PROCEEDINGS OF THE 11TH U.S.

/NORTH AMERICAN MINE VENTILATION SYMPOSIUM,

THE PENNSYLVANIA STATE UNIVERSITY, UNIVERSITY PARK, PENNSYLVANIA, USA,
5–7 JUNE 2006

11th U.S./North American Mine

Ventilation Symposium 2006

Editors:
Jan M. Mutmansky
Professor Emeritus of Mining Engineering, The Pennsylvania State University, University
Park, PA

Raja V. Ramani
Professor Emeritus of Mining and Geo-Environmental Engineering, The Pennsylvania State
University, University Part, PA

London/Leiden/New York/Philadelphia/Singapore

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Sponsored by
Underground Ventilation Committee of SME, Inc.
Mine Safety and Health Administration
National Institute of Occupational Safety and Health
Pennsylvania Department of Environmental Protection
U.S. Environmental Protection Agency
The Pennsylvania State University
Howden Buffalo Inc.
Mine Ventilation Services, Inc.
Schauenburg Flexadux Corporation
Shaft Drillers International
Spendrup Fan Company
TLT-Babcock, Inc.

Cover illustration: Courtesy of ACME Group Pty. Ltd.

Copyright © 2006 Taylor & Francis Group plc, London, UK

All rights reserved, except for
M. Gangal, B. Rubeli, D. Young & J.S. Stachulak
Post-field evaluation of diesel particulate filters
© Her Majesty the Queen in Right of Canada, as represented by the Minister of Natural Resources, 2006.

All rights reserved. No part of this publication or the information contained herein may be reproduced, stored
in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying,
recording or otherwise, without written prior permission from the publisher.

Although all care is taken to ensure the integrity and quality of this publication and the information herein, no
responsibility is assumed by the publishers nor the author for any damage to property or persons as a result of
operation or use of this publication and/or the information contained herein.

Published by: Taylor & Francis/Balkema

P.O. Box 447, 2300 AK Leiden, The Netherlands
e-mail: Pub.NL@tandf.co.uk
www.taylorandfrancis.co.uk/engineering, www.crcpress.com

ISBN10 Book: 0-415-40148-8 ISBN13 Book: 978-0-415-40148-7

Printed in Great Britain

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Table of Contents

Foreword XI

Advisory Committee XIII

Dedication XV

Keynote address
The Methane to Markets Partnership: Opportunities for coal mine methane utilization 3
D. Kruger & P. Franklin

1 Ventilation planning
Ventilation on demand: A preliminary study for Ballarat Goldfields NL 11
M.A. Tuck, C. Finch & J. Holden
Justifying ventilation-on-demand in a Canadian mine and the need for process
based simulations 15
S. Hardcastle, C. Kocsis & D. O’Connor
Ventilation planning for mining of the Sarvisuo Lode in the Orivesi Mine 29
A.L. Martikainen
Ventilation upgrade of Minerales Monclova’s Mine 6 35
J.R. Ponce Aguirre
Transition of the Henderson Mine ventilation system to the new lower levels 41
D.M. Loring & B.V. Nelson
Ventilation planning at Minerales Monclova’s Mine 7 47
J.R. Ponce Aguirre
Maximizing the ventilation of large-opening mines 53
R.H. Grau III, R.B. Krog & S.B. Robertson
Optimization of the Red Lake Mine ventilation system 61
K.G. Wallace, Jr., M. Tessier, M. Pahkala & L. Sletmoen
Remote monitoring and automation of a large mine ventilation network 67
K.R. Mutama & M.A. Meyer
Underground mine ventilation planning, heat loads, and diesel equipment 75
C.A. Rawlins

2 Diesel emissions control

West Virginia diesel regulations: A model for the coal industry 83
P.C. Thakur, C. Hamilton & D. O’Dell
Pennsylvania’s diesel equipment program for underground bituminous coal mines 87
W. Bookshar

Copyright © 2006 Taylor & Francis Group plc, London, UK

 An overview of the MSHA industrial hygiene laboratory 91
R.A. Haney, J.P. Seiler, M.H. Wesolowski, M. Fenlock & J. Swank
Instrumentation for diesel particulate matter emissions research 99
S.E. Mischler, A.D. Bugarski & J.D. Noll
Measuring diesel particulate matter in underground mines using submicron elemental carbon
as a surrogate 105
J.D. Noll, S.E. Mischler, G.H. Schnakenberg, Jr. & A.D. Bugarski
Emissions-based maintenance program 111
S. Forbush
USA salt mines and DPM control strategies 115
E.Z. Manos & M. Liby
An integrated approach for managing diesel emissions controls for underground metal mines 121
G.H. Schnakenberg, Jr.
Implementation of diesel particulate filter technology in underground metal and nonmetal mines 127
A.D. Bugarski, G.H. Schnakenberg, Jr. & L.D. Patts
Post-field evaluation of diesel particulate filters 135
M. Gangal, B. Rubeli, D. Young & J.S. Stachulak
DEEP project on evaluation of diesel particulate filters at Inco’s Stobie Mine 143
J.S. Stachulak, B.R. Conard, A.D. Bugarski & G.H. Schnakenberg, Jr.
DPM reductions at underground metal and nonmetal mines using alternative fuels 151
M.J. Schultz, D.M. Tomko & R.L. Rude
The effects of water emulsified fuel on diesel particulate matter concentrations in underground mines 159
J.D. Noll, S.E. Mischler, L.D. Patts, G.H. Schnakenberg, Jr., A.D. Bugarski, R.J. Timko & G. Love

3 Mine dusts
Evaluation of a new real time personal dust meter for engineering studies 167
A.D.S. Gillies & H.W. Wu
Evaluation of SKC Inc. Dust Detective 175
E.E. Hall, R.P. Vinson & J.C. Volkwein
Comparison of three side-by-side real-time dust monitors in a duct using average and peak
display dust levels as parameters of performance evaluation 179
B.K. Belle
Lowering respirable dust at an iron ore concentrator plant through improved ventilation practices 189
A.B. Cecala, J.A. Zimmer, J.F. Colinet & R.J. Timko
A new method to clean dust from soiled work clothes 197
D.E. Pollock, A.B. Cecala, J.A. Zimmer, A.D. O’Brien & J.L. Howell
Impact of fan type for reducing respirable dust at an underground limestone crushing facility 203
G.J. Chekan, J.F. Colinet & R.H. Grau III
Emerging technologies control respirable dust exposures for continuous mining and roof
bolting personnel 211
G.V.R. Goodman, T.W. Beck, D.E. Pollock, J.F. Colinet & J.A. Organiscak
A field demonstration of an innovative spray system on continuous miners for dust control 217
Y.P. Chugh, A. Patwardhan, H. Gurley & R. Carty
Dust control on longwalls – assessment of the state-of-the-art 225
J.P. Rider & J.F. Colinet

VI

Copyright © 2006 Taylor & Francis Group plc, London, UK

 A field demonstration of the JOY Wet-Head Miner technology 233
Y.P. Chugh, A. Patwardhan, H. Gurley, A. Moharana, R. Saha & R. Carty

4 Heat and humidity

Controlling heat loads from cemented mine backfill 243
E. De Souza
Thermo- and psychrometric properties of intake air passing through fragmented strata 251
O.S. Johnson
Controlling personnel heat exposure in Canada’s deep and highly mechanized mines 259
S. Hardcastle
Optimum insulation for engineering control of mine thermal regime around a mine airway
in permafrost 271
H. Wu, S. Bandopadhyay & V.U. Izaxon
Design of energy efficient mine ventilation and cooling systems 279
W. Marx, F.H. von Glehn & R.W. Wilson
Integration of strategic ventilation and heat management planning with 3D virtual reality
simulation of underground mines 285
R. Moreby, J. Whitmore & P. Dunn
Alternative fog removal methods in mine ramps 295
A.L. Martikainen

5 Mine fires
Reducing belt entry fires in underground coal mines 303
W.J. Francart
Mine fires – measuring the efficacy of gas-enhanced foam 309
M.A. Trevits, A.C. Smith, T.P. Mucho, A. Ozment, J.B. Walsh & M.R. Thibou
Fire-extinguishing and explosion-suppression technology for a super-large and shallow working
underground fire zone 317
F. Zhou, D. Wang, Yongjiu Zhang, Yuliang Zhang, X. Li & J.C. Tien
Correlation between methane and fire hazards in abandoned workings of longwall mining 325
B. Madeja-Strumińska & E. Widzyk-Capehart
Sealing a recent United States coal mine longwall gob fire 331
R.T. Stoltz, W.J. Francart, L. Adair & J. Lewis
The successful recovery of the Dotiki Mine after a major mine fire 337
C.R. Wesley III, T.M. Wynne, J.E. Urosek & K.S. Diederich
First Mexican coal mine recovery after a mine fire, Esmeralda Mine 343
M. Santillan
Mine fire management, case study in Pongkor Gold Mine, PT Aneka Tambang Tbk., Indonesia 347
Risono, N.P. Widodo & R.S. Gautama
Queensland mine emergency level exercises assisted by fire simulation 351
H.W. Wu & A.D.S. Gillies

6 Mine seals
Explosion evaluation of mine ventilation stoppings 361
E.S. Weiss, K.L. Cashdollar, S.P. Harteis, G.J. Shemon, D.A. Beiter & J.E. Urosek

VII

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Evaluation of remotely installed mine seals for mine fire control 367
M.A. Trevits, T.A. Gray, P. Glogowski & L.M. Crayne

7 Coalbed/coal mine methane

Prediction of longwall methane emissions and the associated consequences of increasing longwall
face lengths: A case study in the Pittsburgh Coalbed 375
S.J. Schatzel, R.B. Krog, F. Garcia, J.K. Marshall & J. Trackemas
Predicting methane emissions from longer longwall faces by analysis of emission contributors 383
R.B. Krog, S.J. Schatzel, F. Garcia & J.K. Marshall
Effect of scrubber operation on airflow and methane patterns at the mining face 393
C.D. Taylor, J.E. Chilton, E.E. Hall & R.J. Timko
Effect of water sprays on airflow movement and methane dilution at the working face 401
J.E. Chilton, C.D. Taylor, E.E. Hall & R.J. Timko
Coal mine methane utilization options 407
P.M. Sööt, D.R. Jesse & M.E. Smith
Methods and means of recovery and utilization of methane from mines 413
A. Tor, A. Jakubow, N. Szlazak & S. Wasilewski
VAM to electricity – 1st large-scale installation 421
R. Mattus
Development and application of reservoir models for the evaluation and optimization of
longwall methane control systems 425
C.Ö. Karacan, W.P. Diamond, S.J. Schatzel & F. Garcia
Optimum widths of longwall panels in highly gassy mines – Part I 433
P.C. Thakur
Methane and methane control in Chinese coal mines 439
J.C. Tien
Modern CMM drainage strategies 449
D.J. Brunner, J.J. Schwoebel & J.S. Brinton
Cross-linked polymer gel seals horizontal degas boreholes greater than 4,000 feet long 455
S.J. Kravits, G.M. DuBois, J.M. Reilly & J. Kirley
Target Drilling’s long boreholes maximize longwall dimensions 463
G.M. DuBois, S.J. Kravits, J.M. Reilly & T.P. Mucho
Sinking a 990 m surface borehole for methane drainage of a longwall panel in a UK coal mine 469
J.R. Leeming & P.R. Shead

8 Numerical modeling
Ventilation simulation programs MineVent and MFIRE: Updates to advance the technology of
simulation programming 477
R.J. Hardy & Dr. K.A. Heasley
Fire-generated smoke rollback through crosscut from return to intake – experimental and CFD study 483
G.F. Friel, L. Yuan, J.C. Edwards & R.A. Franks
Case study of CFD simulation in mining accident investigation 491
E. Delic, A. Basic & R. Sisic

VIII

Copyright © 2006 Taylor & Francis Group plc, London, UK

 9 Mine gases
Hydrogen sulfide control on a longwall face 499
M.P. Valoski
Early detection of flammable gas danger in hard rock mines to protect miners 503
F.J. van Zyl, J.J. Dekker & J. Haarhoff
CO dispersion from a coal fire in a mine entry 511
J.C. Edwards, R.A. Franks, G.F. Friel & L. Yuan
The application of open path infra-red detectors in underground coal mines 519
C. Fowler & D. Chalmers

10 Mine fans
Selection of mixed-flow fans for PT Freeport’s DOZ Mine, West Papua, Indonesia 527
I. Duckworth, I. Loomis, K. Karmawan, C. Chmura & K. Lownie
Fan selection for large-opening mines: Vane-axial or propeller fans – which to choose? 535
R.B. Krog & R.H. Grau III
Design and operational aspects in the use of booster, circuit and auxiliary fan systems 543
D.J. Brake & C.A. Nixon
Recirculation in the potash mines of Saskatchewan 555
M. Detharet
Blind shaft drilling as a safer alternative to conventional sinking 559
A.J. Zeni & W.J. Maloney

11 Mine stoppings
Stoppings: Technology developments and mine safety engineering evaluations 565
H.C. Verakis
Development of new protocols to evaluate the transverse loading of mine ventilation stoppings 569
T.M. Barczak & T.J. Batchler
Novel stopping designs for large-opening metal/nonmetal mines 579
R.H. Grau III & G.M. Meighen

12 Leakage and gob flows

Ventilation surveying and modeling of longwall bleeder and gob areas 587
B.S. Prosser & N.L. Oswald
Computational fluid dynamics study on the ventilation flow paths in longwall gobs 591
L. Yuan, A.C. Smith & J.F. Brune
Study of leakage flow in US underground coal mines 599
F. Calizaya & M. Stephens

13 Ventilation education and training

Developments in mine ventilation legislation, statutory qualifications and ventilation
training in Australia 609
D.J. Brake

IX

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Ventilation education and training 619
D. Chalmers
Observations on mine ventilation education in the United States 623
C.R. Hairfield & K.G. Wallace, Jr.

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Foreword

Since its inception in 1982, the U.S. Mine Ventilation Symposium has become a major venue for the presentation
of technical papers in the mine ventilation field. Biannual meetings of the ventilation professionals from around
the world, these symposia provide a forum for exchanging information on technical and practical issues that are
relevant to mine ventilation engineers and mine managers. During the last decade, the symposium was renamed
the U.S./North American Mine Ventilation Symposium to include other ventilation engineers from around North
America. This new designation took on new meaning as Queens University in Kingston, Ontario, Canada, hosted
the first of this symposia series to be held outside the United States in 2002. The Pennsylvania State University
is pleased to host this prestigious symposium, the eleventh in the series, for the second time. In 1987, the third
symposium was held at University Park with 87 papers and over 200 participants.
Mine ventilation is a very broad field encompassing the identification and control of contaminants in the
mine atmosphere, the planning and design of mine ventilation systems to achieve proper air quality and air
quantity at all locations, and the considerations of ventilation and safety in the event of emergencies. Events in
the mining industry in recent years have reinforced the important role that mine ventilation plays in health, safety
and productivity of miners, and the tremendous costs that may be incurred when adequate attention is not paid
to the planning, designing, monitoring and maintenance of the ventilation system. The last two decades have
witnessed profound changes in the mining industry with regard to mining operations and mine environment.
These in turn have major implications for mine ventilation engineers.
It was our aim to incorporate discussions of these emerging trends and technology in both the symposium
activities and the technical program and proceedings. We have programmed three keynote speeches on major
issues: mine health and safety, methane drainage, and ventilation in the future. The keynote speech on methane
drainage, presented by Dina Kruger and Pamela Franklin, is included in the proceedings. The other two keynote
addresses are “Challenges Facing Today’s Coal Industry” by Peter Lilly and “Recent Advances in Ventilation and
Ventilation Down the Road” by Stephen Hardcastle. We have planned four field trips to various sites where new
technology and research methods are being developed, used or implemented for more effective mine ventilation.
Also planned are four pre-symposium short courses focusing on advances in four distinct areas of ventilation.
The technical sessions are a blend of traditional mine ventilation topics such as mine gases, respirable dust
control, heat and humidity issues, ventilation planning and case studies, and major emerging areas such as coal
mine methane capture and utilization, diesel issues for underground coal and metal mines, and mine ventilation
education. The program reflects the international nature of mine ventilation and our authors represent eleven
countries. As has become a practice, we welcome a number of exhibitors who provide literature and assistance to
the participants with regard to their products and services. We hope that this symposium and these proceedings
will help mining engineering personnel to achieve a greater level of effectiveness in their professions.
Anyone who has been involved in the planning and organizing of one of these symposiums knows that it
requires the co-operation of a large number of persons and organizations to make it a success. First, we were
delighted to be selected by the Underground Mine Ventilation Committee of SME to host the eleventh symposium
and thank the members for the confidence they had exhibited in our ability to host the symposium. All of these
committee members were available to provide assistance and advice during our work on the symposium. As you
can see, we have a number of sponsors for our symposium. We are pleased to report that these sponsors provided
advice and support with regard to all symposium activities. We thank them for their vital input in several areas
of our program such as seminars, field trips, technical programming, and exhibits.
We had an outstanding team of volunteers who have rendered yeomen service to our activities. We extend our
sincere thanks to the members of our Advisory Committee who have contributed to the various tasks of reviewing
abstracts and program planning by being responsive in a timely manner. We were pleased that John Kelleher,
President of Schauenburg Flexadux Corporation, agreed to undertake the tasks associated with managing the
exhibits at this symposium, for which we are much obliged. We also thank the exhibitors for their financial
support for some of the social functions and activities as well as providing materials and exhibits for a more
informative and congenial meeting. The services of the short course instructors and field trip sponsors who
went out of their way to provide supplemental activities to symposium participants are much appreciated and
recognized.

XI

Copyright © 2006 Taylor & Francis Group plc, London, UK

 It would not be too far from the truth to acknowledge the fact that without the help of Rachel Altemus, our
Writer/Editor and Alumni Coordinator, we would not have been able to organize this symposium. We extend our
sincere gratitude for her devoted service with regard to all of the correspondence, editorial needs, and coordi-
nation activities required for this conference. We also thank Ms. Anna Morrison and the Department of Energy
and Geo-Environmental Engineering for their support of our efforts and Ms. Katie Frieden, Conference Planner,
and her staff at the Penn Stater Hotel and Conference Center, for handling all of the logistics of the meeting, the
field trips, and the short courses. Finally, we thank the staff at A.A. Balkema Publishers who worked with us in
producing these proceedings. Their assistance throughout made it a pleasurable experience.

Raja V. Ramani
Jan M. Mutmansky
Symposium Co-Chairs

XII

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Advisory committee

Sukumar Bandopadhyay University of Alaska Fairbanks

Bharath Belle Anglo American plc
Bill Bookshar Bureau of Mine Safety
Pennsylvania Dept. of Environmental Protection
Jürgen Brune NIOSH Pittsburgh Research Laboratory
Aleksandar Bugarski NIOSH Pittsburgh Research Laboratory
Felipe Calizaya University of Utah
Duncan Chalmers University of New South Wales
Euler De Souza Queen’s University
Ian Duckworth PT Freeport Indonesia
Pamela Franklin U.S. Environmental Protection Agency
Mahe Gangal NRCan/CANMET
Bob Haney MSHA (Retired)
Stephen Hardcastle CANMET MMSL
Dariusz Holod BHP Billiton Diamonds Inc.
John Kelleher Schauenburg Flexadux Corp.
Ian Loomis PT Freeport Indonesia
Bill Maloney Shaft Drillers International
Kirk McDaniel Earth Tech Inc.
Tom McNider Jim Walter Resources Inc.
John Marks Consultant
Pierre Mousset-Jones University of Nevada-Reno
Thomas Mucho Consultant
Kuda Mutama Barrick Goldstrike Mines, Inc.
Richard Ray Parsons Brinckerhoff Inc.
Joanne Reilly Pennsylvania Services Corp.,
an affiliate of Foundation Coal
Andy Schissler Penn State University
Lee Schultz BCS Incorporated
Pramod Thakur Consol Energy
Jerry Tien University of Missouri-Rolla
Ed Thimons NIOSH Pittsburgh Research Laboratory
John Urosek MSHA
Andrzej Wala University of Kentucky

XIII

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Dedicated to the memory of

Howard L. Hartman
[1924–2001]

Howard Hartman is a familiar name to over five decades of min-

ing engineers through his authoritative text on mine ventilation and
air conditioning. In 1961, Howard dedicated the first edition of his
book, entitled Mine Ventilation and Air Conditioning, to George
McElroy, Ray Mancha and John Warren, whom he termed as mod-
ern giants of mine ventilation. We dedicate these proceedings of
the Eleventh U.S./North American Mine Ventilation Symposium to
Howard’s memory in recognition of the fact that he is a worthy addi-
tion to this illustrious list, and we take this opportunity to provide a
brief background of his outstanding career.
Born in Indianapolis, IN, in 1924, Howard began his college edu-
cation at the Colorado School of Mines in 1942 but joined the U.S.
Navy in 1944. Howard received his B.S. and M.S. degrees from
Penn State in 1946 and 1947, and his Ph.D. from the University
of Minnesota in 1953, all in Mining Engineering. Howard was
a registered professional engineer in the states of Colorado and
Pennsylvania. In the 1948–50 period, he worked for Phelps Dodge
Corporation in Bisbee, AZ, and for the Arizona Mine Inspector’s Office, in Phoenix, AZ.
Howard embarked on an academic career of over four decades as an Instructor in Mining Engineering at Penn
State in 1947. After his Ph.D., he returned to Colorado School of Mines as an Assistant Professor and quickly
rose to the ranks of Associate Professor and Acting Head of the Department of Mining Engineering. In 1957, he
came back to Penn State as Professor and Head of the Department of Mining Engineering. Between 1963 and
1980, Howard held the titles of Associate Dean in the College of Engineering at Penn State [1963–67], Dean
of the School of Engineering at the California State University at Sacramento [1967–71], and Dean, School of
Engineering, at Vanderbilt University [1971–80]. In all these assignments, Howard was an innovator, creating
the first work-study program in mining engineering at Penn State and helping to establish the new program in
Socioengineering at Cal Sacramento and the Technology and Public Policy program at Vanderbilt. He joined
the University of Alabama in 1980 as the first holder of the Garry Neil Drummond Endowed Chair in Mining
Engineering, retiring in 1989 with the Emeritus title.
Continuing his academic endeavors well into his retirement, Howard edited the monumental SME Mining
Engineering Handbook over a period of five years. As the senior editor, he co-coordinated the efforts of six
co-editors, 20 section editors, and 242 authors, producing a two-volume, 2300-page, highly-referenced hand-
book in 1992 for mining engineers worldwide. He produced the third edition of the Mine Ventilation and Air
Conditioning text [John Wiley & Sons, 1997], which he originally authored in 1961. He also worked on the
revision of his 1982 Introductory Mining Engineering text [John Wiley & Sons, 2002] until his death in 2001
from the ravaging effects of Parkinson’s disease. These texts reveal Howard’s attention to the details of science,
engineering and technology, and his dedication to solved examples, mind-broadening exercises, and references
for further study and research. Students’ needs, particularly the support for their education, were never far away
from Howard’s mind. He established the Howard Hartman Honor Scholarship Fund in Mining Engineering at
Penn State which benefits students to this day from the royalties on his mine ventilation texts.
Howard did pioneering research in the areas of rock mechanics, ventilation, rapid excavation and drilling
and served as a consultant to a number of mining and drilling companies and government agencies. During
his distinguished career, he advised 28 students to advanced degrees, including eight who received the Ph.D.
He organized the first symposiums in the U.S. in rock mechanics, rapid excavation, surface mining, technology
and public policy and mine ventilation through the Society for Mining, Metallurgy, and Exploration [SME]. He
is an author or editor of over 100 papers, books, and symposium proceedings. Howard was appointed as the
Chair of the Federal Metal and Nonmetallic Mine Safety Board of Review [1971–75] by President Nixon in
1971. He was a member of the U.S. National Academy of Sciences Committee on Tunneling Technology from

XV

Copyright © 2006 Taylor & Francis Group plc, London, UK

 1974 to1977. He has visited many universities around the U.S. and abroad to lecture on mine ventilation, rock
mechanics and tunneling, and was cited in Who’s Who in America [1964].
Numerous awards for his distinguished contributions to the science and technology of mining engineering
were presented to Howard. He received the AIME’s Mineral Industry Education Award in 1965 and CIMM’s
Distinguished Lecturer Award in Rock Mechanics in 1966. He won several prestigious awards from the Society
for Mining Metallurgy and Exploration [SME] including the Distinguished Member Award [1982], the Daniel
Jackling Award [1990], and the Howard Eavenson Award [1993]. In 1989, SME created the Howard L. Hartman
Award to honor Howard and to recognize individuals who have made outstanding contributions to the field of mine
ventilation. He was the first recipient. In 1994, Howard was elected to the National Academy of Engineering, a
distinction accorded to few engineers and scientists to recognize their contributions to the advance of engineering
sciences and technology.
During his student days at Penn State, Howard met a fellow student, Bonnie Lee Sherrill, a home economics
major from Bentleyville, PA. Ever since their marriage in 1948, Bonnie has been the number one supporter of
Howard, and her contributions to his professional endeavours are enormous. Howard and Bonnie had a great many
shared interests including hiking, opera, and vacationing in the national parks, particularly Yosemite National
Park, where many of the rangers and staff knew them on a first-name basis.
It is our pleasure to have known Howard in all of his many roles – professor, mentor, author, leader, adminis-
trator, and most of all, friend.

Jan M. Mutmansky
Raja V. Ramani
Symposium Co-Chairs

XVI

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Keynote address

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

The Methane to Markets Partnership: Opportunities for

coal mine methane utilization

D. Kruger
Climate Change Division, U.S. Environmental Protection Agency, United States

P. Franklin
Coalbed Methane Outreach Program, U.S. Environmental Protection Agency, United States

ABSTRACT: Coal mine methane (CMM) emissions are an important source of greenhouse gas (GHG) emis-
sions globally and represent a significant opportunity for developing economically-viable energy resources. The
Methane to Markets Partnership, an international initiative joining 17 nations, seeks to overcome key barriers to
CMM project development in the Partner countries. This paper outlines some of the Partnership’s key initiatives
in the coal mine methane sector and summarizes CMM project opportunities and developments, especially in
capturing and recovering emissions from coal mine ventilation systems.

1 INTRODUCTION High GWP

Through an international initiative – the Methane to
Markets Partnership – the global community is focus-
ing on reducing methane emissions to achieve both
economic and environmental benefits. Wide-spread
support for and interest in methane exists for several
key reasons. This paper provides an overview of the
available opportunities and technologies for methane
recovery and use in the coal mining sector and of the
Partnership activities in this sector.
1.1 Importance of global methane emissions
Methane is the primary component of natural gas and
is an important clean energy source. It is also a potent
greenhouse gas (GHG), second in global importance
only to carbon dioxide, that accounts for 16% of all
global GHG emissions resulting from human activi-
ties (Figure 1). Methane is considered a potent GHG
because on a kilogram for kilogram basis, it is 23 times
more effective at trapping heat in the atmosphere than
CO2 over a 100-year time period. Methane is also
a short-lived GHG, with an atmospheric lifetime of
approximately 12 years. Because methane is both a
powerful greenhouse gas and short-lived compared
to CO2 , reducing its emissions would have a rapid
and significant effect on reducing the atmospheric
warming potential.
Globally, a large amount of methane that could
be recovered and used for fuel is emitted to the
Methane Gases
16% 1%
Nitrous
Oxide 9%
CO2 - Land CO2 - Fuel &
Use Change Cement 55%
& Forestry
19%
Figure 1. Global anthropogenic greenhouse gas emissions.
atmosphere unnecessarily. Methane is emitted from
a variety of both anthropogenic (human-influenced)
and natural sources. Anthropogenic emission sources,
which constitute about 60% of global methane emis-
sions, include coal mining, natural gas and oil systems,
landfills, and agriculture (Figure 2). As this figure
shows, global coal mine methane emissions account
for nearly 10% of global human-related methane emis-
sions. The remaining 40% of global methane emis-
sions are from natural sources, principally wetlands,
gas hydrates and permafrost, and termites. The leading
global emitters of methane are China, India, the United
States, Brazil, Russia, and other Eurasian countries;
collectively, these nations are responsible for almost
half of all anthropogenic methane emissions.

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Enteric
Manure fermentation
Rice
4% 28%
11%
Natural gas
15% Biomass
Burning
5%
Coal
Biofuel
combustion
8% Oil 4%
1% Solid waste Waste water
13% 10% Fuel stat. &
mobile
1%
Figure 2. Source allocation of global anthropogenic
methane emissions.
Methane emission sources vary significantly among
countries. For example, the two key sources of methane
emissions in China are coal mining and rice pro-
duction, whereas landfills are the largest source of
methane emissions in the United States.
1.2 Utilizing methane
Many of the available emission reduction opportuni-
ties involve the recovery and use of the methane as
fuel for electricity generation, on-site use, or off-site
gas sales. Even with currently available technology and
the benefits of mitigation, methane recovery and use
is not widespread for several reasons. First, methane is
generally a secondary issue in the industrial processes
from which it is emitted. Coal mines, for example,
want to vent methane from the mine workings because
it is explosive and historically, mining companies have
not viewed methane as an energy resource in its own
right. Second, those responsible for the emissions
may not be familiar with the technologies available
for methane recovery or the potential for profitable
projects. This is especially true in developing countries
where increased information exchange and technical
training would be beneficial to generating support for
methane recovery projects. Finally, poorly functioning
energy markets and financially-insolvent utilities and
municipalities within many countries fail to provide
the private sector with a climate that will attract their
investment in projects to capture and utilize methane.
2 METHANE TO MARKETS PARTNERSHIP
Launched in 2004, the Methane to Markets Partnership
is a multilateral initiative uniting public and private
interests to advance the recovery and use of methane
as a clean energy source. Currently, 17 national gov-
ernments (Table 1) and almost 200 Project Network
member organizations are working collaboratively to
advance project development in four major methane
emission source areas: landfills, underground coal
Copyright © 2006 Taylor & Francis Group plc, London, UK
Table 1. Methane to Markets Partner Countries∗ .
Partner Countries:
Argentina Ecuador Russia
Australia India S. Korea
Brazil Italy Ukraine
Canada Japan United Kingdom
China Mexico United States
Colombia Nigeria
∗
As of December 2005.
mines, natural gas and oil systems, and livestock
waste management systems. The Partnership seeks
to overcome key barriers to project development in
the Partner countries. By utilizing otherwise-wasted
methane, this initiative will produce important ben-
efits, such as reduced greenhouse gas emissions,
enhanced economic growth, energy security, coal mine
safety, and improved local air quality. The Partnership
brings the private and public sectors together to achieve
cost-effective methane emission reduction projects
through cooperative technical assistance, technology
deployment and project implementation.
By 2015, the Partnership has the potential to deliver
annual methane emissions reductions of up to 50
million metric tons of carbon equivalent (MMTCE),
totaling 14.2 billion cubic meters of natural gas (US
EPA 2004c). This would be equivalent to removing 33
million cars from the roadways for one year, plant-
ing 55 million acres of trees, or eliminating emissions
from fifty 500 MW coal-fired power plants. These
ambitious but attainable and measurable results could
lead to stabilized or even declining levels of global
atmospheric concentrations of methane.
The Methane to Markets Partnership focuses on
addressing the key barriers to advancing methane
recovery and use. Through the use of private/public
partnerships, this initiative brings together the techni-
cal and market expertise, financing, and technology
necessary for project development. Member countries
work in collaboration with the private sector, multilat-
eral development banks, and other governmental and
non-governmental organizations.
As part of its commitment to the Partnership, each
Partner nation agrees to undertake a variety of activ-
ities aimed at advancing methane recovery and use
internationally in the target sectors. Each Partner coun-
try manages their own financial contribution and assis-
tance mechanisms based on their national interests and
their areas of expertise.
A Steering Committee guides the work of the
Partners. Sector-specific Subcommittees (landfills, oil
and gas systems, coal mining, and agriculture) are
responsible for developing action plans that iden-
tify and address key barriers and issues for project
development, address market assessment and reform
 issues, facilitate investment and financing opportuni-
ties, and identify cooperative activities and projects
that advance the recovery and use of methane as an
energy source.
Organizations outside of the Partner govern-
ments play a central role in Subcommittee activities.
As a public-private partnership, Methane to Markets
encourages active involvement by private sector enti-
ties, financial institutions, and other governmental and
non-governmental organizations as a means of build-
ing capacity, transferring technology, and promoting
private investment. To this end, the Partnership has
created a Project Network to serve as an informal
mechanism to reach out to and facilitate communi-
cation and coordination among these organizations.
Interested organizations can become a member of the
Methane to Markets Project Network by signing a one-
page, non-binding agreement that is available on the
Partnership’s web site. To date, over 150 organizations
have joined these efforts.
2.1 The US government commitment
The US government intends to commit up to $53 mil-
lion over the next five years to facilitate the develop-
ment and implementation of methane projects in both
developing countries and countries with economies
in transition. This will be accomplished through the
implementation of a range of activities, such as train-
ing and capacity building, market development, fea-
sibility assessments and technology demonstrations.
Central to the US commitment will be leveraging the
efforts of fellow Partner countries as well as the exper-
tise and investment of the private sector and other
members of the Project Network.
The US Environmental Protection Agency (EPA)
leads the efforts of the US government. EPA is build-
ing on the success of the Agency’s voluntary domestic
methane partnership programs. These programs work
directly with the private sector to provide technical
assistance, facilitate technology transfer, and address
barriers to methane recovery and use project develop-
ment.As of 2004, these Programs have been successful
in reducing methane emissions in the US to 10% below
1990 levels.
Other Departments are also playing a central role in
implementing the US commitment, including: Depart-
ment of State, which takes the lead on the US gov-
ernment’s international climate change policy and
activities; Department of Energy, which has valuable
expertise in natural gas and coal mine methane (CMM)
technologies; and the US Agency for International
Development, which provides important technical
expertise in the economic reform of energy sectors
to create markets that support private sector projects
in developing countries and those with economies in
transition.
Copyright © 2006 Taylor & Francis Group plc, London, UK
2.2 Methane to Markets Activities in
the Coal Mine Methane Sector
The Coal Mine Methane Subcommittee promotes
project-based methane emissions reductions, giving
the highest priority to activities that have the great-
est chance to achieve emissions reductions in the near
term. The Subcommittee also recognizes the impor-
tance of laying appropriate groundwork and capacity
building.
In the first year of the Partnership, the Coal Mine
Methane Subcommittee identified four key types of
barriers that inhibit CMM project development:
1. Lack of appropriate information about project
opportunities;
2. Lack of appropriate technology and technical
knowledge;
3. Lack of demonstration of the technical or economic
feasibility of these projects in a specific setting; and
4. Lack of financing or understanding of how to obtain
financing.
The Subcommittee further identified potential Part-
nership activities that address these barriers:
1. Information-based activities such as reports, publi-
cations, and websites;
2. Technology-transfer activities, such as workshops
and training;
3. Technical and economic feasibility studies and
technology demonstrations; and
4. Activities that build capacity for project financing
and investment.
3 COAL MINE METHANE RECOVERY
OPPORTUNITIES
3.1 Global issues
In order to reduce explosion hazards, methane is
removed from underground mines either in advance
of mining, during mining activities, or after mining
has occurred. Instead of releasing this methane to
the atmosphere, profitable uses can be identified and
implemented. Some of these options include natural
gas pipeline injection, power production, co-firing in
boilers, district heating, coal drying, and vehicle fuel.
There are numerous benefits to recovering and uti-
lizing coal mine methane: increasing mine safety,
reducing greenhouse gas emissions, improving local
air quality, providing local energy resources, and
increasing mine productivity.
CMM project development potential is substan-
tial in many Methane to Markets Partner countries.
Development of this resource depends on access to
appropriate technology and technical knowledge. It
also depends on several market factors that may dif-
fer considerably for each country: (1) the price of gas;
 (2) the price of electricity; (3) the condition of, or lack
of, infrastructure; (4) the potential for greenhouse gas
emissions reduction credit sales; and (5) the country’s
regulations, energy subsidies and tax structure.
3.2 CMM recovery and use in the United States
The United States is one of the founding members of
the Methane to Markets Partnership and a world leader
in coal mine methane development. The USA is the
world’s second-largest producer of hard coal, behind
China. Total US production in 2004 was 1,111.5 mil-
lion short tons, slightly below the 2001 record level
of 1127.7 million short tons (Freme 2005). Partly in
response to environmental concerns with the high sul-
fur content of many Eastern coals, in the past few
decades the majority of US coal production has shifted
to the Western Basins, where low-sulfur coal is abun-
dant. In 2003, underground coal mines accounted for
33% of total US coal production, and mines in the
Post-mining
(surface) Abandoned mines
2%
Post-mining
(underground)
11%
Surface
mining
14%
Underground –
drainage (not
recovered and
used) 11%
Figure 3. US Coal Mine Methane emissions sources.
Coal Mine Methane (million cubic meters)
Eastern US provided 44% of the national total (US
EIA 2005).
The US ranks second globally in CMM emissions,
behind China. In 2003, total US CMM emissions were
134 billion cubic feet or Bcf (3.8 billion cubic meters)
for all active operating underground and surface min-
ing activities. Figure 3 illustrates the relative contribu-
tion to net CMM emissions from different sources of
mining-related activities. As the figure shows, under-
ground mines contribute the greatest amount (62%)
to total US CMM emissions from drained gas that is
not recovered and from ventilation system emissions
(excluding post-mining emissions). In 2003, net CMM
emissions from underground US mines were 93 Bcf
(2.6 billion cubic meters).
Since the early 1990s, methane emissions from US
coal mining activities have decreased by 34%, even
though total coal production has risen by about 10%
during this period. These emission reductions have
been achieved primarily from the capture and utiliza-
tion of coal mine methane. Many US coal companies
have realized the important benefits of degasifying
their mines, namely improved mine safety and
increased productivity. Several companies have also
11%
profited financially from the recovery and use of
coal mine methane, especially for sale to natural gas
pipelines.
Figure 4 presents the trends in annual US under-
ground CMM emissions since 1990. U.S. coal mines
liberated 4.9 billion cubic meters in 2003, with under-
ground mining activities liberating the largest amount
Underground –
ventilation air
with 3.8 billion cubic meters (or 77% of the total). Of
methane the 3.8 billion cubic meters of CMM liberated, about
51% 1.6 billion cubic meters were emitted through drainage
systems, while the remainder was emitted from venti-
lation systems. As Figure 4 indicates, emissions from

3,500

3,000

2,500
Ventilation air
emissions (vented
2,000 to atmosphere)

1,500
CMM Drainage
1,000 (vented to
atmosphere) CMM Recovered
and Used
500 (Avoided
emissions)
-
90

91

92

93

94

95

96

97

98

99

00

01

02

03
19

19

19

19

19

19

19

19

19

19

20

20

20

20

Figure 4. US Coal Mine Methane emissions since 1990.

Copyright © 2006 Taylor & Francis Group plc, London, UK

 ventilation systems constitute the majority of CMM
emitted from underground mines. Methane concentra-
tions in the ventilation air are very dilute (typically less
than 1% methane), which presents formidable chal-
lenges to economic recovery and use. At the present
time in the US, none of the methane in the ventilation
air is captured and utilized.
Currently in the US, only drained methane from
underground coal mines is recovered and utilized.
Since 1990, increasing proportions of CMM liberated
from drainage systems (also referred to as degasifi-
cation systems) have been recovered for use. Corre-
spondingly, the CMM emissions to the atmosphere
from mine degasification (drainage) systems have
decreased.
Even though average drainage efficiency at US
mines has increased overall since 1990, the number
of active US coal mines with mine degasification sys-
tems has declined. This decrease is due primarily to the
closure of mines that had operating degasification sys-
tems. By 2003, only 18 US coal mines had operating
drainage systems.
The US leads the world in coal mine methane cap-
ture and utilization at active mines. In 2003, over
70% of all drained coal mine gas was used, primar-
ily for injection into the natural gas pipeline system.
As indicated by the profitability of US coal companies
engaged in these projects, there are numerous environ-
mental, economic, and ancillary benefits associated
with CMM recovery and utilization. There are cur-
rently ten (10) utilization projects at active US mines,
accounting for about 40 Bcf or 1.1 billion cubic meters
(16 million metric tonnes CO2 equivalent) of avoided
methane emissions in 2003 (US EPA 2004a). The total
CMM captured and used is only a small fraction of
the total US coalbed methane (CBM) production. For
instance, in 2003, the total utilized CMM represented
only about 2.5% of the US total CBM production that
year. The largest share of this utilized gas, 0.8 billion
cubic meters, was recovered from two Consol Energy
mines.
The majority of recovery and utilization projects
are located in the Appalachian Basins of the East-
ern United States (in Alabama, Virginia, and West
Virginia). This is due in large part to the greater
occurrence of degasification systems in the East com-
pared to the West. In many Western mines, ventila-
tion systems are sufficient to maintain safe in-mine
methane concentrations and therefore drainage sys-
tems remain relatively rare. A few projects are under-
way in the Western US (Colorado and New Mexico).
Geography, especially surface terrain, plays a large
role in the extent of CMM recovery and use in the
US. In many parts of the US, the mineable coal seams
have high gas contents, relatively high permeability,
and favorable surface conditions (i.e., reasonably flat
terrain).
Copyright © 2006 Taylor & Francis Group plc, London, UK
In the United States, the majority of CMM that is
utilized is sold for injection into natural gas pipeline
systems. There are two main reasons for this. First, the
strong US natural gas market, which in recent years
has offered relatively high prices for natural gas com-
pared to the low cost of electricity, has encouraged
recovery of coal mine methane for sale to natural gas
pipelines. Secondly, the US has an extensive and well-
capitalized natural gas pipeline network, especially
in the Eastern third of the country. At sites where
pipelines are located close to coal mines and CMM
recovery operations, gas gathering and pipeline injec-
tion are very economically attractive. In some cases,
coal mine methane project operators have actually
built pipelines to transport gas to the pipeline.
US natural gas pipeline standards demand high
quality gas (usually 95% or greater methane with
minimal contaminants). To meet these requirements,
generally only gas from wells drilled into virgin seams
in advance of mining (“pre-mining drainage”) is suit-
able. Some mines are able to inject drained CMM
directly into the pipeline without any pre-treatment.
Other mines must upgrade their gas (especially if it is
lower quality, e.g., gob gas) by removing contaminants
or enriching the gas to increase its heating value to
meet pipeline specifications. For example, Jim Wal-
ter Resources in Alabama is employing a nitrogen
rejection and cryogenic processing facility to upgrade
72 million cubic meters per year of lower quality gas
into 41 million cubic meters of pipeline-quality gas.
At this time, there are only a few CMM-based
power generation projects in the US. The scarcity of
CMM-based power projects is due primarily to the
relatively low cost of electricity in the US. Low elec-
tricity prices offer insufficient profit incentives for
investment in CMM projects. Furthermore, there are
currently no regulatory or price-based incentives to
use CMM-generated power. The two power generation
projects currently operating at active US underground
coal mines total 90 MW, the majority of which is used
infrequently for peaking loads only.
There are several key areas for future expansion
in recovery and utilization of coal mine methane in
the United States. First, the use of drained gas at
active underground mines can be increased to tap the
nearly 16 Bcf (453 million cubic meters) of drained
gas that is currently not being utilized. This can be
both by increasing the amount of CMM drained, and
by increasing the rate of recovery and utilization. The
increased use of methane from abandoned mines will
also reduce CMM, albeit on a smaller scale. Gassy
surface mines, particularly in the Western US, may
present important opportunities to recover and utilize
high-quality gas. The most critical element of the long-
term emissions reductions in the US coal mining sector
is utilization of the untapped resource intrinsic in venti-
lation air methane, which constitutes the single largest
 source of methane emissions from US underground
coal mines.
3.3 The role of mine ventilation emissions
Although practically all CMM projects worldwide
have utilized drained gas, the largest source of CMM
emissions is from ventilation shafts. Although ventila-
tion air methane (VAM) contains very low concentra-
tions of methane (typically below 1 percent), the sheer
volume of ventilation air makes this the single largest
source of coal mine methane. Global VAM emissions
in 2000 were 230 million tonnes of CO2 equivalent
(16 billion cubic meters) and are projected to increase
to over 300 million tonnes (21 billion cubic meters)
by 2020.
Recent technological advances have created the
potential for a substantial portion of these emis-
sions to be reduced cost-effectively (USEPA 2000).
A market for VAM oxidation is emerging in many
of the world’s coal producing countries. Collectively,
Methane to Markets Partnership countries are esti-
mated to account for approximately 85 percent of
global VAM emissions. Currently, Australia is the only
country with operating VAM recovery and utiliza-
tion projects, although US EPA is investigating VAM
utilization in the US and China.
4 CONCLUSIONS
Methane is the primary component of natural gas and
an important clean energy source. It is also a potent
greenhouse gas, second only to carbon dioxide. The
Methane to Markets Partnership offers a unique oppor-
tunity for governments and organizations around the
world to work together to address methane emissions
while achieving economic, environmental, and energy
benefits. The US Government believes that significant
progress can be made and is committed to working
with its domestic and international public and pri-
vate sector partners. The US estimates that by 2015
Methane to Markets has the potential to deliver annual
methane emissions reductions of up to 50 million met-
ric tons of carbon equivalent (MMTCE) or recovery of
500 billion cubic feet (Bcf) of natural gas. If achieved,
these reductions could lead to stabilized or even declin-
ing levels of global atmospheric concentrations of
methane.
Coal mine methane plays a critical role in global
methane emissions and in the success of the Methane
to Markets Partnership. A vital component to achiev-
ing substantial emissions reductions from the coal
Copyright © 2006 Taylor & Francis Group plc, London, UK
mine methane sector can be achieved through address-
ing opportunities to mitigate coal mine methane from
drainage and from ventilation systems.
REFERENCES
Freme, F. 2005. US Coal Supply and Demand: 2004 Review.
US Energy Information Administration. http://www.eia.
doe.gov/cneaf/coal/page/special/feature04.pdf.
Schultz, K.H, and C.C. Talkington. 2003. An Overview of the
Global Coal Mine Methane Industry. November, 2003.
3rd International Methane & Nitrous Oxide Mitigation
Conference. Beijing, China.
US Energy Information Administration (US EIA). 2005.
Annual Energy Outlook 2005. Market Trends – Natu-
ral Gas Demand and Supply. http://www.eia.doe.gov/oiaf/
aeo/gas.html (accessed July 8, 2005).
US Environmental Protection Agency (US EPA) Office of
Air and Radiation. 2000. Technical and Economic Assess-
ment: Mitigation of Methane Emissions from Coal Mine
Ventilation Air.
US Environmental Protection Agency (US EPA) Office of Air
and Radiation. 2001. Coalbed Methane Extra. February
2001.
US Environmental Protection Agency (US EPA) Office of
Air and Radiation. 2002a. Emissions and Projections of
Non-CO2 Greenhouse Gas Emissions from Developing
Countries: 1990–2020 (Draft Report). June 2002.
US Environmental Protection Agency (US EPA) Office of Air
and Radiation. 2002b. Non-CO2 Greenhouse Gas Emis-
sions from Developed Countries: 1990–2010. February
2002. EPA 430-R-01-007.
US Environmental Protection Agency (US EPA) Office of
Air and Radiation. 2003. Assessment of the Worldwide
Market Potential for Oxidizing Coal Mine Ventilation Air
Methane. EPA-430-R-03-002. July 2003.
US Environmental Protection Agency. 2003. International
Analysis of Methane and Nitrous Oxide Abatement
Opportunities: Report to Energy Modeling Forum,
Working Group 21. June. http://www.epa.gov/methane/
appendices.html
US Environmental Protection Agency (US EPA) Office of
Air and Radiation. 2004a. Greenhouse Gas Emissions
Inventory 1990–2003.
US Environmental Protection Agency (US EPA) Office of
Air and Radiation. 2004b. Methane Fact Sheet July, 2004.
US Environmental Protection Agency (US EPA) Office of Air
and Radiation. 2004c. Methane to Markets Partnership
Fact Sheet. July, 2004.
US Environmental Protection Agency. 2004d. Methane
Fact Sheet: The Significance of Methane and US Activ-
ities to Reduce Methane Emissions. June. www.epa.gov/
methanetomarkets.
US Environmental Protection Agency (US EPA) Office
of Air and Radiation. 2004e. Methane Emissions from
Abandoned Coal Mines in the United States: Emis-
sions Inventory Methodology and 1990–2002 Emissions
Estimates. EPA 430-R-04-001. May 2004.
 1. Ventilation planning

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Ventilation on demand: A preliminary study for Ballarat Goldfields NL

M.A. Tuck
School of Science and Engineering, University of Ballarat, Ballarat, Victoria, Australia

C. Finch & J. Holden

Ballarat Goldfields NL, Ballarat, Victoria, Australia

ABSTRACT: Ballarat Goldfields NL are currently developing a new goldmine in the historical East Ballarat
goldfield under historic old workings. The ventilation of the mine poses a number of challenges from both a
mining and ventilation perspective. From a ventilation perspective on challenge is being able to ventilate the
proposed stopes given the long intake and return air routes imposed due to the mine being under the regional
urban centre of Ballarat requiring that the surface connections are relatively divorced from the working stopes.
The current mine plan is for a number of sub-mines within the Ballarat east goldmine. As the proposed method
of working is by overhand cut and fill mining the ventilation of the stopes can be divided into a number of phases
each of which requires different levels of ventilation. This study was undertaken to determine if the principles
of ventilation on demand could be applied to the Ballarat east mine and if so what the requirements to control
such a system are.

1 INTRODUCTION operations of Ballarat. Ballarat is located 120 km

The Ballarat East mine is currently under develop-
ment by Ballarat Goldfields NL and plans to obtain
first production from the mine in late 2005/early 2006.
The mine is located in the historic Ballarat East gold-
field and is to mine ore underneath the historic mining
North West of Melbourne as shown in Figure 1. The
ventilation of the mine poses a number of challenges
from both a mining and ventilation perspective. From
a ventilation perspective on challenge is being able to
ventilate the proposed stopes given the long intake and
return air routes imposed due to the mine being under
the regional urban centre of Ballarat requiring that the
surface connections are relatively divorced from the
working stopes.
The current mine plan is for a number of sub-mines
within the Ballarat east goldmine. As the proposed
method of working is by overhand cut and fill mining
the ventilation of the stopes can be divided into a num-
ber of phases each of which requires different levels
of ventilation. This study was undertaken to determine
if the principles of ventilation on demand could be
applied to the Ballarat east mine and if so what the
requirements to control such a system are.

2 CURRENT PROPOSED MINE LAYOUT AND

VENTILATION SYSTEM

An appreciation of the proposed mine layout can be

gained by referring to Figure 2.
There are essentially four semi-independent “sub-
mines”, each accessed by its own internal decline.
Fresh air enters the workings via the surface decline
and Golden Point Intake Shaft, with some minor flows
Figure 1. Location of Ballarat. supplied by the North Woah Hawp Shaft. This intake

11

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 2. Mine layout.
air will be distributed to the four internal declines via
the main decline.
Each internal decline has its own exhaust raise sys-
tem (shown in red in Figure 2). These exhaust raises
are linked together by the Sulieman decline forming
a common lateral “manifold”. Exhaust air flows from
the Sulieman decline to the surface via the 6 m diam-
eter North Prince Extended (NPE) surface ventilation
shaft.
Each of the four “sub-mines” will produce at a nom-
inal rate of 200,000 tonnes per annum, giving the mine
a total production rate of 800,000 tonnes per annum.
The mine plans to produce 100,000 oz of gold per
year initially, ramping up to 200,000 oz per year at full
production, with a planned mine life of 22 years.
The proposed method of working is by overhand cut
and fill mining. At full production this will require a
total of 20 stopes to be available at any one time, 10
in production and 10 in the filling phase. The main
ventilation pollutants identified are:
1. Diesel exhaust emissions
2. Blasting fumes
3. The waste rock used as fill is sulphidic, combined
with water this could produce H2 SO4 . This may be
exacerbated by inflow of water from the historic
workings
4. The water from old workings could also provide
H2 S and SO2 as possible ventilation pollutants
5. In addition the exothermic reaction as the fill cures
in the stopes may provide a heat problem. This is
likely to be more of a concern later in the mine
life as the mine extends to deeper levels when the
combined effect of virgin rock temperature increase
and enhanced diesel pollution in the form of heat in
the intake system combine to form a potential heat
problem for the mine.
The current ventilation design recognizes diesel
pollution as the dominant ventilation pollutant and as
such this is reflected in the current design.
The proposed equipment fleet for the mine is
detailed in Table 1.
12
Copyright © 2006 Taylor & Francis Group plc, London, UK
The proposed equipment utilization in each of the
for sub mines is provided in Table 2.
A number of scenarios have been investigated to
ventilate the mine. The scenario adopted is to ven-
tilate each of the sub mines with a dedicated fan.
Simulations using VENTSIM indicate each of the sub
mines requires two fans operating in parallel providing
50 m3 /s operating at a pressure of 2.5 kPa. This is the
design adopted that employs fixed speed drive motors.
The current design is based on full utilization of
the diesel fleet underground, whilst the current design
does take into account the variability of equipment uti-
lization underground to an extent does not account for
the full variability in demand that will occur practi-
cally. The current design is based on the approach that
all equipment will be in use at any one time. In real-
ity the airflow requirements in each of the sub mines
will associate with the production stage of each of the
stopes located in the sub mines, this has prompted the
study of ventilation in demand for the mine to reduce
the overall power requirements and as a consequence
the operating costs for the mine.
3 VENTILATION ON DEMAND
Ventilation on demand is a simple idea and requires
only the minimum ventilation flow to be circulated
to a particular area of a mine at a particular point in
time to satisfy the ventilation requirement of providing
adequate oxygen and to dilute and remove ventilation
pollutants whilst providing miner comfort. Tradition-
ally the approach has been to determine the maximum
airflow requirement and to circulate this flow through
the whole mine or mine section at all times. From
the perspective of energy efficiency the traditional
approach has a number of failings.
4 THE APPLICATION OF VENTILATION ON
DEMAND TO BALLARAT GOLDFIELDS
The current ventilation planning for Ballarat East is
based on complete 100% utilization of the diesel fleet
 Table 1. Global diesel equipment loadings and aggregate airflow rates for mining stages assuming 0.5 m3 /s per
100 kW diesel rated capacity.

Engine Stage 1 Stage 2 Stage 3 Stage 4 Stage 5

power
Equipment (kW) No m3 /s No m3 /s No m3 /s No m3 /s No m3 /s

2 boom jumbos * 1 2 2 3 6
1 boom jumbos * 1 2 3 6
R1700 loaders 231 1 12 2 23 2 23 4 46 7 81
R1300 loaders 123 1 6 1 6 3 18 4 25
chargeup units 100 1 5 1 5 1 5 2 10 3 15
shotcreters 80 1 4 1 4 1 4 2 8 3 12
Agitator Trucks 120 2 12 2 12 2 12 3 18 5 30
cable bolters * 1 2 3
services platform 100 1 5 1 5 1 5 1 5 1 5
40 tonne trucks 350 1 18 2 35 2 35 4 70 8 140
Grader 120 1 6 1 6 1 6 1 6 1 6
Bobcats 17 1 1 1 1 2 2
personnel carriers 80 1 4 1 4 2 8 4 16
supply transporter 100 1 5 1 5 1 5 1 5
fuel truck 100 1 5 1 5
lube vehicle 100 1 5 1 5
light vehicles * 11 14 16 25 32
Contingency (30%) 18 32 32 62 104
TOTAL 79 137 138 267 450

Table 2. Diesel equipment for each of the 200,000 tpa sub stopes in production with three stopes being filled.
mines. Diesel loading is at its maximum in those stopes under
production, this implies that two of the sub mines are
Minimum
airflow (m3 /s)
over ventilated from the minimum requirement from
Unit Engine (kW) No. @0.05 m3 /s/kW* a diesel exhaust perspective. If an assumption is made
that the ventilation requirement during filling is for a
40t Truck 350 2 35.0 minimum air velocity of 0.25 m/s to remove possible
Large LHD 231 1 11.6 heat and gaseous emissions then three of the sub mines
Small LHD 123 2 12.3 require a flow of 100 m3 /s whilst two will only require
Grader 120 1 6.0 100 m3 /s.
Service (e.g. I.T.) 100 2 10.0 Analysis of the mining system also reveals other
Subtotal 74.8 potential savings that can be made in terms of airflow
30% Contingency Margin 22.8 circulated. At any one point in time in the mining cycle
TOTAL (rounded) 100 only a proportion of the active production stopes will
have heavy diesel loading associated with mucking
operations. The remaining stopes will be drilled and
underground. Analysis of the mining system indicated charged with minimal diesel loadings as the drill rigs
that at any particular time the following scenario is are electro-hydraulic.
likely to occur. Ignoring this later factor it is possible for the two
sub mines with only two from a possible of five stopes
1. 10 stopes will be producing ore for processing with with active diesel loadings to require a total of 80 m3 /s,
associated waste rock a saving of 20 m3 /s from the current design. Assum-
2. 10 stopes will be in the filling phase. ing ventilation pressure to remain constant at 2.5 kPa
this provides an air power saving of 50,000 kW for
The current mine plan assumes this and is reflected each section or a cost saving of $2,000 per hour on the
in the current ventilation design to an extent. At a sim- basis of an electricity cost of $0.04 per kW/hour. Given
plistic level this implies that for each of the four sub ventilation costs can provide a major element of total
mines two will have three stopes in production with underground operating costs this could be significant
two stopes in the fill cycle, with two only having two in ensuring maximum profit.

13

Copyright © 2006 Taylor & Francis Group plc, London, UK

 5 WHAT NEEDS TO BE DONE TO ACHIEVE
VENTILATION ON DEMAND?
Ventilation on demand will require close control over
the entire ventilation system. Specifically for each of
the sub mines this will require the ventilation flow to
be regulated closely to match the ventilation flow with
the required air volume. Two possible methods exist to
achieve this:
1. Passive regulation using regulators, or increased
resistance to control the flow. This is unlikely to
achieve the savings indicated as energy is being
destroyed.
2. Active regulation by changing the operating point
of the fans supplying each sub mine.
Neither provides a complete solution as mine ven-
tilation systems are highly dynamic, thus a hybrid
system using both is optimal, with the dominant ele-
ment being active regulation by the fans. This can
be achieved by employing either of the following
separately or in combination:
1. variable speed drives for the fans
2. guide vane control at inlet and/or outlet to the fan
to regulate the flow through the fans
It is unlikely that this will achieve the desired level
of control over the airflows in each sub mine, implying
that a certain element of passive regulation will also
be required.
An essential element will therefore be to monitor
both the airflow and pollutant levels within each sub
mine as well as the mine as a whole to ensure both
statutory and mine air quality standards are adhered
to. This will require a comprehensive mine monitor-
ing system to be developed with reporting to a central
surface facility to allow adjustment to the ventila-
tion system preferably in an automated manner to be
delivered. Under currently available systems this will
require a highly sophisticated monitoring a control
system to be developed at the mine site.
14
Copyright © 2006 Taylor & Francis Group plc, London, UK
6 WHAT NEEDS TO BE DONE TO ACHIEVE
VENTILATION ON DEMAND
Changes to the current ventilation and other mining
activities include:
1. Variable speed motors to be fitted to all sub mine
ventilation systems
2. If 1 above not followed fit inlet and or outlet guide
vanes to the fans
3. Development of automated air regulation facilities
to each stoping area
4. Development of a mine wide monitoring and con-
trol system. This represents and additional cost,
however this cost can be offset against the control
requirements for mine drainage and other elements
within the mine that require similar levels of control
5. Establish a separate ventilation department within
the mine
6. Further studies regarding the transient nature of
mine airflow and the impact on production.
7 CONCLUSION
The Ballarat East Mine presents an opportunity to
allow full development of the ventilation on demand
philosophy. The nature of the mine incorporating four
sub mines with highly variable ventilation demands
at each stage of the mining process could potentially
provide high levels of ventilation and hence produc-
tion cost savings to the mine despite the requirement
of a costly monitoring and control system.
Ventilation on demand is a departure from classi-
cal ventilation thinking. Its applicability depends to
a large extent on variability associated with variable
ventilation requirements within the mining cycle and
as such is not applicable to all mining systems. It is
a reflection of the fact that current ventilation design
thinking is coal industry based where ventilation flows
are reasonably constant over time, in the underground
metalliferous industry this is not always the case.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Justifying ventilation-on-demand in a Canadian mine and the need for

process based simulations

S. Hardcastle & C. Kocsis

CANMET-MMSL, Natural Resources Canada, Sudbury, Canada

D. O’Connor
INCO Limited, Ontario Division, Sudbury, Canada

ABSTRACT: Justifying a major change in how large and complex mechanized metal mines are ventilated is
difficult and can be extremely time consuming. Mine ventilation design has improved greatly with the use of
network simulation programs, however their use in the design process is normally based upon achieving a desired
flow, such with diesel exhaust dilution requirements, at specific times in a mine’s lifespan. A limitation of this
approach is that the results are often based upon peak production demands and assume continuous production;
consequently, there can be significant and costly redundancy in the solution. This is because, few areas in a
mine continuously need to be ventilated as each area employs a discontinuous process. Ventilation-on-demand
(VOD) is a way to address the redundant use of air, the concept is simple, but the mine-wide benefits can be
difficult to assess using conventional techniques. This paper through an industrial example at INCO’s Creighton
mine, follows the traditional approach to implementing change, but also provides the justification and data
for higher level analyses using process based simulation to determine the ventilation requirements specific
to any point in time in the transient mining process. The process based simulation methodology, as being
promoted by CANMET-MMSL, will also enable the mechanized base and precious metal mining industry to
obtain individualized solutions to mitigate increasing ventilation energy usage, costs and green house gas related
emissions. Site specific solutions are required when one considers the unique set of controlling influences that
impact ventilation design and its operating costs at each mine against the potential ventilation reduction options
such as VOD, changing the diesel powered equipment, or mining remotely.

1 BACKGROUND options as teleremote mining from surface and the

Providing ventilation is a significant and increasing
cost in mechanized base and precious metal mines.
In Canada, the need for some mines to re-evaluate
their ventilation systems and methodology is becom-
ing all the more critical in light of increasing mining
depth, escalating energy prices, the environmental
implications associated with energy consumption, and
continuing pressure to remain competitive. This chal-
lenge to the industry was shown in a study funded by
the Deep Mining Research Consortium (DMRC) in
Canada (Campbell, 2005), indicating that continuing
with current practices, ventilation and ore transporta-
tion would experience the greatest escalation in cost
with increased depth.
With respect to minimizing ventilation costs, the
general options are either changing the design drivers
such as the need to dilute diesel exhaust, or improv-
ing the utilization efficiency of the air supplied within
the system. To move away from diesel criteria, such
introduction of fuel cells to primary equipment, have
both been considered. To improve a mine’s ventilation
efficiency controlled recirculation and “ventilation-
on-demand” have been tried. However any change in
the way mines are ventilated needs to be evaluated
technically and proven financially beneficial.
This paper describes the progress of CANMET-
MMSL and INCO, both members of the DMRC, in
introducing ventilation-on-demand (VOD) from fea-
sibility through to the pilot stage at Creighton Mine
and its contribution to a new process simulation based
methodology to evaluate the economic and environ-
mental benefits of a change of practice. For INCO,
VOD not only offers a way to minimize its day-to-
day operating cost, it has the potential to optimize
the mine’s natural cooling, hence postponing the need
to introduce costly mechanical refrigeration. How-
ever, all the while, maintaining both the maximum
capacity within the system and the same or better air
quality for the workforce. For CANMET-MMSL, part

15

Copyright © 2006 Taylor & Francis Group plc, London, UK

 of Natural Resources Canada, in addition to main- Table 1. Temperatures and flow requirements as a function
taining the sustainability of its industry, VOD and of depth for Creighton mine.
process based ventilation simulation can contribute to
the Canadian Government’s climate change initiative Depth below Virgin Intake air Required
surface strata temperature, airflow
to control greenhouse gases. Consequently, this work
datum temperature tw
was in part funded through the Government of Canada
Action Plan 2000 on Climate Change. (m3 /s (cfm
(m) (ft) (◦ C) (◦ F) (◦ C) (◦ F) /tpd) /tpd)

1.1 Creighton mine 2134 7000 16.35 61.43 0.099 209

2195 7200 43.3* 110.0* 16.58 61.85 0.106 225
Creighton is INCO’s oldest mine and one of the deep-
2256 7400 44.4 112.0 16.67 62.00 0.114 241
est operations in Canada. Despite being in operation 2335 7660 45.9 114.6 17.11 62.80 0.128 272
since 1901, it still has considerable reserves at depth 2371 7780 46.6 115.8 17.41 63.34 0.135 286
that could last well into the future. Over its 100 years 2408 7900 47.2 117.0 17.72 63.90 0.142 300
of production, its mining methods have changed with 2444 8020 47.9 118.2 18.01 64.42 0.148 314
the introduction of mechanization, increased heat and 2481 8140 48.6 119.4 18.33 65.00 0.155 328
ground control issues with additional depth. Today 2499 8200 48.9 120.0 18.56 65.40 0.170 361
production relies upon mechanized vertical retreat
mining (VRM) with delayed backfill and is concen- * Datum condition
trated below 1500 m.
Since 2001, INCO has been mining Creighton Deep at Creighton Mine. Comparatively, the intake wet-bulb
orebody, starting at the 7400 Level (2250 m below sur- air temperatures are low; these are the result of the
face); the original timeline, as provided for ventilation mine using a natural heat exchanger, a mass of bro-
simulation, detailed production continuing through ken rock, to cool (or heat) the intake air according
2019 by which time, the mine could have reached the to the season (Stachulak, 1989). The net result at the
8180 Level (2495 m below surface). 800 Level (250 m) is an air temperature of 2.8◦ C +/−
To facilitate current and future mining at ever 1.7◦ C year round. However, this effectively free source
increasing depths, Creighton Mine’s ventilation sys- of cooling has a limited capacity.
tem has already undergone two major upgrades. The
first, starting in the late 1970’s (Stachulak, 1979) resul- 1.3 Cost and environmental concerns
ted in a new 6.4 mØ concrete lined primary exhaust
For the industry as a whole, the growing need for
raise with new surface fans that expanded its volu-
energy in the form of electricity or refined petroleum
metric capacity from 280 m3 /s (593 kcfm) to 613 m3 /s
products for heating, and the increasing unit cost
(1.3 Mcfm). The second included upgrading surface
of that energy are a continual concern. Mitigating
exhaust fans (O’Connor et al, 2002); additional intake
those costs is not straightforward and each mine
boosters and improvements to the intake raise deliv-
could require an individualized solution because of
ery system, further increased the mine’s capacity to
its unique combination of controlling parameters,
755 m3 /s (1.6 Mcfm).
such as:
• the electricity rates being fixed or demand based,
1.2 Ventilation requirements • the form and associated cost of heating fuels,
• the energy breakdown between the primary venti-
For Creighton, as with other deepening mines, heat
lation system, the auxiliary ventilation system and
mitigation is increasingly becoming more dominant
the air heating or cooling systems,
than diesel exhaust dilution in the overall design
• the degree of duplication of the discontinuous drill-
requirements of the primary ventilation system. Within
blast-muck cycle required to maintain a continuous
the production areas, the air volumes supplied by the
process plant feed,
local auxiliary or secondary ventilation systems are
• the true continuity of activities within and between
still based upon diesel criteria.
shifts in a 24-hour a day, 7-day a week operation,
Table 1 shows that between 2134 m and 2481 m, an
• the actual timing of activities,
increase in vertical distance of 347 m or 16%, the pre-
• the different local regulations governing the dilution
dicted air volume needed for heat mitigation increases
requirement of diesel exhaust with ventilation,
disproportionably by 57%. These airflow requirements
• the different ventilation requirements to mitigate
listed in terms of production rate are based upon cli-
heat with increasing depth, and
matic predictions limiting the wet-bulb temperature to
• mining method.
a maximum of 23.9C (75F).
Table 1 also lists the virgin strata temperature and For example, the power generating industry in Ontario
the intake wet-bulb air temperature for each elevation was deregulated in May 2002. Since then, industrial

16

Copyright © 2006 Taylor & Francis Group plc, London, UK

 users have been charged based upon the Hourly
Ontario Electricity Price (HOEP), which being sub-
ject to demand, has been both volatile and gradu-
ally increasing (IESO, 2003–2005). The initial price
before transmission and delivery charges was $23.0
(Can)/MWh, within the first six months, during Sep-
tember 2002, the hourly price momentarily reached
$1028.4(Can)/MWh. This spike resulted in one mine
postponing all but essential safety activities that
required electricity. Since then the price volatility
has decreased slightly but within the last six months,
the hourly price peaked during September 2005 at
$640(Can)/MWh. In more general terms, the aver-
age monthly basic electricity charge in May 2002
was $29.2(Can)/MWh. In the 44-months since then,
through to December 2005, the monthly price never
dropped below this initial value and has reached as
high as $93.7(Can)/MWh, 220% higher. Throughout
this period, the underlying trend was an annual increase
rate of 7.7%. Although, INCO is to some degree pro-
tected through supply contracts, exceeding contract
arrangements and the generally increasing unit cost
of the energy remains a concern; hence the company’s
interest in minimizing electricity consumption. In con-
trast to this, for the mines in northern Manitoba, where
there is no natural gas pipeline and propane is expen-
sive, using more electricity supplied at a fixed price is
being considered as a heating option.
Another parameter that needs consideration is that
each form of energy used can have different attri-
butable green house gas emissions (GHGs). In addi-
tion, across all energy sources, the energy benefits,
GHG reductions and cost savings are neither common
nor equal.
It is in recognition of the numerous interrelated and
variable parameters that can control a potential ven-
tilation reduction’s benefit, that CANMET-MMSL is
developing the process simulation based approach.
2 VENTILATION-ON-DEMAND (VOD)
The concept of providing ventilation in underground
mines, only to the areas requiring air, in the appropri-
ate amount, and then only as long as necessary would
seem to be logical to minimize the cost associated with
ventilation. Despite its apparent simplicity, for a vari-
ety of reasons, the concept has not been widely adopted
by the industry. As a result, many mines operate their
ventilation systems assuming maximum production,
supplying all potentially active areas, and continuous
operation, 24 hours/day, 7 days/week etc.
2.1 Canadian applications
In Canada, CANMET-MMSL has been promoting
some form of ventilation control to optimize mine
ventilation as a way to reduce costs, yet maintain or
17
Copyright © 2006 Taylor & Francis Group plc, London, UK
improve working conditions since the early 1990’s.
To show the concept, a demonstration system was
installed at CANMET’s Experimental Mine in Val
d’Or, Quebec (Hardcastle et al, 1996). Through 1996-
1998, a pilot system was installed nearby on one
operational level of Barrick’s Bousquet Mine, Que-
bec (Hardcastle et al, 1999). Despite both installations
showing the potential of ventilation-on-demand the
system was not widely adopted.
At Bousquet, the mine was considering selectively
controlling their ventilation to match production activ-
ities as an alternative to having to increase the mine’s
ventilation by 50%. This increase would be needed
for additional production levels and a larger engine-
sized production fleet if they continued with their
normal practice of ventilating all potential production
areas. The pilot system was generally a success, hav-
ing the ability to recognize and selectively react to
mobile equipment, control auxiliary ventilation and an
exhaust raise regulator accordingly, and monitor envi-
ronmental parameters to ensure compliance. Notably,
the trial also showed the system could fail to protect the
worker, the importance of controlling/limiting leakage
through open stopes, and the need for continuous envi-
ronmental monitoring for the worker in addition to the
area exhaust. Despite the success, a significant change
in ventilation design requirements pertaining to the
dilution of diesel engine exhaust resulted in VOD not
being introduced throughout the mine. The change to
Quebec’s mining regulations was adopting the CSA air
quality approach with engine specific dilution require-
ments as opposed to a generalized requirement based
upon engine power. The net result of this change was
that some diesel engines with clean engine technology
required 40% less ventilation.
Although not documented, attempts have also been
made elsewhere in Canada, such as the control of aux-
iliary fans with timers. Apparently, these were success-
ful upon installation but the systems would gradually
deteriorate with lack of maintenance and long-term
commitment from production management. Some of
the reasons for such failures were:
• inadequate additional resources (personnel) to sup-
port the systems,
• lack of integration with routine mine functions, such
as communication systems,
• deficient reporting of the cost benefit, or loss
thereof, and
• insufficient or flawed justification with respect to
the cost benefit.
2.2 European VOD applications
In Sweden (Rönnblom, 1996), the Swedish Min-
eral Industry Research Organization, in co-operation
with Boliden, LKAB, Outukumpu and Union Miniere
 Table 2. The increasing airflow and power requirements, and associated operating costs for the development of Creighton
Deep Orebody (Assumes continuation of current constant delivery system).

Tonnage Airflow Power Cost

Time
frame Deep Deep
conceptual Deep orebody Mine orebody Mine orebody Primary Auxiliary Primary Auxiliary Total
schedule activity Levels (tpd) (tpd) (m /s) (m3 /s) (kW)
3 (kW) ($MCan) ($MCan) ($MCan)

Base 1 Production Level, 7400 to 4000 1300 626 150 5,559 3,200 2.42 1.40 3.82
Model 1 Development Area 7530
2003–2007 3 Production Levels, 7400 to 4000 1700 634 315 5,775 3,360 2.52 1.46 3.98
1 Development Area 7720
2007–2011 3 Production Levels, 7720 to 1500 645 382 6,112 3,528 2.66 1.54 4.20
1 Development Area 7900
2011–2015 3 Production Levels, 7900 to 1500 662 433 6,811 3,704 2.97 1.61 4.58
1 Development Area 8080
2015–2019 3 Production Levels 8080 to 1500 670 447 8,082 3,890 3.52 1.69 5.21
8200

initiated a development project in 1990–92 to
introduce “Ventilation-by-demand” as a pilot project
at the small Kankberg mine using radio transmitters
to control dual duty auxiliary fans and a variable
frequency drive (VFD) surface fan. The system was
successful in reducing the overall ventilation by 40%
and energy consumption by 33% while maintaining
the same or better air quality. Further development of
the system was to include a more efficient auxiliary
fan with built-in VFD and flow monitoring along with
more flexible flow demands from the radio transmitter.
Following on from this pilot project, starting in
1999, VOD was gradually introduced at LKAB’s
Malmberget Mine, Sweden, with similar decreases in
energy consumption by the fan and heating systems
(Nensén & Lundkvist, 2005). That mine’s fans have a
comparable energy usage to INCO’s Creighton mine.
3 VOD JUSTIFICATION AT CREIGHTON
In a well established mine, such as Creighton, installing
widespread monitoring, control and communications
purely for ventilation purposes, would be very expen-
sive and could only be justified if there was a proven
cost benefit. Consequently, starting in 2001, stud-
ies were commissioned to predict how the operating
cost could increase, determine more realistic opera-
tional requirements for the mine’s ventilation system,
and explore what savings may be possible with the
implementation of VOD.
The following sections describe the traditional form
of justification employed at INCO’s Creighton mine.
It starts with a prediction of how the mine’s costs
will increase continuing “business-as-usual” as the
mine gets deeper. It then explores the potential for
cost savings with VOD through temperature and gas
monitoring activity studies, followed up with a pilot
proof of concept trial, as currently underway before
full-scale implementation.
3.1 Operating cost predictions
The energy used by Crieghton’s primary and auxiliary
fan systems account for the majority of the ventilation
systems operating costs. Its heating costs, unlike most
Canadian mines, are minimal by virtue of its natural
heat exchanger.
The following predictions do not take into account
any increase in the price of electricity over time.
3.1.1 Modelling the primary delivery system
Creighton utilizes 13 primary fans in its ventilation
system, five sets of two fans in parallel acting as boost-
ers distributed through its underground intake system,
and three fans in parallel as a surface exhaust. The
total combined rated power of these fans is 11,222 kW.
Table 2, derived through ventilation simulations per-
formed in 2002–3, shows the net effects of increased
air volumes with depth and changes due to the con-
centration of production areas at depth through four
planned stages of development. Currently, the mod-
elling underestimates the true power requirements and
costs as observed by the mine, despite this it can still
be used to provide relative changes.
Table 2 shows that although the mine’s total air-
flow only increases by 10%, from 608 to 670 m3 /s, the
flow supplied to the Deep Orebody increases much
more dramatically firstly from 150 to 315 m3 /s as
it comes fully on-line, and then to 447 m3 /s with
additional depth. Consequently, the associated power
requirement and operating costs of the primary sys-
tem increase by 54%. In reality, the increase may be
even greater as no allowance has been given for pro-
duction from higher regions within the mine beyond
2003–2007.

18

Copyright © 2006 Taylor & Francis Group plc, London, UK

 3.1.2 Underground auxiliary systems
At the time of this analysis, Creighton employed ≈200
auxiliary fans, in either single or two in series configu-
ration, with motor sizes ranging from 37 to 56 kW (50
to 75 hp). However, their operating costs are difficult
to analyze due to the lack of data as to when any of
these fans are operating. For example, using an aver-
age of 46.5 kW per fan, the total installed power would
be of the order of 9,300 kW, however, mine data shows
a total connected load of 6,200 kW. This indicates that
on average only ≈130 of these fans are operating at any
time. Furthermore, if this connected load was contin-
uous the annual operating cost would be $2.7 M(Can),
but the mine reports a cost of $1.4 M(Can), which
would tend to indicate only ≈70 of these fans are oper-
ating at any time and an average connected load of
≈3200 kW.
However, it could be assumed that at least the major-
ity of the fans in the active and potential active areas
run continuously. Across the production levels, this
would represent 65 fans, which at an average fan size
46.5 kW would account for ≈3000 kW or 94% of the
average connected load. This seems to be high, as lit-
tle power is left for the remaining fans elsewhere in
the mine.
Overall, this shows that the auxiliary fans warrant
a more detailed study to determine their operational
profile especially as the mine has increased its max-
imum fan to a 187 kW (250 hp) motor. In Table 2,
$1.4 M(Can), the annual operating cost supplied for
2002 has been used as the base model condition, this
indicates power usage of 3200 kW.
Furthermore, it is also difficult to predict how
the power requirements of the auxiliary system will
change with each stage of the mine’s development. It
would be unrealistic to assume they remain constant
with the addition of more and possibly larger fans,
or that their increasing power requirements match the
54% for the primary system, as not all of the fans are
operational. Consequently, in Table 2 an increase of
5% per stage has been assumed in determining the
auxiliary requirements.
3.1.3 Combined cost estimation
Despite the limitations of the analyses for the pri-
mary and auxiliary ventilation systems, Table 2 still
indicates a 36% increase in energy/cost to operate
the ventilation system as originally planned through
to 2015–2019. This was sufficient for the mine to
continue the feasibility study.
3.2 Activity analyses
At the time of the activity study Creighton mine’s oper-
ational schedule was two 10-hr back-to-back shifts
per day followed by a 4-hr blast clearance break, and
19
Copyright © 2006 Taylor & Francis Group plc, London, UK
of the 14 shifts per week, two week-end shifts were
designated as maintenance and service shifts.
The activity within the production shifts and general
need for ventilation to be supplied to specific mining
areas was successfully determined through a combi-
nation of continuous temperature and gas monitoring
at a 1-minute sampling rate.
3.2.1 Temperature studies
During 2001 and 2002, the following four mining
activities were monitored with temperature/humidity
sensors for a total of 125 days:
• two scooptram drawpoints for 28 & 33 days,
• a scooptram backfilling site for 28 days,
• a scooptram ore dump for 8 days, and
• a long-hole drilling site for 26 days.
Where possible, the temperature/humidity sensors
were generally installed to monitor intake air condi-
tions, working area conditions, and return air condi-
tions. This monitoring identified when activities were
taking place, plus the conditions they generate; that the
auxiliary systems for the most part ran continuously
and the temperature changes when fans were stopped
and restarted. This latter aspect of the monitoring was
very significant as it showed an important effect, or
lack thereof, within the activity locations.
When the airflow supplied by auxiliary fans or assi-
sted by local booster fans was terminated or reduced
for a short period, such as across a weekend, there
was no dramatic deterioration of temperature condi-
tions. In certain instances, local conditions improved
due to the cessation of 50 to 100 kW addition of fan
heat in the immediate intake. Even over prolonged
periods, i.e. after activity had ceased and a local fan
had been turned off the temperatures did not appear to
dramatically increase, however it was not possible to
ascertain the magnitude of any residual airflow under
these conditions.
What is important, when considering VOD, is that in
well-established mining areas (i.e. 100 days old), in the
short-term, there were no detrimental effects that could
not be eliminated quickly upon the fans restarting.
Figure 1, a sample of this monitoring, shows the
temperature conditions at five locations around a
mucking operation. In general, they all agree, showing
to varying degrees when activity started and ceased.
When such data are studied on an expanded time scale
it is even possible to discern the cycle-time of the
production equipment, and also how long it typically
takes for temperature conditions to return to the non-
active levels. Figure 2, for the same mucking operation,
shows an average load-haul-dump cycle-time of 8 to 9
minutes, and that 10 to 15 minutes are required for the
majority of the equipment’s temperature influence to
dissipate. From this type of monitoring, it was found
that an auxiliary ventilation system is only needed to
 36
NIGHT SHIFT BLAST DAY SHIFT NIGHT SHIFT
CLEARANCE
Mucking Activity
35 Mucking Activity

01:08 08:55 15:30 18:08 00:20

34 467 mins 158 mins 450 mins
break break break

33
Temp. (°C)

32

31

30

29
22:00 0:00 2:00 4:00 6:00 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 0:00 2:00 4:00
Time (hh:mm)

Figure 1. Typical plot of dry-bulb temperatures around a 250 kW scooptram mucking drawpoint.

18:12 21:25 22:27 00:37

35
Temperature (°C)

33

31

23 Mucking Cycles - 17 Mucking Cycles -

average 8.77mins average 8.12mins
29
17:45 18:15 18:45 19:15 19:45 20:15 20:45 21:15 21:45 22:15 22:45 23:15 23:45 0:15 0:45 1:15
Time (hh:mm)
Face 1st Access Return 2nd Access Area Exhaust

Figure 2. Expanded dry-bulb temperature showing cycle times and decay characteristics.

20

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Blast Clearance Day Shift Night Shift Blasts
06-Jan
Monday 15:50 17:43
Start of Monitoring
08-Jan Tuesday
Wednesday 1.9 hrs
16.2 hrs 1:07
10-Jan Thursday
Saturday
Sunday 2.4 hrs
12-Jan
Monday
Tuesday 41.8 hrs
14-Jan
Wednesday 4.9 hrs 8:24
16-Jan Friday
Thursday
18-Jan Friday
Saturday
Sunday 47.1 hrs Miscellaneous Activity
20-Jan 50.1 hrs
Monday
22-Jan Tuesday
Wednesday
24-Jan Thursday
Friday
Saturday
26-Jan
Sunday
Monday 55.4 & 78.8 hrs
28-Jan
Tuesday
Wednesday
30-Jan Miscellaneous Activity
Thursday
Friday
01-Feb
Saturday 161.2 hrs
Sunday
03-Feb
Monday 66.0 hrs
Tuesday
05-Feb
Wednesday
Thursday
07-Feb
Friday
End of Monitoring
09-Feb
3:30 5:30 7:30 9:30 11:30 13:30 15:30 17:30 19:30 21:30 23:30 1:30 3:30

Figure 3. Activity log derived from temperature monitoring at a mucking drawpoint operation.

operate continuously from the arrival of the vehicle,
through its active period, up to 10 minutes after its last
departure.
3.2.2 Activity log & control implications
Based upon such temperature monitoring it is possible
to obtain an activity log. Figure 3 shows that the activ-
ities at a mucking location are far from continuous
during 767 hrs of monitoring. Sixty periods of scoop-
tram activity, totalling 123.7 hrs were identified, along
with another 32.6 hrs miscellaneous/mucking activity
in the immediate area. On allowing for temperature
clearance, the auxiliary system was only required to
operate 167.4 hrs or 21.8% of the monitored time; this
is also a measure of the systems utilization. Figure 3
also shows the variability of activity: there were 19
shifts with no activity; on the day shift activity could
start as early as 8:24 am or as late as 14:30 pm; and
then end as early as 10:30 am or as late as 15:50 pm.
Also, the active periods in any day could be as short as
45 minutes or as long as 358 minutes. Breaks in activ-
ity ranged from 40 minutes through to 161 hrs, plus
there was no apparent consistently scheduled mid-shift
meal breaks.
With regard to possible ventilation control, for this
location, the primary ventilation system would only
have to operate at its maximum level for two periods
of ≈7.5 hrs per day, 6 days a week to cover all potential
operating scenarios (i.e. 53% of the regular working
week). During the remaining period, it could operate
at a reduced level.
Similar analyses were performed for all the mon-
itored operations and all showed similar “random”
activity patterns. However, they all indicated firstly,
periods of time where the primary ventilation system
could be reduced, and secondly, the overall opera-
tional requirement for auxiliary systems within such
production areas.
3.2.3 Explosive blast monitoring
As shown in Figure 3, blasting was captured with
the temperature monitoring, however it only provi-
ded information on their scheduling. Draeger Multi-
warn™ gas monitors were used to determine blast gas
clearance times with a typical sample shown in Fig-
ure 4. With the exception of carbon dioxide, all the
other sensors went off-scale during the blast. On
excluding this region, it was found that all the moni-
tored gases: carbon dioxide, carbon monoxide, nitro-
gen dioxide and nitric oxide, upon re-scaling, all dis-
played the same profile. So one gas could be used as
a surrogate of the others.
Throughout the clearance evaluations, carbon mon-
oxide was always the last gas to clear to 50% of its
time weighted average limit; consequently, it would

21

Copyright © 2006 Taylor & Francis Group plc, London, UK

 400 8000
BLAST 02:15
350 7000
CO2 TLV = 5000ppm
300 Time to 1/2 TLV= 02:15 to 02:30 6000

CO2 Concentration (ppm)

Gas Concentration (ppm)

NO TLV = 25ppm
250 Time to 1/2 TLV= 02:15 to 02:37 5000

NO2 TLV = 3ppm

200 Time to 1/2 TLV= 02:15 to 02:44 4000

150 CO TLV = 25ppm 3000

Time to 1/2 TLV= 02:15 to 03:02

100 2000
CO NO NO2 CO2
50 1000

0 0
2:00 2:30 3:00 3:30 4:00 4:30 5:00
Time

Figure 4. Sample of blast contaminant decay monitoring.

be the most reliable gas to monitor to protect the
workforce.
Due to the variable nature of the blasts, off
and unknown auxiliary fan conditions, the clearance
results were very inconsistent. However, the stud-
ies showed it could be as short as 35-40 minutes as
opposed to the mine’s current 4-hr window between
shifts plus the allowances in the shifts.
With respect to blasting, Table 3 also lists the earli-
est and latest monitored detonation times at 42 minutes
apart; this, when combined with the 35–40 minute
clearance time requirement could result in the primary
system running unnecessarily with a pre-programmed
operation. For optimum benefit, the blast-time would
have to be fixed or the ventilation flush cycle tied to
the blast.

3.2.4 Primary system control 3.3 Optimized primary system requirements

Table 3 summarizes the earliest start and latest finish
times of activity on both night and day shifts for the
monitored operations These times generally show that
there was a greater delay in starting activity on the
day shift than on the night shift. This would have been
the result of ensuring blast clearance. Both shifts had
comparable active periods, and activity stopped earlier
on the night shift to accommodate blasting. The tem-
perature based activity monitoring also showed two
anomalies that would have to be discouraged if venti-
lation control was introduced. In Table 3 it can be seen
that drilling started slightly early on one night shift.
Not shown, but important to manage, is that backfill-
ing was performed on a Saturday night shift, typically
a non-productive shift, as opposed to the next day’s
(Sunday) scheduled productive night shift.
Overall, the activity monitoring data shows that the
primary ventilation system need only be programmed
to run at maximum capacity only for 7.75 hrs each
productive shift, or 93 hrs per week, to cover all the nor-
mal activities observed. In addition to this, the primary
system is also required for blast gas clearance.
Table 3 concludes with a suggestion for an opti-
mized operational schedule for the primary system.
This would provide twelve maximum flow periods of
7.75 hrs, one per production shift, and three maxi-
mum flow periods per week of 35 minutes for blast
clearance. In combination, this “ideal” schedule would
require the primary ventilation system to run at its
maximum capacity only 94.75 hrs per week, or 56%
of the time.
3.3.1 Secondary system control
For the auxiliary ventilation systems there are var-
ious control options. The relative benefits of each
will be discussed specifically for the mucking activity
shown in Figure 3. The increasing levels of auxiliary
control are:
• Their operation would match the “ideal” schedule of
the primary system, i.e. 460 minutes per shift with
simple timers. This increases their utilization from
22 to 39%,

22

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 3. Summary of observed activity start/finish times and derived “ideal” primary ventilation maximum duty operational
schedule.

Observed ventilation requirements

Day shift (7:30–17:30) Night shift (17:30–3:30) Blast clearance

Observed # Active Earliest Latest Duration Earliest Latest Duration Earliest Latest Spread
activity shifts start finish (hrs) start finish (hrs) detonation detonation (hrs)

Drilling 27 8:16 15:53 7.62 17:211 1:06 4.58

Mucking #2 43 8:24 15:50 7.43 17:43 1:07 7.40
Mucking #1 20 8:33 16:00 7.45 17:48 1:08 7.33
Backfilling 17 8:33 15:40 7.12 17:34 23:29 5.92
Dumping 15 8:36 15:34 6.97 17:55 0:43 6.80
Blasting 8 1:53 2:35 0.70

Maximum capacity operation programming

Day shift Night shift Blast clearance

Fans max Fans min Duration Fans max Fans min Duration Fans max Fans min Duration

Ideal Times (hh:mm) 8:15 16:00 7.75 17:30 1:15 7.75 2:00 2:35 0.58
# Shifts/ 6 6 3
Blasts
Required (hrs) 46.5 46.5 1.75
Max.
Total Max. (hrs) 94.75
Relative 56%
to 24/7

1
Prior to regular shift start-time

• They would only start with the arrival of a vehi-
cle, i.e. machine operator initiated, and then run for
370 minutes (the maximum observed operating time
plus clearance) or to the pre-programmed end of
shift. This increases their utilization to 71%, or
• They would start and stop as dictated by vehicle
arrival and departure plus clearance delay, with this
higher level of VOD control 100% utilization could
be achieved.
The relative merit of each level of control is depen-
dant upon the production activity. In Table 4, which
summarizes the activity monitoring, it can be seen
that auxiliary ventilation utilization is not very high
under continuous 24-hr operation, with the average
being 20%. It is greatest for ore dumping and drilling,
respectively at 31% and 24%, and then drops down to
7% for backfilling.
Based upon the small number of activities mon-
itored, with simple time of day controls that match
the primary system’s 7.75 hrs per production shift, the
auxiliary ventilation utilization increases on average
to 34%, but is 54% for ore dumping and still only 12%
for backfilling. Consequently, ore dumping may only
need the lowest level of control, whereas backfilling
may need the highest level of on/off control.
4 COST SAVINGS WITH VOD
The analyses presented in preceding section have
shown that there is potential to control both the primary
and auxiliary ventilation systems.
4.1 Primary ventilation system savings
The activity logs have shown that maximum primary
ventilation is required 56% of the time in a stan-
dard working week. For the remaining 44% of the
time, it could be reduced to a lower rate. The rel-
ative benefit of this reduced operational period is
very dependant upon the minimum flow specified.
In the cost reduction analysis, it has been assumed
that all flows in productive areas fall back to a vol-
ume suitable for service vehicles. Table 5 compares the
results of ventilation simulations based upon full and
reduced flow requirements for the original schedule
provided for the 2003–2007 and 2015–2019 periods.
This table shows that as the mine deepens, the net
reduction in flow decreases, as more areas have to
be ventilated. Despite this, the cost of running the
primary ventilation system could be reduced by on
average 30%, and the majority of the operating cost

23

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 4. Summary of observed activity and relative auxiliary ventilation operation requirements.

Auxiliary Ventilation Operation Programming

Operation matched to Primary System (Time-of-day) Maximum Requirement 56%

Minimum requirement versus (%)

Operation matched Time required Time monitored
to Activity (hrs) (hrs) 24 hr Operation Time of day

Mucking #2 134 (167)1 767 17 221 31 411

Mucking #1 83 670 12 22
Mucking Average 17 31
Dumping 59 192 31 54
Drilling 152 623 24 43
Backfilling 47 671 7 12
Overall Average 20 35

1
Includes miscellaneous activity and possibly mucking at adjacent location

Table 5. Primary ventilation system operating costs/savings with “ideal” time of day controls.

Working Mine Deep Primary Annual Operating Adjusted Total

Time depth Operating airflow airflow power cost time cost cost
period (m) condition (m3 /s) (m3 /s) (kW) ($MCan) factor(%) ($MCan) ($MCan)

2003 2250 Maximum 634 315 5,775 2.52 54 1.36 1.68

to to Minimum 369 137 1,575 0.69 46 0.32
2007 2350 Reduction 42% 57% 73% 73% Net Saving 33%

2015 2450 Maximum 670 447 8,082 3.52 54 1.90 2.56

to to Minimum 509 280 3,297 1.44 46 0.66
2019 2500 Reduction 24% 37% 59% 59% Net Saving 27%

(>75%) would be attributable to the full ventilation
period.
4.2 Auxiliary ventilation system savings
Determining the mine-wide cost savings for the auxil-
iary systems is just as problematic as determining the
actual cost in section 3.1.2. However, unlike the pri-
mary system, the fans would be either on, or off. The
activity logs have shown that auxiliary fans may be, on
average, only required 20% of the time. Furthermore,
according to the mine, most of the operational fans are
in the production areas and could be subject to some
form of time/activity based control. In Table 6 it has
been assumed that 80% of the installed auxiliary fan
power would be controlled, that it would firstly oper-
ate 56% of the time in-line with the primary system,
and then further optimized to operate only 20% of the
time. The net result in operating cost is a 35% reduc-
tion with shift-based time of day controls, and a 64%
reduction with activity based control.
4.3 Overall system savings
Table 7 combines the cost benefit analyses of the
primary and auxiliary fan systems. However as both
independent analyses are based upon certain assump-
tions, the final savings only serve to give an indi-
cation of the potential savings. Until further work is
performed, the following is indicated:
• Simple Time-of-day controls on both the primary
and auxiliary fans could produce at least a 30%
reduction in power consumption and cost.
• Simple Time-of-day controls on the primary fans
and Activity-based control of auxiliary fans could
produce at least a 39% reduction in power consump-
tion and cost.
This table also shows that the average primary air-
flow with time-of-day based controls is less than the
626 m3 /s supplied in the initial model. Consequently,
if the mine’s natural heat exchanger retains the same
capacity and is currently satisfactory; it should be able

24

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 6. Auxiliary ventilation system operating costs/savings with “ideal” time of day/activity based controls.

Auxiliary fans
Total Controlled Annual Operating Adjusted Total
Time Depth power Control power cost time cost cost Net
period (m) Operating condition (kW) factor (kW) ($MCan) Factor ($MCan) ($MCan) Saving

2003 2250 Fixed 20% 672 0.29 100% 0.29

to to
2007 2350 Controlled Time-of-day 3,360 80% 2688 1.17 56% 0.66 0.95 35%
Activity 20% 0.23 0.53 64%
24/7 Total 1.46

2015 2450 Fixed 20% 778 0.34 100% 0.34

to to
2019 2500 Controlled Time-of-day 3,890 80% 3112 1.36 56% 0.76 1.10 35%
Activity 20% 0.27 0.61 64%
24/7 Total 1.69

Table 7. Combined ventilation system operating costs and the potential savings with the “ideal” time-of-day/activity based
controls.

Average Power
Average Net
primary Net flow Primary Auxiliary Total savings
Time Depth Operating flow reduction system system Combined cost $M(Can)
period (m) condition (m3 /s) (%) (kW) (kW) (kW) $M(Can) (%)

2003 2250 24/7 634 5,775 3,360 9,135 3.98

to to
2007 2350 Time-of-Day 3,869 2,184 6,053 2.64 1.34
34
517 18
Time-of-Day/Activity 3,869 1,210 5,079 2.21 1.77
44

2015 2450 24/7 670 8,082 3,890 11,972 5.21

to to
2019 2500 Time-of-Day 5,900 2,528.5 8,428 3.67 1.54
30
599 11
Time-of-Day/Activity 5,900 1,400 7,300 3.18 2.03
39

to maintain the same conditions throughout the devel-
opment of the Deep Orebody plus be able to support
other areas. Based upon the overall flow reduction
potential the project proceeded.
4.4 Early implementation & additional auxiliary
ventilation system savings
Following the temperature studies, the mine has started
introducing simple time-based controls to its fans in
the Deep Orebody. This will result in the fans being
off for 48 hrs, 29% of the standard week. Although
notably less than the “ideal” potential identified in the
monitoring studies, this introductory schedule still has
an estimated saving of nearly $480 k(Can) per year.
Another factor not considered in the previous cost
analysis is the life-cycle of the auxiliary fan system
within the production area. At Creighton, the develop-
ment cycle typically lasts 2 years and the production
cycle typically lasts 8 years. During this period the aux-
iliary system is regularly extended and then gradually
retracted, consequently the system’s resistance and
leakage is continually varying. In addition, the opera-
tions and associated diesel powered equipment is also
changing. Despite these shifts in required duty, the
installation is normally designed to and operated at
maximum capacity.

25

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Assuming the use of variable frequency drives
(VFDs) on each auxiliary fan another $280 k(Can)
potential saving per year was identified, or $166 k(Can)
per fan lifetime installation. This was sufficient for the
mine to justify using VFDs.
4.5 Further work, further benefits
The analysis of activity based ventilation requirements
at Creighton has already shown the potential for attrac-
tive savings, but there could be even greater potential.
The analysis has specifically shown the need for more
detailed information regarding both the current auxil-
iary fan operation and the simultaneous requirements
of all possible fans mine-wide. This is important in
a mine where potentially 10–12 activities are contin-
uously taking place for a specific period of a shift
somewhere in a multi-level production area. If this sort
of data were available, the feasibility of controlling the
duty of the primary fans could be determined.
This analysis, to some extent, has also assumed that
all areas need to be ventilated to a certain degree; con-
sequently, a concurrent activity/equipment movement
and auxiliary fan operation benchmark data set would
also be beneficial.
Furthermore, the quantities assigned to production
activities have been based upon the tonnage. The tem-
perature investigations have shown to some degree that
conditions in the general travel areas of the mine are
independent of activity. In addition, if the production
equipment has air-conditioned cabs, the tonnage spe-
cific airflow requirements if based upon the resultant
conditions of activity at a workplace may be greater
than needed. Therefore, these requirements also need
to be re-verified.
5 PILOT TRIAL & ADVANCE JUSTIFICATION
The introduction of VOD to the mine has been justified
based upon the potential savings within the operation
of the auxiliary systems.
Consequently, INCO has reached an agreement
with Delta RCS to install a pilot “proof of con-
cept” system on their 7680 level, which would include
vehicle and personnel identification, airflow, carbon
monoxide and temperature monitoring, VFD fan con-
trol, and general fan status/current monitoring: this
system will be installed in early 2006.
The system being installed includes more vehicle
and personnel tracking plus fan monitoring than would
normally be required. However, this is being installed
to provide baseline information with regard to power
usage and a more detailed global activity map. These
items were identified as some of the limitations in the
traditional justification.
In parallel to this pilot installation, CANMET-
MMSL and Penguin ASI are continuing with the
26
Copyright © 2006 Taylor & Francis Group plc, London, UK
development of a graphical process simulator to model
the variation in ventilation requirements within a
dynamic mining operation (Hardcastle et al, 2005).
The current development although being designed
to be flexible to represent any mechanized mining
operation will be based upon the Creighton operation.
In combination, the activity data already gathered
from the temperature monitoring, and that forthcom-
ing from the vehicle tracking and fan monitoring will
be used to calibrate and validate the process modelling
methodology.
6 CONCLUSIONS
For Creighton mine, it was shown through simula-
tion modelling, that if they continue with their current
strategy of providing ventilation continuously, their
power requirements and cost would increase signif-
icantly. In addition there is the potential for the mine’s
natural cooling capacity to become insufficient. For
the Canadian government and INCO, this is a concern,
not only due to cost that could affect mining’s sustain-
ability, but also for the potential environmental impact
of increased GHG generation, which is contrary to the
country’s GHG reduction obligation.
As a result of temperature, blast clearance and activ-
ity monitoring, it has been shown that the flow through
the primary system could be reduced to a lower opera-
tional level 44% of the time. This monitoring has also
shown that on average an auxiliary fan in the produc-
tion area may be required as little as 20% of the time.
In combination, these potential reductions formed the
basis of the mine considering some form of ventilation
management or VOD.
For example, with the appropriate control of the
primary and auxiliary fans, it has been shown that
through time of day and activity based controls, the
mine’s annual power requirements and hence cost can
be reduced by 30–40% compared to an extrapolation
of their current “constant” delivery system. Further-
more, and most importantly, this could be achieved
without a reduction in flow during productive periods.
The net effect of the management of flow, namely
flow reductions in non-productive periods, results in
average flows through to the year 2019 that are less
than those currently supplied. Hence, the need for
introducing mechanical refrigeration may be avoided.
To date the mine has already started to introduce
time-based controls on their auxiliary fans. The mine
will also be installing the pilot VOD system early in
2006.This system will be evaluating a higher level con-
trol strategy based upon activity tracking and will also
be gathering data to better determine the potential of
controlling the mine’s primary ventilation system. This
data would then be available for a higher-level analy-
sis of the benefits using the process simulation based
prediction of ventilation requirement and their spatial
 variation as being developed by CANMET-MMSL and
Penguin ASI.
Once fully developed, the process simulation based
approach will be available to evaluate the ventilation
reduction potential of other mines and the industry as
a whole, as it could readily account for the diverse
and varying controlling parameters, such as energy
pricing and regulatory requirements, that dictate the
final volume required in a mine.
ACKNOWLEDGEMENTS
The authors would like to express their gratitude to
INCO Limited’s staff and management for their coop-
eration in this work and their permission to present
it findings. The authors would also like to acknowl-
edge the funding support of the Government of Canada
Action Plan 2000 on Climate Change towards the
development of the production based simulator and
implementation of the pilot installation.
REFERENCES
Campbell, J.G. 2005. DMRC Cost-Risk Models – Set-
ting Research and Development Priorities for Deep
Mining, Deep Mining Research Consortium, http://
www.deepmining research.org/, Internal Report.
Hardcastle, S.G., Hackwood, J. and Lambert, L. 1995. An
Integrated Mine Ventilation Management System, Proc.
3rd Canadian Conference on Computer Applications in
the Mining Industry, Montreal, Canada, October 22–25,
Ed. H.S. Mitri, Vol. 2, pp. 1–13.
Hardcastle, S.G., Gangal, M.K., Schreer, M. and Gauthier, P.
1999. Ventilation-on-Demand – Quantity or Quality – A
Pilot Trial at Barrick Gold’s Bousquet mine, Proc. 8th
27
Copyright © 2006 Taylor & Francis Group plc, London, UK
U.S. Mine Ventilation Symposium, Rolla, Missouri, June
11–17, Ed. J. Tien, on CD, University of Missouri-Rolla,
pp 31–38.
Hardcastle, S.G., Kocsis, C., Bissiri, Y. and Baiden, G.
2005. Optimizing Mine Ventilation Through The Use Of
Life-Cycle Production Models. Procs. 8th International
Mine Ventilation Congress, Brisbane, Australia July 6–8,
Ed. A.D.S. Gillies, AUSIMM, Carlton, Victoria, ISBN
1-920806-32-6, pp. 373–361.
IESO 2002–2005, Market Summaries – Hourly Ontario
Electricity Price, http://www.ieso.ca/imoweb/marketdata/
market Summary.asp, Independent Electricity System
Operator.
Nensén, P. and Lundkvist, P. 2005. From 167 GWh To
72GWh – Ventilation On Demand In LKAB’s Iron Ore
Mine, Malmberget, Proc. IPPC Conference, Innovative
Energy Efficiency Examples of Different Industrial Sec-
tors – Energy Efficiency in the Cement, Metal and
Petrochemical Industry, http://www.umweltbundesamt.at/
fileadmin/site/ uweltthemen/industrie/IPPC_Konferenz/
Nensen.pdf
O’Connor, D.F., Gibson, D., MacKay, B. and Grupp, D.R.
2002. Creighton Mine, #11 Shaft Exhaust Fan Upgrade,
Proc. 1st North American/9th U.S. Mine Ventilation Sym-
posium, Kingston, Ontario, June 8–12, Ed. E. DeSouza,
Balkema, ISBN 90 5809 387 5, pp. 137–144.
Rönnblom, W. 1996. Ventilation-by-Demand, Proc. 9th
Canadian Institute of Mining (CIM) Operator’s Confer-
ence, Val d’Or, Quebec, March 3–6, CIM, Session 1, Paper
#2, 5 pages.
Stachulak, J.S. 1989. Ventilation Strategy and Unique Air
Conditioning at INCO Limited, Proc. 4th U.S. Mine Ven-
tilation Symposium, Berkeley, California, June 5–7, Ed.
M.J. McPherson, SME, Littleton, Col., ISBN 0-87335-
082-0, pp. 3–9.
Stachulak, J.S. 1979. Computer Network Calculation of
Creighton Mine Mass Flow and Natural Ventilation,
Proc. 2nd International Mine Ventilation Congress, Reno,
Nevada, November 4–8, Ed. P. Mousset-Jones, SME,
Littleton, Col., ISBN 0-89520-271-9, pp. 750–764.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Ventilation planning for mining of the Sarvisuo Lode in the Orivesi mine

A.L. Martikainen
Helsinki University of Technology, Espoo, Finland

ABSTRACT: Orivesi mine is a 720 m deep gold mine in Finland. The mine is currently, on August 2005, in a
standby mode, to which it was placed in the beginning of 2004, when the steeply dipping ore bodies narrowed with
increasing depth and mining was not considered feasible anymore. However, more ore was soon found nearby dur-
ing an exploration program and a feasibility study concerning the Sarvisuo Lode was started in autumn of 2004.
Mining plans for the Sarvisuo Lode consist of establishing access from existing drives that extend from
285 m and 525 m levels from the main Orivesi decline to the vicinity of the lode. As a part of the mine planning,
preliminary ventilation study was conducted during the winter 2005. This study addressed the issues of ventilation
requirements as well as alternative implementation possibilities for ventilation.
In this paper the survey and simulation results for the study are presented and the functionality of suggested
different shaft alternatives is evaluated. Possibilities to combine the old infrastructure with new workings were
considered essential and thus the subject attracted a lot of attention.
During planning two different mining options called ‘small project’ and ‘large project’ were investigated.
Shaft alternatives were tried out for both. Also a set of different requested detail studies either concerning both
or one of the mining options were performed and compared.
A decision to continue mining in the Orivesi mine was released on July 2005. Mining of the Sarvisuo Lode
is planned to start in spring 2006.

1 INTRODUCTION 2 ORIVESI GOLD MINE

As a part of the mine planning for Sarvisuo orebod- 2.1 Location and geology
ies the requirements for ventilation during stoping and
Orivesi gold mine is located in southern Finland, about
alternative implementation possibilities for ventilation
220 km north of Helsinki. The mine site belongs to the
were studied from autumn 2004 to early 2005 to give
early Proterozoic Tampere schist belt located in the
a basis for feasibility calculations. Base material was
central part of the Svecofennian domain. The Tam-
obtained from Sarvisuo mining plans, maps, and coor-
pere schist belt has an almost continuous east-west
dinates obtained from the Orivesi mine as well as
strike with a length of 200 km and a width of around
estimates on the usability of existing infrastructure.
20 km.
Two different mining options were given, the small
The gold deposit consists of five known, almost ver-
project and the large project. The small project consists
tical gold ore pipes, which are located near to each
of a plan reaching from level +220 m to level + 364 m
other. The size of the pipes varies strongly. The biggest
with one level access from the old mine at level
pipe has a cross-section measuring 50 m by 10–15 m
+285 m to the Sarvisuo Lode. In the large project
over a height of at least 400 m. The other pipes have
there are two level accesses above each other on
smaller dimensions. (Outokumpu 2002)
levels + 285 m and +525 m. The mining height of the
large project reaches from level +202 m down to level
+525 m.
2.2 Mining
An old ventilation model made with the Ventsim
program in 2002 was used as a basis for planning. Mining started as open pit mining with underground
This was then supplemented with preliminary Sarvi- test mining in 1994. The underground mine was
suo mining plans. Different ventilation alternatives opened in October 1994. Since 1996 all production has
with either one or two new shafts were finally tested come from underground. Orivesi mine site is presented
to model the ventilation. in Figure 1.

29

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 1. Orivesi mine site.
The mining method used in Orivesi mine is mech-
anised bench and fill stoping with 20 m spaced sub-
levels /6/. Mining starts from the bottom of 80 m high
blocks and continues upwards in 20 m benches. The
stope size varies from 2000 t to 25000 t depending
on the length and the width of the ore pipes. (Out-
okumpu 2002) The mine was placed to a standby in
the beginning of 2004, when the steeply dipping ore
bodies narrowed with increasing depth and mining was
not considered feasible anymore. However, more ore
was soon found nearby. The exploration drilling was
then focused on defining the new Sarvisuo Lode area
located 400 metres from the existing workings and the
decline.
Mining plans for the Sarvisuo Lode are to establish
access from existing drives that extend from 285 metre
and 525 metre levels from the main Orivesi decline
to the vicinity of the lode. An internal decline will
provide level access to the Sarvisuo Lode. (Dragon
Mining 2005)
3 VENTILATION SYSTEM OF THE
ORIVESI MINE
3.1 General
There are three shafts used for ventilation purposes at
the Orivesi mine. Fresh air is conducted to the bot-
tom of the mine and to the production areas via two
shafts, ivn1 and ivn2 assisted with fans, flexible ducts
and ventilation walls. Ivn1 is a raisebored shaft with a
diameter of 1.8 m down to level +235 m.This shaft was
later piecewise lengthened with long-hole method all
the way to level +700 m. Later another series of raises,
which were originally mined to be waste rock pass
for backfill material, was rebuilt for ventilation use.
The diameter of this fresh air shaft, ivn2, is 2 m times
30
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 2. Cross-section and longitudinal section of Orivesi
mine in 2002.
2 m and it has also been constructed using long-hole
technique.
Ventilation walls with doors and without openings
go down to the level +625 m in front of ivn1. Walls
with about (0.6*0.6) m2 openings or tight shotcrete
walls in front of ivn2 and pn have been installed down
to the level of +590 m with pn and to the level of
+610 m with ivn2. (Martikainen 2002) The choice
between a tight wall and a wall with an opening was
made based on the ventilation demand on the level
at the time of decision-making. Flexible ventilation
ducts are used to direct the air to the production areas
from ventilation walls with openings. The ducts have
diameters of 1000 mm in main tunnels and 800 mm
in drifts. The amount of fresh air entering the Orivesi
mine is about 90 m3 /s. (Martikainen 2005b)
The exhaust air comes out of the mine along the
exhaust shaft, pn, and upwards the incline. The cross-
sectional area of the incline ranges from 20 m2 to 25 m2
depending on the depth and the inclination is 1:7. The
cross-section of the exhaust shaft pn is 2 m times 2 m.
Fans and walls help in directing exhaust air.
A cross-section and a longitudinal section of the
Ventsim model of the mine in 2002 are presented in
Figure 2.
 During the winter the downcast fresh air is heated
up. Heating is controlled by temperature sensors and
thus works automatically.

3.2 Current ventilation status

The ventilation situation of the standby mode was
investigated at a surveying visit. The fresh air shaft
ivn1 was functioning well. The amount of fresh air
arriving to the level +700 m was about 40 m3 /s. On
the other hand, the fresh air shaft ivn2 had some prob-
lems. Three of the shaft fans were not functioning,
and there was a cave-in in the shaft, which partly had
plugged the shaft at the level +490 m. In the year 2002
survey ivn2 was found to import about 31 m3 /s of air
into the level +645 m, when working properly. This
can be estimated to respond the shaft capacity when
repaired and equipped with similar equipment.
On the basis of the old model it can be stated that
when both shafts are functioning, they bring altogether
about 88 m3 /s of fresh air into the level +310 m, and
about 87 m3 /s of air into the level +510 m, of which
a small part only is circulating air. This circulating air
is caused by some observed holes in the ventilation
walls at the shafts, and other similar leakages. On the
basis of the year 2002 measurements the real fresh air
amount arriving into the levels can be estimated to be
about 85 m3 /s.

4 PLANNING VENTILATION FOR SARVISUO

4.1 Ventilation requirements

Ventilation requirement analysis started with informa-
tion gathering. The Orivesi mine personnel gave the
details on the planned yearly use of the machinery for
both the small project the large project and the old
mine. The air volumes for the estimated mine life of
three years were calculated based on these estimates.
– Large project + old mine:
– 1st year: 69 m3 /s
– 2nd year: 49 m3 /s
– 3rd year: 36 m3 /s
– Small project + old mine:
– 1st year: 65 m3 /s
– 2nd year: 45 m3 /s
– 3rd year: 35 m3 /s
4.2 Comparing alternative ventilation options
Three different main shaft alternatives were compared
for both the large and the small project. These were
both inlet and exhaust in the new area, exhaust only,
and inlet only in the new area.The diameter of all shafts
was estimated to be 2.4 m, unless otherwise stated.
Several additional versions to these shaft alternatives
Figure 3. Large project with two shafts at Sarvisuo.
and special cases were modelled and analyzed, too.
Altogether six ventilation options were compared. In
Figures 3 and 4 the large project with two shafts and
the small project with two shafts are presented.
1. Only inlet with a shaft having a diameter of 2.4 m
in the new area equipped with the 1900 mm inlet
fan of the mine, no changes in the old mine, caused
serious problems in both projects. The exhaust of
dirty air was difficult and severe air circulation
existed in many areas. Pressure increased exces-
sively throughout the whole mine. The air velocity
in the ramp increased in the vicinity of the exhaust
shaft cave-in up to 5 m/s.
2. Both inlet and exhaust in the new area with 2.4 m
shafts worked almost without problems, when the
inlet air shaft was equipped with the 1900 mm fan
of the mine. No ventilation changes were required
in the old mine. The dirty air was relatively well
controlled, and the pressure differences between
the new area and the old mine were small. The dirty
air routed through to the exhaust shaft even without
an exhaust fan. The amount of fresh air directed to

31

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 4. Small project with two shafts at Sarvisuo.
Sarvisuo Lode area was in both projects, large and
small, about 51 m3 /s.
3. Alternative 3 was actually alternative 2 with 3 m
diameter shafts for both the large and the small
project. Both air velocities in the shafts and the
drift costs of the shafts decreased clearly, espe-
cially when the modelled air volumes were large.
Otherwise this alternative worked similarly to the
previous alternative.
4. An alternative where in the large project a 3 m
diameter inlet raise for fresh air is driven at Sarvi-
suo to conduct all fresh air into the mine, together
with a 2,4 m exhaust raise, was studied, too. So in
this alternative also the old mine part was venti-
lated through Sarvisuo. Alternative 4 was simu-
lated by removing all fans from the old mine, and
directing different fresh air amounts into Sarvi-
suo. The tested air flows were 70 m3 /s, 80 m3 /s or
90 m3 /s.
32
Copyright © 2006 Taylor & Francis Group plc, London, UK
5. Alternative 5, only exhaust in the new area worked
well for the large project, even without a fan in
the large project, when the fresh air was taken in
through ivn2 from the level +530 m. The air flow
directions were well in control in the new areas
and the blind fresh air shaft planned in Sarvisuo
area worked expectedly. The only problem was the
small amount of fresh air obtained into the new
areas, which was 26 m3 /s. With this alternative also
another option, taking fresh air directly from ivn2
into the connecting tunnel at the level +530 m, was
also tested with the large project. In this case the
whole (5,0 × 4,6) m2 connecting tunnel was con-
verted into fresh air route with a booster fan. Then
the maximum obtained air flow was about 64 m3 /s
into the Sarvisuo area. In this case alternative 5 can
be ranked usable. This alternative worked with the
small project, too. All the fresh air was taken in via
ivn2 from level +285 m. The air flow control into
the blind fresh air shaft, up along the ramp and up
via the exhaust raise, appeared to work properly.
The amount of fresh air taken from ivn2 into the
new areas was about 36 m3 /s. In the small project
the best result was obtained, when air was taken
from ivn1 at the level +310 m, and led via a sepa-
rate connecting tunnel into Sarvisuo, assisted with
a booster fan. In this case transfer of 59 m3 /s of air
was achieved, which is satisfactory with the given
usability grades.
6. Inlet only, presented as the alternative 1, was mod-
ified through increasing the shaft diameter up to
3 m in the new areas to get alternative 6. In addi-
tion all the fans in the old mine were removed to
decrease pressure, so that all of the old mine acted
as a single exhaust route. (Martikainen 2005a)
5 RESULTS
5.1 Alternative 1
Alternative 1 with a 2.4 m inlet shaft only in the new
areas was found unusable in both the small and the
large project, because of severe pressure problems and
too high air flow velocities.
5.2 Alternative 2
In the studied alternatives the best usability is obtained
with alternative 2, namely both the inlet and the
exhaust shaft, 2.4 m diameter, in the new areas for both
the large and the small project. The fresh air needs were
met with the old fans of the mine.
The best result in the large project was obtained
through installing the existing 2 × 1600 mm fans of
the mine parallel on the surface, and a booster fan
on the level +430 m. Then about 62 m3 /s of fresh
air was obtained to the bottom of the Sarvisuo area.
 The same arrangement even without the booster fan
directed about 58 m3 /s of fresh air into the mine with
the small project.
Respectively using the 1900 mm fan gave in both
projects about 51 m3 /s of air into the mine, which is,
according to the usability grade calculations, satisfac-
tory for Sarvisuo, if the ventilation of the old mine will
be done with the old ventilation equipment. In this case
ivn2 can be taken out of use in the old mine, and its fans
be moved into the new areas, if needed, because the
old mine can be ventilated with fresh air through ivn1
only.The proposed change of ivn2 into an exhaust shaft
will underpressurise the old mine, which will markedly
complicate the control of exhaust air in the new areas.
Because with alternative 2 some air was noticed to
move between the old mine and the new areas, the con-
necting levels were equipped with air curtains. These
were found to work well and nearly all air movement
was prevented between the new areas and the old mine.
5.3 Alternative 3
Good results were obtained as well with alternative 3,
where the inlet shaft diameter was 3 m. With compara-
ble fans some more cubic metres of air per second was
obtained into Sarvisuo than in the 2.4 m case. With two
parallel 1600 fans on the surface more than 77 m3 /s of
fresh air was blown into the mine in the large project.
In the small project the results were alike, even if the
cost savings were smaller than in the large project.
However, the less fresh air is blown into Sarvisuo,
the less savings are obtained through the expansion
of the shaft. If only about 50 m3 /s of air is decided
to be imported into Sarvisuo, the air velocity will be
fully acceptable even with the smaller diameter shaft.
Then the cost savings obtained through increasing the
diameter to 3.0 m are negligent. In this case increasing
the shaft diameter up to 3.0 m cannot be recommended.
5.4 Alternative 4
Alternative 4 was found to be a functioning solution as
well. In the large project at Sarvisuo a 3.0 m diameter
fresh air raise was driven, through which all the fresh
air was taken into the mine.
The problems of this alternative were airflow direc-
tions. In the internal ramp of Sarvisuo airflow direction
varied greatly in between levels and thus uncontrolled
air recirculation was observed. In addition the exhaust
shaft did not function correctly any longer as there was
not enough pressure. This problem could be mitigated
by the addition of two 1600 mm booster fans in series
into the level +510 m.
More than a half of the fresh air moved directly
into the old mine without any restrictive measures with
the large project. When the air flow into the old mine
33
Copyright © 2006 Taylor & Francis Group plc, London, UK
was restricted by modelling flap walls into the con-
necting tunnels the amount of air sank at both levels
down to about 12 m3 /s, when the fresh air amount was
80 m3 /s. This mostly covers the fresh air need for the
old mine during the first year. In this case the air in
the upper connecting tunnel is mainly circulating air,
which should be taken into account in planning. The
air in the lower connecting tunnel is, on the contrary,
fresh air. In this case the air velocity in the ramp of the
new areas will be from 0.3 m/s to 1.1 m/s upwards, so
no recirculation of air will take place in Sarvisuo.
In the old mine the airflow is split so that half of
the exhaust air came to surface via ramp, half via the
shafts. The distributive measures of air with walls and
ducts in the old mine were not modelled due to lack
of time.
Alternative 4 can be regarded possible, if the
required detailed planning and the inflexibility of fresh
air amount in the old mine are taken into account. Fresh
air could possibly be transported into the old mine
with a ventilation duct. In any case, the implemen-
tation of this alternative, even if possible in theory,
will require complementary investigations that tar-
get to both distribution and adequacy of air in the
old mine.
5.5 Alternative 5
The next alternative, alternative 5, was the exhaust
shaft only in the new areas, which worked well in the-
ory with the both projects. In practice the problem was
the amount of fresh air, which could not be increased
to meet the requirements easily. Despite considerable
effort no remedy was found. These consisted of tak-
ing air from both fresh air shafts in the old mine and
changes for ivn1. The first trial caused uncontrolled
air circulation in the whole mine and thus the prereq-
uisites were no longer met for the operation in the old
mine. The change concerning ivn1 was to take the fresh
air into the new areas from it only. The expansion of
the shaft diameter from 1.9 m to 2.4 m increased the
amount of fresh air, but unfortunately due to the fan
restrictions about 5 m3 /s only, and thus even this did
not solve the problem.
If about 50 m3 /s to 60 m3 /s will be a satisfactory
fresh air amount, the solution is usable in the large
project, when the connecting tunnel between Sarvisuo
and the old mine at the level +530 m will be used as an
airway only, and when a booster fan will be installed
there. In this case almost all fresh air coming to the
level +550 m from ivn1 can be transported into Sarvi-
suo. The airflow down from the level +550 m should
be prevented, too, and ivn2 be repaired to serve the
old mine.
In the small project this alternative could be possible
with similar assumptions. Ivn2 should be repaired to
serve the old mine, ivn1 expanded and equipped with
 a larger fan or twin fans, and the air for the new areas This knowledge can only be used as a basis for
taken from ivn1 at the level +310 m. the final ventilation planning. The approximate fan
In addition in both cases one of the connecting tun- sizes can be determined and the ventilation plan can
nels should be sacrificed fully for ventilation purposes be defined on the basis of this study. There is not yet
or a separate parallel tunnel should be driven to allow enough information for a complete, detailed ventila-
an adequate amount of air to be transported. A duct tion plan, which will be required after the completion
will not transport enough air. In the ventilation tun- of the feasibility study.
nel a booster fan would also be required as well as a
large exhaust fan on surface level in the exhaust shaft.
To prevent leakages to the new areas should the air ACKNOWLEDGEMENTS
brought in via ivn1 be isolated from other spaces with
sturdy ventilation walls. The air flow downwards ivn1 Personnel of the Orivesi mine are appreciated for giv-
must be prevented. ing the opportunity to learn so much new about the
mine. Special thanks go to the contact persons Taito
Ahola and Jaakko Kilponen. Support from Professor
5.6 Alternative 6
Pekka Särkkä is appreciated. Also technical and finan-
In spite of the variations done in alternative 6 the fresh cial aid from Helsinki University of Technology is
air always got too contaminated to be used for venti- acknowledged.
lation before arriving into the old mine. Besides, the
exhaust of dirty air was difficult, and recirculation was
observed in many areas. REFERENCES
Apcom 2001.Technical excursions. http://www.hut.fi/events/
apcom2001/tech_excursions.html, 15.1.2002.
6 CONCLUSIONS
Dragon Mining 2005. http://www.dragon-mining.com.au/
pages/projects/finland/finland.htm, 20.10.2005.
The goal of this study was to product information for Outokumpu 2002. http://www.outokumpu.com/mining/www
a feasibility study. The shaft alternatives presented mining.nsf/(Follow)/E7B593432D5B1694C22569F4002
contain information on routing of main air flows, 9729E, 15.1.2002.
general principles of air distribution, required airflow Martikainen, A.L. 2002.Ventilation planning for Orivesi gold
volumes based on given usability grades, indicative mine. Master’s thesis.
air volumes simulated with different alternatives, and Martikainen, A.L. 2005a. Expansion of ventilation system
possible problems, advantages and disadvantages of and ventilation modelling; Mining of the Sarvisuo ore bod-
ies in the Orivesi mine – final report. Internal report. In
different alternatives.
Finnish.
Out of the six alternatives, two can be considered Martikainen, A.L. 2005b. Comparative evaluation of fogging
good, two usable with a fair amount of extra work and phenomenon in the ramp of three mines in Finland. In pro-
planning, and the remaining two completely useless. ceedings of 8th International Mine Ventilation Congress.
Alternatives 2 and 3 were the best. In both of these Burwood: BPA Digital: pp: 103–110.
alternatives inlet and exhaust shafts were constructed
at Sarvisuo. Of the presented alternatives numbers 1
and 6 can be regarded useless, in other words a single
shaft alternative, the shaft as a fresh air shaft, is not a
functioning solution. Alternative 4, where all fresh air
is taken in through Sarvisuo, seems as well promis-
ing, but needs additional investigations. Alternative 5,
exhaust only through Sarvisuo, can be used if multiple
excess changes and improvements are made. These are
as follows:
– the inlet shaft ivn2 of the old mine will be repaired
to satisfy the fresh air needs of the old mine,
– ivn1 will be expanded to 2.4 m diameter
– one of the connecting tunnels between the old mine
and Sarvisuo will be sacrificed to be used only for
ventilation or an extra tunnel is built for ventilation
purposes

34

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Ventilation upgrade of Minerales Monclova’s Mine 6

J.R. Ponce Aguirre

Minerales Monclova, S.A. De C.V, Palau, Coahuila, Mexico

ABSTRACT: MIMOSA’S Mine 6 started its operations in 1969 with the name of Mine 1 and in 1997, when
the operations of an old area finished, the name of the mine was changed to Mine 6. The methane concentration
in the new area of the mine goes from 8 to 10 m3 /ton. in the coal seam. The main fan since 1997 has been a JOY
M-96-65 with a 900 horsepower and 900 rpm electric motor. The total air volume from 1997 to June 2002 was
an average of 90 m3 /sec with a pressure of 2.7 Kpa and a methane generation of 0.300 m3 /sec. Because a high
methane generation was expected, a ventilation project was designed to bring a safe environment to the coal
miners and to the mine. Besides this, a methane drainage program was started in 2004. This paper will explain
all the activities performed to accomplish the project with success.

1 INTRODUCTION
Minerales Monclova, S.A. de C.V. (MIMOSA) oper-
ates four underground longwall coal mines in the North
of Mexico. Besides these mines, the company has an
open pit working near the outcrop. The coal is medium
to high volatile in rank and is used to supply steel-
making operations in the city of Monclova, located
100 miles from the mines.
The mines develop approximately 30,000 meters/
year and the longwall retreat rate is an average of
4,000 meters/year for a total coal production of
4000,000 tons/year ROM.
Actually, the total air volume for the MIMOSA
mines is 700 m3 /sec. and the methane generation is
340,000 m3 /day in the ventilation and 30,000 m3 /day
in the methane drainage systems (there are methane
drainage systems with vacuum pumps in Mine 5
and 6).
In the last two years the operation of Mines 2 and 4
finished after 20 years operating. Currently, the com-
pany has Mines 3,5,6 and 7 working and the future
projects are Mines 8,9 and 10. This paper will explain
a mine ventilation upgrade performed at Mine 6.
2 VENTILATION BACKGROUND
2.1 Main ventilation system
Mine 6 operates using the longwall system. The layout
of the mine in 2001 is shown in Figure 1.
The access from the surface to the mine is pro-
vided with two parallel slopes with an inclination of 28
percent a length of 1,000 m supported with steel arches
4.0 m wide by 3.0 m high. Both slopes are the fresh air
intake to the mine; also, one of them is the access for
men and supplies. The other is used to bring the coal
to the surface using the main conveyor belt. The mine
uses the exhausting ventilation system with the main
fan on the surface connected to the mine trough a circu-
lar shaped vertical shaft of 4.5 m diameter and a depth
of 250 m.
The main fan is equipped with devices for continu-
ous measuring and recording the fan pressure, bearing
vibrations and temperature, amperes and methane con-
tent in the discharge of the air. The main fan in use
since September 1997 is a JOY M-96-65 with a 900
horsepower and 900 rpm electric motor. This mine has
a high ventilation resistance.
All the electric equipment underground are permis-
sible. There are methane monitors installed at the main
returns, at the development faces and the tailgate in
the longwall. These monitors shut down the electric
power if the methane concentration in the environment
reaches more than 1.0% at the working faces and 1.5%
or more at the main returns.
All the monitors are connected to a computer sys-
tem on the surface continuously recording all the
measurements.
There is a ventilation technician in each mine in
charge of checking every week the airflows in the mine
and also ventilation supervisors in each shift, including
weekends.
2.2 Airflows and methane generation
From 1997 to 2001 when the mine was working with
longwalls 200 m wide by 700 to 1,000 m long and the

35

Copyright © 2006 Taylor & Francis Group plc, London, UK


New Area
0 1500 mts.
500
Figure 1.
developments in an area with a methane concentration
in the coal seam of 4 to 5 m3 /ton. The total airflow in
the ventilation system of the mine was 90 to 95 m3 /s,
fan pressure of 2.700 Kpa with a methane generation
of 0.45 m3 /s and an average daily production of 4,000
metric R.O.M. coal.
No methane drainage system was used in this mine.
2.3 Longwall ventilation
In the longwall ventilation, the gob area is isolated
with concrete seals built in each crosscut left behind
the longwall. A 1 m2 opening with a metallic door is
used in each seal in order to bleed methane from the
36
Copyright © 2006 Taylor & Francis Group plc, London, UK
N
1000
gob to the ventilation system from the immediate seal
behind the advance. This opening is closed when the
longwall gets far from the crosscut where the seal is
located. The ventilation volumes were 10 m3 /s across
the face and 15 m3 /s for methane dilution, for a total
volume of 25 m3 /s.
2.4 Development ventilation
In each development section with two “Alpine AM50”
roadheaders advancing the dual entry system and
crosscuts every 60 meters, the air volume used was
20 to 25 m3 /s. The ventilation of the faces was per-
formed with JOY 36-26-1770 auxiliary fans with a 75
 1,000 0.6
CH4 generation m3/sec.

0.5

CH4 generation m3/sec.

750
0.4

500 0.3

0.2
250
0.1

0 0
3,000 4,000 5,000 6,000 7,000 10 15 20 25 30
Longwall coal production tons/day Daily advance rate (m)

Figure 2. Figure 3.

horsepower, blowing system and fiberglass ducting of

Table 1. Friction factors used for airways.
30 diameter (0.762 m).
Airway type Friction factor, Kg/m3

3 METHANE GENERATION FORECAST Belt entry 15 m2 0.0228

Materials 15 m2 0.0106
3.1 Longwall Return 15 m2 0.0141
For the longwall methane generation prediction, a Belt & materials in parallel 15 m2 0.0037
2 returns in parallel 15 m2 0.0035
statistic of other longwalls in MIMOSA with similar
conditions of depth and coal seam methane concentra-
tion was used.
In this case, we used the information of coal pro-
duction vs. methane generation of the 6 South longwall 4 TOTAL VENTILATION REQUIREMENT
of MIMOSA’s Mine 4 (Figure 2). The methane emis-
sion rate was 650 l/s for a coal production level of With the statistical information the total ventilation
4,500 tons/day. needed for the mine was defined as 160 m3 /s.
According to this information we determined that a
total air volume of 60 m3 /s would be enough, because
our maximum methane concentration allowed in the 5 VENTILATION PROJECT
longwall return is 1.5% and with this air volume we
• The ventilation projections for the new area of
would have between 1.0 and 1.1% of methane.
the mine were reviewed using the VNETPC 2000
software.
3.2 Development • In order to avoid the high resistance problem of the
mine, a new support type and shape for the future
For the development methane generation prediction,
entries in the new area was proposed, this was the
a statistic of a development system of a five-month
roof support with steel arches of 5 m wide by 3.5 m
period of other sections in MIMOSA with similar
high, this support provides an effective cross section
conditions and methane drainage system was used.
of 15.3 m2 . This has been a standard for the last years
In this case we used the information of daily
in mines 2,3,5 and 7.
advance rates vs. methane generation of the 4 North
• Two new connections of the main returns to the
entries at MIMOSA’S Mine 4 (Figure 3).
vertical shaft were proposed also.
The methane emission rate was between 0.4 and
• We used the average friction factor for MIMOSA’S
0.45 m3 /sec.for an advance rate of 25 m/day with two
entries supported with these arches; this has been
“Alpine AM-50” roadheaders.
obtained from ventilation surveys performed in all
With this information we determined that a total air
the mines using the gauge and tubing method also
volume of 40 m3 /s would be enough for the operation
(see Table 1).
of a section with two “Alpine AM-50” miners, because
our maximum methane concentration allowed in the The information for the main fan supplied to the
developments return is 1.5% and with this air volume fan data bank of the software is obtained from the
we would have between 1.0 and 1.12% of methane. manufacturer’s fan curves.

37

Copyright © 2006 Taylor & Francis Group plc, London, UK


Vertical shaft
0 1500 mts.
500 1000
Figure 4.
Several situations of the mining plan were ana-
lyzed. For this paper we will explain the results of one,
with all the equipment of the first longwall and the
development section working in the new project area.
The mine layout proposed is shown in Figure 4.
6 VENTILATION SIMULATION RESULTS
The ventilation projection for the middle of 2004 was
simulated; the schematic is in Figure 5.
• With the proposed arrangement, the fan operating
point predicted was 156.31 m3 /s with a pressure of
2.646 Kpa, for the blade setting of 35 degrees for
the main fan JOY M-96-65.
Copyright © 2006 Taylor & Francis Group plc, London, UK
N
New Area
• The air volume predicted for the longwall was
55.28 m3 /s and for the development sections 43.30
and 45.04 m3 /s.
7 REAL RESULTS
• The first longwall in the new area of the mine started
in April 2004 with a measured total air volume of
62.00 m3 /s, the total air volume of the mine in April
2004 was 154 m3 /s with a fan pressure of 2.69 Kpa.
• The longwall methane generation between April to
October 2004 was from 0.520 to 0.650 m3 /sec.
• The longwall operated from the beginning to the end
with only 12 hours of delays because methane in the
7 months period.
38
 Figure 5.

• The longwall production fromApril 2004 to October

Delays Hours Apr May Jun Jul Aug Sep Oct Total
was an average from 3,400 to 4,500 tons/day, this
Methane 1 6 5 0 12 was according with the mining schedule.
Equipments 167 142 190 226 183 213 223 1,344
• The methane generation of the mine was increased
from 0.450 m3 /sec. to 1.146 m3 /sec.
Operation 85 80 76 47 55 60 50 453
The statistics of the longwall operation are in
LW tons./shift 1,218 1,498 1,426 1,389 1,185 1,319 1,132 1,311
Figure 6.
LW CH4 gen
0.617 0.638 0.609 0.645 0.624 0.541 0.521 0.600
m3/sec.
% CH4 LW
0.8 0.8 0.7 0.7 0.7 0.6 0.6 0.7
return
8 METHANE
Figure 6.
The methane in the ventilation system was increased
also from 0.45 m3 /s in 2002 to a maximum of
• The ventilation of the mine was increased from an 1.146 m3 /s in April 2004. In October 2004, a methane
average 93 m3 /s in 2001 to 154 m3 /s in April 2004, drainage program started. This was performed with in-
a 66% increase. seam horizontal holes drilled with pneumatic drilling
• Even though the total airflow of the mine was machines, the average length of the holes was 120
increased a lot, the fan pressure was almost the same meters.
before and after (2.70 vs. 2.69 Kpa). The average methane volume handled by this sys-
• The resistance of the mine to the ventilation tem in 2002 was 0.200 m3 /sec. The behavior of the
decreased 64%. methane generation for all the mine in the ventilation

39

Copyright © 2006 Taylor & Francis Group plc, London, UK

 160 1.600 10 CONCLUSIONS
CH4 degas. system m3/s
CH4 vent. system m3/s
Air volume m3/s • The ventilation project designed for MIMOSA’S

CH4 generation m3/sec

120 1.200
Mine 6 allowed us to increase the ventilation of the
Vent. m3/sec

mine 66% with the same fan pressure.

80 0.800 • The correlation between the predicted airflows with
the VNETPC 2000 software and the real numbers
40 0.400 is been more than 95%.
• Even though the methane generation in 2004 was
2.5 times bigger than in 2002, we have been able
0 0.000
to accomplish and maintain the advancement and
Jun-02
Aug-02
Oct-02
Dec-02
Feb-03
Apr-03
Jun-03
Aug-03
Oct-03
Dec-03
Feb-04
Apr-04
Jun-04
Aug-04
Oct-04
Dec-04
production goals of the mine with almost no delays
because of methane.
Figure 7.

ACKNOWLEDGMENTS
and methane drainage systems from 2002 to 2004 is
in Figure 7.
The author wish to thank the ventilation supervisors
of the MIMOSA mines for the information provided
to help in the preparation of this paper.
9 FUTURE PLANS

The long term project for Mine 6 includes the require-

ment of other vertical shaft for exhausting, in that way,
the current shaft will be used like intake to reduce the
resistance of the mine to the ventilation.
A methane drainage system is required also.

40

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Transition of the Henderson Mine ventilation system to the

new lower levels

D.M. Loring & B.V. Nelson

Climax Molybdenum Company, Henderson Operations, Empire, Colorado, USA

ABSTRACT: The Climax Molybdenum Company’s Henderson Mine, owned by Phelps Dodge Corporation,
is a 27,200 tonne per day (30,000 ton per day) molybdenum mine located 69 km (43 miles) west of Denver,
Colorado. Henderson maintains a relatively large and complex ventilation system that has undergone many
changes through the years.This paper describes the history of the Henderson ventilation system and the challenges
encountered in the transition to the new production level that is currently underway. Main fans were installed in
underground exhaust drifts, main surface exhaust fans were optimized, and mine air heating systems have been
upgraded. Completion of the Henderson 2000 project expanded the mine to levels below which the ventilation
system was originally designed. The transition to the new lower production level has required full utilization
of capacity and optimization of the system, as two production levels and numerous development areas must
be supported simultaneously. Creative use of the 7655/35 ventilation level, installation of numerous ventilation
raises, and continuous modifications and adjustments to the ventilation system have allowed Henderson to
transition into the new production area.

1 INTRODUCTION

The Henderson Mine is located 69 km (43 miles)

west of Denver, Colorado, at an elevation of 3,170 m
(10,400 feet) above sea level. The mill is located 24 km
(15 miles) west of the mine, on the other side of
the Continental Divide, at an elevation of 2,800 m
(9,200 feet) above sea level. The ore body is located
more than 900 m (3,000 feet) below the surface, and is
accessed through an 8.5 m (28 foot) diameter service
shaft. Ore is crushed underground and transported to
the mill site on three 1.2 m (48 in.) wide conveyor belts.
An LHD panel-caving system has been used to extract
approximately 170 million tonnes (187 million st) of
molybdenum ore to date.
Henderson Mine utilizes an exhaust ventilation sys-
Figure 1. Henderson mine surface ventilation facilities.
tem. Ventilation is supplied to the mine through the
8.5 m (28 foot) diameter No. 2 service shaft, the 7 m
(23 foot) diameter No. 3 intake shaft, and the 17 km
2 HISTORY OF HENDERSON VENTILATION
(10 mile) long ore haulage tunnel. Air is exhausted
through the 7 m (23 foot) diameter) No. 1 exhaust
2.1 The original ventilation system
shaft and the 9.75 m (32 foot) diameter No. 5 exhaust
shaft. An old exhaust shaft – No. 4 shaft – is no longer The original Henderson ventilation system was a push-
used. Three 930 kW (1,250 hp) fans are installed at pull system consisting of the three 930 kW (1,250 hp)
the collar of No. 1 shaft; only one of these is oper- parallel exhaust fans on No. 1 shaft and two identical
ated under normal conditions. Six 225 kW (300 hp) 930 kW (1,250 hp) fans serving as surface intake fans
main underground fans feed the No. 5 exhaust shaft. on the No. 3 shaft. These five fans were 3 m (10 foot)
In addition, booster fans ranging from 37 kW (50 hp) diameter Joy 120-72-880 axial mine fans. In 1985 the
to 112 kW (150 hp) are utilized underground to direct #3 shaft intake fans were removed, as they did not ben-
air as needed. efit the system and design modifications made them

41

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 2. Henderson Mine overview.
unnecessary. This change resulted in a net loss of only
24 m3 /s (50,000 cfm) to the system, out of a total mine
airflow of about 873 m3 /s (1.85 million cfm).
2.2 The haulage tunnel
Two 375 kW (500 hp) exhaust fans were originally
installed at the base of No. 4 shaft, halfway through
the ore haulage tunnel. These fans pulled approxi-
mately 132 m3 /s (280,000 cfm) from the portal and
66 m3 /s (140,000 cfm) from the mine, exhausting up a
3.4 meter (11 foot) diameter raise. The electric trains
suffered maintenance problems due to the condensa-
tion caused by the boundary between cold, dry intake
air and warm, humid exhaust air. In 1985, the fans
were removed from service and covered. This con-
verted the haulage tunnel entirely to intake, gaining
approximately 137 m3 /s (290,000 cfm) of intake air for
the mine.
The tunnel intake air was used primarily to ventilate
the 7500 rail haulage level, but was also redirected to
the 7655 intake ventilation level. A system of CO sen-
sors and air doors was installed in the haulage tunnel
to minimize the risk from a fire. The system would
automatically redirect contaminated air directly to the
exhaust system.
2.3 Mine air heating
All three mine air heating systems were replaced
between 1994 and 1997. The haulage tunnel portal
system was originally a 2,000 kW electrical resistance
42
Copyright © 2006 Taylor & Francis Group plc, London, UK
system. It was replaced in 1994 with a direct-fired nat-
ural gas system consisting of four burners with a total
capacity of 2,350 kW (8 million BTU per hour). Por-
tal air heating costs were reduced by about 70 percent
with this system.
The No. 3 shaft system burned diesel fuel and drain
oil, and included a glycol recirculation system. It was
replaced in 1996 with a direct-fired natural gas sys-
tem, consisting of two 2,930 kW (10 million BTU/hr)
heaters.
The No. 2 shaft heating system was originally
designed as a natural gas system with a glycol heat
exchanger, similar to that used in the No. 3 shaft. The
potential for glycol leaks in the No. 2 shaft – the main
access to the mine – caused environmental concerns.
This system was also becoming increasingly unreliable
and expensive to maintain. In 1997 this system was
replaced with a direct-fired natural gas system con-
sisting of four 2,930 kW (10 million BTU/hr) heaters.
2.4 Henderson 2000
The Henderson 2000 Project converted the mine from
train haulage with a surface crusher to a conveyor
haulage system with an underground crusher. The
installation of the underground crusher and truck
haulage ore gathering system required an increase in
volume and a shift of ventilation to lower levels. Ramps
were driven from the 7500 rail haul-age level to the
7065 truck level and 6920 reclaim level.
Two 5 m by 5 m (17 foot by 17 foot) drifts were
mined to access the bottom of the No. 5 exhaust shaft
from the 7635 ventilation level. Each of these drifts
 Figure 3. Henderson Mine ventilation cross section.
Figure 4. 225 kW (300 hp) fans in the 8XD exhaust drift.
had two 2.1 m (7-foot) diameter, 225 kW (300 hp) fans
installed in parallel. This increased mine airflow by
140 m3 /s (300,000 cfm) to a total of 990 m3 /s (2.1 mil-
lion cfm) with a mine power consumption increase of
225 kW (300 hp), or nine percent.
In 2001 ventilation needs declined as the 7700 pro-
duction area decreased in size. The 7 m (23 foot) diam-
eter No. 1 exhaust shaft was handling approximately
57 percent of the total 990 m3 /s (2.1 million cfm) mine
airflow with only 50 percent of the cross-sectional area
of the 9.75 m (32-foot) diameter No. 5 shaft. Two more
225 kW (300 hp) fans were purchased and installed
on the 8035 ventilation level to shift volume to the
No. 5 shaft. Only one of the 930 kW (1250 hp) surface
fans was utilized following this change. The net result
reduced mine volume by about 70 m3 /s (150,000 cfm),
but reduced power usage by 520 kW (700 hp) or 19
percent.
3 TRANSITION TO 7210 LEVEL PRODUCTION
3.1 Introduction
The first bell shot was taken on the new 7210 pro-
duction level in August 2004, which initiated the
43
Copyright © 2006 Taylor & Francis Group plc, London, UK
caving process on the new production level. The first
regularly scheduled production draw assignments
began in March 2005. Production will transition to
7210 from the 7700 level between now and the middle
of 2007. Those drifts will be bulkheaded and removed
from the ventilation system as they are exhausted.
Development activity continues on the 7210 produc-
tion level, 7270 undercut level, 7150 ventilation level,
and 7065 truck haulage level.
3.2 Design
The original Henderson mine design did not envision
levels below 7500, so the shafts were not sunk any
deeper. The Henderson 2000 project allowed the mine
to expand below this level. Access and ventilation
was achieved with a number of ramps and ventila-
tion raises from the ventilation levels and access drifts
above. Intake air reaches the lower levels from the #2
man/materials shaft via LA ramp. This feeds 7210,
7150, and 7065 levels. Air from the #3 intake shaft
flows into the 8100 level, drops down intake raises to
the 7655 ventilation level, then is distributed to the
lower levels through three intake raises.
The PC1 conveyor ramp from the underground
crusher is ventilated by both intake air from the
Haulage Tunnel and from a fresh air raise from the
7655 vent level. The PC1 conveyor drift has a direct
exhaust raise part way up the ramp.
Nearly all exhaust air leaves the lower levels
via fanned exhaust raises to the 7635 exhaust level
(Fig. 5). Constrictions limit the capacity of certain
exhaust drifts, where in at least one case over 378 m3 /s
(800 kcfm) is sent through one (16 by 16 ) entry. In
addition, portions of these levels are within the new
7210 cave area, necessitating bulkheading of multiple
drifts and loss of parallel intake and exhaust air paths.
3.3 Transition
The transition period to 7210 has had its difficul-
ties with regard to ventilation. During the transition,
 Table 2. Post-transition design airflows by area.

Airflow
Level m3 /s kcfm

8100 Cave leakage 33 70

7700 Production area 57 120
7700 Maintenance shop 47 100
7500 Shops & Settlers 47 100
7210/7270 Production & Dev. 198 420
7150 & Other development 71 150
7065 Truck haulage level 165 350
Figure 5. 7655/35 levels with lower level raise connections. PC1 Conveyor ventilation 94 200

Table 1. Transition airflows by area.

to the lower production levels from the upper levels
Airflow
as activity on the 7210 production area increased. The
Level m3 /s kcfm total flow off the 7150 level, which primarily ventilates
the 7210 production level and 7270 undercut level,
8100 Cave leakage 33 70 increased from about 155 m3 /s (330 kcfm) to about
7700 Production area 123 260 203 m3 /s (430 kcfm), with the capacity to increase
7700 Maintenance shop 61 130 further to over 245 m3 /s (520 kcfm).
7500 Shops & Settlers 94 200
7210/7270 Development 179 380
As 7700 production drifts continue to exhaust, drifts
7150 & Other development 47 100 will be bulkheaded and removed from active ventila-
7065 Truck haulage level 184 390 tion. Jet fans used to direct air into active drifts will
PC1 Conveyor ventilation 94 200 be removed when quantities in all production drifts
exceed the design criteria. Air will be redirected to
the lower levels to meet design criteria as activity
transitions to new areas of the mine.
large ventilation quantities were required on multi-
ple levels to support production needs on two levels,
development on five levels, as well as the normal activ- 4 RADON CONTROL
ities such as truck haulage and maintenance shops.
Table 1 identifies the ventilations requirements during 4.1 Source
the transition period. Radon has been a significant issue at Henderson. Cav-
Additional ventilation was required for special ing and fracturing of the orebody liberates radon gas
projects such as TDR (cave monitoring) cable instal- from the granites, which contain trace amounts of ura-
lation on 7700 and 7500 levels, infrastructure removal nium, mostly in the form of rare earth minerals such as
and mud settling drifts on 7500 level, and other special euxenite [(Y, Ca, Er, La, Ce, U, Th) (Nb, Ta, Ti)2 O6 ],
needs such as exploration drilling. These demands, in aeschynite [(Y, Ca, Fe, Th) (Ti, Nb)2 (O, OH)6 ], and
addition to the normal maintenance shop and conveyor ilmenorulie [(Fe, Mn) (Nb, Ta)2 O2 ]. Thorite [(Th,U)
system ventilation as well as ordinary leakage, placed SiO4 ] and uraninite (UO2 ) have also been noted in
occasional strain on the ventilation system. As produc- trace amounts in tails assays. Drawpoints create path-
tion drift exhaust raises and ventilation controls were ways for the gas to flow through, and overpressure
established in the lower levels, this became less of a from auxiliary fans in the undercut level allow some
problem. of the radon to escape to the production area.

3.4 System modifications 4.2 Sampling

In September 2005, three 1.8 m (6 foot) diameter
112 kW (150 hp) booster fans were added to two par-
allel raises exhausting from the 7150 ventilation level.
Two of these fans replaced existing 45 kW (60 hp)
exhaust booster fans. The third was added to an exist-
ing installation that already included one 112 kW
(150 hp) fan. These fans help to divert additional air
30 CFR 57 requires quarterly radon sampling when
radon concentrations of 10 to 30 pCi/l [0.1 to 0.3 Work-
ing Levels (WL)] are found in active work areas, and
weekly sampling if concentrations are found above
30 pCi/l (0.3 WL) in work areas. Henderson performs
sampling at least monthly to identify potential prob-
lem areas and take corrective action. Between January

44

Copyright © 2006 Taylor & Francis Group plc, London, UK

 2004 and June 2005, production areas have averaged
about 10 pCi/l (0.1 WL) radon.
4.3 Control measures
A number of steps have been taken to reduce radon
concentrations in active work areas. The first is main-
taining a negative pressure on the cave, the second is
isolating the caved muck from active ventilation, and
the third is maintaining positive pressure on the active
undercut level.
The primary method for radon control is maintain-
ing negative pressure on the cave. The 4X drift, a
regulated exhaust drift on the old 8035 ventilation level
pulls about 28 m3 /s (60 kcfm) from the 8100 Level
cave. Intake drifts feeding into the cave on 8100 were
bulkheaded in 2004, allowing more air to be pulled
through the cave from 7700 Level. This negative pres-
sure prevents a majority of the radon from seeping
down into the production areas. The air exhausted
through 4X from the cave generally measures over
1000 pCi/l (1 WL). The regulator is gated and locked,
and ventilates directly to an exhaust raise.
Bulkheads have been installed on the 7700, 7635,
and 7500 levels to isolate the new 7210 cave area from
the ventilation system. This prevents intake air from
entering the cave, picking up radon, and contaminating
the work areas.
Pressurizing the undercut level with auxiliary fans
does two things: ventilates the undercut drift for the
ring drillers and cave crew loading bell shots, and pre-
vents cave air containing radon from contaminating
the undercut level.
The drawback to this method is that it tends to force
air through the production drawpoints below, which
may contain sufficient radon to cause concern. As of
the writing of this paper, the 30 pCi/l (0.3 WL) thresh-
old had not been exceeded under normal ventilation.
It is worth noting that the cave broke through to
7500 level in late August 2005. This resulted in a short
period of fogging on the 7210 production level, and
higher than average radon levels. Airflow through the
cave from 7500 was found to be the cause, and was
quickly rectified by isolating the area. Radon levels
dropped soon after, and the fogging was almost entirely
eliminated.
5 FUTURE CHALLENGES
The Henderson ventilation system will remain a chal-
lenge. Following completion of the 7700 production
level in mid-2007, re-optimization of the ventilation
45
Copyright © 2006 Taylor & Francis Group plc, London, UK
system will be performed. Ventilation needs will shift
to the lower levels. Booster fans will be readdressed.
A remote monitoring project has been proposed to
add real-time pressure monitoring to all underground
fixed fans (main and booster fans), as well as upgrade
and add to the airflow monitoring system. This will
add to the fan monitoring and control system already
in place to provide alarm capability.
Excessive leakage and underutilized air in the ven-
tilation system are persistent problems. Improvements
to address these situations will make better use of
intake air and increase the overall efficiency of the
ventilation system.
Diesel particulate matter is a continuing issue that is
being addressed in order to meet the 160 µg/m3 MSHA
standard. All new underground diesel equipment pur-
chases include Tier 2 engines. Tier 3 engines are
currently being tested. A Caterpillar C11 engine has
been installed on an evaluation basis in one produc-
tion R1700G LHD. Results are being assessed. Fuel
additives and engine throttle limiters are other options
currently under consideration.
6 CONCLUSION
The Henderson Mine ventilation system has under-
gone many modifications through the years. These
changes have improved the efficiency and reduced the
cost of ventilation and mine air heating, while allowing
the system to transition effectively between production
areas.
The implementation of the Henderson 2000 project
and initiation of production from the 7210 level
brought new challenges as the system was reconfig-
ured and extended. Success depended on careful bal-
ancing based on ventilation requirements, and regular
monitoring of conditions.
REFERENCES
Keskimaki, K.W. & White, R.C. 1989. Optimizing Mine Ven-
tilation at the Henderson Mine. Proceedings of the 4th US
Ventilation Symposium. Littleton: SME.
Nelson, B.V. & Harney, D.C. 1995. Remote Control and Mon-
itoring of Ventilation at the Henderson Mine. Society of
Mining Engineers 1995 Annual Meeting preprint, Denver.
Nelson, B.V. & Fronapfel, L.C. 2002. Recent Ventilation
Improvements at the Henderson Mine. Society of Mining
Engineers 2002 Annual Meeting preprint, Phoenix.
White, W.H. et al. 1981. Character and Origin of Climax-
Type Molybdenum Deposits. Economic Geology 75th
Anniversary Volume: 270–316.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Ventilation planning at Minerales Monclova’s Mine 7

J.R. Ponce Aguirre

Minerales Monclova, S.A. De C.V., Palau, Coahuila, Mexico

ABSTRACT: Minerales Monclova, S.A. De C.V. (MIMOSA) is owned by GRUPO ACERERO DEL NORTE
(GAN). This company has iron and coal mines to feed it’s steelmaking facilities located in the city of Monclova
in the North of Mexico in the state of Coahuila. It also has coal mines to supply steam coal for two power
generation stations near the border with Texas. These two stations are currently producing 8% of the country’s
electricity. MIMOSA has a new project in the Saltillito coal basin (Mine 7) located 40 kilometers far from the
current mines. The Mine 7 project has a total amount of 30 million tonnes of recoverable coal resources and the
first longwall is scheduled to start at the middle of 2006. This paper explains the ventilation planning process of
the mine from the beginning in November 2001 to the current date and the future projections.

1 INTRODUCTION

The history of the coal industry in Mexico, as in other

natural commodities, has two epochs, which are sep-
arated by the Mexican social revolt of 1910. The first
record of commercial mining of coal dates back to
1850 in the state of Coahuila.
Coal has contributed importantly in the development
of industry in Mexico, starting with the development
of the railroad industry.
Although coal was substituted by oil in the period
of 1910 to 1954 it remained a major raw material for
the steel industry. The steel industry started formally
in 1930, and with it the mining of iron ore and coal.
Coal mining in Mexico increased considerably with
the opening of coking facilities in 1954.
Starting in 1960, the CFE (Mexican Electricity Figure 1.
Commission) developed a program on the generation
of electricity through two coal-operated thermoelec- MIMOSA’S annual run of mine coal production is
tric power stations, which were finally set up in Nava, an average of 5 million metric tons, and the company
Coahuila, from the late 1970s to early 1990s. coal reserves in the Sabinas sub-basin are 400 million
This program generated a very extensive coal explo- metric tons.
ration activity in the Fuentes-Rio Escondido Basin of MIMOSA is presently operating the underground
Northern Coahuila and a consequent increase in coal mines 3,5,6,7 and one open pit. (Mines 2 and 4 finished
mining activities (Verdugo and Ariciaga, 1991). Sev- operations in the last two years).
eral attempts have been made to identify and calculate The coal is medium to high volatile in rank and
all the coal resources of Mexico since then, but most is used to supply steelmaking operations in the city of
of the efforts now have been focused on the largest, Monclova which is 100 miles from MIMOSA’S mines.
most economical 10 basins that are located in the state Each mine is equipped with a longwall system and
of Coahuila in northern Mexico (Fig. 1). the development of the longwall panels is performed
The current total annual production of steam and withAlpineAM-50 and IBS SM130 roadheaders using
metallurgical coal of Mexico is approximately 15.4 a dual-entry system.
million metric tons; 6 million is metallurgical coal and The total methane liberation in the ventilation sys-
9.4 is steam coal used locally in Coahuila to generate tems of the underground mines is an average of
electricity. 400,000 m3 of methane per day.

47

Copyright © 2006 Taylor & Francis Group plc, London, UK

 MIMOSA implemented methane drainage systems
in Mines 2, 4, 5 and 6 from 1992 to 1999 under a
horizontal directional drilling contract with Resource
Enterprises (REI) from the United States, under this
contract, REI drilled almost 90,000 ft of in-seam bore-
holes with an average length of 2000 ft per hole. Since
1999, MIMOSA has been doing its own degasifica-
tion drilling with short in-seam holes with an average
depth of 600 ft, and this year a directional drilling pro-
gram has been started with MIMOSA’S own people
after a training period with REI. This paper explains
the ventilation planning process of the mine from the
beginning in November 2001 to the current date and
the future projections in order to reach the total air
volume of 260 m3/sec expected to be required for the
operation of one longwall and five road-headers. Figure 2.

2 BACKGROUND

MIMOSA’S Mine 7 is a new project in the Saltil-

lito coal basin located 40 km from the current mines
(Fig. 2).
Actually this mine is driving the entries for its first
longwall panel. The elevation of the mine is 420 m
above the sea level. This mine is different of the other
mines of the group because in the Saltillito basin the
coal seam has an inclination of 9 degrees vs. the seam
inclination of the other MIMOSA mines, which is 1 to
2 degrees. Another difference is the coal methane con-
centration in the coal seam. In the Sabinas Basin our
records indicate a range from 4 to 12 m3/ton and in the
Saltillito Basin this amount goes from 8 to 22 m3/ton, Figure 3.
and the permeability is much lower also (Fig. 3).
The Mine 7 project has a total amount of 30 mil-
lion tonnes of recoverable coal resources and the first Table 1. Friction factors used for airways.
longwall is scheduled to start at the middle of 2006. It
Airway type Friction factor, Kg/m3
will be located in a depth of 120 m and the maximum
operation depth of the deepest longwall is projected Belt entry 15 m2 0.0228
to be 450 m. The access to the mine is through three Materials 15 m2 0.0106
slopes supported with steel arches 5 meters wide and Return 15 m2 0.0141
3.5 meters high. The slopes were started near the out- Belt & materials in parallel 15 m2 0.0037
crop left by and old open pit. The first works on surface 2 returns in parallel 15 m2 0.0035
took place in November 2001 and at the end of June
2005, almost 12,000 m of entries have been developed.
The operation will be performed with one longwall entries have been obtained from ventilation surveys
using the two entry system and five road-headers. performed in all the mines using the gauge and tubing
method (Table 1).
At the beginning, the mine started driving the slopes
3 VENTILATION SYSTEM EVOLUTION with three roadheaders and the ventilation was pro-
vided with 2 Spendrup fans installed in parallel on the
The ventilation projections for the mine have been surface in one of the slopes. These fans supplied a total
reviewed using the VNETPC 2000 simulation soft- air volume of 43 m3/sec.
ware, this software has been in use in MIMOSA since The mine operated with these fans from September
1989 for the ventilation planning of all the mines and 2002 to December 2003. The ventilation calculation
also for analyzing any major change in the ventilation for this situation projected 45.11 m3/sec of ventilation
systems. The average friction factors for MIMOSA’S air (Fig. 4).

48

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 5. Picture of the fan installation.

Figure 6. Operating point of the fan in January 2004 and

performance chart.

Figure 4.

In January 2004 two JOY M-72-36 fans with 350

horsepower motors were installed on the surface in
one of the slopes, one of these fans was the main and
the other served as a backup. Each one of these fans
supplied 90 m3/sec with a pressure of 1.1 Kpa.
In 2004, the development advance rate of the mine
was increased because a new system of 10-hour shifts
was implemented. With this system, the mine has been
operating seven days a week.
This was good for the operation of the mine, but the
methane generation was bigger also.
This condition required improvement of the venti-
lation volume. To deal with this situation, a new and
bigger main fan (JOY M-9650) was installed at the end
of 2004 with a 900 horsepower motor. The fan started Figure 7. This chart shows the operating points before and
operation in January 2005. after the new fan was installed.

49

Copyright © 2006 Taylor & Francis Group plc, London, UK

 10.000 1.500 3.1 Methane generation, ventilation and
Developments mts./year

CH4 generation m3/sec.

8.000 1.200
developments advance rate
6.000 0.900 In October 2005, the development of the first longwall
4.000 0.600
panel was accomplished with success.
The maximum methane liberation of the mine to
2.000 0.300 the ventilation system was 1.3 m3/sec. and since April
0 0.000 2005 it was more or less in this range (Fig. 8). The
2002 2003 2004 2005
Developments mts. 669 899 5.176 8.837
total air volume of the mine was increased also from
CH4 m3/sec. 0 0.04 0.62 1.129 43 m3/sec. in 2002 to 182 m3/sec. in 2005 (Fig. 9).

Figure 8. 4 LONG-TERM VENTILATION PROJECTION

The ventilation projections for Mine 7 are to handle

more air than the other mines of the group due to the
200 1.500
higher methane concentration in the coal seam. The
average for the current mines is 160 m3/sec. and for
CH4 generation m /s
160 1.200
3
Ventilation m3/s

Mine 7 this amount is expected to be between 250 to

120 0.900
300 m3/sec. even though a gob well program will be
80 0.600 used to capture methane from the longwall gobs.
A shaft for ventilation exhaust will start to be devel-
40 0.300
oped in 2006 and it is scheduled to be ready by the
0 0.000 middle of 2007.
2002 2003 2004 2005
Vent m3/s 43.0 43.5 96.5 182.2
Once this shaft is ready, the three slopes will be
3
CH4 m /s 0 0.04 0.62 1.129 intake instead of two.
The layout for the whole project with the coal
Figure 9. seam depth curves is in Figure 10 and the ventilation
schematic for the situation in 2008 is in Figure 11.

Figure 10.

50

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 11.

Several simulations have been done for the project 6 CONCLUSIONS

but, for this paper the one for 2008 is going to be used.
The assumptions for the ventilation simulations
were as follows:
• Friction factors for entries are according to Table 1.
• Entry cross section is 13 m2 . (The original size was
15 m2 , but the reduction was considered necessary
due to possible floor heaving.)
• Entries perimeter is 13.5 m.
• Cross section for overcasts 10 m2 .
• The fan data supplied to the software was the 20
degrees curve for a JOY M-120-58-880 main fan.
• Shaft diameter is 5 m.
5 VENTILATION SIMULATION RESULTS
The results obtained with the simulation above were:
• Total air volume for the mine 273.54 m3/sec.
• Fan pressure 1.603 Kpa.
• Total air volume in the longwall return was
89.6 m3/sec.
• Development sections were supplied an average of
50 m3/sec each.
• The ventilation arrangements provided for Mine 7
between 2002 to 2005 allowed MIMOSA to accom-
plish the development schedule of the first longwall
panel on time.
• The development rate of the mine has been 70%
higher in 2005 vs. 2004 even though the methane
generation of the mine was 80% larger in 2005.
• The projected layout of the mine allowed MIMOSA
to increase the ventilation quantity in 2005 by 90%
over the quantity in 2004.
• This has been required for methane dilution
purposes.
• The ventilation project for upcoming years will pro-
vide 50% more air to the mine; however, a methane
capture program must be started.
REFERENCE
Querol Suñe, Francisco. 2005. Mexico Country Profile.
Methane to Markets Partnership.

51

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Maximizing the ventilation of large-opening mines

R.H. Grau III, R.B. Krog & S.B. Robertson

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute for
Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh PA, USA

ABSTRACT: The National Institute for Occupational Safety and Health (NIOSH) has conducted research
to improve the ventilation of large-opening mines. Large-opening mine ventilation is unique for the following
reasons: (1) it is challenging to keep airflow velocities high enough to effectively remove or dilute airborne con-
taminants, (2) large air volumes can be moved through the mines with little static pressure drop, and (3) stoppings
to direct ventilation airflows are costly to construct and maintain. The research results suggest that by incor-
porating ventilation planning into the mine planning process, using propeller fans, developing new stopping
materials and construction methods, and using long pillars to eliminate crosscuts where possible, the ventilation
of large-opening mines can be significantly improved. The ventilation improvements created by incorporating
these various techniques into the ventilation plan will help reduce the exposure of mine workers to airborne
contaminants in underground large-opening mines.

1 INTRODUCTION – moving adequate ventilation airflow volumes to

Large-opening, room-and-pillar mines (i.e., those
with openings larger than about 92.9 m2 (1000 ft2 ),
or 12.2 m (40 ft) wide and 7.6 m (25 ft) high)
accounted for 135 of the 230 metal/nonmetal mines
operating in the United States in 2004, and they
employed about 2600 underground workers (MSHA
2004). The commodities produced from these large-
opening mines and the corresponding number of
mines include: limestone for aggregate (103), lead/
zinc (15), salt (14), marble (7), lime (6), stone (4),
and sandstone (1) (MSHA 2004). Virtually all large-
opening mining operations use a fleet of diesel-
powered equipment that typically includes several
large haulage trucks, one or more front-end loaders,
drills, scalers, dozers, and various utility vehicles.
A typical diesel-powered fleet frequently approaches
3730 kW (5000 hp), necessitating ventilation airflows
of more than 354 m3 /s (750,000 ft3 /min) to dilute
diesel particulate matter (DPM) and diesel exhaust
gases. A stone production blast will commonly use
454 kg (1000 lb) or more of high explosives and will
liberate about 425 m3 (15,000 ft3 ) of blasting gases and
fumes, thereby imposing additional ventilation airflow
requirements.
2 VENTILATION CHALLENGES IN
LARGE-OPENING MINES
Large-opening mines face ventilation challenges in the
following areas.
dilute airborne contaminants to statutory levels
– controlling and directing the airflow to where it is
needed the most
– planning ventilation systems to integrate with pro-
duction requirements
2.1 Moving adequate ventilation airflow volumes
Large-opening mines pose unique ventilation chal-
lenges simply to move sufficient air quantities for
dilution of all airborne contaminants. In mature
stone mines, the total open-space volume can reach
7,079,200 m3 (250 MMft3 ). With these large open-
space volumes, even fans capable of moving 472 m3 /s
(1.0 MMft3 /min), take hours to complete a total air
change. Therefore, ventilation air should be directed
and coursed to the production face areas to concentrate
and minimize the ventilation airflow volumes neces-
sary to provide sufficient fresh air to the active working
areas.
2.2 Controlling and directing the airflow
The size of large-opening underground stone mines
creates problems for controlling and directing ventila-
tion airflows. Stopping designs and techniques suitable
for use in coal mines generally cannot be applied in
stone mines because of the entry sizes and the associ-
ated cost of construction. Ventilation curtain materials,
such as mine brattice, can be problematic because of
significant air leakage due to its deterioration caused

53

Copyright © 2006 Taylor & Francis Group plc, London, UK

 by repetitive flapping (especially near the main mine
fan) and blast pressures.
2.3 Planning ventilation systems
Stone production from most large-opening, room-
and-pillar underground stone mines comes from mul-
tiple working faces located on the perimeter of the
mine. In the initial mine development stages when
few operating faces are present, it is relatively sim-
ple to provide adequate ventilation airflow across the
production faces. Natural ventilation and an auxiliary
fan may be sufficient to reduce airborne contaminants
to acceptable levels at this stage of mine development.
However, as a mine expands, ventilating all the work-
ing areas becomes more difficult as the number of
working faces and the distances between the main mine
fan and the working faces increase.
3 NIOSH LARGE-OPENING MINE
VENTILATION RESEARCH
NIOSH is conducting a research effort to improve
the ventilation airflow quantity and quality in large-
opening underground mines. Investigations have
focused on a four-point approach:
– Developing techniques to estimate the required air
quantity to dilute airborne contaminants.
– Identifying alternative fans capable of efficiently
moving the required air quantity to dilute airborne
contaminants to statutory levels.
– Developing improved stoppings to direct and con-
trol the ventilation airflow to the production faces.
– Evaluating improved mine designs to deliver and
distribute the required ventilation airflows without
interfering with production requirements.
3.1 The Air Quantity Estimator
Mine ventilation planners must determine the air quan-
tity needed to meet the statutory DPM concentration
limits by taking into account the emissions charac-
teristics of the site-specific fleet of engines at their
operation. To meet this challenge, NIOSH developed
the Air Quantity Estimator (AQE) to assist mine opera-
tors with ventilation planning (Robertson et al. 2004).
The AQE provides an initial estimate of the required
air quantity needed to dilute DPM contaminates to
statutory levels in the main air stream of the mine.
The AQE is a user-friendly, stand-alone computer pro-
gram that uses diesel engine performance test data
from both the Environmental Protection Agency (EPA
2002) and the Mine Safety and Health Administration
(MSHA 2002). The AQE is available upon request to
NIOSH at: http://www.cdc.gov/niosh/mining/products/
analysissoftware.htm#AQE.
54
Copyright © 2006 Taylor & Francis Group plc, London, UK
3.2 Improving ventilation airflow volumes with
propeller fans
Due to the opening size, the resistance to flow in
large-opening drift mines is minimal. Low-pressure
propeller fans can efficiently move large airflow
volumes in such mines as demonstrated by NIOSH
(Grau et al. 2002, 2004, Krog et al. 2004). Grau
(2002, 2004) found that historically, large-opening
drift mines almost exclusively used vane-axial fans for
main mine fans, where in most cases, a better choice
would have been propeller fans. The difference in the
air quantity-pressure loss relationship between a large-
opening mine and a typical coal mine is significant, as
shown in Figure 1. In both cases, the mine resistance
pressure increases as the square of the ventilation air-
flow volume. The required ventilation pressure for an
underground stone drift mine rarely exceeds 249 Pa
(1 in w.g.) (Grau et al. 2004a) since the resistance is
so small.
Figure 1 also shows the fan curves for a typical
high-pressure, vane-axial fan commonly used to ven-
tilate a coal mine and the fan curve for a typical 3.66-m
(12-ft) propeller fan. The operating point for the coal
mine (Point A, Fig. 1), occurs at a flow volume of
118 m3 /s (250,000 ft3 /min) and a static pressure of
1.77 kPa (7.1 in w.g.). The same coal mine vane-axial
fan used in a large-opening mine would operate at
153 m3 /s (325,000 ft3 /min) and 52 Pa (0.21 in w.g.)
(Point B, Fig. 1). However, at that operating point, the
static fan efficiency is less than 6%, well outside the
manufacturer’s normal operating envelope.
The operating point for the 3.66-m (12-ft) propeller
fan (Point C, Fig. 1) occurs at 175 m3 /s (370,000 ft3 /
min) and 67 Pa (0.27 in w.g.) however, the static fan
efficiency is much higher at 14%. Although this effi-
ciency is low for the propeller fan, it is 2.3 times more
effective than the axial-axial fan used in the same mine.
3.3 Improving stoppings for directing ventilation
airflows
Using stoppings to direct ventilation airflows in large-
opening mines has been limited due to their high
capital cost and construction and maintenance diffi-
culty. The most challenging problems exist in older,
extensive mines that have never incorporated stop-
pings into the mining plan. Retrofitting an older mine
with stoppings requires considerable labor, material,
and maintenance costs (Grau 2004). Major design
criteria for stoppings are:
– withstanding fan pressure differentials
– withstanding production blast pressures
– minimizing leakage between the intake and return
air courses
– having high durability and requiring minimal
maintenance
 Fan Flowrate (x 1000 cfm)
0 50 100 150
3.0
3.66 m Propeller
Large Mine Resistance
2.5
2.13 m Vane-Axial
Coal Mine Resistance
Fan Static Pressure (kPa)
2.0
1.5
1.0
0.5
0.0
0 20 40 60
Figure 1. Pressure and quantity curves for two fans and corresponding resistance curves.
The relative importance of these design criteria is
associated with three distinct areas of a mine, the main
entries, intermediate entries, and the face areas. The
stoppings in the main entries will typically see low
blast pressures. However, due to their close proxim-
ity to the main mine fan, they experience the highest
pressure differential and thus have the greatest poten-
tial for leakage. The stoppings in the main entries
must last throughout the life of the mine with lit-
tle maintenance, and thus have the highest durability
requirement. If curtain stoppings are installed close to
the main fan, they must have sufficient durability to
withstand repetitive flapping (Krog et al. 2004). Grau
et al. (2002) measured pressure differentials gener-
ally less than 62 Pa (0.25 in w.g.) at the main mine
fan in large-opening stone drift mines with multiple
entries. Pressures from face production blasts can far
exceed these ventilation pressures and from observa-
tion, depending on the mine layout, stoppings may
be located near face blasts for several years. Mucho
et al. (2001) measured blast pressures from two dif-
ferent production-face shots, ranging from 8.27 kPa
(1.20 psi) to 9.38 kPa (1.36 psi) at distances of 61 m
(200 ft) to 152 m (500 ft) from the shot. Both test shots
were similar to typical production blasts in under-
ground stone mines, and they each consisted of 181 kg
(400 lb) of ANFO, 77 kg (169 lb) of dynamite, and
23 kg (50 lb) of Detagel.
Copyright © 2006 Taylor & Francis Group plc, London, UK
200 250 300 350 400
12
10
Fan Static Pressure (in. w.g.)
8
A
6
4
2
B C
0
80 100 120 140 160 180 200
Fan Flowrate (m3/s)
Stoppings located close to face areas or even in
intermediate areas may require a blast pressure relief
mechanism to prevent damage from production blasts.
Timko and Thimons (1987) discussed a method to pro-
vide blast relief using VELCROTM fasteners on typical
brattice stoppings. NIOSH recently performed tests
where twelve 23-kg (50-lb) sandbags were placed on
the bottom of a curtain stopping that was laying on the
mine floor. It was found that this method held the stop-
ping in place for lower pressure shots, but allowed for
pressure relief during higher pressure shots by allow-
ing the curtain to slip out from under the sandbags.
Experimentation on a site-specific basis may be nec-
essary to determine the appropriate sandbag pressure
relief weight for stoppings in intermediate areas of
the mine.
The most efficient ventilation barrier separating
intake from return air is a long stone pillar. These bar-
riers are created by eliminating at least the last face
shot that would normally break through two adjoin-
ing entries, thus keeping a natural connection between
the pillars. A series of these connected pillars creates
a long stone air wall that is an effective and practi-
cal method for directing ventilation air. This technique
reduces maintenance and the expense of building stop-
pings. The pillars can eventually be mined when the
barrier is no longer needed, such as just prior to mine
closure.
55

Figure 2. Schematic of super stopping.
Figure 3. Schematic of EZ-Up stopping.
Since stone stoppings cannot always be used,
NIOSH has developed and tested two new stopping
types, the “Super Stopping” and the “EZ-Up Stop-
ping” (Figs 2, 3) for use in large-opening mines (Grau
et al. 2006). The Super Stopping is a long-term or
permanent ventilation control structure, and is con-
structed from low density, composite cement and fly
ash blocks measuring 1.22 m (48 in) wide by 1.22 m
(48 in) high by 0.81 m (32 in) thick, and weighing
approximately 544 kg (1200 lb) each. The blocks are
laid and positioned using an extended reach fork lift.
The EZ-Up stopping is a temporary, portable, curtain
stopping which is raised to the mine roof using a winch
system. In-mine tests at NIOSH’s Lake Lynn Labora-
tory have demonstrated the viability of the proposed
construction methods and materials.
3.4 Mine design considerations for improving
ventilation airflows
Three improved mine designs and associated venti-
lation plans have been identified for large-opening
56
Copyright © 2006 Taylor & Francis Group plc, London, UK
mines: perimeter, unit, and split. Selecting the appro-
priate mine design and associated ventilation plan
is a factor of mine size. Grau et al. (2002, 2004),
Mucho et al. (2001), and Krog et al. (2004) dis-
cuss the applications of these proposed mine designs
for improving the ventilation of large-opening mines.
Perimeter ventilation is best suited for older mines
with little previous ventilation planning. New mines
are generally more readily adaptable to split and unit
mine ventilation practices. Achieving an optimal ven-
tilation system layout that works well with production
is relatively easy for a newer operation. However,
retrofitting the new concepts into extensive existing
operations may present unique challenges. NIOSH
research has identified key considerations for select-
ing the best mine design for optimal ventilation while
meeting basic production requirements (Krog et al.
2004). These considerations include:
– maximizing ventilation system efficiencies
– determine the maximum distance between the face
area and a fresh air source while still providing
adequate ventilation
– locating truck haulage routes underground to min-
imize contamination of the fresh air supply
– direct fresh air to the faces along the shortest path
possible
A mine ventilation system can be rated by calcu-
lating ventilation efficiency. Ventilation efficiency is
the percent of useful ventilation air quantity passing a
specific point compared to the total possible air quan-
tity available. When evaluating different designs for
large-opening, room-and-pillar mines, three related
ventilation efficiency measures have been used by
NIOSH:
Where:
– Ed = delivery efficiency, air delivered to the work-
ing area
– Ef = face ventilation efficiency, air available at the
face
– Et = total mine efficiency
– Qm = air quantity entering the mine
– Qlop = air quantity exiting the face area at last
opening at a pillar
– Qf = air quantity at face.
An example of increasing ventilation efficiency by
incorporating long stone pillars in the design of a new
mine is shown in figures 4 and 5. This example is
a variant of a split-mining ventilation system where
a propeller fan blows 143 m3 /s (303,000 ft3 /min) into
 three dedicated intake airways located in the northwest
portion of the mine.
A long stone pillar separates these intake air-
ways from the working areas. Two crosscuts within
Figure 4. Ventilation airflows of new mine with long stone
pillars. (Bold numbers are ventilation airflow volumes, m3 /s,
(x 1000 cfm).
this pillar line have been sealed with permanent stop-
pings as the mine was developed. Of the 143 m3 /s
(303,000 ft3 /min) entering the mine, about 122 m3 /s
(258,000 ft3 /min) ventilated the working area and was
measured passing through the last opening in the stone
pillar, giving a ventilation delivery efficiency of 85%
(see Equation 1).
About 95 m3 /s (202,000 ft3 /min) of the ventilation
airflow entering the working area was measured at the
working faces, resulting in a calculated face efficiency
of 78% (see Equation 2). Total mine efficiency (see
Equation 3), is the working face ventilation airflow
divided by the mine intake volume, or 67% for this
case. Alternatively, total mine efficiency is the product
of delivery efficiency multiplied by the face venti-
lation efficiency (85% × 78% for this example), or
again, 67%.
Separating ventilation efficiency into its component
parts provides a useful way to compare and evaluate
alternative mine ventilation systems layouts for large-
opening underground stone mines. Recent work by
Grau et al. (2004) illustrates the utility of the ven-
tilation efficiency concept for planning a ventilation
system that delivers more air to the working faces.
Figure 6 shows a 15-year-old mine that relied
primarily on natural ventilation supplemented with
three auxiliary fans. The working faces required an

Distance from fans (ft)

0 500 1000 1500 2000 2500
F 800
100%
350
C
Brattice stopping line 700
87% Long stone pillars
300
Future long stone pillars
76% 600
Air Quantity (x 1000 cfm)

Revised Ventilation System Original auxiliary fan

250
Air Quantity (m3/s)

66% 500

200 53%
400
A Long Stone Pillars G 43%
150 100%
88% B 85% 300
Future Long Stone Pillars H
D 33%
100 200
Required Airflow 108 m3/s (230,000 cfm) 83% 81% 80%
50 100
E Original Auxiliary Fan

0 0
0 100 200 300 400 500 600 700 800 900
Distance from fans (m)

Figure 5. Ventilation delivery efficiencies of new mine using long stone pillars, and 15-year-old mine before and after
ventilation improvements.

57

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 6. Initial ventilation airflows for a 15-year-old mine
with partial brattice curtain stopping line and auxiliary fans.
Figure 7. Improved ventilation airflows for a 15-year-old
mine after installation of brattice curtain stopping line and
propeller fans.
air quantity of about 109 m3 /s (230,000 ft3 /min) to
dilute the expected DPM. An air quantity of 330 m3 /s
(700,000 ft3 /min) was measured downwind of the
second auxiliary fan. Unfortunately, only 116 m3 /s
(245,000 ft3 /min) flowed to the turn in the stopping
line (Location D), giving a ventilation efficiency of
less than 35% (116/330) to that location. Less than
24 m3 /s (50,000 ft3 /min) is available at the last stop-
ping (Location E), implying a delivery efficiency of
less than 7% (24/330), and a very low total mine effi-
ciency, as evidenced by air velocities that were too low
to measure at the production faces.
Figure 7 shows the measured ventilation airflows
for the same 15-year-old mine that is shown in Fig-
ure 6 after it has been upgraded with two exhausting
propeller fans (Location F) and 23 additional brat-
tice curtain stoppings. Figure 5 shows the calculated
ventilation efficiencies for both the old auxiliary fan
ventilation system and the improved revised ventila-
tion systems in the 15-year-old mine, as well as the
results of using long stone pillars in the new mine
shown in Figure 4. At distances greater than 91 m
(300 ft) from the fan, the original auxiliary fan ventila-
tion system could not deliver the required air quantity
58
Copyright © 2006 Taylor & Francis Group plc, London, UK
of 108 m3 /s (230,000 ft3 /min) to dilute the expected
DPM in the 15-year-old mine, and the total efficiency
decreased rapidly with increasing distances from the
fan. As indicated on Figures 5 and 7, measured air-
flows for the upgraded ventilation system exceed the
required airflow at the working faces, and uses less fan
energy (Grau et al. 2004). However, total efficiency
beyond about 305 m (1,000 ft) (Location G) from the
propeller fans falls to less than 50% due to stopping
leakage. This upgraded ventilation system was a sig-
nificant improvement from the original auxiliary fan
ventilation system, but as the mine expands, the loss
of ventilation airflow due to leakage will result in
inadequate fresh air being supplied to the working area.
Preliminary research by NIOSH suggests that using
long stone pillars instead of brattice curtain stoppings
will maintain the total efficiency over 60% at distances
greater than 305 m (1000 ft) from the propeller fans
in similar mining situations as shown in Figures 4,
6, and 7. Long stone pillars eliminate the construction
and maintenance costs associated with brattice curtain
stoppings. Long stone pillars also eliminate leakage,
resulting in both an increase in total mine, and deliv-
ery ventilation efficiencies, as illustrated for the new
mine in Figure 5 by the long stone pillar curve. Using
long stone pillars can be a key factor in maintaining
adequate ventilation airflows at increasing distances
from the fan to the active production faces where it is
most critical.
4 SUMMARY
Improvements in ventilation efficiencies can be
achieved in underground large-opening mines by uti-
lizing a systematic four point approach to the problem.
First, it is imperative to estimate the air quantity
required for adequate dilution of airborne contami-
nants. A thorough understanding of the airborne con-
taminant sources such as DPM and fumes from diesel
equipment and blasting is necessary for the selection of
the appropriate fan for the site-specific conditions. An
air quantity estimator has been developed by NIOSH
to perform that task.
Second, appropriate fan selections should be based
upon the most effective method to move the required
ventilation air quantity for proper dilution. Propeller
fans produce large air quantities at a higher mechanical
efficiency than vane-axial fans for the typical low-
pressure conditions of large-opening drift mines, and
with relatively low initial capital and operating cost.
The third component is the use of improved stop-
ping designs to direct and control the ventilation
airflow to where it is needed the most. Long stone
pillars are recommended to reduce air leakage, and
they also eliminate the stopping maintenance issue.
In situations where long stone pillars are not feasible,
 proper stopping construction techniques are vital for
the delivery of a high percentage of the ventilation
airflow to the desired location.
Fourth, improved mine designs should be used to
deliver and distribute the required ventilation airflows
without interfering with production requirements.
When properly addressed, these four key ventilation
factors will ultimately result in effective, practical, and
cost effective increases in the ventilation airflows to
improve the air quality in large-opening mines.
5 DISCLAIMERS
The findings and conclusions in this report are those of
the authors and do not necessarily represent the views
of the National Institute for Occupational Safety and
Health.
Mention of any company or product does not con-
stitute endorsement by NIOSH.
REFERENCES
EPA. 2002. U.S. Environmental Protection Agency. Web site
at http://www.epa.gov/otaq/certdata.htm#largeng.
Grau, III R.H., Mucho, T.P., Robertson, S.B., Smith, A.C. &
Garcia, F. 2002. Practical techniques to improve the air
quality in underground stone mines. In E. De Souza (ed.),
North American/Ninth U.S. Mine Ventilation; Proc.intern.
symp. Kingston, Ontario, Canada, June 8–12, 2002.
Netherlands: Balkema.
Grau III, R.H., Robertson, S.B., Krog, R.B., Chekan, G.J.
& Mucho, T.P. 2004. Raising the bar of ventila-
tion for large-opening stone mines. In: R. Ganguli,
59
Copyright © 2006 Taylor & Francis Group plc, London, UK
S. Bandopadhyay, (eds), Mine ventilation: Proceedings
of the 10th U.S./North American Mine Ventilation Sympo-
sium,Anchorage,AK, May 16–19, 2004: 349–355. Leiden,
Netherlands: Balkema.
Grau III, R.H., Robertson, S.B., Mucho, T.P., Garcia, F. &
Smith, A.C. 2004a. NIOSH ventilation research address-
ing diesel emissions and other air quality issues in non-
metal mines. Transactions 2004 316: 149–158. Littleton,
CO: Society for Mining, Metallurgy, and Exploration, Inc.
Grau III, R.H., Krog, R.B. & Robertson, S.B. 2006. Novel
stopping designs for large-opening metal/nonmetal mines.
Proceedings of the 11th U.S./North American Mine Venti-
lation Symp. University Park, PA, June 5–7, 2006.
Krog, R.B., Grau III, R.H., Mucho, T.P. & Robertson, S.B.
2004. Ventilation planning layouts for large-opening
mines. SME preprint 04–187: 1–9. Littleton, CO: Society
for Mining, Metallurgy, and Exploration, Inc.
MSHA. 2002. Nonpermissible diesel engines approved under
Part 7. U.S. Mine Safety and Health Administration.
http://www.msha.gov/s&hinfo/deslreg/1909a.htm.
MSHA. 2004. Quarterly employment and coal production:
accidents/injuries/illnesses reported to MSHA under 30
CFR Part 50. Denver, CO: U.S. Department of Labor,
Mine Safety and Health Administration, Office of Injury
and Employment Information.
Mucho, T.P., Grau, R.H. & Robertson, S.B. 2001. Practical
mine ventilation. Presentation at the Safety Seminar for
Underground Stone Mines, Louisville, KY, Dec 5.
Robertson, S.B., Grau III, R.H., Dolgos, J.G. & Mucho, T.P.
2004. A computer software program that estimates air
quantity requirements in large-opening stone mines. In:
R. Ganguli, S. Bandopadhyay (eds), Mine ventilation:
Proc. of the 10th U.S./North American Mine Ventila-
tion Symp., Anchorage, AK, May 16–19, 2004: 363-369.
Leiden, Netherlands: Balkema.
Timko, R.J. & Thimons, E.D. 1987. Damage resistant brattice
stoppings in mines with large entries. Eng. Min. J.:188(5).
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Optimization of the Red Lake Mine ventilation system

K.G. Wallace, Jr.

Mine Ventilation Services, Inc.

M. Tessier, M. Pahkala & L. Sletmoen

Goldcorp Inc.

ABSTRACT: The Red Lake Mine is the richest gold mine in North America with over 5 million ounces of
reserves and a 2004 production of over 550,000 ounces of gold at a cash cost of US$92 per ounce. Starting
in 2003, Goldcorp initiated an expansion program at the mine. This expansion included a new 6.55 m (21.5 ft)
diameter hoisting shaft currently under construction. The shaft will be driven to a length of approximately 2,000 m
(6,500 ft). The shaft will be used as the primary intake to the mine. The original ventilation design was to have a
“push/pull” ventilation system. In 2005, a detailed review of the mine expansion program was performed. This
review resulted in a proposal for significant modifications to the mine ventilation system. This paper describes
the proposed changes in the ventilation system designed to optimize this system.

1 INTRODUCTION

Figure 1 shows the location of the Goldcorp Inc. Red

Lake Mine. The mine was started in 1945 with the
first bar of gold poured in 1949. In the early years, the
primary mining method was small open stopes mined
using jacklegs and slushers. The principal access to the
mine was through the No. 1 Shaft. This shaft extended
to approximately 1,050 m (3,450 ft) below surface. At
this depth, the original orebody was determined to con-
tinue, but was offset from the No. 1 shaft location by
approximately 800 m (2,625 ft). A transfer drift was
constructed and an internal No. 2 Shaft was driven
from 23 Level to 37 Level. The mine total depth at this
time was approximately 1,690 m (5,550 ft).
Both shafts are about 2.13 m × 7.62 m (7 ft × 25 ft).
The No. 1 Shaft is equipped with two skips and one
dedicated man and material conveyance. The No. 2
Shaft conveyance incorporates the two skips in com-
bination with the man and material handling. The first
diesel equipment, a small loader, was purchased in
1979. Around this time, the mining method changed
to cut and fill with mobile diesel equipment. How-
ever, given the small material conveyance, it has always
been a challenge to bring mechanical equipment and
materials into the mine.
The original ventilation system consists of intake air
passing through small raises connecting open stopes.
This has the benefit of not requiring air heating during
the cool winter months. Winter temperatures can be as
cool as −40◦ C.
Figure 1. Location of Red Lake Mine.
The existing system also has challenges because
many of the upper mine areas are inaccessible and any
problems in these zones cannot be easily repaired. At
the end of 1994, the Red Lake Mine was a marginal
operation. In 1995, a $7 million exploration program
was initiated that lead to the discovery of a High Grade
Zone (HGZ) of gold mineralization. The HGZ was dis-
covered at a depth and location previously thought to
have no potential. The average reserve grade of this
mineralization is presently 80.6 gpt (2.22 opt). This
discovery made the Red Lake Mine the richest gold
mine in the world and the largest gold mine in Canada.
In April, 2005 Goldcorp Inc. performed a friendly
takeover of Wheaton River Minerals Ltd.
Following the new HGZ discovery, and three years
of production, Goldcorp began a major deepening

61

Copyright © 2006 Taylor & Francis Group plc, London, UK

 program at the mine. This program consisted of con- against measured airflows with the percent error calcu-
structing a new 6.55 m (21.5 ft) hoisting shaft (called lated to be 6%. The basic model was used extensively
the McEwen Shaft). In addition, a 3.66 m (12 ft) return to determine future fan requirements, and ventilation
shaft was proposed from surface to the 38-2 Level. In system upgrades required to support the mining rates
the summer of 2005, a complete review of the proposed predicted for the Red Lake Mine. The thermal rock
mine and ventilation plan was performed. During this temperatures measured in the mine showed the inter-
review several major changes to the design were ana- esting result that the rock is actually cooler than the
lyzed and recommended for inclusion in the design. air on the lower mining horizons measured during the
These changes included the removal of intake booster survey. These results are shown graphically on Fig-
fans, increasing the size and number of exhaust air- ure 2. Recent data collected on the deepest level of
ways, including the primary exhaust raise to surface, the mine confirm the geothermal gradient being used
and modifying the mine plan to allow for a rapid at the mine. Figure 3 shows the predicted temperature
completion of the new McEwen Shaft. gradient in relation to depth. The data collected during
the survey was used to predict the future ventilation
options at the Red Lake Mine.
2 VENTILATION SURVEY

The Goldcorp Red Lake Mine ventilation system 3 ORIGINAL VENTILATION DESIGN
design is based on a ventilation survey performed in
2003. This survey included measurements of airflow, Figure 4 shows the original predicted ventilation
pressure differential, fan performance, and psychro- design for the year 2012 for the Red Lake Mine. For
metric properties. In addition, rock thermal proper- ventilation purposes, the mine was divided into three
ties of the virgin rock temperature, conductivity and Zones. Zone 1 is from 30 to 37 Level, Zone 2 is from 37
thermal diffusivity were measured. to 43 Level, and Zone 3 is from 43 to 47 Level. The year
The survey and rock property measurements are 2012 was used as the maximum ventilation require-
described in a paper given in the Eighth International ments for the mine since each Zone was assumed to
MineVentilation Congress. In summary, the ventilation be in production at this time.
survey resulted in a basic model that was correlated

Rock Temperature in Three Boreholes

Depth into rock (ft)
0.00 5.00 10.00 15.00 20.00 25.00 30.00
20.5 68.7
Rock Temperature (deg. C)

Rock Temperature (deg. F)

20.0 67.7
19.5
66.7
19.0
65.7
18.5
64.7
18.0
17.5 63.7
17.0 62.7
16.5 61.7
0.0 2.0 4.0 6.0 8.0 10.0
Depth into rock (m)
Left Vertical Right

Figure 2. Rock temperature in borehole.

Rock Temperature Against Depth

Temperature (˚C)
10.0 12.0 14.0 16.0 18.0 20.0 22.0
0
-110
Depth Below Collar (ft)

Depth Below Collar (m)

-1000
-515
-2000
-3000 -920
-4000
New Data Point -1325
-5000
-1730
-6000
-7000 -2135
50.0 55.0 60.0 65.0 70.0 75.0
Temperature (˚F)
Figure 4. Original ventilation system design (predicted for
Figure 3. Rock temperature against depth. 2012).

62

Copyright © 2006 Taylor & Francis Group plc, London, UK

 From the results of the various field tests and ven- The airflow requirements for the new design were
tilation survey, and from data provided by the Red based on the following:
Lake Mine engineering team, a set of design criteria
– Zone 1: 70.8 to 84.9 m3 /s (150,000 to 180,000 cfm)
were established. This criteria included typical fric-
– Zone 2: 70.8 m3 /s (150,000 cfm)
tion factors, resistance per length factors, shock losses
– Zone 3: 70.8 m3 /s (150,000 cfm) plus 33.0 m3 /s
associated with raises, rock thermal properties, and
– (70,000 cfm) for shop and trackless haulage.
heat loads from existing stopes. Airflow requirements
were based on providing sufficient airflow for the The Zone 2 and 3 requirements combined to
expected diesel equipment fleet, fixed facilities, leak- give a total volume across 43 Level of 174.6 m3 /s
age and travelways. Based on the original design, the (370,000 cfm). The total predicted mine volume was
airflow requirements were predicted to be 70.8 m3 /s calculated to be 264.2 m3 /s (560,000 cfm) which is an
(150,000 cfm) per Zone with each active sublevel ven- increase of 47.2 m3 /s (100,000 cfm) over the original
tilated with between 14.2 and 18.8 m3 /s (30,000 to ventilation modeling scenarios. To achieve this air-
40,000 cfm). flow distribution and to remove the intake booster fans
Climatic simulations conducted for the original from the system, the preliminary ventilation modeling
mine design showed that in primary airways the pre- showed that the mine required a number of changes.
dicted environmental conditions should be below the These changes included:
design criteria of 26.6◦ C (80◦ F) wet bulb temperature.
1. 5.5 m (18 ft) diameter return air raise (RAR) is
This is possible because of the low rock temperatures.
constructed from surface to the 23 Level.
Nevertheless, there may be areas in stopes that can
2. From 37 to 23 Level two 4.0 m (13 ft) RARs are
exceed this criterion. However, the new design should
driven in parallel and connect to the 5.5 m (18 ft)
have more airflow per stope which will help reduce
RAR on 23 Level.
temperatures in these areas.
3. 43 Level mined at 5.5 m × 5.5 m (18 ft × 18 ft)
A significant benefit is realized by bringing
dimension (to achieve the total flow requirements
un-heated intake air in the existing, stopes in the upper
for Zones 2 and 3).
portion of the mine. The stopes ice up in the win-
4. New return air raise (RAR) internal to Zones 2 and
ter months which results in heating the air in winter
3 at bored 4.0 m (13 ft) diameter. This raise is in
months and cooling the air in summer. For the ven-
parallel with a 2.7 m (9 ft) Alimak RAR raise in
tilation analyses, two scenarios were analyzed. One
Zone 2.
was using the existing surface intake fan to push air
5. Both Zones 2 and 3 have a 2.7 m (9 ft) Alimak fresh
into the mine to 16 Level where it would combine
air raise (FAR). The FAR in Zone 3 will be used as
with intake air in the McEwen Shaft. This results in
a second egress route.
66.1 to 70.8 m3 /s (140,000 to 150,000 cfm) entering
6. 38-2 Level slashed to 5.5 × 5.5 m (18 × 18 ft).
the shaft on 16 Level. Since the intake air in the old
7. Adjust Zone 1 area fans to achieve 42.5 m3 /s
mine workings could be affected by ground falls, a sec-
(90,000 cfm) from 37-1 to 34 Level and 70.8 m3 /s
ond intake scenario was analyzed where only 47.2 m3 /s
(150,000 cfm) from 34 to 30 Levels.
(100,000 cfm) reached the McEwen Shaft on 16 Level.
8. 2.7 m (9 ft) raised borehole from shaft bottom to
38-2 Level.
4 PRELIMINARY MODIFICATIONS TO THE
Figure 5 is a sketch of the bottom levels of the mine
MINE VENTILATION SYSTEM
showing this infrastructure.
The original design required intake booster fans near
the McEwen Shaft on the 43 and 47 Levels, respec-
tively. These fans were required because of the limited
return capacity in the proposed 3.66 m (12 ft) return
raise. During the review of the overall mine plan in
the summer of 2005, Goldcorp engineers decided that
intake booster fans were undesirable from an opera-
tional and ventilation perspective. In addition, the mine
plan was modified to not drive the McEwen Shaft to
the 47 Level but, rather, to the 43 Level. By reducing
the shaft length, the lower mining horizon could be
brought on line sooner. Options will be maintained to
sink the McEwen Shaft to below 47 Level after initiat-
ing production from Zones 2 and 3. Production from
Zone 3 would be achieved by truck haulage up to 43
Level. Figure 5. Preliminary modifications to lower mine.

63

Copyright © 2006 Taylor & Francis Group plc, London, UK

 5 OPTIMIZED VENTILATION SYSTEM
From the initial mine plan, further optimizations were
performed. Many of these changes were designed to
expedite development schedules, to better utilize exist-
ing mine infrastructure for ventilation purposes, or for
geotechnical reasons. The changes include:
1. A 3.05 m × 3.05 m (10 ft × 10 ft) track drift will be
driven from the base of the ramp in the HGZ at 42
Level across to the McEwen Shaft. This drift will
be used to assist in the development of the shaft
and the infrastructure around the shaft location
(shop, loading pocket, etc.) Once the shaft and
ramp are in, the track will be pulled for greater
ventilation area.
2. A 4.6 m × 4.9 m (15 ft × 16 ft) ramp will be driven
at a slight down grade from the 42 Level near the
McEwen Shaft to the 43 Level at the HGZ ramp.
3. Below 43 Level (Zone 3) in the HGZ, the ramp
will be driven at 4.3 m × 4.3 m (14 ft × 14 ft) to
accommodate the truck haulage system below the
43 Level. Exhaust from this area will be through
a 4.0 m × 4.0 m (13 ft × 13 ft) raise.
4. Above 43 Level to 38-2 Level (Zone 2), the intake
air will be provided in the ramp and in the FAR
parallel to the ramp. Twin 3.05 m (10 ft) diameter
exhaust raises will be driven to return air from
Zones 2 and 3.
5. Booster fans will be required on the return on level
43 to pull air up from Zone 3. In addition, a booster
fan is required on the top of the two 3.05 m (10 ft)
raises on 37 Level. For this time phase, each sub-
level connection to the RARs is regulated with the
exception of the lowest level. For this scenario,
the lowest level was taken as 47 Level. The mine
is expected to go deeper, but at this time, it was
assumed that mining would be nearing comple-
tion in the upper part of Zone 2 allowing greater
airflow to Zone 3.
6. The return raise system comprises of one 4.0 m
(13 ft) diameter raise from 37 Level to 31-1 Level.
On 31-1 the air transfers across existing drifts to
the base of another 4.0 m (13 ft) diameter return
raise up to 23 Level. A second 4.0 m (13 ft) diam-
eter raise connects 37-1 Level to 23 Level. From
37-1 to 38-2 Level a 4.0 m × 4.0 m (13 ft × 13 ft)
drop raise is driven to exhaust air from these two
levels.
7. On 38-2 a 1.83 m (6 ft) diameter raise connects
from the shaft bottom area (42 Level). This raise
was initially designed to take air from the conveyor
drift and shop. This design may change as it may
not be possible to place the raise near the conveyor
level.
8. The top of the 5.5 m (18 ft) RAR from 23 Level to
Surface is reduced to 4.4 m (14.5 ft) in diameter for
the last 15 m (50 ft). This reduction was necessary
64
Copyright © 2006 Taylor & Francis Group plc, London, UK
to accommodate construction limitations caused
by an existing reinforced concrete pad.
9. Trucks will be used to haul ore and waste from the
base of Zone 2 and from each sublevel in Zone 3.
To meet existing production requirements, only
two trucks were calculated to be in operation at
any one time.
10. A conveyor drift will be installed near the McEwen
Shaft. A flow of about 9.4 m3 /s (20,000 cfm) is
designed to exhaust directly to the shaft bot-
tom RAR.
11. The shop/fuel bay is moved to near the McEwen
Shaft. The facility is designed for an airflow of
about 14.2 m3 /s (30,000 cfm) exhausted to the
shaft bottom RAR.
Two future time phases were considered for this
study. The first time phase was at the end of 2007. At
this time, Zone 2 is in full operation, but only one
3.05 m (10 ft) RAR is in place to 37 Level. The com-
plete return system above 37 Level is assumed to
be in place. Zone 3 was ventilated with 35.4 m3 /s
(75,000 cfm). This air is required for driving the ramp
and initial sublevel development.
The maximum airflow scenario was assumed to be
around 2009 (and for several years after) when Zone
3 is in full production in addition to Zones 1 and 2.
At this time, all ventilation infrastructure is in place.
Only the 2009 model was run with a limited 47.2 m3 /s
(100,000 cfm) entering the McEwen Shaft at 16 Level.
5.1 Model results
The 2007 model showed that the primary surface
exhaust fan should be capable of pulling air through
the shop raise and across 37 Level without the need of
booster fans. Zone 1 is ventilated with the existing 30
Level booster fan, the 37-1 Level intake booster fans,
one exhaust fan on the 31-826 ramp at the 30 Level
exhaust and a second exhaust fan on the 30-867-1
Ramp. The surface intake and exhaust fans are also
operating at this time. The existing 22 Level booster
fans are removed at this time. The predicted results
of the modeling for 2007 are shown on Figure 6 (in
Imperial Units).
By 2009 it is assumed that the ventilation system is
near full capacity. All Zones are fully ventilated with
the deepest point in the mine development being to 47
Level. At this time, each sublevel in Zones 2 and 3 are
ventilated with 21.2 m3 /s (45,000 cfm). Booster fans
are required on the exhaust on 43 Level and 37 Level
to help pull the air through the sublevels. An exhaust
fan is also required on the shop exhaust raise. Zone 1
is ventilated as described in the 2007 model. Figure 7
depicts the predicted airflow distribution at this time
(in Imperial Units).
The 2007 airflow distribution was developed assum-
ing the maximum airflow available with the existing
 surface intake fan pushing air to 16 Level. Two addi-
tional models were developed which showed the
impact on main fans if only 47.2 m3 /s (100,000 cfm)
were available to 16 Level and if no air was available
to 16 Level. These models did not change the airflow
distribution significantly, but did impact the fan duties.
5.2 Fan requirements
Table 1 shows the predicted air heating and fan operat-
ing costs for the models developed for this study. The
model reflecting a complete loss of 16 Level intake
was included in this table. For this analysis, it was
assumed that the power cost is 8.1 cents per kWhr
and that the propane cost is $0.58 per liquid liter. The
results indicate that using the 16 Level as an intake
saves significant air heating costs, however, limiting
the intake to only 100,000 cfm does not have a large
impact on the operating costs.
Fan duties were determined for each scenario mod-
eled. These results are shown on Table 2. It is important
to note that the fan pressures shown for new fans does
not include internal fan losses, duct work, etc. The

Figure 6. Predicted airflow distribution and fan locations

Figure 7. Predicted airflow distribution and fan locations
for ventilation system at the end of 2007.
for ventilation system at the end of 2009.
Table 1. Summary of analyses.

Number of Total fan Fan op. cost Heating op. Total Cost
Scenario Year fans in mine horsepower* ($/yr)* cost ($/yr)* ($/yr)

Single intake for mine 2006 n/a n/a n/a n/a n/a
through McEwen Shaft 2007 n/a n/a n/a n/a n/a
(assumes dual intake will be 2009 13 3759 $1,929,952 $2,688,550 $4,618,502
available through 2007)
Dual intake to 16 Level then 2006 7 1021 $540,425 $0 $540,425
through McEwen Shaft 2007 10 2091 $1,088,282 $1,827,783 $2,916,066
2009 14 3656 $1,899,886 $2,148,729 $4,048,615
2009* 14 3331 $1,737,134 $2,292,334 $4,029,468

*Assumes 100 kcfm through 16 Level to McEwen Shaft. Electrical power assumed at 8.1 cents/kWhr and propane at
0.58 $/liquid liter.

Difference Between Dual Intake System and no airflow to 16 Level.

Number of Total fan Fan op. cost Heating op. cost Total Cost
Year fans in mine horsepower ($/yr)* ($/yr)* ($/yr)

2009 1 −428 −$30,066 −$396,216 −$426,282

Results indicate that fan operating costs are similar for the dual system, but that air heating costs are
lower. Net result is total cost is less for the dual intake system.

65

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 2. Summary of predicted fan duties. 6 SUMMARY

Year Pressure (in. w.g.) Quantity (kcfm) The Goldcorp Red Lake Mine is undergoing a major
expansion project. Two new shafts are presently under
Shop Conveyor Fan on 6 ft RAR construction. As part of this project, a significant ven-
2006 n/a n/a
tilation study was initiated. This study consisted of
2007 n/a n/a
2009A 4.8 53.6 surveying the existing mine system, measuring in situ
2009B 4.6 49.4 rock properties, establishing a basic ventilation model
of the mine, and using the model to evaluate future
Primary Surface Exhaust Fans
2006 n/a n/a ventilation requirements. In addition to ventilation
2007 14.5 398.3 analyses, climatic simulations were performed to pre-
2009A 14.5 522.7 dict the working environment in the major access
2009B 15.0 526.2 airways in the mine. Throughout the study, a major
43 Level Booster consideration was utilizing the existing infrastruc-
2006 n/a n/a ture as much as possible. The new ventilation system
2007 n/a n/a reflects this by using the existing airways on 31-1, 37,
2009A 0.5 246.4 and 38-2 Levels.
2009B 0.5 253.0 The key results of this study show that by allowing
37 Level Booster (2 fans) about 30% of the summer intake air to course through
2006 n/a n/a the existing old workings, the temperature in the shaft
2007 n/a n/a is predicted to be reduced by 4◦ C when compared to
2009A 6.8 426.4 having all the air entering the McEwen Shaft. This
2009B 6.8 433.0 is a significant reduction and translates to a reduced
37-1 Intake Boosters (2 fans) temperature throughout the mine.
2006 n/a n/a Goldcorp’s Red Lake Mine is a world class opera-
2007 7.3 87.8 tion. A state of the art ventilation system is presently
2009A 9.0 75.2
being incorporated to provide a safe and reliable envi-
2009B 9.1 66.0
ronment for the mine’s workforce. The initial system
30-867-1 Ramp Fan (1 fan) components should be operational in late 2006.
2006 n/a n/a
2007 8.0 105.8
2009A 8.5 104.7 ACKNOWLEDGEMENTS
2009B 9.5 101.9
30 Level Raise Fan (1 fan) The authors wish to thank the management of Gold-
2006 n/a n/a corp Inc. for permission in publishing this paper.
2007 8.0 75.3
2009A 8.5 74.7
2009B 9.5 83.8 REFERENCES
#3 Shaft Air Heater Duckworth I. J. 1999. Rapid Evaluation of Rock Thermal
2006 n/a n/a Parameters at the Lucky Friday Mine. Proceedings of the
2007 0.8 424.2 Eighth US Mine Ventilation Symposium. Rolla: University
2009A 0.8 498.7 of Missouri.
2009B 0.9 532.0 McPherson, M. J. 1993. Subsurface Ventilation and Environ-
Temp Booster 31-1 mental Engineering. Chapman & Hall.
2006 5.4 150.02 Rocque, P. and Sletmoen, L. 2002. Ventilation Design for
Increased Production at the Red Lake Mine. Mine Ventila-
Note: “A” indicates dual intake system to 16 Level and “B” tion, Proceedings to the North American/Ninth US Mine
indicates 100 kcfm to 16 Level with the remaining intake Ventilation Symposium.
air through the McEwen Shaft. Fan pressures shown above Stewart, J. M. 1982. Environmental Engineering in South
indicate delivered static pressure across fan bulkhead or to African Mines. Cape & Transvaal Printers (Pty) Limited.
the collar of the exhaust shaft. It does not include internal fan 495–533.
velocity losses, silencers, duct losses, etc. Wallace, K. G. and Prosser, B. D. 1999. Practical Values
of Friction Factors. Proceedings of the Eighth US Mine
Ventilation Symposium. University of Missouri, Rolla.
Wallace, K. G. 2001. General Operational Characteristics and
Industry Practice of Mine Ventilation Systems. Seventh
cited pressures are the static pressure across the fan International Mine Ventilation Congress. Krakow, Poland.
bulkhead or to the shaft collar. Any additional losses Wallace, K. G., Tessier, M., Pahkala, M., Sletmoen, L. 2005.
associated with the fan and duct need to be added to Ventilation Planning at the Red Lake Mine. Eighth Inter-
these duties. national Mine Ventilation Congress. Brisbane, Australia.

66

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Remote monitoring and automation of a large mine ventilation network

K.R. Mutama & M.A. Meyer

Barrick Goldstrike Mines Inc., Carlin, Nevada, USA

ABSTRACT: When an underground mine ventilation system becomes very large and complex, remote mon-
itoring of key parameters becomes essential to provide first hand knowledge of what is happening in the mine
without any physical presence of a ventilation engineer. The system in use at Barrick Goldstrike Mines Inc.,
Underground Division monitors all aspects of the mine ventilation system. Main fans, booster fans and auxiliary
fans can be monitored and controlled remotely to change the airflow in the mine. The mine is a relatively large
complex spread over more than 1850 m in length. The rock generates oxygen depleting gases such as carbon
dioxide and sulfur dioxide and, therefore, to maintain a healthy underground atmosphere, an extensive monitor-
ing system was required. Air flow, air temperature and gas concentrations are measured with fixed instruments
and displayed on computer screens, and trended for historical data. Alarms are activated automatically for any
abnormal situation.

1 INTRODUCTION exhaust. The haulage drift from the Betze-Post open

The Meikle Mine underground complex, which also
includes Griffin and Rodeo zones, has seen large
expansion since going into production in 1996. A num-
ber of satellite mining areas apart from these major
zones also contribute to the productivity of the Under-
ground Division of Barrick Goldstrike Mines Inc. The
mine is now extensive with a large foot print cover-
ing an area measuring 1850 m by 125 m on the Carlin
Trend in North Eastern Nevada, USA. There are two
major haulage drifts at different elevations connect-
ing the major mining zones. Ramp systems connect
all levels of the mining areas from the highest to
the lowest level. The lowest levels are now approxi-
mately 610 m below surface at an elevation of 1097 m
above sea level. Gold production of this mine has
ranged from 1,000,000 oz/year, when the mine was
new and smaller with very high grade ore, to the
current 650,000 oz/year at lower grades.
2 THE VENTILATION SYSTEM
The original ventilation and cooling system for Meikle
Mine was described by Van der Walt et al (1996),
later by Mutama (2002) and Mutama (2005). The total
amount of airflow exhausted out of the mine is approx-
imately 1081 m3 /s. The mine has four shafts, a surface
borehole and a decline haul drift from the nearby
Betze-Post open pit. Two shafts are used as ventilation
intakes, and two shafts and a borehole are used for
pit enters the Rodeo Mine at the 4100 level and also
acts as a fresh air intake into the mine.
At the original Meikle main zone there are three cen-
trifugal exhaust fans; two 2.74 m 1300 kW on the main
exhaust shaft and a 1.93 m 520 kW fan on a bore-
hole (or vent raise) which together pull 517 m3 /s.
Approximately 1.6 km away at the Rodeo Mine are
two axi-vane 2.74 m 1120 kW 1200 rpm variable speed
fans on an exhaust shaft pulling 564 m3 /s. A schematic
of the ventilation system is shown in Figure 1 and the
extent of the mine can be clearly seen.
In addition to the surface main fans there are numer-
ous booster fans located throughout underground
ranging from 75 to 375 kW of entirely the axi-vane
type. The fan sizes vary from 1067 to 2134 mm. Most
large booster fans operate at a speed of 1200 rpm but
1800 rpm is also common for the smaller fans. The
booster fans are essential in distributing airflow to the
various levels and zones of the mine where the required
quantity dictates the size of the fan.
Vent raises exist on almost every level where some
of these fans are located to move airflow on the main
haulage or footwall drifts. Air flow per level can vary
from 37.6 to 188 m3 /s or more. Variable airflow quan-
tity demand on some levels is achieved by using
controllable pitch fans and variable flow regulators.
The ventilation system is very dynamic in that changes
in conditions in one area affect other parts of the mine,
thus requiring a carefully balanced system to main-
tain safe working conditions throughout the mine in
real time.

67

Copyright © 2006 Taylor & Francis Group plc, London, UK

Copyright © 2006 Taylor & Francis Group plc, London, UK

68

Figure 1. Schematic diagram of the current Meikle and Rodeo Mine Ventilation System.
 The mine keeps an up to date ventilation network
model which is used for long term planning because
new mining horizons are being added continually. The
ventilation network model is also used to predict air-
flow changes in the mine when certain actions are
taken, e.g. main or booster fan operating settings or
regulator positions are changed or simulating tempo-
rary drastic changes to the ventilation system such as
fan failures or open stopes between levels.
Since the mine has become so large and complex
ventilation models alone are insufficient to predict or
inform the engineers concerning the current status of
the mine when changes do occur. For this reason alone
a mine monitoring system was developed to help in the
safety and productivity of the underground division.
3 ORIGINS OF THE VENTILATION
MONITORING SYSTEM
At the Meikle Mine the remote monitoring system is
a valuable tool in determining real time mine environ-
mental conditions. The system monitors airflow, mine
gases, temperature, and status of fans and regulators.
When the Meikle Mine went into production in 1996
it was discovered that some parts of the mine gener-
ated sulfur dioxide (SO2 ) from mining sulfide rock.
At times SO2 concentrations reached over 100 ppm
on the 975 and 1025 levels of the mine, and despite
cooling the ventilation air, in some areas working tem-
peratures reached unsafe or very uncomfortable levels.
As the mine continued to expand it became obvious
that the mine needed a means to monitor conditions
remotely in order to provide information which could
make decision making easier in case of an emergency.
The surface mine cooling system which employs
ammonia as the refrigerant had a very good monitor-
ing system for plant monitoring and control. A reliable
instrumentation system to detect ammonia and mon-
itor air quality entering the mine through the Bulk
Air Cooler (BAC) was installed in the plant room
and at the mine intake shaft. The same monitoring
concept was soon extended underground. The mon-
itoring system is based on a software package known
at the mine as CIMPLICITY (GE Fanuc Automation,
CIMPLICITY® HMI) which monitors all aspects of
mine services.
4 THE MINE VENTILATION MONITORING
SYSTEM
4.1 Control networks and hardware
Several data and control networks are employed at
the mine to interface with various fixed plant, elec-
trical substations, motor control centers and the like,
since no one system can satisfy all requirements. An
69
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 2. Typical four gas monitoring station.
extensive fiber optic network is installed from surface
and throughout the underground to all main levels of
the mine. Several copper or twisted pair networks are
also installed with media converters, routers, etc. to
interface the local devices with the fiber optic network
as needed.
The ventilation monitoring system of choice is the
MSA Ultima Gas Monitor which uses the LonWorks
digital communication network. Configuration of this
network is undertaken using LonMaker for Windows
and Microsoft Visio (packaged with LonMaker).
The main mine control network is based on a com-
bination of Allen-Bradley Data Highway Plus (DH+)
and DeviceNet which interface to numerous Pro-
grammable Logic Controllers (PLCs) throughout the
mine. Fixed equipment such as crushers and loading
pockets, shotcrete plants, backfill plants, refrigeration
plant, mine hoists, air compressors and main exhaust
fans are all controlled by Allen-Bradley or GE PLCs.
Substations, MCCs and MLCs are also monitored
and each switchgear bucket is controlled by PLC for
remote and local control of fans, pumps, etc.
To date there are in excess of 100 PLCs on the
mine control network on surface and underground, 98
gas monitors, airflow and temperature monitors at the
Meikle Mine and 95 gas monitors, airflow and tem-
perature monitors at the Rodeo Mine. A typical four
gas monitoring station using the Ultima Gas Moni-
tors is shown in Figure 2 and includes a Remote Relay
Module with alarm reset button.
4.2 Supervisory monitoring and control software
CIMPLICITY HMI is a robust supervisory monitor-
ing and control system which uses communication
interfaces to collect data from the LonWorks, DH+,
Modbus and Ethernet networks. This data can then be
displayed on the users’ desktop computer, logged to a
database, and used for trending, alarming and report-
ing. Furthermore, a web based viewer called WebView
is also available for remote monitoring across the
company intranet.
 The CIMPLICITY HMI system comprises of five
Runtime Servers – three on surface and two under-
ground – and 23 Viewers in plant-, hoist- and dispatch
control rooms, and mine offices. Up to five concur-
rent users can also access the system using WebView.
The latter is a useful tool for remote monitoring via
the company intranet. Secure remote access allows
troubleshooting of the system to be undertaken dur-
ing off shift periods. This is particularly important
since most employees live one hour away from the
mine site.
5 FAN PERFORMANCE MONITORING
Ventilation fan monitoring is an essential part of the
system. The purpose of monitoring main fans is to
look for any abnormal conditions and to know from a
remote office computer the status of the operating data
at any given time. All five surface fans are monitored
for operating data on the exhaust shaft or vent raise
system. The data comprises mechanical and electri-
cal operating information as well as actual ventilation
data. Data gathered on the main fans includes vibra-
tion, bearing temperature, motor winding temperature,
motor amperage, power consumption, fan static and
total pressure, air exhaust temperature, airflow quan-
tity and inlet vane position or fan operating speed
(whichever is applicable).
Other mines monitor fan performance but the sys-
tem in use at the Meikle Mine is very novel because
of the way the data is displayed on personal comput-
ers. Systems are displayed visually as they would be
seen in the field, as shown in Figure 3 for example.
When a value goes outside the operating range or set
point, an appropriate change of color for that parame-
ter, sometimes followed by an alarm shown as text on
the screen, explains the problem.
The monitoring of the fans ensures that perfor-
mance is maintained and if there is a problem this
can be addressed before the situation deteriorates.
When a fan shuts down unexpectedly it can be brought
back online sooner because the system will report the
problem immediately. Furthermore, the availability of
operating fan data is helpful for ventilation network
modeling.
Underground booster and auxiliary fans are also
monitored from surface as shown in Figure 4. The sys-
tem provides fan status information, whether running,
stopped or failed. A green fan symbol indicates fan
is running and a red symbol indicates a fan fault. If
the symbol shows a white color, the fan is off and
is ready to be started remotely. Fans can be started
remotely from the computer (if set to remote motor)
or from the local electrical mobile load center where
they are connected. Fan amperage provides informa-
tion on how hard the fan is working. If a fan is no
70
Copyright © 2006 Taylor & Francis Group plc, London, UK
longer pulling the same electrical current, this means
the blades are broken or something changed drastically
in the system. The system also enables electrical power
consumption to be monitored. Auxiliary fans that are
not in use where mining is not scheduled can be turned
off thereby saving electrical power.
6 VENTILATION MONITORING
6.1 Airflow monitoring
Monitoring mine airflow on a continuous basis has
become an absolute necessity at the Meikle mine. Key
haulage drifts, main ramps and other locations are
monitored for airflow. The airflow monitor is an ultra-
sonic device using two transducers installed diagonally
in a drift. The difference in ultrasonic signal travel time
from sensor A to sensor B and sensor B to sensor A
(TAB – TBA ) is directly proportional to the air velocity.
When the air flow is zero the sound pulse travel time
is the same. Airflow quantity is then calculated from
the air velocity measured and cross-sectional area at
the survey station. The analogue output signal is inter-
faced with an input module on the LonWorks network.
Where airflow may reverse under certain conditions, a
bidirectional output is configured accordingly.
In order to ensure the accuracy of the remote airflow
monitoring stations the ventilation simulation model is
consulted for mine ventilation conditions at the time. A
partial or full ventilation survey is carried out in order
to confirm data from the model. Survey data and the
model output data is compared with readings from the
fixed instruments. In all cases the required accuracy
should be within 15%. Airflow instruments are cali-
brated during installation and compared with airflow
measurements from a hand held vane anemometer.
Table 1 shows the results from such a survey and com-
pares modeled flows, survey measured flows and fixed
monitor flows.
Real time airflow data including air direction is dis-
played on the CIMPLICITY screens. On many occa-
sions the instruments capture abnormalities occurring
in the mine ventilation system. Low airflows or reverse
flows will be indicated in a red color. Warning flags
and alarms messages are generated appropriately to
notify the user of the fault. In this case anybody look-
ing at the screen will notify the ventilation engineer
or qualified person to investigate why the airflow has
changed drastically.
In many cases when unexpected airflow changes are
observed in major ventilation supply drifts it is because
an important air door is not fully closed, or because a
major booster or main fan has failed. The status of
all air doors and fans in the mine are also monitored.
When an air door is open the icon will show a green
color and when it is shut it will show a white color,
and if partially open it will be indicated by a flashing
 Figure 3. CIMPLICITY Screen for Rodeo Mine Exhaust Fans.
71
Copyright © 2006 Taylor & Francis Group plc, London, UK
Copyright © 2006 Taylor & Francis Group plc, London, UK

72

Figure 4. CIMPLICITY Screen for Meikle Ventilation System showing operating main fans, booster fans and air flows.
 Table 1. Typical comparison of modeled, measured and
continuous monitored airflow quantity.
Modeled Measured Monitor
Location (m3 /s) (m3 /s) (m3 /s)
Rodeo to Upper 127 132 123–142
Griffin
Betze Portal intake 307 302 293–330
Meikle Extension 94 90 80–94
Intake
Figure 5. Real time trending of air flow and air quality data.
yellow color. If the air door has malfunctioned it will
be indicated by a red color.
A dispatcher in the mine control room monitors all
mining systems on the computer screen. In the event of
a fault or system problem, the dispatcher will inform
responsible personnel to check and repair the systems.
A problem that would normally take hours or days to
discover and then solve can be addressed in a very
short time.
Occasionally an airflow measuring device will fail
and wrong airflow data will be shown on the screen. If
mine ventilation is stable but a displayed ventilation
reading appears to be faulty, the ultrasonic trans-
ducer has most likely developed a problem. The most
common issues are transducer misalignment or dam-
aged transducer or transmitter communication cables.
Excessive dust, mud or any other material build up on
the sensors will also cause them to display inaccurate
data, although they are very tolerant to dirt buildup.
All these problems are easily solved with periodic
maintenance and calibration.
An important feature of the monitoring system is
being able to follow historical trends through time,
refer to Figure 5. Air flow and air quality information
is trended and can be reviewed at any time for up to
one year. The ventilation system performance can be
evaluated and important decisions can be made based
on the data provided.
73
Copyright © 2006 Taylor & Francis Group plc, London, UK
Table 2. Summary of gas alarms for monitoring system.
Warning HMI O2 CO SO2
Light U/G Screen (%) (ppm) (ppm) Action
None White 20.8 <34 <2 None
Magenta Orange <19.5 >35 >3 Caution
Red/Horn Red >100 Evacuate
6.2 Air quality monitoring and fire monitoring
Underground safety considerations make it paramount
to monitor air quality and to be able to tell when a fire
occurs. Air dry and wet bulb temperatures are mea-
sured at a few selected stations. The gases monitored
are oxygen, carbon monoxide, carbon dioxide and
sulfur dioxide. Some parts of the underground mine
generate naturally occurring carbon dioxide, sulfur
dioxide and steam from the rock. Mutama (2002)
describes in detail the generation of these gases from
the rock and their impact on mining. When the con-
centration of carbon dioxide rises, the oxygen levels
fall below the TLV of 19.5% in some mining zones
and because of this problem oxygen is monitored in
the affected areas of the mine.
One of the key reasons for expanding the ventilation
monitoring system was for underground fire emergen-
cies. The mine is very large, spread over almost 2000
metres including the Meikle Main Zone, Griffin and
Rodeo areas. Because of the complex nature of the ven-
tilation network it is important to detect a fire before
it spreads throughout the mine.
Monitoring for carbon monoxide gas would be a
good indicator of a fire occurring underground. If the
concentration of carbon monoxide was reported by
more than one sensor in concentrations 100 ppm or
higher, this would be a true indication of a fire. Indeed
this was the main reason for providing extensive cov-
erage of underground carbon monoxide monitoring to
be able to detect a fire before it went out of control
and issue a warning to mining personnel to evacu-
ate the mine immediately or go to refuge chambers.
If there was a fire, the monitoring computer screens
would show carbon monoxide levels continuing to
rise in the mine at which point the mine dispatcher
could release stench gas into the mine in order to alert
underground employees to evacuate. This system, it is
believed would save lives and property by providing
an early warning. Table 2 summarizes some of the gas
alarm levels for the monitoring system.
MSA Ultima Gas Monitors are used for gas detec-
tion, typically in the active mining area or footwall
drifts. Ultima Relay Modules and Remote Displays
allow for remote display and alarming, say at the
entrance to the mining area. A magenta beacon will be
triggered for any gas value above the threshold limit

Figure 6. Alarm station with coloured beacons and horn.
value (TLV) and will flash continuously until the con-
centration falls to a safe level. Toxic gas concentrations
above the short term exposure limit (STEL) are taken
seriously and if it is carbon monoxide the red beacon
will flash and a horn will sound to evacuate personnel.
Figure 6 shows a typical alarm station with red and
magenta beacons and horn.
There have been occasions when the system was
put to the test with various gases. On one such occa-
sion, a mined stope continuously generated SO2 levels
in excess of 50 ppm for a few days. Mining in the
affected mining zone was suspended while remedial
work was done to eliminate or decrease the gas con-
centration. On another occasion, carbon dioxide levels
exceeding 2% in the Griffin zone affected O2 levels
for some time until plans were introduced to correct
the situation (Mutama, 2002). The monitoring system
played an important part in helping to assess the situa-
tion remotely. The system is an invaluable tool during
mine fire drills, and in the event of an actual mine fire,
to assess all aspects of ventilation during the event. It
also can be used during mine rescue involving a fire
to prevent personnel being in harm’s way.
7 SUMMARY
It has been demonstrated over the years at the Meikle
Mine that ventilation monitoring is a valuable tool
in continuously assessing the mine atmosphere. The
mine has become very large and complex such that
remote monitoring of key ventilation parameters is
the only way to obtain first hand knowledge of under-
ground conditions. All aspects of main fan operation
are captured in real time and trended. Underground
74
Copyright © 2006 Taylor & Francis Group plc, London, UK
booster and auxiliary fans are also monitored and can
be started remotely from surface computers.
Throughout the mine, airflow monitoring stations
provide ventilation data that would take days or weeks
for a ventilation engineer alone to obtain. The airflow
data from the system can be compared with measured
data and the mine ventilation computer model.
Naturally occurring pockets of gases such as carbon
dioxide, generation of sulfur dioxide from sulfide rock,
hot water vapors from rock fissures, can quickly create
a dangerous mine atmosphere. With the monitoring
system in place these gases are monitored and can be
prevented from building up to unsafe concentrations.
The major benefit of the monitoring system is an
early warning system for underground fire emergen-
cies by detecting carbon monoxide gas. The system
has proven invaluable during fire drills and mine res-
cue exercises by providing real time status of the mine
atmosphere.
The mine ventilation monitoring system in use at
the Meikle Mine can be compared to systems used in
industrial process plants where every aspect of plant
performance is monitored and visualized.
ACKNOWLEDGEMENTS
The authors would like to thank the Barrick Goldstrike
Mines Inc. management for their support of the CIM-
PLICITY monitoring system. Great thanks are due to
Meikle Mine’s first class instrument technicians Sid
Chambless, Mark Koppe and Don Skinner who spend
much of their time making sure the readings from the
system are accurate and instruments are maintained
and calibrated regularly. Thanks are also due to Ed
Gosling and Paul Smith in the Electrical Department.
REFERENCES
Mutama K.R. 2002. Mitigating the impact on the ventilation
system of strata carbon dioxide, steam and sulfur dioxide
at Barrick Goldstrike’s underground Nevada Mines.
Mutama K.R. 2002. Ventilation at the Barrick Goldstrike
Nevada underground complex.
Mutama K.R. 2005. Assessing new airflow requirements at
the Meikle Mine: A case for an additional main centrifu-
gal fan and the inclusion of evaporative cooling spray
chambers for air flow quality improvement.
Van der Walt J., Pye R., Pieterse, H. & Dionne, L. 1996.
Ventilation and cooling at Barrick’s Meikle underground
gold mine. Mining Engineering, April 1996: 36–39.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Underground mine ventilation planning, heat loads, and diesel equipment

C.A. Rawlins
School of Mining Engineering, University of the Witwatersrand, Johannesburg, Gauteng, South Africa

ABSTRACT: The use of diesel equipment underground is a factor to take cognizance of when the heat load and
other environmental issues form part of the overall planning methodology. Trackless equipment or diesel pow-
ered machinery is an important aspect especially in massive mining operations. Increasing efficiency and their
specific application into deep level mines are of great importance. Trackless equipment re-directs the emphasis
on the surrounding environment and the ambient condition thereof. Remote control of diesel equipment under-
ground can be done, however the surrounding environment should not be neglected with regards to the ambient
temperature in those areas. Ventilation and the cooling of deep underground mines is a necessity and on the list
of current mine design strategies. The usage of diesel equipment, i.e. LHD’s, trucks, etc., or in some instances
hybrid type systems (combinations of conventional and trackless equipment) could place an additional burden
on such mines.
When planning and designing an underground mine, heat loads are included and the long term implications in
the selection of the correct trackless equipment to be used in a mine cannot be overemphasized. This paper will
discuss the implications, design issues, and ventilation aspects including an overall design philosophy.

1 INTRODUCTION sometimes based on historical and or measured data

Mining in South Africa takes place over a large part of
the land and the different areas each have geologically
different rock properties and geothermal gradients.
The international mining community is aware that var-
ious different mineral deposits are mined in S.A.
Deposits such as, Gold, Coal, Diamonds, Platinum,
Chrome, and Base Metal mining, i.e. Copper, Zink,
etc., are included in the variety.
As anywhere in the world, mining operations are
only viable in relation to the economic extraction of
the mineral/s intended. Mining at depth is an envir-
onmental challenge because of the need to control
firstly the heat flow into the mine and secondly man
made/induced conditions, such as equipment (Track-
less and other machinery) use, blasting, etc. In general,
conditions in the ambient enclosed tunnels and pro-
duction areas of a mine, is greatly affected by the rock
properties of the surrounding strata, i.e. VRT (Virgin
Rock Temperature). The mining method, i.e. trackless
systems, further dictates the ventilation strategy that
needs to be applied to counter this additional heat load
contribution.
This paper intends to show a ventilation planning
methodology followed to determine a real mines
requirements.
Planning of a mines ventilation and cooling require-
ments are certainly not new to the fraternity and are
gathered from local and surrounding mines. The plan-
ning methodology described herein takes a system-
atic approach in determining the overall ventilation
requirements for a mine over a 20-year project period.
2 MINE DESCRIPTION
The mine extracts diamonds as its primary mineral. It is
an existing mine.The mine uses the Block Cave mining
method on its current mining horizon. The intention
of this planning strategy followed was to determine
the ventilation requirements for a new mining block
which will be deeper and using diesel machinery to
assist in the extraction of the orebody. Figure 1 shows
the current mine layout.
The current mining depth is 650 mbc (metres below
shaft collar). The mining block (Block C) consists of
the following levels:
– Conveyor and ventilation collector level (650 mbc)
– Extraction level (630 mbc)
– Undercut level (610 mbc)
In the evaluation of the new mining block there were
several mining methods (Block cave, Incline cave,
Front cave, sub-level cave, etc.) studied to determine
each ventilation system requirements. This paper dis-
cusses the block cave mining method to reflect on the

75

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 1. Schematic view of the current mine.
Figure 2. Schematic view of the mine extension.
ventilation planning methodology followed which was
similar for the other mining methods analysed.
The mine was initially an opencast mine whereafter
underground mining commenced for economic min-
ing reasons. There was a Block A and Block B which
was mined with a sub-level cave mining method. For
the initial underground mine, the fresh air intake facil-
ity is the Main Shaft (9 m Ø - lined) with two upcast
shafts (3 m Ø each) to return air to surface via exhaust
fans positioned on top of the upcast shafts.
A new mining block, Block C, was designed below
Block B and included a Block Cave mining method
application. This is the current mining operation from
where 4 million tones per annum (at full production)
will be extracted.
The Block cave mining method of the following
block, Block D, is below Block C and includes a ver-
tical length of about 250 m block height. The lowest
76
Copyright © 2006 Taylor & Francis Group plc, London, UK
operational level (Conveyor level) of Block D will be
at a depth of 930 mbc. Figure 2 shows the inclusion of
the new block into the current mine.
Due to the fact that the mine had to change from a
surface to an underground mining operation, the mine
inherited some challenges that had to be over come.
One such a challenge was, that as the block of ground
gradually sunk deeper into the pit due to the extraction
of the orebody, so did upper levels “daylight”. The ore-
body forms cracks from the open pit (surface) through
to the drift loading positions meaning that upper levels,
as they retreated to the orebody limits, were connected
directly to the open pit. This posed a problem as fresh
air would intake from these points and less from the
main intake shaft. To counter this problem, pressurisa-
tion fans were installed at strategic places within the
mine. The mine is thus over-pressurised by some 15%
thereby “pushing” air out into the open pit rather that
intake fresh air from the pit.
In other deep mines such as gold mines, pressurisa-
tion or booster fans are used underground purely to
overcome the increased pressure drop over the system
and thus reduce leakage.
3 VENTILATION PLANNING SYSTEM
A planning methodology would include the following
steps:
– Determine the mining method and rate of produc-
tion.
– Define acceptable environmental conditions.
– Calculate the heat load and other contaminants such
as gas emissions, dust and fumes.
– Calculate the air and cooling requirements.
– Select the planning methodology or strategy
applicable.
There were three planning phases that the project
adopted namely, (a) the conceptual planning phase, (b)
the pre-feasibility phase, and (c) the feasibility phase.
All mining methods mentioned previously were
taken through the conceptual planning phase where
after two mining methods (Block cave and Incline
cave) were selected for the pre-feasibility stage. Fig-
ure 3 shows the planning phases.
For this new mining block, the levels were
defined as:
– Conveyor level (930 mbc)
– Ventilation collector level (900 mbc)
– Extraction level (880 mbc)
– Undercut level (860 mbc)
The design and planning of this new mining
block was somewhat different due to the fact that this
new block was below the existing mining horizon and
current infrastructure/facilities were to be shared as far
 A systematic approach to planning
Determine the method
& rate of production Conceptual

Pre-feasibility
Define acceptable environmental conditions
Feasibility

Calculate heat increases, gas emissions,

dust & fumes

Calculate air &

cooling requirements

Optimise alternatives
(Ventilation and Refrigeration)

Select the planning methodology/strategy OPT

PTIIM
MUUM
M

Figure 3. Planning phase layout.

Figure 4. Geothermal Gradient for the mine area.
Table 1. Mine design parameters.

Description Value and units Africa. The data at a vertical depth of 850 mbs
indicated a maximum temperature of about 26.5◦ C.
Mining method Block cave A best fit curve equation was constructed for the
Production rate 4 Mtpa appropriate data and the result is:
Drift tunnel dimensions 4×4m
Lateral level intake dimensions 5×6m Y = A + Bx + C/x
Production rate 4 Mtpa where A = 20.3 (Constant or surface VRT),
Reject temperature 27.5◦ C B = 0.0077 (Constant)
LHD machinery 187 kW C = 4.184 × 10−7 (Constant),
Loading trucks 224 kW x = Variable – Depth below surface (m), and
Y = indicative VRT temperature (◦ C)
as possible. To refer to the planning system shown in Thus, for the 930 level (Conveyor belt level), the
Figure 3, the following is given. VRT was calculated to be 27.46◦ C. The VRT is rela-
After the conceptual planning phase, the pre- tively low in comparison to a mine for example in the
feasibility phase was conducted which focused on Pilansberg geological area which would be at around
defined and more detailed information and calcula- 44.4◦ C at the same elevation.
tions, together with simulations and “what if” studies. The VRT gradient was determined and the in-mine
heat load determined to be 6944 kW. This heat load
included heat from LHD’s, trucks, strata, conveyor
4 HEAT LOADS AND VENTILATION belt, people working, electric reticulation and com-
PARAMETERS ponents, secondary blasting, etc.

As mentioned, a mines heat load comes from different

sources where the surrounding rock in a deep mine
would be the primary heat source. Some significant
heat load sources are:
– Auto-compression
– Fissure water
– Surrounding rock, and
– Machinery use.
For this mine design the surrounding rock strata heat
load had to be determined. The VRT was measured and
Figure 4 shows the results.
The thermal conductivity of the dolomite rock
(country rock) is 5 W/m◦ C and multiplying this by the
average thermal gradient yields a value of ∼37 mW/m2
for the geothermal heat flux, which is consistent with
reported measurements from this region in South
5 TRACKLESS EQUIPMENT PARAMETERS
AND QUANTITY REQUIREMENTS
For the trackless LHDs and loading trucks there are
three values to determine. Firstly the quantity of air
required to dilute diesel fumes adequately, the heat
load from an LHD vehicle, and the temperature
required before the LHD vehicle thereby not to exceed
the reject temperature of the mine. The calculations
and procedures followed were shows in previous pub-
lication (Rawlins CA, Phillips HR, 2005). The results
were:
– LHD (187 kW) quantity (10.92 m3 /s)
– LHD heat load (325.48 kW)
– Temperature before LHD (23.7◦ C)

77

Copyright © 2006 Taylor & Francis Group plc, London, UK

 The LHD is equipped with a 187 kW Detroit S50
engine. As mentioned in the publication there are five
parameter to be evaluated and shown again for clarity
purposes.
– The diesel unit exhaust outlet pipe dilution fac-
tor quoted as 0.06 m3 /s/kW (general mean value
dependant on the fuel type and vehicle engine type
used) of diesel rated power at the point of operation.
– The amount of air required satisfying a mining
related re-entry period (10 minutes for secondary
blasting). for secondary blasting in massive mining
production areas, etc.
– Sufficient air to dilute (1 m/s drift velocity) any
flammable gas encountered or to dilute dust, gases,
fibres or diesel particulates that could be cre-
ated during normal mining operations, i.e. drilling,
loading, tramming, etc.
– To satisfy the cooling requirements of workers, a
minimum air velocity of 0.5 m/s is advocated.
– Heat load of the mine, which generally relates to
the mining depth, equipment used, etc.
Table 2. Summary of quantity requirements.
Description Quantity (m3 /s)
Re-entry period 12.8
Air velocity quantity 16
LHD heat load 18.1
Diesel fume contaminants 10.92
General fume dilution 11.22
Air quantity requirements were determined using
the following parameters:
– Re-entry period after a secondary blast (10
minutes).
– Minimum air velocity in drifts (1 m/s).
– Heat load quantity required (27.5◦ C reject
temperature).
– Diesel tailpipe air dilution (0.0584 m3 /s/kW)
– General diesel fume dilution (0.06 m3 /s/kW)
The above air quantity requirements were calculated
and the results were:
Table 2 shows that the heat load quantity calcu-
lated from the LHD vehicles are the greatest, namely
18.1 m3 /s. Allowing for 15% pressurisation and 10%
leakage the total quantity required per drift would be
(18.1 + 15% = 20.15 × 10%) 22.9 m3 /s.
In the quantity allocation from a LHD heat load
perspective, it is important to note that the return air
passes (RAP’s) are in the centre of the drifts (double
sided intake system) and only one LHD is allowed to
operate on one side of the RAP and another on the
opposite side.
Figure 5 shows a typical production level layout
with double intake drifts and the centre RAP in each
drift.
In summary the air quantity required for the under-
ground facility was:
For the quantity allocation per level it was given that
there would be thirteen productive drifts on the extrac-
tion level and nine on the undercut level. There would
be a further five non productive drifts on the produc-
tion level and nine non productive drifts on the under-
cut level. For the non productive drifts an air velocity
of 0.5 m/s was allocated and the quantity per drift
amounted to (4 × 4 × 0.5 + leakage and pressurisa-
tion) 10.1 m3 /s.

