Saimm 201406 jun by SAIMM

VOLUME 114

NO. 6

JUNE 2014

For eighty years, Mintek has been a global leader in minerals and metallurgical research and development

TODAY THIS CENTRE OF TECHNOLOGICAL EXCELLENCE, with its teams of highly trained and experienced scientists, engineers, researchers and specialists, continues to develop and provide DGYDQFHG WHFKQRORJ\ IRU WKH PRUH HIIHFWLYH H[WUDFWLRQ XWLOLVDWLRQ DQG EHQHĂ&#x20AC;FLDWLRQ RI RXU mineral wealth. Mintek provides expertise in all commodities and PHWDOOXUJLFDO Ă&#x20AC;HOGV ` One of the worldâ&#x20AC;&#x2122;s longest-running and largest minerals technology facilities â&#x20AC;&#x201D; based in Johannesburg, South Africa. ` Metallurgical support throughout all stages of project development. ` â&#x20AC;&#x153;One-stop shopâ&#x20AC;? service â&#x20AC;&#x201D; spanning all metallurgical disciplines. ` Unique, proprietary technologies and products.

` 250 engineers and scientists, and 500 support staff. ` Very large integrated test and pilot plant facilities. ` Comprehensive accredited analytical and mineralogical services. ` Extensive knowledge database of Southern African and worldwide mineral resources. ` 80-year track record in providing quality work.

Celebrating 80 years of excellence in mineral and metallurgical innovation. 200 Malibongwe Drive, Randburg, South Africa. Private Bag X3015, Randburg 2125, South Africa. Phone: +27 (011) 709 4111 Fax: +27 (011) 709 4326 E-mail: info@mintek.co.za http:// www.mintek.co.za

The Southern African Institute of Mining and Metallurgy OFFICE BEARERS AND COUNCIL FOR THE 2013/2014 SESSION Honorary President Mark Cutifani President, Chamber of Mines of South Africa Honorary Vice-Presidents Susan Shabangu Minister of Mineral Resources, South Africa Rob Davies Minister of Trade and Industry, South Africa Derek Hanekom Minister of Science and Technology, South Africa President M. Dworzanowski President Elect J.L. Porter Vice-Presidents R.T. Jones C. Musingwini Immediate Past President G.L. Smith Honorary Treasurer J.L. Porter Ordinary Members on Council H. Bartlett N.G.C. Blackham V.G. Duke M.F. Handley W. Joughin A.S. Macfarlane D.D. Munro

S. Ndlovu G. Njowa S. Rupprecht A.G. Smith M.H. Solomon D. Tudor D.J. van Niekerk

Past Presidents Serving on Council N.A. Barcza R.D. Beck J.A. Cruise J.R. Dixon F.M.G. Egerton A.M. Garbers-Craig G.V.R. Landman

R.P. Mohring J.C. Ngoma R.G.B. Pickering S.J. Ramokgopa M.H. Rogers J.N. van der Merwe W.H. van Niekerk

Branch Chairmen DRC

S. Maleba

Johannesburg

I. Ashmole

Namibia

G. Ockhuizen

Pretoria

N. Naude

Western Cape

T. Ojumu

Zambia

H. Zimba

Zimbabwe

S.A. Gaihai

Zululand

C. Mienie

PAST PRESIDENTS *Deceased * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

W. Bettel (1894–1895) A.F. Crosse (1895–1896) W.R. Feldtmann (1896–1897) C. Butters (1897–1898) J. Loevy (1898–1899) J.R. Williams (1899–1903) S.H. Pearce (1903–1904) W.A. Caldecott (1904–1905) W. Cullen (1905–1906) E.H. Johnson (1906–1907) J. Yates (1907–1908) R.G. Bevington (1908–1909) A. McA. Johnston (1909–1910) J. Moir (1910–1911) C.B. Saner (1911–1912) W.R. Dowling (1912–1913) A. Richardson (1913–1914) G.H. Stanley (1914–1915) J.E. Thomas (1915–1916) J.A. Wilkinson (1916–1917) G. Hildick-Smith (1917–1918) H.S. Meyer (1918–1919) J. Gray (1919–1920) J. Chilton (1920–1921) F. Wartenweiler (1921–1922) G.A. Watermeyer (1922–1923) F.W. Watson (1923–1924) C.J. Gray (1924–1925) H.A. White (1925–1926) H.R. Adam (1926–1927) Sir Robert Kotze (1927–1928) J.A. Woodburn (1928–1929) H. Pirow (1929–1930) J. Henderson (1930–1931) A. King (1931–1932) V. Nimmo-Dewar (1932–1933) P.N. Lategan (1933–1934) E.C. Ranson (1934–1935) R.A. Flugge-De-Smidt (1935–1936) T.K. Prentice (1936–1937) R.S.G. Stokes (1937–1938) P.E. Hall (1938–1939) E.H.A. Joseph (1939–1940) J.H. Dobson (1940–1941) Theo Meyer (1941–1942) John V. Muller (1942–1943) C. Biccard Jeppe (1943–1944) P.J. Louis Bok (1944–1945) J.T. McIntyre (1945–1946) M. Falcon (1946–1947) A. Clemens (1947–1948) F.G. Hill (1948–1949) O.A.E. Jackson (1949–1950) W.E. Gooday (1950–1951) C.J. Irving (1951–1952) D.D. Stitt (1952–1953) M.C.G. Meyer (1953–1954)

* * * * * * * * * * * * * * * * * * * * * * * *

L.A. Bushell (1954–1955) H. Britten (1955–1956) Wm. Bleloch (1956–1957) H. Simon (1957–1958) M. Barcza (1958–1959) R.J. Adamson (1959–1960) W.S. Findlay (1960–1961) D.G. Maxwell (1961–1962) J. de V. Lambrechts (1962–1963) J.F. Reid (1963–1964) D.M. Jamieson (1964–1965) H.E. Cross (1965–1966) D. Gordon Jones (1966–1967) P. Lambooy (1967–1968) R.C.J. Goode (1968–1969) J.K.E. Douglas (1969–1970) V.C. Robinson (1970–1971) D.D. Howat (1971–1972) J.P. Hugo (1972–1973) P.W.J. van Rensburg (1973–1974) R.P. Plewman (1974–1975) R.E. Robinson (1975–1976) M.D.G. Salamon (1976–1977) P.A. Von Wielligh (1977–1978) M.G. Atmore (1978–1979) D.A. Viljoen (1979–1980) P.R. Jochens (1980–1981) G.Y. Nisbet (1981–1982) A.N. Brown (1982–1983) R.P. King (1983–1984) J.D. Austin (1984–1985) H.E. James (1985–1986) H. Wagner (1986–1987) B.C. Alberts (1987–1988) C.E. Fivaz (1988–1989) O.K.H. Steffen (1989–1990) H.G. Mosenthal (1990–1991) R.D. Beck (1991–1992) J.P. Hoffman (1992–1993) H. Scott-Russell (1993–1994) J.A. Cruise (1994–1995) D.A.J. Ross-Watt (1995–1996) N.A. Barcza (1996–1997) R.P. Mohring (1997–1998) J.R. Dixon (1998–1999) M.H. Rogers (1999–2000) L.A. Cramer (2000–2001) A.A.B. Douglas (2001–2002) S.J. Ramokgopa (2002-2003) T.R. Stacey (2003–2004) F.M.G. Egerton (2004–2005) W.H. van Niekerk (2005–2006) R.P.H. Willis (2006–2007) R.G.B. Pickering (2007–2008) A.M. Garbers-Craig (2008–2009) J.C. Ngoma (2009–2010) G.V.R. Landman (2010–2011) J.N. van der Merwe (2011–2012)

Honorary Legal Advisers Van Hulsteyns Attorneys

Corresponding Members of Council Australia:

I.J. Corrans, R.J. Dippenaar, A. Croll, C. Workman-Davies

Auditors Messrs R.H. Kitching

Austria:

H. Wagner

Secretaries

Botswana:

S.D. Williams

Brazil:

F.M.C. da Cruz Vieira

China:

R. Oppermann

The Southern African Institute of Mining and Metallurgy Fifth Floor, Chamber of Mines Building 5 Hollard Street, Johannesburg 2001 P.O. Box 61127, Marshalltown 2107 Telephone (011) 834-1273/7 Fax (011) 838-5923 or (011) 833-8156 E-mail: journal@saimm.co.za

United Kingdom: J.J.L. Cilliers, N.A. Barcza, H. Potgieter USA:

J-M.M. Rendu, P.C. Pistorius

Zambia:

J.A. van Huyssteen

▲

JUNE 2014

The Journal of The Southern African Institute of Mining and Metallurgy

Editorial Board

Editorial Consultant D. Tudor

Typeset and Published by The Southern African Institute of Mining and Metallurgy P.O. Box 61127 Marshalltown 2107 Telephone (011) 834-1273/7 Fax (011) 838-5923 E-mail: journal@saimm.co.za

Printed by Camera Press, Johannesburg

Advertising Representative Barbara Spence Avenue Advertising Telephone (011) 463-7940 E-mail: barbara@avenue.co.za The Secretariat The Southern African Institute of Mining and Metallurgy ISSN 2225-6253

THE INSTITUTE, AS A BODY, IS NOT RESPONSIBLE FOR THE STATEMENTS AND OPINIONS A DVA NCED IN A NY OF ITS PUBLICATIONS. Copyright© 1978 by The Southern African Institute of Mining and Metallurgy. All rights reserved. Multiple copying of the contents of this publication or parts thereof without permission is in breach of copyright, but permission is hereby given for the copying of titles and abstracts of papers and names of authors. Permission to copy illustrations and short extracts from the text of individual contributions is usually given upon written application to the Institute, provided that the source (and where appropriate, the copyright) is acknowledged. Apart from any fair dealing for the purposes of review or criticism under The Copyright Act no. 98, 1978, Section 12, of the Republic of South Africa, a single copy of an article may be supplied by a library for the purposes of research or private study. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means without the prior permission of the publishers. Multiple copying of the contents of the publication without permission is always illegal. U.S. Copyright Law applicable to users In the U.S.A. The appearance of the statement of copyright at the bottom of the first page of an article appearing in this journal indicates that the copyright holder consents to the making of copies of the article for personal or internal use. This consent is given on condition that the copier pays the stated fee for each copy of a paper beyond that permitted by Section 107 or 108 of the U.S. Copyright Law. The fee is to be paid through the Copyright Clearance Center, Inc., Operations Center, P.O. Box 765, Schenectady, New York 12301, U.S.A. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale.

VOLUME 114

NO. 6

JUNE 2014

Contents Journal Comment—The increasing role of computers in engineering by R. Paul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . President’s Corner by M. Dworzanowski . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

iv-v vii

Special Articles The Danie Krige Memorial Lecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SANCOT News by H.J. Tluczek . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Opening of Mining Exhibition at Sci-Bono by C. Kotze . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Joburg Indaba by A. Spratley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v vi viii ix

General papers A computational fluid dynamics model for investigating air flow patterns in underground coal mine sections by D.D. Ndenguma, J. Dirker, and N.D.L. Burger. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

419

A heuristic sublevel stope optimizer with multiple raises by X. Bai, D. Marcotte, and R. Simon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Risk indexing tool for mine planning by W. Abdellah, H.S. Mitri, D. Thibodeau, and L. Moreau-Verlaan . . . . . . . . . . . . . . . . . . . . . . Dealing with open fire in an underground coal mine by ventilation control techniques by N. Sahay, A. Sinha, B. Haribabu, and P.K. Roychoudhary . . . . . . . . . . . . . . . . . . . . . . . . . . Empirical observations of dilution in panel caving by R.L. Castro and P.S. Paredes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Design principles for optimizing an established survey slope monitoring system by N. Mphathiwa and F.T. Cawood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Estimation of rock strength from quantitative assessment of rock texture by C.A. Ozturk, E. Nasuf, and S. Kahraman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A new process for the recovery of iron, vanadium, and titanium from vanadium titanomagnetite by S.Y. Chen and M.S. Chu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A methodology for determining the erosion profile of the freeze lining in submerged arc furnace by H. Dong, H.-J. Wang, and S.-J. Chu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

427 435 445 455 463 471

481

489

Life-of-mine ventilation and refrigeration planning for Resolution Copper Mine by S. Bluhm, R. Moreby, F. von Glehn, and C. Pascoe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

497

Invitation to the SAIMM members. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

496

VOLUME 114

NO. 6

JUNE 2014

International Advisory Board R. Dimitrakopoulos, McGill University, Canada D. Dreisinger, University of British Columbia, Canada E. Esterhuizen, NIOSH Research Organization, USA H. Mitri, McGill University, Canada M.J. Nicol, Murdoch University, Australia H. Potgieter, Manchester Metropolitan University, United Kingdom E. Topal, Curtin University, Australia

The Journal of The Southern African Institute of Mining and Metallurgy

JUNE 2014

iii

▲

R.D. Beck J. Beukes P. den Hoed M. Dworzanowski M.F. Handley R.T. Jones W.C. Joughin J.A. Luckmann C. Musingwini R.E. Robinson T.R. Stacey R.J. Stewart

