposit is affected by different degrees of dissolution up to between 20 and 30 m under the surface (cavity, filling of cracks, small cavities, and re-crystallizations). The major caves, several meters wide, were observed at the foot of the pinnacles at the north part of Limestone area. The presence of karsts in the underground (possibly filled with clay) is expected. Total 2500 m of exploratory drilling has been completed. Rock samples were tested and results are given in Table 8. Exploratory borehole logs were analyzed. The details

34

R. M. Bhatawdekar et al.

are given in Table 9. Other features observed in bore hole logs are total cavities which are 4% of total drilling. Major cavities observed are in Cherty limestone and Argelleous limestone. Figure 11 shows the cavity filled and empty for both Argellous and Cherty limestone. Around 67% cavities are filled with clay and 33% cavities are not filled with any clay. Table 8. Test results of limestone samples from Limestone Deposit at Cambodia UCS - σC (Mpa)

Block

Dry density (gm/cc)

Modulus of elasticity - E (Gpa)

Block 1

2.67

43.6 ± 7.1

45.2 ± 18.9

Block 2

2.65

52.8 ± 18.4

33.8 ± 18.2

Block3

2.73

52.42 ± 12.4

48.9 ± 11.4

Block 4

2.69

75.3 ± 7.7

34.8 ± 1.9

Block 5

2.60

51.5 ± 9.5

37.0 ± 9.4

Table 9. Summary of cavities at limestone at Cambodia Cavity fill (m)

Cavity empty (m)

ARG2

20.77

24.52

45.29

CH1

46.19

8.87

55.06

SM2

0

Total (m)

66.96

Total (m)

0.2

0.2

33.59

100.55

Fig. 11. a Cavity filled and empty for Argellous limestone, b Cavity filled and empty for Cherty limestone

6.3 Site Study of Limestone Quarry at Sri Lanka Aruwakkalu Limestone is a part of Sri Lanka’s Jaffna limestone, which underlies the whole of Jaffna Peninsula and extends southwards mostly along the west coast as

Rock Mass Classification for the Assessment

35

Fig. 12. Jaffna Limestone Deposit [84]

shown in Fig. 12. Limestone deposit is of Miocene age in the southwestern part of the Aruwakkalu which is approximately 40 km from Puttalam Cement Works. The limestone deposit has occurred in the Miocene period. Millions of years ago, Sri Lanka and India were together and before it started to divide. A big lagoon was created between Sri Lanka and the southern part of India. In that lagoon, there were a large number of coral reefs which contained fossils. The lagoon then dried followed by the drying of corals. It is believed that Kala Oya flowed through this location. Then clay layers and sand layers were deposited on this dead limestone. Then red earth entered the cavities which were present in the limestone. It is believed that red earth is an Aeolian deposit which has come with wind from South India. Red earth has very fine particles. The deposit consists of 6 layers as shown in Table 10. Table 10. Layers in Aruwakkalu limestone deposit Layer number Description First

Red earth

Second

Low grade limestone

Third

Clay

Fourth

Low grade limestone

Fifth

High grade limestone

Sixth

The base

Limestone in Sri Lanka has fragmented remains of marine organisms such as coral or foraminifera. Various organisms which include ooids, peloids, intraclasts and extraclasts which secrete shells and leave them behind after they die. In some of the limestone, calcite or aragonite is precipitated through a chemical process and grains are not present.

36

R. M. Bhatawdekar et al.

Limestone often contains a fluctuating amount of silica in the shape of cherts, a fragmented skeleton with silica. Rivers also carry a varying degree of clay or silt which forms part of the limestone. Sudden supersaturated waters can deposit secondary calcite. Granular form of calcite which is found in speleothems, also known as stalagmites or stalactites and oolithic limestone. The first layer of red earth contains Al2 O3 , Fe2 O3 and Illuminate etc. In the second layer, low-grade limestone has pockets where the red earth is found. The third layer consist of clay. The fourth layer of low-grade limestone is free from red earth. However, occasionally clay is found in pockets. The fifth layer is high-grade limestone which is sedimentary in nature containing calcite and aragonite minerals. These minerals have different crystalline forms of calcium carbonate. The sixth layer or bottom-most layer

Fig. 13. a Bird’s eye view of limestone pit in Sri Lanka, b Loading limestone face, c 3rd clay layer of limestone deposit, d 6th layer of limestone deposit

Rock Mass Classification for the Assessment

37

is at the base of the deposit, which is close to groundwater and difficult to mine due to higher moisture content. Figure 13 shows the limestone deposits in Sri Lanka.

7 Comparison of Tropically Weathered Limestones in Thailand, Cambodia and Sri Lanka The limestone deposits evaluated under this study are located in the tropical regions of Thailand, Cambodia and Sri Lanka. However, the origin of the limestone deposits and the topography in each area is different. The limestone deposit in Thailand is above the nominal ground surface, and the limestone is of metamorphic origin. The limestone deposit in Cambodia lies within cone-shaped hills and is of sedimentary origin. Finally, the limestone deposit in Sri Lanka is a marine-based sedimentary deposit. The effect of different types of weathering is observed in each limestone deposit. Generally, the degrees of weathering of tropical limestones are classified as W1, W2, W3 and W4. The characteristics of the different limestone deposits are shown in Table 11. Table 11. Comparison of limestone deposits in Thailand, Cambodia and Sri Lanka Particulars

Limestone deposit in Thailand

Limestone deposit in Cambodia

Limestone deposit in Sri Lanka

Tropical region

Yes

Yes

Yes

Period of limestone formation

Ordovician age (505–438 million years)

Permian age (251–293 million years)

Miocene period (5.3–23 million years)

Engineering karst classification

kII or kIII

kIII

Not able to classify

Tower and Cone karst classification system for limestone

Not able to classify, Comparable with Stage 2: Doline karst

Stage3: Fengcong karst or Stage 4: Maturity of Fengcong karst

Not able to classify

Type of rock

Recrystallized and metamorphosed

Sedimentary

Sedimentary

Tropical weathered limestone classification

W1, W2, W3, W4

W1, W2, W3, W4

W2, W3, W4

Thickness of each type of limestone

W1 (1–100 m) W2 (2–30 m) W3 (2–20 m) W4 (1–5 m)

W1 (5–80 m) W2 (3–25 m) W3 (4–20 m) W4 (2–10 m)

W2 (2–5 m) W3 (3–8 m) W4 (5–15 m)

Type of limestone

Argillaceous

Argillaceous, Cherts, Micritic, Semi-marble

Marine limestone

Type of impurities

Clay

Cherts, clay

Cherts, Clay

Method of excavation

Drilling and blasting

Drilling and blasting

Drilling and blasting, ripping

UCS

77

55

Too brittle to determine

GSI Rock type (Range)

W1 (40–70%) W2 (30–60%) W3 (20–50%) W4 (15–45%)

Quarry under development. Can be developed later.

Unsuitable method as most of the limestone is W3, W4

RQD (%)

50–90

30–85

15–50

Cavity Range

Nil to 12%

Nil to 14%

Average Cavity (%)

2%

4%

Not detected as cavities filled with clay

Remarks

(continued)

38

R. M. Bhatawdekar et al. Table 11. (continued)

Particulars

Limestone deposit in Thailand

Limestone deposit in Cambodia

Limestone deposit in Sri Lanka

Cavity filling with clay Cavity – Type of rock

All cavities empty

67% cavities filled with clay

All cavities filled with clay

Argillaceous limestone

Argillaceous and upper cherty limestone, Rare in semi-marble

Marine limestone

Cave in deposit

Not found

One cave identified

Not found

Range of density (t/m3 )

2.5–2.7

2.6–2.73

2.2–2.5

Remarks

The engineering karst classification or the cone-and-tower karst classification is not a suitable method of rock mass assessment within tropical regions. However, it may be suitable for the assessment of the local complexity in limestone deposit. The limestone deposits in Thailand and Cambodia are harder as compared to those in Sri Lanka, and hence drilling and blasting are essential in these two deposits. On the other hand, the deposits in Sri Lanka have lower strength limestone with a weathering index of W3/W4 and thus can be excavated with ripping. The areas with a weathering index of W2 are harder and requires drilling and blasting prior to excavation. The UCS of the limestone from the deposits in Thailand and Cambodia were evaluated through laboratory testing, while the limestone from the Sri Lankan deposits could not be tested in the laboratory since the rock is too brittle. However, various researchers have carried out point load strength studies of the field samples from Sri Lanka. The compressive strength of the rock evaluated in the laboratory or the field can be correlated with blastability and is thus an important parameter for the assessment of blastability. GSI was evaluated for each type of limestone in Thailand where the distribution of weathering index was W1 (40–70%); W2 (30–60%); W3 (20–50%); W4 (15–45%). The GSI method can be used for the evaluation of the Cambodian limestone deposit. However, in Sri Lanka, most of the limestone has a weathering index of W3 and W4. Several researchers have reported that GSI is not a suitable method for rock mass assessment for W3 and W4 type of rock. Hence, GSI is not considered as an input parameter for the assessment of blastability as a part of the study. Rock density influences the blastability of rock and impacts the ease with which the rock can be broken by blasting. Rock density affects the propagation of explosive energy through the rock, the energy needed for the deformation of the rock and its displacement from the in-situ location. The lower the density of the rock, the less is the energy required for the deformation and displacement of the rock which results in better fragmentation. However, porous rocks, even though have a lower density, are more challenging to fragment because of higher energy absorption. Rocks of higher density need explosives with a higher velocity of detonation and higher detonation pressure. There is a variation of density for each limestone deposit at Thailand, Cambodia and Sri Lanka. Limestone deposits in Sri Lanka have lower density and require less explosive energy for blasting. With further local studies at each deposit, rock density can be correlated with the degree of weathering of the limestone (W1, W2, W3 and W4). The presence of cavities in the rock directly impacts the blastability of limestone as reported by various research studies. Limestone deposits in Thailand and Cambodia have an average of 2–4% cavities reported during drilling. Cavities within limestones in

Rock Mass Classification for the Assessment

39

Thailand are empty, while 67% of the cavities in the limestone deposits in Cambodia are filled with clay. The majority of limestones in Sri Lanka have been subjected to W3/W4 weathering, which has resulted in the cavities being filled with clay. Cavities are also reported in the argillaceous limestone and upper cherty limestone in Cambodia. These cavities can be correlated with the type of limestone and the degree of weathering. The limestone deposits in Cambodia have caves at the bottom of one of the hills which results in the formation of more complex limestone. The percentage of cavities is, therefore, an important assessment parameter for blastability. Various research studies have been carried out where water absorption and porosity can be correlated with different degrees of weathering and strength of rock, and hence these are also important parameters for the assessment of blastability. BI developed by [43] includes rock mass description, joint plane spacing, joint plane orientation, specific gravity index and hardness as input parameters. Various researchers have used BI for assessment of blasting performance. Hence, BI is one of the most important parameters for assessing blastability of limestone.

8 Need for Future Research Limestone deposits in this study are located within the tropical region. Each of the limestone deposits studied has unique characteristics. The deposits in Thailand are metamorphic limestones, those in Cambodia has are of sedimentary origin, while the deposits in Sri Lanka are classified as marine limestones. Each limestone deposit is affected by different degrees of weathering which influence blastability. Further research is required in the following areas for completing the assessment of the appropriate rock mass classification systems: 1. Weathering profile of each limestone deposit should be evaluated, and rock mass properties such as joint spacing, orientation and aperture, RQD (%) and GSI should be determined 2. Laboratory tests should be conducted for determining the physicomechanical properties of limestone samples from different weathering zones to help classify the laboratory test results in the four zones 3. Limestone cavities and their distribution needs to be identified during the exploration stage, and the data should be correlated with the observations found at the working quarry faces 4. The system of rock mass classification needs to be strengthened for improving the assessment of blastability.

9 Conclusion 1. The studied limestones are classified into four classes, namely W1, W2, W3, and W4 depending on the extent of weathering. 2. Various RMC systems were reviewed. Each application considers rock mass properties based on application. Geomechanical and geological features of rock mass are important factors when dealing with any type of sites especially in the tropically weathered limestone.

40

R. M. Bhatawdekar et al.

3. An engineering karst classification or cone and tower karst classification system are not suitable for blastability. Both the systems consider the change in topography throughout a couple of centuries and for the wider area. On the other hand, blasting area is small. 4. UCS of limestones tested are 55 MPa and 77 MPa. For the third site, UCS could not be evaluated as limestone being friable. Point load strength index is a suitable parameter to represent limestone strength. 5. Rock mass properties (Range of RQD, range of cavity (%) for limestones at the studied sites are (15–50%, 0%), (30–85%, 0–14%) and (50–90%, 0–12%) respectively. 6. Range of GSI varies from 40 to 70, 30 to 60, 20 to 50 and 15 to 45. BI is a better parameter as it considers joint orientation and other rock mass properties as compared to GSI. 7. W1, W2 type limestone and W3, W4 type limestones are comparable with intermediate spaced joints and closely spaced joints of BQS. 8. Assessment of rock mass classification for blastability is proposed for tropically weathered limestone based on limestone density, degree of weathering, RQD, cavities, porosity, water absorption, point load strength index and BI.

References 1. Fookes, P.G., Hawkins, A.B.: Limestone weathering: its engineering significance and a proposed classification scheme. Q. J. Eng. Geol. 21(1), 7–31 (1988). https://doi.org/10.1144/gsl. qjeg.1988.021.01.02 2. Edmond, J.M., Huh, Y.: Chemical weathering yields from basement and orogenic terrains in hot and cold climates. In: Tectonic Uplift and Climate Change, pp. 329–351. Springer (1997) 3. Sousa, L.M.O., del Río, L.M., Calleja, L., de Argandoña, V.G.R., Rey, A.R.: Influence of microfractures and porosity on the physico-mechanical properties and weathering of ornamental granites. Eng. Geol. 77, 153–168 (2005). https://doi.org/10.1016/j.enggeo.2004. 10.001 4. Gupta, A.K.: Effect of particle size and confining pressure on breakage and strength parameters of rockfill materials. Electron. J. Geotech. Eng. 14, 1–12 (2009) 5. Gupta, V., Sharma, R.: Relationship between textural, petrophysical and mechanical properties of quartzites: a case study from northwestern Himalaya. Eng. Geol. 135–136, 1–9 (2012). https://doi.org/10.1016/j.enggeo.2012.02.006 6. Sousa, L.M.O.: The influence of the characteristics of quartz and mineral deterioration on the strength of granitic dimensional stones. Environ. Earth Sci. 69(4), 1333–1346 (2013). https:// doi.org/10.1007/s12665-012-2036-x 7. Selby, M.J.: Hillslope Materials and Processes, 2nd edn. Oxford University Press (1993) 8. Hasanipanah, M., Amnieh, H.B.: Developing a new uncertain rule-based fuzzy approach for evaluating the blast-induced backbreak. Eng. Comput. (2020). https://doi.org/10.1007/s00 366-019-00919-6 9. Mohamad, E.T., Armaghani, D.J., Hasanipanah, M., Murlidhar, R.B., Alel, M.N.A.: Estimation of air-overpressure produced by blasting operation through a neuro-genetic technique. Environ. Earth Sci. 75(174) (2016). https://doi.org/10.1007/s12665-015-4983-5

Rock Mass Classification for the Assessment

41

10. Mohamad, E.T., Murlidhar, B.R., Armaghani, D.J., Saad, R., Yi, C.S.: Effect of geological structure and blasting practice in fly rock accident at Johor, Malaysia. J. Teknol. 78(8–6), 15–21 (2016). https://doi.org/10.11113/jt.v78.9634 11. Mohamad, E.T., Yi, C.S., Murlidhar, B.R., Saad, R.: Effect of geological structure on flyrock prediction in construction blasting. Geotech. Geol. Eng. 36(4), 2217–2235 (2018). https:// doi.org/10.1007/s10706-018-0457-3 12. Ye, J., Dalle, J., Nezami, R., Hasanipanah, M., Armaghani, D.J.: Stochastic fractal searchtuned ANFIS model to predict blast-induced air overpressure. Eng. Comput. (2020). https:// doi.org/10.1007/s00366-020-01085-w 13. Bhandari, S.: Engineering Rock Blasting Operations. A.A. Balkema, Rotterdam (1997) 14. Hasanipanah, M., Amnieh, H.B.: A fuzzy rule-based approach to address uncertainty in risk assessment and prediction of blast-induced flyrock in a quarry. Nat. Resour. Res. 29(2), 669–689 (2020). https://doi.org/10.1007/s11053-020-09616-4 15. Lu, X., Hasanipanah, M., Brindhadevi, K., Amnieh, H.B., Khalafi, S.: ORELM: a novel machine learning approach for prediction of flyrock in mine blasting. Nat. Resour. Res. 29(2), 641–654 (2019). https://doi.org/10.1007/s11053-019-09532-2 16. Murlidhar, R.B., Kumar, D., Armaghani, D.J., Mohamad, E.T., Roy, B., Pham, B.T.: A novel intelligent ELM-BBO technique for predicting distance of mine Blasting-Induced Flyrock. Nat. Resour. Res. (2020). https://doi.org/10.1007/s11053-020-09676-6 17. Hasanipanah, M., Keshtegar, B., Thai, D.-K., Troung, N.-T.: An ANN-adaptive dynamical harmony search algorithm to approximate the flyrock resulting from blasting. Eng. Comput. (2020). https://doi.org/10.1007/s00366-020-01105-9 18. Faradonbeh, R.S., Armaghani, D.J., Abd Majid, M.Z., Tahir, M.M., Murlidhar, R.B., Monjezi, M., Wong, H.M.: Prediction of ground vibration due to quarry blasting based on gene expression programming: a new model for peak particle velocity prediction. Int. J. Environ. Sci. Technol. 13(6), 1453–1464 (2016). https://doi.org/10.1007/s13762-016-0979-2 19. Amiri, M., Hasanipanah, M., Amnieh, H.B.: Predicting ground vibration induced by rock blasting using a novel hybrid of neural network and itemset mining. Neural Comput. Appl. (2020). https://doi.org/10.1007/s00521-020-04822-w 20. Bhatawdekar, R.M., Edy, M.T., Danial, J.A.: Building information model for drilling and blasting for tropically weathered rock. J. Mines Met. Fuels 494–500 (2019) 21. Venkatesh, H.S., Bhatawdekar, R.M., Adhikari, G.R., Theresraj, A.I.: Assessment and mitigation of ground vibrations and flyrock at a limestone quarry. In: Proceedings of the Annual Conference on Explosives and Blasting Technique, pp. 145–152. International Society of Explosives Engineers (1999) 22. Vyas, A., Pancholi, A.: Environmental degradation due to mining in south rajasthan: a case study of Nimbahera, Chittorgarh, India. J. Environ. Res. Dev. 4(2), 405–412 (2009) 23. Haule, H., Ringo, J., Luvinga, K., Kawonga, S., Mayengo, G., Morsardi, L.: Effects of limestone mining on deforestation and land degradation in Mbeya Region, Tanzania. Int. J. Mod. Soc. Sci. 5(2), 117–132 (2016) 24. Lamare, R.E., Singh, O.P.: Limestone mining and its environmental implications in Meghalaya, India. 24, 87–100 (2016) 25. Madden, D., Slaughter, G.: Sustainable Development Initiatives at Oparure Quarry. Waitam (2009) 26. Perez, F., Sanchez, L.E.: Assessing the evolution of sustainability reporting in the mining sector. Environ. Manag. 43(6), 949–961 (2009). https://doi.org/10.1007/s00267-008-9269-1 27. Mutti, D., Yakovleva, N., Vazquez-Brust, D., Di Marco, M.H.: Corporate social responsibility in the mining industry: perspectives from stakeholder groups in Argentina. Resour. Policy 37(2), 212–222 (2012). https://doi.org/10.1016/j.resourpol.2011.05.001 28. Sarkar, A.N.: Review of strategic policy framework for re-evaluating ‘csr’ programme impacts on the mining-affected areas in India. Adv. Sustain. Environ. Justice 1, 217–261 (2013)

42

R. M. Bhatawdekar et al.

29. Mudd, G.M.: Global trends in gold mining: towards quantifying environmental and resource sustainability. Resour. Policy 32(1–2), 42–56 (2007). https://doi.org/10.1016/j.resourpol. 2007.05.002 30. Yu, J., Zhang, Z., Zhou, Y.: The sustainability of china’s major mining cities. Resour. Policy 33(1), 12–22 (2008). https://doi.org/10.1016/j.resourpol.2007.10.001 31. Prno, J., Slocombe, D.S.: Exploring the origins of ‘social license to operate’ in the mining sector: perspectives from governance and sustainability theories. Resour. Policy 37(3), 346– 357 (2012). https://doi.org/10.1016/j.resourpol.2012.04.002 32. Wickham, G.E., Tiedemann, H.R., Skinner, E.H.: Support determinations based on geologic predictions. In: N Am Rapid Excav & Tunnelling Conf Proc, pp. 43–64 (1972) 33. Bieniawski, Z.T.: Engineering Rock Mass Classifications: A Complete Manual for Engineers and Geologists in Mining, Civil, and Petroleum Engineering. Wiley, Canada (1989) 34. Deere, D.U., Deere, D.W.: Rock Quality Designation (RQD) After Twenty Years. Colorado (1989) 35. Singh, B., Goel, R.K.: Rock Mass Classification: A Practical Approach in Civil Engineering. Elsevier (1999) 36. Terzaghi, K.: Rock Defects and Loads on Tunnel Supports. Harvard University, Massachusetts (1946) 37. Romana, M.: New adjustment ratings for application of Bieniawski classification to slopes. In: Proceedings of the International Symposium on Role of Rock Mechanics, pp. 49–53. Zacatecas (1985) 38. Romana, M.: The geomechanical classification SMR for slope correction. In: 8th ISRM Congress. International Society for Rock Mechanics and Rock Engineering (1995) 39. Romana, M.R.: A geomechanical classification for slopes: slope mass rating. In: Hudson, J. (eds.). pp. 575–600. Pergamon (1993) 40. Pettifer, G.S., Fookes, P.G.: A revision of the graphical method for assessing the excavatability of rock. Q. J. Eng. Geol. 27(2), 145–164 (1994). https://doi.org/10.1144/gsl.qjegh.1994.027. p2.05 41. Rakishev, B.R.: A new characteristic of the blastability of rock in quarries. Sov Min. 17(3), 248–251 (1981). https://doi.org/10.1007/bf02497198 42. Gupta, R.N.: A Method to Assess Charge Factor Based on Rock Mass Blastability in Surface Mines. MINTECH Publications (1990) 43. Lilly, P.A.: An Empirical Method of Assessing Rock Mass Blastability. The Aus (1986) 44. Ghose, A.K.: Design of drilling and blasting subsystems - a rock mass classification approach. In: Mine Planning and Equipment Selection. AA Balkema, Rotterdam (1988) 45. Scott, A., Cocker, A., Djordjevic, N., Higgins, M., La Rosa, D., Sarma, K.S., Wedmaier, R.: Open Pit Blasting Design Analysis and Optimization. Julius Kruttschnitt Mineral Research Centre, Queensland (1996) 46. Latham, J.P., Lu, P.: Development of an assessment system for the blastability of rock masses. Int. J. Rock Mech. Min. Sci. 36(1), 41–55 (1999). https://doi.org/10.1016/s0148-9062(98)001 75-2 47. Azimi, Y., Osanloo, M., Aakbarpour-Shirazi, M., Bazzazi, A.A.: Prediction of the blastability designation of rock masses using fuzzy sets. Int. J. Rock Mech. Min. Sci. 47(7), 1126–1140 (2010). https://doi.org/10.1016/j.ijrmms.2010.06.016 48. Chatziangelou, M., Christaras, B.: Blastability index on poor quality rock mass. Int. J. Civ. Eng. (IJCE) 2(5), 9–16 (2013) 49. Chatziangelou, M., Christaras, B.: A geological classification of rock mass quality and blast ability for widely spaced formations. J. Geol. Resour. Eng. 4, 160–174 (2016). https://doi. org/10.17265/2328-2193/2016.04.002

Rock Mass Classification for the Assessment

43

50. Christaras, B., Chatziangelou, M.: Blastability quality system (BQS) for using it, in bedrock excavation. Struct. Eng. Mech. 51(5), 823–845 (2014). https://doi.org/10.12989/sem.2014. 51.5.823 51. Ash, R.L.: The Influence of Geological Discontinuities on Rock Blasting. Dissertation, University of Minnesota (1973) 52. Vutukuri, V.S., Bhandari, S.: Some aspects of design of open pit blasts. In: National Symposium on Rock Fragmentation, p. 55. Institution of Engineers, Australia (1973) 53. Larson, W.C., Pugliese, J.M.: Effects of Jointing and Bedding-Separation on Limestone Breakage at A Reduced Scale, Washington, DC (1974) 54. Bhandari, S.: Studies on fragmentation in rock blasting. Dissertation, University of New South Wales (1975) 55. Hagan, T.N.: The effects of some structural properties of rock on the design and results of blasting. In: Third Australia-New Zealand Conference on Geomechanics Wellington, p. 2. Institution of Professional Engineers, New Zealand (1980) 56. Belland, J.M.: Structure as a control in rock fragmentation. Can. Min. Metall. Bull. 59(647), 323–328 (1966) 57. Bhandari, S.: Influence of joint directions in blasting. In: 9th Conference Explosives and Blasting Techniques, pp. 359–369. Dallas (1983) 58. Singh, D., Sastry, V.: Influence of structural discontinuity on rock fragmentation by blasting. In: Proceedings of the 6th International Symposium on Intense Dynamic Loading and Its Effects, Beijing, (1986) 59. Bhandari, S., Badal, R.: Post-blast studies of jointed rocks. Eng. Fract. Mech. 35(1–3), 439– 445 (1990). https://doi.org/10.1016/0013-7944(90)90221-2 60. Hagan, T.N., Reid, I.W.: Performance monitoring of production blasthole drills-a mean of increasing blasting efficiency. In: International Surface Mining and Quarrying Symposium, pp. 245–254 (1983) 61. Hoek, E., Brown, E.T.: Practical estimates of rock mass strength. Int. J. Rock Mech. Min. Sci. 34(8), 1165–1186 (1997). https://doi.org/10.1016/s1365-1609(97)80069-x 62. Dey, K., Sen, P.: Concept of blastability - an update. Indian Min. Eng. J. 42(8–9), 24–31 (2003) 63. Murthy, V.M.S.R., Dey, K., Raitani, R.: Prediction of overbreak in underground tunnel blasting: a case study. J. Can. Tunneling Can. 109–115 (2003) 64. Chatziangelou, M., Christaras, B.: A new development of BQS (Blastability Quality System) for closely spaced formations. J. Geol. Res. Eng. 1, 24–37 (2017). https://doi.org/10.17265/ 2328-2193/2017.01.003 65. Franklin, J.A., Maerz, N.H.: Empirical design and rock mass characterization. In: Proceedings of the FRAGBLAST 5 Workshop on Measurement of Blast Fragmentation, pp. 193–201. Taylor & Francis, Montreal (1996) 66. Atkinson, T.: Selection of open pit excavating and loading equipment. Trans. Inst. Min. Metall. 80, A101–A129 (1971) 67. Weaver, J.M.: Geological factors significant in the assessment of rippability. Civil Engineer = Siviele Ingenieur 17(12), 313–316 (1975) 68. Bieniawski, Z.T.: Estimating the strength of rock materials. J. South Afr. Inst. Min. Metall. 74(8), 312–320 (1974) 69. Kirsten, H.A.D.: A classification system for excavating in natural materials. Civ. Eng = Siviele Ing. 24(7), 293–308 (1982) 70. Kaiser, P.K.: A new concept to evaluate tunnel performance - influence of excavation procedure. In: The 22nd U.S. Symposium on Rock Mechanics (USRMS). American Rock Mechanics Association, Cambridge (1981)

44

R. M. Bhatawdekar et al.

71. Abdullatif, O.M., Cruden, D.M.: The relationship between rock mass quality and ease of excavation. Bull. Int. Assoc. Eng. Geol. – Bull l’Association Int Géologie l’Ingénieur. 28(1), 183–187 (1983). https://doi.org/10.1007/bf02594813 72. Scoble, M.J., Muftuoglu, Y.V.: Derivation of a diggability index for surface mine equipment selection. Min. Sci. Technol. 4, 305–322 (1984). https://doi.org/10.1016/s0167-9031(84)903 49-9 73. Gribble, C., McLean, A.: Geology for Civil Engineers, 2nd ed. George Allen and Unwin (1985) 74. Hoek, E., Karzulovic, A.: Rock mass properties for surface mines. In: Hustrulid, W.A., McCarter, M.K., Zyl, D. (eds.) Society for Mining, Metallurgy, and Exploration, Littleton, pp. 59–70 (2000) 75. ISRM: Rock characterization testing and monitoring. In: Brown, E.T. (ed.), pp. 113–116. Pergamon Press Ltd., Oxford (1981) 76. Tsiambaos, G., Saroglou, H.: Excavatability assessment of rock masses using the geological strength index (GSI). Bull. Eng. Geol. Environ. 69(1), 13–27 (2010). https://doi.org/10.1007/ s10064-009-0235-9 77. Hoek, E., Kaiser, P.K., Bawden, W.F.: Support of Underground Excavations in Hard Rock. AA Balkema, Rotterdam (1995) 78. Kirsten, H.: Case histories of groundmass characterization for excavatability. In: Rock Classification Systems for Engineering Purposes. ASTM International (1988) 79. Pantelidis, L.: Rock slope stability assessment through rock mass classification systems. Int. J. Rock Mech. Min. Sci. 46(2), 315–325 (2009). https://doi.org/10.1016/j.ijrmms.2008.06.003 80. Basahel, H., Mitri, H.: Application of rock mass classification systems to rock slope stability assessment: a case study. J. Rock Mech. Geotech. Eng. 9(6), 993–1009 (2017). https://doi. org/10.1016/j.jrmge.2017.07.007 81. Hack, R.: RQD, slope stability classification, weathering. Eng. Geol. Course. (2011) 82. Waltham, T.: Fengcong, Fenglin, cone karst and tower karst. Speleogenesis Evol. Karst Aquifers 35(3), 77–88 (2008) 83. Tangchawal, S.: Quarry evaluation and control of bench blasting. ASEAN J. Sci. Technol. Dev. 23(1&2), 19 (2017). https://doi.org/10.29037/ajstd.85 84. Epa, R., Perera, N., Manamendra-Arachchi, K., Meegaskumbura, M.: Sri Lanka’s aruwakkalu fossil deposit dates to the Burdigalian age. Ceylon. J. Sci. (Biological. Sci.) 40(2), 163–174 (2011). https://doi.org/10.4038/cjsbs.v40i2.3933

Study on the Reasonable Parameters of the Concentric Hemisphere-Style Shaped Charge for Destroying Rock Trong Thang Dam1(B) , Xuan-Nam Bui2

, Tri Ta Nguyen1 , and Duc Tho To3

1 Le Quy Don Technical University, 236 Hoang Quoc Viet Ward, Bac Tu Liem District,

Hanoi 100000, Vietnam [email protected] 2 Hanoi University of Mining and Geology, 18 Pho Vien, Duc Thang Ward, Bac Tu Liem District, Hanoi 100000, Vietnam 3 Vietnam Mechanics Association, 64 Doi Can Ward, Ba Dinh District, Hanoi 100000, Vietnam

Abstract. Currently, increasing the ratio of useful energy for external charge to destruct rock is a necessary research direction. Therefore, using an analytical method, this study established the dependence of the explosive force and the productivity of using energy for the concentric spherical shaped charge on the uniform-geometric characteristic parameters of this shaped charge. A computational programme was developed in the MATLAB software to calculate, investigate and analyse the abovementioned rule of dependence. The analysis and evaluation of this rule combined with the analysis of the purpose of the rock obstacle destruction allow to select reasonable uniform-geometric characteristic parameters of the concentric hemisphere shaped charge. The model is designed and created for selected concentric hemisphere shaped charge to fill 40 g of plastic explosive C4. The surface of the inner hemisphere of this shaped charge is a hemisphere lining funnel that is made of steel with a thickness of 1 mm. Tests were conducted, and the explosion efficiency of the concentric hemisphere concentrated shaped charge model was compared with that of the normal concentrated charge with the same 40-g explosive amount. The obtained results show that the productivity of concentric hemispheres shaped charge is 2.7 times higher than that of the concentrated charge. Furthermore, in the experiments for grade M300 concrete samples, the destructive zone volume of the proposed hemispherical shaped charge was approximately 2 times higher than that of the conventional concentrated charge. Keywords: Charge · Shaped charge · Destroy · Blasting rock · Concentric hemisphere

1 Introduction Currently, during rock blasting work, the method of external charges is often used when borehole blasting is difficult to apply such as in deep excavation of the seabed under complex hydrological conditions and when the size of the project is not large or when © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 45–68, 2021. https://doi.org/10.1007/978-3-030-60839-2_3

46

T. T. Dam et al.

breaking oversized rocks or breaking construction structures on land or underwater [1, 3, 4, 7, 8, 11, 12, 14–16]. The external charge method has low efficiency because most of the energy is lost into air or water [1, 5, 7, 8, 11, 15]. To improve the rock destruction efficiency of the concentrated external charge, the following solutions are available. – Direction 1: Optimise the shape and the relative ratio among sizes of the concentrated charge. This direction is based on the harmony of the unit explosive pulse and the total explosive pulse of the explosive products impacting directly on the obstacle [1, 5, 6, 10, 21]; – Direction 2: Use layers of inert material such as soil, sand, or water to cover the charges to reduce the loss of explosive energy into the air and increase the energy transferred into the obstacle [7, 8, 11, 14, 15, 19–21]; – Direction 3: Design and apply different types of shaped charge to increase the useful energy to break obstacles [1, 2, 5, 8, 9, 11–16, 20]. This direction is mainly applied worldwide in the military for the purpose of cutting or piercing the obstacles made of steel or concrete; thus, the only criterion for evaluating the effect of breaking the obstacle is the depth of the cut; however, the width of the cut has to be as small as possible. Therefore, the volume of destruction is ignored [2, 5, 9, 11, 15]. However, for soil and rock destruction, the volume of the destruction area is considered to be the most important criteria for all types of charges [1, 17, 19]. The published research results on shaped charges to break the rock are limited and are mainly published in the form of commercial information. Examples include the Russian oversized rock blasting volumes, KZP-5, KZP-100, KZP-200, KZP-300, KZP-400 [13], and FRACMEX Nitro Nobel underwater shaped charge is used in Sweden and Spain [18]. Thusfar, there has not been a comprehensive theory on the soil destruction effect of shaped charge. The publication of calculation and experimental methods on shaped charges to destruct rock or soil has not been determined. Therefore, it is necessary to study solutions to use concentric hemisphere shaped charge to improve the rock destruction efficiency.

2 Analysing Theoretical Basis of Direct Mechanical Effects of Explosive Products Typically, explosive detonation speed is very large (i.e., 5000–8000 m/s). Therefore, when studying the direct effect of explosive products on obstacles, it is possible to assume that the charge is immediately detonated. In addition, when explosion is performed in air, the air pressure can be ignored because it is smaller than the pressure of the explosive product at the start of bursting out. From the basic abovementioned assumptions, it can be assumed that the entire initial energy of the explosive is the potential energy when the exploded material has completely transformed into the kinetic energy of the explosive product. During an instant charge detonation, first, the explosive product completely occupies the volume of the charge prior to detonation, and all explosive particles are immobile. The movement of explosive product particles starts from the outermost layer

Study on the Reasonable Parameters

47

on the surface of the charge. Therefore, a surface appears between the moving explosive product element and the stationary element. This surface moves deeply into the charge at the rate called the splash surface. The movement speed of the splash surface during all splash times remains the same. Therefore, at each moment, the volume of moving explosive product will be similar to the initial volume shape of the charge. Therefore, based on geometric characteristics, it is easy to observe explosive products flying in a certain direction. The interface between two adjacent splash surfaces is the bisectoral plane of two adjacent explosive surfaces. The ratio of the mass of explosive gas product that flies in a given direction and the total mass of charge is called the coefficient of explosive use in that direction (Fig. 1). If the initial density at all points of the charge is the same, the coefficient of explosive energy use in a certain direction is [1, 2, 6, 10, 21]: η=

Va V

(1)

where: Va - volume of the explosive product flying in the direction to be determined; V - volume of the total charge; η - coefficient of explosive energy use in a certain direction also known as explosive performance or the relative volume of the active charge.

t0 t1

t2

a

a a

a

Fig. 1. Explosion product splash diagram of a cube charge t0 , t1 , t2 - The location of the splash surface at three different times

Explosive products can only fly from the open side of an explosive. If any portion of the surface of the charge is in contact with an absolute hard and stationary obstacle, the explosive products will only fly in the direction of the remaining free surfaces. By applying the laws of conservation of energy, the momentum and mass allow to determine the full impulse effect on the obstacle of the square-box bottom shaped charge with the length edge of b, height H, and the cylindrical charge with diameter b and height H [1, 5, 6, 10, 21]:  (2) I = Qo Cμ where: C - weight of the charge, kg; Qo - specific heat of the explosive, J/kg; μ - factor of the form charge.

48

T. T. Dam et al.

Thus, to increase the productivity of obstacle destruction, it is necessary to increase the explosion pulse applied to the obstacles. When the explosive type and weight is fixed, the explosive impulse propagating into the obstacle increases when the energy use coefficient increases in the direction of striking the obstacle. An increase in the use of explosive energy is made by changing the geometric structure of the charge. Explosive theory has confirmed that [1, 5, 9, 10] when using external charges, if the concave surface with reasonable parameters is placed to face towards obstacles to be destructed, the destructive performance will be considerably improved.

3 Determining the Reasonable Congruent Parameters of the Shaped Charge 3.1 Establishing a Rule Depends on the Explosive Energy Efficiency of Concentrated Sphere–Shaped Explosive Volume with a Concentric Conical Cone Based on the theory presented in Sect. 2, in a general case, the charge is a sphere that is shaped with concave spheres, spheres with concentric O, and concave bottom facing towards the required environment surface to be destructed (Fig. 2). The charge has a diameter of D = 2R, height H = HD (H = H / D relative height of shaped charge). The outer sphere has radius r2 , and the inner sphere has radius r1 .

Fig. 2. Concentric spherical charge

Let rp be the radius of the sphere that is spaced evenly inward and outward: rtb =

r1 + r2 2

(3)

Let J be the point on the cross section of the charge, and the sphere is separated so that J is equidistant from the inner, outer and top surfaces of the charge. Thus, J is the intersection point of circle O with radius rtb with the line that is parallel to AA -spaced AA distance r1 2- r2 . The dispersion separator of the explosive product on the charge section is the curve AJJ A , where JJ is the centre circle O of the radius rtb , and AJ and

Study on the Reasonable Parameters

49

J A are the curves of points that are equidistant from the outer sphere and the top of the charge. To establish geometric relations, the following relation is used: ϕ D D D 2 sin2 22 ϕ2 = tan (1 − cosϕ2 ) = ϕ2 2 sin ϕ2 2 2 sin 2 cos ϕ22 2 2   2H = 2 arctan 2H ϕ2 = 2 arctan D

H = r2 (1 − cosϕ2 ) =

(4)

The radius of outer and inner spheres are: r2 =

D ; 2 sin ϕ2

r1 = ηd r2

(5)

Inside: ηd relative bottom concave sphere radius is the ratio between the radius of the bottom sphere r1 and the radius of the outer sphere r2 , which are: ηd = ηgh ÷ 1; ηgh = cosϕ2 From the relationship: r2 cosϕ2 = r1 cosϕ1 , get: 

r2 cosϕ2 ϕ1 = arccos r1





cosϕ2 = arccos ηd

 (6)

From the position of J, the following relations are obtained: r1 − r2 r1 (1 − ηd ) = r2 cosϕ2 + 2 2 r2 (2cosϕ2 + 1 − ηd ) cosϕj = 2rtb r2 (2cosϕ2 + 1 − ηd ) ϕj = arccos 2rtb rtb cosϕj = r2 cosϕ2 +

(7)

Explosive volume: V = Vc2 − Vc1 Where:

 π 3 r2 2 − 3 cos ϕ2 + cos3 ϕ2 3  π 3 = r1 2 − 3 cos ϕ1 + cos3 ϕ1 3

Vc2 = Vc1 Active explosive volume:

ϕ2 Va =

π(Rp sin ϕ)2 d(−Rp cos ϕ) − Vcl ϕ=0

(8)

50

T. T. Dam et al. Va

0.5

Va

0.5

0.45

0.45

0.4

0.4

0.35

0.35

0.3

0.3

0.25

0.25

0.2

0.2

0.15

0.15

Va

0.1

Va

0.1

0.05

0.05

0 0

0.1 0.2

0.3

0.4

0.5 0.6

0.7

0.8

0.9

0 0.2

0.3

0.4

0.5

0.6

H =0.5 Va

0.5

0.7

0.8

0.9

1

H =0.4 Va

0.5

0.45

0.45

0.4

0.4

0.35

0.35

0.3

0.3

0.25

0.25

0.2

0.2

0.15

0.15

Va

0.1

Va

0.1

0.05

0.05

0 0.4

0 0.5

0.6

0.7

0.8

0.9

1

0.7

0.8

0.75

H =0.3

0.85

0.9

0.95

1

H =0.2

Fig. 3. Graph of the power and productivity of explosive use of concentric spherical charge

Where Rp is the distance from centre O to the separating surface of the explosive product. When, ϕ ≤ ϕj , Rp = r2 ; when, ϕ > ϕj , Rp = RM : r2 − RM = RM cos ϕ − r2 cos ϕ2 r2 (1 + cos ϕ2 ) RM = (1 + cos ϕ) d(RM cos ϕ) = −RM sin ϕdϕ + cos ϕdRM r2 (1 + cos ϕ2 ) r2 (1 + cos ϕ2 ) sin ϕdϕ + cos ϕ d(RM cos ϕ) = − sin ϕdϕ (1 + cos ϕ) (1 + cos ϕ)2 

sin ϕ cos ϕ sin ϕ dϕ d(RM cos ϕ) = r2 (1 + cos ϕ2 ) − (1 + cos ϕ)2 (1 + cos ϕ) We obtain: ϕj Va =

3 πrtb

ϕ2 sin ϕdϕ + r2 (1 + cosϕ2 ) 3

ϕ

ϕ=ϕj

 2 π Rp sin ϕ



 sinϕ sinϕcosϕ dϕ − Vc1 − (1 + cosϕ) (1 + cosϕ)2

Study on the Reasonable Parameters

ϕ2 Va =

Vcj − Vc1 + πr23 (1 + cos ϕ2 )3 ϕ=ϕj

sin3 ϕ (1 + cosϕ)3



 cosϕ 1− dϕ (1 + cosϕ)

51

(9)

  3 2 − 3 cos ϕ + cos3 ϕ Inside: Vcj = π3 rtb j j The productivity of explosive energy use is the coefficient of using explosive energy in that direction, which is determined by formula (1). The force of the charge applied to an obstacle (referred to as the force of the charge) is characterised by the relative volume of the active explosive block [2, 4]: Va =

Va D3

(10)

The general force of the charge impacting on an obstacle is characterised by its relative volume [2, 4]: V=

V D3

(11)

The formulas (8) to (11) can be unified to the diameter of the charge D. Thus, the explosive uses a factor that depends only on two parameters H and ηd , where V < 0, 5 and ηd = ηgh ÷1. Figure 3 shows the power and productivity of explosive energy using different values of numbers and ηd . The graph in Fig. 3 shows that: – Explosive power varies inversely with explosive performance; – The bigger is the concave (ηd big), the more efficiently the charge increases and the more explosive power it decreases; – The higher is the height of the charge, the higher is the explosive power and the lower is the efficiency of the charge. 3.2 Analysis and Selection of Reasonable Uniform Parameters of the Concentric Concave Spherical Charge The comparison of this concentric spherical cone-shaped shape charge model with the research results of the cylindrical concave explosion model with the conical concave funnel [4] allows us to make the following conclusions: The powerful concentric spheroidal blasting form (Fig. 2) of the active charge fraction (the relative volume of the active charge, Va ) is less than 0.1; however, the efficiency (coefficient of using explosive energy, η) is usually higher than 0.4 (see Fig. 3). The force of active detonation is the highest when the detonation height is equal to half of the detonation diameter (detonation size); then, the outside of the explosive quantity is the highest, i.e., hemispherical. The shape charges used to cut steel focus only on creating holes deeply into the obstacles; the smaller is the hole width, the better the charge performs. These types of shape charges are usually designed to ensure that the charge base is located far away, i.e.,

52

T. T. Dam et al.

2-3 times the obstacle diameter of the charge. This distance is called the focal length. The obstacle breaking principle of this shape charge is based on the kinetic effect of the penetrating flow formed from the lining hopper into the obstacle [1, 2, 5, 9, 17]. Compared to the concave blasting volume for steel cutting, the shape charge used in rock breaking needs to harmonise both the kinetic effects of the penetrating flow and the explosive product pressure, which is the shock wave pressure acting on the obstacle. The criterion for evaluating the effectiveness of rock breaking is the general volume of the destruction zone. Thus, if the principle of steel cut concave blasting used to break rock is applied, the efficiency will be very low. The above-mentioned analysis shows that to improve the rock breaking efficiency of concentric spherical shaped charge, this type of charge structure is chosen not only to harmonise the power factor and explosion efficiency, but also the design of the structure closer to the obstacle surface to enhance the effect of the shockwave and explosive product pressure on the obstacle. For the hemisphere charge, to make the power and distance to the obstacle surface (stone slabs) as small as required by the above-mentioned analysis, it is best to choose the concentric spherical charge with the highest height H = 0.5. According to the graph in Fig. 3, the parameters are selected in the region: H = 0.5;

= 0.6… 0.7;

= 0.45…0.46; Va = 0.09…0.08.

Thus, we can obtain and analyse one of the following five cases, with the same parameters described in Table 1. Table 1. Identical parameters of concentric hemispherical charge TT Parameters

CA1

CA2

CA3

CA4

CA5

1

Relative bottom concave sphere radius ηd

0.6

0.62

0.64

0.66

0.68

2

Relative height of charge H

0.5

0.5

0.5

0.5

0.5

3

Relative volume of charge V

0.2053

0.1994

0.1932

0.1865

0.1795

4

Force of a charge Va

0.0914

0.0894

0.0872

0.0847

0.0820

5

Productivity of explosive η, %

44.54

44.83

45.12

45.42

45.70

Natural rocks have properties that vary in a wide range with 4 levels, which correspond to the solidity ranges from 1 to 20. For the optimum rock breaking effect, an appropriate charge needs to be chosen for each type of rock or mine. The analysis of the above-mentioned 5 cases shows that an increase in explosive power and mine performance has an inverse relationship. The stronger is the rock, the stronger is the explosive force and vice versa. Therefore, for practical applications, this study chooses case 3 as the semi-concentric shaped charge to be tested on the model. According to case 3, the productivity of concentric shape charge is 45.12%, which is 2.7 times higher than that of concentrated charge with the same mass (the optimal

Study on the Reasonable Parameters

53

concentration charge has a diameter that is twice the height and efficiency of 16.7%). However, the productivity of rock destruction of the shaped charge, besides the dependence on the efficiency and power of the shaped charge, also depends on the momentum of the penetrated metal flowing formed from the conical cone. Therefore, the combined efficiency of the shaped charge and concentrated charge should be compared and evaluated through experimental results.

4 Experimental Framework 4.1 Purpose • Compare the destructive volume of the test sample of the concentric hemisphere shape charge to that of the concentrated charge with the same mass; • Compare stress and deformation values in the sample of concentric hemisphere on cave charge compared to the concentrated charge with the same mass. 4.2 Describe the Experimental Model and the Method of Conducting the Experiment The Experimental Model Includes the Following. The concentric hemisphere shape charge is designed in the form of a miniature model containing 40 g of flexible plastic explosive C4 with properties shown in Table 2; the geometric structural parameters of concave mines are described in Table 3 and Fig. 4 below. The shape charge is made of an aluminium shell to ensure a smaller impact of the durability of the shape charge on

Fig. 4. Design drawing of the concave hemispherical concentrated charge model 1. Plastic explosive C4; 2. Concave explosive aluminium shell; 3. The hemispherical lining funnel is made of steel; 4. The ring holding hopper liner

54

T. T. Dam et al.

the explosive effect; the hemispherical lining funnel is made of a 1-mm thick CT3 steel. The shell details of the concave explosion and hemispherical lining funnels are made by the lathe method on CNC equipment. The geometrical parameters of concave explosion quantity are shown in Table 3 and Fig. 3. Table 2. Characteristics of the C4 plastic explosive No

Parameters

Unit

Values

1

Moisture content and volatile matter

%

≤0.3

2

Flexibility (Needle penetration at 25 °C, uncompressed drug forming cakes)

10−1 mm

44–85

3

Impact sensitivity

%

8–32

4

Density

g/cm3

1.45–1.64

5

The ability to reproduce according to the magic pendulum

%TNT

116–136

6

Explosive speed (at density of 1.45 g/cm3 )

m/s

7300–7700

7

Lead cylinder pressure (25 g, density 1.45 g/cm3 )

mm

≥22

Table 3. Comparison of dimension parameters of concentric spherical charge and concentrated charge containing 40 grams of C4 No

Parameters

The concentric spherical charge

The concentrated charge

1

ηd

0.64

–

2

D, mm

50

D = 2H

3

H, mm

23.5

4

r 1 , mm

15

–

5

η, %

45.12

16.7

6

Force of a charge, Va

0.0872

0.261

– The amount of control concentrated charge is also 40 g of C4 plastic explosive with the parameters listed in Table 3. The concentrated charge is placed in a plastic pipe. The two ends of the explosive were left open. – The shaped charge is made of an aluminium shell, a 1-mm-thick steel funnel; the geometric parameters of the concave pillar explosion are shown in Table 3. – The material to be demolished is an M300 concrete sample with the size of 40 × 40 × 40 cm. – A set of concrete deformation measuring instruments includes: strain gauge PL-6011 from Japan (Tokyo Sokki Kenkyujo Co., Ltd) (see Fig. 5a) and a multichannel oscilloscope CSI-1000DC National Instruments (NI) from the USA, which has 24 channels and the maximum frequency of 200,000 signals/sec (see Fig. 5b). Strain

Study on the Reasonable Parameters

55

gauges are connected to a multichannel oscilloscope to receive distortion signals as a function of time after detonation.

Fig. 5. A set of measuring instruments for the deformation of concrete samples a. Multichannel oscilloscope measurement CSI-1000DC; b. Strain gauge PL-60

Method of Conducting the Experiment. We measured the destruction areas after explosion to compare and evaluate the destruction effect of soil and rock by shape charge to the concentrated charge. The instrument was used to measure the dimensions of the destruction areas by a ruler with millimetre accuracy. Strain gauges (PL-60 paste 10 cm) were used to assess the effect of load increase when detonating a shape charge compared to the concentrated charge from the concrete sample surface. Accordingly, strain gauge 4 were applied to concrete samples used to test the amount of concave explosion. Strain gauges 1 and 3 were arranged horizontally and glued to the concrete sample on two symmetrical sides of the concrete sample, while strain gauges 2 and 4 were arranged vertically along the specimen and stuck to the concrete sample on the remaining two symmetrical sides of the concrete sample (Fig. 6). Meanwhile, strain gauge 2 was applied to the concrete samples used for concentrated charge control. Strain gauge 5 was arranged horizontally, while strain gauge 6 was arranged vertically along the specimen. Both were attached to the two symmetrical sides of the concrete sample (Fig. 7). Before applying the strain gauges to the concrete samples, the part where the strain gaugepaste was placed was cleaned to smoothen the concrete surface and ensure that the sensors attached to the concrete samples well contacted and adhered to the surface of the concrete samples during explosion. This was to ensure that the true deformation value of the concrete sample was obtained.

56

T. T. Dam et al.

Fig. 6. Diagram of deformation sensors on a sample explosive for the explosion of the concave concrete sample

5

6

Fig. 7. Diagram of the deformation sensor on the explosive for the test explosion of the concentrated sample

The strain gauges were placed horizontally to consider the effect of the explosion on the deformation and stress of the specimen in the horizontal direction of the sample surface and vertically to consider the effect of the explosion on the deformation and stress of the sample along the sample surface. During testing, the amount of concave explosion was arranged in the explosive surface at 5 mm distance from the concrete sample surface to ensure the mechanical design of the shell structure. Moreover, the optimal amount of explosive for the explosion must be tested at other distances to determine the optimal focal length. The concentrated

Study on the Reasonable Parameters

57

charge will be arranged in the explosive surface in direct contact with the concrete sample surface. Each test charge (i.e., shape and concentrated charges) was placed on the concrete sample surface (Fig. 8) following the procedure below: – Perform the detonation with electric detonator No. 8. – After each explosion, measure the size and the volume of the destruction funnel on the sample surface with a ruler. Count the number of cracks and measure their depth. – The number of testing samples includes 3 concrete samples using explosive tests corresponding to 3 shape charges, of which 01 concrete sample is used to measure deformation, and 3 concrete samples using explosive test corresponding to 03 concentrated charges of control focus, including one sample for the deformation measurement. 4.3 Experimental Results Sample Destruction Areas. The results after the explosion showed that the test area received the effects of destroying the concrete samples of the concave and concentrated explosions, both of which have the following general form: – In the position in direct contact with the concave or concentrated charge, the destruction area had the form of an explosion funnel. Radial cracks appeared around the mouth of the funnel. Many cracks extended to the side surface of the sample and along the concrete body (Figs. 8 and 9). – Destruction funnel area: This area is a strong demolition-zone concrete shot out of the sample to master the demolition area in the shape of a funnel. This area is formed by the direct effect of the explosive pulse and the compressive stress in the explosion wave symbolised by index 3 in Fig. 8. – Cracking area: This area shows cracks denoted by symbols 5 and 6 in Fig. 8. These fractures are formed by the components of the effect of the tangential tensile stress in the explosion wave formed around the explosion funnel. The volume of this region is calculated to include a strong destruction zone. – The destruction characteristics of the concrete samples include the volume of the cracking area, area of cracks caused by fracture, number of cracks around the crater, crack depth along the body, and crack width in each explosive pattern. Table 3 depicts the results.

Fig. 8. Experimental diagram (a) and shape of areas of failure after explosion (b1, b2, and b3): 1) concrete sample M300; 2) location of the concave or concentrated charge; 3) funnel demolition; 4) piercing hole; 5) cracks on the face; and 6) cracks on the side

58

T. T. Dam et al.

Fig. 9. Some images of experimental explosion: a) sample before explosion; b) sample after explosion by shape charge; c) sample after explosion by concentrated charge; d) data taking after the explosion; e) strain gauge attached to the concrete sample horizontally; and f) strain gauge attached to the concrete sample vertically

Table 4. Demolition-zone characteristics TT

Type of charge, experimental conditions

1

The shape charge 2,360 is located at a 2,727 distance 5 mm 3,109

2 3

The average value 1 2 3

The volume of the funnel area destroyed, Vn (cm3 )

2,732

The concentrated 390 charge to 400 compare is 380 located close to the obstacle

The average value

390

Total number Crack of cracks depth/Crack around the width, Ln /b(cm) surface, N(cracks)

Volume of the fractured area in comparison, Vk (cm3 )

14

15–18/≤0.3

26,400

13

13–19/≤0.1

25,600

15

14–20,4/≤0.2

27,200

14

16.6

26,400

8

7–9/Mini-crack

12,800

7

8÷10/Mini-crack 14,400

6

6÷11/Mini-crack 13,600

7.0

8.5

13,600

Study on the Reasonable Parameters

59

Deformation Values on the Concrete Slope Surface Cracks appeared at the glued position of strain gauges 4 and 6 on the concrete sample after the explosion. On the contrary, at the glued position of strain gauges 1, 2, 3 and 5, no cracks were observed in the concrete sample after the explosion. Table 5 and Figs. 10 and 11 present the results of the deformation measurement obtained after the shape charge and concentrated charge experiments. The time relative strain ε (t) has the following characteristics:+ point with strain value ‘0’ corresponding to the time before explosion;+ score of the corresponding maximum relative strain value ε1 during the explosion impact; and + the point has the smallest relative strain value after the impact of explosion ε2. To analyse the mechanical and physical properties of the explosion impact process of the concrete sample, we included the stress and the specific energy quantities corresponding to the strain values obtained in the test. The dynamic stress value was calculated by the following formula: σ = Ed .ε, Mpa

(12)

where, ε is the relative strain value at the survey point; Ed is the dynamic elastic modulus of the concrete sample material used in the experiment determined as Ed = kd . Es ; Es is the static elastic modulus of the concrete sample material used in the experiment for the concrete graded M300; Es = 35,000 MPa; kd is the dynamic coefficient; and kd = 1.1–1.15. kd = 1.1 is chosen according to [22]. Table 5 presents the calculation results.

Table 5. Results of the relative strain measurements and the calculated stress values Strain Gauges title

Dynamic modulus of concrete E (MPa)

Maximum (Peak)

Ending Blast

Relative deformation, ε1

Stress value calculated, σ (MPa)

Relative deformation, ε2

Times of deformation

Unit energy

Stress value calculated, σ (MPa)

τ1/ τ2 (ms)

W/ W (kJ/m3 )

Note

The experimental shape charge Strain Gauges1

3.85E+04

0.00070573

27.17

0.000498577

19.195

43/265

9.59/ 4.80

Strain Gauges 3

3.85E+04

0.000871042

33.539

0.000860864

33.143

48/224

14.61/ 0.34

Strain Gauges2

3.85E+04

0.001882254

72.468

0.000217564

8.371

27/152

68.20 /67.29

Strain Gauges4

3.85E+04

0.009248257

356.059

0.009248257

356.059

2/-

1646.46/ 0.00

Appear cracks through the Strain Gauges

Appear cracks through the Strain Gauges

The experimental concentrated charge Strain Gauges5

3.85E+04

0.009014276

347.05

0.009014276

347.05

2/-

1564.20/ 0.00

Strain Gauges6

3.85E+04

0.000846098

32.571

0.000298204

11.484

61/481

13.78/ 12.07

60

T. T. Dam et al.

RelaƟve deformaƟon following Ɵmes series (Strain Gauges 1) 0.0008

τ1

τ2

Relative deformation

0.0007 0.0006 0.0005 0.0004 0.0003

0.0002 0.0001

Times (ms)

-0.0001

1 20 39 58 77 96 115 134 153 172 191 210 229 248 267 286 305 324 343 362 381 400 419 438 457 476 495 514 533 552 571 590

0

a) RelaƟve deformaƟon following Ɵmes series (Strain Gauges 3) 0.001

τ1

τ2

0.0009

Relative deformation

0.0008 0.0007

0.0006

0.0005 0.0004 0.0003 0.0002

TimesTimes (ms)

0.0001

(ms)

-0.0001

1 20 39 58 77 96 115 134 153 172 191 210 229 248 267 286 305 324 343 362 381 400 419 438 457 476 495 514 533 552 571 590

0

b) Fig. 10. Relative deformation diagram of the concrete sample received during the shape charge explosion: (a, b) strain gauge 1 and 3 horizontally attached to the concrete sample; and (c, d) strain gauges 2 and 4 attached vertically to the concrete sample

Study on the Reasonable Parameters

61

RelaƟve deformaƟon following Ɵmes series (Strain Gauges 2) τ2

Relative deformation

τ1 0.002

0.0015

0.001

0.0005

Times (ms) 1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301 321 341 361 381 401 421 441 461 481 501 521 541 561 581 601

0

-0.0005

c)

RelaƟve deformaƟon following Ɵmes series (Strain Gauges 4) 0.01

τ1

Relative deformation

0.008

0.006

0.004

0.002

Times (ms) 0 245

246

247

248

249

250

251

252

253

-0.002

d) Fig. 10. (continued)

254

255

256

257

258

259

260

62

T. T. Dam et al.

RelaƟve deformaƟon following Ɵmes seri (Strain Gauges 5) 0.01 0.009

Relative deformation

0.008 0.007 0.006

A

0.005 0.004 0.003 0.002 0.001

Times (ms)

-0.001

1 20 39 58 77 96 115 134 153 172 191 210 229 248 267 286 305 324 343 362 381 400 419 438 457 476 495 514 533 552 571 590

0

(A) in detail 0.01

τ1

0.009

Relative deformation

0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001

Times (ms) 52

51

50

49

48

-0.001

47

46

0

a) RelaƟve deformaƟon following Ɵmes seri (Strain Gauges 6) 0.0009

τ1

τ2

0.0008

Relative deformation

0.0007 0.0006 0.0005 0.0004 0.0003 0.0002 0.0001

Times (ms)

1 20 39 58 77 96 115 134 153 172 191 210 229 248 267 286 305 324 343 362 381 400 419 438 457 476 495 514 533 552 571 590

0 -0.0001

b)

Fig. 11. Relative deformation diagram of the concrete samples received when the explosive concentration is concentrated: a) strain gauge 5 horizontally adhering to the concrete pattern and b) strain gauge 6 vertically attached to the concrete sample

Study on the Reasonable Parameters

63

The energy (W) in the detonation stress wave transmitted to the concrete sample to reach the maximum relative strain value is calculated as follows: W=

1 Ed .ε2 2

(13)

The maximum value of the specific explosion energy (W) transferred to the concrete sample at the position of strain gauges is the energy value used to reach the maximum relative strain value ε1. The maximum specific energy value is calculated using as follows: W=

1 Ed . ε21 2

(14)

Table 5 lists the calculation results W corresponding to the maximum relative strain value. The part of the specific energy recovered from the maximum strain value ε1 to the value ε2 reflects the elastic potential of the concrete sample material in the explosion test (W) calculated as follows: W =

1 Ed .(ε21 − ε21 ) 2

(15)

Table 5 shows the calculation results W.

5 Analysis of Experimental Results Using the Excel section, we analysedthe data set in Table 4 in terms of the volume of the concrete demolition areas shown in the graphs in Fig. 12.

Fig. 12. Graph of the two cases for testing the concentric hemispherical and concentrated charges according to the volume of the destruction zone caused by compression and drag waves

64

T. T. Dam et al.

The analysis of the data in Table 4 and the graph in Fig. 12 showed that when detonating an explosion at a focal length of 5 mm, the destructive efficiency caused by the concentric hemispherical concave was approximately 2 times more effective in the area volume demolition funnel and 2 times in the volume of the fractured area in comparison with the corresponding volume of the concentrated explosion. The average number of the cracks and crack depth was approximately 2 times higher than that of the concentrated charge. The data analysis in Table 5 and the graph in Figs. 10 and 11 draw the following conclusions: – The calculated deformation values and the stresses received by strain gauge 6 when measuring the transverse deformation and pasting the samples both received positive values. This result reflects the calculated stress values as tensile stress. As regards the law of explosive mechanics, these stress values are the tangential tensile stress components of explosive waves. – The rule of the relative deformation change for all the strain gauges versus time has a common form: the relative deformation ε increases from ‘0’ (initial value) to the maximum value (ε1) in a very short period (τ1) of about some milliseconds to some 10 ms. During the time after the relative deformation reached the maximum value, one of the two following cases can happen at each location of the strain gauges: • Case 1: Relative deformation continues to keep the maximum value, indicating that the saturation value has not changed over time. The horizontal strain curve is ε (t) = ε1 = maximum = constant. This includes strain gauges 4 and 5. The corresponding maximum values of stress and the corresponding maximum specific energy values of strain gauges 4 and 5 are 356.059 MPa, 1646.46 kJ/m3 and 347.5 MPa, 1564.20 kJ/m3 , respectively. The values of the elastic strain potentials of the concrete sample materials in the explosion test (W) of strain gauges 4 and 5 are equal to 0. The analysis of the obtained relative strain graph of strain gauge 4 in the concave explosion test (Fig. 10d) and strain gauge 5 in the reference focused explosion test (Fig. 11a) showed that the relative strain value reached the maximum value when the strain increased in very short period of time τ1 = 2 ms. The relative strain value reached saturation the whole time the zone was greater than 2 ms. We observed cracks of the concrete sample cut across the two locations of strain gauges 4 and 5. This reflects that at these locations the effect of the blast stress wave exceeded the strength of the concrete sample, thereby forming a deformation zone beyond the elastic and plastic thresholds. Therefore, although strain gauge 4 was used in the concave blast test (Fig. 10d), strain gauge 5 was used in a controlled concentrated blast explosion test with two independent concrete samples corresponding to two attached different strain gauges to the concrete sample in both vertical and horizontal directions. Note, however, that the resulting maximum relative strain values were almost identical (ε1 = 0.009248257 and 0.009014276). • Case 2: The relative strain decreased from the maximum value ε1 to the value ε2 corresponding to time period τ2 . After time 2, the relative deformation reached the saturation value that did not change over time. This was a horizontal curve computed as ε (t) = ε2 = minimum = const. In this case, the pairs of the maximum stress

Study on the Reasonable Parameters

65

values and the corresponding maximum specific energy values for strain gauges 1, 2, 3 and 6 were 27.17 MPa, 9.59 kJ/m3 ; 72.468 MPa, 68.20 kJ/m3 ; 33,539 MPa, 14.61 kJ/m3 ; and 32,571 MPa, 13.78 kJ/m3 , respectively. The values of the potential elastic strain of the concrete sample materials in the explosion test (W) of strain gauges 1, 2, 3 and 6 were 4.48 kJ/m3 , 67.29 kJ/m3 , 0.34 kJ/m3 , and 12.07 kJ/m3 , respectively. We analysed the obtained relative strain plots of strain gauges 1, 2, and 3 in the concave explosion test (Figs. 10a, b and c) and strain gauge 6 in the reference concentration burst test (Fig. 11b). The relative strain value reached the maximum value in the time of strain increase τ1 = 27–61 ms and the entire next time zone from τ1 to τ2 (152–481 ms) relative deformation decreasing from the maximum value ε1 to the value ε2 and after the time τ2, the relative strain ε2 does not change over time. At the two locations, no cracks appeared at the concrete sample cut across strain gauges 1, 2, 3 and 6. The relative strain recovery from the maximum value ε1 to the value 2 reflected at the location of these strain gauges, the concrete sample is not destroyed. The relatively saturated strain 2 = const at times greater than τ2 denotes the irreversible plastic deformation zone, while the relative strain value ε2 is the residual strain in the concrete sample after the operation was completed. We then analysed the change of the relative strain graph over time when the relative strain value reached its maximum (i.e., it still appeared in the time zone τ2 of strain Gauges 1, 2, 3 and 6). The convex and concave points are the submaximum and minimum points, respectively, appearing in the process of reducing the relative strain from ε1 to ε2. Gauges 6 and 2 in Figs. 11 and 10 showed the most obvious points, reflecting the recovery process to a semi-head state of concrete when deformed beyond the elastic threshold under the effect of load. Gravity is the process of diminishing oscillation. The value of the relative residual strain ε2 reflects the intensity of the explosion stress wave acting at the survey point. The range of the relative deformation of strain gauge PL-60 was 0.02. The maximum relative deformation gained of strain gauges 4 and 5 was also less than 0.02. Therefore, the maximum specific energy obtained at the crack location in the concrete sample (strain gauges 4 and 5) was reflected as the limited value of the specific energy required to destruct the concrete sample material graded M300. To verify this problem, we must perform a comparative analysis with the explosive index required for the destruction of the concrete sample material by explosion. The specific energy of the explosive number 6 ammonite was Qv = 1028 kcal/kg = 4307 kJ/kg. Thus, the calculated maximum specific energy values of strain gauges 4 and 5 (i.e., 1646.46 kJ/m3 and 1564.20 kJ/m3 , respectively) were converted equivalent to the unit ammonite consumption of 0.38 kg/m3 and 0.36 kg/m3 respectively. The tested concrete sample graded M300 had a compressive strength of 30 MPa (300 kg/m2 ). This strength was equivalent to the conglomerate rock in the rock classification table with the unit explosive consumption value of ammonite 6 as 0.35-0.45 kg/m3 when blasting [17, 19]. This result coincides with the calculated unit ammonite explosive consumption converted from the experiment, confirming that the maximum specific energy value corresponding to the case of the crack is the critical specific energy required to destruct the M300 concrete sample. Comparing the maximum relative strain value or maximum tensile stress value in the corresponding non-destructive area of concrete when detonating the shape charge

66

T. T. Dam et al.

(strain gauge 2) with the explosion of the concentrated charge (strain gauge 6) draws the following result: 72.468Mpa ε1.2 σ1.2 0.001882254 = ≈ 2.2 = = ε1.6 σ1.6 0.000846098 32.571Mpa

(16)

where, ε1.2 and ε1 are the maximum relative strain values obtained from strain gauges 2 and 6, respectively, and σ1.2 and σ1.6 are the calculated maximum stress value corresponding to the corresponding maximum relative strain value of strain gauges 2 and 6, respectively. Comparing the corresponding maximum specific energy value of concrete when detonating the shape charge (strain gauge 2) with the concentrated charge (strain gauge 6) yields: ε2 0.0018822542 68.20 kJ W1.2 = = 1.2 = ≈ 4.5 2 W1.6 13.78 kJ 0.0008460982 ε1.6

(17)

in which: W1.2 , W1.6 is the corresponding maximum specific energy value of concrete when detonating the shape charge (Strain Gauges 4) compared with concentrated charge (Strain Gauges 6) respectively. From the comparative analysis of the results of the destruction volume area and the stress–deformation value of the concrete samples (i.e., when the hemispherical shape charge is compared with the control concentrated charge), the concentric spherical shape charge will destruct by approximately 2 times the destructive power of the concentrated charge with the same explosive mass. This research result is consistent with that obtained by Professor A.N. Khanukaev of the Leninsky Mining Institute of Russia: ‘stress wave parameters appear in granite under the effect of concave explosion with hopper lining in almost all experiments are more than two times higher than the amount of normal charge’ [12, 13]. Although the volume of the fracture zone in the fracture form when detonating the concentric hemisphere concave shape charge was only 2 times higher than that of the concentrated explosion, the corresponding maximum specific energy value was 4.5 times. This difference reflects the excess energy of the concentric hemispherical shape charge, which increases the number of cracks and shows a wider crack width in the concrete sample compared with the concentrated charge. When conducting experiments with the type of concave cylindrical charge with conical lining with angle at the top of funnel 78°, having the same volume of plastic explosive C4 is 40 g and tested to destroy concrete samples under the same conditions as described in above, the average value of 3 experimental samples was obtained: the volume of the funnel destroy Vn = 567 cm3 . The average volume of the total sample destruction by crack is Vk = 51,466 cm3 [4]. The comparison of the volume of the destruction areas of the concentric hemispherical concave volume with the number of cylindrical shape charge with a conical-shaped hopper showed that the value of the destruction funnel area of the concentric amount of the larger hemispherical concave was approximately 4.8 times higher than the amount of the cylindrical concave explosion with the conical concave. The mean volume of the entire sample destruction in the cracks of the hemispherical shape charge was approximately 2 times smaller than that of the

Study on the Reasonable Parameters

67

cylindrical concave with the conical concave. This result again showed that the volume of the concave- and funnel-shaped blast has a better destruction rate of the explosion area than the conical-shaped cone. The shape charge with a conical lining had a higher energy concentration and transmitted deeper than the amount of the concave-shaped hemispherical funnel.

6 Conclusions The following conclusions are drawn from this study: • Using the principle of the concentric concave hemispherical shape charge allowed the increase of the coefficient of useful explosive energy and explosive force in the direction toward the rock surface to be broken. With this shape charge (case 3), the efficiency of the concentric shape charge was 45.12%, which is 2.7 times higher than the concentrated charge with the same explosive mass. Although the power of the selected hemispherical concave shape charge was reduced by approximately 3 times smaller than the concentrated charge, the area of the explosive effect was reduced because of the effect of the energy convergence of the lining funnel in the shape charge, which increased the explosive effect of the shape charge. • The selected configuration of the concentric concave blasts H = 0.5 and ηd = 0.67 corresponding to the outer hemisphere height is half the diameter and the radius of the inner crest (i.e., 0.67 out). • By dint of the high efficiency and the increase of the explosive power of the shape charge, the volume of the destruction zone increased because of the direct effect of the explosion pulse and the compression wave 2 times destruction areas of the concentrated charge, when burst to break the grade M300 concrete sample. • The value of stress and deformation in the concrete samples caused by the impact of the concave explosion was approximately 2.2 times higher than the effect of the concentrated detonation. • The maximum specific energy value in the concrete sample caused by the effect of the concave shape charge was approximately 4.5 times higher than that of the concentrated charge. • The results of the theoretical and empirical studies emphasised that the application of a concentric hemispherical shape charge will increase the efficiency of breaking out of rock by using external chargeswhen using contact explosion by a common concentrated charge. Recommendation: We recommend the usage of the uniform parameters of the concentric hemisphere shape charge in case 3 to design and manufacture various types of final explosive with different capacities for rock and soil destruction on land or underwater. The calculated geometric parameters of the other capacity charge will be equal to the size of the shaped charge model multiplied by the explosive uniformity factor.

68

T. T. Dam et al.

References 1. Giao, H.S., Thang, D.T., Quyen, L.V., Chung, H.T.: Chemical explosion - theory and practice, pp. 149–154, 239–245, 413–424. Scientific and Technical Publishing House, Hanoi (2010). (in Vietnamese) 2. Tan, T.B., Doanh, T.V.: The method of calculating the force of theconcave bullet concern the detonation position. J. Sci. Technol. MTA, Hanoi (2016). (in Vietnamese) 3. Charter of explosive work, pp. 113–118. Command of the Engineers, Hanoi (1986). (in Vietnamese) 4. Thang, D.T., Doanh, T.V., Viet, T.D., Tho, T.D.: Study on the reasonable parameters of the cylinder shaped charge with tapered liner funnel to destroy stone. In: International Symposium on Rock Mechanics and Engineering – The 35th VSRM Anniversary ISRM 2019 Specialized Conference, 22–24 November 2019, Hanoi, pp. 118–130 (2019) 5. Baym, F.A., Ctankoviq, K.P., Xextep, B.I.: Fizika vzpyva, pp. 367–440,629– 637. Gacydapctvennoe izdatelctvo fiziko-matematiqecko litepatypy, Mockva (1975) 6. Blacov, O.E.: Ocnovy tepii dectvi vzpyva. BIA, Mockva (1957) 7. Bapon, B.L., Kantop, B.X.: Texnika i texnologi vzpyvnyx pabot v CXA. Hedpa, Mockva (1989) 8. Galkin, B.B., Gilmanov, P.A., Dpogoveko, I.Z.: Bzpyvnye paboty pod vodo. Hedpa, Mockva (1987) 9. Pokpovcki, G.I.: Bzpyv, pp. 48–56. Hedpa, Mockva (1980) 10. Calamaxin, T.M.: Pocobie dl pexeni po teopii mexaniqeckogo dectvi vzpyva, pp. 24–55, 98–99. BIA, Mockva (1967) 11. Pykovodctvo dl inenepnyx voick – podpyvnye paboty, pp. 23–39, 371–384. Boenno izdatelctvo minictepctva obopony coza CCCP, Mockva (1959) 12. Xanykaev, A.H.: Fiziqeckie ppoceccy ppi otboke gopnyx popod vzpyvom, pp. 161– 166. Hedpa, Mockva (1974) 13. Cofponov, A.A., Taqev, A.A.: Pazpyxenie gopnyx popod nepgie vzpyva kymyltivnyx zapdov – Ppoblema pazpyxeni gopnyx popod vzpyvom, pp. 189– 198. Hedpa, Mockva (1967) 14. Surface drilling and blasting, pp. 258–278. Tamrock (2001) 15. Explosives and demolition. Department of the army. Washington (1967) 16. Drilling and blasting of rocks, pp. 272–280. Geomining technological institute of Spain (1995) 17. Henrych, J.: The dynamic of explosion and its use, pp. 105–137, 278–295. Academia Prague (1979) 18. Olofsson, S.: Applied explosives technology for construction and mining, pp. 242–257. PrintedbyNoraBoktryckeriABNora (1975) 19. Kytyzov, B.H.: Pazpyxenie gopnyx popod vzpyvom - Bzpyvnye texnologii v ppomyxlennocti. MGGU, Mockva, 450 c (1994) 20. Pykovodctvo po podpyvnym pabotam, pp. 21–37, 371–384. Boenno izdatelctvo minictepctva obopony coza CCCP, Mockva (1969) 21. Calamaxin, T.M.: Pazpyxnenie vzpyvom lementov konctpykci. BIA, Mockva (1967) 22. Evaluation of dynamic modulus of elasticity of concrete using impact – echo method published in Construction and Building materials 47(20133), 231–239

Effect of Carbon Nanotubes on the Chloride Penetration in Ultra-High-Performance Concrete Pham Manh Hao1 , Nguyen Van Tuan2(B) , Nguyen Cong Thang2 , Nguyen Van Thao1 , Luong Nhu Hai1 , Pham Sy Dong3 , Nguyen Xuan Man4 , and Ngo Ngoc Thuy5 1 Center for High Technology Development, Vietnam Academy of Science and Technology,

Hanoi, Vietnam 2 Faculty of Building Materials, National University of Civil Engineering, Hanoi, Vietnam

[email protected] 3 Faculty of Building and Industrial Construction, National University of Civil Engineering,

Hanoi, Vietnam 4 Hanoi University of Mining and Geology, Hanoi, Vietnam 5 Institute of Techniques for Special Engineering, Le Quy Don Technical University,

Hanoi, Vietnam

Abstract. The improvement of Carbon Nanotubes (CNTs) addition on the properties of concrete has been recently investigated, e.g. mechanical properties and durability, but is rarely reported for Ultra-High-Performance Concrete (UHPC) with a very low water to binder ratio. This study evaluates the effect of CNTs on chloride penetration in UHPC and some mechanical properties of this material. The results of experimental tests show that the addition of CNTs in UHPC with contents from 0% to 0.5% by weight of binder in UHPC will reduce the workability of the concrete mixture, and not significantly improve the compressive strength of UHPC under both standard curing and heat curing conditions. However, the results showed that the addition of CNTs improves the dense microstructure of both the UHPC matrix and interfacial transition zone, and resulting in reducing the chloride penetration in UHPC. This is very important in cases of using UHPC for the constructions working under extreme conditions such as on a coastal, island, or underground structures with water erosion. Keywords: UHPC · Carbon Nanotubes · Chloride penetration · Compressive strength · Microstructure

1 Introduction Ultra-high-performance concrete (UHPC) is a concrete made from a mixture of quartz sand with a size of about 100–600 µm, cement, mineral admixture, superplasticizer, and water, in which the water to binder ratio (W/B) is very low (less than 0.25) [1–3]. This material has superior properties compared to conventional concrete, such as high workability, very high mechanical properties, i.e., the compressive strength of usually © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 69–80, 2021. https://doi.org/10.1007/978-3-030-60839-2_4

70

P. M. Hao et al.

over 150 MPa, low permeability, and superior durability. Due to these advantageous properties, UHPC is capable of opening the door for potential applications in the future, such as in the construction of thin, large structures, large-span bridge structures, highway roads, skyscrapers, corrosion-resistant structures, etc. [2, 4, 5]. In fact, the production of UHPC is based on the essential principles, including (i) Improving microstructure of concrete, (ii) Enhancing homogeneity, (iii) Optimizing the packing density, (iv) Promoting hydrate reaction, and (v) Bettering the toughness [6, 7]. Under special conditions of materials and manufacturing technology, the compressive strength of UHPC can be achieved over 800 MPa [1]. The application of the above principles will increase mechanical properties and especially the durability of UHPC, and this will contribute to improving the possibility of resisting the impact of aggressive erosion agents and deterioration of structures. For concrete structures when exposed to aggressive environmental conditions such as physical, chemical, or mechanical factors will lead to deterioration of concrete quality, these impacts often influence simultaneously, and accelerate the quality degradation of concrete [8, 9]. One of the reasons leading to corrosion on reinforcement and damage to construction works is the chloride ion penetration. When the concentration of that in concrete exceeds the safety threshold, it will break the passive protective layer of steel bar and causing corrosion [10–12]. Until now, many studies have been conducted to consider the effects of mineral admixtures, curing regime, W/B, corrosive environment, etc., on the extent of chloride ion diffusion into concrete [12–14]. Of those, the use of nanomaterials to improve the durability of concrete is also a topic of concern and research. Among materials in nanoscale, carbon nanotubes (CNTs) are relatively new materials with unique properties. The size of CNTs at the nanoscale, as well as the high specific surface area, allow this material to become an ideal reinforcing agent at the micro-level. It can improve the quality of concrete even at low contents by the ability to fill voids [15, 16], combine with hydration products, and bridge effect for microcracks to prevent the crack expansion propagation of concrete [17, 18]. With minimal content added, CNTs show an excellent interaction with binder paste to make a denser microstructure and resulting in better resistance to the penetration of corrosive agents compared to the reference without CNTs [16]. Studying the effect of CNTs on its possibility to resist the chloride ions penetration into UHPC and improvement of some other properties of concrete is an excellent approach to confirm the role of this material for concrete, especially UHPC intended for use in particularly harsh conditions, such as underground structures, structures on the coast or islands.

2 Materials and Methods 2.1 Materials In this study, some materials were used including quartz sand (S) with a mean particle size of 300 µm; Portland cement (C) PC40 with the properties meeting the requirements of Vietnamese standard TCVN 2682: 2009 (Table 1); Condensed Silica fume (SF) having a mean particle size of 0.15 µm; Fly ash (FA) with the particle size in the range of 0.05–50 nm; Polycarboxylate-based Superplasticizer (SP) with 30% solid content by weight; CNTs, used in concrete mixtures are long multi-wall carbon nanotubes, the

Effect of Carbon Nanotubes on the Chloride Penetration

71

inner diameter of those is from 5 to 10 nm and outer diameter from 10 to 30 nm, the density of 0.14 g/cm3 , and some physical properties of CNTs are given in Table 2. To increase the dispersion and reactivity of CNTs in concrete, CNTs were dispersed into the SP solution. The SEM images of cement and SF used in these experiments are shown in Fig. 1. The mixture of CNTs and SP solution was then dispersed for 30 min in an ultrasonic bath before using (Fig. 2). Table 1. Physical properties of Portland cement. Properties Retained on 0.09 mm sieve Fineness (Blaine) Standard consistency Compressive strength - 3 days - 28 days

Unit

Value

Specification

% 2 cm /g %

0.6 3870 29.5

≤ 10 ≥ 2800 -

MPa

29.8 52.2

≥ 21.0 ≥ 40.0

TCVN 4030-2003 TCVN 6017-2015 TCVN 6016-2011

Fig. 1. SEM images of (a) cement, (b) SF

Table 2. Physical properties of CNTs No Properties

Unit

Value

1

Color

2

Average Inside Diameter

3

Average Outside Diameter nm

10–30

4

Length

µm

10–30

Specific Surface Area

m2 /g

>140

Density

g/cm3 0.14

5 6

Black nm

Test methods (Vietnamese standard)

5–10

72

P. M. Hao et al.

Fig. 2. CNTs dispersing (a) in SP solution and (b) in an ultrasonic bath

2.2 Experimental Methods The determination of the influence of CNTs on the workability of UHPC was conducted by measuring the flow of the concrete mixture. The flow diameter of UHPC mixtures was controlled from 200 to 250 mm under the test method ASTM C1437. The compressive strength of concrete was determined at 28 days using cube samples with a size of 100 × 100 × 100 mm. The apparent chloride diffusion coefficient of UHPC was determined according to ASTM C1556 based on the cylinder samples of 100 mm diameter and 200 mm height. The evident chloride diffusion coefficient of UHPC was tested with variable CNTs contents from 0% to 0.5% by weight of the binder (the binder used here is a mixture of cement, FA and SF). At the same time, the other material components are kept the same. The principle of the method for determining the chloride ion diffusion coefficient is described as follows: before being immersed in a solution containing chloride ions, the UHPC sample is separated into two sub-samples, one for the test sample and the other for determining the initial chloride-ion content (Ci ). The initial chloride-ion content sample was cruised, and then the initial chloride-ion content in the acid-soluble fluid is specified. All surfaces of the test sample, except the finished surface, were sealed with a suitable coating. The sealed sample was then soaked to saturation in a calcium hydroxide solution with a concentration of 165 g per liter at 23 °C for 60 days, then rinsed with tap water and placed in sodium chloride solution. After a defined exposure time, i.e., 35 days, the test sample was removed from the sodium chloride solution and cut into thin layers parallel to the contact surface of the sample. The acid-soluble chloride content of each sample layer is determined. Finally, the apparent chloride diffusion coefficient (Da ) and the expected surface chloride-ion concentration (Cs ) were calculated based on the initial chloride ion content. At least six values of chloride-ion content and related depth from the exposed surface are determined.

Effect of Carbon Nanotubes on the Chloride Penetration

73

2.3 Concrete Mix Proportions In this study, the influence of CNTs on some mechanical properties and chloride ion diffusion coefficients of UHPC was determined. The ratio of sand to the binder (S/B) was fixed at 1/1 by weight. The superplasticizer dosage was adjusted to keep the flow of UHPC mixtures from 200 to 250 mm; the ratio of water to binder (W/B) was fixed at 0.16. The concrete mixture proportion is given in Table 3. Table 3. Mixture proportions of UHPC No W/B S/B SF, % FA, % CNTs, SP, % % 1

0.16

1

10

20

0

0.52

2

0.16

1

10

20

0.05

0.51

3

0.16

1

10

20

0.1

0.53

4

0.16

1

10

20

0.3

0.64

5

0.16

1

10

20

0.5

0.85

2.4 Mixing Procedure, Curing Conditions In this study, concrete mixtures were mixed in a 20 liter-Hobart mixer, and the mixing procedure is provided in Fig. 3.

Sand + Cement + SF + FA

3-5 min.

Dry mixture + 70% water

5-7 min.

SP+CNTs + 30% water

5-7 min.

UHPC mixture

Fig. 3. Mixing procedure of UHPC mixtures

The test samples were cured under the standard curing condition (27 ± 2 °C, RH ≥ 98%), and then de-molded after a cast of 24 h. After that, the samples were continuously cured under two different curing conditions: – Standard curing condition: (27 ± 2 °C, RH ≥ 98%) until defined testing ages; – Heat curing condition: in hot water (90 ± 5 °C) for 48 h, followed by standard curing conditions until defined testing ages. The samples were tested for compressive strength at the ages of the 3rd , 7th , and 28th days.

74

P. M. Hao et al.

3 Experimental Results and Discussion 3.1 Flow of Freshly Mixed UHPC Experimental results on the effect of the CNTs’ content on the workability of fresh UHPC mixtures are shown in Fig. 4, in which the flow evaluated the practicability indicated in Sect. 2.2. It can be seen that when the amount of CNTs increases up to 0.1%, the flow of the UHPC mixture is similar to that of the reference mixture without CNTs. However, when increasing the CNTs’ content of over 0.1%, the flow of the concrete mixture is decreased, e.g., with the CNTs content of 0.3%, the dosage of SP is increased by 23% compared to that of the reference sample. It is clear that when the CNTs’ content is increased to 0.5%, the SP dosage is increased significantly, up to 73% compared to that of the reference sample to attain the flow ranging from 200 to 250 mm.

SP, %

Flowability

1

240

0.8

230

0.6

220

0.4

210

0.2

200

Superplastisizer dosage, wt. % by binder

Flowability, mm

250

0 0 0.05 0.1 0.3 0.5 CNTs concentration, wt. % by binder

Fig. 4. Relationship between the CNTs content with the SP dosage and the flow of UHPC mixtures

CNTs can explain the effect of the CNTs content on the flow of fresh UHPC mixtures with a fibrous type, but in nanostructures, when added in concrete with a reasonable content, CNTs will fill in the voids between cement, SF and FA particles releasing a certain amount of water, as a consequence, the amount of free water may be increased, and this will improve the flow of the concrete mixture. However, as the content of CNTs is increased, the amount of water needed to wet the nanoparticles increases, and the dispersion of nanoparticles in the mixture will be more difficult, easy to create agglomerations, and forming more massive particles. Therefore, a certain amount of water is retained inside these agglomerations, thereby reducing the flow of the concrete mixture. 3.2 Compressive Strength The influence of the CNTs’ content on the compressive strength of UHPC samples is shown in Fig. 5. The experimental results show that the addition of CNTs does not

Effect of Carbon Nanotubes on the Chloride Penetration

75

improve the compressive strength of samples under both standard curing and heat curing conditions. It should be noted that the curing state influences the development of the compressive strength of specimens. The compressive strength of samples is enhanced significantly at the age of 7 days under the heat curing condition and much higher than that of samples curing under the standard condition. However, the compressive strength of samples at 28 days is not shown a significant difference for both curing conditions.

(a)

160

Rn7

180

Rn28 Compressive strength, MPa

Compressive strength, MPa

180 140 120 100

80 60 40 20

(b)

160

Rn7

Rn28

140 120 100

80 60 40 20 0

0

0

0.05 0.1 0.3 CNTs content, wt. % of binder

0.5

0

0.05 0.1 0.3 CNTs content, wt. % of binder

0.5

Fig. 5. Effect of CNTs content on compressive strength of UHPC under (a) standard curing (27 ± 2 °C), (b) heat curing (90 ± 5 °C)

3.3 The Microstructure of UHPC Using CNTs UHPC, with a meager W/B ratio, using large amounts of cement and mineral admixtures, applying heat curing condition, will fully promote the pozzolanic reaction and leading to minimize pore size, very low porosity and CH content [19–21]. In this study, the characterization of the microstructure of UHPC will be evaluated in the hardened cement paste and the interfacial transition zone (ITZ). It can be observed from SEM images that for the reference samples (Figs. 6 and 7), the hardened cement paste area exhibits a dense structure and the ITZ with a vast improvement compared to normal and high strength concrete, i.e., the width of the ITZ of about 40–500 nm. However, when adding CNTs, especially with heat curing conditions, it is possible to observe that the microstructure of concrete is improved both in the hardened cement paste and the ITZ to make a denser microstructure. As a consequence, the compressive strength of UHPC is enhanced. UHPC with a meager W/B ratio shows a very dense microstructure, and a resulting very low porosity when compared to normal and high strength concrete. Therefore, the space for the development of hydration products is limited, leading to a denser C-S-H and CH structure. The compact C-S-H structure and low CH content significantly improve the ITZ between aggregates and cement. In UHPC, the thickness of the ITZ between cement and sand is minimal compared to that of standard concrete. This is because the elimination of large aggregate particles results in a decrease in the natural retaining wall effect that appears around the surface of large aggregates. The use of CNTs has been shown to solidify both of the hardened cement paste and the ITZ of UHPC, especially

76

P. M. Hao et al.

Sand ITZ

Matrix

(a)

(b)

Fig. 6. SEM images of microstructure of UHPC reference sample without CNTs with a magnification of (a) ×50, and (b) ×50.000

UHPC matrix

Sand

CNTs UHPC matrix

(a)

Sand

(b)

Fig. 7. SEM images of microstructure of UHPC using CNTs with a magnification of (a) ×2.000, and (b) ×50.000

under the heat curing condition. The thickness of this ITZ of UHPC is observed from 40 to 500 nm, while for standard concrete and high strength concrete, this thickness is about 20–50 µm [22], 100 times and 40 times higher than that of UHPC with CNTs, respectively. 3.4 Chloride ion Concentration and Chloride ion Diffusion in Concrete The diffusion of chloride ion in concrete occurs within the void spaces, fluid-filled pores, and cracks. Also, the mixture composition, curing, finishing, the environment, age, and artistry influences the resistance to chloride penetration of concrete. Besides, the rate of chloride diffusion is also influenced by the valence and concentration of other chloride ions in the pore fluid. Therefore, the apparent diffusion coefficient as determined by the same test procedure. According to Fick’s second law of diffusion, this diffusion coefficient is used to evaluate chloride penetration into mortar or cement concrete under saturated conditions.

Effect of Carbon Nanotubes on the Chloride Penetration

77

In this study, the chloride ion concentration in UHPC samples at different depths and with different content of CNTs are determined and shown in Fig. 8. It should be noted that the reference sample without CNTs before exposure to chloride ion has a chloride ion concentration of 0.0532%, and this value is considered as the initial chloride ion concentration of all samples. The experimental results show that for the reference sample (0% CNTs), when exposure to chloride ion, the largest concentration of chloride ion of 0.0859% can be reached with a depth of 1 mm. However, at a depth of 2 mm, the value of chloride ion concentration shows a massive reduction of about 38.4%. At a depth of 4 mm, no chloride ion penetration can be observed.

Chloride content, %

0.10

0%CNTs 0.1%CNTs 0.5%CNTs

0.08

0.05%CNTs 0.3%CNTs

0.06 0.04 0.02 0.00 Ref

1

2

4 Depth, mm

6

8

Fig. 8. The chloride content in UHPC at different depths

The addition of CNTs significantly reduces the concentration of chloride ion in UHPC samples. It can be observed that with the CNTs content of 0.05%, the chloride ion concentration in the UHPC samples is decreased at different depths, but not significant compared with the samples using 0% CNTs. However, when adding 0.1% CNTs, the concentration of chloride ions in the UHPC sample at the depths from 1 mm to 2 mm is decreased sharply compared to the samples using 0% CNTs, for example, at a depth of 1 mm, the reduction of chloride ion concentration can be reached up to 35%. As the content of CNTs increases to 0.3% and 0.5%, the chloride ion concentration in the UHPC sample is decreased significantly compared to the sample using 0.1% CNTs. It should be noted that no increase in chloride ion concentration can be seen with different CNTs’ contents at a depth of the 4 mm. These results demonstrate that the dense microstructure of UHPC with CNTs gives perfect prevention of chloride ion penetration. In this study, it is almost impossible to see the penetration of chloride into the concrete at a depth of 2 mm. According to ASTM C1556, a regression analysis of a function expressed as chloride concentration was measured at different depths for 60 days, and the apparent diffusion coefficient of chloride (Da) was determined. Through the analysis of experimental results, the chloride ion diffusion coefficient was identified (Fig. 9). The results show

78

P. M. Hao et al.

that when adding CNTs in concrete, the chloride ion diffusion coefficient is significantly reduced compared to the sample using 0% CNTs. However, when CNTs content increased above 0.3%, chloride ion diffusion coefficient tended to increase slightly. This value indicates the need for further study of chloride ion diffusion coefficients in UHPC at later ages.

Diffusion coefficient, 10-14 m2/s

5 4 3 2 1 0 0

0.05 0.1 0.2 0.3 0.4 CNTs content, % by weight of binder

0.5

Fig. 9. The apparent chloride diffusion coefficient of UHPC with different CNTs contents

The addition of CNTs above 0.3% increasing the Da coefficient can be explained that CNTs with a fibrous type but in nanostructures, when added in concrete with a reasonable content, it will be filled in the voids in the matrix and ITZ region, improving the dense structure of concrete. However, when more CNTs are added, the dispersion of nanoparticles in the mixture will be more difficult, the CNTs will easily be agglomerated and form larger particles, which affects the consistency of concrete.

4 Conclusions Based on the experimental results of the effect of CNTs concentration on some mechanical properties and the calculated apparent chloride diffusion coefficient of UHPC, some conclusions can be drawn as follows: – The addition of CNTs influences the flow of the UHPC mixture when the CNTs content is less than 0.1% by weight of the binder, the flow of the UHPC mixture is similar to that of the reference sample, but above this content, the flow of the UHPC mixture is decreased. – The addition of CNTs does not significantly improve the compressive strength of UHPC under both standard curing and heat curing conditions. When the CNTs content is higher than 0.3% by weight of the binder, the compressive strength of concrete tends to decrease. – The addition of CNTs improves both in the hardened cement paste and the ITZ of UHPC. The thickness of ITZ of UHPC observed from SEM images is from 40 to 500 nm, i.e., much improved in comparison with that of standard concrete and high strength concrete, the ITZ of those are about 20–50 µm.

Effect of Carbon Nanotubes on the Chloride Penetration

79

– The addition of CNTs reduces the chloride ion penetration in UHPC. The chloride ion concentration in the UHPC sample using 0.1% CNTs is decreased sharply compared to the sample using 0% CNTs at the depth from 1 mm to 2 mm, i.e., up to 35% at 1 mm depth. No chloride ion penetration can be observed at a depth of 4 mm.

Acknowledgments. The authors would like to thank the Ministry of Science and Technology, Vietnam, for providing financial support to implement the project coded ÐTÐLCN.37/18, and also thanks to the National University of Civil Engineering (NUCE) for supporting laboratory tests.

References 1. Richard, P., Cheyrezy, M.H.: Reactive powder concretes with high ductility and 200–800 MPa compressive strength. In: Mehta, P.K. (eds.) Concrete Technology: Past, Present and Future, Proceedings of the V. Mohan Malhotra Symposium (1994) 2. AFGC-SETRA: Ultra High Performance Fibre-Reinforced Concretes, Paris, France: Interim Recmmendations, AFGC Publication (2020) 3. Schmidt, M., Fehling, E., Geisenhanslüke, C.: Ultra high performance concrete (UHPC). In: International Symposium on Ultra High Performance Concrete, Kassel, Germany (2004) 4. Nematollahi, B., Voo, Y.L.: A review on ultra high performance ‘ductile’ concrete (UHPdC) technology. Int. J. Civil Struct. Eng. 2(3), 1003–1018 (2012) 5. Buitelaar, P.: Ultra high performance concrete: developments and applications during 25 years. In: International Symposium on UHPC, Kassel, Germany (2004) 6. Schmidt, M., Fehling, E.: Ultra-high-performance concrete: research, development and application in Europe. In: Seventh International Symposium on the Utilization of HighStrength/High-Performance Concrete, Washington, D.C., USA, SP-228-4 (2005) 7. Spasojevic, A.: Structural implications of ultra-high performance fibre-reinforced concrete in bridge design. Thèse de doctorat N4051, Ecole Polytechnique Fédérale de Lausanne, Suisse (2008) 8. Neville, A.M.: Properties of concrete. Fourth Edition ed.: PEARSON (2002) 9. Nguyen, V.T., et al.: Ultra-High Performance Concrete- Fundamentals, Experimental results, Applications, Construction Publishing House, 300 p. (2017). ISBN 978-604-82-2288-8 (in Vietnamese) 10. Tuutti, K., Corrosion of Steel in Concrete, Swedish Cement Concrete Research Institute, Fo 4, Stockholm (1982) 11. Malhotra, V.M., Carette G.G.: Durability of Concrete Containing Supplementary Cementing Materials in Marine Environment, Concrete Durability, Katharine and Bryant Mather International Conference, SP-100, J.M. Scanlon ed., American Concrete Institute, Farmington, Hills, Mich. (1987) 12. Roux, N., Andrade, C., Sanjuan, M.A.: Experimental study of durability of reactive powder concretes. ASCE J. Mater. Civil Eng. 8(1), 1–6 (1996) 13. Scheydt, J.C., Müller, H.S.: Microstructure of ultra high performance concrete and its impact on durability. In: International Symposium on Ultra High Performance Concrete and Nanotechnology for High Performance Construction Materials, Kassel, Germany (2012) 14. Thomas, M.: Marine performance of UHPC at treat Island. In: International Symposium on UHPC and Nanotechnology for High Performance Construction Materials, Kassel, Germany (2012)

80

P. M. Hao et al.

15. Li, G.Y., Wang, P.M., Zhao, X.: Mechanical behavior and microstructure of cement composites incorporating surface-treated multi-walled carbon nanotubes. Carbon 43(6), 1239–1245 (2005) 16. Loh, K.J., et al.: Multifunctional layer-by-layer carbon nanotube-polyelectrolyte thin films for strain and corrosion sensing. Smart Mater. Struct. 16(2), 429–438 (2007) 17. Kowald, T.: Influence of surface-modified carbon nanotubes on ultra-high performance concrete. In: International Symposium on Ultra High Performance Concrete, Kassel, Germany (2004) 18. Kowald, T., et al.: Influence of carbon nanotubes on the micromechanical properties of a model system for ultra-high performance concrete. In: 2nd International Symposium on Ultra High Performance Concrete, Kassel, Germany (2008) 19. Nguyen, V.T., Ye, G., van Breugel, K.: Internal curing of ultra high performance concrete by using rice husk ash. In: Rilem (eds.) Proceedings of the International Conference on Material Science and 64th RILEM Annual Week, Aachen, Germany, 6–10 September, vol. III, pp. 265–274 (2010) 20. Nguyen, V.T., Ye, G., van Breugel, K.: Mitigation of early age shrinkage of ultra high performance concrete by using rice husk ash. In: Proceedings of Hipermat, 3rd International Symposium on Ultra High Performance Concrete and Nanotechnology for High Performance Construction Materials, Session 3A.7, pp. 341–348 (2012) 21. Pham, S.D., Nguyen, V.T., Le, T.T., Nguyen, C.T.: Possibility of using high volume fly ash to produce low cement ultra high performance concrete. In: Proceedings of the International Conference on Sustainable Civil Engineering and Architecture (ICSCEA 2019). LNCE, vol. 80, pp. 589–597. Springer (2019) 22. Le, T.T., et al.: Mineral admixtures for cement and concrete. Construction Publishing House, 278 p. (2019). ISBN 978-604-82-2851-4 (in Vietnamese)

Estimating the Radial Displacement on the Tunnel Boundary Within Efficient Working Area of Rock Tunneling Quality Index (Q-System) Van Diep Dinh1 , Ngoc Anh Do1(B) , Amund Bruland2 , and Daniel Dias3,4 1 Hanoi University of Mining and Geology, Hanoi, Vietnam

[email protected] 2 Norwegian University of Science and Technology, Trondheim, Norway 3 School of Automotive and Transportation Engineering, Hefei University

of Technology, Hefei, China 4 Laboratory 3SR, Grenoble Alpes University, Grenoble, France

Abstract. Rock Tunneling Quality Index (Q-system) was developed based on a high number of practical engineering cases in tunneling. Q-system illustrated instructions of how to design tunnel support structure using rock bolts and shotcrete. However, these instructions were obtained on the basis of observations of constructed tunnel. In addition, the stability of supported tunnels designed based on the Q-system has not been quantified and clarified in the literature. The plenty of research pointed out that only a part of the Q-system worked most efficiently, the other parts demanded to aid some supplement methods to determine the appropriate rock support parameters. In this paper, twenty-eight cases of supported tunnels within the effective area of Q-system were surveyed by using RS2 software (Rocscience). The goal is to estimate the strength of rock mass in the efficient working area of Q-system based on the Radial Displacement (RD) value of tunnel boundary obtained from numerical models. The performance research would significantly contribute to Q-system application in tunneling, especially in predicting the stability of a tunnel. Keywords: Q-system · Tunneling · Rock support · Stability · Radial displacement

1 Introduction The classification of the Rock Tunnelling Index (Q-system) developed by Barton et al. [1] is a useful tool for estimating rock mass quality and required support in tunnel and rock caverns. Over the last 10–15 years, a high number of papers have been issued to expand the application of Q-system. It is undeniable the benefit of the Q-system, but it has still existed restriction as rock mass classification systems. Plamstrom et al. [2] pointed out the limitation of Q-system that there was an efficient working area on the whole © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 81–90, 2021. https://doi.org/10.1007/978-3-030-60839-2_5

82

V. D. Dinh et al.

Q-system varied from 0.1 to 40 in Q value, corresponding to the (Span or Height)/ESR ratio changed from 2.5 to 35. This range represents the highlighted rectangular area plotted at the center of Fig. 1 [3]. Besides, the stability of supported tunnels estimated based on Q-system, especially in the efficient working area of Q-system mentioned above, has not been considered adequately. Barton et al. [4] conducted a back-calculate method to estimate the average deformation modulus of rock mass classified by Q-system in vast caverns. By using the finite element method, the results gave the relationship between Q value and the average deformation modulus to predict the deformation of tunnels. Furthermore, the deformations of the arch, wall, and the invert of tunnels were determined by the Q/Span or Q/Height ratio. It was pointed out that the dimensions of excavations will also govern the magnitude of deformations. It is worth noting that these results were obtained from well-instrumented underground power stations. Unfortunately, these data did not include the deformation of small tunnels excavated in a rock mass with high Q values. In the research by Bieniawski [5], the number of in-situ deformation measurements was analyzed by classified rock mass following Q-system. The deformation modulus of an in-situ rock mass was utilized as a useful guide based on Q-system. In other words, the deformation modulus was chosen according to the relationship with Q value. This modulus required the statistical research related to the distribution of stress and displacement surrounding tunnels. However, these results just illustrated the deformations modulus without mentioning the stability of tunnels in the efficient working area of the Q-system. This study aims to estimate the strength of rock mass in the efficient working area of Q-system based on the Radial Displacement (RD) value of tunnel boundary obtained from numerical models. The performance research would significantly contribute to Q-system application in tunneling, especially in predicting the stability of a tunnel.