6 PRODUCTION BUILD-UP,
AIR REQUIREMENTS, AND
INFRASTRUCTURE REQUIRED

The air quantity for the new Block D has been deter-
mined (705 m3 /s) and illustrates the requirement when
the mine is at full production.
Block D would however go through a build-up
phase to reach the required production. The build-up
Figure 5. Double sided intake drift system. phase entails two specific stages the Block has to pass
in order to establish the mine at maximum production.
The build-up phases were named the following:
Table 3. Summary of underground mine quantity.
– Development phase
Description Quantity (m3 /s) – Pre-production phase
– Full production phase
Undercut area 348.7 – Production decline phase.
Extraction/production area 296.1
Workshops, diesel bays, etc. 60
Overall there were four phases identified over the
Total quantity required 704.8, say 705 m3 /s life of the project. The development phase is the stage
when the declines, intake shafts, lateral intake, upcast

78

Copyright © 2006 Taylor & Francis Group plc, London, UK

 shafts, undercut level, production level, conveyor Here after the pre-production phase is entered where
level, RAP’s, FAP’s (fresh air passes) are established. secondary development together with a production
Naturally, all the infrastructure is not fully established build-up is initiated. More and more draw bells are
before production starts, although a great deal thereof established and the amount of LHD’s increase to cater
is completed (primary development phase). for production requirements. After the full production
stage the production decline phase follows. If a mine
BLOCK C & BLOCK D AIR REQUIREMENTS has no further reserves will the mine closure phase be
1400

1300
entered. If there are however more reserves will the
1200
INTAKE FACILITY Old #
INTAKE ONLY
same process repeat as described before.
1100
BLOCK C & D
AIR REQUIREMENTS
When the infrastructure for the new block is estab-
1000 Old #

900
(Rock hoisting
intake) BLOCK C AIR
REQUIREMENTS RETURN FACILITY
lished and another block is still in operation above
QUANTITY (m3/s)

800
to upper level
the new block, problems can occur. The challenge is
700

600
to provide sufficient air for the development phase
500
MRAP 2
in conjunction with supplying the required air to the
400
MAIN #
block above. Figure 6 shows a typical build-up process
300

200
& Ramp
MRAP 1
BLOCK D AIR
REQUIREMENTS where one block is fully established and the other is in
100 the build-up phase to establish the new block.
0 Development phase Pre-production phase Full production phase
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Production decline
Figure 7 illustrates the different phases of a project.
YEAR
It further shows the phases when two and more blocks
are planned. The different phase periods can vary
Figure 6. Air build-up and facility requirements over time.
depending on a variety of parameters such as the size
of vehicles used, decline length development, Down-
YEAR 2003 2004 2004 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
and-upcast shaft development rates, etc.
Block C Pre-Production Phase Full production Production decline Phase Block worked-out
Phase

Main Development Phase Pre-Production Full production

Production decline Phase
Block D
Phase Phase
7 OTHER PARAMETERS
Pre-Production
Block E Block not existing at the moment Main Development Phase
Phase

In South Africa different diesel fuel types are available.

Figure 7. Phases with different block to be developed. The different fuel type are generally referred to as low

Table 4. Ventilation parameters.

No. Parameters Value Unit

1 Production rate 3–6 Mtpa

2 Ventilation system Push–Pull(1)
3 Intake air velocity (Shaft; Men & Material) 10–12 m/s
4 Intake air velocity (Dedicated intake shaft) 12–15(2) m/s
5 Intake air velocity (Lateral haulage; men & material movement) 4–6 m/s
6 Intake air velocity (Dedicated intake haulage) 10–15 m/s
7 Intake air velocity (Conveyor belt ore transport) 1–3 m/s
8 Return air velocity (Upcast/downcast shaft/s) 13–22 m/s
9 Return air velocity (Return lateral haulage/s) 10–15 m/s
10 Shaft resistance (K-value; men & material system; Lined) 0.02 Nm2 /m4
11 Shaft resistance (K-value; lined/Raise bored) 0.01 Nm2 /m4
12 Intake haulage resistance (K-value; blast tunnel) 0.0158 Nm2 /m4
13 Diesel vehicles 0,06 m3 /s/kW
14 Diesel LHD size (Production level) 187 kW
15 Diesel LHD size (Undercut level) 187(3) kW
16 Minimum air velocity in drifts where diesel vehicles operate 1.0(4) m/s
17 Minimum air velocity in drifts where other work is done (No LHD’s) 0.5 m/s
18 Re-entry periods (Secondary blasting) 10 Minutes
19 Ground passes (2 on either side of the pit) 2 × South side & 2 × North side
20 Tipping/crusher/apron feeder operations Air/dust filtration & reuse of air
21 Workshops/Diesel bays (60 m3 /s) Air into RAP or back into workings
22 Crusher rock blasting (20 m3 /s) Air directly into RAP as required

(1) Ventilation system operates on pressurisation/booster intake fans and main exhaust fans situated on the return
side of the mine. (2) Velocities could be higher although pressure drop would increase (P = RQ2 ; R = resistance,
Q = quantity, P = pressure drop over a system). (3) Loading of “Swell” at the rim tunnel positions on undercut level.
(4) Velocity requirement for dust/fume dilution purposes.

79

Copyright © 2006 Taylor & Francis Group plc, London, UK

 emission and high emission diesel. This in turn refers
to the sulphur content of the diesel fuel. High sulphur
fuel would refer to a sulphur content of about 5000 ppm
(parts per million). Low sulphur fuel would indicate a
sulphur content of around 500 ppm and less.
Most mines in South Africa used high sulphur fuel
due to reasoning that it is less expensive than the low
sulphur fuel. This is true although more air per vehicle
would be required to adequately dilute the exhaust
gases. In an economic study done in 2005 (Rawlins,
2005), the results indicated that fuel price is relatively
insensitive in determining a cost per m3 /s.
From January 2006 is a maximum of 500 ppm
sulphur diesel fuel content available in South Africa,
thereby striving towards a low emission diesel require-
ment. Some fuel providing companies can provide
100 ppm sulphur fuel, although at an increased cost.
Furthermore, the ventilation parameters in gen-
eral and applicable to this Block D is summarised in
Table 4.
The parameters shown in Table 4 were used in the
ventilation design methodology and network simula-
tion models.
8 CONCLUSIONS
The ventilation practitioner requires knowledge of the
mining method in the design and planning of existing
and new mines. The mining method is one parameter
in a range of variables in a mine ventilation planning
philosophy.
This paper aimed to illustrate the planning strat-
egy followed for a diamond mine ventilation system.
Diesel vehicles are an important parameter to recon
with in such a design, especially due to a total track-
less environment. The evaluation of a hybrid system,
i.e. diesel and electric combination, should also form
part of such a design, although due to the planning
time allowed for in this project, it was excluded as an
option. The ventilation planning and design of Palab-
ora Copper mine included such a hybrid system and
ultimately operates as such today.
It is important to audit ventilation planning design
system/s. It is further recommended that an external
independent audit be done whereby all aspects of the
design can be evaluated. The design team was fortu-
nate to have an external international and independent
audit done on the project ventilation design including
the mine design whereby important information sur-
faced indicating alternative and other components to
improve the system.
80
Copyright © 2006 Taylor & Francis Group plc, London, UK
In summary, the planning phases were:
a) Conceptual phase
b) Pre-feasibility phase and
c) Feasibility phase (not complete)
The planning phases are further related to a build-up
strategy namely:
a) Development phase,
b) Pre-production phase,
c) Full production phase, and
d) Production decline phase
As a project progresses from conceptual through to
feasibility study so is there an increase in the detail
information and design required.
The application of all relevant parameters in a mine
design is important thus the need for planning to
be done by team members to ultimately achieve the
most suitable system. The most suitable system is not
always the most economic due to the geology of the
underground orebody, the heat load of the mine, etc.
ACKNOWLEDGEMENTS
The authors wish to acknowledge with gratitude the
support of team members at Anglo American plc and
the mine in the design.
REFERENCES
Anglo Technical Division, Anglo American plc. 2003. Ven-
tilation Design and Simulation Studies (Block Cave &
Incline Cave mining methods).
Mine Environmental Engineering in South African Mines
1989. Handbook of the Mine Ventilation Society of South
Africa (MVSSA).
Rawlins, C.A. 2004. Ventilation Design Strategy For Track-
less Deep Underground Mines. 10thAnnual Mining Diesel
Emissions Conference (MDEC). Canada.
Rawlins, C.A. 2005. Diesel powered vehicles, pollution, and
an economic balance thereof. 11th Annual Mining Diesel
Emissions Conference (MDEC). Canada.
Rawlins C.A., Phillips, H.R. 2005. Underground mine venti-
lation planning & design with regards to heat load and
cooling mechanisms. Eight International Mine Ventila-
tion Conference (IMVS): Chapter 50, p381–189. Brisbane,
Australia.
 2. Diesel emissions control

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

West Virginia diesel regulations: A model for the coal industry

P.C. Thakur
CONSOL Energy Inc.

C. Hamilton
W.V. Coal Association

D. O’Dell
United Mine Workers Association

ABSTRACT: In 1997, the Governor of West Virginia established the W.V. Diesel Equipment Commission
comprising three representatives from the coal industry and three representatives from the United Mine Workers
Association (UMWA). They were responsible for writing safe and cost-effective regulations for the use of diesel
engines in the W.V. coalmines. The state also provided funds for laboratory testing of most commonly used diesel
engines with permissible and non-permissible after-treatment devices. After seven years of data gathering and
analysis of federal and other state’s diesel regulations, the final draft of W.V. Diesel Regulations was unanimously
approved by the Commission in 2004. The proposed regulations were approved by W.V. legislature and they
became immediately effective as title 196, Series 1 of W.V. Underground Mine Laws and Regulations. W.V.
Diesel Regulations advocate the use of clean engines, clean fuels and engineered after- treatment systems to
achieve; (a) an ambient diesel particulate matter concentration of 120 microgram/m3 with nameplate ventilation
and, (b) a carbon monoxide concentration of 100 ppm in the tail pipe. Great emphasis is placed on maintenance
of engines and training of operators of diesel engines. The W.V. Diesel Regulations provide maximum safety for
mine workers in a cost-effective manner and as such, they may serve as a model for the entire U.S. coal industry.

1 INTRODUCTION
Coal is the most abundant and economical energy
resource in the world today. In U.S. coal provides 33
percent of all energy consumed and about 52 percent of
all electrical energy. Sixty countries around the world
mine about 5000 million tons of coal per year at present
but coal production is likely to reach 6500 million tons
in 2020. China, U.S. and India alone are projected to
produce nearly 4000 million tons of coal by 2020. Bar-
ring a major break through, e.g., successful nuclear
(fusion-reaction) technology, it is very likely that coal
will remain a dominant source of energy in the 21st
century.
In U.S. about 45 percent of coal production is real-
ized from underground mining. Diesel engines were
introduced in these mines nearly 35 years ago and their
numbers have steadily increased to more than 3000
at present. The driving force behind this phenomenal
growth is mine safety and productivity. Diesel equip-
ment improves safety by eliminating electric shocks,
(Members of W.V. Diesel Equipment Commission)
fire and explosion hazards from trolley wires and elec-
trical cable equipment. It can also prevent personal
injury by eliminating the need to carry small loads
by hand in the working area. Diesel equipment also
improves ventilation in the face area by ending the
need to regulate air velocity (less than 80 m/s) in trolley
wire entries. The main draw back of using diesel equip-
ment in underground coalmines is the contamination
of mine ventilation air with diesel exhaust emissions.
Major components of diesel exhaust are (a) gaseous
emissions and (b) diesel particulate matter. The
gaseous emissions consist of carbon monoxide, carbon
dioxide, nitrogen oxides, sulfur oxides and unburnt
hydrocarbons. The diesel particulate matter consists
of solid carbon particles, liquid and solid hydrocar-
bons, sulfates and moisture. Threshold limit values
(TLVs) for all gaseous components of diesel exhaust
have been established by the American Conference
of Governmental Industrial Hygienists (ACGIH) and
are incorporated into federal Mine Safety and Health
Administration (MSHA) regulations. Many field stud-
ies confirm that meeting these TLVs for gaseous
components of diesel exhaust has not been a prob-
lem in the U.S. coalmines (Johnson, 1980; Reinbold,

83

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 1. Current DPM standards.
Country Ambient DPM concentration
Canada (Metal Mine) 1.5 mg/m3
Germany 0.2 mg/m3
USA (Metal Mine) 0.16 mg/m3
USA (Coal Mine)
Light Duty out by Engines <5 gm/hr
Heavy Duty out by Engines <2.5 gm/hr
Heavy Duty Permissible Engines <2.5 gm/hr
∗
National Institute for Occupational Safety and Health.
∗∗
An ambient DPM standard for coalmines is not feasible because, as yet, there is no way to distinguish between coalmine
dust and DPM.
1981). However, there are no such standards for diesel
particulate matter (DPM) as yet that are based on
sound epidemiological studies correlating sustained
exposure to DPM and resulting health effects. Most
coal mining countries, therefore, have set DPM stan-
dards that can be technically achieved with the present
state-of-the art. Table 1 shows some of these standards.
2 DPM CONTROL STRATEGY
Realizing that diesel equipment can substantially
improve underground mine safety, The Governor of
West Virginia established a W.V. Diesel Equipment
Commission in July 1997, and provided funds to
investigate how DPM emissions in the mines can be
minimized. Diesel engines most commonly used for
diesel equipment, such as, locomotives, shuttle cars
and personnel transport were studied at the diesel
laboratory of West Virginia University. Various after-
treatment devices that were efficient, small in size
and cost-effective were studied for their collection
efficiencies using the IS08178-1 8-mode test.
Based on these findings and the contents of federal
and other state’s diesel regulations, the Commission
reached a consensus that the best strategy to con-
trol DPM in coalmines and tunneling was to take an
integrated approach as described below.
Similar approaches to control diesel emissions were
also advocated by others (Thakur and Patts, 1999;
Schnackenberg, 2001).
1 Reduce DPM generation: use only MSHA approved
clean engines and EPA approved/ASTM D975
diesel fuels.
2 Collect/combust generated DPM: a properly
designed catalytic converter on all diesel engines
and, if needed, soot filters to minimize DPM
emissions.
84
Copyright © 2006 Taylor & Francis Group plc, London, UK
Analytical techniques
Combustible Respirable Dust,
Sample Collected Gravimetrically
Total Carbon by NIOSH∗ 5040 Method
Sample Collected Gravimetrically
Total Carbon by NIOSH∗ 5040 Method,
Sample Collected Gravimetrically
∗∗
Limits DPM Emissions
3 Dilute DPM: the minimum quantity of ventilation
air provided to each engine would be the MSHA
approval – plate air based on gaseous components
of engine exhaust.
4 Monitor gaseous emissions from the engine peri-
odically: carry out “engine stall test” and check
gaseous emissions, particularly carbon monoxide
emissions which should not exceed 100 ppm in the
tail pipe.
5 Routine maintenance and record keeping.
6 Adequate training for safe operations and proper
maintenance.
Based on the above premises, W.V. diesel regula-
tions were drafted in 2003 and the final draft was
approved unanimously by the Commission in March,
2004. The full text can be found in W.V. Underground
Mining Laws, Rules and Regulations under title 196,
series 1. The main structure of the regulations is
presented in Table 2.
3 HIGHLIGHTS OF DIESEL REGULATIONS
It is not possible here to go into details but following
are the essential highlights of these regulations.
1 Clean Engines: all diesel engines will be MSHA
certified and when tested with maximum fuel/air
ratio, it will not require an MSHA Part 7, approval
plate ventilation rate exceeding 75 CFM/rated
brake horsepower. At this ventilation rate, all
gaseous components in the exhaust are diluted to
their respective TLVs.
2 The diesel fuel will meet the standards of the
most recently approved EPA guidelines for over-
the-road fuel. Additionally, the fuel shall also meet
the ASTM D 975 fuel standards with a flash point
of 100 degrees Fahrenheit or higher at standard
temperature and pressure.
 Table 2. W.V. Diesel Regulation: Sub Section. (c) Where multiple engines are in operation, the
minimum ventilation quantity will be the sum
196-1-2 Definition of MSHA Part 7 approval plate ventilation
196-1-3 Underground use quantities. Air quantity measurements shall be
196-1-4 Diesel-Powered Equipment Package
made at the most down-wind diesel unit in the
196-1-5 Exhaust Emissions Control
196-1-6 Ventilation air split.
196-1-7 Exhaust Gas Monitoring and Control
196-1-8 Fuel Storage Facilities
196-1-9 Transfer of Diesel Fuel 4 DIESEL-POWERED EQUIPMENT PACKAGE
196-1-10 Containers
196-1-11 Fire Suppression for equipment and
APPROVAL PROCESS
Transportation
196-1-12 Fire Suppression for Storage Areas A mining company intending to use diesel equipment
196-1-13 Use of Certain Starting Aids Prohibited in their mines must submit an application to the Direc-
196-1-14 Fueling tor of the W.V. Office of Miner’s Health, Safety and
196-1-15 Fire and Safety Training Training. The director will distribute copies of the
196-1-16 Maintenance application to all six Commissioners who generally
196-1-17 Records meet at least once a month to review and approve
196-1-18 Duties of Operator all applications. A typical application is required to
196-1-19 Scheduled Maintenance
196-1-20 Emissions Monitoring and Control
include the following information:
196-1-21 Diagnostic Testing 1 An inventory sheet listing all diesel-powered equip-
196-1-22 Training and General Requirements ment packages to be used at a given mine. A copy
196-1-23 Equipment-Specific Training of this inventory is also kept at the mine.
196-1-24 Diesel Mechanic Training
196-1-25 Operation of Diesel-Powered Equipment
2 The inventory sheet for each diesel-powered equip-
196-1-26 Diesel Inspectors; Employment; Training ment package is required to provide the following
196-1-27 Diesel Inspector-Training Course information.
(a) Name, address and permit number of the mine.
(b) The name and phone number of the person
responsible for maintenance and the testing of
the diesel equipment.
3 All underground diesel-powered equipment shall (c) Manufacturer, model and serial number of the
include an exhaust emissions control and condition- equipment using the diesel power package.
ing system that has been laboratory tested with the (d) Manufacturer, model and serial number of the
diesel engine using ISO 8178-1 8-mode test and has diesel engine.
resulted in DPM emissions that do not exceed an (e) MSHA 7E approval number.
average concentration of 120 microgram/m3 when (f ) Rated horsepower and rpm.
diluted with the MSHA approval plate ventilation. (g) DPM emission in gram/hr; a copy of the ISO
4 Exhaust emission control system includes, (a) a 8178-1 8-mode test should be attached.
DPM filter with at least 75% efficiency, (b) an (h) Ambient DPM concentration when diluted with
oxidation catalyst that reduces carbon monoxide MSHA-approval plate ventilation. This must
concentration in tail pipe to less than 100 ppm. The not exceed 120 microgram/m3 .
system also reduces the exhaust temperature below (i) Type of fuel used.
320 degrees Fahrenheit. Automatic shut down, a (j) Ventilation rate.
spark arrestor, a flame arrestor, sampling ports and, (k) Manufacturer and model number of the after
on-board performance and diagnostic systems are treatment devices, i.e., the catalytic converter
also required. and the soot filter.
5 Ventilation requirements: (l) Efficiency rating of the after-treatment devices
(a) Adequate ventilation shall be provided to
as provided by MSHA or an acceptable third
ensure that the ambient concentration of
party testing facility.
exhaust gases in the mine atmosphere shall not
(m) Manufacturer, type and composition of a pas-
exceed 35 ppm for carbon monoxide, 25 ppm
sive regeneration system, if used (many ceramic
for nitric oxide and 3 ppm for nitrogen diox-
soot filters are regenerated off-board using
ide. Corrective actions will be taken when the
electrical systems).
gas concentrations reach 75 percent of their
respective TLVs. Most applications, in addition to above mine-
(b) Minimum ventilation required would be the specific information, also include a copy of the reg-
MSHA approval plate ventilation for each ulations sub-sections 196-1-8 to 196-1-27 but this
engine. appears redundant. A statement saying that they will

85

Copyright © 2006 Taylor & Francis Group plc, London, UK

 strictly follow the requirements of the regulations with
regards to the specifics of these sub-sections should
be sufficient. A simpler application form is being
prepared for approval by the Commissioners.
So far nearly sixty diesel-powered equipment have
been approved for use in W.V. underground coalmines.
It is hoped that their numbers will multiply and result
in gradual replacement of all trolley wire equipment in
West Virginia coalmines. Besides West Virginia, many
other states have formulated diesel laws/regulations
to supplement Federal laws. Pennsylvania diesel laws
are very similar to West Virginia diesel regulations.
It is hoped that other states without diesel regulations
can use W.V. diesel regulations as a model to write
their own.
REFERENCES
Johnson, J.H. 1980. An Overview of Monitoring and Con-
trol Methods for Diesel Pollutants in Underground
Mines Using Diesel Equipment. CIB Bulletin. July 1980,
pp 73–87.
86
Copyright © 2006 Taylor & Francis Group plc, London, UK
Reinbold, E.O. 1981.Ambient Pollutant Concentration inTwo
Underground Mines Using Diesel Equipment. Mining
Engineering. 33 (1), pp 57–67.
Schnakenberg, G.H. 2001. Estimate of Technically Feasible
DPM Levels in Underground Metal and non Metal Mines.
Mining Engineering. pp 45–51.
Thakur, P.C. & L.D. Patts. 1999. An Integrated Approach
to Control Diesel Particulate Matter in Underground
Coal Mines. Proceeding of the 8th U.S. Mine Ventilation
Symposium. Ed. J.C. Tien, University of Missouri at Rolla,
Rolla, MO, pp 273–279.
Thakur, P.C et al. 2001. Federal Regulations vs Diesel
Exhaust Control in U.S. Coal Mines. Proceeding of the 9th
U.S. MineVentilation Symposium. Ed. E. DeSouza, Queen’s
University, Kingston, Ontario, Canada, pp 419–422.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Pennsylvania’s diesel equipment program for underground bituminous

coal mines

W. Bookshar
Bureau of Deep Mine Safety, Uniontown, PA

ABSTRACT: The 1961 Bituminous Coal Mine Act banned the use of internal combustion engines without the
express written consent of the Secretary. No internal combustion engines were approved for use in underground
mines. In the early 1990’s the Pennsylvania Bituminous Mine Safety Association Committee developed a plan
to bring the interested parties together to discuss the use of diesels in Pennsylvania. Talks were held between
the Pennsylvania Coal Association PCA, the UMWA and equipment manufacturers. These discussions led to
the involvement of various organizations such as NIOSH, Penn State and other lesser know groups. As a result
of these talks, a law was crafted for the controlled use of diesel-powered equipment in Pennsylvania coalmines.
This law was passed in 1996 as an amendment to the Bituminous Coal Mine Act.
The Draft Federal Law was used as a guideline for the initial Pa law and in fact, the Pa law references the
MSHA law in a number of areas. The Pa law was passed by the legislatures with an enactment date in the future.
Meanwhile, the draft federal law was modified resulting in the requirement of changes to the final approved law.
The changes were developed through the Pa Technical Advisory Committee for immediate implementation.
The Pa diesel program regulates emission standards and utilizes carbon monoxide changes to monitor DPM
emissions. Maintenance and training programs are the backbone of the regulation. Article II established a
technical advisory committee to evaluate issues in the interpretation and implementation of the act.
The current Pa diesel fleet has not been without problems but the TAC has worked with industry and labor
to craft standards that are acceptable to all parties. The number of diesel equipment being used in Pa mines is
growing rapidly.

1 INTRODUCTION emergency firefighting equipment used specifically

Prior to the adoption of Article IIA on December 19,
1996, internal combustion engines were not permitted
underground in bituminous coalmines of Pennsylva-
nia. With the adoption of Article IIA, the Bituminous
Coal Mine Act was modified to allow for depart-
ment approval of complete diesel powered equipment
packages utilizing internal combustion engines.
The use of diesel engines in Pennsylvania requires
the following; the use of approved engine and emission
control systems, in depth training of the workforce, and
maintenance of various systems on a routine basis. A
technical advisory committee was established to evalu-
ate unforeseen issues. Diesel engines must be certified
or approved by MSHA for use underground. The Pa
Department of Environmental Protection, Bureau of
Mine Safety, must also grant approval for the engine
and emission package. The Pa regulations place a high
degree of emphasis on specific training issues related
to diesel equipment and the use of diesel fuel under-
ground. The only exception to these requirements is
for that purpose.
This paper will review the basic concepts and some
of the regulations adopted to permit the use of diesel
engines in underground coalmines of Pennsylvania.
During this review, test requirements, ventilation,
training, and unforeseen issues will also be discussed.
2 DIESEL EQUIPMENT PACKAGE
REQUIREMENTS
The basic diesel package starts with an MSHA
approved diesel engine combined with an exhaust
emission control and conditioning system. Each diesel
engine is required to complete a series of laboratory
emission tests based upon the requirements of 30 CFR
Part 7, Subpart E. The standards for these emission
tests parallel ISO 8178 test requirements. The results
of this test provide a diesel particulate matter (DPM)
concentration and a series of gaseous emission out-
puts. In Pennsylvania, the DPM cannot exceed .12 mg

87

Copyright © 2006 Taylor & Francis Group plc, London, UK

 per cubic meter when diluted by the MSHA approval
plate ventilation rate for that engine.The gaseous emis-
sions are used to determine the ventilation plate for that
particular engine based upon Pennsylvania’s threshold
limits for each gas. A list of MSHA approved engines
can be found on the MSHA website at www.msha.gov.
ISO 8178 is an international emission-testing stand-
ard for non-road diesel engine applications. This test
is a sequence of several steady state operating modes,
each weighted to produce a complete emission test.
The test produces an output of DPM and gaseous
emissions for each of the 8 modes tested. Engines are
tested at three different speeds; rated speed, interme-
diate speed and low idle speed. The rated speed is the
speed at which the manufacturer specifies the rated
engine power and the torque is a percentage of the
maximum available torque at a given engine speed.
The weighted average of DPM is used to determine
whether or not the engine meets the requirement of the
Bituminous Coal Mine Act. The gaseous emissions are
used to determine the ventilation rate for that engine.
An exhaust emission control and conditioning sys-
tem must be included with each engine package. Each
exhaust emission control and conditioning system
should be tested with the diesel engine prior to submit-
ting a diesel powered equipment package for approval.
If the engine and system have not been tested as a unit,
an application can be sent to the Technical Advisory
Committee for evaluation. Sufficient data should be
submitted to prove that the combined diesel engine
and emission control system would meet the .12 mg
per cubic meter diesel particulate standard for Penn-
sylvania. A diesel particulate matter filter is required
as part of the emission control system.
3 GASEOUS VENTILATION RATES
Each specific engine model must have an approval
plate that specifies the minimum volume of air
required to ensure that the exhaust emission levels
do not exceed the exposure limits for NO, NO2 , CO,
and CO2 . The Pennsylvania exposure limits for NO2
and CO are significantly lower than those required by
MSHA, therefore the ventilation rate may be different
than the ventilation rate required by MSHA.
The ambient concentration of exhaust gases in the
mine atmosphere is determined by measuring the air
quality and quantity at each specific piece of equip-
ment. If multiple units are operated on the same split,
the volume of air must equal the sum of all of the venti-
lation plate requirements for each piece of equipment.
The volume of air is measured in by the last piece of
equipment on that split.
When contaminant levels of any of the exhaust gases
exceeds 75% of its exposure limits, changes must be
made to the equipment, ventilation, or mining process
88
Copyright © 2006 Taylor & Francis Group plc, London, UK
to reduce the concentration to an acceptable level. If
the concentration of any of the exhaust gases exceeds
its exposure limits, the diesel equipment operated on
that split must be removed from service.
4 MINE WIDE TRAINING
All training course instructors and all training plans
shall be approved by the department. All underground
employees are required to receive special instruction
relating to fighting fires involving diesel fuel. This
training is to be included in the annual refresher train-
ing or included in the fire drills required under 30 CFR
Part 48.
All miners shall be trained in precautions for safe
and healthful handling and disposal of diesel powered
equipment filters.
Refresher training, separate from that required by
MSHA regulations at 30 CFR Part 48, shall be required
annually.
5 EQUIPMENT SPECIFIC TRAINING
Training shall be conducted in the basics of the oper-
ation of a diesel engine, Federal and State regulations
governing their use, company rules for safe operation,
specific features of each piece of equipment and the
ability to recognize problems and shall be provided
to each equipment operator and the mine health and
safety committee if one exists. This training shall be
designed to bring every operator to a level of good
understanding of diesel equipment operation. Each
operator will be qualified by attending a minimum
eight-hour course, including classroom training on
diesel fundamentals and equipment-specific hands-on
training on the job.
Equipment-specific hands-on orientation training
shall be given in an area of the mine where the
equipment will be operated. This orientation shall be
specific to the type and make of the diesel machine
and shall be presented in small groups.
Upon successful completion of both training ses-
sions, the operator shall be issued a Certificate of
Qualification that qualifies him or her to operate a
specific type of diesel-powered equipment. An opera-
tor may be qualified to operate more than one type
of equipment by completing additional equipment-
specific training covering differences specific to each
additional type of equipment.
6 DIESEL MECHANIC TRAINING
Mechanics who work on diesel powered equip-
ment must undergo extensive training. They must be
 qualified to perform maintenance, repairs, and testing
and be certified by both MSHA and the State. They
must undergo at least 16 hours of training in a program
approved by the department. Engine specific training
must also be included. Subjects include federal and
state regulations, company policies and rules, emission
control system design, on-board engine performance
and maintenance diagnostics, service procedures, fuel
system maintenance, fire protection systems, and
engine shutdown devices.
To be qualified, a diesel mechanic must success-
fully complete a minimum of 16 hours of a training
program approved by the department regarding the
general function, operation, maintenance and testing
of emissions control and conditioning components.
The diesel mechanic must be qualified to perform
these tasks on the specific machines used at the mine
or mines where they are employed.
Additional engine-specific training shall be pro-
vided to diesel mechanics in accordance with a plan
approved by the department.
Annual retraining programs for diesel mechanics
shall be required and approved by the department.
The annual retraining shall include refresher train-
ing as well as new procedure and new technology
training as necessary. Such training shall be separate
from refresher training pursuant to MSHA regulations
at 30 CFR Part 48 (relating to training and retraining
of miners) and electrical training required by MSHA.
7 ROUTINE MAINTENANCE AND TESTING
An initial untreated CO reading establishes the base-
line emission level that all future tests will be evaluated
against. This emission test is completed whenever a
piece of equipment first enters operation.
Routine maintenance checks are planned at inter-
vals not to exceed 100 hours of engine operation. If
any subsequent untreated reading is greater than twice
the baseline value, the equipment must be removed
from service. Additionally, if any treated CO read-
ing exceeds 100 ppm, the equipment must be removed
from service.
8 TECHNICAL ADVISORY COMMITTEE ON
DIESEL-POWERED EQUIPMENT
The regulations established a Technical Advisory
Committee on Diesel-Powered Equipment to advise
the secretary regarding implementation of this arti-
cle and evaluation of alternative technology or
methods for meeting the requirements for diesel-
powered equipment. Any alternative technology or
methods recommended by the advisory committee
and/or approved by the secretary shall not reduce or
89
Copyright © 2006 Taylor & Francis Group plc, London, UK
compromise the level of health and safety protection
afforded by this article. The advisory committee con-
sists of two members, one representing the coal oper-
ators and the other representing the working miners.
Requests for an evaluation by the TAC can come
from various sources. A coal miner, coal mine oper-
ator, diesel-related technology manufacturer or on its
own motion, the TAC shall consider requests for the
use of alternative diesel-related health and safety tech-
nologies with general underground mining industry
application. Site-specific requests may be made by
coalmine operators for use of alternative diesel-related
health and safety technologies. All recommendations
on applications submitted under this subsection shall
be on a mine-by-mine basis.
Any action taken by the secretary to either approve
or reject the use of an alternative technology or method
shall be final and binding and not subject to further
review except where a decision by the secretary may
be deemed to be an abuse of discretion or contrary to
law. If any party affected by a decision of the secre-
tary believes that the decision is an abuse of discretion
or contrary to law, that party may file a petition for
review with the Commonwealth Court in accordance
with Pa.R. A.P. Ch. 15 (relating to judicial review of
governmental determinations). The court, in finding
that any decision made by the secretary is an abuse
of discretion or contrary to law, shall vacate and, if
appropriate, remand the case.
The actions and recommendations of theTAC can be
found at the Mine Safety website on the diesel page at:
http://www.dep.state.pa.us/dep/deputate/minres/dms/
website/diesel/actions.htm
9 ISSUES REFERRED TO THE TECHNICAL
ADVISORY COMMITTEE
The following include a list of many of the items
referred to the TAC for evaluation.
• Engine and emission systems not tested as a unit
• Alternate Test procedures for CO monitoring
• Definition of “Specially Trained” as it relates to fuel
spills
• Ventilation Rates as they refer to MSHA Part 7
• Diesel Particulate Filter requirements as they relate
to “clean and new”
• Use of diesel fuel additives
• Test procedures under load
• Alternate method for pre-operation checks
• Fueling equipment in an intake escapeway
• Use and position of flame arrestors
• Fire fighting equipment
• Diesel record books and emission checks
The Technical Advisory Committee has provided
good practical solutions to a number of issues that were
 Table 1.

Quantity Equipment Engine Emission system Vent rate, cfm

1 Diesel Tractor Deutz MWM 916 DST 3868

2 Rubber Tired Mantrips Deutz MWM 916 DST 3868
8 20 Ton Locomotives Cat 3306 PCNA M85 DST 7316
8 9 Ton Mantrips Cat 3304 PCNA M135 DST 4897
2 5 Man Personnel Carriers Isuzu C240 M141 M141 DST 2305
3 MicroTrax Lister Petter LPU 2 DEC 10-001 1684
1 Mine Rover Isuzu C240MA M141 DST 2305
2 10 Man Personnel Carriers Isuzu C240 M141 DST 2305
4 10 Man Personnel Carriers Isuzu C240 M141 DST 7316
4 20 Ton Locomotives Cummins QSB-C240 5.9 M295 DST 13,350
2 18 Man Personnel Carriers Cummins 4BTA3.9C M310 DST 6500
2 10 Man Personnel Carriers Cummins 4B3.3 M276 DST 3500

not clear when the law was drafted. They have worked
with the industry and regulatory departments to ensure
that the protection afforded by the Act is incor-
porated into their recommendations for individual
issues.
Diesel equipment use in Pa has varied from small
MicroTrax units to large turbo charged locomotives.
Following is a summary of equipment, engine and
emission systems that have been used in Pa (Table 1).
The successful implementation and use of diesel-
powered equipment in Pennsylvania is a result of
the operator’s commitment to maintenance and train-
ing and the Departments’ ability to solve critical
operational issues through the use of the Diesel Tech-
nical Advisory Committee. The implementation of
new technology and newer, cleaner engines will only
enhance the Pennsylvania diesel equipment program
and further reduce diesel emissions in our mines.

90

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

An overview of the MSHA industrial hygiene laboratory

R.A. Haney, J.P. Seiler, M.H. Wesolowski, M. Fenlock & J. Swank

Mine Safety and Health Administration, Pittsburgh Safety and Health Technology Center, Pittsburgh,
Pennsylvania, USA

ABSTRACT: The Mine Safety and Health Administration is responsible for monitoring exposures to airborne
contaminants in the mining industry. Samples requiring laboratory analysis are sent to the Pittsburgh Safety and
Health Technology Laboratory for processing. This laboratory is one of two U.S. Department of Labor accredited
industrial hygiene laboratories.
Automated robotic systems are used to weigh both coal and metal and nonmetal mine dust samples. Coal mine
dust samples are analyzed for quartz by Fourier Transform Infrared spectroscopy (FTIR). Metal and nonmetal
mine (MNM) dust samples are analyzed for quartz by x-ray diffraction. Metal dusts and fumes are analyzed by
Inductively Coupled Plasma Atomic Emission Spectroscopy (ICPAES). Diesel particulate samples are analyzed
for carbon content. Gasses and organic vapors are analyzed by gas chromatography. Fiber samples are analyzed
by Phase Contrast Microscopy and Transmission Electron Microscopy.
This paper reviews the general sampling and analytical procedures, discusses the specific laboratory techniques
used for analysis, and explains the application of the sample results.

1 INTRODUCTION than 50 percent of the contaminant standard, controls

A mine ventilation system serves several major func-
tions. One of the main functions that a mine ventilation
system serves is to deliver fresh air to, and remove
contaminated air, from working areas of a mine. The
amount of contaminant generated dictates the amount
of airflow required to be delivered by the ventilation
system and the type of localized or contaminant cap-
ture system needed. While other contaminant controls
such as water sprays, dust collectors, enclosures, or
work practices are available, ventilation is the pri-
mary control used to reduce exposure and remove
contaminants from the working areas of a mine.
Contaminants vary with the type of mining. Res-
pirable coal mine dust and methane are the primary
concerns in an underground coal mine. Respirable sil-
ica, diesel particulate and diesel exhaust gases are the
primary concerns in underground metal and nonmetal
mines. Respirable coal mine dust and silica are the
primary contaminants found at surface mines. The
Mine Safety and Health Administration (MSHA) is
responsible for monitoring exposures of these airborne
contaminants in the mining industry.
In addition to assessing compliance with the vari-
ous exposure standards, the sample results can be used
to evaluate where contaminant controls are adequate
and where controls have deteriorated and need to be
improved. Typically, if exposures are consistently less
are considered adequate and less frequent contaminant
monitoring is needed. When exposures are between 50
and 80 percent of the standard, consideration for and
planning of control upgrades should be made and
frequent monitoring of contaminant levels is needed.
When contaminant concentrations are greater than 80
percent of the standard, controls should be upgraded
and very frequent monitoring of contaminant levels is
needed.
Industrial hygiene samples collected by MSHA
inspectors and mine operators to assess compliance
with the Mine Safety and Health Act of 1977 are sent to
the Pittsburgh Safety and Health Technology Labora-
tory for processing. The Pittsburgh Safety and Health
Technology Laboratory is one of two U.S. Department
of Labor accredited industrial hygiene laboratories.
The Pittsburgh Laboratory is accredited for dust
weighing, silica, metals, and organics analysis. As
an American Industrial Hygiene Association (AIHA)
accredited Laboratory, the MSHA Laboratory must
have a Laboratory Quality Assurance Manual; must
prepare and follow laboratory procedures for each
analysis performed; must have a quality assurance pro-
gram; and must have a sample tracking system. If
samples are sent to an outside laboratory for analysis,
the outside laboratory must also be accredited.
The purpose of this paper is to review the general
laboratory procedures at the Pittsburgh Laboratory,

91

Copyright © 2006 Taylor & Francis Group plc, London, UK

 discuss the specific laboratory techniques used for
analysis, and explain the reports used to disseminate
the results of the analysis.
2 DUST SAMPLING
Mineral dusts are associated with various respiratory
tract diseases. Coal dust causes coal workers pneu-
moconiosis (CWP). Silica dust causes silicosis, and
other mineral dusts can cause irritation of the upper
respiratory tract. Controls for dust typically include
ventilation for dilution of the contaminant, exhaust
ventilation for removal of the contaminant, and water
sprays for suppression of dust generation. Addition-
ally, for coal mines, the Ventilation Plan contains the
specific dust control parameters used at a mine.
Dust exposures for coal mines are monitored under
two separate programs. The Coal Operator Sampling
Program is administered by the coal mine operators;
and the Coal Inspector Sampling Program is adminis-
tered by MSHA. In both coal programs, dust samples
are collected on pre-weighed filter cassettes manufac-
tured by Mine Safety Appliances (MSA). The samples
are collected at a flow rate of 2.0 Lpm using an MSA
pump and a Dorr-Oliver cyclone as a dust presepara-
tor. Inspector samples are pre-weighed to 0.001 mg by
MSHA. Operator samples are pre-weighed to 0.01 mg
by MSA. For both sampling programs, there are coal
production requirements for a valid sample.
For the coal operator dust sampling program, coal
mine operators collect and submit five respirable dust
samples on each underground mechanized mining
unit (MMU) designated occupation (DO) and one
sample for each surface designated work position
(DWP), every two months (bi-monthly samples). Sam-
ple results are averaged and the results are used to
assess compliance with the dust standard. The dust
standard is 2.0 mg/m3 if the quartz content is less than
5 percent and the standard is 10 mg/m3 /% quartz, if
the quartz content is greater than 5 percent.
For the coal inspector dust sampling program,
MSHA inspectors collect samples on five face occupa-
tions during a shift. These samples are used to assess
compliance with the applicable dust standard, verify
that the dust control parameters are effective, and con-
firm that the proper DO has been assigned. If the
average of these five samples is less than the applicable
standard, and there are no individual samples greater
than the standard, the mining unit is in compliance. If
the average of the five samples exceeds the applicable
standard, a citation for excessive dust is issued and the
operator must make changes to the control parame-
ters and resample. If an individual sample exceeds the
standard, the inspector continues to sample to make
a compliance determination based on the average of
five samples.
92
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 1. Respirable dust sampling train.
Dust exposures in NMN mines are monitored only
by an Inspection Program administered by MSHA.
MNM mine inspectors collect two types of dust sam-
ples; respirable dust samples are analyzed for quartz
and total dust samples are evaluated as nuisance dust.
The samples are collected at a flow rate of 1.7 Lpm nor-
mally using an SKC pump (any pump is permitted) and
a Dorr-Oliver cyclone as a dust preseparator. Total dust
samples are collected without the cyclone. The quartz
standard for metal and nonmetal mines is 10 mg/m3 /(%
quartz + 2). The nuisance dust standard for selected
minerals is 10 mg/m3 . Metal and nonmetal mine oper-
ators are not required to submit dust samples to MSHA
for compliance determinations. All metal and non-
metal compliance determinations are based on a single
sample with an “error factor” to account for sampling
and analytical error.
A blank sample is submitted with all samples col-
lected by both coal and MNM inspectors. A respirable
dust sampling train is shown in Figure 1.
3 DUST WEIGHING
The Pittsburgh laboratory has three Automated Weigh-
ing Systems (AWS) that carry out the weighing of these
respirable dust filters. These systems are contained
in a Class1,000 Clean Room that is environmentally
regulated to maintain a temperature between 20 to 23
degrees C ± 2 degrees, and a Relative Humidity (RH)
of 50% ± 10% RH. Each AWS consists of a Zymark
XP robotic arm, turntable, and system controller, a
Metanetics barcode scanner, two Mettler/Toledo MX-
5 microbalances, and one desktop computer. Each
robotic arm is equipped with a specially designed two
pronged hand capable of picking up a filter cassette
without damaging the delicate foil cone shell. The
turntable on the AWS can accommodate up to 8 sample

Figure 2. Automated weighing system.
racks, each of which holds 75 filters, for a maximum
of 600 filters. A robotic weighing system is shown in
Figure 2.
There are two microbalances used by each system.
The balances weigh the filters to a thousandth of a
milligram. One of the balances is the main balance
which weighs every filter in the batch. The second bal-
ance is the quality control balance that weighs every
twelfth filter in the batch. The weights for this filter are
compared and if the difference is greater than ± 10 µg,
the process is halted, both balances perform an inter-
nal self-calibration, and the system reweighs the last
twelve filters. Each filter is identified by a number
that is also contained in a barcode printed on the filter
itself. The filter is placed over a barcode scanner once
it is removed from the balance and before it is returned
to the sample rack. The filter number and weight read-
ing is then stored in a text file on the desktop PC. The
normal run time is about two to two and a half hours
per rack of 75 filters.
After samples are weighed, sample filters are
checked for sample validity based on their weight
gain. Those having a weight gain less than 0.100 mg
or greater than 1.400 mg are visually examined. A
computer program automatically calculates the sam-
ple filter weights and prints out the filters that meet
the above criteria.
Any filters submitted by coal mine operators or
MSHA coal inspectors showing a weight gain below
0.100 mg are checked for evidence of a dust deposit.
If a uniform deposit is not apparent, then a particle
count must be performed. In order to validate that the
filter was used for sampling, the filter must contain 20
or more particles in four randomly selected 1.0 mm2
fields. The filter is considered not valid if less than 20
particles are counted in the four fields and voided for
Insufficient Dust Observed (IDO).
Filters for both the coal and MNM programs are
checked for the presence of over-sized particles. Any
93
Copyright © 2006 Taylor & Francis Group plc, London, UK
samples that appear on the High Weight Sample Report
(weight gain greater than or equal to 1.400 mg) are
opened and visually examined for any characteris-
tics of an invalid deposit. Laboratory procedure also
requires that any coal filter with a weight gain greater
than 6.000 mg, or MNM filter with a weight gain
greater than 3.000 mg must be examined under the
stereo microscope. A sample containing 20 or more
over-sized particles, regardless of its weight gain, is
voided either CON for contaminated (samples weigh-
ing under 6.000 mg), or OSP for over-sized particles
(samples weighing 6.000 mg or more). Any other pres-
ence of foreign material inside the filter results in
a CON void for that sample. Every coal filter that
is received from mine operators and MSHA inspec-
tors is accompanied by a dust data card that has all of
the information pertaining to that sample listed on it.
These cards include information on the mine identifi-
cation number, date the sample was taken, the entity
and occupation that was sampled, and the certification
number of the person that took the sample. Samples
received for the metal-nonmetal program are accom-
panied by a Request for Laboratory Analysis (RLA)
form. An entire MNM inspector survey can be listed
on one RLA. This data is entered into the Laboratory
Information Management System (LIMS) database,
by the laboratory technicians.
When all of the required analyses are completed,
the weight data and sample information are linked by
cassette number and the results are exported from the
LIMS database and formatted into text files which are
then uploaded via the internet into the MSHA Stan-
dardized Information Systems (MSIS) database. The
database holds all of the mine information and data
results for MSHA. The concentration calculations for
the coal program are done in the MSIS after the data is
uploaded. The concentration calculations for the Metal
and nonmetal program are done in the LIMS prior to
data export. Mine operators can obtain certain results,
information and latest news via the date retrieval link
on MSHA’s public website, www.MSHA.gov. Table 1
shows the number of dust and other samples processed
by the Pittsburgh Laboratory.
4 COAL QUARTZ ANALYSIS BY FTIR
Inspector coal mine dust with sufficient weight gains
are sent to the quartz lab for analysis. Inspector dust
samples weighing 0.200 mg or more collected on the
designated occupation (DO), a roof bolter, a desig-
nated area (DA) or designated work position (DWP), a
non-designated entity or a Part 90 miner constitute the
majority of the samples analyzed. Samples weighing
0.100 mg from a select few surface occupations with
historically high potential quartz exposures are also
analyzed.
 Table 1. Summary of annual analytical analysis conducted Instruments are calibrated using suspensions of
by the MSHA pittsburgh laboratory. NIST Standard Reference Material 1878a for quartz
and Hydrite UF for kaolinite deposited on vinyl acrylic
Number copolymer membrane filters. Response factors are
processed
derived for peak area measurements using zero inter-
Type of sample annually
cept linear regression. Samples are analyzed in sets
Dust Weighing of 25.
Inspector Coal Respirable Dust 50,000 When the DO quartz is greater than 5 percent, a
Operator Coal Respirable Dust 40,000 dust standard based on quartz is calculated for the min-
Inspector MNM Respirable Dust 17,000 ing section. This standard is 10 mg/m3 divided by the
Inspector MNM Total Dust 800 percent quartz. If this standard is not protective for
Quartz other occupations, a DA may be established for those
Coal Analysis (FTIR) 15,000 occupations with a standard based on quartz content.
Metal and Non Metal Analysis (XRD) 8,000
DPM Analysis (Carbon) 1,000
Fumes and Metal Dust Analysis (ICP) 800
Fibers and Asbestos Analysis (PCM/TEM) 200 5 METAL AND NON METAL QUARTZ
Organic Vapor Analysis 160 ANALYSIS BY X-RAY DIFFRACTION
Gas Analysis 12,000
After weighing, all metal and nonmetal mine respirable
dust samples weighing 0.100 mg or more are ana-
lyzed for quartz; if cristobalite analysis is requested
this threshold is reduced to 0.050 mg. Dust sam-
ples are prepared for analysis by low temperature
ashing (to remove the PVC filter matrix and carbona-
ceous minerals), ultrasonic suspension of the resultant
residue in isopropyl alcohol, and redeposition of the
sample slurry, using machined funnels and sample
holders designed to center the deposit on a silver
membrane filter.
Because of the wide variation in mineral dust com-
positions seen in the metal and nonmetal mining
Figure 3. FTIR Scan. industry, x-ray diffraction is used for the analysis of
enforcement samples. The MSHA x-ray equipment
Dust samples are prepared for analysis by low tem- utilizes radiation emitted by a copper target x-ray tube
perature ashing (to remove the PVC filter matrix and operated at 1.75 KW. The x-ray beam is focused on the
carbonaceous minerals), ultrasonic suspension of the respirable dust sample deposit on the silver membrane
resultant residue in isopropyl alcohol, and redeposition filter. The crystalline materials in the sample reflect
of the sample slurry, using machined funnels and sam- the radiation, producing diffraction maxima at angles
ple holders designed to center the deposit on a DM-450 which are a function of the interplanar spacing of the
filter. Low temperature ashing is used because it does sample crystallites. The reflected radiation is passed
not alter the quartz content of the sample. through a graphite monochromator to limit fluores-
MSHA utilizes Fourier Transform Infrared Spec- cence effects, and detected with an ionized gas point
troscopy (FTIR) to analyze respirable coal mine counter.
dust samples for quartz. After ashing, the filters are MSHA’s Method P-2 is functionally similar to the
mounted in customized sample holders which are NIOSH Method 7500. Instruments are calibrated using
designed to optimize the alignment of the sample in a range of standards prepared from suspensions of
the infrared beam. The samples are scanned across the NIST Standard Reference Materials. Calibration fac-
frequency range of 1000 cm−1 to 700 cm−1 . An incan- tors are derived by zero intercept linear regression.
descent broad band source emits a beam of infrared An external standard (Permaquartz) is used to correct
radiation. Sample spectra are ratioed to that obtained for long term instrument drift. A silver peak produced
from a blank. An example of an FTIR quartz scan is from the analytical filter serves as an internal standard
shown in Figure 3. to correct for sample self absorption. Sample spinners
Though a number of minerals could potentially are used to minimize preferred orientation effects and
interfere with this measurement, only kaolinite is theta compensating divergence slits are employed so
known to be present in American coal mine dust. An that the irradiated sample area is constant through-
absorbance ratio technique is used to compensate for out the scanning range. Goniometers are arrayed in a
the kaolinite interference. horizontal theta/two-theta configuration.

94

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 4. Primary angle XRD scan.
A preliminary screening analysis is performed
where samples are scanned across a range which
includes the (1,0,1) reflection for quartz and the (2,0,0)
reflection for silver. Peak areas are determined through
profile fitting, and the mass and percentage of quartz,
and a preliminaryThreshold LimitValue (TLV) are cal-
culated. An example of the primary quartz angle scan
is shown in Figure 4. If a potential overexposure is
indicated, a confirmatory analysis is performed using
the (1,0,0), (1,1,2), and (2,1,1) reflections. The lowest
result of the first three peaks is used for enforcement
calculations, with the degree of agreement among the
results and the analyte peak location used to assess the
validity of the analysis. A similar rationale is applied
for cristobalite analysis.
6 DIESEL PARTICULATE MATTER
ANALYSIS (DPM)
In underground MNM mines where diesel equipment
is used it is necessary to have proper ventilation to
ensure that the health risk of workers using this type
of equipment is kept to a minimum. Diesel exhaust is a
very complex mixture containing organic compounds
and elemental carbon. The interim exposure limit for
MNM mines is 308EC µg/m3 . MSHA is in the process
of establishing a final DPM exposure limit.
The controls for DPM include ventilation, low
emission engines, environmental cabs, alternative
fuels, diesel particulate filters and work practices.
From engine test data, a Particulate Index (PI) can be
determined for an engine. The PI is the amount of
airflow required to dilute the whole diesel particulate
emissions to 1000 µg/m3 . A 1000 µg/m3 concentra-
tion of whole particulate is equivalent to 616 EC µg/m3 .
An airflow of double the PI would dilute the engine
emissions to 308EC µg/m3 .
DPM samples are collected using Dorr-Oliver
cyclones and sub-micron impactors manufactured by
SKC. The cyclone and impactor eliminate interference
of larger particles. When a sample is collected at a flow
rate of 1.7 Lpm the impactor cut point is approximately
95
Copyright © 2006 Taylor & Francis Group plc, London, UK
Respirable and Total Dust Diesel Particulate
Metal Dust and Fumes Fibers
Figure 5. Dust sampling cassettes.
0.9 micrometers. The filter cassette used contains tan-
dem quartz fibers filters which are analyzed by the
laboratory for organic and elemental carbon using
NIOSH Method 5040. The first filter is the sample
filter. The second filter in the cassette serves as the
control filter. The filter area is 8.04 square centime-
ters. Figure 5 shows the sample filter cassette for a
DPM sample.
The NIOSH 5040 method is a thermal-optical
method which involves heating the sample to various
temperatures in different atmospheres and monitor-
ing the transmittance of the sample though the use
of a laser. A segment of the sample is cut using a
punch and placed in a sample oven.The analysis occurs
in two stages. In the first stage organic and carbon-
ate carbon are evolved in a helium atmosphere where
temperatures are increased step-wise to about 850◦ C.
The evolved carbon is catalytically oxidized to carbon
dioxide in a bed of granular manganese dioxide, and
then reduced to methane in a nickel-firebrick reduc-
tion column. Methane is then quantified by a flame
ionization detector. A portion of the organic carbon
is pyrolyzed in this first stage, resulting in a reduced
sample transmittance as monitored by a laser.
In the second stage, the oven temperature is reduced,
an oxygen-helium mix is introduced, and the temper-
ature is stepped to about 900◦ C. As oxygen enters the
oven, pyrolytically generated elemental carbon is oxi-
dized and a concurrent increase in filter transmittance
occurs. The carbon evolved until the transmittance
returns to the baseline value established at the begin-
ning of the first stage is deemed to be organic carbon.
 Table 2. Elements and compounds analyzed by ICP and
associated TLV.

Elements and TLV Elements and TLV

compounds (mg/m3 ) compounds (mg/m3 )

Calcium Oxide 5 Iron Oxide 10

Silver Dust 0.01 Lead 0.15
and Fumes
Aluminum Oxide 10 Magnesium 10
Oxide
Arsenic 0.5 Manganese 5
Beryllium 0.002 Molybdenum 10
Cadmium 0.1 Nickel 1
Chromium 1 Titanium Oxide 10
Cobalt 0.1 Vanadium 0.05
Copper 0.1 Zinc Oxide 5

by Inductively Coupled Plasma (ICP). To perform an

Figure 6. Thermogram from DPM Analysis.
Any carbon oxidized after this is considered elemental
carbon. Figure 6 shows the thermogram from a DPM
analysis.
The analytical results are blank corrected and mul-
tiplied by the filter area to obtain the elemental and
organic carbon content of the sample. The final results
are calculated as both organic and elemental carbon
which can be added together to obtain total carbon.
7 METAL DUST AND FUMES ANALYSIS
BY ICP
Toxic metal dusts and fumes may be present in the min-
ing environment.The primary ones of concern are lead,
silver, calcium oxide and welding fumes. The exposure
limits that are the basis of MSHA enforcement activi-
ties are listed in the 1972 TLV handbook for coal mines
and in the 1973 TLV handbook for metal and nonmetal
mines. The primary control for these contaminants is
ventilation.
The laboratory processes approximately 1000 metal
dust and fume samples per year. Samples are collected
on Mixed Cellulose Ester (MCE) filters using a total
dust (no cyclone) sampling configuration. A metal
dust and fume sampling cassette is shown in Figure 5.
Samples are collected at a flow rate of 1.7 Lpm for
metal mines and 2.0 Lpm for coal mines and analyzed
ICP analysis, the sample filter is first digested in
acid(s). The digestion process dissolves the filter and
the various mineral compounds of interest.
The acid solution is introduced by a peristaltic
pump into the ICP where it becomes nebulized into an
aerosol (very fine mist). This aerosol then enters into
the torch area (high temperature plasma) where the gas
temperature is about 6800 K. Desolvation occurs and
the particles are vaporized and dissociated into atoms.
The atoms are then either excited and/or ionized and
emit radiation at wavelengths specific to the elements
of interest as they transit through lower temperature
zones in the plasma. An eschelle polychromator sepa-
rates the complex spectrum into discrete wave lengths
which fall onto a planar array of photoreceptors on the
surface of a silicon chip. The electronic signals pro-
duced are proportional to the intensity of the incident
radiation. Aqueous solutions of NIST-traceable inor-
ganic compounds are used for instrument calibration.
Table 2 shows the elements and compounds typically
analyzed by ICP and their associated TLV.
The final step in an ICP analysis is to convert the
results back to the contaminant compound in question.
If the contaminant is an element such as lead or silver,
no conversion is required. However, if the contaminant
is a compound, such as calcium oxide (CaO) or fer-
ric oxide (Fe2 O3 ) a stoichiometric multiplier must be
applied to the measured elemental value. The stoichio-
metric multiplier is the ratio of the atomic weight of the
compound to the atomic weight of the element. For cal-
cium oxide the stoichiometric multiplier is 1.399 and
for ferric oxide the stoichiometric multiplier is 1.43.
8 FIBERS AND ASBESTOS ANALYSIS
There are many fibrous minerals associated with min-
ing activities. However, there are only six specific
minerals that are regulated as asbestos. These six

96

Copyright © 2006 Taylor & Francis Group plc, London, UK

 minerals are: chrysotile, amosite (cummingtonite/
gunerite), crocidolite, anthophylite asbestos, tremo-
lite asbestos and actinolite asbestos. Typically mining
operations avoid crushing and processing these min-
erals when they are present in the ore body. They are
removed, placed in an abandoned part of the quarry and
covered to prevent contamination of the work place.
In addition to removal, controls for asbestos include
water and ventilation. The current MSHA exposure
for asbestos minerals is 2.0 fibers per cubic centime-
ter of air sampled (f/cc). The current OSHA exposure
limit is 0.1 f/cc.
Fiber samples are collected on black, 25-mm cel-
lulose ester membrane cassettes with a 50-mm con-
ductive extension cowl. The filters have a 0.8 µm pore
size. A fiber sampling cassette is shown in Figure 5.
When sampling in the presence of mineral dust, the
sampler must take care not to overload the sampler
with mineral dust that can occlude the fibers in the
sample. The sample should be collected for at least 2
hours to get enough sample volume for an accurate
analysis. Fiber samples can be collected at either 1.7
or 2.0 Lpm.
Samples are received by the MSHA laboratory and
then sent to contract laboratories for analysis. The sam-
ples are first analyzed by phase contrast microscopy
(PCM, NIOSH Method 7400) for fibers. Mineral par-
ticles greater than 5 micrometers in length with an
aspect ratio greater than 3:1 are counted as fibers. If
the fiber count exceeds 0.10 f/cc, the sample is then
analyzed by transmission electron microscopy (TEM,
NIOSH Method 7402) to identify the fibers. The per-
cent asbestos byTEM is then multiplied times the PCM
fiber count to get the concentration of asbestos in f/cc.
Polarized light microscopy (PLM) is used to analyze
bulk samples for asbestos.
9 ORGANIC ANALYSIS BY GAS
CHROMATOGRAPHY
Organic vapors are found where solvents are being
used to degrease equipment and in some coal prepara-
tion processes. Increasing ventilation typically dilutes
and removes the organic vapors. Typical hazards asso-
ciated with exposure to organic vapors include irrita-
tion to the respiratory tract, compromise of the central
nervous system and in some cases cancer.
Organic samples are collected on either passive
badges or on charcoal tubes. The passive organic
vapor monitors (OVM) are opened and organic vapors
are collected by diffusion on charcoal wafers located
inside the badge. Charcoal tubes also can be used to
collect organic vapors. They are used in conjunction
with a pump to draw air through the tube. Typical
pump flow for a charcoal tube ranges from 0.150 to
0.250 Lpm. These samples are submitted by MSHA
97
Copyright © 2006 Taylor & Francis Group plc, London, UK
Table 3. Measured Organic Contaminants.
Hexane Propyl acetate Ethyl benzene
Octane Trichloroethylene p-xylene
Acetone 4-methyl-2-penanone m-xylene
1,1,1- Tetrachloroethylene Dibromethane
trichloroethane
Nonane Chloroform o-xylene
2-butanone Toluene Mesitylene
Benzene 1,2-dichloroethane Cyclohexanone
inspectors and technical experts. Typically MSHA
analyzes a suite of 21 organic compounds which
include the most common solvents found in mining
operations. These compounds are listed in Table 3.
Organic sample analysis utilizes gas chro-
matographs equipped with flame ionization and mass
selective detectors. Solvents are used to extract air
contaminants from the sample media. An aliquot of
the extract is injected onto the capillary column and
a preliminary result is calculated based on flame
ionization detection. Any sample that contains a con-
taminant whose measured level approaches or exceeds
the permissible exposure limit is analyzed using the
gas chromatograph equipped with a mass selective
detector to confirm the identity of the contaminant.
10 MINE GAS ANALYSIS BY GAS
CHROMATOGRAPHY
The Gas Laboratory has four benchtop gas chro-
matographs equipped with thermal conductivity detec-
tors to analyze mine air samples for asphyxiants and
flammable gases. Mine air is required to contain not
less than 19.5 percent oxygen, and not more than 5000
ppm of carbon dioxide or 50 ppm of carbon monoxide.
Limits for other noxious or poisonous gases are given
in the 1972TLV’s for coal mines and the 1973TLV’s for
metal and nonmetal mines. When the methane concen-
tration in the work place exceeds 1 percent, changes
in the ventilation must be made.
This laboratory also provides analyses of mine air
samples from fires and/or explosions. These analy-
ses are used to determine the status of an underground
mine fire. The trend in concentration of carbon monox-
ide and various ratios of mine gases help to establish
whether a fire is increasing or decreasing in size,
and whether inerting agents are impacting a fire
area. Methane and oxygen relationships establish the
explosiveness of a mine atmosphere.
Gas samples are collected in evacuated containers.
A complete analysis can be conducted on samples col-
lected in 50-cc bottles or BD 60-cc syringes with a
luer-lok tip. A methane, oxygen and carbon dioxide
analysis can be made on samples collected in 10-cc
vacutainers.
 Table 4. Mine gases measured. exposures to harmful or toxic airborne contaminants.
Meeting the requirements for AIHA certification
Gas Analysis range assures that sample analysis is performed in accor-
dance with proper laboratory procedures and that the
Oxygen 0.01–21% accuracy of sample analysis is maintained. The pri-
Nitrogen 0.01–100%
mary focus of the laboratory is mine dusts and gases.
Carbon dioxide 0.03–20%
Carbon monoxide * 10 ppm–10.0%
Methane 0.01–100%
Ethane 5 ppm–0.5% REFERENCES
Ethylene 5 ppm–0.5%
Acetylene 10 ppm–20.0% American Conference of Governmental Industrial Hygien-
Hydrogen * 5 ppm–5.0% ists. TLV’s Threshold Limit Values for Chemical Sub-
stances in Workroom Air, 1972 and 1973.
* Range from reduction gas analyzer. Federal Register 30 CFR, July 1. 2005.
MSHA, 1999. Metal and Nonmetal Health Inspection Proce-
dures Handbook.
MSHA. 2002. Organic Vapors in Air collected on Charcoal
Nine gases, listed in Table 4, are determined by Tubes, Method P30, October 2002.
chromatography and one gas, Argon, is calculated. In MSHA. 2004a. Gravimetric Determination of Mine Dust
addition, reduction gas analyzers are used to measure Collected on Tared Filter Cassettes Method P19, July
low concentrations of carbon monoxide and hydro- 2004.
gen. The argon concentration is calculated based upon MSHA. 2004b. Infrared Determination of Quartz in Res-
the nitrogen concentration. Approximately 870 air pirable Coal Mine Dust, Method P7, January 2004.
samples are analyzed annually in the laboratory. MSHA. 2004c. Xray Diffraction Determination of Quartz
and Cristobalite in Respirable Mine Dust Method P2, July
MSHA also maintains a Mobile Gas Laboratory that 2004.
responds to mine fires and/or explosions. Air samples MSHA. 2005. Metal and Trace Metal Analysis by Inductively
are analyzed on–site at these events. An average of Coupled Plasma-Atomic Emission Spectroscopy, Method
more than 11,000 samples were analyzed at mine sites P3, February 2005.
from 2001 to 2005. NIOSH. 1994. Manual of Analytical Methods, Asbestos and
Other Fibers by PCM, Method 7400.
NIOSH. 1994. Manual of Analytical Methods, Asbestos by
SUMMARY TEM, Method 7402.
NIOSH. 1999. Manual of Analytical Methods, Elemental
Carbon (Diesel Particulate) Method 5040, September
The MSHA analytical laboratory supports a range
1999.
of health programs designed to protect workers from

98

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Instrumentation for diesel particulate matter emissions research

S.E. Mischler, A.D. Bugarski & J.D. Noll

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

ABSTRACT: Measurement and sampling of diesel particulate matter in mine air presents a challenge due
to the complexity of the diesel aerosol. The results of a series of tests carried out by the National Institute for
Occupation Safety and Health, Pittsburgh Research Laboratory, in an underground mine were used to evaluate
several measurement methods that are currently used to characterize particulate matter emitted by diesel-powered
equipment. This paper presents an overview of these techniques and examples of the results along with a
discussion of the advantages and information gained by each technique. The measurement methods discussed
include particle size measurements using a scanning mobility particle sizer, total particulate matter measurements
using a tapered elemental oscillating microbalance, a size selective sampling method for workplace and personal
exposure assessment and a size-selective high volume sampling method for carbon analysis.

1 INTRODUCTION same study, a scanning mobility particle sizer (SMPS)

Measurement and sampling of diesel particulate mat-
ter (DPM) presents a challenge due to the complex
nature of diesel aerosols. There are a number of
methods for DPM measurements and each method’s
relevance is dependent upon the information required.
For instance, the Environmental Protection Agency
(EPA) defines DPM as the mass of material collected
on a filter at a temperature of 52◦ C or less after dilu-
tion of exhaust with air (EPA, 2002). The EPA is
interested in defining the concentrations of DPM in
ambient air. The Mine Safety and Health Administra-
tion (MSHA) currently regulates (MSHA 2005) the
exposure of underground metal/nonmetal miners to
DPM (MSHA, 2005) based upon the mass based con-
centration of elemental carbon (EC). EC was found
to be a suitable surrogate for DPM in underground
metal/nonmetal mines, since diesel engines are prac-
tically the only source of submicron EC particles in
this occupational environment (MSHA, 2005). Other
measurement methods for DPM in occupational envi-
ronments have also included, respirable particulates
(<3.5 µm), smoking corrected respirable particulates
and combustible respirable particulates among other
methods (EPA, 2002).
The mass measurement methods discussed above all
collect the DPM on a filter, which is then sent to a labo-
ratory for mass determination. There are also methods
available for almost real-time DPM measurements.
Jimriska et al. (2004) used a tapered element oscillat-
ing microbalance (TEOM) to continuously measure
PM2.5 during a study of diesel bus emissions. In this
was also used to measure semicontinuous size distri-
bution and particle number concentration of the DPM.
Kittleson (2004) used a photoemission aerosol sen-
sor (PAS) to measure diesel particulate emissions.
These measurements were made when the engine was
operated at five steady-state conditions. The author
compared the PAS measurements with results from a
SMPS, a diffusion charger and a condensation particle
counter.
In this paper, the results of a series of compre-
hensive tests carried out by the National Institute for
Occupation Safety and Health, Pittsburgh Research
Laboratory, in an isolated zone of an underground
metal/nonmetal mine, will be used to evaluate sev-
eral sampling and measurement methods that are able
to characterize particulate matter emitted by diesel-
powered equipment. The results from measurements
using a SMPS, and TEOM as well as results from a
high volume and low volume filter method will be
discussed and compared.
2 METHODOLOGY
These tests were conducted in an isolated zone, a
long underground mine entry ventilated by fresh air
(see Figure 1). Isolated zone tests were designed to
be a compromise between the genuineness of in-situ
measurements of concentrations and the repeatability
and accuracy of the emission measurements obtained
under research laboratory conditions. These tests
allowed the operation of vehicles under conditions and

99

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 1. Isolated zone.

over duty cycles that closely mimic actual production

duty cycles in an area that was not contaminated by
emissions from other vehicles as would occur in tests
conducted in real production areas. In addition, arti-
facts usually generated under laboratory conditions
while attempting to simulate real-life conditions and
processes do not compromise the results of isolated
zone tests. A detailed description of the isolated zone
methodology can be found in Bugarski et al. (2006a
and 2006b).
Simultaneous measurements using each of the sam-
pling methodologies described above were used to
measure DPM concentrations downstream of the iso-
lated zone (see Figure 1) as the vehicle was operated
over a simulated duty cycle between the two load/dump
points. The results obtained from tests conducted for Figure 2. High-volume sampling train.
two types of control technologies, water-fuel emul-
sions and diesel particulate filters, were used for this sampling times are short. Therefore, this method was
analysis. not used to collect samples during the tests on the
filtration systems.
3 INSTRUMENTATION

A description of various methods used in this study
to collect particulate samples and directly measure
concentrations of particulates is given below.
3.1 Standard sampling of DPM for carbon analysis
A standard sampling method (SSM) similar to the one
used by the Mine Safety and Health Administration
(MSHA) for DPM compliance monitoring [MSHA
2005] was used to collect DPM samples for carbon
analysis. All samples were collected in triplicate.
The standard sampling method uses a sampling train
consisting of a flow controlled pump, a 10 mm Dorr-
Oliver cyclone, and an SKC DPM cassette from SKC,
Inc., Eighty-Four, PA.The SKC DPM cassette contains
a single stage impactor and two stacked 37 mm diam-
eter tissue quartz fiber filters. The pumps are operated
at 1.7 l/min. At this flow rate, only aerosols with an
aerodynamic diameter (D50 ) smaller than 0.820 µm
reached the collection filter (Olson 2001).
The 1.7 l/min sampling flow rate used by this
method is inadequate for obtaining DPM samples in
cases where the concentration of DPM is low and
3.2 High-volume (HV) method for sampling DPM
for carbon analysis
A HV sampling method was used to collect DPM
samples. This high volume sampling train, described
in detail elsewhere (Bugarski 2006a), was used to
enhance the collection of samples for carbon analysis
by increasing the sampling flow rate and decreasing the
collection area of the filter. A schematic of the high-
volume sampling train is presented in Figure 2. The
sampling flow rate was increased by merging flows
from five preclassifiers, each consisting of a 10-mm
Dorr-Oliver cyclone followed by a U.S. Bureau of
Mines (USBM) single stage diesel impactor, into a sin-
gle stream. A flow rate of approximately 2.0 l/min was
maintained through each cyclone and impactor pair.
At this sampling flow rate only particles with geomet-
ric mean smaller then 0.775 µm were deposited on the
filters (Olson 2001).
All five preclassifiers were attached to a sym-
metrical plenum that distributed a total flow rate of
approximately 10.0 l/min uniformly among the five
streams. Each of the preclassifier assemblies was con-
nected to the plenum chamber by a 1-m long section

100

Copyright © 2006 Taylor & Francis Group plc, London, UK

 of electrically conductive tubing. The outlet of the
plenum was directly connected to a stainless steel
25 mm diameter filter holder containing two stacked
25 mm tissue quartz fiber filters.
The total sampling mass flow rates were main-
tained using a mass flow controller in the each of the
three sampling lines from the 25 mm diameter filter
holder. The total volumetric flow rates through each
of the sampling streams were measured periodically
by inserting a bubble flow meter inline, between fil-
ter holders and mass flow controllers. The volumetric
flow rates measured during this study were all cor-
rected to ambient conditions. A high volume rotary
vane pump was used to draw the sample though the
filter.
3.3 DPM concentration measurements with a
Tapered Element Oscillating Microbalance
(TEOM) series 1400a ambient particulate
monitor
The TEOM 1400a (Rupprecht & Patashnick Co.,
Albany, NY) draws air through a filter at a constant
flow rate, while continuously measuring the mass
accumulating on the filter and calculating near real-
time mass concentrations. The sample stream is drawn
through this filter from a hollow tapered element that
is vibrated at a precisely controlled frequency and
connected to the suction side of the sampling sys-
tem. The vibration frequency of the tapered element
decreases as particulate mass collected on the filter
increases. By frequently measuring the tapered ele-
ment frequency, the TEOM calculates the increase in
mass of the sample that has accumulated on the filter.
The concentration of TPM can be calculated by divid-
ing the accumulated mass by the volume of airflow
across the filter during the time period over which the
frequency change is measured.
The flow through the instrument is maintained at
a constant rate by a mass flow controller. The flow
is corrected for temperature and barometric pressure.
The internal instrument temperatures are controlled
to minimize the effects of ambient temperatures. To
prevent condensation and ensure that the sample filter
always collects particulates under similar conditions,
the TEOM intake is heated to maintain the sampling
stream at 50◦ C.
During this study, the flow rate on the TEOM was
set at 1.7 l/min. A cyclone and impactor were used as
preclassifiers to the TEOM, allowing only particles
with an average aerodynamic diameter (D50 ) smaller
than 0.820 µm to reach the collection filter. The aver-
age ambient concentrations of TPM were recorded and
saved every 10 seconds. The reported average concen-
trations for a test were obtained from the difference in
filter masses recorded at the same start and stop times
used by the particulate samples for carbon analysis.
101
Copyright © 2006 Taylor & Francis Group plc, London, UK
3.4 Measurement of size distribution and particle
numbers using a scanning mobility particle
sizers (SMPS)
An SMPS was used to periodically measure size dis-
tribution and number of particles in the range between
10 and 392 nm. The SMPS consisted of an electrostatic
classifier (TSI Inc. Shoreview, MN, Model 3080L) and
a Condensation Particle Counter (TSI, Model 3025A).
The flows of monodispersed aerosol were main-
tained at 0.6 l/min throughout the study. At the estab-
lished polydispersed aerosol flow rate of 6.0 l/min, the
inlet impactor had a cut-off point of 460 nm. The con-
densation particle counter was operated in high-flow
mode to minimize diffusion losses. The samples were
collected using 90-second up-scans and 15-second
down-scans. The instrument was operated using a
dedicated laptop computer and Aerosol Instrument
Manager Software from TSI Inc.
The effects of the tested control technologies on
size distribution and count concentrations of aerosols
in mine air were assessed based on the measurements
conducted while the vehicle was performing the por-
tion of the duty cycle at a point nearest to the location
of the SMPS.
4 DISCUSSION OF RESULTS
In this section the instrument results from the isolated
zone tests will be presented along with a discussion
on the applicability of each type of data. The results
presented are from tests when the vehicle was outfitted
with several diesel particulate filter systems and when
the vehicle was fueled with two blends of a water-fuel
emulsion.
Figure 3 presents the results from the HV, SSM
and TEOM sampling methods observed during the
baseline (#1 diesel/muffler) test as well as tests with
cold-weather (C-W) and warm-weather (W-W) water-
fuel emulsions. As discussed previously, the TEOM
results shown here are an average over the sampling
period for the HV and SSM methods.
This figure shows that the HV and SSM meth-
ods measured similar EC concentrations. The TEOM
results show concentrations that are higher then the
HV or SSM methods. The additional mass concen-
trations can be attributed to other DPM compounds
such as organic carbon, metals, and sulfates that are
not measured by the EC analysis. This additional mass
may also contain any particulate, with an aerodynamic
diameter smaller than 0. 820 µm, generated during the
transfer of ore/waste rock during the test.
Figure 4 presents the results from the HV and
TEOM sampling methods observed during the base-
line test as well as tests when the vehicle was outfitted
with diesel particulate filters.
 500 700

TPM Mass Concentration [␮g/cm^3]

600
400 388
500
Concentrations [µg/m^3]

400
300 275 300
262

207 200
200
100
129 0
90 99 0 1000 2000 3000 4000 5000 6000
100
Time [s]
40 40
#1 Diesel/Muffler C-W/Muffler W-W/Muffler
0
#1 Diesel/Muffler C-W/Muffler W-W/Muffler
Figure 5. TEOM semi-continuous data during the water-
EC HV TPM TEOM EC SSM
fuel emulsion tests.

Figure 3. HV, SSM and TEOM results for water-fuel emul-

sion tests. 350

TPM Mass Concentration [ug/cm^3]

300
200 250
180
180
200
Concentration (␮g/m^3)

160
140 150

120 100
105
100 50
80
62 0
60 51 0 1000 2000 3000 4000 5000 6000 7000 8000
43
40 31 Time [s]

20 Muffler Filter 1 Filter 2 Filter 3

8 9
0
Muffler Filter 1 Filter 2 Filter 3 Figure 6. TEOM semi-continuous data during filter tests.
EC HV TPM TEOM

The semi-continuous data also aids in the detection of

Figure 4. HV and TEOM results for filter tests.
The lack of SSM data on Figure 4 suggests one
advantage of the HV method over the SSM. Previous
isolated zone tests using vehicles outfitted with diesel
particulate filters showed that the SSM was not able to
collect enough EC for accurate analysis. However, the
increased collection volume of the HV method allowed
for successful measurement of these lower EC concen-
trations. In these tests the TEOM DPM measurements
were again higher than the HV EC concentrations.
The TEOM was able to effectively measure the DPM
concentrations during the filter tests.
Figure 5 presents the semi-continuous (one-minute
average) data from the TEOM during the water-fuel
emulsion tests. Kelly and Morgan (2002) compared the
real-time TEOM measurements with the filter based
EPA method and found the ability of the TEOM to
measure the transient values of DPM during testing
to be a major advantage. Likewise, having a semi-
continuous stream of data during these tests is an
advantage because this extra data allows for a more
thorough inspection of the test data. Evaluation of
the semi-continuous data presented in Figure 5 shows
the repeatability of the duty cycles during these tests.
potential problems during the test. For example if at
3000 seconds the TEOM data showed a major spike
or decrease in concentrations, than that is evidence
of a potential problem with the test. Data from the
HV or SSM methods would not obviously show that a
problem occurred.
Figure 6 shows the semi-continuous TEOM data
collected during the filter tests. Again this figure
shows the repeatability of each duty cycle. This data
also shows an interesting aspect of the test using Fil-
ter 2. This data shows the DPM concentrations steadily
decreasing during the test, indicating that the filter
efficiency was improving over the duration of the
test. This important piece of information would be
impossible to discern using only the HV or SSM
results.
The results presented thus far are from mass based
measurement methods. However, many researchers
have measured DPM using other metrics including
particle number concentration or surface area. The
advantages and disadvantages of these different met-
rics for measurement of DPM is currently an area of
much research (Kittleson, 2004). One instrument used
during the current study to measure particle number
concentration was the SMPS.

102

Copyright © 2006 Taylor & Francis Group plc, London, UK

 1.0E+07 Table 1. Summary of reviewed measurement methods.
dN/dlog(D_p) [#/cm^3]

1.0E+06
Method
1.0E+05
(Measurement Consideration/
1.0E+04 metric) Applications limitation
1.0E+03

1.0E+02
SSM (1) Personal (1) Long sampling time
(Mass con- sampling. may be required at low
1.0E+01 centration) (2) Time integrated DPM concentrations
10 100 1000
D_p [nm] sampling. (<80 ug/m3 ).
#1 Diesel with Muffler C-W with Muffler W-W with Muffler
(2) Delayed results.
HV (1) Time-limited (1) Occupational
Figure 7. SMPS results from the water-fuel emulsion tests. (Mass con- sampling at low sampling not possible.
centration) DPM con- (2) Delayed results.
centrations
1.0E+07
(< 80 ug/m3 ).
1.0E+06 (2) Time integrated
dN/dlog(D_p) [#/cm^3]

1.0E+05
sampling.
1.0E+04
TEOM (1) Near real-time (1) Occupational
(Mass con- sampling. sampling not possible.
1.0E+03 centration) (2) Time integrated
1.0E+02 sampling.
1.0E+01 SMPS (1) Particle size (1) Occupational
10 100 1000 (Size distri- distribution sampling not possible.
D_p [nm] bution and sampling. (2) Time integrated
Muffler Filter 1 Filter 2 Filter 3 particle (2) Real-time sampling not possible.
number con- sampling.
Figure 8. SMPS results from the filter tests. centration)

Figure 7 shows the effects of the two water-fuel

emulsions on the aerosol size distributions of the DPM. different. The size distribution of the particles mea-
The size distributions of the particles observed sured during the test with the Filter 2 was characterized
during the tests with water-fuel emulsions are char- by much lower concentrations in the larger particle
acterized by lower geometric mean diameter (GMD) size range. On the contrary, the size distributions of
and higher peak concentrations than the size distribu- the particles during the Filter 3 test had significantly
tions observed during the baseline test with #1 diesel. fewer smaller size particles. This is in agreement with
The GMD of the baseline particles was 95 nm, while the results of the EC, and TEOM measurements, which
the GMD for the cold and warm weather water-fuel showed higher EC and TPM reductions for Filter 2 than
emulsions were 68 and 55 nm, respectively. The peak for Filter 3.
particle number concentration for the baseline test Table 1 presents a summary of the measurement
with #1 diesel was 1,065,000/cm3 and was lower than methods presented in this paper and shows their
those observed for cold weather (1,449,000/cm3 ) and appropriate applications and major considerations
warm weather (1,324,000/cm3 ) emulsions. The rela- and limitations. As can be seen each method dis-
tive increases in total aerosol number concentrations cussed has appropriate applications along with certain
were (14%) and (16%) for the cold and warm weather limitations.
water fuel emulsions, respectively. As shown in Figure 3 and Figure 4, the mass
Figure 8 presents the results of the SMPS data measurement results provide evidence that both the
from the filter tests. This data shows that the parti- water-fuel emulsion and the diesel particulate filters
cle number concentrations were higher for the smallest reduced the airborne DPM concentrations. However,
particles and lower for the larger particles. The size dis- the SMPS results tell a slightly different story than
tributions measured during the Filter 1 test were found that derived from the mass measurements. The SMPS
to be characterized with a bimodal distribution hav- results for both the water-fuel emulsion tests and the
ing peak concentrations higher than those observed filter tests show that while the mass is decreasing the
for the baseline case. The net contribution of the particle number concentration is actually increasing.
system to the total particle number was also found Overall, the data in this study emphasize the impor-
to be 105% higher than that of the muffler. tance of using instruments that measure different met-
The effects of Filter 2 and Filter 3 on size distribu- rics when characterizing DPM and evaluating DPM
tion and aerosol concentration were found to be quite control technologies. It is important to realize the

103

Copyright © 2006 Taylor & Francis Group plc, London, UK

 necessity of measuring several different metrics to
achieve a true understanding of DPM and the effects
of DPM control. As shown, control technologies can
work very well in controlling one metric but actually
increase the concentration of a different metric.
REFERENCES
Bugarski, A.D., Schnakenberg, G.H. Jr. , Noll, J.D., Mischler,
S.E., Patts, L.D., Hummer, J.A., & Vanderslice S.E. 2006a.
In-situ effectiveness of selected diesel particulate matter
control technologies for underground mining applications
– Part I. U.S. Department of Health and Human Services.
Report of Investigation. In Press.
Bugarski, A.D., Schnakenberg, G.H. Jr., Mischler, S.E.,
Noll, J.D., Patts, L.D., & Hummer, J.A. 2006b. In-situ
effectiveness of selected diesel particulate matter control
technologies for underground mining applications – Part
II. U.S. Department of Health and Human Services Report
of Investigation. In Press.
Environmental Protection Agency 2002. Health Assessment
Document for Diesel Engine Exhaust, National Center for
Environmental Assessment. EPA/600/8-90/057F. May.
104
Copyright © 2006 Taylor & Francis Group plc, London, UK
Jamriska, L.M., Morawska, S., Thomas, C. He. 2004. Diesel
Bus Emissions Measured in a Tunnel Study. Environ. Sci.
Technol. Vol 38, pg. 6701–6709, 2004.
Kelly, A.N., Morgan, C. 2002. An Evaluation of the Tapered
Element Oscillating Microbalance Method for Measur-
ing Diesel Particulate Emissions. J. Air & Waste Manage.
Assoc. 52: 1362–1377. 2002.
Kittelson, D.B., Watts, W.F., Sanvstrom, J.C., Johnson, J.P.
2005. Influence of Catalytic Stripper on the Response of
real time aerosol instruments to diesel exhaust aerosol.
Journal of Aerosol Science. 36: 1089–1107. 2005.
Mine Safety and Health Administration. 2005 30 CFR Part
57 Diesel Particulate Matter Exposure of Underground
Metal and Nonmetal Miners; Final Rule. Fed. Reg. vol.
70, No. 107, 32868.
Mine Safety and Health Administration 2002. 30 CFR Part
57 “Diesel Particulate Matter Exposure of Underground
Metal and Nonmetal Miners; Proposed Rule,” Fed. Reg.
2002, 30, 47297.
Olson B. 2001. Particle Calibration of the Disposable Per-
sonal Diesel Aerosol Sampler (PDAS). Particle Calibra-
tion Laboratory, University of Minnesota. Report sub-
mitted to National Institute of Occupational Safety and
Health.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Measuring diesel particulate matter in underground mines using

submicron elemental carbon as a surrogate

J.D. Noll, S.E. Mischler, G.H. Schnakenberg, Jr. & A.D. Bugarski
U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control
and Prevention, National Institute for Occupational Safety and Health, Pittsburgh, PA

ABSTRACT: Elemental carbon (EC) is used as a surrogate for regulating the exposure to diesel particulate
matter (DPM) of underground metal/non-metal miners. EC was chosen as a surrogate because EC is selective to
DPM and is a major component of DPM. Using EC as a surrogate also gives one the advantages of no sampling
artifacts and being able to sample at all locations in the mine. Currently, EC represents DPM well in underground
mines. Some control technologies have been shown to possibly alter the relationship between DPM and EC and
characteristics of DPM. Therefore, future work will investigate the relationship between DPM and EC as new
control technologies are implemented.

1 INTRODUCTION
1.1 Background
Long-term exposure to diesel exhaust has become a
concern because diesel emissions are believed to be
a potential carcinogen (NIOSH, 1988). In addition,
acute overexposure to diesel exhaust has been linked
to deleterious health effects such as eye and nose irrita-
tion, headaches, nausea, and asthma (Kahn and Orris,
1988; Rundell et al., 1996; Wade, 1993). Measure-
ments have shown that underground miners can be
exposed to over 100 times the typical environmental
concentrations of diesel exhaust and over 10 times the
concentrations measured in other work environments
where diesel engines are common (Cantrell and Watts,
1997; Nauss, 1998; Haney, 1992).
In the United States, the Mine Safety and Health
Administration (MSHA) has promulgated rules to
limit the exposure of metal/non-metal underground
miners to diesel particulate matter (DPM) to
500 µg/m3 for an interim limit and 200 µg/m3 for a
final limit (MSHA, 2001, 2005). One issue that had to
be overcome was how to measure for DPM.
1.2 Difficulties in measuring DPM in
underground mines
DPM is a complex mixture of particulate elemental
carbon (EC) or soot, particulate and particle bound
organic carbon (OC), sulfates, some metals, etc.
(Kittelson, 1998). Mass measurements of DPM are
prone to interferences from other sources of aerosols
(mineral dust, cigarette smoke, etc.) in the mining
environment and are not sensitive enough for the
concentrations near the proposed final limit (below
200 µg/m3 ). Therefore, a surrogate was needed to
determine DPM exposure.
Initially, total carbon (TC) was considered by
MSHA to be the most adequate surrogate for
DPM because TC accounts for over 80% of the
DPM (MSHA, 2001; Pierson and Brachaczek, 1983;
Kittelson, 1998). However, the EC and OC particles
from mineral dust and OC aerosols from the other
sources commonly present in underground mines,
such as environmental tobacco smoke and oil mist,
were found to interfere with the TC analysis.
A size selective sampler (Cantrell and Rubow, 1991;
McCartney and Cantrell, 1992; Cash et al., 2003, Noll
et al., 2005) has been shown to effectively segregate
the coarse mineral dust from the generally submicron
DPM. Unfortunately, the size selective samplers are
not very efficient in removing cigarette smoke and oil
mist (OC aerosols that generally belong to the same
size category as diesel aerosols). Therefore, cigarette
smoke and oil mist cannot always be avoided when tak-
ing personal samples. For example, when TC vs. EC
was plotted from personal samples below 400 µg/m3
from smokers and non-smokers from two stone mines
(samples corrected for adsorbed vapor phase OC),
there is greater variability in the TC and EC rela-
tionship as samples are potentially exposed to more
interferences (smokers). The R2 goes from about 0.7
to 0.9 when the data from smokers are taken out.
Figure 1 also shows that at the same EC values when

105

Copyright © 2006 Taylor & Francis Group plc, London, UK

 450
400
350
300
TC (␮g/m3)
250
200
150 non-smokers
smokers
100
50
0
0 50 100 150 200 250 300
EC (␮g/m3)
Figure 1. TC vs. EC of personal samples from smokers and
non-smokers in two stone mines.
the majority of the samples were exposed to poten-
tially more cigarette smoke that a larger TC value was
measured. This is probably due to the additional OC
from the cigarette smoke.
At this time, no method to correct for these inter-
ferences has been found. NIOSH did some studies on
using solanesol as a possible surrogate for cigarette
smoke but solanesol was found to be too unstable
(Tucker and Pretty, 2005).
As a result, MSHA turned to using EC as a sur-
rogate for DPM for the interim limit since no other
sources of submicron EC are known to exist in the
metal/non-metal mining environment and EC is a
major component of DPM (MSHA, 2005). This paper
gives an overview of the sampling and analytical
methods presently used to measure EC and some issues
when using EC as a surrogate.
2 SAMPLING AND ANALYTICAL METHODS
2.1 Sampling methods
2.1.1 Personal sampling
For taking personal samples, a personal sampling
pump, a 10-mm Dorr-Oliver cyclone, a submicron
impactor, and two quartz filters placed in tandem (in
series) are used. The pump is operated at 1.7 lpm.
The submicron impactor segregates the mineral dust
from the DPM. This is necessary when using EC as a
surrogate because it avoids overloading a filter with
mineral dust, some mineral dusts contain EC, and
it provides a cleaner sample. When impactors were
first being used, the mining industry brought up con-
cerns with cassette performance and possible defects
(MSHA, 2003). The submicron impactor has been
shown to be efficient in collecting DPM while allow-
ing less than 10% of mineral and coal dust to penetrate
in both laboratory and field studies (Noll et al., 2005;
Copyright © 2006 Taylor & Francis Group plc, London, UK
Birch and Noll, 2004; Noll and Birch, 2004). For exam-
ple, at a dust concentration equivalent to 8-hr time
weighted average of 3 mg/m3 , a sub micron impactor
was shown to collect only 2% of the respirable dust
(Noll et al., 2005).
2.1.2 High volume sampling
When testing the efficiencies of control technolo-
gies in underground mines, the concentration of DPM
can be low. To collect enough material to be at and
above the limit of quantification for NIOSH Analyt-
350 ical Method 5040, one might have to sample for many
hours. When doing research in the field, one does not
always have this luxury. A high-volume (HV) sam-
pling train was developed by Bugarski et al. (2003,
2005) to collect sufficient material for the analysis of
TC and EC using the Method 5040 while still avoiding
mineral dust for control technology studies. The HV
flow rate was achieved by merging flows from five
classifiers, each consisting of a 10-mm Dorr-Oliver
cyclone followed by a U.S. Bureau of Mines (BOM)
single stage diesel impactor, into a single stream. Flow
rates between 1.7 and 2.0 lpm were maintained through
each cyclone and impactor pair. At this sampling flow
rate only particles with geometric mean smaller then
0.8 µm were deposited on the filters. All five clas-
sifiers were attached to a symmetrical plenum that
distributed a total flow rate between 8.5–10 lpm uni-
formly among the five streams. Each of the classifier
assemblies was connected to the plenum chamber by a
3-foot long section of conductive tubing. The outlet of
the plenum was directly connected to a stainless steel
25 mm diameter filter holder containing two stacked
25 mm tissue quartz fiber filters. As seen in Table 1,
the high volume method gave similar results as the per-
sonal sampling methods when collecting samples in an
isolated zone of a metal mine (these were when no con-
trol technologies or just alternative fuels were used).
This high volume sampling method used mass flow
controllers or critical orifices to control the flow rate
and needed electricity to run. However, electricity may
not always be available at sampling locations in the
underground mine being tested. Therefore, we are
currently testing a more portable assembly using a
high flow personal pump with a cyclone with a 1 µm
cutpoint at around 8 lpm (see Figure 2).
2.2 NIOSH analytical method 5040
After a sample is collected, it is sent to a laboratory and
analyzed for elemental and total carbon using NIOSH
Method 5040 (Birch, 2004). This analytical method
analyzes for OC and EC in two different stages. In the
first stage, the OC is measured by ramping the oven
temperature over four progressively higher tempera-
ture steps programmed into the instrument, with the
last step being at about 870 ◦ C in a pure helium (He)
106
 Table 1. High volume (9.7 lpm) sampling vs. standard sam- 3 EC AS A SURROGATE
pling at 1.7 lpm.
3.1 Selectivity of EC
EC concentration between high volume sampling and
personal sampling method As mentioned earlier, submicron EC (after using an
impactor) is selective to DPM in the underground
HV Standard method metal/non-metal mining environment. This gives us
(µg/m3 ) (µg/m3 ) % difference the advantage of sampling for DPM and taking per-
sonal samples in all areas of the mine. If one tried to use
181 172 5 TC as a surrogate, one might have to develop a com-
205 167 20 plicated sampling strategy to avoid the interferences.
287 257 11 A complicated sampling strategy could preclude areas
319 289 10
401 353 13
with the highest concentration of DPM.
414 376 10 In addition to avoiding interferences, known OC
311 296 5 sampling artifacts when using quartz filters (Eatough
213 198 7 et al., 1995; Kirchstetter et al., 2001; Turpin et al.,
1994) do not affect EC based methods. Quartz fil-
HV – high volume sampling ters are known to adsorb vapor-phase OC (not part
Standard method – personal sampling method of DPM). We routinely observe between 30–40 µg/m3
% difference = (HV − standard method)/average(HV − vapor-phase OC adsorbed by the quartz filter in min-
standard method) × 100 ing samples. This could cause a 19–25% error in a TC
result at the proposed final limit of 160 µg/m3 TC if
not corrected. There is also a possibility of some OC
semi-volatiles evaporating and causing a negative bias
relative to the TC results. These artifacts do not affect
the EC results.
Coal dust also has less of an effect on EC results
than TC or DPM mass results since a smaller por-
tion of coal dust is EC (Birch and Noll, 2004; Noll
and Birch, 2004). In both field and laboratory studies,
when using a sub micron impactor, coal dust con-
tributed relatively minimal (<25%) to the EC results
when measuring 160 µg/m3 TC concentrations, but in
a few instances, even when using the impactor, the
coal dust contributed significantly (>25%) to the TC
values when measuring DPM at 160 µg/m3 TC (Birch
and Noll, 2004; Noll and Birch, 2004).

Figure 2. A high volume setup consisting of a cyclone with

a 1 µm cutpoint at 8 lpm, filter to collect particulate and a
high flow rate (5–15 lpm) personal pump.
atmosphere. The EC does not evolve because there is
no oxygen (O2 ) available for it to react. The evolved
OC is oxidized to carbon dioxide (CO2 ), reduced to
methane (CH4 ), and finally measured using a flame
ionization detector (FID). In the second stage, the EC
is measured by reducing the oven temperature to about
600 ◦ C and then again raising the temperature to about
900 ◦ C in a He/O2 atmosphere (O2 is now present to
react with the EC to form CO2 ). The EC is then mea-
sured in the same way as the OC. TC is simply the
sum of OC and EC. The NIOSH Method 5040 also
optically corrects for pyrolysis (charring) of OC.
NIOSH 5040 has been shown to meet the NIOSH
accuracy criteria and has given good results for inter-
laboratory testing (Birch, 2004).
3.1.1 How is DPM represented by EC
The strong points of using EC as a surrogate are that it
is a major part of DPM, is selective to DPM, and can be
sampled and measured accurately. One potential issue
when using EC as a surrogate for DPM is that the EC
fraction of DPM may change depending upon various
factors such as engine duty cycle, fuel type, control
technology, etc.
Some preliminary data demonstrates that, at this
time, EC represents DPM in underground mines well
but may change as new control technologies are imple-
mented. Bugarski et al. 2003, 2005 collected DPM
mass (measured using TEOM), TC, and EC samples
in an isolated zone in a metal mine. The isolated zone
was a section in the mine where only clean air entered
and ventilation could be controlled. This allowed mea-
suring the diesel emissions from a vehicle without
interferences. As seen in Figures 3a and 3b, a lin-
ear relationship existed between DPM and EC and

107

Copyright © 2006 Taylor & Francis Group plc, London, UK

 1400 1800
y = 1.04 x + 90 y = 1.1212x + 41.58
2 1600 R2 = 0.99
1200 R = 0.91
1400
DPM mass (␮g/m3)

1000
1200

TC (␮g/m3)
800 1000
600 800

400 600
400
200
200
0
0 200 400 600 800 1000 1200 0
0 200 400 600 800 1000 1200 1400 1600
(a) EC (␮g/m3) EC (␮g/m3)

600 Figure 4. TC vs. EC for data from four different mines dur-
y = 1.10x + 52.15
R2 = 0.98 ing actual production (trying to avoid interferences).
500

400 between TC and EC increases at the lower concentra-

TC (␮g/m3)

tions (<230 µg/m3 ) of TC which could be the result

300
200
100
0
0 100 200
(b) EC (␮g/m3)
Figure 3. (a) DPM (mass) vs. EC for vehicles in the isolated
zone. (b) TC vs. EC for vehicles in the isolated zone study.
between TC and EC for the types of vehicles tested
in this mine (Noll et al., 2005).
This data was obtained when the vehicles ran
No. 2 diesel fuel, diesel oxidation catalytic convert-
ers (DOCs), biodiesel, or water emulsified fuels.
These conditions are present in operating underground
mines.
This relationship was confirmed when sampling in
four mines during actual production (see Figure 4)
(Noll et al., 2005). The sampling was done at the
main exhausts of the mine where oil mist and cigarette
smoke should be so diluted compared to diesel that
they have minimal effect on the TC results. SKC DPM
cassettes were used to prevent the collection of min-
eral dust. The tandem filter was used to correct for the
adsorption of vapor phase OC (Eatough et al., 1995;
Kirchstetter et al., 2001; Turpin et al., 1994).
DPM mass was not measured during these field
studies but since DPM consists of over 80% TC, most
DPM and EC samples were shown to have a linear
relationship.
It is important to note that the relationship between
TC and EC looks very well with the above data
when interferences were minimized. However, as
shown in Figure 1, when taking personal samples,
cigarette smoke can provide a substantial interfer-
ence. Even when taking precautions, the variability
of interferences or at least partially.
Some preliminary data has shown that as certain
control technologies are implemented in underground
mines, the DPM and EC relationship and even the char-
acteristic of the DPM may change (Bugarski et al.,
300 400 500 2003). In the Bugarski et al. 2003, 2005 studies, diesel
particulate filter (DPF) systems were shown to pref-
erentially reduce EC over OC and non carbon DPM
(e.g. a platinum (Pt)-catalyzed DPF reduced EC by
95% and OC by 80%) which could cause the relation-
ship between DPM and EC to change when these DPF
systems are implemented. The characteristics of DPM
were also shown to change (e.g. the mass of particulate
decreased but the number of nanoparticles increased
when some DPF systems were used). Even though the
DPM and EC relationship changed, EC still could be
measured accurately and followed the trend of DPM
(as DPM decreased so did the EC). The DPF system
that preferentially reduced EC over OC reduced both
substantially; therefore the concentration of DPM or
the contributions of the DPM from vehicles with these
control technologies may be low enough that the poten-
tial change in DPM and EC relationship may not be a
concern.
We do not know at this time the magnitude that the
different control technologies effects will contribute to
the future DPM and EC relationship in underground
mines. We do not know what type of control tech-
nologies will be eventually implemented in the mines,
how the characteristic of DPM could change, and at
what concentration. Future work will be needed as new
control technologies are introduced into the mines to
determine the characterization of DPM.
4 CONCLUSION
EC is selective to DPM in underground mines and
can be sampled and measured accurately. Using EC

108

Copyright © 2006 Taylor & Francis Group plc, London, UK

 as a surrogate allows one to measure in all areas of
the mine. It also gives the benefit of being free of OC
sampling artifacts and makes it easier to sample in coal
mines. Under present conditions, EC represents DPM
well and is a major component. As some control tech-
nologies are implemented, such as diesel particulate
filter systems, the relationship between DPM and EC
may change, may vary between mines, and the char-
acteristics of DPM itself may change. EC can still be
measured accurately and it will still follow the trend of
DPM. Future work will be needed as new control tech-
nologies are introduced into the mines to determine the
characterization of DPM. At this time the benefits of
using EC as a surrogate outweigh the problems.
Disclaimer: Mention of a company name or prod-
uct does not constitute endorsement by the Centers
for Disease Control and Prevention. The findings and
conclusions in this report are those of the authors and
do not necessarily represent the views of the National
Institute for Occupational Safety and Health.
REFERENCES
Birch, M. E. 2004. In NIOSH Manual of Analytical Methods
(NMAM ), ed. O’Connor P.F.,Third Supplement to NMAM,
4th Edition. Cincinnati, OH: Department of Health and
Human Services, Public Health Service, Center for Dis-
ease Control and Prevention, National Institute for Occu-
pational Safety and Health. DHHS(NIOSH) Publication
No. 2003-154.
Birch, M. E. & Noll, J. D. 2004. Submicrometer elemental
carbon as a selective measure of diesel particulate matter
in coal mines. Journal of Environmental Monitoring. 6,
799–806.
Bugarski, A., Schnakenberg, G., Noll, J. D., Mischler, S.,
Crum, M. & Anderson, M. 2005. Evaluation of Diesel
Particulate Filter Systems and Biodiesel Blends in an
Underground Mine. Society of Mining and Metallurgical
Engineers Transactions, vol. 318.
Bugarski, A., Schnakenberg, G., Noll, J. D., Mischler, S.,
Patts, P., Hummer, J., Vanderslice, S., Crum, M. &
Anderson, R. 2003. The Effectiveness of Selected Tech-
nologies in Controlling Diesel Emissions in Underground
Mine – Isolated Zone Study at Stillwater Nye Mine.
http://www.msha.gov/01%2D995/dpmdocs/stillwater.pdf
Cantrell, B. K. & Watts, W. F. 1997. Diesel Exhaust Aerosol:
Review of Occupational Exposure. Appl. Occup. Environ.
Hyg., vol. 12, pp 1019–1026.
Eatough, D. J., Tang, H., Cui, W. & Machir, J. 1995.
Determination of the Size Distribution and Chem-
ical Composition of Fine Particulate Semivolatile
Organic Material in Urban Environments Using Diffu-
sion Denuder Technology. Inhalation Toxicology, vol. 7,
pp 691–710.
EPA. Impacts of Lubrizol’s PuriNOx Water/Diesel Emul-
sion on Exhaust Emissions from Heavy-Duty Engines
EPA420-P-02-007, 2002.
Haney, R. A. 1992. Diesel particulate exposures in under-
ground mines. Mining Engineering, pp 173–176.
Haney, R. 2000. Sampling and Interferences for Diesel
Particulate Matter in Underground Metal/Non-metal
109
Copyright © 2006 Taylor & Francis Group plc, London, UK
Mines. http://www.msha.gov/S&HINFO/TECHRPT/
diesel/haneyimvc7.pdf
Heywood, J. B. Internal Combustion Engine Fundamentals.
McGraw-Hill, Inc. New York (1988).
Howes, P. 2000. City of Houston Diesel Field Demonstration
Project. ERMD Report 01–36.
Kahn, G., Orris, P. & Weeks, J. 1988. Am. J. Ind. Med.,
13, 405.
Kirchstetter, T. W., Corrigan, C. E. & Novakov, T. 2001.
Laboratory and field investigation of the adsorption of
gaseous organic compounds onto quartz filters. Atmo-
spheric Environment, vol. 35, pp 1663–1671.
Kittelson, D. B. 1998. J. Aerosol Sci., 29, 575–588.
Mine Safety and HealthAdministration. 1998. Diesel Particu-
late Matter Exposure of Underground Metal and Non-
metal Miners; Proposed Rule. Federal Register 30:57. pp
58103–58270.
McCartney, T. C. & Cantrell, B. K. 1992. A cost-effective
personal diesel exhaust aerosol sampler. In: Diesels in
underground mines: Measurement and control of partic-
ulate emissions (Information circular 9324). Proceedings
of the Bureau of Mines Information and Technology
Transfer Seminar, Minneapolis, MN, September 29–30,
pp 24–30.
Mine Safety and Health Administration. (January 19, 2001)
30 CFR Part 57 Diesel Particulate Matter Exposure of
Underground Metal and Nonmetal Miners; Final Rule.
Fed. Reg., vol. 66, No. 13, 5706.
Mine Safety and Health Administration. (June 6, 2005) 30
CFR Part 57 Diesel Particulate Matter Exposure of Under-
ground Metal and Nonmetal Miners; Final Rule. Fed. Reg.,
vol. 70, No. 107, 32868.
Mine Safety and Health Administration. Diesel Particulate
Matter Exposure of Underground Metal and Nonmetal
Miners; Proposed Rule. Federal Register. 30:57 (2003).
pp 48668–48721.
Nauss, K. 1998. Diesel exhaust: A critical analysis of emis-
sions, exposure, and health effects. HSA Bulletin, Novem-
ber, pp 10–13.
NIOSH. Carcinogenic Effects of Exposure to Diesel Exhaust.
Current Intelligence Bulletin No. 50. U.S. Department
of Health and Human Services, Public Health Service,
Centers for Disease Control, National Institute for Occu-
pational Safety and Health, DHHS (NIOSH) Pub., 1988,
No. 88–116.
Noll, J. D., Bugarski, A. D., Schnakenberg, G. H., Patts, L. D.,
Mischler, S. E. & McWilliams, L. 2005. The Relationship
between Elemental Carbon, Total Carbon, and Diesel Par-
ticulate Matter in Several Underground Metal/Non-metal
Mines. Submitted to Mining Engineer.
Noll, J. D., Timko, R. J., McWilliams, L., Hall, P. & Haney, R.
2005. Sampling Results of the Improved SKC Diesel Par-
ticulate Matter Cassette. Journal of Occupational and
Environmental Hygiene, 2, 29–37.
Noll, J. D. & Birch, M. E. 2004. Evaluation of SKC DPM
Cassettes for Monitoring Diesel Particulate Matter in Coal
Mines. J. Environ. Monit., 6, 973–978.
Pierson, W. R. & Brachaczek, W. W. 1983. Aerosol Sci.
Technol., 2:1.
Rideout, G. Impact of Lubrizol’s PuriNOX Low-emissions
Diesel Fuel on the Exhaust Emissions from a Heavy-duty
Off-road Diesel-powered Vehicle, ERMD Report 99–37,
October, 1999.
Rundell, B, et al. Occup. Environ. Med., 1996, 53, 658.
 Tucker, S. P. & Pretty, J. R. 2005. Identification of Angeles Basin. Atmospheric Environment, vol. 28, pp
oxidation products of solanesol produced during air 3061–3071.
sampling for tobacco smoke by electrospray mass Wade, J. F., III & Newman, L. S. J. Occup. Med., 1993,
spectrometry and HPLC. The Analyst, vol. 130, issue 35, 149.
10, pp 1414–1424.
Turpin, B. J., Huntzicker, J. J. & Hering, S.V. 1994. Investi-
gation of organic aerosol sampling artifacts in the Los

110

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Emissions-based maintenance program

S. Forbush
Arch Western Bituminous Group

ABSTRACT: Arch Western Bituminous Group (AWBG) operates three underground coal mines in Utah and
one in Colorado. In 1997, MSHA passed the “New Diesel Regulations” (30 CFR 75: 1900–1915), including 30
CFR 1914 g, which requires that mine operators check each piece of Heavy Duty and Permissible equipment for
carbon monoxide emissions and perform a vehicle safety inspection. AWBG took this regulation very seriously
and established its Emission Based Maintenance Program. The EBMP came in four steps. Step one established
a base line for each vehicle engine for which testing was required. Step two established the CO emission levels
or its certified emission levels for each engine. In step three, AWBG technicians worked to reach the certified
CO emission levels for each engine type. In step four, technicians strove to improve on these levels.
This paper will describe in greater detail how AWBG achieved improvements in the emissions of its diesel
engine fleet. For example, the author will include a detailed explanation on how to use emission contour maps
and drive train torque absorption characteristics to achieve optimum horsepower and speed while maintaining
low emissions. All mines employing diesel power can achieve lower emissions by applying the knowledge and
experience contained in this paper.

1 INTRODUCTION elevations. The elevations range between 6,500 feet

In 1996, the Mine Safety and Health Administra-
tion (MSHA) published regulations related to the
health and safety of miners who work in underground
coal mines which employ diesel powered equipment.
These regulations established rules designed to mit-
igate the health and safety issues related to the use
of diesel-powered equipment in underground coal
mines.
A section of this regulation (30CFR75: 1914g)
requires mine operators to check weekly the level of
carbon monoxide (CO) in the undiluted exhaust stream
of each of their permissible and heavy-duty vehicles.
MSHA’s weekly test will alert mine operators to sig-
nificant changes in the operating conditions of diesel
engines in their fleets. A significant change in the
CO level of a diesel engine’s undiluted exhaust stream
indicates a change in the operating condition of that
engine worthy of further investigation. CO concentra-
tions or changes in CO concentrations are limited to
2,500 parts per million by this regulation.
2 AWBG MINES
AWBG produces 20 million tons of low sulfur coal
from 4 mines in the western United States. All 4
operations mine coal from seams that outcrop at high
and 8,700 feet above sea level. AWBG’s Skyline mine
produces from a seam that outcrops at 8,700 feet above
sea level. This makes the mine the highest under-
ground coal mine in the US. Trivial this elevation is not
because as AWBG discovered it significantly impacts
the performance of diesel engines.
In their underground operations, the 4 mines use
420 diesel-powered units, which range from 23 horse-
power to 500 horsepower. Diesel-powered units work
in all facets of the mining operations.
3 AWBG’S APPROACH
The initial testing at AWBG’s first site proved that
diesel engines in poor condition could produce CO
in excess of the 2,500 parts per million limit. The test-
ing proved that the engine emission data available to
AWBG was inadequate. AWBG did not know what the
CO content in the emissions from its engines should
be. During the first few weeks of testing, AWBG tech-
nicians found the fleet average for one of our mines was
1597 ppm CO and 997 ppm for NOX. These original
concentrations served as base lines to determine if the
changes made to engines were effective or not. AWBG
trained technicians at each mine to do the weekly tests.
This ensured that all testing procedures were consistent
and that the results were accurate.

111

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Once AWBG had consistent repeatable results, its
technicians needed to know what the emissions from
each engine family should be. Engine builders in
general prefer to keep emission information to them-
selves. With some help from the Underground Coal
Diesel Partnership, MSHA released most of the emis-
sions information for the engines used at AWBG’s
mines. AWBG technicians compared this published
data to actual test data to determine engines’ condi-
tions. The comparisons alarmed the technicians. Most
of AWBG’s initial emissions tests indicated that the
fleet’s engines operated at emission levels 5 to 10 times
higher than the respective engine’s certification data.
AWBG began an investigation to determine the rea-
sons for the high emission levels. At that point in
time, AWBG’s technicians determined that all of the
tested engines’ fuel rates were set according to the
engine builders’ specifications. Regardless of spec-
ifications, AWBG’s management group committed
to limit engine emissions to values less than those
stated in a respective engine’s MSHA certification
information.
4 AWBG’S PROCESS TO REDUCE EMISSIONS
AWBG’s technicians knew they needed information if
they were to achieve emission levels given in MSHA’s
certification data for the various engines in AWBG’s
fleet.
AWBG went first to engine builders for help. Their
elevation derate schedules were insufficient for the
conditions under which AWBG operated the various
engines. AWBG decided to develop its own engine
settings for derate and timing. This decision forced
AWBG into its own testing program. Technicians were
determined to find the correct fuel and timing settings
for every engine family.
AWBG technicians discovered that trial and error
was the only way they could accomplish this. The
first engine AWBG tested this way, the Caterpillar
3306 PCNA. This engine is used in most of the opera-
tions’permissible equipment.The first round of testing
required 36 changes to find the correct settings for fuel
and timing. The time and effort paid off. The engine’s
CO concentration was reduced by 90% and its NOX
was reduced by 50%. This family of engines now oper-
ates at an average of 160 ppm CO and 447 ppm NOX .
AWBG continued the same process for most of the
rest of the fleet. The fleets’ averages are now below
certification values.
5 CHANGING HOW ENGINE POWER IS
TRANSFERRED TO THE DRIVE TRAIN
Achieving improved emissions did not come without
a compromise in diesel powered unit performance.
112
Copyright © 2006 Taylor & Francis Group plc, London, UK
Reduced fuel rates decreased the power. The power
loss was not equal to the reduction in fuel, but the
loss meant that most of the vehicles were now under-
powered. AWBG hired a consultant to help solve this
problem. Just like other vehicles, AWBG’s diesel units
transmit engine torque through a torque converter to
the rest of the drive train components. The drive trains
have fixed ratios, which are not easily changed. Torque
converters come in a variety of ratios and it was to these
ratios that AWBG turned to solve the power problem.
Torque converter ratio changes permitted the rest of
the units’ drive trains to consume the power from their
engines at the speed for optimum power and minimum
emissions. This change had a very positive effect on
the overall performance of the vehicles.
6 AWBG’S NEXT STEP
Once AWBG’s technicians had optimized emission
and power performance of the diesel fleet, they wanted
to maintain the engines at that level. AWBG’s techni-
cians planned for frequent timing and fuel adjustments
to maintain this improvement. Such adjustments would
have been labor-intensive. This was not the case. The
technicians were able to maintain the diesel engines’
improved performance levels with very few follow-up
adjustments.
MSHA’s new diesel regulations appeared onerous,
butAWBG’s implementation plan to comply, improved
the overall efficiency of its diesel fleet. As a result,
AWBG has established emission baselines for all of
its engines. All of the baselines fell below certification
values.
AWBG uses the emission baseline values to deter-
mine when technicians should remove a unit from
service. When weekly test results indicate a unit’s
emissions exceed its baseline by 20%, technicians
remove the unit from service. Typically, they repair
95% of the units at the time of the test. The other 5%
of the vehicles go to the shop for repair.
7 BENEFITS
AWBG’s Maintenance Managers constantly study the
cost for operation, maintenance, and repair of their
diesel fleets. The EBMP at AWBG has helped lower
these costs through reduced fuel consumption. Fuel
savings are the direct result of running engines under
optimum fuel settings and torque converter ratios
rather than in a lug and over fueled condition. These
modifications produce a substantial savings in operat-
ing cost.
AWBG’s mines’ operating elevation required
enhancements to the engine builders’ derate tables.
The trial and error process AWBG employed enabled
its technicians to derate engines beyond the original
equipment manufacturer’s table limits.
 AWBG’s diesel unit maintenance and repair costs
have improved because average engine life has
increased from 3,000 hours to over 10,000 hours.
Additional cost savings came from improved
exhaust filter life. On 20 January 2000, MSHA pub-
lished regulations that required exhaust filter after-
treatment on all heavy-duty and permissible vehicles
used underground. After-treatment for diesel-powered
equipment is costly and maintenance labor intensive.
Extending the life of a filter reduces per unit operating
cost of a diesel-powered unit. AWBG diesel units fit-
ted with after treatment average 90 hours of operation
between filter replacements.
The EBMP AWBG implemented in combination
with the required after treatment for some diesel units
provides a health benefit, which is difficult to value.
When vehicles operate in a confined area, someone
within that area will breathe a diluted concentration of
the engines’ exhausts. You can dilute these pollutants
further with greater ventilation volumes, but the more
diesel emission pollutants a mining operation produces
the higher the ventilation level required to provide
sufficient dilution. Diesel units operating at optimum
performance levels mean minimum emissions, max-
imum power, lower ventilation requirements, and
improved air quality for the personnel working in the
underground environment.
8 HOW DOES EBMP WORK?
Most of the maintenance decisions for engines and
drive trains are based on the emissions of the engine.
113
Copyright © 2006 Taylor & Francis Group plc, London, UK
When the weekly emissions check shows that an
engine is producing higher than normal emissions,
technicians troubleshoot the engine based on which
gas concentration in the emission exhaust stream is
out of specification. After the technician locates the
trouble, he consults with his supervisor. They analyze
the history of the engine to make the most economi-
cal repair to reduce the emissions. That decision may
involve the engine or the drive train. This puts the
maintenance department in proactive mode instead of
a reactive mode.
The EBMP permitsAWBG’s technicians to find fuel
system, combustion cylinder, head, valve, and drive
train problems before they cause catastrophic failures.
At this point, the cost to repair is minimal. If these
problems were not uncovered, the cost to repair the
resultant catastrophic failures would be much greater.
9 CONCLUSION
A successful EBMP requires full support from every
manager within an organization. Without this support,
an EBMP will fail. Implementing such a program to
meet mandated regulations would not seem at first
to provide other benefits. AWBG’s experience proved
that other benefits accrued to the organization from its
efforts to meet the mandated regulations.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

USA salt mines and DPM control strategies

E.Z. Manos
Detroit Salt Company L.C. Detroit, Michigan, USA

M. Liby
Hutchinson Salt Company Hutchinson, Kansas, USA

ABSTRACT: The U.S. federal regulations limiting diesel particulate matter (DPM) exposures in underground
non-coal mines have compelled mines to consider a range of DPM control strategies. The first line of defense
for most mines is to dilute the DPM with more air, by improving the mines’ ventilation systems. However, these
ventilation upgrades have been shown to provide only a limited amount of improvement. In order to comply with
the final exposure limit of 160 micro-grams of total carbon per cubic meter, due to come into effect in January
2006, mines need to limit the source of the emissions of DPM at the engines themselves. Two salt mines –
one in Detroit, Michigan and the other in Hutchinson, Kansas – have reduced their DPM exposures by taking
both approaches. Ventilation improvements essentially brought the mines into compliance with the current DPM
exposure limit of 400 micro-grams of total carbon per cubic meter. Further reductions have been achieved with
the use of 100% soy fuel. With this strategy, both of these mines are close to the final exposure compliance limits
of 160 micro-grams of total carbon per cubic meter.
This paper will discuss and compare the two salt mines DPM strategies including economics, ventilation
upgrades, equipment modifications and the soy fuel trials and results.

1 HISTORY OF THE MINES for ventilation purposes with the Hutchinson Salt
Mine also sharing parts of their ventilation air with
Both the Hutchison and the Detroit Salt Mines share a record storage company (Underground Vaults &
a long history of underground salt mining. The Storage Inc.) and with the Kansas Underground Salt
Hutchinson Mine was started in 1922, while the Mining Museum.
Detroit Mine started in 1906, making it one of the old-
est operating mines in the USA. (Manos 2003)
From their past reliance on labor and inefficient 2 VENTILATION SYSTEMS
electric equipment, the mines today have evolved into
modern and efficient operations, largely relying upon Both salt mines operate using the room and pillar
diesel equipment for their production, maintenance mining method with a fresh air intake and return air
and service requirements. However, both mines do
utilize some electrical production equipment under-
Table 1. Lists the diesel and electric underground produc-
ground where economically possible; for instance,
tion equipment used at the mines.
both mines currently use an electrical under-cutter and
drill jumbo. For the Detroit Mine, diesel equipment Diesel & Electrical Mobile Equipment
became prominent during the 1950’s and 1960’s about Detroit Salt Hutchinson Salt
the same time our national highway system was being Company Company
developed and de-icing salt became the primary sales
product for the mine. The Hutchinson Salt Mine how- Equipment Type No. hp/ea No. hp/ea
ever did not purchase their first piece of production
diesel equipment until 1984 when they began to use Undercutter 3 NA 2 NA
load/haul dump (LHD) units for their mucking cycle. Drills 2 200 3 NA
Both salt mines run two shifts a day, 5 to 6 days Powder Rig 1 113 1 160
Scaler 1 113 2 85
per week. The mining method for both mines is room Loaders 4 325 6 250
and pillar and production is achieved by drilling and Trucks 0 0 4 150
blasting. Both salt mines have small restrictive shafts

115

Copyright © 2006 Taylor & Francis Group plc, London, UK

 systems. The fresh air is coursed throughout the mine Table 2. MSHA Baseline DPM Exposures.
to the working faces by a brattice wall system and
auxiliary fans; where blast gases and diesel exhaust TC EC
are swept into a return air system that is directed out Date Location Job µg/m3 µg/m3
of the mine.
10/17/2001 UG LHD Operator 30 271
The Hutchinson Salt Mine operates in a blower sys-
10/17/2001 UG LHD Operator 415 349
tem with the main mine fan on the surface. This mine 10/17/2001 UG Laborer 22 22
is ventilated using a 48 Peabody axial vane fan oper- 10/17/2001 UG Powder man 368 305
ating at 6.2 WG generating 75,000 cfm. (Hartman 10/18/2001 UG Crusher Operator 192 139
1995) 10/18/2001 UG Cutting Mach 402 345
The ventilation system at the Detroit Salt Mine is Operator
a combination; forcing and exhausting system with 10/18/2001 UG Laborer 172 125
the main mine fan placed underground. The Jeffrey’s 10/18/2001 UG Hoist Operator 247 192
Model 8HU60 Aerodyne axial vane fan operates at 10/18/2001 UG Mine Exhaust 968 872
10/18/2001 UG LHD Operator 758 674
7.1 WG of total pressure and generates 127,000 cfm.
(Manos & Wala 2004)

Table 3. MSHA Baseline DPM Exposures.

3 MSHA BASELINE DPM STUDY
TC EC
Prior to the promulgated interim DPM exposure lim- Date Location Job µg/m3 µg/m3
its of 400 micro-gram of total carbon per cubic meter
set forth on July 20, 2002; the Mine Safety and 12/4/2001 UG Powder Gang 719 366
Health Administration (MSHA) conducted diesel par- 12/4/2001 UG Powder Gang 656 329
12/4/2001 UG Air Intake 72 42
ticulate sampling at various mines as part of a June,
12/4/2001 UG Wagon 1024 458
2001 joint MSHA/Industry Protocol settlement to the 12/4/2001 UG Cross Cut 435 305
challenged January 19, 2001 MSHA rule to limit 12/4/2001 UG Cross Cut 466 305
DPM exposure Diesel Particulate Matter Exposure of 12/4/2001 UG Wagon 926 458
Underground Metal and Nonmetal Miners, 66 Fed. 12/4/2001 UG Wagon 803 383
Reg. 5706, 2001 (codified at 30 C.F.R. Part 57). 12/4/2001 UG Wagon 708 383
(MSHA 2001) 12/4/2001 UG Cross Cut 393 283
The Hutchinson Salt Mine was one of 31 metal/non
metal underground mines that DPM samples were
taken and studied for their baseline study. On October
17 & 18, 2001 DPM samples were taken for various job 4 VENTILATION IMPROVEMENTS
classifications and locations in the mine. The results
of this study were given as total carbon and elemental The first line of improvements that were undertaken at
carbon. (MSHA 2001) the Hutchinson Salt Mine was a comprehensive pro-
On December 4, 2001, the Detroit Salt Mine took gram to upgrade various ventilation controls through-
part in a MSHA baseline study to assess Ammonia out the mine. These works included repairing/reducing
Nitrate and Fuel Oil (ANFO) as an interferent to diesel leakage around brattice walls, and keeping their fresh
particulate sample matter. The DPM samples were air brattice lines closer to the working faces. This
collected with SKC, Inc samplers equipped with sub- mine also installed auxiliary fans in key positions
micron impactors. The DPM samples were analyzed throughout the mine to move and redirect the air.
according to NIOSH Method 5040 (MSHA Method In addition to brattice line improvements, The
P-13) to determine elemental, organic and total carbon Detroit Salt Mine performed a comprehensive ven-
values. (MSHA 2001) tilation study in September 27, 2003 to determine
Several conclusions can be ascertained from these the strengths and weaknesses of the current ventila-
early MSHA tests at these salt mines in relation to tion system at the mine. From this study, work was
the promulgated rule: 1) several samples in critical performed on the following ventilation short comings:
areas and job classifications failed the interim limit of 1) new high pressure walls were installed underground
400 micro-grams of total carbon per cubic meter 2) and down the shaft to eliminate leakage on the intake
the early elemental carbon to total carbon relationship air side of the main mine fan 2) a high pressure loss
varies with each sample and 3) to use single samples obstruction where the intake air takes a sharp turn from
for compliance purposes is inaccurate due to DPM the shaft plenum was structurally modified to reduce
exposure concentrations that vary widely from area to the shock loss and the extremely high air velocity that
area and day to day. was present. (Manos & Wala 2004)

116

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 4. MSHA DPM Exposures from April 15, 2003.

Date Location Job Contam. Conc’n PEL

4/15/2003 UG Drill Oper. Jumbo TC 227 400

4/15/2003 UG Utility Man TC 204 400
4/15/2003 UG Cutting Mach Oper. TC 168 400
4/15/2003 UG Utility Man TC 76 400
4/15/2003 UG Drill Oper. Rotary TC/EC 229 400
4/15/2003 UG Utility Man TC/EC 201 400

Table 5. MSHA DPM Exposures from 12/17/03 & 3/2/04.

Date Location Job Cont. Conc’n PEL

12/17/2003 UG Powder Gang TC 446 400

12/17/2003 UG Drill Operator TC 445 400
12/17/2003 UG Loader Operator TC 241 400
12/17/2003 UG Loader Operator TC 201 400
12/17/2003 UG Powder Gang TC 510 400
12/17/2003 UG Drill Operator TC 497 400
12/17/2003 UG Loader Operator TC 273 400
12/17/2003 UG Loader Operator TC 221 400
3/2/2004 UG Powder Gang TC 407 400
3/2/2004 UG Drill Operator TC 366 400
3/2/2004 UG Loader Operator TC 289 400
3/2/2004 UG Cutting Mach Oper. TC/EC 259 400
3/2/2004 UG Powder Gang TC/EC 469 400
3/2/2004 UG Drill Operator TC/EC 431 400
3/2/2004 UG Drill Operator TC/EC 276 400
3/2/2004 UG Cutting Mach Oper. TC/EC 105 400

Upon completion of the ventilation improvements
both mines collected new DPM exposure data.
Further MSHA DPM testing at the Hutchinson
Mine during an inspection on April 15, 2003 shows
the ventilation improvements performed at the
Hutchinson Salt Mine resulted in compliance with the
400 micro-grams of total carbon per cubic meter limit.
Subsequent MSHA DPM testing at the Detroit Salt
Mine during inspections on December 17, 2003 and
March 2, 2004 showed the Detroit Mine to have over
exposures on 7 of 16 total carbon samples taken and
compared to the 400 micro-grams of total carbon per
cubic meter limit.
Several conclusions can be reached from the sub-
sequent DPM data collected after the ventilation
improvements: 1) the ventilation improvements made
at the Hutchinson Mine had a greater impact upon
the subsequent MSHA DPM sampling when com-
pared to the Baseline MSHA Study done in 2001 2)
the Detroit Mine was compliant with the interim limit
exposures except for the drilling and powdering opera-
tions (To remedy the DPM exposure problems within
these operations; the mine purchased a new powder
machine with a Tier 3 – Deutz engine and the Tamrock
Drill Jumbo’s engine was rebuilt). 3) neither mine;
even with the ventilation and equipment improvements
would not comply with the 160 micro-grams of total
carbon per cubic meter exposure limit due to in effect
January of 2006.
5 SOY FUEL
Hutchinson Salt Mine in early June of 2003 made
the decision to try an alternative fuel as part of their
strategy to reach the 160 micro-grams of total car-
bon per cubic meter limit. On June 17, 2003 a 100%
soy fuel made from soybean oil was used in #7
LHD an Eimco loader equipped with a Detroit Diesel
D-50 engine. To the authors’ knowledge; the Hutchin-
son Salt Mine is recorded as the first underground
mine to use 100% soy fuel as the primary fuel in a
piece of diesel equipment in the United States. From
the results of the controlled testing which resulted in
no damage to the engine; the Hutchinson Mine man-
agement made the decision to use the fuel in all the
underground diesel equipment. An early precursor to
success was quoted from a visual observation when

117

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 6. MSHA DPM samples from 11/18/2003 & 3/2/2005.

Date Location Job Cont. Conc’n PEL

11/18/2004 UG Powder Gang TC 97 400

11/18/2004 UG Cutting Mach Oper. TC 78 400
11/18/2004 UG Drill Operator TC 66 400
11/18/2004 UG LHD TC 66 400
11/18/2004 UG Powder Gang TC/EC 42 400
11/18/2004 UG LHD TC/EC 42 400
11/18/2004 UG Drill Operator TC/EC 34 400
11/18/2004 UG Cutting Mach Oper. TC/EC 32 400
3/2/2005 UG Loader Operator TC 90 400
3/2/2005 UG Cutting Mach Oper. TC 72 400
3/2/2005 UG Drill Operator TC 58 400
3/2/2005 UG Laborer TC 49 400

Table 7. MSHA DPM samples from 8/10/2005 & 12/8/2004.

Date Location Job Cont. Conc’n PEL

8/10/2005 UG Cutting Mach Oper. TC 142 308

8/10/2005 UG Powder Gang TC 112 308
8/10/2005 UG Loader Operator TC 60 308
12/8/2004 UG Powder Gang TC 159 400
12/8/2004 UG Powder Gang TC 136 400
12/8/2004 UG Drill Operator TC 71 400
12/8/2004 UG Loader Operator TC/EC 147 400
12/8/2004 UG Loader Operator TC/EC 114 400
12/8/2004 UG Powder Gang TC/EC 68 400

one underground miner remarked “the diesel soot in
my nose has disappeared”.
However, the use of soy fuel presents several logistic
and mechanical hurdles to overcome; 1) soybean oil is
a fatty acid and as such it will dissolve accumulations
that have formed in fuel tanks and injectors thus plug-
ging up fuel filters until the system is clean 2) the soy
fuel has a gel temperature of 38 degrees Fahrenheit;
when used during the winter time, multiple tanks must
be used to allow the fuel to adjust to the underground
temperature and 3) the fuel will penetrate rubber hoses
due to its high lubricity, rubber hose fuel lines will have
to be replaced with Teflon lined hoses that will han-
dle the fatty acid esters in the fuel. On 11/18/2003 &
3/2/2005 MSHA performed DPM sampling of the
Hutchinson Mine with the following tabled results.
The Detroit Salt Mine’s use of the soy fuel occurred
in early May of 2004 as another engineering alterna-
tive to lower their DPM exposure limits. The Detroit
Mines trial included using 50% #1 Low Sulfur fuel and
50% Soy Fuel as a mix in their #3 980 G Caterpillar
front-end loader with a 3406 engine in it. After a three
day trial period with out any harmful side effects other
then previously stated above; a 100% soy fuel was used
in this piece of equipment. Within a four week period
all underground equipment was running soy fuel as
the primary fuel except for the John Deere Gator util-
ity vehicles, which have a difficult time starting when
this fuel is used. On 12/8/2004 & 8/10/2005 MSHA
performed DPM sampling of the Detroit Mine with
the following tabled results.
From tables 6 and 7 it is apparent that the use of
soy fuel within these two salt mine’s operations has
had a major impact upon the DPM exposure samples.
Correlation between the amount of equipment using
soy fuel and the corresponding DPM exposure values
within the individual ventilation systems can also be
seen. The Hutchinson Salt Mine uses approximately
25,000 gallons per year compared to approximately
60,000 gallons per year at the Detroit Mine.
6 DISCUSSION OF DPM STRATEGY RESULTS
MSHA’s DPM exposure samples beginning with the
baseline study in 2001 including DPM samples from
inspections up to the year 2005 can be followed
with the Detroit and Hutchison salt mines strategy to
accommodate the promulgated DPM rule. From this
history, incremental changes can be identified with

118

Copyright © 2006 Taylor & Francis Group plc, London, UK

 each DPM control strategy enactment. The costly ven-
tilation and equipment improvements coupled with the
use of soy fuel by both mines has decreased the DPM
exposure levels in their respective underground mines
but does not completely solve the problem. As seen
from the data, without the use of soy fuel as an alter-
native fuel, neither of these mines would comply with
final exposure limit of 160 micro-grams of total car-
bon per cubic meter due to be enforced in January
of 2006. Even with the soy fuel assisting the Detroit
Mine as a DPM control strategy; it is close to violat-
ing the lower 2006 enforcement limit on single sample
overexposures.
The reliance of soy fuel to meet the final exposure
limit is not without concerns about reliable long-term
supply and pricing. If soybean oil has any draw backs,
it is the lack of supply. As stated by the Michigan Soy-
bean Association, “the refiners of Soy Bean oil have an
annual production capacity of 175 million gallons. To
date only 25 million gallons per year has been pro-
duced”. Hutchinson Salt Company has experienced
rationing in the form of monthly quotas, due to the
limited supply in Kansas of soybean refining. Sup-
pliers are unable, or unwilling to enter into long term
supply agreements, calling into question the reliability
of the supply of soybeans in the future.
Concurrently with the supply issues, the cost of
soy fuel has almost tripled during the eighteen months
Detroit Salt Company has used the product. Elimina-
tion of the tax credit, due to expire in 2006 will almost
certainly result in substantial increases in the delivered
price, well above current #1 low sulfur prices.
Even with the success of the soy fuel as a DPM con-
trol strategy; there is a lack of data associated with the
ever changing underground mining systems to predict
119
Copyright © 2006 Taylor & Francis Group plc, London, UK
if it brings these two mines into compliance with the
final exposure limit for the future. Both the Detroit and
Hutchinson Salt Mines remain committed to working
with MSHA to reduce DPM exposure levels for their
miners in the future and as can be seen by this short
history, additional control alternatives may have to be
implemented.
ACKNOWLEDGEMENTS
The authors would like to express their gratitude to
the Detroit Salt Company and Hutchinson Salt Com-
pany for their permission to publish this paper. Special
thanks to Mr. Pete Graham – Mine Manager of the
Detroit Salt Mine and Myron Marcotte – Mine Super-
visor with the Hutchinson Salt Company for all their
help with this paper.
REFERENCES
Hartman, H.L. 1995. Mutmansky, J.M., Ramani, R.V. and
Wang, Y.J., Mine Ventilation and Air Conditioning, third
Edition. New York: John Wiley & Sons.
Manos, E.Z. 2003. Detroit Salt Mine-Past and Future. Mining
Engineering February 2004.
Manos, E.Z. & Wala, A.M. 2004. A Ventilation Study of the
Detroit Salt Mine In R. Ganguli & S. Bandopadhyay (ed.),
Tenth U.S./North American Mine Ventilation Symposium;
Proc. Symp., Anchorage, Alaska, May 16–19 2004.
MSHA 2001. Diesel Particulate Matter Exposure of
Underground Metal and Nonmetal Miners. Federal Regis-
ter, Vol. 66, 19 January 2001 US Mine Safety and Health
Administration.
MSHA 2001. Joint MSHA/Industrial Protocol Base. Line
Diesel Particulate Sample.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

An integrated approach for managing diesel emissions controls for

underground metal mines

G.H. Schnakenberg, Jr.

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh,
Pennsylvania, USA

ABSTRACT: The National Institute for Occupational Safety and Health (NIOSH) has been working with
diesel emissions control technology and the underground mining industry for a number of years. Underground
isolated zone studies provided the emissions reduction performance of all of the alternative technologies currently
available. A NIOSH/MSHA guide is available which provides mines with information and a process for selecting
diesel particulate filters (DPF). Recent NIOSH experience suggests that more attention may be needed in
managing the issues associated with diesel emissions and their control. The author argues that a holistic approach
is needed which involves several key mine operations departments or disciplines by positing that workplace air
can be considered as the product of a complex manufacturing process and examining how that process could be
managed.

1 INTRODUCTION systems using wallflow monoliths of cordierite or sili-

For the past seven years, the National Institute for
Occupational Safety and Health (NIOSH) has been
working closely with the underground mining indus-
try, labor organizations, and the US mine enforcement
agency – the Mine Safety and Health Administration
(MSHA) – to assist in their understanding, develop-
ment, and implementation of the alternative technolo-
gies to control and measure workplace concentrations
of diesel exhaust contaminants. Control and mea-
surement of diesel particulate matter (DPM) became
the preeminent concern upon MSHA’s issuing pro-
posed rules, and in January 2001 the promulgation of
the two distinct diesel rules – one for underground
coal (30 CFR 72) and one for underground metal and
nonmetal mines (30 CFR 57) – that were designed
to limit the exposure of underground workers to
DPM. As one of its early efforts, the fledgling diesel
team at NIOSH Pittsburgh in 2000 produced an
overview of the currently available control technolo-
gies in which their operating characteristics, perfor-
mance characteristics, and limitations were described.
This document emerged as a NIOSH publication IC
9462 (Schnakenberg & Bugarski 2002b) and is avail-
able for download from the NIOSH mining web site.
The technologies that were covered included clean
engine technology, alternative fuels (synthetic diesel,
biodiesel, and water-fuel emulsions), diesel oxi-
dation catalysts, disposable filter elements (DFE),
and variously configured diesel particle filter (DPF)
con carbide. Some of these DPF systems, when engine
exhaust temperatures were favorably high, would pas-
sively burn off the DPM as it was collected; other
systems designed for application at prevailing exhaust
temperatures, required an active application of a pro-
cedure to burn off the collected DPM. The process of
burning off of the collected DPM in a filter is termed
regeneration.
In devising its DPM rule for metal and nonmetal
mines, MSHA assumed that DPF systems would be
necessary (66 Fed. Reg 2001, p 5713,). Because the
implementation of this technology was not simple and
because the entire concept of using filters was new to
most of the mining industry, MSHA asked NIOSH in
the fall of 2002 to help them develop a diesel particu-
late filter selection guide (DPFSG). The DPFSG was
placed on the MSHA and NIOSH web sites in January
2003 with a minor revision in May 2003. This guide
has received high praise from those with diesel control
technology experience.
During the period shortly after promulgation of
the rules, NIOSH diesel researchers at Pittsburgh
conducted several well controlled underground exper-
iments designed to determine the effects of available
control technology on diesel equipment emissions
with the focus on DPM. These experiments were per-
formed by operating a single unit with and without the
control technology in place as it performed a repeat-
able work cycle that closely emulated a typical task
performed by the vehicle. The equipment was run in

121

Copyright © 2006 Taylor & Francis Group plc, London, UK

 an isolated part of a mine where the incoming ven-
tilation rate and air quality were well controlled and
could be measured. In sum, the results of these tests
confirmed the beneficial effects of proper altitude
adjustment of engine fueling rate and power match-
ing for inby coal mine engines, and the laboratory
performance for DPM and elemental carbon reduc-
tions obtainable with alternative fuels, disposable filter
elements (DFE) and DPF systems. These tests also
confirmed that DPF systems that utilized Pt-catalyzed
washcoats to enhance regeneration also produce sub-
stantially elevated nitrogen dioxide (NO2 ) emissions
(Schnakenberg & Bugarski 2002a, Bugarski et al.
2006a, b). Conducting these tests did not, however,
throw much light on the issues concerning the imple-
mentation of these technologies into an actual produc-
tion situation. Although the implementation is quite
technology and mine specific, NIOSH is currently
executing a study to expose many of the implementa-
tion issues in order to assess their impact and develop
solutions.
The implementation issue study currently under-
way, as well as a reflection on the situations experi-
enced during the in-mine control technology testing,
has emphatically shown that engine maintenance to
keep emissions low is not a general or well understood
practice. Aberrant engine emissions owing to malad-
justment, use of an inappropriate replacement part, and
a general failure to comprehend the concept were evi-
dent in the underground experiments noted above. The
DPFSG recognized that emissions-based maintenance
is a necessary prerequisite for DPF implementation,
but assumed that most mines had the resources and
comprehension to readily adopt or had in place a suf-
ficiently adequate engine maintenance program that
controlled DPM emissions and as a result did not
present the details of such a program other than by
reference.
Additionally, the DPFSG stated that the manage-
ment of worker exposures to DPM emissions was not
the exclusive responsibility of equipment maintenance
department but required an integration and interac-
tion of several departments including health and safety,
ventilation, and production, in addition to equipment
maintenance. However, the DPFSG did not go into
much detail on this subject. This paper is an attempt
to provide these details.
2 AIR QUALITY MANAGEMENT
One must consider that the deployment of diesel pow-
ered equipment into an underground mine creates a
complex system involving several mobile and varying
emission sources emitting of a variety of contaminants
which are diluted and transported by the prevail-
ing ventilation at the point of emission (equipment
122
Copyright © 2006 Taylor & Francis Group plc, London, UK
tailpipe). To this system is added the worker who is
mobile (often coincident with a local source of con-
taminant) and is the consumer of the local contaminant
concentrations, which may be mitigated by an envi-
ronmental cab. One can take the view that this entire
system can be modeled as a “manufacturing” process
in which the final product is the air inhaled by the
worker, and that air is “produced” by the interactions
of the systems and process components mentioned.
Considered thusly, the air quality, as measured by
the level of diesel exhaust contaminants present in the
inhaled air, is the responsibility of the manager of this
“manufacturing process or plant” who must produce
acceptable air quality at the lowest possible cost. This
manager will have a team of process stream supervi-
sors or specialists who are responsible for controlling
the quality of their product as it is utilized and as
they individually contribute to the quality of the final
product.
Although the “factory” for the production of “con-
taminated” workplace air is well established by virtue
of operating just a single unit of diesel powered equip-
ment underground, it is unlikely that any management
system to control and assure the quality of workplace
air was established at the same time. Continuing with
the analogy, underground mining operations usually
have the requisite management departments for the
individual production processes contributing to the
final product, but they are not organized or focused on
integrating their efforts towards producing the work-
place air product; rather they are focused on other
objectives. Equipment maintenance, for example, is
focused on maximizing equipment availability and is
repair oriented. The tailpipe emissions, the very “raw”
ingredient so to speak of workplace air quality, are not
emphasized or treated as an important and necessary
output of their efforts.
3 DEPARTMENTAL ROLES
The initial task for a manager of an underground mine
utilizing diesel powered equipment is to recognize
that controlling worker exposures to diesel emissions
contaminants will require an integrated approach and
is not the sole responsibility of one department or
specialist. Because of the complexity and interactive
nature of the underground diesel operation system,
all departments or specialists which influence exhaust
contaminant concentrations must work together. In
many mines, this approach will be novel and most
likely require an effort to shift the culture and focus of
the affected departments. Mine management must be
aware of the effort required and must prepare convinc-
ing arguments to effect the required changes for they
have no other alternative if diesel emissions are to be
controlled efficiently and effectively.
 3.1 Engine maintenance and involved, and the dealing with diesel emissions
may be only one of the objectives. The details are too
The major contributor to workplace air quality, and a
extensive to go into in this paper. For example, there
negative one at that, is the tailpipe emissions from the
are several diesel exhaust contaminants – the gases
diesel equipment. It should be obvious that fresh air
CO2 , CO, NO, and NO2 , and diesel particulate mat-
is being contaminated by the diesel emissions and, as
ter (DPM) – each with a spatially varying emission
a consequence, lower emissions are better and con-
level as the equipment loads the engine differently as
sistent with the air quality goal. Thus maintaining the
it moves throughout the mine with different prevailing
tailpipe emissions of each piece of equipment to at or
airflows and each with its own acceptable air quality
below its prescribed level should be a major objective
level. The tailpipe emissions may be altered by control
of this department. Note that each piece of equipment
technology which affect an entire fleet (in the case of an
should be assigned a prescribed level for each of the
alternative fuel) or an individual piece of equipment
diesel contaminants. What these levels are will or may
as in the case of oxidation catalyst or exhaust filter.
have been set by the analysis performed by production
There will be a balancing act between the amount of
and ventilation departments in consort with the engine
ventilation that can be supplied to an area (and the qual-
emission specialist. However, it should be rather clear
ity of that air if it has been contaminated upstream)
from the cost of ventilation that in practice the max-
and the number of production vehicles that simulta-
imum level for a particular engine’s emissions will
neously occupy that area. The ventilation department
not be much greater than that obtained for a prop-
must work closely with the production department who
erly running engine including properly installed and
may want three trucks in a loading area when there is
operating control technology. When exhaust aftertreat-
only ventilation enough to support two.
ment systems are required, their maintenance is the
Additionally, ventilation allocation is planned on
responsibility of engine maintenance department as
paper, but must be carried out underground. Having
is the maintenance of any ancillary systems such as
proper ventilation at the required locations demands
exhaust back pressure gages or monitors and filter
that the hardware which supplies that air not only be
regeneration systems required for the proper operation
installed properly, but functionally verified and prop-
of aftertreatment systems.
erly maintained. Torn tubing must be replaced, fans
The above additional requirements may be consid-
kept running, etc. In coal or gassy non-metal mines,
ered to affect the “culture” of existing maintenance
ensuring and maintaining adequate ventilation to sup-
operations. The focus must shift from “doing what it
port diesel operation may be straightforward or at least
takes to keep an engine running” to “doing what it takes
has been a necessary and common practice because
to keep an engine running and running clean.” It is the
of the requirement to ventilate for methane. In most
responsibility of management to ensure that this lat-
metal and nonmetal mines, however, ensuring ade-
ter goal is understood, implemented and rewarded. It
quate ventilation for the control of diesel contaminants
is probably safe to state that, in general, the current
at all times and locations can be a novel and difficult
maintenance departments are not adequately staffed
challenge.
with personnel sufficiently trained to perform the tasks
As noted above, diesels emit a limited variety of
required nor are adequate resources allocated to pro-
contaminants. The ventilation engineer must know
vide the required equipment and training. It is quite
what each of these are and how they are affected
possible that finding and hiring suitable individuals to
by control technology. He/she must know the occu-
train is a problem in some areas of the country.
pational limits of the contaminants and work closely
Achieving the goals of engine maintenance can-
with the industrial hygienist or health and safety
not be done in isolation. The production department
department. The ventilation engineer should be aware
must understand that they cannot use a piece of
of the engine’s nameplate ventilation rate provided
equipment that runs but is not running cleanly. Their
for MSHA certified engines, know how this rate is
equipment operators need to be sensitive to their equip-
obtained, and know how to use such information to cal-
ment’s emissions and report any changes such as the
culate ventilation air quantities. The use of aftertreat-
development of visible black smoke.
ment device is not accounted for in the MSHA
ventilation rate determination unless in a rare instance
it is part of the engine. The almost ubiquitous use
3.2 Ventilation
of diesel oxidation catalysts (DOC) which effectively
The role and objective of the ventilation department is reduce CO and hydrocarbons (when they are working)
to provide dilution and transport of the diesel exhaust means that a low or undetectable level of workplace CO
contaminants released from the vehicle tailpipe so that is no assurance that ventilation is adequate. On the con-
the concentrations of these contaminants are within trary, the introduction of a new DOC or a platinum cat-
accepted limits. Despite the simple appearing objec- alyzed exhaust filter may temporarily or permanently
tive, the efforts needed to achieve it are highly complex increase tailpipe NO2 concentrations to the point of

123

Copyright © 2006 Taylor & Francis Group plc, London, UK

 requiring more ventilation air than that specified by The above only touches the surface of the complex
MSHA for the engine without the DOC control. role of the ventilation department. They must work
The ventilation engineer should also realize that one closely with those responsible for selecting new diesel
critical contaminant would remain if all other tailpipe powered equipment and control technology as well as
emissions were reduced to an insignificant level. That with production department who need to utilize the
contaminant is carbon dioxide. CO2 is an irreducible equipment and with the health and safety department.
product of the burning of fuel diesel fuel, and as such The position is so pivotal for DPM control that it seems
leads one towards the rule that there always must be as if the ventilation department should assume the
enough ventilation air to dilute the CO2 from the fuel lead role for developing the overall mine strategy for
being consumed. The monitoring of workplace CO2 controlling worker exposure to DPM.
concentrations is a little used but exceedingly use-
ful tool for assessing the adequacy of localized (or
general) ventilation to deal effectively with the local- 3.3 Health and safety department
ized (or general) burden of diesel equipment exhaust.
The health and safety department or industrial hygien-
Additionally, because CO2 is unaffected by emissions
ist (IH) performs the quality assurance function by
control technology yet is affected by ventilation along
measuring workplace air quality and personal expo-
with the other tailpipe contaminants emitted with it,
sures to ensure that the air quality meets expectations.
one can determine the effectiveness of a control strat-
They must also work very closely with the ventilation
egy by simultaneously measuring CO2 along with the
department to convey to them the workplace contami-
measurement or sampling of a controlled contami-
nant limits for exposure and the relative toxicological
nant such as elemental carbon. For example, the use
importance of each. The IH and safety officials should
of a diesel exhaust filter on one or more pieces of
be involved on an ongoing basis with those who are
equipment in a loading area should result in a dras-
responsible for selecting control technology. In the
tic reduction in the EC to CO2 ratio. The ratios are
process of considering a control technology, the IH
independent of ventilation.
needs to ask probing questions of the technology sup-
Since tailpipe emissions and ventilation are insep-
plier as to how and what emissions are affected and let
arably linked, the ventilation engineer must be cog-
the ventilation engineer determine the impact on ven-
nizant of vehicle deployment and utilization. The
tilation. The safety officials need to determine whether
relationship between fuel burned and DPM emissions
there are any safety issues with the technology.
must be assessed for every vehicle or vehicle class. It is
There is an obvious overlap between worker expo-
becoming more commonly known that the ubiquitous
sure measurements and workplace air quality measure-
utility vehicle using low horsepower but rather dirty
ments performed by the IH and ventilation department.
engines can be a major contributor to DPM burden
The objectives of each department’s measurements are
and exposures. It seems that the ventilation department
different and both are important. The measurements
should perform the engineering calculations which
should not be the exclusive purview of one department;
will prescribe engine emission limits and, if needed or
they should be coordinated when possible. Personal
possible, the appropriate control technology for these
exposure measurements by the IH could be consid-
vehicles. These limits are then transmitted to engine
ered as proof that the air quality “production system” is
maintenance department who perform the necessary
functioning properly; that is, tailpipe emissions, vehi-
tasks to maintain engines to the prescribed emission
cle deployment and ventilation together are controlling
levels.
worker exposures. If exposures are not acceptable, then
Since the objective of the ventilation department is
workplace or worker exposures to CO2 can be used to
to mitigate the effects of tailpipe emissions on work-
assess the adequacy of ventilation, and contaminant
place air quality, it will probably be their role to indi-
to CO2 ratios can be used to determine if control of
cate when the ventilation system cannot adequately
tailpipe emissions (a malfunctioning engine or con-
control DPM emissions to meet the compliance stan-
trol technology failure) is the cause for unacceptable
dard and the reduction of DPM emitted by the vehicle
exposures.
must be reduced using control technology. The selec-
tion of the most cost effective and appropriate control
technology can be quite a complex decision which
3.4 Production
involves fleet-wide solutions (alternative fuels) and/or
equipment-specific emission controls such as replac- The obvious objective of production department is to
ing the engine with a cleaner one or the addition of get the most rock to the surface as safely and quickly
an exhaust filter. Once the need to control equipment as possible by efficiently utilizing available equipment
emissions is realized, the development of a control and operators.After all, production is the money maker
strategy and selection of specific control technologies for the mine. Equipment availability is a prime concern
must be a collaborative effort of all departments. and thus production supervisors push the maintenance

124

Copyright © 2006 Taylor & Francis Group plc, London, UK

 department to service or repair equipment quickly.
They may push the limits of the ventilation by wanting
more pieces of equipment in an area than the local ven-
tilation can support. However, when production goals
cause engine maintenance and ventilation to less than
adequately fulfill their objectives, the underground
workers are the ones most affected by the decrease
in air quality.
The production department, as users of equip-
ment, must be involved in air quality management.
Their input on equipment deployment and utilization
is essential for designing the ventilation system and
equipment emission levels. On the other hand, the
equipment operators must be instilled with an appre-
ciation of the necessity to control air quality and DPM
in particular. Altering driving habits (such as a using
gentle application of the throttle) may reduce exces-
sive DPM emissions. Having the equipment operators
understand the importance of correct and continual
operation of exhaust filters would increase equipment
availability and filter longevity, for example. Equip-
ment operators should be encouraged to perform a key
function in detecting and reporting equipment mal-
function such as excessive smoke, inoperative back
pressure gages, etc.
The production department is most likely heavily
involved in decisions to rebuild old equipment and to
select new equipment for purchase. These decisions
should strive to obtain equipment using diesel engines
with the lowest DPM emissions possible. This goal
should be applied to any diesel-powered equipment
being purchased. The intended purchase should be dis-
cussed with MSHA engine specialists, and the mine’s
maintenance and ventilation departments.
The production department must be involved when
circumstances necessitate the utilization of diesel par-
ticulate filters because retrofitting of filters cannot
be characterized as a “fit and forget” solution. How
equipment is used, and the variability of use, greatly
affects the selection and ultimate performance of DPF
systems. DPF systems which passively regenerate
can be successfully applied to relatively few vehicles
and there is a consequence to ventilation from sec-
ondary emissions of NO2 . DPF systems which must be
“actively” regenerated – through either the application
of a regeneration cycle to the DPF when the equipment
is not in use or by a DPF exchange – have a significant
effect on operational logistics, mine infrastructure, and
equipment operator responsibility. The selection of fil-
ters obviously requires the input and agreement of all
departments.
4 SUMMARY
This paper has presented a limited discussion sup-
porting the need for an integrated interdisciplinary
125
Copyright © 2006 Taylor & Francis Group plc, London, UK
approach to control worker exposures to DPM in
underground mines, and that the control of workplace
air quality and worker exposure to diesel exhaust con-
taminants is a complex task and must be approached
holistically. Using a factory or production analogy
with the workplace air as the final product, the author
identifies the typical mining operations departments
involved in producing that product, discusses their
involvement and roles, and shows that they all need
to work together. It is the author’s view that the need
for such an approach is not widely recognized within
the metal and nonmetal underground mining industry,
yet such an approach is necessary in order to control
worker exposure to DPM to MSHA compliance lev-
els. Management needs to form an integrated team of
principals from equipment maintenance, ventilation,
health and safety, and production departments and
other areas which affect or are affected by measures
required to control air quality. The ventilation depart-
ment appears to be the most appropriate department
to lead and manage the coordinated effort.
REFERENCES
30 CFR 57. Safety and health standards–underground
metal and nonmetal mines. Code of Federal Regulations.
Washington, DC: U.S. Government Printing Office,
Office of the Federal Register.
30 CFR 72. Health Standards for Coal Mines. Code of Federal
Regulations. Washington, DC: U.S. Government Printing
Office, Office of the Federal Register.
66 Fed. Reg. 5706 (2001) Mine Safety and Health Adminis-
tration: 30 CFR 57, diesel particulate matter exposure of
underground metal and nonmetal miners; final rule.
Bugarski AD, Schnakenberg. GH Jr., Noll JD, Mischler SE,
Patts LD, Hummer JA & Vanderslice SE. 2006a. In-situ
Effectiveness of Selected Diesel Particulate Matter Con-
trol Technologies for Underground Mining Applications-
Part I. U.S. Department of Health and Human Services
Report of Investigation RI. In Press.
Bugarski AD, Schnakenberg. GH Jr., Mischler SE, Noll JD,
Patts LD & Hummer JA. 2006b. In-situ Effectiveness
of Selected Diesel Particulate Matter Control Technolo-
gies for Underground Mining Applications – Part II. U.S.
Department of Health and Human Services Report of
Investigation RI. In Press.
Schnakenberg GH & Bugarski AD. 2002a. Effects of Fuel-
ing Rate and Filters on PM Emissions from Diesel
Vehicles at High Elevations. In Proceedings of the Min-
ing Diesel Emissions Conference (MDEC). Markham,
Ontario, Canada: CANMET/NRCAN, 555 Booth Street,
Ottawa, Ontario K1A 0G1.
Schnakenberg G and Bugarski A. 2002b. Review of Tech-
nology Available to the Underground Mining Industry for
Control of Diesel Emissions. U.S. Department of Health
and Human Services Information Circular IC 9262.
[http://www.cdc.gov/niosh/mining/pubs/pdfs/ic9462.pdf]
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Implementation of diesel particulate filter technology in underground

metal and nonmetal mines

A.D. Bugarski, G.H. Schnakenberg, Jr. & L.D. Patts

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh,
Pennsylvania, USA

ABSTRACT: Achieving substantial reductions in the exposure of underground miners to diesel particulate
matter in a number of metal and nonmetal mines in the U.S. depends on the ability of the industry to widely
implement advanced diesel emissions control technologies, primarily diesel particulate filter (DPF) systems.
Recent field studies showed that diesel particulate filter systems with Cordierite and silicon carbide ceramic
filter elements are capable of reducing concentrations of diesel particulate matter and elemental carbon by
more than 70 and 90 percent, respectively. But those studies and several other attempts to implement DPF
systems in underground mines revealed a number of relatively unique technical and operational challenges that
are limiting industry-wide implementation of this technology. This paper provides detailed analysis of some of
those challenges and short overviews of several projects launched by NIOSH Pittsburgh Research Laboratory
in an attempt to provide the underground mining industry with more adequate DPF systems. These systems
employ advanced technologies to overcome the most pronounced challenge, the regeneration of DPF elements
installed on mining engines that generate relatively low exhaust temperature. This paper presents essential
findings obtained through a long-term evaluation of DPF system with diesel fuel burner technology installed
on a heavy-duty load-haul dump vehicle in an underground metal mine. In addition, the essential results of the
laboratory evaluation of a popular on-highway DPF system which was adapted and optimized for underground
mining applications will be summarized in this paper.

1 INTRODUCTION cases, solving these issues will require a multifaceted

Exposures of underground (UG) metal/nonmetal min-
ers to diesel particulate matter (DPM) are signif-
icantly higher than those of any other occupation
(Schnakenberg and Bugarski 2002). Compliance sam-
pling conducted by the Mine Safety and HealthAdmin-
istration (MSHA) in 178 mines between 2002 and
2005 (Cash and Baughman 2005) showed that UG
miners in precious metal mines are exposed to DPM
concentrations as high as 3300 µg/m3 . The magnitude
of those exposures was found to be directly related
to the extent of use of diesel-powered equipment and
the ability of mines to provide fresh air in the work
zones. Ever since MSHA promulgated rule 30 CFR
57.5060 limiting exposures of UG metal and non-
metal miners to DPM, the mining community has
been working on identifying technically and econom-
ically feasible strategies and technologies to reduce
exposures to particulate matter and noxious gases
emitted by diesel powered equipment. The challenge
is designing, optimizing, and implementing techni-
cal and economical strategies and technologies that
can accommodate requirements unique to UG min-
ing environments and operations. In the majority of
approach, a coordinated effort from several mine
departments, and integral solutions.
The strategies available to the UG mining industry
to reduce the exposure of miners to DPM can be gen-
erally divided into two groups: (1) strategies that are
designed toward controlling the DPM once it has been
released in the atmosphere; and (2) strategies that rely
on controlling DPM emissions before they are released
from the tailpipe.
The leading strategies for improvements in con-
trolling diesel emissions released in the atmosphere
are improvements in ventilation and more extensive
use of environmental cabs and personal protective
equipment (PPE). Ventilation is traditionally used to
control concentrations of various pollutants in UG
mines. The results of aforementioned MSHA com-
pliance sampling indicate a need for improvements
in ventilation in a number of U.S. metal and non-
metal UG mines. The first step toward making a
safer work environment for UG miners with respect
to diesel emissions is to provide sufficient quantities
of fresh air to dilute the emitted carbon monoxide
(CO), carbon dioxide (CO2 ) and nitric oxide (NO),
to bellow their respective 1973 American Conference

127

Copyright © 2006 Taylor & Francis Group plc, London, UK

 of Governmental Industrial Hygienists (ACGIH) time-
weighted average (TWA) threshold limit values (TLV)
and the nitrogen dioxide (NO2 ) below its 1973 ACGIH
TWA TLV and short term exposure limit (STEL) val-
ues. The 1973 ACGIH TWA TLVs and STELs are
currently used by MSHA to regulate the exposure
of underground metal and nonmetal miners (30 CFR
57.5001 1995). Diluting DPM emissions sufficiently
to reduce ambient concentrations below interim and
eventually final U.S. standards (30 CFR 57.5060)
requires that substantial additional quantities of fresh
air are supplied to the work zones in underground
mines (Schnakenberg 2001). Unfortunately, increased
ventilation capacity is linked to a substantial increase
in capital and operational costs (McPherson 1993).
Some local improvements can be achieved with more
effective management and distribution of available air
quantities.
Environmental cabs are successfully used in UG
and surface mining operations to control operator’s
exposures to dust, noise and some other pollutants.
Nevertheless, MSHA surveys conducted in metal and
nonmetal UG mines (30 CFR 57.5060) showed that
adequately designed and maintained air conditioned
and filtered environmental cabs might reduce personal
exposures of underground miners to DPM by as high as
80%. Unfortunately environmental cabs do not provide
protection to the workers outside of those enclosures.
Costs associated with retrofitting existing vehicles and
ordering new vehicles with environmental cabs are
relatively high.
According to the current regulations (30 CFR
57.5060) the third available option to control DPM
exposures, PPE should be only used to supplement
engineering and administrative controls when those
are found unfeasible or insufficient to reduce expo-
sures to DPM.
A number of in-mine studies (McGinn 2004,
Bugarski et al. 2005, Stachulak et al. 2005, Bugarski
et al. 2006a,b) showed that a viable alternative to con-
trolling concentrations of diesel pollutants released
in UG environments is to reduce DPM and gaseous
emissions from their diesel-powered fleet. In general,
the concentrations of pollutants emitted by current
UG mining diesel fleets can be significantly reduced
by reducing engine-out emissions and implementing
various exhaust aftertreatment technologies. In a num-
ber of instances an integrated approach is needed to
achieve desired emissions reductions.
The adopted strategy should preferentially address
vehicles that are highest contributors to the miner’s
exposure to DPM and gases. Typically, the largest con-
tributors are heavy-duty production vehicles such as
load-haul-dump (LHD) vehicles and haulage trucks
that are often operated in relatively poorly venti-
lated areas. The contribution of increasingly popular
light-duty vehicles, often powered by relatively high
128
Copyright © 2006 Taylor & Francis Group plc, London, UK
polluting engines, to the total DPM and gaseous con-
centrations in mine air should not be neglected (Rubeli
et al. 2003).
Instituting a comprehensive emissions-based main-
tenance program for heavy-duty and light-duty diesel-
powered vehicles should be the first step toward
reducing concentrations of particulates and gases
emitted by a diesel fleet (McGinn 2004). Establish-
ing the in-use emissions from an existing diesel fleet
is essential to the selection of an adequate emissions
control strategy, the optimization of selected control
technologies, and is a prerequisite for implementation
of advance diesel emissions control technologies. Peri-
odic emissions measurements should be also used to
evaluate the effectiveness of the implemented control
strategies and technologies.
Reductions in engine-out emissions can be also
achieved by the use of reformulated fuels. The effects
of water-fuel emulsions, blended biodiesel fuels, ultra
low sulfur diesel fuel, and #1 diesel on concentra-
tions of particulate matter and gases in mine air were
evaluated by Bugarski et al. (2006b) during isolated
zone testing at Stillwater Mining Company Nye Mine.
This study showed that using the Lubrizol PuriNOx
cold-weather and warm-weather water-fuel emulsions
reduced EC concentrations by approximately 70%
and 85%, respectively. The tested 20% and 50% soy
biodiesel blends showed EC reductions of 49% and
66% respectively. The reductions were slightly less
pronounced for the 20% and 50% yellow grease
blends, 33% and 56% respectively. Several attempts to
implement alternative fuels in underground nonmetal
mines revealed problems related to compatibility of
certain fuels and engine components.
Although substantial reductions in exposures of
underground miners to DPM in a number of metal and
nonmetal mines in U.S. can be achieved implement-
ing the aforementioned strategies and technologies,
the best results can be achieved by complementing
those strategies with a widespread implementation
of advanced aftertreatment diesel emissions control
technologies. Such control technologies include diesel
particulate filter (DPF) systems and filtration sys-
tems with high-temperature disposable filter elements
(DFE). Recent field studies (Bugarski et al. 2006a,b)
showed that diesel particulate filter systems with
Cordierite and silicon carbide wall flow monolith fil-
ter elements are capable of reducing concentrations
of DPM and EC in mine air by more than 70%
and 90%, respectively. The same study showed that
the disposable filter elements (DFE) from Donaldson
and Filter Services reduced the vehicle EC contribu-
tion by 92% and 70%, respectively. The same DFEs
reduced downstream TPM concentrations determined
using gravimetric analysis by 84 and 58%, respec-
tively. Despite their very high effectiveness in removal
of EC and TPM, DPF systems have not been widely
 accepted by the UG mining industry. Various imple-
mentation issues were found to limit industry-wide
implementation of this technology.
2 DPF IMPLEMENTATION ISSUES
The high demand for aftertreatment systems for the
efficient control of particulate emissions from on-
highway and off-road diesel-powered vehicles resulted
in a rapid development of DPF technology. Two com-
prehensive long-term studies sponsored by the Diesel
Emissions Evaluation Program (DEEP), one con-
ducted at Noranda’s Brunswick Mining and Smelting
Mine (McGinn 2004) and one at the International
Nickel Company’s Stobie Mine (Stachulak 2005)
demonstrated the potential of modern DPF systems
for controlling DPM emissions in underground mines.
Those studies also offered insight into the prob-
lems associated with the use of those systems on
underground mining vehicles. Short-term studies con-
ducted in SMC Nye mine (Bugarski et al. 2006a,b)
also revealed a number of technical and operational
challenges unique to UG mining applications.
The relatively large number of different types of
diesel-powered vehicles currently operated in UG
mines and the diversity of mining methods complicate
the selection and implementation of this technology.
Implementation is complicated by the fact that DPF
systems are currently available to UG mining indus-
tries only as aftermarket retrofits for existing and
newly introduced vehicles. Retrofitting aftermarket
DPF systems on a wide range of UG mining vehi-
cles requires a relatively large level of optimization
and customization for individual applications.
One of the major technical difficulties associated
with DPF systems is filter medium regeneration,
which involves periodic oxidizing or burning of the
accumulated DPM on the filter. In general, the filter
can be regenerated using heat and temperature gener-
ated by the engine or by using heat and temperature
provided by various external sources of energy. Sys-
tems which are using the former concept are known
as passive systems. The latter concept is used in
active systems. Due to their relatively lower complex-
ity and operational and logistics requirements, passive
DPF systems are the more desirable for underground
operations.
Surveys of duty cycles for UG mining vehicles
conducted across the mining industry revealed that
temperature generated by engines that power a maj-
ority of light-duty and large a number of heavy-duty
vehicles over regular duty cycles is insufficient to
support passive regeneration of currently available
uncatalyzed, and quite often catalyzed, filters. Data
indicates that mining vehicles spend a significant per-
centage of time at engine operating conditions which
129
Copyright © 2006 Taylor & Francis Group plc, London, UK
do not favor passive regeneration. Data obtained at
Noranda’s Brunswick Mining and Smelting show that
their vehicles spend an average of over 30% of time
at low idle (McGinn 2004), engine operating mode
that produce DPM emissions but does not support
regeneration.
Several UG metal mines are currently operating
heavy-duty vehicles such as LHDs and haul trucks
with platinum coated passive DPF systems. Platinum-
based catalyst formulations are found to be efficient
in facilitating passive regeneration of filters by lower-
ing the regeneration temperature from approximately
550◦ C (uncatalyzed DPF) to approximately 330◦ C
(Jelles et al. 1999). Those formulations are also effi-
cient in lowering CO and hydrocarbon emissions by
oxidizing them to CO2 . Unfortunately they are also
promoting conversion of NO to NO2 (MSHA 2002).
Isolated zone studies conducted by Bugarski et al.
(2006a,b) showed a two or even three fold increase
in emissions of NO2 when certain types of platinum
catalyzed DPF and DPF/DOC systems were used. The-
oretically, vehicles with platinum coated DPF systems
should be operated only in areas of UG mines where
sufficient quantities of fresh air are supplied to dilute
NO2 concentrations to or below a ceiling limit of 5
ppm as established in MSHA standards at 30 CFR
57.5001.
The alternative solution is to use active DPF systems
which are designed to collect DPM over a prede-
termined period of engine operation, e.g. one shift.
Depending on design, regeneration is performed with
the filter element remaining on-board of the vehicle or
with it removed from the vehicle. The on-board regen-
eration process commonly uses an electrical heater
but can use a diesel fuel burner. The regeneration of
such systems can only occur with the vehicle parked
with the engine off. Again depending on system
design, regeneration can take anywhere from a half-
hour or less (fuel burner) to eight hours. The off-board
regeneration commonly uses electrical heaters or kilns
designed for the purpose.
Stachulak et al. (2005) found that active systems
with off-board and on-board electrically regenerated
systems can be successfully used to control DPM
emissions from light- and heavy-duty UG mining vehi-
cles. The authors attributed difficulties with operating
on-board regenerated system in underground mine to
human error and stressed the importance of discipline
and education. A majority of other mine operators
found on- and off-board electrically regenerated sys-
tems to be too complex, delicate, and expensive to be
widely implemented in underground operations.
Therefore, there is an apparent need for efficient
DPF systems which will satisfy the following criteria:
(1) regeneration spontaneously regenerates at rela-
tively low exhaust temperatures; (2) does not produce
secondary emissions of NO2 and other pollutants;
 (3) operates “transparently” to the operator; and (4)
is robust and reliable.

3 OPTIMIZATION OF DPF SYSTEMS

FOR UNDERGROUND MINING
APPLICATIONS

NIOSH PRL has partnered with selected manufactur-

ers of control technologies and the mining industry to
develop and optimize several promising DPF concepts
for UG mining applications and therefore facilitate the
implementation of DPF technology in underground
mines. This paper describes two of those projects
which may provide mining industry with DPF systems
that satisfy the aforementioned criteria.

3.1 ArvinMeritor fuel burner DPF system Figure 1. ArvinMeritor DPF system installation.
The objective of the first project, conducted in coop-
eration with ArvinMeritor (AM), Commercial Vehicle
Emissions, Troy, MI, was the field evaluation of a pro-
totype DPF system that uses a diesel fuel burner to
periodically increase the temperature of the full flow
of diesel exhaust in order to initiate and support filter
element regeneration.
Under NIOSH contractAM designed and fabricated
the prototype fuel burner system for a heavy-duty pro-
duction LHD from the underground mining fleet at
the Stillwater Mining Company’s (SMC’s) Nye Mine
located in Nye, MT. The vehicle was powered by a
modern electronically-controlled turbocharged Deutz
BF4M1013 FC MVS engine. The system shown in
Figure 1 consisted of a vertically mounted uncat-
alyzed ceramic wallflow monolith filter element, the
DuraTrap RC (Corning, NY), followed by a metal sub-
strate diesel oxidation converter (DOC). The DOC was
Figure 2. ArvinMeritor DPF system control box.
designed to control CO and hydrocarbon emissions
during normal vehicle operation and the regenera-
tion process. The original DOC had a platinum (Pt)
based formulation wash-coated onto a substrate. After long-term evaluation. The objective of the study was to
approximately 470 operating hours, the Pt DOC was evaluate the suitability of the AM DPF system for con-
replaced with a similar DOC with a palladium (Pd) trolling diesel particulate matter emissions generated
based washcoat. The computer controlled diesel fuel during heavy-duty underground mining applications.
burner is integrated on the inlet side of the DPF The long-term field evaluation provided an opportu-
element. nity to evaluate various aspects of implementation of
The control box with fuel pump, air blower, and this DPF system, such as installation and regenera-
controller is mounted in the engine bay (see Figure 2). tion, in an underground mining application. During
When activated, the burner ignites diesel fuel injected the evaluation the fuel burner, acting automatically
into the combustion chamber to heat the engine at preset intervals, successfully regenerated the filter
exhaust and filter element to about 650◦ C, a tem- element.
perature sufficient to support DPF regeneration. The After 470 hours in production the system was tested
regeneration process was initiated at predetermined in the isolated zone at the Nye Mine. The objective of
time intervals. this testing was to establish the effects of two configu-
The system was installed on the LHD in January rations of the AM DPF/DOC system on concentrations
2004. After the installation the vehicles was returned of aerosols and selected gases in mine air. Three tests
into production and DPF system was subjected to were conducted, two with tested configurations (DPF

130

Copyright © 2006 Taylor & Francis Group plc, London, UK

 with Pt- and Pd- DOCs) and a baseline in which the
host LHD was equipped with a muffler. For each of
the tests the vehicle was operated over the same sim-
ulated duty cycle and at a similar ventilation rate. The
vehicle was fueled with #1 diesel supplied by a local
refinery. This higher quality fuel was used by the mine
in underground operations as a part of their strategy to
reduce the exposure of miners to diesel emissions. The
effects on tailpipe emissions were estimated from the
results of aerosol and gas measurements performed at
the upstream and downstream end of the isolated zone
and at the vehicle. Several methods and instruments
were used to determine concentrations of the aerosols
in the incoming fresh air at the upstream station and
in the contaminated air at the downstream and vehicle
sampling stations. The gas monitors were used for real-
time measurements and recording of concentrations of
CO, CO2 , NO, and NO2 at all three sampling locations.
The details on the methodology and instrumentation
used during this test are available in Bugarski et al.
(2006b). The effects of each configuration were quan-
tified by comparing the contaminant concentrations
observed for the test when the vehicle was operated
with the test configuration to those observed for the
baseline test.
With the exception of some difficulties in the devel-
opment phase the operation of this system was found
to be relatively transparent to the vehicle operator.
Despite the complexity this prototype system was
found to be relatively reliable.The system required reg-
ular maintenance and couple of unscheduled repairs.
The isolated zone testing revealed undesired effects
of the Pt catalyzed DOC on the tailpipe emissions of
NO2 . The test on the AM DPF system with the Pt
catalyzed DOC had to be terminated prematurely to
prevent an overexposure of the vehicle operator to
NO2 . During the initial stages of that test, the gas
monitor at the vehicle sampling station recorded con-
centrations of NO2 that were close to, or exceeded, the
1973 ACGIH TLV of 5 ppm. This limit is currently
used by MSHA as a Ceiling Limit to regulate expo-
sure of underground metal and nonmetal miners to
NO2 (30 CFR 57.5001). The Pt-catalyzed DOC was
then replaced with one with a Pd-based catalyst. The
AM DPF with the Pd DOC raised the average and
peak downstream NO2 concentrations approximately
30%, yet at the prevailing ventilation rate of 19 m3 /s
(40,258 ft3 /min) used in the tests, the concentrations of
NO2 were well below 1973 ACGIH TLVs and the test
was successfully completed. This test was extended for
an additional three full cycles to assess the effects on
emissions during the diesel fuel burner regeneration
process.
The tests showed that the AM fuel-burner DPF
system with a Pd catalyzed DOC reduced the EC con-
centrations by 92%. The same system reduced down-
stream concentrations of TPM by 69% (gravimetric
131
Copyright © 2006 Taylor & Francis Group plc, London, UK
measurements) and by 72% (TEOM 1400a measure-
ments). The Pd-catalyzed DOC increased the sulfate
concentration from 3.4 to 15.3 µg/m3 . Particle size dis-
tribution results indicated a 105% increase in the total
particle number, possibly corroborating the hypothe-
sis of increased sulfate aerosol concentrations. During
the regeneration of the AM DPF system, the peak
aerosol number concentration increased from about
1,666,000/cm3 , observed during the normal operation,
to approximately 3,177,000/cm3 .
NO concentrations at the downstream and vehicle
sampling locations were found to be generally lower
during tests with the AM DPF having Pt and Pd DOCs.
The slight reduction in NO is partially attributable to
the conversion of NO to NO2 by the DOCs.
The net contributions to the concentration of CO at
the downstream and vehicle sampling stations of the
LHD, equipped with the AM DPF with either a Pt or
Pd DOC were found to be negligible. CO reductions
can be attributed to the DOC’s catalytic processes.
The system was successfully operated until March
2005. In fall 2004 ArvinMeritor decided to stop pro-
viding this system as a retrofit and to offer this system
only as part of OEM solutions.
The DPF system with somewhat similar fuel burner
concept is also offered by HUSS Umwelttechnik
GMBH, Nürnberg, Germany. This system is using
an on-board diesel fuel burner to regenerate a silicon
carbide monolith element. In contrast to the AM sys-
tem, the regeneration is performed while the engine is
turned off. However, it requires less than 30 minutes to
perform and can be done anywhere there is normally
adequate ventilation (Rembor 2005). Thus regenera-
tion can be performed at shift change or “lunch-time”
periods. It appears to be suitable for retrofit applica-
tions for light- and heavy-duty UG mining vehicles
and hence worthy of serious consideration.
3.2 Low NO2 Johnson Matthey Continuously
Regenerating Trap system
Johnson Matthey (JM) Royston, England developed
a Continuously Regenerating Trap (CRT) system to
control DPM emissions from on-highway trucks and
urban buses (O’Sullivan et al. 2004). The core of
the system is a Cordierite filter element. A platinum
catalyst upstream of the filter is used to promote con-
version of NO to NO2 . The exhaust gases, enriched
with NO2 , pass to the filter element where they oxidize
the collected particulate matter, thereby regenerating
the filter. Since this reaction occurs at much lower tem-
peratures than the reaction between particulate matter
and oxygen, it is capable of passively regenerating
when temperatures exceed approximately 250◦ C for at
least 40% of the operation time. When the NO2 reacts
with the particulate matter, the products formed are
carbon dioxide and NO. Inevitably, some NO2 passes

Figure 3. Johnson Matthey CRT system.
Figure 4. Johnson Matthey CRT system control box.
through the filter (Ayala et al. 2001, O’Sullivan et al.
2004). This process, commonly known as NO2 slip,
has prevented widespread application of the CRT sys-
tem in underground mining, where the low exhaust
temperatures needed for regeneration would be greatly
appreciated.
JM engineers (O’Sullivan et al. 2004) proposed two
potential solutions to the problem. The first solution
uses an urea-based selective catalytic reduction (SRC)
system to control both NO and NO2 emissions. The
second solution uses hydrocarbon (HC) injection onto
a catalyst element after the filter which decomposes
the NO2 . The former solution was deemed too complex
and delicate for UG mining applications. The latter
concept was adopted and pursued further.
NIOSH PRL sponsored construction of the proto-
type system shown in Figure 3. The NO2 decompo-
sition catalyst is shown on the right-hand side of the
picture.
The operation of the system is controlled continu-
ously using a HC injection controller that is designed
to be mounted on-board of the vehicle (see Figure 4).
The fuel is injected at the predetermined rates at exh-
aust temperatures above 250◦ C. The controller box
contains the fuel pump, air compressor, and controller.
132
Copyright © 2006 Taylor & Francis Group plc, London, UK
The test cell at University of Minnesota Center for
Diesel Research Laboratory was used to optimize and
evaluate this system. The system was tested using a
1999 Cummins ISM engine rated at 370 hp. An
eddy-current dynamometer was used to test the engine
using the ISO 8178 8-Mode Steady State Test and
Inco Transient Cycle. The engine was fueled with the
required ultra low sulfur fuel. This fuel, supplied by
Consumers’ Co-operative Refineries Ltd. Saskatoon,
Saskatchewan, had average sulfur content less than
15 ppm by weight.
The particulate matter and gaseous emissions were
established for three test configurations: (1) engine
only, (2) engine equipped with JM CRT system, and
(3) engine equipped with JM CRT system with HC
injection and NO2 decomposition catalyst. The emis-
sions from the JM CRT system were used to optimize
the rate of hydrocarbon injection. After the system was
optimized the effects of hydrocarbon injection on NO2
slip from the CRT system were evaluated.
The effects of the three configurations on the con-
centrations of NO, NO2 , CO, as well as non-methane
hydrocarbons (HC), and the effects on the mass emis-
sions of diesel PM (gravimetric) and elemental carbon
(NIOSH 5040 analysis) were established using an
array of instrumentation.
The tests with the JM CRT system (Zarling et al.
2005) showed that the system reduced on average DPM
emissions by more then 90% and EC emissions by
95%. At the several test modes the system substantially
increased fraction of NO2 in total NOx emitted by the
test engines. When engaged the HC injection and NO2
decomposition catalyst reduced NO2 emissions out of
the CRT system below those produced by the engine
alone. The remaining issue is control of NO2 slip at
exhaust temperatures below 250◦ C.
After completion of laboratory development and
evaluation, the JM CRT system with HC injection will
be subjected to long-term field evaluation. The system
will be installed on a heavy-duty underground mining
vehicle and evaluated in production for at least 1500
hours.
4 CONCLUSION
Implementation of advanced DPF technology plays
an important role in efforts to reduce the exposure
of UG miners to DPM in a number of U.S. metal
and nonmetal mines. Developing and optimizing DPF
systems for specific UG mining applications are essen-
tial for successful implementation of this relatively
novel technology in underground mines. Two conce-
pts described in this paper showed that it is possible
to design DPF systems that will be able to operate
and regenerate at relatively low exhaust temperatures
without introducing the risk of exposing miners to
elevated NO2 concentrations. Field evaluation of the
 prototype AM system showed that engineering of a
rather complex DPF system with a fuel burner to a
reliable and dependable system transparent to a vehi-
cle operator requires concerted efforts of both the
filter supplier and mine operator. If successfully com-
pleted, the JM CRT system with NO2 slip control will
provide the mining community with a much-needed
viable solution for controlling DPM emissions without
an adverse increase in NO2 production. The findings
from these two studies should facilitate the imple-
mentation of DPF technologies in underground mines
and ultimately lead to reducing miners’ exposure
to DPM.
REFERENCES
30 CFR 57.5001. Safety and health standards for under-
ground metal and nonmetal mines. Exposure limits for
airborne contaminants. Code of Federal Regulations.
Washington, DC: U.S. Government Printing Office,
Office of the Federal Register.
30 CFR 57.5060. Diesel particulate matter exposure of
underground metal and nonmetal miners. Limit on con-
centration of diesel particulate matter. Code of Federal
Regulations. Washington, DC: U.S. Government Printing
Office, Office of the Federal Register.
Ayala, A., Kado, N. & Okamoto, R. 2001. ARB Study of
emissions from late-model diesel and CNG heavy-duty
transit buses. California Air Resources Board.
Bugarski, A., Schnakenberg, G. Jr., Noll, J., Mischler S.,
Crum, M. & Anderson, R. 2005. Evaluation of diesel
particulate filter systems and biodiesel blends in an
underground mine. Society for Mining, Metallurgy, and
Exploration Transactions 2005, Vol. 318, pp.27–35.
Bugarski, A.D., Schnakenberg, G.H. Jr., Noll, J.D.,
Mischler, S.E., Patts, L.D., Hummer, J.A. &
Vanderslice, S.E. 2006a. In-situ effectiveness of selected
diesel particulate matter control technologies for under-
ground mining applications- Part I. U.S. Department of
Health and Human Services. Report of Investigation. In
Press.
Bugarski, A.D., Schnakenberg, G.H. Jr., Mischler, S.E., Noll,
J.D., Patts, L.D. & Hummer, J.A. 2006b. In-situ effective-
ness of selected diesel particulate matter control tech-
nologies for underground mining applications – Part II.
U.S. Department of Health and Human Services Report
of Investigation. In Press.
Cash, D.A. & Baughman, W. 2005. DPM exposure in metal
and nonmetal mines in the United States 2002–2005 and
the 2005 final rule on the interim limit. Mining Diesel
Emissions Conference (MDEC) 2005, Markham, Ontario,
October 12–14.
DieselNet 2005. Diesel filter regeneration. Diesel particulate
filters. DieselNet Technology Guide. [URL=http://www.
dieselnet.com/tech/dpf_reg.html].
133
Copyright © 2006 Taylor & Francis Group plc, London, UK
Jelles, S.J., Makkee, M., Moulijn, J.A., Acres, G.J.K. & Peter-
Hoblyn, J.D. 1999. Diesel particulate control. Application
of an acticated particulate trap in combination with fuel
additives at an ultra low dose rate. SAE Paper 1999-010-
0113.
McGinn, S. 2004. Noranda Inc. – Brunswick Mine diesel
particulate filter (DPF) field study, Final report of inves-
tigation to the Diesel Emissions Evaluation Program.
[http://www.deep.org/reports/nordpf_final.pdf].
McGinn, S. 2004. The relationship between diesel engine
maintenance and exhaust emissions. Final report to Diesel
Emissions Evaluation Program. [http://www.deep.org/
reports/mtce_report.pdf].
McPherson, M.J. 1993. Optimum size of airway. Subsur-
faceVentilation and Environmental Engineering: 307–308
London: Chapman & Hall.
MSHA 2002. Potential health hazard caused by platinum-
base catalyzed diesel particulate matter exhaust filters.
Program Information Bulletin No. P02-4. May 31 [URL=
http://www.msha.gov/regs/complian/PIB/2002/pib02-04.
htm].
NIOSH 1999. Elemental carbon (Diesel Particulate): Method
5040, Issue 3 (Interim). In NIOSH Manual of analyt-
ical methods. 4th rev. ed. Cincinnati, OH: U.S. World
Wide Web [URL=http://www.cdc.gov/niosh/nmam/pdfs/
5040f3.pdf].
O’Sullivan, R., Stevenson, J., Walker, A. & Werth, P. 2004.
Low NO2 CRT particulate filter systems for diesel engine
exhaust in mining applications. Proceeding of Mining
Diesel Emissions Conference 2004, Toronto, Canada,
October.
Rembor, H.J. 2005. Diesel particulate filters with active
regeneration in mining – Practical experiences from mines
in North America and Europe. Proceeding of Mining
Diesel Emissions Conference 2005. Markham, Ontario,
Canada, October 12–14.
Rubeli, B., Gangal, M., Butler, K. & Aldred, W. 2003.
Evaluation of the contribution of light-duty vehicles to
the underground atmosphere diesel emissions burden,
Phase II. Final Report to Diesel Emissions Evaluation
Program. [http://www.deep.org/reports/ldv2.pdf].
Schnakenberg, G.H. Jr. 2001. Estimate of technically feasible
DPM levels for underground metal and nonmetal mines.
Mining Engineering: 45–51.
Schnakenberg, G.H. Jr. & Bugarski, A. 2002. Review of tech-
nology available to the underground mining industry for
control of diesel emissions. U.S. Department of Health and
Human Services Information Circular IC 9262. [http://
www.cdc.gov/niosh/mining/pubs/pdfs/ic9462.pdf].
Stachulak, J.S., Conard, B.R., Bugarski, A.D. &
Schnakenberg, G.H. Jr. 2005. Long–term evaluation of
diesel particulate filter systems at Inco’s Stobie Mine. Pro-
ceedings of 8th International Mine Ventilation Congress,
Brisbane, Queensland, Australia, July 6–8.
Zarling, D., Waytulonis, B. & Kittelson, D. 2005. Testing a
Low NO2 CRT DPF system. Proceeding of Mining Diesel
Emissions Conference 2005. Markham, Ontario, Canada,
October 12–14.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© Her Majesty the Queen in Right of Canada, as represented by the Minister of
Natural Resources, 2006, ISBN 0-415-40148-8

Post-field evaluation of diesel particulate filters

M. Gangal, B. Rubeli & D. Young

Natural Resources Canada, CANMET-MMSL, Ottawa, Ontario, Canada

J.S. Stachulak
Inco Limited, Copper Cliff, Ontario, Canada

ABSTRACT: The paper discusses the laboratory performance evaluation of diesel particulate filters (DPF)
after a long-term field trial at INCO’s Stobie underground mine in Sudbury. These DPFs, used at the mine with
heavy-duty production and non-production vehicles, were evaluated at the Diesel Emissions Research Labora-
tory of CANMET-MMSL in Ottawa. The laboratory evaluation included emissions bench testing, particulate
characterization, and DPF inspection for internal and/or external damages. The laboratory testing utilised three
types of protocols, (1) ISO 8178-C1 8-mode, (2) 3-mode test cycle, and (3) the SAE J1667 transient snap
acceleration. During testing, measurements were made for gaseous emissions (CO, CO2 , NO, NOx , and THC),
and particulates by mass and size, and PAS (Photoelectric aerosol sensor) by using a NanoMet Instrument. The
DPFs were inspected by visual, x-ray and borescope methods. All the DPFs tested significantly reduced NO2
(varying from 38% to 67%) and diesel particulate matter (greater than 85% on mass basis, greater than 93%
on number basis, and greater than 74% based on the PAS measurement). The lowest efficiency was measured
for the HD213A DPF, perhaps due to some damage in mine-field trials. All DPFs significantly reduced SOF,
PAH and sulphates. In general, laboratory measured DPF efficiencies are similar to or a little lower than those
measured during field trials.

1 INTRODUCTION the application of high temperature diesel particulate

Diesel equipment has been very practical and useful
to the productivity of mining operations. The perfor-
mance advantages of diesel equipment in terms of
durability, flexibility and productivity are well under-
stood. Heavy duty diesel equipment has relieved the
worker of a great deal of physical labour and stress,
while light duty vehicles have improved mobility in
deep and large underground mines. During its opera-
tion, diesel equipment produces many toxic gases and
particulates. However, diesel particulate matter (DPM)
is considered to be a potential human carcinogen, and
therefore it must be controlled.
There has been a number of technologies available
to reduce the concentration of toxic gases and par-
ticulates in mines (Schnakenberg & Bugarski 2002).
In underground operations, dilution of toxic gases by
supplying fresh air has been a common technique to
reduce pollutants. But, the cost of ventilation increases
as mines get deeper. Furthermore, in cold winter
months the fresh air is heated at great expense. There-
fore, the most effective way to control diesel emissions
is at the source, i.e. before it is released in the environ-
ment from the vehicle tailpipe. One such technology is
filters (DPF), whichcan reduce DPM by over 85%.
Earlier field studies to evaluate DPF technologies
in Canadian underground mines were conducted in
the early nineteen eighties and have been summarized
elsewhere (Dainty et al. 1986). One of the prob-
lems these earlier application had was the lack of
DPF regeneration. The regular regeneration of DPF
is essential, otherwise the engine exhaust backpres-
sure will increase to unacceptable levels, and could
damage the engine. The DPM is combusted at about
550◦ C, therefore for DPF regeneration the exhaust
temperature during regular operation should be at
this level. However, in many underground applications
such a temperature during normal vehicle operation is
not achievable. Nevertheless there are other regener-
ation techniques available (e.g. fuel borne catalysts,
catalyzed particulate filters, on-board fuel burners,
electric heaters) that could assist in regeneration.
In 1997, a Diesel Emissions Evaluation Program
(DEEP 1997) research consortium was formed by
the collaboration of industry, labour, government and
researchers, with a goal of reducing underground
miner’s exposure to diesel emissions. Two of the
DEEP studies included extensive in-mine field trials to

135

Copyright © 2006 Taylor & Francis Group plc, London, UK

 evaluate diesel particulate filter technologies in under-
ground mines (McGinn 2004 & Stachulak et al.
2005a). Prior to installation at the mine, all DPFs were
carefully selected to represent various design, regener-
ation concepts, suitability and availability. One of the
objectives of the program was to evaluate these DPFs
in a controlled laboratory environment at the end of
field trials. At the end of the long term evaluation of
DPFs at Inco’s Stobie mine, the six DPFs (four from
heavy-duty and two from light-duty vehicles) were
cleaned and shipped to the Diesel Emissions Research
Laboratory of CANMET-MMSL in Ottawa. One of
the heavy-duty vehicle DPFs was damaged and could
not be bench tested for emissions. The performance
evaluation of other three DPFs used with heavy-duty
vehicles is presented in this paper. The performance
evaluation of DPFs for light duty vehicles is described
elsewhere (Gangal 2005).
2 DIESEL PARTICULATE FILTERS
The relevant details on the three DPFs discussed in
this paper are given in Table 1. All of these DPFs in
the field were on Wagner STB8 heavy-duty LHDs.
The vehicle #820 powered with a Deutz F12L413FW
engine was primarily used to carry out a broad range
of non-production support tasks. The vehicle had
dual exhaust with two vertical DPFs installed, one
on each of the vehicle’s exhaust bank. For the labora-
tory evaluation only one of these DPFs was evaluated.
The vehicle #213 powered with a DDEC Series 60
engine was primarily used for production activities.
This vehicle had a single exhaust bank with two ver-
tical DPFs installed in parallel. For this reason both
DPFs from this vehicle were evaluated. All these
DPFs are made of silicon carbide, and during the
field trial used on-board electric heaters for regener-
ation purpose. It should be noted that vehicle #820
also used cerium-Eolys as a fuel borne catalyst for
regeneration.
Table 1. Details on diesel particulate filters.
DPF HD820 HD213A/HD213B
Number of DPFs Two vertical Two in parallel, vertical
DPF brand JMC ECS Unikat
DPF model DPF 201 Combifilter C18
DPF hours 2138 2084
in field
Engine make Deutz DDEC Series
and model F12L43FW 60,11.1 L
Engine power, 207 213
KW
Vehicle # 820 213
136
Copyright © 2006 Taylor & Francis Group plc, London, UK
3 EXPERIMENTAL
The DPFs were evaluated at the Diesel Emissions
Research Laboratory of CANMET-MMSL. This labo-
ratory was established in the early 70’s and has evolved
over the years to provide diesel emissions expertise for
mines, and emissions evaluation services for off-road
diesel engines and emissions control technologies. The
diesel research facility is registered to ISO 9001:2000
standards. In addition, this test facility is recognized
as an Accredited Testing Laboratory by the Standards
Council of Canada to comply with the requirements of
ISO/IEC 17025 for specific tests.
For this study, DPF performance evaluation was
divided into three sections, (1) engine bench testing
for gaseous and particulate emissions, (2) characteri-
zation of particulate matter, and (3) inspection of DPFs
for internal and/or external damages.
3.1 Engine bench testing
A series of tests on an engine dynamometer were con-
ducted in a controlled laboratory environment. During
the tests, engine parameters and regulated exhaust
gas emissions were measured using standard equip-
ment and methods (ISO 1996a). This ISO standard
specifies instrument calibration requirements, mea-
surement and evaluation methods for gaseous and par-
ticulate exhaust emissions from reciprocating internal
combustion engines on a test bench.
The DPFs were evaluated using the following three
types of test cycles on the engine dynamometer:
3.1.1 8-mode test cycle
This cycle is a steady state test with 8-mode points
as per the ISO 8178-C1 requirements (ISO 1996a,b).
This test cycle is commonly used world wide for off-
road vehicles and industrial equipment. Test modes
and weighting factors for this test cycle are given in
Table 2.
The integrated value for the 8-mode cycle is calcu-
lated by using the above weight factor for each mode,
while an average value is calculated by taking the
arithmetic mean of the 8-mode values.
3.1.2 3-mode test cycle
This type of test cycle is mostly used in the field for the
purpose of engine maintenance and/or performance
Table 2. ISO 8178-C1 8-mode test cycle.
Mode 1 2 3 4 5 6 7 8
Speed Rated speed
Intermediate speed Low
idle
Torque % 100 75 50 10 100 75 50 0
Weight .15 .15 .15 .10 .10 .10 .10 .15
factor
 evaluation of devices, where laboratory test equipment
is not available or cannot be used. These modes are
described below:
– FTC, Full torque converter stall, engine speed is
taken at 93% of the rated speed, load is full (in
field the vehicle service brake is set and wheels are
chocked, engine is loaded against torque converter
with its transmission in the highest gear)
– HI: High idle speed, no load
– LI: Low idle speed, no load
3.1.3 Snap acceleration test cycle
This is a transient test, and repeated three times, with-
out load. For this test, the laboratory AC dynamometer
was programmed to run the test automatically, thereby
ensuring repeatability. This test was performed using
the SAE J1667 snap acceleration test protocol, which
is commonly used for testing of on-highway vehicles
(SAE 1996).
3.1.4 Test engine and fuel
For field evaluation at the Stobie mine, two DPFs were
installed on each heavy duty engine, vehicle #820 had
an engine of 207 kW, while vehicle #213 had an engine
of 213 kW. Therefore, in order to match about the same
engine power per DPF for laboratory testing also, a
Liebherr engine of 105 kW rated power was used. The
engine used for this testing was a rebuilt engine with
the following characteristics:
Manufacturer: Liebherr
Model: D914T
Displaced volume: 6.11 L
Rated power: 105 kW @ 2000 rpm
Peak torque speed: 1400 rpm
Supercharging: Turbocharger without cooling
The diesel fuel used for this study was a low sul-
phur Mining Diesel Fuel CAN/CGSB-3.16-99 similar
to one used at the mine (CGSB 1999). The fuel was
analysed for some parameters and the analysis results
are given below:
Sulphur: 0.033 wt%
Density @ 15◦ C: 826.2 kg/m3
Density @ 25◦ C: 819.1 kg/m3
Carbon: 86.72 wt%
Hydrogen: 12.84 wt%
Nitrogen: 0.24 wt%
3.1.5 Gaseous emissions
The concentrations of gaseous emissions from the
exhaust were measured for carbon monoxide (CO) and
carbon dioxide (CO2 ) using a non-dispersive infrared
detector, and for nitric oxide (NO) and nitrogen oxides
(NOx ) using and a chemiluminescence detector. The
concentration of nitrogen dioxide (NO2 ) was cal-
culated by subtracting the simultaneously measured
137
Copyright © 2006 Taylor & Francis Group plc, London, UK
value of NO from the value of NOx . The concentra-
tions of oxygen (O2 ) and total hydrocarbon (THC)
were measured using paramagnetic and heated flame
ionization detectors (HFID) respectively.
3.1.6 Particulate measurements
The particulates were measured using various meth-
ods and instruments. These were based on particulate
mass, size and number, polycyclic aromatic hydrocar-
bons (PAH), and opacity. These are described below.
The Diesel Particulate Matter (DPM) by mass is
measured gravimetrically using a partial flow sam-
pling system. This is a fully automated test stand with
a micro-dilution chamber which draws the exhaust
sample directly from the engine exhaust stack and
dilutes the entire sample fraction. The system’s two
mass flow controllers provide the required control of
flow rates and dilution ratios. These microprocessor
based flow controllers provide sample flow accuracy
and measurement reliability. The particulate sample is
collected on a pair of Pallflex T60A20 (Teflon coated
glass fibre filler) 90-mm filters. All sample filters are
stored and weighted in an environmentally controlled
balance chamber (temperature at 22 ± 3◦ C and rela-
tive humidity at 45 ± 8%). The DPM concentration in
mg/m3 is calculated by using sample mass, exhaust
gas sample flow rate and sampling duration.
The particle size distribution and number of par-
ticles (range 17 nm to 583 nm) in the exhaust gas
were measured using a Scanning Mobility Particle
Sizer (SMPS) system. A sample of raw exhaust gas
was diluted, and them passed through a thermode-
nuder prior to analysis by SMPS. The thermo-denuder
system was set at a temperature of 300◦ C to remove
moisture and volatile particles from the sample in order
to measure dry particles.
The PAS concentration in real time was measured
using a NanoMet instrument. The unit consists of
a heated sampling line, a mini-dilution system with
tuneable dilution ratio, a diffusion charging (DC) sen-
sor, a photoelectric aerosol sensor (PAS), and the
manufacturer’s software. It is a technique for on-line
characterization of nanoparticle size and composition
and their diffusion behaviour. The time resolution and
the sensitivity of the instrument facilitates transient
measurements in a wide concentration range. The PAS
data is considered proportional to the elemental carbon
portion of DPM.
The opacity meter (Bosch Model RT 100 with 0
to 100% opacity range) was used with transient snap
acceleration testing. The smoke meter used for this
study is based on the partial flow method, where
a pump draws a sample of the exhaust through the
sampling probe and fills the measuring chamber com-
pletely and quickly. The partial sampling method
reduces the influence of the exhaust pressure and its
variations on the result of the measurements.
 3.1.7 Engine test run
Prior to each test run the engine intake air restriction
and exhaust backpressure at the rated power were set
at the maximum values allowed by the engine manu-
facturer. For the Liebherr engine used, the intake air
restriction was set at 3.75 kPa (15 in.H2 O), and the
exhaust backpressure was set at 7.5 kPa (30 in. H2 O).
In order to determine the performance of DPFs, the
engine dynamometer was run on all three types of test
cycles with and without DPFs. The testing performed
without DPFs is referred to as the engine baseline test.

3.2 Characterization of particulates

The DPM samples collected on 90-mm filters during
8-mode test cycles were analysed for the following Figure 1. Cracking around the retaining ring of DPF
items: HD820.
– Soluble organic fractions (SOF) by accelerated
solvent extractor (ASE) extraction method
– Polycyclic aromatic hydrocarbon (PAH) by Gas chro-
matography (GC) and Mass spectrometry (MS)
– Sulphate by ion chromatography
All sample filters for an 8-mode test cycle were
combined together for the analysis. Therefore, analy-
sis was done for each test, and average analysis results
were available for each test only. Individual mode
samples were not analysed.

3.3 Inspection of diesel particulate filters

The DPFs were inspected in the laboratory to deter-
mine if they were damaged during field trials. Any
internal damage will also reduce the particulate fil-
tration efficiency of the DPF. The evaluation included
visual and x-ray inspection prior to engine bench test-
ing, and borescope and destructive inspection after the
bench testing.
The rigid borescope used for this project is devel-
oped by the Karl Storz Industrial group. These are
used to view inside equipment and machinery such as
engines, turbines, pipes and condensers without hav-
ing to take them apart. The borescope uses fiberoptic
light guides to illuminate the inspection area and a
series of lenses to relay the images to the operator.
4 RESULTS AND DISCUSSIONS
4.1 Inspection of diesel particulate filters
The DPF HD820 suffered a surface dent in the out-
side jacket which did not penetrate the inner shell. The
inlet side shows some cracking near the retaining ring.
The outlet side (Fig. 1) has a large amount of crack-
ing around the retaining ring, and some pieces were
broken off and had fallen out. The intumescent mat
Figure 2. Deterioration of mat and DPM blowthrough for
DPF HD820.
was deteriorated and had allowed DPM to bypass filter
(Fig. 2).
The outlet side of DPF HD213A had some sur-
face damage and abrasion. There are large sections of
refractory material missing (Fig. 3) from between the
silicon carbide (SiC) blocks to a depth of 10–20 cm.
There is some soot blow through in the dam-
aged areas. The inlet side is intact aside from some
minor abrasions. The borescope inspection found
some cracks in adjacent channel walls of individual
cell blocks (Fig. 4). This could indicate that cell blocks
were under high axial stress possibly caused by uneven
soot loading and regeneration, or vibration. This DPF
had some internal damage to channels, which affected
its efficiency.
The inlet and outlet faces of DPF 213B were in good
condition with some surface abrasions. Borescope
inspection was not required for this unit, and it was
considered undamaged.

138

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 3. Percent reduction in emissions for the integrated
8-mode test cycle.

Emissions HD820 HD213A HD213B

CO 7.8 −3.2 −3.8

CO2 −0.3 0.9 −0.2
NO 1.8 0.5 0.7
NO2 67.3 38.0 62.5
THC 27.7 8.4 16.0
DPM 89.1 84.7 92.9

1.0E+08

Concentration dW/dLogDp (cm3)

1.0E+07
Baseline
Figure 3. Loss of cell bonding material from the outlet side 1.0E+06 (Liebherr)
of DPF HD213A.
1.0E+05
DPF HD820
1.0E+04 Mode 1
1.0E+03

1.0E+02

1.0E+01
10 100 1000
Diameter (nm)

Figure 5. Particle size distribution for engine baseline and

with DPF HD820 at rated speed and full load (Mode 1).

field, and HD213A and HD213B were base metal cat-

Figure 4. Borescope photos of cracks in channel walls and
penetration of DPM in DPF HD213A.
4.2 8-mode test cycle
During testing, engine parameters, exhaust gaseous
and particulate measurements were made. The test
data was analysed according to ISO 8178-1 calcu-
lations (ISO 1996a), and 8-mode integrated values
were calculated using the appropriate weighing fac-
tors (Table 2). The percent reduction or efficiency of
a DPF is calculated in comparison to the engine base-
line data. The percent reduction in emissions due to
DPFs for the integrated 8-mode test cycle are shown
in Table 3. The negative values represent increase in
emissions, while positive values indicate emissions
reduction attributable to the DPFs.
The percent reduction or increase due to DPFs for
CO2 , CO, and NO is small (<8%), while there is a sig-
nificant reduction in THC (8–28%), NO2 (38–67%),
and DPM (85–93%). All these DPFs were catalyzed,
HD820 used fuel borne cerium-Eolys catalyst in the
alyzed DPFs. The HD820 DPF may have had some
cerium left in it. The inlet temperature of the DPFs
were measured at about 480◦ C for mode 1 and 5
(Table 2), this temperature is high enough to burn
soot in these DPFs. This is perhaps the reason for
the CO increase with the HD213 DPFs. The corre-
sponding CO increase in HD820, perhaps, is masked
by much larger CO reductions due to the cerium cat-
alyst. Hence, it is evident that some regeneration took
place for all the DPFs evaluated. This is perhaps the
reason for a significant reduction in NO2 , 67% for
HD820, 38% for HD213A, and 63% for HD213B. This
reduction seems to be due to NO2 reaction with soot
during regeneration periods. The DPFs were found to
be very effective in reducing DPM, by 89% for HD820,
85% for HD213A, and 93% for HD213B. The DPM
reduction by HD213A is 8% lower than by HD213B,
perhaps due to the slight damage to HD213A in field
as discussed earlier.
The size distribution data (range 17 nm to 583 nm)
for engine baseline and with DPFs were collected at
all 8-mode points. The data for engine baseline and
with HD820 at mode 1 is shown in Figure 5. This data
shows an excellent DPF efficiency in all size ranges.
The results of DPF efficiency based on the num-
ber of particles measured by the SPMS are shown in
Table 4. The results show that for most of the mode

139

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 4. DPF percent efficiency for ISO 8178-C1 8-mode Table 6. Percent reduction in PAH, sulphate and SOF.
test cycle based on number of particles.
Sample HD820 HD213A HD213B
Mode # HD820 HD213A HD213B
PAH 83 70 89
1 99.0 91.8 99.5 Sulphate 34 30 44
2 99.5 91.9 99.7 SOF 44 26 30
3 99.0 94.7 99.9
4 99.0 96.0 99.9
5 98.7 95.7 99.9 Table 7. Breakdown of DPM in terms of sulphate, SOF and
6 98.8 94.6 99.9 solids.
7 99.2 96.8 99.9
8 94.4 83.7 99.7 Baseline HD820 HD213A HD213B
(mg) (mg) (mg) (mg)
Average 98.4 93.2 99.8
SO4 0.27 0.18 0.19 0.15
SOF 3.65 2.04 2.71 2.54
Table 5. DPF percent efficiency for ISO 8178-C1 8-mode Solids 11.08 2.85 3.70 0.85
test cycle based on PAS measurements.
Total DPM 15.00 5.07 6.60 3.54
Mode # HD820 HD213A HD213B

1 95.5 93.7 99.8

particulate filters are given in Table 6. It should be
2 90.8 77.6 99.3 noted that blank filters were also analysed and the
3 90.6 80.1 99.5 results were corrected for the blanks, when quanti-
4 65.6 34.9 96.5 ties in the blanks were higher than the detection limit.
5 79.1 70.8 98.5 The method’s detection limits for PAH and sulphate
6 56.0 35.3 96.0 were 6 nanogram, and 10 microgram, respectively.
7 55.9 33.3 94.8 These reductions are very significant, and all the DPFs
8 47.5 19.6 93.0 tested were found to be very effective in reducing PAH,
Average 72.6 55.7 97.7 sulphate and SOF.
The breakdown of DPM in terms of SOF, sulphate,
SOF and solids is given in Table 7. The solids are cal-
points, DPF efficiency is higher than 99%. The lower culated by subtracting SOF and sulphate from the total
efficiency at all mode points for DPF HD213A is due DPM mass. The solids include elemental carbon, and
to some internal damage as determined during the ini- other solids.
tial inspection. Similarly, efficiency for HD820 DPF
is slightly lower due to small damage to the DPF as
discussed earlier. Overall, the DPF efficiency on a 4.4 Snap acceleration test cycle
number basis is 98.4% for HD820, 93.2% for HD213A The opacity data for snap acceleration testing with the
and 99.8% for HD213B. engine baseline (Liebherr) are shown in Figure 6. The
The results for DPF efficiency based on the PAS opacity (peak value) is measured at 15%. The opacity
(Photoelectric Aerosol Sensor) as measured by the data with DPFs were very small, close to zero, and
NanoMet instrument (Kasper et al. 2000) at all 8 hence are not reproduced in Figure 6. The opacity data
mode points are given in Table 5. It should be noted with DPFs, close to zero, indicated that DPFs were
that PAS data is frequently used to estimate EC (Ele- very effective in reducing opacity. However opacity
ment Carbon) which is part of DPM. The efficiency of testing was found to be too insensitive to detect the
HD213B based on the PAS measurement is higher at damages observed diesel particulate filters.
all 8-mode points compared to HD820 and HD213A.
Overall, the DPF efficiency based on the PAS is 73%
for HD820, 56% for HD213A and 98% for HD213B. 4.5 3-mode field test cycle
The DPF efficiency data also indicates internal leaks The engine data, exhaust gaseous and DPM concen-
within HD213A, and HD280 DPFs as revealed during trations were measured during 3-mode test cycle. The
inspection. data were used to calculate percent reduction due
to DPFs in comparison to engine baseline data, and
results for NO2 and DPM are presented in Table 8.
4.3 Particulate characterization
The exhaust gas temperature at the inlet of DPFs was
The DPM samples collected on 90-mm filter papers about 100◦ C for low idle (LI), 200◦ C for high idle (HI),
during 8-mode test cycles were analysed for PAH, and 475◦ C for full torque converter stall (FTC) modes.
sulphate and SOF. The percent reduction due to diesel It means that regeneration only took place during FTC

140

Copyright © 2006 Taylor & Francis Group plc, London, UK

 20 Table 9. DPF percent efficiency based on particle number
and PAS data for 3-mode cycle.

15 HD820 HD213A HD213B

Opacity (%)

Modes # PAS # PAS # PAS

10
LI 94.4 – 83.7 – 99.7 93.0
HI 99.2 93.1 95.9 96.2 99.9 99.7
5 FTC 98.6 81.8 95.9 78.4 99.9 98.8

0
63 74 86 97 108 120 131 142 154 165 Table 10. Comparison of laboratory and field DPF percent
Time (sec) efficiency for 3-mode test cycle.

Figure 6. Opacity snap acceleration data for engine base- DPF & LI HI FTC
line. particulate
data type Lab Field Lab Field Lab Field
Table 8. Percent reduction in NO2 and DPM for 3-mode
cycle. HD820, # 94 97 99 99 99 99
HD213B, # 100 – 100 99 100 91
HD820 HD213A HD213B HD820, PAS – 100 93 100 82 100
HD213B, PAS 93 – 100 100 99 100
Modes NO2 DPM NO2 DPM NO2 DPM

LI 18 92 15 84 28 91 testing was done inside a mechanical shop at the mine,

HI 42 86 43 68
FTC 42 90 37 86
mode testing. In general, there was a slight increase in
CO for all DPFs.
It is very clear from Table 8 that DPFs made a signif-
icant reduction in NO2 and DPM. The NO2 reduction
at HI and FTC modes is much higher than the reduc-
tion at low idle mode. The DPM reduction in all cases
is high, and above 85% at FTC mode for all DPFs. It
should be noted that DPM reduction for HD213A is
lowest perhaps due to internal damage to the DPF. The
lower DPM reduction by HD213A is consistent with
the previous results of 8-mode data.
The total number of particles in range of 17 nm
to 583 nm was measured using a SMPS instrument,
and PAS data was measured using a NanoMet instru-
ment. The measurements were made at all the modes,
with and without the DPFs. The results of efficien-
cies based on the laboratory data are given in Table 9.
The efficiency for the undamaged DPF HD213B was
over 99 percent based on particulate number, and over
93% based on PAS data. The efficiency based on par-
ticle number and the PAS instrument for the slightly
damaged DPFs were slightly lower. The data clearly
shows that HD213A had some internal damage as the
efficiency is lower compared to HD312B, while both
DPFs were used in parallel on the same equipment.
46 82 where the vehicles with these DPFs were brought for
36 94
testing. Even though the engines used in the field were
different than the one installed in the laboratory, and
field conditions were not controlled as in the labora-
tory environment, it is still worth while to compare the
efficiencies measured in the field with those obtained
in the laboratory.
Efficiencies for HD820 and HD213B based on par-
ticle number and PAS data as measured in laboratory
and the field are given in Table 10. The field mea-
surements for HD820 were obtained in May 2002,
while field measurements for HD213B were obtained
in June 2004 (Stachulak et al. 2005b). Since field
based efficiencies for both DPFs are mostly above
99%, this indicates that DPFs trapped particulates with
a high efficiency and these DPF were not damaged
at that time in the field. In general, laboratory mea-
sured efficiencies are similar to or a little lower than
those measured during field trials. This could be due
to some damage to the DPFs tested after these field
measurements.
The NO2 percent reduction as measured in field and
laboratory are given in Table 11. The field measure-
ments for HD820 were obtained in May 2002, while
field measurements for HD213B were obtained in June
2004 (Stachulak et al. 2005b).The data shows a signifi-
cant reduction in all cases. The field data show a higher
reduction in NO2 than observed in the laboratory.

4.6 Field data for 3-mode test points 4.7 Comments on the DPFs tested in the laboratory
Prior to laboratory performance evaluation these DPFs Overall all the DPFs tested in the laboratory were
were tested in field at the 3-mode test points. The field found to be effective in reducing particulates in spite

141

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 11. Comparison of laboratory and field measured
percent reduction in NO2 for 3-mode cycle.
HD820 HD213B
Modes Lab Field Lab Field
LI 18 79 28 81
HI 42 65 46 61
FTC 42 50 36 36
of some damage to HD820 and HD213A that occurred
during field testing. The DPF efficiency, in compari-
son to the engine baseline data, was calculated using
emission test results from the ISO 8178-C1 8-mode
test cycle.
The DPFs had a small (<8%) impact (increase or
decrease) on CO2 , and CO, perhaps due to regeneration
at some modes. However, all DPFs tested signifi-
cantly reduced NO2 emissions (varying from 38% to
67%), perhaps due to reaction of NO2 with soot during
regeneration periods.
All the filters tested showed significant reductions
in DPM, greater than 85% on a mass basis, greater than
93% on a number basis, and greater than 74% based
on the PAS measurement. The lowest efficiency was
measured for HD213A, perhaps due to some damage
during field trials.
Smoke opacity was not sensitive enough to identify
the damaged DPFs. It is likely it would detect gross
failures of the monolith.
All DPFs, damaged or not, reduced PAH by more
than 70%, sulphate by more than 30%, and SOF by
more than 26%.
In general, laboratory measured efficiencies are
similar to or a little than those measured during field
trials. The slightly lower efficiencies found in labora-
tory testing could be due to damage to DPFs after field
measurements.
DPF HD820 was slightly damaged when the mono-
lith moved inside the can. Deterioration and loss of the
intumescent mat was the likely cause. DPF HD213A
was damaged internally. Better control of vibration or
regeneration would prevent this failure. Due to the fact
that both filters in the dual set up were likely affected
by the same vibration levels, loss of regeneration con-
trol is the more likely cause of failure. HD213B was
not damaged, and showed the highest efficiency of all
the DPFs tested.
5 CONCLUDING REMARKS
In general, diesel particulate filters still remain the
most effective way to control DPM at the source. How-
ever, DPF monoliths are sensitive to damage in the
underground environment from impact, vibration, and
142
Copyright © 2006 Taylor & Francis Group plc, London, UK
mishandling. Some of these problems can be addressed
by the mine, but control technology manufacturers
need to develop more robust units that will protect
the filter monolith from damage.
ACKNOWLEDGEMENTS
The authors would like to express their gratitude to
the DEEP consortium for funding this study. Thank
you to Vince Feres of CANMET-MMSL for his assis-
tance in bench testing of particulate filters, and to
Gianni Caravaggio, Pamela MacDonald and Jean-
Pierre Charland of CANMET Energy Technology
Centre in Ottawa for characterization of particulate
samples.
REFERENCES
CGSB.1999. Mining diesel fuel, CAN/CGSB-3.16-99, pub-
lished by the Canadian General Standards Board, Ottawa,
Ontario, Canada.
Dainty, E.D. et al.1986. A summary of underground mine
investigations of ceramic diesel particulate filters and cat-
alytic purifiers. Heavy-duty diesel emission control: A
review of technology. The Canadian Institute of Mining
and Metallurgy. Special volume 36:54–77.
DEEP 1997. Diesel emissions evaluation program. www.
deep.org.
Gangal, M. et al. 2005. Laboratory evaluation of field trial
dpfs used for light duty vehicles. Paper S1P2. Mining
Diesel Emissions Conference. Toronto. Ontario. Canada.
ISO 1996a. ISO 8178-1. Reciprocating internal combus-
tion engines – Exhaust emission measurement. Part 1:
Test-bed measurement of gaseous and particulate exhaust
emissions.
ISO 1996b. ISO 8178-4. Reciprocating internal combustion
engines – Exhaust emission measurement. Part 4: Test
cycles for different engine applications.
Kasper, M. et al. 2000. Nano-Met: On-line characterization
of nanoparticle size and composition. 4th International
ETH-Conference. BUWAL, CH-3003 Bern, Switzerland.
McGinn, S. 2004. Noranda Inc.-Brunswick mine diesel
particulate filter (DPF) field study, Final Report of Inves-
tigations to the Diesel Emissions Evaluation Program.
http:/www.deep.org/reports/
SAE. 1996. Surface vehicle recommended practice J1667.
Snap acceleration smoke test procedure for heavy-duty
diesel powered vehicles.
Schnakenberg, G.H.Jr. & Bugarski, A. 2002. Review of tech-
nology available to the underground mining industry for
control of diesel emissions. U.S. Department of Health and
Human Services, NIOSH, IC 9262, Pittsburgh Research
Laboratory.
Stachulak, J. et al. 2005a. Particulate filter systems at Inco.
Mining Diesel Emissions Conference. Toronto. Ontario.
Canada.
Stachulak, J. et al. 2005b. Long-term evaluation of diesel par-
ticulate filter systems at Inco’s stobie mine. Proceedings
of 8th International Mine Ventilation Congress, Brisbane,
Queensland, Australia.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

DEEP project on evaluation of diesel particulate filters at Inco’s Stobie Mine

J.S. Stachulak & B.R. Conard

Inco Limited

A.D. Bugarski & G.H. Schnakenberg, Jr.

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory

ABSTRACT: The objective of the Diesel Emissions Evaluation Program (DEEP)-sponsored project at Inco’s
Stobie mine was to conduct a long term field evaluation of diesel particulate filter (DPF) systems available to
the underground mining industry. Some of the major issues studied were criteria for selecting the filter media,
means of DPF regeneration, efficiency of DPF systems and the occurrence of unwanted secondary emissions,
and reliability and durability of DPF systems. Nine state-of-the-art diesel particulate matter (DPM) filtration
systems have been retrofitted to heavy-duty and light-duty vehicles and have been subjected to extensive long-
term in-mine evaluation. Periodic efficiency tests were conducted at various stages of the study (2001, 2002 and
2004) to establish in-use efficiencies and durability of the tested DPF systems. The results were used to assess
the effects of the filter systems on the concentrations of particulate matter, nitric oxide, nitrogen dioxide and
carbon monoxide in the vehicle exhaust. The variety of filtration systems and regeneration concepts used in this
study offered the opportunity to investigate their advantages and disadvantages. This paper also offers a review
of selected experiences with the installation and operation of the DPF systems on underground mining vehicles
at the Inco Stobie Mine.

1 INTRODUCTION Brunswick mine (McGinn 2004) and the other hosted

Since their introduction into underground mining
operations in the mid-1960s, diesel powered equip-
ment has become increasingly employed and recog-
nized as the workhorse in mining. Diesel engines
proved to be relatively durable, easy to maintain and
generally to have relatively low operating costs. Inco
Limited employs over 800 diesel-powered units at
its Ontario mining operations in the Sudbury basin.
Although the use of alternative sources of power (e.g.,
electricity, fuel cells) is being explored, Inco and most
other Canadian deep-rock mining companies realize
that diesel engines will continue to be a very impor-
tant source of power for growing fleet of heavy- and
light-duty vehicles for many years to come. In view of
this, it is imperative to lessen adverse effects of diesel
emissions on quality of the air in underground mines.
A Canadian industry/labor/government consor-
tium, Diesel Emissions Evaluation Program (DEEP),
was formed in 1999 to address issues related to expo-
sure of underground miners to diesel particulate matter
(DPM). One of the major objectives of this program
was to identify strategies and controls that will attain a
tenfold reduction in miners’ exposure to DPM. DEEP
sponsored two projects, one hosted by Noranda’s
by Inco’s Stobie mine (Stachulak et al. 2005a) that
tested existing diesel particulate filter (DPF) tech-
nologies in operating mines. The long-term study,
conducted under DEEP auspices at Inco’s Stobie mine
from April 2000 to December 2004, assessed the suit-
ability and effectiveness of selected DPF systems for
control DPM emissions from heavy- and light-duty
underground diesel mining equipment.
Diesel exhaust is complex mixture of gases, liquids,
and solids. The gaseous phase includes carbon monox-
ide (CO), carbon dioxide (CO2 ), nitric oxide (NO) and
nitrogen dioxide (NO2 ) (in combination termed NOx ),
sulfur dioxide (SO2 ) and sulfur trioxide (SO3 ) and a
number of low molecular weight hydrocarbons (HC).
The liquid or vapor phase includes condensed hydro-
carbons of varying molecular structures and sulfuric
acid aerosols, which are formed by the combina-
tion of sulfur trioxide with water. The solid phase,
commonly known diesel particulate matter, is predom-
inantly made of elemental carbon (that arises from
partially burned fuel), ash and sulfate aerosols (see
Figure 1). The metals contained in these compounds
originate from lubricating oil and impurities in the
fuel. The majority of these diesel-produced particles
are submicron in size and therefore respirable. Due to

143

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 1. Schematic of diesel exhaust particulate.
Figure 2. Typical constitution of diesel exhaust (after
Burtscher, 2005).
the small particle sizes, the elemental carbon particles
provide a large surface area suitable for adsorption
and absorption of volatile and semi-volatile hydro-
carbons. These hydrocarbons constitute the soluble
organic fraction of diesel particulate matter. Typical
chemical composition of diesel particulate matter is
shown in a pie chart in Figure 2 and a typical particle
size distribution, summed over the operating modes of
a diesel engine, is shown in Figure 3.
Physical properties and chemical composition of a
diesel engine’s exhaust were found to vary as a function
of a number of parameters including engine design,
age, operating conditions, maintenance, type of fuel,
etc. It is important to recognize that there are certain
limitations and trade-offs in controlling these parame-
ters in an effort to minimize potentially harmful emis-
sions from diesel engines. It is apparent that substantial
reduction in emissions of a number of components of
diesel exhaust are achievable only by various combi-
nations of exhaust after-treatment technologies such
as diesel oxidation converters (DOCs), diesel particu-
late filter (DPF) systems, selective catalyst reduction
(SCR) systems, etc.
DOCs are widely deployed by the underground min-
ing industry for control of CO and HC emissions from
diesel engines (Stachulak 2003). Those devices were
found to be relatively easy to operate and maintain. On
the contrary, controlling DPM emissions was found to
be substantially more challenging (Schnakenberg and
Bugarski 2002). Improvements in fuel quality (Watts
et al. 1998) and maintenance, as well as improve-
ments in engine design, operation and ventilation, were
shown to be insufficient to accomplish a desired 90%
reduction in DPM. It became apparent that removal
144
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 3. Typical particle size distribution of DPM (after
Kittelson, 1998).
of DPM from the exhaust stream using diesel partic-
ulate filter systems was likely to be an integral part
of the strategy to achieve dramatic reductions in DPM
emissions (Conard and Stachulak 2000).
The first ceramic monolith DPF for use in under-
ground applications was developed by Corning Glass
in collaboration with Natural Resources Canada
(CANMET) scientists. In the early 1980s prototype
systems were laboratory-tested and then taken under-
ground for field trials (McKinnon 1989). Inco was
involved in testing a prototype at Little Stobie Mine
(Stachulak and Conard 2001) and in implementing
what was then considered to be a successful tech-
nology via the National Research Council’s Program
for Industry/Laboratory Projects (PILP) run under the
auspices of the Collaborative Diesel Research Advi-
sory Panel. However, initial success was afterward
impeded by long-term implementation issues. Fre-
quent problems with filter regeneration – burning-off
of the collected DPM – often resulted in unaccept-
able engine backpressures. As a result, many of these
systems were eventually rejected by vehicle operators.
Mine management, as well as workers, became skep-
tical of the technology. It is now recognized that the
performance of a filter is intimately associated with
successful periodic regeneration of the filter. The early
implementation of filters on underground equipment
relied on this removal being done by the heat in the
engine exhaust itself and, as now realized, sufficient
temperatures were not achieved for many DPF sys-
tems because of a combination of lower than expected
exhaust temperatures and/or a lack of sufficient time
at high exhaust temperature to complete regeneration.
VERT, a European project, evaluated the best tech-
nology to curtail DPM emissions from engines used
in construction and tunneling operations (Mayer et al.
1998). Both laboratory testing and field evaluations
of various control technologies were conducted under
auspices of this project (Mayer et al. 1999, 2002). Dis-
cussions between DEEP and Andreas Mayer, VERT’s
technical director, assisted greatly in developing the
Stobie project’s scope of work. VERT’s list of tested
and approved DPFs for retrofitting diesel engines
(SAEFL 2003) was extensively used during the DPF
 selection process for the Stobie project. The knowl-
edge gained by VERT in successfully matching engine
and DPF system characteristics was particularly valu-
able to the DEEP researchers.
As a result of the high demand for after-treatment
systems to control diesel particulate emissions from
on-road and off-road vehicles, DPF technology rapidly
advanced in recent years. Some of these technologies
are in wide use in on-road applications, but have not yet
been tested and proven for underground mining appli-
cations. Therefore, in light of the increased concern
on the effects of DPM on human health and recent
developments in control technology, DEEP decided
to sponsor long-term evaluation of six DPF systems
retrofitted on production and support vehicles in the
Inco Stobie mine fleet. The specific objectives of the
project were to:
1 develop methods for selecting DPF systems for
underground mining vehicles;
2 determine the ability of tested DPF systems
to reduce DPM emissions without significantly
increasing secondary emissions of other noxious
substances;
3 evaluate the long-term durability and reliability of
the tested DPF systems;
4 establish operational and maintenance costs for
such systems;
5 develop Canadian expertise on the DPF technology
and DPM measurement methods.
The project was initiated in early 1998 with the
generation of a proposal for conducting the study.
The proposal was peer reviewed by DEEP’s Technical
Committee and outside reviewers, and the project was
officially approved in mid-1999. Originally planned
to be conducted over roughly 2.5 years, the testing
was extended to December 2004 primarily to enable
reasonable operating hours to be accumulated by the
systems so that an assessment of long-term reliability
and durability was possible.
2 METHODOLOGY
2.1 Test vehicles
Five heavy-duty load/haul/dump (LHD) vehicles were
selected as representing the primary heavy-duty
workhorse in underground mining. One of these units
had a dual exhaust Deutz engine and the other four had
Detroit Diesel Series 60 engines. The engines spanned
a range of age and service use.
Two Kubota tractors powered by Kubota F2803B
engines were selected as being representative of light
duty vehicles, which are increasingly being used in
transporting underground personnel. DPF service on
light-duty underground mining vehicles had not been
studied anywhere at the time the Stobie Project was
145
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 4. Exhaust temperature histogram for production
LHD vehicle.
started. Their inclusion in the Stobie testing was con-
sidered important, because the contribution of light-
duty vehicles to concentrations of diesel pollutants in
underground mines is increasing (Rubeli 2003).
Temperature and pressure sensors were installed
upstream of the muffler on the each candidate vehi-
cle and data were collected at 1 Hz frequency using
on-board data acquisition systems. The temperature
and engine backpressure data were collected for six
months prior to the final selection the DPF systems for
testing. The temperature traces were used to generate
exhaust temperature histograms. Temperature traces
and histograms were needed in DPF system selection
to assess whether the engine-produced exhaust tem-
peratures and heat were sufficient to sustain partial or
even complete passive regeneration of filters. These
data were found to be absolutely essential for selection,
installation and optimization of DPF systems (Mayer
2000).
As illustrated on the example histogram shown in
Figure 4, the engines in heavy-duty LHD vehicles did
not routinely achieve high enough exhaust tempera-
tures (∼600◦ C) for long enough periods to support
complete regeneration of uncatalyzed DPF systems.
This meant that DPF systems retrofitted to LHDs
needed to be catalyzed to support passive regeneration
at relatively lower exhaust temperatures. The alterna-
tive to those systems were active systems regenerated
with help of external sources of energy. Similarly,
data on light-duty vehicles clearly showed the need
for active regeneration. The data on all vehicles indi-
cated that engine backpressure would be an important
diagnostic tool for monitoring the status of the DPF
systems.
2.2 Selection of DPF systems
The temperature traces and other information relevant
to the selection and optimization of the DPF systems
were supplied to potential DPF system suppliers. The
 DPF systems were required to meet criteria devel-
oped by VERT (SAEFL 2003). The resulting proposals
from suppliers were thoroughly reviewed by DEEP
Technical Committee, Inco’s Stobie team and external
reviewers.
Two of the LHD vehicles powered by Detroit Diesel
Series 60 engines were retrofitted with completely
passive systems: an Oberland-Mangold DPF system
with a knitted glass fiber filter and a dosing system
for fuel-borne catalyst on one, and an Engelhard DPX2
DPF system with a catalyzed cordierite monolith on
the other. One of the remaining LHDs, also pow-
ered by Detroit Diesel Series 60, was retrofitted with
a completely active system from ECS/Unikat with a
SiC monolith filter and an on-board electrical heater
for off-duty regeneration. A Johnson Matthey pas-
sive/active system DPF 201 with a SiC (or cordierite)
monolith was retrofitted to the dual exhaust Deutz
engine. This system was partially regenerated pas-
sively with help of a fuel-borne catalyst. The complete
regeneration of these filters was secured with on-board
electrical heaters. Two light-duty Kubota tractors were
retrofitted with active DPF systems: one with an ECS/
3M with a ceramic fiber filter and an on-board elec-
trical heater; the other with a DCL system with a
removable SiC monolith that was regenerated using
an off-board electrical heater.
At a later stage of the project the Oberland-Mangold
DPF system was replaced, due to technical problems,
with an ECS/Unikat DPF system with a SiC mono-
lith filter and an on-board electrical heater. Simi-
larly, the ECS/3M DPF system was replaced with an
ECS/Unikat with a SiC monolith filter and an on-board
electrical heater. In the third year of the project one
additional LHD powered by a Detroit Diesel Series 60
engine was fitted with a prototype active DPF system
from Arvin Meritor. This system was designed around
a cordierite monolith filter and an on-board diesel
fuel burner for regeneration during normal vehicle
operation.
More detailed descriptions of the evaluated DPF
systems, test vehicles, and DPF installations are avail-
able in Stachulak et al. 2005a.
2.3 Measurement of emissions and other relevant
parameters
Each vehicle was brought in for general maintenance at
least every 250 hours of operation. The effects of DPF
systems on the emissions were measured at torque con-
verter stall steady-state engine operating conditions.
Emissions of NO, NO2 , CO, and CO2 were examined
using an ECOM AC portable gas analyzer. That ana-
lyzer was also used to determine the Bacharach smoke
number upstream and downstream of each system.
More extensive emission tests were conducted in
July 2001, May 2002, and June 2004 under three
146
Copyright © 2006 Taylor & Francis Group plc, London, UK
reproducible steady-state engine operating conditions:
(1) torque converter stall, (2) high idle, and (3) low idle.
The concentrations of NO, NO2 , CO, and O2 were mea-
sured upstream and downstream of the systems using
ECOM KL and AC portable gas analyzers. Particulate
concentrations were measured upstream and down-
stream of the systems using a photoelectric aerosol
sensor (PAS 2000 from EcoChem Analytics) and an
opacity meter (DiSmoke from AVL). Particle size dis-
tributions and concentrations were measured using a
Scanning Mobility Particle Sizer (Model 3926 from
TSI Inc.).
Industrial hygiene (workplace) measurements were
conducted for “with filter” and “without filter” con-
ditions while selected test vehicles were performing
normal production duties. Ten particulate samples
were collected for each test: three samples for RCD
analyses and three samples for elemental carbon (EC)
analyses were collected on the test vehicle just behind
the driver; two samples for EC analyses were col-
lected at the fresh air supply; and two EC samples
were collected at the exhaust air outtake. Air flow
measurements were performed during each test period.
All test vehicles were equipped with temperature
(TS-200 EGTS from Heraeus) and pressure (Series
BT8000 from SensorTechnics) sensors and data acqui-
sition systems (Model FCD-001 from Paul Nöthiger
Electronics). The exhaust temperature upstream and
downstream of the DPF systems and engine back pres-
sure data were collected throughout study at 1 Hz
frequency.
3 RESULTS AND DISCUSSION
3.1 Selection of DPF systems
The process of critically evaluating DPF system design
with a vehicle’s duty cycle was found to be essen-
tial for selecting the optimum DPF system. Gathering
adequate information specific to each individual appli-
cation was found particularly important for making
educated decisions on selection and optimization of
each DPF system with regard to both its regener-
ation method and the filter’s proper sizing. Actual
exhaust temperature profiles and good estimates of
in-use DPM emissions are critical inputs for selecting
DPF systems.
The experience gathered during this project was
used to complement VERT filter selection criteria
(SAEFL 2003) to cover specifics of underground
mining applications:
1 The system should have efficiency of 95% in
removal of both particulate number and elemental
carbon mass (NIOSH 5040);
2 The system should provide adequate regeneration
over mining cycle;
 3 Backpressure must meet vehicle/engine warranty:
New DPF 50 mbar
Regenerated DPF 60 mbar
Before cleaning <150 mbar.
4 The system should be supplied with adequate
temperature and engine backpressure monitoring
transducers, data acquisition system, and appro-
priate displays;
5 Secondary emissions (catalyzed traps are the con-
cern):
6 preferably no increase of NO2 , but in any event
the system should not increase ambient NO2 con-
centrations above 0.8 ppm at a ventilation rate of
100 ft3 /hp;
7 Copper is not permitted in any catalyst formulation
(fuel or filter) due to the potential for increases in
formation of dioxins;
8 The system should be rugged enough for under-
ground mining applications;
9 The system should be CSA & CEC approved and
equipped with fire suppression systems;
10 Useful life of the system should be 3 years or 9000
hours.
3.2 DPF sizing
Elevated engine backpressures, substantially exceed-
ing the manufacturer’s recommended maximum val-
ues, were recorded during operation of several tested
systems. For example, the backpressure on the engine
equipped with Engelhard DPX2 DPF system fre-
quently exceeded its manufacturer-recommended val-
ues, periodically exceeding 300 mbar for extended
periods. The turbocharger on that engine eventually
failed and caused an oil fire. At this time is not
clear what role the DPF system and elevated engine
backpressure may have played in the turbo failure. In
general, the effects of prolonged elevated backpressure
on performance and durability of a diesel engine are
not very well established and more research is needed
to address this issue. It is clear, however, that failure
to regularly regenerate an filter and maintain engine
backpressure within recommended limits eventually
resulted in failure of the filter element.
DPF sizing was found to be particularly impor-
tant for active systems in which an inadequately sized
filter could potentially impact vehicle usability and
productivity. It is important to consider the limitations
imposed by restricted space on the vehicle, together
with safety and maintenance issues.
3.3 Effects of DPF systems on particulate and
gaseous emissions
The emissions tests showed that the evaluated DPF sys-
tems, if properly installed and maintained, can provide
reductions in DPM emissions targeted by DEEP. The
147
Copyright © 2006 Taylor & Francis Group plc, London, UK
majority of tested DPF systems provided reductions in
tailpipe DPM concentrations measured by PAS 2000
in excess of 90%. Similarly, elemental carbon anal-
ysis showed twenty-fold reductions for several tested
systems. Size distribution and concentration measure-
ments in the exhaust systems showed reductions in
total number of diesel aerosols in excess of 95%. The
effects of DPF systems on gaseous emissions were also
closely investigated. The uncatalyzed DPF systems
were found to have minor effect on tailpipe concen-
trations of CO, NO, and NO2 . Very low concentrations
of CO downstream of the filter and a two-fold increase
in NO2 emission were observed for the platinum-
catalyzed DPF system from Engelhard. Consequently,
personal exposures to NO2 should be closely moni-
tored. More details on the results of emissions mea-
surements are available in Stachulak et al. 2005a.
3.4 Durability and reliability of DPF systems
One of the DPF systems from ECS, and the systems
from Johnson Matthey and Engelhard accumulated
over 2000 hours of operation. Tailpipe emissions mea-
surements verified their durability. The other system
from ECS failed after approximately 870 hours of
operation due to human error. The initial round of
emissions testing revealed deficiencies in the Ober-
land Mangold system design, and it was consequently
removed from the vehicle. The DPF systems installed
on the Kubota tractors accumulated somewhat less
operating time due to lower utilization of those vehi-
cles. The active systems from DCL (∼730 hours)
and ECS/Unikat (∼460 hours) proved to be durable
and reliable. The ECS/3M system (∼430 hours) was
replaced with ECS/Unikat systems after 3M decided
to stop supplying their wound fiber filter media. It is
crucial to have equipment suppliers committed to an
on-going maintenance and service issues.
Close attention should be given to the design, spec-
ifications, and installation of auxiliary equipment for
the DPF systems. Such equipment including electrical
circuits, fuses, heating elements, timers, air flow con-
trollers, fuel-borne catalyst dosing systems and diesel
burner and other special equipment often required spe-
cial attention to ensure that Canadian standards for
safety were met, that consumables could be purchased
easily, and that maintenance was straight-forward.
The reliability of vehicle-mounted air compressors
and air pumps on active systems was found to be poor.
When possible, compressors and pumps on a vehicle
should be replaced by compressed mine air with the
moisture removed using conventional moisture sep-
arators. Flow rates of the air supply during active
regeneration using electric heating elements were
found to play an important role in regeneration pro-
cess. Excessive air flow can reduce the temperatures
being attained and adversely affect the regeneration
 process. Therefore, air flows should be periodically
verified and adjusted if needed.
3.5 Installation of DPF systems
The project offered an opportunity to investigate sev-
eral issues related to the installation of DPF systems
on underground mining equipment. Ideally, the oper-
ation of DPF systems should be transparent to the
vehicle’s operator and should require minimum addi-
tional maintenance effort. The DPF systems evaluated
in this study did not meet this ideal situation. In gen-
eral, tested systems required a relatively high level of
operator attention and engagement from maintenance
personnel. Some of the systems proved to be particu-
larly complex and demanding. Some of the technical
problems with those systems can be attributed to the
fact that manufacturers had limited experience with
deploying their products to underground applications
and used this study as a platform for development and
optimization of their systems.
All of the DPF systems tested on the relatively large
LHD vehicles were successfully installed in the engine
compartments of those vehicles. It was important to
consider all potential problems associated with instal-
lation and engine maintenance. Installation of DPF
systems on Kubota tractors was somewhat more chal-
lenging. The systems on the tractors were installed on
the front fenders outside of the engine compartment.
Continuous monitoring and recording temperature
and engine backpressure was found to be a very useful
tool. The continuous and instantaneous information on
engine backpressure proved to be valuable for diag-
nostics of regeneration and overall performance of
DPF system. A redundant system with simple, but
effective, dashboard alarm signal is needed in order
to give information to the vehicle operator about the
engine backpressure. Operators must be attentive to
non-conventional alerts and alarms for high backpres-
sure or else serious harm could be done to the engine.
Analyzing temperature and engine backpressure data
requires expertise and time. For proven DPF systems,
the data need only be collected at a minimum fre-
quency so as to minimize the time spent processing
and analyzing data.
Thermal insulation of exhaust pipes and filters,
where possible, was found to be effective in improving
DPF performances. The length of exhaust manifolds
and number of bends were minimized for optimum
performance.
Using shock absorbing mounts on the DPF assem-
bly was found to improve the durability of the systems.
Mounting DPF system high on a vehicle was found
to reduce potential damage from the intrusion of mud
and water in the DPF housing. An incident with an
LHD which was used for cleaning sumps showed that
mud and water can compromise performance of the
148
Copyright © 2006 Taylor & Francis Group plc, London, UK
filter element and electrical connections. An alterna-
tive to elevating the complete system is designing an
enclosure or redesigning the DPF system so that the
electrical heater is not part of the system, but can be
quickly installed when regeneration is needed.
It is important to install the DPF system components
and associated exhaust manifolds at a safe distance
from fire suppression actuators on the vehicle. A fire
suppression system on one of the evaluated vehicles
was actuated as a result of temperatures generated from
an exhaust manifold that was routed too close to the
actuator by mistake.
The operation of the engine with an overloaded DPF
system might cause serious damage to the DPF system
and engine. The ideal solution would be integration of
the DPF system sensors into an electronic control mod-
ule (ECM) of electronically-controlled engines and
thus would allow engine output to be limited when
excessive engine backpressure is sensed. This option
needs to be explored with engine manufacturers.
The use of metal electrical connector housings
instead of the often supplied plastic housings was
found to improve their ruggedness, resulting in fewer
maintenance problems.
3.6 Operation of DPF systems
This project showed that an emissions/maintenance
program is essential for implementation and opera-
tion of DPF systems. The performance of DPF sys-
tems should be evaluated as part of routine preventive
maintenance (Stachulak et al. 2005b). The portable gas
analyzers, such as those from ECOM America used
in this study, were found to be helpful in performing
diagnostics on DPF systems. These measurements are
most effectively carried out by skilled mechanics dur-
ing scheduled vehicle maintenance, and it is relatively
easy to include this in routine vehicle servicing. Edu-
cation and training of maintenance personnel in the
specifics of each DPF system plays an important role
in the process.
Proper education of vehicle operators, mechanics
and others involved with DPF systems implementation
and operation is essential for their successful, long-
term use.
ACKNOWLEDGMENTS
The authors would like to express their gratitude to the
DEEP consortium for funding the project, and par-
ticularly to the members of its Technical Committee.
Also, recognition is due to the Ontario Workers Safety
and Insurance Board and the New Brunswick Work-
place Health, Safety and Compensation Commission
for their funding support of the project. An enor-
mous number of people assisted the project, including
 Andreas Mayer and Paul Nothinger of Switzerland,
Sean McGinn, Win Watts, Michel Grenier, Bill Howell
and DPF system/diesel engine manufacturers. A very
large team of employees at Inco contributed to the suc-
cess of the project including vehicle operators, main-
tenance and instrumentation personnel, along with
shift supervisors and foremen. Acknowledgement
is also given to the Stobie Mine manager, Mike
McFarlane, to senior mine personnel Greg Nault and
Rick Mayotte. Special thanks to Giles Bedard, Bob
Coupal and the other members of Stobie DEEP Team
as well as to the Ontario Division of Inco for their
support and encouragement.
REFERENCES
Burtscher, H. 2005. Physical characterization of particulate
emissions from diesel engines: a review. Aerosol Science,
36, pp. 896–932.
Kittelson, D.B. 1998. Engines and nanoparticles: A review.
J. Aerosol Sci. Vol 29. No. 5/6, pp. 575–588.
Mayer, A., Heeb, N., Czerwinski, J., and Wyser, M. 2002. Sec-
ondary Emissions from catalytic active particle filter sys-
tems. Society of Automotive Engineers. 2003-01-0291.
Mayer, A., Matter, U., Scheidegger, G., Czerwinski, J.,
Wyser, M., Kieser, D., and Weidhofer, J. 1998. VERT:
Diesel nano-particulate emissions: Properties and reduc-
tion strategies. Society of Automotive Engineers 980539,
pp. 127–138.
Mayer, A., Matter, U., Scheidegger G., Czerwinski, J., Wyser,
M., Kieser, D., and Weidhofer, J. 1999. Particulate traps for
retrofitting construction site engines. VERT: Final mea-
surements and implementation. Society of Automotive
Engineers. 1999-01-0116, pp. 107–121.
Mayer, A. 2000. Datalogging review, Trap Selection and in
Mine trap Test Planning. Workshop on DEEP/Inco in Mine
Test at Stobie Mine, Nov.
McGinn, S. 2004. Noranda Inc.- Brunswick Mine Diesel
Particulate Filter (DPF) Field Study, Final Report of Inves-
tigation to the Diesel Emissions Evaluation Program.
[http://www.deep.org/reports/nordpf_final.pdf ].
149
Copyright © 2006 Taylor & Francis Group plc, London, UK
McKinnon, D.L., Wheeland, K.G., and Gavel, K. 1989. Diesel
Particulate Filter Underground Trial at Brunswick Min-
ing and Smelting’s No. 12 Mine. SAE Tech. Paper Ser.
891846.
Rubeli, B., Gangal, M., Butler, K., and Aldred, W. 2003.
Evaluation of the Contribution of Light-duty Vehicles to
the Underground Atmosphere Diesel Emissions Burden,
Phase II. Final Report to Diesel Emissions Evaluation
Program. [http://www.deep.org/reports/ldv2.pdf ].
SAEFL 2003. VERT filter list: Tested and approved particle
trap systems for retrofitting diesel engines. Swiss Agency
for Environment, Forest and Landscape (SAEFL), Bern,
June.
Schnakenberg, G.H. Jr. and Bugarski, A. 2002. Review of
Technology Available to the Underground Mining Indus-
try for Control of Diesel Emissions. U.S. Department
of Health and Human Services Information Circular
IC 9262. [http://www.cdc.gov/niosh/mining/pubs/pdfs/
ic9462.pdf ].
Stachulak, J.S., Diesel Emission Control Strategy 2003.
NIOSH Mining: Workshop on Diesel Emissions Control
Technologies in Metal/Nonmetal Mines/CDC/NIOSH,
U.S.A; Cincinnati, Ohio, Feb. 27 and Salt Lake City, Utah,
Mar. 4.
Conard, B.R., Stachulak, J.S. 2000. Underground Evaluation
of Diesel Particulate Filter Systems at Inco’s Stobie Mine.
Proceedings of the Ventilation 2000, 6th International
Symposium on Ventilation for Contaminant Control,
Helsinki, Finland, June 4–7.
Stachulak, J.S, Conard. B.R. 2001. Diesel Emission Controls
strategy: Underground Evaluation of Particulate Filter
System at Inco’s Stobie Mine. Proceedings of Seventh
International Mine Ventilation Congress, Cracow, Poland,
June 17–22.
Stachulak, J.S., Conard, B.R., Bugarski, A.D., and
Schnakenberg, G.H. Jr. 2005a. Long–Term Evaluation
of Diesel Particulate Filter Systems at Inco’s Stobie
Mine. Proceedings of 8th International Mine Ventilation
Congress, Brisbane, Queensland, Australia, July 6–8.
Stachulak, J.S., Conard, B.R., Nault, G., Bugarski, A.D., and
Schnakenberg, G.H., Jr. 2005b. Evaluation and Expe-
rience with Particulate Filter Systems at Inco. Mining
Diesel Emission Conference, Toronto/Markham, Ontario.
Canada, Oct.12–14.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

DPM reductions at underground metal and nonmetal mines using

alternative fuels

M.J. Schultz, D.M. Tomko & R.L. Rude

Mine Safety and Health Administration, Pittsburgh Safety and Health Technology Center,
Pittsburgh, Pennsylvania, USA

ABSTRACT: The Mine Safety and Health Administration (MSHA) has been assisting mine operators in
meeting both current and future diesel particulate matter (DPM) regulations. To meet these requirements, mine
operators can choose the controls that are best suited to their operation. Typical controls for DPM include:
ventilation, clean engines, environmental cabs, alternative fuels, after-filters, and work practices. In an effort to
evaluate the effectiveness of alternative fuels, MSHA, in cooperation with mine operators, has been conducting
studies at underground mines to evaluate various alternative fuels encountered in the mining community. The
alternative fuels tested include various biodiesel fuel blends of both recycled vegetable oil and soy bean oil-
based and both summer blend and winter blend water emulsified fuels. The surveys have been conducted at four
separate mines with similar surveys conducted at different underground mines to confirm results. The results of
the surveys indicate significant reductions in both DPM emissions and personal exposures at the mines.