Journal Comment The increasing role of computers in engineering

â&#x2013;˛

JUNE 2014

The Journal of The Southern African Institute of Mining and Metallurgy

JUNE 2014

â&#x2013;˛

Journal Comment (continued)

ews N T O C N SA

SANCOT and the International Tunnelling Association (ITA) On the resignation of Andrew Wilson as the outgoing Chairman of SANCOT, Ron Tluczek was elected as the new SANCOT Chairman on 9 April 2014. Ron is a Professional (Geotechnical) Engineer with more than 36 years’ experience, primarily in consulting engineering. He has experience in all aspects of project management and administration of civil and geotechnical related projects, including investigation, field and laboratory testing, analysis, risk assessment, reporting, contract documentation, monitoring, quality assurance, supervision, and budget control. He has acted as an expert witness, as review consultant, and is involved with claims evaluation. He is currently an executive with AECOM SA (Pty) Ltd and is responsible for all tunnelling and geotechnical work in Africa. Ron is a member of the Association of Arbitrators and the Institute of Waste Management. He is also a member of Working Group 2 (Research) with the International Tunnelling Association (ITA). He represented South Africa at the ITA General Assembly which was recently held in Iguassu, Brazil. Ron was born in Broken Hill (now called Kabwe) in Northern Rhodesia (now Zambia), where he attended junior school. His family later relocated to Bulawayo in Southern Rhodesia (now Zimbabwe) where he attended Gifford Technical High. At high school he accumulated seven ‘A’ Level subjects, a record which he still holds today. On leaving high school, Ron studied civil engineering at the University of Cape Town where he graduated in 1975, after which he did his military training in what was then the Rhodesian army. Ron’s first job was on the Du Toitskloof Pilot Bore Tunnel where he found his passion for rock mechanics, and specifically tunnelling. He has since worked on some 16 tunnelling projects, some of the major underground projects including the following: • The Richards Bay Coal Line, where Ron was responsible for the site investigation, rock support design, and writing relevant specifications for two single-track and one double-track railway tunnel and approach works • The Palmiet Pumped Storage Scheme for ESKOM near Grabouw in the Western Cape. Ron was instrumental in the site investigation, design, tender documentation, and site supervision for the underground works, including tunnels and shafts • The Lesotho Highlands Water Scheme, where he was initially involved with the site investigation, design, and contract documentation for the delivery tunnels. Later, Ron was involved on-site with the rock support assessment for the galleries, footprint, and appurtenant works for Katse Dam • The Gautrain Rapid Rail Link in Gauteng, which comprised the construction of 80 km of new rail link, 10 stations (four of which are underground), 15 km of tunnelling (3 km of which were excavated by means of tunnel boring machine), and 20 km of viaducts (the majority over dolomite / karst geology). Ron was a member of the Technical Advisory Team from inception and was responsible for all geotechnical and tunnelling aspects. He still acts in an advisory capacity to the Gautrain Management Agency. Ron is the author and co-author on numerous locally and internationally published papers on tunnel design, pumped storage schemes, and rock support in general. Ron is married to Merryn and they have two daughters, Heather and Meryl.

H.J. Tluczek

▲

JUNE 2014

The Journal of The Southern African Institute of Mining and Metallurgy

JUNE 2014

vii

▲

entʼs d i s e Pr er Corn

Opening of Mining exhibition at Sci-Bono

peaking at the official opening of a permanent mining exhibition at the Sci-Bono Discovery Centre in Newtown, Johannesburg, in April, Marek Dworzanowski, President of the Southern African Institute of Mining and Metallurgy (SAIMM) said that ‘the mining industry in South Africa still has a bright future ahead of it’, and pointed out that part of the SAIMM’s role was to encourage learners to consider a career in the mining industry. To ensure continuity within the mining industry, the Institute aims to establish a pipeline of mining and metallurgy graduates, and Dworzanowski believes that the exhibition is one of the ways people can be made aware of the opportunities offered by the industry. At the opening ceremony which attracted 35 attendees, Sci-Bono CEO David Kramer stressed that the SAIMM-sponsored mining exhibition aims not only to provide learners with new insights into the mining industry, but also to educate the general public on the role of mining and the way in which it influenced their daily lives. The role of the exhibition, while improving people’s understanding, is also to encourage enthusiasm about mining. ‘We need to attract talented young men and women into mining by helping them understand the opportunities that the mining industry offers,’ he said. Kramer used the occasion to appeal to educators in the mining industry to provide their input on how to expand the exhibition and create a mobile mining exhibition that could, in future, be taken into Gauteng’s schools. Dworzanowski says SAIMM’s involvement in the establishment of the exhibition was in line with its commitment to ensure that the industry is able to assist and guide learners interested in entering, and those who have already entered the miningsector. The exhibits, which are interactive and were built in-house by Sci-Bono, include a 4-ton tyre from an off-highway earthmoving vehicle; an interactive periodic table with images and descriptions; a map of the mining regions in South Africa highlighting where specific metals and minerals are found; and an display featuring mineral and metal manufactured products with descriptions of the mineral or metal. There is an interactive wall with touch-screen televisions that show videos covering the entire mining process, from ore extraction to beneficiation. Mannequins clad in personal protective equipment and a re-creation of an underground mining environment are also on display, along with a core table that features drill cores from various mineral and metal deposit, as well as a percussion drill. Sponsors of the exhibition include the Chamber of Mines of South Africa, the Council for Geoscience, the Gold Reef City theme park, computer equipment manufacturer Dell, information technology infra structure consulting firm Smart Computer Solutions, earthmoving equipment dealer Barloworld Equipment, mine support systems company New Concept Mining, resource and infrastructure-focused engineering project house TWP Projects, electrical contractor Dow’s Electrical & Lighting Supplies, mining equipment manufacturer Joy Global, and miners Impala Platinum, AngloGold Ashanti, Kumba Iron Ore, and Zincor. Kramer says the exhibit is not yet complete, as further displays will be added over the next few years with the aim of turning the mining exhibition into a visitor attraction for Sci-Bono and the City of Johannesburg. This contribution was written by Chantelle Kotze of Creamer Media's Mining Weekly, and first printed in Mining Weekly on 18 April 2014.

C. Kotze

The Joburg Indaba Mark Cutifani, Chief Executive of Anglo American joins an esteemed group of mining and investor heavyweights as the keynote speaker at the 2014 Joburg Indaba in October.

utifani, with more than 38 years in the mining industry, has working experience in 25 operating environments and with 20 commodities and is widely recognized as an industry innovator. The Joburg Indaba is aimed at building a competitive, sustainable, and legitimate mining industry through strategic conversations. ‘Mark’s broad industry, technical, and commercial knowledge is a prime example of the calibre of speakers joining the critical conversations at this year’s event’ states Bernard Swanepoel, moderator at the 2014 Indaba. Opportunities and solutions that could outline the route to a competitive and sustainable mining industry will be thrashed out at this year’s event amongst various leaders, game-changers, and entrepreneurs in both the mining and investment sectors. Robert Friedland will be back again to join colleagues in thought-provoking discussions alongside Graham Briggs, Valli Moosa, Mike Teke, Noah Greenhill, Peter Leon, Michael Spicer, and many others. If anything, discussions from the inaugural event were open, honest, and outright blunt. The vast majority (91%) of attendees citied that a strong, supportive, and visible political leadership of the mining sector was vital, while concerns were voiced on current regulatory issues, and the dire need for local government to clearly state its position on mining as SA’s economic driver. Special focus will be given at this year’s event to the labour issues that have crippled the industry and the challenging trade-offs between cost, productivity, and unemployment, highlighted by the mechanization versus labour debate. The majority of the attendees recognized that the industry needs to fundamentally change direction in order to prosper. To join the conversation on 8-10 October 2014, visit www.joburgindaba.com <http://www.joburgindaba.com>

A. Spratley Cutifani joins mining and investor heavyweights for some critical conversations to be held at the 2014 Joburg Indaba

Trusted screen media solutions since 1905 No matter what your screen media requirements, FLSmidth offers the most comprehensive and innovative range of screen surface media in the industry! Our ﬁeld-proven and time-tested Meshcape brand has dominated the South African market since 1905 and continues to increase in international popularity since joining the FLSmidth family of brands back in 2012. We offer a full line of standard, modular and customizable screen media including woven wire screens, KASON (Sweco) machines and spares, expanded metal, perforated materials, polyurethane products, wedge wire, welded mesh and wire conveyor belts. If you’re looking for durability, versatility and superior product performance in media recovery, then look to Meschape branded screen media from FLSmidth. For more information, visit us at www.ﬂsmidth.com 8JUIJO 4PVUI "GSJDB t &EFOWBMF $BQF 5PXO #SBODI t %VSCBO #SBODI 1& #SBODI t ,JNCFSMZ #SBODI

SAFETY AND PRODUCTIVITY Collision Avoidance Systems

Safety Access Systems

Safety Isolation Boxes

Corner Hans Strydom & Van Rensburg Street Witbank, Gauteng Tel: +27 (0) 13 692 4161 Fax: +27 (0) 13 692 3337 www.probeIMT.co.za

Outotec provides leading technologies and services for the sustainable use of Earthâ&#x20AC;&#x2122;s natural resources. As the global leader in minerals and metals processing technology, we have developed many breakthrough technologies over the decades for our customers in metals and mining industry. We also provide innovative solutions for industrial water treatment, the utilization of alternative energy sources and the chemical industry. www.outotec.com

For the whole story, please visit our YouTube channel.

MINING

...Safety and productivity through accuracy and reliability...

Reutech Mining is an international company with a global

We pride ourselves on our customer support philosophy, which

presence in all the major mining regions. The company has more

offers our customers continuous software updates, remote

than 25 years experience in radar development and

monitoring of system vital statistics and unmatched

manufacturing. Over the past decade we have established

maintenance support. This provides our customers with peace of

ourselves as specialists in producing world leading, state-of-the-

mind that high-risk areas are constantly monitored.

art technology solutions that has become indispensable in modernized profit driven mining operations.

Seamlessly integrating our unique technology with established systems and methodologies enables our customers to improve

Our fully customized user-solutions significantly decrease slope

various aspects of their current operations. In addition to this,

stability risk factors in open cast mines and improve productivity

valuable data and experience are constantly generated allowing

by offering around-the-clock monitoring of various elements of

users to incorporate this into their future planning.

the mining operation. By offering a superior value proposition, Reutech Mining has The MSR (Movement and Surveying Radar) outperforms its

positioned itself as a market leader, offering products and

competitors in average global Operational Availability â&#x20AC;&#x201C;

services considered to be cornerstones of successful mining

achieving more than 99% uptime in the harshest of operating

operations in the modern world.

conditions.

www.reutechmining.com

MSR

PAPERS IN THIS EDITION These papers have been refereed and edited according to internationally accepted standards and are accredited for rating purposes by the South African Department of Higher Education and Training

General papers by D.D. Ndenguma, J. Dirker, and N.D.L. Burger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419 The results from an experimental and numerical study of air flow in a scaled-down underground mine model were compared to determine if numerical analysis can be utilized to identify ventilation solutions for controlling high concentration of coal dust and methane in underground coal mines. The agreement between the experimental and numerical results indicates that numerical modelling can be a useful tool in this regard. by X. Bai, D. Marcotte, and R. Simon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427 A new heuristic sublevel mining stope optimizer is presented which seeks the best locations and lengths of a series of vertical raises that, together with the blocks linked to each raise, define a mining stope. Two synthetic cases and one real deposit are used to evaluate the new algorithm and compare the results with the single-raise optimizer. The multiple raises approach leads to improved economics as well as a reduction in dilution. by W. Abdellah, H.S. Mitri, D. Thibodeau, and L. Moreau-Verlaan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435 This paper presents a qualitative method to estimate the risk level that results from mining activity. An assessment scale matrix, based on probability of failure and cost of consequences, is used to assign a risk index value that is directly proportional to the potential for excavation instability. The method is illustrated by means of a case study at an underground base metal mine. by N. Sahay, A. Sinha, B. Haribabu, and P.K. Roychoudhary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445 A methodology has been developed for dealing with advanced-stage open fires in underground coal mines by means of a modified ventilation control technique. The methodology is based on a better understanding of the behaviour of an open fire, proper diagnosis of the problem, application of judicious ventilation control techniques, and selection of suitable fire indices for assessing the status of the fire. The method was used successfully to control a fire in an underground mine in central India by R.L. Castro and P.S. Paredes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455 This paper presents a back analysis of extraction and dilution behaviour at two of Codelcoâ&#x20AC;&#x2122;s operations where panel caving is used as the mining method. Three sources of dilution were identified from the analysis. by N. Mphathiwa and F.T. Cawood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463 This paper provides guidelines on designing an optimal survey slope monitoring system. An open pit diamond mine is used as a case study. The design strategy can be used as a guideline for developing a new slope monitoring system or to optimize an existing one.

