Mine Ventilation: Proceedings of the 18th North American Mine Ventilation Symposium, 12-17 June, 2021, Rapid City, South Dakota, USA [1 ed.] 9781032036793, 9781032036816, 9781003188476, 1032036796

Table of contents :
Cover
Half Title
Title Page
Copyright Page
Table of contents
Preface
Organizing committees
Sponsors
Auxiliary ventilation
Auxiliary fan selection considering purchasing and energy costs based on fan curves
Case studies of mine ventilation
Ventilation tradeoff study considering switch to battery electric vehicles
Calibration of LKAB’s Konsuln test mine ventilation model using barometer Pressure-Quantity (PQ) survey
Computational fluid dynamics applications in mine ventilation
Investigation of explosion hazard in longwall coal mines by combining CFD with 1/40th scaled physical modeling
Integration of conjugate porous media model into mine ventilation network software
Airflow characteristic curves for amature block cave mine
Scale modeling, PIV, and LES of blowing type airflow in adeep cut continuous coal mining section
Diesel particulate control
Transient-flow modelling of DPM dispersion in unventilated dead-end crosscuts and control strategy using curtain
Estimating diesel particulate matter using apredictive technique for use in underground metal mine production scheduling
Diesel aerosols in an underground coal mine
Preliminary field study to evaluate two thermal correction methods for removing VOC interference from DPM sample analysis
Evaluation of methodology for realtime monitoring of diesel particulate matter in underground mines
Optimizing secondary fan location and air quantity to control DPM recirculation in underground workings using Discrete phase modelling
The results of the evaluations of the instrumentation used to monitor concentrations of DPM
Electric machinery in mine ventilation
Considerations of the overall impact on costs in the Battery Electric versus Diesel-powered equipment selection and trade-off
Mine cooling and refrigeration
Development of energy efficient and sustainable cooling strategies for hot underground mines
Design of mine bulk air cooling systems: Numerical, empirical and experimental validation
Numerical evaluation of aspot cooling technique for underground metal mines
Mine dust moni toring and control
Respirable dust characterization using SEM-EDX and FT-IR: Acase study in an Appalachian coal mine
Characterization of respirable dust samples generated from picks at differing stages of wear
Acomprehensive roof bolter drilling control algorithm for enhancing energy efficiency and reducing respirable dust
An analysis of coal mine lung diseases in the US coal mines
Effects of different shapes of drill shroud on dust control for surface mine drilling operation
Development of areal time respirable coal dust and silica dust monitoring instrument based on photoacoustic spectroscopy
Development of non-regulatory runtime respirable coal and silica dust monitor
Respirable Coal Mine Dust (RCMD) research: Characterization, deposition, monitoring, and mitigation of RCMD
Comparison of mineral content in respirable coal mine dust samples estimated using FTIR, TGA, and SEM-EDX
Mine fans
A simple electro-mechanical controllable pitch fan
Optimizing vertically-mounted jet fans in ventilation shafts for arail overbuild
Mine fires and explosion prevention
Goaf gas distribution profiles near the longwall tailgate area
Challenges and solutions in the development of the VentFIRE mine network fire simulator
Water spray suppression of leaked oil fires: Anumerical study
Anti-caking treated rock dust and its effect on downwind respirable dust measurements
Scaling and flow similarity considerations to develop a1/40th scaled aerodynamic model of alongwall coal mine for methane hazards investigation
Mine gases
Review on development technologies and research status of coalbed methane industry in China
Mine heat
Incorporating ventilation and heat in an underground mine production scheduling model
Spray freezing for mine heating a statistical perspective
Mine management and organization of ventilation
Procedures for mitigating safety risks associated with post-blast re-entry times
An innovative methodology for the assessment and maintenance of e-learning courses using the Community of Inquiry model
Mine ventilation and automation
Numerical modeling of longwall-induced permeability under shallow cover
Occupational health and safety in mine ventilation
Application of machine learning to determine underground hazard location
Activity-based respirable dust prediction in underground mines using artificial neural network
Mine ventilation — studies and environmental interventions for Artisanal and Small-scale Mining (ASM) in the Arequipa region, Perú
Evaluating the effect of coal seam height and mine size on coal workers’ pneumoconiosis prevalence in the United States coal mines, 1986-2018
Ultra fine dusts– BEV fleets and the challenges for ventilation planning
Design and characterization of canopy air curtain for protecting against diesel particulate matter exposures in underground mines
Renewable/Alternative energy in mine ventilation
Waste heat recovery from diesel generator jacket water for mine intake air pre-heating in cold climates: A numerical study
Ventilation monitoring and measurement
Comparison of air velocity measurement techniques
AMS-based accident alarm system: A field case study
Field survey of mine ventilation system for large opening underground mines: Pressure, relative humidity, and temperature
Towards atmospheric monitoring data analysis in underground coal mines
Ventilation network analysis and optimization
Ventilation model calibration from limited survey data
Improving calibration of amine ventilation network using continuous airflow monitoring
Ventilation planning and design
Optimizing ventilation design through discrete event equipment simulation
Evolution of the Henderson mine ventilation system
Underground hard-rock mine ventilation considerations for battery electric mobile equipment
Scale model investigation of ventilation parameters in ablock cave mine
Author index

Citation preview

Mine Ventilation EDITED BY

Purushotham Tukkaraja

MINE VENTILATION

PROCEEDINGS OF THE 18TH NORTH AMERICAN MINE VENTILATION SYMPOSIUM (NAMVS 2021), JUNE 12-17, 2021, RAPID CITY, SOUTH DAKOTA, USA

Mine Ventilation

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

CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business © 2021 selection and editorial matter, Purushotham Tukkaraja, individual chapters, the contributors “Auxiliary fan selection considering purchasing and energy costs based on fan curves” authored by Enrique Acuna-Duhart and Michelle Levesque from Natural Resources Canada; and Juan Pablo Hurtado (non public servants). Copyright to Her Majesty the Queen in right of Canada as represented by the Minister of Natural Resources, 2021. 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. Library of Congress Cataloging-in-Publication Data A catalog record has been requested for this book Published by: CRC Press/Balkema Schipholweg 107C, 2316 XC Leiden, The Netherlands e-mail: [email protected] www.routledge.com – www.taylorandfrancis.com ISBN: 978-1-032-03679-3 (Hbk) ISBN: 978-1-032-03681-6 (Pbk) ISBN: 978-1-003-18847-6 (eBook) DOI: 10.1201/9781003188476

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Table of contents

Preface

xi

Organizing committees

xiii

Underground Ventilation Committee (UVC)

xiii

Review committee

xiii

Sponsors and Exhibitors

xv

Auxiliary ventilation Auxiliary fan selection considering purchasing and energy costs based on fan curves E.I. Acuña, M. Levesque & J.P. Hurtado

3

Case studies of mine ventilation Ventilation tradeoff study considering switch to battery electric vehicles N.D. Wineinger Calibration of LKAB’s Konsuln Test Mine ventilation model using barometer PressureQuantity (PQ) survey S. Gyamfi, A. Halim & A. Martikainen

15

23

Computational fluid dynamics applications in mine ventilation Investigation of explosion hazard in longwall coal mines by combining CFD with 1/40th scaled physical modeling A. Juganda, H. Pinheiro, F. Wilson, N. Sandoval, G.E. Bogin & J.F. Brune

37

Integration of conjugate porous media model into mine ventilation network software P.H. Agson-Gani, L. Amiri, S.A. Ghoreishi Madiseh, S. Poncet, F.P. Hassani & A.P. Sasmito

47

Airflow characteristic curves for a mature block cave mine R. Bhargava, P. Tukkaraja, A. Adhikari, S.J. Sridharan & V.V.S. Vytla

56

Scale modeling, PIV, and LES of blowing type airflow in a deep cut continuous coal mining section A.R. Kumar, K.M. Henderson & S. Schafrik

65

Diesel particulate control Transient-flow modelling of DPM dispersion in unventilated dead-end crosscuts and control strategy using curtain R. Morla, S. Karekal & A. Godbole

v

77

Estimating diesel particulate matter using a predictive technique for use in underground metal mine production scheduling J.A. Buaba & A.J. Brickey Diesel aerosols in an underground coal mine A.D. Bugarski, S. Vanderslice, J.A. Hummer, T.L. Barone, S.E. Mischler, S. Peters, S. Cochrane & J. Winkler

86 95

Preliminary field study to evaluate two thermal correction methods for removing VOC interference from DPM sample analysis P.M. Guse, C. Keles & E. Sarver

105

Evaluation of methodology for realtime monitoring of diesel particulate matter in underground mines A. Habibi, A.D. Bugarski, D. Loring, A. Cable, L. Ingalls & C. Rutter

115

Optimizing secondary fan location and air quantity to control DPM recirculation in underground workings using Discrete phase modelling R. Morla, S. Karekal, A. Godbole, M. Sriwas, J. Jacobs, P. Tukkaraja & B. Chapula

124

The results of the evaluations of the instrumentation used to monitor concentrations of DPM S. Sabanov, A. Zeinulla, J. Brune & M.A. Torkmahalleh

133

Electric machinery in mine ventilation Considerations of the overall impact on costs in the battery electric versus diesel-powered equipment selection and trade-off I.M. Loomis

141

Mine cooling and refrigeration Development of energy efficient and sustainable cooling strategies for hot underground mines J.E. Fox & K.C. Kocsis

153

Design of mine bulk air cooling systems: Numerical, empirical and experimental validation A.F. Kuyuk, S.A. Ghoreishi-Madiseh & A.P. Sasmito

168

Numerical evaluation of a spot cooling technique for underground metal mines N.K. Dumakor-Dupey, S. Arya, D. Atambila & M. Anselmi

177

Mine dust monitoring and control Respirable dust characterization using SEM-EDX and FT-IR: A case study in an Appalachian coal mine J. Gonzalez, N. Pokhrel, L. Jaramillo, C. Keles & E. Sarver Characterization of respirable dust samples generated from picks at differing stages of wear S. Slouka, J. Rostami & J. Brune A comprehensive roof bolter drilling control algorithm for enhancing energy efficiency and reducing respirable dust H. Jiang & Y. Luo An analysis of coal mine lung diseases in the US coal mines E. Rahimi, Y. Shekarian & P. Roghanchi

vi

189 198

208 218

Effects of different shapes of drill shroud on dust control for surface mine drilling operation Y. Zheng, W.R. Reed & J.D. Potts Development of a real time respirable coal dust and silica dust monitoring instrument based on photoacoustic spectroscopy P. Nascimento, S.J. Taylor, W.P. Arnott, K.C. Kocsis, X.L Wang & H. Firouzkouhi Development of non-regulatory runtime respirable coal and silica dust monitor C.C. Harb, R.D. Rajapaksha, X. Moya, J. Roberts, P. Hemp, L. Uecker, T. Rawson & P. Roghanchi

225

233 242

Respirable Coal Mine Dust (RCMD) Research: Characterization, deposition, monitoring, and mitigation of RCMD Y. Shekarian, E. Rahimi, P. Roghanchi, M. Rezaee, K.C. Kocsis & W.C. Su

248

Comparison of mineral content in respirable coal mine dust samples estimated using FTIR, TGA, and SEM-EDX. N. Pokhrel, E. Agioutanti, C. Keles, S. Afrouz & E.A. Sarver

255

Mine fans A simple electro-mechanical controllable pitch fan J. McBain

271

Optimizing vertically-mounted jet fans in ventilation shafts for a rail overbuild R.E. Ray, E. Fuster & S. Lee

280

Mine fires and explosion prevention Goaf gas distribution profiles near the longwall tailgate area R. Balusu, K. Tanguturi & B. Belle

291

Challenges and solutions in the development of the VentFIRE mine network fire simulator C.M. Stewart

300

Water spray suppression of leaked oil fires: A numerical study W. Tang, L. Yuan, D. Bahrami & J. Rowland

309

Anti-caking treated rock dust and its effect on downwind respirable dust measurements M.L. Harris, S. Klima, C.B. Brown, I.E. Perera, J.A. Addis, L.L. Chasko & J. Myers

317

Scaling and flow similarity considerations to develop a 1/40th scaled aerodynamic model of a longwall coal mine for methane hazards investigation H. Pinheiro, A. Juganda, N. Sandoval, F. Wilson, K. Gallagher, G. Bogin & J. Brune

326

Mine gases Review on development technologies and research status of coalbed methane industry in China M. Yang, L. Liu, J. Gao & J. Liu

337

Mine heat Incorporating ventilation and heat in an underground mine production scheduling model E. Smoorenburg, O. Ogunomedede, S. Nichols, A. Newman & G. Bogin

vii

349

Spray freezing for mine heating – A statistical perspective S. Akhtar, M. Xu & A.P. Sasmito

357

Mine management and organization of ventilation Procedures for mitigating safety risks associated with post-blast re-entry times E. De Souza An innovative methodology for the assessment and maintenance of e-learning courses using the Community of Inquiry model J.D. Stinnette & K. Luxbacher

369

379

Mine ventilation and automation Numerical modeling of longwall-induced permeability under shallow cover K.M. Ajayi, Z. Khademian, S.J. Schatzel, E. Watkins & V. Gangrade

389

Occupational health and safety in mine ventilation Application of machine learning to determine underground hazard location D. Bahrami, L. Zhou, Y. Xue & L. Yuan

401

Activity-based respirable dust prediction in underground mines using artificial neural network R. Amoako & A. Brickey

410

Mine ventilation — Studies and environmental interventions for Artisanal and Small-Scale Mining (ASM) in the Arequipa region, Perú J. Brune, K.H. McDaniel, L.J. Zamalloa, M.R. Figueroa & J, F. Enríquez

419

Evaluating the effect of coal seam height and mine size on coal workers’ pneumoconiosis prevalence in the United States coal mines, 1986-2018 Y. Shekarian, E. Rahimi, P. Roghanchi & N. Shekarian

428

Ultra fine dusts – BEV fleets and the challenges for ventilation planning A. Hutwalker, O. Langefeld & J. Kegenhoff

436

Design and characterization of canopy air curtain for protecting against diesel particulate matter exposures in underground mines J.D. Noll, W.R. Reed, J.D. Potts & M.R. Shahan

444

Renewable/Alternative energy in mine ventilation Waste heat recovery from diesel generator jacket water for mine intake air pre-heating in cold climates: A numerical study H. Kalantari, D. Baidya, S.A. Ghoreishi-Madiseh & A.P. Sasmito

457

Ventilation monitoring and measurement Comparison of air velocity measurement techniques K.G. Wallace

469

AMS-based accident alarm system: A field case study L. Fan, S. Liu, N. Gendrue & L. Ang

479

viii

Field survey of mine ventilation system for large opening underground mines: Pressure, relative humidity, and temperature N. Gendrue, S. Liu, S. Bhattacharyya & C. Spellman Towards atmospheric monitoring data analysis in underground coal mines J.C. Diaz, Z. Agioutantis, S. Schafrik & D.T. Hristopulos

489 498

Ventilation network analysis and optimization Ventilation model calibration from limited survey data M.D. Griffith & C.M. Stewart

509

Improving calibration of a mine ventilation network using continuous airflow monitoring L. Zhou, R.A. Thomas, L. Yuan & D. Bahrami

518

Ventilation planning and design Optimizing ventilation design through discrete event equipment simulation K.G. Wallace & P. Labrecque

531

Evolution of the Henderson mine ventilation system K. Kolobe, C. Rutter & N. Shea

538

Underground hard-rock mine ventilation considerations for battery electric mobile equipment N. Saeidi & C. Allen

548

Scale model investigation of ventilation parameters in a block cave mine A. Jha, Y. Pan, P. Tukkaraja & S.J. Sridharan

556

Author index

563

ix

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Preface

This volume contains the proceedings of the 18th North American Mine Ventilation Sympo­ sium held, on a virtual platform, June 12-17, 2021. This symposium was organized by South Dakota Mines, Rapid City, South Dakota, 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 Explor­ ation Divisions of the SME. The purpose of the UVC is to promote engineering interest and technological 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 transactions and proceedings. The UVC offers an affiliation home for SME members and others engaged in the practice of underground ventilation. In these ways, 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 in 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 2004 – University of Alaska Fairbanks – Sukumar Bandopadhyay and Rajive Ganguli 2006 – Pennsylvania State University – Jan Mutmansky xi

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 With the help of the organizing committee, a solid 4-day program was assembled, with tech­ nical papers and presentations organized in 19 sessions. A total of 97 abstracts and 57 final papers were received. Session themes include auxiliary ventilation, case studies of mine venti­ lation, computational fluid dynamics applications in mine ventilation, diesel particulate con­ trol, electric machinery in mine ventilation, mine cooling and refrigeration, mine dust monitoring and control, mine fans, mine fires and explosion prevention, mine gases, mine heat, mine management and organization of ventilation, mine ventilation and automation, occupational health and safety in mine ventilation, renewable/alternative energy in mine venti­ lation, ventilation monitoring and measurement, ventilation network analysis and optimiza­ tion, and ventilation 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

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Organizing committees Underground Ventilation Committee (UVC) Arash Habibi, Freeport-McMoRan, USA Jack Trackemas, PMRD/NIOSH Mining, USA Arun Rai, Genesis Alkali, USA John Bowling, SRK Consulting, USA Craig Stewart, MinWare, Australia Purushotham Tukkaraja, South Dakota Mines, USA Heather Dougherty, PMRD/NIOSH Mining, USA

Review committee Adrianus Halim Agus Sasmito Aleksandar Bugarski Alex Hatt Alex Rawlins Andrea Brickey Ankit Jha Arash Habibi Ashish Kumar Bharat Belle Biswajit Samanta Brian Prosser B S Sastry Calen Beaune Candace Tsai Charles Kocsis Cheryl Allen Craig Stewart Daniel Stinnette Darryl Witow David Loring Davood Bahrami Dean Millar D P Mishra

Emily Sarver Emanuele Cauda Euler De Souza Felipe Calizaya Frank von Glehn Gary Li Gerrit Goodman Guang Xu Guangyao Si Heather Dougherty Hua Jiang Ian Loomis Jacinta Klabenes Jack Trackemas Jianwei Cheng John Bowling Jon Fox Joseph Finn Jozef Stachulak Jurgen Brune Keith Wallace Kray Luxbacher Lihong Zhou Liming Yuan

xiii

Mike Tuck Moe Momayez Paul Meisburger Pedram Roganchi Peter Cain Pierre Mousset-Jones Ramakrishna Morla Rao Balusu Rick Brake Robert Randolph Sampurna Arya Shimin Liu Srivatsan Jayaraman Sridharan Stephen Hardcastle Steven Schafrik Sukumar Bandopadhyay Sunil Vytla Vaibhav Raj Vasu Gangrade Yi Luo

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Sponsors Platinum MECANICAD

SRK Consulting, Inc.

Gold Freeport-McMoRan Accutron Instruments Silver G+ Plastics Maestro Digital Mine

Exhibitors MECANICAD SRK Consulting, Inc. Freeport-McMoRan Maestro Digital Mine Accutron Instruments G+ Plastics

xv

xvi

Auxiliary ventilation

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Auxiliary fan selection considering purchasing and energy costs based on fan curves E.I. Acuña & M. Levesque NRCan-CanmetMINING, Sudbury, Canada

J.P. Hurtado School of Mines, Faculty of Engineering, Universidad de Santiago de Chile, Santiago, Chile

ABSTRACT: Mine ventilation practitioners are commonly tasked with estimating operating set point and selecting an auxiliary fan to supply the air required for underground operations. This static decision could be considered a subset of the larger fan selection of main and auxil­ iary fans, which is dynamic over the course of the mine life. It is common to have multiple fan suppliers that can deliver the same operating point with different purchasing cost and expected efficiencies as indicated in the fan curves. Historically, for the same duty point, the decision has been driven by minimum purchasing cost; however, this may not be the best option when including energy costs over a certain period. This paper presents a methodology to simultan­ eously evaluate the capital and energy cost of a fan based on the fan curve inputs, as a decision support tool for purchasing the cost effective fan. A case study is presented to illus­ trate the fan selection methodology and the impact of the energy cost and the time horizon as compared to just the minimum purchasing cost. Finally, a sensitivity analysis is performed on the key inputs to understand the extent and the validity of the conclusions derived.