2 The Efficient Working Area of Q-system Q-system was constituted by the plenty of data that was collected from tunnels in Norway and other countries. Based on Q-system, the parameters of rock support as bolts and shotcrete were determined by rock mass quality in terms of Q value and (Span or Height)/(Excavation Support Ratio, ESR), (Equivalent Dimension, De). Palmstrom and Broch [3] conducted a survey about Q-system elaborately and showed that the Qsystem worked best within a specific range of parameters. This range was illustrated by a rectangular in Fig. 1. The best working area of Q-system fluctuated between 0.1 and 40 in Q value corresponding to the (Span or Height)/ESR ratio varied from 2.5 to 35. If the data was outside this area, it was necessary to use other supplementary calculated methods. Those methods will enhance the reliability for determining the proper rock supports in tunneling. One of the requirements of rock support is to ensure the stability of the tunnel. Although rock support was framed followed instructions by empirical methods of Qsystem, the weakness of Q-system is the fact that the stability of the tunnel was not quantified. One of the specific behaviors of tunnel stability is the displacement of the tunnel boundary. Typically, the removal of supported tunnel boundary reflects the instability of rock mass surrounding the tunnel. In reality, displacement is usually determined

Estimating the Radial Displacement

83

Fig. 1. Limitation of Q-system for rock support. Outside this area, supplementary methods/evaluations/calculations should be applied (reproduced from Palmstrom and Broch [3])

by the convergence measurement method of by extensometer installed in the rock mass. Unfortunately, the Radial Displacement (RD) degree on the tunnel boundary within the efficient working area of the Q-system has not been taken into account adequately. To solve the problem, the author conducted a numerical investigation using RS2 software [6] to determine the stability of tunnel in terms of radial displacement (RD) measured at three points on the tunnel boundary, which are (1) crown of the tunnel; (2) top of tunnel wall and (3) tunnel floor (see Fig. 3).

3 Cases Study and Numerical Model Based on the efficient working area of Q-system proposed by Palmstrom and Broch [3], adopted cases of rock mass quality and a support structure used in this study have been selected on the mutual boundary between categories as seen in Fig. 2. Each case in Fig. 2 was determined by two parameters of Q value and (Span or Height)/(Excavation Support Ratio, ESR). There were a total of 28 numerical calculations conducted in this study. The parameters of all cases are indicated in Table 1. The parameters of tunnel support were selected based on instruction by Palmstrom and Broch [3] (see Fig. 1). In reality, tunneling is a complicated three-dimensional (3D) issue depending on the advance of the tunnel face. However, the length of the tunnel in this study is much larger than the dimensions in the cross-section. For the sake of simplicity, a two-dimensional (2D) model could, therefore, be used instead of a 3D model [7].

84

V. D. Dinh et al.

Fig. 2. Adopted case studies

The 2D model adopted in this study has dimensions of 160 m in both the height and width. These dimensions of the 2D model were selected based on a parametric analysis to eliminate the effect of the boundary condition on the numerical calculation results. The surveyed tunnels in this numerical model are in D shape (see Fig. 2) in which the height of tunnel (H) equals to the width (B). The ESR value was set as 1 (categories D) for power stations, major road, and railway tunnel, civil defense chambers, porta intersections according to suggestions of Barton et al. [1]. The numerical model was discretized and meshed into finite elements. Finite elements in the model were formed as triangles with six nodes since the model size was large enough to eliminate the effect of model size on the stress and displacement in the rock mass surrounding the tunnel. The external boundary of the model was restricted in X and Y directions, respectively (Fig. 3). The effect of gravity on the initial stress-induced in rock mass was taken into consideration. In this study, it was assumed that the depth of tunnel is 100 m, rock’s unit weight (γ) equals 0.026 MN/m3 , and the lateral earth pressure (Ko ) was set as 0.5 for all investigated cases.

Estimating the Radial Displacement

85

Table 1. Parameters of the case study Case

Q value

GSI

B/ESR

Bolt spacing (m)

Bolt length (m)

Thickness of shotcrete (m)

1 2

0.10

30

25.0

1.3

5.7

0.246

0.30

37

35.0

1.4

7.4

0.237

3

0.10

30

10.0

1.3

3.0

0.147

4

0.30

37

20.0

1.4

5.0

0.150

5

0.55

41

35.0

1.6

7.4

0.168

6

0.10

30

4.5

1.3

2.3

0.115

7

0.40

39

14.0

1.5

3.8

0.124

8

1.00

45

25.0

1.7

5.7

0.122

9

2.00

50

35.0

1.8

7.4

0.122

10

0.10

30

2.5

1.3

1.7

0.091

11

0.40

39

5.0

1.5

2.4

0.087

12

1.00

45

10.0

1.7

3.0

0.090

13

3.00

52

20.0

2.0

5.0

0.090

14

6.00

57

35.0

2.7

7.4

0.088

15

0.50

40

2.5

1.5

1.7

0.052

16

1.00

45

4.0

1.7

2.1

0.049

17

3.50

53

10.0

2.0

3.0

0.052

18

6.00

57

16.0

2.2

4.2

0.050

19

10.00

60

25.0

2.3

5.7

0.055

20

17.00

63

35.0

2.3

7.4

0.066

21

1.00

45

2.5

1.7

1.7

0.047

22

4.00

54

4.0

1.6

2.1

0.000

23

10.00

60

10.0

2.3

3.0

0.040

24

30.00

67

20.0

2.4

6.5

0.049

25

40.00

69

25.0

2.5

5.7

0.040

26

10.00

60

5.0

2.0

2.4

0.000

27

40.00

69

19.0

2.5

4.8

0.000

28

40.00

69

9.5

2.5

2.9

0.000

Fig. 3. The vertical displacement determined the layout of the numerical model and monitored points (1), (2) and (3) at point (1), (3) and horizontal displacement at point (2)

86

V. D. Dinh et al.

4 Evaluation of Rock Mass and Rock Support Parameters The constitutive model using the Hoek-Brown failure criterion has been adopted for the rock mass surrounding the tunnel [8]. The deformation modulus of intact rock Ei was evaluated as follows [9]: Ei = MR.σci

(1)

Where: MR - Modulus ratio, MR = 500; σci - Uniaxial compressive strength, σci = 50 MPa. The deformation modulus of the rock mass (Erm ) was calculated based on the following relationship:   1 − D/2 Erm = Ei 0.02 + (2) 1 + e((60+15D−GSI)/11) Where: D - Disturbance factor, assumed D = 0; GSI - Geological Strength Index The reduced value of material constant (mb ) was calculated based on the Hoek Brown failure criterion [8]:   GSI − 100 (3) mb = mi .exp 28 − 14D Where: mi - Material constant. It should be noted that Fig. 2 just adopt Q value. It is, therefore, necessary to estimate the GSI value based on the corresponding Q value. Bieniawski [10] introduced a relationship between the Q value and Rock Mass Rating (RMR) value of rock mass as follows: GSI = RMR89 − 5

(4)

Also, the relationship between Q value and RMR value was determined by a logarithmic function as following [11]: RMR = 15 log Q + 50

(5)

Therefore, GSI value can be calculated as: GSI = 15 log Q + 45

(6)

Bolts and shotcrete were used as rock support for tunnels applied Q-system. The parameters of rock support in Q-system include bolt spacing, bolt length, the thickness of the shotcrete. These parameters were determined according to studied cases in Fig. 2. Moreover, the other parameters of bolts and shotcrete used in models were illustrated in Table 2.

Estimating the Radial Displacement

87

Table 2. Rock support parameters Properties

Unit Value

Fully Bonded Bolts Bolt Diameter

mm

Bolt Modulus (E)

MPa 200,000

20

Tensile Capacity

MN 0.5

Residual Tensile Capacity MN 0.5 Pre-Tensioning Force

MN 60

Shotcrete Young’s Modulus

MPa 45,000

Poisson’s Ratio

–

0.25

Fig. 4. Vertical displacement at Point 1

5 Results and Discussion Twenty-eight cases have been numerically investigated by using RS2 software [6]. The Radial Displacement (RD) values determined at 3 points on the boundary including, (1) top of crown, (2) top of the wall, and (3) middle of the floor were illustrated in Figs. 4, 5 and 6, respectively. The results of RD at points (1), (2) and (3) in the efficient working area of Q-system presented that the RD values have a descending trend gradually when Q value is ascending and (Span or Height)/(Excavation Support Ratio, ESR) is decreasing simultaneously.

88

V. D. Dinh et al.

Fig. 5. Vertical displacement at Point 2

Fig. 6. Horizontal Displacement at Point 3

In other words, the tunnels supported by bolts and shotcrete following suggestions of Q-system is more stability when rock mass quality increase and span of tunnels decrease.

Estimating the Radial Displacement

89

As an example, Fig. 4 showed that the RD at point 1 (vertical displacement) when Q value is 0.1 and B/ESR is 25, equals 4.4 cm, whereas if Q value is 0.3 and B/ESR is 20, the RD drops to 3 cm. The results in Fig. 6 pointed out that the RD values at point 3 are higher than those at other locations. When the Q value was 0.3, and (Span or Height)/ESR ratio was 20, the RD at point 3 (the middle of the floor) was 7.2 cm, whereas the RD at positions 1 and 2 were 3.0 cm and 1.4 cm, respectively. Figure 7 illustrated the effect of the Q/De (Equivalent Dimension) ratio on RD induced at point 1, point 2, and point 3 on the tunnel boundary. Generally, there was a downward trend in the RD at three different positions when the Q/De ratio increased following exponential functions. The dependency degree of RD at point 3 had a substantial higher in comparison with RD at other locations (i.e., points 1 and 2).

Fig. 7. Radial displacement at Point 1, Point 2 and Point 3

It could be demonstrated that any structure did not support the floor of the tunnel so that the rock mass could move into tunnel space freely without any restriction. Whereas at points 1 and 2, the rock mass also had the radial movement, but the displacement magnitude is smaller than at point 3 owing to the support of the bolts and shotcrete layer installed on the tunnel boundary.

6 Conclusions In this paper, a numerical investigation has been conducted to estimate the Radial Displacement (RD) induced on the tunnel boundary within the efficient working area of the Rock Tunneling Quality Index (Q-system). Some conclusions could be derived as follows:

90

V. D. Dinh et al.

– An instruction of estimating the stability of supported tunnel in terms of tunnel’s radial displacements based on Q-system applied for D shaped tunnel has been introduced which allow preliminarily predicting the behavior of tunnel at the design phase; – It is undeniable that Q-system exists an efficient working area that has more reliable due to a large number of practical measurements. Beyond this area of Q-system, it is required to use supplemental methods to estimate the proper rock supports in designing rock supports for tunneling; – The RD at point 3 is more dependent on the Q/De (Equivalent Dimension) ratio than that measured at other positions. Additional numerical calculation and in-situ measurement are necessary to be conducted to enlarge the estimate of the behavior of the supported tunnel using Q-system for other tunnels with different shapes and dimensions.

References 1. Barton, N., Lien, R., Lunde, J.: Engineering classification of rock masses for the Design of Tunnel support. Rock Mech. 6(4), 189–236 (1974) 2. Palmstrom, A., Blindheim, O.T., Broch, E.: The Q-system - possibilities and limitations (in Norwegian). In: Norwegian National Conference on Tunnelling, pp. 41.1–41.43. Norwegian Tunnelling Association, Norwegian (2002) 3. Palmstrom, A., Broch, E.: Use and misuse of Rock mass classification systems with particular reference to the Q-system. Tunnels Underground Space Technol. 575–593 (2006) 4. Barton, N., Loser, F., Lien, R., Lunde, J.: Application of Q-System in design decisions concerning dimensions and appropriate support for underground installations. In: ISRM International Symposium - Rockstore 80, 23–27 June, Stockholm, Sweden. ISRM-Rockstore-1980-073 (1981) 5. Bieniawski, Z.T.: Estimating the strength of rock materials. J. South African Inst. Mining Metal. 74(8), 312–320 (1974) 6. Rocscience. “Software Manual” (2016). https://www.rocscience.com 7. Do, N.A., Dias, D., Oreste, P.P., Djeran-Maigre, I.: 2D Tunnel Numerical Investigation The Influence of the Simplified Excavation Method on Tunnel Behaviour. Geotechnical and Geological Engineering, pp. 43–58 (2014) 8. Hoek, E., Carranza-Torres, C., Corkum, B.: Hoek-Brown failure criterion - 2002 edition. In: Proceedings of the NARMS-TAC Conference, pp. 267–273. Toronto (2002) 9. Hoek, E., Diederichs, M.S.: Empirical estimation of rock mass modulus. Int. J. Rock Mech. Mining Sci. 43(2), 203–215 (2006) 10. Bieniawski, Z.: Engineering Rock Mass Classifications: A Complete Manual for Engineers and Geologists in Mining, Civil and Petroleum Engineering. A Wiley-Interscience publication, New York (1989) 11. Barton, N.: Some new Q-value correlations to assist in site characterisation and tunnel design. Int. J. Rock Mech. Mining Sci. 39(2), 185–216 (2002)

Two Novel Predictive Networks for Slope Stability Analysis Using a Combination of Genetic Programming and Artificial Neural Network Techniques Hossein Moayedi1,2(B) 1 Institute of Research and Development, Duy Tan University, Da Nang 550000, Vietnam

[email protected] 2 Faculty of Civil Engineering, Duy Tan University, Da Nang 550000, Vietnam

Abstract. The stability of slopes is an important parameter which can affect many engineering projects. In this study, we employed genetic programming (GP) and artificial neural network (ANN) techniques, based on upper bound (UB) limit analysis, for the problem in designing solution charts for slope stability. Existing theories of genetic programming predictive network models have not been applied in the area of slope stability. Accordingly, the main objective of this research is to propose a new GP model to estimate the factor of safety parameter and providing design solution charts in a two-layered cohesive slope. A dataset containing 400 UB analysis models was used to train and test the GP and ANN networks. Variables of the GP algorithm training network parameters and weights such as population size, number of genes, and tournament size were optimized. The input includes d/H, (depth factor), the undrained shear strength ratio (Cu1 /Cu2 ), and slope angle (β), where the output was taken as a dimensionless stability number (N 2c ). The predicted results for both datasets (training and testing) from the GP and ANN models were evaluated based on two statistical indexes (root mean square error, RMSE, and coefficient of determination, R2 ). Besides, the obtained results were compared with actual values of N 2c , in the form of design charts. The results show that both the GP and ANN models are accurate enough to be used in this field. Also, ANN performed slightly better than the GP. As a result, a formula was derived for each GP and ANN models to assess the slope stability behaviors of two-layered cohesive soils. Keywords: Artificial neural network · Genetic programming · Optimization algorithm · Slope stability

1 Introduction The slope stability problem has been studied for years, being of considerable concern due to its effects on geotechnical designs [1]. Generally, slopes are categorized as a cut slope, natural slope or fill slope, and physical and geometric factors influence their © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 91–108, 2021. https://doi.org/10.1007/978-3-030-60839-2_6

92

H. Moayedi

stability. Also, the parameters which are related to soil strength affect the slope behavior too. The safety factor (F s ) of a cohesive slope is related to a well-recognized parameter named the dimensionless stability number (N 2c ). One of the first studies that led to the development of useful design charts for the slope stability issue was implemented by Taylor [2]. More recently, various research studies have concentrated on providing design chart solutions (e.g., Qian et al. [3], Abuel-Naga et al. [4], Aksoy et al. [5] and Moayedi and Hayati [6]). Some time and cost constraints are constraints of using the new advances in mathematical and computer modeling tools. Therefore, the importance of providing the design charts is becoming more highlighted. As a novel idea, intelligent predictive tools (i.e., artificial neural networks (ANNs) and particle swarm optimization (PSO)) are being employed for the subject of slope stability evaluation [7, 8]. Jiang et al. [9] applied the UB method to investigate a rock slope’s seismic behavior and safety factor with a tunnel. As a result, they found three factors, namely the horizontal seismic force coefficient, the slope height, and the internal friction angle, to be the most influential parameters influencing the safety factor’s sensitivity. In another study, Pan and Dias [10] assessed the face stability for the case of a non-circular tunnel through an upper bound limit analysis model, synthesized with the technique of strength reduction. According to their results, this approach presents conservative results and can be applied in safety factor evaluation in a particular design. The number of studies concerned about the feasibility of neural network modeling to solve slope stability is limited. The hybrid GP model that is presented in this study has not been proposed for the problem of slope stability previously. There is almost no study on using hybrid GP-based learning systems to predict the factor of safety and its key parameters. In the following, the GP prediction models’ results were compared with the results of an optimized ANN feedforward learning system. Similarly, to find the best ANN structure, all the proposed models were evaluated with a trial and error process on their influential parameters. Finally, each model’s design chart solution was depicted; and the formula was presented for both the GP and ANN models to be used for slopes with the same conditions. In this paper, optimal forms of genetic programming (GP) and ANN models have been utilized to predict the N 2c in a two-layered cohesive slope, based on the upper bound (UB) limit analysis method. The UB model is broadly applied for similar engineering problems. The slope proposed for this work was constructed from two separate clayey layers, lying on a rigid rock bed. The ratio of the first soil layer thickness to the full height of the slope, d/H, the slope angle, β, and the ratio of the undrained shear strength of the first soil layer to that of the slope layer, C u1 /C u2 , were considered as three efficacious parameters that affect the slope situation. These parameters also form the input dataset. Meanwhile, the N 2c was taken as the output of the networks. 80% of the dataset was devoted to the training part, and the remaining 20% was selected to test the networks.

2 Artificial Intelligence Systems Genetic Programming. Genetic programming is a powerful computing technology that has been initiated based on genetic algorithms (GA). This model was first introduced in 1985 by Cramer [11]. Another scholar, Ref Koza, later improves it in 1984 [12]. It

Two Novel Predictive Networks for Slope Stability Analysis

93

is widely used in solving engineering problems by numerous scholars such as Johari et al. [13], Makkeasorn et al. [14], Garg et al. [15], and Garg et al. [16]. During the GP process, particular computer programs are applied to predict the problem’s potential outputs by linking the input data layers and main output(s). One of the critical issues in using the GP algorithm is that the user will have a simpler finalized algorithm. The overall procedure of GP prediction performance is illustrated in Fig. 1. The complete detail of the applied GP technique is well described in Ref. [12, 15–17].

Fig. 1. Typical GP tree presented for the function [(X1-5)/(X2 + X3)]2

Artificial Neural Network. The artificial neural networks (ANNs), which mainly inspired by a biological neural network, are well established as one of the most applicable approaches that have been employed in the last two decades is, which were first introduced by McCulloch and Pitts [18]. These tools are widely employed for prediction purposes [19]. Numerous scholars such as Rao [20], El-Bakry [21], Samui and Kumar [7], and Moayedi and Hayati [22] are experiences successful use of ANN in predicting complex engineering solutions. Overall, the primary objective is to establish a non-linear equation (i.e., trained based on the initial learning process) between the inputs and output(s) dataset [20]. In this sense, a neural network’s typical architecture is prepared according to components of so-called neurons. As illustrated in Fig. 2, the input layer includes layers of the input(s) data. As a predefined model, the number of nodes is considered the same as the input parameters. During the neural network training processes, there can be one or more hidden layer(s) finally, and the calculation process will end to one or more output layer(s). More precisely, for each node, if we assume the term X as the main input and the term W as the interconnected weight (which is also shown as I w ), the bias (which is shown by the term of β) will be added to the summation ofWXs. In this regard, an activation function (f (I)) will be applied to the acquired term ( WX + β) to produce the outputs. The activation function for this case was considered as Tan-sigmoid (Tansig) which is defined by Eq. (1):

Tansig(x) =

2 −1 1 + e−2x

(1)

94

H. Moayedi

Fig. 2. Typical structure and operation of ANNs

3 Data Collection and Problem Statement In this research, the results from the upper bound (UB) type of limit analysis method were employed to assess the short-term stability situation of a cohesive slope. This method is well discussed in other studies (e.g., Florkiewicz [23], Donald and Chen [24], Karkanaki et al. [25], and Jiang et al. [9]). The slope is constructed from a maximum of two different soil layers with separate material properties lying on a bedrock layer in the present case. Besides, Fig. 3 shows a graphical description of the input data range versus the data numbers for d/H, the slope angle of β° and the C u1 /C u2 ratio. The analysis is performed with new computer software called OptumG2. It is based on finite element limit analysis and has been widely used in other studies (e.g., Karkanaki et al. [25]; Caër et al. [26] and Zhou et al. [27]). The analytical method used and the 2D boundary conditions applied in this study, i.e., to illustrate the pure cohesive slope, are similar to the research performed by Qian et al. [3]. A view of the slope model is presented in Fig. 4. As can be obtained from this figure, the proposed slope has been formed from two cohesive soil layers having only consistent undrained strength (cu1 ) with a zero undrained internal friction angle. Cu1 and Cu2, respectively define the undrained cohesive strength for the top and bottom soil layers. The influential parameters and an example of output from OptumG2 are illustrated in Figs. 4a and 4b, respectively.

4 Model Development for Prediction of N 2 c A proper estimation process, which is used by hybrid ANN models, should be formed from several steps such as (i) data processing and normalization, (ii) selecting a suitable hybrid model, and finally (iii) finding an appropriate hybrid structure for the proposed model, which can be achieved through a trial and error procedure. For the aim of producing the design chart solutions, N 2c (a dimensionless stability number, which was first investigated by Taylor [2]) was obtained from Eq. (2): N2c = cu1 /γ HFs

(2)

Two Novel Predictive Networks for Slope Stability Analysis

95

80

6

70

5

60 50

β

d/H

4 3

40 30

2

20 1

10 0

0 0

100

200

300

400

0

100

200

300

400

dataset number

dataset number (a)

(b)

6 5

Cu1/Cu2

4 3 2 1 0 0

100

200

300

400

dataset number (c )

Fig. 3. Graphical description of the range of input data versus data numbers for (a) d/H, (b) slope angle, (c) C u1 /C u2

(a)

(b)

Fig. 4. A view of the model for cohesive slope (a) schematic model, (b) OptumG2 stability output for upper bound limit analysis

where N 2c , Cu1 , and Fs define the dimensionless stability number, the undrained shear strength of the second soil layer, and the factor of safety obtained from the OptumG2 modeling, respectively. Also, γ describes the soil unit weight, which is taken as 20 kN/m3 . The dataset used is constructed from three inputs: the three influential parameters affecting the slope stability situation. The first factor is related to depth (d/H). It defines the ratio of both soil layer thicknesses, d, to the slope’s top layer height, H. The angle

96

H. Moayedi

of the present slope (β) was considered the second parameter. The third effective factor defines the undrained shear strength ratio (C u1 /C u2 ). An example of the mentioned data is presented in Table 1. Note that these values are derived from the OptumG2 simulation, and, as illustrated, the three useful parameters of d/H, β, and C u1 /C u2 are employed to estimate the N 2c in the three FEM methods of LB, UB, and LEM limit analysis. Table 1. Example of inputs and output dataset applied for modeling purpose Model number

Inputs

Output

Model number

Inputs

Output

d/H

β°

Cu1 /Cu2

N2c -UB limit analysis

d/H

β°

Cu1 /Cu2

N2c -UB limit analysis

1

5

75

5

0.7779

26

3

75

4.5

0.6295

2

5

60

5

0.7663

27

5

45

4

0.6276

3

5

45

5

0.7605

28

5

30

4

0.6227

4

5

30

5

0.7547

29

4

15

4.5

0.6179

5

4

75

5

0.7450

30

4

75

4

0.6140

6

4

60

5

0.7275

31

3

60

4.5

0.6101

7

5

15

5

0.7246

32

4

60

4

0.6053

8

4

45

5

0.7217

33

3

45

4.5

0.6024

9

4

30

5

0.7139

34

5

15

4

0.6004

10

5

75

4.5

0.7091

35

4

45

4

0.5995

11

5

60

4.5

0.6994

36

2

75

5

0.5966

12

5

45

4.5

0.6945

37

4

30

4

0.5936

13

3

75

5

0.6936

38

3

30

4.5

0.5898

14

5

30

4.5

0.6887

39

3

15

5

0.5839

15

4

75

4.5

0.6780

40

5

75

3.5

0.5733

16

4

15

5

0.6722

41

3

75

4

0.5684

17

4

60

4.5

0.6664

42

5

60

3.5

0.5645

18

5

15

4.5

0.6635

43

4

15

4

0.5626

19

4

45

4.5

0.6615

44

5

45

3.5

0.5597

20

3

60

5

0.6615

45

3

60

4

0.5578

21

4

30

4.5

0.6548

46

5

30

3.5

0.5548

22

3

45

5

0.6518

47

3

45

4

0.5510

23

5

75

4

0.6412

48

4

75

3.5

0.5510

24

3

30

5

0.6383

49

2

60

5

0.5451

25

5

60

4

0.6324

50

2

75

4.5

0.5432

Two Novel Predictive Networks for Slope Stability Analysis

97

5 Results and Discussion The present study intends to evaluate the stability of a two-layered cohesive slope by using two intelligent techniques. An MLP neural network and a GP mode were applied to approximate the stability of the slope. The important parameter of the dimensionless stability number (N 2c ) was considered the networks’ output. It was estimated to be influenced by three effective factors, which were considered as (i) the layer’s height ratio (d/H), (ii) the slope angle (β), and (iii) the undrained cohesive strength ratio (C u1 /C u2 ) as the input dataset. Similar to previous research, the stock dataset was randomly divided into two parts to train and test the networks, with a ratio of 80% and 20%, respectively (e.g., Moayedi and Hayati [28], Moayedi and Hayati [22], Koopialipoor et al. [29] and Moayedi and Hayati [6]). Also, for each model, the performance was measured by the index of the statistical error terms RMSE and R2 . Equations (3) and (4) explain these indices:   N 1   2 RMSE =  Yiactual − Yiproduced (3) N i=1

2 N  − (Y) (Y) actual,j produced,j j=1 R2 = 1 −  

2 N j=1 (Y)actual,j − (Y)mean

(4)

Where Y i actual and Y i produced stand for the actual and predicted values of N 2c , respectively. N is the indicator of the number of data, and Y mean is the average of the actual slope stability values. The mentioned indices have been widely used in other earlier studies (e.g., Momeni et al. [30], Armaghani et al. [31], Mohamad et al. [32] and [33]). Optimal Hybrid GP Model Predicting N 2c. Determining proper network architecture is a necessary step in the utilization of artificial intelligence. For GP optimization, many trial and error processes, including 36 various GP models, were performed to find an appropriate GP structure for estimating the N 2c . The feasibility of the GP technique was evaluated for different numbers of generations, values of swarm sizes, and selection tournament sizes. For all three mentioned parameters, 12 values, including 50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, and 1000 were considered, and the performance was evaluated by means of the RMSE reduction procedure. Regarding the trial and error processes provided in Figs. 5, 6 and 7, the GP model with the values of 750, 200, and 250 respectively for the population size, some generations and selection tournament size showed the best performance, as indicated by its lower RMSE value. Therefore, this structure was introduced as the optimal architecture of the GP for any further N 2c estimation.

98

H. Moayedi

0.06 0.05

RMSE

0.04 0.03 0.02 0.01 0 0

50

Populaon Size=50 Populaon Size=200 Populaon Size=350

100 Number of generaons Populaon Size=100 Populaon Size=250 Populaon Size=400

150

200

Populaon Size=150 Populaon Size=300 Populaon Size=450

Fig. 5. GP network performance results for different population sizes

Optimal Artificial Neural Network Predicting N2c. Like the GP model, the best ANN architecture was obtained by assessing several ANN structures’ performance. For the ANN, a multilayer perceptron network was verified with various numbers of neurons in its single hidden layer. After a trial and error process, the network that reported the lowest RMSE and the highest R2 was selected as the optimum model. The results of this are depicted in Fig. 8. Note that every structure was performed for six iterations. As a general deduction, the MLP network with at least eight neurons in its hidden layer could present an accruable performance for modeling the problem.

Two Novel Predictive Networks for Slope Stability Analysis

99

0.05

RMSE

0.04

0.03

0.02

0.01

0 0

50

tournament Size=50 Tournament Size=200 Tournament Size=350

100 150 Number of generaons Tournament Size=100 Tournament Size=250 Tournament Size=400

200

Tournament Size=150 Tournament Size=300 Tournament Size=450

Fig. 6. GP network performance results for different tournament sizes

0.04

RMSE

0.03

0.02

0.01

0 0

200 Number of generaons=50 Number of generaons=150 Number of generaons=250 Number of generaons=350 Number of generaons=450

400

600

800

Number of generaons Number of generaons=100 Number of generaons=200 Number of generaons=300 Number of generaons=400 Number of generaons=500

Fig. 7. GP network performance results for a different number of generations

1000

100

H. Moayedi 1.00

1.00

0.99

Tesng R2

Training R2

1.00 0.99 0.99 0.98 0.98

TR1

TR2

TR3

TR4

TR5

TR6

0.98 TS1 TS3 TS5

0.97

0.97

TS2 TS4 TS6

0.96

0.97 0.95

0.96 0.96 0

2

4

6

8

0.94

10

0

2

Nodes in hidden layer` (a)

6

8

10

(b) 0.06

0.06 0.05

TR1 TR3 TR5

0.04

0.05

TR2 TR4 TR6

Tesng RMSE

Training RMSE

4

Nodes in hidden layer`

0.03

0.04

0.02

0.01

0.01

0

2

4

6

Nodes in hidden layer` (c)

8

10

TS2 TS4 TS6

0.03

0.02

0.00

TS1 TS3 TS5

0.00 0

2

4

6

8

10

Nodes in hidden layer` (d)

Fig. 8. Sensitivity analysis for ANN, based on (a) R2 and (b) RMSE values reported for training and testing datasets

Two Novel Predictive Networks for Slope Stability Analysis

101

6 Models Evaluation and Design Solution Charts After obtaining the optimum structures of both the GP and ANN methods, they were applied to the prepared dataset for the proposed estimation of the dimensionless slope stability number. The training dataset did a training operation, and the performance of each model was evaluated using the testing data. Two usual types of charts were used to describe and analyze the results. Firstly, the accommodation of the predicted and actual values of N 2c was depicted in the form of Fig. 9. Also, Table 2 lists the calculated RMSE and R2 . As illustrated in Fig. 9 and Table 2, the GP and ANN models had a robust prediction and sufficient reliability in the slope stability assessment. The high values of R2 can prove this claim, and the low values of RMSE obtained from both model performances. The training results show RMSE of 0.010274 and 0.006112 and R2 of 0.9968 and 0.9989 for the GP and ANN methods. The results of the test phase show a good approximation for these models too. The RMSE and R2 values are 0.011146 and 0.005927, and 0.9967 and 0.9990, respectively, for the GP and ANN methods. Based on the reported results, the optimized ANN had a slightly better performance compared to the optimized GP model, as can be concluded from the lower RMSE and higher R2 values in both the training and testing phases.

(a)

(b)

(c)

(d)

Fig. 9. Training (a and b) and testing (c and d) results of GP and ANN models for predicting N 2c

In the following, the measured results from the upper bound analysis of slope stability are compared with the predicted values obtained from the optimal GP and ANN

102

H. Moayedi Table 2. Obtained RMSE and R2 values for GP and ANN results Model Dataset RMSE

R2

RMSE

R2

GP

0.010274 0.9968 0.011146 0.9967

ANN

0.006112 0.9989 0.005927 0.9990

algorithms. Consequently, Figs. 10-a to 10-e present the measured N 2c as well as the results obtained from the GP and ANN training networks for slope angles between 15 to 75°, respectively. In this regard, according to the different values of the slope angle (β) (i.e., 15°, 30°, 45°, 60° and 75°), separate solution of designing slope stability charts were provided, noting that in each figure the vertical and horizontal axes stand for the N 2c and d/H ratio, respectively. To include the C u1 /C u2 ratio parameter, C u1 /C u2 ratios of 0.2, 0.8, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, and 5 were considered through the calculation steps. Therefore, ten different curves are plotted for each of the predefined slopes. The curves showing C u1 /C u2 = 0.2 to 0.8 are drawn in a relatively straight path, and the N 2c does not show noticeable changes as the d/H ratio rises. In contrast, for the remaining eight curves (especially for C u1 /C u2 = 3 to C u1 /C u2 = 5), the N 2c values have an ascending direction as the d/H ratio rises. Note that the distance between these two curves decreases as β rises so that in the last graph (β = 75°), they have covered each other. Also, there is another difference in the last graph. It relates to the fluctuations of the curves (especially for the curves determined by C u1 /C u2 = 3.5, 4, and 5). In graphs, a – d, a steeper tangent can be observed for d/H values less than 2.5, while it differs for graph e, and camber can be seen in the vicinity of d/H = 2 (especially for C u1 /C u2 = 5). As can be obtained from Figs. 10a–e, both the GP and ANN models gave a satisfactory prediction for estimating the slope stability number. This can be concluded from the GP and ANN curves’ appropriate proximity to the measured N 2c (target data). It is worth noting that each target curve’s general direction is well approximated by both the GP and ANN tools. Therefore, there is good coverage for all of them. Focusing on the measured datasets, considering the black and red curves in Figs. 10a– e (C u1 /C u2 = 4.5 and 5), there are fluctuations and sudden changes for the N 2c results provided for d/H ratios, which are lower than 2.0. To evaluate the GP and ANN’s flexibility, it is necessary to compare their reactions against these changes. In cases a–d, it is observed that the GP model had maintained its main direction without significant changes, even for critical d/H ratios (see red and black dash-dot curves). On the contrary, the changes are well distinguished by the ANN curves, especially for the black curves in Figs. 10a–d. In each case, the fluctuation is well followed by the ANN, and there is excellent accommodation between them. Also, for the changes occurring in Fig. 10e (β = 75°), the ANN gave a better approximation, and this is because its curve is closer to the target curve. For another example, the ANN’s higher precision can be deduced considering the grey curve (C u1 /C u2 = 2) in cases a–c, while more distance between the measured and predicted curves are reported for the GP. Regarding these descriptions and also the obtained RMSE and R2 values, it can be concluded that both the GP and

Two Novel Predictive Networks for Slope Stability Analysis

103

0.9 0.8 0.7

N2c

0.6 0.5 0.4 0.3 0.2 0.1 0 1

1.5

2

2.5

3

3.5

4

d/H Measured -Cu1/Cu2 = 0.8 Measured - Cu1/Cu2 = 2.5 Measured - Cu1/Cu2 = 4.0 ANN - Cu1/Cu2 = 0.2 ANN - Cu1/Cu2 = 2.0 ANN - Cu1/Cu2 = 3.5 ANN - Cu1/Cu2 = 5.0 GP - Cu1/Cu2 = 1.5 GP - Cu1/Cu2 = 3.0 GP - Cu1/Cu2 = 4.5

Measured - Cu1/Cu2 = 0.2 Measured - Cu1/Cu2 = 2.0 Measured - Cu1/Cu2 = 3.5 Measured - Cu1/Cu2 = 5.0 ANN - Cu1/Cu2 = 1.5 ANN - Cu1/Cu2 = 3.0 ANN - Cu1/Cu2 = 4.5 GP - Cu1/Cu2 = 0.8 GP - Cu1/Cu2 = 2.5 GP - Cu1/Cu2 = 4.0

4.5

5

Measured -Cu1/Cu2 = 1.5 Measured - Cu1/Cu2 = 3.0 Measured - Cu1/Cu2 = 4.5 ANN - Cu1/Cu2 = 0.8 ANN - Cu1/Cu2 = 2.5 ANN - Cu1/Cu2 = 4.0 GP - Cu1/Cu2 = 0.2 GP - Cu1/Cu2 = 2.0 GP - Cu1/Cu2 = 3.5 GP - Cu1/Cu2 = 5.0

(a) 0.9 0.8 0.7 0.6

N2c

0.5 0.4 0.3 0.2 0.1 0 1

1.5

2

Measured - Cu1/Cu2 = 0.2 Measured - Cu1/Cu2 = 2.0 Measured - Cu1/Cu2 = 3.5 Measured - Cu1/Cu2 = 5.0 ANN - Cu1/Cu2 = 1.5 ANN - Cu1/Cu2 = 3.0 ANN - Cu1/Cu2 = 4.5 GP - Cu1/Cu2 = 0.8 GP - Cu1/Cu2 = 2.5 GP - Cu1/Cu2 = 4.0

2.5

3

3.5

d/H Measured -Cu1/Cu2 = 0.8 Measured - Cu1/Cu2 = 2.5 Measured - Cu1/Cu2 = 4.0 ANN - Cu1/Cu2 = 0.2 ANN - Cu1/Cu2 = 2.0 ANN - Cu1/Cu2 = 3.5 ANN - Cu1/Cu2 = 5.0 GP - Cu1/Cu2 = 1.5 GP - Cu1/Cu2 = 3.0 GP - Cu1/Cu2 = 4.5

4

4.5

5

Measured -Cu1/Cu2 = 1.5 Measured - Cu1/Cu2 = 3.0 Measured - Cu1/Cu2 = 4.5 ANN - Cu1/Cu2 = 0.8 ANN - Cu1/Cu2 = 2.5 ANN - Cu1/Cu2 = 4.0 GP - Cu1/Cu2 = 0.2 GP - Cu1/Cu2 = 2.0 GP - Cu1/Cu2 = 3.5 GP - Cu1/Cu2 = 5.0

(b)

Fig. 10. Chart solutions for GP and ANN models for (a) β = 15°, (b) β = 30°, (c) β = 45°, (d) β = 60°, (e) β = 75°

104

H. Moayedi 0.9 0.8 0.7 0.6

N2c

0.5 0.4 0.3 0.2 0.1 0 1

1.5

2

2.5

Measured - Cu1/Cu2 = 0.2 Measured - Cu1/Cu2 = 2.0 Measured - Cu1/Cu2 = 3.5 Measured - Cu1/Cu2 = 5.0 ANN - Cu1/Cu2 = 1.5 ANN - Cu1/Cu2 = 3.0 ANN - Cu1/Cu2 = 4.5 GP - Cu1/Cu2 = 0.8 GP - Cu1/Cu2 = 2.5 GP - Cu1/Cu2 = 4.0

3

3.5 d/H Measured -Cu1/Cu2 = 0.8 Measured - Cu1/Cu2 = 2.5 Measured - Cu1/Cu2 = 4.0 ANN - Cu1/Cu2 = 0.2 ANN - Cu1/Cu2 = 2.0 ANN - Cu1/Cu2 = 3.5 ANN - Cu1/Cu2 = 5.0 GP - Cu1/Cu2 = 1.5 GP - Cu1/Cu2 = 3.0 GP - Cu1/Cu2 = 4.5

4

4.5

5

Measured -Cu1/Cu2 = 1.5 Measured - Cu1/Cu2 = 3.0 Measured - Cu1/Cu2 = 4.5 ANN - Cu1/Cu2 = 0.8 ANN - Cu1/Cu2 = 2.5 ANN - Cu1/Cu2 = 4.0 GP - Cu1/Cu2 = 0.2 GP - Cu1/Cu2 = 2.0 GP - Cu1/Cu2 = 3.5 GP - Cu1/Cu2 = 5.0

(c) 0.9 0.8 0.7 0.6

N2c

0.5 0.4 0.3 0.2 0.1 0 1

1.5

2

Measured - Cu1/Cu2 = 0.2 Measured - Cu1/Cu2 = 2.0 Measured - Cu1/Cu2 = 3.5 Measured - Cu1/Cu2 = 5.0 ANN - Cu1/Cu2 = 1.5 ANN - Cu1/Cu2 = 3.0 ANN - Cu1/Cu2 = 4.5 GP - Cu1/Cu2 = 0.8 GP - Cu1/Cu2 = 2.5

2.5

3

d/H

3.5

Measured -Cu1/Cu2 = 0.8 Measured - Cu1/Cu2 = 2.5 Measured - Cu1/Cu2 = 4.0 ANN - Cu1/Cu2 = 0.2 ANN - Cu1/Cu2 = 2.0 ANN - Cu1/Cu2 = 3.5 ANN - Cu1/Cu2 = 5.0 GP - Cu1/Cu2 = 1.5 GP - Cu1/Cu2 = 3.0

(d)

Fig. 10. (continued)

4

4.5

5

Measured -Cu1/Cu2 = 1.5 Measured - Cu1/Cu2 = 3.0 Measured - Cu1/Cu2 = 4.5 ANN - Cu1/Cu2 = 0.8 ANN - Cu1/Cu2 = 2.5 ANN - Cu1/Cu2 = 4.0 GP - Cu1/Cu2 = 0.2 GP - Cu1/Cu2 = 2.0 GP - Cu1/Cu2 = 3.5

Two Novel Predictive Networks for Slope Stability Analysis

105

0.9 0.8 0.7 0.6

N2c

0.5 0.4 0.3 0.2 0.1 0 1

1.5

2

2.5

3

3.5

4

4.5

5

d/H Measured - Cu1/Cu2 = 0.2 Measured - Cu1/Cu2 = 2.0 Measured - Cu1/Cu2 = 3.5 Measured - Cu1/Cu2 = 5.0 ANN - Cu1/Cu2 = 1.5 ANN - Cu1/Cu2 = 3.0 ANN - Cu1/Cu2 = 4.5 GP - Cu1/Cu2 = 0.8 GP - Cu1/Cu2 = 2.5 GP - Cu1/Cu2 = 4.0

Measured -Cu1/Cu2 = 0.8 Measured - Cu1/Cu2 = 2.5 Measured - Cu1/Cu2 = 4.0 ANN - Cu1/Cu2 = 0.2 ANN - Cu1/Cu2 = 2.0 ANN - Cu1/Cu2 = 3.5 ANN - Cu1/Cu2 = 5.0 GP - Cu1/Cu2 = 1.5 GP - Cu1/Cu2 = 3.0 GP - Cu1/Cu2 = 4.5

Measured -Cu1/Cu2 = 1.5 Measured - Cu1/Cu2 = 3.0 Measured - Cu1/Cu2 = 4.5 ANN - Cu1/Cu2 = 0.8 ANN - Cu1/Cu2 = 2.5 ANN - Cu1/Cu2 = 4.0 GP - Cu1/Cu2 = 0.2 GP - Cu1/Cu2 = 2.0 GP - Cu1/Cu2 = 3.5 GP - Cu1/Cu2 = 5.0

(e)

Fig. 10. (continued)

ANN models have sufficient applicability. Besides, the higher flexibility of the ANN in design solution charts can be concluded due to its response against the changes. To provide a reliable solution equation that reflects the presented design charts, an equation was derived from both the GP and ANN trained networks to be presented in respect of the slope stability issue. The GP and ANN formulas are presented in Eqs. (5) and (6), respectively.

(5) Where, X1 = d/H; X2 = β; and X3 = Cu1/Cu2.

(6) Where, Y1, Y2 … Y8 are defined by Eqs. 7–14: Y1 = Tansig (−1.0821 × X1 − 0.4887 × X2 + 2.6243 × X3 + 3.1157)

(7)

Y2 = Tansig (0.1386 × X1 − 0.1071 × X2 + 0.6546 × X3 − 0.4376)

(8)

106

H. Moayedi

Y3 = Tansig (0.0804 × X1 − 0.0947 × X2 + 0.2168 × X3 − 0.0942)

(9)

Y4 = Tansig (2.3293 × X1 − 2.0853 × X2 + 2.2883 × X3 − 0.0371)

(10)

Y5 = Tansig (−0.0285 × X1 + 0.3638 × X2 − 3.0985 × X3 − 1.5416) (11) Y6 = Tansig (0.9039 × X1 + 0.3419 × X2 − 0.7021 × X3 + 1.9361)

(12)

Y7 = Tansig (−1.6865 × X1 + 3.3371 × X2 + 0.3808 × X3 − 1.6126) (13) Y8 = Tansig (−3.1807 × X1 + 0.8723 × X2 + 0.8388 × X3 − 4.1277) (14) where in each equation, X1 = d/H; X2 = β; X3 = Cu1/Cu2. It is important to note that the proposed formula can be used in the slope stability problem to calculate the dimensionless stability number for slopes with a maximum of two layers with similar material conditions. Besides, there are two crucial advantages for the GP model equation compared with that of the ANN. One of them is the GP equation, which is easier to use than the ANN equation (i.e., the intermediate equations of Y1 to Y8 need to be considered before Eq. (6) can be used). The other is the GP equation that can be used directly and requires no normalization process (while in the ANN, the input layers need to be normalized before any further process).

7 Conclusions Regarding the importance of the slope stability issue in many engineering projects, the main objective of this research was to evaluate the capability of two artificial intelligence techniques in the assessment of cohesive slope stability. Optimized GP and optimized ANN methods were applied to the dataset collected from a finite-element procedure called upper bound (UB) limit analysis. Three effective factors to the cohesive slope’s stability were considered as input data to produce the dimensionless stability number. The first factor was the d/H ratio, where d is the slope thickness, and H is an indicator of the topsoil layer’s height. Slope angle (β) was the second factor, and the third parameter was the ratio of the undrained shear strength of the first soil layer to the slope layer’s (C u1 /C u2 ). Besides the use of the statistical indices of RMSE and R2 , the results were compared by the form of design charts solution. From the results, both the RMSE and R2 values showed a slightly better approximation for the ANN than for the GP model in training (RMSE of 0.010274 and 0.006112, and R2 of 0.9968 and 0.9989, respectively for the GP and ANN methods) and testing (RMSE of 0.011146 and 0.005927, and R2 of 0.9967 and 0.9990, respectively for the GP and ANN methods) phases. We also found the ANN to be the more reliable model based on the reported results from the design charts. Regarding these charts, the ANN showed higher flexibility in fluctuations. In the following, two formulas were extracted based on each of the optimized GP and ANN

Two Novel Predictive Networks for Slope Stability Analysis

107

models, for use in the field of slope stability assessment, in place of the traditional formula of the UB limit analysis method. Note that the GP formula was introduced as the more applicable formula due to its greater brevity and simplicity. Compliance with Ethical Standards. Conflict of Interest: The authors declare that they have no conflict of interest.

References 1. Griffiths, D.V., Lane, P.A.: Slope stability analysis by finite elements. Geotechnique 49, 387– 403 (1999) 2. Taylor, D.W.: Stability of earth slopes. J. Boston Soc. Civ. Eng. 24, 197–246 (1937) 3. Qian, Z.G., Li, A.J., Merifield, R.S., Lyamin, A.V.: Slope stability charts for two-layered purely cohesive soils based on finite-element limit analysis methods. Int. J. Geomech. 15, 06014022 (2014) 4. Abuel-Naga, H.M., Bergado, D.T., Gniel, J.: Design chart for prefabricated vertical drains improved ground (reference no 2869). Geotextiles Geomembranes 43, 537–546 (2015) 5. Aksoy, H.S., Gor, M., Inal, E.: A new design chart for estimating friction angle between soil and pile materials. Geomech. Eng. 10, 315–324 (2016) 6. Moayedi, H., Hayati, S.: Artificial intelligence design charts for predicting friction capacity of driven pile in clay. Neural Comput. Appl. 31 (2018, in press) 7. Samui, P., Kumar, B.: Artificial neural network prediction of stability numbers for two-layered slopes with associated flow rule. Electron. J. Geotech. Eng. 11, 1–44 (2006) 8. Jellali, B., Frikha, W.: Constrained particle swarm optimization algorithm applied to slope stability. Int. J. Geomech. 17, 06017022 (2017) 9. Jiang, X., Niu, J., Yang, H., Wang, F.: Upper Bound Limit Analysis for Seismic Stability of Rock Slope with Tunnel. Advances in Civil Engineering 2018 (2018) 10. Pan, Q., Dias, D.: Upper-bound analysis on the face stability of a non-circular tunnel. Tunnelling Underground Space Technol. 62, 96–102 (2017) 11. Cramer, N.L.: A representation for the adaptive generation of simple sequential programs. In: Proceedings of the First International Conference on Genetic Algorithms (1985) 12. Koza, J.R.: Genetic programming as a means for programming computers by natural selection. Stat. Comput. 4, 87–112 (1994) 13. Johari, A., Habibagahi, G., Ghahramani, A.: Prediction of soil–water characteristic curve using genetic programming. J. Geotech. Geoenviron. Eng. 132, 661–665 (2006) 14. Makkeasorn, A., Chang, N.B., Beaman, M., Wyatt, C., Slater, C.: Soil moisture estimation in a semiarid watershed using RADARSAT-1 satellite imagery and genetic programming. Water Resourc. Res. 42 (2006) 15. Garg, A., Garg, A., Tai, K., Sreedeep, S.: An integrated SRM-multi-gene genetic programming approach for prediction of factor of safety of 3-D soil nailed slopes. Eng. Appl. Artif. Intell. 30, 30–40 (2014) 16. Garg, A., Garg, A., Tai, K.: A multi-gene genetic programming model for estimating stressdependent soil water retention curves. Comput. Geosci. 18(1), 45–56 (2013). https://doi.org/ 10.1007/s10596-013-9381-z 17. Khandelwal, M., Marto, A., Fatemi, S.A., Ghoroqi, M., Armaghani, D.J., Singh, T.N., Tabrizi, O.: Implementing an ANN model optimized by genetic algorithm for estimating cohesion of limestone samples. Eng. Comput. 34(2), 307–317 (2017). https://doi.org/10.1007/s00366017-0541-y

108

H. Moayedi

18. McCulloch, W.S., Pitts, W.: A logical calculus of the ideas immanent in nervous activity. Bull. Math. Biophys. 5, 115–133 (1943) 19. Jain, A.K., Mao, J., Mohiuddin, K.M.: Artificial neural networks: a tutorial. Computer 29, 31–44 (1996) 20. Rao, S.G.: Artificial neural networks in hydrology. II: Hydrologic applications. J. Hydrol. Eng. 5, 124–137 (2000) 21. El-Bakry, M.Y.: Feed forward neural networks modeling for K-P interactions. Chaos, Solitons Fractals 18, 995–1000 (2003) 22. Moayedi, H., Hayati, S.: Applicability of a CPT-based neural network solution in predicting load-settlement responses of bored pile. Int. J. Geomech. 18 (2018) 23. Florkiewicz, A.: Upper bound to bearing capacity of layered soils. Can. Geotech. J. 26, 730–736 (1989) 24. Donald, I.B., Chen, Z.: Slope stability analysis by the upper bound approach: fundamentals and methods. Can. Geotech. J. 34, 853–862 (1997) 25. Ranjbar Karkanaki, A., Ganjian, N., Askari, F.: Stability analysis and design of cantilever retaining walls with regard to possible failure mechanisms: an upper bound limit analysis approach. Geotech. Geol. Eng. 35(3), 1079–1092 (2017). https://doi.org/10.1007/s10706017-0164-5 26. Caër, T., Souloumiac, P., Maillot, B., Leturmy, P., Nussbaum, C.: Propagation of a fold-andthrust belt over a basement graben. J. Struct. Geol. (2018) 27. Zhou, H., Liu, H., Yin, F., Chu, J.: Upper and lower bound solutions for pressure-controlled cylindrical and spherical cavity expansion in semi-infinite soil. Comput. Geotech. 103, 93–102 (2018) 28. Moayedi, H., Hayati, S.: Modelling and optimization of ultimate bearing capacity of strip footing near a slope by soft computing methods. Appl. Soft Comput. 66, 208–219 (2018) 29. Koopialipoor, M., Jahed Armaghani, D., Hedayat, A., Marto, A., Gordan, B.: Applying various hybrid intelligent systems to evaluate and predict slope stability under static and dynamic conditions. Soft. Comput. 23(14), 5913–5929 (2018). https://doi.org/10.1007/s00500-0183253-3 30. Momeni, E., Armaghani, D.J., Hajihassani, M., Amin, M.F.M.: Prediction of uniaxial compressive strength of rock samples using hybrid particle swarm optimization-based artificial neural networks. Measurement 60, 50–63 (2015) 31. Armaghani, D.J., Tonnizam Mohamad, E., Momeni, E., Monjezi, M., Sundaram Narayanasamy, M.: Prediction of the strength and elasticity modulus of granite through an expert artificial neural network. Arabian J. Geosci. 9(1), 1–16 (2015). https://doi.org/10.1007/ s12517-015-2057-3 32. Mohamad, E.T., Armaghani, D.J., Momeni, E., Yazdavar, A.H., Ebrahimi, M.: Rock strength estimation: a PSO-based BP approach. Neural Comput. Appl. 30(5), 1635–1646 (2016). https://doi.org/10.1007/s00521-016-2728-3 33. Moayedi, H., Jahed Armaghani, D.: Optimizing an ANN model with ICA for estimating bearing capacity of driven pile in cohesionless soil. Eng. Comput. 34(2), 347–356 (2017). https://doi.org/10.1007/s00366-017-0545-7

A Review of Artificial Intelligence Applications in Mining and Geological Engineering Xuan-Nam Bui1,2(B)

, Hoang-Bac Bui3,4 , and Hoang Nguyen1,2

1 Department of Surface Mining, Mining Faculty, Hanoi University of Mining and Geology,

18 Pho Vien, Duc Thang Ward, Bac Tu Liem District, Hanoi 100000, Vietnam [email protected] 2 Center for Mining, Electro-Mechanical Research, Hanoi University of Mining and Geology, 18 Pho Vien, Duc Thang Ward, Bac Tu Liem District, Hanoi 100000, Vietnam 3 Faculty of Geosciences and Geoengineering, Hanoi University of Mining and Geology, 18 Vien st., Duc Thang Ward, Bac Tu Liem District, Hanoi 100000, Vietnam 4 Center for Excellence in Analysis and Experiment, Hanoi University of Mining and Geology, 18 Vien st., Duc Thang Ward, Bac Tu Liem District, Hanoi 100000, Vietnam

Abstract. Artificial intelligence (AI) is well-known as a robust technique that can support and improve the quality of human life. In the mining industry, applications of AI changed the sciences and technologies, as well as the performance of the mining industry, especially in mining and geological engineering. Smart mines were introduced and widely applied around the world with advanced technologies based on the applications of AI. This paper aims to provide a comprehensive view of AI applications in mining and geological engineering, as well as the ideas for studies in the future. The paper focuses on the published papers of AI applications in rock mechanics, mining method selection, mining equipment, drilling-blasting, slope stability, environmental issues, and relevant geological engineering. The advantages and disadvantages of AI applications in mining and geological engineering will be analyzed and discussed in detail. Keywords: Mining industry · Geo-engineering · Artificial intelligence · Machine learning · The fourth industrial revolution

1 Introduction “Mining is not everything, but without mining, everything is nothing.” Max Planck - a famous English philosopher, said that. It has affirmed the vital role of the mining industry in the world. It is considered as an essential key and has a significant impact on many other industries. Indeed, large industries such as aviation, automobile manufacturing, mechanics, energy, electronics, to name a few, are all supplied with raw-fuel by-products of the mining industry [1–5]. With the significant demand for raw materials from other industries, advanced mining technologies also need to be applied to improve mining efficiency and minimize negative impacts on the surrounding environment. Of those, artificial intelligence (AI) is considered as a robust tool for mining problems [6]. In fact, © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 109–142, 2021. https://doi.org/10.1007/978-3-030-60839-2_7

110

X.-N. Bui et al.

mining activities and geological engineering are taken into account as important issues and attract much attention from scholars as well as engineers. The problems related to the environment are phrased as impacts induced by mining activities [7]. In actuality, mining and geological engineering is a combination of underlying sciences, and they are not that simple. The primary objective of mining and geological engineering is natural resources, rock, soil, groundwater, surface water, and most of them are uncertainty factors [8–11]. Surveying, exploration and geological mapping for mines have encountered many difficulties and inaccuracies due to these uncertainty factors. Therefore, managers of mines are often faced with many complex decisionmaking problems without certainties [12, 13]. They can cause severe damage to people and property as well as significantly affect the production efficiency of the mine if not properly assessed. Issues related to working efficiency and optimal planning for mining are also inherently unsatisfied by investors [14]. That led to a series of changes to the improvement and optimization of mine design [15, 16]. In open-pit mines, the dangers of blasting operations are always of great concern to engineers and scientists. Blast-induced ground vibration, air over-pressure, fly-rock, back-break, dust, and toxic are always the challenges of engineers and researchers in the effort of reduction of blast-induced issues [17–19]. Many mines were forced to close due to the adverse effects of blasting. Besides, natural hazards are also considered a rival of engineers and managers. Research efforts to ensure the stability of benches and slopes are non-stop due to the influence of many uncertainty factors, as well as impacts from blasting and earthquakes [20–22]. For underground mines, similar difficulties have also been mentioned. Besides, due to the effect of mine pressure, heterogeneous rock environment and blast-induced ground vibration, as well as earthquakes, can lead to the collapse of the underground mines [23, 24]. Toxic gases, such as methane, CO, SO2 , are also potential dangers of fire and suffocation in underground mines [25–27]. In recent years, AI is considered a powerful tool capable of solving practical problems related to mining and geo-engineering. This paper aims to provide a review of AI applications in mining and geological engineering, discussion of these AI applications, as well as the chances for future in the mining industry.

2 Applications of Artificial Intelligence in Mining and Geological Engineering 2.1 Rock Mechanics Rock mechanics is one of the indispensable fields in mining engineering. It is considered as a foundation to choose methods of mining, drilling, blasting, or assessing slope stability. However, it is an uncertainty problem and very hard to evaluate precisely. Currently, many soft computing techniques have been introduced to analyze and evaluate the uncertainty of rock mass (Fig. 1), and their performance has been significantly enhanced [28]. Herein, the FLAC3D is numerical modeling software for geotechnical analyses of soil, constructs, groundwater, rock, and ground support. It can analyze or design many engineering problems, research and testing, the factor of safety prediction, and the backanalysis of failure. Whereas the 3DEC is a three-dimensional numerical modeling code

A Review of Artificial Intelligence Applications

111

for advanced geotechnical analysis, and its features are similar to the FLAC3D. Unlike the FLAC3D or 3DEC software, the PFC3D can be customized and applied to an extensive range of numerical investigations. It has been successfully applied for geoscience investigation, such as brittle rock fracturing, hydraulic fracturing, slope stability, soiltool interactions, cave mining. And bulk material flow/mixing. On another side, the RS3 software is a 3Dfinite element analysis program for tunnel and support design, modeling slopes, foundation design, embankments, surface and underground excavations, groundwater seepage, consolidation, and more. Similar to the RS3 , but the Examine2D is a 2-dimensional plane strain boundary element program for the elastic stress analysis of underground excavations. Phase2 software was also designed for similar purposes, but it is an extremely versatile 2D elastoplastic finite element stress analysis program. Whereas, the RESOBLOCK software was designed to investigate the stability of excavations and the impact of different rock bolting patterns. However, these techniques are often time-consuming and expensive.

Fig. 1. Some applications of numerical analyses for rock mass [6, 29].

In recent years, AI has been studied and widely applied in evaluating and predicting rock mechanics-related problems. Along with the soft computing techniques for numerical simulation, as mentioned above, AI is taken into account as a robust tool to predict/estimate/forecast the rock mechanics and rock engineering. It can overcome the drawbacks of the numerical analysis methods with low cost and save time. Sonmez, Gokceoglu, Nefeslioglu, and Kayabasi [30] applied an artificial neural network (ANN)

112

X.-N. Bui et al.

to estimate rock modulus in intact rocks. Accordingly, the modulus ratio ranges of intact rock (Ei ) are extensive, and they are even overlapped, as illustrated in Fig. 2. Therefore, the use of average values of Ei may be too rough since the uncertainty of rock mass. To overcome this drawback, Sonmez, Gokceoglu, Nefeslioglu and Kayabasi [30] used ANN with the uniaxial compressive strength and unit weight to estimate Ei with high reliability (i.e., correlation coefficient R = 0.82).

Fig. 2. The modulus ratio ranges of intact rock (with different intact rocks) [30].

In another study, Ocak and Seker [31] also developed an ANN model to predict the elastic modulus of intact rocks. A promising result was introduced in their study with a root-mean-squared error (RMSE) of 0.191 and R of 0.926. It should be noted that the uniaxial compressive strength and unit weight was also used to predict Ei as those previous study of Sonmez, Gokceoglu, Nefeslioglu, and Kayabasi [30]. Herein, core samples were collected and analyzed for the data collection (Fig. 3).

Fig. 3. Data collection and analysis [31]. (a) The specimen tests and loading period-stress; (b) Distribution of the dataset

A Review of Artificial Intelligence Applications

113

Majdi and Beiki [32] also applied ANN to predict the deformation modulus of rock masses. However, the genetic algorithm (GA) was also applied to optimize the ANN model aiming to obtain better performance. The flowchart of the GA-ANN model for predicting the deformation modulus of rock masses is shown in Fig. 4. In another study, Rukhaiyar and Samadhiya [33] applied ANN to predict the strength of intact sandstone. Unlike the previous studies, Rukhaiyar and Samadhiya [33] used three input parameters, such as uniaxial compressive strength, intermediate, and principal stresses. One hundred ninety-two samples were collected from available databases of previous studies; however, the performance of the ANN was very high (i.e., RMSE = 10.73, R2 = 0.97). This result showed the insights of AI (i.e., ANN) in rock mechanics and rock engineering. Similarly, Rashidi, Hajipour and Asadi [34] also applied ANN to predict the strength of intact limestone with a promising result. Based on the literature review, it is clear that ANN is a reliable method to estimate/predict/forecast the rock mechanics, as well as effectiveness supporting in rock engineering.

Fig. 4. Flowchart of the GA-ANN model.

On another side of rock mechanics, rockburst is taken into account as one of the critical behaviors of rock mass in deep openings. In this regard, AI techniques have also been applied to evaluate, classify, and predict the ability of rockburst, as well as the intensity of it. Indeed, Dong, Li, and Kang [35] classified rockburst using the

114

X.-N. Bui et al.

Random Forest (RF) algorithm. Their classification results showed that RF was a good algorithm in classifying the intensity of rockburst. Zhou, Yun, Deng, Li and Liu [36] also applied various AI techniques, such as RF, Bayes, K-nearest neighbors (KNN), and cloud model (CM), for classification of rockburst intensity. Finally, they found that the CM technique can predict rockburst better than those of the other models. In another study, Faradonbeh and Taheri [37] also applied three AI techniques, i.e., emotional neural network (ENN), decision tree C4.5, and gene expression programming (GEP), for predicting the intensity of rockburst with a promising result. Zhou, Koopialipoor, Li, and Armaghani [38] considered and developed a novel hybrid AI model based on ANN and artificial bee colony (ABC) optimization algorithm for predicting rockburst, called ABC-ANN. They claimed that the ABC-ANN model can predict rock burst with high accuracy. Figure 5 illustrates the vertical stress distribution around a workface and the accuracy of the AI techniques used for predicting rock burst.

Fig. 5. Illustration of vertical stress distribution around a workface and the accuracy of the AI techniques used for predicting rockburst [38, 39].

A Review of Artificial Intelligence Applications

115

2.2 Mining Method Selection Mining method selection is one of the most stages of mining. It decides economic, technical, and environmental efficiency. Mining method selection is taken into account as a problem with multiple decision making, as mentioned in Fig. 6. In fact, some mines use only one mining method. However, some of the mines with more complex conditions require the use of a combination of different mining methods. Therefore, choosing the optimal mining method is a challenge for engineers and researchers.

Fig. 6. Framework of mining method selection based on criteria.

To solve these above problems, AI has been studied and introduced as a useful tool in decision making and mining method selection early. Indeed, Yun and Huang [40] applied the theory of fuzzy set to select the optimal mining method for an underground mine. Three stages were applied for this aim, including initial selection, technical and economic evaluation, as well as the final decision. In another work, mining method selection was also defined as the most critical point in practical engineering. It was recommended as a significant effect on productivity, safety, and economics. Therefore, Guray, Celebi, Atalay, and Pasamehmetoglu [41] proposed an AI model for mining method selection based on 13 different expert systems. This system can support engineers as much as possible to select a suitable mining method with the highest efficiency. Naghadehi, Mikaeil, and Ataei [42] also successfully applied the fuzzy analytic hierarchy process (FAHP) model for mining method selection in an underground bauxite mine of Iran. This model is based on the practical and majority criterial to make a decision. Another approach for mining method selection was also proposed by Azadeh, Osanloo, and Ataei [43] based on an enhanced method of the Nicholas technique [44]. Accordingly, a similar model (i.e., AHP) was applied for this task. However, fuzzy trapezoidal factors were also used to model the mining method. Furthermore, an algorithm with two steps (e.g., hierarchical technical–operational model

116

X.-N. Bui et al.

(HTOM) and hierarchical economic model (HEM)) was proposed in their study for selecting the suitable mining method. Considering the mechanization criteria, Özfırat [45] proposed a fuzzy model for mining method selection with a promising result. Taking into account other criteria (i.e., geological and geometrical characteristics), Dehghani, Siami, and Haghi [46] applied the grey and Tomada de Decisão Interativa Multicritério (TODIM) methods for mining method selection. Their results showed that the approaches applied are better than the previous methods. Fu, Wu, Liao, and Herrera [47] showed that hesitant fuzzy linguistic gained and lost dominance score is a robust technique to select the suitable mining method (HFL-GLDS). The framework of this method is shown in Fig. 7. In a new study, Liang, Zhao, and Hong [48] proposed the MULTIMOORA model for mining method selection. This model combined three methods, including stepwise weight assessment ratio analysis (SWARA), Heronian mean (HM) operators, and a combination of SWARA and HM, to making a final decision.

Fig. 7. Framework of the HFL-GLDS technique for mining method selection [47].

2.3 Mining Equipment Selection In mining, equipment is indispensable for mining activities, such as drilling, blasting, loading/unloading, transporting, dumping, crushing, to name a few. However, the selection of proper mining equipment is not easy. Different types of equipment with different attributes have a complicated relationship and significantly affect the productivity of

A Review of Artificial Intelligence Applications

117

mines [49]. Also, the selection of inappropriate mining equipment can significantly affect the economic efficiency of mines [50], even affect the surrounding environment [51]. There are two approaches to select mining equipment, including traditional and AI methods. As recommended by previous researchers, AI methods often provide better performance than traditional methods, and they should be used to select the optimal mining equipment [52, 53]. In this study, we focus on some AI techniques for mining equipment selection as a review for the feature of AI, as well as future studies. The concept of mining equipment selection using AI techniques is described in Fig. 8.

Fig. 8. The concept of mining equipment selection.

A review of the literature shows that AI techniques have been successfully applied in the selection of mining equipment. Indeed, genetic algorithms were applied to assess the feasibility as well as the compatible mining equipment [54]. Ba¸sçetin, Özta¸s, and Kanli [55] also reviewed computer software (e.g., EQS) based on an AI technique (i.e., fuzzy set theory) for selecting mining equipment. It was then introduced as a useful tool for multiple attribute decision-making, as well as mining equipment selection. In another study, Bazzazi and Karimi [56] used a fuzzy system to decide for mining equipment selection in an open-pit mine, i.e., loading-haulage. The AHP and entropy methods were combined and applied to calculate the weights of attributes. Finally, an optimal loading-haulage was selected and introduced for an open-pit mine of Iran with high reliability. In another study, AGHAJANI BAZZAZI, Osanloo, and Karimi [57] studied and proposed a new fuzzy system to select mining equipment based on three main common factors (i.e., critical, objective, and subjective factors). In this way, time consumption has significantly reduced. Based on the advances in computer science, Ahmad and Mondal [58] also developed an AI system for mining equipment selection with a combination of AHP and mixed-integer non-linear programming (MILP). This system includes two stages, calculating relative weights and finding allocations of each spare-part. Their results showed that the mining equipment with optimal parameters would be better than those of ranking. Based on machine learning algorithms and AI techniques, Samatemba, Zhang, and Besa [59] proposed an optimization algorithm for evaluating and optimizing the effectiveness of mining equipment. Sensitive inputs of the life cycle and overall

118

X.-N. Bui et al.

equipment effectiveness were taken into account and analyzed. Eventually, the effectiveness of mining equipment (e.g., drill rigs, loaders, dump trucks) was optimized aiming to improve the production of the mine. 2.4 Drilling-Blasting Drilling-blasting is an essential operation for rock fragmentation or rock movement in mining, especially in open-pit mines [60, 61]. According to scientists, the effectiveness of rock fragmentation is high, depending on the drilling and blasting parameters [62, 63]. Low effectiveness of rock fragmentation can cause significant effects on the economic, environmental, and other operations (e.g., loading/unloading, transporting, crushing). Hence, to optimize the effectiveness of drilling-blasting operations, AI techniques have also been studied and applied as state-of-the-art tools with high performance. Regarding drilling operation-a high-cost operation in open-pit mines, the performance, as well as the parameters of drilling operations, are primary concerns of engineers. Akin and Karpuz [64] applied an artificial neural network (ANN) to estimate the drilling parameters in shallow carbonates and sandstones formations. Their results were then compared with conventional methods. Finally, the ANN model was introduced as a satisfactory model for this aim. In another study, Bhatnagar and Khandelwal [65] used ANN and multivariate regression analysis (MVRA) methods to evaluate drilling performance. The results with an R2 of 0.985 and MAE of 0.355 of the ANN model revealed that it is a robust AI technique for predicting drilling performance. Drilling troubles were also predicted by Lind and Kabirova [66] using ANN. The researchers found that ANN was a state-of-the-art method for predicting the drilling troubles (Fig. 9). Fattahi and Bazdar [67] proposed various ANN-based models for predicting the drilling rate in open-pit mines. Accordingly, ANN and optimization algorithms (i.e., firefly algorithm (FA), simulated annealing algorithm (SAA), shuffled frog leaping algorithm (SFLA), an invasive weed optimization algorithm (IWO)) were combined for the aim of drilling rate prediction. Drilling parameters, as well as rock mass properties, were taken into account as the influent parameters for predicting the drilling rate. Finally, their results revealed that the SAA-ANN model provided better accuracy than those of the other models. AI techniques have also been taken into account to predict the penetration rate of driller in open-pit mine by Al-AbdulJabbar, Elkatatny, Mahmoud, and Abdulraheem [68]. The fact that the penetration of rate was improved 22% compared with the conventional methods in their study. Sabah, Talebkeikhah, Wood, Khosravanian, Anemangely, and Younesi [69] also applied several AI models to estimate the drilling rate, including ANN, support vector regression (SVR), and ANN-PSO (particle swarm optimization). A superior accuracy was demonstrated on the ANN-PSO model in predicting the drilling rate in their study. However, its performance is rivaled by the SVR model. It is clear that both ANN-PSO and SVR are robust AI techniques for this aim. In a recent study, Liao, Khandelwal, Yang, Koopialipoor and Murlidhar [70] developed a novel hybrid AI model for predicting the drilling rate based on the artificial bee colony (ABC) optimization algorithm and ANN (i.e., ABC-ANN). They claimed that the effectiveness of drilling parameters could be improved using the proposed ABC-ANN model.

A Review of Artificial Intelligence Applications

119

Fig. 9. Drilling points and the trouble predictions of new well using ANN [66]. (a) Clustered oilfield map; (b) Prediction of a new well

Regarding blasting operations in mining, many studies successfully applied AI techniques for optimizing blasting parameters and predicting blast-induced issues. Some applications of AI in blasting operations, as well as their performances, are listed in Table 1.

120

X.-N. Bui et al.

From Table 1, it is clear that AI techniques have been successfully applied in blasting operations. Of those, blast-induced ground vibration was taken into account as the most concern of the researchers since its side effects on the surrounding environment [103– 107]. Based on these AI techniques, blast patterns have been designed and optimized to reduce the side effects of blasting operations [108]. Nevertheless, a review of the literature shows that most of the published studies focused on predicting/estimating the problems induced by blasting only, as illustrated in Fig. 10, and the optimization problems have not been appropriately handled.

Table 1. Applications of AI techniques in blasting and their performances Reference

Blasting issues

AI techniques

Performances

–

ANN

R2 = 0.986; MAE = 0.196

–

–

MLP, GRNN

RMSE = 0.031

–

–

–

ANN

R2 = 0.927; RMSE = 0.071

✔

–

–

–

MLP, BRNN, HYFIS

✔

–

–

–

PSO-ANN

R2 = 0.961; RMSE = 2.319 R2 = 0.890;

Ground vibration

Air over-pressure

Fly-rock

Rock Fragmentation

Back-break

Khandelwal and Singh [71]

✔

–

–

–

Monjezi, Ahmadi, Sheikhan, Bahrami and Salimi [72]

✔

–

–

Monjezi, Ghafurikalajahi and Bahrami [73]

✔

–

Nguyen, Bui, Bui and Mai [74]

–

Hajihassani, Armaghani, Monjezi, Mohamad and Marto [75]

✔

Nguyen and Bui [76]

–

✔

–

–

–

RF-ANN

Marto, Hajihassani, Jahed Armaghani, Tonnizam Mohamad and Makhtar [77]

–

–

✔

–

–

ICA-ANN

Trivedi, Singh and Raina [78]

–

–

✔

–

–

ANN

Nguyen, Bui, Tran, Le and Do [79]

✔

–

–

–

–

ANN

Amiri, Amnieh, Hasanipanah and Khanli [80]

✔

✔

–

–

–

ANN-KNN

R2 = 0.950; RMSE = 1.700

Saadat, Khandelwal and Monjezi [81]

✔

–

–

–

–

MLP

Hasanipanah, Monjezi, Shahnazar, Armaghani and Farazmand [82]

✔

–

–

–

–

SVM

R2 = 0.957; MSE = 0.000722 R2 = 0.957;

MSE = 0.038

R2 = 0.985; RMSE = 0.847 R2 = 0.981; RMSE = 6.582

R2 = 0.983; RMSE = 0.990 R2 = 0.964; RMSE = 0.738

RMSE = 0.340

(continued)

A Review of Artificial Intelligence Applications

121

Table 1. (continued) Reference

Blasting issues

AI techniques

Performances

–

ANN, ANFIS

–

–

ICA-ANN

R2 = 0.980; RMSE = 1.170 R2 = 0.984

–

–

–

HKM-ANN

R2 = 0.983; RMSE = 0.554

–

–

–

–

GP

–

–

–

–

CART

R2 = 0.970; RMSE = 4.700 R2 = 0.950;

Ground vibration

Air over-pressure

Fly-rock

Rock Fragmentation

Back-break

Trivedi, Singh and Gupta [83]

–

–

✔

–

Armaghani, Hajihassani, Marto, Faradonbeh and Mohamad [84]

–

✔

–

Nguyen, Drebenstedt, Bui and Bui [85]

✔

–

Dindarloo [86]

✔

Hasanipanah, Faradonbeh, Amnieh, Armaghani and Monjezi [87]

✔

Zhang, Nguyen, Bui, Tran, Nguyen, Bui and Moayedi [88]

✔

–

–

–

–

PSO-XGBoost

R2 = 0.968; RMSE = 0.583

Monjezi, Baghestani, Faradonbeh, Saghand and Armaghani [89]

✔

–

–

–

–

GEP

R2 = 0.878; RMSE = 3.470

Hasanipanah, Faradonbeh, Armaghani, Amnieh and Khandelwal [90]

–

–

✔

–

–

RT

R2 = 0.872; RMSE = 27.459

Hasanipanah, Shahnazar, Arab, Golzar and Amiri [91]

–

–

–

–

✔

PSO-ANFIS

R2 = 0.922; RMSE = 0.130

Armaghani, Hajihassani, Sohaei, Mohamad, Marto, Motaghedi and Moghaddam [92]

–

✔

–

–

–

ANFIS

R2 = 0.971; RMSE = 2.329

Nguyen, Bui, Bui and Cuong [93]

✔

–

–

–

–

XGBoost

R2 = 0.952; RMSE = 1.742

Arthur, Temeng and Ziggah [94]

✔

–

–

–

–

GPR

Nguyen [95]

✔

–

–

–

–

SVM

R = 0.834; RMSE = 0.157 R2 = 0.924;

Faradonbeh, Armaghani, Amnieh and Mohamad [96]

–

–

✔

–

–

GEP, FFA

R2 = 0.924; RMSE = 29.956

Hasanipanah, Amnieh, Arab and Zamzam [97]

–

–

–

✔

–

PSO-ANFIS

R2 = 0.890; RMSE = 1.310

RMSE = 0.170

RMSE = 0.396

(continued)

122

X.-N. Bui et al. Table 1. (continued)

Reference

Blasting issues

AI techniques

Performances

–

GPR

R2 = 0.948; RMSE = 2.010

✔

–

FFA-ANN

R2 = 0.940; RMSE = 0.100

–

✔

–

FFA-ANFIS

R2 = 0.980; RMSE = 0.520

–

–

✔

–

ACO-BRT

R2 = 0.962; RMSE = 1.643

–

–

✔

–

FFA-ANFIS, GA-ANFIS

R2 = 0.989; RMSE = 0.974

Ground vibration

Air over-pressure

Fly-rock

Rock Fragmentation

Back-break

Gao, Karbasi, Hasanipanah, Zhang and Guo [98]

–

–

–

✔

Asl, Monjezi, Hamidi and Armaghani [99]

–

–

✔

Mojtahedi, Ebtehaj, Hasanipanah, Bonakdari and Amnieh [100]

–

–

Zhang, Bui, Trung, Nguyen and Bui [101]

–

Zhou, Li, Arslan, Hasanipanah and Amnieh [102]

–

Note: MLP (Multiple layers perceptron); GRNN (General regression neural network); BRNN (Bayesian regression neural network); HYFIS (Hybrid fuzzy inference system); PSO (Particle swarm optimization); RF (Random forest); ICA (Imperialist competitive algorithm); KNN (K-nearest neighbors); SVM (Support vector machine); ANFIS (Adaptive neuro-fuzzy inference system); HKM (Hierarchical K-means clustering); GP (Genetic programming); CART (Classification and regression tree); XGBoost (extreme gradient boosting machine); GEP (Gene expression programming); RT (Regression tree); GPR (Gaussian process regression); FFA (Firefly algorithm); ACO (Ant colony optimization); BRT (Boosted regression tree); GA (Genetic algorithm).

Fig. 10. Framework of AI techniques for blasting issues in mining.

A Review of Artificial Intelligence Applications

123

2.5 Slope Stability As one of the critical problems in open-pit mines, slope stability is necessary to be evaluated and predicted carefully and precisely. Similar to the side effects of blasting operations, instability of slope in open-pit mines can make serious effects, even fatal to humans. It disrupts the production and severely affects the economy of mines [109], especially deep open-pit mines [110]. In actual, there are many methods to evaluate the stability of slope in open-pit mines, such as Kinematic [111], 2D and 3D Finite Element Analysis (FEM) [112–114], elastoplastic finite elements [115, 116], limit equilibrium and strength reduction methods [117], upper bound approach [118], numerical manifold [119], Monte Carlo [120], to name a few. Of those, most of them are the numerical and simulation methods, as illustrated in Fig. 11.

Fig. 11. Some evaluations of slope stability based on the 3D FEM technique.

In recent years, AI techniques have also been successfully applied in evaluating and predicting slope stability [121–123]. Indeed, grey systems and ANN have been applied to predict slope stability by Lu and Rosenbaum [124]. They were introduced as a robust tool to predict the movement of the ground in the future with high reliability based on geotechnical properties. Other scholars also conducted similar studies based on ANN for predicting slope stability with a promising result [125–128]. In another study, Kang, Li, and Ma [129] also applied the artificial bee colony (ABC) algorithm to analyze the slope stability. Hoang and Pham [130] also proposed another hybrid AI model to assess slope stability based on the FFA and the Least Squares Support Vector Classification (LS-SVC). The accuracy was then improved by roughly 4% compared with the conventional methods. Based on the optimization algorithms, Luo, Bui, Nguyen, and Moayedi [131] developed a novel hybrid model for predicting slope stability under the combination of the PSO algorithm and the Cubist algorithm (CA), i.e., PSO-CA model.

124

X.-N. Bui et al.

Koopialipoor, Armaghani, Hedayat, Marto, and Gordan [132] developed various hybrid AI models for a similar purpose, including PSO-ANN, GA-ANN, ICA-ANN, and ABCANN models. Finally, the PSO-ANN model was found as the best model in their study for predicting slope stability. Similar studies can be found in the following papers with similar approaches [133–139]. For predicting slope stability using AI techniques, there are two approaches based on the classification and regression problems. For the classification approach, the datasets are often the cases in actual and the history datasets are collected based on the phenomena (i.e., stability or instability). For the regression approach, the AI models were developed based on the analysis results from numerical models or software. Normally, the safety of factor (SOF) is used as the criteria to evaluate the stability of the slopes or benches [140], and it can be modeled and extracted from 2D or 3D finite element analysis. According to Zhou, Cheuk, and Tham [141], the slopes and benches are stable with the SOF > 1. However, Sakellariou and Ferentinou [127] recommended that the SOF should be more than 1.2 to ensure the stability of the slopes. Figure 12 illustrates the approaches (i.e.,

Fig. 12. Two AI approaches in evaluating slope stability (i.e., classification and regression) [131, 142].

A Review of Artificial Intelligence Applications

125

classification and regression) for predicting slope stability, as well as the framework and model assessment methods of these approaches. 2.6 Environmental Issues Regarding environmental issues in mining, they are taken into account as significant concerns of managers, as well as residential areas [143]. Nowadays, environmental issues are more carefully considered in terms of economic efficiency for sustainable development in mining [144–146]. Many countries have closed down open-pit mines because of their massive impact on the environment, especially soil, water, air, and ecosystems [147–149]. The environmental impacts of mining activities seriously have a significant effect on human health, even cause fatal if affected for a long time. Therefore, to overcome these drawbacks, AI techniques have been successfully applied in many fields, as well as smart mines. The concept of AI techniques for environmental issues in mining is presented in Fig. 13.

Fig. 13. Concept of AI techniques for environmental issues in mining.

In underground mines, methane emissions were modeled and predicted by ANN with a promising result [150]. It was considered as a foundation for ventilation in underground mines aiming to reduce gas and toxic, as well as the risk of fire in underground mines. Another similar study was also conducted to predict methane emissions in longwall using ANN [151]. Coal dust and methane explosions were also warned as the most critical factors in underground coal mines [152]. Therefore, the strategies for removing dust were proposed using AI techniques [153–155]. Also, strategies for ventilation in underground mines using AI techniques have been proposed and applied as well. Karacan [156] developed an ANN model for predicting and optimizing ventilation air in underground mines. Hua and Liangshan [157], ZHANG, and DOU [158] also used ANN to assess and control the ventilation system for underground mines with high performance.

126

X.-N. Bui et al.

In open-pit/open-cast mines, environmental impacts include not only air quality but also water and soil quality. Concerning air quality in open-pit mines, dust, gas, toxic, even radiation, are those of the severe impacts on the air environment. Unmanned aerial vehicle (UAV) was used to monitor dust in operations in open-pit mines (e.g., blasting, transporting, crushing, to name a few) [159–162]. Subsequently, hazardous mininginduced were assessed and mapped/predicted by AI techniques based on the datasets collected from UAV [163, 164]. In another study, Patra, Gautam, Majumdar, and Kumar [165] predicted dust concentration in open-cast mine using ANN with high reliability. Bui, Lee, Nguyen, Bui, Long, Le, Nguyen, Nguyen, and Moayedi [166] also proposed a novel AI technique based on the Support Vector Machine for regression problems (SVR) and the Particle Swarm Optimization (PSO) algorithm (i.e., PSO-SVR) to estimate PM10 concentration from drilling operations in open-pit mines with high accuracy. Then, an integrated life cycle inventory and ANN model were established to manage air pollution in the open-pit mine [167]. Nagesha, Kumar, and Singh [168] used ANN to predict PM10 and PM2.5 in open-cast mining operations with high accuracy and reliability. For water pollution, wastewater is considered as a big concern of mining companies, especially metal or mining. It can make dangerous for ecology as well as the health of humans [169, 170]. In recent years, many techniques have been proposed to treat wastewater in mining industries. For example, nanofiltration technology has been applied to treat mining wastewater with high performance [171]. The low-cost absorption techniques for removing heavy metals were also studied and applied in the treatment of mining wastewater [172–175]. Then, AI techniques have been used to estimate or predict the efficiency of heavy metals absorption [176–180]. They were evaluated as successful in predicting, as well as assessing the efficiency of heavy metals absorption with high accuracy. Similar to water pollution, heavy metals can also appear in the soils of mining industries, and it became a severe problem in many countries around the world [181–184]. Although heavy metals can occur in natural soils; however, they were found in industrial, residential, agricultural, and mining areas as well with high contributions [182]. Of those, mining is considered as the most significant sources of heavy metals in soil [185–188]. To improve the quality of soils in mine sites, many techniques were applied, especially heavy metals uptake techniques [189–193]. Subsequently, AI techniques have been applied to predict and optimize the reduction of heavy metals in soils [194–197]. Also, intelligence models were applied to estimate the number of heavy metals in soils with high accuracy, for example, ANN, Adaptive Neuro-Fuzzy Inference System (ANFIS), and Multiple Linear Regression (MLR) [198]. Generally, AI techniques are taken into account as useful techniques in the management and improvement of soil pollution in the mining industry. 2.7 Mineral and Groundwater Potential Mapping In geological engineering, AI techniques have also been widely applied in many areas, such as mineral potential mapping, estimating the components of minerals, landslide susceptibility map, modeling groundwater, to name a few. Of those, mineral and groundwater potential mapping is the most effective application of AI techniques with high

A Review of Artificial Intelligence Applications

127

reliability. An example of mineral and groundwater mapping using AI techniques is shown in Fig. 14.

Fig. 14. Framework of AI models for mineral and groundwater potential mapping (a) Mineral potential mapping [199]; (b) Groundwater potential mapping [200]

In this regard, AI techniques were combined with Geographic Information Systems (GIS) to determine the locations of minerals as well as their components and contents. Spatial data and geo-informatics, remote sensing, properties of the rock mass, geological conditions, to name a few, were used as the input variables for modeling and establishing mineral and groundwater potential maps. Accordingly, Porwal, Carranza, and Hale [201] applied a fuzzy system to predict potentially mineralized zones. Their results confirmed

128

X.-N. Bui et al.

that AI techniques are good candidates to construct mineral and groundwater potential maps. A similar study was also conducted by Porwal, Carranza, and Hale [202] using a radial basis functional link net (RBFLN) with a promising result. In another study, a variety of AI techniques, such as ANN, wavelet neural network (WNN), and SVM, were also developed and used for predicting Cu mineralization [203]. Chen and Wu [204] developed an extreme learning machine (ELM) model for mapping polymetallic prospectively with an accuracy of 82%. Also, many other similar studies were conducted for mineral potential mapping using AI techniques. The studies mainly use different AI models to create mineral maps and compare them as well as evaluate their performance. For groundwater mapping, many AI techniques are also developed and applied with high accuracy. Naghibi, Pourghasemi, and Dixon [205] used a boosted regression tree (BRT), RF, and CART models to produce groundwater spring potential maps. Their results showed that the BRT model provided the highest accuracy in the groundwater potential mapping with an accuracy of 81%. Naghibi, Ahmadi, and Daneshi [206] also applied SVM, RF, and RFGA Genetic Algorithm) to assess groundwater potential by spring locations. The accuracy of the RFGA model obtained 85.6% for potential groundwater mapping. Nhu, Rahmati, Falah, Shojaei, Al-Ansari, Shahabi, Shirzadi, Górski, Nguyen, and Ahmad [207] developed a novel AI system based on bivariate, and multivariate models to establish groundwater spring potential maps with accuracy was up to 85%. Through several above studies, it is clear that AI techniques can be used for potential groundwater mapping with high performance and accuracy. They can explain the relationship between uncertainty factors very well and providing outstanding results.

3 Discussion From the above reviews, it can be seen that AI techniques have been widely applied in mining and geological engineering with greatly improved efficiency. AI applications have leveraged the power of rock mechanics as powerful tools to simulate and predict rock components and stresses. Nowadays, one can simulate the rock mass and its properties with high accuracy using AI techniques [208–210]. This technology is considered as the foundation for designing, developing, and optimizing other technologies related to rock mechanics, such as rock fragmentation, design tunnels, and underground mines, assessment of landslide and slope stability, to name a few. However, some drawbacks of AI techniques have also been shown in previous studies. The accuracy of AI techniques is highly dependent on data collection and analyses [74, 211]. The problems related to over-fitting and under-fitting are big concerns of developers in AI technology [212–214]. Besides, some other aspects of rock mechanics have not been studied and applied using AI techniques to predict and optimize, or the ability to apply has not satisfied scientists. Indeed, AI models for predicting/forecasting and optimizing rockfall seem to be very rare, and the accuracy is lowly [215]. For cracks in the rock, this is an essential parameter in assessing the properties of the rock mechanics as well as predicting the stability of the slopes and other works-related. However, previous researches just only stopped at simulating the fracture system of rock mass without prediction and evaluation by AI techniques. As for the rockburst phenomenon, the accuracy of the AI methods was just over 65% and had not satisfied scientists. Meanwhile, rock mechanics and uncertainty factors play an essential role in the classification and prediction of rockburst intensity.

A Review of Artificial Intelligence Applications

129

Concerning mining method selection, even though the concept was shown in Fig. 6 and many researchers have successfully applied AI in this field. Nevertheless, previous and current studies have not been thoroughly combined with the criteria (i.e., main criteria, sub-criteria, and mine criteria) to select the most proper mining method. To solve this problem, a detailed database with high precision of criteria, as well as operations in mining, is necessary. Also, they should be combined with AI techniques in mine planning, as well as optimization of mine planning and boundaries [216–218] to select the most proper mining method. For mining equipment selection, although the reviews showed that AI techniques had been successfully applied in this field. However, previous and current studies just only focused on the pairs of equipment related to loading and transporting, e.g., loadinghaulage, truck-shovel. Whereas, many types of equipment can be used in mining, as mentioned in Fig. 7, and they have a significant effect on the productivity of mines. Thus, researches on the mining equipment selection with various types of equipment (e.g., driller, truck, shovel, conveyer, railway) using AI techniques is challenging for engineers and researchers. Remarkable, mining equipment selection and optimization using AI techniques are taken into account as a useful solution for hybrid mines (combining open-pit and underground mine sites) aiming to increase productivity, decrease costs, and reduce the impacts on the surrounding environment. For drilling-blasting operations, it can be seen that AI technologies are pretty complete. They provided AI models that can predict blast-induced ground vibration, air over-pressure, fly-rock, rock fragmentation, and back-break with high accuracy. However, the geological conditions-related parameters still seem to be a hard problem with scientists in predicting blasting issues since the uncertainty factors of the rock environment. Besides, the development of novel AI models for blasting issues is still necessary to contribute to the knowledge of AI, and providing novel AI models that can predict blasting issues in many areas with high accuracy. Similar to drilling-blasting operations in mining, applications of AI in evaluating and predicting slope stability has met the demands of the assessment and forecast of slope stability with high accuracy. However, novel AI models and technologies using satellite image, UAV combined with AI models to evaluate and predict slope stability is necessary for the future. Also, the combination of GIS with AI techniques in drillingblasting operations is a potential solution for other studies in the future aiming to improve the accuracy of drilling-blasting operations, as well as their impacts on the surrounding environment. The problems related to the environment in mining were also investigated, evaluated, and forecasted by AI models, as reviewed above. However, AI models for predicting/forecasting air quality in mines (i.e., open-pit and underground mines) seem not to be conducted. Air quality controlling systems for mining using the Internet of Things (IoT) and AI techniques have not also been proposed. These systems are considered as an essential problem in assessing and forecasting air quality of mines. It is an important criterion to reduce the negative impacts on the surrounding air environment and human health. Soil environment has not been paid much attention in addition to heavy metal pollution. AI models for evaluating and predicting the concentration of heavy metals

130

X.-N. Bui et al.

in the soil, as well as the effectiveness of soil heavy metal handling systems are imperfect. Therefore, future research should focus on analyzing and forecasting heavy metal pollution and the ability to treat them with heavy metals uptake systems based on AI techniques. Water environment issues also need more attention and in-depth study. AI techniques in assessing, analyzing, and forecasting the quality of wastewater, as well as the quality of water after being treated in mines, are necessary to ensure the impacts of wastewater in mines on the surrounding environment are minimal. Concerning AI applications in geological engineering, scientists have been very successful in using them for mineral and groundwater potential mapping, as reviewed above. The accuracy of AI techniques was demonstrated to be higher than traditional methods. Also, combining GIS and AI models have increased the accuracy of AI models not only in terms of mineral/groundwater content but also in terms of spatial and location accuracy, as well as their distribution positions. However, researches on the development of novel AI models for geological engineering is still needed in the future, especially new AI applications for geological engineering in deep-sea mining, prediction of composition and content of minerals, groundwater, to name a few.

4 Conclusion In this paper, applications of AI in mining and geological engineering were reviewed. Their advantages and disadvantages were mentioned and discussed in this paper. Furthermore, the potential solutions, as well as future research, have been indicated as well. This paper showed that AI techniques had been widely applied with high performance in the mining industry. The ANN-based and hybrid AI models are state-of-the-art techniques with high accuracy that is the goal of researchers in recent years. Besides, precise data and information are the indispensable requirement of AI techniques. It decides the accuracy and reliability of AI models in practical engineering. However, engineers and researchers often have difficulty in this regard due to uncertainties and impreciseness previously described. Therefore, data needs to be accurately collected with high reliability to improve the accuracy of AI models in mining and geological engineering. In general, AI techniques are advanced techniques that play a large role in mining and geological engineering. They should be used in practical engineering to improve the performance, as well as minimize impacts on the surrounding environment. Acknowledgments. The authors would like to thank the Center for Mining, Electro-Mechanical research of Hanoi University of Mining and Geology (HUMG), Hanoi, Vietnam, and the research team of Innovations for Sustainable and Responsible Mining (ISRM) of HUMG.

References 1. Balzani, V., Armaroli, N.: Energy for a Sustainable World: From the Oil Age to a SunPowered Future. Wiley, Hoboken (2010) 2. Fujii, H., Takahashi, K., Yamashita, Y.: Application of titanium and its alloys for automobile parts. Shinnittetsu giho 62–67 (2003)

A Review of Artificial Intelligence Applications

131

3. Schauerte, O.: Titanium in automotive production. Adv. Eng. Mater. 5, 411–418 (2003) 4. Liedl, G., Bielak, R., Ivanova, J., Enzinger, N., Figner, G., Bruckner, J., Pasic, H., Pudar, M., Hampel, S.: Joining of aluminum and steel in car body manufacturing. Phys. Procedia 12, 150–156 (2011) 5. Wu, S., Yan, T., Kuai, Z., Pan, W.: Thermal conductivity enhancement on phase change materials for thermal energy storage: a review. Energy Storage Materials (2019) 6. Jang, H., Topal, E.: A review of soft computing technology applications in several mining problems. Appl. Soft Comput. 22, 638–651 (2014) 7. Azcue, J.M.: Environmental Impacts of Mining Activities: Emphasis on Mitigation and Remedial Measures. Springer, Heidelberg (2012) 8. Tacher, L., Pomian-Srzednicki, I., Parriaux, A.: Geological uncertainties associated with 3-D subsurface models. Comput. Geosci. 32, 212–221 (2006) 9. Baghbanan, A., Kefayati, S., Torkan, M., Hashemolhosseini, H., Narimani, R.: Numerical probabilistic analysis for slope stability in fractured rock masses using DFN-DEM approach. Int. J. Mining Geo-Eng. 51, 79–90 (2017) 10. Ehlers, L.B.: Uncertainties in groundwater-surface water modelling for the HOBE catchment. Department of Geosciences and Natural Resource Management, Faculty of … (2018) 11. Nguyen, N.V.: Building DEM for deep open-pit coal mines using DJI Inspire 2 (in Vietnames). J. Mining Earth Sci. 61, 1–10 (2020) 12. Bitarafan, M., Ataei, M.: Mining method selection by multiple criteria decision making tools. J. Southern Afr. Inst. Mining Metal. 104, 493–498 (2004) 13. Domingues, M.S., Baptista, A.L., Diogo, M.T.: Engineering complex systems applied to risk management in the mining industry. Int. J. Mining Sci. Technol. 27, 611–616 (2017) 14. Newman, A.M., Rubio, E., Caro, R., Weintraub, A., Eurek, K.: A review of operations research in mine planning. Interfaces 40, 222–245 (2010) 15. Dimitrakopoulos, R., Martinez, L., Ramazan, S.: A maximum upside/minimum downside approach to the traditional optimization of open pit mine design. J. Mining Sci. 43, 73–82 (2007) 16. Gali´c, I., Jankovi´c, B., Mrakovˇci´c, I.: An another way for open pit mine design optimization– floating slopes method. Rudarsko-geološko-naftni zbornik 21, 103–111 (2009) 17. Kahriman, A., Ozer, U., Aksoy, M., Karadogan, A., Tuncer, G.: Environmental impacts of bench blasting at Hisarcik Boron open pit mine in Turkey. Environ. Geol. 50, 1015–1023 (2006) 18. Duan, B., Xia, H., Yang, X.: Impacts of bench blasting vibration on the stability of the surrounding rock masses of roadways. Tunnelling Underground Space Technol. 71, 605–622 (2018) 19. Torres, V.N., Silveira, L.G., Lopes, P.F., de Lima, H.M.: Assessing and controlling of bench blasting-induced vibrations to minimize impacts to a neighboring community. J. Cleaner Prod. 187, 514–524 (2018) 20. Duncan, J.M., Wright, S.G., Brandon, T.L.: Soil Strength and Slope Stability. Wiley, Hoboken (2014) 21. Sjöberg, J.: Large scale slope stability in open pit mining: a review. Luleå tekniska universitet (1996) 22. Zahri, F., Boukelloul, M.L., Hadji, R., Talhi, K.: Slope stability analysis in open pit mines of Jebel Gustar career, NE Algeria–a multi-steps approach. Mining Sci. 23 (2016) 23. Hobba Jr., W.A.: Effects of underground mining and mine collapse on the hydrology of selected basins in West Virginia. USGPO; US Geological Survey, Map Distribution (1993) 24. Stevens, P.R.: A review of the effects of earthquakes on underground mines. US Geological Survey (1977)

132

X.-N. Bui et al.