1 INTRODUCTION used to emulsify the water into diesel fuel. MSHA

personnel involved in these studies were from the
Diesel particulate surveys were conducted at four Pittsburgh Safety and Health Technology Center, Dust
underground mines. Table 1 shows the mines and Division and Metal and Nonmetal Mine Safety and
locations studied. Health.
The purpose of the study was to evaluate the effect Tables 2 through 5 shows the dates and fuel mix-
of utilizing alternative fuels on diesel particulate emis- tures studied at the Maysville Limestone Mine, Black
sions and personal exposures in underground nonmetal River Limestone Mine, Stone Creek Brick Mine, and
mines. All mines were surveyed using the No. 2 low Durham Mine.
sulfur diesel (LSD) fuel. The alternative fuels con- Each study period lasted for two to three weeks.
sisted of biodiesel fuel mixtures and water-blended During the study period, all the underground diesel
diesel fuel emulsion mixtures. Two types of biodiesel equipment was fueled using the appropriate blend of
fuels were studied. They were a recycled vegetable oil alternative fuel. The first week of each study was used
(RVO) consisting of different fuel mixtures and a virgin to purge the equipment and fuel distribution system
soy oil (VSO). The water-blended diesel fuel emulsion from the previous fuels used. Sampling was conducted
studied was a PuriNOx™ fuel consisting of two differ- during the second or third week of the study.
ent blends. The PuriNOx™ winter blend fuel contains
10 weight percent (wt%) water. The PuriNOx™ sum- Table 2. Maysville Mine fuel mixtures.
mer blend fuel contains 20 wt% water and an additive
Date Fuel
Table 1. Mines and locations studied.
DEC 10–12, 2002 RVO 20–80%
Mine Location JAN 7–9, 2003 RVO 50–50%
FEB 4–6, 2003 No. 2 LSD (biodiesel baseline)
Maysville Limestone Mine Maysville, Kentucky APR 1–3, 2003 VSO 50–50%
Black River Limestone Mine Butler, Kentucky JAN 6–7, 2004 RVO 35–65% (PuriNOx™ baseline)
Stone Creek Brick Mine Stone Creek, Ohio FEB 2–3, 2004 PuriNOx™ 10% water
Durham Mine Pella, Iowa MAY 25–26, 2004 PuriNOx™ 20% water

151

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 3. Black River Mine fuel mixtures. 2.1.2 Black River Limestone Mine
The Black River Mine, located in Pendleton County,
Date Fuel Kentucky, is an underground limestone mine, owned
and operated by Carmeuse Lime & Stone, Inc. The
MAR 18–19, 2003 No. 2 LSD (biodiesel baseline)
APR 8–9, 2003 RVO 35–65% Camp Nelson Limestone formation is mined.The mine
APR 29–30, 2003 VSO 35–65% operates two 10-hour production shifts, 4 days per
week to produce approximately 2.5 million tons of
MAR 16–18, 2004 RVO 35–65% (PuriNOx™ baseline)
APR 13–14, 2004 PuriNOx™ 10% water limestone annually. The limestone deposit is mined
MAY 25–26, 2004 PuriNOx™ 20% water using a room-and-pillar, heading-and-bench mining
method. The headings are approximately 24 feet high
and benches range from 30 to 40 feet in height.
Table 4. Stone Creek Brick Mine fuel mixtures. During the biodiesel fuel studies, primary airflow
was induced into the mine at two intake fan locations.
Date Fuel Air was then coursed to the working areas by air walls
and auxiliary fans. The free-standing auxiliary fans,
JUN 10–11, 2003 No. 2 LSD (PuriNOx™ baseline) which had no ductwork or tubing, assisted ventilating
JUL 29–30, 2003 PuriNOx™ 10% water the working panels.
Between the biodiesel surveys and PuriNOx™ sur-
veys, the mine had realigned their ventilation system
Table 5. Durham Mine fuel mixtures. and the south intake was changed to an exhaust (south
side exhaust). During the PuriNOx™ fuel studies, the
Date Fuel intake air entered the mine at the north side intake
and the 6-foot escape shaft intake. Air was coursed to
APR 6–7, 2004 No. 2 LSD (PuriNOx™ baseline)
MAY 25–26, 2004 PuriNOx™ 20% water the working areas by air walls and auxiliary fans. The
free-standing auxiliary fans, which had no ductwork
or tubing, assisted in ventilating the working panels.
Intake air was coursed throughout the mine to three
2 BACKGROUND exhaust areas: the south side exhaust, the main exhaust,
and the belt slope.
2.1 Mines
2.1.1 Maysville Limestone Mine
The Maysville Mine, located in Mason County, 2.1.3 Stone Creek Brick Mine
Kentucky, is an underground limestone mine owned The Stone Creek Brick Mine, located in Tuscarawas
and operated by Carmeuse Lime & Stone, Inc. The County, Ohio, is an underground clay mine owned and
Camp Nelson Limestone formation is mined.The mine operated by Stone Creek Brick Company. The mine
operates two 10-hour production shifts per day to operates as a captive mine to supply clay for the making
produce approximately 3.8 million tons of limestone of bricks. The mine operates one 8-hour shift per day,
annually. The mine is approximately 1,000 feet deep. 5 days per week. Five employees operate the mine.
Mined entries are approximately 45 to 50 feet wide The mine uses a conventional room-and-pillar mining
with the final mining height ranging from 50 to 60 method.
feet. The limestone deposit is mined using a regular Mine ventilation is provided by a blowing system
room-and-pillar, heading-and-bench mining method. powered by a surface installed 42-inch Joy fan. Air
The headings are approximately 20 to 24 feet high and enters by being forced down a shaft into the mine and
the bench ranges in height from 30 to 40 feet. exits through the drift mouth of the mine. Brick walls
Primary airflow was induced into the mine using were used as stopping through most of the mine and
ventilation fans located underground at the base of maintained in excellent condition.
two vertical intake shafts. Each fan installation con-
sisted of a set of fans that worked in parallel with
each other, installed side-by-side. Intake air entered 2.1.4 Durham Mine
the mine at the elevator shaft and the 1 West intake The Durham Mine, located in Marion County, Iowa, is
shaft. Air was then coursed to the working areas by an underground aggregate mine owned and operated
air walls. Free-standing auxiliary fans, which had no by Martin Marietta Aggregates, Inc. The mine con-
ductwork or tubing, assisted ventilating the working sists of three levels. Gypsum is mined on the top level
panels. Intake air was coursed throughout the mine and referred to as the GYP Level. Limestone is mined
to two exhaust areas: an exhaust shaft and the belt on Level 1 and Level 2. A total of 12 underground
slope. Airflow quantities varied slightly from survey miners including supervisors work one production
to survey due to natural ventilation pressures (NVP). shift per day in Level 1 and Level 2. The mine uses

152

Copyright © 2006 Taylor & Francis Group plc, London, UK

 conventional mining methods in all levels, blasting
daily at approximately 4:00 p.m.
One intake shaft and two exhaust slope openings
connect the underground mine to the surface. The mine
used dual fans in parallel, mounted on the surface, to
power a positive pressure ventilation system pushing
intake air down the intake shaft. The 8.5-foot diameter
intake air shaft discharged air into Level 1 of the mine.
From this level, air is coursed throughout the mine to
the exhausting slopes.
2.2 Fuels
2.2.1 Biodiesel
Biodiesel is a methyl ester product produced by com-
bining methanol oil or feedstock, then adding a cata-
lyst. Glycerin is spun off during the refining process
with the remaining product being termed a mono-alkyl
ester known as biodiesel. Biodiesel can be made from
a variety of feedstocks including soybeans, rapeseed
canola, and palm oil as well as from recycled vegetable
oils. The RVO, more commonly referred to as “yellow
grease”, was supplied by Griffin Industries, Inc. The
VSO was supplied by Peter Cremer North America.
2.2.2 PuriNOxTM
The PuriNOx™ technology is an EPA-approved fuel
developed by The Lubrizol Corporation. PuriNOx™
is a formula of additives, deionized water, and com-
mercial diesel fuel that is created using automated
blending units installed at fuel terminals. The fuel is
an invert emulsion, meaning the water droplets are dis-
persed within the diesel fuel. The average water droplet
size is less than one-micron in diameter. PuriNOx™
is available from April through September in a sum-
mer formulation containing 20 wt% water. During the
cooler months of October through March, PuriNOx™
is available in a winterized formulation containing
10 wt% water and an antifreeze to protect the water
droplets from freezing. The flash points are 161 degree
Fahrenheit (◦ F) for the summer blend and 168.8◦ F for
the winter blend. Power loss can occur, since water
does not have any energy content.
3 SAMPLING LOCATIONS
3.1 Maysville Limestone Mine
Six area and five personal samples were collected dur-
ing each day of the study. Area samples were collected
at two main intake locations: one at the bottom of the
elevator shaft and one at the outlet end of the dual
intake fans. Return samples were also taken: two side-
by-side samples at the bottom of the return shaft and
another two side-by-side samples approximately 400
feet up the slope entry. Five personal samples were col-
lected on each shift and included drillers, roof bolters,
153
Copyright © 2006 Taylor & Francis Group plc, London, UK
scalers, loaders, truck drivers, a powderman, and an
ANFO truck driver.
3.2 Black River Limestone Mine
During the biodiesel fuel sampling, seven area sam-
ples and five personal samples were collected during
each day of the study. Area samples were collected at
the two main intake locations: near the north intake
airshaft and the south intake airshaft. Return side-by-
side area samples were also taken at the bottom of the
return airshaft and the exhaust slope. The north jaw
crusher dump point was also sampled. Five personal
samples were collected on each shift. Although indi-
vidual miners were selected based on availability, the
same occupations were sampled for each phase of the
study. Occupations sampled were a loader operator,
two truck drivers, a scaler, and a roof bolter.
During the PuriNOx™ fuel sampling, thirteen area
samples were collected during each day of the study.
Two samples were taken from the intake areas, six sam-
ples were taken from the return areas, and five samples
were placed on operating equipment. The samples col-
lected at the two intake locations consisted of one at
the north side intake and one at the 6-foot escape shaft
intake.The samples collected at the six return locations
consisted of two side-by-side samples each at the south
side exhaust, the main exhaust, and the belt slope. The
five samples placed on the equipment consisted of
a powder truck, a haul truck, a face drill, a loader,
and a roof bolter. The samples were placed outside the
enclosed equipment cabs on the haul truck, face drill,
and loader. The sample for the roof bolter, whose oper-
ator drills from inside the cab and bolts from outside of
the cab, was collected outside the enclosed equipment
cab in the basket. The sample for the powder truck,
was collected in the basket.
3.3 Stone Creek Brick Mine
Three area and three personal samples were collected
during each day of the study. An area sample was
collected from the intake located at Intersection 37
and A. Two return samples were collected approxi-
mately 15 feet from the entrance of the mine. Personal
samples were collected on the driller, transloader, and
locomotive operator.
3.4 Durham Mine
Thirteen area samples were collected during each
study to monitor the intake and exhaust air in the
mine. Side-by-side sampling was used in the mine’s
main exhaust entries. Single samples were used in the
intake entries and one single sample was taken beside
the dumping point at the crusher.
Personal samples were collected on the mucking
loader, powder crew, drill operator, and haul trucks.
 These samples were taken to show effects of the fuel Table 6. Maysville Mine average area samples.
change in the work area.
Weighted %
Return Reduction
Airflow Total Normalized from
4 ANALYTICAL
Fuel (cfm) Carbon (µg/m3 ) No. 2 LSD
Individual area and personal samples were collected No. 2 LSD 892,000 EC × 1.3 350 –
with SKC, Inc., diesel particulate sampling cassettes. (biodiesel EC + OC 318 –
This cassette includes a submicron impactor and baseline)
a quartz fiber filter. All sampling units used 10- RVO 20–80% 883,000 EC × 1.3 228 35%
millimeter nylon preseparator cyclones. Samples were EC + OC 215 32%
collected using SKC pumps and Mine Safety Appli- RVO 50–50% 858,000 EC × 1.3 103 71%
ances (MSA) Elf® pumps calibrated and operated at EC + OC 112 65%
1.7 liters per minute (Lpm) of airflow. VSO 50–50% 748,000 EC × 1.3 179 49%
EC + OC 175 45%
The airborne carbon samples were analyzed by
MSHA, Pittsburgh Laboratory, and National Insti- %
tute for Occupational Safety and Health (NIOSH), Reduction
from RVO
Pittsburgh Research Laboratory, according to NIOSH
35–65%
Method 5040. Elemental carbon (EC), organic carbon
(OC), and total carbon (TC) values were determined RVO 35–65% 977,000 EC × 1.3 243 –
(PuriNOx™ EC + OC 217 –
from the samples collected. This method uses a ther-
baseline)
mal/optical carbon analyzer to determine the OC and PuriNOx™ 918,000 EC × 1.3 117 52%
EC matter per square centimeter of filter surface. 10% water EC + OC 110 49%
Separation of different types of OC is accomplished PuriNOx™ 842,000 EC × 1.3 97 60%
through temperature ramping over time and controlled 20% water EC + OC 89 59%
atmospheric conditions. Carbonaceous minerals are
separated at a temperature of 750 degrees Centigrade
(◦ C) (fourth OC peak). The carbonaceous mineral con-
tent, evolved at the 750◦ C peak, was subtracted from
the OC portion of the analysis using the software capa- 5 RESULTS AND DISCUSSION
bility of the analytical program. OC and EC were added
together to obtain the TC. A field blank correction was 5.1 Area DPM sampling
also applied to the carbon measurements. If the field
5.1.1 Alternative fuel compared to baseline
blank correction resulted in a negative carbon mea-
sampling
surement, the carbon measurement was defaulted to
Table 6 contains the TWA summary of the average area
zero. Concentrations of carbon (micrograms per cubic
DPM sampling with average airflows for the Maysville
meter, µg/m3 ) were calculated from the following
Mine survey.
formulas:
The weighted average DPM concentrations were
obtained by multiplying each location’s air quantity by
its respective DPM concentrations, and then dividing
the sum of these products by the combined total air
quantity. The return DPM concentrations were then
adjusted for the intake concentrations. Between the
various studies, the quantity of air ventilating the mine
changed due to the NVP changes. There was a decrease
where: in airflow from the biodiesel baseline survey to each
C = The corrected OC or EC, concentration measured of the biodiesel fuel surveys as well as a decrease in
in the thermal/optical carbon analyzer. airflow from the PuriNOx™ baseline survey to each
A = The surface area of the filter media used. The of the PuriNOx™ fuel surveys. Since the quantity
surface area of the filters is 8.04 cm2 . of air ventilating a mine is inversely proportional to
All area sample concentrations were based on actual the weighted return DPM concentrations, the results
sampling time resulting in time weighted averages of the studies were normalized to the baseline air
(TWAs). For MSHA enforcement activities, normal quantities.
MSHA Metal and Nonmetal protocol is to base all Tables 7 through 9 contain the TWA summary of the
personal samples as shift weighted averages (SWAs). average area DPM sampling with the average airflows
SWA calculations use 480 minutes as the sample time for the Black River Mine, Stone Creek Brick Mine,
regardless of the actual time sampled. and Durham Mine surveys.

154

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 7. Black River Mine average area samples. Table 9. Durham Mine average area samples.

Weighted % Weighted %
Return Reduction Return Reduction
Airflow Total Normalized from Airflow Total Normalized from
Fuel (cfm) Carbon (µg/m3 ) No. 2 LSD Fuel (cfm) Carbon (µg/m3 ) No. 2 LSD

No. 2 LSD 259,000 EC × 1.3 670 – No. 2 LSD 253,000 EC × 1.3 127 –
(biodiesel EC + OC 601 – (PuriNOx™ EC + OC 125 –
baseline) baseline)
RVO 35–65% 256,000 EC × 1.3 451 33% PuriNOx™ 240,000 EC × 1.3 31 76%
EC + OC 413 31% 20% water EC + OC 47 62%
VSO 35–65% 214,000 EC × 1.3 560 16%
EC + OC 501 17%
%
Reduction Maysville Mine
from RVO Interpolated RVO 35-65%
35–65% 75
RVO 35–65% 464,000 EC × 1.3 263 – 70
(PuriNOx™ EC + OC 265 – 53%
65
Percent Reduction (%)
baseline)
PuriNOx™ 415,000 EC × 1.3 143 46% 60
10% water EC + OC 146 45%
PuriNOx™ 376,000 EC × 1.3 103 61% 55
20% water EC + OC 122 54% 50
48.5%
45
40
RVO 35-65%
Table 8. Stone Creek Brick Mine average area samples. 35
30
Weighted %
Return Reduction 25
Airflow Total Normalized from RVO 20-80% RVO 50-50%
Fuel (cfm) Carbon (µg/m3 ) No. 2 LSD Biodiesel Fuel Type

No. 2 LSD 49,300 EC × 1.3 201 – TC=ECx1.3 TC=EC+OC

(PuriNOx™ EC + OC 312 –
baseline) Figure 1. Graph of Maysville Mine interpolated RVO
PuriNOx™ 48,800 EC × 1.3 90 55% 35–65%.
10% water EC + OC 252 19%

of the RVO 50–50% fuel percent reduction from Jan-

All concentrations and percent reductions based
on TC = EC + OC or TC = EC × 1.3 were similar,
except at the Stone Creek Brick Mine. TC = EC + OC
had resulted in a higher concentration and a smaller
reduction than TC = EC × 1.3.
5.1.2 Alternative fuel compared to No. 2 LSD fuel
The PuriNOx™ sampling for both the Maysville Mine
and Black River Mine utilized the RVO 35–65% fuel
as their baseline to compare the DPM reductions. Cal-
culations were performed to make comparisons to the
No. 2 LSD fuel for both mines.
5.1.2.1 Maysville Limestone Mine
The percent reduction for the No. 2 LSD fuel was
calculated by interpolating the data from the area
sampling of the RVO 20–80% fuel percent reduction
from December 10–12, 2002 and the area sampling
uary 7–9, 2003. Figure 1 shows a graph of the results
of the RVO 35–65% fuel interpolation. The interpo-
lated RVO 35–65% percent reductions were 53% for
TC = EC × 1.3 and 48.5% for TC = EC + OC.
The interpolated RVO 35–65% percent reduction
was used to calculate the estimated RVO 35–65% con-
centration. Table 10 shows the total concentrations
and percent reductions from the 2002 and 2003 stud-
ies (No. 2 LSD, RVO 20–80%, and RVO 50–50%
fuels) and the interpolated RVO 35–65% with percent
reductions using TC = EC × 1.3 and TC = EC + OC.
Changes had been made in the mine operations,
which would affect mine DPM concentrations from the
2002 and 2003 studies. This included newer equipment
with cleaner burning engines, complete engine tuning,
and improved air walls during the past year. Also, the
quantity of air ventilating the mine changed due to the
NVP changes.

155

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 10. Maysville Mine total concentrations and percent Table 12. Maysville Mine total concentrations and percent
reductions for the area sampling of the No. 2 LSD, RVO reductions for the area sampling of the No. 2 LSD fuel equiv-
20–80%, interpolated RVO 35–65%, and RVO 50–50% fuels. alent compared to the PuriNOx™ 10% water and PuriNOx™
20% water fuels.
% %
TC = Reduction TC = Reduction % %
EC × 1.3 from EC + OC from Reduction Reduction
(µg/m3 ) No. 2 LSD (µg/m3 ) No. 2 LSD TC = from TC = from
EC × 1.3 No. 2 LSD EC + OC No. 2 LSD
No. 2 LSD 350 – 318 – (µg/m3 ) Equivalent (µg/m3 ) Equivalent
(FEB 2003)
RVO 20–80% 228 35% 215 32% No. 2 LSD 515 – 421 –
(DEC 2002) equivalent
Interpolated 165 53% 164 48.5% PuriNOx™ 117 77% 110 74%
RVO 35–65% 10% water
RVO 50–50% 103 71% 112 65% PuriNOx™ 97 81% 89 79%
(JAN 2003) 20% water

Table 11. Maysville Mine total concentrations and airflows Maysville Mine
for the area sampling of the No. 2 LSD fuel conducted Alternative Fuel versus % DPM Reduction
February 2003, interpolated RVO 35–65%, RVO 35–65% fuel 100
conducted January 2004, and No. 2 LSD fuel equivalent. 77 81 79
DPM Reduction (%)

80 71 74
65
TC = EC × 1.3 TC = EC + OC Airflow 60 49
(µg/m3 ) (µg/m3 ) (cfm) 35 45
40 32
No. 2 LSD 350 318 892,000
20
(FEB 2003)
Interpolated 165 164 892,000 0
RVO 35–65% RVO RVO VSO PuriNOx PuriNOx
20-80% 50-50% 50-50% 10% Water 20% Water
RVO 35–65% 243 217 977,000
(JAN 2004) Alternative Fuel
No. 2 LSD 515 421 977,000 TC=ECx1.3 TC=EC+OC
equivalent
Figure 2. Graph of Maysville Mine alternative fuel versus
percent DPM reduction.

A No. 2 LSD fuel equivalent was calculated to

allow for a comparison of the No. 2 LSD fuel to the
PuriNOx™ 10% water and 20% water blend fuels. The
No. 2 LSD fuel equivalent was calculated by multiply-
ing the No. 2 LSD fuel from the February 2003 study
by the ratios of the RVO 35–65% fuel from January
2004 and the interpolated RVO 35–65% concentra-
tion. This was performed for both TC = EC × 1.3 and
TC = EC + OC. Table 11 shows the concentrations
and airflows for the No. 2 LSD equivalent along with
the No. 2 LSD from February 4–6, 2003, the inter-
polated RVO 35–65%, and the RVO 35–65% from
January 6–7, 2004.
Table 12 shows the total concentrations and percent
reductions for the area sampling of the No. 2 LSD
fuel equivalent (DPM concentrations from Table 11)
and both the PuriNOx™ 10% water and 20% water
blend fuels (DPM concentrations from Table 6) using
TC = EC × 1.3 and TC = EC + OC.
Figure 2 shows a graph of the alternative fuel type
versus percent DPM reduction for the Maysville Mine.
5.1.2.2 Black River Limestone Mine
Significant changes had been made in the mine oper-
ations, which would affect mine DPM concentrations
from the 2003 study. In particular, the airflow rates
were greatly increased from 2003 to 2004. To deter-
mine if the increased airflow resulted in reduced
concentrations, a comparison was made between the
RVO 35–65% April 2003 study and RVO 35–65%
March 2004 study. From 2003 to 2004 there was
approximately an 81% increase in airflow and a
resultant 73% decrease in RVO 35–65% concen-
tration for TC = EC × 1.3 and a 58% decrease for
TC = EC + OC. This result shows that the change in
baseline RVO 35–65% concentrations were not only
due to the substantial increase in airflow rate, but also
would include other factors such as realignment of
the ventilation system, newer equipment with cleaner
burning engines, complete engine tuning and diagnos-
tic testing, changes in production, and mining practices
during the past year.

156

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 13. Black River Mine total concentrations and air- Black River Mine
flows for the area sampling of the No. 2 LSD fuel conducted Alternative Fuel versus % DPM Reduction
March 2003, RVO 35–65% fuel conducted April 2003, RVO 74
35–65% fuel conducted March 2004, and No. 2 LSD fuel 80 68
63 62

DPM Reduction (%)

equivalent.
60
TC = EC × 1.3 TC = EC + OC Airflow
33 31
(µg/m3 ) (µg/m3 ) (cfm) 40
16 17
No. 2 LSD 670 601 259,000 20
(MAR 2003)
RVO 35–65% 451 413 259,000 0
(APR 2003) RVO VSO PuriNOx PuriNOx
RVO 35–65% 263 265 464,000 35-65% 35-65% 10% Water 20% Water
(MAR 2004) Alternative Fuel
No. 2 LSD 391 386 464,000
equivalent TC=ECx1.3 TC=EC+OC

Figure 3. Graph of Black River Mine alternative fuel versus

percent DPM concentration.

Table 14. Black River Mine total concentrations and percent

reductions for the area sampling of the No. 2 LSD fuel equiv- 5.2 Personal DPM sampling
alent compared to the PuriNOx™ 10% water and PuriNOx™
20% water fuels. Personal samples had many variables affecting their
exposure to DPM concentrations; therefore, they are
TC = TC = not as useful in determining the effectiveness of the
EC × 1.3 % EC + OC % biodiesel and PuriNOx™ fuels as are the area samples.
(µg/m3 ) Reduction (µg/m3 ) Reduction

No. 2 LSD 391 – 386 –

equivalent
PuriNOx™ 143 63% 146 62%
10% water
PuriNOx™ 103 74% 122 68%
20% water
A No. 2 LSD fuel equivalent was calculated to
allow for a comparison of the No. 2 LSD fuel to the
PuriNOx™ 10% water and 20% water blend fuels. The
No. 2 LSD fuel equivalent was calculated by multiply-
ing the No. 2 LSD fuel from the March 2003 study by
the ratios of the RVO 35–65% fuel from the March
2004 and the RVO 35–65% fuel from the April 2003
studies. Table 13 shows the concentrations and air-
flows for the No. 2 LSD equivalent along with the
No. 2 LSD from March 18–19, 2003, the RVO 35–
65% from April 8–9, 2003, and the RVO 35–65% from
March 16–18, 2004.
Table 14 shows the total concentrations and percent
reductions for the area sampling of the No. 2 LSD
fuel equivalent (DPM concentrations from Table 13)
and both the PuriNOx™ 10% water and 20% water
fuel blends (DPM concentrations from Table 7) using
TC = EC × 1.3 and TC = EC + OC.
Figure 3 shows a graph of the alternative fuel
type versus percent DPM reduction for the Black
River Mine.
6 FINDINGS
6.1 Maysville Limestone Mine
The Maysville Mine reductions in the weighted
exhaust TWA DPM biodiesel concentrations from the
baseline (No. 2 LSD fuel) survey using TC = EC × 1.3
were as follows:
• RVO 20–80% indicated a 35% reduction
• RVO 50–50% indicated a 71% reduction
• VSO 50–50% indicated a 49% reduction
The reductions in the weighted exhaust TWA DPM
PuriNOx™ concentrations from the baseline (RVO
35–65% fuel) survey using TC = EC × 1.3 were as
follows:
• PuriNOx™ 10% water indicated a 52% reduction
• PuriNOx™ 20% water indicated a 60% reduction
The reductions in the weighted exhaust TWA DPM
PuriNOx™ concentrations from the calculated No.
2 LSD fuel equivalent using TC = EC × 1.3 were as
follows:
• PuriNOx™ 10% water indicated a 77% reduction
• PuriNOx™ 20% water indicated a 81% reduction
6.2 Black River Limestone Mine
The Black River Mine reductions in the weighted
exhaust TWA DPM biodiesel concentrations from the

157

Copyright © 2006 Taylor & Francis Group plc, London, UK

 baseline (No. 2 LSD fuel) survey using TC = EC × 1.3
were as follows:
• RVO 35–65% indicated a 33% reduction
• VSO 35–65% indicated a 16% reduction
The reductions in the weighted exhaust TWA DPM
PuriNOx™ concentrations from the baseline (RVO
35–65% fuel) survey using TC = EC × 1.3 were as
follows:
• PuriNOx™ 10% water indicated a 46% reduction
• PuriNOx™ 20% water indicated a 61% reduction
The reductions in the weighted exhaust TWA DPM
PuriNOx™ concentrations from the calculated No.
2 LSD fuel equivalent using TC = EC × 1.3 were as
follows:
• PuriNOx™ 10% water indicated a 63% reduction
• PuriNOx™ 20% water indicated a 74% reduction
6.3 Stone Creek Brick Mine
The Stone Creek Brick Mine reduction in the weighted
exhaust TWA DPM PuriNOx™ 10% water concentra-
tions from the baseline (No. 2 LSD fuel) survey using
TC = EC × 1.3 indicated a 55% reduction.
6.4 Durham Mine
The Durham Mine reduction in the weighted exhaust
TWA DPM PuriNOx™ 20% water concentrations
from the baseline (No. 2 LSD fuel) survey using
TC = EC × 1.3 indicated a 76% reduction.
7 CONCLUSIONS
The use of alternative fuels reduces DPM concentra-
tions in underground nonmetal mines:
• The Maysville Mine and Black River Mine are
two limestone mines that switched over to various
158
Copyright © 2006 Taylor & Francis Group plc, London, UK
blends of biodiesel fuel mixtures and PuriNOx™
fuel blends. Subsequent DPM sampling has shown
reductions in the weighted exhaust DPM concentra-
tions at both mines from all alternative fuel mixtures
tested. The PuriNOx™ fuel blends have shown
greater reductions in DPM concentrations than the
biodiesel fuel mixtures.
• The Stone Creek Brick Mine is a clay mine that
switched over to the PuriNOx™ 10% water. The
Durham Mine is a mine with one level of gypsum
and two levels of limestone that had switched over
to the PuriNOx™ 20% water. Subsequent DPM
sampling has shown reductions in the weighted
exhaust DPM concentrations at both mines from the
PuriNOx™ fuels tested.
REFERENCES
MSHA. 2003a. Environmental Diesel Particulate Matter
Investigation at the Black River Mine, Carmeuse Lime
& Stone, Inc., Butler, Kentucky. PS&HTC-DD-03-316,
PS&HTC-DD-04-420.
MSHA. 2003b. Environmental Diesel Particulate Matter
Investigation at the Maysville Mine, Carmeuse Lime &
Stone, Inc., Maysville, Kentucky. PS&HTC-DD-03-308,
PS&HTC-DD-03-311, PS&HTC-DD-04-416, PS&HTC-
DD-04-429.
MSHA. 2003c. Environmental Diesel Particulate Matter
Investigation at the Stone Creek Brick Mine, Stone
Creek Brick Company, Stone Creek, Ohio. PS&HTC-DD-
03-322.
MSHA. 2004. Environmental Diesel Particulate Matter
Investigation at the Durham Mine, Martin MariettaAggre-
gates, Inc., Pella, Iowa. PS&HTC-DD-04-423.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

The effects of water emulsified fuel on diesel particulate matter

concentrations in underground mines

J.D. Noll, S.E. Mischler, L.D. Patts, G.H. Schnakenberg, Jr., A.D. Bugarski, & R.J. Timko
U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control
and Prevention, National Institute for Occupational Safety and Health, Pittsburgh Research Lab,
Pittsburgh, PA, USA

G. Love
Carmeuse North America, Maysville, KY, USA

ABSTRACT: In this study, we evaluated the ambient diesel particulate matter (DPM) concentrations (at the
intakes and exhausts of the mine) as the entire vehicle fleet of a stone mine switched from using 35% biodiesel
to using a water emulsified fuel (PuriNOx). Elemental carbon (EC) was reduced by 45% when a PuriNOx blend
containing 10% water replaced 35% biodiesel and by 57% when a PuriNOx blend containing 20% water was
used. Other factors such as engine duty cycle and production changes could potentially cause some day-to-day EC
fluctuations, and no direct comparison to commonly used diesel fuels (No. 1 or No. 2) was achieved. Therefore,
a second study was performed in a controlled “isolated zone” environment of a metal mine, comparing PuriNOx
to No. 1 diesel fuel in a load haul dump vehicle. The EC fraction of the DPM was reduced by about 71% when the
PuriNOx blend containing 10% water was used and by about 85% when the PuriNOx blend containing 20% water
was employed. This study did not determine how the water emulsified fuel would affect the engine or power.

1 INTRODUCTION
Long-term exposure to diesel exhaust has become a
concern because diesel emissions are believed to be a
possible carcinogen by several organizations (IARC
1989, NIOSH 1988, EPA 2002). In addition, acute
overexposure to diesel exhaust has been linked to dele-
terious health effects such as eye and nose irritation,
headaches, nausea, and asthma (Kahn & Orris 1988,
Rundell et al. 1996, Wade 1993). Measurements have
shown that underground miners can be exposed to over
100 times the typical environmental concentrations of
diesel exhaust and over 10 times the concentrations
measured in other work environments where diesel
engines are common (Cantrell & Watts 1997, Nauss
1998, Haney 1992).
In the United States, the Mine Safety and Health
Administration (MSHA) has promulgated rules for
regulating miners’ exposure to diesel particulate mat-
ter (DPM) in underground metal/non-metal (MSHA
2001). In metal/non-metal mines, the MSHA rule
limits personal exposure to DPM, measured by collect-
ing air samples on quartz fiber filters and analyzing
the filters for elemental carbon (EC) or total carbon
(TC) (summation of organic carbon (OC) and EC)
by NIOSH Method 5040. For the interim rule, miner
exposure is limited to an average of 308 µg/m3 EC for
an entire shift (MSHA 2005). Due to this regulation,
many types of control technologies are being investi-
gated to reduce the DPM emitted from diesel-powered
underground mining equipment.
Water-emulsified fuel is one such control technol-
ogy being tried by the industry. The water in the fuel
is believed (1) to reduce the combustion temperature
which reduces the nitrogen dioxide (NO2 ) production
and (2) to alter combustion to inhibit soot formation.
Water emulsified fuels have been tried on several
different types of vehicles, such as buses, fire trucks,
tractors, etc. and mixed results were achieved (Rideout
1999, Rosenblatt 2000, Rosentblatt & Ainslie 1999,
Howes 2000, EPA 2002, Matthews 2002, Matthews
et al. 2002). For example, in a study done by Envi-
ronmental Canada, a 24% total DPM reduction was
observed in an International 4600 flatbed when the
diesel fuel, used by the City of Houston, was replaced
with PuriNOx (a water emulsified fuel made by Lubri-
zol) (Howes 2000). In the same study, a 69% reduction
in total DPM was observed when PuriNOx was used
in a John Deere road sweeper.
Common problems observed when using PuriNOx
were power loss and increased fuel consumption

159

Copyright © 2006 Taylor & Francis Group plc, London, UK

 (Rideout 1999, Rosenblatt 2000, Rosentblatt & Ainslie
1999, Howes 2000, EPA 2002, Matthews 2002,
Matthews et al. 2002). In some cases, PuriNOx
affected the performance of the vehicle so much that it
prevented its use, but in other cases PuriNOx was suc-
cessfully used for years (Matthews et al. 2002). Due
to the variation in these results, each type of engine
with the duty cycles being used needs to be tested
to determine the effects on the emission by the water
emulsified fuels for that type of vehicle and use.
For the underground mining industry, Environment
Canada showed some preliminary success in using
PuriNOx. In an underground salt mine, Rosenblatt and
Ainslie, from Environment Canada, reported a 46%
reduction in DPM by mass when a PuriNOx blend
replaced No. 2 diesel fuel in a front-end loader oper-
ating under a normal cycle (Rosenblatt 2000). This
study does not give the effects on elemental carbon by
the water emulsified fuel and was performed only on
one vehicle. Therefore, in order to obtain additional
information on the effects of PuriNOx on DPM from
metal/nonmetal underground mining vehicles, we con-
ducted two studies evaluating the effects of PuriNOx
on DPM generated by diesel-powered underground
mining equipment.
One study looked at the effect that PuriNOx had on
DPM concentrations in a limestone mine. The lime-
stone mine used a 35% biodiesel blend to reduce DPM
exposure. The 35% biodiesel gave about a 30% reduc-
tion in EC and TC from DPM when compared to No. 2
diesel (Gerbec & Fields, 2003, Bugarski et al. 2003).
The mine management had decided to switch from
35% biodiesel to PuriNOx blends. To evaluate this
change, samples of EC and TC were obtained at the
intakes and exhausts of the mine for three days when
35% biodiesel was used and again for three days when
warm weather and cold weather blends of PuriNOx
were being used. In this study, other factors such as
engine duty cycle and production changes could poten-
tially cause some day-to-day EC fluctuations. Three
days of data were averaged in an attempt to minimize
any effects that these factors would have on EC reduc-
tion. Nevertheless, these factors could not be totally
accounted for. PuriNOx was also only compared to
35% biodiesel, and no direct comparison to commonly
used diesel fuels (No. 1 or No. 2) was achieved.
Therefore, another study was conducted to confirm
the results in the field by comparing PuriNOx to No. 1
diesel in a controlled environment. A loader was oper-
ated in an isolated zone (a long, sealed, entry supplied
with fresh air). The isolated zone provided a controlled
environment where the only source of DPM was the
piece of equipment being investigated. The loader was
operated over a duty cycle based on that used during
actual mining. Downwind of the operating path of the
vehicle, ambient EC, TC, and DPM mass measure-
ments were obtained when No. 1 diesel fuel was used
160
Copyright © 2006 Taylor & Francis Group plc, London, UK
in the loader (baseline case) and then when a PuriNOx
blend was used.
2 METHODS
2.1 Fuels
No. 1 Diesel: Cenex, Columbus, MT
Biodiesel: A 35% blend of Bio G-3000 biodiesel
from Griffin Industries, produced from recycled
restaurant grease with No. 2 diesel fuel. This fuel was
used as the baseline.
Cold weather PuriNOx blend: 10% Water, 2.15%
Lubrizol Emulsifier Additive, 2.15% Methanol, and
85.7% No. 2 diesel fuel.
Warm weather PuriNOx blend: 20% water, 3%
Lubrizol Emulsifier Additive, and 77% No. 2 diesel
fuel.
Sulfur content and other properties of the fuels can
be found in publications by Bugarski et al. (2003,
2005).
2.2 Sampling and analytical methods
2.2.1 EC-TC sampling train
The EC-TC sampling train used for DPM sampling
was identical to the one used by MSHA for DPM com-
pliance monitoring (MSHA 2001). It consisted of a
flow-controlled MSA Elf Model pump (Mine Safety
Appliances Company, Pittsburgh, PA), a 10-mm Dorr-
Oliver cyclone, and an SKC DPM cassette (SKC, Inc.,
Eighty Four, PA). The SKC DPM cassette contained a
single-stage impactor and two 37 mm diameter tissue
quartz filters, mounted in series. The pumps were oper-
ated at 1.7 lpm. The flow rate for each of the sampling
pumps was measured and recorded at the beginning
and end of each day using a Gilibrator II bubble flow
meter (Sensidyne, Inc., Clearwater, FL). If the mea-
sured flow rates deviated more than 5% from 1.7 lpm,
the pumps were recalibrated.
Exposed SKC DPM cassettes were taken to NIOSH
Pittsburgh Research Laboratory (PRL) and analyzed in
the laboratory for EC and TC content using the NIOSH
Method 5040.
2.2.2 High-volume sampling procedure
When testing the efficiencies of control technolo-
gies in underground mines, the concentration of DPM
can be low. To collect enough material to be at and
above the limit of quantification for NIOSH Analyti-
cal Method 5040, one might have to sample for many
hours. When doing research in the field, one does not
always have this luxury. A high-volume (HV) sam-
pling train was developed by Bugarski et al. (2003,
2005) to collect more material in less time. The HV
flow rate was achieved by merging flows from five
 classifiers, each consisting of a 10-mm Dorr-Oliver
cyclone followed by a U.S. Bureau of Mines (BOM)
single stage diesel impactor, into a single stream.
A flow rate of between 1.7–2.0 lpm was maintained
through each cyclone and impactor pair. At this sam-
pling flow rate only particles with geometric mean
smaller then 0.8 µm were deposited on the filters. All
five classifiers were attached to a symmetrical plenum
that distributed a total flow rate of between 8.5–10 lpm
uniformly among the five streams. Each of the classi-
fier assemblies was connected to the plenum chamber
by a 3-foot long section of conductive tubing. The
outlet of the plenum was directly connected to a stain-
less steel 25 mm diameter filter holder containing a
stack of two 25 mm tissue quartz fiber filters (Tissue-
quartz 2500QAT, Pall Corporation, Ann Arbor, MI).
This design is described in detail in the Bugarski et al.
reports (2003).
2.3 Stone mine survey
After vehicles in a stone mine had used 35% biodiesel
for almost one year, we sampled EC, TC, and NO2
one shift per day for three days. Each vehicle was
equipped with a diesel oxidation catalyst. The fuel for
the entire fleet was then changed from 35% biodiesel
to the cold weather PuriNOx blend (supplied by Lubri-
zol). After running the new fuel for approximately one
month, we again took three days of measurements.
Following this evaluation the fuel was changed to the
warm weather PuriNOx blend and run for about one
month. Once again we sampled for one shift for three
consecutive days.
This limestone mine had one main intake entry
and two main exhaust entries. In the one main intake,
one EC-TC sampling train (described earlier) was set
up. The EC-TC sampling trains were chosen instead
of the high volume sampling since we did not have
electricity to run the high volume sampling and we
were sampling for a long enough period (about 8
hours) to collect enough material using the EC-TC
sampling train. Three EC-TC sampling trains were
placed in each of the two main exhaust entries. In the
exhaust entries, an ITX multi-gas monitor (Industrial
Scientific Corp., Oakdale, PA) measured NO2 and an
anemometer measured air velocities.
Most of the vehicles in the mine (about 61%) had
Caterpillar engines. A complete list of diesel engines
is found in the report by Gerbec and Fields (2003).
EC and TC were chosen as surrogates for DPM,
which consists of over 80% carbon (Heywood 1988,
Pierson 1983). We could not directly measure the DPM
since we believed that gravimetric or mass measure-
ments of DPM on a filter would not be sensitive or
precise enough to obtain reliable data and that even
the small portions of dust that penetrate through the
impactor (about 4–10%) may cause interference to
161
Copyright © 2006 Taylor & Francis Group plc, London, UK
mass measurements. As described later in this paper,
we believed that we could account for TC interfer-
ences (e.g. cigarette smoke, oil mist, and vapor-phase
OC) usually seen in underground mines. Unlike TC,
EC is DPM-selective and would allow us to monitor
DPM concentrations without being concerned about
interferences.
2.3.1 Ventilation normalization
Factors other than fuels can cause different day-to-day
DPM concentrations in a mine. These include venti-
lation changes, varying production, and different duty
cycles on engines. To best evaluate the effects from the
fuel change, we took measurements over three days
to minimize the variable in-mine conditions and nor-
malized the EC and TC concentrations to average
ventilation rates to account for any ventilation changes.
The normalization calculation is described in a report
by Bugarski et al. (2003).
2.3.2 Interferences
There were no known interferences for the EC mea-
surements. For OC from DPM there were possible
interferences from mineral dust, cigarette smoke, oil
mist, and vapor-phase OC.
The SKC DPM cassette was used for EC and OC
measurements. This device allows only particles less
than 0.8 µm to collect on the filter, eliminating any
significant interference from airborne mineral dust.
We were also concerned with vapor-phase OC
adsorbing onto the quartz filter and causing a posi-
tive bias to the particulate TC. To account for this, a
second filter was placed in tandem (in series) with the
primary filter. In theory, the second filter was exposed
to the same concentrations of vapor OC as the first
filter but does not collect any particulate. Thus, to cor-
rect for the adsorbed vapor-phase OC, results from the
second filter were subtracted from the first filter OC
values. In the literature, this is referred to as the tan-
dem filter correction (Kirchstetter et al. 2001, Eatough
et al. 1995, Turpin & Huntzicker 1994).
Oil mist should not be a significant interference by
the time it arrives at the samplers in the exhaust entries.
When compared to DPM, any cigarette smoke should
be so diluted that interference should be minimal at the
sampling locations used in this study (Haney 2000).
2.4 Isolated zone study
An isolated zone was set up in a metal mine as
described in reports by Bugarski et al. (2003). In the
isolated zone, the only source of diesel exhaust was the
vehicle being tested. Air flowed in through one intake
entry and out through one return entry.
A 344 Load Haul Dump (LHD) with a Caterpillar
3126B DITA engine was run with No. 1 diesel fuel
using a duty cycle typical to this vehicle in production.
 This was the same duty cycle described in reports by mine study. The 95% confidence limits for the car-
Bugarski et al. 2003 for an LHD. bon analysis for both the isolated zone and stone mine
Duplicate EC and TC samples were collected studies were calculated using the following equation
upstream and triplicate samples were obtained down- (Skoog 1986):
stream of the operating vehicle using high-volume
(HV) samplers described earlier and analyzed using
NIOSH Method 5040 (Birch 2003). HV samplers were
used in this case since we could only sample for about
two hours. Therefore, we were not sure if we could
where t is the factor for the confidence interval depen-
collect enough material on the filter using the EC-TC
dent on the degrees of freedom, s is the standard
sampling trains. An ultra sonic anemometer was set up
deviation, and N is the number of samples.
to measure entry air velocities.
Since the DPM mass for the isolated zone study
Duplicate DPM mass samples were collected using
was only done in duplicate, the above equation can-
a design similar to the HV sampler described below.
not be used to give a reliable confidence limit. In this
Instead of a quartz filter, a Teflon filter (Pall Corpora-
case, we calculated an accepted confidence limit for
tion, East Hills, NY) was used. Instead of mass flow
gravimetric samples by using the standard deviation
controllers, critical orifices were used to control the air
obtained when hundreds of samples were weighed
flow rate. Prior to and after each test gravimetric sam-
using the same procedure and balance used in this
ples were desiccated and equilibrated in a controlled
study. The 95% confidence limit for the DPM mass
environment (72◦ F and 50% relative humidity) before
measurements was then calculated using the following
weighing. Balance precision was better than 6 µg.
equation (Skoog 1986):
No. 1 diesel fuel was replaced with a cold weather
PuriNOx blend and the test was repeated. The vehicle
was run for some time to allow the engine to consume
the old fuel before the actual test began. Following this
series, another evaluation was started using the warm where s is the standard deviation.
weather PuriNOx blend.

2.4.1 Interferences 3 RESULTS AND DISCUSSION

In the isolated zone a BOM impactor was used for
EC, TC, and DPM mass measurements. This sampler 3.1 Stone mine study
should eliminate any significant interference from
Significant reductions were observed when a mine
ambient airborne mineral dust in the mine. Samples
used PuriNOx in its entire vehicle fleet. Table 1 shows
were taken in the intake entry, upwind of the isolated
the average concentrations of EC and TC for the three
zone, so that any analyte present in the incoming
days of measurements at the two main exhausts of the
air could be determined and subtracted from its
mine.
value measured downwind. Based on the experimental
When the cold weather PuriNOx blend was used
design, in the isolated zone the only source of DPM
instead of the 35% biodiesel, the average of the two
should be from the vehicle in this area.
locations showed a 45% reduction in the EC from
Vapor-phase organic carbon (OC), generated from
DPM. There was no significant effect on the OC con-
engine exhaust but not considered to be part of DPM,
centration, and the TC concentration was reduced by
can adsorb onto the quartz filters and cause a posi-
about 36%.
tive bias of the particulate TC results in both upstream
Substituting the warm weather PuriNOx blend for
and downstream samples. As previously described, OC
the 35% biodiesel, a 57% reduction in EC and a 43%
quantities on a secondary filter, located in tandem with
reduction in TC was observed. There was again no
the primary filter, were subtracted from the OC values
significant effect on the OC concentration.
to correct for the adsorption of vapor-phase OC.
These emission reductions were observed when
Exposed filters were taken to NIOSH Pittsburgh
using 35% biodiesel as a baseline fuel. The diminu-
Research Laboratory (PRL) and analyzed in the labo-
tions would most likely be higher if PuriNOx blends
ratory for EC andTC content using the NIOSH Method
were compared to No. 1 or No. 2 diesel fuel.
5040.
These numbers do have limitations. Other factors
such as engine duty cycle and production changes
could result in day-to-day EC fluctuations. Three days
2.5 Confidence limits
of data were averaged in an attempt to minimize any
Each value for OC, EC, and TC was the average of effect that these factors would have on EC reduction,
triplicate samples for the isolated zone study and the but nevertheless, these factors could not be totally
average of three days of triplicate samples for the stone accounted for.

162

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 1. Reduction in DPM from PuriNOx compared to Table 2. The reduction of DPM in the isolated zone study.
35% biodiesel at Black River Mine.

* There was no significant (<3 µg/m3 ) EC or TC particulate

collected on the upstream samples. Therefore, they were not
subtracted from the data.
Confidence limits were calculated as described in methods
section.
∗
There was no significant (average concentration <5 µg/m3 )
EC or TC particulate collected on the intake samples.
The OC, EC, and TC concentrations were the concentra- determine how the fuel would affect the engines. An
tions measured after tandem filter correction without being odor was observed when the warm weather blend was
normalized for ventilation. The reductions are after being nor- used.
malized for ventilation.
Confidence limits were calculated as described in methods
section. 3.2 Isolated zone study
Use of the cold weather PuriNOx reduced EC by
71% relative to No. 1 diesel fuel; the warm weather
2
35 % biodiesel formulation reduced EC by 86% (see Table 2).
1.8
The cold weather PuriNOx increased OC by 41%
cold weather
1.6
PuriNOx
(no DOC was used) but the warm weather PuriNOx
1.4
warm weather
showed no significant change in OC. DPM mass was
reduced by 44% with the cold weather PuriNOx and
NO2 (ppm)

1.2 PuriNOx
1 by 57% with the warm weather PuriNOx.
0.8
The ventilation did not change between tests. The
duty cycle was the same and the same number of
0.6
cycles were run for each test. Therefore, the effects
0.4
of PuriNOx fuels should be the only cause of DPM
0.2 reduction.
0 The operator and researchers did experience some
0 1 2 3 4 5 6 7
burning of the eyes when the PuriNOx blends were
Figure 1. NO2 concentrations observed in Black River used. This was not observed when running PuriNOx
Mine when each fuel was used. Each point represents the blends at Black River stone mine. This may have been
concentration at a location for each day. because diesel oxidation catalysts were equipped on
all vehicles in the stone mine but were lacking on the
vehicle used in the isolated zone study. Some power
3.1.1 NO2 measurements loss was observed but not enough to prevent the vehicle
Figure 1 shows the daily NO2 concentrations at each from performing the duty cycle.
location and for each fuel. A reduction in NO2 was
observed.
4 CONCLUSION
3.1.2 Problems
A reduction in engine power was observed by opera- PuriNOx substantially reduced EC from DPM in two
tors with the use of PuriNOx fuels but did not seem studies. The OC concentration was either not affected
to affect the production during the days of the study. or increased (when no DOC was used). An odor and
However, this study was too short to determine how burning of the eyes can occur, but might be mitigated
much this power loss would affect the use of the vehi- if DOCs (which probably should be used with Puri-
cles. That would have to be determined after a longer NOx) are also used on the vehicle. Power loss was
use of the fuel. It would also take a longer study to observed and may inhibit the use of a water emulsified

163

Copyright © 2006 Taylor & Francis Group plc, London, UK

 fuel on certain vehicles. These studies were too short
to determine how this fuel will affect the engine.
Disclaimer: Mention of a company name or prod-
uct does not constitute endorsement by the Centers
for Disease Control and Prevention. The findings and
conclusions in this report are those of the authors and
do not necessarily represent the views of the National
Institute for Occupational Safety and Health.
REFERENCES
Birch, M.E. 2004. In P.F. O’Connor (ed.), NIOSH Man-
ual of Analytical Methods (NMAM ), Third Supplement
to NMAM, 4th Edition. Cincinnati, OH: Department of
Health and Human Services, Public Health Service, Cen-
ter for Disease Control and Prevention, National Insti-
tute for Occupational Safety and Health. DHHS(NIOSH)
Publication No. 2003–154.
Bugarski, A., Schnakenberg, G., Noll, J.D., Mischler, S.,
Crum, M. & Anderson, M. 2005. Evaluation of Diesel
Particulate Filter Systems and Biodiesel Blends in an
Underground Mine. Society of Mining and Metallurgical
Engineers Transactions, vol. 318.
Bugarski, A., Schnakenberg, G., Noll, J.D., Mischler S.,
Patts, P., Hummer, J., Vanderslice, S., Crum, M. &
Anderson, R. 2003. The Effectiveness of Selected Tech-
nologies in Controlling Diesel Emissions in Underground
Mine – Isolated Zone Study at Stillwater Nye Mine.
http://www.msha.gov/01%2D995/dpmdocs/stillwater.pdf.
Cantrell, B.K. & Watts, W.F. 1997. Diesel Exhaust Aerosol:
Review of Occupational Exposure. Appl. Occup. Environ.
Hyg. 12: 1019–1026.
Eatough, D.J., Tang, H., Cui, W. & Machir, J. 1995. Inhalation
Toxicology 7: 691.
U.S. Environmental Protection Agency (EPA). 2002. Health
Assessment Document for Diesel Engine Exhaust. Pre-
pared by the National Center for Environmental Assess-
ment, Washington, DC, for the Office of Transportation
and Air Quality, EPA/600/8-90/057F. Available from:
National Technical Information Service, Springfield, VA;
PB2002-107661.
EPA, 2002. Impacts of Lubrizol’s PuriNOx Water/Diesel
Emulsion on Exhaust Emissions from Heavy-Duty
Engines. EPA420-P-02-007.
Gerbec, E.J. & Fields, K. 2003. Diesel Particulate Con-
centrations from DPM Study at the Black River Mine,
Carmeuse Lime and Stone Inc., Butler, KY. Mine Safety
and Health Administration report PS&HTC-DD-03-
316. http://www.biodiesel.org/resources/reportsdatabase/
reports/min/20030430_min-016.pdf.
Haney, R.A. 1992. Diesel particulate exposures in under-
ground mines. Mining Engineering: 173–176.
Haney, R. 2000. Sampling and Interferences for Diesel
Particulate Matter in Underground Metal/Non-metal
Mines. http://www.msha.gov/S&HINFO/TECHRPT/
diesel/haneyimvc7.pdf.
164
Copyright © 2006 Taylor & Francis Group plc, London, UK
Heywood J.B. 1988. Internal Combustion Engine Fundamen-
tals. New York: McGraw-Hill, Inc.
Howes, P. 2000. City of Houston Diesel Field Demonstration
Project. ERMD Report 01-36.
IARC. 1989. Diesel and Gasoline Exhausts and Some
Nitroarenes. IARC Monographs on the Evaluation of Car-
cinogenic Risks to Humans. Lyon, France: International
Agency for Research on Cancer (IARC), World Health
Organization, p 458.
Kahn, G., Orris, P. & Weeks, J. 1988. Am. J. Ind. Med. 13:
405.
Kirchstetter, T.W., Corrigan, C.E. & Novakov, T. 2001.
Atmospheric Environment. 35: 1663.
Matthews, R.D. 2002. Emulsified Diesel Fuels: Literature
Review and Consolidation of Lubrizol Recommendations.
FHWA/TX-03/4576, http://www.utexas.edu/research/ctr/
pdf_reports/4576_1.pdf.
Matthews, R., Hall, M. & Prozzi, J. 2002. Emulsified
Diesel Emission Testing, Performance Evaluation, and
Operational Assessment. FHWA/TX-02/4576-2.
Mine Safety and Health Administration. 2001. 30 CFR Part
57 Diesel Particulate Matter Exposure of Underground
Metal and Nonmetal Miners, Final Rule, January 19, 2001.
Fed. Reg. vol. 66, No. 13, 5706.
Mine Safety and Health Administration. 2005. 30 CFR Part
57 Diesel Particulate Matter Exposure of Underground
Metal and Nonmetal Miners, Final Rule, June 6, 2005.
Fed. Reg. vol. 70, No. 107, 32868.
Nauss, K. 1998. Diesel exhaust: A critical analysis of
emissions, exposure, and health effects. HSA Bulletin
November: 10–13.
NIOSH. 1988. Carcinogenic Effects of Exposure to Diesel
Exhaust. Current Intelligence Bulletin No. 50. U.S.
Department of Health and Human Services, Public Health
Service, Centers for Disease Control, National Institute
for Occupational Safety and Health, DHHS (NIOSH)
Pub., No. 88–116.
Pierson W.R. & Brachaczek W.W. 1983. Aerosol Sci Technol.
2(1).
Rideout, G. 1999. Impact of Lubrizol’s PuriNOx Low-
emissions Diesel Fuel on the Exhaust Emissions from
a Heavy-duty Off-road Diesel-powered Vehicle. ERMD
Report 99–37, October, 1999.
Rosenblatt, D. 2000. Emissions Field-Testing of a Front
End Loader Operating on Lubrizol’s PuriNOx Summer
and Winter Grade Fuel Blends. ERMD Report 99–51,
February, 2000.
Rosenblatt/Ainslie 1999. Impact of Lubrizol’s PuriNOx Low-
Emission Diesel Fuel on the Exhaust Emissions and
Fuel Consumption during Cold and Standard Temperature
Operations. ERMD Report 99–37, September, 1999.
Rundell, B, et al. 1996. Occup. Environ. Med. 53: 658.
Skoog, D.A. & West, D.M. 1986. Analytical Chem-
istry, Fourth Edition. Philadelphia: Saunders College
Publishing.
Turpin, B.J. & Huntzicker, J.J. 1994. Atmospheric Environ-
ment. 28: 3061.
Wade, J.F, III & Newman, L.S. 1993. J. Occup. Med. 35: 149.
 3. Mine dusts

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Evaluation of a new real time personal dust meter for engineering studies

A.D.S. Gillies & H.W. Wu

Gillies Wu Mining Technology Pty Ltd, Brisbane, Queensland, Australia

ABSTRACT: A new personal respirable dust monitor developed by Thermo Electro Corporation under a project
funded by the National Institute of Safety and Health (NIOSH) has generated promising results in underground
coal mine testing. An Australian Coal Association Research Project (ACARP) funded study has been undertaken
to evaluate the new real-time personal respirable dust monitor particularly in engineering studies. It can quickly
highlight high dust situations and allow the situation to be corrected. The instrument has been tested for robustness
and potential to be used as an engineering tool to evaluate the effectiveness of dust control strategies. This project
has evaluated the personal dust monitor’s (PDM) ability to quickly and accurately measure changes to longwall
dust levels after implementation of an improvement. This has been done at two Australian longwall underground
mines. Results of the tests demonstrate the ability of the instrument to increase understanding of the respirable
dust exposure levels faced at underground manned points.

1 INTRODUCTION The instrument has potential to be used as an

A new personal respirable dust monitor developed by
the company Rupprecht and Patashnick in the US
under a project funded by National Institute of Safety
and Health (NIOSH) has generated promising results
in underground coal mine testing performed in the US
recently (Volkwein et al. 2004a & 2004b).
The study described in this paper has been under-
taken to evaluate this realtime personal dust monitor
(PDM) using the tapered element oscillating microbal-
ance (TEOM® system) for personal respirable dust
evaluation particularly in engineering studies. The
study is part of an international program of evaluation
being undertaken by NIOSH in the US.
The technology that forms the heart of the PDM,
the TEOM® system, is unique in its ability to collect
suspended particles on a filter while simultaneously
determining the accumulated mass. The monitor inter-
nally measures the true particle mass collected on
its filter and results do not exhibit the same sensi-
tivity to water spray as optically based measurement
approaches. The technique achieves microgram level
mass resolution even in the hostile mine environment,
and reports dust loading data on a continuous basis.
Using the device, miners and mine operators have the
ability to view both the cumulative and projected end-
of-shift mass concentration values, as well as a short-
term 15 or 30 minute running average. It is believed to
be the first personal dust monitor instrument that reli-
ably delivers a near-real-time reading. It can quickly
highlight high dust situations and allow the situation
to be corrected.
engineering tool to evaluate the effectiveness of dust
control strategies. In a US evaluation performed by
Consol Energy the mine operator evaluated the bene-
fit of a proposed new water spray-based dust control
system. Engineers measured the dust concentration
upstream and downstream of a production location
under different dust control scenarios using two of
the personal dust monitor units. By evaluating the
change in the dust loading between the upstream and
downstream monitoring sites, the company was able to
determine in a few hours which hardware configura-
tion would yield the greatest benefit to the workplace
environment.
Being a personal dust monitor, the instrument mea-
sures the airborne dust from the breathing zone region
and so has many advantages over instruments that mea-
sure from a fixed-point location. It delivers a near-real-
time reading and so can quickly highlight high dust
situations and allow the situation to be corrected.
A recent ACARP supported research project com-
pleted by one of the authors, (Gillies 2001) entitled
“Dust Measurement and Control in Thick Seam Min-
ing” ACARP C9002 has highlighted some areas for
new approaches and research to allow improvement
of dust conditions within extraction panels within
Australia’s emerging thick seam coal industry. Indus-
try, management, technical engineering staff and the
workforce all give strong recognition to the challenge
of dust as an increasing hazard particularly as higher
production levels are achieved.
The underground workplace particularly along
the longwall face is an environment with varying

167

Copyright © 2006 Taylor & Francis Group plc, London, UK

 respirable dust conditions due to aspects such as venti-
lation conditions and air velocity, shearer activity and
design, chock movement, AFC movement, manning
position, face time of individual personnel, outbye
conditions and dust levels in intake air and measure-
ment instrument behaviour.
The objective of this study is to evaluate the new
PDM in its capability as an engineering tool to quickly
and accurately measure changes to longwall dust levels
at manned points after implementation of an improve-
ment. This has been done in conjunction with the
support of mining companies through measurement
at two longwall mines.
Many mines have observed a lack of repeatability
in dust monitoring that is not easily explained. This
study has evaluated the instrument as an engineering
tool that can assess the effectiveness of one change
to improve dust levels in sufficiently short a time that
other aspects have not changed.
The measurement program on the longwall face has
focused on recognised areas for improvement through
different engineering approaches such as spray usage
and man positioning. The project aims to assist in
developing an approach to sampling measurement that
is reliable and repeatable and can ascertain efficiency
or effectiveness of engineering improvements.
The project was undertaken over four months
from April to July 2005. Some preliminary activities
occurred before this period. To ensure reliability of
testing for the international evaluation program train-
ing in use of the TEOM® system, PDM instruments
was given by Rupprecht and Patashnick personnel
in April 2005. Some technology transfer from the
project has taken place in publications, workshops and
conference deliveries during and after the project.
2 EVALUATION OF THE PDM AS AN
ENGINEERING TOOL
In the US the incidence of coal workers pneumo-
coniosis (CWP) has been declining for the past 35
years. Production levels at mines have been continually
increasing and the development of dust control tech-
nologies to protect workers has become more difficult
and complex. Improved dust monitoring of coal mine
dust concentrations offers a new means of protecting
miners’ health by more quickly identifying anomalous
dust conditions.
Despite the decline in CWP, coal mine dust is still
implicated in the US in the premature deaths of miners.
In response, the US Secretary of Labour and the Fed-
eral Advisory Committee on the Elimination of Pneu-
moconiosis among Coal Mine Workers recommended
that better monitoring of coal miner dust exposures be
used as a method to improve miner health. In consul-
tation with labour, industry, and government, NIOSH
issued a contract to Rupprecht & Patashnick Co., Inc.
168
Copyright © 2006 Taylor & Francis Group plc, London, UK
(R&P), to develop a one-piece PDM. The objective
of this work was to miniaturise the TEOM® technol-
ogy into a form suitable for a person-wearable monitor
that would enable accurate end-of-shift dust exposure
information to be available to miners. Furthermore,
any person-wearable dust monitor should minimize
the burden to the wearer by incorporating the monitor
into the mine worker’s cap lamp battery, with exposure
data continually displayed during the shift to enable
workers and management to react to changes in dust
exposure.
The PDM is configured to provide accurate res-
pirable dust personal exposure information in a form
that is convenient to wear by a miner. Respirable dust
exposure data displayed by the device has two main
objectives:
– Providing the miner and mine operator with timely
values to avoid overexposure to dust by making
any necessary changes during the course of a work
shift, and
– Computing an accurate end-of-shift statistic for a
miner’s average respirable dust exposure.
The mass sensor in the PDM, holds the key
to the accurate, time-resolved dust concentration mea-
surements. The inertial, gravimetric-equivalent, mass
measurement technique used in the device typically
provides a limit of detection on par with that of the most
sensitive laboratory-based microbalances. Similar to
the integrated sampling method, the PDM contains
a sampling system that collects particles on a filter
located downstream of a respirable cyclone. In contrast
to the current lapel worn personal method, however,
the PDM mass measurement is performed continu-
ously during a working shift in a mine instead of being
delayed by the days or weeks required for a laboratory
analysis.
The PDM is a respirable dust sampler and a gravi-
metric equivalent analysis instrument that is part of a
belt-worn mine cap lamp battery. The main compo-
nents of the device include a cap lamp and sample
inlet located on the end of an umbilical cable, a
belt-mounted enclosure containing the respirable dust
cyclone, sampling, and mass measurement system,
and a charging and communication module used to
transmit data between the monitor and a PC while
charging its lithium ion batteries for the next shift.
Figure 1 illustrates the components typically carried
by the miner. The PDM is designed to withstand the
harsh conditions found in the mine environment, with
the system designed to meet MSHA intrinsic safety
type approval requirements.
A 2.2 litre per minute flow of particle-laden air
from the mine atmosphere enters an inlet mounted on
the bill of the miner’s hard hat, and passes through
conductive tubing before reaching the Higgins and
Dewell (HD) cyclone at the entrance of the PDM.
 Figure 1. Major components of the PDM.
Figure 2. Installing a sample filter in the mass sensor.
The sample stream with respirable particles that exits
from the cyclone is then conditioned in a heated sec-
tion of tubing to remove excess moisture. As the air
stream subsequently passes through the mass sensor,
an exchangeable filter cartridge collects the respirable
particles. The mass sensor can be removed from the
PDM by a mine’s dust technician (Fig. 2), who changes
its particle collection filter and cleans the unit after the
end of each work shift.
Downstream of the mass sensor, the filtered air sam-
ple flows through an orifice used in conjunction with
a differential pressure measurement to determine the
volumetric flow rate. The system computer uses this
information to maintain a constant volumetric sample
flow by varying the speed of a DC pump.
At the heart of the TEOM mass sensor is a hollow
tube called the tapered element that is clamped at its
base and is free to oscillate at its narrow end (Fig. 3).
The exchangeable filter cartridge mounted on its nar-
row end collects the respirable particles contained in
the air stream that pass from the entrance of the mass
sensor through the tapered element. Electronic compo-
nents positioned around the tapered element cause the
tube to oscillate at its natural (or resonant) frequency.
169
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 3. Tapered element with exchangeable filter
mounted on narrow end.
As additional mass collects on the sample filter, the
natural oscillating frequency decreases as a direct
result. This approach uses first principles of physics
to determine the mass change of the filter, and is not
subject to uncertainties related to particle size, colour,
shape or composition.
Built-in sample conditioning to remove excess mois-
ture minimizes the PDM’s response to airborne water
droplets. The PDM determines the mass concentration
of respirable dust in the mine environment by divid-
ing the mass (as determined by the frequency change)
collected on its filter over a given period of time by
the volume of the air sample that passed through the
system during the same time frame.
The two battery packs inside the PDM provide
power to the miner’s cap lamp and the particle sampling
and analysis system. Laboratory and in-mine testing
conducted to date indicates a battery lifetime of 12
hours or more for both the cap lamp and mass monitor
batteries.
The PDM internally stores the readings from its
built-in environmental sensors and mass sensor for lat-
ter downloading, and provides summary information
on a continuous basis to the miner through the display
located on top of the battery case. The display contin-
uously shows the latest values for the cumulative mass
concentration, the current dust concentration, and the
miner’s end-of-shift projected exposure. Through this
interface, miners can gauge their current dust expo-
sure, as well as the effectiveness of actions taken to
reduce the in-mine dust concentration.
The PDM also allows miners and management
to initiate secondary dust loading measurements for
specific monitoring objectives without affecting the
shift-based statistics.The averaging time used for these
readings is user-selectable prior to the start of the work
shift, and can be set to a time base as short as 15
minutes for maximum instrument responsiveness.This
 capability enables the monitor to be a powerful engi-
neering tool to gauge the effectiveness of various dust
or ventilation engineering control techniques.
In 2003, NIOSH conducted extensive testing of the
PDM to determine its laboratory and in mine perfor-
mance. In addition, anecdotal data collected in the
mines showed several occasions where miners were
able to identify situations that were producing high lev-
els of respirable coal mine dust. Volkwein et al. (2004a
& 2004b) discussed examples where the PDM’s
response to unplanned changes resulted in action being
taken to keep dust exposures as low as possible.
Figure 4. Development Face PDM results – 30 minute
average.
3 AUSTRALIAN EVALUATION OF THE PDM

The objective of the study was to evaluate the PDM

for use particularly in engineering studies to evaluate
the effectiveness of dust control strategies. This project
was designed to evaluate the ability of the new PDM
to quickly and accurately measure changes in mine
section dust levels at manned points before and after
implementation of a change or an improvement.

3.1 Australian tests – Mine A

Mine A is a Hunter Valley underground longwall
mine producing low sulphur bituminous thermal coal
exported for power generation and industrial appli-
cations. PDM evaluation tests undertaken included
operator positions at development faces and long-
wall face, Beam Stage Loader (BSL) and belt transfer
points at various locations.
A total of three PDM units were made available
for engineering evaluation tests. Tests were initially
undertaken at a development face to monitor the dust
exposure levels of various equipment operators. The
PDM units give both 15 minutes (MC1) and 30 minute
(MC0) rolling averages of dust concentration and it
was decided initially that each monitoring test under-
taken should last for at least 45 to 60 minutes. PDM
units were put on continuous miner (CM), bolter and
shuttle car (SC) operators at 8:15 pm. During the tests
the face crew was replaced at 9:10 pm by the second
crew as the first crews were released for crib break.
The results of the PDM tests are shown in Figure 4
as 30 minute average and in Figure 5 as 15 minute
average dust levels.
By comparing the dust levels shown in Figures 4
and 5 it is concluded that for the engineering evaluation
purpose it is better to use the 15 minute rolling average
dust concentration then 30 minute rolling average dust
concentration as the 15 minute average gives a quicker
response to monitored changes during the test and
shows more significant dust concentration variations.
During the tests an unplanned event took place. The
end cap of ventilation ducting in the inactive face at
A heading of the development section was sucked in
Figure 5. Development Face PDM results – 15 minute
average.
and caused reduction in the ventilation air quantity
available to the face being monitored from 7.5 m3 /s to
4.3 m3 /s. This caused a significant loss of suction head
in the ventilation ducting at the face resulting in the
dust-laden air at the face billowing back onto opera-
tors. All PDMs worn by the three operators have reg-
istered sharp rises in dust level. In fact this unplanned
event was first noticed by one of the operators who had
checked the real time display on the PDM he was wear-
ing at the time. The failure of the end cap piece in the
inactive face was soon rectified and the normal ven-
tilation flow re-established. Readings from all PDMs
show the immediate reduction in duct concentration.
Tests were carried out at Mine A longwall face to
monitor the dust suppression efficiency of sprays in
the BSL and at a belt transfer point where the longwall
belt and the main trunk belt met. For the BSL test, one
PDM was placed outbye of BSL, the second PDM was
placed on top of the BSL inbye of the spray and the
third PDM further inbye of the BSL at Chock No 8.
Results of BSL tests are as shown in Figure 6.
During the test, BSL sprays were on initially and
then disconnected for about 30 minutes and then con-
nected again. The results show that with the sprays
off dust concentration levels downstream of BSL were
dramatically increased while the dust concentration
level upstream of BSL remained constant with little
variations.
It was found that the fluctuations in dust levels mea-
sured by the PDM upstream of the BSL correlated well

170

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 6. Mine A Longwall BSL PDM results – 15 minute
average.
Figure 7. Mine A Longwall belt transfer point PDM
results – 15 minute average.
with whether there is coal loaded on the conveyor belt
or not. When there is no coal loaded on the belt the dust
levels of intake air upstream of the BSL were measured
at less than 0.2 mg/m3 . It is possible to draw a horizon-
tal line as shown in Figure 6 to indicate whether there
is coal on the belt or not.
Results of Mine A longwall belt transfer point to the
mains belt PDM measurements show that an increase
in dust levels of more than 0.2 mg/m3 was observed
at a belt transfer point (Fig. 7). This increase doubled
the concentration of respirable dust levels in the intake
air. At the time of measurements the sprays at the belt
transfer point were not connected.
A second evaluation of a belt transfer point occurred
where one Mains belt met another Mains belt at
90 degrees. Similar results were observed as with
the previous test at the Longwall belt transfer. Dust
levels were doubled as air passed across the belt trans-
fer point. Again, there were no dust sprays operating
during the test.
Measurements were carried out at Mine A longwall
face to monitor the dust level experienced by shearer
and chock operators in a unidirectional mining cutting
sequence. Results of these tests are shown in Figure 8.
For the shearer operator test, one PDM (Unit #139)
was worn by a mine person who shadowed one of the
shearer operators for about 30 minutes during unidi-
rectional cutting. The other two PDM were measuring
dust levels outbye of the BSL (Unit #118) and at
171
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 8. Mine A Longwall shearer operator PDM results
under unidirectional cutting.
Figure 9. Belt transfer point PDM results – 15 minute
average.
Chock No 8 position (Unit #134). The shearer posi-
tion data was downloaded from the mine monitoring
system and indicated that the shearer was cutting from
maingate to tailgate first and then cutting from tailgate
back to maingate during the test. The results indicated
that the shearer operator was subjected to high dust
level exposure when cutting from maingate to tailgate.
When cutting from tailgate to maingate the dust level
experienced by the shearer operator was much lower.
3.2 Australian tests – Mine B
Mine B is also located in the Hunter Valley and is
considered to be one of the lowest cost longwall oper-
ations of its kind in Australia. Currently Mine B has an
annual production of about 6.5 million tonnes. PDM
evaluation tests undertaken included operator posi-
tions at development faces and longwall face, BSL,
belt transfer points and belt tripper drives at various
locations. Tests on effectiveness of an air stream hel-
met for isolation of miners to dusty atmospheres were
undertaken.
Results of PDM tests on a belt transfer point are
shown in Figure 9. Information about the tonnage on
the belt during the tests was also obtained from the
mine control and monitoring system. It should be noted
that the tonnage was measured about 1 km away from
the belt transfer point. Therefore, the tonnage on belt
data was shifted horizontally along the timeline to take
this into account.
 Figure 10. PDM results at development face during cut
through holed through.
Figure 11. Photographs of air stream helmet and the filters
used.
It can be seen that the dust concentration measured
correlates well with the amount of coal transported
on the belt. The more coal transported on the belt,
the higher dust concentration levels resulted at belt
transfer point.
Tests were undertaken at a development face to
monitor the dust exposure levels of various equipment
operators.
During the shift, due to a cable problem with a face
shuttle car, cutting was delayed for two hours. Ventila-
tion at the development face was well maintained and
dust levels appeared consistent for all face operators.
Towards the end of shift, a hole-through in mining the
cut through from Heading A to B occurred. Ventila-
tion at the face was disturbed when the hole-through
occurred and dust concentration levels experienced by
CM operator, bolter and SC driver were increased.
These high dust results can be seen in Figure 10.
In Figure 10, it also can be seen that before hole-
through, face ventilation condition was deteriorating,
as the ventilation ducting was not extended. The dust
levels experienced by both CM operator and bolter,
as they were standing right behind the machine were
gradually increased. However, dust levels experienced
by the SC driver remained fairly constant before the
hole-through.
Tests on stream helmets were carried out at the
belt transfer point discussed previously. Two air stream
helmets were used with one worn under normal oper-
ating condition and the other worn with both the pre
and main filters (as shown in Fig. 11) removed. All
three PDMs were used, one sampling the background
atmospheric dust level and the other two sampling the
air inside the two test air stream helmets. The results
of the air stream helmet tests are shown in Figure 12.
An average dust concentration of 0.05 mg/m3 was
measured inside the normal operating air stream
172
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 12. Air steam helmet PDM results – 15 minute
average.
helmet during the 40 minutes test period. This demon-
strates that the filters used by air stream helmet can
filter out most of the respirable dust. Without the fil-
ters in place, average dust concentration inside the
air stream helmet was similar to that of the outside
atmospheric were consistently higher than the dust
levels measured in background atmosphere. A similar
phenomenon was reported by others when attempting
to measure dust levels inside and outside air condi-
tional cabs (Volkwein 2005). It was concluded that an
enclosed space acts as a dust trap when a jet stream
injects dust laden air into a constrained space leading
to higher than background dust level. In addition the
jet stream in the enclosed space would keep the dust
suspending longer.
Caplan et al. (1973) maintain that in air streams
with velocities up to 1.5 m/s neither the air velocity
nor the cyclone inlet orientation has any impact on the
dust concentration measured by a sampler. However,
at air velocities over 1.5 m/s, both the air velocity and
the cyclone inlet orientation have an impact. Cecala
et al. (1983) found that when the Dorr-Oliver cyclone
inlet is pointed directly into the wind, it over sam-
ples when the air velocity exceeds 4 m/s. At very high
velocities of 10 m/s it over-samples by 35 percent.
When the cyclone inlet is at a right angle to the wind
or pointed downwind it under-samples when the air
velocity exceeds 1.5 m/s.
Cecala et al. also tested a shielded cyclone to see if
a shield would reduce the over- and under-sampling.
The shield was a 25 mm wide strip of aluminum sheet
bent into a cylinder. This cylinder was then wrapped
around the top of the cyclone and bolted to the hole
in the back of the vortex finder clamp. Testing showed
that the shield successfully reduced both the over- and
under-sampling to within 14 percent of the true value
when tested at 10 m/s.
These evaluations were done with the traditional
lapel worn personal samplers with ordinary pumps
operating across the normal range of flow rates. Flow
rates from these pumps are affected by conditions such
as the resistance of the filter as it is loaded during
sampling and hose arrangement. The pump used by
 Figure 13. Shearer operators and the chockman PDM
results – under bidirectional cutting test 1.
Figure 14. Shearer operators and the chockman PDM
results – test 2 (Shearer position plotted).
the PDM has a self regulating flow rate function to
correct the response to external conditions and main-
tains a constant flow rate throughout the measurement
period. Examination of flow rates recorded in PDM
data files during the air stream helmet tests showed
that through out the tests the flow rates of the three
PDM units remained at a constant of 2.2 litres per
minutes. Therefore it should not be either over or under
sampling as suggested by Cecala et al.
Two PDM Tests were undertaken to examine the
dust exposure levels of shearer operators and the
chockman along longwall face during bidirectional
cutting. As shown in both Figures 13 and 14 it was
found that when the shearer was cutting from maingate
to tailgate, both maingate and tailgate shearer opera-
tors can experience higher dust concentration levels
than when snaking at either end of the face or when
cutting from tailgate to maingate.
In general the chockman experienced less dust
than shearer operators during cutting as the chockman
usually stands outbye of the shearer. However when
snaking at the tailgate end the chockman may experi-
ence short periods of high exposure as he stands inbye
of the shearer. Advances in automation of shearer cut-
ting and chock advance and reliability of systems will
influence man positing and exposure levels.
Measurements were also taken inbye and outbye of
the belt tripper drive located about 500 m outbye the
longwall face as shown in Figure 15. There was a spray
in operation but inappropriately placed for dust control
purpose. Spray operation was not affecting dust levels
inbye the tripper drive. The results in the following
173
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 15. Belt tripper drive PDM results – 15 minutes
average.
Figure 16. BSL sprays PDM test 1–15 minutes average.
figure showed higher dust level inbye this dust transfer
point and the benefits of installing dust control at the
tripper drive.
Efforts were also done in examining the effects
of sprays installed within the Mine B longwall BSL
on dust suppression efficiency. Two tests were under-
taken. All three PDMs were used with one placed 20 m
outbye the BSL, another at about 1m in front of the
BSL outlet and the other at Longwall face Chock No 8
position. The results of measurements around the BSL
with spray on and off are as shown in Figure 16.
It was found that as the Longwall face was located
near a cut through less than 20 percent of the total
Longwall air was passing across the BSL in the A
heading and the rest of the intake air flowed directly to
Longwall face by the cut through. This dilution led to
a very mild increase in dust levels measured at Chock
No 8 position with the sprays at the BSL turned off
even though the PDM placed close to the BSL outlet
had measured a dramatic increase in dust levels after
the sprays were turned off.
One of the PDM used ran out of power and switched
itself off during the test as the PDMs had been in use
for more than 12 hours. A second test on BSL sprays
with better placement of the PDM units to take into
consideration of intake air split in the cut through was
conducted and the results are shown in Figure 17.
It is shown that with the BSL sprays switched off,
the dust concentration levels inbye of the BSL have
increased dramatically. Dust concentration levels at

Figure 17. BSL sprays PDM test 2–15 minutes average.
Figure 18. Comparison of measured and calculated dust
concentrations at Chock No 8.
Chock No 8 position have also increased but more
modestly. This is because the dusty air from the BSL
was mixed with about twice the quantity of relatively
clean intake air from the cut through.
About one third of the total intake air available to
the Longwall face was from A heading across the
BSL. Dust concentration levels measured by PDM
confirmed the air stream ratio is about 1 to 2. Based on
the ratio, it is possible to calculate the expected dust
concentration at downstream location (Chock No 8)
by knowing the upstream dust levels in intake air (out-
bye of BSL) and dusty air (inbye of BSL). Figure 18
shows a good correlation between the measured and
calculated mixed air dust levels at Chock No 8 position.
4 CONCLUSIONS AND RECOMMENDATIONS
Based on the tests conducted, it is concluded that the
PDM has demonstrated its potential use as an engi-
neering tool to locate and assess various sources of dust
during normal mining operations. The principles and
concepts used to identify and fix some of the higher
dust levels are generally common sense and would be
easy for most miners to understand.
However, to make the most effective use of this
information, training and experience in using this type
of technology will be very important. Experience with
the data from the unit will help miners gain confidence
to use the information to maintain reduced or safe dust
174
Copyright © 2006 Taylor & Francis Group plc, London, UK
levels during mining. Further in-mine trials are neces-
sary to determine the long term durability, stability and
maintenance requirements for this new dust monitor.
Mine managements from several mines have shown
interests in obtaining the units when they meet IS
standards in Australia and are commercially available.
Possible improvements to PDM for Australian mine
applications which could be considered include provi-
sion of an option to display histograms of 15 minute
instead of or as well as 30 minute results, use of an
adaptor to recharge battery through an Australian style
cap lamp bracket. Exchange cap lamp for Australian
type lamp that can power an Air stream helmet and
expand keypad to allow the keying in underground
location. Incorporation of “PED” and “Tag” options
would also be an advantage.
ACKNOWLEDGEMENTS
The authors are grateful for the support of personnel
from Thermo Electron Corporation and NIOSH. They
acknowledge the assistance of Tim Harvey, Senior
Ventilation Engineer, Anglocoal, Jim Sandford, Group
Safety Manager, Xstrata Coal, Bevan Kathage, Con-
sultant and the various Australian mine site managers,
engineers and ventilation officers who supported this
ACARP project. They thank the organizations ACARP
for project funding and Coal Services for significant
cooperation and support. The efforts of these individu-
als and groups ensured that the principal development
and mine site testing aims of the project were accom-
plished and a significant contribution made to future
mine health and safety in Australia and overseas.
REFERENCES
Caplan, K.J., Doemeny, L.J. & Sorenson, S. 1973. Evalua-
tion of coal mine dust personal sampler performance. U.S.
Department of Health, Education, and Welfare contract
PHCPE-R-70-0036.
Cecala, A.B., Volkwein, J.C., Timko, R.J. & Williams, K.L.
1983. Velocity and orientation effects on the 10-mm Dorr-
Oliver cyclone. Pittsburgh, PA: U.S. Department of the
Interior, Bureau of Mines. RI 8764. NTIS No. PB 83-
214387.
Gillies, A.D.S. 2001. Dust measurement and control in
thick seam mining. Australian Coal Association Research
Program Grant C9002, August 2001.
Volkwein, J.C., Thimons, E., Yanak, C., Dunham, D.,
Patashnick, H. & Rupprecht, E. 2004a. Implementing
a New Personal Dust Monitor as an Engineering Tool.
Proc 10th US Mine Ventilation Symposium. Rotterdam:
Balkema.
Volkwein, J.C., Vinson, R.P., McWilliams, L.J., Tuchman, D.P.
& Mischler, S.E. 2004b. Performance of a new personal
respirable dust monitor for mine use. Report of Investi-
gations 9663, National Institute for Occupational Safety
and Health, Pittsburgh Research Laboratory.
Volkwein, J.C. 2005. Personal communication.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Evaluation of SKC Inc. Dust Detective

E.E. Hall, R.P. Vinson & J.C. Volkwein

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute
for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

ABSTRACT: Assessment of workplace exposures is an important tool for helping to minimize dust-related
occupational illness and diseases. Real-time particulate monitoring can effectively locate areas where dust
controls are needed and determine how well they are working. An affordable, person-wearable, real-time dust
monitor is needed. The SKC Inc. Dust Detective (SKCDD) fulfills those requirements (Disclaimer: Mention of
any company or product does not imply endorsement by NIOSH). The SKCDD consists of a disposable sampling
tube (DST) connected to a small hand-held sampling pump. The SKCDD was developed through a Cooperative
Research and Development Agreement (CRADA) between the National Institute of Occupational Safety and
Health (NIOSH) and SKC Inc.The relationship of dust concentrations measured by personal gravimetric samplers
to those measured by the SKCDD was determined in a laboratory aerosol chamber. A comparison of the means
and the relative standard deviation of triplicate measurements of each type of sampling device demonstrated the
SKCDD to be a viable alternative means for measuring dust. While it provides accurate measurements for specific
coal types, it requires correction factors (those correction factors still need to be calculated) for other coal types.

1 INTRODUCTION monitors (Cantrell et al. 1993). The filter method

Medical studies have shown that prolonged inhalation
of excessive levels of respirable coal mine dust can
lead to coal workers’ pneumoconiosis, a disease that is
irreversible and can be progressive, disabling, and ulti-
mately fatal. Thousands of underground and surface
mine workers who are potentially at risk of develop-
ing adverse health effects from breathing respirable
coal and silica dusts. The Coal Mine Health and Safety
Act of 1969 included a Black Lung Benefits Pro-
gram. Over 104,000 beneficiaries received benefits in
2004. The Black Lung Disability Trust Fund provided
over $292 million in monthly benefits to beneficiaries
while the Division of Coal Mine Worker’s Compen-
sation administered another $380 million in monthly
benefits (Office of Workers’ Compensation Programs
2004). Research and development in the measurement
and control of mine dust will help reduce the number
of pneumoconiosis cases as well as minimize worker
compensation expenses.
The assessment and sampling of dust levels can
be challenging because of the variable composition
of the dusts and the constantly moving work area
(Hearl & Hewett 1993). Currently, dust levels in min-
ing are either measured gravimetrically (Raymond
et al. 1987), using filters and the accumulated dust
mass in a given quantity of air, or through the
use of instantaneous electronic or direct-reading dust
can take several weeks to process before results are
reported to the mine, which creates a delay in iden-
tifying and correcting the problem. Current available
instantaneous or direct-reading monitors are not prac-
tical for routine personal monitoring because of their
complexity, size and expense (Volkwein et al. 2000).
The U.S. Department of Labor’s Advisory Com-
mittee on the Elimination of Pneumoconiosis Among
Coal Mine Workers recommended the development of
both fixed site and portable continuous respirable dust
monitors that could be used to enhance their compli-
ance monitoring program. In response to this request,
NIOSH examined several approaches. A personal dust
monitor (PDM) has been designed, built and at present
is proving to be an accurate and reliable dust moni-
tor (Volkwein et al. 2000). However, when the PDM
becomes available to industry, it will cost several thou-
sand dollars (present estimates range from $7,000 to
$15,000). A more affordable, person-wearable, real-
time dust monitor is now available that will provide
the coal miner with a good estimate of his/her expo-
sure to dust. The monitor is called the SKC Inc. Dust
Detective (SKCDD) and will be less than $1,000 to
purchase.The SKCDD empowers miners to self-assess
their exposure to respirable dust in real time.
The PDM is a quantitative instrument while the
SKCDD is a qualitative tool. The PDM is a more
expensive instrument yielding greater accuracy and

175

Copyright © 2006 Taylor & Francis Group plc, London, UK

 more user options, while the SKCDD provides fewer sensor, air pump, control switches, and a digital dis-
options and still maintains accuracy at a lower cost. play.The DST dimensions are 1/2-in in ID by 4 inches in
Both instruments will be very useful in providing length. The tube contains two cylinders of porous foam
mine operators the information they need to make con- that are particle size pre-separators at the inlet end and
trol technology and administrative changes to reduce a glass fiber filter at the outlet end.The outlet end of the
miner exposure when needed. DST is connected to a sampling pump/monitor with
a section of 1/4-in ID flexible tubing. The disposable
single-use tube contains a respirable size classifier and
2 DESCRIPTION OF INSTRUMENTS the pressure drop filter media and can be assembled
for a few dollars per tube (Volkwein et al. 2000).
2.1 Personal gravimetric samplers As the respirable dust collects on the glass fiber
filter there is a corresponding increase in pressure
Flow-controlled coal mine dust personal sampler units drop across the filter. The pressure drop and air flow
(CMDPSU) are currently used in mines; they are man- through the pump are continuously monitored and
dated by 30CFR part 74. A CMDPSU consists of used by the pump’s programmable microprocessor
a Mine Safety Appliances (MSA) Escort Elf pump to calculate the dust concentration. The back pres-
operated at 2 lpm, connected to a sampling head by sure increases linearly with dust particle buildup on
a section of 1/4-in inside diameter (ID) flexible tub- the filter. The instruments are equipped with only
ing. The sampling head consists of a 37-mm diameter three buttons, which allow for simplicity in program-
MSA filter cassette connected to a 10-mm diameter ming the units, entering a calibration factor, and test
Dorr-Oliver cyclone size preselector. The CMDPSU parameters (Fig. 2).
were used to sample coal dust aerosols from the
Marple chamber, a laboratory dust exposure cham-
ber designed to evenly disburse dust throughout the
chamber (Marple & Rubow 1983). Filters were pre-
and post-weighed at the Pittsburgh Research Lab-
oratory (PRL) in a controlled atmosphere. NIOSH
receives filters unassembled; therefore the filters were
prepared without the tamper-resistant backflow valve
or the inner stainless steel support wheel since they
were not necessary for laboratory testing. Pump flows
were checked weekly with a Gilian Bubble Flow
Meter, a primary standard flow measurement device.
Twelve personal samplers were arranged for each test
in groups of three, so that each grouping was evenly
spaced about the central portion of the chamber at
about the same elevation.
Figure 1. SKC Dust Detective disposable sampling tube.

2.2 SKC Dust Detectives

One of the main components of the SKCDD is the dust
detector tube. The dust detector tube was developed to
provide inexpensive, short-term measurement of the
cumulative personal dust exposure of a worker during
a shift. The dust detector tube models itself after the
concept of a radiation dosimeter or, more precisely,
after sorbent detector tubes used to measure exposure
to various gases. The SKCDD software converts differ-
ential pressure to cumulative respirable dust exposure
and short term exposure.
The SKCDD consists of a disposable sampling tube
(DST) connected to a small hand-held sampling pump.
Flow-controlled sampling pumps manufactured by
SKC Inc. (Dust Detective) were operated at a flow
rate of 0.250 lpm to draw coal dust aerosols into the
dust detector tubes. The sampling pump consists of
programmable microprocessor, flow sensor, pressure Figure 2. SKC Dust Detective.

176

Copyright © 2006 Taylor & Francis Group plc, London, UK

 3 METHODS AND MATERIALS
Dust concentrations were measured by personal gravi-
metric samplers and by SKC Dust Detective (SKCDD)
in a laboratory aerosol chamber. By comparing the
means and the relative standard deviation of tripli-
cate measurements of each type of sampling device,
a correlation could be determined. A regression anal-
ysis was done to obtain the relative standard deviation
(RSD), slope, and regression coefficient (R2 ) values.
3.1 Test aerosols
Tests for this study were conducted in the Marple
Chamber. They were comprised of three separate trials
of five different coal types for a total of fifteen coal
dust samplings, where MMD is the mass median diam-
eter and GSD is the gravimetric standard deviation.
The coal types were:
• Keystone Black, 4.66 MMD, 3.04 GSD
• Illinois #6, 5.72 MMD, 3.15 GSD
• Pittsburgh Coal 18 µm, 11.01 MMD, 2.79 GSD
• Pittsburgh Coal 10 µm, 4.05 MMD, 1.98 GSD
• Pittsburgh Coal 5.77 µm, 3.68 MMD, 2.59 GSD
The size distribution measurements were made in
the Marple Aerosol Chamber using Marple Personal
Cascade Impactors 290 Series.
3.2 Test procedure
The guidelines for each set of three tests are as follows:
Test 1: The Marple Aerosol Chamber was brought
to an MRE equivalent (MRE is a multiplier of 1.38
required to meet federal standards (30CFR)) concen-
tration of 2.0 mg/m3 . Three SKCDDs and 12 CMDP-
SUs were started at the beginning of the test. The
SKCDDs ran for 8 hours. Three CMDPSUs were
turned off at each of the MRE equivalent concen-
trations of 0.3, 0.5, 1.0, and 2.0 mg/m3 -to compare
concentrations at different mass loadings.
Test 2: The Marple Aerosol Chamber was brought
to an MRE equivalent concentration of 2.5 mg/m3 .
Three SKCDDs and 12 CMDPSUs were started at the
beginning of the test. The SKCDDs ran for 8 hours.
Three CMDPSUs were turned off at each of the
MRE equivalent concentrations of 0.6, 1.2, 1.8, and
Table 1. Summary of laboratory correlation data.
Coal type Power function y= R2 CMDPSU STD
Keystone Black 1.1x + 0.25 0.98 0.02
Illinois #6 0.72x + 0.06 0.96 0.02
Pittsburgh 10 µm 0.35x + 0.03 0.89 0.03
Pittsburgh 18 µm 0.39x + 0.03 0.94 0.03
Pittsburgh 5.77 µm 0.72x + 0.12 0.99 0.03
177
Copyright © 2006 Taylor & Francis Group plc, London, UK
2.5 mg/m3 to compare concentrations at different mass
loadings.
Test 3: The Marple Aerosol Chamber was brought to
an MRE equivalent concentration of 4.0 mg/m3 . Three
SKCDDs and 12 CMDPSUs were started at the begin-
ning of the test. The SKCDDs ran for 8 hours. Three
CMDPSUs were turned off at each of the MRE equiv-
alent concentrations of 1.0, 2.0, 3.0, and 4.0 mg/m3 to
compare concentrations at different mass loadings.
The sampling inlets for both the SKCDD and
CMDPSUs were positioned so that they were all
facing the center of the chamber. The dust concen-
trations in the aerosol chamber were monitored with
a TEOM1400a operated at 2.0 L/min. The inlet of the
TEOM was connected to a 10-mm cyclone with a 5-ft
section of ¼-in diameter tubing.
4 RESULTS AND DISCUSSION
Tests were run using concentrations as a guide for com-
parison with federal standards however for simplicity
in data analysis dust mass was used. Data was analyzed
and the standard deviation and relative standard devi-
ation for both the CMDPSUs and the SKCDDs were
calculated and are shown in Table 1.
The RSD for the CMDPSU for all coal types,
except Pittsburgh 18 µm, was lower than that of the
SKCDD, but the differences were generally minimal.
The regression coefficient was 0.86 or greater for all
coal types.
The averages of the three SKC Dust Detective
instruments were compared with the CMDPSU data.
The SKCDD dust masses were plotted against the
CMDPSU dust masses. Linear regression was used
to find the RSD, slope and R2 values. As shown above
in Figure 3, the Dust Detective response for each coal
type was linear. The slopes for the coal types varied.
The Illinois #6 and Pittsburgh 5.77 µm had similar
slopes as did Pittsburgh 18 µm and Pittsburgh 10 µm.
This is attributed to the differences in how each coal
type loads the filter and changes the pressure drop.
A regression analysis was performed for the data
for all coal types. Based on the slope of the line that
was derived from the data from all coal types, a range
of ±50% is shown above in Figure 4. The ±50% is
the European Standard criteria for survey instruments
(EN 1994). Four of the five coal types fall within the
SKCDD STD CMDPSU RSD SKCDD RSD
0.14 0.02 0.13
0.10 0.02 0.11
0.06 0.03 0.06
0.17 0.03 0.14
0.19 0.03 0.16
 3.0
Dust Detective (mg)
2.5
2.0
1.5
1.0
0.5
0.0
0.0 0.5 1.0
Figure 3. Dust Detective responses to coal types.
Figure 4. Dust Detective responses optimized for calibration factor.
parameters, therefore suggesting that one calibration
factor would adequately represent those coal types.
Keystone Black was the only coal type whose slope
was out of the range. This could be due to that it is a
manufactured product and not naturally occurring; it
is a more finely ground coal material with size distri-
bution less than naturally occurring coal. Based on the
R2 values, the SKCDDs can be programmed for coal
seams specific to mine operations.
5 CONCLUSIONS
Initial work indicates that the SKC Dust Detective will
be a reliable instrument that will provide a low cost
alternate means for estimating dust concentrations.
The instrument can be calibrated for specific coal types
and is easy to program. Laboratory comparisons to
CMDPSUs show good correlation for individual coal
types. These instruments will be most applicable when
big differences in size distribution are not expected.
Future work will include water spray response (for
longwall operations), determining correction factors,
and conducting underground mine studies. The instan-
taneous feedback on dust level exposures that the
SKC Dust Detective provides will help empower the
worker and mine management to improve the mining
conditions and protect the worker’s respiratory health.
6 DISCLAIMER
The findings and conclusions in this report are those of
the authors and do not necessarily represent the views
Copyright © 2006 Taylor & Francis Group plc, London, UK
Keystone
R2 = 0.98 Illiniois #6
R2 = 0.96 Pittsburgh 5.77␮m
R2 = 0.99
Pittsburgh 18␮m
R2 = 0.94
Pittsburgh 10␮m
R2 = 0.89
1.5 2.0 2.5 3.0 3.5
Reference filter (mg)
of the National Institute for Occupational Safety and
Health.
REFERENCES
Cantrell, B.K., Williams, K.L., Watts, W.F. Jr., et al.
1993. Mine Aerosol Measurement, Aerosol Measure-
ment. Principles, Techniques and Application, Chap. 26,
pp. 591–611. K. Willeke and B.A. Baron, Eds. Van
Nostrand, New York (1993).
European Standard. EN 482 (1994).
Hearl, F.J. & Hewett, P. 1993. Problems in Monitoring:
Dust Levels Within Mines. Occupational Medicine: State
of the Art Reviews 8(1):93–109 (January–March 1993).
Marple, V.A. & Rubow, E.L. 1983. An Aerosol Cham-
ber for Instrument Evaluation and Calibration. American
Industrial Hygiene Association Journal 44(5):361–367.
Office of Workers’ Compensation Programs: Black Lung
Benefits Act 1996. DOL, Washington, D.C. 1997.
Raymond, L.D., Tomb, T.F. & Parobeck, P.S. 1987. Respirable
Coal Mine Dust Sample Processing. MSHA IR 1156.
U.S. Department of Labor: Report of the Secretary of
Labor’s Advisory Committee on the Elimination of
Pneumoconiosis Among Coal Mine Workers, recom-
mendations 8 and 17. DOL, Washington, D.C. (October
1996).
Volkwein, J.C., Schoeneman, A.L. & Page, S.J. 2000.
Laboratory Evaluation of Pressure Differential Based
Respirable Dust Detector Tube. Applied Occupational and
Environmental Hygiene 15(1):158–164.
178
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Comparison of three side-by-side real-time dust monitors in a duct

using average and peak display dust levels as parameters of performance
evaluation

B.K. Belle
Johannesburg, South Africa

ABSTRACT: In South Africa, the monitoring of dust in the mines is a requirement in terms of Section 12.2
and 12.3 of the Mine Health and Safety Act (MHSA) of 1996. In order to ascertain the magnitude and range of
dust levels and to react when an unhealthy dust exposure occurs, a real-time personal monitoring instrument for
mineworkers is undoubtedly required. This paper discusses a comparative study of the three real-time (PDR units)
monitors in a duct using coal and sandstone dust. The Higgins-Dewell (HD) and Dorr-Oliver (DO) type cyclone
operated in accordance with the international size-selective curve were used as ‘true samplers.’The average and
peak display levels recorded by the three PDR units positioned randomly side by side, in the duct were analyzed
using statistical techniques. The results of the study have showed that the dust levels measured with the three
PDR units were not significantly different to the HD sampler data. Interestingly, the results showed significant
differences in measured dust levels between HD and DO cyclones positioned side-by-side. The implication of this
finding is that the majority of real-time monitors (e.g., Tapered Element Oscillating Microbalance (TEOM)) use
these as a ‘reference sampler.’ This means that, based on measured differences found between the two cyclones,
the introduction of TEOM for legal monitoring purposes may create ambiguity in its current state, i.e., agreement
on the use of ‘true cyclone.’ The study demonstrated that, if the DO cyclone were used in the TEOM, it would
measure significantly lower dust levels than the HD cyclone. Therefore, consensus on a ‘true sampler for use in
real-time monitors’ must be established in the mining industry.
Pair-wise t-test analyses were performed to compare the three PDR units using the average and peak recorded
level. The study indicated that when peak value is used to evaluate the performance between instruments, resulted
in different inferences on the recorded levels when compared with the average value. The implication of this is
that in practice, the random selection and use of a real-time monitor for engineering dust control application may
be in favor or against the seriousness of the dust problem. Although the recorded levels show the differences in
dust levels, ANOVA results showed the contrary: dust type, monitoring units or position were not the sources of
variation in the measured average and peak dust levels between the three PDR units. Light scattering monitors
depend solely on air movement to move the dust particles into the sensing zone. It is unknown, if the particle
charges have any specific effect in terms of their movement towards the sensing chamber that could have
contributed to the recorded differences. It is proposed that, for real-time monitor evaluation, the use of ‘peak
display’ level may ascertain the probable sources of variations. The intention of this paper is not to suggest that
the peak levels should be used in place of average levels for exposure monitoring, rather an evaluation parameter
in understanding of variations experienced by researchers.

Keywords: Peak dust levels, real-time monitor, coal dust, silica dust, evaluation, mining

1 INTRODUCTION undoubtedly required. Mainly in the USA, the need

In South Africa, the monitoring of dust in the mines
is a requirement in terms of Section 12.2 and 12.3
of the Mine Health and Safety Act (MHSA) of 1996.
In order to ascertain the magnitude and range of dust
levels and to react when an unhealthy dust exposure
occurs, a real-time dust monitor for mineworkers is
for the development of a real-time continuous res-
pirable dust monitor has resulted in a new product
based on Tapered Element Oscillating Microbalance
(TEOM) principle. In South Africa, the quest for real-
time monitoring has resulted in a number of research
projects that have focused on the issues pertaining to
the assessment of dust hazards in mining operations

179

Copyright © 2006 Taylor & Francis Group plc, London, UK

 (Unsted, 1997; Biffi et al., 2000). The research
work has shown that the use of direct-reading light-
scattering instruments is not reliable due to their inher-
ent sensitivity to particulate matter other than dust.
Against this background, the search for an improved
or an alternative instrument capable of measuring dust
more accurately and reliably is continuing. Therefore,
any new information on the real-time dust monitoring
techniques or their performance evaluation would be
beneficial to the mining industry worldwide.
Figure 1. Pictorial view of the Polley duct with a rectangular
rolling sampling table (right).

2 REAL TIME MONITORS

Direct-reading instruments or real-time monitors
based on light scattering are available to estimate expo-
sure to dust in underground mines. Real-time direct-
reading instruments for mine dust have been used
worldwide for routine engineering control and risk
assessment purposes over two decades due to their
added benefits when compared with the gravimetric
samplers. All the available real-time monitors are cal-
ibrated using ‘mono-disperse’ particles (Arizona road
dust). However, each monitor to be used underground
requires a user-determined ‘correction factor’obtained
from a side-by-side gravimetric size-selective sam-
pler, evaluated with ‘poly-disperse’mine specific dust.
There is no ‘absolute correction factor’ available for an
individual real-time monitor. The ‘correction factor’
changes with the history of the sampling data obtained
in side-by-side comparisons of the real-time moni-
tor and the type of gravimetric size-selective sampler
used.
Direct reading instrument evaluation is not new to
the mining industry. Various studies (Williams and
Timko, 1984; Page and Jankowski, 1984; Gero and
Tomb, 1988; Tsai et al., 1996, Baldwin et al., 1997;
Tarkington et al., 1997; Thorpe and Walsh 2002) have
evaluated different types of real-time monitors for their
usage as personal or area monitors. The conclusions
from these studies are similar in terms of their recom-
mendations on the usage of the real-time monitors, but
with varying degree of certainty. It is well known that
the use of a real-time monitor as a stand-alone unit is
not recommended for personal exposure assessment
purposes but rather it is more suited to the identifica-
tion of dust trends during a working shift. Currently,
there is no consensus standard on the selection of a
suitable real-time instrument for use to the industry.
Field trials using real-time dust monitors used in con-
junction with the visualization system revealed that
its response could vary significantly from one day to
the next (Thorpe and Walsh, 2002). The sources of
variability of the real-time monitors can be attributed
to dust levels, dust type, dust size, air velocity, moni-
tor orientation and contamination of optics, etc. This
7,0 m
Air movement in closed circuit
Dust
briquette
2,0 m
Sampling instruments
Dust
generator
Figure 2. Line diagram of the Polley dust duct operation.
paper investigate the results of three real-time moni-
tors positioned side-by-side in a duct and evaluates the
instruments based on the average and peak measured
dust levels as parameters of evaluation.
3 LABORATORY STUDY
This section of the paper discusses the laboratory eval-
uation of three real-time monitors (PDR) positioned
side by side along with gravimetric samplers in a
laboratory known as the Polley duct (Figure 1).
3.1 Polley duct
The Polley duct consists of a closed-circuit duct and
two dust generators (Figure 2). The closed-circuit duct
consists of two sections: a horizontal section and a ver-
tical section. The horizontal section is the main section
and measures 7.0 m long by 2.0 m high by 0.7 m wide.
The air flows along the horizontal channel into and
along the top half of the large horizontal section. It
returns along the bottom half of the large horizontal
section through a flow-straightening section and flows
along the lower, small horizontal channel into, and
upwards in the vertical section to close the circuit. The
duct also has other auxiliary parts such as time relays,
two fans to circulate the air and a third to exhaust the
dust-laden air through a filter to atmosphere, and a
dust briquette press.

180

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 3. Instrument positions on the sampling table.
3.2 Test instruments and methodology
The three real-time monitors that were used for evalu-
ation purposes were commonly known as PDR or MIE
DataRam (USA). The units operate on forward light-
scattering particle detection principle, which relies on
ambient air movement to introduce particles into the
sensing chamber. The PDR real-time monitor displays
dust level in mg/m3 in addition to TWA, Max, Min,
STEL and sampling time on the display readout. The
instrument has a preliminary Intrinsically Safe (IS)
certificate obtained from the South African Bureau of
Standards (Grupping, 2001).
For the evaluation purposes of real-time monitors,
Higgins-Dewell (HD) and Dorr-Oliver Cyclones were
used as these are the commonly used size-selective
devices used worldwide. It is assumed that HD cyclone
and DO cyclone or sampler gave zero or negligible
errors and is a representative sample of ‘true’ mea-
sured dust level in the chamber. The HD cyclone was
operated at 2.2 L/min and DO cyclone was operated
at 1.7 L/min in terms of new international harmoniza-
tion (ISO/ACGIH/CEN) size-selection curve. For each
test, the samplers were positioned side by side inside
the lower chamber of the duct (Figure 3). Each of the
three real-time monitors was randomly positioned on
locations D or E or F, while HD and DO cyclones were
positioned at location A or C.
A low air velocity (∼0.8 m/s) in the chamber was
maintained consistently for all the tests. For the study,
the instruments were exposed to two types of dust,
viz. coal and sandstone briquette dust. The quartz con-
tent of the sandstone briquette dust was 50.63%. The
real-time monitors were calibrated (zeroing) using an
airtight polythene bag supplied by the manufacturer
after each test. The test chamber did not have any
instruments to measure the size distribution of the
airborne dust in real-time. The detailed experimental
procedures are discussed elsewhere (Belle, 2002). The
cyclone inlets faced the direction of the airflow in order
to avoid the effect of nozzle inlet orientation on sam-
pler performance. New Gillian constant volume flow
pumps were used and were calibrated to the nearest
ml per minute flow rate, using a digital Gillibrator.
The gravimetric sampler dust levels were determined
in accordance with the established Department of
Minerals and Energy procedures (DME, 1997). A pho-
tographic view of the real-time monitors and their
181
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 4. Pictorial view of test real-time monitors (left) and
gravimetric samplers (right).
positions during tests in the chamber is shown in
Figure 4.
4 LABORATORY RESULTS
Initially tests were carried out to confirm that the
dust levels across the Polley duct were uniform. Both
DO and HD cyclones were positioned side-by-side in
the dust chamber and tests were conducted for both
coal and sandstone dust. Preliminary results indicated
that there is no significant difference in the measured
dust levels across the chamber. For example, the mea-
sured coal dust levels at positions B and H were
4.94 mg/m3 and 4.91 mg/m3 , respectively. Similarly,
the measured sandstone dust levels at positions B and
H were 33.35 mg/m3 and 31.90 mg/m3 respectively.
Figure 5 shows the respirable dust levels obtained
in the side-by-side comparisons of similar types of
cyclones. The correlation coefficient (r) between the
two side-by-side DO cyclones was 0.993. Similarly,
the correlation coefficient (r) between the two side-
by-side HD cyclones was 0.998. A combined plot of
the two data (r = 0.998) indicates a strong linear rela-
tionship between the two side-by-side cyclones. The
two data sets of dust values showed that concentra-
tion across the chamber was uniform during the test
conditions.
Figure 6 show the relationship between the mea-
sured dust levels using the DO cyclone and the HD
cyclone positioned randomly, side-by-side in the test
chamber. From the plot and the regression equations it
is noted that the DO cyclone measured approximately
16% less respirable dust than the HD cyclone for a
personal coal dust compliance limit of 2 mg/m3 .
The implications of this finding is that majority of
the real-time monitors use cyclones as a ‘reference
sampler’ operated in accordance with the accepted
size-selective curve. The newly developed Tapered
Element Oscillating Microbalance (TEOM) real-time
 Measured respirable dust concentration (mg/m3)

35 PDR-1 PDR-2 PDR-3 Ideal

y = 0.9912x
R2 = 0.9965
30 20
18

Real-time dust level (mg/m3)

25
16
20 14
15 12
10
10 8
5 6
4
0
0 5 10 15 20 25 30 35 2
Measured respirable dust concentration (mg/m3) 0
0 2 4 6 8 10 12 14 16 18 20
Gravimetric dust level (mg/m3)
Figure 5. Combined data of two side-by-side cyclones (HD
and DO).
Figure 7. Relationship between measured gravimetric and
recorded real-time monitor dust levels.
18
Coal Sandstone
Respirable dust measured with the

P1 P2 P3 Gravimetric
Dorr-Oliver sampler (mg/m3)

15
y = 0.7544x + 0.0612 y = 0.7335x + 0.4248 18
R2 = 0.9212 R2 = 0.8127
12 16
Average dust level (mg/m3) 14
9
12
6
10
3 8
6
0
0 3 6 9 12 15 18 4
Respirable dust measured with the SA sampler (mg/m3) 2
0
Figure 6. Relationship between measured dust levels using 1 2 3 4 5 6 7 8 9 10 11 12 13
side-by-side DO and HD cyclones in the test chamber. Test #

monitor uses the HD cyclone for its operation. This Figure 8. Relationship between average dust levels recorded
by PDR units and gravimetric sampler.
means that, based on measured differences found
between the two cyclones for the US or SA industry, the
P1 P2 P3
introduction of TEOM for legal monitoring purposes
60
may create ambiguity at its current state, i.e., agree-
ment on the use of ‘true cyclone’in real-time monitors.
Average dust level (mg/m3)

50
This study did not evaluate any imprecision of either
HD or DO cyclones. Also, from South African experi- 40

ence, by switching over to the new size-selective curve, 30

the measured coal dust levels were 11% lower than
before at 2.0 mg/m3 personal exposure limit (Belle, 20
2005). Currently there are no changes to the personal
10
exposure limit due to the change over to the new size-
selective curve. For this paper the HD sampler was 0
1 2 3 4 5 6 7 8 9 10 11 12 13
used as a reference sampler.
Test #

4.1
The results of the variation between the dust lev-
els measured by three PDR dust monitors positioned
side-by-side in conjunction with the gravimetric sam-
plers are discussed below. Figure 7 show the rela-
tionship between gravimetric and real-time monitors
for two different dust types. The solid line represents
1:1 relationship between gravimetric and real-time
monitors. The coefficient of variation (CV) of the
mean correction factors is in an increasing order for
PDR results Figure 9. Peak real-time levels recorded by PDR units for
coal and sandstone dust.
PDR instruments, P1 (8%), P2 (14%) and P3 (17%).
The lower the CV of the mean correction factor, the
more linear the response of the monitor.
Figures 8 and 9 show the average and peak (max-
imum) dust levels recorded by the PDR instruments
positioned side-by-side, randomly, in the test chamber
for two dust types respectively. Tables 1 and 2 show the
summary statistics of the average and peak respirable

182

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 1. Average and Peak dust levels using Coal Dust. 3.83 mg/m3 , respectively for coal dust. From the sand-
stone dust, it is noted that the average measured dust
Average dust levels, mg/m3 levels by the units P1, P2, P3 and gravimetric sam-
Test # P1 P2 P3 HD*
pler were 7.44 mg/m3 , 8.57 mg/m3 , 8.67 mg/m3 , and
8.40 mg/m3 , respectively. Using the combined data,
45 1.30 1.50 1.48 1.56 the average measured levels by the units P1, P2, P3 and
46 3.34 3.70 3.67 5.22 gravimetric samplers were 5.36 mg/m3 , 6.13 mg/m3 ,
47 0.91 1.05 1.07 0.55 6.19 mg/m3 , and 6.29 mg/m3 , respectively.
49 0.13 0.17 0.19 0.09 Similarly, peak dust levels recorded by the three
50 5.63 6.25 6.30 7.86 real-time units were compared. From the data it is
51 6.32 7.04 7.19 7.69 noted that the average of peak recorded dust levels
recorded by the units P1, P2, and P3 were 11.64 mg/m3 ,
Peak dust levels, mg/m3
12.86 mg/m3 , and 13.05 mg/m3 , respectively for coal
dust. From the sandstone dust data, it is noted that the
Test # P1 P2 P3 HD*
average of peak dust levels recorded by the units P1,
45 8.80 9.29 8.09 1.56 P2, and P3 were 27.63 mg/m3 , 31.95 mg/m3 , and 32.69
46 14.2 15.33 15.32 5.22 respectively. Using the combined data, the average of
47 2.18 2.72 2.63 0.55 peak levels recorded by the units P1, P2, and P3 were
49 0.23 0.26 0.33 0.09 20.25 mg/m3 , 23.14 mg/m3 , and 23.65 respectively.
50 23.84 26.62 27.18 7.86 From the plots and tables it is observed that the
51 20.57 22.96 24.77 7.69 average and peak display dust levels recorded by the
three PDR units positioned side-by-side would dif-
* Higgins-Dewell gravimetric value fer when the instruments were exposed to the same
dust cloud with inherently the same size characteristics
Table 2. Average and Peak dust levels using Sandstone Dust. (the dust source, dust generation and airborne mech-
anism). Comparison of the PDR and the HD sampler
Average dust levels, mg/m3 dust levels indicate that there is no statistically signif-
icant difference in measured levels for both dust types
Test # P1 P2 P3 HD* for all three units (high p-value). It appears that the
difference between recorded levels by the real-time
45 5.56 6.15 6.44 6.11 monitors is slightly pronounced when peak value is
46 2.99 3.43 3.50 3.35 used as a performance indicator.
47 13.78 16.37 16.18 16.77
49 4.43 5.14 5.22 4.94
50 11.30 12.83 13.17 10.99
51 7.61 8.65 8.77 9.12 5 STATISTICAL ANALYSES
52 6.41 7.45 7.41 7.49
This section of the paper discusses the analyses of the
Peak dust levels, mg/m3 data using appropriate statistical techniques. A paired
t-test was performed on the set of real-time pair dust
Test # P1 P2 P3 HD* data to determine whether there was a statistical dif-
ference in the ratio of dust levels measured between
45 31.51 34.02 6.44 6.11 real-time monitor and HD cyclone. A paired t-test of
46 7.33 8.78 8.95 3.35
47 41.99 49.95 49.61 16.77
hypotheses was developed to compare the concentra-
49 14.69 17.87 18.27 4.94 tion level ratios (mean and peak) measured with three
50 26.36 29.54 31.59 10.99 real-time monitors (P1, P2 and P3) and HD cyclone.
51 38.20 45.15 45.91 9.12 The null and alternative hypotheses for the sample
52 33.33 38.35 36.99 7.49 pairs tested were:

* Higgins-Dewell gravimetric value H0 : CR P1 = CR P2

Ha : CR P1  = CR P2

dust values obtained from the side-by-side compari-
son of the PDR monitors using coal and sandstone
dust respectively.
From the data it is noted that the average measured
levels by the units P1, P2, P3 and gravimetric sam-
pler were 2.94 mg/m3 , 3.29 mg/m3 , 3.32 mg/m3 , and
For example, in the paired t-test, hypothesis H0
states that the average dust ratios between two real-
time monitors (P1 and P2) are equal. On the other hand,
the alternative hypothesis states that the two real-time
monitors, in fact, measure different average dust lev-
els. The results of the paired t-test statistical analyses

183

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 3. Results of paired t-test (average transformed levels by the three PDR units when compared with the
values). average value.
Pair Dust # T-value p-value Hypothesis
5.1 ANOVA
P1-P2 Coal 6 −4.46 0.007 Reject
P1-P3 Coal 6 −3.59 0.016 Reject Upon noting the differences between the recorded lev-
P2-P3 Coal 6 −1.24 0.270 Accept els by the three real-time units, an analyses of variance
P1-P2 Sandstone 7 −16.30 0.000 Reject (ANOVA) were performed. Typical sources for these
P1-P3 Sandstone 7 −48.01 0.000 Reject variation are sampling type (active or passive), dust
P2-P3 Sandstone 7 −2.07 0.083 Accept types, monitor orientation, size distributions of dust,
P1-P2 Combined 13 −9.18 0.000 Reject
air velocity, sensor contamination etc. The measured
P1-P3 Combined 13 −7.68 0.000 Reject
P2-P3 Combined 13 −2.22 0.046 Reject dust concentration ratio between the real-time dust
monitors’ data (average and peak) and the reference
HD sampler data was used to perform an ANOVA
using MINITAB 13.2 statistical software.
Table 4. Results of paired t-test (peak transformed values). For this study, the sources of variation quantified
were the influence of dust type (coal and sandstone),
Pair Dust # T-value p-value Hypothesis monitoring units (P1, P2 and P3) and the position of the
dust-monitoring units in the dust chamber. Essentially,
P1-P2 Coal 6 −4.71 0.005 Reject the measured dust concentration ratio data that were
P1-P3 Coal 6 −2.42 0.060 Accept used for the analysis were in the form of CAijk (mg/m3 )
P2-P3 Coal 6 −0.54 0.613 Accept
P1-P2 Sandstone 7 −9.42 0.000 Reject and CPijk (mg/m3 ) with the following definitions:
P1-P3 Sandstone 7 −12.98 0.000 Reject CA = Ratio between real-time dust monitor and
P2-P3 Sandstone 7 −1.53 0.176 Accept average dust level measured using HD cyclone
P1-P2 Combined 13 −9.20 0.000 Reject CP = Ratio between real-time peak value and aver-
P1-P3 Combined 13 −5.93 0.000 Reject age HD cyclone
P2-P3 Combined 13 −1.09 0.295 Accept i = dust type (DT), i = 0 is a coal dust, i = 1 is a
sandstone dust
j = monitoring unit (MU), j = 0 is unit-P1, j = 1 is
unit-P2, and j = 2 is unit-P3
(for average and peak data) are given in Tables 3 and 4. k = unit position (UP), k = 0, 1, and 2 indicate the
For the analyses, a cut-off p-value of 0.05 was used sampling positions (randomly selected) across the dust
(95% confidence level). chamber respectively.
Using the average recorded level as a performance The results of the analysis of variance (ANOVA)
evaluation parameter (Table 3), a large p-value (>0.05) on the average and peak concentration ratio data are
is observed suggesting that the measured mean con- summarized in Tables 5 and 6.
centration ratios are consistent with the null hypothe- The ANOVA tables give for each term in the model,
sis. That is, the dust levels recorded by P2 and P3 are the degrees of freedom, the sums of squares (SS), the
not affected at the 95% level of confidence for coal adjusted mean squares (MS), the F-statistic from the
and sandstone dust. For the combined data, there was a adjusted mean squares and its p-value.
significant difference between the recorded dust (coal In the ANOVA table some p values were less
and sandstone) levels between all the three units, viz., than 0.05, indicating that these factors are signifi-
P1 and P2; P1 and P3 (p-value of 0.000); and P2 and cant in influencing the concentration values. From the
P3 (p-value < 0.05). ANOVA results using average and peak level data,
Similarly, using the peak recorded level as a perfor- the conclusions are summarized hereafter. The effect
mance evaluation parameter (Table 4), a large p-value of dust type on the dust concentration ratio between
(>0.05) is observed suggesting that the measured peak the real-time monitors positioned side-by-side is sig-
and mean concentration ratios are consistent with the nificant (p-value of 0.017). There is slight evidence
null hypothesis. That is, the dust levels recorded by (p-value of 0.096) of the effect of unit position on the
P2 and P3 are not affected at the 95% level of confi- measured dust levels when the units are exposed to the
dence for both types of dusts. For the combined data, same dust cloud using average value as the reference
there was a significant difference between recorded parameter of evaluation. However, the dust-monitor’s
dust (coal and sandstone) levels between the units, viz., performance is not significantly affected by the posi-
P1 and P2; P1 and P3 (p-value of 0.000). The study tion of the monitoring unit within the chamber or dust
indicates that, when peak value is used as a param- type for peak dust data (p-value >0.50). As we note
eter to evaluate the performance between monitors, from the table, the interactions between the main fac-
different inferences could be drawn on recorded dust tors do not have any significance on the measured dust

184

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 5. Results of ANOVA for average values. considered to be ‘true samplers’ is that the majority of
real-time monitors use them as a ‘reference sampler.’
Sources of variation Df SS MS F-value Pr > F The newly developed Tapered Element Oscillating
Microbalance (TEOM) real-time monitor uses the
Monitoring Unit (MU) 3 0.39 0.13 1.07 0.359 HD cyclone for its operation as a real-time monitor.
Dust Type (DT) 1 0.35 0.82 6.59 0.017
This means that, based on measured differences found
Unit Position (UP) 2 0.27 0.32 2.57 0.096
MU * DT 2 0.03 0.11 0.89 0.425 between the two cyclones, the introduction of TEOM
MU * UP 4 0.57 0.18 1.44 0.251 for legal monitoring purposes may create ambiguity
DT * UP 2 0.62 0.31 2.51 0.102 in its current state, i.e., agreement on the use of ‘true
High order Interactions 25 3.11 0.12 cyclone’. The study has demonstrated that, if the DO
Total 38 5.31 cyclone were used in the TEOM, it would measure
significantly lower dust levels than the HD cyclone
(although the HD cyclone is beneficial in terms of its
Table 6. Results of ANOVA for peak values. sensitivity to higher flow rates). Therefore, owing to
the differences observed in this study, the need for a
Sources of variation Df SS MS F-value Pr > F consensus on a ‘true sampler for usage in real-time
monitors’ which operates according to the proposed
Monitoring Unit (MU) 3 2.26 0.26 0.21 0.816 new international size-selective curve exist in the min-
Dust Type (DT) 1 0.06 0.45 0.35 0.560 ing industry. Furthermore, from the South African
Unit Position (UP) 2 2.68 0.41 0.32 0.727 experience, by switching over to the new size-selective
MU * DT 2 0.16 1.11 0.88 0.429 curve from Johannesburg curve using the HD cyclone,
MU * UP 4 7.52 2.77 2.17 0.102 the measured coal dust levels were 11% lower than
DT * UP 2 5.70 2.85 2.23 0.128
before at 2.0 mg/m3 (Belle, 2004). Currently there are
High order Interactions 25 31.9 1.28
Total 38 50.33 no changes proposed to the personal coal dust exposure
limit due to the change over to the new size-selective
curve.

levels for both average and peak data. Overall, the

ANOVA conclusively indicated that the factors such
as dust type, monitoring unit or its position are not the
sources of variation in the measured average and peak
dust levels between the three PDR units.
6 DISCUSSIONS
The following paragraphs discusses the results of the
study in light of the use of appropriate ‘true reference
sampler’ for real-time monitor and probable unknown
sources of variation in measured levels between three
PDR units.
6.1 Use of appropriate reference sampler
Using cyclones as ‘true reference’ samplers for real-
time monitor evaluation is not new to the mining
industry (Kissell et al., 2002). The conclusions from
the historic real-time studies are similar in terms of
their recommendations on its usage. In past decades,
researchers used the MRE 113a, which followed the
Johannesburg curve, as a benchmark ‘true sampler.’ In
general, the correction factors of the real-time moni-
tors could be explained by the size-dependent light-
scattering characteristics of the sensors with respect
to any of the respirable size-selective sampling con-
ventions and reference samplers.
The implications of the findings on the differences
in measured levels between HD and DO cyclones
6.2 Sources of variation between PDR units
The results of the study have showed that the dust levels
measured with the three PDR units were not signifi-
cantly different to the HD sampler data. Historically,
sources of variations in measured dust levels in real-
time monitors have been evaluated for parameters such
as dust types, dust levels, monitor orientation, particle
size, air velocity, and sensor contamination. In this
study, sources of variations evaluated in recorded lev-
els between three PDR units were dust type, monitor-
ing unit and monitor position. Although the recorded
levels show the differences in dust levels, dust type
or monitoring units or position were not the sources
of variation. Therefore, probable sources, which is not
known or understood, may provide answers to differ-
ences in measured levels between three PDR units.
Parameters such as air velocity, monitor orientation,
particle size were not the sources of variation, as they
remained constant for all the tests.
It is often noted in studies that one of the major
sources of variations in measured dust levels by the
dust monitors could be the size distribution of the par-
ent dust (Soderholm, 1989, Volkwein, 2002; Ramani,
2004). However, in these tests, the size distribution of
the parent dust source, dust generation and airborne
mechanism has been consistent for all the tests. All
units were exposed to similar temporal and spatial
environmental conditions. Therefore any differences

185

Copyright © 2006 Taylor & Francis Group plc, London, UK

 in their responses were due to the sampling characteris-
tics of the dust monitors alone.The monitor differences
can be attributed to the differences in sensor detection
range of units, which is ‘in-built’ to the calibration fac-
tors of the PDR units as these units does not have any
manual calibration feature. Usually if the optics of the
sensor is contaminated the calibration of the monitor
gives ‘high background’ reading. Interestingly, dur-
ing the tests, none of the three PDR units gave any
‘high background’ or ‘calibration problem’ conditions
despite exposure to high dust levels. Also, time had no
significant influence on the sensors or lenses, or on
the correction factor of the real-time monitors as all
the three units had the equal exposure period.
A study by Thorpe and Walsh (2002) showed that
effects of three separate PDR orientation to the air-
flow (upright, on its back on its side) and its influence
on measured concentration showing the variation in
correction factor of 0.69 to 0.92. In this study, for the
same orientation of three PDR units had the correction
factors of 0.95, 1.10, and 1.13. Therefore, the PDR
orientation to the airflow may not be the source of
variation. Lastly, light scattering real-time dust moni-
tors depend solely on air movement to move the dust
particles into the sensing zone of the monitor. It is not
known, if the particle charges of airborne respirable
dust have any specific effect in terms of their move-
ment towards the sensing chamber that could have
contributed to the recorded differences in dust levels
by the three PDR units.
7 CONCLUSIONS
The following conclusions can be drawn from the labo-
ratory evaluation of real-time monitors evaluated using
average and peak recorded dust levels. From the sta-
tistical analysis of side-by-side comparison of three
PDR units and HD cyclone indicate that there is no
significant difference in measured dust levels. The
use of peak-recorded levels indicates that the differ-
ences in recorded dust levels between monitors exist.
The evidence from the study suggested that, while the
‘average’ dust level is a commonly used parameter
for evaluating monitor performance, the use of the
‘peak display’ parameter may lead to different conclu-
sion on the variations between measured dust levels.
The implication of this is that, in practice, the random
selection and use of a real-time monitor for engineer-
ing related dust control application may be in favor or
against the seriousness of the dust problem and could
impact the decision making process on the appropriate
allocation of financial and administrative resources. It
is proposed that for any new real-time monitors the
use of peak display level as a parameter of evaluation
may ascertain the probable sources of variations in
recorded levels. These additional data analyses steps
186
Copyright © 2006 Taylor & Francis Group plc, London, UK
may facilitate the adjudging of the sources of varia-
tions in measured dust levels in real-time monitors.
The intention of this paper is not to suggest that the
peak levels be should used in place of average levels
for exposure monitoring, rather an evaluation param-
eter in the understanding of variations experienced by
researchers worldwide. It is recognized that Tapered
Element Oscillating Microbalance instrument devel-
oped by Rupprecht and Patashnick (USA) may be a
step closer in minimizing errors in dust measurement
than the light scattering instruments.
ACKNOWLEDGEMENTS
The author wishes to express his sincere gratitude and
appreciation to Mr. ValliYousefi and Dr. Du Toit whose
help and support during the initial stages of the labo-
ratory work at NIOH; Tony Grupping of AMS Haden
(SA) for supplying the three PDR units.
REFERENCES
Baldwin, P.E.J., Maynard, A.D. & Northage, C. 1997. An
investigation of short-term gravimetric sampling in pig
farms and bakeries. Appl. Occup. Environ. Hyg., 12(10):
662–669.
Belle, B.K. 2002. Evaluation of Newly Developed Real-time
and Gravimetric Dust-Monitors for Personal Dust Sam-
pling For South African Mines, SIMRAC Report, SA,
pp 136.
Belle, B.K. 2004. International Harmonisation Sampling
Curve (ISO/CEN/ACGIH): Background and its influ-
ence on dust measurement and exposure assessment in
the South African mining industry, The Journal of the
Mine Ventilation Society of South Africa, Vol. 57, No. 2,
pp 55–58.
Biffi, M., Belle, B.K. & Unsted, D. 2000. Proposed ratio-
nal criteria for routine dust sampling of respirable dust in
South African mines. Project HEALTH 604b, SIMRAC
Research Report, South Africa, 110 p.
DME (SA Department of Minerals and Energy). 1997.
Measurement Guidelines. Pretoria, South Africa.
Gero A.J. & Tomb, T.F. 1988. Miniram Calibration Differ-
ences, Appl. Ind. Hyg., 3: 110–4.
Grupping, T. 2001. Personal Communications, AMS Haden,
South Africa.
Kissell, F.N., Volkwein, J.C., & Kohler, J. 2002. Historical
Perspective of Personal Dust Sampling in Coal Mines, 9th
North American Ventilation Symposium, pp 619–623.
Page, S., & Jankowski, R.A. 1984. Correlations Between
Measurements with RAM-1 and Gravimetric Samplers on
Longwall Shearer Faces, Am. Ind. Hyg., Assoc. J., 45(9):
610–616.
Tsai, C.J., Shih, T.S. & Lin, J.D. 1996. Laboratory Testing
of Three Direct Reading Dust Monitors, Am. Ind. Hyg.,
Assoc Journal, 57: 557–63.
Thorpe, A. & Walsh, P.T. 2002. Performance Testing of Three
Portable, Direct-Reading Dust Monitors, Ann. Occup.
Hyg., Vol. 46, No.2, pp 197–207.
 Tarkington, S.B., Rimmer, T.W., Keller, R.J. & Fowler, C.F.
1997. A comparison of two direct reading aerosol instru-
ments to each other and various gravimetric samplers in
several different aerosol environments. AIHA Abstract,
USA.
Ramani, R.V. 2004. Personal Communications, PSU, USA.
Soderholm, S.C. 1989. Proposed International Conventions
for Particle Size-Selective Sampling, Ann. Occup. Hyg.,
33(3): 301–320.
187
Copyright © 2006 Taylor & Francis Group plc, London, UK
Unsted, D. 1997. Dust sampling for engineering control
purposes. Project GEN 417, SIMRAC Research Report,
South Africa, 47 p.
Volkwein, J.C. 2002. Personal Communications, PRL,
NIOSH, USA.
Williams & Timko. 1984. Performance Evaluation of a
Real-time Aerosol Monitor, Bureau of Mines Information
Circular 8968, pp 20.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Lowering respirable dust at an iron ore concentrator plant through

improved ventilation practices

A.B. Cecala, J.A. Zimmer, J.F. Colinet & R.J. Timko

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

ABSTRACT: A cooperative research effort was established between the Tilden Mining Company LC, the
United Steelworkers of America, and the National Institute for Occupational Safety and Health to lower respirable
dust levels in an iron ore grinding and concentrator plant. This cooperative effort involved three field studies
to evaluate different techniques to lower respirable dust levels. The greatest impact on lowering respirable
dust concentrations throughout this facility was achieved through ventilation changes. The most significant
improvement was realized by changing the ventilation profile throughout the entire facility. The ventilation to
this grinding and concentrator facility was provided by approximately 40 roof fans operating in either the intake
or exhaust mode. The operation of these fans was significantly changed in an attempt to provide more of a
directional flow pattern throughout the facility, as well as to more closely balance the intake and exhaust air
volume. This change accounted for a 31 pct reduction in respirable dust levels in the primary grinding area of
the mill. Another successful modification was to improve the sealing of a reclaim tunnel to the outside and by
barricading the access point from the reclaim tunnel to the concentrator plant. This minimized the dust liberated
in the conveyor tunnels from traveling through the reclaim tunnel and into the plant. These two modifications
illustrate the impact that improvements in ventilation can have on lowering respirable dust concentrations in iron
ore processing facilities.

1 INTRODUCTION submitted a proposal for a cooperative research effort

In August 2000, the United Steelworkers of America
held a health and safety workshop in Eveleth, Min-
nesota to discuss methods and techniques to improve
working conditions for its membership in the iron ore
mines in Minnesota and Michigan. For this workshop,
the National Institute for Occupational Safety and
Health (NIOSH) was requested to provide a 90-minute
presentation on dust control research applicable to the
iron ore industry. Shortly after this workshop, NIOSH
was contacted by the Tilden Mining Company LC who
stated their interest in working with NIOSH and the
United Steelworkers of America union membership
at their operation to lower respirable dust concentra-
tions in the grinding and concentrator mill. Shortly
after receiving this request, a trip was scheduled (April
2001) for two individuals from NIOSH’s Respiratory
Hazards Control Branch to visit the Tilden opera-
tion in Ishpeming, Michigan. The first part of this
visit was a meeting with all three parties to discuss a
number of potential dust control areas of interest. After
this meeting, NIOSH was given a tour of the entire
facility with special emphasis given to the grinding
and concentrator mill. Shortly after this visit, NIOSH
to reduce respirable dust levels in the grinding and con-
centrator facility to the Tilden Mining Company and
the United Steelworkers of America. A short time later,
NIOSH received notification that both parties were in
agreement and wanted to proceed on this cooperative
research effort.
This cooperative effort has been composed of three
different studies at the Tilden grinding and concentra-
tor facility: March 11–13, 2002, March 5–7, 2003, and
March 9–11, 2004. The first study was mainly a base-
line analysis to determine dust levels throughout this
huge facility and determine research areas that could
have the most significant impact on lowering dust. The
next two visits were directed towards specific areas of
dust control technology.
All three studies were performed in March for a
number of reasons. First, the Tilden grinding and con-
centrator mill would switch from producing magnetite
to hematite throughout the course of the year. Hematite
is a dustier product and was typically produced from
November through March. All parties agreed that this
research should be performed during the production
of hematite in an effort to have the greatest impact
on lowering respirable dust concentrations. Another

189

Copyright © 2006 Taylor & Francis Group plc, London, UK

 factor in the timing of this research involved travel-
ing in the winter months. The Tilden operation is in
the upper peninsula of Michigan, where winters can
be, and usually are very severe. In March, winter con-
ditions usually begin to ease somewhat, facilitating
personnel going onto the roof of the structure to check
on fan operation. The last factor for performing this
research during the winter months is because the facil-
ity is typically closed to limit the inflow of cold air into
the structure.
The grinding and concentrator mill at the Tilden
operation was an 87,300 m2 (940,000 ft2 ) structure
with an internal air volume of 1,274,000 m3
(45,000,000 ft3 ).The ventilation provided to this build-
ing was through a network of roof fans that can be
used, for the most part, in either an intake or exhaust
ventilation mode. In addition to the numerous roof-
powered fans, each of two replacement air systems
supplied 4250 m3 /min (150,000 cfm) of heated intake
air in the basement of the structure, one on the east-
ern and the other on the western side of the structure.
With this setup, the building had an exhaust capacity of
approximately 121,776 m3 /min (4,300,000 cfm) and
an intake air capacity of 23,500 m3 /min (830,000 cfm)
based upon the rated capacity of the fans. Again,
8500 m3 /min (300,000 cfm) of this air was from the
intake air heaters. The roof-mounted fans are the only
source of powered exhaust ventilation to the facility,
although there may have been some natural ventilation
as the warm air within the structure rises and escapes
through cracks or openings in the exterior walls or the
roof. The amount of natural ventilation was minimal
in comparison to the amount of air exhausted through
the roof fans.
The ventilation setup at this facility changed some-
what over the course of the year based on outside air
temperatures. Typically, the air exhausted from the
structure was significantly less during the winter
months, with the goal being to provide enough ventila-
tion to remove contaminants while trying to maintain
an adequate temperature within the structure. Dur-
ing the summer months, the facility is more open to
the outside air and the exhaust air from the roof fans
is maximized to remove the heat generated from all
the electrical motors and other processes within the
facility.
One last area to consider for this research effort
is the location of various ore processes within the
structure. Because some of the processes are more
inherently dusty, certain areas of the grinding and
concentrator facility typically have higher dust levels.
Figure 1 shows an overview of the major processes
in this facility. The raw ore mined at the operation is
taken to a primary crusher where it is sized to less than
25.4 cm (10 inches). The ore is then conveyed to a cov-
ered storage building that is located along the entire
northern portion of the grinding and concentrator
190
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 1. Location of major processes in grinding and
concentrator building.
facility. When the ore is ready to be processed, it flows
by gravity into the facility onto one of 12 conveyors
lines which brings the ore up a slope that feeds a pri-
mary grinding mill. Each of these 12 primary grinding
mills then feed 2 pebble mills that further reduce the
size of the material. After the ore exits the pebble mills
and has been properly sized, it enters thickener tanks,
then flotation tanks, and then back into concentrator
thickener tanks. From this point, it enters concentrate
slurry tanks and is filtered. The ore exits the structure
and is delivered to the palletizing building.
The 12 conveyor lines and the primary grinding
mills are located in the northern portion of the struc-
ture, which encompasses approximately half of the
total area of the building. This is also the dustiest
area of the structure and Tilden management requires
that all personnel wear personal protective respiratory
equipment when in this portion of the facility. The 24
pebble mills encompass approximately a quarter of the
facility. The last quarter is made up of the concentrator,
slurry, thickener tanks, and filtering area. The southern
portion of the building is much less dusty and does not
require work personnel to wear respiratory protection.
2 SAMPLING INSTRUMENTATION
The following section describes some of the instru-
mentation used in this cooperative research effort.
Although many types of monitors and instruments
were used, this discussion will be limited to the dust
monitoring and air flow instruments used in the two
ventilation research efforts discussed in this report.
Gravimetric dust sampling packages and PDR instan-
taneous dust sampling devices will be discussed,
followed by the two types of instruments used for
measuring air velocity.
Gravimetric dust sampling packages consisted of
two sampling units, each containing a sampling pump,
flexible tygon tubing, a respirable dust classifier, and
a filter. All dust samples were collected with the
 10-mm Dorr-Oliver cyclone, which classifies the res-
pirable portion of dust, (usually considered having an
aerodynamic diameter of 10 microns or less). Each
gravimetric pump was calibrated to a flow rate of 1.7
liters/min, (the Mine Safety and Health Administra-
tion’s (MSHA) required flow rate for the metal/non-
metal industry). The respirable dust classified by the
10-mm cyclone was deposited on a 37-mm dust filter
cassette. These filters were pre- and post-weighed to
the nearest microgram on a microbalance at the Pitts-
burgh Research Laboratory (PRL) and the results of
each sampling unit’s filters were averaged together to
determine an average respirable dust mass. From this
mass value, the average respirable dust concentration
over the sampling period was mathematically derived.
The instantaneous monitors used for this testing
were the Personal Data RAM (PDR) from MIE, Inc.
of Bedford MA. This is a real-time aerosol sampler
that measures the respirable dust concentration based
upon the light scattering of particles drawn through
an internal sensing chamber. The respirable dust lev-
els obtained by light-scattering were recorded on an
internal data logger and then downloaded to a lap-
top computer at the end of each sampling shift. After
the dust traveled through the PDR, it was deposited
on a filter cassette, identical to that of a gravimet-
ric sampler. A new filter cassette was used for each
shift of testing and was analyzed with the other gravi-
metric samples upon returning to PRL. For each
sampling location, the PDR dust value was divided
into the average gravimetric concentration to deter-
mine a correction factor, which was then applied to all
the measurements taken with the corresponding PDR
for that shift. This allowed the instantaneous respirable
dust concentrations to be corrected to equivalent mea-
surements taken using the gravimetric method. All
PDR instrument values shown in this report have been
corrected in this manner. Using both of these types
of respirable dust monitoring equipment provided a
good profile of average dust concentrations through-
out the sampling period, as well as variations and
changes in respirable dust concentrations relative to
time throughout the workday.
In addition to the respirable dust measurements,
airflow measurements were taken using a Solent Velo-
meter from Gill Instruments Limited of Hampshire,
England to evaluate the airflow at different locations
throughout the grinding and concentrator mill build-
ing. This velometer is an ultrasonic anemometer that
operates by transmitting an ultrasonic pulse along the
wind path and measuring the time taken to reach
the receiver. In essence, the faster the wind speed, the
faster the pulse reaches the receiver.The velometer was
held stationary and a 10-sec sample period was started.
At the end of the 10 seconds, the instrument provided
average, maximum and minimum air velocities for that
specific time period.
191
Copyright © 2006 Taylor & Francis Group plc, London, UK
In addition, velocity measurements were also taken
with a vane-anemometer from Davis Instruments of
Baltimore, MD. The Davis anemometer was a 10 cm
(4-in) diameter, eight-blade instrument that can
measure air velocities from 0.15–25 m/s (30 to
5000 ft/min). It contains a balanced vane wheel that
rotates in response to airflow. Rotations are mechani-
cally transmitted to an indicator through a clock-like
movement.
3 TESTING AND RESULTS
In this cooperative research effort, a number of dif-
ferent areas were investigated in an effort to lower
respirable dust concentrations at the Tilden iron ore
grinding and concentrator facility. It should be noted
that, although the goal was to lower respirable dust
concentrations throughout the entire facility, the main
priority was the primary grinding area. In this report,
we will describe the two most successful efforts, both
of which were directly tied to ventilation changes.
The first area to be presented is the impact that low-
ering the amount of ventilation flowing through the
reclaim tunnel had on lowering respirable dust levels
in the western portion of the structure. The second
area is the efforts to improve the ventilating airflow
pattern throughout the entire grinding and concentra-
tor mill using the roof-mounted fans in a more direct
and balanced flow pattern and the impact that this had
on respirable dust levels, specifically in the primary
grinding mill area.
3.1 Improving ventilation and dust in reclaim tunnel
When performing our baseline testing during the first
evaluation in March 2002, it was determined that dust
levels increased from the eastern to the western portion
of the reclaim tunnel. The reclaim tunnel is a 2.1 m
(7 ft) wide by 2.3 m (7½ ft) high opening that runs the
entire length of the facility along the northern portion
of the structure. The function of the reclaim tunnel is
to provide access for workers and equipment to the 12
conveyor lines. Workers are not normally in the reclaim
tunnel unless they need to access one of the discharge
chutes or conveyors, which moves iron ore from the
storage area to one of the 12 primary grinding mills.
During a walk-through, it appeared that there was a
substantial amount of air moving through this tunnel
and there appeared to be no practical reason for this
amount of ventilation.
To focus on the reclaim tunnel, a test sequence was
performed on the daylight shift of March 6, 2002.
Eleven different dust sampling locations provided a
profile of respirable dust concentrations along the tun-
nel, as well as in some of the conveyor slopes. These
monitors were started over a one-and-a-half hour time

Figure 2. Respirable dust and airflow measurements during
baseline testing in 2002.
period and sampled for between five and six hours.
Six of these samples were located along the reclaim
tunnel, with a seventh sample location at the mid-
point of a lower north-south tunnel that connected the
reclaim tunnel to the main building basement. This
connecting tunnel was located between primary mills
10 and 11 and connected the reclaim tunnel to the pri-
mary grinding and concentrator mill about 9 m (30 ft)
below the main working elevation. The remaining four
sampling locations were at conveyor transfer locations
in tunnels 6, 7, 9, and 10. These 11 dust sampling
locations provided a good overview of respirable dust
concentrations for the tunnel. In addition, velocity
measurements were also taken along the entire length
of the reclaim tunnel, in the connecting tunnel to the
basement of the grinding and concentrator building,
and in each of the 12 conveyor slope tunnels.
The dust concentrations and ventilation measure-
ments taken are shown in Figure 2. The values listed in
this figure show a correlation between respirable dust
concentrations and airflow measurements. From a dust
control standpoint, respirable dust levels increased as
one moved from the eastern to the western side of
the reclaim tunnel. Once the airflow traveled a short
distance past tunnel 11, it would make a 90 degree
turn and flow down the connecting tunnel to the main
structure. As seen, the respirable dust concentration
in this connecting tunnel was lower than the eastern
dust values (tunnels 1–11) along the reclaim tunnel
because the dust was being diluted with an additional
144 m3 /min (5,100 cfm) of ventilating airflow that was
coming from conveyor tunnel 12 and was traveling
in the reclaim tunnel from west to east. Although no
dust measurements were taken, it is assumed that the
respirable dust concentration of this air was very low.
When this airflow was combined with the airflow trav-
eling from east to west, it diluted the respirable dust
concentration down to 0.65 mg/m3 . The velocity mea-
surements in the reclaim tunnel show the magnitude
192
Copyright © 2006 Taylor & Francis Group plc, London, UK
of airflow in this tunnel. The airflow direction in the
conveyor slopes varied with some air moving up the
slope towards the grinding and concentrator building
and others with air moving down towards the reclaim
tunnel. For those tunnels that were ventilating down-
wards, they were introducing respirable dust generated
during conveying and at transfer points into the reclaim
tunnel.As this air moved from east to west in the tunnel,
each additional conveyor slope would add additional
dust to the airstream.
After analyzing the data from this test, researchers
concluded that the 620 m3 /min (21,900 cfm) of air
being discharged into the basement of the structure
from the connecting tunnel needed to be reduced. This
volume of air had a respirable dust concentration of
0.65 mg/m3 , which was introducing a significant quan-
tity of dust into the grinding and concentrator mill
building. This volume of dusty air caused respirable
dust concentrations to be higher in the basement and
first floor of the structure in the western portion of the
building, as compared to the center or right portions
of the structure.
A factor in the amount of air moving through the
reclaim tunnel was that the seal at the eastern end of
the reclaim tunnel was badly damaged and was allow-
ing a significant quantity of air into the tunnel. Air
measurements at this location indicated that approxi-
mately 144 m3 /min (5,100 cfm) of air was coming into
the reclaim tunnel from this damaged seal.
The other area of concern was the amount of air
flowing from the reclaim tunnel into the grinding and
concentrator structure basement from the connecting
tunnel. Through discussions with all the cooperating
parties in this research effort, it was decided to replace
the damaged seal at the eastern end of the tunnel and
to install a seal where the connecting tunnel accessed
the main structure.
When returning for the 2003 study, NIOSH was
pleased with the quality of the seals installed at both
locations by Tilden personnel, shown in Figure 3a,b.
The first test performed during this study was to repeat
the evaluation of the reclaim tunnel area. Identical
respirable dust sampling and airflow measurement
locations were used as performed in the previous study.
The results of this test can be seen in Figure 4. As
shown, dust levels constantly increased when moving
from the east to west in the reclaim tunnel. For a
number of areas, one could visually see dust flow-
ing down the conveyor slope and into the reclaim
tunnel. Average respirable dust levels increased from
0.21 mg/m3 between tunnels 2 and 3 to 2.44 mg/m3
between tunnels 10 and 11. Respirable dust levels in
the lower connecting tunnel averaged a concentration
of 1.27 mg/m3 over the entire sampling period.
When comparing the results from pre to post-seal
conditions, there are a number of areas that standout.
First, during pre-seal testing, there was 144 m3 /min
 Figure 3a. Seal replaced at eastern end of reclaim tunnel to
access outside.
Figure 3b. Seal installed where connecting tunnel accesses
main structure in basement of facility.
(5100 cfm) of air flowing from the door at the east-
ern most portion of the reclaim tunnel; this was
reduced by 76 pct to a level of 34 m3 /min (1200 cfm)
under post- seal conditions. In addition, there was a
reversal in the airflow direction in a number of the
sloped conveyor tunnels. For the lower numbered tun-
nels, the air quantity moving up or down the tunnel
during post-seal was substantially lower than during
pre-seal conditions. Conveyor tunnel 9 showed a sig-
nificant increase in airflow quantity traveling up the
tunnel with 215 m3 /min (7600 cfm) for post-seal, as
compared to 144 m3 /min (5100 cfm) for pre-seal. The
effectiveness of the barricade on the lower level con-
necting tunnel is also verified by these values. The
airflow measured in the connecting tunnel during post-
seal was 22.7 m3 /min (800 cfm), which is a 96 pct
reduction from the 620 m3 /min (21,900 cfm) measured
during pre-seal conditions.
When considering the dust values, it may appear that
improving the seal in the reclaim tunnel was counter
productive because respirable dust concentrations
193
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 4. Respirable dust and airflow measurements during
2003 testing after installation of seals.
increased at some locations. This increase was first
noticed at the midpoint for mills 9 and 10 and remained
elevated at the next two sample locations. The dust lib-
erated in the various conveyor slope areas and drawn
into the reclaim tunnel was diluted by a much lower
quality of air, and thus, the reason for the higher dust
concentration at the western portions. To consider the
true impact of the seals on lowering respirable dust
flowing into the grinding and concentrator building
from the reclaim tunnel, both the airflow and the res-
pirable dust concentrations must be considered. For
pre-seal test conditions in 2002, there was 620 m3 /min
(21,900 cfm) of air traveling into the main structure at a
respirable dust concentration of 0.65 mg/m3 . Through
a simple calculation, it was determined that 24,188 mg
of respirable dust was being delivered into the main
building every hour under pre-seal conditions. And
although the respirable dust levels are higher under
post-seal conditions at a level of 1.27 mg/m3 , when
one considers this value with the reduction in airflow,
only 1729 mg of respirable dust is being delivered into
the main building every hour, a 93 pct reduction over
pre-seal conditions.
The most significant benefit from the installation of
seals and minimizing the airflow through the reclaim
tunnel is the significant reduction in the amount of
respirable dust being introduced into the western por-
tion of the building. In the pre-seal testing, there was a
67 pct increase in respirable dust concentrations at pri-
mary mill 12 (0.67 mg/m3 ) on the western side of the
building as compared to respirable dust levels at mill 1
(0.44 mg/m3 ) on the eastern side. When the exact study
was performed again with post-seal conditions, both
dust measurements were almost identical: 0.84 mg/m3
at mill 12 and 0.87 mg/m3 at mill 1. Although the res-
pirable dust concentrations at both locations went up,
this is a relative value based on the amount of produc-
tion and ore being processed at any given time period.
Respirable dust levels being identical on the eastern
 Figure 5. Baseline or normal ventilation setup.
and western side of the structure indicates the impact
that the seals had on minimizing the amount of airflow
traveling through the reclaim tunnel.
3.2 Entire structure ventilation
During the testing performed during 2003, an entire-
structure ventilation change was performed which
indicated promising results. This led to a more in-
depth study performed during 2004 which involved
changing the quantity of air that entered and was
exhausted from the structure. During the first day
of testing, the operation’s normal ventilation setup
was evaluated. Figure 5 shows a summary of this
ventilation setup with the exhaust and intake fans dif-
ferentiated by shading of the fan symbol. To evaluate
this original ventilation design, the structure was bro-
ken down into three distinct areas: primary grinding
(northern-most portion), secondary grinding with peb-
ble mills, and filtering and flotation (southern-most
portion). Respirable dust samplers were located in
these three areas, with the greatest focus being on the
primary grinding mills. This original ventilation setup
included twelve exhaust fans in the primary grinding
area, eight exhaust fans and three intake fans in the
secondary grinding area, and nineteen exhaust fans in
the filter and flotation area. The only other mechan-
ical ventilation provided was from two intake heater
fans located at the base of the facility, one located on
the east and the other on the west side of the building.
The volume of air being exhausted from the structure
was 54,670 m3 /min (1,930,500 cfm), as compared to
17,000 m3 /min (600,000 cfm) being brought into the
structure. These ventilation air volumes were calcu-
lated based upon the fan quantities provided by the
facility and the assumption that all the fans were oper-
ating at their rated output. The two intake heater fans
accounted for 8,500 m3 /min (300,000 cfm), or 50 pct
of the total intake airflow. Dust sampling under this
ventilation setup was performed for approximately 20
hours of testing.
The imbalance between the intake and exhaust air
volumes with this ventilation setup was substantial;
therefore it was decided to modify the ventilation to
a more balanced design. It should to be noted that
194
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 6. Ventilation setup #2.
although the goal was to lower respirable dust concen-
trations throughout the entire facility, the main priority
was to lower respirable dust concentrations in the pri-
mary grinding area. Dust levels in the primary grinding
area were higher than in the other areas because of the
amount of dust liberated from the numerous conveyor
lines feeding the primary grinding mills, as well as the
amount of dust liberated from the mills themselves.
To focus on this area, a ventilation change was ini-
tiated as shown in Figure 6. The two intake heater fans
continued to bring a significant portion of intake air
into the facility. For this modified design, there were
eleven exhaust fans operating in the primary grind-
ing area, twelve intake fans in the secondary grinding
area, and eight exhaust fans operating in the filtering
and flotation area. The volume of air being exhausted
from the structure was approximately 26,600 m3 /min
(940,500 cfm), with approximately 29,600 m3 /min
(1,045,000 cfm) brought into the structure through the
roof intake fans and two heater fans. This ventilation
design was again tested for approximately 20 hours.
During this phase of testing, a smoke flare was
released into one of the intake fans over the pebble
mill. Since properly operating intake fans should move
an air parcel 10 fan diameters before being reduced to
10 pct of the fan’s air velocity, the roof-mounted fans
should introduce a parcel of air into the lower parts
of the building (Industrial Ventilation, 2001). These
roof-mounted fans are all 1.2 m (4 ft) in diameter and
should introduce a parcel of air at least 12.2 to 15.2 m
(40 to 50 ft) into the structure. Since this testing was
performed during the winter months, introducing cold
outside air into the structure should also assist the air
parcel’s ability to flow down into the lower levels of the
structure. From visual observations during this smoke
flare release, it appeared that the intake fan chosen
for this testing was not operating at the calculated air-
flow. The smoke remained within the top one-quarter
of the structure and tended to drift towards the south-
ern portion of the building (toward the flotation and
filtration area). As test personnel walked through the
secondary grinding area, it was very difficult to feel
the intake air being delivered into the structure by the
intake fans. Since the walkway was only 30 ft below
the intake fans, airflow should have been perceivable
at this level.
 of the three different ventilation designs tested. Obvi-
ously, the last ventilation design was the most effective
at lowering respirable dust levels, particularly in the
primary grinding area, which was the main area of con-
cern. When compared to the average of the first two
ventilation designs, the last design lowered respirable
dust concentrations in the primary grinding area by
31 and 25 pct, respectively. This is a very significant
reduction when one considers it was achieved by using
the existing fans and simply varying the intake and
Figure 7. Ventilation setup #3. exhaust flow quantities. Additionally, respirable dust
levels were also slightly lower in the pebble mill and the
Table 1. Average respirable dust concentrations in flotation and filtration areas with the last ventilation
mg/m3 for three ventilation setups. setup. This study indicates the impact that ventilation
can have on lowering respirable dust levels throughout
Setup #1 Setup #2 Setup #3 a large-volume iron ore processing structure.
Location Day Night Day Night Day

Primary # 2 1.06 1.21 0.66 0.38 0.44 4 CONCLUSIONS

Primary # 4 0.72 0.70 0.59 0.59 0.45
Primary # 6 1.15 1.34 1.37 1.49 0.98 A cooperative research effort was established between
Primary # 8 1.23 1.34 1.18 1.23 0.86
Primary # 10 – 0.97 1.05 0.91 0.72
the Tilden Mining Company LC, the United Steel-
Primary # 12 0.67 0.72 0.79 0.88 0.75 workers of America, and the National Institute for
Average 0.97 1.05 0.94 0.91 0.70 Occupational Safety and Health in an effort to lower
Pebble # 1 0.32 0.38 0.38 0.31 0.29
respirable dust levels in Tilden’s grinding and concen-
Pebble # 3 0.42 0.42 0.47 0.51 0.42 trator facility located in Ishpeming, Michigan. The two
Pebble # 6 0.34 0.35 0.48 0.50 0.35 most significant impacts in this cooperative effort dealt
Pebble # 9 0.49 0.50 0.51 0.56 0.49 with ventilation changes. The first change was to min-
Pebble # 12 0.38 0.41 0.36 0.43 0.38 imize the amount of air traveling through a reclaim
Average 0.39 0.41 0.44 0.46 0.38 tunnel before flowing into the basement of the struc-
Flotation # 3 0.25 0.32 0.35 0.34 0.24 ture. As the air moved through this reclaim tunnel, it
Flotation # 9 0.27 0.34 0.45 0.40 0.31 entrained a substantial quantity of respirable dust gen-
Average 0.26 0.33 0.40 0.37 0.27 erated from the twelve conveyor lines. As this dust
laden air then entered the main structure through a
connecting tunnel, it caused respirable dust levels to
be 61 pct higher on the western side of the building
Because the intake air appeared to be drifting as compared to the eastern side. In an effort to elim-
towards the flotation and filtration area, it was decided inate this occurrence, a seal on the very eastern side
to modify the ventilation setup to provide a greater of the structure, where the reclaim tunnel accessed the
quantity of the intake air moving in the northern direc- outside environment, was rebuilt. In addition, a new
tion towards the primary grinding mills. Figure 7 indi- seal was installed where the reclaim tunnel accessed
cates the second modification to the ventilation design the main structure. Both of these seals substantially
which was evaluated for approximately 8 hours during lowered the amount of air flowing through the reclaim
the last day of testing. This ventilation setup was com- tunnel and entering the main structure. Although this
posed of 24 exhaust fans and 12 intake fans, along with lower air flow caused dust levels in the reclaim tunnel
the two heater intake fans. The volume of air exhausted to be higher at some locations, this was not a signif-
from the structure was approximately 32,000 m3 /min icant factor since personnel are not normally in this
(1,128,000 cfm), with 29,600 m3 /min (1,045,000 cfm) tunnel. The benefit of this change was that respirable
being brought into the structure through the roof intake dust levels were lowered in the western side of the main
fans and two heater fans. The amount of intake air grinding area to a point that they were the same as on
was identical to the previous test, but the exhaust was the eastern side of the structure.
increased in an attempt to decrease the flow toward The second and more significant improvement
the flotation and filtration area while increasing the identified in this cooperative research effort was acco-
airflow to the primary grinding mill area. mplished by modifying the way the entire structure was
Table 1 shows the average respirable dust concentra- ventilated. The effort involved monitoring respirable
tion as determined by the instantaneous respirable dust dust levels throughout the entire structure for three
monitors used in this study and provides a comparison ventilation schemes. The first setup tested was as the

195

Copyright © 2006 Taylor & Francis Group plc, London, UK

 structure was normally ventilated (by Tilden). For this
setup, the exhausted air was over three times greater
than the intake air, which caused the structure to be
under negative pressure. For the second setup, the ven-
tilation was changed to a more balanced design and
improved airflow pattern within the structure. Based
upon the configuration of the fans, it was decided to
bring the intake air into the building at the center of the
structure and exhaust it northward toward the primary
grinding mills and southward toward the flotation and
filtering areas. A smoke flare released into one of the
intake fans indicated that a substantial portion of the air
was drifting south towards the flotation and filtering
areas, than was flowing north. Moreover, although the
intent was to lower respirable dust levels throughout
the entire structure, the main priority was to lower lev-
els in the primary grinding area, where workers were
required to wear respirators.
A third ventilation setup was tested which increased
the number of exhaust fans in the northern portion
and reduced them in the southern portion of the struc-
ture. This last ventilation design was the most effective
of the three tested and provided a 31 pct reduction
in respirable dust levels in the primary grinding area
when compared to average levels measured for the
first setup. Average respirable levels were also slightly
196
Copyright © 2006 Taylor & Francis Group plc, London, UK
lower in the pebble mill and flotation and filtration
areas. This is a substantial reduction when consider-
ing that it came about by creating a more balanced flow
pattern, but still using the same amount of fans.
In closing, both of these research efforts indicate the
significant and cost-effective impacts that ventilation
changes can have on lowering respirable dust levels in
a very large iron ore facility.
ACKNOWLEDGEMENT
The authors would like to acknowledge Leonard
R. Parker, Manager-Safety & Environmental, Dave
Wickstrom, (Previous) Union Safety Chair, Bill Vivian
(Union Safety), Greg Magnuson (Ventilation), Dan
Sleeman (Ventilation), and Scott Rasmussen, Area
Manager-Concentrator Operations, for all their help
and assistance in this research effort.
REFERENCE
Industrial Ventilation – A Manual of Recommended Prac-
tices, 2001 – 24th edition (American Conference of
Governmental Industrial Hygienist: Cincinnati).
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

A new method to clean dust from soiled work clothes

D.E. Pollock, A.B. Cecala & J.A. Zimmer

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

A.D. O’Brien
Unimin Corporation, Winchester, VA, USA

J.L. Howell
Unimin Corporation, Marston, NC, USA

ABSTRACT: Dust from worker’s clothes has been shown to be a problematic source of personal dust exposure.
A recently completed effort between NIOSH and the Unimin Corporation has resulted in a quick and effective
way to clean worker clothing. The process involves a booth under negative pressure with an air spray manifold
to supply compressed air to blow off the dust from the clothing. The overall system is designed to meet MSHA
and OSHA requirements. Results of field testing indicated that the manifold cleaned the clothes 10 times faster
and removed 50% more dust than cleaning methods used by workers today.

1 INTRODUCTION contained HEPA filter, or ducted outside. The booth

New methods and techniques to lower respirable dust
exposures to workers in the mining industry are con-
stantly being investigated by health and safety special-
ists. One area of known worker exposure throughout
all industries is from contaminated work clothing. For
the mining industry, a U.S. Bureau of Mines report
documented up to a 1 mg/m3 increase in worker’s dust
exposures on a number of separate occasions from
dusty work clothes (Cecala & Thimons, 1986). These
cases indicated that respirable dust levels were ele-
vated to the extent that workers could be over their
exposure limit in less than two hours. As the individu-
als performed their work duties, dust was continuously
emitted from their clothing. The most effective way
to eliminate this dust source was to clean or change
their work clothing. In the past, the only MSHA-
approved method to perform clothes cleaning was to
use a HEPA-filter vacuuming system which is very
difficult and time-consuming task to perform. Work-
ers will sometimes use an air hose to blow off the
dust on their clothing. This method is prohibited by
both MSHA and OSHA and may cause exposure to
co-workers by liberating dust to the surrounding area.
A cooperative research effort between NIOSH and
the Unimin Corporation has resulted in an improved
method to clean the workers clothing. The new method
utilizes an enclosed booth which can either be ducted
to a baghouse dust collector, cleaned by a self
is under constant negative pressure therefore, no dust
liberated by the cleaning process contaminates the
surrounding area. A compressed air spray manifold
system was developed to remove the product from the
workers clothing. The system consists of a ball-valve
actuated steel pipe manifold with flat-fan air sprays
supplied by a 0.45 or 0.90 m3 (120 or 240-gallon)
air reservoir tank. The pressure to the system is regu-
lated to 206.8 kPa (30 psi) to comply with MSHA and
OSHA regulations. The new method simply requires
the worker to don the required PPE, enter the booth,
actuate the air valve, slowly spin in front of the air
spray manifold (taking roughly 17 seconds), and exit
the booth with clean clothing.
On-site testing by NIOSH researchers has shown
that the new method is 10 times faster and removed
50% more dust than the currently used clothes cleaning
methods. It is both simple and cost effective and has
applications in the mining industry as well as other
industries where particulate contamination of clothing
is an issue. This new technique was recently approved
by MSHA under a petition for modification.
2 CURRENT REGULATIONS AND
PROCEDURES
There are two federal regulations that affect the clean-
ing of clothes during the work day for the United

197

Copyright © 2006 Taylor & Francis Group plc, London, UK

 States mining industry. The first is a mining reg-
ulation established by the Mine Safety and Health
Administration (MSHA) in 30 CFR Part 56.13020,
which states: “At no time shall compressed air be
directed toward a person. When compressed air is
used, all necessary precautions shall be taken to pro-
tect persons from injury.” A second regulation is a
general industry standard established by the Occu-
pational Safety and Health Administration (OSHA)
in 20 CFR 1910.242(b), stating that: “Compressed
air shall not be used for cleaning purposes except
where reduced to less than 206.8 kPa (30 psi) and then
only with effective chip guard and personal protective
equipment.”
Currently, the only approved method by MSHA to
perform clothes cleaning is to use a HEPA-filter vac-
uuming system. To perform this technique, a worker
uses the vacuum hose and manually moves the noz-
zle over his/her soiled clothing in an attempt to
remove the contamination. This is a very difficult
and time-consuming task to perform. Because of this,
some workers actually would prefer to use a single
compressed air hose to blow dust from their work
clothing, even though this is not an approved method
of cleaning. Using this technique can have numer-
ous drawbacks to include: the use of higher than
OSHA approved pressures; and being performed in
the open work areas which not only contaminates the
worker, but co-workers as well. While investigating a
new approach to perform this clothes cleaning process,
it was critical to be able to meet the federal regulations
and standards and come up with a process that workers
would want to use.
3 A NEW CLOTHES CLEANING PROCESS
The initial step of the clothes cleaning process design
was to develop a safe area to clean clothing. UNIMIN
Corporation purchased an enclosed booth and installed
it at their Marston facility, which provided the worker
sufficient space to effectively perform the cleaning
operation. Above the door was an open grate that pro-
vided an intake for the ventilation airflow. A return air
plenum located on the bottom-back wall of the booth
was ducted to the mill building baghouse dust collector
system, which provided a constant flow of air through
the enclosure. See figure 1.
The exhaust flow rate was measured at 2.17 m3 /sec
(4,600 cfm).The booth had a negative differential pres-
sure of 37.3 Pa (−0.15 inches w.g.). Since the booth
is under constant negative pressure, it proved to be an
effective area for clothes cleaning because it did not
allow any dust leakage into the plant.
The next critical aspect was to develop an effective
method to remove the product from the clothing. To
do this, air nozzles were installed in a spray manifold
Figure 1. Cleaning booth showing airflow.
that used compressed air to remove the dust from
workers’ clothes. Considerable design effort went into
determining the most effective spray nozzle mani-
fold configuration and numerous laboratory tests were
conducted at the NIOSH Pittsburgh Research Labora-
tory. Researchers also evaluated the impact of vary-
ing cleaning distances, clothing type, nozzle types,
nozzle spacing, air pressure, and spraying duration to
optimize the cleaning effect.
The air spray manifold was fabricated from 63.5 mm
(1– 1/2 inch) schedule 40 steel pipe that was capped at
the base. The air spray manifold was actuated by the
worker performing the cleaning process by operating
a timer-set pneumatic valve located on the top of the
manifold. The pneumatic valve had a safety interlock
option which would automatically shut the air sup-
ply to the manifold if the exhaust ventilation system
failed to keep the booth under sufficient negative pres-
sure. Twenty-six (26) flat fan air nozzles were mounted
along the manifold, spaced on 50.8 mm (2-inch) cen-
ters. With this spacing, the flat fan nozzles seemed to
provide the most uniform cleaning. The bottom nozzle
was a circular design located 152.4 mm (6 inches) from
the floor. This nozzle was used in coordination with
a ball-type adjustable fitting that was directed down-
wards to clean the individual’s work shoes or boots.
See figure 2.
At a pressure of 206.8 kPa (30 psi), the air spray
manifold system expels 4.7 m3 (166 cubic feet) of air
for the typical cleaning period. In order to supply this
compressed air volume to the air nozzles for effec-
tive cleaning, a 0.45 m3 (120-gallon) air reservoir tank
was necessary. This tank was installed at the opera-
tion and was typically pressurized to the 1,034.2 kPa

198

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 2. Air nozzle manifold design.
Figure 3. Booth, manifold and reservoir.
(150 psi) level. The air reservoir was located directly
behind the cleaning booth and hard-piped to the air
spray manifold located inside the booth. Supply air to
the manifold was regulated down to 206.8 kPa (30 psi).
The air regulator was located in a lock-box enclosure
to prohibit anyone from tampering with the air pres-
sure. Figure 3 shows the cleaning booth, air reservoir,
and air manifold configuration.
The worker performing the cleaning process is
required to wear a half-mask fit-tested respirator
with an N100 filter, hearing protection, and full-seal
goggles.
199
Copyright © 2006 Taylor & Francis Group plc, London, UK
4 EVALUATION OF THE THREE CLEANING
METHODS
Field evaluation consisted of randomly testing the
three different cleaning methods. The HEPA vacuum-
ing system, the single air nozzle regulated to 206.8 kPa
(30 psi), and the air nozzle manifold system which
was also regulated to 206.8 kPa (30 psi). All of these
methods were performed in the booth.
Two gravimetric dust sampling racks were con-
structed to sample inside and outside of the booth.
Each rack consisted of two pumps (calibrated to 1.7
liters/min), two 10-mm Dorr-Oliver cyclones and two
37-mm pre-weighed dust filter cassettes. One rack was
hung on the inside of the booth adjacent to the spray
manifold. The other rack was hung outside of the booth
near the door. The weight gains on the two filters at
each location were averaged to provide an average
respirable dust mass for each location.
The instantaneous monitor used during this testing
was the Personal Data RAM (pDR) by Thermo Elec-
tron Corporation which was set to active sampling
mode. One pDR sampler was hung on the outside
booth rack and one on the inside booth rack to enable a
real-time dust concentration track inside the booth and
monitor for outside contamination during the testing.
In order to test for possible contamination of the
worker by leakage around the 1/2-mask respirator,
researchers utilized a barbed fitting which is com-
monly used during fit testing of respirators. This fitting
was installed between the 1/2-mask respirator and one
of the filter cartridges. A piece of flexible tubing con-
nected to 1/2-mask respirator to an air-tight box which
housed a 10-mm Dorr-Oliver cyclone. This cyclone
was connected to another pDR in active mode.This set-
up enabled real-time monitoring for contamination in
the respirator during testing of the cleaning methods.
A matrix of tests was performed at UNIMIN’s
Marston plant to evaluate the effectiveness of this
newly developed technique. For this field testing,
the new clothes cleaning technique was compared to
the vacuuming system and the single handheld com-
pressed air nozzle. In addition, two different coverall
types were tested, with one being 100 pct cotton and
the other a cotton-polyester blend. Prior to each test,
the coveralls were soiled with inert limestone dust to
a degree that represented an extreme case of soiling.
The weighing procedure consisted of pre-weighing
the clean coveralls and placing them in a pre-weighed
bag. Once the coveralls were soiled, they were placed
in the bag and weighed again. The researcher then
removed the coveralls from the bag (which was post-
weighed) and put on the coveralls while standing on a
pre-weighed piece of brattice cloth. The brattice piece
was then weighed to account for any dust lost while
donning the coveralls. After the test method was per-
formed, the coveralls were removed while standing on
 Table 1. Amount of dust remaining on coveralls after Multiple test dust concentration
cleaning time for cotton and polyester/cotton blend coveralls. Outside booth Inside booth Inside respirator

Dust concentration, mg/m3

100.0
Poly/Cot Poly/ 90.0
80.0
Cotton Cotton Blend Cot Blend 70.0
60.0
50.0
Dust on Clean Dust on Clean 40.0
30.0
Cleaning Coveralls Time Coveralls Time 20.0
Method (grams) (sec.) (grams) (sec.) 10.0
0.0
08:54:50 08:57:45 09:00:40 09:03:35
Vacuuming 63.1 398 45.5 346 Time, hr:min:sec
Air Hose 68.8 183 48.4 173
Manifold 42.3 17 21.9 18 Figure 5. pDR results showing dust concentrations during
tests.

Single test dust concentration

Outside booth Inside booth Inside respirator

Dust concentration, mg/m3

100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
08:54:50 08:55:05 08:55:20 08:55:35 08:55:50
Time, hr:min:sec

Figure 6. pDR results showing a single test.

shows the relative effectiveness of the cleaning tech-

Figure 4. Effectiveness of cleaning methods.
a pre-weighed piece of brattice and placed in a pre-
weighed bag. The coveralls and bag were weighed
together and the brattice was weighed to account for
any dust lost while removing the coveralls. This exten-
sive weighing regimen was developed to account for
all dust lost during each test.
Each test was timed by a stopwatch to determine
the actual cleaning time. Results of testing indicated
that the manifold cleaned the clothes 10 times faster
and removed 50% more dust than the single air nozzle
or vacuuming methods. Table 1 provides the average
cleaning times and the remaining dust weights on the
coveralls from the three different techniques evaluated.
These values represent averages calculated for two
NIOSH test personnel and a total of 96 tests. Figure 4
niques tested.
Respirable dust samples taken inside the respirator
of the test personnel performing the clothes cleaning
process showed minimal to no respirable dust expo-
sure. In more than half of the 48 tests performed
with the air spray manifold, the test subject’s respirable
dust concentration remained at 0.00 mg/m3 inside the
half-mask respirator. See figure 5.
In the remainder of the tests, the value remained
very low with an overall average respirable dust con-
centration of 0.02 mg/m3 for the entire test group.
Figure 6 shows the effectiveness of the booth to remove
the dust which is liberated from the spray manifold.
Note that there was no contamination to the out-
side environment during the testing. Figure 6 also
shows the short time-frame to bring the inside booth
concentration back to zero.
Another factor evaluated during this study was the
cleaning effectiveness of the process on two different
coverall fabrics. As Table 1 shows, there was a signif-
icant improvement with the cleaning effectiveness of
the polyester/cotton blend coveralls when compared
to the pure cotton type. This needs to be considered
by operations implementing this new clothes cleaning
process.
With the air spray manifold design, the flat fan noz-
zles extend 79.4 mm (3-1/8 inches) from the supply
pipe and could easily be broken off if struck forcefully.
Because of this, it is recommended that side barriers

200

Copyright © 2006 Taylor & Francis Group plc, London, UK

 be installed to protect the air nozzles. During field test-
ing, 25.4 mm (1-inch) wood sheeting was used along
both sides of the nozzles, providing an effective barrier
to minimize the potential for nozzle damage.
The air spray manifold was designed for a person
177.8 cm (5 –10 ) in height, which was chosen based
upon the 50-percentile height for a male worker. Taller
workers will have to stoop and drop their shoulders to
effectively clean their upper body. When a person is
shorter, the top air nozzles can be covered with deflec-
tors to prevent the air sprays from directly hitting the
individual’s face. During the final field test, the top
four nozzles were modified with deflectors attached
to the side barriers fabricated from 101.6 mm (4-inch)
PVC pipe that was cut in half and then into 50.8 mm
(2-inch) wide strips. Latches were attached on both
sides of these deflectors so they could be locked in
either the open or closed position.
A primary concern regarding any type of new tech-
nology is the cost of implementation. The total cost
of the clothes cleaning unit should be in the $3,000
to $4,000 range, excluding the cost for the exhaust
volume of air and ductwork cost. The clothes clean-
ing process utilizes compressed air as the cleaning
medium. The compressed air utility available at the
operation must be analyzed to ensure that critical pro-
cesses are not starved due to the operation of the
clothes cleaning process. A dedicated compressor to
supply the necessary air may be an option in this case.
At the UNIMIN Marston Operation test site, an
excess exhaust volume was available in the baghouse
and thus was used for this system. Most operations will
not have this luxury and this will have to be built into
the cost. UNIMIN and NIOSH are in the process of
testing a cleaning system for operations without avail-
able baghouse capacity which will utilize an exhaust
fan to blow the dust-laden air up a stack to the outside
of the facility. Since the amount of dust removed from
201
Copyright © 2006 Taylor & Francis Group plc, London, UK
a worker’s clothing will be relatively minor in rela-
tion to the amount of air necessary to place the booth
under negative pressure, the respirable dust concentra-
tion of air coming out of this stack most likely would be
insignificant. Testing of this system will be performed
at the UNIMIN Elco Operation located in Elco, IL, in
the near future.
5 CONCLUSIONS
The new clothes cleaning process proved to be very
efficient since the worker only needed to don the
required PPE, enter the booth, actuate the automatic
valve, slowly spin in front of the air spray manifold
(taking roughly 17 seconds), and exit the booth with
clean clothing. This process has been demonstrated
to be a much more effective method to remove dust
from a worker’s clothing than methods currently used
by workers. Although this process was designed for
workers in the mining industry, it is applicable to
any industry where contaminated work clothes are a
problem.
It must be noted that this newly designed clothes
cleaning technique is not currently blanket approved
by MSHA for U.S. mining operations. A Petition of
Modification has been granted to UNIMIN Corpora-
tion by MSHA for use of the clothes cleaning process
at the Marston plant. Operations wanting to use this
technique may receive MSHA approval on a case by
case basis.
REFERENCE
Cecala, A.B. & Thimons, E.D. 1986. Impact of Background
Sources on Dust Exposure of Bag Machine Operator. US
Bureau of Mines Information Circular. IC 9089.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Impact of fan type for reducing respirable dust at an underground limestone

crushing facility

G.J. Chekan, J.F. Colinet & R.H. Grau III

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

ABSTRACT: Researchers from the National Institute for Occupational Safety and Health have demonstrated
that mine-wide ventilation in large-opening stone mines can be improved by using low-pressure, high-volume
propeller fans and constructing stoppings in key locations to direct and control airflow. In light of these findings,
a comparative study was conducted to determine if a portable diesel-powered propeller fan could perform more
efficiently for dust dilution and transport than an axial vane fan for localized ventilation. The objective of
this study is to evaluate both fan types for ventilation efficiency at an underground dump/crusher facility in a
limestone mine. Results showed an improvement with the propeller fan to dilute both the respirable dust and the
respirable silica dust around the dump/crusher facility. Overall, an average reduction of 20% in respirable dust
and silica was observed at the dump/crusher location with the propeller fan.

1 INTRODUCTION include mobile workers, crusher operators, and front

The National Institute for Occupational Safety and
Health (NIOSH) at the Pittsburgh Research Labora-
tory is currently involved in various research projects
related to worker health and safety in underground
metal/nonmetal mines. A primary area of research
involves health issues in underground stone mines, a
growing segment of the aggregates industry. Methods
for reducing worker exposure to noise, silica dust, and
diesel particulate matter are being addressed through
various research programs. One approach to reducing
worker exposure to silica dust in large-opening mines
is to adopt ventilation practices that will increase air-
flow and dilute and transport harmful particulate from
major dust sources.
Baseline sampling surveys have shown that under-
ground dumps and crushers in limestone mines can
be a significant source of dust generation (Chekan
et al, 2002). Inhalation of excessive levels of respirable
silica dust can lead to the development of silico-
sis in mine workers. Although the health hazards
from silica dust have been well documented for many
years, the problem of overexposure still persists for
a number of job occupations in both underground
stone and metal/nonmetal mining operations. Com-
pliance sampling from the Mine Safety and Health
Administration (MSHA) database, for a 5-year period
from 2000 through 2004 shows that, on average, 14%
of all occupational samples exceed the permissible
exposure limit. Occupations that exceed the average
end loader operators (MSHA, 2000–2004).
Typically, underground stone operations are drift
mines developed after the economical quarry reserves
have been exhausted. Room-and-pillar mining meth-
ods are utilized, with pillars of either square or rectan-
gular dimensions ranging from 10.6 to 18.3 m (35 to
60 ft). The entries are considered large mine open-
ings with entries widths ranging from 9.1 to 18.3 m
(30 to 60 ft) and entry heights on development rang-
ing from 4.9 to 13.8 m (16 to 45 ft). After benching,
entries can be over 18.3 m (60 ft) high. Due to these
large openings, ventilation fan pressure is very low,
even if significant quantities of air move through the
mine. Ventilation pressures of less than 24.9 Pa (0.1 in
w.g.) are common whether airflow is induced by fans,
natural ventilation pressure, or a combination of both
(Grau et al, 2002a). Depending on the extent of the
workings, air velocities less than 0.13 m/s (25 fpm)
are common, or in some idle areas of the mine, virtu-
ally nonexistent. In addition, airflow in the entries can
be stratified, multi-directional, or readily affected by
the movement of mine equipment.
These large-opening mines require ventilation sys-
tems that are designed to effectively dilute airborne
particulate by supplying sufficient ventilating air to
primary locations, i.e., working faces, worker loca-
tions, etc. To accomplish this, studies show (Grau et al,
2002a; Grau et al, 2002b; Head, 2001; Kissell and
Volkwein, 2002; Timko and Thimons, 1987; Chekan
et al, 2004a) that three key parameters should be

203

Copyright © 2006 Taylor & Francis Group plc, London, UK

 included in the mine ventilation plan to improve air-
flow: (1) a main mine fan to establish air circuits
on a mine-wide scale; (2) either permanent or brat-
tice stoppings in key locations throughout the mine
to more efficiently direct and control the airflow;
and (3) the application of booster fans, to improve
local ventilation in areas requiring a more direct and
controlled volume of airflow to remove harmful par-
ticulate. These fans include axial vane fans, jet fans,
and more recently low-pressure, high-volume pro-
peller fans. Studies comparing fan performance have
shown that propeller fans are more efficient at lower
pressures and can produce larger air quantities at
lower horsepower than axial-vane fans (Grau et al,
2002a; Grau et al, 2002b; Krog and Grau, 2006). In
addition, because of operational characteristics related
to lower fan pressure and larger diameter, propeller
fans provide better regional air coverage in large-
opening mines than axial vane fans. In light of these
findings, a study was conducted at an underground
dump/crushing site to compare the effectiveness of
an electric-powered axial vane fan versus a diesel-
powered propeller fan for the dilution and transport
of respirable dust from this facility.

2 FAN CHARACTERISTICS AND SAMPLING

STRATEGY

The two fans compared in the study are shown in

figure 1. Air volume for each fan was measure
using a moving traverse immediately downstream
of the exhaust opening with a 10 cm (4 in) Davis
vane anemometer. Both fans had approximately a
23.6 m3 /s (50,000 cfm) output, keeping this opera-
tional parameter constant in the study. The axial vane
fan is Joy Model AMF-100-50-26H. It is 460 volts,
37.1 kw (50 hp) with a 0.91 m (3 ft) outlet opening and
equipped with a flow straightener. The propeller fan is
a Spendrup Model 152-764-HVLD. It has a 20.0 kw
(27.2 hp) Duetz Model F2L1011F diesel engine. The
fan is belt driven with a 1.52 m (5 ft) diameter out-
let opening. Figure 2 is a plan view of the study area
showing information concerning the location of the
fans, the dump and crusher, the operator control booth,
ventilation stoppings, and dust sampling stations.
The fans were positioned in the same location,
on the left side of the haul road entry, approxi-
mately 30.3 m (100 ft) from the dump/crusher. The fan
discharge blew across the haul road entry and was ori-
ented toward the center of the dump area and crusher.
The width and height of the entry at the dump were
13.7 m and 18.3 m (45 ft and 60 ft), respectively. Gen-
erated dust was diluted and transported away from
the crusher via the belt entry, into a return airway,
and out of the mine. The belt entry is isolated from
the main developments using both permanent and
Figure 1. Axial vane fan and propeller fan used in study.
curtain stoppings in crosscuts along its entire length
of approximately 152 m (500 ft) as shown in figure 2.
The crusher is a 222.6 kw (300 hp) jaw type rated at
907 metric tph (1000 short tph).
The dust survey consisted of two parts. The dust
survey was first conducted with the axial vane fan,
routinely used by the mine to ventilate the facility.
After completion of the survey, the axial vane fan
was replaced by the propeller fan which was posi-
tioned in the same location. The sampling survey was
then repeated. As shown in figure 2, the sampling
strategy was to collect dust samples at six key loca-
tions around the dump/crushing facility. This sampling
array provided dust concentrations at the same loca-
tions to determine the effectiveness and performance
of each fan as it relates to the dilution and transport of
respirable dust and respirable silica dust.
The instrument package at each sampling loca-
tion varied depending on the information required at
each location. Two types of instruments were used.
The first instrument type that was used for mea-
suring respirable dust was the gravimetric sampler

204

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 1. Dust samplers used in study.

Sampling Instruments

Site Location Gravimetric pDR

1 Intake Entry
2 Dump
3 Crusher
4 Belt
5 Return Entry
6 Adjacent Entry
4 0
2 1
2 1
2 1
4 0
2 0

and records the concentration of respirable airborne

Figure 2. Plan view of the study area showing the location
of the blowing fans (axial vane or propeller), the dump and
crusher, the operator control booth, ventilation stoppings, and
dust sampling stations.
operated at 1.7 liters/minute which complies with
MSHA requirements for metal/nonmetal sampling.
The respirable fraction of the airborne dust (<10
microns) was obtained with a 10 mm Dorr-Oliver
cyclone and deposited onto a 37 mm PVC filter. The
filters used for dust were weighed before and after
sampling to calculate overall respirable dust concen-
trations (which includes all dust types and particulate)
based on the sampling rate and time. The filters were
then analyzed by an independent laboratory for silica
using NMAM-Method 7500 to determine the silica
weight. From this information, silica concentrations
could be calculated.
The second type of sampling instrument was the
MIE personal DataRAM (pDR). The pDR measures
particulate from 0.1 to 10 microns (which includes all
dust types, DPM, and water mist) using a light scat-
tering technique. Light-scattering instruments offer
a relative measure of concentrations, but provide a
continuous record of particulate levels so that concen-
trations can be evaluated over any time interval during
the sampling period.
Table 1 identifies the types and numbers of dust
samplers that were positioned at each location. Sam-
pling packages on the ribs were positioned about
1.82 m (6 ft) above the floor. Figure 2 shows a plan
view of the location of these sampling stations. Sam-
ples for both parts of the study were collected on
3 consecutive days for an average sampling time of
about 5 to 6 hrs per shift. Other information related to
dust production and migration was collected each day
during the sampling period. During this time, the num-
ber of trucks that dumped and the tonnage processed
through the crusher were recorded.
To monitor airflow from the crusher to the return
airway, vane anemometer readings were taken at the
end of the belt entry (site 4 in figure 2) at a 0.91 m by
2.3 m (3 ft by 7.5 ft) regulator in a permanent stopping
leading to the return. Vane anemometer readings of
the air velocity were also taken in front of the dump
to compare the ventilation coverage provided by each
fan. To determine the bulk percentage of silica in the
rock, samples were collected from the belt during both
parts of the study and sent to an independent laboratory
analysis. Percentages of silica in the rock ranged from
37.8% to 44.4%.
3 GRAVIMETRIC SAMPLING
Figures 3 and 4 summarize the concentrations based
on filter weights from gravimetric sampling at all six
locations for the 3 sampling days for the respirable dust
and silica dust, respectively. Both the respirable and
silica dust concentrations were examined to determine
if fan effectiveness varied depending on the type of
dust. In figure 3, the respirable dust concentration is
most likely composed of three main components: inert

205

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 3. Average respirable dust concentrations from
gravimetric sampling at all six locations for the 3 sampling
days.
Figure 4. Average respirable silica dust concentrations from
gravimetric sampling at all six locations for the 3 sampling
days.
limestone dust or calcite, diesel particulate, and silica.
Filters were sent to an independent laboratory for XRD
silica analysis of three primary mineral components:
quartz, cristobalite, and tridymite. The analysis only
found quartz mineral on the filter. Figure 4 shows the
concentrations of silica at the six locations around the
crusher.
For each sampling location, the mean and 95% con-
fidence interval were calculated using a t-distribution.
The graphs in figures 3 and 4 plot the mean concentra-
tion and the upper and lower confidence limit (UCL
and LCL) for each sampling location. In examining
all locations, the following is notable in comparing
the axial vane to the propeller fan:
Site 1 – Intake: Gravimetric sampling showed a
slight increase in both respirable dust and silica dust.
This station was not influenced directly by the fans,
but the low dust levels indicate that very little if any
dust is migrating from the dump and crusher back into
the main developments on the intake side. Observed
concentration increases are most likely related to dust
generated from vehicle traffic in the area.
Site 2 – Dump: This station recorded a reduction
in both respirable and silica dust of 23.1% and 24.8%
206
Copyright © 2006 Taylor & Francis Group plc, London, UK
respectively, with the propeller fan. When compared
to the crusher, the dump has low respirable and silica
concentrations. This suggests that both fans are pre-
venting dust rollback from the crusher as the trucks
dump.
Site 3 – Crusher: This location had the highest con-
centrations of both respirable and silica dust. Of inter-
est is the fact that respirable concentrations increase
threefold from the dump to the crusher location, a dis-
tance of roughly 18.2 m (60 ft). This station recorded
a reduction in both respirable and silica dust of 17.5%
and 17.9% respectively, with the propeller fan. Obser-
vation from inside the operator’s booth showed that
a large plume of dust was created as trucks dumped
regardless of the fan in use. The propeller fan appeared
to be more effective in diluting the dust than the axial
vane fan. Stratification or layering of the air may be
causing this effect as both fans are lifting and entrain-
ing the dust, but the propeller fan is more effective in
dust transport.
Site 4 – Belt: Both the respirable and silica dust
concentrations at the belt location are less than half of
the levels at the crusher, at a distance of approximately
152 m (500 ft). Dust reductions were 3.7% and 12.8%
for the respirable and silica dust, respectively. How-
ever since the concentrations mean values show little
change, this confirms that the dust is well diluted and
uniform when it reaches the end of the belt before it
passes through the regulator in the stopping. This indi-
cates that both fans are slowly moving the air down the
entry with about the same efficiency.
Site 5 – Return: This location behaved much the
same as the intake location with low respirable and
silica concentrations showing that very little dust is
migrating from the dump/crusher back into the main
developments. This site is the only station to record
conflicting concentration levels in that the respirable
dust decreased with the propeller fan but the silica
dust increased. Similar to site1 intake, this station was
not directly influenced by the fans and observed con-
centration increases are most likely related to dust
generated from the tramming of haulage trucks past
the sampling station.
Site 6 – Adjacent Entry: Since dust concentrations
were similar to the dump and the belt locations, this
indicates that considerable air leakage is occurring
through the line curtains along the belt entry. As a
result, dust is being transported through the stopping
line and has the potential to be carried toward the work-
ing faces. Reductions in both respirable and silica dust
were recorded when the propeller fan was in use but
dust concentrations in this adjacent entry can easily be
reduced by tightening the curtain line to prevent the
escape of dust.
Figures 3 and 4 show that the propeller fan out-
performed the axial vane fan in the dilution of respi-
rable dusts as evidence for the lower mean values
 at sampling sites most influenced by the fans. This
includes the dump and crusher, belt, and the adjacent
entry. However, since at all sampling sites the error
bars (upper and lower confidence intervals) overlap,
the dust reductions are not statistically significant for
the amount of data collected. For larger dust particles
or nuisance dust (>10 microns), visible dust from the
crusher still appeared to linger in the air with rela-
tively slow removal from this area regardless of fan
type. Although total dust measurements were not col-
lected, the primary reason for this observation is that
the area of the entry at the dump increases dramati-
cally at the crusher, which results in a significant drop
in air velocity.
Figure 5. Dust patterns for axial vane fan at the dump,
crusher, and belt for a typical day of sampling.

4 PERSONNEL DATA RAMS

The pDR graphs in figures 5 and 6 illustrate the dif-

ference in dust patterns between dump, crusher and
belt for a typical day of sampling. These graphs com-
pare pDR data from the third day of sampling with
the axial vane fan, and the first day of sampling with
the propeller fan. These days were select because ton-
nage processed through the crusher those days were
similar at 3890 metric tons (4290 short tons) and 3820
metric tons (4210 short tons), respectively. The pDR
concentrations in the figures have been corrected to the
average gravimetric concentrations at those locations.
This is simple correction by calculating a ratio of aver-
age gravimetric concentration divided by the average
pDR concentration for the entire sampling period. This
ratio is then used to adjust each individual concentra-
tion in the pDR log. The pDR was configured to log
a concentration every 10 seconds over a 285-minute
sampling period.This equates to 1710 sampling points.
Two primary observations can be made from these
graphs. First, unlike the concentrations from the gravi-
metric samplers (figure 3), the average concentrations
at each of these three locations were similar, regard-
less of fan type. These results are not uncommon,
since the pDR a light scattering instrument which
yields a relative concentration value depending on the
aerosol being sampled. Differences in the content of
the aerosol, (in this case limestone dust, silica dust, and
diesel particulate matter) can vary during the sampling
period and may account for the similarity in corrected
concentrations.
Second, although these light-scattering instruments
offer only a relative measure of concentrations, they
provide a continuous record of dust levels so that con-
centrations can be evaluated over any time interval
during the sampling period. Examining concentration
trends over this time period is the most useful way
to interpret the data. In comparing the two graphs,
concentration trends both the dump and belt sampling
Figure 6. Dust patterns for propeller fan at the dump,
crusher, and belt for a typical day of sampling.
stations tend to have a similar pattern during the dump
and crushing cycle. What is most notable in the figures
is the differing trend at the crusher. The peak and val-
ley concentrations are consistently higher and lower
with the propeller fan. The dump and crushing cycle
was a continuous process with only about a 30-second
delay between trucks arriving to dump. The crusher
was usually behind the dump cycle and trucks had to
wait for the crusher to process most of the rock before
dumping into the hopper. The trucks would idle at the
dump before being signaled to dump. Figure 7 inset
picture shows a truck waiting at the dump.
Figure 7 compares the three-shift average of
anemometer readings taken directly in front of the
dump. The three fixed point measurements were
spaced approximately 4.6 m (15 ft) apart and 1.5 m
(5 ft) above the floor. These reading were taken before
the shift or when no trucks were at the dump. Due to
the operational characteristics of each fan, the axial
vane fan did not provide air coverage at the center
and right side of the dump as did the propeller fan.
As a result, with a truck waiting at the dump, it is
likely that the airflow became more turbulent with

207

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 7. (1) Three shift average of anemometer readings taken across the dump; (2) Three shift average air velocity at the
regulator at the end of the belt; (3) Inset picture showing a truck waiting at the dump.

Table 2. Daily production for each sampling shift.

Axial Vane Fan Propeller Fan

Shift number 1 2 3 1 2 3

Gravimetric Sampling time, min 216 322 298 325 312 297
Number of trucks dumped 55 110 110 108 107 108
Measured tonnage, metric tons (short tons) 1945 3890 3890 3820 3785 3820
(2145) (4290) (4290) (4210) (4170) (4210)

the propeller fan. As airflow moved around the truck,
increase turbulence may have entrained more dust,
leading to higher concentration values logged by the
pDR (Chekan et al, 2004b; Xu and Bhaskar, 1995).
When the truck left the dump, the airflow provided
by the propeller fan diluted the dust produced by the
crusher more effectively, as evident by the immediate
reduction in dust levels.
Figure 7 also shows the three-shift average air veloc-
ity at the regulator at the end of the belt. The air flows
are comparable, indicating that air losses through the
brattice stoppings isolating the belt are similar for each
fan. Figures 5 and 6 show that dust levels at site 4 (belt)
are comparable in trend and concentration. This indi-
cates that both fans are equally effective in transporting
the dust once it has been diluted at the dump/crusher
area.
5 RESPIRABLE DUST GENERATION AND
PRODUCTION
Another method to evaluate the data, other than con-
centration, is to calculate dust generated based on the
tonnage produced during the sampling period. This
will normalize dust generation to production, and offer
an alternative approach to compare the effectiveness
of each fan. This value is calculated by taking the aver-
age milligram weight gain collected on the filter for
a shift, and dividing it by the metric tons processed
through the crusher for the duration of the sampling
period. It is expressed as milligrams of respirable dust
per metric ton of rock crushed per minute. Table 2 pro-
vides specific details about sampling time and tonnage
processed by the crusher for each sampling shift. For
this analysis, the following four sampling sites most

208

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 8. Three-shift average comparing respirable dust
generated based on the tonnage produced during the sampling
period.
Figure 9. Three-shift average comparing respirable silica
dust generated based on the tonnage produced during the
sampling period.
influenced by fan airflow were examined: the dump,
the crusher, the belt, and the adjacent entry.
Figures 8 and 9 compare the 3 shift average of both
the respirable dust and the respirable silica dust for
each fan. The decrease in respirable dust based on
tonnage is apparent when the propeller fan was used.
These results are similar to the percent decease in dust
concentrations at the dump, crusher, belt, and the adja-
cent entry shown in figures 3 and 4 and offer further
evidence that the propeller fan outperformed the axial
vane fan for dust dilution.
6 CONCLUSIONS
Based on gravimetric sampling, the study showed an
improvement with the propeller fan to dilute both the
respirable dust and the respirable silica dust around
the dump/crusher facility. Although the dust reduc-
tions are not statistically significant at a 95% confi-
dence, concentration values show that the propeller
209
Copyright © 2006 Taylor & Francis Group plc, London, UK
fan outperformed the axial vane fan in the dilution of
respirable dusts as evident by the percent reduction
in mean concentration values at sampling sites most
influenced by the fans (i.e., the dump, the crusher,
the belt, and the adjacent entry). In addition, when
examining dust generation as it related to production
the percent reductions offer further support that the
propeller fan outperformed the axial vane fan. Since
reductions for both the respirable dust and the res-
pirable silica were approximately the same whether
comparing concentrations (figures 3 and 4) or pro-
duction (figures 8 and 9), this shows that the propeller
fan’s effectiveness did not vary depending on dust type.
However, one observation needs to be addressed.
For larger dust particles or nuisance dust (>10
microns), visible dust from the crusher still appeared
to linger in the air with relatively slow removal from
this area regardless of fan type. The reason for this
observation is that the area of the entry at the dump
increases dramatically at the crusher, which results in
a significant loss of air velocity.
The pDR concentration showed that concentration
trends at the dump and belt sampling stations tended
to have a similar pattern during the dump and crush-
ing cycle. However, when comparing the concentration
trends at the crusher, the peak and valley values are
consistently higher and lower with the propeller fan.
Due to the operational characteristics of each fan, the
axial vane fan did not provide air coverage at the center
and right side of the dump as did the propeller fan. As
a result, with a truck idling at the dump, it is likely that
the airflow became more turbulent with the propeller
fan. When the truck left the dump, the airflow pro-
vided by the propeller fan diluted the dust produce by
the crusher more effectively, as evident by the imme-
diate reduction in dust levels. Data indicates that both
fans are equally effective in transporting the dust once
it has been diluted at the dump/crusher area. This is
based on similar air velocities measured at the regu-
lator and comparable concentration values and trends
from the gravimetric and pDR samplers.
This survey was conducted to evaluate two different
types of fans for dust dilution and transport from an
underground crushing facility during normal produc-
tion activities. Dust concentrations around the crusher
and down the belt entry were higher than desired
for either fan and could be reduced with improved
dust capture. The current fan location, approximately
30.3 m (100 ft) from the dump/crusher, is perform-
ing a function by diluting the dust at the dump and
keeping it from recirculating back to the main develop-
ments. The use of brattice around the crusher to better
direct and increase airflow would provide an inexpen-
sive method to possibly reduce dust levels. Another
alternative to consider would be to increase fan capac-
ity using either a push-pull system with two auxiliary
fans or a fan-powered dust collector. The push-pull
 system would require an axial vane fan to be placed
outby the crusher in the belt entry with exhaust tubing
positioned as close to the crusher as possible to max-
imize dust capture. Tubing would then be attached to
the exhaust side of the fan to transport captured dust
directly to the return airway. The second alternative
would involve the installation of a fan-powered dust
collector with filtration system to remove airborne dust
and discharge clean air. With the additional fan capac-
ity, either system should increase dust capture at the
crusher, thus lowering dust levels at the crusher and in
the belt entry.
REFERENCES
Chekan, G.J., Colinet, J.F. & Grau, R.H. 2004a. Evaluat-
ing Ventilating Air Movement in Underground Lime-
stone Mines by Monitoring Respirable Dust Gener-
ated from Production Shots. Proceedings of the 10th
US/NorthAmerican MineVentilation Symposium.Anchor-
age, Alaska, May 16–19, 2004, pp. 221–232.
Chekan, G.J., Colinet, J.F. & Grau, R.H. 2002. Silica Dust
Sources in Underground Metal/Nonmetal Mines. SME
Transactions 2002, Vol. 312, pp. 187–193.
Chekan, G.J., Listak, J.M. & Colinet, J.F. 2004b. Factors
Impacting Respirable Dust Entrainment and Dilution in
High Velocity Airstreams. SME Transaction 2004, Vol.
316, pp. 186–192.
210
Copyright © 2006 Taylor & Francis Group plc, London, UK
Grau, R.H., Robertson, S.B., Mucho, T.P., Garcia, F. &
Smith, A.C. 2002a. NIOSH Ventilation Research Address-
ing Diesel Emissions and other Air Quality Issues in
Nonmetal Mines. SME Pre-print 02-187, SME Annual
Meeting, Phoenix, Arizona, February 25–27, 2002, pp. 7.
Grau, R.H., Mucho, T.P., Robertson, S.B., Smith, A.C. &
Garcia, F. 2002b. Practical Techniques to Improve the Air
Quality in Underground Stone Mines. Proceedings of the
9th North American/US Ventilation Symposium Kingston,
Ontario, June 8–12, 2002, pp. 123–129.
Head, H.J. 2001. Proper Ventilation for Underground Stone
Mines. Aggregates Manager, January, 2001, pp. 20–22.
Kissell, F.N. & Volkwein, J.C. 2002. Improving Ventilation
in Underground Stone Mines. Aggregates Manager, April,
2002, pp. 20–25.
Krog, R.B. & Grau, R.H. 2006. Correct Fan Selection for
Large Opening Mines: Axial Vane or Propeller Fans –
Which to Choose? Proceedings of the 11th North Amer-
ican/Ninth US Ventilation Symposium. University Park,
Pa., June 5–7, 2006.
Mine Safety and Health Administration (MSHA). 2000–
2004. Metal/Nonmetal Mine Inspection Data (MNMID),
available from the MSHA Metal and Nonmetal Safety and
Health Office in Arlington, Va.
Timko, R.J. & Thimons, E.D. 1987. Damage Resistant Brat-
tice Stoppings in Mines With Large Entries. Engineering
Mining Journal, Vol. 188, No. 5, pp. 34–36.
Xu, L. & Bhaskar, R. 1995. A Simple Model for Turbulent
Deposition of Mine Dust. Proceedings of the 7th US Mine
Ventilation Symposium. Lexington, Ky., June 5–7, 1995,
pp. 337–334.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Emerging technologies control respirable dust exposures for

continuous mining and roof bolting personnel

G.V.R. Goodman, T.W. Beck, D.E. Pollock, J.F. Colinet & J.A. Organiscak
Pittsburgh Research Center, National Institute for Occupational Safety and Health, Pittsburgh, PA, USA

ABSTRACT: This work presents the findings from a number of NIOSH studies evaluating the impacts of
emerging technologies that may reduce dust exposures for continuous mining and roof bolting personnel. These
technologies include use of a wet head cutting drum on a continuous mining machine that places water sprays on
the cutting drum instead of placing them on a manifold outby the drum. Evaluations at two separate operations
showed that the use of the wet head sprays reduced dust levels at the miner operator in the return air, although
these reductions were quite variable. NIOSH also assessed the effectiveness of a canopy air curtain for protecting
roof bolting personnel. The data showed that dust levels were lower beneath the air curtain than outside the air
curtain. This study also suggested several future modifications to the air curtain to increase its effectiveness.
Finally, NIOSH evaluated mist drilling technology for its effectiveness in controlling respirable dust exposures
of bolting personnel. A recent study showed that higher dust levels existed around a machine using the mist
system compared to a machine using a conventional vacuum system for dust control. Unfortunately, the true
impact of mist drilling was confounded by this machine operating downwind of the continuous miner for much
of the study.

1 INTRODUCTION Continuous mining machines typically are equipped

The Federal Coal Mine Health and Safety Act requires
that the concentration of respirable dust in the coal
mine atmosphere not exceed 2.0 mg/m3 for a working
shift. If quartz content on the dust filters exceeds 5%
by weight, the dust standard is reduced according to the
following expression: 10/(% quartz). Compliance with
the standard or a reduced standard maintains quartz
dust exposures at or below 100 µg/m3 .
Over 17,000 samples collected by Mine Safety and
Health Administration (MSHA) inspectors during the
period 2000–2004 (MSHA 2004) were analyzed for
both respirable coal and quartz dusts. This included
over 12,000 samples at the miner operator occupation
and 5,000 samples at roof bolter occupations. Nearly
2,000 samples of the miner and bolter showed exces-
sive occupational exposure to respirable coal dust
while 3,000 samples at the same occupations showed
quartz exposures in excess of 100 µg/m3 .
Such levels present unacceptable health risks to
the U.S. coal miner. The high incidences of coal and
quartz overexposures suggest that existing control sys-
tems do not consistently limit respirable quartz dust
exposures at roof bolter and miner operator occupa-
tions. Research must, therefore, focus on evaluation
and improvement of these needed controls.
with external water spray systems and a fan-powered
dust collector (scrubber), while roof bolting machines
are equipped with vacuum (dry) dust collection sys-
tems. Novel technologies have been developed that
offer the potential to further reduce dust exposures. A
“wet head” continuous mining machine (water sprays
situated in the cutting drum), a filtered air supply
(canopy air curtain) for a roof bolting machine, and
mist drilling technology are available for implementa-
tion on continuous mining operations. Each of these
technologies offers the potential for reducing the dust
exposures of underground coal miners.
This paper documents NIOSH evaluations of these
emerging technologies for controlling worker expo-
sures to respirable dust. Recommendations are pro-
vided for effective use of these methods.
2 CONTINUOUS MINER DUST CONTROL
2.1 Wet head cutter drum
Water sprays are the most widely used technique
for limiting worker exposure to respirable dust in
underground coal mining. Water sprays wet surfaces
to prevent airborne dust generation, which is most
effective when the water sprays are placed as close

211

Copyright © 2006 Taylor & Francis Group plc, London, UK

 as possible to these surfaces. Typical mining machines
place water sprays approximately 30–40 cm away from
cutting bits to protect the spray manifolds. Wet head
cutting drum designs, on the other hand, introduce
water via sprays located directly behind each cutting
bit on the cutting drum. This places water at each
cutting bit and in the region where the dust cloud
forms, a significant advantage for limiting formation
and growth of this cloud.
Early studies of a prototype wet head machine
showed 25–40% reductions in respirable dust levels
at the mining machine operator (Strebig 1975, Kost
et al. 1976). Return dust levels showed similar reduc-
tions, an important consideration for miners working
downwind of the continuous miner. However, com-
prehensive studies of wet head machines were not
conducted to evaluate the impacts of external spray
and scrubber use on dust reduction potentials.
Poor mechanical reliability and high maintenance
requirements of the wet head drum delayed its initial
acceptance by the underground coal mining industry
(Belle et al. 2002). A major manufacturer of under-
ground mining equipment has recently introduced a
redesigned version of the wet head cutting drum that
improves both reliability and maintenance.
NIOSH conducted two multi-shift studies of this
new design to evaluate the effectiveness of the wet
head cutting drum for limiting respirable dust expo-
sures for the miner operator and downwind personnel.
Pairs of gravimetric samplers were hung in the intake
and return air of the continuous mining machine
while the miner operator wore a single sampler. These
samples were not collected for a full shift; hence
the results could not be correlated to compliance
sampling.
2.2 Wet head study at Mine A
This study compared dust levels around a single
machine that could be configured with either a wet
head spray system or a regular spray system. The
machine ran with three configurations: (1) with the
regular drum sprays and with the flooded-bed dust
scrubber, (2) with the wet head sprays and with the
scrubber, and (3) with the wet head sprays but without
the scrubber. NIOSH evaluated the first and sec-
ond configurations for five shifts each and the third
configuration for two shifts.
Mining height averaged 2.0 m for all test configu-
rations with rock partings accounting for half of the
height in some entries. A combination of a flooded-
bed dust scrubber rated at 3.3–3.5 m3 /s and an exhaust
line curtain hung along the right rib ventilated the face.
Exhaust curtain flow averaged 7.1 m3 /s for configura-
tion 1, 6.2 m3 /s for configuration 2, and 5.6 m3 /s for
configuration 3. Shift production averaged 500 tons
(range 290–575 tons) for configuration 1,620 tons
212
Copyright © 2006 Taylor & Francis Group plc, London, UK
Respirable dust, mg/m3
8
6
4
2
0
Regular machine + Wet head + scrubber Wet head without
scrubber scrubber
Sampling location
Intake Miner operator Return
Figure 1. Gravimetric dust levels for Mine A, with and
without wet head sprays.
(range 360–810) for configuration 2, and 420 tons
(range 330–500 tons) for configuration 3.
The regular spray system consisted of twenty-seven
hollow-cone water sprays for dust control mounted
externally on the frame of the continuous miner, with
an equivalent number located on the cutting boom.
Water spray pressure on the external sprays varied
from 1,034 to 1,276 kPa measured with a dial gauge
on a right-hand end-ring spray. The wet head drum
contained 73 small orifice solid stream sprays (each
delivering 1.1 liters/min at 689 kPa) oriented to spray
directly on the 73 cutter bits. These sprays could be
turned on or off depending on required test condi-
tions. Wet head spray pressures varied from 448 to
655 kPa when measured by company personnel prior
to testing. Water flow rates on the continuous miner
were 174.3 liters/min with the wet head sprays oper-
ating and 181.9 liters/min with the regular sprays. In
addition to the wet head sprays, several sprays on the
cutting boom remained open to provide cooling water
to the cutting motors and contributed to the total flow
measured for the wet head spray system.
Average gravimetric dust levels measured at the
miner operator and in the return airway were slightly
lower using the wet head cutting drum and the dust
scrubber (Fig. 1). Removing the contribution of intake
dust from miner operator and return dust levels showed
that operation of the wet head sprays reduced dust lev-
els by 0.5 and 0.3 mg/m3 , respectively at these two
locations. However, the true impact of these sprays on
respirable dust levels was likely hidden by the variabil-
ity of the gravimetric data collected at these locations.
For instance, miner operator dust levels ranged from
2.20 to 4.52 mg/m3 , 0.94 to 4.26 mg/m3 , and 1.94 to
6.15 mg/m3 for configurations 1, 2, and 3, respectively.
Gravimetric dust levels in the return exhibited similar
variations.
Relatively high quartz dust levels in the intake
airway were observed for configurations 1 and 2
(Table 1). These levels tended to be higher when rock
was being crushed in the feeder breaker and its dust
entered the intake air. This and the cutting of a sand-
stone rider above the coal seam elevated miner operator
 Table 1. Quartz dust levels for three configurations during
wet head study, Mine A.
Respirable quartz levels,
Average (range), µg/m3
Miner
Configuration Intake Operator Return
1 103 (0–262) 356 (140–613) 413 (230–739)
2 82 (0–192) 311 (89–753) 359 (161–579)
3 20 (0–39) 372 (191–553) 621 (436–806)
quartz dust levels on several shifts. The highest quartz
exposures for configuration 1 were recorded at the
miner operator and in the return while cleaning up
a roof fall from the previous shift.
The data showed much higher dust levels for config-
uration 3, indicating that the wet head sprays could not
control dust levels without the dust scrubber. The large
amount of dust around the continuous miner made it
difficult for the operator to see the cutter head. Based
upon the sampling results, NIOSH recommended that
the wet head machine not operate without the scrubber.
2.3 Wet head study at Mine B
This study compared dust levels around two contin-
uous mining machines: one equipped with wet head
sprays and the other equipped with a conventional
spray system. Cutting height varied with face loca-
tion but averaged 1.8 m for the regular machine with
an average rock thickness of 0.5 m. Cutting height and
rock thickness averaged 1.7 m and 0.3 m, respectively,
for the wet head machine. Evaluation of the regular
and wet head machines took place over four shifts.
Section ventilation was single split with the wet
head continuous miner operating upwind of the reg-
ular mining machine during the study (operation was
not simultaneous). Blowing face ventilation for both
machines was provided by an intake line curtain hung
along the right rib and by a flooded-bed dust scrub-
ber rated at 3.2 m3 /s on the wet head machine and
3.0 m3 /s on the regular mining machine. Airflow quan-
tities behind the intake curtain averaged 4.6 m3 /s for
the wet head mining machine and 4.4 m3 /s for the reg-
ular machine. Production for the wet head averaged
610 tons (range 430–740 tons) and 680 tons (range
580–910 tons) for the regular miner.
The wet head continuous miner contained a total
of 89 sprays, 63 of which were small orifice sprays
located on the cutter head. The wet head sprays (each
delivering 1.5 liters/min at 689 kPa) were oriented to
spray directly on the 63 cutting bits. Wet head spray
pressure was approximately 621 kPa as measured by
company personnel prior to testing. With the scrubber
213
Copyright © 2006 Taylor & Francis Group plc, London, UK
5
Respirable dust, mg/m3
4
3
2
1
0
Wet head machine Regular Machine
Sampling location
Intake Miner operator Return
Figure 2. Gravimetric dust levels for Mine B, with and
without wet head sprays.
operating, water flow on the wet head machine varied
from 144 to 159 liters/min measured by a flow meter
on the right side of the machine frame. External spray
pressure was 1,034 kPa measured with a dial gauge
on this machine. Six sprays were open on the cutting
boom to provide cooling water to the cutting motors.
The regular machine contained a total of 41
dust control sprays, each delivering 2.6 liters/min at
689 kPa. Water pressure on the sprays was 1,034 kPa
when measured by a dial gauge on the mining machine.
A flow rate measurement was not possible on this
machine because of a broken flow meter.
The intake and return samplers assigned to each
machine ran only when that mining machine operated.
The results of this study indicate lower dust levels at
the three sampling locations when using the wet head
(Fig. 2). After removing the intake dust contribution,
the reduction at the miner operator location was only
0.2 mg/m3 when operating the wet head sprays. Inter-
estingly, the reduction in the return was nearly 2 mg/m3
when using the wet head sprays.
Respirable quartz dust levels varied between the wet
head and regular mining machines. Measurable lev-
els were not detected on the intake airway or miner
operator samples of the wet head machine. The return
airway showed a respirable quartz level of 81 µg/m3
on the first shift, while return airway samples from
other shifts showed no detectable quartz mass. Mea-
surable levels of respirable quartz were not found in the
intake airway samples of the regular mining machine.
Quartz dust levels for the miner operator of the regu-
lar machine averaged 70 µg/m3 (range 42 to 97 µg/m3 )
while return airway levels averaged 143 µg/m3 (range
108–193 µg/m3 ).
The true impacts of the wet head sprays on
respirable coal and quartz dust levels were likely
confounded by excessive variability in production,
ventilation, and geologic factors between the two
machines. It is the contention of the authors that dust
reductions shown in figure 2 and quartz reductions
described earlier may not have been due solely to the
wet head sprays, but also to these variations.
 Bolter operator's
Inlet filter canopy

Centrifugal fan Flex tubing Air curtain

flow

Figure 3. Schematic of canopy air curtain.

3 ROOF BOLTER DUST CONTROL

Figure 4. Canopy air curtain installed on roof bolter.
3.1 Canopy air curtain

Quartz dust levels, ug/m3

3 140

Coal dust levels, mg/m3

120
Previous work by NIOSH researchers showed that dust
100
levels for roof bolter operators could increase when 2
80
working downwind of the continuous mining machine 60
(Goodman et al. 2002). A canopy air curtain system 1 40
was developed by NIOSH researchers and tested as a 20
means for reducing these exposures. This device con- 0 0
sists of a hollow metal plenum with a perforated sur- Beneath air Tool tray- Tool tray-
face on one side. The plenum is connected by flexible curtain left side right side

tubing to the output of a small centrifugal fan. A filter Sampling locations

is placed on the intake side of the fan. Placed on the Coal dust Silica dust
underside of the operator’s canopy, filtered air flows
from the perforated plate and over the breathing zone Figure 5. Respirable coal and quartz dust levels beneath air
of the person working beneath the air curtain (Fig. 3). curtain and around roof bolter.
Laboratory testing of this device showed that its
tramming cab. A 6.1 m length of 10.2 cm diameter spi-
effectiveness in limiting dust levels beneath the canopy
ral tubing connected the air curtain to the centrifugal
was mostly a function of the velocity of the air mov-
fan. Airflow to the air curtain was cleaned by an EPG
ing across the face of the air curtain. Higher velocity
model G110120 filter (Donaldson, Inc., Minneapolis,
air disrupted the air flowing from the air curtain, thus
MN) mounted on the intake side of the fan. The canopy
reducing its effectiveness. Testing of the air curtain
air curtain system provided a steady stream of filtered
showed dust reductions of 40% to 60% at a low air
air over the operator’s breathing zone.
velocity of 0.3 m/s, a typical value in the bolter head-
Dust sampling was conducted for three shifts
ings of many underground coal mines (Goodman et al.
beneath the air curtain and on the left and right sides of
2001).
the bolting machine on the tool trays. Coal and quartz
An underground study by NIOSH evaluated the
dust levels beneath the air curtain were 1.2 mg/m3
canopy air curtain for its effectiveness in controlling
and 40 µg/m3 lower than dust levels on the tool trays
respirable dust exposures for the bolter operators. Res-
(Fig. 5). However, these samples were collected close
pirable dust levels were measured around a twin boom
to the perforated plate of the air curtain because of
roof bolting machine equipped with a canopy air cur-
headroom constraints and, as such, represent an “opti-
tain mounted on the underside of the canopy of the
mum” scenario for dust protection effectiveness. Data
downwind operator (Fig. 4). The canopy air curtain
collected during this study suggested several ways to
was constructed of 14-gauge steel and measured 61 cm
improve the efficiency of the canopy air curtain, such
long, 25 cm wide, and 5 cm thick. The underside of
as increasing air curtain size to increase coverage area
the air curtain was a section of perforated plate steel
and reducing tubing lengths from the fan to the air cur-
(2.4 mm diameter holes, 4.8 mm staggered spacing).
tain (requiring positioning of the fan closer to the drill
The inside of the air curtain contained a 3.2 mm cell
head).
polycarbonate honeycomb material to straighten and
baffle the air flowing into the air curtain. A small cen-
3.2 Mist drilling
trifugal fan (American fan, model AF-10-R12327-6)
supplied 0.11 m3 /s to the air curtain and was mounted A mist head bolter injects a combination of water and
at the rear of the roof bolting machine adjacent to the compressed air through the drill steel to the drill bit to

214

Copyright © 2006 Taylor & Francis Group plc, London, UK

 6 9.0% (range 7.6 to 10.4%) around the vacuum bolter.
Respirable dust, mg/m3
5 The higher quartz content around the mist bolter was
4 likely a result of this machine working downwind of
3
2
1
0 levels led to higher quartz dust concentrations around
Mist
Bolter type
Intake Bolter area
Figure 6. Impact of mist drilling on gravimetric dust levels.
control drill dust, instead of using a conventional vac-
uum system with canister or bag filters. Air is supplied
by an on-board compressor while the water is supplied
by an on-board water tank or by a water hose dragged
behind the machine.
An underground study compared dust levels around
two roof bolting machines, one using a mist system
and the other using a conventional vacuum system to
control drill dust. The mist bolting machine injected
water and compressed air to the left bolting arm at
rates of 2.6 liters/min and 75.8 liters/min, respectively.
Water and air were supplied to the right arm at rates
of 1.1 liters/min and 75.8 liters/min, respectively. The
vacuum bolter used a four-compartment permissible
vacuum dust collection system with a single canis-
ter filter to collect drill dust. These machines did not
operate simultaneously.
Two gravimetric samplers were attached near the
left and right operator’s work location on each machine
approximately 30–60 cm outby their controls. The
bolter area dust level for each machine was the aver-
age of the dust concentrations from the left and right
samples. Gravimetric samplers were also hung in the
intake air of each machine. The sampling pumps on
each machine operated only while that machine ran.
Three shifts of data were collected for each bolting
machine.
For much of this study, the mist bolter worked down-
wind of the continuous mining machine, resulting in
higher intake dust levels for this bolter than the vacuum
bolter that typically worked in the clean air upwind
of the continuous miner. As a result, gravimetric dust
levels measured around the mist bolter were signif-
icantly elevated compared to dust levels around the
vacuum machine (Fig. 6). Even after removing
the contribution of intake dust, dust levels around the
mist bolting machine still exceeded those around
the vacuum machine.
Quartz contents on the intake filters averaged
10.6% (range 10.3 to 10.9%) at the vacuum machine
and 9.5% (range 7.7 to 11.1%) at the mist machine for
a three-shift study. Quartz contents averaged 12.3%
(range 9.6 to 16.3%) around the mist machine and
Copyright © 2006 Taylor & Francis Group plc, London, UK
the continuous miner as it cut rock (rock thickness
varied with face location, but ranged from 0.6 to over
1.0 m out of a mining height of 1.7 m). This high quartz
dust content coupled with the high gravimetric dust
Vacuum
the mist bolter.
4 CONCLUSIONS
NIOSH evaluated emerging technologies that limit
respirable dust exposures for operators of continuous
mining and roof bolting equipment. These technolo-
gies include use of a wet head cutting drum on a
continuous mining machine that places water sprays on
the drum and closer to the dust cloud. An evaluation at
one operation showed that use of the wet head sprays
reduced dust levels 0.5 mg/m3 at the miner operator
and 0.3 mg/m3 in the return when adjusted for intake
dust levels. Dust reductions at a second operation were
about 0.2 and 2.0 mg/m3 at the miner operator and in
the return airway, respectively. In both studies, the true
impacts of the wet head sprays were likely confounded
by variations in gravimetric data, production, ventila-
tion, and geologic factors. NIOSH also assessed the
protection afforded roof bolter personnel by use of
a canopy air curtain, a device that blows filtered air
over the operator’s breathing zone. The data showed
a reduction in dust levels of over 1.2 mg/m3 beneath
the air curtain compared to levels around the bolter.
This study also revealed several ways to improve the
efficiency of the air curtain, including increasing air
curtain size to increase coverage area and reducing tub-
ing lengths from the fan to the curtain. Mist drilling
controls dust by injecting a combination of water and
compressed air through the drill steels to the drill bit,
instead of using a conventional vacuum system to draw
dust back through the steel. A recent NIOSH study
compared two bolting machines, one using a mist sys-
tem and the other using a conventional vacuum system.
The mist bolter worked downwind of the continuous
miner for much of the study causing higher intake dust
levels for this bolter. As a result, dust levels averaged
1.1 mg/m3 higher around the mist machine than around
the vacuum machine, even after removing the intake
dust contribution.
REFERENCES
Belle, B.K. & Clapham, S. 2002.An Improved Wet-Head Sys-
tem: Prevention of Incendive Ignitions and Dust Control.
In E. DeSouza (ed.), Proceedings, 9th US MineVentilation
Symp., Kingston, Ontario, 8–12 June, 2002. Abington:
Balkema.
215
 Goodman, G.V.R. & Organiscak, J.A. 2001. Laboratory Eval-
uation of a Canopy Air Curtain for Controlling Occupa-
tional Exposures of Roof Bolters. In S. Wasilewski (ed.),
Proceedings, 7th International MineVentilation Congress,
Krakow, Poland, 17–22 June, 2001. Krakow: EMAG.
Goodman, G.V.R. & Organiscak, J.A. 2002. Evaluation of
Methods for Controlling Silica Dust Exposures on Roof
Bolters. Transactions of Society for Mining, Metallurgy,
and Exploration, 312: 133–137.
Kost, J.A., Du Breuil, F. & Saltsman, D. 1976. Design, Devel-
opment, and Demonstration of a Technique to Water-Flush
216
Copyright © 2006 Taylor & Francis Group plc, London, UK
the Bits on a Fixed Drum Continuous Mining Machine.
Final Report H0210050, US Bureau of Mines. Open File
Report: 105–76.
Mine Safety and Health Administration. 2003. Respirable
Coal Mine Dust Data. Available from the MSHA Tech-
nology Center, Dust Division, Pittsburgh, PA 15236.
Strebig, K.C. 1975. ‘Wet-head’ Tests on Miners Concluded.
Coal Mining and Processing. 12(4): 78–80, 88.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

A field demonstration of an innovative spray system on

continuous miners for dust control

Y.P. Chugh, A. Patwardhan & H. Gurley

Department of Mining and Mineral Resources Engineering, Southern Illinois
University, Carbondale, IL, USA

R. Carty
Illinois Clean Coal Institute, Carterville, IL, USA

ABSTRACT: SIU researchers have developed an innovative spray system on continuous miners for improved
dust control. The spray system locates different types of sprays spatially to develop three lines of defenses
to minimize respirable dust emissions in and around continuous miner face areas. The system also utilizes
sprays to maximize scrubber efficiency for dust control. The SIU spray system utilizes approximately the same
number of sprays and the same amount of water as a conventional spray system on a continuous miner. This
redesigned spray system was tested extensively in the field on a cut-by-cut basis and compared side-by-side
with the conventional spray system at a Central Illinois mine. The results indicated a reduction in respirable
dust compared to the conventional system at the miner operator location and in the immediate return to be at
least 33.6% and 26.9%, respectively at 90% statistical confidence. The authors believe that there is margin for
even greater improvement with better face ventilation, higher water pressures and further modifications to the
spray system.

1 INTRODUCTION
Appropriately located water sprays on and around a
continuous miners have been extensively used to con-
trol dust exposures to the miner operator, batch haul-
age unit operator, and in the last open crosscut return.
A typical spray system employed on the continuous
miner has 15–30 sprays located across the top and the
sides of the boom. In addition under-boom sprays pro-
vide a water curtain to contain dust in the face area
and to wet it upon contact with the curtain. Based
on a review of pertinent literature and discussions
with mine operators, it appears there is no consen-
sus on the type and spatial location of sprays, and
volume of water and water pressure to be used in
sprays. Although general guidelines have been devel-
oped by researchers based on laboratory and field
studies, mine operators tend to use their own intu-
itive thinking in using the spray system to suit their
conditions.
Despite excellent research by the U.S. Bureau of
Mines and National Institute of Occupational Safety
and Health over the last 30 years, the authors realized
that a holistic approach to dust control in the face area
was lacking and there was a need to revisit the design
concepts of sprays on continuous miners to control
respirable dust in and around the face area.
1.1 Research goals
The overall goal of this study was to develop and
demonstrate in the field, concepts for dust control
in the face areas of room-and-pillar coal mining
operations with blowing face ventilation with partic-
ular emphasis on water spray systems on continuous
miners.
2 PREVIOUS STUDIES ON SPRAYS
The use and importance of hydraulic sprays to con-
trol airborne dust in underground coal mining are well
known. Numerous studies have reported improved
dust capture efficiency as a result of appropriate place-
ment of sprays and optimization of spray size, volume
and pressure. An excellent review of dust control in
coal mining, including the use of water sprays, was
provided by Kissel (2003). Kissel summarized that

217

Copyright © 2006 Taylor & Francis Group plc, London, UK

 water can be used to control dust through (1) Wetting
of broken material and (2) Capture of airborne dust.
Cheng & Zukovich (1973) indicated that a large
amount of dust created during coal cutting does not
become airborne and stays attached to the broken
material. Therefore, spraying more water on the cut
coal tends to reduce dust (Jankowski & Organiscak
1983, Ruggieri & Jankowski 1983, Zimmer et al. 1987,
Colinet et al. 1997). In addition to the volume of water,
mixing the water uniformly with broken coal is very
important (Hamilton & Knight 1957, Mukherjee &
Singh 1984, Volkwein et al. 1983). Proper mixing can
be accomplished by appropriate placement of sprays
on the miner or better still by adding water directly
at the bits that gets mixed with fragmented coal. For
example, Matta (1976) and Courtney et al. (1978)
have shown that sprays mounted under the boom are
more effective in reducing respirable dust than sprays
mounted on top of the boom. Jayaraman et al. (1984)
used a system of sprays consisting of top, side, and
underboom sprays to reduce dust roll back on the
operator by as much as 40%. Goodman (2000) demon-
strated the use of additional flat-fan sprays on the left
and right side scrubber inlets to better direct the dust
into the scrubber. However, Goodman et al. (1998)
found that sprays on and around the miner must be
used with caution and their improper application may
lead to reduced wet scrubber efficiency and result
in serious non-compliance with the dust standards.
The air entrainment properties of sprays also needs
to be considered when locating sprays to avoid dust
roll-backs (USBM 1981).
Recent Studies on the JOY Wet-Head miner and
HydraTool Wet-Head system have also documented
improved dust reduction compared to the conventional
sprays on continuous miners through the use of bit-
mounted water sprays (Chugh & Patwardhan 2005,
van Zyl 2005, van Zyl & Eroglu 2004).
Tien and Kim (1997) studied the effect of spray pres-
sure among other parameters on coal dust suppression
in a scale-model study. The influence of spray pres-
sure on dust suppression was found to be quadratic.
Dust suppression increased up to 60 psi spray pres-
sure beyond which it started decreasing. The results
are indicative of findings reported by Courtney and
Cheng (1977) that typical water sprays operating at
100 psi do not capture more than 30% airborne dust
in an open environment. Furthermore, inappropriately
designed sprays can displace dust clouds and cause
dust rollback rather than wetting or capturing air-
borne dust. Thimons et al. (1999) however reported
greatly improved air flow to a 40-ft box cut face
by the operation of spray systems on the miner at
120 psi.
The importance of fine size droplets is also well
known to be important for effective dust capture
(McCoy et al. 1983, Bigu & Grenier, 1989). In this
218
Copyright © 2006 Taylor & Francis Group plc, London, UK
regard, Bazzanella et al. (1986) suggested that larger
number of smaller size sprays is better for dust con-
trol than smaller number of larger sprays. Nozzles
operating at higher pressures are more likely to be
efficient in the use of water while providing similar air-
borne dust capture efficiency. However, high-pressure
sprays tend to disperse more dust. Therefore, their
use is more appropriate in relatively confined envi-
ronment. Atomizing nozzles are most efficient in
airborne dust capture followed by hollow cone, full
cone, and flat sprays. However, full cone and hollow
cone sprays are less likely to clog due to larger orifice
area. Also, another problem with atomizing sprays has
been pointed out by McCoy et al. (1983). They have
reported that reducing the droplet size through the
use of atomizing or fogging sprays may temporarily
improve the airborne dust capture efficiency. How-
ever, small droplets tend to collapse/evaporate easily
and release the captured dust.
Use of wetting agents for improved dust control has
also received a lot of attention in the past. Various
studies have shown improved respirable dust con-
trol compared to plain water, averaging about 25%
and ranging from zero (MRDE 1981, Chander et al.
1991) to 25%–30% (Kost et al. 1980) to more than
40% (Meets & Neethling 1987, Chugh et al. 2005a).
The differences in dust reduction performance may
be related to the type of wetting agent, type of coal,
dust particle size, dust concentration, water pH, and
water mineralogy (Hu et al. 1992, Kim and Tien 1994,
Kim 1995, Tien and Kim 1997). However, no univer-
sal methodology existed till recently that would allow a
mine operator to select and use wetting agents for dust
control (Kissel 2003). However, Chugh et al. (2004)
devised a rapid wettability test as a modification to the
“sink test” (Walker et al. 1952, Feldstein 1981) using
the specific coal and mine water to select the most
effective surfactants as well as determine the surfac-
tant dosage to be used. However before surfactant use
becomes widespread, issues related to the downstream
effects of the surfactant in cleaning plants need to be
overcome.
In light of this discussion on the need for well-
designed spray systems for dust control, it needs to
be pointed out that once the dust is airborne and
moves away from the face, the most effective dust
control mechanism in the face area is actually the wet-
scrubber. In this regard, Goodman et al. (2004) did
laboratory studies on a directional spray system (spray-
fan), fitted on the miner, to compare its effectiveness
in reducing dust and tracer gas concentrations as com-
pared to the wet scrubber. It was found that use of
sprayfan significantly increased both the operator and
return dust levels, apparently pushing the dust beyond
the scrubber inlets. The scrubber was also found to be
more effective in controlling respirable dust than the
sprayfan.