1 INTRODUCTION Auxiliary fan purchasing is commonly considered a capital expense (CAPEX) and as such it is usually done in two phases that firstly includes a technical feasibility evaluation and then an economical evaluation. The common process consists of deciding which fans can supply the operating point required (technical) and then from the subset of fans that pass the technical evaluation; the one with the best price (lowest bidding price) is normally purchased. There can be a number of variables included in both the technical aspects such as motor type, efficien­ cies, fan construction, materials, warranties, frequencies of delivery, time of the year delivery for remote sites, and supply management among others. On the economical side there could also be other considerations such as shipping cost, volume discounts, and timelines for deliv­ ery, among others. Nevertheless, the main decision is normally based on minimal capital expenditure. From a short term decision point of view, the decision process only including CAPEX could be labeled as reasonable. However, when considering that the energy costs incurred by a fan working all the time can amount to its CAPEX value in less than a year, considering both CAPEX and OPEX (energy cost) in the economical analysis over a certain period of time, could have merits as an alternative decision process. This idea becomes even more relevant when considering that different fan suppliers can offer better fan efficiencies normally associ­ ated with higher fan capital prices. The question is then how to decide if it is worth it to spend

DOI: 10.1201/9781003188476-1

3

more capital upfront with the expectation of reducing the operating expenditure associated with the energy cost over time. The energy consumption estimation of auxiliary fans is usually done considering that the fan is left running for most of the time. In practical terms this approximation could be con­ sidered fair as auxiliary fans are normally turned on to ventilate dead end headings and usu­ ally turned off by a loss of power across the mining area or a result of a failure. In both cases the fan would be off until it was needed during the same or next shift, or until repaired. How­ ever, fan operation management has been gaining interest in the last decades under the theme of ventilation control systems (VCS) with the ultimate goal of delivering ventilation on demand (VOD). The interest with VCS has been mainly driven by health and safety, flexible production and energy savings. Some of the economical benefits from a VCS system result from energy efficiency in terms of using the fans only when needed with the associated energy cost savings that this strategy generates. However, when considering fan selection in an envir­ onment that could be equipped with a VCS, the OPEX (energy cost) is expected to be lower than a mine without a VCS and as such this could impact the fan selection decision. Energy price plays an important role in calculating the cost of operating auxiliary fans and as such can also play a major role in the fan selection decision. A number of factors can influ­ ence electricity price such as: geographical location, energy sources, regulations, policies, and available incentives. Alternatively if a mine site is remote, it might be required to produce elec­ tricity locally, and could not have an important impact in its resulting price. Hence, there could be significant differences in electricity prices from one mine to another, as well as changes over time which would impact the economical analysis. Therefore, it is not feasible to assess all options for this study, the price will be considered fixed and a sensitivity analysis will be provided to assess the impact of the energy price in the final fan decision. Because energy cost (OPEX) savings is generated over time compared to a baseline, the longer the evaluation period considered, the chances of a higher efficiency fan to offset higher capital expenditure with energy cost savings at a certain discounted interest rate increases. Unfortunately, this has at least two main drawbacks. If the higher efficiency fan has a significantly higher price than the “regular” fan, the additional capital requirement might not be approved due to budget constraints or other reasons, even if it can be offset in a given period of time. Also, due to the harsh conditions to which auxiliary fans are exposed, their life span is commonly not as long as a booster or main fans. Auxiliary fans have a tendency to be considered as consumable supplies, and when one fails another one is taken from storage to replace it to avoid production losses, which could be significantly higher than the fan replace­ ment cost. Thus higher capital expenditure for a more efficient fan may not be appealing for a fan with a shorter life. Without a clear business case, additional capital and improved effi­ ciencies may only be prioritized for main and booster fans. This paper presents a methodology to estimate the OPEX (energy cost) of an auxiliary fan over a certain period of time which can then be combined with the CAPEX to determine which fan presents the lower overall cost as opposed to considering CAPEX only. A sensitivity analysis is performed in terms of three main variables: VCS impact to energy use, energy price, and time period for evaluation. The results are compared over ranges of these different variables to estimate the impact that they could have on the final decision.

2 BACKGROUND When talking about fan power, energy consumption, energy prices and total capital and oper­ ating cost, the orders of magnitude are not always understood by everyone. To introduce the subject, a practical example is presented considering a 1.22 m (48 in) auxiliary fan equipped with a 112 kW (150 HP) motor. It will also be assumed for this example that the CAPEX cost of the fan is $18,000 CAD, which includes supply and transportation to site. The cost pre­ sented is just to illustrate the example. CAPEX cost could also include other components such as installation and commissioning, depending on the site arrangement and if the work is done with internal resources or contractors. Additionally, it will be assumed that the fan would 4

operate at 80% of the motor plate value (89.6 kW) and an annual utilization of 90%. Using these inputs, the fan annual energy consumption is estimated at 706.4 MWh. Furthermore, if an electricity price of 7.5 cents per kWh is considered ($75 CAD/MWh), based on O’Connor (2008) 6 cents per kWh and considering a 25% price adjustment over time, then the associated energy cost for that fan during the year would be $52,980 CAD. This means that energy costs incurred by the mine site will be more than twice the fan capital cost over the first year of use. This demonstrates the importance of the OPEX within the fan operating life, especially con­ sidering that a fan’s useful life can span several years. In addition, if it is considered that different fans can have different performance, usually reflected in their efficiency curves and different energy consumption levels for the same oper­ ating point, then the dynamics between energy and capital costs change, which is the import­ ant dimension considered in this study. Mine auxiliary ventilation systems design and optimization work found in the literature has mostly concentrated on design of auxiliary ventilation systems for long single underground openings (Calizaya and Mousset, 1993), modelling leakage (Duckworth and Lowndes, 2003), measuring leakage to estimate the effects of system performance and cost of ventilation ducts (Millar et al, 2016), optimization of the use of the fan over multiple development periods (Acuña et al, 2010) and operational optimization through ventilation control strategies (Acuña et al, 2016). However, to the best of the authors’ knowledge, a gap remains for fan selection in terms of assessing the impact of fan efficiencies, energy price and fan usage according to different control strategies at the auxiliary fan level. Furthermore, Papar et al (1999) stated the need to “determine some best practice approaches for the selection, installation and operation of mine ventilation systems”to increase energy efficiency. Babu et al (2015) highlighted that “development of cost-effective, reliable, maintenance-free and energy-savings techniques is therefore essential today as the cost of electrical energy is being increased day by day” and concluded that “only a few research works show the optimization of ventilation fan speed and energy consumption in mine”. De Villiers et al (2019) also stressed the need for “maintaining good underground fan practice such as optinal fan selection, ducting design and maintenance is crucial for the effi­ cacy of a mine ventilation network”. It is expected that a more efficient fan could provide energy savings that in some cases could compensate for the additional capital cost. However, the fan efficiency will provide savings only during the time the fan is running. If a ventilation control strategy is con­ sidered, the total energy consumption is expected to be reduced and as such it might impact the opportunity to justify the more efficient fan. Another key component is the energy price as it will influence the amount of time in which an investment in the more efficient fan could be justified. Auxiliary fans can also be made more efficient when coupled to efficient motors. However, that alternative is not included in this study, but the methodology suggested could be applied to consider more efficient fan motors as well.

3 METHODOLOGY The suggested methodology includes the following three steps: 1. Define airflow requirements over time at the working face and at the fan considering a reasonable leakage factor 2. Determine the fan operating point and consumed power by using the pressure drop and fan efficiency 3. Define energy consumption for a given period and establish the baseline (including energy cost) Airflow requirements for a working face have to be estimated taking into account local regulations for each mine site, with the objective of diluting and removing contaminants. In this particular example, and as a simplification of the process, only the diesel horse power will 5

be used to estimate the airflow requirements for the working face. Table 1 presents the activ­ ities and equipment expected to be present at the working face. Airflow requirements in Table 1 are based on 0.06 m3/s/kW of diesel powered equipment, as per article 183.1 (3) of the Ontario Regulation 854. The equipment list considered: • Scoop, 187 kW (250 HP) • Truck, 298 kW (400 HP) • Jumbo or bolter or explosive loader 112 kW (150 HP), only one at the time at the working face Based on the equipment expected to be present based on activity at the working face and the airflow requirement per kW of diesel power, the airflow requirement at the working face is estimated. Additionally a 10% leakage factor is considered for the ventilation duct that directs the airflow from the fan to the face, which is used to estimate the fan airflow requirement over time for each activity. For a fan that stays on all the time, the normal procedure is to provide the highest airflow requirement, which means that the fan airflow requirement would be 32.0 m3/s for this example. With the definition of the airflow requirement at the fan and ventilation duct, the next step consists of defining the fan operating point which is a combination of pressure, airflow and fan efficiency. The Atkinson formula (McPherson, 1993) can be used to estimate the static pressure drop.

The fan pressure was estimated using a round straight duct (300 m) of 1.37 m diameter, with a friction factor (k) of 0.0037 kg/m3, air density ρ of 1.2 kg/m3 and an equivalent shock loss Leq of 10%. The resulting static pressure is 1.67 kPa. With the fan airflow requirement and the estimated pressure drop specified, a fan chart that includes dynamic pressure losses can then be used to define the total pressure and efficiency of the fan operating point. In the example proposed to illustrate the methodology, the estimated fan operating point corresponds to an airflow rate of 32.0 m3/s (67.8 kcfm) and a total pres­ sure of 2.14 kPa (8.6 in wg). Three fans that could satisfy this operating point were used in this example; these fans will be referred to as Fan 1, Fan 2 and Fan 3. Table 2 presents the fan efficiency and capital cost for each. Fan efficiencies are estimated based on the efficiency curves presented in the fan chart. It is good practice to test the fans previous to their installation to confirm that each fan was

Table 1. Example activities, equipment, duration and airflow requirement of mining cycle (12 hours shift).

Activity

Equipment

Duration (h)

Diesel power (kW)

Clear blasted face Ground support Drill and load

Scoop and truck Bolter Jumbo or Loader -

2 3 3

485 112 112

2 2

0 0

Idle Blast and clearance*

Working face airflow requirement (m3/s)

Fan airflow requirement (m3/s)

29.1 6.7 6.7

32.0 7.4 7.4

0.0 0.0/29.1

0.0 0.0/32.0

* The number on the left indicates airflow requirements when blast clearance does not occur, and the

number on the right when it does. For the energy calculation it is assumed that the working face is blasted and cleared only once in every four shifts.

6

Table 2.

Fan efficiencies and supply cost.

Fan

Fan efficiency

Fan capital cost (CAD)

Fan 1 Fan 2 Fan 3

75% 70% 65%

18,000 16,500 15,000

performing as expected before being installed in the mine. It is also good practice to document the fan efficiency as it is the main argument used in this study to justify a higher capital invest­ ment, based on the savings it could provide in energy consumption over time. The efficiency numbers in Table 2 are provided for illustrative purposes only in this study; these can be found in commercially available fan charts and are not the result of a particular fan performance test. Hypothetical values for cost are included to complete the steps of the methodology and also to reflect that a higher efficiency fan can have a higher capital cost. As can be observed in Table 2, Fan 1 is the most efficient fan, but also the more costly. It can also be observed that the energy efficient increments are proportional to the cost increment. However, if a mine adopts this methodology, it is recommended to obtain actual and current input values; those included herein are for demonstrating the proposed methodology. Considering a fan motor efficiency of 90%, 90% fan utilization during the year, and the pre­ viously introduced energy price of 7.5 CAD per kWh (75 CAD/MWh), the estimated energy consumption and cost is calculated and presented in Table 3. Table 3 presents the annual energy consumption estimate for each fan and its associated energy cost in the first two columns on the left. Fan 3 is defined as the more energy intensive fan and used as the base for comparison. The annual energy cost difference column is calcu­ lated as the difference between each fan’s annual energy cost and that of Fan 3 (energy savings as compared to Fan 3). The supply cost difference is established using the same procedure, with Fan 3 as the base case and the values indicating the additional capital investment for each fan. Then the first year balance is calculated by subtracting the supply cost difference (investment) to the energy savings (annual energy cost difference minus fan supply cost differ­ ence). The first year balance values indicate to which extent the energy savings can offset the higher capital costs of more efficient fans. The annual energy cost difference can also be inter­ preted as the maximum premium that a customer would be willing to pay for in a certain period of time to obtain a more efficient fan as compared to the baseline fan. Using values from Table 2 the selected fan would likely be Fan 3 based on the minimum supply cost criteria due to its lower cost compared to the other fans in this example. When the criteria is changed to take into account the energy savings for the first year and the increased supply cost, it can be observed that for both Fan 1 and Fan 2, it is cost effective to consider purchasing a more expensive option than Fan 3, as shown in Table 3. The reason is that in both cases the energy savings over the first year are larger than the premium capital that needs to be paid for the supply of Fan 1 or 2 as compared to Fan 3. A multi-year analysis with dis­ counted cash flow was not required in this case as the investment was offset within the first year. Using the criteria outlined in this methodology that considers energy savings and capital cost, Fan 1 is the recommended fan that would deliver the lowest cost of ownership over its useful life. Table 3. Energy and supply cost comparison. Fan

Annual energy con- Annual energy Annual energy cost sumption (kWh) cost (CAD) difference (CAD)

Fan supply cost difference (CAD)

First year balance (CAD)

Fan 1 Fan 2 Fan 3

800,124 857,276 923,220

3,000 1,500 0

6,233 3,446 0

60,009 64,296 69,242

9,233 4,946 0

7

4 SENSITIVITY ANALYSIS As presented in the previous section, the fan selection results can vary depending on the cri­ teria used for the decision. However, it is unclear how key variables could impact the validity of these decisions. Two key variables such as fan operating parameters and energy price will directly influence the expected energy savings and will be further studied to better understand their impact on the economics used to make a decision. It is becoming more popular to have mines implementing control systems for ventilation. It’s a sensible energy conservation measure given that ventilation is usually the largest energy consumer in underground mines (Hardcastle and Kocsis, 2007). At the auxiliary fan level, the control considered in this study was divided in two options, firstly manual control to turn on and off the fan, and secondly, fan speed control to modulate the airflow using a variable fre­ quency drive (VFD). For these two control strategies, this results in a lower fan utilization and lower airflow, resulting in lower energy consumption. The base case for comparison (baseline) is a system with no control measures where the fan stays on all the time as presented in the previous section. Table 4 illustrates the annual energy savings of each control strategy for Fans 1 and 2 as compared to Fan 3. It can be seen from Table 4 that the energy savings are greater for the more efficient fan (Fan 1) and that the savings are reduced with the introduction of ventilation control measures. Thus as savings are eroded, CAPEX can become a more important factor in the decision making process, especially when including the investment cost of a VCS. Depending on the location, the observed energy price in a country can vary significantly. For a mine, the remoteness can also have a big impact. To account for the range in energy prices, two additional prices were added to the initial analysis, 25 CAD per MWh and 125 CAD per MWh. Using the three energy prices (25, 75 and 125 CAD/MWh), the difference between the fan premium and the energy savings was calculated for each year over a five year period,using a discount rate of 10%. Tables 5, 6 and 7 present the results of the differential (energy savings minus fan capital premium) discounted cash flow. A negative value indicates that for an analysis up to that year, Fan 3 is a lower cost option. For example in year 1 of the manual control strategy, at an energy price of 25 CAD/MWh, Fan 3 is a better option than Fan 1 and Fan 2, when considering capital and energy costs. After the second year for the manual control strategy, both Fans 1 and 2 are better alterna­ tives than Fan 3 in terms of overall costs, as indicated by the positive values.

Table 4.

Fan annual energy savings considering manual control and VFD control.

Annual Energy savings (kWh)

Baseline

Manual control

VFD control

Fan 1 Fan 2 Fan 3

123,096 65,944 0

87,193 46,711 0

26,901 14,411 0

Table 5. Differential annual discounted balance – energy price 25 CAD/MWh. Year Energy price 25 CAD/MWh

Fan

1

2

3

4

5

Baseline (fan always on)

Fan 1 Fan 2 Fan 1 Fan 2 Fan 1 Fan 2

70 135 -746 -302 -2,116 -1,036

2,614 1,498 1,056 663 -1,560 -738

4,926 2,736 2,694 1,540 -1,055 -468

7,028 3,862 4,182 2,338 -595 -222

8,939 4,886 5,536 3,063 -178 2

Manual control VFD control

8

Table 6. Differential annual discounted balance – energy price 75 CAD/MWh. Year Energy price 75 CAD/MWh

Fan

1

2

3

4

5

Baseline (fan always on)

Fan 1 Fan 2 Fan 1 Fan 2 Fan 1 Fan 2

5,666 3,133 3,218 1,821 -893 -381

13,296 7,220 8,622 4,716 774 512

20,232 10,936 13,535 7,349 2,290 1,324

26,538 14,314 18,002 9,741 3,668 2,062

32,270 17,385 22,062 11,917 4,921 2,734

Manual control VFD control

Table 7. Differential annual discounted balance – energy price 125 CAD/MWh. Year Energy price 125 CAD/MWh

Fan

1

2

3

4

5

Baseline (fan always on)

Fan 1 Fan 2 Fan 1 Fan 2 Fan 1 Fan 2

11,261 6,130 7,181 3,944 330 274

23,977 12,942 16,189 8,770 3,109 1,763

35,538 19,136 24,377 13,157 5,635 3,116

46,047 24,766 31,822 17,145 7,932 4,347

55,602 29,884 38,589 20,770 10,020 5,465

Manual control VFD control

When control strategies are considered, results can change significantly. For example, for the first year of the baseline, Fans 1 and 2 are better options than Fan 3. But when a manual control strategy is considered, Fan 3 (the least efficient) is a better option than both Fans 1 and 2 for the first year. However, after the second year both Fans 1 and 2 are a better option than Fan 3. When considering the VFD control, Fan 3 would likely be the selected fan, as VFD control is the option with the minimum energy consumption within the three control strategies (always on, on/off, VFD). In this particular energy price example, Fan 1 cannot compensate for the additional capital premium even after five years of operation, and Fan 2 until the end of the fifth year of operation. In this case, the use of VFD is disadvantaged (as compared to selecting a more efficient fan) because energy consumption is minimal and the price of energy is low. As the energy price increases, to 75 and 125 CAD/MWh it can be observed that most values become positive. In Table 6 (energy price 75 CAD/MWh) only two values are negative, which means that only for the first year of the VFD control, Fan 3 would be the best choice, but in any other case Fan 1 and Fan 2 would be better options. Also, Fan 1 outperforms Fan 2 according to the capital and energy criteria for all cases. In Table 7 all the resulting values are positive which indicates that for an energy price of 125 CAD/MWh, there is value in considering both alternatives, energy efficient fans and applying control strategies. Three variables were included in this sensitivity analysis: energy price, control strategy and time period for payback. If the minimum supply capital cost criteria are applied, then the selection is Fan 3. If the criteria is then expanded to include energy cost, for the original base­ line (fan stays on all the time), then the selection is Fan 1, with both Fans 1 and 2 outperform­ ing Fan 3 and offsetting the additional investment in the first year for all three energy prices considered.

9

For the criteria that includes supply capital and energy cost, if the baseline changes due to ventilation control strategies such as manual control and VFD control, the results also become a function of the energy price and the timeline for payback. In the case of the manual control strategy, two years of operation are required at the low energy price for the higher efficiency fans to compensate for the capital premium. For the mid and high energy prices, the return on investment is observed in the first year. When considering the VFD control strategy, as expected, it takes longer to justify the additional expenditure on higher energy efficiency fans; for the low energy price, the payback is not observed for Fan 1 within the five year period, and only at the fifth year of operation for Fan 2. For the mid energy price the payback is observed in the second year and for the high energy price in the first year. The values presented in Tables 5, 6 and 7 could also be interpreted as the maximum capital premium that a mine would be willing to pay to purchase each fan as compared to buying Fan 3. That is also the reason why a separate sensitivity analysis was not performed on fan price as it is implicit in the values obtained in Tables 5, 6 and 7 when considering the supply costs provided in Table 2.

5 DISCUSSION OF RESULTS The results obtained in this study show that there is merit in not only considering the capital component to select a fan but also the energy consumption as both are paid over time by the mine site. Although the results are positive and encouraging for most cases, it does require that the ventilation practitioner performs more work to estimate energy consumption over time and also to provide an input for the economic evaluation as opposed to just providing input for the technical evaluation. However, it could also be argued that the additional work required can be minimized significantly by developing a model for these calculations that could be quickly updated based on key inputs. It could be argued that a more efficient fan could also have additional CAPEX impact in terms of the electrical infrastructure required, but in practical terms, unless the fan can pro­ vide the same duty point with a smaller motor, there would not be additional measurable sav­ ings on the CAPEX side (motor, starter and other components) resulting from a higher efficiency fan. The results of this study are a function of obtaining accurate fan efficiency values from the fan charts or the fan performance test. If the fan is expected to perform at different operating points and efficiencies over time, the proposed methodology should be extended to account for the changes over the considered period. Both higher efficiency fans and control strategies can be useful alternatives to use the venti­ lation energy in a more efficient form, however, using both alternatives at the same time might not be the recommended decision in all cases of energy use and price. From an energy perspective, a control strategy has a higher potential for savings than a higher efficiency fan as it can stop the fan when not needed (control strategy) as opposed to only reducing a fraction of the energy used (higher efficiency fan). Unfortunately a ventilation control system requires additional capital and operating cost. If the fan is expected to stay on all the time, then higher efficiency fans without control measures should be considered as the option. The system within this study considered a single working face connected to a fan and as such it could be argued that it is the case with the higher potential payback. Although this could be the case, the example is only presented to illustrate the use of the method­ ology. If a fan is expected to provide airflow to multiple faces, in theory it should effect­ ively work more often as compared to a single face fan. This situation is plausible but does not necessarily apply to all cases. As a result the methodology has to be applied on a case by case basis to determine what is the best alternative; this applies to all situations given that the values used as inputs can vary over time and for different locations which could impact the results.