25. Karacan, C.Ö., Ruiz, F.A., Cotè, M., Phipps, S.: Coal mine methane: a review of capture and utilization practices with benefits to mining safety and to greenhouse gas reduction. Int. J. Coal Geol. 86, 121–156 (2011) 26. Diamond, W.P.: Methane Control for Underground Coal Mines: Chapter 11 (1993) 27. Kurnia, J.C., Sasmito, A.P., Mujumdar, A.S.: CFD simulation of methane dispersion and innovative methane management in underground mining faces. Appl. Math. Modelling 38, 3467–3484 (2014) 28. Jing, L.: A review of techniques, advances and outstanding issues in numerical modelling for rock mechanics and rock engineering. Int. J. Rock Mech. Mining Sci. 40, 283–353 (2003) 29. Nguyen, A.T., Pham, V.V., Le, V.Q., Nguyen, T.A., Le, T.H.: Determining for an output capacity of dimension stone exploitation from the computer simulations to generate the fracture network in 3D: case study in some dimensional stone quarries in Vietnam. In˙zynieria Mineralna 21 (2019) 30. Sonmez, H., Gokceoglu, C., Nefeslioglu, H., Kayabasi, A.: Estimation of rock modulus: for intact rocks with an artificial neural network and for rock masses with a new empirical equation. Int. J. Rock Mech. Mining Sci. 43, 224–235 (2006) 31. Ocak, I., Seker, S.E.: Estimation of elastic modulus of intact rocks by artificial neural network. Rock Mech. Rock Eng. 45, 1047–1054 (2012) 32. Majdi, A., Beiki, M.: Evolving neural network using a genetic algorithm for predicting the deformation modulus of rock masses. Int. J. Rock Mech. Mining Sci. 47, 246–253 (2010) 33. Rukhaiyar, S., Samadhiya, N.: A polyaxial strength model for intact sandstone based on Artificial Neural Network. Int. J. Rock Mech. Mining Sci. 95, 26–47 (2017) 34. Rashidi, M., Hajipour, M., Asadi, A.: Correlation between static and dynamic elastic modulus of limestone formations using artificial neural networks. In: 52nd US Rock Mechanics/Geomechanics Symposium. American Rock Mechanics Association 35. Dong, L.-J., Li, X.-B., Kang, P.: Prediction of rockburst classification using Random Forest. Trans. Nonferrous Metals Soc. China 23, 472–477 (2013) 36. Zhou, K., Yun, L., Deng, H., Li, J., Liu, C.: Prediction of rock burst classification using cloud model with entropy weight. Trans. Nonferrous Metals Soc. China 26, 1995–2002 (2016) 37. Shirani Faradonbeh, R., Taheri, A.: Long-term prediction of rockburst hazard in deep underground openings using three robust data mining techniques. Eng. Comput. 35(2), 659–675 (2018). https://doi.org/10.1007/s00366-018-0624-4 38. Zhou, J., Koopialipoor, M., Li, E., Armaghani, D.J.: Prediction of rockburst risk in underground projects developing a neuro-bee intelligent system. Bull. Eng. Geol. Environ. (2020) 39. Dou, L., Lu, C., Mu, Z., Gao, M.: Prevention and forecasting of rock burst hazards in coal mines. Mining Sci. Technol. (China) 19, 585–591 (2009) 40. Yun, Q., Huang, G.: A fuzzy set approach to the selection of mining method. Mining Sci. Technol. 6, 9–16 (1987) 41. Guray, C., Celebi, N., Atalay, V., Pasamehmetoglu, A.G.: Ore-age: a hybrid system for assisting and teaching mining method selection. Expert Syst. Appl. 24, 261–271 (2003) 42. Naghadehi, M.Z., Mikaeil, R., Ataei, M.: The application of fuzzy analytic hierarchy process (FAHP) approach to selection of optimum underground mining method for Jajarm Bauxite Mine Iran. Expert Syst. Applications 36, 8218–8226 (2009) 43. Azadeh, A., Osanloo, M., Ataei, M.: A new approach to mining method selection based on modifying the Nicholas technique. Appl. Soft Comput. 10, 1040–1061 (2010) 44. Nicholas, H.: Method selection-A numerical approach, Design and operation of mines methods. AIME 45. Özfırat, M.: A fuzzy method for selecting underground coal mining method considering mechanization criteria. J. Mining Sci. 48, 533–544 (2012)

A Review of Artificial Intelligence Applications

133

46. Dehghani, H., Siami, A., Haghi, P.: A new model for mining method selection based on grey and TODIM methods. J. Mining Environ. 8, 49–60 (2017) 47. Fu, Z., Wu, X., Liao, H., Herrera, F.: Underground mining method selection with the hesitant fuzzy linguistic gained and lost dominance score method. IEEE Access 6, 66442–66458 (2018) 48. Liang, W., Zhao, G., Hong, C.: Selecting the optimal mining method with extended multiobjective optimization by ratio analysis plus the full multiplicative form (MULTIMOORA) approach. Neural Comput. Appl. 31(10), 5871–5886 (2018). https://doi.org/10.1007/s00 521-018-3405-5 49. Burt, C.N., Caccetta, L.: Equipment selection for surface mining: a review. Interfaces 44, 143–162 (2014) 50. Blackwell, G.H.: Estimation of large open pit haulage truck requirements. CIM Bull. 143– 149 (1999) 51. Bozorgebrahimi, E., Hall, R., Blackwell, G.: Sizing equipment for open pit mining–a review of critical parameters. Mining Technol. 112, 171–179 (2003) 52. Bascetin, A., Kesimal, A.: The study of a fuzzy set theory for the selection of an optimum coal transportation system from pit to the power plant. Int. J. Surface Mining Reclamation Environ. 13, 97–101 (1999) 53. Haidar, A., Naoum, S., Howes, R., Tah, J.: Genetic algorithms application and testing for equipment selection. J. Constr. Eng. Manag. 125, 32–38 (1999) 54. Vagenas, N., Nuziale, T.: Genetic algorithms for reliability assessment of mining equipment. J. Qual. Maintenance Eng. (2001) 55. Ba¸sçetin, A., Özta¸s, O., Kanli, A.: EQS: a computer software using fuzzy logic for equipment selection in mining engineering. J. South Afr. Inst. Min. Metall. 106, 63–70 (2006) 56. Bazzazi, A.A., Karimi, B.: Optimal open pit mining equipment selection using fuzzy multiple attribute decision making approach. In: Seventeenth International Symposium on Mine Planning and Equipment Selection (MPES 2008)(第十七届矿山规划和设备选型国际研 讨会), pp. 253–268. 中国有色金属工业协会; 中国钢铁工业协会; 中国冶金矿山企业协 会 (2008) 57. Aghajani Bazzazi, A., Osanloo, M., Karimi, B.: A new fuzzy multi criteria decision making model for open pit mines equipment selection. Asia-Pacific J. Oper. Res. 28, 279–300 (2011) 58. Ahmad, M.T., Mondal, S.: Dynamic supplier selection approach for mining equipment company. J. Model. Manag. (2019) 59. Samatemba, B., Zhang, L., Besa, B.: Evaluating and optimizing the effectiveness of mining equipment; the case of Chibuluma South underground mine. J. Clean. Prod. 252, 119697 (2020) 60. Gokhale, B.V.: Rotary drilling and blasting in large surface mines. CRC Press (2010) 61. Hrehová, D., Cehlár, M., Rybár, R., Mitterpachová, N.: Mining technology with drillingblasting operations. Int. Multidisc. Sci. GeoConference: SGEM: Surv. Geol. Mining Ecol. Manag. 1, 675 (2012) 62. AyalaCarcedo, F.: Drilling and Blasting of Rocks. Routledge (2017) 63. Rustan, A., Cunningham, C., Fourney, W., Spathis, A., Simha, K.: Mining and Rock Construction Technology Desk Reference: Rock Mechanics, Drilling & Blasting. CRC Press (2010) 64. Akin, S., Karpuz, C.: Estimating drilling parameters for diamond bit drilling operations using artificial neural networks. Int. J. Geomech. 8, 68–73 (2008) 65. Bhatnagar, A., Khandelwal, M.: An intelligent approach to evaluate drilling performance. Neural Comput. Appl. 21, 763–770 (2012) 66. Lind, Y.B., Kabirova, A.R.: Artificial neural networks in drilling troubles prediction. In: SPE Russian Oil and Gas Exploration & Production Technical Conference and Exhibition. Society of Petroleum Engineers, (2014)

134

X.-N. Bui et al.

67. Fattahi, H., Bazdar, H.: Applying improved artificial neural network models to evaluate drilling rate index. Tunn. Undergr. Space Technol. 70, 114–124 (2017) 68. Al-AbdulJabbar, A., Elkatatny, S., Mahmoud, M., Abdulraheem, A.: Predicting rate of penetration using artificial intelligence techniques. In: SPE Kingdom of Saudi Arabia Annual Technical Symposium and Exhibition. Society of Petroleum Engineers (2018) 69. Sabah, M., Talebkeikhah, M., Wood, D.A., Khosravanian, R., Anemangely, M., Younesi, A.: A machine learning approach to predict drilling rate using petrophysical and mud logging data. Earth Sci. Inf. 12(3), 319–339 (2019). https://doi.org/10.1007/s12145-019-00381-4 70. Liao, X., Khandelwal, M., Yang, H., Koopialipoor, M., Murlidhar, B.R.: Effects of a proper feature selection on prediction and optimization of drilling rate using intelligent techniques. Eng. Comput. 36(2), 499–510 (2019). https://doi.org/10.1007/s00366-019-00711-6 71. Khandelwal, M., Singh, T.: Prediction of blast-induced ground vibration using artificial neural network. Int. J. Rock Mech. Min. Sci. 46, 1214–1222 (2009) 72. Monjezi, M., Ahmadi, M., Sheikhan, M., Bahrami, A., Salimi, A.: Predicting blast-induced ground vibration using various types of neural networks. Soil Dyn. Earthquake Eng. 30, 1233–1236 (2010) 73. Monjezi, M., Ghafurikalajahi, M., Bahrami, A.: Prediction of blast-induced ground vibration using artificial neural networks. Tunn. Undergr. Space Technol. 26, 46–50 (2011) 74. Nguyen, H., Bui, X.-N., Bui, H.-B., Mai, N.-L.: A comparative study of artificial neural networks in predicting blast-induced air-blast overpressure at Deo Nai open-pit coal mine Vietnam. Neural Comput. Appl. 32, 3939–3955 (2020) 75. Hajihassani, M., Jahed Armaghani, D., Monjezi, M., Mohamad, E.T., Marto, A.: Blastinduced air and ground vibration prediction: a particle swarm optimization-based artificial neural network approach. Environ. Earth Sci. 74(4), 2799–2817 (2015). https://doi.org/10. 1007/s12665-015-4274-1 76. Nguyen, H., Bui, X.-N.: Predicting blast-induced air overpressure: a robust artificial intelligence system based on artificial neural networks and random forest. Nat. Resour. Res. 28, 893–907 (2019) 77. Marto, A., Hajihassani, M., Jahed Armaghani, D., Tonnizam Mohamad, E., Makhtar, A.M.: A novel approach for blast-induced flyrock prediction based on imperialist competitive algorithm and artificial neural network. Sci. World J. 2014 (2014) 78. Trivedi, R., Singh, T., Raina, A.: Prediction of blast-induced flyrock in Indian limestone mines using neural networks. J. Rock Mech. Geotech. Eng. 6, 447–454 (2014) 79. Nguyen, H., Bui, X.-N., Tran, Q.-H., Le, T.-Q., Do, N.-H.: Evaluating and predicting blastinduced ground vibration in open-cast mine using ANN: a case study in Vietnam. SN Appl. Sci. 1, 125 (2019) 80. Amiri, M., Amnieh, H.B., Hasanipanah, M., Khanli, L.M.: A new combination of artificial neural network and K-nearest neighbors models to predict blast-induced ground vibration and air-overpressure. Eng. Comput. 32, 631–644 (2016) 81. Saadat, M., Khandelwal, M., Monjezi, M.: An ANN-based approach to predict blast-induced ground vibration of Gol-E-Gohar iron ore mine Iran. J. Rock Mech. Geotech. Eng. 6, 67–76 (2014) 82. Hasanipanah, M., Monjezi, M., Shahnazar, A., Armaghani, D.J., Farazmand, A.: Feasibility of indirect determination of blast induced ground vibration based on support vector machine. Measurement 75, 289–297 (2015) 83. Trivedi, R., Singh, T., Gupta, N.: Prediction of blast-induced flyrock in opencast mines using ANN and ANFIS. Geotech. Geol. Eng. 33, 875–891 (2015) 84. Jahed Armaghani, D., Hajihassani, M., Marto, A., Shirani Faradonbeh, R., Mohamad, E.T.: Prediction of blast-induced air overpressure: a hybrid AI-based predictive model. Environ. Monit. Assess. 187(11), 1–13 (2015). https://doi.org/10.1007/s10661-015-4895-6

A Review of Artificial Intelligence Applications

135

85. Nguyen, H., Drebenstedt, C., Bui, X.-N., Bui, D.T.: Prediction of blast-induced ground vibration in an open-pit mine by a novel hybrid model based on clustering and artificial neural network. Nat. Resour. Res. 29, 691–709 (2019) 86. Dindarloo, S.R.: Prediction of blast-induced ground vibrations via genetic programming. Int. J. Mining Sci. Technol. 25, 1011–1015 (2015) 87. Hasanipanah, M., Faradonbeh, R.S., Amnieh, H.B., Armaghani, D.J., Monjezi, M.: Forecasting blast-induced ground vibration developing a CART model. Eng. Comput. 33(2), 307–316 (2016). https://doi.org/10.1007/s00366-016-0475-9 88. Zhang, X., Nguyen, H., Bui, X.-N., Tran, Q.-H., Nguyen, D.-A., Bui, D.T., Moayedi, H.: Novel soft computing model for predicting blast-induced ground vibration in open-pit mines based on particle swarm optimization and XGBoost. Nat. Resour. Res. 29(2), 711–721 (2019). https://doi.org/10.1007/s11053-019-09492-7 89. Monjezi, M., Baghestani, M., Shirani Faradonbeh, R., Pourghasemi Saghand, M., Jahed Armaghani, D.: Modification and prediction of blast-induced ground vibrations based on both empirical and computational techniques. Eng. Comput. 32(4), 717–728 (2016). https:// doi.org/10.1007/s00366-016-0448-z 90. Hasanipanah, M., Faradonbeh, R.S., Armaghani, D.J., Amnieh, H.B., Khandelwal, M.: Development of a precise model for prediction of blast-induced flyrock using regression tree technique. Environ. Earth Sci. 76(1), 1–10 (2016). https://doi.org/10.1007/s12665-0166335-5 91. Hasanipanah, M., Shahnazar, A., Arab, H., Golzar, S.B., Amiri, M.: Developing a new hybrid-AI model to predict blast-induced backbreak. Eng. Comput. 33(3), 349–359 (2016). https://doi.org/10.1007/s00366-016-0477-7 92. Jahed Armaghani, D., Hajihassani, M., Sohaei, H., Tonnizam Mohamad, E., Marto, A., Motaghedi, H., Moghaddam, M.R.: Neuro-fuzzy technique to predict air-overpressure induced by blasting. Arab. J. Geosci. 8(12), 10937–10950 (2015). https://doi.org/10.1007/ s12517-015-1984-3 93. Nguyen, H., Bui, X.-N., Bui, H.-B., Cuong, D.T.: Developing an XGBoost model to predict blast-induced peak particle velocity in an open-pit mine: a case study. Acta Geophys. 67(2), 477–490 (2019). https://doi.org/10.1007/s11600-019-00268-4 94. Arthur, C.K., Temeng, V.A., Ziggah, Y.Y.: Novel approach to predicting blast-induced ground vibration using Gaussian process regression. Eng. Comput. 36(1), 29–42 (2019). https://doi. org/10.1007/s00366-018-0686-3 95. Nguyen, H.: Support vector regression approach with different kernel functions for predicting blast-induced ground vibration: a case study in an open-pit coal mine of Vietnam. SN Appl. Sci. 1, 283 (2019) 96. Faradonbeh, R.S., Armaghani, D.J., Amnieh, H.B., Mohamad, E.T.: Prediction and minimization of blast-induced flyrock using gene expression programming and firefly algorithm. Neural Comput. Appl. 29(6), 269–281 (2016). https://doi.org/10.1007/s00521-016-2537-8 97. Hasanipanah, M., Amnieh, H.B., Arab, H., Zamzam, M.S.: Feasibility of PSO–ANFIS model to estimate rock fragmentation produced by mine blasting. Neural Comput. Appl. 30(4), 1015–1024 (2016). https://doi.org/10.1007/s00521-016-2746-1 98. Gao, W., Karbasi, M., Hasanipanah, M., Zhang, X., Guo, J.: Developing GPR model for forecasting the rock fragmentation in surface mines. Eng. Comput. 34(2), 339–345 (2017). https://doi.org/10.1007/s00366-017-0544-8 99. Asl, P.F., Monjezi, M., Hamidi, J.K., Armaghani, D.J.: Optimization of flyrock and rock fragmentation in the Tajareh limestone mine using metaheuristics method of firefly algorithm. Eng. Comput. 34(2), 241–251 (2017). https://doi.org/10.1007/s00366-017-0535-9 100. Mojtahedi, S.F.F., Ebtehaj, I., Hasanipanah, M., Bonakdari, H., Amnieh, H.B.: Proposing a novel hybrid intelligent model for the simulation of particle size distribution resulting from blasting. Eng. Comput. 35, 47–56 (2018)

136

X.-N. Bui et al.

101. Zhang, S., Bui, X.-N., Trung, N.-T., Nguyen, H., Bui, H.-B.: Prediction of rock size distribution in mine bench blasting using a novel ant colony optimization-based boosted regression tree technique. Nat. Resour. Res. 29, 867–886 (2020) 102. Zhou, J., Li, C., Arslan, C.A., Hasanipanah, M., Bakhshandeh Amnieh, H.: Performance evaluation of hybrid FFA-ANFIS and GA-ANFIS models to predict particle size distribution of a muck-pile after blasting. Eng. Comput. 2, 1–10 (2019) 103. Bui, X.-N., Choi, Y., Atrushkevich, V., Nguyen, H., Tran, Q.-H., Long, N.Q., Hoang, H.-T.: Prediction of blast-induced ground vibration intensity in open-pit mines using unmanned aerial vehicle and a novel intelligence system. Nat. Resour. Res. 29(2), 771–790 (2019). https://doi.org/10.1007/s11053-019-09573-7 104. Bui, X.-N., Jaroonpattanapong, P., Nguyen, H., Tran, Q.-H., Long, N.Q.: A novel Hybrid Model for predicting Blast-induced Ground Vibration Based on k-nearest neighbors and particle Swarm optimization. Sci. Rep. 9, 1–14 (2019) 105. Nguyen, H., Bui, X.-N., Tran, Q.-H., Mai, N.-L.: A new soft computing model for estimating and controlling blast-produced ground vibration based on hierarchical K-means clustering and cubist algorithms. Appl. Soft Comput. 77, 376–386 (2019) 106. Nguyen, H., Choi, Y., Bui, X.-N., Nguyen-Thoi, T.: Predicting blast-induced ground vibration in open-pit mines using vibration sensors and support vector regression-based optimization algorithms. Sensors 20, 132 (2019) 107. Shang, Y., Nguyen, H., Bui, X.-N., Tran, Q.-H., Moayedi, H.: A novel artificial intelligence approach to predict blast-induced ground vibration in open-pit mines based on the firefly algorithm and artificial neural network. Nat. Resour. Res. 29(2), 723–737 (2019). https:// doi.org/10.1007/s11053-019-09503-7 108. Amnieh, H.B., Siamaki, A., Soltani, S.: Design of blasting pattern in proportion to the peak particle velocity (PPV): artificial neural networks approach. Saf. Sci. 50, 1913–1916 (2012) 109. Naghadehi, M.Z., Jimenez, R., KhaloKakaie, R., Jalali, S.-M.E.: A new open-pit mine slope instability index defined using the improved rock engineering systems approach. Int. J. Rock Mech. Min. Sci. 61, 1–14 (2013) 110. Stacey, T., Xianbin, Y., Armstrong, R., Keyter, G.: New slope stability considerations for deep open pit mines. J. South Afr. Inst. Min. Metall. 103, 373–389 (2003) 111. Kliche, C.A.: Rock slope stability (1999) 112. Griffiths, D., Lane, P.: Slope stability analysis by finite elements. Geotechnique 49, 387–403 (1999) 113. Nian, T.-K., Huang, R.-Q., Wan, S.-S., Chen, G.-Q.: Three-dimensional strength-reduction finite element analysis of slopes: geometric effects. Can. Geotech. J. 49, 574–588 (2012) 114. Jeremi´c, B.: Finite element methods for 3D slope stability analysis. Slope Stability 2000, 224–238 (2000) 115. Griffiths, D., Lu, N.: Unsaturated slope stability analysis with steady infiltration or evaporation using elasto-plastic finite elements. Int. J. Numer. Anal. Meth. Geomech. 29, 249–267 (2005) 116. Zheng, H., Liu, D., Li, C.: Slope stability analysis based on elasto-plastic finite element method. Int. J. Numer. Meth. Eng. 64, 1871–1888 (2005) 117. Cheng, Y., Lansivaara, T., Wei, W.: Two-dimensional slope stability analysis by limit equilibrium and strength reduction methods. Comput. Geotech. 34, 137–150 (2007) 118. Chen, Z., Wang, X., Haberfield, C., Yin, J.-H., Wang, Y.: A three-dimensional slope stability analysis method using the upper bound theorem: Part I: theory and methods. Int. J. Rock Mech. Min. Sci. 38, 369–378 (2001) 119. Ning, Y., An, X., Ma, G.: Footwall slope stability analysis with the numerical manifold method. Int. J. Rock Mech. Min. Sci. 48, 964–975 (2011)

A Review of Artificial Intelligence Applications

137

120. Cao, Z., Wang, Y., Li, D.: Practical reliability analysis of slope stability by advanced Monte Carlo simulations in a spreadsheet. Probabilistic Approaches for Geotechnical Site Characterization and Slope Stability Analysis, pp. 147–167. Springer, Heidelberg (2017) 121. Sari, P.A., Suhatril, M., Osman, N., Mu’azu, M.A., Dehghani, H., Sedghi, Y., Safa, M., Hasanipanah, M., Wakil, K., Khorami, M., Djuric, S.: An intelligent based-model role to simulate the factor of safe slope by support vector regression. Eng. Comput. 35(4), 1521– 1531 (2018). https://doi.org/10.1007/s00366-018-0677-4 122. Toghroli, A., Darvishmoghaddam, E., Zandi, Y., Parvan, M., Safa, M., Abdullahi, M.M., Heydari, A., Wakil, K., Gebreel, S.A., Khorami, M.: Evaluation of the parameters affecting the Schmidt rebound hammer reading using ANFIS method. Comput. Concrete 21, 525–530 (2018) 123. Safa, M., Sari, P.A., Shariati, M., Suhatril, M., Trung, N.T., Wakil, K., Khorami, M.: Development of neuro-fuzzy and neuro-bee predictive models for prediction of the safety factor of eco-protection slopes. Physica A: Stat. Mech. Appl. 124046 (2020) 124. Lu, P., Rosenbaum, M.: Artificial neural networks and grey systems for the prediction of slope stability. Nat. Hazards 30, 383–398 (2003) 125. Choobbasti, A., Farrokhzad, F., Barari, A.: Prediction of slope stability using artificial neural network (case study: Noabad, Mazandaran, Iran). Arab. J. Geosci. 2, 311–319 (2009) 126. Yuanyou, X., Xinping, L., Ruigeng, Z.: Engineering geology evaluation method for slope stability based on artificial neural network. Rock Soil Mech. 3 (1996) 127. Sakellariou, M., Ferentinou, M.: A study of slope stability prediction using neural networks. Geotech. Geol. Eng. 23, 419 (2005) 128. Chakraborty, A., Goswami, D.: Prediction of slope stability using multiple linear regression (MLR) and artificial neural network (ANN). Arab. J. Geosci. 10(17), 1–11 (2017). https:// doi.org/10.1007/s12517-017-3167-x 129. Kang, F., Li, J., Ma, Z.: An artificial bee colony algorithm for locating the critical slip surface in slope stability analysis. Eng. Optim. 45, 207–223 (2013) 130. Hoang, N.-D., Pham, A.-D.: Hybrid artificial intelligence approach based on metaheuristic and machine learning for slope stability assessment: a multinational data analysis. Expert Syst. Appl. 46, 60–68 (2016) 131. Luo, Z., Bui, X.-N., Nguyen, H., Moayedi, H.: A novel artificial intelligence technique for analyzing slope stability using PSO-CA model. Eng. Comput. 17, 1–12 (2019) 132. Koopialipoor, M., Jahed Armaghani, D., Hedayat, A., Marto, A., Gordan, B.: Applying various hybrid intelligent systems to evaluate and predict slope stability under static and dynamic conditions. Soft. Comput. 23(14), 5913–5929 (2018) 133. Chebrolu, A., Sasmal, S.K., Behera, R.N., Das, S.K.: Prediction of factor of safety for slope stability using advanced artificial intelligence techniques. In: Advanced Engineering Optimization Through Intelligent Techniques, pp. 173–181. Springer, Heidelberg (2020) 134. Zhou, Y.: Application of artificial neural network in slope stability analysis. Comput. Softw. Media Appl. 2, 16–19 (2019) 135. Gordan, B., Jahed Armaghani, D., Hajihassani, M., Monjezi, M.: Prediction of seismic slope stability through combination of particle swarm optimization and neural network. Eng. Comput. 32(1), 85–97 (2015). https://doi.org/10.1007/s00366-015-0400-7 136. Kang, F., Xu, B., Li, J., Zhao, S.: Slope stability evaluation using Gaussian processes with various covariance functions. Appl. Soft Comput. 60, 387–396 (2017) 137. Mishra, M., Ramana, G.V., Maity, D.: Multiverse optimisation algorithm for capturing the critical slip surface in slope stability analysis. Geotech. Geol. Eng. 38, 459–474 (2020) 138. Qi, C., Tang, X.: Slope stability prediction using integrated metaheuristic and machine learning approaches: a comparative study. Comput. Ind. Eng. 118, 112–122 (2018)

138

X.-N. Bui et al.

139. Moayedi, H., Tien Bui, D., Gör, M., Pradhan, B., Jaafari, A.: The feasibility of three prediction techniques of the artificial neural network, adaptive neuro-fuzzy inference system, and hybrid particle swarm optimization for assessing the safety factor of cohesive slopes. ISPRS Int. J. Geo-Inf. 8, 391 (2019) 140. Fu, X., Sheng, Q., Zhang, Y., Chen, J., Zhang, S., Zhang, Z.: Computation of the safety factor for slope stability using discontinuous deformation analysis and the vector sum method. Comput. Geotech. 92, 68–76 (2017) 141. Zhou, Y., Cheuk, C., Tham, L.: Numerical modelling of soil nails in loose fill slope under surcharge loading. Comput. Geotech. 36, 837–850 (2009) 142. Zhou, J., Li, E., Yang, S., Wang, M., Shi, X., Yao, S., Mitri, H.S.: Slope stability prediction for circular mode failure using gradient boosting machine approach based on an updated database of case histories. Saf. Sci. 118, 505–518 (2019) 143. Bell, F.G., Donnelly, L.J.: Mining and its Impact on the Environment. CRC Press (2006) 144. Nurmi, P.: Green mining–a holistic concept for sustainable and acceptable mineral production. Ann. Geophys. 60 (2017) 145. Zhengfu, B., Inyang, H.I., Daniels, J.L., Frank, O., Struthers, S.: Environmental issues from coal mining and their solutions. Mining Sci. Technol. (China) 20, 215–223 (2010) 146. Qi, R., Liu, T., Jia, Q., Sun, L., Liu, J.: Simulating the sustainable effect of green mining construction policies on coal mining industry of China. J. Clean. Prod. 226, 392–406 (2019) 147. Oltean, I., Goldan, T., Nistor, C.: Prevention and monitoring environmental impact of open pit coal mining activities. Res. J. Agric. Sci. 50, 259–264 (2018) 148. Lilic, N., Cvjetic, A., Knezevic, D., Milisavljevic, V., Pantelic, U.: Dust and noise environmental impact assessment and control in Serbian mining practice. Minerals 8, 34 (2018) 149. Amirshenava, S., Osanloo, M.: A hybrid semi-quantitative approach for impact assessment of mining activities on sustainable development indexes. J. Clean. Prod. 218, 823–834 (2019) 150. Karacan, C.Ö.: Modeling and prediction of ventilation methane emissions of US longwall mines using supervised artificial neural networks. Int. J. Coal Geol. 73, 371–387 (2008) 151. Karacan, C.Ö.: Forecasting gob gas venthole production performances using intelligent computing methods for optimum methane control in longwall coal mines. Int. J. Coal Geol. 79, 131–144 (2009) 152. Smith, G., Du Plessis, J.: Control strategies for coal dust and methane explosions in underground coal mines: current South African research and development initiatives. J. South Afr. Inst. Min. Metall. 99, 117–121 (1999) 153. Li, W., Wu, X., Jiao, W., Qi, G., Liu, Y.: Modelling of dust removal in rotating packed bed using artificial neural networks (ANN). Appl. Therm. Eng. 112, 208–213 (2017) 154. Chen, D., Xu, P., Li, R.: Comprehensive depress & removal of dust caused by mining activities in underground coal mine. Mining Eng. 3 (2010) 155. Li, Y., Zhang, Y., Xu, X., He, L., Serikawa, S., Kim, H.: Dust removal from high turbid underwater images using convolutional neural networks. Opt. Laser Technol. 110, 2–6 (2019) 156. Karacan, C.Ö.: Development and application of reservoir models and artificial neural networks for optimizing ventilation air requirements in development mining of coal seams. Int. J. Coal Geol. 72, 221–239 (2007) 157. Hua, F., Liangshan, S.: An optimized control of ventilation in coal mines based on artificial neural network. J. Sci. Eng. (China) 8, 80–83 (2002) 158. Zhang, X., Dou, S.: The assessment of ventilation system for underground mines based on neural network. Non-ferrous Mining Metallurgy 4 (2005) 159. Alvarado, M., Gonzalez, F., Fletcher, A., Doshi, A.: Towards the development of a low cost airborne sensing system to monitor dust particles after blasting at open-pit mine sites. Sensors 15, 19667–19687 (2015)

A Review of Artificial Intelligence Applications

139

160. Alvarado, M., Gonzalez, F., Erskine, P., Cliff, D., Heuff, D.: A methodology to monitor airborne PM10 dust particles using a small unmanned aerial vehicle. Sensors 17, 343 (2017) 161. Lee, S., Choi, Y.: Reviews of unmanned aerial vehicle (drone) technology trends and its applications in the mining industry. Geosystem Eng. 19, 197–204 (2016) 162. Le, C.V., Cao, C.X., Le, V.H., Dinh, T.: Volume computation of quarries in Vietnam based on Unmanned Aerial Vehicle (UAV) data (in Vietnames). J. Mining Earth Sci. 61, 21–30 (2020) 163. Suh, J., Choi, Y.: Mapping hazardous mining-induced sinkhole subsidence using unmanned aerial vehicle (drone) photogrammetry. Environ. Earth Sci. 76(4), 1–12 (2017). https://doi. org/10.1007/s12665-017-6458-3 164. Beretta, F., Rodrigues, A., Peroni, R., Costa, J.: Automated lithological classification using UAV and machine learning on an open cast mine. Appl. Earth Sci. 128, 79–88 (2019) 165. Patra, A.K., Gautam, S., Majumdar, S., Kumar, P.: Prediction of particulate matter concentration profile in an opencast copper mine in India using an artificial neural network model. Air Qual. Atmos. Health 9(6), 697–711 (2015). https://doi.org/10.1007/s11869-015-0369-9 166. Bui, X.-N., Lee, C.W., Nguyen, H., Bui, H.-B., Long, N.Q., Le, Q.-T., Nguyen, V.-D., Nguyen, N.-B., Moayedi, H.: Estimating PM10 concentration from drilling operations in open-pit mines using an assembly of SVR and PSO. Appl. Sci. 9, 2806 (2019) 167. Asif, Z., Chen, Z., Zhu, Z.H.: An integrated life cycle inventory and artificial neural network model for mining air pollution management. Int. J. Environ. Sci. Technol. 16(4), 1847–1856 (2018). https://doi.org/10.1007/s13762-018-1813-9 168. Nagesha, K.V., Kumar, H., Muralidhar Singh, M.: Development of statistical models to predict emission rate and concentration of particulate matters (PM) for drilling operation in opencast mines. Air Qual. Atmos. Health 12(9), 1073–1079 (2019). https://doi.org/10.1007/ s11869-019-00723-7 169. Daughton, C.G.: Monitoring wastewater for assessing community health: Sewage ChemicalInformation Mining (SCIM). Sci. Total Environ. 619, 748–764 (2018) 170. Cook, N., Sarver, E., Krometis, L.-A.: Putting corporate social responsibility to work in mining communities: Exploring community needs for central Appalachian wastewater treatment. Resources 4, 185–202 (2015) 171. Wei, L., Liang, L.: Treatment of mining wastewater by nanofiltration. J. Beijing Univ. Chem. Technol 1, 006 (2011) 172. Iakovleva, E., Sillanpää, M.: The use of low-cost adsorbents for wastewater purification in mining industries. Environ. Sci. Pollut. Res. 20(11), 7878–7899 (2013). https://doi.org/10. 1007/s11356-013-1546-8 173. Benavente, M., Moreno, L., Martinez, J.: Sorption of heavy metals from gold mining wastewater using chitosan. J. Taiwan Inst. Chem. Eng. 42, 976–988 (2011) 174. Burakov, A.E., Galunin, E.V., Burakova, I.V., Kucherova, A.E., Agarwal, S., Tkachev, A.G., Gupta, V.K.: Adsorption of heavy metals on conventional and nanostructured materials for wastewater treatment purposes: a review. Ecotoxicol. Environ. Saf. 148, 702–712 (2018) 175. Feng, D., Van Deventer, J., Aldrich, C.: Removal of pollutants from acid mine wastewater using metallurgical by-product slags. Sep. Purif. Technol. 40, 61–67 (2004) 176. Yetilmezsoy, K., Demirel, S.: Artificial neural network (ANN) approach for modeling of Pb (II) adsorption from aqueous solution by Antep pistachio (Pistacia Vera L.) shells. J. Hazardous Mater. 153, 1288–1300 (2008) 177. Allahkarami, E., Igder, A., Fazlavi, A., Rezai, B.: Prediction of Co (II) and Ni (II) ions removal from wastewater using artificial neural network and multiple regression models. Physicochem. Probl. Mineral Process. 53 (2017) 178. Madhloom, H.M.: Modeling of Copper removal from simulated wastewater by adsorption on to fungal biomass using artificial neural network. Global J. Adv. Pure Appl. Sci. 5 (2015)

140

X.-N. Bui et al.

179. Lin, W., Jing, L., Zhu, Z., Cai, Q., Zhang, B.: Removal of heavy metals from mining wastewater by Micellar-Enhanced Ultrafiltration (MEUF): experimental investigation and Monte Carlo-based artificial neural network modeling. Water Air Soil Pollut. 228, 206 (2017) 180. Aryafar, A., Gholami, R., Rooki, R., Ardejani, F.D.: Heavy metal pollution assessment using support vector machine in the Shur River, Sarcheshmeh copper mine. Iran. Environ. Earth Sci. 67, 1191–1199 (2012) 181. Li, Z., Ma, Z., van der Kuijp, T.J., Yuan, Z., Huang, L.: A review of soil heavy metal pollution from mines in China: pollution and health risk assessment. Sci. Total Environ. 468, 843–853 (2014) 182. Facchinelli, A., Sacchi, E., Mallen, L.: Multivariate statistical and GIS-based approach to identify heavy metal sources in soils. Environ. Pollut. 114, 313–324 (2001) 183. Solgi, E., Esmaili-Sari, A., Riyahi-Bakhtiari, A., Hadipour, M.: Soil contamination of metals in the three industrial estates, Arak, Iran. Bull. Environ. Contamination Toxicol. 88, 634–638 (2012) 184. Sulaiman, M., Salawu, K., Barambu, A.: Assessment of concentrations and ecological risk of heavy metals at resident and remediated soils of uncontrolled mining site at Dareta Village, Zamfara, Nigeria. J. Appl. Sci. Environ. Manag. 23, 187–193 (2019) 185. Liu, H., Probst, A., Liao, B.: Metal contamination of soils and crops affected by the Chenzhou lead/zinc mine spill (Hunan, China). Sci. Total Environ. 339, 153–166 (2005) 186. Fryer, M., Collins, C.D., Ferrier, H., Colvile, R.N., Nieuwenhuijsen, M.J.: Human exposure modelling for chemical risk assessment: a review of current approaches and research and policy implications. Environ. Sci. Policy 9, 261–274 (2006) 187. Acosta, J., Faz, A., Martínez-Martínez, S., Zornoza, R., Carmona, D., Kabas, S.: Multivariate statistical and GIS-based approach to evaluate heavy metals behavior in mine sites for future reclamation. J. Geochem. Explor. 109, 8–17 (2011) 188. Fernández-Caliani, J., Barba-Brioso, C., González, I., Galán, E.: Heavy metal pollution in soils around the abandoned mine sites of the Iberian Pyrite Belt (Southwest Spain). Water Air Soil Pollut. 200, 211–226 (2009) 189. Alloway, B.J.: Heavy Metals in Soils: Trace Metals and Metalloids in Soils and Their Bioavailability. Springer, Heidelberg (2012) 190. Silveira, M.L.A., Alleoni, L.R.F., Guilherme, L.R.G.: Biosolids and heavy metals in soils. Scientia Agricola 60, 793–806 (2003) 191. Clemente, R., Walker, D.J., Bernal, M.P.: Uptake of heavy metals and As by Brassica juncea grown in a contaminated soil in Aznalcóllar (Spain): the effect of soil amendments. Environ. Pollut. 138, 46–58 (2005) 192. Jung, M.C.: Heavy metal concentrations in soils and factors affecting metal uptake by plants in the vicinity of a Korean Cu-W mine. Sensors 8, 2413–2423 (2008) 193. Amusan, A., Ige, D., Olawale, R.: Characteristics of soils and crops’ uptake of metals in municipal waste dump sites in Nigeria. J. Hum. Ecol. 17, 167–171 (2005) 194. Raj, K., Kardam, A., Arora, J., Srivastava, S.: Artificial Neural Network (ANN) design for Hg–Se interactions and their effect on reduction of Hg uptake by radish plant. J. Radioanal. Nucl. Chem. 283, 797–801 (2010) 195. Ghosh, A., Sinha, K.: Optimization of Reduction of Copper Using Stenotrophomonas maltophilia PD2 Biomass and Artificial Neural Network Modeling. Environ. Eng. Manag. J. (EEMJ) 14 (2015) 196. Roohi, R., Jafari, M., Jahantab, E., Aman, M.S., Moameri, M., Zare, S.: Application of artificial neural network model for the identification the effect of municipal waste compost and biochar on phytoremediation of contaminated soils. J. Geochem. Explor. 208, 106399 (2020)

A Review of Artificial Intelligence Applications

141

197. Zhou, P., Zhao, Y., Zhao, Z., Chai, T.: Source mapping and determining of soil contamination by heavy metals using statistical analysis, artificial neural network, and adaptive genetic algorithm. J. Environ. Chem. Eng. 3, 2569–2579 (2015) 198. Bazoobandi, A., Emamgholizadeh, S., Ghorbani, H.: Estimating the amount of cadmium and lead in the polluted soil using artificial intelligence models. Eur. J. Environ. Civil Eng. 1–19 (2019) 199. Wang, Z., Yin, Z., Caers, J., Zuo, R.: A Monte Carlo-based framework for risk-return analysis in mineral prospectivity mapping. Geosci. Front. (2020) 200. Panahi, M., Sadhasivam, N., Pourghasemi, H.R., Rezaie, F., Lee, S.: Spatial prediction of groundwater potential mapping based on convolutional neural network (CNN) and support vector regression (SVR). J. Hydrol. 125033 (2020) 201. Porwal, A., Carranza, E., Hale, M.: Knowledge-driven and data-driven fuzzy models for predictive mineral potential mapping. Nat. Resour. Res. 12, 1–25 (2003) 202. Porwal, A., Carranza, E., Hale, M.: Artificial neural networks for mineral-potential mapping: a case study from Aravalli Province Western India. Nat. Resources Res. 12, 155–171 (2003) 203. Saljoughi, B.S., Hezarkhani, A.: A comparative analysis of artificial neural network (ANN), wavelet neural network (WNN), and support vector machine (SVM) data-driven models to mineral potential mapping for copper mineralizations in the Shahr-e-Babak region, Kerman, Iran. Appl. Geomatics 10, 229–256 (2018) 204. Chen, Y., Wu, W.: Mapping mineral prospectivity using an extreme learning machine regression. Ore Geol. Rev. 80, 200–213 (2017) 205. Naghibi, S.A., Pourghasemi, H.R., Dixon, B.: GIS-based groundwater potential mapping using boosted regression tree, classification and regression tree, and random forest machine learning models in Iran. Environ. Monit. Assess. 188(1), 1–27 (2015). https://doi.org/10. 1007/s10661-015-5049-6 206. Naghibi, S.A., Ahmadi, K., Daneshi, A.: Application of support vector machine, random forest, and genetic algorithm optimized random forest models in groundwater potential mapping. Water Resour. Manag. 31, 2761–2775 (2017) 207. Nhu, V.-H., Rahmati, O., Falah, F., Shojaei, S., Al-Ansari, N., Shahabi, H., Shirzadi, A., Górski, K., Nguyen, H., Ahmad, B.B.: Mapping of groundwater spring potential in karst aquifer system using novel ensemble bivariate and multivariate models. Water 12, 985 (2020) 208. Singh, V., Singh, D., Singh, T.: Prediction of strength properties of some schistose rocks from petrographic properties using artificial neural networks. Int. J. Rock Mech. Min. Sci. 38, 269–284 (2001) 209. Kaunda, R.: New artificial neural networks for true triaxial stress state analysis and demonstration of intermediate principal stress effects on intact rock strength. J. Rock Mech. Geotech. Eng. 6, 338–347 (2014) 210. Fathipour-Azar, H., Wang, J., Jalali, S.-M.E., Torabi, S.R.: Numerical modeling of geomaterial fracture using a cohesive crack model in grain-based DEM. Comput. Particle Mech. 7(4), 645–654 (2019). https://doi.org/10.1007/s40571-019-00295-4 211. Nguyen, H., Bui, X.-N.: Soft computing models for predicting blast-induced air overpressure: a novel artificial intelligence approach. Appl. Soft Comput. 92, 106292 (2020) 212. Duan, J., Asteris, P.G., Nguyen, H., Bui, X.-N., Moayedi, H.: A novel artificial intelligence technique to predict compressive strength of recycled aggregate concrete using ICA-XGBoost model. Eng. Comput. 4, 1–18 (2020) 213. Nguyen, H., Mehrabi, M., Kalantar, B., Moayedi, H., Abdullahi, M.A.M.: Potential of hybrid evolutionary approaches for assessment of geo-hazard landslide susceptibility mapping. Geomatics Nat. Hazards Risk 10, 1667–1693 (2019) 214. Nguyen, H., Moayedi, H., Foong, L.K., Al Najjar, H.A.H., Jusoh, W.A.W., Rashid, A.S.A., Jamali, J.: Optimizing ANN models with PSO for predicting short building seismic response. Eng. Comput. 36(3), 823–837 (2019). https://doi.org/10.1007/s00366-019-00733-0

142

X.-N. Bui et al.

215. Borella, J., Quigley, M., Krauss, Z., Lincoln, K., Attanayake, J., Stamp, L., Lanman, H., Levine, S., Hampton, S., Gravley, D.: Geologic and geomorphic controls on rockfall hazard: how well do past rockfalls predict future distributions? Nat. Hazards Earth Syst. Sci. 19, 2249–2280 (2019) 216. Lamghari, A., Dimitrakopoulos, R.: Hyper-heuristic approaches for strategic mine planning under uncertainty. Comput. Oper. Res. 115, 104590 (2020) 217. Ahmadi, M.R., Shahabi, R.S.: Cutoff grade optimization in open pit mines using genetic algorithm. Resourc. Policy 55, 184–191 (2018) 218. He, Y., Gao, S., Liao, N., Liu, H.: A nonlinear goal-programming-based DE and ANN approach to grade optimization in iron mining. Neural Comput. Appl. 27(7), 2065–2081 (2015). https://doi.org/10.1007/s00521-015-2006-9

River Sand Mining Vis a Vis Manufactured Sand for Sustainability Ramesh Murlidhar Bhatawdekar1 , Trilok Nath Singh2 , Edy Tonnizam Mohamad1 , Danial Jahed Armaghani3(B) , and Dayang Zulaika Binti Abang Hasbollah1 1 Geotropik-Centre of Tropical Geoengineering, Department of Civil Engineering,

Universiti Teknologi Malaysia, Skudai, 81310 Johor Bahru, Malaysia [email protected], {edy,dzulaika}@utm.my 2 Earth Science Department, Indian Institute of Technology, Bombay, Powai, Mumbai 400076, India [email protected] 3 Department of Civil Engineering, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia [email protected]

Abstract. Sand and gravels are basic construction raw material globally. With increased economic activity, rate of extraction of sand and gravels has gone up three folds during last two decades reaching 40 billion tonnes per year. Critical hot spots for sand extraction are China and India. Uncontrolled illegal sand mining is common – reduction in ground water level and water transparency, increased water turbidity, adverse impact on birdlife. It is estimated that every year 10 million cubic meters of sand is extracted in Morocco by sand mafias converting large sand beaches into rocky landscape. Singapore has extended its area by 20% by importing 517 million tonnes of sand over last 20 years from neighbouring countries. Some of the countries like India and Malaysia have developed river sand mining guidelines to ensure equilibrium of river, avoid aggradation of hydraulic structures, protection of rivers against erosion, and avoid water pollution. Principal of sustainability is based on equilibrium of meeting business objectives, environmental compliance and social compliance of various stake holders. Tipple bottom line pillars for sustainability include financial stability, environmental stewardship and social equity. For minimizing environmental effect due river sand mining, many countries have adopted technology of manufacturing sand through series of crushers – primary jaw crusher, secondary cone crusher and tertiary vertical shaft impact (VSI) crusher. Japan has developed V7 dry sand manufacturing technology. Manufacturing Sand (M-Sand) has comparable gradation with river sand. Various technical specifications of aggregates and M-Sand adopted by India and US are provided in this paper. There is need of creating R&D facilities in every region for sustainability to develop specifications of M-Sand based on available resources and meet local sand requirement. Keywords: Economic activity · Environmental stewardship · M-Sand · River sand mining · Social equity · Sustainability · V7 dry sand

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 143–169, 2021. https://doi.org/10.1007/978-3-030-60839-2_8

144

R. M. Bhatawdekar et al.

1 Introduction Water is considered as lifeline of human being, likewise sand is inseparable ingredient of development of infrastructure for any nation. It is used in various modes in different purposes; like mixing with cement produced concrete, admixing with bitumen resulted into asphalt under control temperature, developed various type of glass as per its purity and many more [1, 2]. In the present era, one cannot think about development without sand and its compact form aggregates. Sandstones are widely geographically distributed in different geological horizons having variable age and characteristics also [3–5]. They are the main source of sand generation. The river flowing through these rock strata produce sand due to disintegration and destruction of solid rock mass into fragments. The transpiration makes size reduction and then deposition. Sand is available either in riverbed, deserts or seabed and some time in creeks [6–8]. Due to weathering and erosion sand grains developed rounded shapes have different binding nature with concrete, whereas sea sand is also rounded in geometry due to long transport history. They also contain seashell and chlorides which deteriorates the quality and durability of concrete [9]. Uses of this type of concrete many do not have longer life. Sand and gravel are formed by the long geological process of weathering and erosion by various geological agents like water, wind, glaciations, storms, flooding etc. [10, 11]. The river sand mining is some places banned due to environmental rules and regulation. This put construction industries some time under stress due to unavailability of sand or even available but cost is very high. This is estimate that 40 billion tons of sand and gravel are consumed by different countries as indicated by united nation environmental program committee [12]. This clearly demonstrates that the demand is always high as compared to the availability. India and china, both highly populated countries are going through the large-scale development activities in term of housing, road, airport, dams and other civic projects. The demand of sand has been growing exponentially high and compel to think other possible resources to meet the growing demand of the industries. One of the estimates says that the 18 kg of sand is requirement for per person per day on the planet on the basis of total consumption over the total population [13–17]. The production of sand needs alternative ways to fulfill the demand of the national meet. India and China required more sand than the rest of the world. One of the ballpark figures indicates that the China has used concrete particularly in last 3 years which is equivalent to the concrete used by USA in the last ten years. Indonesia is another country which required much more sand as compare to previous years. Sand is not only used in construction but is also used as reclamation for land and port development [18]. Many development are taking a place where people have used island for making hotels, tourist spot for example - Dubai, Palm, Jumeirah, Singapore have increase its’ land area by 20-25% since 1960 by reclaiming the watery area into land area. They have brought sand from neighbour country like Malaysia and Indonesia, similarly Hong Kong have development its international airport at Chek Lap Kok on the

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

145

reclaimed area using huge quantity of sand [19, 20]. In India, many rocky beeches have been converted into sandy beach by pouring sand from outer area. The development will be adversely affected in the want of sand in larger quantity and due to an environmental problem [21–23]. The scientist should think to suggest an alternative solution to resolve this outstanding problem. The miniature sand is one of the alternatives which can sustain and reduce the demand supply gap and bring a new dimension to this problem and curtail the challenges. Sand mining is very often to be systematic because in sediments bed it is digging and filling process. This type of digging may damage the sidewall. Haphazard and non-scientific excavation techniques pose serious problem to surrounding ecosystem and environmental. Sometime river bed sand mining or coastal area sand excavation with rapid rate, can disturbed and destroy the existing biodiversity as well as ecological system [24–29]. Secondly, it might aggravate the erosion and undercutting of riverbank and collapse of coastal slopes [30, 31]. It has been also observed that the bridge pillars are also affected by rapid erosion and transportation of base material [27, 32]. Sometime river flow system also changes and developed a new catchment area. There is possibility of flooding also. Cumulatively, this can adversely affect the river carrying capacity leading to destroy the hydrological dynamics system. In India, a huge amount of sand is excavated by unauthorized people without mining lease [33–35]. It reveals that an estimated amount of 150 billion rupees made by unethical ways and huge loss of revenue to the state. Domestic demand has been increased many folds from 1.4 billion tonnes to 4.8 billion tonnes. Uttar Pradesh, in India is highest consumer of sand as compared to other state, approximately 101 million tons of sand is consumed by the Uttar Pradesh followed by Maharashtra, Tamil Nadu etc.

2 Global Scenario of Sand In many developing countries, sand extraction operations do not always follow the environmental regulations and norms. This has a huge negative impact on the environment and society at large, as reported in India, China and other countries in Africa, South America and Asia [36]. Figure 1 shows the extraction of river sand at Dibagmba River, Cameroon by local people. India and China head the list of major environmental impact in coastlines, rivers and lakes [27]. One reason could be that these countries also head the list in respect of infrastructure and construction. Many other countries, like Vietnam, Cambodia and Indonesia are also similarly affected by the construction boom in their countries. Further, for land reclamation and development of regional economic corridor in Southeast Asia, these countries - Vietnam, Cambodia and Indonesia - have become suppliers of aggregate materials - through legal or illegal sources [37]. Figure 2 shows the global consumption of cement, sand and gravel for whole world, China and rest of the world. Columbia in South America identified as a place presenting challenges of sustainability [38]. Similarly, Sierra Leone, Kenya, Tanzania in Africa determined as places presenting sustainability issues [12]. Figure 3 shows the degradation of riverbed, Tanzania. The above cases show how economic factors like livelihood and biodiversity can be badly affected by uncontrolled extraction of sand. Disasters like subsidence and landslides and drowning of workers have resulted in extraction areas [12].

146

R. M. Bhatawdekar et al.

Fig. 1. River Dibagmba, Cameroon - Extraction of river sand by local people [12]

Fig. 2. Global consumption of cement, sand and gravel for whole world, China and rest of the world [12]

Both legal and illegal operations are happening at an increased rate, right next to biodiversity parks and protected areas [37]. These are unique places where citizens at large have agreed to preserve the ecology and culture; they should not tolerate economic activities that are so damaging to the biophysical aspects of this ecosystem. Sadly, marine

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

147

Fig. 3. Degradation of riverbed, Tanzania [12]

life, fisheries and birdlife have been badly affected and so also the survival of threatened species like turtles and freshwater dolphins [37]. Extraction of sand and aggregates in rivers has resulted in significant pollution and fluctuation in pH levels [28]. Riverbanks have become unstable and this has led to frequent and severe floods [27]. The groundwater levels in aquifers have become lower [39]. Occurrence of drought and its severity has increased [40]. Dam construction and sand extraction have reduced the deposits of sediments by the rivers along the coastal areas and deltas; at the same time erosion of beaches has increased [41]. These effects are further aggravated when sand is directly extracted from coastal sand dunes and marine dredging and may have long-term impact on the ecosystem (US: [42], UK: [43]). In places like coastal New Zealand despite warnings of uncertainty of environment sustenance due to mining, urbanisation of the coast and climate change [44], extraction of sand nearshore and offshore continues. A Note on Global Sand Extraction Data. Other than water, aggregates like sand, gravels and crushed rock are extracted in larger quantities than any other mineral in nature. But comprehensive and reliable data and information of their extraction is not available in most countries. Even though sand, gravels and crushed rock is used in a number of applications where cement is not used, the most reliable estimates about global production of sand, gravels and crushed rock can be obtained from cement production data.

148

R. M. Bhatawdekar et al.

China is the largest producer of cement in the world at 2.4 billion tonnes in 2017, followed by India (270 million tonnes) and USA (86.3 million tonnes). Using a ratio of 10 tonnes of aggregate per tonne of cement as in USA, a global production of 4.83 billion tonnes forecast for 2030 would require about 50 billion tonners of aggregate per year in 2030. Sand Mafias, Illegal Sand Extraction and Smuggling in Morocco. In Morocco, almost 50% of the 10 million cubic metres of sand per year comes from illegally exploited coastal sands. A large stretch of sand which formed a beach between Safi and Essouira has completely been transformed into a rocky mass, thanks to sand smugglers. In Asilah, Northern Morocco, due to regulatory issues and pressured of tourism, severe erosion of beaches has taken place and many structures that were responsible for creating such a massive erosion are themselves in danger due to erosion [12]. A Growing International Trade in Sand? Singapore, a city-state hailed as a miracle of development, continues its growth in size. The area of the city-state is small and the population is growing fast. In order to create more space for its growing population, Singapore started reclaiming land from the sea, by using aggregates. Over last 40 years, Singapore has added 130 square kilometres amounting to an increase of 20% of land (See Fig. 4). For this and similar activities, Singapore has imported a reported 517 million tonnes of sand over the last 20 years, making Singapore the largest importer of sand in the world.

Fig. 4. Satellite imagery position of Singapore from 1973 to 2018 showing in-creased territory of Singapore [12]

Originally, sand was imported mainly from Indonesia and a few other countries in the neighbourhood such as Malaysia, Cambodia and Thailand. However, in Indonesia, 24 sand islands where sand was being mined for Singapore disappeared, and this caused major political tensions between the two countries as they struggled to lay maritime

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

149

boundaries between the two countries [12]. In 2002, a temporary ban was imposed on export of sand from Indonesia to Singapore, and since then other neighbouring countries are meeting the demand from Singapore. A study of the international traded volumes shows that there is a mismatch between the total quantity of sand being imported into Singapore (517 million tonnes) and the sum of the export volumes from four neighbouring countries (637 million tonnes). This shortfall of 120 million tonnes are obviously due to illegal import of sand to Singapore, showing the existence of an illegal sand trade and throws up a need for better control and monitoring. Another interesting feature of international sand trade is that whenever the price of sand increases, there is an increase in the traffic of sand by local mafia. The kind of price increases can be judged by the fact that during the period 1995 to 2001, the price of sand imported by Singapore was US$ 3 per tonne; during the period 2003 to 2005, the price had increased to US$ 190 per tonne. Based on literature review and discussion with experts on sand mining five priority areas were identified which are stated below: Lack of Awareness. The public, governments and traders and users of sand do not know why extraction of sand and the state of the sand resources around the world are connected to their lives. Till recently, sand was freely available and cheap, so widely used in construction projects and hence the public had no idea that sand was a limited resource. The general perception of sand as available in deserts in limitless quantity was faulty; the public was not aware that desert sands are highly polished by desert winds and cannot be used in construction or in land reclamation. Endless availability of sand was generally assumed. Economics. Information about sand availability and impact of extraction is scarce. But planning and management of this valuable resource requires basic data and answers to the following questions: 1. Considering all external and internal factors and the impact on ecology, what would the real cost of sand extracted from our coasts that has to be borne by our economies and societies? 2. Since sand is not uniformly available at all consumption points in the world, how would the availability meet the projected demand? What would be the impact on trading and pricing? 3. National governments and business owners must address the crucial question as to who owns the rights to the resources. Who are the others who are directly and indirectly involved? How will they be affected? What can they do? Responsibility. There is no clear responsibility among governments, construction companies, builders, users, contractors who are involved in extraction, transportation and usage of sand. In the entire supply chain, there is no transparency, no information system from start to finish and hence impossible to track data and value addition. Governance. Communities, companies and local governments need to work in a coordinated fashion, so that anybody deviating from the laid down norms is identified and

150

R. M. Bhatawdekar et al.

asked to correct themselves. This would be constructive governance by all concerned working together.

3 Environmental Impacts of Sand Mining Deposits of sand, although occurring in different locations, are environmentally linked with one another; any depletion of sand in one location has an environmental impact on sand deposits at all other locations. Sand mining gives rise to a host of problems and given below is an overview of the problems being faced in sand deposits of Malaysia [26]. Turbidity. Turbidity, caused by presence of particulate matter in water, is due to discharge of wastewater or storm water into streams or rivers and also due to dredging operations for extracting sand or gravel from riverbeds. Turbidity is highest at dredging sites or points where wastewater is being discharged into the streams or rivers. Habitat, Flora and Fauna on Riverbanks. All species require appropriate habitat conditions for survival over long term. By nature, such habitats are aesthetically appealing. But mining activities in river streams not only impact the habitats in the immediate vicinity but extend far beyond. Sediment flow becomes erratic on riverbeds and banks and drastically affect the channel configuration. Fertile land is lost, timber resources are depleted and wildlife in riparian sites are severely affected. Downstream, fisheries lose their productivity; biodiversity is affected and even human recreational facility loses its potential. Ultimately land value and aesthetic value is depleted. Structural Stability. Mining operations for sand and gravel in riverbeds cause channel incisions and weaken piers of bridge piers and other infrastructure. Buried pipelines get exposed, exposing the neighbourhood to danger. There is a huge risk of damage to public and private property due to sand mining. There are actual examples of collapse of bridges due to weakening of piers due to sand mining on riverbeds. Water Quality. Deterioration in water quality is one of the most significant negative impacts of sand mining in or dredging in river streams. Turbidity is the biggest problem caused by a combination of effects like: sediment resuspension, sedimentation caused by dumping of excess material mined, organic matter and oil leaks or spillages from excavation or transportation vehicles. Turbidity caused by suspended solids affects usage of water by people living downstream, particularly if it is used for domestic purposes and that of local biota [45]. 3.1 A Case Study on Sand Policy in Telangana State, India In India sand is considered ‘minor’ mineral and controlled by state government. Telangana state government has taken various steps for monitoring river sand supply, curbing illegal sand mining and promotion of manufactured sand. Water, Land and Tress (WALT) Rules (2004) of Andhra Pradesh. Sand monitoring committees are formed at each district level and central committee at state level. Various information is collected through various departments:

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

1. 2. 3. 4.

151

Ground Water Department. Irrigation and Conservator of River Department. Revenue Department. Mines and Geology Department.

Sand is extracted and supplied through Telangana State Mineral Development Corporation (TSMDC) to avoid excessive exploitation and systematic extraction. The major source of sand is through desiltation of irrigation projects. TSMDC follows environment management plan, mining plan, etc. Mechanism Evolved for Curbing Illegal Mining of Sand. The mechanism is as following: 1. Established check-posts at vulnerable locations and as well as at the interstate borders as per Telangana State Sand Mining Rules (2015). 2. To properly regulate the sand activity by arresting the unauthorized transportation. 3. District wise mobile squads comprising Mines and Geology, Revenue, Police and Transport Department have been constituted and activity being directly monitored by the District Collectors. 4. Stringent mechanism evolved to curb illegal extraction and transportation of sand by other than TSMDC Ltd. by imposing heavy penalty. Promotion of Manufactured Sand. It is as following: 1. Manufactured sand from the rocks or crushed stone sand shall be promoted as an alternative to natural/river sand by giving manufactured sand units industry status and extension of subsidy/concessions with preference in allotment of quarry lease. 2. Presently, 45% of sand requirement for Hyderabad city is being met from the manufactured sand units. 3. Manufactured Sand and Natural Sand in the ratio of 50:50 in all the mixes of cement concrete and in the items of works where sand is used earlier to reduce the burden on natural resources and to make the sustainable ecosystem. The Sand Mining Policy of Telangana State as Model Sand Mining Policy. The policy is as following: 1. The officials at the helm of sand regulation in Chhattisgarh, Maharashtra, Jharkhand, Madhya Pradesh, Haryana and Karnataka States have visited Telangana to study the Sand Policy and its implementation in their states. 2. Although sand being a minor mineral comes under State domain but due to unabated/uncontrolled sand extraction leading to multifarious problems, the Ministry of Mines, Government of India constituted a committee to study the Sand Mining Policies under implementation by various states and prepare a comprehensive policy duly adopting the best practices. 3. Telangana state being a member has projected its policy and implementation before the Committee constituted by Ministry of Mines, Government of India.

152

R. M. Bhatawdekar et al.

4. Madhya Pradesh state is already in the process of adopting Telangana State Mining Policy and the Rules made there under. 5. The Ministry of Mines, Government of India issued Sand Mining Framework as comprehensive guidelines for regulation of sand mining in the country duly incorporating the salient features and best practices from Telangana State Sand Mining Policy. 3.2 River Sand Mining Management in Malaysia [46] General guidelines for Mining gravel and sand from river in Malaysia. Following steps are taken as per guidelines: 1.

Initially location of sand and gravel deposition or aggradation is identified. Mine operator is permitted to extract material where problem of aggradation is least. 2. Sand and gravel mining are decided based on replenishment rate of the river. Historical sediment rating curves are developed for each season and based on this experience, mining operator is permitted to extract quantity of sand and gravel. 3. Thickness of sand and gravel bed which can be removed is dependent on river width and replenishment rate as a particular location in the river. 4. Extraction of sand gravel where there is a chance of erosion or concave bank taking place is prohibited. 5. Extraction of sand and gravel is not permitted from any critical hydraulic structure such as bridges, buildings, pumping stations, canals etc. 6. Mining is permitted from the downstream of the sand bar at river bends. To promote channel stability, retaining of 33% to 66% of riparian vegetation is acceptable practice. 7. Areas having hydraulic structures or infrastructures, flood discharge capacity is necessarily maintained. 8. Location of proposed river sand River survey is carried out 1 km downstream and 1 km upstream side. Sediment sample is collected from riverbed during every stage of mining. Historical curve is river sand is studied during peak season during monsoon. 9. Extracted volume is replenished during suitable period based the historical or gauged flow rating curve. 10. The extraction volume and allowable mining depth is determined based on the sediment rating curve and high flow period. 11. The exact volume of extraction is determined by carrying out cross section survey before, during and after the sand mining operation in riverbed. 3.3 Individual Government’s Role as Per United Nations Guidelines for Sustainability [12] Individual government has a major role to play by coordinating with international agencies such as United Nations (UN) to understand global standard and modify these standards to suit local available resources and local demand of aggregates and sand. Control, monitor and restrain for occurrence of illegal sand mining. Promote responsible

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

153

sand mining by local operators to meet environmental standards. Individual government should investigate to understand geological and hydrogeological aspects of sand resources. Government should also invest their resources for carrying out studies for having long term benefit on society. River sand mining can be avoided with already known solutions: 1. To avoid overdesign or excessive size of infrastructures. 2. To promote construction sector for using manufactured sand, recycling of aggregates or alternative material to sand. 3. To implement best practices for reduction in consumption of sand. What is Sustainability? While the English Dictionary has many definitions of ‘Sustainability’, the most popular one being ‘the ability to continue with a given lifestyle indefinitely’, in the context of environment and business, the word has taken on a deeper meaning. In the context of human life on earth, Sustainability has come to mean ‘the ability to meet the needs of the present, without sacrificing the needs of future generations to meet their needs’. In the context of business and industry, Sustainability has come to mean a business strategy that maximizes long-term profitability, while maintaining, restoring and improving environmental quality and building social equity. Business Case for Sustainability. Companies have made a happy discovery (Fig. 5) that when they pursue sustainable growth initiatives, they can innovate better business ideas and enjoy a competitive advantage. The overall benefits of an organization practicing sustainability are listed below: 1. 2. 3. 4. 5. 6. 7.

Creating newer revenue channels. Improved profitability. Foreseeing and controlling risks. Reducing environmental impact. Creating social goodwill to operate smoothly within the community. Improved market penetration, globally. Providing major advantage against competition.

Environmental Stewardship – Going Beyond Compliance. A study showing performance of companies in terms of expenses incurred by them on environment and overall performance has resulted in some very interesting findings. When companies just manage to obey environmental laws, assess the impact of their industrial activities, preparing and submitting reports to authorities and wherever called for, obtaining legal opinion – the expenses incurred could be significant, but the return on investment (ROI) in such expenses can be minimal. The financial gains enjoyed by companies when they incur expenses in carrying out the bare minimum requirements under law are the least. On the other hand, when companies invest substantially in activities that go beyond the bare minimum required for compliance, it has been found that the economic performance of the companies increase substantially, correlating with the additional investment in environmental activities.

154

R. M. Bhatawdekar et al.

Fig. 5. Systemization of value drivers and value constructs

4 Manufactured Sand + V7 Sand Manufactured sand is fine aggregates produced by crushing suitable quarry stones such as granite, andesite, sandstone, limestone. The larger aggregates pieces or coarse aggregates are converted into sand-sized particles. Manufactured sand or M-Sand is used for manufacturing concrete for various construction applications. M-Sand is the most idle and alternative of river and sea sand which has got all kind of acceptance not only in India but also in other countries. M-Sand can be produced while crushing the rocks quarry stone, sometimes it will be present and disposed as dump material also to a desire size of 150 micron, which is desire size for the mixing with aggregate and concrete. The crushing processes are complicated but scientifically proven one. The hand rock material goes with various capacity of crusher, like primary crusher, secondary crusher, tertiary crusher etc. The time and hardening characteristics’ influence the economy and utility of aggregates. It also depends on grading, durability, shape and size, surface structures and texture, skid resistance, unit weight and voids as well as surface moisture. The aggregates crushing values should satisfy in accordance with IS 2386 (Part 4): 1963 with factor like aggregate impact values, abrasion value and soundness. All these tests indicate that the M-Sand is very reliable, economical, cheaper and scientifically superior as compared to river sand.

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

155

China, Malaysia, Indonesia started production of M-Sand very aggressively as compared to India. Karnataka, state has initiated the production of M-Sand. This state produces nearly 20 million tonnes/year with the help of 164 units. Other states need to encourage and start manufacturing M-Sand to meet the need of the state. Scientist, planner, developers, and engineers all should promote M-Sand production at national level. This approach will serve two purposes, first it will be very useful and helpful to protect the ecosystem of river basins as well as nearly stop illegal sand mining, which is very effective in India, particular in Indo-gangatic plan. 4.1 Manufactured Sand Plant Manufactured sand plant (Figs. 6 and 7) consists of primary jaw crusher, secondary and tertiary cone crusher. Vertical Shaft Impact Crusher (VSI) is commonly installed for getting better shape of M-Sand.

Fig. 6. Manufacturing sand plant 1000 TPH showing overall layout of the plant

4.2 V7 Dry Sand Making Machine Sand is one of the important constituents for manufacturing concrete. Sand as building material is important constituent for quality construction. Due to environmental reasons, river sand which was being used many couple of decades is replaced in many metro cities by M-Sand. River sand size, gradation and quality may vary from where it is extracted. Due to non-availability of quality river sand and environment reasons, VSI technology is adopted worldwide to produce manufactured sand (M-Sand) [47–52] during last decade. Still various aspects of M-Sand such as angularity, shape and gradation and their application in ready mixed concrete are under various stages of experimentation. A study on workability of M-Sand in ready mixed concrete replacing fine aggregate has been done by [53]. Investigation on usage M-Sand in self compacting concrete by [54] found that due to higher percentage of finer particles in M-Sand, water absorption increases and improves hardening property of SCC. Experimental studies by [55, 56] exhibited

156

R. M. Bhatawdekar et al.

Fig. 7. Manufacturing sand plant 1000 TPH showing close up view

results that replacing 60% and 50% natural sand with M-Sand in concrete and cement mortar respectively showed higher strength. Thus, usage of M-Sand in concrete depends upon various factors such as sourcing quarry, hardness of aggregates, size and gradation achieved. However, in many cities, metros including Mumbai RMC plants are using 100% M-Sand for last couple of years. M-Sand faces problem of increased production of grains in the range of 0.6–0.15 mm. Various experimentation have been done of combination of natural sand and M-Sand to obtain maximum strength of concrete. In order to understand product performance during similar conditions (Apple to Apple comparison) between natural sand, locally available VSI crushed sand & V7 sand. This was to understand the different parameter and properties of all three fine aggregate which would help to meet requirement for different applications. For each type of fine aggregate, criteria of comparison are strength, initial workability, slump retention, pump ability, water absorption, particle size distribution, specific gravity or density and temperature of concrete etc. 4.2.1 Dry Sand Technology Kemco V7 dry sand technology has been used worldwide and has received favourable results [57–60]. Crushed sand similar to natural sand in terms of shape and grain size can be manufactured by V7 dry sand-making system. It converts excess crushed stone into great quality sand. Also, it makes use of VSI’s special qualities such as cheapness, interested grain shape and reliable gradation. The 5-port impact rotor aids in enhancing performance and reducing cost. The common hitch of higher production of grains with 0.6–0.15 mm size shall be solved Kemco V7 VSI. It returns some of the crushed material back to crusher as it is a closed-circuit air screen system. The V7 plant produces crushed sand

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

157

to be used as an individual component for concrete and other applications, a practical possibility. It provides controlled grain size and shape almost similar to natural sand. Various benefits are obtained from new sand making machine (Fig. 8). Low operating cost due to proven VSI crusher. Fineness modulus can be easily achieved in the range of 2.6–3.0 mm. The use of water-adding mixer in the technique as well as unswerving grading during stockpiling and transportation shall yield good quality sand. It would be similar to the wet processed sand. Also, interested grain size can be achieved better than natural sand. This system is noise and dust free.

Fig. 8. Latest sand making machine - V7 dry sand

For selecting type and design of system, information required are geology, bulk density, SIO, hardness, feed gradation and moisture content. System is capable of receiving material from 20 to 100 tph and producing 17 to 85 tph based on gradation and fineness of input feed material. Concept of V7 dry making system and air screen are illustrated in Fig. 9 and final product example with gradation curve is shown in Fig. 10. 4.2.2 Summary of V7 Dry Sand Making Machine The V7 plant makes a single sand production component for concrete and other applications. Laboratory level tests were carried out by selecting natural sand, M-Sand and V7 dry sand supplied by Kemco. Physical properties of each type of sand was tested in laboratory for particle size distribution, loose bulk density, specific gravity, water absorption, workability of concrete and retention, pump ability and cohesiveness, temperature of concrete. Compressive strength of concrete with natural sand, M-Sand and V7 dry sand in 7 days was found to be 28.41, 26.04 and 33.90 kg/cm2 . Particle size distribution in V7 sand is better than other two sand which helps in better cohesivity and pump ability of concrete.

158

R. M. Bhatawdekar et al.

Fig. 9. a Air screen, b V7 dry sand making system

Fig. 10. Output product V7 dry sand

5 Aggregates and Sand Specification A lot of infrastructure projects are ongoing consisting of highway roadways, metro rails, extension of state highways and airports. With upcoming infrastructure projects, residential colonies are coming up in metro, two tier cities, near railway metro junctions and airports in south East Asia. The aggregates and sand are used for different construction applications. Aggregates and sand demand have gone up many fold times. There is shortage of natural sand across many parts of various countries. Manufactured sand produced from aggregate rock or waste is an alternative to meet shortage of natural sand for construction purpose. Figure 11 shows coarse aggregates and manufactured

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

159

sand. The properties of aggregate required for each application are different. While considering aggregates for engineering and construction, properties such as particle size, plasticity, mineral composition, strength, durability and permeability are important. In addition, certain specific application, angularity of sand particles, electrical conductivity, thermal conductivity and corrosion potential become more significant. Highway, metro rail and airfield projects include drainage structures, retaining wall structures, erosion control structures and deep foundation supported structures. In India for development of infrastructure projects for building new roadways, airports and rail facilities, tremendous amounts of aggregates are required. Various constructions are done for retaining wall and different types of foundation. During present infrastructure development consisting of bridges, flyovers, underground metro tunnels should have minimum life span of 50 to 100 years with least maintenance. Thus, building material – aggregates and sand used for this infrastructure projects should have desired quality to provide prolonged life.

Fig. 11. Coarse aggregates and manufactured sand

Indian Standard [61] provides guidelines based on mechanical properties and gradation of M-Sand. M-Sand which is utilized for concrete is just not only physical reaction but also chemical and hydrothermal reaction. Aggregates and M-Sand contribute to major volume of concrete. Alkali Silica Reaction in concrete which is caused by silica in aggregates, alkalis from cement and moisture from near bye concrete structure may appear in couple of weeks to after 10 to 20 years. During present infrastructure development consisting of bridges, flyovers, underground metro tunnels should have minimum life span of 50 to 100 years with least maintenance. Standard quality building material will ensure prolonged life of any infrastructure. Natural aggregates occurring near infrastructure project varies in physical, mechanical and chemical quality. Indian Standards for Aggregates and M-Sand needs to be reviewed for identifying Alkali Silica Reactivity, Chloride content, Clay content, Organic impurities. Standards also need to consider different construction applications. Indian Standards for aggregates and MSand are compared with American ASTM and AASTO Standards, Chinese Standards from Shenzhen and Japanese standards for crushed stone and manufactured sand. Properties of Aggregates. Aproperty is qualitative indicative of specific characteristics of material. A number of properties exist that identify the behavior of material under

160

R. M. Bhatawdekar et al.

different conditions [62]. A property may or may not be identified by numerical value, preferably it can be. Aggregates properties can be grouped into three categories: 1. Physical properties: Particle angularity and shape, maximum and nominal maximum particle size, particle surface texture, pore structure, absorption, permeability, specific gravity, particle grading, unit weight and voids in aggregates, thermal value change, thermal conductivity, integrity during heating, electrical conductivity, reflection, glare, colour, volume change wetting and drying, resistance to wetting/drying, deleterious substances. 2. Chemical Properties: Solubility, slaking, base exchange, surface change, coatings, resist to attract by chemicals, chemical compound reactivity, oxidation and hydration reactivity, organic material reactivity, chloride content. 3. Mechanical Strength: Particle strength, mass stability, particle stiffness, wear resistance, resistance to degradation, particle shape of abraded fragments, resilient modulus. All of the above properties of aggregate are not required to be identified or tested for selecting suitability of aggregates. However, based on application, each property can play significant role. Common Physical Characteristics of Conventional Aggregates. Granite, limestone, quartzite and sandstone are common types of aggregates. Some of the physical properties of these aggregates’ particles are tabulated in Table 1. Physical properties of aggregate are tested for (i) density (ii) compressive strength (iii) tensile strength (iv) shear strength (v) modulus of rupture (vi) modulus of elasticity (vii) water absorption (viii) average porosity (ix) linear expansion (x) specific gravity. Table 1. Common physical characteristics of conventional aggregates [63–65] Physical characteristics

Granite

Limestone

Quartzite

Sandstone

Density (g/cc)

2.60–2.75

1.85–2.80

2.65–2.70

1.90–2.70

Water absorption (% by weight)

0.07–0.31

0.50–24.1

0.10–2.1

2.0–12.1

Average porosity (%)

0.4–3.81

1.1–31.1

1.5–1.91

1.9–27.2

IS Specifications. The terminology is as following [66]: 1. 2. 3. 4.

Crushed Stone Sand (CSS) - Fine aggregates (sand) produced by crushing hard stone. Crushed Gravel Sand (CGS) - Fine aggregates produced by crushing natural gravel. Mixed Sand - Fine aggregates produced by blending natural sand and CSS or CGS. Manufactured Sand - Fine aggregates manufactured from other than natural sources by processing materials using thermal or other processes such as crushing, separation, washing and scrubbing.

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

161

Any of above varieties of sand can be used in construction, provided that mechanical and chemical properties specified in [61] are satisfied. CSS is most commonly used as alternative to natural sand. Manufactured Sand. In latest version of [61], “Manufactured Sand” terminology is added considering the scarcity of natural sand. Considering the scarcity in the availability of natural resources, the Indian Standards now allows the use of industrial waste materials such as recycled concrete aggregates (RCA). However, Indian Standards have put certain maximum limits while using as aggregates for concrete as shown in Table 2. Table 2. Extent of utilization of manufactured fine aggregates [61] Type of aggregate

Maximum utilization Plain concrete (%)

Reinforced concrete (%)

Lean concrete (%) (less than MIS grade)

Fine aggregate Iron slag aggregate

50

25

100

Steel slag aggregate 25

Nil

100

Copper slag aggregate

40

35

50

Recycled concrete aggregate (RCA)

25

20 (Only up to M2S grade)

100

Permissible limits for mechanical and chemical properties as per [61] are shown in Table 3. Frequency of testing aggregate properties is illustrated in Table 4. Grading of aggregates specified in [61] is presented in Table 5. 5.1 ASTM and AASHO Standards for Aggregates ASTM. The American Society for Testing and Materials (ASTM) is well known globally not for profit organization. The members of ASTM have diversified representations from users, manufacturers, contractors and groups with common interest. The objective of organization is to develop independent, free-willed standards for materials, products, services and systems. ASTM develops standard methods to establish conformity with their own standard specifications. ASTM standards are published in 65 individual volumes covering a wide variety of materials. Many of them are available in separate reprints. AASHO. The American Association of State Highway and Transportation Officials (AASHO) is a comparable organization to ASTM. Representatives from every 50 states of department of highway or transportation, the District of Columbia, Puerto Rico and US Department of Transportation form AASHO. On the basis of ASTM Standards,

162

R. M. Bhatawdekar et al. Table 3. Permissible limits for mechanical and chemical properties [61]

Property

Permissible limit as specified [61]

Aggregate crushing value (For wearing surfaces)

A maximum of 30%

Aggregate crushing value (For non-wearing surfaces)

In case the value exceeds 30%, text for ‘10% fine’ should be conducted with maximum load for 10% being 50 kN

Abrasion value

(i) Not more than 30% for wearing surface (ii) Not more than 45% for non-wearing surface

Impact value

(i) Not more than 30% for wearing surface (ii) Not more than 45% for non-wearing surface

Soundness

(i) For fine aggregates, - 10% when tested with Na2 SO4 - 15% when tested with MgSO4 (ii) For coarse aggregates, - 12% when tested with Na2 SO4 - 18% when tested with MgSO4

Combined flakiness and elongation index

Three suggested test methods: (i) Chemical method (ii) Mortar bar method - Using 38 °C temperature regime - Using 30 °C temperature regime (ii) Accelerated mortar bar method

Table 4. Frequency of testing aggregate properties [61] Property

Test frequency suggested by IS 4926 for normal monitoring

Gradation

Monthly

Moisture content

–

Silt content for fine aggregates

Monthly

Water absorption

3 Monthly

Particle density

3 Monthly

Bulk density

6 Monthly

Particle density

6 Monthly

Chloride content

6 Monthly

AASHTO also develops their both specifications and test methods. A listing of available ASTM and AASHO standards for aggregates is tabulated in Table 6. ASTM and AASHTO standard test methods are listed by title to serve as an aid in helping to locate appropriate test methods is shown in Table 7.

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

163

Table 5. Grading of aggregates specified in [61] IS Sieve designation Percentage passing for grading zone (%) I

II

III

IV

10 mm

100

100

100

100

4.75 mm

90–100 90–100 90–100 90–100

2.36 mm

60–95

75–100 85–100 95–100

1.18 mm

30–70

55–90

75–100 90–100

600 µm

15–34

35–59

60–79

80–100

300 µm

5–20

8–39

12–40

15–50

150 µm

0–10

0–10

0–10

0–15

Table 6. A list of available ASTM and ASSHO standards for aggregates [62] Application for aggregates

ASTM

ASASHTO

Aggregate base and subbase and soil aggregates

ASTM D 2940

AASTO M-283 AASHTO M-147 AASTO M-155

Aggregates for bituminous applications

ASTM D 3515 ASTM D 693 ASTM D 1139

AASHTO M-43 AASHTO M-29 AASHTO M-17 AASHTO R-12

Aggregates for Portland cement

ASTM C 33

AASHTO M-6 AASHTO M-80 AASHTO M-195

Practices - General

ASTM D 8 ASTM C 125 ASTM C 3665

AASHTO M-145 AASHTO M-146 AASHTO R-1 AASHTO R-10 AASHTO R-11

There are also additional tests for following objectives by ASTM and AASHTO: 1. 2. 3. 4. 5. 6.

Testing aggregates in bituminous applications. Aggregates base moisture - Density - Permeability relationships. Strength parameters of aggregates base. Specific absorption and unit weight of aggregates. Frictional properties of aggregates and pavements. Measurements and indices of particles shape and texture.

164

R. M. Bhatawdekar et al.

Table 7. A list of available test procedure under ASTM and ASSHO standards for aggregates [62] Application for test procedures for aggregates

ASTM

ASASHTO

General testing

ASTM E 11 → ASTM D 3666 ASTM C 1077

AASHTO M-92 AASHTO M-231 AASHTO R-18

Sampling and sampling preparation

ASTM D 75 → ASTM C 702 → ASTM D 421 →

AASHTO T-2 AASHTO T-248 AASHTO T-87 AASHTO T-146

Particle size analysis of aggregates

ASTM C 136 → ASTM C 117 → ASTM D 5444 → ASTM D 422 → ASTM D 546 →

AASHTO T-27 AASHTO T-11 AASHTO T-30 AASHTO T-38 AASHTO T-37

Properties of fines in aggregates

ASTM D 2419 → ASTM 4318 → ASTM D 3744 →

AASHTO T-176 AASHTO T-89, T-90 AASHTO T-210 AASHTO T-330

Tests to evaluate general quality of aggregates (unconfined or in concrete)

ASTM C 88 → ASTM D 4792 ASTM C 666 → ASTM C 131 or C 535 → ASTM D 6928 → ASTM D 7428 ASTM D 4791

AASHTO T-104 AASHTO T-109 AASHTO T-161 AASHTO T-96 AASHTO T-327

Deleterious materials in aggregates

ASTM C 40 → ASTM C 87 → ASTM C 142 → ASTM C 123 → ASTM C 294 ASTM C 295

AASHTO T-21 AASHTO T-71 AASHTO T -112 AASHTO T-113

Test to evaluate potential alkali aggregate reactivity

ASTM C 227 ASTM C 289 ASTM C 586 ASTM C 441 ASTM C 1260 ASTM C 1293 ASTM C 1105 ASTM C 1567

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

165

6 Need of Future Research Mining of gravel and sand from river of seabed is conventional method. With increased in economic activity, more infrastructure development will take place in Southeast Asia and demand of sand and gravel will increase. Transportation cost is major cost in transportation of bulk material and thus local material needs to be used to convert into M-Sand. A study was carried out in for finding alkali silica reactivity in granite of Malaysia [67]. For every local material it is necessary to find out physical properties. Research and development centre should be developed in every region with the following objectives: 1. Resource availability for next 50 years for producing M-Sand and availability of alternative materials for production of M-Sand for sustainability. 2. Research required for compared for testing aggregates and M-Sand with different standards and comparing results. 3. To develop new standard for each local material based on application. 4. Testing of physical, chemical, mechanical and petrography for aggregates and MSand. 5. To create national database of such tests which will be useful for future research. 6. Artificial intelligence (AI) /Machine Learning (ML) application for prediction of aggregate size using image analysis, asphalt performance based on image analysis on fine aggregates and prediction of Ready mix concrete (RMX) concreter performance based on cement, aggregate and M-Sand.

7 Conclusions 1. The topic of river sand and M-Sand is quite vast. In this chapter, examples are limited from India and Malaysia and explained. 2. Illegal river sand mining is global problem where examples are given. Each nation needs to address separately. 3. Guidelines for river sand mining in India and Malaysia have been developed. If such guidelines acre discussed across different nations, best practices can be adopted. 4. Study of comparing results with V7 was carried out at Mumbai, Maharashtra, India. The river sand from nearby source was used to compare with sand produced from Kemco V7 dry sand making machine. The product was developed by using coarse aggregate from basalt in nearby area. 5. IS383:2016 [61] considers four types of gradation which can be suitable for construction. 6. Indian standard does not specify regarding cut off limit for chemical content, chloride content, impurities, organic matter in manufactured sand. 7. Alkali-silica reactivity is matter of concern and the same is not specified in IS 383:2016 [61] American standard has standard test procedure to determine potential alkali-silica reactivity of the material. European RILEM standard, ASTM 295 can identify minerals in rock which are potentially reactive through petrographic examination. Bar mortar test as per ASTM can confirm alkali-silica reactivity. 8. Based on various application for usage of manufactured sand, criteria for acceptance and rejection should be specified in Indian Standard.

166

R. M. Bhatawdekar et al.

9. At regional level, R & D centre needs to be developed for M-Sand and promotion of usage which will be great step for sustainability.

References 1. Cortes, D.D., Kim, H.K., Palomino, A.M., Santamarina, J.C.: Rheological and mechanical properties of mortars prepared with natural and manufactured sands. Cem Concr Res. 38(10), 1142–1147 (2008). https://doi.org/10.1016/j.cemconres.2008.03.020 2. Ahmad, T., Ahmad, R., Kamran, M., Wahjoedi, B., Shakoor, I., Hussain, F., Riaz, F., Jamil, Z., Isaac, S., Ashraf, Q.: Effect of thal silica sand nanoparticles and glass fiber reinforcements on epoxy-based hybrid composite. Iran Polym. J. (English Ed) 24(1), 21–27 (2014). https:// doi.org/10.1007/s13726-014-0296-x 3. Johansson, M.: Facies analysis of the plateau sandstones (eocene to early miocene?), Bako National Park, Sarawak, Malaysia. J. Asian Earth Sci. 17(1–2), 233-246 (1999). https://doi. org/10.1016/s0743-9547(98)00061-0 4. Darman, H.: An Outline of The Geology of Indonesia. PT. Lereng Nusantara Consultant, Jakarta (2000) 5. Decelles, P.G., Kapp, P., Gehrels, G.E., Ding, L.: Paleocene-eocene foreland basin evolution in The Himalaya of Southern Tibet and Nepal: implications for the age of initial India-Asia collision. Tectonics 33(5), 824–849 (2014). https://doi.org/10.1002/2014tc003522 6. Nanson, G.C., Rust, B.R., Taylor, G.: Coexistent mud braids and anastomosing channels in an Arid-Zone River: Cooper Creek, Central Australia. Geology 14(2), 175–178 (1986). https:// doi.org/10.1130/0091-7613(1986)14%3c175:cmbaac%3e2.0.co;2 7. Muhs, D.R., Vance, T.H.: Evidence of active dune sand on the great plains in the 19th century from accounts of early explorers. Q. Res. 43(2), 198–208 (1995) 8. Atkinson, B.L., Grace, M.R., Hart, B.T., Vanderkruk, K.E.: Sediment instability affects the rate and location of primary production and respiration in a sand-bed stream. J. North Am. Benthol. Soc. 27(3), 581–592 (2008). https://doi.org/10.1899/07-143.1 9. Xiao, J., Qiang, C., Nanni, A., Zhang, K.: Use of sea-sand and seawater in concrete construction: current status and future opportunities. Constr. Build. Mater. 155, 1101–1111 (2017). https://doi.org/10.1016/j.conbuildmat.2017.08.130 10. Bradley, W.C.: Effect of weathering on abrasion of granitic gravel, Colorado River (Texas). Geol. Soc. Am. Bull. 81(1), 61–80 (1970) 11. Hudson, J.: Rock Engineering Systems: Theory and Practice. High Plains Press (JAH), Chichester (1992) 12. Gallagher, L., Peduzzi, P.: Sand and Sustainability: Finding New Solutions for Environmental Governance of Global sand resources. Geneva (2019) 13. BMI Research: Global Industry Overview: Sand Miners on Solid Ground - October 2014 (2014). http://www.mining-insight.com/. Accessed 8 Aug 2020 14. Giljum, S., Hinterberger, F., Lutz, C., Meyer, B.: Accounting and modelling global resource use. In: Suh, S. (eds.) vol. 23, pp. 139–160. Springer, Dordrecht (2009) 15. Giljum, S., Lutz, C., Jungnitz, A., Bruckner, M., Hinterberger, F.: European resource use and resource productivity in a global context. In: Ekins, P., Speck, S. (eds.). p. 27. Oxford Uniersity Press, Oxford (2011) 16. UNEP GEAS.: Sand, Rarer Than One Thinks- March 2014 (2014). https://wedocs.unep.org/ bitstream/handle/20.500.11822/8665/GEAS. Accessed 9 Aug 2020 17. Chilamkurthy, K., Marckson, A.V., Chopperla, S.T., Santhanam, M.: A statistical overview of sand demand in Asia and Europe. In: International Conference UKIERE CTMC 2016. Goa (2016)

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

167

18. Yong, K.Y., Lee, S.L., Karunaratne, G.P.: Coastal reclamation in Singapore: a review. In: Urban Coastal Area Management: The Experience of Singapore. ICLARM Conference Proceedings, pp. 59–67 (1991) 19. Jefferson, T.A., Hung, S.K., Würsig, B.: Protecting small cetaceans from coastal development: impact assessment and mitigation experience in Hong Kong. Mar Policy 33(2), 305–311 (2009). https://doi.org/10.1016/j.marpol.2008.07.011 20. Jiang, L., Lin, H.: Integrated analysis of SAR interferometric and geological data for investigating long-term reclamation settlement of chek lap kok airport, Hong Kong. Eng. Geol. 110(3–4), 77-92 (2010). https://doi.org/10.1016/j.enggeo.2009.11.005 21. Khalaf, F.I., Al-Ajmi, D.: Aeolian processes and sand encroachment problems in Kuwait. Geomorphology 6(2), 111–134 (1993). https://doi.org/10.1016/0169-555x(93)90042-z 22. Arnold, C.L., Gibbons, C.J.: Impervious surface coverage: the emergence of a key environmental indicator. J. Am. Plan. Assoc. 62(2), 243–258 (1996). https://doi.org/10.1080/019443 69608975688 23. Padmalal, D., Maya, K., Sreebha, S., Sreeja, R.: Environmental effects of river sand mining: a case from the river catchments of Vembanad Lake, Southwest Coast of India. Environ. Geol. 54(4), 879–889 (2008). https://doi.org/10.1007/s00254-007-0870-z 24. Patten, D.T.: Riparian ecosystems of semi-arid North America: diversity and human impacts. Wetlands 18(4), 498–512 (1998). https://doi.org/10.1007/bf03161668 25. Arun, P.R., Sreeja, R., Sreebha, S., Maya, K., Padmalal, D.: River sand mining and its impact on physical and biological environments of Kerala Rivers, Southwest Coast of India. EcoChronicle 1(1), 1–6 (2006) 26. Ashraf, M.A., Maah, M.J., Yusoff, I., Wajid, A., Mahmood, K.: Sand mining effects, causes and concerns: a case study from Bestari Jaya, Selangor, Peninsular Malaysia. Sci. Res. Essays. 6(6), 1216–1231 (2011) 27. Sreebha, S., Padmalal, D.: Environmental impact assessment of sand mining from the small catchment rivers in the Southwestern Coast of India: a case study. Environ. Manag. 47(1), 130–140 (2011). https://doi.org/10.1007/s00267-010-9571-6 28. Saviour, M.N.: Environmental impact of soil and sand mining: a review. Int. J. Sci. Environ. 1(3), 125–134 (2012) 29. Gari, S.R., Newton, A., Icely, J.D.: A review of the application and evolution of the DPSIR framework with an emphasis on coastal social-ecological systems. Ocean Coast. Manag. 103, 63–77 (2015) 30. Lawler, D.M.: The measurement of river bank erosion and lateral channel change: a review. Earth Surf Process Landforms 18(9), 777–821 (1993). https://doi.org/10.1002/esp.329018 0905 31. Hubble, T.C.T.: Slope stability analysis of potential bank failure as a result of toe erosion on weir-impounded lakes: an example from the Nepean River, New South Wales, Australia. Mar Freshw Res. 55(1), 57–65 (2004). https://doi.org/10.1071/mf03003 32. Binti Sa’adin, S.L., Kaewunruen, S., Jaroszweski, D.: Operational readiness for climate change of Malaysia high-speed rail. In: Proceedings of the Institution of Civil Engineers - Transport, pp. 308–320. Thomas Telford Services Ltd (2016) 33. Sharma, M.L.: Organised crime in India: problems & perspectives. Resour. Mater. Ser. 54, 82–129 (1999) 34. Martinez-Alier, J., Temper, L., Demaria, F.: Social metabolism and environmental conflicts in India. In: Nature, Economy and Society, pp. 19–49. Springer, India (2015) 35. Rege, A.: Not biting the dust: using a tripartite model of organized crime to examine India’s Sand Mafia. Int. J. Comp. Appl. Crim. Justice. 40(2), 101–121 (2016). https://doi.org/10. 1080/01924036.2015.1082486 36. Gavriletea, M.: Environmental impacts of sand exploitation. Analysis of sand market. Sustainability 9(7), 1118 (2017). https://doi.org/10.3390/su9071118

168

R. M. Bhatawdekar et al.

37. Koehnken, L., Rintoul, M.: Impacts of sand mining on ecosystem structure, process and biodiversity in rivers. WWF Rev. 159 (2018) 38. Villas-Bôas, R.C., Barreto, M.L.: Cierre de Minas: Experiencias En Iberoamérica (2000) 39. Myers, N., Mittermeler, R.A., Mittermeler, C.G., Da Fonseca, G.A.B., Kent, J.: Biodiversity hotspots for conservation priorities. Nature 403(6772), 853–858 (2000). https://doi.org/10. 1038/35002501 40. John, E.: The Impacts of sand mining in Kallada River (Pathanapuram Taluk), Kerala. J. Basic Appl. Biol. 3, 108–113 (2009) 41. Kondolf, G.M.: PROFILE: hungry water: effects of dams and gravel mining on river channels. Environ. Manag. 21(4), 533–551 (1997) 42. Thornton, P.K., Jones, P.G., Owiyo, T., Kruska, R.L., Herrero, M.T., Kristjanson, P.M., Notenbaert, A.M.O., Bekele, N., Omolo, A.: Mapping climate vulnerability and poverty in Africa. ILRI, Nairobi (2006) 43. Pye, K., Neal, A.: Coastal dune erosion at formby point, North Merseyside, England: causes and mechanisms. Mar. Geol. 119(1–2), 39-56 (1994). https://doi.org/10.1016/00253227(94)90139-2 44. Hilton, M.J., Hesp, P.: Determining the limits of beach-nearshore sand systems and the impact of offshore coastal sand mining. J. Coast Res. 12(2), 496–519 (1996). https://doi.org/10.2307/ 4298500 45. DuBois, R., Towle, E.L.: Coral Harvesting and Sand Mining Management Practices (1984) 46. Salahuddin, J.S.: River Sand Mining Management Guideline, pp. 1–47 (2009) 47. Kaya, T., Hashimoto, M., Pettingell, H.: The development of sand manufacture from crushed rock in Japan, using advanced VSI technology. In: Proceedings of the International Center for Aggregates Research Symposium, pp. 4–5. Texas (2009) 48. Hämäläinen, E.: Manufactured sand success with Velde Pukk in Norway. In: Metso’s Customer Magazine for the Mining and Construction Industries-Results: Minerals & Aggregates, pp. 6–7 (2010) 49. Westerholm, M., Lagerblad, B.: Filler and filler quality of crushed rocks in concrete production. In: International Conference on Building Materials “18.Ibausil.” Wemar (2012) 50. Cepuritis, R.: Sand from the rocks. new type of crushed sand to replace natural sand in concrete production. Bulk Solids Handling 3, 42–48 (2014) 51. Cepuritis, R.: Building on sand. Int. Cem. Rev. 81–83 (2014) 52. Cepur¯ıtis, R.: Manufactured sand crushing process parameters: short review and evaluation for sand performance in fresh concrete. Nord Concr Res. 48, 27–48 (2013) 53. Prakash, V.M.: Ready mixed concrete using manufactured sand as fine aggregate. In: 32nd Conference on Our World In Concrete & Structures. Singapore (2007) 54. Nanthagopalan, P., Santhanam, M.: Fresh and hardened properties of self-compacting concrete produced with manufactured sand. Cem. Concr. Compos. 33(3), 353–358 (2011). https://doi. org/10.1016/j.cemconcomp.2010.11.005 55. Jadhav, P.A., Kulkarni, D.K.: Effect of replacement of natural sand by manufactured sand on thee properties of cement mortar. Int. J. Civ. Struct. Eng. 3(3), 621–628 (2013). https://doi. org/10.6088/ijcser.222012030130577 56. Jadhav, P.A., Kulkarni, D.K.: An experimental investigation on the properties of concrete containing manufactured sand. Int. J. Adv. Eng. Technol. 3(2), 101–104 (2012) 57. Pettingell, H.: The Kemco RC7. Quarr. Manag. 36(12), 11–14 (1975) 58. Pedersen, B., Hotvedt, O.: Production and Utilisation of Manufactured Sand: State-Of-TheArt Report. Blindern (2009) 59. Wigum, B.J.: Summary report. Production of High Quality Manufactured Aggregate for Concrete. Blindern (2015)

River Sand Mining Vis a Vis Manufactured Sand for Sustainability

169

60. Pilegis, M., Gardner, D., Lark, R.: An investigation into the use of manufactured sand as a 100% replacement for fine aggregate in concrete. Materials 9(6), 440–440 (2016). https://doi. org/10.3390/ma9060440 61. IS 383:2016, I.: Indian Standard: Coarse and Fine Aggregate for Concrete - Specification (Third Revision) (2016) 62. The National Stone, Sand, Gravel Associatiom.: The Aggregates Handbook: Second Edition. Mercury Publishing Services, Alexandria (2013) 63. Kessler, D.W., Insley, H., Sligh, W.H.: Physical, mineralogical and durability studies on the building and monumental granites of the United States. J. Res. Natl. Bur. 24, 161–206 (1940) 64. Powers, T.C.: The Properties of Fresh Concrete. Wiley, New York (1968) 65. Roy, R.F., Beck, A.E., Touloukian, Y.S.: Thermophysical properties of rocks. In: Touloukian, Y.S., Judd, W.R., Roy, R.F. (eds.). pp. 409–502. McGraw-Hill, New York (1981) 66. QRating Concrete: Newsletter Bulletin, vol. 2. Ready Mixed Concrete Manufacturers’Association (RMCMA). Mumbai (2017) 67. Murlidhar, B.R., Mohamad, E.T., Armaghani, D.J.: Potential alkali silica reactivity of various rock types in an aggregate granite quarry. Measurement 81, 221–231 (2016). https://doi.org/ 10.1016/j.measurement.2015.12.022

Evaluating the Effect of Meteorological Conditions on Blast-Induced Air Over-Pressure in Open Pit Coal Mines Quang-Hieu Tran1(B) , Hoang Nguyen1 , Xuan-Nam Bui1 , Carsten Drebenstedt2 , Belin Vladimir Arnoldovich3 , Victor Atrushkevich3 , and Van-Duc Nguyen4 1 Department of Surface Mining, Mining Faculty, Hanoi University of Mining and Geology,

Duc Thang Bac Tu Liem, Hanoi 10000, Vietnam [email protected] 2 Institute for Mining and Civil Engineering, Technische Universität Bergakademie Freiberg, 09599 Freiberg, Germany 3 Mining Institute, National University of Science and Technology “MISiS”, 117279 Moscow, Russia 4 Department of Energy and Mineral Resources, College of Engineering, Dong-A University, Busan 49315, South Korea

Abstract. Air blast waves are recognized as one of the negative effects induced by blasting operations in open pit mines. The intensity of air over-pressure is taken into account as the primary parameter to determine the damages on the surrounding environment. Many researchers commented on the effects of meteorology on blastinduced air overpressure, such as temperature, relative air pressure, wind direction and speed, air humidity, to name a few. However, they were not fully addressed. Therefore, this study aims to fully address the effect of meteorology conditions on blast-induced air over-pressure in open pit mines through the air over-pressure predictive models. Nui Beo open pit mine in Quang Ninh province of Vietnam, where suffers significantly from tropical climate with two generally rainy and dry seasons, was selected as a case study for this aim. The results revealed that the meteorological conditions have a great effect on blast-induced air over-pressure, especially air humidity and wind speed. These contribute to enhance the effect of blasting operation for Nui Beo coal mine in particular, and for all coal mines in Vietnam. Keywords: Blast · Meteorological conditions · Over-pressure · Open pit mine · Nui Beo · Quang Ninh · Vietnam

1 Introduction The influence of intense monsoons characterizes Vietnam’s climate, but with much sunshine, high rainfall, and high humidity. The regions near the tropics and the mountainous areas have a temperate climate. Monsoon climate also affects the change of © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 170–186, 2021. https://doi.org/10.1007/978-3-030-60839-2_9

Evaluating the Effect of Meteorological Conditions

171

tropical moisture. In Vietnam, there are generally two seasons: the hot season is from May to October and the cold season is from November to next April. The difference in temperature between the two seasons in the South is almost imperceptible, with an average of 3 °C. Most notably, in the North, the difference of 12 °C is observed. There are four distinct seasons, most evident in the Northern provinces (from Hai Van toward the North): Spring, Summer, Autumn, Winter. The northeastern region includes the northern and northeastern provinces in Vietnam such as Lao Cai, Yen Bai, Hoa Binh, Ha Giang, Tuyen Quang, Phu Tho, Cao Bang, Lang Son, Bac Kan, Thai Nguyen and Quang Ninh. Northeast monsoons strongly influence the climate. A cold, cloudy winter (little sunshine) is characterized by drizzle. The cold wave came earlier than other provinces. Summer is hot and rainy coincides with the rainy season. However, unlike in the northwest, drought conditions are rare due to low westerly wind frequency. The rainy season usually lasts from May to September, although it can vary from 4 to 10 months. The average annual temperature in the coastal areas is around 22 °C, with the coldest month having an average temperature of 15 °C and the hottest month having an average temperature of 38 °C. The average annual rainfall in coastal areas is about 1,800 mm. The annual average humidity is 84%, the lowest 81%, and the average humidity of many years is 82.3% [1]. For open pit mining, the blasting method can be considered as the most economical technique used to fragment rock blocks. However, only 20–30% of the energy is used for the fragmentation and movement of the rock. At the same time, the rest is wasted in the form of ground vibrations, air explosions, noise, and flying rocks [1]. Both ground vibrations and air explosions are of great concern as they will damage existing surface structures and cause a nuisance to residents near the mine areas, which have too much access to the populated area. To analyze problems related to vibrations, it is necessary to consider the combined effects of several factors such as site characteristics, surface propagation, underground muscle waves, and structural response. The best approach to estimate the weight of the load, which at a given distance generates vibrations below the safe limit, is to use instrumentation on bursts to determine the constants of the actual blasting conditions. Besides, the effective control of vibration-related problems requires the development of a reliable vibration monitoring system and the evaluation of the attenuation characteristics of various vibrations [2, 21, 22] (Fig. 1). Airblast waves are air pressure waves produced by explosion operation. As vibrations on the ground, these pressure waves can be described by time histories where the amplitude is the air pressure instead of the particle’s velocity [11–13]. The high-frequency portion of the pressure wave is audible and the sound accompanying the blasting; the low-frequency portion is inaudible but stimulates the structure and causes a secondary click and sound in the structure. Airblast may affect crack propagation in walls, windows, and human reaction [3, 27, 28]. The degree of damage depends on the air pressure waves. Since it is influenced by climatic conditions such as air temperature, air humidity, and wind speed, etc. The determination of the safe distance when blasting at surface coal mines in Quang Ninh near residents requires research on the effect of climatic conditions on air blasting [5–7]. Airblast waves are recognized as one of the negative effects induced by blasting operations in open pit mines [23–26]. The intensity of air over-pressure is taken into account

172

Q.-H. Tran et al.