These papers will be available on the SAIMM website

http://www.saimm.co.za

General papers (continued) Estimation of rock strength from quantitative assessment of rock texture by C.A. Ozturk, E. Nasuf, and S. Kahraman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471 A method is presented for quantifying rock texture in order to estimate compressive strength. A texture coefficient (TC) is determined from a statistical assessment of images of thin sections, and correlations derived between the TC and the experimentally determined value of compressive strength. The results show that classification of the intact rock based on lithology increases the reliability of the prediction model derived from regression analysis. A new process for the recovery of iron, vanadium, and titanium from vanadium titanomagnetite by S.Y. Chen and M.S. Chu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481 A new process comprising metallizing reduction, magnetic separation, and electrothermal melting separation is proposed for the recovery of iron, vanadium, and titanium from vanadium titano-magnetite. The optimal process parameters were established, which led to recoveries of up to 80% for titanium, 95% for iron, and 71% for vanadium. The reaction mechanisms were elucidated by means of SEM and EDS analysis. A methodology for determining the erosion profile of the freeze lining in submerged arc furnace by H. Dong, H.-J. Wang, and S.-J. Chu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489 This paper presents a heat conduction model to determine the erosion profile of the freeze lining in a ferronickel submerged arc furnace. The model is validated by comparing the heat loss through the lining with the value calculated from industrial data. The optimal solution for the erosion profile can be acquired when the difference in heat losses calculated by these two methods is at a minimum. Life-of-mine ventilation and refrigeration planning for Resolution Copper Mine by S. Bluhm, R. Moreby, F. von Glehn, and C. Pascoe. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497 This paper is reprinted here by kind permission of the AusIMM

These papers will be available on the SAIMM website

http://www.saimm.co.za

A computational fluid dynamics model for investigating air flow patterns in underground coal mine sections by D.D. Ndenguma*, J. Dirker*, and N.D.L. Burger*

This investigation compares the results from an experimental and numerical study of air flow in a scaled-down underground coal mine model. This was done in order to determine if numerical analysis can be relied upon when searching for ventilation solutions to control high concentration of coal dust and methane gas in underground coal mines. Steady state analyses were used to identify flow patterns and recirculation regions within the mining section while transient state analyses were used to determine the time taken to extract the initial air from the model. The agreement between the experimental and numerical results indicates that numerical modelling is useful in this regard. The study went further to devise a method of determining the optimum position of the jet fan that is responsible for mine-gas dilution at different stages of mining. Keywords numerical modelling, computation fluid dynamics, CFD, mine ventilation.

Introduction Coal is the most widely used primary fuel, accounting for about 36% of the total fuel consumption of the world’s electricity production. According to Thopil and Pouris (2010), 68% of South Africa’s primary energy needs are provided by coal. Coal can be extracted either by opencast or underground operations. Approximately 50% of South Africa’s coal production is by underground mining. In underground coal mines different extraction methods are used depending upon the geological formations and other factors. The two major extraction methods are the bord-and-pillar and the longwall. Bord-andpillar mining is the most common type of underground coal mining globally (Peng and Chiang, 1984). Cutting operations are performed using a continuous miner (CM) that makes extended cuts, known as box cuts, into the mining face. To complete a cutting operation the CM makes four cuts. In some South African coal mines these cuts may each be 17.5 m long and 3.5 m wide. Figure 1 shows a plan view of an underground bordand-pillar layout. The final depth of the inheading might be up to 35 m, the width up to 7 m, and the height depends on the thickness of the coal seam. The Journal of The Southern African Institute of Mining and Metallurgy

* Department of Mechenical and Aeronautical Engineering, University of Pretoria, Pretoria, South Africa. © The Southern African Institute of Mining and Metallurgy, 2014. ISSN 2225-6253. Paper received r Jun. 2011; revised paper received Jan. 2014. VOLUME 114

JUNE 2014

419

▲

Synopsis

The extraction off coal, especially by underground mining, is associated with a number of hazards. Airborne dust and methane gas are a risk to the miners and mining equipment. High concentrations of respirable coal dust are known to be a cause of lung disease (Chander et al., 1990; Cheng and Zukovich, 1973), and should be avoided. Methane gas and coal dust can cause explosions that result in injuries and fatalities. In South Africa, since the first coal mine explosion at Elandslaagte colliery in 1891, a further 333 explosions resulted in 1 034 deaths by the end of 1990 (Flint, 1990). According to Nundlall (1996) 78 flammable gas explosions, some of which were in coal mines, occurred in South Africa between 1988 and 2005. Liu et al. (2010) conducted an analysis of coal mine accidents in China from 2005 to 2009. More than 3000 gas accidents were reported in this period. Several attempts have been made, with success, to improve coal mine safety. Esterhuizen and Gürtunca (2006) outlined the positive impact that research can have on reducing coal mining accidents. They reported improvements with regards to the reduction of respiratory diseases, noiseinduced hearing loss, musculoskeletal disorders, traumatic injuries, mine disasters (including reducing and monitoring the risk of flammable atmosphere ignition), and fatalities and injuries due to ground failure. It was shown that a significant reduction in coal mining fatalities has been achieved by implementing a wide range of strategies and solutions. Since coal dust and methane gas cause coal mine accidents, their concentrations must be kept low. The Department of Mineral Resources (DMR) in South Africa issued a

A computational fluid dynamics model for investigating air flow patterns

Figure 1—A plan view of underground coal mine layout during mining

directive that the dust and methane concentration in underground coal mines should not exceed 2.0 mg/m³ and 0.5% by volume, respectively. Ventilation is a popular technique that is used to control dust and methane gas in underground coal mines to allowable concentrations. A jet fan is frequently employed as a source of air flow to ventilate the in-heading. A higher air speed may, however, increase airborne coal dust concentrations and increased ventilation rates should therefore be employed with care. A scrubber that is mounted on the CM also plays a big role in reducing dust and methane gas, since if operating properly it will remove particles from the air stream flowing over the machine. Due to tight production schedules and harsh environmental conditions, it is difficult and costly to carry out experiments in operational coal mines. Therefore, this investigation was done using a comparatively inexpensive, and safer, computational fluid dynamics (CFD) modelling technique. This approach is aimed at aiding the search for solutions to problems associated with airborne dust and high methane gas concentrations in underground coal mines. CFD techniques have been used in the past for mining ventilation studies, but experimental validation is needed (Wala et al., 2007). In this investigation a scaled-down model of an underground coal mine was used to validate the CFD model, which allows predictions using the CFD model to be extended to full-scale applications. Steady-state and transient-state investigations were done numerically and experimentally for a scaled-down underground coal mine model. The experimental results were used to validate the numerical analysis. Further numerical analysis was carried out in order to develop a method for determining the optimum position range of the jet fan at different lengths of in-heading.

improve smoke visibility. Important dimensions off the model in-heading were as follows: 2 333 mm deep, 420 mm wide, and 333 mm high. These were representative of an actual mine section, which in some cases may have a depth of 35 m, width of 7 m, and a height of 5 m (see also Figure 1). Air flow rates were scaled using a percentage volume flow method to resemble conditions in full-scale mining sections during mining. In this method, the number of air changes per hour for the scaled model was the same as for the general full-scale application. Other suitable non-dimensional parameters that would have led to a reasonable representation of the full-scale application could not be identified. Air flow in the through-road was induced by a fan at the tapered outlet of the through-road. This air velocity was set at 0.0665 m/s at the centre of the through-road, representing 2 m/s in an actual mine. The jet fan was represented by means of a nozzle connected to a compressed air system. The nozzle was designed to achieve a jet-stream penetration of 1.733 m (equivalent to the 26 m obtained with actual mining equipment in a full-scale simulation, as measured experimentally at the Kloppersbos testing facility in South Africa) into the in-heading with a velocity of 10 m/s at the nozzle exit. The nozzle had a 10 mm exit diameter. The air flow was regulated with a pressure control valve while the air velocity was measured by an anemometer. The flow rate of the scrubber on the continuous miner was implemented in the scaled model by means of a speed-controlled fan. This represented a common operational configuration where the filtered air is rejected directly backwards, and not towards the sidewall of the in-heading. Conditions both with and without the scrubber in operation were considered. Transient-state tests required measurement of the smoke concentration. A measuring system consisting of photoelectric sensors, a data acquisition unit, and a computer, shown schematically in Figure 3, was used. Photoelectric sensors, (shown schematically in Figure 4) were constructed to detect the intensity of light reflected from smoke particles. A higher light intensity related to a higher smoke concentration. For each sensor, a light-emitting diode (LED) was used as light source in one compartment while a phototransistor was located in the other compartment. As light was

Test facility Steady-state and transient-state experiments were conducted using a 15% scale underground coal mine model, shown in Figure 2. The intention in using a scale model was not to represent a full-scale mine, but rather to validate the CFD model. This technique can later be used to simulate a fullscale mine. All the walls of the model were made from hardboard except for the top and one side of the in-heading, which were constructed from clear Perspex to enable observation. Smoke was used to visualize air motion during steady-state tests and to represent methane gas concentration during transient tests. The inside surface of the walls and floor were black to

▲

420

JUNE 2014

VOLUME 114

Figure 2—The coal mine model scaled down to 15% that was used for the verification experiments The Journal of The Southern African Institute of Mining and Metallurgy

A computational fluid dynamics model for investigating air flow patterns

[3]

Figure 3—Block diagram of a measuring system for transient-state experiment

Figure 4—Retro sensor

reflected from smoke particles, a signal was produced proportional to the smoke concentration. The sensor structure, which was black in colour, was 12 mm wide, 6 mm deep, and 18 mm high. Eight such sensors were placed at different locations in the scale model as shown in Figure 5. Table I indicates the actual positions of the sensors with respect to the origin point (as defined in Figure 5). The signal sent by the phototransistors in form of an electrical current (I ( ) was converted to electrical resistance (R) by the data acquisition unit. Before use, all sensors were calibrated and standardized according to light intensity under controlled conditions. In this process, the read-out deviations for each sensor, compared to the averaged read-out from all sensors combined, were recorded. These deviations were used to derive corrective adjustment functions that were employed later during the experimental data-processing phase.

Numerical computation

Three geometries representing three different mining progressions were examined in the steady-state investigation, while for the transient-state investigation, only one mining progression was used. The mining progression shown in Figure 1 (scaled down) is reported on for both the steadystate and transient-state analyses. The geometry, consisting of the in-heading, a portion of the through-road, a jet fan, a scrubber, and a continuous miner, was defined using CAD software. Provision was made to define numerical sensors at the same locations as the photoelectric sensors in the real laboratory-scale model. The geometry was imported into Star CCM+ for meshing and boundary conditions definition. An automated polyhedral-type meshing scheme was used. After a meshindependence study had been conducted, a suitable mesh size was selected. Appropriate inlet and outlet conditions were defined to simulate flow at the jet fan inlet and outlet, scrubber inlet and outlet, and the through-road inlet and outlet. Pressure outlet type was used for the outlets. The air velocity of the jet fan was set at 10 m/s, the scrubber air outlet velocity at 2.3 m/s, and through-road inlet velocity at 0.0665 m/s. These velocities were scaled down from 46 m/s, 35 m/s, and 2 m/s for actual mining conditions, respectively. Standard air properties were used at a temperature of 293K. For transient-state numerical investigation, where smoke

Figure 5—Position of sensors in experimental model

Star CCM+ version 3.02.006 was used to solve the following Navier-Stokes governing equations within the model for numerical analysis.

Table I

Location of the sensors in the test section

[2]

The Journal of The Southern African Institute of Mining and Metallurgy

Sensor

S1 S2 S3 S4 S5 S6 S7 S8

Position, mm X

2280 2280 2280 2280 1167 1167 1167 150

282 282 70 70 282 282 70 210

36 174 174 36 36 384 36 166

VOLUME 114

JUNE 2014

421

▲

[1]

A computational fluid dynamics model for investigating air flow patterns concentration had to be modelled, ffresh air ffrom the jet ffan and smoke were represented by a molecular weight of 29 kg/kmol and 800 kg/kmol respectively. The properties of the smoke phase in the numerical model were chosen to match the real properties as closely as possible. Table II summarizes the CFD modelling settings used to define the physical phenomenon of steady- and transientstate analyses.

5. Once sufficient ff smoke flow f was established, image capturing was initiated 6. After capturing data at one position, the three tubes were set again at a different level or the next set of holes. Steps 3 to 5 were repeated until all the designated points in the test section had been investigated.

Comparison of steady-state experimental and numerical results

Steady-state tests In the steady-state investigation the aim was to identify air flow patterns in the in-heading. Three different mine geometries representing different mining progressions were studied, but only one geometry is presented here. For all tests the jet fan in the scale model was located 80 mm into the inheading, 67 mm from the sidewall, and 47 mm above the floor. Smoke was released from a smoke-producing device into the in-heading through plastic tubes 5 mm in diameter. These were inserted through the floor of the model at number of grid points as shown in Figure 6. The tube ends were also set at three different heights in order to observe as many points as possible. Smoke was produced by burning a glycol/water mixture in a well-ventilated laboratory. The mixture was pumped from the tank into a container where it was sprayed onto a heater element to produce smoke. The smoke was introduced into the test section through the tube network. Video clips of the flow patterns of air in the test section were recorded for analysis. Figure 7 shows a still image of smoke entering the test section. The video clips were recorded from the top and the side of test section. The experimental procedure was as follows. 1. The valve on the compressed air system was opened to induce the correct volumetric flow rate and velocity at the nozzle representing the jet fan 2. The smoke tubes were inserted into three consecutive holes along the width of the in-heading and set at same level or height at a single time. The heights that were used were 47 mm, 165 mm, and 250 mm, measured from floor level. All other holes on the floor were sealed off to avoid air leakages 3. A digital camera was set to capture data in form of video clips 4. The smoke-producing device, the through-road fan, and observation light were all switched on and the laboratory lights were switched off

The smoke direction observed from video clips was converted into vector format. In Figures 8A and 8B arrows are used to indicate the smoke direction obtained in this fashion. The arrows indicate direction only, and not velocity. Experimental results on arbitrary planes are shown here for graphical representation purposes only. Also shown are the numerical steady-state results from the top and the side view. Although

Figure 6—Positions of smoke tubes in the floor of the heading

Figure 7—Smoke entering the test section at points E1, E2, and E3 for observation of air flow direction

Table II

Physics model of continuum for steady-state and transient conditions

Space Motion Material Flow State of equation Time Viscosity regime Reynolds-averaged turbulence Optional physics model

▲

422

JUNE 2014

VOLUME 114

Steady state

Transient state

Three–dimensional Stationary Gas Segregated Constant density Steady Turbulent K-epsilon Segregated fluid temperature

Three-dimensional Stationary Multi-component gas Segregated Constant density Transient Turbulent K-epsilon Segregated fluid temperature

The Journal of The Southern African Institute of Mining and Metallurgy

A computational fluid dynamics model for investigating air flow patterns 10. The smoke extraction process was left f to run until all the smoke had been removed from the test section. This was determined when the readings from the smoke sensors stabilized 11. Logged data was stored and analysed.