10

6 CONCLUSIONS A methodology to estimate the capital and energy cost over a period of time was presented considering different energy prices, control strategies and evaluation periods. The results show that in most cases it is cost effective to consider both energy cost and capital cost as opposed to only capital costs for fan purchasing decisions. The example herein also showed that higher efficiency fans could improve the ventilation energy use and reduce the overall mine operating cost. Based on the results obtained, a higher efficiency fan may be a better option in applica­ tions where a fan stays on all the time. If ventilation control strategies are considered, such as manual control or variable frequency drive control, then the energy savings to offset the higher fan efficiency investment are reduced, but could still be significant over time depending on the energy price and time horizon. Performing the energy analysis is a good practice and should be part of the ventilation practitioner standard duties. However, as the results will vary based on input, the analysis needs to be performed on a case by case basis. REFERENCES O’Connor, D.F. 2008. Ventilation on Demand (VOD) auxiliary fan project – Vale Inco Limited, Creighton Mine, in Proc. of the 12th US North American Mine Ventilation Symposium, (ed. K. Wallace), 41–44, Reno, NV, June 9–11. McPherson, M.J. 1993. Subsurface Ventilation and Environmental Engineering, Chapman & Hall. Occupational Health and Safety Act. 2019. R.R.O 1990, Regulations 854 Mines and Mining Plants from: https://www.ontario.ca/laws/regulation/900854. Acuña, E.I., Alvarez, R.A., and Hurtado, J.P. 2016. Updated Ventilation On Demand review: implemen­ tation and savings achieved. Proc. of the 1st International Conference of Underground Mining, San­ tiago, Chile, October, 2016. Millar, D., Levesque, M. and Hardcastle, S. 2016. Leakage and air flow resistance in mine auxiliary ven­ tilation ducts: effects on system performance and cost, Mining Technology, DOI: 10.1080/ 14749009.2016.1199182. Calizaya, F. and Mousset-Jones, P. 1993. A method of designing auxiliary ventilation systems for long single underground openings, in Proc. of the 6th US North American Mine Ventilation Symposium, (ed. R. Bhaskar), 245–250, Salt Lake City, UT, June 21–23. Duckworth, I. J. and Lowndes, I. S. 2003. Modelling of auxiliary ventilation systems, Transactions of the Institutions of Mining and Metallurgy, Section A: Mining Technology, 112, (2), A105–A113. Acuña, E, Hardcastle, S, Fava, L and Hall, S. 2010. The application of a MIP model to select the opti­ mum auxiliary fan and operational settings for multiple period duties, INFOR, Vol. 48, No. 2, 2010, pp. 89–96. Hardcastle, S.G. and Kocsis, C.K. 2007. The ventilation challenge – A Canadian perspective of maintain­ ing a good working environment in deep mines. Challenges in deep and high stress mining, 2007, pp. 519–525. Papar, R., Szady, A., Huffer, W.D., Martin, V., and McKane, A. 1999. Increasing Energy Efficiency of Mine Ventilation Systems, in Proc. of the 8th US North American Mine Ventilation Symposium, 611–617, Rolla, MO, June 14–17. Babu, V.R., Maity, T. and Prasad, H. 2015. Energy Saving Techniques for Ventilation Fans Used in Underground Coal Mines—A Survey, Journal of Mining Science, Vol. 51, No. 5, 2015, pp. 1001–1008. De Villiers, D.J., Mathews, M.J., Maré, P., Kleingeld, M., and Arndt, D. 2019. Evaluating the impact of auxiliary fan practices on localised subsurface ventilation, International Journal of Mining Science and Technology, 29, 2019, pp. 933–941.

11

Case studies of mine ventilation

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Ventilation tradeoff study considering switch to battery electric vehicles N.D. Wineinger SRK Consulting, (U.S.), Inc., Clovis, USA

ABSTRACT: With the latest advances in battery technology, fast charging capabilities, and other technologies such as regenerative braking and more efficient battery electric systems, the use of Battery Electric Vehicles (BEVs) in the mining industry is becoming more advanta­ geous. In 2019, SRK Consulting, performed a BEV analysis for a mine in Colombia. The pur­ pose of this study was to consider an alternative ventilation system design for a previously completed feasibility study in which an all-diesel fleet was planned to be used for the mine. For this study, the mine requested that only the haul trucks and load-haul-dump loaders (LHD) be considered for conversion from diesel to battery electric. All other equipment would be left as diesel as in the previous study. This paper presents the ventilation assump­ tions which were made to complete this analysis and a discussion of the results. The results of the study showed significant reductions in the ventilation system power and infrastructure requirements with battery electric haul trucks and loaders compared to their diesel equiva­ lents, but these savings depend on the changes implemented and the level of risk willing to be taken for a switch to BEVs.

1 INTRODUCTION In 2019, a feasibility level ventilation study was conducted for a mine in Colombia where prescriptive airflow regulations regarding airflow for diesel equipment are more stringent than much of the world by requiring equivalent diesel dilution factors of between 0.09 and 0.13 m3/s per kW. Due to these regulations and a large fleet of diesel equipment, ventilation power costs calculated during the study were higher than the parent company expected. To pursue alternatives to reduce this demand, a preliminary tradeoff study was completed which compared the differences in ventilation system costs if the mine uses primarily battery electric vehicles (BEV) instead of diesel equipment. The BEV study revealed the reduction in ventilation and infrastructure that could be achieved along with the risks of implementing such a system.

2 BRIEF SUMMARY OF FEASIBILITY STUDY USING DIESEL EQUIPMENT The mining footprint had approximate dimensions of 500 meters wide and 1800 meters long with a depth near 1000 meters. Having an average estimated mining rate of 10,000 tonnes per day, the mine plan was to use longhole stoping methods with paste backfill. The primary pieces of underground mining equipment included 60t trucks and 18t loaders. At peak equip­ ment demand it was estimated the ventilation system would need to account for 15 active haul trucks and 12 active LHDs. It was estimated that the ventilation system would draw a peak

DOI: 10.1201/9781003188476-2

15

power demand of 6MW. The average power demand over the life of the mine was calculated to be approximately 5MW.

3 KEY ASSUMPTIONS AND DESIGN CONSIDERATIONS Before determining the ventilation system changes that would be needed to switch to battery electric vehicles, several key assumptions were made. • The client requested that an airflow requirement of 0.06 m3/s per kW be used for the ori­ ginal diesel fleet. This value is common in many countries as a minimum design value. º The use of this requirement assumed the mine would seek to get a variance once the mine designs are confirmed for development as Columbian regulations require an equivalent value of 0.09 to 0.13 m3/s per kW of diesel power. Justification for this vari­ ance could include a commitment to using Tier 4 equipment with low sulfur diesel making additional changes to the mine when possible to lower diesel emissions. • Engine requirements for the secondary and maintenance equipment were considered. º In the diesel equipment study, additional airflow for these pieces of equipment was not as heavily emphasized because it was assumed that secondary and maintenance equip­ ment would not be operating at the same time and in the same locations as haulage equipment and that the large airflows required for the original diesel fleet would be suffi­ cient for this equipment. • BEV substitutions were only considered for LHDs and Haul Trucks. º All other equipment were assumed to remain diesel. This is a key assumption since it lowers, but does not eliminate, the need for a ventilation system which removes diesel emissions and particulates from the mine. • Assumed that as advances are made in BEV technology, equivalent 60-tonne haul trucks and 18-tonne LHDs will be made available. º In the original study, 60-tonne diesel haul trucks and 18-tonne diesel loaders were sized for the mine to meet production goals. At the time of this study, 60-tonne BEV haul trucks and 18-tonne BEV loaders did not exist as the largest battery-electric under­ ground haul trucks and loaders available during this study were 42-tonne and 14-tonne, respectively. The study assumes larger pieces of BEV equipment would become available in the future to match those used in the diesel study. This is a key assumption because if BEV 60-tonne trucks and 18-tonne loaders are not available, the mine schedule and mine plan would likely need to change and adapt to smaller existing BEV equipment. Because this equipment did not yet exist at the time of this study, a scaling factor was used based on smaller existing BEV equipment. • Effects from potential equipment fires were not considered at this stage of the study. º The heat from a fire can have a higher buoyancy effect in lower air velocities. This can cause airflow reversals around the fire event. There is also a potential for high concentra­ tions of Hydrogen Fluoride gas and other hazardous substances to be released during BEV equipment fires. This gas is highly lethal. Unfortunately, real world data on this is sparse as it is an emerging technology. It is estimated the battery electric equipment has a reduced risk of catching on fire compared to diesel equipment, since the fuel is sealed from the surrounding environment. • The general flow of traffic and mine development from an operations perspective is assumed to not change with the switch to BEV trucks and loaders. º The type and style of BEV equipment chosen (quick charging, battery swap-out, equip­ ment charge and swap, etc.) can change the layout, traffic patterns, and hauling strategy in a mine. For instance, one of the benefits often touted with BEV equipment is in regard to using regenerative braking to partially recharge the batteries. This benefit is often recommended by BEV manufacturers to be used in combination with hauling

16

downgrade. In this way, more energy is regenerated by retarding a heavier loaded truck going downgrade while using less energy to drive the empty trucks upgrade. This increases the time between stationary charges or decreases the number of battery swaps. This consideration also requires that the locations of battery charging stations be stra­ tegically planned to optimize productivity. For this study, these changes are not ana­ lyzed, and ventilation designs are not altered to include these types of considerations. The purpose of this study was to complete a preliminary analysis of using BEV trucks and loaders to replace their diesel equivalents and evaluate the effects on the required ventilation infrastructure. • A minimum air velocity value of 0.5 m/s is used to comply with Colombian regulations. º In this design, air velocities between 1.5 m/s and 3.0 m/s (Figure 1) were used in dusty airways with conveyors and other equipment. For other areas including active levels, ramps, and drifts, air velocities were maintained between 0.5 m/s and 6 m/s. Dust is pri­ marily a limiting factor in some of the ramp areas and around underground crushers where minimum velocities are maintained to exhaust the dust expected to be generated from the conveyor. All other areas of the ventilation system are maintained above these minimum thresholds and below recommended maximum velocities (except near main fan installations). • This study only considered an analysis of the primary ventilation system. The secondary ventilation system was not considered but is still important and could be affected by the use of BEVs. º By using BEV equipment, there is the potential for decreasing these secondary ventila­ tion costs and requiring less airflow delivered to each face. However, one of the down­ sides of decreasing ventilation for auxiliary headings is increased time for clearing blast fumes. For this mine, blast clearing times were significant due to long development drives. For this reason, it was assumed that the auxiliary systems designed during the diesel case would remain the same for the BEV study to maintain blast clearing times and therefore meet existing production goals.

Figure 1.

Relative dust concentration versus air velocity (McPherson).

17

4 CALCULATIONS AND RESULTS Four sets of ventilation models were developed using VentsimTM Design software in succes­ sion based on the original study. The mine consisted of ventilating several districts with several primary ventilation shafts. Mine scheduling resulted in fluctuating ventilation demands throughout the life of mine. The models developed were based on minimum and maximum velocities in all major areas of the ventilation system for the control of dust in active areas. To account for diesel emissions and diesel particulate matter (DPM), it was determined that besides LHDs and Haul Trucks, diesel maintenance trucks and personnel carriers may be util­ ized in active areas during the shift. The presence of these two pieces of equipment require the largest remaining ventilation allocation. In addition, dedicated ventilation splits for fixed facil­ ities such as underground maintenance shops, fuel bays, blasting magazines, crushers, and parking bays were also accounted for as in the previous all diesel study. Leakage in the mine was calculated through simulated bulkhead resistances. Models were developed for select stages in the mine life noted by significant activity or ventilation changes. Modeling results showed that the decreases in primary fan airflow and power costs were approximately 50% and 80%, respectively. Figure 2 compares the overall results of each stage developed for the BEV analysis with the base diesel results.

Figure 2.

Comparison of diesel equipment versus BEV modeling results.

18

The base study did not make any modifications to the original mine infrastructure other than redistributing the required airflow in the mine to meet various mining demands. Because of the significant reduction in fan power consumption, it was determined that additional options could be considered to reduce the required infrastructure in the mine. Since it was assumed that LHD and Haul Truck sizes were largely unchanged, adjustments to ramp, drift, and level sizes would not be considered as these are based on equipment envelops. However, because of significantly lower airflow demand, it was determined either the number of ventila­ tion raises or both the size and number of raises could be reduced to lower development costs. Therefore, two additional modeling analyses were performed for optimizing both the number and size of raises. The results of this analysis as compared with other models are shown in Figure 3. Table 1 shows a comparison of the total volume of material removed to develop the raises in each life of mine (LOM) model.

Figure 3.

Comparison of diesel and all cases using BEVs.

Table 1.

Comparison of total material removed to develop raises for each LOM model.

Model Description

Total Raise Volume Saved (m3)

Original BEV equip./Diesel equip. study BEV equip., optimized size of raises BEV equip., optimized number and size of raises

– 10,000 110,000

19

By optimizing the number and size of raises, airflows were maintained from the first BEV study while total main fan motor powers still report a savings of nearly 60% from the original diesel equipment study. Moreover, the smaller raises (or fewer raises) resulted in a total amount of raise material removed over the LOM is reduced by 50% which would result in a significant development cost and schedule savings.

5 CLIMATIC MODELS The developed ventilation models considered airflow requirements for secondary diesel equip­ ment, minimum velocities for control of dust and removal of blasting fumes, and air required for personnel. In addition, heat models were developed for each of the three BEV cases con­ sidered. With the all diesel study, heat in the mine was comparatively high due to the number of large diesel equipment. With the switch to BEV haul trucks and LHDs, it was expected that much of these heating effects would be reduced. Climatic models were developed to quantify the improvements. Several different factors were considered in this study including the heat load of under­ ground crushers, conveyors, BEVs, secondary or maintenance diesel equipment, auxiliary fans, and battery charging stations. For underground crushers, a scaling factor was used for dilution of heat based on similar studies with similar mining rates. When comparing the air­ flow requirements for heat generated by the crusher with the minimum airflow requirements to dilute dust in the area, the dust airflow requirements were higher and therefore were incorp­ orated in the models. In determining the heat loads for the anciliary pieces of underground diesel equipment, a series of assumptions were made. Heat loads calculated in the original climatic model were based on the diesel haul trucks and LHDs. No anciliary equipment was modeled since it was generally assumed that the secondary equipment would be in either limited use or add a negligible amount of additional heat when compared with the primary equipment. For the BEV study, these secondary pieces of equipment needed to be considered since their heat con­ tribution is more significant than the BEV haulage equipment (though still with limited util­ ization and availability). For the BEV heat models, a key assumption for heat loads was taken from a 2014 white paper entitled Electrification of Loaders and Trucks – A Step Towards More Sustainable Underground Mining. This paper states that “[battery electric equipment] generate only a third of the heat emitted by a diesel having the same power” (Paraszczak et al., 2014). This is related to the overall efficiency of the energy produced by diesel internal combustion engines. This heat energy appears in three different ways: one third as heat from the radiator and body of the engine, another third from the exhaust gases and the remaining third as useful shaft power which is transformed into heat by frictional losses as the equipment performs its task (Mcpherson 2009). It is assumed that the heat generated by larger BEV equipment will scale similarly to what is already on the market. It is also unclear which makes/models of BEV equipment will be available when the mine is developed since manufacturing capacity of such equipment was still growing. The BEV heat models were developed based on the primary equipment types and locations used in the diesel study. To be conservative and account for unknown heat scaling issues as well as build in a factor of safety for other secondary diesel equipment running in the mine, it was assumed that BEV haul trucks and loaders would pro­ duce 50% of the heat of their diesel equivalents. Additional heat sources were not added to account for secondary equipment separately since these were assumed to operate intermittently. For the climatic study, an exhaust wet bulb temperature of 28°C was used as a design limit for acclimated workers in areas where personnel will be active. Climatic modeling was com­ pleted for all three sets of BEV models. The results for each set of models were very similar. For most areas of the mine, temperatures were moderately warm. In certain areas, wet bulb temperatures do exceed 28°C, but temperatures generally decrease when diluted with air in main ventilation airflow streams. A buildup of equipment closely together on levels or in the 20

ramp results in higher temperatures, but the use of BEVs significantly reduces the temperat­ ures in the mine as well as the mine’s sensitivity to proximity of mining equipment. Climatic modeling did not result in significant changes to the ventilation design based on the airflow allocation for the mining and development ventilation models previously discussed in this paper. Airflow requirements for this study were based primarily on dilution of secondary diesel equipment and maintaining minimum velocities instead of climatic issues.

6 DISCUSSION OF RESULTS When considering an underground mine with all diesel equipment, the amount of airflow required to dilute diesel emissions and particulate matter for trucks and loaders is often the primary driver in determining airflow requirements. With BEVs replacing the largest diesel mining equipment, the airflow requirements become focused on controlling dust, diesel emis­ sions and DPM from smaller equipment, and heat from both fixed and mobile equipment. This study considered all these aspects to determine the control factors for the ventilation system. The study showed significant electricity savings for the ventilation system using BEVs to replace the primary equipment instead of using all diesel equipment. Reducing the airflow for BEVs and smaller diesel equipment resulted in a significant reduction of airflow and hence, main fan duties and fewer or smaller ventilation raises. It also resulted in a significant reduc­ tion in heat from equipment. Climatic modeling showed that heat was not a concern. Certain challenges need to be addressed in order to thoroughly understand the benefits of using BEVs. While there will be a decrease in ventilation power demand due to the need for less airflow, additional electrical power will be needed for charging BEVs. This may incur cyc­ lical loading of the electrical power grid, which could be problematic for grids with time of use cost structuring as well as capacity issues for existing mines. This additional cost would have to be compared with the maintenance and fuel costs associated with equivalent diesel equip­ ment for a more complete picture of the potential savings. Another factor is related to auxil­ iary ventilation requirements. The use of BEVs can result in a reduction in secondary ventilation system costs, but alternate factors should be considered such as dust control in the heading and how blast clearing times would be affected. One of the main goals for electrifying the equipment fleet is to reduce costs. Some of the biggest capital savings can be realized from reducing the size of ventilation raises and airways. However, BEVs of the size used in this study, are more costly than their diesel counterparts and have no track record regarding maintenance, reliability, dependability, availability and longevity. The capital and maintenance cost of a BEV fleet needs to be considered in a more detailed study. Specialized mechanics, electrictians and other technical staff are required to support a BEV fleet. Because of this, using BEVs in a remote mine will be more challenging. If a battery swap out system is chosen, recent improvements by some manufactures have shown that swapouts are comparable to the time needed to refuel diesel equipment of similar size. Extra time may be needed for BEVs if multiple swaps are needed per shift which is a function multiple variables including elevation change, shift length, haul distance, etc. Furthermore, if the mine sizes all infrastructure to support BEVs, then elects to move to diesel equipment, the mine infrastructure would be undersized, resulting in significant capital and operating cost increases. With recent advances in BEV technology, significant benefits and savings for mine ventila­ tion systems may be possible. To date, only a limited number of mines have implemented BEVs. As more data becomes available on the costs of installing and operating a fleet of BEVs, the use of these machines is likely to increase. The significant advantage of cleaner air combined with lower ventilation and cooling demand will drive this technology. However, for significant implementation of BEVs to occur, challenges such as battery degradation, potential production losses due to charging and/or battery swap outs, power grid requirements for char­ ging batteries etc. may need to be more thoroughly analyzed. 21

REFERENCES McPherson, M.J. 2009. Subsurface Ventilation Engineering. Fresno: Mine Ventilation Services, Inc. Paraszczak, J. & Svedlund, E & Fytas, K. & Laflamme, M. 2014. Electrification of Loaders and Trucks – A Step Towards More Sustainable Underground Mining, Renewable Energy and Power Quality Jour­ nal, 81–86, 10.24084/repqj12.240.

22

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Calibration of LKAB’s Konsuln test mine ventilation model using barometer Pressure-Quantity (PQ) survey S. Gyamfi Luossavaara-Kiirunavaara Aktiebolag (LKAB), Kiruna and Malmberget, Sweden

A. Halim Luleå University of Technology, Luleå, Sweden

A. Martikainen Luossavaara-Kiirunavaara Aktiebolag (LKAB), Kiruna and Malmberget, Sweden

ABSTRACT: Ventilation projects such as installation of primary fans, refrigeration system, heating systems, and Ventilation of Demand (VOD) system require some simulations to ascer­ tain their benefits and fulfilment of their purpose before the actual financial commitment is made to execute such projects. Ventilation models used for the simulations should provide some degree of accuracy to ensure that their results will reflect the actual mine ventilation cir­ cuit. This paper outlines a barometer Pressure-Quantity (PQ) survey that was done to cali­ brate the existing ventilation model of LKAB’s Konsuln test mine before it was used in a VOD design study to determine whether it will be feasible for Konsuln mine to install the system. The results show a good correlation between the simulated airflows in the calibrated model and the actual underground measured airflow quantities and primary fans pressure. This good correlation has validated the model for its use in Konsuln VOD design study.

1 INTRODUCTION Ventilation simulation softwares enable ventilation engineers to carry out airflow, radon, pres­ sure, heat, fire and several other simulations in a given mine ventilation model. These soft­ wares also help engineers to model the ventilation circuit of a mine, have a thorough understanding of how the airflow will behave, the fan pressures and effects when certain activ­ ities such as fan installation are carried out in the mine. Several simulation softwares have therefore been developed over the years such as VentSim Design, VuMa, and VnetPC. They have been proven to be of great benefit in the initial design of a mine ventilation circuit. They have also been useful in modelling the spread of blast fumes, heat and contaminants within the mine ventilation circuit. These softwares have also played key roles in accident control system by simulating several scenarios to support deci­ sion-making process in the event of incident in the mine to address various issues such as gas or fire control problems underground (Wu et al., 2019). However, there is a high risk in using a ventilation model that has not been calibrated for ventilation planning and for ventilation projects that involve millions of dollars in capital or operating costs. Not only will there be losses in terms of cost, but it may also affect future mine production or result in serious occupational health and safety consequences associated with gases, fires, dust etc. (Brake, 2015). These produced ventilation models from simulation softwares should therefore be calibrated to ensure that the models are reasonably accurate to

DOI: 10.1201/9781003188476-3

23

give reliable simulation predictions of the performance of the actual mine ventilation system in their current state and for future ventilation planning, such as the situation in LKAB’s Konsuln test mine. Konsuln mine is owned and operated by Luossavaara-Kiirunavaara Aktiebolag (LKAB), a Swedish state-owned iron ore mining company. The mine is located just south of the famous Kiruna mine, the largest underground iron ore mine in the world in term of production, with the output of about 26.9 million tonnes per annum (mtpa) (LKAB, 2021). Konsuln orebody was extracted using open pit mining methods in the 1970s and underground mining in the 1980s. The mine was then closed because of unfavourable economic conditions. When LKAB set up its Sustainable Underground Mining (SUM) project (SUM, 2021), a new block of the mine underneath the old workings was developed 2018-2020 as a test mine for the project. Besides Kiruna and Konsuln, the company also operates 2nd largest underground iron ore mine in the world in term of production, Malmberget, and an open pit mine operation, Svap­ pavaara. All of these operations are located in Malmfälten area in Northern Sweden inside the Arctic Circle. Products from these three mines are transported by rail to the exporting ports of Narvik in Norway and Luleå in Sweden where they are subsequently shipped to LKAB’s customers. Figure 1 shows the map of all mining operations and exporting ports. Besides functioning as a test mine, the newly reopened Konsuln mine contributes ore pro­ duction. The mine currently produces about 0.8 mtpa of iron ore. It is planned to increase it to a rate between 1.8 and 3 mtpa in the future. At the time of the writing of this paper, the new production rate was still being decided, but it will not exceed 3 mtpa. The mine uses the same mining method as in Kiruna mine, Sublevel Caving. The new block of the mine has three production levels: 436, 486, and 536. The deepest level, level 536, is about 390 m below

Figure 1. Map of Northern Sweden, Norway, and Finland, which shows LKAB’s mining operations (Kiruna, Malmberget, and Svappavaara) and exporting ports (Narvik and Luleå).

