Underground Ventilation: Proceedings of the 19th North American Mine Ventilation Symposium ((NAMVS 2023, 17-22 June 2023, Rapid City, South Dakota, USA) [1 ed.] 1032551461, 9781032551463

Citation preview

Underground Ventilation EDITED BY

Purushotham Tukkaraja

UNDERGROUND VENTILATION

Underground Ventilation contains the proceedings of the 19th North American Mine Ventilation Symposium held at the South Dakota School of Mines & Technology (South Dakota Mines) in Rapid City, South Dakota, June 17-22, 2023. South Dakota Mines organized this symposium in collaboration with the Underground Ventilation Committee (UVC) of the Society for Mining, Metallurgy & Exploration (SME). The Mine Ventilation Symposium series has always been a premier forum for ventilation experts, practitioners, educators, students, regulators, and suppliers from around the world to exchange knowledge, ideas, and opinions. Underground Ventilation features sixty-seven selected technical papers in a wide range of ventilation topics including: auxiliary and primary systems, mine fans, case studies, computational fluid dynamics applications, diesel particulate control, electric machinery, mine cooling and refrigeration, mine dust monitoring and control, mine fires and explosion prevention, mine gases, mine heat, mine ventilation and automation, occupational health and safety, renewable/ alternative energy, monitoring and measurement, network analysis and optimization, and planning and design.

PROCEEDINGS OF THE 19TH NORTH AMERICAN MINE VENTILATION SYMPOSIUM (NAMVS 2023), 17-22 JUNE 2023, RAPID CITY, SOUTH DAKOTA, USA

Underground Ventilation

Edited by Purushotham Tukkaraja, Ph.D., QP Mining Engineering & Management, South Dakota Mines, Rapid City, SD, USA

Front Cover Image: © Zitrón - Ventilation solutions for underground mines

First published 2023 by CRC Press/Balkema 4 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN and by CRC Press/Balkema 2385 NW Executive Center Drive, Suite 320, Boca Raton FL 33431 CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business © 2023 selection and editorial matter, Purushotham Tukkaraja; individual chapters, the contributors Typeset by Integra Software Services Pvt. Ltd., Pondicherry, India The right of Purushotham Tukkaraja to be identified as the author of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Although all care is taken to ensure integrity and the quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to the property or persons as a result of operation or use of this publication and/or the information contained herein. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record has been requested for this book

ISBN: 978-1-032-55146-3 (hbk) ISBN: 978-1-032-55147-0 (pbk) ISBN: 978-1-003-42924-1 (ebk) DOI: 10.1201/9781003429241

Underground Ventilation – Tukkaraja (Ed) © 2023 The Editor(s), ISBN 978-1-032-55146-3

Table of contents

Preface

xi

Organizing committees

xiii

Underground Ventilation Committee (UVC)

xiii

Review committee

xiii

Sponsors and Exhibitors

xv

Auxiliary ventilation An investigation of booster fan placements in a large opening underground stone mine utilizing CFD N. Gendrue, S. Liu & S. Bhattacharya Quantifying assemblage losses in auxiliary ventilation systems E. De Souza

3 15

Case studies of mine ventilation Kucing-Liar (KL) mine LOM preliminary ventilation design – PT Freeport Indonesia A.A. Habibi, I. Setiawan, R. Prasojo & C. Stewart

25

A case study in successful use of spot cooling for underground shaft sinking M. Brown, D.W. Durieux, C. McGuire & D. Witow

34

Computational fluid dynamics applications in mine ventilation CFD study of cavern ventilation M.A. Carvajal-Meza & J.P. Hurtado-Cruz

45

Model order reduction of high-fidelity underground mine model M. Vaze, J. Nyqvist & S. Dasgupta

54

Minimizing the aerodynamic impact of a new cooling plant installed upstream of an existing surface fan and heater arrangement through CFD analyses J.K. Shaw, L.K. Falk, C. Allen & M. Kaufman CFD modeling of a large-opening stone mine using COMSOL multiphysics K.V. Raj & V. Gangrade An investigation of shock loss factors at ventilation raise junctions in underground hard rock mines using computational fluid dynamics H. Zhang, L.K. Falk & C. Allen

v

60 69

83

Diesel particulate control Comparing diesel and GDiesel® exhaust exposures in an underground mining laboratory R.J. Reed, J.L. Burgess & E.A. Lutz Improvement of size-selective sampling of diesel aerosols in underground mines A.D. Bugarski, T.L. Barone, J.A. Hummer, T. Lee, S. Vanderslice & S. Friend Results of diesel exhaust nanoparticle experimental sampling in a cabin of LHD loader operating in an active ore heading area S. Sabanov, N. Magauiya, A. Zeinulla, A. Abil, A. Qureshi, M. Torkmahalleh, G. Nurshaiykova & D. Rakhimov Importance of using real-time and microscopic analysis techniques to characterize DPM in underground mines A.A. Habibi, K.O. Homan & A.D. Bugarski Development and evaluation of innovative diesel particulate filter technology J. Stachulak, B. Rubeli, D. Young, K. Watson & B. McLean DPM reduction through emission assisted maintenance - PT Freeport Indonesia program update A.A. Habibi, E. Pinto, M. Mardon, K. Wijayanto & C. Rose

95 104

115

120 129

135

Electric machinery in mine ventilation Underground mine ventilation design: Diesel vs. electrical equipment C.A. Rawlins

149

Mine cooling and refrigeration Selection, design challenges and construction of Vale’s Coleman mine 10 MW surface refrigeration plant S.G. Hardcastle, J.K. Shaw & C. Allen

163

Mine refrigeration using geothermal energy – is this a viable decarbonization strategy? D.W. Durieux

174

Onaping Depth Project underground refrigeration plant update D.W. Durieux, C. McGuire, T. Mehedi, D. Witow & E. Pilkington

183

A direct equation for sigma heat and wet-bulb temperature for underground ventilation applications A. Pandey, S. Jayaraman Sridharan & B.S. Sastry

192

Case study: Refrigeration requirements during mineshaft excavation as a function of heat stress index K. Tom

200

Development of energy efficient and sustainable cooling strategies for hot underground mines J.E. Fox

208

Mine dust monitoring and control Effect of auxiliary scrubbers on respirable coal mine dust particle size and composition F. Animah, A. Greth, C. Keles & E. Sarver Recovery of respirable dust from fibrous filters for particle analysis by scanning electron microscopy A. Greth, S. Afrouz, F. Animah, C. Keles & E. Sarver

vi

221

230

Effects of vertical air-blocking ring of drill shroud on dust control for surface mine drilling operation using CFD Y. Zheng, J.D. Potts & W.R. Reed

239

Investigation on the effect of water pressure on spray performance for removal of respirable dust H. Jiang, S. Klima, T. Beck & Y. Zheng

247

Real-time measurements of respirable crystalline silica, kaolinite, coal, and calcite W. Arnott, C. Kocsis, X. Wang, B. Osho, P. Nascimento, S. Taylor, B. Bingham, C. Murphy & M. Sandink

256

Comparing respirable dust characteristics from full-scale cutting tests of three rock samples with conical picks at three stages of wear S. Slouka, E. Sidrow, C. Tsai & J. Brune

264

Accuracy of low-cost particulate matter sensor in measuring coal mine dust- a wind tunnel evaluation M.M. Zaid, G. Xu & N.A. Amoah

274

Parametric studies to maximize the dust protection performance of the two-level manifold canopy air curtain and computational fluid dynamics modeling N.A. Amoah, G. Xu & A.R. Kumar

285

Respirable coal mine dust research: Characterization and toxicity analysis based on dust sources V.P. Salinas, M.C. Das, G. Rubasinghege, P. Roghanchi & K. Zychowski

296

Comparison of respirable coal and silica dust monitoring systems for underground mining applications A. Medina, A. Vanegas, E. Madureira, P. Roghanchi, R. Rajapaksha, L. Uecker, T. Rawson & C. Harb

305

Evaluation of different surfactants’ performance in varying coal dust concentration through logistic regression analysis Z. Zhao, A. Ghosh, P. Chang & Y. Liu

313

Process ventilation solutions for mitigation of combustible and non-combustible dust hazards at mining operations J. Finn

322

Development of VR-CFD-based training tool for dust control in gateroad development M. Qiao, T. Ren, J. Roberts, J. Hines, C. Chow & A. Clayton

330

Mine fans Stall impact on axial fans and testing of anti-stall rings J. Fernandez

341

Practical values for the evaluation of fan system efficiencies J. Bowling, G. Schult & J. Van Diest

353

Mine fires and explosion prevention Longwall maingate and tailgate proactive sponcom and gas management strategyAn operational safety share on risk management B. Belle & R. Balusu Real-time methane prediction with small dataset in underground longwall coal mining using AI D.C. Demirkan, H.S. Duzgun, A. Juganda, J. Brune & G. Bogin

vii

363 377

Evaluation of different suppression techniques for lithium-ion battery fires L. Yuan, W. Tang, R.A. Thomas & J. Soles

384

Characterization and preliminary assessment of diesel fire prior to setting up large size battery fire experiment R.I. Pushparaj, G. Xu, A. Iqbal & O.B. Salami

393

Fire-induced temperature attenuation under the influence of a single ceiling smoke extraction point in a bifurcated drift O.B. Salami, G. Xu, A.R. Kumar, R.I. Pushparaj & A. Iqbal

399

Application of pressure balancing techniques at the West Elk coal mine C. Kocsis, F. Calizaya, J. Johnson, T. Dias, N. Nunes, E. Lindgren, G. Atchley & J. Poulos

411

Mine gases Underground coal methane gas forecasting using multivariate time series with one and two auxiliary variables J.C. Diaz, Z. Agioutantis, S. Schafrik & D.T. Hristopulos Complete degasification of longwall panels in U.S. coal mines P.C. Thakur Study of in-situ coal seam gas content for Australian coal and gas outburst management: A field data analysis and laboratory experiment Z.B. Li, T. Ren, M. Qiao, D. Black & J. Juric

423 431

439

Airflow patterns and blast fume dispersion in different mining methods S. Jayaraman Sridharan, A. Adhikari, P. Tukkaraja & J. Connot

451

The oxiperator for Ventilation Air Methane (VAM) E. Prabhu, M. Prabhu & J.E. Fox

462

Mine heat Findings and learnings from thermal parameter studies at four LKAB sites F.K.R. Klose, A.L. Martikainen & T.H. Jones

469

Scenario-driven evaluation of heat sources in underground production scheduling J.A. Buaba, E. Udofia, A. Newman & A.J. Brickey

484

Mine ventilation and automation Permeability determination for potential interaction between shale gas wells and the coal mine environment due to longwall-induced deformations under deep cover M.L. Harris, S.J. Schatzel, K.M. Ajayi, M. Van Dyke, P. Zhang, V. Gangrade, J.D. Addis, H. Dougherty & E. Watkins Analysis of variation in longwall-induced permeability under different mining depths K.M. Ajayi, Z. Khademian, S.J. Schatzel, M.L. Harris & J.D. Addis New applications of jet fans in underground mines for haulage ramps and block cave ventilation control C. Stewart

viii

499

507

515

Occupational health and safety in mine ventilation Ventilation research findings for enhanced worker safety when mining near unconventional gas wells in longwall abutment pillars S.J. Schatzel, K.M. Ajayi, Z. Khademian, R. Kimutis, M.L. Harris, M. Van Dyke, J.D. Addis, H. Dougherty & E. Watkins Optimization for fire evacuation applying maximum flow cost algorithm to improve the time-response in underground coal mines S. Lotero, H. Khaniani, V. Androulakis, M. Hassanalian, S. Shao & P. Roghanchi

529

541

Heat mitigation for underground coal mine refuge alternatives D.S. Yantek & L. Yan

550

Occupancy derating for underground coal mine refuge alternatives D.S. Yantek & L. Yan

559

Renewable / alternative energy in mine ventilation Incorporating droplet dynamics to improve the reduced-order model of spray freezing for mine heating applications M. Mohit, S. Akhtar, M. Xu & A.P. Sasmito Analysis of small-scale lithium-ion batteries under thermal abuse A. Iqbal, G. Xu, R.I. Pushparaj & O.B. Salami

569 578

Ventilation monitoring and measurement Gauge and tube surveys: What is their future and that of underground measurements generally as mines transition towards greater use of Big Data and Artificial Intelligence systems? D.J. Brake Improving the accuracy of field airflow measurements for tunnel ventilation fans R.E. Ray & E. Fuster

589 595

Ventilation network analysis and optimization Case study on the abnormal airflow diagnosis method using atmospheric monitoring data L. Zhou, D. Bahrami & R.A. Thomas

607

Ventilation system upgrades at the Waste Isolation Pilot Plant K.G. Wallace & I. Peña

615

Ventilation network optimization: Realizing energy savings while promoting worker health and safety A.K. Ngcibi, M. Mochubele & F.S. Bergh

624

Ventilation planning and design Design highlights for Agnico Eagle’s Macassa 4 Shaft primary ventilation systems K. Boyd, T. Mehedi, D. Witow & M. Pinheiro-Harvey

635

Upper Keel mine ventilation strategy at Eagle mine K. Boyd, D. Witow, C. McGuire & C. Gobbs

642

Empirical and numerical investigation on the optimal length of eddy airflow in dead-end tunnel R. Morla, J. Chen, S. Karekal, A. Godbole, P. Tukkaraja & P. Chang

649

Author index

559

ix

Underground Ventilation – Tukkaraja (Ed) © 2023 The Editor(s), ISBN 978-1-032-55146-3

Preface

This volume contains the proceedings of the 19th North American Mine Ventilation Sympo­ sium held at the South Dakota School of Mines & Technology (South Dakota Mines) in Rapid City, South Dakota, June 17-22, 2023. South Dakota Mines organized this symposium in collaboration with the Underground Ventilation Committee (UVC) of the Society for Mining, Metallurgy & Exploration (SME). The North American Mine Ventilation Symposium series was initiated by the UVC in 1982. The UVC is a Joint Technical Committee of the Coal and Energy and the Mining and Exploration Divisions of the SME. The purpose of the UVC is to promote engineering interest and techno­ logical progress in the ventilation of mines, tunnels, and other subsurface openings. The UVC accomplishes its purpose by conducting technical sessions at SME-AIME meetings, sponsoring the North American Mine Ventilation Symposium with host universities and other organizations, and soliciting papers for publication in Mining, Metallurgy & Exploration Journal and SME con­ ference proceedings. The UVC offers an affiliation home for SME members, and others engaged in the practice of underground ventilation. In these ways, the UVC seeks to encourage research, education, publications, and technology transfer in the field of underground ventilation. The North American Mine Ventilation Symposium, held every two to three years since 1982, provides a forum for practitioners, educators, and researchers to exchange the latest informa­ tion on the ventilation of mines, tunnels, and other underground facilities. Past Symposia were held and organized by… 1982 – University of Alabama – Howard L. Hartman 1985 – University of Nevada, Reno – Pierre Mousset-Jones 1987 – Pennsylvania State University – Jan Mutmansky 1989 – University of California, Berkeley – Malcolm J. McPherson 1991 – West Virginia University – W. J. Wang 1993 – University of Utah – Ragula Bhaskar 1995 – University of Kentucky – Andrzej Wala 1999 – University of Missouri-Rolla – Jerry Tien 2002 – Queen’s University, Canada – Euler De Souza

xi

2004 – University of Alaska Fairbanks – Sukumar Bandopadhyay and Rajive Ganguli 2006 – Pennsylvania State University – Jan Mutmansky 2008 – University of Nevada, Reno – Pierre Mousset-Jones 2010 – Laurentian University – Stephen Hardcastle and Dale McKinnon 2012 – University of Utah – Felipe Calizaya 2015 – Virginia Tech University – Kray Luxbacher and Emily Sarver 2017 – Colorado School of Mines – Jurgen F. Brune 2019 – McGill University and University of British Columbia, Canada – Ali Madiseh, Agus Sasmito, Ferri Hassani, and Jozef Stachulak 2021 – South Dakota School of Mines & Technology – Purushotham Tukkaraja With the help of the organizing committee, a solid 3-day program was assembled, with tech­ nical papers, panel discussions, and keynote presentations organized in 21 sessions. A total of 82 abstracts and 67 final papers were received. Session ventilation themes include case studies, computational fluid dynamics applications, diesel particulate control, electric machinery, mine cooling and refrigeration, mine dust monitoring and control, primary and auxiliary systems, mine fans, mine fires and explosion prevention, mine gases, mine heat, ventilation management, automation, occupational health and safety, renewable/alternative energy, monitoring and measurement, network analysis and optimization, and planning and design. I would like to thank the UVC and Review Committee members for their help with peerreviewing papers, chairing technical sessions, and advice to make this symposium a success. Finally, I would like to thank the Center for Alumni Relations and Advancement (CARA) and the Office of Marketing and Communications at South Dakota Mines for assisting with the symposium registration, advertisement, and website services. Purushotham Tukkaraja, Ph.D., QP Symposium Chair

xii

Underground Ventilation – Tukkaraja (Ed) © 2023 The Editor(s), ISBN 978-1-032-55146-3

Organizing committees Underground Ventilation Committee (UVC) Arash Habibi, Freeport-McMoRan, USA Arun Rai, Compass Minerals, USA Bharath Belle, Anglo American, Australia Craig Stewart, Minware, Australia

David Brokering, Freeport-McMoRan, USA Jack Trackemas, CDC-NIOSH, USA John Bowling, SRK Consulting, USA Purushotham Tukkaraja, South Dakota Mines, USA

Review committee Adrianus Halim Agus Sasmito Aleksandar Bugarski Alex Hatt Ali Haghighat Alex Rawlins Andrea Brickey Ankit Jha Arash Habibi Ashish Kumar Bharat Belle Brian Prosser B S Sastry Calen Beaune Charles Kocsis Cheryl Allen Craig Stewart Daniel Stinnette Darryl Witow David Brokering D P Mishra Duncan Chalmers Emanuele Cauda Emily Sarver

Euler De Souza Felipe Calizaya Frank von Glehn Gerrit Goodman Guang Xu Heather Dougherty Hua Jiang Ian Loomis Jack Trackemas John Bowling Jon Fox Joseph Finn Jozef Stachulak Jurgen Brune Kayode Ajayi Keith Wallace Lihong Zhou Liming Yuan Marcia Harris Matthew Gray Moe Momayez Myriam Francoeur Pedram Roganchi Pierre Mousset-Jones

xiii

Pramod Thakur Ramakrishna Morla Rao Balusu Richard Ray Rick Brake Rohit Pandey Rudrajit Mitra Sekhar Bhattacharyya Shimin Liu Srivatsan Jayaraman Sridharan Stephen Hardcastle Steven Schafrik Vaibhav Raj Vasu Gangrade William Reed Yi Zheng Zach Agioutantis

Underground Ventilation – Tukkaraja (Ed) © 2023 The Editor(s), ISBN 978-1-032-55146-3

Sponsors Platinum Freeport-McMoRan G+ Plastics Hatch Howden Minetek SRK Consulting, Inc. Gold Accutron Instruments Turnstone – ABC VS & JENNMAR Silver BBE Consulting Maestro Digital Mine Mecanicad Spendrup Fan Co. Exhibitors Accutron Instruments BBE Consulting CDC Dust Control & Air Blow Fans CFT Compact Filter Technic G+Plastics Howden Hyperflo Maestro Digital Mine Mecanicad Minetek Pinssar Quick Supply Co Spendrup Fan Co SRK Consulting TLT-Turbo Turnstone – ABC VS & JENNMAR Zitron

xv

Auxiliary ventilation

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

An investigation of booster fan placements in a large opening underground stone mine utilizing CFD N. Gendrue, S. Liu & S. Bhattacharya Department of Energy and Mineral Engineering, G3 Center and Energy Institute, The Pennsylvania State University, University Park, PA, USA

ABSTRACT: The optimization and planning of the mine ventilation system is a key compo­ nent of mine operation given that ventilation related costs can range between 20% and 50% of the total operating cost of the mine. In large opening mines (LOMs) utilizing perimeter venti­ lation schemes minimal guidance is available for determining optimal booster fan (BF) place­ ments. A ventilation survey was conducted and published previously by the author which established a CFD model for a section of an underground room and pillar large opening lime­ stone mine. In this work the previously created CFD model was utilized to investigate a total of 15 BF positions with a focus on recirculation patterns, overall airflow within the BF entry, and airflow around face area. It was found that the maximum airflow around the face areas can be achieved with fan positioned on the same side of the entry as face area; with the max­ imum airflow through the BF entry being achieved when the fan is placed in the center of the entry on the upstream side of the pillar line. The recirculation percentages were similar in all cases reaching a maximum between 35%-40% of the total air movement. However, the highest recirculation percentages also facilitated, via air entrainment, the highest airflow magnitudes through the BF entry. The booster fan’s ability to stimulate airflow through adjacent entries was found to be reduced by approximately 30% for each adjacent entry. Therefore, the recom­ mendation was given to position the BF within 3 entries of the face to achieve adequate airflow.

1 INTRODUCTION Mine ventilation is an essential part of any underground mining system and accounts for between 20% and 50% of the total operating cost of the mine (Carter, 2018, Leonida, 2019, Babu et al., 2015). In large opening mines (LOMs, mines with cross-sections >~1000 ft2), par­ ticularly in stone mines utilizing perimeter ventilation schemes, the placement of booster fans is typically empirically determined by rule of thumb or trial and error until satisfactory condi­ tions are achieved. Many LOMs do not utilize typical auxiliary ventilation systems with an auxiliary fan and bag. Therefore, proper BF placement is imperative considering the lack of other ventilation controls. In place of typical auxiliary systems many LOMs utilize large diameter fans that produce high-volume with low-pressure; and commonly referred to as box fans or propellor fans. Com­ pared to jet fans these fans provide the increased airflow needed to maintain regulatory com­ pliance; however, the low-pressure tradeoff creates other concerns over the increased role of natural ventilation plays in the system due to the low system pressure, ventilation efficiency, recirculation, and short circuiting of mining sections. Specifically for large opening stone mines there is little quantitative guidance on how to combat these issues. This work will inves­ tigate the effect of BF placement within the entry and its influence on recirculation and face ventilation efficiencies. DOI: 10.1201/9781003429241-1

3

In Gendrue et al. (Gendrue et al., 2023), a ventilation survey was conducted and used to create a computational fluid dynamics (CFD) model of a section of a large opening stone mine. The simulations resulted in a discussion of the mines selected BF placement and its effectiveness as well as discussions on measurement locations for ventilation surveys. In this work, the same model will be utilized to simulate different BF placements and discussions on the BF effectiveness around the face area and airflow directions and quantities will be given. This work aims to provide a foundation for future CFD BF placement studies on perimeter ventilation systems and discuss important considerations when selecting a position for a BF. 2 BACKGROUND Previous mine ventilation studies related to face ventilation focus on auxiliary ventilation of drifts or dead-end headings with an emphasis on contaminant distributions and removal. Vari­ ables typically considered when evaluating face ventilation are fan selection, curtain or auxiliary tubing placement, airflow penetration depth, entrainment ratios, recirculation, and jet fan angles being common among the studies listed below. However, large opening stone mine venti­ lation schemes are unique and must be evaluated separately. Studies specifically in LOM envir­ onments are limited however, the same quantitative variables are used to evaluate the systems. Dunn et al. (Dunn et al., 1983) studied how BF inclination, fan elevation and ultimately air­ flow around open entries and the face area in large and medium sized airways. However, their results and discussions were derived from field data which inherently limits the analysis due to the smaller number of sampling locations when compared to numerical simulations. Nonethe­ less, their recommendations for optimizing face ventilation were to: (i) utilize a pushing system rather than a pulling system, however, no pulling systems were investigated, (ii) place fans to direct flow diagonally across the face, (iii) place fans one pillar inby the general flow to minimize recirculation. They also noted that larger capacity fans ventilate more effectively and inclination and elevations of the fans have little effect on face ventilation. Krog and Grau (Krog and Grau, 2006) discussed the use of vane axial or propellor fans in terms of their recirculation, overall airflow, and entrainment ability for use in LOMs. Ultimately giving recommendations that pro­ pellor fans be placed in the fresh air stream and work best for regional ventilation applications while vane axial fans be placed 1 entry behind the airstream and work best for dead end ventila­ tion i.e., headings. They also tested pushing and pulling systems for a new mine utilizing a textbook example of a split mine ventilation scheme concluding that a BF positioned just outby a portal is effective for mine airflow. Later, Grau and Krog (Grau III and Krog, 2009) discussed BF placement and mine design criteria such as long pillars in terms of ventilation effi­ ciency for LOMs. Recommendations were given to place the booster fans outby the last open crosscut. Furthermore, it was noted that fans positioned in the middle of the entry will increase recirculation and that “not all recirculation is detrimental.” These three studies, however, do not cover all the common ventilation schemes used today in LOMs. The study by Krog and Grau (Krog and Grau, 2006) studied only one adjacent entry when three to five entries are common in perimeter schemes. They also investigated a split mine ventilation scheme for a new mine. Grau and Krog (Grau and Krog, 2009) utilized a 3-entry heading which operated under a pseudo split ventilation scheme with fresh air on one side and return air on the other. The BF as directed perpendicular to the face opposite what is done in perimeter ventilation where the fan is directed parallel with the face. Furthermore, the study by Dunn et al. (Dunn et al., 1983) may initially seem like a perimeter ventilation scheme but all the tested fan positions were at a 90° angle from the general airflow direction which is more analo­ gous to a unit ventilation scheme. These studies also utilized smaller fans than are typically seen today with the largest studies being an 8 ft. in Krog and Grau, 2006. Therefore, these studies are fundamentally different not only in terms of fan technology but also in terms of their fundamen­ tal ventilation system design as defined by Krog et al. (Krog et al., 2004) as split, unit, and per­ imeter schemes. BF placement and ventilation system designs have been noted in many studies as the most or one of the most important factors in LOM ventilation systems, however, there are minimal 4

studies that discuss optimal BF placement (Thimons et al., 1985, Dunn et al., 1983, Grau and Krog, 2009, Grau III et al., 2006). Thus, BF placement must be reexamined with current fan capacities and in a perimeter ventilation context with the fan directing airflow parallel to the face area rather than at a 90° angle to it. 3 METHOD 3.1 Partner mine description and CFD model creation A full-scale geometry was developed based on the mine map from a limestone mine located in Pennsylvania, USA. The entry widths and heights average 12.2 m (40 ft.) and 9.1 m (30 ft.) respectively, with 24.4 m (80 ft.) pillar centers. The mine utilizes curtains, stoppings, and booster fans to deliver fresh air across the mine with a perimeter ventilation scheme, with curtains estab­ lished approximately 5-6 entries back from the mining face. Figure 1 shows the plan view for the mine with the curtains, stoppings, and booster fans highlighted in black. The general airflow pattern is also shown with the red and green arrows with the different colors on the mine map indicate the different levels of benching. The yellow highlighted section in Figure 1 indicates the section of the mine that was utilized for regional CFD modeling. This section of the mine was chosen for simulations because the shape is consistent with a typical in a perimeter ventilation scheme. Furthermore, no benching was done throughout the region allowing for a consistent mining height which is common around the face areas of perimeter schemed mine. These two factors allow the results to be broadened to other LOMs utilizing perimeter schemes. To summarize, the model creation that was reported in Gendrue et al. (Gendrue et al., 2023), the boundary conditions for the model were measured directly from the mine during a ventilation survey utilizing a velocity (mass) inlet condition with a pressure outlet. Curtain lines on the east and north side of the geometry as well as the mine boundary on the south and west were utilized as the edges of the geometry. Two booster fans are located within the model domain, one near the center of the geometry labeled inner BF (IBF) and the second near the exhaust portal labeled exhaust BF (EBF). The EBF was held consistent throughout all simulations with only the location of the IBF changing. The specific BF location that was utilized by the mine at the time of the survey can be seen in Figure 2A as location 1 in the original fan locations callout. Table 1 shows the measured airflow velocity from the mine survey as well as the BF settings and descriptions. The measurement locations and labeling system used in Table 1 can be seen in Figure 2B

Figure 1. Plan view of the partner limestone mine with curtains, stoppings, and booster fans highlighted in black, with the selected CFD simulation region highlighted in yellow.

5

Table 1. Measured airflow velocities and fan settings observed during the ventilation survey that were used for the boundary conditions of the CFD simulation, measurement locations and names are shown in Figure 2B.

3.2 Booster fan placement scenarios To evaluate multiple BF locations two scenarios were created totaling 15 BF locations as shown by the blue fan symbols in Figure 2A. Scenario one hereby called original fan location consisted of the original mine selected BF location, on the floor in the center of the entry the 6N and pillar line 7W in Figure 2B. Eight other fan positions within the same entry and pillar line were investigated in a 3x3 grid pattern around the mines original position for a total of 9 positions in the original fan location scenario. The 3x3 grid pattern was set up within the entry so there was 5 ft. of air space between the edge of the IBF and the pillars edge. Given the fans are 12 ft. diameters the center of the fan was aligned 11 ft. from the pillar and crosscut edge. The center of the entry and center of the pillar line was then used for the positions in the middle of the grid. All nine positions are marked as ‘original fan location’ can be seen in the callout of Figure 2A. This 3x3 grid pattern enables fan positions at the front, back, and center of the pillar line to be examined while also investigating positions nearer to each pillar edge and the center of the entry. The second scenario was created because two different Groups Dunn et. al., and Krog and Grau (Dunn et al, 1983, and Krog and Grau, 2006) came to simi­ lar conclusions about placing the fan in a forcing fashion one pillar inby the general airflow without providing any data in their works for pulling fans. Therefore, two new locations were investigated one that was three pillars upstream to examine a more forcing fan location and another location three pillars downstream to study the effect of pulling airflow through the region. These two locations are hereby called inby fan location and outby fan location respect­ ively and are labeled as such in Figure 2A. Only 3 fan positions were investigated for each of the inby and outby locations positioned in a similar fashion to the original scenario: in the center of the pillar line 11 ft. from the pillar edge in the outer positions and in the center of the entry for the middle position. The same boundary conditions were used for all simulations with only the location of the BF changing within the domain. 4 RESULTS AND DISCUSSION 4.1 Calculation of airflow from simulation files The airflow through north to south entries labeled 1N-8N and east to west entries labeled 1W-12W in Figure 2B was calculated from the simulation file at the centerline of pillars. Figure 2B shows the XZ and YZ planes that were used to calculate the airflow in each simula­ tion along with the corresponding labels. The airflow values were calculated directly from the simulation file by velocity vector integration multiplied by the calculated cross-sectional area of the entry. Since the cells in the mesh are not guaranteed to be uniform in size and the fluid was treated as incompressible, an area weighted average was used for vector integration. The volume of airflow entering the model excluding inlet 9 due to its location in the domain based on the ventilation survey was ~280 m3/s (~487,000 CFM) including the airflow measured from the S-CC and N-CC crosscuts which are 3 entries outby and 5 entries outby the original BF pillar line. Thus, any airflow values calculated from the simulations above 280 m3/s indicates 6

some level of recirculation. The air flowing South-North (1N-8N) will be referenced as ‘entry’ airflow while the air flowing West-East (1W-12W) will be referenced as ‘crosscut’ airflow. Air­ flow that is opposite the BF direction (flowing from north to south) in the entries and from east to west in the crosscuts is labeled as negative airflow and is used to calculate a recirculation ratio. The recirculation ratio is defined as the sum of negative airflow across a pillar line divided by the total airflow across the pillar line.

Figure 2. A) Booster fan placements inside the model geometry, 15 different booster fan locations were simulated with 9 arranged in a grid pattern between the same pillar marked original fan location, another 3 locations are 2 pillars inby the original location and the final group of 3 locations being 2 pillars outby the original location B) Location of airflow measurements from the simulation files, entries parallel to the booster fan direction are labeled 1N-8N while the crosscut entries are labeled 1W-12W with the original booster fan placement shown as the blue fan symbol.

4.2 Scenario one: Original fan location The calculated airflow values from the entries and crosscuts of the nine original BF positions are shown in Figures 3 and 5. The positive and negative airflow values are shown for each pillar line segregated by the entry, 1N through 8N in Figure 3. The BF entry label on the y-axis is the pillar line on the 7W plane shown in Figure 2B. Location 7 shows the highest positive airflow in entry 6N of ~680 m3/s two pillars outby the BF with location 3 showing the lowest airflow of 407 m3/s in the BF pillar line. Locations 3 and 5 show the highest absolute airflow across the pillar line with the BF entry while all other locations reach a maximum one to 3 pillars downstream the BF. Location 2, 3, 5, and 9 all show similar patterns with high airflow at the BF pillar line followed by a decrease in airflow one pillar downstream. This is due to the placement of the BF and the pillar geometry. As noted in Krog and Grau (Krog and Grau, 2006), when higher velocity is observed through the main entry the airflow catches the corners of the pillars and is directed perpendicularly through the crosscuts. This is what was observed at the first crosscut in location 3,4,5,6 and 9 causing a lower airflow value to be observed along the first pillar line downstream. The velocity contour around the breathing height (Z=2m) for original locations 6 and 8 are shown for comparison in Figure 4 with the airflow around the BF circled in black. In Figure 4A the airflow hits the first pillar corner downstream in entry 6N leading to immediate and intense recirculation around the left adja­ cent BF pillar. comparatively, the airflow in Figure 4B fully engulfs entry 6N leading to higher 7

airflows and the traditional entrainment pattern at the first crosscut. Furthermore, the airflow entering the model through inlets 1-8 in Figure 4A is mainly pulled though entry 8N while the inlet airflow in Figure 4B is more uniformly distributed throughout all entries. This shows the importance of BF placement within the context of pillar geometry. The pillar geometry was created from the mine map, and when pillars are offset, or not perfectly aligned due to blast­ ing, then BF airflow and effectiveness can be minimized due to the airstream disruption as described above. In typical mine ventilation surveys that an operator would perform daily it is common to measure airflow downstream of the BF and in an adjacent entry when time per­ mits. With these minimal measurement locations, the region cannot be fully quantified. In Figure 4A and 4B the red circled areas on the left upstream side of the region have minimal airflow of under 0.5 m/s and are insignificantly affected by all fan positions in scenario one.

Figure 3. Airflow through entries 1N-8N for scenario one, fan locations are marked 1 through 9 in a counterclockwise direction, positive airflow indicates flow in the booster fan direction (downstream) and negative airflow indicates flow opposite the booster fan direction (upstream).

The right fan positions in the entry (locations 2, 3, and 9) show the lowest total air move­ ment across the pillar lines and through entry 6N in Figure 3; these locations also have the highest downstream positive airflow through entries 7N and 8N which is likely hindering air entrainment as noted in Krog and Grau (Krog and Grau, 2006) that a more uniform distribu­ tion of airflow will not induce entrainment. Since this uniform airflow does not occur when the BF is on the right side of the entry it is likely due to the fan positioned on the same side of the entry as the adjacent curtain line. Indicating that a fan positioned on the same side of the entry as a curtain/stopping line in adjacent entries will minimize BF airflow due to the loss of the booster fans entrainment abilities. Table 2 shows the recirculation percentage across the pillar lines 1W-12W as defined in Figure 2B in section 4.1. Recirculation percentages above 30% are highlighted in orange while percentages below 10% are highlighted in blue. When no recirculation occurs, the cell is marked with a dash (-). With the BF positioned in the center of the entry (positions 1, 4, and 8) there are similar airflow magnitudes in Figure 3 as to when the BF is positioned on the 8

left side of the entry but with less recirculation as seen in Table 2. An increase in airflow and recirculation can be seen in the first two pillars upstream from the BF when the fan is posi­ tioned on the left-side of the entry (positions 5, 6, and 7) which is not seen with the center positioning (Org 1). The left-side locations (Org 5, 6, and 7) also account for the highest recir­ culation percentages in Table 2. With the fan positioned on the left side the cone effect of the BF is maximized allowing for the airflow to be distributed across the entry without pillar/rib interaction; however, with these fan positions the recirculation around the left adjacent pillars are the highest compared to all other positions in Figure 3. Due to the large airflow (velocity) gradient between the BF entry and the adjacent entries in the left positions it is unsurprising these locations also have the highest airflow due to recirculation and entrainment properties. Interestingly, in Figure 3 the position 8 graph is more similar to the left-side graphs in terms of total airflow, recirculation amounts, and face airflow. Indicating that more total air move­ ment can be achieved when the BF is positioned in either position 8, in the middle of the entry in line with the upstream side of the pillar edge, or in the left positions, on the same side of the entry as the open/face area. Table 2. Recirculation percentages across the Pillar lines (1W-12W) from scenario one showing all 9 booster fan positions.