Figure 8â&#x20AC;&#x201D;Plan view of air flow patterns for the experiment (top) and numerical analysis (bottom)

It was found that smoke from the glycol/water mixture was not suitable due its high moisture content, which condensed inside the scale model. As an alternative, smoke was produced by a smouldering process, using different wood-based solids as fuel, in a low-oxygen environment for safety reasons.. This produced smoke with little or no moisture content at a fast enough rate to fill the scale model. A well-ventilated laboratory was used.

Transient-state experimental and numerical results

Transient-state tests Only a single mine geometry was used in the transient-state analysis, compared with three geometries in the steady-state analysis. Eight photoelectric sensors, as described earlier, were used to monitor smoke concentration levels at different locations in the in-heading. The following sequence of activities was carried out to complete the transient-state experiment. 1. Phototransistors were connected to the data card 2. The smoke-producing device was placed inside the test section 3. All inlets and outlets of the test section were sealed off. 4. Laboratory lights were switched off to avoid affecting the phototransistors 5. The light-emitting diodes were switched on 6. Scans were initiated on the data acquisition unit and readings were taken at intervals of 1 second 7. The smoke-producing device was switched on to fill the test section with smoke 8. After 5 minutes the smoke-producing device was removed carefully to prevent smoke from escaping from the test section. 9. The jet fan, scrubber fan, and the through-road fan were switched on. At the same time the seals on the through-road inlet and outlet were removed. This marked the beginning of smoke extraction process The Journal of The Southern African Institute of Mining and Metallurgy

1. Friction between air and model wall in the experimental set-up was disregarded in the numerical analysis 2. During the experiment there was a gap of about 3 seconds between the opening of the in-heading inlet and outlet and the switching on of the fans, while in the numerical model these activities took place simultaneously 3. The KK Ń&#x201D; model used in numerical analysis was based on recommended settings for general boundary layer calculations. More suitable setting may exist. VOLUME 114

JUNE 2014

423

â&#x2013;˛

the scrubber was involved in both methods, the air flow pattern of the scrubber is not included in the steady-state results for graphical representation purposes. As can be seen, the air flow patterns for the experiment and simulation were similar. The air flow in both cases is recirculating. This is not desirable for the purpose of diluting methane gas. The results for the other two geometries showed similar results. From this it was concluded that the numerical and experimental results are in agreement. Based on full-scale experimental measurements at the Kloppersbos facility, similar recirculation patterns and penetration depths would be present with actual mining conditions and equipment. It could be argued that the flow pattern distribution is non-dimensional in nature.

Results for three of the eight sensors are presented and discussed here. These sensors are numbers 3, 7, and 8 as defined in Figure 5. To make the comparison more meaningful, both the experimental and numerical results are plotted on the same graphs. For this reason, the smoke concentration scale, which was measured proportionally as electrical resistance from the smoke sensors, was adjusted to match the mass-fraction scale obtained from the numerical results on the graphs when all the smoke has been evacuated. The scaling factor employed was determined from the smoke sensor readings at the beginning of a test (when the whole model was filled with smoke), and at the end of a test (when all smoke were cleared and steady-state smoke concentration was reached). These points were matched with the numerical mass fraction of air values at the beginning and at the end of the numerical simulation test run. Only the relative smoke extraction rate, as determined experimentally and numerically, is thus of interest. Figure 9 shows the plots of mass fraction of air against time as the fresh air from the jet fan diluted and expelled the smoke from the in-heading. At mass fraction zero and time zero the in-heading was filled with smoke, but as the air was blown into the in-heading the mass fraction of gas mixture increased until the in-heading was almost completely filled with air. The matching experimental smoke concentration is plotted using the same time axes as for the numerical results. As can be seen, for each sensor the time-dependent shape of the plots for both experiment and numerical results appears to be similar. Of greatest importance is that the numerical results predicted the time needed to clear the model from smoke with relatively good accuracy. Some deviations between the two sets of results were found (Figure 9) and can be attributed to the following:

A computational fluid dynamics model for investigating air flow patterns A ffull in-heading length, Lmh = 2333 mm in the scaled model, is used here to illustrate how the searching exercise for optimum fan positions for different in-heading lengths was accomplished. The results for the four positions Lf = 0 mm, Lf = 80 mm, Lf = 500 mm, and Lf = 1000 mm are shown in Figure 11. The extraction times that were obtained from these positions were 285, 165, 560, and 765 seconds respectively. The extraction time is defined as the time needed to reach a mass fraction of air of 0.99 (a mass fraction of smoke of 0.01). Additional search cases were conducted, which confirmed that Lf = 80 mm was the optimum position for the in-heading length of 2333 mm. The searching exercise revealed the optimum fan position for different in-heading lengths, as shown in Table III.

Figure 10—Plan view of the mine showing fan position. Lmh and Wmh are the scaled model heading length and width respectively, while Lf is the fan position with respect to the heading entrance

Figure 9—Comparison of the rate of extraction of initial volume of air by experiment and CFD methods at sensor locations 3, 7, and 8

Development of a method for determining optimum fan position for a simplified geometry Owing to the relatively good correlation between experimental and numerical results for both the steady and the transient states, numerical-based analyses can be used to determine, for instance, the optimum geometry of a jet fan in a blind in-heading. Fan positions that result in minimum smoke extraction times for five different in-heading lengths and two different widths were established. A full width (Wmh = 420 mm) and half width (Wmh = 210 mm) were explored. The results for the width of 210 mm are included only for purposes of illustration, since in this case there would be no space for the jet fan. The fan position was varied along the in-heading length, but its position with respect to the floor and the sidewall remained fixed. Since the last through-road is the most problematic area as far as dust and methane gas extraction is concerned, the wall at the blind end was used as a sensor position.

▲

424

JUNE 2014

VOLUME 114

Figure 11—Graph of mass fraction of air against time, showing the time taken to extract smoke at four different fan positions

Table III

Optimum fan position for five different heading lengths with two different heading widths Heading Length in mm (Lh) 467 933 1400 1867 2333

Fan Position in mm (Lf) Heading Width = 420mm Heading Width = 210mm 15 35 80 70 80

20 50 110 95 100

The Journal of The Southern African Institute of Mining and Metallurgy

A computational fluid dynamics model for investigating air flow patterns Symbols

Figure 12—Optimum fan position for in-heading lengths ranging between 467 mm and 2333 mm

Some optimum results From the results in Table III, a graph of fan position against in-heading length was plotted as shown in Figure 12. This graph shows that for widths of both 420 mm and 210 mm, the trends for the optimum Lf value are similar. For a given in-heading length, the optimum Lf value was greater for a width of 210 mm than for a width of 420 mm. This indicates that for narrower blind tunnels, the optimum jet fan position appears to be deeper into the tunnel. The optimum fan position is, however, also influenced by factors such as the tunnel height and volume flow rates, but a discussion of these falls beyond the scope of this paper. Similar behavioural trends regarding the optimum position of the jet fan could be extrapolated to full-scale scenarios, provided that the in-heading aspect ratios in terms of depth, height, and width remain the same.

Lf Lh P SMx SMy SMz t V x, y, z u, v, w

Jet fan position from in-heading entrance [mm]

Density [kg/m3]

Viscosity due to linear deformation [kg/m.s]

Viscosity due to volumetric deformation [kg/m.s]

In-heading length [mm] Pressure [Pa] Body force in directions x [N] Body force in directions y [N] Body force in directions z [N] Time [s] Velocity vector (column) [m/s] Space coordinates [m]

x, y, and z Velocity components [m/s]

References CHANDER, S., ALABOYUN, A.R., and APLAN, F.F. 1990. On the mechanism of capture of coal dust particles by sprays. Proceedings of the Third Symposium on Respirable Dust in the Mineral Industries, Pittsburgh, PA, 17-19 October 1990. CHENG, L. and ZUKOVICH, P.P. 1973. Respirable dust adhering to run-of-face bituminous coals. US Department of the Interior, Bureau of Mines, Pittsburgh, PA. RI 7765. NTIS No. PB 221-883. ESTERHUIZEN, G.S. and GÜRTUNCA, R.G. 2006. Coal mine safety achievements in the USA and the contribution of NIOSH research. Journal of the Southern African Institute of Mining and Metallurgy, vol. 106, no. 12. pp. 813–820. FLINT, J.D. 1990. Mine Gas and Coal Dust Explosions and Methane Outbursts – Their Causes and Prevention. MSc dissertation. University of the

Ventilation techniques are currently used to mitigate the problems of coal dust and methane gas concentrations. The objective of this study was to investigate whether a numerical technique can be used to investigate the best methods of controlling dust and methane gas concentrations in underground coal mines. Due to practical limitation, CFD modelling and experiments were conducted on a scaled-down underground coal mine model instead of at an actual mine. The numerical and experimental investigations yielded similar results for both steady- and transient-state cases. From this it was concluded that a numerical technique could be used to establish optimized ventilation conditions in scaled-down blind in-headings. The study further explored whether an optimum position of the jet fan exists for a simplified in-heading. Optimum jet fan positions do in fact exist, which indicates that an optimum jet fan position in full-scale mine application may also exist. Since the study proved that numerical technique can be used to solve ventilation problems in the scaled-down coal mine model, this technique could be extended to other mine ventilation and related studies, for example air conditioning. The Journal of The Southern African Institute of Mining and Metallurgy

Witwatersrand, Johannesburg. pp. 3–8. LIU, X., GUO, L., and ZHANG, Z. 2010. Statistics analysis of death accident in coal mines from January 2005 to June 2009. Proceedings of the AsiaPacific Power and Energy Engineering Conference, APPEEC 2010, Chengdu, China, 28-31 March 2010. pp. 5794–5796. NUNDLALL, A.R. 2006. Case study: lessons learnt from recent flammable gas explosions in South African hard rock mines. Journal of the Mine Ventilation Society of South Africa, vol. 59, no. 2. pp. 64–69. PENG, S.S. and CHIANG, H.S. 1984. Longwall Mining. Wiley–Interscience, New York. ISBN 0-471-86881-7. THOPIL, G.A. and POURIS, A. 2010. An overview of the electricity externality analysis in South Africa within the international context. South African Journal of Science, vol. 106, no. 11–12. Article no. 248 . WALA, A.M., VYTKA, S., TAYLOR, C.D., and HUANG, G. 2007. Mine face ventilation: a comparison of CFD results against benchmark experiments for the CFD code validation. Mining Engineering, g vol.59, no. 10. pp. 49–55.

◆

VOLUME 114

JUNE 2014

425

▲

Conclusions

A TRUSTED INDUSTRY LEADER For 90 years, SCAW, a South African industry leader, has partnered with leading mining houses to ensure safety critical

standards

and

productivity

targets are consistently met. Whether it’s hoisting, grinding or excavating, Scaw produces an extensive range of products that drive mining safety and productivity. %DFNHG XS E\ WHDPV RI TXDOL¿HG engineers with extensive experience in all aspects of grinding media, steel wire rope, chain and cast products, Scaw’s

product

specialists

are

available to advise on the design, selection, installation and maintenance of safety critical and mining products manufactured by Scaw.

EXPERIENCE MATTERS. Join Scaw’s global safety and productivity drive, call: Scaw South Africa Grinding Media: Haggie ® Steel Wire Rope:

Tel +27 11 621 1555 Tel +27 11 842 9548 Tel +27 11 620 0000

Chain Products: Cast Products: Rolled Steel Products:

Tel +27 16 428 6000 Tel +27 11 842 9303 Tel +27 11 842 9359

www.scawmetals.com

S12127/12128/3718

The Scaw Metals Group is a South African company serving international markets.

A heuristic sublevel stope optimizer with multiple raises by X. Bai*, D. Marcotte*, and R. Simon*

A new heuristic sublevel mining stope optimizer is presented. The optimizer seeks the best locations and lengths of a series of vertical raises that, together with the blocks linked to each raise, define a mining stope. Five design constraints – the footwall angle, the hangingwall angle, the number of raises, the maximum distance of a block from a raise, and the minimum width required to move the farthest block towards the raise – allow the shape of the sub-stopes associated with each raise to be controlled. The optimization is done on the locations and lengths of raises using a genetic algorithm to efficiently sample the parameters’ space. For each raise, a local network is defined in cylindrical coordinates around the raise such as to impose the design constraints. A maxflow algorithm on the local network is used to determine the optimal sub-stope for each raise. All substopes are combined to define the global stope for the entire deposit. The best global stope is obtained using a genetic algorithm to find the raise parameters providing the best profit over the entire deposit. Two synthetic cases and one real deposit are used to evaluate the new algorithm and compare the results with the single-raise optimizer. The multiple raises approach leads to significantly improved economics compared with the single-raise stope optimizer, and the dilution is also substantially reduced compared to the single-raise case. Keywords underground mining, stope optimization, mining constraints, maximum flow algorithm, cylindrical coordinate transformation, multiple raises.