24

the surface. The ore is transported to the surface stockpile by diesel trucks via a ramp. This is different than the ore handling system used in Kiruna mine where shaft hoisting is used. The ventilation and heating system is crucial to ensure safe and efficient operation of Kon­ suln mine. At the moment, the primary ventilation system is designed for a production rate of 0.8 mtpa. Due to a planned increase in the production rate mentioned above, a study was car­ ried out to see whether employing VOD would avoid an expensive upgrade of the existing pri­ mary fans and the subsequent increase of their power cost. A major part of this study was to calibrate the existing ventilation model of the mine in order to make it fit for the purpose of designing a VOD system in Konsuln test mine. The original model was found to have some significant discrepancies with the ventilation survey results and therefore was deemed as not fit for the purpose. This paper outlines the calibration work of the Konsuln ventilation model, which was done using a barometer Pressure-Quantity (PQ) survey method from April until May 2020. The aim of this PQ survey was to obtain the actual friction factor and shock loss factor of Kon­ suln’s airways, and also to check the accuracy of the primary fan curve that was used in the original model. Experience shows that these discrepancies are generally caused by the fact that the actual friction factor and shock loss factor are often different than the preset values gener­ ated by the simulation softwares. Simulation softwares have preset values of these parameters, which were obtained from literatures. However, the actual values are often different because these values were obtained from measurements in certain mines, which may have different geometry of airways compared to the simulated mine. Using inaccurate fan curves was also found as another main cause for these discrepancies. The barometer method was chosen over the gauge and tube method because of its simplicity and practicality in the measurements, in particular when surveying airways that have some traffic in them such as the main ramp and footwall drives. Another reason for choosing this method was the time limitation to complete this calibration work. Due to its simplicity, the barometer method requires less time to complete the measurements than the gauge and tube method; and therefore was the better option to meet the deadline of this calibration work.

2 KONSULN MINE VENTILATION SYSTEM Konsuln mine employs a combination of force and push-pull primary ventilation systems. The push-pull system is only used during the clearance of production blasting fumes. After production blasting is done on a certain level, two 22 kW fans, bolted to a bulkhead located in the connection drive to the exhaust raises, are turned on to pull the fumes from that level. After the fumes are cleared, these fans are turned off and the system runs only on “push” mode, which is the mode during the steady-state production. The primary intake fans are two 75 kW EOL Vent system inline axial fans installed in parallel. Both fans are equipped with Variable Frequency Drives (VFD) to vary their speed. These fans are located 52 m below surface and were designed to deliver 100 m³/s of fresh air to the mine. Due to extreme winter climate conditions where ambient surface temperatures can be as low as -40°C dry bulb (DB), the intake airflow must be heated during winter months. A direct-contact heating system, using electric coils, is installed on top of the fan intake raise. Figure 2 shows a schematic of the primary ventilation system, which includes dimensions of the primary airways. Figure 3 shows the ventilation model of the mine, which was created using VentSim Design software. Primary airflow on each level is provided by 30 kW auxiliary fans bolted to a bulkhead located in the connection drive to the intake raises. Each fan is connected to a 1,000 mm diam­ eter duct that extends to the level footwall drive. This air is then distributed to each crosscut (production drive) using a 11 kW auxiliary fan (called crosscut fan) connected to a 800 mm diameter duct installed in the access to each drive. Figure 4 shows a view of level ventilation in Konsuln mine.

25

Figure 2.

Schematic of Konsuln primary ventilation system, with dimension of primary airways shown.

Figure 3.

VentSim model of Konsuln ventilation system.

3 KONSULN VENTILATION MODEL CALIBRATION PROCESS The existing ventilation model of Konsuln mine as of 2 December 2019 was obtained as the base case model for calibration. The model was created using VentSim Design software, a popular ventilation simulation software. Due to time limitation, the survey was not done in all airways within the mine. Rather, some representative airways with similar characteristics were measured as follows: – Airways that have the same dimensions and ground support on each level. For example, on level 436 and 486 only some parts of the footwall drive that have the same dimension and ground support were surveyed at the same elevation, instead of the whole footwall drive. – Some sections in the main decline/ramp. 26

Figure 4.

A plan of Konsuln level ventilation.

The survey was not done in the ventilation shafts and raises due to the limited number of personnel that could be involved. Only two personnel were available to complete the survey program; and they were required to stay together during the survey to follow safety protocols. Hence, surveying ventilation shafts and raises, which would have required two personnel at both ends simultaneously (total of four personnel), could not be done. Therefore, the calibra­ tion of their friction factor was done based on the calibration in other airways in the vicinity (i.e. after these airways were calibrated, the model was run and then the values of the raise airflow quantities were compared with those measured. If the difference was found to be less than 10%, it was then concluded that the preset friction factors in VentSim Design for shafts and raises was acceptable and could be used in the calibrated model). It must be noted that level 536 was still being developed during the survey and no measure­ ments could be done on this level. However, because parameters on level 536 are similar with those on level 436 and 486, the calibrated friction factor and shock loss factor on these levels can be applied.

4 THEORY OF BAROMETER PRESSURE-QUANTITY (PQ) SURVEY PQ surveys play an important role in obtaining the frictional pressure drop and the corres­ ponding airflow quantities needed for calibration and validation of a model. The parameters that are measured are airflow quantity, air pressure, and air temperatures. A guideline for survey execution to build and calibrate a given ventilation model has been presented by some practitioners (Prosser and Loomis, 2004; Prosser, 2020; Rowland, 2009, 2010, 2011). 27

In a barometer PQ survey, air pressure at two points within a uniform airway is measured simultaneously using a barometer. Airflow quantity, air temperatures (dry and wet bulb), and distance between these two points are also measured. When the two points are on the same elevation (i.e. inside a horizontal airway such as footwall drives), the pressure difference between these two points is the pressure loss of the airway between them. When the two points are not at the same elevation (i.e. inside an inclined airway such as ramps), a correction to the pressure difference must be made to exclude the weight of the air column. The correc­ tion is calculated by multiplying air density with elevation difference and gravitational acceler­ ation. The elevation of each measurement point was obtained from the mine surveyors. The air density is calculated using psychrometric equations with air temperatures and barometric pressure as the input data. The airway resistance can therefore be calculated using the following equation:

Where: P is the pressure loss of the airway between two measurement points (Pa) R is the resistance of the airway between two measurement points (Ns2/m8, nicknamed as Gaul) Q is the airflow quantity inside the airway between two measurement points (m3/s) After calculating the resistance of the airway between two measurement points, friction factor and shock loss factor between these two point can therefore be calculated using the fol­ lowing equation:

Where: k is airway friction factor between two measurement points (Ns2/m4 or kg/m3) C is the circumference (perimeter) of the cross-section of the airway between two measure­ ment points (m) L is the length of the airway (m) or the distance between two measurement points A is the area of the cross-section of the airway between two measurement points (m2) ρ is the density of the air that flows inside the airway between two measurement points (kg/m3) 1.2 is the density of air at sea level. The unit is kg/m3

X is the airway shock loss factor between two measurement points (dimensionless)

5 MEASUREMENT OF PRIMARY FANS QUANTITY AND PRESSURE Another PQ survey was done on the primary fans to determine their actual duty point and to check the accuracy of their curve in the original model. As mentioned previously, using inaccurate fan curves is another general cause of the discrepancies between the model and the actual condition. Figure 5 shows the layout of the Konsuln primary fans. It can also be seen in Figure 5 that the airflow delivered by the primary fans encounters two 90° (approximately) bends after the regulator, which will be removed in the future in order to achieve the design quantity of 100 m3/s. The fan quantity was measured using a vane anemometer at the regula­ tor in front of the outlet of the fans’ diffuser; and Fan Total Pressure (FTP) was determined by measuring the pressure across the fan bulkhead using a digital manometer. The pressure was then adjusted for shock loss caused by the expansion of the ejected air from the fans into the chamber using the following relationship:

28

Figure 5.

Plan view of the installation of Konsuln primary fans during the PQ survey.

Where:

A1 is the cross sectional area of the outlet of fans’ diffuser (m2)

A2 is cross sectional area of the fan chamber (m2)

FVP is Fan Velocity Pressure (Pa)

6 INSTRUMENTS THAT WERE USED IN THE SURVEY Following are the instruments that were used in the PQ survey to calibrate Konsuln ventila­ tion model: – 2 x GE Druck DPI800 electronic barometers, used to measure barometric pressures. – 1 x Dwyer series 475 Mk. III digital manometer, used to measure pressure across primary fans bulkhead. – 1 x TSI VelociCalc 5725 vane anemometer, used to measure air velocity in horizontal air­ ways and in primary fans diffuser outlet. – Kimo AMI 310 hot wire anemometer, used to measure air velocities in horizontal airways that have very low air velocity. – 1 x Leica Disto D210 laser distancemeter, used to measure cross-sectional dimension of horizontal airways. – 1 x Kestrel 5000 pocket weathermeter, used to measure air temperatures. – 1 x Zeal whirling hygrometer, used to measure air temperatures. 29

7 RESULTS FROM THE MODEL CALIBRATION After the model was calibrated using the calculated friction factor and shock loss factor, and the corrected primary fan curve, the simulation was ran and some of its simulated values were compared against the measured values and the simulated values from the original (uncali­ brated) model. 7.1 Primary fans The measurement of primary fans quantity and pressure found that an inaccurate plot of the supplied curve was used in the original model. The plot was subsequently corrected and then re-entered into the calibrated model. Figure 6 shows the comparison between the measured quantity and FTP of Konsuln primary fans and their simulated value in uncalibrated and cali­ brated models. It shows that the difference between simulated and measured values has been reduced substantially. However, the FTP in the calibrated model is still 135.2 Pa higher than the measured value (corresponds to 13% difference relative to the measured value). When the measured duty point was plotted on the supplied fan curve, it lays below the curve. This indi­ cates that the actual fan curve is lower than the one supplied by the fans’ manufacturer. This situation is not unusual because the supplied fan curves are derived from the testing program done in a test chamber that is clean and dry; and during testing the air that flows through the fans is relatively clean (contains very little dust and moisture). However, in the mine, air that flows through the fans is usually dusty and moist, which can impact the actual fan duty point where it intersects the supplied fan curve. 7.2 Level 436 and 486 airflow comparison The level survey data were also compared to the simulated modelled values, as shown in Figure 7 and 8. These figures show a comparison between airflow quantities in the uncali­ brated model, calibrated model with flows that were measured at various measurement sta­ tions on level 436 and 486. It is clear from these figures that there is an improved correlation between airflow quantities in the calibrated model and the measured data after the calibration.

8 DISCUSSION It can be seen in Figure 6 to 8 that the model has been successfully calibrated. The difference in primary fan quantity (relative to the measured value) has been reduced from 3.8% in the uncalibrated model to 1% in the calibrated model, whilst the difference in primary FTP (relative to the measured value) has been reduced from 24.6% in the calibrated model to 13% in the calibrated model. The difference between airflow quantities on level 436 and 486 has also been reduced substantially. For example, on level 436 (station S03), the difference in model quantity (relative to the measured value) has been reduced from 716.4% in the uncalibrated model to 0.6% in the calibrated model. Similarly, on level 486 (station S02), the difference has been reduced from 79% in the uncalibrated model to 3.1% in the calibrated model. It must be noted that the main disadvantage of barometer method is the requirement to apply correction to the readings in inclined airways. This correction depends on the accuracy of the measurement of measurement points elevation and barometric pressure. Although this was not an issue in this calibration work, the use of gauge and tube method should be con­ sidered in the future calibration work because there is no need to apply correction when this method is used. However, a longer timeframe and more personnel than what was available for this calibration work would be required should gauge and tube method be selected in the future. 30

Figure 6.

Comparison of primary fan performance before and after calibration.

31

Figure 7.

Comparison of underground measured and modelled values on Level 436.

Figure 8.

Comparison of underground measured and modelled values on Level 486.

9 CONCLUSION AND RECOMMENDATION The work outlined in this paper shows the improvement of the Konsuln ventilation model after its calibration using the barometer PQ survey method. Even though the primary FTP in the calibrated model is still 135.2 Pa higher than the measured value, the airflow quanti­ tites in the calibrated model have become closer to the measured values after the calibra­ tion. The lower measured FTP compared to simulated value in the calibrated model indicates that the actual fan curve is lower than the supplied one, which is not uncommon as explained in Section 6.2. Therefore, the model can be considered as acceptable for designing the VOD system and other future ventilation design and planning work in Kon­ suln mine. 32

It is recommended that the model continue to be calibrated as conditions change in the future as mining progresses. The barometer PQ survey is useful not only to calibrate the venti­ lation model but also to get a snapshot of the actual condition in the mine ventilation circuit. ACKNOWLEDGEMENTS The authors are grateful to LKAB for the support in conducting this study and for approval to publish this paper. The authors also acknowledge the following personnel at Konsuln test mine and Kiruna mine for their support in planning and carrying out the PQ survey: Mr. Michael Lowther, Konsuln mine manager, Dr. Matthias Wimmer, manager of Kiruna mine mining technology, Ms. Stina Klemo, Kiruna mine ventilation engineer, and Mr. Michal Grynienko, mining technology research engineer. REFERENCES Brake, D J, 2015. Quality assurance standards for mine ventilation models and ventilation planning, in Proceedings The Australian Mine Ventilation Conference, pp 221–228 (The Australasian Institute of Mining and Metallurgy: Melbourne). Luossavaara Kiirunavaara Aktiebolag (LKAB) website, 2021. https://www.lkab.com/en/about-lkab/ from-mine-to-port/mining/ Prosser, B., 2020. Conducting Airflow Measurements For Ventilation Surveys, Internal Presentation, SRK consulting, USA. Retrieved from https://www.srk.com/en/publications/conducting-airflow-meas urements-for-ventilation-surveys Prosser, B. S., & Loomis, I. M., 2004. Measurement of frictional pressure differentials during a ventilation survey. In 10th US Mine Ventilation Symposium (pp. 59–66). Rowland, J A., 2009, Moving the Ventilation Report into the 21st Century, in Proceedings of 9th Coal Operators’ Conference. University of Wollongong and the Australasian Institute of Mining and Metal­ lurgy, 2009, pg. 155–174. Rowland, J. A., 2010. Survey Execution to Build a Ventilation Model, Australian Style. In 13th United States/North American Mine Ventilation Symposium. Rowland, J A., 2011, Ventilation Surveys and Modelling - Execution and Suggested Outputs. 11th Under­ ground Coal Operators’ Conference, University of Wollongong & the Australasian Institute of Mining and Metallurgy, pg. 214–224. Sustainable Underground Mining (SUM) project website, 2021. https://sustainableundergroundmining. com/news/ Wu, F., Chang, X., & Dan, Z., 2019. A Mine Ventilation Program Integrated with Gob Flow Field Simulation. In Proceedings of the 11th International Mine Ventilation Congress (pp. 888–898). Springer, Singapore.

33

Computational fluid dynamics applications in mine ventilation

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Investigation of explosion hazard in longwall coal mines by combining CFD with 1/40th scaled physical modeling A. Juganda, H. Pinheiro, F. Wilson, N. Sandoval, G.E. Bogin & J.F. Brune Colorado School of Mines, Golden, USA

ABSTRACT: To evaluate methane explosion hazards in longwall coal mines, researchers at the Colorado School of Mines have developed a Computational Fluid Dynamics (CFD) model, along with a 1/40th scaled, optically accessible model of an underground longwall mining section. In this project, CFD models assisted in the design of the physical model to ensure specifications were met for accurately representing the scaling physics as well as assist in narrowing the experimental matrix and identifying key locations for sensor placement to measure, flow, pressure, and gas concentrations. This procedure will help develop a mine wide methane monitoring system to improve methane ignition and explosion mitigation strategies in longwall coal operation.

1 INTRODUCTION Adequate ventilation and mine atmospheric monitoring system are keys in preventing explo­ sions in underground coal operations. A major concern in longwall coal operations is methane gas explosions in the face area that could potentially transition to a more severe coal dust explosion, such as the Upper Big Branch (UBB) mine explosion in 2010 (Page, 2011; Phillips, 2012). Explosion risks always exist in longwall operation, primarily in the form of face igni­ tions in the longwall face and gob areas. Several factors, such as ventilation setup and gob characteristics, can significantly impact the formation of explosive gas mixtures inside the longwall face and gob area. The use of point-type methane sensors to detect the explosion risk relies heavily on the sensor placement (Juganda, 2020; Krog, 2016). In addition, due to its inaccessibility, methane monitoring of the gob area is difficult. Mine operators usually rely on methane drainage through Gob Ventilation Boreholes (GVB) and nitrogen injection as a preventive method. These methods are commonly used for progressively sealed gobs, but nitrogen injection is less effective for bleeder-ventilation panels (Marts et al., 2015; Saki, 2015). Researchers at the Colorado School of Mines (CSM) have built a 1/40th scaled physical model of a longwall coal mine. To complement and guide the development of this model, Computational Fluid Dynamics (CFD) modeling is used to help identify critical ventilation parameters, such as flow scaling, gas mixture distribution, and sensor placement. These parameters are first optimized in the CFD model, then implemented in the scaled physical model. The physical experiments are used to validate the CFD model, including gob perme­ ability. The validated CFD model can help reduce the time and number of experiments required for different ventilation scenarios. The combination of Computational Fluid Dynamics (CFD) and scaled physical modeling can be used to develop more reliable atmospheric monitoring practices and provide an improved understanding of the gas mixtures formation inside the gob for different ventilation

DOI: 10.1201/9781003188476-4

37

scenarios. The goal of this project is to develop early detection method to improve methane explosion prevention and mitigation strategies in longwall coal operation.

2 PHYSICAL MODEL OVERVIEW The 1/40th scaled physical model has a dimension of 7 m long by 6 m wide and 1.7 m high. The modeled active longwall panel includes the longwall face, gob area, the surrounding mine entries, and the ventilation control required to simulate different ventilation setup. The scaled version has optical access on the sides and tops to visually observe the airflow pattern. Flow and gas sensors are installed in critical areas to analyze the flow distribution and gas mixtures inside the gob area and surrounding mine entries. The mine entries dimensions are 144.8 mm wide and 72.4 mm high, which are equivalent to 5.8 m wide and 2.9 m high in full scale. The longwall face is 5.5 m long, equivalent to 220 m in full scale, and separated into 11 face carts consisting of 10 shields per cart. These face carts can be advanced individually to simu­ late different shearer cutting scenarios. Automation is integrated into the shearer using a Bluetooth connection to allow movement across the face, and adjustment to each shearer arm height and drums rotation depending on the simulated scenario (McMack et al., 2018, Pinheiro et al., 2020). The gob cart dimension is 60 cm long, 50 cm wide, and 61 cm high. The model can accom­ modate 11 columns and up to 6 rows of gob carts. Each gob cart can be filled with different ball sizes to simulate the different gob porosity and permeability distribution observed in the real longwall gob. Figure 1 shows the schematic of the physical scaled model, while Figure 2 shows the picture of the built model. The model allows methane inflow into the model through two injection systems. Figure 2 (right) shows the gas injection manifold at the face carts that will be connected to the injection

Figure 1. Isometric view of the 1/40th scaled physical longwall model schematic. (Modified from McMack et al., 2018).

38

Figure 2. Left (a), (b), (c): Assembled physical scaled model (Pinheiro et al., 2020). Right: Longwall face injection system.

line. These manifolds consist of 6 mm diameter holes spaced every 6 mm. Behind the mani­ fold, 25 mm thick porous foam is used to simulates the uncut coal that also diffuses the incom­ ing simulated methane. Due to safety concern, 70%He/30%CO2 gas mixtures, which has similar molecular weight with pure methane gas, is used to simulate methane in the physical model. The second injection system, shown in Figure 3, is located on top of the model and used to deliver the simulated methane inflow from the fractured strata above the gob. The gas lines are 12.7 mm in diameter and each injection point to the gob contains a needle valve to allow gas inflow adjustment. Two flow sensors and a CO2 gas sensor are installed on six the gob carts to analyze flow and CO2 distribution inside the gob area. The location of these six gob carts is interchangeable with other gob cart to analyze different part of the gob. Several more flow and CO2 sensors are installed on the longwall face and mine entries. 3 CFD MODEL ANSYS Fluent v. 18.2 is used for the CFD simulation. The CFD model is developed based on the scaled physical model dimensions and features, such as adjustable face carts, gob carts, and ventilation controls. The model is separated into multiple segments and meshed separately before combining them in the ANSYS Fluent software for the simulation. The model is separ­ ated as follows: • Longwall face consists of 11 face carts, 110 shields, armored face conveyor, and a shearer, which have been simplified compared to the physical model. • Mine entries and ventilation control, such as stoppings and regulators • Gob carts, consists of front gob, back gob, and middle gob with adjustable gob rows • Strata gas injection system consists of the gas injection system and strata

39

Figure 3. Strata simulated methane gas injection system. a) overview of the gas system. b) and d) close-up view of a single distribution line. c) close-up view of an injection point into the strata.