Figure 1. Sources of dust around a continuous miner,
(a) elevation view and (b) plan view.
3 NEW CONCEPTS AND HYPOTHESES
3.1 Sources of dust in and around a continuous
miner face area
An overview of sources of dust around a contin-
uous miner is shown in Figure 1. Some pertinent
observations are given below.
– High concentration of respirable dust occurs near
and along the roof level (Chugh et al. 2005b). Loca-
tion of boom sprays, type of sprays used and water
spray pressure may result in displacing dust–laden
air along the roof level, including roll back toward
the miner operator and batch haulage unit operator.
– Side sprays on the miner on the operator side tend
to contain the dust in the face area. However, due to
interaction between side sprays and boom sprays,
dust may again be pushed towards the roof level.
– Dust-laden air along the roof level is moving at a
high velocity based on water pressure used, seam
height, and the cutting rate. This air is not captured
by suction inlets of scrubbers.
– Most of the dust load in the scrubber is from the
bottom scrubber suction inlet over the material
transport conveyor.
– Even if the scrubber does an excellent job of wet-
ting the dust, the dust generated during the material
discharge from the conveyor into the haul unit sig-
nificantly increases dust concentration in the last
open crosscut (LOXC) return.
– Movement of batch haulage units around the face
area causes turbulence and further complicates dust
219
Copyright © 2006 Taylor & Francis Group plc, London, UK
propagation, intake air flow to the face area, and the
efficiency of dust control technologies.
3.2 Development of concepts for modified sprays
on a continuous miner
An extensive literature review, field observations on
dust control in the face area, and conceptual analyses
of the efficiency of dust suppression sprays employed
on the miner, led to identification of several avenues
for improving spray efficiency. Four (4) key concepts
were developed by the authors for improving the dust
suppression using hydraulic sprays on the continuous
miner. These can be summarized as follows.
1. Prevent dust generation, and minimize generated
dust from becoming airborne by increasing water
at the cutting bits, and cutting face, with appropriate
spatial distribution of sprays.
2. Wet and surround the airborne dust to allow the
scrubber to capture it.
3. Further wet the airborne dust escaping the scrubber
inlets area before it enters the area behind the miner
and the last open crosscut.
4. Maximize the residence time in the face area to
improve wetting of coal dust.
For each concept, spatial distribution and type of
sprays should be different depending upon expected
nature of dust. Due to intellectual property reasons,
detailed discussion on the concepts and specific spray
configurations utilized is curtailed here.
4 EXPERIMENTAL AND DATA ANALYSIS
PROCEDURES
4.1 Field experimental procedures
The modified spray system and baseline studies were
conducted on the same miner operating in a belt entry
development section at a Central Illinois mine. The
single miner section had three 20-ft wide entries with
200-ft × 100-ft center-to-center pillars heading east
and 300-ft × 125-ft staggered pillars heading south. A
map of the section indicating locations of the sampled
cuts is provided in Figure 2. The sampling program
was conducted over three rounds of testing spanning
over a 7-week time period. Sampling was conducted
for a total of 8 shifts as 25 cuts (11 conventional and
14 modified sprays) were sampled. Of these, six cuts
(three each for the conventional and modified system)
were cross-cuts, one was a straight initial cut and the
rest were cuts where the face was located 150+ feet
from the LOXC. The modified spray configuration uti-
lized approximately the same number of sprays and
volume of delivered water as the conventional spray
system. This ensured that the difference in dust capture
 Modified sprays
Cut sequence of baseline and modified Cut sequence
spray system.
Baseline sprays
Cut sequence

1
2
9
1
1 4
0
1
1 1
3
8
1
0
9
7

1
1
125 ft
8

125 ft
5

4 1 Entry #3 (Intake)
300 ft
5
3
7
5
1 2 Entry #2

6 2 3 Entry #1 (Return)

Figure 2. Cut sequence map of the section indicating the

sampled cuts in this study for the conventional and modified
spray system.

between the two spray systems was not confounded by

the impact of additional water. The number of sprays Figure 3. Typical cut encountered during this study. Sam-
was maintained at 29 and 28 for the modified and pling pump locations are indicated and marked 1, 2, 3, 4.
conventional system while the volumetric flow rate of
water was recorded to be the same at 25 gpm for both
Face area ventilation involved blowing system using
the systems.
a curtain. The plan requires 6,000 cfm of air at the end
of the line curtain with scrubber operating. The spray
system provided by Joy on the miner consists of 28
4.2 Description of mine
hollow cone sprays operating at 85 psi pressure.
The central Illinois mine mines about 5.6-ft thick No
5 coal seam at 350-ft depth. About 3.0 million tons of
run-of-mine coal is mined each year. Partial extraction
4.3 Data collection and analysis
room-and-pillar mining system is used with extraction
ratio ranging 50–55%. The pillar sizes typically are 4.3.1 Data collection
70-ft square (c-c) with 20-ft entry width. Joy 14 CM The data collection in this study utilized cut-by-cut
continuous miners are used in super-section layouts sampling. The pump locations for dust sampling
for mining. shown in Figure 3 were as follows:
The demonstration area in the mine was a three entry
1. Intake.
layout as shown in Figure 2. This anomalous layout
2. Near Miner Operator.
was to develop a new main belt entry for the mine.
3. Immediate Return. (∼80 ft. outby the face)
The pillar centers in this area are 300-ft long and 125-
4. Return.
ft wide. The proposed entry height in this area was
7-ft but due to poor roof condition, mining height was The pumps were hung from the roof at all locations
typically 10-ft. as shown in Figure 4.

220

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 4. Gravimetric dust sampling and instantaneous dust
measurements using Personal Data Rams.
Other measurements taken for all the cuts included
the following:
• Line Curtain (LC) Air.
• LOXC Air.
• Car Count.
• Cut Dimensions. (Height, Width, Depth)
• Cut cross-section. (Roof, Seam, Floor)
• Cut Time.
• Water Pressure and Water Volume.
4.3.2 Data transformations
The filter cassettes from the dust pumps were weighed
by Mine Safety and Health Administration (MSHA).
The weight gain of the cassettes and the pump opera-
tion time along with the MRE equivalent correction
multiplier of 1.38 were used to compute the raw
dust concentrations at the four sampled locations.
These dust measurements were then transformed to
isolate the impacts of cut-time variability (down-
time during cuts), varying LC and LOXC air flows
and the differences in intake dust concentrations. The
221
Copyright © 2006 Taylor & Francis Group plc, London, UK
transformations were accomplished using the follow-
ing methodology.
• To correct for different intake air dust concentra-
tions, the Location 1 measurement was subtracted
from the measurements for the other three locations.
• The cut-times for cuts of same dimensions may dif-
fer due to haulage delays as well as unexpected
delays encountered during the cuts. These down-
times (delays) were estimated as the difference
between cut time (pump run time) less average time
for mining the particular cut volume. The Location
2, 3, and 4 dust concentration numbers were then
adjusted for this down-time for changing the divisor
time for the dust data from the actual pump run time
to the expected run time. In effect, this adjustment
accounts for varying production.
• Varying ventilation air flows to the face and in the
LOXC directly impact the measured dust concentra-
tions as the generated dust is either diluted by excess
air or concentrated at lesser airflow. To correct for
these variations, all the measured dust concentra-
tions were normalized for 7,500 cfm of LC air and
34,000 cfm of LOXC air. The LC normalization was
applied to the Location 2 and 3 samples while the
LOXC normalization was applied to the Location
4 samples. It is recognized that due to recircula-
tion of air in the face, the applied corrections for
Locations 2 and 3 (particularly Location 2) are not
truly indicative of the conditions near the face. Loca-
tion 4 readings are however largely unaffected by
recirculation at the face.
It is also to be noted that all data presented here is
transformed cut-by-cut data and hence is not indicative
of full-shift dust exposures to occupations.
5 RESULTS AND DISCUSSION
To minimize the impact of variability in mining con-
ditions on the measured dust concentrations, appro-
priate corrections for intake dust concentration, air
flows, production and haulage and unexpected delays
were used. The nuisance variables which were not
corrected by mathematical transformations included
varying mining height and cut roof height, types of
cuts and different testing periods. However, appropri-
ate statistical analyses were performed to isolate the
confounding effects of these nuisance variables. This
was accomplished by comparing the mining height
and cut roof height encountered during testing of the
two spray systems which revealed that on an aver-
age, these two variables were almost identical (Table
1). Since a positive correlation (at 99.5% confidence)
level was observed between respirable dust and mining
height (Table 2), this factor was included as a covariate
in the Analysis of Covariance (ANCOVA) procedure.
 Table 1. Average mining conditions (cut roof and mining 14
Conventional Sprays
height) during the testing program. 12 Modified Sprays

Dust Concentration (mg/m3)

Cut Mining Cut roof 10

Sprays sampled height (inches) (inches) 8

6
Conventional 11 96.96 23.45
Modified 14 96.93 24.35 4

Immediate

Immediate

Immediate
Operator

Operator

Operator
Table 2. Effect of mining height and cut rock height on res-

Return

Return

Return

Return

Return

Return
Miner

Miner

Miner
pirable dust concentrations in the face area. Reported results
are Partial Correlations controlling for Testing Round, Spray Round I Round II Round III Round I Round II Round III Round I Round II Round III
System and Sampling Location. Testing Round - Sampling Location

Mined Figure 5. Results of the modified spray systems compared

height Cut roof Dust to the conventional spray system over three rounds of testing.

Mined Correlation 1 0.816 0.330

height Significance – 0.000 0.005 systems (conventional and modified) were signifi-
df 0 69 69
cantly different. Further evaluation of the data compar-
Cut Correlation 0.816 1 0.332 ing the average dust levels and the standard deviations
roof Significance 0.000 − 0.005 for different locations and over different blocks for the
df 69 0 69
two spray systems supported the finding that the mod-
Dust Correlation 0.330 0.332 1 ified spray system performed significantly better in
Significance 0.005 0.005 – controlling respirable dust at all three sampled loca-
df 69 69 0
tions, viz. near miner operator, immediate return and
return (Figure 5). The reductions ranged from 4% to
80%. Most typical dust reductions were in the 40–50%
Table 3. ANCOVA for testing the significance of treatments range (Table 4).
(one of which is the spray system used). Dependent variable: Since only a limited amount of data was collected,
respirable dust concentration. the variances in measurements were high as is typical
for field testing in underground mining conditions.
Type III However, the variances encountered in Round II of
sum of Mean testing were low, indicating more controlled conditions
Source squares df square F Sig.
encountered during this phase of testing. The low vari-
Corrected 1054.5∗ 6 175.7 20.2 0.000
ances allowed statistical comparisons to be made for
Model this period of testing. A “t-test” comparing the means
Intercept 2.4 1 2.4 0.3 0.598 of dust concentrations revealed that for the miner
Testing Phase 87.3 2 43.6 5.0 0.009 operator and immediate return sampling locations, the
Spray System 75.5 1 75.5 8.7 0.004 modified spray system provided better dust control at
Pump Location 685.6 2 342.8 39.4 0.000 a statistically significant 99.7 and 99.0% confidence
Mining Height 31.5 1 31.5 3.6 0.061 levels, respectively. It could also be concluded with
Error 583.0 67 8.7 90% confidence that the improvements were at least
Total 3297.7 74 33.6% and 26.9%, respectively (Table 5).
Corrected Total 1637.4 73
Though there was also an approximately 50% reduc-
∗
R Squared = .644 (Adjusted R Squared = .612). tion in dust at the return location, this could not be
established with statistical confidence. The reason was
that the variance encountered during return sampling
To isolate the uncontrolled effect of testing over three of the conventional system was high. Also, there was
discontinuous time periods, the data was blocked in concern in statistical comparisons for the return loca-
three groups for statistical analyses. This blocking also tions since the LOXC air flow was very high resulting
eliminated the impact of cut-types as in one of the time in dilution of the dust generated at the face to very low
periods, only X-cuts were sampled while in the other levels. Though the data was corrected for intake dust
two time periods only deep cuts were made. concentrations, the return sampling location is very
The ANCOVA procedure revealed that the “Spray sensitive to intake dust concentrations in this case
System” variable was significant at 99.6% level which may have caused the aberration in variance
(Table 3). This implies that the two tested spray during the conventional spray testing in Round II.

222

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 4. Summary respirable dust concentration measurements (in mg/m3 ) for conventional and modified spray systems
sampling. Presented data is transformed cut-by-cut data and is not indicative of a full-shift sample.

Conventional Modified
Reduction
Average SD N Average SD N (%)

Round I
Miner operator 6.47 2.79 4 5.22 5.29 2 19
Immediate return 10.5 4.23 3 5.90 5.99 2 44
Return 0.64 0.52 4 0.12 0.17 2 81
Round II
Miner operator 4.35 0.37 3 2.07 0.66 4 52
Immediate return 3.09 0.53 3 1.44 0.53 4 53
Return 1.04 1.12 3 0.52 0.27 4 49
Round III
Miner operator 10.5 6.81 4 4.18 1.64 8 60
Immediate return 13.3 3.14 4 10.7 1.78 8 20
Return 1.06 0.53 4 1.01 0.48 8 4

Table 5. Comparison of differences between the conven- REFERENCES

tional and modified spray systems during Round II of
testing. Bhaskar R., Gong R. & Jankowski R.A. 1991. Studies on
Under-boom Dust Control to reduce Operator Exposure
H0: Difference Percentage to Dust. Proceedings, 5th U.S. ventilation Symposium,
Conventional = (mg/m3 ) reduction Morgantown, WV, pp. 197–206.
Modified at 90% at 90% Bigu J. & Grenier M.G. 1989. Reduction of airborne radioac-
p-value confidence confidence tive dust by means of a charged water spray. AIHA J
50:336–345.
Miner 0.003 1.461 33.6% Chander S., Alaboyun A.R. & Aplan F.F. 1991. On the mech-
operator anism of capture of coal dust particles by sprays. In:
Immediate 0.01 0.831 26.9% Proceedings of the Third Symposium on Respirable Dust
return in the Mineral Industries (Pittsburgh, PA, October 17–19,
Return 0.52 – – 1990). Littleton, CO: Society for Mining, Metallurgy, and
Exploration, Inc.
Chugh Y.P. & Patwardhan A. 2005. A Field Demonstration
of a JOY Wet-Head Miner in Illinois. Interim Report,
6 CONCLUDING REMARKS Project No. 04-1/US-2, Illinois Clean Coal Institute,
Carterville, IL.
Based on field dust sampling studies on the con- Chugh Y.P., Alam M.M., Patwardhan A. & Thatavarthy
ventional and modified spray system and subsequent K.K. 2005a. A Field Demonstration of a Modified Wet
statistical analyses of the data, the authors believe Scrubber for Dust Control in an Illinois Coal Mine. Pro-
ceedings, 8th International Mine Ventilation Congress,
that the modified spray system is significantly bet-
July 6–8, Brisbane, Australia, Editor: A.D.S. Gillies.
ter than the conventional sprays on the miner. SIU Publisher: The Australasian Institute of Mining and Met-
researchers also believe that the performance of the allurgy, Carlton Victoria, Australia, ISBN 1-920806-33-4,
modified spray system can be further improved by pp. 143–150.
varying water volume and water pressure for sprays Chugh Y.P., Patwardhan A., Gurley H. & Alam M.M. 2005b.
associated with different hypotheses outlined above. Advanced Mining Technologies to Reduce Underground
The performance should also improve with improved Coal Production Costs. Final Report, Project No. 04-
ventilation conditions at the face. 1/1.1A-2, Illinois Clean Coal Institute, Carterville, IL.
Chugh Y.P., Thatavarthy K.K. & Patwardhan A. 2004. Devel-
opment of a Wettability Test for Assessing Effectiveness of
Surfactants for Dust Control. Tenth U.S./North American
ACKNOWLEDGEMENTS
Mine Ventilation Symposium, May 16–19, Anchorage,
Alaska.
In addition to the funding support by the Illinois Clean Cheng L. & Zukovich P.P. 1973. Respirable dust adhering to
Coal Institute, the authors greatly appreciate the sup- run-of-face bituminous coals. Pittsburgh, PA: U.S. Depart-
port of cooperating mines staff for their exceptional ment of the Interior, Bureau of Mines, RI 7765. NTIS No.
support during this study. PB 221–883.

223

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Colinet J.F., Spencer E.R. & Jankowski R.A. 1997. Sta-
tus of dust control technology on U.S. longwalls. In:
Ramani RV, ed. Proceedings of the Sixth International
Mine Ventilation Congress. Littleton, CO: Society for
Mining, Metallurgy, and Exploration, Inc., pp. 345–351.
Courtney W.G. & Cheng L. 1977. Control of respirable dust
by improved water sprays. In: Respirable Dust Control –
Proceedings of Technology Transfer Seminars, Pitts-
burgh, PA, and St. Louis, MO, IC 8753, pp. 92–108. NTIS
No. PB 272 910.
Courtney W.G., Jayaraman N.I. & Behum P. 1978. Effect of
water sprays for respirable dust suppression with research
continuous mining machine. U.S Department of Interior,
Bureau of Mines, Pittsburgh, P.A, NITS No. PB 279,881.
Feldstein N. 1981. Surface Chemical Technology for
Improved Wetting of Coal Dust, U.S. Bureau of Mines,
Report OFR-127-81, 47 pp.
Goodman G.V.R. & Jankowski R.A. 1998. Optimal Spray
System Designs for Continuous Miner Dust Control.
Advances in the Prevention of Occupational Respiratory
Diseases, July, 1–9.
Goodman G.V.R. 2000. Using water sprays to improve perfor-
mance of a flooded-bed dust scrubber. Applied Occupa-
tional and Environmental Hygiene, vol. 15(7), 550–560.
Goodman G.V.R., Pollock D.E. & Beck T.W. 2004. A com-
parison of a directional spray system and a flooded-bed
scrubber for controlling respirable dust exposures and face
gas concentration. In R. Ganguli and S. Bandopadhyay
(Eds.), Proceedings of the 10th US/North American Mine
Ventilation Symposium (pp 241–248). Anchorage, Alaska:
A.A Balkema Publishers.
Hamilton R.J. & Knight G. 1957. Laboratory experiments
on dust suppression with broken coal. London, U.K.:
National Coal Board, Mining Research Establishment
Report No. 2083.
Hu Q., Polat H. & Chander S. 1992. Effect of surfactants
in dust control by water sprays. In: Proceedings of the
Symposium on Emerging Process Technologies for a
Cleaner Environment. Littleton, CO: Society for Mining,
Metallurgy, and Exploration, Inc., pp. 269–276.
Jankowski R.A. & Organiscak J.A. 1983. Dust sources and
controls on the six U.S. longwall faces having the most
difficulty complying with dust standards. Pittsburgh, PA:
U.S. Department of the Interior, IC 8957. NTIS No.
PB84-142058.
Jankowski R.A., Jayaraman N.I. & Babbitt C.A. 1987. Water
Spray System for Reducing Quartz Dust Exposure of
the continuous Miner Operator. Proceedings of the 3rd
U.S. Mine Ventilation Symposium, Pennsylvania State
University, pp. 605–611.
Jayaraman N., Kissel F.N. & Schroder W.E. 1984. Modify
Spray Heads to Reduce Dust Rollback on Miners. Coal
Age, June 1984.
Jayaraman N.I. & Jankowski R.A. 1988. Atomization of
water sprays for quartz dust control. Appl Ind Hyg
3:327–331.
Kim J. 1995. Respirable Coal Dust Control Using Surfac-
tants – With Special Emphasis on a Liquid Spray System.
University of Missouri – Rolla Dissertation.
Kim J. & Tien J.C. 1994. Effect of added base on coal wet-
ting ability of nonionic surfactant solutions used for dust
control. Min Eng (London) 154(399):151–155.
224
Copyright © 2006 Taylor & Francis Group plc, London, UK
Kissel F.N. 2003. Handbook for Dust control in Mining.
National Institution for Occupational Safety and Health,
Information Circulation (IC 9465), 2003, 131 p.
Kost J.A., Shirey G.A. & Ford C.T. 1980. In-mine tests for
wetting agent effectiveness. Bituminous Coal Research,
Inc. U.S. Bureau of Mines contract No. J0295041. NTIS
No. PB82- 183344/ XAB.
Matta J.E. 1976. Effect of location and type of water sprays
for respirable dust suppression on a continuous mining
machine. U.S Department of Interior, Bureau of Mines,
Pittsburgh, P.A, NITS No. 254,677, 11 p.
McCoy J., Melcher J., Valentine J., Monaghan D., Muldoon T.
& Kelly J. 1983. Evaluation of charged water sprays
for dust control. Waltham, MA: Foster-Miller, Inc. U.S.
Bureau of Mines contract No. H0212012. NTIS No.
PB83-210476.
Meets E.J. & Neethling A.F. 1987. Some experiences in the
use of wetting agents to suppress dust at Sigma colliery.
J Mine Vent Soc S Afr Oct:126–133.
MRDE. 1981. Methods of reducing dust formation and
improving dust suppression on longwall faces: final
report on ECSC research project 7256-12/003/08. Mining
Research and Development Establishment (U.K.).
Mukherjee S.K. & Singh M.M. 1984. New techniques for
spraying dust. Coal Age June:54–56.
Page S.J., Mal T. & Volkwein J.C. 1994, Unique Water Spray
System Improves Exhaust Face Ventilation and reduces
Exposure to Respirable Dust on Continuous Miners in
Laboratory Tests, Journal of Mine Ventilation Society of
South Africa, October, pp. 196–205.
Ruggieri S.K. & Jankowski R.A. 1983. Fundamental
approaches to longwall dust control. In: Proceedings of
the Symposium on Control of Respirable Dust (Beckley,
WV, Oct. 4–8, 1983).
Thimons E.D., Taylor C.D. & Zimmer J.A. 1999. Ventilating
the box cut of a two-pass 40 ft extended cut. Journal of the
Mine ventilation Society of South Africa, July/September,
1999.
Tien J.C. & Kim J. 1997. Respirable coal dust control using
surfactants. Apppl.Occup.Environ.Hyg., December 1997,
12(12), 957–963.
U.S. Bureau of Mines. 1981. Technology News 118: Lower
dust exposure of longwall shearer operator by relocating
the machine cooling water sprays. Pittsburgh, PA: U.S.
Department of the Interior, Bureau of Mines.
Volkwein J.C., Cecala A.B. & Thimons E.D. 1983. Use of
foam for dust control in minerals processing. Pittsburgh,
PA: U.S. Department of the Interior, Bureau of Mines, RI
8808. NTIS No. PB84-131184.
Walker P.L., Peterson E.E. & Wright C.C. 1952. Surface
Active Agent Phenomena in Dust Abatement. Ind. And
Eng. Chem., Vol. 44, No. 10, Oct., pp. 2389–2393.
Zimmer R.A., Lueck S.R. & Page S.J. 1987. Optimization
of overburden drill dust control systems on surface coal
mines. Int J Surf Min 1:155–157.
Zyl F.J.van & Eroglu H.N. 2004. Operational compari-
son between a wet-head and directional spray dust-
suppression systems on a continuous miner at Kriel
Colliery. Report No. 2004-0203, CSIR MiningTek.
Zyl F.J.van. 2005. Operational evaluation of JOY’s wet-head
system at Goedehoop Colliery with regard to respirable
dust control. Report No. 2005-0327, CSIR MiningTek.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Dust control on longwalls – assessment of the state-of-the-art

J.P. Rider & J.F. Colinet

National Institute for Occupational Safety and Health, Pittsburgh, Pennsylvania, United States

ABSTRACT: Operating practices and resulting production from longwall faces are continually improving.
Consequently, longwall operators are continually seeking to improve the dust control capabilities on longwalls.
The National Institute for Occupational Safety and Health (NIOSH) has been conducting dust surveys at longwalls
throughout the country in an effort to document the types of controls that are being used and to quantify the
relative effectiveness of these controls. To date, surveys have been completed at longwalls in Alabama, Colorado,
Pennsylvania, Utah, and West Virginia. In addition NIOSH has been reviewing longwall dust parameters provided
by mine operators to MSHA to assess the status of dust control technology being used throughout the industry.
Results from the underground dust surveys and information obtained from MSHA will be summarized in this
paper to provide an update on longwall dust control technology and practices.

1 INTRODUCTION compliance dust samples, respectively. Analysis of

Improvements in longwall mining equipment and min-
ing practices have led to substantial gains in long-
wall production levels, resulting in longwall mining
accounting for 51% of the coal produced underground
in the United States. Overall production from U.S.
longwall mines in 2004 was over 189.5 million tons.
The average output from U.S. longwall mines has
increased approximately 115% over the last 15 years,
from 1.9 to 4.12 million tons. Longwall panels aver-
aged 288 m (940 ft) wide and nearly 3062 m (10,000 ft)
long in 2002, compared to an average of 230 m (750 ft)
wide and 2135 m (7000 ft) long in 1994. New longwall
panels, currently in the planning stage, may be as wide
as 457.2 m (1500 ft) and as long as 4572 m (15,000 ft).
Cutting heights in 2002 averaged 215 cm (86 in) and
ranged between 121.9 and 396.2 cm (48 and 156 in)
(Fiscor, 2003).
Although significant gains in longwall dust control
have been made, they have been challenged by sub-
stantial increases in coal extraction rates, which result
in the potential to generate more dust (Webster et al.,
1990). Average production during compliance sam-
pling by mine operators has increased from an average
of 2800 tons per shift in 1990 to an average 5600 tons
per shift in 2004. Consequently, longwall operations
continue to have difficulty in maintaining consistent
compliance with the federal dust standard of 2 mg/m3 .
During a five-year span from 2000 through 2004,
mine operators and Mine Safety and Health Adminis-
tration (MSHA) inspectors collected 7421, and 1587
these samples show that 14% of the mine operator
samples and 15% of the MSHA samples were equal to
or exceeded 2.1 mg/m3 , the level at which a citation is
issued (Niewiadomski, 2004).
Health surveillance data also indicate that overex-
posure to respirable dust in underground coal mines
continues to lead to the development of lung disease.
Results from the most recent full round (1992–1996)
of the Coal Worker’s X-ray Surveillance Program indi-
cate that approximately 8% of the examined miners
with at least 25 years of mining experience were
diagnosed with Coal Worker Pneumoconiosis (CWP)
(category 1/0+). The continued development of CWP
in underground coal mine workers and the magnitude
of respirable dust over exposures in longwall min-
ing occupations illustrate the need for improved dust
control technology on longwalls.
Ventilating air and water sprays are the primary
means used to control dust and methane in longwall
operations. This paper describes the on-going research
effort to identify the current status of longwall min-
ing and dust control technology being applied in the
industry.
2 LONGWALL DUST PARAMETERS
Longwall dust parameters provided to MSHA as part
of the mine’s ventilation plan were reviewed to assess
the status of control measures being used through-
out the industry. Ventilation plans between the years

225

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 1. Range of velocity readings.
2000 to 2003, representing 44 longwall panels, were
analyzed to identify and quantify the different types
of longwall control features used to limit the miners’
exposure to respirable dust. A variety of dust control
measures were encountered when the dust parameters
were reviewed. Each longwall operator had a different
and unique approach to control respirable dust levels.
Figure 1 shows the range of minimum mean air
velocities required at the headgate and tailgate. The
average minimum headgate velocity was approxi-
mately 2.2 m/sec (430 ft/min) for all longwall opera-
tions. Eastern mines averaged 2.2 m/sec (440 ft/min),
while western operations averaged 2.0 m/sec (390 ft/
min). The average minimum volume of air reported
in the dust plans was approximately 24.5 m3 /sec
(52,000 ft3 /min). As expected, minimum air quanti-
ties were significantly higher for western operations,
32.1 m3 /sec (68,000 ft3 /min), compared to 21.1 m3 /sec
(45,000 ft3 /min) in the east. Approximately 28% of the
mine operators reported the use of beltway air as a
means to supplement the total volume of air reaching
the face. Nearly one-half of the longwall operators in
eastern U.S. mines utilize air deflectors on the first 3
or 4 shields as a means of helping move the air onto
the face and controlling dust levels when the shearer
was cutting out at the headgate.
The quantity, type, and pressure of the shearer cut-
ting drum sprays reported by longwall operators varied
with each longwall operation. The number of sprays
ranged between a low of 30 and a high of 62. Accord-
ing to the ventilation plans, between 75 and 90% of the
drum sprays are required to be operational when min-
ing. The type of spray preferred by the majority of the
operators was the Conflow 2801 with either full-cone,
flat-fan, or hollow-cone spray patterns. The range
and frequency of drum spray pressures are shown in
figure 2. Sixty-one percent of the longwalls reported
drum spray pressures between 620.5 and 758.4 kPa (90
and 110 psi). The average drum spray pressure was
226
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 2. Pressure ranges for shearer drums sprays and
splitter arm sprays.
620.5 kPa (90 psi). Thirteen operations reported using
crescent sprays on the ranging arms as a means of con-
trolling fugitive dust. The number of crescent sprays
ranged between 4 and 11, with an average spray pres-
sure of 586.1 kPa (85 psi), and varied between 448.2
and 689.47 kPa (65 and 100 psi).
The ventilation plans showed that 39 out of 44 oper-
ations utilized a headgate splitter arm as a means of
either suppressing or moving fugitive dust away from
the personnel operating the shearer. The length of the
splitter arm, type of sprays, distance between sprays,
angle of the sprays, and spray pressure varied greatly
between mines. The number of sprays mounted on the
splitter arm ranged between 2 and 20. The majority
of the ventilation plans stated that at least 90% of the
splitter arm sprays are required to be operational when
mining.Approximately 15% of the operations reported
utilizing an extension of the splitter arm that angled
between 30 and 45 degrees toward the face. A variety
of spray patterns were used, including full cone, hollow
cone, vee-jet and flat fan. Approximately 30% of the
operators reported using venturi sprays. Splitter arm
sprays were angled 20 to 30 degrees up or down and
directed toward the shield tips or pan line. The angle
of the sprays directed with the airflow ranged between
10 and 45 degrees. The majority of the splitter arm
sprays were aimed at the headgate drum, especially
the sprays positioned on the splitter arm upwind of the
cutting drum. Headgate splitter arm spray pressure var-
ied greatly, as shown in figure 2. Sixty-eight percent
of the sprays operated with spray pressure at or above
689.5 kPa (100 psi).
Manifolds located on the tailgate side of the shearer
or tailgate splitter arms were reported on nearly all
the longwall operations. The majority of the western
operations used the splitter arm technology to direct
fugitive dust downwind, while most eastern opera-
tions utilized spray manifolds to suppress or move
dust generated by the tailgate drum. The number of
 sprays on the tailgate splitter arm ranged between 3
and 10, while the spray pressure varied between 551.6
and 689.5 kPa (80 and 100 psi). The sprays were either
directed at the tailgate drum or angled from 10 to 45
degrees in the direction of the airflow. Nearly 60% of
the operations utilized a manifold located on the tail-
gate side of the shearer or in the lump breaker area.
Manifolds positioned on the tailgate side of the shearer
directed the spray plume downwind. Spray pressures
ranged between 344.8 and 861.8 kPa (50 and 125 psi).
All but 4 operations operated with spray pressure
below 689.5 kPa (100 psi), with the average pressure
of approximately 551 kPa (80 psi). Lump breaker man-
ifolds consisted of 4 to 6 sprays with pressures ranging
between 620.5 and 827.4 kPa (90 and 120 psi) and
directed toward the pan line.
Thirty-five operations reported using sprays on the
top deck or on the face side of the shearer to confine
and move dust away from the mine operators. Nineteen
of the thirty-five operations utilized between 2 and 6
spray manifolds spaced evenly across the shearer. Each
manifold was equipped with a range of 2 to 16 sprays.
Shearer body spray configurations consisting of 3
spray manifolds with 3 or 4 sprays per manifold were
the most commonly reported spray system and were
utilized by nearly 25% of operators. One out of five
operations used between 4 and 6 venturi sprays spaced
evenly and positioned either at the shearer centerline
or the gob side of the shearer. Shearer spray pressure
ranged between 482.6 and 1034.2 kPa (70 and 150 psi).
Approximately 40% of the shearer sprays operated
with pressures at or above 689.5 kPa (100 psi).
3 LONGWALL SURVEYS
3.1 Sampling methodology
Gravimetric dust samplers, identical to those used in
compliance sampling, were operated at 2L/min in con-
junction with a 10-mm Dorr-Oliver nylon cyclone.
Samplers were utilized at stationary and mobile sam-
pling locations to quantify the levels of respirable dust
generated at prominent sources along the longwall
face. Gravimetric sampling was conducted on-section
for 4 to 6 hours, and mine personnel did not wear any
of the samplers. Calculated concentrations were not
converted to Mining Research Establishment (MRE)
equivalent dust levels and should not be compared to
compliance sampling concentrations.
Personal DataRAMS (PDRs) were used adjacent to
the gravimetric samplers at select sampling locations
to obtain a time-related profile of dust levels generated
during each sampling period. The PDR is an MSHA-
approved, instantaneous dust measuring device where
dust-laden air passes through a sampling chamber and
a light source. The amount of light deflection in the
227
Copyright © 2006 Taylor & Francis Group plc, London, UK
chamber is measured and provides a relative measure
of the dust concentration. The PDR samples were uti-
lized in an active sampling mode where dust laden air
was pulled through a 10-mm Dorr-Oliver cyclone at
a flow rate of 2 L/min. Instantaneous dust levels were
stored at 10 second intervals in an internal data log-
ger and then downloaded onto a computer for analysis.
Dust levels measured with the PDR can be calculated
for any time period of interest (e.g., head-to-tail or
tail-to-head passes).
Mobile dust sampling to determine the amount of
dust generated by the shearer and by movement of
advancing shields was conducted by a three-member
NIOSH sampling team. Ideally, the UPWIND sam-
pling location was approximately 5.2–7.6 m (25–50 ft)
upwind of the headgate cutting drum and measured
intake dust levels reaching the shearer. The SHEARER
sampling location was located between mid-shearer
and the tailgate end of the shearer. The sampling crew
member was usually positioned within a shield or two
of the tailgate shearer operator. This data provided an
indication of the amount of dust that has migrated from
the face between the cutting drums. When possible,
the DOWNWIND sampling location was approxi-
mately 5.2–7.6 m (25–50 ft) downwind of the tailgate
drum. Due to shield movement patterns, this sampling
location had to be adjusted at certain mines. Each
team member maintained their relative position with
the shearer as it moved across the face. The difference
in dust levels between the UPWIND and DOWN-
WIND locations was dust generation attributed to the
shearer. In addition, similar mobile sampling was con-
ducted upwind and downwind of shield movement on
selected head-to-tail passes to isolate dust liberated
during shield advance.
At each of these mobile sampling positions, each
sampling crew member wore a sampling vest that con-
tained two gravimetric pumps and four cyclones with
appropriate filter cassettes. Two of the four cyclones
[attached to the sampling vest on the left and right sides
of the chest area] were connected to the gravimetric
pumps and used to sample dust levels during head-
to-tail passes. The other two cyclones, also attached
to the sampling vest on the left and right sides of the
chest area, were used for sampling tail-to-head passes.
If the shearer was stopped for an extended period (>3
minutes), the gravimetric pumps were paused so that
mobile sampling along the face was representative of
dust levels during active mining. Along with the gravi-
metric sampling package, members of the sampling
crew carried a PDR sampler.
Mobile sampling was augmented with stationary
sampling packages. At each stationary sampling loca-
tion, two gravimetric samplers were located adjacent
to one another and operated over the same sam-
pling period. Stationary sampling locations included
the intake, belt entry, shield 10, and approximately
 10 shields from the tailgate. Intake samplers were
typically located in the last open crosscut and used to
isolate the dust contamination from sources outby the
longwall face. If the mine was utilizing the belt entry
as an additional intake, gravimetric samplers were
located outby the last open crosscut and the stage
loader to monitor dust levels liberated in the belt entry.
The shield 10 samplers were used to monitor the dust
concentrations of air coming onto the face. The dif-
ference between dust levels measured at shield 10 and
the intake and belt sources represent an estimate of
dust liberated by the stage-loader/crusher dust source.
The tailgate sampling package provided an indica-
tion of the total dust generated along the face. The
samplers were typically started after arrival upon the
longwall face and operated continuously until sam-
pling was completed. PDR samplers were also placed
at the shield 10 and tailgate sampling locations.
In addition to dust measurements, sampling per-
sonnel monitored airflow and water quantities on the
longwall section. During each shift of sampling,
spot air velocity readings were taken with hand-held
anemometers at 10-shield intervals down the face.
These measurements were one-minute readings taken
approximately one foot above the spill plate of the
face conveyor. Also, a rough estimate of the area at
each velocity sampling location was calculated to esti-
mate the air quantity present. If possible, water flow
meters were installed in the water line supplying the
shearer and the line supplying the stageloader/crusher
sprays. Periodic readings were taken from each of these
meters to monitor the quantity of water being used to
suppress dust.
3.2 Survey results
A variety of operating conditions were encountered
when surveys were conducted at the different mines
throughout the country. Six of the eight mines sur-
veyed to date utilized a bidirectional cutting sequence,
with two mines employing unidirectional cutting. Face
panel widths ranged between 237.7 and 304.8 m (780
and 1000 ft), and the average face width was 272.2 m
(893 ft). Comparing panel widths observed in this lat-
est series of surveys with panel widths reported in
a 1995 longwall study (Colinet et al., 1995) showed
close to a 25% increase. The average cutting height
was 2.7 m (9 ft) with a range between 2.1 and 3.4 m
(7 and 11 ft).
During the surveys, daily spot velocity readings
were recorded approximately every 10 shields along
the longwall face and were used to calculate the
average face velocity. The average velocity of the sur-
veyed longwalls was 3.4 m/sec (665 ft/min), which
represents an increase of over 0.8 m/sec (50 ft/min)
when compared to average air velocities reported in
228
Copyright © 2006 Taylor & Francis Group plc, London, UK
the 1995 longwall study. Seven of eight longwalls
had average air velocities greater then 3.0 m/sec
(600 ft/min), with two mines averaging over 4.1 m/sec
(800 ft/min). Along with air velocity calculation, a
rough estimate of the area under the shields at each
velocity sampling location was used to calculate aver-
age air quantity for each face. The average volume
of air moving down the face was approximately
31.6 m3 /sec (67,000 ft3 /min), with a range between
24.3 to 39.1 m3 /sec (51,000 to 83,000 ft3 /min). Air
quantity observed for six of the eight longwalls was
greater than 30.2 m3 /sec (64,000 ft3 /min). Average air
quantities calculated in this current series of longwall
surveys increased approximately 65% when compared
to the mid-1990 longwall study.
Along with an escalation of air down the face, the
use of water to the shearer has increased in an effort
to control dust liberating from the face. The average
water usage at the shearer observed during the eight
most recently completed surveys was 492.0 L/min
(130 gpm). The number of shearer drum sprays ranged
between 35 and 62, and the average drum spray pres-
sure was approximately 1034.2 kPa (150 psi). In the
1995 study, the average water usage at the shearer was
378.5 L/min (100 gpm) with an average drum spray
pressure of 965.3 kPa (140 psi). One-half of the sur-
veyed mines utilized six to eight crescent sprays on
the ranging arms.
Studies (O’Green et al., 1994) have shown that
shearer drum water sprays are very effective at mini-
mizing dust generated at the point of coal fracture but
can increase airborne respirable dust levels if operated
at overly high water pressure. Instead of suppressing
dust, the drum sprays may force the dust out away from
the cutting drum causing excessive turbulence around
the cutting drums. The increased turbulence leads to
higher dust levels at the shearer and downwind. Larger
orifice sprays could be utilized to increase water vol-
ume and reduce operating pressure. Increased flow
to the drum-mounted sprays should result in a more
uniform wetting of the coal product, which may also
reduce conveyor-generated dust.
External spray systems with air-moving directional
sprays were observed at the headgate side of the
shearer on seven of the eight surveyed longwalls. The
splitter arm technology used to keep fugitive dust away
from the shearer operators was unique to each min-
ing operation. The lengths varied between 2.7 and
4.6 m (9 and 15 ft) with the number of sprays ranging
between 6 to 19. Spray spacing on the splitter arm and
the directional angle of the sprays varied greatly at each
operation. Extension arms attached to the end of split-
ter arms and angled between 30 and 45 degrees toward
the face were utilized at three mines. Venturi sprays
located on top of the splitter arm were observed at two
operations. Splitter arm spray pressures were approxi-
mately 689.5 kPa (100 psi) when hollow cones sprays

Figure 3. Range of dust concentrations measured at the
stationary sampling locations.
were utilized and in excess of 1551.3 kPa (225 psi)
when venturi sprays were operated.
A variety of spray configurations were employed
on the body of the shearers during the mine surveys.
Deflector or sloughing plates were observed at three
western mine sites. One of the three operations used
a single plate spanning the length of the shearer with
6 venturi sprays evenly spaced across the top of the
deflector plate. Deflector plates were split into three
independent sections on two of the longwall opera-
tions. Each section had five hollow cone sprays evenly
spaced across the top of the plate. At eastern mine sites
a series of spray manifolds were used the majority of
the time. Three or four manifold consisting of four or
five sprays were evenly spaced across the length of the
shearer. The manifold were either located on the face
side of the shearer or on the top of the shearer close
to the face. One mining operational moved the spray
manifolds toward the middle of the shearer and ele-
vated the manifolds 15.24 to 30.48 cm (6 to 12 in) off
the shearer body.
The majority of the surveyed mines made use of
manifolds located above the lump breaker or on the
shearer body to control dust in the tailgate drum area.
A minimum of 4 and maximum of 16 sprays were
directed toward the cutting drum or down onto the
conveyor. One mining operation utilized a series of
higher pressure sprays and directed them downwind
just inside the spill plate. The sprays formed a water
curtain that was effective at keeping fugitive dust out
of the walkway. A tailgate splitter arm was observed
at one mine.
The minimum, average and maximum dust levels
for stationary gravimetric sampling locations are
shown in figure 3. Intake dust levels were consis-
tently low indicating very little dust contamination
was occurring from outby sources. The maximum dust
level measured was 0.34 mg/m3 with the majority of
229
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 4. Range of mobile sampling measurements.
the operations below 0.20 mg/m3 . Half of the long-
walls surveyed used belt air as a supplementary air
source. For these longwalls, the average dust level
in the belt entry was 0.41 mg/m3 , while the average
intake concentration was 0.18 mg/m3 , approximately
2.5 times less than the belt entry dust concentrations.
Although the average dust levels in the belt entry lev-
els are relatively low on average, the belt entry has the
potential to add to face dust levels. However, according
to past research studies (Potts and Jankowski, 1992)
potential increases in face dust levels seemed to be
negated by the potential for increased dilution with
the additional air reaching the face.
The dust level monitored at shield 10 is a good
indication of the dust entering the face from the
stageloader/crusher along with intake and belt outby
sources. Average dust concentration found at shield 10
was 0.67 mg/m3 . The difference between intake/belt
dust levels and shield 10 concentrations is primarily
the dust contributed by the stageloader/crusher. On
average, the amount of dust that was attributed to
the stageloader/crusher was 0.48 mg/m3 . Dust levels
measured at the tailgate sampling location provided
a good indication of the amount of total dust gener-
ated along the face that reaches the tailgate area. Dust
levels ranged between 1.04 mg/m3 to 3.88 mg/m3 and
averaged 2.56 mg/m3 .
The volume of air introduced on longwall faces in
recent years has had a major impact on dust levels at
the stationary sampling locations. Average dust levels
at the intake sampling location were reduced by 53%
when compared to the average intake samples mea-
sured during the previous (1995) longwall surveys.
Reductions in dust levels at the other stationary sam-
pling locations ranged between 27% and 39% when
compared to the 1995 study.
Average dust concentrations at the UPWIND and
SHEARER mobile sampling locations were higher for
the head-to-tail passes (figure 4). Dust levels at these
two sampling locations reflect dust liberated from
 Figure 5. Adjusted PDR concentrations for a typical head-to-tail pass.
shields being advanced when mining head-to-tail.
At the SHEARER sampling location dust generated
by the shearer, primarily the headgate drum is also
included. Isolating dust generated by the shearer is
accomplished by subtracting the UPWIND sampling
concentrations from the DOWNWIND concentra-
tions. Average shearer-generated dust was found to be
2.06 mg/m3 when mining head-to-tail and 2.82 mg/m3
while cutting tail-to-head. Higher dust levels produced
by the shearer were evident while mining tail-to-head
when the headgate drum is the primary cutting drum.
Shearer operators downwind of the headgate drum
may be exposed to fugitive dust caused by the cut-
ting action of the drum. This is not the case when
mining head-to-tail when the dust generated by the
primary cutting drum (tailgate drum) is downwind of
the headgate shearer operator and possibly the tailgate
shearer operator.
Comparing dust levels at shield 10 with UPWIND
samples for tail-to-head passes showed an increase of
0.49 mg/m3 . Dust liberated by face spalls, from the
face conveyor, and dust migrating from the gob may
be causing the increase in dust levels.
When comparing average dust levels from the
UPWIND and SHEARER sampling locations, dust
levels increased 0.51 mg/m3 for tail-to-head cuts and
0.31 mg/m3 for head-to-tail cuts. In the 1995 study, the
increase in dust levels was 1.85 mg/m3 and 1.29 mg/m3
for tail-to-head and head-to-tail cuts when compar-
ing the two sampling locations. This suggests that the
increased air velocity down the face in conjunction
with the directional spray systems have had a positive
effect of confining fugitive dust close to the face and
away from the walkway.
230
Copyright © 2006 Taylor & Francis Group plc, London, UK
Instantaneous dust samples measured with the PDR
were used to calculate relative dust concentrations
at each sampling location along the face. These rel-
ative dust measurements calculated from PDR data
were based upon time study data obtained on a per-
pass basis. Extended downtimes and wedge cut dust
concentrations were excluded from the average PDR
dust concentration. The PDR dust levels were adjusted
based upon the ratio between the average dust con-
centrations obtained from the two gravimetric sam-
plers divided by the dust concentration obtained from
the PDR.
Figures 5 and 6 illustrate adjusted PDR mobile
sampling results from a typical head-to-tail and tail-
to-head pass. Figure 5 shows that OUTBY shield
movement dust levels are relatively low [averaging
approximately 0.50 mg/m3 ]. Significant increases in
dust levels are found at the UPWIND sampling loca-
tion. This supports the hypothesis that much of the dust
liberated during head-to-tail passes is generated by
advancing shields. Examining data at the SHEARER
sampling location shows that the directional spray sys-
tem in conjunction with the high velocity/volume of air
moving down the face has a positive effect at diluting
high dust levels generated by shield movement. Dust
levels at the DOWNWIND sampling location [approx-
imately 1 shield inby the tailgate shearer operator and
1 to 2 shields outby shield movement] are approxi-
mately equivalent to the dust levels recorded by the
SHEARER samplers.
Figure 6 shows dust levels to be relatively low at the
three shearer sampling locations when mining tail-to
head. These dust levels are very similar to the outby
dust levels measured at Shield 10 and the OUTBY
 Figure 6. Adjusted PDR concentrations for a typical tail-to-head pass.
sampling location. The spike in dust levels at the
SHEARER and DOWNWIND sampling locations as
the shearer approached the headgate is explained by
shield advances. Because of unstable roof conditions,
shields were pulled in advance of the shearer. The
SHEARER and DOWNWIND samplers were inby the
advancing shields and significant increases in dust lev-
els were evident. These dust levels closely correspond
to dust levels at the UPWIND sampling location when
mining head-to-tail.
4 CONCLUSION
Although the number of longwall sections has
decreased over the past decade, the production and
productivity from today’s operating longwalls is much
greater. Consequently, the potential to liberate signifi-
cantly more respirable dust exists. Mine operators have
made substantial increases in the application of dust
control technology in an effort to control dust exposure
on longwalls.
Longwall dust parameters obtained from MSHA
ventilation plans and sampling data from eight sur-
veys conducted by NIOSH indicate that a wide variety
of control technology applications exists across the
industry. These differences may in part be a reflec-
tion of differing operating conditions and equipment.
Regardless, a number of similarities also exist and
indicate that these controls are viewed as keys for
successful dust control:
• Intake dust levels are generally well controlled.
• The velocity and quantity of air being utilized has
been increased.
231
Copyright © 2006 Taylor & Francis Group plc, London, UK
• The volume of water supplied to shearer dust control
sprays has also been increased.
• All operations are using a version of directional
water sprays (shearer clearer) in an effort to confine
dust near the face and away from workers.
• The vast majority of operations are using some form
of splitter arms on the headgate side of the shearer
to contain dust liberation from the headgate drum.
• Splitter arms and/or directional sprays are being
used successfully at the tailgate end of the shearer to
provide protection for the tailgate shearer operator.
In addition, sampling results indicate that signifi-
cant dust liberation and worker exposure is resulting
during shield advance. Given that the majority of
sampled longwalls are utilizing bidirectional cutting
sequences, the shearer operators are being exposed to
this dust during the head-to-tail cutting pass. Addi-
tional research is needed to identify effective control
technologies for limiting dust exposure to shield dust.
REFERENCES
Colinet, J.F., Spencer, E.R. & Jankowski, R.A. 1997. Status of
Dust Control Technology on U.S. Longwalls. Proceeding,
6th International Mine Ventilation Congress. May 17–22,
Pittsburgh, Pa., Chapter 55, pp. 345–351.
Fiscor, Steve. 2003. Total Number of Longwalls Drop to
52 – US Longwall Census 2003. CoalAge. February, 2003,
pp. 20–24.
Niewiadomski, G.E. 2004. Mine Safety and Health Admin-
istration, private communication.
O’Green, J.E., McNider, T.E., Gallick, J.M., Jarrett, R.J. &
Jankowski, R.A. 1994. An Overview of General Operating
 Experience as It Relates to Longwall Dust Control, Pro-
ceedings of Longwall USA, Pittsburgh, Pa., pp. 221–233.
Potts, J.D. & Jamkowski, R.A. 1992. Dust Considerations
When Using Belt Entry Air to Ventilate Work Areas. U.S.
Bureau of Mines RI 9426, 12 pp.
Webster, J.B., Chiaretta, C.W. & Behling, J. 1990. Dust Con-
trol in High Productivity Mines, SME Annual Meeting,
Preprint 90–82, Salt Lake City, Utah, 9 pp.

232

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

A field demonstration of the JOY Wet-Head Miner technology

Y.P. Chugh, A. Patwardhan, H. Gurley, A. Moharana, R. Saha

Department of Mining and Mineral Resources Engineering, Southern Illinois University, Carbondale, IL, USA

R. Carty
Illinois Clean Coal Institute, Carterville, Illinois

ABSTRACT: A field demonstration of the JOY Wet-Head Miner (WHM) technology was performed at a South-
ern Illinois mine. The WHM technology utilizes sprays mounted behind each bit to reduce ignition potential and
increase bit life by cooling the “hot-spots” and the bit itself and reduce respirable dust by controlling dust genera-
tion at the source. The study was conducted in three phases. In Phase I, limited pilot studies were conducted on the
standard miner and the WHM to determine the variability in measurements and to formulate a scientific test plan
for more elaborate studies in Phase II and Phase III. In Phase II studies, sampling was conducted to obtain statis-
tically significant comparisons of the WHM with the standard miner. Phase III studies were aimed at improving
the performance of the WHM by modifying the sprays on the miner with respect to the pressures and locations.
Bit usage and productivity of the WHM and the standard miner were also monitored throughout the testing. The
results of these studies indicate that the WHM technology significantly improves visibility and appears to reduce
respirable dust to a limited extent in and around the face area. It also appears that there is a significant improvement
in bit life due to the cooling effect of the wet-head water sprays. The industrial engineering studies have confirmed
that there is no impact on productivity from WHM use. Overall, improvements in visibility, reduction in bit usage
and the ignition control benefits related to the use of WHM have the potential to improve mine productivity and
profitability which in turn should improve coal competitiveness and the health and safety of miners.

1 INTRODUCTION AND PROBLEM

STATEMENT

With an increased emphasis on high production rates in

the coal industry, more and more US coal operators are
having difficulty meeting stipulated dust control stan-
dards. The high out-of-seam dilution mining (20–25%
of seam thickness), and associated quartz content have
only exacerbated the problem. Recently, dust control
in the face and out-by areas has become a bottleneck
to achieve competitive coal production rates at several
mines in the USA.
To reduce ignition potential, improve bit life and
better control and suppress dust generation at the
source itself, Joy Mining Machinery (JOY) has devel-
oped a novel wet-head miner technology. A “wet head
miner” (WHM) is a continuous mining machine in
which water under pressure (∼100 psi) is injected
behind each bit on the cutting drum. The water is
directed at the cutting point to (a) minimize dust gener-
ation at the cutting point, (b) cool the bits to minimize Figure 1. Joy Wet-Head Miner J35 bit and block spray.
potential for methane ignition and extend bit life, (c)
quench any hot-spot trails left behind by the cutting bit, The WHM has been demonstrated at only 2–3 loca-
and, (d) more effectively wet the coal prior to loading tions around the world since the early 1990s. The
by the miner conveyor to reduce dust during transport. technology appears to have proven very effective of
A WHM cutting bit with sprays is shown in Figure 1. late after initial issues regarding water seals design

233

Copyright © 2006 Taylor & Francis Group plc, London, UK

 were addressed. A recent study in South-Africa doc-
umented large dust reductions using this technology
(Zyl 2005). These results were however not tested for
statistical significance. In the USA, a mine in West
Virginia (WV) recently demonstrated the WHM tech-
nology with National Institute of Occupational Safety
and Health (NIOSH) monitoring dust generation and
control (NIOSH 2005). The results indicated a quali-
tative but significantly improved visibility and limited
improvement in respirable dust.
Comparative tests conducted by SIU on the wetta-
bility (Chugh et al. 2004) of coal from WV mine with
different Illinois coals indicated that the WHM should
be much more effective in Illinois mines. Hence, due to
the interest of the Illinois coal industry a WHM demon-
stration was conducted at a Southern Illinois coal mine.
2 SIU DEMONSTRATION STUDIES
2.1 Mine description
The WHM demonstration was conducted at a Southern
Illinois mine extracting Illinois No. 5 seam coal. This
mine uses room-and pillar mining system with an aver-
age mining height of approximately 6.5-ft and the
average entry width of about 20-ft. The mining layout
consists of a walk-between super-section with one
intake and one return in a nine-entry system being
mined with two continuous miners. A blowing face
ventilation system with line curtains is used. The ven-
tilation plan requires minimum 7,500 cfm at the end
of the line curtain when the scrubber is operational.
The scrubber discharges air on the left hand side
of the miner. Typical airflow in the last open cross-
cut is 15,000 cfm. In the section sampled, mining is
done using one JOY 14CM15 wet-head miner on the
intake side of the section and another Joy 14CM15
conventional miner on the return side of the section.
Both continuous miners are equipped with flooded-
bed dust scrubbers. On the wet-head miner, chassis
and boom sprays operated at about 100 psi and the
wet-head sprays operated in the range of 80–110 psi
depending on test conditions. On the standard miner,
the plan required chassis/boom sprays to be operat-
ing at 100 psi. The total water usage on the wet-head
and standard miners was approximately 40–42 gpm
and 29–32 gpm respectively. Thirteen (13) ton ram
cars were used to haul coal from the face to the
feeder/breaker. The ram car operator was located on
the right hand side of the car.
2.2 Demonstration studies
The WHM field demonstration program was devel-
oped to answer the following major questions.
1. How does productivity and production capacity
of the WHM compare with standard continuous
234
Copyright © 2006 Taylor & Francis Group plc, London, UK
miners currently in use at the demonstration mine?
This comparison was achieved through industrial
engineering studies on the WHM and standard min-
ers in a single super-section with relatively similar
mining conditions.
2. How does the dust control efficiency of the WHM
compare with the standard miners currently being
utilized? This comparison involved comparative
dust measurements in the face area and out-bye
areas as well as on selected occupations as required
by the regulatory agencies.
3. What is the impact of bit cooling in WHM on
bit life? To evaluate this, a cut-by-cut bit change
count on all the sampled cuts was maintained and
aggregated for comparisons.
The overall study was structured into three phases.
In Phase I, limited pilot studies were conducted on the
standard miner and the WHM to determine the vari-
ability in measurement and to formulate a scientific
test plan for more elaborate Phase II studies. In Phase
II studies, additional sampling of the standard miner
and WHM was conducted to obtain statistically signif-
icant comparisons. Fifty eight (58) cuts were sampled
using the WHM and thirty four (34) cuts using the stan-
dard miner. Phase III studies focused on improving the
WHM performance by reorienting the boom sprays to
increase water addition in the center of the drum at
the expense of reduced water at the drum extremities.
The wet-head spray pressures were also increased to
improve dust capture.
2.3 Miner configurations tested
The following different miner configurations were
tested in this study.
1. Standard Miner (REG) – A conventional continu-
ous miner without wet-head spray cutting drum.
Used smaller shank diameter bits at this mine.
Had only boom and chassis sprays. Used about
29–32 gpm water for dust control.
2. Wet-Head Miner (WHM) – A conventional contin-
uous miner with wet-head spray cutting drum. At
this mine, it used larger shank diameter bits. Had
wet-head sprays in addition to the boom and chas-
sis sprays. Used 40–42 gpm water for dust control.
Boom/Chassis sprays operated at about 100 psi,
while wet-head sprays operated in the 80–110 psi
range depending on test conditions.
3. Dry-Head miner (DHM) – Same as WHM but oper-
ated with the wet-head sprays turned off and boom
and chassis sprays installed similar to the REG
miner. Tested in Phase III of this study only.
4. Wet-Head Miner – Modified (MOD) – WHM oper-
ated with modified spray configuration and spray
pressure. Used 40–42 gpm water for dust con-
trol. Boom sprays operate at about 100 psi, while
 wet-head sprays operate at about 100–110 psi. Straight Cut -
Tested in Phase III of this study only.
Deeper

2.4 Data collection and analysis ~30’

Roof Bolter
2.4.1 Industrial engineering studies
Industrial engineering studies were performed on both 2
the WHM and the standard miner to compare the
40’+
production potential of WHM vis-à-vis the standard
miner. Comparison of production potential of the two
miners required the collection of production character- 41’
istics of both the miners and haulage system, geologic
and panel characteristics. Both the miners were placed 4
1 Intake Air
in the same super section with WHM as the intake Return
miner and the standard miner as the return miner. Air
This allowed calculation of the variability in produc- 3 19’
tion potential due to the miners while assuming similar Curtains
geologic and panel characteristics.
Collection of equipment performance data (produc-
tion characteristics) was done by continuous watch Ram Car In
time study of the miner-haulage system. A two-person & Out
time study was planned with one person near the miner
and the other near the feeder. Data was collected on
time study forms such that all the time during a shift
could be accounted. The following times were col- Figure 2. Dust sampling locations indicated by 1, 2, 3, 4.
Locations 1 and 4 are the intake and the return.
lected (a) Load Time, (b) Haulage travel time from
(i) Face Change Out to Face, (ii) Face to Face Change
Out, (iii) Face Change Out to Feeder Change Out,
(iv) Feeder Change Out to Feeder, and (v) Feeder to 2.4.3 Dust data collection
Face Change Out, (c) Dump time, and (d) Delays due Since, only one miner in the section was the WHM, a
to breakdowns. Simultaneously, haulage paths for the cut-by-cut sampling strategy was adopted to study the
ram cars were collected for each cut and haulage dis- dust control impact of the WHM.The data was grouped
tances computed. Apart from the above time study, into one of 7-different type of cuts which are typical in
a time study on the roof bolters was also conducted continuous miner mining. These were, initial straight
during initial phase of the study. However, since the (starting at 0–10-ft from the face), deep straight (start-
data indicated that roof bolters were not a bottle- ing at 11–40-ft from the face), deeper straight (starting
neck in the production system, the study was later at 40+-ft from the face), right turn, right blow-through,
discontinued. left turn and left blow-through. Figure 2 depicts a
The above collected data was statistically analyzed “deeper” cut plan with indications of dust sampling
to compare the production potential of the standard locations. Comparisons were made between the stan-
miner and WHM. Data analyzed consisted of the load- dard miner and the WHM and in between types of cuts.
ing capacity of the ram car, loading rate (haulage The pump locations for dust sampling were as
capacity/load time) of the continuous miner, and follows:
haulage unit travel speeds. 1. Intake – IN, Location 1
2. Near Miner Operator – MO, Location 2
2.4.2 Bit usage data collection 3. Near Ram Car Operator – RC, Location 3
A bit usage study was conducted on all the sampled 4. Return – RET, Location 4
cuts. A record was kept of the number of bits replaced
Other measurements taken for all the cuts included
after every cut was made for the standard and the wet-
the following:
head miner. Even though the number of bit changes
after a particular cut is not related to that cut itself, 1. Line Curtain (LC) Air
it does represent an aggregate of several previous cuts 2. Last Open Cross-Cut (LOXC) Air
made by that miner. Hence, after data was collected for 3. Car Count
a sufficiently large number of cuts, averages and stan- 4. Cut Dimensions (Height, Width, Depth)
dard deviations were computed and the bit usage/cut 5. Cut cross-section (Roof, Seam, Floor)
was statistically compared for the standard and the 6. Cut Time
wet-head miner. 7. Water Pressure and Water Volume

235

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Mine Safety and Health Administration (MSHA)
approved gravimetric sampling equipment and proce-
dures were utilized for dust sampling.
2.4.3.1 Data measurements and transformations
The filter cassettes from the dust pumps were sent
to MSHA for weighing at a resolution of 1 micro-
gram. The weight gain of the cassettes and the pump
operation time were used to compute the raw dust con-
centrations at the four sampled locations. These dust
measurements were then transformed to isolate the
impacts of cut-time variability (down-time during
cuts), varying LC and LOXC air flows and the dif-
ferences in intake dust concentrations. The trans-
formations were accomplished using the following
methodology.
– To correct for different intake air dust concentra-
tions, the Location 1 measurement was subtracted
from the measurements for the other three locations.
– The cut-times for cuts of same dimensions may dif-
fer due to haulage delays as well as unexpected
delays encountered during the cuts. These down-
times (delays) were estimated as the difference
between cut-time (pump run time) less average time
for mining the particular cut volume. The Location
2, 3, and 4 dust concentration numbers were then
adjusted for this down-time for changing the divisor
time for the dust data from the actual pump run time
to the expected run time. In effect, this adjustment
accounts for varying production.
– Varying ventilation air flows to the face and in the
LOXC directly impact the measured dust concentra-
tions as the generated dust is either diluted by excess
air or concentrated at lesser airflow. To correct for
these variations, all the measured dust concentra-
tions were normalized for 8,000 cfm of LC air and
15,000 cfm of LOXC air. The LC normalization was
applied to the Location 2 sample while the LOXC
normalization was applied to the Location 3 and
Location 4 samples. It is noted here that the normal-
ization of Location 3 data is actually cut dependent.
Depending on the type of cut, this data should be
normalized with LC or LOXC air. Also, air flow
corrections on Location 2 samples may not be com-
pletely accurate due to recirculation of line curtain
air prevalent when fresh air intake is lower than the
scrubber capacity. Hence for the sake of robustness,
subsequent statistical analyses were performed on
both the corrected and uncorrected measurements.
The authors believe that Location 4 (Return) sam-
pling position and the corrections on it are free of
any major shortcoming and the measurements at this
location are the best indicator of the efficiency of
the employed dust control system. This assessment
is further validated by the low variance of measure-
ments encountered at this particular location.
236
Copyright © 2006 Taylor & Francis Group plc, London, UK
2.4.3.2 Data diagnostics and screening
The data diagnostics involved Post-Hoc comparisons
between the measured dust concentrations and the
cut-types. These comparisons indicated no statistically
significant difference between cut-types.
To eliminate the extreme observations from skew-
ing mean computations, the means and standard devia-
tions of dust concentrations were calculated separately
for Locations 2, 3 and 4. Observations lying outside
the ±2 Standard Deviations range were eliminated as
outliers.
2.4.3.3 Statistical data analysis
Two approaches of statistical evaluation were used in
this analysis on the screened data. The first approach
involved computation of the means and standard devi-
ations of the data for Locations 2, 3 and 4 separately for
the standard and the wet-head miners. For each loca-
tion, “t-tests” were conducted comparing the means
of the standard and the wet-head miner under the
null-hypothesis that the means were the same. The sig-
nificance level for a two-tailed test was established at
0.10 and the null hypothesis was rejected for t-values
lower than the significance level.
Bivariate and partial correlations between the
dependent and independent variables were also com-
puted to gain better insight on the true factors affecting
the measured dust concentrations.
3 RESULTS AND DISCUSSIONS
3.1 Industrial engineering studies
The load and dump times and haulage tram speeds
for the ram cars were collected for both the WHM
and standard miner. Statistical comparisons indicated
lower loading times and higher dump times for WHM
as compared to the standard miner. Lower loading
times (higher loading rates) can be attributed to better
visibility and somewhat better dust control thus allow-
ing the miner operator to cut and load more effectively.
This also allows the ram car operator to position his
car behind the miner conveyor easily. The higher dump
time of WHM could be due to higher water content of
coal thus slowing down the dump process. The unit
haulage speeds were statistically similar.
The above data was modeled using SIU-Suboleski
Production (SSP) model (Chugh et al. 2005) assum-
ing no difference in haulage unit speed over shorter
distances (feeder change-out to feeder and face to
face change-out). The results indicated ROM unit pro-
duction of 2,221 tons for conventional miner and
2,195 tons for WHM. The difference is very small
to be significant. The collected data was also used
in a stochastic production planning model (Moharana,
2004) to assess production in the mining cut sequence.
The production comparison outputs are indicated in
 Table 1. Production comparison using stochastic model. Table 2. Cut-cross-section comparisons.

Mean SD Standard (REG) WHM

Production Standard 2481 108 Avg No. of Avg No. of

WHM 2478 75 (inches) samples (inches) samples

Cut height 78.3 31 78.3 48

2900 Rock height 20.9 22 21.6 47
Floor 6.5 2 4.3 27
2700
Expected Production

2500
Table 3. Summary PHASE II results. Comparison of trans-
2300
formed and corrected dust measurements.

2100 Sample size

Regular Miner requirement*
Wet Head Miner
1900
Location REG WHM p-value** Total Sampled
1700

2 Mean 1.99 1.43 0.32 438 47

1500
0 10 20 30 40 50 60 70 80 90 100 SD 1.68 1.58
Simulation # Samples 11 36
3 Mean 1.35 2.66 0.08 355 41
Figure 3. Comparison of expected production of Standard SD 1.67 1.77
and Wet-Head Miner based systems. Samples 7 34
4 Mean 4.41 3.65 0.05 100 92
Table 1. Figure 3 shows the expected variability in SD 2.02 1.62
production using the WH and the standard miner. Samples 34 58
Although WHM production is almost equal to the con-
ventional miner production, the production variability * Sample size required to resolve a difference of 15% between
is lower for WHM system than standard miner system. WHM and REG at 90% confidence.
This is expected to be a result of lower delays encoun- ** Values lower than 0.10 indicate statistical significance at
tered with the WHM system allowing for a smoother atleast 90% confidence.
production cycle.
the WHM at Location 3 which was statistically sig-
nificant. However when measurements not corrected
3.2 Dust control studies
for air-flow variations were considered (Table 4), the
3.2.1 Phase I studies difference between the two miners was not statistically
The Phase I dust sampling program involved limited significant. Location 4 experienced lower dust concen-
pilot tests on the standard miner and the WHM to deter- trations with WHM. This difference was significant
mine the variability in measurements and to formulate both for the corrected and uncorrected comparisons.
a scientific test plan for more elaborate studies in Phase Also, it must be noted that the minimum sample
II and Phase III. Phase I data as such was however size requirements were marginally satisfied only for
excluded from subsequent comparative analyses. Location 4 measurements.
As observed in Phase I studies, WHM significantly
3.2.2 Phase II studies improved wetting of fugitive dust that contributes to
In Phase II studies, sampling of standard miner and the visibility in the face area. Most workers in the
WHM was done to obtain statistically significant com- face area commented favorably on WHM based on
parisons. Fifty-eight (58) cuts were sampled using visibility.
the WHM and thirty-four (34) cuts with the stan-
3.2.3 Phase III studies
dard miner. The cut geometries for standard miner and
Phase III dust control studies were designed to improve
WHM sampled cuts are compared in Table 2. The data
the dust control performance of the WHM by rear-
indicates that geometry of sampled cuts was very sim-
rangement of the boom sprays and increasing the
ilar for the two types of miners. The impact of WHM
wet-head spray pressures from approximately 80 psi to
in reducing respirable dust at different sampling loca-
100 psi. JOY professionals and the authors identified
tions is summarized in Table 3. The results indicate that
and utilized two hypotheses for improving WHM dust
Location 2 experienced slightly less average dust con-
control performance.
centration with WHM compared to the standard miner.
The difference however is not statistically significant. – Operate WHM bit sprays at higher pressure of 100
Higher average dust concentration was recorded with psi. This will result in reduced gas ignition potential

237

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 4. Dust control results from five configurations/phases of testing on the conventional miner
and wet-head miner at three different sampling locations in the face area. Raw respirable dust con-
centration data has been corrected for intake dust concentration and production. Impact of varying
ventilation conditions has been ignored at this time. MO – Near Miner Operator Location, RC – Near
Ram Car Location, RET – Return Location. Boom/Chassis Spray Pressure at 100 psi.

Dust concentration (mg/m3 )

average, standard deviation
number of samples
Miner type, Height Water WH sprays
Phase (inches) (gpm) pressure (psi) MO RC RET

REG P-II 78.3 32 – 1.51,1.3311 1.98,3.237 4.91,2.2 34

WHM P-II 78.3 41 80 1.22,1.3436 2.88,2.0534 3.65,1.6858
REG P-III 72.4 32 – 0.51,0.5312 2.52,1.7412 3.28,1.0513
DHM P-III 74.9 32 – 0.99,1.1113 1.60,0.9913 2.86,1.2013
MOD P-III 69.7 41 105 0.66,0.6944 1.36,0.9943 2.38,1.0346

and increased dust reduction potential. JOY has rec-
ommended this pressure in the past based on studies
done in UK (Hole 2000) and Australia (Allanson &
Roxborough 2001).
– WHM dust reduction potential can be improved
through increasing the amount of water in the center
portion of the drum/cutting area to minimize dust
generation.
Seventy-two additional cuts were sampled in Phase
III of which forty-six (46) were on the modified WHM
(MOD), thirteen (13) on the standard miner (REG) to
allow a side-by-side comparison and the thirteen (13)
on the Dry-Head Miner (DHM) which involved oper-
ation of the WHM machine as a standard miner by
turning off the wet-head sprays and installing chas-
sis/boom sprays similar to the standard miner. The
results of these studies, tabulated in Table 4, indicate
the following:
– Performances of REG and DHM were not sta-
tistically different. This supports the comparisons
between WHM and REG in Phase II where REG
was considered as the baseline.
– Comparison of REG with MOD indicated that the
MOD configuration provided better dust control at
a statistically significant 95% confidence level at
both the RET and RC sampling locations. Dust con-
centrations in the MOD configuration were 21%
and 27% lower at the RC and RET location com-
pared to the REG configuration. At 90% statistical
confidence, these translate to about 9% and 11%
reductions. The differences in performances at the
MO sampling location were not statistically signif-
icant for any of the tested configurations. This is
due to higher variability typically encountered while
sampling this location. There are reasons explaining
this variability which are not discussed here.
– The performance of MOD was found to be
significantly better than the WHM performance in
Phase II studies (35% dust reduction; 23% reduction
at 90% confidence). However, this result should be
accepted with some trepidation as the time frames
of the two series of testing were different and the
mining height with MOD was lower at 69.7-inches
compared to 78.3-inches for the WHM (Phase II).
This stresses the importance of side-by-side testing
as in Phase II studies which indicated improved per-
formance with the WHM, albeit, at a much smaller
significance.
3.2.4 Summary dust control comparisons
The studies on the WHM conducted during Phases II
and III involved testing the following configurations
in different phases.
1. REG (Phase II)
2. WHM (Phase II)
3. REG (Phase III)
4. DHM (Phase III)
5. MOD (Phase III)
Phase I testing was excluded in these comparisons
as the testing and sampling methodologies were being
formulated during that time. The comparisons between
all combinations of the five configuration/phases
listed above are presented in Table 5. Only com-
parisons at the RET location are presented as these
observations were found to be the most reliable in
terms of lower variability and also satisfied sample
size requirements for the desired resolution.
In summary, the results indicate the following.
– From Table 5, it appears that the controlling factors
for improved dust reduction performance at the RET
location may be lower mining height and higher
water usage or the WHM technology. The available
data does not permit differentiation of the contribu-
tions of higher water usage from that of the WHM
technology since the higher water use was only
associated with WHM technology.
– Table 6 supports the previous observation indicating
strong correlations of mining height, water usage

238

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 5. Statistical comparisons of the differences in dust Table 6. Bivariate correlations between important parame-
concentration at the return sampling location for the five dif- ters affecting dust concentration in the RET location in the
ferent configurations/phases of sampling on the conventional face area.
miner and the wet-head miner. Impact of varying ventilation
conditions has been ignored at this time. Height GPM Pressure

% p- % Height ρ+ 1 −.095 −.475*

Comparison Diff.+ value* Diff. # Remarks Sig. (2-tailed) .229 .000
N 164 164 104
REG PII – 33.2 0.001 16.9 Lower Mining Height GPM ρ 1 (# )
REG PIII in Phase III Sig. (2-tailed) .000
REG PII – 25.6 0.006 10.7 Higher Water Usage N 164 104
WHM PII with WHM Dust ρ .442* −.274* −.406*
Sig. (2-tailed) .000 .000 .000
REG PII – 41.8 0.000 24.5 Lower Mining Height
N 164 164 104
DHM PIII
REG PII – 51.5 0.000 37.6 Lower Mining Height & +
Pearson Correlation.
MOD PIII Higher Water Usage * Correlation is significant at the 0.01 level (2-tailed).
with MOD #
Cannot be computed because at least one of the variables is
REG PIII – 10.1 0.321 – Higher Mining Height & constant.
WHM PII Higher Water Usage
with WHM
Table 7. Partial correlation between wet-head spray pres-
REG PIII – 12.9 0.350 – Higher Mining Height & sure and dust concentration (while controlling for mining
DHM PIII Higher Water Usage height).
with DHM
REG PIII – 27.5 0.007 10.9 Lower Mining Height & Dust
MOD PIII Higher Water Usage
with MOD Correlation −0.237*
WHM PII – 21.7 0.060 2.9 Lower Mining Height Pressure Sig. (2-tailed) 0.016
DHM PIII with DHM df 101
WHM PII – 34.8 0.000 22.6 Lower Mining Height
* Correlation is significant at the 0.02 level (2-tailed).
MOD PIII with MOD
DHM PIII – 16.8 0.159 – Lower Mining Height
MOD PIII with MOD correlations between dust concentration and wet-
head spray pressure are estimated while controlling
* This column gives the likelihood that there is no statistical for mining height (Table 7). The partial correlations
difference between the two comparisons (H0 :µ1 = µ2 ) For indicate that higher wet-head pressure does indeed
example, 0.017 value means that there is no more than a reduce dust.
1.7% chance the REG(Phase II) and Reg (Phase III) results
are not different. Another way of presenting the same result
is that we are 98.3% (100-1.7) confident that Phase II and 3.3 Bit usage studies
Phase III result for REG (Standard Miner) are different.
+
Percent difference between the second testing compared to The WHM technology uses sprays mounted behind
the first testing. Thus, in the first row, dust control in Phase each bit to suppress dust generation at the source. This
III testing using the standard miner was 25.6% better in Phase spray arrangement results in cooling of the bits which
II testing. can improve bit life. In this study, bit usage compar-
#
Percent difference at 90% confidence. isons were conducted between the REG, WHM, DHM
and MOD configurations described above. The REG
and wet-head spray pressure with dust concentra- was a standard miner using smaller shank bits at this
tion. The correlation of dust however is strongest mine while DHM was essentially a standard miner
with mining height followed by the water use (GPM) but with larger shank bits as used on the WHM at
or the WHM technology. this mine. MOD was the WHM with larger shank bits
– Within the WHM observations, it appears that but with a different boom spray arrangement than that
higher wet-head spray pressure is negatively and used on the WHM. The bit usage comparisons for these
significantly correlated with dust concentration. configurations during their test periods in Phase II and
However, there happens to be a significant corre- Phase III studies are provided in Tables 8 and 9.
lation between wet-head spray pressure and min- In Phase II, WHM bit usage was about 46% lower
ing height as the conducted experiments were than the standard miner. However, the two miners used
not orthogonal. Hence, to evaluate the effect of different size bits making this comparison somewhat
wet-head spray pressure on dust reduction, partial questionable. Hence, Phase III studies were conducted

239

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 8. Bit usage data collected in Phase I and Phase II.
REG WHM
Average bit usage/cut 5.10 2.73
Std dev 4.40 3.61
Samples 49 77
Table 9. Bit usage data collected in Phase III.
DHM REG MOD
Average bit usage/cut 2.62 2.33 3.23
Std dev 3.97 3.94 3.39
Samples 13 12 39
to obtain a more direct evaluation of the impact of
wet-head sprays on bit usage where the WHM with
the larger bits was operated with and without the wet-
head sprays on. Due to the relatively small number of
data points collected in Phase III studies, no concrete
statistical analyses on bit usage could be conducted.
However, a few general observations could be made
observing the trends of the bit usage. These were:
– The DHM and REG in Phase III studies have sim-
ilar bit usages. This indicates that the difference in
bit usage observed between the WHM and REG
in Phase I and II studies is due to the wet-head
effect and not due to the different bit type used on
the WHM.
– The bit usage for MOD is however not statistically
different than the bit usage for REG or the DHM.
This could be possibly due to the reduced water
usage on the Wet-Head in the MOD configuration
compared to the original WHM configuration. The
reduction in wet-head spray water from WHM to
MOD was from 25.2 to 21.2 gpm.
4 CONCLUDING REMARKS
Based on the field demonstration studies conducted
on the JOY wet-head miner, the authors conclude the
following:
– WHM does not appear to drastically reduce res-
pirable dust concentration in the face area. Although
about 10% reduction in the last open cross-cut return
sampling location is indicated, it can not be stated
with confidence whether it is due to WHM or due to
ventilation or operations differences in the sampled
cuts for standard and WHM cuts.
– The cooling effect of the wet-head sprays on the bits
appears to significantly improve bit durability.
– Coal discharge into batch haulage units seems to
produce less fugitive dust from WHM because of
240
Copyright © 2006 Taylor & Francis Group plc, London, UK
higher moisture content in the coal. As a result, most
workers in the face area have consistently reported
improved visibility.
– Higher capital and maintenance costs of the WHM
would require higher production rates, lower dust
levels or a lower likelihood of a potential methane
ignition for it to be an economical alternative to stan-
dard miner. Since the estimated production rates for
WHM and standard miner are very similar, and res-
pirable dust reduction is currently not large, efforts
must be directed to improve dust control and quan-
tify the benefits of a lower ignition potential with the
use of WHM for it to have significant penetration
into the market.
– The authors believe that WHM technology has
potential for significantly higher respirable dust
control. This can be achieved with some modifi-
cations to the design and operation of WHM.
ACKNOWLEDGEMENTS
In addition to the funding support by the Illinois Clean
Coal Institute, the authors greatly appreciate the sup-
port of cooperating mines staff for their exceptional
support during this study. The authors also acknowl-
edge MSHA support in providing and processing the
dust sampling cassettes. In addition, the inputs pro-
vided by MSHA and NIOSH professionals are also
appreciated.
REFERENCES
Allanson, C.N. & Roxborough, F.F. 2001. Integrated Pick
Back Spray to Eliminate Frictional Ignitions when Using
Continuous Miners.Australian CoalAssociation Research
Program. ISBN 0 7334 1823 6.
Chugh, Y.P., Thatavarthy, K.K. & Patwardhan, A. 2004.
Development of a Wettability Test for Assessing Effec-
tiveness of Surfactants for Dust Control. Tenth U.S./North
American Mine Ventilation Symposium, May 16–19,
Anchorage, Alaska.
Chugh, Y.P., Moharana, A. and Patwardhan, A. 2005, Devel-
opment of simple production modeling software for
continuous miner production systems. International Sym-
posium on Mine Planning and Equipment Selection
(MPES-2005), November 1–3, Banf, Canada.
Hole, B.J. 2000. Assessment of a Continuous Miner “Wet-
Head” System. IMI Technical Services Limited. Contract
Research Report 301/2000. ISBN 0 7176 1864 1.
Moharana, A. 2004. Simulation of a continuous miner
room-and-pillar system. unpublished. Southern Illinois
University, Carbondale, Illinois.
NIOSH, 2005. Dust Study of Wet Head Miner at Performance
Coal’s Upper Big Branch Mine. Confidential.
Zyl, F.J. van, 2005. Operational evaluation of JOY’s wet-head
system at Goedehoop Colliery with regard to respirable
dust control. Report No. 2005-0327, CSIR MiningTek.
 4. Heat and humidity

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Controlling heat loads from cemented mine backfill

E. De Souza
Queen’s University, Ontario, Canada

ABSTRACT: During stope placement and the early stages of hydration, cemented backfill generates relatively
large amounts of heat and may create relatively high, undesirable temperatures in the underground environment.
In a series of investigations using scaled stope models instrumented with thermistors, the heat generated by
cemented backfill during hydration has been successfully quantified. Heat transfer to the hanging wall, orebody,
footwall and cross-cut and production drifts was assessed during the backfill curing process and relationships
between cement content, cure time, and heat generation were developed. Having determined the potential heat
contribution from cemented backfill, the author proposes the incorporation of post-consumer glass as an alternate,
equally effective and lower cost binder agent strategy for cement in backfill that may lead to cement consumption
reductions and reductions in heat load. Testing, using scaled stope models, has been implemented in order to
demonstrate the reductions in heat load from ‘glassfill’. The results of this work will assist mine operators in
predicting the heat generated by cemented backfill, and introduce glass based backfill as an effective means of
controlling backfill heat in the underground environment.

1 INTRODUCTION 40–50%; auto-compression 10–15%; equipment 20–

In hot and deep underground mines, prediction of
climate conditions is essential for planning air require-
ments and the cooling capacity necessary to provide
adequate working conditions. Although the dominant
method of heat removal in mines is by ventilation air,
the heat emitted in deep hardrock mines may become
intensified to require the installation of a refrigeration
system. The selection and design of a mine climate
control system is very complex. Identification and
quantification of each source of heat emitted into
the mine atmosphere is required in order to evaluate
the airflow volume needed to remove the heat or to
size a refrigeration plant. Extensive research has been
devoted by many to identify the sources of heat in
underground mines, and a number of models have been
developed to estimate the total heat load into the mine
air. Elaborate mine climate simulation programs have
also been developed to help evaluate such complex
models.
The major sources in mines include: heat from
exposed strata; heat due to auto-compression; heat
from machinery; heat from fissure water; heat from
transported rock; heat from personnel; heat from
explosives; heat from oxidation; heat from compressed
air pipes; heat from electric cables; and heat from
lighting equipment. Although the contribution of each
source to the heat load of a mine is site dependent, a
percent range in heat contribution is provided: strata
25%; water 15–20%; all other sources 2–5%.
Although cemented backfill is an important source
of heat, no studies concerning its contribution and
effect have been produced. Under certain circum-
stances, cemented backfill may become a critical
source of heat. Mines operating in the high arctic
for example, operate in permafrost ground conditions.
Although frozen rock exhibits superior strength and
demonstrates low susceptibility to ground disturbance,
mining development may result in thawing of the rock
mass due to heat transfer, with subsequent reduction
in rock strength. A number of such mines use frozen
backfill as a means of ground support, whilst some
have employed cemented backfill when secondary
filling is required. Under such circumstances, heat pro-
duced from cemented backfill, if not controlled, may
adversely affect the integrity of such mines. As the
thaw front advances into joints and fractures in the
rock mass, reduction in rock strength would result,
thus compromising the stability of the workings.
In highly productive, hot and deep mines utilizing
bulk mining methods and employing large quanti-
ties of cemented backfill, the backfill mass may
also become a major contributor of heat, signifi-
cantly affecting the underground climate. Workers
exposed to high temperatures are known to pro-
duce reduced work output and to have high accident
rates. Long-term exposure of workers to heat stress is
naturally unhealthy and unproductive and may result in

243

Copyright © 2006 Taylor & Francis Group plc, London, UK

 fatalities. In such case, accurate quantification of the
heat emitted by the fill mass into the mine atmosphere
is required in order to evaluate the most efficient
method of heat removal.
The author has recently introduced the first known
effort in quantifying the contribution of cemented
backfill to the heat load in mines (De Souza, 2005).
The study presented the results of a series of inves-
tigations aimed at measuring the heat generated by
cemented backfill during hydration. Scaled stope
models, instrumented with thermocouples, were used
in the investigation. Heat transfer to the hanging wall,
orebody, footwall, overcut, undercut and production
drifts were measured during the backfill curing pro-
cess and relationships between cement content, cure
time, and heat generation were developed.
Having determined the potential heat contribution
from cemented backfill, the author proposes in this
paper the incorporation of post-consumer glass as
an alternate, equally effective and lower cost binder
agent strategy for cement in backfill that may lead
to cement consumption reductions and reductions in
heat load. Extensive work has already been completed
to demonstrate the strength properties, pipeline trans-
portation behavior and economic feasibility of glassfill
(De Souza & Archibald 2005, De Souza et al. 2005,
De Souza & Archibald 2004).
In this paper, engineering testing has been imple-
mented in order to demonstrate the reductions in heat
load from ‘glassfill’. Such tests have incorporated
scaled stope models, instrumented with thermistors,
similar to the previous work with pure cement backfill.
2 THE TESTING PROGRAM
2.1 Physical model development
Cemented backfill may contribute to the heat load in
mines in different ways: exposed fill face in empty
stopes; exposed fill in overcut drifts; exposed fill in
drawpoints; and heat transfer to adjacent rock.A scaled
physical model of a backfilled stope was constructed
to determine the different contribution processes of fill
to the overall heat load, as presented in Figure 1.
The model consisted of 5 granite blocks to sim-
ulate the orebody and stope walls. Blocks A and B,
used to represent unmined stopes, were 0.56 m high ×
0.15 m wide × 0.27 m deep. Block B had an under-
cut drift 2.54 × 2.54 cm in cross-section. Between
blocks A and B was the backfilled stope, 0.56 m
high × 0.15 m wide × 0.27 m deep. Blocks C and D,
0.08 m thick × 0.5 m wide × 0.36 m deep, were used
to represent the rock mass below and above the
mining area, respectively. Block D had an overcut
drift 2.54 × 2.54 cm in cross-section. Block E, used
to represent the footwall, was 0.56 m high × 0.47 m
244
Copyright © 2006 Taylor & Francis Group plc, London, UK
wide × 0.11 m deep. Block E had a haulage drift and
two cross-cuts, each 2.54 × 2.54 cm in cross-section.
One cross-cut connected the haulage drift to the block
B undercut and the other cross-cut connected the
haulage drift to the backfilled stope bulkhead.
Twelve thermocouple probes were installed in the
model to monitor the temperature profile within the
mine area. The thermocouple probes were Type T flex-
ible insulated wire probes, 0.254 mm in diameter (30
gauge) and 1.5 m long. Figure 2 illustrates the tem-
perature probe position configuration in longitudinal
and cross-sectional views. Probes 1, 2, 3, 4 and 5
were positioned within the backfill and ore to pro-
vide the temperature profile along a horizontal line
passing from the center of the backfilled stope to the
center of the orebody. Probes 6, 9, 10 and 11 were
positioned to provide the vertical temperature profile
within the fill and in the overcut drift. Probes 7 and
8 were positioned to provide temperatures within the
undercut and haulage drifts. Probe 12 was installed to
provide temperatures at the footwall. An external ther-
mocouple (Omega precition fine wire teflon type T
thermocouple), connected to a Omega Programmable
Digital Thermocouple Meter, was also used to monitor
the temperature in the control room.
A 12 channel Digi-Sense scanner was used for mon-
itoring of temperatures. The scanner had such features
as user defined scan and log rates, individually pro-
grammable channels and alarm output. The scanner
was connected to a computer via bidirectional RS-
232 for data transmission, storage, and real-time data
graphing. Data acquisition was set at 1 sample every 10
minutes per channel to provide a detailed temperature
profile over the testing period.
2.2 Backfill preparation
Cemented backfill was prepared using unclassified
tailings supplied by an underground gold mine from
its mill operation. The received material had a mois-
ture content approximating 19%. Particle size analysis
conducted using a Laser Particle Size Analyzer indi-
cated a uniformity coefficient (D60 /D10 ) of 1.28 and
a coefficient of curvature (D230 /D10 × D60 ) of 0.96.
Approximately 80% of the size distribution is minus
200 mesh (75 microns). Material characterization tests
indicated the dry tailings material to have a bulk
density of 1760 kg/m3 , a porosity of 36%, cohesion
averaging 35 kPa and an angle of friction of 36 degrees.
Three cemented backfill recipes were designed at
25, 5 and 10% cement contents and 76% solids.
Approximately 58 kg of backfill was required for
each batch to fill the model test stope (0.028 m3 ).
For the 2.5% cement backfill, the recipe included
40.35 kg of dry tailings, 1.01 kg of portland cement
and 13.06 kg of water. For the 5% cement backfill,
the recipe included 40.35 kg of dry tailings, 2.02 kg of
 Figure 1. Scaled physical model.
portland cement and 13.38 kg of water. For the 10%
cement backfill, the recipe included 40.35 kg of dry
tailings, 4.04 kg of portland cement and 14.02 kg of
water.
Glassfill was prepared using the same tailings mate-
rial to cemented backfill. Three backfill recipes were
designed at 25, 5 and 10% binder contents, at 50%
cement and 50% glass contents, and 76% solids. Waste
glass, pulverized to minus 325 mesh (45 µm) was
used. For the 2.5% binder backfill, the recipe included
245
Copyright © 2006 Taylor & Francis Group plc, London, UK
40.35 kg of dry tailings, 0.5 kg of portland cement,
0.5 kg of glass and 13.06 kg of water. For the 5%
binder backfill, the recipe included 40.35 kg of dry
tailings, 1.01 kg of portland cement, 1.01 kg of glass
and 13.38 kg of water. For the 10% binder backfill,
the recipe included 40.35 kg of dry tailings, 2.02 kg
of portland cement, 2.02 kg of glass and 14.01 kg of
water.
Figure 3 presents the different phases of backfill
preparation, placement and curing.

Figure 2. Temperature probe position configuration.
3 BACKFILL HEAT TRANSFER
The flow of exposed cemented backfill heat into mine
airways depends on a number of parameters including,
cement type and cement content; tailings material;
water to cement ratio; cure time; thermal properties of
246
Copyright © 2006 Taylor & Francis Group plc, London, UK
the backfill; the length and geometry of the opening;
and the volume of air flow. The treatment of heat
transfer in mines is very complex; the basic processes
of heat transfer include conduction, convection and
radiation.
3.1 Heat flow from backfill to the surrounding rock
The heat flow from backfill to the surrounding rock is
due to conduction,
where Qf = heat flow in Watt; A = area in m2 ;
L = thickness in m; K = thermal conductivity in
W/m.K; and T = temperatures in K.
The heat flux (Qf /A) is given by,
where H = heat flux in Watt/m2 .
Using the above model, the thermal conductivity of
the backfill can be determined as,
where Kfill and Krock are the thermal conductivities of
the fill and surrounding rock and T1 , T2 and T3 are
temperatures within the fill, at the interface between
the fill and the stope wall and within the rock mass.
3.2 Heat flow from backfill to the mine air
The heat flow process to the mine air is due to
convection,
where Qf = heat flow in Watt; hc = heat transfer coef-
ficient in W/m2 K; A = rock surface area in m2 ;
Ts = surface backfill temperature in K and Ta = air
dry-bulb temperature in K.
The heat pick up by the air, Qf , can be calculated
from,
where Mf = mass flow of dry air in kg/s; Cp = specific
heat of air in J/kg◦ C and t = backfill and air tempera-
tures in ◦ C.
4 EVALUATION OF BACKFILL HEAT LOAD
The 10%, 5% and 2.5% cement backfill materials
were monitored over 28 day, 56 day and 32 day cure
processes, respectively, and the glassfill products were
monitored over 7 day cure processes. Trends in model
profile temperatures were analysed for each backfill
 Figure 3. Backfill preparation, placement and curing.
product. Using the developed thermal models, varia-
tions in thermal conductivity of the backfill during the
curing process were determined and the backfill heat
flux was estimated. Static heat transfer from the fill to
the undercut, overcut and haulage drift air were also
estimated. A summary of the observations and cal-
culations of cemented backfill and glassfill heat load
properties is presented in Table 1.
The testing program indicated that the 10%, 5%
and 2.5% cement backfill materials experienced a
247
Copyright © 2006 Taylor & Francis Group plc, London, UK
change in temperature of 5.2◦ C, 4.1◦ C and of 5.0◦ C
during the hydration process, respectively. Average
thermal conductivities of 2.65 W/m.K, 2.47 W/m.K
and 2.88 W/m.K were determined for the 10%, 5%
and 2.5% cement fill materials. The heat flux for the
10% cement backfill was estimated at 38 kW/m2 , for
the 5% cement backfill at 33 kW/m2 , and for the 2.5%
cement backfill at 16 kW/m2 .
In comparison, the 10%, 5% and 2.5% glass-
fill materials experienced a change in temperature
 Table 1. Summary of backfill heat contribution.

10% binder backfill 5% binder backfill 2.5% binder backfill

Property C C/G C C/G C C/G

Backfill temperatures (◦ C) max. 27.5 24.5 max. 25.1 25.8 max. 24.8 24.6
min. 22.3 20.3 min. 20.5 21.1 min. 19.8 20.6
diff. 5.2 4.2 diff. 4.1 4.7 diff. 5.0 4.0
Ore temperatures (◦ C) max. 26.6 23.7 max. 24.8 25.1 max. 24.8 24.0
min. 23.5 19.3 min. 20.8 22.1 min. 21.8 20.6
diff. 3.1 4.4 diff. 4.0 3.0 diff. 3.0 3.4
Footwall temperatures (◦ C) max. 26.2 23.8 max. 24.8 25.0 max. 24.4 23.3
min. 23.7 19.3 min. 20.7 21.8 min. 22.2 20.5
diff. 2.5 4.5 diff. 4.1 3.2 diff. 2.2 2.8
Overcut temperatures (◦ C) max. 27.6 24.7 max. 25.8 25.6 max. 25.1 22.9
min. 22.6 19.8 min. 21.3 22.3 min. 22.0 21.0
diff. 5 4.9 diff. 4.5 3.3 diff. 3.1 1.9
Undercut temperatures (◦ C) max. 25.6 28.3 max. 24.3 24.1 max. 24.1 23.6
min. 23.3 19.2 min. 20.3 22.0 min. 22.0 20.1
diff. 2.3 4.1 diff. 4.0 2.1 diff. 2.1 3.5
Backfill thermal conductivity (W/m.K) 2.65 1.67 2.47 2.02 2.88 2.29
Heat flux from fill to ore (kW/m2 ) 38 7.9 33 13.8 16 7.1
Heat flux from fill to footwall (kW/m2 ) 10.8 6.6 13 9.5 14.5 5.7
Heat transfer to overcut drift (kJ/kg) 5 2.9 4.5 1.7 3.1 3.5
Heat transfer to undercut drift (kJ/kg) 2.3 1.5 4 0.9 2.1 1.8
Heat transfer to haulage drift (kJ/kg) 2.3 4.6 4.3 0.9 5.2 13.2

of 4.2◦ C, 4.7◦ C and of 4.0◦ C during the hydration
process, respectively. Lower average thermal conduc-
tivities of 1.67 W/m.K, 2.02 W/m.K and 2.29 W/m.K
were determined for the 10%, 5% and 2.5% glass-
fill materials. The heat flux for the 10% glassfill was
estimated at 7.9 kW/m2 , for the 5% cement backfill
at 13.8 kW/m2 , and for the 2.5% cement backfill at
7.1 kW/m2 . The above information clearly indicates
the potential reduction in heat load using glassfill.
For the 10% cement backfill, the overcut drift expe-
rienced a maximum temperature of 27.6◦ C; the static
heat transfer from the fill to the air was estimated at
5 kJ per kg of dry air. The undercut drift experienced a
maximum temperature of 25.6◦ C; the static heat trans-
fer from the fill to the air was estimated at 2.3 kJ per kg
of dry air. The footwall rock experienced an increase
in temperature of 2.5◦ C; the heat flux into the footwall
was estimated at 10.8 kW/m2 .
In comparison, for the 10% binder glassfill, the
overcut drift experienced a maximum temperature of
24.7◦ C; the static heat transfer from the fill to the air
was estimated at 2.9 kJ per kg of dry air. The undercut
drift experienced a maximum temperature of 28.3◦ C;
the static heat transfer from the fill to the air was esti-
mated at 4.6 kJ per kg of dry air. The footwall rock
experienced an increase in temperature of 4.5◦ C; the
heat flux into the footwall was estimated at 6.6 kW/m2 .
For the 5% cement backfill, the overcut drift
experienced a maximum temperature of 25.8◦ C; the
static heat transfer from the fill to the air was estimated
at 4.5 kJ per kg of dry air. The undercut drift experi-
enced a maximum temperature of 24.3◦ C; the static
heat transfer from the fill to the air was estimated at
4 kJ per kg of dry air. The footwall rock experienced
an increase in temperature of 4.1◦ C; the heat flux into
the footwall was estimated at 13 kW/m2 .
Comparatively, for the 5% binder glassfill, the
overcut drift experienced a maximum temperature of
25.6◦ C; the static heat transfer from the fill to the air
was estimated at 1.7 kJ per kg of dry air. The undercut
drift experienced a maximum temperature of 24.1◦ C;
the static heat transfer from the fill to the air was esti-
mated at 0.9 kJ per kg of dry air. The footwall rock
experienced an increase in temperature of 3.2◦ C; the
heat flux into the footwall was estimated at 9.5 kW/m2 .
For the 2.5% cement backfill, the overcut drift expe-
rienced a maximum temperature of 25.1◦ C; the static
heat transfer from the fill to the air was estimated at
3.1 kJ per kg of dry air. The undercut drift experi-
enced a maximum temperature of 24.1◦ C; the static
heat transfer from the fill to the air was estimated at
2.1 kJ per kg of dry air. The footwall rock experienced
an increase in temperature of 2.2◦ C; the heat flux into
the footwall was estimated at 14.5 kW/m2 .
In comparison, for the 2.5% binder glassfill, the
overcut drift experienced a maximum temperature of
22.9◦ C; the static heat transfer from the fill to the air
was estimated at 3.5 kJ per kg of dry air. The undercut

248

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 2. Percent change in backfill heat properties. cemented backfill and by glassfill during hydration
and the heat flux to the hanging wall, orebody, footwall
10% binder 5% binder 2.5% binder and cross-cut and production drifts of a scaled stope
backfill backfill backfill model. For the tested cemented backfills, heat flux
to adjacent rock reaching 38 kW/m2 and heat trans-
Thermal −37% −18.2% −20.5%
fer to mine air reaching 5 kJ per kilogram of air were
conductivity
Heat flow from −79.2% −58.2% −55.6% determined. Such values indicate that backfill materi-
fill to ore als may represent a major contributor to the heat load
Heat flow to −38.9% −26.9% −69% in underground mines and should be incorporated in
footwall mine climate engineering planning exercises. Glassfill
Heat transfer of −42% −62.2% +12.9% was introduced as a possible means of reducing the
overcut drift heat load in mines. For the tested glass based back-
Heat transfer −34.8% −77.5% −14.3% fills, heat flux to adjacent rock reaching 13.8 kW/m2
undercut drift and heat transfer to mine air reaching 4.5 kJ per kilo-
Heat transfer to +100% −79.1% +153.8%
gram of air were determined. Average reductions in
haulage drift
heat flux of 55% and in heat transfer to mine air of
52% were achieved.
drift experienced a maximum temperature of 23.6◦ C;
the static heat transfer from the fill to the air was esti- REFERENCES
mated at 1.8 kJ per kg of dry air. The footwall rock
experienced an increase in temperature of 2.8◦ C; the De Souza, E. & Archibald, J.F. 2005. The Strength and
heat flux into the footwall was estimated at 5.7 kW/m2 . Flow Performance of Glassfill Technology. Toronto Rocks
Table 2 shows percent changes in thermal con- 2005 – Mining Industry Conference and Exhibition –
ductivity, heat flux and heat transfer when cemented Annual General Meeting of CIM, Toronto, Ontario. pp. 8.
fill is replaced with glassfill. Positive values repre- De Souza, E. Archibald, J.F., Dirige, P. & Sargeant, A.
sent increase in thermal properties and negative values 2005. Glassfill – A Promising Mine Backfill Alternative.
Alaska Rocks 2005 – Rock Mechanics for Energy, Mineral
represent reductions in thermal properties.
and Infrastructure Development in the Northern Regions,
Tables 1 and 2 clearly indicate that backfill can Anchorage, Alaska.
represent a major contributor to the heat load in under- De Souza, E. & Archibald, J.F. 2004. Environmental Ben-
ground mines, with heat loads equivalent to other efits Deriving from the Application of Post-Consumer
major sources of mine heat, and that glassfill may con- Glass as a Cementing Agent in Mine Backfill. COM 2004,
stitute an effective means of reducing backfill heat in Second International Symposium on Greenhouse Gases
the underground environment. in the Metallurgical Industries: Policies, Abatement and
Treatment. Hamilton, Ontario: 27–42.
De Souza, E. & Hewitt, A. 2005. The Contribution of
Cemented Backfill to Heat Loads in Mines. 8th Inter-
5 CONCLUSIONS
national Mine Ventilation Congress, Brisbane, Australia:
87–93.
This paper presented the results of a series of inves-
tigations aimed at quantifying the heat generated by

249

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Thermo- and psychrometric properties of intake air passing through

fragmented strata

O.S. Johnson
Johnson Consulting, Sudbury, Ontario, Canada

ABSTRACT: This paper shows the prediction of thermo- and psychrometric properties of the intake air passing
through a mass of pit muck which acts as a heat exchanger. Such area’s temperature field makes influence on
the temperature of the intake air heating this air in winter and cooling it in summer. The study was conducted for
an area between ground surface and mine opening. The intake naturally conditioned air passing into the mine
through this area changes its properties along airways. The ground surface temperature changes according to the
weather conditions. The temperature in the mine is constant during the year. The amount of the heat exchanged
between broken rock and the conditioned air in capillaries is governed by the general theory developed to assess
heat transfer at dry and wet surfaces, also heat and mass transfer in capillary-porous bodies theory. The results
shown in this paper are available for Sudbury (Canada) region.

1 INTRODUCTION capillaries, changes relative humidity of the air and the

This study shows the prediction of the change of ther-
mal and psychrometric properties of the intake mine
air under the air temperature and pressure influence.
The gathering area acts as a heat exchanger. The
fresh air is drawn down from the surface through a
mass of broken rock. The broken rock acts as a heat
exchanger warming air in winter and cooling it in sum-
mer. The temperature on the surface is variable and
depends on the weather conditions. The temperature in
the mine is stationary all the year round and varieties
about +2◦ C in winter to +4◦ C in summer. In winter
the air which has negative temperature goes from the
surface to the mine. This air is naturally warmed by
a mass of pit muck which temperature is positive. In
summer the hot air from the surface goes through cap-
illaries to the mine. It is cooled by fragmented rock
which temperature is positive but lower than the air
temperature. When air is driven down its thermal and
psychrometric properties change with the change of
its temperature.
This study investigates the intake air’s thermo-
and psychrometric properties change under different
temperature influence. Temperature of the air was
determined according to McPherson (1992) general
theory involving the Reynolds, Biot, Nusselt, and
Fourier numbers. This theory is applied to air flow-
ing through the system of capillaries located in the
fragmented rock. The water from melted ice wets the
amount of latent heat accompanied by condensation
or an evaporation process. CLIMSIM for Windows
(2000) software program was used for calculations.
The results shown here are available for average
weather conditions in Sudbury (Canada) region.
2 PROBLEM DESCRIPTION
The study consists of three parts: determination of the
air temperature influence on the air thermal properties
for winter and summer conditions, determination of
the air temperature influence on the air psychrometric
properties for summer conditions and determination
of the air pressure influence on the air psychromet-
ric properties for summer conditions. In winter the
upper layer of the broken rock contains ice and has
negative temperature. The lower layer has positive
temperature. All the broken rock area has positive tem-
perature in summer months. The air is driven down
from the surface to the mine through the broken rock
area and changes its temperature and properties under
the pit muck temperature influence. The empirical
equations were developed for all parts to give the pos-
sibility of the intake air properties determination. All
the parts include graphs which show the relationship
between thermal and psychrometric air properties, air
temperature and pressure.

251

Copyright © 2006 Taylor & Francis Group plc, London, UK

 3 THE INFLUENCE OF THE AIR
TEMPERATURE ON ITS THERMAL
PROPERTIES

3.1 Empirical equations

Study of the air temperature influence on its proper-
ties (such as: specific heat, thermal conductivity, and
thermal diffusivity) was conducted using the John-
son (2004) method and Luikov (1966) numerical data
for winter and summer weather conditions: tempera-
ture intervals are from +10◦ C to +100◦ C and from
−20◦ C to −100◦ C; the air pressure is constant and
equal 101.325 kPa.
The thermal conductivity change was studied for
negative and positive air temperature. The thermal
conductivity λ for any negative temperature Tn can
be determined according to the developed empirical Figure 1. Thermal conductivity change depending on air
formula: temperature.

Thermal diffusivity for positive temperature can be

where λn = thermal conductivity for the negative air determined as:
temperature; Tn = negative temperature of the air.
For positive temperature the formula for the deter-
mination of the thermal conductivity is:
where ap = thermal diffusivity for the positive air
temperature.

where λp − thermal conductivity for the positive air

temperature; Tp − positive temperature of the air.
Formulas (1) and (2) show the air thermal conduc-
tivity change under the air temperature influence and
constant air pressure.
This study also includes the air specific heat change
according to the air temperature and constant pressure.
Specific heat for the negative air temperature zone can
be determined according to the developed empirical
equation in the interval: from −20◦ C to −100◦ C:
where cn = specific heat for the negative air
temperature.
For positive air temperature the equation in the
interval: from +10◦ C to +100◦ C can be written as:
where cp = specific heat for the positive air
temperature.
Thermal diffusivity change can be described
according to the developed empirical formula:
where an = thermal diffusivity for the negative air
temperature.
3.2 Results
The results of calculations according to the developed
empirical formulas from (1) to (6) using created by
O. S. Johnson Microsoft programs and CLIMSIM for
Windows (2000) software program are shown below
for the following data for winter and summer weather
conditions:
air temperatures: from −100◦ C to +100◦ C;
air pressure = 100 kPa; rock density = 2000 kg/m3 .
Figure 1 shows the air thermal conductivity change
under the air temperature influence and constant air
pressure.
The graph consists of two parts: positive and neg-
ative temperature areas. Both parts are considered
straight lines but the angles between the axis of ther-
mal conductivity λ and the right and the left parts
of the graph are different. It investigates that posi-
tive and negative temperature make different influence
on the air thermal conductivity change. The graph
shows that the air thermal conductivity grows as the
air temperature is increased.
Figure 2 presents the air specific heat change
according to the air temperature and constant pressure.
The graph consists of two unsymmetrical parts: neg-
ative air temperature part and positive one. The left part
of the graph shows the decreasing of the specific heat
when the air temperature is close to zero. The right

252

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 2. Specific heat change depending on air
temperature.
Figure 3. Thermal diffusivity change depending on air
temperature.
part of the graph shows the increasing of the air spe-
cific heat as the positive air temperature is increased.
Between +10◦ C and −20◦ C specific heat of the air is
stationary and equal 0.00028 W/kg◦ C.
Figure 3 shows the air thermal diffusivity change
depending on the air temperature.
The graph consists of two parts: positive and nega-
tive temperature zones. In both zones the air thermal
diffusivity is growing when the air temperature grows.
The angles between thermal diffusivity axis and upper
and lower parts of the graph are different what inves-
tigates the different speed of air temperature change
influence on its thermal diffusivity.
The developed empirical formulas from (1) to (6)
give the possibility to calculate thermal characteristics
of the intake air for winter and summer conditions.
253
Copyright © 2006 Taylor & Francis Group plc, London, UK
4 THE INFLUENCE OF AIR TEMPERATURE
ON ITS PSYCHROMETRIC PROPERTIES
4.1 Empirical equations
The way the psychrometric properties (moisture con-
tent, density, humidity, enthalpy, and sigma heat) of
the air change when the air temperature changes was
studied using Johnson (2004) method. New empirical
equations were developed to calculate psychrometric
properties of the intake air. The air pressure is constant
and equal 100 kPa.
The newly developed following equation describes
the air moisture content change according to the air
temperature:
where X = the air moisture content; T = the air
temperature.
The equation for air density is the following:
where D = the air density.
Several temperature intervals of air humidity deter-
mination were studied. The air humidity depending on
the air temperature for the temperature interval: from
+1.5◦ C to +4.0◦ C can be described by the following
empirically developed formula:
where H1 − the air humidity for the temperature inter-
val: from +1.5◦ C to +4.0◦ C.
For the temperature interval from +4◦ C to +5.5◦ C
humidity H2 is constant.
For the temperature interval from +5.5◦ C to
+6.5◦ C the newly developed equation shows how the
air humidity changes according to the temperature
change:
where H3 = the air humidity for the temperature inter-
val: from +5.5◦ C to +6.5◦ C.
The air enthalpy changes depending on air temper-
ature according to the formula:
where En = the air enthalpy.
The empirical formula presented below describes
the connection between the air sigma heat and the air
temperature:
where Si = the air sigma heat.
The presented equations show the strong connection
between the psychrometric properties of the intake air
and the air temperature.
 Figure 5. Air density change depending on temperature.
Figure 4. Moisture content change depending on
temperature.

4.2 Results
Calculations to determine the air moisture content,
density, humidity, enthalpy, and sigma heat change
under the influence of different positive temperatures
(April to July) but the constant air pressure have
been done using Johnson (2004) method and data for
summer weather conditions in Sudbury (Canada) area:
• thickness of the gathering area = 50 m;
• air thermal conductivity = 3.0 W/m◦ C;
• air thermal diffusivity = 1.4 · 10−7 m2 /s;
• wetness factor = 0.01;
• geo step = 50 m/◦ C;
• atmospheric pressure = 100 kPa;
• capillary area = 0.5 m2 ;
• airflow = 0.1 m3 /s;
Figure 6. Humidity change depending on air temperature.
• mine opening rock temperature = +4◦ C;
• dry intake air bulb temperature for:
• April = +3◦ C;
three parts: 1 – upper left part, 2 – horizontal part and
• May = +10.8◦ C;
3 – lower right part. Upper left part is a straight line
• June = +15.8◦ C;
for the following temperature interval: from +1.5◦ C to
• July = +19.1◦ C.
+4.0◦ C. The horizontal part of the graph exists for the
The air temperature in the mine is lower than the air following temperature interval: from +4◦ C to +5.5◦ C.
temperature on the ground surface. Humidity here is constant and equal H2 = 91.29%. The
Figure 4 shows the air moisture content change right lower part of the graph is a straight line for the
according to the air temperature. temperature interval: from +5.5◦ C to +6.5◦ C.
The graph reflects the air moisture content increas- Figure 7 presents the air enthalpy change according
ing when the air temperature grows. to the air temperature.
Figure 5 presents the air density change under the The graph is a straight line. The air enthalpy grows
air temperature influence. when the air temperature grows.
The graph consists of a straight line showing the The graph of the air sigma heat change is also a
air density decreasing when the air temperature is straight line what is shown in Figure 8.
growing. Figure 8 shows the growth of the air sigma heat
The air humidity change according to the air tem- according to the growing temperature.
perature is shown in the Figure 6. The presented in this part of the paper results show
The graph shows that the air humidity is decreasing the strong connection between the air psychrometric
when the temperature grows. This graph consists of properties and the air temperature.

254

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 7. Enthalpy change depending on air temperature.
Figure 8. Sigma heat change depending on air temperature.
5 THE INFLUENCE OF THE AIR PRESSURE
ON ITS PSYCHROMETRIC PROPERTIES
5.1 Empirical equations
The following empirically developed equations show
the air psychrometric properties dependence on the air
pressure. The studied interval of atmospheric pressure
is from 40 kPa to 300 kPa. The intake air temperature
is considered stationary for July in Sudbury (Canada)
area.
The air moisture content can be determined accord-
ing to the empirical formula:
where Xp = the air moisture content; P = the air
pressure.
The formula written above can be used for moisture
content determination under different atmospheric
pressure starting 100 kPa.
255
Copyright © 2006 Taylor & Francis Group plc, London, UK
The developed empirical formula for air density
determination is:
where Dp = the air density; P = the air pressure.
The formula for the air humidity determination is
the following:
where Hp = the air humidity; P = the air pressure.
This formula can be used for the air pressure from
100 kPa to 250 kPa.
The conducted research shows that there are three
intervals of the air pressure which are satisfied by
equations listed below. The first equation presents
the connection between the air enthalpy and the air
pressure in the interval of 40 kPa to 110 kPa:
where Ep1 = the air enthalpy; P = the air pressure.
For the interval of the air pressure 230 kPa to
300 kPa the formula is:
where Ep2 = the air enthalpy; P = the air pressure.
In the interval of the air pressure 110 kPa to 230 kPa
the enthalpy is constant and can be determined using
formula (16).
The same intervals of the air pressure are avail-
able for sigma heat determination. For the air pressure
40 kPa to 110 kPa the equation is:
where Sp1 = sigma heat; P = the air pressure.
For the interval of the air pressure 230 kPa to
300 kPa the equation is:
where Sp2 = sigma heat; P = the air pressure.
In the interval of the air pressure 110 kPa to 230 kPa
the sigma heat is constant and can be determined using
formula (18).
This study also includes the determination of the
connection between the air temperature and the air
pressure. This formula is shown below:
where T = the air temperature; P = the air pressure.
The empirical formulas developed in this part of the
paper give the possibility to determine the air psychro-
metric properties for the stationary air temperature and
the changeable air pressure.
 Figure 9. Moisture content change depending on air
pressure.
Figure 10. Density change depending on air pressure.
5.2 Results
The data used in these calculations is the same as in
the part 4.2. The air pressure is variable: from 40 kPa
to 300 kPa.
Figure 9 shows the air moisture content change
under the different atmospheric pressure conditions.
The curve investigates the decreasing of the air
moisture content when the atmospheric pressure is
increasing.
Figure 10 presents the air density change under the
different atmospheric pressure conditions.
The graph is a straight line. The air density is
increasing when the air pressure grows.
The graph of the air humidity change under the
different atmospheric pressure is shown in Figure 11.
The curve can be considered a straight line between
the air pressure: 100 kPa to 250 kPa. It can be seen
from the graph that the air humidity is decreasing with
the atmospheric pressure increasing.
256
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 11. Air humidity change depending on pressure.
Figure 12. Enthalpy change depending on air pressure.
The next graph presents the connection between the
air enthalpy and the air atmospheric pressure.
Figure 12 shows a curve.
The curve has two parts: 1 – the upper left part;
2 – the lower right part which is close to a horizon-
tal line. This graph investigates that the air enthalpy
is decreasing when the air pressure grows in the first
part of the graph before the point P = 110 kPa. This
point can be called the extreme point and the further
increasing of the air pressure leads to a small increas-
ing of the air enthalpy in the second part of the graph
after this point.
Figure 13 shows the air sigma heat change under
the influence of the air pressure.
The picture of the graph looks like the previous one.
The extreme point here is also P = 110 kPa.
Figure 14 shows the connection between the air
temperature and the air pressure.
This is a straight line. The air temperature increases
when the air pressure grows.
 weather conditions. Positive and negative air tem-
perature does not make the same influence on the
thermal characteristics of the air. Thermal conductiv-
ity and thermal diffusivity grow when the temperature
is growing. Specific heat is decreasing when the air
negative temperature is close to zero and the opposite
effect exists when the positive air temperature grows.
In this case the air specific heat is increasing.
The air psychrometric properties also depend on the
weather conditions. Moisture content, enthalpy, and
sigma heat of the air grow when the air temperature is
growing. The air density and humidity are decreasing
when the air temperature is growing. When the air
pressure is growing the air density also grows. But
the air moisture content, humidity, enthalpy and sigma
heat are decreasing.
Figure 13. Sigma heat change depending on air pressure.
The equations presented in this paper give possibil-
ities for the easy determination of the air thermo-and
psychrometric properties.

REFERENCES
Gulberg, J. 1997. Mathematics from the birth to numbers.
New York, London: W.W. Norton & Company Inc.
Johnson, O.S. 2004. Determination of natural heating and
cooling of intake air in underground mines, M.A.Sc. thesis
(published). Sudbury: Laurentian University.
Luikov, A.V. 1966. Heat and mass transfer in capillary-
porous bodies. Oxford, London, Edinburgh, New York,
Paris, Frankfurt: Pergamon Press Ltd.
McPherson, M.J. 1992. Subsurface Ventilation and Environ-
mental Engineering. London, Glasgow, New York, Tokyo,
Melbourne, Madras: Chapman and Hall.
Mine Ventilation Services Inc. 2000. CLIMSIM for Windows
Version 1.0, Users Manual and General Theory. Fresno:
Figure 14. Relationship between air pressure and Mine Ventilation Services Inc.
temperature.

The shown results prove the strong dependence of

the air psychrometric properties on the air pressure.

6 CONCLUSIONS

The research results presented in this paper show that

the air thermal characteristics strongly depend on the

257

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Controlling personnel heat exposure in Canada’s deep and

highly mechanized mines

S. Hardcastle
CANMET-MMSL, Natural Resources Canada, Sudbury, Canada

ABSTRACT: Canadian mining companies, with ore reserves at depths of >3000 m below surface, have iden-
tified controlling heat exposure as a major technical concern. At issue are an appropriate guideline and/or
standard that can be used to determine whether a working environment is safe and the magnitude of ventilation
and/or refrigeration that may be required. Consequently, exploring heat exposure issues is a major project within
Canada’s Deep Mining Research Consortium (DMRC). The DMRC is jointly funded by mining companies
and all levels of government, municipal, provincial and federal. Additional members of the Consortium include
Federal research organizations, university academia and Provincial health and safety regulators. In Canada, the
wet-bulb globe temperature (WBGT) is the most common index that is currently being used to determine the
suitability of a work environment and the need for a work-rest regimen. However, in mines, its application is
being questioned where workers benefit from the cooling effect of a significant air velocity, have the ability
to self-pace their work, and could be working in air conditioned enclosures. Other questions/concerns include:
adequate protection for short-lived intense exertion; the effect of personal protective equipment; reduced cooling
potential of denser air at depth; and acclimation, especially in winter when workers experience diurnal variations
of up to 60◦ C. This paper describes a four year project that is underway including: a review of heat exposure
guidelines; a mechanical and energy expenditure characterization of standard mining tasks; a survey of the
environmental conditions in mines; laboratory simulation of the tasks under controlled conditions; laboratory
evaluation of heat guidelines, their field validation; acclimation studies and an instrumentation evaluation.

1 BACKGROUND managed by CANMET-MMSL as described in this

In Canada, the increasing need to combat personnel
heat exposure in mines, as a function of depth, is
becoming an increasingly important issue to the extent
that it can become the overriding criterion in ventila-
tion system design. Hence, due to the cost of ventila-
tion and refrigeration control measures, heat exposure
issues could limit the depth to which Canada’s exist-
ing mines can continue to be economic and potentially
rich reserves may be left in the ground. In addition,
the optimization of power usage, either for ventilation
or for refrigeration, would benefit Canada as a whole
through reduced greenhouse gas emissions.
Consequently, ways to prolong a deep mine’s viabil-
ity and the northern economies they support becomes
very important. This and other depth related chal-
lenges to the industry resulted in the formation,
in Canada in 2003, of the Deep Mining Research
Consortium (DMRC) with a $7M(Can) budget over
five years. One of its major projects, currently in
progress, is a $1.5M(Can) four year heat research study
paper.
1.1 Canada’s Deep Mining Research Consortium
(DMRC)
The DMRC consists of individual companies inter-
ested in solving problems related to mining at depth
in Canada. The member mining companies include:
Agnico-Eagle, Barrick Gold, Falconbridge (Sudbury
operations & Kidd Creek), Goldcorp, Inco (Ontario
& Manitoba operations), Placer Dome and Rio Tinto.
An interesting feature of the DMRC is the non-
industry support which has been pledged by all three
levels of government – municipal, provincial and
federal, through the following: Natural Resources
Canada’s Mining and Mineral Sciences Labora-
tories (CANMET-MMSL), Industry Canada’s fed-
eral regional development organizations for northern
Ontario (FedNor) and Développement économique
Canada in Quebec, Ontario’s Ministry of Northern
Development and Mines’ Northern Ontario Heritage

259

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Fund Corporation (NOHFC) and the City of Greater
Sudbury, a mining based community in Northern
Ontario.
1.2 DMRC project justification
Hot environments will almost be impossible to avoid
in deep mining (Marx, 1998). The challenge is to
understand and control that environment.
At this time, the economic benefits of a better
understanding of the heat exposure and stress issues
are unknown. However, the ever-increasing need to
address the “heat” is evident in how ventilation
demands are changing, and the need for refrigera-
tion, as a function of increasing depth. For example, at
Inco’s Creighton mine which has a virgin rock temper-
ature (VRT) of 48◦ C at 2400 m (Thibodeau, 1999), to
combat additional heat from the strata, it has been esti-
mated that airflows would have to increase by 20% as
mining moves from 2000 m to 2300 m below surface.
If the supply system remained unchanged, the associ-
ated power and operating costs could increase by 73%,
consequently the mine is exploring demand-based ven-
tilation (Hardcastle et al, 2006). Accommodating such
increases can also involve significant capital outlay in
infrastructure namely as primary fans, as has already
taken place at Creighton mine (O’Connor et al, 2002),
or alternatively or in conjunction with improvements
in the delivery system, i.e. ventilation raises. At Fal-
conbridge’s Kidd Mining Division operation with pro-
jected VRTs of 35◦ C at 3000 m (MacIssac, 2001), the
mine has spent $12M(Can) upgrading its existing ven-
tilation system to permit mining of its deep D-orebody
(Howes & Sedlacek, 2001; Hortin & Sedlacek, 2002)
and is also installing a 12 MWR cooling system
(Howes & Hortin, 2005) at an additional cost of about
$9.0M(Can). A similar analysis was performed for
Falconbridge’s Onaping Depth (Howes & Stockhaus,
2002), which has a VRT of 45◦ C at 2400–2500 m. At
Agnico Eagle’s Laronde mine, with a VRT of 38◦ C at
2000 m, a total of 5MWR of mechanical cooling was
installed underground in 2002.This was the first instal-
lation in Canada of a mechanical refrigeration system
to control heat exposure. In the summer of 2004,
this was relocated to surface and supplemented to
12 MWR at an additional cost of $6.6M(Can). How-
ever, depending upon the heat design criteria used, it
may be necessary to increase the capacity by another
50% to 18 MWR in the near future.
These numbers highlight the degree of financial
commitment needed to maintain an equitable work-
ing climate to continue mining at depth and how it can
be dependant upon the design criteria. However, the
major capital outlay is not the only issue. The routine
operating cost of the ventilation/refrigeration systems
can be $5–10M(Can) per year, and these costs are also
a function of the heat stress design criteria used to
determine worker comfort.
260
Copyright © 2006 Taylor & Francis Group plc, London, UK
The significant cost involved was confirmed in
another study funded by the DMRC (Campbell, 2005),
indicating that continuing with current practices, ven-
tilation (including refrigeration) and ore transportation
would experience the greatest escalation in cost with
increased depth.
1.3 Heat exposure, stress and strain
Heat exposure is an issue because of the potential
for heat stress/strain amongst Canada’s underground
workforce with a resultant deterioration of safety and
production. The heat levels experienced, which could
lead to the heat stress/strain, are a result of the increas-
ing heat contributions from the strata, autocompres-
sion of air with depth, and from the mining machinery.
Typically, this heat addition is addressed through either
dilution as per other common contaminants, or through
cooling/refrigeration of the air.
In Canada, similar to other mining countries the
information available on the incidence or frequency
of heat stress within the industry is extremely limited
despite increasing numbers of workers being exposed
to hot, potentially stressful, environments. To some
degree, the failure to recognize a heat related condi-
tion is limited by the non-specific symptoms, and most
treatment would only be reported as a “first-aid”. In
Ontario which has Canada’a largest number of mines,
statistics from 1988–2004 listed between 5 and 98 inci-
dences per year of heat related injuries resulting in
lost time across all industries (WSIB, 1997, 2004).
Across Canada, from 1994–96, Workers Compensa-
tion Boards reported 224 lost-time claims, again across
all industry attributable to heat (Chandler, 1998).
In regard to mining, the studies showing the
increased incidence of injuries, or loss of performance
with increasing temperature, which can be a function
of depth, include:
• Van der Walt et al (1981) found the incidence of heat
stroke was related to wet-bulb temperatures.
• Howes and Nixon (1997) found an association
between air cooling powers of less than 300 W/m2
with a decrease in mine productivity.
• Misaqi et al (1976) noted that dexterity, co-ordina-
tion, observation, alertness and the decision process
were negatively influenced by heat stress. Further-
more, safety was further compromised with adverse
effects in worker irritation and anger.
Currently, most jurisdictions in Canada follow the
American Conference of Governmental Industrial
Hygienists TLVs® based upon the wet-bulb globe tem-
perature (WBGT) as a heat stress screening criteria
and to determine work/rest regimens (ACGIH, 2005).
Within its application to the mining industry, what is at
issue is whether the WBGT provides adequate protec-
tion to the worker and accounts for the working condi-
tions specific, and possibly unique, to Canada’s deep
 and highly mechanized mines. Furthermore, when
these WBGT guidelines are used to design the mine-
wide ventilation/refrigeration requirements, through
guaranteeing safe conditions, they may greatly overes-
timate the actual heat stress potential (Howes & Nixon,
1997). Another question is whether Canadian mines
need mine-wide refrigeration for their small mobile
workforces potentially working from a machine with
a controlled environment.
2 THE DMRC “HEAT STRESS” PROJECT
The three-phase four-year $1.5M(Can) project was
developed by a sub-committee consisting of the
following:
• Mine representatives from the DMRC members.
• External technical expertise – mining: RHP Con-
sulting, CANMET-MMSL; medical: University of
Ottawa, Faculty of Health Sciences, School of
Kinetic Engineering.
• Labour health and safety representatives from the
United Steel Workers of America and the Canadian
Auto Workers.
• Regulatory and industrial spokespersons, from the
Ontario Ministry of Labour and the Mines and
Aggregates Safety and Health Association.
The project was submitted to the DMRC Technical
Committee and upon their acceptance forwarded to
the DMRC Management Committee for final approval
in 2004. The industry champion for this project was
Agnico-Eagle, its breakdown is as follows.
2.1 Phase 1 – background information
Phase 1 consists of a literature review of current
and emerging heat exposure guidelines, their basis,
required instrumentation, complexity, level of protec-
tion etc. and their suitability to the Canadian mining
scene. The sub-phases of this element are as follows.
2.1.1 Phase 1.1 – review of heat exposure
definitions/standards/instrumentation/
previous research
This review will comprise:
• A detailed listing of terms and definitions pertain-
ing to heat exposure/stress/strain and the medical
terminology.
• A literature review of historical and emerging heat
exposure standards, their applicability to the Cana-
dian mining scene, and level of protection afforded
workers.
• A document review of current heat exposure proce-
dures at the DMRC partner mines and an assessment
of their applicability, and safety.
261
Copyright © 2006 Taylor & Francis Group plc, London, UK
• A review of previous heat exposure research/ assess-
ments in the mining industry as pertinent to the
Canadian Industry.
• A search and review of commercial instrumentation
to evaluate the thermal environment and monitor
biomedical parameters on mining subjects.
• A written summation report identifying, the more
pertinent heat stress indices and the most appropri-
ate instrumentation to use.
2.1.2 Phase 1.2 – quantification of working
conditions and initial acclimation
assessment
This element will be a field based determination of the
range of working conditions that may be experienced at
the host mine, Agnico-Eagle’s Laronde property. This
determination will not only provide the data under
which the thermal indexes will be evaluated in the
laboratory within Phase 2, but will also be used at a lat-
ter stage to justify specific condition evaluations and
assess the impact of new thermal indices upon current
working conditions. This determination will include:
• A detailed listing of heat exposure conditions that
may contribute to heat stress.
• An evaluation of the environmental conditions with
respect to temperature (dry/wet/globe) humidity,
barometric pressure, air speed etc.
• An evaluation of the most stressful tasks with
respect to task intensity.
• An evaluation/categorization of the contributing
factors that play a role in creating the potentially
stressful condition.
• An initial evaluation of acclimation (this element
has yet to be defined pending the definition of
acclimation).
2.1.3 Phase 1.3 – evaluation of commercial
environmental instruments to assess
heat exposure
This phase will be both a laboratory and field based
evaluation of the most promising instruments to assess
heat exposure as a spot measurement or through con-
tinuous monitoring. It will evaluate each environmen-
tal instrument with respect to its ability to accurately
determine temperature, humidity, barometric pressure
and air velocity as applicable. The evaluation will then
determine the accuracy/suitability of the instruments
to determine more specific heat stress parameters. This
assessment will not exclude instrumentation that has
barometric pressure limitations; in this instance it will
assume that such instruments may be modified to meet
the demands of industry. This evaluation will include:
• A detailed listing of instruments tested along with
their associated direct measurements and indirect
determinations.

Figure 1. The Snellen Air Calorimeter with Human Subject.
• An evaluation of the instrument’s accuracy with
respect to measuring basic environmental param-
eters such as temperature (dry, wet, globe etc.),
humidity, barometric pressure, air speed, etc.
• An evaluation of the instrument’s accuracy with
respect to measuring any derived heat stress indices
along with the inclusion of such modifiers as cloth-
ing values, air velocity considerations, work rest
regimens, etc.
2.2 Phase 2 – validate an existing thermal stress
index that is appropriate and relevant to
assessing thermal stress in miners working in
canada’s deep mechanized mines
The second phase of the DMRC “heat stress” project
contains five sub-phases. This phase includes exten-
sive field measurements categorizing and measuring
underground tasks, their replication under controlled
laboratory conditions and then the laboratory sim-
ulation of extreme environments, concluding with
the field validation of a thermal index guideline, or
monitoring approach.
Unlike previous research into heat stress in mining,
this work will use a Snellen air calorimeter (Fig-
ure 1). Through controlling the volume, temperature
and humidity of the airflow entering the calorimeter,
and measuring the conditions of the exhaust air, the
calorimeter can accurately determine the sensible and
262
Copyright © 2006 Taylor & Francis Group plc, London, UK
evaporative heat loss of humans performing specific
tasks under specific conditions.
2.2.1 Phase 2.1 in-situ time/motion analysis of
selected mechanized mining tasks
Six occupational groups with tasks associated to typ-
ical mechanized mining practices were selected for
the in-situ time/motion analyses. The targeting of spe-
cific tasks was only to limit the size of the study;
furthermore, it did not reflect any assumption of heat
exposure, rather, it was orientated towards the more
common tasks/group with the better potential for sta-
tistical significance. The targeted occupational groups
were as follows:
i. Production drill (Jumbo/longhole) operation.
ii. Production scooptram (no a/c).
iii. Manual bolting (stoper).
iv. Manual shotcrete application.
v. Heavy construction.
vi. Conventional mining (track drift).
The time/motion analysis, performed by personnel
with a background in human kinetic/health sciences
would evaluate each of the tasks as they are performed
by typical workers under typical working conditions. If
possible, work locations where heat exposure is not an
issue were selected. The host mine for this analysis was
Agnico-Eagle’s Laronde mine. The analysis, spanning
a normal working shift (collar-to-collar), would entail
on-site monitoring, video recording, and mechanical
characterization as follows:
• Video record task from one/two angles.
• Time task and sub-task duration.
• Time rest/work intervals.
• Assess load resistance (equipment weight etc.).
• Track task/sub-task frequency, repetition, inten-
sity, etc.
• Characterize static/dynamic contraction type and
body/muscle movement.
• Note body segment implication (upper/lower body),
• Perform a complete anthropometric inventory of
each subject worker.
• Carry-out a clothing/personnel protective equip-
ment inventory with respect to fabric, weight etc.
• Monitor the work and rest areas’ environment with
respect to temperature (dry, wet, globe etc.) humid-
ity, barometric pressure, air speed etc.
Ideally, for statistical validity, a minimum of nine
subjects would be required in each group. This is not
practical for most Canadian mines, which due to their
mechanized nature, have small workforces. Conse-
quently, it was proposed that only three of the tasks be
evaluated in detail with nine subjects and the remain-
ing three tasks would be evaluated with only three
subjects to determine their relative position to the tasks
studied in detail.
 2.2.2 Phase 2.2 – in-situ task energy cost analysis
An in-situ energy analysis will be performed on
the same selected tasks/personnel as subjected to a
time/motion analysis by personnel with a background
in human kinetic/health sciences. This analysis will
also be performed with the same density, i.e. three
tasks in detail with nine subjects and three tasks with
three subjects to determine their relative position.
Similarly, the energy expenditure will evaluate each
of the six tasks as they are performed by typical
workers under typical working conditions. This again
should be in a location where heat exposure is not an
issue. The analysis which will span a normal working
shift (collar-to-collar) will entail periodic open-circuit
oxygen consumption plus the continuous monitoring
of selected biomedical parameters such as heart-rate,
skin heat flux and core temperature. The energy
analysis would include:
• Oxygen expenditure with a portable metabolic sys-
tem; this would entail a subject wearing a mask for
short period of times during their tasks.
• Skin heat flux measurement with temperature sen-
sors attached to the skin that transmit remotely to a
data-logger. At a minimum, this monitoring will be
required during the oxygen expenditure analysis
but in certain instances it may be required for the
full-shift or potentially 24 hrs.
• Core temperature monitoring through a variety of
methods – radio-pill, tympanic and rectal probes as
acceptable to the monitored subjects.
• Supporting information/observations to correlate
the biomedical assessment with conditions mea-
sured during the time/motion evaluation.
• Evaluation of the task’s working environment, again
to standardize the conditions.
2.2.3 Phase 2.3 – laboratory task simulation,
energy expenditure replication and
resultant temperature measurement
Prior to starting this phase, the potential of using the
data gained in the field specific to Canadian mine
operations, along with research elsewhere, will be
evaluated to determine the depth of this and subse-
quent elements. This and the subsequent laboratory
phases allow for the evaluation of a thermal index and
its measurement under strictly controlled conditions
plus in greater detail than achievable in-situ.
The task simulation would be based upon the data
collected during the time/motion and energy expen-
diture analyses. From that data a compressed work
protocol will be developed for each of the analyzed
tasks. This protocol will replicate the in-situ mine
monitored tasks with respect to loads, movement and
wearing the same clothing, and be performed under
similar conditions of temperatures, humidity and wind
speed: the only condition not simulated would be the
increased air pressure experienced with depth. These
263
Copyright © 2006 Taylor & Francis Group plc, London, UK
tests will be performed in the University of Ottawa’s
Snellen air calorimeter.
While performing the work protocol, the subjects
will be monitored in greater detail than achievable in
the mine situation; the measurements being: oxygen
consumption, changes in superficial and deep body
temperature, heat loss, sweating, skin blood flow and
various cardiovascular parameters. Within these mea-
surements, the various core temperature methods and
skin heat flux sensors will be evaluated with respect
to their relative response and cross-correlation. The
test environments will also be assessed according to
common heat stress indices.
The personnel used for these evaluations would be
neither mine-based subjects, nor have any acclimation
to working in hot environments.
2.2.4 Phase 2.4 – evaluation of the extreme as
opposed to normal environmental
conditions on physiological response
In this phase, the same subjects, as used in the pre-
vious evaluation, will perform the same tasks under
what may be the most extreme mine conditions either
generally or specifically for that task. These extreme
conditions will have been established in the earlier
review of mining conditions. The monitoring of the
test subjects will also be identical; however, it will
be focussing on whether the conditions generate a
stressful situation, which heat stress indices to use
in order to provide protection and which commercial
environmental instruments or personnel monitors are
best suited to those heat stress indices.
2.2.5 Phase 2.5 – in-situ thermal index/
monitoring validation
In this phase, the assessment of the thermal stress
indices returns to the field for mine validation and
the six tasks/groups identified initially will be re-
evaluated. Here, the indices and instrumentation iden-
tified as offering the best protection will be assessed
under two thermal conditions at the mine. These two
situations will be: under normal non-stressed working
conditions and under higher temperature conditions,
such as that which would be considered stressful
according to the existing ACGIH based protocol used
at the mines.
2.3 Phase 3 – mitigating factors or controls that
would affect the application of the thermal
stress index
This phase currently has one identified component,
whether Canadian miners are acclimated to working in
hot conditions and if so, how does that affect the appli-
cation of a thermal stress index. Furthermore, a study
of how long it takes to become, or lose, acclimation.
During the course of this evaluation it could be
found that certain elements of a miner’s personal
 protective equipment can be detrimental to heat loss
and would warrant more specific investigation.
2.3.1 Phase 3.1 – physiological acclimation
evaluation
This is a very important area to investigate with respect
to conditions in Canada. However as the stress indices
to be used and what constitutes acclimation have not
been defined, the description of this phase is very
preliminary.
To date this evaluation/development of heat indices
has been based upon using non-acclimated subjects.
It is known that people can to some degree become
accustomed to working in hot conditions. However, it
is not necessarily known to what level Canadian miners
are acclimated.
This phase would evaluate whether an acclimation
process takes place within miners, establish the dif-
ference in their heat tolerance, and the subsequent
application of any heat index.
Another issue that may require evaluation is whether
there is any seasonal difference in acclimation between
winter and summer.
2.4 Project timeline and current work summary
The project as described contains elements that are
on-going, currently in the planning stage, and future
directions. The project started in 2004 and is expected
to be completed in 2007/8. At this time the review of
heat exposure guidelines and instrumentation is near-
ing completion. The time and motion field evaluation
of the six tasks have been completed and the data is
being analyzed. The energy expenditure field analy-
ses are underway with fifteen of thirty-six assessments
complete.
The following sections provide some information
with respect to each element.
3 HEAT STRESS GUIDELINES IN CANADA
As mentioned previously, Canada uses the ACGIH
TLV® screening criteria based upon the WBGT to
assess heat stress and work rest regimens. The WBGT
index was developed as part of work directed by
the U.S. Navy aimed at evaluating factors surround-
ing exertional heat-related injuries in young military
recruits during training (Yaglou & Minard, 1957). The
simplicity of this index has led to its widespread use,
acceptance (NIOSH, 1972) (ISO 7243, 1982, 1989)
and being the basis of several environmental moni-
tors; however it has limitations when used to simulate
human response. With respect to mining these include:
• The TLV only contains five generalized metabolic
categories of activities that are difficult to align with
mining tasks.
264
Copyright © 2006 Taylor & Francis Group plc, London, UK
• The WBGT can be adjusted to account for the influ-
ence of clothing according to only three categories.
• It is not designed for use with clothing that does not
permit evaporation or layered clothing ensembles;
in Canada, miners can wear clothing that provides
a partial or complete barrier to evaporation and
layered clothing is common.
Where the WBGT criteria is not applicable, the
ACGIH recommend the use of a more detailed anal-
ysis such as calculating the required sweat rate (ISO
7933, 1989); this has not been extensively employed
in Canada. The ISO 7933 required sweat rate is a more
complicated method but may be more applicable to
mechanized mining as it considers such inputs as:
• Air velocity: this can be significant considering
the dilution requirements of large diesel powered
equipment in some Canadian mines.
• The fraction of the body area exposed: this can be
minimal with full personnel protective equipment
(PPE) – full coveralls, gloves, boots, hard-hat, ear-
muffs, safety glasses with side-shields and a T-shirt
covering the chest area.
• A clothing insulation value, which although not as
severe as for non-permeable clothing, should con-
sider rubber boots and gloves, multiple layers and
the use of foam inserts in hard-hats as prescribed by
some jurisdictions for side impact protection.
The ISO 7933 method, however, does not include
any consideration of barometric pressure and how it
affects the air density as included in McPherson’s
(1992) formulation of Air Cooling Power (ACP) and
the Thermal Work Limit (TWL) (Brake & Bates,
2002).
At this time no single index appears to be ideal for
deep mining conditions in Canada. It is for this reason
that this study has gone back to some of the basics to
determine work load and energy output, quantify the
environment and consider clothing.
4 INSTRUMENTATION EVALUATION
This section provides details of the overall instrument
selection criteria, followed by the instruments chosen
for evaluation and comments on their initial operation.
4.1 Selection criteria
As the final heat stress index (or indices) is unknown,
the study did not limit itself solely to instrumentation
marketed for heat stress measurement. In the instru-
ment selection, in order of priority, the following were
considered.
4.1.1 Dry-bulb temperature
The direct measurement of dry-bulb temperature was
deemed to be essential.
 4.1.2 Natural and aspirated wet-bulb temperature
Either a direct or indirect measurement of wet-bulb
temperature would be acceptable. The indirect method
of determining a wet-bulb temperature from dry-bulb
temperature, relative humidity, barometric pressure
and air velocity was considered as a way to avoid a
“wet” probe and it becoming dry with extended use
under continuous monitoring conditions.

4.1.3 Relative humidity

The direct measurement of relative humidity was con-
sidered as an alternative to a direct determination of
the wet-bulb temperatures.

4.1.4 Globe temperature

Although measuring globe temperatures is an integral
part of determining the WBGT, there is the possibility
that it may be redundant in underground environ-
ments as globe temperatures tend towards the dry-
bulb temperature. This assumption would be checked
through an evaluation of radiant heat sources in deep
mechanized mines.
4.1.5 Barometric pressure
Non-standard (sea-level) barometric pressure affects
the determination of natural wet-bulb temperature, air
density and numerous other heat stress related param-
eters with respect to an environment’s ability to permit
cooling. Consequently, elevated barometric pressures,
as observed in deep mines, can become important.
Therefore instruments that include a barometric sen-
sor could prove advantageous. This need to consider
barometric pressure also has the potential to reduce
the advantages of some units that internally process
their measurements into a heat stress index assuming
a standard barometric pressure.
4.1.6 Air velocity
Certain heat indices such as effective temperature and
those based upon the air’s cooling power require the
measurement of air velocity. An air velocity measure-
ment is also needed to correct between the natural and
aspirated wet-bulb temperatures.
4.2 Test instruments
The instrument descriptions include their direct mea-
surements, calculated measurements and pertinent
limitations.
4.2.1 Calor Heat Stress Meter
The Calor Heat Stress Meter (HSM), Figure 2, incor-
porates dry-bulb and black globe temperature, relative
humidity, barometric pressure and air speed sen-
sors. The unit calculates the wet-bulb temperature
from some of these measured parameters. From these
primary values, along with, where applicable, user
input values the unit determines the wet-bulb globe
Figure 2. Calor Heat Stress Monitor (HSM).
temperature (WBGT), plus provides guidance with
respect to work/rest cycles, maximum work duration
and water replacement requirements. The HSM also
uses a modified air cooling power calculation to deter-
mine the Thermal Work Limit (TWL), the limiting
metabolic work rate that can be performed in the
measured environment. With the inclusion of an air
velocity probe and a barometer, the HSM currently
has the most comprehensive range of measurements
and subsequent determination of heat exposure related
parameters.
With respect to deep mining in Canada where baro-
metric pressures can reach 140 kPa, one notable limi-
tation of this unit is its 40–115 kPa barometric pressure
range which currently restricts wet-bulb temperature
determinations at depth.
4.2.2 Casella Microtherm Heat Stress WBGT
The Casella Microtherm Heat Stress WBGT, Figure 3,
has the typical three probes for heat stress determi-
nations. The unit directly measures dry-bulb, natural
wet bulb and a black globe temperature. From these
measurements it calculates the dew point tempera-
ture and relative humidity assuming standard pressure.
With respect to heat stress, the unit calculates and dis-
plays the WBGT, and post analysis software can be
used to determining work/rest regimes. For deep min-
ing applications, its potential limitations are the lack of
a barometric pressure and air velocity measurements
to determine other heat stress parameters.
4.2.3 IST-AIM RSS-214 WibGet & HeatStress Lynk
The WibGet (Figure 4) and HeatStress Lynk (Figure 5)
both have the three typical probes to directly measure

265

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 3. Casella Microtherm Heat Stress WBGT.
Figure 4. IST AIM RS-214 WibGet.
dry-bulb, natural wet bulb and black globe tem-
peratures. Post measurement software can calculate
metabolic rates and clothing type factors. Both the
WibGet and HeatStress Lynk have the same potential
pressure and velocity limitations as the Casella unit to
calculating other heat stress indices.
4.2.4 QUESTemp◦ 32 or QUESTemp◦ 34
These Quest units, (Figure 6) are also common heat
stress meters with the three typical probes to directly
266
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 5. IST AIM HeatStress Lynk.
Figure 6. QUESTemp◦ 34.
measure dry-bulb, natural wet bulb and black globe
temperatures. An added feature of these units is a
humidity sensor mounted internally. The unit as tested
had the instrument standard 50 mm (2 ) diameter black
globe. A 150 mm (6 ) diameter globe was optional.
These units can calculate the WBGT plus the Heat
Index used by the U.S. National Weather Service
or the Humidex used by Environment Canada. Both
the Heat Index and Humidex equate the dry-bulb
temperature and relative humidity into a single “equiv-
alent” temperature. Similar to the IST-AIM units, post
measurement analysis can produce a variety traditional

Figure 7. Kestrel 4000 Weather Tracker.
thermal comfort parameters. Again these units have
the same pressure and velocity limitations; to some
degree, these are addressed with the QUESTemp◦ 36
with its optional air velocity probe.
4.2.5 Kestrel 4000 Pocket Weather Tracker
The Kestrel 4000 (Figure 7) incorporates dry-bulb
temperature, humidity, barometric pressure and air
velocity measurements into a small pocket-sized and
comparatively cheap instrument. The unit, although
not marketed as a heat stress meter, and lacking a
globe temperature measurement, was thought to have
some potential because of its temperature and baro-
metric pressure capabilities. Similar to the Calor HSM,
the unit calculates the wet-bulb temperature from
the measured parameters. It also calculates the Heat
Index and Wind Chill equivalent temperatures and
the dew point temperature. This unit, or its associ-
ated post measurement software, does not determine
any of the more traditional industrial thermal comfort
parameters.
Again, similar to the Calor HSM, for deep min-
ing applications, the unit is currently limited by its
barometric pressure range and subsequent wet-bulb
temperature determination.
4.2.6 Vaisala HMI 141
The Vaisala HMI 141 is a more basic dry-bulb tem-
perature and relative humidity meter which calculates
the wet-bulb and dew point temperatures, and absolute
humidity. Despite not being marketed as a heat stress
meter, one advantage of this unit is that the calculated
267
Copyright © 2006 Taylor & Francis Group plc, London, UK
values can be compensated for variations in barometric
pressure through the user input of local values.
4.2.7 ACR SmartReader Plus Dataloggers
Combined dry-bulb temperature, relative humidity
and barometric pressure sensors were considered due
to their ability to remotely and continuously monitor
the underground environment and potentially transmit
that information through underground communication
systems. This type of monitoring ability was deemed
attractive as it could be used to determine what activ-
ities may be performed in an environment prior to
dispatching the workforce and/or troubleshooting for
adverse conditions.
4.3 Preliminary instrument findings
The analysis pertaining to the accuracy of the instru-
ments, subsequent calculations and suitability to a heat
stress index is still on-going. However, the following
has been discovered.
4.3.1 Temperature measurement (general)
As expected all the instruments generally agree when
measuring a dry-bulb temperature, and all instru-
ments with multiple temperature sensors (wet-bulb
and globe) agree when all were reduced to a dry-bulb
measurement.
4.3.2 Globe temperature measurement
Of the four instruments that measure globe tempera-
ture, three units generally agreed, while the fourth was
up to 2◦ C lower when placed in a common outdoor
environment. At this time, the manufacturer had no
explanation for this difference. All of the instruments
have globes with a smaller diameter than the Stan-
dard recommended 150 mm globe to speed up their
response; however their measurements are supposedly
corrected for the size difference.
Although this difference is a concern when deter-
mining WBGT values, it may not be relevant under-
ground. The evaluation of this measurement is now
being performed against standard sized globes.
4.3.3 Barometric pressure measurement
As expected both instruments with built-in barometric
pressure sensors failed to operate/calculate correctly
once their apparent operational range was exceeded.
Up to that point the measurements were suitable
for calculating the pyschrometric properties of air at
different barometric pressures.
However, based upon the performance of an ACR
SmartReader Plus 4 with a 0–30 PSI (200 kPa) abso-
lute pressure module, and discussions with pressure
sensor manufacturers, the instruments could already
have a suitable sensor and only require the response
calculation range extending. Regardless, these units
could be modified for deep mine applications.
 4.3.4 Relative humidity measurement
The quoted accuracy of the instruments to measure
relative humidity ranges between ±1% to ±3% after
calibration against Standard salt solutions. This upper
order of difference, and larger, was routinely observed
in the instrument evaluation. This could limit the
accuracy of such units to determine the wet-bulb tem-
peratures. Currently, the accuracy of the direct reading
humidity meters is being re-evaluated against a high
accuracy chilled mirror dew-point meter.
5 MINE ASSESSMENTS
To date, the data collection for the time and motion
study is complete, and the energy expenditure mea-
surement along with body temperature monitoring is
underway. Consequently, the analysis of the combined
data is in its early stages and few results are available.
However, the following may be reported:
• 40 subjects were observed in a time/motion study
across six occupational groups.
• 8000 minutes of video recording was collected on
the various tasks for subsequent analysis.
• Based upon the video records, 23 sub-tasks have
been identified to characterize the occupation.
• At this time, a relative intensity from 1 through
5, based upon the observer’s experience has been
assigned to each sub-task. This assessment will be
evaluated during the energy expenditure analyses.
• 3500 minutes of concurrent body core and sur-
face temperature and oxygen consumption has been
performed on 15 subjects through their various
sub-tasks.
With respect to the instrumentation being used:
• The energy expenditure analysis was delayed by the
portable metabolic apparatus needing modification
for the higher barometric pressures encountered in
the mine.
• Radio pills and wireless contact sensors are per-
forming well with little loss of data.
6 CONCLUSIONS
Controlling heat exposure in Canada’s deep mecha-
nized mines is becoming increasingly important and
costly. This is reflected in the number and magnitude
of the ventilation system upgrades and the introduc-
tion of mechanical cooling systems taking place. In the
future it could affect the viability of individual mines
and sustainability of the communities they support.
Consequently, for the DMRC, a major project
within their research goals is investigating personal
heat exposure as it applies to the Canadian mechanized
mining industry.
268
Copyright © 2006 Taylor & Francis Group plc, London, UK
ACKNOWLEDGEMENTS
The author would like to recognize the support of the
Deep Mining Research Consortium in this on-going
research thrust and its permission to publish this paper.
REFERENCES
ACGIH, 2005, Threshold limit values for chemical and
physical agents and biological exposure indices, Amer-
ican Conference of Governmental Industrial Hygienists,
Cincinnati, ISBN 1-882417-58-5.
Brake, D.J. and Bates, G.P. 2002. Limiting Metabolic Rate
(Thermal Work Limit) as an Index of Thermal Stress,
Applied Occupational Environmental Hygiene, March,
17(3), pp 176–186.
Campbell, J.G. 2005. DMRC Cost-Risk Models – Setting
Research and Development Priorities for Deep Mining,
Deep Mining Research Consortium, http://www.deep-
mining research.org/, Internal Report.
Chandler, H. 1998. Heat Stress Monitoring; Hot enough for
ya?, OHS Canada, June/July.
Hardcastle, S.G., Kocsis, C. and O’Connor, D. 2006. Jus-
tifying ventilation-on-demand in a Canadian mine and
the need for process based simulations. To be included
in the Procs. 11th U.S./North American Mine Ventilation
Symposium, Penn State, Pennsylvania, June 5–7, 2006.
Hortin, K.M. and Sedlacek, J. 2002. Change of the push-pull
ventilation system at Kidd Creek mine of Falconbridge
Ltd. (challenges and logistics), Proc. 1st North Amer-
ican/9th U.S. Mine Ventilation Symposium, Kingston,
Ontario, Canada, 8–12 June, Ed. E. DeSouza, Balkema,
ISBN 90-5809-387-5, pp 69–75.
Howes, M.J. and Hortin, K. 2005. Surface cooling at
Kidd Creek mine, Proc. 8th International Mine Ventila-
tion Congress, Brisbane, Australia, 6–8 July, Ed. A.D.S.
Gillies, AUSIMM, Carlton, Victoria, ISBN 1-920806-32-
6, pp 56–63.
Howes, M.J. and Nixon, C.A. 1997. Development of pro-
cedures for safe working in hot conditions, Proc. 6th
International Mine Ventilation Congress, Pittsburgh, May
17–22, Ed. R.V. Ramani, SME, ISBN 0-87335-146-0,
pp 191–197.
Howes, M.J. and Sedlacek, J. 2001. Kidd Creek mine – Ven-
tilation and cooling planning far an extension to 3100 m,
Proc. 7th International Mine Ventilation Congress,
Krakow, Poland, 17–22 June, Ed. S. Wasilewski, EMAG,
ISBN 83-913109-1-4, pp 991–1003.
Howes, M.J. and Stockhaus, F.J. 2002. Onaping Depth
project – Ventilation and cooling planning for mining
to 2600 m below surface, Proc. 1st North American/9th
U.S. Mine Ventilation Symposium, Kingston, Ontario,
Canada, 8–12 June, Ed. E. DeSouza, Balkema, ISBN
90-5809-387-5, pp 391–400.
ISO 7243, 1982, rev. 1989. Hot environments – Estimation of
the heat stress on working man, based on the WBGT- index
(wet bulb globe temperature), International Organisation
Standardisation, Geneva.
ISO 7933, 1989, Hot environments – Analytical determi-
nation and interpretation of thermal stress using calcu-
lation of required sweat rate, International Organisation
Standardisation, Geneva.
 MacIssac, H. 2001. Kidd Mine D Project, Canadian Institute
of Mining and Metallurgy (CIM) 15th Mine Operators
Confernce, February 11–14, Sudbury.
Marx, W.M. Providing an Acceptable Working Environment
in Ultra Deep Mines, Journal of the Mine Ventilation Soci-
ety of South Africa, Vol. 51, No. 2, April/June, 1998,
pp 57–60.
McPherson, M.J. 1992. The Generalization of Air Cooling
Power, Proc. 5th International Mine Ventilation Congress,
Johannesburg, S. Africa, 25–30 October, Ed. R. Hemp,
MVSSA, ISBN 0-620-17156-1, pp 27–36.
Misaqi, F.L., Inderberg, J.G., Blumenstein, P.D. and
Naiman, T. 1976. Heat Stress in Hot U.S. Mines and
Criteria for Standards for Mining in Hot Environments,
Mining Enforcement and Safety Administration, United
States Department of the Interior, MESA Report No. 1048,
pp 1–47.
NIOSH 1972. Criteria for a Recommended Standard-
Occupational Exposure to Hot Environments, National
Institute for Occupational Safety and Health, HSM 72–
10269.
269
Copyright © 2006 Taylor & Francis Group plc, London, UK
O’Connor, D.F., Gibson, D., MacKay, B. and Grupp,
D.R. 2002. Creighton Mine, #11 Shaft Exhaust Fan
Upgrade, Proc. 1st North American/9th U.S. Mine Ven-
tilation Symposium, Kingston, Ontario, June 8–12, Ed.
E. DeSouza, Balkema, ISBN 90-5809-387-5, pp 137–144.
Thibodeau, D. 1999. Deep Mining Considerations at
Creighton Mine, INCO Ltd. Ground Control Bulletin,
Issue No. 5, October.
Van der Walt, W.H., Schutte, P.C., Hitge, J.J., Kielblock, A.J.
and Strydom, N.B. 1981. A Survey of the incidence of heat
stroke in the gold mining industry over the period 1969
to 1980, Research report No. 15/8, Chamber of Mines of
South Africa, Johannesburg.
WSIB 1997–2004. Statistical supplements to the Annual
Reports, Workplace Safety and Insurance Board, Ontario,
Canada.
Yaglou, C.P. and Minard, D. 1957. Control of heat casualties at
military training centers, American Medical Association
Archives of Industrial Health, 16:302–316 and 405.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Optimum insulation for engineering control of mine thermal regime

around a mine airway in permafrost

H. Wu & S. Bandopadhyay
University of Alaska Fairbanks, Fairbanks, AK

V.U. Izaxon
Kemerov Mining Institute, Russia

ABSTRACT: For mine opening in permafrost, if the temperature increase in the surrounding rock mass is
large enough during the mine operation due to heat transfer between the ventilation air and the rock mass, the
overall strength of the rock mass may decrease, causing deformation to increase over time. Large deformation of
a mine opening will create potential stability problems in “warm”, ice-rich silt and gravel placer gold deposits.
To prevent the heat transfer from ventilation air into the surrounding rock mass, one commonly used method is
to insulate the walls of mine airways. In this paper, analysis of several configurations for a completely insulated
airway surrounded only by frozen silt, including three types of insulation and four insulation thicknesses is
presented. Results from numerical simulations for a time interval of ten years was used to optimize insulation
thickness for engineering control.

1 INTRODUCTION a series of complications which can be prevented or

During excavation of a mine opening and subsequent
exploitation of mineral resources, extensive venti-
lation is required. Large seasonal variations in the
temperatures of the ventilating air cause changes in the
original thermal field around a mine airway, through
heat and mass exchange between the air and the
surrounding medium. This is especially true in the
summer, when the temperatures between the air enter-
ing the mine is far above the freezing point of water,
causing thawing of the ice contained in the surround-
ing rock. In the winter, in order to avoid icing up of the
intake airway and to protect mine workers from cold
air, the intake air is often heated, and the velocity of
the heated air must be kept fairly high. These thermal
interactions have major influence on climatic quality
as well as on the stability of the mine openings.
The nature and intensity of heat exchange between
the air and the walls of an underground working
are functions of a number of independent variables,
including the thermal properties of the permafrost and
air, the velocity and distribution of the air stream, the
size, shape and cross section of the airways, the ice
content and temperature of the permafrost, the temper-
ature and moisture content of the ventilation air, and
time (Udd 1989). The method of ventilation involving
heated air, which is commonly used at present, does
not correspond to the natural conditions prevailing in
permafrost regions; it is not economical and leads to
greatly reduced by lowering the temperature of the air
stream entering the mine until it is equal to or very
close to the temperature of the frozen ground.
Solution of mine heat regime of a single excavation
requires determination of the temperature field sur-
rounding the opening and the basic parameters such as
velocity, quantity of air, temperature and humidity of
the ventilation stream. Mine heat regime is controlled
by the coupled process of heat and mass exchange. The
influence of phase change of water/ice in the rock mass
surrounding mine openings increases with increasing
ice content in the permafrost and with increasing rock
mass temperatures during the cold season.
Therefore, for planning of ventilation systems and
for engineering control of the mine thermal regime in
a mine in the Arctic, it is necessary to predict tem-
perature variations in the incoming air as well as the
temperature variations along the length of the airway.
A relationship between the deformation of the airway
and the temperature can be determined based on the
thermo-dynamic considerations. This relationship can
be developed based on the coupled theory of both
temperature and thermal stress. Thus, to evaluate the
stability of a mine opening in the Arctic, it is neces-
sary to solve a heat transfer problem generally known
as the “ Stephan Problem”, taking into account the
varying thermal-physical properties of air and the rock
mass (Bandopadhyay et al. 1995). The solution to the
Stephan problem is an iterative process, and requires

271

Copyright © 2006 Taylor & Francis Group plc, London, UK

 data on the initial mine thermal regime. Finite element
formulation of the thermal problem for deep mine air-
way as well as for airways located close to the surface
is presented elsewhere (Bandopadhyay et al. 1998).
The rock mass bodies considered here are frozen
deposits of silt. The stability and deformation of frozen
materials are functions of their temperatures. For mine
openings in permafrost, if the temperature increase in
the surrounding rock mass is larger during the oper-
ation due to heat transfer between the ventilation air
and the rock mass, the overall strength of the surround-
ing rock mass may decrease and the deformation will
increase over time (Bandopadhyay et al. 1995). Large
deformation of a mine opening will create potential
stability problems in “warm”, ice-rich silt and gravel
deposits.
A simple preliminary analysis of the governing
equations of the analytical models for heat trans-
fer in an un-insulated airway presented elsewhere
(Bandopadhyay et al. 2001, Bandopadhyay & Zhang
1990) clearly indicates that a thawing area would
develop around a mine airway. A thawing region as
large as 2 m deep in the frozen ground is expected
based on the results of the finite element analysis
(Bandopadhyay et al. 1998). If the thermal regime is
not controlled, the thawing area would advance out-
ward and finally would reach a steady state condition.
In comparison with frozen soil, the thawed soil has
less bearing strength. With an increased temperature,
the rock mass looses its strength due to the melting
of ice-bounding and resulting moisture migration is
accompanied by a breakdown of the ice and structural
bonds of the soil grains. Two different methods can be
considered to prevent or reduce instability caused by
the thermal redistribution surrounding a mine airway.
First, the use of roof bolts to control the insta-
ble roof. This can be viewed as an indirect approach.
The reinforcement does not prevent or reduce the heat
transfer around a mine airway but may even intensify
the thawing process in the presence of roof bolts which
is made of steel. Secondly, the use of thermal insulation
(Fig. 1) to prevent or reduce heat flux transferred into
the frozen rock mass surrounding the mine opening.
Figure 1. Physical model of an insulated mine airway.
272
Copyright © 2006 Taylor & Francis Group plc, London, UK
Ground freezing may also be applied to any soil
and rock materials regardless of structure, grain size,
or permeability. Cyro-anchors, also referred as ther-
mosyphons and thermotubes, are used throughout
Alaska and at many locations in Canada to remove heat
and depress the subsurface temperature in permafrost
regions. During arctic winter season, Cryo-anchor sta-
bilizers can remove large quantities of heat from the
surrounding rock mass, and reject that heat to the air.
In the summer, when the air and the environment are
warmer than the rock mass, the device becomes inac-
tive, minimizing heat input to the ground. Engineering
control of thermal regime around a mine airway with
thermosyphons placed on the surface is presented
elsewhere (Bandopadhyay et al. 1996).
To prevent the heat transfer from ventilation air
into the surrounding rock mass, one commonly used
method is to insulate the walls of mine airway. The
mechanism of heat transfer around a frozen under-
ground airway and its behaviors are much different as
compared to an un-insulated airway. In an insulated
mine airway, the ventilation air is not on direct con-
tact with the frozen ground. In this situation, the heat
transfer process can be viewed as heat flowing through
different mediums with different thickness and varying
thermal-physical properties.
The ensuing part of the paper is organized into
four sections. Section 2 briefly discusses the governing
equations of the model and the finite element formu-
lation of the problem. Section 3 lists the properties
of materials and conditions under which the model
is simulated. Section 4 briefly lists the key results
with appropriate discussion. Section 5 summarizes the
research with the major conclusions.
2 MODELING OF HEAT TRANSFER IN
DEEP MINE AIRWAY
When two different materials are in contact, an inter-
face exists between them. If the contact between the
different materials is considered as a complete one,
then convective and radiative heat transfers can be
totally neglected. In such cases, the equations repre-
senting the heat transfers in the composite materials
are the same as that in a single medium. At each inter-
face, however, the interfacial conditions (additional
boundary conditions) must reflect two facts:
1 In the absence of thermal contact resistance, the
temperature of each body at the contact interfaces
will be the same;
2 Heat flow from the body of high temperature must
be equal to the heat flow into the body of lower
temperature.
The governing equations for an insulated mine air-
way with additional boundary conditions are presented
here for ready reference.
 Previous analysis of the roof thawing depths both and
for the deep and shallow airways indicated that a
two-meter thawing after ten year’s ventilation can be
reasonably expected (Bandopadhyay et al. 1995). The
only heat source considered for the analysis is the heat
flow from the ventilation air to the frozen soil. To where, k1 is the heat conductivity of the insulation
reduce the transfer of heat flux from the ventilation material.
air, one possible approach is to insulate the airway For a partially insulated mine airway (Figs. 3 and 4),
completely. In this situation, the heat is transferred the boundary condition at the contact interfaces of
from the ventilation air to the insulation material rather airway roof “A-B” is:
than to the frozen ground. If it is assumed that the
contact between the airway walls and the insulation
is complete, then only the conductive heat transfer
would exist on the interface of the insulation and the The boundary condition at the contact interfaces of
frozen soil. Under this assumption the physical model airway roof “B-C”:
of a completely insulated airway can be simplified as
shown in Figure 2.The mathematical model of the ther-
mal system is the same as that of an insulated airway
as shown below:
For a completely insulated mine airway (Figs. 5
and 6), the floor is also needed to be provided with
insulation. The additional contact, at interface, “C-D”,
In an insulated mine airway, the general interfa-
cial boundary conditions between the insulation and
airway walls expressed mathematically are:

Figure 2. Physical models of heat transfer around insulated Figure 3. Domain of heat transfer around a partially insu-
mine airways. lated mine airway.

273

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 4. Finite element mesh for heat transfer simulation
around a partially insulated mine airway.
is described identically to contact interface, “A-B”,
above.
Equations (4) and (5) above define the additional
interface conditions for an insulated mine airway. With
those additional boundary conditions, the temperature
distribution around a deep insulated airway can be
analyzed in conjunction with the mathematical model
developed for an uninsulated airway.
In the following paper, the models of heat trans-
fer and subsequent finite element treatments already
presented elsewhere (Hibbit et al. 1992) was used
with additional boundary conditions (Fig. 2) to reflect
the presence of insulated boundaries around the mine
cross-section. Here again, the finite element program
ABAQUS (Hibbit et al. 1992) was used for the heat
results analysis because ABAQUS also provides a
more convenient way to conduct a thermal analysis
for an insulated mine airway.
3 MODELING SCENARIO
Several types of insulating materials can be identified
for many industrial applications. For mining applica-
tions, only three insulation types have been considered
here. To make the parameters more representative, the
274
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 5. Domain of heat transfer around a fully insulated
mine airway.
thermal property values of three selected insulation
materials used in the ABAQUS analysis is defined in
Table 1.
Since the thickness of insulation as well as thermal
properties vary widely, to examine the influence of
insulations with different thicknesses, a total of twelve
variants were analyzed for a completely insulated air-
way. Three types of insulation materials (Polyurethane
foam, Fiber glass, and Cement fiber) and four insula-
tion thicknesses (30 mm, 50 mm, 80 mm, and 100 mm)
were considered as shown in Table 2.
The finite element meshes for a completely insu-
lated airway (both deep and shallow) are nearly the
same (Fig. 6). Triangle elements were used to mesh
the airway wall-interface layer. To provide a better
focus, the analysis was limited to airways in frozen
silt. Numerical simulations were conducted for a
time interval of ten years. Results of the numerical
simulation for different insulation alternatives were
compared mine airway with no-insulation scenario.
4 RESULTS AND DISCUSSION
It can be seen from Figure 7 that polyurethane
foam is highly effective. With polyurethane foam as
 Table 2. Analysis of thawing depths & temperature at node
1235 in a fully insulated deep airway.

Insulation material Without insulation

Soil Silt
Thickness (mm) 0
Total time (year) 10
Time step (day) 4
Roof Thawing (m)
Thawing time (year)
Roof temp (◦ C)
Insulation material Polyurethane foam
No. of case 1 2 3 4
Soil Silt Silt Silt Silt
Thickness (mm) 30 50 80 100
Total time (year) 10 10 10 10
Time step (day) 4 4 4 4
Roof Thawing (m) 0.23
Thawing time (year) 10.12
Roof temp (◦ C) 0.3 −0.8 −1.7 −2
Insulation material Glass fiber
No. of case 5 6 7 8
Soil Silt Silt Silt Silt
Thickness (mm) 30 50 80 100
Total time (year) 10 10 10 10
Time step (day) 4 4 4 4
Roof Thawing (m) 0.63 0.05
Thawing time (year) 4.68 4.37
Roof temp (◦ C) 1.0 0.07 −1 −1.9
Insulation material Cement fiber
No. of case 9 10 11 12
Soil Silt Silt Silt Silt
Thickness (mm) 30 50 80 100
Figure 6. Finite element mesh for heat transfer simulation Total time (year) 10 10 10 10
around a fully insulated mine airway. Time step (day) 4 4 4 4
Roof Thawing (m) 1.23 1.05 0.48 0.3
Thawing time (year) 4.17 6.92 3.73 2.65
Table 1. Heat transfer coefficient values of insulation mate- Roof temp (◦ C) 2.08 1.46 1.05 0.8
rials used in ABAQUS input files.

Insulation material PF GF CF

Density, (kg/m3 ) 32 112
Conductivity, W/(m-◦ C) 0.25 0.36
Specific Heat, kJ/(kg-◦ C) 0.96 0.96
PF-Polyurethane foam, GF-Glass fiber, CF-Cement fiber
insulation, thawing of roof appeared only when the
insulation thickness is 30 mm and during a ten-year
period, the thawing depth was limited to a depth of
0.21 m. In comparison with uninsulated airway, thaw-
ing of roof is ten times more than that could be
observed with only 30 mm insulation. No thawing
occurs when the thickness of polyurethane foam was
increased to 50 mm of more.
Similarly thawing conditions were also observed
with fiber glass as insulation material (Fig. 8). Dur-
ing the time ten years span, thawing depths between
480 0.6 m ∼ 0.8 m for the 30 mm thick insulation and
0.87 less than 0.2 m for the 50 mm thick insulation were
1.3 observed. On the other hand, un-insulated airway walls
thawed 3 ∼ 10 times more than the insulated airway
under similar ventilation conditions.
With cement fiber as insulation material, the depths
of thawing are 1.2–1.4 m, 1.0–1.2 m, and 0.4–0.6 m
for insulation thickness of 30 mm, 50 mm, and 80 mm
respectively in the same time period. When the insula-
tion thickness was increased to 100 mm, no thawing
was observed (Fig. 9). The depth of roof thawing
is considerably larger than the two other insulation
materials considered before.
From Figures 7 through 9, it can be seen that roof
thawing decreases with the increasing thickness of
insulation and for the same insulation thickness, insu-
lation with smaller conductivity has lesser thawing
depth. From the results shown in Figure 10, it can be

275

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 7. Relationship between polyurethane thickness and
thawing depth.
Figure 8. Relationship between fiber glass thickness and
thawing depth.
seen that cement fiber is not a very effective insulation
material.
From the results in Table 2, it can be seen that the
thermal regime surrounding an airway can be con-
trolled by insulating the airway. The roof temperature
can be kept at approximately −0.7◦ C by applying a
50 mm thickness of polyurethane foam or an 80 mm
276
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 9. Relationship between cement fiber thickness and
thawing depth.
Figure 10. Relationship between thawing depth and insula-
tion thickness.
thickness of fiber glass insulation. If the roof tem-
perature is required to be kept at −1.6◦ C, an 80 mm
thickness of polyurethane foam may, however, be
required. Similarly for keeping the roof temperature
at −1.3◦ C a 100 mm thickness of fiber glass and at
−1.8◦ C by 100 mm of polyurethane foam would be
needed.

Figure 11. Temperature variations at mid-points over deep
airway roof.
Figure 12. Temperature distribution for one of the insulation
conditions.
Thawing depths in Figure 10 describe the relation-
ship between thawing depth and ventilation time for
different thickness and different types of insulations.
It was seen that the roof temperature of an insulated
airway tends to a steady state again for each case even
though the time needed to reach a steady state varies
from 5 years to 10 years. Figure 11 shows the variations
in roof temperature for one of the combinations. The
temperature distribution in each case has reached its
277
Copyright © 2006 Taylor & Francis Group plc, London, UK
steady state after an interval of 3 years or more years.
This with increasing ventilation time, no changes in
temperature would be expected. Figure 12 shows the
temperature distribution for one of the insulation con-
ditions. Readers are referred to Bandopadhyay et al.
(1998) for all the temperature profiles.
5 SUMMARY AND CONCLUSIONS
The development and installation of mining infras-
tructure in frozen ground such as the Arctic needs to
take into account the significant differences in rock
mass properties under different thermal regimes. In
other words, roof stability is a crucial component of
an efficient mine design.
Mine thermal regimes of deep airways and opening
were analyzed. A finite element formulation was used
to estimate the propagation of heat through the frozen
rock mass. Results of both deep and shallow mine
ventilation situations show that without any thermal
control, there would be a stable change in temperature
along the ventilation path.The need for thermal control
including refrigeration of intake air was emphasized by
previous researchers to ensure stability. In this paper
one of the alternative methods of thermal control: the
use of thermal insulation in mine airways was inves-
tigated. Different insulation on the walls of airways
such as polyurethane, fiber glass and cement fiber
was investigated for their efficiency. It was found that
cement fiber provided the best thermal insulation for
deep mine airways.
With different insulations on the walls of airways,
roof thawing can be reduced and in certain cases, com-
pletely controlled. Roof thawing decreases with the
increasing insulation thickness, and for the same insu-
lation thickness, insulation with smaller conductivity
results in smaller thawing depth.
Analysis of the model suggested that roof thaw-
ing increases by a small amount if the airways are
partially insulated. In some situations, roof thaw-
ing did not appear for partially insulated airways. It
would be worth noting that airways with partial insu-
lation are more viable due to different engineering
considerations.
The current research has been performed on a two-
dimensional basis and can be easily extended in to a
three dimensional case. Some other strategies such as
controlled recirculation of air in the mine airways need
to be investigated for minimizing thawing of frozen
rock mass along the mine airways.
REFERENCES
Bandopadhyay, S. & Y. Zhang, “Influence of Ventilating Air
Fluctuations on Stress Distribution Around an Airway
 in Frozen Ground,” Transactions, SME/AIME, Vol. 288,
pp. 1819–1825, 1990.
Bandopadhyay, S., Wu, H., Nelson, M.G., & V. Izaxon. 1995.
Design of Ventilation Systems for Underground Placer
Mines in the Arctic: Analysis of Mine Thermal Regime. In
A. Wala (ed.), Proceedings of the 7th U.S. MineVentilation
Symposium, A. SME/ AIME: 27–32.
Bandopadhyay, S., Hanguang, Wu, Nelson, M.G. &
V. Izaxon. 1996. Thermosyphons for Removal of
Ventilation-Induced Heat in an Underground Placer Mine
in the Arctic. Transactions, SME/AIME. 300: 1915–1921.
Bandopadhyay, S., Hanguang, Wu, Nelson, M.G. & V. Izaxon.
1998. A Finite Element Model of Heat Transfer in shallow
placer mine airway in the Arctic. In J. Udd & A.J. Keen
278
Copyright © 2006 Taylor & Francis Group plc, London, UK
(eds), Proceedings of the 5th International Symposium on
Mining in the Arctic: 135–145. A.A. Balkema Publishers.
Bandopadhyay, S., Hanguang, Wu, Nelson, M.G. & V. Izaxon.
2001. Analysis of Thermal Regime of a Deep Mine Airway
in Permafrost. In H.K. Olsen, L. Lorentzen, & O. Rendal
(eds), Proceedings of the 6th International Symposium on
Mining in the Arctic: 107–117.
Hibbit, Karlson & Sorensen, Inc. 1992. ABAQUS Users
Manuals, Version 5.3.
Udd, J. 1989. Mining in Canada’s North-A Technical
Overview. In S. Bandopadhyay, & F. Skudrzyk (eds), Pro-
ceedings of the 1st International Symposium on Mining in
the Arctic:15–25. Balkema Publishers.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Design of energy efficient mine ventilation and cooling systems

W. Marx, F.H. von Glehn & R.W. Wilson

Bluhm Burton Engineering, Sandton, South Africa

ABSTRACT: A concern over global energy resources has resulted in an increased emphasis being placed on
energy efficient designs of all energy consuming equipment. Electricity is the main mode of energy supply and
all major industries are affected by the increasing cost of this commodity.
In South Africa with its deep and hot mines some of the main electricity consumers are mine ventilation
and cooling systems. In this regard, a number of projects are being undertaken to implement energy efficient
mine ventilation and cooling/refrigeration designs. These include new generation surface cooling installations
which incorporate thermal storage systems, active control and cyclical operation of ventilation and cooling
systems, and systems in which in-mine ventilation parameters are monitored at a few strategic locations to
predict conditions throughout the rest of the mine. This paper will describe the concepts and will discuss results
from actual implementations of mine systems.
The principles of design of energy efficient ventilation and cooling systems apply to all of the system design
and operating stages. It starts as early as setting primary design criteria and continues to the active operating of
supply systems throughout the life-of-mine. System energy efficiency is maximised by determining optimum
design airflow quantities and wet-bulb temperatures, using innovative system components, cleverly integrating
already efficient systems, and finally by cyclical and on-demand operation.
The necessary knowledge and technology exist in most cases to drastically improve the energy efficiency
of ventilation and cooling systems, and it is a matter of applying sound principles and being energy conscious
during the design phase. International pressure to preserve global energy resources dictates this approach in all
future mine designs.

1 INTRODUCTION designs for mine ventilation and cooling systems.

A concern over global energy resources has resulted in
an increased emphasis being placed on energy efficient
designs of all energy consuming equipment. Electric-
ity is the main mode of energy supply and all major
industries are affected by the increasing cost of this
commodity.
In South Africa with its deep and hot mines some
of the main electricity consumers are mine ventilation
and cooling systems and the operating costs associ-
ated with these systems have increased significantly
over the last decade (Bluhm et al. 2003). The increase
is not only due to electricity costs, but also as a
result of increasing water costs and a higher demand
for ventilation and cooling. Electricity and water
costs are increasing at rates higher than inflation and
this is a world-wide trend that is likely to continue.
Higher demand for ventilation and cooling is driven
by increased mechanization, increase mining depth
and the general drive to improve underground working
conditions.
As a result of increased operating costs, consid-
erable emphasis is being placed on energy efficient
There are several stages during the design, implemen-
tation and operation of mine ventilation and cooling
systems where energy efficiency could and should be
optimised. The initial stage is planning and optimisa-
tion of primary variables for the proposed system, next
individual system components are optimally designed,
followed by the efficient integration of these compo-
nents. Finally energy efficiency is improved during
operation through cyclical control of systems and
ultimately by supplying ventilation and cooling on
demand. Any reduction of electrical power cost must
be evaluated in relation to the effect on production and
the overall mine operating cost per ton mined.
A number of projects are being undertaken in South
Africa to implement energy efficient mine ventilation
and cooling/refrigeration designs. These include for
example, modern surface cooling installations which
incorporate ice bank thermal storage (Wilson et al.
2003), monitoring systems to allow prediction of con-
ditions throughout the mine based on only a few
strategic measurements (von Glehn & Ox. 2004), as
well as energy management through active control
and cyclical operation of ventilation and cooling

279

Copyright © 2006 Taylor & Francis Group plc, London, UK

 systems (Gundersen et al. 2005, Wilson et al. 2005).
The following sections describe these concepts accord-
ing to the stages of energy efficiency optimisation,
from setting design criteria to active operational sys-
tem control, using recent projects as examples.
2 PLANNING AND OPTIMISATION OF
PRIMARY VARIABLES
Ventilation planning has to account for all known fac-
tors which may have an impact on the underground
environment. Any constraints and standards to be
applied (pollutant levels, temperatures, velocities, etc.)
should be specified before developing any ventila-
tion distribution system. Design constraints provide
a check to ensure that all ventilation requirements are
accounted for in the planning process. A formal doc-
umented list has to be prepared early in the design
process as these constraints form the basis for heat and
pollutant limits and environmental acceptance criteria.
Typical design criteria are the following (Bluhm et al.
2003):
• Existing mine characteristics and constraints.
• Mining method, production and development
schedules.
• Service ventilation requirements (workshops, pump
stations, sub-stations, etc).
• Main and secondary shaft engineering requirements
and infrastructure.
• Underground equipment, vehicles and work cycles.
• Geothermal data.
• Fissure water expectation and strata gas emission
expectation.
• Existing thermal criteria (temperature, cooling
power) and airborne contaminant design criteria
(dust, gases, diesel exhaust, radiation).
• Service water quality and usage.
• Power cost data and expected project rate-of-return.
• Surface weather data.
Efficient planning is a process that optimises the
primary system variables. Energy efficient planning
requires, for instance, a fine balance between ven-
tilation air quantity and applied cooling to achieve
the desired dilution of heat and other pollutants, eco-
nomic airway sizing ensures optimised air speeds and
optimum ventilation/cooling vs. production ratios. A
typical example of an energy efficiency optimisation
is shown in Figure 1, (Bluhm et al. 2003). Carefully
specifying ventilation network aerodynamic and air
thermodynamic characteristics plays a major role in
an ultimate energy efficient design.
It must be noted that the classical optimum is not
always the best solution and systems must be designed
to provide flexibility for possible unforeseen changes
280
Copyright © 2006 Taylor & Francis Group plc, London, UK
Power cost
Total power cost
Main and recirculation fans
Refrigeration and pumping
for spray chamber
Refrigeration for BAC
0 200 400 600 800 1000 1200
TOTAL DOWNCAST (kg/s)
600 400 200 0
RECIRCULATED FLOW (kg/s)
Figure 1. Typical optimisation for energy efficiency.
to the design constraints. These changes could typ-
ically be changes to the mining method, production
and development schedules.
3 DESIGNING ENERGY EFFICIENT SYSTEM
COMPONENTS
System components include the airflow network (air-
ways), ventilation controls, fan stations, refrigeration
plants, cooling distribution network (chilled water pip-
ing), individual air coolers (heat exchangers), etc.
There are basic rules and standard practises regard-
ing the efficiency of each of these components, which
should be applied as a matter of principle. In addition to
generic efficiency optimisation, innovative and intel-
ligent engineering design and operation can signifi-
cantly improve the energy efficiency of components
further.
An example of engineering design aimed at energy
efficiency is thermal storage, which is often engi-
neered into mine cooling systems. Traditionally ther-
mal storage systems have consisted of large chilled
water dams that are typically filled with chilled water
at night during periods of low demand and low ambient
temperature. The chilled water is then used for mine
cooling via air-cooling installations or as chilled mine
service water.
An improvement over chilled water dams is the use
of ice coil banks for thermal storage.
 Warm
ambient air
Air cooling heat exchanger
Cold
By-pass valve
downcast air
Water chiller
Ice coils
Glycol chiller
Figure 2. General concept of ice thermal storage air cooler.
Figure 3. Ice on submerged ice coil banks.
Ice thermal storage is well developed in air condi-
tioning systems especially in district-cooling utilities.
The concept is to produce ice on the outside of sub-
merged heat exchanger coils during periods of low
power cost and melt this ice during periods of high
power cost and peak demand. Installations making and
melting over 1000 tons of ice per day are common in air
conditioning systems in North America and Europe. A
schematic of a similar application that recently proved
successful for mine cooling is shown in Figure 2
with a photograph of the ice formed shown in Fig-
ure 3 (Wilson et al. 2005). Although of similar design,
the mine cooling application is significantly different
when compared to the district-cooling approach.
Differences lie mainly in the objectives of these sys-
tems. Mine cooling applications aim to produce 0◦ C
water, which is used in direct contact heat exchangers
to produce very cold air. Another advantage of ice ther-
mal storage systems is that refrigeration plants could
be sized for average loads and these plants can operate
continuously at full load (high efficiency). Mine ice
thermal storage also allows for power consumption
profiling.
Ice thermal storage installations can store up to
500 kg of ice per external cubic meter of sub-
merged coil, relating to a thermal storage capacity
of 170 MJ/m3 . Equivalent systems using chilled water
281
Copyright © 2006 Taylor & Francis Group plc, London, UK
storage require four times the volume to provide the
same thermal store. Ice thermal storage systems can
provide water chilled to almost 0◦ C that can be used in
simple direct contact heat exchangers to cool ambient
air to temperatures as low as 2◦ C. Cold water can be
piped underground for use in underground air cool-
ers or as chilled service water. In addition, ice thermal
storage units have a more compact design than their
water storage counterparts.
The concept of ice thermal storage is further ide-
ally suited for the local electricity supplier’s (ESKOM)
Demand Side Management (DSM) programme, which
is aimed at shifting electrical load out of peak demand
periods. Ice is produced during ESKOM standard and
off-peak power periods and subsequently, this ice is
melted by diverting water flow to the ice coils to pro-
vide cooling during ESKOM peak power periods thus
reducing the load on all the water chillers.
The benefit to the mine is lower power costs where
the benefit to ESKOM is the power shifting from peak
periods.
Similar innovations exist for almost all individual
ventilation system components and should be consid-
ered when designing a new system.
4 INTEGRATION OF ENERGY EFFICIENT
COMPONENTS
The previous example highlights the importance of
properly integrating energy efficient components.
Mine heat loads drive the need for the use of cold air
and there exists a critical interdependence between the
cooling and ventilation systems.An example is a recent
study to optimise the ventilation and cooling system
for a block cave operation.
Characteristics of the specific mine are:
• Full calendar operation (24 hour/day, 7 days/week)
• Virgin rock temperature on loading level of approx-
imately 38◦ C
• Large concentration of diesel fleet in loading area
• Underground crushing
• Conveyor belt rock transport to shaft area
• Upto 1 million tons production per month
• Large bulk air cooler (BAC) on surface.
The point to note is that the mine requires a signifi-
cant amount of air and cooling to dilute heat and diesel
pollutants. System optimisation and integration started
with the crusher and workshop ventilation systems.
These were traditionally ventilated in series and
semi-series with production cross-cuts, resulting in
dust and heat from these areas being introduced to
the main production area (and causing major risk in
case of fire). Optimisation was done by determin-
ing the optimum air quantity required by the crusher
stations and workshop area and altering the system
layout to ventilate these areas directly to return.
 Simulated Airflow Distribution indicates cross-cut wet-bulb temperatures for different
40 cooling strategies.
35
30 In this case, the ultimate system will not use less
Airflow m /s
3

25 electrical energy or reduce the electrical power cost.

20
15
However, cross-cut flows have been doubled, exces-
10 sive wet-bulb temperatures reduced and dust and diesel
5
0
pollutant levels reduced ensuring improved productiv-
x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c ity and health of the workforce. This will result in an
1 2 3 4 5 6 7 8 9 10 12 13 14 15 16 17 18 19 20 increase in production and ultimately improved effi-
Cross-cut Number
ciency through a reduction in overall mine operating
Simulated Ave Simulated
cost per ton mined.
Figure 4. Optimized cross-cut air flow distribution.
5 SYSTEM MONITORING FOR ENERGY
BAC Only vs. BAC and Spray-chamber
EFFICIENCY CONTROL
Temperature ˚C Wet-bulb

34
32
It is often stated that, ‘you can’t control what you can’t
30
28
measure’.
26
Due to the availability of improved instrumentation
24 and communication systems, it is becoming easier for
22 mine operators to monitor underground environmental
x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c x|c conditions. However, the extensive number of airways
1 2 3 4 5 6 7 8 9 10 12 13 14 15 16 17 18 19 20
Cross-cut Number and large range of variables and influencing factors
BAC Only BAC & U/G Average Average
make it impractical to locate remote instruments all
over a mine. It is more practical to have a few critical
Figure 5. Optimized cross-cut temperature distribution. measurement sites that are fully reliable and to use sim-
ulation software to extrapolate from this data to predict
conditions throughout the mine. The VUMA-network
Next the airflow distribution to the cross-cuts
simulation software was recently improved to include
(20 off ) was optimised. This involved regulation of
a feature to incorporate critical measurements into the
some intakes, increasing the number of return passes
simulation network, which is used to predict condi-
and optimising the positions of these (for new passes).
tions everywhere else in the mine. Software modules
The result was a much-improved distribution with
that were developed to enable this are capable of:
cross-cut quantities not differing by more than 20%
throughout the system. • Communicating between the mine monitoring sys-
Following this the study concentrated on increas- tem and VUMA-network
ing the cross-cut flow quantities to the required level • Real-time solving and calibration of mine models
(almost double the current flow). Minimising system • Anticipating potential problems by displaying warn-
pressure losses via simulation and optimising fan per- ings when changes in measured parameters exceed
formance and air speeds in in-takes increased cross-cut specified limits.
flows significantly but did not achieve the required
By linking these software modules to VUMA-
level of increase. A careful analyses of the system indi-
network, users can use existing simulation network
cated the only option to achieve this was to re-use some
information and there is no need to switch between
air. The proposed re-use air currently ventilates the
two different programs. It is obviously important that
conveyor belt drive and the crushers and therefore has
the standard VUMA-network file accurately repre-
to be re-conditioned before re-use. Simulations indi-
sents the current mine layout. The system increases
cated an increase in cross-cut quantities to the required
the coverage provided by existing instruments through
level. Figure 4 shows cross-cut flow distribution after
extrapolation of measured values in VUMA-network
optimization and re-use of air.
to provide an expanded real-time view of the mine.
Finally a heat energy balance indicated a poor com-
The modules include a Linker to a mine control
bination of cooling and ventilation distribution. In
and data acquisition (SCADA) system, Diagnostic
general intake airways and the first few production
faulty-instrument checker, Comparison-solver (which
cross-cuts are overcooled and cross-cuts at the extreme
includes calls toVUMA-network solver) and Reporter.
of the mine are undercooled causing excessively high
The following briefly describes the operation of the
wet-bulb temperatures. The study showed that some of
system:
the available cooling should be applied in the surface
air cooler (but less than the current level) and some • The Linker interacts with the mine SCADA sys-
in a secondary air cooler near the cross-cuts. Figure 5 tem through a common open interface (OPC).

282

Copyright © 2006 Taylor & Francis Group plc, London, UK

 The Linker matches measurement locations with
nodes within a VUMA-network data file and passes
measured data to the Comparison solver.
• The Linker screens the input data for ‘noise’and also
carries out some basic smoothing of data to min-
imise false alarms as a result of sudden but short-
lived extreme changes in measured conditions.
• The Diagnostic faulty-instrument checker scans the
measured data for validity and consistency. If the
measured reading is not valid or consistent with pre-
vious readings or with neighbouring measurements,
a report is produced and the program is returned to
the control of the Linker. The user will be requested
to check the relevant instrument(s). Checks carried
out include whether the instrument is displaying a
fault code, whether the measured data are within
the range of the instrument and whether the mea-
sured value is within acceptable range for current
conditions.
• The Comparison solver calls the VUMA-solver to
establish a reference prediction based on user input.
The predicted values at the first measurement node
are compared with the actual measurements and
the difference displayed. For subsequent steps the
parameters for that node are defined to be the same
as the measured parameters so that downstream
branches in the computer network will have the
same starting point for calculation as the actual
ventilation system. The VUMA solver is repeatedly
called until the parameters at each measured node
have been compared with the predicted values.
• When the final measurement point has been com-
pared or calibrated, all predicted values at mea-
surement nodes should correspond exactly with
measurements.
• Control is then handed back to the Linker to await
the next series of measurements.
The concept of limited measuring points and pre-
diction of the rest not only gives a real-time view of the
complete mine ventilation network, but is also a power-
ful alarm system highlighting any abnormal conditions
as well as the downstream effect of these. Although
the system is only used for monitoring currently, the
ultimate aim is to allow full network monitoring and
control. Network control is the subject of the next two
sections.
6 ENERGY EFFICIENCY THROUGH
CYCLICAL VENTILATION AND COOLING
CONTROL
Conventionally all areas of a mine are been ventilated
and cooled on a fulltime basis. This approach does
not acknowledge mining as a cyclical operation, or
cyclical changes in ambient conditions or variations
283
Copyright © 2006 Taylor & Francis Group plc, London, UK
in electrical power costs during peak periods. It is evi-
dent that the design of energy efficient ventilation and
cooling systems requires new thinking in this regard.
Cyclical ventilation control strategies include reduc-
ing air cooler duty and ventilation airflow rates during
periods of low air-cooling and ventilation demand. Ice
bank thermal storage was earlier used as an example
of an energy efficient system component. Surface bulk
air cooler duty can be easily controlled using this tech-
nology, which is ideally suited for use with cyclical
ventilation and cooling strategies.
A typical mining cycle in conventional hard rock
mining consists of two eight-hour shifts per day, one
mainly for drilling and charging up and one for remov-
ing the broken rock from the production zone. The
third eight-hour period is dedicated to face blast-
ing, normally in the afternoon. Mining crews in this
case occupy the production zone for a maximum of
two-thirds of the day. Although there will be person-
nel underground during the remaining third of the
day, they will be in intake airways close to the main
shaft. There is another shorter period between the
two “active” shifts where the production zones are
unoccupied when shift change takes place.
This cyclical schedule lends itself perfectly for
energy efficient ventilation and cooling system oper-
ation. When production zones are unoccupied there
is no need to continue supplying the normal air
quantity and quality to this area. It is therefore pos-
sible to reduce fan power and refrigeration during
these times in a controlled manner. The design will
require heat flow simulations to predict transient ther-
mal effects and to ensure average thermal conditions
remain acceptable. In addition it will have to be con-
firmed that a reduced airflow quantity could clear all
dust, gases and blasting fumes during the unoccupied
period.
The benefits of reducing overall fan power
and refrigeration/cooling supply lie in the reduced
absorbed electrical power. If a cyclical control strat-
egy is adopted, the electrical power associated with
ventilation and cooling systems can be reduced by
12% for ultra-deep mines (Ramsden et al. 2001) and
by 20% for deep mines with air coolers located on
surface (Wilson et al. 2005). These systems have to
be carefully designed due to the increased safety risk
and more demanding control and maintenance, as well
as existing constraints such as diesel pollutant and
blast fume removal and acceptable face velocities.
However, operation is cyclical and therefore reason-
ably predictable and personnel and mining systems
will adapt to it rapidly.
Cyclical control of total airflow quantity can be
achieved by using main fans fitted with variable speed
drive motors or variable inlet guide vanes. Cycli-
cal control of air cooling duty is easily achieved by
using automated valves to change cooling water flow
 rates and/or temperatures. Conventional refrigeration
machines could be switched off during these periods
or thermal storage could be used to store cooling for
periods of high cooling demand.
Although the effect of thermal storage in the sur-
rounding rock might require a slight increase in overall
required refrigeration duty by no higher than 10%
(Gundersen et al. 2005, Ramsden et al. 2001), there is
still great merit in cyclical ventilation control. Similar
systems are currently being installed in South African
mines as part of a national drive by the electricity sup-
ply body to reduce peak electricity demand. Cooling is
generated during periods of low electricity demand and
used during peak demand periods when refrigeration
plants are switched off.
7 ULTIMATE ENERGY EFFICIENCY BY
VENTILATION/COOLING ON DEMAND
Ventilation-on-demand and cooling-on-demand have
the potential to reduce both capital and operating costs
of mine ventilation and cooling systems and the mech-
anisms required have proved to be technically feasible.
These systems are generally more flexible than the
cyclical systems described previously and entail sup-
plying the necessary airflow quantity and quality as
and when needed. As with cyclical control these
designs have to cater for increased risk and demands
on management and control.
Supply-on-demand systems operate on a concept
of monitoring and control with air pollutant levels and
wet-bulb temperatures typically triggering supply. An
example is secondary ventilation systems that are re-
directed to follow the diesel fleet, reducing flow and
cooling when the fleet leaves an area and increasing
resources when the fleet enters another area.
Complementary to the innovative monitoring sys-
tem described previously, strategic placement of sen-
sors when using VUMA-live can assist the active
control of ventilation and cooling systems. The system
could identify whether operators are correctly switch-
ing fans when they enter/leave an area, and it would
be able to continuously assess whether sufficient air
is being supplied to working places. A central con-
trol station would be alerted whenever conditions do
not satisfy minimum requirements. If the program is
linked to a central vehicle despatch system, it would
ensure that sufficient air follows equipment as it moves
through the mine.
Gundersen et al. (2005) discuss potential energy
and cost savings achieved in a number of case studies.
By implementing a ventilation-on-demand strategy
running cost savings of approximately 10% can be
achieved. Implementing a cooling-on-demand strat-
egy‘running cost savings of 15% can be achieved, or
10% additional cooling can be provided during the
on-period for the same running cost.
284
Copyright © 2006 Taylor & Francis Group plc, London, UK
The potential improvement in energy efficiency
through on-demand ventilation and cooling is obvi-
ous. These systems require additional effort during
the design phase, but benefits are realised over the
life-of-mine making the effort more than worth while.
8 CONCLUSION
The principles of design of energy efficient ventilation
and cooling systems apply to all of the system design
and operating stages. It starts as early as setting pri-
mary design criteria and continues to the active oper-
ating of supply systems throughout the life-of-mine.
System energy efficiency is maximised by deter-
mining optimum design airflow quantities and
wet-bulb temperatures, using innovative system com-
ponents, cleverly integrating already efficient systems,
and finally by cyclical and on-demand operation.
The necessary knowledge and technology exist in
most cases to drastically improve the energy efficiency
of ventilation and cooling systems, and it is a matter of
applying sound principles and being energy conscious
during the design phase. International pressure to pre-
serve global energy resources dictates this approach in
all future mine designs.
REFERENCES
Bluhm, S.J., von Glehn, F.H. and Smit, H. 2003. Important
basics of mine ventilation and cooling planning, Proceed-
ings of the MVS Annual Conference, Managing the Basics,
February 2003, Pretoria, South Africa, Mine Ventilation
Society of South Africa.
Gundersen, R.E., von Glehn F.H. and Wilson, R.W. 2005.
Improving the efficiency of mine ventilation and cool-
ing systems through active control. Proceedings of the
8th International Mine Ventilation Congress, June 2005,
Brisbane, Australia, AusIMM.
Ramsden, R., Sheer, T.J. and Butterworth, M.D. 2001. Design
and simulation of ultra-deep mine cooling systems.
Proceedings of the 7th International Mine Ventilation
Congress, Krakow, June 2001, Poland, EMAG.
von Glehn, F.H. and Ox, I. 2004. Monitoring and control of
underground ventilation systems using VUMA-network.
Proceedings of the 10th US Mine Ventilation Symposium,
May 2004, Anchorage, Alaska, USA, Balkema.
Wilson R.W, Bluhm, S.J., Smit, H. and Funnel, R.C. 2003.
Surface bulk air cooler concepts producing ultra-cold air
and utilising ice thermal storage. Proceedings of the MVS
Annual Conference, Managing the Basics, February 2003,
Pretoria, South Africa, Mine Ventilation Society of South
Africa.
Wilson, R.W., Bluhm, S.J. and von Glehn F.H. 2005. Ther-
mal storage and cyclical control of mine cooling systems.
Journal of the Mine Ventilation Society of South Africa,
October–December 2005.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Integration of strategic ventilation and heat management planning

with 3D virtual reality simulation of underground mines

R. Moreby
University of New South Wales, Sydney, NSW, Australia

J. Whitmore & P. Dunn

MIRARCO, Laurentian University, Sudbury, Ontario, Canada

ABSTRACT: This paper describes methods employed to integrate strategic ventilation and heat management
planning into a virtual reality (VR) simulation tool to enable various scheduling scenarios to be investigated.
This can be performed even when infrastructure details are unknown or not yet finalised. The approach taken is
to apply conventional activity based determination of ventilation requirements for each stage of the project. In
addition, the spatial location of each element of the model, combined with allocated ventilation rates, geothermal
properties of the ore body and surface climatic profiles, are used to predict wet and dry bulb temperatures in
all intake airways and working locations. Consequently, both the nature of, and schedule for, heat management
strategies can be determined.

1 INTRODUCTION a means of producing a ‘strait-jacket’ ”, (Burrows et al.

Ventilation network simulation software, including
heat gain, refrigeration and fire simulation routines,
are now readily available for use on desk top com-
puters, for example Ventsim, VnetPC, Vuma and
Ventgraph. This software provides for detailed and
sophisticated analysis of most ventilation networks,
including graphical user interfaces, from simple single
shaft, three panel coal mine circuits to large metallifer-
ous mines involving perhaps 20 to 30 shafts, multiple
development and stope locations together with 3000 to
4000 m3 /s of air, for example the Olympic Dam oper-
ation in South Australia (Hall and Clarke 1988) and
the Mt Isa mine in Queensland.
However, with respect to strategic planning at the
feasibility stage, it is necessary to determine how ven-
tilation and or heat management requirements will
change with time as the mine expands laterally and
in depth. In smaller mines this may be straightfor-
ward and conventional analysis remains appropriate.
However, in larger mines the sometimes overwhelm-
ing complexity can make decision making difficult
particularly when the mine geometry changes rapidly
with time and infrastructure details are not yet
finalised.
A prevailing risk is that mines will embark on overly
optimistic production schedules with ventilation plans
not having sufficient contingency for variations that
may occur. Or, with respect to detailed analysis, “Siz-
ing airways according to a neat ‘economic’size can be
1989) and “Although mine plans continually evolve, it
is important for far sighted decision making to identify
overall design constraints and critical design scenar-
ios for the longer term” (Bluhm & von Glehn 2004).
In these circumstances, future ventilation related
problems, such as inadequate shaft dimensions or
inadequate cooling capacity, often do not require
sophisticated software for identification. These prob-
lems may be further exacerbated when persons in key
decision making or capital allocation roles do not have
sufficient ventilation expertise to fully understand the
inevitable consequences of their design choices.
In 2001, MIRARCO (Laurentian University)
opened a state-of-the-art Virtual Reality Laboratory
(VRL) and have developed a Mining Exploratorium
(Figure 1) that is unique in the world and offers an
exceptional data interpretation environment for a myr-
iad of purposes. As a result of developing this unique
capability, MIRARCO has gained invaluable expertise
in the area of data integration and has demonstrated
how value can be derived in the decision-making pro-
cess when the Collaborative Immersive Virtual Reality
(CIVR) environment is used. Some key benefits of
using CIVR in the minerals industry are:
• Data fusion (all digital data in one common VR
model)
• Knowledge integration (getting the relevant expert-
ise to work on data) and transfer (speed of compre-
hension/evaluation of subsurface models)

285

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 1. Collaborative Immersive Virtual Reality facility.
• Technical conflict resolution, and
• Multi-disciplinary team collaboration (consensus
building).
Dunn & van Wageningen (2003) have developed
the CIVR into a 5D mining exploratorium where the
3D data can be viewed with time (4th dimension) and
any other mining property (5th dimension) tracked and
visualised with time. It was therefore an obvious pro-
gression to create a VR ventilation simulation tool to
enable the tracking of ventilation requirements with
time and integration with all other mine planning data
for decision making purposes.
The overall objective of the VR ventilation sim-
ulation model described in this paper is to reduce
complexity and highlight consequences of a mine’s
production schedules using a systematic approach
combined using fundamental design parameters.
Guideline design parameters include those used world-
wide for limiting velocities, volumetric requirements
for diesel equipment, sources of heat and limiting air
temperatures described by, inter alia, Burrows et al.
1989, Wallace 2003, Bluhm & von Glehn 2004, and
Fourie 2004. In addition, site specific parameters are
obtained by selected climate simulation using geother-
mal properties of the ore body, for example wet bulb
depression related to rates of evaporation for the
degree of airway wetness.
This model is a continuation of work previously
undertaken in Australian coal mines that has suc-
cessfully provided faster and more simplistic analysis
of networks without the need for overly complex
simulation (Moreby 2002).
The fundamental difference between this approach
and more detailed analysis is that ventilation and
heat management requirements, including volumetric
capacity, geometry of the primary circuit and refriger-
ation load profiles, can be determined to tolerances
required for feasibility stage capital budgets without
286
Copyright © 2006 Taylor & Francis Group plc, London, UK
detailed infrastructure being known or time consum-
ing detailed network simulation being undertaken. The
purpose is therefore to supplement, rather than replace,
more sophisticated analysis that will of course still be
required for final design and project implementation.
2 VOLUMETRIC REQUIREMENTS
At any point in time, the volumetric requirements of
a mine are dependent on the number and type of
mining activities, quantity of air allocated to each
activity type together with an allowance for fixed
locations such as ramp systems, crushers, fuel bays
and workshops. The quantity of air allocated to each
activity type typically, in the absence of significant
strata gas emission, being determined by the amount
of diesel equipment employed together with accept-
able air velocities and air residence times for dilution
and clearance of blasting fumes.
InAustralian metalliferous mines, the quantity of air
required for diesel equipment ranges 0.04 to 0.05 m3 /s
per rated kW (over the engine) with values of 0.06
to 0.10 m3 /s per rated kW currently being typical for
overall or mine section ventilation in metalliferous
mine design worldwide. Quite simply, the amount
of air required for management of diesel exhaust
emission is determined by the rated power of diesel
units present and the appropriate ventilation factor, for
example, 10 to 12.5 m3 /s per 250 kW rated unit.
For ramps, draw points or stock piles the quantity of
air is calculated pro rata for the rated power of diesel
equipment present. That is, at any point in time, it is
the number and location of mining or rock handling
activities that determines the number and location of
diesel units in the mine for which adequate ventilation
must be supplied.
The increasing concern about management of diesel
particulate may cause a problem with respect to spe-
cific volumetric requirements (m3 /s per kW) possibly
increasing to 0.16 m3 /s per kW if dilution alone is
used to meet ever decreasing threshold limiting values
(Wallace 2003). This is a variable that can be readily
changed in the model to assess the effects of such a
reduction.
With respect to clearance of blasting fumes and
heat management in most standard length (100 to
400 m) development headings under auxiliary ventila-
tion, face velocities should be between 0.5 and 0.75m/s
(Howes 1988) or ranging 12 to 25 m3 /s for 5 m × 5 m
or 5 m × 6 m excavation profiles. These requirements
are therefore consistent with that for diesel equipment
normally employed in such locations.
During the development phase of a project it is
invariably required to ventilate several development
faces at the same time in order to maximise utilisa-
tion of equipment in the charge, blast, bolt & load
 Table 1. Example activity based volumetric analysis. Table 2. Guideline volumetric requirements for mining
types.
m3 /s per Quantity
Mining activity Locations location m3 /s m3 /s per 1000 tpy Tonne air/tonne ore
Mining type Low High Low High
Ramp development 2 35 70
Level development 7 35 245
Loading horizons 4 20 80 Block cave 0.05 0.1 1.9 3.8
Drilling horizons 4 15 60 with good
Ring firing locations 3 15 45 control
Stope filling 3 15 45 Large mines, 0.2 0.3 7.6 11.4
Diamond drilling 2 15 30 long hole
Raise boring 2 15 30 stopes
Ground support 2 15 30 Smaller mines, 0.3 0.6 11.4 22.7
Service/maintenance 3 30 90 various
bays degrees of
Rock handling 125 control
and crushing Pillar recovery 0.6 0.8 22.7 30.3
operations,
Total 850 poor control

Underground 850 m3 /s
quantity
Production 3,000,000 tpy An analysis of this type provides a quantification of
Specific ventilation 0.28 m3 /s per 1000 tpy gross ventilation requirements with time, often indi-
Mean depth of 1200 m cating the volume of air required for the development
mining phase of a project being similar to, or exceeding, that
Volume of air on 944 m3 /s
surface
for the production phase. Importantly, these are the
Intake shaft 94.4 m2 minimum volumetric requirements for management of
area at 10 m/s gaseous or particulate contamination. Higher volumes
Intake level 189 m2 may be required for management of heat subject to
area at 5 m/s practicable cooling capacities at the depth of mining.
Exhaust shaft 55.6 m2 In addition, total volumetric requirements can then
area at 17 m/s be reconciled with “rule of thumb” or guideline values
for generic mining types, Table 2.

then drill production cycle. In these circumstances a 3 SOURCES OF HEAT

certain degree of series ventilation may be accept-
able provided that the accumulation of contaminants,
including heat, remains below statutory or design
limits. That is, the amount of air required at a certain
location will be a defined fraction of the activity based
total. A similar situation arises if the mine employs
PLC control of auxiliary fan duties to modify demand
in response to the activities being undertaken.
The VR ventilation simulation model identifies the
number, location and type of active “elements” present
in the mine at any point in the production schedule
from which all of the above volumetric requirements
are calculated using simple arithmetic summation.
A summary of analysis for a 3.0 Mtpy long hole
stope operation, 1000 m to 1400 m below surface is
shown in Table 1. In this plan, higher ventilation rates
were allocated to development locations for manage-
ment of heat and to provide for loading stock piles
under auxiliary ventilation. The cross sectional area
of intake hoisting shafts were identified as being inad-
equate and were later supplemented with dedicated
intake ventilation shafts.
Assuming that atmospheric contaminants resulting
from blasting operations, emission of strata gases and
the use of diesel equipment are adequately addressed
by provision of acceptable volumetric capacity, the
next stage of design in many projects, as a result
of depth of workings or surface climatic temperature
profiles, is to quantify heat management requirements.
Computational techniques for the quantification of
mine heat loads resulting from various natural (sur-
face climate, auto compression and strata) or artificial
(mine equipment and explosives) sources of heat are
widely documented and readily calculated from basic
thermodynamic and psychrometric analysis (Hemp
1989, Marais 1992, van der Walt 1986, McPherson
1993, Rawlins 2004).
These techniques are not new to traditionally “hot”
mines such as those in Southern Africa or north-
ern latitudes of Australia. However, the operational
consequences of increasing heat loads in a number
of “cool” mining locations, due to increased depth
(Canada and Sweden) or increasing equipment heat

287

Copyright © 2006 Taylor & Francis Group plc, London, UK

 loads kW per m3 /s (Australian coal mine longwall 25
Queensland (Emerald)
faces), is that mines previously considered not to have 20
significant problems are now having to manage heat 15 95% design =
to increasing degrees. In some locations, non refriger-

Surface Wet Bulb°C

24.0°C WB
10
ation based controls such as equipment segregation or
5 95% design =
increased volumetric capacity are adequate while in 12.0°C WB
other, typically deeper, operations refrigeration must 0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
be introduced. For example, the Onaping depth project -5
(Howes & Sockhaus 2002). -10
Canada (Sudbury)
Recognising that the stated aim of the VR venti-
-15
lation simulation model is to provide ventilation and
-20
heat management requirements without sophisticated
thermodynamic network simulation, the techniques
Figure 2. Australian and Canadian annual surface wet bulb
and rationale for calculation of air temperatures temperature profiles.
employed by the model are described in the following
sections. Depth
40 3000m
3.1 Surface temperature profiles 2000m

U/G Wet Bulb Temperature°C

30 Design U/G intake 26°C WB 1000m
Surface climatic temperatures are obtained from
20
weather bureau statistics to determine average, maxi- +4°C WB per km
mum and 95% design surface wet bulb temperatures 10
for the location being considered. For example, the Critical surface
seasonal variation in average surface wet bulb tem- 0 y=x temperatures
peratures for Emerald in Queensland, Australia and -10
Sudbury, Canada are shown in Figure 2. Sudbury Emerald
+8°C WB per km
For the purposes of the VR ventilation simulation -20
model, wet bulb temperatures are used as the foun- -20 -10 0 10 20 30

dation for heat simulation with dry bulb temperatures Surface Wet Bulb Temperature°C
determined from climate simulation software for vari-
ous areas of a mine. That is, in intake areas the dry Figure 3. Theoretical increase in wet bulb temperature with
depth.
bulb depression will typically be 5 to 10◦ C and 1 to
5◦ C in face areas subject to how wet the airways are.
Air Temperature°C Air Temperature°C
This issue is discussed further below.
10 15 20 25 30 10 15 20 25 30
0
3.2 Auto compression Damp Dry
shaft DB =1.0°C/km shaft
For any specified surface temperature, the wet bulb 200
Dry
gradient in intake shafts is calculated as a function bulb
of depth and surface temperature i.e. surface plus 400
Dry
Depth m

9.8 kJ/kg per km. Notably, this relationship is not linear bulb
390m3/s
with the actual wet bulb temperature rise (◦ C per km) 600
Wet
reducing with increasing intake wet bulb temperature 320m3/s bulb
800
(Fig. 3). Wet
The adiabatic increase in dry bulb temperature with bulb DB = 7.0°C/km
depth is about 9.8◦ C per km providing airway surfaces 1000 WB = 3.9°C/km

are perfectly dry. In practice, the presence of water in DB = 9.5°C/km

intake shafts and certainly in intake ramp systems, will
reduce the rise in dry bulb temperature. For example,
the observed increase in air wet and dry bulb tempera-
tures in two shafts (visibly dry and visibly damp) in an
Australian metalliferous mine are shown in Figure 4.
In the case of intake ramps systems with dust
suppression sprays and or hot dry surface climatic
conditions, the dry bulb gradient can in fact be neg-
ative as it is in wet intake shafts. These factors are
incorporated in the model by pre determining values
for wet bulb depression from climate simulation. As
1200
WB = 4.5°C/km
Figure 4. Observed increase of air temperature with depth.
identified by Lambrechts (1985), calculation of air dry
bulb temperature requires a site specific and subjective
judgment of degree of wetness to be made regardless
of the sophistication of simulation software employed.
It is for this reason that wet bulb temperatures form
the basis of this model with air dry bulb temperatures
being more of an approximation requiring a certain

288

Copyright © 2006 Taylor & Francis Group plc, London, UK

 100 10 Heat from strata 3.5

Heat W/m per °C (tvr - WB)

90 9
3.0

Wet Bulb Gradient°C/km

80 8
Rock Temperature°C

Queensland Australian Sudbury Increasing convective heat

70 coal metalliferous 7 transfer coefficient 2.5
Canadian
60 metalliferous 6 2.0
50 5 tair = 20°C tvr = 44°C
40 Onaping 4 1.5
Kidd Creek
30 3 1.0
20 2
10 Resultant wet bulb gradient 0.5
1
0 0 0.0
0 500 1000 1500 2000 2500 3000 0.0 1.0 2.0 3.0 4.0 5.0
Surface range Depth m Air velocity m/s

Figure 5. Australian and Canadian geothermal gradients. Figure 7. Effect of air velocity on heat transfer from strata
and resultant wet bulb temperature.
40
Heat W/m per°C (tvr - WB)

35
30
Mean “working” airways
25
20
Mean “new” airways
15
10
5
0
0 6
Figure 6. Time dependent change in heat transfer from
strata.
degree of tuning to observed data. It has been found
that a relationship between evaporation rate and heat
load (g/s per kW) can also be determined to assist
in these calculations. In this respect errors are likely
to reduce in face areas in any event due to increased
evaporation rates and reduced wet bulb depression.
3.3 Strata surfaces
Heat transfer from airway surfaces is determined by a
number of parameters, but most significant are thermal
properties of strata (diffusivity m/s2 and conductivity
W/m◦ C), virgin strata temperature tvr, air temperature,
age of airway and air velocity (Hemp 1985).
Virgin strata temperatures are determined from the
specified surface temperature and geothermal gra-
dient, for example those shown for Australian and
Canadian conditions in Figure 5. It is interesting to
note the effect of the surface temperature offset of
some 15◦ C between Australia and Canada on strata
temperatures at a given depth.
Using techniques described by Hemp (1985) and
Howes (1988) or climate simulation software, such as
Climsim, it is relatively straightforward task to cal-
culate generic site specific strata heat transfer curves
(kW per km or W per m per ◦ C (tvr – tair ), (Fig. 6)).
That is, knowing the age of an airway element
and the difference between intake air temperature and
strata temperature at its known depth, the heat trans-
fer rate can be determined with a reasonable degree of
accuracy. For the purposes of this work, and assum-
Mean “old” airways
ing partially wet mining conditions, the value of tair
is that of the ambient wet bulb temperature. In some
cases it may be more appropriate to use tair = dry bulb
12 18 24 30 36 42 48 or tair = average of wet and dry bulb temperatures. In
Airway Age months
any event, even with the most sophisticated climate
simulation software, the degree of airway wetness is
an input parameter that has to be obtained from site
specific observation and tuning of models.
A further analysis of generic heat transfer relation-
ships for a given mine will also provide an indication
of the importance of air velocity (Fig. 7).
Again, remembering that the objective of VR venti-
lation simulation is not to employ network simulation
or detailed climate simulation, the issue to note is that
the range of heat transfer (W per m per ◦ C (tvr – tair ))
is relatively insensitive to air velocity within the range
of that employed in working places. For a given differ-
ence between air and strata temperature, the resultant
wet bulb gradient is therefore dependent on the rate of
heat transfer and the quantity of air in the airway. For
the example given in Figure 7, the wet bulb gradient
is about 1.0◦ C per km at 2.0 m/s and about 0.5◦ C per
km at 4.0 m/s i.e. the velocity determines the quantity
of air present rather than significantly altering the heat
transfer rate.
3.4 Equipment loads
Although the efficiency of diesel equipment does
change with engine load, engine type and useful
thermodynamic work done, using an efficiency of
30% provides a reasonable approximation of potential
heat output with analysis of fuel consumption required
to quantify actual heat load. For example, a 250 kW
rated loader will, at 60% engine load, emit about
250/0.3 × 0.6 = 500 kW of heat, which, if ventilated

289

Copyright © 2006 Taylor & Francis Group plc, London, UK

 110kW Jumbo 270kW Loader Table 3. Example diesel fleet analysis.
12
@ 60% load 10m3/s @ 60% load
Total
10 rated Trucks & Other
Wet Bulb Increase °C

Number power loaders vehicles

8 3
25m /s Application of units (kW) (kW) (kW)
6
65m3/s Development 24 3169 1482 1687
4 Production 27 5668 5198 470
Drilling 14 848 515 333
2 Electrical 3 306 306
Logistics 9 840 840
0 Total 77 10,831 7195 3636
0 200 400 600 800 1000 1200 1400
Heat Load kW

Figure 8. Potential wet bulb increase due to diesel units.
by 0.05 m3 /s per rated kW or 12.5 m3 /s would result
in a circa 7◦ C increase in wet bulb temperature. The
maximum potential heat output for this unit would be
about 800 kW resulting in a 15◦ C WB increase in wet
bulb temperature if ventilated to the same standard.
This issue is shown graphically for various equipment
loads in Figure 8.
The temperature of development faces under aux-
iliary ventilation are the most sensitive to simulation
tolerances reflecting the fact that high strata and equip-
ment heat loads are superimposed on relatively low
ventilation rates.
In an existing Australian metalliferous mine, recent
heat load surveys have shown an increase of 9 to
13◦ C WB where diesel equipment (235 kW loader
and 475 kW dump truck) is used under auxiliary
ventilation conditions even if ventilated to statutory
requirements of 0.05 m3 /s per rated kW. In these
circumstances ventilation requirements for the man-
agement of heat far exceeds that for dilution of gaseous
or particulate contamination that may arise.
It is also the case that, superimposed on intake air
temperature determined by surface climate and depth,
the effect of diesel equipment on work place tempera-
tures is both profound and variable (Howes 1988). For
this reason, and regardless of the climate simulation
software employed, it is impossible to predict abso-
lute face temperatures and an “operating envelope”
approach is required.
As an example of overall mine heat loads, the
diesel fleet employed in the mine section described
in Table 1 comprised 77 units with a total rated power
of 10.8 MW, Table 3.
Based on 850 m3 /s (1105 kg/s at 1200 m) the over-
all ratio of air to rated diesel power was 0.078 m3 /s
(0.10 kg/s) per kW. The observed specific diesel fuel
consumption was 1.5 litre/tonne equivalent to a heat
load of circa 5.5 MW averaged over the year.
The input parameter to the model is equipment heat
load (electrical or diesel) as a point, kW, or linear, kW
per km, source for active elements with actual wet bulb
temperatures calculated from the assigned air volume.
At any point in active or inactive elements, negative
heat sources can be assigned to represent refrigeration
plant.
3.5 Design face temperatures and cooling
capacity
Regardless of the heat stress index employed at a mine
design limiting face temperatures will typically be 27.0
to 29◦ C WB for persons working outside air condi-
tioned cabs and not more than 32.0◦ C WB (Australia)
or 32.5◦ C WB (Republic of South Africa) in any event.
For example, at 0.5 m/s and 5◦ C wet bulb depression
the limiting wet bulb temperature for an air cooling
power of 140 W/m2 is about 30.5◦ C WB.
The VR ventilation simulation model represents air
temperatures by colour gradients, allowing ranges to
be identified. The cooling capacity of underground
ventilation (kW) is then calculated from the mass flow
rate and difference between the sigma heat at actual
and design air temperatures. Overall, in warm con-
ditions (>25◦ C WB) cooling capacity with increas-
ing depth amounts to 410 to 350 kW per 100 kg/s
per ◦ C WB.
This, for example, allows a comparison to be made
between management of heat with increased vol-
umetric capacity, when there is sufficient cooling
capacity at depth of mining, and the introduction of
refrigeration.
4 PRIMARY VENTILATION CIRCUIT
CAPACITY
The analysis of volumetric requirements can of course
be undertaken without network simulation or visuali-
sation of mine geometry. However, it is the provision
of primary circuit intake and exhaust capacity that
requires location dependent ventilation requirements
to be identified i.e. at any point in time how much

290

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 4. Reduction of intake air volume with depth. Surface WB°C (Figure 2)
Strata temp tvr °C (Figure 5)
x,y,z
Depth Quantity
(m) (m3 /s) 3
Air quantity m /s cor-
Shaft Q m3/s
(Table 2,5) rected for depth mass
0 100 flow rate analysis
Depth Zm (Table 4)
1000 92
2000 85 Plus auto com-
3000 79 pression kJ/kg
(Figure 3)
Plus strata new and
Plus strata (old) and face heat sources, kW
heat sources, kW (Figures 6,7,8)
(Figures 6,7,8)
Table 5. Guideline design air velocities.
Face Q m3/s
Intake Q m3/s (Table 5)
Velocity x,y,z (Table 5) x,y,z x,y,z
Airway type range m/s
Shaft Intake Face
WB & DB˚C WB & DB˚C WB & DB˚C
Intake shaft velocities with hoisting <12 tvr°C tvr°C tvr°C
Intake ventilation shaft velocities <20
Intake airways for vehicles and pedestrians 4 to 6 Figure 9. VR ventilation simulation calculation process.
Large workshop excavations >0.25
Auxiliary ventilated headings & drill horizons >0.5
Loading horizons >0.75
Exhaust airways with pedestrian access <10 are increasing prior to appropriate intake and exhaust
Exhaust shaft velocities 16 to 22 capacity being installed. This is particularly true of the
development phase where volumetric requirements are
typically high but establishment of primary exhaust
circuits have not yet been completed.
ventilation is required and where in the mine. It is in
this aspect of mine design that visualisation of circuit
expansion in more complex mines is required. 5 VR VENTILATION SIMULATION
The VR ventilation simulation model can identify
what activities are active at any point in time and there- The calculation process for each element of the mine
fore summate volumetric requirements. The effect of model (Fig. 9) proceeds as follows:
variations in air density are taken into account with
a simple depth relationship calculated with respect to 1. Surface temperature is defined by climatic profile.
air density on surface. For example the relationship 2. Intake shaft temperatures determined by surface
for air leaving surface at 100 kPa 20◦ C WB 25◦ C DB temperature plus auto compression.
1.16 kg/m3 is shown in Table 4. 3. Level air quantity defined by design velocity e.g.
It would be equally valid to undertake this same 5 m/s.
analysis on a mass flow (kg/s) instead of volumetric 4. Air temperatures entering work places calculated
(m3 /s) basis. from shaft temperature plus heat transfer from
Prior to network simulation being undertaken, vol- strata (old) and mine equipment heat load in
umetric ventilation requirements will determine the intakes. Calculated at design intake air velocity, air
minimum cross sectional area of intake and exhaust temperature and virgin strata temperatures.
shafts together with the number and size of airways in 5. Air temperatures in work places (active elements)
the primary intake and return ventilation circuit. These are calculated from intake airway temperature plus
design parameters are determined again by “rule of heat transfer from strata (new) and defined face
thumb” or guideline design air velocities (Table 5), equipment heat load.
which can be modified for site specific factors such as 6. Air required in parts of the mine or mine as a whole
the financial structure of the project, cost of devel- given by the sum of activities and allocation to fixed
opment or ground stability dictating limiting shaft locations.
dimensions.
With respect to mine geometry, this means that not Quite simply, the wet bulb temperature at any point
only are total volumetric requirements required to be in the mine is calculated from the cumulative heat
known with time but also the mean depth at which they (kJ/kg) added to the defined airflow rates. The wet bulb
must be delivered. Problems may arise when, for exam- temperature is calculated from curve fitting routines
ple, an existing mine is being extended in depth and with similar result to that described by McPherson
volumetric requirements of deeper mining horizons (1984).

291

Copyright © 2006 Taylor & Francis Group plc, London, UK

 An example calculation for three 200 m develop- A further advantage of undertaking analysis in this
ment faces, 600 m from an intake shaft at 1000, 2000 manner is that each of the individual sources of heat
and 3000 m below surface is shown in Figure 10. Sur- can be summated.
face intake temperatures are 15◦ C WB and 20◦ C DB, The overall model for a particular mine can then
strata heat transfer being calculated from Figure 6 and be examined in large scale 3D stereo (Fig. 11) with
150 kW allocated to the development faces. colours or graphical output used to identify various
ventilation or heat management issues at any particular
stage in the project (Fig. 12).
Surface
15˚C WB 20˚C
tvr C Depth m
600 m 200 m
20 0m 5mx5m 5mx5m

500 m “Cool”
19.6 / 27.6 20.1 / 24.4 23.2 / 26.0
35 1000 m
45 m3/s 15 m3/s 150 kW
1500 m “Limit”
24.6 / 30.8 27.9 / 32.1
23.9 / 34.3
50 2000 m
45 m3/s 15 m3/s 150 kW
2500 m “Hot”
28.1 / 40.3 29.0 / 36.8 32.2 / 37.9
65 3000 m
45 m3/s 15 m3/s 150 kW
Figure 11. Reviewing mine plans in the Virtual Reality
Figure 10. VR ventilation simulation calculation example. Laboratory.

Figure 12. VR ventilation simulation visualisation of a base metal project. Air temperature shown by colour and volumetric
requirements shown graphically.

292

Copyright © 2006 Taylor & Francis Group plc, London, UK

 6 CONCLUSION
Ventilation and heat management requirements are
critical aspects of underground mine design and can
constrain operational decisions or design choices if not
properly addressed in both short and long term mine
plans.
The VR ventilation simulation approach to venti-
lation and heat management design described in this
paper allows the following issues to be evaluated
without recourse to network simulation or detailed
knowledge of mine infrastructure;
1. Volumetric requirements with time.
2. Distribution of volumetric requirement within the
mine.
3. Number and or size of intake and exhaust airways
or shafts.
4. Number of main intake and exhaust airways.
5. Gross mine heat load with time.
6. Seasonal variations in mine air temperatures with
expansion of mine geometry.
7. Cooling capacity of the ventilation system.
8. Effect of increased volumetric capacity on mine air
temperatures.
9. Point at which refrigeration is required due to the
inevitable consequence of depth superimposed on
surface intake temperatures.
10. The effect of main intake refrigeration plant on
mine air temperatures.
Simulation of the network in greater detail and to
determine characteristics of surface of underground
booster fans is still required for implementation of
the project as is detailed climate simulation to assess
refrigeration strategies if and when required. As the
capabilities of VR ventilation simulation are contin-
uously evolving, these may be incorporated in future
models together with means of real time interaction
with operating mines.
REFERENCES
Bluhm, S., & von Glehn, F. 2004. Important basics of mine
ventilation and cooling planning. Journal MineVentilation
Society South Africa. 57(1): 15–24.
Burrows, J., Hemp, R., Holding, W. & Stroh, R. 1989.
Planning Ventilation And Refrigeration Requirements.
In J. Burrows (ed.), Environmental Engineering In
South African Mines: 953–974. ISBN 0 620 06258 4,
chapter 38.
293
Copyright © 2006 Taylor & Francis Group plc, London, UK
Dunn, P. & van Wageningen, A. 2003. The Development
of Mine Plan Audit Capability for the Mining Industry.
CIM Operators Conference (paper no. 36), Saskatoon,
Saskatchewan, October 19–22.
Fourie, G. 2004. An integrated approach to mine planning
and design. Journal Mine Ventilation Society South Africa.
57(4): 129–132.
Hall, G & Clarke, B. 1988. Ventilation design and practice at
the Olympic Dam, South Australia. In A.D.S Gillies (ed.),
Proceedings of the Fourth International Mine Ventilation
Congress, Brisbane, Australia, 3–6 July 1988: 575–584.
Melbourne: AusIMM.
Hemp, R. 1989. Thermodynamic aspects of mine airflow,
Psychrometry. In J. Burrows (ed.), Environmental Engin-
eering In South African Mines: 29–48 and 435–463. ISBN
0 620 06258 4, chapters 2 and 6.
Hemp, R. 1985. Air temperature increases in airways. Journal
Mine Ventilation Society South Africa. 38(1, 2).
Howes, M. 1988. Heat and moisture exchange in mine
airways. Proceedings 4th International Mine Ventilation
Congress: 257–264. Brisbane.
Howes, M.J. & Nixon, C.A. 1999. Development of proced-
ures for Safe Working In Hot Conditions. Proc. 6th
International Mine Ventilation Congress.
Howes, M. and Sockhaus, F. 2002. Onaping Depth Project –
Ventilation and cooling planning for mining 2600 m below
surface. In E. De Souza (ed.), Proceedings of the North
American 9th US Mine Ventilation Symposium, Kingston,
Ontario, Canada, 8–12 June 2002: 391–400. Rotterdam:
Balkema.
Lambrechts, J. 1985. Written contribution to Air Temperature
Increases in Airways. Journal Mine Ventilation Society
South Africa. August 1985: 94–95.
Marais, D. 1992. Development of a model for the prediction
of the overall heat load on a Southern African colliery. 5th
International Mine Ventilation Congress: 131 to 137.
McPherson, M. 1984, A direct relationship between sigma
heat and wet bulb temperature. Journal Mine Ventilation
Society South Africa. November 1984: 71–72.
McPherson, M. 1993. Subsurface Ventilation And Envir-
onmental Engineering. Chapman and Hall, ISBN
04122353008.
Moreby, R. 2002. Sources Of Heat And Heat Management
In Australian Longwall Coal Mines. In E. De Souza (ed.),
Proceedings of the North American 9th US Mine Ventila-
tion Symposium, Kingston, Ontario, Canada, 8–12 June
2002: 363–370. Rotterdam: Balkema.
Rawlins, C. 2004. Underground mine heat loads and asso-
ciated reduction methodologies. Journal Mine Ventilation
Society South Africa. 57(1): 25–30.
van der Walt, J. 1986. Predicting the nominal heat load in a
mine. Journal Mine Ventilation Society South Africa. May
1986: 61–74.
Wallace, K.G. 2003. General Operation Characteristics and
Industry Practices Of Mine Ventilation Systems. 7th Inter-
national MineVentilation Congress, Krakow, Poland, June
2001. Available from www.mvsengineering.com.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Alternative fog removal methods in mine ramps

A.L. Martikainen
Helsinki University of Technology, Espoo, Finland

ABSTRACT: Fog forms in mine ramps when warm and humid exhaust air moves upwards the ramp and temper-
ature decreases. As fogging is a fairly common problem, many different fog removal alternatives are known. For
example, heating, cooling, scrubbing, chemically drying the air, and increasing air velocity are used to remove fog.
In this study fogging problems of some Finnish mines are reviewed. Furthermore, possible fog removal
methods are compared to identify most feasible and sensible solutions. The most common fog removal methods
in subarctic countries are air heating and increasing air velocity, so the main interest is directed for comparison
of these methods. In cases where there have already been attempts for fog removal, user experiences of the used
fog removal methods are presented.
To enable comparison of the methods tests were conducted and measurements taken. Results of the comparison
indicate that increasing air velocity by air rerouting is not a satisfactory fog removal method. Installing an
additional fan to increase the air velocity yielded better results, but calculations showed that the resulting
changes in temperature and humidity are directly proportional with the heating capacity of the fan. Best fog
removal results were achieved with air heaters.

1 INTRODUCTION roads, fogging comprise a potential for a safety hazard.

As fogging is often a seasonal problem, it is considered
in many mines more of a nuisance than a safety haz-
ard. Only when visibility is decreased to extremely low
or fogging becomes a continuous problem, corrective
measures are taken. Even then it is not easy to decide,
how to deal with the problem, as every alternative
approach has its drawbacks.
In Finland as well as in other subarctic or arctic
countries another thing to consider is the cool climate
that prohibits use of fog removal methods designed
for warmer climates. The drastic seasonal changes of
weather conditions and thus fogging should be taken
into account. Also the occurrence of fog mainly or
only in the decline of the mine is typical for countries
with an aforementioned climate. This restricts also fog
removal method selection.
Earlier studies have indicated problems in fog
removal by increasing air velocity. Now this issue is
addressed in a greater detail. More thorough field
research with measurements is completed concern-
ing both air velocity increasing methods and heating.
Obtained results are then presented and analysed.
2 FOGGING AND REMOVAL ALTERNATIVES
2.1 Fogging problems in Finnish mines
Fogging occurs mainly in the declines of the mines
in Finland. As the declines are often used as haulage
In ventilation point of view the declines are used mai-
nly as secondary exhausts. As the exhaust air typically
carries high particle concentration, gaseous pollutants
and plenty of water, it is not easy to avoid fogging.
Weather influences fogging especially in shallow
underground mines. If the weather is humid or the
outside temperature changes rapidly, fogging is more
probable. Thus thickest fogs are observed during
spring and autumn in most mines in Finland.
Choosing a fog removal method in subarctic coun-
tries is difficult, as the fogging problem is usually
seasonal, and thus fog removal is only required part-
time. Also methods developed for warm climates are
seldom practical. Another thing to consider is the indi-
vidual character of fogging in a mine. In some mines
fogging occurs because of the high humidity of air,
in some other mines the main influencing factor is the
high particle concentration. Thus the same fog removal
method may not be the most effective in every mine
even in a region with similar weather conditions.
2.2 Comparison of fog removal alternatives
Fog removal alternatives used worldwide are heating,
cooling, scrubbing, increasing air velocity and chemi-
cal drying of air. Sometimes two or even more of these
methods are used in a combination. Usually best fog
removal results are achieved by these combinations.
The fog removing effect of increasing air velocity
is said to be based on air mass mixing and promoting

295

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 1. Measurement results with full operation.

Depth Particles Air velocity Temperature Relative humidity Dew point Fogginess
(m) (mg/m3 ) (m/s) (◦ C) (%) (◦ C) (estimate)

230 0.314 0.95 9.1 84.4 6.6 No

222 0.319 1.1 8.7 87.2 6.7 No
206 0.505 1.2 9 86.7 7.2 Light fog
175 0.176 1.3 8 89.8 6.5 Fog
155 0.197 1.2 8 90 6.5 Fog
130 0.273 1.7 8.7 89 7 Light fog
110 0.296 3.1 8.9 88.7 7.2 No
70 0.366 1.5 8.9 88.6 7.2 Light fog
40 0.427 1.1 8 90.8 6.7 Fog

evaporation from fog droplets. In addition, an increase
in airflow velocity creates a psychrometrically uni-
form air mass and diminishes the number of potential
condensation points. Economically this method is sug-
gested to be the most plausible solution. (Tien 1999)
Increasing air velocity is currently one of the most
popular ways for fog removal, but its efficiency is
questioned. (Martikainen 2005, Calizaya et al. 2001)
Centrifugal scrubbers work well in level workings
according to literature, but they are not necessarily
suitable for clearing declines. Their effectiveness in
fog removal is based on their multiple function char-
acteristics. They increase air velocity, reduce particle
concentration as well as humidity.
Heating is often criticized because of the added heat
to the mine environment, but in cool climates with
cool or temperate underground temperatures increas-
ing temperature does not cause problems. Heating is
unfortunately quite expensive as it consumes a lot of
energy. It does not reduce the amount of particles,
either.
Cooling by refrigeration is the most expensive fog
removal method and in a cool climate it may also cause
difficulties as the temperatures are already low and
freezing is definitely not recommended.
Chemical drying is considered impractical in min-
ing industry because the amount of drying material
needed is huge. Spreading the material and gathering
it is also extremely troublesome.
Removing humidity by preventing leakages is not
mentioned in literature as a fog removal method, but
can also be used as such to some extent. Unfortunately
treating a long ramp is costly. As there are also open
ditches and pumping stations along the decline, pre-
venting leakages may not be enough for fog removal.
3 FOG REMOVAL ATTEMPTS AND USER
EXPERIENCES
3.1 Fog removal trials and errors in Finland
In most Finnish underground mines fog removal by
means of rerouting air in order to obtain higher air
velocity in the decline has been tried out in some
decree. The results vary, but it can be concluded, that
fogging problems have not disappeared. In some mines
the fog has moved further up along the ramp or long
fog fronts have been split to shorter ones. Fog removal
has not been accomplished.
In a paper by (Martikainen 2005) fogging problems
of three Finnish mines have been analyzed and results
obtained from increasing air velocity by rerouting were
presented. As high velocity as 2.5 m/s was not able
to scatter fog, even if the literature value suggested
enough to greatly reduce the widespread occurrence
of the underground fog is only one tenth of this.
Best air rerouting results were obtained acciden-
tally at the Pyhäsalmi mine as most of the humid air
escaped from the decline to the old workings and thus
took the fog away from the ramp entirely. Unfortu-
nately the air velocity in the decline dropped close to
zero, which was not appreciated. The air reentered the
decline some hundred meters above with lower humid-
ity. It was concluded that either the filling of the old
workings was adsorbent or water condensed on the
cool surfaces of the filling material and rock surfaces
while working its way through.
Because increasing air velocity is one of the most
popular fog removal methods, further studies were
conducted. An additional fan was installed to the
decline-level crossroad of the Louhi mine on level
+130 m. Measurements covered both operating mode
and the situation in the ramp with the fan turned off.
Measurements for comparison purposes were taken
with full operation and after the shift. Results are pre-
sented in Tables 1 and 2. After the shift the main fan
operated only with 80% capacity and the level work-
ing fans were turned off. This can be seen from the
velocity values.
The fan installation and the fan itself are shown in
Figures 1 and 2. The idea was to take fresh air from
the level and direct it through the ventilation wall to
the decline with a duct. Airflow direction is presented
with arrows in Figure 2.
Analysis of results showed most noticeable change
in the depth of 110 m. The temperature rise was only

296

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 2. Measurement results after the shift.

Air Relative Dew

Depth Particles velocity Temperature humidity point Fogginess
(m) (mg/m3 ) (m/s) (◦ C) (%) (◦ C) (estimate)

230 0.314 0.5 8.7 86.7 6.8 No

222 0.319 0.8 8.4 88.1 6.7 No
206 0.505 1 8.2 89.3 6.4 light Fog
175 0.176 0.6 8.3 88.1 6.5 ligth Fog
155 0.197 0.9 8.4 88.6 6.5 light Fog
130 0.273 1.8 8.3 88.5 6.5 light Fog
110 0.296 1.7 8.5 88.1 6.8 light Fog
70 0.366 1.6 8.4 88.4 6.7 light Fog
40 0.427 1.2 8 89.6 6.5 fog

Figure 1. An additional fan installed on level +130 m.