Introduction In underground mining, stope design aﬀects ﬀ the profit and safety of the operation. Stope design requires: (1) a prior ore reserve model as input data, usually obtained by estimation or simulation using geostatistical tools (David, 1988; Journel and Huijbregts, 1978), and (2) the geotechnical constraints, including the hangingwall and footwall angles, the stope dimensions, the in situ stress tensor, the rock strength, and the local geological structures. The general procedure of stope optimization is to decide which volumes are included in the stope and which are not so that, under the geotechnical constraints, the resulting stope produces the greatest profit possible. In the last few decades, several approaches have been developed for stope optimization. These methods were reviewed by Ataee-Pour (2005) and Alford et al. (2007). The dynamic The Journal of The Southern African Institute of Mining and Metallurgy

* Department of Civil, Geological and Mining Engineering, École Polytechnique de Montréal, Canada. © The Southern African Institute of Mining and Metallurgy, 2014. ISSN 2225-6253. Paper received r Jun. 2013; revised paper received Sep. 2013. VOLUME 114

JUNE 2014

427

▲

Synopsis

programming method (Riddle, 1977) and branch and bound technique (Ovanic and Young, 1995) were used to optimize a stope in one or two dimensions. However, these methods fail to produce realistic stopes for complex three-dimensional (3D) deposits that cannot be simplified to two-dimensional (2D) mining problems. Some 3D techniques were also reported, including mathematical morphology tools (Serra, 1982; Deraisme et al., 1984), floating stope technique (Alford, 1996), maximum value neighborhood method (Ataee-Pour, 2000), and octree division approach (Cheimanoﬀ et al., 1989). These heuristic methods cannot directly integrate the geotechnical constraints. Recently, Manchuk and Deutsch (2008) provided a simulated annealing-based algorithm, with some mining constraints incorporated. However, simulated annealing is very slow and the convergence to a global optimum is therefore not ensured in practice. Moreover, the restriction of perturbations to moves respecting all constraints on slopes in a three-dimensional setting can seriously hamper the capacity of the algorithm to find a good solution. Bai et al. (2013) developed a stope optimizer based on graph theory. The basis of the approach is the vertical raise initiating the opening necessary for blasting, which plays a similar role as the surface in open-pit mining. A cylindrical coordinate system is defined around the raise. Then a network is built where cylindrical blocks are linked towards the raise such as to impose de facto the geotechnical constraints. The optimal block selection is obtained by applying eﬃcient maximum flow methods over the network. Two important design parameters were defined: the maximum distance a block can be

A heuristic sublevel stope optimizer with multiple raises to reach the raise (R) (or distance off influence f off the raise), and the minimum width required to move this farthest block to the raise (yR) (see Figures 1 and 2). The stope obtained is optimal for the raise location and extent chosen, and for the R, yR, and footwall and hangingwall angles imposed. The global stope optimization then simplifies to finding the best raise location and extent within the orebody. The optimizer was shown to provide good results on a number of simple deposits, both synthetic and real. Although quite appealing, the single-raise optimizer has a few drawbacks. Firstly, it cannot provide a satisfying solution for large deposits or lenses where more than one raise is needed. In that case, repetitive application of the single-raise optimization needs to take into account the interactions between the raises. Also, for small deposits or lenses with curved shapes (e.g. following folds), the single-raise optimal solution could provide more dilution and less profit than a manual solution obtained with more raises, each raise having a smaller distance of influence R. In this paper, the authors aim at solving these drawbacks. The algorithm of the single-raise optimizer is extended to multiple-raise situations, keeping the core component of generating a sub-stope for each raise. In the multiple-raise framework, each sub-stope is a feasible geometry with controllable maximum dimensions. Optimization is done on the set of raise parameters using a genetic algorithm to efficiently sample the parameter space. For a given set of raise parameters, each raise is optimized with the single-raise optimizer, thus defining as many sub-stopes as there are raises. Each sub-stope meets the design parameters and is optimal for that particular raise location and length and set of design constraints. The union of all points within one or more of the sub-stopes defines heuristically the global stope. Three deposits, two synthetic and one real, are considered where the results obtained with the single raise and multiple raises are compared and discussed. For the sake of simplicity, the true grade values of the deposit are assumed known everywhere or obtained using a conditionally unbiased estimator (David et al., 1984), so the effect ff of the uncertainty on grades with regard to the stope design is not considered in this study.

Methods Stope optimization with a single raise The stope optimizer using a single raise proposed by Bai et al. (2013) is reviewed. After a brief summary of the required graph theory for mining optimization, the network construction used to represent mining constraints is described and the stope optimizer workflow is presented.

Graph theory in mining optimization The ore block model and the mining constraints are represented as a weighted directed graph (or network) G = (V, A), where the vertices V denote the ore blocks and the oriented arcs A define the precedence relations between blocks so as to incorporate the mining constraints. The profit from mining block i is pi. It is computed from the block i ore grade and tonnage, the recovery factor, the mining and processing costs, and the mineral price (Lane, 1988).

▲

428

JUNE 2014

VOLUME 114

The stope optimization amounts to ffinding the closed set of nodes V′ ⊆ V such that ∑i∈V′ pi is a maximum. Let Γi be the subset of immediate successor nodes to node i, representing the set of blocks that need to be mined prior to block i. The maximum closure problem is formulated as: [1] [2] [3] where xi equals 1 when the block is selected, 0 if not, and index j refers to a successor of block i. For a typical deposit, the integer program involves a lengthy computational time due to the large number of ore blocks N. The Lerchs-Grossman algorithm (LGA) (Lerchs and Grossman, 1965) presented an effective ff tool to solve the open-pit mining problem implemented in some commercial software. Even more efficient methods appeared after the seminal paper of Picard (1976) proving that the maximum closure problem of the open-pit mine is equivalent to the minimum cut problem, hence allowing the application of maximum flow algorithms (e.g. Goldberg and Tarjan (1988); King et al. (1992)), which are substantially more efficient than the LGA (Hochbaum, 2001, 2002). The high efficiency enables the repetitive application of the algorithm while still keeping computing costs realistic.

Implementation of stope geometric constraints in network To apply the network flow concept to stope optimization, the key is to find the free surface to start stoping, similar to the ground surface in an open-pit mine. Actually, in sublevel stoping, the raise, the vertical or sub-vertical tunnel, plays the role of the initial free surface. The introduction of cylindrical coordinates starting from the raise (Figure 1a) facilitates the control of geometric constraints as the stoping sequence can be expressed by linkages of the cylindrical blocks toward the raise. Bai et al. (2013) indicated how the common stope constraints are implemented with the diﬀerent ﬀ linkages in a graph. The hangingwall and footwall slope constraints define the precedence links in the vertical direction (Figure 1b). For a cylindrical system with blocks defined by Δr, Δθ, Δz (see Figure 1b), and Δz/Δr = 1, one link upward and two links downward define a hangingwall slope of 45° and a footwall slope of 63.4°. Stope width is controlled by linkages in the horizontal plane, specifically one radial link and two side links for each block (see Figure 2a) and by two design parameters: the maximum extent of the stope from the raise R (or distance of influence of the raise) and the minimum width (yR) needed to move, by gravity, the farthest block to the raise. For example, with R = 30 m, Δθ/Δr=1 degree per metre, three horizontal links towards the raise provide yR = 7.7 m, (see Figure 2). A third type of geotechnical constraint, the maximum stope height, is simply controlled by the length of the raise. The blocks above the top of the raise or under its bottom are not part of the network, hence are not contained in the stope.

Algorithm for a single raise The optimization algorithm consists of two main parts. The first part, the stope optimizer, is the core of the approach. It The Journal of The Southern African Institute of Mining and Metallurgy

A heuristic sublevel stope optimizer with multiple raises

Figure 1—Block model under cylindrical coordinates (a), and typical arcs in vertical section in the proposed method (b)

The second part is to search the best raise location and height. This is done by global optimization on the raise location and height parameters, using as the objective function the stope value found with the stope optimizer. The single-raise approach has some limitations. For example, when R is large, a relatively wide stope is produced as many blocks have to be mined before the farthest blocks are accessed. When the deposit is curved or inclined, this can lead to the mining of a substantial amount of waste as shown in Figure 3. In other scenarios, isolated clusters of ore could be left in the ground because the ore clusters do not pay for the additional waste included. In these cases, a better approach would be to use more than one raise so as to define smaller sub-stopes, hence diminishing the eﬀects ﬀ due to curvature or inclination of the orebody, and simultaneously allowing more flexibility to reach isolated clusters of ore.

Stope optimization with multiple raises Similar to the single-raise algorithm, the multiple-raise algorithm is comprised of two main parts: (1) the stope generator with multiple raises based on a series of separate network flow problems, one for each raise; and (2) the optimization of the best parameters for the raises’ locations, extents, and zones of influence.

Stope generator with multiple raises

Figure 2—Horizontal plane showing (a) blocks and links defined in the cylindrical system and (b) corresponding blocks and links in the Cartesian system. Shaded blocks represents blocks to be removed to access block A. Trace of the envelopes defined by the lateral links in the cylindrical system (c) as they appear in the Cartesian system (d)

Each raise is first treated separately with the single-raise optimizer described previously. For each raise, a cylindrical coordinate system is defined, and the stope constraints are implemented through the precedence relations in the associated network. The maximum extent that a raise can access, or distance of influence Ri, is defined to control the maximum size of the sub-stope for raise i. As a result, an optimal sub-stope is generated for each raise (see Figure 4). The sub-stopes in cylindrical coordinates are converted to a common regular grid in Cartesian coordinates. For the conversion, the status of a grid point in Cartesian coordinates (in or out of the stope) is identified by the status of the nearest cylindrical block centroid in each sub-stope. It suﬃces that a grid point belongs to any of the sub-stopes to be identified as being in the global stope. Hence, the global stope is the union of all the sub-stopes. The profit from the global stope is calculated on the Cartesian grid, so as to avoid counting any part of the stope twice (or more).

generates an optimal stope for a specified raise location and height, with chosen design parameters R and yR. The stope optimizer includes the following steps:

The Journal of The Southern African Institute of Mining and Metallurgy

Figure 3—Illustration of possible problems with one raise: (a) in a horizontal section, the envelope from A to the raise includes a large quantity of waste; (b) in a vertical section, waste has to be mined in the upper part due to the network associated with the single raise VOLUME 114

JUNE 2014

429

▲

1. Construct an economic block model in cylindrical coordinates with given raise location and height as the reference axis 2. Build the graph with vertical arcs to impose slope constraints, and horizontal arcs to impose width constraints 3. Construct the flow network by adding the source and sink nodes to the graph 4. Solve the maximum flow problem. The generated stope is conditionally optimal to the raise location and height.

A heuristic sublevel stope optimizer with multiple raises mutations are generated by introducing, in the child, chromosome - genes that do not come from the parents. The mutations enable new areas of the parameters’ space to be explored. The less fit individuals in the population are eliminated so as to keep the size of the population constant. With the iterations, the average fitness of the population increases, until the convergence or another stopping criterion is reached (see Figure 5). The genetic algorithm proposed follows the following steps:

Figure 4 – Conceptual model of the stope generator with multiple raises in a horizontal section: (a) two ore models in cylindrical coordinates, one for each raise, are established; (b) and (c) first and second substopes in a cylindrical coordinate obtained by the maxflow method on the two separate networks; (d) and (e) the sub-stopes (b) and (c) converted on the Cartesian grid; (f) the final stope in the Cartesian grid from (d) and (e)

Optimization of multiple raises parameters with a genetic algorithm b t In the model, each raise is parameterized as, (xi, yi, zi , zi , and Ri), i = 1, . . . , n, where xi and yi denote the coordinates b t of raise i in horizontal section, zi and zi represent its bottom and top elevation, Ri is the maximum distance a block can be

from the raise i, and n is the number of raises. The global stope is obtained as the union of sub-stopes that are each optimal in their local cylindrical system. A good set of parameters for the raises is found using a genetic algorithm (Holland, 1975) to allow an eﬃcient sampling of the parameters’ space. The genetic algorithm tries to mimic the natural evolution of a population. An instructive example of the application of genetic algorithms to mining optimization is presented in Armstrong et al. (2012). Here, we define an individual as a single set of multiple raise parameters. Starting from a population of individuals, one creates new individuals in the population by crossover and mutations. In our algorithm, a vector of multiple raise parameters represents the individual chromosome with 5n genes, since there are 5 parameters to optimize for each raise. The profit from the global stope associated with the raise parameters measures the fitness of the individual to its environment. The algorithm is initiated by generating an initial population. The fitness of each individual in the population is evaluated. A certain proportion among more fit individuals is randomly selected as parents. Each set of parents mates and creates a child whose genes are inherited from them (crossover). Moreover, a certain proportion of