Figure 4 shows an example of the model geometry separation for meshing purposes. The arrows indicate the interface connection once the model is assembled in the Fluent software. In the physical scaled model, a porous foam will be used to represent the uncut coal, as shown earlier in Figure 2 (right). In the CFD model, the uncut coal is modeled as a porous medium with a source term assigned to supply a certain amount of simulated methane into

Figure 4. right).

Model geometry separation- Mine Entries (left), Longwall face (top right), and gob (bottom

40

the longwall face area. The gob and strata height can be adjusted within the 61 cm available space inside the model. The strata injection system can also be adjusted to deliver the methane gas either to the top of strata, or directly into the top of the gob. In the physical model, porous foam is used to represent the strata. In the CFD model, a porous zone with adjustable viscous resistance is used to model the strata. Figure 5 (left) shows a close-up view of the longwall face area and porous zone that repre­ sents the uncut coal, while Figure 5 (right) shows the strata injection system. Depending on the focus area of the study, certain sections of the model will be re-meshed to refine the area of interest. For example, if the focus of the study is to analyze flow pattern and gas distribution in the longwall face area, the longwall face cart mesh will be refined and the strata injection system will be simplified to reduce the computational time. Table 1 shows the mesh summary for the CFD model with the gob modeled as a porous medium. The model can be re-assembled to simulate different gob sizes and ventilation scenarios, as shown in Figure 6. The following assumptions are currently used for the CFD model, and will be refined as more data becomes available from the experiments in the scaled physical model: • The stopping is modeled as a porous medium with an adjustable viscous resistance value of 1x1013/m2, used to represent stopping with minimal leakage. It can be changed to solid objects to represent a seal in U-type ventilation. • The regulators are modeled as a porous medium with adjusted viscous resistance value to reduce mesh requirement the computation effort of having to redraw the model each time to adjust the flow. The viscous resistance value for each regulator is adjusted to match the flow passing through the regulator in the physical model for the tested ventilation scenario. • The gob and strata zone are modeled as a porous medium with adjusted viscous resistance. For a detailed study on a specific part of the gob, partial discrete gob modeling can be done by replacing one or more gob carts with a discrete model. Table 2 shows the model settings used for the simulation.

Figure 5. Geometry showing a porous medium that represents uncut coal (left) and cross-section view of the strata injection system (right).

Table 1. Mesh summary. Segment

Cell size

Mesh control

Cell number

Mine entries Face carts Uncut coal face- 11 carts Gob carts (for one row) Strata + injection system

2 – 16 mm 1 – 15 mm 5 – 15 mm 5 – 20 mm 0.75 – 24 mm

6 mm for stoppings and regulators 5 mm for coal face 5 mm for connection with face cart 5 mm for interface with face carts 3 inflation layers for injection line

2.4 million 7.5 million 40,000 4 million 6.6 million

41

Figure 6. Schematic of 5 rows gob for a bleeder (left) and progressive sealed (right) ventilation system. The red line represents face curtain extending across the first face cart, while the green and yellow colored rectangle represent regulators and stoppings, respectively.

Table 2. ANSYS model setting for CFD simulation. Parameter

Setting

Time Solver Flow density Species transport Turbulent model Solution methods Convergence Residuals

Steady-state Pressure based Incompressible Methane – Air mixtures and simulated methane (He/CO2) Realizable k-ε with standard wall function SIMPLE scheme with second-order discretization 1 x 10-4 for continuity and momentum, 3 x 10-5 for turbulence, 1 x 10-5 for gas species, and 1 x 10-6 for energy Mass flow inlet and Pressure outlet Gob, strata, and uncut coal face Pressure: 80,560 Pa; Temperature: 297 K

Boundary condition Porous zone Operating condition

4 CFD SIMULATION OF METHANE DISTRIBUTION For the model setup, mass flow inlet boundary condition is used for the inlets at mains and strata injection line, while pressure outlet boundary is used for the model outlet at the bleeder. The main inlet provides 0.11 m3/s of fresh air. The amount of 100% CH4 gas sup­ plied from the strata injection lines is set to 0.0011 m3/s, which is approximately 1% volume of the fresh air supplied from the main inlet. An additional 0.0011 m3/s methane gas quan­ tity is set to be supplied from the uncut coal face in a uniform manner, which results in 0.0022 m3/s total methane flow into the model. The modeled gob height is 30.5 cm, while the strata is 7.6 cm high. Both the gob and strata are modeled as porous medium zones. For the gob resistance, the gob edge area resistance is assigned a viscous resistance value of 6.5 x 106/m2, while a viscous resistance value of 1.05 x 107/m2 is assigned to the 5 rows of gob in the middle. These viscous resistance values are based on the physical test done with 58 mm and 38 mm diameter balls packed in random manners. Different gob resistance values will be tested to check the trend. The strata is assigned a viscous resistance value of 1 x 108/m2, which is based on a test done on foam material used to represent strata in the physical scaled model. 42

Figure 7. Velocity streamlines (left) showing flow distribution for bleeder ventilation model, only show­ ing the path of fresh air supplied from the main inlet. Red line represents the face curtain extending across the first face cart, while the green and yellow colored rectangle represent regulators and stoppings, respectively. Volume rendering of methane mole fraction (right).

Figure 7 (left) shows the streamlines of methane from the injection pipes colored by methane mass fraction, while Figure 7 (right) shows the methane distribution inside the model in the longwall face and in the gob and strata region. The results show that continuous leakage occurs across the longwall face as the airflow travels from the headgate to the tailgate side of the face. Parametric study varying flow rate from the main inlet and gob resistance show different leakage rates across the face ranging from 40% - 80% leakage. This leakage rate is within the expected leakage rate typically observed in the real mine, which is around 50% (Krog, 2016) or up to 70% Thakur (2006). Other factors such as gob resistance can also significantly affect the leakage rate, which will be tested using both the physical and CFD models. Some of this leaked air re-enters the bleeder entries through the crosscuts inby the face on both side of the gob, while the rest travel to the back of the gob, which is the typical flow pattern in bleeder ventilation system (Juganda, 2020; Gangrade et al. 2019). Methane distribution in the model, Figure 7 (right), shows a high concentration of methane on the strata and top part of the gob, where the methane is injected. Most of the methane in the lower gob area, especially on the headgate side, has been diluted and swept by the fresh air leaking from the longwall face, while accumulation at lower concentration can still be seen on the tailgate side of the gob.

5 CFD SIMULATIONS OF SIMULATED METHANE Injecting a large volume of methane gas into the physical scaled model can create a safety hazard. As an alternative solution, the methane gas is replaced with a non-explosive gas mix­ ture when conducting experiments to analyze airflow distribution and simulated explosive gas concentrations inside the gob and longwall face. An important parameter that requires verifi­ cation, is the behavior of the alternative non-explosive gas transport compared to the methane gas when used in the physical model. As the methane replacement, a 70%He/30%CO2 gas mix­ ture was chosen due to having a similar molecular weight (16 g/mol) with pure methane gas. The simulated methane gas transport was investigated using the CFD model and the results are shown in Figure 8 and Figure 9. Note that although the strata were simulated, for visual clarity the methane in the strata is not included in the volume rendering to emphasize the gas distribution inside the gob area where the gas sensor will be installed. The methane concentra­ tion in the strata zone can be seen in Figure 7 (right). The results with the 70%He/30%CO2 show a similar gas distribution trend inside the gob and longwall face area, especially where the shearer is located, when 100% methane is injected 43

Figure 8. Comparison of CH4 (left) and CO2 (right) distribution for 70% He/30% CO2 simulated methane.

Figure 9.

Close-up view of longwall face area showing volume rendering of CH4 (left) and CO2 (right).

into the model. Since 70%He/30%CO2 gas mixture has approximately similar molecular weight with 100% CH4 gas, the equivalent carbon dioxide concentration by volume should be approximately 30% of the methane concentration by volume, which is shown by the predicted 1.9% CH4 and the equivalent 0.57% CO2 reported at the model outlet in Figure 8. Looking at the methane concentration, the area of concern in the longwall face section is around the tail­ gate corner. As methane continues to accumulate across the longwall face, the highest methane concentration can be found in the last face cart at the tailgate corner. With the current ventila­ tion scenario and shearer located in the middle of the longwall face (face cart-6), Figure 9 shows methane concentration within the explosive range, ~5 – 15% CH4 by volume, forming around the shearer tailgate drum. The volume of this explosive gas zone is expected to increase as the shearer travel towards the tailgate. The methane distribution inside the gob shows that the leaked fresh air from the longwall face seems to be sufficient to dilute the methane in the headgate side of the gob. However, high concentration of methane can be observed around the top and edges of the gob, along within some crosscuts on the back corner of the gob. This flow pattern and gas distribution will be verified using the physical model. Different airflow scaling can significantly change the overall methane distribution inside the gob and longwall face area. For this project, the flow scaling was done using geometric, kinematic, and dynamic scaling based on Reynolds number for the flow inside the face area (McMack et al., 2018). 44

These initial model results indicate the viability of replacing the methane gas with 70%He/30%CO2 gas mixtures for safety reasons. Instead of CH4 sensors, CO2 sensors are installed in the physical model, and the equivalent CH4 concentration can be calculated from the CO2 readings. 6 CONCLUSIONS AND FUTURE WORK A CFD model is used to aid in the design and development of a 1/40th scaled physical model of a longwall coal mine. Simulation results are used to check critical sensor placement loca­ tions and the viability of using other gases to simulate methane. The applicability of the modeling technique and assumption used in the current CFD model, such as modeling gob as a porous medium, is currently being investigated along with modeling the gob as discrete obstacles. Following the physical experiments using the physical scaled model, the resulting airflow pattern, leakage rate, and gas distribution will be used to validate the CFD model. Once validated, the CFD model can be used for parametric studies to check critical parameters and reduce the time and cost required to run physical experiments. This coupled modeling approach can be used be scaled up to develop more robust full-scale longwall coal mine CFD models that aid in the improvements of ventilation and monitoring practices to reduce the risk of longwall mine explosions. For future work, the CFD model can be used to simulate different ignition scenarios, such as face ignition by the shearer cutting drum or in-gob ignition behind the longwall shields. Different gob modeling methods, such as partial or full discrete gob modeling (Juganda et al., 2020), can be incorporated into the CFD model for more accurate gob representation in the physical scaled model and allow in-gob ignition simulation. ACKNOWLEDGEMENT This research was conducted with financial support by the National Institute for Occupational Safety and Health (NIOSH) under contract number 200-2017-94491, and the use of Colorado School of Mines Wendian High Performance Computers. REFERENCES Juganda, A., Brune, J.F., Bogin, Jr., G.E., Strebinger, C. 2020. Discrete Modeling of a Longwall Coal Mine Gob for CFD simulation. International Journal of Mining Science and Technology, Vol 30, p. 463–469 Juganda, A. 2020. Evaluation of Point-based Methane Monitoring and Proximity Detection for Methane Explosive Zones in Longwall Faces of Underground Coal Mines. Dissertation, Colorado School of Mines. Gangrade, V., Schatzel, S.J., Harteis, S.P., and Addis, J.D. 2019. Investigating the Impact of Caving on Longwall Mine Ventilation Using Scaled Physical Modeling. Mining, Metallurgy & Exploration. Krog, R.B. 2016. Critical analysis of longwall ventilation systems and removal of methane. Dissertation, West Virginia University. Marts, J., Gilmore, R., Brune, J., Bogin, G., Grubb, J., and Saki, S. 2015. Optimizing nitrogen injection for rogressively sealed panels. SME annual meeting. Denver, CO. McMack, J., DeRosa, C., Zurhorst, M., Bogin, G.E., and Brune, J. 2018. Design and construction of a 1/40th scale longwall mine model for physical testing of methane explosions. SME Annual Meeting. Denver, CO. Pinheiro, H., DeRosa, C., Juganda, A., Wilson, F., Bogin Jr., G., Brune, F., Gallagher, K., Sandoval, N., Gilmore, R., Shapen, N., and Rozendaal, J. 2020. An optically accessible 1/40th scaled dynamic ventilation model of a longwall coal mine. SME Annual Meeting. Phoenix, AZ. Page, N.G., Watkins, T.R., Caudill, S.D., Cripps, D.R., Godsey, J.F., Maggard, C.J., Moore, A.D., Morley, T.A., Phillipson, S.E., Sherer, H.E., Steffey, D.A., Stephan, C.R., Stoltz, R.T., Vance, J.W. and Brown, A.L. 2011. Report of Investigation, Fatal Underground Mine Explosion, April 5, 2010,

45

Upper Big Branch Mine-South. Performance Coal Company, Montcoal, Raleigh County, West Vir­ ginia. ID No. 46-08436. Mine Safety and Health Administration. Arlington, VA, 2011, p. 965. Phillips, C.A. 2012. Report of Investigation into the Mine Explosion at the Upper Big Branch Mine April 5, 2010, Boone/Raleigh Co., West Virginia, West Virginia Office of Miners’ Health, Safety & Training, Charleston. Saki, S.A. 2015. Gob ventilation borehole design and performance optimization for longwall coal mining using computational fluid dynamics. Doctor of Philosophy Dissertation, Colorado School of Mines. Thakur, P.C. 2006. Optimum widths of longwall panels in highly gassy mines-Part 1. 11th U.S. Mine Ven­ tilation Symposium. Pennsylvania State University, PA.

46

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Integration of conjugate porous media model into mine ventilation network software P.H. Agson-Gani Department of Mining and Materials Engineering, McGill University, Montreal, QC, Canada

L. Amiri* Department of Mechanical Engineering, Université de Sherbrooke, Sherbrooke, QC, Canada

S.A. Ghoreishi Madiseh NBK Institute of Mining Engineering, University of British Columbia , Vancouver, BC, Canada

S. Poncet Department of Mechanical Engineering, Université de Sherbrooke, Sherbrooke, QC, Canada

F.P. Hassani & A.P. Sasmito Department of Mining and Materials Engineering, McGill University, Montreal, QC, Canada

ABSTRACT: Understanding the behavior, characteristic and interaction of airflow and air/ gases leakage through broken rock zones, tunnels and underground mine openings is still a challenging research area due to the complex flow structure through the dynamic nature of underground mine environment. It is crucial to accurately calculate the ventilation parameters when modelling the underground mine ventilation systems. The Atkinson equation along with Kirchhoff laws solved using the Hardy-Cross algorithm has been commonly used in mine ven­ tilation network (MVN) software. However, this model cannot be directly applied for coupled airways with flows through porous media, i.e. air leakage in broken/fragmented rocks, gob of coal mine and caved zone. This paper develops a novel friction factor model that can be easily integrated into typical MVN software to predict the conjugate flow through a broken rock structure. The model is further verified against conjugate 3D computational fluid dynamics model. The results show that the model enables accurate prediction of interaction between flow in underground opening and flow through broken rocks that is extremely useful for prac­ tical mine ventilation application. Keywords:

porous media, mine ventilation, CFD, broken rocks, underground mines

1 INTRODUCTION Block (gravity) caving and longwall are two of the most cost-effective and high efficient under­ ground mining practices. In both methods, understanding the characteristics of the blasted/ broken zone is crucial to design the proper mine ventilation system (Amiri et al. 2020). In under­ ground coal mines, the main production takes place in the longwall face (Wang et al. 2018). Thus, the majority of gases and dust are accumulated in this area which requires to be properly ventilated to prevent any gas explosion disaster. Additionally, the gas leakage through the gob and caved zone is of vital importance to achieve safety in underground coal mines (Yueze et al. *The first and second authors contributed equally. DOI: 10.1201/9781003188476-5

47

2017). In the literature, (Karacan 2008) studied the 20 years of ventilation emission data from 63 longwall underground mines. They stated that a large number of variables such as coal and rock properties are involved in predicting the gas emissions in these mines. It should be noted that the calculation of the airflow resistance factor, along with the friction factor in porous media, is one of the most challenging tasks in designing and evaluating the underground mine ventilation sys­ tems. According to the previous study (Amiri et al. 2019, 2020), rock size, porosity, velocity, and dimension of the shaft/stope are the key parameters in defining the friction factor and flow resist­ ance in blasted shafts. In addition, the phenomena of fluid flow through the blasted part is not well understood due to the dynamic and complex nature of underground ventilation. For example, the permeability and porosity of blasted stope as well as broken rock size change during the block caving procedure and effects on the airflow patterns in underground openings (Poulsen, Adhikary, and Guo 2018). The initial permeability is increased logarithmically by increasing the particle size, despite the exponentially decreasing of pore compressibility in the same situation (Zhang and Zhang 2019). Mine ventilation network (MVN) software such as VNetPC, Ventsim, or Vuma are widely used to design and evaluate underground mine ventila­ tion systems. Atkinson equation, along with Kirchhoff laws solved using Hardy-Cross algorithm, has been commonly used in mine ventilation network software. However, the commercial soft­ ware fails to address the porous structure and assign accurate properties to the zones filled by broken rock (Liu et al. 2020; Pan et al. 2019; Wang et al. 2018). Moreover, this model cannot be directly used to couple airways with flow through porous media, i.e., air leakage in broken/frag­ mented rock, gob of coal mine, and caved zone. Inaccurate assumptions of airflow resistance and behavior through the porous zones lead to some fundamental problems in designing the mine ventilation network, such as unknown gas leakage stream, accumulation of flammable gases, and higher ventilation cost. Generally, the characteristics of the porous media can be deter­ mined by i) direct in situ measurement; ii) analytical models; and iii) Pore-scale Computational Fluid Dynamics (CFD) models. It should be noted that the direct in-situ measurement in the caved zone is practically very challenging due to the inaccessibility of this zone in the under­ ground mine. Hence, the analytical models and pore-scale CFD models are the principle tools to study the airflow behavior inside the broken rocks in blasted shafts. CFD solver, such as Fluent, is proven to model fluid flows under complex condition accurately. In exchange for more detailed airflow data, the 3D solver needs more time and computational cost to run. Therefore, the know­ ledge gaps in this field and the limitations of MVN software define the main objective of this study. Hence, the objective of this study is to propose an innovative strategy by using the analyt­ ical solution into CFD solver and the MVS software. This study aims to bridge between CFD simulation and MVS software, which needs less time and computational power into the CFD simulation, which produces more accurate simulation. The bridging between those simulations is found by comparative analysis with the proposed analytical solution derived from pore-scale CFD reported in our previous study (Amiri et al. 2019, 2020) as input data to improve VentSim simulation by converting Fluent data into a correction factor. In this paper, we use the proposed analytical solution of (Amiri et al. 2020) to estimate the characteristics of fractured zones during the mining activities by considering a wide common range of porosity (ε) from 0.2 to 0.6 and rock diameter from 4 cm to 55 cm in the mining, oil and gas and construction industries. Next, the inertial and viscous resistance coefficients for 20 differ­ ent porous media are used in the CFD model to evaluate the airflow behavior, pressure drop, and air leakage according to practical requirements. Then, the friction factor of porous (fpor) correl­ ated to the porous media is encapsulated into an MVN software, Ventsim, suite to evaluate the gas leakage and airflow behavior. Lastly, the results of these solutions are quantitatively evaluated and compared. The results of this study provide useful suggestions on how to model the blasted/ fractured zone in MVN software to accurately estimate the amount of gas leakage and pressure drop through the mentioned zone and to ensure safety in underground mines.

48

2 MODEL DESCRIPTION This study compared the numerical simulations using ANSYS Fluent and Ventsim in order to validate the friction factor for porous media inside the mine ventilation systems. For this study, a haulage drift with a drawpoint is chosen as the physical domain of flow through porous media in an underground mine simulation, as shown in Figure 1(a). The drawpoint in the model acts as porous media, which creates resistance in the mine ven­ tilation system. The airflow goes from the duct inlet into the drawpoint, then a certain amount of air seeps into the blasted zone, and some of the air returns along the outlet drift. The duct diameter is 0.6 m, and the dimension of the drift and the drawpoint are 4�5 m² and 5�5 m², respectively. The process of integrating the conjugate porous media into the mine ventilation software is shown in Figure 2. 2.1 Analytical solution Friction factor determination is needed during mine ventilation airways calculation. Mine ventilation engineers currently use the established friction factor from the Moody chart or the compilation of mine ventilation surveys (McPherson, 2009). Additionally, a new friction factor correlation for the flow through broken rocks was recently developed (Amiri et al. 2020). The friction factor for porous media, fpor, is defined as a function of the local Reynolds number, Rek, and Forchheimer coefficient, F. This friction factor is valid for porosity (ε) in the range of 0.2-0.7 and particle diameters (dp) ranging from 0.04-1.2 m, which could emulate the conditions of broken rocks inside the mine caused by blasting or other mining operations. The fpor correlation can be obtained as:

To utilize the friction factor for porous media in the mine ventilation software, a more gen­ eral Atkinson friction factor is needed. Thus, by applying the geometric characteristic correc­ tion, the Atkinson friction factor could be calculated as:

where α is the specific surface area based on the solid volume, and γ (=Perimeter/Area) is the geometric characteristics which is a function of the perimeter and cross-section area of the airway.