◦
0.3 C, but humidity decrease was more pronounced
3.4%. The increase in air velocity was somewhat less
than 1 m/s.
As the air was taken from a level, also the charac-
teristics of the air there were considered important and
thus measured. Furthermore, the effect of the fan to the
airflow was studied. The results showed that in passing
the fan the airflow dried considerably and was heated.
The temperature increased by 1.3◦ C and the decrease
of relative humidity was as high as 5%. Results are
presented in Table 3.
Calculations showed that the changes in humidity
and temperature with the fan on corresponded exactly
with the heating capacity of the fan. The calculated
temperature of the united airflows of 14 m3 /s from the
fan and of 46.8 m3 /s from lower levels on level +130 m
was 8.376◦ C when the measured value was 8.4◦ C. The
correspondence was similar for the humidity.
Also air heating in combination with the fan was
studied. The installation of the heaters is shown in Fig-
ure 2. Air heating was already evaluated to be suitable
for fog removal of this mine earlier. (Martikainen &
Särkkä 2004) The heaters were not tested without fans
Figure 2. Fan and heater installation on level +130 m.
on during operation, because satisfactory results were
received with the fan and turning it off would have
had a negative effect on ore transport. Heaters were
installed close to the fan. The overall heating capacity
of these heaters was 94 kW, which was slightly lower
than anticipated from the calculations, which are pre-
sented later. With heating a temperature rise of 2◦ C
was sought after.
On the installation location measured increase in
temperature was 1.4◦ C. Unfortunately in the depth
of 110 it was already decreased to only 0.6◦ C. A
slight increase in temperature was observed all over
the decline. This was not, however, enough for com-
plete fog removal. On the other hand the effect of
heating on the measured relative humidity values was
drastic. On the crossroad of +130 m the decrease of
relative humidity was over 7%. The smallest difference
in relative humidity was measured near the surface, but
even there relative humidity decreased over 1%. These
measurement results are shown in Table 4.
The reason for failure to remove fog completely was
that the calculations for the heater dimensioning were

297

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 3. The effect of fan to the airflow.

Air Relative Dew

Depth Particles velocity Temperature humidity point Fogginess
(m) (mg/m3 ) (m/s) (◦ C) (%) (◦ C) (estimate)

Fan on
110 0.243 2.8 8.8 87.6 6.7 No
130 0.197 1.2 8.4 89.8 7 No
Fan off
110 0.336 2 8.5 91 7.3 Fog
130 0.281 2.2 8.1 90.7 6.6 Fog
Duct
Duct end 9.3 82.7 6.5 No
Fan inlet 0.039 1.7 8 87.7 6.1 No

Table 4. Measurement results with heaters on and off.

Air Relative Dew

Depth Particles velocity Temperature humidity point Fogginess
(m) (mg/m3 ) (m/s) (◦ C) (%) (◦ C) (estimate)

Heaters on
230 0.678 0.9 9.3 82.9 6.5 No
222 0.69 1 9.2 83.7 6.6 No
206 0.667 1 9 86 6.5 No
185 0.509 1.4 8.6 88.4 6.6 Light fog
130 0.338 1.4 9.6 84.9 6.8 No
110 0.299 2.3 9 87.2 6.9 No
70 0.244 2.3 8.7 89.1 6.7 No
40 0.336 1.2 8 91.6 6.8 Light fog
Heaters off
230 0.577 0.8 9 87.8 7.2 No
222 0.678 1 8.3 90.6 7 No
206 0.543 0.7 8.3 91.2 7.1 Light fog
185 0.592 1.4 8 91.8 6.9 Light fog
130 0.413 1.7 8.2 92 7.1 No
110 0.334 2.2 8.4 89.3 6.9 Light fog
70 0.211 2.4 8.4 93 7.4 Fog
40 0.265 1.3 7.8 92.7 6.9 Fog

very basic and did not take into account all factors that The other calculation way is presented as Equa-
affect the temperature. Two sets of calculations were tion 3.
done and both of these gave similar results. The first
set of formulas was meant for interior climate design
of houses. It is presented with Equations 1 and 2:

In this equation BTU is British thermal unit, 0.24 is

where φ represents the required heater capacity in kW,
ϕ represents air density, qv airflow and h1 and h2 air
enthalpies. Enthalpies are calculated by:
in which h is the enthalpy in kJ/kg, t is the temperature
of humid air and x is the absolute humidity of air. (Sep-
pänen 1996) With these formulas obtained required
heater capacity was 115 kW.
the specific heat of air, 0.0746 is the weight of one
cubic foot of air, CFH is the volume of air to be heated
and T is the number of degrees rise desired. (Kennedy
1996) After converting all units to SI units, a value of
106 kW was obtained.
For example leakages, rock thermal conductiv-
ity and virgin rock temperature were not taken into
account at all in any of the formulas. As all these effec-
tively decrease the temperature, it is easy to understand
why the heater capacity was not sufficient.

298

Copyright © 2006 Taylor & Francis Group plc, London, UK

 The worst leakages were observed between the mea-
surement points of 40 m and 70 m, where the smallest
decrease of relative humidity was measured. Also
other measurement results correspond well with the
observations.
If dimensioning for complete fog removal was
made, more complicated analysis of the relevant
parameters is required. It would probably be best to
use a computer program designed for this purpose. As
an example a mine climate simulation program CLIM-
SIM can be mentioned. The program distributor, Mine
Ventilation Services, states that this program takes into
account geothermal gradient, rock thermal conduc-
tivity and diffusivity, airflow, air quality, age of the
excavation, wetness of the rock surfaces and the siting
and capacity of machinery, heat exchangers or other
local or disseminated sources of heat and humidity.
(Mine Ventilation Services 2003)
3.2 User experiences
Increasing air velocity by air rerouting has not been
able to solve fogging problems anywhere. Informa-
tion concerning changed fog movement patterns and
fogging locations have been obtained from personnel,
but actual improvements have not been observed. So
it can be concluded that increasing air velocity by air
rerouting is not a sensible fog removal method.
Increasing air velocity by installing a booster fan at
the Louhi mine had a positive effect on the situation.
The personnel of the mine were pleased as the fog
got lighter near the fan. Air velocity increased about
1 m/s upstream of the fan and humidity values changed
locally.
Installing heaters also gave positive results. Foggi-
ness decreased noticeably upstream from the instal-
lation. Measurement results showed conditions less
favourable for fogging also downstream, but unfortu-
nately the personnel did not respond accordingly. The
feedback contained comments about better visibility
in the upper part of the decline, but also complaints
about decreased visibility below the heater installation
location.
It was obvious that the capacity of the heater instal-
lation was not enough to treat the long decline. Even
if the temperature was higher in every measurement
point, it was still not high enough to prevent fogging
completely.
The reason for the feedback was not obvious,
so it was questioned whether it originated from the
improved conditions of the upper section of the decline
or changes below the heaters. When the visibility
upstream the heaters increased noticeably, could it be
possible that the almost unchanged situation down-
wards of the heaters was considered worse? As the eye
adjusts to the better visibility light fog that was not
even observed earlier after driving through a heavily
299
Copyright © 2006 Taylor & Francis Group plc, London, UK
fogged part surprises the eye and seems worse than
before.
Leakage prevention used at the Pyhäsalmi mine pre-
vents fog formation, but the situation was already quite
good in the treated part of the incline before these mea-
sures were taken. As the temperature there was well
above the dew point the worst problems were likely
to occur in the upper parts of the ramp anyway. In
there the fog situation did not change. (Pulkkinen &
Martikainen 2004) Personnel are almost completely
working in the deep section of the mine where fogging
does not occur, so they are quite happy with the results.
In the upper part of the decline, from about depth of
1050 m upwards, however, fogging is still a nuisance.
4 CONCLUSIONS
Increasing air velocity by rerouting is highly unlikely
to solve fogging problems.This method has been tested
in many mines with similar results. The only observed
changes were location changes of the fog fronts as well
as some fog movement variations even if the achieved
air velocities reached 2 m/s and above. No fog removal
actualized.
Air rerouting may give positive results if foggy air
is rerouted through absorbent material or through tun-
nels with cold walls in which water condenses. As
the humidity of the air decreases, probability of fog-
ging also decreases. This method is, however, not well
tested.
Increasing air velocity in one of the mines with
an additional fan installation gave slightly different
results when compared to increasing air velocity by
rerouting. Local fog removal was observed. Calcu-
lations showed that the measured temperature and
humidity decreases were directly proportional to the
heating capacity of the fan. Thus it can be concluded
that the fog removal effect of a fan is due to its effect
as a heater.
Altogether it can be concluded that increasing air
velocity does not give acceptable results in removing
fog from a mine. If fans are used for fog removal, their
effect is not based on the air velocity increase, but
rather on their heating capacity.
Heater installation gave fair results. It decreased rel-
ative humidity and increased temperatures all around
the decline. This was by far the most satisfactory of
the tested methods.
Unfortunately the dimensioning of the heaters was
not enough for complete fog removal in the test
case. Formulas used for the capacity dimensioning
were straightforward and did not take into account
every parameter that affects temperature in a mine
climate. More refined numerical modelling is sug-
gested whenever a detailed quantitative understanding
is required.
 ACKNOWLEDGEMENTS
Feedback from the mine personnel of Louhi, Orivesi
and Pyhäsalmi mine is appreciated. Also other sup-
port received from the personnel during these studies
was essential for obtaining useful and reliable results.
Special thanks go to the visit hosts. Financial and tech-
nical support of the Helsinki University of Technology
is acknowledged.
REFERENCES
Calizaya, F., Karmawan, K. & Wallace, K.G. 2001. Utilization
of heater fans to control mine atmospheric fogging. In
Proceedings of 9th US/North American Mine Ventilation
Symposium.
Kennedy, W.R. 1996. Practical Mine Ventilation. Chicago:
Intertec Publishing.
300
Copyright © 2006 Taylor & Francis Group plc, London, UK
Martikainen, A.L. & Särkkä, P.S. 2004. Revision of the venti-
lation system in Louhi mine – a case study. In Proceedings
of 10th US/North American Mine Ventilation Symposium:
503–508. London: AA Balkema Publishers.
Martikainen, A.L. 2005. Comparative evaluation of fogging
phenomenon in the ramp of three mines in Finland. In Pro-
ceedings of 8th International Mine Ventilation Congress:
103–110. Burwood: BPA Digital.
MineVentilation Services. 2003. http://www.mvsengineering.
com/climsim.htm.
Pulkkinen, M.S. & Martikainen A.L. 2004. Ventilation sys-
tem of Pyhäsalmi mine, Finland – a case study. In
Proceedings of 10th US/North American Mine Ventilation
Symposium: 67–72. London: AA Balkema Publishers.
Seppänen, O. 1996. Air conditioning and interior climate.
Published in Finnish as: Ilmastointitekniikka ja sisäil-
masto. Helsinki: Kirjapaino Kiitorata Oy.
Tien, J.C. 1999. Practical Mine Ventilation Engineering.
Chicago: Intertec Publishing.
 5. Mine fires

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Reducing belt entry fires in underground coal mines

W.J. Francart
US Department of Labor, USA

ABSTRACT: The use of belt air to ventilate working sections coupled with improved fire detection capability
have been the major reasons for mine operators installing carbon monoxide detection systems in belt conveyor
entries of underground coal mines in the United States. The advent of carbon monoxide monitoring systems
may have contributed to reducing the frequency of reportable belt fires, but prevention of the conditions that
allow fires to occur is more effective than detecting and extinguishing fires after ignition occurs. An analysis of
reported mine fires in belt entries shows many belt fires could have been prevented with proper maintenance
of the haulage system. This paper reviews the reported belt fires in underground coal mines for the period from
1980 to 2005. Also included in this paper is an analysis of ignition sources as identified in accident reports issued
by MSHA investigators, as well as the factors involved in creating ignition sources.

1 INTRODUCTION A delay in evacuating miners contributed to five peo-

The use of belt air to ventilate working sections has
been a major reason for mine operators installing
carbon monoxide (CO) monitoring systems in belt
conveyor entries of underground coal mines in the
United States. Until the publication of regulations
regarding the use of belt air in 2003, mine operators
were permitted to use belt air only through a petition
for modification of federal regulations, all of which
required the use of CO monitoring systems. In many
other mines, CO systems have been installed in lieu of
typical point-type heat sensors (PTHS) because of the
increased detection capability of the systems. Many
mine operators have determined the additional cost of
maintenance and calibration of CO monitoring to be
an invaluable investment for improving fire protection
for miners and mine property.
The majority of coal mine disasters have been due
to explosions of methane and coal dust. Mine fires
have been the cause of fewer disasters. However, the
Wilberg mine fire in 1984 took the lives of 27 min-
ers (Huntley 1986). Since 1980, only one other death
has occurred during a belt conveyor fire. However, as
miners escaping from the Marianna mine belt fire will
relate, there is a real danger and potential for disaster
when a fire occurs in a coal mine belt entry.
The Marianna belt entry fire (Strahin 1988) most
likely resulted from frictional ignition of coal and coal
dust at a belt drive which was not properly maintained.
Compounding the dangers of the fire was the deluge
fire suppression system at the drive being disabled,
and unavailability of a water supply to fight the fire.
ple being treated for smoke inhalation. Even though
the mine was using a CO monitoring system as well as
a PTHS system for fire detection, the fire apparently
developed so quickly that miners were still in jeopardy.
The accident points out that prevention of the condi-
tions that cause fires is more effective than detecting
and extinguishing flames.
2 HAZARDS OF BELT FIRES
Since 1980, belt fires have accounted for 15–20 per-
cent of all fires reported to MSHA by coal mine
operators. By definition, a mine fire is reportable to
MSHA if it is not extinguished within 30 minutes
of discovery. After the mine operator reports a fire,
MSHA accident investigators will determine a possi-
ble cause of the fire. An investigative report, including
a description of the accident, details of the detec-
tion method, extent of property damage, and method
used to extinguish the flames is generated by MSHA.
Depending on the extent of the fire, this report can con-
sist of a two-page standard form (MSHA 7000-50a) up
to a multiple-page report complete with Root-Cause
Analysis (Paradies 2005). For this paper, accident
reports on 62 belt entry fires for a 26-year period (1980
to 2005) were reviewed to identify and tabulate fire
ignition sources, and to determine if maintenance of
the belt conveyor was a factor in allowing the fire to
occur. One fire was still under investigation and could
not be included in this analysis.

303

Copyright © 2006 Taylor & Francis Group plc, London, UK

 The belt conveyor is the lifeline of the modern coal
mine. If the belt stops, the flow of coal stops and mining
ceases. Depending on the location of a fire within the
belt, coal production can be disrupted for days, weeks,
and historically has even been the death knell to a coal
mine. In addition to the loss of production, fires can
destroy mine infrastructure often requiring extensive
rehabilitation to resume production. Besides replac-
ing belting, structure and drives, mine roof conditions
must be restored and in extreme cases, development of
a new belt slope bottom has been necessary to restore
coal haulage to the surface. Costs for containing, extin-
guishing and recovering a mine fire can run into the
tens of millions of dollars.
The bad news can be rather grim, but there is
encouraging good news. New fire detection system
technology is evolving, allowing mine operators to
detect fires before they cause extensive damage. A
small fire is easier to control and extinguish than a
large fire. In addition, many belt fires are preventable,
as suggested by the findings of accident investigation
reports. Proper maintenance of the belt and belt com-
ponents may have prevented many of the sixty-one
reported fires included in this analysis.
3 FREQUENCY OF REPORTED FIRES
Figure 1 shows the number of reported belt entry fires
by calendar year from 1980 to 2005. The general trend
is downward for the period, with the exception of four
reported fires in 2002 and three in 2003. Except for
the period from 2002–2003, the belt entry fire rate per
1000 mines is generally one to three fires per year,
with an average of 1.8. In 2002–2003, however, the
rate grew to nearly twice the typical rate and more than
three times the average rate, to 5.7 per 1000 mines.
One company experienced two major fires in 2003
causing considerable interruption to production and
damage to underground facilities. MSHA investiga-
tors attributed both fires to frictional heating caused by
the belt contacting belt support structure. As a result
of these fires and potential for additional fires, the
company implemented a program to improve protec-
tion against major underground mine fires. Included
in the program were increased emphasis on training
of mine examiners to identify potential fire ignition
sources, and emphasis of maintaining and testing fire
suppression systems (Tolliver 2004).
4 IGNITION SOURCES
The majority of mine fires attributed to mining equip-
ment ignition sources are due to equipment misuse,
lack of maintenance, compromising safety controls
and running equipment for long periods of time
304
Copyright © 2006 Taylor & Francis Group plc, London, UK
unattended (McPherson 1993). In the case of conveyor
belts, these factors would include inadequate design,
spillage and coal accumulations, failure to lubricate
moving parts, failure to replace bad rollers, malfunc-
tioning slippage switches at belt drives, inadequate
examinations for hazards in the belt entry, misalign-
ment of the belt, and improper installation. In most of
these cases, frictional heating becomes the likely igni-
tion source. Careless welding and cutting practices are
other identified ignition sources (Mitchell 1996).
The belt drives and conveyor transfer points are con-
sidered by many to be primary ignition point sources
for a number of reasons. Electrical sources are more
likely to exist at drives, with drive motors and cables
located nearby, and transfer of coal can cause spillage
at these locations which, if not removed, can eventually
build up to a point where frictional contact between
coal and moving conveyor components occurs. MSHA
acknowledged the increased potential in these areas
when promulgating the 2003 regulations on the use of
belt air on working sections by requiring fire detection
sensors to be installed downwind of all belt drives.
The most common ignition source identified in the
accident reports is frictional heating. Other sources
included electrical equipment and conductors, weld-
ing and cutting operations, diesel equipment and
hydraulics. Table 1 shows the breakdown of the igni-
tion sources for reported belt entry fires for the period
from 1980–2005. The ignition source could not be
identified in 17 fire investigations due to extensive
damage to the belt and belt entry. In most accident
reports, the most likely ignition source was determined
by investigators based on interviews of miners and
conditions discovered underground. In 34 of the acci-
dent reports (56 percent), inadequate maintenance was
identified as a contributing factor in the occurrence of
the fire.
It is interesting to note that the number of frictional
ignitions along the belt was equal to the number iden-
tified as originating at belt drives. While it is probable
that many of the indeterminable sources were also at
or near drives, the numbers indicate ignition along the
belt is a significant concern. These fires have been
attributed to friction between the moving rubber belt
and wooden posts, friction between the belt and dam-
age belt idlers (rollers) and friction between the belt
and support structure.
5 FIRE DETECTION
The method of fire detection in belt entries is required
by regulation to include sensors to measure the rise in
temperature, or other types of sensors which provide
equivalent detection capability. Commonly known as
Point-Type Heat Sensors (PTHS), bi-metallic sensors
measuring rise in temperature, were state-of-the-art
 Figure 1. Number of reported belt fires by year, 1980–2005.

Figure 2. Number of active mines, number of belt entry fires reported, and fire rate per 1000 mine.

when hearings on the Federal Coal Mine Health and
Safety Act of 1969 were held. These sensors are
designed to activate when a predetermined tempera-
ture is reached, commonly 57◦ C. Because of the low
ambient temperature in a coal mine, the size of fire
needed to create such an increase in temperature must
be rather large. The fire’s heat intensity required for
detection has been compared to the number of fur-
naces operating in an 8 room home (Mitchell 1996).
For a nominal flow of 9.5 m3 /sec, the heat of more

305

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 1. Ignition sources of belt entry fires 1980–2005.
Ignition source Number of fires Percent of total
Cutting and welding 5 8%
Friction at belt drive 11 18%
Friction along belt 11 18%
Hot rollers 6 10%
Electrical 8 13%
Diesel and hydraulic 2 3%
Indeterminable 18 30%
than 10 furnaces would be required to cause an alarm
at a sensor only 15 meters downwind of the fire. This
fire would be difficult-to-impossible to control at this
size. The relationship is supported by a 1988 study
by MSHA on belt fires, where only three of 29 fires
were successfully detected by PTHS systems (Stephan
1989).
Current technology used for detecting fires in belt
entries includes PTHS systems as well as carbon
monoxide (CO) sensor systems. The CO systems are
considered superior to PTHS systems for coal mine
belt entries, as supported by an investigation report
of a fire in one mine where both PTHS and CO fire
detection systems were installed (Wilson 1996). As
evidenced in the report of this fire, the CO sensors
detected a fire early in development at a distance
of 460 meters from the fire source. Heat sensors
located within 9 meters of the fire did not activate at
any time.
Numerous anecdotal incidents have been known by
MSHA personnel over the years, showing CO sen-
sor systems are effective for providing early warning
for miners to escape hazardous situations, combating
developing fires and avoiding or reducing losses due
to fire damage. The implementation of these systems
has been steadily increasing since the first installation
in 1984, with over 20 percent of active mines cur-
rently using CO detection for compliance with federal
standards (Francart 2005). In many cases the systems
are not specifically required by MSHA, but rather are
installed by mine operators as an alternative to heat
sensor systems.
In promulgating regulations regarding the use of
belt air to ventilate working sections, an economic
analysis comparing the installation of CO-based detec-
tion with PTHS systems was completed. System pur-
chase and installation costs for the two systems are
equivalent at approximately 9,000 feet of belt entry
development. Beyond 9,000 feet, CO system purchase
and installation costs are less than PTHS systems.
Maintenance costs of the CO system are significantly
higher than PTHS systems due to calibration and repair
costs. However, the potential savings in rehabilitation
costs and lost production by detecting a fire in a timely
306
Copyright © 2006 Taylor & Francis Group plc, London, UK
manner can exceed the maintenance costs many times
over.
For the period from 1980 to 2005, approximately
25 percent of all belt fires occurred in mines using
belt air. Since the highest percentage of mines using
CO systems during the period was approximately 21
percent, it may seem that mines using belt air may have
a higher fire incidence rate. However, other factors
complicate this analysis, the most significant being
the size of the mine. Until the year 2000, most mines
using belt air were multiple section mines, often with
miles of belt conveyors. With a greater haulage infra-
structure comes a greater opportunity for fire sources
as the number of belt drives and length of conveyor
belt increases.
Recently smaller mines have begun to use belt air on
working sections. Because this ventilation method was
made available to all mines without need for a petition
for modification, this trend may strengthen. Federal
regulations promulgated in 2003 require the use of CO
monitoring systems in mines electing to use belt air
to ventilate working sections. Technological advances
are improving the detection capabilities of CO sys-
tems. The uses of diesel discriminating sensors and
hydrogen discriminating sensors have reduced signif-
icant interference of diesel exhaust and hydrogen gas
on CO sensors.
While extensive data on non-reportable fires is not
readily available, there is little doubt that the use of belt
air as a section intake has reduced the incidence and in
some cases the severity of belt fires. For an 18-month
period in 2002–2003, an MSHA survey indicated a
total of 32 non-reportable fires were detected by CO
monitoring systems (Francart 2005). Due to the size of
fire needed to activate heat sensors, these fires would
have likely grown into reportable fires if CO systems
had not provided early detection.
6 REDUCING BELT FIRE POTENTIAL
While ignition sources for 18 belt entry fires could not
be identified in MSHA accident investigations, after a
review of the accident reports it was determined that
many of the 17 could have been a result of frictional
heating at a belt drive. It is probable that one in four
fires in belt entries is related to frictional heat at a belt
drive.
Fires at belt drives are often due to accumulations
of coal which are ignited by heat produced by friction
between the belt fabric and drive roller due to belt slip-
page, or friction between the belt and structure or other
object such as a rock, post, crib or coal. Belt slippage
can be caused by a number of factors, including inade-
quate belt tensioning; increased drag on the belt due to
idler damage; spillage or roof fall; improper slippage
switch installation or operation; and decreased friction
 Figure 3. Slippage switch installation. Figure 4. Belt cutting through structure.

coefficient between the belt and roller lagging due to

improper or inadequate removal of contaminants or
worn lagging.
Current federal regulations for underground coal
mines require installation of slippage and sequence
switches on underground belt conveyors. Slippage
switches are often installed at belt drives to monitor the
belt and drive roller movement. When the drive roller
turns without belt movement, the switch is designed
to de-energize the drive and eliminate the slippage.
Most conveyor slip switches are designed to oper-
ate on one of three principles; magnetism, centrifu-
gal force, or photo-electricity. A common choice for
underground use is a system utilizing a slip switch
proximity sensor. These units work by detecting inter-
ruptions of a magnetic field by targets installed on a Figure 5. Damaged belt roller.
rotating shaft or roller. Slippage is detected by a change
in the speed of rotation of the shaft or roller. In addi-
tion, a new technology that uses a computer generated Other potential sources of heat for fire ignition are
signal to monitor motor current and voltage to detect belt misalignment and damaged belt rollers. Figure 4
a slow down in the belt is available (MSHA 2005). shows conditions in one belt entry where the belt was
Figure 3 shows the installation of a magnetic field slip- found to be cutting through the belt structure. The
page switch on a belt roller. Regardless of the type of frayed edge of the belt is visible on the right side of
detection system used, the unit should be tested fre- the vertical support. In some cases the belt has been
quently. MSHA recommends that systems be checked allowed to completely cut through the structure. These
at least once per shift. occurrences are often due to belt misalignment, and
Spillage along beltlines is sometimes attributed to have been identified as potential ignition sources in
misalignment of the belt (Swinderman 2002). Spillage accident investigation reports. One fire was ignited as
provides a ready fuel for a fire, and thus is not a result of belt misalignment, as the loaded top belt
permitted to accumulate in underground coal mines. ran out over the rollers contacting two wooden posts
Regulations require a program for regular cleanup and which had been ignited by the frictional heat.
removal of these accumulations. Frequent and repeated Another potential ignition source is damaged belt
spillage at particular locations has also been attributed rollers. Figure 5 shows a damaged roller which needs
to damaged belt structure and missing components. to be replaced because the roller bearings have been
Often when a belt roller is found to be bad, it is removed dislodged. Continued contact with the moving belt can
from the belt without being immediately replaced. cause wear and heat which can ignite coal dust, grease
Immediately replacing the roller would likely inter- or other combustible materials. In extreme cases, belt
rupt production, so rollers are removed to be replaced rollers have been found with holes worn in the contact
at a later time. surface due to friction with the moving belt.

307

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Electrical sources have accounted for approxi-
mately 13 percent of the ignition sources for belt entry
fires reported in the study period. Cutting and welding
along belt entries has been identified as the ignition
source for just under 10 percent of the reported belt
entry fires. These fires were often the result of care-
lessness on the part of the miner operating the torches.
Proper work practices and adequate examinations can
greatly reduce the potential for fire as a result of cutting
and welding.
7 CONCLUSIONS
The downward trend in the number of reported belt
fires has followed the trend in the number of active
underground coal mines for the past 26 years. The rate
of fires per 1000 active mines has been rather steady,
with the rate ranging from one to three. The excep-
tion is seen in 2002 and 2003, when the rate rose to
nearly six fires per 1000 mines. In only four of the
26 years were there no reported belt entry fires. While
many reportable fire sources could not be determined,
the most common ignition source reported is frictional
heating.
The use of belt air to ventilate the working section
has directly improved the fire detection capability in
these coal mines due to the required use of CO monitor-
ing systems. Technological advances have continued to
improve early-warning capabilities of these systems.
Intuitively, one could conclude that the use of belt
air has reduced the number of fires reported by mine
operators to MSHA.
However, mine operators cannot rely on the use of
CO monitoring systems alone for early fire detection
to minimize economic losses from a fire. Significant
308
Copyright © 2006 Taylor & Francis Group plc, London, UK
losses of production and property damage can occur
when a fire is discovered in early stages of develop-
ment. Only prevention through proper maintenance
and examination can effectively prevent reportable belt
entry fires.
REFERENCES
Francart, W. 2005. 2003 MSHA Survey – Atmospheric Mon-
itoring Systems in U.S. Underground Coal Mines. SME
Annual Meeting, Salt Lake City, UT.
Huntley, D. et al. 1986. Report of Investigation – Wilberg
Mine Fire. Mine Safety and Health Administration.
McPherson, M.J. 1993. Subsurface Ventilation and Environ-
mental Engineering. New York: Chapman and Hall.
Mine Safety and Health Administration. 2005. Accident
Prevention Program Safety Ideas: Is Your Belt Slippage
Protection Working? www.msha.gov.
Mine Safety and Health Administration. 1980–2005. Acci-
dent Investigation Reports.
Mitchell, D.W. 1996. Mine Fires. Chicago: Primedia Business
Media.
Paradies, M. 2005 Root Cause Analysis Meets Goals for Safe,
Reliable and Efficient Operations. System Improvements
Inc.: Knoxville, TN.
Stephan, C., & Desautels, L. 1989. Underground Coal Mine
Conveyor Belt Fires 1970-1988. Report Number 04-304-
89, Mine Safety and Health Administration, Industrial
Safety Division of Technical Support, Pittsburgh, PA.
Strahin, R., Wolfe, D., & Pogue, C. 1988. Report of Inves-
tigation – Marianna No. 58 Mine Fire. Mine Safety and
Health Administration
Swinderman, R., Goldbeck, L., & Marti, A. 2002. Founda-
tions 3. Martin Engineering, Neponset, IL.
Tolliver, W. 2004. Mine Emergency Preparedness. Presen-
tation at SME/PCMIA Annual Joint Meeting, Greens-
burg, PA.
Wilson, W. 1996. Report of Investigation – Dilworth Mine
Fire. Mine Safety and Health Administration.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Mine fires – measuring the efficacy of gas-enhanced foam

M.A. Trevits & A.C. Smith

National Institute for Occupational Safety and Health, Pittsburgh, Pennsylvania, USA

T.P. Mucho
Thomas P. Mucho & Associates, Inc., Washington, Pennsylvania, USA

A. Ozment & J.B. Walsh

US Foam Technologies, Inc., Longview, Texas, USA

M.R. Thibou
On Site Gas Systems, Newington, Connecticut, USA

ABSTRACT: The National Institute for Occupational Safety and Health (NIOSH), with US Foam Technologies,
Inc. and On Site Gas Systems, Inc., conducted research on the remote application of extinguishing agents with
the intent of improving deployment strategies to limit miner exposure and to help ensure the best possible
outcome during a mine fire. Full-scale in-mine were conducted at the NIOSH Lake Lynn Experimental Mine to
determine the flow characteristics, stability and fire-suppression capability of gas-enhanced foam. The in-mine
experimental work was designed to evaluate movement of gas-enhanced foam through mine workings closely
simulating an underground coal mine environment and how long gas-enhanced foam would remain stable. This
paper presents results of the experiments and provides valuable insight into mine fire deployment strategies for
gas-enhanced foam technology.

DISCLAIMER: The findings and conclusions in this report are those of the authors and do not necessarily
represent the views of the National Institute for Occupational Safety and Health.

1 INTRODUCTION

The occurrence of mine fires continues to be one of

the most persistent problems facing the mining indus-
try. Since 2000 there have been 19 major mine fires
(including thermal events) in underground mines in
the United States (Fig. 1). The National Institute for
Occupational Safety and Health (NIOSH), in partner-
ship with the Mine Safety and Health Administration,
is conducting on-going field and laboratory investi-
gations that are focused on the remote application of
extinguishing agents, deployment strategies for fire- Figure 1. Major underground mine fires in the United States
fighting equipment, and permanent and temporary (2000–2005).
fire containment sealing technology. The goals of this
research are to evaluate new and existing technologies Omans 1993b, Omans 1993c). The application of fire
and to limit miner exposure by developing or improv- fighting foam in underground mines was developed
ing control and extinguishment strategies to help in the 1950’s (Hartman et al. 1958, Nagy et al. 1960).
ensure the best possible outcome during a mine fire. Its use has been mostly limited to direct fire fight-
Fire fighting foam technology has been available ing application using high-expansion foam generators
for many years, and has been used extensively to fight located underground to push the foam and flood the
flammable and combustible liquid fires on the surface fire area (Scott & Nagy 1968, Banerjee & Acharya
(Colletti 1992, Brackin et al. 1992, Omans 1993a, 1986, Conti 1995, Conti & Weiss 1998). Other

309

Copyright © 2006 Taylor & Francis Group plc, London, UK

 applications include pumping of nitrogen-enhanced
high expansion foam into gob areas to control sponta-
neous combustion (Komai et al. 1989, Voracek 1993).
More recently, the methods that employ compressed
gas foams have been developed, resulting in smaller,
more uniform bubbles (Grady 1994). The chemicals
used to create foam concentrate are claimed to reduce
the surface tension of water, greatly increasing the pen-
etrating and wetting abilities of the water used, and
significantly increases the effectiveness of the water
supply (US Foam 2004).The addition of gas to the mix-
ture increases the resulting foam volume to between
five and fifteen times the volume of water used.
Foam addresses a fire condition through evapora-
tion of contained water and cooling by energy removal.
Foam serves to blanket the combusting material and
isolate it from oxygen. As the foam collapses, water is
released and the temperature of the water increases by
absorbing heat and eventually turns into the water into
steam. Water is released from foam either through bub-
ble rupture or through the effects of gravity distorting
the bubble walls. Because this process takes time, foam
can act as a water reservoir, releasing water at a rate that
allows absorption into the fuel, rather than running off
the surface (Snuffer Corp 2004). The four components
of a fire include: the oxidizing agent (any of the vari-
ous gases that support combustion), the reducing agent
(any material that is reducible to combustible materials
thus becoming a fuel), heat from within or without the
material, and a self-sustaining chemical chain reaction
(the action of the first three) (Gottschalk 2002). If the
foam is enriched with nitrogen gas, then it can serve
to remove two components of the fire tetrahedron by
robbing the fire of heat and removal or displacement
of oxygen. Given sufficient stability and the capability
to efficiently move and migrate through a mine open-
ing, it is thought that nitrogen-enriched foam could
provide an ideal technology for addressing mine fire
conditions.
The use of remotely-applied foam from the surface
through boreholes is a relatively new concept. Because
of the lack of experience with this new technology,
there are uncertainties with respect to the foam’s ability
to maintain its composition when pumped under pres-
sure through deep boreholes and its ability to move
away from the downhole location. In order to quan-
tify these and other parameters, a series of full-scale
experiments were conducted in NIOSH’s Lake Lynn
Experimental Mine (LLEM) located approximately
60 miles southeast of Pittsburgh, PA. The LLEM is
a world-class, highly sophisticated underground facil-
ity where large-scale explosion trials and mine fire
research is conducted. This work is being conducted
under the tenets of the NIOSH Research to Practice
(r2p) initiative which is aimed at reducing or elimi-
nating occupational disease and injury by increasing
the use of research findings in the workplace. During
310
Copyright © 2006 Taylor & Francis Group plc, London, UK
the in-mine experiments, the components of the foam
were mixed together on the surface and the result-
ing foam was injected through a borehole into the
mine void. Experiments were designed to measure the
foam’s stability, flow speed and length of flow. In addi-
tion, the ability of the foam to flow through non-linear
mine void configurations, through and around obstruc-
tions, and over pooled water was evaluated. Foam’s
fire suppression effectiveness was measured against
a deep-seated coal fire and a diesel pool fire. Lastly,
its ability to flow up a sloped entry was measured.
2 FOAM DELIVERY SYSTEM
US Foam Technologies, Inc. together with On Site Gas
Systems provided nitrogen gas-enhanced foam ser-
vices during the experiments. US Foam Technologies
Inc. offers a specialized gas-enhanced foam generating
system and On Site Gas Systems provides nitrogen gas
for the system using portable skid-mounted nitrogen
plants that extract nitrogen gas from the atmosphere
using membrane technology (On Site Gas Systems
2005).
US Foam Technologies Inc. owns a proprietary
nitrogen-enriched foam delivery system known as the
Hellfighter™ which includes a pipeline manifold and
a sophisticated mixing chamber. Water is pumped at
a controlled rate and pressure into a line that is con-
nected to one of the inlet ports on the Hellfighter™
Mine Fire Fighting Foam™ (MFFF) concentrate, a pro-
prietary formulation of chemicals designed to produce
long-lasting foam that can withstand higher temper-
atures than Class-A Fire Fighting Foam (AFFF), is
added to the waterline using precision injection pumps.
Nitrogen gas from On Site Gas Systems’ membrane
plant is then injected at a controlled rate into another
inlet port on the Hellfighter™ . The fluid (water and
foam concentrate) and nitrogen gas is mixed within
the Hellfighter™ and creates nitrogen-enhanced foam
that flows from the outlet port of the Hellfighter™ to
wellhead assembly attached to the injection borehole
and into the mine void. A schematic of the foam system
is shown in Figure 2.
Up to three small Hellfighter™ units (or one large
unit) can be connected to a single borehole depending
on the diameter of the hole. Water, foam concentrate
and nitrogen gas flow rates can be adjusted to produce
a variety of foam mixtures from a thin water-foam
blend to a thick froth similar to shaving cream. In addi-
tion, the material being injected can be switched (with
no downtime) from foam to only nitrogen gas simply
by closing a valve. The gas-enhanced foam and nitro-
gen membrane plant combination provides an ideal
platform for addressing mine fires because it can be
readily moved from one borehole location to another
and can be deployed using off-road equipment. Bulk
 Figure 2. Schematic of the foam system.
liquid nitrogen trucks can deliver more nitrogen in a
unit of time but their use is limited by rugged road
conditions.
For these experiments, water was stored in a 10,000
gallon tank at the surface location of the borehole and a
fire engine was used to pump the water. The foam con-
centrate was supplied in 250 gallon plastic “tote” con-
tainers. The nitrogen was supplied by a 1,000 ft3 /min
membrane separation plant. Three Hellfighter™ units
were attached to the injection borehole.
3 LAKE LYNN EXPERIMENTAL MINE
A layout map of the LLEM is shown in Figure 3.
The underground mine entries were developed adja-
cent to an abandoned commercial limestone quarry
and underground limestone mine. The entries of the
abandoned limestone mine, labeled as the old work-
ings, are approximately 49 ft wide by 33 ft high. The
LLEM contains 5 drifts openings called A, B, C, D,
and E as shown in Figure 3. These entries were devel-
oped to approximate the size of a typical Pittsburgh
Coalbed mine, about 20 ft wide by 6 ft high, and range
from 500 to 1,500 ft in length. The entries, in con-
junction with the use of two explosion-proof bulkhead
doors that can be positioned to open or close an entry,
can be made to simulate room-and-pillar and longwall
mine configurations (NIOSH 1999).
The injection borehole penetrated the mine void at
a depth of 197 ft. The injection borehole intercepts the
mine workings in the first crosscut outby the face area
between the B and C drifts. The borehole is completed
to the mine opening with 6-in diameter casing. At the
location of the borehole, the mine opening is 7-ft high
and slopes towards B drift on a 1.13 percent gradient.
4 EXPERIMENT NO. 1
Experiment No. 1 was designed to evaluate the phys-
ical characteristics and flow behavior of foam. Air
was used as the gas in this experiment so researchers
311
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 3. Layout map of the Lake Lynn Experimental Mine.
Figure 4. Layout map of mine workings for Experiment
No. 1.
could observe foam movement without the need to use
self-contained breathing apparatuses. The parameters
studied in this experiment included measuring the con-
dition and volumetric changes of the foam as it was
pumped into the borehole and entered the mine, foam
stability, flow speed, length of flow, stacking charac-
teristics, movement through and around obstructions,
movement over a pool of water, and flow through non-
linear mine configurations. To accomplish this, the
multiple entry section of the mine was configured as
shown in Figure 4. To evaluate the long-term stability
of the foam under the typical mine temperature and
humidity conditions, a 4.3-ft high concrete block stop-
ping was construct in the closed end of the A drift to
create an isolated area that could be filled with foam.
This area measured 92 ft-long by 20 ft-wide. To trans-
port the foam to this area, a flexible hose was attached
to the bottom of the injection borehole and extended to
this enclosure. The volume of this chamber to the top
of the stopping (7,912 ft3 ) was then filled with foam.
A combination of block, plywood, and brattice stop-
pings was used to direct the foam as it exited the
borehole. A plywood stopping at the outby end of
the B drift stub and a concrete block stopping in the
 first crosscut between A and B drifts were installed
to direct the foam down B drift. The second cross-
cut between A and B drifts was filled rib-to-rib and
floor-to-roof with broken rock to simulate compacted
gob. This area was used to determine if foam would
penetrate the broken rock mass or if the rock only
served to block the flow path. Another block stopping
was placed in the B drift just outby the third cross-
cut. One block was replaced with a window and a
camera was located behind the stopping window to
record the foam movement. Brattice stoppings were
placed in the second and third crosscuts between B
and C drifts and in the fourth, fifth, and sixth cross-
cuts between A and B drifts. Because of the slope of the Figure 5. Rock pile heater element temperature profiles.
mine floor, this combination of stoppings directed the
flow of the foam across a broken rock pile in crosscut 3
between the A and B drifts. This rock zone was approx-
imately 3-ft high and spanned the width of the opening.
Three electric resistance heating elements were buried
at the center of this rock pile to simulate a deep-seated
hot spot, to determine if the foam could infiltrate the
broken rock and reduce the temperature of the hot spot.
Two rows of four crib blocks were constructed in
the A drift between the third and fourth crosscuts and
a pool of water was created in A drift between the
fourth and fifth crosscuts. The purpose of the pool of
water was to determine whether foam would move over
a body of water, if the pool created a barrier to foam
flow, or if the pool served to degrade the foam. The
water pool was created by constructing a watertight Figure 6. Foam flow and build-up in the crib block area.
2-ft high block stopping across the width of A drift to
act as a dam. Because of the slope of the mine floor, a The mine stopping configuration then caused the
pool of water approximately 50 ft long was made in the foam to turn into the third crosscut between A and B
A drift. Lastly, a 26 ft long conveyor belt structure was drifts, where the foam encountered the 3-ft high rock
placed 3.5 ft from the rib area in the A drift between pile with the heating elements. Figure 5 shows the
the fifth and sixth crosscuts to observe the behavior of temperature profiles for the thermocouples attached
the foam as it flows through or around mine structures. to the four heating elements located in the rock pile.
The initial attempts to deliver high quality foam Unfortunately, the heating elements failed after reach-
to the mine entry required some adjustment to liquid ing temperatures ranging from 28◦ to 34◦ C, just 15◦ to
components and the rate of gas injection.This is typical 20◦ C above ambient temperature. Unfortunately, the
of foam use and is usually done on the surface by the limited exposure of rock to the heating elements did
equipment operator. In this case, the adjustment was not raise the rock temperature enough to measure the
made through communications between in-mine per- effect of the foam on the rock temperature. However,
sonnel and personnel on the surface at the pump site. it is clearly obvious that the foam quickly infiltrated
After adjustment, foam having a shaving cream-like the rock pile and cooled the heating elements within
consistency was achieved at the bottom of the bore- minutes.
hole. The foam moved down and around the corner After turning the corner into the A drift, the foam
of the first crosscut into B drift. The foam reached a moved until it encountered the sets of crib blocks. Fig-
height of 3-to 4-ft through B drift as it moved down ure 6 shows that the crib blocks formed an effective
slope and away from the borehole. barrier to the foam, causing it to build up to roof height
The first obstacle encountered in B drift was the behind the crib blocks. Once past the crib blocks, the
rock pile in the second crosscut between A and B drifts. foam height in the entry was reduced to about 2-ft.
The camera mounted on the A drift side of the rock pile The foam then approached the water pool in the A
showed that foam did not penetrate the broken rock drift. The water had no apparent effect on the foam
mass. Water, however, that apparently settled out from as it moved over it without any apparent affect on
the foam was observed flowing from the bottom of the foam quality. Finally, the foam encountered the con-
rock pile. veyor belt structure. The structure acted much like the

312

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 7. Rate of decay of foam volume versus time.
sets of crib blocks in the A drift, causing foam build-
up upstream of the structure. Since the structure only
occupied 4.8 ft of the entry width, the foam moved
through the unobstructed part of the entry similarly to
the way it moved through the B drift.
As mentioned previously, the closed end of the A
drift was filled with foam to a height of 4.3 ft. Figure 7
shows the rate of decay of the foam versus time. The
foam lasted for 9 days and the rate of decay was linear
over the 9 day period at about 900 ft3 . It should be noted
that this area was not exposed to mine ventilation air
flow and was strictly a measure of the rate of foam
degradation.
5 EXPERIMENT NO. 2
Experiment No. 2 was designed to evaluate the fire
fighting effectiveness of foam against a deep-seated
coal fire and a flammable liquid (diesel fuel) pool
fire. In this experiment, nitrogen was used as the gas.
The mine configuration to direct the flow of the foam
was the same as in Experiment No. 1. The obstruc-
tions in the A drift were removed since the purpose
of this experiment was to evaluate the ability of foam
to extinguish the test fires. A concrete block stopping
was placed in the third crosscut to contain the foam
to B drift. Because of this configuration, B drift was
not ventilated. The diesel fuel fire was located in the
center of the third crosscut in B drift. Five gallons of
diesel fuel was floated on a 1-in water layer in a 3-ft by
3-ft square metal tray. This fire was ignited just prior
to the arrival of the foam front. The size of fire was
approximately 600 kW. The deep-seated coal fire was
placed just up slope of the diesel fuel fire in the center
of the mine entry. To create the coal fire, approximately
250 lb of coal was placed on top of 25 lb of commer-
cial cooking charcoal in a 3-ft by 3-ft square metal tray.
This fire was ignited approximately 60 minutes before
the foam flow was initiated. Figure 8 shows a layout
map of the mine for Experiment No. 2.
313
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 8. Layout map of mine workings for Experiment
No. 2.
Figure 9. Temperature–time trace for thermocouples
located above the diesel fuel fire.
The diesel fuel fire was monitored using two ther-
mocouples located 4-in above the fuel layer. The coal
fire was monitored using seven thermocouples located
throughout the coal pile. Both fuel and coal trays were
also monitored by a thermal imaging infrared cam-
era that enabled researchers to remotely view the fire
through the foam. Figure 9 shows the temperature–
time trace for the thermocouples located above the
diesel fuel tray fire. The fire was ignited and allowed to
burn for 10 minutes before the foam reached the fire.
Temperatures of about 800◦ C were recorded 4 inches
above the fuel layer. When the foam reached the fuel
tray, it easily enveloped and extinguished the fire.
Figure 10 shows the temperature–time plot for the
thermocouples located in the coal fire tray. Tempera-
tures near the top of the coal pile reached about 850◦
and 625◦ C. Other areas of the coal pile were between
100◦ and 250◦ C, with the exception of the two ther-
mocouples located on the outer edge of the coal pile
which only reached about 50◦ C. The foam reached and
enveloped the coal fire tray and quickly cooled the hot
spots. Temperatures deeper inside the coal pile slowly
decreased over the next hour.
 Figure 10. Temperature–time trace for thermocouples
located in the coal fire.
Data from the infrared camera that was used to
image the fuel and the space above the coal tray showed
the space above the fire tray was reduced to about
15◦ C. The temperature of the foam downstream from
the fuel tray was above the ambient temperature of
the foam showing that it carried heat away from the
fire. At the same time, the maximum temperature in
the coal pile, as measured by the thermocouples, was
about 90◦ C. This suggests that foam be used to quickly
cool the airspace near a deep-seated fire, reducing the
chance for spreading the fire. Furthermore, the nitro-
gen gas used to create the foam reduces the oxygen
concentration in the airspace, thereby reducing the
chance of a methane ignition.
During this experiment, foam reached the roof in
B drift because of the stopping configuration, which
essentially closed off B drift. In this confined exper-
iment, foam filled the mine void and then moved
upslope towards the injection borehole. Foam was also
observed penetrating through the top of the rock pile
located in the second crosscut between A and B drifts.
6 EXPERIMENT NO. 3
Experiment No. 3 was designed to evaluate the abil-
ity of the foam to fill a single mine entry. The 500 ft
long E drift was used for this experiment. A layout
map of the LLEM for this experiment is shown in Fig-
ure 11. The slope of E drift (6.2 percent) is much more
severe than in the multiple entry section of the mine. In
this experiment, concrete stoppings were constructed
in the crosscut between B and C drifts and in C drift
to isolate E drift, creating a long single entry. Several
blocks were removed from the first course of blocks in
these stoppings to allow water to flow from the degra-
dation of the foam. It was thought that water could
collect in this area possibly causing the block structure
to fail. A longwall shearer was located near the dis-
tal end of E drift and was utilized to evaluate foam’s
Figure 11. Layout map of mine workings for Experiment
No. 3.
Figure 12. Progression of the body of foam along the E drift.
ability to move around the obstruction after flows a
long distance.
The foam was delivered through the borehole and
filled the closed area (crosscut between B and C drift
and area that was closed at the opening to the C drift)
and then began moving up into the E drift. The foam
reached the mine roof quickly because of the severe
slope of the entry in this area. The rate of advance of
the foam plug, as measured by the lowest point of the
leading edge of the foam, was constant at 1.5 ft/min
once the foam reached the mine roof and began mov-
ing up the E drift. Over the first 50 minutes, the foam
had not yet reached the roof since it was filling the
first crosscut at the intersection of C and E drifts. The
pressure exerted on the concrete stoppings was mea-
sured and was approximately equal to the foam density.
Figure 12 shows the progression of the body of foam
along the E drift.
The foam filled the entire length of E drift in five
hours. The experiment was stopped when the foam
reached the fan shaft at the end of E drift to prevent
disruption of the ventilation system. The rate of foam
dissipation was measured over the next seven days, and
is shown in Figure 13. At the time the foam injection

314

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 13. Rate of foam dissipation in the E drift.
was stopped, the length of the E drift filled with foam
was 469 ft. Initially, the foam was quite stable, decay-
ing at a rate of only about 8.8 ft/day (1,145 ft3 /day)
after the first day. The rate eventually increased to
about 64.7 ft/day (8,410 ft3 /day) at seven days. The rate
of decay over the first three days is comparable to the
results from the stability test conducted earlier in the
A drift stub. The rapid increase in decay rate after three
days is possibly attributed to air pathways developing
over the top of the body of foam as it degraded allow-
ing airflow to pass over the foam. This likely increased
the rate of dehydration of the foam.
7 SUMMARY
The foam experiments were very successful in evalu-
ating many of the parameters that can affect the use of
foam for remotely fighting mine fires. However, these
experiments were also limited in scope with respect to
the type of foam used, the geometry and slope of the
mine entries, and the size and types of fires. Therefore,
these results should be used as guidelines for the use
of foam, and not as design specifications for its use.
The foam used in Experiment No. 1 showed good
stability over time. Foam stability is a function of foam
concentrate, water volume, and gas flow, and these
parameters can be varied based on the particular appli-
cation. Observations during the experiments suggested
that the life of gas-enhanced foam is highly depen-
dent on the quality of the foam (relative ratio of water,
foam concentrate and nitrogen gas). “Wet” foam had
shorter life (from a few minutes to a few hours) and
“dry” foam lasted as long as several days. During the
LLEM experiments when “wet” foam was created and
the system was shut-down the foam degraded quickly
and flowed away from the injection borehole. In addi-
tion, the life of the foam can be dependent on the mine
temperature and wetness or dryness of the environ-
ment, water hardness, and pH. However, the effect
315
Copyright © 2006 Taylor & Francis Group plc, London, UK
of varying these parameters was not measured in this
experiment.This experiment also showed that the foam
quality can be adequately managed from the surface.
However, periodic monitoring of the foam consistency
on the surface is recommended to ensure stable foam
is entering the mine.
Foam flow, speed, and ability to reach the mine roof
and fill the entire cross-section of the mine opening
were shown to be highly dependent on the slope of
the mine floor. In the first two experiments, the foam
flowed quickly and did not reach the roof of the mine
until it was obstructed by the sets of crib blocks. In
the third experiment, the foam reached the mine roof
very quickly because of the upslope of the mine floor
and the obstructions that did not permit down slope
flow, but the rate of advancement of the foam front
was much slower than in the first two experiments.
In confined applications where the objective is to
flow foam from the borehole to a fire location some
distance away, this parameter is extremely important.
In a situation where the elevation of the fire is below
the foam entry point into the mine, the foam will flow
quickly, however an obstruction, such as a remotely
installed mine seal, will be needed to cause the foam
to completely fill the mine void from the floor-to-roof.
In situations where the foam will need to move up
slope, it will be critical to get the location of the foam
injection borehole as close as practicable to the fire
and position obstructions down slope from the foam
injection point. Again, the use of remotely installed
mine seals could be necessary. In situations where the
mine entry is level, foam should flow in all directions
away from the injection borehole, but the flow will be
highly dependent on obstructions. Unfortunately, this
scenario was not tested at the LLEM.
The nonlinear configuration of the mine open-
ings used in the first two experiments demonstrated
that foam will flow around corners, but again the
major influence on its effectiveness was elevation and
obstructions. In the first two experiments, foam flowed
past the rock rubble pile in the second crosscut between
A and B drifts, essentially treating it as a stopping. The
stopping in B drift acted to turn the foam into the cross-
cut between A and B drifts. However, when the foam
reached A drift, which was open in both directions, the
foam only flowed in the down slope direction.
The ability of foam to suppress a liquid pool fire
was shown in the second experiment. However, foam’s
ability to cool a hot, deep-seated fire was not com-
pletely addressed in these experiments. Foam did cool
the high temperatures in the coal fire, but this was
a relatively small fire. Foam was able to infiltrate the
rock pile and quickly cool the heating elements in the
second experiment, but the temperatures were not high
enough to truly test the foam’s cooling ability. Foam’s
inability to penetrate through the rock pile in the first
experiment was noted. In the second experiment, foam
 moved through the rock pile near the mine roof when
it was confined to the B drift. This indicates that the
compaction of gob material and other possible path-
ways will determine the foam’s ability to infiltrate and
penetrate deeply into a gob area.
ACKNOWLEDGMENTS
The authors acknowledge the support of Frank A.
Karnack, William Slivensky, Donald D. Sellers,
Kenneth Jackson, and Frances J. Goff, and Richard A.
Thomas in preparation of the experimental set-ups and
data acquisition and analysis, and William Monaghan
for gathering the thermal imaging data.
REFERENCES
Banerjee, S.C. & Acharya, A.K. 1986. High Expansion
Foam – Its Use in Mine Fire. J. Mines, Metals, & Fuels
October: 448–451.
Brackin, M., Bublitz, D., Crosley, S., Farmer, K.,
Wallace, L. & Wiggins, T. 1992. The Pros and Cons of
Class A Foam. Fire Engineering July: 59–66.
Colletti, D.J. 1992. Class A Foam for Structure Firefighting.
Fire Engineering July: 47–57.
Conti, R.S. 1995. Inflatable Partitions for High-Expansion
Foam Generators. J. Mining Eng June: 561–566.
Conti, R.S. & Weiss, E.S. 1998. Inflatable Devices for
Combating Underground Mine Fires. Proc. Minesafe
International September 28: 388–393.
DeRosa, M.I. 2004. Analysis of Mine Fires for All
U.S. Underground and Surface Coal Mining Categories:
1990–1999. NIOSH Information Circular 9470.
Grady, C. 1994. Compressed Air Foam – An Idea Whose Time
Has Come. American Fire Journal January.
Gottschalk, J. 2002. Firefighting. New York: DK Publish-
ing, Inc.
316
Copyright © 2006 Taylor & Francis Group plc, London, UK
Hartmann, I., Nagy, J., Barnes, R.W. & Murphy, E.M. 1959.
Studies With High-Expansion Foams For Controlling
Experimental Coal Mine Fires. U.S. Bureau of Mines
Progress Report August.
Komai, T., Isei, T., Shikada, N., Kinoshita, M., Suzuki, T.,
Zong-cheng, L., Guang-yang, H., Yun-guang, F. &
De-chang, X. 1989. Underground Fire Extinguishing
by the Combined System of Inert Gas Generator and
Foam Generator. Shigen: National Research Institute for
Pollution and Resources.
Nagy, J., Murphy, E.M. & Mitchell, D.W. 1960. Controlling
Mine Fires With High-Expansion Foam. US Bureau of
Mines Report of Investigations 5632.
National Institute for Occupational Safety and Health. 1999.
Lake Lynn Laboratory (flyer), DHHS (NIOSH) Publica-
tion No. 99–149.
Omans, L.P. 1993a. Fighting Flammable Liquid Fires,
A Primer, Part 1: The Family of Foams. Fire Engineering
January: 50–61.
Omans, L.P. 1993b. Fighting Flammable Liquid Fires,
A Primer, Part 2. Fire Engineering February: 50–58.
Omans, L.P. 1993c. Fighting Flammable Liquid Fires,
A Primer, Part 3: Fire Engineering March: 93–104.
On Site Gas Systems, Inc. 2005. Home Page, http://www.
onsitegas.com. April 4.
Scott, F.E. & Nagy, J. 1968. Fighting Fires with High-
Expansion Foam. Coal Mining and Processing June:
42–45.
The Snuffer Corporation. 2004. Home Page, http://www.
snuffer.com. July 23.
U.S. Foam Technologies, Inc. 2004. Home Page, http://www.
usfoam.com. July 18.
Voracek, V. Necessary Data on High-Pressure Foam Gen-
erator of a Nitrogen Foam For Its Utilization In the
Field of Both Prevention and Suppression of Sponta-
neous Combustion of Coal In Caved Areas of Longwall
Faces. Ostrava – Radvanice, Czech Republic CZ-716 07:
Research Mining Institute, Inc.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Fire-extinguishing and explosion-suppression technology for a

super-large and shallow working underground fire zone

Zhou Fubao & Wang Deming

School of Mineral and Safety Engineering, CUMT, Xuzhou, China

Zhang Yongjiu, Zhang Yuliang & Li Xiang

Ningxia Mining Industry Group, Yinchuan, China

Jerry C. Tien
Department of Mining Engineering, University of Missouri-Rolla, Rolla, MO, USA

ABSTRACT: The Baijigou Mine fire in Ningxia Province, which broke out on October 24, 2003, affected
more than 10 Mm3 of the mine and was probably the largest underground fire in China in recent years. In
addition to its size, the fire was also characterized by excessive air leakage and potential for violent methane
explosions. A series of new measures were employed to fight the fire, including sealing intake tunnels with
water, injecting three-phase foam through boreholes, and flushing with a large volume of nitrogen. The fire was
successfully extinguished and production resumed soon afterwards; not one single methane explosion occurred
during fire-fighting and afterwards.

1 INTRODUCTION

Mine fires are one of the top five disasters in under-

ground coal mines in China. Records show that 2,006
miners lost their lives due to coal mine fires between
1949 and 1995, of which, 138 fires had three or more
deaths that killed a total of 1,670 miners (Wang & Li
2002).
Baijigou Coal Mine, a part of the Ningxia Min-
ing Industry Group, is located in the middle of Helan
Mountain, Ningxia Province (Fig. 1). It is one of the
top Taixi anthracite production bases in China.
Low in ash, sulphur and phosphor, coal in Baijigou
has high mechanical strength with a good capacity for
absorption. Coal is exported to over 20 countries in
Europe, the Americas, and Asia. Baijigou started pro-
duction in 1972 producing 1.2 million metric tonnes Figure 1. Location of Baijigou Mine.
annually (Mt/a) at the time. After several expansions in
subsequent years, its 2003 annual production exceeded quantities of air are drawn through cracks from the
2.0 Mt. surface to underground workings.
The mine is highly gassy, having an absolute and rel- At 19:00 on October 24, 2003, a gas explosion
ative methane emissions of 45.7 m3 /min and 31.8 m3 /t, occurred in the goaf area of No. 2421-1 working face
respectively. The coal is susceptible to spontaneous and the explosion damaged the permanent stoppings
combustion and has an incubation period of approx- in No. 1700 return airway. Electricity underground
imately 12 months; it is also susceptible to dust was deenergized and miners were evacuated imme-
explosions (Zhou et al. 2005). Baijigou has an aver- diately. Meanwhile, the mine ventilation system was
age mining depth of approximately 100 m and is fully adjusted, and a rescue team was dispatched under-
mechanized, using both standard longwall and long- ground to assess damages in specified sites. More than
wall caving methods. Because of subsidence, large one hundred gas explosions occurred subsequently in

317

Copyright © 2006 Taylor & Francis Group plc, London, UK

 (a) Shockwave out of the mine openings
Figure 2. With continuingly expanding fires underground (right), smoke was visible from surface.
the goaf during the following 11 hours. Despite vari-
ous attempts, explosions and fires could not be brought
under control. Moreover, the fire in areas affected by
the explosions continued to expand. With underground
conditions continuing to deteriorate, both the provin-
cial government and the State Administration of Coal
Mine Safety (SACMS) in Beijing were notified on
October 29, 2003.
On November 2, 2003, six experienced fire-fighting
experts and two officials from SACMS arrived at
the mine site. However, the underground situation
appeared to be out of control: shock waves from the
explosions had reached the mine entrance, and the
command center was forced to seal the entire mine
(Fig. 2). Of the over 10 Mm3 of space sealed under-
ground, the fire was believed to have affected more
than 3 Mm3 . Only two mine fires in recent years
were larger in terms of the size of affected area, the
Meiyukou Coal Mine fire in Datong Coal Mining
Group (1983; 5 deaths) and the Duerping Coal Mine
fire in Xishan Coal Mining Group (1985; 48 deaths).
Both were eventually controlled through sealing and
pressure-balance, which reduced their influence on
other part of the mine. The affected areas were not
opened for reproduction.
Fire area underground continued to expand because
of a continuing supply of oxygen due to leakage from
the surface through ground cracks. This increased
chances of additional explosions, and the mine man-
agement was faced with serious decisions.
2 SUPPRESSION OF GAS EXPLOSION
THROUGH REMOTE WATER SEALING
Despite the sealing of the entire mine, gas explosions
were not controlled during the first rescue attempt.
Intake airways could not be effectively sealed off ini-
tially and they continuously supplied air to the fire
area; these intake airways had to be more tightly sealed
318
Copyright © 2006 Taylor & Francis Group plc, London, UK
(b) Seals collapsed after explosion
in order to avoid likely explosions underground. How-
ever, continuously expanding fire zones made sealing
operations almost impossible; an alternate method was
urgently needed and a remote “water-sealing” measure
was proposed.
The method was to flush water into the intake air-
ways through mountain cracks from the surface and
from the gas drainage pipes leading to the airways
(Fig. 3). Since the airways were at a lower level, water
was easily introduced to fill these airways due to grav-
ity, which effectively forced air out of the fire zone
that subsequently suffocated the fire. Besides, water
that seeped through the gas-drainage boreholes helped
reduce the further methane emissions from adjacent
coal seams.
After “water sealing,” the methane explosions
slowly stopped. This gained precious time for fur-
ther extinguishing of fire and led to subsequent mine
recovery (Fig. 3).
3 FIRE-FIGHTING WITH THREE-PHASE
FOAM
The three-phase foam is composed of noncombustible
material (fly ash or mud), inert gas (nitrogen) and
water, which, used properly, have effectively con-
trolled mine fires previously (Michaylov 1995, Wang
2004). Following nitrogen infusion into the slurry con-
taining the foam agent, the three-phase foam expanded
rapidly in volume. This effectively filled the mine goaf
and covered the loose coals in the area. The nitro-
gen encapsulated in the foam was finally released and
continued to “suffocate” the mine fire. In addition,
three-phase foam contains solid materials such as fly
ash or mud, which become part of the foam surface.
This helps keep the foam stable for a longer period
of time. As the foam cells rupture, the fly ash and
mud can cover the coal in the goaf area uniformly as
 5000

4500

4000

3500

3000

2500

2000

1500

1000

500

11-12
11-13
11-14
11-15
11-16
11-17
11-18
11-19
11-20
11-21
11-22
11-23
11-24
11-25
11-26
11-27
11-28
11-29
11-30
12-1
Figure 3. Flushing water through mountain cracks (left) and water consumption vs. time in the fire-fighting operation (right).

Foam stabilizer Foaming agent Nitrogen gas

Fly ash Slurry

Pump Mixer Foaming generator
or mud tank

Fire zone Boreholes Three-phase foam

Figure 4. Schematic showing the introduction of the three-phase foam to underground fire zone.

a result of its viscosity, that have kept coal from further

oxidation.
While fighting the Baijigou fire, it was difficult to
locate the fire sources because of the large area of min-
ing and a lack of direct access to underground workings
because the mine has been sealed after the fire. This
makes foam infusion through the surface boreholes the
only practical means of fire control.

3.1 Injecting three-phase foam from

surface boreholes
Steps for foam preparation and subsequent injection
are shown in Figure 4 (Zhou et al. 2005). The initial
slurry, composed of water and fly ash or mud, was
prepared in the slurry-mixing tank. Foam stabilizer
was then added to the tank to provide a more uni- Figure 5. Drilling of T685WS gadding machine.
form foam mix. The prepared slurry was then pumped
into the grouting pipes where the foaming agent was
3.2 Borehole drilling
introduced by a volumetric screw pump in a proper
proportion. The slurry and foaming agent were fully The T685WS gadding machine (Fig. 5), manufac-
mixed and then pumped into the foaming generator. tured by Schramm Corporation, USA, was used to
Nitrogen was then added to the foaming generator drill boreholes (Song et al. 2004). This machine over-
through a rubber pipe, and three-phase foam was came many difficulties including rugged terrain, low
formed following the interaction between the nitro- temperatures, insufficient water supply, etc. Despite
gen and slurry. Gravity and pumping delivered the these obstacles, eleven boreholes were rapidly drilled
three-phase foam to the fire zone through boreholes. successfully (Fig. 6).

319

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 6. Boreholes and connecting pipes on the surface at Baijigou Coal Mine.

(a) The slurry-mixing tank (b) The foaming equipment

Figure 7. Foam pumping operation at the mine site.

0.60 0.25
Concentration of CO (%)
Concentration of CO (%)

0.50 0.20
0.40 0.15
0.30
0.10
0.20
0.05
0.10
0.00
4.8

5.2
5.4
5.6
4.10
4.12
4.14
4.16
4.18
4.20
4.22
4.24
4.26
4.28
4.30

0.00
4.1

4.2

4.3
3.23

3.26

3.27

3.28

3.29

3.31

4.13

Date
Date
(a) CO concentration vs. time at Borehole No. 12 (b) CO concentration at Borehole No. 5

Figure 8. CO concentration vs. time after foam injection at Boreholes Nos. 5 and 12.

3.3 Foam injection 4 The foaming multiple is 30;

5 The foaming agent was supplied at 0.06 t/h.
Large quantities of three-phase foam were pumped
into the underground fire zone through drill holes
(Fig. 7). The following were some of the relevant The entire on-site pumping operation was under
operating parameters: the guidance of the research team from the China
University of Mining & Technology. Pumped foam
1 The slurry was mixed at 20 m3 /h (with a water to fly quickly hardened in the fire zone and appeared to
ash or mud ratio at 3:1 to 5:1); cover the entire fire zone rather rapidly, which effec-
2 Nitrogen was supplied at 600 m3 /h; tively stopped the mine fire from further spreading.
3 On-site foam was generated at 620 m3 /h; The nitrogen released from the three-phase foam kept

320

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 9. On-site nitrogen generators.
1200000
1071500
1000000
800000
N2/m3
580332
600000
477850
400000
149450
200000
21475
0 COMPLICATED CONDITIONS
1# 3# 4# 5# 6#
Boreholes
Figure 10. Various amount of N2 were introduced through
different boreholes.
the goaf smothered and subsequent analysis of gas
samples from the fire area indicated that concentration
of O2 was brought down to below 5 percent rela-
tively quickly (Fig. 8), suggesting that the fire has been
brought under control.
4 FLUSHING LARGE QUANTITY NITROGEN
As mentioned in an earlier section, there was signifi-
cant air leakage from surface to underground openings
through surface cracks, this caused oxygen concen-
tration in the goaf area to be at a dangerously high
level which, in turn, caused the fire to continue burn-
ing in the goaf area and increased the possibility for
further methane explosions underground. To further
quench this burning fire, additional inert gas/material
has to be introduced. Four nitrogen generators with a
total capacity of 2,400 m3 /h were eventually brought
to Baijigou Coal Mine (Fig. 9).
To introduce nitrogen into the underground fire
zone, grouting materials were first applied to seal
as much as possible all cracks in ribs and roofs and
surface cracks with mud to reduce air leakage. To
maximize the nitrogen flushing effect underground,
higher than normal oxygen in the flushed material was
Copyright © 2006 Taylor & Francis Group plc, London, UK
provided. By the end of July 2004, more than 10 Mm3
of nitrogen had been flushed into the underground fire
zone (Fig. 10). As a result, oxygen concentration in
the fire area was gradually reduced to below 5 percent,
which made the goaf effectively inert. That prevented
further methane explosions underground.
189020 165600
5 REDUCTION OF SEALED AREA UNDER
12# 13#
No. 1660 Entry is the main intake tunnel, connected
to the fire zone through three other tunnels: Nos. 1, 2
and 3; it was also one of the locations where explosion
occurred (Fig. 11). Methane drainage collecting pipes
underground were all destroyed during the explosions.
At the first phase of mine rescue, portions of the under-
ground airways were deliberately flooded using large
amount of water to keep the airflow from entering into
the fire zone. Although effective, the large amount of
water in Entry 1660 was also a big nuisance as it has
effectively reduced area for travel in the sealed area.
After extensive discussion among members of the
command center, it was determined to use nitrogen to
replace water to inertize the entire sealed area. First,
water was drained in three different stages and seals
constructed in Entries No. 1, No. 2 and No. 3, one at a
time. Large quantities of three-phase foam were then
introduced from the surface boreholes to inertize air-
flow in the No. 1 Entry. Methane in No. 1660 Entry
was drained naturally through broken pipes, which
prevented methane from getting into the fire area;
slowly, water and methane in No. 1660 Entry were
successfully removed.
6 UNSEALING FIRE AREA AND SECONDARY
EXPLOSION PREVENTION MEASURES
It is critical that all fires in the sealed area are totally
extinguished before the area is unsealed. As a rule
321

Middle raise
N2
el
Tunn
Three-phase
No.1
1660.8 1661.4
1660.2 1661.6 No.1
Connecting
Way
1640 North Entry
Figure 11. Illustration of technical plan.
Figure 12. Illustration of local pressure-balancing
unsealing.
of thumb, a CO concentration of less than 10 ppm
in the fire zone has been commonly used as such an
indicator. However, because of the massive amount
of CO that had accumulated in the sealed fire zone
due to extended burning, removal of CO remained
a significant challenge.
Furthermore, two other indices, temperature and
O2 concentrations, were also used to gauge the status
of the fire behind the seals. Air temperatures behind
the fire zone were measured through boreholes with
thermal sensors and results showed that air tempera-
ture ranged between and 12◦ C behind the fire zone,
nearly the same as the temperature prior to the fire
started. This, coupled with normal oxygen concentra-
tion readings, it was thought the fire was already been
distinguished and the sealed area unsealed.
As shown in Figure 12, the haulage entry was sep-
arated from the workface by stones and caved coals
due to gas explosions. With the workface also filled
with the remaining grouting slurry in the lower cor-
ner, it could only be unsealed from the return entry.
Since a large amount of gases still remained in the goaf
area, a so-called “local pressure-balancing” method
was applied to unseal the fire zone in order to keep the
methane from escaping into entries.
To properly apply this method, two regulating doors
were first installed in the return entry; then a local
ventilator was placed to provide airflow to the work-
face with the ducts going through the doors; the doors’
opening areas were adjusted to allow the concentration
of CH4 , O2 and CO to be within a predetermined range.
Copyright © 2006 Taylor & Francis Group plc, London, UK
N 166
1685.5 0 In
take
No.
2 Tu
nne
l
1662.5 1660 Entry 1664.2
1662.3 No.2 No.3
Airflow Connecting Connecting
Way Way
1640 South Entry Gas-drainage Pipes
The purpose of this is to reduce the airflow into the
fire zone and suppress the gas emission. The method
was proven to be quite effective in controlling further
methane emissions.
Since there might be isolated spots with high tem-
peratures that might rekindle further methane explo-
sions, precautionary measures were taken as well.
These included:
1 continuing to seal other passages that were con-
nected to the goaf;
2 flushing nitrogen at the workfaces that have been
unsealed;
3 injecting three-phase foam into the suspected high-
temperature spots; and
4 reducing air quantity below normal stage.
These precautionary measures have been proven
effective as there have been no further after-
combustions since the fire zone was unsealed.
7 APPLICATION AND RESULTS
Fire-fighting efforts at Baijigou Coal Mine have been
very successful. This not only distinguished one of
China’s major fire disasters in years, but provided
prompt and proper decisions to allow the mine to
quickly resume coal production.
Part of Baijigou Coal mine was unsealed in mid-
November 2003, which allowed coal production to be
partially resumed less than one month later. The No.
2421-1 full-mechanized caving face was unsealed suc-
cessfully on September 28, 2004, which allowed the
mine to resume full production on December 6, 2004.
Most significant of all, despite of the several hun-
dred methane explosions after the fire, not a single life
was lost in one of the largest underground coal mine
fires in recent Chinese mining history.
ACKNOWLEDGEMENTS
The authors would like to thank the engineering
technicians of Ningxia Mining Industry Group for
their devoted help.
322
 REFERENCES
Michaylov, M. 1995 Preventing and Fighting Spontaneous
Combustion by Foam Pulp in Bovov Dol Coal field. In
A.M. Wala (ed.), Proc. 7th US Mine Ventilation Sympo-
sium, Lexington, Kentucky, June 5–7: 185–190.
Song, Y.M., Liu, Z.J., & Wang, W.S. 2004 Practice of Quick
Drill in Coal Mine Emergency Rescue. China Safety
Science Journal. 14(6): 63–65.
Wang, J.F. & Li, W.J. (eds) 2002 Chapter 5: Coal Mine
Fires. China Coal Mining Accidents, with Special Expert
323
Copyright © 2006 Taylor & Francis Group plc, London, UK
Comments and Analysis, Vol. II. Beijing: Coal Mining
Industry Press. p. 1511.
Wang, D.M. 2004 A new technology of three-phase foam for
mine fire control. Safety in Coal Mine 35(7): 16–18.
Zhou, F.B., Shui, X.Y. & Zhang, Y.J. 2005 New Technol-
ogy of Injecting Large Quantity Three-phase Foam into
Boreholes. In J.H. Fu (ed.), Theory and Practice of Pre-
vention and Control of Coal Mine Gas: 279–283. Xuzhou,
Jiangsu: China University of Mining & Technology Press.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Correlation between methane and fire hazards in abandoned

workings of longwall mining

B. Madeja-Strumińska
Wroclaw Institute of Technology, Department of Mining Engineering, Wroclaw, Poland

E. Widzyk-Capehart
The University of Queensland, CRCMining, Brisbane, Australia

ABSTRACT: In underground coal mines, fire and methane explosion hazards are considered the most perilous.
Those hazards are especially risky when exploitation of methane coal seams susceptible to self-ignition is
undertaken. In this paper, an evaluation of fire and methane explosion hazards based on Graham’s, La Chatelier’s
and oxygen indexes as well as an approximate temperature of burning gases is presented. It is proposed that during
exploitation of methane coal beds susceptible to self-ignition, the probabilities of gas explosion or ignition are
estimated based on systematic sampling of gas composition in abandoned workings. This approach would allow
a numerical value to be assigned to the risk of gas explosion or ignition and would provide basis to undertake
preventive action to reduce or eliminate those hazards.

1 INTRODUCTION conditions. It is also crucial to predict the fire and

Instances of methane burning or explosion occurring
due to spontaneous fires and of coal burning as a result
of methane burning are known throughout the mining
industry (Kabiesz 2002 and Yearly Report 2001). This
indicates that fire and methane hazards are interre-
lated, that is, fire affects the initiation, intensity and
characteristics of methane explosion/ignition and vice
versa. This is especially applicable in longwall mining
operations with high methane content and coal beds
susceptible to self-ignition.
Evaluation of these hazards can be undertaken using
various methods. Most frequently, indexes such as:
increase in carbon monoxide CO, amount of carbon
monoxide V̇CO , Graham’s index G, Young’s YR, respi-
ratory index RQ, Willet’s W, Trickett’s Tr, Bystron’s F
and Morris’M measures are assessed based on compo-
sition of gases in abandoned workings (Bystroń 1997).
Recently, the determination of fire indexes has been
introduced in coal mining, including hydrogen, ethy-
lene and propylene (Yearly Report 2001). All of
those indexes characterize the instantaneous con-
ditions of fire or explosion hazards in abandoned
workings.
In practical application, it is vital to determine how
the fire or explosion will develop in the future and,
especially, what probabilities of explosion or ignition
of gob gases could be experienced in particular mining
methane explosion hazards prior to their occurrence.
In this paper, an assessment of probabilities of fire
or explosion events of mine gases is proposed, based on
systematic measurements of chemical composition of
mine gases. This approach allows for the numerical
estimation of the magnitude of the occurring hazards
with probability values used as basis for preventive
actions. An application of the proposed method to pre-
dict gas explosion and ignition in longwall mining is
also demonstrated.
2 EVALUATION OF FIRE HAZARD IN
ABANDONED WORKINGS OF
LONGWALL CAVING OPERATION
Coal self-ignition depends on coal susceptibility to
self-ignition and ambient temperature in the area
where heat accumulation occurs. Favorable conditions
to coal self-ignition may occur during coal seam
exploitation; those conditions depend on geological
and mining characteristics of the operation. Factors
such as type of mining method, purity of coal extrac-
tion, direction of exploitation, method of isolation
of abandoned workings, moisture content, rock mass
pressure, degree of discontinuities of the coal seam,
rate of ventilation and others influence the coal self-
ignition to various degrees.

325

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Generally, a high probability of spontaneous fires
in abandoned workings exists with currently used
mechanized roof-caving longwall mining methods. In
particular, as illustrated by actual observations, the
existence of methane in abandoned workings is highly
dangerous when exploitation of coal seams highly
susceptible to self-ignition is undertaken. In such con-
ditions, an explosive atmosphere may easily be created
and the developed spontaneous fire may cause an
ignition and explosions of the gases present.
To evaluate hazards due to spontaneous fires,
Graham’s index, G, which is determined based on
gas samples obtained from specified locations in
abandoned workings, is frequently used. According
to several publications (Bystroń 1997, Maciejasz &
Kruk 1977, Strumiński 1996, Strumiński & Madeja-
Strumińska 1997), this index is calculated as:
where CO, N2 , and O2 represent percentages of car-
bon monoxide, nitrogen, and oxygen content in the gas
volume, respectively.
The amount of nitrogen N2 in Equation 1 is usually
calculated from:
where CO2 , CH4 , H2 , Cm Hn , and H2 S represent per-
centages of carbon dioxide, methane, carbons, and
hydrogen sulphide content, respectively.
According to Bystroń (1997) and Strumiński
(1996), the fire hazard criteria based on Graham’s
index are as follows:
• for normal conditions i.e., fire hazard does not
exists:
• for an increase in sampling frequency of air for
chemical analysis:
• for precautionary actions to commence, with
normal operation continued in the region of
investigation:
• for fire fighting action to commence:
If Graham’s index increases to G = 2, coal heating
in abandoned workings is already very advanced, while
for G = 3, the existence of open fire is almost certain.
326
Copyright © 2006 Taylor & Francis Group plc, London, UK
An approximate temperature of the fire center in
old workings T (◦ C) can be calculated as:
where Tp is the temperature of coal self-ignition (start
of fire) and Tp = 300◦ C; Tg is the original temperature
of the rock, ◦ C; and α is a coefficient of proportionality,
1/ppm.
The original temperature of the rock Tg can be
calculated from:
where T0 is the average yearly air temperature in the
region where coal deposit is being mined, ◦ C; H is
the depth of forming fire center, m; H0 is the depth to
which the changes in seasonal temperature reach in the
specified region, m; and m is the average geothermal
rate in the specified region, m/◦ C.
The coefficient of proportionality α, according to
(Strumiński & Madeja-Strumińska 2004) is:
If an open fire exists in abandoned workings, i.e.
when G > 3, an approximate temperature of the fire
center is calculated according to Bystroń (1997) as:
where Gpp is Graham’s index for fully developed
spontaneous fire at a temperature of 1200◦ C. Follow-
ing Bystroń (1997), Gpp is assumed equal to 30 in this
paper.
3 EVALUATION OF LIKELIHOOD OF GAS
EXPLOSION OR IGNITION DUE TO
EXISTING FIRE IN ABANDONED
WORKINGS
Combustible gases, such as methane, carbon monox-
ide, hydrogen, hydrocarbons and hydrogen sulfide are
present in abandoned workings during mining. In par-
ticular, when mining methane coal seams susceptible
to self-ignition, and especially in the areas of forming
fire centers. These gases can ignite or explode when
favorable conditions occur.
Explosion of combustible gases is possible when
the following conditions are present:
• specific concentration of explosive gases
• sufficient amount of oxygen
 • source of high temperature
• sufficient space in abandoned workings for the
explosion to unfold.
In mining practice, the potential for a combustible
gas mixture to explode is generally evaluated based on
Le Chatelier’s index, L, (Strumiński 1996):
It is assumed that the hazard of gas explosion exists
in abandoned workings if the index L is not less than 1.
For a gas mixture to be explosive it must contain
a specific (minimum) concentration of oxygen, which
depends on the concentration of explosive elements in
the gas mixture.
The safe lower limit of oxygen concentration, O2min
in explosive gas mixture below which an explosion
does not occur, can be calculated as (Strumiński 1996):
The danger of a gas explosion exists when L ≥ 1
and when oxygen concentration in the gas mixture is
O2 ≥ O2min.
It is further assumed that a gas explosion may occur
when the temperature is approximately Twg = 600◦ C
and if there is sufficient space in abandoned workings
for the explosion. (Strumiński 1996, Strumiński &
Madeja-Strumińska 1997). If the concentration of
explosive gases is greater than the maximum explosive
limit, then these gases will ignite and non-explosive
burning will occur for the temperature of the fire cen-
ter equal to the temperature of the gas ignition and for
O2 ≥ O2min.
To determine whether gob gases can undergo com-
plete burning, a modified Le Chatelier’s index, L∗ can
be used:
where CH4 , CO, H2 , Cm Hn , and H2 S are as speci-
fied for Equation 2. The values in the denominators
in Equation 13 represent the maximum limits of
explosivity for each listed gaseous element.
Furthermore, it is assumed that the possibility of
gas ignition exists when L∗ ≥ 1 and the danger of gas
explosion if L∗ < 1 or L ≥ 1. It is thus apparent that
even for L∗ ≥ 1, gas ignition does not always occur. Sit-
uations may transpire when the temperature of the fire
center is too low or the oxygen concentration prevents
the self-ignition process of gases.
327
Copyright © 2006 Taylor & Francis Group plc, London, UK
4 EVALUATION OF LIKELIHOOD OF
GAS EXPLOSION OR IGNITION IN
ABANDONED WORKINGS BASED ON
CHEMICAL ANALYSIS
When fire occurs in abandoned workings, the likeli-
hood of gas explosion or ignition can be determined
based on chemical analysis of individual gases in the
gas mixture. Using relationships presented in Section
3 of this paper, it is possible to establish whether the
explosion/ignition can occur or whether those haz-
ards are not present. However, this prediction only
applies to the time when gas sampling is under-
taken. From a practical point of view, an estimation
of the probability of gas explosion/ignition occurring
when all conditions influencing these events are ful-
filled simultaneously is more important. This could
be accomplished by undertaking a systematic exam-
ination of gas composition in specified abandoned
workings. The location of gas sampling stations should
be established to ensure the samples are representative
of the entire area under investigation.
The composition of gases in the abandoned work-
ings depends on many factors, such as state of fire
development, atmospheric pressure, air volume flow-
ing through the mine (ventilation area, footwall) and
others. All these factors have random characteristics
and thus it is assumed that the concentrations of O2 ,
CO, H2 , CH4 , Cm Hn , H2 S and the gas temperature in
the gob are continuous random variables taking on any
value from the specified interval [a,b] (Bystroń 1984,
Madeja-Strumińska 1990). Based on these assump-
tions, all values specified in Section 3 are continuous
random variables X, because they are calculated based
on the composition of gob gases.
The main purpose of this article is to determine the
probability that the random variable X is greater than
the limit value xg , that is, P(X > xg ), using statistical
analysis. This probability can be calculated from:
where f(x) is the probability density function for ran-
dom variable X; and x is the value of the random
variable X.
Equation 14 is equivalent to:
where F(xg ) is the distribution function for random
variable X calculated for X = xg .
In the evaluation of the distribution of the ran-
dom variable X two basic parameters are used; an
expected value E(X) and variance D2 (X). If the num-
ber of observations N = 40÷60 or more then E(X) and
D2 (X) are determined on the basis of distribution series
 and a histogram or graph of the variable X (Madeja-
Strumińska 1990, Strumiński 1996). Based on the
analysis of the histogram or a graph, a hypothesis can
be formulated with respect to the variable distribution
functions; usually those are normal or close to nor-
mal distributions. However, it is important to note that
normal distributions are not always present. Therefore,
statistical verification must be completed to prove the
hypothesis that the random variable X has a normal
distribution (Madeja-Strumińska 1990).
To estimate the probability of gas explosion in aban-
doned mine workings on the basis of systematic studies
of the composition of the gas mixture, a large number
of comprehensive chemical analyses should be con-
ducted. During each analysis, Graham’s index G, Le
Chatelier’s indexes L and L∗ , minimum oxygen con-
centration O2min and temperature of individual gases
in the fire center are calculated.
The distribution and the probability of a random
variable exceeding the limit value are calculated as
follows:
• for index G > 3
• for index L > 1
• for index L∗ > 1
• for minimum oxygen concentration O2 > O2min
where O2min is the highest value of O2min calculated
according to Equation 3 for a specified number of
statistical samples.
The probability of exceeding the ignition tempera-
ture of the gob gases, i.e., 600◦ C, is calculated as:
Field observations confirmed that the described events
can be treated independently. Therefore, the prob-
ability can be assessed as the product of mutually
independent events (Madeja-Strumińska 1990):
In addition, the explosion of gob gases is only
possible when the following conditions are fulfilled:
• spontaneous fire in the gob does not exist
• gob gases are explosive
328
Copyright © 2006 Taylor & Francis Group plc, London, UK
• oxygen concentration in the gas mixture is not less
than the minimum concentration for which the gas
explosion is possible
• gas temperature reaches or exceeds 600◦ C.
Assuming that: A1 = P(G > 3), A2 = P(L > 1),
A3 = P(O2 > O2min ), A4 = P(Tp > 600) and
the probability of gas explosion is calculated as:
To fulfill the set of conditions for gas explosion
to occur, the variables must assumed the following
values: P(G > 3) = 1, P(L > 1) = 1, P(O2 > O2min ) = 1
and P(Tp > 600) = 1.Therefore, according to Equation
22, the explosion of gob gases occurs for P(W) = 1.
On the other hand, the ignition of gases due to
existence of spontaneous fire takes place when:
• fire centers exist in the gob
• concentration of gases exceeds the maximum limit
of explosivity
• concentration of oxygen in the gas mixture is
conducive to gas burning
• temperature in the gob exceeds the ignition temper-
ature of gob gases.
To determine the probability of ignition of gob
gases, the following notations are substituted in
Equation 21: A1 = P(G > 3), A2 = P(L∗ > 1), A3 =
P(O2 > O2min ), A4 = P(Tp > 600) and
The probability of gas ignition is calculated then as:
Ignition of gob gases will take place when all
of the following conditions are met: P(G > 3) = 1,
P(L∗ > 1) = 1, P(O2 > O2min ) = 1 and P(Tp > 600) =
1. Therefore, the probability P(Z) must equal to 1.
Hence, for the probabilities P(W) = 1 and P(Z) = 1,
gob gas explosion or ignition is possible, while for
P(W) = 0 and P(Z) = 0, the risk of ignition or explo-
sion practically does not exist. Consequently, the prob-
abilities calculated according to Equations 22 and 23
can be treated as indexes (measures) of the possibility
of explosion or ignition; the higher the probabili-
ties the higher the risk of explosion/ignition of gob
gases.
 Figure 1. Longwall 608 in methane coal seam 405 at mine S.
It is suggested that a 40% safety margin is used
and thus it is assumes that when P(W) ≥ 0.6 and
P(Z) ≥ 0.6, the risk of explosion or ignition exists.
For probabilities P(W) ≈ 0.4 or P(Z) ≈ 0.4, preventive
action should be taken by conducting neutralization
of the gob atmosphere, reducing air flow to the gob
and/or other means.
5 CASE STUDY
The case study illustrates the applicability of the
presented method to evaluate methane-fire hazards
correlations. Figure 1 shows the abandoned workings
for methane coal seam 405 at mine S. Air samples
were taken from behind the barrier T1 and, using
57 results of chemical analysis, the following val-
ues were obtained: P(G > 3) = 0.90, P(L > 1) = 0.79,
P(O2 > O2 min ) = 0.98, and P(Tp > 600) = 0.70. The
probability of gas explosion was calculated accord-
ing to Equation 22 as P(W) = 0.49. This indicates a
high risk of gas explosion since P(W) > 0.4. Thus,
active preventive measures were undertaken by pump-
ing liquid nitrogen through barrier T1 , which reduced
the probability P(W) to a safe level of 0.26.
329
Copyright © 2006 Taylor & Francis Group plc, London, UK
6 DISCUSSION
Observations made in Section 3 and 4 of this paper are
applicable to situations where methane and explosive
gases are present in the center of the spontaneous fire.
Practical observations show that methane, being the
main element in methane coal seams, where sponta-
neous fires do not exist, can ignite or explode. For
example, as a result of sparking produced by friction
between rock masses during caving operation (Cimr &
Trenczek 2002). In addition, the burning methane may
initiate a fire of the coal remains (Kajdasz & Buchwald
1999, Maciejasz & Kruk 1977. Matuszewski 2002,
Yearly Report 2001).Thus, it is important that the prob-
abilities of explosion and/or ignition of gob gases, even
when the center of the spontaneous fire does not exist,
are estimated. These probabilities can be calculated
based on Equations 22 and 23 after modifications:
– Probability of gas explosion (mainly CH4 )
– Probability of gas ignition
 If caving occurs in a strong rock and sparks are
produced during the caving process, which might
ignite mine gases, a value of P(Tp > 650) = 1, where
Tp = 650◦ C is the temperature of methane ignition,
should be assumed for Equations 24 and 25. To ensure
safety, the probabilities P(W) ≥ 0.6 and P(Z) ≥ 0.6
should also be considered.
7 CONCLUSIONS
In this paper, an evaluation of hazards associated with
explosion and/or ignition of gas mixtures in abandoned
coal workings based on a likelihood of occurrence
method was presented. The probability estimations of
gas explosion/ignition are based on routine yet sys-
tematic examinations and chemical analysis of the
gas mixtures and concentrations of individual gases.
The accuracy of the method depends on the air sam-
ples’ collection stations and the number of samples.
Sampling should be conducted in locations that best
characterize the section of the gob under study and,
in general, the number of statistical samples collected
should be greater than forty.
The likelihood evaluation method for explosion/
ignition of gases in abandoned workings based on
chemical analysis was applied for a real mining condi-
tion. The method proved beneficial in preventing these
events from occurring by indicating when preventive
actions should be taken to minimize the risks.
REFERENCES
Bystroń, H. 1984. Active fire suppression and elimination of
fire field. Mining Review No. 12.
Bystroń, H. 1997. Fire status, ignition and explosion of
gases during active and passive fire liquidation. Mining
Archives, Vol. 4, Series 1.
330
Copyright © 2006 Taylor & Francis Group plc, London, UK
Cimr, A. & Trenczek, S. 2002. Risk of methane explosion
due to sparks induced by rock friction during longwall
caving operation. 2nd School of Mining Aerology, Mining
Aerology Section. Zakopane, Poland: Mining Committee
of the Polish Academy of Science.
Kabiesz, J. 2002. Evaluation of associated mining haz-
ards and selection of preventive action. Science Series
Studies/Thesis/Monographs, No. 849. Katowice, Poland:
Central Mining Institute.
Kajdasz, Z. & Buchwald, P. 1999. Use of nitrogen in elim-
ination of fire hazards in gobs of active excavations and
innovative usage of nitrogen as an inert gas during fire
action. 1st School of Mining Aerology, Mining Aerol-
ogy Section. Zakopane, Poland: Mining Committee of the
Polish Academy of Science.
Maciejasz, Z. & Kruk, F. 1977. Underground mine fires.
Vol. 1. Katowice, Poland: Slask.
Matuszewski, K. 2002. Preventive actions against methane
initiation in underground polish coal mines. 23rd Tech-
nical Days ROP 2002 and 19th Seminar on Methane
Hazards and Coal Dust Explosion in Underground Coal
Mines – Theory and Practice. Rybnik, Poland.
Madeja-Strumińska, B. 1990. Evaluation of underground
mine ventilation. Scientific Works of Mining Institute
of Wroclaw Polytechnic. Monograph No. 25. Wroclaw,
Poland: Wroclaw Polytechnic.
Strumiński, A. 1996. Elimination of fires in underground
mines. Katowice, Poland: Slask.
Strumiński, A. & Madeja-Strumińska, B. 1997. Evaluation
and elimination of spontaneous fire hazards in under-
ground mines. Wroclaw, Poland: Lower Silesia Education
Publishing House.
Strumiński, A. & Madeja-Strumińska, B. 2004. Temperature
evaluation at the center of spontaneous fires. 30th Techni-
cal Days ROP 2004 & 21st Seminar on Elimination of Fire
Hazards in Mines – Theory and Practice. Rybnik, Poland.
Yearly Report. 2001. Status of natural and technical basic haz-
ards in underground mining. In W. Konopko (ed.), Yearly
Report. Katowice, Poland: Central Mining Institute &
Main Center for Mining Safety.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Sealing a recent United States coal mine longwall gob fire

R.T. Stoltz & W.J. Francart

US Department of Labor, Pittsburgh, PA, USA

L. Adair & J. Lewis

Genwal Resources, Inc., Price, UT, USA

ABSTRACT: On February 20, 2004, a spontaneous combustion fire was identified in the active longwall gob
at the West Ridge underground coal mine. Control of the spontaneous combustion event, along with the safe
removal of the longwall equipment, were the priorities of the mine fire control and recovery operation. The
longwall face was advanced approximately 6 m to enable removal of face equipment, while ventilation rates
and airflow patterns in and around the gob were maintained as much as possible. While recovery operations
continued, air samples were collected for analysis by gas chromatograph. Final sealing and pressure-balancing
of the area required coordination of teams of miners completing seventy-six seals while injecting carbon dioxide
into the gob. This paper discusses the details concerning the fire, gas sampling, seal construction and ventilation
adjustments, which enabled the mine to control the fire and resume coal production in 24 days after its discovery.

1 INTRODUCTION growth rate of the fire.Although no visible flames were

The West Ridge Mine is operated by West Ridge
Resources, Incorporated. The mine produces bitumi-
nous coal at a level of approximately 2.7 Mtpa. The
mine, opened in 1999, was developed from four drift
openings into the Lower Sunnyside coal seam. The
main entries dip 10–15% percent causing the inby end
of the mains to be approximately 305 m lower in eleva-
tion. The mine developed longwall panels for second
mining of the seam. Due to ground control concerns,
the panels were developed using a two-entry mining
system. Figure 1 represents the extent of mining at the
time the fire was discovered.
encountered, the presence of fire gases indicated a low-
level combustion was occurring.
Based upon underground examination and gas
samples collected along the perimeter of the gob, the
suspected fire area was identified as being in the
vicinity of the back of the 4th Right longwall gob.
Figure 2 shows the suspected location of the heating
area. Hydrogen, carbon monoxide, ethylene and other
fire gases were being produced within the gob and
were measured in samples collected from the 1st Right
sampling locations.
3 FIRE FIGHTING

2 FIRE DETECTION
On February 20, 2004, elevated levels of carbon
monoxide (160 ppm) were measured in air also con-
taining traces of hydrogen at a bleeder evaluation point.
MSHA’s Coal Mine Safety and Health district office
was contacted to report the findings, and production
at the mine ceased.
MSHATechnical Support personnel were contacted
and directed to assist all involved parties in response to
the fire. In addition to supplying personnel to the mine
site, a gas chromatograph was delivered to analyze gas
samples from the underground atmosphere. Technical
Support personnel used the sample analyses to assist
with trying to determine the location, the extent and
The mine, primarily the longwall gob, liberated a sig-
nificant methane volume (70 000 m3 /day). Due to the
unknown location of the heating/fire (speculated in the
4th Right longwall gob) and the amount of methane
produced, typical direct fire fighting activities were
not a viable option. West Ridge and MSHA personnel
cooperatively worked to develop plans to seal the gob.
Three sealing plans were developed and each was
prepared to be implemented depending on the fire’s
status. The plans consisted of sealing the entire mine,
sealing the longwall gob at the mouth of the panels,
which would have included the longwall equipment,
or removing the longwall equipment and sealing along
1st Right and at the mouth and in by ends of 1st Right
through 5th Right.

331

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 1. West Elk mine development.

Figure 2. Area of suspected spontaneous combustion.

332

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 3. Locations for seal construction.
As long as the fire did not increase in intensity, it
was decided to allow implementation of the third plan,
which entailed removal of the longwall equipment and
sealing the perimeter of the gob. The longwall face
needed to be mined approximately 6 m to allow for
screening the roof at the face, which was necessary to
recover the face equipment. Sealing activities would
start, but ventilation was to be maintained to the extent
possible. To maintain pre-sealing ventilation patterns
along the gob perimeter, 0.2 m diameter steel pipe(s)
or 0.8 m door frames were designed, approved and
inserted into seals during their construction. To permit
eventual sealing, the pipes were prepared to accept
vitaulic caps or fitted with valves. The door frames
were prepared to accept covers which could be bolted
into place to cover the opening.
Seal construction began on February 26 and was
completed on March 7 except for the temporary seals
in 5th Right. Approximate seal locations are shown
in Figure 3. Temporary seals were constructed at the
mouth of 5th Right. These temporary seals would be
upgraded with permanent explosion-resistant seals to
later be built on the outby side.
A total of 76 explosion-resistant seals were built
around the active gob, utilizing three approved seal
designs. One design was a cementitious-foam seal.
This seal consisted of an inby and outby temporary
333
Copyright © 2006 Taylor & Francis Group plc, London, UK
stopping wall, approximately 1.2 m apart. A cemen-
titious foam mixture was then pumped into the con-
structed void through at least three injection ports
installed into this form. No ventilation devices, pipes
or door frames, could be installed in this seal design.
Eighteen of these seals were constructed; six had been
completed prior to the spontaneous combustion event.
A second approved seal design was a polyurethane-
gravel core seal, formed by constructing two surface
bonded, dry-stacked solid or hollow core concrete
block walls, hitched into the ribs and bottom. A poly-
urethane and limestone core thickness of at least 0.4 m
is pumped into the void to complete the seal. Twenty
of these seals were constructed.
The third approved seal design was constructed sim-
ilar in manner to a solid concrete block seal except that
instead of hitching into the ribs and bottom, inflatable
bags are installed along each rib and along the top of
the seal. These inflated bags preload the seals, which
are constructed of 15 or 20 cm solid concrete block.
Thirty-five of these seals were constructed.
Three different seal designs were utilized to facil-
itate completion of the project. No one single seal
construction method could be implemented due to the
contractor and seal supply limitations. After complet-
ing removal of all of the longwall face equipment, the
final temporary seal was completed and all personnel
 were withdrawn from the mine for a 72-hour waiting
period.
As is the custom in many sealing operations, a
waiting period is observed following completion of
the final temporary seal to allow the sealed area to
stabilize. A 72-hour period is commonly observed.
During this time, atmospheric samples were collected
from three underground locations from the surface
to evaluate the conditions underground. Methane lev-
els began to increase as would be expected due to
the methane liberation, estimated to be 70 000 m3 /day.
After 72 hours the conditions would be re-evaluated to
determine if re-entry into the mine would be permitted.
Temporary sealing was completed by closing the
fire door in the mouth of 5th Right (temporary seal)
and closing all openings installed in the seals along
the perimeter of the gob. The closing of these openings
consisted of placing the cover over the 0.76 m frame
opening and securing it, or closing the victalic valves
or caps in the pipes.
The original ventilation system moved intake air
from the mains to the back bleeder and return. The
existing pressure differential was significant, and the
company decided to adjust the ventilation system to
provide intake air splits on both the mains and bleeder
sides of the gob. This in theory would balance the
pressure differential across the sealed area and limit
leakage into the fire area. Limiting the leakage would
in turn allow faster reductions in oxygen levels by
increasing methane concentrations. Following com-
pletion of the seals and modifications to the ventilation
system, initial measurements of the pressure drop from
the intake into the sealed area from the back bleeder
were approximately 90 Pa.
The heating/fire never significantly grew in inten-
sity throughout the sealing period. The gas analyses of
the mine atmosphere and statistical evaluation showed
no significant ascending overall trends, consistent with
the steady to descending carbon monoxide trends dep-
icted for three of the seven monitoring points shown
in Figure 4.
Figure 4. Trends for CO concentrations.
334
Copyright © 2006 Taylor & Francis Group plc, London, UK
4 RE-ENTRY
After the 72 hour waiting period, samples collected
from the mouth of the 4th Right panel indicated the
atmosphere was explosive (Figure 5).
Based on sample trends for fire gases and calcu-
lated indices, it was determined that re-entry into the
1st Right entries only would be permitted in order to
remotely collect samples from bleeder connectors to
evaluate conditions at locations on the inby end of the
gob. Prior to sealing, sample lines were installed from
these bleeder connectors to a location in 1st Right to
permit remote sampling of these locations. When ana-
lyzed, these samples indicated nearly fresh air. An
evaluation of pressure drops on the seals indicated a
pressure drop of 348 Pa into the back end of the gob
from the bleeder. This pressure was measured at 90 Pa
at the time temporary seals were completed. The pres-
sure drop was expected due to the air flow within the
bleeder system. Even though the air was an intake split,
the pressure loss within the bleeder would create a
significant pressure drop for the parallel split through
the gob.
As a result of sample analyses, it was determined
that methane liberation would not inert the entire
sealed area atmosphere due to the existing pressure
differential. A plan was developed to inject carbon
dioxide into the gob from the outby side in 5th Right
through a two-inch fresh water supply line while
the remaining permanent seal was completed under-
ground. In addition, pressure drops around the sealed
area would be balanced to reduce the leakage of
fresh air into the sealed area. Pressure differentials
Figure 5. Explosibility curve following seal completion.
 were reduced from ingassing 348 Pa to less than 75 Pa
ingassing.
Carbon dioxide was injected at a rate of approxi-
mately 42 m3 /min from tankers for a period of seven
and one-half hours. During this time a sediment trap
was installed in the No. 1 entry, and the permanent
seal location on the outby side was prepared. While
the final seal was being constructed outby the tempo-
rary seals, carbon dioxide injection was continued at
a rate of approximately 18.8 m3 /min for eleven and
one-half hours.The purpose of the injection was to pro-
vide an added measure of safety for miners by reducing
oxygen in the headgate entries and adjoining areas in
the gob. The project was completed on March 15 at
approximately 1:45 pm.
5 CONCLUSIONS
Control of the spontaneous combustion event at the
West Ridge mine was accomplished by constructing
seals around the entire gob. As seals were constructed,
atmospheric monitoring locations were established to
335
Copyright © 2006 Taylor & Francis Group plc, London, UK
allow determination of the fire growth rate and
explosibility of the atmosphere. Ventilation through
the gob was maintained to the extent possible while
seals were constructed to avoid methane accumula-
tions. Doors and pipes were installed in key seals to
facilitate continued ventilation. Following construc-
tion of the final temporary seals, adjustments to the
ventilation around the sealed area were needed to
balance pressures across the gob, thereby reducing
leakage into the gob and allowing the gob to inert.
Based on completion of the permanent seals, the
operation was completed and production resumed in
24 days following the discovery of the elevated carbon
monoxide levels.
Pressure differentials across sealed areas have often
been the most difficult obstacle for inerting gob areas.
The West Ridge fire was no exception. Pressure differ-
entials can induce considerable airflow through sealed
areas, as seals are not air-tight ventilation controls.
This project reminds us that accounting for pressure
balancing across the gob is critical for ultimate success
in extinguishing similar fires.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Sealing a recent United States coal mine longwall gob fire

R.T. Stoltz & W.J. Francart

US Department of Labor, Pittsburgh, PA, USA

L. Adair & J. Lewis

Genwal Resources, Inc., Price, UT, USA

ABSTRACT: On February 20, 2004, a spontaneous combustion fire was identified in the active longwall gob
at the West Ridge underground coal mine. Control of the spontaneous combustion event, along with the safe
removal of the longwall equipment, were the priorities of the mine fire control and recovery operation. The
longwall face was advanced approximately 6 m to enable removal of face equipment, while ventilation rates
and airflow patterns in and around the gob were maintained as much as possible. While recovery operations
continued, air samples were collected for analysis by gas chromatograph. Final sealing and pressure-balancing
of the area required coordination of teams of miners completing seventy-six seals while injecting carbon dioxide
into the gob. This paper discusses the details concerning the fire, gas sampling, seal construction and ventilation
adjustments, which enabled the mine to control the fire and resume coal production in 24 days after its discovery.

1 INTRODUCTION growth rate of the fire.Although no visible flames were

The West Ridge Mine is operated by West Ridge
Resources, Incorporated. The mine produces bitumi-
nous coal at a level of approximately 2.7 Mtpa. The
mine, opened in 1999, was developed from four drift
openings into the Lower Sunnyside coal seam. The
main entries dip 10–15% percent causing the inby end
of the mains to be approximately 305 m lower in eleva-
tion. The mine developed longwall panels for second
mining of the seam. Due to ground control concerns,
the panels were developed using a two-entry mining
system. Figure 1 represents the extent of mining at the
time the fire was discovered.
encountered, the presence of fire gases indicated a low-
level combustion was occurring.
Based upon underground examination and gas
samples collected along the perimeter of the gob, the
suspected fire area was identified as being in the
vicinity of the back of the 4th Right longwall gob.
Figure 2 shows the suspected location of the heating
area. Hydrogen, carbon monoxide, ethylene and other
fire gases were being produced within the gob and
were measured in samples collected from the 1st Right
sampling locations.
3 FIRE FIGHTING

2 FIRE DETECTION
On February 20, 2004, elevated levels of carbon
monoxide (160 ppm) were measured in air also con-
taining traces of hydrogen at a bleeder evaluation point.
MSHA’s Coal Mine Safety and Health district office
was contacted to report the findings, and production
at the mine ceased.
MSHATechnical Support personnel were contacted
and directed to assist all involved parties in response to
the fire. In addition to supplying personnel to the mine
site, a gas chromatograph was delivered to analyze gas
samples from the underground atmosphere. Technical
Support personnel used the sample analyses to assist
with trying to determine the location, the extent and
The mine, primarily the longwall gob, liberated a sig-
nificant methane volume (70 000 m3 /day). Due to the
unknown location of the heating/fire (speculated in the
4th Right longwall gob) and the amount of methane
produced, typical direct fire fighting activities were
not a viable option. West Ridge and MSHA personnel
cooperatively worked to develop plans to seal the gob.
Three sealing plans were developed and each was
prepared to be implemented depending on the fire’s
status. The plans consisted of sealing the entire mine,
sealing the longwall gob at the mouth of the panels,
which would have included the longwall equipment,
or removing the longwall equipment and sealing along
1st Right and at the mouth and in by ends of 1st Right
through 5th Right.

331

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 1. West Elk mine development.

Figure 2. Area of suspected spontaneous combustion.

332

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 3. Locations for seal construction.
As long as the fire did not increase in intensity, it
was decided to allow implementation of the third plan,
which entailed removal of the longwall equipment and
sealing the perimeter of the gob. The longwall face
needed to be mined approximately 6 m to allow for
screening the roof at the face, which was necessary to
recover the face equipment. Sealing activities would
start, but ventilation was to be maintained to the extent
possible. To maintain pre-sealing ventilation patterns
along the gob perimeter, 0.2 m diameter steel pipe(s)
or 0.8 m door frames were designed, approved and
inserted into seals during their construction. To permit
eventual sealing, the pipes were prepared to accept
vitaulic caps or fitted with valves. The door frames
were prepared to accept covers which could be bolted
into place to cover the opening.
Seal construction began on February 26 and was
completed on March 7 except for the temporary seals
in 5th Right. Approximate seal locations are shown
in Figure 3. Temporary seals were constructed at the
mouth of 5th Right. These temporary seals would be
upgraded with permanent explosion-resistant seals to
later be built on the outby side.
A total of 76 explosion-resistant seals were built
around the active gob, utilizing three approved seal
designs. One design was a cementitious-foam seal.
This seal consisted of an inby and outby temporary
333
Copyright © 2006 Taylor & Francis Group plc, London, UK
stopping wall, approximately 1.2 m apart. A cemen-
titious foam mixture was then pumped into the con-
structed void through at least three injection ports
installed into this form. No ventilation devices, pipes
or door frames, could be installed in this seal design.
Eighteen of these seals were constructed; six had been
completed prior to the spontaneous combustion event.
A second approved seal design was a polyurethane-
gravel core seal, formed by constructing two surface
bonded, dry-stacked solid or hollow core concrete
block walls, hitched into the ribs and bottom. A poly-
urethane and limestone core thickness of at least 0.4 m
is pumped into the void to complete the seal. Twenty
of these seals were constructed.
The third approved seal design was constructed sim-
ilar in manner to a solid concrete block seal except that
instead of hitching into the ribs and bottom, inflatable
bags are installed along each rib and along the top of
the seal. These inflated bags preload the seals, which
are constructed of 15 or 20 cm solid concrete block.
Thirty-five of these seals were constructed.
Three different seal designs were utilized to facil-
itate completion of the project. No one single seal
construction method could be implemented due to the
contractor and seal supply limitations. After complet-
ing removal of all of the longwall face equipment, the
final temporary seal was completed and all personnel
 were withdrawn from the mine for a 72-hour waiting
period.
As is the custom in many sealing operations, a
waiting period is observed following completion of
the final temporary seal to allow the sealed area to
stabilize. A 72-hour period is commonly observed.
During this time, atmospheric samples were collected
from three underground locations from the surface
to evaluate the conditions underground. Methane lev-
els began to increase as would be expected due to
the methane liberation, estimated to be 70 000 m3 /day.
After 72 hours the conditions would be re-evaluated to
determine if re-entry into the mine would be permitted.
Temporary sealing was completed by closing the
fire door in the mouth of 5th Right (temporary seal)
and closing all openings installed in the seals along
the perimeter of the gob. The closing of these openings
consisted of placing the cover over the 0.76 m frame
opening and securing it, or closing the victalic valves
or caps in the pipes.
The original ventilation system moved intake air
from the mains to the back bleeder and return. The
existing pressure differential was significant, and the
company decided to adjust the ventilation system to
provide intake air splits on both the mains and bleeder
sides of the gob. This in theory would balance the
pressure differential across the sealed area and limit
leakage into the fire area. Limiting the leakage would
in turn allow faster reductions in oxygen levels by
increasing methane concentrations. Following com-
pletion of the seals and modifications to the ventilation
system, initial measurements of the pressure drop from
the intake into the sealed area from the back bleeder
were approximately 90 Pa.
The heating/fire never significantly grew in inten-
sity throughout the sealing period. The gas analyses of
the mine atmosphere and statistical evaluation showed
no significant ascending overall trends, consistent with
the steady to descending carbon monoxide trends dep-
icted for three of the seven monitoring points shown
in Figure 4.
Figure 4. Trends for CO concentrations.
334
Copyright © 2006 Taylor & Francis Group plc, London, UK
4 RE-ENTRY
After the 72 hour waiting period, samples collected
from the mouth of the 4th Right panel indicated the
atmosphere was explosive (Figure 5).
Based on sample trends for fire gases and calcu-
lated indices, it was determined that re-entry into the
1st Right entries only would be permitted in order to
remotely collect samples from bleeder connectors to
evaluate conditions at locations on the inby end of the
gob. Prior to sealing, sample lines were installed from
these bleeder connectors to a location in 1st Right to
permit remote sampling of these locations. When ana-
lyzed, these samples indicated nearly fresh air. An
evaluation of pressure drops on the seals indicated a
pressure drop of 348 Pa into the back end of the gob
from the bleeder. This pressure was measured at 90 Pa
at the time temporary seals were completed. The pres-
sure drop was expected due to the air flow within the
bleeder system. Even though the air was an intake split,
the pressure loss within the bleeder would create a
significant pressure drop for the parallel split through
the gob.
As a result of sample analyses, it was determined
that methane liberation would not inert the entire
sealed area atmosphere due to the existing pressure
differential. A plan was developed to inject carbon
dioxide into the gob from the outby side in 5th Right
through a two-inch fresh water supply line while
the remaining permanent seal was completed under-
ground. In addition, pressure drops around the sealed
area would be balanced to reduce the leakage of
fresh air into the sealed area. Pressure differentials
Figure 5. Explosibility curve following seal completion.
 were reduced from ingassing 348 Pa to less than 75 Pa
ingassing.
Carbon dioxide was injected at a rate of approxi-
mately 42 m3 /min from tankers for a period of seven
and one-half hours. During this time a sediment trap
was installed in the No. 1 entry, and the permanent
seal location on the outby side was prepared. While
the final seal was being constructed outby the tempo-
rary seals, carbon dioxide injection was continued at
a rate of approximately 18.8 m3 /min for eleven and
one-half hours.The purpose of the injection was to pro-
vide an added measure of safety for miners by reducing
oxygen in the headgate entries and adjoining areas in
the gob. The project was completed on March 15 at
approximately 1:45 pm.
5 CONCLUSIONS
Control of the spontaneous combustion event at the
West Ridge mine was accomplished by constructing
seals around the entire gob. As seals were constructed,
atmospheric monitoring locations were established to
335
Copyright © 2006 Taylor & Francis Group plc, London, UK
allow determination of the fire growth rate and
explosibility of the atmosphere. Ventilation through
the gob was maintained to the extent possible while
seals were constructed to avoid methane accumula-
tions. Doors and pipes were installed in key seals to
facilitate continued ventilation. Following construc-
tion of the final temporary seals, adjustments to the
ventilation around the sealed area were needed to
balance pressures across the gob, thereby reducing
leakage into the gob and allowing the gob to inert.
Based on completion of the permanent seals, the
operation was completed and production resumed in
24 days following the discovery of the elevated carbon
monoxide levels.
Pressure differentials across sealed areas have often
been the most difficult obstacle for inerting gob areas.
The West Ridge fire was no exception. Pressure differ-
entials can induce considerable airflow through sealed
areas, as seals are not air-tight ventilation controls.
This project reminds us that accounting for pressure
balancing across the gob is critical for ultimate success
in extinguishing similar fires.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

The successful recovery of the Dotiki Mine after a major mine fire

C.R. Wesley III & T.M. Wynne

Alliance Resource Partners, L.P., Lexington, KY, USA

J.E. Urosek & K.S. Diederich

Mine Safety and Health Administration, Pittsburgh, PA, USA

ABSTRACT: When a fire occurred on a diesel-powered supply tractor, the operation of the Dotiki Mine was
in jeopardy. Direct attack of the fire was unsuccessful, and the mine was evacuated. Sealing became the only
viable option for controlling the fire. Due to the extent of the mine development, sealing at the surface mine
openings would have required a long atmosphere-inerting period, as well as an extensive recovery operation after
breaching the seals. A plan was developed to install remote seals from the surface at strategic locations which,
in conjunction with existing ventilation controls, provided a means for confining inert gas injected into the fire
area. The success of the operation is credited to the implementation of newly developed seal technology and the
cooperative efforts of all parties involved in fighting the fire. Proper planning and execution of the plan were
key components to resuming production 26 days after the fire started.

1 INTRODUCTION miners who were underground when the fire started

Dotiki Mine, operated by Webster County Coal, LLC,
a subsidiary of Alliance Resource Partners, L.P., was
located near Nebo, Kentucky in the United States
of America. Dotiki Mine employed 380 miners and
used six mechanized mining units (MMUs) to produce
approximately 4.5 million tonnes of coal annually. The
mine had three air shafts and one dual compartment
shaft (belt conveyor/track) that penetrated the Ken-
tucky 9 coal seam. The coal seam averaged 1.5 meters
in thickness and was 180 meters deep. Dotiki Mine,
ventilated with two exhausting fans and one blowing
fan, liberated approximately 13,600 cubic meters of
methane in a 24-hour period. The Dotiki IV shaft was
the then-current main portal.
The six MMUs, each having two continuous min-
ing machines, operated two production shifts per day
five days per week. Coal was transported from the
MMUs to the surface by a belt conveyor system.
Diesel-powered equipment was the primary mode of
transportation for miners and supplies. Production
averaged about 22,000 tonnes per day.
2 FIRE DETECTION
On February 11, 2004, at 4:30 am, a fire was reported
on a diesel-powered supply tractor in a major inter-
section approximately 800 meters from the bottom
of the Dotiki IV intake and return shaft. The fire
quickly spread to the surrounding coal pillars. Seventy
were safely evacuated in less than an hour. Company
officials on site and those traveling to the mine from
home for the morning shift were notified of the fire.
Efforts to extinguish the fire directly were unsuccess-
ful. Thirteen hours after the discovery of the fire it was
determined that all of the active areas of the mine had to
be sealed. Even before the area was completely sealed,
company officials were already reviewing options to
recover the mine back into full production.
3 EMERGENCY PLAN IMPLEMENTATION
3.1 Notification of regulatory agencies
Notifications were made to the local offices of the
KDMM and MSHA. Personnel from each of the agen-
cies immediately responded to the mine to provide
assistance. Corporate personnel were notified and
dispatched to the scene. A command center was estab-
lished in the mine’s engineering office. Contact was
made with the owner of the surface property over the
fire area, and permissions to drill boreholes into the
mine were obtained. Drilling contractors were con-
tacted and the first two of many boreholes were marked
on the map. The first two boreholes were projected to
intersect the mine very near the fire area.
3.2 Direct firefighting
Personnel on site at the fire attempted to extinguish the
fire directly with rock dust, portable fire extinguishers,

337

Copyright © 2006 Taylor & Francis Group plc, London, UK

 and water. As additional personnel arrived to help fight
the fire, foam generating machines were put into use.
Dotiki officials estimated that 2 kiloliters (530 gallons)
of water and 10 liters (2.5 gallons) of foam were being
used every minute. Firefighters wearing self-contained
breathing apparatus approached the original fire area
using foam generators and fire hoses. Doors in stop-
pings were opened to the return in order to reduce
the amount of smoke in the intake entries. Firefighters
were able to reach the diesel tractor that had been the
starting point of the fire, but the surrounding coal was
now the primary fuel source.
In order to evaluate the safety of the miners fight-
ing the fire, air quality readings were taken downwind
of the fire. Handheld gas detectors were used to eval-
uate the gases traveling from the fire in the return air-
course. Methane (CH4 ) readings were noted as varying
from 0.8% to 1.2% and carbon monoxide (CO) and car-
bon dioxide (CO2 ) were either over range or unavail-
able. The exhaust of the nearby Dotiki IV fan was also
monitored. At 1:18 pm, MSHA personnel arrived with
electro-chemical and infrared gas detection instru-
ments and began continuous monitoring the Dotiki IV
Fan exhaust for CH4 , oxygen (O2 ), CO, and CO2 .
Smoke and high temperatures hampered direct fire
fighting efforts. The gas readings at the Dotiki IV Fan
exhaust continued to rise and by 6:00 pm, had reg-
istered as high as 2% CH4 and 2000 ppm CO. Since
the CH4 and CO were climbing and the fire had not
been brought under control, instructions came from
the command center for fire fighting crews to evacuate
the mine.
3.3 Planning for sealing
During the initial hours of firefighting, KDMM,
MSHA, and company officials decided that prepa-
rations should be made to seal the mine if the fire
could not be controlled. Both underground sealing
and surface sealing were considered. The mine had
been developed 16 km from the slope to the farthest
working section. Sealing underground could not be
readily accomplished near the fire area. The best loca-
tion to seal off the active portion of the mine was
between Dotiki III and Dotiki IV shafts in the life of
mine (LOM) entries. If the mine was sealed at the sur-
face, the mine’s atmosphere would inert very slowly
which would extend the recovery process. Alternate
plans were developed, which included drilling into the
mine and injecting water or inert gasses to extinguish
the fire.
3.4 Mine sealing
After the direct fire fighting efforts were abandoned,
seal locations had to be determined. The decision was
made by the command center to seal Dotiki IV intake
and return shaft on the surface and underground at a
point between the Dotiki III and Dotiki IV portals in the
LOM entries. The underground personnel were given
the general location of the seals they were to build,
but also the latitude to select the best location for the
seals based on underground conditions. The selected
location was confirmed by the command center at
11:39 pm. The 16 seals were constructed by 3:33 am
on February 12, day 2. Figure 1 shows the Dotiki Mine

Figure 1. Dotiki Mine complex.

338

Copyright © 2006 Taylor & Francis Group plc, London, UK

 complex with the slope, shafts, sections, fire area, and
the location of the LOM seals used to isolate the active
area of the mine.
Within 30 minutes of the completion of the LOM
seals between Dotiki III and Dotiki IV, all personnel
were evacuated from the mine. After confirmation of
the complete evacuation, the command center ordered
the Dotiki IV fan to be de-energized. Samples lines
were extended down both the intake and return side
of the Dotiki IV shaft and plumbed back to a pump-
ing station a safe distance from the shaft. The cage at
Dotiki IV was positioned such that it could act as part
of the seal. Steel plates and plywood were placed on
the cage and the entire area was sprayed with expand-
ing foam to form the seal on the intake side of the
shaft. Anti-reverse doors inside the fan housing, the
return side, were braced and foamed. The sealing of
the Dotiki IV shaft was completed by 5:15 a.m. For 72
hours, only those persons necessary for sampling were
allowed within 150 meters of the Dotiki IV shaft.
3.5 Inert gas injection
Fortunately, the surface area over the fire was a cul-
tivated field that was readily accessible. A road was
constructed and the first two boreholes were posi-
tioned and drilled near the original fire area. As the
holes were being drilled, there was discussion in the
command center as to how to utilize them upon com-
pletion. Water injection was considered, but ruled out
since the topography of the mine would allow the water
to run away. Several inert gasses were considered, car-
bon dioxide, nitrogen, and argon. Carbon dioxide was
used, but nitrogen was the primary inert gas injected.
4 REMOTE SEALING
4.1 Utilizing existing ventilation controls
A remote sealing procedure was adopted and approved
as the plan of attack to reduce the size of the sealed
area. Arrangements were made with Halliburton
Energy Services to install the remote seals. The remote
seals were commonly referred to as Halliburton seals.
The fire area was at an intersection of five sets
of entries. Initial plans included five possible sets
of remote seals to surround the fire area. However,
through strategic planning of the seal locations and
the use of existing ventilation controls the final seal
plan required three sets of remote seals. On the east
side of the fire area, using existing ventilation controls
helped to reduce the number of remote seals to four.
Nineteen remote seals were required to surround the
fire area. All of the potential drill holes were named
based on their location. For example SE27, was the
27th seal in the “E” group of seals. Since there were
four different proposed locations for the “E” group of
339
Copyright © 2006 Taylor & Francis Group plc, London, UK
seals that crossed 7 entries, there were 28 potential
drill holes named in the “E” seals. As the boreholes
were drilled and while the seals were being pumped,
they were referred to by these names. After the remote
seal installation, the seal groups were simply referred
to as the North, West, and East seals.
The drilling and pumping of the remote seals took
8 days. In addition to drilling the holes that were to
be used for remote seals, observation holes and moni-
toring holes were drilled. Video cameras were lowered
into the boreholes to evaluate the mine conditions for
fire damage, evaluate the air flow patterns, determine
if the remote seal holes were in the center of the entry,
and to observe a remote seal installation. Three mon-
itoring holes, MI 1, MI 2, and MI 3, were drilled on
three sides of the fire area, but within the boundaries
of the remote seals. Sampling lines were run from the
three monitoring boreholes to MSHA’s gas monitoring
truck for continuous monitoring through the end of
the recovery. After the permanent seals were built and
the mine emergency personnel departed, these holes
became sample points to evaluate the conditions within
the sealed area.
4.2 Command center activities
While the remote seals were being installed, the com-
mand center personnel continued to evaluate and con-
sider modifications to the approved plan. The company
always had one person in the command center des-
ignated to handle the information flow. This person
would communicate with all of the parties that were
to report information to the command center. In order
to keep everyone informed, dry erase boards were uti-
lized to post the status of the drilling, remote seals
pumped, loads of nitrogen and carbon dioxide injected,
and which downhole camera videos were available.
Having this information readily available at a glance
allowed the person in charge of communication from
the command center to concentrate on the job. Com-
puter network specialists also worked on connecting
the command center operations to the MSHA gas
analysis computers to more readily transfer informa-
tion. Mine atmosphere samples were collected from
the boreholes and analyzed by gas chromatography.
MSHA then used the results of the gas analyses to
determine fire gas trends. Mine ventilation simulations
were produced to plan the best method to re-ventilate
and re-enter the mine. Natural ventilation currents
develop within the sealed area due to ventilating pres-
sures developed by the heat of the fire. Inerts were
still being injected into the mine since these currents
continue to deliver oxygen to the fire area. Mine atmo-
sphere sampling was done through boreholes on all
sides of the fire area. Samples were collected for gas
chromatograph analysis and instantaneous readings
were taken using handheld gas detectors. Oxygen
 Figure 2. Halliburton seal after foam applied.
readings on the handheld gas detectors were indicators
of the inert gas flowing through the mine.
5 RE-ENTRY AND TEMPORARY SEALS
After the Halliburton seals were all in place and gas
sampling points were established on both sides of
the seal lines, the re-entry plan was approved and
adopted. The seals in the LOM entries between Dotiki
III and Dotiki IV were breached. At the same time,
the intake side of Dotiki IV shaft was opened. Venti-
lation simulations confirmed that the mine could be
adequately ventilated for recovery efforts, without use
of the Dotiki IV mine fan. After the hoist cables were
inspected and approved, rescue teams established a
fresh air base at the bottom of Dotiki IV shaft. Rescue
teams advanced along the mains to the North remote
seals. The teams directed intake air with them to a point
inby where the permanent seals were to be constructed.
Between each remote seal and future permanent seal
a substantial seal was constructed using posts, boards,
and canvas secured to the roof, ribs and floor with
nails from Hilti guns. These substantial seals were
commonly referred to as Hilti seals.
It was suspected that the Halliburton seals would not
completely fill the mine entry in all cases. The remote
seals were considered as regulators designed to restrict
air flow, and not as permanent seals. Their mission was
to slow the exchange of inerted air around the fire area
with oxygen rich air in a ventilated mine. The remote
seal at location DF4 was suspected to be the least
effective since it took more material than expected.
340
Copyright © 2006 Taylor & Francis Group plc, London, UK
Rescue teams estimated that it blocked about 75% of
the entry. Figure 2 shows one of the remote seals. The
material from the remote seal slopes upward at about
30 degrees. The upper corners of the entry had been
sealed off with expanding foam.
After the completion of the north Hilti seals, the
teams advanced to the west side of the fire area.
The east seals were the last to be approached. While
approaching the seals to the east of the fire area, the
teams encountered water that had to be pumped in
order to proceed. The water was pumped into a pre-
viously sealed area father to the east. The Hilti seals
were completed on all three sides of the fire in less
than two days.
6 PERMANENT SEALING
After the fire area had been surrounded by substantial
seals, the flow of air through the area was greatly
reduced. It was now time to build permanent,
explosion-resistant seals behind the substantial seals.
Water was flowing through the substantial seals from
the fire area. It was discovered that some of the uncased
boreholes were allowing water to pass into the mine.
After casing the holes, the flow of water slowed and the
seal areas were pumped dry. The explosion-resistant
seals were completed on all three sides of the fire in
less than four days. Figure 3 is a map of the area of the
mine around the fire. The locations of the Hallibur-
ton temporary seals used in conjunction with existing
stoppings to contain the inert gas for the east set of
seals are indicated on the map.
 Figure 3. Halliburton remote seals used in conjunction with existing stoppings as temporary seals.
7 PRODUCTION RESUMES
On March 8, 2004, just 26 days after the fire was dis-
covered production resumed at the Dotiki IV Mine.
Webster County Coal utilized all of their miners dur-
ing this time either at the other company mines or
helping with the recovery of the Dotiki IV Mine. No
injuries occurred during the recovery period. A rap-
port among company, KDMM, and MSHA had been
established and continued for ongoing monitoring of
341
Copyright © 2006 Taylor & Francis Group plc, London, UK
sealed fire area. During the entire time the company
never lost sight of their goals to restore the mine to nor-
mal operations quickly and safely. Even areas enclosed
by explosion-resistant seals breathe due to pressure
differentials caused by barometric pressure changes,
ventilating pressure from mine fans, and natural ven-
tilation pressure. The configuration of the seals was
such that the North Seals were directly off of the intake
side and the East seals were directly off of the return
side of the Dotiki IV shaft. This positioning created
 a high pressure differential through the sealed area,
making it difficult to pressure balance the seals. But
after adjusting and monitoring ventilation controls for
a period of time, seal leakage was minimized. After the
cessation of the inert injection, higher oxygen content
was found in the monitoring holes in the sealed area.
It was discovered that the ground around the injection
holes had frozen to form a seal while the super-cooled
liquids were passing through the pipes. After thaw-
ing, they provided a conduit for fresh air to pass into
the mine.
8 CONCLUSION
The cooperation and free exchange of ideas among
the interested parties were factors that made for a suc-
cessful operation. This created a positive atmosphere
to formulate and execute the plan of attack. The ability
342
Copyright © 2006 Taylor & Francis Group plc, London, UK
to have ample supplies and a skilled workforce capa-
ble of installing the LOM seals in a short period of
time was critical. This enabled the Dotiki IV fan to be
de-energized and sealed to begin depleting the fire of
oxygen. Having full access to the surface above the
fire area allowed the placement of the remote seals
in the best locations based only on underground fac-
tors. Keeping a continuous flow of inert gasses into the
mine at the fire area kept the fire from getting further
out of control or possibly setting off an explosion. The
company’s decision to keep much of their workforce
available to help with recovery efforts aided in sample
collection, instantaneous gas readings, and delivery of
supplies and information.
The use of remote seals should continue to become
more common as the contractors and mine operators
become more familiar with their applications. They
provided a safe means to isolate dangerous areas and
conditions from the rest of the mine.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

First Mexican coal mine recovery after a mine fire, Esmeralda Mine

M. Santillan
Minerales Monclova, S.A. de C.V., Palaú, Coahuila, México

ABSTRACT: On May 8, 1998, a fire started in the development section of the Esmeralda Mine from methane
released into the mine through a roof-bolt hole. The flames spread quickly as the coal was ignited. After eight
hours, the Safety Department decided to seal the vertical ventilation shafts and the slopes. The quality of coal in
the Esmeralda mine is very high quality, and MIMOSA decided to recover the facilities. However, the Esmeralda
mine coals have a very high gas content of 12 m3 /ton. Immediately after the mine was sealed, MIMOSA staff
implemented a plan to monitor the gas levels underground and contacted specialists for assistance in bringing
the mine back on line.
The entire recovery operation required seven months before the Esmeralda mine was back in operation in
December 1998. Although bringing the mine back into operation was important for MIMOSA from a business
standpoint, the most important thing was the recovery of the mine without any injuries to the personnel who
were involved in the recovery.

1 INTRODUCTION 4 meters wide by 3 meters high. Mining operations are

The three underground longwall mines operated by
MIMOSA exploit coals of the Upper Cretaceous Los
Olmos Formation in the state of Coahuila in north-
ern Mexico. This region contains Mexico’s largest
coal reserve, 311,391 ktons, of which an estimated
273,808 ktons is recoverable. MIMOSA is presently
operating five of those mines in the Sabinas sub-basin
and just beginning another one in the Saltillo Basin.
The Olmos formation contains two distinct coal
seams (locally known as the Double seam) that are
mined commercially where the rock parting between
them is between 0.10 and 0.20 m (4–8 inches), for
a combined thickness of approximately 2 m (6.5 ft).
The coals are medium to high volatile in rank, and for
MIMOSA, supply related steelmaking operations in
the city of Monclova. In addition, for the last 8 years,
the MIMOSA mines have provided steam coal for elec-
tric generation that provides 8% of the total electricity
generating capacity of the country.
The coals in both the Sabinas and the Saltillo Basins
have very high gas contents with averages of 10 m3 /ton
in the Sabina’s Basin and 12 m3 /ton in the Saltillo
Basin.
2 VENTILATION
MIMOSA utilizes a longwall mining system at Mine
Esmeralda. The opening has nominal dimensions of
carried out by Alpine miners and JOY miners. Follow-
ing cut, the coal is loaded on shuttle cars that transport
the coal to the belt conveyors.
The layout of the mine at the time of fire included
two slopes, seven entries in the mains (300 meters)
from the slope, and two development sections. The
mine operated exhausting system of ventilation with
the main fan located on the surface. The fan was a JOY
M-96-50 with a motor of 900 hp and 1,200 rpm, and
a flow capacity of 155 m3 /sec of mine ventilation air.
A JOY fan was also used for auxiliary ventilation of
the mine.
3 THE FIRE
The fire began at the face of the number three entry
in the General Sur. Within three hours, it had spread
to the first, second and fifth entries. When the safety
personal and the miners were evacuated from the mine,
the ventilation shaft and portals were sealed with con-
crete. MIMSOA left some points open for sampling
gas to know the extent of methane concentrations and
migration within the mine. According to the section
foreman, the fire started around 1:00 am on May 8,
1998. The mine personnel were drilling a hole in the
mine roof with the roof drill to install the third bolt.
The drill rod worked properly the first 0.30 meters
but the last 1.20 meters were very hard rock. The roof
bolting team then removed the drill bit from the hole to

343

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 1. Location of the MIMOSA mines, Coahuila area, Mexico.
install a drill rod appropriate for the hard rock. When
removing the drill bit, a flame burst from the roof hole
and spread immediately to the face and ribs.
4 THE FIRE
The fire began at the face of the number three entry in
the General Sur. Within three hours, it had spread to
the first, second and fifth entries. When the safety per-
sonal and the miners were evacuated from the mine,
the ventilation shaft and portals were sealed with con-
crete. MIMISOA left some points open for sampling
gas to know the extent of methane concentrations and
migration within the mine. According to the section
foreman, the fire started around 1:00 am on May 8,
1998. The mine personnel were drilling a hole in the
mine roof with the roof drill to install the third bolt.
The drill rod worked properly the first 0.30 meters
but the last 1.20 meters were very hard rock. The roof
bolting team then removed the drill bit from the hole to
install a drill rod appropriate for the hard rock. When
removing the drill bit, a flame burst from the roof hole
and spread immediately to the face and ribs.
When the mine fire occurred, the mine was just
preparing the first longwall panel. After sealing the
mine, the company administration decided to recover
the mine, and spent the remainder of May interviewing
344
Copyright © 2006 Taylor & Francis Group plc, London, UK
specialists in mine fire recovery. On May 11, 1998,
the company administration established contact with
consultants from the USA (Patrick Graham, Jack
Stevenson, Donald Mitchell) and decided not to con-
tain the fire and develop procedures to seal the mine.
After review of the situation, the consultants recom-
mended drilling a vertical hole in the intersection of
the main entries diagonal #12. That hole provided
the opportunity to assess the mine-atmosphere gases
by sampling the mine environment. During the mine
recovery, MIMOSA staff used three locations for sam-
pling, the vertical ventilation shaft, the portals, and the
borehole.
On June 14, 1998, the consultants presented a pre-
liminary recovery plan to the mine personal. The plan
consisted of the following tasks:
• mine rescue team training (at least 6 teams);
• a chromatograph for sampling around the clock;
• flooding the mine with water;
• keeping the oxygen level lower than 2.0%; and
• daily review of the CO/CO2 and H/CH4 ratios.
The recovery plan was based on the “Step-by- Step”
system which consists of a method of air-lock recov-
ery to avoid a reignition of the coal. If the coal is still
smoldering in a gassy mine, there is a very high prob-
ability of a methane explosion during implementation
of the recovery plan.

Figure 2. Mine rescue team during the first phase of the
recovery plan.
5 MINE RESCUE TRAINING
This was a personal experience for the author because
he was involved with the training of the rescue team
in Minera Carbonifera Rio Escondido (MICARE) and
Minerales Monclova (MIMOSA). People were very
skeptical that a group of Mexican miners with no
previous experience in mine recovery could perform
a successful recovery operation in the face of a mine
fire. During June and July 1998, the mine recovery
team started the training with Mr. Don McBride, of
the Illinois Department of Natural Resources, who
was responsible for developing the training program
for the rescue team. The training included one week at
MICARE and a second week at the MIMOSA prop-
erty. The program devised by Mr. McBride focused on
the physical condition of the personnel and the training
primarily covered use of the self-contained breath-
ing apparatus. Then Mr. Patrick Graham, independent
consultant, took his place. His participation was crit-
ical, because he was able to successfully transfer his
knowledge to the rescue teams while maintaining the
enthusiasm and respect of the miners. Mr. Graham said
this about the mine recovery personnel, “I was very
satisfied with the competence of the benchman, fresh-
air baseman, and the MIMOSA and MICARE team
members. Never in my years of mine recovery have
I experienced greater dedication than given by them.”
In a meeting of the consultants and the technicians
from MIMOSA on July 28, 1998, the decision was
made to proceed by flooding the mine with water. The
final plan was ready on August 1, 1998.
6 MINE RECOVERY
The first phase consisted of injecting water through
borehole #2 to the 104 m elevation shoreline. Once
345
Copyright © 2006 Taylor & Francis Group plc, London, UK
the recovery team finished pumping the water, the
final recovery plan was developed and addressed
recovery of:
• inclines A & B,
• the ventilation shaft bottom,
• the shaft, and
• the 1 & 2 east entries.
The command center and engineers from the mine
operation were concerned that water might cause seri-
ous damage to the number 1 and 2 entries because
those entries are supported by steel beams and wood
post. When the team recovered the slopes, shaft, and
the main entries, the number 1 east entry did not show
any damage because the 104 m shoreline was previ-
ously located near those entries. Within the number 2
east entry, however, roof falls did not allow the team
to continue up to cross-cut five during the rescue team
inspection.
It is suspected that the mine still contained smolder-
ing coal prior to sealing the vertical shaft because the
temperatures were reportedly 1350◦ C. It is believed
that this was due to hot carbons at elevations as high
as 107 m according to graphics of hot carbons pro-
duced during the recovery. During water injection, the
precision of the testing was crucial and was made pos-
sible because the consultants maintained a mobile lab
on-site.
Esmeralda is noteworthy for having been the first
coal mine in Mexico that was recovered and put back in
operation after a mine fire started in the development
section in May 1998 (Hookhan 2004).
Mine workers were involved in the mine recovery
which entailed sealing the vertical ventilation shaft and
pumping water into the mine to reduce any probability
of fire during the recovery process.
During May to August 1998, the command center
took gas samples from the mine and with the help of
a chromatograph analyzed the gas content of the envi-
ronment underground. During this period management
developed a training program for rescue teams who had
no previous experience of mine recovery. The most
important thing was the recovery of the mine without
any injuries.
7 CONCLUSION
Officially the mine recovery finished in November
1998, and the mine resumed operation in December
1998. This mine is very important for MIMOSA. As
noted at the beginning of this paper, the recovery work
at La Esmeralda mine, was accomplished without any
injuries to the rescue teams, Everyone involved in the
project is very proud to have been a part of the first
coal mine recovery in Mexico.
 Esmeralda mine is still in operation. Following the REFERENCE
mine fire, MIMOSA management implemented a very
aggressive methane gas drainage plan utilizing hor- Hookhan, Marian, 2004. Gas Major challenge for Mexi-
izontal boreholes. In addition, MIMOSA improved can mine. In International Lonwall News Magazine (Ed.
M Hookhman).
the auxiliary ventilation to protect its operations and
to avoid a problem with methane gas concentrations.
These new measure provide a safe environment to
MIMOSA’s workers and facilities.

346

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Mine fire management, case study in Pongkor Gold Mine,

PT Aneka Tambang Tbk., Indonesia

Risono
Pongkor Underground Gold Mine, PT Aneka Tambang, Indonesia

N.P. Widodo
Department of Earth Resources Engineering, Kyushu University, Japan

R.S. Gautama
Department of Mining Engineering, Institut Teknologi Bandung, Indonesia

ABSTRACT: This paper describes the management of fire fighting in case of fire accident that was occurred
on March 3rd, 2004 in Kubang Cicau area, Pongkor underground gold mine Indonesia. The fire successfully
ceased by closing the oxygen supply as well as decreasing the air pressure and temperature using exhaust fans.
By this method the underground mine could be re-open only within one-month period since the fire occurred.

1 INTRODUCTION
Pongkor gold mine owned by PT Aneka Tambang Tbk
(PT Antam) is one of underground mines in Indonesia.
The mine is located in West Java Province, approxi-
mately 90 km from Jakarta the capital city of Indonesia
(Figure 1).
At the moment Pongkor mine has three primary
veins of gold which are being mined; Ciguha Utama,
Kubang Cicau and Ciurug. A conventional cut and fill
stopping method with hydraulic backfill of tailing is
used for the first two of these veins. A mechanized cut
and fill method, using hydraulic jumbo drill and load
haul dump (LHD), is implemented at Ciurug.
The fire accident, which will be discussed in this
paper, was occurred at Kubang Cicau area. This mine
consists of five levels, those are: Level 500, Level 550,
Level 600, Level 650 and Level 690. Each level is
connected by some raises that have a function as ore
passes, ventilation as well as man-way (service shaft).
On each level, there are cross cuts directed to the stopes
(see Figure 2). An exhaust ventilation system is con-
trolled by an axial-exhaust fan, located at KKRB-3.
Intake air to the mine is supplied from: KKRB 1,
KKRB 4, Service Shaft and Lubang Ventilasi shafts,
as well as Main Haulage Level (MHL) (see Figure 2).
On March 3rd, 2004, at Level 690 Block Cepu
Kubang Cicau – one of the mining blocks in Kubang
Cicau area – a mine fire accident was occurred. The
illustration of mining layout and the area of mine fire
is shown in Figure 2.
LUBANG KKRB 1
The locations
VENTILASI
KKRB 3 of the fire accident
KKRB 4
SERVICE
SHAFT
LEVEL 690
water pump
LEVEL 650

backfill
LEVEL 600 pump

LEVEL 550 Ma
in
Ha
ula
t rth ge
LEVEL 500 es No Le
ve
W l 50
0
th
S ou

Figure 2. Mine layout of Kubang Cicau area and location

Figure 1. Location of Pongkor Gold Mine. the fire accident was occurred.

347

Copyright © 2006 Taylor & Francis Group plc, London, UK

 The fire had burned set of timber; supporting sys- Mine Technical Manager
tems applied in Kubang Cicau and had caused several
fatalities. Ventilation treatments and actions had been
Mine Technical Vice Manager
conducted to fight the fire and it was successfully con-
ducted to stop the fire within 20 days.The ventilation
analysis and fire fighting actions are described in this Field Coordinator

paper.
Emergency Response Group Service Crew Technician

2 FIRE OCCURRENCE AND THE EFFECT

Figure 3. Structure of fire fighting and rescue team.
The fire was firstly detected visually by the occurrence
of smoke at the mine level 690 in Kubang Cicau mine
area at Block Cepu. The fire was located close to the
mine air intake Lubang Ventilasi shaft (Figure 2). From
this area, the toxic gases, produced from fire were eas-
ily spreading to the whole Kubang Cicau area. At that
time, as reported by the mine operation section, there
were 28 miners working in this area. All miners could
escape using skip from working level to the MHL 500
Level. Accident report explained that the fire initiated
by the illegal mining activities in Kubang Cicau stopes.
There were smokes came out from KKRB IV shaft
and from adits made by the illegal miners from the
surface of the hill. Gases detected at KKRB IV shaft
were as follows: the CO concentration was more than Figure 4. Kubang Cicau mine ventilations in normal
0.1% (the maximum limit of the CO gas detector in condition.
Pongkor mine) and the CO2 was 0.8%. It indicated
that the fire had burned sets of timber support. Since – Mine Technical Vice Manager: coordinator of man-
timber support is the main support system in this mine, agerial and technical actions
the fire became bigger and difficult to control. – Field Coordinator: coordinator of field actions
Fatality recorded from this accident was 12 peo- – Emergency Response Group: gas monitoring, evac-
ple from illegal miner and one from company’s miner uation, emergency actions, and fire fighting
(Ciurug mine area). Illegal miners were working close – Service Crew: installation and un-installation of
to the area where fire was initiated so they were fac- ventilation equipment, transportation sections and
ing high concentration of CO, while the miner was logistic supply
found at the Main Haulage Level 500. It was reported – Technician: electrical installation section.
he did not realized that the CO had been spread onto
the MHL airways. He was walking from Ciurug to the Plans of action consisted of mine fire location and
mine portal, from West to East (see Figure 5). distribution survey, mine airflow quantity and its qual-
ity measurement, fire fighting action and equipment
planning, actions and monitoring.
3 THE ACTION IN FIRE FIGHTING AND
3.2 Fire and ventilation network analysis
RESCUE
Ventilation analysis had been conducted after the fire
3.1 The rescue and fire fighting team occurred in order to check the change in airflow due to
preparation the fire and to understand the intake/oxygen supply of
the fire. The comparison between normal conditions
Team for fire fighting and rescue had been set up
(before the fire) and after mine fire can be seen in
immediately after the fire occurred. The main task was
Figures 4 and 5.
to stop the fire and evacuate the people in the mine.
From the analysis, it was understood that the oxy-
The team’s structure command was shown in Figure 3.
gen supply of the fire came from the intake of mine
The job description and responsibilities of team
air at Lubang ventilasi, KKRB I, KKRB III and Ser-
were as follows:
vice Shafts. With this supply, the fire could occur until
– Mine Technical Manager: took the responsibility of all timber supports being burned up. This conditions
all actions for fire fighting and rescue, as stated in could be very dangerous to the mine, because the rock
Indonesian mining law was very weak and it might be collapsed.

348

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 5. Kubang Cicau mine ventilations after the mine
fire (simulation result by Ventsim software).
Having this consideration and to evacuate the ille-
gal miners trapped in the fire locations, it was a first
priority to stop the fire. No one knew the number and
conditions of illegal miners trapped in the mine stopes.
Since the fire had been occurred for several days it
could be predicted that the illegal miners trapped in
the stopes were all died.
3.3 Field measurement
The simulation of airflow direction after mine fire
using Ventsim software matched with the measure-
ment results. The measurement results on March 17th,
2004 (14 days after fire accident) were as follows,
– Smoke came out from KKRB IV shaft with air-
flow quantity of 0.9 m3 /s, CO content was more than
1000 ppm and CO2 content was 8000 ppm
– Smoke came out from four illegal miners’ adits on
the surface (see “Lubang PETI” in Figure 5)
– The fresh air entered into the mine from KKRB I,
Lubang Ventilasi, and KKRB III shafts with the air-
flow quantity of 1.3 m3 /s, 2.1 m3 /s, and 2.25 m3 /s,
respectively
– The fresh air with very small airflow quantity
entered the mine from “Service Shaft”.
3.4 The fire fighting method and actions
The fire will be occurred if the burnable material, oxy-
gen and fire initiation present. If one of these factor is
removed, the fire will be stopped. The method of fire
fighting to stopped the fire in Kubang Cicau was:
1 Stop the initiation of fire by decreasing the pres-
sure and temperature at fire locations; that could
be done by installing an exhaust fans close to the
fire locations
2 Remove the oxygen supply to the fire by sealing
the intake of air into the Kubang Cicau Mine.
349
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 6. The seal positions to prevent oxygen supply to the
fire at Kubang Cicau mine.
Figure 7. Fan position at the Level 690, Kubang Cicau.
The actions of fire fighting were as follows:
– Installing an exhaust fan 15 kW at KKRB IV shaft
and at the illegal miner’s adit “Lubang Seng”. The
purpose was to create the negative pressure on the
stope. The stope’s intake air should be sealed to
support this action. If the conditions met, the tem-
perature might be decreased with the decreasing of
the pressure
– Sealing the air intake at the KKRB I, KKRB III,
Lubang Ventilasi, Service Shafts, and all of the ille-
gal miner’s adits (except the Lubang Seng). The seal
positions and airflow pattern after this actions can
be seen in Figure 6
– Turned-off the fan at Pondok Batu (note: by this
action, the air did not flow into the MHL airway)
– After those succeeded, the actions continued
by sealing all of the airways to the stope at the
Level 690
– Installing auxiliary fans at Level 690 (1 unit of 5
kW fan at cross cut MW 2S, 1 unit of 15 kW fan
near the Service Shaft, and 1 unit of 37 kW fan at
Vein D). Illustration of fans positions is in Figure 7
– After the temperature reduced and toxic gas
removed, the fire fighter could enter the place of
 Table 1. The concentrations of gas after the actions of miners mining activities in the area, both their open-
fire fighting. ings and the number of people working. It was believed
that the fire was initiated from the illegal miners’work-
Concentrations of gas ing area. Their cooking activities were supposed to be
the source of fire.
CO2 CO O2
Date Location ( ppm) ( ppm) (%) Although the possibility of mine fire in an ore
underground mines is small, it is important to have
March 19, 2004 Lubang PETI 1000 30 – a well planned mine ventilation system including its
KKRB IV 8000 120 – contingency plan especially for mine fire. Continu-
March 21, 2004 Lubang PETI 300 10 – ous monitoring and evaluation as well as necessary
KKRB IV 205 60 – improvement should be conducted regularly. On the
other hand, illegal miners should not be allowed to
March 23, 2004 Lubang PETI – 5 –
KKRB IV – 47 – mine in this area.

fire and water spraying was conducted to stop the
fire. The water also sprayed from the surface from
illegal miner’s adits to the stope.
3.5 The results
After the actions of fire fighting completed, the gas
concentration monitoring on the exhaust fan was
carried out. Table 1 describes the measurement data.
From data on Table 1, it could be analyzed that in the
negative oxygen reactions, the gases produced from
the fire reaction were CO, moist, and small part of CO2 .
It means the CO2 concentrations had been decreased
and the actions succeeded. This measurement was
needed to make sure that the actions were confirmed
with simulation and calculations as planned.
The measurement results also showed the decreas-
ing of CO concentrations over time. This result gave a
confident to the team that the actions had given a good
condition in stopping the fire.
Twenty days after the fire accident occurred, the
Emergency Response Group entered the mine and took
direct fire fighting actions by water spraying from the
Level 690 and from the surface. Accordingly it must
be ensured that fans were working in order to remove
the toxic gas from the fire locations.
After the fire stopped, inspection had been con-
ducted to check the mine condition. It was found
that in the locations there were many illegal miner’s
equipments and logistics remained.
4 DISCUSSION
The mine fire management and analysis that conducted
on the accident in Pongkor mine was successfully
stopped the fire. The actions of fire fighting con-
ducted were actually simple in concept and actions.
The problems were the un-identification of illegal
5 CONCLUSION
Lesson learned from this fire accident, even in the
underground gold mine there is also a potential of the
fire accident occurred. The fire accident in an under-
ground mine is very dangerous to the miner and in
many cases it is a complex problem. The standard pro-
cedures for fire prevention as well as fire fighting in
a mine should be developed. Regular training should
be conducted to ensure that the procedures have been
well understood by the miner.
The well planned ventilation system and contin-
gency strategy, the competent personnel as well as suf-
ficient infrastructures are important factors for a mine
to respond in a case of mine fire.
ACKNOWLEDGEMENT
The author would like to acknowledge all of the Mine
Fire and Rescue Team of the Pongkor Gold mine, and
the Management of PT. Antam Tbk., Indonesia.
REFERENCES
Hartman, Howard L. 1997. Mine Ventilation and Air Condi-
tioning, 3rd ed. New York: John Wiley & Sons.
Health Center Occupational Department, PT Antam Tbk
UBPE Pongkor, Report on Working Environment Moni-
toring, March 2004.
Kepmen Pertambangan dan Energi Indonesia. 1995. No.:
555. K/26/M.PE/1995.
http://www.pikiran-rakyat.com/cetak/0304/16/0804.htm
McPherson, Malcolm J. 1993. Subsurface Ventilation
and Environmental Engineering. London: Chapman &
Hall Inc.
Mining Department, PT Antam Tbk UBPE Pongkor. 2004.
Daily Mine Operation Report, March 2004.
Ventsim Software Version 3.4, Manual Handbook, CMS
Software, 2001.

350

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Queensland mine emergency level exercises assisted by fire simulation

H.W. Wu & A.D.S. Gillies

Gillies Wu Mining Technology Pty Ltd, Brisbane, Queensland, Australia

ABSTRACT: A mine’s safety and health management system must provide for managing emergencies in
Queensland coal mines under the Coal Mining Safety and Health Regulations 2001. The system must include
provision for carrying out aided rescue and self-escape of persons from the mine in an emergency and also
auditing and reviewing the emergency exercises. For a number of years mines have been participating in what
are termed Level 1, 2, 3 and 4 evacuation exercises on a regular basis. Examples of work undertaken focused on
mine fire incidents in each of these categories are described. Meeting requirements for annual involvement in
and management of these Level exercises is a significant effort for underground mines. The type of work that
has been undertaken in assisting mines with “fire” based exercises and management approaches to development
and implementation is discussed.

1 INTRODUCTION It is difficult to predict the pressure imbalance and

Many people consider that mine fires remain among
the most serious hazards in underground mining. The
threat fire presents depends on aspects such as the
nature and amount of flammable material, the ven-
tilation system arrangement, the duration of the fire,
the extent of the spread of combustion products, the
ignition location and the reaction of personnel present.
An Australian Coal Association Research Program
supported project incorporating a number of mine site
exercises, as described by Gillies et al. (2004a) and
Wu et al. (2004) has been undertaken focused on the
application of mine fire and ventilation software pack-
ages for contaminate tracing and fire modeling in coal
mines. This paper in particular examines the type of
work that has been undertaken in assisting mines with
“fire” based exercises and management approaches to
development and implementation of these.
The study into this complex area has utilized the
recently upgraded Polish mine fire simulation soft-
ware, “Ventgraph”. There is a need to understand the
theory behind the simulation program and to allow
mine site use by those already familiar with the
main existing mine ventilation analysis computer pro-
grams currently popular within the Australian, United
States and South African industries such as “Ventsim”,
“VnetPC” and “Vuma”. “Ventsim”, “VnetPC” and
“Vuma” were not designed to handle fire effects
on mine networks. Under the project a small sub-
routine has been written to transfer the input data
from the existing mine ventilation network simulation
programs to “Ventgraph”.
leakage created by a mine fire due to the complex
interrelationships between the mine ventilation sys-
tem and a mine fire situation. Depending on the rate
and direction of dip of the entries (dip or rise), rever-
sal or recirculation of the airflow could occur because
of convection currents (buoyancy effect) and con-
strictions (throttling effect) caused by the fire. This
reversal jeopardizes the functioning of the ventila-
tion system. Stability of the ventilation system is
critical for maintaining escapeways free from contam-
ination and therefore available for travel. Reversal of
air following fires can have a tragic outcome (Wala
1999).
Simulation software has the great advantage that
underground mine fire scenarios can be analyzed and
visualized. A number of fire simulation packages have
been developed to allow numerical modeling of mine
fires, such as Greuer (1984), Stefanov et al. (1984),
Deliac et al. (1985), Greuer (1988), (Dziurzyński et al.
(1988). The Ventgraph fire simulation program has
been described in detail by Trutwin et al. (1992). The
software provides a dynamic representation of a fire’s
progress in real time and utilizes a color-graphic visu-
alization of the spread of combustion products, O2 and
temperature throughout the ventilation system. Dur-
ing the simulation session the user can interact with
the ventilation systems (e.g. hang brattice or check
curtains, breach stoppings, introduce inert gases and
change fan characteristics). These changes can be sim-
ulated quickly allowing for the testing of various fire
control and suppression strategies. Validation studies
on Ventgraph have been performed using data gathered

351

Copyright © 2006 Taylor & Francis Group plc, London, UK

 from a real mine fire as undertaken by Wala et al.
(1995).
The primary objective of the part of the study
described in this paper is to use mine fire simulation
software to assist in the design of emergency training
evacuation exercises.This paper examines work under-
taken to assist mines in the planning and undertaking
of emergency training exercises and related workforce
training.
Work undertaken with appropriate bodies during
preplanning and subsequently during the course of
mine emergency exercises is discussed. Some com-
ments have been made on the ventilation aspects of
the emergency exercise from observations made dur-
ing the course of the incidents. Some of these are set
down as observations and some were personal com-
ments from participating individuals. A key aspect of
the software is the ability to model fires in a mine
and the consequent effects of control measures such as
ventilation changes and the introduction of means of
atmosphere inertisation. Management is provided with
a pre-emptive tool that gives ability to have control
measures such as emergency seals or doors in place,
as well as a predictive tool for analyzing actions prior
to implementation in the event of a fire.
Broad conclusions from work undertaken at indi-
vidual Australian coal mines are discussed as exam-
ples. The effort is built around the introduction of the
fire simulation computer program “Ventgraph” to the
Australian mining industry and the consequent model-
ing of fire scenarios in selected different mine layouts.
In Queensland coal mines are requires to participate
in what are termed Level 1, 2, 3 and 4 evacuation exer-
cises on a regular basis. Similar systems are in place in
other states. Examples of work undertaken across dif-
ferent categories of evacuation exercises are described.
2 EMERGENCY EXERCISES
Emergency training exercises have been undertaken in
Queensland since implementation of the recommen-
dations of the Inquiry Report from the Moura No 2
coalmine disaster of August 1994. Under the state’s
Coal Mining Safety and Health Regulations 2001, the
mine’s safety and health management system must
provide for managing emergencies at the mine. The
system must include provision for carrying out aided
rescue and self-escape of persons from the mine in an
emergency and also auditing and reviewing the emer-
gency exercises. For a number of years mines have
been participating in what are termed Level 1, 2, 3
and 4 evacuation exercises on a regular basis. The
categories of exercises are:
– Level 1. Whole industry exercise organized by
Mines Inspectorate and from which a report on
outcome is widely circulated.
352
Copyright © 2006 Taylor & Francis Group plc, London, UK
– Level 2. Once a year full mine simulated evacuation
followed by de-brief.
– Level 3. Done every six months for each sec-
tion crew. Crews presented with a scenario and
undertake evacuation followed by de-brief.
– Level 4. One hour training on emergency
scenario.
Reece (2005) summarized aspects of these
exercises:
– Every Queensland underground coal mine has a
mandatory requirement for a range of emergency
exercise simulations each year.
– At least one full mine evacuation must be under-
taken.
– Everyone must participate in at least one exercise
each year.
– Two Mines Inspectorate state run exercises are
organized each year.
– One of these includes a full mine evacuation.
– One targets management and inertisation and is
largely a desktop exercise.
Scenarios are generally simple and:
– Require escape using self contained self rescuers
in limited visibility.
– Focus on fire, explosion or major energy release
that must be controlled.
– Need formation and response by a formal Incident
Management Team.
– Showcase emerging technology.
– Employ borehole cameras and remote monitoring.
– May include use of automated unmanned surveil-
lance vehicles.
– Have a need for Incident Control Systems and
Dynamic Decision Making techniques.
– Utilize of the order of 15 industry assessors.
The outcome is the production of a report for indus-
try with recommendations for the individual mine and
general industry. Generally results have been good, but
individual testing reveals that many miners could not
escape. Of mine employees present 20 percent know
what to do and take the lead while the remaining are
not as focused. There has been a reluctance to mitigate
an incident even from a safe location, the emphasis
being on escape.
Queensland emergency exercises are not just to test
that a mine has a system and to satisfy a gazetted
regulation or to embarrass mine management and
make them feel uncomfortable. They are to assist
in improving responses, provide learning opportuni-
ties, reinforce good practice and showcase emerging
technology.
The paper briefly examines a number of exercises
that have recently been constructed and run.
 3 LEVEL 1 DESK TOP 2004: MAINS FIRE
WITH TWO MINERS TRAPPED

Under this exercise a simulated fire was burning in

a Mains intake airway with two miners trapped inbye
the fire. They were barricaded in a cut-through where a
500 mm ballast borehole was providing fresh air from
the surface. The scenario has been described in detail
by Gillies et al. (2004b).
The Ventgraph ventilation and fire simulation soft-
ware was used to model the fire and mine conditions
as various mitigation strategies were assessed. This
assisted in determining the ultimate best solution to
the problem. The mine used the GAG inertisation unit Figure 1. Ventgraph simulation of part of pit inbye fire (red
to reduce the intensity of the fire, whilst theoretically dot) showing CO levels after 6 hours of fire burning. This
extricating the miners via the borehole and single rope shows air quantity inbye the fire of 26.5 m3 /s, 0.3 m3 /s of
and harness rescue. A borehole camera was used to fresh air flowing down the borehole and contaminated air is
view the men and the surrounding area. At the same leaking into the refuge from B Heading.
time actions were taken to control the fire by drilling
a series of boreholes from the surface to seam level,
just outbye (in the same roadway) the last known posi-
tion of the fire as well as in cut-throughs around the
expected perimeter; then dropping flyash into the road-
ways. This was to be undertaken to block off the air to
these roadways and starve the fire of O2 .
The incident tested the mines aided response capa-
bility. That is, following on from first response princi-
ples, personnel had been evacuated, as far as possible
and the Incident Management Team (IMT) must bring
the situation under control utilizing any resources that
may be available.
The scenario commenced with a fire burning 10 Figure 2. Ventgraph simulation of part of pit 12 hours after
pillars outbye from where two miners were trapped placement of flyash outbye fire GAG operating at Drift Portal
and barricaded below a ballast borehole site. The GAG showing CO levels. Air quantity outbye the fire is reduced to
3.6 m3 /s (O2 is 12%, very high CO levels) and the heat output
inertisation unit was at the mine site but had not been
from fire reduced to 5 MW.
connected or activated. Mines rescue teams were also
on site but not operational, due to the mine measured
gas levels from products of combustion approaching diesel vehicles with acetylene diesel ignition and sub-
80% of the Lower Explosive Limit (LEL). sequently ignition of standing coal. The mine has been
For the purposes of the exercise, the mine was evacuated except for two miners (one with a suspected
advised of these details three days prior to the event, broken lower arm) who have barricaded themselves
in order to replicate some degree of preparation – as if at 61c/t A-B Hdg at the (600 mm) ballast borehole.
the incident had happened a day or so beforehand. The Attempts had been made by evacuating personnel and
mine were also advised that the IMT would be utiliz- a mine rescue team (using high expansion foam) to
ing the Ventgraph software and that as the mine also extinguish the fire but these were unsuccessful.
had access to this software it would be able to utilize
both the model and expert operators who were present
at the mine to model the situation. 3.2 Control action
This description focuses only on the fire simulation
Drilling of borehole into B Heading outbye of the fire
modeling during the exercise. Examples from the con-
is proposed with fly ash used to obstruct the airway
cise scenario development and simulation results are
(R = 5). Figure 2 shows Ventgraph simulation of part
set down.
of the pit after flyash placement outbye fire. GAG oper-
ating at Drift Portal showing CO levels. Air quantity in
B heading outbye the fire is reduced to 3.6 m3 /s (O2 is
3.1 Scenario
12%) and the heat output from fire reduced to around
As illustrated in Figure 1 the fire is in the Mains 5 MW. The fire is now controlled and small enough to
inbye 51c/t B Heading caused by collision of two fight with foam or water.

353

Copyright © 2006 Taylor & Francis Group plc, London, UK

 3.3 Alternative rescue approaches simulated
A number of alternative scenarios were simulated. The
mine normally has three fans running.
After starting up the GAG, an attempt was made to
gradually shut off fans to reduce the O2 supply to the
fire. It was found that after shutting off the second fan
(with the third fan still running), the airflow is reversed
at the fire site and an explosion may occur. This is an
unsatisfactory solution.
After starting up the GAG, an attempt was made
to shut off one fan and then progressively close off
two of the highwall entries to reduce the O2 supply
to the fire. Shutting off one fan successfully reduced
the airflow to the fire but due to a lack of segrega-
tion, closing off the highwall entries had no effect in
further reducing the flow to the fire. This alternative
approach did not appear to cause a situation under
which a gas explosion would be likely. However it is
not effective as the fire would not be starved from lack
of air.
3.4 Observations on the emergency exercise
The following comments can be made on the ventila-
tion aspects of the emergency exercise from obser-
vations made during the course of the incident.
Some comments are set down as observations and
some were personal comments from participating
individuals.
1. The Ventilation Officer (VO) was directed to
analyze a number of potential solution options, uti-
lizing Ventgraph, for the IMT. This process took
significant time mainly due to the need to care-
fully document each step as it was modeled, so that
an effective plan was generated and key activities
were not missed.
2. VO was close to the IMT (in an adjacent room) but
in effect was sufficiently removed from view that
the IMT didn’t adequately utilize the Ventgraph
model.
3. The Ventgraph model was not well understood and
as a result was very underutilized as a conceptual
aid and management tool.
4. The ventilation models and options being run
in another room was effective in that it allowed
ventilation experts to concentrate on developing
options unencumbered. However, the drawback
was that it became disconnected from the IMT
processes.
5. A number of key functions such as data entry
and validation were under resourced. Time had
not been devoted to understanding Ventgraph or
preparing the IMT room for operation. Ventgraph
is a very recent addition to site capabilities and
therefore no doubt will take time to be assimi-
lated. The assimilation process may require much
354
Copyright © 2006 Taylor & Francis Group plc, London, UK
more extensive training by the providers including
scenario training.
6. It was clear that the purpose, capabilities and lim-
itations of Ventgraph were not fully understood by
pit staff. This may in part be due to the short lead
time and also limited training in its operation by
the software supplier.
7. A lot more work needs to be done to adequately
incorporate Ventgraph into mine ventilation anal-
ysis system and current ventilation models should
be integrated into Queensland Mines Rescue Ser-
vice functions as part of GAG operations. Difficult
to evaluate operation of GAG without this. There
is no doubt that using Ventgraph will signifi-
cantly enhance the IMT’s ability to deal with mine
fires.
8. There is a need to refresh key personnel under-
standing of mine fires and their characteristics,
– for example the effect of reducing O2 whilst
maintaining airflow and effects of recirculation.
This should be interfaced with actual examples
of mine fires and successful and unsuccessful
treatment of them.
9. There is a need to objectively demonstrate the
behavior of mine fires including their products and
how they are influenced – effects of air flow, O2
concentration and inert gas.
10. Ventgraph modeling was a valuable tool but
became intense due to the number of options being
run – the IMT with more experience should ask
for a prioritized assessment of the key options.
A comprehensive emergency training exercise held
at an operating colliery incorporating a fire incident
has been examined. The Ventgraph simulation soft-
ware was used to model the fire and mine conditions
as various mitigation strategies were assessed. This
assisted in determining the ultimate best solution to the
problem. Use of the Ventgraph simulation approach
highlights actions that mine management can under-
take to improve their position in the event of a fire.
Examples of these include the following.
– Many fires occur associated with conveyor belts.
Segregation of belt headings assists in preventing
fumes ingressing escape intake roadways. Vent-
graph gives a clear indication of poor or non-existent
segregation stoppings.
– Short circuiting of fire fumes to return can keep
escape ways open. This could be achieved by open-
ing stopping man doors or vehicle doors. This can
be very difficult if these are inbye a fire source.
Remote control of these doors from a surface mine
control room can overcome this difficulty.
– Many mines have change over locations (or refuge
chambers in metalliferous mines). Ventgraph sim-
ulation can assist in determining the optimum
location for these.
 4 LEVEL 2 DESK TOP 2004: SIMULATION OF
A SPONTANEOUS COMBUSTION INDUCED
MINE FIRE
Ventgraph fire simulation software has been used to
examine and illustrate the effects on the mine venti-
lation network of an open fire on a pillar sidewall rib
induced by a spontaneous combustion heating devel-
oped from within the pillar. The simulation illustrates
the effects of the fire on the whole mine ventilation
network after an incubation period of about 700 hours
following the outbreak of the pillar fire following a
long incipient period and a migration phase upwind.
The pillar under examination is positioned separating
a mains intake heading from a return heading and so
the heating has initially migrated toward the intake air.
A fire development is examined in two stages within
the case study mine.
1. An open fire that has broken out on the intake side
of the pillar.
2. A subsequent stage when an open fire has broken
out on the return side of the pillar (the charring
phase, when the heating has continued to develop
until the hotspot and charline encounter the pillar
rib on the return side).
This hypothetical spontaneous combustion incident
is reported as a simulation scenario that focuses on
effects across the whole mine network. It is written
up as a series of developments against time from the
outbreak of the open fire in the pillar rib. The scenario
has been described in detail by Gillies & Wu (2005).
All spontaneous heatings require that certain condi-
tions are satisfied for the coal temperature to continue
to rise. Primary amongst these conditions is that, at
some point within the pile or solid mass of coal, the rate
of heat generation from oxidation exceeds the rate of
heat loss due to conduction and convection. If ever this
condition is not fulfilled, the heating will have reached
a maximum temperature and there will be no further
increase. The temperature in the pile or mass of coal
will henceforth begin to decrease. Whilst the require-
ment for this condition is well known, it is difficult to
predict the characteristics of coal mass or size, coal
reactivity, airflow flux and other parameters that will
allow the development of a high temperature heating.
Humphreys (2004) developed a numerical model
to examine the development of a heating within a coal
mass, pillar or pile. For the purposes of modeling, it
has been assumed that the starting conditions in the
coal pile are homogenous; that is with all coal at the
same particle size, reactivity and initial temperature.
An underground coal pillar or solid mass will have a
permeability that allows passage of air as controlled
by the mine ventilation air pressure across the pillar.
This permeability is likely to be lower than that exhib-
ited by loose coal in a pile although the spontaneous
355
Copyright © 2006 Taylor & Francis Group plc, London, UK
600 6000
Migration Phase
Charring Phase
Position (mm from upwind surface)
Maximum Temperature (deg C)
500 5000
Incipient Phase
400 4000
300 3000
200 2000
100 1000
0 0
0 100 200 300 400 500 600 700 800
Time (hours)
Maximum Temperature Position of Max. Temperature Position of Min. Oxygen
Figure 3. Spontaneous combustion heating development –
temperature, time and position of minimum oxygen
(Humphreys 2004).
combustion development characteristics will follow
the same trend. In Humphreys’ analysis the airflow
flux is constant across the model although obviously
there is consumption of oxygen as air passes through
the model. For the purposes of examining the nature
of a spontaneous heating as it occurs in a pile of coal,
a quasi three dimensional model has been run for a
representative coal.
The development of this heating can be summarized
in Figure 3. The peak temperature in the pile, the posi-
tion of the peak temperature (the hotspot) and the
position of the minimum oxygen concentration in the
pile are shown. At the very start of the heating, there
is a moderately rapid increase in temperature, with the
“hotspot” located at the upwind surface of the pile
and the minimum oxygen concentration at the back
surface. The rate of temperature rise moderates and
the position of the peak temperature moves gradually
downwind.
The position of the minimum oxygen remains at the
back surface of the pile, although the minimum oxy-
gen concentration is decreasing. After 275 hours, the
peak temperature has moved to the furthest downwind
position at about 1900 mm. The hotspot remains in this
position until its temperature exceeds 125◦ C. This trig-
gers a change in the behavior of the heating and the
hotspot begins to migrate forward. Shortly afterwards,
the minimum oxygen concentration in the pile falls to
zero, as does the oxygen concentration at the hotspot.
Despite this, the peak pile temperature is increasing
rapidly, at approximately 8◦ C/hour.
Once the positions of the peak temperature and
minimum oxygen concentration coincide, they begin
to migrate together toward the upwind surface. This
can only begin when the temperature profile in the
coal ahead of the hotspot is sufficient to consume
all the oxygen entering that part of the pile. The for-
ward migration of the heating is limited by the upwind
surface which triggers another increase in the coal tem-
perature. A short while after this, the temperature of
 the coal is sufficient to cause charring and a charline
is formed in the pile. The final phase of the heating is
the lateral expansion and downwind migration of the
charline, as all the reactive elements in the coal are
consumed by oxidation.
From this analysis, it is possible to divide the
development of this heating into three distinct phases:
1. The incipient phase characterized by peak tempera-
tures up to about 125◦ C. During this phase a hotspot
develops from the upwind surface, migrates down-
wind to a maximum depth and remains static in that
position.
2. The migration phase characterized by the forward Figure 4. Smoke distribution after 22 minutes. Some smoke
migration of the hotspot. During this phase the oxy- is reaching surface exhausting main fans outlets.
gen concentration falls to 0 percent and there is a
very rapid increase in the peak coal temperature.
Without remedial action, the heating continues to
develop and could lead to the outbreak of fire at the
upwind surface of the pile.
3. The charring phase, when the temperature in the
pile is sufficient to cause the formation of unreac-
tive char. Without remedial action the heating will
continue to develop until the hotspot and charline
encounter the downwind surface when an open fire
could break out.
The fire scenario developed for the emergency exer-
cises focuses on stages 2 and 3 above during which an Figure 5. Smoke distribution after 5 minutes.
open fire may be in existence.
Deputy has hose ready to fight by 60 minutes. Coal fire
grows. Deputy cannot extinguish fire at 90 minutes and
4.1 Fire scenario development
drives out reaching the surface at 105 minutes. Fire is
Spontaneous combustion fire in fractured pillar coal out of control. Decision reached that underground ven-
in the rib of a Mains Heading. There is a very tilation control will be ineffective. Decision made to
high pressure of about 1200 Pa across the adjacent shut down the two underground booster fans and one
(Intake/Return) pillar. Heating started as deep-seated of the two Main fans.
oxidation. In the initial stages of heating, moisture At 120 minutes point no CO sensors in mine have
transfer and coal oxidation predominate. alarmed (exceeded 8 ppm) yet.
Coal fire continues to grow. CO sensor on Develop-
ment Dog Leg return first alarms at 130 minutes. Coal
4.2 Intake side fire following the migratory phase fire continues to grow past 150 minute point. Seam
As the coal dries out, a substantial local hot spot devel- methane is present in mine. CH4 never reaches the
ops and begins to migrate upwind. The heating front fire zone during the simulation. CO sensor at Longwall
has moved upstream in search of oxygen to the Head- Dog Leg return has not alarmed after 180 minutes.
ing pillar rib. It has just developed to the point of an
open fire as simulation begins. Some highlight points
are set down. 4.3 Return side fire following the charring phase
Smoke first reaches the development face at 22 Following the charring phase, when the heating has
minutes as shown in Figure 4. continued to develop until the hotspot and charline
Hypothetical actions of Development face crew are encounter the pillar rib on the return side an open fire
predicted. Crew see smoke and phone Outbye Deputy has broken out on the downwind side of the pillar.
at 30 minutes. Crew contact Control Room Operator A spontaneous combustion initiated fire is present
and commence to evacuate mine. Crews drive out in in fractured rib pillar coal in the Mains Heading return
smoke. Crews reach surface at time since fire outbreak near the No 2 booster fan. There are no electronic
of 45 minutes. sensors inbye the fire.
Hypothetical actions of Deputy are predicted. Figure 5 shows location of fire and smoke
Deputy finds fire source at 45 minutes after fire start. distribution.

356

Copyright © 2006 Taylor & Francis Group plc, London, UK

 4.4 Summary
Both these intake and return side scenario simulation
could be undertaken for much longer on the assump-
tion that coal within the mine continues to burn and
no remedial action such as flooding or introduction of
gas inertisation occurs. It has shown how a relatively
common form of mine fire, a spontaneous combus-
tion initiated coal pillar fire (with the pillar separating
intake and return air and with substantial pressure dif-
ferences) can affect the mine workings. It can be shown
how CO levels in mine airways increase over time for
a specific fire build up scenario.
In the intake side fire significant CO levels reach
the Mains and Development faces early but also even-
Figure 6. Smoke distribution at 5 minutes.
tually reach the longwall face if the fire is not stabilized
and extinguished. The fumes from the fire have only
limited effect on the Longwall face as it receives most
he checks Panel antitropal beltway and finds CO
of its intake air from unaffected airways.
less than 100 ppm. Crew continue on foot via Panel
In the return side fire significant CO levels build up.
antitropal intake beltway having passed through stop-
However these pollutants are restricted to the return
ping man door.
airways and so do not directly imperil miners who are
While face crew evacuating up Panel antitropal belt-
evacuating the mine.
way, CO levels of 20 to 100 ppm will be experienced
while they walk out. Distance dictates that next cache
of rescuers is outbye fire. (Normal rate in smoke is
5 LEVEL 3/4 2005: SECTION EVACUATION
1000 m walk in 60 minutes.)
AND TRAINING
Longwall tailgate return Dog leg CO sensor alarms
at 30 to 35 minutes. Control room phones or PEDs
Ventgraph fire simulation software has been used to
the Outbye Deputy if not done already. Control room
examine and illustrate the effects on the mine ventila-
can interrogate and see rising CO level. Control room
tion network of an open fires caused by vehicle acci-
phones the Shift Coordinator if not done already.
dents leading to fuel ignition. The example illustrated
As time moves on face crew are evacuating using
here was developed for a mine with three working
Maingate antitropal belt road checking stopping doors
crews. The assignment was to develop three separate
for fresh air in travel road (gas reading taken). At each
PowerPoint training sequences, namely one for each
DAC along the belt they buzz Mine Control. CO levels
working place. These form Level 4 training packages
continuously dangerous at 20 to 100 ppm while they
that can be delivered by a company trainer. Follow-
walk out. Outbye Deputy will have arrived at intake
ing training of all crews the company will implement
side of fire by about 45 minutes or earlier if he has
a Level 3 exercise under which the same incident is
a vehicle. Otherwise his arrival is a little later. Panel
used and a mine evacuation is required.
intakes inbye the fire is greater than 100 ppm CO.
Some sequences from one example are given for
Conclusion points highlighted in the training are
illustration.
The scenario presented is of a vehicle fire in a long- – Evacuating face crew in belt air should pass through
wall Maingate. Driver is stunned and observing but not to intake air outbye the fire and then exit mine
fighting fire. Face seam methane gas source is illus- through any intake Mains headings. Time escaping
trated by use of measured mine monitored readings. walking up beltway may be almost an hour.
Figure 6 shows fire after 5 minutes. – Smoke reaches face within 5 to 10 minutes and
The crew on the longwall face smell smoke within completely fills panel within 15 minutes.
10 minutes and raise the alarm by phoning Control. – CO levels rise rapidly and so movement out of LW
Men stop work, don 30 minute self rescuers and move panel into clear intake air in the mains should be as
to MG first rescuer cache. Men don 50 minute self fast as possible.
rescuers at cache and proceed to crib room at 25 minute – Preference to evacuate through Beltway as contam-
point. Evacuate at 35 minute point, drive up intake for inant level is lower. Man doors and segregation
300 m in smoke. Deputy is continuously monitoring stoppings must be kept closed and in good order
atmosphere with Personal Gas Detector. Smoke is too to achieve segregation.
thick at 300 m outbye face and they continue on foot. – Escape headings should be kept clear to assist evac-
Deputy reads CO at 300 m point from face and uation which may be over a long distance (up to full
finds CO in excess of 100 ppm in intake roadway, panel length).

357

Copyright © 2006 Taylor & Francis Group plc, London, UK

 – Seam methane vents through returns and so should
not pass near fire.
– Fight fire as soon as possible to stabilize atmo-
sphere and assist the men to escape.
– Evacuation exercises allow crews to become famil-
iar with the self escape equipment and procedures
and experience an escape in potentially non clear
and difficult conditions.
Fire incidents cause many important mine emer-
gencies. Training for section crews and “practice”
evacuation exercises are important. Planning of exer-
cises and production of training material using mine
fire simulation software is valuable. This approach is
being adopted by an increasing number of mines.
6 CONCLUSIONS
A study has examined the potential for simulation of
the effects of a variety of relatively common forms of
mine fires in developing level 1, 2, 3 and 4 emergency
exercises. The approach emphasizes the ability of
“Ventgraph” mine fire simulation software to be pro-
ductively used to preplan for mine fires and possible
emergency evacuations.
The background to this approach to simulating the
effects of mine fires on the mine ventilation network
has been examined. The anatomy of a number of fires
have been analyzed. Mine fires are recognized across
the world as a major hazard issue. New approaches
allowing improvement in understanding their conse-
quences have been developed as an aid in handling
this complex area.
The mine fire simulator Ventgraph has been shown
to be an important tool in planning for mine fires
developed from spontaneous combustion heatings.
The capability to visually display the spread of effects
of a fire quickly and reliably provides a strong aid to
those involved in developing emergency plans or con-
tributing to emergency management. The active use
of mine fire simulation in emergency planning should
continue to be encouraged.
ACKNOWLEDGEMENT
The support of the University of Queensland, the
Queensland Mines’ Inspectorate, Australian Coal
Association Research Program and a number of oper-
ations within the Australian coal mining industry in
funding and contributing to this study are acknowl-
edged.
REFERENCES
Deliac, E.P., Chorosz, G. & D’Albrand, N. 1985. Develop-
ment of Ventilation Software on Personal Computers in
358
Copyright © 2006 Taylor & Francis Group plc, London, UK
France and the Application to the Simulation of Mine
Fires. In P. Mousset-Jones (ed.), Proceedings 2nd US
Mine Vent. Symp.: 19–27. Littleton, Colorado: Society of
Mining Engineers.
Dziurzyński, W., Tracz, J. & Trutwin, W. 1988. Simulation of
Mine Fires. In A.D.S. Gillies (ed.), Proceedings 4th Int.
Mine Vent. Cong.: 357–363. Melbourne: The Australasian
Institute of Mining and Metallurgy.
Gillies, A.D.S., Wala, A.M. & Wu, H.W. 2004a. Mine fire
simulation in Australian mines using computer software.
Australian Coal Association Research Program Grant
C12026, Final Report, November.
Gillies, A.D.S., Wu, H.W., Reece, D. & Hosking, R.S. 2004b.
Use of mine fire simulation for emergency prepared-
ness. Queensland Mining Industry Health and Safety
Conference, Townsville, 13–22 August.
Gillies, A.D.S. & Wu, H.W. 2005. Spontaneous combus-
tion and simulation of mine fires and their effects on
mine ventilation systems. Coal Operators’ Conference,
Aus.: 225–236. Melbourne: The Australasian Institute of
Mining and Metallurgy.
Greuer, R.E. 1984. Transient-State Simulation of Ventilation
Systems in Fire Conditions. In M.J. Howes & M.J. Jones
(eds), Proceedings 3rd Int. Mine Vent. Cong.: 407–410.
London: Institute of Mining and Metallurgy.
Greuer, R.E. 1988. Computer Models of Underground Mine
Ventilation and Fires. US Bureau of Mines Information
Circular: IC 9206 (Recent Developments in Metal and
Non-Metal Mine Fire Protection: 6–14.
Humphreys, D. 2004. The Application of Numerical Mod-
eling to the Assessment of the Potential for, and the
Detection of, Spontaneous Combustion. Underground
coal Mines. Ph.D thesis (unpublished), University of
Queensland, 2004.
Reece, D. 2005. Lessons Learnt from Emergency Exer-
cises. Mine fires Simulation Workshop. In A.D.S. Gillies
(ed.), Eighth International Mine Ventilation Congress.
Melbourne: The Australasian Institute of Mining and
Metallurgy.
Stefanov, T.P., Asenyan, E.E. & Vlasseva, E.D. 1984.
Unsteady-state Processes during An Open Fire in a Ven-
tilation Network. In M.J. Howes & M.J. Jones (eds), Pro-
ceedings 3rd Int. Mine Vent.: 417–420. London: Institute
of Mining and Metallurgy.
Trutwin, W., Dziurzyński, W. & Tracz, J. 1992. Computer
Simulation of Transients in Mine Ventilation. In R. Hemp
(ed.), Proceedings 5th Int. Mine Vent. Cong.: 193–200.
Johannesburg: Mine Ventilation Society of South Africa.
Wala, A.M., Dziurzyński, W., Tracz, J. & Wooton, D. 1995.
Validation Study of the Mine Fire Simulation InA.M. Wala
(ed.), Model. Proceedings 7th US Mine Vent. Symp.: 199–
206. Littleton, Colorado: Society of Mining Engineers.
Wala,A.M. 1999.Three Underground Coal MineExplosions –
Twenty Miners Killed – One Reason. In J. Tien (ed.),
Proceedings 8th US Mine Vent. Symp.: 395–404. Rolla,
Missouri: University of Missouri.
Wu, H.W., Gillies, A.D.S. & Wala, A.M. 2004. Case Studies
from Application of Numerical Simulation Software to
examining the effects of fires on mine ventilation systems.
In R. Ganguli & S. Bandopadhyhy (eds). Proceedings 10th
US Mine Vent. Symp.: 445-455. Rotterdam: Balkema.
 6. Mine seals

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Explosion evaluation of mine ventilation stoppings

E.S. Weiss, K.L. Cashdollar & S.P. Harteis

U.S. Department of Health and Human Services, Centers for Disease Control, National Institute
for Occupational Safety and Health, Pittsburgh, PA, USA

G.J. Shemon, D.A. Beiter & J.E. Urosek

U.S. Department of Labor, Mine Safety and Health Administration, Pittsburgh, PA, USA

ABSTRACT: The National Institute for Occupational Safety and Health (NIOSH) and the Mine Safety and
Health Administration (MSHA) conducted joint research to evaluate explosion blast effects on mine ventilation
stoppings at the NIOSH Lake Lynn Experimental Mine (LLEM). After mine explosion accidents, MSHA con-
ducts investigations to determine the cause(s) as a means to mitigate or eliminate future occurrences. As part of
these post-explosion investigations, the condition of underground stoppings, including the debris from damaged
stoppings, are documented as evidence of the strength and the direction of the explosion forces. The LLEM data
showed that a total pressure of 52 kPa (7.6 psi) destroyed the solid concrete block stopping, ∼36 kPa (∼5.2 psi)
destroyed the hollow-core concrete block stopping, and 9 kPa (1.3 psi) destroyed the steel panel stopping. These
results will assist investigators in determining the explosion forces that destroy or damage stoppings during
actual coal mine explosions.

1 INTRODUCTION Unlike mine ventilation seal structures (30 CFR

Permanent stoppings are utilized to control and dir-
ect the ventilation air flow through underground coal
mines to dilute and render harmless methane, entra-
ined coal dust, and other contaminants at the working
face and other areas of the mine. The Ventilation con-
trols section (§75.333) of Title 30 Code of Federal
Regulations (30CFR) requires that permanent stop-
pings be built and maintained between intake and
return air courses beginning at the third connecting
crosscut outby the working face, and separate other
air courses and direct air as specified. To perform
the intended function and meet the requirements of
30 CFR §75.333, permanent stoppings are to be con-
structed in a traditionally accepted method and of
materials that have been demonstrated to perform ade-
quately or in a method and of materials that have been
tested and shown to have a minimum strength equal
to or greater than the traditionally accepted in-mine
controls. A few examples of traditionally accepted
(Federal Register Preamble 1996) stopping construc-
tion methods are as follows: 1) 20-cm and 15-cm (8-in
and 6-in) concrete block (both hollow-core and solid)
with mortared joints; 2) 20-cm and 15-cm concrete
blocks, dry-stacked and coated on one or both sides
with a strength enhancing sealant suitable for dry-
stacked stoppings; and 3) steel stoppings (minimum
20-gauge) with seams and perimeter sealed with a
suitable mine sealant.
§75.335; Greninger et al. 1991, Mitchell 1971, Weiss
et al. 2002) that are commonly used to isolate unused
sections of a mine, stoppings are not intended to with-
stand even moderate explosion overpressures. Unfor-
tunately, mine explosions do occur. Depending on the
location and severity, explosions can result in fatali-
ties and injuries to underground mining personnel, and
cause considerable damage underground to equipment
and structures. MSHA personnel conduct investiga-
tions of these explosion accidents to determine the root
cause(s) as a means to mitigate or eliminate future
occurrences. As part of these post-explosion inves-
tigations, the location and condition of underground
structures and debris are mapped. This information
assists in determining the strength and the direction of
the forces of the explosion.
The research evaluations at the LLEM involved var-
ious full-scale stoppings subjected to known overpres-
sures generated from methane and/or coal dust explo-
sions. Since the construction details and explosion
pressure forces were known for these tests, the post-
explosion observations and data will be useful for
future investigations of mine explosion accidents.
2 EXPERIMENTAL MINE AND TEST
PROCEDURES
The explosion evaluation tests on the various perma-
nent stoppings were conducted within the experimental

361

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 1. Plan view of the Lake Lynn Experimental Mine showing multiple entry area and stopping locations.
mine at Lake Lynn Laboratory (Mattes et al. 1983,
Triebsch & Sapko 1990). Lake Lynn Laboratory is
located about 80 km (50 miles) southeast of Pittsburgh,
near Fairchance, Fayette County, PA, and occupies
more than 1.6 km2 (400 acres) at a former limestone
mine.
The LLEM consists of nearly 2,300 m (7500 ft)
of underground mine entries developed for mining
research. Figure 1 shows an expanded view of the
stopping test area in the multiple-entry section of the
LLEM. The faces, or closed ends, of A-, B-, and C-
drifts are at the right of the figure. The three drifts are
each about 520 m (1700-ft) long and extend far to the
left of the portion shown in the figure. The stoppings
were built in the crosscuts between B- and C-drifts,
and the test explosions were ignited in C-drift. The
permanent stoppings were constructed in crosscut 4
(X-4) at 108 m (355 ft), X-5 at 138 m (452 ft), X-6
at 168 m (550 ft), or X-7 at 198 m (649 ft) from the
face of C-drift. Each of the crosscuts is approximately
6.1 m wide by 2.2 m wide (20.0 ft wide by 7.3 ft high)
with a cross-sectional area of about 13.4 m2 (145 ft2 ).
Explosion-resistant seals from a previous study were
located in X-1 through X-3. Before each explosion
test, a 60-t pneumatically operated, track-mounted,
concrete and steel bulkhead was positioned across
E-drift to contain the initial explosion pressures within
C-drift.
Most of the tests utilized a ∼9% methane-air conc-
entration within a 3-m-deep by 3.7-m-wide (10-ft-deep
by 12-ft-wide) ignition zone (∼27 m3 or ∼955 ft3 )
contained in the C-drift face area with a clear plas-
tic diaphragm (Fig. 1). An explosion-proof fan mixed
the natural gas and air prior to ignition. Electrically
activated matches located at the face (closed end) or
outby the face within the gas ignition zone, depend-
ing on the explosion overpressure desired, were used
to ignite the flammable natural gas and air mix-
tures. In some of the tests, shelves of coal dust were
located within a distance of 3 to 12 m (10 to 40 ft)
362
Copyright © 2006 Taylor & Francis Group plc, London, UK
from the face as a means to increase the explosion
overpressures.
Pressure transducers were mounted in every data
gathering panel in A-, B-, and C-drifts. The pressure
transducers of particular interest for these stopping
evaluations were along the C-drift rib at 93, 123, 153,
182, and 231 m (304, 403, 501, 598, and 757 ft) from
the face at the data gathering panel positions shown
in Figure 1. These rib transducers were perpendicu-
lar to the direction of the propagating explosion, and
therefore measured static pressures. There was also a
pressure transducer mounted on the C-drift side of each
of the stoppings.These transducers faced the explosion
forces propagating into the crosscut, and measured the
total pressure (static plus dynamic) at the stoppings.
The static pressure is the pressure that is exerted in all
directions; the dynamic pressure is the pressure asso-
ciated with the explosion wind or gas flow. Attached at
the center (mid-height and mid-width) of the B-drift
side of each stopping was a linear variable differen-
tial transducer (LVDT) that measured the movement
of the center of the stopping during each explosion. A
high-speed, PC-based computer data acquisition sys-
tem collected the data from the various instruments at
a sampling rate of 1500 per second. The reported data
were averaged over 10 ms.
3 STOPPING CONSTRUCTION
3.1 Concrete block stoppings
Four hollow-core concrete block stoppings were con-
structed in X-4 through X-7 between C- and B-drifts
at the LLEM (Fig. 1). Each stopping was located
approximately 1.5 m (5 ft) toward B-drift from the
midpoint of the crosscut (approximately 7 m or 23 ft
deep into the crosscut, as measured from C-drift).
The blocks were 3-core concrete blocks, with nomi-
nal dimensions of 15-cm by 20-cm by 40-cm (6-in by
8-in by 16-in). The uniaxial compressive strength of
 Figure 2. Miner applying sealant to concrete block
stopping.
the block material was 13 MPa (1900 psi). A concrete
foundation was installed along the width of each cross-
cut, and a small amount of mortar was used under the
first course of block for leveling purposes. The remain-
ing blocks were dry-stacked (no mortar between the
block joints) with staggered joints. Wood wedges were
used to tighten each block course at the mine ribs.
Wood header boards and wedges were used between
the top block course and the mine roof to tighten the
structure. An approximately 6-mm (¼-in) thick coat-
ing of an approved sealant was applied to both sides
of the stoppings as shown in Figure 2.
After the completion of the explosion test evaluation
and the removal of the first set of stoppings, addi-
tional concrete block stoppings were constructed in
X-4 and X-5 using solid concrete blocks, with nomi-
nal dimensions of 15-cm by 20-cm by 40-cm (6-in by
8-in by 16-in). These stoppings were constructed in the
same manner as the previously evaluated hollow-core
concrete block stoppings.
3.2 Steel panel stoppings
Jack Kennedy Steel Stoppings manufactured by Jack
Kennedy Metal Products and Buildings, Inc., (here-
inafter referred to as the steel panel stoppings) were
constructed within the LLEM as per the written Ins-
truction Guide provided by the manufacturer. (Dis-
claimer: Mention of any company or product does not
constitute endorsement by NIOSH or MSHA). These
stoppings consisted of a series of 30-cm wide by 183-
cm high by 5-cm thick (12-in wide by 72-in high by
2-in thick) vertical telescoping steel panels (formed
from 20 gage galvanized steel sheeting) that could be
lengthened or shortened to accommodate roof heights
between 1.8 and 3.1 m (72 and 120 in).
The stoppings were constructed on level concrete
foundations in X-6 and X-7. Since the LLEM instal-
lation required the use of panel heights in excess of
363
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 3. Nearly completed Kennedy steel stopping show-
ing the back side of the panels attached to the horizontal steel
angle bars.
1.5 m (5 ft), three rows of steel angle bars, extending
horizontally from one rib to the other, were required
as per the manufacturer’s Instruction Guide. Figure 3
shows the nearly completed steel panel stopping in X-
6 as shown from the B-drift side. The steel angles were
positioned into small holes that were cut into each rib.
The angle bars were located at 0.6, 1.0, and 1.7 m (22,
39, and 68 in) from the mine floor in accordance with
the manufacturer’s specifications.
To start the stopping installation, the first steel panel
was installed near the center of the crosscut. As part
of the installation, a piece of rigid foam was manu-
ally inserted at the top and bottom of each panel to
provide a better seal with the roof and floor. Then,
a specially-designed installation jack was positioned
within the panel’s top and bottom grooves and used to
exert a roof-to-floor pressure according to manufac-
turer’s specification to temporarily hold the stopping
panel in place. A wire twist clamp was fastened around
the horizontal rib-to-rib angle, inserted into the inside
flanges of the panel, and then tightened. Six clamps
were used on the panel to attach it to the three hori-
zontal angles. The temporary jack was then removed
from the panel. The steel angles and clamps maintain
compression and keep the panels aligned.
After the installation of the center panel, a second
panel was then installed against the outby rib, jacked
into place, and secured with twist clamps. This pro-
cess was repeated until all of the panels were installed
across the crosscut. The panels were alternated such
that the intersection of the telescoping panel sections
offset each other. This was accomplished by turning
every other panel upside down. Lapover panels were
used to cover any gaps between panels, and side exten-
sions were used to cover any large gaps between the
panels and the mine rib. A polyurethane sealant was
applied to the perimeter on both sides of the stop-
ping and between all of the vertical panel joints on the
 smooth, or closed panel side (which is the high ventila- #428. The LVDT data showed that both the X-4 and
tion pressure side in an actual coal mine installation). X-5 stoppings moved more than 8 cm (3-in or the maxi-
All of the steel stoppings were constructed such that mum displacement that can be measured by the sensor)
the high ventilation pressure side of each stopping was as they were destroyed.
on the C-drift side, which was the side that would be The pressure data recorded for the intact hollow-
directly impacted by the explosion forces. core concrete block stopping in X-6 showed an inter-
polated static pressure of 22 kPa (3.2 psi); the LVDT
showed almost no movement of this stopping. At the
4 EXPLOSION TEST RESULTS intact stopping in X-7, the total pressure was 23 kPa
(3.4 psi), the interpolated static pressure was 20 kPa
4.1 Hollow-core concrete block stoppings
The four hollow-core block stoppings were evaluated
in a series of LLEM explosion tests (#427–#434). In
test #428, the hollow-core concrete block stoppings
in X-4 and X-5 were destroyed by the explosion. The
hollow-core concrete block stoppings in X-6 and X-7
showed little damage from this explosion. The pressure
traces associated with the stopping in X-4 are shown in
Figure 4. The pressure measured by the transducer in
front of the stopping is listed as “X-4C, 108 m” and is
shown as the solid black curve in the figure. The max-
imum measured pressure value was 36 kPa (5.2 psi).
The pressure traces from the two transducers in the
C-drift rib at 92.7 and 122.8 m (304 and 403 ft) were
adjusted in time so that the peaks would match that
of the transducer in X-4. The interpolated (weighted
average) static pressure from these two transducers Figure 5. Debris from block stopping in X-4 after LLEM
was 25 kPa (3.6 psi) at the X-4 stopping location. The test #428, looking toward B-drift.
B-drift pressure behind the stopping was near zero
until after the stopping was destroyed.
The debris field from the destroyed X-4 stopping is
shown in Figure 5. The plan view debris map is shown
in Figure 6, with whole or partial blocks shown as solid
squares and pieces of blocks shown as small x’s. The
original stopping location between C- and B-drifts is
shown by the double horizontal line in the crosscut
near the bottom of the figure. Note that these pressure
traces, photograph, and debris map for the X-4 hollow-
core concrete block stopping are representative of the
type of data collected on each of the stoppings during
the three evaluation programs described in this paper.
The interpolated static pressure was 23 kPa (3.4 psi)
at the X-5 stopping that was also destroyed during test

Figure 6. Plan view debris map for the X-4 stopping after
Figure 4. Pressure traces for the block stopping in X-4 LLEM test #428, with original stopping location shown by
during LLEM test #428. double line near bottom of figure.

364

Copyright © 2006 Taylor & Francis Group plc, London, UK

 (2.9 psi), and the LVDT showed a movement of less
than 0.5 cm (0.2-in).
4.2 Steel panel stoppings
The two steel panel stoppings in X-6 and X-7 were
evaluated during a series of two LLEM explosion tests
(#457–#458). During test #457, there was little or no
observable damage to either of the steel panel stop-
pings. Both the total and static pressures were ∼6 kPa
(∼0.8 psi) at the steel panel stoppings in X-6 and
X-7. The pressures in B-drift behind the stoppings
were negligible. The LVDT showed ∼1 cm (∼0.4-in)
maximum displacement for the two steel panel stop-
pings, but returned close to its original position after
the explosion.
During test #458, the steel panel stoppings in
X-6 and X-7 were destroyed by the explosion. The
maximum total pressure measured at X-6 was 9 kPa
(1.3 psi). The interpolated static pressure was 8 kPa
(1.2 psi) at the X-6 stopping location. The B-drift pres-
sure behind the stopping in X-6 was near zero until
after the stopping was destroyed. The debris from the
destroyed X-6 steel panel stopping is shown in Fig-
ure 7. The stopping in X-6 was displaced from its
original position as a unit.
Similar pressure data were collected for the steel
panel stopping in X-7. The total pressure at the stop-
ping in X-7 was 9 kPa (1.25 psi), and the interpolated
static pressure was 8 kPa (1.2 psi). After the test, one
side of the X-7 stopping was still partially attached to
the outby rib, but the remainder of the stopping was
on the floor.
4.3 Solid concrete block stoppings
The two solid-concrete-block stoppings in X-4 and X-5
were evaluated during a series of seven explosion tests
(#457–#463). During test #460 (fourth explosion test
Figure 7. Debris from the X-6 steel panel stopping after
LLEM test #458, looking toward B-drift.
365
Copyright © 2006 Taylor & Francis Group plc, London, UK
against these stoppings), the total pressure was 46 kPa
(6.7 psi) and the static pressure was 31 kPa (4.4 psi)
at the X-4 solid-block stopping. At the X-5 solid-
concrete-block stopping, the total pressure was 33 kPa
(4.7 psi) and the static pressure was 26 kPa (3.8 psi).
The LVDT’s showed movements of about 3 cm (1.2-in)
for the X-4 stopping and about 2 cm (0.8-in) for the
X-5 stopping. These maximum LVDT displacements
were recorded ∼0.02 s after the occurrence of the peak
total pressures on each stopping. After the explosion
test, both stoppings were essentially intact. However,
pronounced cracking was evident on the X-4 stopping
between the first and second block course, with a 1.3-
to 2.5-cm (½- to 1-in) total displacement of the entire
second block course toward B-drift. Additional hori-
zontal cracking above the entire top block course was
noted, and a hairline crack extended nearly rib-to-rib
across the centerline of the stopping. For the X-5 stop-
ping, a pronounced crack was observed between the
first and second block courses with a 0.6 cm (¼-in) dis-
placement of the entire second course toward B-drift.
During test #461, the pressures were not higher than
during test #460, and no significant additional dam-
age was observed. During test #462, the total pressure
was 52 kPa (7.6 psi) and the static pressure was 39 kPa
(5.7 psi) at the X-4 solid concrete block stopping. This
stopping was essentially destroyed. Most of the stop-
ping blocks were scattered to and beyond the B-drift
intersection. At the X-5 block stopping, the total pres-
sure was 46 kPa (6.7 psi) and the static pressure was
38 kPa (5.5 psi). The stopping withstood the pressure
pulse. The X-5 stopping exhibited a more pronounced
horizontal crack between the first and second block
courses, with an approximately 1.3 cm (½-in) displace-
ment of the entire second block course toward B-drift.
A new hairline crack extending from the right center of
the stopping to the outby floor corner was also evident.
At the center of the X-5 stopping, the LVDT showed
movement of slightly over 2.3 cm (0.9-in) during the
explosion, but returned close to its original position
after the explosion.
A summary of the explosion pressure data for the
various stoppings is shown in Tables 1 and 2. Note
that for all of these stopping evaluations, the failure
pressures are based on the construction conditions
and explosion tests within the non-yielding limestone
strata at the LLEM. Stopping strengths in mines may
vary from these LLEM results when the stoppings are
constructed in a coal seam and if subjected to roof
convergence and/or floor heave.
5 CONCLUSIONS
NIOSH and MSHA conducted full-scale evaluation
studies to determine the explosion pressures required
to destroy typical coal mine ventilation stoppings at
 Table 1. Static pressures necessary to destroy stoppings. the explosion forces that destroy or damage stoppings
during actual coal mine explosion accidents.
Static pressure at Static pressure at
which stopping which stopping
survived was destroyed
DISCLAIMER
Stopping type kPa psi kPa psi
The findings and conclusions in this report are those of
Kennedy Metal 6 0.8 8 1.2 the authors and do not necessarily represent the views
Hollow-Concrete- 22 3.2 23 3.4 of the National Institute for Occupational Safety and
Block Health.
Solid-Concrete- 38 5.5 39 5.7
Block
REFERENCES
CFR. Code of Federal Regulations. Washington, DC: U.S.
Table 2. Total pressures necessary to destroy stoppings. Government Printing Office, Office of the Federal
Register.
Total pressure at Total pressure at Federal Register Preamble, Rules and Regulations, Vol. 61,
which stopping which stopping No. 18, pg. 9783, 1996.
survived was destroyed Greninger, N.B., Weiss, E.S., Lusik, S.J., & Stephan C.R.
1991. Evaluation of solid-block and cementitious foam
Stopping type kPa psi kPa psi seals. Pittsburgh, PA: U.S. Department of the Interior,
Bureau of Mines, RI 9382.
Kennedy Metal 6 0.8 9 1.3 Mattes, R.H., Bacho, A., & Wade, L.V. 1983. Lake Lynn Lab-
Hollow-Concrete- ∼23 ∼3.4 ∼36 ∼5.2 oratory: construction, physical description, and capability.
Block Pittsburgh, PA: U.S. Department of the Interior, Bureau of
Solid-Concrete- 46 6.7 52 7.6 Mines, IC 8911.
Block Mitchell, D.W. 1971. Explosion-proof bulkheads: present
practices. Pittsburgh, PA: U.S. Department of the Interior,
Bureau of Mines, RI 7581.
Triebsch, G.F. & Sapko, M.J. 1990. Lake Lynn Labora-
the LLEM. In summary, a static pressure of 39 kPa tory: a state-of-the-art mining research laboratory. In
(5.7 psi) and a higher total pressure of 52 kPa (7.6 psi) Proceedings of the International Symposium on Unique
destroyed a typical solid concrete block stopping as Underground Structures, Vol. 2: 75-1 to 75-21. Golden,
installed within the LLEM. In comparison, a static CO: Colorado School of Mines.
pressure of 23 kPa (3.4 psi) and a higher total pressure Weiss, E.S., Cashdollar, K.L., & Sapko, M.J. 2002. Eval-
uation of explosion-resistant seals, stoppings, and over-
of ∼36 kPa (∼5.2 psi) destroyed a typical hollow-core
cast for ventilation control in underground coal mining.
concrete block stopping. A static pressure of 8 kPa Pittsburgh, PA: U.S. Department of Health and Human
(1.2 psi) and a slightly higher total pressure of 9 kPa Services, Public Health Services, Centers for Disease
(1.3 psi) destroyed the typical steel panel stopping. Control and Prevention, National Institute for Occupa-
The results from the recent LLEM evaluations will tional Safety and Health, DHHS (NIOSH) Publication No.
assist investigators in more accurately determining 2003-104, RI 9659.

366

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Evaluation of remotely installed mine seals for mine fire control

M.A. Trevits
National Institute for Occupational Safety and Health, Pittsburgh, Pennsylvania

T.A. Gray & P. Glogowski

GAI Consultants, Inc., Homestead, Pennsylvania

L.M. Crayne
Howard Concrete Pumping Company, Cuddy, Pennsylvania

ABSTRACT: Of the 19 major mine fire events (including thermal events) that have occurred in the last 6 years
(2000–2005), it is estimated that remotely constructed mine seals could have been used at 65% of the events to
control or suppress the fire. Underground observations of seals that have been remotely installed during mine
fire events show that they often do not fully achieve mine roof-to-floor and rib-to-rib closure. Unfortunately, the
inability to reliably close the mine void has limited or precluded the regular use of this technology. The National
Institute for Occupational Safety and Health is conducting full-scale underground experiments at the Lake Lynn
Experimental Mine to identify and remedy existing remote sealing technology shortcomings, to develop novel
technologies, and to transfer the new or improved technologies to the mining industry. This paper will discuss
the remote mine seal testing program and will provide the results of the in-mine experiments.
Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent
the views of the National Institute for Occupational Safety and Health.

1 INTRODUCTION

It has been reported that from 1990–1999 there were

81 coal mine fires in the United States (DeRosa, 2004).
Moreover, in the last 6 years, 19 major mine fires
and thermal events have occurred in underground coal
mines in the United States. On average, about three
events have occurred each year with a maximum of
five events over this time period. These statistics sug-
gest that mine fires are occurring with alarming fre-
quency. It is estimated that remotely constructed mine
seals could have been used at 65% of the events to
control or suppress the fire.
The need to evaluate, improve, and develop new
technology to remotely construct mine seals was iden-
tified jointly by National Institute for Occupational Figure 1. Layout of the Lake Lynn Experimental Mine.
Safety and Health (NIOSH) and the Mine Safety and
Health Administration (MSHA) in 2001. This need versus downhole mixing) (Trevits and Urosek, 2002).
resulted in a NIOSH research project (NIOSH, 2001). The second phase of the work involves the remote
In addition, MSHA agreed to serve as a technical part- construction of mine seals. The research was con-
ner in this effort. The first phase of the work involved ducted at NIOSH’s Lake Lynn Experimental Mine
the qualitative review of existing technology used to (LLEM) located approximately 60 miles southeast of
remotely construct mine seals. The review included Pittsburgh, PA. The LLEM is a world-class, highly
materials used to construct mine seals, including sophisticated underground facility where large-scale
cement and polyurethane foam, and an analysis of explosion trials and mine fire research is conducted
the available material mixing technologies (surface (NIOSH, 1999) (figure 1).

367

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 2. Underground layout of the seal construction site.
Howard Concrete Pumping Company (Howard) of
Cuddy, Pennsylvania and GAI Consultants, Inc. (GAI)
of Monroeville, Pennsylvania served as research part-
ners with NIOSH in this effort. This paper describes
the development of novel grout-based technology for
remote mine seal construction, evaluation of the mate-
rials used, construction practices, and follow-up test-
ing. An in-depth discussion of this work is described
by Gray et al. (2004).
2 SEAL DESIGN CONSIDERATIONS
The objective of this work was to develop a specialty
grout product and a method for placing the product
through a borehole into a mine opening to form a
mine seal. There were several additional factors that
were included in the engineering design process. These
factors are listed as follows:
• The methodology developed must be capable of
being deployed quickly.
• The mine seal must be rapidly installed.
• The seal material used must be locally available.
• The grout material must be of a consistency to allow
placement in a free space without excessive flow if
the mine is open and unobstructed but must also
be capable of filling a mine opening containing
roof fall debris, cribbing, equipment or conveyor
structures.
• The grout and the methods of application must
facilitate mine roof-to-floor and rib-to-rib closure.
• The seal must be strong enough to withstand the
force of a mine explosion (up to 20 psi).
Previously, a 6-in diameter cased borehole was com-
pleted in the first cross-cut between the B and C Drifts
of the LLEM and it was determined that this bore-
hole was suitable for the seal construction research
(figure 2).
The thickness of the overburden in the area of the
borehole is 197 ft. The cross-cut in the mine measured
368
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 3. Directional elbow for bulk fill placement.
19 ft wide, 40 ft long and 7 ft high, with a mine floor
slope gradient of 1.13 percent. A second borehole,
located about 30 ft away, was available for viewing the
mine seal installation using a downhole video cam-
era. Testing conditions for this technology was not
designed as a “blind test” so in-mine to surface com-
munication was facilitated through the use of a mine
pager phone system.
3 IN-MINE MATERIAL PLACEMENT
TECHNOLOGY
Prior to installing the mine seal at LLEM, a model mine
opening was constructed at Howard’s facility for test-
ing and direct observation of the performance of the
downhole nozzle and pumping equipment. The model
mine opening consisted of a small excavation in a hill-
side. The roof of the model mine was formed using
crane mats so a drill rig could be located over the mine
void to hold the pipe for the downhole equipment.
Two series of tests were performed at the model mine
along with an initial test at the LLEM before the final
placement strategy and grout mixture was developed.
Changes were made to the cement content, admixtures
and additive ratios to improve stickiness, time-of-set
and application uniformity. Some laboratory work was
also conducted to improve the grout blends by modi-
fying admixtures and additive ratios. After each test,
modifications were made to the materials, equipment
and equipment usage.
The final technique developed included a special-
ized directional elbow for directional placement of
bulk fill material (figure 3) and a spray nozzle for
material to address the remaining open areas in the
mine void (figure 4).
The specialized nozzle required the use of two
strings of pipe (one inside of the other) to convey two
streams of material to the nozzle. The spray nozzle
permitted the blending of the two-part grout accel-
erator mix while allowing sufficient air velocity to

Figure 4. Spray nozzle.
transport the grout to the mine roof-and-rib areas. The
bulk grout was pumped to the borehole using a positive
displacement pump and compressed air. The sprayed
grout was moved to the borehole using a conventional
grout pump and compressed air.
4 GROUT MATERIAL
Constructing an effective mine seal through a single
borehole is a difficult engineering challenge.The grout
mixture cannot be too fluid or it will merely flow away
from the borehole. If the grout mixture is too stiff, it
will tend to build quickly forming a mound at the bot-
tom of the borehole and will not flow and fill the mine
roof-rib areas. It was decided that the first material
to be placed in the mine would be a bulk fill mate-
rial designed to fill most of the open space. This was
also the least costly component of the fill material and
would help to lower the overall cost of the seal. The
bulk fill material used a mixture consisting of fly ash,
Portland cement, and 2A (3/4-in minus) crushed lime-
stone aggregate. A conventional concrete admixture
was used to accelerate the set of the grout. The mate-
rial was blended to achieve a pumpable mixture that
had adequate strength and rapid setting properties. The
amount of fly ash added was sufficient to produce a
mix that could be pumped to the borehole, travel down
the borehole without segregation and provide a moder-
ate degree of flowability. Once the grout was in-place,
the aggregate would provide sufficient shear resistance
for the grout to be somewhat immobile until the mix
set. Typical initial set time for this mixture could be
achieved in 15 to 20 minutes and would support foot
traffic in 30 to 45 minutes.
The second material to be used to fill any remain-
ing open space above the bulk fill along the roof-rib
line was a two-part grout blend that was developed
with the assistance of Master Builder’s Concrete Prod-
ucts Laboratory in Cleveland, Ohio. The basic grout
369
Copyright © 2006 Taylor & Francis Group plc, London, UK
was to be a blend of ASTM Class-F fly ash and
Portland cement. The initial testing of the grout indi-
cated that a conventional shotcrete accelerator would
not produce sufficient stiffening in the desired time
frame. Additionally, it did not exhibit suitable rheo-
logical and hardening properties required for the grout
application. Further testing determined that Master
Builder’s TCC system was more effective in provid-
ing the desired grout characteristics than conventional
admixtures. The Master Builder TCC System is made
up of two-parts. Part A improves the pumping char-
acteristics and provides a reaction platform for Part B
and is mixed with the grout before it is pumped into
the borehole. Part B is a liquid, high-performance
shotcrete and grout accelerator that reacts with Part A
to create an immediate stiffening of the grout. Part B
is added to the grout mixture using the spray nozzle
(positioned at the mine level) using the stream of air
that also transports the grout to the mine roof-and-rib
surface. The reaction between the Part A and Part B
admixtures essentially provides the initial stiffening
through a flocculation process that is unrelated to the
chemical hydration of the cement products in the grout.
Therefore, a concrete accelerator was also added at the
nozzle to accelerate the hydration process. The addi-
tion of the accelerator along with the cement content of
the grout facilitated rapid strength development of the
in-place grout spray. To improve the stiffening prop-
erties of the grout and produce the required stickiness
for the grout spray to adhere to the mine roof-and-
rib areas, the water content of the mix was adjusted
while retaining the fluidity and pumpability of the mix
through the addition of a high-range water-reducing
additive.
As the work on the seal material development
progressed, it became apparent that the uniform, con-
sistent blending of the constituents in the sprayed grout
was critical to the grout performance. The final portion
of the grout mix design work focused on a sensitivity
study that identified the grout’s reaction to deviations
in the blending process. It was concluded that it would
be necessary to very finely meter the ingredients in
the grout mix to achieve the desired performance from
the sprayed grout. After a series of field and laboratory
tests, adjustments were made to the equipment used to
control material feed and a significant improvement of
the material mix was achieved by the GAI engineers.
5 MINE SEAL CONSTRUCTION
Pumping of the first part of the remote seal (bulk mate-
rial) began using a sand, fly ash and cement mixture.
This material was pumped into the mine opening using
the directional elbow. The bulk material was pumped
in a series of lifts to fill the mine most of the opening.
Pumping was terminated after approximately 55 yd3 of
 Figure 5. View of bulk fill placement for mine seal.
Figure 6. Use of spray nozzle during mine seal construction.
material had been placed into the cross-cut. It should
be noted that that communication with underground
personnel was allowed to orient the directional elbow
and complete the construction of the base. Under-
ground examination revealed that seal material was
placed to within 1.5 ft of the mine roof below the bore-
hole and within 2.5 to 3 ft of the mine roof near the rib
areas (figure 5).
It was decided to remove 6-in of material near the
top of the seal to allow sufficient room to test the capa-
bility of the spray nozzle. After conducting a 10 yd3
surface test of the seal mixture (fly ash, cement and
accelerators), a dual string of drill pipe and casing
affixed with the spray nozzle was then placed into the
6-in diameter borehole in preparation for the second
part of the seal construction. Once the nozzle pene-
trated the mine opening, seal material was sprayed in
a back-and-forth motion along the mine rib areas to
fill in the gaps. Interaction between observers under-
ground and engineers on the surface ensured that the
nozzle was aimed in the proper direction. Good mine
roof-and-rib contact was made with the sprayed mate-
rial. The problematic corner areas at the mine roof-rib
intersection were filled before the grout began to build
up and migrate towards the spray nozzle (figure 6).
370
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 7. View of completed mine seal.
Filling of the remaining area near the borehole was
accomplished by lowering the spray nozzle into the
wet material below the nozzle and then rotating the
operating spray nozzle through a 360 degree arc. Even-
tually, the material built-up around the nozzle and
closed the mine opening (figure 7). In all, a total of
22.5 yd3 of sprayed material was used to close the mine
opening. An underground examination showed that the
mine seal material (both bulk and sprayed material)
had flowed about 12 ft from the borehole towards the
B-Drift and only about 9 ft from the borehole towards
the C-Drift. The shape of the seal approximated a trun-
cated pyramid whose base measured 19 ft wide (the
width of the cross cut) by 21 ft deep and whose top
measured 19 ft wide (the width of the cross cut) by 3
to 5 ft deep.
6 MATERIAL AND MINE SEAL TESTS
Unconfined compressive tests were conducted on 3-in
diameter cylinder samples (cylinder area – 7.07 in2 )
that were collected during seal construction. Samples
were collected underground as the material was being
placed in the mine void. The results of the tests are
shown in Table 1. The compressive strength of the bulk
fill material is substantially higher than that of the
sprayed fill material. The reason for the lower com-
pressive strength of the sprayed material is that the
sprayed mix does not contain sand and had air bubbles
trapped in the mixture from the mine seal material
placement process.
Although the major thrust of this research effort was
aimed at development of material mixes and mine seal
construction techniques, the benefits of constructing
the seal at the LLEM included the option of testing the
seal’s ability to confine mine air and also to withstand
the forces of a mine explosion. Air leakage tests were
conducted by building a frame on one side of the mine
seal and covering that frame with brattice cloth. Next
 Table 1. Results of compressive strength tests on cylinder
samples.
Bulk fill material Spray fill material
Compressive Age, days Compressive
Age, days strength, psi strength, psi
9 2403 5 230
28 3818 7 270
– – 28 468
– – 56 765
Table 2. Results of air-leakage tests on mine seal.
Differential pressure across
the seal, inches of water gage 0.52 1.05
Air-leakage rate, ft3 /min 252 322
an opening was made in the brattice cloth the size of an
anemometer to facilitate air velocity measurements.
Once this work was completed, air flow in the mine
was adjusted to produce a desired differential pressure
and the air leakage through the seal was measured. Air
leakage tests were conducted on the mine seal and the
results are shown in Table 2 (Weiss, 2003).
Prior to conducting the air leakage tests, several
holes (on the order of about 1 inch diameter) were
observed in the seal near the mine roof area. Therefore,
the air leakage values observed in the table were not
totally unexpected.
To conduct the explosion test, a known quantity of
methane gas was injected in the end of the C-Drift
near the cross-cut where the seal was installed. This
area was temporarily closed with a frame and brattice
cloth to confine the gas. The gas was diluted with air to
achieve an explosive concentration. The gas was then
ignited producing an explosion. The mine seal with-
stood a static load pressure of 18 psi from the explosion
with no visible signs of damage (Weiss, 2003).
7 RESEARCH FINDINGS AND
RECOMMENDATIONS
The overall objective of the work was to determine if
a mine seal could be constructed remotely from the
ground surface. This objective was achieved as a seal
was successfully built through a borehole and was
confined to the cross-cut of the mine opening. The
technology used to build the seal was tested and
the correct material mix design was developed. The
results of follow-up testing showed that a strong and
robust seal was constructed as required in the design
Copyright © 2006 Taylor & Francis Group plc, London, UK
constraints. The issue of air-leakage can be addressed
by slowing the rotation of the spray nozzle to allow
for a more substantial build-up of seal material. As
an additional remedy, it may be also be possible to
insert the spray nozzle into the observation borehole
and spray the entire face of the seal to close and fill
any remaining holes.
Results of the work to date suggest that this remote
seal construction system may have merit for isolating
a mine fire. This technique however does require addi-
tional trials since considerable communication with
the subsurface personnel was needed to achieve rib-to-
rib and roof-to-floor closure. One of the fundamental
keys to successful in-mine construction is the abil-
ity to directly observe the progress of construction.
Because this was a research and demonstration project,
communication between the surface operation and the
1.52 underground seal location was permitted. This will
426 not be the case when a mine fire occurs. Additional
research is therefore planned to further refine the con-
struction method. A mine seal should be constructed
at the LLEM without voice communication with the
surface. The only means of observing the progress
of construction should be via the nearby borehole
equipped with a downhole video camera with suffi-
cient resolution capabilities and lighting. Experience
gained during this work also suggests that a downhole
laser or radar imaging device should be developed that
offers real-time imaging and is capable of penetrating
smoke, dust or the fog that forms in the mine opening
as the seal material begins to set.
A 6-in borehole was used during the trials at LLEM
and the downhole equipment was designed to meet
this need. The issue of working with this equipment
in smaller diameter boreholes should be addressed
along with the fact that deeper overburden depths will
undoubtedly be encountered. Perhaps an additional
spray nozzle should be constructed to facilitate remote
seal construction in small-diameter boreholes.
Finally, it is suggested that this technology should
be further evaluated through construction of a mine
seal at LLEM in a mine entry that is obstructed with
debris (roof fall material) and mine structures (pos-
sibly cribbing, track, or conveyor structures). This
approach will test the ability of the seal material to
flow around obstructions and still form a seal while
closely matching the conditions most likely found in
an underground mine.
ACKNOWLEDGMENTS
The authors would like to recognize John Urosek and
Clete Stephan, MSHA, for their input and support
throughout the design and conduct of this research
project. Special thanks are also made to Alex Smith,
Eric Weiss and the NIOSH staff at the LLL facility for
371
 their professionalism, dedication, and assistance in the
conduct of this research effort.
REFERENCES
DeRosa, M.I. 2004. Analysis of Mine Fires for all U.S.
Underground and Surface Coal Mining Categories: 1990–
1999. NIOSH Information Circular 9470.
Gray, T.A., Trevits, M.A., Crayne, L.M., and Gloglowski. P.,
2004. Demonstration of Remote Mine Seal Construction.
2004 SME Annual Meeting Denver, CO February 23–25
Preprint No. 04–194.
National Institute for Occupational Safety and Health. 2001.
A Compendium of NIOSH Mining Research 2002. DHHS
(NIOSH) Publication No. 2002–110.
372
Copyright © 2006 Taylor & Francis Group plc, London, UK
National Institute for Occupational Safety and Health. 1999.
Lake Lynn Laboratory (flyer), DHHS (NIOSH) Publica-
tion No. 99–149.
Trevits, M.A. and Urosek, J.E. 2002. Technology for Remote
Mine Seal Construction. 2002 SME Annual Meeting
Phoenix, AZ February 25–27 Preprint No. 02-185.
Weiss, E.S. 2003. Supervisory Mining Engineer, Disaster
Prevention and Response Branch, NIOSH. Personal Com-
munication October 23.
Weiss E.S. 2003. Supervisory Mining Engineer, Disaster
Prevention and Response Branch, NIOSH. Personal Com-
munication December 10.
 7. Coalbed/coal mine methane

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Prediction of longwall methane emissions and the associated consequences

of increasing longwall face lengths: A case study in the Pittsburgh Coalbed

S.J. Schatzel, R.B. Krog, F. Garcia & J.K. Marshall

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute for
Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

J. Trackemas
Pennsylvania Services Corporation, Foundation Coal Company, Waynesburg, PA, USA

ABSTRACT: In an effort to increase productivity, many longwall mining operations in the U.S. have continually
increased face lengths. Unfortunately, the mining of larger panels may increase methane emissions. The National
Institute for Occupational Safety and Health (NIOSH) conducted a mine safety research study to characterize
and quantify the methane emissions resulting from increasing face lengths in the Pittsburgh Coalbed. The goal of
this research effort was to provide the mine operator with a method to predict the increase in methane emissions
from the longer faces for incorporation of additional methane control capacity into the mine planning process,
if necessary. Based on measured methane emission rates of 0.066 m3 /s (140 cfm) for a 315 m (1032 ft) face,
projected longwall face methane emission rates were 0.090 m3 /s (191 cfm) for a 366 m (1200 ft) face, 0.106 m3 /s
(225 cfm) for a 426 m (1400 ft) face, and 0.124 m3 /s (263 cfm) 488 m (1600 ft) face.

1 INTRODUCTION gob areas represents an increase in permeability. Most

Continuous enhancements in longwall mining equip-
ment have significantly improved face advance rates.
This increasing longwall advance rate has generally
outpaced the continuous miner development section
advance rate. One potential solution is to decrease
the relative amount of development mining required
by increasing the size of longwall panels, in particu-
lar face lengths. However, increases in longwall face
length can create problems such as increased cumula-
tive face methane emissions and increased potential
for methane-related production delays (Krog et al.
2006). This may be further exacerbated by the com-
plex airflow movements along the face itself. Although
airflow movement along the longwall face is generally
assumed to be linear, evidence of exchanges between
face and gob atmospheres have been noted (Balusu
et al. 2001, Wendt & Balusu 2001). Other researchers
have suggested an increased level of opportunity for
air exchanges between face and gob regions result
from resistant roof units and associated void spaces
behind the longwall face (Noack 1998, Balusu et al.
2001, Wendt & Balusu 2001).This hypothesis suggests
that under certain roof conditions, increased longwall
face lengths could also increase the void space behind
the face. This increase in void space in the face and
prior research regarding gob permeability addresses
airflows in the bleeder entries between mined-out pan-
els (Mucho et al. 2000) or characterizations within the
longwall gob (Sing & Kendorski 1981, Brunner 1985,
Karacan et al. 2005) and not the near face locations.
This research effort was designed to address issues
in predicting and configuring adequate longwall face
ventilation when face lengths are increased. One goal
of this study was to predict the magnitude of the
increase in the face methane emission rates due to an
increase in face length. A second goal was to char-
acterize air and methane movements along the face,
as well as the influence of the mining direction of
the shearer and production delays, including methane
delays, on face emission rates. With respect to the
methane delays, an effort was made to identify any
potential patterns or causes relative to how, when,
and where they occur. A final goal was to learn how
significant changes in the geology, mining practices
(e.g. methane drainage) and mining conditions might
affect methane emission rates. Achieving the goals of
the study would increase the industry’s understanding
of longwall face emissions and associated ventilation
issues. Ultimately, the most important outcome sought
in this research project was to enhance miner safety in
the underground mining environment by reducing the

375

Copyright © 2006 Taylor & Francis Group plc, London, UK

 frequency of hazardous accumulations of methane on
longwall faces.
2 METHODOLOGY
2.1 Study site
The field site for this study was a longwall coal mine
operating in the Pittsburgh Coalbed in southwestern
Pennsylvania. The mine operated three continuous
miner development units and one longwall system.The
annual production from the mine is approximately 5.9
million t (6.5 million short tons) per year, of which
5.0 million t (5.5 million short tons) are produced
from the longwall mining system. Coal thickness at
the mine ranged from 2.0 m (6.5 ft) to 2.4 m (8.0 ft).
The depth of cover in the study area ranged from
183 m (600 ft) in the headgate area to 244 m (800 ft)
in the tailgate area. The dimensions of the study panel
were 3250 m (10,650 ft) long by 315 m (1032 ft) wide
(Fig. 1) (dimensions of the outlined coal block). The
study began with 1250 m (4100 ft) of panel length
remaining. This panel was the third panel mined in a
longwall district containing five longwall panels. The
longwall bleeder system was ventilated by the use of
a centrifugal bleeder fan, and four vertical gob venti-
lation boreholes per panel placed at regularly spaced
intervals of approximately 610 m (2000 ft) to provide
additional methane control capacity. Four horizontal
in-seam methane drainage boreholes were also present
adjacent to the gateroads of the study panel.
2.2 Experimental procedure
The basic methodology for this study was first devel-
oped by Diamond and Garcia (1999) for a longwall
face emissions investigation at two mines operating in
the Pocahontas No. 3 Coalbed, VA. The goal of the Dia-
mond and Garcia (1999) study was to predict methane
emissions rates for 305 m (1000 ft) faces based on face
Figure 1. Longwall face emission study area.
376
Copyright © 2006 Taylor & Francis Group plc, London, UK
emissions monitoring on a 229 m (750 ft) wide face
at the VP-1 and VP-3 mines (Fig. 2). In that study,
continuous methane emissions monitoring was con-
ducted near the headgate and tailgate corners. The
face was divided into three equal segments of about
76 m (250 ft). Average cumulative methane emissions
data for each of the three segments were plotted as a
function of face length. Curves were fit to the actual
emission data and were then extrapolated to the 305 m
(1000 ft) face widths to predict methane emission rates
on the longer faces. The data showed that the two
mines would likely experience significantly different
emission rate consequences in response to increas-
ing the face lengths by about 76 m (250 ft). This was
due to variations in mine design and methane control
practices between the two mine sites.
For this study in the Pittsburgh Coalbed, continu-
ously recording methane monitors were installed along
the 157 shields of the longwall face. The monitors
were sampled at 5 sec. intervals, and 1-minute average
methane concentrations were recorded by data loggers.
Airflow measurements were made at various locations
on the face multiple times per shift to provide ven-
tilation airflow measurements for the calculation of
emission rates based on the methane concentration
data. A production time study consisting of shearer
location on the face (recorded as shield numbers) and
shearer mining direction (head-to-tail or tail-to-head)
was also conducted throughout the three days of the
face emissions monitoring. One shift was monitored
on each of the three days of the study. The duration
and cause of all production delays, the face position
at the start and end of each shift, the presence of
any discernable geologic discontinuities or conditions
encountered along the face, and any other pertinent
data or observations were also noted as part of the
time study record.
To analyze the movement of methane emissions
in the longwall ventilation airflow, the face was
divided into four segments of equal length. Thus each
Figure 2. Prediction of methane emissions at greater long-
wall face lengths (Diamond & Garcia 1999).
 face segment measured approximately 79 m (258 ft)
(Fig. 3). Methane emission rates were determined for
each face segment of each pass of the shearer using the
associated methane concentration, ventilation airflow,
and time study data. For this study, three continuously
recording methane sensors were used for face methane
monitoring. Since the principal focus of this study was
longwall face ventilation and methane emissions, the
decision was made to position the monitors away from
the headgate and tailgate corners to avoid the inclu-
sion of ventilation air in these areas which did not
traverse the face. The methane sensors were installed
at shields 20, 80, and 145 (Fig. 3). The distances from
the headgate corner to each monitoring location were
35, 158, and 290 m (114, 517, and 953 ft) for the 315 m
(1032 ft) face (Fig. 3).
To analyze the acquired methane emissions data, the
longwall face segments were designated as 1, 2, 3, and
4, as numbered from the headgate for head-to-tail (H-
T) passes, and were designated as 1a, 2a, 3a, and 4a, for
tail-to-head (T-H) passes as numbered from the tailgate
as shown in Figure 3. Average methane emission rates
were computed for each of the face segments using
simple algebraic formulas to determine the methane
volume emitted, which was then divided by the time
required to mine each face segment. Consequently, a
total of four methane emission rates were computed
for each pass of the longwall shearer in each direction.
A sample face segment methane emission rate calcu-
lation is shown in Table 1. This sample shows methane
emissions data from a shearer pass through face seg-
ment 3 with no delays. Methane emissions measured
during the mining of this segment are determined as
the difference in emission rates at two continuously
recording methane sensors, while the shearer position
for each minute of mining is shown relative to the long-
wall shields. From this data, average methane emission
rates can be determined for each segment.
Gob
All three methane sensors were used in this analy-
sis, with the nearest downstream methane sensor used
to compute the methane emission quantities. The tran-
sit time for ventilation air to flow along the longwall
face was determined, and the duration of airflow from
the end of a face segment to the sensors ranged from
0 seconds to about 1 minute in computing methane
emission quantities. This airflow transit time along
the face was ignored in the computations for the fol-
lowing reasons: 1) even the longest transport time
(one-minute) to a methane sensor is small compared to
the time required for the shearer to mine the segment,
2) according to the protocol, the methane produced in
the mining of the face segment was measured after the
shearer actually passed the methane sensor to allow
time for the instrument to respond to methane in the
airstream, and 3) the recorded methanometer data were
averaged over each minute of monitoring.
Due to the location of the recording methanometers
(Fig. 3), face segments 1, 4, 1a, and 4a are actually
shorter than the specified 79 m (258 ft) length. In seg-
ments 1 and 4a, all ventilation airflow passing across
the segment reaches one or more of the methane sen-
sors. For face segments 1a and 4, all of the ventilation
airflow does not reach the sensors (Fig. 3). Conse-
quently, for face segments 4 and 1a, some assumptions
were made based on general trends in the methane
emissions data to determine methane emission quan-
tities when the shearer was on the tailgate side of
the methane sensor located on shield 145 (Fig. 3). In
applying this methodology, the following assumptions
have been made to project methane emissions to longer
faces: 1) the mine advance rate and the frequency of
mining delays will occur at about the same rate; 2)
face methane emissions are assumed to be constant
within each segment; 3) all sources of methane emis-
sions change at a constant rate with increased face
length; and 4) the solutions are site specific for the
Pittsburgh Coalbed, the ventilation configuration, and
the methane drainage systems applied at the study
mine site.

Table 1. Sample calculation of emissions data for face

290 Instrument 157 35
segment 3 acquired during day 1 of the face emissions study.
(953) locations (516) (114)
CH4
315 236 157 79 0
(1032) (774) (516) (258) Time Recorded CH4 CH4 emission
Face segment endpoints, distances from headgate cutting shearer emission emission during pass
corner, m (ft) face position, rate at start rate at end in face
segment, shield (shield 80), (shield 145), segment 3,
4 3 2 1(H-T) h:min number (m3 /s) (m3 /s) (m3 /s)
(T-H)1a 2a 3a 4a
Face segment identities
20:39 88 0.021 0.079 0.059
Headgate
Tailgate

20:40 – 0.021 0.099 0.078

20:41 – 0.021 0.099 0.078
20:42 – 0.021 0.099 0.078
20:43 116 0.021 0.099 0.078
Figure 3. Pittsburgh Coalbed longwall face emission study 20:44 120 0.041 0.099 0.058
site with instrument locations and face segment identities.

377

Copyright © 2006 Taylor & Francis Group plc, London, UK

 3 RESULTS AND DISCUSSION The cumulative methane emission rates for the face
segments for each of the three days of the study are
3.1 Analysis of the methane emission data from the shown in Figure 5 and Table 2. Face methane emis-
315 m (1032 ft) longwall face sion rates were consistently higher by about 41% in
the H-T passes than in the T-H passes, in part because
On the study panel, belt air was used on the longwall
of the longwall face mining sequence which produced
face (Fig. 4). After the shearer completed a T-H pass,
a much longer “wedge” cut towards the tailgate than
sumped in, and began an H-T pass, methane carried to
towards the headgate entry and the slower shearer cut-
the face from the belt line made up the great majority
ting rate in the T-H passes due to panline loading.
of the methane measured at the shield 20 methanome-
Methane emissions were lowest on day 2, [0.042 m3 /s
ter. Under these circumstances, the data showed that
(88 cfm)] of the methane emission monitoring study
methane emissions at the shield 20 location frequently
and highest [0.061 m3 /s (129 cfm)] on day 3 at shield
exceeded those at the shield 80 location (i.e. non-linear
145. One contributing factor to the changes in methane
airflow along the longwall face). Consequently, for
emission rates over the course of the study is thought
analytical purposes, face emissions were assumed to
to be the proximity of gob vent boreholes (GVB’s)
be zero for the first 20 shields of an H-T pass (up to
located near the tailgate side of the panel. The study
35 m, 114 ft from the headgate corner). Although this
area was approaching the maximum distance between
may not be a totally accurate assumption, data analy-
gob vent boreholes (GVB’s), where the next borehole
sis indicated that the level of methane emissions in this
would be intercepted two days after the study was
portion of the face was quite low, usually measured as
completed. This may account for a general increase in
between 0.0 and 0.021 m3 /s (0 to 44 cfm). This portion
of the longwall panel is near the previously developed
headgate entry, so there is a significant opportunity Ft
for the migration of methane out of this portion of the 0 200 400 600 800 1000
0.09
unmined solid coal block prior to mining.
Determined Cumulative Methane (CH4)

Day 1 H-T 180

0.08
The sums of gas emissions from the eight face seg- Day 1 T-H 160
0.07
ments for the H-T and T-H passes were normalized Day 2 H-T
140
Emissions, m3/s

0.06 Day 2 T-H

to match actual gas emission totals measured at the 120

Cfm (CH4)
0.05 Day 3 H-T
shield 145 methane emissions monitor. A total of 24 100
Day 3 T-H
passes were analyzed and three of those passes were 0.04 80
not complete data sets. No passes were eliminated from 0.03 60
the data set due to production delays. A total of 27 0.02 40
methane-related delays were noted during the study: 0.01 20
4 on the first day, 11 on the second day, and 12 on 0.00 0
0 50 100 150 200 250 300 350
the third day. Methane delays were typically of short Distance From Headgate Corner, m
duration, averaging about 7 minutes. Consequently,
separation of data with methane delays to view only Figure 5. Cumulative longwall face methane emissions for
data showing uninterrupted passes was not possible the three days of monitoring. Showing the face segments
because of the high percentage of passes affected by endpoints.
short duration methane delays.
Table 2. Cumulative face segment emission rates deter-
mined for each day of study.
Gob
Face
S

segments, Day 1 values Day 2 values Day 3 values

H-T m3 /s (cfm) m3 /s (cfm) m3 /s (cfm)
Mining Direction
1 0.007 (14) 0.010 (22) 0.020 (42)
Fresh Air 2 0.016 (34) 0.016 (34) 0.053 (112)
Return/Bleeder Air 3 0.050 (106) 0.035 (74) 0.067 (143)
Headgate

4 0.081 (171) 0.046 (98) 0.088 (187)

Tailgate

Face Air
Mining Direction
Face
setments,
T-H
1a 0.045 (96) 0.037 (78) 0.069 (147)
Belt entry 2a 0.036 (77) 0.032 (67) 0.059 (125)
3a 0.008 (17) 0.015 (32) 0.051 (109)
4a 0.000 (0) 0.009 (19) 0.027 (57)
Figure 4. Ventilation configuration at study site.