Figure 5 shows the recirculation percentage through the entries as in Section 4.1. When the BF is positioned in the center of the entry (positions 1, 4, and 8) there are similar airflow mag­ nitudes in Figure 3 as to when the BF is positioned on the left side of the entry but with less recirculation as seen in Figure 5. An increase in airflow and recirculation can be seen in the first two pillars upstream from the BF when the fan is positioned on the left-side of the entry (positions 5, 6, and 7) which is not seen with the center positioning. The left-side locations also account for the highest recirculation percentages for all pillar lines (1W-12W) in Figure 5. With the fan positioned on the left side the cone effect of the BF is maximized allowing for the airflow to be distributed across the entry without pillar/rib interaction; however, with these fan positions the recirculation around the left adjacent pillars is the highest compared to all other positions in Figure 3. Due to the large airflow (velocity) gradient between the BF entry and the adjacent entries in the left positions it is unsurprising these locations also have the highest airflow due to recirculation and entrainment properties. Interestingly, in Figure 3 the position 8 graph is more similar to the left-side graphs in terms of total airflow, recirculation amounts, and face airflow. Indicating that more total air movement can be achieved when the BF is positioned in either position 8, in the middle of the entry in line with the upstream side of the pillar edge, or in the left positions, on the same side of the entry as the open/face area. Figure 5 shows the airflow through the crosscuts labeled1W-12W with each bar representing a north to south pillar lines labeled as right or left of the BF entry. Again, positive airflow in Figure 5 represents air moving from west to east in the geometry in Figure 2B. The total bar height represents the magnitude of airflow through the crosscuts or the exchange between the 9

Figure 4. Velocity contour of positions 6 and 8 of scenario one near the breathing plane (Z=2m), A) location 6, B) location 8.

entries. The BF location in scenario one is between crosscuts 6W and 8W. Given the large amounts of air entrainment seen in positions 1, 4, 6, 7, 8, in Figure 3 it is expected that the air­ flow in the first and second crosscuts (7W and 8W) are high in magnitude and positive on the left and negative on the right side of the BF entry indicating air flowing inwards towards the BF entry i.e., entrainment. Furthermore, the maximum airflow at these locations in Figure 3 is reached a maximum either one pillar (80 ft.) or two pillars (160 ft.) downstream the BF which is consistent with the finding of Krog and Grau (Krog and Grau, 2006) as described by Figure 6 in their work that maximum airflow for propeller fans being reached ~ 52 m (170 ft.) downstream. Moving to ventilation effectiveness across the face area the ability to dilute contaminates generated near the face is also related to crosscut flow due to the geometry, the only way to transport containments out of the mine is towards the general airstream in the BF entry which must be achieved through crosscut airflow. Therefore, it is important to maximize the crosscut airflow near the face (5 and 4 pillars to the left of the BF entry). Locations 5, 6, and 8 in Figure 5 shows similarly high values for airflow 5 entries to the left of the BF entry (between entries 1N and 2N) compared to all other fan positions. This may be due to the higher overall airflow generated by the BF across the region however, location 7 also has high absolute air­ flow of ~804 m3/s which does not translate to higher face airflow. Interestingly at position 7 in Figure 5, marked by the red box, the first left adjacent entry (1 left) has the highest positive and overall airflow through it which does not translate to high face airflow. Thus, leading to the conclusion that not only is high regional airflow necessary to ventilate further adjacent entries but some amount of airstream disruption via air diversion from pillar corners is needed to maximize crosscut exchange between entries. To quantify the influence of the BF to adjacent entries the percent reduction in airflow exchange (crosscut airflow) between the BF entry and adjacent entries can be calculated. When face airflow is higher, as seen in positions 5, 6, and 8 a reduction in crosscut airflow can be seen starting from 1 pillar to the left of the BF entry until approximately 4 pillars to the left as shown by the red arrows in position 5 on Figure 5. When averaged across these positions a total decrease in airflow of 91% on the positive side and 84% on the negative side is seen. Indicating that crosscut airflow is reduced by approximately 46% to 55% for every consecutive adjacent entry from the BF entry. Therefore, the BF effectiveness to facilitate airflow past four adjacent entries is seen to be less than 15% of the BF entry and may not be sufficient to be relied upon for mine ventilation planning. 4.3 Scenario two: Three pillars inby and outby the original fan location Six more simulations were conducted three pillars inby and three pillars outby the original BF location. Booster fan positions 1, 2, and 6 were selected for these inby and outby locations. The calculated airflow values were obtained in the same way as the previous section. 10

Figure 5. Airflow through the crosscuts (1W-12W) from scenario one, fan locations marked 1 through 9 in a counterclockwise direction, positive airflow indicates flow from west to east in the crosscuts (1W to 12W) while negative airflow values indicate airflow in the opposite direction.

Figure 6 shows the calculated airflow values through entries 1N through 8N with the ori­ ginal inby and outby BF locations marked on the y-axis. Inby location 1 shows the highest airflow through entry 6N with a value of ~ 575 m3/s with inby position 6 showing the lowest airflow through entry 6N of ~330 m3/s. Inby location 1 showed significant negative airflow values in entries 3N to 5N throughout the middle portion of the geometry, which facilitates the high airflow in entry 6N through air entrainment as discussed in Section 4.2. The outby fan locations in Figure 6 shows minimal airflow through the majority of the geometry due to the BF being in the outby position. However, similar maximum airflow values are still reached downstream from the BF. All three of the outby fan positions did not have negative airflow until 3 pillars upstream from the BF location which allows the majority of the sec­ tion to have positive airflow. This leads to all three outby simulations (similar to original location 2) having positive airflow throughout the face area (entries 1N-3N). However, the magnitude of the positive airflow seen around the face area was not significantly larger than any of the other simulations that saw negative airflow. Indicating that if avoiding negative airflow is a necessity then a more outby BF location would be desired with a preference towards a fan placement in the middle of the entry or on the same side as the adjacent cur­ tain/stopping lines. The recirculation percentages in Table 3 of the outby and inby fan locations are similar in magnitude to that of the original locations. The recirculation at the original, and outby loca­ tion 6 have the highest magnitude and occurrence while inby location 1 has the highest mag­ nitude and occurrence. Recirculation across the BF pillar line can be thought of as the ‘overall sections recirculation’ given that the air is directly recirculating from downstream to upstream of the BF. Location 6 has the largest magnitude in all positionings, 36% for outby, 30% for original, and 33% for inby positions. This is consistent with the results from Table 2 showing the highest recirculation percentages in original locations 5, 6, and 7 of scen­ ario one. Lastly, the crosscut airflow from the inby, outby, and original locations are shown in Figure 7. A similar reduction in airflow in adjacent entries can be seen in both the inby and outby locations. Inby location 6 shows a significantly different pattern than 11

the rest of the locations due the high airflow through entry 4W. Similar to Figure 4A, the airflow in entry 6N in the inby location 6 simulation caught the corner of a pillar and split the air stream between entry 6N and crosscut 4W. This air stream continued towards the face and split resulting in positive airflow downstream and negative air­ flow upstream through entry 1N as shown by the red circle in the inby location 6 graph in Figure 6. Even with this extreme example of airflow being directed towards the face the average reduction in crosscut airflow still remains and is 79% and 74% for positive and negative airflow respectably in the inby locations. The average reduction for outby locations is 91% and 77% for positive and negative airflow respectably. These reductions correspond to a 25-30% reduction per pillar line in all cases, indicat­ ing that positioning the BF at an inby or outby locations does not have a significant effect on crosscut airflow.

Table 3. Recirculation percentage of the inby, original, and outby fan locations, fan locations 1, 2, and 6.

Figure 6. Airflow through entries 1N-8N for booster fan locations 1, 2, and 6 at the inby, original, and outby fan locations, positive airflow indicates flow in the booster fan direction (downstream) and nega­ tive airflow indicates flow opposite the booster fan direction (upstream).

12

Figure 7. Airflow through the crosscuts (1W-12W) from the inby, original, and outby fan locations, posi­ tive airflow indicates flow from entries west to east in the crosscuts (1W to 12W) while negative airflow values indicate airflow in the opposite direction.

5 SUMMARY AND CONCLUSION A CFD model was created based on a mine ventilation survey of a typical perimeter ventilation schemed large opening room and pillar underground mine. 15 BF positions were investigated in two scenarios: the first scenario consisted of nine fan positions around the original BF placement selected by the mine operator with the second scenario consisting of three fan positions three pil­ lars inby and outby the original location. The airflow around the BF was investigated through recirculation patterns and overall airflow within the BF entry and the room and pillar region with a focus around the face velocity. The following conclusions were made based on the simulations: (1) (2)

(3) (4) (5)

The highest total airflow in the section may not always be obtained when the BF is posi­ tioned at the center of the entry but rather when the BF is positioned on the same side of the entry as the open room and pillar areas. As noted in Krog and Grau (Krog and Grau, 2006) and confirmed in these simulations air entrainment is maximized when high velocity gradients or non-uniform distribution of airflow is seen through the adjacent BF entries. Furthermore, the maximum airflow through the BF entry was confirmed through the simulations to be ~ 160 ft downstream the BF which is consistent with their findings of ~170 ft. Not only is high regional airflow necessary to ventilate further adjacent entries from the BF but some amount of airstream disruption via air diversion from pillar corners is needed to maximize crosscut exchange between entries. The airflow through the crosscuts observes a 46%-55% reduction per entry indicating that booster fans in this geometry have minimal effect past 4 adjacent entries. Leading to a recommendation to keep the BF within 3-4 entries of the face. In situations where recirculation must be avoided a BF located more outby from the center of the section is preferable. This outby fan location was shown to create ‘positive’ airflow around the face areas i.e., non-recirculating airflow.

While all mine geometries are unique and BF positioning should be investigated on a caseby-case basis with the purpose of the BF in mind, these simulations and recommendations can act as a starting point for LOM operators who utilize a perimeter ventilation scheme. 13

ACKNOWLEDGEMENT This work was financially supported by The National Institute of Occupational Safety and Health (NIOSH) under contract No. 75D30119C05743. REFERENCES Babu, V.R., Maity, T., Prasad, H., 2015. Energy saving techniques for ventilation fans used in under­ ground coal mines—A survey. J. Min. Sci. 51, 1001–1008. https://doi.org/10.1134/S1062739115050198. Carter, R., 2018. Focusing the Flow. Eng. Min. J. 219, 50–56. Dunn, M., Kendorski, F., Rahim, M., Mukherjee, A., 1983. Testing Jet Fans in Metal/nonmetal Mines With Large Cross-sectional Airways.pdf. Gendrue, N., Liu, S., Bhattacharyya, S., Clister, R., 2023. An investigation of airflow distributions with booster fan for a large opening mine through field study and CFD modeling. Tunn. Undergr. Sp. Technol. 132. Goodman, G., Taylor, C., Thimons, E., 1992. Jet Fan Ventilation in Very Deep Cuts-A Preliminary Analysis. Grau III, R.H., Krog, R.B., Robertson, S.B., 2006. Maximizing the ventilation of large-opening mines. Proc. 11th U.S./North Am. Mine Vent. Symp. - 11th U.S./North Am. Mine Vent. Symp. 2006 53–59. https://doi.org/10.1201/9781439833391.ch8. Grau III, H., Krog, R., 2009. Using mine planning and other techniques to improve ventilation in large-opening mines. Min. Eng. 61, 46–50. Hargreaves, D.M., Lowndes, I.S., 2007. The computational modeling of the ventilation flows within a rapid development drivage. Tunn. Undergr. Sp. Technol. 22, 150–160. https://doi.org/10.1016/j. tust.2006.06.002. Krog, R.B., Grau III, R.H., Mucho, T.P., Robertson, S.B., 2004. Ventilation planning layouts for large opening mines. Soc. Mining, Metall. Explor. 1–9. Krog, R.B., Grau, R.H., 2006. Fan selection for large-opening mines: Vane-axial or propeller fans Which to choose? Proc. 11th U.S./North Am. Mine Vent. Symp. - 11th U.S./North Am. Mine Vent. Symp. 2006 535–542. Leonida, C., 2019. Changing the Face of Mine Ventilation. Nguyen, V.D., Heo, W.H., Kubuya, R., Lee, C.W., 2019. Pressurization ventilation technique for con­ trolling gas leakage and dispersion at backfilled working faces in large-opening underground mines: CFD analysis and experimental tests. Sustain. 11. https://doi.org/10.3390/SU11123313. Parra, M.T., Villafruela, J.M., Castro, F., Méndez, C., 2006. Numerical and experimental analysis of different ventilation systems in deep mines. Build. Environ. 41, 87–93. https://doi.org/10.1016/j. buildenv.2005.01.002. Thimons, E., Kohler, J., 1985. Measurement of air velocity in mines.

14

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Quantifying assemblage losses in auxiliary ventilation systems E. De Souza AirFinders Inc., Kingston, Canada

ABSTRACT: While mine ventilation systems may account for 40% to 50% of the energy con­ sumption of a mine operation, auxiliary ventilation alone may be accountable for half of this con­ sumption. In effect, auxiliary ventilation systems comprise a significant portion of a mine operation’s base energy demand and is consequently responsible for a large percentage of the total mine operating costs. This paper presents how engineering design principles can be applied to design efficient and reliable auxiliary ventilation systems, especially focusing on assemblage losses. Case studies are presented to demonstrate the effect of design, installation and maintenance prac­ tices on system reliability and operating costs. In particular, the effect of assemblage losses is quantified in terms of operating efficiencies, energy consumption and costs.

1 INTRODUCTION Increasing costs of electricity have resulted in emphasis on energy-efficient designs and oper­ ation for all energy-consuming systems in mining. Since ventilation systems normally account for 25-40% of the total energy costs and 40-50% of the electrical consumption of a mine oper­ ation (De Souza, 2018, 2013), the optimization of ventilation systems is today a ventilation engineer’s top priority. While modern duct-fan systems require precise engineering design, meticulous attention to installation and regular maintenance practices, many installations are often designed based on outdated rules of thumb and with disregard to best installation practices. In the course of many years of investigations of duct-fan systems, the author has found them to be, in general, fairly energy inefficient, with many systems operating at efficiencies below 65% and with air leakages ranging between 25% and 75% (De Souza, 2004). Auxiliary ventilation systems comprise a significant portion of a mine operation’s base energy demand and may be accountable for half of a mine ventilation system energy consumption. Typical auxiliary ventilation systems are fairly energy inefficient; factors affecting mechanical and ventilation efficiencies loss include design flaws, installation practices and air leakage. General solutions and tactics for improving auxiliary ventilation systems as presented in this paper come from multiple ventilation audits performed by the author. They target subsystem components which affect shock losses and mechanical and ventilation efficiencies. By increasing the efficiency of auxiliary ventilation system components and by correcting inappropriate designs or system degradation caused by poor maintenance, the overall capacity of the system in delivering air to the required active faces can often be improved. 2 CASE APPLICATION A case application associated with extensive engineering work conducted by the author is pre­ sented in this section to demonstrate how, by conducting detailed ventilation efficiency audits, simple low-cost solutions can be devised to increase efficiency, reduce power consumption, and lower operating costs. DOI: 10.1201/9781003429241-2

15

The case study is based on the auxiliary ventilation system illustrated in Figure 1. A series of analysis are performed to quantify the contribution of each component of the complete system to energy consumption and costs: inlet bell, screen, silencers, system friction, couplings, bends, duct exit. The effect of air leakage and of installation practices is also quantified. The analysis was based on analytical procedures and scientific guidelines developed by the author (De Souza & Dirige, 2022) and use of specialized software (AirFinders, 2022). The auxiliary ventilation system, to be installed in a development drift to supply 18.88 m3/s air for a diesel production fleet rated at 298.3 kW, has the following design characteristics: – layflat duct - new, 1.2192 m diameter, provided in 15.24 m long sections, 10 sections of ducting. Multi clip joints. – spiral duct, new, 1.2192 m diameter, 1 section 15.24 m for bend. Multi clip joints. – total column length - 167.64 m. – fan: 1.219-0.80-1780 (fan diameter-hub diameter-rpm) with a 149.14 kW motor. 600 V. Motor efficiency 95%. Power factor 0.84. – inlet bell - 1.651 m diameter, 0.2159 m long. – screen - wire mesh screen of 95% net free area. – silencers - two podless flow through silencers. – bend - right angle normal bend of 1.524 m radius. – cost of power: $0.08/ kW.hr. – operation - 24 hours/day, 365 days/year. – standard density conditions. Three cases, with operational features presented in Table 1, are considered. Case 1 represents the system design to meet the flow requirements of at the face of 18.88 m3/s and considers a design air leakage of 20%. The fan supply flow is 23.6 m3/s. An ‘installation quality factor’ of 10%, representing a ‘good’ installation, is used to adjust the system static resistance pressure. Case 2 represents the system as installed. Attained flows at the face of 15.56 m3/s do not meet requirements, and is based on an air leakage of 30%. The fan supply flow is 22.23 m3/s. An ‘installation quality factor’ of 30%, representing a ‘poor’ installation, is used. Case 3 represents the system with the fan blade setting adjusted to achieve the required face flow of 18.88 m3/s. Air leakage is not constrained, remaining at 30%. The fan supply flow is 26.97m3/s. The system installation quality is not improved; an ‘installation quality factor’ of 30% is used. Table 1. Operational features of three duct system cases. Case

1

2

3

Face Flow Leakage Fan Flow Installation Quality Inlet Bell Screen Silencers Friction Head Loss Coupling Losses Bend Duct Exit

18.88 m3/s 20% 23.6 m3/s Good Installed 95% Net Free Area Proper Connection Same Resistance Same Resistance Proper Installation Full Section

15.56 m3/s 30% 22.23 m3/s Poor No Bell 30% Net Free Area Non-Aerodynamic Same Resistance Same Resistance ‘Kinked’ Reduced Section

18.88 m3/s 30% 26.97 m3/s Poor No Bell 30% Net Free Area Non-Aerodynamic Same Resistance Same Resistance ‘Kinked’ Reduced Section

The contribution of each duct system component to power consumption and operating costs is presented in the following sections. 2.1 Inlet bell The fan inlet bell ensures smooth air flow through the fan intake and serves to minimize entrance losses. Case 1 has a proper inlet bell installed and in Cases 2 and 3 the fan is installed 16

without an inlet bell. Table 2 presents a summary of results for the 3 cases. For cases 2 and 3, substantial increases in operating cost of 723% and of 1,370% are noted when an inlet bell is not used. Table 2 clearly shows that, when an inlet bell is installed, significant energy and cost savings can be achieved. Table 2. Head losses, power, and operating costs for inlet bell. Case

Head Loss (Pa)

Power (kW)

Cost/Year ($/y)

% Change in Cost

1 2 3

14.73 108.90 160.27

0.59 4.84 8.64

412.28 3,393.06 6,058.39

– 723 1,370

Figure 1.

Auxiliary ventilation system configuration.

2.2 Screen The fan screen prevents debris from entering the fan. Case 1 has a screen of 95% net free area installed and in Cases 2 and 3 the fan screen is partially blocked by the deposition of debris. Table 3 presents a summary of results for the 3 cases. For cases 2 and 3, substantial increases in operating cost of 340% and of 686% are noted when the screen is partly blocked with debris and not well maintained. Table 3. Head losses, power, and operating costs for screen. Case

Head Loss (Pa)

Power (kW)

Cost/Year ($/y)

% Change in Cost

1 2 3

11.29 44.65 65.71

0.45 1.99 3.54

315.99 1,391.19 2,483.91

– 340.26 686.06

2.3 Silencers Silencers provide a level of noise reduction to meet specific needs and for compliancy with regulations. Case 1 has silencers properly connected to the fan and in Cases 2 and 3 the silen­ cers have a non-aerodynamic connection to fan. Table 4 presents a summary of results for the 3 cases. For cases 2 and 3, relatively large increases in operating cost of 19.7% and of 45% are noted when the silencers are not properly installed. 17

Table 4. Head losses, power, and operating costs for silencers. Case

Head Loss (Pa)

Power (kW)

Cost/Year ($/y)

% Change in Cost

1 2 3

49.77 53.50 53.50

1.99 2.39 2.89

1,392.95 1,666.93 2,022.36

– 19.67 45.19

2.4 Friction head losses for layflat duct For the three cases, the layflat duct has the same frictional resistance, however the resistance pressures vary as a function of the airflow volumes passing through the duct column due to leakage effects. Table 5 presents a summary of results for the 3 cases. Because of the reduced flows, case 2 has a reduced operating cost of 19.1% and, for case 3, because of the increased fan flow, an increase in operating cost of 44.4% is noted. Table 5. Head losses, power, and operating costs for layflat duct friction losses. Case

Head Loss (Pa)

Power (kW)

Cost/Year ($/y)

% Change in Cost

1 2 3

604.66 469.54 691.04

21.60 17.47 31.19

15,136.12 12,240.01 21,854.54

– –19.13 44.39

2.5 Friction head losses for spiral duct For the three cases, the spiral duct has the frictional resistance, however the resistance pres­ sures vary as a function of the fan flow. Table 6 presents a summary of results for the 3 cases. Because of the reduced fan flow, case 2 has a reduced operating cost of 1.2% and, for case 3, because of the increased fan flow, an increase in operating cost of 76.4% is noted. Table 6. Head losses, power, and operating costs for spiral duct friction losses. Case

Head Loss (Pa)

Power (kW)

Cost/Year ($/y)

% Change in Cost

1 2 3

226.75 201.23 296.16

9.06 8.95 15.97

6,346.37 6,269.83 11,195.12

– –1.21 76.40

2.6 Coupling losses for layflat duct For the three cases, the layflat duct has the same resistance due to couplings, however the resistance pressures vary as a function of the airflow volumes passing through the duct column due to leakage effects. Table 7 presents a summary of results for the 3 cases. Because of the reduced flows, case 2 has a reduced operating cost of 19.1% and, for case 3, because of the increased fan flow, an increase in operating cost of 44.4% is noted. Table 7. Head losses, power, and operating costs for layflat duct coupling losses. Case

Head Loss (Pa)

Power (kW)

Cost/Year ($/y)

% Change in Cost

1 2 3

64.27 49.91 73.46

2.30 1.86 3.31

1,608.96 1,301.10 2,323.12

– –19.13 44.39

2.7 Coupling losses for spiral duct For the three cases, the spiral duct has the same resistance due to couplings, however the resist­ ance pressures vary as a function of the fan flow. Table 8 presents a summary of results for the 18

3 cases. Because of the reduced fan flows, case 2 has a reduced operating cost of 1.2% and, for case 3, because of the increased fan flow, an increase in operating cost of 76.4% is noted. Table 8. Head losses, power, and operating costs for spiral duct coupling losses. Case

Head Loss (Pa)

Power (kW)

Cost/Year ($/y)

% Change in Cost

1 2 3

71.42 63.38 93.28

2.85 2.82 5.03

1,998.86 1,974.75 3,526.02

– –1.21 76.40

2.8 Bend Case 1 has a properly designed bend and, in Cases 2 and 3, the installed bend is ‘kinked’, resulting in a higher resistance pressure. Table 9 presents a summary of results for the 3 cases. For cases 2 and 3, relatively large increases in operating cost of 97.6% and of 252.8% are noted when the bend is not properly installed. Table 9. Head losses, power, and operating costs for bend losses. Case

Head Loss (Pa)

Power (kW)

Cost/Year ($/y)

% Change in Cost

1 2 3

45.40 80.59 118.60

1.81 3.58 6.40

1,270.75 2,510.85 4,483.27

– 97.59 252.80

2.9 Exit losses Case 1 has a properly installed duct end, with its full cross-section open, and in Cases 2 and 3 the installed duct end has its exit reduced in section, resulting in a higher resistance pressure. Table 10 presents a summary of results for the 3 cases. For cases 2 and 3, relatively large increases in operating cost of 109.2% and of 273.5% are noted when the duct end is not prop­ erly installed. Table 10. Head losses, power, and operating costs for exit losses. Case

Head Loss (Pa)

Power (kW)

Cost/Year ($/y)

% Change in Cost

1 2 3

157.07 337.29 496.41

5.02 10.50 18.74

3,516.58 7,356.40 13,134.63

– 109.19 273.51

2.10

System component contribution summary

Table 11 presents the contribution of each individual component, relative to Case 1, to annual operating costs. For case 2 the overall system component contribution to costs is 19.1% and for case 3 it reaches 109.6%. This clearly indicates the potential for cost savings when an auxil­ iary ventilation system is properly designed, commissioned, and maintained. 2.11

Fan operation

Figure 2 presents the fan curve and operating points for the 3 cases and Table 12 presents details of the fan operation. For cases 2 and 3, relatively large percent increases in fan input power and operating cost of 10% and of 94.5%, relative to case 1, are noted. It is pointed that, for case 2, the system is not in compliance since the face supplied flow does not meet regula­ tory requirements. Also, for case 3, the fan operates very close to stall. To reduce risk in both cases, the installed system components should be improved including, installing an inlet bell, maintaining the fan screen, correcting the silencer connections, installing a proper bend, and 19

Table 11. Component percent contribution to annual operating costs. Component

Percent $/year change relative to case 1

Inlet bell Screen Silencers Friction layflat Friction spiral Joints layflat Joints spiral Bend Exit Overall

Case 2 723.0 340.3 19.7 –19.1 –1.2 –19.1 –1.2 97.6 101.2 19.1

Case 3 1,369.5 686.1 45.2 44.4 76.4 44.4 76.4 252.8 273.5 109.6

the correcting duct outlet. Also, the duct column installation quality should be improved, and air leakage controlled. These actions will bring the system operation close to the design specifications.

Figure 2.

Fan characteristics and operating points for the three cases.

Table 12. Fan operation for the three cases. Flow Case (m3/s) 1 2 3

23.6 22.23 26.97

TP (Pa)

Blade Angle (degrees)

Input Efficiency Brake Power Power Cost/Year % Increase (%) (kW) (kW) ($/y) in Cost

1,615.30 2,049.48 2,983.49

19.5 19.5 28

64.5 70 70

59.1 65.09 114.94

20

62.2 43,773 68.51 48,160 120.99 84,797

– 10 94.5

2.12

Operational costs

Table 13 presents annual fan operating costs as a function of the number of duct installations. Independent on the number of installations, fan operating costs increase by 10% and 93.7% for cases 2 and 3, relative to case 1. Typical hard rock mines may have well over 40 auxiliary ventilation systems installed to support development and production activities. For the case study presented, annual cost savings reaching some $1.65M can be realized when a system is installed and maintained according to design. It is noted that the comparative analysis was based on direct costs only. It is recognized that ventilation training, asset management and preventive maintenance produce significant reduc­ tions in ventilation operating costs. While indirect costs are site dependent, the author has observed that when workers acquire practical ventilation training prior to working under­ ground, which normally comes at a very nominal cost, considerable improvements in ventila­ tion system performance are readily realized. Table 13. Annual operating costs for multiple duct system installations. Number of Duct System Installations Case

1

5

10

20

30

40

1 2 3

43,773 48,160 84,797

218,865 240,800 422,983

437,730 481,600 847,965

875,460 963,200 1,695,930

1,313,190 1,444,800 2,543,895

1,750,920 1,926,400 3,391,860

3 CONCLUSIONS This paper has demonstrated how the use of proper engineering design for optimal duct system installations would dramatically result in reduced energy consumption and in reduced operating costs. When properly designed, installed, and maintained, an auxiliary ventilation system can operate efficiently with substantial power and operating cost savings. In a case study presented, it has been illustrated that savings in fan energy consumption exceeding 94% can be achieved by correctly commissioning and maintaining a properly designed auxiliary ventilation system. REFERENCES AirFinders, 2022. AirFinders Force Auxiliary Ventilation System Design. Engineering Design software. De Souza, E. 2004. Auxiliary ventilation operation practices. Proceedings of 10th U.S./ North American Mine Ventilation Symposium. Anchorage:Balkema. 341–348. De Souza, E. 2013. Improving the energy efficiency of mine fan assemblages. Proceedings of the 23rd World Mining Congress, Montreal, QC. pp 9. De Souza, E. 2018. Cost saving strategies in mine ventilation. CIM Journal. Volume 9. Issue 2. De Souza, E. & Dirige, P. 2022. Auxiliary mine ventilation manual. Version 2.0. Workplace Safety North. Pp 261.

21

Case studies of mine ventilation

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Kucing-Liar (KL) mine LOM preliminary ventilation design – PT Freeport Indonesia A.A. Habibi Technical Expert, PT Freeport Indonesia

I. Setiawan Chief Ventilation Engineer, PT Freeport Indonesia

R. Prasojo Mine Engineer, PT Freeport Indonesia

C. Stewart Principal Engineer, Minware Consulting

ABSTRACT: A ventilation study was conducted on Kucing Liar (KL) mine. KL is the add­ ition to PT Freeport Indonesia (PTFI) panel cave operations and is anticipated to operate in parallel to Grasberg Block Cave (GBC). The study consisted of model construction, budget preparation and investigating multiple scenarios to meet the minimum budget requirements. Utilization of new engine technologies controlled partial air reuse, push-pull and utilization of Battery Electric Vehicles (BEVs) were considered during this feasibility study. A main intake and exhaust drift size optimization study was also conducted to determine the number of add­ itional main fans and drift sizes required to support the KL mine. The results show the requirement for two additional 5MW main fans, two intake drifts and two exhaust drifts (7 m x7 m). The results also show the requirement for eight 6m ventilation raises to support primary air delivery to KL mine. Proof of concept trials are deemed necessary to determine the feasibility of applying some of the proposed concepts in underground operations (such as BEVs). The trial results will be used to complete a detailed ventilation network design. 1 INTRODUCTION GBC mine development started in 2004 with initial drifting and the first drawbell blast in 2018. GBC mine is expected to achieve full production rates of 130 kt/d to 160 kt/d of ore by 2025. Currently, the three active underground mining operations in the Grasberg Mining District are the GBC block cave mine (target 130-160 kt/d ore), the Deep Mill Level Zone (DMLZ) panel cave mine (target 80 kt/d), and the Big Gossan open stoping operation (7 kt/d). The Deep Ore Zone (DOZ) panel cave mine (80 kt/d) has been closed and is no longer in operation. The KL orebody is a new large caving operation where development started in 2021 and pro­ duction is due to begin in 2027. The Grasberg Open Pit commenced production in 1990 with active pit production ending in late 2019. Figure 1 is a schematic view of the Grasberg District orebodies. The concentrating plant has a peak capacity of about 240 kt/d. Peak production targets from the underground district will remain at 240 kt/d and will be entirely from underground. The GBC will be the primary source of mill feed targeted at 130-160 kt/d (Brannon et al. 2020). 1.1 Existing ventilation infrastructure and network layout KL mine is located at a lower elevation from GBC. The main exhaust and intake drifts (labled the Grasberg Vent Drift (GVD) system) will be shared between the existing GBC mine and future DOI: 10.1201/9781003429241-3

25

KL mine. Five Howden mixed flow rated at 5 MW each are currently installed in a parallel in an exhaust configuration to support the GBC mine. Fresh air from the mill valley is downcast to the mine through four parallel primary intakes (GVD 1-4) mined at 6.8 mW × 9.0 mH, at 2.6 km from the portals to the footprint. Air is distributed to the working areas of the mine using ventila­ tion drifts and raises mined at various dimensions. The return air reports back to the under­ ground main exhaust fans and out into the valley. Figure 2 displays the main components of the ventilation system. The system capacity is 3500 m3/s when all five fans are running. Based on the recent LOM baseline estimate, the current GBC mine (which will be operating in parallel with KL block cave) consumes approximately 25 m3/s/ktpd. Other block cave mines typically have ratios from 17 to 40 m3/s/ktpd (Brannon et al. 2020). while the ratio cur­ rently sits at the lower end of the scale, a Ventilation On Demand (VOD) system has been put in place to improve the efficiency of air distribution.

Figure 1.

PTFI mining district overview.

Air velocity and diesel exhaust dilution requirements form the main design criteria. The air­ flow velocity requirements air primarily for dust control in the ore flow system, and diesel dilution for development and production activities.

Figure 2.

GBC and KL mines network layouts.

26

1.2 Ventilation budget The KL mine shares primary intake and exhaust drifts with the GBC mine the ventilation cap­ acity of this system is insufficient for the combined GBC and KL development and produc­ tion, and will need to be expanded by the year 2026 with additional intake and exhaust drift(s) and additional primary ventilation fan(s). Based on current activity forecasts, GVD ventilation capacity is required to increase from 3500 m3/s to 5100 m3/s, proportional to planned production increases from the combined mines. KL mine requirements peak briefly at 2400 m3/s, although as this peak is short-term, 2200 m3/s is used as an economic design target for long-term requirements. Freeport PT mine airflow requirements are based on diesel equipment and minimum air vel­ ocity criteria, depending on location and activity. • All working areas must maintain a minimum 0.3 m/s velocity standard. • Fixed facility infrastructure with working personnel adopts a higher 0.5 m/s velocity standard. • Remotely operating diesel equipment (MineGEM) requires a minimum air velocity of 0.8 m/s to ensure equipment cooling. • Diesel activities with personnel present, until recently were required to meet an 0.08 m3/s/kW standard at a nominal machine utilization factor (typically 80%). This has since been simpli­ fied to 0.067 m3/s/kW at 100% machine installed engine power. • For BEV ventilation budget calculations, minimum velocity of 0.6 to determine the budget, which exceeds the minimum velocity requirement. This approach resulted in higher airflow requirement compared to kW-based approach. The above criteria represent minimum flows that should be met or exceeded at the work location. To allow for operational loss and leakage, a leakage factor is applied to some activ­ ities, and a ‘balancing’ factor to maintain minimum airflow delivery to multiple locations is also applied. In addition, an air density factor at the primary fans is applied to account for density differences due to elevation. 2 DESIGN SCENARIOS The GBC mine production is scheduled to increase to 136 ktpd by 2024. KL mine develop­ ment has commenced, with steady-state full production planned for 90 ktpd by 2033. The mine will supplement GBC mine production which reduces to 103 ktpd by 2033, giving a combined total tonnage from both mines from 2033 of 193 ktpd. Five scenarios were con­ sidered as case studies for this design. 2.1 Scenario 1. Base case - diesel mine using existing strategies The base case represents a diesel equipment mine, operated to existing strategies and airflow guidelines. To meet planned production requirements, the proposed baseline ventilation design increases total airflow through the combined GBC + KL primary system from 3500 m3/s to 5100 m3/s, an increase of 42%, in line with the production increase. Economic analysis using mining and fan capital costs and discounted life of asset operating costs revealed there was no prospect of increasing airflow through existing mining infrastructure due to high system resistance, existing fan performance limitations, and other practical consid­ erations (such as mine and ventilation disruption during construction). The primary surface system design requires an additional two (2) exhaust drifts, with pri­ mary fans (two 5MW mixed flow fans) totaling 5.2 km development drifts. In addition, two intake drifts totaling 4 km DEQ are required. Air to KL mine is delivered from the GVD exhaust and intake system via four (4) intake raises and four (4) exhaust raises totalling 2.9 km at 6m in diameter.

27

2.2 Scenario 2 - KL mine air resuse An air reuse option redistributes 600 m3/s of cleaned infrastructure exhaust to the extraction level, effectively reducing the airflow drawn to and from the surface by 600 m3/s. This permits a reduction in primary main fan infrastructure requirements, and a resultant reduction in cap­ ital and operating costs. Infrastructure level exhaust is proposed for reuse as it is lower (per unit volume) in diesel activity (DPM is more difficult to scrub and reduce). Reintroduction of the scrubbed air to the extraction horizon is a form of ‘series’ ventilation meaning gases will not continue to build as may be the case in a recirculated design. This proposal is speculative, in that no effective validated air-reuse scrubber arrangement has been proven for this mine. To do this, the characterization of dust and contaminants would need to be performed, and available technologies evaluated for cost and efficiency. A proposed arrangement is to parallel 80 m3/s capacity scrubbing units together in at least two locations on the infrastructure air return level, linked by raises into the extraction level. Several proven wet and dry technologies exist for removing fine silica and nuisance dust from the air. Removal of DPM from the air requires dry filter scrubbing which should be avoided from a likely cost and maintenance perspective. The most likely robust fine dust removal technology is ultra-fine wet spray chambers, using water atomized with the assistance of air compressors. For design purposes, power totaling 400 kW is assumed to overcome 2 kPa of filter media or sprays with collectors, and 1.5 kPa air reinjection into the mine extrac­ tion level. 2.3 Scenario 3 - BEV utilization at extraction level combined by air reuse The use of BEVs for loader extraction activities is expected to reduce airflow requirements by 275 m3/s. To achieve a step change to 600 m3/s airflow reduction (required to eliminate an intake and exhaust drift), a further 325 m3/s is assumed to be re-used – a limited version of the previous option. The option reduces infrastructure requirements similar to the previous air reuse case (one less intake and exhaust drift/fan) but is slightly cheaper and lower in power cost due to the reduction of air scrubbers. The capital and productivity assumptions of BEV loaders are excluded from this cost. 2.3.1 BEV design details Consideration of BEVs for production mucking and loading activities was undertaken to determine potential ventilation improvements to the KL mining zone. BEVs are considered zero-emission vehicles, emitting no discernible gases or diesel particulates. Oxygen levels remain constant and heat emissions are estimated at less than 20% compared to diesel equip­ ment (McGuire, et. Al. 2022). The minimum airflow requirement for BEVs is therefore limited by only: • Minimum air velocity requirements • Build-up and removal of dust that would otherwise cause health & safety issues, or visibility problems. • Other noxious gases that enter the atmosphere – e.g., blasting fumes or oxidation of ores. • Heat build-up from all sources, but primarily from external sources such as rock strata or hot water, instead of BEV machinery. Fire risk and resultant toxic gas emissions from BEV fires have been flagged as serious haz­ ards. Despite several mining BEV fires reported worldwide, no evidence to date suggests BEV fires in underground mining are more common or more hazardous than diesel machine fires (Stewart, 2022). Nonetheless, hydrogen fluoride emissions during lithium battery fires have been flagged as a potentially lethal atmospheric hazard, and when coupled with burning hydrocarbon products (such as tires) and potentially reduced airflows for BEV use, the con­ centrations of fire combustion products may be more hazardous than an equivalent well28

ventilated diesel machine fire. A BEV LHD trial is underway (in 2023) as a follow-up to hazard and risk assessment strategy identifying BEV-specific safety issues, controls and mine design changes. 2.3.2 BEV design strategy The prime potential for ventilation reductions in BEVs comes from the removal of the diesel airflow allocation rule. For PTFI, 0.067m3/s/kW would no longer be the minimum airflow standard for operating machines. BEV heat emissions are unlikely to cause heat build-up even with reduced airflow. Moisture and humidity, partial products of diesel exhaust, are also reduced contributing to dryer atmos­ pheres and potentially reduced wet bulb temperatures. Currently, PTFI uses a diesel airflow requirement of 0.067 m3/s/kW to satisfy Government regulations and maintain suitable working temperature of the mine. This reduction in sug­ gested airflow requirements is further supported by BEV field test work in Canada which revealed equivalent heat output of electric loaders in production activities is reduced by a factor of 5.3 or to 18.7% of an equivalent diesel loader (McGuire, et. Al. 2022), which could arguably support reductions in flow rates to as low as 0.013 m3/s/kW to provide the same WBGT increase. Alternatively, a minimum airflow velocity approach can be taken, ensuring there is suffi­ cient capacity to remove BEV and strata heat, dust and gas build-up. Applying a minimum airflow velocity of 0.5 m/s to KL Mining schedules and airway sizes for example, in most cases would result in an equivalent airflow ranging from 0.025 to 0.040 m3/s/kW of equivalent electric power. Being zero-emission vehicles, the type of BEV is largely irrelevant for ventila­ tion considerations. Apart from diesel activity areas, other active areas at PTFI have a budget airflow allocation of 0.5 m/s while passive areas can be reduced to 0.3 m/s. Assuming dust and heat can be man­ aged at reduced airflow, a conservative estimate of BEV airflow can be made as follows. An active airflow allocation of 0.5 m/s is assumed for BEV-operated areas, including Mine­ GEM remote loaders. This allows for reasonable dust and blast fume removal and provides adequate cooling for exposed working personnel The proposed guide reduces extraction panel ventilation to 7 m3/s while development with electric trucks and loaders could be reduced to 17-25 m3/s (depending on drift size). Figure 3 shows budgeted estimates for airflow requirements for different BEV options. An extraction level loader option is considered, reducing extraction airflow requirements from 730 m3/s to 460 m3/s (a reduction of 270 m3/s or 12% of total KL Mine flow). To create a step reduction in ventilation infrastructure (i.e., one less intake and exhaust drift), an additional 330 m3/s reduction in demand is still required, most likely through some air reuse. 2.4 Scenario 4. full BEV development and extraction This scenario engages the use of BEVs for all major mining activities. This permit airflow reductions in both development and production areas, although haulage is maintained at simi­ lar airflow levels due to the 0.5 m/s requirement to ventilate active chutes (meaning diesel trucking for haulage transport could continue). While reduced air usage for BEVs does have risks (greater potential concentration of air­ borne dust, slower removal of gases and blast fumes), the reduction in total airflow for KL mine (29%) is seen as realistic and achievable and avoids the need to rely on speculative scrub­ ber technology for air reuse. A full BEV option (trucks and loaders) considering both development and extraction activ­ ities, results in a 650 m3/s airflow reduction (29% of total KL mine flow). The lower airflow requirements (if combined with mine air reuse and scrubbing for the partial BEV case) will reduce capital mining infrastructure by one GVD intake and exhaust drift, and one less pri­ mary fan. The comparison highlighting airflow versus BEV options is shown below in Figure 3.