Figure 1. The haulage drift with a drawpoint: (a) Illustration of the flow inside the model (not to scale); (b) Ventsim model; (c) ANSYS Fluent model.

49

Figure 2. software.

Flowchart for integrating porous media friction factor correlation into mine ventilation

2.1.1 Governing equations The model consists of two zones, namely the drift and the drawpoint. The airflow in the drift is a turbulent flow due to the high Reynolds number (89,000-295,000). The equations for the conservation of mass and momentum in the drift write:

where ρ is the density, U is the fluid velocity, P is the pressure, μ is the fluid dynamic viscosity, g is the gravity, and subscripts f and t stand for fluid and turbulent, respectively. – Turbulence model: The current study employs the standard k-epsilon turbulence model. The model requires a two-equation model that solves for the turbulence kinetic energy K, and its rate of dissipa­ tion 2, expressed as:

50

where σK and σ2 are the Prandtl numbers for K and 2, respectively. GK is the turbulence kinetic energy due to the mean velocity gradients. The turbulent viscosity μt is formulated by combin­ ing K and 2 as:

where C 12 , C 22 , and C μ are constants (Nield and Bejan 2017). Meanwhile, the airflow in the drawpoint, which is filled with broken rocks, is assumed to be laminar and steady. The equations for the conservation of mass and momentum in the drawpoint write:

where u is the superficial velocity and the SD and SF are the Darcy term and Forchheimer term, respectively, which can be calculated as:

Permeability K is the inverse of the viscous resistance coefficient α, and can be defined as a function of the particle diameter dp and porosity ε, while β is the inertial resistance coeffi­ cient. They are expressed as:

where η is the passability, constants A and B are Ergun coefficients and were plotted from the graphs developed by (Amiri et al. 2019). 2.2 Initial and boundary conditions

The initial and boundary conditions are listed below:

• At the inlet: the injection flow rate is 20 m3/s. • At the tunnel, drawpoint, and duct walls: the no-slip condition is applied (Eq. 14).

• At the outlet in the drawpoint and drift: the pressure is set as 0 (Eq. 15).

51

2.3 Model development The numerical model for both Ventsim (Figure 1(b)) and ANSYS Fluent (Figure 1(c)) was created based on the schematic of haulage drift with a drawpoint shown in Figure 1. To cap­ ture the correlation between both models, this study runs 20 simulations varying the porosity and the broken rock diameters. These parameters will affect two user-inputs, which are: (1) porous media friction factor correlation (Eq. (2)) to simulate the drawpoint in Ventsim model; and (2) viscous and inertia resistances (Eq. (12) and (13)) to find the viscous and inertia resist­ ance of the drawpoint in ANSYS Fluent model. 2.4 Numerical simulations The computational domain was created and meshed with ANSYS 20.1. A 3D model is used for the simulation, and the mesh-independent solution was ensured by conducting several tests, with a final mesh size of 4�105 elements. The governing equations with the initial and boundary con­ ditions were solved using the finite volume method by ANSYS Fluent. The model was com­ puted with the Semi-Implicit Pressure-Linked Equation (SIMPLE) algorithm and second-order upwind discretization. All residuals were set to 10-6 as the convergence criteria for all equations.

3 RESULTS AND DISCUSSION 3.1 Air leakage and pressure losses in the drawpoint The simulation results of the air leakage and pressure drop for the drawpoint area with 5m height at rock size ranging from 0.04m to 0.55m and porosity (ε) between 0.2 and 0.6 are presented in Figure 3 and Figure 4, respectively. Considering the effect of porosity and rock size, it is evident that increasing porosity gives rise to air leakage through the blasted zone. This phenomenon is attributed to the reduced resistance to airflow. Contrarily, the higher airflow leads to a higher pressure drop through the porous media. Its effect can be observed in Figure 5; more airflow is

Figure 3. Air leakage in the drawpoint of CFD and Ventsim models for a range of porosity ε of 0.2 and 0.6 and rock size dp of 0.04 to 0.55 m.

52

Figure 4. Pressure drop in the drawpoint of CFD and Ventsim models for a range of porosity ε of 0.2 and 0.6 and rock size dp of 0.04 to 0.55 m.

seeping into the drawpoint for the model with the ε value of 0.6 compared to the model with ε value of 0.2. Therefore, the required fan power is larger in the case of larger porosity and rock size due to more leakage through the ventilation network. This can be explained by the fact that larger porosity and rock size lead to smaller friction factor and larger pathways between rocks. On closer inspection, the average absolute percentage deviation for the air leakage and pres­ sure drop observed between the results of the two software (i.e., Ventsim and Ansys Fluent) are less than 12% and less than 3%, respectively. It should also be mentioned that the results from the ANSYS Fluent model reveal higher air leakage and pressure drop compared to the Ventsim model. These discrepancies might be ascribed to different parameters, including (i) placement of the duct. Fluent enables precise modeling of the duct position, while in Ventsim, there is less control regarding the placement; (ii) bending between the drift into the drawpoint. Fluent numerically simulates the resistance when the flow encounters a bending airway, while Ventsim uses shock loss as a user-input to determine the resistance; and (iii) the area of the outlet. In Fluent, some part of the outlet area is intersecting with the area of the duct, which leads to a decrease in the outlet area. However, Ventsim does not assume taking up the outlet area when building the duct. Therefore, it can be stated that the proposed friction factor cor­ relation (Amiri et al. 2020) can accurately predict the porous media friction factor for airflow through broken rocks and can directly be used in 1D-MVN software to estimate air leakage and pressure drop in the blasted/fractured zone in underground mining.

53

Figure 5. Velocity streamlines of CFD simulation for airflow drawpoint with a porosity ε of: (a) 0.2; and (b) 0.6.

4 CONCLUSION This study presented a novel analytical approach to quantify the friction factor in the blasted/ fractured zones in an underground mine. This approach can directly be plugged into 1D-MVN software such as Ventsim in order to accurately estimate the air leakage and pressure drop through such porous media. An analytical solution was employed to calculate the inertia and viscous resistance coefficient as well as the friction factor in the drawpoint by considering a wide range of porosity and rock size. The air flow's behavior through the haulage drift and the drawpoint was investigated by developing a 3D-Ansys Fluent model and 1D-Ventsim model. Good agreement was obtained between the 3D and 1D simulations. The results showed that the model is highly applicable in accurate prediction of the interaction between flow in underground open­ ing and flow through broken rocks (assuming the accurate friction factor). Hence, the proposed analytical correlation in our previous study can accurately estimate the Atkinson friction factor and be plugged into the 1D-MVN software like Ventsim. It was observed that the air leakage and pressure loss (i.e., required ventilation fan power) through the drawpoint increase with the expansion of the porosity and rock size. The results revealed that the porosity of the blasted/ fractured zone has more meaningful effects on the airflow resistance than the rock size. The implementation of such an innovative approach may assist engineers to accurately estimate the air/gas leakage through the blasted/fractured zone in underground mines. ACKNOWLEDGMENT Leyla Amiri acknowledges Claire Deschênes scholarship and Fonds de recherche du Québec ­ Nature et technologies (FRQNT) for supporting this research. Putra Agson-Gani would like to thank Indonesia Endowment Fund for Education for supporting this work.

54

REFERENCES Amiri, Leyla, Seyed Ali Ghoreishi-Madiseh, Ferri P. Hassani, and Agus P. Sasmito. 2019. “Estimating Pressure Drop and Ergun/Forchheimer Parameters of Flow through Packed Bed of Spheres with Large Particle Diameters.” Powder Technology 356: 310–24. Amiri, Leyla, Seyed Ali Ghoreishi-Madiseh, Ferri P. Hassani, and Agus P. Sasmito. 2020. “Friction Factor Correlation for Airflow through Broken Rocks and Its Applications in Mine Ventilation.” International Journal of Mining Science and Technology 30(4): 455–62. Karacan, C. Özgen. 2008. “Modeling and Prediction of Ventilation Methane Emissions of U.S. Longwall Mines Using Supervised Artificial Neural Networks.” International Journal of Coal Geology 73(3–4): 371–87. Liu, Ang, Shimin Liu, Gang Wang, and Derek Elsworth. 2020. “Predicting Fugitive Gas Emissions from Gob-to-Face in Longwall Coal Mines: Coupled Analytical and Numerical Modeling.” International Journal of Heat and Mass Transfer 150: 119392. Mcpherson, Malcolm J. 2009. M. J. McPherson Subsurface Ventilation Engineering. http://www.mvsengi neering.com/ (March 1, 2021). Nield, Donald A., and Adrian Bejan. 2017. Convection in Porous Media Convection in Porous Media. Pan, Y. et al. 2019. “An Investigation of the Effects of Particle Size, Porosity, and Cave Size on the Air­ flow Resistance of a Block/Panel Cave.” In Proceedings of the 11th International Mine Ventilation Congress, Springer Singapore, 82–91. https://doi.org/10.1007/978-981-13-1420-9_8 (March 1, 2021). Poulsen, Brett A., Deepak Adhikary, and Hua Guo. 2018. “Simulating Mining-Induced Strata Perme­ ability Changes.” Engineering Geology 237: 208–16. Wang, Zhongwei, Ting Ren, Liqiang Ma, and Jian Zhang. 2018. “Investigations of Ventilation Airflow Characteristics on a Longwall Face—A Computational Approach.” Energies 11(6): 1564. http://www. mdpi.com/1996-1073/11/6/1564 (February 26, 2021). Yueze, Lu, Saad Akhtar, Agus P. Sasmito, and Jundika C. Kurnia. 2017. “Prediction of Air Flow, Methane, and Coal Dust Dispersion in a Room and Pillar Mining Face.” International Journal of Mining Science and Technology 27(4): 657–62. Zhang, Cun, and Lei Zhang. 2019. “Permeability Characteristics of Broken Coal and Rock Under Cyclic Loading and Unloading.” Natural Resources Research 28(3): 1055–69. https://doi.org/10.1007/s11053­ 018-9436-x (February 26, 2021).

55

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Airflow characteristic curves for a mature block cave mine R. Bhargava, P. Tukkaraja, A. Adhikari & S.J. Sridharan South Dakota School of Mines and Technology, Rapid City, SD, USA

V.V.S. Vytla MSC Software, USA

ABSTRACT: Block/panel caving is an underground bulk mining method that utilizes gravitational force for mining massive, steeply dipping and deep-seated ore deposits at a lower operating cost. Since caving is a dynamic process, the design of an effective ventila­ tion system is a challenging task and therefore, estimation of airflow resistance offered by the broken rock inside the cave is critical. The complex and dynamic nature of caving also makes it difficult to predict the airflow resistance by using traditional approaches. This study investigates the effect of changes in the bulk porosity of the broken rock on the cave airflow resistance using Computational Fluid Dynamics (CFD) approach. The results show an inverse relationship between the cave airflow resistance and the bulk cave porosity. Keywords: media

1

Porosity, Cave airflow resistance, Panel/Block cave mining, flow through porous

INTRODUCTION

Mass mining methods such as block/panel caving tend to be an ideal choice for massive, steeply dipping and deep-seated ore deposits with an ability to extract deposits at high produc­ tion rates and low operating costs. The deposits mined by the caving mining method are gen­ erally disseminated or low grade in nature [1]. Previous studies reported the advantage of block/panel caving mining method in terms of the maximization of net present value for lowgrade ore deposits [2]. Thus, the economics for a low-grade ore deposit also tend to tilt in favor of block/panel caving. An optimally designed block/panel cave mining method can have the lowest operating cost as compared to other underground mining methods subject to condi­ tions of keeping ore dilution under checked limits [3, 4]. Super caves have been defined as those underground cave mines which have a production rate exceeding 25 Mt per annum or approximately 70 kt/day [5, 6]. The evolution of production rates over years with respect to conventional and super caves is shown in Figure 1. The cave mining methods demands a robust ventilation system to support the high produc­ tion rates of the future super caves. Therefore, this research study contributes to the know­ ledge base of block/panel cave mine ventilation. This study investigates a less explored concept of cave airflow resistance that plays an important role in the ventilation system with regards to requirements of air quantity.

DOI: 10.1201/9781003188476-6

56

Figure 1.

Production rates: Open pits, conventional underground mining & super caves [5, 6].

2 LITERATURE REVIEW In panel caving, the cave is assumed to be fully developed when the broken rock reaches the economical ore boundary. Airflow resistance of a cave is defined as the resistance offered by the broken rock pile to the airflow when airflows through the rock pile inside the cave. The porosity of the rockpile is defined as the percentage of air/void space inside the total volume occupied by the rock pile. The porosity of a cave varies as the cave propagates; it plays an important role to buffer an air blast event. Thus, porosity and the broken rock resistance are important param­ eters for modeling airflow inside the cave [7]. Leakage of airflow into the cave and ore passes from the production drifts is also an important phenomenon; it has been mentioned that around 40 percent of air supplied into the production drifts leaks into the cave and the ore passes, so the production drift airflow quantity should be adjusted for accounting these leakages [8]. Several attempts have been made in the past for characterizing airflow through broken rock [9, 10]. Porous media can be modeled as a discrete or continuum model. Discrete modeling offers certain advantages over the continuum model with regards to accuracy and realistic approach but the meshing of the geometry becomes a tedious task in discrete modeling. The advantage of the continuum model is that it is computationally inexpensive but accuracy gets sacrificed in the process [11]. CFD modeling has been successfully used in the past for under­ ground mining applications involving gob characteristics for a longwall gob involving spon­ taneous heating and airflow patterns in bord and pillar mining [12, 13]. In a multi-lift caving operation, the older working areas might be connected to a mature cave[11]. Attempts for pre­ dicting the airflow resistance of a mature panel cave have also been made in the past with limited success. The previous study did not consider the connection of the cave to the older workings which is very unlikely for a mature cave although the study indicated that the flow inside the cave is neither laminar nor turbulent [11].

3 MODEL LAYOUT AND RESEARCH APPROACH This study considered a panel cave continuum model with a cave dimension 375 m x 256 m x 356 m (height x length x width) and nine production drifts, three undercut inlet ducts, eight 57

undercut drifts, eight exhaust drifts. The model also includes 94 drawbells, broken rock region and uncaved in-situ rock inside the cave. These regions are simulated as porous zones. The isometric, side and front views of the panel cave model are shown in Figures 2 and 3. The cave advancement direction is assumed from right to left. Each of the production drifts, undercut drifts and exhaust drifts have dimensions of 4.3 m x 4.3 m. There are three undercut inlet ducts (inside the undercut drifts) of size 1 m x 1 m and are used for ventilating undercut drifts. The computational time has been kept in mind while modeling the airflow through cave zones. Therefore, broken rock zones and in-situ waste rock zones are modeled diagonally in the model consisting of 16 broken rock zones and 2 uncaved in-situ waste rock zones. Each broken rock zone is divided into 3 subregions laterally to conform to the obser­ vation in the field that the boundary of the cave has higher porosity values with larger par­ ticle sizes as compared to the mid-region. Sub-regions for broken rock zone 10 are shown as an example in Figure 3. The bulk cave porosity is calculated by dividing the sum of the porous volume of broken rock from all the regions by the total volume of the cave (exclud­ ing intact rock zone). This study modeled broken rock zones as porous media. The methodology used for this study is shown in Figure 4. Different scenarios of cave resistance are simulated by changing

Figure 2.

Isometric view of panel cave model [14].

Figure 3.

Side and front view of panel cave model [14].

58

Figure 4.

Airflow resistance calculation methodology.

cave porosity values in the model, and the pressure loss (the pressure difference is measured across the cave) obtained from the CFD analysis is plotted against airflow to obtain the equa­ tion for airflow resistance under different porosity conditions. The pressure difference is meas­ ured across the cave. Figure 5 shows the schematic of air leaking into the cave while all the outlets except the exhaust outlet have been closed to calculate the airflow resistance by allow­ ing the air to leak through porous media of broken rock. The undercuts have not been con­ sidered in this study for calculating the airflow resistance as their effect would be minimal for a mature panel cave. This study considers that the matured cave is connected to old workings through the exhaust drifts in the model. A reliable result from the CFD model is highly dependent on the boundary conditions applied to the model. A summary of boundary conditions applied to the panel model is pre­ sented in Table 1. Static pressure condition is applied to the exhaust drift outlet. Natural inflow/outflow condition is used for the rooftop. This condition assumes that velocity and pressure do not change in the normal direction. It is used because the flow is assumed to con­ tinue into other sections of the mine [15]. SC/Tetra (CFD software program) consists of three sets of programs. These include the pre-processor (for creation of computational mesh and setting the boundary conditions for the simulation), solver (for the execution of analysis) and post-processor (for visualization and analyze the results). The simulation study considers a steady-state incompressible and turbulent airflow through the panel cave model. Standard k-ε turbulence model is used for the study to consider the effects of turbulence [16]. Flow field can be obtained by solving momentum and mass conservation equations. Therefore, for obtaining velocity and pressure

59

Figure 5.

Schematic of air leaking into the cave.

Table 1. Boundary conditions. Region

Boundary Condition Type

Value/Condition

Production Drifts (9 Nos.)

Fixed velocity

Undercut inlet Drift duct (3 Nos.) Undercut Drift (8 Nos.) Exhaust Drift (8 Nos.) Roof Top

Wall condition Wall condition Static Pressure Outlet

0.75 m/s, 1 m/s, 1.25 m/s, 1.5 m/s and 1.75 m/s n/a n/a zero Pa Natural inflow/outflow

fields, momentum and mass conservation equations have been solved respectively. A porous media condition is also used to consider the pressure drop of flow through the broken rock [17]. For obtaining reliable results from the CFD simulations, it is also important that the simu­ lation results are grid-independent. Mesh independent study was also conducted for the mature panel cave model. From the analysis of the results with mesh elements ranging from 5.5 million to 15.8 million, it was observed that a further increase in the mesh elements from 11.8 million did not have a significant effect on the results. Hence, 11.8 million mesh elements were used for this study.

4

RESULTS AND ANALYSIS

For calculating the airflow resistance of the broken rock, the mature panel cave model was analyzed under six different bulk cave porosities and five different air quantities through the production drift. Figure 6 represents the measurement of airflow pressure at 60

Figure 6. Pressure measurement at entry and exit of the cave in the production and exhaust drifts at 35% and 56% cave porosity condition (top to bottom).

the entry of the cave inside the production drifts and at the exit of the cave inside the exhaust drifts for cave porosities of 35 and 56 % and an airflow rate of 291 m3/s in the production drifts. For a given porosity of the cave and airflow in the production drifts, the pressure dif­ ference across the cave is calculated and then plotted against the airflow quantity flowing through the cave as shown in Figure 7. For example, for a bulk cave porosity of 35 %, five different air quantities were simulated. Therefore, for six different bulk cave poros­ ities, a total of 30 simulations were performed to develop the pressure-quantity (P-Q) characteristic curves for a mature panel cave mine. Figure 8 shows the variation of the airflow resistance value with respect to the bulk cave porosity; these values are tabulated in Table 2. From Figure 7, the relationship between the pressure difference (Pa) across the cave and the airflow resistance (Nsa/mb), quantity supplied (m3/s) can be summarized by

where, 1:8 � a � 1:9; 7:4 � b � 7:7 and 1:8 � c � 1:9

Equation 1 suggests that the flow inside the cave is neither fully laminar nor fully turbulent.

61

Figure 7.

Pressure-Quantity characteristic curves.

Figure 8.

Airflow resistance vs. bulk cave porosity.

62

Table 2. Cave porosity and airflow resistance. Cave Porosity (%)

Cave Airflow Resistance (Nsa/mb)

21 28 35 42 49 56

0.17 0.08 0.04 0.02 0.01 0.005

5 CONCLUSIONS The airflow characteristics of a block cave mine is examined with the help of CFD using a continuum approach. This study reveals that porosity plays an important role in changing the resistance offered by the broken rock to the airflow leaking into the cave. The airflow resistance increases as the porosity of the broken rock pile decreases. The pressure-quantity relationship for the airflow through the broken rock is different from the Atkinson’s law for the turbulent flow in a regular mine airway. The resistance of the block cave mine changes dynamically with the bulk porosity of the broken rock. This study is valid for a mature panel cave (with no air gap) where the cave is connected to the older working areas. ACKNOWLEDGMENTS The authors acknowledge the financial support from the National Institute for Occupational Safety and Health (NIOSH) (200-2014-59613) for conducting this research. REFERENCES 1. Lovejoy, C., Block caving: Keeping up with caving. Mining magazine, 2012(6): p. 46–64. 2. Lei, Q., et al., Effects of geomechanical changes on the validity of a discrete fracture network represen­ tation of a realistic two-dimensional fractured rock. International journal of rock mechanics and mining sciences, 2014. 70: p. 507–523. 3. Laubscher, D., Cave mining-the state of the art. Journal of The Southern African Institute of Mining and Metallurgy, 1994. 94(10): p. 279–293. 4. Trueman, R., R. Castro, and A. Halim, Study of multiple draw-zone interaction in block caving mines by means of a large 3D physical model. International Journal of Rock Mechanics and Mining Sci­ ences, 2008. 45(7): p. 1044–1051. 5. Brown, E., Block Caving Geomechanics. 2nd edn. The International Caving Study, 2007. 6. C deWolfe and I. Ross. Super caves - benefits, considerations and risks. in Proceedings of the 7th Inter­ national conference and exhibition on mass mining, Sydney. 2016. 7. Vejrazka, C., Northparkes Mines’ Current Air Blast Risk Assessment Practices for Block Caving Oper­ ations. 2016. 8. Brokering, R.D., D.M. Loring, and C.J. Rutter. Practical Implementation of VOD at the Henderson Mine. in 16th North American Mine Ventilation Symposium. 2017. Golden, Colorado. 9. Schafrik, S., The use of packed sphere modelling for airflow and heat exchange analysis in broken or fragmented rock. 2015, Laurentian University of Sudbury. 10. Schafrik, S. and D.L. Millar, Verification of a CFD code use for air flow simulations of fractured and broken rock. Applied Thermal Engineering, 2015. 90: p. 1131–1143. 11. Baysal, A., et al. Prediction of Airflow Resistance of a Matured Panel cave. in 16th North American Mine Ventilation Symposium. 2017. Golden. 12. Yuan, L. and A. Smith. Effects of ventilation and gob characteristics on spontaneous heating in longwall gob areas. in of: Proceedings of the 12th US/North American Mine Ventilation Symposium. 2008.