378

Copyright © 2006 Taylor & Francis Group plc, London, UK

 methane concentrations prior to the next GVB coming
on line, GVB 46-4 (Fig. 6).
The measured average daily methane emission
rates varied from about 0.046 m3 /s (98 cfm) to about
0.088 m3 /s (187 cfm) on head-to-tail passes over the
three days of the study (Fig. 5) (Table 2). Produc-
tion delays due to increasing methane concentration
were most common when the shearer was near the
tailgate side of the longwall face of the H-T pass,
but began occurring with increasing frequency at loca-
tions nearer the headgate side of the face as the study
progressed over the three days. An analysis of methane
emissions from individual face segments (Table 3)
suggests potential causes of this behavior.
Average face methane emissions rates were cal-
culated for each of the four face segments and each
pass direction. These values varied during each pass
and varied over successive days of the study. This
variation in behavior can be viewed more clearly in
Figure 7, which shows the average methane emissions
from each segment alone (H-T and T-H combined data)
during the three days of monitoring (Table 3). Figure 7
shows that the methane emissions produced during
passes on day 1 were relatively low [from 0.003 to
0.009 m3 /s (7 to 18 cfm)] in the segments nearest to
Figure 6. Actual portions of the study panel monitored for
face methane emissions and associated methane drainage
borehole locations.
Table 3. Face segment emission rate data shown for com-
bined averaged H-T, T-H passes.
Face Day 1 Day 2 Day 3
segments m3 /s (cfm) m3 /s (cfm) m3 /s (cfm)
1, 4a 0.003 (7) 0.010 (20) 0.023 (49)
2, 3a 0.009 (18) 0.006 (13) 0.029 (61)
3, 2a 0.031 (66) 0.018 (37) 0.011 (24)
4, 4a 0.020 (42) 0.008 (18) 0.016 (33)
379
Copyright © 2006 Taylor & Francis Group plc, London, UK
the headgate (1, 2, 3a, 4a). Emissions increased sig-
nificantly in segments 3 and 2a [0.031 m3 /s (66 cfm)],
and diminished somewhat from that level toward the
tailgate in segments 4 and 1a. On day 2, methane
emission rates in the near headgate segments 1 and
4a were higher than on day 1 [0.010 m3 /s (20 cfm)],
and then decreased in the next segments, 2 and 3a
[0.006 m3 /s (13 cfm)]. Similar to day 1, methane emis-
sions increased in face segments 3 and 2a [0.018 m3 /s
(37 cfm)] and then decreased slightly in the face seg-
ments towards the tailgate, i.e. face segments 4 and
1a [0.008 m3 /s (18 cfm)]. On day 3, methane emis-
sions were much higher [0.023 m3 /s (49 cfm)] near the
headgate face segments 1 and 4a than during the prior
two days of the study (Table 3). The methane emissions
rate increased in the next face segments [0.029 m3 /s
(61 cfm)] 2 and 3a, but then decreased significantly
in face segments 3 and 2a [0.011 m3 /s (24 cfm)], with
a slight increase [0.023 m3 /s (49 cfm)] in the tailgate
segment face segments 4 and 1a (Fig. 7).
Methane drainage via horizontal boreholes near the
face emission study area may have contributed to the
variable methane emissions characteristics observed
on the face. Two horizontal boreholes produced gas
from the Pittsburgh Coalbed on the longwall panel near
the study area. One hole was drilled on each side of
the panel, and the holes were oriented parallel to the
gateroad entries, about 30 m (98 ft) from the respective
gateroads (Fig. 6). The holes were drilled towards the
advancing face, and terminated in the study area. Prior
to interception by mining, the holes were filled with
water.
On day 2 of the study, the presence of a horizontal
borehole on the face was noted near the tailgate side of
the panel at shield 140. On day 3 of the study, the inter-
ception of the horizontal borehole near the headgate
at shield 23 was observed. It is likely that coalbed
Distance, ft
0 200 400 600 800 1000
0.04 Day 1 Avg
Day 2 Avg 80
Day 3 Avg
Methane emissions, m3/s
3 Day Avg H-T pass
0.03
3 Day Avg T-H pass
60
Methane, cfm
0.02 40
0.01 20
0.00 0
0 50 100 150 200 250 300 350
Distance from headgate corner, m
Figure 7. Average methane emission rates for each day of
study determined for each face segment (the zero emission
value assumed for the start of segment 1 and the end of
segment 4a have been omitted).
 methane production from the first borehole encoun- Ft
tered near shield 140 on the tailgate side of the face 0.14
0 200 400 600 800 1000 1200 1400 1600

Determined Cumulative Methane (CH4)

had decreased the methane content of the Pittsburgh 3 Day Avg all passes
y = 0.0002784x - 0.0123
280
0.12 260
Coalbed in the vicinity of the borehole. Consequently, 3 Day Avg H-T pass
R2 = 0.983 240
after the borehole was intercepted on day 2 of the 0.10 3 Day Avg T-H pass 220

Emissions, m3/s
200
study, methane emissions dropped towards the tailgate Trend line A (3 Day Avg

Cfm (CH4)
180
0.08 All passes)
160
side, as shown by the comparison of face segments 4 Trend line B (3 Day Avg
140
H-T pass)
and 1a on day 1 [0.020 m3 /s (42 cfm)] versus those on 0.06
y = 0.0002329x - 0.00783 120
R2 = 0.993 100
day 2 [0.008 m3 /s (18 cfm)] (Fig. 7). A similar drop in 0.04 80
methane emission rates was not observed in face seg- 0.02
60
40
ments near the head gate from day 2 and day 3 when the 20
borehole was intercepted near shield 23 (Table 3). It 0.00
0 100 200 300 400
0

is not known if methane emissions would have dimin-
ished in the region of face segments 1 and 4a if the
study had been continued for another day. Not all of
the horizontal boreholes have the same time to produce
gas prior to interception by longwall mining.
Production delays also affected face methane emis-
sion rates. Production delays, including those due to
high methane emissions result in lower calculated pass
segment methane emission rates because the time to
complete the pass segment increases while the long-
wall face equipment is idle. Therefore, the increased
number of methane-related production delays on day
3, and to a lesser extent on day 2, resulted in lower
average methane emission rates on some individual
pass segments, particularly on tailgate side face seg-
ments 4 and 1a (Fig. 7). Although day 3 of the study
produced the highest cumulative methane emission
rates for complete passes, these values would have
been even higher had the methane-related delays not
occurred, which resulted in reduced average emission
rates on those individual pass segments.
3.2 Predicting methane emission rates for longer
longwall faces in the Pittsburgh Coalbed
Graphs of the cumulative measured average methane
emission rates for the 315 m (1032 ft) panel face seg-
ments in both the H-T and T-H directions are shown
in Figure 8. From these data, two least-squares linear
regression curves were calculated to predict methane
emissions for longer face lengths of 366 m (1200 ft),
426 m, (1400 ft) and 488 m (1600 ft) in the Pittsburgh
Coalbed. To create trend line A, all H-T and T-H passes
were averaged and then an overall average emission
pass plot was created. Trend line A was fit to this over-
all average emission pass data (Fig. 8). Trend line A
predicts face emission rates of 0.077 m3 /s (163 cfm),
0.091 m3 /s (193 cfm), and 0.106 m3 /s (225 cfm), for
face lengths of 366 m (1200 ft), 426 m, (1400 ft) and
488 m (1600 ft), respectively (Table 4). The equation
for trend line A is given (in metric units) and the
R2 value of the curve fit to the data is 0.993. It
should be noted that the emission rate for a 315 m
(1032 ft) face is based on a projection of the cumula-
tive emissions data from the face segments projected
Distance From Headgate Corner, m
Figure 8. Methane emission prediction curves for the Pitts-
burgh Coalbed; trend line A based on average emission data;
trend line B based on H-T pass emission data.
from shield 145 to shield 157 or 290 m (953 ft) from
the headgate corner using data from all of the long-
wall passes. Trend line B was fit to the H-T passes
only (Fig. 8). Since most of the delays occurred and
higher face methane emission rates occurred on the
H-T passes, this plot may be more representative
of problematic concentrations of face gas than trend
line A, which includes the generally lower T-H pass
face emission rate data (Fig. 8). Using trend line
B, with an R2 value of 0.983, to predict methane
emissions on longer longwall faces for lengths of
366 m (1200 ft), 426 m, (1400 ft), and 488 m (1600 ft)
yields 0.090 m3 /s (191 cfm), 0.106 m3 /s (225 cfm),
and 0.124 m3 /s (263 cfm), respectively, as compared
to the 0.066 m3 /s (140 cfm) value for the base 315 m
(1032 ft) panel (Table 4).
4 SUMMARY AND CONCLUSIONS
It should be noted that the results of this study to predict
longwall face methane emissions at increasing face
lengths is site specific and includes all delays which
occurred during the monitoring period. Due to the
higher rate of face methane emissions measured during
H-T passes, linear trend line B was fit to only the aver-
aged H-T passes, and is considered a better predictor
of hazardous methane emission conditions on longer
faces at the study site than an analysis using all of the
data. Using trend line B developed from the H-T pass
data, longwall face lengths of 366 m (1200 ft), 426 m
(1400 ft), and 488 m (1600 ft) were predicted with an
R2 value of 0.983. Using trend line B, the predicted
face methane emissions represent increases of 36%,
61%, and 88% for 366 m (1200 ft), 426 m (1400 ft),
and 488 m (1600 ft) faces respectively, as compared to
the base 315 m (1032 ft) face.
This methodology allowed for the computation of a
total of eight methane emission rates from head-to-tail

380

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 4. Face methane emission predictions based on projections made from
face segment methane emissions determinations.
Distance from
Face headgate corner
segments m (ft)
1, 4a 79 (258)
2, 3a 157 (516)
3, 2a 236 (774)
4, 1a 290 (952)
projection 315 (1032)
projection 366 (1200)
projection 426 (1400)
projection 488 (1600)
(H-T) and tail to head (T-H) passes. Analysis of
the monitoring data suggests that horizontal methane
drainage boreholes appeared to be effective in reduc-
ing methane emissions from the face in the vicinity
of the borehole near the tailgate. A similar borehole
intercepted near the headgate on day 3 did not appear
to reduce emissions, but a subsequent day of face
emissions monitoring was not conducted that could
have confirmed that boreholes impact. The data also
showed evidence of non-linear airflows, i.e. interac-
tions of the longwall face air with airflows behind
the shields. This was evident where belt air, carry-
ing methane from the transported cut coal, increased
concentrations near the headgate, while lower methane
concentrations were measured at the mid-face loca-
tion. Methane delays were most frequent when the
shearer was mining near the tailgate. Methane delays
on the longwall face increased in number from the first
to third day of the study as the longwall face reached
the approximate maximum distance from the nearest
operating gob vent borehole (GVB) prior to intercep-
tion of the next GVB, two days after the completion of
the study.
With the determination of multiple emission rates
over the length of a longwall face, the described
method provides insight into changing emissions rates
along the face due to all influential factors. The authors
suggest that for projected increased methane emis-
sions on wider panels, the methodology described
by Krog et al. (2006), using this same data set in a
companion paper, represents a more rigorous analy-
sis of the methane source components contributing to
the projected increased methane emissions. Thus, this
companion analysis provides additional insights on
appropriate control measures to address the projected
increases in methane emissions, although it should be
noted that both are empirical in nature and site spe-
cific. The longwall face methane emission predictions
made by Krog et al. (2006) using the same data set
are of greater magnitude than predictions appearing
Copyright © 2006 Taylor & Francis Group plc, London, UK
Average 3 day pass
combined H-T, T-H Average 3 day
data H-T passes only
m3 /s (cfm) m3 /s (cfm)
0.012 (26) 0.012 (26)
0.026 (56) 0.028 (60)
0.047 (99) 0.051 (107)
0.061 (129) 0.072 (152)
0.066 (140) 0.075 (159)
0.077 (163) 0.090 (191)
0.091 (193) 0.106 (225)
0.106 (225) 0.124 (263)
in this report. The primary reasons for differing face
emissions predictions are: 1) the predictions made by
Krog et al. were for peak load conditions and the pre-
dictions appearing here are for averaged data, and 2)
the predictions made by Krog et al. were for an “ideal-
ized” cut sequence with no delays where the methane
predictions in this analysis included all production
delays occurring in the data set.
Providing scientific predictions of the methane
emission consequences of implementing mine design
changes, such as increasing longwall panel face
lengths, was the primary goal of this research effort.
With these methane emission predictions, adequate
engineering and administrative controls can be imple-
mented in advance to ensure that the underground
workforce is not exposed to potentially hazardous con-
centrations of methane. Mine designers and ventilation
engineers have several options at their disposal to deal
with the higher emissions associated with increased
longwall face length, such as increasing face ventila-
tion airflow, utilizing or increasing methane drainage
capacity, and ventilation system changes, or as a last,
reducing the mining rate on the longwall.
5 DISCLAIMER
The findings and conclusions in this report are those
of the authors and do not necessarily represent the
views of the National Institute for Occupational Safety
and Health.
REFERENCES
Balusu, R., Deguchi, G., Holland, R., Moreby, R., Xue, S.,
Wendt, M. & Mallet, C. 2001. Goaf gas flow mechanics
and development of gas and sponcom control strate-
gies at a highly gassy coal mine. Papers and presen-
tations from the Australia-Japan Technology Exchange
Workshop, Dec. 2–4, 2001, Hunter Valley.
381
 Brunner, D.J. 1985. Ventilation Models for Longwall Gob
Leakage Simulation. Proceedings of 2nd US Mine Venti-
lation Symposium, September 23–25, Reno, NV : 655–663.
Diamond, W.P. & Garcia, F. 1999. Analysis and prediction of
longwall methane emissions: a case study in the Pocahon-
tas No. 3 Coalbed, VA. NIOSH Report of Investigations,
RI 9649.
Karacan, C.O., Diamond, W.P., Esterhuizen, E. &
Schatzel, S.J. 2005. Numerical analysis of the impact of
longwall panel width on methane emissions and gob gas
ventholes. Proceedings of the 2005 International Coalbed
Methane Symposium, May 18–19, 2005, Tuscaloosa.
Krog, R.B., Schatzel, S.J., Garcia, F. & Marshall, J.K. 2006.
Predicting methane emissions from wider longwall pan-
els by analysis of emission contributors. 11th U.S./North
American Mine Ventilation Symposium, Pennsylvania
State University, PA, USA, June 5–7, 2006, Rotterdam.
382
Copyright © 2006 Taylor & Francis Group plc, London, UK
Mucho, T.P., Diamond, W.P., Garcia, F., Byars, J.D. &
Cario, S.L. 2000. Implications of recent NIOSH tracer
gas studies on bleeder and gob ventilation design.
SME Annual Meeting and Exhibit, Feb. 28-Mar. 1, Salt
Lake City.
Noack, K. 1998. Control of gas emissions in underground
coal mines. International Journal of Coal Geology 35:
57–82.
Singh, M.M. & Kendorski, F.S. 1981. Strata disturbance
prediction for mining beneath water and waste impound-
ments. In Peng (ed.), Proceedings of the First Conference
on Ground Control in Mining, July 27–29, Morgantown:
76–89.
Wendt, M. & Balusu, R. 2001. CFD modeling of longwall
goaf gas flow dynamics. Proceedings of the Australia-
Japan Technology Exchange Workshop, Dec. 2–4, 2001,
Hunter Valley.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Predicting methane emissions from longer longwall faces by

analysis of emission contributors

R.B. Krog, S.J. Schatzel, F. Garcia & J.K. Marshall

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute for
Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, Pennsylvania, USA

ABSTRACT: As part of its mining health and safety research program, the National Institute for Occupational
Safety and Health (NIOSH) conducted a longwall methane emission and mining time study at a mine operating
in the Pittsburgh Coalbed to access the methane emission consequences of mining a longer face. The methane
emission contributors from the mining of a longwall face are: 1) gas released from the coal broken by the shearer,
2) gas emitted from the coal on the face conveyor, 3) gas emitted from the coal transported on the belt, and 4)
background gas emitted from the coal face and from the adjoining ribs. Based on the results of the study, a set
of site-specific mathematical formulas and constants were developed to characterize each of the four longwall
emission contributors. The mathematical formulas were then applied to longer longwall face mining scenarios
to predict the methane emissions from these faces.

1 INTRODUCTION face (Diamond & Garcia 1999, Schatzel et al. 2006).

The coal industry trend of mining wider longwall pan-
els is continuing throughout the world as gateroad
development mining struggles to keep ahead of pro-
duction on the longwall face. The mining of longer
longwall faces has the advantage of less development
mining per ton of coal mined on the longwall, but
can result in increased methane emissions. The ques-
tion asked by ventilation engineers is how much of an
increase in methane emissions can be expected with
the longer longwall face, and how can this be miti-
gated to maintain a safe underground workplace? An
increase in the ventilation airflow to dilute the expected
increase in methane emissions might not be possi-
ble because many modern longwalls are at, or near,
the reasonably practical airflow limits. Therefore, the
extra methane emissions will generally have to be han-
dled by a combination of increased ventilation airflow,
methane drainage and/or production management.
As part of its mining health and safety research pro-
gram, NIOSH conducted a detailed longwall methane
emission and mining time study at a mine operating in
the Pittsburgh Coalbed in southwestern Pennsylvania.
The initial goal of the experiment was to determine
the methane emissions from individual sections of the
longwall face and to extrapolate that data to estimate
emissions from a longer longwall face. Using this
approach, the face is divided into segments to char-
acterize how methane emission rates vary across the
A graphical solution is then used to predict face emis-
sions at longer face lengths based on emissions data
from shorter faces. One shortcoming of this approach
is that it assumes that the defined segments of the
longwall face will continue to emit methane at the
same rate with increasing face length. The assumption
that a 244 m (800 ft) long longwall face will have the
same methane emissions as encountered at the 244 m
(800 ft) point on a 305 m (1000 ft) long face is incorrect
because of the variations in coal production, methane
emission drainage characteristics and coal transport
factors.
A detailed and more meaningful analysis of the
methane sources and their individual contributions to
the total longwall methane emissions can be obtained
from methane concentration data collected at the
beginning and end of the longwall face, along with
the shearer location and other relevant ventilation and
mining data. The methane emission contributors from
the mining of a longwall face that were evaluated for
this study are: 1) gas released from the coal broken
by the shearer, 2) gas emitted from the broken coal on
the face conveyor, 3) gas emitted from the coal trans-
ported on the belt, and 4) background gas emitted from
the coal face and from the adjoining ribs in the intake
airway gateroad entries. Once the methane contribu-
tions from the various sources have been defined for an
actual longwall cutting sequence, the methane emis-
sions from an ideal (i.e. delay-free) cut sequence can

383

Copyright © 2006 Taylor & Francis Group plc, London, UK

 be predicted. The calculated methane emission con-
tributions can then be extrapolated to longer longwall
faces, taking into account the variations in coal produc-
tion and transport factors, to more accurately predict
future methane emission rates from longer longwall
faces.
2 METHODOLOGY
2.1 Underground field study
Methane concentrations were continuously monitored
and recorded on the longwall face for one produc-
tion shift per day (of the 2 production shifts per
day) for three consecutive days. There were a total
of 157 shields on the 315 m (1032 ft) longwall face.
The methane monitors were located at the 20th, 80th,
and 145th shield locations. Multiple methane sampling
concentration values were measured and recorded as
an averaged value for each minute. Periodic ventilation
airflow measurements were made along the longwall
face at the location of the methane monitors and at
other key locations in the study area. The location of
the shearer was recorded as part of the time study,
which also included the time, duration, and reason for
any mining delays due to the following factors: gas,
belt problems, face conveyor problems (rock jam), cal-
ibration of methane sensors, and miscellaneous other
problems.
The methane monitors for this study were located
at shields number 20, 80 and 145. These positions
were chosen based on previous studies that indicated
possible air interactions at the corners of the head
and tailgate (Diamond & Garcia 1999). Some of the
headgate air for the first few shields on the long-
wall face can flow behind the shields before reaching
the longwall face. Airflow from behind the tailgate
shields can be pulled into the main exhaust system
at various locations along the face, or face ventila-
tion air can travel behind the shields into a bleeder
system. This complex interaction of longwall face ven-
tilation airflow and airflow behind the shields was to be
reduced by the placement of methane sensors approx-
imately 37 m (121 ft, shield 20) from the headgate
corner and 21 m (69 ft, shield 145) away from the tail-
gate corner of the panel. The methane contribution
interpretation does not take into account the possi-
ble effect of ventilation airflow interactions with the
gob, which is consistent with the initial goal of the
study.
2.2 Site specific information
The physical aspects, equipment, operational, and ven-
tilation scenarios of the longwall panel (Fig. 1) are
very important to the evaluation of methane sources
384
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 1. Study area for longwall face emission investiga-
tion.
because they determine the mathematical formulas
used in the methane contribution model. The site-
specific variables for this study are as follows:
– Three entry gateroad development [#1 belt (intake),
#2 track (intake), #3 (return)] on the headgate
– Previous longwall panel mined on the tailgate side
– Exhaust ventilation via bleeder fan
– Bi-directional cut sequence
– Shearer depth of cut (web), 1.07 m (42 in)
– Some intake air brought to face via belt air-
way, therefore, belt coal methane component to
account for
– Panel width (edge-to-edge), 315 m (1032 ft)
– Original panel length, approximately 3048 m
(10,000 ft)
– Remaining panel length at the time of the study,
1195 m (3920 ft)
– Longwall face height, 2 m (6.5 ft)
– Shield width, 2 m (6.5 ft)
– Total number of shields, 157 (#1 is at the headgate
and #157 is at the tailgate)
– Shearer average cut speed, 14 m/min (46 ft/min)
– Face conveyor length, 315 m (1032 ft)
– Face conveyor speed, 1.78 m/s (353 ft/min)
– Longwall face average airflow velocity, 2.54 m/s
(500 ft/min)
– Belt speed, 4.06 m/s (800 ft/min)
– Belt airflow average velocity (intake), 1.93 m/s
(380 ft/min)
– The shearer’s location recorded at the location of
the headgate drum.
2.3 Assumptions
Methane content of the coal is assumed to be con-
sistent across the longwall face and consistent along
the #1 belt rib. No decay function for the compo-
nents sources was modeled because the accuracy of the
methane readings did not support such detail. Thus,
 the following assumptions have been applied to this Therefore, when the shearer is cutting coal, the lib-
study: eration of methane from the coal face will be recorded
by the methane monitor located near the tailgate at
– Background methane emissions from the longwall
shield 145, either instantly or up to 2 minutes later,
face (active, not idle) are linearly dependent on
depending on the shearer’s location. If the shearer is
longwall face length.
located on the tailgate side of shield 145, no extra
– Background rib methane emission from the #1
methane liberated by the shearer will be recorded by
belt entry are linearly dependent on the remaining
the methane monitor. If the shearer is located within
length of the longwall panel, but are independent
61 m (200 ft) on the headgate side of shield 145, then
of longwall face width or activity.
methane liberation due to the cutting of the coal will be
– There is no significant interaction between the face
recorded for that minute. If the coal was cut nearer the
air and the air in the gob.
headgate, it would take up to 2 minutes for the face air
– Stage loader located at the headgate was not incor-
to travel to the methane sensor located at shield 145.
porated in the simulation.
The contribution of methane emissions from the
– Methane liberation rate at the shearer is propor-
coal on the face conveyor is more complicated because
tional to the cut coal volume.
coal cut by the shearer can be transported on the face
– Methane liberation rate on the face conveyor is pro-
conveyor for 1 to 3 minutes. To determine the transport
portional to the coal tonnage × elapse time on the
time for methane emissions from the coal on the face
conveyor.
conveyor, the counter flowing longwall ventilation air-
flow and the shearer location/direction-of-travel must
2.4 Construction of formulas be taken into account. As an example, coal cut at
shield 145 will release methane that will be recorded at
The calculations for the methane contribution model
shield 145 instantly (0 minutes). Coal transported on
are empirical, and site-specific for the longwall panel,
the face conveyor will release gas into the face venti-
but the concept can be generally applied to other long-
lation airflow for a total of 3 minutes (0–2 minutes),
wall panels. Since the location of the longwall shearer
during which time the counter flowing face ventila-
is known from the time studies, the change in shearer
tion will take up to 2 minutes to reach shield 145, so
location and the depth of cut can be used to determine
methane emitted from the coal on the face conveyor
the volume of coal mined during each minute. The face
will be recorded at shield 145 for 0–4 minutes in this
conveyor and belt speeds are known, so the transporta-
example.
tion times of the coal after cutting can be determined.
Continuing this example, methane emissions from
The ventilation airflow rates along with the width and
the coal on the belt will be recorded at shield 145 for
remaining length of the longwall panel determine the
5–19 minutes after coal was initial cut at shield 145 (2
transit time of the airflow across the longwall face and
minutes for the face conveyor, 2 minutes for the face
for the belt entry.
airflow transit time, 1–5 minutes for the belt trans-
The following describes the mathematical process
port time, and 0–10 minutes for the belt airflow transit
used in the evaluation. The longwall face was bro-
time). Therefore, coal mechanically cut by the shearer
ken down into 61 meter sections (200 ft) to corre-
will affect methane emission levels near the tailgate
spond with the width of 30 shields. Since the face
instantly, and for as long as 19 minutes after being cut.
conveyor speed was 1.78 m/s (350 ft/min), the trans-
The contributions of methane from individual
port time for coal over 61 m (200 ft) of the face
sources over time lead to a simple set of five linear
conveyor was (61 m/1.78 m/s) = 34 seconds, and the
equations for shearer, face conveyor, and belt back-
transport time for the counter flowing ventilation air-
ground emissions from the coal face and back-
flow was (61 m/2.54 m/s) = 24 seconds. Therefore, the
ground emissions from the adjoining ribs in the intake
time for the coal to be transported by the face conveyor
gateroads, which were solved for each minute of the
and the time required for the counter flowing air to
three shifts monitored. Constants in each equation
transit the same distance adds up to 58 seconds, which
were calculated by least-squares, linear regression
was rounded to 1 minute to match the methane concen-
such that the calculated results best matched the actual
tration readings recorded every minute. The emission
readings at shields 20 and 145.
times for methane from the sources of interest as well
The shield 20 data led directly to calculations of the
as transit times for the associated ventilation airflows
background emission for the #1 belt entry intake rib.
can be summarized as follows:
The total background emission at shield 145 minus
– Coal cut by Shearer, 0–2 minutes the contribution from the belt yields the estimated
– Coal on face conveyor, 1–3 minutes background component from the longwall face alone.
– Airflow along longwall face, 2 minutes Longwall production delays in the cutting sequence
– Coal on belt, 5 minutes are used to define the methane emissions associated
– Airflow along belt, 10 minutes. with individual component sources.

385

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 1. Calculated constants for the emission component.

Background Background
Location used Conveyor Belt Shearer shield 145 shield 20

Day 1 Shield 145 only 6.16 1.85 9.03 119

Shield 20 & 145 6.51 1.63 7.95 129 −6
Shield 20 only 8.82 1.42 2.95 1
Day 2 Shield 145 only 4.20 1.99 5.10 123
Shield 20 & 145 4.07 1.55 3.48 149 5
Shield 20 only 4.69 1.12 4.26 24
Day 3 Shield 145 only 4.97 2.40 7.15 194
Shield 20 & 145 6.64 0.72 7.34 246 86
Shield 20 only 6.20 0.51 7.23 94
3-day average
Shield 145 only 5.11 2.08 7.09 146
Shield 20 & 145 5.74 1.30 6.26 175 28
Shield 20 only 6.6 1.0 4.8 40

3 RESULTS emission component into its two parts (i.e. methane

The calculated emission constants for each of the
methane contributor components for each day of the
study are shown in Table 1 including a 3-day average
value for these constants. Note that when both shield
20 and 145 are used to calculate the constants, the
background methane emissions can be separated into
the longwall face and intake rib components by the
difference between the shield 145 and shield 20 val-
ues. During the evaluation it was determined that the
readings from the methane sensor at the shield 80 loca-
tion were erratic and gave conflicting results indicating
incorrect data received from this sensor. Therefore,
constants have only been calculated for the shield 20
and 145 locations.
The calculated emission constants are consistent for
the three days of the study, except for the background
methane levels at shield 20. The background at shield
20 represents the methane load from the intake ribs,
which should be fairly consistent. However, this value
varied considerably over the three days of the study,
starting out as a minor negative value in day one, to
a high positive value in day three. The reason for this
dramatic increase is most likely caused by the inter-
section of a horizontal degas hole located outby and
up wind of the shield 20 methane sensor. The back-
ground rib value at shield 20 for day one, calculated
using data from both sensors, results in a small negative
value, which translates to only 0.0014 m3 /s (3 cfm).
After considering the consistency of the emission con-
stants for the other three components (face conveyor,
belt, and shearer) from shields 20 and 145 individually
and combined, it was decided to use the three-day aver-
age for shields 20 and 145 for further evaluation of the
background emission component. Using the average
for shield 20 and 145 will also separate the background
emissions from the face and from the intake ribs),
thereby giving a more accurate prediction for longer
face lengths.
The calculated average methane emission rates for
each contributor are shown in Tables 2 and 3. Table 2
shows the methane contributors solved using the daily
constants for shields 20 and 145. Table 3 shows the
methane contributors solved using the three-day aver-
age constants for shields 20 and 145. The ‘total’ values
in Tables 2–3 are the calculated average methane
for the individual days. The ‘actual’ values are the
recorded methane for the individual days, and in
Table 3, they do not exactly match the calculated ‘total’
values.
3.1 Gas delays
The background face and rib emissions represent about
59% of the total daily emissions; however, this is some-
what misleading because these emissions continue
throughout the entire shift, including during mining
delays. In contrast, the shearer, face conveyor, and belt
emissions are intermittent sources (i.e. they are only
a factor during active mining on the face and during
coal transport), but they are the primary contributor to
longwall face gas delays, as shown in Table 4.
The calculated methane contributions (Table 4)
clearly show that the head-to-tail passes experience
higher methane concentrations than the tail-to-head
passes. The time study data for the three days of the
study showed 19 gas delays (shutdowns of mining
equipment due to excessive methane concentrations)
on 10 of the 11 head-to-tail passes. The average loca-
tion for the gas delays on the head-to-tail passes was at
about shield 119, 245 m (804 ft), or 78% of the distance
down the longwall face from the headgate corner.

386

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 2. Daily methane emission contributor averages and percentages using daily shield 20 &
145 constants (top of table 1).

Shield 20 & 145 Conveyor Belt Shearer Face Rib Total Actual
daily results (m3 /s) (m3 /s) (m3 /s) (m3 /s) (m3 /s) (m3 /s) (m3 /s)

Day 1 0.012 0.015 0.005 0.027 −0.001 0.057 0.057

Day 2 0.005 0.014 0.004 0.028 0.002 0.052 0.052
Day 3 0.007 0.006 0.005 0.030 0.019 0.068 0.068
Average 0.008 0.012 0.005 0.029 0.006 0.059 0.059
Shield 20 & 145 Conveyor Belt Shearer Face Rib Total Actual
daily results (cfm) (cfm) (cfm) (cfm) (cfm) (cfm) (cfm)
Day 1 26 31 10 57 −3 121 121
Day 2 11 29 8 60 3 111 111
Day 3 15 14 11 64 40 143 143
Shield 20 & 145 Conveyor Belt Shearer Face Rib Total Actual
daily results (%) (%) (%) (%) (%) (%) (%)
Day 1 22 26 8 47 −3 100 100
Day 2 10 26 7 54 3 100 100
Day 3 10 9 8 45 28 100 100
Average 14 20 8 49 9 100 100

Table 3. Daily methane emission contributor averages and percentages using average three-day
shield 20 & 145 constants (bottom of table 1).

Shield 20 & 145 Conveyor Belt Shearer Face Rib Total Actual
3-day average (m3 /s) (m3 /s) (m3 /s) (m3 /s) (m3 /s) (m3 /s) (m3 /s)

Day 1 0.011 0.012 0.004 0.028 0.006 0.061 0.057

Day 2 0.010 0.012 0.004 0.028 0.006 0.060 0.052
Day 3 0.008 0.010 0.003 0.028 0.006 0.056 0.068
Average 0.010 0.011 0.003 0.028 0.006 0.059 0.059
Shield 20 & 145 Conveyor Belt Shearer Face Rib Total Actual
average results (cfm) (cfm) (cfm) (cfm) (cfm) (cfm) (cfm)
Day 1 23 25 8 60 13 129 121
Day 2 21 25 8 60 13 127 111
Day 3 18 21 6 60 13 119 143
Shield 20 & 145 Conveyor Belt Shearer Face Rib Total Actual
3-day average (%) (%) (%) (%) (%) (%) (%)
Day 1 19 20 7 49 11 106 100
Day 2 19 22 7 54 12 114 100
Day 3 13 15 4 42 9 83 100
Average 17 19 6 48 11 101 100

There were four gas delays recorded on three of
the nine tail-to-head passes. The average location for
the gas delays on the tail-to-head passes was at about
shield 71, 146 m (480 ft), or 46% of the distance down
the longwall face from the headgate corner. The aver-
age duration for all 23 gas delays was seven minutes
irrespective of cutting direction.
The relative contribution of the components of the
average daily methane emissions is not as significant
as it is the relative contribution of the components in
the peak levels that causes gas delays. Table 4 shows
that during the periods of gas delays, the methane
contributions from the face conveyor and shearer
dramatically increase in total percentage.
3.2 Cutting direction and delays
The direction of cutting is very important to methane
emissions on the longwall face because at the end
of a head-to-tail pass, the face conveyor, shearer, and
belt are all contributing gas at or near their maximum
rate. The tail-to-head passes do not have coincidental

387

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 4. Methane contribution percentages from longwall 250
Shearer
emission contributors during gas delays compared to daily Shearer Location
0.10

Shearers' Shield Location

averages for a 315 m (1032 ft) longwall face.

Methane Emissions m3/s

Face Conveyor

Methane Emissions cfm

200

0.08
Methane Conveyor Belt Shearer Background 150
contributions (%) (%) (%) (%) 0.06
100
Three day 17 19 6 59 0.04
averages 50
0.02
Gas delays No.
Cutting gas 23 32 18 9 42 0.00 0
17:00 18:00 19:00 20:00 21:00 22:00 23:00 0:00
delays
Time
Head-to-Tail 19 33 18 9 41
gas delays
Figure 3. Calculated shearer and face conveyor coal trans-
Tail-to-Head 4 26 18 8 48
port methane contribution based on the shearer’s location,
gas delays
study day one, 315 m (1032 ft) face.

Shearer Location 250 250

Measured Methane Shield 145 Shearer Location
Measured Methane Shield 20
0.10 Background
Shearers' Shield Location

0.10
Methane Emissions cfm

Shearers' Shield Location

200 Methane Emissions m3/s Belt

Methane Emissions cfm

Methane Emissions m3/s

200
0.08 0.08
150 150
0.06 0.06
100 100
0.04
0.04

0.02 50 50
0.02

0.00 0
17:00 18:00 19:00 20:00 21:00 22:00 23:00 0:00 0.00 0
17:00 18:00 19:00 20:00 21:00 22:00 23:00 0:00
Time
Time

Figure 2. Recorded methane emission data at shield 20 and Figure 4. Calculated belt coal transport and background
145 and shearer’s location, study day one. methane contributions based on the shearer’s location, study
day one.
maximums for face conveyor, shearer, and belt emis-
sions, so a more consistent emission rate occurs over
the entire cut sequence, which explains the less fre- Shearer
Belt
Calculated
Face Conveyor
Background
250
quent gas delays on tail-to-head passes. Of the 11
0.10
Methane Emissions m3/s

head-to-tail passes over the three shifts that were mon- 200 Methane Emissions cfm
itored for this study, not one full speed cut was made 0.08
without a gas delay. The first head-to-tail pass on 150
day one had two belt delays. The remaining 10 cuts 0.06

all had gas delays. Of the nine tail-to-head passes, 0.04

100

three had gas delays, three were cut at a slower speed

50
for unknown reasons, and three were cut full speed 0.02
without any delays.
0.00 0
Figures 2–6 show the calculated methane contribu- 17:00 18:00 19:00 20:00 21:00 22:00 23:00 0:00
tor components for day one using the three-day average Time
constants of shield 20 and 145 from Table 1. The cal-
culated methane contribution is a multiplication of the Figure 5. Calculated individual methane contributors and
three-day average constant of shield 20 and 145, and total calculated methane emissions for study day one.
the formula results for each minute for day one. Fig-
ure 2 shows the actual methane emissions measured at
shield 20 and 145, and the shearer’s location for day Figure 5 shows all components and the calculated
one. Figure 3 shows the calculated shearer and con- total methane emissions for day one. Figure 6 shows
veyor emissions for day one along with the shearer’s the calculated versus measured methane for shield 145
location. Figure 4 shows the belt and background during day one, along with the four gas delays that
emissions along with the shearer’s location. correspond to the high methane emissions.

388

Copyright © 2006 Taylor & Francis Group plc, London, UK

 250 Calculated Shearers Location
Belt Delays Measured Conveyor Belt
Calculated Shearers Background 350
0.10 0.160
Methane Emissions m3/s

H-T, 20 Min T-H, 20 Min

Methane Emissions cfm

200
0.140 300

Methane Emissions m3/s

0.08

Methane Emissions cfm

Sump, 8 Min Sump, 8 Min
0.120 250
150

Shearer location
0.1104
0.06 0.100
0.0897 200
100 0.080
0.04 150
0.0638
0.060 0.0593
50 100
0.02 0.040
Gas Delays 0.020 50
0.00 0
17:00 18:00 19:00 20:00 21:00 22:00 23:00 0:00 0.000 0
0 10 20 30 40 50 60 70 80 90
Time
Time (minute)

Figure 6. Calculated vs. actual measured methane emis-

sions, 315 m (1032 ft) longwall face, study day one, using Figure 7. Two full-cut passes without delays for 305 m
the three-day average, including four gas delays. (1000 ft) longwall face length.

Calculated Shearer Location

Conveyor Belt
3.3 Predicting methane emissions on longer 0.160
Shearer Background
350
longwall faces 0.140
H-T, 33 Min 0.152 T-H, 33 Min
300
Methane Emissions m3/s

Methane Emissions cfm

Sump, 8 Min 0.124 Sump, 8 Min
Simply determining the methane emissions at the 0.120 250

Shearer Location
corner of the tailgate and then linearly extrapolating 0.100
200
the results to longer faces does not take into account 0.080
0.084
150
the effect of coal production and transport factors. The 0.060
simple assumption that a longer face will only have 0.040
100

increased background face emissions and no con- 0.020 50

tribution from the other components (shearer, face
0.000 0
conveyor, belt) is incorrect, because productivity (coal 0 10 20 30 40 50 60 70 80 90

volume mined/hour) will be increased for a longer face Time (minute)

since the cut cycles are face length dependent. The

mine used a bi-directional cut sequence. At longer face Figure 8. Calculated methane emissions for two full cuts
without delays for a 488 m (1600 ft) longwall face.
lengths, the wedge/sumping times are assumed to be
the same as for the base case, but the cutting times will
increase proportionally to the face length, minus the component on a longer longwall face. The belt emis-
sumping distance. sions are a function of the amount of coal on the
Assuming that the longwall face conveyor can keep belt and the transport time. The belt transport time
up with the shearer and the shearer cuts at the same (dependent on remaining panel length and belt speed),
speed over the greater face length, it then follows that will be the same for different longwall widths, and
the productivity of the shearer will increase because a only the increased amount of coal produced by the
greater percentage of time will be allotted to cutting, shearer working on a longer longwall face will effect
than sumping. Therefore, the total methane liberation the belt emissions. The belt emissions can also reach a
from the mined coal during a shift would increase, but steady-state maximum when the entire belt is full with
the shearer’s emission rate during cutting will remain coal, thereby limiting the upper limit of belt emissions
the same for a longer longwall face, if the cutting regardless of panel width.
speed remains constant. The rib emission will be lin- The calculated results for methane emissions on
early dependent on the remaining length of the panel. longer longwall faces are predicted for a location
The background emissions for the longwall face will approximately 15 m (50 ft) outby the tailgate corner,
increase linearly with face length, however where the before any possible interaction with the gob gas near
belt line in used as a source of face air the background the tailgate. Figure 7 shows two full cuts without delays
emission from the intake rib will remain constant. for a 305 m (1000 ft) wide longwall panel. The pre-
The face conveyor, if operating at the same speed, dicted peak methane emission of 0.110 m3 /s (234 cfm)
will transport a greater volume of coal per hour for closely matches the maximum values recorded dur-
a longer longwall face because of the higher shearer ing the study, 0.099 m3 /s (210 cfm) (Fig. 2). Figure 8
utilization time for coal cutting versus sumping. In shows the predicted methane emissions for two full
addition, the face conveyor will also transport the cuts without any delays for a 488 m (1600 ft) wide
coal over a greater distance and for a longer time, longwall panel. The calculated peak methane emis-
thereby increasing the methane emissions from this sions for the 488 m (1600 ft) wide longwall panel

389

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 5. Calculated rates for methane emission contributors on idealized passes on longer longwall
faces.

Face Width Conveyor Belt Shearer Background Peak

(m) (m3 s) (m3 s) (m3 s) (m3 s) (m3 s)

305 0.040 0.027 0.009 0.035 0.110

366 0.048 0.027 0.009 0.040 0.124
427 0.056 0.027 0.009 0.046 0.138
488 0.064 0.027 0.009 0.052 0.152
Face Width Conveyor Belt Shearer Background Peak
(ft) (cfm) (cfm) (cfm) (cfm) (cfm)
1000 85 58 18 73 234
1200 101 58 18 86 263
1400 118 58 18 98 292
1600 135 58 18 110 322
Percent relative to 305 m (1000 ft) longwall face
Face Width Conveyor (%) Belt (%) Shearer (%) Background (%) Peak (%)
305 m (1000 ft) 100 100 100 100 100
366 m (1200 ft) 120 100 100 117 112
427 m (1400 ft) 140 100 100 134 125
488 m (1600 ft) 160 100 100 150 137

[0.152 m3 /s (322 cfm)] are 37% higher than for a 305 m
(1000 ft) wide longwall panel [0.110 m3 /s (234 cfm)]
(Table 5).
Coal on the face conveyor had the largest calculated
increase in methane emission rates on the longer long-
wall faces, while coal cut by the shearer and on the belt
had no increase (Table 5). The face conveyor’s methane
emission increase is due to the increased length and
time that the coal will be carried by the conveyor. Keep-
ing the length of the remaining panel (and hence, the
length of the belt) constant at 1195 m (3920 ft) for each
of the increased face length emission calculations pre-
cludes any extra peak methane load being emitted by
coal on the belt.
The background emissions increase with increases
in the face length due to the increase in exposed long-
wall face area. The methane contribution from the ribs
in the gateroads does not increase because the length
of the gateroads remains constant in these calculations.
One other consideration that should be noted in the
evaluation of the influence of increased face length on
longwall methane emissions is that the longer longwall
faces will theoretically have higher coal productivity
because a greater percentage of time will be spent cut-
ting coal and not sumping. As an example, the 488 m
(1600 ft) wide longwall panel will emit 37% more
methane at its peak (Table 5, Fig. 8), but will also
increase coal productivity by 11% (Table 6).
3.4 Reducing shearer speed to reduce peak
methane emissions on head-to-tail passes
As mentioned previously, the highest predicted
methane emissions for longer longwall faces are near
Table 6. Percent in total methane emissions and coal pro-
duction for longer longwall faces.
Face width Peak Average Productivity
methane (%) methane (%) tonnage (%)
305 m (1000 ft) 100 100 100
366 m (1200 ft) 112 113 104
427 m (1400 ft) 125 128 108
488 m (1600 ft) 137 141 111
the end of the head-to-tail cuts when the emissions
from the coal on the face conveyor and belt are at their
highest (Table 4). During the three shifts monitored
for this study, not one of the 11 head-to-tail passes
was completed at full speed without a gas delay. The
average gas delay occurred around shield 119, 245 m
(804 ft) from the headgate and averaged seven minutes.
Slowing down the shearer for the second half of the
face traverse will reduce the peak methane emissions
and give a more consistent emission level, as demon-
strated by decreasing the shearer speed factor in the
emission equations generated for this study.The results
of the influence of shearer traverse speed on emissions
for the 485 m (1600 ft) panel are shown graphically in
Figure 9.
The reduction in methane emission rates resulting
from slowing down the shearer transit time is dramatic,
with only 86% [0.152/0.131 m3 /s (278/322 cfm)] of
the peak methane emissions being encountered. The
five extra minutes required to cut the head-to-tail pass
due to the reduced shearer transit time (Fig. 9) is still
less than the observed average gas delay of seven min-
utes for head-to-tail passes. With the slowing of the

390

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Calculated
Conveyor
Shearer
0.160
H-T, 38 Min
0.140
Methane Emissions m3/s
0.131
0.120
10 m/min
12 m/min
14 m/min
0.100
0.080
0.060
0.040
0.020
0.000
0 10 20
Figure 9. Calculated influence of reduced shearer transit
time pass for 485 m (1600 ft) longwall face, shearer traveling
at 14 m/min for the first 250 m of the head-to-tail cuts, then
at 12 m/min, and finally at 10 m/min to the tailgate.
shearer transit time on head-to-tail passes, the cal-
culated peak methane emission values based on the
cutting direction are now within 7% of each other,
as compared to the 20% difference for the full speed
head-to-tail pass (Fig. 8).
3.5 Converting belt entry to return to eliminate
belt methane emissions
The conversion of the #1 belt entry to return airflow,
and the conversion of the #3 return entry to intake
could increase total airflow at the longwall face, as
well as eliminate the belt coal methane emission com-
ponent from the total methane emission load reaching
the tailgate corner of the face. The primary drawback
to this arrangement is that the #3 entry ribs tend to have
a higher background methane emission rate than the
#1 entry, due to the virgin coal along the #3 entry’s rib.
4 CONCLUSION
Coal production and transport factors have a dynamic
effect on methane emissions experienced on the long-
wall face. With a detailed mining time study and
associated methane concentration data across a long-
wall face, the methane emission contributions from the
major coal production and transport components can
be mathematically determined. The site specific lay-
out and position of the longwall panel determines the
regression constants for the calculations since methane
emission contributions change throughout the min-
ing of the panel. The mathematical concept can be
used to estimate the methane emissions for wider and
longer longwall panels. In addition, the concept can
be used for evaluating mine design, ventilation as well
as engineering and operating measures to control the
expected increases in methane emissions. The compo-
nent source approach can be applied to other coalbeds
Copyright © 2006 Taylor & Francis Group plc, London, UK
Shearer Location
Belt
provided that coalbed-specific methane emission char-
Background 350 acteristics are known or can be measured.
Sump, 8 Min
T-H, 33 Min
Sump, 8 Min
The expected peak methane emission increases for
300
wider longwall panels result primarily from the coal
Methane emissions cfm
0.127 0.123
250 transported on the face conveyor and the background
Shearer Location
200 emissions from the exposed coal on the face. The
0.078
methane emission increases related to the transport of
150
coal on the face conveyor is unavoidable for longer
100 longwall faces, unless the shearer’s transit speed is
50 reduced at the end of a head-to-tail pass. Methane
emissions associated with the cutting of coal on the
0
30 40 50 60 70 80 90 face by the shearer will remain constant at longer long-
Time (minute) wall face lengths, so long as the transit speed remains
the same.
Methane emissions from the coal transported on
the belt is the only component that can be altered by
engineering and ventilation practices such as cours-
ing the belt air away from the face. The background
face methane emission contribution can be reduced by
the use of or more extensive methane drainage tech-
niques. Long-term methane drainage will also reduce
the methane content in the coal and thereby reduce the
shearer and conveyor methane components.
In summary, mine designers have several options
to reduce or dilute methane emissions expected from
wider longwall panels, such as the following:
– Increasing ventilation airflow quantities to the
longwall face.
– Reducing the shearer transit speed, especially on
head-to-tail passes.
– Utilizing or increasing methane drainage tech-
niques will reduce emissions from all the consid-
ered sources by reducing the methane content of
the coal.
– Implementing ventilation design changes, e.g. not
coursing the #1 belt entry’s ventilation airflow to
the face.
DISCLAIMER
The findings and conclusions in this report are those of
the authors and do not necessarily represent the views
of the National Institute for Occupational Safety and
Health.
REFERENCES
Diamond, W.P. & Garcia, F. 1999. Analysis and prediction of
longwall methane emissions: a case study in the Pocahon-
tas No. 3 Coalbed, VA. NIOSH Report of Investigations
(RI 9649).
Gillies, S. 2000. Maximising coal production in the pres-
ence of H2 S seam gas, Final report. The University of
Queensland, Queensland, Australia: The University of
Queensland.
391
 Karacan, C.O., Diamond, W.P., Esterhuizen, E. &
Schatzel, S.J. 2005. Numerical analysis of the impact
of longwall panel width on methane emissions and gob
gas ventholes. Alabama Coalbed Methane Symposium,
Tuscaloosa.
392
Copyright © 2006 Taylor & Francis Group plc, London, UK
Schatzel, S.J. Trackemas, J., Krog, R.B., Garcia, F. &
Marshall, J.K. 2006. Prediction of the methane emission
consequences of increasing longwall faces widths: A case
study in the Pittsburgh coalbed. 11th U.S./North American
Mine Ventilation Symposium, Penn State.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Effect of scrubber operation on airflow and methane patterns at the

mining face

C.D. Taylor, J.E. Chilton, E.E. Hall & R.J. Timko

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute for
Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

ABSTRACT: The National Institute for Occupational Safety and Health (NIOSH) has conducted research to
determine the influence of mining machine-mounted scrubbers on airflows and methane concentrations at the
face when blowing ventilation systems are used. Tests were conducted in a full-scale ventilation gallery with a
model mining machine that simulated airflow created by a dust scrubber. During the tests, ultrasonic anemometers
were used to measure airflow speed and direction at several locations near the face. For the same test conditions,
methane was released from the face and gas concentrations were measured at 21 locations above the machine
using fixed point methanometers. Changes in airflow speed and direction are correlated with scrubber airflow
and the measured methane distribution above the mining machine. The research results showed that operation
of machine-mounted scrubbers improved face ventilation when blowing ventilation is used by increasing both
the intake flow and the quantity of air reaching the face.

1 INTRODUCTION
Effective face ventilation requires that the intake air
delivered to the end of the tubing or brattice must reach
the face. For cuts deeper than 6.1 m (20 ft), as little
as 5 pct of the air delivered to the end of the curtain
reaches the face (Thimons 1999). The amount of air
delivered generally decreases as the curtain setback
distance increases. Flooded bed scrubbers are widely
used on continuous miners to reduced airborne dust
levels during cutting. Previous research by Volkwein
(1986), Gillies (1982), andTaylor (1996, 1997) showed
that scrubber operation increases the quantity of air
reaching the face. The increased airflow can result
in lower methane levels. However, the mechanisms
whereby scrubber operation increases face airflows are
not clearly understood.
The objective of the current work is to use a full-
scale model of a mining machine at a simulated mining
face to examine how scrubber use affects face airflow
and methane concentrations. Ultrasonic anemometers
were used to measure airflow near the face with and
without a scrubber operating. Changes in methane
concentrations resulting from use of the scrubber were
determined by releasing methane at the face and mea-
suring the methane concentrations above the machine.
2 TEST PROCEDURES
2.1 Gallery and mining machine
Testing was conducted in NIOSH’s Pittsburgh
Research Laboratory, Ventilation Test Gallery. One
side of the “L shaped building” is designed to model
an underground mining entry, which is 5 m (16.5 ft)
wide by 2.1 m (7 ft) high (Fig. 1). A 1.2 m (3.5 ft)
wide by 12.2 m (40 ft) long box was built along the
right side of the face to simulate an uncut slab of
coal. The resulting entry simulated a 4 m (13 ft) wide
box cut.
A brattice and wood curtain constructed 0.6 m (2 ft)
from the left side of the entry directed airflow toward
the face for the scrubber evaluation tests. The curtain
setback distance for all the tests was 10.5 m (35 ft).
The intake flow quantity was calculated by multiplying
the velocity times 1.3 m2 (14 ft2 ), the cross-sectional
area of the airflow pathway behind the curtain. Regula-
tor door openings were adjusted to deliver the desired
flows of 1.9 or 2.8 m3 /s (4000 or 6000 ft3 /min) to the
end of the curtain.
A full-scale model continuous miner was located at
the center of the entry to simulate a mining machine at
a box-cut face (Fig. 2). The continuous miner model

393

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 1. NIOSH ventilation test gallery.
Figure 2. Continuous miner model used in ventilation
gallery.
includes a simulated scrubber system that consists of
the following:
– Two inlet openings 25 by 36 cm (10 by 14 in.), one
on each side of the mining machine, and 2.7 m (9 ft)
from the front edge of the cutting drum.
– An exhaust opening 38 by 38 cm (15 by 15 in.) at the
right rear of the machine chassis. Louvers over
the exhaust opening direct the airflow away from
the face.
394
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 3. Ventilation gallery airflow sampling locations.
– Ducting to provide flow between the inlets and
exhaust openings.
– An axial fan to move the air from the scrubber inlet
openings to the exhaust.
Scrubber airflow is adjusted by inserting orifice
plates with different size openings into the ducting,
just upstream of the fan.
2.2 Airflow measurement
Two, three-axis anemometers (Windmaster) and one,
single-axis anemometer (Solent), manufactured by
Gill Instruments Ltd. Great Britain (Disclaimer: Men-
tion of any company or product does not imply
endorsement by NIOSH), were used to make the
airflow measurements. Both instruments measure
velocities of 0.01 m/s (2 ft/min) and greater. Airflow
measurements were made at seven locations as marked
in Fig. 3. The single-axis instrument measures flow in
the direction in which the instrument is oriented. An
arrow on top of the instrument indicates the direc-
tion of the flow measurement. Behind the curtain, the
single-axis anemometer was aligned with the airflow
that is directed toward the face.
The three-axis instruments were used to measure
flow velocity between the curtain and the face where
flow direction frequently changes. The instruments
were positioned vertically so the flow velocity mea-
surements were made only in a plane parallel to the
top and bottom of the entry (Fig. 4). The vector com-
ponents in this horizontal plane, measured by the
anemometer, were used to calculate flow in the hor-
izontal plane in the direction of the flow. Although the
three-axis instrument also measures flow in the ver-
tical direction, velocities were only calculated for the
airflow in a horizontal plane.
To simplify the comparison of velocities measured
between the curtain and the face, measurements with
the three-axis anemometers were calculated to give
flow velocities either perpendicular or parallel to the
face. Specifically, velocities perpendicular to the face
were determined for locations 3–6. By convention, if
 Figure 4. Vertically mounted three-axis anemometer sensor
head.

the airflow was toward the face it was positive, and

away from the face it was negative (Fig. 4). Velocities
parallel to the face were determined for locations 1 and
2. The sign for flow parallel to the face was positive to
the right and negative to the left.
The three-axis sensor heads at locations 1 and 2 were
positioned 61 cm (2 ft) from the roof and at locations
3–6 the sensor heads were 30 cm (1 ft) from the roof.
The sensor head for the one-axis instrument (location
7) was located at the center point behind the curtain.
All anemometers were programmed to record aver-
age airflow data once a second. A software program
(ANEMVENT 2003) written by NIOSH recorded the
three-axis instruments data, and Hyper Terminal soft-
ware (Hilgraeve, Inc.) was used to record data from Figure 5. Ventilation gallery methane sampling locations
the single-axis instrument. Data was transferred to above mining machine.
EXCEL spreadsheets for analysis. The average veloc-
ities were calculated for the duration of each test (2 to [approximately 0.01 m3 /s (21 ft3 /min)] was monitored
3 minutes). and maintained by passing the gas through a rotameter.
For some tests, however, the gas flow was reduced to
2.3 Methane measurements
prevent concentrations in the gallery from exceeding
To evaluate the distribution of methane concentrations 2.5 pct. For the reduced flow tests, methane concentra-
over the mining machine, natural gas (approximately tions were adjusted by using the ratio of the flows.After
96 pct methane) was released into the gallery through mixing, methane concentrations were measured for the
four connected horizontal 3.7 m (12 ft) long by 3.8 cm next 5 minutes. The concentrations were recorded once
(1½ in.) diameter copper pipes. Holes were drilled every second. Data was downloaded to a computer via
2 mm (0.06 in.) in diameter and 6.3 cm (2½) apart on an analog-to-digital conversion board using commer-
top and bottom of each pipe. The four pipes were cial data acquisition software. Average concentrations
equally-spaced vertically, and located 0.1 m (4 in.) were then calculated for each sampling location.
away from the face to provide a relatively uniform
release of gas. 2.4 Test conditions
Methane concentrations were measured at 21 loca-
Airflow and methane readings were taken for the
tions above the mining machine (Fig. 5). All 21
following six operating conditions:
locations were 43 cm (17 in.) from the roof. A vac-
uum pump pulled air samples through plastic tubing – Scrubber off, intake flow 1.9 m3 /s (4000 ft3 /min)
from each of the sampling locations to one of 15 – Scrubber off, intake flow 2.8 m3 /s (6000 ft3 /min)
heat-of-combustion methane sensors. Fifteen loca- – Scrubber flow 1.9 m3 /s (4000 ft3 /min), intake flow
tions (1–15) were monitored simultaneously. Tests 2.8 m3 /s (6000 ft3 /min)
were then repeated to obtain data for locations 16–21. – Scrubber flow 1.9 m3 /s (4000 ft3 /min), intake flow
To allow the methane and air to mix thoroughly, 1.9 m3 /s (4000 ft3 /min)
methane was released into the gallery for 5 minutes – Scrubber flow 2.8 m3 /s (6000 ft3 /min), intake flow
before making gas measurements. A constant flow rate 2.8 m3 /s (6000 ft3 /min)

395

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 6. Airflow parallel to face (Location 1).
Figure 7. Airflow parallel to face (Location 2).
Figure 8. Airflow perpendicular to face (Location 3).
– Scrubber flow 2.8 m3 /s (6000 ft3 /min), intake flow
1.9 m3 /s (4000 ft3 /min)
Tests were repeated for each of the six test conditions,
and the results averaged for each set of conditions.
3 RESULTS AND DISCUSSION
3.1 Airflow measurement results
As described above, airflow velocities were calculated
for flows moving either parallel or perpendicular to the
face. The velocities of the airflow as it moved parallel
to the face (locations 1 and 2) are shown in Figures 6–7.
396
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 9. Airflow perpendicular to the face (Location 4).
Figure 10. Left side airflow perpendicular to face (4000
ft3 /min intake).
With the scrubber operating the velocities varied from
0.07 to 0.18 m/s (15 to 35 ft/min). Without the scrub-
ber, velocities were practically zero. In most cases
velocities increased when the intake airflow volume
was increased.
With no scrubber, the air moved toward the face on
the right side of the machine and away from the face on
the left side of the machine. With the scrubber operat-
ing, airflow directions were reversed on each side of the
machine. Although directions were reversed, the flow
velocities were similar on the right side of the machine,
with and without the scrubber operating. Increasing
the scrubber flow decreased airflow velocities moving
toward the left side of the face.
Figures 10 and 11 show the airflow velocities mea-
sured at locations 3 and 5–7 along the left side of the
entry. These locations correspond to distances of 1.5 m
(5 ft), 4.6 m (15 ft), 6.1 m (20 ft), and 10.7 (35 ft) from
the face. Airflow velocities decreased as air moved
away from the curtain. With the scrubber off, velocities
4.6 m (15 ft) from the face had decreased to practically
zero. Increasing either the intake or scrubber airflows
increased air velocities, except at sampling location
3, which was 5 feet from the face. At this location,
the flow velocity is reduced because a portion of the
air moving toward the face is drawn into the scrubber
inlet, which is just outby location 3.
Flow velocities at the end of the curtain increased
when the scrubber was turned on. A test was conducted
 Figure 11. Left side airflow perpendicular to face (6000
ft3 /min intake). Figure 12. Airflow changes measured behind the curtain
and at the regulator door due to scrubber use.
to evaluate the source of the increased flow. Two of
the three regulator doors were closed and the third
left completely open (Fig. 1). An outside door was
opened enough to adjust intake flows to either 1.9 m3 /s
(4000 ft3 /min) or 2.8 m3 /s (6000 ft3 /min) behind the
blowing curtain.
For each intake flow, the airflow velocity was mea-
sured at the end of the curtain and at the regulator door
opening with the scrubber operating. The air quantities
were determined by multiplying the measured veloci-
ties by either the area behind the curtain or the area of
the regulator opening. The same airflow measurements
were made at the end of the curtain and at the regulator
door opening, but with the scrubber operating with the
following conditions: Figure 13. Methane concentrations above mining machine
(4000 ft3 /min intake flow).
3 3
– Scrubber flow 1.9 m /s (4000 ft /min), intake flow
2.8 m3 /s (6000 ft3 /min)
– Scrubber flow 2.8 m3 /s (6000 ft3 /min), intake flow
1.9 m3 /s (4000 ft3 /min)
In each case with the scrubber operating, the air-
flow behind the curtain increased and the flow at the
regulator door decreased (Fig.12). The flow quantity
measured at the end of the curtain was 24 to 51 per-
cent greater than the flow quantity measured at the
regulator door. The additional airflow at the end of the
curtain could be due either to additional intake air-
flow diverted from the flow at the regulator doors, or
air leaking from the return to the intake side of the
blowing brattice.
Figure 14. Methane concentrations above mining machine
(6000 ft3 /min intake flow).
3.2 Methane measurement results
Methane concentrations measured above the mining the machine where the curtain was located. The high-
machine are shown in Figures 13 and 14 for 1.9 m3 /s est concentrations generally occurred in the corner of
(4000 ft3 /min) and 2.8 m3 /s (6000 ft3 /min) intake air- the face opposite the blowing curtain. Operation of
flows. The distribution of the methane over the mining the scrubber reduced methane levels at most of the
machine indicates airflow patterns that are distinctly sampling locations.
different for the various intake and scrubber flow com- The methane sampling locations (Fig. 5) were
binations. The methane concentrations measured over arranged in seven rows with three locations in each
the mining machine were higher closer to the face row. The methane concentrations were averaged for
where the gas was released, and lower on the left side of each of the seven rows of samples to evaluate how

397

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 15. Methane concentrations with distance from face
(4000 ft3 /min intake).
Figure 16. Methane concentration with distance from face
(6000 ft3 /min intake).
concentrations varied with distance from the face. The
averages for each row were plotted versus the distance
from the face (Figs. 15–16). For example, rows 1 and
7 correspond to distances of 0.6 m (2 ft) and 7.9 m
(26 ft), respectively, from the face. With or without
the scrubber operating, methane concentrations gen-
erally decreased as sampling distance from the face
increased. The concentrations decreased faster when
the scrubber was on, and intake airflow was higher
(i.e. 6000 cfm).
Methane released from the pipe manifold simulated
methane liberation at the mining face. Some of the gas
liberated at the face will pass through the scrubber and
be exhausted in the direction of the entry return. Part
of the gas from the scrubber exhaust, however, can
move back toward the face and over the machine. A
test was conducted to determine how methane from
the scrubber exhaust would affect methane levels over
the machine. A 3.8 cm (1½ in.) inner diameter hose
was placed in the scrubber ducting, just upstream of
the scrubber outlet, to inject methane into the scrub-
ber airflow (Fig. 2). The maximum possible methane
flow [0.016 m3 /s (34 ft3 /min)] was used to obtain a
minimum gas concentration in the ventilation airflow
needed for detection using the methane monitors. Test
conditions included:
– Intake flow of 2.8 m3 /sec (6000 ft3 /min) and a
scrubber flow of 1.9 m3 /sec (4000 ft3 /min), and
398
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 17. Methane from scrubber exhaust.
– Intake flow of 1.9 m3 /sec (4000 ft3 /min) and a
scrubber flow of 2.8 m3 /sec (6000 ft3 /min).
After allowing time for the gas and air to mix, methane
concentrations were measured at methane sampling
locations 16 to 21 (Fig. 5). Methane from the scrub-
ber exhaust was present above the machine for both
test conditions. Higher concentrations were measured
when scrubber flow was greater than the intake airflow
(Fig. 17).
4 CONCLUSIONS
Scrubber operation increased the airflow velocity
behind the blowing curtain and at all sampling loca-
tions along the intake side of the entry. Consequently,
airflow volumes reaching the face were greater when
the scrubber was on. Scrubber operation also affected
flow direction. With the scrubber on, air drawn into
the scrubber reduced flow velocities at sampling loca-
tions in by the inlets, especially when the intake airflow
was 1.9 m3 /s (4000 ft3 /min). However, airflows in by
the inlets were still higher and methane concentrations
were lower than when the scrubber was off. With no
scrubber operating, air moved from right to left across
the face and reversed when the scrubber was on.
Changes in airflow velocities and direction due to
scrubber operation affected methane concentrations
measured above the mining machine. Methane con-
centrations generally were lower at most sampling
locations when the scrubber was turned on. Methane
concentrations were reduced the most on the left side of
the face where air velocities were highest. Regardless
of whether the scrubber was operating or not, methane
concentrations decreased as distance from the face
 increased. The concentrations decreased more quickly
when the scrubber was operating and the intake airflow
was 1.9 m3 /sec (4000 ft3 /min).
Methane levels closest to the face (locations 1–3)
were also affected by airflow direction. Concentrations
were higher with the scrubber off because the right to
left flow interfered with the movement of air to the
return side of the entry. With the scrubber on, velocities
decreased as the air moved from left to right across the
face. The direction of the flow and the reduction in
velocity resulted in higher methane concentrations at
the right corner of the face.
Because the scrubber operation increased airflow
behind the curtain, more air was directed toward the
face. This study showed that part of this air was intake
air drawn up behind the curtain. The remainder of the
increase in intake airflow most likely came from return
air that was drawn through openings in the blowing
curtain.
Test results showed that methane in the scrubber
exhaust can move back over the mining machine. More
methane passed back over the machine when the scrub-
ber flow was greater than the intake airflow. For the
conditions tested, any gas from the scrubber exhaust
would have had only a small effect on methane levels
at the face. Moreover, the increase in intake airflow
reaching the face due to scrubber operation would
have further reduced methane gas from the scrubber
exhaust.
5 DISCLAIMER
The findings and conclusions in this report are those of
the authors and do not necessarily represent the views
of the National Institute for Occupational Safety and
Health.
399
Copyright © 2006 Taylor & Francis Group plc, London, UK
REFERENCES
Gillies, A.D.S. 1982. Studies in improvements to coal face
ventilation with mining machine – mounted dust scrubber
systems. SME-AIMEAnnual Meeting, Preprint No. 82–24,
13 pp.
Taylor, C.D., Rider, J. P. & Thimons, E.D. 1996. Changes in
face methane concentrations using high capacity scrub-
bers with exhausting and blowing ventilation. In: Pro-
ceedings of the SME Annual Meeting. Phoenix, Arizona,
Preprint 96–167.
Taylor, C.D., Rider, J.P. & Thimons, E.D. 1997. Impact
of unbalanced intake and scrubber flows on methane
concentrations. In: Proceedings of the 6th International
Mine Ventilation Congress. Pittsburgh, PA, Chapter 27,
pp. 169–172.
Taylor, C.D., Timko, R.J., Senk, M.J. & Lusin, A. 2003.
Measurement of airflow in a simulated underground mine
Environment using an ultrasonic anemometer. In Proceed-
ings of the SMEAnnual Meeting. Cincinnati, Ohio Preprint
03–065, 6 pp.
Taylor, C.D. & Zimmer, J. A. 2001. Effects of water sprays
and scrubber exhaust on face methane concentrations.
In: Proceedings of 7th International Mine Ventilation
Congress. Krakow, Poland, Chapter 65, pp. 465–470.
Thimons, E.D. Taylor & Zimmer. 1999. Ventilating the box
cut of a two-pass 40 ft extended cut. J. Mine Vent S.A.,
Vol 52, No. 3, 8 pp.
Volkwein, J.C. & Thimons E.D. 1986. Extended Advance of
Continuous Miner Successfully Ventilated With a Scrub-
ber in a Blowing Section. SME Fall Meeting, SME
Preprint 86–308, 4 pp.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Effect of water sprays on airflow movement and methane

dilution at the working face

J.E. Chilton, C.D. Taylor, E.E. Hall & R.J. Timko

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National
Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

ABSTRACT: The National Institute for Occupational Safety and Health (NIOSH) has been conducting research
to determine the influence of mining machine mounted water sprays on airflows and methane concentrations
at the face when blowing ventilation systems are used. Tests were conducted in a full-scale ventilation gallery.
Airflow speeds and directions were measured at several locations near the face with ultrasonic anemometers.
Methane was released from the face and concentrations were measured in the entry at locations above the mining
machine using fixed point methanometers. Changes in airflow speed, direction, and methane concentrations were
correlated with water spray operations. The test results using different spray arrangements and water pressures
showed that operation of the machine-mounted sprayers can improve face ventilation effectiveness by increasing
the velocity of airflow moving toward and away from the face. The improved ventilation resulted in reduced
methane levels near the face.

1 INTRODUCTION

Mining machine-mounted water sprays help provide

dust control by wetting the coal surface and agglomer-
ating the dust particles. Past NIOSH studies (Volkwein
1985) have shown that machine-mounted water sprays,
in addition to reducing dust, also move considerable
amounts of air, and can have a significant effect on
the dilution and redistribution of methane liberated at
the face. Water sprays can be especially helpful for
maintaining adequate airflow at the face when cutting
depths exceed 6.1 m (20 ft).
Sprays can reduce methane by increasing airflow
velocities and improving the mixing of methane and
intake air. Individual water sprays move air like a small
fan. When sprays at the front of the mining machine
are directed 30 degrees toward the return side of the
entry, they can move air across the face and help clear
methane gas (Ruggieri 1984, Taylor 2001).

2 TEST EQUIPMENT AND PROCEDURE

2.1 Gallery and mining machine

Testing was conducted in the NIOSH, Pittsburgh
Research Laboratory, Ventilation Test Gallery. One
side of the “L” shaped building is designed to model
an underground mining entry that is 5 m (16.5 ft) wide
by 2.1 m (7 ft) high (Fig. 1). A box, 1.2 m (3.5 ft) by Figure 1. Ventilation test gallery.

401

Copyright © 2006 Taylor & Francis Group plc, London, UK

 12.2 m (40 ft) long, was built along the right side of the
face to simulate an uncut slab of coal. The resulting
entry simulated a 4 m (13 ft) wide box cut.
A brattice curtain constructed 0.6 m (2 ft) from the
left side of the wall directed airflow toward the face.
The curtain setback distance for all the tests was 10.5 m
(35 ft). Two intake airflow quantities were used for
these tests, i.e. 1.9 or 2.8 m3 /s (4000 or 6000 ft 3 /min).
Airflows at the end of the curtain were varied by
adjusting regulators.
A full-scale model continuous miner was located at
the center of the entry to simulate a mining machine at a
box-cut face (Fig. 2). Ten Spraying Systems Company
model 3/8-BD-3 hollow cone nozzles (Disclaimer:
Mention of any company or product does not imply
endorsement by NIOSH) were aligned and mounted
in each of the two 3.0 m (10 ft) long spray bars. The
spray bars were constructed of 3.8 cm (1.5 in) diame-
ter plastic pipe and mounted on top of the miner boom
behind the cutting drum (Fig. 3). The nozzles were
Figure 2. Model mining machine at gallery face.
Figure 3. Straight and angled water sprays on model
mining machine.
402
Copyright © 2006 Taylor & Francis Group plc, London, UK
placed approximately 0.3 m (1 ft) apart. At this posi-
tion on the boom the spray nozzles were approximately
1 m (41 in) back from the face.
Water pressure was measured with a gauge mounted
near the spray manifolds. Water spray tests were con-
ducted with “high” and “low” water pressures. Water
line pressure, 483 kPa (70 psig), was used for the low
pressure tests. A centrifugal pump provided a pres-
sure of 1200 kPa (174 psig) for the high pressure tests.
Water flow rates from the sprays were measured for two
nozzles in each of the two pipes. The average water
flow rate for a nozzle was 0.0025 m3 /s (0.7 gal/min)
at 483 kPa (70 psig) water pressure and 0.004 m3 /s
(1.1 gal/min) at 1200 kPa (174 psig) water pressure.
Therefore, the total water flow rate for each set of
10 nozzles was 0.025 m3 /s (6.7 gal/min) at 483 kPa
(70 psig) and 0.040 m3 /s (10.6 gal/min) at 1200 kPa
(174 psig) pressure.
2.2 Airflow measurements
Two, three-axis anemometers (Windmaster) and one,
single-axis anemometer (Solent), manufactured by
Gill Instruments Ltd. Great Britain, were used to make
the airflow measurements. The three sampling loca-
tions are shown in Figure 4. The three-axis sensor
heads at locations 1 and 2 were positioned 61 cm (2 ft)
from the roof and 1.5 m (5 ft) from the face. The sensor
head for the single-axis instrument was located at the
center-point behind the curtain (location 3).
The three-axis instruments were used to measure the
airflow velocity between the curtain and the face where
flow direction frequently changes. The instruments
were positioned vertically so that velocity measure-
ments were made in plane parallel to the top and
bottom of the entry (Fig 5). The vector components
in the horizontal plane, measured by the anemome-
ter, were used to calculate flow in this plane and
in the direction of the flow. Although the three-axis
instrument also measures flow in the vertical direc-
tion, velocities were calculated only for the airflow in
a horizontal plane.
Figure 4. Airflow sampling locations (1–3) on model
mining machine.
 To simplify the comparison of the velocities mea-
sured between the curtain and the face, measurements
from the three-axis anemometers were used to cal-
culate flow velocities perpendicular to the face. By
convention, if the airflow was toward the face it was
positive, and away from the face it was negative.
All anemometers were programmed to record aver-
age airflow data once per second. A software program
written by NIOSH (ANEMVENT 2003) recorded the
three-axis instrument data, and Hyper Terminal soft-
ware (Hilgraeve, Inc.) was used to record data from
the single-axis instrument. Data was transferred to
EXCEL spreadsheets for analysis. The average veloc-
ities were then calculated for the duration of each test
(2 to 3 minutes).
2.3 Methane measurements
To evaluate the distribution of methane concentrations
over the mining machine, natural gas (approximately
96 pct methane) was released into the gallery through
four interconnected horizontal 3.7 m (12 ft) long by
3.8 cm (1.5 in) diameter copper pipes. Holes were
drilled 2 mm (0.06 in) in diameter and 6.3 cm (2.5 in)
apart on top and bottom of each pipe. The four pipes
were equally-spaced vertically, and located 0.1 m (4 in)
away from the face to provide a uniform release of gas.
Methane concentrations were measured at 15 loca-
tions above the mining machine (Fig. 6). All 15 loca-
tions were 43 cm (17 in) from the roof.A vacuum pump
pulled air samples through plastic tubing from each of
the sampling locations to one of 15 Bacharach methane
heat of combustion sensors. The fifteen locations
(1–15) were monitored simultaneously.
A methane flow rate of 0.015 m3 /s (32 ft3 /min) was
set with a globe valve and monitored by passing the
gas through a rotameter. Gas flows were reduced to
Figure 5. Vertically mounted three-axis anemometer sen-
sor head.
403
Copyright © 2006 Taylor & Francis Group plc, London, UK
0.0095 m3 /s (20 ft3 /min) for some tests to prevent
methane concentrations in the gallery from exceeding
2.5 pct. For the reduced methane flow tests, mea-
sured concentrations were adjusted by multiplying the
recorded concentrations by the ratio of the methane
flows (i.e. 0.015/0.0095 = 1.6) so that comparisons of
effects could be made at equal methane input flows.
Before making measurements, methane was released
into the gallery for 5 minutes to allow it to mix with
air and reach a relatively constant concentration. After
mixing for five minutes, methane concentrations were
recorded each second for the next 5 minutes.
Methane concentration data was downloaded to a
computer via an analog-to-digital conversion board
using Labtech commercial data acquisition software.
Time averaged concentrations were then calculated for
each sampling location.
2.4 Test conditions
Straight sprays (directed perpendicular to the face),
and angled sprays (directed 30 degrees to the right)
were tested at high and low water pressures [483 and
1200 kPa (70 and 174 psig)] and intake airflows of 1.9
and 2.8 m3 /s (4000 and 6000 ft3 /min).
Figure 6. Methane sampling locations (1–15) above model
mining machine.

Figure 7. Effect of water sprays on air intake velocities.
Figure 8. Airflow movement at the face without sprays
(+ toward face, − away from face).
3 RESULTS AND DISCUSSION
3.1 Airflow measurement results
Intake airflow velocities were measured at location 3
behind the blowing curtain (Fig. 7). Operation of the
sprays did not change the airflow behind the intake
curtain. However, spray operation did affect airflow
near the face at locations 1 and 2.
Airflow direction towards and away from the face
varied depending on whether the sprays were on or
off, and if straight or angled sprays were used. With
the sprays off, airflows moved toward the face on
the right side of the machine (location 2), and away
from the face on the left side of the machine (loca-
tion 1) (Fig. 8). Increasing the intake airflow from 1.9
to 2.8 m3 /s (4000 to 6000 ft3 /min) increased the air
velocities a small amount at both locations.
Figure 9 compares the effects of spray configura-
tions, water pressures and intake airflows on airflow
velocities at location 1. Both water spray systems
caused airflows on the left side of the mining machine
(location 1) to move toward the face. The airflow
velocities were 27 to 60 pct higher when the angled
sprays were used. When using the higher water pres-
sure, the airflow velocities were 103 to 209 percent
higher. Increasing the intake airflow had only a small
effect on airflow velocity at location 1.
404
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 9. Effect of spray configurations, intake airflows,
and water pressures on face air velocities on the left side
(Location 1) of the model mining machine.
Figure 10. Effect of spray configurations, intake airflows,
and water pressure on face air velocities on the right
side (Location 2) of the model mining machine (+ toward
face, − away from face).
Figure 10 compares airflow velocities on the right
side of the entry (location 2) for different intakes, spray
configurations and spray pressures. Airflow veloci-
ties increased, either toward or away from the face,
when the intake airflow was increased. Airflow direc-
tion changed depending on the spray configuration.
Air moved away from the face with the angled sprays,
and toward the face with the straight sprays. Increased
water spray water pressure resulted in higher velocities
for both the straight and angled nozzle orientations.
3.2 Methane measurement results
Methane concentrations measured above the mining
machine for tests with 1.9 m3 /s (4000 ft3 /min) and
2.8 m3 /s (6000 ft3 /min) intake airflows are shown on
Figures 11–12. Methane concentrations were averaged
for each of the 5 rows of samples to evaluate how these
concentrations varied with distance from the face. The
distribution of the methane over the mining machine
shows how airflow patterns are distinctly different for
the various intake and water spray combinations. In
general, the methane concentrations measured over the
mining machine were higher closer to the face where

Figure 11. Methane distributions (4000 ft3 /min intake).
Figure 12. Methane distributions (6000 ft3 /min intake).
Figure 13. Effect of intake airflow on methane
concentrations.
the methane was released, and lower on the left side of
the machine where the intake air curtain was located
(Figs. 11–12).
Additionally, methane concentrations versus dis-
tances from the face were plotted (Figs. 13–15). The
data (Figs. 13–15) show that increasing the intake air-
flow reduced methane levels approximately the same
405
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 14. Effect of nozzle direction on methane
concentrations.
Figure 15. Effect of spray pressure on methane
concentrations.
amount at all distances from the face (Fig. 13). Angled
sprays were more effective for reducing methane lev-
els at all sampling locations, but were more effective at
locations closer to the face (Fig. 14). The higher water
pressure had the greatest effect on methane levels at
the locations nearest [0.6 m (2 ft)] the face (Fig. 15).
4 CONCLUSIONS
This NIOSH study evaluated how the use of machine-
mounted water sprays installed on the front of a model
continuous mining machine can increase face airflow
movement and improve the dilution and removal of
methane released at the mining face.Airflow velocities
toward and away from the face were measured with
ultrasonic anemometers located on each side of the
machine.
With angled sprays, airflow velocities increased as
air moved toward the face on the left side of the
machine and away from the face on the right side of the
machine. Airflow movement across the face was left
to right. The straight sprays increased air velocities
moving toward the face on both sides of the min-
ing machine. Airflow away from the face was over
the top of the machine. Increasing the water pressure
 for angled and straight spray configurations more than
doubled the air velocities moving perpendicular to the
left (intake) side of the face. Increasing the intake air
quantities had minimal effects on airflow velocities
moving toward the face. The angled water spray, high
water pressure, and high intake airflow configuration
had the greatest effect near the face where air velocities
were highest resulting in efficient methane dilution.
In general, the use of angled sprays with higher
water pressure resulted in the highest airflows near the
face, and the best methane dilution over the mining
machine. Methane levels were lower for both straight
and angled sprays when water pressure was increased.
Angled water sprays operating at the higher pressure
reduced methane levels the most at the locations clos-
est to the face. Farther from the face, the sprays had
less of an effect on airflow velocities and the methane
levels above the machine were approximately the same
with either spray system.
The effective use of water sprays assumes that all
nozzles can be kept clear of blockage from particles
in the water supply. For these tests, each of the noz-
zles was frequently removed from the tubing and its
orifice cleaned. In actual use in the mining environ-
ment, the addition of an in-line water flow indicator
and a stainless steel mesh particulate filter is strongly
recommended (Southern 2003).
406
Copyright © 2006 Taylor & Francis Group plc, London, UK
5 DISCLAIMER
The findings and conclusions in this report are those of
the authors and do not necessarily represent the views
of the National Institute for Occupational Safety and
Health.
REFERENCES
Ruggieri, S. K., Doyle, D. M. & Volkwein, J. C. 1984.
Improved spray fans provide ventilation solutions. Foster
Miller Inc. Waltham MA, Coal Mining and Processing,
April, pp. 94–98.
Southern, P. 2003. Quick-change water spray system. Repair
King Inc., Shinnston, WV http://www.msha.gov/illness_
prevention/initiatives/jetspray/jetspray.htm
Taylor, C. D. & Zimmer, J. A. 2001. Effects of water sprays
and scrubber exhaust on face methane concentrations.
In: Proceedings of the 7th International Mine Ventilation
Congress. Krakow, Poland, Chapter 65, pp. 465–470.
Volkwein, J. C., Ruggieri, S. K., McGlothlin, C. &
Kissell, F. N. 1985. Exhaust ventilation of deep cuts
using a continuous-mining machine. Pittsburgh, PA: U.S.
Department of the Interior, Bureau of Mines RI 8992.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Coal mine methane utilization options

P.M. Sööt, D.R. Jesse & M.E. Smith

Northwest Fuel Development, Inc., Lake Oswego, OR, USA

ABSTRACT: Coal mine methane (CMM) utilization activities have been accelerating over the last 20 years.
The technologies range from power generation and pipeline delivery of the CMM to uses such as combustion
for thermal heating or simply flaring to mitigate anthropogenic emissions of methane into the atmosphere.
NW Fuel has had design and operating experience with the following CMM utilization technologies over the
last two decades: electricity generation, gas processing and delivery to a natural gas pipeline, direct combustion
of methane for mine air heating, replacement of propane use at a mine with CMM, and flaring of CMM to
generate carbon offset (greenhouse gas) credits.
A broad spectrum of technologies is available for capture and use of CMM. The coal operator needs to consider
all of these in the context of business and environmental constraints. The operator must also study the CMM
resource to make sure that enough gas is available over the life of the project.

1 INTRODUCTION CMM may be virtually free, the capital cost of new

Northwest Fuel Development, Inc. (NW Fuel) has
been studying and developing coalbed methane
(CBM) and coal mine methane (CMM) for over two
decades. This work has included both commercial
projects as well as technology R&D. Projects have
ranged from demonstration of three different power
generation prime movers, gas processing and delivery
to pipeline, direct combustion for mine air heating and
flaring of the gas to mitigate greenhouse gas emissions.
NW Fuel also proposed the use of abandoned under-
ground coal mines as sites for processing of CO2
emissions from power plants. The concept was to
utilize the coal left behind after mining as an adsor-
bent for processing the stack gas to remove the CO2 .
The removed CO2 could eventually be sequestered in
deeper, unmined coal seams. The initial process review
was funded by the Ohio Coal Development Office.
Although the process was technically feasible, the slow
kinetics of the adsorption step forced an abandonment
of the concept for economic reasons.
2 POWER GENERATION
The occurrence of coal resources generally provides
that geographic area with an economic advantage
over other parts of a country – low electricity costs.
Coal-fired power plants supply those areas with low-
cost power. That creates a challenge for new power
generating plants based on coal mine methane. One
must have a very low capital cost. Even though the
power generation makes it difficult to justify. In
considering these cost elements of power genera-
tion, one needs to be innovative in order to develop
economically viable power generating projects.
In its early power generation projects, NW Fuel
realized that a market survey would help in estab-
lishing the best type of prime movers to use. It
became evident that there would be a limited mar-
ket for large (>10 MW) power plants using CMM as
fuel. There are two examples of large CMM power
plants in the world. In Australia, 94 MW of generating
capacity were installed at the Appin and Tower Col-
lieries. In Buchanan County, Virginia, Consolidation
Coal Company is operating an 88 MW peaking plant.
Unlike these projects, most coal mines do not have
the capacity of producing enough methane for such
large installations. Even at the Australian operation,
natural gas is used to supplement the available coal
mine methane. This approach would not be economic
in most cases. As noted above, most coal mine areas
are typically in areas of low electricity prices.
A market survey of domestic coal mines and their
methane production potential provided a quantifica-
tion for the generating unit size. There are very gassy
mines that could theoretically produce a considerable
amount of power, but 8 of the 10 gassiest mines in the
US already capture methane from their best sources
and deliver it to pipelines. Most of the remaining
unused methane at these mines is emitted from the ven-
tilation shafts. That source can be used to supplement
higher quality methane as generator fuel, but it cannot
be used as the sole source of fuel. That means that the

407

Copyright © 2006 Taylor & Francis Group plc, London, UK

 ten gassiest mines provide only spotty market poten-
tial for power generation. Looking at other gassy mines
where the methane capture is not as well developed,
one immediately sees that the total methane emis-
sions are limited. The second 10 gassiest mines in
the US emit an average of 7 MMCFD (14 m3 /min)
from their ventilation shafts. If aggressive methane
drainage practices (e.g. pre-mining vertical and hor-
izontal wells) are instituted at such mines, they could
expect to recover 40 to 50% of that methane evolu-
tion in advance of mining and provide it as fuel for
power generation. That means that the power gener-
ating capacity at such active mines would be 10 MW
or less at each mine. Since it is unlikely that many
mines will institute aggressive methane drainage sys-
tems, one can see that a more realistic goal is 5 MW or
less per mine for power generation. Scores of smaller
mines will have even less potential. This is the reason
that NW Fuel had selected small, low cost generators
as the basic module for its coal mine power genera-
tion systems. These units can readily be combined to
provide 1 MW, or more, of capacity.
The same situation applies at abandoned under-
ground mines. Such mines do not emit as much Figure 1. NW Fuel engine for generators.
methane as they did while coal mining was active,
hence the power generating potential is even less. The
power generating potential will be 1 to 3 MW at even The small size of the units flies in the face of the con-
the gassiest abandoned mines. That meant that NW ventional wisdom of “economies of scale”. The advan-
Fuel should develop a generating plant concept of tage that small equipment has is the economies of
1 MW or less. If a smaller generator were built, mul- mass production. The production levels of the engines
tiple units could be installed to achieve the 1+ MW are orders of magnitude greater than the industrial
range. engines, or gas turbines, that are typically considered
The downside of NW Fuel’s approach is the signif- for power generation options at coal mines. Even with
icant overhead associated with each small installation. higher “per kW” costs for other parts of the gensets, the
This overhead burden applies to issues from design and modular skids have a lower total price than comparable
fabrication to acquisition of air emission permits. By generating units based on industrial engines.
designing each installation around basic modular con- Multiple units have been installed at two locations
cepts, NW Fuel was able to make small installations at the Nelms Mines in Ohio and at one location at the
economically competitive with large facilities. Federal No. 2 Mine in West Virginia. The West Virginia
location is host for 18 engines with a total capacity
2.1 Internal combustion (IC) engines of 1.2 MW. One of the Ohio sites has had as many
as 14 units operating, see Figure 2. The second Ohio
As a result of the market evaluation, NW Fuel decided location is being expanded to an eventual capacity of
to utilize small internal combustion (IC) engines as 1.2 MW.
the prime mover. The generating equipment has been
developed by the Company based on more than a
2.2 Gas turbine
decade of R&D. The basic module fabricated by NW
Fuel is an engine/generator skid package that can Gas turbines are generally not economically viable.
produce 75 kW of electricity as shown in Figure 1. The 1 MW–3 MW turbines would cost $800 to
The first units built by the Company operated 8,000 $1,000/kW for just the bare equipment. The site devel-
to 10,000 hours before requiring significant mainte- opment and electrical connection hardware would add
nance, i.e. valve replacement. The complete engines more cost to the total installation. NW Fuel is able to
were replaced after about 20,000 hours of operation. fabricate and install small IC units for a total installed
The latest gensets operated by NW Fuel have signif- cost of less than $700/kW, including site development.
icantly improved on these benchmarks as the valve The only type of gas turbines that would be compet-
replacement is now at over 12,000 hours and some itive are used units. Even that approach still has the
engines are lasting over 30,000 hours before needing downside that gas turbines need to have fuel delivered
a “lower-end” overhaul. at much higher pressures than IC engines. That would

408

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 2. NW Fuel Ohio gensets.
require the installation of additional compressors for
fuel delivery.
The one aspect of gas turbines that are superior to
IC engines is the potential for such prime movers to
use the methane emitted from mine air shafts as the
combustion air in the turbine. That would combust the
extremely dilute methane in the mine ventilation air.
The reason that the gas turbine has an advantage is
the stoichiometric ratio of fuel to air in the respective
prime movers. Gas turbines typically burn 4% methane
in air within the combustion canister. IC engines
require much higher concentrations of methane in air
mixtures, nominally 8–10% methane. Given the 1%,
or less, methane in mine ventilation air, it is obvious
that gas turbines can offset a larger fraction of the
fuel requirement by using the mine ventilation air as
combustion air.
NW Fuel has physically demonstrated this concept
at the Nelms No. 1 Mine in Ohio. Some CMM from the
Mine was diluted with air to simulate mine ventilation
air. This was delivered to a gas turbine as combustion
air. The test showed the displacement of conventional
fuel by the methane in the combustion air as the fuel
controller automatically reduced the amount of con-
ventional fuel delivered to the combustion canister
while maintaining a constant output from the gener-
ator. This is the only known test of its kind in the
world. Due to the poor economics of gas turbine prime
movers, this concept has never been commercially
applied.
NW Fuel also has direct experience with the use
of micro-turbines for power generation using CMM
as the fuel. This approach was applied in Japan. Sum-
itomo Coal Company invited NW Fuel staff to help
them set up such units at their sealed Akabira Mine on
Hokkaido Island in Japan, as seen in Figure 3. NW Fuel
evaluated the CMM resource at that Mine and fabri-
cated the fuel supply compressor that provides CMM
to the pictured micro turbines.
409
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 3. Micro turbines at Akabira Mine, Japan.
2.3 Fuel Cell
Fuel Cell Energy (FCE) of Stamford, CT installed and
operated a demonstration fuel cell at the Rose Valley
(sealed) Shaft of the Nelms No. 2 Mine in Ohio. This
$6 MM project was co-funded by FCE, US Depart-
ment of Energy, and NW Fuel. The fuel cell used CMM
from the Nelms 2 Mine and delivered electric power
to the utility grid. The unit was installed in May, 2003
and operated successfully until early 2004.
3 CMM DELIVERY TO NATURAL GAS
PIPELINES
With recent high natural gas prices, it is often more
profitable to deliver CMM to pipeline than to burn
the gas for power generation. This is the case even if
the gas must be processed to bring it to pipeline qual-
ity standards. In some cases, pipeline companies are
wiling to accept sub-quality gas as long as the sub-
sequent mixture of gases in their pipeline meets their
specifications. The latter is the case in a project in
Ohio.A 10-mile pipeline was installed to deliver CMM
gas to a Dominion/East Ohio (DEO) gas transmission
line. CMM has been delivered through that pipeline
since 1998.
3.1 Carbon dioxide and water removal
The DEO example is not the norm. Most pipeline
companies require that the delivered CMM meet their
pipeline specifications, which limit the inert gas con-
tent to 4% or less and the water content to 7 pounds
per million cubic feet (MMCF). CMM normally
 contains CO2 , N2 , and water that would not meet these and it loses 30% to the waste stream. Again, the waste
specifications. The CO2 plus N2 must be reduced to stream can be fed to generators for power generation.
less than 4% of the total gas stream. Ninety-nine per- NW Fuel’s philosophy for both of these N2 rejec-
cent (99%) of the water must be removed since the gas tion processes was to develop low cost options for
coming from an underground mine typically contains small-scale CMM applications. They are not as effi-
700 lb/MMCF. cient as cryogenic plants nor large-scale PSA systems,
Neither the CO2 removal nor dehydration are par- but they are cost competitive for projects that need
ticularly difficult. One can install an amine scrubbing partial N2 rejection and that have only 1–2 MMCFD
system to remove virtually all the CO2 in any gas of CMM available. The processes can be run for less
stream. Alternatively, a pressure swing adsorption sys- than $0.50/MCF, which includes capital and operating
tems can remove CO2 . And, glycol scrubbing equip- costs. The Engelhard system is more efficient but has
ment is commercially available to remove the water larger capital costs.
from the CMM. The only modifications needed for
these commercial processes relate to the operating
pressure. Both of these processes are oil-field pro- 4 DIRECT COMBUSTION APPLICATIONS
cesses that are normally designed for high pressure
applications. Process engineering adaptations have to 4.1 CMM flaring
be made to operate them optimally at the low pressures Recent interest in global climate change has provided
that are typical at CMM production facilities. the motivation to add another technology to the list
Another significant consideration is oxygen (O2 ). If of CMM utilization options. Carbon Dioxide (CO2 ) is
there is O2 present in the CMM, which is often the case, the largest greenhouse gas emitted by anthropogenic
then one has to allow for that in both the amine scrub- sources, but the methane (CH4 ) in CMM is about 21
bing for CO2 removal and glycol scrubbing for dehy- times as potent on a weight basis (7.6 times on a volu-
dration. Process engineering and operational changes metric basis). That means that every cubic feet of CH4
have to be made. that one can burn to form one cubic foot of CO2 would
reduce CO2 -equivalent greenhouse gas emissions by
a net value of 6.6 cubic feet of CO2 .
3.2 Nitrogen rejection In reflection of that desire to limit methane emis-
Nitrogen rejection is not as easy as CO2 or water sions, NW fuel has developed flares to burn the
removal. Cryogenic processing can separate the N2 methane in CMM so it is not emitted into the atmo-
from methane, but such plants need to be very large sphere. This design was developed through R&D at the
in order to be economically competitive. Smaller scale Nelms Mines where the combustors were designed,
CMM production (1 MMCFD of CH4 or less) requires fabricated and tested. Several units were then fabri-
the use of a different process – pressure swing adsorp- cated for “commercial” application (i.e. to produce
tion (PSA). NW Fuel has developed two processes carbon offset credits) and placed at the Black Vent and
for nitrogen rejection from CMM and Engelhard also the Main North Shaft (Fig. 4) of the Nelms No. 2 Mine.
has a process that is commercially viable for CMM
applications.
The PSA processes are batch operations.They use an
adsorbent that selectively adsorbs one gas component
over another during a pressurization step. The process
releases the adsorbed gas during a depressurization
step. There are numerous intermediate steps that can
be used to optimize the process for either improvement
of separation or product recovery efficiency.
NW Fuel’s conventional PSA unit at the Nelms 1
Mine is capable of removing about 50% of the nitro-
gen in the feed stream while losing 20% of the feed
gas to a “waste” stream. The “waste” stream need not
be thrown away. It can be delivered to a generating
equipment to produce electricity.
In an advancement to allow for high throughput and
low cost, NW Fuel developed the Continuous PSA
(CPSA) that can process CMM streams of 1 MMCFD
and higher. The higher capacity of this system comes at
the cost of less N2 rejection and a larger waste stream.
The CPSA can remove about 30% of the feed nitrogen Figure 4. Main North Shaft Flares, Nelms No. 2 Mine, Ohio.

410

Copyright © 2006 Taylor & Francis Group plc, London, UK

 4.2 Mine air heating
This same burner design was subsequently modified
for application as an air heater at an active coal mine.
Coal mines in cold climates have problems in the win-
ter with the mine inlet air arriving in the underground
workings at extremely low temperatures. NW Fuel’s
flare equipment is now being successfully used at an
operating mine. It is producing the desired amount
of heat for the mine inlet air heating while meeting
MSHA guidelines for carbon monoxide (CO) concen-
trations inside the mine. Four heaters were installed.
Each one is rated at 12 MMBtu/hr (12 × 106 Btu/hour).
The heaters are shown in Figures 5 and 6.
Figure 5. Mine air heater.
Figure 6. Heater installation, West Elk Mine, Colorado.
411
Copyright © 2006 Taylor & Francis Group plc, London, UK
4.3 Propane fuel offset
Many coal mines use propane as a fuel for space
heating or water heating. NW Fuel is currently work-
ing at an active coal mine to install a CMM pipeline
and modify combustion equipment to accommodate
CMM. Once the conversion is complete, the mine will
be able to save about $200,000 per year in propane
costs.
5 CNG AND LNG DEVELOPMENTS
If none of the foregoing options are viable, there is
still another way of making use of CMM – converting
it into a fuel that can be transported to a point of use.
This requires that the CMM be modified from a light
gas to a product with a significant energy density. This
can be accomplished by compressing the gas to very
high pressures to produce a compressed natural gas
(CNG) or by cryogenically producing a liquid natural
gas (LNG).
The CNG application is going to be made at a
project in India, where the CNG will be used as a fuel
in coal mining trucks. The diesel truck engines will
be converted to bi-fuel capability. Some diesel will be
needed to provide the basis for igniting the CMM in
the engine, but most of the fuel requirement will be
met by the CMM.
NW Fuel will also be involved in a plant that pro-
duces LNG from CMM. Appalachian Pacific (AP) is
working under an US DOE contract to demonstrate this
application for CMM. The AP plant will be installed
at NW Fuel’s West Virginia site.
6 CMM RESOURCE ASSESSMENT
Technology selection is not necessarily the major
decision in CMM project development. Many CMM
developers have made major investments, only to find
that they do not have a sufficient CMM supply for the
facility that was installed. CMM resource issues are
beyond the scope of this current paper. Those matters
have been covered in depth in other publications by
NW Fuel.
In developing CMM projects, one needs to under-
stand factors such as remaining CMM resources and
the producibility of each CMM well. NW Fuel has
developed a CMM resource prediction model and
identified several different types of CMM wells. Each
type of well must be produced in an unique manner
in order to optimize the amount of CMM that can be
delivered to a plant that utilizes CMM.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Methods and means of recovery and utilization of methane from mines

A. Tor & A. Jakubow

Jastrzebie Coal Company, Jastrzebie, Poland

N. Szlazak
AGH University of Science & Technology, Cracow, Poland

S. Wasilewski
Research and Development Centre for Electrical Engineering and Automation in
Mining EMAG, Katowice, Poland

ABSTRACT: Methane occurring in coal seams and emitted into the space of underground workings is a
potential source of hazard for underground mines. Emission of methane from the seams of the surrounding
rocks and selected space creates a serious hazard of explosion or methane ignition.
Recent years distinctly show that in Polish mines due to carrying of extraction from ever deeper and deeper
seams the methane hazard is a source of serious threat for the lives and health of miners and can also disturb the
continuity of mining operations. Monitoring of the methane hazard and prevention against methane explosion
is carried out by the monitoring systems with automatic switching off of electric energy supply in underground
workings when methane concentration exceeds the allowable limit. Combating of methane hazard with ventilation
methods by supplying of a respective volume of air into the workings to dilute and lower methane concentration
to a safe level with a high methane emission may be not enough and therefore the necessity of methane drainage.
Degasification is an important and effective means of prevention measures and combating of methane hazard
therefore it is used at present in 23 Polish mines. The paper characterises the system of methane drainage from
the workings applied in Polish mines. Irrespective of safety aspects methane drainage can be very important
because of economic and environmental reasons provided it is then utilized. Because there are no possibilities of
management of the gas all the year round some of the Polish mines do not utilize methane at all flaring the gas into
the air. In the remaining mines the gas is used mainly for power purposes as electric energy and heat generation.
Today about 60% of methane recovered from deep mines is utilized. An interesting application consists in the
use of a Combined Heating, Cooling and Power Plants (CHCP) At Pniowek mine generating three types of Energy
from the recovered methane: heat, cooling and electric. This novel solution applied in the most gassy mines in
Poland with absolute gassiness of about 240 m3 /min, allowed to increase the utilization of the gas recovered in
this mine from 64% to 71%. The installation of an energy-cooling combined system with the power 6, 4 MW at
Pniowek mine is the first installation in Poland with the application of a gas engine which allows generating low
and high temperature heat. In order to provide proper parameters of the gas on its inlet to gas motor a precise
system was introduced at Pniowek mine of automatic and continuous monitoring of gas parameters in the gas
pipelines. Knowledge of concentration distribution and also the volume of gas in underground branch pipelines
allow for such an adjustment of valves as to obtain in a collective pipeline on the surface proper parameters of
the gas on its inlet to gas engine.

1 INTRODUCTION without oxygen, which can result in suffocation. Many

Methane present in coal seams, when emitted into min-
ing headings, is the source of a serious hazard for
miners working underground and for the continuous
works of a coal mine. Methane emission from mined
coal and goaf is a serious hazard as far as work safety
conditions are concerned, both in the form of methane
explosion or ignition and in the form of atmosphere
disasters, which took place in mines, are the result of
the hazards just mentioned.
Mining coal seams in methane mines requires using
special technical solutions in order not to allow exceed-
ing the safe level of methane concentration in mining
air. The basic method is the usage of suitable ventila-
tion systems with an intensive stream of air. However,
ventilation methods are frequently not very effective.

413

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Therefore methane drainage from coal seams and
enclosing rocks is necessary.
Reducing methane hazard by means of methane
drainage helps to improve miners’ work safety and
the continuity of mining equipment’s work at the
same time limiting the number of machine stoppages
resulting from power cuts caused by exceeding the
critical values for methane concentration. Methane
drainage of mines is a significant and effective means
of methane hazard prevention and at present is carried
out in 23 Polish mines. In 2003 227.1 million m3 of
methane was recovered by means of methane drainage,
which is 28.5% methane emitted to underground work-
ings. Approximately 131.5 million m3 , that is 59.9%,
was utilized.
Methane drainage together with ventilation air does
not form a possibility of energetic usage and increases
a greenhouse effect. Methane belongs to six green-
house gases whose emission, in accordance with Kioto
Protocol, should be controlled and reduced by 25%.
Methane life is relatively short (about 12 years com-
pared to 114 years for NO2 and from 5 to 200 years for
CO2 ), however, its greenhouse potential is 21 times
bigger than for CO2 . The most common sources of
anthropogenic methane emissions are: garbage dumps
(∼30%), rice paddies (∼18%), fuel chain of natural
gas (∼9%), cattle breeding (∼29%) and others (2%).
As mining industry is one of the main industries
in Poland and there are a lot of methane mines, the
amount of methane emitted into atmosphere is quite
big. Using a methane drainage method on a large scale
makes the energetic utilization of methane possible
and helps to reduce a greenhouse effect provided it is
utilized. Although the total amount of methane emitted
from mines is increasing all the time (from 763 mln m3
in 1998 to 779.9 mln m3 in 2003), the amount of
drained methane is also increasing (from 203.6 mln m3
in 1998 to 228.6 mln m3 in 2003). Unfortunately, the
amount of methane utilized fell from 141.5 mln m3 in
1998 to 128.0 mln m3 in 2003, that is by about 10%,
which means an increase in the amount of methane
emitted into atmosphere. Therefore, action needs to
be taken in order to improve the utilization of methane
recovered by means of methane drainage system.
The level of methane utilization in mines using
methane drainage systems varies a lot. There are some
mines where methane is not utilized at all (e.g. Coal
Mines: Chwalowice, Slask, Sosnica, Szczyglowice)
but there are also some, such as Brzeszcze Coal Mine,
where methane is utilized in 100%.

Table 1. Indexes of methane bearing capacity and methane drainage for 2003 in coal mines.

Methane quantity, m3 /min Methane

quantity
From Methane emitted to Indexes, %
From methane utilization, atmosphere,
No. Mine Total ventilation drainage m3 /min m3 /min Of drainage Of utilization

1 Brzeszcze 199,8 147,6 52,2 52,1 0,1 26,1 99,8

2 Halemba 53,64 50,74 2,9 2,31 0,59 5,4 79,7
3 Pokoj 3,34 1,09 2,25 1,78 0,47 67,4 79,1
4 Silesia 62,07 53,26 8,81 4,99 3,82 14,2 56,6
5 Staszic 36,16 24,22 11,94 5,86 6,05 33,0 49,1
6 Wesola 86,2 59,8 26,4 11,6 14,8 30,6 43,9
7 Bielszowice 59,68 47,1 12,58 4,54 8,04 21,1 36,1
8 Budryk 87,3 62,3 25 8,4 16,6 28,6 33,6
9 Sosnica 77,78 70,6 7,18 0 7,18 9,2 0,0
10 Slask 16,71 14,28 2,43 0 2,43 14,5 0,0
11 Szczyglowice 33,85 25,57 8,28 0 8,28 24,5 0,0
12 Makoszowy 23,47 21,11 2,36 0 2,36 10,1 0,0
13 Wujek 13,31 10,81 2,5 0 2,5 18,8 0,0
14 Zofiowka 82,88 50,49 32,39 31,02 1,36 39,1 95,8
15 Jas - Mos 54,5 36,18 18,32 17,11 1,21 33,6 93,4
16 Pniowek 239 145,9 93,1 66,09 27,01 39,0 71,0
17 Marcel 46,03 34,68 11,35 6,84 4,51 24,7 60,3
18 Krupinski 112,8 54,75 58,05 27,36 30,74 51,5 47,1
19 Borynia 30,35 26,67 3,68 1,93 3,37 12,1 52,4
20 Jankowice 40,26 21,42 18,84 1,54 17,3 46,8 8,2
21 Chwalowice 13,05 7,42 5,63 0 5,63 43,1 0,0
22 Anna 21,21 16,46 4,75 0 4,75 22,4 0,0
23 Rydultowy 90,51 66,46 24,05 0 24,05 26,6 0,0
TOTAL 1483,9 1048,91 434,99 243,47 193,15

414

Copyright © 2006 Taylor & Francis Group plc, London, UK

 The mines belonging to Jastrzebie Coal Company
are the ones with the greatest methane hazards and
Pniowek Coal Mine has both the highest methane
emission and the greatest level of methane drainage
(Table 1). In 2000 there was 18.5 million m3 of not
utilized methane in Pniowek Coal Mine, but due to the
introduction of a combined heating – cooling – power
system (CHCP) the utilization of drained methane
increased from 64% to 71% and the quantity of not
utilized methane fell down to 12.3 million m3 .
Mines obtain gas as a by-product during coal
mining; therefore methane drainage methods aim at
improving work safety and intensifying coal extrac-
tion. Due to changeable mining conditions, gas param-
eters including its content and quantity change, which
makes assuring the gas parameters required by its
recipients more difficult, especially the quantity of gas
with the right calorific value and concentration. This
fact makes the effective utilization of mining gas as
a natural source of energy difficult.
2 METHANE DRAINAGE METHODS USED
IN POLISH MINES
Methane release when the structure of rock mass is
not disturbed during mining operations is small due to
geological conditions of methane occurrence in a coal
deposit and small desorption of Polish coal.
Negative experiences of foreign companies as far as
methane drainage through bore-holes from the surface
in the unstressed parts of rock mass is concerned con-
firm the thesis that there is little methane release from
such a rock mass. The quantity of methane released is
strictly connected with the range of mining operations,
both opening and proper mining of coal seams.
Mining, at present conducted in 100% with caving,
additionally releases methane present in coal seams.
In the mines characterized by high methane content,
ventilation methods for controlling methane hazards
have not been sufficient since their construction was
started. Therefore, they were forced to use differ-
ent methods for controlling methane hazards, that
is methane drainage of opening, development and
mining workings.
Methane drainage of rock mass by means of bore-
holes consists in making bundles of holes from an
excavation or stables drilled earlier and connecting
them with a drainage pipe in order to reduce pressure
below the pressure present in an excavation. As rock
mass close to an excavation is cracked a bore-hole in
this zone is sealed by means of a cemented pipe-holder
in order to limit airflow into this hole. However the
sealing is not perfect and some air gets into a bore-
hole. During the existence of a borehole some extra
cracks are formed as a result of mining stresses that
cause an increase in airflow into a borehole, which in
415
Copyright © 2006 Taylor & Francis Group plc, London, UK
turn causes the fall in methane concentration in the
mixture drained.
Methane drainage of rock mass is the most effective
way of controlling methane hazard ensuring the fall
in methane outflows into workings and preventing or
limiting phenomena such as: blow-outs, abrupt break-
outs of methane and coal etc.
Among the methods of methane drainage used so
far, the following distinction can be made:
– methane drainage of development headings,
– pre-drainage – conducted before mining in an
unstressed rock mass with an original state of
stresses is started,
– methane drainage during mining – conducted at
the same time as coal seams are mined in a
rock mass, where the original equilibrium of the
deposit was disturbed by mining and the process
of methane drainage is conducted in rocks with
variable stresses,
– methane drainage from behind stoppings.
Methane drainage through bore-holes is conducted
from development headings, entries and operating
faces. At deeper levels of the mines methane drainage
by means of drilling takes place only in working
faces. Pre-drainage is hardly or not used at all in
Polish coal mines as coal permeability is too low;
therefore a methane drainage method is not effec-
tive enough. Moreover an increase in concentration,
intensification of mining and shortening the time of
drilling development workings makes obtaining pos-
itive results of pre-drainage of seams without putting
into use processes simulating liberation and flow of
methane impossible.
In the mines of Jastrzebie Coal Company about
59% of drained methane was obtained from drainage
of rock mass through boreholes; the remaining quan-
tity was drained from sealed cavity.
Methane and air mixture with concentration of
50–60% CH4 is drained. This method of methane
drainage results in steady stream of mixture and the
concentration of CH4 varies depending on:
– barometric pressure oscillations,
– methane inflow from bore-holes and from behind
stoppings,
– leak tightness of underground drainage pipe-lines.
The following aspects should be taken into con-
sideration when considering arguments for methane
drainage of coal deposits:
– safety as methane is drained into pipelines at the
same time reducing the quantity of methane emitted
into streams of airflow and limiting burst hazard,
– ecological aspect as natural environment is pro-
tected due to the fall in methane emission into
atmosphere,
 – economic aspect resulting from the fall in the quan-
tity of air necessary to ventilate excavations and
profit from the sale of two carriers of energy – coal
and methane.
3 METHANE UTILIZATION FROM
METHANE DRAINAGE
Although methane drained during the process of
methane drainage of mines is the fuel of full value,
containing even 30% CH4 , its utilization runs into
many problems, which is caused by contradictory
requirements of a gas producer, which is a mine and
its recipients. Mines generate gas in the process of
methane drainage as a by-product of coal mining. As
mining conditions of methane drainage change all the
time, the composition of gas and its quantity change
as well. However for a recipient the condition of the
proper process of gas utilization is its calorific value,
constant chemical constitution and keeping deliveries
of methane on a fixed level.
Methane and air mixture cannot be delivered to a
pipe network without purification and enrichment to
achieve the parameters of network natural gas. There-
fore at present the only possibility of its utilization is
its use in industrial installations. It is utilized in: boiler
– houses and driers belonging to the mine, processing
plants, heat and power generating plants, gas engines
installed in Krupinski C.M. and Pniowek C.M. Jas-
Mos C.M., Pniowek C.M. and Zofiowka C.M. are the
only ones that have opportunities of gas transmission
to recipients outside and there are no pipe networks
in other mines, so drained methane can be utilized
only on the spot. Methane utilization varies in winter
and summer due to its seasonal character and great
[mln m3]
120
drainage utilization
100
80
60 124 88
108 75
96 72 92 71
87 68
40
20
0
2000 2001 2002 2003 2004
Figure 1. Methane drainage and utilization in the mines of
Jastrzebie Coal Company from 2000 to 2004.
416
Copyright © 2006 Taylor & Francis Group plc, London, UK
differences between demand and supply as the recip-
ients of drained methane and air mixtures are mainly
heat and power plants. The figure below and table 2
present methane drainage and utilization.
3.1 Central cooling system in Pniowek Coal Mine
Since the beginning of 2001 in Pniowek Coal Mine
a combined engineering-cooling system, which is a
modern engineering trigeneration power plant, has
been in use. Trigeneration consists in generating three
different forms of energy from one main source. Power
engineering, generating heat, cooling and electric
energy from methane original energy is called CHCP
(Combined Heating, Cooling and Power Plants). At
present it is common all over the world to design (tens –
hundreds KWe ) cogeneration plants CHCP using nat-
ural gas obtained from biomass gasification, biogas
from sludge gas or waste-refuse gasification.
A combined engineering – cooling system in
Pniowek Coal Mine with power of 6.4 MWe is the first
of its kind in Poland; it uses a gas engine that enables
generating the so-called low- and high-temperature
heat, The system consists of two identical blocks oper-
ating simultaneously. One block contains the following
pieces of equipment:
– four-stroke gas engine of type TBG 6322 V 16 by
DEUTZ ENERGY GmbH company with rotational
speed of 1000 m−1 ,
– electric generator by A. Van. Kaick company with
rated power of 3993 kVA, voltage 6.3 kV and fre-
quency 50 HZ,
– one set of chillers consisting of:
– warm-water absorption chiller of type YIA HW
3B3 by YORK company (cooling power – 600 kW)
called AKM1 in block 1 and AKM3 in block 2,
– hot-water absorption chiller of type YIA HW 6C4
by YORK company (cooling power – 1730 kW)
called AKM2 in block 1 and AKM4 in block 2,
– compression chiller of type YLC 717 SE-SD 64
WCOC by YORK company (cooling power –
570 kW) called SKM1 in block 1 and SKM2 in
block 2.
Volumetric intensity of cooled water in a closed sys-
tem is equal to 150 m3 /h and the total cooling power
of one block is equal to 2.86 MWch . The second block
consisting of identical equipment is characterised by
the same parameters. A simplified scheme of an instal-
lation of a combined system is presented in figure 2.
In order to transport a cooling power from surface
chillers to an underground level, a thermally isolated
pipeline of 300 mm internal diameter was used. High
hydrostatic pressure is reduced in a three-chamber
pressure exchange system DRKA 200, called a pres-
sure lock, by Siemag company. A system of water
circulation on the surface together with pressure lock
Siemag is called a surface circulation.
 Table 2. Methane drainage and utilization in Jastrzebie Coal Company.

Total quantity and participation

Total quantity % of methane utilization
of drained methane
Mine [thousands. m3 /year] [thousands. m3 /year] [%] Way of utilization

Borynia 266,4 96,4 36 Gas boilers 2 × 1.2 MWt

Jas-Mos 8.670,0 8.388,6 97 EC Moszczenica
Krupinski 20.124,0 11.586,1 58 Aggregate
TBG 632 V16
Boiler WR
Floto-concentrate drier
Pniowek 42.100,5 29.842,1 71 EC Zofiowka
EC Moszczenica
Boilers of CHP
Plant Pniowek
Aggregate
TBG 632 V 16
Zofiowka 21.119,9 20.821,2 99 EC Zofiowka
JSW S.A. 92.280,8 70.734,4 77 Plants of JCC including:
Gas Boilers
Floto-concentrate drier
SEJ SA including:
EC Moszczenica
EC Zofiowka
Aggregate
TBG 632 V 16
Boilers WR
Total

heating network heating network

gas from drainage system Flue chimney
3
25 m CH4/min
mechanical
high-temperature cooling
tower
from the second block of chillers

heat exchanger
electric power KOWT
3,2 MW 100°C
6,3 kV
125°C

generating plant
A.Van.Kaick
G gas engine
TBG 6322 V16
warm-water hot-water compression
85°C absorption absorption chiller
low-temperature chiller 14,5°C chiller 4,5°C 1,5°C
heat exchanger 70°C 600 kW 1730 kW 570 kW
MOWT
3
to the second block of chillers 18,5°C 150 m /h

installations with water

3
coolers of air 17,5°C 300 m /h three-chamber
in mining areas 3°C pressure sluice
Siemag

Figure 2. Simplified scheme of combined engineering-cooling system in Pniowek Coal Mine.

417

Copyright © 2006 Taylor & Francis Group plc, London, UK

 An underground system of isolated and unisolated
pipelines, from pressure lock Siemag to water cool-
ers of air installed at workplaces, is an underground
circulation. After reducing the temperature and humid-
ity of air in these coolers, warm water returns through
unisolated pipelines to a Siemag. The detailed tech-
nical parameters of an air-conditioning system are
discussed in the works.
Assuming 18884 kW of chemical energy in the
drained stream of mining gas with the content of 50%
vol. CH4 , approximately 80% of fuel energy is trans-
formed into electric power and heat in a combined
power system. The participation of electric power
amounts to approximately 38% and heat to app. 42%.
Production of electric power amounts to app. 6400 kW.
As energy consumption of a combined system is equal
to 788 kW a mining network can be supplied with app.
5600 kW of electric power.
4 POSSIBILITIES OF INCREASING
METHANE UTILIZATION
Our experience shows that drained methane is used
at maximum in mines which have a net of pipelines
enabling the transport of gas to recipients and its uti-
lization is at minimum when gas is used only for a
mine’s own purposes. Drained methane in its natural
form with 80% concentration can be used as a sub-
stitute for natural gas or can be added to natural gas
(Roszkowski, Szlazak, 1999). The condition for assur-
ing the proper gas utilization for a recipient is a steady
calorific value, steady chemical constitution and keep-
ing gas supplies on the same level. Drained methane,
which is a mixture of methane and air, is characterised
by a changeable character of its parameters depending
on air content. Therefore, in order to be used, it must
be stabilised to keep the parameters on the same level.
A lot of steps have been taken for many years to
increase drained methane utilization in Polish mines.
The steps involve not only the modernisation of present
power systems but also new investments, including,
among others, central cooling system in Pniowek Coal
Mine, enrichment of methane mixtures to the parame-
ters of natural gas, attempts at storing methane excess
for some time and control over drained methane in the
process of continuous monitoring. All these factors are
necessary to utilize drained methane more effectively
and finally to intensify underground methane intake so
as to reduce the quantity of methane emitted together
with ventilation air.
4.1 Storage of drained methane excess
The technical problem to be solved is the adjustment
of the quantity of drained methane to the possibili-
ties of its utilization by its recipients during the whole
418
Copyright © 2006 Taylor & Francis Group plc, London, UK
year. The distribution of drained methane can be more
effective by storing methane excess in closed-down
galleries. That is why, the problem of adaptation of
closed-down headings for storing methane seems very
interesting.
The first attempts at storing methane were taken in
the mines of Jastrzebie Coal Company in the early
90-ties (Nawrat and Jakubow, 2000). In Morcinek
Coal Mine methane recovered from a deposit was
transformed into thermal energy. The main recipients
of the energy were: the drying house of flotation con-
centrate and the local thermal power station. Until
1993 the heat and power station of Moszczenica Coal
Mine was the main recipient of methane excess. When,
for technical reasons, a segment of transmission grid
was switched off, methane excess was emitted to atmo-
sphere. Then the idea of storing methane excess came
up, for which shaft V was used (depth – 650 m, con-
crete and tubbing support) and a 275-meter blind
heading of that shaft (the total capacity was 35,625 m2 )
in the closed down Morcinek Coal Mine. These gal-
leries were not connected with any others and no
mining works were conducted nearby. Storing methane
in closed down galleries protected against explo-
sions aims at limiting disproportions in utilization of
drained methane, which result from seasonal demand
for methane for economy and changeable methane uti-
lization connected with the production cycle involving
working and free days.
Methane storage in closed down headings is an
enterprise not only having specific economic effects
but also is an environmentally-friendly action as it
contributes to the reduction in harmful emission of
methane into atmosphere.
4.2 Enrichment of methane mixtures to the
parameters of natural gas
In the 90-ties research into recovering methane from
mining gases was conducted at the University of Sci-
ence and Technology in Cracow (Roszkowski and
Szlazak, 1999).
The process of variable pressure adsorption (PSA)
is one of the most effective methods for separation
of gases slowly liquefying. Using it for enrichment of
mining gases with methane concentration amounting
to 30–50% enables obtaining fuel with calorific value
comparable to the one of natural gas. The research con-
ducted so far determined the adsorption and structural
properties of coal adsorbents as far as their useful-
ness for separation of mixtures containing methane by
means of PSA method is concerned. On the basis of
the research it can be concluded that mainly carbon
molecular sieve and to a smaller extent assets can be
used for separation of methane mixtures. Some ini-
tial research into recovering methane from a mixture
with low methane concentration in flow conditions was
 carried out. Adsorbed methane was recovered in the
process of vacuous desorption.
The conclusion was drawn that when using our Pol-
ish active coal a nearly triple enrichment of the product
in the last stage of desorption could be obtained. In the
laboratory stage a method for conducting PSA pro-
cess was determined, which enabled obtaining a gas
stream considerably enriched in methane, even with
its low content in the main gas. At present the research
into enriching methane in a methane and air mixture is
still experimental and conducted only in laboratories.
However, its continuation may allow for eliminating
methane from atmospheric air and its industrial uti-
lization. Independently of the research into methane
recovery from mixtures of low concentration, methane
drainage should be conducted so as to obtain the high-
est concentration of methane recovered in a net of
pipelines.
4.3 Continuous monitoring system of
methane drainage
The methods for controlling methane drainage used so
far and based on periodical measurements by hand are
not sufficient at the moment.
To assure high effectiveness of a methane drainage
system requires a continuous monitoring and con-
trol over the parameters of the medium in pipelines.
Practically, it is necessary to control a network of
methane drainage system in order to keep the assumed
parameters on the same level, independently of the
changeable mining and geological conditions as well
as the environmental influence. For example, changes
in barometric pressure influence considerably methane
drainage efficiency (fig. 3). This phenomenon is well-
known and described in literature. The fall in pressure
leads to an increase in methane emission and an
increase in pressure results in a decrease in methane
emission.
The latest solutions concerning gasometrical sys-
tems allow for the continuous measurement of
methane concentration and its quantity in pipelines
of a methane drainage network (Tor and others, 2004).
Figure 3. Dependence of pure methane intake and its
concentration on barometric pressure.
419
Copyright © 2006 Taylor & Francis Group plc, London, UK
The idea of the control system is based on the measure-
ment of the medium parameters at selected points of a
methane drainage pipeline, in particular methane con-
centration, fall in pressure on orifice, absolute pressure
and temperature. Additionally, the level of carbon
monoxide in methane drainage pipelines is controlled
in order to monitor fire hazard. The parameters mea-
sured in the system allow for estimating methane
quantity drained in the pipeline. The control and mea-
surement functions mentioned above are ensured by a
sub-system based on a detector of methane drainage
parameters of CPO-1 type, which can be included
in the system monitoring gas hazards and ventilation
conditions, commonly used in Polish mines.
CPO-1 detector contains the following sub-systems
(fig. 4):
– CH4 – detector of high methane concentrations (0–
100%)
– T – detector of temperature (0–50◦ )
– Pb – detector of absolute pressure (0–120 kPa)
– P – detector of differences in pressure on a
measurement orifice (0–200 Pa)
– CO – detector of carbon monoxide content in a
pipeline (0–1000 ppm) – optionally.
The signals measured by detectors are processed in
a detector’s system and later on they are transmitted to
the central monitoring system on the surface.
Computer programmes allow for the visualisation
and registration of data obtained from detectors in
the proper areas of methane drainage network. The
system gives signals in emergency situations. In a
methane drainage network the emergency value, as
far as methane concentration is concerned, is the fall
below 30% in a pipeline. The warning value is the
fall below 40% of methane concentration at a mea-
surement point. Special computer programmes for the
system allow for current calculations of the medium’s
parameters, including values of methane flow rate as
well as for making balances for a given period of time
on the basis of instantaneous values obtained from spe-
cific detectors. An important element in presenting
the condition of a controlled fragment of a methane
Figure 4. Diagram of a detector for methane drainage
parameters CPO-1.
 drainage network is its visualisation in a spatial dia-
gram of some areas of a mine. On the basis of methane
flow rate, shift and day balances are calculated as well
as values of drained methane weekly and monthly.
5 CONCLUSIONS
In Polish conditions methane drainage is indispensable
and the most effective method for methane drainage
is by means of bore-holes from underground head-
ings. The most effective methane drainage is obtained
through long bore-holes drilled in headings located
over mining areas.
In the mines of JCC a great quantity of methane is
emitted during mining, out of which a substantial part
(over 60%) is emitted into atmosphere by means of
a ventilation system. During the whole year approx-
imately 90–100 million m3 is drained by a methane
drainage system. The highest methane bearing capac-
ity can be observed in Pniowek Coal Mine, where
approximately 42.1 million m3 of methane is drained
while the effectiveness of methane drainage system
amounts to app. 37%.
An increase in the effectiveness of methane
drainage and full utilization of gas drained may yield
significant economic results. For instance, an increase
in effectiveness by 1% results in app. one million m3
of methane drained worth 36 thousand USD. An
improvement in effectiveness of methane drainage
and utilization of drained methane lead to greenhouse
effect and affect ozone concentration in upper layers
of atmosphere. These phenomena significantly affect
global climatic changes.
Trigenerational power system (CHCP) in Pniowek
Coal Mine, designed especially for mining air-
conditioning, achieved fully satisfying parameters and
high economic effectiveness.
This paper presents rational and environmentally
friendly technological solutions applied by the mines
of JCC, aiming at limiting methane emission into
420
Copyright © 2006 Taylor & Francis Group plc, London, UK
atmosphere and, firstly, at economically profitable
methods of its recovery, excess storage and proper
utilization in electrical power engineering.
The article was written within Statutes Research
AGH, No. 11.11.100.663
REFERENCES
Nawrat St. & Jakubów A. 2000. Materialy na posiedzenie
Komisji ds. Zagrozen Metanowych, Pozarowych,
Wybuchem Pylu Weglowego oraz Przewietrzania i
Klimatyzacji w Podziemnych Zakladach Gorniczych dla
zaopiniowania projektu pt.: “Technologiczne gromadze-
nie metanu w nieczynnych wyrobiskach dolowych kopaln
“Krupinski” i “Pniowek” Suszec - KWK “Krupinski” 12
stycznia 2000.
Roszkowski J., Szlazak
˛ N., Szlazak
˛ J. 1997. Methane drainage
as a method of hazard combating and recovery of fuel.
Proceedings of the school of underground mining 1997,
Pub. CPPGSMiE PAN, Cracow, Poland (polish text).
Roszkowski J. & Szlazak N. 1999. The selected problems
of methane drainage in coal mines. Dissertations, Mono-
graphies, AGH-UST Publishers, Cracow, Poland (polish
text).
Szlazak
˛ N., Borowski M., Obracaj D. 2001. Methane drainage
of mine as a method for improving mine safety during
closing down a mine. Proceedings of the school of under-
ground mining 2001, Pub. CPPGSMiE PAN, Cracow,
Poland (polish text).
Szlazak
˛ N., Tor A., Jakubów A. 2002. Analysis of methane
drainage and utilization in mines of Jastrzebie Coal
Company. Proceedings of the second school of mine ven-
tilation. 7–11 October 2002. Zakopane, Poland (polish
text).
Szlazak
˛ N., Nawrat S., Jakubów A., 2000.: First in Poland
central air conditioning plant at Pniowek Coal Mine of
the Jastrzebie Coal Company. Przeglad Gorniczy, No. 10
(polish text).
Tor A., Jakubow A., Wasilewski S., Mroz J., Berger J.
2004. Monitoring of methane drainage process; eco-
nomic and safety aspects. Proceedings of the third school
of mine ventilation 12–15 October 2002, Zakopane,
Poland (polish text).
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

VAM to electricity – 1st large-scale installation

R. Mattus
MEGTEC Systems, Gothenburg, Sweden

ABSTRACT: At a coal mine in Australia, the very dilute methane in 150 000 scfm of ventilation air (a side
stream of 20% of the total volume) is to be converted into 6 MW of electricity. One single typical ventilation
shaft annually emits about 50 000 tons of pure methane. Besides being an environmental greenhouse problem,
this represents a major waste of energy. Project WestVAMP in Australia is likely to be the first large-scale project
in the world to utilize ventilation air methane as primary fuel. The world’s first VAM Power Plant is scheduled
for commissioning during 2006. A full-scale plant, taking the entire flow of ventilation air, would generate a net
of around 20 MW of electricity.

1 INTRODUCTION
By combining a special form of oxidizer technol-
ogy with traditional power plant boiler technology,
MEGTEC Systems has developed a patented system
for utilizing the methane of extremely low concentra-
tion as primary fuel in a power plant. The system is
based on the Vocsidizer, a regenerative thermal oxi-
dizing process with flameless, in-bed combustion and
with no generation of thermal NOx .
MEGTEC Systems is a leading global supplier of
emission control equipment for low concentrations of
hydrocarbons to air. Since the 1970s, the company
has supplied over 3000 installations, whereof over
700 are of Vocsidizer type. MEGTEC has over 800
employees and is headquartered in Wisconsin, USA.
Center of competence for the Vocsidizer is located in
Gothenburg, Sweden.
When methane in air is sufficiently heated, oxida-
tion will take place – as with all oxidizable gases. The
rate of oxidation is relatively slow until the tempera-
ture reaches (at typical retention time) levels around
800◦ C. At around 900◦ C, all methane has oxidized.
Basically, the Vocsidizer consists of a well-insulated
steel container, with a bed of ceramic material inside.
There is an air plenum on top and another one below
the ceramic bed.
For starting up the system, electrical coils heat the
center of the ceramic bed to 1000◦ C (approximately
1800◦ F). The electrical coils are then turned off prior
to starting a fan to blow the coal mine ventilation
air vertically through the ceramic bed. Being well
insulated, the ceramic material maintains the high tem-
perature and as the coal mine ventilation air is passed
through the bed, the air increases in temperature until
the methane in the air is oxidized. Heat is released
when the methane is oxidized. Continuing through the
ceramic bed, the hot air meets colder sections of the
ceramic bed material, which by efficient heat trans-
fer absorbs the heat. In order to maintain the energy
necessary for oxidation of incoming air, making the
system self sustainable with oxidizing energy, a mere
0.15% methane is required. Comparing the tempera-
ture of incoming and outgoing ventilation air, the net
increase in self sustainable mode is approx 40◦ C –
the same air having been heated to 1000◦ C. In order to
keep the heat zone centered in the bed of ceramic mate-
rial, the direction of flow of ventilation air through the
bed is altered every few minutes.
Given sufficient energy value of oxidizable gases in
air, the energy released that is in excess of the energy
required to keep the oxidization going can be retrieved.
When the energy content of methane in coal mine
ventilation air is sufficient, the excess energy can
be retrieved by embedded steam tubes at appropriate
levels in the ceramic material.
The energy from the ventilation air methane of a
typical coal mine could generate around 70 MW of
thermal energy. With a total system efficiency of 30%,
this would convert to around 20 MW of electricity.
Deep coal mines normally have a cooling requirement.
In that case, some of the electricity could be sacri-
ficed for allowing a somewhat higher temperature of
the water of the condenser. This hot water could then be
utilized in providing the driving energy for an absorp-
tion chiller that could supply cold water for chilling
the air in the mine.

421

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 1. Open cut illustration of the Vocsidizer, an industrially well proven single bed regenerative, thermal oxidizer.
Embedded tubes retrieve oxidation energy released, utilizing it to convert water to superheated steam.

Figure 2. Taking the ventilation air from a coal mine, the Vocsidizers can oxidize the extremely low methane content and
retrieve the energy released, to have it utilized for the generation of heating, of cooling and of electrical energy.

2 REFERENCE INSTALLATIONS AT COAL
MINE SITES
The first installation supplied by MEGTEC at a coal
mine site was made in the UK in 1994. The pur-
pose was to demonstrate the ability to efficiently abate
methane at the very low concentrations present in coal
mine ventilation air. 8000 Nm3 /h of ventilation air with
methane concentration mostly varying between 0.3
and 0.6% was successfully treated.
The second installation supplied by MEGTEC was
made at a coal mine site of the Appin Colliery in
Australia in 2001-2002. The purpose was two-fold; to
demonstrate that the Vocsidizer with embedded steam
tubes could utilize the energy of the low-grade Ventila-
tion Air Methane to boil water, and to demonstrate that
the system can handle normal variations of methane
concentration in coal mine ventilation air. The demon-
stration project was operated for 12 months and was
partly financed by ACARP, Australian Coal Associ-
ation Research Programme. On 5 April 2005, this
demonstration project was awarded as being the best
Green House Gas project funded by ACARP.
Based on the successful results at the Appin Col-
liery, project WestVAMP was agreed upon in 2004.
As was the case with the Appin project, the cus-
tomer is BHP Billiton in Australia, one of the world’s
largest resource companies, with part financing from
the AGO, Australian Greenhouse Office. Located at
the WestCliff Colliery, south of Sydney, WestVAMP
(WestCliff Ventilation Air Methane Project) will oxi-
dize the methane of 250 000 Nm3 /h (150 000 scfm)

422

Copyright © 2006 Taylor & Francis Group plc, London, UK

 of ventilation air and utilize the inherent energy to
generate 6 MW of electricity. The air volume repre-
sents around one fifth of the full volume flow of the
ventilation shaft.
Following the site installation of the four Vocsi-
dizers in July, the power plant is completed with
equipment belonging to the boiler steam cycle and to
the steam turbine-generator system. Start up of the
power plant is planned during 2006.
Figure 3. Demonstrating VAM abatement in 1994.
Figure 4. Award winning “water boiler” 2001–02.
Figure 5. Principle layout of project WestVAMP, converting Ventilation Air Methane to electricity.
423
Copyright © 2006 Taylor & Francis Group plc, London, UK
WestVAMP will most likely be the first large-scale
installation to utilize VAM as the primary source of
energy.
An installation taking the full flow of ventilation air
from a major coal mine generates a net of 15-25 MW
of electricity, abating around 50 000 tons of methane,
which corresponds to around 1 million tons of CO2 .
Oxidizing and utilizing the VAM emissions from such
a coal mine has the same annual effect on global warm-
ing as closing down a coal-fired power plant of several
hundred MW, or as taking up to ½ million cars off the
streets.
3 CONCLUSIONS
• Coal mines are major sources of emissions for
methane. The major part of emissions to atmosphere
is in very low concentrations, typically below 1%,
but in very large volumes of ventilation air, often
around one million Nm3 /h from one coal mine shaft.
• By combining proven technologies, MEGTEC has
demonstrated the ability to abate the methane emit-
ted and to convert the energy into useful form to be
utilized for heating, for cooling or as electricity.
• Project WestVAMP being supplied to BHP Billiton
in Australia, planned for startup in 2006, will most
likely be the first large scale installation to utilize
VAM (Ventilation Air Methane) as primary fuel.
ACKNOWLEDGEMENTS
Special thanks to Mr. Ake Kallstrand, Manager for
Research and Development at MEGTEC, for having
 mastered the technical opportunities of this new field
of application, which has the potential of contribut-
ing to a major reduction in emissions of the Green
House Gas methane. Thanks to Mr. Mohit Uberoi,
President of MEGTEC Systems, who with the patient
understanding of the importance of this technology
has provided sufficient internal resources. Thanks to
424
Copyright © 2006 Taylor & Francis Group plc, London, UK
Mr. Richard Danell and colleagues of BHP Billiton,
having contributed with important input and discus-
sions in the development of the technology. And thanks
to Dr. Andrew Weavers and colleagues of Australian
Greenhouse Office (AGO) for supporting the world’s
first large-scale Power Plant Demo fuelled primarily
by Ventilation Air Methane.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Development and application of reservoir models for the evaluation and

optimization of longwall methane control systems

C.Ö. Karacan, W.P. Diamond, S.J. Schatzel & F. Garcia

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute for
Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, Pennsylvania, USA

ABSTRACT: Methane explosions have historically been one of the major causes of fatalities and injuries in
underground coal mining operations. Advanced numerical models and predictive modeling approaches have
the potential to offer optimized methane control solutions for general mine planning purposes and to address
specific methane-related operational problems. This paper describes the development of reservoir models for
the longwall mining environment and their application for investigating the influence of various completion
design parameters on the methane drainage effectiveness of gob gas ventholes. The influence of increasing
longwall panel width on the effectiveness of current gob gas venthole completion and placement strategies in the
Pittsburgh Coalbed were evaluated and optimized designs developed to capture the expected increase in methane
emissions on the larger panel.

1 INTRODUCTION The thickness of the fractured zone can vary up to

Comprehensive assessments of the need for additional
methane control capacity beyond ventilation often
require both an empirical and theoretical approach for
an adequate or timely control of increased methane
emission levels. Thus, the prediction of methane emis-
sions and optimization of methane control systems
prior to starting a new mining operation will be a major
improvement towards eliminating the explosions in the
underground workplace.
During longwall mining, the caving of immedi-
ate strata and stress relief create horizontal fractures
along bedding planes and vertical fractures in the strata
overlying the caved zone. These fractures provide an
extensive pathway for gas migration from the sur-
rounding coalbeds and other gas bearing strata into
the longwall mining environment (Fig. 1).
Figure 1. Schematic cross-section view of subsided strata
zones and methane control system influence zones above
mined-out longwall panel (Mucho et al. 2000).
100 times the height of the mined coalbed (Palchik
2003). The fractured and caved rock mass left behind
the advancing longwall face is generally referred to
collectively as “gob” (Fig. 1). The methane that origi-
nates and accumulates in the gob above the mined-out
longwall panel is the main source of potential gas
emissions during longwall mining.
Gob gas extraction in the Northeastern U.S. is
almost exclusively accomplished using ventholes that
are drilled from the surface to within a short distance
[typically 10–15 m (30–45 ft)] of the coalbed being
mined (Diamond 1994). Commonly, the bottom sec-
tion of the well casing [generally about 60 m (200 ft)]
is slotted. The gob gas ventholes generally become
productive only when the mining-induced fractures
are created as mining advances under the venthole
(Diamond 1994).
Numerical models offer effective capabilities for
predicting methane emissions, and designing drainage
systems accordingly. There have been reported studies
using computational fluid dynamics (CFD), bound-
ary element and finite element modeling (FEM)
techniques to better characterize the parameters for
gas emission prediction (Ren & Edwards 2002,
Lunarzewski 1998, Tomita et al. 2003).
Reservoir simulators developed over the years can
represent the complex reservoir flow mechanisms in
coalbeds (King & Ertekin 1991). However, a com-
prehensive reservoir model capable of realistically
representing various aspects of mining operations and

425

Copyright © 2006 Taylor & Francis Group plc, London, UK

 production from gob gas ventholes has not previ-
ously existed. In one study, Zuber (1997) modeled
the face and rib emissions during development min-
ing. Karacan et al. (in prep., 2005) and Esterhuizen &
Karacan (2005) have developed “dynamic” reser-
voir models that include the subsided strata above
the mined panel during longwall mining to evaluate
methane emissions and various gob gas venthole
design factors for their impact on gas drainage
efficiencies.
2 MODEL DEVELOPMENT
The models summarized in this paper were developed
for mine sites operating in the Pittsburgh Coalbed in
the Southwestern Pennsylvania section of the North-
ern Appalachian Basin. The reservoir models were
constructed using Computer Modeling Group’s (2003)
GEM compositional reservoir simulator.
2.1 General description of the mine sites
Overburden depths in the area range between 152
and 274 m (500 and 900 ft). Longwall panels in the
old mining districts of this mine were around 253 m
(830 ft) wide and were increased to 305 m (1000 ft).
In the new districts, the panel widths were originally
430 m (1250 ft); however, the first 480 m (1450 ft)
wide panel is currently being mined. Thus, all the
panels, particularly the recent ones, are super-critical,
which results in a more complete caving of the
overburden strata into the mine void.
A generalized stratigraphic section for the study
area is shown in Figure 2. Several coalbeds with a
combined thickness of almost 3 m (10 ft) are present
in the 26 m (85 ft) of strata immediately above the
Figure 2. A generalized stratigraphic section of the strata
above the Pittsburgh Coalbed in the study area (Mucho et al.
2000).
426
Copyright © 2006 Taylor & Francis Group plc, London, UK
Pittsburgh Coalbed. Within this interval, the thickest
coalbed is the Sewickley, which is about 25 m (75 ft)
above the Pittsburgh Coalbed. Between these two
major coalbeds, there are comparably thin Pittsburgh
rider coals and the discontinuous Redstone Coalbed,
which are contained in the caved zone after panel
extraction (Fig. 1). The gas emissions associated with
the caved zone report to the bleeder ventilation system
(Mucho et al. 2000). In some parts of the study area
a sandstone paleochannel replaces the shale unit
usually present above the Pittsburgh Coalbed. The
thickness of the paleochannel varies between 0–13 m
(0–40 ft). Regionally, the Sewickley Coalbed may split
into two separate benches, and its height above the
Pittsburgh Coalbed may vary. Gas contained in the
fractured zone (Fig. 1), in particular, gas in the Sewick-
ley Coalbed, primarily reports to the gob gas ventholes,
if they are present and operational (Mucho et al. 2000).
2.2 Grid model generation for longwall sites
In order to model the longwall mining process and
analyze the associated methane control systems, a
three-dimensional grid model of the mine site has to
be created. The horizontal dimension of the grid mod-
els were usually determined based on the problem type
and the total area of interest. The number of vertical
layers and their thicknesses were based on generalized
stratigraphic sections for the mine site.
The Pittsburgh Coalbed (mining) layer was con-
structed differently from the other layers in the grid
model to host both the mined and unmined Pittsburgh
Coalbed, and the entries surrounding the longwall pan-
els. An example showing this structure for one of the
study sites is given in Figure 3.
2.3 Gob gas ventholes and the pseudo-ventilation
system
A simplified version of the ventilation system was
incorporated into the model. For each panel, a set of
wells injecting air into the entries with a rate con-
straint of 1700 m3 /min (60,000 cfm) represented the
Entries
Gob
Unmined
coal
Mining direction
Figure 3. Pittsburgh Coalbed layer as represented in the
models.
 air intake part of the ventilation system. The exhaust- Case-1 (Fig. 4) was developed for a previous, multi-
ing bleeder fan at the top of a 1.8 m (6 ft) diameter air panel mining district, and was used to evaluate the
shaft was modeled with a large-diameter vertical well influence of various gob gas venthole completion
on the tailgate side of the panel and operated with a parameters on methane capture. The grid model shown
bottom-hole pressure constraint of 1.36 kPa (0.2 psia) in Figure 5 (Case-2) was developed for the new mining
negative pressure. district that started with 381 m (1250 ft)-wide panels,
The locations of gob gas ventholes in the model but would be switching to 442 m (1450 ft)-wide pan-
were determined based on their locations on the study els. The focus of the Case-2 study was to estimate
panels, and they were configured based on their actual the increase in expected methane emissions and to
reported completion data. The ventholes are usually investigate alternative gob gas venthole completion
drilled to within about 12–13 m (40–45 ft) of the top and placement scenarios on the larger panel.
of the Pittsburgh Coalbed at this mine site, and 17.8 cm
(7 in) casing, with 61 m (200 ft) of slotted pipe on
3.1 Case-1, evaluation of gob gas venthole
the bottom is installed as shown for Venthole A, in
completion parameters
Figure 1. However, in some case, the ventholes were
drilled closer to the mining horizon and into the caved 3.1.1 Effect of slotted-casing diameter
zone (Venthole B, Fig. 1), which affects their perfor- The standard casing diameter for the gob gas ventholes
mance, as will be discussed in the following sections. in the study area was 17.8 cm (7 in). The effects of
different slotted-casing diameters [25.4 cm (10 in) and
2.4 Geomechanical calculations for strata 10.2 cm (4 in)] on methane production were evaluated.
response and permeability changes The length of the slotted casing and its setting depth
above the top of the Pittsburgh Coalbed were held
The geomechanical, fast lagrangian analysis of con-
tinua (FLAC) model (Ithasca Consulting Group, Inc.
2000) was used to evaluate the effects of longwall
mining on the surrounding rock mass and to calculate
mining-related permeability changes.
Calculation of permeability changes was accom-
plished using the final stress and rock failure distri-
butions from the FLAC model runs, and employing
empirical relations (Ren & Edwards 2002, Lowndes
et al. 2002). Details of the model development, perme-
ability calculations, and geomechanical analysis are
given in Esterhuizen & Karacan (2005).

2.5 Simulation and model-calibration strategy

During longwall mining, the strata disturbances and
the onset of production in successive gob gas ventholes
move along with the face. This leads to a moving-
boundary problem in modeling, which was addressed Figure 4. Study mine Case-1, grid model of a multi-panel
with “restart” models. Each model restart run was mining site, Pittsburgh Coalbed.
performed so that it would progress up to either the
Gob Gas Ventholes
next venthole location or to a defined location on the
panel for the distance and time characterizing the inter- Air Intake and Return
VH 1-4
VH 1-3 VH 1-2 VH 1-1
vening face movement (Karacan et al. in prep.). The
reservoir-parameter changes were incorporated into Bleeder
the model as the face was advanced between restarts
during calibration and prediction runs.

3 MODEL APPLICATIONS

Figures 4 and 5 show the two grid models constructed

for this study. In both figures, only the coalbeds are
Pittsburgh Coalbed
depicted to improve the visualization. The models Sandstone Channel Sewickley Coalbed
were calibrated by matching the measured gas pro-
duction rates, methane concentrations in the produced Figure 5. Study mine Case-2, grid model of a new mining
gas stream, and the flowing bottom-hole pressures. district, Pittsburgh Coalbed.

427

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 1. Predicted effect of casing diameter on cumulative
methane and total gas production for 910 days of simulated
mining.
CH4 Air
Casing Cum. compared Cum. Gas compared
diameter CH4 , to 7-in (CH4 + Air) to 7-in
inch MMscf casing, % MMscf casing, %
4 391.8 −6.7 609.4 −15.2
7 419.8 – 676.5 – 50
10 440.1 +4.9 728.4 +12.3
constant at their original design values, 61 m (200 ft)
and 12 m (40 ft), respectively.
The modeling results given in Table 1 for the
simulated mining period (910 days) predict that the
cumulative methane production using the 25.4 cm
(10 in) casing will increase 4.9%, as compared to the
17.8 cm (7 in) standard diameter casing. The amount of
methane produced with the 10.2 cm (4 in) casing was
about 6.7% less than that produced with the 17.8 cm
(7 in) diameter casing. However, the amount of mine
air produced with the 10 in casing was 12.3% more,
which resulted in a lower predicted methane concen-
tration. Conversely, the amount of mine air produced
with the 10.2 cm (4 in) diameter casing was 15.2% less,
as compared to the standard casing, which resulted in
higher methane concentrations.
The predicted increase in cumulative methane pro-
duction with the larger diameter wellbore was due to
the increase in the open-to-flow area of the wellbore.
Also, with larger diameter wellbores, the calculated
pressure losses were less compared to smaller diameter
wellbores. The predicted reduction in methane concen-
tration with the 25.4 cm (10 in) diameter casing is most
likely the result of more mine air being captured due
to an expanded pressure sink and associated deple-
tion radius created by the production of gas from the
larger diameter casing. However, since the total pre-
dicted gas production (methane and air) was higher for
the 25.4 cm (10 in) diameter casing, it still resulted in
higher cumulative methane production, even though
the methane concentration was less.
3.1.2 Effect of slotted casing length
To evaluate the influence of the length of the com-
pletion interval on gob gas venthole performance, the
length of the slotted casing section was changed in
the model to 30.5 m (100 ft) and to 76.2 m (250 ft),
as compared to the original 61 m (200 ft) length.
The casing diameter was kept at 17.8 cm (7 in), and
the setting depth of 12 m (40 ft) above the top of
the Pittsburgh Coalbed was maintained. The modeling
results predict that the cumulative methane production
will increase with increases in slotted casing length
428
Copyright © 2006 Taylor & Francis Group plc, London, UK
500
Cumulative Methane Production, MMSCF
100-ft slotted casing
450
200-ft slotted casing
400 250-ft slotted casing
350
300
250
200
150
100
0
0 200 400 600 800 1000
Time from Start of Mining (Days)
Figure 6. Simulated cumulative methane production from
ventholes with varying lengths of slotted casing.
(Fig. 6). The methane production with 76.2-m (250-
ft) of slotted casing was 459.4 MMscf, as compared
to 391.8 MMscf with the standard 61 m (200 ft) of
slotted casing. This difference corresponds to a 9.5%
increase in methane capture from the four panels mod-
eled in Case-1 (Fig. 4). However, when the slotted
casing length was shortened to 30.5 m (100 ft), the pre-
dicted methane production decreased to 314.7 MMscf,
which was about 25% less than what was captured with
61 m (200 ft) of slotted casing.
3.1.3 Effect of slotted-casing setting depth
The effect of casing setting depth (distance from the
top of the mining layer) on gas production was inves-
tigated by modeling alternative completion depths of
19.8 m (65 ft), 7.6 m (25 ft) and 4.6 m (15 ft), as com-
pared to the original, 12 m (40 ft) depth. In these
alternative cases, the 7.6 m (25 ft) completion depth
generally corresponded to a close proximity to the
caved zone, which was modeled as 7.3 m (24 ft) above
the Pittsburgh Coalbed for the Case-1 study site, and
the 4.6 m (15 ft) depth corresponded to circumstances
where the venthole was drilled into the caved zone. The
19.8 m (65 ft) completion depth corresponds strati-
graphically to a depth slightly below the Sewickley
Coalbed (Fig. 2). For these scenarios, the casing diam-
eter and slotted casing lengths were kept at their
original design values, 17.8 cm (7 in) and 61 m (200 ft),
respectively.
Raising the slotted casing setting depth to 19.8 m
(65 ft), as compared to 12 m (40 ft), above the
Pittsburgh Coalbed resulted in a 4% predicted cumu-
lative methane production increase. The predicted
cumulative methane production declined by about 5%
and 29% when the casing was set to within 7.3 m (25 ft)
and 4.6 m (15 ft) of the top of the mining layer, respec-
tively. In the 15 ft setting depth scenario, the lower
slots of the casing were in the caved zone influenced
by the mine ventilation system where flow resistance
was small. Therefore, the ventholes pulled 74% more
mine air, as compared to the operator’s standard ∼12 m
 65-ft from the coalbed 100
90 45-ft from the coalbed
90
25-ft from the coalbed

Methane Concentration, %
80 15-ft from the coalbed 80
Methane Concentration, %

70 70
60
60
50
50 40
40 30
20
30 Depth: 47 ft
10 Depth: 35 ft
20 0
0 200 400 600 800 1000
0 50 100 150 200 250 300
Time from the Start of Mining (Days)
Time from the Start of Mining (Days)

Figure 7. Simulated methane concentrations from vent-

holes completed to varying depths above the Pittsburgh Figure 8. Actual methane concentrations measured from
Coalbed. two gob gas ventholes completed to different depths.

(∼40 ft) setting depth. Since most of the produced gas length of approximately 3350 m (11,000 ft). However,
was mine air at the 4.6 m (15 ft), the average methane starting with the third panel, the panel widths were
concentration in the cumulative produced gas at the to be increased by 61 m (200 ft) to 442 m (1450 ft).
end of mining was about 40%, as opposed to 60–70% Due to the uncertainty of the methane emission con-
average methane concentration calculated for other sequences associated with mining of the larger panel,
depths (Fig. 7). the area in question (Fig. 5) was modeled to estimate
A real-world example of the gas quality conse- the expected increase in gas flow and to investigate
quences of completing gob gas ventholes into the methane control options. The question to be answered
caved zone is illustrated with measured gas concentra- was whether the current number and configurations
tion data from two ventholes continuously monitored of gob gas ventholes would adequately control the
in the Case-2 study area (Fig. 8). For this site, the height projected increase in gob gas on the larger longwall
of the caved zone was estimated to be ∼12 m (∼40 ft), panels (Karacan et al. 2005).
higher than the Case-1 site shown in Figure 4, due to the
presence of the sandstone paleochannel. The first ven- 3.2.1 Evaluation of the increase in methane
thole on the study panel (1-1, Fig. 5) was completed to emissions due to the mining of a wider
a depth of 14.3 m (47 ft) above the top of the Pittsburgh longwall panel
Coalbed, generally within the standard depth range for Modeling the increase in panel width from 381 m
the mine site. However, the second venthole drilled on (1250 ft) to 442 m (1450 ft) results in about 47 MMscf
the study panel (1-2, Fig. 5) was inadvertently drilled of additional methane liberation from the coal mined
deeper to a depth of 10.6 m (35 ft) above the top of on the longwall face and 137 MMscf from the overly-
the Pittsburgh Coalbed, which is in the caved zone. ing disturbed strata over the 268 days of mining sim-
As shown in Figure 8, the methane concentration in ulated for this study. Depending on the availability
the produced gas from the venthole completed into the of additional gob gas drainage capacity, some of the
caved zone averaged about 30% less than that of the additional 137 MMscf of methane originating in the
standard completion depth above the caved zone due overlying strata may report to the ventilation sys-
to the increased production of mine ventilation air. It tem. This would represent the potential of up to
should be noted that even though the estimated caved about 355 cfm of additional methane entering the
zone height for which the predictions for the influence underground workplace.
of different casing setting-depths were made (Case-
1) is slightly different than that at the Case-2 study 3.2.2 Analysis of expanding panel width from
site where these actual field production data were 381 m (1250 ft) to 442 m (1450 ft) with the
measured, a similar methane concentration decrease four actual ventholes in operation
(25–30%) was predicted for the venthole penetrating The four actual gob gas ventholes on the first panel
into caved zone for the Case-1 study (Fig. 7). in this new mining district at the study mine site were
drilled to varying depths [14, 11, 9, and 12 m (47, 35,
30, and 40 ft)] above the top of the Pittsburgh Coalbed,
3.2 Case-2, evaluation of longwall panel width on
and, therefore, two of the ventholes were completed
gob gas venthole performance
into or at the top of the caved zone, as opposed to the
The first two longwall panels in the new mining district preferred ∼12 m (∼40 ft) distance. For this analysis,
at the study mine site were 381 m (1250 ft) wide with a the results of simulation runs for the 381 m (1250 ft)

429

Copyright © 2006 Taylor & Francis Group plc, London, UK

 wide panel with the four actual gob gas ventholes
were compared to the results when the panel width was
increased to 442 m (1450 ft), using the same venthole
placement [91 m (300 ft) from the tailgate side of the
panel], completion, and production histories (Fig. 9).
The performance of the four individual gob gas ven-
tholes on the simulated wider panel was very similar
to those on the original panels since the wellbore flow
model is not dependent on panel width and due to
the similar reservoir permeability, irrespective of the
panel width as confirmed by the FLAC computations
(Karacan et al. 2005). Since the performance (gas pro-
duction potential) of individual gob gas ventholes is
not influenced significantly with an increased panel
width, the additional 137 MMscf (355 cfm) of methane
released from the overlying strata as a result of mining
the larger panel will potentially enter the underground
workplace, if additional methane drainage capacity is
not provided. Thus, the constructed model (Fig. 5) was
used to evaluate multiple scenarios to optimize the
number and locations of the gob gas ventholes on the
wider panel and to minimize the volume of additional
methane entering the ventilation system.
3.2.3 Alternative gob gas venthole placement
scenarios and performance analysis
for a 442 m (1450 ft) wide panel
The simulated alternative gob gas venthole placement
and completion scenarios investigated for this study,
Base Case Base Case
Tailgate 1250-ft
Headgate 1450-ft
Scenario A Scenario B
1450-ft
Scenario C Scenario D
1450-ft
Scenario E Scenario F
1450-ft
Optimal wells
Actual wells Mining direction
Figure 9. A schematic representation of gob gas venthole
configurations for the simulation of methane control options
on wider longwall panels.
430
Copyright © 2006 Taylor & Francis Group plc, London, UK
as shown in Figure 9, were: (A) moving the four actual
ventholes (as drilled and completed on the first panel
in the Case-2 study area) to locations 152 m (500 ft)
from tailgate side, i.e. 61 m (200 ft) closer to the cen-
terline of the panel than on the first panel; (B) adding
four optimal (continuously operating with 2.7 psi suc-
tion pressure and completed to 12 m (40 ft) above the
Pittsburgh Coalbed) infill ventholes located between
each actual venthole; (C) adding four optimal ven-
tholes located 90 m (300 ft) from the gateroads on
the headgate side of the panel, positioned diagonally
to the actual ventholes; and (D) adding four optimal
ventholes located 90 m (300 ft) from the gateroads on
the headgate side and positioned directly opposite the
actual ventholes on the tailgate.
Simulation runs showed that the optimal gob gas
ventholes produced gas that is 85–90% methane
through the entire mining period, as compared to the
65–70% range for the actual wells during most of their
production history (due to close proximity of two of
the holes to the caved zone). Optimal wells were also
predicted to produce about 50% more methane than
the actual wells.
The data presented in Table 2 compare the cumu-
lative methane production volumes from each of the
gob gas venthole configuration scenarios shown in
Figure 9. The lowest predicted cumulative methane
production (Case A) is obtained when the four vent-
holes with the actual completions are located 152 m
(500 ft) from the tailgate entry of the 442 m (1450 ft)
panel. This configuration produces 0.37 × 106 m3
(13 MMscf) less methane compared to the base case
production from the 381 m (1250 ft) panel, thus illus-
trating the importance of near-margin venthole place-
ment (Diamond et al. 1994).
The highest predicted cumulative methane pro-
duction is achieved when additional optimal infill
ventholes are used (Table 2). Case D, with four actual
ventholes on the tailgate side of the panel and four
additional optimal headgate ventholes located directly
across from the tailgate holes produces 3.91 × 106 m3
Table 2. Cumulative predicted methane production differ-
ences from gob gas ventholes on a 442 m (1450 ft) wide
panel.
Venthole Cumulative methane production
placement difference compared to 381 m
scenario ID (1250 ft) base case, MMscf
442 m (1450-ft) −0.9
Base Case
A −13
B 113
C 119
D 138
E 86
F 134
 (138 MMscf ) more methane compared to the four
actual ventholes on the 381 m (1250 ft) panel base
case. Case C, with four actual ventholes on the tailgate
side of the panel and four additional optimal vent-
holes located on the headgate side of the panel (but
diagonal to the tailgate ventholes) is the next high-
est incremental producer of methane at 3.37 × 106 m3
(119 MMscf ). While Cases C and D are similar, Case
D is probably the higher producer because the tail-
gate and headgate ventholes are closer to each other
(since they are directly opposite each other on the
panel) and are intercepted by mining at the same
time. This scenario results in a quicker overlap of
the venthole drainage radiuses, which enhances gas
desorption from the overlying coalbeds associated
with the subsided strata. Also, when headgate and
tailgate ventholes are intercepted at the same time,
the headgate ventholes start producing earlier and
stay on production longer, as compared to the diag-
onal location in Case C, resulting in more methane
production.
One of the main considerations in gob gas vent-
hole design and operation is to locate and drill the
ventholes optimally and operate them continuously to
minimize the number of ventholes while maximizing
the production. Thus, Cases E and F were simulated
to determine the minimum number of optimal vent-
holes that will produce the same maximum amount
of methane as in the highest-producer configuration
(Case D). In Case E, six optimal ventholes were placed
along the tailgate side of the panel, and in Case F, two
of the optimal ventholes were placed on the headgate
side directly opposite the first two ventholes on the
tailgate side (Fig. 9).
The predicted methane production performance of
venthole configurations E and F using six optimal
gob gas ventholes were compared to the performance
of Case D, the highest predicted methane producer
[3.91 × 106 m3 (138 MMscf )] in the previous simu-
lations. Table 2 shows that Case E, where six optimal
ventholes were located along the tailgate side of the
Table 3. Predicted uncaptured methane volumes poten-
tially available for flow for simulated gob gas venthole
configurations on the 442 m (1450 ft) wide panel.
Venthole Cumulative uncaptured methane
placement volume compared to the 1250 ft
scenario ID base case, MMscf (cfm)
442 m (1450-ft) 137 (355)
Base Case
A 148 (380)
B 25 (65)
C 17 (44)
D −1 (−)
E 50 (130)
F 3 (10)
panel, produced about 2.44 × 106 m3 (86 MMscf )
more methane than the 381 m (1250 ft) base case,
whereas Case F, with two of the optimal wells on the
headgate side, produced 3.79 × 106 m3 (134 MMscf)
more methane. Thus, the six optimal ventholes of
Case F produced almost as much methane as the eight
(four actual and four optimal) ventholes of Case D. The
predicted performance differences between Case E
and F are due to the location of the ventholes and to
the length of time they stay on production.
3.2.4 Evaluation of the impact of gob gas venthole
configurations on controlling ventilation
system methane on a wider face
Table 3 presents the predicted amount of uncap-
tured methane that will be available for flow to the
ventilation system on a 442 m (1450 ft) wide panel
with the various gob gas venthole configurations that
have been simulated relative to the gob gas produced
volume for the 381 m (1250 ft) base case panel. The
greatest amount of uncaptured methane available for
potential flow from the overlying disturbed strata to
the ventilation system [4.19 × 106 m3 or 0.179 m3 /s
(148 MMscf or 380 cfm)] is Case A, where the gob
gas ventholes are placed an additional 61 m (200 ft)
towards the center line of the panel. The minimum
amount of uncaptured methane on the wider panel
occurs with Case D (−1 MMscf ) when the four actual
tailgate ventholes are supplemented with four optimal
headgate ventholes. The minimal uncaptured methane
flow into the ventilation system for Case D (eight ven-
tholes) is comparable to the 0.85 × 106 m3 (3 MMscf )
of uncaptured methane for Case F, where only six
optimal ventholes are used.
4 SUMMARY
Comprehensive “dynamic” reservoir models for sim-
ulating gas flows associated with the longwall mining
environment have been developed. The models were
constructed based on reservoir parameter data avail-
able for the study site and through history matching
techniques for those parameters that were not easily
obtainable. Simulating the mining component of the
model, and updating the changing reservoir param-
eters due to strata disturbance, were addressed with
the use of “restart” models, which were configured
based on a schedule built from the actual longwall
face advance and gob gas venthole interception times.
Theoretical results indicated that, while keeping the
other completion parameters constant, increasing the
casing diameter increased cumulative methane pro-
duction from subsided strata. However, a marginal
decrease in the produced methane concentration was
evident, possibly due to increased mine-air extraction.
It was demonstrated that longer slotted casing lengths

431

Copyright © 2006 Taylor & Francis Group plc, London, UK

 produced more gob gas, and thus more methane,
depending on the geology and the presence of gas bear-
ing strata in the horizons adjacent to the extra slotted
casing length. Most importantly, it was demonstrated
that casing setting depth played an important role rel-
ative to the methane concentration in produced gas
stream and the volume of methane captured. Modeling
results showed that when the setting depth was within
the estimated caved zone, the methane concentration in
the produced gas decreased by about 30%. Conversely,
raising the casing setting depth into the fractured zone
above the caved zone increased methane production.
The models were also used to predict the methane
emission consequences of mining a wider longwall
panel in the Pittsburgh Coalbed. It is predicted that
increasing the panel width from 381 m (1250 ft) to
442 m (1450 ft) would result in about 3.88 × 106 m3
(137 MMscf ) from the overlying disturbed strata,
resulting in an average of about 0.167 m3 /s (355 cfm)
of additional methane potentially entering the ventila-
tion system over the 268 days of simulated mining for
this study. The importance of completing and operat-
ing the ventholes optimally to increase their methane
control capability was also demonstrated. The model
simulations predicted that six optimally completed,
placed, and operated ventholes could produce the
same amount of methane on the wider panel as eight
non-optimal wells would produce.
Reservoir modeling has been shown to be a viable
approach for evaluating methane emission and con-
trol issues in the longwall mining environment. This
approach is far superior to the traditional trial-and-
error methods, and has the capability of address-
ing unexpected methane emission problems as they
evolve.
DISCLAIMER
The findings and conclusions in this report are those of
the authors and do not necessarily represent the views
of the National Institute for Occupational Safety and
Health.
REFERENCES
Computer Modeling Group Ltd. 2003. Generalized equa-
tion of state model-GEM. User’s Guide. Calgary, Alberta,
Canada.
Diamond, W.P. 1994. Methane control for underground
coal mines. U .S. Bureau of Mines, Information Circular
No. 9395.
432
Copyright © 2006 Taylor & Francis Group plc, London, UK
Diamond, W.P., Jeran, P.W. & Trevits, M.A. 1994. Evaluation
of alternative placement of longwall gob gas ventholes for
optimum performance. U .S. Bureau of Mines, Report of
Investigations No. 9500.
Esterhuizen, G.S. & Karacan, C.O. 2005. Development of
numerical models to investigate permeability changes
and gas emission around longwall mining panels. Proc.
AlaskaRocks 2005, 40th US Symposium on Rock Mechan-
ics, Anchorage, Alaska, 25–26 June.
Ithasca Consulting Group, Inc. 2000. FLAC-Fast lagrangian
analysis of continua. User’s Guide. Minneapolis,
Minnesota.
Karacan, C.O., Esterhuizen, G.S., Schatzel, S.J. &
Diamond, W.P., in prep. Reservoir simulation-based mod-
eling for characterizing longwall methane emissions and
gob gas venthole production. To be submitted to the
International Journal of Coal Geology.
Karacan, C.O., Diamond, W.P., Esterhuizen, G.S. &
Schatzel, S.J. 2005. Numerical analysis of the impact
of longwall panel width on methane emissions and
performance of gob gas ventholes. 2005 International
Coalbed Methane Symposium, Paper 0505. Tuscaloosa,
AL. 18–19 May.
King, G. & Ertekin, T. 1991. State of the art modeling for
unconventional gas recovery. SPE Formation Evaluation
March: 63–72.
Lowndes, I.S., Reddish, D.J., Ren, T.X., Whittles, D.N. &
Hargreaves, D.M. 2002. Improved modeling to support
the prediction of gas migration and emission from active
longwall panels. In E. De Souza (ed.), Mine Ventilation:
267–272. Balkema.
Lunarzewski, L. 1998. Gas emission prediction and recovery
in underground coal mines. International Journal of Coal
Geology 35: 117–145.
Mucho, T.P., Diamond, W.P., Garcia, F., Byars, J.D. &
Cario, S.L. 2000. Implications of recent NIOSH tracer gas
studies on bleeder and gob gas ventilation design. Society
of Mining Engineers Annual Meeting, 28 Feb.–1 March,
Salt Lake City.
Palchik, V. 2003. Formation of fractured zones in overbur-
den due to longwall mining. Environmental Geology 44:
28–38.
Ren, T.X. & Edwards, J.S. 2002. Goaf gas modeling tech-
niques to maximize methane capture from surface gob
wells. In E. De Souza (ed.), Mine Ventilation: 279–286.
Balkema.
Tomita, S., Deguchi, G., Matsuyama, S., Li, H. &
Kawahara, H. 2003. Development of a simulation pro-
gram to predict gas emission based on 3D stress anal-
ysis. 30th International Conference of Safety in Mines
Research Institutes, South African Institute of Mining and
Metallurgy: 69–76.
Zuber, M.D. 1997. Application of coalbed methane reservoir
simulators for estimation of methane emissions in long-
wall mining. Proceedings of the 6th International Mine
Ventilation Congress, 17–22 May, Pittsburgh: 435–440.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Optimum widths of longwall panels in highly gassy mines – Part I

P.C. Thakur
CONSOL Energy Inc., Morgantown, WV, USA

ABSTRACT: Longwall mining is the safest and the most efficient method of mining coal. Currently a little
more than fifty percent of all underground mined coal is mined by this method. Driven by the demands of higher
productivity, the prevalent trend is for longer and wider panels, higher horsepower equipment and faster rates
of extraction. In mildly gassy mines (gas contents less than 3 m3/t) and even in moderately gassy mines (gas
contents between 3 and 7 m3/t), limits on the width of panel and rates of extraction have not been reached. But
in highly gassy mines (gas contents between 7 and 20 m3/t), there are indications that limits on both the width
of the panel and rates of extraction are being reached. The width of the longwall panel is primarily limited by
the volume of air reaching the tail-end of the face. The ventilation air quantities reaching the tail-end of the face
should be not only enough to keep the methane concentrations below statutory limits (generally one percent)
but also enough to prevent any gas layering. A third criterion for the adequacy of ventilation air is the ability
of ventilation air leaking into the gob to push away the explosive methane-air mixture away from the gob area
immediately behind the longwall face where active roof fall is still taking place (typically 30–45 m). Ventilation
air reaching the tail-end of the longwall face depends on the air quantity reaching the head gate-end of the
longwall face and air leak-off on the face. In general, the wider the longwall face, the higher is the leak-off on the
face. The second group of variables that also controls the width of the longwall panel consists of (a) the rate of
advance, (b) bleeder air quantities (c) specific methane emission for the gob areas and (d) efficiency and cost of
gob gas capture. This paper will discuss only the first group of variables that controls the longwall panel width.
The influence of the second group of variables will be discussed in a separate paper as a sequel to this paper.

1 INTRODUCTION mined coal is produced by longwall faces operating in

Energy whether it is derived from coal, oil, natu-
ral gas, uranium or renewable resources, is essential
for improving the living standards of world human
population. Living standards in countries around the
world are directly related to the energy consumption
on per capita basis. Coal reserves in all known coal
basins around the world constitute about 65% of all
proved fossil fuel reserves (Doerell 2001). In USA,
coal reserves are geographically well distributed and
comprise 90 to 95% of the nation’s energy resources
base. USA coal industry produces over 1.1 billion tons
per year. Ninety percent of all coal produced is used
for electricity production. About 45% of all mined coal
is produced by underground mining.
Underground coal mining has survived many eco-
nomic challenges mainly because of nearly 250 per-
cent improvement in productivity and a substantial
improvement in safety and risk management. Long-
wall mining and coal seam degasification are the two
most significant innovations in the past thirty years
that led to higher productivity and improved safety in
mines. Today a little over fifty percent of underground
USA coal mines. CONSOL Energy is the largest pro-
ducer of underground mined coal and operates twelve
longwall panels in Pennsylvania, West Virginia and
Virginia. The trend in longwall mining is for larger
panels, bigger longwall equipment and, higher produc-
tion capability and productivity. Many longwall panels
today are 330 m wide and 3000 to 5000 m long con-
taining 2.5 to 4 million tons of raw coal. Such extended
longwall panels have many advantages as well as many
disadvantages.
The main benefits are:
1. Improved safety and reduced injury rates because
of improved longwall/development coal ratios and
fewer longwall moves.
2. Improved recovery of in-situ coal.
3. Improved productivity and reduced cost per ton.
On the other hand, main disadvantages are:
1. Ventilation and methane control.
2. Respirable dust control.
3. Unknown geologic anomalies, e.g. faults, wash
outs, sand channels, etc.

433

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 1. Gassiness of coal seams.

Depth Gas content

Category (m) (m3 /t)

Mildly gassy Less than 180 Less than 3

Moderately gassy 180–360 3–7
Highly gassy 360–1000 7–20

4. Ground control.
5. Escape from longwall face in case of an emergency,
such as fire.
The Part I of this paper will discuss only the
ventilation and methane control challenges faced by
super-extended longwalls on the tail-end of the faces. Figure 1. A typical ventilation layout for longwall panels.
Bleeder ventilation and longwall gob gas control will
be discussed in a sequel to this paper.

2 GASSINESS OF COAL SEAMS

Since coal seams and methane contained in them are

both derived from the same plant material, it is log-
ical to assume that all coal seams are gassy but they
vary in their degree of gassiness, i.e. gas contained per
ton of coal. Coal seams mined in USA as well as the
rest of the world can be generally classified in three
categories depending on their gas contents and depths
from surface as shown in Table 1.
The depth of coal seams from surface generally
correlates well with their gas contents but a direct
measurement of gas contents is highly recommended.
Moderately gassy and highly gassy coal seams
must be properly degassed prior to mining and dur-
ing mining otherwise, high productivity of coal cannot
be achieved. During the past thirty years, CONSOL
Energy and the erstwhile USA Bureau of Mines devel-
oped several techniques for coal seam degasification,
namely, (a) in-seam horizontal drilling, (b) vertical gob
wells, (c) vertical wells with massive hydraulic frac-
ing and (d) horizontal boreholes drilled from surface
(Thakur 1997a,Thakur 2005). It is feasible to extract at
least fifty percent of in-situ gas prior to mining from
moderately gassy coal seams. For highly gassy coal
seams, it may be necessary to remove 70–80 percent
of in-situ gas prior to mining.
Extracting methane from coal seams prior to mining
and during mining reduces ventilation requirements
considerably and allows high extraction rates leading
to higher productivity and lower cost of mining. The
drained methane can be processed and marketed to
defray the cost of coal seam degasification.
3 VENTILATION LAYOUTS FOR
LONGWALL FACES
Typical ventilation layouts for extended longwall pan-
els are shown in Figures 1–3. The layout shown in
Figure 2. A typical ventilation layout for longwall panels.
Figure 1 is used in mildly gassy and moderately gassy
mines where the ventilation quantities needed at the
tail-end of the longwall face generally do not exceed
800–1200 m3 /minute. With increasing gas contents of
the coal seams, it becomes necessary to bring addi-
tional air to the tail gate end of the longwall face and
the layout shown Figure 2 is often used. This layout
is also necessitated because the tail-gate roads inby
of the longwall face do not stay open for more than
600 m and are unable to conduct large quantities of
ventilation air.
A preferred ventilation layout is shown in Figure 3
where the tail gate entry next to the longwall block
(usually the most stable entry) is dedicated as an “inter-
nal return” and is regulated to ensure adequate air
quantities at the tail gate end of the longwall face.
As shown in Figures 1–3, extended longwall panels
should be ventilated with a combination of a main fan
and a bleeder fan. With proper ventilation planning and
design, it appears possible to deliver adequate ventila-
tion air to the longwall face and the bleeders. Table 2
shows suggested ventilation quantities for different
coal seams (Thakur 2001).
Ventilation air quantities reaching the tail-end of
the longwall face is dependent on the ventilation air

434

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 3. A typical ventilation layout for longwall panels.
Table 2. Ventilation air quantities on longwall faces *.
Tailgate Bleeder
Category (m3 /min) (m3 /min)
Mildly gassy 700 3000–4500
Moderately gassy 1200 4500–7000
Highly gassy 2000 7000–10,000
* It is assumed that coal seam degasification will remove
about 50% of gas contents in moderately gassy mines and
70-80% in highly gassy coal seams prior to mining.
Figure 4. Air leak-Off on longwall faces as a function of
width, m.
quantities (and therefore, the air velocity) at the head
gate end of the longwall face and the air leak-off on
the face. The wider the longwall panel, the higher is
the leak-off factor. Figure 4 shows a plot of some actu-
ally observed data in coal seams with a thickness of
1.5 to 1.8 m.
It is clear from this observation that the wider the
longwall face, the smaller is the ventilation air quan-
tities reaching the tail-end of the face. It is also true,
as discussed later in the paper, that the wider the long-
wall face, the higher is the methane emissions at the
tail-end of the longwall face. It is, therefore, easy to
435
Copyright © 2006 Taylor & Francis Group plc, London, UK
conclude that for a given set of conditions, a limit on
the longwall face width will be reached when it will
not be possible to dilute the methane emissions enough
to meet the statutory requirements.
4 GAS EMISSIONS ON LONGWALL FACES
4.1 Gas emissions
Gas emissions on the longwall face is dependent on (a)
the residual gas content of degassed coal seam or the
degree of degasification, (b) the rate of coal extraction,
and (c) the diffusivity of coal (Thakur 1977). The latter
is a measure of rate of gas emission from mined coal.
Higher rank coal, e.g. low volatile bituminous coals
have a much higher diffusivity than high volatile bitu-
minous coals (Kissell 1972) and as such they release a
higher fraction of their original gas content on mining.
4.2 Rate of methane emissions
An approximate estimate of methane emissions at the
tail-end of a longwall face can be derived from the
following equation.
Where:
Q = total methane emissions at the tail-end of
longwall, m3 /min
Qo = total methane emitted when no mining is done,
m3 /min
V = average rate of mining in t/min
A & B = gas contents of coal prior to mining and
after mining respectively
C(x) = methane lost to gob areas by air leak-off and
is a function of the distance from the head gate, m3 /min
Assuming a 300 m wide longwall face has a daily
advance rate of 15 m, and a ventilation leak-off of 50
to 70 percent, average methane emissions from highly
gassy coal seams at the tail-end of the longwall face
as a function of degree of degasification is shown in
Figure 5. For high coal productivity, highly gassy coal
seams should be degassed prior to mining to reduce
the gas content of coal to at least below 6 m3 /t but
preferably to 3 m3 /t. It is also prudent to assume a peak
emission rate that is 50 percent higher than the average
emission rate. The minimum ventilation air quantity
should be able to dilute the peak methane emissions to
below statutory requirements, i.e. one percent in USA.
4.3 Gas layering on longwall faces
The second criterion for the adequacy of ventilation on
longwall faces is the prevention of gas-layering near
the roof or floor. Gas layering in any mining road-
way including the longwall face is governed by (a)

Figure 5. Methane emissions on longwall faces as a function
of gas content of coal, m3 /t.
the methane emission rate, (b) ventilation air velocity,
and (c) the effective width of the airway. The gas lay-
ering number (GLN) is mathematically expressed as
follows (UK National Coal Board 1979: Leeming &
Yates 2002).
Where:
Q1 is the methane emission rate in l/sec.
V1 is the air velocity in m/sec
D is the effective width of the longwall face, m
A minimum value of 5 for GLN is considered neces-
sary to prevent layering. The higher the value of GLN,
the less likely it is that gas layering will occur.
A typical calculation is shown here.
Assume, Q1 = 250 l/sec
D = 3m
And a GLN of 6 for safety
The necessary air velocity to prevent gas-layering
from Equation 2 is 4.3 m/sec. Assuming a mining
height of 1.8 m, the minimum ventilation air required
at the tail-end of a longwall face is 1393 m3 /min in
order to prevent gas layering.
5 MATHEMATICAL MODELING OF
METHANE FLOW
The flow of methane and air on longwall faces can
be easily modeled mathematically. Fundamental basis
of such models are already developed (Thakur 1974).
Main assumptions made are:
1. Symmetry in the directions (y-z) perpendicular to
the longwall face (x direction).
2. The density effect of a lighter gas like methane is
neglected.
3. When mining is in progress, a steady state situation
is likely to prevail, i.e. time dependence of methane
concentration is discarded.
436
Copyright © 2006 Taylor & Francis Group plc, London, UK
Taking a mass balance over a small element of the
longwall face and applying the above assumptions, the
turbulent dispersion of methane on longwall faces can
be represented by the following mathematical model,
Where:
Ex is the longitudinal co-efficient of turbulent disper-
sion,
u(x) = uo exp(−ax), where u(x) is the air velocity at
any point on the face; uo is the velocity of air at head
gate;
a is a leakage co-efficient that is experimentally
determined.
q(x) is the methane source and includes both the
steady and transient methane emissions,
p(x) is the loss of methane in the gob at any point
on the face due to air-leakage
c is the concentration of methane at a “small ele-
ment” of longwall face, defined as Q/V where Q is the
total methane emission; V is the volume of air at the
same location.
Boundary conditions are:
At x = 0 (i.e. the headgate); dc/dx = 0, i.e. methane
concentration is a constant.
At x = L (i.e. the tailgate) c = 0.01 (or any other
statutory limit).
Equation 3 is a second order, non-homogenous dif-
ferential equation and no analytical solution can be
obtained in a closed form. However, solutions can be
obtained using finite difference or finite element tech-
niques and computers. The second challenge here is to
accurately measure values of Ex; q(x) and p(x). Fur-
ther discussion of mathematical modeling is beyond
the scope of this paper but it is an excellent topic for
future research.
6 SUMMARY AND CONCLUSIONS
1. In highly gassy mines, longwall panels must be
degassed to reduce methane content to at least
6 m3 /t but preferably 3 m3 /t for safety and high
productivity.
2. The optimum width of the longwall should be cal-
culated such that (a) the methane concentration at
the tail gate does not exceed the statutory limits
and (b) the velocity of air at the tail gate is high
enough to prevent gas-layering. Ventilation simu-
lation will yield the attainable air quantities and
methane emission can be estimated from Equation
1.A check on gas layering should be done using
Equation 2. A safety factor of 1.5 should be used to
allow for peak methane emissions.
3. Mathematical modeling of methane and air flows
on longwall faces is highly encouraged for a more
 accurate prediction of optimum widths of longwall
faces in highly gassy coal seams.
REFERENCES
Doerell, P.E. 2001. Confidence in Coal Confirmed. EM
Journal 2001: 43–44.
Kissell, F.N. & Bielica, R.J. 1972. An In-Situ Diffusion
Parameter for the Pittsburgh and Pocahontas No. 3
Coalbeds. USBM Research Report 7668.
Leeming, J.R. & Yates, C.P. 2002. Current British Practice
for Methane Ignition Prevention in Coal Mine Headings.
In Euler DeSouza (ed.), Mine Ventilation: 487–490. Exton
(Pa): A Balkema Publishers.
National Coal Board (UK). 1979. Ventilation in Coal Mines.
A Handbook for Colliery Ventilation Engineers: 90–92.
Thakur, P.C. 1974. Mathematical Modeling of Tunnel Air
Pollution. In H.C. Pattison & E. D’Appolonia (eds.),
437
Copyright © 2006 Taylor & Francis Group plc, London, UK
Proceedings of the Rapid Excavation and Tunneling
Conference, 24–27 June 197: 283–294. San Francisco,
California.
Thakur, P.C. 1997. Methane Drainage from Gassy Mines – A
Global Review. Proceedings of the 6th International Mine
Ventilation Congress: 415–422.
Thakur, P.C. 2005. Advancing Mine Safety and Energy Pro-
duction through Coalbed Methane Production. The 1st
China International Conference Proceedings on Coal
Mine Gas Control and Utilization: 145–153. Beijing: Coal
Industry Publishing House.
Thakur, P.C. & Davis, J G. 1977. How to Plan for Methane
Control in Underground Mines. Mining Engineering:
41–45.
Thakur, P.C. & Zachwieja, J. 2001. Methane Control and Ven-
tilation for 1000-ft Wide Longwall Faces. Proceedings of
the Conference on Longwall USA: 167–179. Pittsburgh,
PA.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Methane and methane control in Chinese coal mines

J.C. Tien
Department of Mining Engineering, University of Missouri-Rolla, Rolla, Missouri, USA

ABSTRACT: Over 95 percent of China’s coal is mined underground at an average mining depth of 410 m and
increasing in depth at 10 m per year. With increasing mining depth, methane emission increases as well. Gov-
ernment statistics indicate that nearly half of China’s coal mines are classified as gassy, with recorded methane
explosions in the hundreds every year, and are responsible for over 80 percent of all coal mining deaths. Methane is
also a clean energy resource, although only a small percentage of the methane is captured in China today. Captur-
ing methane underground can not only improve mine safety, it also has the added benefits of reducing greenhouse
gas and providing clean energy. In recent years, there have been increased efforts to develop China’s coalbed
methane (CBM) resource. This article describes these control measures, and drainage and utilization efforts.

1 INTRODUCTION
China has been mining coal for thousands years and
today, nearly all its coals (95 percent) are mined under-
ground (Tien 2005). The deepest Chinese coal mines
exceed 1,000 m, with an average mining depth around
400 m, and depth has been increasing at a rate of 10 m
per year (Tien 2006). Government data also shows
that over 48 percent of China’s mines are classified
as gassy (with a specific methane content of at least
10 m3 /metric tonne, or mt); increased mining depth
appears to have increased methane emission in many
areas as well. For example, methane emission exceeds
170 m3 /min (6,000 cfm) on longwall faces at Fushun,
northeastern China (Sun et al. 2005). In addition, 58
percent of mines are prone to spontaneous combustion
and 88 percent of mines are considered at risk for dust
explosion (Anon. 2004a).
According to the State Administration of Work
Safety (SAWS), there were 18,071 coal mining acci-
dents which resulted in 30,924 deaths between 2000
and 2004, an average of 1.71 deaths in each accident.
Since 1949, there have been 19 serious coal mining
accidents with more than 100 deaths each. Of these,
18 were due to methane explosions. All seven serious
mining accidents (with at least 30 deaths each) in 2004,
and ten of the eleven major coal mining accidents were
all due to methane explosions. In fact, methane explo-
sions are responsible for 80 percent of the number of
accidents and deaths in all major coal mining accidents
(Anon. 2005h, Li 2005b).
While methane has caused serious safety prob-
lems underground, it is also a clear energy source.
Seemingly continuous coal mining accidents and
increasing public awareness of the environment have
prompted the government to place added emphasis on
methane control through increased safety measures
and drainage to improve mine safety and use it as an
alternative energy source.
A conservative estimate puts China’s CBM or coal
mining methane (CMM, the same as coalbed methane
but emitted as a result of coal mining activities)
resources at approximately 31.46 trillion cubic meters
(Tm3 , or 1012 m3 ) ranking third in the world (Lin
2005), similar to its natural gas reserves of 30 Tm3 (Li
2005b). Currently, China’s primary energy source is
still coal followed by petroleum, with natural gas com-
prising only 2.8 percent. It has long been recognized
that expanded use of natural gas and CBM or CMM
as a substitute for coal could help address many of the
country’s air pollution problems, including greenhouse
gas emissions.
China’s coal mines emit 1.8 billion cubic meters
(Bm3 ) of methane in 2004 (Lin 2005), and only a
small percentage of the methane is captured with the
remaining 1.3 Bm3 vented directly to the atmosphere
every year (Anon. 2005a). With government’s recent
push for increased methane development and utiliza-
tion, it is hopeful this will change over time, even if it
is changing only slowly.
While important policy changes have occurred
and drilling technologies become available over the
past few years, a number of obstacles could con-
strain development of the CBM/CMM market. This
paper addresses China’s coalbed methane resources,
methane drainage and utilization as well as challenges
to methane recovery and usage.

439

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 1. China’s CBM reserves (after Lin 2005).
2 CBM/CMM RESOURCES
Next to Russia and Canada, China ranks number
three in coalbed methane resources. According to
government data, total CBM resources in shallow
coal seams (less than 2,000 m) is estimated to be
between 31.46 × 1012 m3 , or 45 billion tonnes of coal
equivalent. (The conversion factor is: 1 m3 natural
gas = 1.3300 Standard Coal Equivalent.) This is equiv-
alent to at least 20 years’ annual consumption based
on current consumption. As of the end of 2004, total
proven CBM reserve approved by the Chinese gov-
ernment in four blocks is 102.308 Bm3 while total
methane drained methane from coal mines reached
1.8 Bm3 (Lin 2005).
China has eight coalification periods; 98 percent
of its CBM resources occur in Late Permian coal
seams, the bulk of which is contained in its thirteen
coal-bearing basins. The most notable ones are the
Ordos Basin (western Inner Mongolia), Qinshui Basin
(Shanxi Province), Northern China, Tu-Ha Basin (Xin-
jiang Province in northwestern China), Junggar Basin
(eastern Inner Mongolia), and theYunnan and Guizhou
coal-bearing regions (Fig. 1), with each of these
regions containing CBM resources in excess of 1 Tm3
(Tien et al. 2002). According to the survey of 115 tar-
geted major CBM regions, average methane content
is 9.76 m3 /mt at an average concentration of 90.6 per-
cent, or averaging at 115 M m3 /km2 with a gas content
saturation of 41 percent (Anon. 2005e).
Although China’s experience in CBM recovery
dates back over half century ago, its systematical
development did not start until the 1980s. Today, it is
still in its infancy with no commercialized production
base.
3 METHANE EMISSIONS FROM CHINESE
COAL MINES
Major methane emission in China mainly come from
its tens of thousands of production mines, abandoned
440
Copyright © 2006 Taylor & Francis Group plc, London, UK
Table 1. Total methane emissions by year (Huang et al.
2003).
Year 1987 1992 1993 1994 1995 1996 2000
Total emissions 6.45 8.32 8.55 8.95 8.90 9.28 9.63
(Bm3 )
Total coal 0.93 1.11 1.15 1.23 1.29 1.37 1.00
production (Bt)
Table 2. Methane emission for major state-owned mines,
1994 and 2000 (Huang et al. 2003).
1994 2000
Emission, Emission,
Mm3 % Mm3 %
Low emission mines 647.71 14.1 664.80 13.7
(<5 m3 /mt)
High emission mines 1,895.73 41.3 2,099.13 43.3
(>10 m3 /mt)
CH4 outburst mines 2,050.49 44.6 2,050.49 43.0
mines and process of coal treatment, storage and trans-
portation after mining. Ninety five percent of CMM
comes mainly from ventilation air methane (VAM)
in production mines with the balance from methane
drainage systems. With increasing mining depth and
steady increase of coal output, CMM emission has
been in an upward trend in China (Anon. 2005e).
The exact number of China’s coal mines are
unknown. Estimates by the government range from
an earlier 85,000 in the late 1990s to today’s 24,000
after thousands of small mines were shut down; actual
figures are probably higher, in the mid 50,000s (Tien
2005). As 48 percent of the major state-owned mines
are gassy with varying high methane content (20–
30 m3 /mt), methane emissions as a result of mining
and other activities are huge.
It is estimated that total methane from China’s major
state-owned operating coal mines exceeded 4.815 Bm3
in 2000; includes emissions from abandoned mines
and other sources, this figure reached 9.625 Bm3
(Huang et al. 2003), and it has been increasing due
to increased coal production (Table 1). The amount
of methane emitted depends on the gassiness of the
coal mined, from a low of 5 m3 /mt for low emission
mines to the highest reaching 77.45 m3 /mt at Fuxun
(Sun et al. 2005) (Table 2).
3.1 Methane from abandoned mines
Due to an on-going coal sector restructuring and gov-
ernment’s push to improve coal mine safety, records
show that over 60,000 small coal mines that fail to
meet production and price stability have been closed
 since 1997; and nearly one-third of China’s 600 major
state-owned coal mines were approaching their end of
production as of 1999 (Sage & Creedy 2003). Statis-
tics also show that approximately 459 major coal mines
have already been mined out with an estimated 300 Mt
of coal still remaining in pillars and gobs; the major-
ity of these mines are gassy mines. It is estimated
that by 2020, China’s abandoned major coal mines are
expected to reach 541 (Sun & Jing 2002).
A survey of 82 major coal mining regions shows
that 70 percent of the 150 abandoned coal mines
are gassy mines. For example, total CMM potential
for Fengcheng (Jiangxi), Jixi (Heilongjiang), Nantong
(Sichuan), and Yongrong (Chongqing) regions alone
reached 25.2 Bm3 . The same survey shows that only
less than 30 percent of the mines actually utilized
drained methane (Sun & Jing 2002). A different report
show that more than 120 major state-owned coal
mines with depleted reserves having been either closed
and/or bankrupted (Sage & Creedy 2003).
It is estimated that total emissions from China’s coal
mines was 7.7 Bm3 in 1997 and 9.3 Bm3 in 2000.These
were estimated based on 10 m3 /mt for state-owned
major mines and 1 m3 /mt for township- and village-
owned small operations (Sage & Creedy 2003). These
figures may differ slightly from other stats because
of the source of production statistics, but they pro-
vide an order of magnitude of the amount of methane
emitted.
4 A NATION-WIDE COAL MINE SAFETY
CAMPAIGN
Methane problems have been worsening in recent
years and the number of methane-related accidents and
deaths steadily increasing as the coal industry expands
in coal production and mining depth increases. In 2005
alone, there were at least five major coal mining dis-
asters and associated deaths that were due to methane
explosions (Tien 2005):
02/14/05 Sunjiawan Mine Methane explosion 214
(Liaoning)
03/19/05 Xishui Mine Methane explosion 72
(Shanxi)
07/11/05 Fukang Mine Methane explosion 81
(Xinjiang)
11/27/05 Dongfeng Mine Coal dust/methane 172
(Heilongjiang) explosion
12/07/05 Liuguantun Mine Methane explosion 108
(Hebai)
All these and earlier accidents and fatalities have
prompted the government to vow to improve coal
mine safety and launched a nation-wide safety cam-
paign for all state-owned major coal mines between
May and July 2005 (Anon. 2005f). The purpose was
441
Copyright © 2006 Taylor & Francis Group plc, London, UK
to systematically examine all 45 state-owned major
operating coal mining regions (a total of 402 mines)
for hidden safety concerns. In addition, the govern-
ment also started to implement the following seven
measures in April 2005 (Zhao 2005):
1) Continuous Safety Monitoring at All 45 Major
Gassy Coal Mining Regions, with Emphasis on
Methane Control
The 402 mines in the targeted 45 mining regions
produced 503 million tonnes (Mt) or 25.7% of
China’s total coal production in 2004. Over 60%
of these mines are classified as gassy, highly gassy,
or have potentials for methane outburst; up to 65%
of theses mines are susceptible to spontaneous
combustion.
Specifically designated government personnel
will be responsible for the continuing inspection
and monitoring of each mine’s safety training, pro-
duction planning, and safety and production permit
applications. There will also be designated com-
pany safety personnel at the mine to work in concert
with government inspectors.
2) Nation-wide Safety Inspection and Campaign
In February, the government launched a nation-
wide safety campaign, including a large-scale
safety audit program for the aforementioned 402
mines by a 91-member special task force com-
prised of mining professionals from universities,
research and design institutes, and mining compa-
nies. Specific areas examined at each mine were
mine development and mining method, ventila-
tion system, methane drainage system, mining in
adjacent seams, coal and CH4 outburst control
measures, fire-fighting system, dust control sys-
tem, mine monitoring system, mine electrical safety
system, mine water drainage system, mine safety
program.
3) “Coal Production Determined by Ventilation
Capacity” Strictly Enforced
There has been a long-existing but rarely enforced
“12-character Principle” in ventilation and produc-
tion planning in China’s coal industry: “Drainage
precedes mining, methane measuring/monitoring
and control, production be determined by venti-
lation capacity” (“xian-chou hou-cai, jian-ce jian-
long, yi-feng ding-chan”).
With rising coal prices in the last several years,
many mines have dramatically increased coal pro-
duction beyond their initial designed and/or permit-
ted capacity, with many producing 50 to 100 percent
of their normal capacity (Tien 2005). To control this
practice, the government has tightened up mining
permit issuances and stepped up safety awareness
campaigns and field enforcement on mine safety
issues, and more importantly, shut down thousands
of small, often unsafe, coal operations throughout
the country.
 4) More Emphasis Placed on Mines with Potential of
Methane Outburst
A coal mine is classified as an “Outburst Mine”
if there is at least one occurrence of methane,
rock, and/or carbon dioxide outburst, and is sub-
ject to special provisions as outlined in the Coal
Mining Safety Regulations. These include spe-
cific guidelines dealing with methane forecast and
control (section 1), special planning (section 2),
regional preventative measures (section 3), local
control measures (section 4), and safety measures
(section 5).
Forty-four coal mines in China have been offi-
cially listed as mines that have signs of extraor-
dinary methane migrations, the government will
provide special technical and financial assistance
to help combat methane problems.
5) Mine-wide Monitoring System for All Gassy Mines
The plan is to have mine-wide digital monitoring
systems at all major gassy mines regardless of their
size, to monitor all ventilation-related parameters.
Networking will be county-based for small oper-
ations where they are located, while major coal
operations will be networked centrally in a mining
region.
Except for a few cases, most of the moni-
toring systems are developed in China. Leading
coal producing provinces such as Shanxi, Henan,
Heilongjiang, and Shandong have installed mon-
itoring systems with varying levels of sophisti-
cation. For example, Yangquan Coal Group (near
Taiyuan, Shanxi Province) installed its first moni-
toring system over 20 years ago, perhaps one of the
first in China. The company has invested over 35
M yuan ($4.30 M) until now. Today, Yangquan has
a total of eight systems, 225 substations and 889
sensors and probes, and has a total of 468 km of
wiring in its five underground mines.
There are 119 highly gassy coal mines in
Heilongjiang Province. The use of mine-wide mon-
itoring systems started at least five years ago and
today, they are installed at 56 gassy mines at Jixi,
Hegang, Shuangyiashan, and Qitaihe coal groups.
Costing around $2 M, the system includes local net-
works for small operations and a central system
covering all major coal operations in the province.
Substations and/ or individual modular systems are
connected through a General Packet Radio Service
(GPRS) network to the server.
6) Enhanced Research on Methane Migration Mech-
anism, Forecast and Control
All coalfields in China are subject to vary-
ing degrees of crustal stresses, a legacy of the
Himalayan mountain movement 55 M years ago
(Tien & Ying 2005). Records show that 1,377
coal bumps were measured in 17 medium and
large mines, with the highest reaching 4 on the
442
Copyright © 2006 Taylor & Francis Group plc, London, UK
Richter scale. In 2002, a total of 6,127 bumps
were detected in Laohutai Mine, Fushun (Liaoning
Province, northeastern China) alone.All these seem
to indicate there is a strong correlation between
crustal stress field and methane emission (and/or
outburst). More research considering this stress
field is needed.
7) National Methane Engineering Research Institute
Established
Challenges faced by China’s mining industry are
many. In addition to being gassy, over 51.2 per-
cent of China’s coal reserves are also suscepti-
ble to spontaneous combustion. Other challenges
include coal dust explosions, susceptibility to inun-
dation (China’s fourth cause of deaths), and heat
underground.
Despite these challenges, there has been a lack of
concerted effort at the central government level to
systematically study the various problems encoun-
tered by the industry, except for a individual aca-
demics and several research and design institutes
who specialize in particular areas, e.g., ground con-
trol, methane forecast and control, fire fighting and
mine rescue. This would be a welcome move to con-
duct many critically needed research in the area of
mine health and safety.
5 METHANE DRAINAGE IN CHINA’S COAL
MINES
The earliest methane drainage in China dates back
1637 where bamboo was used to extract and channel
methane out of coal seams (Tien 2006). In mod-
ern times, systematic methane drainage started in
the early 1950s at Fushun (Liaoning Province) in
thick coal seams with relatively high permeability (up
to 220 md) (Tien 2006). Methane was successfully
drained through flank fan boreholes from sealed insets
in roadways ahead of mining. By the second half of
that decade, accessing methane-laden coal seams from
bleeder entries, surface and adjacent seams were suc-
cessfully attempted at Yangquan Mining District in
Shanxi Province; this method was subsequently used in
gassy reserves at several other mining bureaus (Li et al.
1998).
With increasing mining depth in China, the amount
of methane emitted changed as well. Among the
600 some major state-owned mines, average abso-
lute methane emission was 15.07 m3 /min and the
average relative emission rate was 17.10 m3 /mt. In
2003, these values for the same group mines sampled
were 19.8 m3 /min and 13.9 m3 /mt, respectively (Anon.
2005e).
Between 1999 and 2003, the number of drainage
systems in major state-owned mines increased from
140 to 203 in four years, while the volume of CBM
 2,000 250 De-pressurized zone Intake

Number of Drainage Systems

1,800
3
Drained Methane, Mm

1,600 200
1,400

Main Return
1,200 150

Main Intake
1,000
800 100 Gob
600
400 50
200
0 0
52 58 64 70 76 81 87 93 00 04

Figure 2. Amount of methane drained and utilized, Drainage pipeline Return

1952–2004 (Hu 2005, Sun et al. 2005).
Figure 3. Drill holes are overlapped to enhance methane
recovery without additional cost (Wang et al. 2005).
3 3
drainage was doubled, from 740 Mm to 1,521 Mm
from 1998 to 2003 (Hu 2005; Fig. 2); of this amount,
approximately 500 Mm3 was recovered and used,
mostly as boiler fuel and town gas. Average CMM
drainage concentration in 2003 in these mines reached
26%. Total methane drained from ten major coal min-
ing regions (e.g., Yangquan, Huainan and Fushun) was
close to 1 Bm3 in 2003 (Anon. 2005e). With increas-
ing mining depth and expanded coal production in
recent years, there has been an increasing need for
methane drainage which has moved China’s CBM
industry ahead.
Drainage recovery rate at major state-owned mines
is only at 25.8 percent, and the national average rate is
even lower at 10 percent (Huang et al. 2003). Average Figure 4. Drill holes are overlapped to enhance methane
methane concentration is also low, at only 32.2 per- recovery (Wang et al. 2005).
cent in 2000, one of the major reasons for China’s
low methane utilization rate at this point; drainage ground airways with the purposes of reducing some,
technologies must be improved. or all, of the methane into the development headings,
Drained methane is still primarily used as house- and preventing outbursts.
hold gas where prices are low, demands fluctuating,
and the needed pipelines for convenient delivery are 6.1 Drainage hole arrangements
lacking, placing methane in a noncompetitive position
Most pre-drainage operations involve horizontal bore-
when compared to coal.
holes drilled in the working seam from underground
airways to reduce some, or all, of the methane into the
6 DRAINAGE TECHNIQUES COMMONLY development headings, coal seams, and from the work-
USED IN CHINA ing coal seams to lower methane concentrations on the
faces and/or to prevent outbursts.The length of the hole
As of the end of 2004, there are 203 permanent surface will be at least 80 m with 3-5 m spacing between holes;
drainage wells producing a total of 1.8 Bm3 , with 85 this will usually be carried out at least six months ahead
percent produced from the top 20 mines, and one-third of mining. To enhance recovery, holes are often slanted
coming from three gassy coal deposits (Yangquan in and/or overlapped (Fig. 3).
Shanxi, Huainan in Anhui and Fushun in Liaoning). Field applications at Jiaozuo and Pingdingshan
Most coal seams in China can be characterized by mining districts (Henan Province) show that a slanted-
their low permeability, high specific methane content, hole arrangement can avoid hole collapsing and
high outburst susceptibility, multiple seams, and often plugging while provide 1.5 times as much methane
with complicated geological conditions. These make reduction when compared to parallel holes; an over-
coal seam depressurizing during drainage essential. lapped arrangement can increase methane reduction
Because of the variety of different mining condi- in airways without incurring additional cost.
tions, all drainage methods have been applied in China Another layout that has been successfully used in
over the years. The following are typical methods Taiji #10 Mine in Beipiao (Liaoning Province) is a
commonly used in Chinese coal mines. grid-type arrangement (Fig. 4). Taiji’s coal is soft, has
Most pre-drainage operations involve horizontal some of the lowest permeability in China, and is the
boreholes drilled in the working seam from under- most difficult coal seam to pre-drain.

443

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 5. To access coal seam from the adjacent strata above
(Li et al. 1998).
This type of arrangement, given adequate pre-drain
time, has yielded up to 30 percent recovery. The
method has become one of the major methods when
draining in soft coal with low permeability (Wang et al.
2005).
6.2 Adjacent seam drainage
Drilling from an adjacent seam above the coal seams
being mined using longholes has also been commonly
used to effectively drain methane from longwall faces
(Fig. 5). This method has been practiced at over 160
mines in more than 20 mining regions such as Huainan
(Anhui Province), Chong-qing, Pingdingshan (Henan
Province), Fuxin and Tiefa (Liaoning Province).
6.3 Methane drainage in thick seams
In thick coal seams, coals are left behind the gob area
and they can be sources of methane, as well as a sink
where methane emitted from both the adjacent seams
and surrounding areas can accumulate.This makes gob
drainage essential in controlling methane. To facilitate
drainage, pipelines are pre-laid in the gob area as min-
ing faces advance, pumps are then used to drain the
collected methane from the gob; or a sort of bleeder
entry is pre-planned to drain methane from behind the
gobs (Fig. 6).This method has been regularly applied at
mines in Fuxun mining region with methane recovery
reaching 80 percent (Wang et al. 2005).
Laohutai Mine in Fuxun has been draining gobs
since 1954; it is estimated that at least 400 Mm3 of
methane has been drained using this method in the
past fifty years (Wang et al. 2005).
In gassy longwall faces with a potential for methane
outburst, a combination of several drainage methods
are used the same time to enhance drainage. For exam-
ple, Panyi Mine at Huainan (Anhui Province) mines
gassy (>10 m3 /min) thick coal seams (4.5 m) with
outburst potential. Four methods in any one panel
are used simultaneously (Fig. 7): cross-measure long-
holes for accessing overlying seams, cross-panel holes,
444
Copyright © 2006 Taylor & Francis Group plc, London, UK
A
A
Gob
Intake
A – A Cross-section Drilling Site
Return Airway
Figure 6. Pre-planned drainage pipelines in the gob area
(Wang et al. 2005).
Figure 7. Multiple drainage methods used in highly gassy
mines that are prone to methane outbursts (Wang et al. 2005).
methane drainage from gob at headgate using drainage
pumps, pre-fractured holes using explosives in the first
300 m enhanced drainage zone to shorten drainage
time, and gob drainage after the panel is mined out.
This method has been quite successful in controlling
methane emission at the gassy longwall faces at mines
in Huainan. For example, Zhangji Mine mined 5.07 Mt
using this method, with methane recovery exceeding
70 percent in 2002. The mine made a net profit of
500 M yuan ($61.4 M) with a payback period of less
than one year (Wang et al. 2005).
Waterjets have also been used to cut relief slots to
enhance methane emission (Fig. 8). For soft to medium
soft coal, a pressure of 8.5 ∼ 16 Mpa has been used.
High pressure can increase the slotting depth thus
enlarging depressurizing zones to enhance methane
flow. To ensure proper penetration and cuttings being
properly flushed, a penetration rate of 10 ∼ 24 m/hr is
usually adequate. For high pressure jets, this rate could
be as much as 60–90 m/hr. Two to three slot holes are
usually placed from one inset.
6.4 Directional drilling
Shanxi Asian American Daning Energy Co., Ltd
(SAADEC), the first Sino-American joint ven-
ture underground operation in China, reported its
successful use of a VLD 1000 Electro Hydraulic
 Enlarged
Boreholes
Jethead
Inset
Waterjet Mine entry
Cutter
High pressure
waterline
Figure 8. Water jet used to enhance drilling (Li et al. 1998).
Figure 9. VLD 1000 Electro Hydraulic Drill (Courtesy of
Orion Energy International, Inc., Beijing).
Drill (Fig. 9) at its Daning No. 1 Mine in Jincheng,
Shanxi Province. Using an AMT DDM MECCA elec-
tronic guidance system, the unit completed 78,484 m
of directional drilling during the first year (Bahr 2005,
Liu & Freer 2004) (Fig. 10).
On November 27, 2004, Orion Energy Interna-
tional, Inc. successfully completed its first surface to
in-seam (SIS) multilateral horizontal drainage (MLD)
well in Qinshui Basin, Shanxi Province. With a diam-
eter of 150 mm, the DNP-02 well has a total in-seam
borehole length of 7,687 m in 13 laterals covering a
total drainage area of 0.5 km2 , or a total length of
8,018 m including vertical length (Fig. 11). The well
has been producing over 18,000 m3 of methane per day
continuously since then. It is reported that a second
well of over 10,000 m has been completed in October
2005.
PetroChina, a major natural gas producer in China,
has signed a contract with the Dallas-based CDX Gas
445
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 10. Directional drilling underground at SAADEC’s
Daning No. 1 Mine in Shanxi Province (Courtesy of Orion
Energy International, Inc., Beijing).
Figure 11. Numerical modeling of DNP-02 well indicates
that a 70 percent of methane recovery is expected in five years
(Courtesy of Orion Energy International, Inc., Beijing).
Company to explore CBM reserves using the same
technology in Shanxi Province. The contract calls for
two horizontal holes in the area, expecting to produce
1.5 Bm3 by the end of 2008 (Anon. 2005g). China
United Coalbed Methane Co., Ltd. (CUCBM) has sim-
ilar plans to drill seven wells using the same technology
in Qinshui Basin; working with CUCBM, BHPB also
plans to exploit MRD drilling in Zhunge’er area and
Far East Energy plans to drill two similar wells in its
Shouyang Project in Shanxi Province (Lin 2005).
7 METHANE UTILIZATION
Methane utilization started at the end of 1970s, and the
State brought the coal mine methane utilization project
in line with the national plan of energy savings basic
construction investment in 1982. For mines which
have methane drainage systems in place, they often
have established a regional pipeline system to supply
methane in adjacent mining communities. In 1994,
methane used by residents accounted for 74 percent
of total consumption, and by 2020, it is expected that
gas utilization rate will reach 85∼90 percent in large
and medium-sized cities covering 350 M populations,
 and utilization rate expected to reach 45 percent in
towns covering 380 M people, and the gas demand will
be 63∼71.3 Bm3 by then. Two typical such cases are
Fushun in Liaoning Province, northeastern China and
Jiaozuo in Henan Province. (Huang et al. 2003)
Fushun, located 30 km from Liaoning’s capital of
Shenyang (a city of 4.54 M population), has a well-
established surface pipeline network and has had years
of experience in drainage operations. At 126 Mm3
annually, Fushun ranks first in China in total quan-
tities; it has a recovery rate of 78 percent. Methane
was drained mainly from its Laohutai underground
mine during the first stage, providing 58.8 Mm3 of
high quality gas annually. The project has proven to be
financially successful.
In the second phase of the project, the high quality
gas from surface drilling was mixed with extracted gas
from underground drainage, making sure the final gas
met contractual specifications. To produce 61.84 Mm3
pure methane annually, total investment for the project
was 155.6 M yuan (RMB) with an ROR of 23 percent
and a payback period of seven years. This has had a
total drainage of 15 Mm3 annually. (Huang et al. 2003)
7.1 China United Coalbed Methane Co. Ltd.
(CUCBM)
One of the major developments in China’s CBM and
CMM developments was the establishment of the
China United Coalbed Methane Co., Ltd (CUCBM)
in 1996 to be the sole state-owned company respon-
sible for CBM exploration, development, production,
pipeline construction and sale in China (Tien 2006);
this is in addition to the many CMM programs by
individual mining companies throughout the coal
mining fields.
As of the end of 2004, total proven CBM reserves
in four blocks owned by the government reached
102.306 Bm3 and total CMM drainage reached
1.8 Bm3 (Lin 2005). Total CBM wells drilled totaled
287 by yearend 2004, of which CUCBM had the
lion’s share (187 wells or 65 percent; 60 wells were
funded internally, with the remaining 127 by foreign
companies); others included PetroChina (29), China
Coalfield Geology Bureau (25), UNDP (20), Fuxin
Municipality (8), and six wells by Sinopec (Lin
2005). There are also several demonstration projects:
(1) Panhe Project (Qinshui, Shanxi Province; Fig. 12) –
total 950 wells in three phases, total project is expected
to reach 700 Mm3 when completed.
(2) Hancheng-Longting Pilot Project – 30 wells in
2005, expected to reach 200 wells in 2006–2007; will
produce 800 Bm3 to 1 Bm3 CBM by 2010.
7.2 Clean Development Mechanism (CDM)
In addition to CUCBM, the China Coal Information
Institute (CCII) has also been actively involved in
446
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 12. Compressed CBM is transported to Changzhi
City and neighboring Henan Province.
promoting CBM/CMM developments in recent years.
Since 1994, CCII has been working closely with
USEPA on the China Coalbed Methane Clearinghouse
program (Franklin et al. 2004, Tien 2006). Other for-
eign entities such as the Asian Development Bank
(ADB) and the United Nation’s Development Pro-
gram (UNDP) have also been playing key roles in
this area as well. The latest developments also include
the Clean Development Mechanism (CDM), a project-
based flexible mechanism under the Kyoto Protocol. It
is designed to make it easier and cheaper for industri-
alized countries to meet the greenhouse gas emission
reduction targets and to assist developing countries
in achieving sustainable development through carbon
credit sales (Zheng & Liu 2004). Its latest examples in
China are two CBM projects in Jincheng and Fuxin.
At the Jincheng Anthracite Mining group, ADB will
provide a $117.4 M loan to help develop a power plant
to harness the methane produced underground at its
gassy Sihe Mine; power generated using internal com-
bustion engines will be used at the mine. In addition,
the project will capture CMM from other mines in
the area for distribution to residential, commercial,
and industrial consumers in the Jincheng area. When
completed in 2008, about 90,000 households and var-
ious industries and commercial establishments will be
served with at least 166 Mm3 of CMM annually, about
400,000 people in the area will benefit directly from
cleaner energy that reduces indoor and outdoor pol-
lution. Total project cost is estimated at $237 M, of
which $20 M will come from Japan and $37.86 M from
the Industrial Bank of China. The Jincheng munici-
pal government and two mining groups will provide
equity capital totaling $61.24 M and the U.S. Trade
and Development Agency has also provided a grant of
$500,000 (Anon. 2004b).
Under its CDM Facility and Clearworld Energy (a
clean energy development company in Beijing), ADB
also backed a smaller but similar coal mining and CBM
 utilization project in northeastern China in November
2005. Under the transaction structure, Fuxin Mining
Group sells the methane produced from its under-
ground mines to a buyer consortium (ICTJ and Nat-
source) under an innovative pricing structure that will
provide improved mine safety, CMM extraction, distri-
bution, and utilization and economic benefits to Fuxin.
The loan amount is $15.8 M (Anon. 2005b).
7.3 Preferential policies
Because of the increasing publicity on mining acci-
dents and public awareness of the environment, the
government has been becoming more interest in reduc-
ing coal burning and promoting the use of clean energy,
and in pursuing methane control in the field. In the
past five years, there have been series of international
conferences in China on mining safety, methane con-
trol and utilization. The government has allocated a 3
B yuan ($368 M) special “safety foundation” plus a
5–15 yuan/mt coal “safety fee” earmarked for mine
safety improvement, more specifically, methane con-
trol (Li 2005a). Several major laws and specific
policies designed to encourage the development of
CBM/CMM have also been issued in recent years. For
example:
Comprehensive Utilization Catalogue of Resources
in 2003 – projects listed under the catalogue could
enjoy relevant preferential policies and reductions of
income taxes.
Current State-encouraged Key Development Indus-
tries, Products and Technologies Catalogue, revised
in 2000 – tariff and import VAT are exempted for
domestic investors.
Foreign Investor’s Industrial Guiding Catalogue,
March 2002 – lists types of resources (coal-associated
resources) in which foreign companies are encouraged
to participate.
While important policy changes have occurred and
drilling technologies have become available over the
past few years, it is believed that a number of obsta-
cles could still constrain the development of the CBM
market.
1. High entry cost; most of the mines in China are
located in remote mountainous areas. Limited by
the size of local population and economic con-
ditions, it is difficult to construct long distance
pipelines delivering drained CMM to large and
medium-sized cities with dense population. This
has been partially addressed by the commencement
of the West-east gas pipeline, but more investment
incentives are needed.
2. Accurate methods for estimating CMM emissions
during mining and transporting, as well as the
amount of CMM remaining in the gob areas.
3. Low-cost drilling and high efficiency equipment
and techniques to effectively perform in highly
447
Copyright © 2006 Taylor & Francis Group plc, London, UK
fractured seams with low permeability. There is
also a need for remote methane drainage and
releasing technology for multiple seams and low-
concentration ventilating air methane (VAM)
recovery and utilization.
4. Lack of consistent and efficient legal framework
to encourage CBM/CMM utilization. Although
this has been improved since the establishment of
CUCBM in 1996, framework improvements are
still needed.
8 SUMMARY
China has abundant CBM/CMM resources. Statistics
show in its 13 coal-bearing regions between 300–
2,000 m in depth, CBM resources exceed 31.46 Tm3 ,
roughly the same as its land-based natural gas
resources of 30 Tm3 (Anon. 2003, Li 2005). Although
the exploitation of CBM in virgin seams is best
approached from surface-bored wells in large scale
commercial development, low permeability (less than
0.01 md) in most China’s coal seams makes surface
production difficult; underground methane drainage
has been the main method in China (Tien 2006).
Despite China’s annual methane drainage capac-
ity of 4.2 Bm3 , only 1.6 Bm3 or 38.1% was drained
in 2004. On the average, only approximately 10%
of methane underground was drained at state-owned
gassy mines in China, much less in the tens of
thousands of county- and village-owned small oper-
ations; and only less than 2% of the drained methane
is utilized, presenting a huge potential for methane
utilization (Anon. 2005c).
CBM development has multiple benefits in China:
(1) It can significantly improve mine safety by reduc-
ing methane emissions underground, which has been
a major cause of China’s mining accidents and related
deaths. (2) It can significantly reduce methane emis-
sions into the atmosphere, a greenhouse gas which is
20–24 times more harmful than carbon dioxide and
contributes 15 percent to global warming. (3) It pro-
vides a source of clean energy, and its large scale
development may ease the tension of China’s energy
supply and improve its energy structure.
With the gradual deregulation of gas price and
increased government promotion of methane as a clean
substitute for coal, commercial development projects
have been carried out in many mining regions. Despite
these encouraging developments, methane utilization
rate is still low; today, natural gas and CBM/CMM
combined are still playing only a minor role in China’s
energy mix.
Although China’s energy structure will still center
on coal in the foreseeable future, and it is expected to
add 1.1 Bt of new coal production capacity by 2020,
China should and will have to enhance its alternative
energy sources by encouraging the use of natural gas
 and coalbed methane. In fact, China has already started
its “clean and green energy” projects and even passed
a Renewable Energy Law on February 28, 2005. The
government has already had plans to reduce its share
of coal use by 3 percent each year from 2003 to 2020
(Anon. 2005d). With continuous progress, hopefully
CBM/CMM will be in a good position to fill this gap.
REFERENCES
Anon. 2003. Improve Coal Mine Safety Monitoring, with
Emphasis on Methane Control. Heilongjiang SCMSA,
SAWS Website, Nov. 10 (Chinese).
Anon. 2004a. Toward A Sustainable Coal Sector in China. In
Joint NDP/World Bank Energy Sector Management Assis-
tance Programme (ESMAP), Report 287/04. The World
Bank, 228 pp.
Anon. 2004b. ADB Project to Harness Coal Mine and
Coalbed Methane for Power Plant and Distribution in
Shanxi, PRC.
Anon. 2005a. ADB Helps China’s Coal Mine Methane
Project. Xinhua NewsAgency. November 15, www.adb.org,
http://www.china.org.cn/english/BAT/148907.htm.
Anon. 2005b. ADB Helps Seal Carbon Credit Sales for Coal
Mine Methane Project in Fuxin, PRC. Nov. 15, 2005;
www.adb.org, 3 pp.
Anon. 2005c. China Should Increase Its CBM Utilization
Program, SAWS, March 18, 1 pp. April 13. (Chinese)
Anon. 2005d. Coal to Remain Top Energy Source for China.
Shenzhen Daily. March 30.
Anon. 2005e. Methane to Markets Partnership, Coal Mine
Methane Subcommittee – Country Profile: People’s
Republic of China. Action Plan of Coal Mine Methane
Subcommittee, Methane to Markets Partnership, U.S.
EPA, 10 pp.
Anon. 2005f. Methane Control to Improve Coal Mine Safety:
Root Reasons for Coal Mining Accidents. SAWS website,
Beijing, February 23 (Chinese).
Anon. 2005g. Methane Gas Development Eyed. China Busi-
ness Weekly, July 12: 3 pp.
Anon. 2005h. Notice on Safety Audit for State-owned Major
Coal Mines. State Administration of Work Safety – State
Administration of Coal Mine Safety (SAWS-SACMS).
Bahr, Angle. 2005. Drilling Success in China. International
Longwall News Newsletter, November 8: 2 pp.
Franklin, P.M., Fernandez, R. & Talkington, C.C. 2004.
Methane to Markets Partnership: Opportunities for Coal
Mine Methane Project Development. In Proc. 4th Inter-
national Symposium on CMB/CMM in China, December
1–2: 164–171. Beijing, China.
Hu, Yuhong. 2005. China’s Coal Mining Safety, Current Sta-
tus and Prospect. In Proc. Development and Utilization
448
Copyright © 2006 Taylor & Francis Group plc, London, UK
of Coal-bed Gas, September 29 (Chinese). Beijing: China
Coal Information Institute.
Huang, S.C., et al. 2003. Optimal Projects for China’s Coal
Mine Methane Mitigation. In Proc. 3rd Int’l Methane &
Nitrous Oxide Mitigation, Beijing, China, November 17–
21: 15–25.
Li, Yizhong. 2005a. Special speech on mine safety, Tele-
conference on Production Safety, Beijing, December 5
(Chinese).
Li, Yizhong. 2005b. Welcome Speech. 1st China Interna-
tional Seminar on Gas Control and Utilization, State
Administration of Work Safety, Beijing, China, October
26–27 (Chinese).
Li, Xuecheng, et al. (eds) 1998. Part III – Chapter 4.2:
Methane Drainage. In Mine Safety, Encyclopedia of Coal
Mining Safety in China: 323-347. Beijing: Coal Industry
Publishing House (Chinese).
Lin, Jianhao. 2005. The Emerging CBM Industry in China.
6th US-China Oil and Gas Industry Forum (OGIF), New
Orleans, LA, June; PowerPoint Presentation.
Liu, Zilong, Freer, Chris. 2004. The Application of VLD
Directional Drilling in China. In Proc. 4th International
Symposium on CBM/CMM in China, Beijing, December
1–2: 224–229.
Sage, P.W., Creedy, D.P. 2003. Reducing the Environmental
Impacts of Abandoned Coal Mines in China. In Proc. 3rd
Int’l Methane & Nitrous Oxide Mitigation Conference,
Beijing, November 17–21: 8 pp.
Sun, X. & Jing, L. 2002. Methane Utilization in China’s
Abandoned Coal Mines. In Proc. 3rd Int’l CBM/CMM
Symposium in China, November 13–14, Beijing: 162–167
(Chinese).
Sun, Xuehui, Lu, Guojing & Yun, Baozheng. 2005. Methane
Drainage Research and Application in Highly Gassy
Mines. In 1st China International Seminar on Gas Con-
trol and Utilization, State Administration of Work Safety,
Beijing, China, October 26–27: 114–122.
Tien, J.C., Keim, K.S., & Miller, M.S. 2002. China’s Coalbed
Methane Reserves, World Coal, March: 49–54.
Tien, J.C. 2005. Health and Safety Challenges for China’s
Mining Industry. Mining Engineering, April, 15–23.
Tien, J.C. &Yin, An. 2005. Geological Basis for Chinese Coal
Mine Safety Problems. China Coal Mine Safety Develop-
ment and Production Symposium, Beijing, October 29;
PowerPoint presentation.
Tien, J.C. 2006. Coalbed Methane Production and Utilization
in China. SME Annual Meeting Preprint, St. Louis, MO,
March 26–29: 11 pp.
Zhao, Tiechui. 2005. Mine Safety and Methane Control
Forum. opening speech, SWAS, April 20, 6 pp (Chinese).
Zheng, Wenbo & Liu, Wenge. 2004. China CBM/CMM
CDM Project Potential. China Coal Information Institute
(CCII), Beijing, China, PowerPoint presentation, 19 pp.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Modern CMM drainage strategies

D.J. Brunner, J.J. Schwoebel & J.S. Brinton

REI Drilling, Inc., Salt Lake City, Utah, U.S.A.

ABSTRACT: Numerous advances have been made in directional drilling equipment over the last five years.This
paper will review specific improvements in this technology and present how this impacts CMM drainage strategies
for underground coal mines. Long in-seam boreholes can be placed in the working seam far in advance of mining
to significantly reduce in-situ gas contents of virgin reserves and commercially recover CMM. Directionally
drilled boreholes can also be drilled into overlying strata to reduce the in-situ gas content of gas bearing strata
in advance of mining (this gas would otherwise contribute to gob gas make), and subsequently recover gob gas
liberated during under-mining (reduce emissions of CMM into mine workings). Precision placement, ultra-long
boreholes, and higher capacity boreholes enable mine operators to more effectively recover the “total resource”
(coal and CMM), reduce CMM emissions during mining, and improve mine safety.

1 INTRODUCTION a result, directionally drilled boreholes can drain larger

REI Drilling, Inc. (“REI”) is a wholly and employee
owned subsidiary of Resource Enterprises, Inc.,
formed in the early eighties and has evolved into
an internationally recognized directional drilling ser-
vices company. REI operates seven drills domestically
that provide directional drilling solutions to the coal
mining industry for CMM drainage, geologic and
old works exploration, water drainage and transfer,
and de-stressing. CMM drainage solutions include the
application of long directionally drilled boreholes that:
(1) immediately reduce CMM emissions into mine
developments; (2) are placed significantly in advance
of mining and reduce gas contents of large virgin
areas, and; (3) serve dual purposes; reduce the gob gas
potential of overlying source seams and subsequently
recover gob gas liberated during undermining. These
solutions are implemented globally through contract
drilling services and technology transfer, including in
the United Kingdom and the Ukraine slated for 2006.
Recent (last 5 years) developments in directional
drilling technology that are considered “modern” in
this paper include higher capacity underground drills,
peak performing water pumps, higher torque down
hole mud motors, rapid penetration drill bits, and pre-
cise steering tools that can be coupled with geophysical
sensors.
These recent developments impact all CMM
drainage strategies that involve long boreholes drilled
either in advance of mining, or for gob gas recovery.
These developments provide for precision placement,
ultra-long depths, and drilling at larger diameters. As
areas in advance of mining, be placed strategically as
in-fill boreholes between vertical CMM wells to sub-
stantially reduce the gas content of large reserves, and
recover higher volumes of gas.
2 MODERN DIRECTIONAL DRILLING
TECHNOLOGY
2.1 High capacity drills
Since 2000, underground directional drills have been
built with increased axial thrust and pull-back, and
rotational torque capacity. For example, a current proj-
ect funded by various US government entities and
managed by Partnership for Energy and Environmen-
tal Reform (“PEER”) will introduce modern under-
ground directional drilling equipment to Ukraine and
initiate work in 2006 at the Krasnolimanskya Mine.
This drill was built by J.H. Fletcher & Co. under
REI oversight and designed with an axial thrust of
50,000 lbs and rotational torque of 2,400 ft-lbs. This is
an increase from prior models designed with a thrust
and pull-back force of 40,000 lbs, and a rotational
torque capacity of 1,700 ft-lbs.
Additionally, these high capacity units are built with
articulation capability, specifically for the application
of overlying gob boreholes.
2.2 Peak performing water pumps
The high capacity drills are equipped with higher
volume reciprocating water pumps; with 33 percent

449

Copyright © 2006 Taylor & Francis Group plc, London, UK

 more flow capacity. This additional water volume
provides the ability to operate down-hole motors at
higher rpm, and the flexibility to use alternative lobe
configurations or larger diameter motors with higher
torque.
For example, excellent penetration rates have
been achieved with down-hole motors operating at
1050 rpm in very hard rock for the application of
horizontal gob boreholes. The additional water can
produce an increase in down-hole torque from approx-
imately 170 ft-lbs to over 350 ft-lbs with the use of
a larger diameter down-hole motor and enable larger
diameter directional drilling.
2.3 High strength down hole mud motors
Modern down-hole motors are built with heavy duty
tungsten carbide bearing packs that provide 33 percent
more wear surface. These motors are equipped with
transmission shafts made of Astralloy™, which is an
alloy with fatigue limits 25 percent higher than 4140
chrome alloy.
These motors provide increased operational life and
the ability to develop larger diameter boreholes at
longer lengths using mud-motor reaming techniques.
2.4 Rapid penetration bits
Modern directional drilling bits used for the devel-
opment of long in-seam boreholes are crown shaped
and fitted with large 13 mm diameter poly diamond
crystalline cutters. In harder coals, for example those
with a Hardgrove Grindability index of less than 40,
penetration rates over 1,000 ft in a single 8 hour shift
have been achieved. This is particularly advantageous
when drilling in-seam boreholes that immediately
reduce CMM emissions into mine workings (in higher
permeability coals).
2.5 Precise steering tools
The modern down hole survey tool provides high
speed data acquisition from very accurate geometric
sensors (tri-axial magnetometer and accelerometers)
positioned behind the mud motor, on a demand basis.
Recently available steering tools developed specifi-
cally for directional drilling in underground coal mines
provide the ability to integrate geophysical sensors
such as focused gamma and magnetometers in with
the geometric sensors. These steering tools provide
real time data to the drill operator through a PC based
system with graphical interpretation.
Data provided by modern steering tools is inte-
grated into AutoCAD® software for further geologic
interpretation with drilling logs, and placement of
boreholes and geologic features on mine maps for the
coal operator.
450
Copyright © 2006 Taylor & Francis Group plc, London, UK
3 IMPACT ON CMM DRAINAGE
STRATEGIES
3.1 Precision placement
Modern directional drilling technology enables pre-
cise placement of long boreholes. This is critical for
the successful implementation of the following CMM
drainage strategies: (1) placement of horizontal gob
boreholes in specific overlying strata to serve dual
purposes; (2) navigation of horizontal in-fill boreholes
around vertical CMM wells; (3) installation of closely
spaced in-seam boreholes to uniformly reduce gas con-
tent in tight coal seams; and (4) implementation of an
effective pattern of in-seam hydraulic fractures.
3.1.1 Borehole accuracy
Placement accuracies of less than 0.5 degrees hori-
zontally and 0.1 degrees vertically can be achieved
with modern directional drilling steering tools. Prior
to drilling, these tools are calibrated to mine grid to
compensate for the magnetic declination in the region
and the dip of the magnetic field. Accurate placement
cannot be achieved without proper calibration.
Figure 1 illustrates a mine intercept of an in-seam
directionally steered borehole.The borehole was found
to be 8.6 ft left of its location at 2,500 ft, as deter-
mined by borehole surveys obtained during drilling.
This corresponds to a placement accuracy of less than
0.2 degrees in the horizontal plane.
3.1.2 Dual purpose boreholes
From underground mining horizons, in-seam bore-
holes can be directionally drilled and steered through
roof or floor strata and precisely placed in overlying
or underlying coal seams for pre-mining drainage of
gas or water as well as to increase gas recovery for
commercial purposes. Boreholes placed in an over-
lying coal seam in advance of mining can serve two
purposes. If the target coal seam is gassy and con-
tributes to gob gas when undermined, in-seam drainage
in advance of undermining will reduce the gas con-
tent of this coal seam and decrease its contribution to
the gob gas make when undermined. Depending on
the elevation of the coal seam relative to the mined
seam, the overlying boreholes can serve a secondary
purpose as gob gas boreholes. Horizontal gob bore-
holes placed in overlying coal seams between 50 and
120 ft above the mined seam can remain intact in the
fracture zone (they can also be lined). When placed
adjacent to low ventilation pressure entries and in ten-
sion zones along gate roads horizontal gob boreholes
will successfully control gob gas when placed under
vacuum. Lastly, this borehole can provide pre-mining
water drainage of overlying strata and minimize water
production after undermining. Figure 2 illustrates the
dual purpose principal; reduce the gas content of the
gob gas contributing source seam before it is mined,
 Figure 1. Mining intercept of an in-seam borehole illustrates high placement accuracy.

Figure 2. Dual purpose overlying borehole reduces gas contents of gob gas contributing source seam in advance of
undermining and subsequently serve as gob boreholes.

and then use the borehole to reduce gob gas emissions
into workings during mining.
As the vertical elevation of contributing seams is not
known precisely, this strategy typically requires iden-
tification of the seam by direct contact, and then direc-
tionally drilling a tangential borehole to precisely place
the borehole so that it can be maintained in the seam.
3.1.3 Horizontal in-fill boreholes
Some gassy US coal fields contain reserves that were
aggressively drilled with hydraulically stimulated
vertical wells. Although these operations have been
successful, several mine operators have determined
that the residual gas content of the coal seam could
be further reduced to improve safety and the effi-
ciency of underground coal extraction. As shown on
Figure 3, long, directionally drilled in-seam boreholes
can be accurately placed in-filling areas between the
vertical wells to produce CMM and further reduce
gas contents. Precision placement is required to suc-
cessfully navigate in-seam boreholes near the ver-
tical wells. Borehole stability problems arise near
the hydraulic fractures, which can also lead to fluid
circulation problems when intercepted. Although

451

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Existing Mains

ft
Borehole #3, 5100 ft
Borehole #1, 2500 ft

500

Borehole #7, 3700 ft

#6, 2
hole

Borehole #10, 3700 ft

Bore

Borehole #9,
ft
#8, 3700
Borehole

4800
800 ft

Boreh

ft
Bore
#2, 4

ole #5
hole
hole

, 4800
#4, 5
Bore

ft
100 ft

Pl anned L ong w all P ane ls

Vertical Well
80 acre spacing

Figure 3. In fill drilling of horizontal boreholes between hydraulically stimulated vertical wells.

fracture orientations are generally known, the exact technology. In-seam boreholes in excess of 5,500 ft
location and extent are not. This strategy requires in depth have been developed from underground coal
precise placement to adhere to safe drilling zones, mines. These ultra-long boreholes provide remote
particularly at borehole lengths in excess of 5,000 ft. access to large areas of virgin reserves and initi-
ate CMM drainage long before actual mining to
3.1.4 In-Seam boreholes in tight coals significantly reduce in-situ gas contents.
In order to reduce gas contents in advance of mining in
very gassy, tight coal seams (overseas in areas where 3.2.1 Reduce gas contents significantly in
anthracite is mined, for example) in-seam drainage advance of mining
strategies demand very closely spaced boreholes. In Longer in-seam boreholes can effectively reduce gas
some conditions, boreholes spaced between 15 ft and contents of large reserves in advance of mining.
50 ft are required, depending on the time available for Because of their length, these boreholes can extend
drainage, and the in-situ gas content. Drilled along the into areas that will not be immediately mined and typ-
longitudinal axis of longwall panels, or across multi- ically can drain gas for longer periods of time (before
ple longwall panels, closely spaced in-seam boreholes intercepted by mining). Drilled orthogonally to future
need to be placed precisely to achieve the required longwalls as shown on Figure 3, long in-seam bore-
lengths and not intersect, and to reduce gas contents holes can span as many as four (4) typical longwall
uniformly. panels.
Accurate placement of in-seam boreholes is also Longer in-seam boreholes produce more gas as the
necessary to maximize the benefit of hydraulic stim- increased surface area of the borehole intercepts more
ulation of in-seam horizontal boreholes in tight, gassy natural fractures and cleats. In gassy, high permeability
coal seams. Redundant fractures and fracturing into coals, ultra-long in-seam boreholes can initially pro-
adjacent boreholes can reduce the effectiveness of this duce as much as 300 MSCFD of methane per foot
stimulation practice. of hole.

3.2.2 Access large virgin reserves

3.2 Longer boreholes
Longer in-seam boreholes can access large virgin
CMM drainage strategies involving longer boreholes tracts to produce gas commercially, and reduce in-situ
are made possible by modern directional drilling gas contents to enable safer mining. CMM drainage

452

Copyright © 2006 Taylor & Francis Group plc, London, UK

 strategies involving drilling long in-seam boreholes
from mine infrastructure developed in advance, such
as ventilation shafts, can at current market rates for
natural gas, provide a revenue stream to more than
offset the advanced development (depending on reser-
voir conditions). Radial in-seam boreholes developed
into coals at multiple levels from a shaft can produce
as much as 2 BCF, depending on reservoir conditions.
3.2.3 Access deeper virgin reserves
CMM drainage strategies involving directional
drilling into dipping coal seams from surface, for
example, to reduce gas contents in advance of min-
ing, and for commercial recovery, benefit from
longer boreholes. Longer boreholes can access deeper
reserves under higher reservoir pressure and at higher
gas contents, increasing gas production and the area
drained.
3.3 High performance horizontal gob boreholes
Modern directional drilling technology improves the
capability of horizontal gob boreholes. Horizontal gob
boreholes are directionally drilled over the mining
seam in advance of the longwall face and placed ver-
tically in what will be the fracture zone when the gob
forms, and typically along the lower pressure side
of the gob in the horizontal plane. Various param-
eters affect the performance of the horizontal gob
boreholes: (1) diameter and length, (2) vertical and
horizontal placement relative to the working seam and
mine ventilation system, (3) borehole integrity follow-
ing undermining, and (4) borehole production issues
related to vacuum and dewatering.
Horizontal gob boreholes have been applied at sev-
eral mines in the U.S., and overseas in China, Japan,
and Europe, with varied success. Typically a 3 inch to
4 inch diameter borehole, 1,000 ft to 2,000 ft in length,
placed between 75 ft and 100 ft above the mining seam,
produces between 200 and 400 MCFD of gob gas
under vacuum.
Experience in the U.S. suggests that horizontal gob
boreholes are not always as effective as vertical gob
wells, but in some cases vertical gob wells do not work
due to tight packing of gob, or are cost prohibitive. Hor-
izontal gob boreholes, should be given consideration in
areas of deep cover, restrictive surface areas and mul-
tiple seam mining. In these particular cases, horizontal
gob boreholes are likely to be safer, more cost effective
and efficient than traditional systems. In many cases,
horizontal gob boreholes significantly improve CMM
drainage efficiencies at mining operations overseas
that employ cross-measure boreholes.
3.3.1 Higher capacity boreholes
Modern directional drilling technology allows for
development of longer horizontal gob boreholes at
453
Copyright © 2006 Taylor & Francis Group plc, London, UK
larger diameters. Both of these parameters affect the
resistance of the borehole to gob gas flow. Using
actual gob gas composition, vacuum pressure and
flow data from horizontal gob boreholes employed at
the Willow Creek Mine in Utah, correlation exercises
determined that the Mueller equation for gas flow best
represented measured conditions. This relationship
was used to derive flow capacities of varying bore-
hole diameter configurations achievable with modern
directional drilling technology. Figure 4 illustrates the
incremental increase in flow capacity at a vacuum of 6
inches Hg for a 3,280 ft horizontal gob borehole drilled
at 3.8 inches in diameter (standard), relative to the
flow capacity of a borehole partially and completely
reamed to 5.75 inches in diameter using modern
directional drilling techniques.The corresponding pre-
dicted gob gas flows increase from 300 MSCFD to
950 MSCFD.
3.3.2 Maintaining integrity when undermined
Horizontal gob boreholes are typically directionally
drilled to rapidly gain elevation above the working
seam and steered to be placed between 75 to 125 ft
above the top of and oriented parallel to the mined
seam. The exact orientation is dependent on the loca-
tion of gas contributing strata and geo-mechanical
characteristics of the gob. If placed too low and
near the rubble zone, the horizontal gob boreholes
may not maintain their integrity when undermined,
and/or may draw significant volumes of ventilation
air above the face, diluting recovered gas concen-
trations. Horizontal gob boreholes placed too high
intercept fractures that are less conductive and may
not be as effective as boreholes placed at optimal
horizons.
Modern directional drilling technology enables
the development of larger diameter horizontal gob
boreholes that can be lined with perforated steel so
that integrity is maintained irrespective of vertical
placement. A smaller diameter pilot borehole is direc-
tionally drilled and then reamed using conventional
or mud-motor reaming techniques, depending on
depth. Lining ensures that all holes will remain intact
and will produce gob gas even if placed too low in
the gob.
3.3.3 Operating horizontal gob boreholes
Because of their placement in the gob, horizontal gob
borehole performance depends on longwall face activ-
ity and wellhead vacuum. Ideally, operators should
control vacuum based on face conditions. During long-
wall mining, increasing the wellhead vacuum pressure
will result in increased gas production. Unfortunately
high vacuum pressures during face idle periods will
draw excessive ventilation air and reduce recovered
gob gas concentrations.
 Gas Flow Rate (70 Percent Methane in Air) for 3,280 ft Horizontal
Gob Borehole Configurations with Wellhead Vacuum of 6 inches Hg
1200

1000
Gob Gas Flow Rate (MSCFD)

800

600

400

200

0
3,280ft x 3.782in 1,640ft x 5.75in, 2,186ft x 5.75in, 3,280ft x 5.75in
1,640ft x 3.782in 1,093ft x 3.782in

Figure 4. Gob gas flow capacity as a function of borehole diameter and length for a given wellhead vacuum pressure.

4 CONCLUSION boreholes target larger reserves and allow commercial

Developments in directional drilling technology over
the last 5 years improve on CMM drainage strate-
gies that involve long steered boreholes. Stronger,
more powerful equipment, coupled with real-time pre-
cision down-hole navigation allow for longer, more
accurately placed boreholes at larger diameters. Accu-
rate placement provides the ability to safely install
in-fill boreholes among hydraulically stimulated ver-
tical wells, install dual purpose boreholes in overlying
or underlying seams, and implement enhanced CMM
drainage techniques in tight coals. Longer in-seam
recovery of CMM while reducing gas contents for safe
and more productive mining in the future.
There is growing interest in larger diameter horizon-
tal gob boreholes as a viable alternative in certain areas
to vertical gob wells and conventional cross-measure
boreholes. Various case studies allowed a better under-
standing of the technique and have demonstrated
significant gob gas productivity. Advancements in
directional drilling equipment allow larger diameters
to be drilled for installation of slotted and conven-
tional steel liners so that borehole integrity and gas
production is maintained after undermining.

454

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Cross-linked polymer gel seals horizontal degas boreholes

greater than 4,000 feet long

S.J. Kravits & G.M. DuBois

Target Drilling Inc.

J.M. Reilly
Pennsylvania Services Corporation

J. Kirley
Concrete Construction Materials

ABSTRACT: Horizontal boreholes drilled from within the mine or from the surface have proven to be effective
in recovering coal bed methane for degasification and commercialization. However, the inability to completely
plug horizontal boreholes still producing methane gas prior to mine-through has caused potentially hazardous
situations and significant coal production delays. To date, cement slurry has been used predominantly to plug
horizontal degasification boreholes, usually with unsatisfactory results. The purpose of this paper is to describe
the successful sealing of 16 horizontal long boreholes and their abandoned sidetracks totaling 25,960 m (85,148 )
by pumping 349,320 l (92,290 gal) of metal cross-linked polymer gel from the surface to underground horizontal
wellheads through slick-line suspended in cased vertical boreholes installed to the horizontal wellheads. Since
almost two times the volume of gel was pumped compared to the actual volume of boreholes including sidetracks,
the gel effectively squeezed into the fracture system of the coal displacing gas and water. Lastly, with an affinity
to attach itself to everything, except for itself, the gel adhered to the inner wall of the borehole providing an
impenetrable skin preventing gas and water to migrate back into the borehole as evidenced by observations noted
when mining into the boreholes.