Fig. 1. Influence of shock waves, airwaves to the protective structures when conducting blasting

as the primary parameter to determine the damages on the surrounding environment [14–17]. Many researchers commented on the effects of meteorology on blast-induced air overpressure, such as temperature, relative air pressure, wind direction and speed, air humidity, to name a few [9, 10, 29–31]. However, they were not fully addressed. In this study, the data measurements of air blast waves in Nui Beo coal mine of phosphate were analyzed according to several established criteria of damage to evaluate the ground vibrations characteristics resulting from production blasts and to evaluate the vibrations impact on the nearby structures. This study aims to fully address the effect of meteorology conditions on blast-induced airover-pressure in open pit mines through the air over-pressure predictive models. An open pit mine in Quang Ninh, Vietnam was selected as a case study for this aim. The results revealed that the meteorology conditions have a significant effect on blast-induced airover-pressure, especially air humidity and wind speed [8].

2 Materials and Methods 2.1 Study Site The study area is the Nui Beo open pit coal mine located in the Ha Long city, Quang Ninh province, Vietnam, about 160 km east of the Hanoi city. This mine is managed by the Nui Beo Coal Joint-Stock Company belonging to the Vietnam National Coal and Mineral Industries Group (VINACOMIN). The total coal field is around 3.75 km2 for the open pit coal mine and 5.6 km2 for the underground coal mine. In terms of topographical, the Nui Beo coal mine has complicated terrain conditions where the center is the open pit mining area (Fig. 2) [1, 2, 4].

Evaluating the Effect of Meteorological Conditions

173

Fig. 2. Overview of Nui Beo open pit coal mine (Vinacomin)

Quang Ninh province is located in tropical climates [1]. Climate is divided into two seasons, and the rainy season starts from May to October and the dry season from November to April next year. Dry season temperature changes from 16–21 °C, the lowest year 4 °C. The rainy season temperature changes in the range of 24–35 °C, average 28– 30 °C, the highest 38 °C. The average annual humidity is 84% (the highest is 90% in 2007, the lowest is 81% in 2003), and the average humidity is 82.3% for recent years (Figs. 3, 4 and 5).

Air temperature, OC

35 30 25 20 15 10 1

2

3

4 2017

5

6 2018

7

8

9

2019

10 11 12 Month

Fig. 3. The average temperature changes annually in Quang Ninh - Vietnam

174

Q.-H. Tran et al.

Fig. 4. The average humidity changes annually in Quang Ninh - Vietnam

Fig. 5. Average rainfall varies from year to year in Quang Ninh - Vietnam

2.2 Multiple AQ (Air Quality) Monitoring System The AQ sensors are designed and built to monitor the ambient (indoor & outdoor) air quality over the large space such as mining and construction sites. Since the wireless realtime monitoring system is equipped with XBee technology has long-term continuous monitoring capability. The built-in multiple gas and dust modules can measure most of the regulated hazardous pollutants generated from the industrial working sites and monitor the atmospheric conditions simultaneously [8, 18–20]. Multiple sensors for

Evaluating the Effect of Meteorological Conditions

175

indoor and outdoor ambient monitoring is installed and can be used to measure the regulated gases, dust levels, and other parameters such as:

Fig. 6. Multiple sensors and attachments installed for indoor and outdoor ambient monitoring

– – – –

Gases: CO, CO2 , NO, NO2 , SO2 , VOC. Dust: Airborne dust. Atmospheric pressure and temperature/humidity. Wind speed and direction.

The sensors can be also positioned over a large area as each sensor can act as a repeater; sensor networks can span kilometers. GPS module will locate sensors exactly at the position. Multiple sensors can be relayed to the router connected to the PC. Individual sensors can do a repeater.

176

Q.-H. Tran et al.

Multiple sensors and Xbee router are deployed at target locations. Readings are transmitted between the sensors and also to Xbee route. PC will receive all the measured data through USB. If the data is connected to the cellular network, it will be transmitted to mobile devices (Fig. 7).

Fig. 7. Multiple sensors and Xbee router are deployed on Nui Beo coal mine (Vinacomin)

Output management: According to the user’s intention, measurement data can be displayed on the PC monitor and managed in real-time. Optionally, mobile devices can be logged in to show the real-time data (Fig. 8).

Fig. 8. mobile devices can be logged in to show the real-time data

Evaluating the Effect of Meteorological Conditions

177

Fig. 9. Installation of the monitoring systems for temperature, humidity and wind speed at Nui Beo coal mine

2.3 Blasting Vibration and Tilt Wireless Monitoring Solutions Real-Time Wireless Sensors are built to be deployed on mining and construction sites, tunnels, bridges, pathways, and other structures. The system is designed with the ultimate goal of understanding challenges and provide solutions to protect assets by the low-cost system with high accuracy and reliability. Structural Safety Monitoring System are used primarily for structural safety monitoring to ensure assets such as mines, tunnels, buildings, bridges, and other structures are protected. Our sensors also have the ability to monitor both vibration and tilt simultaneously. Thanks to this ability, the effects of blasting vibration on structural safety can be understood in real-time. Sensor & Gateway for monitoring & routing: Tilt & Vibration wireless sensor provides real-time monitoring data. Gateway products quickly route data and alerts to the desired locations (i.e., mobile phones, FTP Server, GIS). USB Management Node is a dongle attached to the PC and establishes a wireless mesh network. It enables bidirectional communication with a maximum 100 sensors. Sensors have been designed for easy installation and wireless remote management with the objective of providing readings and alerts when user-defined allowable limit is exceeded (Fig. 10).

178

Q.-H. Tran et al.

Fig. 10. Real-Time Wireless Sensors

Sensors can be positioned over a large area as each sensor can act as a repeater; sensor networks can span kilometers. Multiple sensors are deployed at target locations. Readings are transmitted to the Management Node (Management dongle) connected to PC. Since our sensors have built-in router function, all readings are sent to the sensor near the Management Node, and, in turn, to the Node (Fig. 11).

Fig. 11. Mining & construction site monitoring on Nui Beo coal mine

3 Results and Discussion The blasting test parameters at the Nui Beo coal mine is shown in Table 1 and the monitoring results of air blast overpressure at this mine is presented in Table 2. From

Evaluating the Effect of Meteorological Conditions

179

these results, we can determine the effect of air humidity on the intensity of air blast waves (Fig. 12); the effect of wind speed on the intensity of air blast waves (Fig. 13); the effect of wind direction on the intensity of air blast waves (Fig. 14). Table 1. Blasting test parameters at Nui Beo coal mine Hole Bench Powder diameter, high, factor, mm H, m q (kg/m3) 105

Burden, m

Hole Row Subdrill, Stemming, Powder spacing spacing LKT, m Lb, m column a, m b, m length Lt, m

8 ÷ 10 0,3 ÷ 0,35 9,5 ÷ 11,5 3,8

3,8

1,5

4,9

6,6

Table 2. The monitoring results of air blast overpressure from the blast at Nui Beo coal mine Blast log number

01-TT3/HCKNM

02-TT3/HCKNM

03-TT3/HCKNM

04-TT3/HCKNM

05-TT3/HCKNM

06-TT3/HCKNM

07-TT3/HCKNM

Air Air Wind Total Charge Distance Scale Air Wind temperature humidity speed charge, hole R, m distance blast, direction √ T, oC ϕ, % v, Q, kg per R/ 3 Q P, Pa m/s delay Qvs, kg 32

30

34

35

29

28,5

22

50

52

59

65

62

48

96

2,34

9212

830

677

72,04

8,4 −

1,26

9212

830

715

76,08

9,7 −

2,74

13,4 +

9212

830

495

52,67

2,86 10060

880

481

50,19

14,6 −

2,42 10060

880

523

54,58

16,9 −

0,68 10060

880

338

35,27

28,8 +

9,04

2889

315

392

57,61

37,3 +

4,20

2889

315

339

49,82

11,5 +

3,38

2889

315

499

73,34

6,0 −

10,36

6874

905

138

14,27

78,7 +

4,09

6874

905

563

58,20

8,3 −

7,11

6874

905

302

31,22

18,0 −

9,95

4919

205

122

20,69

92,0 +

0,75

4919

205

683

115,83

4,8 −

2,45

4919

205

481

81,58

10,2 −

5,03

1943

302

335

49,93

9,8 −

5,11

1943

302

267

39,80

15,5 +

6,23

1943

302

285

42,48

9,3 −

18,0 10171

905

133

13,7

260,0 +

19,7 10171

905

199

20,6

179,0 −

(continued)

180

Q.-H. Tran et al. Table 2. (continued)

Blast log number

Air Air Wind Total Charge Distance Scale Air Wind temperature humidity speed charge, hole R, m distance blast, direction √ T, oC ϕ, % v, Q, kg per R/ 3 Q P, Pa m/s delay Qvs, kg

08-TT3/HCKNM

09-TT3/HCKNM

10-TT3/HCKNM

18

14

15

11-TT3/HCKNM 15, 5

12-TT3/HCKNM

19

92

95

88

80

82

4,5 10171

905

339

35,0

35,0 −

6,9

9199

823

330

35,2

14,3 −

6,9

9199

823

189

20,2

51,3 +

5,7

9199

823

301

32,1

23,0 −

5,1

6901

922

327

33,6

22,8 −

4,2

6901

922

206

21,2

66,5 +

4,7

6901

922

291

29,9

43,5 −

4,7

5114

213

196

32,8

25,0 +

3,8

5114

213

719

120,4

9,0 −

4,7

5114

213

399

66,8

18,0 −

4,3

2562

288

218

33,0

26,0 +

1,2

2562

288

682

103,3

6,3 +

4,0

2562

288

391

59,2

18,3 −

6,2

2134

312

335

49,4

29,0 +

5,0

2134

312

267

39,4

32,0 −

3,4

2134

312

291

42,9

18,0 +

Forward wind direction (+); Weak wind direction (-).

Fig. 12. The effect of air humidity on the intensity of air blast waves

Evaluating the Effect of Meteorological Conditions

181

Fig. 13. The effect of wind speed on the intensity of air blast waves

Fig. 14. The effect of wind direction on the intensity of air blast waves

Base on the monitoring results in Table 2, the authors have built the graphical √ relationship between the air blast and scale distance R/ 3 Q as in Fig. 15.

182

Q.-H. Tran et al.

√ Fig. 15. Relationship between the air blast overpressure and scale distance R/ 3 Q

Then, air blast overpressure can be calculated as follows:   R −1,485 , Pa P = 5598 √ 3 Q

(1)

   R −1,485 , kg cm2 P = 0, 05598 √ 3 Q

(2)

or

When the wind is mentioned (use of the wind scale table), the air blast can be determined by the following equation:    υ2μ R −1,485 , kg cm2 + P0 P = 0, 05598 √ 3 2R1 T Q

(3)

where: P0 is atmospheric pressure during blasting operations, P0 = 101.325 Pa; v is wind speed, m/s; μ is Molar mass of air, μ = 28,9.10−3 kg/mol; R1 is gas constant for air, R1 = 8,31 J/mol.K; T is air temperature during blasting operations (Kelvin, T = 273 + t°C). According to legal documents [4], to protect the civil structures existing near the mining sites, air blast overpressure level must be in permitted limits (airblast overpressure limits depend on the distance from the blast, total charge and civil structures). That means P ≤ Pcp . From Eq. (3), relationship between the safety distance for air blast overpressure and the total explosive charges per delay with any weather conditions can be defined as follows (in case P = Pcp = 130 Pa): Qcp

 = 17, 86Pcp −

 Po ν 2 μ R3 , kg 0.112R1 T

(4)

Evaluating the Effect of Meteorological Conditions

183

Table 3. Total explosive charges per delay estimated by Eq. (4) Distance from blast R, 100 150 m

200

250

300

350

400

450

500

Wind speed Air temperature T = 150 °C v = 5 m/s

420 1.417 3.358 6.559 11.334 17.998 26.865 38.252 52.471

v = 7 m/s

319 1.077 2.553 4.987 8.617

13.684 20.426 29.083 39.895

v = 9 m/s

206 697

1.652 3.226 5.575

8.852

13.214 18.815 25.809

v = 11 m/s

100 337

800

4.287

6.399

1.562 2.700

9.111

12.498

Wind speed Air temperature T = 200 °C v = 5 m/s

422 1.423 3.373 6.588 11.384 18.077 26.984 38.420 52.703

v = 7 m/s

322 1.088 2.579 5.038 8.706

13.824 20.635 29.381 40.304

v = 9 m/s

211 712

1.687 3.294 5.693

9.040

13.493 19.212 26.354

v = 11 m/s

105 353

836

4.483

6.691

1.634 2.823

9.527

13.069

Wind speed Air temperature T = 250 °C v = 5 m/s

423 1.429 3.388 6.617 11.434 18.157 27.102 38.589 52.934

v = 7 m/s

326 1.099 2.605 5.088 8.791

13.960 20.839 29.671 40.701

v = 9 m/s

215 726

1.721 3.361 5.808

9.222

13.766 19.600 26.887

v = 11 m/s

109 368

872

4.676

6.979

1.704 2.944

9.938

13.632

Wind speed Air temperature T = 300 °C v = 5 m/s

425 1.435 3.402 6.645 11.482 18.233 27.217 38.753 53.159

v = 7 m/s

329 1.109 2.630 5.136 8.875

14.093 21.037 29.952 41.087

v = 9 m/s

219 740

1.754 3.426 5.920

9.400

14.032 19.979 27.407

v = 11 m/s

114 383

908

4.866

7.264

1.773 3.065

10.343 14.188

Wind speed Air temperature T = 350 °C v = 5 m/s

427 1.441 3.416 6.672 11.529 18.308 27.329 38.911 53.376

v = 7 m/s

332 1.119 2.654 5.183 8.956

14.222 21.229 30.226 41.462

v = 9 m/s

223 754

1.787 3.489 6.030

9.575

14.292 20.350 27.914

v = 11 m/s

118 398

943

5.054

7.545

1.842 3.183

10.743 14.736

4 Conclusions 1. Studying the combined effect of the temperature factor and wind speed on P and, in general, on the parameters of blast-blowing, which is typical in winter in Vietnam, from the results in Table 3, we note that for P = 130 Pa at a fixed value with wind speed of 7–9 m/s and air temperature of 15 °C–25 °C, the explosion power decreases by 1,02–1,1 times, and when the temperature changes from 25 °C to 35 °C, the explosion power decreases by 1,07–1,2 times.

184

Q.-H. Tran et al.

2. When the air temperature T is lower than 25 °C, in comparison with the normal condition of the standard atmosphere, it is necessary to reduce the power of explosive by at least 20–30%. During blasting operations in wind conditions, there is a significant decrease in the maximum permissible value of excess pressure in the front of the shock wave and the magnitude of the explosion power at various distances to protect objects. In particular, in the case of T = 25 °C in Table 3, it can be seen that for P = 130 Pa and a distance interval of 100 ≤ R ≤ 500 m at wind speeds v = 7–9 m/s, in comparison with the condition of no wind (v ≤ 5 m/s), the maximum allowable explosion power decreases approximately by 1,5–2,2 times. At the wind speeds v = 9–11 m/s, in comparison with the condition of without wind (v ≤ 5 m/s), the maximum allowable explosion power decreases approximately by 2,2–6,2 times. 3. Thus, from the calculated experimental studies carried out in this study, it follows that when carrying out massive explosions at the Nui Beo coal mines in autumnwinter climatic conditions, when the air temperature in the region is around 15 °C, compared to T = 25 °C, the power of explosions must be reduced by about one third. 4. Based on the air blast overpressure monitoring at the residential sites next to Nui Beo coal mine (Vinacomin), it is found that air blast overpressure depends on air temperature and wind speed during blasting operations. According to the results in Table 3 the total explosive charge per delay and safety distance depend on weather conditions. Careful estimation of the blasting charge based on the weather condition is necessary to improve blast efficiency in limestone quarries and minimize adverse impacts on the environment.

References 1. Report on the results of monitoring stations National Hydrometeorology Bai Chay 2. Reporting the results of blasting monitoring at Nui Beo surface coal mine – Vinacomin (2019) 3. QCVN 02:2008/BCT, National technical regulation on safety in the storage, transportation, use and disposal of industrial explosive materials 4. Nui Beo Joint Stock company: The environmental impact assessment report on the project of investments and construction of underground mining of the Nui Beo coal mine 5. Van Bach, N., Van Thanh, N.: Impact of big explosions and some measures to protect buildings. J. Mining Ind. 4(1996), 13–14 (1996) 6. Van Bach, N., et al.: Measures to minimize the impact of shock when blasting in Nui Beo mine. J. Mining Sci. Technol. 14(2006), 58–62 (2006) 7. Tran, Q.H., et al.: Impact of shock waves and blasting on surface works when exploiting at Dong Trang Bach mine, Uong Bi, Quang Ninh. In: National Conference on Earth Science and Resources with Sustainable Development (ERSD 2018), Hanoi, pp. 49–55 (2018) 8. Bui, X.N.: Shock from blasting on open pit - characteristics and influencing factors. J. Mining Ind. 05(2006), 20–22 (2006) 9. Afeni, T.B., Osasan, S.K.: Assessment of noise and ground vibration induced during blasting operations in an open pit mine—a case study on Ewekoro limestone quarry, Nigeria. Mining Sci. Technol. (China) 19(4), 420–424 (2009) 10. Ak, H., Konuk, A.: The effect of discontinuity frequency on ground vibrations produced from bench blasting: a case study. Soil Dyn. Earthq. Eng. 28(9), 686–694 (2008)

Evaluating the Effect of Meteorological Conditions

185

11. Aldas và, G.G.U., Ecevitoglu, B.: Waveform analysis in mitigation of blast-induced vibrations. J. Appl. Geophys. 66(1–2), 25–30 (2008) 12. Nguyen Dinh, A., et al.: Analyzing and evaluating the research result of experimental blasts at Nui Beo surface coal mine to reduce ground vibration and air blast near residential area. In: The 3rd International Conference on Advances in Mining and Tunneling, Publishing House for Science and Technology, Vung Tau, tr. 79 (2014) 13. Uysal, O., Erarslan, K., Cebi, M.A., Akcakoca, H.: Effect of barrier holes on blast induced vibration. Int. J. Rock Mech. Mining Sci. 45, 712–719 (2008) 14. Adhikari, G.R.: Role of blast design parameters on ground vibration and correlation of vibration level to blasting damage to surface structures. S&T Project Report: MT/134/02 (2005) 15. Ozer, U., Kahriman, A., Aksoy, M., Adiguzel, D., Karadogan, A.: The analysis of ground vibrations induced by bench blasting at Akyol quarry and practical blasting charts. Environ. Geol. 54, 737–743 (2008) 16. Azizabadi, H.R.M., Mansouri, H., Fouché, O.: Coupling of two methods, waveform superposition and numerical, to model blast vibration effect on slope stability in jointed rock masses. Comput. Geotech. 61, 42–49 (2014) 17. Saadat, M., Khandelwal, M., Monjezi, M.: An ANN-based approach to predict blast-induced ground vibration of Gol-E-Gohar iron ore mine, Iran. J. Rock Mech. Geotech. Eng. 6, 67–76 (2014) 18. Elsemain, I.A.: Measurement and analysis of the effect of ground vibrations induced by blasting at the limestone quarries of the Egyptian cement company. College of Engineering, Assiut University, ASIUT EGYPT (2000) 19. Ak, H., Iphar, M., Yavuz, M., Konuk, A.: Evaluation of ground vibration effect of blasting operation in a magnesite mine. Soil Dyn. Earthq. Eng. 29, 669–676 (2008) 20. Adhikari, G.R., Singh, M.M.: Influence of rock properties on blast-induced vibration. Mining Sci. Technol. 8, 297–300 (1989) 21. Giraudi, A., Cardu, M., Kecojevic, V.: An assessment of blasting vibrations: a case study on quarry operation. Am. J. Environ Sci. 5, 468–474 (2009) 22. Simangunsong, G.M., Wahyudi, S.: Effect of bedding plane on prediction blast-induced ground vibration in open pit coal mines International. J. Rock Mech. Mining Sci. 79, 1–8 (2015) 23. Kumar, R., Choudhury, D., Bhargava, K.: Determination of blast-induced ground vibration equations for rocks using mechanical and geological properties. J. Rock Mech. Geotech. Eng. 8(3), 341–349 (2016) 24. Faramarzi, F., Ebrahimi, F., Mohammad, A., Mansouri, H.: Simultaneous investigation of blast induced ground vibration and air blast effects on safety level of structures and human in surface blasting. Int. J. Mining Sci. Technol. 24, 663–669 (2014) 25. Nicholls, H.R., Johnson, C.F., Duvall, W.I.: Blasting vibrations and their effects on structures, U.S. Department of Interior, Bureau of Mines Bulletin, vol. 656 (1971) 26. Edwards, M., Rudenko, P.G.D.: Site attenuation and vibramap Study for Hanson aggregates east Crabtree quarry Raleigh, Wake country, North Carolina. Vibra-Tech Engineers report (2011) 27. Ganopolcki M.I, Bapon B.L, Belin B.A., Pypkov B.B., Civenkov B.I. Metody vedeni vzpyvnyx pabot, cpecialnye vzpyvnye paboty. Bzpyvnoe delo. 2007 28. Ganopolcki M.I. K pacqety davleni vo fponte ydapno vozdyxno volny ppi maccovyx vzpyvax ckvainnyx zapdov/ M.I. Ganopolcki, .I. Cetlin// Gopny ypnal. - 1980. - №1. - C. 44–46 29. Dpykovany M.F. Metody yppavleni vzpyvom na kapepax. Hedpa, 1973.-402c 30. Kytyzov, B.H.: Bezopacnoct vzpyvnyx pabot v gopnom dele i ppomyxlennocti.M.: Gopna kniga, Izd-vo MGGU, 2009.-670c

186

Q.-H. Tran et al.

31. Cadovcki, M.A.: Ocenka cecmiqecko opacnocti ppi vzpyvax. Tpydy cecmologiqeckogo in-ta AH CCCP, M., 1941 g 32. Cadovcki, M.A.: Mexaniqeckoe dectvie vozdyxnyx ydapnyx voln vzpyva po dannym kcpepimentalnyx iccledovani // Fizika vzpyva, №1. - M.: Iz-vo AH CCCP. - 1952. - C. 20–110

Development of a Blasting Vibration Monitoring System Based on Tri-axial Acceleration Sensor for Wireless Mesh Network Monitoring Won-Ho Heo1(B) , Jung-hun Kim2 , Van-Duc Nguyen3(B) , Quang-Hieu Tran4 , Hoang Nguyen4 , Xuan-Nam Bui4 , and Chang-Woo Lee3(B) 1 Mining Tech Co., Ltd., Saha-gu, Busan 49315, South Korea

[email protected] 2 Fordev Co., Ltd., Sasang-gu, Busan 49315, South Korea 3 Department of Energy and Mineral Resources, College of Engineering, Dong-A University,

Busan 49315, South Korea [email protected], [email protected] 4 Department of Surface Mining, Mining Faculty, Hanoi University of Mining and Geology, Duc Thang, Bac Tu Liem, Hanoi 100000, Vietnam

Abstract. Recently a variety of vibration monitoring devices based on MEMS (micro electro-mechanical system) 3-axis acceleration sensor has been introduced and is gradually replacing analog wire-type geophones for blasting vibration monitoring. Blasting vibration monitoring tasks generally require frequent movement of the monitoring devices. Since accurate device set along the vertical axis is essential at a new location, acceleration sensors sensitive to the gravitational acceleration are not suitable for accurate monitoring of the blasting vibration. In this study, the vibration monitoring system with a 3-axis MEMS acceleration sensor is developed for wireless mesh network monitoring. Individual monitoring units are equipped with an algorithm for reorientation along the direction of gravity once they are placed on a particular baseline. The algorithm aims at automatically adjusting the z-axis and resetting the zero offset value altered after each blasting vibration monitoring and relocation. With this feature, it shows individual unit can be applied as conventional portable devices as well. In addition, comparative studies are also carried out along with conventional units for 3-axis acceleration and primary frequency analysis. There are several advantages of the developed system. Firstly, this system has been designed for easy installation and wireless remote management to provide readings and alerts when the user-defined allowable limit is exceeded. Secondly, due to remote management, it can improve staff safety, reduce human resources, and save time and cost. Thirdly, this system can be positioned over a large area as each sensor can act as a repeater. Finally, multiple sensors can be installed to measure various locations monitoring at the same time. Furthermore, without the cables to interface with operations or accidental damage, this system improves safety and reduces maintenance costs. The readings from the multiple sensors deployed at target locations are transmitted to the management node connected to the PC. Thus, all the live data can be seen on the PC. This system is built to be deployed on mining and construction sites, tunnel, bridges, and other structures. The system is designed with the ultimate goal of © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 187–202, 2021. https://doi.org/10.1007/978-3-030-60839-2_10

188

W.-H. Heo et al. understanding challenges and provide solutions to protect assets by the low-cost system with high accuracy and reliability. Keywords: Blasting vibration · Vibration monitoring · MEMS · Zero offset · Wireless remote management

1 Backgrounds Excavation using blasting is to remove rocks from the excavation section by using the impact and the gas pressure generated by the explosion of explosives. The impact and the gas pressure generated during blasting work spread deep into the rock in the form of elastic waves, causing ground vibrations and leading to the destruction of the rock. However, this vibration energy is not limited to rock destruction. However, only 5-20% of vibration energy is transmitted to the nearby ground in the form of elastic waves, which is called the vibration pollution caused by blasting. Table 1 shows the factors influencing blasting vibration. To minimize vibration damage caused by blasting, each country regulates the allowable vibration limits to nearby target structures during blasting, and the criteria are listed in Table 2. In Korea, since 2005, most of blasting vibration allowances have been limited under 0.3 cm/sec for any residential structures and also even strictly limited under 0.2 cm/sec in the urban area. And for a case of livestock (any animal), the usual limitation has been kept under 0.09 cm/sec. As shown in the table above, since countries around the world clearly define the acceptance criteria for vibration, accurate blasting vibration measurement is required for each blasting operation. However, internationally-renowned blasting vibration monitoring-only instruments are still expensive measurement procedures that are also laborious and time-consuming. Therefore, demands for wireless communication, IoT utilization, and unmanned automatic measurement are increasing.

Development of a Blasting Vibration Monitoring System

189

Table 1. The influential factors the blasting vibration

∙Geological conditions (the type of rock) ∙ Height of bench ∙ Number of free faces ∙ The angle of the free plane towards the structure ∙ Minimum burden and spacing ∙ Sub-drilling ∙ Stemming and length of stemming ∙ Number of holes and lows in one line ∙ The energy of the explosives ∙ Formation and type of detonator and ignition ∙ Amount of charge per delay ∙ Number of Decks ∙ Distance to structure

Influence factors of blasting vibrations

*Image source: D.W Kang, Applied blasting engineering, book, 1997

Variables

Category ∙Amount of charge per delay ∙Delays ∙Character of explosives

∙Minimum burden and spacing Adjustable factor ∙Stemming ∙Drilling direction ∙Direction of detonating ∙Amount of charges in a round ∙Distance to target structure Nonadjustable factor

∙Terrain condition ∙Thickness and shape of topsoil ∙Condition of rock ∙Condition of air

Impact Serious

Normal

Weak

Remarks

190

W.-H. Heo et al. Table 2. The allowable vibration limits by country

USA [1] Type of construction

Peak Particle Velocity (cm/sec)

Ancient and historic monuments

0.75

Housing in poor repair

1.2

Good residential, commercial, and industrial structures

2.5

Welded gas mains, ground sewers, engineered structures

5.0

Safe Level (cm/sec)

Frequency < 40 Hz

Frequency > 40 Hz

Modern homes-drywall

0.75

2.0

Older homes-plaster on wood

0.50

2.0

Type of structures

Russia(Soviet union) [2] Velocity (cm/sec) Type of structure

Long blasting period

Short blasting period

(repeatable)

(one-shot)

Hospital

0.8

3.0

Kindergarten and residential building

1.5

3.0

Factory, public station, small residential building

3.0

3.0

6.0

12.0

Steel frame concrete structure, mine shaft(over 10 years)

12.0

24.0

Mine shaft(within 3 years)

24.0

48.0

Office, industrial factory, waterway(tunnel), high reinforc ed concrete pipe, elevated construction

(continued)

Development of a Blasting Vibration Monitoring System

191

Table 2. (continued) Portugal [3] Vibration limit (cm/sec) Ground condition

Mixture of non-cohesive

Low cohesive soil, high

soil and gravel

cohesive sand

P2000m/sec

Ruins, hospital, skyscraper

0.25

0.5

1.0

Very usual structures

0.5

1.0

2.0

Reinforced concrete, seismic design

1.5

3.0

6.0

Type of structure

Cohesive soil, rock

Germany [4] Vibration Velocity, Vi, in (mm/sec) Plane of floor of uppermost

Foundation

Type of structure

full story Frequency

< 10Hz

10~50Hz

50~100Hz

Frequency mixture

Business, industrial building and similar

20

20 ~ 40

40 ~ 50

40

Residential and similar

5

5 ~ 15

15 ~ 20

15

3

3~8

8 ~ 10

8

Vibration-prone structure (ruins and major structure) Swiss [5]

Blasting vibration Category

Type of structure

Frequency (Hz)

Steel Structure and Reinforced Concrete Structure 1

Particle velocity (mm/sec)

10~60

30

60~90

30~40

10~60

18

60~90

18~25

Factory building, retaining wall structure, Bridge, an iron tower, Open channel An underground tunnel and underground cavity treated with concrete lining, or w ithout lining. Building with foundation wall and constructed from concrete slab or stone walls.

2

masonry retaining wall Loose stratum tubularis Stone lining treated in underground tunnels and underground cavities

3

A building with a wooden ceiling along with a stone wall

4

Structures of historical value and other vibration-sensitive structures

10~60

12

60~90

12~25

10~60

8

60~90

8~12

(continued)

192

W.-H. Heo et al. Table 2. (continued) 500.0

Great damage occurs to the building.

100.0

Great damage occurs to the building.

5.0

Light damage occurs to the building.

2.0

Extremely light damage occurs to the building. (Man feels like a building is about to collapse)

1.0

The human body feels strongly, but there is no damage to the building.

0.5

In general, many people feel vibrations.

0.1

A very sensitive person feels a vibration.

0.05

Can't feel it in the human body.

Particle velocity (mm/sec)

Human response

Velocity (cm/sec)

Can feel

0.2 ~0.5

Strongly feel

0.5 ~0.95

Feels uncomfortable

0.95 ~2.0

Get pain

2.0 ~3.25

Can not endure

3.25 ~5.0

South Korea [6], [7] Classification Category

sensitive structure

Limit of PPV In detail

ruins, high technology

(cm/sec)

0.3

(continued)

Development of a Blasting Vibration Monitoring System

193

Table 2. (continued) masonry structure(bricks, stone) a structure with wooden walls and ceilings

masonry structure of subbase and concrete slab

Small to medium-sized buildings with reinforced concrete frames and slabs

an old house, low floor house

1.0

low floor slab house, townhouse

2.0

medium and low–rise apartment small to medium-sized shopping mall

Large building with reinforced concrete or steel

seismic design structure

frame and slab

high-rise apartment, huge structure

Type of structure

3.0

factory

5.0

Frequency band over 30 Hz

under 30 Hz

Ruins or very old structure

0.2

0.2

Damaged building, cracked structure

0.5

0.4

Cracked but not damaged

1.0

0.8

1.0 ~ 4.0

0.8 ~ 2.0

Industrial structure with non cement wall

2 Development of New Type Blasting Vibration Monitor This case studied the unmanned automatic measurement of MEMS-based precision tri-axial accelerometers to replace the existing analog geophone. MEMS-based acceleration sensors perform superbly in measuring vibration of high frequency and sinusoidal waves, such as motor vibration, the vibration of belt conveyors, the friction of rotating bearings, etc., but blasting vibration is completely different from those; it is subject to one-time measurement, changing location each time of measurement, coupling with various ground conditions. Thus, it is essential to calibrate the direction of gravitational acceleration for each measurement. To overcome these limitations, the new type monitor induces more accurate three-axis vibration components to be measured by the offset correction of the differences in the fine angle of the measuring sensor during each vibration measurement. Figure 1 shows the program for zero off-settings embedded in the new type monitors, while in Fig. 2, data collected in the case of manual zero settings is compared with those automatically set by the embedded program [9, 10]. In addition, all data is automatically transferred to the server through a mesh network in the 2.4G band in line with IoT, as illustrated in Fig. 3, while simplifying the procedures for installing, measuring, withdrawing, and backing up data by personnel manually.

194

W.-H. Heo et al.

(a) Incorrect zero match

Fig. 1. Re-arm with zero offset command in embedded board

(b) Corrected zero match

Fig. 2. Zero offset corrected by offset command

Fig. 3. Description of the wireless mesh networking for vibration monitor

For the comparative test of this study, the BM series of Canada “I” company, which is known as a worldwide standard model, was selected as the target of comparison, and the three-axis acceleration sensor of Hong Kong “G” company were tested simultaneously and the cross-correlation was analyzed.

3 In-Situ Test for the Comparative Study For comparative experiments at the actual site, a site separated by 20–41 m from the rock blasting location was selected as the site for vibration monitoring; the monitoring location was within a building construction site in the urban district of Busan, Korea. The blasting operation was a typical urban blasting, with the maximum amount of charge per delay being limited to 0.1–0.4 kg/delay. Figures 4 and 5 show the location of the study site and its aerial view, while Figs. 6 and 7 are the scenes for blast hole cleaning and charging. And Figs. 8 and 9 include all the sensors compared in this study.

Development of a Blasting Vibration Monitoring System

195

Fig. 4. Geological location of the field test

Fig. 5. Aerial view of the in-situ test site

Fig. 6. Cleaning up the blast hole

Fig. 7. Charging

Fig. 8. Whole view of the monitor layout

Fig. 9. Monitors for comparative test

The blasting vibration monitors used in this comparison test are pictured in Fig. 9. as follows. 1) Analog type –Instantel(Canada) BlastMate series 2 2) Analog type –Instantel(Canada) BlastMate series 3 3) Analog type –Oyo(Japan) 3-axis geophone

196

4) 5) 6) 7)

W.-H. Heo et al.

Digital(MEMS) type –New type monitor(Korea) _ z axis auto offset Digital(MEMS) type –New type monitor(Korea) - z axis auto offset Digital(MEMS) type –New type monitor(Korea) - z axis auto offset Digital(MEMS) type –Wireless GSS(HongKong) 3GV

Each measuring sensor is fully bonded to the ground with epoxy adhesive for the same coupling with the ground and secured firmly. A total of over 100 vibration-monitoring experiments were performed at the site, but only from the 31 cases, data were found to be complete for comparisons among the five monitors. For all sensors, the trigger vibration level and sampling rate were set to 0.03 Kine (cm/sec) and 1024 per second, respectively.

4 Test Results All of the vibration seismic data were saved in the serial text files and plotted. Matlab (R2015) data analysis tools were applied for signal alignment, analysis of correlation and cross-correlation, Fourier transformation and power spectral density estimation. Since the distance from the blasting site was very close, all seismic signals showed mainly vertical components. Therefore, in this study, only the vertical components that are the basis for PVS(Peak Vector Sum) were compared. 4.1 Peak Particle Velocity(Vertical) Comparisons Figures 10 and 11 show that all the measurements from the analog-type sensors are very similar, and those from the MEMS-based sensors also show similar characteristics. As illustrated in Fig. 12, the determination coefficient, R2 , is 0.97 between the peak particle velocity measurements by the analog and MEMS type units, and this indicated the new type monitoring system could be applied comparably and reliably with the internationally-renowned analog-type blasting vibration monitors.

Fig. 10. Similarity between analog monitors

Development of a Blasting Vibration Monitoring System

197

Fig. 11. Similarity between MEMS(digital) monitors

Fig. 12. Peak Particle Velocity(vertical) comparison between Analog type and MEMS type units

4.2 Correlation Analysis of the Full Waveform The initial vibration wave measurements with different trigger time obtained by different sensors were realigned. Then the correlation analysis was carried out to check the similarity of the full waveforms. The correlation analysis was performed by Matlab. Figure 13 illustrates the vibration waveforms collected from the analog-type sensors with short post-trigger time, while Fig. 14 from MEMS-based sensors with relatively long post-trigger time. Figure 15 shows the seismic forms aligned with respect to the trigger time. Also, cross-correlation analysis was done for the aligned data, and the results plotted in Fig. 16 shows a very high similarity between the waveforms from the different sensors. The correlation analysis results plotted in Fig. 17 shows the correlation between the well-known analog-type sensor and the new type of MEMS-based sensors ranges from 0.80 to 0.87; the average is 0.84. 4.3 Domain Frequency Comparisons In addition, the analysis of domain frequency was conducted using Matlab’s FFT(Fast Fourier Transformation) function to analyze the spectrum and find the domain area of each vibration signal wave, and the results were as follows. In Figs. 18 and 19, the Fourier-transformed data are plotted for the analog-type and MEMS-based sensors, respectively. As shown in Fig. 20, each domain frequency is very similar. If the noises can be filtered effectively, their similarity is expected to be close to the correlation coefficient of 1.0.

198

W.-H. Heo et al.

Fig. 13. Examples of seismic graph from analog-type sensors with short post-trigger time

Fig. 14. Examples of seismic graph from MEMS-based sensors with long post-trigger time

4.4 Results of the Wireless Data Transferring In this study, to overcome some of the inefficiencies of the blasting vibration monitors, which relies on manpower, wireless data transferring devices are assembled to the new-type MEMS-based sensor. This sensor is designed to be used as a portable device like internationally-renowned analog units or as a wireless device and includes several wireless communication modules that can be selected freely with user’s demands. The wireless mesh network can be built with multiple sensors, and LoRA, WiFi, LTE, or even Bluetooth can be selected to cover long and also short-range data communication. In the new devices tested in this study, 2.4 G mesh networks and gateways were used, and LTE was adopted.

Development of a Blasting Vibration Monitoring System

199

Fig. 15. Examples of the graphs aligned for the first trigger(arrival) time

Fig. 16. Three sets of the aligned vibration measurements analyzed by the cross-correlation analysis

Fig. 17. Correlation analysis results between analog-type and MEMS-type sensors

200

W.-H. Heo et al.

Fig. 18. Examples of FFT results for the data from the analog-type sensors

Fig. 19. Examples of FFT results for the data from the MEMS-based sensors

As shown in Fig. 21, all of the full seismic waveforms can be reliably transferred to the FTP server within several seconds to 3 min after event recording. This indicates that most of the laborious operations with conventional devices can be replaced with the new wireless MEMS-based units.

Development of a Blasting Vibration Monitoring System

201

Fig. 20. Domain frequency comparison results

Fig. 21. List of real-time event reports transferred to the FTP server

5 Conclusions In this study, a wireless blast vibration monitoring system based on MEMS was developed to solve the limitations with the conventional analog-type monitoring system, which is expensive analog equipment and also dependent on manpower. To evaluate the accuracy and efficiency of the developed sensors, the new-type devices were tested along with internationally-renowned conventional blasting vibration monitors at a construction site. And the comparisons can be summarized as follows: 1. The PPV(Peak Particle Velocity) value obtained by the MEMS-based tri-axial accelerometer showed very high similarity to the analog instrument. When only the vertical PPV values, the major component of PVS(Peak Vector Sum) in this

202

W.-H. Heo et al.

study, were taken into consideration, the comparison results show the coefficient of determination value is 0.971. 2. For a more accurate comparative analysis, the full waveforms recorded by the analogtype sensors were compared with those obtained by MEMS. The results from two analog and three MEMS devices were cross-compared, and show a very high average correlation coefficient of 0.84 for all waveforms. 3. The domain frequency of all waveforms was derived from the frequency analysis, one of the main evaluation items of the seismic analysis, and also the cross-correlation analysis was performed. The results showed that the domain frequencies in all axial directions were consistent. 4. To solve the limitations with the conventional analog-type monitoring system, one of the most important functions considered in developing the new blasting vibration monitoring system was the wireless data transmission. The new device enables communication with the server through an integrated gateway using the 2.4 G mesh network module and the LTE module, confirming that all data can be transmitted within a maximum of three minutes after the blasting event recorded on each device.

Acknowledgements. This research was partly supported by grants of “Development and Onsite Demonstration of Smart ICT/IoT-Based Mining Smart Ventilation System” (grant No. 20182510102380) funded by the Ministry of Trade Industrial and Energy of the Korean government.

References 1. Siskind, D.E., Stagg, M.S., Kopp, J.W., Dowding, C.H.: Structure Response and Damage Produced by Ground Vibration From Surface Mine Blasting, United States Bureau of Mines Report of Investigations 8507 (USBM RI 8507), pp. 72 (1980) 2. Russian Regulation SanPin 2.1.2.2645-10 3. Esteve, J.M.: Control of Vibrations Caused by Blasting, Porutugal 4. German Standard from DIN 4150 5. Studer, J., Susstrunk, A.: Swiss standard for vibrational damage to buildings. In: Proceedings, X. International Conference on ISSMFE, Stockholm, vol. 3, pp. 307–312 (1981) 6. Allowance of blasting vibration according to structural damage criteria, Korea society of tunnel (1999) 7. Allowance of blasting vibration with frequency band, Seoul Metro (2005) 8. Looney, M.: MEMS vibration monitoring: from acceleration to velocity. Analog Dialogue, 51-06 (2017) 9. Datasheet of ADXL 355, Analog Devices Co. Ltd. 10. Datasheet of LPT5901-IPM, Analog Devices Co. Ltd.

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm to Predict the Dust Concentration in Deo Nai Open-Pit Coal Mine (Vietnam) Xuan-Nam Bui1,2(B) , Hoang Nguyen1,2(B) , Carsten Drebenstedt3 , Hai-Van Thi Tran4 , Ngoc-Bich Nguyen5 , Xuan-Cuong Cao6 , and Qui-Thao Le1,2 1 Department of Surface Mining, Mining Faculty, Hanoi University of Mining and Geology,

Duc Thang, Bac Tu Liem, Hanoi 10000, Vietnam {buixuannam,nguyenhoang}@humg.edu.vn 2 Center for Mining, Electro-Mechanical Research, Hanoi University of Mining and Geology, Duc Thang, Bac Tu Liem, Hanoi 10000, Vietnam 3 Institute of Mining and Civil Engineering, TU Bergakademie Freiberg, Gustav-Zeuner Straße 1A, 09599 Freiberg, Germany 4 Faculty of Information Technology, Hanoi University of Mining and Geology, Duc Thang, Bac Tu Liem, Hanoi 10000, Vietnam 5 Department of Occupational Health and Safety, Faculty of Environmental and Occupational Health, Hanoi University of Public Health, 1a Duc Thang, Bac Tu Liem, Hanoi 11910, Vietnam 6 Department of Mine Surveying, Faculty of Geomatics and Land Administration, Hanoi University of Mining and Geology, Duc Thang, Bac Tu Liem, Hanoi 10000, Vietnam

Abstract. In mining engineering, drilling operation causes a significant impact on the environment, especially dust. In this study, PM10 is investigated and predicted, aiming to assess the risk and its effects on human health comprehensively. Previous studies showed that PM10 from open mines, particularly drilling activities, creates many potential risks to human health. Therefore, accurate prediction of PM10 concentration induced by drilling operation is necessary for protecting human health and the neighboring environment. This study proposeda novel model based on a robust combination of neural network and genetic algorithm (GA) to accurately predict PM10 concentration induced by drilling operation in mine sites. Accordingly, the Levenberg-Marquardt backpropagation neural network (LMBPNN) was developed. Then a robust metaheuristic/optimization algorithm (i.e., GA) was embedded in the LMBPNN model to improve its accuracy; 245 observations of drilling in an open-pit coal mine were measured using the Karomax dust sensor. The results predicted by the proposed GA-LMBPNN model were then compared with the developed LMBPNN model (without optimization) to reach a conclusion. Various performance metrics (e.g., RMSE, R2 , MAE, MAPE, and VAF) and Taylor diagrams were used to evaluate the performance of the developed models. The results indicated that the proposed GA-LMBPNN model provided higher accuracy and reliability than the LMBPNN model with an RMSE of 0.026, R2 of 0.986, MAE of 0.018, MAPE of 0.044, VAF of 98.607, for the training dataset, and an RMSE of 0.028, R2 of 0.979, MAE of 0.023, MAPE of 0.055, VAF of © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 203–223, 2021. https://doi.org/10.1007/978-3-030-60839-2_11

204

X.-N. Bui et al. 97.822, for the testing dataset. Whereas, the LMBPNN model (without optimization) only provided a performance with an RMSE of 0.038, R2 of 0.971, MAE of 0.024, MAPE of 0.056, VAF of 97.123, for the training dataset, and an RMSE of 0.052, R2 of 0.928, MAE of 0.030, MAPE of 0.061, VAF of 92.146, for the testing dataset. In other words, the accuracy of the proposed GA-LMBPNN model is up to 97.8%, whereas the accuracy of the LMBPNN model is only 92.1%. Besides, the results also indicated that moisture content, penetration rate, rebound hardness, compressive strength, and density of rock mass are the most influence parameters that should be used to forecast PM10 in drilling operation in open-pit mines. With the results of this study, PM10 concentrations can be properly predicted and controlled. Keywords: Air pollution · Dust concentration · Green mining · Predict PM10 · Artificial intelligence applications

1 Introduction Mining operations have a positive impact on the economy, energy security, civil engineering, and construction. Max Planck - a famous physicist, said that “Mining is not everything, but without mining everything is nothing.” [1]. However, a large amount of air pollutions is caused by mining activities, such as drilling, blasting, crushing, transporting, and loading/unloading [2, 3]. Of the air pollutants, particles with a diameter of aerodynamic smaller than 10 µm(PM10 ) and total suspended particulate are the major concerns [4–7]. According to the findings of previous researchers, PM10 in open-pit mines is recommended as the cause of black lung disease, silicosis, and increased death rate [8–12]. In addition, cardiopulmonary diseases, respiratory diseases, and lung cancer were also found to be associated with particle matters (PM) [13–15].Previous studies also indicated that PM size has a significant influence on human health, especially PM10 [16]. PM10 consists of particles  10 µm, which can deposit in the respiratory tract (i.e., bronchi and lungs). Therefore, it can be hazardous to health, potentially causing lung cancer and other respiratory diseases. To understand the dangers posed by air pollutants in open-pit mines, especially PM10 , the Colombian environmental authority studied and assessed the impact of mining activities on air quality. Accordingly, PM10 has been defined as one of the major air pollutions [17–19]. High concentrations of toxic elements were found in PM10 , such as Ti, Pb, Cr, Mn, and As [20, 21]. Therefore, spare measurement stations were considered as the primary solution to control the air quality around the mine. However, due to the complexity of the structure and the hazardous components of PM10 , predictive systems are necessary to ensure human health and safety, particularly for the workers at the mines. In recent years, the combined use of sensors and machine learning significant improved most of the real-life problems [22–30]. For instance, Lal and Tripathy [31] predicted dust concentrations in the Haerwusu Surface Coal Mine (China) using an artificial neural network (ANN). Laser and beta-ray particle monitors were used to measure PM10 and PM2.5 at the mine. They found that the dust concentration was very high around the

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm

205

working mining equipment, especially near drilling operations. Finally, an ANN model was developed to predict dust concentrations with acceptable reliability. Focusing on the dust dispersion from drilling operations in open-pit mines, Sastry, Chandar, Nagesha, Muralidhar and Mohiuddin [32] used the Minitab (MIN) model in a MATLAB environment for predicting and analyzing PM2.5 and PM10 . Their findings showed that the MIN model could be used to predict dust concentrations from drilling operations with high accuracy (i.e., 88%). Furthermore, they confirmed that the dust generated within 80 to 100 m is dangerous for workers in the vicinity of drilling operations. In another study, Patra, Gautam, Majumdar and Kumar [33] also applied the ANN technique to predict PM concentrations with different sizes in an open-pit copper mine in India and found a strong agreement for their predicted PM levels. The multiple regression (MR) method was applied to predict the dust dispersion from drilling activities in an open-cast mine by Nagesha, Sastry and Chanda [34]. The results were then compared with an empirical model, and they showed that the MR model could predict dust concentrations better than the empirical model. By the use of an assembly of support vector regression (SVR) and a nature-based optimization algorithm (i.e., particle swarm optimization), Bui, Lee, Nguyen, Bui, Long, Le, Nguyen, Nguyen and Moayedi [35] also predicted PM10 concentrations from the activities of drilling in an open-pit mine with high accuracy. They recommended that artificial intelligence (AI) techniques should be applied and further studied for forecasting dust concentrations in open-pit mines, especially PM10 . AI techniques were also recommended as the potential methods for solving many real-life problems, as well as in the mining industry [36–45]. A review of the literature shows that air pollutants emitted from mining operations cause damage to the surrounding ecology and environment [46]. Flora and fauna around the mining areas can be affected by air pollutants [47, 48]. Therefore, forecasting air pollutants is significant for human health and for the surrounding environment. In this paper, PM10 induced by drilling activities in open-pit mines is predicted. Although previous studies developed and proposed several AI models for predicting PM10 induced by drilling operations, they are scarce. Most of the published papers used simple models for predicting PM10 , such as ANN and SVR. Meanwhile, new models with higher accuracy are always necessary to predict and assess the impact of PM10 on human health and the surrounding environment. Thus, this study proposed a novel approach for predicting PM10 induced by drilling operations based on an improved ANN (i.e., Levenberg-Marquardt backpropagation neural network - LMBPNN) and a metaheuristic algorithm (i.e.,genetic algorithm - GA), called GA-LMBPNN. Accordingly, the GA will solve the drawback of the LMBPNN model (i.e., local minima) and optimize the weights and biases of the LMBPNN model aiming to improve the accuracy of the LMBPNN model. The details of the model’s development, as well as its performance and accuracy, are presented and assessed in the next sections.

2 Methodology 2.1 Levenberg–Marquardt Backpropagation Neural Network ANN is known as the most popularAI technique that is widely used in prediction problems. ANN models with feed-forward propagation were proposed by Werbos [49]) in

206

X.-N. Bui et al.

1975 as a beyond regression tool. Subsequently, Rumelhart, Hinton and Williams [50]) developed a new algorithm based on backpropagation (BP) for training ANN models in the mid-1980s. This algorithm can use parallel distributed processing [51, 52]. Until now, the multi-layer perceptron neural network with the BP algorithm isoften used as the benchmark model for classification, regression, and time-series problems [53–61]. The primary advantages of the BP neural network (BPNN) are flexibility for non-linear issues and not needing a specific form of a particular model [62]. However, it trains very slowly, and a large amount of training data is required for the BP algorithm[63]. In addition, local minima are a major disadvantage of the BP neural network [64]. Therefore, a new algorithm was designed to overcome the defects of the BPNN, namely the Levenberg–Marquardt backpropagation (LMBP) algorithm [65, 66]. In other words, LMBP neural network (LMBPNN) is an enhanced ANN model with higher speed in training, and it overcomes the local minima. The weights of ANN can be optimized by adaptive adjustment between the steepest gradient descent method and the Gauss-Newton algorithm [67]. Accordingly, LMBP is used to train ANN models aiming to minimize the sum-of-squares error. More details of LMBPNN can be referred in the following papers [65, 66, 68, 69]. Herein, LMBPNN is used as the critical model for predicting PM10 concentrations induced by drilling operations. The structure of the LMBPNN model for predicting PM10 concentration in this paper is introduced in Fig. 1.

Fig. 1. Structure of the LMBPNN model for predicting PM10 concentration

2.2 Genetic Algorithm The GA is one of the evolution-based stochastic algorithms. It was developed based on the theory of Darwin [70, 71]. The key operators of the GA include selection, crossover, and mutation [72]. In GA, each gene is represented by a parameter, and each chromosome

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm

207

is denoted for a solution. Like other evolution algorithms, GA uses an objective/fitness function (i.e., root-mean-squared error (RMSE), mean-squared error (MSE)) to evaluate the fitness of each individual in the population. The best solutions can be selected randomly for improving poor solutions based on the proportionality of probability and fitness [73]. The probability of poor solution selection can also avoid local optima [74]. The principle of GA is to provide the best solutions in each generation and improve other solutions by the provided best solutions [75]. Thus, the whole population can become a better generation. Subsequently, crossover is adopted to create a new generation based on two parents. The new generation can be better than the previous generation based on the selection of the best previous solutions [76]. Mutation can also happen in GA when creating new individuals. This stage changes the genes of chromosomes randomly to generate a new population with a diversity of individuals [77]. Then, better solutions can be selected for new individuals towards the global optimum [78]. Figure 2 describes the optimization process of GA.

Fig. 2. GA framework for optimization process

2.3 Proposing the Novel GA-LMBPNN Model for Predicting PM10 Concentrations As stated above, the purpose of this study is to propose a novel AI technique to predict PM10 concentrations induced by drilling activities in mine sites. Indeed, the LMBPNN is considered as an enhanced ANN model for predicting PM10 concentrations herein. To improve the accuracy and performance of the LMBPNN model, the GA is applied to optimize the parameters of the LMBPNN model. Accordingly, the framework of the

208

X.-N. Bui et al.

novel GA-LMBPNN model for forecasting the concentration of PM10 is described in Fig. 3.

Fig. 3. Proposing a novel GA-LMBPNN model for estimating PM10 concentrations

Before developing the GA-LMBPNN model, the dataset of PM10 is divided into two parts, including training and testing datasets. For the set of the training data, 221 drilling operations were randomly selected for training and developing the model. Subsequently, the 24 remaining drilling operations were used for testing the accuracy/performance and accuracy of the developed GA-LMBPNN model. During model development, an initial LMBPNN model is established as the first phase. The weights of the established initial LMBPNN model are the main parameters that need to be optimized to improve the accuracy of the model, as introduced above. Biases are calculated based on the weights of the LMBPNN model. However, it is challenging to know the optimal values for the weights of the LMBPNN models. Therefore, the next phase is to apply the GA to solve this problem, i.e., GA-LMBPNN. The GA operator implements a global search for the values of the weights and biases. For each obtained value of the weight and bias, the accuracy of the GA-LMBPNN model is evaluated through MSE (Eq. 8). The process is repeated until the lowest MSE is determined. Then, the optimal GA-LMBPNN model is defined.

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm

209

2.4 Model Assessment Methods As a regression problem, the performance of the PM10 concentration predictive models can be assessed through MAE (mean absolute error), MAPE (mean absolute percentage error), RMSE, VAF (variance accounted for), and R2 (correlation coefficient). Due to the LMBPNN model using only the MSE metric to evaluate its performance [79–81], the formula for the MSE calculation is also presented in this section. However, it should be noted that it is only used to evaluate the fitness of the MLBPNN model during the optimization process. Five remaining performance metrics (i.e., MAE, MAPE, RMSE, VAF, and R2 ) were computed to assess the predictive models comprehensively. Equations 8– 13 show the formulas for model assessment in this study. Furthermore, as mentioned above, the LMBPNN model (without optimization) is also taken into consideration to compare with the proposed GA-LMBPNN model; therefore, the Taylor diagrams are also constructed to get an overall view of the performance of the models. 2 1  PM10, measured − PM10, predicted MSE = n n

(1)

i=1

 1   PM10, measured − PM10, predicted  n n

MAE =

(2)

i=1

 n  100%   PM10, measured − PM10, predicted    n PM10, measured i=1   n  1  2 RMSE =  PM10, measured − PM10, predicted n

MAPE =

(3)

(4)

i=1



var(PM10, measured − PM10, predicted VAF = 1 − × 100 var(PM10, measured )  (PM10, measured − PM10, predicted )2 i R2 = 1 −  (PM10, measured − PM10, measured )2

(5)

(6)

i

where, PM10, measured , PM10, predicted , and PM10,measured denote the values of PM10 for measured, predicted, and the mean of the measured values.

3 Study Site and Data Acquisition This study was undertaken in an the open-pit coal mine of Vietnam, namely Coc Sau.It is the largest open-pit coal mines in Vietnam, which is located at 107o 19 20 E - 107o 20 50 E and 21o 0 55 N - 21o 2 20 N, as shown in Fig. 4. The annual drilling volume of the mine is approximately 15,000,000 m/year, using drills with a diameter of 200–250 mm, such asD245S, CBIII-250, and DML. By the end of 2019, the elevation of the pit bottomwas 290 m below sea level (−290 m) [82].

210

X.-N. Bui et al.

Fig. 4. Location of the study site and the distance to a residential area

For this study, 245 drilling operations on the mine were collected. PM10 dust dispersion data were recorded by a Kanomax instrument (model 3442) in real-time sensor, as shown in Fig. 5. Accordingly, dust concentration data during the 6 h of drilling was recorded for each period of 15 s. According to some previous studies, meteorological factors have influenced the dispersion of dust in open mines [83–85], such as air humidity, wind speed, and wind direction. However, this study aims to predict the PM10 concentrations induced by drilling activities in mine sites. Therefore, dust measuring devices were placed around the drilling machine (with a distance of 5–10 m) to eliminate dust concentrations from other sources (e.g., blasting, loading, transporting), as well as the influence of wind direction (Fig. 3). Finally, the maximum dust concentration value during the 6 h of drilling was determined as an output value for the intent of this study.

Fig. 5. Real-time PM10 concentration in drilling operations (a) Dust sensor; (b) Results from the dust sensor

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm

211

According to previous studies, the dust caused by drilling activities in open-pit mines depends on the working parameters of drilling equipment and properties of rock [35, 86]. Accordingly, the drilling diameter (d), penetration rate (P), density of rock mass (Do), rebound hardness (R), silt content (S), compressive strength (c), and moisture content (m), are taken into account as having the greatest effect on the PM10 concentration. Therefore, these parameters were considered and selected as the model inputs for predicting PM10 concentrations in this study. An input data collection procedure was established to ensure accuracy. Experimental methods, combined with in-laboratory analyses, were used to collect the values of the inputs as accurately as possible. Finally, a database with seven inputs and one output (i.e., PM10 concentration) was obtained for this study. The details of the attributes are illustrated in Fig. 6.

Fig. 6. Visualization of the PM10 database and the distribution of parameters (a) 3D visualization of m, d, P; (b) 3D visualization of c, S, Do; (c) 3D visualization of c, R, PM 10

4 Results and Discussion In this section, we discuss the model development steps implemented after building a reliable database, as proposed in Fig. 3. An LMBPNN model was developed first as the

212

X.-N. Bui et al.

material for the optimization process. To develop the LMBPNN model, the dataset was normalized by the min-max scale method with the data in the range of [−1, 1], aiming to reduce the skewness of the model, as well as avoid the over-fitting phenomenon. Herein, we used an ANN model with the number of hidden layers, hidden neurons and weights of the neurons are the parameters of the ANN model. A “trial and error” procedure was used to determine the optimal number of hidden layers and neurons, as shown in Fig. 7. The LMBP algorithm was used for training the ANN model. Eventually, an optimal LMBPNN model for PM10 forecasting was defined with the structure of 7-2815-1 based on Fig. 7, and its structure is illustrated in Fig. 8. Note that the weights of the LMBPNN model are illustrated through the lines between the neurons in Fig. 8.

Fig. 7. Trial and error procedure to determine the optimal structure of the MLBPNN model

Once the structure of the MLBPNN model is defined, weights are the parameters of the MLBPNN model, and they were adjusted by the genetic algorithm (GA), as the primary goal of this study (proposing the GA-LMBPNN model). Please be aware that the GA parameters were selected based on recommendations of the previous studies Before optimizing the weights and biases of the ANN model, the GA’s parameters were established as the first step. It is considered as the step of the start for the optimization engine. According to the flowchart of the GA in Fig. 2, the initial population size is needed first for global searching. The optimization performance of the GA models is different for different the number of populations [91]. Thus, different sizes of the population were applied for the GA operator, i.e., 100, 150, 200, 250, 300, 350, 400, 450, and 500. In addition to the population size, the number of variables (n), crossover (cc), and mutation (mc) coefficients are also critical parameters for the GA operator. Herein, n was selected to be 8; cc equaled 1.15, and mc equaled 0.2. Subsequently, the optimization process was performed based on the developed LMBPNN model and the GA. MSE was used to evaluate the accuracy of the GA-LMBPNN model during optimization. Ultimately, the best GA-LMBPNN model (i.e., lowest MSE) was determined, as shown in Fig. 9. Once the LMBPNN (without optimization) and the GA-LMBPNN models were developed, the performance indices of these models were computed according to Eqs. 1– 6. Primarily, the testing dataset with 24 drilling operations was used as a new set of data

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm

213

Fig. 8. LMBPNN model for predicting PM10 concentration

for double-checking the quality and accuracy of the developed models. Table 1 lists the performance indices of the LMBPNN and GA-LMBPNN models based on both training and testing processes. In Table 1, the LMBPNN and GA-LMBPNN models have high reliability in predicting PM10 concentrations. Preliminary results show that the input variables used are significant for predicting PM10 concentration. The data collection method described in this study eliminated meteorological factors’ influence to calculate PM10 for drilling activities. In other words, by placing data collection stations around the drilling area, meteorological factors (i.e., wind speed, wind direction) do not seem to affect PM10 concentration significantly.

214

X.-N. Bui et al.

Fig. 9. Training the LMBPNN model by the GA for predicting PM10 concentrations Table 1. Performance indices of the LMBPNN model with and without optimization by GA in predicting PM10 concentration Model

Training

Testing

RMSE R2 LMBPNN

MAE MAPE VAF

RMSE R2

MAE MAPE VAF

0.038

0.971 0.024 0.056

97.123 0.052

0.928 0.03

0.061

92.146

GA-LMBPNN 0.026

0.986 0.018 0.044

98.607 0.028

0.979 0.023 0.055

97.822

Comparing the two models’ performance LMBPNN and GA-LMBPNN, the GALMBPNN model provided much better performance than the LMBPNN model (without optimization). Although the LMBPNN is considered an enhanced ANN model, GA’s weights and biases were more optimal than the weights and biases calculated by the LMBP algorithm. It improved the LMBPNN model’s accuracy significantly (from 97.123% up to 98.607% on the training dataset, and 92.146% up to 97.822% on the testing dataset). The results in Table 1 show that the proposed GA-LMBPNN model was more stable than the LMBPNN model. Indeed, the GA-LMBPNN model’s performance between the training and testing datasets is smaller than those in the LMBPNN model’s performance. Figures 10 and 11 show the correlation between the predicted and measured PM10 values of the models and their 80% confidence level. In addition to the performance indices, Taylor diagrams were also constructed to get an overview of the developed method, based on the correlation coefficient, standard deviation, and centered RMS difference, as shown in Fig. 12. Accordingly, the green square denotes the center of the optimal value. If a model is closer to the center, its error is smaller than the remaining model. Also, if a model is closer to 0, it has a lower standard deviation than the residual model. In other words, the model with a lower standard deviation is better and more stable. From the Taylor diagrams (Fig. 12), the proposed GA-LMBPNN model is much better than the LMBPNN model. To clarify the

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm

215

Fig. 10. Correlation of predicted and measured PM10 concentration by the LMBPNN model (a) Training phase; (b) Testing phase

Fig. 11. Correlation of predicted and measured PM10 concentration by the proposed GALMBPNN model (a) Training phase; (b) Testing phase

accuracy of the proposed GA-LMBPNN model, the measured and predicted PM10 values by the LMBPNN and GA-LMBPNN models, respectively, are illustrated in Fig. 13. The results show that the predicted PM10 values by the GA-LMBPNN model were closer to the actual PM10 values than those of the LMBPNN model. Considering the estimated results of the proposed GA-LMBPNN model, it is clear that the estimated effects are positive with high accuracy. The use of the seven input variables to determine PM10 concentration in this study is useful. However, to quantitatively evaluate the proposed GA-LMBPNN model, the input variables’ importance was analyzed through a sensitivity analysis. However, to quantitative evaluate the proposed GA-LMBPNN model, the importance of the input variables was analyzed through a sensitivity analysis. The Olden method [92] was applied for this task, and the results are shown in Fig. 14. Accordingly, the moisture content (m) has the highest importance, and

216

X.-N. Bui et al.

Fig. 12. Taylor diagram for model assessment of the GA-LMBPNN and LMBPNN models (a) Training dataset; (b) Testing dataset

Fig. 13. Clustered column chart of the PM10 levels measured and predicted by the GA-LMBPNN and LMBPNN models

penetration rate (P), rebound hardness (R), and compressive strength (c) have a significant effect as well. The remaining variables have not been appreciated in the proposed GA-LMBPNN model for estimating PM10 concentration.

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm

217

Fig. 14. Importance of the input variables for estimating PM10 concentration by the proposed GA-LMBPNN model

5 Conclusion PM10 concentration is one of the dangerous factors that seriously affect human health and workers’ working sites. It has been identified as a significant cause of respiratory diseases, black lung, lung cancer, and related-cardiovascular, eye diseases, to name a few. Therefore, accurate prediction of PM10 concentration in open-pit mines is necessary to challenge environmental science enrollment to protect human health. Based on the investigation and prediction results of this study, we withdraw the following conclusions and remarks: – AI is a robust technique capable of accurately predicting PM10 concentration induced by drilling operations in open-pit mines. It is considered as a significant advance and trend in environmental science and technology. – LMBPNN is an enhanced ANN model which is capable of predicting PM10 concentration with a high confidence level (RMSE = 0.052; R2 = 0.928; MAE = 0.03; MAPE = 0.061; VAF = 92.146). The study’s investigation results show that the LMBPNN is an appropriate technique to predict PM10 concentration in open-pit mines based on the dataset collected. – The proposed GA-LMBPNN model is a robust soft computing model for predicting PM10 concentration induced by drilling operations in open-pit mines. It promoted the role of GA in optimizing the LMBPNN model for predicting PM10 concentration. Although LMBPNN is an improved model from a simple ANN model (i.e., MLP (multi-layer perceptron), BP (backpropagation) -MLP); however, GA continued to optimize the LMBPNN aiming to improve the accuracy of the LMBPNN in predicting PM10 concentration. It should be used in practical engineering to control, forecast, and provide timely solutions to protect human health from PM10 dispersion.

218

X.-N. Bui et al.

– Drilling is only the first phase in mining operations. However, the dust concentration from drilling operations is significant. It is considered one of four operations (i.e., drilling, blasting, loading, and hauling) with the most significant impact on the environment, especially dust (i.e., PM10 ). Because the characteristics of operations are different, the path of dust dispersion, and the ability/method of collecting data for each activity in open-pit mines, are different. Therefore, dust prediction studies for other operations (e.g., blasting, loading, hauling) are necessary to assess the overall dust pollution in open mines. – Although the results of this study are positive with high accuracy and reliability. However, further investigation of the model in other sites (other open-pit mines) in future studies will be attractive to potential engineers and researchers.

Acknowledgments. This work was financially supported by the Ministry of Education and Training (MOET) in Viet Nam under grant number B2018-MDA-03SP. The authors also thank the Center for Mining, Electro-Mechanical research of Hanoi University of Mining and Geology (HUMG), Hanoi, Vietnam, and the research team of Innovations for Sustainable and Responsible Mining (ISRM) of HUMG.

References 1. Dubi´nski, J.: Sustainable development of mining mineral resources. J. Sustain. Mining 12, 1–6 (2013) 2. Chaulya, S.: Air quality status of an open pit mining area in India. Environ. Monit. Assess. 105, 369–389 (2005) 3. Monjezi, M., Shahriar, K., Dehghani, H., Samimi Namin, F.: Environmental impact assessment of open pit mining in Iran. Environ. Geol. 58(1), 205–216 (2008). https://doi.org/10. 1007/s00254-008-1509-4 4. Ghose, M.K., Majee, S.: Assessment of dust generation due to opencast coal mining–an Indian case study. Environ. Monit. Assess. 61, 257–265 (2000) 5. Trivedi, R., Chakraborty, M., Tewary, B.: Dust dispersion modeling using fugitive dust model at an opencast coal project of Western Coalfields Limited, India (2009) 6. Chaulya, S., Ahmad, M., Singh, R., Bandopadhyay, L.K., Bondyopadhay, C., Mondal, G.: Validation of two air quality models for Indian mining conditions. Environ. Monit. Assess. 82, 23–43 (2003) 7. Chaulya, S.: Assessment and management of air quality for an opencast coal mining area. J. Environ. Manage. 70, 1–14 (2004) 8. Sikarwar, S., Guha, S., Singh, K.: Environmental pollution control through geo–botanical means–a case study of Manikpur block, Korba coalfield, MP. India. Environ. Geol. 38, 229– 232 (1999) 9. Huertas, J.I., Camacho, D.A., Huertas, M.E.: Standardized emissions inventory methodology for open-pit mining areas. Environ. Sci. Pollut. Res. 19, 2784–2794 (2012) 10. Wallingford, K.: NIOSH (National Institute for Occupational Safety and Health) indoor air quality in office buildings National Inst. for Occupational Safety and Health Cincinnati, OH (USA) (1987) 11. Li, J., Garshick, E., Al-Hemoud, A., Huang, S., Koutrakis, P.: Impacts of meteorology and vegetation on surface dust concentrations in Middle Eastern countries. Sci. Total Environ. 712, 136597 (2020)

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm

219

12. Kabatas, B., Pierce, R., Unal, A., Rogal, M., Lenzen, A.: April 2008 Saharan dust event: its contribution to PM10 concentrations over the Anatolian Peninsula and relation with synoptic conditions. Sci. Total Environ. 633, 317–328 (2018) 13. Hendryx, M., Ahern, M.M.: Relations between health indicators and residential proximity to coal mining in West Virginia. Am. J. Publ. Health 98, 669–671 (2008) 14. Weichenthal, S.A., Pollitt, K.G., Villeneuve, P.J.: PM 2.5, oxidant defence and cardiorespiratory health: a review. Environ. Health 12, 40 (2013) 15. Kurth, L., Kolker, A., Engle, M., Geboy, N., Hendryx, M., Orem, W., McCawley, M., Crosby, L., Tatu, C., Varonka, M., DeVera, C.: Atmospheric particulate matter in proximity to mountaintop coal mines: sources and potential environmental and human health impacts. Environ. Geochem. Health 37(3), 529–544 (2014) 16. van Zelm, R., Huijbregts, M.A., den Hollander, H.A., Van Jaarsveld, H.A., Sauter, F.J., Struijs, J., van Wijnen, H.J., van de Meent, D.: European characterization factors for human health damage of PM10 and ozone in life cycle impact assessment. Atmos. Environ. 42, 441–453 (2008) 17. Triantafyllou, A.G.: Levels and trend of suspended particles around large lignite power stations. Environ. Monitor. Assess. 89, 15–34 (2003) 18. Guan, Q., Luo, H., Pan, N., Zhao, R., Yang, L., Yang, Y., Tian, J.: Contribution of dust in Northern China to PM10 concentrations over the Hexi corridor. Sci. Total Environ. 660, 947–958 (2019) 19. Zarra, T., Galang, M.G., Ballesteros Jr., F., Belgiorno, V., Naddeo, V.: Environmental odour management by artificial neural network–a review. Environ. Int. 133, 105189 (2019) 20. Srimuruganandam, B., Nagendra, S.S.: Source characterization of PM10 and PM2. 5 mass using a chemical mass balance model at urban roadside. Sci. Total Environ. 433, 8–19 (2012) 21. Espitia-Perez, L., Sosa, M.Q., Salcedo-Arteaga, S., Leon-Mejia, G., Hoyos-Giraldo, L.S., Brango, H., Kvitko, K., da Silva, J., Henriques, J.A.: Polymorphisms in metabolism and repair genes affects DNA damage caused by open-cast coal mining exposure. Mutation Res./Genet. Toxicol. Environ. Mutagenesis 808, 38–51 (2016) 22. Sarigiannis, D.A., Karakitsios, S.P., Gotti, A., Papaloukas, C.L., Kassomenos, P.A., Pilidis, G.A.: Bayesian algorithm implementation in a real time exposure assessment model on benzene with calculation of associated cancer risks. Sensors 9, 731–755 (2009) 23. Nguyen, H., Choi, Y., Bui, X.-N., Nguyen-Thoi, T.: Predicting blast-induced ground vibration in open-pit mines using vibration sensors and support vector regression-based optimization algorithms. Sensors 20, 132 (2019) 24. Nguyen, H., Bui, X.-N.: Soft computing models for predicting blast-induced air over-pressure: A novel artificial intelligence approach. Appl. Soft Comput. 92, 106292 (2020) 25. Zhang, S., Bui, X.-N., Trung, N.-T., Nguyen, H., Bui, H.-B.: Prediction of rock size distribution in mine bench blasting using a novel ant colony optimization-based boosted regression tree technique. Nat. Resour. Res. 29(2), 867–886 (2019) 26. Baek, J., Choi, Y.: Deep neural network for predicting ore production by truck-haulage systems in open-pit mines. Appl. Sci. 10, 1657 (2020) 27. Chenar, S.S., Deng, Z.: Development of artificial intelligence approach to forecasting oyster norovirus outbreaks along Gulf of Mexico coast. Environ. Int. 111, 212–223 (2018) 28. Jeong, J., Choi, J.: Development of AOP relevant to microplastics based on toxicity mechanisms of chemical additives using ToxCast™ and deep learning models combined approach. Environ. Int. 137, 105557 (2020) 29. Al–Dabbous, A.N., Kumar, P., Khan, A.R.: Prediction of airborne nanoparticles at roadside location using a feed–forward artificial neural network. Atmos. Pollut. Res. 8,446–454 (2017) 30. Franceschi, F., Cobo, M., Figueredo, M.: Discovering relationships and forecasting PM10 and PM2. 5 concentrations in Bogotá, Colombia, using artificial neural networks, principal component analysis, and k-means clustering. Atmos. Pollut. Res. 9,912–922 (2018)

220

X.-N. Bui et al.

31. Lal, B., Tripathy, S.S.: Prediction of dust concentration in open cast coal mine using artificial neural network. Atmos. Pollut. Res. 3, 211–218 (2012) 32. Sastry, V., Chandar, K.R., Nagesha, K., Muralidhar, E., Mohiuddin, M.S.: Prediction and analysis of dust dispersion from drilling operation in opencast coal mines. Proc. Earth Planet. Sci. 11, 303–311 (2015) 33. Patra, A.K., Gautam, S., Majumdar, S., Kumar, P.: Prediction of particulate matter concentration profile in an opencast copper mine in India using an artificial neural network model. Air Qual. Atmos. Health 9(6), 697–711 (2015) 34. Nagesha, K., Sastry, V., Chanda, R.K.: Prediction of dust dispersion during drilling operation in open cast coal mines: A multi regression model. Int. J. Environ. Sci. 6, 681–696 (2016) 35. Bui, X.-N., Lee, C.W., Nguyen, H., Bui, H.-B., Long, N.Q., Le, Q.-T., Nguyen, V.-D., Nguyen, N.-B., Moayedi, H.: Estimating PM10 concentration from drilling operations in open-pit mines using an assembly of SVR and PSO. Appl. Sci. 9, 2806 (2019) 36. Armaghani, D.J., Hajihassani, M., Mohamad, E.T., Marto, A., Noorani, S.A.: Blastinginduced flyrock and ground vibration prediction through an expert artificial neural network based on particle swarm optimization. Arab. J. Geosci. 7(12), 5383–5396 (2013) 37. Jahed Armaghani, D., Hasanipanah, M., Tonnizam Mohamad, E.: A combination of the ICAANN model to predict air-overpressure resulting from blasting. Eng. Comput. 32(1), 155–171 (2015) 38. Jahed Armaghani, D., Hasanipanah, M., Mahdiyar, A., Abd Majid, M.Z., Bakhshandeh Amnieh, H., Tahir, M.M.D.: Airblast prediction through a hybrid genetic algorithm-ANN model. Neural Comput. Appl. 29(9), 619–629 (2016) 39. Nhu, H.V., Duong, B.V., Vo, T.A., Pham, K.T.: Using numerical modeling method for design and constructive controlling of excavation wall in Madison building, Ho Chi Minh city. J. Mining Earth Sci. 61, 19–27 (2020) 40. Tran, V.T.T.: Dynamic symbol for electronic map, network map and the ability to create dynamic symbol on the map with the online answering machine. J. Mining Earth Sci. 61, 88–98 (2020) 41. Nguyen, X.V.: Design calculations and simulation dynamics of banana screen (in Vietnames). J. Mining Earth Sci. 61, 50–57 (2020) 42. Dang, C.M., Thai, V.Q., Pham, M.N., Dang, T.T., Phung, M.T.T., Ngo, T.D.: The solution of data transmission security for gateway IoT (in Vietnames). J. Mining Earth Sci. 61, 58–67 (2020) 43. Nguyen, L.Q.: A novel approach of determining the parameters of Asadi profiling function for predicting ground subsidence due to inclied coal seam mining at Quang Ninh coal basin (in Vietnames). J. Mining Earth Sci. 61, 86–95 (2020) 44. Nguyen, N.T.T., Tong, H.T.: Predicting land use change base on GIS and remote sensing (in Vietnames). J. Mining Earth Sci. 61, 106–115 (2020) 45. Pham, L.T., Nguyen, S.P., Nguyen, N.V., Dao, H.V., Doan, L.D., Vo, N.H.T., Nguyen, T.T.T., Tran, H.V., Le, N.T.: Establishment of land cover map using object-oriented classification method for VNREDSat-1 data (in Vietnames). J. Mining Earth Sci. 61, 134–144 (2020) 46. Tripathi, N., Singh, R.S., Hills, C.D.: Reclamation of Mine-impacted Land for Ecosystem Recovery. Wiley Online Library, Hoboken (2016) 47. Singh, R.S., Singh, R.K., Bera, S.: Impact of mining on herbaceous ground cover and wildfauna in Birsha-Damoh forest range of Malanjkhand copper mines of India. Eurasian J. Forest Sci. 5, 1–7 (2017) 48. Tripathi, N., Singh, R.S., Hills, C.D.: Soil carbon development in rejuvenated Indian coal mine spoil. Ecol. Eng. 90, 482–490 (2016) 49. Werbos, P.: Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Sciences. vol. Ph. D. dissertation. Harvard University (1974)

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm

221

50. Rumelhart, D.E., Hinton, G.E., Williams, R.J.: Learning representations by back-propagating errors. Nature 323, 533 (1986) 51. Leonard, J., Kramer, M.: Improvement of the backpropagation algorithm for training neural networks. Comput. Chem. Eng. 14, 337–341 (1990) 52. Antwi, P., Zhang, D., Xiao, L., Kabutey, F.T., Quashie, F.K., Luo, W., Meng, J., Li, J.: Modeling the performance of single-stage nitrogen removal using anammox and partial nitritation (SNAP) process with backpropagation neural network and response surface methodology. Sci. Total Environ. 690, 108–120 (2019) 53. Nguyen, H., Bui, X.-N.: Predicting blast-induced air overpressure: a robust artificial intelligence system based on artificial neural networks and random forest. Nat. Resour. Res. 28(3), 893–907 (2018) 54. Nguyen, H., Bui, X.-N., Bui, H.-B., Mai, N.-L.: A comparative study of artificial neural networks in predicting blast-induced air-blast overpressure at Deo Nai open-pit coal mine, Vietnam. Neural Comput. Appl. 32, 1–17 (2018) 55. Nguyen, H., Drebenstedt, C., Bui, X.-N., Bui, D.T.: Prediction of blast-induced ground vibration in an open-pit mine by a novel hybrid model based on clustering and artificial neural network. Nat. Resour. Res. 29(2), 691–709 (2019) 56. Shang, Y., Nguyen, H., Bui, X.-N., Tran, Q.-H., Moayedi, H.: A novel artificial intelligence approach to predict blast-induced ground vibration in open-pit mines based on the firefly algorithm and artificial neural network. Nat. Resources Res. 29(2), 723–737 (2019) 57. Nguyen, H., Bui, X.-N., Tran, Q.-H., Le, T.-Q., Do, N.-H., Hoa, L.T.T.: Evaluating and predicting blast-induced ground vibration in open-cast mine using ANN: a case study in Vietnam. SN Appl. Sci. 1(1), 1–11 (2018) 58. Mason, C., Twomey, J., Wright, D., Whitman, L.: Predicting engineering student attrition risk using a probabilistic neural network and comparing results with a backpropagation neural network and logistic regression. Res. High. Educ. 59, 382–400 (2018) 59. Moayedi, H., Jahed Armaghani, D.: Optimizing an ANN model with ICA for estimating bearing capacity of driven pile in cohesionless soil. Eng. Comput. 34(2), 347–356 (2017) 60. Dou, J., Yamagishi, H., Zhu, Z., Yunus, A.P., Chen, C.W.: TXT-tool 1.081-6.1 a comparative study of the binary logistic regression (BLR) and artificial neural network (ANN) models for GIS-based spatial predicting landslides at a regional scale. In: Landslide Dynamics: ISDR-ICL Landslide Interactive Teaching Tools, pp. 139–151. Springer (2018) 61. Mishra, N., Soni, H.K., Sharma, S., Upadhyay, A.: Development and analysis of artificial neural network models for rainfall prediction by using time-series data. Int. J. Intell. Syst. Appl. 11, 16 (2018) 62. Zhao, X., Han, M., Ding, L., Calin, A.C.: Forecasting carbon dioxide emissions based on a hybrid of mixed data sampling regression model and back propagation neural network in the USA. Environ. Sci. Pollut. Res. 25(3), 2899–2910 (2017) 63. Pradhan, P., Tingsanchali, T., Shrestha, S.: Evaluation of soil and water assessment tool and artificial neural network models for hydrologic simulation in different climatic regions of Asia. Sci. Total Environ. 701, 134308 (2020) 64. Maru, A., Dutta, A., Kumar, K.V., Mohapatra, D.P.: Effective Software Fault Localization Using a Back Propagation Neural Network, pp. 513–526. Springer Singapore (2020) 65. Levenberg, K.: A method for the solution of certain non-linear problems in least squares. Q. Appl. Math. 2, 164–168 (1944) 66. Marquardt, D.W.: An algorithm for least-squares estimation of nonlinear parameters. J. Soc. Ind. Appl. Math. 11, 431–441 (1963) 67. Reddy, S.V.B.S., Kumar, B., Swaroop, D.: Investigations on training algorithms for neural networks based flux estimator used in speed estimation of induction motor. In: 2019 6th International Conference on Signal Processing and Integrated Networks (SPIN), pp. 1090– 1094. IEEE (2019)

222

X.-N. Bui et al.

68. Ye, Z., Kim, M.K.: Predicting electricity consumption in a building using an optimized backpropagation and Levenberg–Marquardt back-propagation neural network: case study of a shopping mall in China. Sustain. Cities Soc. 42, 176–183 (2018) 69. Kannaiyan, M., Raghuvaran, J.G.T.: Prediction of specific wear rate for LM25/ZrO2 composites using Levenberg–Marquardt backpropagation algorithm. J. Mater. Res. Technol. 9, 530–538 (2020) 70. Mitchell, M.: An introduction to genetic algorithms. MIT press, Cambridge (1998) 71. Carr, J.: An introduction to genetic algorithms. Senior Project 1, 40 (2014) 72. Jean-Luc, R., Gonon, F., Favre, L., Niederhäuser, E.-L.: Convergence of multi-criteria optimization of a building energetic resources by genetic algorithm. In: 2018 International Conference on Smart Grid and Clean Energy Technologies (ICSGCE), pp. 150–155. IEEE (2018) 73. Mirjalili, S.: Genetic algorithm. In: Evolutionary Algorithms and Neural Networks, pp. 43–55 (2019) 74. Seyedpour, S., Kirmizakis, P., Brennan, P., Doherty, R., Ricken, T.: Optimal remediation design and simulation of groundwater flow coupled to contaminant transport using genetic algorithm and radial point collocation method (RPCM). Sci. Total Environ. 669, 389–399 (2019) 75. Sin, I.H., Do Chung, B.: Bi-objective optimization approach for energy aware scheduling considering electricity cost and preventive maintenance using genetic algorithm. J. Clean. Prod. 244, 118869 (2020) 76. Fang, Q., Nguyen, H., Bui, X.-N., Tran, Q.-H.: Estimation of blast-induced air overpressure in quarry mines using cubist-based genetic algorithm. Nat. Resour. Res. 29(2), 593–607 (2019) 77. Le, L.T., Nguyen, H., Dou, J., Zhou, J.: A comparative Study of PSO-ANN, GA-ANN, ICAANN, and ABC-ANN in estimating the heating load of buildings’ energy efficiency for smart city planning. Appl. Sci. 9, 2630 (2019) 78. Liu, L., Moayedi, H., Rashid, A.S.A., Rahman, S.S.A., Nguyen, H.: Optimizing an ANN model with genetic algorithm (GA) predicting load-settlement behaviours of eco-friendly raft-pile foundation (ERP) system. Eng. Comput. 36(1), 421–433 (2019) 79. Nawi, N.M., Rehman, M., Aziz, M.A., Herawan, T., Abawajy, J.H.: An accelerated particle swarm optimization based Levenberg Marquardt back propagation algorithm. In: International Conference on Neural Information Processing, pp. 245–253. Springer (2014) 80. Sapna, S., Tamilarasi, A., Kumar, M.P.: Backpropagation learning algorithm based on Levenberg Marquardt Algorithm. Comput. Sci. Inform. Technol. (CS and IT) 2, 393–398 (2012) 81. Kaur, H., Kaur, M.: Fault classification in a transmission line using Levenberg–Marquardt algorithm based artificial neural network. In: Data Communication and Networks, pp. 119– 135. Springer (2020) 82. Company, C.S.: Summary report of production in 2019, Coc Sau (in Vietnamese) (2019) 83. Arpacıo˘glu, C., Er, C.: Estimation of fugitive dust impacts of open-pit mines on local air quality-a case study: Bellavista gold mine, Costa Rica. In: 18th International Mining Congress and Exhibition of Turkey-IMCET, pp. 220–235 (2003) 84. Huertas, J.I., Huertas, M.E., Izquierdo, S., González, E.D.: Air quality impact assessment of multiple open pit coal mines in northern Colombia. J. Environ. Manag. 93, 121–129 (2012) 85. Silvester, S., Lowndes, I., Hargreaves, D.: A computational study of particulate emissions from an open pit quarry under neutral atmospheric conditions. Atmos. Environ. 43, 6415–6424 (2009) 86. Nagesha, K., Chandar, K.R., Sastry, V.: Prediction of dust dispersion by drilling operation using artificial neural networks. Int.J. Prev. Control Ind. Pollut. 1, 1–13 (2016) 87. El-Mihoub, T.A., Hopgood, A.A., Nolle, L., Battersby, A.: Hybrid genetic algorithms: a review. Eng. Lett. 13, 124–137 (2006)

Utilizing a Novel Artificial Neural Network-Based Meta-heuristic Algorithm

223

88. Juang, C.-F.: A hybrid of genetic algorithm and particle swarm optimization for recurrent network design. IEEE Trans. Syst. Man Cybern. Part B (Cybern.) 34,997–1006 (2004) 89. Robinson, J., Sinton, S., Rahmat-Samii, Y.: Particle swarm, genetic algorithm, and their hybrids: optimization of a profiled corrugated horn antenna. In: IEEE Antennas and Propagation Society International Symposium (IEEE Cat. No. 02CH37313), pp. 314–317. IEEE (2002) 90. Tonnizam Mohamad, E., Jahed Armaghani, D., Hasanipanah, M., Murlidhar, B.R., Alel, M.N.A.: Estimation of air-overpressure produced by blasting operation through a neurogenetic technique. Environ. Earth Sci. 75(2), 1–15 (2015) 91. Gen, M., Lin, L.: Genetic Algorithms. In: Wiley Encyclopedia of Computer Science and Engineering, pp. 1–15 (2007) 92. Olden, J.D., Joy, M.K., Death, R.G.: An accurate comparison of methods for quantifying variable importance in artificial neural networks using simulated data. Ecol. Model. 178, 389–397 (2004)

Evaluating the Air Flow and Gas Dispersion Behavior in a Deep Open-Pit Mine Based on Monitoring and CFD Analysis: A Case Study at the Coc Sau Open-Pit Coal Mine (Vietnam) Van-Duc Nguyen1(B) , Chang-Woo Lee1(B) , Xuan-Nam Bui2,3 , Hoang Nguyen2,3 , Quang-Hieu Tran2,3 , Nguyen Quoc Long2,3 , Qui-Thao Le2,3 , Xuan-Cuong Cao2,3 , Ngoc-Tuoc Do4 , Won-Ho Heo5 , and Ngoc-Bich Nguyen6 1 Department of Energy and Mineral Resources, College of Engineering, Dong-A University,

Busan 49315, South Korea [email protected], [email protected] 2 Department of Surface Mining, Mining Faculty, Hanoi University of Mining and Geology, Duc Thang, Bac Tu Liem, Hanoi 100000, Vietnam 3 Mining, Electro-Mechanical Research Center, Hanoi University of Mining and Geology, Duc Thang, Bac Tu Liem, Hanoi 100000, Vietnam 4 Department of Surface Mining, Institute of Mining Science and Technology, 03 Phan Dinh Giot Street, Phuong Liet, Thanh Xuan, Hanoi 100000, Vietnam 5 Mining Tech Co., Ltd., Saha-gu, Busan 49315, South Korea 6 Faculty of Environmental and Occupational Health, Hanoi University of Public Health, 01A Duc Thang Road, Duc Thang Ward, North Tu Liem District, Hanoi 100000, Vietnam

Abstract. Air quality in the mining industry is a severe environmental issue associated with many health problems. Managing air quality in mining areas has faced many challenges due to the lack of understanding of the climatic factors and physical removal mechanisms of gas contaminants. In order to study the effects of the atmospheric conditions on the pit pollution, the air velocity distribution and gas dispersion behavior were evaluated in the deepest open-pit coal mine in Vietnam based on the monitoring data and numerical modeling. The field study was conducted in Coc Sau open-pit coal mine located in the northeastern of Vietnam. Two fixed monitoring stations were installed at the ground level to measure the wind speed, wind direction, and temperature to evaluate the atmospheric class stability based on the Froude number and Pasquill stability class. These monitoring data were also used for 3D CFD analysis of the polluted gas dispersion behavior. Furthermore, the vertical temperature profile within the pit was measured to determine the existence of the temperature inversion layer. In general, the Froude Number is an estimate of whether the flow can cross high mountains or not and is basically the ratio of the wind perpendicular to the mountain chain to the stability of the atmosphere. The Froude Number was found to be 0.1–0.7 during the 4-day test. With Fr < 1.3, the airflow in the pit is totally decoupled from the airflow above the pit. The existence of the temperature inversion layer was observed. CFD analysis of the air velocity distribution and CO gas dispersion behavior indicates that high © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 224–244, 2021. https://doi.org/10.1007/978-3-030-60839-2_12

Evaluating the Air Flow and Gas D‘ispersion Behavior in a Deep Open-Pit Mine

225

dust and gas concentrations within the pit observed during the study were partly attributed to the stability of the atmosphere. Keywords: Air quality · Deep open-pit mine · Atmospheric stability classes · Temperature inversion · CFD analysis

1 Introduction Mining, as a leading industry, play a vital role in industrialization growth. However, it is true that all significant mining activities directly or indirectly contribute to air pollution issues and related health hazards [1]. The developed emphasis on open-pit mining operations in recent years to achieve ever-increasing production targets has further aggravated the problems of air pollution [2]. The contaminated air generation and its dispersion have been the primary concern in ambient air quality in the deep cavities such as openpit mines. The dispersion mechanism consists of diffusion and advection processes. The movements in the atmosphere transport and diffuse the pollutants released from the sources. A variety of studies have been carried out to understand the transport and diffusion mechanisms. The tracer gas method was used to investigate the dispersion characteristics by Richardson [3]. The temporal and spatial scales of motion serve to disperse pollutants in the atmosphere by mixing and thus lowering the ambient pollutant concentrations [4]. Mikkelson [5] and Hanna et al. [6] have discussed various aspects of dispersion and related parameters influenced primarily by source size, buoyancy and momentum of release, roughness and surrounding terrain, atmospheric stability and large-scale differential heating. The potential air pollutants of mining operations are particulate pollutants, including particles with an equivalent aerodynamic diameter of less than 10 µm (PM10 ) and gaseous emissions (CO, CO2 , SO2 , NOx ). These pollutant sources constitute the main environmental concerns. Notably, the mining operations that generate these particles are drilling, blasting, loading and dumping, road transport on unpaved roads, and losses from tailings dump. Particles reduce air quality and visibility and adversely affect flora and fauna as well as human health. They can be transported over long distances by the wind and then settle on land or water and cause environmental damage to other ecosystems [7]. As a result, the numerical of dispersion characteristics within the deep open-pit coal mines becomes essential to analyze the intricate wind flow patterns [8]. The dispersion equations developed within the deep pit boundary provide a reasonably accurate estimate of PM10 dispersion within the near field region of the deep open pit quarry mines [9]. The fundamental equations of continuity and momentum describe the in the pit dispersion mechanisms within the atmospheric boundary layer [8]. Also, the meteorological conditions in deep open-pit coal mines are significantly affected by atmospheric temperature (atmospheric layer stability), and the roughness conditions, which eventually create complex dispersion phenomena, including layer separation of the atmospheric boundary layer, recirculation, reflux and settling down PM10 [10, 11]. However, the field measurements of PM10 within the deep open-pit coal mines are limited by safety regulations, the complex shape of the pit, uncontrolled airflow, and the different types of operations that follow. Thus, polluted air monitoring becomes extremely difficult [8].

226

V.-D. Nguyen et al.

Moreover, the numerical analysis of dispersion characteristics using the fundamental governing equations may require modifications to incorporate the microclimate effects in the pit into the flow regime [12, 13]. Therefore, it is necessary to analyze and evaluate microclimate parameters, especially the airflow turbulence, to understand the dispersion of polluted air within the pit [14]. This study aims a comprehensive description of the dispersion mechanisms in the deep open-pit coal mines considering the topographic, thermal, and meteorological factors based on the monitoring data and CFD analysis. The atmospheric class stability and temperature inversion layer are discussed to understand dispersion mechanisms. Besides, to help better understand the gaseous dispersion mechanism, one 3D full-scale model of the open-pit domain is simulated using the CFD tool. This study results for atmospheric class stability, temperature inversion layer, and CFD analysis are expected to provide vital information for the pollutant dispersion mechanism in deep open-pit coal mines. It can be used as a piece of useful information to give advanced warning of potential emission problems and providing the basis for future planning in deep open-pit mining [8].

2 Study Area and Methodology The Coc Sau open-pit coal mine is one of the largest and deepest open-pit coal mines in Vietnam at the time of the study, with a depth of 250 m below the sea level (MASL). It is located in northeast Vietnam, and it lies within latitudes 106o 25 00 E-108o 05 00 and longitudes 20o 45 30 N-21o 10 00 N as shown in Fig. 1 [15]. The coal production is over 2 to 3 million tonnes/year, and overburden of 30 to 40 million m3 /year is removed, relating mining operations, the drill machines of D245S and DML with the borehole diameters 200–250 mm can be used. The electrical EKG excavator with a bucket capacity of 4.6

Fig. 1. Location of the study site [15]

Evaluating the Air Flow and Gas D‘ispersion Behavior in a Deep Open-Pit Mine

227

to 8 m3 or hydraulic excavator with a bucket capacity of 5.2–12 m3 was the favorite overburden excavator used in the mine. The huge hard-frame truck with a load of 55–96 tonnes can be used to transport the overburden. For coal transportation, the soft-frame trucks with the loading capacity of 36–39 tonnes were employed. In general, the mining operations contributed to an increase in the impact of polluted air on the surrounding environmental and public health. In this regard, it can be seen in Fig. 1 that the distance from the mine to the residential area is about 700 m. Therefore, the effect of generated contaminant air, such as PM10 concentration, is significant. In addition, the occupational hazard for employees is a particular concern.