29

Figure 3.

KL mine ventilation budget reduction.

2.5 Primary ventilation drift optimization (GVDs) The GVD primary ventilation drifts will need to support both the existing GBC block and the new KL block. An optimization study was performed to determine the additional number and drift dimensions of primary ventilation drifts to support the additional activity in the KL mine for each ventilation scenario. The option with the lowest combined capital cost of mining ventilation drifts and new primary fans, and discounted power costs over the life of the mine is the preferred velocity. Base case scenario for the analysis is assuming diesel powered mobile equipment. The Net Present Value (NPV) analysis has been discounted at 10% for this study and is based on budget PTFI power and mining costs. A 95% total cost variance limit was chosen to reflect variable confidence in the assumptions used and ensure the recommended outcomes are close to the maximum economic value. While larger drift sizes show a slight increase for higher economic velocities, as values for most sizes are close, it is suggested an optimum velocity of around 10.5 to 11 m/s for any drift size should be adopted for planning purposes, with a peak upper limit of 14.0 m/s. Internal shaft velocities can vary between 13.1 m/s and 20.5 m/s. Figure 4 shows an analysis example for the 6.8 m X 9 m drift. Table 1. Calculated optimum velocity for different drift profiles. Drift Size W (m)

H (m)

4.4 4.0 6.8 6.0 7 7.0 6.8 9.0 6.0 m Raise

Optimum Velocity

Upper Limit 95%

Lower Limit 95%

Max Upper Airflow

m/s

m/s

m/s

m3/s

10.8 11.0 11.0 11.8 16.5

13.3 13.6 13.6 14.6 20.5

8.6 8.6 8.6 9.4 13.1

234 555 666 894 580

30

Figure 4.

Velocity optimization for primary ventilation drifts.

Where personnel are required to travel in high-velocity drifts, dust and coarse particle entrainment in the air can create safety hazards, particularly with eye injuries. The KL design for PTFI uses an economic velocity range that may exceed 10 m/s for infre­ quently travelled airways while maintaining 10 m/s or less for common travel ways. Spe­ cial safety precautions and travel restrictions will apply for travelling in airways exceeding 10 m/s. Nonetheless, every effort is made to reduce air velocity to below the maximum economic limit. For example, for Scenario 1 (fully diesel mine, no air re-use) two GVD intakes have been added to the system as optimum velocity analysis indicates that this can be done at a cost-neutral basis, the impact being higher capital costs being offset by longer-term lower operating costs. 2.6 KL Block profile size optimization A service level drift optimization process was conducted for KL footprint to reduce the mining cost in intake and exhaust level (service level). It was determined that at the rear of the KL mining block, it is permissible to decrease the size of the intake and exhaust drifts, where cumulative air requirements decrease. Based on the distribution of open regulators, the rearmost 300m sections of exhaust and intake drifts in the KL footprint could be reduced from 7.0m x 7.0m to 5.5m x 5.5m or smaller, reducing overall mining capital costs without any significant detri­ mental effect on ventilation cost or delivery. The concept is shown in Figure 5. show­ ing the region of reduced airflow and potentially smaller exhaust and fresh air drift sizes (the red shading indicates higher concentration of airflows as block delivery and exhaust system accumulate airflow). The reduction of rear KL Block drifts sizes would only be recommended for the KL block if the mining footprint was not expanded further west, or if the mine was prepared to strip or mine additional drifts if the footprint did extend west in future.

31

Figure 5.

KL service level profile size optimization.

3 SIMULATION AND ANALYSIS RESULTS Figure 6 shows the results for selected scenarios. The base case diesel mine, utilizing existing airflow concepts and designs, is unsurprisingly the highest cost case, requiring at least two full intake and exhaust drifts and fans and higher power for the increased ventilation flow. The full BEV option delivers the lowest overall capital and operating cost, reducing primary infrastructure requirements and reducing power costs. A partial BEV and air reuse option deliver the next lowest capital and operating costs, while an air reuse option with diesel equip­ ment is the third lowest cost option.

Figure 6.

KL service level profile size optimization.

32

4 CONCLUSION No commitment to any specific mine ventilation design scenario could be made until further research is completed on BEV equipment options, and air reuse strategies. An intermediate KL Mine Development plan required for all scenarios is recommended for one additional 7.0m x 7.0m GVD intake, exhaust drift, primary fan and four 6.0m internal shafts each for the intake and exhaust connections. In summary, the intermediate mining and infrastructure plan will increase GVD exhaust capacity from 3750 m3/s to nearly 4500 m3/s. Mining an additional GVD exhaust and intake access of 7.0m x 7.0m linear development meters was recommended. While the optimum development size should technically be 6.8m x 9.0m high, a more rapid single-pass mining option of 7.0m x 7.0m is acceptable which gives near equivalent financial optimization at the loss of 100 m3/s of airflow. The addition of a 5MW primary mixed flow fan was recommended and added to the plan as well as four KL exhaust raises (360m each at 6.0m diameter) and four equivalent KL intake raises. The intermediate plan is required at a minimum for every conceivable scenario until the via­ bility of options of BEV or air reuse is resolved. Further work will be undertaken to validate BEV viability and air reuse viability. An additional intake and exhaust drift with a primary fan and internal raise will only be required if the reuse of infrastructure air or BEV use cannot be made viable. The full BEV option provides the lowest-cost ventilation option and, subject to other viabil­ ity analyses of BEVs, is recommended. Partial air re-use is viewed with some risk and contro­ versy, particularly if diesel equipment is continued to be used, as fine DPM is difficult to remove. Electric vehicles (BEVs) are seen as an inevitable implementation in underground mines given the demonstrable benefits in lower heat, reduced airflow requirement and the elimin­ ation of gas and DPM emissions. As the mining ventilation plan for scenarios 1, 3 and 4 are similar, the implementation of BEVs does not influence the intermediate mine ventilation infrastructure plan, only the later size and type an air-re-use facility. REFERENCES Brannon, C. Brard, D. Pascoe, N. Priantna, A. 2020, “Development and production update for the Gras­ berg Block Cave mine – PT Freeport Indonesia”, 8th International conference and exhibition on Mass Mining virtual conference Proc, pp 747 to 760, 9-11 December 2020, Santiago Chile. ISBN 978-95619-1196-3 McGuire, C, Witow D, Mayhew M, Bowness, K. 2022, “Comparison of heat, noise and ore handling capacity of battery-electric versus diesel LHD”, Australian Mine Ventilation Conference, Gold Coast, QLD, Australia, 10-12 October 2022. Stewart, C.M., 2022. “Ventilation considerations and modelling of lithium battery fires in underground mining.” Australian Mine Ventilation Conference, Gold Coast, QLD, Australia, 10-12 October 2022.

33

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

A case study in successful use of spot cooling for underground shaft sinking M. Brown, D.W. Durieux, C. McGuire & D. Witow Hatch

ABSTRACT: This paper presents engineering design and equipment selection for a successful temporary spot cooling installation to support underground shaft sinking that has been in oper­ ation for the last 14 months: September 2021 to November 2023. The need for cooling was driven by the depth of shaft sink (starting from >1,900m below surface), however this system was subject to many of the common challenges preventing the widespread use of spot cooling, including limited process water & dewatering capability, heat rejection equipment placement in the path of blasting fumes, limited airflow quantity for heat rejection, and layout constraints due to the existing and upcoming mine services installations and construction. Components were selected for mine-duty with consideration for the dusty environment. Use of hybrid cooling towers allow for increased heat rejection capacity from evaporative cooling while maintaining a fully closed-loop condenser water circuit. Skid-mounting of all components allowed for easy placement and relocation. Use of HDPE piping lashed to existing ground support allowed for maximum layout flexibility and minimized installation time. Performance, operational features, and additional lessons learned, including feedback from operations personnel, are shared.

1 INTRODUCTION This paper concerns a case study project which is currently in execution phase. This project required a small-scale underground “spot cooling” system deployment to support shaft sinking operations in the construction of a new winze. Mechanical refrigeration at this relatively small scale is widely understood and deployed in industrial surface applications; however, the ability to effectively reject heat within the confines of the underground workplace requires careful design due to a variety of factors including limited footprint and headroom availability, high air temper­ atures, low available air flow volumes, and dirty/dusty air conditions. Additionally, where these spot cooling installations are proposed for temporary applications, there is often a constraint on available electrical services and water supply/removal capabilities, and little desire to invest in reticulation systems to support the temporary facility. None of these challenges are, on their own, sufficient to disqualify mechanical refrigeration from spot cooling applications. Operational challenges can be attributed to multiple small issues concurrently as opposed to a single, broad failure in design. Instead, careful attention needs to be paid in the design and specification of equipment for underground service, to obtain fit-forpurpose equipment which maximizes the chance of success given the known constraints. This paper presents one such case study in which the selection of spot cooling equipment specifically for the challenging underground service has resulted in a robust and reliable operation. 2 APPLICATION DESCRIPTION The case study considers a project in the process of an expansion to a deep orebody which will be accessed by an internal winze shaft from approximately 1,200m to 2,600m below surface. The DOI: 10.1201/9781003429241-4

34

lower leg (1,900m to 2,600m) of the winze is being developed using conventional “blind” shaft sinking methodology with a Galloway stage and equipping stage progressing in parallel. Prior to starting development of the winze, there was no refrigeration capacity present at the mine. Due to the significant depth of the sinking activity, refrigeration is required to manage workplace temperature conditions and maintain safe, productive advance rates for the largely manual mining workplace. The mine will operate a large refrigeration plant for future mech­ anized development and production activities, but this plant will not be available until the completion of the sinking process. Therefore, a temporary interim solution was required, which was to be fit-for-purpose for the reduced ventilation demands of a shaft sinking applica­ tion as opposed to mechanized drill & blast with multiple crews and diesel vehicles. The appli­ cation is considered to be in-line with typical “spot cooling” applications in underground mining due to the small air flow rate and dedicated cooling of a single workplace, as compared to “bulk air cooling” of a mine’s entire intake air stream. A simplified schematic of the mine’s ventilation system is presented in Figure 1. Custom ductwork was installed to allow intake air from the downcast portion of the shaft, bypassing the heat from diesel mobile equipment operating in the service decline to support development and construction activities throughout the mine.

Figure 1.

Ventilation Schematic (Flows in m3/s).

The cooling system is located at 1,900m below surface at a mid-shaft station. During the plant’s period of operation, none of the mine’s permanent systems and infrastructure (e.g., process water supply, dewatering, and electrical distribution) are available at the plant loca­ tion. Therefore, minimizing the duty with which these services were impacted by the cooling plant was a key design consideration. 3 PROCESS CONDITIONS Ventilation design was undertaken by creating a detailed thermodynamic model in Ventsim software. Surface ambient conditions for the refrigeration design were selected based on the ASHRAE 2% humidification case from a nearby weather station. These surface conditions were used as an input for the Ventsim model, which determines the corresponding air tem­ perature at the intake of the spot cooling system. 35

The outlet temperature from the air-cooling plant was selected by the design engineers based on heat modelling, reflecting conditions to mitigate workplace heat stress risk. The heat stress index of choice at the site is wet bulb globe temperature (WBGT), and the maximum workplace temperature target is 28°C WBGT, based on the ACGIH limit for 100% work with a moderate work intensity. Process conditions for the system are summarized in Table 1. Table 1.

Summary of Spot Cooling Process Conditions.

Parameter

Units

Value

Air Flow (Volumetric) Mass Flow (Dry Air) Inlet Air Temperature Chilled Air Temperature Barometric Pressure Nominal Air Cooling Duty Air Temperature Available for Heat Rejection Maximum Air Flow Available for Heat Rejection

m3/s kg/s °C wb/db °C wb/db kPa kW(R) °C wb/db m3/s

12.0 15.9 28.4/36.0 12.0/12.0 121.3 740 30.0/36.8 98.0

One critical and uncommon characteristic of the ventilation system is the use of fiberglassreinforced polymer (FRP) construction for the sinking ventilation duct. This material has better thermal insulative properties than steel or flexible fabric duct, allowing for warmer chilled air tem­ perature at the cooling coil outlet for a given temperature at the discharge of the duct. This selection, in addition to equipment selections described in Section 4 below, was made with the primary object­ ive of minimizing electrical loading required to achieve a safe workplace for the shaft sinking crews. 4 EQUIPMENT SELECTION Equipment selection for the spot cooling system focused on ensuring robust performance at the required process conditions while attempting to minimize impacts to other mine systems (e.g., water and electrical) during the temporary installation period. A schematic of the final system is shown in Figure 2. The general configuration was largely dictated based on the overall process and mechanical con­ ditions. Specifically, the air-cooling requirements fell well within typical applications for indirect water-cooling coils (i.e., tube and fin heat exchangers). This equipment type also worked well to facilitate tie-ins to the fresh air ductwork supporting the mine development. Although the cooling coil is located in the mine’s fresh air stream, there is still anticipated to be dust present due to the nature of underground operations. Thus, the design was specified to include a wider-than-standard fin spacing of 7 fins per inch (vs. common surface applications as high as 12 fins per inch), which reduces risk of dust accumulation plugging the coils and reducing both airflow and heat transfer efficiency. 7 fins per inch was selected in consultation with the equipment supplier as a compromise between risk of fouling and the reduced heat transfer performance offered by the reduced fin quan­ tity. The increased spacing has the added benefit of reducing the coil air side pressure drop, thereby minimizing the fan power required for the temporary ventilation system. Water is transferred in a closed loop circuit between the air-cooling coil bank and a water-cooled chiller. The large physical space and air-flow requirements associated with an air-cooled chiller were deemed not feasible with an underground installation location. Additionally, presence of dust in underground workplaces presents significant operational and maintenance challenges with air-cooled chilling equipment. Frequent inspection and replacement of air side filters are required which can put strain on owner or contractor maintenance resources. Thus, this equipment type was not considered for the application. A screw compressor with dual compressor arrangement was selected to maximize compressor lift available in the event of challenging heat rejection conditions (i.e., high condenser temperature/pressure), and R-134a refrigerant is used due to its Class A1 (non-flammable and nontoxic) properties. 36

Figure 2.

Schematic of the Cooling System.

The last component of the spot cooling system is the heat rejection circuit. Design of this area was of critical importance in ensuring consistent, robust performance of the overall system. Design of heat rejection systems in underground applications requires careful consideration and can be challenging due to a variety of factors including high air temperatures, low available air flow vol­ umes, and dirty/dusty air conditions. For the spot cooling system, hybrid cooling towers were ultim­ ately selected for heat rejection. The towers operate through indirect heat exchange between the air and a closed-loop water coil connected to the chiller. They also utilize a small spray water stream on the coil exterior to facilitate latent heat transfer, and notably this evaporative cooling water source is completely independent of the chiller system (which prevents contamination of the conden­ ser cooling water circuit). Overall, this heat rejection configuration provided many benefits to the overall installation. First, it enabled maximum thermal performance without introducing potential fouling and water quality concerns with the chiller. The project did consider indirect heat transfer through the use of tube and fin heat exchangers similar to the air-cooling coil; however, this configur­ ation was eliminated. The lack of evaporative cooling and the high dry bulb temperature of the mine air stream available for heat rejection (only approximately 5°C cooler than the condenser 37

water stream) resulted in prohibitive quantities of ventilation air required to achieve the design heat rejection. Secondly, the use of indirect heat exchange with only an external spray design simplified the process water circuit by minimizing make-up and blow-down water requirements as compared to direct evaporative cooling. The range of acceptable water chemistry is also greater in the cooling towers, which are typical galvanized steel construction, in comparison to the special­ ized material present in the chiller heat exchangers which likely requires chemical dosing and water treatment consideration to prolong the life. This was an important consideration as the mine water systems are limited during temporary development phases, both for supply of fresh process water for make-up and for dewatering capacity to accommodate blow-down. Finally, the tower design attempts to minimize concerns with dirty/dusty air conditions and potential heat exchanger fouling (which will worsen heat transfer performance). Although the tower does contain a coil, it is designed to reduce air-side pressure drop while still maximizing area and heat transfer performance. The external spray water system also helps promote coil cleaning and removal of potential dust build-up on the heat exchange surface. It is also noted that the system was designed from a process and mechanical perspective to minimize electrical system impacts, again an important consideration due to electrical system constraints during early mine development periods. For example, reasonable airflow rates and velocities were targeted to minimize pressure drops and resulting electrical motor sizes. In add­ ition, the closed-loop water circuit temperatures were mindful of chiller lift considerations and enabled a coefficient-of-performance of slightly below 4 at maximum design conditions (impacting the compressor motor size and electrical consumption). Because this spot cooling system is directly coupled to the shaft sinking ventilation system, reliable operation is vital to prevent downtime on the project critical path. Improvements in reliability were achieved by installing piped-in standby pumps on both closed-loop water cir­ cuits. Additionally, both the cooling coils and the cooling towers consist of two units installed in parallel. While this does not achieve full redundancy, it removes the risk of a “single point of failure” in the cooling system causing a complete outage. Consultation with equipment sup­ pliers indicated that the system could operate (at reduced capacity) in the event of failure of any one compressor, cooling tower or cooling coil. Preventative maintenance and cleaning of the system was identified as a key enabler of reliable performance and would be performed on the system regularly via a service agreement with a local supplier. 5 INSTALLATION Another key aspect of the equipment design and selection was the ease of installation in an underground mining environment. The system was designed such that installation and connec­ tions could be achieved by tradespeople within the workforce of typical underground mining contractors, and not requiring any specific qualifications. Given the relatively small size and cap­ acity of the chiller, it was shipped to site fully assembled and charged with refrigerant, eliminating the need for any qualified refrigerant technicians to participate in the installation process. All pipe connections were Victaulic grooved, consistent with the mine’s standard for air, water and dewatering lines. Complex piping arrangements including hard-piped headers and tie points were fabricated in the manufacturer’s shop, minimizing the number of field connections required and ensuring that all field connections required identical couplings. An example is shown in Figure 3, in which four parallel inlets and four parallel outlets for the cooling coils are piped back to a single DN150 (6-inch) grooved inlet and single DN150 grooved discharge connection. Due to the mine layout and ventilation system, the cooling towers were installed approximately 200 metres from the chiller and cooling coils, adjacent to the return air raise. The main condenser water piping across this distance was installed using HDPE piping, which offered several advantages: • Pipes were sufficiently lightweight to be supported from existing ground support screen using chains (see Figure 4), eliminating the need to drill any rock anchors for pipe supports for a temporary application, 38

Figure 3.

Photo of shop-installed pipe headers on air-cooling coils.

• Thermal expansion and contraction can be easily accommodated by allowing the pipe to sag between subsequent supports, • Flexibility to adjust layout to work around existing infrastructure and avoid interferences, • Significant reduction in the number of couplings required, improving installation time. Flexible hose was selected for all water lines except the 200m run from the chiller to the cooling towers. This provided flexibility to field-locate equipment skids and adjust to condi­ tions in the underground workplace. Finally, in addition to local installation considerations, the system was specifically designed to be easily transported from surface to underground. The system is operating in a shaft

Figure 4.

Photos of HDPE Pipe Supported by Chain from Mine Ground Support Screen.

39

access mine, and therefore all components were selected and manufactured to be shipped within cage size constraints. Of all the equipment included in the system, only the cooling towers needed to be broken down for transport due to their overall height of 3.9m being unsuitable for ramp transport on flatbed trucks available at the mine. A photo of the cooling tower split in preparation for transport is shown in Figure 5.

Figure 5.

Photo of Cooling Towers in Preparation for Transport.

Custom skids were engineered for each piece of equipment, which allowed for forklift trans­ port and creating a stable base that allowed the equipment to be installed on compacted mine roadway ballast. Skids with fork pockets are visible in Figure 3 and Figure 5. The level of flexibility was so pronounced that the complete chiller, cooling coil and evaporator water system was installed in a temporary location to provide interim cooling in the summer months, and then relocated to the final location only three months later. 6 PERFORMANCE The temporary refrigeration system has been in service for 14 months as of the writing of this paper in October, 2022. The system has been operating reliably and has yet to experience any unplanned downtime. The only unplanned maintenance activity to date resulted from a failed seal on the operating condenser water pump; however, the standby pump was switched into service and the repair was made without any appreciable downtime to the plant. A ventilation fan upgrade in 2022 has increased the total airflow delivery to the sinking crews from 12 m3/s to approximately 20 m3/s since the start of operation. Periodic measure­ ments of the air side differential pressure have seen a negligible change over time, indicating little dust build-up on the coils. The housing was opened after 12 months of service to clean the coils, and no notable dust was observed. Differential pressure remained unchanged after cleaning. A photo of the air-cooling equipment is presented in Figure 6. Cooling tower performance has been stable throughout the operation. Dust filters on the air intakes are cleaned periodically by spraying down with process water - a cleaning process which can be done without shutdown of the units. The self-cleaning characteristics of the 40

Figure 6.

Photo of sinking ventilation fan and air-cooling coil in service.

internal coils with water sprays have maintained consistent operation despite the dusty envir­ onment. A photo of the cooling towers in service is shown in Figure 7. It can be seen that the intakes on the right-hand unit have recently been cleaned at the time of the photo. Operator feedback has been positive, noting that the complete system is operated in a hands-off manner. The chiller automatically modulates duty by adjusting screw compressor speed using its on-board variable speed drive (VSD), and all other components operate con­ tinuously without need for any modulation or intervention. System performance has been measured to align with the engineering design values for the conditions to date. As sinking progresses and ambient temperatures increase, the system per­ formance at the design maximum values will be measured and validated.

Figure 7.

Photo of cooling towers in service with recently cleaned air intakes on the right hand unit.

41

Computational fluid dynamics applications in mine ventilation

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

CFD study of cavern ventilation M.A. Carvajal-Meza & J.P. Hurtado-Cruz Universidad de Santiago de Chile, Santiago, Chile

ABSTRACT: The ventilation of mines and underground works has been studied for more than 100 years, with extensive and rigorous results that currently aid in safely performing underground operations of different kinds and characteristics. However, the study of under­ ground ventilation in large volumes such as caverns has seen relatively little open study, being part of engineering studies without further disclosure. In general terms, it is the final use of the cavern that determines the ventilation system to be used. In this regard, the present study is generated from the need to define elementary ventilation systems for underground caverns linked to the construction system, which allows for maximum pollutant drag with minimum flow rate and reducing recirculating secondary flows. For this, Computational Fluid Mechan­ ics is used as a simulation and analysis tool through the Ansys Fluent software, generating various simulation scenarios of steady and transient flows. The results show different ways of ventilating according to the established boundary conditions and the construction geometries used, which could help to orientate future ventilation designs from an academic exercise.

1 INTRODUCTION 1.1 Cavern design A cavern can be defined as a cavity excavated underground, usually of large dimensions, greater than the dimensions of the tunnels from which it is accessed. When considering the depth at which the caverns can be found, it is possible to detach 3 groups: the first includes caverns on the surface and in depths no greater than 3 times their diameter (Hoek, et al., 2002); the intermediate group covers depths less than 100 meters and greater than 3 times its diameter; and the third group, the deep ones, considers caverns located at depths greater than 100 meters (Sharma & Judd, 1991). Caverns, like all underground works, have different types of shapes and are generated from a series of factors, requirements and parameters, combining to determine the dimensions of the section to be excavated. The functionality and the desired use for the work must be taken into account as well. Chile provides a good example, where buildings tend to have a vaulted and arched roof to better support stresses given existing seismic and tectonic conditions (the shape used in this study). Caverns can be classified according to the type of use that is given to them, such as hydroelectric caverns, oil, hydrocarbon and gas storage, mining industry, for laboratories, and other civil and/or military uses. Specifically, in mining, caverns are one of the most important underground construction works. Their applications cover crushing, grinding, flotation rooms, maintenance workshops for loading and transport equipment, and storage for explosives and/or dangerous chemical elements, among others. These factors make it important that the infrastructure generated is economical, efficiently ventilated, and, above all, safe. Cavern construction design is generally based on two access tunnels that arrive at the upper and lower ends (either on the same face or in the opposite faces –this last is more used-), making a recess from the upper part, and taking advantage of gravity to remove muck (see DOI: 10.1201/9781003429241-5

45

Figure 1). This also aids with fortification as the walls are lowered, leaving the cavern ceiling and walls reinforced. Also, can be find three tunnels, one arrives at the upper and two at the lower opposite ends.

Figure 1. Examples of mining caverns of Codelco: (a) Crusher cavern of Dacita, El Teniente (b) Crusher cavern of Chuquicamata Underground; (c) Crusher cavern of Diablo Regimiento, El Teniente.

1.2 CFD and modelling studies The focus of this work is based on the study of cavern ventilation, an area that has not been openly studied in detail over the years; models that can be replicated or that are capable of estab­ lishing air flow behavior inside the cavern are limited. From this, to develop this work, analogous cases from other areas have been analyzed, with models of similar physical behavior, studying ventilation in closed or semi-closed spaces on a smaller scale. Thus, taking advantage of dimen­ sionlessness at different scales of turbulent phenomena, the behavior of air is similar. Suspended dust particle size is also very similar in all environments, even comparable to the size of small vir­ uses (0.3 μm). Among the studies analyzed, some of the most pertinent considered surgical wards in hospitals (Anuraghava et al., 2021), subway stations (Guan et al., 2007) and mining ventilation systems (Chang et al., 2019) (Zhou et al., 2020). Based on these works, it is considered a common factor that all study models require a forced air injection and/or extraction system in closed spaces. The paper “CFD modeling of ventilation and dust flow behavior above an underground bin and the design of an innovative dust mitigation system” (Ren et al., 2014) establishes a CFD analysis of the behavior of an air-injection ventilation system in coal mining. The K-Epsilon tur­ bulence model is also used together with the transport of particles in a discrete phase model (DPM) and with a geometry similar to that of an underground cavern, where the dimensions of the tunnel under study were considerably greater than those seen in other studies and helped establish the behavior of dust particles within this geometry. Various medical studies on COVID-19 prevention were considered for this work as well, includ­ ing effective ventilation in supermarkets (Mohamadi and Fazeli, 2022), classrooms (Jeong et al., 2022), isolation rooms (Bhattacharyya et al., 2020), and hospitals (Bayatian et al., 2021, Obeidat et al., 2021). The paper “CFD modeling of airborne virus diffusion characteristics in a negative pressure room with mixed mode ventilation” (Anuraghava et al., 2021) shows the modeling car­ ried out for the ventilation of COVID-19 patients’ isolation rooms, using mixed ventilation sys­ tems for this purpose. The importance of this study lies in the fact that it establishes a mixed cross ventilation system using a K-Epsilon turbulence model, together with a discrete phase particle transport model with geometry similar to that of an underground cavern. These elements help provide the bases to generate the mixed models that are applied. All these works have been consulted in order to observe ventilation systems with 1 or 2 air out­ lets with different dimensions. This combination helps establish a convergence between them all in the use of a turbulence model for particle transport, but especially in the meshing and spatial dis­ tribution, thereby validating similarities with the case of mining caverns and allowing the replicat­ ing of said models. 46

2 GEOMETRY DESIGN 2.1 Geometry design First, we must establish the layout of the geometry to be simulated and its dimensions. For this study, as the most of the caverns, maintain a parallelogram shape, the length being the longest side, with extensions close to 50 meters in the case of mining (Villablanca, 2014). However, these can reach lengths of up to 900 meters, as is the case of the Jinzhou project oil storage cavern (Zhang et al., 2018). On the other hand, the height and width are limited depending on the geome­ chanical characteristics of the rock in the project. For this particular case, the study of cavern model is based on two important works carried out in Chile, the Diablo Regimiento crushing cavern and the Phase I grinding cavern of Andina, where it is assumed that in the upper part the cavern has arc shape in order to better support the geomechanical conditions (Villablanca, 2014). A form factor (f) was also established that uses the orthogonal dimensions of the cavern, coming from rock fragmentation studies. Dimensionless factor f obeys a criterion used in mining in order to provide rock fragments with a size factor, and to be able to compare them with each other. The applied formula is shown in Equation 1. The factor is in the range 0 to 1, where if f is near 1 all dimensions are the same (a cube), while if f is near 0 at least one dimension is very stretch. In this work is applied to visualize how stretch or similar a cube is the cavern geometry.

This provides the results of factors shown in Table 1, from which measures with a factor equal to the average of both (0.31) were chosen for this study. The value of 0.31 converges for other case studies carried out, mainly those in health and social areas, including operating rooms (Anuraghava et al., 2021), supermarkets (Mohamadi and Fazeli, 2022) and schools (Jeong et al., 2022). Therefore, the dimensions chosen to carry out the base design according to the geometry of the cavern are 30 m high, 21 m wide, 45 m long, and a size factor of 0.31. Table 1. Cavern dimensions. Study Case

Height (m)

Width (m)

Length (m)

Shape Factor (f)

Grind Cavern Phase I Andina Crushing Cavern Diablo Regimiento Average Value Applied case

37 22 29.5 30

28 13 20.5 21

58 31 44.5 45

0.31 0.30 0.31 0.31

2.2 Arrangement of air inlet and outlet In many cases, construction accesses to the cavern are taken into account to generate ventila­ tion designs. As mentioned, the construction of the cavern generally begins by connecting an upper tunnel with a lower tunnel by means of a raise in order to carry out the developments by recess. Therefore, only one air inlet is considered, and 1 or 2 outlets depending on the con­ struction design or subsequent use. Inlet and outlet drifts have square section 4x4m. All designs were devised to have a direct connection between the air inlet and outlet, in order to avoid any significant air vortices, stagnant air, or dead air spaces (Figure 2). – Design 1: The air inlet is through an injection tunnel that is located in the upper center of the cavern, while the outlet is on the opposite wall, lower right sector, and consists of an extraction tunnel of the stale air found inside the cavern. – Design 2: The air inlet is through an injection tunnel located in the upper left sector of the cavern, while the outlet is on the opposite wall, lower right sector and consists of a tunnel that is used to extract stale air found inside the cavern. – Design 3: The air inlet is through an injection tunnel situated in the upper center of the cavern, while the outlets (2) are located in the lower right and left sectors of the opposite wall, and consist of two galleries used to extract the stale air that is inside the cavern. 47

Special design: a design that considers the air inlet and outlet from the same face or wall, for cases where there is no possibility of creating ventilation paths on the opposite wall. This is considered in a separate analysis because, constructively, the excavation would only be from below, when access to the cavern construction is limited (for example, disassembling a TBM in a dead end).

Figure 2.

Isometric view of designs 1 (left), 2 (center) and 3 (right).

3 CFD SETUP 3.1 Mesh, initial and boundary condition The mesh is a structured hexagonal domain in the shape of regular parallelepipeds. A mesh size analysis is performed, establishing two critical sizes with lengths of 20 and 25 cm, for which a final size of 20 cm is established, generating 1.8 million cells (Table 2). Table 2.

Cells number by design.

Case Design

Cell Number

1 2 3

1,852,108 1,892,369 1,882,534

Boundary conditions established are velocity inlet (upper tunnel as inlet) and pressure outlet (lower tunnel), as shown in Table 3. The same table includes initial conditions, velocities, and time steps used in the unsteady cases. Time independence tests were made to validate conver­ gence, based on Courant number less than 1. Unsteady cases were made to test flow instabilities and some DPM cases to evaluate dust residence time, but most cases were a steady simulation. Table 3. Boundary and initial conditions. Condition

Value

Unit

Air Density Gravitational acceleration Inlet velocity Inlet Temperature Pressure Outlet (gauge) Wall Temperature Time step for 0.5 m/s Time step for 2.5 m/s Time step for 5.0 m/s Time step for 10 m/s

1.23 -9.8 0.5 – 2.5 – 5 – 10 277 0 303 0.4 0.08 0.04 0.02

kg/m3 m/s2 m/s K Pa K s s s s

48

3.2 Turbulence model In order to choose the turbulence model to be used, the researchers’ experience was considered and an exhaustive bibliographical analysis was carried out, searching in various areas such as transportation, energy, mining, health, and others (17 studies overall). Additionally, for this work it not needed heat transfer. So that, the chosen turbulence model is k-epsilon Realizable, with second order upwind spatial discretization and Simplec pressure-velocity coupling scheme. 3.3 Order of design-cases for results The designs used have the distribution presented in Table 4 in the inlet and outlet positions of the airflow. Designs 1 and 2 take the sides specular response into account, for one side to another or in an inverse way, as the same behavior. In turn, the cases are classified according to flow speed (see Table 5). Some authors (Vutukuri & Lama, 1986) suggest a minimum air velocity of 0.5 m/s is needed to manage dust and consider that a maximum of air velocity of 2.5 m/s is required to minimize the dust concentration in the air. Additionally, the Chilean law establish 2.5 m/s as the maximum allowed in working zones in Chile. 5 m/s is the air speed recommended to conveyor drifts (6 m/s for main haulage routes) (McPherson, 2009). Finally, 10 m/s is the speed limit fixed by authors because is a exceed velocity which exceeds more than double human comfortable limits and is near to economical airflow limit. Table 4. Designs and spatial distribution. Design

Inlet

Outlet

Design 1 Design 2 Design 3 Special Design

Center Left Center Center

Left Right Left-Right Center

Table 5.

Cases and airflow.