1 INTRODUCTION proven ineffective filling the entire void of the bore-

The Mine Safety and Health Administration (MSHA)
has determined that Coal Bed Methane (CBM) wells
are subject to the same mine operator ventilation plan
and mapping requirements that apply to methane gas
wells McKinney (2005). A mine operator that uti-
lizes degas boreholes has to describe what precautions
they will take prior to mine-through, usually indicat-
ing water infusing or plugging the boreholes in their
MSHA approved ventilation plan. More recently, this
includes specifying how Surface-To-In-Seam CBM
or Coal Mine Methane (CMM) wells will be plugged
according to the mine operator’s MSHA approved
Ventilation Plan.
If a mine operator is not permitted to water infuse
a horizontal degas borehole prior to mine-through
by their MSHA approved ventilation plan, cement
slurry has been used predominantly to plug the
horizontal degasification borehole. However, pump-
ing cement slurry in horizontal boreholes has often
hole and the borehole sidetracks. The ineffectiveness
of cement slurry results from a) dilution of the cement
in the slurry caused by mixing with water in the bore-
hole, b) heat and friction generated during the mixing
and pumping of the cement slurry causing pressure
build-up in the borehole before the entire borehole
and sidetracks have been filled with cement slurry,
and c) inherent shrinkage of the cement when it cures
or sets Aul and Cervik (1979). Consequently, plug-
ging degasification boreholes with cement slurry that
have not had adequate degasification time prior to
mine-through, have resulted in ignitions at the face
and coal production delays on development, and obvi-
ously much more devastating, when intercepted by
the longwall face. Consequently, when degasification
boreholes, that have not had adequate degasification
time prior to mine-through, have been plugged with
cement slurry, the result has sometimes been ignitions
at the face and coal production delays on development
sections, and obviously much more devastating, when

455

Copyright © 2006 Taylor & Francis Group plc, London, UK

 intercepted by the longwall face. Given the overall
safety and productivity contributions of degasifica-
tion boreholes to mining, a more full-proof plugging
material or methodology was required.
2 GEL BACKGROUND
Target Drilling Inc. (TDI) has directionally drilled over
100 in-mine horizontal degas boreholes greater than
1,220 m (4,000 ft) and several Surface Directional
Horizontal CBM Wells with horizontal laterals in the
coalbed for Foundation Coal’s (Foundation) Coal Gas
Recovery, LP (CGR) at Foundation’s Cumberland Coal
Resources, LP and Emerald Coal Resources, LP coal
mines. As an alternative to either water infusing or
plugging these in-mine or surface directional horizon-
tal boreholes with cement slurry prior to mine-through,
CGR and TDI solicited the technical support and expe-
rience of Concrete Construction Materials (CCM) to
develop an alternative to sealing or plugging these
boreholes with cement slurry by using a gel. Gener-
ally, the gel would be designed to be highly viscous,
similar to water, so that it could be pumped very long
distances, and at a specifically designed timeframe,
cure to a semi-solid state that could be maintained for
greater than 6 months. The gel was designed as a rela-
tively thin liquid with viscosity of 1000 cp at shear so
the gel could be pumped distances greater than 3 km
(10,000 ). As the borehole fills with gel and becomes
pressurized, the gel would squeeze into and infiltrate
the natural fracture system of the coal forcing gas and
water into the natural fractures of the coal. Finally, as
the water soluble gel mix, cures or “gels” into a semi-
solid, insoluble gel at the designed “gel time,” it would
form a skin or filter cake on the inside wall of the bore-
hole preventing migration of gas and water back into
the borehole (Figure 1).
3 CHEMICAL DESCRIPTION OF THE GEL
CCM developed a man made, proprietary inorganic,
metal cross-linked polymer gel consisting of: 1) 97–
99% water by weight, 2) 2–3% by weight liq-
uid high molecular weight water-soluble, partially
hydrated, poly acrylimide polymer product (PHPA)
called VMA-007; 3) less than 1/2% by weight valiant
chromium ion XLR-C, used to “cross-link” or complex
the poly acrylimide material (VMA 007) to change the
polymer from water soluble to insoluble and; 4) less
than 1/4% by weight liquid accelerator and conditioner
to control reaction rate (Activator-M). In simple terms,
the poly acrylimide, in excess of 2 million molecular
lengths, is allowed to hydrate with water creating a
mixture similar to a bowl of noodles. Then at a spe-
cific time, based on the mix design, chromium III ions
456
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 1. Polymer gel forces gas and water into fracture
system.
attach at bonding points in between the noodles cre-
ating extremely over cooked noodles. CCM hired an
independent laboratory to develop specific concentra-
tions of the individual components for the gel mix
based on the desired times required for mixing, pump-
ing, setting and maintaining its semi-solid state by
conducting tests in their laboratory.
Table 1 is an Information Document for the metal
cross-linked polymer gel mix structured as an MSDS
sheet, but it is NOT an actual MSDS sheet. Slight vari-
ations in mixing in the field make it difficult to develop
a representative MSDS. Before the gel was used ini-
tially to seal degas boreholes, meetings were held
between representatives from the local work force’s
safety department, mine management, MSHA District
2; an MSHA Toxologist, CCM, and TDI. The meet-
ings resulted in concluding that the gel ingredients and
mixed gel were not toxic or hazardous providing ade-
quate training was conducted for safe handling, mixing
and pumping of the gel ingredients and that the final
gel mix – either as a viscous liquid or semi-solid –
would not be hazardous or jeopardize the safety and
well being of the local workforce if they came in con-
tact with the gel. It is important to note, this product
combination of gel mix was developed specifically for
this application and to overcome the deficiencies of
using a modified guar. The material is resistant to bac-
terial attack and stable to decay for long periods of
time. Other natural gel systems (guar and low cost oil-
field viscosity agents) consist of food type materials
which present bacteria issues and long-term stability
problems.
4 LABORATORY AND SURFACE TRIAL
GEL TESTS
Laboratory tests and trial field tests were conducted
prior to sealing or gelling the horizontal boreholes. The
gel mix can be modified to fit different combinations
of gel time, temperature and required pumping and
setting viscosity. The purpose of the laboratory tests
 Table 1. Metal cross-linked poly acrylimide gel information safety data sheet (not an MSDS).

Section Number Metal Cross-Linked Poly Acrylimide Gel

1. Product Identification Proprietary Blend

2. Ingredients Non-Hazardous
3. Hazard Identification Light green to powder blue gel with no distinct odor
4. First Aid Measures Eye Contact-water flush; Skin Contact-soap & water; Ingestion-physician; Inhalation-
fresh air; artificial respiration
5. National Fire Prevention Code Health = 0; Flammability = 0; Reactivity = 0; Special Hazard = None
6. Accidental Release Measures Wear personal protective equipment and remove with absorbent materials
7. Handling and Storage Keep away from heat and incompatible materials
8. Personal Protection Chemical resistant personal protective equipment
9. Physical and chemical pH:6–9; SG:1.00–1.05 g/ml; Insoluble to water; Boiling Point: >100 deg. C; Freezing
properties Point: 0 deg. C; light gray to blue green gel with no odor
10. Stability and Reactivity Hazardous polymerization will not occur
11. Toxological Information No information available
12. Ecological Information No information available
13. Disposal Not an RCRA hazardous material. Discard according to regulatory agencies
14. Transportation Primary Hazard Class/Division: Not restricted
15. Regulatory Information OSHA Hazard Communication Status: Non-Hazardous
16. Other Communication HMIS Ratings: Health = 1; Flammability = 0; Reactivity = 0

Figure 2. Elevation sideview of horizontal borehole with sidetracks.

was to simulate various formulations representative Table 2. Dosage of gel ingredients per batch.
of the desired gel mix for sealing the horizontal bore-
holes. The primary criteria of the successful laboratory Volume By Weight Volume By Liquid
Gel
tests, using water that first was tested for pH, were
Ingredient Kilograms Pounds Liters Gallons
to develop: 1) a gel liquid viscosity of approximately
1000 cp for pumping several km; 2) a gel setting time
Water 3.846 8.480 3.785 1.000
of 5 to 8 hours including mixing and pumping time; VMA-007 (tote) 78 173 76 20
and 3) a viscosity of about 20 cp for the cured gel. A XLR-C (blue 155 41 15 4
Bradford programmable viscometer was used to plot pail)
the induction, build, gel and cure cycles for the gel Activator M 9 21 7 2
tests. Subsequently, the gel formulations developed in (round white
the laboratory were used to gel two sections of high pail)
density polyethylene pipe (HDPE) 61 m (200 ) long
and 76 mm (3 ) diameter laid on the ground connected
to 76 mm (3 ft) wyes and shorter sections of 76 mm that a set or cure time of 7 to 8 hours would be used to
(3 ) HDPE pointing up to simulate abandoned side- seal the horizontal boreholes.
tracks that intercepted the roof rock (Figure 2), capped
off with closed 76 mm valves. Holes, 1.6 mm (1/16 )
diameter, were drilled periodically in the wall of the 5 GEL DESIGN FOR THE BOREHOLES
HDPE pipe to simulate the coal’s face and butt cleat
fracture system. The results sealing the two 61 m (200 ) Table 2 provides the final dosage of gel ingredi-
sections of 76 mm (3 ) HDPE pipe with gel indicated ents designed for the horizontal boreholes based on

457

Copyright © 2006 Taylor & Francis Group plc, London, UK

 the laboratory and HDPE trial gel tests with mix-
ing gel batches with 3,785 l (1,000 gal) of water. Due
to potential problems underground out of our con-
trol such as a mine fan going down, it was decided
to mix 3,785 l (1,000 gal) batches at a time to mini-
mize potential losses if pumping had to be terminated
prematurely. It was arbitrarily established to apply a
volume Factor of Safety of 2 for the volume of gel mix
to 1) fill the boreholes, including sidetracks, 2) force
gel into the fracture system of the coal displacing gas
and water, and 3) develop a maximum gel pressure of
690 kPa (100 psi) at the horizontal wellhead to form a
skin or filter cake on the inner borehole wall, which
would prevent gas and water to flow back into the bore-
hole. A very positive feature of the metal cross-link
poly acrylimide gel mix was that the gel mix displays
an affinity to attach or adhere to everything, except
for adhering to itself. The gel pressures at the bore-
holes were not allowed to exceed 690 kPa (100 psi), the
water tested capability of the wellhead control collar
pipes. Trimming line or conduit would not be installed
in the horizontal boreholes to transport the gel mix
to the end of the horizontal boreholes because of the
high viscosity of the liquid gel. It was very important
that a sample of the water to be used to mix with the
gel ingredients was tested for pH so that CCM could
make final adjustments to the concentration of the gel
ingredients if necessary.
additive was supplied in 18.9l (5 gal) blue pails or
208 l (55 gal) drums. The Activator M conditioner/
accelerator is supplied partially hydrated in 11.35 l
(3 gal) white pails (Figure 4). As one batch of gel mix
is being pumped, the next batch is being mixed. The
preparation and mixing of a batch is timed to coin-
cide with the batch being pumped so that a continuous
pumping operation is achieved to keep the mixed liq-
uid gel flowing into the borehole (Figure 5). Viscosity
measurements were taken and recorded for samples of
each batch of the mixed gel before pumping into the
slick-line (Figure 6). Precautions were taken to handle
the gel components and the mixed gel which included:
1) chemical impermeable gloves and safety glasses; 2)
hands and arms were washed with soap and water after
handling; 3) if the gel or gel ingredients were spilled
onto clothing, the clothing was washed with soap and
water; and 4) it was prohibited to taste the mixed gel
or the individual gel ingredients.

6 GEL MIXING AND PUMPING

PROCEDURES

After Target’s staff reviewed MSDS for the individ-

ual gel ingredients and the Information Safety Data
Sheet for the gel, they were required to wear approved
eye, hand and foot chemical personal protective equip-
ment. Prior to mobilizing to each vertical vent hole Figure 3. Slick-line suspended down vent hole with pager
surface gel site, the 31.75 mm (1.25 ) plastic slick-line phone wire and devoted pager phone at vent hole.
was suspended down the vertical vent hole attached to
steel cable and telephone cable. A pager telephone on
the surface was wired directly (not through the mine
wide pager phone system) to pager phones at the bot-
tom of the vertical vent hole and at the horizontal
wellheads for continuous communication during the
gelling process (Figure 3). The slick-line was then
installed underground to the horizontal borehole well-
heads where it was hooked up with a bypass wye,
two valves and a pressure gage. Water was then water
pumped from the surface through the slick-line to the
horizontal wellheads to verify that the slick-line did
not leak. A sample of the mix water was to be used to
mix with the gel ingredients was tested to determine
for pH and the presence of any other chemicals that
might require slight adjustment to the concentration
of the gel ingredients.
The poly acrylimide polymer VMA-007 was sup- Figure 4. A tote of VMA-007, blue pails of XLR-C and
plied in totes and the chromium III metal cross-linker white pails of Activator M at the vent hole gel site.

458

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 5. Simultaneously pumping a batch of mixed gel
(right) while circulating and mixing gel ingredients (left).
Figure 6. Taking a viscosity measurement of the mixed gel
prior to pumping the mixed gel down the slick-line.
Figure 7. B1 longwall panel boreholes sealed with metal cross-linked poly acrylimide gel.
459
Copyright © 2006 Taylor & Francis Group plc, London, UK
7 RESULTS OF SEALING B1, B2 AND B3
LONGWALL PANEL BOREHOLES
WITH GEL
Sixteen (16) in-mine horizontal degasification bore-
holes drilled within 3 longwall panels were sealed
using metal cross-linked poly acrylimide gel (Fig-
ures 7, 10 and 11). All of the boreholes were sealed by
mixing and pumping the gel from the surface through
plastic slick-line suspended down vertical vent holes
and then traversed to the horizontal borehole well-
heads following the mixing and pumping procedures
described earlier. It was believed trimming pipe or con-
duit installed in the boreholes to convey the gel to the
end of the boreholes would not be required because
of the designed high viscosity of the liquid gel. It was
pumped directly from the start of the 101.6 mm (4 )
inner diameter wellhead plumbing the slick-line, gage,
valves and bypass wye to the wellhead. Table 3 sum-
marizes the results of sealing 16 in-mine horizontal
coal degasification boreholes.
7.1 Results of sealing B1, B2 and B3 longwall
panel boreholes with gel
The first two boreholes sealed with the metal cross-
linked poly acrylimide gel were B1-5 and B1-6
(Figure 7). The longwall face mined into the end of
both boreholes to find the ends filled with the gel
including the sidetracks resulting in zero methane
caused longwall production delays (Figure 8). Cure
time for the gel was designed for 7 hours.
The actual volume of gel pumped into B1-5 and B1-
6 of 22,712 l (6,000 gal) was 1.67 times the volume
of the two boreholes, including sidetracks. Pump pres-
sure gradually increased as the boreholes’ filled with
gel as expected. Pumping was stopped when the hor-
izontal wellhead pressure reached 483 kPa (70 psi).
 Table 3. Results of sealing 16 horizontal boreholes with gel mixed and pumped from the surface.

Borehole Borehole length Gel volume Hor. wellhead Distance gel pumped
length with sidetracks estimated F.S. 2 Actual pumped pressure To end of borehole

# Meters Feet Meters Feet Liters Gallons Liters Gallons kPa Psi Meters Feet

B1-6 741 2430 1134 3720 16214 4278 483 70 1240 4067
B1-5 637 2090 783 2567 11174 2952 22712 6000 483 70 1362 4467
B1-4 1253 4110 1744 5720 24898 6578 18925 5000 552 80 2337 7666
B1-3 1232 4040 1832 6010 26160 6912 22710 6000 1035 150 2309 7575
B2-6 1186 3890 1752 5745 25007 6607 24603 6500 414 60 3174 10412
B2-5 1049 3440 1601 5251 22856 6039 22710 6000 359 52 3063 10046
B2-4 1250 4100 1933 6340 27596 7291 20818 5500 276 40 2009 6591
B2-3 1244 4080 2125 6970 30339 8016 26495 7000 345 50 2013 6601
B2-2 1241 4070 1913 6274 27309 7215 23467 6200 483 70 1868 6127
B2-1 1220 4000 1951 6400 27858 7360 15897 4200 400 58 1823 5978
B3-6 1040 3410 1259 4131 17981 4751 20818 5500 414 60 2936 9631
B3-5 991 3250 1140 3740 16279 4301 18925 5000 518 75 2897 9501
B3-4 1265 4150 1725 5658 24628 6507 37850 10,000 290 42 1941 6366
B3-3 1204 3950 1779 5836 25403 6711 30280 8000 242 35 1890 6199
B3-2 1000 3280 1664 5457 23753 6276 18925 5000 483 70 2256 7400
B3-1 1293 4240 1625 5329 23196 6128 24186 6390 690 100 2549 8360
Tls. 17845 58530 25960 85148 370628 97920 349320 92290 N/A N/A 35667 116987

Figure 8. B1-5 mined into by longwall face without Figure 9. Cross-linked polymer gel forms skin on borehole
incident. wall.

When the longwall intercepted the B1-5 borehole it
was full of cured gel. Minimum sloughing of the gel
occurred after mine-through. Where the gel sloughed
from the top of the inner borehole wall, it was noted
that the gel had formed a thin layer or skin sticking to
the inner wall. There was not any gas and water pro-
duction from the gelled B1-5 borehole when it was
initially mined into by the longwall face. Sealing the
B1-5 and B1-6 boreholes was considered a major suc-
cess because the boreholes did not cause any longwall
production delays. Consequently, it was decided to seal
B1-3 and 4 boreholes with the gel several months later.
B1-3 and B1-4 boreholes were sealed with the
same gel formula used for B1-5 and B1-6 boreholes.
However, while pumping the first batch of gel in the
B1-3 borehole it began to thicken or set much faster
than it was anticipated, increasing pumping pressure
quickly to 552 kPa (80 psi). It was quickly determined
that more than the prescribed amount of Activator M
was used on the first batch.Additional water was added
to thin the second batch that was already mixed and
waiting to be pumped because the first batch was tak-
ing longer to pump than it should. A total of 5 batches
were pumped in B1-3 or 1.7 times the volume of
the B1-3 borehole including sidetracks when pump-
ing was stopped at a wellhead pressure of 1,035 kPa
(150 psi). The last 3 batches of gel were mixed with
the exact amount of Activator M, as it was determined

460

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 10. B2 longwall panel boreholes.

Figure 11. B3 longwall panel boreholes.

that the premixed containers of Activator M had too 7.2 Results of sealing the B2 boreholes
much Activator M in them. B1-4 was then sealed with
The B2-5 and B2-6 longwall panel boreholes were
18,925 l (5,000 gal) of gel or 1.5 times the volume of
sealed with gel pumping a total distance from the sur-
the B1-4 borehole obtaining a wellhead pressure of
face to the end of the boreholes over 3 km (10,000 ).
552 kPa (80 psi).
The actual gel volume pumped for these two boreholes
The B1-3 borehole was full of gel when the longwall
was 1.99 times the volume of the boreholes including
first intercepted it. It remained full of gel for 199 m
sidetracks. Overlap of boreholes B2-3, B2-4 and B2-1
(649 ) of B1-3. However, B1-3 was then found to be
probably had an impact on the gel volumes actually
void of gel for 485 m (1,591 ). Fortunately, the bore-
pumped totaling 1.66 times the calculated borehole
hole was producing little gas upon interception and
volume, including sidetracks, in addition to leakage
was easily diluted with existing face ventilation. To
through the coal ribs near the wellheads at an average
eliminate any potential hazard, the longwall crew foam
gel wellhead pressure of 366 kPa (53 psi). Likewise,
packed the B1-3 borehole every 6.1 m (20 ). At a bore-
only 1.14 times the calculated borehole volume with
hole depth of 549 m (1,800 ), B1-3 was found to be full
sidetracks for B2-1 of gel was pumped even through
of gel. B1-3 and B1-5 were drilled to overlap as illus-
experiencing gel leakage at the rib near the wellhead
trated in Figure 7 to provide continuous coverage from
with a wellhead gel pressure of 400 kPa (58 psi). The
methane emissions on development. It is believed that
B2 longwall panel boreholes were either found to be
gel pumped into B1-5 and B1-6 migrated into the bot-
full of gel or if not full, the gel had formed a skin on
tom end of B1-3. It is also conceivable that the B1-3
the inner wall of the borehole. Consequently, the B2
gel did not fill the B1-3 borehole because the borehole
longwall panel production was not interrupted due to
caved and/or the first batch of gel with too much Acti-
gas delays caused by the B2 longwall boreholes.
vator M cured before it flowed to the bottom of the
B1-3 borehole. B1-4 borehole was either completely
filled with gel when it was mined into by the longwall
7.3 Results of sealing the B3 boreholes including
face or if not entirely filled, a skin of gel was found
mine through
to be adhered to the inner wall of the borehole pre-
venting gas and water migration back into B1-4. It is More than two times the calculated volume of gel
also interesting to note that at times, slugs of gel were required was pumped into all of the B3 longwall panel
being forced onto the longwall face as breaks in the boreholes, except for B3-2. A different form of XLR-C
gob directly behind the shields were occurring, pos- cross-linker was used on B3-4 which thinned the mix
sibly indicating instantaneous increases in abutment resulting in 3 times the volume of gel pumped into
loading out by the longwall face. B3-4. CCM did not charge CGR for the additional

461

Copyright © 2006 Taylor & Francis Group plc, London, UK

 volume of gel pumped in B3-4 and resumed using
the original form of liquid XLR-C for the remain-
ing B3 boreholes. The B3-5 and B3-6 boreholes were
mined into when with the continuous miner developed
the longwall setup entries. The boreholes were found
to be full of gel. The B3 boreholes have not been mined
into with the longwall face at the time this paper was
written.
8 CONCLUSIONS
Sixteen in-mine horizontal boreholes and their aban-
doned sidetracks totaling 25,960 m (85,148 ) were
sealed by pumping 349,320 (92,290 gal) of metal
cross-linked polymer gel from the surface to under-
ground, through slick-line suspended in cased, vertical
vent holes installed to the horizontal wellheads and
then plumbed to the wellheads. Due to the high viscos-
ity of the liquid gel pumped, conduit was not required
to be installed in the boreholes. Pumping almost two
times the volume of gel pumped than the calculated
volume of boreholes, including sidetracks, the gel
effectively squeezed into the fracture system of the
coal displacing gas and water. The gel attached itself
to the inner wall of the borehole providing an impen-
etrable skin preventing gas and water to migrate back
into the borehole, as evidenced by observations when
mining into the boreholes.
Consistency problems in mixing and pumping the
gel were overcome when more boreholes were sealed
with the cross-linked poly acrylimide gel.
Improvements currently being implemented include
constructing a more comprehensive gel system
designed specifically for mixing and pumping the gel
ingredients and liquid gel mix. By more uniformly
mixing the precise formulation of individual ingredi-
ents, the ability to pump the polymer gel through the
462
Copyright © 2006 Taylor & Francis Group plc, London, UK
slick-line and into the boreholes to distances greater
than 3 km (10,000 ) will be increased. Importantly,
the polymer gel will be displaced into the horizon-
tal degasification boreholes, either drilled from within
the mine or from the surface, consistently with the
designed pumping viscosity and cure viscosity to keep
the borehole filled with cured gel when intercepted by
mining. The development of this technology, which
permits safe and productive mine-through degas of
boreholes, helps ensure the continued growth in the
utilization of degasification for overall safety and
productivity benefits when mining gassy coalbeds.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the continued
support and perseverance of Pennsylvania Services
Corporation’s and Emerald Coal Resources, LP’s
senior management. Special thanks also go to Albert
Giocondi, Emerald Coal Resources, LP’s Degas Coor-
dinator, whom without his daily commitment with
his boots on the ground to work out the bugs to
seal the boreholes with the cross-linked polymer gel
would never have been successful including acting as
a liaison between management, the work force safety
committee and MSHA during the initial phases of gel
development.
REFERENCES
Aul, G. & Cervik, J. 1976. Grouting Horizontal Drainage
Holes In Coalbeds. U.S Bureau of Mines RI 8843, 16 pp.
U.S. Department of Labor, Mine Safety and Health Adminis-
tration, Program Information Bulletin No. P05-10, Effec-
tive Date: May 5, 2005, Ray McKinney, Administrator for
Coal Mine Safety and Health.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Target Drilling’s long boreholes maximize longwall dimensions

G.M. DuBois & S.J. Kravits

Target Drilling Inc.

J.M. Reilly
Pennsylvania Services Corporation

T.P. Mucho
Thomas P. Mucho & Associates, Inc.

ABSTRACT: To address increasing ventilation and methane emission issues underground horizontal long
boreholes (longholes) are utilized in the mine plan at a large Pittsburgh coal bed longwall mine. Considerations in
designing the placement of long boreholes included existing mine ventilation design, reserve recovery, ventilation
limitations of three entry development, belt air direction, federal ventilation criteria, longwall panel dimensions,
and estimated methane liberations.
Evolution to the current in-mine longhole drilling pattern, shielding gate road development is reviewed.
Utilizing this technology has allowed the mine operator to successfully and profitably mine the largest longwall
panels (416 acres) in the Pittsburgh Coal Seam, extending both length and width of the panels.
Target Drilling Inc. employed their directional drilling technology and an experienced staff, to successfully
drill all longholes utilizing two drills simultaneously. Target has directionally drilled one hundred sixteen in-mine
longholes greater than 1,219 m (4,000 ), in the Pittsburgh coal seam primarily shielding gate road development.

1 INTRODUCTION coal, thereby lowering overall production costs has

Many US Longwall mines are located in relatively
gassy coal beds. Mine planning has aimed to increase
the percentage of longwall coal produced versus con-
tinuous miner development coal tons. This need was
evident and was a result of the coal industry’s need to
stay competitive in a world energy market. At long-
wall mines, these economic forces have fueled the
continuing expansion of longwall panel dimensions,
improvements in reliability of longwall equipment.
These actions have increased productivity and pro-
vided lower costs to the consumer. As we review the
past 15 years of increasing longwall panel dimensions
and productivity, it became apparent, that ventilation
alone may not be sufficient to dilute the ever increas-
ing volumes of methane that are liberated into the
mine’s ventilation system without adversely impacting
advancement rates of continuous miner development
and longwall sections. Due to higher and higher long-
wall productivity and advancement rates, coal oper-
ators continued to expand panel length and width to
maintain a balance between section development rates
and longwall retreat.
The coal industry’s goal to maximize the ratio of
longwall coal to more expensive continuous miner
been very instrumental in the continued expansion
of longwall panel size. It is this expansion of long-
wall panel size that has driven the mine engineer to
incorporate the use of longhole technology into the
design of mine layouts and mine ventilation systems.
Compliance with Federal Ventilation Regulations is
mandatory. Ventilation design always incorporates a
safety factor in its design, and all available options
and technology needs considered. This paper discusses
how the use of, underground horizontal directional
drilled longholes were effectively utilized to increase
longwall panel dimensions, reduce longwall move fre-
quency, enhance safety of the employees and the oper-
ation, reduce citations and increase the profitability of
a Southwestern Pennsylvania underground longwall
mine.
2 HISTORY OF LONGWALL PANEL
DIMENSIONS
During the early 1990’s typical dimensions of long-
wall panels at this mine were 2,438 m × 229 m
(8,000 × 750 ). This area encompassed 558,302

463

Copyright © 2006 Taylor & Francis Group plc, London, UK

 square meters (147 acres) with a Maximum Ventilat-
ing Perimeter (MVP) of the longwall panel = 5,105 m
(2,438 m + 2,438 m + 229 m). As global energy sup-
plies continued to influence the downward trend of
coal prices, coal operators looked to offset the down-
ward price spiral with improvements in productivity.
The obvious choices were to increase longwall panel
size, and at the same time increase the ratio of long-
wall coal to continuous miner coal, thus lowering
production costs. By the mid to late 1990’s longwall
panel dimensions had increased to 3,810 m × 283 m
(12,500 × 930 ) with a MVP = 7,903 m (25,930 ),
encompassing 1,078,230 square meters (269 acres).
Longwall panel dimensions in a new reserve at
the mine were projected to be 4,419 m × 382 m
(14,500 × 1250 ) and a MVP = 9,219 m (30,250 )
encompassing 1,688,058 square meters (416 acres).
Comparing panel size of the early 1990’s to the pro-
jected panels in the new reserve MVP’s and acreage
are projected to increase 81% and 183% respectfully.
Monitoring of methane levels and air flow quantities
indicated there was a direct one to one correlation
between MVP and methane liberated into the section’s
ventilation system. This indicated the requirement for
additional section ventilation capacity. It was appar-
ent that completion of longwall panel development
would be significantly hampered by an inadequately
designed ventilation system, especially since section
advancement is always critical at a longwall mine.
Computer ventilation modeling of the 3 entry gate
road development sections indicated that those current
systems would probably not be capable of deliver-
ing sufficient quantities of air to adequately ventilate
the development section without significantly affect-
ing the section advancement rates. With the projected
increasing MVP’s new concepts were evaluated. Mine
Wide Atmospheric Monitoring Systems and Under-
ground Horizontal Directional Longhole Drilling were
two potentials to supplement the ventilation system.
3 REVIEW OF PANEL SIZE, METHANE
LIBERATIONS AND CONCENTRATIONS
In the early 1990’s when the typical longwall panel was
229 m (750 ) wide and 2,438 m (8,000 ) long typical
volumes at this mine’s section regulators were set at
2,831 cubic meters per minute, (cmm) (100,000 cfm)
and methane levels were just below the legal limits
of 1.0%. The total methane being liberated from a
single three entry development section was approx-
imately 25.5 cmm (900 cfm) of methane when panel
development was near completion.
By the mid 1990’s typical longwall panel’s MVP had
increased to 7,903 m (25,930 ). To meet the require-
ments of the federal and state regulations this operator
began utilizing Atmosphere Monitor Systems (AMS)
which permitted methane levels to increase by 50%, to
1.50% in section return entries. As longwall develop-
ment section’s MVP increased the quantity of methane
liberated by a three entry system also increased. When
reviewing the percent rise in methane generated by a
3 entry development sections compared to the percent
rise in MVP, both entities were increasing at similar
rates (see Tables 1 and 2).
After evaluating Tables 1 and 2 it is apparent
the mine’s ventilation system in a very long three
entry development section will marginally meet the
requirements of the federal ventilation criteria. What
options are available to the mine engineer? This paper
discusses utilization of an aggressive underground
degasification program and how it resolved numerous
regulatory and operational issues.
4 IN-MINE LONGHOLE DEGASIFICATION
In the mid 1990’s the coal operator began to utilize
longhole drilling technology to shield longwall devel-
opment sections from the impacts of methane delays

Table 1. Comparison of actual section methane produced with increasing MVP.

Air quantity @ Methane Total section, methane

MVP regulator (cmm) concentration liberated (cmm)

Early 1990’s (Act) 5,105 m (act) 2,831 0.90% 25.5 (No de-gas)
Late 1990’s (Act) 7,903 m (act) 2,831 1.30% 36.8 (Limited de-gas)
% Increase +54.8% Same +44.4% +44.4%

Table 2. Estimated section methane produced with largest MVP.

Air quantity @ Methane Total section, methane

MVP regulator (cmm) concentration liberated (cmm)

Actual Readings 5,105 m (act) 2,831 0.90% 25.5 (No de-gas)

Estimated Readings 9,220 m (act) 3,398 (Est.) 1.36% (Est.) 46.2 (No de-gas)
% Increase +81% +20% (Est.) +51.1% +81% (Est.)

464

Copyright © 2006 Taylor & Francis Group plc, London, UK

 at the mining face. The initial drilling pattern involved
a single horizontal longhole being drilled parallel to
the return entry of the active section.
This initial drilling pattern provided some relief of
methane accumulations at the mining face. However
because the underground drilling site is developed by
section advancement the horizontal longhole had min-
imal time producing “free gas” that is trapped in the
coal cleats. Within a few months section development
advances passed the end of the horizontal longhole and
the developing section was back to the same problems
associated with methane accumulations at the face. It
was obvious that more time was needed to allow the
horizontal longhole to degasify the coal seam.
Drilling patterns evolved to maximize degasifica-
tion time and drain the “free gas” trapped in the
coal cleats by utilizing cross panel boreholes, this
allowed up to 18 months degasification time prior to
the continuous miner advancement.
It also became apparent the entire development
section had to be shielded with horizontal longholes
to effectively degas the entire mining area; therefore
branches were drilled from the main longhole to obtain
continuous coverage along both the return and belt
entries.
5 MINE PLAN DESIGN – TIMING AND
RESOURCE UTILIZATION
Planning for the new reserve, given the goals of
increasing the mine’s viability, proved to be a depar-
ture from the past. There are many considerations and
Figure 1. Single longholes parallel to return entry.
Figure 2. Cross panel longholes parallel to return and belt entries.
465
Copyright © 2006 Taylor & Francis Group plc, London, UK
issues that must be evaluated and properly engineered
when developing and designing a mine layout for an
existing mine. For a longwall mine to be a viable entity
the first overriding goal that must always be recon-
ciled is that section development must always stay
ahead of longwall retreat. This ensures maximizing
the longwall productivity without major interruptions
caused by section development not being completed
in a timely manner. A second criterion for mine lay-
out is to maximize the available coal reserves. Other
design criteria include minimizing the impact from
roof control and ventilation issues.
Utilizing Surv Cadd Software, numerous mine plan
timing models and configurations were evaluated and
it became apparent that the remaining coal reserve
would not be split into two equal east and west
halves off the mains section development. To meet
the timing model projections, longwall panel length
had to be reduced in the short term to ensure sec-
tion development would be completed on the west
side of the mains. This accomplished the first goal
of the mine engineer: moving the longwall equip-
ment from one recovery to the next set-up without a
delay. Because the mine plan timing issues divided the
coal reserve in unequal halves, the projected longwall
panel’s MVP exceeded any previous developed panel’s
MVP by 17%. However the proposed mine plan layout,
maximized recovery of the coal reserve.
The new mine plan layout proposed longwall
panels 4,419 m (14,500 ) long × 382 m (1250 ) wide,
mining 416 acres per panel, with a MVP = 9,220 m
(30,250 ), Again it was assessed that these larger
panels would create a situation where the presently
 projected ventilation system design might not be suffi-
cient to meet the legal ventilation requirements. With
greater MVPs two ventilation issues were identified
that would be detrimental to executing the mine devel-
opment plan in a timely manner. The two ventilation
issues were
• Methane levels at the mining faces exceeding 1.0%
methane, causing operational delays to the continu-
ous mining process.
• Methane levels in the Section Return exceeding
1.5% methane, resulting in suspension of mining
in the entire operating section.
The strategies to minimize these two ventilation
risks associated with the mine layout utilizing very
large panels were:
1 Increase mine ventilation capacity by increasing fan
horse power 33%.
2 Reduce section return methane concentrations
below 1% thus eliminating the required AMS sys-
tem. This goal would allow a 50% margin of safety
in the ventilation requirements (1.0% to 1.5%) to
meet the federal ventilation standards.
3 Develop an aggressive underground degasifica-
tion strategy to capture and safely transport the
“free gas” to the surface before mining activities
released the “free gas” into the Mine’s Underground
Ventilation System.
6 DEVELOPMENT OF IN-MINE
DEGASIFICATION PLAN
The main emphasis of the in-mine degasification strat-
egy was focused on capturing the “free gas” trapped
in the fractures or cleats of the coal and directing it
into an underground gas pipeline gathering system and
ultimately to the surface (through vertical boreholes),
rather than releasing the gas into the mine’s ventilation
system during mining. This would be accomplished
utilizing safety measures approved by the governmen-
tal agencies in a safe and responsible manner. Because
the main line development split the coal reserve into
unequal parts, special emphasis for degasification was
put into the larger portion of the coal reserve where the
longwall panels were 4,419 m (14,500 ) long, 382 m
(1,250 ) wide with Maximum Ventilating Perimeter
(MVP) of 9,220 m (30,250 ). The degasification plan
focused on those issues that would most effectively
impact the reduction of methane in the coal reserve.
They were:
• Drill longholes as soon as possible to maximize
degasification time.
• Plan drilling far enough in advance to insure contin-
uous coverage along both the belt and return entries
in all three entry longwall development sections.
466
Copyright © 2006 Taylor & Francis Group plc, London, UK
• Utilize surface to in mine vertical boreholes to max-
imize gas flow and vent gas to atmosphere or capture
for sale.
7 MAXIMIZE DEGASIFICATION TIME
Longwall panels were projected to be developed from
both sides of the mains. The smaller western panels
were to be mined first while the larger eastern pan-
els were not scheduled to be mined for 2.5 years.
This allowed a unique opportunity to drill longholes
into the eastern reserve as the main’s development
advanced north, maximizing degasification time for
this portion of the reserve. Drilling sites favorable
to longhole orientation were developed to maximize
longhole shielding along both the belt and return
entries. Sumps were cut in the mine floor to allow sep-
aration of drilling fines and water. Longholes drilled
from the mains section would effectively shield the
first 25% of the gate road development and were com-
pleted prior to the longwall section “neck ins” being
developed. The longholes were planned, oriented, and
drilled through the longwall recovery entries in such
a way that the “neck in” of the development section
would not intersect the longholes, again maximiz-
ing the effective time the longholes were producing
methane. This drilling strategy permitted longholes to
be active for 6–24 months prior to any mining develop-
ment near the longholes. Prior to the continuous miner
section intersecting the longholes, the longholes were
water infused, mined through, and reactivated utilizing
a mechanical flow through packer on each side of the
section where the longhole was intersected.
Additional challenges presented themselves when
the first longwall panels in the eastern side of mains
development were started. The orientation of the belt
entries in the tailgate and headgate sections created
a situation where a degasification longhole could not
follow the typical plan of drilling with the wellhead
in the return entry. This would have left 75% of the
section without sufficient number of shielding long-
holes. If degasification longholes were to be drilled to
shield the belt entries of either the headgate or tailgate
sections, it would require the wellheads would have to
be located in an active belt entry. Working with fed-
eral and state agencies, local mine management and
representatives of the UMWA local, a plan was devel-
oped and approved to permit drilling from the intake
entry, with wellheads in an active belt entry of an active
development section.
The approved plan included the following guide
lines:
• Methane monitor sensors were positioned in
the belt entry. These sensors were continuously
monitored on the surface, and included audible and
visual alarms at the section tailpiece.
 Figure 3. Longholes drilled from mains sections.
Table 3. Comparison of section methane produced without/with degasification.
Air quantity @
MVP regulator (cmm)
Without De-gas (Est.) 9,220 m (act) 3,398 (est.)
With De-gas (Actual) 9,220 m (act) 3,822 (act)
% Increase (Decrease) Same +12.5%
• Methane monitors were wired into the belt starter.
This enabled the belt system to be de-energized
when 1% methane was detected in the belt entry
by the methane sensors.
• Gas transmissions lines that crossed intake entries
from the belt entry to the return were encased in
an additional steel or aluminum pipe for added
protection.
• Wellheads were equipped with pneumatic shut off
valves.
• Gas-water separator tank’s drain lines were coursed
to the return airway to avoid methane being released
into the belt entry in the event the water dump valves
did not function properly.
• Belt air flow was directed outby at the wellheads and
into the return entry before it passed over the belt
drive or head roller. While methane drilling opera-
tions were being conducted in the belt entry, the belt
air was never directed inby to the active working
section.
• A pre-shift examination was conducted at the well-
heads and the results recorded in the pre-shift
examination book.
This approved plan was site specific, and was not
replicated in other locations through out the mine.
With the ability to continuously monitor the methane
levels on the belt entry and have this information con-
nected to the section’s belt load center which was
467
Copyright © 2006 Taylor & Francis Group plc, London, UK
Methane Total section, methane
concentration liberated (cmm)
1.36% (Cal) 46.2 (Est. No-degas)
0.80% 30.6 (Actual, With de-gas)
(41.2%) (33.7%)
programmed to immediately de-energized the entire
belt system when methane levels reached 1%, provided
a level of confidence in the plan by mine management,
employees and government regulatory agencies.
8 CONCLUSION
The use of longholes for degasification has been
extremely successful; it has met the criterion for the
project. These goals were:
1 Reduce production delays at the mining face on
continuous miner sections. Delays associated with
excessive methane at the mining faces has all but
disappeared from the section foreman’s reports.
2 Reduce methane levels in section return below
1%; this eliminates the need for the Atmospheric
Monitoring Systems (AMS). Methane levels in the
section return have dropped to an average peak
of 0.8%, primary due to the degasification pro-
gram. Actual methane being liberated has dropped
approximately 30%–35% (see Tables 3 and 4).
3 Most important, utilizing horizontal longholes
drilled by Target Drilling Inc.’s experienced staff,
has allowed the safe and timely development of
the largest current longwall panels in the Pitts-
burgh coal seam. Development was successfully
 Table 4. Comparison of % rise in MVP versus % rise in section methane liberated without/with degasification.

Air quantity @ Methane Total section, methane

MVP regulator (cmm) concentration liberated (cmm)

Without De-gas (Act.) 5,105 m (act) 2,831 0.90% 25.5 (No degasification)
With De-gas (Act) 9,220 m (act) 3,822 0.80% 30.6 (Actual, With de-gas)
% Increase(Decrease) +80.6% +35% (11.1%) +20%

completed without delays associated with exces-
sive methane at the mining faces or in the mine’s
ventilation system.
When analyzing the value of this project to the mine
operation it is difficult to quantify, because of so many
variables, but the following given.
Production delays for excessive methane associ-
ated with continuous miner development approach
zero minutes compared to 4,000 minutes for previous
sections without degasification.
1 Section advancement rates were not negatively
impacted by methane delays.
2 Section methane delays were reduced 30% to 35%.
3 With the belt air direction reversed from intake to a
neutral split, typical belt entry maintenance such as
rock dusting were more easily managed.
4 Reversing the belt air direction eliminated signif-
icant amounts of methane reporting to the mining
faces (14.2 cmm [500 cfm]). This methane reduc-
tion helped eliminate excessive methane delays.
In summary without an aggressive degasification
strategy the methane delays would have severely
impacted the development of the longwall develop-
ment sections, and the mine plan’s timing schedule
would not have been met. As a result there would
have been two choices, shorten the longwall panels
and lose significant portions of the coal reserve or
idle the longwall for numerous months and lose sig-
nificant amounts of money. Either choice would have
been unacceptable.

468

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Sinking a 990 m surface borehole for methane drainage of a longwall

panel in a UK coal mine

J.R. Leeming
Health and Safety Executive, Sheffield, UK

P.R. Shead
UK Coal Mining Ltd, Harworth, UK

ABSTRACT: Methane drainage is employed in gassy coal mines to capture sufficient of the methane produced
to maintain an atmosphere that is within legal limits. Traditionally, in the UK this has been by a network of
underground boreholes drilled above the waste from the return gate. In early 2005, one already gassy colliery
experienced a major methane emission from the waste that caused production to be suspended. This paper
describes the event and the effects it had, details the work done to control the situation, then shows how the
decision was taken to drill the UK’s first ‘gob well’ as a longer term solution. With the workings lying at a
depth of 987 m, oil well technology had to be employed. Water-bearing strata was passed through and the paper
presents the procedures put into place to ensure the safety of the mine and its personnel as the borehole thirled
with the waste.

1 INTRODUCTION other seams in the UK, it is relatively impermeable, and

Harworth mine produces coal from the Deep Soft
seam, the current production area is at a depth of
987 m. The seam is some 1.9 m thick, and won by a
retreat longwall unit. The mine has always been a gassy
mine, with methane levels controlled by a combination
of ventilation and local methane drainage.
In early 2005, production changed from the
exhausted 19s panel to its replacement, 30s. This face
was 320 m long, with a face run of about 1500 m, at
essentially level gradients.
2 NORMAL VENTILATION PARAMETERS
Mine ventilation was generally homotropal because
of heat and humidity considerations, but to mitigate
the high gas levels, district ventilation was antitropal
utilising 28 m3 /s of clean intake air. With production
rates expected to reach 30,000 tpw, and assuming 50%
methane capture, it had been recognised in the plan-
ning stages that this ventilation quantity would not be
sufficient to maintain less than 1.25% methane in the
tailgate. Steps had been put in hand to improve venti-
lation quantities but at the time of the incident these
had not yet been realised.
The Deep Soft seam has one of the highest methane
contents of any seam in the UK, at 7.8 m3 /t and a
specific emission rate of 20 m3 /t. In common with
pre-drainage although attempted in the past has proven
to be ineffective in reducing the methane content of
the panel of coal to be mined. Gas is predominantly
evolved during cutting, such that if production ceases
then the background methane levels in the tailgate
reduce significantly. Methane levels are typically 1.4%
by the end of the week, and reduce to 0.8% by the time
production recommences after the weekend stand.
Local drainage consists of holes drilled at 7 m inter-
vals just behind the face in the tailgate. The holes are
parallel to the faceline and angled upwards at 55o .
Holes are drilled with an hydraulic borer, and con-
sist of a 24 m drilled length at 75 mm diameter, into
which is fitted 21 m of standpipe, sealed at the mouth
with chippings and gas-fast tape. The upper 3 m of
standpipe are slotted. Holes are coupled to twin paral-
lel ranges in the tailgate, usually of 350 mm diameter,
and hence to twin 350 mm diameter ranges local to
the district. The trunk range in the mine consists of
3 parallel ranges back to the surface.
Suction is applied at the surface by up to eight Nash-
Hytor CL3002 water-sealed pumps, developing 20 kPa
of vacuum at the outbye end of the tailgate. On the pre-
vious face, captures of 300 l/s had been achieved, rep-
resenting 56% of the total district make. On 30s, as the
goaf had not properly developed and not many holes
had been drilled, initially 40 l/s at 15% purity were
being captured, representing 29% of the district make.

469

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Traditionally methane has been used to fire boilers,
for baths water and space heating, or if insufficient
methane was available, vented to atmosphere. With the
identification of methane as a potent greenhouse gas,
mines in the UK are now tending to flare the gas, thus
converting it into the less potent CO2 . In Harworth’s
case, so much methane has been available that some
13 years ago an 18 MW power station was built next
to the mine to utilise the gas.
3 THE EMISSION EVENT
On 12th February 2005, after only 40 m of retreat on
the new district, extremely high methane levels were
experienced in the tailgate of 30s panel. With insuf-
ficient district methane drainage boreholes drilled,
methane was being emitted from the goaf onto the
faceline at approximately the mid point. Methane
levels in the tailgate quickly rose above 2%, at which
point manpower was withdrawn and the road fenced
off. Levels continued to rise, reaching +5% in the
main return in 40 m3 /s. Although 109 l/s was being
captured, this now represented only 5.45% of the total
district make.
The methane also backed up against the district ven-
tilation, preventing access to the face. The 2% fringe
in the maingate was tracked; its worst position found
on 15th February 2005 at the 670 m point, about 900 m
back from the faceline. The fringe had retreated back
to the face entrance at the intake end on the follow-
ing day, and access was gained through the face to the
tailgate end on 17th February 2005.
The event was attributed to an unusually high emis-
sion that showed no signs of abating over the next few
days. To give some indication of the magnitude of the
event, it has been determined that at its peak, 1800 l/s
of pure methane were being emitted; and that over
the initial seven days of the event, the emission rate
averaged 1200 l/s of pure methane.
The face stood from 13th February to 13th March
2005 as a result of the gas levels. By this time, gas
levels had reduced to a point that enabled limited
production to resume. That allowed further methane
drainage holes to be drilled and put onto suction, thus
improving gas capture.
4 POSSIBLE SOLUTIONS
With production curtailed, and return roadways having
experienced an explosive mixture, initial steps were to
contain the situation and to recover those parts of the
mine accessible. Methane levels in the return stabilised
at 1.6%. Development work had also been suspended,
and alterations were made to the ventilation in that
area of the mine to bleed clean air into the returns. The
returns were degassed in sections, with 30s district
470
Copyright © 2006 Taylor & Francis Group plc, London, UK
being left until all the return roadways were back to
their normal methane levels. This enabled electrical
switchgear to be purged.
4.1 Ventilation improvements
4.1.1 Airway resistance
Resistances in some of the outbye roads had increased
as a result of considerable and continuing floor heave.
A pressure survey had been carried as part of the plan-
ning procedure for the new booster fan and a resistance
of 1.12 gauls determined. Roadway repair works insti-
gated before the face commenced production, had
decreased the resistance to 0.70 gauls; however, the
installation of a 1372 mm wide conveyor belt and
400 mm diameter methane pipes, and further floor
heave had raised the resistance to 0.89 gauls. The con-
tinuing repair program, although being concentrated
on, did not hold the prospect of any significant benefit
in the short term.
4.1.2 Booster fan
Ventilation predictions in the planning stages of the
district had identified the need for uprated booster fan
performance. The mine had two booster fans operating
outbye: it had been planned to replace these with a
single custom designed installation sited inbye. The
extensive mining works required had been completed
and the booster fan delivered and installed, but upon
commissioning was found to improve inbye quantities
by only 5 m3 /s rather than the 30 m3 /s predicted, and
required. The cause of this lack of performance was
identified as the fitting of incorrect inlet guide cones,
compounded by the fact that outbye roadway resist-
ances were too high, as noted above. New cones had
to be made, supplied and fitted, and instead of the
new booster running as a stand alone fan, the outbye
boosters had to be restarted and run in parallel.
4.1.3 Local ducted ventilation
To give an immediate improvement in tailgate gas lev-
els, a small fan was installed in intake leakage and
some 7 m3 /s of clean air were ducted into the tail-
gate. The duct was advanced as methane levels were
reduced, thus enabling access into the tailgate to carry
out works to further improvements to the methane
drainage system. The extra clean air diluted tailgate
gas levels to 1.4%, so allowing access. The technique
had been used occasionally and by exception in the
UK, as removal of a significant amount of intake air
tended to starve other places. That said, the technique
is a useful expedient in an emergency.
4.2 Methane drainage
4.2.1 Local drainage
With little prospect of any significant improvements
in district ventilation, improvements to the methane
 drainage system were sought. Provision of extra holes
on the range was not feasible. The holes were already
optimally placed at 7 m intervals to avoid interaction
and the tailgate behind the face, although standing, was
gas-fast and inaccessible. With the impermeability of
the strata, holes in advance of the face do not produce
worthwhile quantities of gas, and are invariably lost as
the face passes.
Suction could be improved by concentrating the
effect of the surface plant onto this district by regulat-
ing other legs in the mine. This was done, and resulted
in the suction at the outbye end of the district rising
to 22 kPa, and raising capture to 262 l/s, or 47% of the
district make.
Extra capacity was also sought, to complement the
additional suction. At the time of the event, only one
350 mm range had been installed in the tailgate, with
the second range just started. This was quite normal, as
the second range would not be required until about 10
or 12 holes were on the first range. With access gained
in stages to the tailgate, and during this stand in pro-
duction, the methane drainage system was improved
by completing the 350 mm diameter second methane
range in the tailgate and coupling it to the second out-
bye range. This took 14 days, but once coupled enabled
a vacuum of 9 kPa to be applied at the inbye end of
each range, improving capture to 913 l/s pure gas at
70% purity.
4.2.2 Remote drainage
A radical solution, as yet untried in the UK, was pro-
posed by the chief executive of the operating company,
from his experiences working in the US. This was to
drain the excess gas to the surface using a dedicated
borehole drilled for the purpose. Although expensive,
and untried at this depth, the fact was that none of the
current efforts looked likely to provide a solution if the
emission continued at its present rate. The suggestion
therefore was felt to have some merit and was pursued.
The concept was to drill, vertically if possible, from
the surface to intersect the goaf near to the face start
line. Suction would then be applied at the top of the
hole, and the gas vented or flared. It was hoped that the
hole would continue to produce gas as the face worked
away, as being vertical it should be relatively immune
to the effects of caving. This decision was made at the
end of February 2005.
5 THE SURFACE BOREHOLE
5.1 Siting
A vertical hole was preferred on the grounds of cost,
and fortunately a disused sand quarry existed directly
over the tailgate half of the face, above the goaf
being formed. Permission was obtained from the
landowner to access the site and drill. In accordance
471
Copyright © 2006 Taylor & Francis Group plc, London, UK
with boreholes regulations1 , the intention to drill was
formally notified to the Health and Safety Executive.
5.1.1 Location
A roadway had to be constructed within the quarry to
allow the mobile drill rig to access the drilling site. The
quarry was disused, consequently there were no mains
services. Water for the mud farm would be brought to
site by bowser from the mine, some 8 miles distant.
Other than that the rig was self contained and diesel
powered.
5.1.2 Specialist contractor
The mine operating company had no in-house expert-
ise in this type of exploration drilling, and a specialist
contracting company (British Drilling and Freezing
Ltd) was engaged to assist with the planning and design
of the hole, and to drill it. It was confirmed that their
No. 28 rig would be available during April and May
2005. Various parameters were specified for the design
of the hole to provide a specified finished diameter and
to avoid various hazards, as described below. Contracts
were agreed and signed in March 2005.
5.2 Inrush hazards
UK Regulations2 require a mine owner and manager
to assess any inrush hazards to their workings; and if
necessary to construct a scheme of work to minimise
the risk from those hazards. Usually risks are presented
by the mine workings approaching a static hazard, e.g.
a previously drilled surface borehole, or old workings.
In this case the mine workings were essentially static,
and the hazard was approaching them. Various con-
cerns were raised: the following specific matters were
taken into account.
5.2.1 Aquifers
Extensive knowledge of the strata to be drilled through
was held, both from shaft sinking records and also
from an exploration borehole, drilled in 1990 and
only 174 m away from the proposed methane extrac-
tion borehole. Notable aquifers include the 151 m
thick Sherwood Sandstone which outcrops at surface;
and the 53 m thick Lower Magnesian Limestone at
around 300 m depth. These aquifers produced 500 and
170 gpm (38 and 13 l/s) respectively during shaft sink-
ing. Other aquifers are present within the upper Coal
Measures, the lowest significant being the 15 m thick
Mexborough Rock at about 600 m depth. It was essen-
tial the borehole would not present a flow path for
these waters to access the mine workings; this would
be achieved by a double layer of casings and cement
grouting, as detailed below.
5.2.2 Old workings
Harworth mine was sunk to exploit the Top Hard seam
at a depth below surface of 870 m and old workings are
likely to contain gas and/or water. Production moved
 to the lower Deep Soft seam in 1981. Fortunately the
Top Hard seam had not been worked in the area above
30s panel.
5.2.3 Mud flush on thirling
As the bottom of the hole approached 30s goaf, at
some point the head of mud within the hole would
overcome the strength of the intervening strata and
the hole would then empty into the goaf. It was veri-
fied that the 209 barrels (33.25 m3 ) of mud contained
within the hole was unlikely to be seen on the face, and
would certainly not pose a hazard from the liquid. This
mud would however, displace a similar volume of gas
from the goaf. If high purity methane lay immediately
behind the powered supports, there was a danger that
this would be displaced onto the faceline, to pass over
electrical equipment or shearer cutter picks. To remove
the ignition risk, a thirling procedure was drawn up that
detailed communication from the rig to the coal face,
and suspension of work at the point of breakthrough.
5.2.4 Effect on mine ventilation
There were further potential dangers resulting from
the connection of the hole to the goaf. Firstly, the pos-
sibility of fresh air being pulled down the borehole
by the low relative pressure within the mine caused
by the mine ventilating fans. This fresh air could then
dilute goaf gasses into the flammable range. Secondly,
the seam is very susceptible to spontaneous combus-
tion: a further critical reason to prevent air ingress into
the goaf. To prevent any airflow in that direction, it
was arranged that on loss of mud at thirling, the hole
would be sealed at the surface until suction could be
connected and applied.
5.3 Design of the borehole
The coal company proposed to install a portable
methane drainage plant that was available and could
develop 60 kPa. It was fitted with a vent to atmos-
phere. Mine staff estimated that drawing 500 l/s of
pure methane would enable the tailgate to be kept
below 1.25%, and it was assumed that methane would
be drawn up the borehole at 50% purity, thus requir-
ing 1000 l/s of mixture to be exhausted. Over a
1000 m pipe, this would require a pipe diameter of
203 mm (8 ).
From the required finished size of the hole, and
length needing to be sealed against the aquifers,
drilling and casing sizes could then be determined. The
proposal was to sink a borehole as follows.As an added
precaution, a blow out preventer was fitted as soon as
the hole was established, as is normal practice. (Imper-
ial units used as per oil drilling standard practice).
– Drill 24 diameter 0–30 m.
– Install and cement to surface 185/8 diameter casing
0–30 m.
472
Copyright © 2006 Taylor & Francis Group plc, London, UK
– Drill 17½ diameter from 30–600 m.
– Install and cement to surface 133/8 diameter casing
0–600 m to seal off Sherwood Sandstone, Magnes-
ian Limestone, Basal sands, and Coal Measures
aquifers to below the Mexborough Rock.
– Drill 12¼ diameter from 600–910 m.
– Install and cement to surface 95/8 casing 0–910 m.
– Drill 8½ hole to Deep Soft seam at approx. depth
of 972 m. This length not cased.
5.4 Thirling procedure
The thirling procedure was devised to minimise the
hazard to the workforce from the thirling of the bore-
hole with the goaf and the consequent sudden release
of the mud flush. The emphasis was placed on good
communications from the drill rig to the mine, and to
the underground district. During drilling and casing to
the 910 m point, daily written reports were presented
from the Bore Master to the mine, specifically to the
manager, undermanager, surveyor and environmental
engineer. The surveyor monitored the position of the
borehole and of the working face.
Once drilling below 910 m started, a member of the
colliery supervisory staff was present at the drill rig
at all times. An operational control was established at
the mine, to continually communicate with the rig site
and manned by a mine official. All communications
were logged. In addition an underground control was
established, managed by the shift manager. Progress
of the hole was monitored closely and when loss of
drilling fluid was expected, all personnel on the face
were to be moved to the intake side.
5.5 Drilling history
The rig site was commissioned and drilling com-
menced on Sunday 3rd April 2005, with the hole
established by drilling at 24 diameter for 43 m, and
grouting in an 185/8 casing. This was completed on
5th April. The hole was then drilled at 17½ to 628 m,
to pass through all the water bearing strata. Penetra-
tion rates varied as different strata were encountered:
only 20 m were drilled from 11th–12th April but 80 m
were drilled from 14th–15th April. Drilling continued
beyond the planned 600 m depth to 628 m depth, to
ensure a 30 m clearance below the bottom of the
Mexborough Rock. The 628 m point was reached on
20th April, whereupon the hole was lined with a 133/8
casing, grouted to seal the water bearing strata.
Drilling recommenced at 12¼ diameter towards
a point 100 m above the seam. This section of hole
passed through numerous coal seams, which caused
problems with mud loss. At 839 m depth, on 27th
April, coincident with the 0.45 m thick Main Smut
seam, circulation was lost for the first time. Remedial
action was to thicken the mud with coconut husks; this
proved successful. The second circulation loss was at
 863 m depth on 28th April and was coincident with the
2.17 m thick Top Hard seam. Similar remedial action
was applied and the hole continued. A third circulation
loss then occurred at 869 m depth on 29th April, at a
point between the Top Hard and 1st Waterloo (0.23 m)
seams. Again the mud was thickened in a similar man-
ner, and the hole reached 896 m depth on 30th April.
In view of the continuing difficulties with mud loss,
the decision was taken to insert the 95/8 casing at this
point. During sealing of the casing, grout losses were
again high and only the bottom 25 m was eventually
confirmed as grout sealed to the strata.
Drilling recommenced at mid-day on 4th May 2005
at 8½ diameter. With the base of the hole some 95
metres above the seam, and the face having retreated
to a position 160 metres outbye of the borehole, the
thirling procedure was put into effect. Water was used
for the final stages of drilling as a flushing medium,
rather than mud, and losses were incurred. In the first
12 hours, 1188 barrels (189 m3 ) were lost; in the 24
hours of 5th May 2005 a further 4032 barrels (641 m3 );
in the 24 hours of 6th May 2005 a further 1560 barrels
(248 m3 ) were lost. All these losses were total circula-
tion losses i.e. there was no return to the surface at all.
In an effort to stop the continuing circulation losses,
on 6th May 2005 at a depth of 965 m, the hole diameter
was reduced to 6¾ . On 7th May 2005, a further 1390
barrels (221 m3 ) were lost, plus a further 209 barrels
(33 m3 ) on thirling, from the volume contained in the
hole. The hole completed at 992 m depth, some 5 m
below the base of the seam.
Steering the hole was considered not necessary.
Low order surveying was undertaken at 150 metre
intervals followed by a higher order directional survey
at 870 metres hole depth. This determined a 9-metre
deviation to the south-southwest and extending that
degree and direction of deviation to total depth, it
was estimated that the hole would be 8 metres inside
30’s goaf.
6 RESULTS
6.1 Effects on thirling
On thirling, with the loss of water from the hole, no
effect was noticed on the face from a flush of goaf
gasses, and there was no measurable change to the dis-
trict ventilation or gas levels. The portable methane rig
and venting stack were ready at the drill rig site, and on
the afternoon of 7th May 2005 these were connected
to the hole, and 300 l/s pure gas drawn out. On Sunday
8th May 2005 the suction had to be turned off to enable
the drill string to be tripped out and the drilling equip-
ment de-camped: this entailed suspension of suction
for four days.
Vacuum was re-applied on the following Friday,
13th May 2005. Good purity but very little flow was
473
Copyright © 2006 Taylor & Francis Group plc, London, UK
achieved, and the conclusion was reached that the hole
had blocked.
6.2 Attempts to clear blockage
The initial intention was to re-drill the hole to clear the
blockage; however the drill rig had moved to another
job and would not be available for some weeks. On
27th May 2005, in an effort to clear any blockage,
18,000 gallons (82 m3 ) of water were released into the
hole. This had no effect, and on 3rd June 2005 the
hole was plumbed to acquire data. A camera survey
had been contemplated but in the event the camera
was not used, as simple plumbing of the hole would
be quicker, cheaper and more effective. The plumbing
exercise produced quite conclusive evidence. Standing
water was encountered at 864.6 m depth, and the hole
was found closed at 898 m depth −5 m below the base
of the casing. On Monday 6th June 2005 a further
82 m3 of water were released into the hole. Later that
day 1000 gallons (4.55 m3 ) were released, followed by
10,000 gallons (45.5 m3 ) of continuous discharge, then
a further seven, 1000 gallons (4.55 m3 ) discharges. All
this work was subject to further delays as the water had
to be brought to site by bowser.
None of this water was seen on the district, which by
now had worked away: on 4th June 2005 the face was
approx. 250 m outbye of the borehole. A final attempt
to free the blockage was made using compressed air.
The hole was pressurised to 1.9 MPa, to no avail.
Gas samples were taken and analysed: the ethane/
methane ratios are known for the local seams and
confirmed that the gas was from the Deep Soft seam.
Efforts to open the hole were then abandoned, and
although the hole was left with suction applied, it never
regained any productivity. The hole was eventually
cement sealed to the surface in early October 2005.
7 SUBSEQUENT METHANE DRAINAGE
IMPROVEMENTS
With the face working outbye, more holes could be
drilled from the tailgate and put on the range. Suc-
tion had already been improved by the addition of
the second range. Eventually, up to 1300 l/s of pure
methane was captured, at that point some 82% of the
total district make.
The most significant factor in obtaining improve-
ments in underground methane capture was the fact
that more holes could be maintained on the system.
This was achieved essentially by better protection of
the standpipes, thus enabling more holes to remain pro-
ductive. As a consequence of roof support difficulties,
extra support in the form of rib dowelling, cribs and
truss bolting was being set outbye the face in the
tailgate to counteract front abutment pressures. This
additional support was found to have a significant
 effect on maintaining the integrity of the local drainage
boreholes, as the tailgate road was kept open behind
the face.
For the next face, the drainage system will have a
third pipe range installed to provide even more cap-
acity. It is intended that a 250 mm range be added.
This will be kept for the normally unproductive floor
holes and to enable cross coupling of the older holes;
to leave the two 350 mm ranges available for the most
productive holes near to the face.
8 CONCLUSIONS
This hole was considered unique, particularly in UK
mining and possibly in the world for the depth of the
hole for this application and for its close proximity to
a working face.
The hole was initially productive, and can be
deemed to have been an initial success.
The inner casing was not properly grout sealed back
to surface, so some of the intended protection against
the aquifers was therefore not installed. As a further
474
Copyright © 2006 Taylor & Francis Group plc, London, UK
consequence, an extra 270 m or so of the hole was
essentially unsupported, although outside the casing.
The finished hole was only 6¾ diameter, not the
8½ diameter originally planned. This will have some-
what reduced the potentially achievable drained flow.
The hole had collapsed below the casing. Coal
Measures strata in the UK are relatively weak and not
likely to be self-supporting, so such holes need a cas-
ing for their full depth. To aid gas capture, a perforated
casing for the last 50 m or so would be advantageous,
as has been found to be the case with the underground,
district methane drainage standpipes.
There may be scope for repeating the exercise, but
at a cost of +£700,000 it is likely to be used as an
emergency measure only.
REFERENCES
1. The Borehole Sites and Operations Regulations. 1995.
SI 1995 No. 2038. London, HMSO.
2. The Mines (Precautions Against Inrushes) Regulations.
1979. SI 1979 No. 318. London, HMSO.
 8. Numerical modeling

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Ventilation simulation programs MineVent and MFIRE: Updates to

advance the technology of simulation programming

R.J. Hardy & K.A. Heasley

Mining Engineering, West Virginia University, Morgantown, WV, USA

ABSTRACT: Advances in computer technology and programming styles have allowed increased safety and
productivity in the mining industry. Most of these technological advances have already been implemented
regarding various ground control and roof support prediction and modeling programs; however, such is not the
case with respect to ventilation. The scope of ventilation programs currently available is limited, and those useful
ones that are still accessible, such as MineVent and MFIRE, are quickly becoming outdated while in their present
state. This particular research project aims to update these two specific ventilation prediction programs using the
object-oriented programming language Visual C++ and increase their functionality to the mining community
as well as their user-friendliness, maintainability and portability.

1 BACKGROUND outputs color-coordinated AutoCAD drawings that

There are several ventilation simulation software pack-
ages available for research and private industry use
with respect to mining applications; however, since
the time when these programs were originally devel-
oped, there have been very few updates to keep them
current with technology. In addition, these software
packages have been very useful with reference to their
own specific area and function, but a combination of
their capabilities would offer a wider range appliance
for which all users would benefit.
1.1 MineVent
MineVent is an integrated computer-aided mine plan-
ning software module offered by Ohio Automation
that was originally based on the Penn State University
ventilation program PENVENT and some steady-state
aspects of the US Bureau of Mines ventilation pro-
gram MFIRE (Ohio Automation 2003). The MineVent
solution algorithm uses a classic “network” analysis.
MineVent is designed to run from within the Auto-
CAD interface working with the as-mined and pro-
jected timing maps. These planning maps serve as the
basis for drawing the ventilation schematics to-scale.
Also, forms within the AutoCAD interface allow the
user to input critical airway properties (such as length,
size, perimeter, surface roughness, k-factor, etc.) and
fan characteristics.
Once the input file is created, the MineVent solver
is run from within AutoCAD. For output MineVent
illustrate the branch airflow quantities and node pres-
sures, either separately or superimposed on the mine
map itself. The available output also provides informa-
tion for setting regulators.
1.2 MFIRE
MFIRE was developed for the Bureau of Mines
under contract to Michigan Technological University
to predict the spread of fumes and other products-
of-combustion in a multilevel mine network scenario
under real-time constraints (Edwards & Greuer 1982).
This mine ventilation computer program was derived
from an earlier program released by Michigan Techno-
logical University and restructured to include real-time
prediction calculations for contaminant spread in addi-
tion to steady-state analysis using a least squares
interpretation of fan characteristics (as opposed to the
Lagrangian interpolation of fan data).
The procedure for computation takes into account
airway resistances, fan interactions, and thermal
exchange with airway walls. Buoyancy-induced nat-
ural ventilation is directly calculated using airway
temperatures which changes according to thermal dif-
fusion into the airway walls. Fire is then quantified by
the heat generation in determining the effects on the
natural ventilation (Edwards & Greuer 1982).
The program simulates airflow in a multilevel mine
network by assigning each junction, or crosscut and
intersection, and branch or airway with a specific and
unique identification. The program then automatically

477

Copyright © 2006 Taylor & Francis Group plc, London, UK

 forms closed paths, or meshes, throughout the net-
work. Conservation of energy is applied with each
mesh; and similarly, conservation of mass is applied
with each junction (Kirchoff’s laws).The program then
iteratively determines the airflow rates and tempera-
tures throughout the mine network (Edwards & Greuer
1982).
The advantage of MFIRE over a more basic mine
ventilation analysis program, such as MineVent, is its
capability to model thermal events and fire situations
predicting the propagation of products-of-combustion
and contaminants. The real-time concentration calcu-
lations initially assume a steady ventilation flow, as this
is a reasonable approximation for fires in their initial
development stages (Edwards & Greuer 1982). At this
particular point in time, the ventilation changes only
occur within close proximity to the fire. Time plays a
crucial role in concentration distributions as gases nor-
mally travel with the ventilation currents. Simulation
of these distributions is governed by the law of mass
conservation. Because of this, the mine atmosphere
in MFIRE is divided into homogeneous concentration
control volumes (Laage et al. 1995). These control vol-
umes are difficult to handle when programming, due to
separate volumes meeting at an intersection from dif-
ferent pathways to create new control volumes. This
creates a problem of sorting, for which object-oriented
programming becomes ideal.
1.3 FORTRAN
FORTRAN, or Formula Translator, is what was known
as a high-level computer programming language and
was developed by IBM Corporation between 1954
and 1957 for scientific engineering purposes requir-
ing complex mathematical computations (Metcalf &
Reid 1999). Hundreds of high-level languages have
been created; however, few have had the recognition
that FORTRAN has earned. While FORTRAN is still
widely used in many engineering applications, it is dif-
ficult to update and not easily maintained, since it is
quickly losing support by the newer and more popular
operating systems currently released for computers.
Both MineVent and MFIRE in their current states
employ the FORTRAN programming code. MineVent
was written in the earlier FORTRAN 77 language,
while MFIRE was updated in August 1995 from the
original FORTRAN 77 to use the visual FORTRAN
90/95 language.
2 OBJECT-ORIENTED MODELING AND
VISUAL C++
Object-oriented software takes a real world approach
to modeling “objects” using a language-independent
design organized around those objects. This technique
478
Copyright © 2006 Taylor & Francis Group plc, London, UK
for simulation focuses primarily on understanding
clear requirements when stating a problem, and plan-
ning a logical solution method in the design process
to ultimately implement a user-friendly program in a
particular programming language – in this case, Visual
C++. The emphasis when using object-oriented pro-
gramming is upon the object constructs as models
of real entities, rather than just simple programming
techniques. Using this approach results in realistic pre-
dictions relating to real-world objects and is typically
regarded as being more user-friendly because these
programs are easier to understand, correct, modify and
update.
2.1 Object-oriented modeling and analysis
The actual term “object-oriented” refers to the orga-
nization of a collection of discrete objects which
incorporates both data structure and behavior. The
characteristics required by this methodology include:
identity, classification, polymorphism, and inheri-
tance (Rumbaugh et al. 1991). Identity can be defined
as quantifying the data into discrete, distinguishable
items known as objects. These objects may be solid
or conceptual entities; but each object is unique and
possesses its own distinct identity, even if two objects
have attributes that are identical.
The term classification indicates that all similar
objects, or those that have the same attributes (or data
structure) and behavior (or operations), are grouped
together in a class. A class is an abstraction that
describes properties important to an application while
ignoring any others. Each class is a (possibly infi-
nite) set of individual objects, or instances, within that
particular class.
Polymorphism means that one operation may
behave differently when referencing different classes
(Rumbaugh et al. 1991). Here, an operation refers to
an action or transformation to which an object is sub-
ject or that it may perform. A specific operation by
a certain class is known as a method and because an
object-oriented operator may be polymorphic, it may
have more than one method associated with it.
Inheritance may be defined as the sharing of
attributes and operations among classes based on a
hierarchical relationship. Typically, classes are defined
broadly and then successively refined into subclasses.
Each of these subclasses incorporates, or inherits, all
the properties of its specific superclass while adding
its own unique properties.
Each of these aforementioned properties char-
acterizes mainstream object-oriented languages and
together, they complement one another synergistically.
The emphasis is placed on what the object actually is,
rather than its function or how it is used. This illustrates
the benefit of using the method of object-oriented pro-
gramming. Even though object-oriented development
 may involve more time as opposed to conventional
development methods, it is intended to promote future
utilization and reduce downstream errors and main-
tenance as subsequent revisions are more localized
requiring fewer debugging iterations.
2.2 Visual C++
The programming language C++ evolved from C
which was formulated by Dennis Ritchie of Bell Lab-
oratories in association with Brian Kernighan, and
came from the earlier languages BCPL – established
by Martin Richards in 1967 for writing operating sys-
tems software and compilers, and B – developed by
Ken Thompson (Deitel & Deitel 2003). C++ is an
extension of C produced in the early 1980s by Bjarne
Stroustrup at Bell Laboratories and is useful because
of its capabilities for object-oriented programming
(Deitel & Deitel 2003).
Using the modular object-oriented design and
implementation approach to software development
promotes a much more productive and easily main-
tained software program as opposed to earlier
programming techniques, such as structured program-
ming. The advantage of object technology is essen-
tially a packaging scheme that encourages creation
of meaningful software units that are focused on
particular application areas.
The actual world is comprised of real objects and
thus, makes sense to organize software programming
in such a manner. Before object-oriented languages
became popular, programming languages, such as
FORTRAN, were used with key focus being upon
actions rather than real things or concrete objects.
The paradigm shift of a programmer, living in a
world of real concrete things, having to focus pri-
marily on actions, was a difficult transition and
made the development of software awkward and dif-
ficult to manipulate. Now with its availability, the
shift to object-oriented style has resulted in signifi-
cant productivity enhancements and for programs like
MineVent and MFIRE, it promotes a more natural
procedural programming process.
3 IMPLEMENTATION AND DESIGN
In order to effectively incorporate MineVent and
MFIRE into a combined ventilation modeling pro-
gram, the individual programs must be dissected to
determine what objects are contained therein, as well
as the specific function of each. These objects each
have their own identity and therefore, must be orga-
nized into classes for more streamlined algorithmic
calculations. In each of the classes, multiple checks
are performed routinely to ensure data consistency and
that values calculated or entered are within reasonable
ranges for the different variables.
479
Copyright © 2006 Taylor & Francis Group plc, London, UK
MAIN
NETWORK CLASS
BRANCH CLASS
FAN CLASS
CONTAMINATION CLASS
Figure 1. Diagram of object-oriented programming style.
3.1 Classes
The term class refers to the abbreviation object class
(Rumbaugh et al. 1991). Objects within a class, such
as branches or airways, have the same attributes and
similar behavior patterns – each branch object within
the branch class has an associated starting and ending
node, resistance, airflow quantity, etc. Most objects
achieve individuality from differences in their attribute
values and relationships to other object. Different
branches will have different lengths, resistance, etc.
The main body of the program is divided into sepa-
rate classes for coherent programming structure – here,
they include: the network class, the branch class, the
fan class, and the contamination class (Fig. 1). All
classes, calculations, printing subroutines, and other
major algorithms are called within the main portion
of the program; however, the actual solution method-
ology and iteration processes are contained within the
classes themselves.
3.1.1 Network class
The network class interprets the input data and orga-
nizes it into a coherent ventilation system usable
for modeling purposes. The network class is capa-
ble of analyzing the network topology and calculating
topology parameters, testing the branch parameters
for consistency, switching branches if necessary, and
ordering the branches by type and increasing value of
resistance (Fig. 2).
The spanning tree, created from the normal and fan
branches, is also developed within the network class
and is ordered consistent with the branch nesting level.
Here, the meshes are also determined along with the
corresponding direction of each mesh and a positive
or negative direction for each branch. The branch air
quantities are then initialized in order to perform the
Hardy-Cross calculations.
The Hardy-Cross method for iterative network solu-
tion is also executed within this class.The Hardy-Cross
method for solving network systems has long been the
convention method and is used in this case. This par-
ticular calculation loops through the input data using
successive iterations until: a satisfactory solution is
 The network fundamentally manages the branch,
START
fan, and contamination classes as well as specific
structures or items that use only values, rather than
Read Input Data a combination of parameters with methods or func-
tions that is typical of a class (Fig. 3). These struc-
Create Network Topology tures include the average/default values used for input
parameters as well as the real-time modeling parame-
ters for the dynamic simulations.
Order Branches The real-time parameters are used only once per
simulation. Variables used within the real-time struc-
Create Spanning Tree
ture are: the time duration and interval for the event, the
threshold value for critical contamination, the accu-
racy of the contamination calculations, volume flow
Order Tree Branches rate of the contaminants, concentration of the con-
taminants, and heat entering the airway, in addition
to fume and heat production of oxygen delivery for
Create Meshes
fuel-rich fires and oxygen concentration for oxygen-
rich fires. The average values for rock temperature,
Initialize Air Quantities thermal diffusivity, conductivity, friction factor, air-
way length, perimeter, and cross-sectional area are also
stored within the average value structure.
Perform Hardy-Cross

3.1.2 Branch class

Figure 2. Program flow diagram within network class. The branch class uses the input data pertaining to
all branches connected with the network. This class
defines all the specific types of branches available
NETWORK CLASS for network structure, utilizes the properties associated
with each branch, checks the values for consistency,
and calculates the pressure differentials (Fig. 4). The
BRANCH CLASS branch class also deals with the volume of methane for
that particular airway, as well as the rock temperature,
the rock thermal diffusivity, the rock thermal conduc-
NODES tivity, and the elevation change associated within that
particular airway.
There are numerous variables or attributes associ-
ated with the branches or airways contained within
FAN CLASS
a ventilation network. The sample illustration shows
many, but is not limited to just these (Fig. 4). Branches
each have starting and ending junctions, a specific
CONTAMINATION type – whether it be a normal branch, a fan branch,
CLASS a leakage branch, a limited quantity branch, a fixed
quantity branch, or a dummy branch.
For each of these branch types, there are specific
STRUCTURE: attributes inherent to each; for example, a fixed quan-
AVERAGE VALUES tity branch has a fixed airflow quantity in addition
to the other properties associated with all the other
branches in the network. Essentially, the fixed quan-
STRUCTURE: REAL- tity branch inherits the attributes of the Branch class;
TIME PARAMETERS as does the fan branch which in addition has a corre-
sponding reference number. The leakage branch has a
connected coefficient for leakage, the limited quantity
Figure 3. Breakdown of network class. branch has a limited airflow quantity, and the dummy
branch has a related gas concentration (Fig. 5).
achieved within a user-specified tolerance level, or the
maximum number of iterations is reached—a numeric 3.1.3 Fan class
value also defined by the user through visual input The fan class contains all the properties associated
screens. with the fan or fans within the ventilation system. This

480

Copyright © 2006 Taylor & Francis Group plc, London, UK

 BRANCH Class
Start Node
End Node
Branch Name
Branch Code
Branch Type
Pressure
Quantity
Resistance
Natural Ventilation
Pressure
Methane Production
Rock Temperature
Thermal Diffusivity
Rock Conductivity
Elevation Change
Figure 4. Attributes of the Branch Class.
class is responsible for obtaining the air quantities and
determining from the data the appropriate fan curve
or curves. If there are multiple fans within a single
branch, this class performs calculations to determine
airflow quantities and pressures, whether the fans are
in series or parallel.
The fan class has a complete class behavior contain-
ing both attributes of the fans and methods correspond-
ing to them (Fig. 6). It determines the characteristic
curves for the fan or fans while also storing the fan
481
Copyright © 2006 Taylor & Francis Group plc, London, UK
BRANCH Class
Branch Type
Normal Branch
Fan Branch
Reference Number
Leakage Branch
Leakage Exponent
Limit Quantity
Limit Air Quantity
Fixed Quantity
Fixed Air Quantity
Dummy Branch
Gas Concentration
Figure 5. Illustration of different branch types in the Branch
Class, and specific variable linked with each.
specific information, such as fan make, model number,
blade setting, axial speed, air pressure and air density
among other relevant parameters. One unique aspect in
this class is the methods for calculating an equivalent
fan from known fans in parallel or series.
3.1.4 Contamination class
The contamination class handles all the contaminant
and product-of-combustion input parameters (Fig. 7)
associated with the steady-state and real-time simu-
lations. Here they may be applied for use with both
the steady-state calculations as well as the real-time
simulation calculations.
The gases monitored include methane and oxygen
as well as the concentration of the contaminant flow.
The fume production and heat production of oxygen
delivery for fuel-rich fires, as well as oxygen con-
centration with which fumes leave the fire zone for
oxygen-rich fires is also taken into consideration.
The methane is measured by volume in cubic feet
per minute and/or as a production rate per unit surface
area (cfm per square foot). This class also references
 4 SUMMARY AND CONCLUSIONS
FAN CLASS
MineVent and MFIRE are highly useful tools for ven-
tilation modeling in the mining industry; however, as
written in FORTRAN, they are in danger of becoming
ATTRIBUTES unsupported by newer computer technology. It is cru-
cial that these valuable simulators do not get lost in the
steady march of computer operating systems.Also, it is
FAN NUMBER important that these programs be maintainable so they
are compliant with recent standards and upgradeable
to accommodate new supplemental features useful for
mining applications.
MAKE / MODEL Currently the updating and merging of the Mine-
Vent and MFIRE ventilation simulators is still in the
coding and debugging stages. However, when the pro-
grams are fully converted to Visual C++, they will
BLADE SETTING be more easily sustained and user-friendly for future
users. The desired end product will be capable of
steady-state ventilation simulation for “ideal” mine sit-
AXIAL SPEED uations, as well as prediction for fires and/or thermal
events with products-of-combustion and contaminants
calculations in a dynamic simulation.
The ability for a mine to possess this modeling
CALCULATION capacity will prove invaluable as proactive planning
and emergency training programs are improved by use
METHODS: SERIES of this prediction software. Utilization of this type of
& PARALLEL software may better forecast and strategize for volatile
emergency situations. It is the hope that this new pro-
gram combination will inevitably be able to assist mine
Figure 6. Illustration of the Fan Class. planners and ventilation experts in their daily routines
as well as in emergency situations.
CONTAMINATION Class
REFERENCES
Methane Volume Laage, L.W., Greuer, R.E. & Pomroy, W.H. 1995. MFIRE
Users Manual. U.S. Bureau of Mines Training Workshop
on the “MFIRE” Mine Fire and Ventilation Simulator.
Contaminant Inflow Minneapolis, MN, Twin Cities Research Center, U.S.
Bureau of Mines.
Edwards, J.C. & Greuer, R.E. 1982. Real-Time Combus-
Heat Entering tion of Product-of-Combustion Spread in a Multilevel
Mine. Information Circular, United States Department
of the Interior, Bureau of Mines, 8901. Pittsburgh, PA:
Pittsburgh Research Center, U.S. Bureau of Mines.
Oxygen Concentration Hartman, H.L., Mutmansky, J.M., Ramani, R.V. & Wang, Y.J.
1997. Mine Ventilation and Air Conditioning, 3rd ed: 152–
153. New York: John Wiley & Sons, Inc.
Fume Production Ohio Automation. 2003. Solutions for Electronic CAD, Mine
Planning, and Terrain Modeling – Engineering Software
since 1985. Ohio Automation Inc. 24 September 2005 at
Heat Production http://www.ohioautomation.com/index.html
Rumbaugh, J., Blaha, M., Premerlani, W., Eddy, F. &
Lorensen, W. 1991. Object-Oriented Modeling and
Figure 7. Illustration of Concentration class. Design. New Jersey, Prentice Hall.
Deitel, H.M. & Deitel, P.J. 2003. C++ How to Program, 4th
the information pertaining to the rock data, such as the ed. New Jersey, Prentice Hall.
average rock temperature in the airway, the thermal Metcalf, M. & Reid, J. 1999. FORTRAN 90/95 Explained,
diffusivity of the rock, and the thermal conductivity of ed.2. New York, Oxford University Press.
the airway rock walls.

482

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Fire-generated smoke rollback through crosscut from return to

intake – experimental and CFD study

G.F. Friel, L. Yuan, J.C. Edwards, & R.A. Franks

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

ABSTRACT: Two mine-fire experiments were conducted in the National Institute for Occupational Safety
and Health’s (NIOSH) Safety Research Coal Mine (SRCM) at the Pittsburgh Research Laboratory (PRL) which
demonstrated that smoke from diesel-fuel fires of 500 kW and 660 kW heat-release rates in a return airway
can develop, without causing a complete air flow reversal, into a roof layer that can migrate upwind forming
a counter flow to the primary airflow in a crosscut. Subsequently, smoke can penetrate into an intake airway
and create a hazardous atmosphere in the intake airway upwind from the fire. Visibility conditions less than
13 m were created by the smoke in the intake airway downwind from the crosscut. Modeling of the event with
a three-dimensional, time-dependent, computational fluid dynamics (CFD) program correctly represented the
smoke movement.

Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent
the views of the National Institute for Occupational Safety and Health.

1 INTRODUCTION once the fire has evolved to sufficient intensity, the

Smoke rollback in a mine entry from a fire can pose
hazardous consequences for mine evacuation, miner
rescue, or firefighting. Inhalation of fire-generated
and toxic products-of-combustion (POC) can be inju-
rious or fatal to miners. The low visibility associated
with smoke impedes safe miner escape and rescue.
The greatest hazard of smoke rollback is in fighting
fires. Smoke rollback can occur directly upwind in
the entry with the fire, or if it penetrates a crosscut,
it can be transported downwind in the adjoining air-
way. If a mine rescue team or firefighting brigade
should approach a fire in the fresh air of an intake
airway before the fire intensity becomes sufficient to
push smoke through a breached crosscut into the intake
entry upwind of the mine rescue team or firefighting
brigade, these miners could then become enshrouded
in low visibility and toxic POC.
Initially the thermal buoyancy forces generated by a
fire in a mine entry will produce an ascending plume
of fire smoke. In the absence of ventilation, the hot
smoke will impinge upon the roof and migrate radially
outward. Any entry ventilation will bend the ascend-
ing smoke plume. Initially, the dominant transport of
the smoke at the roof will be in the direction of the
ventilation. The smoke will be transported downwind
from the fire along the roof until the smoke is cool
and is mixed over the entry cross-section. However,
buoyancy forces associated with the temperature of
the smoke plume will overcome the inertial forces of
the ventilation, and some of the smoke will migrate
upwind along the roof counter to the existing venti-
lation direction. Another hazard of smoke rollback,
noted by Mitchell (1996), is that the smoke can con-
tain combustible gases in a low-oxygen mixture which
are subject to ignition by the mine fire when diluted
with fresh air.
Previous research (Edwards et al. 2005, Hwang &
Edwards 2005) has focused upon smoke rollback in a
single entry and the air velocity required to prevent the
smoke rollback. For a sufficiently intense fire relative
to the airflow, it is possible to develop complete airflow
reversal in an airway that can be modeled accurately by
a network mine-fire simulator such as MFIRE (Chang
et al. 1990).
Smoke rollback in a mine entry can be complicated
by the connectivity of the entry to other mine entries
by open crosscuts and leakage paths. The smoke layer
which accumulates near the roof can migrate into par-
allel airways depending upon the air velocity and the
amount of leakage between the airways. Permanent
stoppings are used to separate intake and return air-
ways in a coal mine except for the first two connecting
crosscuts outby the working face. A check curtain may
be used at the second crosscut outby the working face
to provide for vehicular traffic movement. Any breach

483

Copyright © 2006 Taylor & Francis Group plc, London, UK

 of the stoppings due to an open man door or stopping
destruction by a fire or explosion, or significant leak-
age around a stopping or check curtain, creates the
potential for smoke flow from a fire in a return airway
into an intake airway.
The objective of this work was to demonstrate with
in-mine fire experiments the potential for smoke roll-
back from a fire in a return airway into an intake airway
and to model this experimental work with CFD simula-
tion. CFD provides the opportunity to expand the range
of prediction of smoke spread for a wide spectrum of
airflows, fire intensities, and mine-section inclinations
and layouts.

Figure 1. View of the mine section in the NIOSH SRCM

2 EXPERIMENTAL CONFIGURATION (Not drawn to scale).

Experiments were conducted in parallel entries in the

SRCM shown in Figure 1.
In the crosscut upwind of the fire, Room 11,
The average entry height and width at the location of
and 4.8 m from B-Butt, a one-meter-long, optical,
the fire in B-Butt were 1.81 and 2.79 m, respectively.
smoke-obscuration meter (light meter) was suspended
Diesel fuel was selected as the combustible material
horizontally 30 cm from the roof at S3. A thermocou-
for the fires because of its relatively uniform burn-
ple was attached 10 cm from the roof above this light
ing rate from containers. The three containers that
meter. The upwind edge of the crosscut downwind
were used were horizontally square steel pans with
from the fire zone, Room 13, was 25 m from the fire
sides of 46 cm, 61 cm, and 76 cm long. The heights
zone. Within Room 13, 5.2 m from B-Butt, and at S4
of the pan sides were about 13 cm. The fire zone was
was another light meter 30 cm from the roof. A ther-
protected from thermal radiation by lining the roof
mocouple was also placed 10 cm from the roof above
and ribs with a ceramic fiber insulation. The smoke
this light meter.
concentration in B-Butt upwind of the fire was mea-
In A-Butt, 2 m downwind from Room 13, and at S5
sured by an ionization detector at sensor station, S1,
was placed the first of two posts to hold gas-velocity
near the middle of B-Butt with its inlet about 20 cm
sensor units and optical, smoke sensor units similar to
from the roof and its position at 15.5 m from the
the configuration in B-Butt. The spacing between the
upwind edge of the fire zone and 4.2 m from the
gas-velocity units was 8.9 m and the spacing between
upwind edge of Room 11 (See Figure 1). A thermo-
the smoke sensor units was 9.4 m.
couple was also placed near the ionization detector
In A-Butt, at 25 m from the downwind edge of
and about 8 cm from the roof. A one-meter-long, laser,
Room 11, and at S6 were placed an ionization smoke
smoke-obscuration meter (laser-light meter) was also
sensor and a laser-light meter near the middle of A-
positioned near the ionization detector and 30 cm from
Butt to detect smoke emanating from either or both
the roof. Along the rib of B-Butt at S2 and 12.5 m
of Rooms 11 and 13. The laser-light meter was 30 cm
upwind from the ionization detector was the first of
from the roof and the inlet to the ionization smoke
two steel mine posts which supported a pair of path-
sensor was about 20 cm from the roof.
averaging sensors. The second post was 21.8 m from
the ionization detector on the opposite rib of B-Butt.
An infrared smoke sensor, composed of a transmitter
and a receiver, was attached to the posts about 60 cm 3 EXPERIMENTAL PROCEDURE
from the roof with the infrared beam directed horizon-
tally and diagonally 10.2 m across B-Butt. Two units The air velocities at the six positions on Figure 1
of a gas-velocity sonic sensor also were attached to were measured using a vane anemometer. At positions
these two posts. One unit was attached 20 cm from the 1, 3, 5, and 6 the average velocity was measured at
roof on the first post and the other unit was attached each of five points in the cross section of the air-
30 cm from the floor on the second post. The length way. At positions 2 and 4 the average velocity was
of the path between these units was 9.7 m. The gas- measured at each of 3 points in the cross section of
velocity sensor units were attached on the sides of the the doorways next to A-Butt. These velocity averages
posts closest to each other while the smoke sensor units were determined by measuring the velocity over at
were attached on the sides farthest from each other. least one cycle length of atmospheric changes to mine

484

Copyright © 2006 Taylor & Francis Group plc, London, UK

 air velocities which usually lasted about 30 s. Cross- Table 1. Measured resistances through crosscuts.
sectional areas were calculated from measurements
of height and width using an electronic, ultrasonic Static Relative
distance-measuring tool. Room Resistance, Pressure, Temperature, Humidity,
2 8 ◦
The fire pans were placed level, next to each other, Number Ns /m cm Hg C %
and on the floor near the middle of B-Butt in the
11 0.1709 73.56 2.39 89.8
fire zone. Attached to a laboratory stand next to the 13 0.1407 73.53 1.78 92.6
pans was a thermocouple positioned horizontally and
about 10 cm above the middle of the downwind pan
to indicate the end of combustion. The diesel fuel
was poured into a measuring container before being Table 2. Airflow rates at locations in Figure 1.
poured into each pan. Before ignition of the diesel
fuel using a propane torch, about 250 mL of denatured, Experiment 1 Experiment 2
95% ethanol was poured into each pan. The cold diesel Location on (500 kW Fire) (660 kW Fire)
fuel was difficult to ignite without the ethanol. After Figure 1 Airflow Rate, m3 /s Airflow Rate, m3 /s
ignition, exiting of personnel from B-Butt was through
Room 11 into A-Butt and then along A-Butt to a point 1 4.2 4.5
upwind of Room 13. A digital camcorder was attached 2 1.8 1.6
to a tripod about 1.5 m from the floor of A-Butt and 3 3.2 3.5
opposite Room 13 to record the smoke emanating 4 1.7 1.6
5 2.2 2.4
from Room 13 and also Room 11 by pivoting the
6 1.7 1.8
camcorder 90◦ . 1, During 3.5 3.7
Experiments

4 MEASUREMENT OF ROOM RESISTANCES

The static pressure drops between A-Butt and B-Butt
through Room 11 and Room 13 were measured using
a manometer placed level on a pedestal within the
rooms with tubes extended into A and B-Butts. The
airflow rates through the rooms were calculated from
the average of three velocity measurements at different
heights within doorways in each room using the vane
anemometer and the dimensions of the doorways. The
static pressures were measured using an altimeter. The
temperatures and relative humidities were measured
using a digital psychrometer. The resistances were cal-
culated by dividing the air static pressure drop by the
square of the volumetric airflow rate. The results of
those measurements at mine static air pressures and
temperatures are listed below in Table 1. The resis-
tances can be converted to a standard set of conditions
for use in simulators such as MFIRE.
5 EXPERIMENTAL RESULTS
Table 2 shows the volumetric airflow rates in the mine
network prior to each experiment. Also shown are
the air velocity measurements at position 1 that were
repeated after each experiment because of changes
to the airflow rate into A-Butt during each exper-
iment. These velocity changes were made after a
semi-equilibrium position of smoke had occurred in
Room 13 and to facilitate penetration of smoke into
A-Butt before the diesel fuel was all consumed. These
ventilation decreases were accomplished by opening a
door in a crosscut fromA-Butt to the return and upwind
of the entrance to Room 13. The airflow rates for the
two experiments were similar, although the inaccura-
cies in the volumetric balances could have been caused
by the vortices produced by the edges of nearby rooms
and leakage around temporary brattices.
In the first experiment, 23.7 L of diesel fuel was
burned in the 46 cm and 61 cm square pans. Based
upon the fire duration and the heat of combustion
for diesel fuel, the fire intensity was estimated to be
500 kW. A large quantity of smoke passed through
Room 11 upwind of the fire and from there downwind
through A-Butt and back through B-Butt and the fire
zone. This fire yielded no penetration of smoke from
B-Butt through Room 13 into A-Butt until a reduc-
tion in airflow rate through A-Butt was made. Figure
2 shows the light meter response in millivolts (mV),
which decreased with an increase in the concentra-
tion of smoke at the light meter, and the temperature
at the thermocouple within Room 13 during the first
experiment. The POC’s penetrated to the doorway next
to A-Butt. All sensors were sampled every two sec-
onds. Figure 3 shows the light meter and thermocouple
response within Room 11 during the first experiment.
Figure 4 shows the temperature and indicates smoke
intensity at the point ionization smoke sensor in B-
Butt (S1), and indicates the smoke intensity at the
point ionization smoke sensor in A-Butt (S6) located
25 m downwind from Room 11 during the first exper-
iment. About a minute was required for the smoke to
backup 12 m.

485

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 2. Temperature and light meter response in Room 13
for experiment no. 1.
Figure 3. Shows the light meter and thermocouple response
within Room 11 during the first experiment.
Figure 4. Sensor responses in A-Butt and B-Butt for exper-
iment no. 1.
In the second experiment, 33.1 L of diesel fuel were
burned in the 46 cm and 76 cm square pans. The esti-
mated fire intensity was 660 kW. This fire yielded a
brief penetration of smoke from B-Butt through Room
13 into A-Butt. A large quantity of smoke also passed
through Room 11 upwind of the fire and from there
downwind through A-Butt. The quantity of smoke
from Room 13 was much less than the quantity of
smoke emanating from Room 11 into A-Butt. Figure
5 shows the light-meter response to the smoke and the
temperature of the smoke at the thermocouple within
Room 13 during the second experiment. These POC’s
penetrated from Room 13 into A-Butt continuously
during the fire after the flow rate through A-Butt was
reduced. The greater smoke rollback into Room 11
486
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 5. Temperature and light meter response in Room 13
for experiment no. 2.
Figure 6. Temperature and light meter response in Room 11
for experiment no. 2.
was associated with the maximum roof temperature of
78◦ C in Room 11 compared with the maximum roof
layer temperature of 42◦ C in Room 13. For the first
experiment, the maximum roof-layer temperatures in
Rooms 11 and 13 were 59◦ C and 29◦ C, respectively.
Figure 6 shows the light-meter response to the
smoke and thermocouple measurement of the smoke
temperature within Room 11 during the second exper-
iment.
Figure 7 shows the temperature and indicates smoke
intensity at the point, ionization smoke sensor in B-
Butt and indicates the smoke intensity at the point,
ionization smoke sensor in A-Butt located 25 m down-
wind from Room 11 during the second experiment.
The gas temperature may have caused the tempo-
rary failure of the thermocouple’s electronic interface
which also was supported near the roof.
6 VISIBILITY
As the light-meter responses in Figures 2, 3, 5, and 6
show, smoke obscuration is total near the roof within
the crosscuts. Smoke movement through Room 11
from the return, B-Butt, into the intake, A-Butt, cre-
ated a reduced visibility downwind from Room 11 in
A-Butt. The visibility was determined from Jin’s rela-
tionship (Jin 1977), which, in turn, was based upon
 Figure 7. Sensor response in A-Butt and B-Butt for experi-
ment no. 2.
Figure 8. Smoke optical density and visibility in A-Butt
25 m downwind from Room 11 during experiment no. 1.
the smoke optical density determined from the reflec-
tion of light from a reflecting sign – as might occur
when a miner’s cap lamp is focused upon a reflect-
ing mine-escape marker. Figure 8 shows the optical
density (OD) at the laser light monitor located 25 m
downwind from Room 11 and the associated visibility
for the 500 kW fire of the first experiment. The opti-
cal density is defined from the light transmission at
the laser light monitor by equation 1:
T is the light transmission between the transmitter and
the receiver in smoke-laden air, T0 is the light trans-
mission in clear air, and d is the distance between the
transmitter and the receiver.
The minimum visibility was 9.4 m. Also shown are
the 4 m and 13 m visibility requirements reported by
Jin (1981) for minimum visibility of someone familiar
and not familiar with the surroundings, respectively.
For the 660 kW fire of the second experiment, the min-
imum visibility was 5.5 m. Someone unfamiliar with
the SRCM would not have the minimum visibility for
escape from the fire zone through B and A-Butts. A
scenario could exist of a firefighting brigade enter-
ing B-Butt through A-Butt upwind of the fire zone to
extinguish the fire. Initially the fire intensity might not
be adequate to force smoke rollback through Room 11
into A-Butt. As the fire intensity increases, the smoke
487
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 9. Smoke density distribution near the roof at 600 s
for the 500 kW fire.
rollback could not be confined to B-Butt but would
penetrate through Room 11 and soon be upwind of the
firefighting brigade.
7 COMPUTATIONAL STUDY
Smoke reversal through crosscuts from a return air-
way into an intake airway was simulated using the
Fire Dynamics Simulator (FDS), a CFD program
developed by NIST (McGrattan et al. 2002). FDS
is a three-dimensional, large-eddy simulation model
developed for studying the transport of smoke and hot
gases during a fire in an enclosure. It is a widely used
large-eddy simulation model in the field of fire science
and has demonstrated good agreement with experi-
mental data in numerous validation studies. The model
uses finite difference techniques to estimate solutions
of the Navier-Stokes equations for fluid flow with a
mixture-fraction combustion model.
The physical model for the numerical simulation
consists of the intake airway, the return airway, two
crosscuts, and an airway connecting the intake and the
return airways – as shown in Figure 1. In Figure 9
below the connecting airway was not drawn.
Because of the memory limitations and speed of the
computer running FDS, Room 11 was made perpen-
dicular to both intake and return airways to increase
the mesh size of the simulation. The two crosscuts each
have two doors and a plenum between the doors. All
dimensions for the airways and crosscuts were entered
from a mine-section survey. Each diesel fire was mod-
eled using its average heat-release rate estimated in the
experiment. The airflow velocity measured upwind of
Room 13 in A-Butt was entered as the initial flow con-
dition. In both experiments, the airflow velocity was
reduced after about 15 minutes from the ignition of the
diesel fuel. In the simulation, the smoke movement
became quite stable after 10 minutes, so the airflow
 Figure 10. Smoke density distribution near the roof at 670 s
for the 500 kW fire.
velocity change was made after 10 minutes. For the
500 kW fire, the airflow velocity was 0.96 m/s at the
beginning and was changed to 0.79 m/s after 10 min-
utes. For the 660 kW fire, the airflow velocity was
0.99 m/s at the beginning and was changed to 0.80 m/s
after 10 minutes.
A cold airflow simulation was conducted prior to
the fuel ignition. The airflow rates in A-Butt and the
two crosscuts obtained from the simulation agreed well
with the rates calculated from the pre-experiment ven-
tilation survey. However, the measured airflow rate in
B-Butt was smaller than the simulated rate because of
leakage around brattices along one of the ribs in B-
Butt into other crosscuts and the rest of the SRCM.
Simulations were then conducted for the experimental
fires. Figure 9 shows the smoke (soot) density dis-
tribution near the roof at 600 s for the 500 kW fire,
immediately before the change of the intake air veloc-
ity. Smoke passed through the open door of Room 11
adjacent to B-Butt and filled the Room-11 plenum,
but could not pass through the open door adjacent to
A-Butt. Smoke also passed through the open door of
Room 13 adjacent to B-Butt and only partially filled
the Room-13 plenum. The smoke inside Room 13 was
also less dense than the smoke in Room 11.
Figure 10 shows the smoke density distribution near
the roof at 670 s for the 500 kW fire 70 s after the
reduction of intake air velocity.
Smoke passed through the open door of Room 11
adjacent to A-Butt and leaked into A-Butt but could
not roll upwind in A-Butt. Smoke filled the entire
plenum in Room 13 but could not pass through the
open door adjacent to A-Butt. As time increased, more
smoke leaked into A-Butt from Room 11 and slightly
rolled upwind next to the rib closest to Room 11. A
small quantity of smoke passed into A-Butt through
the open door of Room 13 adjacent to A-Butt. These
results are consistent with the observations in the first
experiment.
488
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 11. Smoke density distribution near the roof at 600 s
for the 660 kW fire.
Figure 12. Smoke density distribution near the roof at 660 s
for the 660 kW fire.
Figure 11 shows the smoke density distribution near
the roof at 600 s for the 660 kW fire and immediately
before the change of the inlet velocity.
Some smoke already passed intoA-Butt from Room
11 through the open door adjacent to A-Butt but was
quickly diluted and could not roll upwind in A-Butt.
Smoke passed through the open door of Room 13
adjacent to B-Butt and filled most of the plenum.
Figure 12 shows the smoke density distribution near
the roof at 660 s for the 660 kW fire 60 s after the
ventilation reduction.
Dense smoke leaked into A-Butt from Room 11
and rolled upwind in A-Butt. Smoke filled the entire
plenum of Room 13 and less dense smoke leaked
into A-Butt. As time increased, more smoke leaked
into A-Butt from Room 11 and rolled upwind against
the intake airflow, while no more smoke leaked into
A-Butt from Room 13. These results also are qualita-
tively consistent with the observations in the second
experiment.
It can be seen from the FDS simulation that the
smoke leakage from Room 11 into the A-Butt only
occurred near the roof. Close to the floor in Room
 11 the ambient air still moved in against the smoke
leakage. This is one advantage of the FDS’s three-
dimensional simulation. In one-dimensional model-
ing, one airway is either occupied by clear air or an
air-smoke mixture but cannot be both. In a real situa-
tion partial smoke leakage into an airway could occur
and poses a threat to the safety of underground miners
and mine rescue personnel.
8 CONCLUSIONS
The following conclusions were determined from the
experimental results and the CFD simulations.
• The experimental mine fires in a return airway
produced sufficient buoyancy to establish a smoke-
laden roof layer that flowed through connecting
crosscuts counter to the direction of fresh air from
the intake entry.
• The density of the smoke that leaked into the intake
was shown to yield insufficient visibility downwind
from the last connecting crosscut in the intake entry
for someone unfamiliar with the mine to find their
way out easily.
• The CFD simulations showed good agreement with
the experimental observations of smoke movement.
Smoke rollback along the roof from a fire coun-
tercurrent to the cooler airflow near the floor can be
a mechanism for smoke to move from a mine return
into a mine intake in low airflow sections. This move-
ment of smoke can impede firefighting efforts. The
489
Copyright © 2006 Taylor & Francis Group plc, London, UK
realization of this possibility would not be predicted
from a mine-network ventilation program which is
based only upon unidirectional flow. CFD modeling
is a viable method for analyzing potential visibility
hazards associated with smoke from an in-mine fire.
REFERENCES
Chang, X., Laage, L.W. & Greuer, R.E. 1990. A user’s man-
ual for MFIRE: a computer simulation program for mine
ventilation and fire modeling. U.S. Bureau of Mines IC
9245, 171 pp.
Edwards, J.C., Franks, R.A., Friel, G.F. & Yuan, L.L. 2005.
Experimental and modeling investigation of the effect of
ventilation on smoke rollback in a mine entry. SMEAnnual
Meeting, Feb 28–Mar 2, 2005, Salt Lake City, UT, paper
05-14, 6 pp.
Hwang, C.C. & Edwards, J.C. 2005. The critical ventilation in
tunnel fires – a computer simulation. Fire Safety Journal
40: 213–240.
Jin, T. 1977. Visibility through fire smoke. Journal of Fire
and Flammability 9: 135–155.
Jin, T. 1981. Studies of emotional instability in smoke from
fires. Journal of Fire and Flammability 12: 131–142.
McGrattan, K.B., Forney, G.P., Prasad, K., Floyd, J.E. &
Hostikka, S. 2002. Fire Dynamics Simulator (Version 3)
– User’s Guide. U.S. Department of Commerce, National
Institute of Standards and Technology.
Mitchell, D.W. 1996. Mine fires prevention, detection, fight-
ing: 19-202. Chicago, IL: Intertec Publishing Inc.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Case study of CFD simulation in mining accident investigation

E. Delic, A. Basic & R. Sisic

Faculty of Mining, Geology and Civil Engineering, University of Tuzla, Bosnia and Herzegovina

ABSTRACT: Hazard event precautions, explosion ignition and fire progression are essential facts for mining
fire and explosion accident investigations. Identification of causes in post accidental analysis is based on mine
site investigation, available documentation and interviews. Tests of proposed hypothesis in traditional investi-
gation methods are based on expert interpretation and utilization of conventional engineering calculations with
limited numerical simulation capabilities.
Finite elements based “computational fluid dynamic” (CFD) calculations in investigation of mining accidents
are comprehensive method for identification of accident critical conditions and consequences. It may be essential
key in estimation of pressure and thermal variations in each point of domain volume in case the critical amount
of reliable data for simulation are available. CFD analysis, based on finite volumes method, shows potential to
become one of essential investigation technique in validation of hazards event hypothesis and extended report
verification.
Paper is based on real case-study based on utilization of “Fluent” CFD software package as an additional tool
for investigating major accident in Bosnia and Herzegovina mining history that took place in underground coal
mine “Dobrnja” in 1990, on which occasion 180 employees got killed.
It was clear from the very beginning how series of coal dust explosions took place in mine areas, but difficult
to find and confirm how the first explosion occurred and what was the main ignition source. Combining with
conventional engineering approach and available data, simulation model for CFD is generated and ignition
process is simulated according to circumstances in underground workings at that time. Findings of CFD
simulation gave necessary link between evidences, conventional calculation methods and hypothesis.

1 INTRODUCTION these differences are caused by alterations on fans or

Accident in underground lignite mine Dobrnja (1990)
is the most tragic accident ever in mining his-
tory of Bosnia and Herzegovina (BH) and former
Yugoslavia. Lignite mine Dobrnja is located in north-
east Bosnia, in vicinity of Tuzla city. Coal is deposited
in 4 major layers, inclined from 22 to 38 degrees.
Dobrnja mine consisted of 2 underground mining
operations called “1st Roof Layer” and “2nd Roof
Layer”. Both underground operations had own main
fans and intakes and were connected by 2 underground
roadways. Mine operation “1st Roof Layer” was con-
sidered methane-free, whereas operation “2nd Roof
Layer” was categorized as potentially dangerous from
methane.
Main fans in both operations were of the same
type (4AV-1600). Declared working points are: engine
power 90 kW, airflow 36 m3 /s, and pressure 1177 Pa.
Ventilation survey and ventilation design indicate that
effective airflow was close to declared, but effective
pressure was 844 and 822 Pa, what differs significantly
from declared. It was not possible to determine if
measuring technique applied.
Having effective orifice 1.48 and 1.43 m2 at main
fans, both mines were tight and had enormous ventila-
tion resistance, regardless of the fact that legal obliga-
tions were fulfilled. Regulations of former Yugoslavia
stipulate a minimal allowed orifice of 1 m2 .
Lignite was mined applying a method which was
combination of “short wall” and “room and pillar”
methods, using blast operations to remove coal.
Drilling, blasting and loading were completely man-
ual. Very low level of mechanization is reason why
large number of workers (more than 180) were engaged
to run process of extraction and transportation. Work-
ings were ventilated by auxiliary fans.
Despite the fact that several small explosions with
participation of coal dust had previously been regis-
tered, general consideration before accident was how
lignite dust was not prone to combustion. In morning
hours on 26th of August 1990 several explosions of
coal dust killed 180 workers, with only one survivor
who was at that moment located close to main mine
opening (decline).

491

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 1. Simulation domain divided in III zones.
2 ACCIDENT INVESTIGATIONS
First accident investigations were conducted immedi-
ately upon the disaster. Due to complexity of accident
investigation and break out of war in formerYugoslavia
(1992–1996) investigations were not completed to the
present time. In the year 2005 investigations restarted
with aim to get final report according to prosecutor’s
request.
Initial hypothesis during accident pre-cautions
reconstruction was that at one mine working explo-
sive was used to remove steel support that was not in
compliance with minimal safety requirements. Con-
centration and characteristics of deposited coal dust
in the working areas enabled the coal dust to ignite
and spread fire in direction opposite to fresh airflow.
According to evidence collected at the mine site it
was clear how coal dust did not explode on the very
beginning. Initial area of a few hundred meters was
affected by fire (deflagration). Taking into account
the fact that explosion of coal dust will occur only if
dust is swirled from surfaces in oxygen sufficient area
before initial ignition and that pressure in area must
exceed 0.15 MPa, it was unexpected that mine fire
would turned to dust explosion in actual working area.
During the accident investigation conducted in
1990, propagation conditions were calculated based
on determined quantities of burned coal dust and fresh
air intake. Generated heat was used to calculated pres-
sure increase along the critical path. Calculation con-
firmed possibility of pressure increase to critical level
for first explosion, according to available evidence.
3 CFD SIMULATION OF IGNITION PROCESS
In order to test initial hypothesis on accident root
causes and to compare results with conventional
investigation from year 1990, simulation model for
finite volume CFD simulation was created. Simulation
domain was defined and model was developed using
mining maps, technical documentation and accident
investigation documentation. Properties of deposited
coal dust and explosive characteristics were deter-
mined from chemical analysis and explosion tests in
experimental mine “Barbara” (Katowice, Poland).
492
Copyright © 2006 Taylor & Francis Group plc, London, UK
Main concern was to confirm possibility to mine
fire transit to dust explosion in specific situation.
Domain size and boundaries were adjusted to include
all area from first initial fire to first explosion, which
occurred some 150 meters after explosive ignition.
Simulation model was created using GAMBIT
mash generator. For CFD a simulation a Fluent soft-
ware package was used. Premixed combustion model
was selected, where coal dust and oxidant were pre-
viously mixed in domain based on calculations of
deposited dust quantity and portion of dust which
participated in ignition.
By hypothesis, explosive used for breaking steel
support ignited and swirled coal dust from sur-
faces. Dust continued to burn opposite to fresh air
intake, gradually creating conditions for transition
from deflagration to detonation. Problem how to simu-
late swirling of coal dust in front of fire and feeding of
fire was solved using method “cold flow-hot flow”.
First was simulated the expected air and dust flow
through workings without ignition.Then, in area where
explosive was used to break steel support, high surface
temperature was produced sufficient to ignite fire.
4 MODEL GEOMETRY AND BOUNDARY
CONDITIONS
CFD simulation of coal dust fire process up to the first
detonation was divided into 3 zones and simulated in
3 phases.
Phase I is in area where miners used explosive to
break steel support up to the first intersection. Phase
II is extended from the first to the following intersec-
tion, while phase III include vicinal workings which
may interfere with fire process and first detonation
alongside the transportation line.
Boundary conditions were represented by initial
conditions in a manner that final result of a phase
represents initial condition for the following one.
Physical and chemical properties of coal dust in
simulation model are:
– Coal dust thermal value 23 959 kJ/kg
– Specific thermal capacity cp = 1000 J/kgK
– Thermal conductivity λ = 0.0454 W/mK
– Viscosity ν = 1.72 × 10−5 kg/ms
Air and dust mixture in fire process is considered
as ideal gas. Because of high temperature and pressure
error from this assumption is insignificant.
Taking into account the fact that initial ignition
occurred in auxiliary ventilated “dead end” working,
calculation domain was developed as a pipe system,
completely closed on one side (head face in working
area) and open towards air inlets and outlets. Ade-
quate equivalent orifice was implemented on open
ends of pipes, thus simulating influence of remaining
ventilation system to process.
 Table 1. Maximal flame and detonation velocity was registered
in case when one end of pipe is sealed and other end
Boundary opened. From point of sealing type, auxiliary ven-
Phase type Zone type Parameters
tilated “dead end” mining workings are exposed to
I Inlet dust inlet (→) v = 1 m/s maximal hazard from rapid increase of flame velocity
boundary dust concentration and its transition to detonation.
30 g/m3
T = 325 K 5 SIMULATION RESULTS
fresh air inlet Q = 2,08 m3 /s
from booster T = 298 K Coal dust swirling and burning resulted from inap-
ventilator (→) propriate usage of explosives for crashing steel sup-
Outlet domain exit – on Aeq = 0,17 m2 port at a mine working, level “XXI” in lignite mine
boundary 40 m from p = 100 “Dobrnja”. From head face flame propagation speed
working site – 601,32 Pa was around 2.5 m/s all the way up to the first inter-
point 56 (→) section. From the first intersection flame velocity
II Inlet dust inlet (→) v = 1 m/s increased and after 21 s conditions for first explosion
boundary dust concentration were reached. During that time approximately 3,5 kg
30 g/m3 of coal dust were burned, that was sufficient to gener-
T = 325 K ate critical pressure of 0.15 MPa after just 130 to 180
fresh air inlet Q = 2,08 m3 /s meters from location where explosive was used.
from booster T = 298 K Simulation based on finite element method was
ventilator (→) compared to conventional heat-pressure calculations,
Outlet domain exit – Aeq = 0,44 m2 which was calculated 15 years ago, (see table 2).
boundary on 115 m from p = 100 In the first phase from head-face up to the first
working site – 600,97 Pa intersection, simulated pressure and temperature gra-
point 56 (→) dient was in accordance to conventional calculated
III Inlet dust inlet (→) v = 1 m/s values. On next section T42–T43 (from the first to
boundary dust concentration the second intersection) CFD simulated pressure was
30 g/m3 5% lower, compared to conventional calculation, but
T = 325 K gas temperature was significant lower (40%).
fresh air inlet Aeq = 2,25 m2 In the third section (T43–T49) temperatures were in
from GTH (main p = 101 325 Pa similar range for both kinds of calculations. Pressure
transportation road) T = 295 K differences at end of domain significantly differed due
(→) – point 17 to different size of domain and different “end point”
Outlet exit from domain Aeq = 0,57 m2 in simulation. As for CFD simulation, “combustion”
boundary in UR-1, point 30 p = 101 325 Pa model was used, and for the final point was selected a
exit from domain in Aeq = 0,67 m2 location where pressure increased above critical level
TN-1 (transportation p = 100 600,7 Pa for explosion. In conventional heat-pressure calcula-
swage) – point 43 tion in 1990 end point was the last intersection. First
explosion occurred before second intersection (T49).
Conventional heat-pressure calculation did not con-
sider flame velocity and position of transition from
deflagration to detonation. Aim was to prove possibil-
ity to coal dust be swirled, burned, and to explode in
conditions according to hypothesis.
First detonation produced shock wave which gener-
ated high pressure front, which progressed with sonic
speed and was able to lift and burn coal dust. Each
explosion was stopped when all available oxygen was
burned, but after explosion flame propagated opposite
to fresh airflow decreasing speed to subsonic values.
Figure 2. Boundary conditions. High flame speed prevented possibility of all available
oxygen participating in fire, thus gradual increases of
Bartknecht (6) deduced based on flame propagation oxygen concentration in direction opposite to fresh
experiment in pipe 40 m length and 1.4 m diameter, airflow was caused. Following explosion arose when
that type of pipe sealing had significant influence oxygen concentration increased to the level sufficient
on flame propagation and transition to detonation. for transition to sonic velocity. The final explosion

493

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 2. Results of CFD simulation in year 2005 compared to conventional thermodynamic
calculation from accident investigation in 1990.

Conventional calculation CFD simulation

Pressure Temperature Pressure Temperature

Location (MPa) (◦ C) (MPa) (◦ C)

Headface-42 0,116–0,119 42,14–52,67 0,117 44–49

T42–T43 0,144–0,158 129–171 0,137 77
T43–T44 0,197–0,227 317–407 0,129 137
T43–T49 0,203–0,237 342–443 −0,14 to 1,66 412

Figure 3. Static pressure contours in mining workings at the level XXI – zone I.

Figure 4. Static pressure contours in mining workings at the level XXI – zone II.

494

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 5. Static pressure contours in mining workings at the level XXI – zone III.
occurred just several meters from surface, in main air
intake.
Unfortunately transportation and communication
lines were positioned in main fresh air intakes,
therefore all miners were close to the most affected
area, and only one of them survived this tragical
accident.
Differences in pressure and temperature in both
calculations are results of different methodology and
approach. In both cases, scenario from hypothesis and
evidence was confirmed: from the initial explosive
ignition to the first coal dust explosion. Following
explosions were not requested to be simulated, because
of complex conditions and minor significance for
required objective.
6 CONCLUSION
Computing simulation and analysis of fire propagation
by initial hypothesis based on evidence, proved pos-
sibility that inappropriate usage of explosives ignited
coal dust, swirled it from surfaces and generated fire
close to road head face, that can produce conditions
for transition from deflagration to detonation after only
495
Copyright © 2006 Taylor & Francis Group plc, London, UK
130 to 180 meters in conditions existing in coal mine
“Dobrnja” when accident took place in 1990.
Whole accident, from the first fire occurrence to
the last explosion, was difficult to simulate, and it was
not required.
Presented case confirms capability of CFD sim-
ulation for successful investigation of mining acci-
dents. Critical phase in whole process is creating of
appropriate model and selecting of proper boundary
conditions.
Once model is developed, numerical calculations
can be used for testing of different hypothesis. Experi-
ences from real mining accidents investigation are also
good base for risk assessment for expected hazards and
hazardous situations.
REFERENCES
Bartknecht, W. 1988. Dust Explosions: Course, Prevention
and Protection. Springer-Verlag.
Bjørkhaug, M. and Hjertager, B.H. 1984. The influence of
obstacles on flame propagation and pressure develop-
ment in a radial vessel of ten metre radius. Final Report.
CMI Report No. 843403-9, Chr. Michelsen Institute,
Bergen.
 Bjørn Johan Arntzen 1998. Modelling of Turbulence
and Combustion for Simulation of Gas Explosions in
Complex Geometries», Thesis for The Dr. Ing. Degree,
The Norwegian University Of Science And Technology.
FLUENT software documentation, 2004. USA: Fluent Inc.
Gardner, B.R. Winter, R.J. and Moore, M.J. 1986. Explosion
Development and Deflagration-to-Detonation Transition
in Coal Dust/Air Suspensions. Twenty-first Symposium
(International) on Combustion. The Combustion Institute.
Pineau, J.P. and Ronchail, G. Propagation of Coal Dust Explo-
sions in Pipes. Industrial Dust Explosions. ASTM STP
958. Cashdollar & Hertzberg (eds).

496

Copyright © 2006 Taylor & Francis Group plc, London, UK

 9. Mine gases

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Hydrogen sulfide control on a longwall face

M.P. Valoski
Pittsburgh Safety and Health Technology Center, Technical Support, Mine Safety and Health Administration,
US Department of Labor

ABSTRACT: Exposure to hydrogen sulfide, while mining coal, is potentially hazardous to the health of coal
miners. Overexposure to hydrogen sulfide can cause respiratory failure or even death. Acute exposure can rapidly
cause a coma with or without convulsions.
Hydrogen sulfide has the distinct odor of rotten eggs. This strong, foul-smelling odor can warn miners when
they first encounter hydrogen sulfide. Unfortunately, hydrogen sulfide can quickly cause olfactory paralysis.
Once this occurs, miners no longer “smell” hydrogen sulfide and will continue working in high concentrations
of hydrogen sulfide without realizing it. Moreover, lethal concentrations may not be detected by a victim.
The odor threshold for hydrogen sulfide is 0.13 parts per million (ppm). This is far lower than Mine Safety and
Health Administration’s (MSHA’s) Coal Mine Safety and Health’s (CMS&H) permissible exposure limit (PEL).
CMS&H’s PEL is the 1972 American Conference of Governmental Industrial Hygienists Threshold Limit Values
(TLV). CMS&H enforces an 8-hour time-weighted-average (TWA8 ) of 10 ppm and an Excursion Factor of 2.
The Excursion Factor is defined by the ACGIH as the magnitude in level permitted above the TWA8 for a short
time provided the TWA8 is less than the TLV.
In the western United States some coal mines are encountering hydrogen sulfide. The concentrations of
hydrogen sulfide can be high. This obligates mine operators to evaluate miners’ exposures and control hydrogen
sulfide liberations using acceptable and appropriate techniques.
This paper describes an engineering evaluation of a multi-faceted approach for controlling hydrogen sulfide
exposures. The operator uses a combination of personal monitoring, engineering controls, and work practices to
minimize exposure to hydrogen sulfide.

1 INTRODUCTION hydrogen sulfide concentration without realizing it.

Hydrogen sulfide, often called “stink damp” by miners
because of its distinctive odor, is a naturally occurring
air contaminant found in some coal mines in the west-
ern United States. This contaminant is not a typical
hazardous gas in underground coal mines. Because
hydrogen sulfide potentially can cause serious harm
or even death, mine operators must evaluate exposure
of their miners to this hazardous gas. If the hydro-
gen sulfide levels are high then the mine operator
must develop acceptable and effective control mea-
sures to minimize exposure of miners to hydrogen
sulfide.
Hydrogen sulfide has the distinct odor of “rotten
eggs.” Humans can detect this odor at 0.13 parts per
million (ppm) (Stahl). While this odor can warn the
miner of the presence of hydrogen sulfide, this gas
quickly causes olfactory paralysis. Once this occurs,
the miner quits “smelling” hydrogen sulfide. With the
miner incapable of detecting an increasingly strong
odor, the miner can work in an area with increasing
The hydrogen sulfide may overcome the miner.
High concentration of hydrogen sulfide can lead
to a coma with or without convulsions. Continued
exposure to the atmosphere will lead to death.
Other adverse health effects of hydrogen sulfide
exposure include eye and respiratory system irritation.
2 EXPOSURE LIMITS
Coal Mine Safety and Health (CMS&H) obligates
mine operators to control noxious and hazardous
gas(es) using administrative and engineering controls.
These controls must lower the miner’s exposure to or
below the permissible exposure limit (PEL). CMS&H
enforces the 1972 Threshold Limit Values (TLVs® )
published by theAmerican Conference of Governmen-
tal Industrial Hygienists (ACGIH) as the PELs.
The CMS&H PEL for hydrogen sulfide is an
8-hour time-weighted-average of 10 ppm. The Excur-
sion Factor is 2.

499

Copyright © 2006 Taylor & Francis Group plc, London, UK

 The National Institute for Occupational Safety and Table 1. Hydrogen sulfide concentrations in ppm along the
Health (NIOSH) recommends a ceiling limit for hydro- longwall face measured with the iTX.
gen sulfide of 10 ppm and the Immediately Dan-
gerous to Life or Health Level (IDHL) is 100 ppm Day 1 Day 2 Day 3
(NIOSHa).
Location TWA EL TWA EL TWA EL
Both the PEL and NIOSH recommended levels
are higher than the odor threshold. Detecting hydro- Last open 0 0 0 0 0 0
gen sulfide can cause people to become concerned. crosscut
People can smell hydrogen sulfide without being over- Headgate 2.4 5.8 1.9 5.5 0 0
exposed to it. Midface 5.3 21.2 4.9 24.4 0 0
Tailgate 10.4 21.1 10.6 18.4 0 0
Return 1.2 2.1 0.8 3.0 0 0
3 DESCRIPTION OF THE MINE

The mine, which began operation as strip mine before

the overburden became too thick, has one longwall Table 2. Hydrogen sulfide concentrations in ppm along the
section and several continuous mining sections. The longwall face measured using NIOSH method.
longwall has a remote controlled double drum shearer.
The shearer cuts in both directions. Continuous miners Day 1 Day 2 Day 3
are used to develop the longwall panels. The long- Location TWA TWA TWA
wall panel is approximately 1000 feet in width. The
Last open crosscut 0.03 2.4 0.0
height of the coal seam is about 114 inches. During
Headgate 2.2 1.0 0.02
the investigation, approximately 14,400 tons of coal Midface 7.3 7.0 0.02
was produced on each production shift. No coal was Tailgate 8.5 14.0 0.02
mined during the maintenance shift. Return 2.5 2.4 0.04

4 SURVEY TECHNIQUES

An engineering survey was conducted over three day- Industrial Hygiene Association accredited laboratory
light work shifts. The first two were production shifts analyzed the samples using ion chromatography.
and the last was a maintenance shift. No personal In addition, ventilation measurements were con-
exposures were determined. Only area samples were ducted with a calibrated, handheld vane anemometer,
collected. a stop watch, and a tape measure. The ventilation mea-
The hydrogen sulfide concentrations were mea- surements were conducted at the same locations as the
sured at five locations on the longwall panel: last open hydrogen sulfide measurements.
crosscut, headgate (shield #4), midface (shield #83),
tailgate (shield #175), and mouth of the return. Two
methods were used to measure the hydrogen sulfide
5 RESULTS
concentrations.
The first method used Industrial Scientific Corpo-
The hydrogen sulfide levels varied as to the position of
ration iTX1 electronic instruments. With these instru-
the shearer along the longwall face. Once the shearer
ments the time-weighted-average for each 60 second
passed the sampling location, the hydrogen sulfide
interval was recorded. The data were downloaded onto
concentration decreased.
a laptop computer for later data analysis.
As expected for the work shifts, the tailgate had
The second method was NIOSH Method 6013
the highest concentration of hydrogen sulfide and the
(NIOSHb). This method collects hydrogen sulfide on
headgate the least. Table 1 summarizes the data from
activated charcoal tubes preceded by Zeflour filters.
the electronic instrument and Table 2 presents the data
MSA Escort ELF sampling pumps with Gemini twin
using the NIOSH method. Both the time-weighted-
port sampler flow regulators were set to draw 0.125
average (TWA) for the work shift and the highest
liter per minute of air through the sampling train. The
Excursion Level (EL) recorded for the iTX are listed
Zeflour filters and charcoal tubes were changed after
in the tables. The EL is the highest concentration mea-
approximately four hours of sampling. An American
sured during any one 15 consecutive minute period.
The NIOSH method only gives the average exposure
1
Reference to brand names is made to facilitate understand- over the measurement duration. The NIOSH results
ing and does not constitute an endorsement by the author or were combined to yield time-weighted-averages for
the Mine Safety and Health Administration (MSHA). the work shift.

500

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 3. Average air flow measurements.
Average air velocity Average air quantity
Location in feet per minute in cubic feet per minute
Last open 740 133, 500
crosscut
Headgate 740 128, 600
Midface 720 110, 800
Tailgate 650 91, 000
Return 310 53, 500
Both the iTX and NIOSH method determined sim-
ilar levels of hydrogen sulfide in the air considering
the measurement locations were in close proximity
and not in the same location. Essentially, no hydro-
gen sulfide was entering the longwall panel in the
intake air. The highest hydrogen sulfide concentra-
tions for the work shift were measured at the tailgate.
The returns had lower hydrogen sulfide concentra-
tions than the tailgate because air from other locations
besides the longwall face diluted the hydrogen sulfide
concentration.
On the maintenance shift (Day 3) very low hydrogen
sulfide concentrations were measured. This argues that
the hydrogen sulfide is contained within the coal.
Table 3 presents the air flow measurement results.
Please note that the return entry air quantities are at
the instrument location and not the total air quantities
in the returns.
Based upon the measured hydrogen sulfide con-
centrations, the peak liberations were 280 cubic feet
per minute (cfm) and 310 cfm on Day 1 and Day 2,
respectively.
6 CONTROL MEASURES
The mine operator has adopted a multi-faceted
approach for controlling hydrogen sulfide exposures.
The approach uses personal monitoring, administra-
tive controls, and engineering controls.
Mine management requires each miner and visi-
tor (including mine inspectors) to the longwall face to
wear an Industrial Scientific Corporation T82 hydro-
gen sulfide detector. At the end of the work shift each
individual’s exposure to hydrogen sulfide is down-
loaded, recorded, and maintained in a computer file.
The T82 continuous measures the hydrogen sulfide
concentrations. When the hydrogen sulfide concen-
tration reaches 10 ppm, the T82’s alarm (sound and
vibration) activates. This informs the miner that he/she
needs to move to an area with lower hydrogen sulfide
concentrations.
As an administrative control, mine management
does not permit anyone, including the shearer operator,
on the tailgate side of the shearer when the shearer is
501
Copyright © 2006 Taylor & Francis Group plc, London, UK
cutting coal. This has the added advantage of reduc-
ing the miners’ exposures to respirable coal mine dust
because the miner is not in the dust cloud. Further, the
mine operator requires all personnel in the returns to
carry an appropriate, MSHA-approved respirator for
hydrogen sulfide. The respirator is for emergency use.
Two engineering controls are employed to protect
miners from high hydrogen sulfide exposures.The first
is ventilation. The air flow sweeps the hydrogen sulfide
and other noxious and harmful gases and dusts from
the face. The ventilation alone was not sufficient to
keep the hydrogen sulfide concentrations at the tailgate
to below 15 ppm.
The second engineering control supplements the
ventilation by regulating the maximum shearer speed
based upon the hydrogen sulfide concentration mea-
sured by a sensor on the tailgate. Once the hydrogen
sulfide concentration exceeds 15 ppm, the maxi-
mum shearer speed automatically slows based on
an algorithm. The equation will control the actual
shearer speed by linearly tapering the shearer speed
beginning when the hydrogen sulfide concentration
reaches 15 ppm.A shearer operator cannot override the
decreased maximum shearer speed. A shearer opera-
tor can operate the shearer slower than the decreased
maximum speed. The reduced shearer speed allows
the liberated hydrogen sulfide to be removed by the
ventilation.
As the hydrogen sulfide concentration increases
above 15 ppm, the maximum shearer speed decreases
further. If the hydrogen sulfide concentration
decreases, the maximum shearer speed increases,
based upon a mathematical function. In the range from
15 ppm to 25 ppm, the algorithm gradually reduces the
shearer speed until the hydrogen sulfide concentration
reaches 25 ppm. At 25 ppm the shearer speed is cutting
at its slowest maximum cutting speed. If the hydrogen
sulfide concentration reaches 27.5 ppm, the algorithm
is set to stop the shearer completely so that no coal is
being cut.
When the hydrogen sulfide concentration reaches
30 ppm, the power to the shearer is cut and the shearer
cannot be restarted until the hydrogen sulfide concen-
tration decreases to less than 10 ppm at the tailgate.
7 CONCLUSIONS
The hydrogen sulfide is in the coal at this mine. When-
ever the coal is sheared from the face or crushed, hydro-
gen sulfide is released. By slowing or stopping the
shearer, the ventilation can sweep the hydrogen sulfide
from the face. The work practice of keeping all person-
nel upwind of the shearer while cutting coal is good.
Using a multi-faceted approach of engineering and
administrative controls the exposure of coal miners to
hydrogen sulfide can be limited.
 REFERENCES
Code of Federal Regulations, CFR Part 75.322.
NIOSHa 1997. NIOSH Pocket Guide to Chemical Hazards.
DHHS (NIOSH) Publication No. 97-140 U.S. Govern-
ment Printing Office.
NIOSHb 2003. NIOSH Pocket Guide to Chemical Hazards
and Other Databases. DHHS (NIOSH) Publication No.
2004-103 National Institute for Occupational Safety and
Health.
Stahl, W.H. (editor) 1973. Compilation of Odor and Taste
Threshold Values Data. Philadelphia: American Society
for Testing and Materials.

502

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Early detection of flammable gas danger in hard rock mines to

protect miners

F.J. van Zyl, J.J. Dekker & J. Haarhoff

CSIR, Johannesburg, South Africa

ABSTRACT: According to the South African Department of Minerals and Energy’s (DME) database,
flammable gas accidents accounted for 7,3% of fatalities in the South African mining industry between 1998 and
2001. Nearly 97% of these accidents occurred in mines other than collieries. Most research in South Africa over
the past decade has focused on collieries, but five of the last six devastating flammable gas-related events (four
or more deaths) occurred in hard rock mines. The aim of this paper is to review the location of these accidents,
the operations associated with them, and, most importantly, to propose suitable interventions to address the
identified shortcomings.

1 INTRODUCTION Table 1. Number of accidents, injuries and fatalities per

commodity with respect to ignitions and ignitions of
The official DME data (SAMRASS) indicated flammable gas, for the period January 1988 to April 2005.
flammable gas accidents accounted for 7,3% of fatal-
ities in the mining industry as a whole, with nearly Commodity Accidents Injuries Fatalities
97% of these occurring in mines other than collieries
Gold 23 25 67
for the period 1998 to 2001. Coal 15 52 8
The number of accidents, injuries and fatalities per Chrome 1 1 0
commodity, with respect to ignitions and ignitions of Diamonds 8 23 6
flammable gas, for the period January 1988 to April Platinum 30 40 8
2005, is depicted in Table 1. Prospecting Minerals 1 3 0
The Mine Health and Safety Council’s (MHSC) Total 78 144 89
research over the past decade has focused mainly on
collieries, with only one of the 33 research projects
completed directed towards hard rock mines.
place in a longwall stoping section and claimed the
lives of 7 employees and caused damage to two up-dip
ledging panels.
2 THE MOST RECENT HARD ROCK
The last major incident occurred on 8 May 2001
FLAMMABLE GAS INCIDENTS IN
at the Beatrix Gold Mine No. 2 Shaft. The incident
SOUTH AFRICA (1999 TO 2005)
occurred in 16G Drive West and the crosscuts of
two partially mined-out stopes. Thirteen employees
2.1 The last three big flammable gas incidents in
perished in this incident.
South Africa
Since 1999 there have been three major flammable gas
2.2 Findings from official investigations
incidents in the South African underground hard rock
mining industry. In terms of Section 65(1) of the Mine Health and Safety
The first incident occurred on 29 July 1999 at Act (MHSA), the Chief Inspector of Mines directed
AngloGold’s Mponeng Gold Mine. This flammable an Inspector to conduct an inquiry into each of the
gas explosion occurred at a twin haulage development disasters.
section (2 737 metres below surface) and resulted in Although very specific recommendations were
the death of 19 employees. made for each incident based on the shortcomings
The second incident occurred on 15 May 2000 at identified and lessons learnt, definite similarities
the Beatrix Gold Mine No.1 Shaft. This incident took between the incidents can be drawn. In all three

503

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Other Shaft
11% 2%
Shaft Stope
7% Development 23%
50%

Stope
32%

Development
Figure 1. Location of flammable gas incidents in hard rock
75%
mines - gold and platinum (Cook 1998).
Figure 2. Location of flammable gas related fatalities in
cases two common primary causes, which led to the hard rock mines – gold and platinum (Cook 1998).
flammable gas explosions can be identified. These are:
– Unforeseen changes to the ventilation system, with- 100%
Accidents contributed to

out the consent of the ventilation officer, or due to 80%

natural changes, and
– Poor detection of the presence of flammable gas. 60%

2.3 Research findings
During 1999 the MHSC initiated a research project
titled “GAP 504 – The occurrence, emission and igni-
tion of combustible strata gases in Witwatersrand gold
mines and Bushveld platinum mines, and means of
ameliorating related ignition and explosion hazards”.
The primary objective of the project was to improve
the understanding of gas emissions in general. Two
of the areas researched were the working area where
flammable gas incidents occurred in underground hard
rock mines, and the actions associated with flammable
gas ignitions.
When looking at the location where flammable gas
were detected in hard rock underground mines, it was
found that half of all reported flammable gas occur-
rences occurred in development ends and approxi-
mately a third in the stopes.
This is not an unexpected result, as in develop-
ment ends ground is broken into virgin areas of the
mine with an associated higher risk of intersecting new
flammable gas sources. In the case of stopes, it is the
area where the most rock is broken on a regular basis
and in large quantities, again raising the possibility to
intersect gas sources (Figure 1).
If this information is now compared to the working
areas where most fatalities occur, it is not surprising
that nearly all fatalities occurred in these two areas of
a mine (Figure 2).
The presence of flammable gas cannot in itself lead
to a flammable gas explosion. For this to occur enough
gas needs to accumulate undetected, and an ignition
40%
20%
0%
Ventilation Testing Contraband
Figure 3. Major causes of flammable gas accidents (Cook
1998).
source needs to be present. GAP 504 identified three
main causes leading to flammable gas accidents:
– unplanned changes to ventilation,
– not detecting the presence of flammable gas, and
– contraband, including tampering with cap lamps to
make igniters (Figure 3).
When looking at the findings of the research report
completed in 1998, and comparing it to the accident
investigation report findings, it can be seen that the
identified potential causes need to be addressed in
order to reduce the risk of underground flammable
gas explosions.
3 CURRENT MINE PRACTICES
3.1 Codes of practice
One of the tools utilised by mine management to min-
imise the risk of underground gas explosions is to
compile a code of practice (CoP). It is a legal require-
ment of the South African Government instructing
mine operators to develop a CoP that specifically deals

504

Copyright © 2006 Taylor & Francis Group plc, London, UK

 with the detection and control of flammable gas in
underground hard rock mines.
To assist the mines in this regard the DME has com-
piled and distributed a “Guideline for the Compilation
of a mandatory Code of Practice for the Prevention of
Flammable Gas Explosions in Mines other than Coal
Mines (DME 2002)”.
The objective of this guideline is to assist the
employer of every mine, other than a coal mine, to
compile a CoP which, if properly implemented and
complied with, would considerably reduce the risk of
an ignition of flammable gas.
In summary, the aspects that need to be addressed
in the CoP include:
– Identification of flammable gas sources, occur-
rences and composition.
– Control of gas emissions.
– Detection of flammable gas.
– Dilution, removal and dispersion of flammable gas.
– Identification of hazardous locations.
– Preventing ignition of flammable gas.
Although the CoP guideline document comprehen-
sively focuses the attention of the mine operator to the
minimum standards, for various operations and loca-
tions, required, it is not prescriptive as to how to do this.
Specifically looking at the findings of the acci-
dent investigations and the research conducted around
flammable gas incidents, the guideline does address
development ends and stopes. It also addresses proce-
dures for the detection of flammable gas in mines, and
how to ventilate for control.
3.2 Flammable gas monitoring instrumentation
Although various instruments are used in underground
hard rock mines to detect flammable gas, the operating
principle of all the detection equipment is the same,
i.e. pellistor based flammable gas detectors calibrated
for methane. Numerous versions of these instruments
are available, from spot reading alarm instruments, to
continuous reading measuring instruments.
Although pellistor-based flammable gas detection
technology is well established, on reviewing current
mine practices during this project, interesting findings
where made. These were:
– Although mixtures of hydrocarbon gasses are found
underground, all instruments are calibrated to alarm
at the lower explosive limit (LEL) of methane.
But in the presence of hydrogen, this can lead to
dangerous under reading of the LEL.
– Contract labor force is not well trained on the use
of the instruments.
– Sensor problems/failure and faulty readings due
to moisture, water ingress, barometric pressure
change, cross-sensitivity and zero settings drifting
are common.
505
Copyright © 2006 Taylor & Francis Group plc, London, UK
– Alarm levels of the instruments are too soft and
difficult to hear.
– Displays are too dim on some instruments.
– Generally, the lamp room personnel and train-
ing department personnel are trained by Original
Equipment Manufacturers (OEMs) and the work-
force are then trained by the training personnel.
– Instrumentation is mainly issued to personnel in
supervisory positions only, and in some cases face
drillers.
– Battery problems encountered with most of the
units. High temperatures affect the batteries. Bat-
tery life is too short and does not last for eight hours.
– Age of instrumentation varies between eight
months and nine years.
3.3 Potential short comings
Although technology exists to detect the presence of
flammable gas, and mine operators are legally required
to specifically address the risk of flammable gas in
their mines, three major incidents claimed the lives
of 39 miners. With the benefit of hindsight poten-
tial shortcomings were identified that need to be
addressed.
– Firstly, the erratic occurrence of flammable gas in
underground mines (DME, 2002) means high risk
areas need to be monitored constantly.
– Secondly, dedicated equipment/systems need to
be developed to focus on the high risk opera-
tions, that not only monitor for the presence of the
risk (flammable gas), but also monitor the control
(ventilation).
– Thirdly, workers in immediate danger need to be
empowered by reliable information to withdraw
immediately from a danger area. This withdrawal
is not only applicable for the team detecting the
hazard, but everybody in the immediate ventilation
district.
– Lastly, supervisory personnel need to get a history
of the build-up of the hazard and the operation of the
control measures to be able to take the appropriate
action.
4 POTENTIAL SOLUTION
4.1 Confines of required solution
In light of the potential shortcomings identified in
section 3.3, a solution was needed. In order for the
solution to be practical some boundaries where set.
These where:
– Only current off-the-shelf technology is to be used.
– The system must inform the workers in immediate
danger to take action.
 – The system must inform supervisory personnel of
the hazard, or potential hazard.
Based on the identified potential shortcomings and
within the set boundaries, the CSIR has developed
a methodology and equipment to address the prob-
lem, by not only reliably detecting the presence of
flammable gas, but also taking into consideration the
state of the control measures before activating the
evacuation alarm.
4.2 Scope of proposed solution
The proposed solution, utilising CoPs requirements,
current flammable gas-detection technology, air flow
monitoring equipment, underground radio frequency
(RF) communication systems, and communication
software in a cost-effective manner, permits miner/s
in high risk areas in an underground mine, finding
themselves in immediate danger of a gas explosion, to
take the appropriate and correct action. The solution
aims to allow a section-wide timely withdrawal, and
a record of where the event was recorded. This will
allow supervisory personnel to pin point the event in
space and time and initiate the correct chain of events
in time to prevent a gas explosion from occurring.
The developed system aims to detect flammable gas,
or situations that can lead to sudden flammable gas
build up, in a workplace. This is done by monitoring
both flammable gas and the ventilation conditions in
identified high risk areas.
As stated earlier, flammable gas can only be diluted
with fresh air. Therefore, it is important both the pres-
ence of flammable gas and ventilation be monitored.
The level of flammable gas detected gives an indi-
cation of the immediate safety of the miners, while
the ventilation flow is an indication of the prepared-
ness of the area to deal with an ingress or outburst of
flammable gas.
This information needs to be used at two different
levels, i.e. employee level and the supervisor level.
– At the employee level a go/no go signal is required,
indicating if it is safe to work or not in their section,
not just at their workstation. This message needs to
be clear, visual and reliable in the hot, dark and
noisy underground environment.
– At supervisor level, more information is required
to be able to be pro-active in “predicting” potential
danger situation and taking preventative actions.
This information must be central, be fixed in space
and time, and indicate the status of contributing fac-
tors (gas levels) and controlling factors (ventilation
conditions).
4.3 Operating principle of the system developed
To be able to allow section wide awareness of the sta-
tus of the risk of flammable gas, and allow enough
506
Copyright © 2006 Taylor & Francis Group plc, London, UK
information to supervisory personnel to take action,
an in-section portable telemetry system was identified
as the most appropriate way to go forward.
The system has to be portable as blasting operations
conducted in working areas can potentially damage
a fixed system. Portability also allows the different
teams, e.g. drilling, mucking, meshing, etc., to take an
instrument to the area where they were “disturbing”
rock. It also has the advantage that the system can be
checked and calibrated after each shift.
The devised prototype system comprises of a base
station and four monitoring units. The focus of the
system is to monitor flammable gas and ventilation
conditions, in a stope or development end, by means
of the monitoring units, and sending this information
to the base station. The base station will monitor the
in-section conditions by processing the information
based on pre-described logic. If a “no-go” situation is
identified, the base station will activate the alarm on all
the monitoring units, affecting a section wide retreat.
The base station will also record which monitoring
unit alarmed, when it alarmed and what the cause of
the alarm was.
The base station will also record non-critical events,
i.e. a steady rise in flammable gas in the section or lack
of sufficient ventilation. This information will be dis-
played on the base station as potential warning signals
for the supervisor to review on inspection and to take
appropriate preventative or corrective action.
All no-go and warning signals can be sent to an
appropriate person, by linking the base station into the
existing mine communication system, to alert them
of any condition requiring their attention. In the same
manner control rooms or any other person or system
which needs to be informed can be informed. (The
current prototype does not have this feature built in).
The no-go and warning signals can reside at the base
station, but be password protected to identify the per-
son and the time the preventative or corrective action
was taken.
4.4 Practical considerations associated with the
developed system
Due to the density of rock and the current status of off-
the-shelf RF technology, telemetry is only possible in
line of sight. As each of the monitoring units needs to
be poled by the base station and are not always in line
of sight (Figure 4), each unit must also act as a repeater
to relay information from other units in its line of sight.
To minimise the impact of a unit or base station
failure, each unit is capable of working autonomously.
This means a unit will keep on working even if con-
tact with the base station is lost. Unit failure will be
displayed on the base station when detected.
When an alarm or warning signal is raised it stays
active until it is reset by the appointed responsible
 Figure 4. Graphical representation of a potential application of the prototype system

Figure 5. The prototype base station and a monitoring unit.

person. This is done in a password protected manner Figure 6. A monitoring unit.

to record the person and the time of the intervention.
To ensure workers know the system is working,
but these features could not be made to work and even-
flashing LEDs will flash green in a go situation and
tually all the communication was done in firmware.
red in a no-go situation. The same flashing principle
The computation unit communicates to the radio mod-
is present on the base station display.
ule using a SPI based protocol for configuration and
one additional communication line for the radio data.

4.5 Technical specifications of the system

4.5.1 Communication
To keep the price down, standard 433 MHz radio
IC’s (IA4420) were selected for providing the com-
munications link. The chosen radios have 100 m of
line-of-sight range, and are half-duplex transceiver
modules. The IC’s advertised automatic band rate gen-
eration for transmission and automatic sync on receive,
4.5.2 Sensing
A complete off-the-shelf airspeed sensor was procured
for sensing the current airspeed. The sensor chosen
(Davis Windscribe) provides an I2 C based interface to
access the current airspeed.
A Pellistor-based flammable gas sensor calibrated
with a sensing range of 0%–100% of the lower explo-
sive limit of methane is used in the system.

507

Copyright © 2006 Taylor & Francis Group plc, London, UK

 4.5.3 Power
Power to the whole system is provided by a 6V sealed
lead-acid battery equipped with a low-voltage cut-out
circuit.
4.5.4 Computation
The PIC16F876A was chosen to integrate the com-
munication and sensing functions. The PIC16F876A
is a low-cost, low-power device with enough com-
putational throughput to handle the communications.
It also has enough memory to hold all the firmware
needed to provide the communications functions with-
out additional hardware.
4.6 Protocol for mesh networking
One of the most challenging tasks in this design was
to provide a very lightweight communications proto-
col which would allow the monitoring nodes to mesh
network, yet be able to fit on a standard PIC16F type
device. There are two parts to the protocol – one the
message handling by the monitoring nodes and the
other the network detection performed by the base sta-
tion. The next two sections will both refer to Figure 4
that shows a sample underground setup.
There are four monitoring nodes with ID’s 1, 2, 3
and 4. The base station reserves the 0 ID. For clarity,
nodes 1–4 will be referred to as N1-N4. As can be seen
from Figure 4, the base station has a line-of-sight link
to N3 and N4 and can communicate to these nodes
directly. N3 has a line-of-sight link to N1, N1 has a
line-of-sight link to N2. Thus, for the base station to
communicate to N2, the message must go to N3, then
to N1 and then to N2. The section below describes how
each node handles the messages passed to it.
4.6.1 Message handling
Assume for now the base station has discovered the
layout of the network and knows that to communicate
to N2, for example, it has to relay the message via N3
and N1 respectively. Assume the base station wants to
receive a simple acknowledge reply from N2 to check
for its existence, it would transmit a message such as
the following:
"<N3><N1><N2><ACKNOWLEDGE><N0>"
This message will be received by N3 and N4. N4
discards the message as it is addressed to N3. Because
the message is addressed to N3, N3 then inspects the
rest of the message. The next part of the message is
not a command but the ID of another node (N1), so
N3 rotates the command to the left and transmit the
following message:
"<N1><N2><ACKNOWLEDGE><N0><N3>"
This command is received by N4, N1 and the base
station. The base station and N4 ignore the command
508
Copyright © 2006 Taylor & Francis Group plc, London, UK
because it is not addressed to them. N1 receives the
command, looks at the next part and discovers it is an
ID of another node (N2), and not an instruction. N1
then proceeds to rotate the command to the left and
transmits the following:
"<N2><ACKNOWLEDGE><N0><N3><N1>"
N3 and N2 receive this command, but it is dis-
carded by N3. N2 inspects the rest of the command
and discovers that the next part of the command is
the ACKNOWLEDGE instruction. It knows that the
response to ACKNOWLEDGE is ANSWER, and then
proceeds to reverse the command string it received. It
then transmits:
"<N1><N3><N0><ANSWER><N2>"
This command then travels back to the base station
in exactly the same way; each node rotates the com-
mand if it is not the intended recipient. This approach
allows the communication functions on the monitor-
ing nodes to be very lightweight. Each node only has
to be able to rotate a command string, and to reverse it.
4.6.2 Network discovery
In order for the base station to discover the topology
of the network, it first sends a direct ACKNOWL-
EDGE command to all four nodes. If it receives a
reply from any of the nodes it knows there is a line-of-
sight connection to that node. It then proceeds to send
ACKNOWLEDGE commands to the undiscovered
nodes via the discovered nodes until all nodes have
been found, or all possibilities have been exhausted.
The discovery sequence for the system in Figure 4
would be:
"<N1><ACKNOWLEDGE><N0>
<N2><ACKNOWLEDGE><N0>
<N3><ACKNOWLEDGE><N0> (reply)
<N4><ACKNOWLEDGE><N0> (reply)
<N3><N1><ACKNOWLEDGE><N0> (reply)
<N3><N2><ACKNOWLEDGE><N0>
<N4><N2><ACKNOWLEDGE><N0>
<N3><N1><N2><ACKNOWLEDGE><N0> (reply)"
After successful completion of the above discov-
ery sequence the base station would have discovered
all four nodes. This discovery process is run when
the system is powered up and thereafter whenever the
base station loses communications with a node it had
previously established communication with.
4.7 User interface
A PC-based user interface was developed to show and
store the information collected from the monitoring
nodes as required. The PC connects to the base station
with a RS232 based serial protocol.
 5 CONCLUSIONS & FUTURE
CONSIDERATIONS
5.1 Operational considerations
During underground evaluation of the prototype sys-
tem the concept was well accepted, as the workers felt
empowered to make decisions regarding their safety.
Supervisors also found the system helpful as the his-
tory of ventilation and flammable gasses condition
could be quickly reviewed, and corrective action taken
if required. Although the system did not have the com-
munication built in to inform the supervisor of an
emergency condition, it was felt that this will add value
to the system.
It was requested the system be made less bulky than
the prototype. This can be done by using an alternative
power source to the sealed lead acid batteries. The use
of alternative airspeed monitors will also significantly
reduce the size of the monitoring units.
The software interface also needs to be developed
to be modular in construction. This will help to cus-
tomise the interface, which is important as different
mine groups have different terminology and CoPs.
5.2 Technical considerations
A problem with the current system is that the half-
duplex communication system is relatively slow and it
places a heavy burden on the PIC16F876A because all
509
Copyright © 2006 Taylor & Francis Group plc, London, UK
the low-level communications need to be handled by
software functions. In a future revision it would be a
good idea to move to a full-duplex radio system which
can perform most of the communications functions in
hardware.
The current airspeed sensors are proprietary, have
difficulty communicating with other systems and are
bi-directional. It would be beneficial to replace them
with a sensor that more easily communicates with
other systems and is omni directional. This will also
assist with the size consideration of the monitoring
units.
REFERENCES
Cook, A.P. 1998. GAP 504 – The occurrence, emission and
ignition of combustible strata gases in Witwatersrand gold
mines and Bushveld platinum mines, and means of ame-
liorating related ignition and explosion hazards. MHSC
Final Report, South Africa.
Department of Mineral and Energy: Mine Health and Safety
Inspectorate. 2002. Guideline for the Compilation of
a mandatory Code of Practice for the Prevention of
Flammable Gas Explosions in Mines other than Coal
Mines. Republic of South Africa.
Mponeng.1999.http://www.miningweekly.co.za/min/features/
health/?show=49794 (Sited 09/02/2006).
Nundlall, A.R. 2005. Case Study: Lessons Learnt From
Recent Flammable Gas Explosions in South African Hard
Rock Mines. IOHA Conference, Pilansberg, South Africa.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

CO dispersion from a coal fire in a mine entry

J.C. Edwards, R.A. Franks, G.F. Friel & L. Yuan

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

ABSTRACT: Five mine fire experiments were conducted in a 2.08 m high and 2.90 m wide, ventilated mine
entry in the National Institute for Occupational Safety and Health (NIOSH)’s Safety Research Coal Mine (SRCM)
to determine the effect of the dispersion of carbon monoxide (CO) on mine fire detection. CO measurements
were made at distances from 7.6 m to 45.2 m downwind from the fire with diffusion mode CO sensors positioned
near the roof. For small intensity fires, less than 30 kW heat release rate, generated by 14 kg coal in a 0.61 m
square tray, it was determined that air flow and sensor spacing were significant for fire detection at the 10 ppm CO
alarm level. Within 15.0 m downwind distance from the fire, 10 ppm CO alarm values occurred for volumetric
air flows less than 11.5 m3 /s. However, the 10 ppm CO alarm value did not occur 30.0 m downwind from the fire
for air quantities greater than 6.2 m3 /s due to dilutive mixing of the CO in the air stream. The criterion that the
mine fire alarms occur within 15 min of the onset of flaming combustion could not be consistently met with the
10 ppm CO alarm. This suggests the use of lower CO alarm values, or reduced CO sensor spacings for mine fire
protection. It is demonstrated how computational fluid dynamics (CFD) can be used to model the CO dispersion
downwind from the fire in support of a plan to optimize sensor spacings.
Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent
the views of the National Institute for Occupational Safety and Health.

1 INTRODUCTION fire. This dispersion process is enhanced by the ther-

The deployment of carbon monoxide (CO) sensors in a
mine entry to achieve early and reliable fire detection
is important for miner safety. The Code of Federal Reg-
ulations (2004) specifies CO sensor spacing distances
for a belt air-course. The fire source location for a fire
in a belt air-course typically occurs on the entry floor,
or near the entry half-height associated with the con-
veyor belt structure. Fires could occur due to frictional
heating at the belt drive and along the belt. Coal which
has accumulated on a belt drive can be heated by roller
slippage. Another possible fire source is the ignition
of spilled coal on the floor from cutting and welding
activity. Previous research (Litton, et al., 1991) has
established that approximately 15 min (average time
for 15 experiments was 14.3 min with a standard devi-
ation of 7.3 min) after a smoldering coal fire evolves
to the flaming stage, a conveyor belt in contact with
the fire can be ignited. Early and reliable fire detection
is important within this time frame. When a plume of
combustion products rises towards the mine roof due
to the thermally induced density differences between
the hot products-of-combustion (POC) and the ambi-
ent air, the POC will also be convected with the bulk
air flow downwind from the fire source. The POC will
be dispersed by the dilutive mixing with the fresh air
over the entry cross-section downwind from the source
mal equilibration of the POC with the ambient air. It
is important to know the expected distribution of CO
near the mine roof along the entry in the smolder-
ing and flaming coal fire stage to provide guidance
for sensor site location, and the effect of volumetric
air flow upon the CO concentration. The objective of
this research is to analyze experimentally and compu-
tationally the dispersion of CO from small coal fires
in a mine entry and determine the sensitivity of CO
concentration to ventilation and sensor site location.
2 EXPERIMENTAL CONFIGURATION
The experimental location selected for the fire exper-
iments to determine the effect of air flow and sensor
spacing upon fire detection was an entry in NIOSH’s
SRCM at the Pittsburgh Research Laboratory (PRL).
The entry height and width at the location of the fire
pan were 1.96 m and 2.79 m, respectively. Since in a
mine environment the source fire could occur on a belt
drive at entry mid-height, the experiments included
both floor and mid-height fire source locations. Coal
was selected as the fire fuel source. Five small coal fire
experiments were conducted. The CO concentration
was measured with diffusion mode Conspec Controls
Inc CO monitors near the roof at locations 7.6, 15.0,

511

Copyright © 2006 Taylor & Francis Group plc, London, UK

 30.0, and 45.2 m downwind from the fire pan. (Refer- Table 1. Mine fire experimental conditions.
ence to a specific product is for informational purposes
and does not imply endorsement by NIOSH.) At the Volumetric Average air
7.6 and 15.0 m locations two CO sensors each were Experiment air flow Q, velocity, Fire source
located at equally spaced intervals across the entry number m3 /s m/s location
near the roof. At the 30.0 and 45.2 m stations a sin-
1 9.34 1.71 Floor
gle CO sensor was positioned at the entry center near 2 2.74 0.50 Floor
the roof. Calibration of the CO monitors was con- 3 6.19 1.13 Floor
ducted before each experiment with hydrocarbon free 4 11.5 2.10 Mid-height
air and 25 ppm CO in air. The entry average height 5 5.84 1.07 Mid-height
and width were 2.08 m and 2.90 m. The entry associ-
ated hydraulic diameter was 2.42 m. Table 1 lists the
fire source location and entry air flow for each experi-
ment. Sensor data was collected every 2 seconds with a 20
mine monitoring system. For experiment nos. 1–3 the 7.6 m

FIRST FLAME
fire pan was positioned on the mine entry floor, and for 15.0 m
experiment nos. 4 and 5 the fire pan was positioned on 30.0 m 15.0 m
45.2 m
a set of blocks 0.89 m above the floor, which is approxi- 15
EVENT
mately half the 1.96 m entry height at the fire zone. For
each experiment about 14 kg of run-of-the-mine Pitts-

FIRST SMOKE
burgh Seam coal contained in a 0.61 m square tray was CO, PPM

heated by embedded electrical strip heaters. A small 10

quantity, about 0.5 kg, of pulverized Pittsburgh Seam 7.6 m

coal was added to the coal. Heating of pulverized coal

provided an early source of CO. Electrical power was
supplied to the strip heaters over a 30 min period with
an increase from 1.7 to 2.8 kW. The slow increase in
heating was to produce incipient CO emissions at the
lowest power source available, and to simulate the slow
heating to be expected from a heated element, such as
at a belt drive or along the belt. The linear air flow mea-
surements were made with a Solomat Neotronics hot
wire anemometer for experiment nos. 1–4. An aver-
age linear air flow was determined from a five point
average measurement over the entry cross-section. The
Solomat was not temperature compensated. A correc-
tion to the Solomat measured air flow was made based
upon a calibration of the Solomat with an Edra 6 Air
flow Developments Ltd vane anemometer over a tem-
perature range from 25 to 73 deg F. For the four temper-
ature evaluations a linear fit was made which had an
R-square value, coefficient of determination, equal to
0.9973. The corrected air flow values at the fire pan
are listed in Table 1. The volumetric air flow refers to
the entry cross-section at the fire pan. For experiment
no. 5 an ultrasonic flow monitor from El-Equip Inc
was used. The ultrasonic flow monitor measures the
average air flow along a diagonal between the entry’s
opposing ribs.
3 RESULTS
For all five experiments the CO measurements at the
7.6 and 15.0 m stations indicated the CO dispersion
was uniform over the midsection near the roof. Con-
sequently, the CO concentration at these two locations
POWER OFF
5
0
9:30 9:43 9:56 10:10 10:23
TIME
Figure 1. CO concentration at sensor stations in response
to coal fire on entry floor for experiment no. 2 at 7.6, 15.0,
30.0, and 45.2 m distance from fire.
was represented as an average. Experiment nos. 2 and
5 provide a comparison of the time dependent CO
concentration at the sensor stations with a fire on the
mine entry floor and at entry mid-height. These two
experiments are representative of the two experimen-
tal configurations for source fire location. Figures 1
and 2 show the measured CO concentration at each
station for experiment nos. 2 and 5, respectively. For
experiment no. 2 the heat source was increased from
1.7 kW to 2.8 kW over the time period 9:32 to 10:02.
For experiment no. 5 the heat source was increased
1.7 kW to 2.8 kW over the time period from 9:39 to
10:09. In each experiment the first flames occurred
approximately when the increased electrical power to
the strip heaters reached 2.8 kW. The measured CO
concentration at the 30.0 m and 45.2 m stations tracked
each other closely for all five experiments. These sta-
tions are approximately 12 and 19 hydraulic diameters
downwind from the source fire. The implication is that,
beyond 10 hydraulic diameters, the source fire product
emissions are well mixed over the entry cross section.

512

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 2. Measured values of maximum CO, and CO and

APPLY H2O
50 smoke optical density, OD, 15 min after flaming combustion

REMOVE HEATERS
45.2 m downwind from the fire source.

FIRST FLAME
7.6 m
40 15.0 m Exp. MAX. HR CO∗ , H∗R ,
30.0 m
45.2 m
No. CO, ppm kW, ppm OD∗ , m−1 kW
EVENT
30 1 6.8 17 6.84 —– 17
CO, PPM

2 15.8 11 14.06 —– 10
SMOKE
FIRST

3 7.8 13 6.64 0.024 11

20
4 7.2 22 2.54 0.0072 8
15.0 m 5 18 28 7.03 0.022 11
∗
10
15 min after visible flames

7.6 m At the 7.6 m and 15.0 m stations the 10 ppm CO

0 alarm value occurred for all five experiments. The CO
9:40 9:56 10:13 10:30 10:46
10 ppm alarm value was not achieved at the 30.0 m
TIME
and 45.2 m stations for experiment nos. 1, 3, and 4.
Figure 2. CO concentration at sensor stations in response
These latter experiments are associated with the air
to coal fire at entry mid-height for experiment no. 5 at 7.6, quantity greater than 6.19 m3 /s, and consequently, CO
15.0, 30.0, and 45.2 m distance from fire. is subjected to much greater dilution.

The CO concentrations at the 7.6 m and 15.0 m sta-
tions, which were three and six hydraulic diameters
downwind from the fire, had more divergent CO con-
centrations than those at the 30.0 and 45.2 m stations.
The instability of the CO production due to fluctu-
ations in the fire intensity was not averaged out by
the air flow in the 15 m zone immediately downwind
from the fire. This was observed for all five exper-
iments. In each experiment an approximately linear
increase in CO concentration occurred in the smolder-
ing stage, with an increased rate in the flaming stage.
For experiment no. 1 there was an electrical power dis-
ruption after initial flaming combustion. This resulted
in a cooling of the coal and the extinguishment of an
early flaming combustion. Power was reapplied, and
steady flaming combustion followed.
3.1 CO alert and alarm
For each experiment the 5 ppm CO alert value was
attained at each of the four sensor stations. Only for
experiment no. 2 did the CO alert occur in the smol-
dering combustion stage. This was a consequence of
the lowest volumetric flow rate of 2.74 m3 /s for exper-
iment no. 2. To assure early fire detection in the
smoldering stage for volumetric air flow rates greater
than 2.74 m3 /s for these small coal fires, either the
CO alert values could be lowered or the sensor spac-
ing could be decreased. Reference to volumetric air
flow rate must be understood with caution. The air
velocity can affect the fire production rate of CO. In
addition, the turbulent mixing of the fire POC plume
will be affected by the confinement height of the tun-
nel. However, for early detection of small fires, these
effects should be relatively small.
3.2 Fire heat production
The heat release for coal combustion can be estimated
from the CO generated in the flaming combustion
mode. The CO generated by the flaming combustion
of coal is 4.8 mg/kJ (Egan, 1990). Formally, the heat
release HR (kW) is determined from
where [CO] is the CO concentration in ppm, and Q is
the volumetric air flow in m3 /s.
Since eq (1) depends upon the volumetric air flow,
more complete mixing in a mine entry results in a bet-
ter estimate of the heat production rate when a CO
sensor near the roof is used to detect a concentration
representative for the cross-section. The observation
that the measured CO concentration at the 30.0 m and
45.2 m stations tracked each other provides the CO
concentration at the 45.2 m station as an estimate of the
heat production rate based upon uniform POC mixing.
The maximum CO concentration 45.2 m downwind
from the fire is used to represent the maximum heat
release rate in the flaming stage. The CO maximum
concentration and the associated heat release rate are
shown in Table 2. For experiment nos. 1–3 with the
fire source on the entry floor the highest maximum CO
concentration is associated with the lowest ventilation
of experiment no. 2. The maximum CO concentration
at the 45.2 m station occurred after the onset of flam-
ing combustion for each experiment. For experiment
nos. 1–2 the time at which the maximum CO occurred
at the 45.2 m station was approximately 15 min after
flaming combustion. The maximum CO at the 45.2 m
station occurred at least 30 min after flaming com-
bustion for experiment nos. 3–5. As shown in Table 2

513

Copyright © 2006 Taylor & Francis Group plc, London, UK

 the maximum fire intensities were less than 30 kW for 1.71 m/s 0.50 m/s
the fire sources used in these experiments. These fire
20 1.13 m/s 2.10 m/s
intensities are indicative of a small fire source.
1.07 m/s
3.3 CO and smoke 15 min after flaming
combustion 15

Previous research (Litton, et al., 1991) established the

CO, PPM
significance for fire detection within 15 min of flam- 10
ing combustion based upon ignition of a belt fire by
a small flaming coal fire. For these experiments, the
measured CO concentration 45.2 m downwind from 5
the source fire where the fire POC are well mixed is
listed in Table 2 at 15 minutes after flaming combus-
tion. Only for experiment no. 2 with a volumetric air 0
flow of 2.74 m3 /s was the CO concentration greater 0 10 20 30 40 50
than the 10 ppm alarm level 15 minutes after flaming DISTANCE, m
combustion. In experiment no. 5, the other experiment
Figure 3. CO concentration 15 min after flaming combus-
in which a CO alarm was reached, the CO did not reach tion dependence upon sensor location for average air flows
alarm level until 24 min after flaming combustion. of 0.50, 1.07, 1.13, 1.71, and 2.10 m/s.
In proximity to the fire, at the 7.6 m station, where
the contaminants are less diluted, the CO alarm value
20
was reached for experiment nos. 1 and 5 within 15
min after flaming combustion. For experiment no. 2 30 m 45.2 m

the CO alarm value occurred within the smoldering 15 m 7.6 m

Linear fit
combustion stage. For experiment nos. 3 and 4 the CO 15
alarms did not occur until 34 and 65 min after flam-
ing combustion. This inconsistency of the CO alarm
CO, PPM

time with air flow further supports the need to lower 10

the CO alarm level or place the sensor closer to any
in-mine locations identified as high risk for a potential
fire source.
5
A light obscuration monitor was located 1.6 m
upwind from the 45.2 m CO sensor location for experi-
ment nos. 3–5. The light obscuration monitor consisted
of an incandescent light source and a photovoltaic cell 0
separated by one meter. The measured light transmis- 0 4 8 12 16

sion through the light monitor was used to determine VOLUMETRIC AIR FLOW, m3 / s

the smoke optical density. Table 2 shows for experi-

ment nos. 3 and 5 that 15 minutes after visible flames
the CO concentration was in excess of the 5 ppm alert
value and the smoke optical density value was greater
than the smoke sensor optical density alarm value of
0.022 m−1 . The higher ventilation in experiment no. 4
resulted in CO alert and smoke alarm values occurring
65 min and 51 min after visible flames. For this latter
case the greater POC dilution suggests that a lower CO
or smoke sensor alarm value should be used for earlier
fire detection, or the placement of the smoke sensor
should be closer to any location identified as high risk
for a fire source.
Figure 3 shows the CO concentration 15 min after
flaming combustion at the four sensor locations for
the five cases. The separation in CO concentration val-
ues is more significantly dependent upon ventilation
velocity than whether the fire source is on the floor or
at mid-height. The lowest CO concentration is associ-
ated with the highest air flow of 2.1 m/s. The highest
Figure 4. CO concentration 15 min after flaming combus-
tion at 7.6, 15.0, 30.0, and 45.2 m distance from fire.
CO concentrations are associated with the lowest air
flow of 0.5 m/s. The intermediate CO concentrations
are associated with the range of air flows from 1.07 m/s
to 1.71 m/s. There is not a significant variation in the
CO concentration between the 30.0 and 45.2 m sensor
stations for a specific air flow. Because of the small
fire intensities, the air flow has little effect upon the
CO production.
Figure 4 shows the dependency of CO concentra-
tion upon the volumetric air flow for all four sensor
stations 15 min after flaming combustion for all five
air flow conditions. The best fit of a linear rela-
tionship to the data with an R square value of 0.73
shows that the CO concentration is dependent upon
the volumetric air flow, but there is a wide degree of
uncertainty. For fires of nearly equal intensity a linear

514

Copyright © 2006 Taylor & Francis Group plc, London, UK


CO
30 EVENT
OPTICAL
DENSITY
FIRST FLAME
CO, PPM
20
SMOKE
FIRST
10
0 0.00
9:40 9:56 10:13 10:30
TIME
Figure 5. Comparison of CO and smoke optical density at
45.2 m for experiment no. 5.
dependence of CO concentration upon volumetric
flow rate is expected. Extrapolation of the linear fit
of the data to non-detectable, near zero, CO concen-
tration occurs at a volumetric flow rate of 14 m3 /s. For
the 2.08 m by 2.90 m airway considered, the linear air
flow rate is 2.32 m/s. These results show the need to set
lower CO alarm values for higher air flow conditions
in a mine.
3.4 Correlation of CO and smoke
Figure 5 shows a comparison of the CO concentration
and the smoke optical density at the 45.2 m station
for experiment no. 5. The smoke optical density and
CO exhibit similar trends over the course of the exper-
iment. An analysis of the optical density D as it is
correlated with the CO concentration at the 45.2 m
station determined the linear correlation over the com-
bined smoldering and flaming combustion range of the
coal fire
where A is a proportionality constant.
The values for A and the coefficient of determi-
nation R2 are listed in Table 3 for experiment nos.
3–5. The R square value is not less than 0.79 for these
experiments.
The optical density D, measured in units of inverse
meters, m−1 , is related to the smoke mass concentra-
tion Cm by the relationship (Mullholland, 1988)
where the specific extinction coefficient Km depends
upon the size distribution and optical properties of the
smoke. A combination of equations (2) and (3) yields
Copyright © 2006 Taylor & Francis Group plc, London, UK
Table 3. Proportionality constant for optical density depen-
APPLY H2O
0.10 dence upon CO concentration.
Experiment number A R2
0.08
OPTICAL DENSITY , 1/m
3 0.0036 0.87
4 0.004 0.79
0.06 5 0.0045 0.86
REMOVE HEATERS
0.04
Equation (4) provides an interpretation of the smoke
concentration as linearly dependent upon the CO
0.02
concentration. The fit of the smoke optical density
with CO concentration did not distinguish between
smoldering and flaming combustion.
10:46 The value for A can be selected to be the average
of the three values for experiments 3–5. The aver-
age value of A is 0.004. This value in equation (2)
implies that a 5 ppm CO concentration is associated
with an optical density equal to 0.02 m−1 , and a 10 ppm
CO concentration is associated with an optical density
equal to 0.04 m−1 . The smoke sensor optical density
alarm value of 0.022 m−1 would be associated with the
5 ppm CO alert value.
An estimate can be made of Km based upon its
definition in terms of the extinction coefficient Qe ,
the smoke particulate mass density ρ, and smoke
particulate average diameter d.
For smoke particulate density 1,400 kg/m3 , diame-
ter of 0.3 micron, and extinction coefficient of 2.5,
the value of the specific extinction coefficient is
8,900 m2 /kg. The specific extinction coefficient is
comparable to the value of 7,600 m2 /kg (Seader &
Einhorn, 1976) for flaming combustion of wood and
plastics. The relationship between Cm and [CO] is
approximately
3.5 Simulation of CO concentration using FDS
CO concentrations from the small coal fires were
simulated using the Fire Dynamics Simulator (FDS),
an existing computational fluid dynamics program
(McGrattan, et al., 2002). The chemical structure of
the Pittsburgh coal is simplified as CH0.74 O0.08 . CO is
assumed to be created with constant yield at the flame
and transported with the combustion products. The
fraction of fuel mass converted into carbon monox-
ide is 0.078 obtained for Pittsburgh coal (Egan, 1990).
Other important parameters for the CO simulation are
the amount of energy released per unit mass of oxygen
consumed and the fire heat release rate. The amount of
energy released per unit mass of oxygen consumed for
515
 16
MEASURED
MEASURED
60 FDS
FDS
12

CO MEASURED 45
CO, ppm

CO, ppm
8
30
FDS

4 FDS
15

CO MEASURED
0 0
0 400 800 1200 1600 2000 0 500 1000 1500 2000 2500
(a) TIME, s (a) TIME, s

20
MEASURED MEASURED
FDS 12
FDS FDS
16

CO, ppm
CO, ppm

12 8

8
CO MEASURED
CO MEASURED
4

4 FDS

0 0
0 500 1000 1500 2000 2500
0 400 800 1200 1600 2000
(b) TIME, s
(b) TIME, s

Figure 6. CO concentrations predicted by FDS for experi- Figure 7. CO concentrations predicted by FDS for experi-
ment no. 2 with fire source on floor (a) at 7.6 m station; (b) ment no. 4 with fire source at mid-height (a) at 7.6 m station;
at 45.2 m station. (b) at 45.2 m station.

after 1,000 s at 7.6 m station. This is probably because

Pittsburgh coal is 11,900 kJ/kg derived from the heat of
combustion for the coal. From the CO measurements
in the experiment, it was determined that the whole
coal combustion process can be approximately divided
into three stages: the smoldering combustion stage,
the flaming combustion stage, and the peak combus-
tion stage. The heat release rates at these three stages
increased nearly linearly with the time. For the simula-
tions the heat release rate in each stage was simplified
as a linear increase with time. The heat release rates
at the end of each stage were estimated by equation
(1) with the CO concentration measured at 45.2 m sta-
tion. For the purpose of early CO alert and alarm, the
peak combustion stage may be not important but can
be used for the comparison between the simulation and
the experiment.
Figures 6 and 7 show the CO concentrations pre-
dicted by the FDS for experiments nos. 2 and 4,
respectively. For experiment no. 2 with the fire source
at the floor, FDS prediction was in good agreement
with the experiment at 45.2 m station while the FDS
predicted lower CO concentration than the experiment
the fraction of fuel mass converted into carbon monox-
ide was not always constant in the experiment. For
experiment no. 4 with the fire source at the mid-height,
FDS prediction was much higher than the experiment
after 1,500 s at 7.6 m station while a peak occurred in
FDS simulation before 1,000 s at 45.2 m station. Gen-
erally speaking, FDS simulation agrees well for the
experiment with the fire source at the floor and can be
used to optimize the CO sensor spacing. Further mod-
eling work is needed to improve the simulation with
the fire source at the mid-height.
4 CONCLUSIONS
For the small, less than 30 kW heat release rate inten-
sity, coal fires in a 2.08 m high and 2.90 m wide entry,
it was determined that:
• At distances greater than 10 hydraulic diameters
from the source fire, the CO concentration was
well-mixed over the entry.

516

Copyright © 2006 Taylor & Francis Group plc, London, UK

 • To assure detection in the coal fire smoldering stage,
it would be beneficial to decrease the alert and alarm
CO concentration values for volumetric air flows
greater then 2.74 m3 /s.
• Relatively close to the potential fire source, within
15 m of the fire, CO 10 ppm alarm values occurred
for the volumetric air flows less than 11.5 m3 /s. For
air quantities greater than 6.19 m3 /s, the 10 ppm
CO alarm did not occur at distances greater then
30.0 m downwind from the fire. This is consistent
with the requirement of the Code of Federal Regula-
tions (2004) that any CO sensor be less than 30.5 m
(100 ft) downwind from each belt drive unit.
• For volumetric air flow rates greater than 2.74 m3 /s,
a CO alarm was not achieved within 15 min after
the onset of flaming combustion 45.2 m distance
from the source fire. At a distance of 7.6 m from
the fire the occurrence of a CO alarm value was not
consistent with the air flows. This suggests the use
of lower CO alarm values for a CO sensor downwind
from a potential fire source, such as a belt drive.
• Extrapolation of CO concentration 15 min after
flaming combustion with air flow showed that the
CO concentration would not be detectable for air
flows greater than of 2.32 m/s at distances greater
than 7.6 m from the source fire.
• CO concentration correlates with smoke mass con-
centration and optical density.
• FDS simulation agreed well for the experiment with
the fire source at the floor and can be used to
517
Copyright © 2006 Taylor & Francis Group plc, London, UK
optimize the CO sensor spacing. Further modeling
work is needed to improve the simulation with the
fire source at the mid-height.
REFERENCES
Code of Federal Regulations. 2004. 30 CFR, Part 75.351 (e).
Office of the Federal Register, NationalArchives and Stan-
dards Administration, U.S. Government Printing office,
Washington, D.C.
Egan, M.R. 1990. Summary of combustion products from
mine materials: their relevance to mine fire detection. U.S.
Bureau of Mines Informational Circular 9272, 12 pages.
Litton, C.D., Lazzara, C.P., & Perzak, F.J. 1991. Fire detection
for conveyor belt entries. U.S. Bureau of Mines RI 9380,
21 pages.
McGrattan, K.B., Forney, G.P., Floyd, J.E., Hostikka, S., &
Prasad, K. 2002. Fire Dynamics Simulator (Version 3)
User’s Guide, U.S. Dept. of Commerce. National Institute
of Standards and Technology.
Mullholland, G.W. 1988. Smoke production and properties.
The SFPE Handbook of Fire Protection Engineering. Sec-
tion 1/Ch. 25, (Ed.-in-chief, P.H. DiNenno, National Fire
Protection Association, Quincy, MA).
Seader, J.D., & Einhorn, I.N. 1976. Some physical, chemical,
toxological and physiological aspects of fire smokes. 16th
Symp. (Int.) on Combustion, MIT, Cambridge, MA. August
15–20, 1976, (The Combustion Institute, Pittsburgh, PA),
pp.1423–1445.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

The application of open path infra-red detectors in underground coal mines

C. Fowler & D. Chalmers

School of Mining Engineering, The University of New South Wales, Sydney, Australia

ABSTRACT: The overall aim of the project was to identify and select the most ‘promising’ of the existing,
commercially available infra-red ‘open path’methane detection and measurement systems for use in underground
coal mines and to demonstrate it in a simulated mine environment.
Research issues that were to be addressed included the cross-sensitivity of such systems to gases other than
methane, their speed of response, sensitivity to airborne coal dust and to water droplets and films, and the effects
of vibration. Issues of safety including their ‘Certifiability’ for use in explosive gas atmospheres in Australian
underground coal mines and their ‘Approvability’ for use in hazardous zones in New South Wales coal mines
were also to be considered.
In addition, potential applications of infra-red ‘open path’ gas detection and measurement systems in
underground coal mines were to be identified.

1 INTRODUCTION Coal industry interest in employing the ‘open path’

Infra-red methane detection and measurement devices
have recently been introduced into underground coal
mines in New South Wales and Queensland. They
utilise the principle of the absorption by methane gas
of infra-red radiation. All are ‘point’ devices and work
by determining the concentration of methane after it
has passed from the mine atmosphere into an enclosed
cell or ‘cuvette’.
Infra-red adsorption devices afford several advan-
tage over the catalytic methanometers which have long
been employed in coal mines.
– They do not ‘burn’ the methane and, consequently,
are not dependent upon the concentration of oxygen
in the mine atmosphere.
– They can read up to 100% methane concentration.
– They respond more quickly to changes in methane
concentration than do catalytic devices but their
potential is still limited by the rate at which methane
can pass into the ‘cuvette’ through the necessary
filters.
A more recently developed class of infra-red detec-
tion and measurement device does not employ a closed
‘cuvette’but determines in-situ methane concentration
over a path between two points. The distance between
the points may vary from as little as 500 mm to more
than 200 metres. This so called ‘open path’ methane
detection and measurement has been widely adopted
by the oil, gas and petrochemical industries.
technique has centred on those areas of the mine
which are not readily accessible for conventional
‘point’ detection and measurement devices or where
adverse conditions limit their usefulness. However,
none of the existing commercially available infra-red
‘open path’ methane detection and measurement sys-
tems has been trialed in underground coal mines. Nor
has an assessment been made as to whether the under-
ground environment precludes the safe use of such
systems.
1.1 Objectives of the research
The overall aim of the project was to identify and select
the most ‘promising’ of the existing, commercially
available infra-red ‘open path’ methane detection and
measurement systems for use in underground coal
mines and to demonstrate it in a simulated mine
environment.
Research issues that were to be addressed included
the cross-sensitivity of such systems to gases other
than methane, their speed of response, sensitivity to
airborne coal dust and to water droplets and films, and
the effects of vibration. Issues of safety including their
‘Certifiability’ for use in explosive gas atmospheres in
Australian underground coal mines and their ‘Approv-
ability’for use in hazardous zones in New South Wales
coal mines were also to be considered.
In addition, potential applications of infra-red ‘open
path’gas detection and measurement systems in under-
ground coal mines were to be identified.

519

Copyright © 2006 Taylor & Francis Group plc, London, UK

 1.2 Main findings and conclusions
The objectives of the project have been achieved.
It has been demonstrated that, subject to detailed
safety modifications that have been identified, at
least one commercially available infra-red ‘open path’
methane detection and measurement system is suitable
for use in certain potential applications in Australian
underground coal mines.
Because of the relatively small market compared
with the oil, gas and petrochemical industries, high
entry costs (including Certification and Approval) and,
in particular, a lack of knowledge as to how ‘open path’
systems might be expected to behave in the under-
ground environment, manufacturers have shown little
interest in adapting such technology for coal mining.
Even in the case of infra-red ‘point’ methane moni-
toring devices, only a very few manufacturers have
taken the trouble to make their products available to
the Industry.
Matters may be expected to progress further now
that it has been demonstrated that infra-red ‘open path’
systems can function safely and effectively in the
underground mine environment. Further development
of such systems will be required, however, before they
will be useful in all potential underground coal min-
ing applications. These necessary developments have
been identified.
The comprehensive final report includes all find-
ings relevant to the application of remote methane
detection and measurement to Australian underground
coal mines.
1.3 Work programme description
Manufacturers of infra-red ‘open path’ methane detec-
tion and measurement systems were identified and
details of all commercially available systems obtained.
A set of selection criteria was developed specifically
for Australian underground coal mines, all of the
systems assessed against the criteria and the most
suitable selected for testing and appraisal including
demonstrating it in a simulated mine environment.
– Research issues that were addressed included,
– cross-sensitivity to hydrocarbons other than
methane and to oxygen and carbon dioxide,
– speed of response compared with ‘point detectors’,
– sensitivity to both respirable and inspirable air-
borne coal dust,
– sensitivity to airborne water droplets and water
films, and
– effects of vibration, including differential vibration.
Issues of safety which were considered included
‘Certifiability’ for use in explosive gas atmospheres in
Australian underground coal mines and ‘Approvabil-
ity’ for use in hazardous zones in New South Wales
coal mines. Potential industrial applications were also
identified and are listed below.
520
Copyright © 2006 Taylor & Francis Group plc, London, UK
1.4 Potential for industrial applications
Coal industry interest in employing the ‘open path’
gas detection and measurement technique has centred
on those areas of the mine that are not readily access-
ible for conventional ‘point’ gas monitoring devices
or where the adverse conditions limit their useful-
ness. In addition, it is considered that the fast response
time of such devices is a valuable feature that may
be exploited. The following potential applications of
‘open path’ gas detection and measurement systems in
underground coal mines have been identified.
– The detection and measurement of rapidly chan-
ging, transient methane concentrations during wind
blasts and outbursts.
– The monitoring of the open goaf behind the roof
supports when ‘hang ups’ occur during longwall
mining.
– The monitoring of the face side of the longwall
shearer on faces subject to frictional ignition.
– The replacement of multiple ‘point’ devices for the
detection and measurement of methane along belt
roads in ‘gassy’ coal mines.
– Methane monitoring in the vicinity of the face
during remote-controlled, deep cut (extended cut)
operations associated with ‘place changing’.
At least one of the commercially available infra-red
‘open path’ methane detection and measurement sys-
tem would be suitable for the first four application
listed above, subject to the detailed safety modifica-
tions that are identified in the body of the report. It is
considered, however, that current systems are unsuit-
able for the final application which must await the
development of devices that are ‘single ended’ and
capable of resolving methane concentration along the
path of the infra-red beam.
‘Open path’ gas monitoring systems, in general,
also have a potential role to play in the detection of
fire and in the detection and measurement of gases
from spontaneous combustion, fires and explosions.
1.5 Recommendations
For infra-red ‘open path’ methane detection and meas-
urement systems to realise their full potential in the
underground coal mine environment, further develop-
ment work needs to be undertaken.
Commercially available infra-red ‘open path’detec-
tion and measurement systems are certified for use
in explosive gas atmospheres. Unfortunately, the cer-
tification is restricted to ‘places with a potentially
explosive atmosphere, other than mines susceptible
to firedamp’. However, it has become evident, as a
result of the research project, that at least one commer-
cially available system is capable of being modified
to meet the requirements of ‘intrinsically safety’ in
‘mines susceptible to firedamp’. It recommended that
 the underground coal mining industry encourage the Systems are currently available from several manu-
development of such a modified system. facturers. They differ in matters such as
Although, in some underground coal mining appli-
– infra-red radiation source,
cations, both ends of the ‘open path’ are accessible,
– active and reference wavelengths,
‘single-ended’ access is a feature of other potential
– method of ‘beam splitting’ at the receiver,
applications. In such situations, the positioning of a
– ‘open path’ length,
receiver or reflector at the ‘far end’ of the ‘open path’
– sensitivity to gaseous hydrocarbons,
opposite the transmitter is often difficult or impossible.
– upper limit of detection range,
An alternative arrangement would be for the infra-red
– speed of response and
beam to be reflected back from the coal rib or face to
– explosion protection and certification.
a transmitter/receiver.
It is predicted that, due to a combination of diffuse
and specular reflectance, the reflectivity of the coal
3 ‘TOTAL PATH’ CALIBRATION
will be of the order of one percent. Consequently, for
the intensity of the infra-red signal at the receiver is
In order to facilitate the introduction of gases into the
be similar to that of currently available infra-red ‘open
full length of the measurement path, the basic testing
path’methane detection and measurement systems, the
rig was built to totally enclose the infra-red beam.
output power level of the infra-red source would need
The enclosure comprised a polycarbonate tube of
to be increased by factor of around one hundred. It
123 mm bore, sufficient to accommodate the 100 mm
is envisaged that this could be achieved by substitut-
nominal diameter of the infra-red beam. The polycar-
ing a distributed-feed-back laser diode, with a high
bonate tube formed the upper section of a closed circuit
in-band power, for the incandescent globes that are cur-
into which gases could be introduced via an inlet port.
rently employed as infra-red sources. It recommended
An outlet port was also provided in order to purge gases
that the industry support research into the applicabil-
from the closed circuit. Differential pressure between
ity of laser diodes for remote methane measurement
the gas in the circuit and the outside air was limited
in underground coal mines.
to ±5 kPa and monitored by a Dwyer Model 2300
Another avenue of investigation that is commended
Magnahelic Pressure Gauge of ±5 kPa range.
for support is the issue of the full spatial resolution
For the purpose of ‘total path’ calibration, the
of methane concentration or, at least, resolution of the
closed circuit was filled with two percent methane
methane concentration within small, discrete ranges
‘calibration gas’. This was confirmed by the methane
along the total path. It is envisaged that this might be
concentration displayed on the computer monitor.
achieved by scanning the laser diode beam across the
ribs & face of a section of the mine and then combining
the data in a tomagraphic-like algorithm.
4 SENSITIVITY TO AIRBORNE COAL DUST

2 INFRA-RED ‘OPEN PATH’ METHANE
DETECTION AND MEASUREMENT
SYSTEMS
In essence, the system consists of a transmitter and
a receiver, which may be separated by distances ran-
ging from 500 mm to more than 200 metres, and which
are optically aligned. The transmitter projects an infra-
red beam through the area under surveillance to the
receiver where the beam is split and photo-detectors
measure its intensity at frequencies corresponding to
the active, or sample, wavelength and the reference
wavelength. The quotient of the active signal and the
reference signal provides an output that is proportional
to the total hydrocarbon gas content in the beam path.
Because it is an ‘open path’ system, what is being
detected and measured is the mean gas concentration
along the entire path length. Consequently, the out-
put is a function of gas concentration times the length
of the surveillance path and is commonly expressed
in the units of ppm · m (parts per million × metres) or
LEL · m (lower explosive limit × metres).
Infra Red Monitor is designed in such a way that
when the intensity of the infra-red ‘open path’ beam
is reduced by 45%, it ceases to output a 4–20 mA
signal that is a linear function of the methane con-
centration and instead outputs a nominal 2 mA ‘beam
block’ signal. This is intended to prevent possible erro-
neous readings caused by a weak infra-red signal at the
Receiver.
The effect of the presence of airborne coal dust
in the beam is to reduce its intensity. Consequently,
a series of experiments was undertaken to determine
the concentration of airborne coal dust that would give
rise to ‘beam block’. Finely milled coal dispersed in
the air was used to simulate airborne coal dust. Two
grades were employed: Type 1, approximately analo-
gous to inspirable dust and Type 2, very finely ground
to simulate respirable dust.
Gravimetric sampling was employed to measure the
concentration of dust in the ‘open path’ beam and,
hence, determine the concentration which reduced the
intensity of the infra-red signal by 45% and so resulted
in the monitor indicating ‘beam block’.

521

Copyright © 2006 Taylor & Francis Group plc, London, UK

 4.1 Determination of dust concentration by
gravimetric sampling
The gravimetric sampling protocol followed the rele-
vant standards, Australian Standard AS 2985–1987
Workplace atmospheres – Method for sampling and
gravimetric determination of respirable dust and AS
3640–1989 Workplace atmospheres – Method for sam-
pling and gravimetric determination of inspirable dust.
Essentially, a known volume of dusty air was drawn
from the ‘open path’ through a pre-weighed filter
membrane and the weight of dust determined from
the increase in weight of the membrane. The dust con-
centration could then be obtained from the quotient of
the weight of dust and the volume of air.
The procedure used to determine the critical coal
dust concentration to just ensure beam block when
using Type 1 (inspirable) dust was essentially the same
as that used for Type 2 (respirable) dust. The minimum
concentrations of airborne dust in the ‘open path’beam
that affected its functionality by causing ‘beam block’
were 300 mg per cu.m for ‘respirable’ coal dust and
500 mg per cu.m for ‘inspirable’ coal dust.
To put these figures into context, the upper limit
for airborne respirable coal dust at the working place
in New South Wales underground coal mines is 3 mg
per cu.m while the lower explosive limit for airborne
coal dust is of the order of 30 g per cu.m.
5 WATER
5.1 Water films
Testing was undertaken to determine whether water
films that could settle on the external optical surfaces
of infra-red ‘open path’ methane detection and mea-
surement systems in underground operation would
affect their functionality.
In order to generate a water film, a sub-vertical sheet
of Robax glass-ceramic was positioned so that it inter-
sected the axis of the infra-red ‘open path’beam. Water
was arranged to flow down the sheet in such a way that
the water film occupied the whole of the cross section
of the beam.
The testing confirmed that the functionality of the
monitor was unaffected by the presence of the water
film and, in particular, that ‘beam block’ was not
instituted.
5.2 Airborne water droplets
An experiment was undertaken to determine whether
water sprays, which are widely used in underground
coal mines to suppress dust, would affect the function-
ality of the monitor.
The testing arrangement, utilising the basic testing
rig from which it will be observed that the infra-red
‘open path’ beam was arranged to pass through two
522
Copyright © 2006 Taylor & Francis Group plc, London, UK
‘portholes’ of 150 mm nominal diameter situated on
the transverse axis of a 1500 mm high, 650 mm diam-
eter cylindrical enclosure. The enclosure contained a
water spray generated by a modified air atomising
nozzle.
The spray characteristics were as follows.
Flow rate: 500 mL/m
Spray droplet volume median diameter (VMD):
45 µm
Length of spray cone traversed by infra-red beam:
325 mm
The testing confirmed that the functionality of the
monitor was unaffected by the presence of airborne
water droplets and, in particular, that ‘beam block’ was
not instituted.
6 DIFFERENTIAL VIBRATION TESTING OF
TRANSMITTER/RECEIVER SYSTEM
It has been reported (Jessel 1998) that, in some cir-
cumstances, differential movement of the Receiver
relative to the Transmitter may result in interruption
to the operation of an infra-red ‘open path’ methane
detection and measurement system.
In order to confirm whether this was the case for
the system, the transmitter/tilt assembly was bolted
to the shaking table while the Receiver was mounted
in a fixed location and optically aligned with the
Transmitter at a path length of 3.22 metres.
The testing confirmed that the functionality of
the system was unaffected by the differential
vibration of the intensity and duration specified in
AS/NZS 61779.1-2000. It must be understood, how-
ever, that if the system is mounted on an item of
underground machinery, careful attention must be paid
to the design of the mountings to ensure that dif-
ferential movement does not result in intermittent
misalignment which could affect the functionality of
the system. This is particularly so when the path length
is greater than that of the test system.
7 POTENTIAL DEPLOYMENT IN
AUSTRALIAN UNDERGROUND COAL
MINES
The deployment of the System in hazardous zones New
South Wales (NSW) underground coal mines would be
subject to the requirements of the Coal Mines (Gen-
eral) Regulation 1999 and Coal Mines (Underground)
Regulation 1999 made under the NSW Coal Mines
Regulation Act 1982.
The Coal Mines (Underground) Regulation 1999
(pp 13–14) defines a hazardous zone as
(a) a return airway in a mine, or
(b) that part of an intake airway in a ventilation district
in a mine that is on the return side of such points as
 are 100 metres outbye the most inbye completed
line of cutthroughs or 100 metres from, and on the
intake side of, a longwall or shortwall face, or
(c) a part of a mine in which there is a methane con-
centration of 1.25% or greater in the general body
of air, or
(d) a part of a mine specified as a hazardous zone.
In addition, the DMR classification of hazardous
areas (Department of Mineral Resources 1996) is
more proscriptive than that of Australian Standard AS
2430.1-1987. It states (sec. F, p. 9f ) the following
definitions.
Zone 0
1 All underground areas where no mechanically
induced ventilation exists.
2 Any area where firedamp exists in concentrations
greater than 1.25%.
Zone 1
1 Any place within twenty (20) metres of a face sub-
ject to mechanically induced ventilation able to
continuously dilute firedamp to below 1.25%.
2 All return airways.
3 Any area in [which] a mixture of firedamp is likely
to occur.
Zone 2
Any area further outbye than twenty (20) metres
by less than one hundred (100) metres from a face
subject to mechanically induced ventilation able to
continuously dilute firedamp to below 1.25%.
The Coal Mines (General) Regulation 1999
includes an ‘approvals’ requirement that is applied
to non-portable items of electrical apparatus in haz-
ardous zones. This implies, in practice, assessment
and approval by an accredited assessing authority. It is
understood, however, that it is the policy of the NSW
Department of Primary Industry (formerly Mineral
Resources, DMR) to afford approvals only to those
items of equipment that have been tested and certified
as being in conformity with the appropriate Australian/
New Zealand Standards.
The general conclusion, based upon the detailed
assessment of the PLMS GD4003, is that infra-red
‘open path’ methane detection and measurement sys-
tems are, in principle, ‘Certifiable’ for use in explosive
gas atmospheres in Australian underground coal mines
and ‘Approvable’ for use in hazardous zones in New
South Wales coal mines.
The general conclusion, based upon the detailed
assessment of the System, is that infra-red ‘open path’
methane detection and measurement systems are, in
principle, ‘Certifiable’ for use in explosive gas atmos-
pheres in Australian underground coal mines and
‘Approvable’ for use in hazardous zones in New South
Wales coal mines.
523
Copyright © 2006 Taylor & Francis Group plc, London, UK
8 CONCLUSIONS REGARDING CURRENT
SYSTEMS AND THEIR FURTHER
DEVELOPMENT
As a result of the study, it is concluded that, sub-
ject to safety modifications, at least one commercially
available infra-red ‘open path’ methane detection and
measurement system is suitable for use in certain
applications in Australian underground coal mines.
However, further development of such systems is
required before they will be useful in all potential
applications.
8.1 ‘Single ended’ systems
Although, in some underground coal mining appli-
cations, both ends of the ‘open path’ are accessible,
‘single-ended’ access is a feature of many potential
applications. In such situations, the positioning of a
receiver or reflector at the ‘far end’ of the ‘open path’
opposite the transmitter is often difficult or impossible.
An alternative arrangement would be for the infra-
red beam to be reflected back from the coal rib or face
to a transmitter/receiver.
It is predicted that, due to a combination of diffuse
and specular reflectance, the reflectivity of the coal
will be of the order of one percent. Consequently, for
the intensity of the infra-red signal at the receiver is
be similar to that of currently available infra-red ‘open
path’methane detection and measurement systems, the
output power level of the infra-red source would need
to be increased by factor of around one hundred. It is
envisaged that this could be achieved by substituting
a distributed-feed-back laser diode, with an in-band
power of around 5 mW, for the incandescent globes
that are currently employed as infra-red sources.
The Ventilation Research Group in the University
of New South Wales (UNSW) School of Mining
Engineering has begun research into the applicabil-
ity of laser diodes for remote methane measurement
in underground coal mines. It recommended that
the underground coal mining industry support such
research.
8.2 Resolution of methane concentration along the
path of the infra-red beam
Another avenue of investigation that is commended
for support is the issue of the full spatial resolution
of methane concentration or, at least, resolution of the
methane concentration within small, discrete ranges
along the total path. It is envisaged that this might be
achieved by scanning the laser diode beam across the
ribs & face of a section of the mine and then combining
the data in a tomagraphic-like algorithm.
 REFERENCES
Department of Mineral Resources. 1996. Electrical com-
pendium of guidance notes to mines and manufacturers,
for coal mines in New South Wales, Australia, MDG
2001. Sydney: New South Wales Department of Mineral
Resources.
Dräger. 1998. Polytron Transmitter IR Ex – Instructions for
Use, 9th edn. Lübeck: Dräger Sicherheitstechnik.
Fowler, J.C.W. 2001. The Displacement of Methane from the
Goaf into the Working Place as a Result of Wind Blasts in
Underground Coal Mines, UMRC Report RR8/01. Syd-
ney: The University of New South Wales Mining Research
Centre. ISBN 0 7334 1816 3.
Jessel, W. 1998. ‘OpenPath’ detectors – photoelectric barriers
for detecting flammable gases. Dräger Review, no. 81:
34–37.
524
Copyright © 2006 Taylor & Francis Group plc, London, UK
Millipore. 2001. Laboratory Catalogue 2001–2002. Bedford,
MA: Millipore Corporation.
PLMS. 1998a. Installing and Operating the GD4003
Enclosed Path Gas Detector, issue 1.4. Plymouth: Pre-
cision Light Measurement Systems.
PLMS. 1998b. Instructions for the Use of the PLMS GCK400
Gas Check Kit, issue 2.0. Plymouth: Precision Light
Measurement Systems.
Redman, M. 2000. Application of Infra-Red Methane Detect-
ors in Underground Coal Mines. unpublished BE (Hons)
Thesis, The University of New South Wales School of
Mining Engineering.
Schott Glass. 1999. Schott Technical Glasses. Mainz: Schott
Glass.
 10. Mine fans

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Selection of mixed-flow fans for PT Freeport’s DOZ Mine,

West Papua, Indonesia

I. Duckworth, I. Loomis, K. Karmawan & C. Chmura

PT Freeport Indonesia, West Papua, Indonesia

K. Lownie
Howden Australia, New South Wales, Australia

ABSTRACT: As part of the analyses to advance the DOZ block cave mine from 35,000 to 50,000 tonnes
per day, the need for additional airflow and ventilation infrastructure was identified. The primary ventilation
for the DOZ Mine is currently provided by five centrifugal fans located on the surface in the Dom valley
(4,000 m above sea level). The DOZ Mine is approximately 3,100 m above sea level; hence the exhaust air is
drawn 900 m up through ventilation shafts to the fans. In order to increase ventilation to the DOZ Mine, the
solution selected was exhaust portals at approximately mine elevation in the valley wall above the mill site.
Faced with limited-space at the portals, and noise considerations, a surface fan installation was discounted and
underground fans were selected. The fans chosen for the DOZ ventilation project will be required to move a
total airflow of approximately 780 cubic meters per second and develop a pressure exceeding 2.5 kilo-Pascals
(3.4 kilo-Pascals sea level equivalent); additionally, the projected life of the fans will exceed 25 years. Careful
consideration of the available fan designs and potential parallel fan installations led to the selection of twin
parallel impulse-bladed (mixed-flow) fans rather than the more typical (for mining) reaction-bladed (vane-axial)
fans. This paper describes the selection process and design for the application of an established fan technology in
a novel underground mining application. The new DOZ exhaust fans are scheduled to be in service in April 2006.

1 INTRODUCTION/PTFI BACKGROUND

The PT Freeport Indonesia (PTFI) Ertsberg District

mine site is located in the Sudirman Mountain Range
of West Papua (formerly Irian Jaya). This is the
Eastern-most province of Indonesia that occupies the
Western end of the island of New Guinea. The general
geography is shown in Figure 1.
PTFI presence in the district originated in 1967 and
the area has been in continuous development, mining
and expansion since that time. The mine site is located
in the mountains, between 2,500 and 4,000 m above
sea level, approximately 100 km north of the South-
west coast of West Papua.
Since the initial development of the Ertsberg open
pit, the GBT block cave, IOZ block cave and the
DOZ open stope mines have been developed, mined Figure 1. Location of PTFI’s mining operations.
and closed. Current production comes from the DOZ
block cave and the Grasberg open pit mines. Through of this paper, approximately 42,000 tonnes per day is
the life of the mining area, the total production rate being produced from the DOZ block cave mine.
has increased from 15,000 tonnes per day in 1978, to An expansion project is currently underway to bring
60,000 tonnes per day in 1992, to the present produc- the designed production rate from the DOZ block cave
tion of around 250,000 tonnes per day. At the time mine up above 50,000 tonnes per day. This project is

527

Copyright © 2006 Taylor & Francis Group plc, London, UK

 intended to increase productivity of the DOZ orebody
with the possibility of combining the adjacent ESZ
orebody within the overall block cave plan. Two key
elements of the expansion are the construction of a sec-
ond primary underground crusher and the expansion
of the exhaust ventilation system.
2 DOZ/ESZ BLOCK CAVE MINE
2.1 50,000 tonnes per day expansion
During 2004, a study to increase the DOZ production
from 35,000 to 50,000 tonnes per day was undertaken.
At that time the mine had been consistently producing
in excess of 40,000 tonnes per day, however such a rate
was considered unsustainable as the mine progresses
further to the West. The expansion study considered
numerous factors associated with increasing the daily
production, including expansion of the truck haulage
system, provision of a second primary crusher, elec-
trical power upgrades, and increase in the capacity of
the ventilation system.
Figure 2. Plan of proposed West side ventilation infrastructure with Extraction level drifting (250 m grid spacing).
528
Copyright © 2006 Taylor & Francis Group plc, London, UK
From the perspective of the ventilation system, the
study showed that the DOZ exhaust capacity needed
to be significantly increased and an additional (fourth)
primary intake drift would be required.
Figure 2 shows a plan view of the West side of the
DOZ Mine as scheduled towards the end of the mine
life. Proposed new ventilation infrastructure is shown
in bold, with the new twin exhaust adits and fans clearly
identified.
The DOZ Mine presently has five main surface
fans (called the Dom Fans) located approximately
900 m above the Extraction level. These fans are
each 3.2 m diameter, 725 kW heavy-duty centrifugal
fans manufactured by Howden Australia. In total, the
fans handle approximately 1,180 m3 /s of air, which is
drawn up through a series of 6.5 m diameter ventilation
raises. Ventilation modeling predicts that a signifi-
cant increase in airflow will be required to support
life-of-mine expansion to the West. Various scenarios
were considered and modeled to increase the mine air-
flow capacity. The existing mine exhaust infrastructure
was determined to be fully maximized and the case of
 Table 1. Fan requirements for 50,000 tonnes per day. new location a subsurface fan installation was selected
for the following reasons:
Pressure Quantity Air Power
Fan (kPa) (m3 /s) (kW) – There will be minimal portal construction. The
portals will be blasted from inside the mine and
Existing Fans supported with shotcrete and bolts.
Dom #11 3.1 240 744 – There will be no requirement for access to the por-
Dom #12 2.7 210 567 tals from the outside. This allows the portals to be
Dom #13 2.6 195 507 located at the most suitable location, and saves on
Dom #14 2.7 210 567 the cost of building and maintaining a road.
Dom #15 2.6 195 507
– The portals will be directly above existing facilities.
New DOZ 2.5 780 1,950
Placing fans underground will eliminate surface
TOTAL – 1,830 4,842 noise, and only a low velocity discharge of exhaust
air will be generated that is expected to mix and
disperse rapidly near the portals.
new twin exhaust adits with primary underground fan Another factor in the selection of underground fans
installations was selected. was to gain experience with that type of configura-
Drifts were chosen over shafts for two main reasons. tion, which will be a necessity for future proposed
The first is that horizontal ventilation infrastructure operations.
will not be threatened by the cave profile. The existing
Dom fans, power, access and associated ventilation
raises will be impacted by the cave crack line towards 3.2 Fan type
the end of the ESZ Mine life. The second reason is that The selection of an underground site for the DOZ main
drifting can be developed in house, more rapidly and fans meant that a conventional centrifugal fan design,
is less costly than vertical infrastructure. which has a history of over 20 years service at PTFI
Table 1 provides a list of the predicted fan operat- was not a practical solution.
ing characteristics for the existing and proposed fans Confronted with an underground main fan location,
during peak demand under 50,000 tonnes per day. In PTFI ventilation engineers investigated the available
this case each Dom fan is operating at a lower airflow technology and design options to achieve the desired
than present due to the operating point being higher airflow quantity and pressure (approximately 3 kPa
on the fan curve (the fans will be required to meet a sea level equivalent). Two available technologies were
higher pressure demand). assessed and compared for application; these were
The ventilation study has been detailed in a previous vane-axial and mixed-flow designs.
publication (Duckworth et al., 2005). From the outside, mixed-flow fans look similar to
conventional vane-axial fans. The fans can be supplied
as open inlet designs where the flow enters and leaves
3 FAN DESIGN ALTERNATIVES the fan axially. The difference between mixed-flow and
vane-axial fans lies in the way pressure is generated.
Following the determination of the required duty
for the new DOZ fan system a number of design 3.2.1 Vane-axial fans (reaction-bladed axial fans)
alternatives were considered. The conventional axial (or reaction) flow fan is a well
tested and efficient design. Fans with acceptably high
efficiency have been produced and used since the
3.1 Surface vs. underground installation
1930’s (McPherson 1992). The theory of these fans
The current Dom main fans are surface installed cen- is based on a plate moving, with an angle of attack,
trifugal units. These fans are located in the Dom through a fluid. As the angle of attack of the plate is
valley near the partially developed, but currently increased there is development of lift by an increase in
moth-balled Dom block cave mine. The elevation of the air pressure below the plate and a decrease in the air
these fans is nearly 4,000 m, compared to the eleva- pressure above it. Since the moving plate is fixed in a
tion of the primary mine levels being approximately single plane, the lift is manifested by the displacement
3,100 m. The available space in the Dom Valley, com- of air perpendicular to the plane of movement. Sev-
bined with the capability of the centrifugal design to eral design enhancements have been implemented that
develop adequate pressure (approximately 4 kPa sea increase the performance and efficiency.These include
level equivalent) made the application of that design the use of an aerofoil blade, rather than a flat plate; the
ideal. introduction of a twist along the length of the blade so
The new DOZ exhaust drifts, however, will portal that the relative angle of attack is maintained for the
through a steep valley wall above the mill site. In this entire blade, and the use of straightening vanes (stators

529

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 3. Typical vane-axial fan impeller.
or impulse plates) to control flow rotation behind the
impellor.
Vane-axial fans require accurate aerofoil blade pro-
files to deliver high efficiencies (see Figure 3). Fan
total efficiencies up to 92% can be achieved at the
Reynolds Numbers normally applicable to most indus-
trial fan applications, but more typically, the efficien-
cies are in the order of 87%. The blades are always
attached to a cylindrical hub such that the annular cross
sectional area of the impeller passage is constant.
These fans are considered to be “reaction-bladed” as
the flow is initially accelerated and then slowed down
(diffused) through the blade passage. Accurate blade
profiles are essential to the efficiency of this process,
and the greater proportion of the static pressure rise is
generated across the actual impeller. Some additional
static pressure is recovered from the rotational flow
component by the downstream fixed stator vanes and
the fan diffuser section.
Reaction-bladed axial-flow fans are generally con-
trolled by speed variation or blade pitch adjustment.
The use of inlet radial vane controls is not acceptable
as the risk of stall operation is high.
The application of reaction-bladed axial-flow fans
is normally restricted to relatively clean air situations.
In these circumstances, the fan will deliver high effi-
ciencies. Reaction-bladed axial-flow fans are often not
suitable for those mine ventilation applications with
heavily contaminated exhaust air (typically dust and
moisture). Severe erosion and build-up on the fan
blades can occur after a relatively short period of time
(see Figure 4). Fans that have even small amounts of
build-up and blade wear will start to fall short of the
original performance specifications.
530
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 4. Build-up and wear on a mine vane-axial Fan.
Figure 5. Fabrication of a typical Howden mixed-flow fan
impeller.
3.2.2 Mixed-flow fans (impulse-bladed fans)
Mixed-flow fans are “impulse-bladed”. Instead of
accurately profiled blades, the impeller makes use of
constant thickness curved steel plate blades which are
fixed (welded) to a conical hub (see Figure 5).
The annular cross sectional area through the blade
passage is reduced as the flow progresses across the
impeller. All of the energy imparted to the air by the
impeller is reflected by increased momentum (veloc-
ity). For this reason, accurate blade profiles are not
required. The reduction in annular area accelerates the
flow axially and radially through the blade passage.
This process produces a significant radial flow com-
ponent, hence the term “mixed-flow”. These effects,
together with the high rotational component imparted
by the impeller, combine to produce a high energy flow
leaving the impeller, which is recovered by the down-
stream stator vanes and fan diffuser section. Fan total
efficiencies up to 89% can be achieved, with around
85% being more common for a well-specified instal-
lation. Impulse-bladed mixed-flow fans are generally
 Table 2. Fan comparison.
Vane-axial fans Mixed-flow fans
Pros
High initial efficiency Durability
Low cost Aerodynamic stability
Track record in mining Inlet vane controllers
Compact No wear/build-up concern
Pressure generation
Cons
Blade wear & build-up Cost
Durability Larger size
Stall (instability) Limited application in mining
Flow control Low efficiency at low pressure
controlled by inlet radial vane controls similar to those
used for centrifugal fans. As with all fan types, speed
control is also an option.
3.2.3 Reasons for selection of mixed-flow fans at
PTFI
The pros and cons associated with both vane-axial
and mixed-flow fans were evaluated (see Table 2). The
information was based upon bids and feedback from
most of the international vendors offering large axial
fans to the mining community.
The primary consideration for selecting mixed-flow
fans was durability. The specification for these fans
was a life of at least 25 years. Mixed-flow fans typically
consist of a massive welded design and are inher-
ently much less susceptible to wear, fatigue, and blade
impact damage when compared to vane-axial fans.
Other considerations are discussed in the following
sections.
3.2.3.1 Pressure generating capacity
Impulse-bladed mixed-flow axial fans deliver substan-
tially higher pressure coefficients than vane-axial fans
(more than double). This allows lower impeller tip
speeds to be used, and results in lower wear rates, rotor
stress levels and noise generation.
3.2.3.2 Impeller wear
Vane-axial fans are sensitive to wear of the blade pro-
file. Even quite minor erosion of the blade noses will
reduce efficiency by 5% to 10% with heavy wear
causing typical efficiency reductions of 20% to 50%.
3.2.3.3 Flow control
Vane-axial fans require either variable pitch in-motion
impellers or variable speed drives to achieve dynamic
flow control through the fans. There were concerns
regarding the maintenance and long term durability
of both of these methods of control. The inlet vane
controllers used on the mixed-flow fan will be similar
to the existing controllers used on the Dom centrifugal
fans and are well proven at PTFI’s operations.
531
Copyright © 2006 Taylor & Francis Group plc, London, UK
3.2.3.4 Fan stability
A mostly ignored consequence of blade wear and
build-up on reaction-bladed vane-axial fans is a reduc-
tion in the pressure generated by the fan. This is impor-
tant because the reduction in peak pressure will result
in a reduced margin to stall. It is probable that some
of the “mysterious” vane-axial fan impeller failures
that happen periodically are caused by stall operation.
The high pressure rise capability of mixed-flow fans
is insensitive to blade wear.
3.3 Number of fans
During the evaluation of the options provided by sev-
eral fan vendors, various parallel fan configurations
were considered. This included both two and four par-
allel fan configurations, as well as mixed-flow fan
design compared to traditional vane-axial design.
Since there are two main exhaust drifts, the two
fan option places one fan in each drift, while the four
fan option would have two fans installed in each drift.
The four fan option had several advantages, includ-
ing smaller and lighter motors, increased redundancy
and reduced fan starting power. However, this option
was discarded due to concerns with the aerodynamic
stability associated with four parallel vane-axial fans.
The four fan mixed flow option was discarded based
on higher cost and the complexity associated with
underground construction.
3.4 Vendor selection
A detailed specification was developed and submitted
to all known qualified vendors for large vane-axial
and mixed-flow fans. The specification included such
details as:
– Fan operating characteristic (nominal, peak and
absolute fan total pressure limit). This excluded fan
losses (requirement of the vendor to provide).
– Maximum negative pressure that the fan installation
must start against.
– Fluid details such as temperature, humidity, eleva-
tion and dust and diesel particulate loading.
– Fan configuration shown on plans. The vendor
was required to provide details such as fan type,
blade fabrication, potential instability associated
with parallel fan configuration, size and weight of
components, all fan electrical systems, starter and
control details, and commissioning spares.
Vendors were asked to expand on issues such as
parallel fan instability (“double-loop”) and the likely
impact and solutions associated with fan blade wear.
Two fan vendors were ultimately short listed for
detailed consideration. It is worth noting that ven-
dors who did not address the specification fully were
rejected (regardless of price). The two finalists were
 carefully considered prior to awarding the contract. larger chambers and heavier lifting, it was recog-
The details for these two fans are provided in Table 3. nized that vane-axial fan components would be less
Ultimately a twin mixed-flow fan installation from forgiving to mishandling and set-up. There was also
Howden Australia was selected. The fan curve (at a confidence that the chambers could be developed.
mine air density of 0.84 kg/m3 ) for one MF 3500 How- Although large subsurface fans have not been con-
den fan is shown in Figure 6. structed at PTFI, similar experience has been gained
Factors considered during final selection included: in developing crusher chambers and constructing
truck loading chutes.
– Construction and service. The components of the
– Fan power and motor specifications. Sufficient
mixed-flow fan design are larger and heavier than
power capacity to drive twin 1,600 kW motors was
equivalent axial-flow fans. Although this means
determined to be available from existing facilities
(with minor upgrade) using auto-transformers to
start. This was critical and ultimately allowed the
Table 3. Details of the two short-listed fans. four fan options to be rejected. Based on previous
Vane-axial Mixed-flow
experience, stainless steel motor coolers and covers
Parameter (Other Vendor) (Howden) were chosen to prevent corrosion.
– Potential development of the MLZ Mine in the
Impellor diameter 3.2 m 3.5 m future. The MLZ deposit is located below the
Impellor material Aluminum – Steel existing DOZ Mine. Pre-feasibility analyses have
Carbide Coat determined that the fan pressure requirement will
Speed 880 rpm 590 rpm be higher than that required for the DOZ Mine. The
Tip speed 147 m/s 108 m/s mixed-flow design will better support these long
Flow control Variable pitch Inlet vortex term goals based on the higher pressure profile and
Motor size 1,600 kW 1,600 kW longer expected fan life relative to the axial-flow
Base cost for 2 fans US$1.2 M US$1.5 M
Min operating cost US$830 k/yr US$965 k/yr
design.
Max operation cost US$1,480 k/yr US$1,325 k/yr – Vendor Track Record. The Howden mixed-flow
axial fan was entirely developed in Australia (almost
30 years ago) in response to the need for a robust
high pressure fan for the gas cleaning systems being
adopted for large power generation boilers. Howden
Australia has supplied more than sixty large mixed-
flow fans. Prior to final selection, a trip was made to
visit the offices and fabrication facilities of Howden
Australia. A field trip was conducted to inspect fans
that have been in continuous service since 1982, and
to hold discussions with Maintenance personnel.

4 FAN LAYOUT

At the time of this writing (November 2005) two fan

Figure 6. Howden MF 3500 fan curve.
chambers have been constructed to house the fans. The
schedule is to install the fans during January–March
2006, with commissioning of both fans at the start of
April, 2006.
Figure 7 shows a Vulcan plot of the fan chambers.
These chambers are located within 5.5 × 6.5 m
exhaust drifts, which are accessed via cross cuts con-
nected to three primary intake airways. The motor con-
trol center/electrical room is located off one of these
cross cuts. Both chambers were developed in horizon-
tal lifts from the Cross Cut #6 side by blasting the top
drift first using a ramp constructed from the broken
rock. Permanent support for the chambers consists of
grouted threadbar with 75 mm of shotcrete.
Figure 8 shows the cross section of one of the fans,
with the associate chamber shown in Figure 9. Each

532

Copyright © 2006 Taylor & Francis Group plc, London, UK

 chamber has a 15 tonne overhead monorail allow-
ing initial installation and future maintenance of the
fans. The ultimate height of the chamber was deter-
mined by the clearance required for the monorail crane,
hook and hanging impeller to move easily over the
motor and inlet bell. For a fan of this size it is not
advisable to remove the fan motor (10 tonne) or vane
controllers (5.5 tonne) to access and replace other crit-
ical components. Furthermore, for ease of assembly it
is critical that adequate clearance is available to drop
components like the impeller into place from a vertical
position. The monorail can be seen to project signif-
icantly in front of the motor pedestal which services
the staging area for the fan components.
The general slope of the fan chambers is 3% to facil-
itate drainage, however the fans and monorail are both
designed level. The fan layout is such that the evasée
discharges directly into the drift rather than the much
larger chamber. This design was necessary to minimize
dynamic losses associated with the fans.
Figure 7. Vulcan plot of fan chambers and associated drifts.
Figure 8. General arrangement drawing of the mixed-flow fan.
533
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 8 provides additional detail showing such
components as the mandoors, pedestal arrangement,
safety screens, and anchor bolts. Not shown are addi-
tional sets of stairs and platforms required for monorail
maintenance. The interior of the fan housing features a
walkway to allow easy access to the non-drive end fan
bearing. A second monorail is provided internal to the
fan housing which allows the 7.9 m shaft to be pushed
through the impeller, thus enabling impeller removal
without having to take out the motor, vane controls or
shaft.
5 BUDGET
The budget for purchasing and installing the two
mixed-flow fans, inclusive of all electrical require-
ments is $4.2M (US). This budget includes:
• Howden Australia Scope
– 2 × Howden MF 3500 fans.
– 3 × 1,600 kW Toshiba motors (1 spare) with
stainless steel heat exchanger tubing.
– Full electrical systems from 4,160 V incoming to
the fans.
– Two auto transformer starters.
– Full control including inlet vanes, vibration,
pressure and flow.
– Power factor correction.
• Fan Construction
• Power Plant & Substation Upgrade
• Motor Control Center (non Howden)
• Engineering & Project Management
Other costs not shown are those associated with
pneumatic equipment doors in both cross cuts (to
isolate the exhaust drifts) and the civil costs for the
chamber and drift development and ground support.

Figure 9. Fan chamber and motor pedestal construction.
Figure 10. Fabrication of fan stator section – Thailand.
6 PROGRESS
Figure 9 shows a photo of one of the fan chambers at
the start of November, 2005. Each chamber has min-
imum dimensions of 13 m high by 8 m wide by 33 m
long. The steel work for the motor pedestal can be seen
in the foreground. The floor of the chamber is a poured
slab which is keyed directly into bed rock. Each of four
fan pedestals and one motor pedestal is anchored into
the floor using a grid of 2 m long grouted rock anchors.
Fabrication of the fans occurred between June
and December 2005. Critical welding, such as fan
impellers, is being conducted in Brisbane, whereas
534
Copyright © 2006 Taylor & Francis Group plc, London, UK
the fan housings are being fabricated in Thailand.
Figure 10 is a progress photo from the Asian welding
facility used by Howden. In this case the stator sec-
tion of the fan housing is undergoing final welding.
Figure 5 shows a typical fan hub being prepared for
welding at the fabrication facility in Brisbane.
7 CONCLUSIONS
The installation of new exhaust fans at PTFI will result
in approximately 60% additional ventilation capacity
through the mine. This step change is required to allow
the DOZ Mine to increase present ore tonnage to above
50,000 tonnes per day.
Following an international bid process 3.5 m diam-
eter mixed-flow fans were ultimately selected for
installation in twin subsurface chambers. The ven-
dor for the fans is Howden Australia. Although less
common in mining, the selected fans have a proven
record of operation throughout Australia, particularly
in power stations. Mixed-flow technology was selected
over traditional vane-axial based on such issues as gen-
eral durability, wear resistance of blades, integrated
flow control, aerodynamic stability and high pressure
characteristic.
The entire exhaust fan installation has been
designed to allow for ease of construction and mainte-
nance. While this has resulted in larger chambers, and
higher construction costs, it is believed that this will
be more than offset in the long term due to improved
availability.
ACKNOWLEDGEMENTS
The authors would like to thank the management team
of PTFI for supporting the publication of this paper.
REFERENCES
Duckworth, I.J., Karmawan, K., & Chmura, C. 2005, Expan-
sion of the DOZ Mine Ventilation System. Society of Min-
ing, Metallurgy and Exploration 2005 Annual Meeting,
Salt Lake City, Utah.
McPherson, M.J. 1992. Subsurface Ventilation and Envi-
ronmental Engineering. London: Chapman and Hall.,
905 pp.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Fan selection for large-opening mines: Vane-axial or propeller

fans – which to choose?

R.B. Krog & R.H. Grau III

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute
for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, Pennsylvania, USA

ABSTRACT: The National Institute for Occupational Safety and Health (NIOSH) has investigated the unique
ventilation requirements of large-opening mines to help identify and evaluate the effectiveness of various fan
types to improve the ventilation and air quality in the underground workplace. Large-opening mines, with their
low airflow resistance factors, can be ventilated with free-standing auxiliary fans because airflow patterns in these
mines are primarily controlled by airflow momentum. The flow characteristics of both vane-axial and propeller
fans were investigated and tested at four large-opening mines to assess the effects of fan location on recirculation
and entrainment. Each fan type has its own airflow, entrainment and penetrating airflow characteristics, and
operating costs that are advantageous for specific applications. Either fan type can be used for most auxiliary
applications. However, this research has shown that the optimum placement and use criteria for propeller fans
differ from those promulgated by the U.S. Bureau of Mines (USBM) for vane-axial fans Brechtel et al. (1985).

1 INTRODUCTION efficient face ventilation; however, similar investi-

Moving adequate fresh air volumes in large-opening,
room-and-pillar mines presents several challenges due
to the large open-space volume of the mine and the
extremely low airflow resistance 0.0005 N·s2 /m8 (4.5 ×
10−13 in·min2 /ft6 ). The low mine resistance is caused
by the large 12 m × 8 m (40 ft × 27 ft) entries. A well
developed underground stone mine can have a void
volume of several million cubic meters (100 million
cubic feet) and require many hours to effect a com-
plete air change. Unlike most underground coal mines
and many metal mines, the resistance to flow in these
large-opening mines is very low. Recognizing this fact,
NIOSH demonstrated the use of low-pressure pro-
peller fans for whole mine ventilation (Grau et al.
2004, Krog et al. 2004). Several large-opening mines
have subsequently installed propeller fans for main
mine ventilation instead of relying on changing natural
ventilation. Mine operators are responsible to meet any
federal guidelines (Title 30 U.S. Code of Federal Reg-
ulations Part 57) regarding placement and installation
of fans.
NIOSH has since focused on the application of pro-
peller fans to regional and face ventilation applications
in large-opening mines. Previous research on venti-
lating large-opening oil shale mines conducted by the
U.S. Bureau of Mines (USBM) in the 1970s and 1980s
tested free-standing vane-axial fans as regional and
face fans (Brechtel et al. 1985, Dunn et al. 1983). The
researchers examined vane-axial fan placement for
gations using propeller fans for auxiliary face and
regional ventilation were required and are discussed
in this paper.
2 FAN AIRFLOW ENTRAINMENT
CHARACTERISTICS
NIOSH researchers conducted studies of auxiliary,
free-standing vane-axial and propeller fans, showing
that each type has different airflow distribution pat-
terns around the fan. An electronic vane anemometer
mounted on an adjustable pole was used to measure
average airflow readings on a grid pattern. A propeller
fan tends to draw air from behind, and entrains the air-
flow only up to 30 m (100 ft) or to the first crosscut, as
the resultant medium-speed airflow expands rapidly.
Conversely, a vane-axial fan, with its lower air quan-
tity and higher fan exit speed, draws minimal air from
behind the fan. However, the vane-axial fan entrains
the airflow for a distance of up to three crosscuts ahead
of the fan because the high-speed airflow takes over
90 m (300 ft) to fully expand. Therefore, propeller and
vane-axial fans have different placement criteria when
used as auxiliary fans.
2.1 Vane-axial fan
Figure 1 shows the vane-axial fan used in the study.The
free-standing vane-axial fan had a diameter of 0.91 m

535

Copyright © 2006 Taylor & Francis Group plc, London, UK

 6.0-7.0 4.0-5.0 2.0-3.0 0.0-1.0
5.0-6.0 3.0-4.0 1.0-2.0
7.3

6.7

5.5

Velocity (m/s)

Height (m)
0-1 4.3

3.0
1-2

1.8
2-3 3-4 4-5 5-6
6-7
0.6
0.6 1.8 3.0 4.3 5.5 6.7 7.9 9.1 10.4 11.6
Width (m)
Figure 1. 0.91 m (36 in) vane-axial fan with 0.58 m (23 in)
reducer. Figure 3. Vertical cross-section A of airflow 28 m (92 ft) in
front of 0.91 m (36 in) vane-axial fan, flow out of page.

6.0-7.0 4.0-5.0 2.0-3.0 0.0-1.0

5.0-6.0 3.0-4.0 1.0-2.0
7.3

6.7
0-1
Velocity (m/s)

5.5

Height (m)
1-2
4.3

3.0
2-3
1.2
3-4
0.6
0.3 1.5 2.7 4.0 5.2 6.4 7.6 8.8 10.1 11.3 12.5
Width (m)

Figure 2. Airflow characteristics for a 0.91 m (36 in)

vane-axial fan with reducer, note recirculation patterns gen-
erated around the second to forth intersections.
(36 in). It was powered by a 19 kW (25 hp) motor, and
mounted with a reducer at the outlet with a discharge
diameter of 0.58 m (23 in). The airflow characteristics
generated by the vane-axial fan are shown in Figure 2.
The patterns indicate that the high exit velocity of
the fan’s reducer [39 m/s (7600 ft/min)] causes air tur-
bulence and entrainment for a distance of over 90 m
(300 ft) in front of the fan. The venturi effect of the
reducer was observed up to 3 meters (10 ft) in front
of the fan, where little air interaction was observed
surrounding the fan’s high-speed airflow.
The airflow through the fan was 10.4 m3 /s
(22,000 cfm). However, due to entrainment effects, the
total airflow 83 m (272 ft) downstream of the fan was
over 17 times greater, at 183 m3 /s (388,000 cfm). The
airflow patterns in the cross-cuts at the 2nd to 4th
intersections in front of the fan were bi-directional
(Fig. 2). High velocity air moving down the main entry
would catch the corners of the pillars and be directed
perpendicularly down the cross-cuts. An eddy formed
Figure 4. Vertical cross-section B of airflow 55 m (180 ft)
in front of 0.91 m (36 in) vane-axial fan, flow out of page.
behind the pillars of the cross-cuts, pulling air into the
main ventilation drift. The net result was little change
in the total airflow moving down the main entry for
the first three intersections, downstream from the fan
(Fig. 2). However, because of the high-speed of the
expanding airflow, there was recirculation with the
surrounding air.
Figure 3 shows the cross-sectional velocity profile
28 meters (92 ft) in front of the vane-axial fan’s out-
let. The airflow is concentrated at the lower half of the
12.5 m × 8.4 m (41.1 ft × 27.4 ft) opening in the cen-
ter of the drift as the high-speed air starts to expand.
The peak air velocity was 6.56 m/s (1290 ft/min),
and the air velocity approaches zero in the upper third
of the cross section. Airflow in the top corners was
below recordable levels.
Figure 4 shows a similar air profile as in Figure 3
recorded 27 m (88 ft) further down the drift. The air
stream is still predominantly flowing along the ground,
and very little airflow was recorded in the upper half of
the drift. The peak velocity was 3.68 m/s (725 ft/min).

536

Copyright © 2006 Taylor & Francis Group plc, London, UK

 4.0-5.0 2.0-3.0 0.0-1.0
3.0-4.0 1.0-2.0
7.3

6.1
1-2 2-3 3-4

Velocity (m/s)
4.9

Height (m)
4.3
1-2
3.0

1.8
3-4
4-5 0.6
0.6 1.8 3.0 4.3 5.5 6.7 7.9 9.1 10.4 11.6
Width (m)
Figure 5. 2.4 m (8 ft) propeller fan used in study.
Figure 7. Vertical cross-section A of airflow 25 m (82 ft) in
front of 2.44 m (8 ft) propeller fan, flow out of page.

2.0-3.0 1.0-2.0 0.0-1.0

7.3

6.7
1-2 2-3
Velocity (m/s)

5.5

Height (m)
4.3
2-3
3.0

1.2

0.6
0.3 1.5 2.7 4.0 5.2 6.4 7.6 8.8 10.1 11.3 12.5
Width (m)
Figure 6. Airflow characteristics for a 2.44 m (8 ft) pro-
peller fan, note recirculation patterns generated at the second Figure 8. Vertical cross-section B of airflow 52 m (171 ft)
intersection. in front of 2.44 m (8 ft) propeller fan, flow out of page.

Figure 4 shows that entrainment will still occur fur-
ther down the drift. The airflow interaction with the
corners of the pillars varies with the height off the
floor because of the different velocities along the ribs.
Airflow interactions in the crosscuts at the intersec-
tions were bi-directional with airflow leaving the main
drift along the lower half and airflow entering the main
drift in the upper half of the crosscuts. This was due
to the difference in airflow velocities in the main drift
(Figs. 3–4 ribs).
2.2 Propeller fan
Figure 5 shows the 2.44 m (8 ft) diameter propeller fan,
powered by a 22.3 kW (30 hp) motor, used in the study.
Both the vane-axial and propeller fans were operating
at close to 17.2 kW (23 hp) during the field inves-
tigations. The propeller fan moves a larger quantity
[58.5 m3 /s (124,000 cfm)] of slower moving air that
interacts differently with the surrounding air than the
airflow from the vane-axial fan. The air leaving the
propeller fan outlet is moving at 13 m/s (2500 ft/min),
and expands rapidly to cover the entire cross-section
of the drift. Figure 6 shows the airflow patterns gen-
erated by the propeller fan. Note that, compared to
the vane-axial fan, a higher airflow quantity, 253 m3 /s
(536,000 cfm) was achieved initially with the propeller
fan. The airflow quantity progressively diminishes
after each intersection. Also, the airflow does not leave
the propeller fan in a straight uniform direction, but
spreads out over the entire cross section of the drift.
Figure 7 shows the cross-sectional velocity profile
25 meters (82 ft) in front of the propeller fan’s out-
let. The airflow is quickly being dispersed across the
entire drift. The peak air velocity was 4.40 m/s (866 ft/
min), and the minimum air velocity was 1.12 m/s
(221 ft/min). All grid sections of the drift had positive
measurable airflow down the drift, with no airflow roll
back being observed.
Figure 8 shows the airflow 52 m (171 ft) in front
of the propeller fan. The air stream completely
enveloped the drift, with a peak air velocity of 2.84 m/s
(559 ft/min) and minimum air velocity of 1.26 m/s
(249 ft/min). The velocity profile is much closer to

537

Copyright © 2006 Taylor & Francis Group plc, London, UK

 uniform (i.e. being more evenly distributed across the
drift) than was observed with the vane-axial fan shown
in Figure 4. Recirculation is only observed in the sec-
ond intersection, and it is less than with the vane-axial
fan. It appears that some of the airflow is leaving the
main drift after the first intersection.
The airflow quantities moving down the main drift
for both fans is shown in Figure 9. The vane-axial
fan with its high exit velocity maintains a level of
entrainment as far as 79.5 m (261 ft) in front of the
fan, whereas the propeller fan only maintains entrain-
ment for about 52 m (170 ft). Both fans have similar
exponential (natural decay) reductions in airflow quan-
tities beyond the effects of entrainment. The propeller
fan acts like a large single source of airflow that
quickly settles into a long natural decay trend after the
first intersection, with an exponential curve of best
fit (R2 = 0.997). The vane-axial fan does not experi-
ence the natural decay trend until after 79.5 m (261 ft)
because of ongoing entrainment. The vane-axial fan
does not act as a single source of airflow, but covers a
larger area of entrainment. The propeller fan pulls air
from behind the fan and only up to the first crosscut.
The vane-axial fan pulls air from behind the fan and
the first three crosscuts.
Statistical data for the velocity profiles of the cross-
sections in front of the fans (Figs. 3–4 and 6–7) are
given in Table 1. The level of possible entrainment is a
function of the non-uniformity in the velocity profiles
across the drift. A uniform distribution of velocities
across the drift will not induce entrainment, but non-
uniform velocity distributions can. As can be seen in
Table 1, the vane-axial fan standard deviation is high,
when compared to the mean velocity for both cross-
sectional locations, indicating that entrainment will
occur further down the drift and work against the natu-
ral decay of airflow quantities. Data from the propeller
fan have a much lower standard deviation, indicating
that additional airflow cannot be entrained beyond the
first crosscut, as shown in Figure 9.
2.3 Use of large propellers fans for regional-area
ventilation
The use of propeller fans for ventilating a large regional
area was investigated. One mine operator has success-
fully used a 3.66 m (12 ft) free-standing propeller fan
for this purpose (Fig. 10). The induced airflow for the
3.66 m (12 ft) fan in this application had a similar pat-
tern to that of the previously discussed 2.44 m (8 ft)
propeller fan application, i.e. all of the air was pulled
from behind the fan or from the first intersection in
front of the fan. Flow rates measured in front of the fan
216 m (710 ft) down the main drift were over 250 m3 /s
(530,000 cfm) (Fig. 11). After the first intersection, the
airflow traveling down the main drift exhibits the same
natural decay as the 2.44 m (8 ft) propeller fan.

Distance From Fan (feet)

0 200 400 600 800
300
600,000
2.44 m Propeller
250
Air Quantity (cfm)

0.91 m Vane-Axial 500,000

Air Quantity (m3/s)

200 400,000

150 300,000

100 200,000

50 100,000

0 0
0 50 100 150 200 250 300
Distance From Fan (m)

Figure 9. Airflow characteristics for free-standing Figure 10. 3.65 m (12 ft) propeller fan used for regional
vane-axial and propeller fans. airflow studies.

Table 1. Mean velocity and standard deviation of cross-

sectional airflow in front of the fans.

Vane-Axial Propeller

Distance to fan (m) 28 55 25 52

Mean Velocity (m/s) 1.80 1.86 2.46 1.97
Standard Deviation (m/s) 1.60 0.94 0.81 0.40
Distance to fan (ft) 92 180 81 169
Mean Velocity (ft/min) 354 366 485 388
Standard Deviation (ft/min) 315 185 158 80 Figure 11. Regional airflow distribution for a 3.66 m (12 ft)
propeller fan.

538

Copyright © 2006 Taylor & Francis Group plc, London, UK

 2.4 Propeller fans used on single entry drifts
NIOSH tested a trailer mounted, diesel powered 1.37 m
(54 in) propeller fan to be used for regional and
face ventilation (Fig. 12). The diesel power propeller
fan performed well in inducing airflow at the last
open crosscut, but performed poorly when ventilat-
ing a single-entry heading. Figure 13 demonstrates
the effect that an incorrect placement of an inappro-
priate fan can have on ventilation efficiency. The low
exit velocity of 10.0 m/s (1970 ft/min) from the pro-
peller fan has a poor penetrating distance, which could
not sufficiently remove the airborne contaminants at
the mucking location located 55 m (180 ft) in front of
the fan. Very little airflow mixing was observed at the
intersection used for loading, and a build up of diesel
contaminants soon caused a shut down of production.
Based on previous USBM studies a better result would
Figure 12. 1.37 m (54 in) diesel powered propeller fan used
for face and regional mine ventilation.
50 m
100
Vane-Axial
Propeller Fan
Test Location 98 Regional Fan
Mucking Area Recommended
Location for
Propeller Fan
22 94
Loading Area
98 = Airflow m3/s
= Airflow Direction
= Fan
Figure 13. Test of a diesel powered propeller fan in a dead
end drift, airflows in m3 /s.
539
Copyright © 2006 Taylor & Francis Group plc, London, UK
have been achieved with a vane-axial fan equipped
with a reducer to create a high-speed jet airflow capa-
ble of penetrating completely into the single entry drift,
as shown in Figure 3. This would have resulted in more
air mixing and dilution at the loading site. Propeller
fans, because of their low speed and quickly expand-
ing airflow patterns, are poor choices for ventilating
single entry drifts.
A good application using a propeller fan to venti-
late an underground crusher site was demonstrated by
Chekan (2006). The low-speed, wide airflow pattern
generated by the propeller fan completely engulfed the
entry and pushed the dust towards the return airway
drift, without disturbing the dust lying on the ground.
2.5 Fan placement recommendations
Vane-axial fans should be positioned in such a way
that they can entrain a maximum amount of fresh air.
In some cases it is best to have the fan blowing across
a fresh air stream because these fans entrain little air-
flow from behind the fan yet have entrainment for 90 m
(300 ft) after the fan before airflow is lost from the
main drift (Fig. 4). Propeller fans should be located
with in the fresh air stream because propeller fans
entrain airflow from behind the fan and from the first
crosscut after which airflow is lost from the main
drift (Fig. 6). The following recommendations are also
important:
– Propeller fans should be situated in the fresh air
stream.
– Vane-axial fans should be placed one entry behind
the fresh air stream.
– Propeller fans work best in regional ventilation
applications.
– Vane-axial fans work best (better penetration,
greater mobility) in face and dead end ventilation
applications while propeller fans are not well suited
for these applications.
3 REPLACEMENT OF VANE-AXIAL MAIN
MINE FANS WITH PROPELLER FANS
As main mine fans for large-opening mines (fans
mounted in a bulkhead), propeller fans are the pre-
ferred choice (based on lower noise levels, capital and
operating costs), so long as the pressure requirements
are low [less than 185 Pa (0.75 in w.g.)]. Generally,
the pressure requirements are low for large-opening
drift mines with several portals. Figure 14 shows typ-
ical pressure versus flow curves for a high-resistance
coal mine and a low-resistance, large-opening mine.
Fan curves are shown for a high-pressure vane-axial
fan and a low-pressure propeller fan. Points A and B
represent the operating points for the high-pressure
vane-axial fan with a high-resistance coal mine and
 Fan Flowrate (x1000 cfm)
0 50 100 150 200 250 300 350 400
3.0 12
Fan Static Pressure (kPa)

3.66 m Propeller

Fan Static Pressure (in. w.g.)