3 Data Collection and Its Characteristics For data collection, three fixed monitoring stations were installed. Specifically, two fixed monitoring stations were installed at the ground level to measure wind speed, wind direction, and temperature to evaluate the atmospheric class stability based on the Froude Number and Pasquill stability class. In addition, the third fixed monitoring station was installed at the pit bottom to measure PM10 and CO gas concentration to evaluate the dispersion of polluted air within the pit. It can be seen in Figs. 2a and 2b that the first fixed monitoring station was installed at +105 m, and the second fixed monitoring station at +195 m. Regarding monitoring sensors, the Ultrasonic anemometer Young 81000 was used for data collection. This model is a 3-axis, no-moving-parts with the sensor. It is perfectly suited for measuring wind speed that requires a quick response, high resolution, and three-dimensional wind measurement. The Young 81000 instrument can measure three-dimensional wind velocity and speed of sound based on the transit time of ultrasonic acoustic signals. Sonic temperature is derived from the speed of sound, which is corrected for crosswind effects. The details of the specification of the Young model 81000 can be found in Table 1. As mentioned above, to evaluate the contaminated air dispersion (i.e., PM10 and CO) in the pit, the third fixed monitoring station was installed at the pit bottom (−250 m compared with seal level) in Fig. 3c. The data collection process, as well as the meteorological data of this study, is illustrated in Fig. 4, and the collected data during the 4-days are summarized in Table 2. Table 1. Specifications of Ultrasonic Anemometer (Young 81000) Specification Wind speed (m/s) Wind direction (Degree)

Range

Precision

Resolution

0–40 ±0.05 m/s 0.01 m/s 0–360 ±2–5°

0.1°

Speed of sound (m/s)

300–360 ±0.1%

0.01 m/s

Sonic temperature (o C)

−50–50 ±2°C

0.01 °C

228

V.-D. Nguyen et al.

Figure 4 shows the wind charts of two fixed stations at the ground level during the 4-day test. It can be seen that at the first fixed station (+105 m), the wind dominantly blew from the North to the South direction during the 4-day test. The wind speed of 0.16–3.34 m/s was measured at this station, as shown in Figs. 4a, c, e, and g. However, the velocity was mainly in the range of 0.5–1.0 m for the first fixed monitoring station, as shown in Fig. 4. In the second fixed station at the elevation of +195 m, the velocity was higher, ranging between 1.06 and 5.85 m/s. In addition, the wind blew from the North-East to the South-West direction at the second fixed station (+195 m). Thus, the wind blew from the first station to the second station during the test. All the collected measurement data at the first fixed station used for the numerical analysis, it will be discussed in the following sections.

(a) Coc Sau open-pit coal mine and monitoring station installation

(b) Cross-sectional view via the mine (A – A)

Fig. 2. Coc Sau open-pit coal mine and monitoring station installation

Evaluating the Air Flow and Gas D‘ispersion Behavior in a Deep Open-Pit Mine

(a) 1st fixed monitoring station (+105 m)

229

(b) 2nd fixed monitoring station (+195 m)

(c) 3rd fixed monitoring station (+250 m)

Fig. 3. Data collection at the Coc Sau open-pit mine

230

V.-D. Nguyen et al.

(a) Day 1 – at the 1st station (+105 m)

(b) Day 1 – at the 2nd station (+195 m)

(c) Day 2 – at the 1st station (+105 m)

(d) Day 2 – at the 2nd station (+195 m)

(e) Day 3 – at the 1st station (+105 m)

(f) Day 3 – at the 2nd station (+195 m)

(g) Day 4 – at the 1st station (+105 m)

(h) Day 4 – at the 2nd station (+195 m)

Fig. 4. Wind chart at two fixed stations during 4 days

Evaluating the Air Flow and Gas D‘ispersion Behavior in a Deep Open-Pit Mine

231

Table 2. Summary of the experiment data Monitoring station

Experiment time

1st fixed station (+105 m)

2nd fixed station (+195 m)

Temperature (o C)

U velocity (m/s) Max

Mean

Min

Max

Mean

Min

Day 1

1.20

Day 2

1.08

0.30

−3.12

40.29

32.50

29.85

0.30

−3.27

40.65

31.97

29.61

Day 3

1.21

0.16

−3.01

39.39

31.87

27.45

Day 4

1.48

0.35

−3.34

39.31

32.50

29.95

Day 1

1.21

−1.45

−4.50

33.01

30.01

28.83

Day 2

1.68

−1.26

−5.41

34.89

30.36

28.61

Day 3

1.61

−1.06

−4.67

36.17

30.57

26.99

Day 4

2.28

−1.45

−5.85

35.23

30.29

29.05

4 Atmospheric Stability Classes Atmospheric stability is defined as an air parcel that can move upward or downward after it has been displaced vertically by a small amount [16]. If the air parcel tends to fall back to its original level after the lifting effect stops, the atmosphere is known to be stable, while if the air parcel tries to increase vertically when the lifting stops, it is known to be an unstable condition. The neutral condition is when the parcel tends to remain in place after the lifting stops. Stability classes depend on thermal disturbance, static stability, and mechanical disturbance. Static disturbances are related to change in temperature with height variation, and the mechanical turbulence is dependent on the effect of wind speed and surface roughness. Lapse speed can be defined as the rate at which the atmospheric temperature decreases or increases with changing elevation. It helps in determining the stability of the atmosphere. If the rate of environmental deviation is higher than those of the adiabatic lapse rate, the atmosphere is considered to be unstable and stable when the rate of environmental deviation is less than the adiabatic lapse rate. The neutral condition can be defined when the adiabatic and environmental deviation rates are the same. In 1961, Pasquill and Gifford developed a method for classifying the amount of turbulence present in the atmosphere, and it was taken into account as the most commonly used method [17]. Passquill stability classes were to classify the stability of a region of the atmosphere in terms of the horizontal surface wind, the amount of solar radiation, and the fractional cloud cover. Finally, Passquill and Gifford classified atmospheric disturbances into six stable layers, named as in Table 3, and their meteorological conditions are listed in Table 4. Of those, the A-class is taken into account as the most chaotic or unstable class, and F-class is considered as the most stable or least disturbing class. It affects the vertical movement of air. Based on the collected data during the 4-day test in the mine, as shown in Fig. 4 and Table 2, it can be seen that the wind direction did not change at both fixed monitoring stations on the ground level. As aforementioned, the wind speed was in the range of 0.16– 3.34 m/s at the first fixed monitoring station; 1.06–5.85 m/s at the second fixed station. Based on the Passquill stability classes in Table 4, the C class is the atmospheric stability

232

V.-D. Nguyen et al. Table 3. The Pasquill stability class [18] Stability class Definition

Stability class Definition

A

Very unstable

D

B

Unstable

E

C

Slightly unstable F

Neutral Slightly stable Stable

Table 4. Meteorological conditions that define the Passquill stability class [18] Surface wind speed (m/s) Daytime incoming solar radiation

Nighttime cloud cover

Strong Moderate Slight >50% 1.6, then inversion is swept away. If 1.3 < Fr < 1.6, the air in the basin is coupled to the air above the basin, and if Fr < 1.3, the air is totally decoupled. Furthermore, the effect of dispersion in the coal mines by the Passquill stability class are summarized in Table 6. In another study, Mark and Jim [22] investigated the temperature inversion in deep open-cut mines. They found that the stable conditions occurred with Fr < 1. Note: 1) 2) 3) 4)

– Aids dispersion if the terrain is flat and restricts dispersion if the terrain is rising. – Impact by particulate deposition. – Impact by airborne particulates. – Higher values of surface heat, flux aids dispersion due to the generation of upward air currents. 5) – Lower values of surface heat flux restrict aerial dispersion.

234

V.-D. Nguyen et al.

(a) Day 1 – at the 1st station (+105 m)

(b) Day 1 – at the 2nd station (+195 m)

(c) Day 2 – at the 1st station (+105 m)

(d) Day 2 – at the 2nd station (+195 m)

(e) Day 3 – at the 1st station (+105 m)

(f) Day 3 – at the 2nd station (+195 m)

(g) Day 4 – at the 1st station (+105 m)

(h) Day 4 – at the 2nd station (+195 m)

Fig. 6. Temperature measurements at two fixed stations during the 4-day test

Fig. 7. The difference in temperature between the pit bottom and the ground level in Day 1

Based on the collected data during the 4-day test at the Coc Sau open-pit coal mine, it is clear that the atmospheric stability class was C in the daytime, and it is the E or F class in the nighttime. The C class in the daytime means the particulate deposition can be impacted, and the E or F class in the nighttime indicates that the particulate deposition and the airborne particulate can be impacted, and also the lower values of surface heat flux restrict aerial dispersion. These results imply that the generated pollutants such as PM10 and toxic gases were hardly dispersed under these kinds of effects. Grainger

Evaluating the Air Flow and Gas D‘ispersion Behavior in a Deep Open-Pit Mine

(a) Day 1

(b) Day 2

(c) Day 3

(d) Day 4

Fig. 8. The estimated Froude numbers during the 4-day test

235

236

V.-D. Nguyen et al. Table 5. Summary of the temperature comparison

Comparing data

1st fixed station (+105 m)

2nd fixed station (+195 m)

3rd fixed station (−250 m)

Day 1

10.44

4.18

3.61

Day 2

11.04

6.28

4.22

Day 3

11.94

9.18

2.56

Day 4

9.36

6.18

3.11

The difference in Day 1 temperature average Day 2 between the pit bottom Day 3 and the ground level Day 4

3.82

Variation of temperature between night and daytime

5.11 4.58 4.22

Table 6. Effects on dispersion in the coal mines by stability class [8] Stability class Effects on dispersion in the coal mines A

(2), (3), (4)

B

(2), (3), (4)

C

(2), (4)

D

(2)

E

(1), (2), (3), (5)

F

(1), (3), (5)

G

(1), (3), (5)

and Meroney [11] observed that inversion layer effects are more predominant in pits rather than on flat terrains. Furthermore, it is well-known as strong nocturnal inversions that occur in valleys and mountain basins on a calm and clear night [23, 24]. The air temperature inversion affects surface mine environments in the form of accumulation of higher concentrations of pollutants (gases and dust) [8, 22]. However, there was almost no data on the nocturnal meteorology in deep open-pit mines [22]. In this study, the vertical temperature profiles were measured to identify the temperature gradient and to determine airflow stability as well as temperature inversion layer by the unmanned aerial vehicle (UAV) to investigate the temperature gradient within the pit; the temperature sensor was attached to the UAV. One UAV was flown vertically from the pit bottom (−250 m) above the ground level, as shown in Fig. 9. The experiment was carried twice a day before the sunrise and after the sunset. The vertical temperature profiles are shown in Fig. 10.

Evaluating the Air Flow and Gas D‘ispersion Behavior in a Deep Open-Pit Mine

237

Fig. 9. Flightpath of UAV for vertical temperature profile measurement

(a) Test 1

(b) Test 2

Fig. 10. Vertical temperature profile

It can be observed that the temperature inversion effect has occurred during the two tests. The temperature inversion layer can be seen in Fig. 10a at the vertical elevation of 20 m and 150 m. A similar pattern can be observed in Fig. 10b at the vertical elevation of 100 m. Furthermore, Pasquill categorized the atmospheric stability classes based on the vertical temperature gradient, as shown in Table 7. Table 7. Pasquill stability classes by the vertical temperature gradient [16] Stability class

Vertical temp gradient, T/Z (DegC/100 m)

A

−1.9

Very unstable

B

−1.9 to −1.7

Unstable

C

−1.7 to −1.5

Slightly unstable

D

−1.5 to −0.5

Neutral

E

−0.5 to 1.5

Slightly unstable

F

1.5 to 4.0

Stable

G

>4.0

Very stable

Definition

238

V.-D. Nguyen et al.

In this regard, the temperature gradient was observed in 3 days. The results showed that the temperature gradient in the range of 0.64–1.53 °C/100 m. Based on the classification in Table 7, the atmosphere stability can be classified into E and F classes for the site measurements. This result is similar to the classification above based on wind speed and Fr number. Due to the stability classes are “slightly unstable” and “stable”. The dispersion of pollutants generated within the pit is hardly dispersed out of the deep pit. It is expected to result in high CO gas and PM10 dust concentrations in the pit. This is what was observed during the study, as shown in Fig. 11. This figure shows high concentrations of CO gas and PM10 after the sunset. It can be seen that due to the atmospheric stability class of “slightly unstable” and “stable” and the existence of the temperature inversion layer in Fig. 10b contributed to high CO gas and PM10 concentrations. The CO concentration kept on increasing when the time went by, while the PM10 concentration was almost constant. These results indicate that the pollutants generated within the pit were not effectively dispersed out of the deep pit due to poor pit ventilation (Table 8).

(a) CO concentration (ppm)

(b) PM10 concentration (ppm)

Fig. 11. Contaminant measurement data at the pit bottom (−250 m)

Table 8. Vertical temperature gradients for the study site Category

Before sunrise

After sunset

Upward flight

Downward flight

Upward flight

Downward flight

1st day measurement

1.17

1.53

0.64

0.68

2nd day measurement 3rd day measurement

0.68

0.84

1.53

1.28

1.21

1.12

1.13

1.23

Evaluating the Air Flow and Gas D‘ispersion Behavior in a Deep Open-Pit Mine

239

5 CFD Analysis of the Atmospheric Flow and the CO Gas Dispersion Numerous studies are using the CFD tool in the environmental aspect. ANSYS-FLUENT is well-known CFD software in the world. In this study, the CFD analysis was carried out using ANSYS FLUENT, which has commonly been used by many researchers to study various fluid-flow and heat transfer problems in the underground mine environment [25, 26]. Based on the real scale of the experiment site in Fig. 2b, the 3D geometry layout for CFD analysis is illustrated in Fig. 12. The initial condition of CFD simulation can be shown in Table 9. The k-E model was employed as the turbulence model in this study, and all computational iterations were solved implicitly to evaluate the gas dispersion behavior under the effect of the atmospheric stability class. As mentioned above, nine polluted air sources were defined based on the location of excavator and truck at working sites. A CO source of 10 ppm was defined as the polluted air sources based on the reference of the diesel engine exhaust emission survey by Arif Susanto [27], as shown in Fig. 12b. In fact, there are other toxic gases in open-pit mines such as methane (CH4 ), Carbon monoxide (CO), Sulphur dioxide (SO2 ), and nitrogen dioxide (NO2 ). However, the gas dispersion behavior under the atmospheric stability class was similar. Thus, in this study, only CO gas was used as a toxic gaseous pollutant to evaluate the dispersion behavior. In terms of the input data of wind speed for CFD analysis, the wind velocity measurements at the first fixed station (+105 m) in Fig. 4c were used.

(a) 3D geometry for CFD analysis

(b) excavator locations as pollutant emission sources

Fig. 12. Geometry for CFD analysis and polluted air sources

Figure 13 shows the temporal distributions of the air velocity by the CFD analysis. It can be observed that the low velocity of 0–0.1 m/s is predominant at the pit bottom of the mine. In Fig. 13, all the temporal distributions from 2 h to 24 h are similar. This result implies that due to low velocity distributed at the pit bottom, the generated contaminated air was hardly dispersed out of the deep pit and remained for a long time. This phenomenon is clearly shown in Fig. 14, where the high CO concentration gas remains near the pit bottom even after one day. From Fig. 15, it can be seen that the low velocity at the pit bottom is thought to contribute to the high CO concentration. The CO concentration kept on increasing, and this trend is quite similar to the experiment data, as shown in Fig. 11a. These results indicate that the pollutants generated within the pit are hardly dispersed out of the deep pit if the climatic situation near the pit remains the same. Based on this study, the air

240

V.-D. Nguyen et al. Table 9. Initial conditions of the CFD analysis Parameters

CFD model

Inlet boundary

Velocity inlet

Wall boundary

Friction wall

Ventilation resistance (k)

0.014 kg/m3

Wall temperature

37°C

Mesh type

Tetrahedron elements

Solution model

Turbulence model (k-E)

Mesh size function

Proximity and curvature

Number of mesh elements 3.000.000 Simulation condition

Transient-state conditions

Fig. 14. CO gas concentrations by the CFD analysis after 24 h

polluted within the pit cannot be effectively dispersed only by natural ventilation at a deeper level. It is the fact that the mining plan of the Coc Sau open-pit coal mine is more in-depth in the future, the dust and gases may have a significant impact on the workers. Therefore, the authors hope that the obtained results in this study are useful for potential solutions in air quality controlling in deep open-pit coal mines.

Evaluating the Air Flow and Gas D‘ispersion Behavior in a Deep Open-Pit Mine

(a) after 2 hours

(b) after 4 hours

(c) after 8 hours

(d) after 16 hours Fig. 13. Temporal distributions of the air velocity by the CFD analysis

241

242

V.-D. Nguyen et al.

(e) after 24 hours Fig. 13. (continued)

Fig. 15. CO concentration at the pit bottom by CFD analysis after 12 h

6 Summaries and Conclusions This study aims at evaluating the effects of the atmospheric conditions on the air quality within the deep pit. The atmospheric class stability at Coc Sau mine, the deepest open-pit coal mine in Vietnam, was investigated in terms of wind speed/direction, Fr number, and vertical temperature gradient. With the experimental meteorological results, 3D CFD analysis was performed to understand the gaseous pollutant physical removal mechanism. The results can give advanced warning of the potential emission problems and provide the basis for future planning. Several significant results can be summarized as follows: 1. The atmospheric stability class during the 4-day test was C class in the daytime and E or F in the nighttime. This result implies that the atmospheric stability classes

Evaluating the Air Flow and Gas D‘ispersion Behavior in a Deep Open-Pit Mine

2.

3.

4.

5.

243

were neutral in the daytime and slightly stable or stable in the nighttime. These observations show that the generated pollutants such as PM10 or toxic gases were hardly dispersed under the poor atmospheric situation. The Fr number was 0.1–0.7 during the 4-day test in the mine. This result implies that the air was totally decoupled. Consequently, the particulate and gaseous pollutants could not be dispersed out of the deep pit. This conclusion can be confirmed by the high CO gas and PM10 concentration data. The vertical temperature gradient measurements showed the existence of the temperature inversion layer. The atmosphere stability classes were classified as E and F based on the vertical temperature gradient measurements, and this result was similar to the classification based on wind speed and Fr number. The low air velocity of 0–0.1 m/s and high concentration of CO gas were distributed at the pit bottom by the CFD analysis. This result is quite similar to the measurement data. Since the air quality at deep open-pit mines is likely to get worse as pits go more in-depth, this study will provide fundamental knowledge for the understanding of the pollutant dispersion mechanisms and developing efficient control.

Acknowledgments. This work was financially supported by the Ministry of Education and Training (MOET) in Vietnam under grant number B2018-MDA-03SP. The authors also thank the Center for Mining, Electro-Mechanical Research of Hanoi University of Mining and Geology (HUMG), Vietnam; the engineers and leaders of the Coc Sau open-pit mine, Quang Ninh province, Vietnam for their help and cooperation.

References 1. Asif, Z., Chen, Z.: Environmental management in North American mining sector. Environ. Sci. Pollut. Res. 23(1), 167–179 (2016) 2. Sahu, R.A.J.A.T., Panda, P.S.: Ambient air quality assessment in opencast metal mines. (Unpublished bachelor dissertation). National institute of technology Rourkela, Odisha, India (2013) 3. Richardson, L.F.: Atmospheric diffusion shown on a distance-neighbor graph. Proc. Roy. Soc. Lond. Ser. A, Contain. Pap. Math. Phys. Character 110(756), 709–737 (1926) 4. Turner, D.B.: Workbook of Atmospheric Dispersion Estimates: An Introduction to Dispersion Modeling. CRC press (1994) 5. Mikkelsen, T., Nielsen, M.: Modelling of pollutant transport in the atmosphere. MANHAZ position paper, Ris∅ National Laboratory, Denmark (2003) 6. Hanna, S.R., Briggs, G.A., Hosker, R.F.: Handbook on Atmospheric Diffusion Technical Information Center, U.S. Department of Energy DOE/TIC 11223 (1982) 7. Morawska, L., Moore, M.R., Ristovski, Z.D.: Impacts of Ultrafine Particles. Australian Government, Department of the Environment and Heritage Health, 9 (2004) 8. Chinthala, S., Khare, M.: Particle dispersion within a deep open cast coal mine. Air Qual.: Models Appl. 81 (2011) 9. Silvester, S.A., Lowndes, I.S., Hargreaves, D.M.: A computational study of particulate emissions from an open pit quarry under neutral atmospheric conditions. Atmos. Environ. 43(40), 6415–6424 (2009)

244

V.-D. Nguyen et al.

10. Bitkolov, N.Z.: Wind and Temperature of quarry atmospheres. Fiziko-Tekhnicheskie Problemy Razrabotki Poleznykh Iskopaemykh, vol. 5, pp. 66—73 (1969) 11. Grainger, C., Meroney, R.N.: Dispersion in an open-cut coal mine in stably stratified flow. Bound.-Layer Meteorol. 63, 117–140 (1993) 12. Markov, V.A., Potashnik, E.L., Rivkind, V.Y.: Two-dimensional mathematical model of natural ventilation processes in opencast mines. Atmos. Oceanphys. 14, 5 (1978) 13. Aloyan, A.E., Baklonov, A.A., Penenko, V.V.: Application of fictious domain method in numerical simulation problems of open cast mine ventilation. Meteorol. Gidrolog. 7, 42–49 (1982) 14. Turner, D.B.: Workbook of atmospheric dispersion estimates. Environmental Protection Agency, AP-26, Warren Spring Laboratory, National survey of smoke and Sulphur dioxide, Instruction manual (1994) 15. Bui, X.-N., et al.: Estimating PM10 concentration from drilling operations in open-pit mines using an assembly of SVR and PSO. Appl. Sci. 9(14), 2806 (2019) 16. Woodward, J.L.: Estimating the Flammable Mass of a Vapor Cloud, vol. 21. Wiley, Hoboken (2010) 17. Jitesh, K.M.: Modelling the dispersion of dust generated from open pit mining (Unpublished bachelor dissertation). National institute of technology Rourkela, Odisha, India (2015) 18. Wikipedia the free encyclopedia: Outline of air pollution dispersion, https://en.wikipedia.org/ wiki/Outline_of_air_pollution_dispersion 19. Sutherland, B.R. Generation mechanisms. In: Internal Gravity Waves, pp. 284–285, Cambridge University Press, Cambridge (2018) 20. Sheridan, P.F., Vosper, S.B.: A flow regime diagram for forecasting lee waves. Rotors Downslope Winds. Meteorol. Appl. 13, 179–195 (2006) 21. Maruhashi, J., Serrão, P., Belo-Pereira, M.: Analysis of mountain wave effects on a hard landing incident in pico aerodrome using the AROME model and airborne observations. Atmosphere 10(7), 350 (2019) 22. Hibberd, M.F., Hondros, J.: Temperature inversion in deep open-cut mines? Centre for Australian Weather and Climate Research, CSIRO Marine & Atmospheric Research, Aspendale 3195 BHP billiton, 55 Grenfell St, Adelaide 5000 (2009) 23. Clements, C.B., Whiteman, C.D., Horel, J.D.: Cold-air-pool structure and evolution in a mountain basin: Peter Sinks. Utah. J. Appl. Meteorol. 42, 752–768 (2003) 24. Hibberd, M.F.: Nocturnal dispersion meteorology in an urban valley. Clean Air Environ. Qual. 37(4), 34–37 (2003) 25. Nguyen, V., Kim, D., Hur, W., Lee, C.: Experimental and CFD study on the exhaust efficiency of a smoke control fan in blind entry development sites. Tunn. Undergr. Space 28(1), 38–58 (2018) 26. ANSYS, Inc.: FLUENT User’s Guide, Version 18.0, USD, ANSYS Inc., Canonsburg (2018) 27. Susanto, A., et al.: Diesel engine exhaust emissions survey of underground mine in Indonesia. J. Indu. Pollut. Control 32(2), 608–616 (2016)

Numerical Investigation of Characteristics of Mine Ventilation Using One or Two Ducts in Underground Mining Faces Minsik Kim, Jongmyung Park(B) , Youngdo Jo, Dongkil Lee, and Huiuk Yi Korea Institute of Geoscience and Mineral Resources, 124 Gwahak-ro, Yuseong-gu, Daejeon, Korea [email protected]

Abstract. In this study, ventilation in mining faces with one duct (blowing) and two ducts (blowing and exhaust) was numerically analyzed. Underground ventilation is very important, particularly around underground mining faces, because many miners work together in these locations. For mine ventilation in the workplace, one duct (blowing) or two ducts (blowing and exhaust) are usually employed to supply fresh air and remove dirty air. The numerical solutions presented in this paper provide analytical, additional, and helpful information on flow structures for building safe and secure ducts for mine ventilation. The numerical code ANSYS CFX 19.1 was employed to solve related equations, and ICEM-CFD 19.1 was used for the numerical grids. The turbulence model was employed for good agreement with the experimental results. The Reynolds-averaged Navier–Stokes equations were resolved using this model. Second-order spatial discretization schemes were used in this simulation to ensure good solutions. The convergence of the numerical simulations was checked based on scaled residuals for the continuity, momentum, and turbulent equations. The distribution of velocity, flow streamlines, velocity vectors, eddy viscosity, and dust concentration are presented for the analysis of flow characteristics in ventilation ducts. Moreover, all flow quantities are explained in view of comparisons between the use of one duct (blowing) and two ducts (blowing and exhaust). The airflow around ventilation ducts and mining faces contains complicated vortices, which capture the air, thereby increasing its age and level of dust contamination. Keywords: Mine ventilation · CFD · Ventilation duct · Mining face

1 Introduction Ventilation systems in mines are essential because of the recent increases in mining depth. Mining workplaces are where drilling, blasting, and diesel equipment generate a substantial amount of dust; hence, mine ventilation is more important in these areas. For example, methane and coal dust from the mining face of underground coal mines cause health problems (such as pneumoconiosis) in workers and are likely to cause © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 245–262, 2021. https://doi.org/10.1007/978-3-030-60839-2_13

246

M. Kim et al.

explosion accidents. Ventilation ducts are the most common method of controlling dust in underground mines. In this study, the ventilation system in the workplace of mining faces was simulated with one blowing duct that provides fresh air from outside to the mine face, and one exhaust duct that draws in the polluted air from the working area and discharges it to the outside. The purpose of this study is to present the primary data of the design of the duct ventilation system by numerical analysis of coal dust and methane distribution for mine ventilation systems of one duct (blowing only), and two ducts (blowing and exhaust), respectively. In particular, coal dust sizes of PM10 and PM2.5 (which cause pneumoconiosis for mine workers) were analyzed in this study to show the ventilation efficiency of fine dust in the working area. Numerous studies have been carried out for experimental and numerical investigation of the ventilation efficiency with ducts in the mining face. Parra et al. [1] compared the field measurement data and the numerical analysis data to the MAA in an underground coal mine. They found that ventilation systems are efficient for the case of the exhaust duct located very close to the mining face and the blowing duct located relatively far from the mining face. Torano et al. [2] investigated the effect on workers by numerical analysis of airflow and dust concentration with field measurement data at the mining face with road headers using two ducts. Torano et al. [3] measured the concentration of methane in an underground coal mine and compared the results with numerical analysis. The measurement data of the mining field indicates there was a high methane concentration at the mining face. The methane concentration decreased to a minimum value at the ventilation duct and then increased with distance from the mining face. Thus, the authors reported that auxiliary ventilation systems are necessary for high ventilation efficiency. Diego et al. [4] developed a method to calculate air pressure loss by CFD modeling and performed 4D simulation for dispersion of pollutants in tunnels with a tunnel boring machine (TBM). Lihong et al. [5] showed CFD modeling using the experimental data at the mining face with a continuous miner in mine tunnel ventilated with a blowing curtain. Sasmito et al. [6] conducted a study on the design of a ventilation system using a blowing duct, exhaust duct, and brattice in an underground coal mine being mined with a room and pillar mining method. According to the numerical analysis results, the ventilation system using a brattice and exhaust system minimizes the appearance of a circular zone. Lu et al. [7] distributed methane and coal dust using blowing and exhaust ducts, and a continuous miner using room and pillar mining. The authors found that high airflow velocity and use of an exhaust duct generally lower the methane and dust concentration when continuous mining is in operation. Hasheminasab et al. [8] performed a numerical simulation to evaluate the distribution of methane concentration using a brattice and exhaust dust at an underground coal mining face. The results show that when using a brattice as a ventilation system, the use of an exhaust duct with a suction fan (auxiliary ventilation equipment) reduces methane concentration in the mine. Torno et al. [9] investigated the behavior of the carbon dioxide generated after a blast in a tunnel with a ventilation duct, and modeled the ventilation efficiency of the tunnel. The authors presented a new mathematical model (not a general mathematical model) that analyzed the statistical correlation with experimental data. Kurnia et al. [10] performed an analysis of the concentration and distribution of methane-based on the number and

Numerical Investigation of Characteristics of Mine Ventilation

247

location of the methane sources in CFD modeling at the mining face. The simulation results show that methane dispersion is affected by the number and location of methane sources. The authors suggested that the method of reducing methane concentration is the focus of airflow at one point to reduce methane concentration. Park et al. [11] used the mean age of air (MAA) to indicate the freshness of the air by quantifying the air quality near the workplace in a mining tunnel. The results show an increase in MAA due to circular motion formed by the discharging of reflected fresh air from the mining face. Park et al. [12] performed a parameter evaluation to design a blowing duct to reduce MAA. CFD modeling was used to simulate the flow characteristics, vectors, and MAA depending on the location of the blowing duct to improve ventilation efficiency near the mining face. Therefore, the authors suggested that the best position of the blowing duct is the top of the center in the mining tunnel. Yi et al. [13] performed a parameter evaluation for optimal ventilation duct design. The optimal location and operation conditions of the ventilation duct are presented through MAA analysis by the location of the duct and the mass flow rate variable. The optimal location of the exhaust duct was suggested as the edge of the lower section of the mining face; the exhaust duct had no effect when the mass flow rate of the exhaust duct is less than 25% of the blowing duct mass flow rate. In this study, the location of blowing and exhaust ducts were in accordance with Park et al. [12] and Yi et al. [13]. This study provides information on the distribution of coal dust and methane in the mining face using one duct or two ducts in the working area.

2 Mining Face and Workplace Arrangement Figure 1 shows the geometry of mining workplaces with (a) one duct (blowing only) and (b) two ducts (blowing and exhaust). There are two types of workplace ventilation in mines: a single duct for blowing, and two ducts for blowing and exhaust. Park et al. [12] and Yi et al. [13] present the velocity distribution, mean age of air, and turbulent properties for the conditions of one and two ducts. The mining tunnel features are the mining face, blowing duct, exhaust duct, outlet, roof, and bottom. The length and height of the mine tunnel were 36 and 2.9 m, respectively. The locations of the blowing and exhaust ducts are in accordance with Park et al. [12] and Yi et al. [13]. The blowing duct was located 0.14 m from the roof of the tunnel, and the exhaust duct was located 0.4 m from the bottom of the tunnel. The blowing and exhaust ducts had diameters of 0.6 and 0.3 m, respectively. The X-axis was along the length from the mining face to the outlet, the Y-axis along the height of the tunnel, and the-Z axis perpendicular to the X- and Y-axes.

248

M. Kim et al.

(a) One duct

(b) Two ducts

Fig. 1. Schematic diagram of geometry to compute flow distribution (a) one duct (blowing only), and (b) two ducts (blowing and exhaust)

3 Numerical Approach and Procedures Numerical simulations using ANSYS CFX 19.1 were conducted to study the flow characteristics under the blowing duct and exhaust duct conditions in the mining tunnel. The software ICEM CFD 19.1 was used to generate the hexahedral grid for the simulation. 3.1 Independency of Numerical Grids The structured meshes of the computational domain near the outlet are shown in Fig. 2. The hexahedral meshes were employed with the o-grid of ICEM-CFD 19.1 to improve the quality of the mesh near the air duct. A fine mesh was generated near the mining face and wall to resolve the high property gradient. In particular, the area where fresh air from the blowing duct hit the mining face is fine mesh because of its high-velocity gradient. Three different grids (0.5, 3.5, and 4.6 M) were checked for grid independency. The number of cells (3.5 M–4.1 M) was chosen to yield good results over a short time. 3.2 CFD Modeling The purpose of this study was to characterize the flow and distribution of dust and methane using a blowing duct or blowing and exhaust ducts in a mine tunnel. The simulation fluid was air, coal dust, and methane. The coal dust diameters were set as 1 × 10−4 m, PM10 (10 × 10−6 m), and PM2.5 (2.5 × 10−6 m), respectively. In numerical analysis, particle movement can be modeled using the particle transport model of commercial code ANSYS CFX 19.1. The flow analysis of coal dust is solved by the Langrangian particle transport model of CFX. The particle transport model was used to analyze the dispersed phase (particles) in the continuous phase (air). The coupling

Numerical Investigation of Characteristics of Mine Ventilation

(a) one duct

249

(b) two ducts

Fig. 2. Portion of the computed grid to compute the continuity and momentum equations: (a) one duct (blowing only), and (b) two ducts (blowing and exhaust)

between the two phases (air and particles) exists as a one-way, two-way, or full coupling; this study assumed a full coupling in which the dust and airflow affect each other [14]. The coal dust (PM10, PM2.5) was injected uniformly at the mining face. Methane was set to be released from the methane source for 5 min at the mining face. The inlet boundary represents the velocity condition, which is 10 m/s, and the outlet boundary represents the condition of ambient pressure. The entire walls of the mine tunnel were set to be non-slip. The mass flow rate of the exhaust duct was set as 30% of the mass flow rate of the blowing duct, referring to Yi et al. [13]. Methane was released at a total flow rate of 0.05 m3 /s at the mining face, in accordance with Torano et al. [3]. The total flow rate of dust was 0.0062 kg/s, in accordance with Lu et al. [7]. All numerical analyses were solved under steady and unsteady conditions. For methane distribution, the continuity and momentum equations were solved in the steady-state to reduce the computational time. Based on their solutions, the methane distribution equation was solved under the unsteady condition. For particle transport, all equations were solved under the unsteady condition. 3.3 Turbulence Modeling and Solution Convergence The k-ε turbulence model was chosen to solve the flow distribution. The SST model was used to solve the transport equation because stable convergence of the solution was needed. The residual of the continuity and momentum equation was 1 × 10−6 .

4 Numerically Predicted Flow Characteristic and Distribution of Coal Dust and Methane 4.1 Verification of Computation Model To verify the numerical analysis in this study, the distribution of the velocity of x = 4 m was compared with the results of Kurnia et al. [10] from the mining face. Figure 3a shows the distribution of the velocity of x = 4 m using FLUENT software and the k-ε model in Kurnia et al. [10]. The result of the Kurnia et al. [10] simulation is based on

250

M. Kim et al.

the experimental data from Parra et al. [1]. The values of P (2.2, 1.8, 1.3, −3.2, and − 4.4 m/s) are the experimentally measured data by Parra et al. [1]. Figure 3b shows the distribution of the velocity of x = 4 m using CFX software and the k-ε model in this study. The fresh air from the blowing duct has a negative value based on the X coordinate. The red area in Fig. 3a and the dark area in Fig. 3b indicate negative values. Fresh air was reflected from the mining face and discharged to the outlet, and thus it has a positive value. Although Kurnia et al. [10] used FLUENT software for numerical analysis, and CFX software was used in this study, the results in Fig. 3 can be assumed identical.

(a)

(b) Fig. 3. Comparison of streamwise velocities in the (a) plane X = 4.0 from the mining face [Kurnia et al., 2014], and (b) same location as the plane in this study

Numerical Investigation of Characteristics of Mine Ventilation

251

4.2 Distribution Characteristics of Airflow Figure 4 shows a schematic diagram of planes in different locations in a mine tunnel. The planes are set as X = 3, 12, and 20 m to compute the air, coal dust, and methane flow quantities. X = 3 m is a plane near the working area.

Fig. 4. Schematic diagram of planes to compute profiles of airflow quantities

Figure 5a shows the streamline flow of air from the mining face with a blowing duct velocity of 10 m/s. Fresh air from the blowing duct hit the mining face, and fresh air reflected from the mining face was discharged into the outlet. When fresh air was discharged into the outlet, vortical flow developed near the blowing duct. The vortical flow prevented fresh air and dust from discharging into the outlet, causing poor ventilation in the working area. Figure 5b shows the streamline flow of the mine tunnel with the blowing and exhaust ducts. The fresh air from the blowing duct was reflected by the mining face and flowed into the exhaust duct and outlet. Some air flows, generating vortical flow, but it can be seen that Fig. 5b is less vortical flow than Fig. 5a, because some air is discharged into the exhaust duct. Since mining works work in areas where the vortical flow exists, this study investigates the effect of vortical flow on the flow of coal dust and methane in the case of one duct or two ducts. 4.3 Distribution of Methane Using One or Two Ducts Figure 6 shows the distribution of the methane mass fraction for time = 10, 20, 110, and 350 s in the case of one duct (blowing only). The distribution of the methane mass fraction is also symmetrical, and the streamline flow developed into asymmetrical flow. In the area of X = 3 m near the working area, the distribution of the methane mass fraction was the highest at time = 110 s. Methane was distributed at the bottom of the

252

M. Kim et al.

(a) One duct

(b) Two ducts

Fig. 5. Streamlines of airflow around the mining face with (a) one duct (blowing only), and (b) two ducts (blowing and exhaust)

working area (X = 3 m) at time = 10 s, and with time, the methane spread with the airflow and was distributed throughout the mine tunnel. This phenomenon is shown in Fig. 7.

T = 10 s

T = 20 s

T = 110 s

T = 350 s

X=3

X = 12

X = 20

Fig. 6. Distribution of methane mass fraction at time = 10, 20, 110, and 350 s, at the X = 3, 12, and 20 planes for the case of one duct (blowing)

Figure 7 shows the distribution of methane mass fraction at the Z = 0 plane of the on duct. The methane was initially distributed on the bottom, and then throughout the mine tunnel. The methane mass fraction is highest in areas where vortical flow exists. This explains how methane is stagnant by vortical flows. The methane mass fraction in the mine tunnel was the highest in all sections at time = 110 s and reduced at time = 350 s. After time = 350 s, the methane mass fraction near the mining face where working is done is low, with most methane are discharged to the outlet.

Numerical Investigation of Characteristics of Mine Ventilation

253

T = 10 s

T = 20 s

T = 110 s

T = 350 s

Fig. 7. Distribution of methane mass fraction at time = 10, 20, 110, and 350 s at the Z = 0 plane for the case of one duct (blowing)

T = 10 s

T = 20 s

T = 110 s

T = 350 s

X=3

X = 12

X = 20

Fig. 8. Methane distribution at time = 10, 20, 110, and 350 s at the X = 3, 12, and 20 planes for the case of two ducts (blowing and exhaust)

254

M. Kim et al.

Figure 8 shows the distribution of the methane mass fraction for the case of two ducts (blowing and exhaust). As in the case of one duct, methane was distributed on the bottom at time = 10 s. The methane mass fraction of the two ducts was also the highest at time = 110 s. However, unlike the case of one duct, the methane was partially concentrated in the mine tunnel.

Z=0 T = 10 s

T = 20 s

T = 110 s

T = 350 s

Z = 1.55

T = 10 s

T = 20 s

T = 110 s

T = 350 s

Fig. 9. Methane distribution at time = 10, 20, 110, and 350 s for the Z = 0, and 1.55 plane for the case of two ducts (blowing and exhaust)

Numerical Investigation of Characteristics of Mine Ventilation

255

Figure 9 presents the distribution of the methane mass fraction at time = 10, 20, 110, and 350 s for the Z = 0, and 1.55 plane. The methane around the exhaust duct was discharged to the outlet. 4.4 Distribution of Coal Dust Using One or Two Ducts The particle transport model was used to resolve the distribution of coal dust generated during coal mining. In particular, PM10 and PM2.5 coal dust are dangerous substances that can cause pneumoconiosis in workers. Thus, flow simulations of PM10 and PM2.5 coal dust are very important in terms of the working environment in underground coal mines. Figure 10 shows the streamline of coal dust for the case of one duct. The black, red, and blue lines are coal dust, PM10, and PM2.5, respectively. Coal dust moved into the bottom of the mine face under the influence of the flow of fresh air discharged from the blowing duct. PM10 and PM 2.5 flowed to the bottom of the mine face, with some scattering throughout the mine tunnel. Because PM10 and PM 2.5 had small diameters, it was judged to be sporadic scattering without the effect of airflow. Coal dust that moved to the bottom of the mine face flowed along the bottom of the mine tunnel to the outlet. PM2.5 flowed to the outlet because of its small diameter, and PM10 fell to the bottom at the midpoint of the mine tunnel. PM10 and PM 2.5 flowed to the vortical flow with the air. Vortical flow causes PM10 and PM2.5 to stagnate, which affects the health of mining workers. Therefore, ventilation systems of mine tunnels are essential to minimize vortical flow around the working area.

Fig. 10. Streamlines of coal dust flow around the mining face for the case of one duct (blowing only)

256

M. Kim et al.

Coal dust (1x10-4 m)

PM10

PM2.5

X=3

X=12

X=20

Fig. 11. Distribution of averaged volume fraction of dust at the planes of X = 3, 12, and 20 for the case of one duct (blowing)

Figure 11 shows the averaged volume fraction of coal dust for the case of one duct. As mentioned, the averaged volume fraction of coal dust (1 × 10−4 m) presents as zero in the contour because the coal dust flowed along the bottom of the mine tunnel.

Numerical Investigation of Characteristics of Mine Ventilation

257

Fig. 12. Streamlines of coal dust flow around the mining face for the case of two ducts (blowing and exhaust)

Figure 12 shows the streamline of coal dust for the case of two ducts, similar to that for the case of one duct. Coal dust flowed to the bottom, and PM10 and PM2.5 flowed throughout the mine tunnel with the airflow. However, the case of the two ducts had less coal dust in the working area than that of the one duct, as some PM10 and PM2.5 were discharged by the exhaust duct to the outlet. As shown in Fig. 13, the averaged volume fraction of the coal dust was lower than that of the one duct. This seems to have caused coal dust to be discharged to the outlet by the exhaust duct.

258

M. Kim et al.

Coal dust (1×10-4 m)

PM10

PM2.5

X=3

X=12

X=20

Fig. 13. Distribution of averaged volume fraction of dust at the X = 3, 12, and 20 planes for the case of two ducts (blowing and exhaust)

Figures 14 and 15 show the averaged volume fraction of coal dust in the iso-surface for the case of one and two ducts, respectively. Coal dust flowed along the bottom of the mine tunnel, and PM10 and PM2.5 were distributed throughout the mine tunnel with the airflow. All coal dust was discharged after time = 350 s. The case of the two ducts had a lower average volume fraction of coal dust than the case of the one duct. The results show that the ventilation system in the mine tunnel was effective with the blowing duct that blew fresh air into the working area, and the exhaust duct that discharged polluted air. The future study plan is to investigate the distribution of coal dust and methane for increasing mass flow rates of the exhaust duct.

Numerical Investigation of Characteristics of Mine Ventilation

259

Coal dust

T = 10 s

T = 20 s

T = 110 s

T = 350 s

PM10 T = 10 s

T = 20 s

T = 110 s

T = 350 s PM2.5 T = 10 s

T = 20 s

T = 110 s

T = 350 s

Fig. 14. Distribution of averaged volume fraction of dust for time = 10, 20, 110, and 350 s at the iso-surface for the case of one duct (blowing)

260

M. Kim et al. Coal dust =

T = 10 s

T = 20 s

T = 110 s

T = 350 s PM10 T = 10 s

T = 20 s

T = 110 s

T = 350 s

PM2.5 T = 10 s

T = 20 s

T = 110 s

T = 350 s

Fig. 15. Distribution of averaged volume fraction of dust for time = 10, 20, 110, and 350 s at the iso-surface for the case of two ducts (blowing and exhaust)

Numerical Investigation of Characteristics of Mine Ventilation

261

5 Summary and Conclusions This research presents an analysis of three-dimensional fluid flow and particle transport for mining workplaces that have a blowing duct only or blowing and exhaust ducts. These results aid in understanding the characteristics of fluid flow and particle transport around the mining face in mines. Moreover, they facilitate more efficient duct design by the ventilation engineers. The length and height of the mining gallery were 36 and 2.9 m, respectively. Two types of ventilation ducts were selected: one duct for blowing only, and two ducts for blowing and exhaust. The commercial code ANSYS CFX 19.0 was chosen to solve continuity, momentum, and particle transport. ICEM-CFD 19.0 was employed to create hexahedral grids for better solutions. In summary, the results show airflow streamlines, methane distribution, dust streamlines, dust volume fraction for the cases of blowing only, and blowing and exhaust. In the case of methane distribution, the highest mass fraction exists under the discharge plane of a blowing duct. Because blowing flow generates circular motions. In the circumstance of the blowing and exhaust duct, the highest mass fraction exists near the suction plane of an exhaust duct. Because methane distribution is identical to airflow, the position of an exhaust duct is very important. It can control the highest point of mass fraction for methane. With or without the exhaust duct, the heavy dust shows the same results. However, the fine dust of PM 2.5 or PM 10 can be easily affected by airflow. The volume fraction of fine dust is lower in the circumstance of an exhaust duct. It means exhaust ducts are needed when fine dust in the workplace of mining face in a mine are dominant. Acknowledgements. (1) This work was supported by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry & Energy (MOTIE) of the Republic of Korea (No. 20182510102380). (2) The project code of this work for Korea Institute of Geoscience & Mineral Resources (KIGAM) is NP2018-027.

References 1. Parra, M.T., Villafruela, J.M., Castro, F., Mendez, C.: Numerical and experimental analysis of different ventilation systems in deep mines. Build. Environ. 41, 87–93 (2006) 2. Torano, J., Torno, S., Menendez, M., Gent, M.: Auxiliary ventilation in mining roadways driven with roadheaders: validated CFD modeling of dust behavior. Tunn. Undergr. Space Technol. 26, 201–210 (2011) 3. Torano, J., Torno, S., Menendez, M., Gent, M., Velasco, J.: Models of methane behavior in auxiliary ventilation of underground coal mining. Int. J. Coal Geol. 80, 35–43 (2009) 4. Diego, I., Torno, S., Torano, J., Menendez, M., Gent, M.: A practical use of CFD for ventilation of underground works. Tunn. Undergr. Space Technol. 26, 189–200 (2011) 5. Lihong, Z., Christopher, P., Yi, Z.: CFD modeling of methane distribution at a continuous miner face with various curtain setback distances. Int. J. Min. Sci. Technol. 25, 635–640 (2015) 6. Sasmito, A.P., Birgersson, E., Ly, H.C., Mujumdar, A.S.: Some approaches to improve ventilation system in underground coal mines environment - a computational fluid dynamic study. Tunn. Undergr. Space Technol. 34, 82–95 (2013)

262

M. Kim et al.

7. Lu, Y., Akhtar, S., Sasmito, A.P., Kurnia, J.C.: Prediction of air flow, methane, and coal dust dispersion in a room and pillar mining face. Int. J. Min. Sci. Technol. 27, 657–662 (2017) 8. Hasheminasab, F., Bagherpour, R., Aminossadati, S.M.: Numerical simulation of methane distribution in development zones of underground coal mines equipped with auxiliary ventilation. Tunn. Undergr. Space Technol. 89, 68–77 (2019) 9. Torno, S., Torano, J., Ulecia, M., Allende, C.: Conventional and numerical models of blasting gas behaviour in auxiliary ventilation of mining headings. Tunn. Undergr. Space Technol. 34, 73–81 (2013) 10. Kurnia, J.C., Sasmito, A.P., Mujumdar, A.S.: CFD simulation of methane dispersion and innovative methane management in underground mining faces. Appl. Math. Model. 38, 3467– 3484 (2013) 11. Park, J., Park, S., Lee, D.: CFD modeling of ventilation ducts for improvement of air quality in closed mines. Geosyst. Eng. 19(4), 177–187 (2016) 12. Park, J., Jo, Y., Park, G.: Flow characteristics of fresh air discharged from a ventilation duct for mine ventilation. J. Mech. Sci. Technol. 32(3), 1187–1194 (2018) 13. Yi, H., Park, J., Kim, M.S.: Characteristics of mine ventilation air flow using both blowing and exhaust dusts at the mining face. J. Mech. Sci. Technol. 34(3), 1–8 (2020) 14. ANSYS Inc., ANSYS CFX-Solver Modeling Guide, ANSYS (2018)

An Experimental Study on the Turbulent Diffusion Coefficients in Large-Opening Multi-level Limestone Mines Chang-Woo Lee(B) and Van-Duc Nguyen(B) Department of Energy and Mineral Resources, College of Engineering, Dong-A University, Busan 49315, South Korea [email protected], [email protected]

Abstract. The turbulent diffusion coefficient was evaluated using sulfur hexafluoride (SF6 ) as a trace gas in an underground multi-level limestone mine with large cross-sectional airways driven by the room-and-pillar mining method. SF6 concentrations can be measured at the downstream of the release point, and turbulent diffusion coefficients is determined along with the average air velocity from the measured concentration profiles, including multiple modes. The famous Taylor equation was found to be inappropriate as the estimation method for the turbulent diffusion coefficient in the underground mine. On the other hand, a curve-fitting method assuming the existence of multiple modes associated with individual concentration distributions was used to evaluate the turbulent diffusion coefficents. In 1912 m long airway, the turbulent diffusion coefficients were estimated to range between 15 and 18 m2 /s. A higher coefficient implies faster dispersion of the contaminants. In blind airways, ventilated by the auxiliary system and employing several movable equipment types, the turbulent diffusion coefficient was relatively low, ranging from 0.45 to 1.60 m2 /s. This study shows that the concentration of gaseous contaminants diffused in large-opening multi-level limestone mines can be predicted relatively precisely based on the turbulent diffusion coefficients estimated by the procedure described in this paper. It also indicates that engineering solutions for improving the air quality in large-opening underground mines can be derived. Keywords: Trace gas · Turbulent diffusion coefficient · CFD analysis · Large-opening mines · Mine ventilation

1 Introduction Room-and-pillar mining is an underground mining method applied to a wide variety of hard-rock deposits worldwide. In Korea, more than 120 limestone mines have started to go underground in recent years due to the depletion of easily-accessible high-grade ore bodies near the surface and strict environmental regulation. Typically, the local room-and-pillar mines entries are 6–9 m high and 10–15 m wide. Each level can be connected by rampways of 10–13% grade. Horizontal drifts with a vertical elevation © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 263–282, 2021. https://doi.org/10.1007/978-3-030-60839-2_14

264

C.-W. Lee and V.-D. Nguyen

difference of 8–20 m are developed toward the strike direction, sometimes extending several kilometers. The number of drifts at each level depends on the seam; generally, 2–5 drifts in parallel are driven. Since it is challenging to achieve proper ventilation to dilute contaminants due to low air velocity in large-opening underground mines, ventilation efficiency has become a critical issue due to pollutants generated by the intensive deployment of diesel equipment. Furthermore, due to the low air velocity in large-opening entries and the complex ventilation network, pollutants’ diffusion mechanism is hard to understand. Since turbulent diffusion is likely to be more critical than advection-diffusion in the low-velocity airflow, the turbulent diffusion coefficient needs to be scrutinized for the air quality control within the large cross-sectional working area. For this study, a series of field experiments were carried out in a large-opening multi-level underground limestone mine in Korea to estimate turbulent diffusion coefficients. The results were compared with the computational fluid dynamics (CFD) analysis. Ultimately, this study’s obtained results aim to develop the ventilation system that can be applied to derive an engineering solution for improving the air quality in large-opening underground mines.

2 Study Area and Methodology 2.1 Contaminant Diffusion Mechanism In general, the diffusion is defined as a mass transfer phenomenon that makes the distribution of a chemical species more uniform in space as time passes. The most straightforward description of diffusion was given by Fick’s law, developed by Adolf Fick in the 19th century. The molar flux due to diffusion is proportional to the gradient concentration. The concentration rate changes at a point in space proportional to the second derivative of the concentration with space [1]. The rule of the diffusion equation from Fick can be shown in Eq. (1). The magnetic diffusion equation Fick illustrates the diffusion of substances caused by progressive diffusion due to the airflow and the Brown motion of molecules by turbulent flow. The solution of concentration C (x, t) at a downstream point x at time t after dispersion is shown in Eq. (2): ∂c ∂ 2y ∂c = −u + D 2, ∂t ∂x ∂x    v 2 C(x, t) = − √ · e− x−¯u · t /(4, D, t) 2A π, , τ In which: C(x, t) - Concentration of SF6 at point x on t V - SF6 diffusion volume (m3 ) A- Cross-sectional area (m2 ) D - Turbulent diffusion coefficient (m2 /s) t - Time (s) x - Distance (m) and u - Average air velocity (m/s)

(1) (2)

An Experimental Study on the Turbulent Diffusion Coefficients

265

Theoretically, the downstream concentration of the substance is decided by ventilation velocity, concentration gradient, and turbulent diffusion coefficient as a particularly important factor. Diffusion velocity is proportional to the concentration gradient and the turbulent diffusion coefficient. Thus, the effect of turbulent diffusion is the most significant factor for substance diffusion in a limestone mine working zone, where the air velocity is kept very low at 0.1 m/s or lower [2]. Widodo et al. [3, 9] can found several studies of the turbulent diffusion coefficient, such as Arpa et al. [4], Sasaki et al. [5], Widiatmojo et al. [6], and Taylor, G. I [7, 8]. However, there can hardly be any study under shallow air velocity [2]. In this study, the turbulent diffusion coefficients were assessed two different airflow paths: a long airway from the intake portal to the exhaust portal and a relatively short blind airway. The study in the long airway through multi-levels was designed to evaluate the total mine ventilation efficiency. Since the majority of working sites in multi-level underground limestone mines are blind sections with shallow velocity. The turbulent diffusion coefficient in blind sections with portable equipment should be investigated to optimize the auxiliary ventilation system for the blind working sites. 2.2 Experimental Site The experiment works were carried out at D mine in Jecheon City, Chungbuk Province, in the South of Korea. The room-and-pillar mining method was applied in this mine. As illustrated in Fig. 1, a large-opening rampway was developed from the 370 ML surface down to the 6th 200 m level. One or two entries drove each level; entries are approximately 6 m high and 10 m wide. One ventilation shaft was developed from the surface at 370 ML to connect the 4th level at 260 ML. The diameter of the ventilation shaft is 2.0 m. Table 1 shows the specifications of the high-pressure fan installed at

Fig. 1. Ventilation network at Mine D

266

C.-W. Lee and V.-D. Nguyen

the top of the ventilation shaft. At the collar of the shafts equipped with the main fan, ventilation gates were installed; the gates were open when the fan is not operated to apply the natural ventilation force. In Fig. 1, the main fan is operated by the exhaust system; the air flows from the portal through the rampway down to the 4th level, reaches the shaft bottom, and then returns to the atmosphere through the shaft. Table 1. Main fan specification Category

Specifications

Vertical fan installation

Diameter (mm)

1250

Pressure (Pa)

1848

at the

Quantity (m /s)

28.9

operating

Power (kW)

67

point

Frequency (Hz)

60

Blade angle (°)

42

3

2.3 Experimental Method Table 2 shows the location of four monitoring stations and the list of monitoring targets of each station. While the first outside station was designed to measure the atmospheric pressure and temperature, the second station installed just inside the portal measures the temperature and velocity. Similarly, the third station was installed at the rampway connecting the third to the fourth level. One more station was installed at the shaft bottom to measure the temperature and velocity. The data collected during three days of the experiment were analyzed to evaluate the natural ventilation and the fan operation effects on the mine ventilation. For assessing the turbulent diffusion coefficient, the tracer gas method with sulfur hexafluoride (SF6 ) was applied. This method is known to be useful in evaluating the effectiveness of low air velocity in large-opening airways. Figure 2 shows the pictures of SF6 dispersion test and the monitoring station at the shaft bottom.

An Experimental Study on the Turbulent Diffusion Coefficients

267

Table 2. Measurement stations and installation Station no.

Location

Monitoring target

STN 1

Outside mine

Atmospheric pressure, temperature

STN 2

Inside the portal

STN 3

Rampway between 290-260ML

STN 4

At the shaft bottom

Site study

Temperature (air, wall surface), Velocity

(a) SF6 dispersion test

(b) The monitoring station at the shaft bottom

Fig. 2. SF6 dispersion test and monitoring station

3 Experimental Results 3.1 Air Velocity Measurement The air movement, either by the natural ventilation force or by the fan operation, creates a pressure difference between two locations due to various factors such as viscous friction and shock loss. This study evaluates both the natural ventilation and fan force effects on the air movement during the three days of the experiment. The natural ventilation force is defined as the air’s weight difference between the two air columns induces the air into the portal and discharge it through the shaft. However, this airflow direction can be reversed due to the atmospheric temperature changes. Once the air enters the shaft and goes down deeper, its pressure increases due to the air column’s density increase. This phenomenon is called the auto-compression, and the temperature also increases. However, the results could vary depending on the hour, day, or season because of the different atmospheric conditions. In this study, the ventilation efficiency was evaluated by measuring the pressure, temperature, and airflow velocity/direction inside and outside

268

C.-W. Lee and V.-D. Nguyen

the mine. Figure 3 shows the results of air velocity monitoring results in the cases of natural ventilation and fan operation. The natural air velocity without the fan operation was in the range of 0.1–0.2 m/s measured at the shaft bottom. On the other hand, the fan operation’s induced air velocity ranged between 0.5 and 0.8 m/s. The direction of airflow is plotted in Fig. 4. This figure shows the estimation concept and equation of the natural ventilation force. The easiest method to explain how the natural ventilation works are through the concept of density difference between two air columns with the same vertical distance.

(a) by the natural ventilation on the first day

(b) by the fan operation on the first day

(c) by the natural ventilation on the second day

(d) by the fan operation on the second day

(e) by the natural ventilation on the third day

(f) by the fan operation on the third day

Fig. 3. Air velocity distributions by the natural ventilation and the fan operation

An Experimental Study on the Turbulent Diffusion Coefficients C

Main exhaust fan

C

E

E

D

4th level

B

(a) Ventilation by the natural ventilation force

A'

8

8

D

A Shaft

Shaft

Rampway

A

Rampway

A'

269

4th level

B

(b) Ventilation by the main fan operation

Fig. 4. Airflow direction in the ventilation network

Fig. 5. Estimation of natural ventilation pressure

As mentioned above, the air is induced into underground mines as the differential pressure created by the natural ventilation and the fan operation. Mine fans can provide pressure difference mechanically, as shown in Fig. 3 with the fan operation, while the natural draft is caused and maintained by the thermal energy. It is because of the different temperatures inside and outside. The pressure difference depends on both the air densities inside and outside of the mine and the vertical depth difference from the surface to the underground workings (Fig. 5). As a result, the higher the pressure difference, the more the induced air quantity. In practice, the natural ventilation pressure undergoes seasonal changes, which vary the amount of airflow induced or, in some cases, even reverses airflow directions. In particular, in the medium-sized mines of 30,000 tonnes/year (as of 2017), such as D mine, which is the subject of this study, due to the frequent vehicle operation, the piston effect, which is the forced-air flow inside a tunnel caused by moving vehicles, is relatively high. In Fig. 3, the instantaneous air velocity fluctuations were observed between 8:00–17:00, the working period when the equipment, including

270

C.-W. Lee and V.-D. Nguyen

vehicles, were moved around. However, the airflow seems to be interpreted as a pseudosteady state flow. Therefore, it was possible to estimate the average air densities in the two air columns in Fig. 5: the first column from the intake portal to the rampway bottom and the second one corresponding to the ventilation shaft. The average temperature inside the shaft was assumed to be the average temperature of its bottom and the outside. The rampway’s average temperature connected from the intake portal to the fourth level was assumed as the average temperature outside, entrance, and the location connecting the rampway and the fourth level. Figure 6 shows the average temperature inside the two air columns and the distribution of the natural ventilation air velocity flowing into the bottom of the shaft. In the natural ventilation without fan operation, the average temperature of the downcast shaft was relatively lower than the average temperature along the rampway. Consequently, the air was induced into the shaft. The air quantity only by the natural ventilation ranged between 0 and 11.7 m3 /s. In Fig. 7, the temperature differences between the upcast and downcast sections and the air velocity at the shaft bottom, the positive correlation was clearly observed, and its coefficient was 0.8. As a result, in the case of D mine, the ventilation velocity prediction was possible by monitoring the temperatures at the locations mentioned above.

a. 1st-day

b. 2nd-day

c. 3rd-day

Fig. 6. Variation of temperature and natural velocity in the up-cast and downcast airway

3.2 The Turbulent Diffusion Coefficient in the Total Mine Ventilation System The diffusion experiment with SF6 was performed to analyze the dynamic characteristics of the airflow in the total mine ventilation system. As a trace gas, about 122–168 L of SF6 was infused into the balloon, dispersed at the intake portal, and its concentration was monitored at the exhaust shaft bottom by INNOVA’s Model 1312 Photo-acoustic gas monitor. The 1312 photoacoustic Multi-gas monitoring is a highly accurate, reliable, and stable quantitative gas monitoring system. Its measurement principles are based on the photoacoustic infrared detection method. The 1312 can measure almost any gas that

An Experimental Study on the Turbulent Diffusion Coefficients

271

Fig. 7. Relationship between the induced velocity and the upcast-downcast temperature differences

absorbs infra-red light. The reliability of measurement results can be ensured by regular self-tests, which the 1312 performs. Accuracy is ensured by the 1312’s ability to compensate for any measurement for temperature fluctuations, water-vapour interference, and other gases known to be present. The SF6 diffusion experiments are outlined in Table 3. In the case of the two tests during the fan operation, the trace gas dispersed at the intake portal was detected at the shaft bottom 34 and 36 min after diffusion, respectively. The monitor measurements reached the peak concentration after 63 and 62 min, respectively, and these results were very similar in both tests. Meanwhile, in the two experiments during the natural ventilation without fan operation, the monitored concentration peaked after 187 and 177 min, respectively. The average air velocities estimated at the peak concentration during the fan operation were 0.50 and 0.52 m/s, respectively; the velocities were almost identical in two different tests. These figures in case of the natural ventilation pressure were 0.17–0.18 m/s. The SF6 concentration was estimated by Eq. (2) to compare with the experimental data to analyze the turbulent diffusion coefficients in the mine’s whole ventilation system. In Fig. 8, the measured and estimated concentrations are compared. In all experiments, similarity can be observed in terms of the initial concentration observation time, the initial concentration increase trend, and the peak concentration arrival time and peak concentration. The relatively higher concentrations observed after the peak concentration arrival seem to be due to the increase of the turbulence intensity due to the equipment movement and partly due to the gas’s arrival, which was diffused into the various spaces in the upstream and re-entrained into the airflow later. In natural ventilation, the concentration was exceptionally well fitted by the estimation model, as shown in Fig. 8(c). There are several studies of the turbulent diffusion coefficient for large-opening limestone underground mines in Korea by Lee, Chang-Woo [10], Lee, Seung-ho [11], and Kim, Doo-young [12]. The turbulent coefficients published in those previous studies are in the range of 0.01–3.4 m2 /s. All of the previous studies were conducted in relatively

272

C.-W. Lee and V.-D. Nguyen Table 3. Outline of the SF6 trace gas experiment

Category

During the fan operation

During the natural ventilation

Test 1

Test 3

Releasing SF6 gas time

3:26:00 PM 2:25:00 PM 10:36:00 AM 11:51:00 AM

SF6 gas volume (L)

122

Test 2 132

168

Test 4 168

Time to detect gas

4:15:00 PM 3:10:00 PM 12:33:39 PM –

Time to reach peak concentration

4:30:45 PM 3:26:00 PM 1:38:14 PM

2:47:02 PM

Total monitoring time (min)

63

62

187

177

0.48

0.17

0.18

Distance from SF6 releasing point to 1912 monitoring point (m) Average velocity (m/s)

(a) 1st SF6 test

0.50

(b) 2nd SF6 test

(c) 3rd and 4th SF6 test

Fig. 8. Comparison of SF6 by theoretical and experimental data

short airways less than 500 m. The results can be applied as the turbulent diffusion coefficients to predict and control the pollutant diffusion in mining workplaces. However, the airways evaluated in this study were long airways, with a length of 1912 m covering the total mine ventilation system. The turbulent diffusion coefficients of the fan operation and natural ventilation are estimated to be between 15 and 18 m2 /s, and between 5.2 and 8.2 m2 /s. The experiment results are summarized in Table 4. In a study of the turbulent diffusion coefficient by Kim et al. [2, 12] in a domestic limestone mine, the distance measured was in the range of 45–141 m, and the equivalent diameter (Da) was 8.76–12.16 m. L/Da was 4–15. The turbulence diffusion coefficient was in the range of 0.01 m2 /s–1.00 m2 /s, while the average was estimated to be 0.25 ± 0.06 m2 /s. Kim et al. [2, 12] stated that the turbulent diffusion coefficient is proportional to L/D (tunnel length/diameter), as shown in Fig. 9. The diffusion coefficient estimated by Kim’s model with L/D (tunnel length/diameter) of 205.5, which is the typical number in this study, is in the range of 1–10 m2 /s. These estimated coefficients range includes the diffusion coefficients of 5.2 m2 /s and 8.2 m2 /s observed in this study during the natural ventilation. The turbulent diffusion coefficient can be quantitatively analyzed in terms of diffusion and migration, making it an indispensable parameter for improving the air quality of the mine by ventilation [13].

An Experimental Study on the Turbulent Diffusion Coefficients

273

Table 4. Summarized results of turbulent diffusion in the total ventilation system Category

During the fan operation

During the natural ventilation

Test 1

Test 2

Test 3

Test 4

Average velocity (m/s)

0.50

0.48

0.17

0.18

Turbulent diffusion coefficient (m2 /s)

15

18

5.2

8.2

Distance to the monitoring location (m)

1912

Gas volume (L)

122

132

168

168

Airway area (m2 )

68.3

Fig. 9. Linear regression analysis of D and L/Da [12]

3.3 The Turbulent Diffusion Coefficient in the Blind Airway A similar diffusion experiment with SF6 was performed at the 6th level (220 ML) of D limestone mine, as shown in Fig. 1, to analyze the turbulent diffusion coefficient in the blind airway. Figure 10 shows the detailed study site at the 6th level. Approximately 78, 157, and 188 L of the trace gas, SF6 , was infused into the mine. The same monitor monitored its concentration at 100 m downstream of the blind working face. At the study site, four-velocity monitoring stations were installed, as shown in Fig. 10. The number of mining equipment such as trucks and drilling machines deployed at the blind face was 3–4 types. In this experiment, the auxiliary ventilation system’s effects and the moving mining equipment on the turbulent diffusion coefficient were analyzed. Two fans were

274

C.-W. Lee and V.-D. Nguyen

installed side by side as free-standing auxiliary fans. The distance from fans to the blind working face was 170 m. On the one side, a blowing fan with the outlet velocity of 38 m/s was installed to blow air to the blind working face. In contrast, on the opposite side, an exhaust fan with an outlet velocity of 35 m/s was installed to exhaust the contaminated air dispersed from the working site. The fan specifications are described in Table 5. Details of the monitoring station installation are illustrated in Fig. 10. The experiment results are compared with the CFD analysis output to evaluate the turbulent diffusion coefficient in blind airway influenced by the auxiliary ventilation and the moving mining equipment.

Fig. 10. Mine study site at the blind entry

Table 5. Fan specifications for the test in the blind entry Categories

37 kW fan 22 kW fan

Discharge velocity (m/s) 35

38

Diameter (m)

1.4

0.91

Pressure (Pa)

555.1

1275

Noise level (dB(A))

89–95

103–105

Power (kW)

37

22

Length (m)

3.0

3.45

Weight (kg)

998

1.090

The air velocity measurement results at the monitoring station 1 to 4 before and after fan operation were plotted in Fig. 11. The monitoring station 2, 40 m downstream of the exhausting fan, and the monitoring station 3, 30 m downstream of the blowing

An Experimental Study on the Turbulent Diffusion Coefficients

275

fan, were directly influenced by the fan operation. After turning on the fans, the air velocity increased to approximately 2.90 m/s and 4.1 m/s. Regarding the SF6 trace gas

(a) First monitoring station

(b) Second monitoring station

(c) Third monitoring station Fig. 11. Data of air velocity measurements

276

C.-W. Lee and V.-D. Nguyen

(d) Fourth monitoring station Fig. 11. (continued)

Fig. 12. Data of SF6 concentration measurements

experiment, two fans were turned on at 13:49:00, and 78 L of SF6 was injected at the face. From then on, SF6 injection was repeated twice at 15:02:18 and 15:43:23, as shown in Fig. 14. The SF6 concentration monitored at the first station, 100 m downstream of the blind working face, is plotted in Fig. 12 and summarized in Table 6. Table 6. Trace gas experiment results of SF6 in the blind airway Category

Case 1

Case 2

Case 3

Gas volume (L)

78

157

188

Peak concentration (ppm)

20.31

41.11

54.33

Total time to reach the peak concentration

18

17

13

Distance from releasing SF6 gas to the monitoring station (m)

100

Average velocity (m/s)

0.092

0.098

0.128

In Case 1, after 18 min, the SF6 reached the peak concentration and was kept stable. The total time to the peak concentration in Case 2 and 3 was 17 min and 13 min,

An Experimental Study on the Turbulent Diffusion Coefficients

277

respectively. As a result, the shallow average velocity of 0.092–0.128 m/s was estimated in this experiment. The detailed experiment results of SF6 trace gas are summarized in Table 6. The geometry in Fig. 13 was used for CFD analysis to evaluate the turbulent diffusion coefficient in the blind airway. The initial conditions of the CFD analysis are shown in Table 7. SF6 was dispersed at the working face, as described in Table 6. Two fans in Table 5 were installed for this analysis.

Fig. 13. Geometry for the CFD analysis Table 7. The conditions for CFD analysis Parameters

CFD model

Boundary of inlet

Velocity inlet

Boundary of wall

Friction wall

Ventilation resistance (k)

0.014 kg/m3

Wall temperature

20 °C

Type of mesh

Tetrahedron elements

Solution model

Turbulence model (k − ε)

Mesh size function

Proximity and curvature

Number of mesh elements 1.000.000 Simulation condition

Transient-state conditions

Figure 14 shows the U-velocity (axial) profile at 1.0 m height by the CFD analysis. The low air velocity distribution can be observed at the working face. Further, as shown in Fig. 14, the CFD analysis results were reasonably similar to the experimental measurements. Since the monitoring stations 2 and 3 are the locations directly influenced by the

278

C.-W. Lee and V.-D. Nguyen

fan operation, the measurements are noticeably fluctuating with the averages of 2.9 m/s and 4.1 m/s, respectively. On the other hand, the average velocities at the monitoring station 1 and 4 are relatively low; 0.22 m/s and 0.25 m/s.

Fig. 14. U velocity profile by the CFD analysis

(a) First monitoring station

(b) Third monitoring station

(c) Second monitoring station

(d) Fourth monitoring station

Fig. 15. Validating the CFD analysis results

An Experimental Study on the Turbulent Diffusion Coefficients

(a) Case 1

(b) Case 2

(c) Case 3

Fig. 16. Comparison of the theoretical model, experiment data and CFD analysis results

279

280

C.-W. Lee and V.-D. Nguyen

Similarly, to analyze the turbulent diffusion coefficient, the SF6 concentration was calculated by Eq. (2) to compare with the experiment data. The curves in Fig. 16 were estimated by Eq. (2), and the figures also include the SF6 gas concentration in the blind airway by the CFD analysis. In general, due to the low air velocity at the working site, higher SF6 concentration can be observed at a distance of 100 m from the blind working face. Particularly in Fig. 16 (a) of Case 1, the bi-modality analysis of SF6 concentration distribution shows high overall goodness of fit between the experimental and numerical results. The turbulent diffusion coefficient was 0.3 m2 /s for mode 1, and this figure was 0.15 m2 /s for mode 2. A similar pattern can be found in Figs. 16 (b), (c) for Case 2, and Case 3. The estimated turbulent diffusion coefficient in case 2 was 1.4 m2 /s, and this figure was 0.6 m2 /s in Case 3. The detailed turbulent diffusion analysis can be seen in Table 8. Further, a study on the turbulent diffusion coefficient of contaminants in an underground limestone with a broad cross-section using tracer gas by Kim et al. [2, 12] found that the turbulent coefficient estimated by Taylor’s equation was 0.195.57 m2 /s. This value was not appropriate for applying in limestone mines in Korea, adopting the room-and-pillar method. The study results by Kim et al. [2, 12] indicated that the concentration profile of SF6 by Fick’s equation with multiple modes in Eq. 2 demonstrated a good fit for the experiment data. The turbulent diffusion coefficient was estimated to be 0.01–1.80 m2 /s. This study’s results for analyzing the turbulent diffusion coefficient in the blind airway was 0.45–1.6 m2 /s. Table 8. Analysis of turbulent diffusion in the blind airway Category

Case 1 Case 2 Case 3

Average velocity (m/s)

0.092

0.098

0.128

Turbulent diffusion coefficient (m2 /s)

0.45

1.4

0.6

157

188

Distance to the monitoring station (m) 100 Gas volume

78

Airway area (m2 )

65

4 Conclusion This study was conducted to analyze the turbulent diffusion coefficients of at a largeopening underground limestone mine with 67 kW main fan installed on the top of the 110 m-long ventilation shaft. The turbulent diffusion coefficient was analyzed for the cases of the total mine ventilation and natural ventilation. Additionally, the turbulent diffusion coefficient in the blind airway was influenced by the auxiliary ventilation system, and the moving equipment was experimentally estimated and compared with the CFD analysis. The results are summarized as follows: 1. At the study site, the natural ventilation was induced the air quantity ranging from 0 to 11.7 m3 /s. Since the natural ventilation is a phenomenon due to the difference in

An Experimental Study on the Turbulent Diffusion Coefficients

2.

3.

4.

5.

281

the air density between the intake and returns air columns, the relative correlation coefficient between the mean temperature and the natural ventilation rate was 0.8. On the other hand, with the main fan operation, the exhaust air velocity in the lower part of the shaft is 0.45–0.55 m/s, and the exhaust air quantity was 20.3–24.8 m3 /s. The turbulent diffusion coefficients in 1912 m-long airway from the intake portal to the exhaust ventilation shaft were 15 m2 /s and 18 m2 /s during the fan operation. In the case of the natural ventilation, the coefficients were 5.2 m2 /s and 8.2 m2 /s. Thus, relatively sizeable turbulent diffusivity observed in this study coincides with previous research on the long tunnels. Therefore, the gas dispersed from the intake portal reached the shaft bottom 29 and 26 min faster than the airflow moving at 0.5 m/s and 0.52 m/s. Since pollutants diffuse relatively faster than the air stream, it is necessary to exhaust air rapidly to control the air quality. The exhaust shaft can be an effective solution. In the blind airway, the turbulent diffusion coefficient was estimated to be 0.45– 1.60 m2 /s, and it demonstrated higher goodness of fit for the estimates by Fick’s equation with multiple modes. Even under the effect of the auxiliary ventilation system and moving equipment at the blind airway, this study shows that air velocity was under 0.1 m/s at the working environment. The higher SF6 concentration was confined at the working site. Thus, in some cases, heavy diesel mining equipment can improve the working condition in mines. The CFD results were well validated with the experiment data in case of the blind airway. The CFD analysis results of the SF6 concentration profile demonstrated a good fit for the experiment data. Since the turbulent diffusion coefficient was in the range of 0.45–1.60 m2 /s, higher SF6 concentration was confined near the blind working face. The ventilation shaft installed in the limestone of this study showed significant effects on natural ventilation and mechanical ventilation. However, in the blind airway at the lowest level of the mine, the low ventilation efficiency was observed with low turbulent diffusion coefficient and superficial velocity in the blind entry. Thus, these results imply that it is necessary to improve the ventilation efficiency and effectively maintain the working condition in the large-opening underground limestone mines by alternative methods

Acknowledgment. This research was partly supported by grants of “Development and Onsite Demonstration of Smart ICT/IoT-Based Mining Smart Ventilation System” (grant No. 20182510102380) funded by the Ministry of Trade Industrial and Energy of the Korean government.

References 1. Tyrrell, H.J.V.: The origin and present status of Fick’s diffusion law. J. Chem. Educ. 41(7), 397 (1964) 2. Kim, D.Y., Lee, S.H., Jung, K.H., Lee, C.W.: Study on the turbulent diffusion coefficients of contaminants in an underground limestone mine with large cross section using tracer gas. Geosyst. Eng. 16(2), 183–189 (2013)

282

C.-W. Lee and V.-D. Nguyen

3. Widodo, N.P., et al.: Mine ventilation measurements with tracer gas method and evaluations of turbulent diffusion coefficient. Int. J. Min. Reclam. Environ. 22(1), 60–69 (2008) 4. Arpa, G., et al.: Tracer gas measurement and simulation of turbulent diffusion in mine ventilation airways. J. Coal Sci. Eng. (China) 14(4), 523–529 (2008) 5. Sasaki, K., Dindiwe, C.: An integrated mine ventilation simulator MIVENA Ver.6 with application. In: Proceedings of the North American 9th US Mine Ventilation Symposium, Ontario, Canada (2002) 6. Widiatmojo, A., Sasaki, K., Sugai, Y., Suzuki, Y., Tanaka, H., Uchida, K., Matsumoto, K.: Evaluation of mine size effect on turbulent diffusion from tracer gas measurement data and numerical simulation. In: The 14th U.S./North American Mine Ventilation Symposium, pp. 4– 13. Taylor and Francis Group, London (2012) 7. Taylor, G.I.: Dispersion of soluble matter in solvent flowing slowly through a tube. Proc. R. Soc. Lond. Ser. A 219, 186–203 (1953) 8. Taylor, G.I.: The dispersion of matter in turbulent flow through a pipe. Proc. R. Soc. Lond. Ser. A 233, 446–468 (1954) 9. Widodo, N.P., Sasaki, K., Sugai, Y., Gautama, R.S., Widiatmojo, A.: A laboratory experiment for tracer gas diffusion. In: Proceedings of 4th International Workshop on Earth Science and Technology, pp. 279–284 (2006) 10. Lee, C.W., Yang, W.C.: A study on diffusion coefficients of diesel exhaust in coal mine airways using a racer gas. Korean Soc. Miner. Energy Resour. Eng. 31, 483–490 (1994) 11. Lee, S.H.: Analysis of the ventilation characteristics and turbulent diffusion coefficients in a local underground limestone mine with large cross-section. (Unpublished MS thesis) Dong-A University, Busan, North Korea (2012) 12. Kim, D.Y.: A study on the turbulent diffusion coefficients for the air quality control in domestic limestone mines. (Unpublished Ph.D. thesis) Dong – A University, Busan, North Korea (2018) 13. Lee, C.W., Nguyen, V.D.: Development of a low-pressure auxiliary fan for local large-opening limestone mines. Tunn. Undergr. space 25(6), 543–555 (2015)

Increasing Productivity and Safety in Mining as a Chance for Sustainable Development of Vietnam’s Mining Industry Duong Duc Hai1(B) , Le Duc Nguyen1 , Nguyen Duc Trung1 , Marian Turek2 , and Aleksandra Koteras2 1 Institute of Mining Science and Technology (IMSAT), Hanoi, Vietnam

[email protected] 2 Central Mining Institute (GIG), Katowice, Poland

Abstract. According to the Master Plan of Coal Industry Development in Vietnam, the demand of coal for domestic consumption keeps increasing annually. Meanwhile, its mining capacity has nearly reached the maximum value. In addition, working conditions become increasingly hostile as mining depth increases. On solving above difficulties, Vietnam’s mining industry is strongly implementing the restructuring program and applying technological advances to improve safety and drive mining efficiency and productivity gains. However, the proposed solutions have not helped mining companies to thoroughly solve the problem of increasing productivity yet. Besides, the concept of productivity is still understood in a narrow sense, which is labor productivity. The purpose of this article is to study and give the basic definitions and measuring methods of productivity for underground coal mines in Quang Ninh province. Comparative evaluations of the technical and economic efficiency, as well as the relation between productivity and labor cost, in thirteen underground coal mines, will be analyzed. From these obtained results, the paper aims to propose the most important solutions to increase the productivity and safety of Vietnamese coal companies. Keywords: Productivity · Safety · Labour costs · Management · Mining

1 Introduction As we move towards the second decade of the 21st century, thanks to the application of scientific and technical advances to all stages in the production process, Vietnam’s coal industry has had a strong growth to meet the development needs of the national economy. Especially in underground coal mines, the improvement of levels and efficiency of mechanization in the longwalls has created a great motivation to increase their coal production, productivity and the standards of safety, gradually replacing coal production from the surface mines. However, in recent years, underground coal mines have been exploited deeper and further, leading to increase in the costs of transportation, ventilation, drainage… Meanwhile, labor productivity is still low. The above reasons cause © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 283–307, 2021. https://doi.org/10.1007/978-3-030-60839-2_15

284

D. D. Hai et al.

the unit cost of coal production to rise, which is making it difficult to compete with the imported coal prices. To solve these problems, the research of solutions for increasing productivity and safety in underground coal mines is extremely necessary for Vietnam’s coal industry. In a market economy, productivity is an important factor that enhances the competitiveness of enterprises in general and mining enterprises in Vietnam’s coal industry in particular. In the current period, when the price of minerals is much lower on the world market than ten years ago [1], increasing productivity, which aims to improve competitiveness, is the target of all businesses. In the coal industry, there is a buzz phrase “Today it is not hard to excavate, but to sell what you excavated”. Their simple words perfectly characterize the most important problem which mining enterprises are facing now because you can sell, with profit, only what has a competitive cost [2]. The fundamental condition is the effective functioning of a mining enterprise. One of the measures of effectiveness is productivity. For many years Vietnam’s coal mining industry has been undergoing restructuring processes with the aim of increasing the effectiveness of coal mines. There are many reasons (geological and mining, technical, infrastructural, social, and political) why the efforts have not been achieved as expected. However, productivity improvement, while maintaining high work-related safety, is a top priority, as it is an indispensable condition if the industry is to survive and develop. The purpose of this paper is to identify and propose solutions to improve productivity and safety in underground coal mines of Quang Ninh coalfield. The paper aims to answer the following research questions: • What is productivity, and how is it expressed in underground coal mines in Quang Ninh coalfield? • What are the key drivers of productivity in the Vietnam’s coal mining industry? • What is the state of work safety in the Vietnam’s underground coal mines? • How can the Vietnam coal mining sector be improved in future productivity and safety? In addition, two research hypotheses are made: • H1: Mines with a high share of labor costs in the total cost structure are characterized by low productivity, • H2: Mines with a high intensity of natural hazards have low productivity. Research method: The measurement of productivity in the underground coal mines will be made using the indicators calculated on the basis of available data from the Vietnam National Coal - Mineral Industries Holding Corporation Limited (VINACOMIN) and the unpublished data from the years 2011–2019. In addition, the paper will be undertaken a benchmarking study of the productivity performance of Vietnam’s underground coal mines. In this research, primary documentation - economic and technical - from the underground coal mines examined, and reports prepared at the Institute of Mining Science and Technology -Vinacomin (IMSAT) were used.

Increasing Productivity and Safety in Mining

285

2 What Is Productivity? In market economy conditions, there is no room for an enterprise, also a mining one, which functions “at all cost”, ignoring the balance of revenue and expenditure associated with its operation [2]. It has to be effective. Productivity, as one of the measures of the efficiency of the enterprise, is commonly defined as a ratio of the amount of products produced and sold in a given period to the number of input resources used or consumed. This means that productivity can be interpreted as the efficiency of using the system’s input resources such as energy, materials, human labor, capital, information, space, time in the production of goods and services that constitute its output. The technical concept of productivity is illustrated in Fig. 1.

INPUT QRI r = 1, 2, ..., R Labour Capital Material Energy Fixed assets ...

PROCESSES

PRODUCTIVITY T

P

=

OUTPUT QTO t = 1, 2, ..., T - The amount of made/sold product - The value of product sold ...

∑Q t =1 R

O t

∑Q r =1

I r

Fig. 1. Technical concept of productivity [3]

In some countries with the developed coal mining industry in the world, there have been many pieces of research on productivity, including works [4–9]. However, in Vietnam, there are very few research papers on productivity in the coal industry, such as [1, 10, 11] in which the authors mainly focused on labor productivity or mention a small aspect of productivity. For a coal mining enterprise, the effect is the amount of produced coal [12]. The applied mining systems, which consider deposit conditions within given mining areas and the need to provide work-related safety (first of all preventing natural hazards) and protecting the surface, ought to ensure the followings (Fig. 2): • possibly maximal surface and environmental protection; • minimal losses of deposit; • maximal profitability.

286

D. D. Hai et al.

Fig. 2. Conditions and effects of productivity of a mining enterprise [2]

3 The Yardsticks of Productivity in the Coal Industry The purpose of productivity measurement is to define the relationship between the system’s outputs and inputs. Knowing this relationship helps the mining companies to make better decisions and effectively manage the production system. Thus, it can be said that the measurement of productivity is an important element of system performance control. According to the traditional method, productivity is measured in underground mines as the term “coal output per man shift”, which means The amount of produced coal in tons/The number of workers x Shifts. It is clear that the above definition only deals with labor productivity. Each underground coal mines have specific geotechnical conditions, and they will have different control systems for their operations (e.g., cost, price, finance, production, inventory, quality, etc.). Therefore, the productivity measurement should be considered in combination with economic and technical factors, not just labor productivity. To solve this problem, since 1970, countries with the developed coal industry in the world, such as the USA, Canada, the EU, and Australia, have introduced the method of measuring the total factor productivity (TFM) for the coal industry. TFM most often considers all factors of production of an expenditure stream - technological level, skills of personnel, organization of an enterprise, changes in capital expenditure, relations in planning and organizing a production process - thus providing a fuller picture of a mining enterprise. In the study [7] on productivity in underground coal mines in the USA, he stated that to achieve a measure of overall productivity, including all inputs, was near impossible. Furthermore, [1] and [2] showed that total productivity index is not used in hard coal mining. Due to significant differences in geological and mining conditions in given coal mines, there is no point in creating a “universal” index, which certainly would have to be “modified” in each mine to adjust it to its specific conditions. According to [11], there are two basic indicators in underground coal mines, such as technical efficiency and economic efficiency, which were adopted as measures of productivity for this way.

Increasing Productivity and Safety in Mining

287

The technical efficiency is measured as labor productivity. Meanwhile, the economic efficiency is determined as the quotient of sales revenues to the number of persons of the underground coal mine. In underground coal mines, productivity is usually considered partially, in the following aspects: • Effectiveness of coal production to the number of used resources, e.g., materials, energy, capital investment; • Determination of the current state referred to what could be achieved as far as production and used resources are concerned; • Comparability of indices calculated to analyze a given coal mine with other ones; • Trends were determining changes in productivity indices and their tendencies over time. Partial productivity focuses on the fragmentary scope, which may result in unfavorable consequences for an enterprise itself. For instance, focusing on improving general or underground productivity may lower total productivity, e.g., by wasting the effective time of work of expensive equipment in extraction workings (Table 1). Table 1. Examples of partial productivity measures in a mining enterprise [2] No. Partial productivity measures

Example

1

Labor productivity

– Number of kilograms of coal to the number of worked working days of all the personnel (so-called general productivity, expressed in coal) – Number of kilograms of run-of-mine coal to the number of worked working days of the underground personnel (so-called underground productivity, expressed in run-of-mine coal) – Number of produced tonnes of coal to the number of all the personnel (so-called productivity per an employee) – Value of sold coal (revenue) to the costs of remuneration with overheads of all the personnel – Value of sold coal (revenue) to the number of personnel/underground personnel

2

Machine and equipment effectiveness – Number of tonnes of run-of-mine coal to the number of worked hours most often referred to a given type of equipment, e.g. shearers, conveyors - so-called equipment effectiveness – Number of worked minutes per shift - the so-called effective time of work in a working – Number of minutes of downtime due to a failure to the number of ton of run-of-mine coal - so-called failure index

3

Energy productivity

– Number of tonnes of run-of-mine coal/coal to the installed/used power so-called energy productivity index

4

Capital productivity

– Number of produced/sold coal to the costs of used or employed resources. Revenue from the sales of coal to the costs of used or employed resources – Value of produced coal to the value of “frozen” materials and stock or fixed assets (buildings, structures, machines, equipment, installations, networks)

5

Mining face productivity

– Number of produced tonnes of coal - in one longwall, per shift, day, month – Length of driven roadways - per one shearer, per shift, day, month

288

D. D. Hai et al.

4 Productivity and Safety in Vietnam’s Underground Coal Mines 4.1 The Data and Method of Measuring the Productivity of Underground Coal Mines in Quang Ninh Coalfield An essential element of the productivity improvement process is its measurement, especially the proper selection of productivity measures. This part of the paper discusses various approaches to measuring productivity in underground coal mines. Productivity and safety were analyzed in twelve underground coal mines of VINACOMIN, operating in 2015 in the Quang Ninh coal basin. As for indicators of the productivity determined in this way, two indicators were adopted, common in the hard coal mining industry, labor productivity and capital productivity (partial productivity measures). When measuring it in the mines, two basic parameters were used, which included the volume of extraction, traditionally expressed in tonnes and in the so-called tonne of conventional fuel tpu (the amount of energy generated by burning 1 metric ton of coal − 1 tpu = 29.302 GJ), which also allowed for taking into account the calorific value of the extracted raw coal, and thus not only quantitative but also qualitative parameters. The parameters listed, on a scale of one year, were successively referred to as the total employees and the underground workers. In this way, the following indicators were used in the assessment: • Overall labor productivity, calculated as the quotient of the volume of extraction in tones or tpu to the total number of personnel of the underground coal mine; • Partial labor productivity, calculated as the quotient of the extraction volume in tones or tpu to the number of underground personnel of the underground coal mine; • Overall capital productivity, calculated as the ratio of sales revenues to the total number of personnel of the underground coal mine; • Partial capital productivity, calculated as the quotient of sales revenues to the number of underground personnel of the underground coal mine. In addition to the above ratios, the share of labor costs in total costs was also expressed as a percentage. In order to analyze the level of safety in underground coal mines, commonly recognized categories of natural hazards occurring were used. Methane, fire, flood, collapse, coal spontaneous combustion potential, and dust or gas explosion are the most critical hazards specifically linked to underground mining. Therefore, the assessment included (1) methane hazards - classified in five categories (from I to super emitters) according to the Department of Industrial Safety and Environmental Technology, Ministry of Industry and Trade and (2) the level of coal spontaneous combustion potential - classified in five categories (from I to V) based on research results of the Central Mining Institute (GIG). Currently, there is no standard to classify hazards for the remaining factors, so this paper will not be included in the assessment.

Increasing Productivity and Safety in Mining

289

4.2 The Results of Productivity Measurement of Underground Coal Mines in Quang Ninh Coalfield The value of all listed indicators of productivity in underground coal mines is presented in Tables 2 and 3. In addition, basic statistical indicators are shown in Table 4 to systematize the research data. Table 2. Indices of labor and capital productivity in the tested underground coal mines in 2015 Indicator

Unit

Underground coal mines Mao Khe

Nam Mau

Vang Danh

Uong Bi

Ha Lam

Hon Gai

Overall labour productivity

tonnes/all the persons

372

412

533

338

529

416

Partial labour productivity

tonnes/underground persons

555

566

772

638

871

693

Overall labour productivity in calorific value

tpu/all the persons

239

380

533

346

386

297

Partial labour productivity in calorific value

tpu/underground persons

357

522

772

652

635

495

Overall capital productivity

Sales revenue/all the persons

129

122

152

126

121

123

Partial capital productivity

Sales revenue/underground persons

192

168

220

237

199

205

According to the data contained in Tables 2 and 3, the tested underground coal mines are characterized by a very large variation in the area of partial labor productivity. They are defined by a high range value and a coefficient of variation. Besides, the partial labor productivity expressed in tpu shows a slightly higher variability than the values expressed in the metric ton, which further emphasizes the diversified qualitative varieties of extracted raw coal in individual mines. The capital productivity in the underground mines is less diverse than labor productivity, which is reflected by more than twice lower the values of coefficients of variation. In the three best underground coal mines such as Vang Danh, Ha Lam, Khe Cham, overall labor productivity exceeds 455 tonnes per year for one person in total and 700 tonnes per one underground worker. The worst mine (Mong Duong) has overall labor productivity below 315 tonnes and below 500 tonnes for one underground worker. For eight other underground coal mines, the overall labor productivity is in the range from 330 to 470 tonnes, and partial labor productivity is in the range from 500 to 700 tonnes per worker/year.

290

D. D. Hai et al.

Table 3. Indices of labor and capital productivity in the tested underground coal mines in 2015 Indicator

Unit

Underground coal mines Quang Hanh

Thong Nhat

Ha Long

Duong Huy

Khe Cham

Mong Duong

Overall labour productivity

tonnes/all the persons

373

455

393

466

455

315

Partial labour productivity

tonnes/underground persons

571

721

587

719

705

479

Overall labour productivity in calorific value

tpu/all the persons

320

423

292

499

539

293

Partial labour productivity in calorific value

tpu/underground persons

490

670

436

771

834

445

Overall capital productivity

Sales revenue/all the persons

128

127

127

117

155

128

Partial capital productivity

Sales revenue/underground persons

196

201

189

181

240

195

Table 4. Statistical measures of indices of labor and capital productivity in the tested underground coal mines in 2015 Indicator

Unit

Statistical measure Max

Min

Range

Average

Standard deviation

Coefficient of variation (%)

Overall labour productivity

tonnes/all the persons

533

315

218

422

66

15.63

Partial labour productivity

tonnes/underground persons

871

479

392

656

105

16.06

Overall labour productivity in calorific value

tpu/all the persons

539

239

299

379

96

25.43

Partial labour productivity in calorific value

tpu/underground persons

834

357

477

590

147

24.98

Overall capital productivity

Sales revenue/all the persons

155

117

38

130

11

8.66

Partial capital productivity

Sales revenue/underground persons

240

168

72

202

20

10.14

Although the overall labor productivity in tpu is slightly different, the order of the top three mines with the best results was changed. The Khe Cham coal mine is ranked best, followed by Vang Danh and Duong Huy coal mines. However, Mao Khe coal mine remains the worst. In terms of partial capital productivity, three underground coal mines (Ha Lam, Uong Bi, Khe Cham) achieve a result of over 220 million VND per person/year. It is highly rated in previous categories by Khe Cham coal mine, and Uong Bi coal mine rated lower. Nam Mau is the worst coal mine because it is the only mine in the studied coal mine group to generate income per one person below 170 million VND.

Increasing Productivity and Safety in Mining

291

Table 5. The ranking of tested underground coal mines in 2015 No.

Partial labor productivity

Partial labor productivity in calorific value

Partial capital productivity

1

Ha Lam

Khe Cham

Khe Cham

2

Vang Danh

Vang Danh

Uong Bi

3

Thong Nhat

Duong Huy

Vang Danh

4

Duong Huy

Thong Nhat

Hon Gai

5

Khe Cham

Uong Bi

Thong Nhat

6

Hon Gai

Ha Lam

Ha Lam

7

Uong Bi

Nam Mau

Quang Hanh

8

Ha Long

Hon Gai

Mong Duong

9

Quang Hanh

Quang Hanh

Mao Khe

10

Nam Mau

Mong Duong

Ha Long

11

Mao Khe

Ha Long

Duong Huy

12

Mong Duong

Mao Khe

Nam Mau

- the best mine

- the worst mines

Table 6. Share of labor costs in total costs of the tested underground coal mines in 2015 (increasing) Underground coal mine

Hon Gai

Ha Lam

Ha Long

Duong Huy

Thong Nhat

Khe Cham

Share, %

20.37

26.58

28.17

28.45

29

29.5

Underground coal mine

Mong Duong

Uong Bi

Quang Hanh

Vang Danh

Nam Mau

Mao Khe

Share, %

31.04

31.3

31.3

31.76

33.8

36.73

As a summary of the assessment of productivity, the ranking of tested underground coal mines in Quang Ninh coalfield taking into account the partial labor productivity in tonnes and tpu, as well as the capital productivity, as shown in Table 5. Due to the large diversity of assessments in terms of labor and capital productivity, it is quite difficult to identify the underground coal mines with the best final productivity. Total high-output mines do not always generate the highest sales revenue streams. In order to verify the H1 hypothesis in the introduction section, the paper was supplemented by determining the share of labor cost in the structure of coal production cost. Labor cost includes remuneration, employee benefits, compulsory contributions, and trade union fund. The results of the cost calculation are presented in Table 6.

292

D. D. Hai et al.

The research results do not allow to unequivocally confirm the H1 hypothesis that underground coal mines with a high share of labor costs in the total costs are characterized by low productivity. For example, Ha Long coal mine, which is worse in the statement, has the share of labor cost in total costs below 30%, while the better Vang Danh coal mine shows a high share of labor costs above 31%. However, it is worth noting that three of the four coal mines, which have the best productivity, are in the top ten of ranking with the lowest share of labor cost. Therefore, it can be said that the underground coal mine having a lower share of labor cost in the total costs is often consisting of higher productivity.

5 Safety and Productivity in Underground Coal Mines Underground coal mining is a dangerous industry with many natural hazards and challenging work conditions. Therefore, the analysis and improvement of safety levels in underground coal mines is always a top priority [12]. Increasing productivity must also ensure safety in the workplace. Individual costs must be incurred related to the monitoring of threats existing in a given deposit with the specific geotechnical conditions and their control (prevention against hazards). In underground coal mines, the additional costs may also be generated, related to any restrictions in the scope or pace of works resulting from specific rigors of their performance in hazardous conditions, primarily methane. All these aspects should be taken into account at the opening stage of the coal mining project. The schematical problem is illustrated in Fig. 3. An underground coal mine is the totality of the underground workings and surface plant equipped for extracting coal. The production process in underground mines is conducted in order to obtain commercial coal for sale with appropriate quality parameters, which are required by individual consumptions [9]. The underground coal mine can be treated as a technical system in which many elementary processes are carried out, adequately organized, and focused on efficient and stable production of coal. The partial processes of coal production are divided into: • Preparatory processes: making the available deposits for mining (openings - access to the coal deposits, preparing the seams for exploitation by carrying out cutting works); • Basic processes: mining coal in longwalls, processing; • Auxiliary processes: ventilation, transport, ensuring work safety; • Accompanying processes: protection of the natural environment such as sewage treatment plants, sedimentation tanks, reclaiming land after mining operations. Besides, it must be noted that ensuring the safety, health, and welfare of the employee is the primary purpose of coal mines. In Vietnam’s coal industry, underground mining is accompanied by all typical natural hazards, and the intensity of their occurrence increases annually. These are the following hazards: methane, water, gas and rock outbursts, fire hazards. In order to ensure safety during all works carried out in underground coal mines, natural hazards that may occur, there must be identified and fought. Therefore, the process of mining production is selected with the optimal mining technology as

Increasing Productivity and Safety in Mining

293

Fig. 3. Costs connected with the occurrence and combating natural hazards and their impact on the profitability of exploitation [9]

well as maintaining workplace safety in certain conditions. In addition to the costs of mining, coal mines also incur the costs of safety solutions. All these costs affect the selling price of commercial coal. On the other hand, a coal mine would like to achieve mining efficiency when the coal production cost has to be lower than the selling price. As a consequence, the proposed solutions should improve the surrounding conditions of coal companies that affect the obtained indicators of productivity. In order to verify the H2 hypothesis, which stated that underground coal mines with high intensity of natural hazards are characterized by low productivity, an assessment of the scale of natural hazards was carried out according to their categorizations, which were presented in Sect. 4.1. The actual coal mining in the Quang Ninh coal basin shows that the causes of fatal accidents in underground mines in the period of 1997–2019 are methane gas and water inrush, as illustrated in Fig. 4. Although the frequency of occurrence of methane explosion is not many, the level of death rates is so high. For example, in 1999 and 2002, there are only two accidents occurred per year with 19 and 13 victims, respectively. In addition, there are 11 victims in 2008, even though only in one accident. Besides, almost every year there are from 1 to 2 the water inrush accidents occurred with the number of victims from 1 to 3 in most of the underground mines in Quang Ninh coalfield. Some coal mines such as Mao Khe, Hon Gai, Ha Long, Mong Duong and Duong Huy have the serious perilous danger of water inrush. Furthermore, some mining areas such as Khe

294

D. D. Hai et al.

Chuoi, Hong Thai, Thong Nhat, Ha Lam, etc. have shown the risk of spontaneous coal combustion. Although the endogenous fires do not cause fatal accidents, they cause a lot of damage, interrupting production and affecting the mining plans of underground coal mines.

Fig. 4. Number of accidents at the underground mines in Quang Ninh coalfield from 1997 to now [13]

In assessing the level of safety, commonly recognized categories of natural hazards occurring in coal mines were used. The natural hazards in underground coal mines in Vietnam were analyzed, such as methane, water inrush, and coal spontaneous combustion potential. The research results are shown in Table 7. According to the data contained in Table 7, the tested underground coal mines are characterized by high intensity of natural hazards, which is typical for coal mining, including primarily methane hazards. No less dangerous and often there is a water inrush. Two underground coal mines also have the highest level of coal spontaneous combustion. The combination of the above-mentioned risks has caused negative impacts on the level of safety, continuous mining process, and productivity in underground coal mines. In addition, it has created the highest workplace threat in many cases. Referring to the results in Table 7 to the productivity assessment, the following conclusions can be made: • Ha Lam coal mine is characterized by high labor and capital productivity. The methane hazard is not in the highest category, but there are also coal spontaneous combustion and water hazards;

Increasing Productivity and Safety in Mining

295

• In Khe Cham and Hon Gai coal mines, there are great results in terms of productivity, even though they have the risks of methane, coal spontaneous combustion, and water hazards; • Nam Mau coal mine has the lowest productivity, although it is characterized by the low intensity of natural hazards.