Case

Air Velocity (m/s)

Airflow (m3/s)

1 2 3 4

0.5 2.5 5.0 10

8 40 80 160

4 RESULTS 4.1 Designs-case results First, an analysis of the airflow inside the cavern was carried out considering a minimum vel­ ocity limit equal to 0.004 m/s, which is the turbulent flow limit (Re≥4000), in order to avoid gas/dust stratifications and clear away dust. In this way, the effective circulation inside the cavern can be classified according to the percentage of sweeping or dragging of contaminants showed in the Table 6. Figure 3 shows the Streamlines and Isosurface, specifically cases 1 and 4 (due to the extension of the work no presented all analyzed cases). The ranges of color cor­ responds to zero as minimum velocity (blue color) and the maximum velocity (red color) of each case 0.5 and 10 m/s (cases 1 and 4), which emphasizing the differences. Isosurface dia­ grams represents the swept spaces wit turbulent flow (Re≥4000), and the empty volume repre­ sents the unswept spaces (Re29.4°C (85°F)

Derating unnecessary 0%–10% 0%–40% 25%–75% 50%–90% 90%–100% Cooling/dehumidification required

5 CONCLUSIONS Although it would be impractical to conduct testing for all possible mine temperatures and mine strata compositions, our analysis results can be used as a rough guide for planning pur­ poses. In order to ensure compliance with the 35°C (95°F) AT limit, testing or simulations using validated thermal simulation models should be used to estimate the final AT, which pri­ marily depends on initial mine temperature and mine strata composition. In the absence of cooling or dehumidification, our analysis indicates that occupancy derat­ ing of RAs may be required for some mines. For mines below 18.3°C (65°F) with very low mine strata thermal conductivity, the occupancy may need to be slightly reduced. For mines between 18.3°C–21.1°C (65°F–70°F), occupancy derating would be necessary for mines with strata having low thermal conductivity. From 21.1°C–26.7°C (70°F–80°F), the results indicate that occupancy derating would be necessary. Above (80°F), occupancy derating would approach 100%, and cooling or dehumidification would be needed to ensure that the thermal conditions inside an RA stay below the 35°C (95°F) AT limit. 565

6 LIMITATIONS Because the analysis presented here was performed on only two RAs with conditions from only five mines across the U.S. and one nonproduction mine, the results should not be assumed to directly apply to mine-specific RA installations. Additional testing or analysis would be neces­ sary to determine occupancy derating percentages for a specific RA installation. REFERENCES Bernard, T., Yantek, D. & Thimons, E. 2018. Estimation of metabolic heat input for refuge alternative thermal testing and simulation. Mining Engineering, 70, 50. Bernard, T. E. 2011. Physiological analysis of human generated heat in a refuge alternative. NIOSH Con­ tract Report 254-2011-M-40932. Tampa, FL, University of South Florida. Bissert, P., Yantek, D., Klein, M. & Yan, L. 2016. Analysis of heat loss mechanisms for mobile tent-type refuge alternatives. Transactions of Society for Mining, Metallurgy, and Exploration, Inc, 340, 70. CFR. 2019. Code of Federal Regulations, 30 CFR 75.1506, Refuge alternatives. Washington, DC, U.S. Government Publishing Office. https://www.govinfo.gov/content/pkg/CFR-2019-title30-vol1/pdf/ CFR-2019-title30-vol1-sec75-1506.pdf accessed on 2/24/2023. Federal Register. 2008. Refuge Alternatives for Underground Coal Mines; Final Rule, Department of Labor, Mine Safety and Health Administration, 30 CFR Parts 7 and 75. Washington, DC, US Gov­ ernment Printing Office. Klein, M. 2017a. Mine shelter thermal analysis - 23-person inflatable RA TAITherm model validation and analysis. Report 3018–001, Revision 4.1. ThermoAnalytics, Inc., 23440 Airpark Blvd., Calumet, MI. Klein, M. 2017b. Mine shelter thermal analysis - steel RA TAITherm model validation and analysis. Report 3090–001, Revi-sion 4.1. ThermoAnalytics, Inc., 23440 Airpark Blvd., Calumet, MI. Klein, M. & Hepokoski, M. 2017. Human thermoregulation model for analyzing the performance of mine refuge alternatives. Mining Engineering, 69. Klein, M., Yantek, D., Hepokoski, M. & Yan, L. 2017. Prediction of human core temperature rise and moisture loss in refuge alternatives for underground coal mines. Transactions of Society for Mining, Metallurgy, and Exploration, Inc, 342, 29. OSHA. 2003. Permit-required confined spaces. Department of Labor, Occupational Safety and Health Administration, https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.146 accessed on 2/24/2023. Rothfusz, L. P. 1990. The heat index equation (or, more than you ever wanted to know about heat index). Fort Worth, Texas: National Oceanic and Atmospheric Administration, National Weather Ser­ vice, Office of Meteorology, SR 90–23. https://www.weather.gov/media/ffc/ta_htindx.PDF accessed on 2/24/2023 Shumaker, W. A. 2013. Information relayed from Wesley Shumaker, Mechanical Engineer, MSHA Approval and Certification Center, to NIOSH Pittsburgh Mining Research Division regarding the heat used to represent a refuge alternative CO2 scrubber system during heat and humidity testing. U.S. Department of Labor, Mine Safety and Health Administration, Approval and Certification Center. Triadelphia, WV. Steadman, R. G. 1979. The assessment of sultriness. Part I: A temperature-humidity index based on human physiology and clothing science. Journal of Applied Meteorology and Climatology, 18, 861–873. Yan, L. & Yantek, D. 2018. Portable refuge alternatives temperature and humidity tests. Mining engin­ eering, 70, 43. Yan, L., Yantek, D., Klein, M. & Bissert, P. 2016a. Interior thermal environment of a 6-person metaltype refuge alternative (RA). ASME International Mechanical Engineering Congress and Exposition. Phoenix, Arizona, USA, American Society of Mechanical Engineers. Yan, L., Yantek, D., Klein, M., Bissert, P. & Matetic, R. 2016b. Validation of temperature and humidity thermal model of 23-person tent-type refuge alternative. Mining engineering, 68, 97. Yan, L., Yantek, D. & Reyes, M. 2020. Underground mine air and strata temperature change due to the use of refuge alternatives. Mining, Metallurgy & Exploration, 37, 773–781. Yantek, D., Yan, L., Bissert, P. & Klein, M. 2017. Effects of mine strata thermal behavior and mine ini­ tial temperatures on mobile refuge alternative temperature. Mining engineering, 69, 41.

566

Renewable / alternative energy in mine ventilation

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Incorporating droplet dynamics to improve the reduced-order model of spray freezing for mine heating applications M. Mohit McGill University, Montreal, Quebec, Canada

S. Akhtar Polytechnique Montreal, Montreal, Quebec, Canada

M. Xu & A.P. Sasmito McGill University, Montreal, Quebec, Canada

ABSTRACT: Spray freezing technology has been shown its exceptional efficiency, safety, and sustainability for underground mine heating. Designing spray freezing systems requires a mathematical model with the rigorous formulation and fast computation methods, capable of predicting performance indicators. Existing models for spray freezing often take the droplet motion and velocity distribution as a priori, thus making it less feasible in practice. In the present work, a novel reduced-order model is developed to calculate the distributions of droplet velocity and residence time for various spray configurations (namely, flat fan, hollow cone, and full cone) and droplet diameters. The velocity distributions are then incorporated into a robust heat transfer model for mine heating to improve the prediction of the droplets freezing time and overall heat transfer rate (HTR). Consequently, the residence time distribution is used along with droplets diameter distribution to calculate the ice packing factor (IPF) and cooling capacity of the system.

1 INTRODUCTION In recent years, studying different clean technologies for providing the heating demand of underground mines has attracted researchers’ attention due to the importance of employing eco-friendly systems in mining industries. Spray freezing technology as a renewable and effi­ cient method for air heating in under-ground mines of sub-arctic regions is one of these tech­ nologies. In this method, the sub-arctic air stream is heated by using the water latent heat of solidification. An example of spray freezing application for underground mine heating has been introduced by Stachulak (1991). Through spraying water into the cold air stream, water droplets freeze, and the air temperature increases. The schematic of the spray freezing setup is illustrated in Figure 1 (a). To efficiently evaluate the performance of spray freezing setups, it is necessary to develop mathematical models with high accuracy and low computational cost. The two important aspects of spray freezing modeling are the droplet freezing and droplet dynamics. Droplet freezing process has been the topic of study by many scholars due to its importance from both the theoretical and practical point of view. The solid-liquid phase change phenomenon of water droplets is categorized under the subject of Stefan problem by many scholars. Various numerical, analytical, asymptotic, and hybrid methods have been developed in the literature to solve the Stefan problem in different conditions and coordinates (Rubinšteĭn, 2000; Meirmanov, 2011). One powerful method to investigate Stefan problems subject to various boundary conditions is the perturbation series analysis which has been used in many works (Caldwell and Kwan, 2003; Akhtar, Xu and Sasmito, 2021b). It should be DOI: 10.1201/9781003429241-58

569

noted that most results on droplet solidification modeling do not take nucleation and den­ dritic growth stages into account. This shortcoming has been addressed by Akhtar, Xu and Sasmito (2021a) by developing a reduced-order semi-analytical model for droplet solidifica­ tion that can capture all freezing stages. On the other hand, several researchers have studied droplets motion in different conditions. Analyzing droplets motion and dynamics is essential to calculate spray characteristics in various fields like spray cooling, spray freezing, and pesti­ cide spraying. The equation of motion for droplets, bubbles, and solid particles has been dis­ cussed by Michaelides (1997). This equation of motion can be numerically solved for droplets to find their trajectory and velocity (Deshpande, Gao and Trujillo, 2011; Zhang, Zhang and Bai, 2022). However, droplets dynamic has not been considered in the literature of sprayfreezing modeling to the best of authors’ knowledge. Furthermore, to fully capture the drop­ lets dynamic, it is necessary to take the spray shape into account. This issue has not been addressed in the literature as well. Akhtar, Xu and Sasmito (2021) studied the thermal characteristics of the spray freezing setup by developing a reduced-order theoretical-statistical model. However, the effects of droplets motion are not considered in that study. In the present work, a reduced-order model for droplets motion has been combined with the droplet freezing model introduced by Akhtar, Xu and Sas­ mito (2021a) to accurately capture system-scale characteristics of spray freezing setup. Different spray shapes have been studied and the droplets velocity distribution and residence times have been calculated by numerical approaches. After that, the results have been coupled with the drop­ let freezing model to calculate heat transfer coefficient (HTC), HTR, IPF, and cooling capacity. 2 MATHEMATICAL MODELING The mathematical modeling of the spraying freezing system can be divided into three parts, namely, a) Droplet freezing modeling; b) Droplet dynamics modeling; c) Spray modeling. The first two parts are treated as particle-scale processes, while the third one can be seen as a macro-scale process. The following assumptions are made in the present work: i) Droplets are spherical with no deformation and there is no change in the droplet volume during solidification; ii) Droplets break-up and coalescence are neglected; iii) The thermo-physical properties are assumed to be phase-dependent in the droplet freezing model (Sections 2.1.1 and 2.1.2); iv) When a droplet collides with the system boundaries, it does not contribute to the heat transfer anymore. In the following subsections, each modeling part will be discussed. 2.1 Particle-scale parts 2.1.1 Droplet freezing model For droplets solidification process, the model introduced by Akhtar, Xu and Sasmito (2021a) is used in the present work. As discussed by Akhtar, Xu and Sasmito (2021a), the droplet freezing process consists of five different stages, namely, water super-cooling, nucleation, dendritic growth, equilibrium freezing, and solid sub-cooling. Figure 1 (b) illustrates different stages of droplet freezing. The mathematical representation of each stage has been given in the work of Akhtar, Xu and Sasmito (2021a). A similar approach is employed in the present work to calcu­ late the droplet thermal characteristics and freezing time. The following correlation holds between the dimensionless time and the interface position (Akhtar, Xu and Sasmito, 2021a):

where: 570

The dimensionless variables are defined as:

where ri and rI,i are the droplet and interface radiuses, respectively, t is time, and αs is the ice thermal diffusivity. Ste and Bi are the Stefan and Biot numbers, respectively, as defined below:

Equations (1)-(3) will be used to calculate the interface position of each droplet in the spray freezing process. The schematic of the solid-liquid interface motion is shown in Figure 3 (c).

Figure 1. (a) Spray freezing setup (b) Droplet freezing curve (c) Solid-liquid interface motion for a droplet.

2.1.2 Droplet dynamics model The droplets motion is modeled numerically by solving the equation of motion on a Lagrangian mesh. The position and velocity of a droplet are correlated as below:

where xd and ud are the droplet position and velocity vector, respectively. The drag force and gravity are the two significant forces acting on a droplet and other terms in the equation of

571

motion can be neglected (Zhang, Zhang and Bai, 2022). Writing the second Newton’s law for a single droplet gives:

which can be simplified as:

where:

In above equations, md, ρd, rd, and Dd are the droplet mass, density, radius, and diameter, respectively, ug, ρg, μg are the gas velocity vector, density, and dynamic viscosity, respectively, g is the gravitational acceleration vector, and Red is the Reynolds number according to the drop­ let-gas relative velocity. CD is the drag coefficient which can be calculated for different values of Red as below (Zhang, Zhang and Bai, 2022):

From (5) to (8), the droplets position and velocity can be calculated over time using numer­ ical approaches. Furthermore, the residence time of each droplet, i.e., the total time that each droplet remains within the system domain, can be calculated by comparing its position with the system boundaries. 2.1.3 Effects of droplet motion on HTCs While the models proposed in Sections 2.1.1 and 2.1.2 are solved separately, in the present work to improve the computational efficiency, the effects of the droplet motion on the freezing model is taken into account. In fact, droplets motion affects the calculation of the overall HTC. As discussed by Akhtar, Xu and Sasmito (2021a), the overall HTC is calculated by con­ sidering the effects of forced convection, radiation, sublimation, and evaporation:

where indices eq, c, r, s, and e are referring to terms “equivalent”, “convection”, “radiation”, “sublimation”, and “evaporation”, respectively. Coefficient hr is not affected by the droplets velocity because it is not a function of Reynolds number (Red), but hc, hs, and he are func­ tions of Red. In previous works (Akhtar, Xu and Sasmito, 2021; Akhtar, Xu and Sasmito, 2021a), since droplets velocity is neglected, Red is calculated based on the air velocity:

In the present work, Reynolds number is calculated as given below:

This value will be used in the calculation of hc, hs, and he instead of Reold d . It should be noted that ud is obtained from the droplets dynamic model given in Section 2.1.2 and is 572

a function of droplet diameter, droplet initial direction, and spray shape. In the following part, a novel approach for considering the spray shape and droplets initial velocity direction in the droplet’s velocity calculation will be discussed. 2.2 System-scale part: Spray modeling A spray consists of a mass of droplets with different diameters and velocities. Several distribu­ tions have been used in the literature to represent sprays diameter distribution like normal and Rosin-Rammler distributions. In the present work, it is assumed that the flat spray shape is produced by a fan spray and hollow cone and full cone shapes are produced by a pressure swirl atomizer. The experimental correlations reported by Liu (1999) are used to calculate the Sauter mean diameter (SMD) of these sprays and a Rosin-Rammler distribution is then util­ ized to find the droplet diameter distribution. The distribution is discretized by considering n different values for droplet diameter where n is set to 20 in the present work. The probability distribution of diameters for each spray shape is illustrated in Figure 2 (a) and (b). To calculate droplets velocities over time, it is necessary to model their initial direction. Here, it is assumed that all droplets have an equal initial velocity magnitude, but their initial directions are different. The droplets initial velocity directions are different for each spray shape. In the present work, the droplets different initial directions are modeled by assuming a finite number of paths for droplets in each spray shape as depicted in Figure 3. In this figure, each black circle indicates a droplet path which can be seen as a straight line that connects the spray position to that circle. Some of the droplet paths are shown with dotted lines in Figure 3. The x, y, and z-components of the initial droplets’ velocity can be calculated for each droplet path in each spray shape using the spray angle θ which is a known variable based on spray characteristics, and the number of paths. To incorporate droplet diameter distribution in this framework, it is assumed that the distributions in all droplet paths are equal. The residence times and velocities in different paths are then averaged to find a mean value for each droplet diameter. 2.3 Computational framework and other considerations The presented mathematical model is computed by Python programming language using a hybrid method (both analytical and numerical approaches). IPF and HTR are considered as large-scale indicators of the setup performance. The IPF is defined as the ratio of the frozen water over the overall mass of sprayed water:

_ s is the mass flow rate of the ice formed during the operation and m _ t is the total mass where m flow rate of sprayed water. To calculate IPF, the frozen portion of each droplet should be calculated and then integrated for all droplets:

in which f(Di) is the probability of a droplet with diameter Di to exist, N_ is the particle injec­ tion rate of the spray, and ms,i is the mass of the frozen portion of a droplet with diameter Di. N_ can be calculated as:

where mi is the mass of a droplet with diameter Di. For a droplet with diameter Di, variable ms,i in (14) is calculated by finding the interface position based on the droplet resi­ dence time (τr,i) and freezing time (τf,i) using equations (1)-(3). It should be noted that the 573

volume change occurring during the water freezing process is neglected in the droplet freezing model developed by Akhtar, Xu and Sasmito (2021a). This will result in the mass conservation violation due to the phase-dependency of density which yields a considerable error in IPF cal­ culation. Hence, a correction is applied to IPF calculation by using the mass conservation equation. The total mass of water is conserved during the process, hence:

where ml.i is the volume of unfrozen water of the droplet with diameter Di and Δm is the mass difference due to neglecting the droplet diameter movement. Δm can be added to the numer­ ator of the IPF correlation (14) to compensate for the effect of droplet volume increase during the freezing. Therefore:

The HTR is calculated as:

in which ei is the total heat emission of a droplet with diameter Di during solidification and can be calculated as:

Figure 2. Probability distributions (a) Flat spray (Liu, 1999) (b) Cone sprays (Liu, 1999) (c) Experimen­ tal data from Santangelo (2010).

where hst,i is the HTC of each freezing stage, Ti is the temperature profile over time, T∞ is the air temperature, and Ai is the droplet surface area. 3 MODEL VALIDATION To validate the model, the experimental data from the frood-stobie mine reported by Stachulak (1991) has been used. The droplet distribution in this part is adopted from the experimental results proposed by Santangelo (2010) which suitably mimics the situation in frood-stobie mine. This distribution is depicted in Figure 2 (c). As discussed by Stachulak (1991), the total amount of ice produced during the winter operation of the frood-stobie mine is approximately mice,exp =145.000 tons. Using the developed model in the present work with the distribution shown in 574

Figure 2 (c), the IPF will be equal to 92%. Considering 1,200 operating hours of two stopes during the winter and an average water flow rate of 18.9 kg/s as reported by Stachulak (1991) and Trapani and Chen (2017), the total amount of ice formed during winter can be calculated using the model output as below:

The relative error is equal to:

Figure 3.

Modeling droplet initial directions (a) Flat spray (b) Hollow cone spray (c) Full cone spray.

4 RESULTS AND DISCUSSION The developed model has been run using the specifications given by Stachulak (1991) for the frood-stobie mine. The spray shapes discussed in Section 2.2 along with their corresponding diameter distributions have been considered in running the code one by one. For each spray, different parameters have been calculated and averaged using the given distribution. The results are summarized in Table 1. As can be seen from this table, the values of IPF and HTR are greater for hollow cone and full cone sprays compared to flat spray. This difference is due to the different diameter distributions or SMDs of the sprays as reflected in Figure 2. While larger droplets require more time for freezing, their residence time is smaller compared to finer droplets due to the larger gravity force imposed on them. Hence, the ratio of the frozen volume over the droplet volume is smaller for droplets of larger diameter which in turn results in a smaller IPF for sprays with larger SMDs. Comparing the results for hollow cone and full cone spray shapes shows that although the spray shapes and droplet trajectories are different, but the average values listed in Table 1 are quite close. The reason is that the diameter distribution is the same for both cases as they are produced by a specific type of nozzle.

575

Table 1. Average values for different spray shapes. Spray shape Parameter and unit

Symbol

Flat spray

Hollow cone

Full cone

Sauter mean diameter (μm) Average residence time (s) Average freezing time (s) Average nucleation time (s) Average relative velocity (m/s) Average interface radius (μm) Heat transfer rate (MW) Ice packing factor (%) Cooling capacity (GJ)

SMD �τR �τF �τN �vrel �rI HTR IPF Cc

822.83 7.34 19.19 8.79 4.52 540.54 1.98 30.82 8,404.7

258.49 7.46 4.81 2.17 1.72 44.30 5.56 94.11 25,664.2

258.49 7.46 4.80 2.17 1.74 44.24 5.58 94.14 25,672.4

The ice produced during the cold days of winter can be stored and utilized in summer for cooling purposes. The cooling capacity (Cc) of a spray-freezing setup is calculated by multiply­ ing the rate of ice production by the water latent heat of fusion and the total operating time:

where top is the total operating time. As can be inferred from (21), a larger IPF will result in a greater cooling capacity, hence using cone-sprays will provide more cooling capacity during the summer operations. 5 CONCLUSION In the literature of spray freezing modeling, there is a research gap in bridging between the models of small-scale droplet freezing phenomena and system-scale spray characteristics. The present work aimed to address this shortcoming by developing a reduced-order model for cal­ culating droplets velocity and residence times and incorporating this model into the existing droplet-scale models of freezing phenomena and experimental correlations of spray character­ istics. For modeling the droplets dynamic, different spray shapes, i.e., flat, hollow cone, and full cone sprays have been modeled in a computationally efficient manner. Droplets diameter distribution is then included in the model for each spray shape to determine the system-scale characteristics of the setup like the HTR and IPF. The total amount of ice formed during the winter operations has been used as a benchmark to validate the model. The model results indi­ cate that sprays with finer droplets (like cone sprays) will provide greater HTR and IPF. The developed model can be extended to capture other contributing phenomena like dropletdroplet and droplet-wall collisions. ACKNOWLEDGEMENT The authors wish to thank the Fonds de recherche du Québec - Nature et technologies (FRQNT) (PR-300597) and the Natural Sciences and Engineering Research Council of Canada (NSERC) (RGPIN-2021-02901). The first author would also like to show his gratitude to the McGill Engineering Doctoral Award (MEDA) for supporting his doctoral research. REFERENCES Akhtar, Saad, Xu, M. and Sasmito, A. P. 2021a. Development and validation of a semi-analytical framework for droplet freezing with heterogeneous nucleation and non-linear interface kinetics, International Journal of Heat and Mass Transfer. Elsevier Ltd, 166, p. 120734.

576

Akhtar, Saad, Xu, M. and Sasmito, A. P. 2021b. Development and validation of an asymptotic solution for a two-phase Stefan problem in a droplet subjected to convective boundary condition, International Journal of Thermal Sciences. Elsevier Masson SAS, 164(December 2020), p. 106923. Akhtar, S., Xu, M. and Sasmito, A. P. 2021. Spray freezing for mine heating a statistical perspective, Mine Ventilation, (2008), pp. 357–365. Caldwell, J. and Kwan, Y. Y. 2003. On the perturbation method for the Stefan problem with time-dependent boundary conditions, International Journal of Heat and Mass Transfer. Elsevier, 46(8), pp. 1497–1501. Deshpande, S., Gao, J. and Trujillo, M. F. 2011. Characteristics of hollow cone sprays in crossflow, Atomization and Sprays. Begel House Inc., 21(4). Liu, H. 1999. Empirical and Analytical Correlations of Droplet Properties, in Liu, H. B. T.-S. and E. of D. (ed.). Norwich, NY: William Andrew Publishing, pp. 238–314. Meirmanov, A. M. 2011. The stefan problem. Walter de Gruyter. Michaelides, E. E. 1997. Review—the transient equation of iviotion for particles, bubbles, and droplets, Journal of Fluids Engineering, Transactions of the ASME, 119(2), pp. 233–247. Rubinšteĭn, L. I. 2000. The stefan problem. American Mathematical Soc. Santangelo, P. E. 2010. Characterization of high-pressure water-mist sprays: Experimental analysis of drop­ let size and dispersion, Experimental Thermal and Fluid Science. Elsevier Inc., 34(8), pp. 1353–1366. Stachulak, J.S., 1991. Ventilation strategy and unique air conditioning at Inco Limited. CIM Bul-letin (Canadian Institute of Mining and Metallurgy); (Canada), 84(950). Trapani, K. and Chen, Z. 2017. Computational fluid dynamic modelling of the Frood-Stobie ice stope thermal storage for mine ventilation heating, Proceedings of the Eighth International Conference on Deep and High Stress Mining, pp. 289–298. Zhang, Haibin, Zhang, Hao and Bai, B. 2022. Numerical Study on Flow Structures of a Hollow Cone Spray in Crossflow, Heat Transfer Engineering. Taylor & Francis, 43(8–10), pp. 737–753.

577

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Analysis of small-scale lithium-ion batteries under thermal abuse A. Iqbal, G. Xu, R.I. Pushparaj & O.B. Salami Department of Mining and Explosives Engineering, Missouri University of Science and Technology, Rolla, MO, USA

ABSTRACT: Lithium-Ion Batteries (LIB) have dominated the energy market for several decades due to their high energy density and long-life cycle. However, several fire accidents in electric vehicles have raised questions about their safety concerns. Researchers have identified thermal runaways to be the major reason for the fire susceptibility of LIBs. One of the major risks in accidents involving LIB fires is toxic gaseous emissions. Hence, it is necessary to understand those toxic emissions to properly counter them. In this study, small-scale battery thermal abuse tests will be performed, and the toxic gases will be analyzed using different tools such as Fourier Transform Infrared (FTIR) spectroscopy ultimately. 5 different chemis­ tries of roughly the same nominal capacity i.e., 3000 mAh, will be tested and a comparison of the gaseous emissions will be presented in future studies, which has not been done before. However, so far, we have performed preliminary experiments to study the surface temperature of cells. The results of this study showed that the thermal runaway of the cell is triggered when the surface temperature of the cell reaches 190°C, and it keeps on increasing for up to 208°C, even without the use of the heating source, the duration of the experiment varies from 30min to 45min. Based on the study results, suitable experiment scenarios are designed the study the gaseous analysis of the LIB on the cell level. Keywords:

Thermal abuse, Small-scale LIB, Cell surface temperature

1 INTRODUCTION Lithium-ion batteries (LIBs) are the most commonly used batteries in battery electric vehicles (BEV) due to their high energy density and long lifetime (Larsson et al., 2016). However, sev­ eral BEV accidents involving battery fires have raised concerning alarms about the fire risks of LIBs (Sun et al., 2020)(CMTeam, 2022). Since 2006, Federal Aviation Administration (FAA) has recorded about 433 fire accidents involving LIBs in aviation (FAA, 2022). In 2012, A Toyota Prius (hybrid) was recorded to have caught fire after being flooded by hurricane Sandy (Labovick, 2021)(GARTHWAITE, 2012). In 2013, two Tesla Model Ss were reported to have caught fire after being involved in two separate collisions(Musk, 2013)(George, 2013). Further down the timeline, different car makers have experienced BEVs catching fire accidents such as Volkswagen in 2017 (Traugott, 2021), BMW i8 (Bruce, 2019), and two separate events involving Porsche Panamera in 2019 and a Taycan in 2020 happened in Portugal and USA respectively (Labovick, 2021). Several other events have happened when BEV caught fire while charging or parking idle (Wayland, Micheal. kolodny, 2021)(Yoney, 2021). Existing research has identified thermal runaways to be the probable reason for most of the fire accidents in BEVs (Sun et al., 2020)(Liu et al., 2021)(Liu et al., 2016). Thermal runaway happens when the battery pack ignites spontaneously due to overheating, according to the lit­ erature. It is caused by either electrical abuse, mechanical abuse/collision, or thermal abuse (Sun et al., 2020)(Liu et al., 2016)(Zhong et al., 2019)(Zhao et al., 2021). Once thermal runaway is DOI: 10.1201/9781003429241-59

578

triggered, chemical reactions start to take place inside the cell/pack, causing the emission of toxic fumes. Due to the presence of fuel and oxidizer inside the cell, these chemical reactions will most likely continue spontaneously, giving rise to heat and toxic gases during the process (Doughty and Roth, 2012). Several studies have analyzed the thermal characteristics of LIB fires, for instance, Yangyang Fu et.al (Fu et al., 2015) concluded from their experimental study that the stored electrical energy has a significant impact on the burning behaviors of LIBs. In the contrast, Larsson et.al (Larsson et al., 2016) stated that the stored electrical energy has nom­ inal dependence on the thermal properties of the battery. One of the most under-explored areas of BEV fires is the amount and types of gases emitted during fires. These toxic gases pose an immense threat to human health (Peng et al., 2020), as fatalities and injuries are mostly caused by the inhalation of toxic gases in fire accidents (Stec, 2017). Several researchers have performed experiments on the fire behavior and propagation of the fire in LIB cells (Fu et al., 2015)(He et al., 2020). However, very little information is present on the types and amount of toxic gases as suggested by a review performed by Sun et al. (Sun et al., 2020). The presence of toxic gases such as CO, CO2, HCl, HF, SO2, and POF3 has been confirmed by different studies(Larsson et al., 2017) (Ribière et al., 2012), How­ ever, the amount of the gases has still not been well understood. Larsson et al (Larsson et al., 2017) indicated the presence of HF, POF3, and PF5, However, the cells used by Larsson et.al (Larsson et al., 2017) were of different chemistries and the number of cells in each battery sample varied. In a separate study (Lecocq et al., 2016), the fire behavior of Li-ion batteries with LiPF6 chemistry was analyzed and the presence of toxic gases such as HF, and SO2 was confirmed, however, they were also unable to quantify the amount of the gases and also the experiment was done on only one chemistry of LIB. This research is also endorsed by a different study in which LIB cells were dissembled and the heat value and gases analysis was performed, confirming the presence of PF5, CO2, and HF (Kriston et al., 2019). However, the quantification and complete analysis of the gases is unavailable. Besides, reignition is another major problem related to LIB fires and several researchers have studied the effect of different extinguishers on LIB fires, and so far, a specifically effective extinguisher against LIB fires is non-existent (Maloney, 2014)(Luo et al., 2018)(Su et al., 2014). The main objective of this study is to evaluate the toxic fumes emitted during fire incidents involving Li-ion batteries. However, so far, only preliminary tests are performed, and the results are used as a guideline for the future study of the project. Similar studies have been performed by other researchers; however, the difference lies in the number of chemistries of LIB used, the types of gases analyzed, and the SOC relative to the LIB chemistries tested. The objective of the research will be achieved by first identifying and quantifying the toxic fumes by performing fire abuse tests. Applying fire directly to the battery cells, the cells will be left to burn out completely at different states of charge (SOC). The smoke of the batteries will be analyzed for the type and amount of gases, with the help of a multi-gas detector and also spectroscopy using FTIR. This analysis will be useful in identifying and quantifying the gases emitted during battery fires, which can further be used in designing specialized protective equipment and fire extinguisher designs, to elevate the health and safety standards of battery-powered vehicles. 2 METHODOLOGY 2.1 Lithium-ion battery incinerator To conduct the experiments, we designed and built the Lithium-ion battery Incinerator. The dimensions of the incinerator are 20×20×60 inches. It is made up of 26-gauge meta sheets and ¾” metal cube rods screwed together, and it is equipped with a 20×20×24” chamber for fire, and other divisions for the data logger and DC power supply. We fixed caster wheels to the bottom of the structure to easily move it in and out of the lab to protect ourselves from toxic exposures. The top of the incinerator is transitioned into a 4” duct with the help of an alumi­ num transit. Which is extended with a flexible plastic duct for about 5 feet. At the end of the duct, a 4” inline duct fan is fixed to pull the smoke out of the chamber Figure 1. 579

Figure 1.

Lithium-ion battery incinerator.

2.2 Cell selection LIBs are characterized by medium to high energy density, with a variety of chemistries. These cells have layers of Anode (copper foil coated with a specialty carbon) and Cathode (aluminum foil coated with a lithiated metal oxide or phosphate) separated by a microporous polyolefin separator. The electrode is an organic solvent and a dissolved lithium salt provides a medium for lithium-ion transport (Williams and Back, 2014). After a careful review of the literature, we found out that the emission of toxic gases is majorly influenced by the state of charge (SOC) and the cathode chemistry of the LIB. Hence, we decided on the five chemistries that are most common among electric vehicle manufacturers i.e. Lithium Iron Phosphate (LFP), Lithium Nickle Manganese Cobalt (NMC), Lithium Nickle Cobalt Aluminum (NCA), Lithium Titanate (LTO), and Lithium Manganese Oxide (LMO) at five different SOC i.e. 0%, 25%, 50%, 75%, and 100%. The detailed information on the cells can be seen in. So far, NiCad cells at 100% SOC were tested as preliminary experiments. Table 1.

Detailed specifications of the cells.

Chemistry

Voltage (V)

Capacity (Ah)

Packaging

No of cells

Manufacturer

LFP NMC NCA LTO LMO

3.3 3.6 3.6 3.3 3.75

2.5 2.5 2.25 3.0 2.8

26650 Cylindrical 18650 Cylindrical 18650 Cylindrical 18650 Cylindrical 18650 Cylindrical

10 20 10 10 10

Lithiumwerks Samsung Panasonic Titanate batteries Samsung

2.3 Abuse initiation Different studies have used different fire sources including an electric heater (Fu et al., 2015), a thermal oven (Larsson and Mellander, 2014), and a propane burner (Larsson et al., 2016). In this study, small-scale battery fire tests are conducted, and we used 18 Gauge Kenthal A-1 Annealed round, Nickle Chromium (Ni-Chrome) resistance wire, which is powered by NICE POWER DC power supply to heat the cells. The wore is a wire wound around the cell as 580

shown in Figure 2 and is gradually heated up at a rate of 4°C per min until the cell starts smoking off. A high-resolution thermal camera (FILR E4) Figure 7a, is used to capture the burning process of the battery based on the flame profile and temperature data, through the transparent window of the chamber.

Figure 2.

a: the Nickel Chromium resistance wire, b: application of heat using Ni-Chrome resistance wire.

2.4 Emissions analysis For thermal management of the cell, we made use of the cell surface temperature which was recorded using J-thermocouples attached to the surface of the sample with the help of heatresistant aluminum adhesive tape. While the data was recorded using the Anabi AT44532 tem­ perature data logger Figure 3 Thermocouples (TC) location on the cell is shown in Figure 4, TCs are designed to be attached to the cell at 1/4th of the length from each end.

Figure 3.

a: J thermocouples, b: the AT4532 temperature data logger.

For the analysis of the smoke, we made used Fourier Transform InfraRed (FTIR) spectros­ copy. Thermo Scientific Antaris IGS analyzer Figure 5, which is also reported to have been used by F Larsson (Larsson et al., 2017). The sample is burned in the chamber, and the smoke is pulled out by the inline duct fan at an average speed of 40 L/sec, however before the smoke is released into the atmosphere, the sample is sucked into the FTIR gas chamber using an external vacuum pump at 0.05 L/sec. The gas cell conditions are set at 180°C and 650 mmHG, and these were the conditions at which the FTIR Fire Science Calibration Library will be used for gaseous analysis. The experiment lasted for 45 min and the results are based on real-time analysis of the smoke. 581

Figure 4.

Location of thermocouples on the cell.

Figure 5.

Thermo scientific antaris IGS analyzer.

Fire will be directly applied to a battery cell in the chamber, once it starts to smoke, it will be collected and analyzed in by the multi-gas detectors and the sample of the gas will also be collected for FTIR analysis for the POF5, PF3, and HF. The experimental setup was bor­ rowed and modified from a study performed by F. Larsson (Larsson et al., 2017) in 2017. Once the gas sample is analyzed, and based on those analyses, PPEs designs, and Fire suppres­ sion techniques will be recommended. Figure 6 shows the illustration of the experimental setup used for the analysis of Toxic emis­ sions of LIB cells. Starting from Step 1 includes lithium-ion battery cells indicated by A1, which are first charged to the desired state of charge with a MIBOXER Cell charger indicated by A2 after that in step 2, J thermocouples (B1) are attached to the cell surface which is con­ nected to Anabi AT44532 temperature data logger (B2). In the next step, the 18 Gauge Kenthal A-1 Annealed round Nickle Cobalt resistance wire (C1) is wound into a coil and powered with a DC power supply (C2), and the system is placed in the incinerator. The cell is then heated till it starts smoking and the smoke is directed toward the FTIR (D1) in step 4, and the gases are analyzed with OMNIC software using a personal computer (PC). The data from the temperature data logger is also stored in the PC using AT4500 software, as shown in Figure 6, Step 4 and step 5 are not a part of this report. Compared to the similar setup used by other researchers, the advantage is that this setup is easily and commercially available in the market, this setup is much easier and economical to build and use for battery experiments. Other researchers have used a similar setup, however, they have some equipment such as SBI apparatus, which is much more lab specific and cost too 582

Figure 6.

Experimental setup used for toxic gas analysis of the LIB emissions.

much. This setup is a modified version of that used by Larsson et.al (Larsson et al., 2017) How­ ever, the difference is that the designed experiment can collect the temperature data using both thermocouples and a thermal imaging camera Figure 7(a), which will be further processed for fire profiling analysis. This will help us to better understand the fire profile and devise effective suppression solutions and help in the development of emergency PPEs for LIB fires. Another aspect of the novelty of this research is that so far, no researcher has identified and analyzed all the gases at once while studying five different LIB chemistries with the same power and charge capacities. For instance, F Larsson et.al (Larsson et al., 2017) studied three chemistries but with variable capacities, P. Ribière et.al (Ribière et al., 2012) and Yangyang Fu et.al (Fu et al., 2015) studied only one chemistry. Hence studying different chemistries at the same capacity level will give us an insight into how different chemistries behave under the same circumstances, and it will be a realistic comparison. 3 PRELIMINARY RESULTS Initially, trial experiments were performed to find a suitable method and apparatus for the experiments. The experiments were repeated 6 times, and most of the results are the same as described below. Different types of abuse methods were applied including propane burner, heating strip, and Ni-Chrome resistance wire. Tests indicated problems with controlling the heat applied to the sample using the propane burner. However, the heating strip and the NiChrome resistance wire techniques performed better. Ni-Chrome resistance wire with a DC voltage controller, as it is economical compared to the heating strip, is recommended, and will be used further in this research. The surface temperature of the cell was measured with the help of thermocouples, TC1 is attached to the top (valve) end while TC2 is connected to the bottom end Figure 4, data were recorded continuously every 10 sec. In these preliminary experiments, it was observed that the cell starts smoking when the surface temperature of the cell exceeds 190°C. The electrolyte is ejected from the cell. A mild hissing sound is evident that intensifies over the remainder of the experiment until all the gases are emitted from the cell. The burning process was also captured with a thermal camera to analyze the flame profile and surface temperature of the cell Figure 7. 583

Figure 7. cell.

A: FLIR E4 Thermal Imaging Camera, b: Thermal image of the burning process of the

The power voltage of the DC power supply was increased after each 10 min. which can be seen in Figure 8. The power of the supply was kept constant at 20W for the first 20 min. It can be seen on the surface of the temperature profile of the cell. Here, the rate of temperature increase doubles in the third section of the experiment. When the temperature of the cell reached 220°C, it started smoking vigorously, and the power supply was turned off. The cell kept smoking for 10 more mins and the temperature of the cell decreased thereafter. Figure 8 shows the data captured by only two thermocouples because only two of them were attached to the cell.

Figure 8.

Data Logger results of the surface temperature of the cell trial 1; TC1 (red), TC2 (yellow).

In another trial was repeated with the same procedure and different results were achieved this time for the surface temperature of the battery. The experiment lasted for more than one hour and the liquid was still confined within the cell. The maximum surface temperature was seen to be up to 280 °C. At the time, t=34.0 min, the power source was turned off. The surface temperature was seen to rise to 190 °C. It further rose to 208 °C without any external source. It can, therefore, be inferred that thermal runaway occurs when the temperature of the cell reaches 190 ° Figure 9. 584

Figure 9.

Data logger results of the surface temperature of the cell.