▲

430

JUNE 2014

VOLUME 114

➤ Initial population—The initial population comprises two parts: more fit individuals and random individuals. The more fit individuals ensure the inclusion of good genes. They can be obtained in two ways: (a) intuitive good raise parameters selected by the user; or (b) fast optimization with an initial low-resolution ore model. Random individuals covering the whole range of raise parameters are added to the initial population to allow suﬃcient genetic diversity so as to better explore the parameters’ space and identify interesting area (Haupt and Haupt, 2004). ➤ Parent selection—The basic principle of parent selection is to give more fit individuals a higher priority to be parents. Less fit individuals have less probability to be selected for mating. Although the chance is small, it is important to allow the less fit individuals to pass down their genes for the purpose of diversifying the population. The fitness proportionate selection (or roulette wheel selection) method (Back, 1996) is used for parents selection. The individual i with fitness value (profit) fi has the probability of being selected probi =

fi/∑ ∑M j=1 fj, where M represents the number of individuals in the population. To implement this, all the fitness values are normalized to be located in [0,1]. The normalized values are sorted in ascending order and then transformed to a cumulative normalized fitness value (CNFV). A random number Rnd is drawn

Figure 5—Genetic algorithm diagram to search for the best raises’ parameters The Journal of The Southern African Institute of Mining and Metallurgy

A heuristic sublevel stope optimizer with multiple raises ➤ Termination of iterations—The s loop stops when a series of successive iterations does not improve the best individual fitness or the average fitness of the population, or when a maximum number of iterations has been reached.

f from [0,1], and the ffirst one with CNFV greater than Rnd is selected for the mating pool. The individuals in the mating pool are randomly paired. ➤ Genetic operator. Two mating methods are employed: crossover and mutation. The crossover is done by picking genes randomly from parents and combining them to define a new individual, following the formula:

Results

[4]

Parameters in the algorithm where Xnew denotes the new individual, Xi represents the parent i, and βi is a 0-1 variable indicating whether the gene is inherited or not. In this way, a child is basically the recombination of genes from its parents. Mutation can be applied to a child to allow the child genes to depart substantially from those of its parents. This is done using (Haupt and Haupt, 2004):

To test and evaluate the proposed methods, three orebody models are used: two synthetic deposits and one real deposit (ore block model estimated by kriging). The two synthetic models illustrate typical scenarios where the single-raise algorithm partly fails and where the multiple-raise algorithm is expected to perform better. The initial synthetic block model is expressed on a Cartesian grid of spacing 1 m × 1 m × 1 m. For the larger real deposit model, the Cartesian grid is defined at every 2 m so as to save some computing time for the interpolation (it was checked that the results obtained are robust to this choice). The networks have one link vertically upward and two links downward, and three links horizontally. Therefore, for the three cases, the hangingwall angle is 45°, and the footwall angle is 63°. The Δθ is computed so as to ensure approximately the desired yR (see Table I).

[5] where ◦ denotes the Hadamard (or element-wise) product between the vectors; Z is a column vector of random numbers drawn from the standard normal distribution; I is a 0-1 column vector indicating the genes to be mutated, and the σ scalar controls the extent of the mutation. A posterior check is applied to the mutated genes so as to ensure they remain in feasible ranges for the raise parameters.

Table I

Geometric and design parameters, discretization, and optimized raise parameters Parameters Raise type Economic parameters Mining and processing costs ($/t ore) Metal price ($/kg) Recovery rate Rock density Mean ore grade (%)

Case 1 M1

Case 2

Case 3

← ← ← ← 0.25

50 10 0.9 3 0.25

→ → → → 0.54

0.54

45 63 40 15

→ → 150 50

150 50

Geometric parameters for stope Minimum footwall angle (deg) Minimum footwall angle (deg) Maximum height (m) Minimum height (m)

50 10

← ← 40 15

Discretization dz (m ( ) dr (m ( )

0.5 0.5

1 1

Optimized raises parameters Raise 1 Location X (m) Location Y (m) Bottom level (m) Top level (m) Maximum radius R (m)

34.4 -20.9 -127.3 112.3 16.2

18.9 20.9 -132.5 -107.0 27.3

30.9 21.4 -124.5 -105.0 12.7

33.1 18.7 -124.2 -106.1 33.7

3164.7 -76.6 -493.8 -397.8 18.9

3097.5 -72.5 -497.3 -352.8 78.7

Raise 2 Location X (m) Location Y (m) Bottom level (m) Top level (m) Maximum radius R (m)

10.2 20.2 -132.9 -107.1 16.1

Raise 3 Location X (m) Location Y (m) Bottom level (m) Top level (m) Maximum radius R (m) 2S:

3121.7 -73.7 -498.0 -352.4 46.4

12.0 17.5 -125.2 -105.7 14.7

3121.6 -74.5 -438.5 -352.9 45.9

Optimization with multiple raises Optimization with single raise

The Journal of The Southern African Institute of Mining and Metallurgy

VOLUME 114

JUNE 2014

431

▲

1M:

48.0 18.9 -125.2 -105.2 14.8

A heuristic sublevel stope optimizer with multiple raises For the genetic algorithm, the initial population size is 40 × n, where n is the number of raises. Two more-fit individuals are included in the initial population. The first one is the solution obtained by optimization at a lower resolution. The second is obtained by spreading the raises uniformly within the deposit. Three parents are used to create a new individual. The mutation rate is selected to be 0.1 for each gene, so that many offspring ff will include mutations of their parent genes. In each iteration, 20 × n new individuals are created and the same number of least fit individuals are eliminated to keep the size of the population stable. The optimization stops when the number of iterations reaches 100, or when the most fit individual among the population does not improve in 10 successive iterations. In the examples tested, these choices constitute a good tradeoff to ensure simultaneously a good final solution and to keep the computation time tractable. Increasing significantly the number of offspring ff or the population size, or making the termination rules more stringent, would possibly provide a slightly better final solution but at the cost of additional computation time. Admittedly, these values might have to be adapted to the particular deposit being studied. Bai (2013) verified with two simple cases that the preceding choices for the genetic algorithm were sufficient to ensure retrieving the known optimal value with high probability. The design parameter yR controls the minimum width the sub-stope must have for a block located at distance R from the raise. This value is likely to vary according to the rock mechanics condition of the deposit in the area where the stope is created. A smaller yR allows dilution to be reduced and profit increased. However, it should not be too small, otherwise there is the risk of ore jamming within the stope. Rock mechanical conditions and experience with mining in a particular geological environment should guide the choice of this parameter. With the examples tested, a value around yR = R/3 seems to provide visually sensible shapes. Here, to diminish the number of factors to study, we choose yRi = Ri/3 for all cases. The effect ff of yR is further discussed later. The first synthetic case represents two distinct mineralized lenses (Figure 6). The optimal single-raise solution (c and d) locates the raise in the waste approximately at mid-distance of the lens centroids. In contrast, the multiple-raises solution locates, as expected, the two raises close to the centroid of each lens (Figures 6e and f). Moreover, the radius of influence of each raise Ri is correctly identified as larger for the larger lens. This solution provides 13.5% more profit than the single-raise solution and the dilution of ore is reduced significantly from 21.6% to only 2.3% (Table II).

The second case is an ore vein with changing direction in horizontal section (Figures 7a and b). The vein is approximately 60 m long by 10 m wide by 20 m high. The stope is designed with three raises (see Figure 7). The value of the stope with multiple raises is 10.7% higher than with the single raise ($976 000 vs. $882 000). It includes less waste (cost $4 400 vs. $57 700) and misses less ore ($4 100 vs. $45 000). The dilution of the multiple raises solution is onethird the dilution of the single raise (3.1% vs. 10.4%) (Table II). The kriging block model of a metal deposit in Canada is used as the third case study (the name and location of deposit are undisclosed for confidentiality reasons). A portion

Figure 6—Case 1, simulated ore model and created stopes: (a) 3D-view of the orebody; (b) x-y horizontal section of the orebody at z=-120 m; (c) 3D view of the optimized stope with a single raise; (d) x-y horizontal section of the single raise-stope at z=-120 m, showing ore in stope (blue), waste in stope (red), and ore out of stope (green); (e) 3D view of the optimized stope by multiple raises; (f) x-y horizontal section of the multiple raises’ stope at z=-120 m. Raises in black. Design parameters as in Table I

Table II

Economic evaluation of the case studies Cases

Raise type

Stope profit (000$)

Missed ore value (000$K)

Waste value in stope (000$)

Dilution volume rate 1

Profit improved

Case 1

Multiple Single

2622 2312

24.2 84.6

-24.0 -273.6

2.3 % 21.6 %

13.41 %

Case 2

Multiple Single

976 882

4.1 45.0

-4.4 -57.7

3.1 % 10.4 %

10.68 %

Multiple 8933 90.4 Single 8327 543.8 1Dilution volume rate = Volume of waste in stope / Volume of stope

-44.2 -197.0

3.4 % 8.1 %

7.28 %

Case 3

▲

432

JUNE 2014

VOLUME 114

The Journal of The Southern African Institute of Mining and Metallurgy

A heuristic sublevel stope optimizer with multiple raises

Figure 7—Case 2, simulated ore model and created stopes: (a) 3D-view of the orebody; (b) x-y horizontal section of the orebody at z=-120 m; (c) 3D view of the optimized stope with a single raise; (d) x-y horizontal section of the single raise stope at z=-120 m, showing ore in stope (blue), waste in stope (red), and ore out of stope (green); (e) 3D view of the optimized stope by multiple raises; (f) x-y horizontal section of the multiple raises’ stope at z=-120 m. Raises in black. Design parameters as in Table I

therefore are respected in the global stope. The effect ff of the yRi parameter is case-specific. For a single raise and a given R, diminishing yR necessarily increases the profit. However, yR should not be taken too small otherwise there is a risk of ore jamming within the stope. Rock mechanical conditions and experience with the mining in that particular geological environment should guide the choice of this parameter. A value around yR = R/3 seems to provide visually sensible shapes in the tests conducted. The sensitivity of the best solution to this parameter is not expected to be high. As an example, in case 3 the profits obtained with yR = R/2, R/3, and R/4 are respectively $8 909 000, $8 933 000, and $8 948 000, showing differences ff of only 0.4% between R/2 and R/4 and 0.17% between R/3 and R/4. Similar small differences ff were observed for the ore and the waste in the stope. The computation time of the algorithm depends of two main factors: the time required for one iteration of the multiple-raise stope generator and the time required to explore the raise parameters’ space by GA (or eventually an alternate search method). To give a rough idea, case 2 has 40 x 60 x 31 blocks in a Cartesian grid, and the stope generator with three raises takes between two and ten seconds (on a laptop) to produce a stope. The exact time depends on the raise extent and on Ri which, for a given discretization, control the number of blocks. To find the best raise locations, the GA evolves during 57 generations, which takes around three hours of computation, the stope generator being called a total of 57 × 20 × 3 + 40 × 3= 3540 times.

of the deposit of size 108 m × 68 m × 148 m (Figures 8a, b, and c), is selected for stope design for the sublevel stoping method. Three raises are used to optimize the stope as shown in Figures 8g, h, and i. The optimized raise parameters are given in Table I. This time, the profit of the multiple-raise stope is 7.3% higher than the profit of the single-raise stope. The dilution for the multiple raises is 3.4% of the stope volume compared to 8.1% for the single-raise solution.

Discussion

The Journal of The Southern African Institute of Mining and Metallurgy

Figure 8—Case 3, test with a real ore deposit: (a) 3D-view of the orebody; (b) y-z vertical section of the orebody at x=3130 m; (c) x-y horizontal section at z=-424 m; (d) 3D view of the optimized stope with a single raise; (e) y-z vertical section at x=3130 m; (f) x-y horizontal section at z=-424 m; (g) optimized stope with multiple raises, (h) y-z vertical section at x=3130 m; (i) x-y horizontal section at z=-424 m. In (d) and (g), stopes are in red, ore out of stope is in green. (e), f), (h), and (i), ore in blue, waste in red, and ore out of stope in green, raises in black. Design parameters as in Table I VOLUME 114

JUNE 2014

433

▲

We have developed an improved stope optimizer for the sublevel stoping method. The new multiple-raise stope optimizer is an extension of the single-raise optimizer presented in Bai et al. (2013). In all test cases, the multipleraise algorithm provides stopes with higher profit and less dilution compared to the single-raise optimizer. In the three test cases presented, the best global stope from multiple raises generated between 7.3% to 13.4% more profit than the best stope obtained with a single raise. The improvement in the multiple raise heuristic solution compared to the single-raise approach is due to its increased flexibility. The distance of influence of a raise Ri is the parameter controlling the size of a sub-stope. It is ensured that the geotechnical constraints are respected within each sub-stope,

A heuristic sublevel stope optimizer with multiple raises One limitation off the proposed approach is the restriction to vertical raises. It would cause a higher dilution rate for scenarios of inclined deposits, which usually adopt inclined raises in reality. The generalization of the method to an inclined raise is far from evident due to the loss of symmetry with respect to the gravity force vector. The solution to the problem needs further investigation. Also, in the proposed approach, the cost of development of access to the top and bottom levels of the raises was neglected. When the multiple raises are located at different ff levels, the relative additional costs would reduce the benefit of this approach compared to the single raise. Moreover, it was supposed that the raises could be located rather freely within the deposit without imposing constraints on the elevations of the beginning and the end of the various raises. An alternative strategy, closer to the practice for larger deposits, would be to optimize the common height between levels (within specified bounds) and impose each raise to span the entire height. The optimization would then simplify to find the best elevation for the first level and find the best number and locations of raises within each level. This modification is currently being investigated. At first glance, the adoption of a cylindrical system of coordinates might appear as an unnecessary complication. In fact, it seems difficult (if even possible) to define the network directly in the Cartesian system of coordinates to ensure simultaneously all slope constraints, the raise width, and the distance of influence of a raise. Referring to Figure 2d, it is obvious that the links to impose vary according to the distance of influence considered. For each block, it would be necessary to compute the corresponding envelope of preceding blocks. Moreover, for a given distance of influence, any given internal block would be covered by many such envelopes coming from outer blocks. The network associated with the different ff envelopes for a given block might easily differ ff from one envelope to the other, rendering the network definition impossible. This complexity vanishes with the cylindrical system.