Table 7. Level of natural hazards in the tested underground coal mines in 2015 No.

Underground coal mine

Level of methane hazard

[14]

Level of coal spontaneous combustion potential [15]

Level of water inrush*

Hight water risk

1

Mao Khe

Super emitter

III

2

Nam Mau

I

I

3

Vang Danh

I

III

4

Uong Bi

I

II, III

5

Ha Lam

I

IV

6

Hon Gai

I

I, III

7

Quang Hanh

III

II, III

8

Thong Nhat

II

II, III

9

Ha Long

II

III

Hight water risk

10

Duong Huy

II

II, III

Hight water risk

11

Khe Cham

II, III

III

12

Mong Duong

II

III

Hight water risk

Hight water risk

* based on actual mining situation at the underground coal mine

The above analysis indicates that the H2 hypothesis, which confirms that mines with high intensity of natural hazards are characterized by low productivity, is not confirmed.

6 Technological Change and Productivity in Vietnam’s Underground Coal Mines With the goal of increasing labor productivity, improving working conditions, and ensuring maximum safety for workers, VINACOMIN has continuously focused on investing and promoting the application of new technologies and mining equipment to the production stages. At the same time, we were giving priority to research activities, improving potentials, and developing human resources. Up to now, VINACOMIN has been mastering many advanced technologies in the world, actively manufactured and localized many mining machines, equipment, and spare parts for production. Currently, most of the underground coal mines in Quang Ninh coalfield are applying the longwall system for mining coal seams. It is understood that a proper selection of

296

D. D. Hai et al.

the equipment synchronization in the longwall is one of the key elements for a successful operation of the mining technology. Underground mining companies have been struggling to find the most suitable equipment for the extraction of coal seams by longwall methods. Initially, most of the longwall faces were supported by wooden props, while coal was extracted from the face by the drilling and blasting method. Due to the improvement in coal production, productivity, and the standards of mine safety, VINACOMIN has applied hydraulic single-props since 1997 and semi-mechanized shields model XDY-1T2/LY since 1999. The success of these works has created a “technical revolution” to replace the wooden props hydraulically and improved the overall technical, economic parameters. Coal production in the longwall faces supported by hydraulic support had increased rapidly from 100,000 to 150,000 tonnes per year as compared with the production of 50,000–70,000 tonnes per year when the face was supported by wooden props [16]. Labour productivity in longwalls using the single hydraulic props increased to 2.5–5.37 tonnes per man-shift, an average of 3.0-3.5 tonnes per man-shift. Since 2006, VINACOMIN has continued to apply new hydraulic supports in the underground coal mines, for instance, self-moving hydraulic frame model ZH1600/16/24Z, model GK/1600/16/24/HTD and model ZH1800/16/24ZL. Production increased range from 140,000 to 250,000 tonnes per year, labor productivity from 5.0–7.0 tonnes per man-shift [16]. In addition, working conditions and safety in workplaces has been significantly improved. Mining began more environmentally friendly due to the elimination of wooden support. However, mining operations also required the amount of manual work, and this type of supports is only suitable for extraction by drilling and blasting. As a result, production is still limited. Therefore, the increase of coal mining production according to the VINACOMIN’s plan, which means an increase in the number of longwalls as well as the number of workers. This is hardly feasible in the current mining situation in Vietnam. This is the opening step for the renovation of underground mining technology, which is a premise for the introduction of mechanized equipment into underground mines in Quang Ninh coalfield. In 2005, Khe Cham coal company had applied the first mechanized mining technology in a similar condition, using shield support ZZ-3200/16/26 combined with the shearer MG150/375-W. The mechanized longwall has a hight output of 233–388 thousand tonnes/year, averaging 289 thousand tonnes/year; labor productivity from 9.9–11.4 tonnes/per man shift, an average of 10.3 tonnes per man shift [16]. The success of the longwall was also a premise for technological innovations in the next exploitation for medium thick seams in Duong Huy and Quang Hanh coal mine since 2015. Compared with the longwall supported by hydraulic support in the same conditions, the average output of the mechanized longwall is 2.3–3.9 times higher, and the average productivity is 2.5–3.0 times higher. Some mines have investigated the mechanized mining equipment for the extraction of thick and gentle slope seams. For example, in November 2007, Vang Danh coal mine applied the LTCC, supported by mechanized equipment,i.e. shield VINAALTA2.0/3.15, coal shearer, one face conveyor, and chain conveyor. As a result, the highest production reported approximately 450,000 tonnes/year. This LTCC system (supporting by VINAALTA-2.0/3.15 and using one face conveyor) was also used in Nam Mau coal

Increasing Productivity and Safety in Mining

297

mine in 2010. Besides, VINACOMIN has applied new LTCC system using double conveyor (front conveyor and rear conveyor) in 4 underground mines with 5 longwalls (two longwalls in Ha Lam coal mine with a capacity of 600,000 tonnes/year and 1,200,000 tonnes/year respectively; one longwall in Khe Cham coal mine with a capacity of 600,000 tonnes/year, one longwall with capacity 450,000 tonnes/year in Vang Danh coal mine (as shown in Fig. 5) and one longwall with capacity 600,000 tonnes/year since February 2020 in Nui Beo coal mine. Labour productivity is relatively high, with an average of 33.534.0 tonnes per man-shift and 4–5 times higher than longwall using semi-mechanized support [16].

Fig. 5. The mechanized longwall in Vang Danh coal mine

The mining system retreats along with the seam dip by coal plough has been applied in the condition of the high steep thin coal seams in Mao Khe and Hong Thai coal mine since 2008, supporting by shield support 2ANSHA. Labor capacity and productivity of coal mining mechanized longwall by self-acting shield support 2ANSHA (average mining capacity is 64,500 tonnes/year, labor productivity reached 5.6 tonnes per man-shift) were two times higher, the expense for preparation roadway was seven times lower, and technology loss (16%) decreased two times than sublevel mining technology, supported by a movable hydraulic beam in the same geological conditions [3]. In addition, since August 2015, Hong Thai coal company (now belong of Uong Bi coal mine) has applied the new diagonal longwall system using hydraulic support model ZRY at seam No.9B, level of +30/+95. By the end of 2015, technology has achieved relatively great results: average coal output of 400 tonnes/day, labor productivity reached 5.5–6.0 tonnes per man-shift (2–3 times higher compared with sublevel mining method), and coal loss only

298

D. D. Hai et al.

between 12.6–16.3% [3]. This technology has improved the level of safety and working conditions. More importantly, the application of the diagonal longwall system with the use of support ZRY has reduced the rate of roadway development per tonne of coal production, with only 16.7 m per 1000 tonnes as compared with average 30–40 m per 1000 tonnes in the sub-level mining method. Thereby, the experimental project in Hong Thai opened a new way to innovate the mining technology for steep seams. Up to now, VINACOMIN has six underground mines applying the mining technology with seven longwalls [16]. The process of mining technological change in the underground coal mines in Quang Ninh, as illustrated in Fig. 6.

Fig. 6. Mining technological change in the underground coal mines in Quang Ninh coal basin

Over the past twenty years [19], underground coal mines in Quang Ninh coalfield have been widely applied mechanized technologies. The results of practical application have shown the superiority in improving mining output, labor productivity as well as improving the safety level and working conditions in comparison with manual longwall, as illustrated in Fig. 7. The number of workers in the mechanized longwall is only 95 persons. However, there are 120–160 persons at the manual longwall. This means the number of workers will decrease from 1.5–2 times when the underground coal mine applies mechanized mining technology; moreover, the labor capacity of the mechanized longwall increases 1.5–2.5 [5]. In recent years, Vietnam’s coal industry has been undergoing restructuring processes, the primary goal of which is to improve productivity. Decisions of the Prime Minister of the Vietnamese Government on 07/02/2013 and 12/12/2017 approving the restructuring program for VINACOMIN - scope of the implementation in 2012–2015 (Stage 1) and 2017-2020 (Stage 2). VINACOMIN restructures its enterprises to improve the quality, productivity, and effectiveness of their operation. With that goal, from 2012 to now, VINACOMIN has merged several companies and reducing the number of workers. Specifically, the total number of employees of the coal industry has decreased from 123,263 in 2012 to 98,000 personnel in 2019 (decreasing about 20%). In addition, VINACOMIN has strongly implemented measures to manage costs, manage resources,

Increasing Productivity and Safety in Mining

299

Fig. 7. The accident rate and productivity from 2010 to 2019 in underground coal mines of Quang Ninh coalfield [17]

and effectively use existing resources, while the production output has increased steadily over the years. That is also the main reason for the increased productivity in recent years.

7 Sources for Improving Productivity and Work Safety for Underground Coal Mines in Quang Ninh Coalfield Undoubtedly, productivity and safety are keys of sustainable development of Vietnam’s coal industry. The last section of this paper presents some proposed solutions to improve these parameters, aimed at stabilizing the situation in coal companies and thus ensuring Vietnamese mining survival, and in the long run development opportunities. The sources of improvement of productivity and safety are shown in Table 8. The solutions are divided into sources of a technical nature, referring primarily to technical infrastructure and technologies used, as well as economic and organizational nature, including proposals for changes in the principles of operation of underground coal mines. In addition, the table also includes methods to ensure workplace safety in underground coal mines. As it results from the above considerations, the level of natural hazards in the tested mines is high. Therefore all activities in this area are considered to be particularly significant and conditioning underground mining. Considering the high level of threats, it is recommended to conduct research and implementation activities in the area of monitoring and combating them. Besides, the expansion of the methane drainage network of mines, as well as the expansion of air conditioning and air cooling networks, are also crucial in order to reduce the increasing climate hazards. In terms of other activities listed in Table 8, it can show that some solutions should be immediately implemented. These undoubtedly include, above all, making efforts to

300

D. D. Hai et al.

Table 8. Sources of improvement of productivity and safety in the Vietnam’s coal industry Technical sources

Economic and organizational sources

Development of the system for monitoring natural hazards and methods to combat them

Implementation of a modern cost management system focused on reducing production costs

Expansion of methane drainage systems for underground coal mines

Restructuring employment

Improving climatic conditions in underground coal mines

Restructuring of non-productive and financial assets

Improving productivity in terms of the basic technical parameters affecting mining costs

Limiting the share of remuneration in total costs. Introduction of an incentive system more related to economic effects

Modernization of existing technical infrastructure to increase the efficiency of the mining equipment

Organizational transformations, merging coal mines

Increasing the quality of coal for sale

Improving relations with consumers from the professional power industry

Expansion of the application of methane drainage system and using them for energy

Rationalization of energy resource management

Application development technologies focused on clean coal production

Implementation of economies of scale in transport and logistics

Acquiring new licenses to increase coal resources and increasing coal mine life

Acquiring additional sources of financing investment outlays

- Activities in the area of work safety

- Activities in the area of productivity

improve the technical parameters affecting mining costs, i.e., the mining level, seam, longwall, shaft. This can only be achieved by simplifying the spatial structure of underground coal mines. The modernization of the existing technical infrastructure will not only contribute to increase productivity and the efficiency of mining equipment but also improve the quality of the produced coal to ensure a good competitive position and a favorable relation between the price of raw materials and its quality. Among the important activities of long-term strategic, technological changes should be mentioned. For example, the expansion of the use of methane drainage systems in underground coal mines as well as using them for energy, the application of mining technological advances, which are focused on clean coal production. However, they will only be able to be done if mining enterprises improve the efficiency of operational management and raise funds for their implementation. As well as that, coal mines can acquire new licenses for mining to increase coal resources and increasing mine life, and thus provide enterprises with long-term development prospects. The analytical results in Sect. 6 show that the application of mechanized mining technology is the great way of Vietnam’s coal industry to improve the efficiency of operation and to ensure the target sustainable development. However, the application of mechanization in underground coal mines is still restricted. For example, the highest coal output from mechanized longwalls in 2019 reached 3.42 million tonnes, accounting for 11.7% of total underground coal output. The main reasons are as follows:

Increasing Productivity and Safety in Mining

301

• The geotechnical conditions of underground coal mines are relatively complicated. For example, the variation of thickness and dip seam is considerable. Moreover, there are many faults in the Quang Ninh coal basin, which divide the coal seams into areas with short direction and slope. Therefore, it is so difficult to apply mechanized mining technologies; • The huge investment cost of mechanized equipment: this makes managers of underground coal mines hesitate to apply them, moreover the fear of investment risk. Therefore, it is necessary to continue completing the mechanized technologies suitable with specific conditions at each underground coal mine for further development. The application of mechanization equipment in underground coal mining is essential in the current period due to the demand for increasing the coal output and productivity, solving difficulties on underground employees, improving the working condition, and working conditions to underground miners. In order to expand the mechanization application, it is necessary to have drastic guidance from VINACOMIN, the determination, and initiative from production units, associated responsibilities of consultant units. Furthermore, some specific solutions have been proposed as in [16]. In the mechanized longwalls, an essential factor is the useful working time of mining equipment. This means we have to use them more efficiently, as well as utilize them for a higher proportion of the available working time. The machines in the longwall, such as shield support, armored face conveyor, shearer, crusher, are so expensive, and due to this fact, it is necessary to take actions that maximize their use. There are still some problems in the mechanization application process, which make the output not achievable as designed. Through the process of monitoring the operation of mechanized longwall in the first four months of trial application at Nam Mau coal mine, it showed that the operation time of longwall (mining time) only accounted for 46% in the total time of production, the remainings 54% are incidents which make the production jammed. The time analysis result on longwall operation is summarized as illustrated in Fig. 8. From the actual experience, the main reasons affecting the efficiency of mechanization application work at underground coal mines in Quang Ninh coalfield are the work of shelf-control, the effective use and maintenance of the equipment, and the qualification level and technical knowledge of underground workers, which are still low. So the equipment maintenance work for periodically mechanized equipment is not done so well. To evaluate the effective working time of mechanized equipment, the Total Productive Maintenance strategy (TPM) evaluation method is commonly used. Thanks for applying this method, the managers of mining enterprises could propose solutions to improve the efficiency of the use of mining, and simultaneously manage employees better in all production stages. In practice, there are many methods and tools to evaluate the use of the mining machines, their failure frequency, and reliability. One such tool is the Overall Equipment Effectiveness (OEE) model, which is successfully applied to the analysis of the use of machines in other industries. The analysis of the working time of machines could be based on the Overall Equipment Effectiveness (OEE) model, as shown in Fig. 9: OEE = Availability ∗ Performance ∗ Quality

302

D. D. Hai et al.

Fig. 8. Coal mining time in the mechanized longwall [18]

Fig. 9. OEE equation [3]

The waveforms currents collected by each engine of the mining machines formed the basis of the analysis [19]. The actual real work time was determined in relation to the theoretical availability time. Without regard to service, the registration of current waveforms of the mining machine engines in a given time enables the identification of all kinds of interruptions associated with breakdowns and other types of downtime. It was also very crucial to record minor stoppages and breaks for which the crew was not able to provide reasons. The results should be an important source of information for maintenance services in mines. In the area of economic and organizational sources, priority is given to reduce unit costs of production. If this condition is not met, Vietnam’s coal industry turns out to be

Increasing Productivity and Safety in Mining

303

uncompetitive in relation to both imported coal and other energy sources. Then even the strategic position of coal in the Vietnamese energy balance is not able to guarantee the survival of coal mines. Activities in this area should be started with the development and effective implementation of a cost management system, allowing to precisely determine the unit costs of longwall before and during the mining operation. Only this approach will enable the actual assessment of the effectiveness and identification of cost reduction opportunities. Significant mining cost reduction opportunities are certainly also in the employment restructuring, meaning in this case, a gradual rationalization of employment, concentration of mines in order to maximize the use of the existing technical and human potential and implementation of an incentive system related to the effects of work. Among the significant measures to improve productivity, one should also mention the strengthening of relations with consumers from the commercial power industry, thereby ensuring a stable market for Vietnam’s coal industry. However, it will not be possible without offering them competitive coal prices, which in turn requires a reduction of unit production costs. Solutions of slightly lesser importance, due to their impact on the level of total costs, include activities for the rationalization of energy resource management and effectiveness in transport and logistics. Nevertheless, they should be undertaken in parallel with the implementation of the cost management system and employment restructuring, as their effects will also allow reducing the total production costs. Besides, the development and effective functioning of an enterprise depends mainly on the quality of management. A well-managed mining enterprise, even in rather unfavorable geotechnical parameters of deposits and rather disadvantageous market conditions (sales possibility, price of coal), can function effectively. Therefore, Vietnam’s coal industry should be studied and applied new management methods in production to improve productivity, reduce product costs but still ensure quality such as production management system “Just-in-Time”, method SMED (Single-digit Minute Exchange of Die), KAIZEN and 5S. This is clearly seen in the diagram of the integrated program for improving productivity and quality, as illustrated in Fig. 8. In this program, there are three phases related to the area of concentration of activities towards improving productivity and quality, development of employees’ awareness, reconstruction of their attitude to processes and changes taking place in the enterprise, and change of their mentality. This is a great way to improve the productivity of Vietnamese mining companies by using operational management tools in the natural conditions of underground coal mines. Integrated productivity and quality improvement program as illustrated in Fig. 10.

304

D. D. Hai et al.

Fig. 10. Integrated productivity and quality improvement program [3]

8 Conclusions The coal industry in the world has conducted research to measure productivity by new methods since the 1970s, such as Total Factor Productivity (TPM) or partial productivity. It has to be emphasized that total productivity most often considers all factors of production of an expenditure stream - technological level, skills of personnel, organization of an enterprise, changes in capital expenditure, relations in planning and organizing a production process - thus providing a full picture of an enterprise. Partial productivity focuses on the fragmentary scope. However, most publications in recent decades have shown that the term productivity in Vietnam’s coal industry is understood only as labor productivity. In this situation, there is a need for a modern understanding of productivity, not only in the aspect of labor productivity but also in the aspect of the efficiency of the use of materials and energy, as well as engaged fixed assets and working capital, both own and borrowed. In order to increase the coal output, labor productivity, as well as improving the safety level and working conditions, VINACOMIN has been orientated towards developing the application of the mechanization technologies in underground coal mines in Quang Ninh coalfield. The results of the practical application over the years have confirmed that technological advances are partly responsible for both improvements in the safety and productivity of Vietnam’s coal industry.

Increasing Productivity and Safety in Mining

305

Summarizing the results of the diagnosis of productivity and safety in the Vietnamese underground coal mines operating in the Quang Ninh coal basin, the following conclusions can be made: 1. The labor and capital productivity of Vietnamese underground coal mines are varied. 2. Coal mines with high overall productivity expressed in tonnes and tpu do not always achieve the best results in the area of capital productivity, measured by the value of sales revenues per person in total and underground. 3. The hypothesis, which states that mines with a high share of labor costs in the total cost structure is characterized by high productivity, is not clearly confirmed. 4. Underground coal mines are characterized by a high level of natural hazards typical for hard coal mining, including the most hazardous methane. However, the hypothesis that mines with high intensity of natural hazards is characterized by low productivity is not confirmed. In connection with the above, a detailed multi-faceted diagnosis of factors influencing the productivity of domestic mines should be made, oriented at its improvement and leveling of barriers hindering its increase. However, in the aspect of the functioning of the studied coal companies, one should strive to reduce the productivity spread in individual mines and better link economic results with the remuneration system. The most important activities to be emphasized for improving workplace safety, and productivity in underground coal mines include: • Conducting research and activities implemented in the area of improving methods of monitoring and combating natural hazards; • Development of mine demethanation network; • Reduction of unit production costs; • Rationalization of employment; • Acquiring funds for development activities; • Implementation of new production technologies: continue to research and apply more suitable mechanized mining technologies in underground coal mines. In addition, automation and computerization technologies in operation and supervision have also been applied thoroughly to modernize the production process, reduce manual labor directly, reduce costs, and improve productivity and business results; • For most coal mines, both in Vietnam and in the world, full use of the technical possibilities of the machines is one of the most critical factors that can determine their economic effectiveness. It is necessary to apply the appropriate methodology such as OEE based on reliable information to evaluate in an actual way the level of use of technical equipment in the mining enterprise. The mining industry in Vietnam, as well as the other countries in the world, is facing problems of productivity decline as mining depth increases. Facing this challenge, VINACOMIN has quickly launched a series of effective solutions such as restructuring the organization and management model improving cost management, human resource management, and applying scientific technology in mining processing. Over the past years, the above mentioned synchronous solutions have improved the productivity of

306

D. D. Hai et al.

mining enterprises. However, the solutions that the paper proposed above are possible as meaningful suggestions for businesses in the mining sector of Vietnam to increase productivity and product competitiveness in terms of price and quality in the market economy. Acknowledgments. We express our deep sense of gratitude to Professor Marian Turek for assistance with methodology, and Dr. Aleksandra Koteras in the Central Mining Institute (GIG) in Poland for comments that greatly improved the manuscript. This research was supported by Dr. Le Duc Nguyen in the Institute of Mining Science and Technology (IMSAT), Vietnam. We would like to thank our co-workers in the Underground Mining Technology Department of IMSAT for their assistance with the economic and technical data of underground coal mines in the Quang Ninh coal basin. Many people, especially our team members themselves, have made valuable comment suggestions on this paper, which gave us the inspiration to improve our assignment.

References 1. Tran, L., Vu, P.: Improving productivity in Vietnam mining industry. Vietnam’s Economy in 2018 and Prospects for 2019, pp. 450–462 (2018) 2. Prusek, S., Turek, M., Dubi´nski, J., Jonek-Kowalska, I.: Increasing productivity – a way to improve efficiency of operational management in hard coal mines. Arch. Min. Sci. 63(3), 567–581 (2018). https://doi.org/10.24425/123675 3. Kosieradzka, A.: Enterprise productivity management (2012) 4. Bradley, C., Sharpe, A.: A detailed analysis of the productivity performance of mining in Canada (2009) 5. Organisation for Economic Co-operation and Development: SourceOECD (online service), measuring productivity: measurement of aggregate and industry-level productivity growth: OECD manual. OECD (2001) 6. Topp, V., Soames, L., Parham, D., Bloch, H.: Productivity in the mining industry: measurement and interpretation. Productivity Commission (2008) 7. Stainer, A.I.: Productivity in longwall mining of hard coal. Loughborough University (1976) 8. Mishra, D.P., Sugla, M., Singha, P.: Productivity improvement in underground coal mines - a case study. J. Sustain. Min. 12(3), 48–53 (2013). https://doi.org/10.7424/jsm130306 9. Dubi´nski, J., Turek, M.: Increase of effectiveness and safety of work in mines as a chance for the operation and development of mining industry of hard coal in Poland. Przegl˛ad górniczy 70(4)(1097), 1–8 (2014) 10. Phuong, V.H.: Total factor productivity growth, technical progress & efficiency change in Vietnam coal industry - nonparametric approach. In: E3S Web of Conferences, vol. 35, March 2018. https://doi.org/10.1051/e3sconf/20183501009 11. Nguyen, V.B.: Research and proposed solutions to increase labor productivity and reducing coal price in VINACOMIN (2015) 12. Dubi´nski, J., Prusek, S., Turek, M.: Key tasks of science in improving effectiveness of hard coal production in Poland. Arch. Min. Sci. 62(3), 597–610 (2017). https://doi.org/10.1515/ amsc-2017-0043 13. VINACOMIN: Summary report on occupational accidents at underground coal mines, Hanoi (2019) 14. M. of I. and Trade: Ranking of the underground mine by methane (2015)

Increasing Productivity and Safety in Mining

307

15. Central Mining Institute: Assess the coal spontaneous combustion potential in underground coal mines in Quang Ninh (2018) 16. Duong, D.H., Dinh, V.C., Koteras, A., Do, V.H., Vu, B.T.: Development orientation of mechanized technology of underground mining in Quangninh coalfield, Vietnam. In: AIP Conference Proceedings, vol. 2209 (2020). https://doi.org/10.1063/5.0000005 17. VINACOMIN: Summary report on mining technical and technological work (2019) 18. Nguyen, A.T., Nhu, V.T., Dao, H.Q.: The trend of mechanization development for underground mining in the Quang Ninh coalfield (2012) 19. Brodny, J., Stecuła, K., Tutak, M.: Application of the TPM strategy to analyze the effectiveness of using a set of mining machines. Int. Multidiscip. Sci. GeoConference SGEM: Surv. Geol. Min. Ecol. Manag 2, 65–72 (2016)

Analytical Study on the Stability of Longwall Top Coal Caving Face Tien Dung Le(B) Department of Underground Mining, Hanoi University of Mining and Geology, Hanoi 100000, Vietnam [email protected]

Abstract. Longwall face instability is a critical geotechnical issue in the operation of longwall with considerable mining height and significant caving height. Of the commonly applied research techniques, analytical solutions can provide fundamental insight into key mechanics of face stability to decide if the more in-depth but costly analysis is to be implemented. The paper first presents a mechanical model of face spall in mechanized longwall top coal caving method in which the interaction between coal seam, shield support and roof strata is taken into consideration. An equation for estimation of face spall extent is then developed from the model, using field observations at two typical mechanized longwall top coal caving faces at Quang Ninh coalfield, Vietnam. The estimation demonstrates that the equation is applicable to the cases where the spall is mainly caused by the ruptures of the immediate roof and top coal beams. When the main roof ruptures, the roof pressure is rapidly increased that may cause shield damage and whole face collapse. The paper’s results provide mining engineers with a low-cost tool for quick evaluation of face instability risk in the preliminary stage of mine design. Keywords: Face stability · Spall extent · Mechanical model · Field observation

1 Introduction Face instability is always a critical geotechnical issue in the operation of longwall coal mining, especially when applying modern technologies such as extra cutting height and/or caving height. The issue may occur in small (spall or slab in centimeters) to a large extent (spall or fall in meters) that threatens worker’s safety, damages expensive equipment, and interrupt normal production. For example, a large face spall occurred at Seam 11 Ha Lam coal mine in Quarter III 2019 caused the face stopped in about two weeks, resulting in an economic loss of 140,000 USD per day. In order to control the problem, face stability mechanics and their driving parameters have been extensively studied for many years through various research techniques, such as field measurement, analytical solution, computational modeling, and physical modeling. Field measurement [1, 2] and physical modeling [3] studies can provide fundamental insight into instability mechanisms. However, they should be carefully performed due to high costs and © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 308–319, 2021. https://doi.org/10.1007/978-3-030-60839-2_16

Analytical Study on the Stability of Longwall Top Coal Caving Face

309

long design time. Computational modeling studies [4, 5], in contrast, can reveal micromechanisms of the instability at reasonable cost and time. To achieve reliable outputs, they require good modeling strategy and sound justification for many unknown input parameters. For cases where a fundamental insight into crucial mechanics of face stability is needed to decide if the more in-depth and but costly analysis is to be implemented, an analytical solution should be first considered as the main study tool. Although the solution requires simplifications of rock mass conditions and failure criterion, it can produce, satisfactory outcomes if a simplified mechanical model is adequately developed, and calculation inputs are reliably estimated. Analytical solution, as typically restricted to two-dimensional problems due to complexity, is well suited to analyzing longwall face-associated geomechanics since the face is long in strike direction and continuous in dip direction [6].

Fig. 1. Conceptual model of longwall face stability [7]. W is abutment loading; t is extraction height; l is tip to face distance; x is faced spall depth; σ h is horizontal stress; du/dt is the rate of shield convergence; P is shield loading.

A number of mechanical models of longwall face stability were developed in previous analytical studies. Frith [7] proposed a conceptual model that was composed of a vertically cleated coal seam, a horizontally bedded immediate roof loaded by overburden strata, and shield support (Fig. 1). Based on the theory of Euler buckling (in particular, critical buckling stress/load in a vertical column), the face spall depth x was found as a function of abutment stress σ v and cutting height t. Similarly, the roof fall potential ahead of shield support was controlled by the mining-induced horizontal stress σ h and unsupported span l + x between shield and face line. Wang, Yang, and Kong [8], based on practical observation of shear spall incidents, presented a shear failure model of a single pass longwall face (Fig. 2). Using the limit equilibrium method and Mohr-Coulomb strength theory, the authors assessed the face stability condition through a safety index, which was the ratio of shear stress to shear strength applied on the coal spalling block. The involved parameters were roof load, the gravity of spalling block, shield guard force, coal seam strength parameters, and heights of cutting and spalling coal. The index was further improved in Kong, Cheng, and Zheng [9] in which the friction f between coal

310

T. D. Le

body and the immediate roof was taken into consideration, as shown in Fig. 3. It is also seen that the ruptured main roof was explicitly incorporated into the analysis in the figure.

Fig. 2. Shear failure model of face spall in single pass longwall [8]. a, b and c are corners of failure block; G is the weight of failure block; L is the length of failure plane, and S is shear strength.

Fig. 3. System of the face–support–roof (left) and its simplification (right) [9]. Q is overlying strata pressure; Qo is acting force on face; P is shield resistance force; G is the gravity of sliding body; S and N are shear and regular forces on sliding surface; T is the shear force on failure face; f is friction between coal and roof, and β is failure angle.

Another approach in the analytical solution for face stability is to consider the equilibrium of face–support–roof system in terms of energy or moment. For longwall top, coal caving (LTCC) face, Wang and Wang [10] applied energy conservation principle to transform the kinetic energy of broken rock block into intact coal, broken top coal and face support in the form of dynamic load, plus surface energy and heat energy of fractures (Fig. 4a–b). The theory of elastic foundation beam was then used to calculate

Analytical Study on the Stability of Longwall Top Coal Caving Face

311

the roof pressure acting on coal wall and shield support. This pressure was finally compared to the strength of both coal wall and shield support to assess the system’s stability. For single-pass large longwall mining (SPLL), the approach was similar, but no energy transformation was considered due to the absence of the top coal section (Fig. 4c–d). In both cases, strength and stiffness of coal seam, as well as stiffnesses of roof strata and shield support, were found as key driving factors of face stability. The energy conservation principle was also applied by Guo, Liu, Dong, and Lv [11] for single-pass longwall in where the main roof formed a voussoiring beam (Fig. 5a). The authors stated that the work done by rotation of primary and immediate roofs must be equal to the deformation energy stored in the immediate roof, coal wall, and shield support. The authors added another model in where the main roof formed cantilever beam (Fig. 5b). In both models, at the instant when the main roof ruptured, the sum of moments of the system about rotating point O should be zero.

Fig. 4. System of face–support–roof in (a–b) LTCC and (c–d) SPLL faces [10]. L 1 and L 2 are distances from support to roof breaking point and coal wall; Q is coal wall force; P is shield force, and q1 is the gravity of overburden.

Previous studies successfully developed several mechanical models of face stability from which key controlling parameters were identified; laws of parameters’ impact on stability were analyzed; spall/fall extent was estimated, and criteria for assessing face stability were proposed. These analytical models were mainly based on the theory of force/moment balance (static solution) or energy conservation principle (dynamic solution). For LTCC face, the models were, however, complicated in use due to the complex calculation of kinetic/potential energy of ruptured/caved top coal and roof strata. Therefore, this paper aims at establishing a simple mechanical model for the fundamental understanding of face stability in the longwall top coal caving method. The paper’s results provide mining engineers with a low-cost tool for quick evaluation of face instability risk in the preliminary stage of mine design.

312

T. D. Le

Fig. 5. System of face–support–roof in (a) voussoir beam-formed main roof and (b) cantilever beam-formed main roof [11]. B is the depth of rib spalling; L is breaking span of the main roof; O is origin point, and θ is the rotation angle of the main roof.

2 Field Observation at Quang Ninh Mechanised Faces Two mechanized longwall top coal caving faces at Quang Ninh coalfield in Vietnam are taken as real databases for development and comparison of face stability mechanical model. They are Face 11-1.14 at Ha Lam coal mine and Face I-8-1 at Vang Danh coal mine. A field observation at Face I-8-1 was presented in Le and Dao [12] in which the periodic intervals of the immediate roof and main roof were reported at 10 and 90 m, respectively. According to the mining engineer at the site [13], minor face spall occurred more frequently in less than 0.5 m deep, while major spall was less frequent but more severe - the spall extended 1–2 m into coal face and 5–7 m above seam floor. For Face 11-1.14 Ha Lam coal mine, general geotechnical conditions can be found in Le, Bui, Pham, Vu, and Dao [14]. It was reported by on-site engineers that the periodic caving interval of the immediate roof was 10 m, while no information on primary roof behavior was revealed. Face spall was observed up to 3.15 m deep into coal face and several meters into the top coal section. To better understand the face behavior at the site, field measurement of shield leg pressure was implemented for the current study. Front and rear legs of three shield support near tailgate (Shield #1), near the main gate (Shield #71), and at mid-panel width (Shield #35) were recorded and displayed in Fig. 6(a– c). The following observations were made from the figure as follows. Firstly, at the mid-panel width, the leg pressure mostly ranged from 15 to 25 MPa, while near two gate roads, the leg pressure mostly ranged from 10 to 20 MPa. The difference in the range was because near two gate roads, and the installed shields had the higher working capacity, while the area was additionally supported by supplemental hydraulic props. At the monitoring locations, the front leg pressure was from time to time less than the rear

Analytical Study on the Stability of Longwall Top Coal Caving Face

313

Fig. 6. Pressure in front leg and rear leg of (a) Shield #1 near tailgate; (b) Shield #71 near maingate; and (c) Shield #35 at mid-panel width

314

T. D. Le

leg pressure. This phenomenon, however, was not seen in Face I-8-1 [12] and other stable coal faces [15, 16]. The reason was mainly the face spall and roof fall ahead of shield support that ruins the tight contact between shield roof canopy and top coal section. The comparison confirms that the face instability at Ha Lam coal face was more significant compared to that at Vang Danh coal face.

3 Mechanical Model of Seam–Support–Roof Interaction 3.1 Mechanical Model Excluding large geological structures which are beyond the scope of this study, face spall in Quang Ninh LTCC faces occurs typically in two typical conditions: (a) coal seam is weal/relatively weak, highly jointed and easy to cave (e.g., Ha Lam coal mine); and (b) coal seam is moderate strength and immediate and main roofs cave/rupture in larger intervals (e.g., Vang Danh coal mine). From these conditions, a model (system) representing the mechanical interaction between coal seam, shield support, and roof strata is established, as illustrated in Fig. 7. In this figure, top coal is the coal seam section above shield support and is allowed to cave; the immediate roof is the rock immediately above top coal and caves as top coal caves; and the main roof is the rock above the immediate roof, which ruptures but does not cave, and can still transmit horizontal force [17]. The system represents a static equilibrium immediately before the face spall. That is, both algebraic sum of all forces and algebraic sum of all turning/bending moments in the system must be equal to zero. It is assumed that the depth of face spall is D and associated origin point is O. The interaction system shows that the load supported by coal wall and shield support consists of weights of top coal W tc , immediate roof W imr and main roof W mr . The top coal component L tc covers the lengths of the coal wall, tip-to-face distance L d, and roof canopy L s . The immediate component is the cantilever beam, which caves at the same time or with a little delay after top coal caving. It is noted that top coal recovery leads to the greater mined-out height that consequently creates sufficient space for the immediate roof to collapse completely. The length of the immediate roof part L imr is dependent on its structure as well as for overburden load. The main roof component is referred to as the periodic rupture interval L mr resting on the immediate roof. The broken but un-caved main roofs can still transmit some horizontal forces, and their weight acts on lower strata and caved rock pile. Below the loading components and immediately before face spall, shield support increases to its maximum working capacity Ps while the coal wall reaches its maximum bearing capacity Pc . The shield applies a guard force Pg on the coal wall and, in turn, receives an opposite force Pg’ from the wall. Note that the weight and force in the current model are corresponding to 1 m in the out-of-plane direction.

Analytical Study on the Stability of Longwall Top Coal Caving Face

315

Fig. 7. Mechanical model of the seam–support–roof interaction for LTCC. H mr , H imr , H tc, and H c are heights of the main roof, immediate roof, top coal and cutting coal; W mr , W imr, and W tc are weights of the main roof, immediate roof, and top coal; Ps , Pc , Pg and Pg’ are the maximum working capacity of shield support, maximum bearing capacity of coal wall, guard force of shield and opposite force of coal wall; L mr , L imr, and L tc are rupture/caving lengths of the main roof, immediate roof, and top coal; O is origin point and D is the depth of face spall.

3.2 Extent of Face Spall As mentioned earlier, at static equilibrium, the mechanical interaction system in Fig. 7 satisfies two conditions at the same time: the sum of all forces and the sum of all moments must be equal to zero. Hence: Wmr + Wimr + Wtc = Pc + Ps Lmr Limr Ltc D Lmr Hmr γmr + Limr Himr γimr + Ltc Htc γc = Pc + 2 2 2 2

(1) 

 Ls + Ld + D Ps 2 (2)

Where γ mr , γ imr and γ c are unit weights of primary roof rock, immediate roof rock, and coal, respectively; H mr , H imr, and H c are heights of main roof strata, immediate roof strata, and top coal beams, respectively. If the intervals of immediate roof caving and main roof rupture are known, the depth of face spall can be estimated from Eq. (2) as follows:   Lmr Limr Ltc Ls + 2Ld Pc Lmr Hmr γmr + Limr Himr γimr + Ltc Htc γc − Ps = D + Ps 2 2 2 2 2 (3)

316

T. D. Le

D=

L2mr Hmr γmr + L2imr Himr γimr + L2tc Htc γc − (Ls + 2Ld )Ps Pc + 2Ps

(4)

Substituting Eq. (1) into Eq. (4) gives: D=

L2mr Hmr γmr + L2imr Himr γimr + L2tc Htc γc − (Ls + 2Ld )Ps Lmr Hmr γmr + Limr Himr γimr + Ltc Htc γc + Ps

(5)

If coal face spalls in conformity with Mohr-Coulomb failure criterion, the angle, and height of the spall can also be estimated. As depicted in Fig. 8, the angle of failure surface compared to vertical direction is α, height of spall is H, weight of spalling block is G, load acting on spalling block is W, internal friction angle of coal is ϕ, and cohesion strength of coal is c. Immediately before the spall, shear resistance force R is equal to shear force S, which means: R=S

(6)

  D = (W + G)cosα − Pg sinα (W + G)sinα + Pg cosα tanϕ + c sinα

(7)

Fig. 8. Coalface spall in conformity with Mohr-Coulomb failure criterion. D is the depth of face spall; G is the weight of spalling block; H is the height of face spall; H c is the height of the coal face; Pg is the guard force of shield; W is acting load; and α is the angle of spall

Since D2 γc , W = Lmr Hmr γmr + Limr Himr γimr + Ltc Htc γc − Ps and G = 21 HDγc = 2tanα then      D 2 γc D 2 γc D W+ sinα + Pg cosα tanϕ + c cosα − Pg sinα (8) = W+ 2tanα sinα 2tanα

Analytical Study on the Stability of Longwall Top Coal Caving Face

317

or      HDγc D2 HDγc HD W+ tanϕ + Pg 2 + cD = W + tanϕ − P g 2 2 2 D +H 2 D + H2 (9) The angle of failure surface and height of the spall can be calculated from Eq. (8) or (9). 3.3 Calculation Result Based on the field observations presented in the previous section, the depth of face spall at the two LTCC faces is calculated by using Eq. (5) for comparison purposes. It is noted that coal seam, roof strata, and shield characteristics are adopted from site and consultant reports [18, 19]. In the first round of calculation, practical main roof interval was assigned to the equation; however, the resultant depth was very unrealistic. This means that Eq. (5) is capable of estimating face spall depth when the spall is mainly caused by the ruptures of the immediate roof and top coal beams. When the main roof ruptures, it increases roof pressure rapidly that can break shield support and destroys the whole coal wall. Hence, Eq. (5) is rewritten, as shown in Eq. (10). The final calculation results are shown in Tables 1 and 2. It can be seen that the calculated depth at Face 11-1.14 falls within the range of the depth reported from the site. This estimated value is about 1.8 times less than the maximum depth recorded. In contrast, the calculated value at Face I-8-1 is 1.8 times greater than the maximum depth recorded in the field. The difference in results can be attributed mainly to the role of strata strength in the calculation. That is, although the developed solution cannot explicitly take into account coal/rock material strength nor small geological structures in strata (e.g., joints) that impact strata stability, it does consider an overall effect of strata strength through the rupture lengths and heights of the immediate roof and top coal beams. In practice, the rupture lengths of the immediate roof and top coal and the shield characteristics are the same for the two faces, as reported in Tables 1 and 2. However, the height of the immediate roof in Face 11-1.14 is 1.7 times less than that in Face I-8-1, while the height of top coal in the first face is 2.5 times greater than that in the second face. The difference here makes the estimated value in Ha Lam coal face less than that in Vang Danh coal face. Furthermore, as earlier stated in this paper, Ha Lam coal seams are weaker than Vang Danh coal seams. This explains why the calculation underestimates the spall depth at Ha Lam coal face while it overestimates the instability at Vang Danh coal face. D=

L2imr Himr γimr + L2tc Htc γc − (Ls + 2Ld )Ps . Limr Himr γimr + Ltc Htc γc + Ps

(10)

318

T. D. Le Table 1. Calculation of face spall depth at Face 11-1.14 Ha Lam coal mine

Parameter

Value

Parameter

Value

Estimated value

Field value

Lmr

0 (m)

Ld

0.63 (m)

D = 1.72 (m)

D ≤ 3.15 (m)

Hmr

16.72 (m)

Ls

3.75 (m)

γmr

26000 (N/m3 )

Ps

2.93E6 (N)

Limr

10 (m)

Pg

0.15E6 (N)

Himr

8.48 (m)

γimr

26000 (N/m3 )

Ltc

4.38 (m)

Htc

8.39 (m)

γtc

15000 (N/m3 )

Table 2. Calculation of face spall depth at Face I-8-1 Vang Danh coal mine Parameter

Value

Parameter

Value

Estimated value

Field value

Lmr

0 (m)

Ld

0.63 (m)

D = 3.55 (m)

D ≤ 2.00 (m)

Hmr

11.3 (m)

Ls

3.75 (m)

γmr

26600 (N/m3 )

Ps

2.93E6 (N)

Limr

10 (m)

Pg

0.15E6 (N)

Himr

14.5 (m)

γimr

26600 (N/m3 )

Ltc

4.38 (m)

Htc

3.41 (m)

γtc

16700 (N/m3 )

4 Conclusions This paper first presents a mechanical model of face spall in mechanized longwall top coal caving method in which the interaction between coal seam, shield support, and roof strata has been taken into consideration. An equation for estimation of face spall extent has been then developed from the model, using the field observations at Face 11-1.14 Ha Lam coal mine and Face I-8-1 Vang Danh coal mine. The estimation demonstrates that the equation applies to cases where the spall is mainly caused by the ruptures of the immediate roof and top coal beams. When the main roof ruptures, the roof pressure is rapidly increased and may subsequently cause shield damage and whole face collapse. Although the equation may either underestimate or overestimate the instability due to the lack of impact of strata strength, it provides mining engineers operating longwall top coal caving face with a low-cost tool for quick evaluation of face instability risk in the

Analytical Study on the Stability of Longwall Top Coal Caving Face

319

preliminary stage of mine design. A more detailed investigation of the impact of strata strength on face stability is being implemented by the author using numerical modeling. Acknowledgments. This research is funded by Hanoi University of Mining and Geology (HUMG) through Project T20-14.

References 1. Frith, R.C.: A holistic examination of the load rating design of longwall shields after more than half a century of mechanised longwall mining. Int. J. Min. Sci. Technol. 25, 687–706 (2015) 2. Prusek, S., Rajwa, S., Wrana, A., Krzemie´n, A.: Assessment of roof fall risk in longwall coal mines. Int. J. Min. Reclam. Environ. 31, 558–574 (2017) 3. Yang, S., Song, G., Kong, D.: An evaluation of longwall face stability in thick coal seams through a basic understanding of shield–strata interaction. J. Geophys. Eng. 16, 125–135 (2019) 4. Bai, Q.-S., Tu, S.-H., Chen, M., Zhang, C.: Numerical modeling of coal wall spall in a longwall face. Int. J. Rock Mech. Min. Sci. 88, 242–253 (2016) 5. Yao, Q., Li, X., Sun, B., Ju, M., Chen, T., Zhou, J., Liang, S., Qu, Q.: Numerical investigation of the effects of coal seam dip angle on coal wall stability. Int. J. Rock Mech. Min. Sci. 100, 298–309 (2017) 6. Galvin, J.M.: Ground Engineering - Principles and Practices for Underground Coal Mining. Springer, Cham (2016) 7. Frith, R.: Half a career trying to understand why the roof along the longwall face falls in from time to time? In: 24th International Conference on Ground Control in Mining, pp. 33–43. West Virginia University (2005) 8. Wang, J., Yang, S., Kong, D.: Failure mechanism and control technology of longwall coalface in large-cutting-height mining method. Int. J. Min. Sci. Technol. 26, 111–118 (2016) 9. Kong, D.-Z., Cheng, Z.-B., Zheng, S.-S.: Study on the failure mechanism and stability control measures in a large-cutting-height coal mining face with a deep-buried seam. Bull. Eng. Geol. Environ. 78(8), 6143–6157 (2019) 10. Wang, J., Wang, Z.: Systematic principles of surrounding rock control in longwall mining within thick coal seams. Int. J. Min. Sci. Technol. 29, 65–71 (2019) 11. Guo, W., Liu, C., Dong, G., Lv, W.: Analytical study to estimate rib spalling extent and support requirements in thick seam mining. Arab. J. Geosci. 12 (2019). Article number: 276 12. Le, T.D., Dao, H.Q.: Field investigation of face spall in moderate strength coal seam at Vang Danh coal mine, Vietnam. VNU J. Sci. Earth Environ. Sci. (2020). (Accepted) 13. Do, X.H.: Personal communication. Vang Danh Coal Company (2020) 14. Le, T.D., Bui, M.T., Pham, D.H., Vu, T.T., Dao, V.C.: A modelling technique for top coal fall ahead of face support in mechanised longwall using discrete element method. J. Min. Earth Sci. 59, 56–65 (2018) 15. Verma, A.K., Deb, D.: Analysis of chock shield pressure using finite element method and face stability index. Min. Technol. 116, 67–78 (2007) 16. Yun, D., Liu, Z., Cheng, W., Fan, Z., Wang, D., Zhang, Y.: Monitoring strata behavior due to multi-slicing top coal caving longwall mining in steeply dipping extra thick coal seam. Int. J. Min. Sci. Technol. 27, 179–184 (2017) 17. Peng, S.S.: Coal mine ground control. West Virginia University, Morgantown (2008) 18. Ha Lam Coal Company: Mining passport for Face 11-1.14. Ha Lam Coal Company (2015) 19. Vinacomin Institute of Mining Science and Technology: Investment and mining project for level 0: 175 Vang Danh site, Vang Danh coal mine (2016)

Recycling Ash and Slag of the Thermal Power Plant to Replace Protective Pillars in Mao Khe Coal Mine, Vietnam Phi-Hung Nguyen1(B) , Vladimir Ivanovich Golik2 , Manh-Tung Bui1 , Thai-Tien-Dung Vu1 , and Van-Chi Dao1 1 Department of Underground Mining, Faculty of Mining, Hanoi University of Mining and

Geology, Hanoi, Vietnam [email protected] 2 Scientific Centre, North Caucasus State Technological University, 362000 Vladikavkaz, Russia

Abstract. The underground mining method is the activity that leads to the surface impacts caused by subsidence movements. The mining Vietnam industry is developing, and step by step, applying advanced technologies to practical production in order to improve mining efficiency. Because of safety issues, optimal exploitation of mineral reserves, and reduction of losses are always the top priorities in mining companies of Vietnam. The paper assesses the ability to utilize the sources of ash and slag from the thermal power plant as backfilling material for underground mining activities in the Mao Khe coal mine, Vietnam. This technological solution helps to release a large amount of ash and slag discharged from the plants and minimizes negative impacts on the environment. Besides, the use of backfill methods also increases the stability of the safety bank and decrease coal losses during the exploitation process in underground coal mines in Vietnam. In this study, the authors performed theoretical and experimental researches with the hydraulic transport method using the materials with various ratios of water and ash in the field. The results were recorded and statistically analyzed to select the most suitable materials that are applicable at the southern wing area of The Mao Khe coal mine. Keywords: Ash and slag · Backfilling · Safety pillar · Natural resource recovery · Loss reduction

1 Introduction Backfilling in underground coal mines in the world has been practiced for more than 100 years, and evidence anticipates the application of mine fill technology at an increasing rate during this decade. The evolution of backfill technology is closely related to the establishment of new mining methods. Hydraulically transported tailings and alluvial fills were introduced in the 1950s, thus permitting the adoption of cut and fill mining where the backfill was used as a working floor. In the early 1960s cemented hydraulic © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 320–343, 2021. https://doi.org/10.1007/978-3-030-60839-2_17

Recycling Ash and Slag of the Thermal Power Plant

321

backfill was introduced, followed by the adoption of undercut and vertical retreat mining methods. The 1990s has been considered by many as the decade of high-density tailings fills and paste fills, during which several mines have successfully introduced paste fill into their operations [1]. Paste fill, relatively new technology in mining, is gaining importance because of its many perceived advantages. The reasons behind the conversion to paste fill include economic, environmental, geotechnical, and safe improvements. The use of backfill in underground mine openings and workings is increasing due to the need for systematic selection [2]. Backfill is defined as the material or materials that are utilized in void openings of underground mines for mining technical or mining safety purposes. Backfill is applied to prevent fires and explosions, to improve mine ventilation, to improve the stability of the rock, to reduce subsidence effects at the surface, as well as for economic and environmental factors. Materials are mainly from the mining industry and mining-related, such as fly ash, Flue Gas Desulfurization (FGD), gypsum, slag, infertile overburden, tailings, filter dust, residues from mineral processing, etc., and other industries (e.g., incineration ash, used building material, old bricks, used foundry sand, furnace blow-out, etc.). Mining with backfill technology helps mining companies achieve many of these goals. The backfilling technology enables a wide range of engineering solutions for particular mine sites and their unique sets of problems and opportunities. William Ross Wayment was well known to backfill excavated or mined out regions of an underground mine by transporting a slurry of sands or mill tailings having 40 to 70% solids, by weight, to the stope areas to be filled [3]. Further, backfill is not utilizable in the mining operations until it has been proved about consistency and strength to permit men and machines to support safely on the surface of the fill. Dimitre Antonov has presented mine backfill design concepts and procedures, physical and mechanical properties of backfill, and their measurement [4]. Concerning the backfilled stope stability, some new methods for minimizing amount of cement used in backfilled stopes have been presented. Manoon Masniyom investigated backfill materials and techniques suited for systematic selection and application of backfill in underground mines [5]. Laboratory tests were carried out on the physical, chemical, and mechanical properties of different backfill materials and mixtures, therefore. Special attention was paid to materials generated as by-products and other cheaply available materials, e.g., fly ash and FGD-gypsum from power plants, natural and synthetic anhydrite. The system selection, application, and placement of backfill in underground mines are truly multidisciplinary processes. One of the investigated material mixtures can be used as a technically and economically viable backfill for underground mines, is the crushed backfill material [5]. In recent years, in Vietnam, some coal mines are in the studying process to convert the backfilling form, especially the possible utilization of entire waste rocks and fly ashes as components of paste fill for environmental benefits [6]. Some studies have indicated that the backfill can enhance the support potential in underground mining operations [7, 8]. The tests at the underground coal mines in Cam Pha and Ha Long areas, Quangninh province in Vietnam, have had many positive results in decreasing the surface subsidence and increasing the coal recovery from the protective pillars in these mines.

322

P.-H. Nguyen et al.

According to the geological documents, down to the level of −400 m, the coal reserve at the southern wing area of the Mao Khe coal mine, under the protected surface works is relatively significant, about 23.2 million tonnes, which accounts for 33.2% of the total capacity of the whole region [9]. This reserve is mainly concentrated in the seams, namely V9A, V9, V9B, V8, and under constructions such as residential areas, power lines, streams and reservoirs, open pits, old underground. The coal capacity by the current mining technology of roof rock control will cause the surface displacement deformations, which can destroy civil and industrial constructions, and increase the water output or the risk of water burst from water-filled objects in old underground tunnels. Currently, to ensure the safety in mining activities, the Mao Khe coal mine has applied the technical solution of leaving the coal reserve as the safety banks to keep the stability of the surface works and mining facilities nearby the area. This solution can meet the protected requirements of surface works but wastes a large volume of coal resources. Thus, it is necessary to study and apply reasonable mining technology to exploit the reserves maximally. At the same time, the surface works are still protected as well as the mining operations must be safe. The mining technology of controlling a collapsed rock wall by insertion is a technological solution that ensures technical requirements. Meanwhile, the Mao Khe Thermal Power Plant invested by Vietnam National Coal and Mineral Industries Holding Corporation Ltd. (VINACOMIN), has been officially put into commercial operation since January 18, 2013. Two thermal generation sets operate by using two circulating fluidized bed boilers CFB, burning with coal dust 6B. After seven years of operation, the plant has produced over 20 billion kWh of electricity for the national electricity system. Annually, it consumes about 1.6 million tonnes of coal from Dong Trieu - Uong Bi areas and emits about 650 thousand tonnes of ash and slag, of which bottom ash is from 30–40%, and fly ash is from 60–70% [10]. According to the analysis, the ash and slag from the Mao Khe Thermal Power Plant have no hazardous components so that they can be used for the production of unburnt construction materials and as cement additives. However, in recent years, the ash and slag from the plant have been partly consumed as additives for manufacturing cement and unburnt construction materials (accounting for nearly 17% of the total amount of ash and slag discharged by the plant). The rest cannot be consumed and still have to be transported to the waste dump of the plant. Hence, environmental issues, waste dump, and treatment costs have cared as significant problems. This study aims to solve two simultaneous problems: • The thorough exploitation of coal resources under the surface works that need to be protected; • Releasing ash and slag from the thermal power plants by thoroughly using them as materials to fill in the mined areas in the underground mines to protect surface works.

2 Introduction of the Coal Reserve at the Southern Wing Area of the Mao Khe Coal Mine Under the Protected Surface Works Based on the geological documents, the coal reserve at the southern wing area of the Mao Khe coal mine down to the level −400 m is 69,938.6 thousand tonnes [9]. In which,

Recycling Ash and Slag of the Thermal Power Plant

323

the reserve under the surface works that need to be protected, such as residential areas, power lines, reservoirs, and streams, is approximately 23,240.6 thousand tonnes, which accounts for 33.2% of the total coal capacity of the southern wing area to the depth of −400 m. This coal reserve is saved to use as the safety banks for the protected surface works at the southern wing area of the Mao Khe coal mine. Most of them are mostly distributed in the coal seams, namely V10, V9, V9a, V9b, V8, V8a, V7 and V6 [10]. The coal reserve for the safety banks mostly concentrates at the following seams: the 9 seam (6,729.1 thousand tonnes, accounts for 29% of the total coal reserves), the 9a seam (5,853.5 thousand tonnes, 25.2% of the total), the 9b seam (3452.1 thousand tonnes, 14.9% of the total), the 8 seam (3,038.9 thousand tonnes, 13.1% of the total), and the 10 seam (3,009.6 thousand tonnes, 12.9% of the total). The coal reserve of the remained seams accounts for 1.1–2.2% the total capacity of the safety bank under the protected surface works at the southern wing area of the Mao Khe coal mine (Fig. 1).

Fig. 1. Coal reserves for the safety bank at the southern wing area of the Mao Khe coal mine

The protected surface works at the southern wing of the Mao Khe coal mine is divided into four groups: (i) residential area, (ii) high voltage power lines (110 kV, 220 kV), (iii) streams, water reservoirs, and (iv) pit mine and old tunnels [9]. In which, the coal reserve under the area of high voltage power lines is the most significant, about 13,788.6 thousand tonnes (59.3% of the total) (Fig. 2). According to the range of thickness and slope angle, the coal reserve in the safety bank area is mainly distributed at the seams having slope angle higher than 4.5°. The coal reserve of these seams is 20,947.3 thousand tonnes, which accounts for 90.1% of the total (Table 1). Thus, the current coal reserve is maintained to protect the surface works at the southern wing area of the Mao Khe coal mine is over 23 million tonnes. With the current production is nearly 2 million tonnes/year, the coal reserve of this area is sufficient for the Mao Khe coal mine to exploit more than ten years. Therefore, it is necessary to find reasonably technical and technological solutions to exploit effectively and safely this reserve [7, 8].

324

P.-H. Nguyen et al.

Fig. 2. Coal reserve for the safety bank at the southern wing area of the Mao Khe coal mine by the protected surface works

Table 1. Coal reserve in the safety bank at the southern wing area of the Mao Khe coal mine by the thickness and slope angle of coal seam Coal reserve

Thickness of seam (m) ≤3.5

Slope angle of seam (degree)

3.5 ÷ 5.0

Total >5.0

≤45

489.2

200.2

1,603.9

2,293.3

>45

7,492.3

5,663.5

7,791.5

20,947.3

7,981.5

5,863.7

9,395.4

23,240.6

Total

3 Determination of Safe Mining Depth Under the Protected Surface Works The safe mining depth is calculated based on the protected surface works. Each type of protected works is characterized by the permissible displacement parameters. They can be determined as follows: 3.1 Safe Mining Depth Under Residential Areas For residential areas, the determination of the safe mining depth under the different conditions that the protected surface works are houses or civil constructions, and in the case of a coal seams, needs to take into account the slippage of soil and rock along its contact surface. The safe mining depth, in this case, is determined by the formula [11]: Ha =

0,9 × m × sin2 α [Dc ]

(1)

where m is mining thickness (meters); [Dc] is the value of permissible horizontal deformation (deformation criteria) or the inclination of work, whose value is taken from Table 2; α is the slope angle of the coal seam (degree).

Recycling Ash and Slag of the Thermal Power Plant

325

Table 2. Criteria of permissible ground deformation and limitation for the group of civil and industrial constructions on the mine surface No.

Type of construction

Type of deformation

Value Permission

Limitation

1

3-storey houses, brick wall 10 m × 15 m, atrophy 30%

[ε]

4 × 10−3

5 × 10−3

2

3-storey houses, frame structure 15 m × 20 m, atrophy 30%

[ε]

5 × 10−3

6 × 10−3

3

Brick barrier wall 0.2–0.3, 1.5 m [ε] high, 40–50 m long.

10 × 10−3

12 × 10−3

4

School buildings, hospital, office, [ε] concrete frame structure, brick wall, 2–3-storey high, atrophy 30%

3.5 × 10−3

4,5 × 10−3

5

Water pipelines a. Main pipeline on the ground: b. Main steel pipeline underground:

[ε] [ε]

10 × 10−3 5 × 10−3

15 × 10−3 8 × 10−3

6

Load of rail under 10 million tonnes/year, speed within 40 km < v < 80 km National highway 18

[i] [ε]

10 × 10−3 8 × 10−3

10 × 10−3 8 × 10−3

7

Substation a. 110 ÷ 400 kV b. 0.3

0,7

0,4

0,4

0,15

0,7

1,0

k0

Horizontal Deformation The horizontal deformation of any point in the subsidence is determined as follows: – At L B1 area: ηm × S  (z) LB1

(8)

k0 × ηm × (1 + kπ ) × F  (z) LB2 + Lπ 1

(9)

k0 × ηm × kπ × S  (z) Lπ 2

(10)

εz = – At L B2 and L L1 area: εz = – At L L2 area: εz =

The curvature The curvature of any point in the subsidence pot the following formula: Km =

im − im−1 lcp

(11)

338

P.-H. Nguyen et al.

where im , im−1 are degrees of inclination at m point and m − 1 point; lcp is average distance length. Based on the selected calculation method, the prediction of surface displacement parameters was calculated when controlling the roof rock by hydraulic backfilling (see Figs. 13, 14, 15 and 16).

Fig. 13. Prediction of the variation in surface subsidence η

Fig. 14. Prediction of variation in inclination of topographic surface I x

Fig. 15. Prediction of variation in horizontal deformation ε z

Recycling Ash and Slag of the Thermal Power Plant

339

Fig. 16. Prediction of variation in curvature K x

The summary of calculated terrain deformation parameters for the proposed pilot application area is shown in Table 8. From this result, the surface displacement area corresponding to cutting lines on the topographic map of the mine can be specified (Figs. 17 and 18).

Fig. 17. Estimated displacement tank on top of the surface

Settlement ηz , mm

42.0

41.0

38.0

32.0

24.0

16.0

9.00

4.00

2.00

0.00

0.00

Relative coordinates z

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0.00

0.06

0.13

0.29

0.45

0.61

0.64

0.54

0.32

0.16

0.00

0.00

0.05

0.10

0.23

0.36

0.49

0.52

0.44

0.26

0.13

0.00

0.00

1.07

2.02

2.97

3.97

4.48

5.11

5.56

5.87

6.12

6.31

0.00

1.02

1.92

2.82

3.6

4.26

4.86

5.28

5.58

5.82

6.00

L 1 side

0.00

1.20

2.40

5.40

8.40

11.4

12.0

10.2

6.00

3.00

0.00

L 2 side

Wall side

L 2 side

L 1 side

The displacement in the horizontal direction ξz (mm)

Inclination iz (× 10−3 )

0.00

0.05

0.10

0.23

0.36

0.49

0.52

0.44

0.26

0.13

0.00

Wall side

0.23

0.21

0.20

0.18

0.15

0.13

0.10

0.08

0.05

0.03

0.00

Areas L B2 and L B1

0.00

0.03

0.06

0.13

0.20

0.27

0.28

0.24

0.14

0.07

0.00

Areas L L1 and L L2

Horizontal deformation εz (× 10−3 )

Table 8. Results of calculated ground deformation parameters in the test area

−0.008 −0.008 −0.008 −0.003 −0.003

−0.012 −0.012 −0.005 −0.005

−0.002

−0.002 −0.012

0.005

0.011

0.008

0.008

0.000

L 1 side

0.007

0.017

0.012

0.012

0.000

L 2 side

Curvature K Z (mm/m2 )

340 P.-H. Nguyen et al.

Recycling Ash and Slag of the Thermal Power Plant

Fig. 18. Some cross-sections of the proposed displacement tank

341

342

P.-H. Nguyen et al.

6 Conclusion The above-calculated results show that when exploiting and backfilling by hydraulic method at the levels −80/+0 of the long wall with a horizontal length of exploitation of 210 m, a surface displacement tank with horizontal dimensions will be formed a length (about 380 m) and a width (about 300 m). However, the parameters of displacement and surface deformation are not significant: maximum settlement ηm = 42 mm, maximum inclination Iz = 0,64.10−3 , maximum horizontal deformation εz = 0,52.10−3 . The predicted values of displacement and deformation parameters are all within the permissible limits for civil and industrial constructions on the surface in Table 4. Therefore, the technology of exploiting and controlling cliff walls using hydraulic furnace inserts meets the requirements of protecting surface works. Flow state of the mixture with the ratio of bottom ash to water is greater than 1:4, and with the component of fly ash is 20%, and bottom ash is 80% is the best chosen. Research results of the calculation and application of waste ash and slag from the Mao Khe Thermal Power Plant as backfilling materials for the underground mining works at the southern wing area of the Mao Khe coal mine are feasible. The proposed technical solution meets not only the technical requirements (mining safety) but also the criteria to maximally utilize coal mineral resources without causing significant impacts on surface works that need to be protected. Besides, the use of fly ash and bottom ash will contribute to the release of a large amount of production output from the Mao Khe Thermal Power Plant, which helps to reduce the area of the waste dump, reduce environmental impacts, and increases economic benefits. Conflict of interest. The authors declare that there is no conflict of interest.

References 1. De Souza, E., Archibald, J.F., Dirige, A.P.: Economics and perspectives of underground backfill practices in Canadian mines. In: 105th AGM-CIM, Montreal, Canada, 15 p. (2003) 2. Grice, A.G.: Recent minefill developments in Australia. In: Proceedings of the 7th International Symposium on Mining with Backfill, Seattle, Washington, pp. 351–358 (2001) 3. Wayment, W.R.: Method of mine backfilling and material therefor (USP1977114059963), USA (1976) 4. Antonov, D.: Mine Backfill Design and Characteristics: New Concept for Backfill Underground Support, 291 p. VDM Publishing, Germany (2009) 5. Manoon, M.: Systematic Selection and Application of Backfill in Underground Mines, Freiberg, Germany, 336 p. (2009) 6. Phung, M.D., Nguyen, A.T, Truong, D.D: Research and select reasonable technical and technological solutions for coal mining in areas with historical and cultural relics, industrial and civil constructions. Institute of Mining Science and Technology-Vinacomin, Hanoi, Vietnam, 165 p. (2011). (in Vietnamese) 7. Nguyen, A.T., Truong, D.D., Dang, H.T.: Studying and applying the technology of backfill for roof control and protecting surface objects in the conditions of underground mines in Quang Ninh. Institute of Mining Science and Technology-Vinacomin, Hanoi, Vietnam, 124 p. (2006). (in Vietnamese)

Recycling Ash and Slag of the Thermal Power Plant

343

8. Dao, H.Q.: Research and apply the technology of backfill in the longwall coal mining systerm in Quang Ninh coal mines areas. Innovating and modernizing technology in the mining industry by 2015, with a vision to 2025 of VietNam Ministry of Industry and Trade Research programme, Hanoi, Vietnam, 190 p. (2015). (in Vietnamese) 9. Mao Khe Coal Mine: Geological literature on the southern wing area, Quangninh, Vietnam, 64 p. (2019). (in Vietnamese) 10. Mao Khe Coal Mine: Plan for exploitation and protection of surface constructions in the South wing area, Quangninh, Vietnam, 80 p. (2018). (in Vietnamese) 11. Le, N.H.: Underground coal mine design. Transport Publisher, Hanoi, 273 p. (2008). (in Vietnamese) 12. Mao Khe Thermal Power Plant: Report on production and business activities in 2018, Quangninh, Vietnam, 37 p. (2018). (in Vietnamese) 13. Le, N.H.: Technology design and parameter calculation systerm for underground mining. Hanoi University of Mining and Geology, Hanoi, Vietnam, 250 p. (1999). (in Vietnamese) 14. Tran, V.T.: The special method for explotation in the underground mining. Hanoi University of Mining and Geology, Hanoi, Vietnam, 200 p. (2008). (in Vietnamese) 15. Zolotapev, G.M., Gopoxov, .D.: Mexanizaci zakladoqnyx pabot na ygolnyx xaxtax. M.: CHIIygol, 85 p. (1974). (in Russian) 16. Ocnovy texnologii podzemno pazpabotki mectopodeni c zakladko. otv. ped. ppof., d-p texn. nayk A. F. Hazapqik. M.: Hayka, 200 p. (1973). (in Russian) 17. Ppimenenie cictem pazpabotki c tvepdewe i betonno zakladko vypabotannogo ppoctpanctva: cb. ct.. ped. H. C. Mypaxova. M., 168 p. (1967). (in Russian) 18. Ppavila oxpany coopyeni i ppipodnyx ob ektov ot vpednogo vlini podzemnyx gopnyx pazpabotok na ygolnyx mectopodenix (PB 07-269-98). BHIMI (1998). (in Russian)

Development of the Global 21st Century Mining Technical Services Professional: The WMI-SAGE Collaborative Model Sarfraz Ali(B) , Frederick Cawood, Tariq Feroze, and Hamid Ashraf Wits Mining Institute (WMI), University of the Witwatersrand, Johannesburg, South Africa [email protected]

Abstract. Mining in 21st century faces with numerous challenges. In addition, mining at depth is faced with extreme geo-stress uncertainties, geomechanics, and cost, demanding sophisticated design and site-specific planning. The extraction of ores necessitates advance delineation of ore bodies to minimize the cost of extraction and processing thereof. Besides, waste and tailings concentrations need to be minimized and preferably eliminated. New underground support systems, once designed to cater to usual stress re-distributions of excavations, are now faced with extremely high demands of structural strength, thus necessitating a paradigm shift in design philosophy. Such improvement of mine support systems from conventionally passive to active and de-stressing concepts must withstand enormously high sequentially transiting loads from static to dynamic stress states. The new complex mining environment is also a challenge for mine health and safety, which issue is of great concern for regulators, mine workers, and mine owners. Ultradeep mines now require automated digital systems to monitor the real-time well being of people, mining equipment, and underground excavations. The authors of this paper believe that the existing knowledge, technologies, and techniques in the mining industry have reached their limits, and it is time to re-imagine the technical services knowledge and skills requirements in the context of the 21st century and innovate for efficient, safe and sustainable mining. To accomplish these complex and urgent needs of the mining industry, international collaboration is required. The success of the WITS-NUST collaboration (detailed in another paper), have encouraged the authors to extend their earlier collaboration to embark upon qualifications design for the development of the global 21st Century Technical Mining Services Professional. Keywords: 21st century mining · Mining challenges · Deep mining · Digital mining · Mining collaborations · Collaborative models · Technical services qualifications · University collaboration

1 Introduction Metals and minerals are essential in all aspects of life, e.g. farming, healthcare, communications, water, energy supply, transport, space technology, and construction of cities © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 344–363, 2021. https://doi.org/10.1007/978-3-030-60839-2_18

Development of the Global 21st Century Mining Technical Services Professional

345

[1]. With the ever-increasing population, the demand for base metals, particularly iron, copper, and aluminum, will double from 2010 to 2025 [2]. To put only the demand for iron ore in perspective, the World needs five more Rio Tinto Pilbara Mine operations producing nearly 200 MT iron ore annually [2]. In addition, the copper demand from 2010 to 2035 will equal all copper consumed in the last century [2]. The greatest challenge the mining industry faces today is to meet such exponential growth in the demand for minerals and metals. With the exhaustion of high grade and easily accessible ore deposits, the challenge of meeting emerging demand is compounded by the need to beneficiate low-grade ore waste stocks or mine at much greater depths. The extraction of high-grade ores up to medium level depths necessitates advance delineation of ore bodies to minimize the cost of extraction and simultaneously developing technologies for cost effective processing of large low-grade ores and stockpiles once left as uneconomical waste. Contemporary underground support systems, once designed to cater for usual stress re-distributions of excavations, are now faced with extremely high demands of structural strength. This necessitates a paradigm shift in the design philosophy of conventionally passive to active and de-stressing concepts to withstand enormously high sequentially transiting loads from the static to dynamic stress states. The purpose of this paper is to highlight the 21st -century geo-mining and technical services challenges and share our experience of collaboration amongst the Wits Mining Institute (WMI) at the University of the Witwatersrand (WITS), Johannesburg, South Africa and the School of Advanced Geomechanical Engineering (SAGE) at National University of Sciences & Technology (NUST), Islamabad Pakistan. The authors have conceptualized the WMI-SAGE Mining Educational Collaboration with a concept to design a curriculum for and research to support a 21st -century mining industry. The paper spans over six sections, inclusive of this Introduction and the Conclusion (Sect. 5); Sect. 2 highlights the challenges faced in the 21st century that inform mining qualifications and skills. Section 3 describes the requirements for qualifications and skills for 21st -century mining. Section 4 explains the structure, framework, and mechanism of WMI-SAGE collaboration model to translate mutual capacities to research and development. Section 5 covers the implementation details of the SAGE postgraduate program in collaboration with the WMI in South Africa.

2 Twenty-First Century Mining Challenges Informing Qualifications and Skills Design 2.1 Mineral Resource Management Challenges The 21st -century mineral industry has changed dramatically and at an exponential rate in the last twenty years. These changes are due to significant advances in technology. The industry today relies on 4IR technologies to gather data on all aspects of mining so that professional occupations can make better decisions. However, despite these improvements, the industry remains under tremendous pressure to do more. The sustainable or responsible mineral development initiatives have made it necessary for mining professionals to become more skilled in the field of digital technology, in addition to the core skills. Digital skills as these apply to mining (often called Digital Mining) are therefore very important for skills design aimed at the Technical Services Professional.

346

S. Ali et al.

Mineral Resources will remain the key driver for economic growth in the 21st century. However, the extent to which these transformations take place depends upon the management of these resources through technology, skills development, and innovation. Mineral resources are national assets of a country and should be managed responsibly by professionals having appropriate qualifications and skills. For this to happen in countries like Pakistan, the industry has to embark on an implementation plan for embracing the challenges of 21st -century mining as part of qualifications design for professional occupations. This implementation of a sustainable development goal requires the building of skill development, and integration of different elements explained below (Fig. 1).

Fig. 1. 21st -century sustainable technological, multi-disciplined skilled development plan

• Pillar One: A sustainable economic growth for the national benefit and greater good of the populace. • Pillar Two: Environmentally responsible mining by reducing tailings and waste by incorporating recycling in the mining value chain. • Pillar Three: A socially responsible mining operation for the corporation to earn its social license to operate. • Pillar Four: Technological advancement in mining, materials, and equipment. • The Foundation: Capacity building for the 21st century and skills development, which act as a foundation for the four pillars highlighted above. At a national level, digital mining necessitates policymakers to incorporate digital or smart mineral resource management into policy formulation frameworks. Professionals must also consider sustainability issues affecting natural assets and resultantly, design policies to obtain both short- and long-term benefits. It is essential that these asset-specific strategies are designed in such a manner that they do not work against one another and are aligned with the overall country strategy. Therefore, the 21st -century mining must

Development of the Global 21st Century Mining Technical Services Professional

347

be based upon smart economic principles and the digital economy. The 21st -century mineral resource management paradigm must, therefore, consider the following three fundamental characteristics for the mining industry [3]: • Increased transparency and capacity building of the institutions to deal with the challenges of 21st century in the resource sector by incorporating skills development of key institutional stakeholders (Fit-for-purpose short executive courses) described later in Sect. 4 (Fig. 10); • Planning and establishment of regulatory institutions to promote trust among all stakeholders to deal with industry-specific 21st -century challenges; and • Development of responsible policies to project the mineral economy into a smart and green economic benefit, skills development must also consider the role of the regulator, mineral law framework, and the link to the broader economy through local content and value addition. In recent times, there has been a global shift towards a circular green economy1 , which warrants the reuse of minerals and re-imagination of mine waste strategies. Volume extraction of minerals will continue to grow in line with global GDP growth, which will continue to put pressure on corporations to appreciate scale effects and cost-efficiency. Industry demands for cost-effectiveness in parallel with the demand for environmentally and socially responsible actions will be the standard for 21st -century mining. This requires a new operating model of recycling, digital mining, and waste management, which further informs qualifications design. Above all, technology will be the key. Mining companies will have to focus on waste treatment optimization and metal companies on the improvement of low-grade processing capabilities. It will become increasingly important to understand better supply chains and consumer preferences, which make multi-disciplinary skill development programs and international collaborations even more critical. Figure 2 explains the requirement for skills development in the 21st century, where the industry will have to strive to cut the flow of the process (indicated with the red arrows) and innovate to increase the flow of the process (represented with green arrows). 2.2 Mining Geomechanical Challenges In the past, rigorous analytical approaches in the analysis and design of complex engineering problems led to the concept of specialization in each sub-discipline of engineering such as geological engineering, soil engineering, rock engineering, geotechnical engineering, tunneling engineering, and mining engineering. The specialization concept 1 Material implications within a green economy include materials used in wind, solar and energy

storage batteries technologies. The key materials examined by World Bank scenarios were aluminum, chromium copper, indium (rare earth), iron, lithium, lead, manganese molybdenum silver, steel and zinc. However, other materials acknowledged as important for green economy include antimony, boron, cadmium cerium, chromium dysprosium europium, gallium, germanium gold, lanthanum, neodymium nickel, niobium platinum, praseodymium, selenium silicon tellurium terbium, tin zinc and vanadium.

348

S. Ali et al.

Fig. 2. Perceived 21st -century mining cycle illustrating the importance of skill development within a collaborative data flow process

necessitated close cooperation, coordination, and harmony of the specialist professionals. Though it worked well in the past and would surely continue to be of value in the future, the computational innovations such as explicit time marching numerical techniques have encouraged unification of engineering disciplines (ranging from partial to complete) and collaboration with non-engineering fields, such as health and safety, economics, management, law, and other technical disciplines. With the relatively shallower reserves already exhausted in most countries, future extractions would be at greater depths requiring autonomous mining systems to reduce worker exposure to risks in deep and ultra-deep hostile environments [4]. Salient features from research paper [4] include: • Uncertainty and variability are characteristic features of rock engineering. • Although the empirical rules for complex rock mass problems are still valid, they should not be applied beyond the bounds of established practice (such as mining at great depth) without more research. • The Hoek-Brown criterion and Q-System were developed for estimating the strength of rock mass with geological features that have extensively been used in the past. However, they are unable to characterize 3D behavior of the jointed rock system, pressurized fluid flow, interaction of dynamic loadings, etc. • With the advancement in computational speed, DEM (Distinct Element Method - a numerical technique for discontinuous materials) is more applicable to rock engineering problems to incorporate factors that may have a critical influence on both the predicted and actual behavior of the rock system.

Development of the Global 21st Century Mining Technical Services Professional

349

• Synthetic Rock Mass (SRM), created by superimposing DFN (discrete fracture network - a representation of large scale discontinuities of rocks from field observations) on Bonded Particle Model (a representation of the intact rock from laboratory tests on cores) is a major advance to consider rock mass as a 3D anisotropic discontinuity [5] (Fig. 3).

Fig. 3. The Synthetic Rock Mass (SRM) [5]; the future of the mining geomechanics

• DEM codes can be used to study the rock fragmentation process at the particle level to develop efficient rock excavation systems. • The establishment of multi-disciplinary R&D groups in mining is the preferred way to embrace 21st -century mining challenges. • Using SRM, Itasca (Fig. 3) has innovated techniques to predict primary fragmentation in the rock mass. Analogous to the modern numerical codes, which enable coupled-modeling, analyzing, designing, and predictive forecasting of diverse engineering projects, innovative engineering technologies have also transformed into multi-purpose functionalities. Current challenges and issues are demanding coupled geomechanical (geological, geotechnical, hydrological, structural, thermal, rheological) static and dynamic analysis of the mining-geotechnical projects are offering an opportunity to amalgamate sub-disciplines and sub-technologies, e.g. geoengineering and geotechnology.

350

S. Ali et al.

2.3 Mine Health and Safety Challenges Mining, irrespective of the type of minerals, mines, and mining methods, is inherently a hazardous profession, and management of Occupational Health and Safety (OHS) is considered one of the primary challenges. Therefore, mines must plan, develop, and implement Codes of Practice (COP) to manage the hazards using best risk management practices as, for example, have detailed guideline documents [6]. The implementation of such standards is a fundamental right of mineworkers since it enables them to carry out their responsibilities and go underground without the fear of hazards and being unable to return to their families every day. Implementation of stringent regulations introduced to reach the goal of zero harm and because of pressure from organized labor has further improved mine workplace safety. However, even with such improvements, countries like South Africa still have too many fatalities because of injuries, occupational diseases, and related deaths (Table 1). The presence of these occupational issues in the 21st century is an indication that more education is required to improve health, safety, and management skills in mining. Table 1. Mine safety and health data south african mines [7, 8] Year

Fatalities

Occupational injuries

Occupational diseases

Medical deaths due to occupational diseases

2016

73

2847

4632

902

2017

90

2664

4483

975

2018

81

2447

3458

999

2019

51

2406

Not available

The quest to achieve the target of zero harm in South Africa proves to be complicated. One possible explanation is that significant efforts are made to regulate such risks. Still, not enough is being done to develop skills and competence in health and safety science and risk management. Also, new knowledge and approaches are required to reach the goal of zero harm. The authors strongly believe that zero harm is within reach, but will require more adoption of fourth industrial revolution technologies, best practices, and technologies for real-time monitoring of hazards. Additionally, it requires an integrated approach of predictive numerical modeling and to aid automated controls and decision making, as illustrated in Fig. 4. Such a real-time monitoring approach to OHS requires a multi-disciplinary approach. Working in silos is not going to help to achieve zero harm. The biggest challenge will thus be to equip professionals with relevant skills so that they cannot only develop, but also understand how new technologies work for effective design and installation of such systems. This is the objective of the WMI-SAGE model (Fig. 4), which is proposed as a way forward to achieve zero harm in Pakistan’s mining industry. Mining researchers and academics will need to improve their skills-sets by closely following developments in other areas of engineering and health sciences, and by participating in developing and

Development of the Global 21st Century Mining Technical Services Professional

351

Fig. 4. Integrated multi-disciplinary (WMI-SAGE) approach to achieve zero harm

implementing the best techniques and technologies to solve the mining and mine OHS problems. 2.4 Effectiveness of the Current BS-MS Mining Education System The significance of mining education systems is usually measured by student progression rates between years of study and employment rates following graduation. In this paper, we are concerned about the shortage of mining engineers, which is considered a ‘critical, scarce skill’ – even in countries like South Africa that are known to supply large numbers of mining engineering graduates to the market [9]. The issue is aggravated at a postgraduate level because of the small pool of mining engineering graduates wanting to do postgraduate qualifications. For this reason, this paper proposes a Geotechnical professional qualification at a Master’s degree level to add mining engineering content to other graduates. Such a system will help in two ways. It will firstly increase the graduate pool feeding the postgraduate program. Secondly, by admitting appropriately qualified first-degree graduates into the program, more mining engineering professionals can be developed for specific technical fields of study. For this to work, the mining education system must add additional mining engineering content to the postgraduate program, hence the need for a different course on Induction/Introduction to Mining Engineering for differently qualified graduates. Such a system allows Universities to issue a Masters qualification in Mining Engineering to someone entering the program with a Bachelor’s degree in e.g. Mechanical Engineering after completing the mining engineering

352

S. Ali et al.

requirement of the program. Such broadening of knowledge also addresses a significant challenge for 21st-century mining, namely the addition of more multi-disciplinary (e.g. digital) content from other disciplines. Keeping this in view, the authors have articulated a Mining Qualifications and Skills Framework described in Sect. 4.3, which suggests the unification of various mining related disciplines into mining engineering qualification, analogous to a concept of “mono-engineering”.

3 Qualifications and Skills for 21st Century Mining This section describes the qualification and skills requirements for first, the Technical Services Professional, and second, implementation of new techniques and technologies. The Society of Mining Professors (SOMP) published a guideline containing learning content for mining qualifications [9]. Also, Cawood [11] proposed the following program for 21st -century mining professionals: • World of work content: Topics include personal mastery, green skills, ethics and value systems for mining professionals; • Science content: Topics include mathematical-, computer-, spatial and geosciences for mining; • Engineering content: Topics include the application of general engineering principles to mining, mining methods, mine planning, and design, mine geotechnical engineering, mine operations management; socio-environment factors and mine modernization; • Mining business and economics: Topics include economic systems, mine finance, and mineral resource management; • Management and leadership: Topics include principles of management, resources-, change-, safety-, technology- and risk management; and • Mining and mineral regulation, with the content on the legal knowledge required for mining professionals. Innovation-driven growth is usually inclusive when the focus is on the creation of relevant skills and qualifications. Such skills and qualifications then support a “procompetitive business environment providing ready access to information, financing, export opportunities, and other essential business services that facilitate the entry and expansion of young firms”. For university capacities to become sustainable, they require collaboration with other research centers, government, industry, organized labor, and start-ups. Such cooperation needs the understanding of each discipline with enough knowledge of other disciplines to address the many new challenges when applying 21st century technologies to mining. A 21st century model for mining must, therefore, add knowledge to traditional qualifications so that innovation can lead to job creation, complex problem solving, manufacturing, and services to support a thriving mining sector and economy. This is difficult to get right within one university. Typical university constraints are the costs associated with developing and offering qualifications and a scarcity of international professional skills in the technical services space. High investment barriers to entry constrain educational budgets, which education budget must compete with other

Development of the Global 21st Century Mining Technical Services Professional

353

country priorities like e.g. health services and defense. For innovation-driven teaching, a database or skills matrix for the mining cluster is essential to understand the migration of skills and where these are needed. If universities and governments can overcome these constraints, value is created when their graduates are capable, can work safely and productively so that value is added to the business and technological innovation becomes possible. Businesses are then more cost-effective, and they have the skills to implement new technologies efficiently. The ultimate value proposition is the sustainable and responsible development of non-renewable resources for the broad benefit and inclusive growth of the country. 3.1 A Multi-disciplinary Approach for the New World of Work for Mining Considerations for mining qualifications and skills that are relevant in the 21st century are competencies to make mines smarter, safer, leaner, and more efficient, in addition to understanding sustainability issues like environmental, social, and economic impacts. Sustainability, health, and safety topics are crosscutting, which require all professionals to have a working knowledge of sustainable development, risk management, and, more importantly, to understand the impact silo disciplines have on others, the environment, and the economy. Technology adoption is fundamental to this, and it affects all disciplines. For technologically advanced businesses, new skills-sets are required as summarized by IISD “Higher-skilled tasks such as those linked to the analysis of data, digital planning or remote central operations will be created”. Such competence supports both innovation and job creation because these are transferable skills allowing for opportunities elsewhere in the economy. Figure 5 shows how recent impact trends on the future of mining. The last two trends identified in Fig. 5 are having a profound impact on the future of mining and implies that the mine of the future will be and look very differently – and staffed by professions that do not exist today. Professionals who use their deep (core or first-degree) discipline as a foundation to build a geotechnical qualification that is future-ready, will demand premium jobs and able to innovate. 3.2 Application to the Technical Services Professional in Mining The South African Mining Qualifications Authority (SA MQA) has developed a career guide with a list of scarce skills within the SA mining and minerals sector [12]. The career guide lists Mining Manager, Mine Planner, Mining Engineer, Rock Engineer, and Mine Surveyor in the ‘hard to fill’ or scarce category. It states that the main reasons for the scarcity are geographical clustering of skills and lack of career awareness while students are still at university. While it is possible to study for a Mining Engineer (University), Mine Manager (Government Certificate of Competency) and Mine Surveyor (Combination of University and Government Certificate of Competency) at the Professional level, there are no formal qualifications at a first-degree level for a Mine Planner and Rock Engineer. However, the career guide does provide for professional qualifications at a Master’s degree level, and universities typically offer such qualifications as an area of specialization in master’s degree programs. The document also gives guidance on occupational titles within the mining industry. In addition to the scarce skills listed earlier, such

354

S. Ali et al.

Fig. 5. Global trends that will affect the future of mining [13]

titles at the professional level include Inspector of Mines, Engineering Manager, Mineral Resources Manager, Environmental Manager, Geologist, Geophysicist, Industrial Engineer, Mine Ventilation Engineer, and Metallurgical Engineer. Because of a combination of skills scarcity and simplified mine organizational structure, it has become standard practice to put several professional occupations on one department, often referred to as ‘Mine Technical Services’. Typical professions in Technical Services are Mine Planners, Rock Engineers, Mine Surveyors, Mineral Resources Managers, Environmental Managers, Geologists, Geophysicists, Industrial Engineers, and Mine Ventilation Engineers. To be promoted to the head of a typical technical services department, it is the first necessary to have a core specialization, e.g. mine planning. Secondly, the head must have sufficient knowledge of the other professions so that there is an understanding of the more prominent departmental function and the roles of the different professionals within the context of a technical service. For these reasons, universities typically offer professional qualifications at firstly, a master’s degree level, secondly, have core areas of specializations and thirdly, allow for a system of electives so that e.g. a Ventilation Engineer have sufficient understanding of rock engineering, surveying, planning, and economics concepts to be promoted within the structure. A multi-disciplinary approach to the technical services profession is not new. The specialized services profession draws from core disciplines like e.g. engineering, geology, and mine surveying. Typical jobs for such professionals are in areas where there are no first degrees, e.g. rock mechanics, mine ventilation, mine planning, mineral resource management, and mine evaluation. The skills-sets of such technical professionals must be such that they can lead people who perform the following roles: • • • •

Service technicians (e.g. Mechanics, Electrics and Electronics) Mining technicians (GeoTech: Economics, Ventilation, Survey, Rock Eng.) Value-based “green” skills (e.g. what is happening in architecture) Business-by-database skills (Smart exploration and sourcing market goods)

Development of the Global 21st Century Mining Technical Services Professional

355

• Thinking forward skills (Statistics, pattern recognition, and creativity) • Risk management using digital systems (GIS and sensing technology skills) Because of their leadership roles, Technical Services Professionals should also possess people skills on multicultural diversity and have the ability to work in a team - and collaborate within a framework of values and trust. Figure 6 (LHS) below illustrates how the WMI unites inside and outside South Africa for research on mine health and safety through the application of digital mining sciences and technologies to mining.

Fig. 6. Partner collaboration of DigiMine (LHS) and Application of digital science and engineering to mine health and safety (RHS) Source: DigiMine Business Plan

4 Implementation of the WMI-SAGE 21st Century Mining Technical Services Framework for Professional Development 4.1 A Government-Academia–Industry Linkage Framework The government is the custodian of mineral growth in a country and a significant source of funding (through ministries, councils, higher education commissions, science foundations, etc.) to the universities. In some cases, mines also collaborate with universities for skills development, research, and technology development. Such initiatives are mostly site or problem-specific and applied to the sponsor’s needs. For many mines, especially junior mines, it is not possible to sustain the full range of skills or qualified engineers in all streams. In the case of international partners, the involvement of the respective consulates also becomes vital as delays in visas can seriously affect timelines and deliverables. Given these resources constraints, the authors propose a meaningful “Government-Academia-Industry Linkage” as illustrated in Fig. 7. However, the content and substance of the proposed matrix is likely to vary from country to country for both commodities and mining methods. The proposed “Government-Academia-Industry framework for Pakistan is described in Table 2 and lies at the heart of the Joint Stakeholder’s Desk (Fig. 7). Before the evolution of this framework, the authors under the WMI-SAGE collaboration (an extension of NUST-WITS collaboration) had formulated the joint qualification and skill framework, as shown in Figs. 8 and 9.

356

S. Ali et al.

Fig. 7. A generalized Government-Academia-Industry linkage Framework for sustainable mineral growth

Development of the Global 21st Century Mining Technical Services Professional

357

Table 2. Agenda and parametric focus of the government-academia-industry framework Agenda

Parameters

Mineral resource management and growth for • Mineral Law and Policy framework social development • Transparency • Local skill development and • Job creation • National, Provincial and Mine-wide information management systems • Mine health and safety compliance, including real-time monitoring of health and safety variables • Others Sustainable development for future generations

• Environmental impact assessment • Efficient processing to yield minimal waste • Post closure mine management (real-time monitoring of ecosystem, acid mine drainage, mining-induced seismicity, etc.) • Res-use/recycling of ores stocks and minerals • Optimizations and innovations • Others

Management of mining risks

• • • •

Mining Geomechanics

• Real-time monitoring (ground behavior, support systems, and infrastructure, etc.) • Predictive forecasting using advanced numerical modeling techniques • Optimizations and innovations • Others

Security of tenure/license to operate Cyber/digital data security New qualification and skills (Fig. 8) Demand, supply, cost, recycling, and alternative materials • Optimizations and innovations • Others

4.2 WMI-SAGE Mining Technical Services Qualification and Skills Framework As highlighted earlier, practicing mining professionals require qualifications to enable them to design, operate and maintain technology-intensive mines, gathering real-time data, evaluate alternative scenarios and carry out predictive forecasting to make sound engineering decisions using many variables. Figure 8 illustrates core and elective courses at a Master’s Degree level of the WMI-SAGE joint mining technical services qualification framework. Candidates have the advantage to select core and elective courses according to first, the Higher Education Commission and Engineering Council(s) credit requirements, and second, specific needs of the mining operation and the professional’s career

358

S. Ali et al.

Fig. 8. WMI-SAGE joint mining technical services qualification framework

path plans. The course work is followed by a project report encompassing 21st -century digital technologies such as digital-wireless sensors, MEMS (micro-electromechanical systems), IoT (internet of things), AI (artificial intelligence), and ML (machine learning) techniques. As illustrated in the Fig. 8, candidates get support in teaching and supervision at both WITS and NUST. The qualification framework is flexible according to the needs of the Government-Academia-Industry stakeholders.

Development of the Global 21st Century Mining Technical Services Professional

359

Fig. 9. The manifestation of the WMI-SAGE collaborative mining geomechanical research project “DigiMine”

4.3 Joint WMI-SAGE Research Agenda The application of the 4IR or digital mining concepts to mining forms the core of the WMI-SAGE research agenda. The project involves research and development of the entire range of 21st -century technologies and the implementation of a wireless real-time ground behavior monitoring system [14] with extensive involvement of SAGE. Phase 1

360

S. Ali et al.

was planned in 2014 to test real-time ground behavior monitoring systems in a mock-up mine environment. This phase involved testing digital sensors to measure deformations, vibrations, stresses and strains, extension and convergence, volatile compounds, weather, and seismicity. Observing the performance and limitations of various sensors installed in the mock-up gives the lessons on the accuracy, connectivity, vendor response, and system reliability, etc. Phase 2 was launched in 2018 with ground monitoring sensors that were installed at the Sterkfontein Caves located some 50 km away from WITS. In Phase 3, the real-time ground Behaviour monitoring concept is currently tested in real mine environments aimed at addressing specific problems. Figure 9 illustrates the connection to the Sibanye-Stillwater Digital Mining Laboratory (DigiMine), which provides some of the opportunities for joint research within the WMI structure. DigiMine’s work and data flows allow for the following: • Real-time time collection of sensor data, which data is stored in the control room. • Analysis of data through identifying trends under varying conditions e.g. season, weather temperature, precipitation, ground water, quality of air, seismic events, etc. • Using geomechanical properties (physical, mechanical, stiffness) of the rocks and rockmass, baseline numerical models (both cave and chamber scale) would be developed and calibrated to simulate/mimic the ground behavior observed from for the real-time monitoring data. • Using calibrated/validated models, futuristic mining excavation design, and modeling to predict the ground behavior. For on-mine studies, more predictive forecasting scenarios are studied, such as rate of advance, support systems, etc., at various scales (regional, mine, and stope scale for the real mine). • Using various iterations of the predictive forecasting of scenarios vis-à-vis ground behavior responses, it is then possible to define and implement optimizations and innovation in layout design, rate of advance, support systems, etc. 4.4 Extension of WMI-SAGE Collaboration to Benefit Pakistan The faculty of SAGE capitalized on WMI capacity by supporting provincial mining department projects and offering courses in numerical modeling techniques and the application of digital mining principles to mining. SAGE, in collaboration with Landfolio South Africa, was contracted for the development of a Mining Cadastre System for a mining department in Pakistan. Benefits of this WMI-SAGE collaboration include the facilitation of meetings, travel, accommodations, contracting, and training on the mining cadaster system. Besides, SAGE has an outreach program to involve the training of Pakistan’s consultants and company staff to do courses on advanced numerical modeling techniques using 3DEC, UDEC, FLAC, etc. Keeping in view the keen interest and appreciation shown by relevant stakeholders in Pakistan and South Africa, WMI and SAGE have developed the skills development plan shown in Fig. 8. It illustrates skills gaps to address 21st-century mining challenges. For example, an advanced numerical modeling short course for WMI students was organized at NUST Main Campus, Islamabad in December 2018. A refined version of the short course is planned in South Africa in 2020, to include a wider audience and industry professionals. To maximize the outcomes of the WMI-SAGE collaboration, a series of

Development of the Global 21st Century Mining Technical Services Professional

361

short courses (illustrated in Fig. 10) were planned for provincial mining departments in both Pakistan and South Africa in April 2020. Unfortunately, these short courses were postponed due to the COVID-19 Pandemic and subsequent international lockdown.

Fig. 10. The skill development plan for 21st -century mining challenges

5 Conclusion The mining industry of the 21st century is faced with numerous challenges. Many of these challenges can be addressed by ensuring that the professionals who design, manage, and maintain mines have up-to-date knowledge and skills. The existing qualifications must be re-imagined, including multi-disciplinary technical services knowledge and skills requirements to stimulate innovations for efficient, safe, and sustainable mining. To accomplish this urgent need, government, industry, and academia must collaborate, and such partnerships must extend internationally. This article summarizes the skills and research rollout program of the WITS-NUST collaboration (detailed in the first article). More specifically, it targets the multi-disciplinary Technical Services Professional to provide technical support to mine managers. The research program is already established, the provision of courses has started, and the first postgraduate intake at NUST happened in the fall of 2019. In summary, the NUST-WITS (later WMI-SAGE) collaboration has proved to be highly successful from capacity building and research perspectives. However, the 21st century brings complex mining challenges of responding to ever-increasing mineral demand, low grade, and mining ultra-deep deposits. Such mining requires more multidisciplinary technical services professionals to support mine managers. These professionals need to equip themselves with more knowledge and technology skills. In the

362

S. Ali et al.

face of international financial, human, and resource constraints, collaborations such as WMI-SAGE collaboration is a model to cluster scarce skills, reduce cost, capacity development timeframes, enhance confidence amongst cross borders stakeholders in the face of ever-changing geo-socio-political environment. Decision making in mining operations must be based on the patterns and trends generated by data acquisition so that Artificial Intelligence and Machine Learning result in mathematical algorithms for mine design optimization. For this, data science knowledge and the skill to do predictive forecasting of possible scenarios using advanced numerical modeling techniques is essential. Concepts such SRM (Synthetic Rock Mass), explicit time marching advanced numerical tools such as UDEC, 3DEC, FLAC, FLAC3D, PFC, self-sensing technologies, virtual health monitoring and controlling, dynamic-yielding-self-healing support systems, etc. are essential inputs to the “Autonomous Mine” of the future. For this to happen, we require more Government-Academia-Industry linkages and partnerships. Acknowledgment. The authors would like to thank and acknowledge the contributions and support provided by SAGE (NUST), the WMI (WITS), and Sibanye-Stillwater for the WMI-SAGE collaboration. Also, the authors want to acknowledge the critical contribution of Professor Emeritus Asghar Qadir (NUST), who provided vision and wisdom as Patron of the collaboration.