4 CONCLUSIONS The preliminary tests run on the small-scale LIB yield some scientific results. They are described below: • It takes 45 mins for the cell to burn out with the 50W DC power supply. • The cell starts to smoke when the temperature reaches 190 °C. This also could be assumed to be the temperature at which the thermal runaway begins. 5 FUTURE STUDIES For future recommendations, this study provides valuable input on cell behavior under ther­ mal abuse. The authors will continue to work on capturing the smoke coming out of the cell and analyze the nature of the smoke for the type and amount of different toxic gases with the help of FTIR, further in the project. NOTE This is a work in progress, and the study is still not complete yet. REFERENCES Bruce, C., 2019. Smoking BMW i8 Dumped In Water By Firefighters [WWW Document]. URL https:// www.motor1.com/news/315476/bmw-i8-fire-reponse-netherlands/ CMTeam, 2022. How Many Electric Cars Catch Fire Every Year: Electric Car Fire Statistics (2022 Find­ ings) [WWW Document]. URL https://www.carsmetric.com/electric-car-fire-statistics/#:~:text=Key Electric Car Fire Statistics 1 Electric-powered vehicles,vehicles and 3.4%25 for hybrid vehicles. More items. Doughty, D.H., Roth, E.P., 2012. A general discussion of Li ion battery safety. Electrochem. Soc. Inter­ face 21, 37. FAA, 2022. Lithium Battery Incidents [WWW Document]. URL https://www.faa.gov/hazmat/resources/ lithium_batteries/incidents Fu, Y., Lu, S., Li, K., Liu, C., Cheng, X., Zhang, H., 2015. An experimental study on burning behaviors of 18650 lithium ion batteries using a cone calorimeter. J. Power Sources 273, 216–222. Garthwaite, J., 2012. Mystery at Port Newark: Why Did 17 Plug-In Cars Burn? NewYork Times.

585

George, P., 2013. Another Tesla Model S Caught Fire After A Crash In Mexico [WWW Document]. URL https://jalopnik.com/another-tesla-model-s-caught-fire-after-a-crash-in-mexi-1453376349. He, X., Restuccia, F., Zhang, Y., Hu, Z., Huang, X., Fang, J., Rein, G., 2020. Experimental study of self-heating ignition of lithium-ion batteries during storage: effect of the number of cells. Fire Technol. 56, 2649–2669. Kriston, A., Adanouj, I., Ruiz, V., Pfrang, A., 2019. Quantification and simulation of thermal decompos­ ition reactions of Li-ion battery materials by simultaneous thermal analysis coupled with gas analysis. J. Power Sources 435, 226774. Labovick, 2021. ELECTRIC VEHICLE FIRE INCIDENTS AND STATISTICS [WWW Document]. URL https://www.labovick.com/blog/electric-vehicle-fire-incidents-and-stats/ Larsson, F., Andersson, P., Blomqvist, P., Mellander, B.-E., 2017. Toxic fluoride gas emissions from lithium-ion battery fires. Sci. Rep. 7, 1–13. Larsson, F., Andersson, P., Mellander, B.-E., 2016. Lithium-ion battery aspects on fires in electrified vehicles on the basis of experimental abuse tests. Batteries 2, 9. Larsson, F., Mellander, B.-E., 2014. Abuse by external heating, overcharge and short circuiting of com­ mercial lithium-ion battery cells. J. Electrochem. Soc. 161, A1611. Lecocq, A., Eshetu, G.G., Grugeon, S., Martin, N., Laruelle, S., Marlair, G., 2016. Scenario-based pre­ diction of Li-ion batteries fire-induced toxicity. J. Power Sources 316, 197–206. Liu, X., Wu, Z., Stoliarov, S.I., Denlinger, M., Masias, A., Snyder, K., 2016. Heat release during thermally-induced failure of a lithium ion battery: Impact of cathode composition. Fire Saf. J. 85, 10–22. Liu, Y., Sun, P., Niu, H., Huang, X., Rein, G., 2021. Propensity to self-heating ignition of open-circuit pouch lithium-ion battery pile on a hot boundary. Fire Saf. J. 120, 103081. https://doi.org/10.1016/J. FIRESAF.2020.103081 Luo, W., Zhu, S., Gong, J., Zhou, Z., 2018. Research and development of fire extinguishing technology for power lithium batteries. Procedia Eng. 211, 531–537. Maloney, T., 2014. Extinguishment of lithium-ion and lithium-metal battery fires. US Fed. Aviat. Adm. Washington, DC, USA 46–51. Musk, E., 2013. Model S Fire [WWW Document]. URL https://www.tesla.com/blog/model-s-fire Peng, Y., Yang, L., Ju, X., Liao, B., Ye, K., Li, L., Cao, B., Ni, Y., 2020. A comprehensive investigation on the thermal and toxic hazards of large format lithium-ion batteries with LiFePO4 cathode. J. Hazard. Mater. 381, 120916. Ribière, P., Grugeon, S., Morcrette, M., Boyanov, S., Laruelle, S., Marlair, G., 2012. Investigation on the fire-induced hazards of Li-ion battery cells by fire calorimetry. Energy Environ. Sci. 5, 5271–5280. Stec, A.A., 2017. Fire toxicity–The elephant in the room? Fire Saf. J. 91, 79–90. Su, C.-H., Chen, C.-C., Liaw, H.-J., Wang, S.-C., 2014. The assessment of fire suppression capability for the ammonium dihydrogen phosphate dry powder of commercial fire extinguishers. Procedia Eng. 84, 485–490. Sun, P., Bisschop, R., Niu, H., Huang, X., 2020. A review of battery fires in electric vehicles. Fire Tech­ nol. 56, 1361–1410. Traugott, J., 2021. Two-Day-Old VW Golf Hybrid Explodes While Driving [WWW Document]. URL https://carbuzz.com/news/two-day-old-vw-golf-hybrid-explodes-while-driving. Wayland, Micheal. Kolodny, L., 2021. Fires, probes, recalls: The shift to electric vehicles is costing auto­ makers billions [WWW Document]. URL https://www.cnbc.com/2021/08/19/fires-probes-recalls-auto makers-spend-billions-in-shift-to-evs.html Williams, F.W., Back, G.G., 2014. Lithium battery fire tests and mitigation. NAVAL RESEARCH LAB WASHINGTON DC CHEMISTRY DIV. Yoney, D., 2021. Hyundai Kona Electric Fires In Norway And Korea Cause Concern [WWW Docu­ ment]. URL https://insideevs.com/news/515983/kona-electric-fire-norway-korea/. Zhao, J., Xue, F., Fu, Y., Cheng, Y., Yang, H., Lu, S., 2021. A comparative study on the thermal runaway inhibition of 18650 lithium-ion batteries by different fire extinguishing agents. Iscience 24, 102854. Zhong, G., Mao, B., Wang, C., Jiang, L., Xu, K., Sun, J., Wang, Q., 2019. Thermal runaway and fire behavior investigation of lithium ion batteries using modified cone calorimeter. J. Therm. Anal. Calorim. 135, 2879–2889.

586

Ventilation monitoring and measurement

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Gauge and tube surveys: What is their future and that of underground measurements generally as mines transition towards greater use of Big Data and Artificial Intelligence systems? D.J. Brake Mine Ventilation Australia

ABSTRACT: Gauge and Tube (also known as Trailing Hose or P-Q) Surveys have been widely used in underground mine ventilation for frictional pressure loss surveys and ventilation model calibrations for a very long time. They are one of the most basic types of underground ventilation measurement. However, there are many factors that have been contributing to and accelerating the decline of this type of survey and underground measurements more generally. In most cases, barometric pressure surveys are now the preferred method for validating ventilation circuits and models and can be completed more quickly, more cheaply, with fewer personnel and with less disruption to mine operations. In addition, the growing use of real-time instrumentation of the ventilation system combined with Big Data analysis techniques is reducing the amount of manual measurements and other data collection required. Artificial Intelligence is also likely to assist in the future with providing statistically reliable advice regarding the location, type and fre­ quency of underground measurements, as well as answers to many other ventilation questions including validating ventilation models and providing site-specific advice on management of upset conditions such as fan or ventilation control failures or outages, fires or explosions. This paper discusses the factors behind these trends and identifies some of the advantages and disad­ vantages of both gauge and tube and barometric pressure surveys, as well as the broader poten­ tial use of Big Data and Artificial Intelligence to assist with decisions regarding underground measurements generally and assisting with other ventilation-related advice.

1 INTRODUCTION Gauge and tube ventilation surveys have been the mainstream ventilation survey technique for measuring underground flow directions and volumes for a very long time. However, this type of survey started falling from favour in Australian hardrock mines from the early 1980s and was almost entirely abandoned by the early 1990s. Gauge and tube surveys persevered somewhat longer in coal mines but have also now been abandoned for many years. The coup de grâce for the decline of this type of survey occurred during the transitional period when underground mines went from 5-day operations to 7-day operations. This major structural change in the industry was usually associated with moving from 8-hour to 12-hour shifts and in many cases in Australia, also associated with moving from residential to fly-in, fly-out (FIFO) operations. The nature of gauge and tube surveys means they need to be undertaken when there is the minimum possible activity in the mine resulting in the minimum disruption to the ventilation cir­ cuits. This was manageable when gauge and tube surveys could be undertaken when the mine was mostly idle over a weekend and ventilation staff lived in a nearby town and could complete the survey on “overtime”. When mines moved to 7-day operations, gauge and tube surveys became increasingly problematic.

DOI: 10.1201/9781003429241-60

589

However, the reasons for their declining use were not only related to the loss of available non-production time underground to conduct the survey but also to the availability of suitable personnel and their time and the rise of alternative measurement technologies and processes. In addition, the rapid advances in the use of real-time instrumentation, Big Data analysis techniques and Artificial Intelligence are likely to become a significant factor in the science and practice of underground measurements and ventilation model validation, fault-finding and decision-making in the medium term. This paper summarises the author’s opinions about Guage and Tube Surveys in particular and extends these reflections to the case of underground measurements more generally. 2 VENTILATION SURVEYS In most cases, the purposes of a P-Q survey in a mine are to: • Check airflow directions for compliance with the required safe and effective circuit config­ uration, fresh air bases and egress routes • Measure windspeeds and hence volumetric airflow rates to ensure statutory or good prac­ tice compliance with minimum or maximum wind speeds, diesel requirements, gas dilution, recirculation, etc. • Measure pressure differentials for management of unsafe pressures on ventilation controls, avoidance of excessive leakage or spontaneous combustion and the like • Measure frictional pressure gradients in a mine to establish high-resistance or high-cost paths for debottlenecking by appropriate circuit changes or installation of circuit fans • Provide a baseline (trending) for fault-finding in the ventilation circuits • Validate the as-built ventilation model to ensure it is a reliable tool for future mine plan­ ning, e.g. modelling major extensions to ventilation circuits or assessing fan duties for the purchase or relocation of primary fans, etc. This last objective has really only arisen since computer modelling of mine ventilation networks became practical in the 1970s. From the mine’s point of view, the method used to measure the parameters required to achieve these purposes is not important, providing the objectives themselves are achieved. In particular, it is not important to obtain a direct measurement of the resistance of individual airways or even the entire mine for the mine ventilation engineer. The critical underground measurements for validating a ventilation model are airflow directions and volume flows, pressure differentials across key parts of the circuit and checking that fans are operating on their curves (Brake, 2015; Rowland, 2011). Even more broadly than this, it could be argued that the future of mine ventilation measure­ ments will be to use real-time instrumentation combined with fewer, but smarter, manual measurements and integrate these very targeted manual measurements with the use of increas­ ingly sophisticated Big Data and mine ventilation Artificial Intelligence technologies to signifi­ cantly improve the timeliness and reliability of mine ventilation decisions and advice. 3 PROBLEMS WITH GAUGE AND TUBE SURVEYS The theory and techniques for gauge and tube (G&T) surveys are well described by McPher­ son (2007). However, there are a number of drawbacks of using gauge and tube surveys in modern mines, despite their historical popularity. • In a gauge and tube survey, the resistance of every section of airway (calculated from every flow and pressure measurement of the gauge and tube as it is advanced) has to be added to the resistance of the previous gauge and tube measurement. Any errors that creep in even­ tually get carried through the system (at least unless a second closed loop can be formed). • Airway resistances often change over time for a variety of reasons, so that just surveying the incremental new parts of the ventilation circuit since the last survey can and will 590

• • • • •

eventually result in problems in the model as older airways are altered and these changes are not necessarily recognised. There is significant manpower and time required and hence significant labour cost. There is the opportunity cost to these time-consuming surveys. While the ventilation team is undertaking a gauge and tube survey, they are not doing other, often critical, ventilation work and this lost time cannot be made up There is the disruption to the operations with restrictions on equipment travel and/or venti­ lation control operation during the period of the survey to obtain accurate results There are potential safety issues for the survey team if surveying in high activity areas, e.g. on main ramps or travelways or in exhaust airways during production periods. It is not possible or very difficult to undertake a gauge and tube survey in many vertical raises or other airways in which travel through is not possible.

Expanding the remit of this discussion from gauge and tube surveys to the role of underground ventilation measurements more generally, it could easily be argued that most of the underground measurements taken by the ventilation professional are never used again and contribute little to the effective running of the mine, or the validation of the ventilation model in the mine. In some cases, the collection of underground measurements actually reduces the ability of the ventilation professional to perform other more value-adding activities and therefore decreases the overall effectiveness of the ventilation system and negatively impacts on the ventilation outcomes on site. 4 CHANGING TECHNOLOGIES There are four trends over the past few decades impacting on the methodology for conducting ventilation measurements surveys underground. Firstly, the development of very comprehensive and accurate ventilation modelling pro­ grams. Compared to the pre-computer era or even the period of early ventilation modelling software, modern mine ventilation modelling programs can comprehensively include all the factors that drive flow and pressure in underground ventilation networks including variable density, compressible airflow, shock losses and full thermodynamic modelling as required. To some extent they can also not only model steady-state conditions but even transient condi­ tions. Any errors in current ventilation models are due to errors in the physical inputs of the airways and leakage resistances etc., rather than due to assumptions or approximations in the software itself, as was previously the case. Secondly, the ready availability of inexpensive high accuracy barometers. In many respects, the traditional method of undertaking ventilation surveys in a mine is analogous to the trad­ itional method of undertaking a topographical survey of surface land: in historical land survey­ ing practices, surveyors would use levels or plane tables and theodolites to comprehensively measure the three dimensions of many surface points and create contour maps from these measurements. However, with the advent of, firstly, accurate aerial mapping and more recently differential GPS, these older methods of land surveying have been mostly superseded. It is very possible today to use aerial photographs to develop high resolution contour maps with only the occasional spot heights (or spot elevations) surveyed by humans, if at all. In a similar way, high accuracy barometers have made similar advances possible with ventilation surveys as will be discussed in the next section. Thirdly, the increasing trend to instrumenting the ventilation system in a mine, often part of a more comprehensive Ventilation-on-Demand system. The use of fibre-optic communication backbones has also made real-time Big Data type of analyses now possible. Instrumentation, VOD and Big Data will continue to decrease the number of manual measurements needing to be taken, assuming suitable software becomes available to process this data into useful information. Fourthly, the development of so-called Artificial Intelligence is beginning to have a major impact on assisting with solving many complex problems in society and industry ranging from medical diagnosis to setting up spreadsheet formulas. AI is likely to have an equally profound impact on the profession of mine ventilation over the decade to come. 591

5 BAROMETRIC PRESSURE SURVEYS The theory and techniques for barometric (or altimeter) pressure (BP) surveys are well described by McPherson (2007). BP surveys are not new and date back to around the 1920s (McElroy & Kingery, 1957; Harris et al, 1973). However, it is the ready available and low cost of high accuracy barometers (Derrington, 2015) that has made the technique a far more attractive option than previously. A barometric pressure survey can be used in at least two ways: • For airways which have a significant frictional pressure loss along their length, a BP survey can be used to directly measure frictional pressure loss, resistance and friction factor for that airway • For validation of a ventilation model, effectively checking the integrity of the model against measured “spot heights” (in the pressure sense). In the latter case, a number of nodes are selected throughout the mine and the barometric pressure (and wet and dry bulb temperatures) are accurately measured at those locations. If the airflows in the mine are known (and these can be measured easily and accurately), and the elevation of the BP positions is accurately known (to within about 0.5 m, which can now usu­ ally just be done simply and easily using the survey floor strings in the as-built survey data­ base), then the pressure loss between surface and those node points can be accurately measured and compared to the predicted values from the ventilation model. In addition to correlating the model by comparing predicted and actual BP values, it is good practice to also measure the differential pressure (by manometer) across key ventilation controls, which can also then be checked against the modelled values of differential pressure (DP) at these loca­ tions. Ventilation controls are also inspected to estimate leakage (m3/s) so that this can be accurately modelled. Measuring DP and flow across ventilation controls (effectively a gauge and tube survey of a single “airway”) remains very useful in terms of measuring the leakage resistances of different types and constructions of ventilation controls (e.g. doors, air cross­ ings). Any discrepancies between measured and modelled values of BP, DP and leakage can be investigated. Once any differences are within a reasonable tolerance, the model can be con­ sidered to be validated and fit for purpose. The main advantages of barometric pressure surveys are: • They are quick and inexpensive (compared to gauge and tube) • There are no cumulative errors as each measurement stands alone as a fixed reference point for validation in the model • Any changes in the resistance of old airways or significant new sources of leakage are detected because each BP survey is, in effect, a full new mine survey. • Very importantly, if the as built ventilation model has failed to include some new piece of development somewhere, or some important development has not been removed from the model (i.e. these changes have inadvertently not been included in the ventilation model) then the BP survey will detect this, but an incremental gauge and tube survey of only the new development in a mine will not. • A BP survey can be conducted by one person, assuming the base (or control) station tech­ nique is used, with the surface barometer at a suitable location in data-recording mode • A BP survey can be conducted with the mine operating as normal with generally little loss of accuracy. • A BP survey results in negligible disruption to mine operations. Barometric pressure surveys are particularly useful for very large operations or those with many airways that are difficult or impossible to traverse. For example, the three integrated Mount Isa underground mines circulated 6000 m3/s of air through over 600 km of ventilated airways over a strike length of 5 km, from surface to 1900 metres below surface and with levels every 40 m (every 20 m in the upper areas).

592

6 THE IMPACT OF VENTILATION SYSTEM INSTRUMENTATION, VENTILATION-ON-DEMAND AND BIG DATA The increasing range of responsibilities of the mine ventilation engineer has already put great pressure on his or her ability to collect manual measurements from underground. Even where these are collected, sometimes for statutory purposes, they are often not used for any useful subsequent analysis. Most new mines are now being set up with a fibre optic communications backbone and larger mines are implementing an increasing degree of real-time instrumenta­ tion and monitoring, and in some cases, adopting a form of remote and semi-autonomous control of the ventilation system (VOD). The ability to collect and usefully manage large sets of data (Big Data) will be an essential precursor to the development and refinement of Artifi­ cial Intelligence systems in mine ventilation engineering. 7 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON MINE VENTILATION ANALYSIS INCLUDING VENTILATION MODEL VALIDATION Artificial Intelligence (AI) has been defined (Copeland 2023) as “the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings. . . developing systems endowed with the intellectual processes characteristic of humans, such as the ability to reason, discover meaning, generalize, or learn from past experience”. Examples would include some robotics, self-driving cars, chatbots, virtual travel agents and many others. So how could AI assist with ventilation decision-making and model validation in the future? With respect to ventilation model validation, an interesting paper by Griffith and Stewart (2021) effectively described using a limited number of gauge and tube survey data to improve the calibration of a mine ventilation computer model by setting up a sensitivity matrix of actual versus measured values to modify the resistance of the most probable airways that need resist­ ance adjustment. It is likely that this approach could be modified to do one of the following: • Use the same limited (or nil) P-Q data, along with accurate barometric pressure data and differential pressure measurements, and with the option for the user to set or relax friction factors and airway sizes and other parameters, to develop the most reliable true estimate of the ventilation model of the mine, or • Assess the model and provide a list of ventilation controls or fan locations and the type of measurements needing to be undertaken at each location (e.g. pressure difference across a ventilation control, pressure and flow through a fan, barometric pressure difference to surface), to provide a statistically-valid confidence level for the model. For example, the AI system could determine that if certain specific measurements are taken, it would then be in a position to use these measurements in the model to determine how much uncertainty is left in the reliability of the overall model, or perhaps even provide advice about the uncer­ tainty regarding a particular question from the ventilation engineer such as this: “Given the existing ventilation model and the results of the last flow and pressure survey and the accur­ acy of these measurements, what additional measurements would need to taken to be 95% confident that a predicted fan pressure and flow at (for a new fan duty at some location) would be within 3% on both flow and pressure”. • In fact, all five questions posed in the dot points in section 2 of this paper could be candi­ dates to put to an AI system in the future. Artificial Intelligence will certainly impact on the role of measurements and data collection and analysis more generally for the mine ventilation professional in the future. However, this author’s view (Brake, 2022) is that the impacts will mostly be positive. AI could provide statis­ tically sound advice answering many of the questions that come up frequently (or less fre­ quently) for the ventilation engineer allowing for faster, more reliable and more timely decisions. In addition, AI could remove the need to take many of the often-tedious and timeconsuming measurements that perhaps add no real value to the mine ventilation monitoring 593

program whilst simultaneously improving the reliability of the analysis, freeing up the ventila­ tion professional for more important parts of the job. 8 CONCLUSION Given the increasing demands on the ventilation engineer’s time, the shrinking size of the ven­ tilation department on most mine sites, and the 24/7 nature of mine operations, barometric pressure surveys, when combined with volume surveys, are increasingly the most practical method of validating ventilation models. This very useful survey technique should be part of the repertoire of every practicing mine ventilation engineer and its specific requirements (e.g. timing) should be included in the mine’s Ventilation Management (Control) Plan. With regard to ventilation measurements more generally, it is likely that Big Data analysis and Artificial Intelligence will play an increasingly useful role in many technical related mat­ ters, including those in mine ventilation, over the next decade. AI is likely to develop the point where it can provide independent and statistically sound advice regarding the types and loca­ tions of measurements to be taken underground to achieve a statistically reliable estimate of the overall validity of a ventilation model or to rapidly answer specific questions regarding fault-finding or diagnosis of the ventilation system including under upset conditions such as power failures, fan failures, fires or explosions. REFERENCES Brake D J, 2015. Quality Assurance Standards for Mine Ventilation Models and Ventilation Planning, Proc 3rd Aust Mine Vent conf (Belle B (ed)). AusIMM. pp. 221–228. Brake D J, 2022. Developing and maintaining mine site ventilation capability and new trends in mine ventilation career paths. The 6th Australian Mine Ventilation Conference (Belle, B (ed). AusIMM. Gold Coast, 10-12 Oct 2022. pp 181–204. Copeland, R J, 2023. Artificial Intelligence. Artificial intelligence | Definition, Examples, Types, Applica­ tions, Companies, & Facts | Britannica. Encyclopaedia Britanica. Derrington, A.S. 2015. Development of a Low-cost Instrument for Barometric Pressure Surveys. The Australian Mine Ventilation Conference. AusIMM. Chalmers, D (ed). Sydney, NSW. AusIMM. 31 Aug-2 Sep 2015. Griffith M D & Stewart C M, 2021. Ventilation model calibration with limited survey data. Proc 18th NAMVS, Tukkaraja, P (ed). 12–17 Jun 2021. Online. Taylor & Francis. pp 509–517. Harris, E.J., Dalzell, R.W, Kline, R.J & Miller, E.J. 1973. A Method for Calculating Mine Ventilation Pressure Losses Using Computers and Desk-Top Calculators, A Supplement to Information Circular 7809. Information Circular 8594. US Bureau of Mines. McElroy, G.E & Kingery, D.S. 1957. Making ventilation-pressure surveys with altimeters, Information Circular 7809. US Bureau of Mines. McPherson, M.J. 2007. Subsurface Ventilation Engineering. http://www.mvsengineering.com/ 835 pp. Rowland, J.R. 2011. Ventilation Surveys and Model Building–A Consultant’s Opinion. Australian Mine Ventilation Conference/Sydney, NSW, 5-6 September 2011.

594

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Improving the accuracy of field airflow measurements for tunnel ventilation fans R.E. Ray, Jr. & E. Fuster WSP USA Inc., New York, NY, USA

ABSTRACT: Accurately measuring the airflow quantity delivered by tunnel ventilation fans during site acceptance testing can be difficult due to the fan configuration and constrained sizes of adjoining plenums. Most axial flow tunnel ventilation fans are reversible with the drive motor attached directly to the fan impeller. Obstructions created by motor supports pre­ clude performing pitot tube traverses in the fan section and the presence of transition duct­ work and sound attenuators on each side of the fan dictate that airflow measurements be taken at the end of the fan equipment train. Point air velocity measurements taken with an anemometer at the fan room and plenum interface are averaged to calculate the fan airflow quantity. Limited floor space in ventilation buildings may result in narrow plenums adjacent to fan rooms that produce non-uniform flow with high velocities concentrated on one side of the fan ductwork, making the accuracy of anemometer point traverses problematic. This paper presents the results of a study of alternative grid spacing of point anemometer traverses to yield more accurate field measurements using CFD modeling of a sample fan room and plenum arrangement.

1 INTRODUCTION The NFPA (National Fire Protection Association) 130 Standard for Fixed Guideway Transit and Passenger Rail Systems (2020) requires passenger rail and transit (subway) tunnels and stations to be equipped with emergency ventilation systems that provide tenable egress condi­ tions for passengers evacuating from a train fire incident. To meet the standard, a series of shafts are typically spaced along the length of the tunnels to facilitate a “push-pull” ventilation system utilizing reversible axial flow fans. For a fire emergency in the tunnel, passengers would walk towards the shaft supplying air, while smoke is exhausted from a different shaft in the opposite direction. Passenger rail stations encompassing multiple tracks with much larger enclosed space than tunnels may be equipped with an exhaust emergency ventilation system to maintain smoke layers above the walking height of passengers. Station platform ventilation fans are also usually reversible to allow them to work in tandem with adjoining tunnel ventila­ tion shafts if required. As noted by Ray and Gamble (2010), since ventilation systems serve a critical life safety function for passenger rail and transit tunnels used by the public, factory acceptance testing of tunnel ventilation fans is required for most projects to confirm that the fans deliver the design duty airflow. Factory testing is conducted in accordance with procedures described in AMCA 210-07, “Laboratory Methods of Testing Fans for Certified Aerodynamic Performance Rating,” with pitot tube traverses performed across a measurement plane located a minimum of 8.5 test duct diameters away from the fan with essentially a uniform air velocity profile across the duct. In addition to factory acceptance tests, site acceptance testing is also performed to verify that tunnel ventilation fans deliver the design duty airflow quantity in the field. As shown in DOI: 10.1201/9781003429241-61

595

Figure 1, an elevation view of a typical reversible axial fan equipment train, round-torectangular transition ducts are attached to each side of the fan with sound attenuators placed downstream of the transitions. Measuring velocity and total pressures with a pitot tube in the fan section is not feasible since the drive motors are coupled directly to the fan impeller and motor support structures create disturbances to airflow and obstruct insertion of the tube. The increase in cross-sectional area through the transition ducts creates angular airflow that also precludes using a pitot tube since the bent tip section of the pitot tube needs to be aligned with the flow. As a result, measurement of fan total pressure in the field is not feasible. Since parallel full height baffles in the sound attenuators create channeled airflow with no access for measurement instrumentation, suitable locations for measuring fan airflow quantity are gener­ ally limited to the ends of the fan equipment train.

Figure 1.

Typical reversible tunnel ventilation fan configuration.

2 MEASURING FAN AIRFLOW QUANTITY AT THE PLENUM INTERFACE Air velocity traverses taken across the end of the fan equipment train in an adjoining air plenum can be performed by direct measurement of velocity using vane anemometers or by measuring velocity pressure with a pitot tube and calculating the velocity. Due to the large size of the duct­ work, pitot tube traverses are cumbersome to perform, requiring custom tubes long enough to reach the center of the ducts and sufficiently stiff to avoid deflection with the fan operating. Digital anemometers with remote readouts and attached to extension rods are easier to use and can allow testing personnel to remain to the sides of the duct for all measurements. Field testing specifications typically require anemometer traverses to be performed twice across the measurement plane and averaged. Conducting continuous traverses at the plenum interface without personnel standing in front of a portion of the duct would be difficult, and repeatability of successive traverses would be problematic since testing contractors do not have experience with this method. Instead, point air velocity measurements taken at the end of the fan equipment train are averaged and multiplied by the duct cross-sectional area to esti­ mate the fan airflow quantity. As noted by Jorgensen (1999), the most frequently used method to locate traverse stations is to divide the duct into equal areas and measure at the centroid of the subdivided areas. In determining the number of point measurements comprising the tra­ verse grid, practical considerations need to be weighed against selecting a sufficient number of locations to establish the airflow distribution exiting (or entering) the fan ductwork. A test grid with too few velocity measurements will not accurately reflect the airflow quantity delivered by fan, while a grid with a large number of measurement points may take an exces­ sive amount of time to complete. For a facility with multiple fans operating in parallel, the time required to complete two traverses for each fan in both exhaust and supply operation can be substantial. McElroy (1935) recommended not less than 16 traverse points for square cross 596

sections and not less than 24 for rectangular cross sections where one dimension is twice the other for “precise traversing methods” in major underground air currents or tests of primary fan performance. He further noted that a minimum spacing of 0.304 m (1 ft) between grid points to be satisfactory in mine airways, and that if closer spacing appears to be required, “it is very probable that the conditions of turbulence would prevent a high degree of accuracy by any method of measurement”.

Figure 2.

Air velocity traverse being performed with digital anemometer.

3 SAMPLE FIELD TEST DATA Fan airflow quantity and power measurements taken during site acceptance testing recently completed for a passenger rail station are summarized in Table 1. Point anemometer traverses were performed at the end of the fan equipment trains for both the exhaust and supply airflow directions for eight reversible axial fans—four, 2.59-m (102-in.)-dia., 184 kW (250 hp) fans and four, 2.16-m (85-in.)-dia., 110 kW (150 hp) fans. Velocity measurements were taken with a digital vane anemometer equipped with a remote readout at a total of 36 points at the end of the 3.66-m (12-ft) by 3.66-m (12 ft) fan ductwork for the 2.59-m (102-in.)-dia. fans and at the end of the 2.44-m (8-ft)-wide by 2.74-m (9-ft)-high ductwork for the 2.16-m (85-in.)-dia. fans. The measurement points were located at the center of 36, 0.61-m (24-in.) by 0.61-m (24in.) subdivided square areas in the larger ducts. This spacing placed measurement points 0.31 m (12 in.) from the duct walls and 0.61 m (24 in.) apart in between. For the smaller ducts, measurement points were located at the center of six-wide by six-high, 0.41-m (16-in.) by 0.46m (18-in.) rectangles. This grid spacing placed measurement points 0.31 m (8 in.) from the duct side walls and 0.41 cm (16 in.) apart across the duct, and 0.23 m (9 in.) from the top and 597

bottom duct walls and 0.46 cm (18 in.) vertically apart. The grid measurement locations were marked on the screens installed across the ends of the ductwork with electrician’s tape. The anemometer was secured to an adjustable painter’s extension pole to allow testing personnel to remain to the side of the ductwork while taking measurements, as shown in the photograph provided in Figure 2. A ladder was needed to help reach some measurement stations in the top center of the larger duct. At each measurement station, the anemometer was set up to read continuously over a 10-second interval and the high and low velocities were averaged. A three-phase power quality analyzer wired directly to the fan motor terminals was used to directly measure current, voltage and power factor and calculate power. In addition to listing the field test airflow quantities and power for each fan in both direc­ tions of flow, Table 1 also displays the fan airflow quantity derived from plotting fieldmeasured power (corrected to the project air density) on the factory test airflow quantity vs. bhp curves. This is considered the most accurate method of determining fan airflow in the field for tunnel ventilation projects. The field-measured airflow quantities ranged from approximately 112% to 132% of the airflow quantities determined from the bhp factory curves for the 2.59-m (102-in.)-dia. fans in exhaust and 114% to 131% in supply. The field-measured airflow quantities for the 2.16-m (85-in.)-dia. fans varied from approximately 99% to 109% of the airflow quantities determined from the bhp factory curves in exhaust and 99% to 115% in supply. It should be noted that the repeatability of the two traverses for each fan test was excellent, with the differences in average air velocity between successive traverses falling in a range of 0.1% to 3.1%. This indicates that the variations seen between the field-measured airflow quantities and the airflow quantities determined from the factory airflow quantity vs. bhp curves was not due to inconsistent handling of the anemometer and extension pole setup or failing to measure the same grid point locations in successive traverses. It is also assumed that the installed fan arrangement does not degrade the factory performance curves since the field airflow measurements exceed those predicted from the factory curves and AMCA 201-02 (R2011) addresses “system effects” causing non-uniform flow conditions at the fan inlet or outlet by increasing the fan system resistance, not de-rating the fan. An additional velocity traverse was subsequently performed for SVF-4 operating in exhaust with an increase from 36 to 64 measurement points to evaluate whether the accuracy Table 1. Site acceptance testing airflow measurements for station ventilation fans. EXHAUST

TVF-1

TVF-2

TVF-3

TVF-4

kW (bhp) Q measured: m3/sec (cfm) Q bhp curve: m3/sec (cfm) Measured Q/bhp Q

138 (185) 145.7 (310,000) 124.1 (264,000) 117.4%

144 (193) 148.1 (315,000) 120.3 (256,000) 123.0%

131 (175) 143.4 (305,000) 127.8 (272,000) 112.1%

156 (209) 152.8 (325,000) 115.6 (246,000) 132.1%

SUPPLY

TVF-1

TVF-2

TVF-3

TVF-4

kW (bhp) Q measured: m3/sec (cfm) Q bhp curve: m3/sec (cfm) Measured Q/bhp Q

138 (185) 136.8 (291,000) 119.9 (255,000) 114.1%

134 (180) 154.6 (329,000) 122.2 (260,000) 126.5%

129 (173) 146.6 (312,000) 127.8 (272,000) 114.7%

145 (194) 154.2 (328,000) 117.5 (250,000) 131.2%

EXHAUST

TVF-5

TVF-6

TVF-7

TVF-8

kW (bhp) Q measured: m3/sec (cfm) Q bhp curve: m3/sec (cfm) Measured Q/bhp Q

62 (83) 70.0 (149,000) 71.0 (151,000) 98.7%

70 (94) 72.9 (155,000) 66.7 (142,000) 109.2%

69 (92) 72.4 (154,000) 67.7 (144,000) 106.9%

66 (88) 70.5 (150,000) 70.0 (149,000) 100.7%

SUPPLY

TVF-5

TVF-6

TVF-7

TVF-8

kW (bhp) Q measured: m3/sec (cfm) Q bhp curve: m3/sec (cfm) Measured Q/bhp Q

66 (89) 64.4 (137,000) 64.9 (138,000) 99.3%

61 (82) 76.1 (162,000) 67.7 (144,000) 112.5%

64 (86) 76.6 (163,000) 66.7 (142,000) 114.8%

68 (91) 70.0 (149,000) 64.4 (137,000) 108.8%

598

of a 36-point traverse was significantly impacted by the spacing of velocity measurement points. For the revised test, measurement points were placed in the center of 64, 0.41-m by 0.41-m (16-in. by 16-in.) subdivided duct sections. These points were within 0.20 m (8 in.) of the duct walls and 0.41 m (16 in.) apart, compared to 0.31 m (12 in.) from the duct walls and 0.61 m (24 in.) apart for the original test traverses. Despite a 78% increase in velocity meas­ urement locations, the airflow quantity measured in exhaust was 151.8 m3/s 323,000 cfm, only 1% less than the original measurement and still 131% of the airflow quantity predicted from power measurements. Velocity contours across the end of the fan equipment train developed from the field test traverse measurements are provided in Figure 3 for TVF-4 operating in supply. Field airflow quantity measurements for TVF-4 varied the most from those predicted from power measure­ ments (132.1% exhaust/131.2% supply) while field airflow quantity measurements for TVF-5 were closest to those predicted from power measurements (98.7% exhaust/99.3% supply). The simplest explanation for these differences is that measurement points were spaced closer together and closer to the duct walls for the smaller ductwork connected to TVF-5 versus the larger ductwork connected to TVF-4. However, the velocity contours displayed in Figure 4 for TVF-4 supply operation indicate that the airflow is more concentrated in the upper center to lower right side of the duct with velocities exceeding 13.2 m/s (2,600 fpm) in the upper center. The overall average measured duct velocity was 8.8 m/s (1,736 fpm). The uneven flow distribution may make it more difficult to accurately measure the average velocity for SVF-4 in supply.

Figure 3.

Air velocity contours in fpm for SVF-4 supply traverse.

4 CFD AIRFLOW MODELING OF SAMPLE FAN ROOM Based on the results of the 64-point traverse for TVF-4 operating in exhaust, field airflow measurements were not repeated for TVF-4 or any other fans with additional grid points. Instead, it was decided to study alternative velocity traverse grid spacings using computational fluid dynamics (CFD) analysis to evaluate the number of points that would provide the closest measurement to the actual airflow rate of a highly skewed flow distribution. The primary goal 599

of this analysis was to use CFD as a tool to easily compute airflow quantity from skewed vel­ ocity contours using different measurement patterns, not to simulate the internal airflow through the fan equipment train with a high degree of accuracy. A three-dimensional model of the TVF-4 fan room and adjoining plenums originally developed during construction using AutoCAD was updated to include more detail of the fan equipment train. As illustrated in Figure 4, representations of the fan dampers, the full height baffles installed in the sound attenu­ ators located on each side of the fan and the fan drive motor and foot-mounted motor pedestal were added to the fan equipment train portion of the CFD model. The only missing features in the geometry compared to the physical installation were the fan impeller and wire mesh screens. Transient simulations of TVF-4 operating in supply were performed using ANSYS Fluent soft­ ware. The cases utilized the realizable k-ϵ turbulence model and the grid size was approximately 10 million polyhedral cells. The TVF-4 supply airflow quantity was set to approximately 117.5 m3/sec (250,000 cfm) to match the quantity determined from the factory performance curves using the field-measured power. The upstream inlet and downstream outlet were modeled as pressure boundaries, and to achieve the desired non-uniform flow at the plane of measure­ ment the Fluent “fan” internal boundary condition utilized was set to swirl the flow to produce a similar velocity pattern to that depicted in Figure 3. An acceptably skewed pattern was devel­ oped after iterating the fan boundary settings. To determine if the solution was grid independ­ ent, two sensitivity cases were run, and the results were nearly identical. Air velocities were measured for four different traverse grids across the end of the ductwork adjacent to the exhaust plenum in the CFD model: (1) 36 points, (2) 49 points, (3) 81 points and (4) 144 points. The 36-point grid was identical to that used in the field testing, while the 81-point and 144-point grids placed the air velocity measurement points at the center of 9 by 9 and 12 by 12 equal subdivided areas of the overall 3.66-m (12-ft) by 3.66-m (12-ft) duct, respectively. Traverse points for the 81-point grid were within 0.20 m (8 in.) of the duct walls and 0.41 m (16 in.) apart, compared to 0.15 m (6 in.) from the duct walls and 0.31 m (12 in.) apart for the 144-point grid. The 49-point traverse was based on the log-Tchebycheff method that utilizes a logarithmic distribution adjacent to the duct walls and a polynomial distribution of velocity for the interior points. Spacing of the 49-point traverse grid points is depicted in Figure 5. As noted by Jorgensen (1999), all points are also equally weighted in this method. The velocity traverse points were represented as 6.99-cm (2.75-in.)-dia. circles in the CFD model that mimic the size of the digital anemometer. Area-weighted averages of velocity for each circle were averaged across the duct cross section to determine the airflow quantity meas­ ured by each traverse grid configuration.