63

13. Ranjan, M. and S.K. Karan Kumar, Mine ventilation in a bord and pillar mines using CFD. 2013. 14. Bhargava, R., et al. CFD Analysis of the Effect of Porosity, Quantity and Emanating Power Variation on Gas Emissions in Block/Panel Cave Mines. in 11th International Mine Ventilation Congress. 2019. China. 15. Ajayi, K., et al., Computational Fluid Dynamics Study of Radon Gas Migration in a Block Caving Mine. 15th North American Mine Ventilation Symposium, 2015: p. 341–348. 16. Hurtado, J.P., et al., Shock losses characterization of ventilation circuits for block caving production levels. Tunnelling and underground space technology, 2014. 41: p. 88–94. 17. Cradle, User’s Guide: Basics of CFD Analysis. 2015.

64

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Scale modeling, PIV, and LES of blowing type airflow in a deep cut continuous coal mining section A.R. Kumar, K.M. Henderson & S. Schafrik University of Kentucky, Lexington, USA

ABSTRACT: Blowing curtain systems are preferable by many continuous miner operations as they are considered superior in their removal and dilution of pollutants such as methane gas. However, numerous field tests, laboratory experiments, and numerical models also show a massive separation of airflows and the formation of vortices close to the blind headings. Air­ flow recirculation prevents fresh air from reaching the active face. Computational fluid dynamics simulations using Reynold’s Averaged Naviers Stokes (RANS) approach have suc­ cessfully reproduced the observed phenomenon. These steady-state models cannot resolve the transient state eddies observed in the entries of coal mines. Moreover, these are computation­ ally expensive when a high resolution is desired. Reduced scale physical prototypes could be used to mimic the full-scale phenomenon when scaling laws are applied. This paper summar­ izes the scale modeling of airflows on a reduced scale model. The particle image velocimetry (PIV) technique was used to validate the flow separation from the ribs. Large eddy simula­ tions (LES), as presented here, also resolve those transient-state eddies and the flow separ­ ation. Results from PIV tests and LES models set up and run on equipment-free box and slab cuts are presented. Keywords: Large eddy simulations (LES), Computational fluid dynamics (CFD), Particle image velocimetry (PIV), mine ventilation, recirculation

1 INTRODUCTION Underground mining operations accounted for approximately a third (275 MT out of a total of 756 MT) of all coal produced in the United States in 2018. [1]. Of this, room and pillar mining operations contributed about 44% (122 MT out of a total of 275) of underground pro­ duction. Mining operations generate dust, often in concentrations, exposure to which could lead to sickness in miners. Despite the implementation of improved remedial measures to con­ trol dust exposure, the National Institute of Occupational Safety and Health (NIOSH) has reported an increased occurrence of coal workers’ pneumoconiosis in younger, inexperienced miners as well [2]. Dust, in addition to the onset of these debilitating ailments, and methane gas have contributed to deadly explosions underground [3]. Coal dust explosions in the Jim Walter Resources and Upper Big Branch mine in the United States led to many fatalities. Many tragic incidences led Congress to enact laws including the Federal Coal Mine Safety Act of 1969 and the MINER Act of 2006 [4, 5]. The new dust rule of 2014 is one of the latest legislations enforcing safer mining practices and limiting the dust exposure levels of the miners [6]. Operators use a variety of measures to reduce air contaminants including wet-head cutting drums for deep-cuts, machine-mounted water sprays, and physical barriers like line brattices or curtains [7]. The main method, however, for diluting and removing contaminants from DOI: 10.1201/9781003188476-7

65

the face is the ventilation system [8]. Delivering fresh air to the currently mined coal face is inherently difficult due to the geometrical layout of the workings. A significant portion of the airstream directed into the active cutting area undergoes recirculation and separation – a phenomenon where the air is forced to turn back without reaching the coal face. A study at the US Bureau of Mines in the 1960s showed that blowing line brattice systems of ventila­ tion is more effective compared to the exhaust curtain-type systems. Experiments also showed that the ventilation system is most efficient when the closest end of the brattice is within 1.5 m (5 ft) from the face and close to the rib [9]. NIOSH investigated methane levels at the mining face and used orthogonally placed ultrasonic anemometers to better under­ stand the near face flows [10]. Several laboratory experiments and numerical modeling efforts including Computational fluid dynamics (CFD) studies have demonstrated the air­ flow pattern in these room and pillar headings [11–15]. Numerous research efforts have been directed towards face ventilation systems using flooded-bed dust scrubbers as well and means of their improvements [16–18]. Increased mechanization and machine availability boost the productivity of continuous mining sections. The practice of taking an extended-cut or deep cut, extending rooms up to 12.2 m (40 ft), can enhance coal extraction time by reducing the number of continuous miner moves, lowering the duration of auxiliary activities like an extension of ventilation curtains, and reducing roof bolter moves. This practice offers additional challenges to the removal of coal dust and methane from the active face due to the distance from the ventilation controls to the newly created face. As the coal mines move towards deep cuts, ventilation will become increasingly difficult. US Bureau of Mines investigated the application of jet fans to ventilate the deep cuts. Research showed that free-standing jet fans could ventilate cuts exceeding 12.2 m (40 ft). The addition of a check curtain to the setup lowered circulation close to the fan [19]. NIOSH researched ventilation of 40 ft. deep, two-pass, extended cut. Advancing the cur­ tain from 15.2 m (50 ft) to 8.8 (28 ft) resulted in a 600 % increase in air quantity reaching the face [20]. Computational fluid dynamics modeling has been proven to be a powerful tool in mine ventilation research and is preferred due to its versatility and ease of use. Rapid development in computing capabilities, efficient algorithms, and ease of setting up models for computations before expensive laboratory experiments have boosted the appli­ cation of CFD modeling in mine ventilation engineering [21–23]. Because of the difficul­ ties associated with detailed experiments in an active coal mine, CFD modeling and experiments were set up for a 1:12 reduced scale model of the typical room and pillar coal mine entry. A particle image velocimetry study was conducted with airflow adjusted using Froude’s number scaling. This article presents detailed CFD modeling studies of the reduced scale model for airflow patterns after the third and fourth lift of a typical extended-cut sequence. The full-scale mine entry measures 2.1 m x 6.0 m (7 ft x 20 ft). MSHA requires a minimum airflow of 4.25 m3/s (9,000 cfm) reaching the intake end of the pillar line. A constant volumetric airflow of about 7.08 m3/s (15,000 cfm) was assumed to be maintained behind the curtain in full-scale operation in this research. This was done to account for dust and methane dilution. Higher airflows are also observed in large super sections underground. Analysis of airflows using PIV and LES is presented in the following sections.

2 DESCRIPTION OF THE REDUCED SCALE MODEL OF THE SECTION Modeling airflow in large openings is computationally expensive. Running tests in a mine might also interfere with the nominal production activities. The technique of scale-modeling allows researchers to work with prototypes of convenient physical dimensions and scales of time [24]. Scale models have been used for research on mine scrubbers and contact forces from equipment [25–28]. An equipment-free 1/12th reduced scale model of the section as shown in Figure 1 of the room and pillar operation was used for computer modeling and laboratory testing. 66

Figure 1.

Setup of box and slab cuts.

Scaling laws were developed by identifying the underlying dominant forces in the process under investigation. The two major governing forces involved in this flow are the inertia of air motion and the gravity in the turbulent flow regime. Forces were computed for a pocket of air. Buckingham’s π-theorem’s postulates one π-number for these forces. The π-number values for full and reduced scale remains the same. Eq. 1 and 2 show the inertial and gravitational forces on a tiny air volume inside the set-up characterized by the length, l.

where ρ denotes the density of the working fluid (air). The parameters, v, and l are charac­ teristic velocity and length respectively. The most relevant π-number for this scenario is, there­ fore, the Froude’s number as shown in eq. 3.

The π number indicates that the velocity in the miner section should be proportional to the square root of characteristic length as indicated in eq. (4).

This scaling relationship when applied to the reduced scale model for a maximum airflow of 7.08 m3/s (15,000 cfm) to a section in a full-scale mine results in a corresponding flow of about 0.01 m3/s (30 cfm) through the reduced scale model. This airflow quantity was used in CFD modeling and laboratory experiments.

3 PARTICLE VELOCIMETRY EXPERIMENTS A physical model of a typical room and pillar coal mine was constructed at a reduced (1:12) scale. The ribs (sides) and top were made from clear acrylic and glass, respectively. Transpar­ ency was critical for the use of the particle image velocimetry (PIV) system. Through the side of the model, a powerful YAG laser emits a thin sheet of pulsating green light illuminating particles that have been seeded into the air at the inlet. A CCD camera mounted above the model captures images, in quick succession, of the illuminated particles, and those images are further processed to determine the displacement of the particles. This displacement is used to 67

Figure 2.

Particle image velocimetry set-up.

determine the air velocity vectors and to track particles. The experiment set-up is shown in Figure 2. A small centrifugal fan with variable speed control was attached to the end of the exhaust ducting to precisely control airflow into the physical model. Neutrally buoyant particles, aero­ solized olive oil droplets, were introduced into the system using a compressed air system. Syn­ chronized with the laser pulses, two hundred raw image pairs were taken in quick succession. Insight4G software was used to process the raw images into particle velocity vector fields for further analysis. Tecplot 360 software was used to process these velocity vector fields into the instantaneous velocity magnitude contour plots at different instances of the experiment as shown in Figure 3 and Figure 4.

Figure 3.

Contours of velocity magnitude in the box cut, obtained from PIV experiments.

68

Figure 4.

Contours of velocity magnitude in the slab cut, obtained from PIV experiments.

4 COMPUTATIONAL FLUID DYNAMICS MODELING CFD modeling technique mimics the flow inside the reduced scale model. A software numeric­ ally integrates the Navier Stokes’ equations of conservation of mass, momentum, energy in a flow domain. The software scFLOW, version 14, was used for this research. This software has excellent unstructured polyhedral mesh generation capabilities, occupies less memory, and is compatible with high-speed computing. Computations were run on 40 cores in parallel on a high-performance workstation with an Intel® Xeon® E5-2687W v4 processor. CFD models can be set-up to establish steady-state flow fields first long after the air has moved into the model, or transient state simulations mimicking the intermittent steps leading up to the steady-state fields. A steady-state flow profile generated using a dense mesh could explain the average flow profile with good accuracy. The converged steady-state models show the average flow field long after fluids have been introduced in the model. These simulations are also usually based on Reynold’s Averaged Naviers Stokes (RANS) turbulence models and suffer from inherent errors arising out of approximations in the turbulence model itself. These may be incapable of representing the real-life phenomenon accurately even after numerical simulations convergence due to the inherent unsteady, transient, and unpredictable nature of turbulent flows. Three-dimensional CFD models were generated to mimic the flow field in the equipment-free reduced scale model for the laboratory experiments with and without the slab. The continuous miner was not considered here, although the presence of moving machinery would increase the turbulence. The drawings were imported into the CFD software, and all relevant volumes and surfaces were registered. Structured octree elements were generated to manually place the mesh elements. Multiple prism layers were imparted to resolve the nearwall flows and boundary layer phenomenon.

69

4.1 Steady-state flow models CFD models representing flow patterns were generated first. Flux conditions were set-up at the inlet and outlet of the model and impermeable wall conditions on other surfaces. The time derivative terms were set at the second order. The steady-state simulations were run until con­ vergence. The normalized wall distances were examined on the surfaces to ensure good quality elements. Meshes with increasing grid density were generated to simulate the flow. These meshes were adapted for expected flow and gradients in flow parameters. The number of mesh elements for the two setups was about 4.47 and 2.97 million respectively and have been used for all the simulations. Figure 5 shows the vectors colored by velocity magnitude on a horizontal plane at mid-height through the models. The airflow is observed to separate from the curtains imme­ diately as it moves towards the face. Airflow patterns resembling figure eight were observed close to the face. Several other vortices could be observed, including the ones close to the face, the curtain, and the one furthest away from the face which are the most prominent and stable. The presence of several vortices in the system could be associated with an inherently highly intermittent and circulatory flow pattern in this geometry. These patterns also depend on the curtain-rib separation. These patterns indicate that it is extremely difficult to push air close to the face. Most of the airflow turns toward the large low-pressure area and move into the exhaust, because of the incoming speed this air contacts the rib and splits. 4.2 Large eddy simulations (LES) Direct numeral simulation (DNS) is the most accurate approach to resolve the flow fields since all turbulence scales are identified [29]. However, an extremely large grid element count required makes this approach extremely expensive in terms of computational resources presently and is used mostly for fundamental research. Large eddy simulation (LES), is an accurate and computationally intensive modeling tool but has excellent cap­ abilities to resolve eddies and near-wall flows [30, 31]. LES is the technique of resolving

Figure 5. Velocity magnitude contour on a converged model for the box cut (top) and slab cut (bottom).

70

all eddies larger than the characteristic length defined by the mesh. Though this length is larger than the Kolmogorov length scale, a very high resolution of the flow field is obtained. LES could be a suitable modeling approach in cases where the fluid flow is not in equilibrium or massive separation of flow from the walls is observed; this is typical of room and pillar workings [32]. Since the LES technique applies suitable filters to resolve larger eddies and models all smaller vortices lower than the chosen sub-grid length scale, this is within computing reaches nowadays. LES is a transient state approach and a better approximation of generation, growth, and transport of eddies could be observed, unlike steady-state RANS models. Therefore, an accurate time history of the flow profile could also be obtained. Further, since all turbulent flows and associated eddies are threedimensional, no simplification in the geometry of the flow domain is permissible. The LES models presented here also have an average normalized wall distance (y+) lower than 1.0. These models have very fine mesh especially close to the walls. The average Courant number for the entire flow duration of 60.0 s and indicates a rapidly changing velocity magnitude. Figure 6 and Figure 7 show the contours of velocity magnitude on a plane through the model in box and slab cut scenarios. The contours show the flow separation from the ribs. The models show the development and transportation of tiny eddies from the primary airstream and secondary airstream close to the coal face after the airflow is established. The images show the velocity contours at 15.0, 30.0, 45.0, and 60.0 s. The images show the major air streamline moving at a high velocity that turns back before reaching the face. They also show secondary eddies swirling in the opposite direction. Smaller eddies that branch off from these streamlines are also visible.

5 RESULTS AND CONCLUSIONS This research was aimed at detailed studies of airflow patterns in a reduced-scale equipmentfree model of an underground mining section. Computer modeling and particle image velo­ cimetry studies on a reduced-scale model reproduced the turbulent airflow phenomenon observed in underground mines. Unsteady airflows in mines and accompanying turbulence intensity encouraged the authors to use large eddy simulation technique instead of the

Figure 6.

Snapshots of the contours of velocity magnitude at instances in the box cut.

71

Figure 7.

Snapshots of the contours of velocity magnitude at instances in the slab cut.

conventional steady-state approach. LES was established as an accurate methodology that explains the separation of airflows at the curtain and close to the active face. Computer modeling of airflows in a much smaller flow volume enabled the generation of very fine meshes to resolve the flow pattern in detail. The numerical models and laboratory studies clearly show the highly turbulent, chaotic, and unpredictable flow patterns in the reduced scale model. Images of the computer models obtained at 15.0, 30.0, 45.0, and 60.0 s into the flow show the fine eddies on a horizontal plane through the slab and box cut mining sections. Particle image velocimetry experiments were also set up and run with a small fan regulated to produce the scaled airflow. Snapshots of particles seeded in the continuous miner section were recorded at specific intervals. These were processed to generate velocity contours. The contours show pockets of swirling airflows and a strong presence of vorticity. These are observed in much higher resolution in the images from the PIV experiments. Airflow stream­ lines being separated from the ribs are detail which has been shown by many in-mine and laboratory-scale research [7,12]. The research presented here shows a systematic method for modeling mine ventilation sys­ tems. A fundamental airflow analysis is crucial for those set-ups. A mine section is an example where traditional steady-state models have been considered accurate to represent airflows in general, but not where most critical which is at the face. A good resolution requires a huge grid packing density. Experiments on full-scale models and mines are often expensive and cause the operations to be affected. In certain circumstances, they can even be too dangerous to perform. The reduced-scale model enables mimicking the full-scale operations quickly when scaling laws are applied correctly. PIV experiments were used to validate the computer models using the scaled flow parameters. REFERENCES 1. U.S. Energy Information Administration. (2019). Annual Coal Report 2018. U.S. Department of Energy. 2. Division of Respiratory Disease Studies, NIOSH. (2008). Work-Related Lung Disease Surveillance Report. Department of Health and Human Services. Morgantown, WV: Centers for Disease Con­ trol and Prevention, NIOSH.

72

3. MSHA. (2010). Historical Data on Mine Disasters in the United States. Retrieved Dec 10, 2017, from US Department of Labor, MSHA: https://arlweb.msha.gov/mshainfo/factsheets/mshafct8.htm 4. US Department of Labor (2006, June 15). Mine Improvement and New Emergency Response Act of 2006 (MINER Act). Retrieved Nov 24, 2017, from United States Public Laws: https://arlweb.msha. gov/MinerAct/2006mineract.pdf 5. MSHA. (1969, December 30). Federal Coal Mine Health and Safety Act of 1969, Public Law 91-173*. Retrieved December 10, 2017, from United States Department of Labor: https://arlweb. msha.gov/solicitor/coalact/69act.htm 6. United States Department of Labor. (2014, May 1). Lowering Miners’ Exposure to Respirable Coal Mine Dust, including Continuous Personal Dust Monitors; Final Rule. 7. Reed, W., & Taylor, C. (2007, Aug). Development of coal mine face ventilation systems during the 20th century. Mining Engineering, 40–51. 8. Reed, W., & Taylor, C. (2007). Factors affecting the developing of mine face ventilation systems in the 20th century. SME Annual Meeting. Denver, CO: Society of Mining, Metallurgy and Exploration. 9. Luxner, J. (1969). Face ventilation in underground bituminous coal mines: airflow and methane dis­ tribution patterns in immediate face area-line brattice. Technical Report, US Bureau of Mines, Washington D.C. 10. Hall, E., Taylor, C., & Chilton, J. (2007). Using ultrasonic anemometers to evaluate face ventilation conditions. SME Annual Meeting and Exhibit. Preprint - 07–096. Denver, CO: Society for Mining, Metallurgy, and Exploration, Inc. 11. Heerden, J., & Sullivan, P. (1993). The application of CFD for evaluation of dust suppression and auxiliary ventilation systems used with continuous miners. In R. Bhaskar (Ed.), 6th US Mine Venti­ lation Symposium (pp. 293–297). Salt Lake City, UT: Society for Mining, Metallurgy, and Exploration. 12. Wala, A., Jacob, J., Brown, J., & Huang, G. (2003, Mar). New approaches to mine-face ventilation. Mining Engineering, 55(3). 13. Wala, A., Jacob, J., Rangubhotla, L., & Watkins, T. (2005, Oct). Evaluation of an exhaust face ven­ tilation system for a 6.1 m (20-ft) extended cut using a scaled physical model. Mining Engineering, 57(10). 14. Wala, A., Vytla, S., Taylor, S., & Huang, G. (2007, Oct). Mine face ventilation: A comparison of CFD results against benchmark experiments for the CFD code validation. Mining Engineering, 49–55. 15. Organiscak, J., & Beck, T. (2013). Examination of redirected continuous miner scrubber discharge configurations for exhaust face ventilation systems. Transactions of the society of Mining, Metal­ lurgy, and Exploration, 334(1), 435–443. 16. G. Goodman, “Using water sprays to improve performance of a flooded-bed dust scrubber,” Applied Occupational and Environmental Hygiene, vol. 15, no. 7, pp. 550–560, 2000. 17. Wala, A., Vytla, S., Huang, G., & Taylor, C. (2008). Study of the Effects of Scrubber Operation on the Face Ventilation. In k. Wallace (Ed.), 12th U.S./North American Mine Ventilation Symposium. Reno, NV. 18. Zheng, Y., Organiscak, J., Zhou, L., Beck, T., & Rider, J. (2015). CFD analysis on gas distribution for different scrubber redirection configurations in sump cut. Transactions of Society of Mining, Metallurgy, and Exploration Inc., 338(1), 423–432. 19. G. Goodman, C. Taylor and E. Thimons, Jet fan ventilation in very deep cuts - A preliminary ana­ lysis (Report of Investigations 9399), Bureau of Mines, 1992. 20. E. Thimons, C. Taylor and J. Zimmer, “Ventilating the box cut of a two-pass 40 ft extended cut,” Journal of Mine Ventilation Society of South Africa, vol. 52, no. 3, pp. 108–115, 1999. 21. Xu, G., Luxbacher, K., Ragab, S., Xu, J., & Ding, X. (2016, Jan). Computational fluid dynamics applied to mining engineering: a review. International Journal of Mining, Reclamation and Environ­ ment, 31(4), 251–275. 22. Xu, G., Luxbacher, K., Ragab, S., & Schafrik, S. (2013). Development of a remote analysis method for underground ventilation systems using tracer gas and CFD in a simplified laboratory apparatus. Tunneling and Underground Space Technology, 33, 1–11. 23. Petrov, T. P. (2014). Development of Industry-Oriented CFD Code for Analysis/Design of Face Ventilation Systems. Doctorate Dissertation, University of Kentucky, Department of Mining Engin­ eering, Lexington. Retrieved Jan 12, 2018, from 24. Saito, K., & Williams, F. (2014). Scale modeling in the age of high-speed computing. In Progress in Scale Modeling, Volume II (pp. 1–16). Springer.