References 1. ICMM: Metals and Minerals (2020). https://www.icmm.com/en-gb/metals-and-minerals 2. Albanese, T., McGagh, J.: Future Trends in Mining, Chapter 1.3 SME Mining Engineering Handbook (2017). https://www.scribd.com/document/376046924/future-trends-inminig-pdf 3. World Bank: World bank policy research working papers: competition and innovation-driven inclusive growth. In: Dutz, M.A., Kessides, I., O’Connell, S., Willig, R.D. (eds.) October 2001 (2011). https://elibrary.worldbank.org/doi/abs/10.1596/1813-9450-5852. Accessed 29 Dec 2018 4. Fairhurst, C.: Some challenges of Deep Mining. Published by Elsevier LTD on behalf of the Chinese Academy of Engineering and Higher Education Press Limited Company (2017). http://www.creativecommons.org/licenses/by-nc-nd/4.0/ 5. Pierce, M., Cundall, P., Potyondy, D., Mas Ivars, D.: A synthetic rock mass model for jointed rock. In: Eberhardt, E., Stead, D., Morrison, T. (eds.) Rock Mechanics: Meeting Society’s Challenges and Demands. Volume 1: Fundamentals, New Technologies & New Ideasp, pp. 341–349. Taylor & Francis Group, London (2007) 6. Department of Mineral Resources: Guideline for the compilation of a mandatory code of practice for the prevention of fires in mines. Mine Health and Safety Inspectorate, Department of Mineral Resources, Pretoria, February 2017 (2017) 7. Department of Mineral Resources: Annual Report 2017/2018. Mine Health and Safety Inspectorate, Department of Mineral Resources, Pretoria, 2018 (2018) 8. Department of Mineral Resources: Mine health and Safety Statistics 2019 released by Minister Gwede Mantashe. Mine Health and Safety Inspectorate, Department of Mineral Resources, Pretoria, 2019 (2019) 9. Department of Home Affairs Critical Skills List. from http://savacancies.co.za/critical-skillslist-in-south-africa/. Accessed 04 Aug 2020

Development of the Global 21st Century Mining Technical Services Professional

363

10. SOMP 2019: Mines of the Future Version 1.0. Saydam, S. (ed.). https://MiningProfs.org. Accessed 15 Jan 2020 11. Cawood, F.T.: Mining in the twenty-first century and its future of work: considerations for universities offering mining qualifications. Submitted to XIV International Research and Practice Conference Ukrainian School of Mining Engineering on Innovative technologies and digital systems in mining. 07 to 11 of September 2020. Paper reviewed and accepted for publication (2020, in progress) 12. MQA, Undated. Career Guide: Mining and Minerals Sector, Mining Qualifications Authority, Johannesburg, South Africa 13. Cawood, F.T.: GeoTech and 4IR: Disciplines Uniting for 21st Century Mining. Presentation at EE-Publishers SA GeoTech 2019, Emperors Palace, Johannesburg, 22 July 2019 (2019) 14. IISD: New Tech, New Deal: Technology impacts review. Intergovernmental Forum on Mining, Minerals, Metals and Sustainable Development. International Institute for Sustainable Development (IISD), Monitoba, Canada (2019)

Assessment of Feasible and Effective Technologies for the Chemical Utilization of Domestic Coal for Value-Added Production in Vietnam Michaela Scheithauer1(B) , Patricio E. Mamani Soliz1 , Roh Pin Lee1,2 , Florian Keller1 , Bernd Meyer1,2 , Xuan-Nam Bui3 , and Tong Thi Thanh Huong4 1 Institute of Energy Process Engineering and Chemical Engineering,

TU Bergakademie Freiberg, Freiberg, Germany [email protected] 2 Fraunhofer IMWS, Branch Lab Freiberg – Circular Carbon Technologies, Freiberg, Germany 3 Department of Surface Mining, Faculty of Mining, Hanoi University of Mining and Geology, Hanoi, Vietnam 4 Department of Chemical Engineering, Faculty of Oil and Gas, Hanoi University of Mining and Geology, Hanoi, Vietnam

Abstract. Vietnam is a country rich in coal resources. Currently, coal is mainly combusted for energy production. However, there is increasing interest to generate additional value from domestic coal via chemical utilization as feedstock for production of chemicals and/or transportation fuels. This article evaluated the chemical utilization of Vietnam’s coal via gasification for the production of syngas, and the subsequent synthesis of syngas via Fischer Tropsch (FT) technology for the production of FT diesel. A technology overview of coal gasification technologies provided insights into different types of gasification processes as well as a comparative evaluation of their advantages and disadvantages. Similarly, a review of FT technologies enabled a comparative technology overview of commercial FT reactors and associated processes, their advantages and disadvantages. Using a case study approach, the suitability of identified commercial gasification and FT technologies are evaluated based on their applicability for the conversion of high ash-containing Vietnamese anthracite with high melting temperature to produce FT-diesel. Evaluation results indicated that the Fixed Bed Dry Ash (FBDA) gasification technology in combination with the medium-temperature Fischer Tropsch (MTFT) synthesis would be the most advantageous technologies for the production of FT-diesel from Vietnamese anthracite. The importance of considering the gas loop and product recovery is also highlighted. Keywords: Vietnamese anthracite · Chemical utilization · Gasification · Fischer tropsch synthesis · Technology assessment

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 364–384, 2021. https://doi.org/10.1007/978-3-030-60839-2_19

Assessment of Feasible and Effective Technologies

365

1 Introduction Vietnam is a coal-rich country with 3,704 million tons (MMst) of proven coal reserves [1]. Although domestic coal is utilized mainly for energy production today, there is growing interest to generate additional value from domestic coal resources. This is observable from projects which have been implemented for alternative coal applications such as coal gasification. Via gasification, coal resources can be converted into a chemical feedstock for production of chemicals and transportation fuels rather than using it for energy production via combustion. Not only could it support value-added production from domestic coal resources, the alternative utilization of coal as a chemical feedstock via gasification – rather than as an energy feedstock via combustion – could also contribute to lowering Vietnam’s carbon footprint by reducing CO2 emissions associated with coal combustion. Moreover, it could also contribute to reducing the country’s dependence on imported materials for the chemical and transport sectors [2–4]. Predominantly, coal gasification projects in Vietnam utilized underground coal gasification (UCG) processes. With UCG, the actual gasification process occurs underground. Subsequently, synthesis gas (i.e. syngas) produced via gasification reactions are collected for further processing and utilization. For instance, underground coal gasification (UCG) and underground coal bio-gasification (UCBG) technologies are proposed for test implementation in the coal area Red River Basin at Thai Binh and Hung Yen [5]. Another project using Russian UCG technology is also recommended to Vietnam’s Dong Duong Corporation [6]. Following gasification, the syngas generated can be used for the production of diverse products. Via Fischer Tropsch (FT) synthesis – a Coal-to-Liquid (CtL) process – syngas from coal gasification can be subsequently synthesized to produce liquid products such as FT-diesel [7].This would provide an alternative coal-based fuel to conventional diesel for Vietnam’s transportation sector. Note that FT synthesis is an indirect coal liquefaction technology. Besides FT synthesis, other CtL technologies include coal pyrolysis technology and direct coal liquefaction technology [8]. These however are beyond the scope of this article. Currently, while CtL technologies are being investigated on a laboratory-scale in the country, no CtL projects have been implemented so far in Vietnam. Application of CtL technologies such as the combination of coal gasification with FT synthesis for the production of FT-fuels has considerable relevance for Vietnam. Not only could it contribute to increasing the domestic coal value-chain and generate new employment opportunities within the country, in providing a domestic alternative to conventional fuels, it could also decrease Vietnam’s dependence on fuel imports. In view of its strategic significance for Vietnam, this article assessed CtL technologies for the production of FT-diesel. Specifically, the investigation focused on gasification and FT technologies. The article has three main objectives namely (1) technology overview and assessment of gasification technologies, (2) technology overview and assessment of FT technologies, and (3) case analysis of suitability of diverse gasification and FT technologies for FT-diesel production from domestic coal in Vietnam. The article is structured as follows: First, a technology overview and assessment of gasification and FT technologies/processes is provided. Subsequently, the technologies/processes are evaluated according to their suitability for FT-diesel production from domestic coal in

366

M. Scheithauer et al.

Vietnam. A case study based on high ash-containing anthracite with high ash melting temperature formed the basis for the technology evaluation. The article then concluded with a summary of insights gained from the technology evaluation for the development of Coal-to-FT-diesel as alternative to conventional imported diesel in Vietnam.

2 Technology Overview Gasification and FT-synthesis technologies represent two key technological components in the Coal-to-FT diesel process chain (see Fig. 1). A technology overview of these two technological components is provided in this section.

Fig. 1. Simplified Coal-to-FT diesel process chain

2.1 Gasification Technologies Gasification as a Thermochemical Conversion Process. Gasification is the thermochemical conversion of a fuel (gasification feedstock) with a reactant (gasification agent) to a combustible gas (syngas). The primarily desired components of the produced gas are hydrogen and carbon monoxide. Other products include carbon dioxide, methane and higher hydrocarbons. There are different potential feedstock materials and target products (see Fig. 2). The main occurring reactions during gasification are:  C + 1 2 O2 → CO Partial oxidation C + H2 O → CO + H2 C + CO2 → 2CO CO + H2 O  CO2 + H2 C + 2H2 → CH4

Heterogeneous water gas reaction Boudouard reaction Homogeneous water gas reaction Heterogeneous methanation

(1) (2) (3) (4) (5)

CO + 3H2  CH4 + H2 O

Homogeneous methanation

(6)

2CO + 2H2  CH4 + CO2

Homogeneous methanation

(7)

Assessment of Feasible and Effective Technologies

367

Fig. 2. Potential feedstock materials and target products of gasification processes

For autothermal gasification, required reaction heat for the endothermic gasification reactions is provided by oxygen supply and subsequent partial oxidation. In allothermal gasification operations, the necessary heat (energy) is supplied from outside sources such as hot flue gas or electrical energy. The principles of allothermal and autothermal gasification are illustrated in Fig. 3.

Fig. 3. Autothermal vs. allothermal gasification

The gasification process takes place at temperatures in the range of 800 °C to 1800 °C. The exact temperature depends on the characteristics of the feedstock, in particular the softening and melting temperatures of the ash, and the gasification technology. There are considerable advantages for gasification under pressure, including savings in compression energy for the generated syngas and reduction of equipment size for the gasification and downstream gas treatment plant [9–13]. Consequently, practically all modern processes are operated at pressures of 10 bar to 100 bar.

368

M. Scheithauer et al.

Classification of Gasification Processes According to Bed Type. In practical realization of gasification processes, a broad range of reactor types and gasification systems have been developed and applied commercially. For most purposes, these reactor types can be classified by the type of contact between fuel and gasification agent into three categories namely moving-bed (also called fixed-bed), fluidized-bed and entrained-flow gasifiers (see Fig. 4) [9–13].

Fig. 4. Main gasification categories [14]

Moving-bed gasifiers (or fixed-bed gasifiers) are characterized by a bed in which the coal moves slowly downwards under gravity. The gasification agents such as oxygen and steam are introduced at the bottom of the reactor – either via a rotating grate for the non-slagging version of the gasifier (Fixed Bed Dry Ash (FBDA) gasifier) or via tuyère nozzles above the slag bath (British Gas/Lurgi (BGL) gasifier). In such a counter-current arrangement, the hot synthesis gas (syngas) from the gasification zone is used to preheat and pyrolyze the downward flowing coal. Therefore, the oxygen consumption is very low. Note that the presence of pyrolysis products in the product synthesis gas requires extensive tar separation steps. Furthermore, the syngas also contains significant amounts of methane. Even though high temperatures are achieved in the heart of the bed – as in the case of the BGL-Gasifier – the outlet temperature of the synthesis gas is comparably low (about 400 to 800 °C). Moving-bed processes operate on lump coal. Advantages of this feedstock preparation method include low cost preparation compared to fluidized-bed and entrained-flow gasifiers and the ability to use waste material as a co-feedstock. A disadvantage of moving-bed gasifiers is that an excessive amount of fines, particularly if the coal has strong caking properties, could block the passage of the up-flowing syngas.

Assessment of Feasible and Effective Technologies

369

Commercial technologies (Licensors): • Non-slagging FBDA gasification technology (SEDIN Engineering/China in Asia; Air LiquideEngineering&Construction(formerlyLurgi/Germany)inEurope),seeFig.5a. • SlaggingBGLgasificationtechnology(ZEMAG/ChinainAsia; Envirotherm/Germany inEurope),seeFig.5b.

Fig. 5. Moving-bed gasifiers: a) Non-slagging version according to the FBDA technology, b) Slagging version according to the BGL technology [10]

Fluidized-bed gasifiers convert crushed feedstock which is fluidized with a gas flow of gasification agents. It offers a good mixing between feedstock and oxidant, which promotes both heat and mass transfer. However, the good mixing of the bed simultaneously leads to the main disadvantage of the technology. The fresh feedstock and the converted material (mainly ash) cannot be extracted separately. Therefore, there will always be residual carbon in the discharged ash which limits the carbon conversion of fluidizedbed processes and requires further ash treatment. The operation of fluidized-bed gasifiers is generally restricted to temperatures below the softening point of the ash, since ash slagging will disturb the fluidization of the bed. Sizing of the particles in the feedstock is critical; material that is too coarse will settle down to the ash discharge and material that is too fine will tend to become entrained with the syngas at the top of the reactor. Such coal fines are usually partially captured in a cyclone or hot gas filter and returned to the bed. Furthermore, the operation window of the process that is defined by the introduced gas flow of gasification agents is narrow and needs to be carefully predicted in advance. The lower operation temperature of fluid-bed processes (800–1100 °C) means that they are more suited for gasifying reactive feedstocks such as low-rank coals and biomass and for feedstock with a high ash content. Due to the challenges in process handling and

370

M. Scheithauer et al.

the ash treatment, the worldwide employment of this technology in an industrial context remains limited. Examples of fluidized-bed technologies are presented in Fig. 6.

Fig. 6. Fluidized-bed gasifiers (examples) [10]

Commercial technologies (Licensors): • Agglomerating Fluidized Bed (AFB) gasification technology (Institute of Coal Chemistry, Chinese Academy of Sciences/China). • High-Temperature Winkler Technology (HTW) gasification technology (ThyssenKrupp Uhde/Germany). • Utility-Gas (U-Gas) gasification technology (Chicago Gas Technology Institute (GTI)/USA). • Transport-integrated gasifier (TRIG) (Kellogg Brown & Root (KBR)/USA). Entrained-flow gasifiers operate with pulverized or liquid (slurry) feedstock in cocurrent flow. The residence time in the gasifier is short (a few seconds). The feed is grounded to a size of 500 µm or less to promote mass transfer and facilitate transport in a dense flow feeding. Feedstock preparation is essential and also more expensive for this type of gasifier. Given the short residence time, high temperatures are required to ensure a good carbon conversion. Hence, all entrained-flow gasifiers operate in the slagging range (above 1200 °C). The high-temperature operation creates a high oxygen demand for this type of process. Entrained-flow gasifiers do not have any specific technical limitations on the type of coal used. However, coals with a high ash melting temperature or high ash content (>30%) will drive the oxygen consumption to levels where alternative processes may be economic advantageous. The generated gas is free of hydrocarbons, has a low amount of methane and needs less efforts for gas cleaning. Examples of entrained-flow technologies are presented in Fig. 7. Commercial technologies (Licensors): • Opposed multiple burner (OMB) gasification technology (Institute of Clean Coal Technology at the East China University of Science and Technology (ECUST)/China). • Hangtian Lu (HT-L) gasification technology (China Aerospace Science and Technology Corporation (CASC)/China).

Assessment of Feasible and Effective Technologies

371

Fig. 7. Entrained-flow gasifiers (examples) [10]

• Thermal Power Research Institute (TPRI) gasification technology (Thermal Power Research Institute (TPRI)/China). • Multicomponent slurry gasification (MCSG) technology (Chinese Northwest Research Institute of Chemical Industry/China). • Tsinghua two-stage oxygen gasification technology (Beijing Tsinghua University/China). • Pressurized entrained-flow (Prenflo) gasification technology (Shell and Uhde/Germany). • Siemens fuel gasification technology (Siemens Fuel Gasification Technology Freiberg/Germany – no longer traded). • General Electric (GE) Energy gasification technology (General Electric, formerly Texaco/USA). • E-Gas gasification technology (Lummus Technology/USA). • Mitsubishi Heavy Industries gasification technology (Mitsubishi Heavy Industries/Japan). • Pratt & Whitney Rocketdyne (PWR) gasification technology (Aerojet Rocketdyne/USA). • Choren Clean Coal gasification (CCG) technology (Choren/Germany). The main properties of the three gasifier categories are summarized in Table 1. Where special characteristics can be classified into very advantageous (++), advantageous (+) or less advantageous (−), the corresponding table cells are marked. However, in many cases this assessment cannot be generalized. For moving-bed (fixed-bed) gasifiers, the non-slagging (FBDA) and slagging (BGL) version of the gasifier should be considered individually. For fluidized-bed and entrainedflow gasifiers, the reactor types and process principles can significantly differ depending on individual solutions of different technology providers. Therefore, it is important to note that process selection is always a complex process taking coal properties, boundary conditions and desired product specifications into account.

372

M. Scheithauer et al.

Table 1. Comparison of different gasification processes (++: very advantageous, +: advantageous or −: less advantageous) Bed type Fuel: Particle size

Moving-bed (fixed-bed) Coarse-grained to lumpy 6 to 60 mm

& Preparation

Fluidized-bed Small-grained ++

None

0.5 to 10 mm

+

Crushing

Oxygen consumption

Low/moderate

++

Steam consumption

High/moderate

-

Ash melting (slagging)

Entrained-flow Pulverized / slurry

No (FBDA) /

Up to 0.5 mm

-

Grinding

Medium

+

High

-

Low

+

Very low or slurry

++

No

+

Yes

-

5 - 50 seconds

+

Yes (BGL) Residence time

15 - 30 minutes

-

High Carbon conversion

Medium ++

Up to 99% Raw gas temperature Synthesis gas ingredients

350 – 800 °C High in hydrocarbons (tar, PAH's); secondary cracking necessary

-

800 – 1000 °C Low in hydrocarbons; secondary treatment necessary

+

+

38 / 32 / 25 / 5

Typical ash

Fine-grained to glassy

Fine-grained

-

Up to 400 MW

Up to 500 MW

+

Main challenges

Clogging of bed / fine particles

Particle agglomeration; erosion; limited operation window

Process examples

FBDA, BGL

AFB, HTW, U-Gas, TRIG

1300 – 1500 °C Free from hydrocarbons

-

++

60/ 35 / 5 / < 0.1

20 / 38 / 28 / 12

+

++ Up to 99%

Typical components CO / H2 / CO2 / CH4

Single unit capacity

++

High -

95 to 96% ++

2 - 10 seconds

Granulated slag Up to 1000 MW Coal preparation; temperature control, oxygen demand OMB, HT-L, TPRI, MCSG, Tsinghua, Prenflo, Siemens, GE, E-Gas

++ ++

2.2 Fischer Tropsch (FT) Technologies FT Synthesis for Liquid Production. FT synthesis is a catalytic process based on the conversion of synthesis gas for the production of liquid hydrocarbons from coal, natural gas and other carbonaceous feedstock such as waste, biomass and CO2 . FT products are characterized by a very broad product spectrum containing up to 50 carbon atoms. While main products are generally transportation fuels (see Fig. 8), in some cases waxes and olefins for the cosmetic and chemical industry are desirable products as well.

Assessment of Feasible and Effective Technologies

373

Fig. 8. Inputs and outputs of FT synthesis including product processing

In chemical terms, this synthesis involves the polymerization of carbon monoxide in combination with hydrogenation. These reactions take place in a three-phase system: gas (carbon monoxide, hydrogen, steam and light hydrocarbons), liquid (hydrocarbons and waxes) and solid (catalyst). The most abundant class of produced compounds via FT synthesis are paraffins and olefins, whose proportion in the product spectrum depends directly on the applied catalyst, process conditions and reactor technology. Three different industrial applications have been developed over the years, namely high-temperature Fischer Tropsch synthesis (HTFT), low-temperature Fischer Tropsch synthesis (LTFT) and the latest medium-temperature Fischer Tropsch (MTFT) synthesis. In recent years, LTFT processes are predominantly used to obtain products with higher molar mass. These technologies are suitable for the production of diesel and waxes [13]. This trend towards LTFT synthesis and especially cobalt-based synthesis can be seen in the latest publications and studies on the techno-economic evaluation of process chains in which FT processes are involved [15–22]. MTFT synthesis produces a similar spectrum of products as LTFT. This synthesis is available for a production of fuels as main product. HTFT synthesis is a process variant mainly used for the production of chemicals (olefins) and gasoline. A co-production of long chains fuels together with chemicals and gasoline through HTFT synthesis is only possible with a suitable short olefin Oligomerization Unit. The following chemical reactions represents the main reactions during FT synthesis, which occur mainly in all industrial applications. The FT reactions are exothermic, producing an average heat of reaction of 10 MJ per kg of hydrocarbon product produced [23]. nCO + (2n + 1)H2 → H (CH2 )n H + nH2 O nCO + 2nH2 → (CH2 )n + nH2 O nCO + 2nH2 → H (CH2 )n OH + (n − 1)H2 O

Paraffins Olefins Alcohols

(8) (9) (10)

374

M. Scheithauer et al.

nCO + (2n − 1)H2 → (CH2 )nO + (n − 1)H2 O nCO + (2n − 2)H2 → (CH2 )nO2 + (n − 2)H2 O, n > 1 CO + H2 O  CO2 + H2

Carbonyls

(11)

Carboxylic acids (12)

Water gas shift

(13)

All these reactions can be catalysed by a variety of active metals capable of catalyzing them. These include among others, Fe, Co, Ni, Ru and Rh. Of these, only cobalt and iron have exhibited good results in industrial applications [24]. The FT-catalyst not only contains one active metal, it also includes several promoters and can be combined with a carrier [23]. The development of promoters and supports in catalysis plays an essential role in the intensification of the FT process and its transition from laboratory scale to commercial plant. Technical aspects such as catalysis, catalyst lifetime, FT crude chemistry, heat integration and process conditions, are of key importance when selecting an appropriate FT technology. The selection of a specific FT technology for an industrial/commercial application has a ripple effect on many of the design decisions [25]. Hence, this assessment focused on the applied metal catalyst, the composition of the FT crude, the quality of the FT crude, the quality of steam produced and the operating conditions of the process. Classification of FT-Reactors & Associated FT-Processes. In current commercial applications, four main multiphase reactors can be identified for FT synthesis (See Fig. 9). On the one hand, there are stirred flow reactors such as the Slurry Bubble Reactor Column (SBRC) and Fluidized Beds (CFB and FFB). On the other hand, there are plugged flow reactors such as the Multi Tubular Fixed Bed reactor (MTR) and Microchannel Reactor (MCR). FT reactors are typically catalyzed at 190–360 °C and 15–50 bar [26]. Subsequently, the different FT reactors and associated FT processes are briefly discussed and their respective commercial licensors specified.

Fig. 9. Slurry Bubble Column (a); Fluidized bed (b); Multi-tubular Fixed bed (c); Two bpd Microchannel plant (d) & Microreactor scheme (e). (a): taken and adapted from [26, 27]; (b) and (c): taken and adapted from [27]; (d): taken from [28] and (e): taken and adapted from [29].

Assessment of Feasible and Effective Technologies

375

Slurry Bubble Reactor Column (SBRC): Slurry reactors are three-phase systems, including solids (catalyst), liquid and gas fractions. The reactor consists of a vertical vessel with a gas distributor for the synthesis gas, a stack of heat exchanger tubes inside for steam generation and a cyclone on top of the vessel. The cyclone is necessary for the separation of the catalyst powder, which is carried in the gas stream. These technologies are developed to overcome the difficulties associated with multi-tubular reactors [30]. Catalyst wear (catalyst deactivation) as well as separation of the catalyst from the liquid and wax are the main challenges for these technologies. This is because for expensive catalysis such as cobalt in LTFT synthesis, wear is a significant negative aspect. Comparing prices of FT catalysts [31], iron catalyst are a thousand times cheaper than cobalt catalyst. Commercial technologies (Licensors): • Slurry Bed Process/Slurry Bed Reactor (SBP/SBR): Slurry technology for the Fe and Co-LTFT synthesis (SASOL/South Africa). • High temperature Fischer Tropsch process (HTSFTP): Slurry technology for the FeMTFT synthesis (Synfuels/China). Fluidized Bed Reactor (CFB and FFB): Fluidized Bed reactors are used for two-phase systems, where gas reacts on a solid catalyst surface. Normally, fluidized beds need a fluidization gas but in the case of FT synthesis this is not necessary because synthesis gas serves as medium for the fluidization [31]. This type of reactor is mainly used for the iron-based HTFT technology. Originally, the technology is based on a recirculation system (circulating fluidized bed CFB), which is further developed over the last years to a fixed fluidized reactor (FFB). The fixed fluidized bed has a simpler design than CFB and includes a gas distributor and a number of heat exchangers and cyclones for the separation of catalysts particles. The design of FFB is similar to the Slurry Bubble Reactor and is developed for the improvement of issues present in the Multi Tubular Fixed Bed Reactor. However fluidized beds present problems such as agglomeration and the formation of blockages that compromise the scale-up on this application [30]. Commercial technologies (Licensors): • Synthol: CFB technology for the Fe-HTFT synthesis (SASOL/South Africa). • Sasol advance Synthol (SAS): FFB technology for the Fe-HTFT synthesis (SASOL/South Africa). Multi-tubular Fixed Bed Reactor (MTR): This multiphase application is available for reactions in the gaseous, solid and liquid phases and is originally developed in Germany for Fe-LTFT synthesis [32]. The reactor contains a defined number of catalyst packed tubes with heat removal through steam generation on the shell side of the reactor. MTR is most robust of all commercial reactors and is the technology with the longest and most proven history of stable and reliable FT operation [23]. Operational problems include a relatively high pressure drop, low heat removal, low catalyst utilization, heat and material transport limitations (filling the pores of the catalyst with wax) and need for periodic replacement of the catalyst [30].

376

M. Scheithauer et al.

Commercial technologies (Licensors): • ARGE Reactor: MTR Technology for the Fe-LTFT synthesis (Lurgi/Germany and SASOL/South Africa). • Shell Middle Distillate Synthesis (SMDS): MTR technology for the Co-LTFT synthesis (SHELL/Netherlands). Microchannel Reactor (MCR): Chemical reactor units based on microchannel technology are characterized by parallel sets of microchannels in a range of less than millimeters (80% of ash content); • Checking possibility of applying air jigs as a supplement for FGX vibrating air tables (amount of grain class 0–6 mm affects separation efficiency in this separator in situation when amount of this grain class is higher than 15%) (Tables 14, 15 and 16).

Modern Methods of Dry Mineral Separation - Polish Experience

421

Table 14. Results of separation grain class 2–20 mm. Fraction

Yield [%]

Ash content [%]

Concentrate I

32.0

8.6

Concentrate II

30.3

9.9

Concentrate III

15.1

10.9

Middling

11.4

17.5

Tailings

11.2

78.2

Table 15. Results of separation grain class 12–20 mm. Fraction

Yield [%] Ash content [%]

Concentrate 55.2

4.8

Middling

29.3

7.2

Tailings

15.5

84.3

Table 16. Results of separation grain class 3–6 mm. Fraction

Yield [%] Ash content [%]

Concentrate 36.0

10.7

Middling I

37.3

14.7

Middling II 13.3

15.6

Waste

47.2

13.4

4 Industrial Applications and Patents Granted Research work carried out by Łukasiewicz Research Network - Institute of Mechanised Construction and Rock Mining led to the concept of creating a technological line based on air-vibrating separators. According to patent PL 223787 dry separation node will be used as supplement into typical coal processing plant. After the initial separation of coarse fractions and gangue, some of the finer fractions with a grain size of less than 100 mm are directed to dry enrichment to separate part of the gangue or a pure commercial product (additional enrichment in wet processes is required to reduce ash and sulphur) (Fig. 4). The main advantage of the proposed solution is the removal of a significant amount of gangue from the processing system of the processing plant, which reduces the load on wet enrichment machines, thus allowing to increase the amount of feed given to them. At the same time, by removing part of the gangue, the feed is averaged, directed to wet enrichment processes, by eliminating large fluctuations in stone content. It should also

422

W. Mijał et al.

Fig. 4. Block diagram of devices for dry enrichment process of energy mineral resources (patent PL 223787).

be noted that the smallest coal grains removed in the dust removal operation do not end up in the water circulation of the processing plant and thus do not create difficult to use sludge for economic.

5 Conclusions Dry separation methods are not so popular in Polish mining industry. After many years of using conventional wet beneficiation methods for mineral processing plants possibility to implement dry separator which are using typical gravity separation forces or optical separators with x-ray sensors are very difficult. Research projects leaded by Łukasiewicz Research Network - Institute of Mechanised Construction and Rock Mining shows potential in using described method as a supplement for wet enrichment processes (last years also research on removing toxic elements) [26–28]. Nowadays two fully equipped technological lines from private companies are using FGX-3 air-vibrating separators. Main task for these separators is to remove clean tailings before selling final coal products to small commercial customers.

Modern Methods of Dry Mineral Separation - Polish Experience

423

Simplified analysis of economic efficiency for a dry separation installation equipped with an air-vibrating separator FGX-3 operating in Poland carried out with the adopted assumptions showed the possibility of achieving a profit of 5.65 $/Mg, which on a monthly basis allows to get income in the amount of approx. 40,000 $. The number of months needed to return the investment [Payback Period] is 12.2 (the payback period of the capital invested expressed in years equals 1.02) [25]. To sum up, it should be stated that installations for dry separation of steam coal are characterized by high profitability and constitute an interesting business offer for hard coal distributors in Poland. The investigation for optic & X-ray sorting separators is ongoing. Results from research in zinc and lead ores shows potential of application but deposit resources are limited, and the mine will soon finish mining (probably at the end of 2020). Last possibility which is also still the main point of research in Polish conditions it to removing sulfur and mercury by using this beneficiation method. In connection with the new requirements of the European Union and the Polish Government regarding fossil fuels, and above all commercial products for small customers, it remains only a matter of time when dry coal deshaling installations will be more widely used. Conflict of interest. The authors declare that there is no conflict of interest.

References 1. Mijał, W., Blaschke, W., Baic, I.: Dry coal beneficiation method in Poland. Polish Min. Rev. 11, 9–18 (2018) 2. Tangshan Shenzhou Machinery Co., Ltd.: Company prospectus of Tangshan Shenzhou Machinery Co., Ltd., China (2012) 3. Baic, I., Blaschke, W.: Coal preparation in Poland – development trends for increasing production efficiency. In˙zynieria Mineralna – J. Polish Min. Eng. Soc. 2(40), 7–14 (2017) 4. Mijał, W., Tora, B.: Development of dry coal gravity separation techniques. In: Mineral Engineering Conference. IOP Publishing Conference Series: Materials Science and Engineering, p. 427 (2018) 5. Blaschke, W.: New generation of air concentrating tables. Zeszyty Naukowe Instytutu Gospodarki Surowcami Mineralnymi Polskiej Akademii Nauk – The Bulletin of The Mineral and Energy Economy Research Institute of the Polish Academy of Sciences, nr 84, pp. 67–74 (2013) 6. Baic, I., Blaschke, W., Szafarczyk, J.: The first FGX unit in the European Union. CPSI J. Mag. Coal Prep. Soc. India. VI(16). 5–12 (2014) 7. Baic, I., Blaschke, W., Buchalik, G., Szafarczyk, J.: The first FGX unit in the European Union. Thesis Collection of FGX Dry Coal Preparation Technology. Tangshan Shenzhou Manufacturing Co., Ltd., China, pp. 21–27 (2014) 8. Baic, I., Blaschke, W., Szafarczyk, J.: Dry coal cleaning technology. J. Polish Min. Eng. Soc. 2(34), 257–262 (2014) 9. Baic, I., Blaschke, W., Sobko, W., Szafarczyk, J., Buchalik, G.: Nowoczesne powietrzne stoły koncentracyjne do wzbogacania w˛egla kamiennego. Czasopismo Techniczne Krakowskiego Towarzystwa Technicznego, nr. 154–161, s. 3–9 (2014)

424

W. Mijał et al.

10. Baic, I., Blaschke, W., Góralczyk, S., Sobko, W., Szafarczyk, J.: Stanowisko badawcze do odkamieniania urobku w˛eglowego metod˛a suchej separacji. Czasopismo Techniczne Krakowskiego Towarzystwa Technicznego, nr. 154–161, s. 15–19 (2014) 11. Blaschke, W., Baic, I.: Wykorzystanie powietrznych stołów koncentracyjnych. Czasopismo Techniczne Krakowskiego Towarzystwa Technicznego, nr. 154–161, s. 20–22 (2014) 12. Blaschke, W., Szafarczyk, J., Baic, I., Sobko, W.: A study of the deshaling of Polish hard coal using an FGX unit type of air concentrating table. In: International Coal Preparation Congress in Saint Petersburg. Russia. pp. 1143–1148 (2016) 13. Baic, I., Blaschke, W., Sobko, W., Szafarczyk, J., Okramus, P.: Badania mo˙zliwo´sci usuwania kamienia z urobku w˛egla kokowego na powietrznych stołach koncentracyjnych. Monografia “Innowacyjne i przyjazne dla s´rodowiska techniki i technologie przeróbki surowców mineralnych” KOMEKO 2014, Wyd. ITG KOMAG, pp. 65–79 (2014) 14. Blaschke, W., Okramus, P., Ziomber, S.: Skuteczno´sc´ suchego odkamieniania w˛egla koksowego metod˛a separacji na powietrznych stołach koncentracyjnych. Monografia “Innowacyjne i przyjazne dla s´rodowiska techniki i technologie przeróbki surowców mineralnych” KOMEKO 2014, Wyd. ITG KOMAG, pp. 81–91 (2014) 15. Góralczyk, S., Blaschke, W., Kozioł, W., Sobko, W.: Wykorzystanie powietrznych stołów koncentracyjnych FGX do oczyszczania kruszyw naturalnych łamanych. Mining Science – Mineral Aggregates, 23(1), 37–46 (2016) 16. Baic, I., Blaschke, W.: Analysis of the possibility of using air concentrating tables in order to obtain clean coal fuels and substitute natural aggregates. Polityka Energetyczna – Ener. Pol. J. 16(3), 247–260 (2013) 17. Kołacz, J.: Advanced sorting technologies and its potential in mineral processing. AGH J. Min. Geoeng. 36(4), 39–48 (2012) 18. Kołacz, J., Wieniewski, A.: Advanced separation methods in the X-ray optical sorting systems. CUPRUM – Czasopismo Naukowo-Techniczne Górnictwa Rud, nr. 2(75), 137–146 (2015) 19. Wieniewski, A., Szczerba, E., Nad, A., Łuczak, R., Kołacz, J., Szewczuk, A.: Ocena mo˙zliwo´sci zastosowania nowoczesnych technik separacji do wst˛epnego wzbogacania rudy Zn-Pb. CUPRUM 2(75), 109–122 (2015) 20. Kołacz, J.: Advanced separation technologies for pre-concentration of metal ores and the additional process control. In: E3S Web of Conferences, vol. 18, p. 01001 (2017) 21. Mijał, W.: Research possibilities to improve the effectiveness of the recovery of coal substances from small coal sortiments. Master thesis, Faculty of Mining and Geology, Gliwice (2016) 22. Mijał, W., Tora, B.: Enrichment of small grain classes: Laboratory scale. In: Congress Proceeding Volume II - XIX International Coal Preparation Congress, Woodhead Publishing India. Pvt 13–15.11.2019 New Delhi Indie, ISBN 978-93-88320-19-1, pp. 47–53 (2019) 23. Baic, I., Blachke, W., Sobko, W., Szafarczyk, J.: Wdro˙zenie innowacyjnej technologii oczyszczania w˛egla kamiennego przy pomocy suchej metody wzbogacania (FGX) drog˛a obni˙zenia kosztów wytwarzania produktów handlowych w krajowym górnictwie. Monografia “Innowacyjne i przyjazne dla s´rodowiska techniki i technologie przeróbki surowców mineralnych” KOMEKO 2016. Wyd. ITG KOMAG (CD version), pp. 60–78 (2016) 24. Blaschke, W., Baic I.: FGX air-vibrating separators for cleaning steam coal – functional and economical parameters. In˙zynieria Mineralna – J. Polish Min. Eng. Soc. vol. XXI. NR 2(44) 19–26 (2019). 2019 r. ISSN 1640-4902 25. Buchalik, G., Motyczka, S., Szafarczyk, J., Baic, I., Blaschke, W.: Economic efficiency of the dry separation process: Polish experience. In: Congress Proceeding Volume II - XIX International Coal Preparation Congress, Woodhead Publishing India. Pvt 13-15.11.2019 New Delhi Indie, pp. 54–63 (2019) 26. Blaschke, W., Baic, I., Sobko, W., Biel, K.: Removal of sulphur from hard coal using the FGX concentration table. Bull. Min. Ener. Econ. Res. Inst. Polish Acad. Sci. 95, 137–144 (2016)

Modern Methods of Dry Mineral Separation - Polish Experience

425

27. Baic, I., Blaschke, W.: Preliminary studies on the reduction of mercury content in steam coal through the use of air-vibrating concentration table. Ann. Set Environ. Protect. 19, 480–496 (2017) 28. Dziok, T., Strugała, A., Baic, I., Blaschke, W.: Analysys of mercury content in the feed coal ´ and rejects for selected hard coal cleaning processes. Rocznik Ochrona Srodowiska – Ann. Set Environ. Protect. 21, 951–961 (2019)

Becher Method Application for Ilmenite Concentrates of Vietnam Tien Thuat Phung1,2(B) and Ngoc Phu Nguyen1 1 Faculty of Mining, Hanoi University of Mining and Geology, Hanoi, Vietnam

[email protected] 2 School of Metallurgy, Northeastern University, Shenyang, China

Abstract. Several methods used to extract iron from ilmenite concentrate to increase TiO2 content of ilmenite concentrates such as Benelite, Ishihara, Murso, Austpac, and Becher methods. Benelite method uses HCl for leaching of reduced iron from ilmenite. Auspac method uses hot HCl while Ishihara uses hot acid H2 SO4 in the leaching stage. Murso method uses 20% HCl acid under atmospheric pressure for the leaching of roasted ilmenite. The significant advantage of Benelite, Auspac, Ishihara, and Murso methods is high-quality product with TiO2 content of up to 95–97%. However, the use of acid, especially hot HCl, is the principal disadvantage of these methods. They may severely impact the surrounding environment and may cause many safety and health problems. These methods have not been applied in Vietnam yet. Unlike the above methods, the Becher method relies on the iron corrosion capability of the less-toxic NH4 Cl solution for separating iron from roasted ilmenite. Therefore, it is considered to be more environmentally friendly. This paper presents the preliminary study results on Becher method applied to ilmenite concentrates of BinhThuan province of Vietnam. From the ilmenite concentrates of Binh Thuan province with TiO2 content of about 52%, the study has produced artificial rutile with TiO2 content of over 85% and a byproduct superfine iron oxide powder that may be suitable for the use in pigment production. The study results show that Becher’s method is suitable for the treatment of the ilmenite concentrates of Vietnam. Keywords: Ilmenite processing · Becher method · Increase TiO2 content · Vietnamese ilmenite concentrate

1 Introduction Titanium is a metal with silver color. It has desirable properties of high-performance metal, including high resistance to corrosion, low density, high specific strength, and superior ballistic properties [5]. Titanium is very highly resistant to corrosion even in extreme conditions such as in seawater, aqua regia, and chlorine. This metal is widely used in most areas [10], and its application is being expanded to diverse fields of engineering. The newer generation of titanium alloys is recognized as being much stronger and lighter than the most widely chosen and used steels. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 426–435, 2021. https://doi.org/10.1007/978-3-030-60839-2_22

Becher Method Application for Ilmenite Concentrates of Vietnam

427

Vietnam titanium ores are mainly of beach placer ore type, in which the leading titanium bearing mineral is ilmenite. Currently, most titanium processing technological lines are mainly concentrated on the cleaning stage, at which the obtained ilmenite concentrates are low grades that make them unusable in many applications [4]. The processing of titanium minerals not only brings economic value but also creates more jobs for local workers and creates momentum for the research and development of specific products from titanium such as pigments, titanium metals, high-strength alloys, etc. Recently, some titanium mineral processing factories have been built, such as titanium slag refining. However, the actual economic efficiency is not high since high-temperature processes consume large amounts of high power [4]. In the world, there are some methods for the production of artificial rutile such as Becher method, Benelite method, and Austpac method, etc. Which method should be used, and which is effective and suitable for Vietnam’s current conditions in the context of strictly environmental pollution control? This report presents preliminary experimental results of a study on leaching of ilmenite concentrates of Vietnam using Becher method. It proposes some suggestions on the methods for processing titanium ores that are currently applied worldwide, thereby saving time and cost in choosing the orientation for the research deployed in Vietnam.

2 Evaluation of Processing Methods Ilmenite processing is divided into two broad groups, which are pyrometallurgical and hydrometallurgical methods. Titanium slag smelting is a pyrometallurgical method. Hydrometallurgical methods include Benelite, Austpac, Ishihara, Murso, Becher methods, etc. Here are some outstanding features of the methods. 2.1 Titanium Slag Smelting This method, developed by Russian, is widely applied globally, particularly in Canada, South Africa, Norway, Ukraine, Japan, and China. According to this method, ilmenite concentrate is mixed with carbon and additives. The mixture is then pressed into briquettes. The briquettes are then smelted in an electric arc furnace. At high temperatures (about 1600 °C), iron is reduced to metal form Fe, melted to form pig iron, while TiO2 is left almost unchanged and thus transferred into slag. As a result, two products are obtained, which include pig iron and rich in titanium slag. The method’s significant advantage is that it does not require high TiO2 content of the feed; besides, the method also obtains valuable pig iron. However, the drawback of the method is high power consumption. According to the original technology, electricity consumption is about 2,700 kWh/tonne. Even today, the two-stage smelting technology has improved electricity consumption, but it still requires about 1,000 kWh/tonne of titanium slag. This method also causes severe pollution to the surrounding environment due to a large number of emissions during the production process (Fig. 1).

428

T. T. Phung and N. P. Nguyen

Ilmenite Bentonite

Pelletizer

Pre-reduction Magnetic separation Electric furnace Smelting Semi - steel

Ti slag

Desulphurization Carburization Iron casting machine Pig iron ingot Fig. 1. Titanium slag method flowsheet [3]

2.2 Hydrometallurgical Methods for Processing of Titanium Concentrates The Benelite Method The Benelite method has overcome operational disadvantages associated with the direct acid leaching of ilmenite [6]. This method is used in several operations, including Malaysian Titanium Corporation’s plant in Ipohand Kerr McGee’s plant in Alabama. Partial reduction roasting of ilmenite is necessary before leaching because the resultant ferrous oxides are easier soluble in a leaching solution such as chloride acid. Moreover, the partial reduction of ilmenite significantly reduces acid concentration (20% HCl) [6]. This process includes the following stages: + Reduction roasting of ilmenite to convert Fe(III) to Fe(II). + Dissolution of the roasted product in HCl solution under low-pressure conditions. + Filtration of leached ilmenite to receive a solid cake richer in TiO2 content. This cake is then calcinated in order to obtain a calcined solid containing up to 94% TiO2 . The iron-containing filtrate follows the regeneration stage to recover HCl and Fe2 O3 simultaneously.

Becher Method Application for Ilmenite Concentrates of Vietnam

429

Austpac method Austpac is a method for the production of high-quality artificial rutile. The method was developed by Austpac Company, and the method was applied in production since 2001. The method consists of two main processes, which are ERMS (Enhanced Roasting and Magnetic Separation) and EARS (Enhanced Acid Regeneration System) [7]. The principle of the two processes can be summarized as follows: The sequence of the ERMS process for producing artificial rutile is as follows: reduction roasting → magnetic separation → dissolution in hot HCl solvent. After filtration and calcination, the final solid product is the artificial rutile with TiO2 content of up to 97% that is most suitable for commercial purposes. Unlike the ERMS process, EARS adds FeCl2 solution to recover HCl and Fe2 O3 as materials suitable for steel making. Artificial rutile produced by Austpac method have higher rutile quality (97–98% TiO2 ) in comparison to other methods. However, hot and dense acids create several issues related to equipment erosion and the environment. Ishihara Method Like Benelite method, ilmenite concentrates are roasted in rotary tube furnaces to convert Fe(III) into Fe(II). Magnetic separation is used for the removal of any residual coke from partially reduced ilmenite. The product is then dissolved in an H2 SO4 solvent at a temperature of 130 °C for several hours. Artificial rutile contains up to 95% TiO2 [8].The leaching step’s reject acid solution can be used as a raw material in an ammonium sulphate plant or it may be re-used by adding make-up acid. Murso Method Before reducing ilmenite at 800–850 °C, the ore concentrate is oxidized at 900–950 °C to convert Fe(II) to Fe(III). The roasted product is dissolved in 20% HCl acid under atmospheric pressure. Impurities such as Mn, Mg, Al and V are also dissolved during dissolution. The obtained rutile contains 95% TiO2 [9]. The method has been proved to be efficient in a pilot-scale [6]. Becher Method Becher method, named in honor of Robert Gordon Becher (Australia), results from the extensive studies on the enrichment of ilmenite concentrates by selective roasting of ilmenite and subsequent corrosion of iron in NH4 Cl solution under air bubbling [10]. This technology’s final products include artificial rutile with TiO2 content of around 90% and iron oxide powder. In 1961 Becher was granted a U.S patent. This method’s main disadvantages are the requirement of high-quality ilmenite feed (TiO2 ≥ 55%) [2], and the processing time is relatively long. 2.3 Evaluation and Selection of Processing Methods As mentioned, titanium slag smelting has the main disadvantage of consuming immense power. Therefore it is only suitable for countries where power sources are abundant and cheap, but not the case for Vietnam. The hydrometallurgical methods such as Benelite, Austpac, Becher, Ishihara, Murso, etc. allow processing of ilmenite concentrates with varying quality in a wide range, relatively complete iron removal, and relatively low

430

T. T. Phung and N. P. Nguyen

power consumption in comparison to slag smelting method, the obtained artificial rutile with high TiO2 content. Except for the Becher method, most hydrometallurgical methods use leaching to solve relatively high concentrations of acidity. Therefore, the need to use high corrosion-resistant and high-pressure equipment became the major disadvantage of these methods. In practice, the applications of these hydrometallurgical methods are mainly limited in laboratories and some pilot-scale use, for example, Austpac method. Currently, Becher method is widely used in industrial-scale in Western Australia, India, and China to produce artificial rutile. This method has many advantages as it is a fast and straightforward process with the environment [2]. The method also requires less energy. Becher method is summarised, according to Fig. 2. This method is based on the possibility of electrochemical corrosion of Fe metal in NH4 Cl solution [1]. The difference in Ilmenite concentrate

Coal Reduction roasting

Magnetic Separator

Oxygen/ozone mixture

NH4Cl NH4Cl make-up

NH4Cl recycle

unburned coals

Leaching

Iron oxide

Gravity sedimentation H2SO4

Thickener Axid leach Iron oxide Filter

Fig. 2. Becher method flowsheet

Synthetic rutile

Becher Method Application for Ilmenite Concentrates of Vietnam

431

the electric potential of oxygen and iron causes the following electrochemical reactions: Fe = Fe2+ + 2e−

(1)

O2 + 2H2 O + 4e− = 4OH−

(2)

In normal conditions, ion Fe2+ are easy to combine with OH− to form Fe(OH)2 that may be deposited on ilmenite particles’ surface and retard further dissolution of Fe from ilmenite. Thanks to the actions of NH4+ and Cl− ions, the formation and deposition ofiron hydroxide on the surface of ilmenite particles are minimized [1, 2]. NH4+ ions dissociated from NH4 Cl reacting with hydroxyl ions according to the reaction: NH4+ + OH− = NH3 + H2 O

(3)

Then, NH3 from the above reaction acts with iron ions to form iron-containing 2+ complexions. Complex Fe(NH3 )2+ x ions transport Fe from the ilmenite surface to the solution and then decompose to form superfine iron oxide or hydroxide particles [1]. As a result, the process is not only separating iron from ilmenite particles to form artificial rutile but also creating a byproduct of superfine iron oxide that can be used as a raw material to produce pigments. According to the above analysis, based on actual electricity, ore reserves, and quality of ilmenite in some mining areas in Vietnam and cared for environmental issues, the Becher method is considered highly feasible in both technical and economic sense also to be an environment-friendly method.

3 Results and Discussions Several experiments were conducted on the ilmenite concentrates of placers of Binh Thuan province to verify the Becher method’s ability. Diffraction analysis of the ilmenite concentrate samples (Fig. 3) show that the main components of concentrates include: ilmenite (FeTiO3 ), pseudorutile (Fe2 Ti3 O9 ), small amounts of rutile (TiO2 ), and hematite (Fe2 O3 ).

Fig. 3. XRD analysis image of ilmenite concentrate sample

432

T. T. Phung and N. P. Nguyen Table 1. Composition of ilmenite concentrate sample (by weight) Oxide % by weight Oxide % by weight TiO2 Al2 O3

52.18

CaO

0.20

0.95

MgO

0.36

Fe2 O3 30.40

MnO

2.77

FeO

11.52

Cr2 O3 0.11

SiO2

1.50

ZrO2

0.33

P2 O5

0.04

SO3

0.36

The analytical results also showed that, in the composition of Binh Thuan ilmenite concentrate, iron existed in ilmenite in both forms of oxide FeO and Fe2 O3 (Table 1). Fe2 O3 accounts for 72.52% of total iron oxide, bringing a benefit because it becomes easier to reconstitute than FeO [1]. The ilmenite is also quite fine, and the size fraction 0.1–0.2 mm was predominant (Table 2). Therefore, there is no need to use the grinding process for this concentrated sample. Table 2. The size distribution of the ilmenite concentrate sample Size (mm)

>0.5 0.5–0.2 0.2–0.1 5 years) about silicosis. It is because the longer people work in a working environment, the more they are exposed to information related to their work, including silicosis. Sources of information could be found through communication programs, health care workers, or senior colleagues. The longer time they work, the more opportunities workers have to see their colleagues with illness and can urge them to learn more about the disease. Thus, for silicosis prevention to be effective, it is necessary to pay attention to advanced interventions on attitude knowledge and practice of entry-level workers to help ensure a high-quality labor force. 4.3 Smoking and Silicosis This study found the link between cigarette smoking and some attitudes toward silicosis. Smokers were more likely to have subjective attitudes than non-smokers when most of them supposed that silicosis was not a severe disease and could be cured. Smokers were

482

V. Nguyen et al.

also less interested in practicing silicosis prevention than non-smokers. The explanation is that smokers are inherently ignoring warnings about possible risks related to smoking, so they also tend to disregard other illnesses. Souza TP’s study (2017) showed that the incidence rate of silicosis of workers who were smoking or had a smoking history was 1.85 times higher compared to the non-smokers’ group (95% CI: 1.41–2.43; p < 0.001) [8]. It is, therefore, critical to warn workers about the harmful effects of tobacco. 4.4 The Importance of Regular Health Checkups Most of the entire workers performed annual health checkups (93%). That can be explained by the fact that periodic health checkup has been included in the regulations of Vietnam Labor Code to ensure the benefits of employees [9]. These results were consistent with those found in a study implemented in Shanghai, China, on the prevention and treatment of silicosis in 2015, with the rate of annual medical examination being 92.4% [10]. The periodic health check is to ensure that the employees have good health when entering the workforce. Besides, in specific work environments where high risks of occupational diseases occur, companies need to continue working with the health authorities to maintain periodic health checks, along with regular checkups of occupational health. Thereby work productivity is enhanced and contributed to the companies’ development. 4.5 Frequent Use of Facemasks As silicosis is a respiratory disease, wearing masks is one of the essential preventive measures to prevent workers from direct contact with silica dust. The rate of facemask used in this study (88%) was higher than that in China study, where it showed just 56.7% among the total of 302 participants [10]. Research has also pointed out an association between the right knowledge and attitude on silicosis with regular use of masks. Workers with good knowledge and attitudes are more likely to be cautious with this disease and understand the importance of preventive measures, such as wearing masks in the work environment. In 2008, Takemura Y. et al. conducted a study on the effects of wearing facemasks and educating workers on the prevention of occupational dust exposure, showing that these interventions could help prevent respiratory function decline among the workers exposed to dust [11]. In addition to that, the study indicated the lack of knowledge among workers about whether the disease was cured. Only 26% were adequately aware that this was an incurable disease. Prevention can be the only treatment for silicosis as there are still no specific drugs to tackle this illness [1]. Therefore, health education should be further implemented, especially paying more attention to the clarification of the harmful effects of silicosis when inhaled, which causes fibrous lung disease and that silicosis is an incurable disease. 4.6 Other Prevention Practices When being interviewed about ways to reduce dust level in the working environment, three given categories such as water spray on the factory floor, clean factory floor, vacuum

KAP on Silicosis Among High-Risk Worker Population in Five Provinces

483

ventilation fan were chosen with the rate of 46%, 55%, 47%, respectively. These results were much lower compared to the measure of using facemasks. It can be explained that employees were aware of properly wearing masks as the easiest and most effective way to prevent direct contact with dust through the respiratory tract, which makes this method constitutes the highest proportion. Moistening the environment and ventilating are two measures that depend on the company’s resources affecting the production process, which is more challenging to undertake. As such, employees were not well aware of that, and as a result, the selection rates for these measures were relatively low. This finding suggests that employers should be more responsible for the prevention of silicosis and for creating a safe environment for their employees [9]. 4.7 Strengths and Limitation of the Study Research on knowledge, attitudes, and practices of workers on the prevention of silicosis in Vietnam and the world has still been inadequate. The study described the workers’ knowledge, attitude, and practices about silicosis and identified some KAP factors and some epidemiological characteristics of workers. The underlying cause of KAP rates among workers and its association with silicosis has not yet been investigated. That needs to be further studied.

5 Conclusion Overall, the majority of workers had the correct knowledge and attitudes on silicosis. However, only 26% of them had a proper awareness that it was an incurable disease. As for practice, the paper pointed out that the majority of workers have experienced annual health checkups (93%) and used facemasks (88%). Other preventative measures, for example, to reduce dust in the working environment, were not so common in practice. The study found that better knowledge, attitude, and practices were associated with higher education levels, longer duration of work, and non-smoking. Therefore, health education should be prioritized for workers, particularly smokers, entry-level workers, or those with lower educational levels. Acknowledgment. The authors thank the Centres of Diseases Control/Preventive Health in 5 provinces, including Hai Duong, Thai Nguyen, Binh Dinh, Dong Nai, and Phu Yen, for their supports during the field visits.

Funding. The research is supported by the Ministry of Science and Technology, Vietnam in the Program: Research on applying and developing advanced technology for public health protection and care code KC.10/16-20. Ethics Approval and Informed Consent. This study was approved by the Ethics Committee of Hanoi Medical University (Code number: 4218/HMU-IRB dated November 16, 2018). Following a one-on-one explanation of the study from trained healthcare workers at the Hanoi Medical University, all participants were given a written informed consent before they participated in the study, acknowledging their full understanding of the study’s purpose, their rights to withdraw from the study at any time, and the protection and confidentiality of the collected data.

484

V. Nguyen et al.

References 1. Duy, K.V., Anh, L.M.: Occupational Diseases. Medical Publishing House, Hanoi (2017) 2. Dal, M., Malak, A.T.: Effects of SiO2 in Turkish natural stones on cancer development. Asian Pac. J. Cancer Prev. 13(10), 4883–4888 (2012) 3. Fedotov, I.: The ILO/WHO global programme for the elimination of silicosis. Occup. Health Southern Africa (2006) 4. Select Research (Pvt) LTD: Knowledge, Attitudes and Practises (KAP) on TB, HIV and silicosis among key populations Aged 15 and 59 years in Southern Africa (2017) 5. Yadav, S.P., Anand, P.K., Singh, H.: Awareness and practices about silicosis among the sandstone quarry workers in desert ecology of Jodhpur, Rajasthan. India. J. Hum. Ecol. 33(3), 191–196 (2011) 6. Aggarwal, B.D.: Worker education level is a factor in self-compliance with dust-preventive methods among small-scale agate industrial workers. J. Occup. Health 55(4), 312–317 (2013) 7. Falk, L., Bozek, P., Ceolin, L., et al.: Reducing agate dust exposure in Khambhat, India: protective practices, barriers, and opportunities. J. Occup. Health 61(6), 442–452 (2019) 8. Souza, T.P., Watte, G., Gusso, A.M., et al.: Silicosis prevalence and risk factors in semiprecious stone mining in Brazil. Am. J. Ind. Med. 60(6), 529–536 (2017) 9. Law no. 84/2015/QH13 dated June 25, 2015 on Occupational safety and hygiene in Vietnam (2015) 10. Wang, L., Liu, X., Yu, D., et al.: [Current situation of prevention and treatment of silicosis in Jinshan District of Shanghai, China]. Zhonghua Lao Dong Wei Sheng Zhi Ye Bing Za Zhi Zhonghua Laodong Weisheng Zhiyebing Zazhi. China J. Ind. Hyg. Occup. Dis. 33(6), 456–458 (2015) 11. Takemura, Y., Kishimoto, T., Takigawa, T., et al.: Effects of mask fitness and worker education on the prevention of occupational dust exposure. Acta Med. Okayama 62(2), 75–82 (2008)

The Impact of Afforestation on Seepage Water Formation on Post-mining Spoil Heaps and Dumps - Results of Water Balance Modeling Christian Hildmann(B) , Lydia Rösel, Beate Zimmermann, Dirk Knoche, and Michael Haubold-Rosar Research Institute for Post-Mining Landscapes, Finsterwalde, Germany [email protected]

Abstract. Mining sites like spoil heaps and dumps often have a geochemical composition that causes harmful effects on the environment - especially on ground and surface water. Water collection and treatment require elaborate technologies and high expenditures. The reduction of seepage water by afforestation, which leads to high evapotranspiration rates, offers an alternative to mitigate negative environmental impacts. To estimate such effects, we investigate the water regime of different post-mining tips and dumps in Germany: potash tailings piles with high salt contents in Thuringia, a slagheap of hard coal mining with heavy metal leaching in Saxony and a dump in the Lusatian lignite mining district, causing the discharge of iron sulfates. We simulate the seepage water formation with the Bowahald and LWFBrook90 software. Up to now, only 10% of the surface area of the potash tailings piles are covered by woody plants. Complete afforestation of the five potash tailings piles would reduce seepage water by 44%. On the sites in Saxony and Lusatia we modeled the water regime for different forest types. The results show differences between the sites characterized in terms of soil properties and between the selected tree species compositions. Hence, we recommend a mixture of tree species to keep a stable forest on spoil heaps and dumps and protect the groundwater at once. Keywords: Mining · Water balance modelling · Spoil heaps · Afforestation · Seepage water

1 Introduction Mining often results in dumps on which overburden or residues are deposited. The material properties affect the environment, especially ground and surface water quality. The subsequent treatment of contaminated mine drainage water is costly or, in the case of groundwater treatment, hardly practicable. For this reason, the heaps are often covered with suitable materials. The vegetation on this coverage protects against erosion and evaporate a part of the precipitation water. As well-known, in the case study region forests have a much higher evapotranspiration rate than herbaceous vegetation or grasses. The © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 485–497, 2021. https://doi.org/10.1007/978-3-030-60839-2_26

486

C. Hildmann et al.

contribution of a dump protection forest to the reduction of leachate depends on the site conditions. In this paper, we investigate the following questions based on case studies: • By what proportion can forests reduce the seepage water amount from waste rock piles in comparison to partial or herbaceous cover? To this end, we are investigating five potash tailings piles in the Südharz district of Thuringia. • What differences are to be expected between different dump protection forests? We modeled the accumulating seepage water for two other forest covers of a hard coal dump and a lignite dump in Saxony.

2 Sites and Methods 2.1 Tailing Piles of Potash Mining For the first part, we investigated five tailing piles of potash mining. They are located in the Südharz region of Thuringia. Over there, in Sondershausen (since 1897), Bleicherode, Rossleben, Sollstedt, and Menteroda (since 1902) and Bischofferode (since 1911, not considered here), the residues of potash fertilizer production have been deposited (Fig. 1). The facilities were closed between 1990 and 1993. About 350 hectares of tailings piles with a volume of 135 million m3 remained.

Fig. 1. Map of the tailings piles, mine dumps, and slag heaps under investigation.

The tailings piles mainly consist of easily water-soluble salts, above all NaCl (about 75% halite). Due to the high salt concentrations, their leaching has an adverse effect

The Impact of Afforestation on Seepage Water Formation

487

on the watercourses nearby (Bode, Wipper, Lower Unstrut) and the groundwater. The tailings piles are covered with a contour layer, which consists of construction waste and other inert wastes and a final reclamation layer (main rooting zone) to reduce the leachate. However, the covering has not yet been completed. First, the surface of every potash tailings pile was divided into hydrologic response units (HRUs), which are assumed to be homogeneous regarding their hydrologic behavior because they have unique cover, exposure, and vegetation. After the evaluation of aerial photographs and digital elevation models provided via LIDAR remote sensing, we recorded all sub-areas in the terrain. In doing so, the state of coverage, divided in uncovered, provisionally covered, and fully covered areas were documented. Furthermore, we recorded the vegetation according to its percentage of coverage with herbaceous vegetation, shrubs, and trees. On each tailings pile, we investigated the topsoil of the different cover classes by randomly distributed samples using pedological field methods. Besides, the saturated hydraulic conductivity of the topsoil layer was determined in situ using an Amoozemeter (well permeameter, [1]). For simulating the seepage water volumes, we used the Bowahald model [2]. The parameterization was carried out with the collected field data. For the required climate data, we used the surrounding weather stations of the German Meteorological Service (DWD) and interpolated them to the five tailings pile sites. We also used the daily values for each of the HRUs individually. The leachate formation was eventually aggregated again to the total heap. Because we simulated 13 hydrological years - from 2000 to 2014 - a range of the expected leachate formation can be given in addition to the mean value. In total, we modeled three different variants: the leachate formation of the tailings piles in their current state, a temporary greening of the areas which are not covered yet, and fully covered tailings piles with a protection forest. 2.2 Legacies of Hard Coal and Lignite Mining The second task considered two other mining legacies quite typical for the region: The first is a tailings pile called “Vertrauenschacht”, which was created from tailings and other residues of the coal mining industry. It is located between Lugau and Ölsnitz in the Saxon Ore Mountains and was operated between 1856 and 1936. Some areas were also flushed with furnace ash and sludge. Today the slag heap is covered with a succession forest dominated by birch trees. In the groundwater measuring stations nearby the slagheap, strongly elevated concentrations of hazardous heavy metals such as Cd, Ni, and Zn were found. Therefore, the question is whether an alternative forest community close to a regional climax forest could reduce the seepage water formation compared to the current state. We divided the slagheap into several HRUs. At randomly selected sampling locations within each HRU, we recorded the most important parameters for the soil water balance of the uppermost 60 cm using field pedological methods. In addition, the saturated hydraulic conductivity was determined by an Amoozemeter. To record the current stock of trees and shrubs, several sample circles were distributed over the slagheap, and the parameters essential for modeling were recorded. A time series of the meteorological data (daily values) was provided to us by the Saxony State Office

488

C. Hildmann et al.

of Environment, Agriculture, and Geology (LfLUG). The data was prepared within the framework of regionalized climate forecasts for various models runs from 1960 to 2100. It is based on the scenario RCP2.6 run 1, but in this article, we only consider the current development until 2020. We used the R software [6] and here, the library LWFBrook90R [7], to run the onedimensional hydrological model LWFBrook90. It is based on the approach of Federer [3], and was further qualified by the Bavarian State Institute of Forestry and Forest Management [4]. Unlike the Bowahald model, this model has not been developed for dumps and landfills. However, it allows a much more detailed parametrization of the vegetation. Thus, we first modeled the water balance of the current succession forest, divided for the different HRUs. The succession forest is then compared to a forest type determined by forest expertise on the basis of the site characteristics of the HRUs (Table 2). For this “alternative” forest, we also modeled the seepage water formation. The second location is situated in the post-mining landscapes of the dump of the Nochten opencast lignite mine, which is still in operation. This location was prepared for reforestation at the time of our study but is not afforested yet. Therefore, we have considered the reforestation plans of the mining company as the basis for the first modeling approach. Alternatively, we have chosen other forest compositions based on the site conditions (Table 3). A detailed soil map was available from the mining company LE-B. With this data, we parametrized the soil. Based on the soil data, we also subdivided the area into HRUs. We supplemented the data with additional measurements of the saturated hydraulic conductivity. For this site, we also received a time series of meteorological data from LfLUG.

3 Results 3.1 Tailings Piles of Potash Mining The mapping of the potash tailings piles shows that the vegetation strongly depends on the coverage with foreign material. Over the years, a loose weathering zone (regolith) forms on the uncovered areas, from which the easily soluble salts are primarily washed out. The regolith consists of more than 90% gypsum, to a smaller extent of anhydrite and other minerals. Due to the loose stratification and the lack of water storage capacity, a close vegetation cover is missing even after decades. 16% of the piles’ surface is uncovered. In the past, the material was provisionally dropped down from a higher position in some areas. However, the thickness of this type of cover is limited and decreases downhill. It is because these areas are not passable by vehicles, and therefore, the material may not be driven over with bulldozers. This accounts for about 12% of the regions (as of 2015). On these sub-areas, we found mainly herbaceous vegetation because the water storage capacity is very limited. The coverage that is common today consists of a technical layer that is used to model the relief of the heaps to the desired final state. So far, about 43% of the piles’ surface is completely covered (125 ha). Construction waste and other inert material are used for this purpose. A re-cultivation layer provides suitable conditions for subsequent vegetation which is placed on top. Due to the high proportion of building rubble, the substrate is often hardened, has low dry bulk densities, and a high ratio of air voids. Nevertheless, the substrates are well-rooted. The hydraulic conductivity differs

The Impact of Afforestation on Seepage Water Formation

489

on a small scale due to the substrate differences. The remaining areas are access roads and earth walls (13%) or are currently underway to be covered (16%). Only about 3% of the piles’ surface is covered with trees and 7% with shrubs to date (status 2015). In total, only 10% of the areas are covered with woody plants (Fig. 2). We found considerable differences between the five tailings piles. The potash tailings pile Bleicherode, for example, already has 25% woody plant cover and the Sollstedt pile about 15%, while we recorded less than 5% cover with woody plants on the other piles. Grasses and herbs were found on 47% of the surface of the dump. These were either sown or had found themselves there. Up to 43% of the dump surfaces did not show any vegetation.

Fig. 2. Sum of coverage of the five potash tailings piles in hectares (present situation 2015)

Based on these results, we calculated the current seepage water formation for each of the HRUs using the Bowahald program. The results are shown in Fig. 4 as a map for the Sondershausen tailings pile (Fig. 3). In total, all five potash tailings piles produce about 600,000 m3 of leachate per year. For the Sondershausen tailings pile, we calculate a seepage amount of 86,500 m3 or 22% of the long-term precipitation average and 168,000 m3 = 40% for the tailings pile Sollstedt. We also examined an intermediate greening without the application of an additional covering or re-cultivation layer. Some shrubs such as elderberry (Sambucus nigra) or the North American black locust (Robinia pseudoacacia), which already grow spontaneously, are in principle suitable for an interim greening, but not a sustainable solution. Our results show that this variant is not very useful in the long term due to the lacking soil water storage. On average, the temporary greening reduced the current leachate volume by only 7% at the five potash tailings piles.

490

C. Hildmann et al.

Fig. 3. Current seepage water formation at the potash tailings pile Sondershausen as a share of precipitation for each HRU

The situation would change with a dump protection forest. Based on the special conditions of the site, such as the steep slopes, strong winds on the western zones, and the intensive sunshine on the south-facing areas, we developed some recommendations for the selection of site-adapted tree species. Native oak trees play a dominant role, which we supplemented with further admixture tree species, in more detail: sessile oak (Quercus petraea), noble deciduous woods, small-leaved lime (Tilia cordata), hornbeam (Carpinus betulus), wild service tree (Sorbus torminalis), mountain ash (Sorbus aucuparia), European beech (Fagus sylvatica) and to a small portion wild fruits for an ecological upgrading of forest edges. Overall, the derived forest types are in good agreement with the specifications of the Thuringian forestry authorities and the potential natural vegetation in the region on similar soil substrates. For the proposed forest cover, the model predicts a reduction of the leachate volume of the five heaps by more than 262,000 m3 per year. Compared to the current situation, the amount of seepage water could be reduced by 44%. The ratio of leachate in the precipitation fluctuates - depending on the precipitation height and the conditions of the heaps. Potash tailings pile fully covered and afforested with deciduous trees reduce the amount of seepage water from 22 to 8 and from 40 to 26% of the annual precipitation. 3.2 Comparison of Different Pile Forests Vertrauenschacht Slag Heap. Different soils have developed on the Vertrauenschacht dump so that the five HRUs are sometimes different (Table 1). For example, they have a

The Impact of Afforestation on Seepage Water Formation

491

Fig. 4. Current and final seepage water formation in the percentage of precipitation, given are the median values. The whiskers represent the range between minimum and maximum (source: [5])

very different skeleton proportion. The two HRUs 4 and 1 have the highest water storage capacity, HRU 3 is the lowest. The other two are in between. We modeled evapotranspiration of the Vertrauenschacht slag heap between 293 mm/a and 359 mm/an on average over the time series considered from 1960 to 2020 (Table 4). The differences reflect both the soil’s physical properties and the vegetation aspect. The comparison with the alternatively proposed oak mixed forest shows only minor differences. The mean values of evapotranspiration are between 295 mm/a and 352 mm/a. In contrast, the differences are more evident in the interception. Currently, these are between 48 and 64 mm/a. For the alternative forest cover, we calculated values between 63 and 79 mm/a. This can be explained mainly by the higher leaf area index. The values for transpiration for the current condition are between 170 mm/a and 227 mm/a. Compared to the alternative forest, they hardly differ in some HRUs. HRU 1, however, currently shows higher transpiration of 225 mm/a than the alternative forest type with 211 mm/a; for HRU 5 the difference is even more significant with 227 mm/a and 197 mm/a. In total, this results in different seepage water quantities. For the current mixed birch forest, the average is between 289 mm/a and 341 mm/a (Fig. 5). In comparison, the seepage water quantities for HRU 4 could reduce from 341 mm/a to 315 mm/a with a mixed oak forest (Fig. 6). In contrast, for HRU 5, we show an increase in the leachate volume from 298 mm/a to 320 mm/a.

492

C. Hildmann et al.

Table 1. Parametrization of the soil conditions found at Nochten mine dump and Vertrauenschacht slagheap Subarea/ HRU

Site Nochten

Vertrauenschacht

Ksat [b] (mm/d)

[a]

Soil texture

Percentage of skeleton (%)

1

sand with 5 to 25% lay

3545

14

2

sand with 0 to 25% clay

3545

14

3

sandy clay loam

2671

14

4

sand with 17 to 25% clay

3545

14

5

sand with 0 to 17% clay

3545

14

6

loamy clay

2671

10

1

silt, coarse sand

2607

12

2

silt, fine sand, silty clay

3852

47

3

fine and coarse sand

9467

63

4

silt, silty sand

9877

59

5

loamy sand, silt

8336

39

a

for Nochten, soil texture was deduced from the available soil map; for the Vertrauenschacht slagheap, the predominant soil textures (which we recorded most often at the sampling locations) b median for the subarea/HRU at 20 cm soil depth

Table 2. Parametrization of the forest vegetation at Vertrauenschacht slagheap. LAI = leaf area index, TAI = trunk area index, cur. = current afforestation plan, alt. = alternative afforestation

Main tree species HRU

Tree hight (m)

Max. LAI (m²/m²) Cur.

Alt.

TAI (m²/m²)

Cur.

Alt.

Cur.

Alt.

Cur.

Alt.

1

Birch

Oak

16.2

13.1

3.6

4.2

0.5

1.4

2

Birch

Oak

15.0

13.1

3.1

4.2

0.6

1.4

3

Birch

Oak

14.7

13.1

3.1

4.2

0.6

1.4

4

Birch

Oak

12.7

8.8

3.7

6.6

0.6

1.2

5

Birch

Oak

14.0

13.1

4.7

4.2

0.6

1.4

Mine Dump Nochten. For reasons of practicability, we have combined the 12 HRUs of the Nochten dump into six subareas. They mainly differ concerning the deposited sediment. Sandy substrates dominate, while in the sub-areas 3 and 6 loamy to clayey soils occur.

The Impact of Afforestation on Seepage Water Formation

493

Table 3. Parametrization of the forest vegetation at mine dump Nochten. LAI = leaf area index, TAI = trunk area index, cur. = current afforestation plan, alt. = alternative afforestation Tree height (m)

Main tree species Site

Max. LAI (m²/m²) Cur.

TAI (m²/m²)

Cur.

Alt.

Cur.

Alt.

Alt.

Cur.

Alt.

1

Oak

Pine

16.5

17.0

2.3

2.7

1.0

2.3

2

Pine

Pine

11.6

17.0

3.3

2.7

2.5

2.3

3

Pine

Pine

11.6

17.0

3.3

2.7

2.5

2.3

4

Birch

Oak

20.5

16.9

1.8

3.2

1.1

1.5

5

Birch

Pine

20.5

17.0

1.8

2.7

1.1

2.3

6

Birch

Oak

20.5

16.9

1.8

3.2

1.1

1.5

Table 4. Model results for the partial areas of the Nochten mine dump and the HRU of Vertrauenschacht slagheap. All values are mean values (1960–2020) given in mm/a (ET = evapotranspiration, T = transpiration, I = interception, SW = seepage water).

Site

Current Forest ET

T

Alternative Forest I

SW

ET

T

I

SW

212 201 191 207 209 192

51 51 51 46 51 46

159 166 166 205 161 232

211 192 167 198 197

63 63 63 79 63

289 332 340 341 298

Nochten Mine Dump (precipitation 680 mm/a) T1 T2 T3 T4 T5 T6

293 302 293 273 278 260

195 212 199 177 178 170

34 57 57 32 32 32

221 148 152 241 235 262

301 293 287 299 299 280

Vertrauenschacht Slagheap (precipitation 841 mm/a) H1 H2 H3 H4 H5

359 319 293 310 344

225 195 170 197 227

52 49 48 53 64

293 322 332 315 320

352 321 295 326 322

In the current reforestation planning, pure stands of the tree species are planned (Red oak, Common Birch, and Scots pine, see Table 3). Regarding the common trend towards near-natural forest management on reclaimed land, the alternative proposals for afforestation are always mixed stands, where additional deciduous tree species are considered.

494

C. Hildmann et al.

Fig. 5. Seepage water formation at Vertrauenschacht slagheap (current forest), calculated yearly values from 1960 to 2020 for the five HRUs.

On average, we modeled values for evapotranspiration between 260 mm/a and 302 mm/a for the partial areas of the currently planned forest (Table 4). The model results for the alternative scenario show a range between 280 mm/a and 301 mm/a. This

The Impact of Afforestation on Seepage Water Formation

495

Fig. 6. Seepage water formation at Vertrauenschacht slagheap (alternative forest), calculated yearly values from 1960 to 2020 for the five HRUs.

becomes somewhat clearer when interception and transpiration are considered separately. The interception is higher in fully-stocked Scots pine forests than in areas cultivated with deciduous trees. The interception of pure birch stands is 32 mm/a on average, but 57 mm/a in pine forests. In the alternatively planned mixed forests the values lie

496

C. Hildmann et al.

in between. The sessile oak-dominated woodland shows an interception of 46 mm/a, and the mixed pine stands by 51 mm/a. The calculated values for transpiration depend primarily on the chosen tree species. We found the highest values for the pure Scots pine stocking with 212 mm/a, followed by the sessile oak mixed forest with 207 mm/a. The pure birch stand on sub-area 6 shows the lowest value (170 mm/a). We determined the lowest seepage water formation for the pristine pine forest on sub-area 3 at 152 mm/a. However, the highest seepage rates are expected under the pure birch forest on sub-area 6 with 262 mm/a – a difference of 110 mm/a after all. Mixed sessile oak forest on subplot 6 could at least reduce the seepage rate to 232 mm/a.

4 Discussion and Conclusions 4.1 Tailings Piles of Potash Mining The model results show that a complete covering and reforestation of the heaps can achieve a significant reduction of leachate volumes. Nevertheless, a constant leachate flow remains, which reaches the heaped body. This does not only lead to continuous pollution of ground and surface water. The dissolution of the salt also leads to uncontrolled subrosion, and thus the formation of sinkholes on the tailings piles. It is, therefore, desirable to reduce leachate formation even more. One possibility is the installation of an additional sealing layer, for example, in or below the technical layer. However, the availability of suitable materials and the associated costs of selective extraction and transport is problematic. Even covering the potash tailings piles takes a very long time due to the limited availability of materials. This applies in particular to the material with a very low water conductivity. However, large quantities of iron hydroxide sediments (EHS) accumulate in the Lusatian mining district. They are a consequence of lignite mining. The iron results from the iron sulfides in the subsoil which have been weathered by aeration. Via the groundwater, it reaches the surface waters where it oxidizes. During the maintenance of the water bodies, it accumulates in larger quantities, e.g. also in the Spremberg dam. Depending on their composition, EHS has a very low hydraulic conductivity. We carried out the first tests with the material and concluded that it is suitable. We, therefore, propose to test the further suitability in a field test. Such a sealing layer cannot build under already covered surfaces. However, the potash tailings pile at Bischofferode is not yet covered and suitable to apply this technology. On the tailings piles that are currently being covered, however, it would be possible to use an additional sealing layer to those areas that are still open. Apart from that, the only remaining task is to implement the proposals for the dump protection forest fully and to carry out the collection of the dump water at the foot of the dump more consistently than before. 4.2 Comparison of Different Pile Forests The modeling does show differences between the tree species. However, they are smaller than we had expected, which we attribute to the parametrization of the vegetation. While

The Impact of Afforestation on Seepage Water Formation

497

our database allowed to implement species-dependent leaf and stem area indices, there is a general lack of knowledge about e.g. the plant physiological properties of the considered tree species. This concerns, in particular, the distribution of fine roots in the soil and the shoot-root ratio. Due to the project framework, we could not validate the model results. This should be possible in future work. In the present evaluation, we have limited ourselves to the period until 2020. However, the two dry years 2018 and 2019 (“summers of the century” with over 30% precipitation deficit as compared to typical climate) show impressively how much the climate conditions are already changing. The different scenarios, especially with a stronger anthropogenic radiative forcing, let us expect further drastic changes for the two locations under consideration. This is another reason why it is imperative to establish mixed stands that will continue to ensure forest functions in the event of the loss of a species due to drought stress or consequential damage such as calamities. In general, we conclude that the high evapotranspiration of forests is a useful instrument to reduce the environmentally hazardous leachate formation of mining legacies. Based on our results, we recommend a functional permanent post-mining forest with siteadapted resilient tree species, especially on sites with high seepage formation combined with the leaching of environmentally hazardous substances. Acknowledgments. We would like to thank the LMBV for commissioning us to investigate the optimization of the greening of potash tailings piles. We would like to thank LfLUG for the second task, which we were also able to carry out within the framework of a project. Thanks to the LE-B for providing data for the location of the Nochten dump. We thank our partners for the excellent cooperation and discussion.

References 1. Amoozegar, A.: Compact constant head permeameter: a convenient device for measuring hydraulic conductivity. In: Advances in Measurement of Soil Physical Properties: Bringing Theory into Practice. Soil Science Society of America Special Publication, vol. 30 (1992) 2. Dunger, V.: Dokumentation des Modells BOWAHALD zur Simulation des Wasserhaushaltes von wasserungesättigten Deponien/Halden und deren Sicherungssysteme. Technical report, TU Bergakademie Freiberg, pp. 1–186 (2002) 3. Federer, C.A.: BROOK90: a simulation model for evaporation, soil water and streamflow (2002). http://www.ecoshift.net/brook/brook90.htm 4. Hammel, K., Kennel, M.: Charakterisierung und Analyse der Wasserverfügbarkeit und des Wasserhaushaltes von Waldstandorten mit dem Simulationsmodell Brook90, vol. 185. Forstliche Forschungsberichte, München (2001) 5. Hildmann, C., Rösel, L., Zimmermann, B., Knoche, D., Hartung, W.-D., Benthaus, F.C.: Reduction of seepage outflow from potash tailings piles by improvement of greening: results of a hydrological simulation. In: Drebenstedt, C., Paul, M. (eds.) Mining Meets Water, Leipzig, pp. 772–779 (2016) 6. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria (2017). https://www.R-project.org/ 7. Schmidt-Walter, P., Trotsiuk, V., Meusburger, K., Zacios, M., Meesenburg, H.: Advancing simulations of water fluxes, soil moisture and drought stress by using the LWF-Brook90 hydrological model in R. Agric. Forest Meteorol. (2020). https://doi.org/10.1016/j.agrformet.2020. 108023

A Conceptual Digital Framework for Near Real-Time Monitoring and Management of Mine Tailing Storage Facilities Iqra Atif1(B) , Hamid Ashraf2 , Frederick Thomas Cawood1 , and Muhammad Ahsan Mahboob3 1 Wits Mining Institute (WMI), University of the Witwatersrand, Johannesburg, South Africa

[email protected], [email protected] 2 School of Advanced Geomechanical Engineering (SAGE), National University of Sciences

and Technology (NUST), Risalpur Campus, Risalpur, Pakistan 3 Sibanye-Stillwater Digital Mining Laboratory (DigiMine), Wits Mining Institute (WMI),

University of the Witwatersrand, Johannesburg, South Africa

Abstract. The demand for mineral products is doubling every 30 to 40 years. With a growing world population, mineral product supply must keep up with the demand because of industrialization, urbanization, and internationalization. With an increase in mining activities and falling mineral grades, more waste production is inevitable. This issue poses tremendous challenges to the world and its people, and efficient storage and management of waste material associated with mineral tailings are considered as part of the global re-imagination of mining waste projects. It is common practice to store the mineral waste in impounding structures called ‘tailing dams’ to manage toxic waste concentrations and acid mine drainage. During the last few decades, the frequency of tailings dam disasters has increased, giving rise to global concern about safety, sustainability, management, and monitoring of Tailing Storage Facilities (TSFs). World-leading sustainable development organizations like ICMM, UNEP, and ICOLD have developed guidelines and policies for efficient management of TSFs. However, with the advancement of new technologies and packaging these within a system or integrated approach, there is a need to design a concise and explicit framework for the efficient monitoring of TSFs as part of a risk management strategy. This research is aimed at formulating a conceptual framework for near real-time monitoring using advanced geospatial technologies, remote sensing, on-site instrumentation, and their integration with the GIS-based system. Such a smart GIS system will assist stakeholders, mine operators and engineers to remotely sense and control the potential failure of TSF, to understand its causes and estimate the consequences of failure. Keywords: Tailing storage facilities · Digital mine · GIS · Near-real-time monitoring · Conceptual framework · Fourth industrial revolution (4IR) · Mining 4.0 · Tailings review

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 498–530, 2021. https://doi.org/10.1007/978-3-030-60839-2_27

A Conceptual Digital Framework

499

1 Introduction The world’s growing population needs minerals, and without mineral products, it cannot survive – let alone sustain itself. The demand for minerals is doubling every 30 to 40 years [1]. With a growing world population, mineral product supply in the 21st century will have to keep up with the increasing demand for an industrialized, urbanized, and internationalization context. With rising volumes and depths of mining, combined with falling mineral grades, more waste production is inevitable with current mining methods. More tons at lower mining grades mean more waste tailings that need to be stored in a Tailings Storage Facility (TSF) [2]. Good design, monitoring, and management of waste material is part of the global re-imagination of mining waste project to ensure the safety and way of life of people living close to these TSFs. It is common practice to contain toxic waste in impounding structures - in mining called tailing dams to avoid the spreading of acid mine drainage contaminating water and land. The risks associated with TSF failure are significant, and impacts range from minor to catastrophic, i.e., loss of the facility, life, and the natural environment. International organizations like International Council on Mining and Metals (ICMM), United Nations Environment Program (UNEP), International Commission on Large Dams (ICOLD) have developed TSFs guidelines and policies assisting with risk and controls for the management of TSFs. Governments have intruded stricter rules and regulations, while mining companies have incorporated it into their risk management strategies. Advancements in digital technologies in an integrated system make near real-time monitoring and risk management a definite possibility. The expected benefits of such a system are apparent. They include a better understanding of the problem because of more (and new) data that becomes available, on-going risk assessment and control, and better productivity made possible by preventative tools and compliant operations. Besides, with central monitoring because of technology, scarce skills can be concentrated at a group level rather than preceding ‘thin’ skills across all operations. This research is part of the Sibanye-Stillwater Digital Mining Laboratory (DigiMine) research agenda and was commissioned by the company as part of its risk management strategy. The organization of the rest of this paper is as follows. Section 2 evaluates typical literature on existing leading practices and regulatory requirements on TSFs. Section 3 lists the considerations and technologies available for TSFs design. In contrast, Sect. 4 proposes a conceptual framework for near real-time monitoring and management of tailings storage facilities for intelligent decision-making. The paper ends with a conclusion and recommendation on how to proceed with the implementation of the system.

2 Background and Leading Practices on the Management of TSFs The conceptual architecture of the study is illustrated in Fig. 1. The objective is to monitor as part of an overall risk management structure. After considering the general regulatory framework and guidelines on TSF management, the leading practices and technologies will become part of the Conceptual Framework (Sect. 4) for monitoring on a GIS platform in near real-time.

500

I. Atif et al.

Fig. 1. Conceptual architecture of tailing storage facilities study

2.1 Background The 21st century holds both challenges and opportunities for mineral-based economies due to massive industrialization, fast urbanization, and the ever-increasing international demand for minerals. With increasing mineral extraction activities across the globe, efficient management of tailings is becoming a global challenge to ensure a safe environment and a sustainable economy based on mining. Over the last two decades, there has been a substantial rise in tailing dam failures giving rise to concern over the monitoring and management of these TSFs. World-leading sustainable development organizations like ICMM, UNEP, and ICOLD have developed TSFs guidelines and policies for their efficient management. However, with the advancement of new technologies by incorporating intelligence-based integrated approaches, there is a need to design an explicit framework for the efficient management of TSFs. Since 1915, the failure of TSFs has emerged as a major public concern and a source of liability risk for the mining industry. TSF instability is not new. A well-known case study is the 1915 TSF collapse at Agu Dulce in Chile [3] where intense rainfall resulted in the failure of the dam and released 180,000 m3 of copper tailings into the immediate environment. By 2020, over 292 tailings dam failures had been reported. These incidents resulted in at least 2100 fatalities, significant financial loss, and severe damage to the environment and global economy (estimated average financial loss of $500 million) [4], as shown in Fig. 2.

A Conceptual Digital Framework

501

Fig. 2. Tailings dams’ failures record as per decades

Hui, Charlebois and Sun [5] studied 106 cases of tailing dam failures concerning their major causes from 1915 to 2016. Table 1 contains an analysis of the major tailings dam failures in the last decade. Unlike other parts of the mining value chain that show downward trends, the number of fatalities associated with TSFs have increased over the last five decades (Fig. 3). Table 1. Major tailings dam failure after 2012. Sr. No

Mine

Location/Country

Year

Damage

1

Padcal Mine

Itogon, Benquet province, Philippines

2012

20.6 Mton released after a strong rainstorm

2

Obed Mountain Coal Mine

North East of Hinton, AB, Canada

2013

0.67 Mm3 of coal tailing water released and 90.000 tons sediments

3

Mount Polley

Canada

2014

24.3 Mm3 of tailings and contacted water released

4

Hpakant

Kachin State, Myanmar

2015

At least 113 people dead, no further information

5

Fundao

Brazil

2015

45 Mm3 . Nineteen fatalities and extensive contamination.

6

New Wales Plant

Mulberry, Polk County, FL, USA

2016

0.84 Mm3 of waste fluid contacted an aquifer source of drinking water (continued)

502

I. Atif et al. Table 1. (continued)

Sr. No

Mine

Location/Country

Year

Damage

7

Kokoya Gold Mine

Bong County Liberia

2017

Gold tailing with cyanide and arsenic content released to the river – the source of drinking water

8

Mishor Rotem

Israel

2017

0.10 Mm3 of tailings and contacted water released

9

Ronglvsham Mine

Hubei province, China

2017

0.20 Mm3 of tailings released

10

Cadia

New South Wales, Australia

2018

Upper basin containment dam collapsed over the lower basin, minor consequences

11

Brumadinho

Brazil

2019

12 Mm3 total tailings were released; 197 people dead and 111 missings. Extensive environmental damage

Fig. 3. Number of fatalities over time.

A standard definition for a TSF is an earth-filled embankment dam used to store byproducts of mining operations after separating the ore from the waste. It can be in the form of liquid, solid, or the form of a slurry of fine particles. These are usually highly toxic and potentially radioactive, which can be catastrophic, in case of contact with the

A Conceptual Digital Framework

503

environment [6]. Table 2 highlights the purpose of tailing dams and its deviation from conventional earth-filled dams. Table 2. Difference between tailings and water dams. Sr. No

Characteristic

Water Dams

Tailings Dams

1

Design Objectives

It contains and controls water for any purpose like irrigation and/or power

It holds wastes from mineral processing (tailings). It is primarily built to contain slurry waste material from the processing of ore and usually is damaging to the environment

2

Engineering Design Criteria

Stability (both static and seismic). Hydrology is based on downstream risk

Stability (both static and seismic). Hydrology is based on the downstream flood, process needs, and environmental risks

3

Environmental Criteria

Limited long-term environmental impacts, apart from short-term flooding

Environmental impact can be severe - so minimize risk to an acceptable level in all phases (construction, operation, and closure)

4

Construction

It is usually constructed in a single phase

It is constructed in a phased (or layered) approach over the life of the mining operation to suit storage requirements

5

Operation

Based on the operational level of the reservoir

Continued design changes as per construction and storage to meet waste disposal, environmental and flood control requirements

6

Closure

No specific closure requirements, except the life of reservoir and monitoring (inspection) requirements

The closure is critical as mining and processing cease. The closure is designed for a longer sustainable duration to cater to environmental and social issues

2.2 Tailings Dam Types The classification of TSFs is based on the methods of construction, i.e., upstream, downstream, and centerline [7] (Fig. 4). The centerline (or layered) approach is the most common structure for TSFs at South African gold and platinum mines.

504

I. Atif et al.

Fig. 4. Different types of tailings dam construction methods.

There are different stages in the life cycle of a tailing dam. It is essential that monitoring is done for all stages of the TSF life cycle. Figure 5 explains the extent of monitoring required during different stages over time. It is necessary to manage tailings within a carefully designed framework of safety during all the phases over the life of a tailing dam. These phases could simply be divided into three distinct stages, namely the commissioning-, operation- and decommissioning stage.

Fig. 5. Different stages in the life of a tailing dam

2.3 TSF Leading Practices To benchmark and analysis of best practices, countries, and regions like the United States, Canada, the European Union, Australia, and South Africa were analyzed. It is crucial

A Conceptual Digital Framework

505

to have a proper engineering design for the construction of tailings dams, which should comply with both local legal requirements and international best practices. Broadly, the tailings dam safety recommendations from technical organizations such as ICMM, ICOLD, and Environmental Social and Governance (ESG) can be summarized as follows [8]: • Proper construction design and operating procedures according to the type of the TSFs. • An all-inclusive, robust and stand-alone online monitoring system. • Planning and design for closure, with post-closure supervision ensuring no “walkaway” from responsibility • An adequate monitoring organizational structure, with clearly defined of roles and responsibilities, and • Procedure for risk management enabling near real-time information feeding decisionmaking processes and protocols. Until recently, the monitoring governance structure of most mining companies was not adequate to ensure the stability of TSFs. The ICMM, CAD, ICOLD, amongst others, have developed guidelines and recommendations to improve governance for tailings management systems. Figure 6 summarizes the basic requirements for such governance systems [8].

Fig. 6. Basic requirements for adequate governance and system for tailings dam monitoring, Source: Abbott, Eldridge, Wates and Marais [8]

2.3.1 Visual Inspections Visual inspections remain an essential aspect of monitoring, but it is not always necessary for ‘boots-on-the-ground.’ Depending on the urgency, suitable remote sensing technologies include drones and satellite images. These inspections could be on an adhoc, continual, or periodic basis, to individually check for early indicators of potential

506

I. Atif et al.

failure modes in the tailings dams [9]. The person carrying out inspection must be trained in protocols and must be able to relate any observed changes to site-specific identified failure modes. The inspecting manager must carry out regular analysis for potential dam failure scenarios for the different loading capacities. Mine operators usually commission external experts for such advice when there is limited internal capacity [9]. The situation is not ideal, because the person carrying out visual inspections is often not dedicated to this task and must schedule it as one of many other duties he or she is responsible for. 2.3.2 Seepage Monitoring Increased pore pressure results in seepage, thereby increasing the risk of failure. To decrease the risk of failure due to seepage, a proper drainage system is essential. Seepage points are monitored on an individual basis, as these provide a good indicator of common failure causes such as internal erosion [10]. Discharges are recorded with weirs, but auto or semi-automatic systems are also available for real-time recording of discharge. This helps for improved quality and frequency of data, which allows better interpretation and behavior in case of abnormalities. The flow of dangerous contaminants entering the water bodies downstream of the dam can also be recorded using the same system. Geophysical surveys are used for selected tailings dams to detect resistivity variations effectively. Measurements can be repeated with a constant time interval to determine changes of possible leakage zones, depending on the electrical properties of the filling materials [11]. 2.3.3 Phreatic Surface Monitoring A low phreatic surface level is important for the stability of tailing dams, and the open standpipe piezometers are the most commonly used instrument for this purpose. The readings form standpipe piezometers can be automated for real-time work. However, their installation in tailing dams is not as convenient as is the case of water storage dams. TSFs are raised continuously, causing the piezometers installed initially at the bottom to be higher than the actual foundation of the facility. Hence, it is important to check their elevation level constantly at the bottom end to ensure that the screening zone is still functioning [12]. The section of the inner diameter of standpipe piezometers also needs to be carefully positioned as it varies with the location of the instrument. The standpipe piezometers with a smaller diameter and less porous surface are usually used for more delicate materials. Closed piezometers are also sometimes used, but they require more maintenance than standpipe piezometers. Automation is achieved with the use of special pressure transducers installed inside the piezometric tubes, which increases data quality and frequency. 2.3.4 Pore Pressure Monitoring Pore pressure is another critical factor to be monitored, particularly in the case of a tailing dam with poor drainage. The standpipe piezometers can also be used to measure pore pressure by reading the water level with capped tubes when the pressures are too high in the inner part of the tailing dams. In this case, different types of piezometers,

A Conceptual Digital Framework

507

such as electrical resistance, vibrating wire, pneumatic, and hydraulics, are used to read pressures remotely for monitoring [13]. However, in the case of upstream tailings dams, the use of piezometers to monitor pore pressure can be ineffective because it fails to predict static liquefaction. Piezometers require extensive maintenance and can become problematic, and in that case, vibrating wires are suitable replacements to piezometers. Vibrating wires are more robust and do not get affected by electrical disturbances like electrical resistances. For the measurement of dynamic pore pressure (critical in case of seismic events) vibrating wire, piezometers are widely used due to the quick response; they can offer. Table 3 presents a summary of the most commonly used monitoring methods and technologies for pore pressure monitoring. Table 3. Monitoring instrumentation for pore pressures or moisture changes. Sr. No Device and Method Parameters

Application

Experience

1

Electric Pore pressure and piezometers with temperature telemetry to process plant or phone

Monitor pore pressure changes due to loading and changes in hydrogeological conditions

Standard practice at many mines. Strings at multiple depths are preferred

2

TDR, Neutron Probes

Saturations levels and temperature

3

Self-Potential

The passive electrical method, which is sensitive to the flow of seepage water

Electrodes are placed on the dam surface for both investigation and monitoring

Research and long-term field measurements have been performed mainly in the US, Canada, France, and Sweden

4

Distributed Fiber Optic sensing

Temperature and strain are measured in optical fibers using laser light

Cables are installed in new or old dams for seepage evaluation using temperature and strain analyses to assess movements

Basic research since 1996 in Germany and Sweden. Further research, especially in France, Austria, the Netherlands, UK, and the US. Challenges are calibrating measurements to site conditions

508

I. Atif et al.

2.3.5 Seismicity and Dynamic Pore Pressure In active seismic regions, large tailing dams can be monitored with “strong motion accelerographs” to register earthquakes above a given threshold. Usually, two to three accelerographs are used to control a particular site in case of any seismic event. They provide the data needed for calibrating seismic models and calculating the earthquake response of the dam as construction advances. Liquefaction may occur when dynamic pore pressure is very high, and the effective stresses are reduced to zero. To monitor dynamic pore pressure, vibrating wire piezometers, which have a rapid response, can be used. 2.3.6 Deformation Monitoring There are two types of movements in the tailing dams, vertical and horizontal movements, namely. • Vertical Movements. Tailings dams do not have permanent surfaces, and there are internal vertical movements due to compaction from the weight of the material and consolidation [14]. Sudden vertical movements are usually an indicator of a failure mode, including internal erosion. There are also some external movements, which are mostly measured using Global Navigation Satellite System (GNSS) and collimating line, small angle, geometric leveling and precise trigonometric leveling methods, etc. It is a survey network consisting of a control network of geodetic datum outside the premises of the dam and then establishing a separate network of points typically located along the crest of the dam and on its berms. Internal vertical movements can also be measured (in terms of differential settlements within the dam) with electromagnetic probes deployed within standpipes equipped with stainless steel plates or rings. • Horizontal movements. These are more conspicuous and can be detected easily with the help of visual inspections and by using GNSS at the same geodetic points that are used for surveying the vertical movements. Horizontal movements are generally caused by settlement of the different sections of the dam [14]. The inclinometers and piezometers are the primary instruments used to measure the horizontal movements, particularly for interlayers movement monitoring. Besides, the geodetic techniques, Terrestrial Laser Scanning (TLS), hydrographic survey data, and photogrammetry are also an effective way to detect changes in the horizontal direction. Satellite images, drones, and aerial photography can monitor vertical and horizontal displacements. It is becoming useful tools to monitor the downstream slope (or sides) of the dam. Proper ground control with sufficient checkpoints must be surveyed first to benchmark the data obtained from drones or satellite maps. The precision obtained from the satellite information provides for sub-cm accuracy (depending on the good control points, weather conditions, and level of vertical movements) when permanent Scattered Synthetic Aperture Radar (SAR) and Interferometry (PSInSAR) technologies are used with highly reflective targets.

A Conceptual Digital Framework

509

2.4 Regulatory Requirements on TSFs Mining related legal and regulatory regimes have become stricter over time. Past environmental regulations were not specific enough to stop catastrophic TSF disasters from happening. Recently, the world environmental bodies and other organizations from developed countries introduced more stringent environmental legislation to ensure sustainable mining practices. The prime objective of this study is to formulate a robust and scientific monitoring system for the management of TSFs. There is a need to design a system based on leading practices and compliant with regulatory requirements. For this purpose, the Laws, guidelines, and best available techniques from the United States, Canada, Australia, the European Union, and South Africa are consulted. 2.4.1 International Frameworks The environmental issues related to the mining industry primarily related to the waste it produces during the processing of ore are considered for this study. Most of these countries are from the (western) developed world, and the main reasons for this are: • Enabling Institutional Framework. These countries have to allow institutional frameworks and enforce specific legislative requirements. There are also regulatory approvals that require monitoring, which is implemented by the regulators as part of the overall regime; and • Corporate Risk Mitigation. There is risk associated with not having effective plans in place, and large corporations know that it can affect their reputation and, ultimately, license to practice. Table 4 contains different organizational initiatives internationally to ensure the safety of the tailing dams. Table 4. Global organizations and their respective guidelines/initiatives Sr. No

Organization

Document

Standard

Phases

1

Global Minerals Professional Alliance (GMPA)

Global Action on Tailings

Initiative

All phases

2

GRID-Arendal

Global Tailings Dam Portal Project

Initiative

All phases

3

International Commission on Large Dams (ICOLD)

Sustainable design and post-closure performance of tailings dams – bulletin 153

Guidance

Planning/design/concept/closure

(continued)

510

I. Atif et al. Table 4. (continued)

Sr. No

Organization

Document

Standard

Phases

4

International Commission on Large Dams (ICOLD)

Tailings Dam Design Technology Update

Guidance

Planning/design/concept

5

International Commission on Large Dams (ICOLD)

Recommendations for operation, maintenance, and rehabilitation – bulletin 168

Guidance

Operational/Maintenance

6

International Commission on Large Dams (ICOLD)

Dam safety Guidance management: operational phase of the dam life cycle – bulletin 154

Operational/Maintenance

7

International Council on Mining and Metals (ICMM)

Global Tailings Review

Initiative

All Phases

8

International Council on Mining and Metals (ICMM)

ICMM 2016 Initiative Tailings – Define an appropriate tailings storage facility governance framework

All Phases

9

International Organization for Standardization (ISO)

Mine Closure and Reclamation Terminology Mine Closure and Reclamation Management Planning

Standard

Closure/Decommissioning

10

International Organization for Standardization (ISO)

Environmental Management Systems – ISO 14001:2015

Standard

All Phases

11

UN Environment Mine Tailings & GRID Storage: Safety Is – Arendal No Accident

Guidance

All Phases

(continued)

A Conceptual Digital Framework

511

Table 4. (continued) Sr. No

Organization

Document

Standard

Phases

12

UNEP (United Nations Environment Programme (UNEP) & International Council on Metals & the Environment (ICME)

The Cyanide Code

Initiative

All Phases

Regulations for tailings dam monitoring for the United States, Canada, Australia, EU, and South Africa are presented in Table 5. Table 5. Guidance/standards/initiatives for management of TSFs Sr. No

Organization

Document

Standard

Phases

US 1

United States Risk Management Society on Dams for Dam (USSD) Construction

Guidance

Planning/Design/Concept/ Pre-Construction/Closure/Decommissioning

2

Alaska Department of Natural Resources (ADNR)

Alaska Dam Safety Program

Guidance

All phases

3

Federal Emergency Management Administration (FEMA)

Federal Guidelines Guidance for Dam Safety

Planning/Design/Concept/ Pre-Construction/Operation/Maintenance/

4

U.S. Environmental Protection Agency

Design and evaluation of Tailings dams

All phases

5

US Environmental Protection Agency

The National Environmental Policy Act of 1969

Guidance

(continued)

512

I. Atif et al. Table 5. (continued)

Sr. No

Organization

Document

6

US Environmental Protection Agency

The Clean Air Act, 42

7

US Environmental Protection Agency

the Clean Water Act, 33

Standard

Phases

Guidance

Operational

Canada 1

Canadian Dam Association

Dam Safety Guidelines

2

The Mining Association of Canada (MAC)

The Tailings Guide Guidance – A Guide to the Management of Tailings Facilities

All phases

3

The Mining Association of Canada (MAC)

OMS Guide – Developing and Operation, Maintenance, and Surveillance Manual for Tailings and Water Management Facilities

Guidance

All phases

4

The Mining Association of Canada (MAC)

TSM – Towards Sustainable Mining

Initiative

All phases

Australia 1

Australian National Committee on Large Dams (ANCOLD)

Guidelines on Tailings Dams

Guidance

All phases

2

Australian Government, Department of Industry, Tourism, and Resources

Leading Practice Sustainable Development Program for the Mining Industry – Tailings Management

Initiative

All phases

3

Government of Western Australia, Department of Mines, Industry Regulation and Safety

Tailings storage facilities in Western Australia – Code of Practice

Guidance

All phases

(continued)

A Conceptual Digital Framework

513

Table 5. (continued) Sr. No

Organization

Document

Standard

Phases

4

Queensland Government

Dam Safety Management

Guidance

All phases

5

Queensland Government & METS Ignited

METS Clustering

Initiative

All phases

1

United Nation Economic Commission for Europe (UNECE)

Safety guidelines and good practices for tailings management facilities

Guidance

Pre-Construction/Operation/Maintenance/ Closure/Decommissioning

2

EIT Raw Materials

Stings Initiative – Supervision of Tailings by an Integrated Novel Approach to combine Ground-based and Spaceborne Sensor data

All phases

3

European Commission

Best Available Techniques (BAT) Reference Document for the Management of Waste from Extractive Industries

Guidance

All phases

4

Environmental Protection Agency

European Waste Catalogue and Hazardous Waste List

Guidance

EU

South Africa 1

Department of Environmental Affairs

The National Environmental Management Act No. 107 of 1998 (NEMA)

Act

All phases except closure & reclamations

2

Department of Mineral & Energy

Mineral & Petroleum Resources Development Act, 28 of 2002

Act

All phases

3

Department of Mineral & Energy

Mine Health & Safety Act, 29 of 1996

Act

All phases except closure & reclamations

(continued)

514

I. Atif et al. Table 5. (continued)

Sr. No

Organization

Document

Standard

Phases

Ministry of Environmental Affairs

National Environmental Management: Waste Act No. 59 of 2008

Act

All phases except closure & reclamations

Department of Mineral & Energy

Radioactive waste Guidance management policy and strategy for the Republic of South Africa 2005

All phases except closure & reclamations

4

Department of Health

National Health Act, 61 of 2003

Act

–

5

Department of Health

National Water Quality Act, 36 of 1998

Act

6

Ministry of Environmental Affairs

NEMA, Air Quality Act, 39 of 2004

Act

7

Ministry of Environmental Affairs

NEMA, Waste Act, 59 of 2008

Act

8

Department of Environmental Affairs

Mining Residue Regulations (MRR) in 2015

Guidance

In Canada, mining is being regulated at the provincial level, and different provinces have their own set of rules and regulations for the regulation of tailings dams. Table 6 presents rules and regulations being followed in all the regions of Canada. Table 6. Province/territory wise mechanism of regulation of tailings dams in Canada. Sr. No

Province/Territory

Ministry/agency

Legislation

Regulation

1

British Columbia

Ministry of Energy Mines Act & Mines

Health, Safety & Reclamation Code

2

Alberta

Alberta Energy Regulator

Water Act

Water Ministerial Regulation

3

Saskatchewan

Environment

Assessment Act

No

4

Manitoba

Sustainable Development

The Water Rights Act

No (continued)

A Conceptual Digital Framework

515

Table 6. (continued) Sr. No

Province/Territory

Ministry/agency

Legislation

Regulation

5

Ontario

Ministry of Northern Development and Mines

Mining Act

No

6

Quebec

Natural Resources

Mining Act

No

7

New Brunswick

Environment & Clean Water Act Local Government

No

8

Nova Scotia

Environment

Environment Act-Regulation

No

9

Newfoundland & Labrador

Department of Municipal Affairs and Environment

Water Resources Act

No

10

Prince Edward Island

Environment, Energy & Forestry

No

No

11

Yukon Territory

Water Board, Water Resources Section

Water Act

No

12

Northwest Territories

Mackenzie Valley Land and Water Board

Mackenzie Valley Waters Regulation Resource Management Act Waters Act - Justice

13

Nunavut

Nunavut Water Board

Nunavut Waters & Surface Rights Tribunal Act

No

14

Canadian Federal Government

Canadian Nuclear Safety Commission

Nuclear Safety and Control Act

No

3 Monitoring Technologies and Considerations for TSFs Design 3.1 Suitable Monitoring Technologies for TSFs Design 3.1.1 Ground Penetrating Radar Ground Penetrating Radar is a geophysical method that determines the sub-surface conditions using radar pulses. It can monitor and detect movements within a surface, such as an embankment of a tailings dams in real-time. Also, it can trigger an alarm. However, the technology is most suited for solid concrete structures and is thus not ideal for dynamic structures such as TSFs that move over time and potentially pose new risks.