Figure 4.

Three-dimensional CFD model of TVF-4 fan equipment train and plenums.

600

Figure 5.

Log-Tchebycheff traverse grid for 3.66-m (12-ft) by 3.66-m (12 ft) cross section.

5 ALTERNATE TRAVERSE GRID SPACING RESULTS Velocity contours at the measurement plane at the end of the fan equipment train adjacent to the exhaust plenum for the CFD simulation of TVF-4 operating in supply with the 36-point traverse grid are shown below in Figure 6. As noted previously, the Fluent “fan” internal boundary condition utilized in the CFD model was set to swirl the flow to attempt to skew the higher air velocities in the 11.2 m/s (2,200 fpm) to 14.2 m/s (2,800 fpm) range to the upper center and lower right portions of the cross section as was measured in the field. Table 2 pro­ vides a summary comparison of the actual airflow quantity set in the CFD model (as close as possible to the quantity derived from the power measurement) and the airflow quantity com­ puted from the traverse of the CFD simulation cross section at the end of the fan equipment train for each of the four alternative grid spacings. An examination of the airflow data shown in Table 2 confirms that the CFD simulations predict that the supply airflow quantity delivered by TVF-4 would still be overestimated by using any of the four alternate grid spa­ cings. The simulation of the “base case” of 36 points was somewhat closer to the assumed air­ flow quantity vs. the 36 points measured in the field (124.1% of assumed flow vs. 131.2%). Using the centroid of subdivided equal areas, increasing the number of measurement points from 36 to 81 would reduce the “overmeasurement” of the TVF-4 supply airflow from 124.1% to 117.2% of the assumed airflow. Increasing to 144 measurement points would further reduce the “overmeasurement” of airflow to 114.9% of the assumed airflow. Using the logTechebycheff grid spacing of 49 points results in a predicted TVF-4 supply airflow quantity of 120.3% of the assumed airflow. It would appear that using the log-Techebycheff method to space measurement locations does not provide advantages over using centroids of subdivided equal areas since the simulation data shows that its overestimated airflow of 120.3% of the assumed airflow for 49 measurement points falls in between the 36 and 81-point equal area centroid spacing overmeasurements. A comparison of the data contained in Tables 1 and 2 indicates that the field measurements for the smaller 2.16-m (85-in.)-dia. fans (TVFs-5, 6, 7 and 8) are all closer to the assumed airflows (98.7% to 114.8%) than the simulated measure­ ments for similar or even closer traverse point spacing for the 2.59-m (102-in.)-dia. fans. 601

Figure 6. supply.

Air velocity contours at inlet end of ductwork for CFD simulation of TVF-4 operating in

Table 2. TVF-4 actual CFD model supply airflow quantity vs. traverse CFD airflow quantity. Traverse Grid

Actual CFD Q m3/sec (cfm)

Traverse CFD Q m3/sec (cfm)

Traverse CFD Q/ actual CFD Q

36 Points 81 Points 144 Points 49 Points (Log-Tchebycheff)

119.2 (253,000) 120.6 (256,000) 120.1 (255,000) 118.2 (251,000)

147.9 (314,000) 141.3 (300,000) 138.0 (293,000) 142.2 (302,000)

124.1% 117.2% 114.9% 120.3%

6 CONCLUSIONS Obtaining accurate measurements of the airflow quantity delivered by tunnel ventilation fans during site acceptance testing is challenging for a number of reasons, including skewed airflow distribution through the fan caused by narrow adjoining plenums and the difficulty locating measurement cross sections sufficiently distant from obstructions and transitions in the duct­ work. Recent field test data presented for a station with eight reversible tunnel ventilation fans showed airflow quantities measured with anemometer traverses overestimating the fan airflow quantity determined from field power measurements applied to factory test curves by 12 to 32 percent for 250,000 cfm fans and as much as 15 percent for 125,000 cfm fans. CFD modeling of the fan with the highest overmeasurement in the field showed a reduction in over­ measurement from 24% to 15% by increasing the number of measurement points from 36 to 602

144 across 3.66-m (12 ft) by 3.66-m (12 ft) fan ductwork. Two complete traverses over the cross section being measured are required for field testing of tunnel ventilation systems, how­ ever, and performing two, 144-point traverses for each direction of flow for a large number of fans would be time prohibitive. In the case of the 117.5 m3/sec (250,000 cfm) fans, increasing the number of velocity measurement points from 36 to 81 points would be more practical and result in more accurate results, although still 17 percent high based on the results of the CFD analysis described above. Since nearly all of the fan airflow quantities measured in the field for the sample station exceeded the derived values using power measurements, it is clear that the lower velocities adja­ cent to the duct walls are underrepresented in the traverses. Based on the CFD analysis, this would also apply to the traverse with the largest number of measurement points (144), with points 0.15 m (6 in.) from the duct walls. Because the CFD simulations also predict overmeasure­ ment of average air velocities for the unequal airflow distribution entering the fan equipment train, it is unlikely that anemometer error is contributing to the overmeasurement of the fan airflow. REFERENCES AMCA. 2011. Publication 201-02 (R2011), Fans and Systems. Arlington Heights, IL: Air Movement and Control Association International, Inc. ANSI/AMCA. 2008. Standard 2010-07 (ANSI/ASHRAE Standard 51-07), Laboratory Methods of Test­ ing Fans for Certified Aerodynamic Performance Rating. Arlington Heights, IL: Air Movement and Control Association International, Inc. Jorgensen, R. (ed). 1999. Fluid flow. In Fan Engineering, Chapter 2: 2–90. Kingsport, TN: Quebecor World Book Services. Jorgensen, R. (ed). 1999. Fluid flow. In Fan Engineering, Chapter 2: 2–94. Kingsport, TN: Quebecor World Book Services. McElroy, G.E. 1935. Engineering Factors in the Ventilation of Metal Mines: 12–13. Washington, D.C.; U.S. Department of the Interior, Bureau of Mines. National Fire Protection Association. 202. NFPA 130: Standard for Fixed Guideway Transit and Passen­ ger Rail Systems: 130-22-130-24. Quincy, MA. Ray, R. E. & Gamble, G.A. 2010. The benefits of conducting factory performance tests for main mine fans. In S. Hardcastle & D.L. McKinnon (eds.), Proceedings of the 13th U.S./North American Mine Ventilation Symposium: 529–539. Sudbury, Ontario: MIRARCO – Mining Innovation.

603

Ventilation network analysis and optimization

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Case study on the abnormal airflow diagnosis method using atmospheric monitoring data L. Zhou, D. Bahrami & R.A. Thomas Pittsburgh Mining Research Division National Institute for Occupational Safety and Health (NIOSH), Pittsburgh, USA

ABSTRACT: A stable and well-maintained mine ventilation system is the key to ensuring a safe and healthy working environment for miners. A sudden, unplanned, and significant change in airflow termed as abnormal airflow is frequently observed in mine ventilation. Some abnormal airflows can return to normal without manual intervention; however, some abnormal airflows may cause catastrophic accidents if left unattended. In addition, abnormal airflow may be a consequence of an accident such as a blocked airflow route due to a roof fall. Promptly diagnosing and locating the cause of abnormal airflow can help prevent acci­ dents. Researchers at the National Institute for Occupational Safety and Health (NIOSH) have developed a method to diagnose the cause of abnormal airflow for underground mine ventilation systems. The purpose of this paper is to verify the developed method using experi­ mental tests conducted at NIOSH’s experimental mine. The airflows were monitored by a real-time atmospheric monitoring system installed in the experimental mine during the tests. As demonstrated in this case study, the developed abnormal airflow diagnosing method, based on the resistance sensitivity and matching method, has been proven to be reliable. Keywords: airflow

Mine ventilation, Network modeling, Atmospheric monitoring system, Abnormal

1 INTRODUCTION Ventilation is the lifeline of underground mines to provide the underground workings with fresh air in sufficient quantity and quality, dilute and remove dust and noxious gases, and regulate tem­ perature and other conditions. A stable and well-maintained mine ventilation system is the key to ensuring a safe and healthy working environment for miners. The efficiency of mine ventilation highly relies on airflow distributions and airflow quantities in the mine ventilation system. The airflow distributions are closely monitored and controlled during the life of an active mine. Changes in airflow quantity or direction are not rare in the routine management of ventilation. The airflow changes can be categorized as: 1) planned changes for the purpose of maintaining a safe and healthy working environment; 2) periodic changes associated with atmospheric condi­ tions such as barometric pressure, temperature, humidity, and so on; 3) expected temporary changes from the interference of moving equipment; 4) sudden, unplanned, and significant changes due to some hidden hazardous conditions or malfunction of ventilation fans. The fourth category of airflow changes can be termed abnormal airflow if the changes in quantity exceed a certain level. The abnormal airflow can be caused by some kinds of failures in ventilation. For instance, unattended opening/closing of a ventilation door, airpath blockage due to roof collapse, fan stoppage, or any other unforeseen hazardous condition. The abnormal airflows may cause catastrophic accidents if left unattended. For example, insufficient air supplies resulting from abnormal airflow may induce elevated levels of hazardous gas accumulation. Therefore, promptly DOI: 10.1201/9781003429241-62

607

diagnosing and locating the cause of abnormal airflow can help prevent accidents. Researchers at the National Institute for Occupational Safety and Health (NIOSH) have developed a method, named as abnormal airflow diagnosis method, to assist the diagnosis of the cause of abnormal airflows for underground mine ventilation systems (Bahrami & Zhou, 2022). Bahrami & Zhou (2022) also verified the abnormal airflow diagnosis method using the cases generated with ventila­ tion network modeling. The method was found to be reasonably accurate in locating an unknown source airway caused by a flow resistance change in a ventilation network. However, the verifica­ tion was based on model-based numerical data which tends to provide “perfect” cases without involving any interferences from the real world. There is a need to test the proposed method using real-world scenarios. For this reason, we have conducted a ventilation test at NIOSH’s experimen­ tal mine by changing airflow distribution while monitoring of the Atmospheric Monitoring System (AMS) and testing the abnormal airflow diagnosis method using the experimental data. In this paper, while we refer to our earlier work (Bahrami & Zhou, 2022), the focus is to experimen­ tally verify and demonstrate how to diagnose an abnormal airflow using the proposed method with the monitored AMS data. 2 METHODOLOGY The abnormal airflow diagnosis method referred to in this paper is based on the availability of the sensitivity matrix for a given mine ventilation network and the matching method which compares the monitored airflow changes with their sensitivities from the sensitivity matrix. For a better understanding of the diagnosis procedure, a brief introduction to the sensitivity matrix and the matching method is necessary. These have been described comprehensively by the authors previously (Bahrami & Zhou, 2022; Zhou & Bahrami, 2022). 2.1 Sensitivity matrix The airflow quantity distributed to each airway of a mine ventilation system depends on the schematic connecting diagram of airways, the resistance of airways, and air pressures pro­ duced by mechanical fans or/and natural ventilation. For a given mine ventilation network, a resistance change in an airway will affect the airflow quantities in other airways more or less. The resistance sensitivity is frequently used to quantify airflow change in a certain airway due to resistance change in another airway (Li et al., 2011; Dziurzynski et al., 2017; Griffith & Stewart, 2019; Jia et al., 2020; Zhou & Bahrami, 2022). Mathematically, the resistance sensi­ tivity can be expressed as (Zhou et al., 2007; Jia et al., 2020):

Where Sij is the sensitivity of airway i to the change of resistance, ΔR, in airway j; Rj is the resistance in airway j, and Qi is the volumetric airflow rate in the airway i. Given a ventilation network with N airways, each airway can have N sensitivities to the rest of the airways and itself. Therefore, an N � N sensitivity matrix S can be formed for the ventilation network (Zhou & Bahrami, 2022):

2.2 Matching method The sensitivity resistance matrix of a ventilation network serves as a benchmark to indicate the interdependence of airflow and resistance of any two airways. If any unexpected airflows 608

are detected in a ventilation system, the airflow changes can be compared to their sensitivity columns in the sensitivity matrix for each airway to find out which airway can cause a similar airflow change pattern and has the closest airflow change quantities. Bahrami & Zhou (2022) employed Root Mean Square Error (RMSE) to measure how well the changed airflows match the sensitivity of each airway. For each airway in the network, a corresponding scale factor is determined by minimizing the corresponding RMSE. It is assumed the airway with the min­ imum RMSE is the airway that has caused the abnormal airflow due to its resistance change. For detailed information about the matching method, refer to Bahrami & Zhou (2022). 3 VENTILATION TESTS AT THE SAFETY RESEARCH COAL MINE The Safety Research Coal Mine (SRCM) is one of the two experimental mines located at the Bruceton Campus of NIOSH in Pittsburgh, PA. Despite the smaller size of entry (6.5 ft high by 14 ft wide) compared to the operating mines nowadays, this room-and-pillar mine has a complete ventilation network and has been frequently used for research regarding mine ven­ tilation, mine fires, and other studies (Zhou et al., 2017, 2018, 2019, 2020). The main fan, installed at the surface above the return shaft, exhausts air from the mine. Stoppings, doors, regulators, and brattices are used in the mine to direct the airflow to the desired routes. The quantity of main airflow getting into the mine is controlled by the main fan and a door at the main return entry. The layout of the SRCM is shown in Figure 1. The SRCM is equipped with a mine-wide AMS to continuously monitor the atmospheric condition of the mine. Nine air-velocity sensors are arranged throughout the mine to monitor the real-time air velocity (as shown in Figure 1) for the ventilation tests conducted in this study. The sensors are all mounted close to the roof in the middle of the entry. The airflows are continuously monitored at these locations with a server on the surface to store air-velocity readings every minute.

Figure 1.

Air-velocity sensor locations within the SRCM.

609

Within the SRCM, a door in the entry connecting B Butt and A Butt remains closed for normal ventilation (as shown in Figure 1). To create an abnormal airflow scenario, we intention­ ally open this door and maintain the open status for approximately 17 hours for our testing. The real-time air velocities at the nine AMS monitoring locations were recorded using air-velocity sensors every minute. Figure 2 shows the monitored air velocities from Sensors S2, S3, S4, and S5 during the time the door was open There was a large air-velocity drop at Sensor S2 from around 500 ft/min to 250 ft/min when the door was opened. The air velocity returned to 500 ft/ min after the door was closed. Upon opening the door, a slight air-velocity increase can be seen at Sensors S3 and S4. Sensors S1, S6, S7, S8, and S9 do not see significant air-velocity changes upon opening and closing the door. Therefore, they are not plotted in the figure.

Figure 2.

Monitored air velocities at Sensor S2, S3, S4, and S5.

4 SRCM VENTILATION NETWORK MODEL CONSTRUCTION AND CALIBRATION As it was mentioned above, the abnormal airflow diagnosis method proposed in this paper con­ sists of two major parts: one is the resistance sensitivity matrix for a ventilation network, and the other is the matching method using RMSE. The generation of the resistance sensitivity matrix of a ventilation network is based on one-time ventilation simulation results. Therefore, the collaboration of the network model and the mine is critical to ensure a good application of the abnormal airflow diagnostic method. Continuous efforts have been made to establish and maintain a well-calibrated ventilation network model for the SRCM (Iannacchione et al., 2015; Zhou et al., 2022). The schematic outline drawing of the SRCM ventilation network including 166 airways is displayed in Figure 3. The source airway of the abnormal airflow, where the door is maintained open for the test, is labeled as Airway #40 in the network drawing. Table 1 displays the comparison of the measured airflows from all nine air-velocity sensors and simulated airflows using MFIRE, which is an open-sourced mine fire simulation program having ventilation network simulation capability. The calculation of resistance sensitivity is 610

Figure 3.

Schematic drawing of the SRCM ventilation network with airway numbers.

implemented in MFIRE as well (Zhou & Bahrami, 2022). Due to the fluctuation in the airvelocity sensor readings, a week-long average air velocity at each sensor was used as the value of the measured air velocity. With the known cross-sectional area for each sensor location, the airflow rate can be obtained by multiplying the average sensor reading by the cross-sectional area at each sensor location. However, the calculated airflow rate is based on the point reading of the air-velocity sensor instead of the average airflow rate through the whole cross-sectional area as the air-velocity sensor measures the air velocity where the sensor is mounted. The airvelocity sensors used in the SRCM are typically installed at the upper section of the entry, close to the roof or ribs, to avoid interference from the moving equipment and miners. A correction factor for each sensor location was introduced and calculated to convert the point airflow rate to the average airflow rate following the same procedure illustrated by Zhou et al. (2017) and by Zhou & Bahrami (2022). The adjusted airflow rates using these correction factors are considered the final airflow rates. The comparisons were made between the adjusted airflow rates and the simulated airflow rates using MFIRE. The average difference for all the sensor locations was 4.63% with the largest difference occurring at Sensor 4.

Table 1. Comparison of measured and simulated airflow at monitoring locations. Sensor ID

Measured Vel­ ocity (ft/min)

Area Airflow (ft3) (kCFM)

Correction Adjusted Air­ Factor flow (kCFM)

Simulated Air­ flow (kCFM)

Difference

1 2 3 4 5 6 7 8 9

436.69 449.13 43.8 95.44 186.11 235.63 182.57 191.58 479.2

65.83 58.00 73.37 76.38 69.30 97.35 62.25 62.25 70.71

1.01 1.01 0.85 0.81 0.94 0.90 0.95 1.10 1.13

29.56 23.23 3.29 7.29 11.96 23.64 12.39 9.43 30.88

3.71% –11.03% –14.91% –23.45% –14.73% –7.82% 3.45% –14.97% 2.90%

28.75 26.05 3.21 7.29 12.90 22.94 11.36 11.93 33.88

28.46 25.79 3.78 9.00 13.72 25.49 11.96 10.84 29.98

Average

4.63%

611

5 RESULTS AND DISCUSSION By applying the derivative method implemented in MFIRE (Zhou & Bahrami, 2022), a 166by-166 resistance sensitivity matrix was obtained for the SRCM ventilation network. In the sensitivity matrix, if we say the rows represent the source airways where resistance changes are made and the columns represent the target airways whose airflow changed according to the resistance change at each airway, an element Sij in the ith row and jth column is the rate of airflow change in airway j responding to the resistance change in row i. As shown in Figure 2, airflow changes occurred at Sensors S2, S3, S4, and S5 while opening the door at Airway #40. Table 2 summarizes the monitored airflow changes at all nine sensors and the correspond­ ing airway number (#) of each sensor in the SRCM ventilation network.

Table 2. Airflow changes at all monitoring locations. Sensor

S1

S2

S3

S4

S5

S6

S7

S8

S9

Airflow change (CFM) Airway # of sensor in network

0 2

13,167 155

3,413 81

2,173 88

–902 77

0 117

0 118

0 157

0 19

In this case, the airflow change at Sensor S2 located in Airway #155 is obviously abnormal due to the significant airflow reduction of 13,167 cfm. To find out the possible source airway­ (s), whose resistance change has caused similar airflow changes in all sensor airways including Airway #155, all the monitored airflow changes, including zero changes, at all nine sensor air­ ways are set as target airflow changes. Then the columns, representing the corresponding sensor airways, in the resistance sensitivity matrix are extracted out to be used to match the input airflow changes. Before applying the abnormal airflow diagnosis method, each row is screened for the sorted sensitivities to eliminate all-zero rows. An all-zero row means that the resistance change in the airway represented by this row has no influence on all monitored air­ ways, and the calculated RMSE of the all-zero row airway will be zero as well. Figure 4 is showing the RMSE for each non-zero sensitivity row. It can be seen that Airway #40 has the least RMSE which indicates Airway #40 is the best matching airway to cause the airflow changes in the monitored airways, as well as the abnormal airflow change in Airway #155. In addition to Airway #40, Airways #145, #155, and a few others can also be the source airways causing the airflow changes in the monitored airways.

Figure 4.

RMSE of non-zero airways for the case.

612

6 CONCLUSIONS The abnormal airflow in underground mine ventilation could be an indication of a malfunction of the ventilation system. If left unattended, it may lead to catastrophic accidents. The cause of the abnormal airflow needs to be diagnosed, located, and mitigated to restore a safe working environment. An abnormal airflow diagnostic method has been developed previously by NIOSH researchers. In this work, the diagnostic method is verified experimentally using a ventilation test conducted at NIOSH’s experimental mine. This paper has clearly shown that the abnormal airflow diagnostic method can diagnose the cause of abnormal airflow. In the ven­ tilation test, air-velocity sensors from an AMS were used to monitor and collect air-velocity data. It has been demonstrated in this research that air-velocity sensors are a practical tool to detect abnormal airflow. From the research that has been carried out, it is possible to conclude that the abnormal airflow diagnostic method can be potentially used in practice in the future. To ensure a successful application of the abnormal airflow diagnostic method in a mine ventilation system, the constructed mine ventilation network model must be as close as pos­ sible to the real operating condition. The number and the selection of the monitored airflow change inputs also plays an important role in the outcome of the proposed method. The min­ imum number of input airflow changes is three. Further research has been planned to study the impact of the number and the selection of the input airflow changes regarding the confi­ dence level of the results from the proposed method. DISCLAIMER The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the National Institute for Occupational Safety and Health, Centers for Disease Control and Prevention. Mention of any company or product does not constitute endorsement by NIOSH. ACKNOWLEDGEMENT The authors wish to thank John Soles of the Pittsburgh Mining Research Division for con­ ducting the full-scale AMS test. REFERENCES Bahrami, D. Zhou, L. 2022. A novel methodology to locate an abnormal airflow in underground mine ventilation networks. Proceedings of the 2022 SME Annual Conference & Expo, Preprint 22-018, Salt Lake City, February 27-March 2, 2022. Dziurzynski, W. Krach, A. Palka, T. 2017. Airflow sensitivity assessment based on underground mine ventiltion systems modeling. Energies, 10(10), 1451. Griffith, M. & Stewart, C. 2019. Sensitivity analysis of ventilation models – where not to trust your simu­ lation. Proceedings of the 17th North American Mine Ventilation Symposium. Montreal, Canada: The Canadian Institute of Mining, Metallurgy and Petroleum, pp. 654–660 Iannacchione A. Joseb A. Hornc, T. Iannacchione, G. Iannacchione, S. 2015. Modeling the ventilation network at NIOSH’s Safety Research Coal Mine. Proceedings of the 15th North American Mine Ven­ tilation Symposium, June 20-24, 2015, Blacksburg, VA. Jia, J. Jia, P. and Li, Z. 2020. Theoretical study on stability of mine ventilation network based on sensi­ tivity analysis. Energy Science & Engineering, 00, 1–8. doi:10.1002/ese3.699. Li, G. Kocsis, C. Hardcastle, S. 2011. Sensitivity anlysis on parameter changes in underground mine ven­ tilation systems. Journal of Coal Science & Engineering, 17(3), 251–255. Zhou, L. Luo, Y. Wu, F. 2007. Identificataion of abnormal airflow quantity in underground coal mines. International Journal of Mineral Resources Engineering, 12(4), 257–265. Zhou, L. Yuan, L. Thomas, R.A. Iannacchione, A. 2017. Determination of velocity correction factors for real-time air velocity monitoring in underground mines. International Journal of Coal Science and Technology, (2017) 4: 322–332.

613

Zhou, L. Yuan, L. Bahrami, D. Thomas, R.A. Rowland, J.H. 2018. Numerical and experimental investi­ gation of carbon monoxide spread in underground mine fires. Journal of Fire Sciences, 36(5): 406–418. Zhou, L. Yuan, L. Bahrami, D. Thomas, R.A. Cole, G. 2019. Study on integration of real-time atmos­ pheric monitoring system monitoring data and MFIRE simulation. Proceedings of the 16th North American Mine Ventilation Symposium, Montreal, Canada, April 28-May 1, pp. 794–803. Zhou, L. Yuan, L. Thomas, R.A. Bahrami, D. Rowland, J.H. 2020. Determination of a mine fire inten­ sity using atmospheric monitoring system in a ventilation network. Proceedings of the 2020 SME Annual Conference & Expo, Preprint 20-005, Phoenix, AZ, February 23–26. Zhou, L. Thomas, R.A. Yuan, L. 2022. Experimental study of improving a mine ventilation network model using continuously monitored airflow. Mining, Metallurgy & Exploration. 39, 887–895. https:// doi.org/10.1007/s42461-022-00574-4. Zhou, L. & Bahrami, D. 2022. A derivative method to calculate resistance sensitivity for mine ventilation networks. Mining, Metallurgy & Exploration. 39, 1833–1899 https://doi.org/10.1007/s42461-022-00630-z.

614

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Ventilation system upgrades at the Waste Isolation Pilot Plant K.G. Wallace, Jr. SRK Consulting, (U.S.), Inc., Clovis, CA, USA

I. Peña Nuclear Waste Partnership LLC, Carlsbad, NM, USA

ABSTRACT: The Waste Isolation Pilot Plant (WIPP) facility, operated by the U.S. Depart­ ment of Energy (DOE), is the only transuranic waste repository in the United States. The facility is designed for the permanent disposal of transuranic radioactive waste generated through nuclear weapons production and research. On February 14th, 2014 a continuous air monitor (CAM) alarmed indicating a radioactive contamination event had occurred in the active emplacement panel underground. The ventilation system automatically switched to a filtration mode where all exhaust air is sent through a bank of filters including High Effi­ ciency Particulate Air (HEPA) filters. The original filtered ventilation system had limited cap­ acity. Interim projects have included adding additional fans and filtration units to the surface and an underground fan to enhance airflow to mining areas. The DOE has decided to main­ tain the WIPP facility for TRU waste disposal and to this end has authorized several signifi­ cant capital projects. These projects, which are currently under construction, includes a dust and water extraction system on surface combined with one of the world’s largest HEPA filtra­ tion systems. The new Safety Significant Confinement Ventilation System (SSCVS) will be capable of filtering up to 540,000 cfm. The dust extraction system will contain six 100,000 cfm capable filtration units to remove salt dust and water before the air reaches the HEPA filters. The second project is a new 28 ft diameter intake shaft 2,150 ft in depth capable of passing 500,000 cfm. New surface intake fans are designed for this shaft. An intake shaft at the site is being reconfigured as an exhaust for the new intake fans. Both capital projects are significant with expenditures expected to exceed $400 million. This paper describes the analyses per­ formed to size the new fans, filtration units and shaft to support work activities and to main­ tain proper flow alignment in the underground. The status of the capital projects will also be described. 1 INTRODUCTION In 1979, the United States Congress authorized the Waste Isolation Pilot Plant (WIPP) as a research and development facility to demonstrate the safe disposal of transuranic radioactive (TRU) waste from defense activities. TRU waste results chiefly from the production of nuclear weapons from plutonium and enriched uranium. The term transuranic indicates that the waste contains radionuclides with atomic numbers greater than 92, that is, greater than that of uranium. TRU waste consists of a wide variety of contaminated materials from laboratory and production operations, including discarded protective clothing, laboratory test equipment and reagents, machine components, and solidified sludge. This waste has accumulated over the past 75 years because of weapons development and production at U.S. defense facilities. The goal of WIPP is to dispose legacy TRU waste from 22 generator sites nationwide. The facility is managed by the Department of Energy (DOE) with Nuclear Waste Partnership, LLC being the operating contractor. WIPP is the first geological repository in the nation for which an application for permanent geological isolation has been approved. WIPP is located approximately 35 miles southwest of Carlsbad, New Mexico (see Figure 1). The repository DOI: 10.1201/9781003429241-63

615

horizon is located 655 m (2150 ft) below surface in the Salado salt horizon. The salt horizon was formed about 250 million years ago and is 610 m (2,000 ft) thick. This salt horizon was selected based on its stable geology, lack of water, high thermal conductivity and because of its “creep” characteristics which will eventually encapsulate the waste in the emplacement rooms. The first shipment of TRU waste was received in 1999.

Figure 1.

Location of WIPP facility.

2 ORIGINAL WIPP VENTILATION SYSTEM DESCRIPTION WIPP currently consists of four shafts and eight panels for the disposal of TRU waste. The original ventilation system at the WIPP facility utilizes three shafts for intake air, the salthandling shaft (SHS), waste handling shaft (WHS) and air-intake shaft (AIS). A single exhaust shaft (ES) is utilized for the return airway at the facility. During normal operations, fresh air is supplied primarily by the AIS with secondary intake from the SHS. Additionally, the SHS is used for personnel and material access as well as removal of the mined salt. Nuclear waste is lowered through the waste handling shaft which is equipped with an enclosed headframe. The 6.1 m (20 ft) diameter waste handling shaft provides access for personnel and equipment to the repository with limited intake downcasting the shaft. The air from the waste-handling shaft is directly routed to the exhaust shaft after ventilating the wastehandling shaft station. Figure 2 gives a cross-section of the WIPP facility and the original infrastructure. Figure 3 illustrates the original surface fan configuration at WIPP. Primary ventilation through the facility was achieved by running of one or two of the 450-kW (600 hp) centrifugal main fans (labeled 700A, B and C). During concurrent mining and waste-handling operations, two of the three fans operate in parallel to provide 230 m³/s (490,000 cfm). The surface fan arrangement also included a filtration mode. In this mode the airflow decreases to 28 m³/s (60,000 cfm). The shift from full airflow to filtration mode is accomplished by turning off the main fans and starting one of the three 175 kW (235 hp) 860 centrifugal standby filtration fans. A series of isolation dampers diverts the air through the filtration system where the air is routed through a series of high-efficiency particulate air (HEPA) filters. The underground ventilation system is divided into four circuits. These are the North area, where the primary equipment shop is located along with a salt disposal investigation (SDI) experimental area, the construction circuit, TRU waste disposal circuit and the waste shaft station circuit. Regulators strategically located in each circuit ensure airflow courses from an area where there is no TRU waste towards the disposal circuit. Figure 4 illustrates the four underground circuits at the beginning of 2014. 616

Figure 2.

Geology at WIPP with original shaft and underground design.

Figure 3.

Repository configuration with air splits shown (in 2014).

3 CONTAMINATION EVENT On February 14, 2014 a radiation detection sensor, called a continuous air monitor (CAM), alarmed in Panel 7. The CAM measures airborne radioactivity within the exhaust of the panel where active waste emplacement operations and final storage occurs. The CAM alarmed in the evening when no personnel were underground at the time of the incident and triggered an 617

Figure 4.

Original fan system at WIPP.

automatic ventilation system switch from one primary unfiltered fan (700C) to filtration mode (860C) where all air exhausting the underground passes through high efficiency filter banks. The cause of the release was due to a single drum that contained incompatible materials resulting in a spontaneous combustion event in Room 7 of Panel 7. Figure 5 shows the damage to the TRU container. The release contaminated certain areas in the underground, namely downstream of where the release occurred. After the release, DOE and NWP have implemented numerous tem­ porary fan systems to increase the available airflow to the underground. The interim ventilation system (IVS) consists of two additional surface HEPA filtration units tied to dedicated exhaust fans. In 2016 the IVS was placed into service, increasing filtered exhaust by an additional 54,000 cfm raising the total filtered exhaust to 114,000 cfm. The Supplemental Ventilation System (SVS) fan was installed in 2017. This fan is located at the base of the AIS and pushes air into the construction circuit. This increased the ventilation system capacity for the construction and the north area of the repository by 46,000 cfm for a total intake volume of 180,000 cfm. With SVS operating, airflow is routed in series from construction to the disposal circuit. The north circuit is ventilated by using the SHS as an exhaust. The IVS and SVS were installed as tempor­ ary measures until a permanent ventilation system could be engineered and installed.

Figure 5.

Waste container that self-combusted.

618

4 FUTURE MINING AND VENTILATION PLANS AT WIPP The original operational life of WIPP, at the time of the radiologic release, was to finish emplacing waste in Panels 7 and 8 and in the main north/south airways between Panels 1 and 4. At that time this would have ended the operational phase for WIPP. However, after the radiologic event, DOE re-evaluated the amount of TRU waste currently emplaced at WIPP and concluded that the facil­ ity needed to continue operation under the current waste operating permit. This resulted in a major study to determine where new potential panels could be mined, a new surface fan config­ uration and an option for a new shaft to access the facility. DOE developed two primary criteria as a design basis, the first was that for lifecycle planning the facility was to operate to the year 2050. The second criterion was that all exhaust air from the disposal circuit will be filtered with HEPA systems. A series of tradeoff studies were performed to determine where additional panels could readily be developed. The Land Withdraw Act of 1992 set aside 16 sections of land for WIPP. The cur­ rent WIPP underground is in the approximate center of this area. The selected alternative was to mine additional panels to the west of the existing repository. The primary reason for this selection was ease of access to the west for mining, faster development to the next panel for waste handling after Panel 8 is filled, straight forward ventilation system design, and sufficient area to the west for panel development to achieve the design goals. The current design is to include two additional panels (11 and 12) to the west with the ability to add additional panels. Once the mining direction was determined, the engineering team began assessing geotechnical considerations for the new design. The cross sectional area of new airways was slightly adjusted as well as the distance between cross-cuts and distance between parallel airways. The panel design was also adjusted with a greater length between the first room in the panel and the main access west drives. The original mains at WIPP are oriented north/south and consist of four entries. South of the SHS, the second main from the west is a common intake to a regulator that controls air to the disposal circuit. Air not entering this regulator continues in this drive to ventilate the construc­ tion circuit where it returns in the furthest west main entry. The third entry from the west is the disposal intake and the fourth is the disposal exhaust. In the new design a second disposal intake is included. The reason for this is for ground control considerations. With a single dis­ posal intake, any ground control function interrupted waste handling operations. With two intakes this challenge is virtually eliminated. The mining of salt at WIPP is primarily by full face continuous mining machines. No dust suppression is applied during mining. Because of this, exhaust airflow from the construction circuit can have significant airborne salt dust. In the original ventilation system design, this dust reported to the surface and was exhausted directly to atmosphere. The new design is to have the air pass through HEPA filters prior to discharge. In addition to dust, the exhaust airflow can have significant moisture. This moisture is introduced during certain times of the year. The facility is in the high desert of New Mexico and monsoonal moisture, typically coming from the south in the summer, is very common. As the air upcasts the exhaust shaft, this moisture condenses resulting in significant water discharging through the fans. Water and salt dust could cause plugging of the pre-filters in the HEPA filtration system. This would result in constant filter change outs. To minimize the impact of salt and moisture, a surface Salt Reduction Building (SRB) is being constructed. The purpose of this system is to remove mois­ ture through a demister and salt dust through filtration before it reaches the HEPA filter system. 5 MINE VENTILATION DESIGN The first ventilation analyses assumed a new surface salt and water reduction system upstream of a HEPA filter system. These structures will replace the system shown on Figure 3. The cur­ rent design included Panels 11 and 12 as shown on Figure 6. Modelling was performed incorp­ orating natural ventilation pressure (NVP) for both cold (winter) and hot (summer) conditions. Air expansion/contraction was accounted for in each intake and exhaust shaft. Airflow requirements were based on diluting diesel exhaust, namely from the small mine 619

Figure 6.

Ventilation system design for future WIPP panel development.

trucks, and the equipment used to transport the TRU waste from the Waste Shaft to the active emplacement panel room. In addition, there are ventilation requirements for each waste handling room of 20 m3/s (42,000 cfm). It is assumed three rooms are actively ventilated in a normal operation. The construction area requires 35.4 m3/s (75,000 cfm) in the active mining zones. Ventilation for the main shop and SDI is assumed at 28.3 m3/s (60,000 cfm) and the waste shaft station at 11.8 m3/s (25,000 cfm). Bulkhead resistance values were based on his­ toric values at WIPP. Airway resistances were based on measured friction factors and pres­ sure/quantity measurements. Regulators were incorporated with fixed quantity branches and adjusted until the airflow design values given above were achieved. The base model for WIPP determined that the maximum exhaust airflow required is 245 m3/s (520,000 cfm) at a static pressure of 2.5 kPa (10.0 in. w.g.) (as measured at the collar of the exhaust shaft. All system losses downstream of the collar were calculated by HVAC engineers assigned to specify the primary fans). 6 SAFETY SIGNIFICANT CONFINEMENT VENTILATION SYSTEM Once the total flow and required delivered pressures were determined, design engineers began to develop concepts for a salt reduction system and parallel HEPA filtration system. The salt reduc­ tion system consists of six independent moisture and salt reduction units. Each unit is comprised of a CFT GmbH demister upstream of a CFT GmbH Model HTKK 1/1500-2S dry scrubber. A Spendrup 152-091-1800-C-1 600 hp fan is connected to each unit. The fan motors are on vari­ able speed drives for airflow control. Each unit is capable of handling a throughput of 120,000 cfm. The fans are fitted with an outlet silencer. Figure 7 shows a SRU during factory acceptance testing. The salt reduction units are in a separate building called the Salt Reduction Building (SRB). The SRB has one primary intake duct tied to a manifold to each SRU. A common exhaust plenum takes the air back to the intake to the Safety Significant Confinement Ventilation System (SSCVS) in the New Filter Building (NFB). Photos of an installed SRU is shown on Figure 8. 620

Figure 7.