73

25. Kumar. A.R., Schafrik, S., & Velasquez, O. (2020). Designing, modeling, and laboratory testing of a non-clogging impingement type filter for mining dust scrubbers. Mining, Metallurgy & Exploration, 37, 1911–1918. 26. Ur Rehman, A., & Awuah-Offei. (2020). Resistive force analysis for design of rubber tire loader buckets. American Society of Mechanical Engineering. Proceedings of the ASME 2020 International Design Engineering Technical Conferences and Computers and Information in Engineering Confer­ ence. Virtual. 27. Ur Rehman, A., & Awuah-Offei. (2020). Understanding how speed, tractive effort, digging height, and rake angle affect bucket penetration and resistive forces for rubber tire loaders. Mining, Metal­ lurgy & Exploration, 37, 1423–1435. 28. Ur Rehman, A., Awuah-Offei, & Sherizaheh, T. (2020). Discrete element modeling of scaled bucket excavation. 54th US Rock Mechanics/Geomechanics Symposium. Virtual. 29. Moin, P., & Mahesh, K. (1998, January). Direct Numerical Simulation: A Tool in Turbulence Research. Annual Review of Fluid Mechanics, 30, 539–578. 30. Bouffanais, R. (2010). Advances and challenges of applied large-eddy simulation. Computers and Fluids, 39(5), 735–738. 31. Choi, H., & Moin, P. (2012). Grid-point requirements for large eddy simulation: Chapman’s esti­ mates revisited. Physics of Fluids, 24(1). 32. Piomelli, U. (2014). Large eddy simulations in 2030 and beyond. Philosophical transactions. Series A, Mathematical, physical, and engineering sciences, 372 (2022). doi: 10.1098/rsta.2013.0320

74

Diesel particulate control

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Transient-flow modelling of DPM dispersion in unventilated dead-end crosscuts and control strategy using curtain R. Morla, S. Karekal & A. Godbole School of Civil, Mining, and Environmental engineering, University of Wollongong, Australia

ABSTRACT: The main drawback of diesel-operated vehicles usage in unventilated deadend crosscut is that they generate diesel particulate matter (DPM), a known carcinogenic agent. Since DPM particles are nanometer in size, they do not settle easily under their own weight and will require longer duration to dilute. This paper presents detailed study of DPM dispersion in unventilated dead-ends and control strategy with curtain using field experiments and CFD modelling investigations. Studies conducted with different dead-end crosscut lengths (10 m,15 m, 20 m, 25 m and 50 m). Results show that longer time durations are required to dilute DPM for greater dead-end crosscut lengths. Studies also show that installing curtains will help to a certain extent at the dead-end crosscuts, and a curtain at 45° is better than a curtain at 90°.

1 INTRODUCTION Although the diesel engine was invented by Dr Rudolf Diesel in 1892, the first diesel-powered locomotive was introduced into the Ruhr Coal mine in Germany in 1927, (Harrington and East, 1947). Later, in 1928, a diesel locomotive began to operate in the Witwatersrand gold mines, South Africa, in 1939 they began operations in French, Belgian and British coal mines, and in 1946, American coal mines introduced diesel-powered vehicles into underground mines (Nundlall, 2014, Belle, 2010). Nowadays, most mines around the world use diesel-powered vehicles to transport men, material, ore, waste rock and coal, and for various other mining operations (Rawlins, 2006). Diesel-powered vehicles are more flexible than electric and battery-operated vehicles because they can travel longer distances and between working sections, (Matsui, 2009, Bugarski et al., 2012). Moreover, diesel vehicles are efficient, as evidenced by their ease of maintenance, consistency and durability, which is why many nations depend on these types of vehicles (Daniel, 1984, Morla et al., 2019). The main concern with diesel-powered vehicles in underground mine environments is their exhaust pollutants, including DPM (MDG, 2008). Since underground environments are neces­ sarily confined and have restricted ventilation and enclosed areas, the contaminants generated by diesel engines cannot readily escape, (Morla et al., 2020b, Morla et al., 2020a). Diesel-powered vehicles sometimes need to operate in isolated areas like parking cuddies, foot-wall drives, cut-throughs and other unventilated ‘dead-end’ areas with either restricted airflow or with no ventilation. As 90% of DPM particles range from 3 nm to 30 nm in size and their density ranges from 0.3 gm/cm3 to 1.2 gm/cm3 (Bugarski et al., 2004), DPM particles do not tend to settle easily under their own weight. Hence, a very long time may be required to dilute DPM to acceptable levels in such locations, so that the operators may be exposed to high concentration DPM. This paper outlines the DPM dispersion study in dead-ends using computational fluid dynamics (CFD) and control strategy using a curtain. DOI: 10.1201/9781003188476-8

77

2 EXPERIMENTAL DETAILS 2.1 Details of the field experiment To study the DPM dilution in dead-end crosscuts, field experiments were conducted in one of the Indian coal mines. 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). During the experiments, the flow rate of the instrument was adjusted to 2.83 × 10-5 m3/s (1.7 l/min). Initially, the dead-end was filled with DPM emitted from an LHD. The LHD was removed and the DPM concentration was monitored over time. The air velocity in the adjacent main gal­ lery/level was maintained at 2.52 m/s. Table 1 shows the measured results for dead-end experiments. 2.2 CFD modelling For modelling, the DPM was considered as solid particles and chemical reactions and colli­ sions were not considered. The Boussinesq approximation was invoked to simulate the effects of buoyancy and the standard k-ε model to simulate turbulence. DPM particles diameter was considered as 1e-9 m to 1e-7 m and mean diameter of 1e-8 m. Discrete-phase modelling was used to model the DPM flow patterns in dead-end crosscuts, (Ajayi et al., 2018, Purushotham et al., 2010, Tukkaraja and Bandopadhyay, 2010). DPM particles were treated as an inert material and Rosin-Rammler diameter distribution was used. For physical models, spherical drag law was used as a drag parameter. For stochas­ tic tracking, discrete random walk model with 10 number of tries and 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 particle-phase (DPM) was tracked using the Lagrangian approach. Particle-to­ particle interactions in the DPM were not considered because the particulate volume fraction was 0.5% (dilute). The experimental site is 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° w.r.t the main gallery. Figure 1 shows the mesh generated for the surfaces of the experimental gallery. The sampling point at 0.5 m from the face, 1.2 m from the floor and 3 m from the sides. The computational domain and mesh consist of almost half a million computational cells. The boundary conditions of the model were considered as having an intake air velocity of 2.52 m/s at 300 K. Airflow in the tunnel was treated as a turbulent flow. To model turbulence, the ReynoldsAveraged Navier-Stokes equation was used. In Reynolds averaging, the solution variables in the exact Navier-Stokes equations consist of time-averaged and fluctuated components for velocity components (ANSYS, 2013).

Table 1. Comparison of simulated DPM results with experimental results. Measured DPM concentration (μg/m3) Time [sec]

Experiment [1] Experiment [2] Experiment [3] Average

Modelled DPM concentration (μg/m3)

0 60 120 180 240 300

817 550 417 350 284 230

820 577 428 352 278 222

825 560 425 358 287 235

819 551 422 352 280 231

820 554 421 353 284 232

Difference (%) N/A +4% +2% 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.

78

Figure 1.

Computational domain and mesh.

where ūi and ui 0 are the mean and fluctuating velocity components (i = 1, 2, 3). The Reynolds-averaged Navier-Stokes (RANS) equation was obtained by substituting time and aver­ age velocity in the momentum equation, (Morla et al., 2018):

where -ρ ui ; uj ; is the Reynolds stress, which can be solved using the Boussinesq hypothesis and Reynolds stress models (RSMs). In the Boussinesq hypothesis, Reynolds stress is related to the mean velocity gradient (ANSYS, 2013):

To determine the turbulent viscosity, μt , the k-ε model was used.

where Cμ is a constant, k is the turbulence kinetic energy, and ε is the dissipation rate of k. The turbulent heat transport is modelled using the concept of Reynolds analogy to turbulent momentum transfer. The modelled energy equation is as follows:

79

( ) where k is the thermal conductivity, E is the total energy and τij eff is the deviatoric stress tensor, defined as:

The standard k-ε model is based on the model transport equations k and ε. 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 mathemat­ ically 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 follow­ ing governing equations (ANSYS, 2013):

where Gb is the generation of turbulent kinetic energy due to buoyancy, and Gk is the pro­ duction of turbulent kinetic energy due to the mean velocity gradient. Figure 2 shows the results of the base case (10 m dead-end crosscut). It shows how the con­ centration of the DPM trapped in the 10 m long dead-end crosscut changes time. Table 1 com­ pares an average of three experimental measurements with the CFD simulation results of DPM concentration. Note that there is good agreement between the simulated results and experimental data, although there are some discrepancies between the simulated and measured results can be due to uneven surfaces in the gallery wall that were not considered while modelling; the overall difference varies from – 5% to + 4%.

Figure 2.

Simulated results of 10 m crosscut DPM dispersion with time.

80

3 DPM DISPERSION IN UNVENTILATED DEAD-END CROSSCUT 3.1 Effect of dead-end crosscut length on DPM dispersion To better understand the characteristics of DPM accumulation and dispersion in dead-end zones, simulation studies have been carried out of different lengths of dead-end crosscuts (10 m, 15 m, 20 m and 25 m from the airflow in the main gallery), as shown in Figure 3. The crosscuts are at 90° w.r.t. the main gallery. The width and height of the crosscut and main gallery are 6 m and 2.7 m, respectively. The main gallery is 100 m long and the airflow velocity in the main gallery is 2 m/s. Initially, crosscuts were filled with 820 μg/m3 of DPM. The effect on the concentration of DPM in the dead-end crosscut was monitored with respect to time. Figure 3 shows the DPM cloud at the different dead-end crosscuts. After 600 sec the DPM concentration at 10 m, 15 m, 20 m and 25m from the airway was 147 μg/m3, 267 μg/m3, 340 μg/m3 and 532 μg/m3, respectively. Figure 4 shows the percentage reduction in spot DPM concentration over 15 mins at 10 m, 15 m, 20 m and 25 m from the dead-end crosscuts. To reduce the DPM by 50% takes 4 min,

Figure 3.

DPM distribution after 10 minutes at 10 m, 15 m, 20 m and 25 m dead-end crosscuts.

Figure 4. DPM concentration percentage of reduction with time at 10 m, 15 m, 20 m and 25 m dead-end crosscuts.

81

Figure 5.

DPM concentration after 10 minutes, one hour and 5 hours at 50 m dead-end crosscuts.

6.5 min, 9.5 min and 13 min for 10 m, 15 m, 20 m and 25 m dead-end crosscuts, respectively. The figure shows that if the dead-end crosscut become longer (>20m), the influence of the air velocity in the main gallery to disperse DPM dispersion is very small. Simulation studies were also carried out with a 50 m long dead-end crosscut (Figure 5). The figure shows that after 10 min, one hour and five hours, the spot concentration of DPM was 820 μg/m3, 540 μg/m3 and 100 μg/m3, respectively.

4 CONTROLLING DPM DISPERSION IN DEAD-END CROSSCUT WITH CURTAINS If a diesel-powered vehicle operates in a dead-end crosscut and no secondary ventilation is available, it takes a long time for the DPM to dilute. To help control DPM in such areas, CFD simulation studies were carried out using curtains at the edge of the crosscut and at the main ventilation drive. Curtains are typically used in underground mines to divert venti­ lation air and dilute gas and dust (Zhou et al., 2015, Babbitt and Ruggieri, 1990, Burgess Jr, 1979). In this simulation, the dead-end crosscut was 15 m away from the main drive and at 90° from the main gallery. The width and height of the crosscut and the main gallery were 6 m and 2.7 m, respectively. The air velocity in the main gallery was 2 m/s. A curtain measur­ ing (2.7 m long × 1 m wide) was incorporated into the model. Two curtain orientations were tested. Curtain at 90° (Figure 6a) to the main gallery and (2) Curtain at 45° with the main gallery (Figure 6b). The initial concentration of DPM in the dead-end was set as 820 μg/m3 and transient flow modelling was carried out to simulate the flow evolving over 15 minutes. Figure 7 shows the changes in the concentration of DPM in the dead-end with no curtain, with curtain at 90° and with a curtain at 45°. Figure 8 shows the velocity contours in the

Figure 6.

Location of dead-end crosscut (a) with curtain at 90°; (b) with curtain at 45°.

82

Figure 7.

Dead-end DPM dispersion with no curtain, curtain at 90° and curtain at 45°.

Figure 8.

The velocity countours in 15 m dead-end crosscuts with and without curtains.

dead-end with no curtain, with curtain at 90° and with a curtain at 45°. With curetins, air velocity in the main gallery is high at downstream side of the curtains. Figure 9 shows the concentration of DPM without and with curtains. Figure 9a shows that after 15 min and without a curtain, the concentration of DPM decreased from 820 μg/m3 to 165 μg/m3. With a curtain at 90° to the main gallery or in line with the dead-end crosscut wall, Figure 9b shows that after 15 min the DPM concentration decreased from 820 μg/m3 to 83

Figure 9. The concentration of DPM after 15 mins in 15 m dead-end crosscuts with and without curtains.

111 μg/m3. With a curtain at 45° to the main gallery, Figure 9c shows that after 15 min. the DPM concentration decreased from 820 μg/m3 to 75 μg/m3.

5 CONCLUSIONS This paper describes parametric studied on the dispersion of DPM in unventilated dead-end crosscuts. Studies were carried out for different dead-end crosscut lengths. DPM concentra­ tion in the dead-end crosscut is influenced by its length. For dead-end crosscut modelling, con­ sidered air velocity in the adjacent gallery of 2 m/s and dead-end crosscut lengths of 10 m, 15 m, 20 m and 25 m. Results concluded that to reduce the DPM concentration by 50% for 10 m, 15 m, 20 m, and 25 m length dead-end crosscuts, it took 4 min, 6.5 min, 7.5 min and 13.5 min, respectively. Results also concluded that for a 50 m long unventilated dead-end crosscut, it took 5 hours to reduce the DPM from 820 μg/m3 to 100 μg/m3. Therefore, it is concluded that the longer the dead-end, the longer it takes for DPM to dilute. If auxiliary ventilation in a dead-end crosscut is not possible, the DPM can be diluted with a curtain at the dead-crosscut entrance. This study considered a 15 m long dead-end crosscut with an air velocity of 2 m/s in an adjacent gallery and the initial DPM concentration in the dead-end crosscut of 820 μg/m3. For this modelling, transient flow modelling studies were con­ ducted for three scenarios: (1) no curtain, (2) curtain at 90° and (3) curtain at 45° angle to the dead-end crosscut. Studies concluded that over a 15 min duration, the DPM concentration reduced to 165 μg/m3, 111 μg/m3 and 75 μg/m3, respectively for the no curtain, curtain at 90° and curtain at 45° models. 84

REFERENCES Ajayi, K., Shahbazi, K., Tukkaraja, P. & Katzenstein, K. 2018. A discrete model for prediction of radon flux from fractured rocks. Journal of Rock Mechanics and Geotechnical Engineering, 10, 879–892. Babbitt, C. & Ruggieri, S. 1990. Evaluate fundamental approaches to longwall dust control: Sub­ program E, Longwall application of ventilation curtains. Foster-Miller, Inc., Waltham, MA (USA). Belle, B. 2010. Use of baseline personal DPM exposure data for mine ventilation planning-A South Afri­ can Journey. J. Mine Vent. Soc. S. Afr., 12–16. 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. Bugarski, A. D., Janisko, S. J., Cauda, E. G., Mischler, S. E. & Noll, J. D. 2012. Controlling exposure to diesel emissions in underground mines, SME. Burgess JR, J. V. 1979. Mine ventilation system and elements thereof. Google Patents. Daniel, J. 1984. Diesels in Underground Mining: A Review and Evaluation of an Air Quality Monitoring Methodology, US Department of the Interior, Bureau of Mines. Harrington, D. & East, J. H. 1947. Diesel equipment in underground mining, US Department of the Inter­ ior, Bureau of Mines. Janisko, S. & Noll, J. 2008. Near real time monitoring of diesel particulate matter in underground mines. Safety And Health, 509–514. Khan, M., Usman. 2017. Real-Time Diesel Particulate Matter Monitoring in Underground Mine Atmos­ pheres, Association With the Standard Method and Related Challenges. Matsui, K. 2009. UNDERGROUND MINING TRANSPORTATION SYSTEMS. Civil EngineeringVolume II, 5, 285. MDG 2008. Guideline for the management of diesel engine pollutants in underground environments. MDG-29. Morla, R., Godbole, A., Karekal, S., Bhattacharjee, R. M. & Balasubrahmanyam, N. 2018. Fun­ damental understanding of diesel-operated man riding vehicle DPM dispersion–a case study. Journal of Sustainable Mining, 17, 105–110. Morla, R., Karekal, S. & Godbole, A. 2019. Analysis of Diesel Particulate Matter Flow Patterns in Dif­ ferent Ventilation and Operational Conditions of Underground Mines. Diesel Emissions. IntechOpen. 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. Nundlall, A. R. 2014. Development of a regulatory mechanism for the control of exposure to diesel particu­ late matter (DPM) in the South African mining industry. Purushotham, T., Sastry, B. & Samanta, B. 2010. Estimation of shock loss factors at shaft bottom junc­ tion using computational fluid dynamics and scale model studies. CIM Journal, 1. Rawlins, C. Underground mine ventilation planning, heat loads, and diesel equipment. 11th United States/North American Mine Ventilation Symposium. Taylor & Francis Group, Penn State Univer­ sity, College Park. USA, 2006. 75–81. Tukkaraja, P. & Bandopadhyay, S. 2010. Analyzing shock losses at air-crossings in a mine ventilation network using CFD simulations. The 13. U.S./North American mine ventilation symposium,. ON (Canada): Laurentian Univ., Sudbury, ON (Canada). Mirarco Mining Innovation. Zhou, L., Pritchard, C. & Zheng, Y. 2015. CFD modeling of methane distribution at a continuous miner face with various curtain setback distances. International Journal of Mining Science and Technology, 25, 635–640.

85

NAMVS 2021 – Tukkaraja (Ed.) © 2021 Copyright the Author(s), ISBN 978-1-032-03679-3

Estimating diesel particulate matter using a predictive technique for use in underground metal mine production scheduling J.A. Buaba & A.J. Brickey South Dakota Mines, Mining Engineering and Management Department, Rapid City, South Dakota, USA

ABSTRACT: Many underground metal mining operations use diesel-powered equipment which emits diesel particulate matter (DPM) into the atmosphere posing a health and safety hazard to expose mine workers. The United States Environmental Protection Agency (EPA) and National Institute for Occupational Safety and Health (NIOSH) have classified DPM as a possible carcinogen creating greater need to control exposure to DPM in the workplace. Within current underground metal mine planning practices, ventilation requirements are often considered after the production schedule has been developed, leading to operational challenges in managing DPM levels. We present a method of estimating DPM using artificial neural net­ work (ANN) for use in underground production scheduling. By incorporating DPM produc­ tion from various underground mining activities, the resulting production schedules can help better utilize ventilation and production resources in addition to allowing operations to meas­ ure the impacts of various non-ventilation DPM reduction methods, e.g., biofuels, electric equipment. The results here show that there is significant potential in predicting DPM concen­ trations for underground production activities. This research aims to improve the mine envir­ onment by providing a tool to estimate DPM that can be incorporated into the production schedule, thereby influencing strategic and tactical-level planning decisions.

1 INTRODUCTION Underground metal mining is often conducted through the use of heavy diesel-powered equip­ ment and explosives in tight spaces with limited ventilation. This environment increases the risk of miners being exposed to diesel particulate matter (DPM) and toxic gases. To mitigate this situation, infrastructure, such as a ventilation system, is needed to dilute and flush out these harmful contaminants from the working areas (McPherson, 1993). Diesel-powered equipment emits diesel exhaust (DE) which is a complex mixture of diesel particulate matter (solid fraction) and hydrocarbon gases generated by the incomplete combustion of fuel within an engine. The gaseous components of DE include carbon dioxide, oxygen, nitrogen, water vapor, carbon monoxide, nitrogen compounds (NO and NO2), sulfur compounds, and numer­ ous hydrocarbons. Some toxic gaseous components are the aldehydes (e.g., formaldehyde, acetaldehyde, acrolein), benzene, 1,3-butadiene, polycyclic aromatic hydrocarbons (PAHs), and nitro-PAHs (EPA, 2002). Solid particles present in DE, i.e., DPM, consists of a center core of elemental carbon (EC) and adsorbed organic compounds (OC) which consist of fine particles (diameter < 2.5 μm) including a subgroup with a large number of ultrafine particles (diameter