2.5 Large Mine Resistance 10
2.13 m Vane-Axial
2.0 8
Coal Mine Resistance
A
1.5 6

1.0 4

0.5 2
B C
0.0 0
0 20 40 60 80 100 120 140 160 180 200
Fan Flowrate (m3/s)

Figure 14. Pressure and quantity curves for vane-axial and

propeller fans in a large opening mine. Figure 16. 3.66 m (12 ft) propeller fan mounted in a steel
bulkhead.

Table 2. Engineering comparison and cost analysis of

replacing a vane-axial main mine fan with a propeller fan.

Vane-Axial Propeller 2 Propeller

Fan Fan Fans

Diameter (m) 2.44 3.66 3.66

Diameter (ft) 8 12 12
Capital cost $14,000 $28,000
Airflow (m3 /s) 127 132 368
Airflow (cfm) 270,000 280,000 780,000
Brake kW 119 25 119
Brake horsepower 160 34 160
Cost per kWh $0.07 $0.07 $0.07
Cost per day $200 $43 $200
Figure 15. 2.44 m (8 ft) vane-axial main mine fan mounted Cost savings per day $157 $–
in a sheet metal bulkhead. Airflow increase (%) 4% 189%
Payback (days) 89
the low resistance of a large-opening mine, respec- Annual cost savings∗ $47,222
tively. Point C is the operating point for the propeller
∗
fan and the large-opening mine. Assumed 300 day of operation per year

annual cost saving is over $47,000, an obvious eco-

3.1 Case study nomic advantage for the use of propeller fans as main
A mature large-opening mine made a direct replace- mine fans in this case. The total ventilation airflow
ment of a 2.44 m (8 ft) vane-axial fan with a 3.66 m of the mine could be increased by the installation of
(12 ft) propeller fan for main mine ventilation. The two 3.66 m (12 ft) propeller fans operating together to
original bulkhead mounted vane-axial fan shown in exhaust 368 m3 /s (780,000 cfm) and operating at the
Figure 15 was operating at 119 kW (160 hp) and same power cost as the original vane-axial fan, 119 kW
exhausting 127 m3 /s (270,000 cfm) of ventilation air- (160 hp). This would represent an increase of approx-
flow. The replacement 3.66 m (12 ft) propeller fan is imately 190% more airflow in the mine for the same
operating at 25 kW (34 hp) and exhausting 132 m3 /s operating costs as the original vane-axial fan (Table 2).
(280,000 cfm) of ventilation airflow. Figure 16 shows
3.2 Parallel fans
a 3.66 m (12 ft) propeller fan installation. Table 2 com-
pares the operating conditions for the two fans along The addition of a second parallel fan to a bulkhead of a
with their operating costs. The replacement of the large-opening mine will create a significant improve-
vane-axial fan with the propeller fan saves the mine ment over the original fan due to the little interference
about $155 per day in operating costs at $0.07/kWh, between the parallel fans. Placing two high-pressure
which results in a payback period of less than three fans in parallel normally causes the two fans to work
months for the propeller fan purchase. The mine oper- against each other, reducing the individual fan’s air-
ates approximately 300 days a year, so the expected flow quantity. However, with the low resistance of

540

Copyright © 2006 Taylor & Francis Group plc, London, UK

 the large-opening mines, the parallel fans only have
to overcome a small static pressure. The result is an
almost free-flow discharge by fans operating in paral-
lel, and the actual minor reduction in airflow quantity
is within measurement errors. Further testing will be
conducted to investigate the possibility of operating
three or four fans in parallel. Figure 16 shows a pro-
peller fan mounted in a steel bulkhead with room for
a second fan to be installed in the future.

4 VENTILATING A NEW MINE

NIOSH collaborated with the operators of a new, dual-

portal large-opening mine to determine suitable venti-
lation systems during the first few years of operation. Figure 17. Site one, 2.44 m (8 ft) propeller fan blowing into
Free-standing fans were recommended to ventilate the a dual-entry large-opening mine.
mine initially because at the early stage of mine devel-
opment, bulkheads would not be able to stand up to Table 3. Fans blowing into mine.
nearby blast pressures. The same vane-axial and pro-
peller fans (Figs. 1, 5) used previously to determine Distance Propeller Vane-Axial
underground airflow patterns were used for this study. m Location m3 /s m3 /s
The research approach consisted of varying the loca-
tion and directions of the fans, and then evaluating Site 1 19.2 Outside 153 80
the resultant airflows into and out of the dual portals Site 2 31.7 Outside 147 98
(Fig. 17). Site 3 25.9 Inside 134 93

ft cfm cfm
4.1 Fans blowing into the mine
Site 1 63 Outside 325,000 169,000
The 2.44 m (8 ft) propeller fan (Fig. 17) was tested at Site 2 104 Outside 311,000 207,000
three locations (Table 3) while blowing into the mine Site 3 85 Inside 283,00 198,000
as shown in Figure 18. The experiment was repeated
again using the vane-axial fan (Fig. 19) at the same
three locations. Table 3 summarizes the site locations
relative to the portal and the recorded airflow.

4.2 Fans blowing out of the mine

Tests were also performed with the fans exhausting
from the mine, but resulted in less airflow the closer
the fans were positioned to the portal entrance because
of mixing and entrainment of the air. Site 3 was used
again, as well as an additional location (Site 4), located
14.0 m (46 ft) inside the portal. The results for the pro-
peller fan are shown in Figure 20 and Table 4. With the
propeller fan blowing out of the mine, a maximum flow
rate of 83 m3 /s (176,000 cfm) was achieved, which
Figure 18. 2.44 m (8 ft) propeller fan blowing into the mine.
is much less than that measured with the same fan
blowing into the mine at any of the test sites (Fig. 18).
The vane-axial fan was also tested at Sites 3 and 4 5 SUMMARY
inside the mine, but the airflows were below the vane
anemometer’s detection limits of 19 m3 /s (40,000 cfm) NIOSH investigated the use of propeller fans in
at both locations. The poor performance of the vane- underground large-opening mines as an alternative to
axial fan when used to exhaust from the mine was vane-axial fans. Free-standing propeller fans only pull
expected because the confined high-speed jet air airflow from behind the fan and from a distance of
stream would be outside the mine before it could 30 m (100 ft) in front of the fan (first intersection).
expand and entrain any appreciable mine airflow. The propeller fan acts as a large single source to push

541

Copyright © 2006 Taylor & Francis Group plc, London, UK

 should be located so that they blow through the fresh
air stream.
Propeller fans clearly have a large advantage when
mounted in a bulkhead for applications in the low mine
resistances common to large-opening multiple-portal
drift mines. Improvements in airflow quantities, oper-
ating costs, and ventilation efficiency can be achieved
with the replacement of high-pressure, vane-axial fans
with low-pressure propeller fans, as shown in Table 2.
For a new large-opening mine, the use of a propeller
fan located just outby a portal and blowing into the
mine generates a large and effective airflow for the
mine. A vane-axial fan mounted with a reducer is far
Figure 19. 0.91 m (3 ft) vane-axial fan blowing into the superior to a propeller fan for ventilation of a single
mine. development heading. For single-entry ventilation pre-
vious USBM work by Brechtel et al. (1985) and Dunn
et al. (1983) on free standing vane-axial fans is still
valid, and is the preferred method.

6 DISCLAIMER

The findings and conclusions in this report are those of

the authors and do not necessarily represent the views
of the National Institute for Occupational Safety and
Health.

REFERENCES
Figure 20. 2.44 m (8 ft) propeller fan blowing out of the
Brechtel, C.E., Adam, M.E. & Agapito, J.F. 1985. Develop-
mine.
ment of effective face ventilation systems for oil shale
mining. Pittsburgh, PA. U.S. Department of the Interior,
Table 4. Fans exhausting out of mine.
Bureau of Mines, (BOM), OFR 14–1.
CFR. U.S. Code of Federal Regulations Part 57. Washing-
Distance Propeller Vane-Axial
ton, DC: U.S. Government Printing Office, Office of the
m Location m3 /s m3 /s Federal Register.
Chekan, G.J., Colinet, J.F. & Grau, R.H. 2006. Impact
Site 3 25.9 Inside 83 <19 of Fan Type for Reducing Repairable Dust an Under-
Site 4 14.0 Inside 52 <19 ground Limestone Crushing Facility. Proceedings of the
11th North American/Ninth US Ventilation Symposium,
ft cfm cfm University Park, Pa., June 5–7, 2006.
Dunn, M.F., Kendorski, F.S., Rahim, M.O. & Mukherjee, A.
Site 3 85 Inside 176,000 <40,000 1983. Using auxiliary “Jet Fans” in large underground
Site 4 46 Inside 110,000 <40,000 metal/non-metal mines. A handbook. Pittsburgh, PA: U.S.
Department of the Interior, Bureau of Mines (BOM), OFR
107-84.
Grau III, R.H. et al. 2002. NIOSH ventilation research
the air in one direction, followed quickly by the natural addressing diesel emissions and other air quality issues
decay of the airflow. in nonmetal mines. 2002 SME Annual Meeting, Preprint
Compared to a propeller fan, a free-standing, vane- 02-187. Phoenix, Arizona, February 25–27.
axial fan mounted with a reducer pulls less airflow Grau III, R.H., Robertson, S.B., Krog, R.B., Chekan, G.J. &
from behind the fan and from the first 30 m (100 ft), but Mucho, T.P. 2004. Raising the bar of ventilation for large-
it has a large area of entrainment and mixing for 90 m opening stone mines. In R. Ganguli, S. Bandopadhyay
(300 ft) in front of the fan before the natural decay of (eds), Mine ventilation: Proceedings of the 10th U.S./
North American Mine Ventilation Symposium, Anchorage,
the airflow occurs. Both free-standing vane-axial and
AK, May 16–19, 2004: 349–355. Leiden, Netherlands:
propeller fans have the ability to effectively ventilate Balkema.
a region or to flush the working areas of a mine. How- Krog, R.B., Grau III, R.H., Mucho, T.P. & Robertson,
ever, propeller fans will have lower operating costs S.B. 2004. Ventilation planning layouts for large-opening
for the same total airflow. Propeller fans should be mines. SME preprint 04-187: 1–9. Littleton, CO: Society
located in the fresh air stream, whereas vane-axial fans for Mining, Metallurgy, and Exploration, Inc.

542

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Design and operational aspects in the use of booster, circuit and

auxiliary fan systems

D.J. Brake & C.A. Nixon

Mine Ventilation Australia, Brisbane, Australia

ABSTRACT: Whilst there have been relatively few innovations in the nature or application of main surface
fans in the past 20 years in Australia, there have been a significant number of notable changes in the size, type
and application of booster and auxiliary fans. In metalliferous mines, these changes relate to factors such as:
increased development size, changes in mining methods brought about by the use of remote and teleremote
equipment, the use of air-conditioned cabins, new developments in refuge chambers, significant increases in the
use of mine refrigeration and the extraction of both “thinner” orebodies (lower tonnes per vertical meter) and
massive low-grade orebodies (caving methods) and overall greater depths. In coal mines, these relate to increased
length and width of longwall panels, much higher extraction rates, more effective gas pre- and post-drainage,
longer main intakes and returns and changes in legislation. In both coal and hardrock, the use of more and/or
much larger diesel engines and changes in the philosophy of egress are also impacting on the use of circuit and
booster fans. This paper describes the factors that are influencing new applications for auxiliary and booster
fans, the ways in which these fans are being used and some of the problems that have been experienced in the
use of booster, circuit and auxiliary fans.

1 INTRODUCTION AND DEFINITIONS
It is interesting to note that there have been serious
and expensive contractual disputes recently in Aus-
tralia regarding the terminology of “primary” and
“secondary” ventilation. In one case, the contractor
ultimately went into liquidation over this matter. A
typical dispute arises where the contract states that
the mine owner is responsible for “primary” venti-
lation and the mining contractor is responsible for
“secondary” ventilation.
In this paper, we propose the following definitions:
– A primary fan is one that impacts significantly on
the total mine airflow. In many mines, there is only
one primary fan. However, in other mines, there
may be multiple exhaust or intake shafts with multi-
ple fans on these shafts. These would all be primary
fans. Primary fans are usually located on surface
but this is not always the case.
– A booster fan is a primary fan that is in series with
another primary fan. Without the booster fan oper-
ating, the total mine airflow will fall significantly.
Booster fans are typically installed as the mine
develops and the original primary fan no longer
has the pressure or flow capability to service the
increased resistance of the operation. In this sense,
the booster fan operates as an integral and “symbi-
otic” unit with the original primary fan to provide
sufficient total airflow for the mine.
– A circuit or district fan is a fan that does not signif-
icantly impact on the overall total mine airflow but
does have a significant effect on a particular district
or circuit within the mine.
– A development or auxiliary fan is a fan that venti-
lates a “blind” (dead end) heading or workplace. In
most cases, the development or auxiliary fan has no
impact on district airflows, although some impor-
tant exceptions are discussed later in this paper.
Note that we have not used and consciously avoid
the term “secondary” ventilation. Whilst it is a term
in common use, there is little agreement about exactly
what is meant by it. Some would perhaps argue that a
“secondary” fan is any fan that is not a surface primary
fan; however, it is difficult to put a rational, consistent
and unambiguous case for this definition and we do
not endorse it.
Under these definitions, it is the context that a fan is
used in that determines its role, rather than the actual
fan design itself. For example, a development fan (nor-
mally used with duct) can be installed without duct
in a wall and effectively becomes a circuit fan. Or
four development or “circuit” fans can be installed

543

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 1. Typical coal mine auxiliary fan.
in parallel on top of an exhaust shaft and effectively
become a primary fan installation.
Circuit, booster and auxiliary fans are a very impor-
tant part of the ventilation engineer’s toolbox. Not only
are they essential to ventilate most dead-end headings
of any length or to boost the primary fan(s), but they
can also control the direction and magnitude of air-
flows in the mine (acting as negative regulators). In
this role, they have two distinct advantages over reg-
ulators. Firstly they do not destroy pressure, so that
they are more energy efficient. Second, they are rela-
tively less affected by changes in ventilation controls
at other locations, so that the district flow is usually
more consistent than with a regulator. In addition,
these fans can be used very deliberately to reduce the
magnitude of or control the direction of leakage in var-
ious parts of the mine. In some larger hardrock mines,
dozens or hundreds of circuit, booster and auxiliary
fans will be in use absorbing many megawatts of elec-
trical power, often in excess of the surface fans. Careful
management of these fans is therefore not only impor-
tant from a safety and operational point of view, but
also in terms of managing mine capital and operating
costs.
This paper concentrates on many of the practical
design and operational issues that cause problems with
booster, circuit and development fans.
544
Copyright © 2006 Taylor & Francis Group plc, London, UK
2 DEVELOPMENT OR AUXILIARY FANS
Development or auxiliary fans in this context are used
to ventilate blind headings using duct. Certainly there
are applications in which fans are used to “throw” air
into a heading without using duct (e.g. a Venturi fan)
or used as an air brake, air curtain or air mover without
duct to reduce or increase airflows through a particular
split of air or to increase the wind speed or “scour” a
particular region of the working place; however, these
“ductless” applications are not discussed in this paper.
An auxiliary fan in this paper has duct on one or both
ends and may have a bellmouth on the intake or an
evasé on the outlet, but not both a bellmouth and an
evasé. By contrast, a circuit fan will normally have
both a bellmouth and an evasé (Fig. 8).
There are substantial differences between auxiliary
fans used in coal mines and those used in metalliferous
mines. Coal mine auxiliary fans (Fig. 1) are normally
centrifugal fans employed in exhaust configuration
and need to have flameproof motors, usually have the
motor out of the air stream, be constructed of special
materials, have provision for de-gassing the heading,
and often have integral trickle dusters at the outlet end
of the fan. They are large, heavy and expensive.
By contrast, hardrock auxiliary fans (Fig. 2) are usu-
ally axial fans and are much simpler, lighter, cheaper
 Table 2. Advantages and disadvantages of forcing system
of auxiliary ventilation.

Advantages Disadvantages

Positive pressure at face If used with

Duct (flexible) is cheap continuous mining,
Duct is easy to handle, transport, may force dust on
store operator which is
Fans are much cheaper, simpler, then not dispersed
Figure 2. Typical metalliferous (hardrock) mine auxil- lighter or extracted
iary fan. Fan is in the intake May require
Provides fresh air to face dust filters
Table 1. Advantages and disadvantages of exhaust system Better for outbursts Duct is very
of auxiliary ventilation. No de-gassing boxes etc. required easily damaged
Can extend longer distances Noise from fan in
Advantages Disadvantages Doesn’t need to be kept as close to intake/travel road
the face Tramming is in
Immediately Tube diameter can constrict Easier to detect leaks (under return air from
removes dust activities in low seam heights pressure) face
and gas produced Due to its weight and rigidity, it is No gas passing through fan
at face more difficult to handle, transport Quieter as longer ducts possible
Gets constant and store (so fan is further from face)
extraction quantity Cost of tubes
(for short ducts) Damage to tubes by passing vehicles
Removes heat and Potential for gas through fan
– Compared to the relatively “set and forget” nature
humidity produced Cost and weight of fan (especially
at face if bifurcated type) of circuit or primary fans, auxiliary systems are
Option for trickle Low face velocity (low scouring very dependent on the operators to maintain stan-
duster at discharge of face) dard and therefore effectiveness
end of fan Difficult to identify leaks as under – In most cases, the systems only ventilate one side
Easily extended suction of the face. It is common for a Venturi fan to be
Fan out of way in Needs electrical interlocks used to improve air distribution in the face region.
return Safety issue of personnel being
sucked up the tube
4 OPEN MOUNTED FANS

and smaller, even though they are frequently much
higher powered due to the high flow requirements for
the larger diesel engines in hardrock mines.
3 FORCING (BLOWING) AND EXHAUSTING
AUXILIARY FAN CONFIGURATIONS
Ducted fans can be used in forcing (blowing) or
exhausting configurations, or various combinations of
these. Both options are in widespread use, although the
general rule is that “continuous” mining methods (such
as roadheaders or continuous miners) use exhausting
ventilation because the contaminants of gas and dust
are being continuously produced, whereas “drill and
blast” mining methods use forcing ventilation because
the gas and dust is largely produced during the blasting
process.
The advantages and disadvantages of the two meth-
ods are shown in Tables 1 and 2.
Common problems with all auxiliary duct systems
(blowing and exhausting) are:
– Leakage reduces face flows below fan flows
Ducted auxiliary fans can therefore be used in two
ways:
– Hung from the back/roof or placed on the ground
(open mounted fans), or
– Installed in or through a wall (wall mounted fans)
A basic design principal for any fan is to avoid recir-
culation of air from the fan outlet back into the fan
inlet. This is especially true of ducted fans where the
nature of a blind heading is that the air leaving the head-
ing is often passing close by to the fan inlet. Several
“rules of thumb” are used to ensure there is sufficient
“bypass” air so as to avoid recirculation:
– The airflow through the district must be equal to
the sum of the “open flow” (i.e. without resistance)
volumes of all fans in the district plus an additional
30% of the “open flow” volume of the largest fan
in the district.
– The airflow through the district must be at least 30%
greater than the sum of the “open flow” volumes of
all fans in the district.
– The airflow should be such that at least 0.5 m/s of
air bypasses each fan.

545

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Each of these approaches has weaknesses. For
example, if an auxiliary fan has an open circuit flow
of 10 m3 /s, then under the first rule, the bypass airflow
would need to be 3 m3 /s. If the heading in which the
fan is located is 3 m × 3 m, then the bypass airflow has
a wind speed of 0.33 m/s – low but perhaps sufficient.
However, if the heading is 6 m × 6 m, then the bypass
wind speed is less than 0.1 m/s, i.e. imperceptible and
unlikely to avoid recirculation.
Note that auxiliary fans and ducts are usually
selected on the basis of meeting the required face
airflow at the “longest” duct run that is required. How-
ever, in terms of preventing recirculation at the fan, the
system should be carefully designed so that the fan will
not short-circuit even when the fan is at its highest flow
in the system (usually when the duct is at its shortest).
Our recommendation (apart from meeting statutory
requirements) is for the bypass flow to be at least
0.5 m/s when the fan is at its highest flow in that
fan/duct application and for the fan to be located at
least 10 m from the return airflow.
Note that technically, the auxiliary fan must be
selected so as to not only overcome the duct friction,
but also the frictional pressure loss of the “return” air
in the dead-end heading outside the duct. However,
in virtually all practical cases, this frictional pressure
loss is so small compared to the duct friction loss that
it can safely be ignored.
5 WALL MOUNTED FANS
A wall-mounted fan is not only a fan that is mounted
in a sealed wall, but also a fan that has duct pass-
ing through a sealed wall (Fig. 3). In both cases, the
fan not only sees the resistance of the duct, but also
some of the system resistance before the wall. This
type of installation creates a “hybrid” auxiliary fan –
Figure 3. Wall mounted fans.
546
Copyright © 2006 Taylor & Francis Group plc, London, UK
one that overcomes the duct resistance and also part
(or sometimes all) of the district or mine resistance.
Wall mounted fans have advantages and disadvan-
tages compared to open mounted fans.
These include:
– Providing they are sealed into the wall, wall
mounted fans cannot recirculate at the wall, elim-
inating a key safety concern with open mounted
fans. Note however, that there is still the potential
for recirculation to occur via parallel connections
around the wall, although the potential for this is
usually low in practice.
– Because they cannot recirculate at the wall, wall
mounted fans do not need a “bypass” airflow
allowance, which often reduces the total mine
airflow requirement, saving capital and operating
costs. This is a particular advantage if the air must
be chilled as chilled air is expensive.
– Because air travels at typically about 20 m/s in a
duct, but at perhaps 1 to 2 m/s in normal develop-
ment, then if a wall mounted fan is a forcing fan,
the air reaches the face must faster than if the fan
is open mounted. The air reaching the face can-
not be contaminated by diesel or other fumes or
gases or moisture in this region and due to the fast
transit time, cannot pick up as much sensible heat.
For this region, wall-mounted installations are usu-
ally much preferable where the mine operates with
refrigerated air as the temperature increase from the
fresh air raise to the face is usually much lower than
if the air is allowed to leave the fresh air raise, drift
along the open development to an open-mounted
fan, and then be ducted to the face.
– Wall mounted fans can be used on a sealed fresh
airway (if blowing fans) or a sealed return airway
(if exhausting fans) to overcome the duct resistance
and the resistance of the airway. An effective self-
closing damper must be installed in these cases to
ensure the airway itself does not recirculate. This
feature of wall-mounted fans is especially impor-
tant if a major airway has many such fans mounted
on it. If the airway is not sealed, then a circuit or
other fan is required to ensure that the airway does
not recirculate. However, sizing this circuit fan can
be problematic as it must be sized for the maxi-
mum number of wall mounted auxiliary fans that
might be on at any time, which means that air will
be wasted when fewer fans are in use. However, if
the raise is sealed and the auxiliary fans properly
selected, then the system is self-regulating without
a circuit fan.
6 NUMBERS OF DUCTS AND FANS
There is now a large variety in diameters, unit lengths,
materials and construction methods for both rigid and
 flexible duct. In general, the cheaper and poorer quality
duct is employed for “one off” situations where the
duct is basically sacrificial. In situations where the duct
will be recovered and reused, or where the installation
is critical (e.g. very long duct runs, then higher quality
duct should be employed.
Exhaust duct must always be rigid and is almost
always recovered. Due to manual handling issues, it is
usually short lengths (typically 3 m) and this also fits
in with the requirement for exhausting duct to be close
to the face at all times.
Forcing (blowing) ducts are usually flexible and
are the preferred option for hardrock mines due to
the lower cost and ease of storage and installation.
They can also be further away from the face due to
the “nozzle” effect of air scouring the face even when
the duct is 15 to 20 m from the face. This is an advan-
tage in hardrock mines as they employ drill and blast
techniques which obviously damage any nearby duct.
Ventilation should always be turned off before blasting
and the duct should either be removed to 50 m from
the face, or a “sacrificial” length kept close to the face,
so that duct extensions are added one duct length back
from the last duct length.
Where the duct runs are short, then it is usually
more convenient to use short lengths of duct (20 m)
but when the duct runs are long (300 m or more) then
it is highly beneficial to use 100 m lengths of duct. This
significantly reduces leakage through the duct as, for
well-maintained duct, most of the leakage occurs at
the joins so that the fewer joins, the lower the leakage.
With long duct runs, it is very desirable to have at
least two parallel ducts. This provides for maintenance
on the ducts and also for a minimum airflow to clear
gases or heat even if one duct (or one fan) is off. For
example, if one duct is damaged, then it must be turned
off for repairs to be made safely and effectively. If there
is no parallel duct, then either a special duct must be
Figure 4. Two stage, two motor, contra-rotating axial devel-
opment fan (courtesy SDS Ausminco).
547
Copyright © 2006 Taylor & Francis Group plc, London, UK
run for the repair job, or the work must be undertaken
without ventilation.
The fact that auxiliary fans of the axial type can
be used so easily in series, or in parallel, or in
series-parallel provides great flexibility in selecting
the correct combinations of fans and ducts for virtually
any hardrock development application.
7 INTEGRATED SERIES FANS
Auxiliary fans purchased as a paired set of two
impellers in series, each with its own motor, are now
common in hardrock applications. An example is the
SDS GAL 14 (twin 90 kW) fan shown in Fig. 4 with
fan curve shown in Fig. 5. In this fan, the two stages
(impellers) are contra-rotating (i.e. spin in opposite
directions). This effectively removes the need for inter-
stage guide vanes, reducing the cost, weight and length
of the fan. A common misconception by operators is
that these fans can be “split” into their two compo-
nents and operated as two identical single-stage fans.
However, in the case of this fan, the impellers have dif-
ferent numbers of blades and the performance of the
fan as two separate stages is very different (and much
poorer) than that of the two stages operating as a single
integral unit. For example and with the fan split, at a
Figure 5. Fan curve for two stage, two motor, contra-rotating
axial development fan (courtesy SDS Ausminco).
 flow of 25 m3 /s, stage 1 will provide 450 Pa and stage 2
(by itself) would provide 1400 Pa but joined together,
the unit would provide a FSP of almost 3800 Pa.
8 GENERAL ISSUES
An important point is to take into account leakage
in the duct, and also to take into account any addi-
tional resistances on the system inlet or outlet that are
not already included in the fan curve. For example, if
an exhausting auxiliary fan is supplied by the man-
ufacturer with an evasé, but this is removed by the
operators, then the fan static pressure curve will be
affected (adversely) and must be adjusted for the addi-
tional losses. A similar situation arises if a bellmouth
is removed, although in this case, both the fan static
pressure and fan total pressure curves are affected.
It is very important that open mounted fans are
installed where they will not be damaged by equip-
ment, especially if the duct is to have a long life and/or
is very long. If the fan is installed as an exhausting
fan, then it is located in the return airway that usually
has lesser traffic. However, if the fan is a blowing fan,
then it will be installed in the intake air way, which
is usually the principle route for people and equip-
ment. The development profile, mobile equipment and
fan/duct combination need to be carefully selected so
that they are compatible. In addition, if the duct is flex-
ible, then it will hang down lower when uninflated and
this also needs to be taken into account. Recommended
clearances are shown in Table 3. These are based on
ventilation duct in the middle of the back/roof. Often,
the duct is placed in one corner of the roof, in which
case, the table can be adjusted.
Hence in a hardrock mine, if the heaped truck
height is 3.5 m and the duct diameter is 1.4 m, then
the design development height should be 3.5 + (4 ×
0.3) + 1.4 = 6.1 m. The finished height (allowing for
road base) would be 5.8 m. It is also very impor-
tant that if the duct must pass across an intersection,
that the heights of the development do not force a
poor installation where the duct is constantly being
rubbed or struck by passing vehicles. An example for
a hardrock application is given in Fig. 6. This clearly
shows the importance of stripping the back/roof where
development of dissimilar heights meets and auxiliary
ventilation is required to pass from one to the other.
A further issue with ducting is that duct should
never be run in close proximity to firing lines (blasting
cables). Incidents have occurred where flexible duct
has rubbed through the protective sheathing on fir-
ing lines and accidental ignitions have subsequently
resulted due to a static electrical discharge from the
duct to the firing line. Coal mines are required by law
to use FRAS ducting (fire resistant, anti-static).
548
Copyright © 2006 Taylor & Francis Group plc, London, UK
Table 3. Recommended duct clearances for critical auxil-
iary ventilation installations (duct in centre of roof/back).
Hardrock Coal
Clearance from design floor to true 0.3 m 0.1 m
floor (due to muck/dirt/coal left
on floor)
Clearance from top of full (heaped) 0.3 m 0.2 m
truck or shuttle car tray to duct
Clearance from top of duct to design 0.3 m 0.2 m
back/roof
Allowance for occasional minor 0.3 m 0.2 m
underbreak or irregularities on floor
or back, or for slightly over-filled
trucks or cars
9 OVERLAP SYSTEMS
As noted earlier, long blind development headings are
best serviced by twin ducts to provide some redun-
dancy and means of repairing each duct. If “leak free”
rigid duct is employed, then fans can be spaced inside
the duct at appropriate intervals to develop the nec-
essary system pressure to provide for virtually any
length of duct run. However, if flexible duct is to be
employed, then fans cannot be located along the duct
as this is generally not practical especially since the
face is constantly advancing. An overlap system can
be effective in these circumstances (Fig. 7, shown at
completion).
The typical operation of the overlap system is:
– A single blowing fan and duct is used to develop a
certain distance
– As the amount of equipment increases in the head-
ing due to its length, a second blowing fan and
duct in parallel with the first is used to develop to
the “overlap” point which is usually about the mid-
point of the expected completed heading length
– The fans are then advanced to the overlap point (the
ducts left in place) and new fans attached to the orig-
inal duct at the overlap point. A “Y” piece and short
length of larger single duct is often used at the over-
lap point to avoid the need for four ducts to be across
the back/roof of the heading. The original ducts are
then operated in “reverse”, i.e. still as blowing ducts
but from the overlap back to the start of the drive.
The face continues to be advanced by installing new
twin ducts from the overlap to the face.
Centre overlap systems such as this result in hotter
delivery temperatures (due to the air travelling more
slowly in the first half of the heading, and also due
to diesels operating in the intake air in this first half),
require more air (to ensure no recirculation at the over-
lap section), require more fans, the overlap section
must frequently be higher (roof stripped) and wider,
 Clearance from back of drives to
top of duct = 0.3 m in all cases
Flexible vent
in collapsed Stripping required here
state

Dimensions of Toro 0010, 2.57 m H x 11.12 m L, 261 kW

Perimeter drive

Drawpoint
Build up on floor = 0.3 m

1. Main drive is 5.7 H x 5.0 W. Assume 1.4 ␾ duct hung centrally in back of perimeter drive. Assume clearance from back to
duct is 0.3 m. Clearance under duct to floor (assuming 0.3 muck on floor) is 3.5 m when duct inflated. If flexible duct is not
inflated, it hangs down 2.2 m, not 1.4 m, reducing clearance from 3.5 to 2.7 m. Hence need for semi-rigid poly duct in this
region.
2. Note that intersection of perimeter drive with every drawpoint will need to have back stripped else “T-piece” will foul on
equipment in perimeter drive.
3. Drawpoint is 4.2 H x 5.0 W. Assume 1.1 ␾ duct. Note that clearance under inflated duct is 2.3 m (insufficient for LHD) and
even less if duct is not inflated. Hence drawpoint must be wide enough to fit vent duct in shoulder so that it is safe whether
duct is inflated or not.

Figure 6. The profiles and dimensions of development must be carefully considered for satisfactory auxiliary ventilation oper-
ation, taking into account intersecting development, duct clearances when inflated and uninflated, and equipment dimensions
loaded and unloaded.

to RAR Walls of drive

Flexible duct Rigid

to RAR Face

Flexible duct

Not to scale Face

Twin fans in series

Figure 7. Push-pull system with centre overlap (plan view).

electrical power must be run to the overlap section for

the fans, and the system will result in at least some
recirculation in the first half of the heading due to
leakage from the return duct directly into the intake.
However, overlap systems do ensure that the “first”
part of the development is still in “fresh” air, which
can be beneficial if diamond drillers or other activities
need to take place while development is continuing.
In addition, they can substantially reduce the re-entry
time after blasting, as only the “face compartment”
needs to be cleared of fumes as the return fan carries
any fumes away within the vent bag itself. The re-entry
time issue can become a key design requirement for
these long headings, especially if they are “high prior-
ity” jobs and set up to have independent ventilation to
the rest of the mine so that they can “fire anytime”.
To avoid recirculation, the length of the over-
lap region is very dependent on the velocity of the
“bypass” air. The higher the velocity of bypass air, the
shorter the overlap can be and vice-versa. A minimum
Figure 8. Circuit fan (note bellmouth and evasé on ends,
and mesh guards on both ends) (courtesy SDS Ausminco).
velocity would be 0.5 m/s and the corresponding over-
lap distance would be about 10 diameters (of the
development heading in which the overlap occurs).
10 BOOSTER OR CIRCUIT FANS
Booster and circuit fans do not have duct connected at
either end of the fan. They therefore normally include
both a bellmouth and an evasé, as shown in Fig. 8.
Note that whilst booster fans are in series with
a primary fan, they are physically separated from
the primary fan. It is very important to ensure that,
if the primary fan goes off, the booster fan does not go

549

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 9. Two methods for providing bypass to an
underground booster fan.
into stall, and vice-versa, i.e. if the booster fan goes off
that the primary fan does not go into stall. A similar
rationale applies for circuit fans.
It is also important to ensure the physical layouts
of the fans allow them to be safely serviced in-situ
and/or removed for repairs without compromising the
other fan. In a hardrock mine, this can be achieved
either by a door in the wall containing the fan, or by a
bypass drive (Fig. 9). In the case of a coal mine where
the exhaust usually consists of multiple roadways in
parallel, the bypass arrangements are more complex,
especially since the motors must be set up in intake air.
11 FAN PRESSURE
Technically, the pressure losses in any ventilation
system such as a mine are: the inlet losses from atmo-
sphere into the intakes, the frictional losses (including
550
Copyright © 2006 Taylor & Francis Group plc, London, UK
the special frictional loss call shock losses) within the
mine, and the unavoidable loss of (velocity) pressure at
the exhaust outlets from the mine back into the atmo-
sphere. In theory, the correct solution for fan pressure
problems is to sum these losses and then apply the fan
total pressure(s) (FTP) to overcome them. However,
this is a complex and usually unnecessarily compli-
cated approach. The more prudent approach is to use
only fan static pressures (FSP) to overcome the system
pressure losses. As FSP is always lower than FTP, this
usually builds a small measure of conservatism into
the fan selections.
It is important when selecting a fan to carefully con-
sider the air density going through the fan, correct size
the motor (and drives, if applicable) and to ensure the
fan curve is shown as FSP and is shown at the correct
impellor speed.
12 SERIES FAN INSTALLATIONS
In a series fan installation (the outlet of one fan is the
direct inlet to the next fan), there is only one loss of fan
velocity pressure (FVP) at any given flow, so that the
“normal” rule of summing the individual FSP values
of each fan is not correct. The true combined fan FSP
at any flow is the sum of the individual FSP values
plus the FVP of the first fan. A further exception to
the normal rule for adding fans in series is when the
fans are specifically designed to operate as a series
unit. An example is the two-stage contra-rotating fans
discussed earlier.
Note that in some circumstances, it is preferable
to purchase two identical axial fans (e.g. two 90 kW
fans) and bolt these in series, rather than purchasing a
two-stage contra-rotating axial fan (180 kW two-stage,
two-motor fan). Advantages of the separate fans is that
they can be “split” into two identical units, and, even
as a pair, they are often quieter than the equivalent
contra-rotating fan.
13 PARALLEL FAN INSTALLATIONS
There are a number of potential problems with fans
installed in parallel, such as in Fig. 10. It is impor-
tant that both the fan inlets and fan outlets are free of
obstructions and provide good flow paths into and out
of the fans. Some manufacturers recommend a baf-
fle be installed between parallel fans to avoid one fan
“starving” the other fan (Fig. 11).
A key issue for fans in parallel is to ensure they
do not operate in the stall region. The situation is
illustrated by the fan curve for four identical fans in
parallel shown in Fig. 12. In this case, all four fans
can be started (one after the other) because the final
system operating point (A) is below the lowest stall

Figure 10. Four 132 kW fans installed in parallel in under-
ground mine.
Figure 11. Baffle between parallel fans.
point of any of the individual fans. Note that if the
ventilation officer did not know the “stall region” for
each of the fans (as is often the case), then he/she
might conclude that this system could safely operate
at point B, but this is not the case. This fan system
cannot operate reliably at point B even though point
B is well below the stall point of any of the individual
fans. Note that this issue is particularly important at
higher fan pressures as the “deep stall” at point C for
any of these fans will become “deeper” (lower) as the
blade angle is made more aggressive and the fan motor
power increased. This means that a parallel installation
of such fans becomes increasingly low pressure as the
551
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 12. Single and combined fan curves for four identi-
cal fans operating in parallel.
blade angle is increased. It is therefore very impor-
tant for the fan manufacturer to be consulted before
proceeding further on a parallel fan system.
14 FAN DUTY
Some discussions have already been provided con-
cerning the important issue of nominating a fan duty.
It is very important that all details be provided to the
fan manufacturer before a fan is selected.
Of particular importance is for ventilation models
to have correct airway dimensions, lengths, friction
(“k”) factors and make adequate provision for shock
losses and leakage. Fan curves should be based on FSP
values.
Where a fan is a booster or circuit fans, it will be part
of a “system” that includes other fans. It is very impor-
tant to note in these cases that the system resistance
curve as seen by this fan will not pass through the “ori-
gin” (0 m3 /s and 0 Pa). It is very important, therefore,
for the ventilation officer to give not only the required
fan duty to the fan suppliers in this case, but also the
expected system operating curve. Consider the situa-
tion shown in Fig. 13 in which a circuit fan is required
to meet a duty of 100 m3 /s at 2500 Pa FSP. Assuming
a normal system resistance curve (curve A), the single
fan appears to be well suited to the duty as the depth of
the stall point is not relevant for a single fan starting up
under normal conditions (along curve A). However, in
this case the fan operates in parallel with others cre-
ating a system resistance for this fan of curve B. The
depth of the stall point (often not known to the ventila-
tion engineer) combined with the fact that the system
resistance seen by this fan does not pass through the
“origin”, will create serious problems as the fan will
have to pass through the stall point as it spins up and
may in fact become locked in stall. If only the fan
duty is provided to the fan supplier, and the fan is only
 4500
A fan and the inner fairing on the evasé was removed
4000
3500
3000
FSP, Pa
2500
2000
1500
1000
500
0
0 20 40 60 80 100 120 140
Flow, m3/s
Figure 13. System resistance curve for underground cir-
cuit fan.
evaluated on the basis of this duty, then once the fan is
installed, it simply may not operate at the duty at all.
15 FAN PIECES
It is important that underground fans be provided with
sufficient spare parts, including spare fan pieces. The
performance of the installation will fall, sometimes
very significantly, if the fan is meant to include an
evasé but it has been removed, or is meant to have a
bellmouth but it has been removed.
From a safety point of view, fan inlets and outlets
(except where duct is attached) should have a safety
screen.
If there is any potential for air to want to flow “back-
wards” through a fan when it is off, then it must be
fitted with a self-closing damper (SCD). These come
in various designs. The cheapest is the simple “but-
terfly” type installed between the fan outlet and the
evasé. This provides for the physically smallest SCD
and therefore lowest cost. However, as the velocity
pressure is high in this region, the SCD is relatively
inefficient. A more efficient design is for the SCD to
be installed at the outlet of the evasé, where the veloc-
ity pressure is already low, so that any shock losses
associated with the SCD (when open) are also low.
When fans are horizontal, then SCD butterfly leaves
mounted vertically work satisfactorily. However, it is
important that they are mounted correctly. If they are
not mounted correctly, then they may not close when
required, or may not open when required, both of which
will result in problems. It is also important that if
the SCD is designed to be in a particular location
on the fan then the manufacturer should be consulted
before mounting it somewhere else on the fan. We are
familiar with a situation in which a circuit fan had a
self-closing damper on the inlet and a evasé with an
inner fairing on the outlet. The operators decided to
move the fan to a different location. At the new loca-
tion, the SCD was changed to the outlet side of the
Copyright © 2006 Taylor & Francis Group plc, London, UK
B to allow the SCD to continue to open and close. The
fan went into stall and the subsequent investigation
revealed that its peak pressure performance had fallen
from 2400 Pa to 1600 Pa!
When fans are mounted vertically, then the SCD
usually requires a counterweight so that it will
work effectively without destroying excessive pressure
when open.
160 180
16 FAN INSTALLATIONS AND MOUNTING
Where fans are mounted into a sealed raise that
requires pedestrian access (e.g. to a ladderway, or to
take refuge from a fire), it is very important that air-
lock pedestrian doors are fitted if the pressures are
excessive, which they often are. In a recent case, a mine
surveyor lost part of his finger when a pedestrian door
with excessive pressure closed on it.
17 FAN IDENTIFICATION
In mines with many underground fans, it is important
that an identification system be adopted. Key elements
include:
– Casings colour coded by fan type
– Manufacturer’s name
– Motor kW
– Fan diameter
– Arrow for airflow direction
– Arrow for impeller rotation direction
– Steel boxing for forklift tynes transport and also for
mounting either from back/roof or on floor
– Note that airflow and impeller directions should be
welded onto casing as well as painted.
The fan diameter and motor kW should be painted
at four locations at equal angles around the external
circumference of the fan casing, so that these can be
easily read from all directions.
18 ON/OFF INDICATION AND RE-STARTING
FANS
In large hardrock mines with dozens and sometimes
hundreds of underground fans, it is common to find
one or more fans off when they should be on. Important
fans should have on/off and amps indication back at a
remote monitoring point, or at the very least, should
have double indicator lights at an easily visible loca-
tion. For example, critical fans in a return airway that
is not easily accessible should have twin lights (per
motor) at an external location that is easily travelled.
552
 In many cases, fans can now be automatically turned
on and off using ControlPED™or other technologies.
Re-entry times after blasting are very dependent
on when auxiliary ventilation can be re-established,
bearing in mind that the re-entry time is actually the
time after ventilation is re-established and not after the
blast occurs as such. If fans need to be started manu-
ally, then this can delay re-entry. In addition, manual
re-start means that fan starters must always be in fresh
air to allow safe access for re-staring, which can be dif-
ficult or expensive to achieve in some circumstances.
Alternative options include automatic re-start based
on a time delay (i.e. a timer is built into the fan starter
so that, if the timer is pressed and the fan is turned
off, then it will automatically re-start (say) 10 minutes
later). ControlPED™and other technologies can also
be used to re-start fans.
Most fan starters used with flexible duct today will
provide an automatic pulse start so that the duct is not
blown apart on start-up.
553
Copyright © 2006 Taylor & Francis Group plc, London, UK
Monitoring fan amps is an important proxy for fan
condition. If amps are much lower than design, then it
is possible that the blades have worn and the fan is no
longer performing.
19 SUMMARY AND CONCLUSIONS
Booster, circuit and auxiliary fans play a critical role in
underground mine ventilation. There are many impor-
tant factors that need to be adequately understood for
safe and effective use of these types of fans. New
opportunities are being developed for using these types
of fans and in some cases, new modes of operation of
these fans are making new mining methods possible or
providing significant potential increases in productiv-
ity or reduction in capital or operating costs of existing
mining methods.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8
Recirculation in the potash mines of Saskatchewan
M. Detharet
Saskatchewan Labour, OHS Division, Mines Safety Unit
ABSTRACT: A detailed study of recirculation in production rooms of potash mines with possible courses of
action to reduce recirculation.
1 POTASH INDUSTRY IN SASKATCHEWAN
Saskatchewan is the first producer of potash in the
Americas with 8 underground mines and 2 solution
mines. Most of these mines were opened in the mid
to late 60’s and have kept expanding away from the
original shafts. Life expectancy for these existing
mines is still very high.
2 MINING METHODS
Besides solution mining, potash extraction is per-
formed by continuous miners (borer, mining machine,
etc.) unloading on a network of belt conveyors. Each
mine developed its own mining method that mostly
falls into the parallel mining or chevron mining cate-
gory. In parallel mining, a borer is advancing in straight
line, opening a room of the same profile as the borer.
Once the first pass is completed, the borer turns around
and cut a second and sometime third or fourth pass,
enlarging the original room accordingly. Most mines
excavate in one single horizon (side-by-sides passes),
but one mine is excavating two passes on the upper
level, followed by two passes on the lower level. In
chevron mining, a multiple-entry system of four to
five parallel drifts is cut and is intersected diagonally
by short section of drifts in order to ensure ventilation
and increase ore recovery.
3 VENTILATION ISSUES
Many issues are encountered in the ventilation of
potash mines:
– Given the cost of shaft sinking, all mines are still
operating today from the original shafts. This lead
to extreme distance from shafts to extraction rooms.
– Given the depth and thermal gradient, the under-
ground airstreams is naturally quite warm.
555
Copyright © 2006 Taylor & Francis Group plc, London, UK
– Potash dust, although considered innocuous, is
quite unpleasant to breathe and is produced in great
abundance by the borers and to a lesser extend by
every vehicle and at every conveyor transfer point.
– Fumes, mostly from diesel equipment, linger in the
air for a long time, given the generally low air speed
resulting from very large cross sections.
– Recirculation is a concern in potash mines but has
been accepted as a fact of life.
4 SASKATCHEWAN LEGAL FRAMEWORK
Saskatchewan’s Mines Regulations have been re-
enacted in 2003 and like its predecessors, and most
other provincial regulations, are mostly based on
hard rock environment. The Mines Regulations, espe-
cially as it pertains to ventilation, is ill-suited to deal
with the specific conditions encountered in potash
mines and a code of practice for underground ven-
tilation mines is being prepared by the Mine Safety
Unit of Saskatchewan Labour.
5 VENTILATION FUNDAMENTALS
Because mine main ventilation occurs in the turbu-
lent range, P = R × Q2 . Practically, increasing head
pressure by 100% only increases airflow by 41%.
Resistance R is directly proportional to the rubbing
surface (=length of the circuit × perimeter of the open-
ing) and is inversely proportional to the square area of
the opening.
However, because leakage (and therefore recircula-
tion) is not taking place in the turbulent range, leakage
is then quasi-proportional to the difference of pressure
on both sides of the structure. Thus, increasing the dif-
ference of pressure by 100% increases recirculation by
about 100%.
 Energy used = k1 × P × Q = k2 × Q3 . In other side, far above what the pollutant source could create
words, the energy used by a fan is proportional to alone.
the cube of the airflow generated. Thus a 26% “avoid- For the values in B to stabilize, it takes:
able leak” translates theoretically into an increased fan
energy consumption of …100%. 2 cycles at 5% recirculation
5 cycles at 25% recirculation
11 cycles at 50% recirculation
6 RECIRCULATION SIMULATION 27 cycles at 75% recirculation

A simplified and idealized simulation has been based The higher the recirculation, the longer it takes to
on a set-up that closely resembles a mining room with stabilize.
one single source of tracer gas (pollutant) and one With a cycle time of 11 minutes and 50% recircula-
single recirculation. tion, values that have been found to be fairly common,
NOTE: In all the following, recirculation is it takes 2 hours for the level of tracer gas (i.e. pollu-
expressed as the ratio of the volume of air escaping tant) to stabilize upwind from the release point (i.e. in
from the return airway to the volume of air available at the work area).
the face (thus including the re-circulated air itself). By Table 2 shows the relationship between recirculation
using that convention, recirculation can never exceed level and increase in pollutant.
100%. Recirculation is sometime expressed as the ratio
of the volume of air escaping from the return airway
Table 1.
to the volume of air entering the room. In this case
recirculation can exceed 100%.
In this simulation (see Figure 1 and Table 1) with
a tracer gas (i.e. pollutant) released at a concentra-
tion of 10 ppv (particle per volume, i.e. 1000 particles
released in a 100 air flow), the ultimate concentration
in B (which ideally should be nil) reaches:
Less than 1 ppv at 5% recirculation:
3 ppv at 25% recirculation
10 ppv at 50% recirculation
30 ppv at 75% recirculation

Recirculation can “amplify” the level of Tracer gas

(i.e. pollutant) to astonishing levels on the fresh air

Figure 1.

556

Copyright © 2006 Taylor & Francis Group plc, London, UK

 7 TESTING METHOD 8 INDUSTRY RESULTS

Only laboratory conditions allow for airflow measure-
ment with a precision of a few percents. Because air
is compressible and subject to turbulence, measure-
ments within a real environment provide results that
can easily be affected by a 10 to 15% imprecision. Most
recirculation is therefore of the same magnitude as
the imprecision of the value measured. Consequently,
recirculation is generally impossible to measure by
volumetric methods.
Because volumetric methods were impractical, a
tracer gas method, originally devised by CANMET,
is being used by the Mine Safety Unit to measure
recirculation in various underground potash mines.
The tracer gas used, SF6, can be detected at concen-
tration as low as 1 particle per billion and can generally
be measured with an imprecision of less than 2% of
the measured value.
Table 3 shows the results of the tests conducted through
all potash mines in Saskatchewan.
It is clear that the type of mining method has a huge
impact on the level of recirculation:
Straight rooms consistently return recirculation
between 7% and 13% while chevron rooms experience
recirculation ranging from 22% to 55%.
9 CONCLUSION AND POSSIBLE COURSES
OF ACTION
With the level of recirculation documented, it is fair to
conclude that if a typical parallel room recirculation
can increase the existing level of pollutant by up to
15%, while a typical chevron room recirculation can
increase the existing level of pollutant up to…122%.
Based on elementary physics, there is several ways
recirculation could be tackled:

Table 2. Increase in pollutant level vs. recirculation. 9.1 Eliminate secondary fans altogether
Secondary fans installed by the face usually have
Recirculation in (%) Increase in pollutant level (%)
others functions than just breathing requirement:
0 0 – Carry dust away from operators.
5 5 – Increase air speed to cool operators.
10 11
15 18 Regulatory airflow could eventually be generated
20 25 without fan by the face, but dust suction and operator
25 33 cooling would not then be provided. Could a smaller
30 43 fan, integrated into the machine, provide the same dust
35 54 control? Could air be cooled?
40 67
45 82
50 100
55 122
9.2 Reduce the physical size of the opening
60 150 In a mining room, air may find its way around the fan
65 186 through:
70 233
– Porosity of the Fabreene material itself.
– Cuts and scratches through the Fabreene.
– Gap between Fabreene and potash.
– Cracks in the surrounding potash.
Table 3. Recirculation in mining room.
Fabreene porosity is undocumented but would
Length Recirculation affect both mining methods in a similar fashion. How-
Pattern (m) (%) ever, comparison between straight rooms and chevrons
rooms using about the same surface of Fabreene, shows
Chevron 1,700 55 tremendous differences. Porosity is therefore not likely
Chevron 1,500 54 to be a significant contributing factor.
Chevron 1,300 50 All rooms tested recently were carefully inspected
Chevron 1,400 30
Chevron 600 22
before testing, some cuts and scratches were eventually
fixed and the Fabreene separation was again inspected
Parallel 800 13 during testing. Although not perfect, the Fabreene sep-
Parellel 1,400 8
Parellel 400 7
aration was always found to be in excellent shape,
Parallel 1,075 7 as good as can be expected underground. Cuts and
scratches need to be routinely monitored and fixed

557

Copyright © 2006 Taylor & Francis Group plc, London, UK

 but there is little to gain by trying to eliminate the very
last ones.
Fabreene seals to the back and floor was always
found to be excellent. However, stress adjustment is
relentlessly pushing blocks apart. Because straight
rooms Fabreene is not in contact with the walls, there is
no consequence. However, for chevron mining, every
panel of Fabreene is being affected, as the surrounding
walls move. Gap between Fabreene and surround-
ing potash is a major contributor to recirculation in
chevron rooms.
In both mining methods, the back and the floor were
found to be free of cracks. However, walls were found
to be heavily cracked. This is a direct result of stress
adjustment and cannot be eliminated. However, for a
straight room, where a continuous Fabreene is being
used, this has no effect on recirculation. For chevron
rooms on the contrary, every piece of Fabreene is con-
nected to two walls and any crack may be a potential
recirculation. Cracks through the surrounding potash
are a major source of recirculation for chevron mining.
9.3 Reduce fan air pressure
Fan pressure is in direct relation to the expected
airflow. As shown at the beginning, pushing a given
558
Copyright © 2006 Taylor & Francis Group plc, London, UK
volume of air through an opening of a given resistance
requires a certain air pressure. This is a physical value
that does not change, regardless of the “type” of fan
used. It is impossible to have the same airflow as before
in a given environment with less pressure than before,
unless…a single fan is replaced by a succession of
fans.
As demonstrated at the beginning, increasing head
pressure by 100% increases leaks by close to 100%
while only increasing air flow by 41%.
Several smaller fans spread along a return airway
could generate the same air flow as a single fan, but
with lower head pressure, therefore, less recirculation.
9.4 Relocate air pressure zone
Fans, or the outlet of the fan (end of the vent duct)
are located in the return airside of the room. Because
of this, the pressure push back polluted air across the
Fabreene separation, thus:
– Could fan(s) be located outside the room?
– Could fan(s) be located on the fresh airside?
– Could the fan, or the end of the vent duct, be further
down the return airside?
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Blind shaft drilling as a safer alternative to conventional sinking

A.J. Zeni & W.J. Maloney

Shaft Drillers International, Morgantown, WV

ABSTRACT: Safety has become a very large issue in recent years in the shaft sinking industry. The disastrous
ignition of methane in a coal mine shaft in January of 2003 has brought the issue of safety in conventional shaft
sinking into closer focus. MSHA, State Regulatory Agencies and Contractors have worked together in an effort
to eliminate such accidents in the future.

1 INTRODUCTION simple demand for more shaft openings has increased

MSHA reports that from 1982 through 2003 there were
15 fatalities in shaft and slope construction. In the same
period of time there were 43 methane ignitions in shafts
(most of these occurred while welding or cutting in
the shaft). Seven of the fatalities were due to methane
ignition, 7 were hoisting related, either falling or get-
ting struck by the conveyance and one falling into a
shaft because he was not tied off.
MSHA proposed in July of 2004 to modify the 30
CFR part 48 rules for miner training to include shaft
and slope workers, requiring them to be trained in the
same way as underground miners. The proposed rule
(Training Standards for Shaft and Slope Construction
Workers at Mines) was reviewed by representatives of
the shaft sinking and drilling industry. These persons
responded to the proposed rule with detailed com-
ments protesting the increased training requirements
because the working conditions were very different
and the labor force available to the shaft sinking
industry is not nearly as stable as it is for mining. The
blind shaft drilling industry had the most convincing
argument because all of the shaft construction work
is done on the surface. No persons are required to go
underground on a typical blind drilling project.
The state of West Virginia Board of Coal Mine
Health and Safety also proposed new regulations
regarding shaft/slope operations. These changes were
proposed in September of 2004 but never enacted. The
changes would have required more detail on design,
especially of water rings and more detail on methane
detectors and equipment.
The focus of this paper is shafts for mining appli-
cations, specifically coal mining because the danger
of methane gas is greatest in that environment. As
long-wall mining has expanded, more methane is liber-
ated, meaning ventilation requirements increase. The
the market for blind drilled shafts particularly bleeder
shafts.
Developments in large diameter drilling technology
will soon make it possible to mechanically drill shafts
of adequate size for major ventilation and access in
underground coal, limestone and other mines. Blind
shaft drilling technology has been improved steadily
over the past 40 to 50 years. The technique has been
limited so far to relatively small diameter shafts for
bleeder ventilation systems. It is now possible to drill
shafts of 20’ diameter and larger with a new generation
of drilling equipment being developed.
In addition to the bleeder shafts, there is a movement
toward separate blind drilled shafts in place of conven-
tional, split compartment intake/return shafts. These
shafts offer similar capacities in air movement and can
be positioned as near or as far apart as desired for the
best placement underground and for surface equip-
ment such as fans and escape hoists. In addition to
some of the design advantages, blind shaft mechanical
drilling also offers significant safety rewards. In spite
of these trends, most of the main shaft construction
for coal mines is still done conventionally.
2 HAZARDS OF SHAFT SINKING
By its nature conventional shaft sinking includes many
safety and health hazards. A partial list of these
hazards is:
• Falling of persons
• Falling objects
• Methane and other gasses
• Hoisting and communications failures
• Equipment operating in a small work area
• Respirable Dust

559

Copyright © 2006 Taylor & Francis Group plc, London, UK

 • Water Innundation 3 CURRENT LARGE DIAMETER SHAFT
• Back injuries DRILLING TECHNOLOGY
• Blasting
• Ignitions in the Shaft Shafts have been drilled using mechanical blind
• Concrete burns drilling in the US for more than 50 years. The tech-
• Escape in an emergency nology has improved gradually but much more rapidly
• Noise in the last 10 to 15 years. Blind drilling is much more
efficient, safer and competitive. The capability to drill
The industry has made great strides over the years in
larger diameters is still improving as market demands
reducing and eliminating these hazards but many still
for very large (>16 feet diameter) shafts increase.
pose significant problems.
Larger rotary tables have been built using AC electric
Mechanical blind boring of shafts can eliminate
or hydraulic power, cutterhead designs that mini-
many of these hazards. Since persons never have to
mize torque and optimize cuttings removal have been
work inside the shaft there is very little danger of
developed. Improved cutters are used that run longer
falling in the shaft or dropping objects and hitting per-
between trips.
sons below ground. The shaft is usually full of water
Drilling 14 to 16 diameter is becoming routine.
and therefore minimizes the possibility of methane
These shafts can be drilled economically and are now
accumulation in the work area. All of the workers are
favorably competitive with the conventional sinking of
in the same general area and are usually visible to
similar size shafts. Larger shafts, up to 24 diameter,
each other hence there is less chance of miscommu-
are achievable using blind drilling technology due to
nication. In conventional sinking, the bottom crew is
the acquisition of larger downhole components such as
working in a very tight space using suspended buckets,
mandrels, weights, drill pipe, cutterheads, cutters and
jumbo drills, muck excavators with swinging booms
stabilizers. This conventional method of shaft drilling
and many other items that can cause pinching or
has served our business well for many years and will
entrapment injuries. Blind drilling needs no bottom
continue to do so because the technology is advanced
crew. Very little respirable dust is generated in the fresh
and well suited to the mining industry. The only sig-
air environment of surface drilling. Back injuries are
nificant problem with this technology is that it is very
less likely because there are less restricted space lift-
equipment intensive.The investment required to obtain
ing requirements. There is no blasting required in the
the specialized tools for this work is so great that it
process of shaft drilling. Ignitions are impossible in a
is difficult to amortize in the niche market that blind
flooded shaft. Noise control is easier on a surface site
drilling occupies.
as is emergency escape.
Another area of development is the downhole
Conventional shaft sinking is very labor intensive.
drilling machine. Zeni Drilling developed a machine
The people needed for this work must be willing to
designated the VDS in the early 1990’s. This machine
work in wet, dirty and dangerous conditions. In today’s
was designed to work in a submerged environment
labor market there are fewer people willing to perform
and be steerable, lightweight and capable of very high
this type of work.
thrust with no parasite weight to push the cutters into
Given the above factors, mechanical shaft drilling
the rock. The machine achieved mixed success but was
appears to be safer than conventional sinking.
a good start into this technology. The weaknesses of the

Figure 1. Typical blind drilling worksite. Figure 2. 14 diameter cutterhead.

560

Copyright © 2006 Taylor & Francis Group plc, London, UK

 VDS were that it used drill pipe and a surface mounted
rotary table to rotate the cutterhead and it required an
electrical umbilical from the surface for control and
hydraulic unit power. The gripper pads that provided
the reaction to the thrust cylinders tended to slip in cer-
tain formations and it had structural weaknesses that
contributed to maintenance problems.
The potential of the VDS is very great. The ability
to apply hydraulic weight to the cutterhead makes the
Figure 3. Composite lining steel shell with 8 concrete.
Figure 4. Installing 8 steel casing into a blind shaft.
561
Copyright © 2006 Taylor & Francis Group plc, London, UK
production of the machine limited only by the abil-
ity to remove the cuttings from the hole. The VDS
must be made to operate in a blind hole environment
(submerged) and must have the thrust and rotary mech-
anisms operate at the machine. The ideal machine
would be similar to the Wirth V-mole or Zeni Full
Hole Drill but submersible. The submersible downhole
machine does not address the problem of installing a
liner in the shaft. The best method for installation of
a liner would be the same as a standard blind hole in
that the shaft would be finished and the liner installed
from the surface in a flooded shaft and cemented in
place.
4 SHAFT LININGS
Lining the shafts remains the most significant problem
with the mechanical drilling method. Conventional
sinking retains the distinct advantage of the abil-
ity to line and complete the shaft concurrently with
excavation.
The goal of the blind shaft drilling industry in
general is to eliminate the need for persons to enter
the shaft at all during construction. In very large
diameter shafts, the typical method of installing pipe
liners either of steel or concrete composite becomes
 impractical. Shotcrete can be applied remotely by
robotics but lacks the capability of sealing water zones.
There has been some thought about remotely placing
concrete but these techniques and proposed equipment
remain untried. Allowing persons to enter the shaft
for concrete lining after the excavation is complete is
somewhat safer but hazards still exist such as methane
gas or a toxic atmosphere.
The practical limit for steel linings is approximately
15 to 16 diameter. Composite linings could conceiv-
ably be larger say up to 20–24 diameter. A composite
liner requires an even larger shaft to be excavated to
allow for the thickness of the liner. In addition, a com-
posite liner must be built on site and presents problems
logistically. Other composite liner designs from the
past include steel, concrete and asphalt designs.
The future of drilling very large diameter shafts
blind depends more on technological developments
in fabrication and installation of linings that can be
installed remotely. Until that time, concrete linings
installed by humans in a shaft environment, appear to
be the most practical method available to use for shafts
over 18’ diameter. The process of installing concrete
liners in large diameter drilled shafts is reasonably
safe by conventional standards. All work is under-
taken inside the lined shaft with a minimal amount
of exposed rock walls. No blasting is done during the
normal lining process, however any other excavation
besides the plain shaft profile must be blasted. Water
562
Copyright © 2006 Taylor & Francis Group plc, London, UK
rings are nearly always necessary for concrete lined
shafts in water bearing areas. Water rings also present
some of the most hazardous environments in the shaft.
They create confined space for methane to accumulate
and sometimes are susceptible to roof and rib falls.
Besides the techniques described above, some of
the possibilities of future shaft linings may be:
• Remote Slipforming Concrete
• Lighweight composite linings made from extruded
fiber reinforced polymer
• Other methods of installing conventional steel or
concrete prefabricated sections
• Lightweight concrete combined with steel or
polymer
• Totally waterproof cast in place concrete or
shotcrete that requires no drainage
5 CONCLUSION
In the blind shaft drilling industry, as with any con-
struction/mining type business, safety of the work-
force is the overriding concern. Technology has
advanced since the early 1950’s to improve efficiency
and at the same time safety in the process. This techno-
logical progress will continue until mechanical shaft
drilling becomes the safest and most efficient way to
install vertical shafts in any ground for any purpose.
 11. Mine stoppings

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Stoppings: Technology developments and mine safety engineering

evaluations

H.C. Verakis
US Department of Labor, Mine Safety & Health Administration, Technical Support, Triadelphia, West Virginia

ABSTRACT: Advances in materials technology have resulted in newer products for the construction of stop-
pings used in USA underground coal mines. Some of the newer materials were developed for the construction
industry and were consequently introduced into the mining industry. Ventilation control devices such as stop-
pings, overcasts and undercasts have traditionally been constructed of concrete block or metal panels. Efforts
to reduce material and labor costs in the construction of stoppings and other ventilation controls have led to
the introduction and use of “nontraditional” materials. The “nontraditional” or newer materials may consist
of lightweight blocks composed of various cement-type compounds or panels composed of gypsum or other
mineral-type compounds, and plastic membrane and steel fabricated designs. Reduced labor and material costs
may be derived from such materials, while required safety properties are maintained.
An MSHA engineering evaluation program was developed to assess the suitability of such newer products
in terms of the safety standards for fire endurance and flexural strength as specified in USA Title 30, Code of
Federal Regulations, Part 75.333. The engineering evaluation program is described and a short description of the
fire endurance and flexural strength tests is presented. A discussion of the newer materials evaluated is included.

1 INTRODUCTION 25 or less. The flame spread index test specified in Part

When permanent ventilation controls such as stop-
pings, overcasts, and undercasts are installed in under-
ground coal mines, they are required to meet the safety
standards specified in USA Title 30, Code of Fed-
eral Regulations, Part 75.333 [1]. This safety standard
requires that permanent ventilation which includes
overcasts, undercasts, shaft partitions, permanent stop-
pings, and regulators be constructed of noncom-
bustible material. Noncombustible material is defined
in Part 75.301 [1] as a material which when used to
construct a ventilation control results in a control that
will continue to serve its intended function for one
hour when subjected to a fire test incorporating an
American Society for Testing and Materials, Interna-
tional, ASTM E-119-88 temperature/time heat input,
or equivalent [2]. Additionally, the ventilation con-
trol must meet a flexural strength that is equal to or
greater than a conventional 20.32-centimeter hollow
core concrete block stopping. The 20.32-centimeter
hollow core concrete block with mortared joints has
been tested and shown to have a minimum strength of
190.42 kilograms per square meter. The ASTM E-72-
80 is used to determine the flexural strength [3]. Also,
sealants or coatings applied to ventilation controls to
reduce air leakage must have a flame spread index of
75.333 is the ASTM E-162-87 [4].
The basis for the safety standard of fire endurance
and flexural strength relates to concrete block. Con-
crete block has been a long time traditional material
for the construction of mine stoppings. However, the
construction of a concrete block stopping is labor
intensive and time consuming. The blocks are heavy
and injuries from carrying and lifting the blocks can
result. Developments in materials technology have
resulted in newer, lighter materials to replace the
use of concrete block, particularly for the construc-
tion of stoppings. As an engineering aid in evaluating
newer materials and designs for ventilation controls to
assess their suitability as replacement for conventional
materials, a program was developed by the MSHA
Approval & Certification Center. Details are subse-
quently presented on the evaluation program and the
newer materials evaluated.
2 NEWER CONSTRUCTION MATERIALS
As a result of advances in materials technology, newer
products have become available for the construc-
tion of ventilation controls, particularly for stoppings
used in underground coal mines. Some of the newer

565

Copyright © 2006 Taylor & Francis Group plc, London, UK

 materials were developed for the building and con-
struction industry and were subsequently introduced
into the mining industry. Another development was the
use of fly ash as a component in the manufacture of
lightweight blocks. Ventilation control devices such as
stoppings, overcasts and undercasts have traditionally
been constructed of concrete block or steel panels.
Efforts to reduce material and labor costs in the
construction of stoppings have led to the introduc-
tion of newer lightweight materials for the construction
of ventilation controls. Traditional stoppings are usu-
ally constructed of hollow-core concrete blocks by dry
stacking or by binding with mortar between the block
joints. A difficulty with using concrete blocks is they
are heavy. A typical 20.32-cm wide by 20.32-cm high
by 40.64-cm long hollow core concrete block weighs
on average about 18 kg. Today, comparable lightweight
blocks weigh much less, which can be one third or less
weight, or around 12 kg or less. The lightweight blocks
are generally available in various sizes from 15.24 cm
thick and 20.32 cm or 30.48 cm high and 60.96 cm
long.
Some of the “nontraditional” or newer materials
consist of lightweight blocks composed of various
cement-type compounds or mixtures which may
include fly ash or inorganic oxides such as calcium
oxide and sand. Sometimes a small amount of plas-
tic fibers are used in cement-type block mixtures
to prevent micro cracks from developing into macro
cracks. There are also block units made from auto-
claved aerated concrete mixture. This type of unit is
based on calcium silicate hydrates in which low density
is acquired from reaction of the formulated materials
to produce macroscopic voids in the finished product.
Some of the “nontraditional” or newer materi-
als consist of lightweight blocks composed of vari-
ous cement-type compounds or mixtures which may
include fly ash or inorganic oxides such as calcium
oxide and sand. Sometimes a small amount of plas-
tic fibers are used in cement-type block mixtures
to prevent micro cracks from developing into macro
cracks. There are also block units made from auto-
claved aerated concrete mixture. This type of unit is
based on calcium silicate hydrates in which low density
is acquired from reaction of the formulated materials
to produce macroscopic voids in the finished product.
3 ENGINEERING EVALUATION PROGRAM
Because of the newer materials and designs being
introduced for the construction of ventilation con-
trols, an engineering program was established by the
MSHA Technical Support, Approval & Certification
Center to assess the suitability of the newer products
[6]. The program has been in effect for several years.
The engineering evaluations are intended to provide
information to mine operators and mine inspection
566
Copyright © 2006 Taylor & Francis Group plc, London, UK
personnel on those “nontraditional” or newer types
of ventilation control materials that meet the safety
standards prescribed in Part 75.333. “Nontraditional”
ventilation controls may consist of, but are not limited
to, flexible plastic membrane materials, unique steel
fabricated designs, or lightweight blocks composed of
fly ash and cement-type materials.
The engineering evaluation program is not intended
to evaluate the “traditional” ventilation controls, such
as hollow core or solid concrete block. The program
is voluntary. The MSHA regulations do not require a
manufacturer of materials to be used for stoppings or
other ventilation controls to be submitted for approval.
The manufacturer or mine operator would have to pro-
duce evidence to MSHA enforcement personnel that
the materials used in the construction of a permanent
ventilation control meet the prescribed requirements.
The MSHA voluntary program is designed as a mech-
anism or aid to the mine operator and MSHA enforce-
ment personnel about those newer materials that have
been found suitable.
The application procedure for an engineering eva-
luation can be acquired from MSHA’s home page at
http: / /www.msha.gov/techsupp/acc/application/asap
5006.pdf. In addition, test guidelines and performance
criteria for the evaluation of “nontraditional” ventila-
tion controls can be acquired from MSHA’s home page
at http://www.msha.gov/techsupp/acc/application/acri
5001.pdf.
An hourly fee is charge by the MSHA Approval
and Certification Center for the engineering evalua-
tions made. If a “nontraditional” ventilation control is
found to be suitable by MSHA from the engineering
evaluation, then a suitability number will be issued to
the applicant.
The type of information and data that an appli-
cant would need to provide consists of a description
of the product and trade name; composition of the
materials used to construct the stopping or ventilation
control; test data from an independent laboratory with
the results of the flexural strength test and the fire
endurance test; instructions for use of the product;
and program to maintain product compliance with the
specified safety requirements.
4 FIRE ENDURANCE & FLEXURAL
STRENGTH TESTS
The American Society for Testing and Materials,
ASTM E-119-88 is used for determining the fire
endurance (noncombustibility) of a “nontraditional”
stopping. The test specimen, which is to be represen-
tative of the proposed in-mine construction, is exposed
to a furnace or heat source of the E-119 test method.
The area of the test specimen exposed to the radiant
heat of the furnace is specified as not less than 9.29 m2 ,
 ASTM E119 Fire Test Heat Input
1000
900
Temperature in degrees C

800
700
600
500
400
300
200
100
0
0 10 20 30 40 50 60 70
Minutes

Figure 1. Time-temperature heat input for the ASTM E-119

test method.

with neither dimension less than 2.74 m. The criteria

to meet the fire endurance requirement is that the test
specimen must remain in place and not permit the pas-
sage of visible flames during the one-hour exposure
period. Also, openings in excess of two square inches
must not develop in the test specimen during the one-
hour exposure period. Figure 1 is an illustration of the
time/temperature heat input in the ASTM E-119 test Figure 2. Flexural strength test, ASTM E-72, on lightweight
block.
method.
The ASTM E-72-80 is used for determining the
flexural strength of a “nontraditional” stopping. Three
test walls of the stopping are prepared and separately
tested using theASTM E-72 test.The width of each test
wall is nominally 1.22 m and the height conforms to
approximately 2.44 m. The stopping must exhibit an
average transverse strength of at least 190.42 kg/m2 .
An example of an E-72 test being conducted on a
lightweight block stopping is shown in Figure 2.

5 ENGINEERING EVALUATIONS OF
NONTRADITIONAL VENTILATION
CONTROLS

Engineering evaluations of several different types of
“nontraditional” ventilation controls had been made
by the MSHA, A&CC. The ventilation controls ranged
from steel studding combined with gypsum or cement
board, plastic membrane material using construction
hardware and sealant, lightweight blocks using fly
ash and cement-type constituents, and an innovative
design using steel in a folding (book) type stopping.An
illustration of a “nontraditional” stopping constructed
and being sprayed with a cement-type material in a
coal mine is shown in Figure 3.
Figure 3. A “nontraditional” stopping constructed &
sprayed with cement-type material in a coal mine.
Those “nontraditional” ventilation controls evalu-
ated by the MSHA, A&CC, and which have been
issued a suitability number under the voluntary pro-
gram are listed on MSHA’s internet site.The list may be
accessed at http://www.msha.gov/TECHSUPP/ACC/
lists/ventcontrol/Stoppingsuitabilitynumbers.pdf
The list does not include those “nontraditional” ven-
tilation controls evaluated and found suitable by the

567

Copyright © 2006 Taylor & Francis Group plc, London, UK

 MSHA, A&CC prior to charging administrative fees
for evaluations.
6 SUMMARY
Newer materials for the construction of ventilation
controls, particularly stoppings, continue to enter the
mining industry. The emphasis for the newer materials
relates to being lighter that conventional hollow core
concrete block and quicker installation. Panels com-
posed of gypsum or other mineral-type compounds
or plastic membrane and blocks containing fly ash
are lighter than standard concrete blocks. Injuries
from lifting and handling may be reduced by using
these lighter weight materials for construction of stop-
pings and other ventilation controls. The steel and
composite-type designs may require more labor to
install. They are not advantageous for use in non-
uniform openings without additional materials and
effort to close openings and gaps. The most promising
of the “nontraditional” ventilation controls evaluated
appear to be the lightweight blocks.
Advancements continue to be made in the devel-
opment of newer, stronger, and lighter materials and
improved joint sealing compounds. Newer materials
that are developed for other industries will continue
to find application as materials for the construction
of ventilation controls. The ease of construction and
568
Copyright © 2006 Taylor & Francis Group plc, London, UK
the reduction in labor and material costs for ventila-
tion controls will continue to be overriding factors for
greater acceptance of “nontraditional” materials.
REFERENCES
[1] USA Code of Federal Regulations, Title 30, Part 75,
Subpart D – Ventilation, 75.301 and 75.333, July 1,
2005.
[2] Standard Test Methods of Fire Tests of Building Con-
struction and Materials, American Society for Testing
and Materials, ASTM E-119-88, 1988.
[3] Standard Methods of Conducting Strength Tests of
Panels for Building Construction, American Society
for Testing and Materials, ASTM E-72-80, 1980.
[4] Standard Test Method for Surface Flammability of
Materials Using a Radiant Heat Energy Source, Amer-
ican Society for Testing and Materials, ASTM E-162-
87, 1987.
[5] CBS News: The Osgood File, CBS Worldwide Inc.,
Developing “smart bricks” for construction, Wednes-
day, August 11, 2004.
[6] Verakis, Harry C., Engineering Evaluation of “Non-
traditional” Ventilation Controls, Society of Mining
Engineers Annual Meeting, 2003, February 24–26,
Cincinnati, OH, Preprints.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Development of new protocols to evaluate the transverse loading of

mine ventilation stoppings

T.M. Barczak & T.J. Batchler

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA

ABSTRACT: The Code of Federal Regulations (CFR) requires that the transverse load capacity of stoppings be
equal to or greater than traditionally accepted in mine controls, which for block stoppings is generally accepted
as 39 psf as referenced in the preamble to the CFR standard. This measure is based on physical testing of a
freestanding wall in accordance with ASTM E 72 specifications, where our past research has shown that the
dominant parameter is the tensile strength of the sealant. A new protocol based on rigid arch loading of the
structure is proposed to determine the true transverse load capacity of block stoppings. Arching is achieved by
the restraint of the stopping against the mine roof and floor, whereby compressive forces are developed within
the wall. A laboratory procedure using the NIOSH Mine Roof Simulator (MRS) to simulate rigid arching of
block stoppings was developed and results verified through full-scale in-mine tests. The rigid arch tests have
demonstrated transverse load capacities more than an order of magnitude higher than the ASTM E 72 evaluation
method from which the 39-psf requirement is derived. More importantly, the rigid arch tests have shown that the
load capability is dependent on the physical properties of the block and geometric properties of the wall and not
the sealant.

Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent
the views of the National Institute for Occupational Safety and Health.

1 INTRODUCTION by taking advantage of the compressive forces that are

In mining, stoppings are used to direct and separate
ventilation air courses and are designed primarily to
withstand air pressure differentials generated by the
mine fan that exert transverse loading against the high-
pressure side or face of the stopping. These pressures
can range from as low as 0.25 psi in the working sec-
tions of the mine to over 1 psi near the area of a bleeder
fan. Air blasts from roof falls, gas ignitions, and other
potential sources can generate localized areas of sig-
nificantly higher pressure that can destroy stoppings.
As a result, it is critical that the ultimate transverse
loading capability of stoppings be given consideration
in their design and evaluation.
There are no full-scale tests required for stoppings
to determine their load capacity. The current Code
of Federal Regulations (CFR) requirement is to test
4 × 8-ft sections of freestanding walls in accordance
with ASTM E-72 specifications (30 CFR, Part 75.333,
July 2005).This test inadequately determines the trans-
verse load capacity of actual in-mine stopping con-
structions when the mine the roof and floor, and ribs of
un-mined areas restrain the stoppings. his restraint cre-
ates significantly greater transverse loading capability
generated as the wall arches between the mine roof
and floor. As a result, the true transverse load capaci-
ties of mine ventilation stoppings are not known using
the current evaluation testing methodology.
Recently, a new generation of low strength, light-
weight blocks has been developed for mine ventilation
stopping constructions. While stoppings utilizing
these blocks have all passed the current CFR criteria,
it is believed that their true transverse load capac-
ity varies considerably. This is because the material
strength of the block types varies by as much of an
order of magnitude, and the material strength of the
different blocks correlates to the arching capability
of the restrained wall in the mine during transverse
loading. Without such knowledge, the design of mine
ventilation systems using these lighter-weight, but
lower-capacity alternative constructions can be mis-
leading, potentially exposing the mine to inadequate
ventilation control under some circumstances.
The objective of this paper is to describe a new
protocol to determine the transverse load capacity of
block stopping constructions. Using the unique biax-
ial loading capabilities of the National Institute for
Occupational Safety and Health’s (NIOSH) Mine Roof

569

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Simulator (MRS), rigid-arch conditions for stoppings
are simulated in the laboratory. Verification of the
procedure has been done through full-scale testing of
stoppings in a pressure chamber in the NIOSH Exper-
imental Coal Mine and in the underground Longwall
Gallery from stopping failure data gathered during
full scale explosion testing at the NIOSH Lake Lynn
facility. Ultimately, this should lead to more appropri-
ate design criteria and allow regulatory statutes to be
developed that will ensure a more accurate evaluation
of stopping transverse loading capability.
2 CURRENT CFR CRITERIA FOR STOPPINGS
Part 75.333 Ventilation Controls of the CFR requires
that stoppings be tested in accordance with ASTM E
72 specifications.
2.1 ASTM E 72 test specifications
The procedure requires testing of a nominal 4x8-ft sec-
tion of wall (ASTM E72, 1981). The wall is to be
constructed in the manner it will be used in the mine,
including the application of sealant when specified.
The test apparatus is shown in the diagram illustrated
in figure 1. The wall is placed on a steel channel which
rests on a cylindrical roller to prevent restrained end
conditions. The axis of the roller is parallel to the
face of the wall, allowing rotation to occur without
restraint, as the wall is flexed from the application of a
transverse load. Two reaction rollers and contact plates
positioned at the top and bottom of the wall allow the
wall to flex under the application of a transverse load
from the opposite face. Again, rollers are utilized to
prevent longitudinal restraint as the wall flexes. Trans-
verse loading is applied across the width of the wall
through a steel contact plate at quarter-height points
of the wall. Rollers in the form of a steel pipe are
again used to transfer the load from a central I-beam
through the contact plates, again to prevent any rota-
tional restraint from occurring. As the load is applied,
it is required that the load be recorded as a function of
the displacement at the mid-span of the wall height.
The maximum load normalized to the square foot area
of the wall is then defined as the transverse load capac-
ity for the wall. It is also required that three separate
walls be tested. The average transverse load capacity
from these three tests must exceed 39 psf to comply
with the CFR statute.
2.2 Inadequacies of current CFR specifications
Examination of the mechanics of the wall response
to transverse loading reveals the inadequacies of the
CFR test procedure. First, it is seen that great care is
taken to ensure that there is no longitudinal restraint
570
Copyright © 2006 Taylor & Francis Group plc, London, UK
SUPPORT ROLLER
SPONGE AND PLATE
RUBBER
TRANSVERSE LOAD
REACTION
FRAME
LOADING ROLLER AND PLATE
ROLLER
Figure 1. Diagram of test apparatus for conducting E-72
testing of stopping walls.
provided to the wall as the load is applied. Essentially,
the wall is considered freestanding and unrestrained
from vertical movement as it flexes or bends from the
application of the transverse load. The objective of the
test is to evaluate the flexural strength of the wall. Any
structure that is subject to bending produces tensile
stresses on one side of the structure and compressive
stresses on the opposite side of the structure. Typically,
the tensile strength of the material, being weaker than
the compressive strength, controls the capability of the
structure to withstand loads that produce bending.
Concrete has relatively low tensile strength, but a dry-
stacked block stopping has no effective tensile strength
because the joints are not bonded. Theoretically, the
transverse load capacity of a freestanding, dry-stacked
stopping would be provided only from the weight of
the block, which acts to provide a superimposed ver-
tical load on the structure. Even the heaviest typical
blocks, which weigh about 55 lbs, would not provide
enough axial loading to meet the 39-psf criteria.
The tensile strength for stoppings constructed in the
mine is actually provided by the application of sealant
to the face of the wall. First, this is obviously not the
primary function of the sealant. In order for the sealant
to be effective for controlling the transverse loading,
it must be applied to the anticipated low-pressure side
of the stopping or the face opposite the side where
the transverse load will be applied. If the ventilation
pressure could be reversed either intentionally or unin-
tentionally, then the sealant should be applied to both
sides of the stopping under these criteria. Since sev-
eral sealants are available, each with different material
properties, then the stopping should only be certified
with the specific sealant used in the evaluation test.
Furthermore, for a given sealant, the thickness of the
sealant contributes significantly to the effective ten-
sile strength and resulting transverse load capacity
of the stopping. How thick the sealant is applied in
the test program compared to the thickness normally
applied to such stoppings in the mine is another issue of
 concern. The test program should exclude abnormally
thick sealant applications, since in-mine constructions
are not likely to apply the sealant any thicker than is
necessary to prevent air leakage.

3 RIGID ARCH LOADING MECHANISM

In the mine, stopping walls are not freestanding struc-

tures as evaluated by the ASTM E 72 test referenced in
the current CFR criteria. Stoppings, as constructed in
the mine, bridge the distance between the mine floor
and the mine roof and are typically wedged in place at
the roof interface to provide a tight fit during installa-
tion. They also span the full entry width, butted against
and typically trenched into the un-mined pillars on
both sides. Hence, if the mine stoppings are restrained Figure 2. Three hinge arch illustration and free body dia-
by the mine roof and floor and pillars, this restraint gram of half-wall.
allows for a completely different loading mechanism
to occur, namely arching.

3.1 Description of arching

Arching is the mechanism that occurs when the cur-
vature of the stopping, specifically the extension of
the tension face of the stopping, as it bends under the
application of transverse loading is prevented by the
rigid contacts of the mine roof and floor. This arch-
ing of the wall produces a thrust force that acts at
the mine roof and floor interface, and produces com-
pressive forces within the wall that can dramatically
increase the transverse load capacity of the wall com-
pared to a freestanding condition. In the unloaded or Figure 3. Examination of the thrust force developed in
minimally transverse loading condition, the ends of the rigid-arching loading conditions.
wall are in full contact with the mine roof and floor and
the individual horizontal joints between the courses of
block are in full contact with each other. As the trans-
verse loading increases, the wall will begin to flex or
bend. Associated with the bending will be the opening
of the block joints along the mid height span of the
wall (location of the maximum positive moment), and
the opening of the interfaces between the blocks and
the mine roof and floor (location of the maximum neg- Where ρ = transverse load, psi,
ative moment). A three-hinged arch is formed where L = height of the wall, in,
the external moment caused by the transverse loading w = width of the wall, in,
(w × ρ × L2 /8 term in equation 1) is resisted by the P = resultant thrust force at the hinge points,
internal force couple (P × r), where r is defined as the lbs, and
width of the arch and P is the thrust force generated r = width of the arch, in.
by the arching. This condition is illustrated in the free-
body diagram in figure 2 and expressed mathemati- If it is assumed that the arching thrust (P) is
cally by equation 2. This equation can then be solved controlled by the compressive strength of the block
for the transverse pressure (ρ) as shown in equation 3. material and the “crush zone” is acting over an area of
the block equal to 2/10 the thickness of the wall (see
figure 3), then an expression for P can be derived as
given in equation 4. As shown in figure 3, this assump-
tion also results in the width of the pressure arch (r)

571

Copyright © 2006 Taylor & Francis Group plc, London, UK

 being equal to 0.8 × t. Substituting this expression for
r and the expression for P from equation 4 into equa-
tion 3 yields a solution for determining the transverse
load capacity of a stopping wall (equation 5).
Where ρ = transverse load, psi,
P = arching thrust, lbs,
t = thickness of wall, in,
L = wall height, in, and
fc = compressive strength, psi
An example is considered using a 6-in-thick wall that
is 72 in high and constructed from concrete blocks
with a compressive strength of 1,000 psi. The term
fc X (t/L)2 equates to 6.94 psi for this example, which
computes a predicted transverse load capacity of 8.9
psi or 1,279 psf.
Using these same relationships, it can be shown that
the transverse pressure acting on a full-scale stopping
can be computed from the measured horizontal force
(HF) at the base of a half-wall as used in the MRS
laboratory testing from equation 6, where w is the
width of the wall and (L/2) is the half-wall height.
3.2 Implications of rigid arching to stopping
design
Rigid arching indicates that the physical properties of
the block and the size of the mine opening must be
considered to evaluate the transverse loading capa-
bility of a stopping for design purposes. Intuitively,
higher strength block will provide greater transverse
loading capability, but increasing the thickness of the
block, or constructing the stopping with the wide side
of the block as providing the contact area, can also
greatly increase the transverse load capacity of the
stopping. Likewise, it is important to recognize that,
for given design parameters, the transverse load capac-
ity will decrease as the entry height increases. These
relationships are illustrated in figure 4 for conven-
tional concrete masonry units that have historically
been used to construct stoppings. These solid blocks
measure nominally 6 × 8 × 16 inches and have a unit
block compressive strength of 1,330 psi. As seen in
the figure, using the 8 in-wide construction nearly
doubles the transverse load capacity of the stopping
for a specific construction height. For comparison, for
572
Copyright © 2006 Taylor & Francis Group plc, London, UK
Narrow (6 in) width Wide(8 in) width
7,000
TRANSVERSE PRESSURE, psf
6,000
Wide
5,000
4,000
Narrow Width
3,000
2,000
1,000
0
40 60 80 100 120 140
HEIGHT OF WALL, inches
Figure 4. Impact of wall thickness and height on transverse
load capacity of a conventional concrete block stopping.
a stopping constructed in a 16 ft-wide by 8 ft-high
opening, the wide-side construction would require
192 blocks while the narrow-side construction would
require 144 blocks. In other words, for a 33 pct increase
in the number of block, the transverse load capacity
can be increased by 100 pct. Also, note that the trans-
verse load capacity would drop by a factor of four if
the construction height were doubled.
4 SIMULATING RIGID ARCHING THROUGH
BIAXIAL LOADING IN THE MINE ROOF
SIMULATOR
NIOSH has a unique load frame that was designed to
simulate the behavior of rock masses in underground
mining operations. It is called the Mine Roof Simulator
(MRS). This unique facility’s capabilities provide an
ideal framework in which to conduct rigid-arch testing
of stopping walls.
Since the load frame platens are 20 ft × 20 ft and
with a maximum vertical opening of 16 ft, the MRS
can accommodate full-scale stopping constructions.
The MRS is capable of providing controlled biaxial
loading in the vertical as well as one horizontal axis.
Up to 3 million lbs of vertical force can be applied
through a 24-in vertical stroke of the lower platen and
up to 1.6 million lbs of horizontal force through a 16-
in horizontal stroke of the lower platen. The loads
or displacements in these two axes can be applied
individually or simultaneously if desired.
4.1 Test protocol for simulating rigid arching
In order to simulate rigid arching, a half-height sec-
tion of a stopping wall was placed in the load frame
in a typical vertical orientation as it would be in the
mine. The upper platen position was adjusted to the
height of the block column and was hydraulically

Figure 5. Simulating rigid arching in the NIOSH mine roof
simulator.
clamped to maintain its position. The vertical posi-
tion of the lower platen was commanded to remain
constant. Hence, the fixed positions of the upper and
lower platen allowed them to act as rigid restraints. The
lower platen was then moved horizontally at a constant
velocity of 0.5 in/min, causing the wall to rotate (fig-
ure 5). As the base of the wall was forced to move
horizontally, crush zones were created at the ends of
the wall on opposite sides, consistent with the rigid-
arch loading mechanism. The horizontal force applied
to the base of the half-wall by the MRS was measured
and was equated to the transverse pressure acting on a
stopping wall using equation 6.
4.2 Transverse load determinations from MRS
half-wall testing
An example of a transverse loading test is shown in
figure 61 . The test consisted of a single column of
lightweight, autoclaved concrete block, stacked four
blocks high with the narrow side contact between
blocks. This block measures 5.875 × 8.375 × 17.250
in with a density of 42.5 lbs/cu ft resulting in a unit
block weight of approximately 21 lbs. Tests conducted
on an individual block indicated that the compressive
strength was 546 psi. The computed transverse pres-
sure determined from the measured horizontal force
during this test was 834 psf as computed by equation 6.
The measured vertical force is equivalent to the
arching thrust (P) as defined in equation 4. The arch-
ing thrust is a function of the elastic properties of the
concrete block and the contact area that develops as
the wall rotates. The kinematics of the wall suggests
that the contact area will decrease with increasing lat-
eral displacement. This essentially causes an increase
1
Mention of company name or product does not consti-
tute endorsement by the National Institute for Occupational
Safety and Health.
573
Copyright © 2006 Taylor & Francis Group plc, London, UK
MRS Half-Wall Test (Horizontal Force )
MRS Half-Wall Test (Thrust Force )
1.8 18
HORIZONTAL FORCE, kips
1.6
THRUST FORCE, kips
15
1.4
1.2 12
1.0 Thrust 9
0.8 Force
0.6 6
0.4 Horizontal
Force 3
0.2
0.0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
LATERAL DISPLACEMENT, inches
Figure 6. Example test result of 4 course half-wall arch test
of Kingsway block stopping wall in the MRS.
in the stress acting on the crushing zone until the
compressive strength of the material is reached. Once
the compressive (or shear strength) of the material is
exceeded, the thrust force will decrease and the trans-
verse pressure capacity of the stopping will decline.
Tests conducted on a similar half-wall constructed
from block made from conventional Portland cement,
sand, and aggregate material with a compressive
strength of 1,330 psi produced a maximum transverse
loading of 2,134 psf, or 2.6 times that of the autoclaved
block used in the previous test, which is consistent with
the difference in material strength. This provides addi-
tional validation for the application of arching theory
to stopping wall behavior.
5 FULL-SCALE LOAD VERIFICATON
TESTING
In order to confirm that arching was the proper loading
mechanism controlling the transverse load capacity of
mine ventilation stoppings and to verify the MRS half-
wall rigid-arch testing protocol, full-scale tests of stop-
ping walls were also administered. These tests were
conducted in the NIOSH Experimental Coal Mine
at the Pittsburgh Research Laboratory. Test data of
full-scale stoppings was also analyzed from NIOSH’s
explosion testing at the Lake Lynn Laboratory.
5.1 NIOSH PRL experimental coal mine tests
An air pressure chamber was constructed in one of the
crosscuts in the mine to provide a facility for static
loading of mine ventilation stoppings. Stoppings were
constructed in a crosscut measuring approximately
16 ft wide and about 80 in high. The test wall was
constructed in a normal dry-stacked fashion. The top
of the wall was tightened with wood wedges and the
gaps were filled with mortar and sealant to prevent air
leakage.
 MRS Half-Wall Test (Kingsway)
Experimental Mine Full-Scale Test (Kingsway)
TRANSVERSE PRESSURE, psf
500
400
300
200
MRS
Test Experimental
100 Mine Test
0 Exp. Mine
0 0.5 1 1.5
LATERAL DISPLACEMENT, inches
Figure 7. Comparison of MRS laboratory test and full-scale
mine test of low strength block stopping.
Pressurized air was injected into the chamber
between the barrier and the test wall through the air
intake port. The air pressure was increased gradually
in increments by adjusting a control valve on the pres-
sure line. This process continued until the air pressure
in the chamber blew out the wall. Three displacement
transducers were utilized to measure the lateral dis-
placement at the mid and quarter point heights of the
wall as the pressure was applied.
5.2 Test results and comparisons to half-wall
MRS tests
Two full-scale wall tests were conducted in the NIOSH
Experimental Coal Mine. For direct comparison pur-
poses, these were constructed using the same block
materials that were utilized in the MRS laboratory
tests. The first test utilized the lightweight autoclaved
concrete blocks, and the second test was a wall con-
structed from the conventional solid concrete aggre-
gate block. Comparisons of the MRS half-wall rigid-
arch test to the full-scale mine test are shown in figures
7 and 8. Figure 7 shows that the peak transverse loading
for MRS half-wall test was higher than that observed
in the full-scale mine test. Examining figure 7, it is
seen that the peak transverse pressure for the MRS
test occurred at a larger lateral displacement, sug-
gesting that the full-scale mine test may have failed
prematurely compared to that observed in the MRS.
The full-scale mine stopping also exhibited a stiffer
response than the MRS test. This may be caused by
loading of the wall from convergence of the mine entry
or by a shorter arching height than was considered in
the MRS test. As shown in figure 8, only a 4-course
and 6-course-high half-wall was tested in the MRS for
the conventional concrete block, while the in-mine test
was constructed with 10 courses (full-height). There-
fore, a direct comparison was not provided. However, it
is seen from the figure that, as expected, the full-scale
mine test fits between the two laboratory tests. It is
Copyright © 2006 Taylor & Francis Group plc, London, UK
Exp. Mine Full-Scale Test --10 course high
MRS Half-Wall Test -- 4 course high
MRS Half-Wall Test -- 6 course high
Theoretical Prediction -- 5 courses high
2,400
TRANSVERSE PRESSURE, psf
4 course
2,000 high test
Theoretical
1,600 Prediction
1,200
2 2.5 800
Test
400
6 course high test
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
LATERAL DISPLACEMENT, inches
Figure 8. Comparison of MRS laboratory test and full-scale
mine test of high strength block stopping.
concluded that the MRS laboratory half-wall tests
reasonably predict the full-scale mine tests in both
cases.
These results are also consistent with the arch
mechanics theory presented in the previous section,
which indicate that higher walls will have less trans-
verse load capacity than shorter walls and that weaker
block materials will provide less transverse load capac-
ity than higher strength block materials. These rela-
tionships were expressed mathematically by the fcX
(t/L)2 term (compressive strength of the material times
the square of the ratio of the wall thickness to the
wall height) in equation 5. Figures 9 and 10 plot the
measured transverse pressure for both the half-wall
laboratory tests and full-scale mine tests as a func-
tion of the fcX (t/L)2 term. The chart shows a strong
correlation of the half-wall transverse pressure mea-
surements to this term. Also shown on the chart are the
full-scale mine tests, and again, it is seen that the mine
tests also fit this correlation very well. The theoretical
design curve produced from equation 5 over-predicts
the measured responses for the lower fc X (t/L)2 values,
and more accurately predicts the transverse pressure
as the value increases. It is believed that the error lies
in the arch thrust moment. The current formulation
does not include the lateral displacement of the wall,
which occurs throughout the loading history. Lateral
displacement reduces the arch thrust moment and the
resulting force couple. This topic is currently being
addressed in the continuing research at NIOSH.
Additional full-scale transverse loading tests of
stoppings were conducted at the NIOSH Lake Lynn
Laboratory (Sapko, 2003 and Weiss, 2004). Both
hollow-core and solid, high-strength, concrete block
stoppings were evaluated in this study. The hollow-
core block had an average material compressive
574
 MRS Half-Wall Test Data (Kingsway)
Experimental Mine Full-Scale Test (Kingsway)
Rigid Arching Theory Predictions Trendline
1,000
900
Transverse pressure, psf

y= 229.43x - 179
800 R2 = 0.98
700
600
500
400
300 Experimental
200 Test
100
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00
fc x (t/L)2

Figure 9. Comparison of MRS laboratory test, full-scale test, and theoretical predictions of transverse load capacity of various
stopping constructions using low strength cellular concrete block.

MRS Half-Wall Test Data (Klondike 1,330 psi)

MRS Half-Wall Test Data (Klondike 1,727 psi)
Experimental Mine Full-Scale Test (Klondike 1,330 psi)
Lake Lynn Full-ScaleTest (Conventional Solid Block 1,900 psi)
Lake Lynn Full-ScaleTest (Modified Block Strength 1,500 psi)
Lake Lynn Full-ScaleTest (Hollow Core Block 1,456 psi)
Rigid Arching Theory Prediction Trendline
3,000
TRANSVERSE PRESSURE, psf

y = 233x - 681
2,500
R2 = 0.99

2,000

1,500

1,000

500

0
0 2 4 6 8 10 12 14 16 18
fc x (t/L)2

Figure 10. Comparison of MRS laboratory tests, full-scale tests, and theoretical predictions of transverse load capacity of
various stopping constructions using conventional solid and hollow core block.

strength of 1,456 psi and the solid block an average shown in the chart in figure 10, the mine test data is
material compressive strength of 1,900 psi. Typically, closer to the MRS laboratory tests when the compres-
the block strength achieved in a unit block or column sive strengths are reduced to from 1,900 to 1,500 psi.
of block measurement is considerably less than the The stopping walls constructed in the crosscuts were
material strength tests that are conducted under ASTM 12 courses high (7.5 ft), 6-in thick, and approximately
specifications on small scale cylinders or cubes. As 20 ft in length.

575

Copyright © 2006 Taylor & Francis Group plc, London, UK

 6 SUMMARY AND CONCLUSIONS
Stoppings are a key component of underground mine
ventilation systems. Permanent stoppings are often
constructed from some form of concrete block, typ-
ically dry-stacked to form a wall, equal in thickness to
the narrow or wide dimension of the block, and bridg-
ing between the mine roof and floor and pillar ribs.
The criteria for approval of block stoppings to be suit-
able for coal mine use in the United States is generally
39 psf of transverse load capacity in a freestand-
ing loading condition. This specification is based on
ASTM E 72 testing requirements. Based on the work
presented, this standard does not provide an accurate
representation of the loading conditions that occur in
the mining situation if the restraint of the mine roof and
floor is considered. For dry-stacked stopping construc-
tions, the transverse load capacity under the ASTM
E 72 criteria is primarily determined by the tensile
strength of the sealant. Any block material, regardless
of its physical properties, can be made to pass this
test criterion for use in underground coal mines pro-
vided the sealant is strong enough and can adhere to
the surface of the block.
The restraint provided by the mine roof and floor
and coal pillars allows the stopping wall to arch
between these abutments as the wall flexes and bends
from the application of transverse loading.Arching has
long been recognized as a valid loading mechanism
that can more realistically show the increased capabil-
ity of jointed structures to resist loading induced by
bending when end restrained. Arching relies on com-
pressive forces within the wall structure to offset the
bending moment induced by the deflection of the wall
from the application of transverse loading. For dry-
stacked stopping constructions, which have no tensile
strength across the joints except for the sealant on the
face of the joint, these compressive forces can increase
the transverse load capacity of a stopping by two orders
of magnitude compared to the freestanding condition.
A static analysis of the arching condition was exam-
ined to develop a theoretical relationship for predicting
the transverse load capacity of a stopping. The analysis
indicates that both the compressive strength (fc ) of the
construction material and the thickness (t) of the wall
have a significant impact on the transverse load capac-
ity of a stopping. Increasing the thickness of the wall
will cause an increase in the force couple developed
by the resultant compressive forces acting on the wall,
and thereby increase its transverse loading capacity.
Conversely, increasing the height (L) of the stopping
will reduce the force couple, and thereby reduce the
transverse load capacity of a stopping. None of these
factors is considered in the current criteria for mine
ventilation stoppings. A theoretical design equation
to predict the transverse loading was developed using
these key factors, expressed by the term fc X (t/L)2 .
576
Copyright © 2006 Taylor & Francis Group plc, London, UK
A laboratory testing protocol to simulate rigid arch-
ing of stopping walls by biaxial loading in the NIOSH
MRS was developed. This process is most easily sim-
ulated in the MRS by testing a half-height section of
wall. The wall is restrained vertically by the fixed verti-
cal position of the load frame platens, thereby acting as
rigid end restraints simulating the mine roof and floor.
The lower platen is then moved laterally, causing the
base of the wall to displace with the platen and causing
the wall to rotate accordingly, similar to the three-hinge
theory. Crush zones are created at edges of the half-
wall in the areas where these two hinges would occur
in a full-height wall. By measuring the horizontal
forces applied to the base of the wall by the MRS, the
transverse load capacity of the wall can be determined.
Tests were conducted in the MRS on two different
block types using this testing protocol. The two block
types were selected for study because of their differ-
ing physical characteristics. One block was made from
a low-density, autoclaved concrete material (546 psi
compressive strength) while the other block was made
from a more conventional material – Portland cement,
sand, and aggregate (1,330 and 1,727 psi compres-
sive strength). Tests were conducted on several wall
heights ranging from 5 to 10 ft. For these tests, the
transverse load capacity ranged from 138 to 834 psf
for the lightweight block and 96 to 2,136 psf for the
conventional block. Comparing these results to the
39 psf requirement clearly shows the disparity caused
by the assumptions made in the boundary conditions,
freestanding as considered in theASTM E 72 test spec-
ification and the rigid arch conditions being proposed
here as a more accurate representation of the actual in
mine conditions.
The MRS half-wall rigid-arch testing methodol-
ogy was verified with two full-scale tests of stopping
walls in the NIOSH Experimental Coal Mine at the
Pittsburgh Research Laboratory and two full-scale
explosion tests at the NIOSH Lake Lynn Laboratory.
Overall, good agreement was shown between the MRS
tests and the full-scale mine tests.
In conclusion, rigid arch stopping design would be
a departure from the current freestanding wall design
assumed in the ASTM E 72 specifications cited by the
CFR. For arch loading conditions, the physical prop-
erties of the block and the size of the mine opening
should be examined to determine the proper design for
a stopping application. The sealant would no longer be
considered to impact the transverse load capability of
the stopping. This approach, which will be presented
and discussed with MSHA and the ASTM, could lead
to a safer mine environment for mineworkers in under-
ground coal mines by distinguishing the transverse
load capacity of different stopping designs; opposed
to the current system that permits stoppings of widely
varying transverse loading capabilities to be employed
in the same environment.
 REFERENCES
ASTM Designation E 72-80. Standard Methods of Con-
ducting Strength Tests for Building Construction. Annual
Book of ASTM Standards, Vol. 04.07, March, pp. 283,
293, 1981.
Code of Federal Regulations, Title 30 Mineral Resources,
Part 75 Mandatory Safety Standards – Underground
Coal Mines, Subpart D – Ventilation, Section 75.333
Ventilation Controls, pp. 471–473, July, 2005.
577
Copyright © 2006 Taylor & Francis Group plc, London, UK
Sapko, M.J., E.S. Weiss & S.P. Harteis. 2003. Alternative
Methodologies for Evaluating Explosion-Resistant Mine
Ventilation Seals. Published in the 30th International Con-
ference of Safety in Mines Research Institutes, South
African Institute of Mining and Metallurgy.
Weiss, E.S., S.P. Harteis & K.L. Cashdollar. 2004. NIOSH-
PRL Evaluation of the Kennedy and Solid-Concrete-
Block Stopping Designs for MSHA during LLEM Explo-
sion Tests, Sept., 40 pp.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Novel stopping designs for large-opening metal/nonmetal mines

R.H. Grau III

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National
Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh, PA, USA

G.M. Meighen
Wabash Holding Co, Foundation Coal Co, Keensburg, IL, USA

ABSTRACT: Improving the air quality in metal/nonmetal mines is important in protecting the health of
miners. Air quality can be improved in metal/nonmetal mines by developing proper ventilation techniques. Mine
ventilation systems require stoppings to direct the airflow and establish pressure differentials throughout the
mine. Due to their large size, stoppings in large-opening mines are cumbersome and costly to construct and
maintain. As part of its mining health and safety program, the National Institute for Occupational Safety and
Health (NIOSH) designed and tested two novel stoppings (Super Stopping and the EZ-Up Curtain Stopping) at
its Lake Lynn Laboratory (LLL). The Super Stopping is designed as a long-term permanent stopping for use in
the main entries of the mine, while the EZ-Up Curtain Stopping, although very durable, is designed more for
portability and ease of installation. This paper describes the design, materials, and construction methods used for
these stoppings as well as their performance and durability when subjected to tests simulating actual production
face blast pressures.

1 INTRODUCTION large numbers of stoppings requiring a significant

Stoppings play a significant role in underground mines
by coursing and separating ventilation airflows. The
utilization of stoppings, combined with proper mine
designs, are essential components of a reliable venti-
lation system (Grau 2002, 2004a, b, Head 2001a, Krog
2004). In addition, supplying sufficient fresh air quan-
tities to work areas to remove airborne contaminants
will significantly improve air quality (Head 2001b,
Robertson 2004, Haney 1998) thereby improving min-
ers health. The design, construction, and maintenance
of effective stoppings in large-opening mines are often
challenging for mine operators. The choices for the
design, materials, and construction techniques depend
on a number of site-specific utilization factors, such
as the intended life and required effectiveness of the
stopping.
Stoppings have not always been widely used in
large-opening stone mines. Capital expense, construc-
tion, and maintenance problems have impeded this
segment of the mining industry from fully utilizing
stoppings. Their use is particularly problematic in the
larger, more established mines where stoppings have
not previously been incorporated into the mining plan.
In those mines, retrofitting the operation to include the
use of stoppings to course the air requires building
investment in time and expense.
The Super Stopping and EZ-Up Curtain Stopping
are designed for use as a practical and cost effec-
tive means to control and direct ventilation airflows
throughout the mine. A reduction in worker exposure
to diesel particulate matter (DPM), gases from pro-
duction blasts and diesel equipment, and silica dust
(Chekan 2002, 2004) can be accomplished by using
an enhanced ventilation system that includes these
stoppings.
2 SUPER STOPPING
The Super Stopping (Fig. 1) was constructed by
NIOSH at the LLL in cooperation with several man-
ufacturers. The stopping is designed for use as a
permanent, long-term ventilation control, and is con-
structed from oversized Omega Blocks, which are low
density, composite cement and fly ash blocks manufac-
tured by Burrell Mining Products, Inc. The individual
blocks are 1.22 m (48 in) long by 1.22 m (48 in)
high by 0.81 m (32 in) wide and weigh approximately
544 kg (1200 lb) each. The Super Stopping constructed
at LLL is 17.4 m (57 ft) wide by 9.1 m (30 ft) high.
Construction of the Super Stopping is conducted in

579

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 1. Schematic of Super Stopping. Figure 2. Setting blocks for construction of a Super Stop-
ping using extended reach fork lift and precast lifting holes.
three phases: Phase I – Site Preparation and Founda-
tion Building; Phase II – Block Stacking; and Phase
III – Capping and Sealing.

2.1 Phase I – Site Preparation and Foundation

Building
For a stable installation, the stopping should be built on
a level surface which is best achieved using a concrete
footing. The approximate footing dimensions are the
distance across the mine opening width by 1.12 m (44
in) in depth, with a minimum 0.15-m (6-in) concrete
thickness. A footing width of 1.12 m (44 in) provides
an extra 0.15 m (6 in) on each side of the 0.97-m (32-in)
wide stopping block. The footing should consist of
20,680-kPa (3,000-psi) standard mix concrete, embed-
ded with standard 0.15-m (6-in) by 0.15-m (6-in) wire
mesh reinforcement material, which is sufficient to
withstand the weight of the wall.
2.2 Phase II – Block Stacking
An extended reach forklift is used to lift and posi-
tion the blocks (Figs. 2, 3) during the stacking phase.
The blocks have two factory cast holes on one side to
receive the forklift tines. Articulation of the blocks is
necessary for their proper alignment in the structure
to maintain stability and vertical trueness. It is critical
to establish a stable structure as the stopping height
increases. The first course of blocks is aligned along a
chalk line struck on the concrete footing.
Each block is secured in place with a high grade
polyurethane construction adhesive. The adhesive
used on the Super Stopping at LLL was the Touch
‘N Seal Mine Mortar Adhesive manufactured by Con-
venience Products. The adhesive is applied in a zigzag
pattern with each line spaced approximately 5.08 cm
(2 in) apart and distributed on all contacting sides of
each block (Fig. 3). The blocks must be positioned
tightly to ensure proper adhesive distribution between
Figure 3. Placing Super Stopping block in place after
applying adhesive.
blocks. The standard masonry technique of staggering
blocks on successive layers is recommended to maxi-
mize the strength of the stopping (Fig. 1). The fly ash
blocks can be scribed and readily cut to fit with a sim-
ple hand saw or axe. Successive layers of blocks are
laid until the remaining gap between the top of the
blocks on the highest row and the roof is less than the
1.22-m (48-in) height of a full block.
2.3 Phase III – Capping and Sealing
To complete the structure, smaller fly ash blocks,
0.61 m (24 in) by 0.41 m (16 in) by 20.32 cm (8 in) or
15.24 cm (6 in), are used to cap the top of the stopping.
As with Phase II construction methods, the smaller cap
blocks are staggered and cut to conform to the roof as
well as the ribs, and secured in place with an adhesive.
The final row of cap blocks should be placed within
approximately 7.62 cm (3 in) of the roof and walls of
the mine opening.
The remaining gaps at the top and sides of the
stopping are sealed using an expansion foam product.
Since the foam is applied using a handheld apparatus,

580

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Figure 4. Schematic of EZ-Up Curtain Stopping that is
constructed using fabric strips.
access to areas beyond reach is achieved using a man-
basket.The stopping constructed at the LLL was sealed
using Touch ‘N Seal Mine Foam manufactured by
Convenience Products. The foam is applied to approx-
imately half the depth of the blocks, or until expansion
of the foam completely fills the voids. The expansion
foam acts to conform to the shape of both the cut blocks
and surfaces of the mine opening, thus increasing
structure stability. Since the fork lift holes do not pass
through the entire block, they do not require sealing.
3 THE EZ-UP CURTAIN STOPPING
The EZ-Up Curtain Stopping (Fig. 4) was also con-
structed by NIOSH at the LLL in cooperation with
several manufacturers. This stopping is designed pri-
marily for use as a temporary or portable stopping,
although test results indicate it may remain useable
for an extended time. The EZ-Up Curtain Stopping
can be constructed from a variety of fabrics includ-
ing mine brattice. Evaluations were performed on
EZ-Up Curtain Stoppings composed of two separate
curtain materials: (1) NOVA-Shield RU88XFR-6, a
high density polyethylene woven fabric, manufactured
by Intertape Polymer Group, Inc. and (2) Dura-Skrim
D15CFB, which is composed of two layers of film
laminated with polyethylene that sandwiches a scrim
reinforcement, manufactured by Raven Industries, Inc.
The EZ-Up Curtain Stopping is manufactured with
a sewn-in loop so it can easily be laced onto tubing
while laid out on the ground. It is then hoisted to the
mine roof with a strap and ratchet mechanism system.
Since the EZ-Up Curtain Stopping is constructed pri-
marily on the ground, workers’ exposure to accidents
related to the typical 6.10-m (20-ft) to 12.20-m (40-ft)
working heights associated with large-opening mines
is reduced.
581
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 5. Ratchet assembly for hoisting EZ-Up Curtain
Stopping.
The EZ-Up Curtain Stopping is constructed in four
phases: Phase I – Preparation and Ratchet Assem-
bly Installation; Phase II – Curtain Preparation; Phase
III – Curtain Hoisting; and Phase IV – Securing the
Perimeter and Floor.
3.1 Phase I – Preparation and ratchet assembly
installation
The walls and roof of the stopping installation area
should be fully scaled to prevent injuries from falling
rock during construction, and to provide a solid foun-
dation to secure the stopping to these surfaces. The
stopping is attached to the mine roof with four ratchet
assemblies anchored by roof bolts (Figs. 4, 5). For the
LLL stopping, a 30.0-cm (12-in) long by 2.5-cm (1-in)
diameter roof bolt that was anchored in competent roof
material provided substantial support to suspend the
curtain.
The ratchet assembly (Fig. 5) was constructed by
welding the ratchets to U-angle steel stock, drilling
two holes into the stock for the roof bolts, and welding
a nut on the ratchet for easier winching. The ratchets
used for these assemblies are standard off-the-shelf
ratchets used for equipment tie-downs on trucks. The
four ratchet assemblies are positioned in a straight
line across the roof of the entry. One ratchet assembly
should be mounted near each entry wall, with the two
additional assemblies spaced evenly across the entry.
After mounting, the ratchet mechanisms are ready to
accept the straps for lifting the curtain.
3.2 Phase II – Curtain preparation
The fabric size used for the EZ-Up Curtain Stopping
should be full entry width and height, plus several feet
excess in both dimensions to allow for attachment to
the rib and for sealing on the mine floor. A stopping
composed of material strips that can also be used as
 a drive through stopping can be constructed by hang-
ing strips of fabric as shown in Figure 4. A similar
designed, but one-piece fabric stopping, can be made
by having the fabric distributor sew the strips together
or, if available, ordering from the manufacturer a cus-
tom designed one piece fabric to fit the entry. Both a
hanging strip stopping and a one piece stopping were
constructed and tested for blast pressures at LLL. In
either case, the fabric along the top edge of the fabric
material is folded over and sewn at the factory to form
a loop, through which 5-cm (2-in) diameter tubing or
conduit is threaded (Fig. 4). The individual lengths
of tubing can be joined together with either threaded
or tapered connections to span the entire length of the
entry. For the hanging strip curtain stopping, to limit air
leakage, each curtain panel is overlapped horizontally
approximately 0.6 m (2 ft) onto the adjoining panel.
The loop at the top of each panel must be cut away on
one side to allow for the 0.6-m (2-ft) overlap.
3.3 Phase III – Curtain hoisting
The lift straps are suspended from the ratchet mecha-
nisms and attached to the tubing through a narrow slot
cut with a small handheld electrical grinder (Fig. 4).
The flat plate hook is fitted into the slot, and bolted
if necessary, to secure attachment to the tubing. After
attaching all lift straps, the EZ-Up Curtain Stopping is
lifted using the ratcheting mechanism.
Lifts should be made approximately 1.5 m (5 ft) at
a time by alternately ratcheting each mechanism. This
will ensure the curtain tubing is kept as level as prac-
tical (Fig. 6). The curtain is hoisted until the tubing is
flush with the pulley, which will place the top of the
curtain near the roof of the mine entry.
3.4 Phase IV – Securing the perimeter and floor
The final step in the installation process is to secure
the EZ-Up Curtain Stopping to the walls using vertical
boards with dimensions of 3.8 cm (1.5 in) by 8.9 cm
Figure 6. Raising the EZ-Up Curtain Stopping into position.
582
Copyright © 2006 Taylor & Francis Group plc, London, UK
(3.5 in) by 3 m (10 ft) (Fig. 4). The first set of boards
should be bolted to the walls from the roof to the floor
in line with the edge of the curtain. Excess fabric on the
sides of the curtain is then wrapped onto a second set
of vertical boards. This wrapping prepares the curtain
for attachment to the existing wall mounted boards to
create a sandwich effect. Care should be taken to avoid
wrapping the curtain too tight, or stretching the fabric,
since the curtain should be able to yield slightly when
subjected to variable pressures. The wrapped curtain
boards on the prototype stopping at LLL were attached
to the wall boards using 0.6-cm (0.25-in) diameter by
10-cm (4-in) long lag screws. Test results, described
below, revealed that larger or different anchors would
be better. When the curtain is anchored to the rib,
excess curtain material along the bottom is trimmed
to the desired length to complete the perimeter.
The curtain stopping can be further secured by lay-
ing sand bags on the excess curtain material draped on
the floor. The bags serve two purposes: they seal the
stopping from excessive leakage, and they provide for
pressure relief to help reduce stopping damage in the
case of a face blast that produces excessive pressures
against the stopping. In an overpressure scenario, the
curtain will be blown free from the sand bags prevent-
ing damage, but the sandbags will need to be replaced
to reestablish the seal.
4 RESULTS OF BLAST TESTING
Stoppings in an underground metal/nonmetal mine
must withstand air blasts associated with development
and production blasting. Evaluation of the structural
integrity of such stoppings can be investigated by sub-
jecting them to unconfined blasts. Performance test-
ing using consecutive charges of 1.4 kg (3 lb), 2.3 kg,
(5 lb), 3.2 kg, (7 lb), and 4.5 kg (10 lb) of 50/50 Pen-
tolite (50% TNT/50% PETN) unconfined explosive
charges were conducted at LLL on the Super Stop-
ping and the EZ-Up Stopping. The explosive charges
were positioned at about 1.83 (6 ft) above the floor
in the entry, approximately 18.29 m (60 ft) from the
stoppings.
The unconfined blast tests confirmed the integrity
the stopping construction and the materials used. The
Super Stopping evaluation included observing how
well the stopping held up to the blasts, i.e. the strength
of the adhesive between the blocks, the effectiveness
of the perimeter foam, and the overall stability of the
stopping. Similarly, the integrity of the EZ-Up Curtain
Stopping was tested for material strength, especially
at the critical high-tension perimeter areas. From the
multitude of blasts test performed, the stoppings were
subjected to an incident maximum peak pressure of
up to 17.2 kPa (2.5 psig) and up to an incident max-
imum pulse pressure of 0.14 kPa/s (0.02 psig-s). The
 incident maximum peak pressure for each test was:
1.4 kg (3 lb), 9.0 kPa (1.3 psig); 2.3 kg (5 lb), 11.7 kPa
(1.7 psig); 3.2 kg (7 lb), 14.5 kPa (2.1 psig); and 4.5 kg
(10 lb), 17.2 kPa (2.5 psig). Although the exact rela-
tionship between the pressures created and acting
on stoppings from free standing blast charges and
actual face production blasts is unknown, actual pres-
sures from production blasts in operating stone mines
have been measured at approximately 6.9 kPa (1 psig)
(Mucho et al. 2001). The Super Stopping showed
no structural damage from the tests, although small
amounts of superficial foam sealant was dislodged
during the 2.3-kg (5-lb), 3.2-kg (7-lb), and 4.5-kg
(10-lb) explosive charge blast tests. The EZ-Up Cur-
tain Stopping had no rips or tears in either material
tested due to blast pressures. However, during the
3.2-kg (7-lb) blast, several lag screws pulled out that
were sandwiching the curtain material to the wood
plank that was secured to the rib but damage was
not increased during the blast using the 4.5-kg (10-
lb) charge and the overall integrity of the stopping
remained intact.
5 CONCLUSION
The Super Stopping and EZ-Up Curtain Stopping
developed by NIOSH, in cooperation with sev-
eral manufacturers, provide a practical and effective
means to direct ventilation airflows within large-
opening mines. The stoppings are durable and with-
stood test pressures that may be well above nor-
mal mine production blast pressures. The stoppings
are constructed using economical, readily available
materials. The stoppings are a practical and effective
means to improve ventilation systems in large-opening
mines, and thus reduce worker exposure to airborne
contaminants.
DISCLAIMERS
The findings and conclusions in this report are those of
the authors and do not necessarily represent the views
of the National Institute for Occupational Safety and
Health. Mention of any product or company does not
constitute endorsement by NIOSH.
583
Copyright © 2006 Taylor & Francis Group plc, London, UK
REFERENCES
Chekan, G.J., Colinet, J.F. & Grau III, R.H. 2002. Silica dust
sources in underground metal/nonmetal mines: two case
studies. Transactions 2002: 312: 187–193. Littleton, CO:
Society for Mining, Metallurgy, and Exploration, Inc.
Chekan, G.J., Colinet, J.F. & Grau III, R.H. 2004. Evaluating
ventilating air movement in underground limestone mines
by monitoring respirable dust generated from production
shots. In R. Ganguli, S. Bandopadhyay (eds), Mine venti-
lation: Proceedings of the 10th U.S./North American Mine
Ventilation Symposium Anchorage, AK, May 16–19, 2004:
221–232. Leiden, Netherlands: Balkema.
Grau, III, R.H., Mucho, T.P., Robertson, S.B., Smith, A.C. &
Garcia, F. 2002. Practical techniques to improve the air
quality in underground stone mines. In E. De Souza
(ed.), North American/Ninth U.S. Mine Ventilation, Proc.
intern. symp. Kingston, Ontario, Canada, June 8–12,
2002. Netherlands: Balkema.
Grau, III, R.H., Robertson, S.B., Mucho, T.P., Garcia, F. &
Smith, A.C. 2004a. NIOSH ventilation research address-
ing diesel emissions and other air quality issues in non-
metal mines. Transactions 2004 316: 149–158. Littleton,
CO: Society for Mining, Metallurgy, and Exploration, Inc.
Grau, III, R.H., Robertson, S.B., Krog, R.B., Chekan, G.J. &
Mucho, T.P. 2004. Raising the bar of ventilation for large-
opening stone mines. In R. Ganguli & S. Bandopadhyay
(eds), Mine ventilation: Proceedings of the 10th U.S./North
American Mine Ventilation Symposium, Anchorage, AK,
May 16–19, 2004: 349–355. Netherlands: Balkema.
Haney, R. & Saseen, G. 1998. Estimation of diesel particulate
concentrations in underground mines. Preprint 98–146,
Society of Mining, Metallurgy and Exploration Annual
Meeting March 9–11, Orlando, FL.
Head, H.J. 2001a. Proper Ventilation for Underground Stone
Mines. Aggregates Manager, January 2001, 60: 20–22.
Head, H.J. 2001b. Calculating Underground Mine Ventilation
Fan Requirements. Aggregates Manager, April 2001, 63:
17–19.
Krog, R.B., Grau III, R.H., Mucho, T.P. & Robertson, S.B.
2004. Ventilation planning layouts for large-opening
mines. SME preprint 04–187: 1–9. Littleton, CO: Society
for Mining, Metallurgy, and Exploration, Inc.
Mucho,T.P., Grau, III, R.H. & Robertson, S.B. 2001. Practical
mine ventilation. Presentation at the Safety Seminar for
Underground Stone Mines, Louisville, KY, Dec 5.
 12. Leakage and gob flows

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Ventilation surveying and modeling of longwall bleeder and gob areas

B.S. Prosser & N.L. Oswald

Mine Ventilation Services, Inc.

ABSTRACT: Longwall bleeder and gob areas have long been one of the most challenging and uncertain areas
in a ventilation model. There is a tendency to “black box” the bleeder and gob areas in a model, however this
can lead to problems when examining the addition of contaminates in the air stream as they pass though these
undefined areas. Additionally, the interaction between the bleeders and the gob is important to understand when
attempting to control gas concentration at bleeder evaluation and mixing points. This paper discusses several
experiences and common problems encountered with developing an accurate model to represent bleeder and
gob areas. Common techniques used to quantify the bleeder systems and the shortcomings associated with these
techniques are also discussed.

1 INTRODUCTION There are two approaches to ventilating gob areas.

In establishing a ventilation network model of a long-
wall coal mine, it is natural to place a high emphasis on
the modeling of the measurable main airways. These
areas include main intake and returns, belt lines and the
headgate, longwall face, and tailgate. Areas difficult to
quantify, such as the gob and bleeder systems, are often
modeled simplistically. However, these areas of the
mine can be a critical component of the overall venti-
lation system.An inaccurately modeled bleeder system
can result in under or overestimating the bleeder
efficiency. This in turn may result in an unreliable
base model.
During a typical coal mine ventilation survey only
a few hours are devoted to gathering measurements
in bleeder areas. This is due to the natural tendency to
place more focus on the higher profile out by areas that
have a greater impact on the accuracy of the resulting
model. However, it is the bleeder and gob ventila-
tion that may directly control the effectiveness of the
ventilation system.
To insure contaminant levels fall below legal limits,
the airflow through the bleeders are monitored for gas
content. When these limits are reached, more air must
circulate through these areas to dilute the contaminates
in order to achieve acceptable levels. If a ventilation
model accurately portrays the bleeder and gob areas,
then it may be used with confidence to determine
the proper ventilation settings required throughout the
mine. Therefore, accurate measurements need to be
taken in these areas to adequately represent them in a
network model.
The first approach is to create a high differential pres-
sure across the gob region in order to induce airflow
through the gob. In doing this, the gas can be drawn
out of the gob and removed from the system though
the bleeder. The second approach is to maintain a low
differential pressure across the gob region in efforts to
draw little or no air though the gob and into the bleeder.
Each approach has different pros and cons as well as
different surveying and modeling techniques.
2 BLEEDER SURVEYS
To model a bleeder system the frictional pressure drop
along the airways and the airflow quantities need to
be determined through out the bleeder system for both
ventilation approaches, Figure 1 shows a simplified
schematic of a high differential pressure bleeder sys-
tem. This type of system actively draws air up the
headgate and through the semi caved tailgate (three
entry system), abandoned gate road entries and the
gob. A traditional gauge and tube type ventilation sur-
vey can be used to determine the differential pressure
around the bleeder areas provided access is not limited
by ground conditions. Due to the high quantities of air
entering the bleeder at points “C”, “D” and “E” it may
be sufficient to only take measurements in the actual
bleeder airways. The difference in the airflow measure-
ments taken before and after these three points should
be suffcient to determine the quantity of air leaking
through the gob.

587

Copyright © 2006 Taylor & Francis Group plc, London, UK


Figure 1. High differential pressure bleeder.
Figure 2. Low differential pressure bleeder.
When the airflow entering through the gob is low
it is very difficult to differentiate between an increase
in airflow across an injection point and the instrument
error encountered with a traditional anemometer. To
minimize this error, direct measurements on gob reg-
ulators may be required. Care should be taken when
taking measurements around gob regulators due to the
poor ground conditions and higher gas contents typ-
ically encountered in these areas. The quality of the
air (e.g. CH4 , O2 , CO, CO2 , etc.) should be contin-
uously monitored because of the possibility of rapid
fluctuations and the measurer’s proximity to the gob
airstream.
Figure 2 illustrates a low differential pressure
bleeder system. Rather than actively drawing air
through the gob, this low pressure bleeder system mon-
itors the gob and the release of gas from the gob area at
position F. With this type of arrangement the airflow
through the bleeder system is substantially reduced.
Therefore, the gauge and tube method for the measur-
ing differential pressure differentials around the gob
area will be difficult to use. A barometer survey can
be used to determine the pressure differentials between
588
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 3. Gauge and tube technique.
the longwall face, mains, and key bleeder locations.
However, because of the very low flows in the bleeder,
measuring differential pressures will be very difficult
to accomplish whether using barometers or the gauge
and tube. Airflow measurements must be taken at each
point where gob air is introduced into the system
(points C and D).
3 GAUGE AND TUBE MEASUREMENTS IN
BLEEDER AIRWAYS
Generally, if there is enough air flow, the gauge and
tube technique can be used to measure frictional
pressure drops though out the bleeder system. This
technique allows the direct measurement of frictional
pressure differentials using a digital manometer or
a magnehelic gauge connected to a length of 6 mm
(1/4 in) nylon tube. The ends of the tube are connected
to the total pressure port of pitot static tubes. Lengths
of up to 300 m (1,000 ft) can be used, depending on
the application. The nylon tube, which is strong, semi-
rigid, and relatively inflexible, is strung along a drift
from one marked station to another marked station.
Both ends of the tube are connected to their respec-
tive pitot tube and are positioned near the middle of
the drift and facing into the airstream. A manometer is
placed in-line with the gauge and tube set-up. Figure 3
shows how this set-up is achieved with two people.
The measurement technique is independent of
minor changes in elevation, psychrometric parameters,
and air velocity measurements. No additional equip-
ment, other than the pitot tubes, nylon tube and the
manometer is necessary to carry out these measure-
ments. The simplicity of the measurement allows for
rapid reduction of data, field accuracy and verification
checks, and is unhindered by the need for additional
parameters to be measured by the survey team.
This method can be used to measure very low pres-
sure differentials, resulting from low airflows. How-
ever, when several measurements are to be combined,
the possible combination of small errors associated
with each measurement can pose a problem. Bleeder
systems can require 10,000 to 15,000 ft of tube mea-
surements in some cases and with a large number
of measurements the instrument associated errors can
inadvertently add up. This could cause a falsely high
or low calculated resistance in the bleeder.
 Using this method of measuring the frictional dif-
ferential pressure drop in the bleeder drifts omits
the need for other measurement techniques, such as
tracer gases. Due to the constant fluctuations in the
airflow though the gob, the higher accuracy measure-
ments obtained with these more expensive techniques
will not contribute much to the accuracy of a model.
Therefore the gauge and tube method is preferred.

4 MODELING OF A BLEEDER SYSTEM

If there is sufficient airflow to obtain good pressure dif-

ferential measurements in the bleeder system, then the
system should be modeled in a conventional manner
Figure 4. Developed bleeder system model.
as with the rest of the mine. Since it is not possible to
obtain measurements in the gob areas, the pressure and
airflow in these areas must be determined using indi-
rect methods.The input values for these branches in the
model must be determined by closing pressure loops
through the known bleeder system, around the gob to
the intake air for the gob areas and then back to the
bleeder. Several things can be inferred by examining
the differential pressures across the gob regulators and
the separation stoppings along the bleeder. These pres-
sures can indicate how tightly the gob is caving or if the
central tailgate entry is open for airflow. Additionally,
gas concentrations and air temperature can be used
to infer leakage routes though the gob areas. Figure 4
shows a developed bleeder system model with the leak-
age branches through the gob. The resistance of these
leakage branches is determined by difference, clos-
ing pressure loops through the longwall tailgate. The
resistance of the branch representing the old gateroad Figure 5. Determination of leakage path pressure differ-
at note D cannot be determined in this model, however, ential.
the resistance of all leakage branches radiating from
the longwall tailgate can be determined. The resistance inferred that the air at point K is coming off of the
values of these branches will change as the position of longwall, as shown above.
the longwall progresses toward the mains, and as the Both a pressure and an airflow balance of this area
gob further settles. can be accomplished using this simple methodology.
Figure 5 shows a simplified bleeder arrangement. This is done to add confidence to all the measurements
The letters represent the pressure measurements for taken. If the pressures or airflows do not balance prop-
each of the branches that will be representing the erly, then the faulty measurement can be identified and
bleeder system in the model. The branches represented retaken as necessary. Also, by balancing the measured
by “C” and “L” are not measured directly and must be airflows, approximate airflows routes though the gob
determined by difference using Kirchoff’s Laws. can be determined. Perimeter airways and former gate
From Kirchoff ’s second law it is known that the roads must be considered as likely airflow routes when
summation of pressures around a closed loop must determining leakage paths.
equal zero. For realistic modeling purposes, a 10%
error is permissible. To determine the differential pres-
sure in branch C a summation around branches G, 5 SUMMARY
H, I, J and B would be calculated. The pressures are
added in the direction of the airflow, or from low Bleeder systems and gob areas can be adequately
to high, therefore C = G + H + I − J − B. Likewise, surveyed and modeled using traditional techniques.
to solve for branch L the following would be true, Problems arise when low airflows are encountered in
L = G + H + I + M − B. If the stopping shown at point low differential pressure bleeder systems. However,
F is blocking leakage into the gob, then it can be due to the fluctuations in airflow though the gob areas,

589

Copyright © 2006 Taylor & Francis Group plc, London, UK

 a traditional and simple method of measuring frictional
differential pressures in the bleeder drifts is preferred.
Because the ventilation system is directly affected
by the bleeder and gob areas, these areas should be a
primary focus during a ventilation survey. They should
also be accurately represented in the ventilation model.
Although a tracer gas survey can be conducted to
determine the absolute origin of air flowing through a
bleeder/gob system, this level of detail may not ulti-
mately be necessary for the development of a viable
ventilation model. The dynamic nature of the airflow
through the gob areas combined with the continuous
movement of the longwall make accurate measure-
ments difficult.
590
Copyright © 2006 Taylor & Francis Group plc, London, UK
REFERENCES
Duckworth, I.J. & Prosser, B.S. 1997. An Analysis of the
Data Obtained from Ventilation Studies of Longwall Pan-
els. Proceedings of the 6th International Mine Ventilation
Congress, Pittsburgh, Pennsylvania.
McPherson, M.J. 1993. Subsurface Ventilation and Environ-
mental Engineering. New York, New York: Chapman &
Hall.
Mucho, T.P., Diamond, W.P., Garcia, F., Byars, J.D. & Cario,
S.L. 2000. Implications of Recent NIOSH Tracer Gas
Studies on Bleeder and Gob Gas Ventilation Design. SME
preprint 00-8. Littleton, Colorado.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Computational fluid dynamics study on the ventilation flow paths in

longwall gobs

L. Yuan, A.C. Smith & J.F. Brune

National Institute for Occupational Safety and Health, Pittsburgh Research Laboratory, Pittsburgh,
Pennsylvania, USA

ABSTRACT: To provide insights and assistances for the optimization of ventilation systems for U.S. under-
ground coal mines facing both methane control and spontaneous combustion issues, a computational fluid
dynamics (CFD) study is being conducted to investigate the effect of the ventilation scheme on the prevention
of spontaneous combustion in longwall gob areas. This report focuses on the flow patterns within the gob under
three different ventilation systems; one-entry and two-entry bleederless systems as well as a three-entry bleeder
system. The gas flow in the caved gob area is simulated as a laminar flow through porous media while the gas
flow in ventilation airways is simulated as a fully developed turbulent flow. The air flow patterns are visualized
using flow path lines. Air velocity contours and vector data are also obtained. The possible location of critical
velocity zones where the gob is most liable to spontaneous heating is discussed.

DISCLAIMER: The findings and conclusions in this report are those of the authors and do not necessarily
represent the views of the National Institute for Occupational Safety and Health.

1 INTRODUCTION regarding detection, fairly precise locating, and fire

Spontaneous combustion continues to be a significant
risk for underground coal mines, particularly in U.S.
western mines where the coal is generally of lower
rank. This hazard is exacerbated in mines with appre-
ciable levels of methane, due to the potential of an
explosion ignited by a spontaneous combustion fire.
The most recent reported data on mine fires in the
U.S. for the period 1990–1999 show that there were
15 reported fires caused by spontaneous combustion in
underground coal mines, accounting for 17% of the 87
total reported fires (DeRosa 2004). Although the num-
ber of fires is relatively low, amounting to an average of
1 to 2 fires per year, the potential for catastrophic con-
sequences is very high. In fact, three of the mine fires
from the reported 1990–1999 period resulted in sub-
sequent methane explosions. In addition, although the
number of spontaneous combustion fires has remained
nearly constant for the last 35 years, there is the possi-
bility that the number of spontaneous combustion fires
increases due to growth in the dimensions of longwall
panels and due to the exhausting of easily-mined coal
beds in the U.S., which results in increased mining of
lower rank coals and deeper coal beds.
Most spontaneous combustion fires have occurred
in the gob areas of U.S. coal mines. Gob areas,
by the very nature, present accessibility problems
extinguishment. The self-heating of coal occurs when
the heat produced by low-temperature oxidation is not
adequately dissipated, resulting in a net temperature
increase in the coal mass. Under conditions that favor
a high heating rate, a fire ensues. The oxidation of
coal requires a fuel, the coal, and an oxidizer, usually
the ventilation air. To reduce the amount of oxidation,
and thus the amount of heat being produced, either the
coal or the air must be removed. On the other hand,
if the airflow is sufficiently high, it can remove the
heat produced by the coal oxidation, and thus prevent
acceleration of coal self-heating that may result in a
fire. The amount of coal in a retreat longwall gob can
be somewhat controlled by the mining practices, such
as pillar design and the amount of coal left in the roof
and floor. Additional coal from un-mined coal bands
immediately above or below the mined seam may be
present in the gob as well.
The amount of airflow through a longwall gob can
be reduced by using a bleederless ventilation design.
In the U.S., bleederless systems are only permitted as a
spontaneous combustion control method in mines with
a demonstrated history of spontaneous combustion.
However, in U.S. mines with appreciable methane lev-
els, ventilating gob areas with bleeder systems to dilute
and remove methane has been traditionally regarded as
the preferred method to deal with the methane hazard.

591

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Bleeder system design can exacerbate the spontaneous
combustion hazard in mines with both methane and
spontaneous combustion issues.
In order to optimize the airflow in the gob to pre-
vent spontaneous combustion and control the methane
content, it is important to understand likely air flow
patterns inside the gob and their effect on both coal
self-heating and methane migration. Because much of
the gob area is inaccessible, it is difficult to measure
the air flow rate in it. Although some field measure-
ment methods such as tracer gas technology have been
used to determine the air flow pattern, the results are
limited. Therefore, flows through the caved area of
the gob and even to some extent along the periphery
have been mainly conjecture based on spotty experi-
ences or investigations. Absent practical methods to
gather empirical data, controlled modeling techniques
are viewed as the only current reasonable way to assess
the ventilation in gob areas.
Computational Fluid Dynamics (CFD) modeling
has been increasingly applied to mine ventilation in
recent years. CFD is the application of numerical
techniques to solve the Navier-Stokes equations for
fluid flow. The Navier-Stokes equations are derived Figure 1. Ventilation schemes: (a) one-entry bleederless
by applying the principles of conservation of mass and system; (b) two-entry bleederless system; (c) three-entry
momentum to a control volume of fluid and are solved bleeder system.
by discretizing the equations using either finite differ-
ence or finite element techniques. In Australia, Balusu
mine ventilation systems. Among these challenges
et al. (2002) and Wendt & Balusu (2002) conducted
is the characterization of flow patterns through and
CFD modeling of longwall gob air flow dynamics with
around the gob, and the effect of these airflows on the
focus on bleederless ventilation systems. In the U.K.,
spontaneous combustion risk.
Ren et al. (1997) performed CFD modeling of methane
Three types of ventilation schemes were investi-
flow around longwall coal faces. All these studies were
gated to demonstrate the air flow patterns in the gob.
based on specific ventilation schemes, and results were
The simulated gob area in the three studies is 1000 m
limited to those ventilation systems. In the U.S., Wala
long, 300 m wide and 50 m high starting from the bot-
et al. (2003) have used CFD to evaluate aspects of
tom of the coal seam, representing a typical longwall
mine face ventilation, but no CFD work has been done
panel layout in the U.S. The caved region may have a
to model flows in the longwall mine gob area. Brune
height of 3∼4 times mining height, but a highly frac-
et al. (2000) identified the significant changes for U.S.
tured region may extend beyond that height. The 50 m
longwall mine ventilation and studied the effects of
height was chosen to cover both the caving and highly
different bleeder configurations and ventilation adjust-
fractured regions. Although actual longwall panel
ments. Longwall gob leakage was also simulated using
lengths extend to 3,000 or 4,000 m, a 1,000 m sec-
a ventilation network analysis program, VNET, by
tion was selected to model the flow paths in this study.
Brunner (1985) and Banik et al. (1995).
The ventilation airway dimensions are 2 m high by 6 m
NIOSH has initiated a new program to develop
wide, and the longwall face is 4 m wide by 2 m high.
CFD models to evaluate ventilation schemes to control
The first ventilation scheme is a simple “U” bleed-
spontaneous combustion and to describe the venti-
erless ventilation system with one entry, as shown
lation pathways through the immediate gob and to
in Figure 1 (a). The second scheme is a two-entry
evaluate their effect on gob methane control systems.
bleederless system as shown in Figure 1 (b). For the
In this paper, a preliminary CFD study of air flow
convenience of simulation, two intake entries were
patterns in the gob is presented.
combined into one entry, and only one crosscut close to
the face was open to the second return entry. The third
ventilation system, shown in Figure 1 (c), is a three-
2 VENTILATION SCHEMES entry bleeder system. The three collapsed entries at
the tailgate side were combined into the gob area and
The increase of longwall panel lengths and widths in were connected to the bleeder fan through an open
the U.S. continues to present challenges to longwall entry. It should be pointed out that this treatment may

592

Copyright © 2006 Taylor & Francis Group plc, London, UK

 be only valid near the end of the panel life. During
the earlier stages of the panel life, the middle entry
remains at least partially open and has a different air-
flow resistance than the collapsed area. It then serves
as a primary route for bleeder air. With the retreat of the
longwall face, the middle entry continues to collapse,
and will have a same order of magnitude resistance as
other collapsed areas approaching the end of the panel
life. On the headgate side, the belt entry was com-
bined with the intake entry and the flow rate in both
entries was taken into account. All crosscuts between
the second and third entries on the headgate side were
open. Although intact stoppings may be maintained
between the second and third entry for some long-
walls for the purpose of methane dilution, it will not
affect this model significantly. The bleeder entries at
the back end of the gob were also combined into one
model entry connecting to the bleeder fan.
3 ESTIMATION OF GOB PERMEABILITY
It is generally believed that the permeability inside a
gob varies in different areas. The global permeabil-
ity distribution within the gob of a single longwall
panel was represented by five constant-permeability
zones, as shown in Figure 2. The division of the gob
area and the assignment of permeability data were
based on geotechnical modeling of longwall mining
and the associated stress-strain changes/rock failure
using FLAC (Fast Lagrangian Analysis of Continua)
code (Esterhuizen & Karacan 2005). In FLAC model-
ing, mining was simulated in increments, starting from
one side of the grid and advancing to the other side.
Extraction of the coalbed was modeled by removing
elements over the height of the coalbed. The process
of gob formation was modeled by first deleting rock
elements in the roof of the coalbed, so that they are
stress relieved, followed by inserting gob properties
in these elements. Gob properties were also inserted
in previously mined coalbed elements, so that the gob
filled the mined void.
In the gob caved area, the re-compaction of the
caved rock has a significant effect on its permeability.
Stress changes in the rock in the gob cause changes
in the fracture apertures which then impact the per-
meability. Immediately behind the advancing face,
the caved rock is loosely stacked and has a porosity
of approximately 0.4. As the mining face advances
away from this caved rock, the weight of the over-
burden gradually increases, which re-compacts the
caved rock. This loading or stress in the caved material
increases exponentially, until the full overburden load
is supported by the caved material. The re-compaction,
estimated with the FLAC model’s assigned strength
and mechanical properties of the gob material, results
in a reduction in the void space within the broken rock
593
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 2. Gob zones for a single longwall panel.
rubble and, therefore, significant changes in perme-
ability. A simple relationship was used to estimate the
changes in permeability in the caved rock based on the
Kozeny-Carmen equation:
where n is the porosity and k is the permeability.
Using the results of the FLAC model, the per-
meabilities for the five zones were determined to
be 1 × 106 millidarcies (md) for zone 1, 2 × 105 md
for zone 2, 7 × 104 md for zone 3, 1 × 104 md for
zone 4, and 5 × 103 md for zone 5. The longwall face
shields were simulated as a thin porous media with
a permeability of 9 × 107 md. Note that the higher
permeability in zone 1 provides for lower resistance
pathways along the periphery of the gob. It should
be pointed out that the permeability data used in this
study are different from those used by Brunner (1985),
1 × 10−7 to 1 × 10−5 m2 (1 × 108 to 1 × 1010 md),
Ren et al. (1997), 1 × 10−15 to 1 × 10−10 m2 (1 to
1 × 105 md), and Wendt et al. (2002), maximum
1 × 10−9 m2 (1 × 106 md). These differences may be
caused by different coal seam geology, different mine
panel layout, and different time periods in which the
permeability was estimated.
4 NUMERICAL MODELING
A commercial CFD software program, FLUENT™
(Reference to a specific product is for informational
purposes and does not imply endorsement by NIOSH),
from Fluent, Inc., was used to simulate the air flow in
the gob. FLUENT is a general purpose CFD solver
for a broad spectrum of flow modeling applications.
It has the capability to simulate laminar and turbulent
porous media flow based on the known permeability
and the inertial resistance factor data. The air flow in
the longwall mine gob area is treated as laminar flow in
a porous media. The real flow inside a gob can be very
complicated, and this laminar flow assumption may not
be valid for every region inside a gob. Further research
will be carried out that may modify this assumption in
the next phase of this program. The air flow in the
 Figure 3. The mesh for the three-entry bleeder system.
ventilation airways was simulated as fully developed
turbulent flow using standard k-ε equations with k the
turbulent kinetic energy, ε the turbulence dissipation
rate. The k-ε equations are derived from the eddy vis-
cosity theory of turbulent flow, and have been used
successfully to model turbulent flow in many different
situations.
The physical model and mesh for the CFD simula-
tion were created using the mesh generation software,
GAMBIT, from Fluent, Inc. The cell size varies from 1
to 4 m in the model. They were selected to resolve the
flow patterns both in the gob and inside typical mine
airways, as well as to achieve an acceptable conver-
gence within a reasonable amount of time. This mesh
is depicted in Figure 3. The total cell number for the
three-entry bleeder system was 644,660, most of which
were hexahedral cells as shown in the figure.
For the simulations with FLUENT, boundary con-
ditions were chosen to represent pressures and flow
quantities found in typical longwall ventilation situa-
tions for the geometry modeled. For one-entry and the
two-entry ventilation systems, the boundary condition
used was 0.12 kPa (0.5 inches water gauge) pressure
differential between the intake and return entry along
the longwall face. To control the air flow quantity to the
longwall face, the wall roughness was adjusted to have
a realistic intake airflow rate of 30 m3 /s (64,000 cfm).
For the three-entry bleeder ventilation system, two
regulators were placed at the end of the second and
third intake entry, respectively, as shown in Fig-
ure 1 (c), and were simulated as surfaces that can have
a pre-defined pressure change. No regulators were
placed at the tailgate side of the gob, and the flow
was controlled by the permeability value assigned.
In actual bleeder ventilation systems, there are usu-
ally regulators in the entries at the tailgate side. In
the model this simplification was justified since the
boundary conditions effectively replace these regula-
tors. The boundary conditions used were −0.62 kPa
594
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 4. Flow path lines colored by velocity magnitude
(m/s) for the one-entry system: (a) in the whole gob; (b) near
the shields.
(−2.5 inches water gauge) pressure at the intake air-
way inlet on the headgate side of the face, −0.75 kPa
(−3 inches water gauge) pressure at the return entry
outlet on the tailgate side of the face, and −3.7 kPa
(−15 inches water gauge) pressure at the bottom of
the bleeder fan shaft. The wall roughness was adjusted
to have the total airflow rate of 43 m3 /s (91,000 cfm)
for all three intake entries. The pressure drops across
the two headgate regulators were also adjusted so that
the airflow quantity along the longwall face was the
same as longwall face quantities used with the one-
entry and two-entry ventilation systems, i.e. 30 m3 /s
(64,000 cfm).
5 FLOW PATTERNS INSIDE THE GOB
The air flow inside a gob is expected to be three dimen-
sional with the flow in the vertical direction weaker
than in the other two directions. In order to visualize the
flow patterns inside the gob, a virtual horizontal refer-
ence surface was created 1 m (3 ft) from the bottom of
the coal seam floor. All results reported hereafter are
with respect to this horizontal reference surface.
5.1 One-entry bleederless system
Figure 4 shows the flow path lines for the one-entry
bleederless system colored by velocity magnitude.
 Figure 5. Flow path lines colored by velocity magnitude
(m/s) for the two-entry system: (a) in the whole gob; (b) near
the shields.
Figure 4 (a) shows the path lines in the entire gob
area, and Figure 4 (b) shows the path lines near the
face. The path lines show that flow through the gob
itself was mainly concentrated behind the shields. At
the headgate side, air leaked through the shields but
some was forced back into the face again through
the shields near the tailgate side. The air velocity
ranged between 1.0 × 10−5 to 1.0 × 10−4 m/s (0.002 to
0.02 fpm) near the shields, and below 1.0 × 10−5 m/s
(0.002 fpm) farther away from the shields.
5.2 Two-entry bleederless system
Results of the modeled flow paths for the two-entry
bleederless ventilation system were similar to the one-
entry system, except for slightly higher air velocities
on the tailgate side created by the second return cross-
cut inby the face shown in Figure 5. For both one-entry
and two-entry bleederless ventilation systems, the air
flow velocity at the back end of the gob was very
low. Figure 6 compares velocity contours close to the
shields and near the tailgate between the one-entry and
two-entry ventilation systems. The velocity contours
were only slightly different between the two systems,
indicating that the second return entry has little effect
on the velocity field close to the shields. Inside the gob
and near the second return entry inside the gob, the air
velocity was about 1.0 × 10−4 to 7.0 × 10−4 m/s (0.02
to 0.14 fpm). The air velocity inside the second return
595
Copyright © 2006 Taylor & Francis Group plc, London, UK
Figure 6. Comparison of colored velocity contours (m/s)
between the one-entry and two-entry ventilation system: (a)
one-entry system; (b) two-entry system.
entry was about 7.0 × 10−4 to 1.0 × 10−3 m/s (0.14 to
0.2 fpm).
5.3 Three-entry bleeder system
Figure 7 (a) shows the flow path lines in the gob with
the three-entry bleeder ventilation system. A bleeder
fan is located at the back of the panel. Gob air flow was
mainly concentrated in three areas; behind the shields,
near the back end of the gob, and along the tailgate
entry, where the permeability is highest compared to
other caved zones in the gob. The air velocity along the
least compacted zone (zone 1) at the tailgate side was
larger than that behind the shields, but smaller than that
near the back end of the gob. Figure 7 (b) shows the
flow path lines near the shields. Air that leaked through
the shields at the headgate side does not flow back into
the face at the tailgate side in this ventilation scheme,
but flows towards the tailgate side and then out of the
gob through the least compacted zone. Some air from
the second and third intake entries enters into the gob
through the crosscuts near the back end of the gob. The
air velocity behind the shields was between 1.0 × 10−3
to 7.0 × 10−3 m/s (0.2 to 1.4 fpm), much higher than
was seen for the one-entry and two-entry ventilation
systems. Figure 8 (a) shows the flow path lines near

Figure 7. Flow path lines colored by velocity magnitude
(m/s) for the three-entry bleeder system: (a) in the whole
gob; (b) near the shields.
Figure 8. Flow path lines colored by velocity magnitude
(m/s) for the three-entry bleeder system: (a) near the back
end of the gob; (b) away from the back of the gob.
596
Copyright © 2006 Taylor & Francis Group plc, London, UK
the back end of the gob. The air velocity was about
1.0 × 10−2 to 3.0 × 10−2 m/s (2 to 6 fpm), even higher
than that behind the shields. Figure 8 (b) shows the flow
path lines near the center of the gob. These path lines
were directed towards the bleeder fan, but with very
low air velocity, about 1.0 × 10−6 to 7.0 × 10−6 m/s
(0.0002 to 0.0014 fpm).
6 CONDITIONS FOR SPONTANEOUS
COMBUSTION
In coalbeds that are reactive to oxidation, critical
velocity airflow over reactive coal increases the risk
of coal heatings. Critical airflow is defined as insuf-
ficient airflow to remove the heat due to oxidation,
but sufficient airflow to maintain the oxidation pro-
cess (Smith et al. 1994). In general, just behind the
shields and along the edges of the gob (in the bleeder
case), the air velocity is high, so that any heat gen-
erated by self-heating is carried away. Deeper in the
gob, the air velocity is too low to provide sufficient
oxygen to support spontaneous combustion. However,
there possibly exists a critical air velocity zone in the
gob where the airflow and oxygen concentration are
suitable for the continual growth of a self-heating to
an eventual fire. The exact range of the critical veloc-
ity and its location depend on several factors such as
the coal’s particle/fragment size, physical and chemi-
cal properties, local pressure, temperature, and water
vapor concentration. To demonstrate the effect of the
ventilation scheme on the critical velocity zone, an
estimation for the minimum velocity to carry away
heat generated from the self-heating of a 150 µm coal
particle was made using the experimental data from
Smith & Lazzara (1987).
The self-heating rate of coal under adiabatic condi-
tion can be expressed as:
where dT/dt is the self-heating rate in K/s, A the pre-
exponential factor in K/s, E the activation energy in
kcal/mol, and T the temperature in K. Smith & Lazzara
(1987) conducted a series of spontaneous combustion
studies and determined the activation energy and the
pre-exponential factor for a number of U.S. coals. The
heat generation rate from the self-heating is:
where Q is the heat generation rate in W, m the mass of
the particle in g, Cp the specific heat of coal in J/(g · K)
at constant pressure. In order to prevent the occurrence
of self-heating, the generated heat needs to be carried
away by the airflow. The convective heat transfer rate
can be expressed as:
 where S is the particle surface area for convective
heat transfer in m2 , h the heat transfer coefficient in
W/(m2 · K), T the temperature differential between
the particle and the airflow in K. The heat transfer
coefficient h is related to a non-dimensional Nusselt
number Nu as:

where l is the diameter of the particle in m, λ the heat

conductivity of air in W/(m · K). For the laminar flow
over a sphere, the Nu number correlates with the Re
number as (Kays 1966):

here Re = ul/v with u the airflow velocity in m/s, ν the

kinematic viscosity of air in m2 /s.
Using an example of the Wyoming No. 80 seam
coal particle with l = 0.15 mm, the maximum parti-
cle diameter used by Smith & Lazzara (1987), A =
6.9 × 106 K/s, E = 16.8 kcal/mol, T = 60◦ C (333 K)
and) T = 40 K for the No. 80 seam coal
particle, coal specific heat Cp = 1.17 J/(g · K), air
heat conductivity λ = 2.59 × 10−4 J/(cm · s · K), and
ν = 15 × 10−6 m2 /s, the upper bound for the critical
airflow velocity calculated using equations (2) to (6)
was 1.1 × 10−4 m/s (0.02 fpm). The minimum airflow
velocity (lower bound for critical velocity) required to
provide sufficient oxygen to sustain the self-heating
will be determined by incorporating the model of
kinetic reaction of coal self-heating into the CFD
modeling in the next stage of this program.
Figure 9 shows the velocity contour of 1.1 ×
10−4 m/s for the three ventilation systems. The criti-
cal velocity zone for the particle size of 150 µm should
start from this contour towards the lower velocity area.
Thus, in this example for the No. 80 seam coal using a
one-entry and two-entry bleederless ventilation sys-
tem, the critical velocity zone started immediately
Figure 9. Velocity contour of 1.1 × 10−4 m/s for the three
behind the shields towards the center of the gob. Closer ventilation systems: (a) one-entry bleederless system; (b)
to the headgate and the tailgate, the zone started a short two-entry bleederless system; (c) three-entry bleeder system.
distance behind the shields. For the three-entry ventila-
tion system, the critical velocity zone was pushed away
both from the shields and from the back end of the gob bleederless ventilation systems, the flow patterns in the
towards the center. Under these simulation conditions, gob were very similar. Only a slightly higher velocity
the velocity contour line was close to the boundary zone occurred near the second return entry with the
between zone 1 and zone 2 as shown in Figure 2. two-entry ventilation system. This demonstrates that
sudden changes in airflow resistance can potentially
constitute zones with critical velocities that support
7 CONCLUSIONS spontaneous combustion.
The air velocity was about 1.0 × 10−5 to 1.0 ×
CFD simulations were conducted to investigate the 10−4 m/s (0.002 to 0.02 fpm) near the shields, and
air pathways in the gob area under different ven- 1.0 × 10−4 to 7.0 × 10−4 m/s (0.02 to 0.14 fpm) near
tilation schemes. The air flow patterns in the gob the second intake entry. The velocity near the back end
were visualized by flow path lines, velocity contours of the gob was very low. For the three-entry bleeder
and velocity vectors. For the one-entry and two-entry system, the air velocity was much higher behind the

597

Copyright © 2006 Taylor & Francis Group plc, London, UK

 shields than with the bleederless systems, about 1.0 ×
10−3 to 7.0 × 10−3 m/s (0.2 to 1.4 fpm). There was also
high velocity air flow near the back end of the gob,
about 1.0 × 10−2 to 3.0 × 10−2 m/s (2 to 6 fpm).
A preliminary calculation for a coal particle size of
150 µm in diameter and similar reactivity to the coal
used in this study shows the potential critical veloc-
ity zones may occur immediately behind the shields
towards the center of the gob for the one-entry and two-
entry ventilation systems. For the three-entry bleeder
ventilation system, the critical velocity zone may occur
farther away from both the shields and the back end of
the gob where a permeability gradient exists.
A gob will have coal particles of variable sizes. For
any single particle size, there are multiple solutions
for particle temperature T and temperature difference
between particle and air. Therefore, it is difficult to
develop a unique solution. However, the goal is to
outline areas with spontaneous combustion potential.
Future work will be done to further investigate the flow
patterns and to simulate the effect of methane con-
trol measures on the air flow patterns inside the gob
in mines with appreciable methane levels. It is also
intended to verify the results of the CFD model by
using tracer gas and quantity/pressure surveys where
possible.
ACKNOWLEDGMENTS
The authors wish to acknowledge the technical help
from Drs. Gabriel S. Esterhuizen and Özgen C.
Karacan with the estimation of permeability in the gob.
REFERENCES
Balusu, R., Deguchi, G., Holland, R., Moreby, R., Xue, S.,
Wendt, M. & Mallett, C. 2002. Goaf gas flow mechanics
and development of gas and Sponcom control strategies
at a highly gassy mine. Coal and Safety 20: 35–45.
598
Copyright © 2006 Taylor & Francis Group plc, London, UK
Banik, J., McPherson, M.J. & Topuz, E. 1995. Ventilation
control of self-heating in retreating longwall coal mines.
Proceedings of the 7th US Mine Ventilation Symposium,
Lexington, 5–7 June.
Brune, J.F., Aman, J.P. & Kotch, M. 1999. Developments in
Longwall Ventilation. Proceedings of the 8th US Mine
Ventilation Symposium, Rolla, 14–17 June.
Brunner, D.J. 1985. Ventilation models for longwall gob
leakage simulation. Proceedings of the 2nd US Mine
Ventilation Symposium, Reno, 23–25 September.
DeRosa, M. 2004. Analysis of mine fires for all U.S. under-
ground and surface coal mining categories, 1990–1999.
NIOSH Information Circular 9470.
Esterhuizen, G. S. & Karacan, C.O. 2005. Development of
Numerical Models to Investigate Permeability Changes
and Gas Emission Around Longwall Mining Panels.
Proc. AlaskaRocks 2005, 40th US Symposium on Rock
Mechanics, Anchorage, 25–26 June.
Kays, W.M. 1966. Convective Heat and Mass Transfer.
McGraw-Hill.
Ren, T.X., Edwards, J.S. & Jozefowicz, R.R. 1997. CFD
modeling of methane flow around longwall coal faces.
Proceedings of the 6th International Mine Ventilation
Congress, Pittsburgh, 17–22 May.
Smith, A.C., Diamond, W.P., Mucho, T.P. & Organiscak, J.A.
1994. Bleederless Ventilation Systems as a Spontaneous
Combustion Control Measure in U.S. Coal Mines. US
Bureau of Mines Information Circular 9377.
Smith, A.C. & Lazzara, C.P. 1987. Spontaneous Combus-
tion Studies of U.S. Coals. US Bureau of Mines Report of
Investigations 9079.
Wala, A., Jacob, J., Brown, J. & Huang, G. 2003. New
approaches to mine-face ventilation. Mining Engineering
March: 25–30.
Wendt, M. & Balusu, R. 2002. CFD modeling of longwall
goaf gas flow dynamics. Coal and Safety 20: 17–34.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Study of leakage flow in US underground coal mines

F. Calizaya & M. Stephens

Mining Engineering Dept., University of Utah, Salt Lake City, Utah, USA

ABSTRACT: This paper presents the results of ventilation surveys that were conducted in two US underground
coal mines to determine a parameter(s) to characterize leakage paths such as stoppings, overcasts, doors, and
regulators. It includes a summary of conditions under which these parameters can be used directly and when
they should be adjusted for external loads. The study also presents the results of computer exercises that were
performed to study the impact of changes in key structures on the overall ventilation efficiency.

1 INTRODUCTION of factors that must be considered in the estimation of

these resistances.
In underground mines, stoppings and doors are used
to separate two types of airways: intakes and returns.
Depending upon the complexity of the mine, the venti- 2 LEAKAGE FLOW FORMULAE
lation system may require hundreds of these structures.
The air pressure differentials created across these In the estimation of leakage quantity and leakage path
structures will inevitably cause some loss of air, i.e. resistances, the following equations are used:
some quantity of air will be short circuited without
being used in the working areas. This loss of fresh air is
referred to as leakage flow. Ventilation specialists have 2.1 Percent leakage flow (%L)
estimated that, in coal mines, this quantity can range This percentage represents a fraction of the total quan-
from 30 to 60 percent of the total quantity induced tity of air passed through the main fans that is short
by the surface fans (McPherson 1993, Richardson circuited before reaching the working areas. It is
1997). calculated by:
In ventilation design, stoppings are represented as
highly resistive air paths. When pressure-quantity sur-
vey data are available, the resistance of each path is
calculated from Atkinson’s equation (P = R Q2 ), and
sometimes this equation is modified to reflect the
flow type. Based on field measurements, researchers where QT = total quantity of air circulated through the
have found that the resistance for a ‘good’ sin- fans, m3 /s; QE = quantity of air utilized effectively at
gle masonry stopping may vary between 560, and the workings, m3 /s.
782,800 Ns2 /m8 (Bruce & Koenning 1987; Duckworth
1995.) It is the range of these values that makes venti-
lation design difficult and the leakage flow estimates 2.2 Equivalent leakage path resistance
questionable. For a number of leakage paths arranged in parallel, this
This paper presents the preliminary results of ven- resistance is approximated by Atkinson’s equation:
tilation surveys that were conducted in two US under-
ground coal mines. The main objective of these studies
was to determine the parameter(s) that may be used to
represent the flow of air through various types of stop-
pings, overcasts, and doors. It describes the conditions
under which these parameters can be used directly, and where Re = the equivalent resistance for a number of
when they should be adjusted to account for changes stoppings, Ns2 /m8 ; P = weighted average pressure
in the flow regime. The study also includes a summary difference, Pa; Qd = total leakage quantity, m3 /s.

599

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Face B
R Leakage Paths
4
Return
3
6 2
XC-3
XC-1
5 Key
1 1
Main Fan
Intake
R
Belt
Figure 1. Mine A representative ventilation schematic.
2.3 Individual leakage path resistance
Once an equivalent resistance has been estimated, an
individual stopping resistance can be calculated by
applying Kirchhoff ’s second law as follows:
where Ri = individual stopping resistance, Ns2 /m8 ;
n = number of stoppings in parallel arrangement.
3 VENTILATION SURVEYS IN TWO
UNDERGROUND COAL MINES
Air pressure – quantity surveys have been conducted in
two US underground coal mines: Mine A and Mine B.
At Mine A, the vent system consists of one intake
portal, a set of parallel entries (of about 1.2 km long
each), two headings, and one exhaust portal. The sys-
tem is powered by a 149 kW (200 HP) axial fan. This
fan is equipped with a variable frequency drive-motor
which allows the fan to change its speed between 500
and 900 rpm. When the fan is set at its highest speed,
it can exhaust up to 94 m3 /s of air at 1200 Pa of static
pressure.
At Mine B, the vent system consists of seven intake
portals, several mains and sub-mains, two active devel-
opment headings, one longwall panel and one return
portal. The system is equipped with over 1000 concrete
and steel stoppings, several airlock doors, overcasts
and regulators, and one 950 kW exhaust fan. When
operated at 890 rpm, this fan can exhaust from this
600
Copyright © 2006 Taylor & Francis Group plc, London, UK
Face A
XC-20
XC-30
Hot
Spot
7 8
Station No. Intake Air
Overcast Return Air
R Regulator Stopping
mine approximately 392 m3 /s of air at 1800 Pa of static
pressure.
In both mines, calibrated manometers, anemome-
ters and barometers were used to conduct the ventila-
tion surveys.
3.1 Description of Mine A
A schematic representing Mine A is shown in Figure 1.
The coal seam is horizontal with an average thickness
of 2.5 m. Although the mine has two headings, at the
time of the visit, only one was operated at a given
time. A five entry system is used to access and gain
the coal from each heading. The mine is ventilated by
a standard U-tube ventilation system in which the belt
line is isolated from other entries by stoppings made of
Omega 384 blocks. Dust is the major contaminant at
the mine, but in an isolated location, strata heat is also
a problem. Ventilation is the primary control method
utilized in this mine, but water is also used.
Ventilation surveys were conducted in May, 2005. In
this case, the main objective was to develop ventilation
strategies to increase the cooling power of the air. It
was hypothesized that the heat flow problem can be
controlled by increasing the capacity of the main fan
and reducing leakage quantity.
3.1.1 Survey data
Ventilation surveys in this mine have been conducted
for two conditions: (1) normal fan speed of 500 rpm,
at which the fan exhausted 60.46 m3 /s of air at 500 Pa
of static pressure, and (2) maximum fan speed of
 Table 1. Ventilation survey data for Mine A (air density (XC-1 through XC-3) blocked by Omega stoppings.
ρ = 0.91 kg/m3 ). These stoppings due to their proximity to the main fan
are subject to high pressure differentials. The South-
Item Input variables Condition 1 Condition 2 east Main section (between stations 7 and 8) includes
10 cross-cuts (XC-20 through XC-30) also blocked
1. Fan duty
by Omega stoppings. The stoppings, located at about
Quantity, m3 /s 60.65 94.00
Total head, Pa 498 784 900 m from the main fan, are subject to lesser pressure
Pressure across 324 560 than those mentioned previously.
explosion door, Pa The stoppings are not sealed completely; some
quantity of air is allowed to pass through them, in some
2. Airflow rates, m3 /s cases intentionally through doors and small windows
Intake 1 (portal) 49.13 68.79 used as pressure equalizers. Further, these walls are
Main intake, St 2 46.62
Main intake, St 3 14.93 23.22
built according to the needs as mining progresses and
Main intake, St 4 10.94 17.01 therefore are subject to deterioration over time.
SE main intake, St 7 47.06 By applying equations 2 and 3 to the measurements
SE main intake, St 8 29.33 43.57 taken from the Main Return section (XC-1 through
Main return, St 5 51.63 80.29 XC-3), the following leakage path resistances were
Main return, St 6 48.22 74.54 estimated:
3. Gage pressure across stoppings, Pa
For condition 1,
Belt to return XC-1 187 423
Belt to return XC-3 149 398
Intake to belt, XC-20 44
Intake to belt, XC-30 34
4. Resistance per stopping, Ns2 /m8 giving Re = 14.44 Ns2 /m8 . And for n = 3, Ri =
Belt to return 130 112 130 Ns2 /m8 .
Intake to belt 320 For condition 2,
St = measurement station; XC = cross-cut.

900 rpm when the fan duty increased to 94.00 m3 /s of

air at 1200 Pa of static pressure. During condition 1,
pressure-quantity measurements were collected from
main intake and return airways located near the surface
fan (between stations 1 and 6 in Figure 1). During con-
dition 2, the measurements were expanded to monitor
the losses of fresh air near the workings (between sta-
tions 7 and 8 in Figure 1). In both cases the objective
was to determine the leakage path resistances.
Table 1 shows a summary of air quantity measure-
ments taken from the intake and return entries and
the pressure differences across a selected number of
stoppings. The data for condition 1 refers to measure-
ments collected when the fan was operated at 500 rpm,
and condition 2 when the fan speed was increased to
900 rpm.
3.1.2 Analysis of results
This analysis is restricted to leakage path resistances
estimated from air pressure-quantity measurements
shown in Table 1. In this case, a path is represented by a
set of parallel stoppings installed in cross-cuts adjacent
to the belt line. They are used to isolate the conveyor
belt line from both the intake and return entries. The
analysis is restricted to two sections: Main Return and
Southeast Mains. The Main Return section (between
stations 5 and 6 in Figure 1) includes three cross-cuts
giving Re = 12.40 Ns2 /m8 . And for n = 3, Ri =
112 Ns2 /m8 .
By applying this procedure to the measurements
collected from the Southeast Main section (between
stations 7 and 8), the following leakage path resis-
tances were estimated for condition 2:
giving Re = 3.20 Ns2 /m8 . And for n = 10, Ri =
320 Ns2 /m8 .
A preliminary evaluation of the above results shows
that the leakage path resistance depends on several
factors including the pressure differential across the
stoppings, the maintenance program utilized and the
age of the stoppings.
3.2 Description of Mine B
A schematic representing Mine B is shown in Figure 2.
The mining method used is a longwall (LW) unit with
two continuous miners (CM).
Dust is the major air contaminant in this mine.A sin-
gle split system is used to ventilate the workings. The
system is powered by an exhaust fan which is centrally

601

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Key
1 Station No.
Main Fan
R Regulator
D Airlock Door
North Mains No. of Intakes: 7
Intake
Portals
XC-29
1 2
4 3
D
East Mains
D
Intake
CM 2
Figure 2. Mine B representative ventilation schematic.
located with respect to the workings (Figure 2). This
is a 3 m diameter axial fan, equipped with a 950 kW
motor, capable of exhausting up to 440 m3 /s of air. Of
the seven intake portals, the one located nearest to the
LW workings, supplies about 40% of the total air flow-
ing into the mine. The fresh air is supplied to the work
areas through a set of parallel entries and the contami-
nated air is returned to the surface through another set.
Concrete, cinder block, and metal Kennedy stoppings
are used to separate the intake and return entries from
the belt entry.
The ventilation survey was conducted in September,
2005. As with Mine A, the objective was to estimate
leakage flow rates and head losses to determine resis-
tance values for various sets of stoppings and other
control devices.
3.2.1 Survey data
Table 2 shows a summary of pressure-quantity mea-
surements collected from this mine. The measure-
ments are divided into three sets: one set to determine
the overall efficiency of the system, and two sets to
determine the leakage flow parameters.
Taking into account that the mine utilizes stan-
dard concrete stoppings, the study on leakage paths
is restricted to a section of the East Mains spanning
between cross-cut 29 and 132. The East Mains section
is made up of five entries: two intakes, two returns and
one belt line. Figure 3 shows the ventilation schematic
used to determine the leakage path resistances.
Copyright © 2006 Taylor & Francis Group plc, London, UK
Intake Air
Return Air
Stopping
CM 1 LW
Overcast
No. of Returns: 1
XC-132
n = 103
R
Intake
3.2.2 Analysis of results
Two factors are evaluated in this section: percent ven-
tilation leakage, and leakage path resistances. The
percent ventilation leakage represents a fraction of
the total quantity of air that is lost through stoppings,
doors, overcasts and other control devices and returned
to the surface without any effective utilization, the
higher the loss the lower the efficiency of the sys-
tem. The amount of loss depends on several factors
including the materials and techniques used to build
stoppings, the maintenance program undertaken, and
the pressure differences created by the main fans.
Based on the quantities of air directed to active
workings, old seals, and underground shops, and the
quantity of air passed through the main fan, the percent
leakage was calculated using equation 1 as follows:
Of this, 29% (114 m3 /s) of the fresh air is lost in old
mining districts.
3.2.3 Leakage path resistances
The leakage path resistances were calculated for two
rows of stoppings located in the East Main section
(between XC-29 an 132). Table 2 shows part of the
pressure-quantity measurements taken from this sec-
tion. Item 2 shows the measurements for the stoppings
located between the intake and belt entries, (far side
from main fan) and item 3 the measurements for those
602
 Table 2. Leakage path survey data for Mine B (air density 900
ρ = 0.93 kg/m3 ).
750 Belt to Return

Pressure Drop, Pa
Pressure Quantity 600 Intake to Belt
Item Survey description drop Pa m3 /s
450
1. Primary ventilation survey
300
Fan operating point 1790 393.00
Fan airlock door 1670 105
Air lock doors by overcast 700
Longwall head gate 30.90 0
36 45 60 74 89 104 118
Bleeder entries 9.50
Continuous miner sections 17.70 Cross-cut Number
Seals (mine out areas) 56.70
Underground shops 9.60 Figure 4. Pressure drop across stoppings.

2. Leakage flow survey – intake to belt

Intake 1 (left of XC-29) 152.15 the air in the belt line. The graphs of this figure show
Path a1, n = 15 168 4.91 that the pressure differences decrease with the distance
Path b1, n = 1 87 6.05 from the surface fan. For similar stoppings, the rate of
Path c1, n = 29 75 7.65
change follows a decay function with high values near
Path d1, n = 29 60 7.65
Path e1, n = 29 62 7.61 the fan and low values near the workings.
Intake 2 (right of XC-132) 118.51 Equations 2 and 3 were used to calculate the equiv-
alent resistances for the two sets of stoppings. For the
3. Leakage flow survey – belt to return far side (intake to belt), this resistance was calculated
Return 4 (right of XC-29) 185.17 as follows: P = [168 × 15 + 87 × 1 + (75 + 60 +
Path a2, n = 15 814 14.17
62) × 29]/103 = 80.8 Pa; Qd = (152.15 – 118.51) =
Path b2, n = 1 712 1.46
Path c2, n = 29 553 8.03 33.64 m3 /s. Re = 0.0714 Ns2 /m8 . For n = 103, equation
Path d2, n = 29 398 8.03 3 yields Ri = 757 Ns2 /m8 .
Path e2, n = 29 321 7.98 By applying the same procedure, for the near
Return 3 (left of XC-132) 145.49 side set of stoppings, the following resistances
were estimated: Re = 0.3071 Ns2 /m8 and for n = 103,
4. Resistance per stopping, Ns2 /m8
Ri = 3258 Ns2 /m8 .
Intake to belt 757
Belt to return 3258 The figures show that the leakage path resistance
also depends on the geometry of the ventilation system
n = number of stoppings per branch. (single or double split), and the relative position of the
fan in relation to the sets of stoppings.
XC-132
XC-29

n = 103 4 ANALYSIS OF AIR LEAKAGE THROUGH

1 2
a1 b1 c1 d1
a2 b2 c2 d2
4 3
940 kW Exhaust Fan
(Q = 392 m3/s, 1790 Pa)
Figure 3. Leakage paths in East Mains of Mine B.
located between the belt and return entries (near side).
Figure 4 shows the pressure profiles for the two sets of
stoppings. Because of the relative position of the fan
as an exhauster, the stoppings on the near side are sub-
ject to higher pressure differences than those located
on the far side. Another reason for the high pressure
differences on the near side is that in the belt line the air
moves inby with the intake air and there are numerous
feed points and small windows that are used to control
Intake
STOPPINGS
e1
Belt
e2 This analysis is restricted to a section of Mine B. For
simplicity the section has been modified to include 90
Return cross-cuts only. To establish a U-tube vent system, it is
assumed that the belt entry is isolated from others by
two rows of stoppings: one that separates the belt from
the intake entries and the other from the returns. For
the analysis, it is assumed that each stopping on the
intake side includes a small window to control the air
in the belt line. The stoppings on the return side do not
have any regulator. All the vent doors on the return side
are installed in concrete walls and kept as airtight as
possible. The resistances of other airways are assumed
to be equal to those determined from pressure-quantity
measurements. Figure 5 shows the vent schematic
used for this analysis. In this graph, each leakage path
represents ten stoppings. The flow rate in the conveyor
entry is controlled by two regulators. The system is

603

Copyright © 2006 Taylor & Francis Group plc, London, UK

 300 m

FQ1 R1 FQ2

R2

R1 = 7.57 Ns2/m8 R2 = 32.58 Ns2/m8 66.5 m3/s

Q = 102 m3/s
FQ1 = 7.1 m3/s FQ2 = 4.7 m3/s
P = 1000 Pa

Figure 5. Sample mine ventilation schematic.

Table 3. Airways parameters for sample ventilation 40

schematic.

Leakage Fraction, %
39
Resistance Fixed quantity
Airway Number Ns2 /m8 m3 /s 38

Intake branch 10 0.00423 37

Belt branch 10 0.0903
Return branch 10 0.00304 36
Face 2 0.07269
Leakage path 1 9 7.57 35
0 1000 2000 3000 4000
Leakage path 2 9 32.58
Fan Pressure, PA
Regulator 1 1 7.08
Regulator 2 1 4.72
Figure 6. Effect of fan pressure on leakage fraction.

1600
powered by a fixed pressure fan. Table 3 shows the 1400
Pressure Drop, Pa

network parameters used in the analysis. 1200

4.1 Network formulation
Based on Figure 5, a computer network was created
using a ventilation simulator, VNETPC. The network
included 50 branches and 32 junctions. Of the 50
branches, 18 were assigned as leakage paths. Two regu-
lators of adjustable size are used to control the quantity
in the belt line. One regulator is located at the entrance
of the belt entry and the other near the face. The flow
quantities passing through these regulators are set at
7.08 and 4.72 m3 /s of air respectively. Initially, the fan
pressure is set at 2000 Pa. Then, this is changed to
produce various flow rates.
4.2 Simulation results
The ventilation simulator was used to study two fac-
tors: (1) effect of fan pressure on leakage quantity
and (2) effect of replacing multiple stoppings by a
fixed number of leakage paths. For the former, the fan
pressure was varied between 500 and 4000 Pa and the
percent leakage evaluated. For the latter, initially for
each row, a single path represents 10 stoppings (9 paths
for 90 stoppings). Next the number of stoppings per
1000 Belt to Return
800
600
IIntake to Belt
400
200
0
0 20 40 60 80 100
Number of Stoppings
Figure 7. Pressure across stoppings (fan P = 2000 Pa).
path was increased to 15 and then to 30, and the leak-
age quantities evaluated. For each path, the equivalent
resistance was calculated by applying equation 3.
Figure 6 shows the changes in percent leakage with
the fan pressure. As expected, the leakage fraction
increased with the fan pressure. Except for some low
pressures, the rate of increase is almost linear. The cur-
vature of the graph depends on the type of stopping,
the lower the resistance, the higher the slope.
Figure 7 shows two graphs depicting the changes
in gage pressure with the number of stoppings. The
graphs show finite pressure gradients that resemble
decay functions which become asymptotic with the
number of stoppings (distance from the main fan).

604

Copyright © 2006 Taylor & Francis Group plc, London, UK

 The stoppings located in the vicinity of the fan are
subject to high air pressure differentials. These stop-
pings, if not designed and maintained properly, will
eventually develop fractures and cracks that will affect
the strength of the stoppings and cause damage, thus
increase the leakage quantity over time.
The equivalent resistance for a set of stoppings was
calculated using equation 2. The pressure difference
for each path was determined based on at least three
pressure readings. For similar stoppings, the pressure
difference decreases with the number of stoppings. As
shown in Figure 7, the rate of change is fairly linear
and for long rows an arithmetic average is an accept-
able approximation. The leakage quantity for a set of
stoppings was determined by flow measurements in
adjacent drifts before and after the set. For a set of
n number of stoppings, the individual stopping resis-
tance was calculated by multiplying the equivalent
resistance by the number of stoppings to the sec-
ond power. This approximation works well for similar
stoppings in parallel arrangement.
For this analysis, the network of Figure 5 was mod-
ified by decreasing the number of leakage paths from
18 to 6 and calculating their equivalent resistances
using equation 3. For instance, if a single stopping
resistance is Ri = 757 Ns2 /m8 , then the equivalent
resistance would be Re = 7.57 Ns2 /m8 for a set of 10
stoppings. The same approach was used to determine
equivalent resistances when the number of stoppings
per path increased to 15 and then to 30.
When the vent system shown in Figure 5 was pow-
ered by a fan pressure of 2000 Pa, using VNETPC, a
leakage fraction of 0.3797 was estimated (37.97% of
quantity loss). When the number of leakage paths on
both sides of the conveyor line was decreased to 6 paths
(each representing 15 stoppings), the leakage fraction
dropped to 0.3784. The same procedure was repeated
when the number of paths decreased to 3 and then to
2. In each case, the leakage fraction remained prac-
tically constant. These results show that for planning
purposes it is fair to replace a number of stoppings by a
leakage path of equivalent resistance calculated using
the square law (equation 3).
5 DISCUSSIONS
In ventilation planning, resistance values are measured
in the field. These measurements are then used in a
simulated model. which is calibrated to match the field
conditions. This calibrated model is used to estimate
the pressure/quantity requirements for future mining
scenarios.
In item 4 of Table 1, the resistance per stopping
between the belt and return stoppings at the Main
Return section remained relatively constant under
varying pressure conditions. Under condition 1,
605
Copyright © 2006 Taylor & Francis Group plc, London, UK
Ri = 130 Ns2 /m8 . For those same stoppings under con-
dition 2, Ri = 112 Ns2 /m8 . These results help to vali-
date the assumption typically made in vent planning
and design.
Measurements were made at both the Main Return
section and the Southeast Main section while the fan
was running at 900 rpm (condition 2). The resistance
of an individual stopping in the Main Return (immedi-
ately inby the main fan) was significantly lower than an
individual one in the Southeast Mains (900 m inby the
main fan). Since the same materials were used in each
location, and assuming that the same quality of work-
manship was applied, the most likely reason for the
difference is the respective ages of the stoppings. The
stoppings in the Main Return are approximately 1 year
older than those of the Southeast Mains. Personnel at
both mines indicated that crucial doors, regulators, and
other vent controls where checked periodically and if
a structure was significantly damaged or deteriorated
then it was repaired, but no rigorous stopping main-
tenance program was being implemented throughout
the entire mine.
The resistance of stoppings decreases over time due
to deterioration caused by external forces such as roof
and floor convergence and high pressure differential.
As stoppings age, small holes and cracks are generated,
especially at the contacts with the roof and ribs. High
pressure differentials cause these openings to increase
in size, causing additional deterioration which results
in lowered resistance.
In item 4 of Table 2, the resistance per stopping on
the intake side is significantly less than on the return
side. The same materials were used to construct both
lines of stoppings and they were constructed at about
the same time. The primary difference between the
measured resistance values is attributed to the fact that
the belt entry is not maintained as neutral air, but that
it has a small amount of intake air moving toward the
working areas. Therefore, the air passing through the
stoppings from the intake to the belt entry should not
necessarily be considered leakage.
In each case, the resistance/stopping values mea-
sured at Mine A were significantly lower than those
measured at Mine B. The most likely cause for this
is the difference in materials used for construction.
Mine B uses standard concrete blocks for stoppings
subject to high pressure differentials (mains sections)
and Kennedy stoppings in low pressure areas. Whereas
Mine A uses Omega blocks in both high and low pres-
sure areas. It should be remembered that the maximum
pressure of the fan at Mine A is roughly half of that at
Mine B.
The circumstances of a real-world mine rarely
exhibit the ‘ideal’conditions needed to obtain the most
accurate measurements. The airflows and pressure
drops in an underground mine are subject to consider-
able variation due to movement of equipment, opening
 vent doors, and other changes. In addition, mine lay-
outs are often extremely complex and may be such
that the airflow profile at the location where a mea-
surement is required is not fully developed. This can
make so that fluid-flow laws are not truly applicable.
Nevertheless, a practical effort must be made.
6 CONCLUSIONS
Stoppings and other control devices used in under-
ground mines can be viewed as air paths with high
resistance. The amount of air that leaks through
these devices depends on the type of construction
materials, workmanship, and maintenance. Improp-
erly constructed and poorly maintained structures will
significantly lower this resistance and cause undue
leakage.
ACKNOWLEDGEMENTS
This paper was prepared with financial support of
the NISOH Pittsburgh research center, project num-
ber: 254-2005-M-12666; their assistance is gratefully
acknowledged.
The authors would also like to thank the mine per-
sonnel for allowing us to conduct the surveys, for
606
Copyright © 2006 Taylor & Francis Group plc, London, UK
their assistance and support during the surveys, and
for sharing their field data with us.
REFERENCES
Bruce W.E. & Koenning T.H. 1987. Computer Modeling of
Underground Coal Mine Ventilation Circuits, Selection
and Application of Airway Resistance Values. Proceed-
ings of the 3rd US Mine Ventilation Symposium, SME,
Littleton, CO, pp. 519–524.
Duckworth I.J., Wallace G.K. & Wise R. 1995. Ventilation
Planning and Design of the Skyline Mines. Proceedings of
the 7th US Mine Ventilation Symposium, SME, Littleton,
CO, pp. 9–14.
Fields K., et al. 2004. Study to Assess The Performance of
Ceramic Diesel Particulate Filters for Reducing Diesel
Emissions, Proceedings of the 10th US Mine Ventila-
tion Symposium. Balkema Publishers, New York, NY,
pp. 127–134.
McPherson M.J. 1993. Subsurface Ventilation and Environ-
mental Engineering. Chapman & Hall, London SE1 8HN,
UK, pp. 312–321.
Richardson M.A., Gilbride L., Adair L. & Glasson M.W.
1997. Ventilation Planning at the Aberdeen Mine. Pro-
ceedings of the 6th International Mine Ventilation
Congress, SME, Littleton, CO, pp. 33–38.
 13. Ventilation education and training

Copyright © 2006 Taylor & Francis Group plc, London, UK

 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Developments in mine ventilation legislation, statutory qualifications and

ventilation training in Australia

D.J. Brake
Mine Ventilation Australia, Brisbane, Australia

ABSTRACT: Australia has seen substantial changes to its framework of mine ventilation legislation, statutory
mining qualifications and ventilation training over the past 15 years. This has been driven by a number of
factors including: advances in underground mining technology, shortages of professionals, fly-in, fly-out (FIFO)
arrangements on many sites, the ascendancy of the “risk assessment” philosophy to control health and safety risks,
increasing cost pressures, globalisation of the mining industry, the changing nature of federal-state relations,
competition policy (anti-trust legislation), the trend towards competency-based qualifications and other factors.
Some of these impacts have been discussed in general terms in an earlier paper by Brake & Nixon (2004).
This paper reviews more specifically the trends in legislation, statutory mine ventilation qualifications and mine
ventilation training and discusses the likely future impacts not just for Australia, but also to other developed or
developing mining nations.

1 INTRODUCTION levels. This has been recognised by the many working

Australia is a country the same physical size as con-
tinental USA, but with only six states and one sig-
nificantly sized continental territory. Mine safety is
currently regulated at the state not federal level. There
is a wide variation in legislative approaches between
the states. Some states (Queensland) have specific and
separate legislation for coal and metalliferous mining;
some (WA) have combined legislation covering both
coal and metalliferous mining; some (NSW, SA) use
general “OH&S” legislation applicable to all indus-
trial activities with supplementary specific legislation
for mines whilst other states use only general OH&S
legislation (Tasmania). This leads to wide variations in
legislative requirements between the states, and also to
wide variations in the industry background of “inspect-
ors” with some states not having mines inspectors at
all, relying on workplace health and safety inspect-
ors who may have no mining experience. Some states
have certain tasks on the mine site that are statutory
appointments (such as ventilation officer); others do
not. Some states have boards of examiners that issue
statutory qualifications; others do not. Some individ-
uals now have statutory appointments issued by states
that have since abolished their boards of examiners
which effectively renders their statutory qualification
obsolete.
These changes are exacerbating what is already a
critical and growing shortage of skilled personnel at all
groups and conferences on recruitment and retention
in the past 2 years and the recruitment of increasing
numbers of overseas-trained mining personnel.
As a result of these factors, the role of ventilation
officer or engineer is undergoing major changes. This
paper reviews particularly the impacts of changing
legislative requirements and statutory appointments
and the development of Australia’s competency-based
qualification system for ventilation professionals.
2 THE CHANGING NATURE OF
FEDERAL-STATE RELATIONS,
COMPETITION POLICY AND
GLOBALISATION
Australian states were independent colonies of Britain
until 1901 when the states “federated” to form an inde-
pendent nation, the Commonwealth of Australia. The
national (federal) government was initially involved
principally with defence and foreign affairs. How-
ever, over the years, and especially in the past 20
years, the federal government is increasingly to pro-
vide nationally consistent legislative frameworks for
many activities that were earlier the prerogative of
the states. This process has often been taking place
with the states approval, although in some instances,
the states have been unwilling to cooperate in this
increasingly “federal” framework of government.

609

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Key drivers of the increasingly federal legislative
framework in Australia have been the much higher
mobility of personnel across states and the demand
by both individuals and organisations for a consistent
legislative framework to promote business opportun-
ities and to facilitate the “portability” of personal
qualifications across state boundaries. The drive for
improved national productivity has also meant that
the chronic and inefficient duplication of services and
government bureaucracies across states is difficult to
justify. Competition policy has also meant that “arti-
ficial” barriers to trade or business activities across
states (e.g. different standards for similar activities)
are being struck down by federal legislation and the
courts, leaving the states with little else other than
to adopt consistent standards. Globalisation is also a
key driver, in that large companies are insisting on
more consistent regulatory frameworks across states
due to the large costs involved in having different
business standards and procedures or different per-
sonnel requirements for similar or identical activities
in different states. The “terrorist” threat is also mak-
ing Australians more concerned about the problems of
inconsistent or “piecemeal” regulations or approaches
providing opportunities for terrorists and, conversely,
making Australians more comfortable with national
approaches to many issues. For example, it is now
widely recognised that the regulations regarding the
transport, storage and use of explosives (which was
legislated at state level with numerous inconsistencies
across the nation) is a flawed approach and a national
framework has now been adopted.
It is therefore likely that, in future, the Australian
state governments will be left as the key “deliverer” of
services, such as police, health, schools, local gov-
ernment and the like, with the federal government
providing a consistent legislative framework for most
aspects of most activities.
The same drivers for more consistent regulatory
frameworks are also likely to start to develop across
national boundaries. For example, if a mining nation
sets up a taskforce to review and make recommenda-
tions for new mine safety legislation, it is likely that the
global mining companies as key stakeholders would
become involved and would set out to recommend
what they believe is “best practice” legislation across
their portfolio of operating mines transglobally. In the
same way that the world is (slowly) moving towards
lower trade barriers, it is likely that the world will
also move towards a more consistent approach on legal
frameworks, including occupational health and safety.
A further example of the impact of globalisation is
the fact that major mining houses are developing their
own internal standards for key hazards, often including
underground ventilation (Anon 2001). It is clearly in
their interests for these standards to be able to be used
in all their operations.
610
Copyright © 2006 Taylor & Francis Group plc, London, UK
Benchmarking of legislation is slowly leading to the
conclusion that some types of legislation are super-
ior to others. For example, a review commissioned
by the NSW Government has found support from
several sources that key elements of the NSW legis-
lation should be changed to be similar to that in the
Queensland legislation (Anon 2000).
As part of this trend in Australia, the Conference
of Chief Inspectors of Mines (a meeting of the Chief
Inspectors of Mines of each state) has recently issued
the National Mine Safety Framework Implementation
Plan (Anon 2003a) and implementation of this plan
has started in all Australian mining jurisdictions. It is
interesting that the first two of the seven strategies in
this plan are provision of:
– A nationally consistent legislative framework, and
– Competency support to ensure that workers are
competent to do their jobs.
This clearly reflects the desire for “harmonisation”
of mine safety regulations across the country and the
concern from the regulators about ensuring all work-
ers in the industry are actually “competent” to do
their jobs.
3 RISK BASED LEGISLATION, APPROVED
GUIDELINES OR CODES OF PRACTICE,
AUDITS, PROSECUTION POLICIES AND
THE CHANGING NATURE OF MINING
INSPECTIONS
In the past 10 years, Australia has moved very com-
prehensively towards the “risk based” duty of care
approach to workplace health and safety legislation.
More recently, several states are seriously considering
adopting the “safety case” style of safety legislation,
which is highly regarded by the regulators of other
hazardous industries such as the North Sea oil and
gas platforms, etc. (Heiler 2005, Anon 2005a, Raman
undated). This general risk-based duty of care is
supplemented by varying amounts of “prescriptive”
regulations. In addition, most legislation also requires
operators to adopt the “ALARA” principle, i.e., to
reduce the risk not only to an acceptable level (defined
in the regulations) but also to the lowest that is reason-
ably achievable. The industry is still struggling to come
to grips with just what is meant in practice by “as low as
reasonably achievable” and how it should be assessed.
In some instances, the legislation defines this. How-
ever, it may well be that once this principle is tested in
a court, that what was “reasonably achievable” in hind-
sight after a serious safety accident may be interpreted
differently to what operators considered to be “reason-
ably achievable” before the incident. The way in which
risk assessments are to be conducted and the results
incorporated into hazard management plans and
 standard operating procedures is now carefully defined
in law. In many instances, ventilation officers are
now involved in these formal risk assessments requir-
ing them to be familiar with these concepts. In
some cases, they are required to facilitate these risk
assessments, which requires a new set of technical
and management skills. For example, in Queens-
land, all coal ventilation officers are now required
by law to have the national competency standard
MNCG2 “Facilitate the risk management process”
(Anon 2002c).
One of the outcomes of the trend away from pre-
scriptive legislation (and the resulting “void” that it
leaves) is that state governments have been developing
“Guidelines” or “Approved codes or practice” to sup-
plement the regulations (e.g. Anon 1997, Anon 2003b,
Anon 2003c, Anon 2003d). Mine operators do not have
to comply with these guidelines or codes, but they are
admissible in law, and if an operator has a related inci-
dent, then the operator must be able to demonstrate
that the level of risk under the operator’s own pro-
cedure was at least as low as that resulting from the
approved guideline or code (Anon 2003c). In addition,
new legislation often no longer specifies allowable
gas concentrations or noise levels etc, but defers to
the most recent national TWA or STEL limits in these
areas. A consequence of these changes is that ventila-
tion officers must be familiar with not only the mining
“regulations” but also a wide range of other standards
and documents.
The traditional role of mines inspectors has been
auditing against detailed regulations and investigating
incidents “after the fact”. The approach now in vogue
focuses more the “systems design”, i.e. the inspector
(often in a team) conducts an audit of the operators
risk assessments, procedures and outcomes, and then
issues binding and/or non-binding improvement or
non-compliance notices. Most of these audit template
documents are publicly available and are quantitative
(i.e. give a point score), so that self-auditing is pos-
sible and desirable (Anon 2003e). Ventilation officers
have always had an important auditing role, but this is
now becoming much more formalized and also more
focused on “systems” as well as the minutiae of the
regulations.
Finally, most states have issued “compliance pol-
icies” or “enforcement and prosecution policies” for
their safety legislation (Anon undated, Anon 1999,
Anon 2004). These clearly set out the factors that the
inspectorate will take into account when determining
whether to launch a prosecution against a company or
an individual.Typical factors that regulators might take
into account when determining whether action should
be taken against an organisation or individual, or the
severity of the action include:
• The sufficiency of safety management plans at the
operation
611
Copyright © 2006 Taylor & Francis Group plc, London, UK
• The implementation of safety management plans at
the operation
• Training of personnel in terms of:
– Content
– Accreditation, and
– Understanding of the safety management plans
• Communications in the organisation in terms of:
– Internal
– External and
– Retention of knowledge base (corporate memory)
• Previous history of incidents in this organisation
• The risk assessment process, especially controls put
in place as a result of the risk assessments
• Good practice across the industry
• What other risk control options were considered,
if any
• Audit systems in place
Taken together, these factors mean that the breadth
and depth of knowledge required by the ventila-
tion officer has increased significantly. The technical
knowledge and skills are increasing in complexity
but in addition, the ventilation officer must manage
technical issues in the light of “duty of care” and
“as low as reasonably achievable” legislation. The
potential shortfall in competency due to the growing
skill set requirement is aggravated by the reduction
in middle management positions (who in the past
had both the time and skills to provide an important
mentoring role for new ventilation officers or engi-
neers), leaving today’s ventilation officer with a sub-
stantial exposure to prosecution if an error of judge-
ment or any negligence were to occur and an incident
resulted.
4 CHANGING TECHNOLOGY
Changing technology is having a significant impact
on the nature of the ventilation officer’s role. For
example, gas monitors are now available that can anal-
yse four or even five gases using the one belt-worn
(handheld) instrument. The relative ease of use of
these devices has meant that legislators and managers
expect more extensive gas testing to be undertaken
than in the past, where “stain tubes” were the only
practical method of gas testing for spot samples.
However, the ventilation officer may now also be
responsible for ensuring these devices are calibrated,
or even doing site calibration, requiring specialised
training.
Mining equipment is getting larger and has much
more powerful engines or motors.
Longwall machines are much more powerful so
that the width of a longwall face can be much longer
than before, and the length of the panel can be much
longer. Gas management is more complex as the higher
cutting rates and longer face width and panel depth
 mean that dilution with intake air is not always suf-
ficient, giving rise to a much stronger reliance on
gas drainage and goaf management. The greater seam
heights and wider faces also increase the risks from
windblast.
Development heading sizes are increasing. Duct
and fan technology is resulting in longer distances
being run on auxiliary ventilation. The advent of
remote-controlled machines has meant that mining
methods are changing, sometimes resulting in mine
design or operational requirements that conflict with
ventilation good practice. The higher productivity of
these machines and methods, combined with bet-
ter metallurgical processing, is making lower grade
orebodies economic, creating a more general trend
towards “caving” methods for hardrock mines.
5 TRAINING AND QUALIFICATIONS
In the past, mine ventilation training in Australia was
at two levels:
– Mining engineers obtained some ventilation know-
ledge during their undergraduate program. How-
ever, this was often not started until third year and
was seen as a follow-on to fluid mechanics taught in
first or second years. It was taught from a substan-
tially theoretical base, with practical experiments
depending on the quality of the local laboratory.
This is not a criticism, but it certainly wasn’t the
intention of most programs for students to be able
to finish their course and immediately practice as
a confident and competent ventilation officer in a
modern mine.
– Mining engineers or non-professional persons
could attend ventilation “courses” offered by pri-
vate individuals. Often these courses were really a
re-hash of a university course, but perhaps cover-
ing additional material. In all cases, these courses
were not assessed (required no assignments, exam-
inations or practical tests), so that neither the
knowledge nor the competence of those attending
was known at the end of the course.
This approach was typical not just of mine ven-
tilation training, but also of much of the training
in the industry. Training was not assessed and the
resulting competence of the individuals was left to
the subjective opinion of peers or management. This
leaves the manager very exposed in the event of a
safety issue occurring in the mine, given the man-
ager’s “duty of care”. Just how does the manager prove
that he has discharged his “duty of care” when send-
ing a young engineer to a non-assessed ventilation
course?
Recognising the shortcomings of these approaches,
Australia has, over the past 10 years, developed or
612
Copyright © 2006 Taylor & Francis Group plc, London, UK
greatly expanded its range of vocational qualifications.
These were developed at the national level by industry
groups or advisory boards but have been subject to rig-
orous consultation and peer-review before being given
federal and state ministerial endorsement and official
legislative backing. All competencies are now listed
in the National Training Inventory Service (NTIS
http://www.ntis.gov.au/) and are legally recognised
throughout Australia providing both true “national
accreditation” across Australia and “national portabil-
ity” of individual qualifications between states. The
Australian Qualifications Framework (AQF) defines
the standards required at each level of competency-
based qualification.
The Resources and Infrastructure Industry Skills
Council (RIISC http://www.riisc.com.au/) [formerly
the National Mining Industry Training Advisory Board
(NMITAB)] has obtained federal and state government
approval for new national industry training packages
for both the coal and metalliferous mining industries.
These new competency-based packages are designated
MNC04 and MNM05 and replace the earlier (and
“first run”) packages MNC98 and MNM99 endorsed
in 1998/1999. The training packages are not meant to
replace academic qualifications, such as degrees, but
they are meant to actually equip mine personnel to
the point of actually being “competent” in a practical
sense to do a task such as manage the ventilation in the
mine, rather than just understand ventilation principles
and theory.
These training packages provide a list of compe-
tencies and career paths for the industry at worker,
supervisor and manager level. Completion of spe-
cific competency units can lead to qualifications
such as the Advanced Diploma in Coal Mining Man-
agement or the Advanced Diploma in Metalliferous
Mining.
All these units are competency-based. This is a
different concept to University-style of teaching and
assessment, which is largely knowledge-based (pass-
ing a knowledge-based exam) or earlier statutory
requirements, which were frequently experience-
based (“must have three years practical experience”,
etc). The adoption of competency-based qualifica-
tions is, over time, resulting in significant changes
to the training environment in the industry and will,
in the future, impact very substantially on health
and safety outcomes. Note that in both Queensland
and NSW, a graduate mining engineer no longer
automatically qualifies for a ventilation officer’s
“ticket”, even after obtaining suitable industry expe-
rience. For example, the Queensland Mining Board
of Examiners has nominated the national qualifica-
tion MNCU1109A Manage, operate and maintain the
mine ventilation system as being the key requirement
for the statutory position of Ventilation Officer in
underground coal mines rather than an undergraduate
 mining engineering degree. It is likely that other Table 1. New ventilation-related competencies in the
regulatory authorities will follow in the future. Simi- Advanced Diploma in Mine Ventilation. For a complete list
larly, Queensland has linked the statutory positions of of competencies, see Table 3.
Underground Deputy, Open Cut Examiner and First
AMV100- Be able to identify, analyse and
Class Mine Manager into the national vocational qual-
Identify, analyse evaluate problems involving mine
ifications and not into academic or experience-based and evaluate air and increasing or decreasing
qualifications. psychrometric amounts of humidity, and to assess
An emerging problem for the industry is that, where heating and surface or underground mine
local state legislation requires specific mine appoint- cooling processes workplace climate (temperatures,
ments (such as ventilation officers) to be statutory and climate humidity, etc)
positions, the same local legislation also sets up a board AMV101- Be able to understand heat illness,
of examiners to accredit such personnel. The board Establish the assess the thermal environment for
of examiners maintains a supervisory role regarding heat stress heat stress in terms of recognised
the certificates issued under its jurisdiction so that, if management Australian and international
subsequent events require, the board will suspend or plan standards, and develop and
cancel such accreditation for an individual. manage an appropriate heat
stress management plan
It is now the case that certificates of competency
issued under mining legislation that has since been AMV102- Be able to assess heat loads in the
revoked may not be recognised by other regulators. Establish the mine and determine the need for
thermal refrigeration and the advantages and
For example, Tasmanian mine safety legislation for
environment disadvantages of the various types
many years provided for a Board of Examiners to management of mine refrigeration and delivery
issues First Class Mine Manager’s Certificates of plan methods
Competency. Many current Australian professionals
AMV103- Be able to create computer-based
have obtained Tasmanian accreditation. Under mutual Establish the ventilation models, audit these models
recognition legislation, this Tasmanian Certificate has mine ventilation against measured underground data,
been recognised in another state if the individual trans- model and conduct and conduct and then use the models
fers to another Australian jurisdiction and obtains network analyses for fault-finding, network analysis
an “endorsement” of the Certificate in that state. and future mine planning or fan duty
However, Tasmania no longer has specific mining AMV104- Be able to understand the various
legislation and has disbanded its Board of Examiners. Establish the types of mine emergencies that may
This means there is now no authority that supervises ventilation impact on the mine ventilation system,
and, if necessary, suspends or revokes a certificate emergencies the likely disruptions to the mine
issued by the Tasmanian Board. Therefore there is the (egress and ventilation system and the
potential for all past Certificates issued by the Tas- entrapment) consequential impacts on egress and
management entrapment, and to be able to
manian Board of Examiners to become effectively
plan select and design egress and
worthless. It is possible that other states may fol- appropriate entrapment options
low Tasmania’s lead and abandon their mining OH&S
legislation, relying instead on general OH&S legis-
lation. If this were the case, then it could be argued
that linking statutory appointments into AQF qualifi-
cations (with re-accreditation if necessary as required) In response to industry demands, another national
would provide a more robust and enduring solution competency-based ventilation qualification has been
to this problem. This is particularly true since new introduced, the Advanced Diploma in Mine Venti-
legislation often requires “refresher training” to be lation, which is designed to integrate seamlessly
completed to an accredited standard every five years. with the existing AQF ventilation competencies. This
Therefore, if a ventilation officer does not practice in Advanced Diploma provides dual-trained ventilation
the area of ventilation and is not re-certified, then specialists in both coal and metalliferous (hardrock)
his/her “ticket” will effectively be cancelled after applications.
five years. The ventilation competencies and their links to
At a more general level, linking statutory appoint- the industry training packages are shown in Table 2.
ments into industry-recognised, competency-based Note that five new ventilation-related competencies
qualifications is also a more “transparent” accredi- (Table 1) have been introduced into the AQF, all at
tation process with very carefully defined outcomes Advanced Diploma level. These were selected on the
and should ensure more consistent and reliable results basis of key ventilation-related issues in Australian
for the industry as a whole. It should also make it even mines.
easier than is currently the case for individuals to be Completion of the Advanced Diploma in Mine Ven-
able to work within different jurisdictions. tilation provides successful students with a single

613

Copyright © 2006 Taylor & Francis Group plc, London, UK

Copyright © 2006 Taylor & Francis Group plc, London, UK

Table 2. AQF mine ventilation competencies and links to other AQF and Queensland statutory mining qualifications.

Advanced 1st class Coal 1st class

Advanced diploma mine Mgr metalliferous
Black coal/ Coal diploma Advanced metalliferous certificate of mine Mgr
metalliferous statutory VO Metalliferous in mine diploma coal mining competency certificate of
Name of unit AQF designation competency VO competency ventilation mining Mgt (underground) (Qld) competency (Qld)

AQF designator MNCU1109A MNMMSM631A 39139QLD MNC60204 MNM60104 n/a n/a

Establish the ventilation MNCU1106A (coal) Mandatory∗ Mandatory Mandatory Mandatory Mandatory Mandatory Mandatory
management plan/system MNMMSM631A
(metalliferous)
Establish the spontaneous MNCU1102A Mandatory∗ Mandatory Mandatory Mandatory
combustion management
plan
Establish the gas MNCU1111A Mandatory∗ Mandatory Mandatory Mandatory
management system
∗
Establish the outburst MNCU1121A Mandatory Mandatory Elective Mandatory
management plan
614

Establish the gas drainage MNCU1116A Mandatory Elective Mandatory

management plan
Establish the mine AMV103 Mandatory
ventilation model and
conduct network analyses
Identify, analyse and AMV100 Mandatory
evaluate psychrometric
heating and cooling
processes and climate
Establish the thermal AMV102 Mandatory
environment management
plan
Establish the heat stress AMV101 Mandatory
management plan
Establish the ventilation AMV104 Mandatory
emergencies (egress and
entrapment) management
plan
∗
MNCU1109A only offered in conjunction with these additional competencies.
 Table 3. Complete list of AQF-ventilation competencies
provided to students completing the advanced diploma in
mine ventilation.
MNCU1106A Establish the ventilation management
plan
MNMMSM631A Establish the ventilation management
system
MNCU1102A Establish the spontaneous combustion
management plan
MNCU1111A Establish the gas management system
MNCU1121A Establish the outburst management
plan
MNCU1116A Establish the gas drainage management
plan
MNCU1109A Manage, operate and maintain the mine
ventilation system
AMV100 Identify, analyse and evaluate
psychrometric heating and cooling
processes and climate
AMV101 Establish the heat stress management
plan
AMV102 Establish the thermal environment
management plan
AMV103 Establish the mine ventilation model
and conduct network analyses
AMV104 Establish the ventilation emergencies
(egress and entrapment) management
plan
qualification comprising 12 individual nationally-
recognised mine ventilation competencies (Table 3)
covering both coal and metalliferous industries.
6 ADVANTAGES OF UNDERTAKING
ACCREDITED VENTILATION TRAINING
Even where an appointment is not “statutory” or does
not require a formal qualification, there are important
advantages in obtaining an accredited qualification
rather than attending an unaccredited course such as
traditionally occurred in the past, including:
– An AQF qualification is nationally recognised and
“portable” across states and territories under the
“mutual recognition” section of the Australian
Quality Training Framework (AQTF).
– The student has been assessed to a consistently
applied national competency “standard”. Unlike
non-assessed courses, the mine manager should
be able to be confident that the student can actu-
ally understand and correctly apply the knowledge
and skills developed during the course in a prac-
tical way. This can be particularly important in
assisting the manager to demonstrate that he/she
has met the company’s “Duty of Care” in this
important technical area, especially in the event of a
subsequent ventilation-related accident or incident.
615
Copyright © 2006 Taylor & Francis Group plc, London, UK
– The student him/herself can also be confident that
he/she meets an assessment standard used across
Australia for that course. The student will have used
their new skills in achieving competency in a situ-
ation away from both the training environment and
the course presenters. This independent assessment
helps the students to develop confidence in their
own abilities.
– Students intending to take up a role as Ventilation
Officer, Planning Engineer or the like, or a longer-
term career as a ventilation specialist, benefit from
obtaining a formal accreditation rather than merely
attending an unassessed course.
As accredited training courses do not attract
Australian GST (10% goods and services tax), the
additional cost for the fully accredited course is
often competitive with the cost for the non-accredited
version.
7 WHY PROVIDE FOR DUAL TRAINING
(COAL AND METALLIFEROUS)?
It is not uncommon to see hardrock-trained mine
surveyors working in coal mines, or coal-trained elec-
tricians working in hardrock mines, or any of a variety
of other situations occurring in which workers in one
industry move their skills and employment across to
the other industry. There have been recent examples
of mine managers and ventilation officers transferring
from coal to metalliferous and vice versa. It makes
sense to be able to draw on workers in one sector
to meet skills shortages in another sector. In addi-
tion, it improves career prospects and may have other
advantages to individual workers in terms of better
meeting their personal requirements (e.g. for a res-
idential mine rather than FIFO), which may avoid
their loss to the industry entirely. Often, the system
of “qualifications” in the past resulted in restraints
on trade in personal services and, effectively, “closed
shops”.
Transferring staff between coal and hardrock also
results in “cross-fertilisation” of ideas and helps with
technology transfer and has other related benefits to
both sectors.
With the rationalisation of Australian mining com-
panies and the growth of global mining houses (many
of which have both coal and metalliferous portfolios),
many of these international mining groups see the
benefit of engineers and other staff having exposure
to a wider range of minerals produced by the mining
house.
Finally, whilst some of the operational details of
mine ventilation are clearly different between a metal-
liferous mine and a coal mine, the basic principles
and many of the fundamental competencies remain the
same. For example, checking a gas reading, measuring
 airflow or pressure across a ventilation control, design- Table 4. Selected major US hardrock mining disasters.
ing a drop-board regulator, establishing a fan pres-
sure/flow specification or investigating the direction Year Mine Mineral Type Deaths
and volume of leakage.
It is often a widely held view that hardrock oper- 1972 Sunshine Mine Silver Fire 91
1971 Barnett Fluorspar Hydrogen 7
ations do not have problems with gas, or explosions,
Complex, Ozark- sulfide gas
or outbursts, or spontaneous combustion. However, the Mahoning Co.
following examples indicate that this is not necessarily 1968 Belle Isle Mine Salt Fire 21
the case. 1963 Cane Creek Mine Potash Explosion 18
1943 Boyd Mine Copper Explosion 9
1942 Sandts Eddy Limestone Explosion 31
– One Australian hardrock mine currently has prob- Quarry (surface)
lems with hydrogen as strata gas. A coal-based gas 1926 Barnes Hecker Iron Flood 51
drainage specialist has been engaged to advise on Mine
options. There is regular “popping and banging” 1917 Granite Copper Fire 163
due to minor ignitions. Mountain Shaft
– Another Australian hardrock mine currently has
problems with CO2 accumulations in the mine. The
surrounding strata are carbonates and CO2 builds
up in poorly ventilated areas. This operation has had with several strata gases and outbursts of CO2
several instances with miners becoming affected by and N2 .
CO2 , even losing consciousness. They have CO2 – Steam outbursts resulting in multiple fatalities have
sensors and a telemetric system installed on major been reported in a Japanese gold mine.
return airways. – A number of hardrock mines in Australia have had
– Many Australian hardrock mines have methane as problems with windblast, with the Parkes mine in
a strata gas. NSW suffering from a major windblast in the past
– Uranium mines have significant issues with radon 5 years that resulted in a multiple fatality.
gas and radon daughter products. A number of ven- – Almost all gold and platinum mines in South Africa
tilation strategies applicable to gassy coal mines experience problems with flammable gas and sev-
have relevance to uranium operations. eral have had problems with gas outbursts. Between
– The Isa mine in Queensland has a large high-grade 1989 and 1999, there were 25 fatalities and 36 ser-
orebody that was subject to spontaneous combus- ious injuries due to flammable gas issues in South
tion when attempts were made to extract it in the African hardrock mines. The average gas concen-
1960s. The ore was so susceptible to spontaneous tration across all mines was 66% methane and 26%
combustion that broken ore would increase its tem- hydrogen.
perature to over 1000◦ C. Large quantities of SO2
Therefore even on just the basis of gas management,
and CO2 were also produced. LHD tyres would
hardrock mines as a group have the potential for sig-
catch on fire. The orebody remains unmined to
nificant safety and cost consequences. To these can be
this day.
added the issues of dust, diesel particulates and egress
– At least three mines in Western Australia have had
and entrapment.
more recent problems with spontaneous heatings in
In addition, note that most of the significant US
ores with high pyrite contents.
hardrock mining disasters (multiple fatalities) were
– Several hardrock operations have had problems
“ventilation related” (see Table 4).
with SO2 and H2 S strata gas. One WA mine had an
“outgassing” of CH4 that contaminated the return
airway for two full weeks.
8 CANDIDATES FOR THE QUALIFICATION
– Many hardrock mines have reported problems
with NH3 (ammonia gas) being produced when
The following groups of individuals are likely to be
ANFO (the most common underground explo-
interested in and benefit from completing the
sive) dissolves in water and comes into contact
advanced diploma in mine ventilation.
with lime from cement (also dissolved in water),
which results in an exothermic reaction producing – A “career” senior ventilation technical person in a
ammonia. mine
– Many hardrock operations have had significant – A ventilation superintendent responsible for a ven-
incidents with sulphide dust explosions, which have tilation department in a mine
had serious cost and safety implications. – A corporate (group) mining engineer responsible
– Potash, Trona and other evaporite mines overseas for oversight of ventilation across several mines
have had serious gas management/dilution issues – A ventilation consultant

616

Copyright © 2006 Taylor & Francis Group plc, London, UK

 – A regulator (e.g. Inspector) with a key “brief” in
the area of ventilation
– Fundamentally, any person looking to develop a
career in mine ventilation or with strong supervi-
sory/technical role mine ventilation.
9 SUMMARY AND CONCLUSIONS
The nature of mine safety legislation is changing in
Australia to a “duty of care” and “as low as reason-
ably achievable” approach. There is also a strong trend
towards harmonisation of safety legislation across the
country. The shortage of skilled personnel in both the
coal and hardrock mining industries at present is likely
to persist. A much stronger focus on educating the
workforce and providing improved opportunities for
movement of personnel between mines, employers and
between industries, along with methods of delivery
that fit in with the multiple requirements of work-
ers (personal situations, residential and FIFO, varying
rosters, etc.) will be required if the industry is to
develop a strong base of competent personnel. Well-
trained, competent personnel are also a key factor in
developing and maintaining a sustainably safe indus-
try. New competency-based ventilation training has
been developed to facilitate recognition of ventilation
qualifications throughout Australian jurisdictions and
provide maximum flexibility to meet both student and
employer requirements.
REFERENCES
Anon, undated. Compliance Policy (Principles and Pro-
cedures for Assessing Compliance and Determining
Response to Non-Compliance. Queensland Government.
Anon, 1997. Underground Ventilation (Metalliferous Mines)
Guideline. WA Department of Industry and Resources.
Document No: ZED569LF, December.
Anon, 1999. The Enforcement of Health and Safety Stand-
ards in Mines – Policy Statements and an Organisa-
tional Approach to Enforcement, Assessment of Mining
Operations, Investigation of Accidents or Incidents and
Prosecution. NSW Department of Mineral Resources.
Anon, 2000. Submission to the Green Paper on Revision
of Health and Safety Legislation in NSW Coal Mines.
CFMEU Mining And Energy Division. Final Report.
October.
Anon, 2001. Underground Ventilation Major Hazard Stand-
ard. WMC Resources.
Anon, 2002a. Minerals Industry Safety Handbook. NSW Dept
of Mineral Resources.
617
Copyright © 2006 Taylor & Francis Group plc, London, UK
Anon, 2002b. National Mine Safety Framework – Realising
a Safe and Healthy Mining Industry: the Contribution of
Government.
Anon, 2002c. Competencies Recognised by the Coal Mining
Safety and Health Advisory Council.
Anon, 2003a. National Mine Safety Framework Implemen-
tation Plan.
Anon, 2003b. Guidance Note QGN03 Healthy Atmos-
phere for Underground Mines. Queensland Government
Department of Natural Resources and Mines. Version 4,
Sept.
Anon, 2003c. Control of risk management practices
Coal Mining Safety and Health Act 1999 Recognised
Standard – 02. Queensland Government Department of
Natural Resources and Mines.
Anon, 2003d. Guidance Note QGN09 Reviewing the
Effectiveness of Safety and Health Management Sys-
tems. Queensland Government Department of Natural
Resources and Mines.
Anon, 2003e. Underground Ventilation Management HIF
Audit. WA Department of Industry and Resources.
Anon, 2004. Enforcement and prosecution policy.WA Depart-
ment of Industry and Resources.
Anon, 2005a. A National Quality Framework for Australian
Transnational Education and Training. Australian Govern-
ment Department of Education, Science and Training. 17
Nov 05.
Anon, 2005b. Education for the Minerals Industry – Report
of Joint Industry/University of Queensland Task Force.
AusIMM “Week in review”. 7 Feb 05.
Anon, 2005c. The case for safety cases? MineSafe. Volume
14 No 3, Sep 2005: 3–5.
Brake, D.J. & Nixon, C.A. 2004. The impact of changing
employment and mining practices on the mine ventilation
profession in Australia.
Brake, D.J. & Greenhill, G. 2005. New integrated pathways
for mine ventilation technical training in Australia. Pro-
ceedings of the 2005 Queensland Mining Industry Safety
and Health Conference. Townsville, Aug.
Hebblewhite, B. 2005. Mining Engineering Education Initia-
tives in Australia. SME Annual meeting. Salt Lake City,
Feb 28–Mar 2.
Heiler, K. 2005. Is the Australian mining industry ready for
a safety case regime? Proceedings of the 31st Interna-
tional Conference of Safety in Mines Research Institutes.
Brisbane October 2–5.
McCarthy, P. 1996. Challenges for staffing new projects. Pro-
ceedings of the 2nd Annual Australian contract mining
conference. Kalgoorlie.
McCarthy, P. undated. Outsourcing the mine planning func-
tion. AMC internal report.
McCarthy, P. undated. The Shortage of Experienced Mining
Professionals. AMC online library.
Raman, R. undated. Underground Mine Safety – are we doing
enough?
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Ventilation education and training

D. Chalmers
School of Mining Engineering, The University of New South Wales, Sydney, Australia

ABSTRACT: The University of New South Wales has provided Ventilation training and education to the Coal
Industry in Queensland, New South Wales, and New Zealand. Courses are offered on several Levels. Additionally
a Graduate Diploma in Ventilation has been offered for several years. All courses have been well attended and
have included candidates from Olympic Dam and from Large Iron ore mines in Sweden.
Approaches have been made to also been made to assist China in training and education with a view to reduce
the incidence and severity of methane blasts.
This paper reviews the delivery of these courses from Australia and explores the need for greater understanding
of ventilation for those practitioners who should be the first line of defence in ensuring that mines are safe. It
discusses the case for and against statutory responsibility for a ventilation officer.

1 INTRODUCTION evidence to meet ventilation competencies determined

The School of Mining Engineering has enrolled 144
students in ongoing professional development courses
in mine ventilation between 1998–2005 with probably
another 18 to be enrolled in February 2006. Addition-
ally, there have been approximately 50 students that
have undertaken a one week short course in mine ven-
tilation focusing on metal mine ventilation techniques.
The course developed from a request made by the
then Chief inspector for mines in NSW to provide
a programme that would allow mine personnel to
upgrade their skills, knowledge and resources to bet-
ter manage the risks associated with or controlled by
ventilation.
Professor Jim Galvin, Head of School assembled
a group of experts to develop this course. This team
included, Drs Roy Moreby, David Cliff, Ray Williams
and Messrs Paul Mackensie- Wood, and the author.
Courses run in 1998 and 1999 were structured in two
parts. Part A (Basic) was designed to provide refresher
training to mining personnel. Part B (Advanced) built
upon the underpinning knowledge and provided the
skills required for the ventilation officer.
It was quickly evident that there was a significant
shortfall in knowledge within the first groups and con-
siderable time was spent covering basic ventilation
theory.
1.1 Ventilation competencies
The programme was developed to provide the under-
pinning knowledge to allow candidates to assemble
by the mining industry training advisory board. These
competencies were:
– Assess and interpret the conceptual and legal basis
for mine ventilation
– Assess and interpret information essential to the
identification, analysis and evaluation of hazards
and risk associated with the mine ventilation system
– Identify, interpret and apply ventilation planning
laws, theorems and criteria
– Assess and interpret information essential to venti-
lation application and practice
– Access and interpret information underpinning
ventilation design
– Access and interpret the essential underpinning
ventilation monitoring systems and operations
– Access and interpret the essential information
underpinning ventilation management systems
– Identify and evaluate hazards and risk of mine
ventilation
– Contribute to mine ventilation system planning
– Develop, prepare and document mine ventilation
system procedures
– Forecast, calculate and plan for future ventilation
requirements
– Implement mine ventilation monitoring, recording
and reporting systems
– Implement the mine ventilation system mainten-
ance program
– Plan, co-ordinate and control changes to the venti-
lation system
– Audit and review the effectiveness of the mine
ventilation system

619

Copyright © 2006 Taylor & Francis Group plc, London, UK

 1.2 Programme structure
The programme is divided into eight modules, each
candidate completes modules 1 through 4, with coal
mining candidates completing modules 5 & 6 and the
metalliferous candidates completing modules 7 & 8.
A brief outline of each module is given below:
– Module 1 Ventilation and Mine Services
– Fluid Flow, Friction and Shock Losses
– Fans and Auxiliary Ventilation
– Ventilation Network Analysis
– Ventilation System Monitoring and Mine Ser-
vices
– Module 2 Environmental Contaminants
– Mine Gases and Gas Laws
– Atmospheric Gas Monitoring
– Airborne and Explosible Dust
– Mine Fires and Explosions
– Module 3 Heat In Underground Mines
– Psychrometry
– Heat Transfer and Sources Of Heat
– Heat Stress Management
– Module 4 Ventilation System Management
– Management Plans and Risk Assessment
– Project Economics
– Module 5 Mine Hazards and Control
– Gas Reservoir Characteristics
– Gas Drainage, Outbursts and Windblasts
– Spontaneous Combustion of Coal
– Module 6 Coal Mine Ventilation
– Coal Mine Legislation
– Coal Mine Ventilation Planning and Practice
– Module 7 Metalliferous Mine Hazards And Control
– Refrigeration
– Ionising Radiation
– Module 8 Metalliferous Mine Ventilation
– Metalliferous Mine Legislation
– Metalliferous Mine Ventilation Planning and
– Practice
1.3 Assessments
All participants will be required to demonstrate com-
petency on an ongoing basis throughout the course.
The range of evidence required to demonstrate com-
petency will be gathered from a diagnostic, formative
and summative assessment.
Diagnostic assessment aims to determine a partici-
pant’s competency gap or deficiencies and to identify
their training needs at the commencement of a topic.
Formative assessment is an assessment process that
establishes a participant’s progress in relation to each
learning outcome. The assessment will be recorded but
only as an aid to the candidate and the person assessing.
Summative assessment is an assessment that veri-
fies that a participant is competent in a learning
outcome. Each of these final assessments will address
620
Copyright © 2006 Taylor & Francis Group plc, London, UK
both the theory and the practical applications related
to the learning outcome.
2 RATIONALE FOR THE COURSE
In any mining process there are core risks, a signifi-
cant number of these can be controlled by ventilation.
Recent events have shown that when these are improp-
erly managed, the results can lead to loss of life, and/or
the loss of mine.
The approach made to the school to develop the
course was in response to the recommendations made
by the Moura Inquiry.
Events such as Moura suggest that a need for one
person who can recognise, provide ventilation solu-
tions and has the power and responsibility to manage
these risks. The ventilation officer would aid the mine
manager in safely managing the mine.
Some of the failings at the mine were the ineffi-
cient communication of the significance of the rising
indicators, coupled with a lack of responsibility, and a
simplistic management plan.
The creation of a statutory position of Ventilation
Officer was designed to provide one person whose sole
responsibility is the mine ventilation system. The fun-
damental thinking was that this person would not be
distracted by other operational constraints.
Previously, one or more of the undermanagers
would be responsible for the ventilation system as
an additional duty. Consequently, crisis management,
such as longwall breakdowns, could distract this
person from the bigger picture.
Another lesson that Moura should teach is that
such responsibility for the ventilation system should
be provided 24/7 with a relieving Ventilation officer
appointed to cover annual leave and illness. Where pos-
sible a handover period should be employed to enable
the oncoming person to be properly briefed, so that the
management is smooth and effective.
3 NSW EXPERIENCE
Shortly after the programme commenced the NSW
government enacted legislation that amongst other
things established the Ventilation Officer’s position.
Although this had been mooted as the direction that
the legislation would take, it was evident that some
mines would not be able to comply as those mines had
not sent anyone to be trained or assessed. Addition-
ally it was also suggested that those that were already
practicing ventilation management should be allowed
to continue to do so.
The legislation then provided for mines to accept
those that had been practicing before enactment to
be appointed ventilation officers without the need for
training, education or assessment. It also stated that a
 person that was to be appointed after the set date must
complete a course in mine ventilation provided by the
University of New South Wales or equivalent.
Other mines contracted a ventilation officer to pro-
vide coverage during significant ventilation changes,
to conduct ventilation surveys, and provide monthly
air readings while being employed on a part time
basis. One mine where this arrangement existed was
Southlands Colliery.
4 QUEENSLAND EXPERIENCE
Queensland also created a statutory position and
required that any person must meet the requirements
of the competency standards before they can be
appointed.
To meet this requirement the school was approached
to provide this training for mines in the Bowen Basin.
To date three courses have been run in Queensland, and
several candidates have joined courses run in NSW.
As the requirements to be appointed in Queens-
land are meeting the competency standard, any ser-
vice provider can register themselves to train and or
assess candidates. This has allowed other providers in
Queensland to assess candidates.
Additionally, people trained in NSW are able to be
appointed in Queensland.
5 CONFLICTS
The transfer of persons from Queensland to NSW
provides for some interesting and complex conflicts.
If the person has completed the UNSW pro-
gramme then they are able to transfer to NSW without
problems.
If they have been assessed by another service
provider then they must demonstrate that the course
that they have undertaken is equivalent to the UNSW
programme.
Although this is a requirement set out in legislation,
the legislation does not stipulate who shall make the
determination of equivalence. The inspectorate does
not believe that it is their responsibility, nor does the
school believe that it should be making judgements as
to the quality or adequacy of another service provider.
6 SUCCESS RATES
Of the 144 candidates that have commenced this
course 85 have received a Ventilation Officers’Certifi-
cate. 27 are in the process of completing. 6 completed
the Basic ventilation modules and either did not wish
to or were not in a financial position to continue.
26 candidates have either dropped out or failed to
complete assessment tasks.
621
Copyright © 2006 Taylor & Francis Group plc, London, UK
Of those that are in the process of completing it is
expected that about 20 of them will complete and be
awarded the ventilation officers’ certificate.
7 THE CASE FOR TRAINING
Prior to the commencement of this course, the per-
son responsible for ventilation was an undermanager.
This undermanager, had several other functions. The
primary function was to supervise the operation of a
shift. The education that they would have received typ-
ically would have been a four year part time course in
coal mining methods. This course included 28 hrs of
ventilation instruction. As such this ventilation train-
ing could only cover the basics to assist in the day to
day running of a coal mine.
Ventilation was seen as a secondary area of study
and only a means to secure a position as undermanager.
Additionally the legislation at the time only required
monthly air readings to be taken. Little emphasis was
given to the effects of ventilation provided that prod-
uction levels were maintained. If a significant event
occurred, then consultants would be used. Larger com-
panies had a ventilation “guru” to provide additional
support as an in-house consultancy.
The downturn in commodity prices led to a reduc-
tion of in in-house expertise, a push for greater pro-
ductivity and the distribution of responsibilities across
fewer people.
The demands for higher production stressed ventila-
tion systems leading to “gas outs” and other significant
incidents. The prevention of most of these incidents
could have been achieved by mine personnel having
a greater understanding of ventilation, seam gas char-
acteristics and spontaneous combustion.
Since the introduction of the programme there have
been fewer incidents and several mines were able to
improve their ventilation efficiency. The focus has
shifted from crisis, or just in time management to
timely considered forward planning.
In summary, the education, training and appoint-
ment of ventilation officers, has better managed
ventilation systems in coal mines.
This has been achieved by having one person
responsible for the gas, ventilation, outburst and
spontaneous combustion management plans, a person
whose primary function is to monitor and maintain the
health of the ventilation system.
8 THE CASE AGAINST HAVING A
VENTILATION OFFICER
The appointment of a ventilation officer provides the
opportunity for the appointee to be held accountable
for the faults and failings of the ventilation system.
 Also the creation of another management position
creates an additional link in the communication chain
with the potential that correct actions may not be taken
if a breakdown does not allow mine managers to make
informed decisions.
However this can also occur if part time ventila-
tion officers are contracted to mines where a man-
agement company has subcontracted the operational
management to a third party.
9 CONCLUSIONS
Ventilation training is essential to the safe running of
coal mines
Having one person who is responsible for the oper-
ation of management plans allows a mine manager to
focus on the overall operation of the mine in the know-
ledge that he has a person that is well trained and
competent to advise him.
622
Copyright © 2006 Taylor & Francis Group plc, London, UK
Although the school has assessed more people than
there are underground mines there is still a high
demand for the course.
Some mines are still requiring new ventilation offi-
cers, either to cover their statutory responsibility of to
provide additional ventilation officers to cover annual
leave and other absences.
Some candidates that completed the courses
decided that after receiving the training that the
position of ventilation officer was not for them.
At least 10 candidates have used the ventilation
officer’s course as a stepping stone to obtain a mine
manager’s certificate.
An educated workforce is able to better recognise
hazards and in a culture that allows for and encour-
ages communication provides the potential for a safe
working environment.
An uneducated workforce is not an option.
 11th U.S./North American Mine Ventilation Symposium 2006 – Mutmansky & Ramani (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-40148-8

Observations on mine ventilation education in the United States

C.R. Hairfield & K.G. Wallace, Jr.

Mine Ventilation Services, Inc.

ABSTRACT: Mine ventilation engineering curriculums vary significantly between the various undergraduate
mining engineering programs in the United States. Each mining program has a different curriculum that stresses
various aspects of the mining/minerals engineering discipline. The respective curriculums have been developed
and modified throughout their lifetimes to teach mining subjects based on the experience of individual instruc-
tors, the influence of surrounding industry and the career paths of the graduating students. Whether a graduating
engineer is entering the industry as a ventilation engineer or as an engineer for an underground mining oper-
ation with some responsibility or oversight for mine ventilation, it is necessary to have adequate knowledge
of ventilation. This paper analyzes the relationship between the varying degrees of ventilation education and
the capabilities of ventilation engineers in the field. The views of a recently graduated ventilation engineer as
to which ventilation topics have proven to be the most useful, and which should be given more attention in
the university setting are also described. In addition, the paper discusses some issues facing mine ventilation
engineering in the United States.

1 INTRODUCTION 2 UNIVERSITY MINING ENGINEERING

There are approximately fourteen Mining Engineer-
ing or mining-related undergraduate programs in the
United States. Based on information listed on each
respective program’s website, which is summarized
in Table 1, all programs offer a required ventila-
tion course for mining engineering students. How-
ever, it is becoming increasingly apparent to the
authors that there may be a general lack of venti-
lation knowledge among younger engineers at some
mining companies. This conclusion is based on the
type of questions being asked by young engineers
relating to ventilation. Fundamental questions, such
as how to perform ventilation surveys and how to
develop a model, are frequently being asked. Hence
it appears that at some university mining engineering
programs, ventilation may not be getting sufficient
attention. As a consequence, a question arises: “Are
graduating mining engineers receiving sufficient ven-
tilation knowledge to meet the needs of the mining
industry?”
The authors give their opinion on the current status
of mine ventilation education in the United States and
offer some insight from a recently graduated mining
engineer in the educational needs in the field of mine
ventilation.
PROGRAM INFORMATION
Using information available from either published
documents or via internet searches, the authors found
fourteen accredited mining engineering programs in
the United States. Each of these programs offers at
least one class in mine ventilation as it is required to
have a ventilation course in order to receive accredi-
tation from the Accreditation Board for Engineering
and Technology (ABET). However, only eleven of
these schools provide a laboratory section with the
class lecture. Ten of the programs offer possibilities
for graduate work in ventilation engineering. Among
these programs, there are only three that offer regular
graduate level ventilation courses.
In addition to the apparent limitations in ventilation
courses, the ventilation expertise among university
professors is another area of concern. Only six of the
mining programs in the U.S. have a professor on staff
that would be considered a specialist in the field of
mine ventilation. To qualify as a specialist, a professor
should have a PhD in the field and/or have practical
mine ventilation experience. In addition, the profes-
sor should have published papers in the subject and be
wholly responsible for the ventilation courses being
taught at the university.

623

Copyright © 2006 Taylor & Francis Group plc, London, UK

 Table 1. Review of mining engineering curriculums.

Offers Offers Graduate Dedicated Near retirement

ventilation ventilation ventilation ventilation age ventilation
University course(s) laboratory courses professor professor

1 x x x x
2 x x x
3 x x
4 x x x
5 x x x x
6 x x x x
7 x
8 x x x x
9 x
10 x
11 x x
12 x x x
13 x x
14 x x x
Totals 14 11 3 6 3

Of the six mine ventilation professors, at least three
are approaching retirement age. This is a concern as
it does not bode well for the future of mine ventila-
tion education in the United States. The entire subject
of mine ventilation is in jeopardy of losing the talent
necessary to train engineers in the field.
3 MINE VENTILATION PRACTICE AND
EDUCATION
Many mining companies in the United States do not
have engineers employed solely to be in charge of
mine ventilation. Furthermore, the majority of the
engineers in mining companies who are responsible
for the ventilation systems at the mines often do not
remain responsible for the ventilation systems for a
long time. Therefore, there is often a general lack of
experience within mining companies that is available
to teach younger engineers ventilation practices. This
is one of the primary reasons why it is imperative for
mining engineering students to receive adequate venti-
lation education in college. Ideally, a newly graduated
engineer in the mining industry should have enough
knowledge to be capable of understanding a ventila-
tion system, perform a basic survey, establish a model,
and use the model to project ventilation demands at a
mine. In addition, the recent graduate should have a
basic understanding of certain thermodynamic pro-
cesses in order to calculate natural ventilation and
other environmental factors.
One of the authors recently graduated with a Bach-
elor’s of Science degree in Mining Engineering and
is now a practicing ventilation engineer. His experi-
ence is probably not unlike most recent graduates. In
reviewing his educational background he noted that the
most useful subjects taught in mine ventilation related
to laboratory work and practical application of the-
ory. He found classroom lectures in ventilation survey
techniques inadequate when compared to the actual
survey requirements. In addition he felt that projects
that tied together a ventilation survey with data reduc-
tion and calculations, creating schematics, applying
fan information and using computer software to solve
the networks observations would have been a valuable
asset.
On the job training has greatly assisted the young
engineer in understanding the field of mine ventilation
engineering. However, what happens when a young
engineer does not have the assistance of an experienced
ventilation engineer at a particular mine? The risk is
high of errors causing an impact on the health and
safety of the work force or increased operating and
capital costs to the company caused by inappropriate
ventilation designs.
4 RECOMMENDATIONS
If underground mining in the United States is to flour-
ish, then it is essential that the health and safety of the
work force be ensured. To a large extent the responsi-
bility of maintaining a safe mine falls on the engineer
responsible for the ventilation system. Therefore, it
is essential that engineers receive the training neces-
sary to properly engineer the ventilation system. This
requirement should not be left to the mining compa-
nies to meet with an on the job training program. The
universities must be capable of effectively teaching the
subject.
It is becoming increasingly obvious to the authors
that among U.S. mining engineering universities,

624

Copyright © 2006 Taylor & Francis Group plc, London, UK

 proper ventilation training may not be occurring.
Unless the mining programs make a concerted effort
to hire professors who have specialized in ventilation
the problem will continue.
One suggestion to mitigate this trend is to ensure a
more uniformed educational system. This would mean
that a standardized course in the field of ventilation be
taught at all accredited mining engineering programs.
Such a course would require the input from each of
the mining curriculums and from the mining industry.
In addition, input should be provided by government
agencies such as the National Institute of Occupational
Safety and Health and the Mine Safety and Health
Administration.
A series of standardized course materials could be
developed by these stakeholders. It is envisioned that
this material would comprise of the minimum required
information for a graduating engineer. Each mining
program could add to the material to reflect specific
goals of their program.
Without standardizing the curriculums, and with the
probability of losing a number of senior ventilation
professors, the U.S. is at risk of graduating engineers
without the knowledge necessary to ensure a proper
and safe mine ventilation system. This is an injustice
to the mining companies that rely on these programs
for their engineering talent.
5 SUMMARY
The authors believe that the United States is in real
jeopardy of graduating mining engineers with little
625
Copyright © 2006 Taylor & Francis Group plc, London, UK
mine ventilation expertise. A mining engineering stu-
dent must have sufficient knowledge to know how to
perform a complete quantitative pressure and airflow
survey in a mine and then be able to solve a ventilation
network. Mine ventilation planning, fan performance,
thermodynamics, and health and safety issues also
need to be part of the curriculum.
Many curriculums offer only one undergraduate
class in ventilation and several programs do not offer
a ventilation laboratory class. This observation rein-
forces the thought that ventilation may not be given
sufficient attention at universities. Of the fourteen
programs reviewed, the authors concluded that six
have professors that would be considered primary ven-
tilation instructors. Of these six, three are nearing
retirement age. Universities need to consider hiring
more professors with ventilation experience if this
trend is to be reversed.
It is also suggested that universities consider
standardizing their mine ventilation education by con-
forming to a minimum educational standard. By estab-
lishing this type of program, universities that do not
have an expert in mine ventilation could at least grad-
uate engineers with a minimum acceptable standard.
This standard could be developed not necessarily by
an accreditation committee, but rather by an industry,
government and university consortium.
Unless mining universities are willing to change
the way mine ventilation is taught, and hire professors
with more ventilation expertise, the education of mine
ventilation engineering, is at risk of being a footnote
at many mining curriculums.