Demister, filter and fan system for a single SRU (at the time of factory acceptance testing).

Figure 8.

Photos of installed filter and fans.

The SSCVS was designed based on each HEPA filter unit being capable of 12.7 m3/s (27,000 cfm). To achieve a design flow of 255 m3/s (540,000 cfm) this resulted in a minimum of 20 filter units being available. The system was designed with 22 units. A filter unit is comprised of four in line filters. The first filter is a MOD filter followed by a HIGH filter then two HEPA filters. With this train contamination can be removed from the airstream. The basic design of the SRB and SSCVS system is shown on Figure 9. This figure also shows a typical HEPA filter unit. Mechanical engineers calculated the pressure losses through the HEPA filters, ducts, dampers and elbows and computed a maximum possible pressure for the SSCVS fans. A total of six fans were selected for the design, each fan capable of 70.8 m3/s (150,000 cfm) at a design pressure of 7.0 kPa (28.1 in. w.g.) (at an air density of 0.96 kg/m3 [0.06 lbm/ft3]). At maximum flow four fans would be in operation each at approximately 63.7 m3/s (135,000 cfm). A Clarage 86.625 SW5731 CHS Fan was selected for the project. A 746 kW (1000 hp) motor on a variable speed drive is included in the fan design. Because the fans are internal to the SRB, the fan motors will be water cooled. Because the resistance in each HEPA filter unit will vary, depending on the amount of load­ ing, the airflow through each unit will be controlled by a flow damper connected to an airflow sensor on the intake to the HEPA filter unit. The damper will not modulate continuously, rather be set and compared by personnel operating the system. Differential pressures will also be monitored across each filter bank within a HEPA unit. Increases in pressure for a constant flow will indicate when filters should be changed out. 621

Figure 9.

Isometric view of SRB and NFB system.

The SRB system is designed to have the internal fans simply overcome the losses of the SRU and duct work. Flow through the system will be balanced with the SSCVS total flow. In the event of a radiological release, the SRB will be isolated from the system with high efficiency dampers and all airflow from the underground will be directed through the SRB bypass duct. The surface SRB and SSCVS are currently under construction. Completion is expected by late 2024. The existing fan system connects the north side of the exhaust shaft. The SSCVS con­ necting duct will be to the east. Figure 3 shows the direction of the SSCVS in relation to the existing fan system. Prior to connecting the SRB and SSCVS to the Exhaust Shaft, a series of tests will be performed on the system. The 4.3 m (14 ft) diameter horizontal duct from the Exhaust Shaft location will be fitted with a regulator. The regulator will be partially closed to simulate the resistance required to support the repository. Testing will include various fan oper­ ations, including full flow, filter unit configurations combined with fan operation, duct integrity, damper integrity, filter flow adjustment other operational considerations. It is expected that by conducting a comprehensive testing program the operation once the system is connected to the Exhaust Shaft will be simplified. 7 NEW SHAFT DESIGN Historically at WIPP extracting salt is a challenge as the SHS is only 3.04 m (10 ft) in diameter and consists of a single 7.25 tonne (8 ton) skip. The shaft and its internal components were constructed in the early 1980s. To expect this infrastructure to last the life-cycle time to 2050 would be a challenge. After the SSCVS study was commenced, DOE embarked on a second study to evaluate the addition of a new shaft. A series of alternative studies were performed to evaluate the location of a new shaft, the size and the long term use of the shaft. The results of these studies concluded that a new 8.54 m (28 ft) diameter shaft would be constructed due west of the AIS and SHS. The location is shown on Figure 6. This location was selected since the new mining panels were being constructed to the west and the location was readily accessible on surface. The shaft diameter was selected to accommodate a possible new hoisting system in the future. This new hoist would replace the SHS for primary salt removal and provide personnel and equipment access to the repository. A large conveyance for personnel and equipment with two salt skips are envisioned for the shaft. The shaft will be an intake to the repository providing air to the construction and disposal circuits. A surface intake fan is to be installed on the shaft. The design has two parallel fans installed on the surface connected to a plenum (see Figure 10). Only one fan will operate at 622

a time with the second being a fully redundant fan. The fans are designed to intake over 235 m3/s (500,000 cfm). With a forcing fan on the collar of the shaft, the airflow from the construction panel will course to the AIS where it will exhaust on surface. This flow path sig­ nificantly reduces the salt dust loading passing to the Exhaust Shaft. To facilitate airflow exhausting the AIS a new horizontal duct with stack was constructed on the collar of the AIS. This design will allow the dust to settle away from adjacent buildings and parking areas. Figure 11 shows the exhaust duct and stack on the AIS. The shaft project includes underground excavations around the shaft station and two airways connecting to the primary north/south drifts. The shaft project completion date is expected to be midyear 2024.

Figure 10. shaft.

Intake fan with plenum to new

Figure 11.

Exhaust stack at the AIS.

8 SUMMARY This paper describes two significant ventilation upgrades at the WIPP facility. The SSCVS con­ sists of a 255 m3/s (540,000 cfm) exhaust fan system. The entire exhaust system will be capable of being filtered for any radiological release. This filter system includes HEPA filters. The result­ ing design is the largest HEPA filtration system in the world. A total of six 746 kW (1000 hp) fans are included in the design with a maximum of four fans operating in parallel. A water and salt dust removal system are being installed upstream of the HEPA filtration units to reduce exhaust contaminants from reaching the filters. This system incorporates six demister and dry dust filters each tied to a vane axial fan. It is designed to filter the entire maximum flow through the system. In the event of a radiological release underground, the water and salt dust filtration system will be turned off and the air diverted directly to the SSCVS. In addition to the primary exhaust system upgrade, WIPP is installing a new 8.54 m (28 ft) diameter intake shaft with surface intake fans. This system will provide upwards of 236 m3/s (500,000 cfm) of intake air to the underground. A new exhaust path will be established to take salt laden air from the construction system and route it to the air intake shaft which will be an exhaust. On the surface of this shaft is an exhaust duct and stack to minimize salt dust over adja­ cent buildings and the parking lot. The system can be adjusted to operate without the intake fans on the new shaft since the SSCVS can handle the maximum expected airflow at the facility. With this system it is expected that WIPP can maintain operations for construction of Panels 11 and 12 and beyond should DOE elect to do so. In addition, the system will be robust and can be managed to adapt to any normal or off-normal operating condition. A future project to install a new hoisting system in the new shaft will further enhance WIPPs operational flexibility. REFERENCES CFT GmbH, Beisenstraße 39-41, 42964 Gladbeck, Germany, www.cft-bmbh.de.

623

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Ventilation network optimization: Realizing energy savings while promoting worker health and safety A.K. Ngcibi & M. Mochubele University of the Witwatersrand

F.S. Bergh Howden Africa

ABSTRACT: The COVID-19 pandemic has certainly shocked the world and made everyone to rethink, rework and reimagine the workplace of tomorrow. The 21st-century engineer will have to work smarter to manage the changing environment and employ technological tools to aid them in the quest for zero harm. This paper outlines the concept of remotely and autono­ mously supplying ventilation to underground workings, commonly referred to as Ventilation-On -Demand (VOD). For ease of reference, the presentation refers to VOD. However, the proposed system should rather be viewed as a Ventilation Optimization Solution. Through remote moni­ toring of working conditions, these systems can be used extensively to improve workplace condi­ tions, whilst offer financial benefits. This is made possible by employing sensors for real-time monitoring integrated with hardware to address sub-standards in real time. 1 INTRODUCTION Globally, the mining industry is grappling with a number of issues which include shortage of electricity supply and associated rising costs. A casing point is a South African metalliferous mining industry which is heavily reliant on electricity and highly sensitive to energy cost. This begs the question: how will mines be able to supply qualitative and quantitative air in the underground working in a cost-effective manner? Historically, South African mines would afford to adopt the ‘brute-force’ ventilation strat­ egy which means ventilating all the underground workings at full capacity 24/7 a day regard­ less of activity or the underground environmental conditions. The year 2007 can be marked as a defining point of paradigm shift as South Africa started experiencing unprecedented electri­ city tariff increase and load-shedding. This is one of the major reasons why the industry is looking at reducing the electricity cost by focusing research on energy efficient projects such as Ventilation On Demand (VOD). The guiding principle of energy cost saving using VOD is based on the cubic relationship between power and air quantity (P α Q3). This means any reduction in quantity will result in significant energy savings. This research will also go a long way in addressing the economic sustainability and indirectly reducing carbon footprint by using energy derived from fossil fuel efficiently. The scope of this paper is on reducing the cost of ventilation using VOD system in under­ ground metalliferous mines and also improving the health and safety. This is a literature review paper, and the layout is made up of research background, litera­ ture review, health and safety, conclusion and recommendations. The literature review pro­ vides views/conclusions from different literature illustrating how real time monitoring, in conjunction with the remote control of ventilation devices can support mines on their journey to zero harm. In addition, the literature review provides the feasibility study which provide the cost of VOD and the payback period. DOI: 10.1201/9781003429241-64

624

2 RESEARCH BACKGROUND AND CONTEXT The dynamic underground environment, and the dependence on human intervention, are major challenges in supplying adequate fresh air to areas where employees are required to travel and/ or work. This can result in a deterioration of workplace conditions and is a constant threat to the health and safety of employees. Covid-19 forced skilled and experienced personnel to work from home, which increased the risk of mismanagement of the ventilation systems. Inadequate supply of fresh air has detrimental effects on the health and safety of the employees. Mine ventilation professionals must empower themselves with the available tools to ensure workplaces remain safe and risk-free to the health of employees. Mines need to become more innovative and utilise the available infrastructure to distribute the air in a more optimised and autonomous fashion by adopting optimisation systems. This approach will minimise the dependency on human intervention and make it possible for skilled and experienced personnel to work from home. 3 LITERATURE REVIEW Historically, South Africa (SA) had very competitive electricity tariffs. The abundance and rates did not warrant an energy efficient approach towards mine ventilation. This notion is supported by Johnson (2012) in his study which concluded that South African mines are pro­ duction driven where energy efficiency is not part of the mind set and culture. Resultantly, the forbearers of mine ventilation in SA could afford to design ventilation sys­ tems based on the worst-case scenario as highlighted by Kocsis, et al (2003). The approach was to “set and forget” as the brute-force ventilation system would always ensure more than the required air supply to the underground workings (Mochubele, 2014). As early as 2003, Kocsis et al recognised the potential of VOD systems for energy savings, with Johnson in 2012 and Mochubele (2013) all reaffirming the same notion. The health and safety improvements of the systems are neglected in the literature, with only claims to ‘main­ tain’ health and safety being made. The literature review focused on VOD implementation and the associated health and safety improvements and energy savings. 3.1 Ventilation on demand implementation The widely accepted VOD definition is that it is a process of adjusting mine airflow to supply qualitative and quantitative airflow in the required area, in real-time, to maintain environmental standards. VOD is a system of sensors and remotely controlled hardware capable of real time adjustments to mine ventilation equipment, such as fans and regulators, to meet underground atmospheric quality requirements for personnel and equipment based on their status and location. VOD systems set out to improve health and safety, reduce energy usage and increase gen­ eral productivity. Even so, not all mines are lining up for the technology. This is due to the high cost associated with installing monitoring and communication systems and the subse­ quent maintenance required to keep the system functional and effective (Wallace, Prosser, & Stinnette, 2014). This cost of implementation is rarely ever stated in literature reports, as most always quantify the savings but not the capital nor operation cost of the systems. 3.1.1 VOD levels of control The high capital and maintenance costs along with the steep learning curve associated with VOD systems has led to mines considering implementing VOD strategies usually take a phased approach. The approach is based on the levels of control of the system shown in Figure 1. Mines can phase-in VOD systems by starting with the basic Level 1 and 2 control, and then scaling up the system over years to the more sophisticated Level 5 control, which yield best results in terms of energy cost saving. Level 1 offers the most basic and simple VOD system, allowing for manual remote control of the ventilation equipment. This is widely used in mines for control over primary fans. 625

Figure 1.

VOD levels of control (Howden, 2018).

Energy savings can be achieved through the system when used with variable speed drives (VSDs) on the fan motors instead of guide vanes which reduce quantity by reducing the fan rotational speed. For health and safety improvements in the mine, it is worthy to invest in level 3 of controls which require extensive monitoring of the environmental conditions and level 4 which track the locations of personnel and machines. Level 5 refers to completely autonomous dynamic ventilation-on-demand systems which incorporates time-based, qualitybased and event-based ventilation optimization. 3.1.2 Remotely controlling VOD systems Nick Holland, former CEO of Goldfields spoke of the possibility of mining at 5km (Mining Weekly, 2020), citing that it is only possible through remote or autonomous mining due to the potential dangers and risks. Mochubele (2014) reported on the potential benefits of automa­ tion, with electrical cost reduction being top of the list. The added and most significant benefit of remote mining is the removal or reduction of personnel from the high-risk underground environment. VOD implementation is a preliminary step towards the goal of ‘ZERO HARM’, which in reality can only be achieved through the complete removal of personnel from the hazardous underground environment. VOD systems allow for tracking of the positions of personnel and machines and real-time monitoring of the airflow properties and gas remotely. Events are scheduled and setpoints are created for desired airflow quantities and velocities, temperatures, percentage gases in the general atmosphere etc. Computer programmes such as Ventsim Control compute the air demand and determine the necessary fan and regulator set points required to achieve these environmental standards. Hardware such as Programmable Logic Controllers (PLC’s) or remote Inputs and Outputs (I/ O’s) enable the remote control of equipment such as fan starters, Variable Frequency Drives (VFD’s) or actuators. 3.2 Health and safety implications The Mine Health and Safety Act (MHSA) of 1996 (No. 29) was the first stand-alone health and safety legislation for mines in South Africa. Under the MHSA, the government, mining industry (employer) and workers are all required to promote health and safety in the work­ place (Mohapi & Zarske, 2018). This means that mining companies are under pressure to pro­ actively protect employees from all present and future risks. Implementation of new technologies such as VOD in underground mines is one way that companies can keep ahead of the health and safety curve. 3.2.1 Early entry examinations Efficient underground ventilation and cooling forms a critical part of sustainable underground mining. Ventilation provides oxygen while removing dust and post-blasting fumes and is used in conjunction with cooling for temperature control (Wallace, Prosser, & Stinnette, 2014). 626

There is a need for re-entry examinations in underground mines after blasting, to ensure the environmental conditions are within acceptable limits before re-entry into working areas. These are usually done manually with hand-held devices, increasing the risk of exposure to the observer, to not just noxious fumes, but seismicity risks as well. Through VOD, examinations can be done remotely by checking the real-time gas readings in the system. This reduces the risk of exposure and hence improve the health and safety position of the mine. A hypothetical mine has a re-entry standard of 45 minutes, with early-entry examinations conducted manually by the re-entry team. The re-entry team involves the shift supervisor (miner); the crew leader; and two safety representatives. Hand-held gas detection instruments are used for this and the following hazards are identified for this process: i) Short term overexposure to CO ii) Short term overexposure to NO2 iii) Flammable gas accumulation and fire risk The time-weighted average occupational exposure limits (TWA.OEL) represent the minimum acceptable exposure over a longer period i.e. 8 hours. The Short Term Exposure Limit (STEL) refer to acceptable exposure over a short period, usually 15 minutes, as long as the timeweighted average is not exceeded. Please see the TWA and STEL values in Table 1 below. Table 1. Occupational Exposure limits (MHSA, 1996). Gas

TWA.OEL

STEL.OEL

CO

30 ppm

100 ppm

The acute effects of CO exposure include but are not limited to mental confusion, vomiting, loss of muscular coordination, loss of consciousness and death. High exposures, above 70ppm, leads to victims rapidly becoming mentally confused and can lose muscle control and consciousness without experiencing any mild symptoms CO takes longer than all the gases to dissipate to the OEL.TWA levels. This occurs after more than 45 minutes from blasting according to the figure below. With a re-entry time of 45 minutes in place at the mine, the probability of acute exposures of the re-entry team is quite high.

Figure 2.

Blast fume dissipation trend (Brake, 2015).

627

Potential accumulations in muck piles and lingering accumulations caused by ineffective and damaged ventilation infrastructure poses a risk to the re-entry team. The likelihood for this event to occur is probable (35% - 65%) Figure 2 illustrates the dissipation trend of blast fumes. 3.2.2 Real time monitoring VOD systems incorporate real-time gas and airflow sensors which allow for monitoring of environmental conditions in real-time. Whilst reducing the risk of exposure, it also increases the monitoring capacity of underground environmental conditions. The real-time monitoring and remote/ autonomous control means changes in environmental conditions can be dealt with timeously by either directing more airflow to the affected areas or evacuating personnel from the areas in time. This improves the health and safety of underground mines. This goes a step further than simply “maintaining” health and safety standards, but suggest that these solutions can improve workplace conditions. A practical example is the VOD solution deployed at Eleonore mine in Canada. The solu­ tion automates the operation of all ventilation equipment including main fans, auxiliary fans and airflow regulators (Howden, 2021). The system includes 30 ventilation monitoring stations (VMS) which determine the quantity and quality of airflow at various positions in the mine. Each VMS is equipped with airflow, carbon monoxide (CO), nitrous dioxide (NO2) and propane (C3H8) sensors. The system incorporates a mine-wide tracking system for locating personnel and vehicles. Both personnel and vehicles are issued with radio frequency identification (RFID) tags that connect to any one of the 250 zone-based WIFI access points. The RFID location data provides enhanced safety and management of the movement of personnel and vehicles, while the VMS system improves the monitoring of environmental con­ ditions in the production zones. The combined data allows for ventilation requirements for each ventilation zone to be calculated by the VOD logic software, based on personnel and vehicle location and environmental conditions. 3.2.3 Remote control and management of a ventilation system In a post-Covid-19 world, more and more skilled employees work from home, and mines will have to follow suit to attract skilled employees to the sector. The danger is that mines will always need to remain ventilated and environmental conditions maintained to ensure safe, healthy and sustainable production. Furthermore, as mines operate on multiple shifts there’s a risk that conditions can deterior­ ate during periods where ventilation experts are not present e.g. exposure to pollutants and/ or fire or influx of gas during a night shift that could endanger the lives of hundreds of persons. Managing a ventilation system remotely will empower mine ventilation engineers to carry out their legal duties even if being on-site is not possible. A control system will assist by log­ ging all changes to the ventilation system, complete with alerts based on varying levels of severity. This function will also fulfil the legal requirement of an early warning system. In add­ ition to this, a control system allows the mitigation of sub-standard conditions in real-time. Naturally, the system and associated hardware needs to be maintained to ensure it operates satisfactorily. The hardware kits installed have a proven record of operating in the harsh underground environment, but care and maintenance are critical in ensuring sustainable and efficient operation thereof. 3.3 Energy cost savings Airflow in underground mines is driven and controlled through pressure differentials created by the use of fans. The associated cost of ventilating comes from the energy dissipated by these fans to create a pressure differential. The absorbed power is based on the relationship between mine resistance, fan pressure and airflow. Costs can be reduced either by using smal­ ler fans where possible or reducing the speed of the fan. VOD allows for the manoeuvrability of fans operation point (speed) based on demand to improve energy efficiency, whilst main­ taining quality of mine environmental conditions. 628

The implementation of a Ventilation Optimisation Solution can: • • • •

Schedule events for reduced flows during shift-change and breaks; Reduce ventilation in zones where no production is occurring Reduce ventilation rates in zones where ventilation exceeds the requirements Monitor and control air quality within regulatory limits to maintain an adequate level of gases, airborne pollutants and temperature • Post-blast ventilation management to minimise blast clearing times. Through these controls, the mine can achieve cost savings and improve health and safety through increased monitoring and control over the airflow to respond to conditions in realtime. A case study at Newmont’s Eleonore Gold Mine in Canada, shows the mine has achieved • a 43% reduction in mine heating costs • a 56% reduction in underground ventilation electricity costs; and • a decline in surface ventilation electricity costs of 76% through VOD implementation (Gleeson, 2019). 3.4 Ventilation optimisation study at a South African mine A feasibility study performed at a South African operation yielded the following results. The operation could not establish additional vertical shafts to cater for their life of mine commit­ ment due to surface topography. The scope of work was to revise the ventilation design and strategy to ensure the ventilation commitments comply with the mine standard. The future mine will use seven (7) working ventilation districts for production. Of these 7 sections: • 3 Sections are production sections with an average intake velocity in the last through road required to ≥ 0.75m/s (2 roadways). • The 4 remaining sections are non-production sections with an average intake velocity in the last through road required to ≥ 0.15m/s (2 roadways). 3.4.1 Optimisation of primary surface fans The study proposed fitting VSDs to the Primary Fans. This will enable the primary fans to be operated at different speeds, and in doing so will enable the mine to capitalise on periods where the ventilation demand is reduced. In 2023, the two primary fans need to run at 100% speed during production (P), however, savings can be realised during non-production (NP) hours. From 2024 onwards, due to the underground activity, primary fans need only run at 85% of their capacity during production time to meet the demand. Tables 2 and 3 below illustrate this: Table 2. Characteristics of primary fans to be integrated 2023. Power Quantity (kW)

Starter Type

Flow (m³/s)

Fan Speed Percentage: Production time (P) and Non- production time (NP)

2

VSD

176

P:100%

700

NP: 1 x fan Off; 1x fan 57%

Table 3. Characteristics of primary fans to be integrated 2031. Power Quantity (kW)

Starter Type

Flow (m³/s)

Fan Speed Percentage: Production time (P) and Non- production time (NP)

2

VSD

176

P:85%

700

NP: 1 x fan Off; 1x fan 57%

629

3.4.2 Ventilation optimisation of underground ventilation network and booster fans The proposed design, with air crossings placed strategically within the underground work­ ings, eliminated the need for underground booster fans. Ventilation control devices comprise of air crossings and automated regulators in the return airway of each section. The running cost of the booster fans add to annual savings for the mine. Figure 3 below depicts an auto­ mated louvered regulator used to automatically regulate airflow flowing into the ventilation districts.

Figure 3.

Automated Louvered regulator.

3.4.3 Ventilation optimization of auxiliary fans The following VOD strategy concerning auxiliary fans was proposed. All auxiliary fans are switched off for the two (2) hours of non-production time i.e. travelling time, vehicle inspec­ tions etc. The auxiliary fans in the non-productive sections remain switched-off for the dur­ ation of the shift. Auxiliary fans in the non –production sections will be switched on automatically when personnel and/ or machinery enter sections, or when real time monitors indicate sub-standard conditions. 3.5 Summary price: VOD solution The system has a capital cost of USD 1.8 million which translates to just over ZAR 28 million. This is based on the primary fans being retrofitted with VSDs and the auxiliary fans being retrofitted with auxiliary links. The inability to retrofit would impact the capital costs of the system as new fans would be required for the system to work optimally. 3.6 Financial analysis Figure 4 graphically represents the financial analysis conducted for the VOD system outlined in this case study: From Figure 4, the system payback period is shown to be just below 2 years. This is the point when the net present value (NPV) is equal to zero, and it increases to just below USD 13 million in 2031. Concerns about the VOD system being capital intensive should be 630

Figure 4.

Financial analysis USD.

discarded. VOD systems have short payback periods of about 2-3 years as they pay for them­ selves through the energy cost savings achievable through this dynamic control of airflow. These systems have proven to be critical not just for energy savings but for improving the health and safety in underground mines as well. 4 HEALTH AND SAFETY IMPLICATIONS VOD systems allow mines to view environmental condition data in real-time, whilst logging changes in the ventilation system and notifying officials of sub-standards for localized evacu­ ations. The system also allows the user to implement controls in real-time that will safeguard worker health. VOD systems help limit the time of exposure to airborne pollutants by preventing a buildup of contaminants in production zones through re-directing noxious gases directly to return airways. This will aid the operations existing controls in creating an environment that will go a long way to achieving “ZERO HARM”. 5 CONCLUSION Ventilation-on-demand consists of different levels of control i.e. manual-based, time-based scheduling, quality-based (requires flow and gas readings), event-based (requires tracking and tagging); and autonomous systems. Adoption of the systems is heavily swayed by the high capital and learning curve costs associated with the higher levels of control. Case studies on mines with VOD systems show that they have the potential to reduce energy costs while main­ taining, and improving, health and safety conditions in underground mines. A feasibility study carried out in one of the South African mines demonstrated that although VOD system has high capital cost of R28 million the payback period is less than two years. Real-time monitoring of environmental parameters is crucial and something that is being employed extensively within the mining industry. A VOD solution will enable mines to act and mitigate exposure to airborne pollutants in real-time. Based on the information contained in this report, it is evident that VOD solutions are advantageous. Mines need to consider util­ ising these technologies to improve workplace conditions and reduce energy costs. 631

6 RECOMMENDATIONS Based on the conclusion and information contained in this technical paper, it is evident that VOD or Ventilation Optimization solutions are advantageous to improving health and safety. The removal of employees from dangerous work places or conditions, such as re-entry exam­ inations, shows to be the most effective way of improving health and safety. The current system only removes the re-entry team from the hazardous area during re-entry examinations. The team will still go and work in the very same areas for an entire shift where environmental conditions can be readily reduced to a sub-standard. Mines should explore modernisation techniques such as remote or automated mining solutions. VOD systems are an enabler to such solutions according to Mochubele (2014). REFERENCES Brake, R., 2015. A Review of Good Practice Standards and re-Entry Procedures after Blasting and Gas Detection Generally in Underground Hardrock Mines. Virginia, 15th North American Mine Ventila­ tion Symposium. Dominguez, C. R., Martinez, I. V., Pena, P. M. & Ochoa, R. A., 2019. Analysis and evaluation of risks in underground mining using the decision matrix risk-assessment (DMRA) technique, in Guanajuato, Mexico. Journal of Sustainable Mining, 18(1), pp. 52–59. Gleeson, D., 2019. International Mining. [Online] Available at: https://im-mining.com/2019/11/07/how dens-eleonore-ventilation-on-demand-solution-wins-award/ [Accessed 25 August 2020]. Howden, 2021. VentSim Control: Eleonore. [Online] Available at: https://www.howden.com/en-us/case studies/ventsim-control-eleonore#pdf [Accessed 16 November 2021]. Howden, V. S., 2018. Levels of control within a ventilation optimization system. Montreal: Howden Ventsim Solutions. Johnson D, F. C., 2012. An overview of energy efficiency in South African Hard Rock Mines, ??: Research gate. Kocsis, C. K., Hall, R. & Hardcastle, S. G., 2003. The integration of mine simulation and ventilation simulation to develop a ‘Life Cycle’mine ventilation system. South African Institute of Mining and Metallurgy, pp. 223–230. MHSA, 1996. Mine Health and Safety Act 29 of 1996. [Online] Available at: Mine Health and Safety Act [Accessed 21 August 2020]. Mining Weekly, 2020. Getting the best out of mining starts with policy certainty, says Gold Fields’ Hol­ land. [Online] Available at: https://www.miningweekly.com/article/getting-best-out-of-mining-startswith-policy-certainty-holland-2020-08-20 [Accessed 21 August 2020]. Mochubele, E. M., 2014. Effects of Increasing Rejection Temperatures on Electricity Demand for Venti­ lation and Cooling in Automated Metalliferous Underground Mines, Johannesburg: University of the Witwatersrand. Mohapi, G. & Zarske, R., 2018. Health and Safety in South African Mines: A Best Practice Report, Frankfurt: Southern African-German Chamber of Commerce and Industry. New Jersey Department of Health, 2000. Hazardous Substance Fact Sheet: Nitrogen Dioxide. New Jersey, New Jersey Department of Health and Senior Services. Oxford Reference, 2021. Quantitative Research. [Online] Available at: https://www.oxfordreference.com/ view/10.1093/oi/authority.20110803100357649 [Accessed 16 November 2021]. US Consumer Product Safety Commission, 2021. Carbo-monoxide Questions and Answers. [Online] Available at: https://www.cpsc.gov/Safety-Education/Safety-Education-Centers/Carbon-MonoxideInformation-Center/Carbon-Monoxide-Questions-and-Answers [Accessed 16 November 2021]. Wallace, K., Prosser, B. & Stinnette, J. D., 2014. The practice of mine ventilation engineering, California: Mine Ventilation Services, Inc. Fresno. Yahnke, K., 2021. Risk Assement Matrix. [Online] Available at: https://www.i-sight.com/resources/riskassessment-matrix/ [Accessed 16 November 2021].

632

Ventilation planning and design

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Design highlights for Agnico Eagle’s Macassa 4 Shaft primary ventilation systems K. Boyd, T. Mehedi & D. Witow Hatch, Canada

M. Pinheiro-Harvey Agnico Eagle, Canada

ABSTRACT: The 4 Shaft Project represents a step change for ventilation capacity at Agnico Eagle’s Macassa Mine in Kirkland Lake, Ontario. This paper outlines the process design, equipment selection, and experience in procuring new primary ventilation equipment for this project. Equipment includes new 64 MMBtu/hr natural gas direct fired mine air heat­ ers with integrated silencers; twin 600 hp vane axial fresh air fans; and twin 3,000 hp centrifu­ gal primary exhaust fans with demisters and high-performance silencing measures. This paper presents the description of obstacles tackled during the engineering design of the primary mine ventilation equipment, which included space limitations on surface and noise concerns due to proximity to local communities.

1 INTRODUCTION In Northern Ontario, Canada, the town of Kirkland Lake has seen seven gold mining oper­ ations in the 20th century. After a successful exploration campaign in the early 2000’s, Kirk­ land Lake Gold re-established gold mining operations at the Macassa mine, which included using the existing 3 Shaft as a service, skipping, and access shaft, and the 2 Shaft as secondary egress. The ventilation circuit was driven by underground booster fans operating in parallel within several of the historic tracked mining levels. The most recent configuration had fresh air downcast in 3 Shaft and exhaust in 1 Shaft and 2 Shaft. After successful exploration at depth, a new access shaft, called 4 Shaft, was approved, and began construction in 2018. To accommodate the increased fresh air capacity, a pair of new exhaust raisebores to surface were also approved for construction in parallel with the shaft. Now merged with Agnico Eagle, the company’s 4 Shaft Project is now in the final stages of construction in 2023. This shaft represents an opportunity to renew and improve mine access in proximity to the deep SMC and Lower North orebodies including ore handling systems; mine services and utilities; and ventilation. Additional fresh air will support mining activities at depth which promises to improve safety and productivity. To support production before the four-year shaft construction period was complete, the SMC230 project was developed at Macassa in 2020 which included three new underground fan stations and refit of three others which increased fresh airflow to operations at depth – achieving the namesake of 230 kcfm airflow to the SMC orebody. New ventilation infrastructure to leverage the new 4 Shaft and raisebore airways was div­ ided among two projects: 1. 4 Shaft Project, which would complete the 4 Shaft Heaters and 160L Booster Fans. 2. Raisebore Exhaust Fans Project, which would construct the new fans for the raisebores.

DOI: 10.1201/9781003429241-65

635

2 PROCESS DESIGN The life-of-mine (LOM) ventilation schematic for Macassa can be seen in Figure 1. The venti­ lation circuit is negatively induced via the twin raisebore exhaust fans with the fresh air coming from 3 Shaft and 4 Shaft. Multiple booster fans are installed mine-wide to direct the airflow through various mining areas. The design for the 4 Shaft Heaters, 160L Booster Fans, and Raisebore Exhaust Fans com­ menced with a studies phase to develop the basis of design including tradeoffs to confirm equip­ ment selection and configuration. New ventilation equipment was sized to maximize the capacity of the new shaft and exhaust raisebores – to provide the highest degree of future flexibility. Key process conditions forming the basis of the 4 Shaft Heater, 160L fans and the Raise­ bore Exhaust Fans are summarized in Tables 1 and 2. Table 1.

Summary of process parameters for 4 Shaft Heaters.

Parameter

Units 3

m /s °C db °C db MMBtu/hr -

Nominal Air Flow (Volumetric) Design Inlet Temperature (Min) Design Temperature Rise Thermal Capacity Heating Fuel Heater Type

Table 2.

Value 307 -44 49 60 Natural Gas Direct Fired

Summary of process parameters for 160L Fans.

Parameter

Units 3

m /s °C db kPa -

Nominal Air Flow (Volumetric)- combined Inlet Air Temperature Mine Static Pressure Fan type Airflow Isolation (Maintenance) Fan Configuration

Value 307 5 to 36 1.04 Vane Axial – Arrangement 4 Louver Dampers (actuated) Two Fans in Parallel

A schematic of the process design for the 4 Shaft Heater and 160L fans are shown in Figure 2. The layout of various components was designed with the following key design aspects: • 4 Shaft Heaters would include an acoustically insulated steel plenum above the Vent Shaft with removable roof panels to allow future access for fan replacement. • Relatively low-pressure application – with the 160L fans only having to overcome pressure losses up to 4 Shaft Station on 160L. • Fan flow slightly higher than 4 Shaft downcast demand to upcast 30 kcfm to 4 Shaft collar which provides some make-up air for the headframe. Prior to selection of equipment, a noise study was carried out to define the maximum allow­ able sound power level to be specified for the heater house intake that will meet community noise limits at the nearest receptors and determine whether the underground fan discharge levels can meet occupational noise exposure limits at 160L and at the Shaft collar work area. Two fan operation conditions were evaluated: • Both fans operating under typical conditions • Maintenance scenario where one fan runs while the second is being maintained.

636

Figure 1.

Macassa life of mine ventilation schematic in long section view process design.

Table 3. Summary of process parameters for Raisebore Exhaust Fans. Parameter

Units

Nominal Air Flow (Volumetric)- combined Inlet Air Temperature Mine Static Pressure Fan Type Fan Configuration

3

m /s °C db kPa -

Value 330 15 7.69 Centrifugal Two Fans in Parallel

The Raisebore Exhaust Fans are planned to be installed on surface, which is the preferred location, however there are concerns regarding the noise impact on the adjacent community due to limited mitigation opportunities with a surface installation. An underground installa­ tion to alleviate these concerns was considered, but ultimately the decision was made to go forward with a surface installation due to technical and economic considerations. This decision required modifications to the design of the fan system to include an independ­ ently supported acoustic enclosure surrounding the fan casing and all ductwork before the existing silencer, and the addition of a second silencer of similar performance in series. Both the considered underground and chosen surface layouts are shown in Figure 3. The VFD cutouts for the 160L fans are located downstream of the fans (as shown in Figure 4 Figure 4). To protect the equipment inside the cutout from high circulating/turbulent air movement induced by the passing airstream, placement of deflection shields was suggested based on CFD analysis. A sketch of predicted velocities and net forces are shown in Figure 5. Figure 5 Finally, a standard operating procedure (SOP) was developed for accessing 160L to promote safety given airflow velocities around the 160L shaft station. Demisters were selected for the Raisebore Exhaust Fan application to avoid contamination of local soils with dirty water droplets. The outlet side of the fan was selected for the demister loca­ tion, which reduced ductwork and footprint while also offering less velocity transition. In other words, the exhaust air only requires one expansion transition to reach a low velocity required for the demister, outlet airflow silencer, and stack. Turbulence from the fan discharge moving through the evasé was a concern and CFD modeling was conducted to demonstrate to project teams and demister vendor that the velocity profile was suitable for equipment integrity and performance. 637

Figure 2.

Schematic showing 4 shaft heaters and 160L Booster fans in elevation view.

Figure 3. Section view of exhaust fan concept in underground chamber (a) and Selected configuration showing elevation view of one fan on surface (b).

3 EQUIPMENT SELECTION AND PROCUREMENT Equipment selection trade-off studies, which included workshops with Operations and vendor consultations, were conducted by the project team to inform specification packages for competi­ tive bidding. After clarifications, the technical bid evaluation process identified compliant

638

Figure 4.

Isolation view of the underground 160L Booster Fans with VFD location on right hand side.

Figure 5. CFD study results indicating high circulating flows in left view and shield walls with improved conditions on the right.

offerings and a single vendor, Howden, a Chart Industries Company, was selected to provide all the packages. Figures 6 and 7 show the general arrangement of the 4 Shaft Heaters and 160L fans. To main­ tain access to the ventilation shaft for material movement, namely for 160L fan components, the heater house is equipped with removable roof panels. Each burner will be connected to its individ­ ual valve train mounted on a steel rack inside a control room adjacent to the heater house. To remove any sort of lifting hazards, the four burner valvetrains were divided into pairs and located along opposite walls in the control room. The control room will house a control panel with a burner management system (BMS) as required by the Technical Safety Standards Authority (TSSA), and an integrated PLC system which will allow the Operators to monitor and control the system either directly from a local operator station (panel mount touchscreen) or indirectly from any HMI location elsewhere on the site. Split baffle-type silencers were selected for the 160L fans so that a low-profile lifting trolley can be installed within the drift excavation which will be used for initial installation and main­ tenance needs. 4 DESIGN LESSONS AND DISCUSSION Highlights from the 4 Shaft Heaters include:

639

Figure 6.

General arrangement for 4 Shaft Heater in Long Section across the Burner Room.

Figure 7.

Elevation view arrangement showing one of the two of 160 Fan Systems.

• To fit the four 15-foot-long grid burners, a two-story heater building was used, making all burners fully accessible without working at heights. • Due to strict community noise limits, noise from combustion blowers and natural gas com­ bustion in the burners was a concern. A field study at an existing heater was conducted so that a realistic sound power level for these components could be incorporated into the noise control monitoring. • Like the 3 Shaft Heaters, the burners and burner control systems utilized similar compo­ nents including flame rods for combustion monitoring, Maxon APX forced combustion air grid burners, and Honeywell 7800 series flame safety relay. • The plenum connecting the Vent Shaft and the heater was constructed as modular acoustic­ ally insulated panels to reduce site construction time. Features of the 160L Fans are summarized: • AMCA Arrangement 4 fans were used to minimize operating complexity. • Low voltage (600 V) fan motors were selected to reduce both motor and VFD cost. • Louvre-type rectangular isolation dampers were used for maintenance isolation – including an open/close actuator.