Conclusions The proposed method was shown to provide good heuristic stope solutions for typical geometries of curved or inclined deposits. The solutions respect the geotechnical constraints of the sublevel stoping method, including footwall and hangingwall slopes, and stope minimum and maximum heights. The best stopes obtained with multiple raises for the three cases considered exhibit significantly larger profit (+7.3% to +13.4% increase) and less dilution (58% to 89% reduction) compared to the best solution by the single-raise method. The gain of the multiple-raises approach is due to the increased flexibility compared to the single-raise case.

Acknowledgement This research was made possible by research grants provided by the Chinese Scholarship Council and the Natural Sciences and Engineering Research Council of Canada (NSERC). The authors are grateful to two anonymous reviewers for their constructive comments and editing suggestions that helped improve the manuscript.

▲

434

JUNE 2014

VOLUME 114

References ALFORD, C. 1996. Optimisation in underground mine design. International Journal of Rock Mechanics and Mining Sciences and Geomechanics Abstracts, vol. 33, no. 5. pp. 220A–220A. ALFORD, C., BRAZIL, M., and LEE, D.H. 2007. Optimisation in underground mining. Handbook of Operations Research in Natural Resources. Weintraub, A., Romero, C., Bjørndal, T., and Epstein, R. (eds.). Springer, New York. pp. 561–577. ARMSTRONG, M., VINCENT, A., GALLI, A., and MHEUT, C. 2012. Genetic algorithms and scenario reduction. Ninth International Geostatistics Congress, Oslo, Norway. ATAEE-POUR, M. 2000. A heuristic algorithm to optimise stope boundaries. Ph.D. thesis, University of Wollongong, Australia. ATAEE-POUR, M. 2005. A critical survey of the existing stope layout optimization techniques. Journal of Mining Science, vol. 41. pp. 447–466. BACK, T. 1996. Evolutionary Algorithms in Theory and Practice: Evolution Strategies, Evolutionary Programming, Genetic Algorithms. 1st edn. Oxford University Press, USA. BAI, X. 2013. Optimization of underground stope with network flow method. PhD thesis, École Polytechnique de Montréal, Canada. BAI, X., MARCOTTE, D., and SIMON, R. 2013. Underground stope optimization with network flow method. Computers and Geosciences, vol. 52. pp. 361–371. Cheimanoff, N.M., Deliac, E.P., and Mallet, J.L. 1989. GEOCAD: an alternative CAD and artificial intelligence tool that helps moving from geological resources to mineable reserves. 21st International Symposium on the Application of Computers and Operations Research in the Mineral Industry. Society for Mining, Metallurgy and Exploration Inc., Colorado. pp. 471–478. DAVID, M. 1988. Handbook of Applied Advanced Geostatistical Ore Reserve Estimation. Elsevier. DAVID, M., MARCOTTE, D., and SOULIE, M. 1984. Conditional bias in kriging and a suggested correction. Geostatistics for Natural Resource Characterization. Part 1. Verly, G., David, M., Journel, A.G., and Marechal, A. (eds.). Reidel, Dordrecht, Netherlands. Volume 122 of NATOASI C. pp. 217–230. DERAISME, J., DE FOUQUET, C., and FRAISSE, H. 1984. Geostatistical orebody model for computer optimization of profits from different ff underground mining methods. Proceedings of the 18th International Conference on the Application of Computers and Operations Research in the Mining Industry (APCOM), London, England. pp. 583–590. GOLDBERG, A. and TARJAN, R.E. 1988. A new approach to the maximum-flow problem. Journal of the Association for Computing Machinery, vol. 35. pp. 921–940. HAUPT, R.L. AND HAUPT, S.E. 2004. The continuous genetic algorithm. Practical Genetic Algorithms. John Wiley & Sons. pp. 51–66. HOCHBAUM, D.S. 2001. A new-old algorithm for minimum-cut and maximumflow in closure graphs. Networks, vol. 37. pp. 171–193. HOCHBAUM, D.S. 2002. Solving integer programs over monotone inequalities in three variables: a framework for half integrality and good approximations. European Journal of Operational Research, vol. 140. pp. 291–321. HOLLAND, J.H. 1975. Adaptation in natural and artificial systems: an introductory analysis with applications to biology, control, and artificial intelligence. University of Michigan Press, Ann Arbor. JOURNEL, A.G. and HUIJBREGTS, C.J. 1978. Mining Geostatistics. Academic Press, London. KING, V., RAO, S., and TARJAN, R. 1992. A faster deterministic maximum flow algorithm. Proceedings of the Third Annual ACM-SIAM Symposium on Discrete Algorithms. Academic Press, Orlando, FL. pp. 157–164. LANE, K.F. 1988. The Economic Definition of Ore: Cut-off Grades in Theory and Practice. Mining Journal Books, London. LERCHS, H. and GROSSMAN, I.F. 1965. Optimum design of open-pit mines. - CIM Bulletin, vol. 58. pp. 47–54. MANCHUK, J. and DEUTSCH, C. 2008. Optimizing stope designs and sequences in underground mines. SME Transactions, vol. 324. pp. 67–75. OVANIC, J. and YOUNG, D.S. 1995. Economic optimisation of stope geometry using separable programming with special branch and bound techniques. Third Canadian Conference on Computer Applications in the Mineral Industry. Balkema, Rotterdam. pp. 129–135. PICARD, J.C. 1976. Maximal closure of a graph and applications to combinatorial problems. Management Science, vol. 22. pp. 1268–1272. RIDDLE, J.M. 1977. A dynamic programming solution of a block-caving mine layout. Proceedings of the Fourteenth International Symposium on the Application of Computers and Operations Research in the Mineral Industry, October 4-8, 1976. Society for Mining, Metallurgy and Exploration Inc., Colorado. pp. 767– 780. SERRA, J.P. 1982. Image Analysis and Mathematical Morphology Academic Press, New York. ◆ The Journal of The Southern African Institute of Mining and Metallurgy

Risk indexing tool for mine planning by W. Abdellah*, H.S. Mitri*, D. Thibodeau†, and L. MoreauVerlaan†

The purpose of this paper is to establish a qualitative method to estimate the risk level (e.g. rating and ranking) resulting from mining activity. Risk is the product of two factors: probability of failure and cost of consequences. A resultant assessment scale matrix is then used to assign a risk index value which is directly proportional to the potential for excavation instability. A case study, the No. 1 Shear East orebody at Vale’s Garson Mine in Sudbury, Ontario will be examined. A three-dimensional, elastoplastic, finite difference model (FLAC 3D) is presented for a mine development intersection situated 1.5 km below ground surface. The developed assessment scale matrix is used to estimate risk index for intersection (2981) located on 5000 level. The results are presented and categorized with respect to risk-index value, probability of instability, cost of consequence, and mining stage. Keywords risk-index tool, cost of consequence, probability of instability, numerical modelling, case study, underground mine developments.

Introduction Stability is a key issue in underground mining. The stability and serviceability of mine developments (e.g. haulage drifts, crosscuts, and intersections) are crucial parameters influencing the success of ore extraction. Unexpected instability is expensive and is a risk to personnel and equipment. Failed or damaged mine developments will require extra expenditures for repair: slashing, rehabilitation costs, costs of adding secondary support, and delay of production. Clearly, delays caused by instability are costly and time-consuming and should be avoided (Ellefmo and Eidsvik, 2009). Engineers have to achieve stability in their design while dealing with uncertain ground conditions. The complexity of the design process increases with the lack of accurate data. Furthermore, safety standards must meet all laws and regulations set by government agencies. Additionally, there are many parameters to be considered in the design process such as: safety, serviceability (e.g. quality of technical solution), economics (e.g. cost), environment, and rock mass properties. For example, rock mass properties alone, are The Journal of The Southern African Institute of Mining and Metallurgy

Dealing with uncertainty Uncertainty and variability govern the geomechanical data collected from the natural environment. Thus, a reliable design approach must include consideration of uncertainties, evaluation of the probability of occurrence of a risk, and measures to be adopted to reduce the risk to an acceptable level. Reducing the risk can involve the narrowing of the uncertainty range (e.g. by collection of additional data). Rock mass properties are significant geotechnical design input parameters. However, these parameters are never known precisely. There are always uncertainties associated with them. Some of these

* McGill University, Montreal, Canada. † Vale Ltd, Sudbury, Canada. © The Southern African Institute of Mining and Metallurgy, 2014. ISSN 2225-6253. Paper received r Aug. 2013; revised paper received Nov. 2013. VOLUME 114

JUNE 2014

435

▲

Synopsis

complex and are associated with uncertainty in deep underground mines. These five factors should be maintained and combined together in the decision-making process. Consequently, wrong decisions may lead to unwanted risks. In order to facilitate decision-making, probabilistic analysis and risk assessment should be adopted (Einstein, 1996; Sturk, Olsson, and Johansson, 1996). The best way to express performance uncertainty is to describe it in the form of a distribution (e.g. probability density function) related to a fixed limit (e.g. threshold of performance criteria) as shown in Figure 1. This study presents a useful methodology for choosing secondary support requirements, both temporally and spatially. The methodology is based on estimation of risk level using risk-index as a mine design tool. This tool is adopted through stability evaluation of mine development intersections with respect to mining steps with delayed backfill.

Risk indexing tool for mine planning

uncertainties are due to lack of knowledge or limited data availability, and some are intrinsic. Furthermore, some may arise from errors in testing (e.g. estimating strength of intact rocks, mapping the joint spacing, assessing the joint surface condition) and random data collection. Thus, a great deal of uncertainty is inherent in the design of underground excavations. Uncertainty can be dealt with in different ways; deterministic, parametric, and probabilistic analysis as shown in Figure 2. In a deterministic analysis, single values are used as model input parameters. Therefore, a resultant single output value does not give any information on the variability of the input parameter (e.g. no distribution). A parametric study (e.g. sensitivity analysis) can be performed by varying a single parameter according to a certain range (e.g. coefficient of variation) while keeping all other variables constant. This gives an understanding of the effect of certain parameters on the overall model behaviour. However, no distribution is obtained using this method. If data is limited, probabilistic methods or statistical simulations are more powerful. These methods quantify the uncertainty and estimate the likelihood (probability) of occurrence. Therefore, engineers can develop more reliable, robust designs and economical solutions (Hammah, Yacoub, and Curran, 2008). In this paper, the modified point-estimate method (Zhou and Nowak, 1988) of (2n2+1) is used to estimate the probability of unsatisfactory performance for intersection no. 2981 on 5000 level at Vale’s Garson Mine in Sudbury, Ontario. Also, the cost of potential consequences from this intersection is calculated. Finally, risk-indices are estimated.

off approximately 610 m (2000 ffeet) each, and a dip off 70º to the south. These orebodies, however, vary in size and shape. An olivine diabase dyke crosses the two orebodies near midstrike on the 5100 level. The dyke is steeply dipping to the south-west and continues with depth. The footwall typically consists of norite (NR) and greenstone (GS) and the hangingwall consists of metasediments (MTSD) as shown in Figure 3. The stability analysis is conducted, using Itasca's FLAC3D (ITASCA, 2009) for the No. 1 Shear East orebody, whereby a planned sequence of 108 stopes over four production levels (5100, 5000, 4900, and 4800) is simulated in the form of 18 mine-and-fill numerical model steps (i.e. each step represents six stopes); see longitudinal strike view in Figure 4. While doing so, the strength-to-stress ratio is monitored on level 5000 at the intersection of 3150 haulage drift with the crosscut at the 2981 location. The rock mass qualities of the main units are summarized in Table I (Bewick et al., 2009). The physical and geomechanical properties of rock mass used in numerical modelling for each geological unit are presented in Table II (Vale, 2009).