516

I. Atif et al.

3.1.2 Terrestrial LiDAR LiDAR is a survey instrument that takes an accurate ground surface survey by illuminating the target area by laser light and calculating the distance it covers to reach back. LiDAR can measure the minutest movements in the dam wall. This is done by generating a topographical survey point cloud model and comparing it with a previous model. The method can be automated by allowing the LiDAR to continuously scan the area and detect any anomaly during the scanning process. The use of LiDAR is new to the mining industry, and on occasion, has been used for tailings dam monitoring [9]. Another reason for LiDAR not being popular is its initial price tag, which is substantial for the purchase of the hardware and software. LiDAR is especially useful for creating 3D models of the area by using the laser light and generating point data. However, due to the capacity of the system to detect minimal changes in the object, it may not be that effective in a downstream tailing dams’ facility, where natural expansion can falsely trigger an alarm and thus causing over-reaction. This problem can be overcome by using an external-expert system for monitoring purposes, and real alarms can be verified with that system as well. Nevertheless, the use of LiDAR could be very beneficial, but the exact effects will be determined by the site-specific requirements of the tailing dams. It is also pertinent to note that LiDAR would be useful for detecting external movements and not internal ones - as the signals cannot penetrate surfaces. Figures 7 and 8 show a typical use of LiDAR technology.

Fig. 7. Some cases of the use of LiDAR technology, Source Minear [15]

3.1.3 Seismic Sensors Tremors and earthquakes can affect the stability of the dam structures. It is, therefore, crucial to record and monitor seismic data. Seismic sensors are specially designed to detect movement caused by earthquakes, so precautionary measures must be taken for the safety of the structure [17]. Seismic sensors also detect movements induced by blasting and heavy plant movements operations. For example, accelerometers and magnetometers can detect minute movements in the dam walls, which can be modeled to set up the triggering of alarms. However, care must be observed in the use of seismic sensors around active mines as they can sometimes produce false alarms due to blasting operations. These sensors

A Conceptual Digital Framework

517

Fig. 8. Laser intensity 3D point cloud showing the three scan positions, retro-reflector targets, and vectors between scan positions and surface marks, (Source: Berberan, Ferreira, Portela, Oliveira, Oliveira and Baptista [16])

could very easily be modeled and coupled with other sensors to provide real-time data for the monitoring of the tailing dams. 3.1.4 InSAR Interferometric Synthetic Aperture Radar, abbreviated InSAR is a technique in which radar waves are used to detect changes in the topography of surfaces. The equipment is carried by a satellite and works on the principle of change in time of returning light from the source [18]. InSAR is also capable of producing high-resolution digital terrain models up to sub-centimeter accuracy. The potential precision of InSAR highly depends on the efficiency of the satellite orbit, atmospheric and weather conditions, and the SAR wavelength. InSAR can easily be used to monitor the deflections in tailings dams’ surface [19, 20]. The issue with InSAR is that the data is produced after a fixed duration of 8– 10 days, and only surface information could be obtained without any knowledge of the underlying sub-surface causal factors responsible for the movements. InSAR is best used with geodetic control points in the area and when coupled with real-time geotechnical sensors reporting the sub-surface movements. 3.1.5 Total Station Automation The mining industry, especially for the monitoring of surface mine slopes, is familiar with total station automation surveys. The principle is to automatically (and accurately) record the changes in the surface of the wall or tailings dams by comparing the already

518

I. Atif et al.

set out survey points with earlier ones (the user sets recording intervals). The total station survey monitoring has similar results compared to LiDAR or InSAR, in that it allows the user to see movement. However, these methods offer little in the way of insight into its causes, and geotechnical interpretation is required before an alarm could be triggered. 3.1.6 Real-Time Monitoring Systems The overall real-time tailings dams monitoring concept involves combining different sets of instrumentations like piezometers, weather stations, ultra-sonic height sensors, seismic sensors, and phreatic board sensors installed at the same time and recording data collectively. This will enable the monitoring managers to audit the data in real-time and make informed decisions according to the set alarm limits. The data collection intervals can be set according to the specific site requirements. This set of arrangements gives a holistic view of a particular situation. If connected to a robust real-time communication system, it provides flexibility for the system to be centralized–and used from anywhere in the world [21]. Each mine has different conditions like different topographical, geological, and mineralogical properties, which necessitates a different approach for various sites. There is also a difference in the level of expertise of the people involved in the monitoring system, which results in each tailings product being unique. Although most mining operators have installed some type of monitoring system to monitor tailings dams, there is no foolproof system to date. A common reason for regular failure is that dams are not operated according to their design criteria, and if monitoring is not done in a holistic manner involving all necessary ingredients of a real-time monitoring system, deformation and failure can happen. Hence, it is vital to design a monitoring system that is based on site-specific research and incorporating leading practices for effective risk management at TSFs. 3.1.7 Mine IoT Based Systems The internet of things (IoT) is an interrelated system allowing computing devices, digital machines, sensors, and humans to interact with each other without depending upon one another and to be able to transfer and process big data. The IoT is an example of a comprehensive independent monitoring system that can help people to make informed decisions. Such a (Mine) IoT based system, when installed at a tailings dams, can intelligently sense deformation, seepage, water level, rainfall, and other critical safety parameters of tailings dam operation. One such a Mine IoT based system was designed by Wang, Yang, and He [22] and implemented for the monitoring of tailing dams in Dexing Copper mine in China. 3.2 Considerations for Tailings Storage Facilities Design 3.2.1 Planning TSF management and monitoring start with the considerations for locating the facility. Table 7 lists the typical considerations during this phase. Individual siting would depend upon the site-specific conditions of the area of the mining operation.

A Conceptual Digital Framework

519

Table 7. Potential impact areas and the issues to consider when planning a new tailings management facility Sr. No 1

Impact Area

Property

Planning Parameter

Land

Topography

The topography of the area

2

Potential impact area

3

Land use

Existing and future possible land use

4

Sites of existing waste rock piles

5

Deposition plane and volume/capacity

6

Geotechnical

Condition of basin and dam foundations

7

Vegetation

Baseline conditions including flora and fauna

8

Potential impact no vegetation, wildlife, aquatic life

9 10

Environment

Reclamation

Mine closure considerations

Water Management

Water Management Scheme

11

Potential of surface and groundwater for contamination

12

Local & regional surface & groundwater flow

13

Complete Water balance with all inputs & outputs

14

The potential impact on downstream communities

15

Atmosphere

16

Climate conditions, including precipitation and lowest and highest annual temperatures Potential airborne releases

The planning for the TSFs must start at the initial stage. Leaving it to the operational phase could lead to severe environmental and financial impacts as the cost of implementing tailings management practices during operation will be significantly higher for problematic locations [23]. TSFs must be regularly inspected to ensure its stability and a regular monitoring mechanism must be in place. Monitoring practices must be documented and conducted according to plan illustrated in Fig. 9. 3.2.2 Legal Framework for TSFs The regulatory requirements for TSFs management usually adhere to the following principles: • The regulations concerning tailings disposal and subsequent management rest with the respective government department or environmental protection agency in the region where the mine exists; and

520

I. Atif et al.

Fig. 9. The plan and method for monitoring

• The regulations for pollution control and TSF water discharge rest with the environmental protection agency. The main focus of all the regulations concerning tailings should be to ensure that tailings are managed, and TSFs are safe. The main aims are first, the stability of tailings dams and second, to contain pollution. This requires continuous consideration of closure design, construction, and aftercare throughout the TSF life cycle. The regulatory body must ensure that all TSFs designs demonstrate that sustainable outcomes are achieved during operations and after the closure of mine by applying leading practices, which include: • Assessment of the risks associated with tailings storage at a particular site; • Reporting of all TSFs, particularly those involving dewatering and/or have no further capacity to hold a surplus amount of water in it; and • Demonstration of the ability that the tailings storage method will manage all risks within acceptable limits [24].

A Conceptual Digital Framework

521

In some cases, the TSF design, construction, and on-going management are regulated by specific legislation issuing specific guidelines for the management of TSFs. For example, in New South Wales, Australia, the Dams Safety Committee monitors the tailings containment under the Dams Safety Act 1978. In Western Australia, the Department of Mines and Petroleum has developed a code of practice (DMP 2013) for TSFs with the following essential factors: • • • •

Fulfilling the requirements approved under mining-related legislation; Demonstration of the fact that TSF is safe, stable and non-polluting; Ensuring the involvement of competent consultants; and Meeting the occupational health and safety requirements.

3.2.3 Best Practices Framework Table 8 presents a framework for a sustainable tailings dam monitoring system after considering the issues in this section. Table 8. Potential impact areas and the issues to consider when planning a new tailings management facility Sr. No

Characteristic

Strategy

1

Design, operation, and maintenance

According to specifications of ICOLD and ANCOLD, or other internationally recognized standards An appropriate independent review should be undertaken at design and construction stage On-going monitoring of both the physical structure and water quality during operation and decommissioning

2

Seismic stability

When located in areas of high-risk seismic zones, a check on the maximum design earthquake assumptions shall be checked Structure’s stability during seismic events must also be incorporated in the design.

3

Geotechnical stability

The specific risks/hazards associated with geotechnical stability must be taken into accoun

4

Mitigation

Environmental considerations should also include emergency preparedness and esponse planning and containment/mitigation measures in case of catastrophic release of tailings or supernatant waters (continued)

522

I. Atif et al. Table 8. (continued)

Sr. No

Characteristic

Strategy

5

Diversion structures

Any diversion drains, ditches, and stream channels to divert water from surrounding catchment areas away from the tailings structure should be buil

6

Seepage management

Seepage management should be a key consideration in design and operation of tailings storage facilities A specific piezometer-based monitoring system for seepage water levels can be installed to monitor seepage

7

Freeboard safely

The design parameters should maximum flood event and the required freeboard to safely in to consideration This must be done for the complete planned life of the tailings dam, including its decommissioned phase

8

Acid Mine Drainage

Planning to isolate acid leachate-generating material from oxidation or percolating water On-land disposal alternatives should be designed, constructed, and operated according to internationally recognized geotechnical safety standards

9

Deep-sea tailings placement (DSTP)

Thickening or formation of paste for backfilling of pits and underground workings during mine progression Riverine (for example, rivers, lakes, and lagoons) or shallow marine tailings disposal is not considered good international industry practice; nor is riverine dredging, which requires riverine tailings disposa Deep-sea tailings placement (DSTP) may be considered as an alternative only in the absence of an environmentally and socially sound land-based alternative and based on an independent scientific impact assessment When DSTP is considered, such consideration should be based on detailed feasibility and environmental and social impact assessment of all tailings management alternatives

A Conceptual Digital Framework

523

3.3 Existing Tailings Monitoring Systems There are very few tailings monitoring systems that exist, e.g., Du, Ge, Ng, Zhu, Horgan and Zhang [20] developed and discussed a tailings dam safety system consisting of modern computer technology, network technology, image, and Video transmission technology and GPS satellite positioning technology. Leonida and Carly [25] also discussed the techniques like Inmarsat consists of terrestrial Lidars, seismic sensors, InSAR, total station automation as a guide to tailings dam monitoring. Another research conducted by Wang, Yang and He [22] has designed and implemented a GIS-based tailings safety monitoring system consists of IoT, ZigBee wireless sensors, and wireless data transmission module. Similarly, Sun, Zhang and Li [26] have developed a tailings dam monitoring and pre-alarm system (TDMPAS) based on the IoT and cloud computing with the abilities of real-time tracking of the saturated line, impounded water level and the dam deformation in near real-time. Besides that, many other systems have been devolved by researchers over time. However, none of them is consisting on all the advanced digital technologies for real-time monitoring. Most of the systems are very expensive, without proper graphic user interfaces, time consuming and bulky in terms of installations, hence, there is a need for an effective globally standardize tailings monitoring systems for better management and decision making.

4 Proposed Conceptual Framework for Near Real-Time Monitoring and Management of TSFs Proper management and efficient monitoring of tailing dams are essential to prevent the probability of their failure and mitigation of associated risks. There is no universal standard procedure for monitoring and management because every TSF has its unique design, complex geomechanical characteristics, and potential hazard index. Based on the extensive literature review and best practices, in this research, a comprehensive framework has been conceptualized and proposed for near real-time monitoring and management of tailings storage facilities for intelligent decision making, as shown in Fig. 10. This framework will also serve as the minimum standard criteria for proper monitoring and management of TSFs. There are four key components, i.e., Geotechnical, Geochemical, Social-Environmental, and Hydro-Meteorological, which are necessarily required to put in place for TSF management, monitoring, and risk assessments. In addition, the legal aspects of the country should be considered carefully but are not included in the framework because of variations from country to country. 4.1 Geotechnical Monitoring Geotechnical parameters are critical and used as early warning signs of any possible risk related to dam stability. The most sensitive parameters which should be monitored for both active and inactive TSFs are seismicity, pore-pressure, uplift pressure, surface movement, and internal movement (seepage, internal erosion, and piping) [27]. When there is seismic activity, the TSF structures experience geo-mechanical stresses that can cause the high susceptibility to liquefaction and high pore pressure in faults, which

524

I. Atif et al.

Fig. 10. Proposed conceptual digital framework for TSFs monitoring and management.

becomes the reason for tailings failure. The on-site installations of seismographs can record the natural and mine induced seismicity in the region. High pore pressure can also cause the material to trigger seepage, internal erosion, and piping that can lead to failure of a TSF. Accurate measurements and nature of the distribution of uplift/pore pressure are necessary as it assumes to be the controlling factor in reducing risk. Pressure transducers, which are vibrating wire point type piezometers, can be used to measure the pore pressure at high susceptible segments of the dam. For stability of the TSF structure, surface movement should be monitored. It can be categorized into four subtypes as the horizontal, vertical, rotational, and lateral movement that occurs due to several geo-mechanical stresses [28]. Real-time extensometers, tiltmeters, and Fibre Optic Sensing (FOS) techniques are commonly used instruments for monitoring of surface movement. However, the problem is that these instrumentations can only provide point specific measurements and are can only install in critical areas. Besides these on-site instrumentations, Satellite Remote Sensing (SRS) technology can also be used for efficient monitoring of geotechnical parameters and evaluation of tailings facility regularly [20, 28, 29]. SRS technology can reduce the cost of instrumentation and should be highly recommended for sustainable, efficient, and regular monitoring of TSF [28]. Several case studies use this technology to monitor the surface deformation, and the analysis of results found good agreement between SRS results and ground truth data. The research conducted by Du, Ge, Ng, Zhu, Horgan and Zhang [20], Carlà, Intrieri,

A Conceptual Digital Framework

525

Raspini, Bardi, Farina, Ferretti, Colombo, Novali and Casagli [30] and Iannacone, Lato, Troncoso and Perissin [31] used satellite-based Interferometric Synthetic Aperture Radar (InSAR) data to monitor multi-temporal surface displacement and subsidence of different tailing dams. Similarly, research conducted by Mura, Gama, Paradella, Negrão, Carneiro, De Oliveira and Brandão [32] and Gama, Paradella, Mura and de Oliveira [33] used Differential InSAR for spatio-temporal evaluation of slope stability, surface deformation and risk assessment mapping of TSFs. 4.2 Geochemical Monitoring The two most significant geochemical parameters that should be considered are first, surface and groundwater quality, and second, silting/sedimentation. In the mining industry, surface and groundwater pollution is a significant concern associated with mining operations and sources that release toxic material [34]. Open-pit mines, TSFs, stockpiles, waste rock dumps, and dump leach piles can be the potential sources of surface water contamination [35]. Water is more contaminated when these mining sources are exposed to precipitation. In contrast, groundwater can be polluted due to the percolation of toxic tailing wastewater through its subsurface and foundation. Indirect transfer of pollutants can also occur when there is a hydraulic/hydrological connection between surface and groundwater. Beside that formation of Acid Mine Drainage (AMD) is a significant threat that can wreak havoc on watersheds and contaminate the entire streams and rivers associated with characteristics of rock material and minerals [36]. Parameters that should be monitored to assess the Water Quality are pH, Total Dissolved Solids (TDS), Total Suspended Solids (TSS), and turbidity. Automatic and continuous monitoring sensors play an integral role in analyzing the variability in WQ relatively over a short period, i.e., before, during and after storm events. Nevertheless, the sensor’s traffic-ability, installation, maintenance, and security are a big concern. Several research studies [37–40] investigated that SRS and Geographical Information Systems (GIS) technologies are beneficial to map, model and assess the spatial and temporal variability of water quality parameters, which are not readily available from situ measurements. SRS can be used for mapping of streams sediments yield, geochemical evidence and their concentration, assessment of turbidity, and acid mine waste monitoring using hyperspectral imagery. SRS is also proved useful in monitoring geomorphological and hydro-chemical characterization of mining catchment. Along with SRS technology, GIS is playing a vital role in assessing the risk by predicting the potential consequences of mining waste from TSFs [41, 42]. These technologies are found to be efficient, robust, and cost-effective as they reduce field visits by experts. 4.3 Social-Environmental Monitoring The impacts of TSFs, such as environmental pollution, changes to ecosystems, and biodiversity that aggravates with anthropogenic activities are harmful to communities and the environment. Environmental pollution, like dust in the mining area, needs to be appropriately monitored and recorded [43]. The traditional method to control the susceptibility of TSF to environmental pollution is the collection of samples by dust sampling instrumentation (DST) [44]. Still, at the same time, it is essential to note that it will represent the

526

I. Atif et al.

small area; other issues should be considered as security, cost, and maintenance of DST. GIS and SRS technologies enable to monitor the land and environmental changes over time that helps to identify the potential sources [45–47]. Several studies [48, 49] have used the SRS data for environmental impact assessment and urban settlement changes because of its high temporal and spatial coverage, affordable cost, and enables to get information on remote areas. Thermal and hyperspectral remote sensing is found to be very useful in monitoring TSF environmental disaster and its susceptibility to dust pollution [45]. Mining areas have more exposure to air pollution particulate matter, which can affect human health. Hence, human health in the mining area should be monitored by collecting the data from mine hospitals. Several researchers [46, 50, 51] found that GIS is an efficient platform to analyze and predict the social-environmental hazards and risk that helps the decision-makers in planning and management to achieve sustainable development goals. 4.4 Hydro-Meteorological Monitoring Meteorological conditions including temperature, evaporation, precipitation (rainfall and snowfall), wind speed, and direction are critical to be monitored regularly for each TSF. Extreme weather conditions can cause severe damage to the TSFs and further to their collapse [4]. The weather-related data can be obtained from national weather/meteorological service providers if their system is installed nearby of TSFs. Otherwise, on-site instrumentation is highly recommended, such as an automatic weather station and elevation gauges to provide real-time data to observe the variability in weather conditions, which ultimately leads to monitoring the health of TSF. Besides that, SRS has been widely used for weather forecasting and monitoring activities [52]. Wang, Xu, Yang, Wang, Yuan and Wang [53] used the high-resolution satellite-based rainfall datasets to improve the rainfall simulation. Suseno and Yamada [54] used satellite-based rainfall estimations as input to land surface models for flash floods hazards and risk assessments. Similarly, Patel [55] proposed a GIS-based early warning system based on satellite data and numerical modeling results to issue alerts to the communities of India in case of potential flood hazards. 4.5 Data Analysis and Decision Making The near-real-time data from geospatial sensor networks integrated with Global Navigation Satellite Systems (GNSS) and SRS can be collected through the Mine Internet of Things (IoT) other reliable communication networks at the centralized database server in the form of the geodatabase. A trigger threshold value must be defined for each parameter, which alerts the behavior and health of TSF. The geodatabase can also be used to understand site-specific conditions for problem-solving, which can be done through advanced numerical modeling on selected parameters. Two and three-dimensional scenarios based modeling of slope stability and dam breach is highly recommended for the risk assessment of TSFs [56]. It will help to understand and simulate any possible scenario that can be happened and to map the zone of influence [57]. The risk will be categorizing into different classes, i.e., high risk (unacceptable), medium risk (conditionally acceptable), and low risk.

A Conceptual Digital Framework

527

Risk assessment results and real-time geospatial data will empower mine management because of easy and centralized access, usage, and manipulation to identify the problems, causes, and make efficient decisions. Geodatabase and modeling results are displayed and analyzed in the geospatial portal for the early warning. The geospatial portal can also be accessed remotely on the internet and mobile apps for senior executives of mine management along with the community. The person responsible at the mine will issue the alert to the broader community in case of risk. This proposed framework for TSFs monitoring and management will enable mine management to be proactive rather than reactive, allowing for intelligent decision-making and emergency preparedness when there is an event.

5 Conclusion and Recommendations The management of the TSFs is crucial, not only for the mining industry but also for nearby communities and the economy. In this paper, the current standards for TSFs monitoring and management have been reviewed. Leading practices on design, technology selection, legal requirements, and guidelines were considered. The available best practices from several national and international organizations such as ICOLD, ICMM, ISO, UNEP, ICME, USSD, ADNR, FEMA, MAC, ANCOLD, UNECE, and others were analyzed, and factors were identified in terms of efficient monitoring of TSFs. Based on the literature and best leading practices, a digital, conceptual framework was designed and developed for near real-time monitoring and management of tailings storage facilities in an efficient way. In the proposed framework, the four key components, which need to be monitored, are geotechnical, geochemical, social-environmental, and hydrometeorological. The monitoring of these components can be done either through on-site instrumentation or remotely through satellite technology. In both cases, the centralized database is developed in a GIS environment and used as input to numerical modeling for dam failure analysis. Potential dam failure analysis provides information on the zone of influence, which can be categorized as high, medium, and low risk. The near real-time geospatial data certainly help mine management to make informed decisions. The proposed framework is based on the minimum implementation criteria that should be considered and enhanced with respect to country, TSF type along with other local conditions. It is recommended as a way forward to implement this proposed framework in the broader mining industry for the safety of communities, the trust-building of investors, and the efficiency of mining operations at a global level. Going forward, more parameters can be monitored, e.g., metal and mineral concentrations in the TSF. It will then be possible to attach an economic value to the TSF at any point in time. It will be a significant contribution with the international objective to reduce (even eliminate) waste from mining while supplying the markets with the byproducts from mining. Acknowledgment. The work presented here was conducted as part of a postdoctoral fellowship at the Wits Mining Institute (WMI), University of The Witwatersrand, Johannesburg, South Africa. The authors would like to thank and acknowledge the administrative and financial support provided by the Sibanye-Stillwater Digital Mining Laboratory (DigiMine), WMI.

528

I. Atif et al.

References 1. Kooroshy, J., Meindersma, C., Rademaker, M., Sweijs, T., Diederen, A., Beerthuijzen, M., Goede, S.D.: Scarcity of minerals. A strategic security issue. The Hague Centre for Strategic Studies (2009) 2. Tayebi-Khorami, M., Edraki, M., Corder, G., Golev, A.J.M.: Re-thinking mining waste through an integrative approach led by circular economy aspirations. Miner. 9, 286 (2019) 3. Bowker, L.N., Chambers, D.M.: Root causes of tailings dam overtopping: the economics of risk & consequence. In: Protections 2016: 2nd International Seminar on Dam Protection Against Overtopping. Colorado State University. Libraries (2016) 4. Rico, M., Benito, G., Salgueiro, A., Díez-Herrero, A., Pereira, H.: Reported tailings dam failures: a review of the European incidents in the worldwide context. J. Hazard. Mater. 152, 846–852 (2008) 5. Hui, S., Charlebois, L., Sun, C.: Real-time monitoring for structural health, public safety, and risk management of mine tailings dams. Can. J. Earth Sci. 55, 221–229 (2018) 6. Lv, S.R.: Discussion on stability analysis method for tailings dam. In: Applied Mechanics and Materials, Trans Tech Publ, pp. 93–96 (2012) 7. Vick, S.: Planning, Analysis, and Design of Tailings Dams. BiTech Publishers, Vancouver (1990) 8. Abbott, M., Eldridge, T., Wates, J., Marais, F.: Review of tailings management guidelines and recommendations for improvement. In: International Council on Mining and Metals (ICMM) (2016) 9. Roca, M., Murphy, A., Walker, L., Vallesi, S., McElroy, C., Gimen, O.: A review of the risk posed by the failure of tailings dams. Minimizing the Risk of Tailings Dams failures through the use of Remote Sensing Data. HR Wallingford Ltd. (2019) 10. Vick, S.G.: Discussion to “Analyses of Flow Failures of Mine Tailings Dams” and “Investigation of Flow Failures of Tailings Dams” by Jey K. Jeyapalan, by J. Michael Duncan, and H. Bolton Seed (February, 1983). J. Geotech. Eng. 110, 454–456 (1984) 11. Mainali, G.: Monitoring of tailings dams with geophysical methods [Ph. D. Thesis] [D]. Sweden: Luleå University of Technology (2006) 12. Albuquerque, R., Braga, M.A., Oliveira, L.A., de Oliveira Dias, L.S., Almeida, L.A.P., de Oliveira, A.H., Brandão, S.: Caracterização de barragens de rejeito usando geofísica rasa: aplicação na barragem B1 de cajati, são paulo. Anuário do Instituto de Geociências 42, 567–579 (2019) 13. Fahrenkamp-Uppenbrink, J.: Improving safety at mine tailings dams. American Association for the Advancement of Science (2019) 14. Chen, G., Lian, Z.: The design and implementation of displacement monitoring system for tailings dam. In: 2018 International Conference on Sensor Network and Computer Engineering (ICSNCE 2018), 27 to 29 April 2018, Xi’an, China (2018) 15. USGS, https://www.usgs.gov/media/images/percy-quin-dam-lidar-scan 16. Berberan, A., Ferreira, I., Portela, E., Oliveira, S., Oliveira, A., Baptista, B.: Overview on terrestrial laser scanning as a tool for dam surveillance. In: 6th International Dam Engineering Conference. LNEC, Lisboa (2011) 17. Olivier, G., Brenguier, F., de Wit, T., Lynch, R.: Monitoring the stability of tailings dam walls with ambient seismic noise. The Leading Edge 36, 350a351–350a356 (2017) 18. Ouyang, Z.H.: Research, development and application of surface displacement automatic monitoring system of tailings dam. In: Advanced Materials Research, Trans Tech Publ, pp. 304–309 (2012) 19. Wang, T., Perissin, D., Rocca, F., Liao, M.-S.: Three Gorges Dam stability monitoring with time-series InSAR image analysis. Sci. China Earth Sci. 54, 720–732 (2011)

A Conceptual Digital Framework

529

20. Du, Z., Ge, L., Ng, A.H.-M., Zhu, Q., Horgan, F.G., Zhang, Q.: Risk assessment for tailings dams in Brumadinho of Brazil using InSAR time series approach. Sci. Total Environ. 717, 137125 (2020) 21. Hu, J., Liu, X.: Design and implementation of tailings dam security monitoring system. Procedia Eng. 26, 1914–1921 (2011) 22. Wang, L., Yang, X., He, M.: Research on safety monitoring system of tailings dam based on internet of things. In: IOP Conference Series: Materials Science and Engineering, IOP Publishing, p. 052007 (2018) 23. UNEC: Safety guidelines and good practices for Tailings Management Facilities (2014) 24. Penman, A.: Tailings dams: risk of dangerous occurrences. In: Geoenvironmental Engineering: Geoenvironmental Impact Management: Proceedings of the third conference organized by the British Geotechnical Association and Cardiff School of Engineering, Cardiff University, and held in Edinburgh on 17–19 September 2001, Thomas Telford Publishing, pp. 150–156 (2001) 25. Leonida, C.: Technology to tackle tailings. Eng. Mining J. 220, 40–45 (2019) 26. Sun, E., Zhang, X., Li, Z.J.: The internet of things (IOT) and cloud computing (CC) based tailings dam monitoring and pre-alarm system in mines. Saf. Sci. 50, 811–815 (2012) 27. Ozcan, N.T., Ulusay, R., Isik, N.S.: A study on geotechnical characterization and stability of downstream slope of a tailings dam to improve its storage capacity (Turkey). Environ. Earth Sci. 69, 1871–1890 (2013) 28. Goff, C., Gimeno, O., Petkovsek, G., Roca, M.: Minimising the risks of tailings dams with remote sensing data (2019) 29. Rauhala, A., Tuomela, A., Davids, C., Rossi, P.M.: UAV remote sensing surveillance of a mine tailings impoundment in Sub-Arctic conditions. Remote Sens. 9, 1318 (2017) 30. Carlà, T., Intrieri, E., Raspini, F., Bardi, F., Farina, P., Ferretti, A., Colombo, D., Novali, F., Casagli, N.: Perspectives on the prediction of catastrophic slope failures from satellite InSAR. Sci. Reports 9, 1–9 (2019) 31. Iannacone, J.P., Lato, M., Troncoso, J., Perissin, D.: InSAR monitoring of active, inactive and abandoned tailings facilities. In: Proceedings of the 5th International Seminar on Tailings Management, Santiago, Chile, pp. 11–13 (Year) 32. Mura, J.C., Gama, F.F., Paradella, W.R., Negrão, P., Carneiro, S., De Oliveria, C.G., Brandão, W.S.: Monitoring the vulnerability of the dam and dikes in Germano iron mining area after the collapse of the tailings dam of Fundão (Mariana-MG, Brazil) using DInSAR techniques with TerraSAR-X data. Remote Sens. 10, 1507 (2018) 33. Gama, F.F., Paradella, W.R., Mura, J.C., de Oliveira, C.G.: Advanced DINSAR analysis on dam stability monitoring: a case study in the Germano mining complex (Mariana, Brazil) with SBAS and PSI techniques. Remote Sens. Appl. Soc. Environ. 16, 100267 (2019) 34. Rey, N.J., Demers, I., Bussière, B., Mbonimpa, M., Gagnon, M.: A geochemical evaluation of a monolayer cover with an elevated water table for the reclamation of the Doyon-Westwood tailings ponds. Can. Environ. Earth Sci. 79, 1–14 (2020) 35. Qi, Z., Xiao, C., Zhang, B., Liang, X., Wang, G.: Risk assessment of groundwater pollution in a mining area based on physical processes. In: E3S Web of Conferences, EDP Sciences, p. 09025 (2019) 36. Rezaie, B., Anderson, A.: Sustainable resolutions for environmental threat of the acid mine drainage. Science of The Total Environment, p. 137211 (2020) 37. Topp, S.N., Pavelsky, T.M., Jensen, D., Simard, M., Ross, M.R.: Research trends in the use of remote sensing for inland water quality science: moving towards multidisciplinary applications. Water 12, 169 (2020) 38. Flores-Anderson, A.I., griffin, R., Dix, M., Romero-Oliva, C.S., Ochaeta, G., SkinnerAlvarado, J., Ramirez Moran, M.V., Hernandez, B., Cherrington, E.: Hyperspectral satellite remote sensing of water quality in lake Atitlán, Guatemala. Frontiers Environ Sci. 8, 7 (2020)

530

I. Atif et al.

39. Abdelmalik, K.: Role of statistical remote sensing for Inland water quality parameters prediction. Egypt. J. Remote Sens. Space Sci. 21, 193–200 (2018) 40. Mahboob, M., Genc, B., Celik, T., Ali, S., Atif, I.: Mapping hydrothermal minerals using remotely sensed reflectance spectroscopy data from Landsat. J. South Afr. Inst. Min. Metall. 119, 279–289 (2019) 41. Mallik, S., Mishra, U.: Assessment of Spatial Variations in Groundwater Quality of Agartala, Tripura for Drinking Employing GIS and MCDA Techniques. Applications of Geomatics in Civil Engineering, pp. 273–288. Springer (2020) https://doi.org/10.1007/978-981-13-70670_22 42. Verma, P., Singh, P.K., Sinha, R.R., Tiwari, A.K.: Assessment of groundwater quality status by using water quality index (WQI) and geographic information system (GIS) approaches: a case study of the Bokaro district India. Appl. Water Sci. 10, 27 (2020) 43. Tian, S., Liang, T., Li, K.: Fine road dust contamination in a mining area presents a likely air pollution hotspot and threat to human health. Environ. Int. 128, 201–209 (2019) 44. Moja, S., Kwata, M., Sebesho, L., Masindi, K., Mtunzi, F.: Characterization of Surface and Trapped Dust Samples Collected around Human Settlements That are in the Vicinity of Old Mine Tailings in Mpumalanga Province, South Africa. J Earth Sci Clim Change 7, 2 (2016) 45. Zwissler, B.: Dust susceptibility at mine tailings impoundments: thermal remote sensing for dust susceptibility characterization and biological soil crusts for dust susceptibility reduction (2016) 46. Kandakji, T., Gill, T.E., Lee, J.A.: Identifying and characterizing dust point sources in the southwestern United States using remote sensing and GIS. Geomorphology, p. 107019 (2020) 47. Mahboob, M.A., Atif, I.: Geospatial modelling for groundwater quality with contamination potential assessment of Three Urban Areas in Pakistan. International Journal Of Water Resources & Arid Environments (IJWRAE) 6, 11 (2017) 48. Che, D., Liang, A., Li, X., Ma, B.: Remote sensing assessment of safety risk of iron tailings pond based on runoff coefficient. Sens. 18, 4373 (2018) 49. Rigina, O.: Environmental impact assessment of the mining and concentration activities in the Kola Peninsula, Russia by multidate remote sensing. Environ. Monit. Assess. 75, 13–33 (2002) 50. Boloorani, A.D., Samany, N.N., Mirzaei, S., Bahrami, H.A., Alavipanah, S.K.: Remote Sensing and GIS for Dust Storm Studies in Iraq. Environmental Remote Sensing and GIS in Iraq, pp. 333–375. Springer (2020) 51. Mahboob, M.A., Atif, I., Iqbal, J.: Remote sensing and GIS applications for assessment of urban sprawl in Karachi, Pakistan. Sci. Technol. Dev. 34, 179–188 (2015) 52. Sheffield, J., Wood, E.F., Pan, M., Beck, H., Coccia, G., Serrat-Capdevila, A., Verbist, K.: Satellite remote sensing for water resources management: potential for supporting sustainable development in data-poor regions. Water Res. Res. 54, 9724–9758 (2018) 53. Wang, J., Xu, Y., Yang, L., Wang, Q., Yuan, J., Wang, Y.: Data assimilation of high-resolution satellite rainfall product improves rainfall simulation associated with landfalling tropical cyclones in the Yangtze river Delta. Remote Sens. 12, 276 (2020) 54. Suseno, D.P.Y., Yamada, T.J.: Simulating flash floods using geostationary satellite-based rainfall estimation coupled with a land surface model. Hydrol. 7, 9 (2020) 55. Patel, A.: Rainfall-Runoff Modelling and Simulation Using Remote Sensing and Hydrological Model for Banas River, Gujarat, India. Advances in Water Resources Engineering and Management, pp. 153–162. Springer (2020) 56. Mahdi, A., Shakibaeinia, A., Dibike, Y.B.: Numerical modelling of oil-sands tailings dam breach runout and overland flow. Sci. Total Environ. 703, 134568 (2020) 57. Das, T., Hegde, A.: A Comparative Deterministic and Probabilistic Stability Analysis of Rock-Fill Tailing Dam. Advances in Computer Methods and Geomechanics, pp. 607–617. Springer (2020)

Research on the Use of Fly Ash for Underground Mine Supports in Quang Ninh Coal Area Trong Hung Vo1 , Van Kien Dang1(B)

, Ngoc Anh Do1

, and Van Ha Truong2

1 Department of Underground and Mining Construction, Faculty of Civil Engineering,

Hanoi University of Mining and Geology, Hanoi, Vietnam [email protected] 2 Underground Construction Engineering K38, Faculty of Civil Engineering, Hanoi University of Mining and Geology, Hanoi, Vietnam

Abstract. The main objective of this research is to propose an optimized proportion of the fly ash of thermal power plants used in the concrete mixture for underground mine supports in Quang Ninh coal area. Fly ash is used as a partial replacement of cement in fresh concrete for underground mine supports to reduce the cost of drift supporting and to increase the effects of the environment protection. This study also shows that the new trend in using new material to make rock/soil support in the underground mines in Vietnam. Fly ash utilization in concrete as partial replacement of cement is gaining importance day by day. Technological improvements in thermal power plant operations, as well as collection systems of fly ash, improved the quality of fly ash. For the use of fly ash in concrete, cement is replaced partially by fly ash in concrete by laboratory tests at the construction laboratory of Hanoi University of Mining and Geology (HUMG). In this experimental work, concrete mix was prepared with the replacement of fly ash by 0%,10%, 20%, 30%, and 40%. Effect of fly ash on workability, compressive strength, and flexural strength of underground concrete are studied. To study the impact of partial replacement of cement by fly ash on the properties of concrete, experiments were conducted on different concrete mixes at the laboratory. The effectiveness of the underground mine support with the concrete insert plate in the SVP steel frame is carried by analytical methods. The comparison of drift support systems using the concrete insert plates in SVP steel frame made by fly ash and conventional support mining shows the advantages of using fly ash. From the result of this study, it can be concluded that cement replacement by fly ash is useful in lower grades of cement, such as M20. It can be stated that at 30% of the replacement of cement by fly ash, there is a considerable increase in strength properties. Incorporation of fly ash in concrete can save the coal and thermal industry disposal cost and produce a “greener” concrete for construction. With the use of mineral admixture, the cost is considerably reduced since mechanical vibrators and viscosity modifying admixtures are not required. The strength of concrete decreases with increases in the percentage of fly ash first. The utilization of fly ash gives benefits in terms of economic and technical aspects. Using support with fly ash can reduce the cost of drift support and improve the effect of environmental protection.

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 531–549, 2021. https://doi.org/10.1007/978-3-030-60839-2_28

532

T. H. Vo et al. Keywords: Fly ash · Underground mining support · Mine drift · Concrete mix · Environment protection

1 Introduction The use of concrete has recently gained popularity as resource-efficient, durable, and cost-effective. Ordinary Portland Cement (PC) is widely used as the main building material; however, its manufacturing process results in high emission of carbon dioxide (CO2 ) to the atmosphere. The production of cement is not environmentally friendly as it damages nature in search of limestone and energy involved. Fly ash, also known as “pulverized fuel ash”, is a coal combustion product composed of fine particles that are driven out of the boiler with the flue gases. Fly ash is an industrial waste. A concrete mix with fly ash can provide environmental and economic benefits. Fly ash concrete enhances the workability, compressive strength, flexural strength, and also increases its permeability, durability, and concrete finishing. It also reduces corrosion, alkalisilica reaction, sulfate reaction shrinkage as it decreases its permeability and bleeding in concrete. Sample of fly ash is presented in Fig. 1. There are two basic types of fly ash: Class F and Class C. Both types react in concrete in similar ways. Both Class F and Class C fly ashes undergo a “pozzolanic reaction” with the lime (calcium hydroxide) created by the hydration (chemical reaction) of cement and water, to create the same binder (calcium silicate hydrate) as cement. In addition, some Class C fly ashes may possess enough lime to be self-cementing, in addition to the pozzolanic reaction with lime from cement hydration. The disposal of fly ash is a severe environmental problem. The fly ash used in concrete industry by partly replacement of cement and in embankment for filling the material. Leaving the waste materials to the environment directly can cause an environmental problem. Hence, the reuse of waste material has been emphasized. These industrial wastes are dumped in the nearby land, and the natural fertility of the soil is spoiled. Cement with fly ash reduces the permeability of concrete and dense calcium silicate hydrate (C-S-H). Past research shows that adding fly ash to concrete, as a partial replacement of cement (less than 35%), will benefit both the fresh and hardened states. While in the fresh state, the fly ash improves workability. This is due to the smooth, spherical shape of the fly ash particle. The tiny spheres act as a form of a ball bearing that aids the flow of the concrete. This improved workability allows for lower water-to-cement ratios, which later leads to higher compressive strengths. In the hardened state, fly ash contributes in several ways, including strength and durability. In comparison, fly ash tends to increase the setting time of the concrete. The pozzolanic reaction is removing the excess calcium hydroxide, produced by the cement reaction, and forming a harder CSH. The fly ash was used in many purposes of underground construction as a lining tunnel and other supports [1, 2]. For example, the new St. Clair River Tunnel was constructed in 1993–1994 between Sarnia in Ontario and Port Huron in Michigan (Fig. 2a). Fly ash is also used in the lining of Delhi Metro Rail Corporation (Fig. 2b). The groundwater contained chlorides (4000 ppm) and sulfates (155 ppm), and this environment, combined with hydrostatic heads of up to 35 m, led to the inclusion of both chloride diffusion and

Research on the Use of Fly Ash for Underground Mine Supports

533

permeability limits in the concrete specification for the precast tunnel lining segments. The requirements for the concrete were [5]: • • • •

Cementitious content from 400 to 550 kg/m3 ; w/cm ≥ 0.36. Compressive strength ≥ 60 MPa at 28 days. Chloride diffusion coefficient, Da ≥ 600 × 10 − 15 m2 /s at 120 days. Water permeability, k ≤ 25 × 10 − 15 m/s at 40 days.

The concrete was produced at a local ready-mixed concrete plant and delivered in a transit mixer to the precast plant. The concrete mixture used at the start of the production process contained 6% silica fume and 30% Class C fly ash with w/cm ranging from 0.29 to 0.32. This mix met specifications including the chloride diffusion coefficient at 120 days. Advantages of fly ash in cement concrete [3, 6, 7, 15]: • Reduction in the heat of hydration and thus the reduction of thermal cracks and improves soundness of concrete mass; • Improved permeability of concrete and workability; • Converting the excess lime into binding material through the hydration process. • Improved impermeability; • Reduced cost of concrete for the same strength due to reduced cement requirement; • Increases the modulus of elasticity of concrete when concretes of the same strength with and without fly ash are compared; • Improved workability; • Improved sulfate resistance; • Greater resistance to attack of aggressive water.

Fig. 1. Sample of fly ash [1, 2]

534

T. H. Vo et al.

a) St. Clair river tunnel precast tunnel lining segments (Courtesy Hatch Mott MacDonald) [5]

b) Using fly ash in Delhi Metro Rail Corporation [20]

Fig. 2. Examples of using fly ash in underground concrete

Effect of using fly ash in fresh and hardened concrete: • Using fly ash in concrete contributes to the strength of concrete due to its pozzolanic reactivity of fly ash; • The pozzolanic reaction also provides with the texture of concrete dense, resulting in a decrease of water permeability and gas permeability; • It should be noted that since the pozzolanic reaction can only proceed in the presence of water, fly ash concrete should be cured for a longer time; • Dams will derive the full benefits of attaining improved long-term strength and water tightness. Various numbers of research have been conducted to examine the effects of the use of fly ash as an additive in cement, admixture in concrete, and as replacement of cement in concrete. The compressive strength of concrete was tested by replacing different proportions of cement with suitable quantities of fly ash, and the results have been found most useful and applicable [4–8]. However, most of the research works have been conducted only for a limited percentage of 5–20% of cement replacement for a lower grade of concrete M200, M300. It is, therefore, necessary to conduct extensive research on compressive strength of different qualities of concrete as well as different proportions of fly ash at different curing periods and for underground concrete. Typically, during the manufacturing of cement at plant designers usually opt for about 15–30% of fly ash of total cementitious material. However, there have been many studies that have been taken place for the past 15 years for using high volume or a high percentage of fly ash in concrete. It is properly designed and constructed there will be some benefits of concrete, which have got 40%, 50%, and 60% fly ash replacements include dramatically reduced concrete permeability and excellent resistance to all forms of premature deterioration. In Vietnam, fly ash was used to as roller compacted concrete of Son La hydroelectricity, rural road concrete, and unburnt bricks. However, they still have not been used for underground coal mine supports. Now, in underground mines of Vietnam National CoalMineral Industries Holding Corporation Limited (VINACOMIN), concrete insert plates with steel arches (SVP steel frame, I frame, U frame, etc.) are the biggest proportion of

Research on the Use of Fly Ash for Underground Mine Supports

535

Table 1. Physical property of fly ash [14] Specific gravity

2.146 to 2.42

Wet sieve analysis (% retained on No. 325 BS sieve) 51.00 (dry) Specific surface (cm2 /g)

2800 to 3250

Lime reactivity (kg/cm2 )

56.25 to 70.3

the type of current structural support (about 70-80%). Concrete insert plates are one part of rock/soil support of SVP steel frame in underground mines. They are responsible for closing the gap between the SVP steel frames in drift, and the dug evenly distributing the pressure of rock and soil on SVP steel frames, preventing roof landslides and sidewalls. Selecting concrete insert plates with dimension of section: b × h = 150 × 50 mm with the length equal to the distance of SVP steel frames: L = 0.7 m, L = 0.9 m, and L = 1.0 m, depending on the characteristic of mine pressure where the drift thought rock mass. The calculation of the dimension of the cross-section of the concrete insert plates is made by treating the as a beam on two supports in which the distance between two supports equals to the distance of SVP steel frames. The SVP steel frame is located in the middle of the roof, subject to the uniformly distributed load of arches destroyed (see Fig. 3).

a)

b)

Fig. 3. Calculation diagram of the concrete insert plate in SVP steel frame. a) Calculation model of concrete insert plate in SVP steel frame; b) Load diagram

At present, many underground mines, due to the quality of concrete insert plates, construction methods, and wet environment of drift, the increase in pressure compared to the design. They make concrete insert plates broken, cracked, such as Ha Lam coal mine, Mong Duong coal mine, Nam Mau coal mine. According to the survey results in some underground mines in Quang Ninh coal area, the concrete insert plate is damaged due to the large deformation of the drifts in complicated geological conditions. Transporting concrete insert plate from the ground to the support position in deep underground mining (−400 m to −500 m) takes quite a bit of time due to the large size and volume of the insert. In addition, because the insert plate has a bar structure, it is easy to be damaged during transportation and use, leading to re-make rock/soil support, increasing costs, and time in drift excavation. Therefore, the reduction of the size and weight of the concrete insert plate will contribute to reducing the difficulty of construction, thereby speeding up drift excavation.

536

T. H. Vo et al.

a) Using concrete insert plate in drifts in underground mining b) The concrete insert plate structure Fig. 4. Structure of concrete insert plates with length of L = 0.7 m in Ha Lam coal mine, Quang Ninh [1]. a) The concrete insert plate in SVP steel frame at a drift of Ha Lam coal mine b) Concrete insert plate structure

According to the results of the re-calculation, the concrete insert plates by the diagram on Fig. 3 show that the most significant bending moment in concrete insert plates is higher than the permissible bending moment. In addition, due to the wet drift condition, the reinforcement steel in the concrete insert plates quickly erodes the durability and causes instability. This paper presents the effect of fly ash replacement on compressive strength and flexural strength of concrete along with the slump and other fresh and hardened properties. A comparative cost investigation with the different replacement of fly ash has been presented. The result of the above comparison will be chosen as the reasonable value of fly ash replacement in the concrete mixture on making support of underground mining in underground mines in Quang Ninh coal area.

2 Methodology The test was carried out in the construction laboratory of Hanoi University of Mining and Geology (HUMG) by Advantest 9 (Control – Italy) system (see Fig. 5). Materials used in the time to time experiments are as under: Cement: Ordinary Portland cement (PCB 30 Nghi Son) was used having specific gravity of 3.10, 31.5% consistency, and compressive strength of 30 MPa. Fly ash: From the combustion of pulverized coal and transported by the flue gases of boilers by pulverized coal, fly ash is produced. It was obtained from Pha Lai thermal power station (in Quang Ninh, Vietnam), dried and subsequently used.

Research on the Use of Fly Ash for Underground Mine Supports

537

Fine aggregate: Natural sand with a maximum size of 0,14–5 mm (by TCVN 75722006) was used with specific gravity of 2.55 and fineness modulus of 2.61. Coarse aggregate: Natural aggregates with a maximum size of 20 mm at Hoa Thach rock mine (in Quoc Oai, Hanoi) were used with a specific gravity of 2.68 and fine modulus of 7.5. Water: Drinking water from HUMG was used for the preparation of concrete. The quality was uniformed and the water samples were potable. The concrete mix was designed for M20 grade, and the mix design was done as per TCVN 10302:2014 [4]: Activity admixture- Fly ash for concrete, mortar, and cement. The properties of constituents of concrete were taken into account for mix design of concrete. Different concrete mixes with varying fly ash content percentages were produced, replacing 0% (reference concrete), 20% cement in terms of weight (Table 1). For compressive strength tests, cube specimens of 150 mm size were cast. The cubes were cast in stainless steel moulds and wet cured at standard temperature until the time of the test. The cubes were cured for a period of 28 days. Three samples from each set of the mix were tested at the age of 28 days for compressive strength 28 days for flexural strength. Preparation of specimen: All concrete mixes were prepared using a motorized mixer of mixed design proportion, 1:1.27:2.83, with a constant water-cement ratio of 0.42. Cube specimens are prepared of size 150 mm × 150 mm × 150 mm and beam specimen of 100 mm × 100 mm × 500 mm (see Fig. 4). The samples were cured in a curing room at 30°C temperature and 90% relative humidity. Fly Ash mix concretes were tested at 28 days of age to get compressive strength and 28 days for flexural strength values. Servo-hydraulic system for static and low-frequency dynamic tests on building materials under control of Load/Stress, Displacement Strain (see Fig. 5) is available on Advantest 9 (Control - Italy) system in the construction laboratory of HUMG. Ideal both for traditional tests, such as compression and flexure on concrete, cement, mortar, blocks, etc. and cyclic tests for the determination of secant elastic modulus (E) according to all relevant international standards, and also for measuring, for example, the ductility and fracture energy of concrete reinforced with fibers (FRC) and lined with polymers (FRP), or the toughness of sprayed concrete slabs (shotcrete) under concentrated load tests can be carried by this system. The console is connectable to up to four test frames.

Fig. 5. Advantest 9 (Control - Italy) system of HUMG [1, 2]

538

T. H. Vo et al.

Features of Advantest 9 (Control - Italy) system are as under: • Performance: User-defined test procedures that can be easily performed under load, specific load, displacement, and strain control. • Flexibility: Possibility to connect up to 4 different testing frames from 15 to 5000 kN load capacity, easily selectable by the user-friendly software. • Accuracy and reliability: Long-life system due, essentially, to the advanced electronic, the efficiency of the closed-loop system, P.I.D. control adapted to the test, and very high resolution. • Interactive software: To perform: – Remote control of the system. – Monitoring and display of all test data and parameter either in graphic or numerical format. – File management by building materials, tests, specifications, clients etc. – Print of standard or customized test certificate. – Real-time variation of the setting including the control method (load, displacement or strain). – User-friendly interface. • Extra channels: In addition to the four channels used for the connection of up to four separate test frames, an additional four channels are provided for connection to the displacement transducers, pressure transducers, load cells, strain gauges or similar sensors, which can be configured by the user conforming to the test requirement on a case by case basis. The compressive strength limits of concrete is determined by the TCVN 3118–1993 standard. The compressive strength of concrete is determined by the following formula:

Rn =

P F

(1)

where: P - destructive load of samples (N); F - cross-section area of the sample (mm2 ). Tests for flexural strength: This test was carried out on a specimen of concrete called as prism or beam and the flexural strength sometimes called as modules of rupture the test was done in accordance to the IS specification, and the strength can be calculated by the formula given below [14]: fb =

 pl , N mm2 bh2

(2)

Where: b - width of the specimen in mm; d - depth in mm of the specimen during of failure in mm; l - length of the specimen in mm on which it was placed in mm; p highest load applied on to the specimen in kN. The number of prisms, tested for different proportions with CVC, fly ash, and activated fly ash are shown in Tables.

Research on the Use of Fly Ash for Underground Mine Supports

539

• Mix Design: The preferred characteristic strength of 30 N/mm2 at 28 days was used in this study. The concrete mix was designed for M20 grade, and the mix design was done as per TCVN 10302:2014 [4]. A total of 45 cubes and 30 beams were prepared for this study in 5 sets. All sets were prepared in the control mix of water-cement ratio 0.42. Three samples from each set of the mix were tested at the age of 7, 14, and 28 days for compressive strength and 7 and 28 days for flexural strength. Table 2. Compare mixtures of M200 concrete in making concrete insert plate of SVP steel frame in underground mines between two methods Materials

Conventional concrete mix

Concrete mix using fly ash to make concrete insert plate

Cement PCB 30 (kg)

312.0

Coarse aggregate(kg)

1251.0

1216.0

Fine aggregate (kg)

540.0

568.7

Water (liter)

195.0

195.0

Fly ash (kg)

0.0

Plasticizer additives (kg) Total: Cement + Coarse aggregate + Fine aggregate + Fly ash; (kg)

193.83

83.07

0.0

3.87

2298.0

2260.47

• Preparation of Specimen: All concrete mixes were prepared using a motorized mixer of mixed design proportion, 1:1.27:2.83 with a constant water-cement ratio of 0.42. Cube specimens are prepared of size 150 mm × 150 mm × 150 mm and beam specimen of 100 mm × 100 mm × 500 mm (see Fig. 6). The specimens were cured in a curing room at 30°C temperature and 90% relative humidity. Fly ash mix concretes were tested at 28 days of age to get compressive strength 28 days for flexural strength values (Fig. 7 and Fig. 8).

Fig. 6. Specimen prepared for flexural strength value tests

540

T. H. Vo et al.

Fig. 7. Curing of Specimens

a)

b)

Fig. 8. Compressive strength test on the cube sample by Advantest 9 (Control - Italy) at HUMG

Table 3. Details of specimens prepared for the test at construction laboratory of HUMG No. Details of Cube Specimen

Details of Beam Specimen

Slump value Name of Fly Weight Name Fly ash Weight of (%) cube ash of fly of (%) fly sample (%) ash in beam ash in mix mix sample (gm) (gm)

1

C0

0

0

B0

0

0

40

2

C 10

10

156

B 10

10

235

38

3

C 20

20

312

B 20

20

470

35

4

C 30

30

468

B 30

30

705

32

5

C 40

40

624

B 40

40

940

30

Saving in cement: 312 – 193.83 = 118.17 kg/m3 and the weight is lighter 37.3 kg/m3 . The results on Fig. 9 show that the workability of concrete decreases with the increase in fly ash, the particles of fly ash reduces the amount of water required to produce a given slump. The circular shape of the fly ash particles and its dispersive ability provide

Research on the Use of Fly Ash for Underground Mine Supports

541

Table 4. Test results of compressive strength No. Cube sample Fly ash (%) 28 days strength name N/mm2 (Average of three samples) 1

B0

0

40.15

2

B 10

10

39.60

3

B 20

20

41.15

4

B 30

30

43.19

5

B 40

40

40.28

Table 5. Test Results of Flexural strength No

Beam sample name

Fly ash (%)

28 days Strength N/mm2 (Average of three samples)

1

B0

0

6.17

2

B 10

10

6.20

3

B 20

20

6.33

4

B 30

30

6.42

5

B 40

40

6.16

water-reducing characteristics. The result on Fig. 10 also shows that the compressive strength and flexural strength increases with the increase of fly ash in concrete up to 30% replacement with cement in the conventional mix, however, the percentage increase of compressive strength more than compared to the percentage increase of flexural strength about 2.8% at 30% replacement with cement (see Fig. 12). Mixing of fly ash in conventional concrete mix has resulted in considerable variation in the properties of fresh concrete. The integration of fly ash in concrete increased the cohesiveness of the mix, prohibited segregation, and resulted in reduced bleeding. Higher percentages of fly ash can cause a change in the color of the mix. Incorporation of fly ash in concrete can save the coal and thermal industry disposal costs and produce a green concrete for underground construction. The result in Figs. 11 and 12 show that the compressive strength of the concrete cubes had been tested at the interval of 28 days. It seems that the strength goes on the increase with the increase in fly ash, but after the replacement of 30% fly ash with cement, the strength decreases. The flexural strength of concrete is tested at the interval of 28 days, and it has seemed that flexural strength goes on increase up to 30% replacement. The strength variation is more on compressive as compared to flexural strength.

542

T. H. Vo et al.

Fig. 9. Workability test of concrete mix with different percentage of fly ash

Fig. 10. Compressive strength test of concrete mix with different percentage of fly ash

3 Method of Fly Ash Concrete Mixing For obtaining the best result, the fly ash concrete should be prepared by the following mixing method: About ¾ quantity of the mixing water be taken in the concrete mixer. Weighted amount of the required amount of fly ash then added to it and mixed for 30 s. To the slurry of fly ash so obtained, weighted quantities of coarse aggregate, fine aggregate, cement, and remaining amount of the mixing water be added and mixed for 90 s. However, if this is not convenient, the normal mixing method may be adopted i.e., Weighted quantities of coarse aggregate, fine aggregate cement, and fly ash should be put together in the concrete mixer and mixed dry for 30 s. The required amount of the

Research on the Use of Fly Ash for Underground Mine Supports

543

Fig. 11. Flexural strength test of concrete mix with different percentage of fly ash

Fig. 12. 28 days percentage increase of strength with different percentage of fly ash

mixing water was then added, and the mixing continued for 90 s. The superplasticizer should be added just before the discharge of the mix from the mixer.

4 The Results of Concrete Mixing Design for Concrete Insert Plates According to the above research results, the determine the aggregate content for 1 m3 of concrete M200 used for manufacturing concrete insert plates (see Table 2). When calculating mortar concrete mix for M200 concrete, insert plates using fly ash instead of a part of cement in concrete gradation of the research team have compared results of two alternatives (see Table 3). Thus, from the total weight between the two methods of concrete mix, we can realize that the use of fly ash to replace a part of cement will

544

T. H. Vo et al.

be saved in cement: 312 – 193.83 = 118.17 kg/m3 and 337.3 kg lighter than normal concrete grade (see Table 4). In addition, this study also calculated the cost of materials for concrete insert plates for drift No .2 level -140 Lo Tri, Thong Nhat coal company with excavation area Sd = 15.3 m2 , the rock consolidating coefficient f = 4–6 (by M. N. Protodyakonov); the later one is related to the mechanical parameters of the ground f = tanϕ, L = 0.7 m (see Table 6) and f = 7–8 (see Table 7) with the L = 0.9 m, 40 mm thickness of concrete insert plates, compared to the concrete insert plates used 50 mm in thickness (Table 5 and Table 8). Table 6. Summing of main material for a concrete insert plate with L = 0.7 m in length, dimensions b × h = 150 mm × 40 mm Type of concrete insert plates

Steel bar N°

Diameter of steel bar (mm)

Length of one steel bar (mm)

Quantity of steel bar

Total of bar steel (mm)

Amount (kg) 1m

Total

Flyash used (m3 )

L= 0.7 m

1

6

730

3

2190

0.222

0.486

0.0042

2

6

150

6

1020

0.222

0.226

Total

0.713

Table 7. Summing of main material for a concrete insert plate with L = 0.9 m in length, dimensions b × h = 150 mm × 40 mm Type of concrete insert plates

Steel bar N°

Diameter of steel bar (mm)

Length of one steel bar (mm)

Quantity of steel bar

Total of bar steel (mm)

Amount (kg) 1m

Total

Flyash used (m3 )

L= 0.9 m

1

6

930

3

2790

0.222

0.619

0.0054

2

6

170

6

1020

0.222

0.226

Total

0.845

The design and calculation of the concrete insert plates of the drift is carried out according to the following steps: Calculation the maximum bending moment in concrete insert plates with L in length and bxh in section: Mmax =

γ .c . H . L2 ; kN .m 8

(3)

where: γ is represents the specific weight of the ground (kN/m3 ); c is width of concrete insert plates (m).

Research on the Use of Fly Ash for Underground Mine Supports

545

Table 8. Statistics of ash and slag discharged in thermal power plants in Vietnam [9] N0

Coal-fired power plants

Allotted ash and slag disposal area (ha)

1

Pha Lai

27.7

0.802

0

2

Uong Bi MR 29.0

0.337

0.75

4.5

3

Quang Ninh

15.0

0.945

4.80

7.0

4

Hai Phong

56.0

0.949

1.0

3.0

5

Nghi Son 1

74.7

0.359

0.85

6.3

6

Vinh Tan 2

38.37

1.371

3.90

2.9

7

Duyen Hai 1

31.0

1.006

1.73

1.4

8

Duyen Hai 3

29.0

1.112

0.065

9.6

9

Mong Duong 24.0 1

1.125

1.70

0.8

10

Na Duong 1

57.6

0.485

2.002

15.0

11

Cao Ngan

4.4

0.260

1.806

10.0

12

Son Dong

2.2

0.605

2.020

4.0

13

Dong Trieu

24.0

1.048

1.255

7.0

14

Cam Pha 1 + 47.0 2

1.530

5.535

6.0

15

Nong Son

0.052

0.108

25.0

12.530

27.521

10.7

Total

Volume of ash, slag discharged (million tonnes/year)

Total volume contained (million tonnes)

Enough years to contain (year) 25.0

H is the height of the equilibrium arch (natural equilibrium arch) formed on the roof of the tunnel according to some hypotheses of pressure arch around the underground structures after excavation. One of the main assumptions of the Protodyakonov’s equilibrium arch theory is that the ground behaves like a loose medium. During the excavation of a tunnel, the strata are disturbed, and an arch-like disturbed zone is developed in the strata, just above the tunnel roof (see Fig. 13) [16]. According to the Protodyakonov’s equilibrium arch theory, the vertical loosening pressure q can be obtained as:  q = γ H, kN m (4) where H is given by: H=

B + 2Ttanθ B = 2f 2f

(5)

Figure 13 illustrates the general scheme of Protodyakonov’s equilibrium arch theory. T and B, respectively, represent the height and the width of the tunnel. H is the height of the equilibrium arch. W is the width of the intersection between the equilibrium arch

546

T. H. Vo et al.

Fig. 13. Schematic illustration of Protodyakonov to calculation the H value [16]

and the horizontal plane at the top of the tunnel. θ is the sliding angle between the shear plane and the vertical plane, which is related to the mechanical parameters of the ground, and could be: θ = (45◦ − ϕg /2)

(6)

where ϕg represents the calculated friction angle of the ground. Note that, ϕg is not the real internal friction angle of the ground. It is mainly a function of the cohesion c and the internal friction angle ϕ of the ground. In other words, the Protodyakonov’s theory simplifies a two-parameter medium (i.e., c and ϕ) as an ideal loose medium identified by only ϕg . The maximum allowed bending moment is according to the formula: Mmax = [σu ]. W, kN.m

(7)

Figure 14 shows a diagram of internal force calculation and design of reinforcement concrete structure concrete insert plate with a rectangular cross-section (see Fig. 14). The steel area in the rectangular cross-section of the insert plate of the tension region is calculated by Mmax value from Eq. (3). Result of design of reinforcement concrete structure concrete insert plate is chosen 3 steel bars: N01 , L = 730 cm, 6 mm in diameter and N02 , 6 steel bars, L = 150 mm, 6 m in diameter. The result of the design of reinforcement concrete structure concrete show in Tables 6 and 7 with the corresponding L = 0.7 m and L = 0.9 m in length. Based on the calculation parameters of concrete insert plates made by apart of fly ash, we carried out to make the support passport for some drifts in Mao Khe coal mine company. By the above calculation result, we carried out to make 4,000 concrete insert plates from fly ash at Cam Pha, Quang Ninh, Vietnam, and tested as a rock/soil support in Mao Khe coal company (see Fig. 15). The result of in-situ testing will be presented on the next future.

5 Effects of Environmental Protection Due to Using Fly Ash to Making Concrete Insert Plate Fly ash can be defined as a waste residue that is released from the coal combustion process in electric power stations. Fly ash is the unburned residue that is carried away from the burning zone in the boiler by the flue gases and then collected by either mechanical or electrostatic separators.

Research on the Use of Fly Ash for Underground Mine Supports

547

a) Calculation diagram of the concrete insert plate

h

h0 a

h

x 2

x

x 2

Ru

b

b) Design of reinforcement concrete structure concrete insert plate

Fig. 14. Reinforcement concrete insert plate structure with rectangular cross-section

a) The finished concrete insert plate

b) Using concrete insert plates as a rock/soil support in Mao Khe Coal Company [22]

Fig. 15. Using concrete insert plates as rock/soil support in Mao Khe Coal Company [17]

Vietnam started the development of coal-fired power plants 20 years ago. Now, there are 16 coal-fired power plants in Vietnam [9]. They produce about 8.1 million tonnes of fly ash [9] (see Table 8). According to statistics, the fly ash production rate is far outweighed consumption due to increased amounts of energy being generated by coal-fired power plants and widely available across the globe. Fly ash contains silica, alumina, ferric oxide, and other oxides material that might turn fly ash to hazardous material. This hazardous material is a contributing factor in air, water, and soil pollution that lead to human health problems and various geo-environmental issue. These adverse situations will interrupt the entire ecological cycles if not correctly disposed. Therefore, good waste management practices are needed to sustain a healthy environment [10]. Fly ash emissions from coal combustion units show a wide range of components with the present of elements below atomic number 92 and are considered as the major source of

548

T. H. Vo et al.

air pollution. The ultrafine particle of fly ash will behave like cumulative poisons after remaining for long periods when it reaches the respiratory region [11, 12]. As a result, several physiological disorders and other related health problems such as respiratory problems, cancer, anaemia, hepatic disorder, gastroenteritis, and dermatitis will arise [10, 11]. Several studies on the present ground showed that wet disposal of this waste causes the migration of metal into the soil [13]. The populations located near the fly ash dumping area facing surface water pollution and underground water pollution. However, surface water pollution is more critical than underground water pollution. The surface water pollution decreases the fish population and other aquatic organisms due to heavy metal material and organic matter content contained in the water. Surface water contamination also causes skin diseases diarrhea and death due to bathing and drinking water from the contamination river [10]. Waste treatment is a vital issue of thermal power plants and the society of Vietnam. It is a huge problem. If treated and used thoroughly, it will solve many goals such as Saving land area used as a waste dump, minimizing environmental pollution, creating more revenue from treated products, increasing investment efficiency of projects, including pit mining, underground mining projects. In addition, it contributes to saving foreign currencies, reducing the burden of trade deficit for the economy, because recycling will create a source of raw materials to replace the raw materials that are still imported to add additives for cement production, making concrete, treating the soil. At the same time, to form a market for buying and selling waste that has been handled for use as raw materials for construction materials.

6 Conclusion This study aims at proposing an optimized proportion of the fly ash of thermal power plants used in the concrete mixture for underground mine supports in Quang Ninh coal area. Several important results can be summarized as follows: • The increase in fly ash reduces the workability of concrete. The particles of fly ash reduces the amount of water required to produce a given slump. The circular shape of the fly ash particles and its dispersion ability provide water-reducing characteristics. • The compressive strength and flexural strength increase with the increase of fly ash in concrete up to 30% cement replacement in the conventional mix. • Mixing of fly ash in conventional concrete mix has resulted in considerable variation in the properties of fresh concrete. The integration of fly ash in concrete increases the cohesiveness of the mix, prohibited segregation, and resulted in reduced bleeding. • Higher percentages of fly ash can cause a change in the color of the mix. Incorporation of fly ash in concrete can save the coal and thermal industry disposal costs and produce a “greener” concrete for underground construction. • The research can be conducted further on higher grades of concrete or integration of waste materials to improve the strength. • This study shows that the cement partially replaced by fly ash is 37.78 kg lighter than normal grade concrete and saves 118.17 kg of cement in 1 m3 of concrete for making concrete insert plate of SVP steel frame in underground mines.

Research on the Use of Fly Ash for Underground Mine Supports

549

• Using fly ash for concrete insert plate will have positive effects on environmental protection in Vietnam.

Acknowledgment. This research was funded by the Nam Mau coal joint-stock company and the construction laboratory of Hanoi University of Mining and Geology. The funding support is much appreciated.

Conflict of Interest. The authors declare that there is no conflict of interest.

References 1. Dang Van, K., Vo Trong, H., Do Ngoc, A., Pham Tuan, A.: Research and manufacture reinforced concrete furnace insert plates from ash slag of thermal power plants in underground mines in Vietnam. National Mining Science & Technology Workshop 2019. Industry and Trade Publishing House. August 2019. ISBN. 978-604-931-849-8 2. Van Dang, K., Van Truong, H., Pham, A.T.: Experimental study on use fly ash in underground construction concrete. J. Min. Earth Sci. 60(3), 60–67 (2020) 3. Kosalram, R., et al.: Mix Design of Fly Ash based concrete using DOE Method 4. TCVN 10302:2014. Activity admixture - Fly ash for concrete, mortar and cement 5. Thomas, M.: Optimizing the Use of Fly Ash in Concrete Microstructural approach in damage. In: The 2nd International Conference on Dynamics and Control. 23–26 January 2007, Norfolk. Janpan, pp. 377–387 (2007) 6. TCVN 3118:1993. Heavyweight concrete - Method for determination of compressive strength 7. Gopalakrishna, S., Rajamane, N.P., Neelamegam, M., Peter, J.A., Dattatreya, J.K.: Effect of partial replacement of cement with fly ash on the strength and durability of HPC. Indian Concr. J. 75(5), 335–341 (2001) 8. Fly ash for cement concrete. Resource for High Strength and Durability of Structures at Lower Cost. Ash Utilization Division. NTPC Limited 9. Quang Chieu, N.: Fly ash: the origin of use and the environment. J. Transp. (7) (2011) 10. Nordin, N., Abdullah, M.M.A.B., Tahir, M.F.M., Sandu, A.V., Hussin, K.: Utilization of fly ash waste as construction material. Int. J. Conserv. Sci. 7(161–166), 2016 (2016) 11. Dwivedi, A., Jain, M.K.: Fly ash - waste management and overview: a review. Recent Res. Sci. Technol. 6, 30–35 (2014) 12. Tiwari, M.K.: Fly ash utilization a brief review in indian context fly ash utilization: a brief review in indian context int. Res. J. Eng. Technol. 3, 949–956 (2016) 13. Senapati, M.R.: Fly ash from thermal power plants-waste management an overview. Curr. Sci. 100, 1791–1794 (2011) 14. Tipraj, B., et al.: Strength characteristics of concrete with partial replacement of cement by fly ash and activated fly ash. Int. J. Recent Technol. Eng. (IJRTE) ISSN: 2277-3878, 8(4), November 2019 15. Patil, J.V.: Partial replacement of cement by fly ash in concrete mix design. Int. Res. J. Eng. Technol. (IRJET), 04(11), November 2017. e-ISSN: 2395-0056 16. Li, P., Wang, F., Fan, L., Wang, H., Ma, G.: Analytical scrutiny of loosening pressure on deep twin-tunnels in rock formations. Tunn. Undergr. Space Technol. 83, 373–380 (2019) 17. Pham, T.A., et al. Research and manufacture on reinforced concrete insert plates from ash slag of thermal power plants. Report of Vinacomin at the basic level project. Ha Noi 7 (2020)

The Potential Use of Waste Rock from Coal Mining for the Application as Recycled Aggregate in Concrete Nguyen Cong Thang1 , Nguyen Van Tuan1(B) , Dao Ngoc Hiep2 , and Vu Manh Thang2 1 Faculty of Building Material, National University of Civil Engineering, Hanoi, Vietnam

[email protected] 2 Vinacomin Industry Investment Consulting Joint Stock Company, Hanoi, Vietnam

Abstract. Along with the development of the coal mining industry, a massive amount of waste rock has been exhausted. It does waste not only plenty of natural resources but also pollutes the environment, as well as occupying a large part of the land use for dumpsites. This paper focuses on the possibility of recycling waste rock in coal mining as an aggregate in producing concrete. Accordingly, the waste rock was crushed to produce fine and coarse aggregates, and then they were used to produce concrete. Besides, the natural aggregates were also taken into account to compare and evaluate the quality of the recycled concrete. The experimental results showed that aggregate samples with lithological composition and indicators of the alkali-silica reaction were completely compatible with natural aggregates. They met requirements in producing mortar and concrete. Another finding in this study is the waste rock with compressive strength of over 100 MPa was suitable for coarse aggregate for producing concrete. Finally, when using recycled aggregate from waste rocks of coal mining, it is possible to make concrete with compressive strength of over 30 MPa. Besides, the flexural strength, bond strength between steel bar and concrete, modulus of elasticity, and shrinkage deformation of concrete at long-term ages are not too dissimilar to those of concrete samples using natural aggregates. Keywords: Natural resources · Coal · Waste rock · Aggregate · Recycled concrete

1 Introduction The production and recycling activities of wastes in producing construction materials are the significant concerns of engineers and researchers. The recycling of waste rocks in coal mining for construction materials has been studied. It contributed to reducing environmental pollution, creating technical value, and economic benefits from waste rock [1–3]. The waste rock at coal mines is a technological waste formed during coal production. It has been wholly or partly removed from the properties of coal used [3]. The process of accumulating waste rock of a mine can lead to environmental pollution. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 550–561, 2021. https://doi.org/10.1007/978-3-030-60839-2_29

The Potential Use of Waste Rock from Coal Mining

551

On the other hand, it is the use of natural resources resulting from inappropriate, limited reserves, and increasing depletion. Therefore, a study on the use of waste rock as potential raw materials for the construction industry is necessary [4, 5]. A review of the literature shows that waste rock was widely used in the production of construction materials [4, 6]. The main chemical compositions of waste rocks include SiO2 , Al2 O3 , Fe2 O3 , and a certain amount of loss on ignition [6]. The landfills formed after coal mining contain a large amount of valuable raw materials for the construction sector, such as stone, clay, sand, and others. Recycling waste rock in mining is not only reducing significant production costs and generating new materials with useful properties but also reducing environmental pollutions significantly [7, 8]. This issue is considered as a big concern in the worldwide, and in Vietnam as well. In addition, the use of natural resources in the production of construction materials is gradually exhausted. In recent years, some areas around the world have become scarce, and artificial materials are taken into account as alternative materials. In some regions, over-exploitation caused environmental pollutions, severe erosion of many river sections, and affecting human life significantly [9]. In Vietnam, the demand for construction materials is increasing, especially in Quang Ninh province. From statistical results show that the need for construction sand is about 130 million m3 per year [10], in which the regular supply only accounts for 50% as desired. Meanwhile, a large number of quarry dumps, for example, Quang Ninh province, has been discharging a large amount of waste rock with lithological components such as sandstone, siltstone, sandstone, etc. It is worth mentioning that these waste rocks are assessed as suitable for the production of construction materials. However, the ratio of processed waste rock in mining for construction materials is low. This reason is due to inadequate research on both technology and economics of the problem. Therefore, research and development of construction materials from the waste rock are necessary, aiming to meet the increasing demand for construction materials. That also will help to save mineral resources, limit the situation of illegal aggregate exploitation, ensure the local supply of aggregate for concrete, and thus, lead to save costs, improve socioeconomic and environmental efficiency. The paper focuses on the evaluation of potential use waste rock from coal mining in Quang Ninh province, Vietnam (Fig. 1), for the application as recycled aggregate in concrete. The study was based on experiments carried out on the concrete with a designed compressive strength of 30 MPa using waste rock incomparable with one using natural aggregate.

2 Materials and Methods 2.1 Materials The material used in this study is Portland cement blended PCB40 with the properties given in Table 1. This cement met the requirements of Vietnamese standard TCVN 6260:2009. The natural sand from Lo River in Vietnam was used as the fine aggregate, and its properties are given in Table 2 following the Vietnamese standard TCVN 7570: 2006. In the study, polycarboxylate-based superplasticizer (ROADCON SSA2000, Silkroad)

552

N. C. Thang et al.

was used with some properties such as pH of 6.5, the density of 1.05 g/cm3 , and its dosage of 0.5–1.2 L for 100 kg cement. Table 1. Physical properties of Portland cement. Properties

Unit

Value

Specification

Test methods (Vietnamese standard)

Retained on 0.09 mm sieve

%

3.3

≤10

TCVN 4030-2003

Fineness (Blaine)

cm2 /g

3150

≥2800

Standard consistency

%

29.0

–

TCVN 6017-2015

Compressive strength −3 days −28 days

TCVN 6016-2012 MPa

≥21.0 ≥40.0

24.1 45.5

Table 2. Physical properties of sand. No

Properties

Unit

Value

Specific gravity

g/cm3

2.65

2

Bulk density

kg/m3

1450

3

Porosity

%

45.5

4

Clay lumps content

%

0.08

5

Organic impurities content

–

pass

No

Cumulative retaining, by weight %

Fineness modules

1

1

Note

0.14 (mm)

0.315 (mm)

0.63 (mm)

1.25 (mm)

2.5 (mm)

Sieve size > 5 mm

96.6

71.2

47.8

28.8

11.5

1.8

2.56

90–100

65–90

35–70

15–45

0–20

–

Specific requirements

Waste rock in coal mining used in the study was taken at Dong Trieu region - Uong Bi, Quang Ninh province, Vietnam. After being taken at the quarry, the waste rock was taken to the laboratory to be homogenized and crushed to the required particle size, and then analyzed the mineral composition, to determine the compressive strength of the original rock. The detailed information is given in Sect. 3.1. 2.2 Concrete Mix Proportioning In this study, the designed compressive strength of concrete was 30 MPa. The waste rock was used as the coarse aggregate incomparable with the natural total with the properties as given in Table 6. The concrete mix proportioning is given in Table 3. Some properties of fresh mixtures and hardened concrete were determined, i.e. the workability of concrete mixtures, compressive strength at different ages, the modulus of elasticity and shrinkage.

The Potential Use of Waste Rock from Coal Mining

(a) Waste rock storage

553

(b) Coarse aggregate after crushing

Fig. 1. The waste rock (a) waste rock sample at the storage area, (b) the coarse aggregate after crushing

Table 3. Concrete mix proportioning used in this study Mixture designation

Concrete ingredients, kg/m3 Cement

Sand

Coarse aggregate

Water

Superplasticizer

Waste rocks (WR)

370

880

980

175

2.22

Natural aggregates (NA)

370

880

980

175

2.22

2.3 Experimental Methods Test methods for determination of petrographic compositions were adopted according to Vietnamese standard TCVN 7572–3: 2006. The potential alkali reactivity of waste rock was tested at the age of 6 months by mortar-bar with the size of 25 × 25 × 285 mm based on the standard ASTM C1260. The compressive strength of original rock was tested using 50 × 50 × 50 mm cubic samples and following the standard TCVN 7572-10:2006. The compressive strength of rock was determined with both dry and water-saturated condition. The mechanical properties of waste rock in coal mining used as coarse aggregate such as particle size, specific gravity, bulk density were determined by the standard TCVN 7572: 2006. The compression test of concrete was carried out at the ages of 3 days, 7 days and 28 days using 150 × 150 × 150 mm cube samples and following the standard TCVN 3118-1995. The flexural strength of concrete was tested using a prismatic sample with dimensions of 150 × 150 × 600 mm and following the standard TCVN 3119:1993. The four-point bending tests were made with the span length (L) of 4500 mm between two lower support points, the bending test diagram is shown in Fig. 2. Elastic modulus of concrete was tested using the cylinder sample of 150 × 300 mm and following the standard ASTM C469.

554

N. C. Thang et al.

P

P

L/3

L/3

L/3

L = 4500

Fig. 2. The four-point bending tests

Shrinkage of concrete was tested for up to 91 days using samples with the size of 100 × 100 × 285 mm and following the standard ASTM C157. The bond strength between steel bar and concrete was determined with using the same type of reinforcement CB300-V (according to the standard TCVN 1651-2: 2008) with a diameter of 14 for the test samples. The reinforcing bars were embedded in concrete and placed through the centre, parallel to the edges of the 150 × 150 × 150 mm concrete sample. The contact length between steel bar and concrete is taken equal to 3d (42 mm). In the non-contact positions, 2 cold-welded plastic pipes were used with outer diameter 21 and inner diameter 14 to cover the steel bar to ensure no contact between the reinforcement and concrete. The samples were tested at the age of 28 days using three pieces for each mixture, which were placed in the test system shown in Fig. 3b. The bond strength between steel bar and concrete was obtained through the Eq. (1) when the steel bar was pulled out from the sample. The values of τmax and δmax were determined through the bond stress τ relative - displacement δ relationship.

5 1

2

3

6

4

(a)

(b)

Fig. 3. Pull-out test set-up with (a) making test samples and (b) test system diagram (1) Test sample (2) Frame (3) Hydraulic jack (4) Load cell (5) LVDT (6) Datalogger TDS530

τ=

P π.1.Φ

(1)

The Potential Use of Waste Rock from Coal Mining

555

where τ: Bond strength between steel bar and concrete, MPa P: Applied load, N F: Diameter of steel bar, mm l: Anchorage length of steel bar, mm

3 Experimental Results and Discussion 3.1 Properties of Waste Rock from Mines Waste rock from mines was washed, and crushed up to the required particle size, and then determining the physical properties of specific stones. Regarding mineral composition, the sample belongs to arkose sandstone type with mineral composition in the sample, including particles with medium roundness, in which quartz mineral occupies the main content and is distributed evenly throughout the sample. The other minerals can be accounted for but with a small container such as felspar, mica, and quartzite. Besides, the binding material accounts for 10% of the mineral content of the sample. The mineral composition of mine waste rock is given in Table 4. Table 4. Mineral composition of mine waste rock. Sample Waste rock

Content of minerals in waste rock, % Quartz

Felspat

Mica

Quarzit

Felzit

Sericit, cholorit

Silic

70–71

4–5

(trace)

8–10

5–7

6–7

3–4

Regarding alkali-silica reaction (ASR), this reaction occurs when alkali and OH− ions in hardened cement paste react with active SiO2 content in the aggregate to create a volumetric expansion that damages the structure of the hardened cement paste. ASR occurs when (1) the alkaline content of the cement or the raw materials is sufficiently high; (2) the activated silica source from aggregate or other sources is sufficiently high and (3) requires an aqueous environment for ASR to occur. Typically, for ordinary concrete, when the amount of cement and active mineral admixtures is relatively low, ASR occurs on activated aggregate with large particle size to produce a quantity of large enough gel product to cause swelling and damage to the structure of the concrete. Therefore, it is necessary to study the ASR in the concrete in case of using the crushed aggregate from the mine waste rock. Experimental results to evaluate the ASR potential of ground mine waste rock by the mortar bar method to check whether it belongs to the hazardous area or not, are shown in Table 5. The results showed that the deformation of the mortar bar increased over time, but the level of increase was not significant. The deformation of the mortar bar at the age of 1 month reached 0.019%, the deformation at the age of 6 months continued to increase and reached 0.04%, less than 0.1% as required, and thus, the waste rock sample during coal mining was assessed as not likely to be effected by ASR.

556

N. C. Thang et al. Table 5. Deformation of mortar bar due to ASR

Sample

Deformation of mortar bar, %

Waste rock

1 day

1 month

2 months

3 months

4 months

5 months

6 months

0

0.019

0.024

0.026

0.033

0.038

0.040

In addition to the indicators of mineral composition, and assessment of ASR, some physical properties of crushed mine waste rock (WR) were also determined (Table 6). Experimental results showed that the compressive strength of the original rock sample in both dry and saturated state was over 100 MPa, its softening coefficient reached 0.74. Besides, the content of chloride ion in aggregate was 0.005%, less than 0.1%, so it is entirely suitable for use in producing concrete. Table 6. Physical properties of crushed waste rocks (WR) and natural aggregate (NA) in coal mining. No

Properties

Unit %

Value WR

NA

0.35

0.18

Lighter than the standard color

Lighter than the standard color

1

Dust, mud content

2

Organic impurities

3

Aggregate Crushing Value (ACV)

%

7.18

8.2

4

Compressive strength under dry state

MPa

144.2

125.0

5

Compressive strength under saturated state

MPa

106.8

108.5

6

Softening coefficient of the original rock

–

0.74

0.86

7

Chloride content

%

0.005

0.004

8

Particle size distribution

No

Cumulative retaining, by weight %

Note

5 (mm)

10 (mm)

20 (mm)

40 (mm)

WR

96.7

52.1

7.3

0.0

NA

97.2

53.6

6.8

0.0

90–100

40–70

0–10

0

70 (mm)

–

Specific requirements

The Potential Use of Waste Rock from Coal Mining

557

3.2 Properties of Concrete Mixtures and Hardened Concrete Using Mine Waste Rock

10 WR

Slump, mm

8

NA

6 4 2 0 0

30

60

Time, minutes

Fig. 4. Workability of the concrete mixture with time

Compressive strength, MPa

Workability and Compressive Strength. The workability and compressive strength of concrete using WR and NA are shown in Figs. 4 and 5. Experimental results showed that the workability of the concrete mixture using WR and NA was not significantly different. Besides, when assessing the effect of coarse aggregate on the slump loss of concrete mixture with time showed that the workability of the concrete mixture decreased with time for both the mixtures using natural stone and mine waste rock, but this decline was not different between the two types of mixtures. 40 30 20 WR

10

NA 0 3

7

28 Age, days

Fig. 5. Compressive strength of concrete with time

The effect of the aggregate type on compressive strength of concrete is shown in Fig. 5. It can be seen that the compressive strength of concrete increased with time. The 28-day compressive strength of concrete was attained over 30 MPa, and the 7-day compressive strength reached 75% compared with the 28-day compressive strength. Test results indicate that there is no difference between the two types of aggregate when used in concrete. Flexural Strength of Concrete Flexural test results of concrete using NA and WR from waste rock are shown in Figs. 6 and 7. The experimental results showed that the sample using NA had about 2.6% higher flexural strength than the sample using WR. Therefore, it can be concluded that the flexural strength of concrete using NA and WR was not significantly different. Elastic Modulus of Concrete The effect of aggregate on elastic modulus was also studied using the testing equipment shown in Fig. 8 to evaluate the possibility of using aggregate from coal mining waste in producing concrete. The experimental results of the influence of aggregates on elastic modulus (E) at the ages of 7 and 28 days are given in Table 7 and shown in Fig. 9.

558

N. C. Thang et al.

Flexural strength, MPa

9

8.48

8.26

NA

WR

8 7 6 5 Coarse aggregate

Fig. 6. Flexural test of concrete

Fig. 7. Flexural strength of concrete using NA and WR

Table 7. Elastic modulus of concrete using WR and NA Sample

WR

Stress, MPa

0 1.86

NA 4.81

6.88

Strain, × 10−3 mm/mm 0 0.117 0.225 0.317 E-, GPa

24.3

9.73

0 1.8

5.19

7.67

0.425 0 0.133 0.283 0.383

9.82 0.467

25.2

The relationship of stress-strain at the initial stage, which was considered up to 40% of ultimate stress, shown in Fig. 9 was linear as indicated in the concrete theory. Experimental results showed that the elastic modulus of concrete using NA and WR was not significantly different. The results of elastic modulus with samples using NA were about 4% higher than those using WR, within the deviation during the experiment. The Bond Strength Between Steel Bar and Concrete Test results on the bond strength between steel bar and concrete are shown in Fig. 10. It can be observed that the bond force between steel bar and concrete can be expressed through the following stages: (1) the adhesion force is mainly the adhesive force of the hardened cement paste and no cracks can be observed in the concrete at this stage, the steel bar has not slipped (δ = 0); (2) crack appears at the bond location between steel bar and concrete, the applied load P of a steel bar on the concrete increases, the displacement δ increases; (3) the cracks continue to develop, bond force reaches the maximum value, the displacement of the steel bar increases rapidly; (4) the amount of the applied load on steel bar is not increased for a short time, then quickly decreased, the steel bar rushed until the concrete-steel bar connection is damaged. The results showed that, for both natural and crushed aggregate samples from the mine waste rock, at the beginning when the bond strength reached 7 MPa, the steel bar’s displacement in the concrete sample remained zero. However, when the load continues

The Potential Use of Waste Rock from Coal Mining

559

Fig. 8. Concrete samples for elastic modulus test

12 NA

WR

Linear (NA)

Linear (WR)

Stress, MPa

10 y = 25269x - 1.0192 R² = 0.9989

8

y = 24015x - 1.4789 R² = 0.9991

NA

6 4

WR

2 0 0.00E+00

1.00E-04

2.00E-04

3.00E-04

4.00E-04

5.00E-04

Strain Fig. 9. Stress-strain relationship of concrete using WR and NA

to increase (bond force increases), the displacement of the steel bar in the concrete sample starts to increase, the maximum bond force reached 17 MPa corresponding to displacement δ of 280 μm. At this stage, the load continues, the bonding force does not increase, but the displacement starts to increase greatly, continues to maintain the load begins to decrease, and the displacement increases strongly, at this time the steel bar is pulled-out from the concrete. For concrete samples using NA and WR, the bond strength is reduced sharply at the displacement of the steel bar reached 600 μm. The experimental

N. C. Thang et al.

Bond strength (MPa)

560

20.0 18.0 16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0 0.0

NS W R NA 0

WR

100 200 300 400 500 600 700 800 Displacement of steel (µm)

Fig. 10. The relationship of bond strength τ and displacement δ of steel bar

results demonstrate that there is no significant difference in the bond strength of steel bar and concrete when using coarse aggregate from WR compared with NA. Shrinkage Deformation of Concrete Effect of aggregate types on shrinkage deformation of concrete is shown in Fig. 11. The experimental results show that the shrinkage deformation of concrete at 28 days reaches about 360 μm/m. It should be noted that at 91 days of shrinkage deformation of concrete reaches 498 μm/m and 535 μm/m with samples using aggregates from mine waste rock and natural aggregates, respectively. It means that there was no significant difference in shrinkage deformation between concrete using NA and concrete using WR. Age, days 0

7

14

21

28

35

42

49

56

63

70

77

84

Shrinkage, µm/m

0 -100

WR

-200 -300 -400 -500 -600

Fig. 11. Shrinkage deformation of concrete using WR and NA

NA

91

The Potential Use of Waste Rock from Coal Mining

561

4 Conclusions Based on the experimental results of the study on the possibility of recycling waste rock during coal mining as aggregate in concrete production, some conclusions can be drawn as follows: • It is possible to use waste rock during coal mining in Dong Trieu and Quang Ninh areas as aggregates in concrete production. Waste rock samples had a high compressive strength of over 100 MPa under both dry and saturated states. • The aggregate sample had a shallow chloride-ion content of 0.005%, smaller than required for aggregates using in mortar and concrete. The shrinkage deformation of the mortar at 6 months reached 0.04%, less than 0.01% as required. Thus, waste rock samples are evaluated as unlikely to be affected by alkali-silica reaction. • It is possible to use recycled aggregate from coal mining waste to produce concrete with a compressive strength of over 30 MPa. Elastic modulus of concrete of over 24 GPa was attained. There is no significant difference between the physical and mechanical properties, and bond strength of steel bar and concrete between concrete using WR and concrete using NA.

Acknowledgements. The authors would like to thank Vinacomin Industry Investment Consulting Join Stock Company and the National University of Civil Engineering for their support of materials and equipment during the experiment.

References 1. Stolboushkin, A.Y., et al.: Use of overburden rocks from open-pit coal mines and waste coals of Western Siberia for ceramic brick production with a defect-free structure. In: IOP Conf. Series: Earth and Environmental Science, vol. 84, pp. 012045 (2017) 2. Stolboushkin, A.Y., Ivanova, A.I., Fomina, O.A.: Use of coal-mining and processing wastes in production of bricks and fuel for their burning. Sci. Direct, Procedia Eng. 150, 1496–1502 (2016) 3. Shchadov, V.M.: Processing of coals in Russia in the 21st century. Coal 8, 28–31 (2007) 4. Skarzynska, K.M.: Reuse of coal mining wastes in civil engineering: utilization of minestone. Waste Manage. 2, 83–126 (1995) 5. Sandecki, M.: Aggregate mining in river systems. Calif. Geol. 42(4), 88–94 (1989) 6. Santos, C.R., Amaral, J.R.: Use of coal waste as fine aggregates in concrete paving blocks. Geomaterials 3, 54–59 (2013) 7. Haibin, L., Zhenling, L.: Recycling utilization patterns of coal mining waste in China. Resour. Conserv. Recycl. 54(12), 1331–1340 (2010) 8. Silva, R.D.R., Rubio, J.: Treatment of acid mine drainage (AMD) from coal mines in South Brazil. Int. J. Coal Prep. Utilization 29(4), 192–202 (2009) 9. Milioli, G.: Mining, environment, and development in South Santa Catarina, Brazil. nongovernmental organization ‘Terra Verde’ and its ideas for sustainability. Environments 33(1), 25–40 (2005) 10. Bac, P.V.: Situation of management, exploitation and use of sand and industrial waste (sandstone, thermal ash) to produce construction materials in Quang Ninh province. In: Proceedings of Crushed Sand Replaces Quang Ninh Natural Sand, Environmentally Friendly Materials: Construction Publishing House, pp. 14–20 (2018) (in Vietnamese). ISBN 978-604-82-2426-4

Author Index

A Ali, Sarfraz, 344 Armaghani, Danial Jahed, 13, 143 Arnoldovich, Belin Vladimir, 170 Ashraf, Hamid, 344, 498 Atif, Iqra, 498 Atrushkevich, Victor, 170 B Baic, Ireneusz, 407 Bhatawdekar, Ramesh Murlidhar, 13, 143 Binti Abang Hasbollah, Dayang Zulaika, 143 Blaschke, Wiesław, 407 Bruland, Amund, 81 Bui, Hoang-Bac, 109 Bui, Luyen K., 1 Bui, Manh-Tung, 320 Bui, Xuan-Nam, 1, 45, 109, 170, 187, 203, 224, 364 C Cao, Xuan-Cuong, 203, 224 Cawood, Frederick Thomas, 498 Cawood, Frederick, 344 Chinh, Vu Thi, 436 D Dam, Trong Thang, 45 Dang, Van Kien, 531 Dao, Van-Chi, 320 Dias, Daniel, 81 Dinh, Van Diep, 81 Do, Ngoc Anh, 81, 531 Do, Ngoc-Tuoc, 224 Dong, Pham Sy, 69

Drebenstedt, Carsten, 170, 203 Dung, Nhu Thi Kim, 436 F Feroze, Tariq, 344 G Golik, Vladimir Ivanovich, 320 Goyal, Ropesh, 1 H Hai, Duong Duc, 283 Hai, Luong Nhu, 69 Hao, Pham Manh, 69 Haubold-Rosar, Michael, 485 Heo, Won-Ho, 187, 224 Hiep, Dao Ngoc, 550 Hildmann, Christian, 485 Huong, Tong Thi Thanh, 364 J Jo, Youngdo, 245 K Keller, Florian, 364 Kim, Jung-hun, 187 Kim, Minsik, 245 Kim, Nhung Ta Thi, 469 Knoche, Dirk, 485 Koteras, Aleksandra, 283 L Le Thi Thanh, Xuan, 469 Le Thi, Huong, 469 Le, Huong Thi, 453

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 X.-N. Bui et al. (Eds.): ISRM 2020 - Volume 1, LNCE 109, pp. 563–564, 2021. https://doi.org/10.1007/978-3-030-60839-2

564 Le, Qui-Thao, 203, 224 Le, Tien Dung, 308 Le, Xuan Thi Thanh, 453 Lee, Chang-Woo, 187, 224, 263 Lee, Dongkil, 245 Lee, Roh Pin, 364 Lien, Huong Nguyen Thi, 469 Long, Nguyen Quoc, 1 Luong, Anh Mai, 453 M Mahboob, Muhammad Ahsan, 498 Mai, Anh Luong, 469 Mai, Huong Phan Thi, 469 Mamani Soliz, Patricio E., 364 Man, Nguyen Xuan, 69 Meyer, Bernd, 364 Mijał, Waldemar, 407 Moayedi, Hossein, 91 N Ngoc, Anh Nguyen, 469 Nguyen, Anh Ngoc, 453 Nguyen, Doanh Quoc, 453 Nguyen, Hoang, 109, 170, 187, 203, 224 Nguyen, Huong Thi Lien, 453 Nguyen, Huyen Thi Thu, 453 Nguyen, Le Duc, 283 Nguyen, Ngoc Phu, 426 Nguyen, Ngoc-Bich, 203, 224 Nguyen, Phi-Hung, 320 Nguyen, Thao Thanh, 453 Nguyen, Tri Ta, 45 Nguyen, Van-Duc, 170, 187, 224, 263 Nguyen, Viet, 469 Nhu, Nguyen Tran, 469 P Park, Jongmyung, 245 Pathak, Pranjal, 13 Peuker, Urs A., 385 Pham, Quan Thi, 453 Pham, Thanh Hai, 385

Author Index Phan, Huong Thi Mai, 453 Phung, Tien Thuat, 426 Q Quoc, Doanh Nguyen, 469 R Rösel, Lydia, 485 S Scheithauer, Michaela, 364 Singh, Trilok Nath, 13, 143 Son, Nguyen Hoang, 436 T Ta, Nhung Thi Kim, 453 Thang, Nguyen Cong, 69, 550 Thang, Vu Manh, 550 Thanh, Thao Nguyen, 469 Thao, Nguyen Van, 69 Thi, Quan Pham, 469 Thu, Huyen Nguyen Thi, 469 Thuy, Ngo Ngoc, 69 To, Duc Tho, 45 Tonnizam Mohamad, Edy, 13, 143 Tran, Hai-Van Thi, 203 Tran, Nguyen Nhu, 453 Tran, Quang-Hieu, 170, 187, 224 Trung, Nguyen Duc, 283 Truong, Van Ha, 531 Tuan, Nguyen Van, 69, 550 Turek, Marian, 283 V Van Duoc, Tran, 436 Van Khuong, Duy, 453 Van, Duy Khuong, 469 Vo, Trong Hung, 531 Vu, Thai-Tien-Dung, 320 Y Yi, Huiuk, 245 Z Zimmermann, Beate, 485