640

• Motor brakes were included to allow any free-wheeling fan to be brought safely to a stop. Disc-type brakes were used with rotors mounted to shaft extensions on the motor nondrive end. • Fan preventative maintenance was prioritized, and a Bently Nevada 1900/65 monitor was selected to monitor a horizontal vibration probe on each fan bearing. With two vibration probes per fan, a single four-channel monitor can monitor both fans. • Temperatures of both motor bearings and windings are also monitored to improve main­ tenance planning. • Variable speed drives were selected for fan modulation and are expected to have a long ser­ vice life in the relatively clean conditions on 160L. • With the relatively unobstructed and short airway these fans are responsible for, stall sen­ sors were not required. Features included on the Raisebore Exhaust Fan are as follows: • Full condition monitoring instrumentation for 3,000 hp fan motors and fan bearing/shafts. • Modulation with variable speed drive and option to also modulate with parallel-blade inlet dampers. The inlet dampers also serve to isolate a fan system during maintenance or repair. • Fitted wash bars were allocated for the stainless steel, blade-type demisters. • Extensive noise engineering was applied to optimize noise control measures for the fans – which includes a combination of baffle-type outlet silencers, acoustic lagging on exposed ducting surfaces, and acoustic enclosure around the fan body.

641

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Upper Keel mine ventilation strategy at Eagle mine K. Boyd, D. Witow & C. McGuire Hatch, Canada

C. Gobbs Lundin Mining Corporation, Eagle Mine, USA

ABSTRACT: Lundin Mining operates the Eagle Mine in the Upper Peninsula of Michigan USA. The mine is preparing to construct the new Upper Keel zone – another high-grade nickel orebody in proximity to the existing mine. This paper provides an overview on several aspects of the ventilation design and strategy including the options for development and con­ struction, life-of-mine operations, and the innovative ability for on-shift blasting during the construction phase while maintaining operations in the existing Eagle and Eagle East zones.

1 INTRODUCTION Eagle Mine is owned by Lundin Mining Corporation, a global diversified base metal company headquartered in Toronto. Eagle Mine started operations in 2014 with an expected life of mine into 2027. The hard rock underground mine is located in the Upper Peninsula of Mich­ igan, USA. Nickel/copper ore is mined using both stoping and cut-and-fill mining methods for the Eagle and Eagle East ore bodies. A satellite ore body known as the Upper Keel deposit is located approximately 800m east of the existing portal and approximately 430m below surface. 2 MINE-WIDE VENTILATION The primary ventilation circuit for the mine uses two fresh air intakes: one through the mine portal that is equipped with fresh air fans and direct-fired heaters and a Fresh Air Raise (FAR) that is equipped with direct-fired heaters. Fresh air from the main decline and FAR is distributed to Eagle by two 700hp exhaust fans located on the surface, at the collar of a Return Air Raise (RAR). The Eagle East ventilation uses a branch circuit that is powered by two 400hp fresh air fans to “push” fresh air from the main decline and FAR to the mining zone. All exhaust air from both zones is collected on the 265L return air drive (RAD) via exhaust raises and exhausted out the RAR. Additional details can be found in (Witow et al., 2019). The Upper Keel deposit is found adjacent to the main decline, at a shallower depth than the other existing zones. During initial construction of this main decline, muck bays were con­ structed to allow the staging of development muck. Muck Bays 2 and 4 will be re-purposed to establish development drives to access the top and bottom portions of Upper Keel, respect­ ively. The proposed ventilation plan will utilize fresh air from the upper portion of the main decline drawn off at Muck Bay 2, circulated through this zone, down a series of FARs, and returned to the Upper Keel ramp to the main decline at Muck Bay 4. A schematic of the over­ all proposed ventilation network is shown in Figure 1.

DOI: 10.1201/9781003429241-66

642

Figure 1. Eagle mine overall ventilation schematic outlining the ventilation of Eagle, Eagle East and Upper Keel zones.

3 UPPER KEEL LIFE-OF-MINE VENTILATION: A ventilation schematic showing the ventilation of the Upper Keel is shown in Figure 2. Air for the Upper Keel will be drawn from the Main Decline at Muck Bay 2 by the new Upper Keel Booster Fans. Fresh air will be delivered to the Upper Keel through a set of internal fresh air ventilation raises. All the air will exit into the Upper Keel haulage ramp and exhaust air will travel up the decline to rejoin the remaining air from the portal that has not been directed to the Upper Keel.

Figure 2.

Upper Keel Life-of-Mine ventilation schematic.

643

The ramp section between Muck Bay 2 and 4 will have a reduced flow rate due to the Upper Keel ventilation supply “bypassing” this area. Appropriate measures will ensure portal air flow is greater than the Upper Keel demand resulting in s sufficient air flow to support trucking in this ramp. The proposed Upper Keel ventilation design draws and returns air to the same branch of the primary ventilation circuit, creating an independent circuit resulting in: 1. No impact on fan flows or pressures within the Upper Keel. 2. Exhaust air from Muck Bay 4 will recirculate back up to Muck Bay 2 and through the Upper Keel zone. 3. No increase or decrease in air flows or pressures in the Eagle primary circuit or fan stations. To control the ventilation within the Upper Keel zone and to maintain a simple control schematic, all the fresh air from the raises will be directed into the Upper Keel ramp on the lowest mining level by constructing timber bulkheads which will block airflow from each of the internal levels with the exception of the lowest one. Air from the lowest level will enter the Upper Keel ramp which will upcast toward the main decline. Auxiliary fans will be placed in the ramp to ventilate each active level. 4 UPPER KEEL BLAST MODE VENTILATION As seen in the schematics above, exhaust air from Upper Keel rejoins the fresh air circuit for Eagle and Eagle East. On-shift blasting could allow development crews to maximize develop­ ment rates in Upper Keel, where crews can blast at the face as soon as drilling and loading are done and not wait until the end of the shift blasting window. Therefore, to enable on-shift blasting in the Upper Keel, consideration must be made to isolate blasting gases to avoid fumes re-entering the main decline and downstream workplaces. To achieve this, a single 54” steel exhaust duct is proposed to pick up the exhaust air from the Upper Keel decline connec­ tion in Muck Bay 4 and direct it to the 265 RAD and through the airlock doors. Planning for blast gas clearing was informed by previous testing (Carriere et al. 2017). This configuration is chosen based on the historical development of both Eagle and Eagle East declines that were completed with a similar single 54” steel duct. Thus, the headroom in the ramp is sufficient so that the duct can be installed without impacting haulage activities. The configuration of the fans at the 265 RAD airlock doors (Figure 3) was used during the development of the Eagle East decline and should be familiar to site operations.

Figure 3.

Eagle East development duct arrangement at 265L RAD Airlock.

644

A ventilation schematic showing the ventilation flows during blast clearing and the general airflow is shown in Figure 4. Following an on-shift blast, the Upper Keel Blast Fans located at the 265L RAD will be turned on to draw all the air from the Upper Keel straight to exhaust. To ensure that no blasting gases enter the main decline, the fresh air booster fans at the top of the raise in Muck Bay 2 will be turned down and a door installed in Muck Bay 4 will be closed. This will completely isolate the Upper Keel and ensure that any fresh air drawn through the portal and down the Eagle decline will be free of blasting gases, allowing produc­ tion activities to continue in either Eagle or Eagle East without delay.

Figure 4.

Upper Keel Blast mode ventilation schematic.

The 80 kcfm airflow volume during blast mode is based on the practical amount of air that can be induced at the 265L RAD by two 54”200 hp fans without the use of inline booster fans. Having no in-line booster fans in the exhaust duct will keep all parts of this system under negative pressure, ensuring that any leakage at duct joints or from mobile equipment damage to a duct would not result in the escape of blast smoke. Instrumentation for remote monitoring and control will be used for the Upper Keel Booster Fans, Blast Door, and Upper Keel Blast Fans to ensure the safe and stable operation of the blast mode system. In the event of any upset conditions, alarms will be programmed in the mine’s control system, and appropriate system interlocks, including the shutdown of Upper Keel Booster Fans, which can be programmed for additional layers of redundancy. 5 UPPER KEEL ACCESS VENTILATION STRATEGY Initial development from Muck Bay 4 into the Upper Keel decline is proposed to occur under a traditional forcing auxiliary ventilation arrangement until the steel exhaust duct would not be damaged by concussion and fly rock. These initial development blasts will require end-ofshift blasting allowing blast smoke to clear through the mine’s primary ventilation circuit, without impacting Eagle and Eagle East operations. The blast mode exhaust duct can be installed after development has commenced; at which time the remaining development can utilize on-shift blasting. A fresh air fan will be required 645

to direct air to the working face and flush blast gases back up the ramp, where the exhaust duct will collect all exhaust and direct it through the 265L RAD Airlock to the RAR. The fresh air fan should be installed in an overlapping configuration with the intake of the exhaust duct to prevent recirculation of blast smoke into the intake duct (Figure 5). The ducting over­ lap distance has not yet been calculated – and may be limited by available space.

Figure 5.

Exhaust overlap ventilation in Muck Bay 4.

After completion of the first internal FAR, the Upper Keel Booster Fans will be installed, and flow-through ventilation will be instituted. The construction of the Blast Door should be scheduled to complete prior to the start-up of the Upper Keel Booster Fans. 6 MAJOR VENTILATION EQUIPMENT CONCEPTUAL DESIGN HIGHLIGHTS The Upper Keel Booster Fan will be the driver of the ventilation circuit that begins at the top of the FARs in Muck Bay 2. An example sketch of a potential installation for the fans is shown in Figures 6 and 7. The fresh air booster fans will be controlled by a variable frequency drive (VFD). The VFD will be networked to allow remote control of the fans from the surface. This will be critical to ensure fan turndown is achieved during blast clearing. In the blast clearing mode, the Upper Keel ventilation flow is matched to the capacity of the exhaust duct at 80 kcfm. Based on this balance, all of the blasting fumes should be cap­ tured before entering the main ramp. As an additional measure of safety, a ventilation door (blast door) is proposed for Muck Bay 4 which will prevent an imbalance of flows and prevent any fugitive blasting gasses from reaching the main ramp. This door will normally be open during production and will not affect truck haulage. In blast mode, the doors will close and all air from the zone will be captured in the exhaust duct. The blast door will be required only after the ramp and Upper Keel FAR are connected and a flow-through circuit is made. An overhead steel sectional door, comparable to those currently used on 265 RAD, is pro­ posed for the blast door. The door system is shown in Figure 7. A higher-pressure door has been selected to withstand upset conditions in the case the Upper Keel Booster Fans speed is 646

Figure 6.

Design Sketch for upper keel booster fan station - Plan View.

Figure 7.

Design sketch for upper keel booster fan station – Section view.

647

not correctly reduced or if there is an exhaust duct fan failure. This pressure would be too high for a rubber roll-up door. 7 CONCLUSIONS Having the Upper Keel zone located adjacent to the primary mine access presents an oppor­ tunity for a simple ventilation circuit both in construction and operations phases. Challenges for minimizing blast clearing time can be overcome with a reduced-flow blasting mode in the zone and an isolated exhaust duct connection to the return air raise. If Operations elect to pursue on-shift blasting, the addition of a blast isolation door can provide a failsafe configuration. REFERENCES Witow, D., Gobbs, C., Lam, J. & Harris, W. (2019) Lundin Eagle East Project – Mine Ventilation Update. 17th North American Mine Ventilation Symposium. Canada. Carriere, R., McGuire, C., McLaren, E., Witow, D. (2017) Studying Operational Improvements in Blast Gas Clearing Using Ventilation Control. 16th North American Mine Ventilation Symposium. Colorado, USA.

648

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Empirical and numerical investigation on the optimal length of eddy airflow in dead-end tunnel R. Morla & J. Chen Norton Gold Fields, Kalgoorlie, WA, Australia

S. Karekal & A. Godbole School of Civil, Mining, and Environmental Engineering, University of Wollongong, Australia

P. Tukkaraja Department of Mining Engineering and Management, South Dakota School of Mines, Rapid City, SD, USA

P. Chang WA School of Mines: Minerals, Energy and Chemical Engineering, Curtin University, Australia

ABSTRACT: This paper uses vortex flow modelling to find the optimal length of eddy air­ flow in dead-end gallery using air velocity, pressure and diesel particulate matter (DPM) simu­ lations and field investigations. Computational fluid dynamics (CFD) modelling conducted for four different dead-end crosscut lengths (10 m, 15 m, 20 m, and 25 m), three different crosscut angles (45°, 90° and 135°) and different air velocities in adjacent galleries revealed that a distinct vortex flow develops in the dead-end crosscut. Results indicated that an eddy airflow revolved in a curved form, while the air velocity and pressure decreased towards the centre of the vortex and DPM concentration increased towards the centre of the vortex. The eddy airflow influence distance in a dead-end crosscut depends on the crosscut angle and air velocity in the adjacent gallery. If the air velocity in the adjacent gallery is one m/s and the crosscut angle is 90°, eddy airflow ventilates up to 20 m from the entrance. Though the air velocity in the adjacent gallery is 4 m/s, eddy airflow is not ventilating the crosscut after 30 m from the crosscut entrance. The eddy flow distance is lower in obtuse-angled crosscuts than the acute-angled crosscuts.

1 INTRODUCTION Dead-end workings are common in underground areas during development and production mining (García-Díaz et al., 2019). Dead-end crosscuts adjacent to the main airways, like decline, incline, tunnel etc., are commonly used in the subsurface to install substations, pumps, stockpiles, etc. (Morla et al., 2021, Toraño Álvarez et al., 2002). Figure 1 shows a part of a mine near portal. From the figure, it can be observed that the substation, stockpile, mono pump, fuel bay, underground (u/g) toilet, sump, u/g store, etc. are in dead-end tunnel/drive/ cuddy/crosscuts/gallery. These areas are located adjacent to the main air gallery and most of them are ventilated by eddy airflow. If the dead-end cuddy length is too long, eddy air flow may not reach the endpoint of the gallery, causing accumulation of higher temperatures or dust or DPM concentrations. Diesel-powered vehicles are sometimes operated in unventilated areas like parking cuddies, stockpiles, etc. (Morla et al., 2020a). As the lower airflow in dead-end workings, it takes longer time to disperse exhaust fumes, including DPM. The DPM particles are made of carbon,

DOI: 10.1201/9781003429241-67

649

elemental (EC) and organic (OC), ash, sulphates and nitrates (Chang et al., 2019), particles are small in size (nm) and their densities are low (0.3 gm/cm3 to 1.2 gm/cm3) (Bugarski et al., 2004), they do not tend to settle quickly under their own weight (Morla et al., 2020a, Morla et al., 2019). Prolonged exposure to DPM causes adverse health effects (Ristovski et al., 2012, AIOH, 2013, Morla and Karekal, 2017, Morla et al., 2018, Chang and Xu, 2017).

Figure 1.

A part of a mine with details.

As per Australian mine regulations (AIOH, 2013, MDG, 2008, WHSR, 2022) to minimize the adverse health effects, the 8-hour time-weighted average exposure of mine personnel to EC is limited to 0.1 mg/m3, total carbon (TC), 0.16 mg/m3 and diesel particulate (DP), 0.2 mg/ m3. As per the Western Australian mines regulations, to dilute diesel emissions, each location where a diesel engine operates must be ventilated with a minimum airflow of at least 0.05 m3/ s/kW of the engine capacity (WHSR, 2022). Western Australian mine regulations also indicate (WHSR, 2022), an average air velocity of at least 0.3 m/s is to be supplied at all working areas of the mine where persons work. The design of eddy airflow ventilated dead-end tunnels is important to maintain regu­ latory ventilation requirements. Mapping airflow patterns in dead-end crosscuts will better understand air distribution and develop the optimum design of eddy flow venti­ lated tunnels. In this paper, studies were conducted using air velocity, pressure and DPM field and modelling investigations to find the optimum length of eddy airflow in a dead-end crosscut. 2 MATERIALS AND METHODS 2.1 Details of the field experiment Field experiments were conducted in an underground mine to study the airflow and DPM distribution in dead-end crosscuts. The airflow of the intake airway was controlled by a regulator located at the return side of the airway. A calibrated ‘Airtec’ real-time DPM monitor was used for this field study (Janisko and Noll, 2008, Khan, 2017). The flow rate of the DPM monitoring instrument was adjusted to 2.83 × 10-5 m3/s (1.7 L/min). During the investigation, the air velocity in the adjacent main gallery was maintained at 3 m/s. Air vel­ ocity was measured with an anemometer. The experimental site was a 100 m long tunnel with a rectangular cross-section (width 6 m, height 2.7 m), with a dead-end crosscut 10 m long and at 90° with the main gallery. For the DPM experiment, a loader was operated in dead-end crosscut up to DPM concentration reached to 820 µg/m3. Though the regulatory limit of EC is 100 µg/m3, the experiment was conducted at a high concentration to precisely monitor the DPM dispersion with time.

650

2.2 CFD modelling CFD studies have been used in the mining industry to solve various critical ventilation issues (Morla, 2013, Morla et al., 2013, Morla et al., 2015, Chang et al., 2020a, Chang et al., 2020b, Chang et al., 2019a, Chang et al., 2019b). The commercially available CFD package ANSYS Fluent was used for this DPM modelling (Tanguturi et al., 2013). The CFD simulations were carried out in the sequence of steps outlined in the following paragraphs. Figure 2 shows a CAD model representing the experimental gallery and the dead-end cross­ cut. The sampling points are at 5 m from the dead-end face. Figure 2 shows the mesh gener­ ated for the volume of the experimental gallery. The finer mesh cells were also used adjacent to gallery walls, with seven layers of cells accommodated in the boundary layers. The compu­ tational domain and mesh consist of almost half a million tetrahedron-shaped computational cells. Convergence and mesh-independent studies were conducted in this modelling. The residual RMS error value is 10-4, and the domain has imbalances of less than 0.01. The min­ imum wall Y+ value for the model is 1. The boundary conditions of the model were considered as having an intake air velocity of 3 m/s at 300 K of dry bulb temperature and the initial DPM concentration in the dead-end crosscut was considered as 820 µg/m3. The transient state modelling was used to model DPM concentration dis­ persion in a dead-end gallery. For this modelling, the diameters of the DPM particles are considered between 1e-9 m to 1e-7 m with a mean diameter of 1e-8 m. DPM particles were treated as inert mater­ ials and the Rosin-Rammler diameter distribution was used. For physical models, the spherical drag law was used as a drag parameter. As the air velocity is low (below 4 m/s), the standard k-ε turbulent model was used. For stochastic tracking, a discrete random walk model with 10 tries and a 0.15-time scale was used. The intake air and DPM are considered as two different phases. The EulerianLagrangian approach is used whereby the gas phase (air) was solved using the Eulerian approach, and the particle-phase (DPM) was tracked using the Lagrangian approach.

Figure 2.

Computational domain and sampling points locations.

The airflow in the tunnel was treated as a turbulent flow. The standard k-ε model, which is one of the most commonly used turbulent models, is applied to simulate the airflow. The model transport equation for k was derived from the exact equation, while the model transport equa­ tion for ε was obtained using physical reasoning and bears little resemblance to its mathematic­ ally exact counterpart. In the derivation of the k-ε model, the assumption is that the flow is fully turbulent, and the effect of molecular viscosity is negligible. As the mine air is considered as fully turbulent flow, the k-ε model is valid for mine air. The turbulent kinetic energy, k, and its rate of dissipation, ε, are obtained from the following governing equations (ANSYS, 2013): 651

Figure 3.

Meshed model of dead- end gallery.

where Gb is the generation of turbulent kinetic energy due to buoyancy, and Gk is the produc­ tion of turbulent kinetic energy due to the mean velocity gradient. The particle flow is modelled using the Euler-Lagrange approach, where particle properties are studied along the particle flow path (Chang et al., 2019a). These models define particle flow by considering the various forces that act on the particle (Thiruvengadam et al., 2016). The forces commonly encountered are the drag force between the fluid and the particle, the lift force, the virtual mass force, etc. The fluid phase is treated as a continuum by solving the Navier-Strokes equations, while the dispersed phase is solved by tracking a large number of particles, bubbles, or droplets dis­ persed through the calculated flow field. The dispersed phase can exchange momentum, mass, and energy with the fluid phase. DPM particles are tracked using the Lagrangian method in the discrete phase, and the particle or droplet trajectories are computed individually at speci­ fied intervals during the fluid phase calculation. The force balance equation relates the particle inertia with the forces acting on the particle and can be written as

Where ~ F is an additional acceleration (force/unit particle mass), FD ~ u force per unit particle mass, and

� �! up is the drag

Here ~ u is the fluid phase velocity, �! up is the particle velocity, µ is the dynamic viscosity of the fluid, ρ is the fluid density, ρp is the density of the particle material, and dp is the particle diameter. Re is the relative Reynolds number, which is defined as:

The additional forces induced on the particle due to the fluid surrounding the particle due to growth in the boundary layer is called the virtual mass force and is given by 652

Where Cvm is the virtual mass factor with a default value of 0.5, the fluid and the particle are coupled together mathematically in the form of slip velocity. 3 RESULTS AND DISCUSSIONS 3.1 Results of field experiment and model validation for air velocity Figure 4 shows the results of the airflow distribution at a cross-sectional area in the crosscut. Lower air velocity is at the middle of the gallery and air velocity towards the gallery sides is increasing. Table 1 compares the results of field experiments and modelled air velocities at vari­ ous locations shown in Figure 4 for validation purposes. To measure air velocity in the mine, a rotating vane anemometer was used. This anemometer can measure the minimum air velocity of 0.15 m/s; therefore, the field measurements did not consider air velocity below 0.15 m/s. It can be observed from Table 1 that the base case simulated results were in good agreement with the measured data in most cases. Some discrepancies between the simulated and measured results can be due to the unevenness in the gallery wall surfaces that was not considered while modelling. Overall, the difference varies from – 10% to + 8.6%, which are acceptable. From the modelling studies, it is also observed that center of the vortex is 5 m from the entrance.

Figure 4.

Results of 10 m crosscut velocity contours.

Table 1. Comparison of simulated DPM results with experimental results. Sample Measured air Modelled air Difference Sample Measured air Modelled air Difference point velocity (m/s) velocity (m/s) (%) point velocity (m/s) velocity (m/s) (%) a b c d e

0.58 Error 0.54 0.42 Error

0.63 0.03 0.53 0.43 0.05

8.6 N/A -1.8 2.3 N/A

f g h i

0.51 0.40 Error 0.5

0.50 0.36 0.06 0.48

-2 -10 N/A -4

Note: Difference % is the difference between simulation results and test results and is calculated as (Simulated value– Experimental value)/ experimental value) × 100.

653

3.2 Vortex flow modelling of air velocity and pressure with different crosscut lengths and angles After the initial dead-end crosscut modelling studies were conducted with a base case model, detailed parametric investigations were carried out on the airflow patterns in unventilated dead-end crosscuts of different geometries (crosscut lengths, orientation with respect to the main airflow gallery and air velocities). Figure 5 shows the CFD modelling results of air velocity distribution for dead-end crosscuts 10 m, 15 m, 20 m, and 25 m deep and two different orientations (45° and 135°) were con­ sidered for the 20 m long dead-end crosscut. For this simulation, intake air velocity was assumed to be 3 m/s. The vortex flow patterns are plotted on a plane 1.2 m above the ground. The air velocity in the crosscut is relatively low and the velocities are lower than 0.1 m/s in some areas, as shown in Figure 5. The lower air velocities are at the centre of the vortex and air velocity increases towards the periphery of the vortex. The locations of the centre of the vortex for the 10 m, 15 m, 20 m and 25 m long crosscuts at 90° to the main gallery were found to be at 4.5 m, 8.5 m, 10 m and 10 m from the crosscut entrance. The centres of the vortices in the 20 m long crosscut oriented at 45° and 135° were found to be at 10 m and 6.5 m, respect­ ively, from the crosscut entrance. Figure 6 shows the pressure distribution in a dead-end crosscut. For this modelling con­ sidered dead-end crosscuts of 10 m, 15 m, 20 m (90°), 25 m, 45° (20 m) and 135° (20 m), and an air velocity in the main gallery of 3 m/s. The results show that air pressure is reducing towards the centre of the vortex. There is higher air pressure in the main air gallery; the pres­ sure difference between the main air gallery and the dead-end crosscut (ΔP) is very low and is below 0.4 Pa. A higher air pressure exists in the main air gallery.

Figure 5.

Air velocity streamline patterns in dead-end crosscut.

3.3 Results of field experiment and model validation for DPM Figure 7 shows the results of DPM distributions at a cross-sectional area in the crosscut. Higher DPM concentration is at the middle of the gallery and DPM concentration towards the gallery sides is decreasing. Table 2 compares the results of field experiments and modelled air velocities at various locations shown in Figure 7(b) for validation purposes. To measure DPM concentration in the field, a real-time DPM monitor was used. It can be observed from Table 2 that the base case simulated results were in good agreement with the measured data in most cases. Some discrepancies between the simulated and measured results can be due to the unevenness in the gallery wall surfaces that was not considered while modelling. Overall, the

654

Figure 6.

Pressure streamlines pattern in dead-end crosscut.

difference varies from – 2.5% to + 2.3%, which are acceptable. From the modelling studies, it is also observed that center of the vortex is 5 m from the entrance Figure 7(a).

Figure 7.

Results of 10 m crosscut DPM concentration contours.

Table 2. Comparison of simulated DPM results with experimental results. Sample point

Measured DPM concentration (µg/m3)

Modelled DPM concentration (µg/m3)

Difference (%)

a b c

551 650 554

564 648 540

2.3 0 -2.5

Note: Difference % is the difference between simulation results and test results and is calculated as (Simulated value– Experimental value)/ experimental value) × 100.

3.4 Vortex flow modelling of DPM concentration with different crosscut lengths and angles Figure 8 shows the DPM distribution in different dead-end crosscuts. For this modelling, dead-end crosscuts of 10 m, 15 m, 20 m (90°), 25 m, 45° (20 m) and 135° (20 m), an air velocity of 3 m/s in the main gallery, and an initial DPM concentration in the dead-end crosscut of 655

820 µg/m3 were considered (Morla et al., 2020b). In all cases, transient flow modelling studies were used for 180 sec for this modelling. A higher DPM concentration is observed at the vortex centre and DPM concentration reduces towards the ends of the vortex. The DPM con­ centration is higher in obtuse-angled crosscuts than the acute-angled crosscuts.

Figure 8.

DPM concentration streamline patterns in dead-end crosscut after 180 sec.

Figure 9 shows the changes in DPM concentration concerning air velocity inside the deadend crosscut. DPM concentration increases with decreasing air velocity. A high DPM concen­ tration is at the centre of the vortex. Figures 6 and 8 show changes in DPM concentration concerning pressure differences between the main air gallery and dead-end gallery (ΔP). DPM concentration is increases with decreasing air pressure. A high DPM concentration is at the vortex centre, and the lower pressure drop is at the centre of the vortex.

Figure 9. Changes of DPM concentration with air velocity for 10 m, 15 m, 20 m, 25 m, 45° and 135° dead-end crosscut.

656

3.5 Vortex flow modelling with different air velocities in adjacent gallery Investigations have been conducted with different air velocities in dead-end crosscut’s adjacent gal­ lery. For this modelling considers a 90° dead-end crosscut of 50 m long, and air velocities in the adjoining gallery are 0.5 m/s, 1 m/s, 2 m/s, 3 m/s and 4 m/s. Figure 10 shows the results of simula­ tions studies; if the air velocity in the adjacent gallery is 0.5 m/s, vortex-shaped airflow is spread up to 15 m in the dead-end gallery. If the air velocity in the adjacent gallery is 1 m/s, 2 m/s, 3 m/s and 4 m/s, eddy airflow distance in the dead-end gallery is 20 m, 23 m, 27 m and 30 m, respectively.

Figure 10. Air velocity contours in 50 m long dead-end crosscut with different air velocities in adjacent gallery.

4 CONCLUSIONS Investigations concluded that eddy airflow was formed near the dead-end crosscut entrance and revolved in a curved form, while the air velocity and pressure decreased towards the centre of the vortex. DPM concentration increased towards the centre of the vortex. The eddy airflow distance and velocity in a dead-end crosscut depend on the crosscut angle and air velocity in the adjacent gallery. If the air velocity in the adjacent gallery is 1 m/s and the crosscut angle is 90°, the eddy flow is up to 20 m from the entrance. The eddy airflow distance is lower in obtuse-angled crosscuts than the acute-angled crosscuts. Though the air velocity in the adjacent gallery is 4 m/s, eddy flow is not ventilating the crosscut after 30 m from the crosscut entrance. This research provided clear information about ventilation distribution and DPM particle flow patterns in dead-end crosscuts. REFERENCES AIOH 2013. Diesel particulate matter Occupational Health Issues.: (The Australian Institute of Occupa­ tional Hygienists, Inc). Bugarski, A., Schnakenberg, G., Noll, J., Mischler, S., Patts, L., Hummer, J., Vanderslice, S., Crum, M. & Anderson, R. The Effectiveness of Selected Technologies in Controlling Diesel Emissions in an Underground Mine - Isolated Zone Study at Stillwater Mining Company’s Nye Mine. 2004. 1–86. Chang, P. & Xu, G. 2017. A review of the health effects and exposure-responsible relationship of diesel particulate matter for underground mines. International Journal of Mining Science and Technology, 27, 831–838. Chang, P., Xu, G. & Huang, J. 2020a. Numerical study on DPM dispersion and distribution in an under­ ground development face based on dynamic mesh. International Journal of Mining Science and Tech­ nology, 30, 471–475. Chang, P., Xu, G., Mullins, B., Abishek, S. & Sharifzadeh, M. 2020b. Numerical investigation of diesel particulate matter dispersion in an underground development face during key mining activities. Advanced Powder Technology, 31, 3882–3896.

657

Chang, P., Xu, G., Zhou, F., Mullins, B. & Abishek, S. 2019a. Comparison of underground mine DPM simulation using discrete phase and continuous phase models. Process Safety and Environmental Pro­ tection, 127, 45–55. Chang, P., Xu, G., Zhou, F., Mullins, B., Abishek, S. & Chalmers, D. 2019b. Minimizing DPM pollution in an underground mine by optimizing auxiliary ventilation systems using CFD. Tunnelling and Under­ ground Space Technology, 87, 112–121. MDG 2008. Guideline for the management of diesel engine pollutants in underground environments. MDG-29. Morla, R. 2013. Optimum Inertisation Techniques for Blasting Gallery Method. 3rd International Work­ shop on Mine Hazard Prevention and Control,. Brisbane: CSIRO. Morla, R., Balusu, R., Tanguturi, K. & Khanal, M. 2013. Prediction and control of spontaneous com­ bustion in thick coal seams. Coal Operators conference. Wollongong: University of Wollongong. Morla, R., Balusu, R., Tanguturi, K. & Ting, R. 2015. Inertisation options for BG method and optimisa­ tion using CFD modelling. International Journal of Mining Science and Technology, 25, 401–405. Morla, R., Godbole, A., Karekal, S., Bhattacharjee, R. M. & Balasubrahmanyam, N. 2018. Fundamen­ tal understanding of diesel-operated man riding vehicle DPM dispersion–a case study. Journal of Sus­ tainable Mining, 17, 105–110. Morla, R. & Karekal, S. 2017. Diesel particulate matter investigations in underground coal mines. Morla, R., Karekal, S. & Godbole, A. 2020a. CFD simulations of DPM flow patterns generated by vehicles in underground mines for different air flow and exhaust pipe directions. International Journal of Mining and Mineral Engineering, 11, 51–65. Morla, R., Karekal, S. & Godbole, A. 2020b. Investigation of DPM dispersion in unventilated dead-ends using transient flow modelling. International Journal of Mining and Mineral Engineering, 11, 121–133. Morla, R., Karekal, S., Godbole, A., Bhattacharjee, R. M., Nasina, B. & Inumula, S. 2019. Effect of ventilation air velocities on diesel particulate matter dispersion in underground coal mines. Inter­ national journal of mining and Geo-Engineering, 53, 117–121. Ristovski, Z. D., Miljevic, B., Surawski, N. C., Morawska, L., Fong, K. M., Goh, F. & Yang, I. A. 2012. Respiratory health effects of diesel particulate matter. Respirology, 17, 201–212. Thiruvengadam, M., Zheng, Y. & Tien, J. C. 2016. DPM simulation in an underground entry: Compari­ son between particle and species models. International Journal of Mining Science and Technology, 26, 487–494. WHSR 2022. Work Health and Safety (mies) Regulations. In: AUSTRALIA, W. (ed.).

658

Underground Ventilation – Tukkaraja (Ed) © 2023 The Author(s), ISBN 978-1-032-55146-3

Author Index

Abil, A. 115 Addis, J.D. 499, 507, 529 Adhikari, A. 451 Afrouz, S. 230 Agioutantis, Z. 423 Ajayi, K.M. 499, 507, 529 Akhtar, S. 569 Allen, C. 60, 83, 163 Amoah, N.A. 274, 285 Androulakis, V. 541 Animah, F. 221, 230 Arnott, W. 256 Atchley, G. 411

Demirkan, D.C. 377 Dias, T. 411 Diaz, J.C. 423 Dougherty, H. 499, 529 Durieux, D.W. 34, 174, 183 Duzgun, H.S. 377

Bahrami, D. 607 Balusu, R. 363 Barone, T.L. 104 Beck, T. 247 Belle, B. 363 Bergh, F.S. 624 Bhattacharya, S. 3 Bingham, B. 256 Black, D. 439 Bogin, G. 377 Bowling, J. 353 Boyd, K. 635, 642 Brake, D.J. 589 Brickey, A.J. 484 Brown, M. 34 Brune, J. 264, 377 Buaba, J.A. 484 Bugarski, A.D. 104, 120 Burgess, J.L. 95

Gangrade, V. 69, 499 Gendrue, N. 3 Ghosh, A. 313 Gobbs, C. 642 Godbole, A. 649 Greth, A. 221, 230

Calizaya, F. 411 Carvajal-Meza, M.A. 45 Chang, P. 313, 649 Chen, J. 649 Chow, C. 330 Clayton, A. 330 Connot, J. 451

Jayaraman Sridharan, S. 192, 451 Jiang, H. 247 Johnson, J. 411 Jones, T.H. 469 Juganda, A. 377 Juric, J. 439

Das, M.C. 296 Dasgupta, S. 54 De Souza, E. 15

Karekal, S. 649 Kaufman, M. 60 Keles, C. 221, 230

Falk, L.K. 60, 83 Fernandez, J. 341 Finn, J. 322 Fox, J.E. 208, 462 Friend, S. 104 Fuster, E. 595

Habibi, A.A. 25, 120, 135 Harb, C. 305 Hardcastle, S.G. 163 Harris, M.L. 499, 507, 529 Hassanalian, M. 541 Hines, J. 330 Homan, K.O. 120 Hristopulos, D.T. 423 Hummer, J.A. 104 Hurtado-Cruz, J.P. 45 Iqbal, A. 393, 399, 578

659

Khademian, Z. 507, 529 Khaniani, H. 541 Kimutis, R. 529 Klima, S. 247 Klose, F.K.R. 469 Kocsis, C. 256, 411 Kumar, A.R. 285, 399 Lee, T. 104 Li, Z.B. 439 Lindgren, E. 411 Liu, S. 3 Liu, Y. 313 Lotero, S. 541 Lutz, E.A. 95 Madureira, E. 305 Magauiya, N. 115 Mardon, M. 135 Martikainen, A.L. 469 McGuire, C. 34, 183, 642 McLean, B. 129 Medina, A. 305 Mehedi, T. 183, 635 Mochubele, M. 624 Mohit, M. 569 Morla, R. 649 Murphy, C. 256 Nascimento, P. 256 Newman, A. 484 Ngcibi, A.K. 624 Nunes, N. 411 Nurshaiykova, G. 115 Nyqvist, J. 54 Osho, B. 256 Pandey, A. 192 Peña, I. 615 Pilkington, E. 183 Pinheiro-Harvey, M. 635 Pinto, E. 135 Potts, J.D. 239 Poulos, J. 411

Prabhu, E. 462 Prabhu, M. 462 Prasojo, R. 25 Pushparaj, R.I. 393, 399, 578 Qiao, M. 330, 439 Qureshi, A. 115 Raj, K.V. 69 Rajapaksha, R. 305 Rakhimov, D. 115 Rawlins, C.A. 149 Rawson, T. 305 Ray, R.E. 595 Reed, R.J. 95 Reed, W.R. 239 Ren, T. 330, 439 Roberts, J. 330 Roghanchi, P. 296, 305, 541 Rose, C. 135 Rubasinghege, G. 296 Rubeli, B. 129 Sabanov, S. 115 Salami, O.B. 393, 399, 578 Salinas, V.P. 296 Sandink, M. 256

Sarver, E. 221, 230 Sasmito, A.P. 569 Sastry, B.S. 192 Schafrik, S. 423 Schatzel, S.J. 499, 507, 529 Schult, G. 353 Setiawan, I. 25 Shao, S. 541 Shaw, J.K. 60, 163 Sidrow, E. 264 Slouka, S. 264 Soles, J. 384 Stachulak, J. 129 Stewart, C. 25, 515 Tang, W. 384 Taylor, S. 256 Thakur, P.C. 431 Thomas, R.A. 384, 607 Tom, K. 200 Torkmahalleh, M. 115 Tsai, C. 264 Tukkaraja, P. 451, 649 Udofia, E. 484 Uecker, L. 305 Van Diest, J. 353 Van Dyke, M. 499, 529

660

Vanderslice, S. 104 Vanegas, A. 305 Vaze, M. 54 Wallace, K.G. 615 Wang, X. 256 Watkins, E. 499, 529 Watson, K. 129 Wijayanto, K. 135 Witow, D. 34, 183, 635, 642 Xu, G. 274, 285, 393, 399, 578 Xu, M. 569 Yan, L. 550, 559 Yantek, D.S. 550, 559 Young, D. 129 Yuan, L. 384 Zaid, M.M. 274 Zeinulla, A. 115 Zhang, H. 83 Zhang, P. 499 Zhao, Z. 313 Zheng, Y. 239, 247 Zhou, L. 607 Zychowski, K. 296