Figure 2—Methods of dealing with uncertainty in the model input parameters

Case study Vale’s Garson Mine is located in the greater Sudbury area of Ontario, Canada. The mine has been in operation for more than 100 years and has produced 57.2 Mt containing an average grade of 1.33% copper and 1.62% nickel (Vale, 2009). The current production is 2400 t/d. Both transverse and longitudinal stope mining methods are employed with delayed backfill in an upward direction. The typical planned stope dimensions are 30×15×12 m (height × length × width: 100 × 50 × 40 ft.). The stopes are extracted in two or three blasts and then tight filled with a mixture of pastefill and waste rock. In the area of the current investigation, two orebodies are found – namely, No. 1 Shear (east and west) and No. 4 Shear (east and west), which have a strike length

▲

436

JUNE 2014

VOLUME 114

Figure 3—Zoom-in view of No. 1 SHE orebody on 5000 level shows the planned stopes and intersection 2981 under study The Journal of The Southern African Institute of Mining and Metallurgy

Risk indexing tool for mine planning

Figure 4—Longitudinal strike view of No. 1 Shear east for four production levels (5100 L to 4800 L) with planned mining steps

Table I

Major rock types and their geomechanical classification (Bewick et al., 2009) Geological unit

Norite Greenstone South limb dyke North limb dyke Massive sulphide (ore) Metasediment

Rock designation index (Q’) range

Geological strength index (GSI) range

11–33 5–17 No observation 20-50 30–38 0.4–2

70–80 65–75 55–75 (estimated) 90–100 65–75 20–35

It should be noted that the value off 0.027 MPa/m represents the average weight of overburden of the different rock masses. This value is determined by MIRARCO to estimate the in situ stresses on 4900 level. Table II, on the other hand, reports the unit weight of each geological unit in the vicinity of the case study. The values of applied boundaries pre-mining stresses are determined by trial and error using the boundary traction method (Shnorhokian, Mitri, and Thibodeau, 2013). In this method, only the bottom of the model is fixed in the -Z direction. The upper boundary of the model (ground surface) is simulated as free surface where the overburden stress (gravitational stress) is applied. Major horizontal stress is applied on the right-hand side and left-hand side of the model in the +X and -X directions. In addition, a minimum horizontal stress is applied on the front and back of the model in the direction of +Y and -Y (Shnorhokian, Mitri, and Thibodeau, 2013). This technique is adopted in order to obtain the same measured initial stresses undertaken by MIRARCO on 4900 level. Pre-mining (initial) stresses are those stresses that would exist prior to any excavations being made on the model boundaries. The calculated in situ stresses from the above relations are then compared with those computed by FLAC3D, as listed in Table IV, on the same location where measurements were

Table II

Geomechanical rock mass properties for Garson Mine (Vale, 2009) Rock mass

Calibration output properties C

σt

MPa

Deg.

MPa

GPa

5.5 5.1 6.0 4.3 5.7 1

52.7 52 55.5 46.7 54.9 30

0.68 0.53 0.60 0.56 0.51 0.01

56.4 45.5 86.3 43.8 65.0 0.01

kg/m3

Deg.

2920 2780 3000 4530 3170 2000

13.18 13.0 13.88 11.68 13.73 7.50

Model calibration

➤ Minimum principal stress (σ3) = 0.027 MPa/m oriented vertically ➤ Maximum principal stress (σ1) = 1.8 * σv oriented flat on an azimuth of 70º ➤ Intermediate principal stress (σ2) = 1.1 * σv oriented flat on an azimuth of 340º. The Journal of The Southern African Institute of Mining and Metallurgy

Norite Metasediments Olivine diabase Massive sulphide Greenstone Backfill

0.25 0.24 0.26 0.30 0.23 0.30

Table III

Measured in situ stresses by MIRARCO (MIRARCO and Golder Associates, 2008) Principal stress

Magnitude, MPa

Trend (º)

Plunge (º)

72 45 40

70 162 157

02 44 -46

σ1 σ2 σ3

Table IV

Comparison between FLAC3D and measured stresses Principal stresses σ3 σ1 σ2

MIRARCO stresses, MPa

FLAC3D stresses, MPa

Difference, %

40 72 45

39.30 73.21 42.86

2.58% -0.84% 3.40%

VOLUME 114

JUNE 2014

437

▲

The model is calibrated through deep analysis using data from the measured in situ stresses in the mine (McKinnon, 2001). A series of in situ stress measurements were done at Garson Mine in 2005 by MIRARCO in the 2670 crosscut off by the 4900 level ramp (MIRARCO and Golder Associates, 2008). Five of the six tests in the norite were successful, but tests in the dyke could not be successfully completed (McKinnon, 2001; MIRARCO and Golder Associates, 2008; Maloney and Cai, 2006). It is assumed that the major and intermediate principal stress components act in the horizontal plane while the minor principal stress component is vertical (Perman, 2011). The in situ stress measurements were taken on 4900 level and are presented in Table III. Table III shows the values of σ2 and σ3 are close in magnitude, thus limiting the accuracy of the orientations. Therefore, it will be assumed that the minimum principal stress is oriented vertically and equal in magnitude to the weight of the overburden. The resulting stress values used in the calibration model are as follows (MIRARCO and Golder Associates, 2008; Perman, 2011):

Risk indexing tool for mine planning undertaken by MIRARCO. From Table IV, the maximum difference between the measured and modelled in situ stresses is acceptable (e.g. 3.4%). It is clear that the numerical modelling results satisfy the measured in situ stresses.

Risk analysis procedures Risk analysis is an integrated part of decision-making. Risk can be defined as the expected consequence, as opposed to expected loss of utility, or the product of probability of unsatisfactory performance and the cost of consequence, as given in Equation [1]: [1] where Pf is the probability of unsatisfactory performance CC is the cost of consequence. According to the above definition of risk, the probabilistic analysis is carried out and discussed in detail in the next section. The estimated probability of unsatisfactory performance is then rated and categorized according to Table V (Abdellah et al., 2012). The cost of consequence, CC, is calculated in dollars and is based on the hypothesis that the intersection has collapsed or has become unusable regardless of the probability of unsatisfactory performance at that intersection. To calculate the CC, three scenarios for re-opening the intersection are studied and the one with the lowest cost is retained as the ‘cost of consequence CC’. In the current study, scenarios such as loss of life and loss/damage of equipment at the intersection were precluded. The three scenarios are discussed in detail later. The ratings and rankings of the cost of consequence are defined in Table VI. Then the risk is calculated as defined in Equation [1]. The resulting risk is called ‘Risk-index’. It has a rating from 1 to 25 and helps to decide temporally (when) and spatially (where) secondary support is required. The risk matrix is shown in Figure 5.

Probabilistic analysis Probabilistic methods provide a rational and efficient means of characterizing the inherent uncertainty that is prevalent in geotechnical engineering. There is high uncertainty in mine design based on such parameters because of the inherent uncertainty associated with parameters such as in situ stress fields, rock mass properties, and geological features around the openings. Due to the high level of uncertainty, there is a

need to develop stochastic analysis techniques capable of defining the statistical variation of model input parameters and to enable better understanding of the risk associated with choosing the design parameters based on uncertain input data. Hence, predicting the probability of unsatisfactory performance using probabilistic analysis approaches together with the developed numerical modelling (deterministic techniques) becomes necessary. While the stability of haulage drifts may depend on a wide range of factors, including shape and size (due, for example, to drilling inaccuracy and blasting quality), mining depth, in situ stress regime, orebody dip, and the standoff distance to the orebody among others, the focus of the probability analysis in this study is on the inherent variability of the footwall rock strength parameters.

Modified point-estimate method (PEM) The point-estimate approach was first developed by Rosenblueth in 1975 and subsequently modified by Zhou and Nowak in 1988. It proposes predetermined points in the standard normal space to compute the statistical parameters of a function with multiple random variables X, by the 2n2+1 formula where n represents the number of random variables (Zhou and Nowak, 1988; Peschl and Schweiger, 2002). In our investigation, three random variables are used, namely the cohesion, friction angle, and modulus of elasticity of the footwall rock mass. Therefore the required number of simulations is 2(3)2+1 = 19, regardless of the mining sequence. The PEM requires the mean and variance to define the input variables. The term ’failure’ has a very general meaning here as it may indicate collapse of a structure or in a general form define the loss of serviceability or unsatisfactory performance associated with the performance function. To minimize simulation time (a single run takes 30 hours), the modified PEM of 2n2+1 is adopted in this analysis. Table VI

Assumed ratings of the cost of consequence Rating

Ranking

Cost of consequence

1 2 3 4 5

Low Minor Moderate Major Severe

<$10 K $ 10 K- $100 K $100 K-$1M $ 1 M – $ 10 M > $10 M

Table V

Rating and ranking of probability of unsatisfactory performance (Abdellah et al., 2012) Rating Ranking

P(f)

Rare

< 5%

May occur in exceptional circumstances

Unlikely

5%-20%

Could occur at some time

Possible 20%-60%

Might occur at some time

Likely

Certain

▲

438

60%-85% Will probably occur in most circumstances >85%

JUNE 2014

Expected to occur in most circumstances

VOLUME 114

Figure 5—Risk matrix chart as design tool for mine planning and decision-making The Journal of The Southern African Institute of Mining and Metallurgy

Risk indexing tool for mine planning In order to assess the stability of the intersection, a performance criterion must first be selected. This may be one of a large number of conditions such as maximum permissible floor heave ratio (e.g. the ratio of the floor heave to the span of the drift) or roof sag ratio (e.g. the ratio of the roof/back sag to the span of the drift), or allowable stress concentration factor (normally associated with linear elastic analyses), or a yielding condition such as Mohr-Coulomb or Hoek-Brown (Abdellah et al., 2012; Zhang and Mitri, 2008). The choice of a performance criterion is dependent on the application and field observations. In the current study, very little deformation is observed at the intersection in response to mining activities, as shown by the extensometer installed at the intersection, thus making it difficult to set a threshold. Details of the field monitoring programme and instrumentation results are beyond the scope of this paper. Therefore, a yield-based criterion has been selected in which MohrCoulomb is used as the failure condition. Adopting a safety factor of 1.4, it was decided that the stability performance of the intersection is considered satisfactory when the MohrCoulomb strength-to-stress ratio is equal to or greater than 1.4. Strength-to-stress ratio is a built-in function in FLAC3D. It should be noted that the numerical analysis carried out is elastic-perfectly plastic. Thus, as a yielding-based function is adopted (e.g. strength-to-stress ratio), yielding theoretically occurs at a ratio ≤ 1. Due to the large size of the three-dimensional model and in order to optimize the computer storage requirements, only the area around the drift has been discretized with a fine mesh based on a mesh sensitivity analysis. Mesh size defines the number of zones in the grid. The dense mesh offers the advantage of accommodating high stress gradients around the underground mine openings and adequate progression of plasticity (Zhang and Mitri, 2008). According to ITASCA (ITASCA, 2009), sizing the grid for accurate results, but with a reasonable number of zones, can be complicated. Three factors should be taken into account when sizing the mesh: (1) a finer mesh (dense) leads to more accurate results as it provides a better representation of high-stress gradients, (2) the accuracy increases as the zone aspect ratio tends to unity, and (3) if different zone sizes are required, then the more gradual the change from the smallest to the largest, the better the results. The mine-wide model created for Garson Mine has about 965 250 zones. Based on the parametric study (sensitivity analyses) that has been conducted, the most influential model input parameters are Young's modulus (E ), cohesion (C ), angle of internal friction (ϕ), and horizontal-to-vertical stress ratio (K). K is not considered in our probabilistic analysis as the in situ stresses were measured in the mine. In this study, only Young’s Modulus (E ), cohesion (C ), and friction angle (ϕ) are considered, with the Mohr-Coulomb strength-to-stress ratio as shown in Table VII. The modified point-estimate method (Zhou and Nowak, 1988) of (2n2+1) is adopted in this analysis (e.g. 19 runs for the three input variables). The coefficient of variation (COV) is taken as ± 20%.

Stochastic results for No. 2981 intersection (5000 level) The average strength-to-stress ratio for intersection No. 2981 The Journal of The Southern African Institute of Mining and Metallurgy

on 5000 level is given in Table VIII. These values are plotted against mining steps in Figure 6. Figure 6 shows that the strength-to-stress ratio decreases as mining progresses. The strength-to-stress ratio for the roof does not exceed the threshold (e.g. threshold =1.4) up to mining step 5. After that, it deteriorates as mining progresses. Thus, secondary support is recommended at early mining stages. On the other hand, the north wall (NW) does not go beyond the threshold as mining proceeds; it closes to the threshold value only at the final stage.

Table VII

Stochastic properties for footwall (GS) rock mass Rock mass property

Mean

S.D

(COV)

Cohesion (C ), MPa Friction angle (Φ), deg. Young's Modulus (E (E), GPa

5.70 54.90 65.01

1.14 10.98 13.0

0.20 0.20 0.20

Table VIII

Average strength-to-stress ratio for the roof and north wall (NW) of the intersection 2981 on 5000 level (probabilistic analysis) Mining step

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Average strength-to-stress ratio Roof

2.34 2.14 1.87 1.64 1.52 1.37 1.33 1.3 1.24 1.21 1.16 1.12 1.12 1.09 1.08 1.06 1.04 1.02

4.02 3.63 3.12 2.67 2.43 2.14 1.99 1.89 1.82 1.77 1.71 1.66 1.6 1.56 1.52 1.5 1.46 1.41

Figure 6—Average strength-to-stress ratio for intersection 2981 on 5000 level VOLUME 114

JUNE 2014

439

▲

Failure evaluation criteria (performance criteria)

Risk indexing tool for mine planning The probability density ffunction (PDF) is shown in Figure 7. It can be seen that the increase in the area under the curve (e.g. on the left of the threshold line) represents unsatisfactory performance (instability) as mining advances. This indicates increasing probability of instability. The probability of unsatisfactory performance is shown in Figure 8. It can be seen from Figure 8 that the roof of intersection 2981 on level 5000 may require secondary support after mining step 11 since the probability of instability is certain (Pf > 85%). However, the north wall (NW) looks stable and its probability of instability lies between rare and possible (highest probability of instability Pf < 60%). According to Figure 6 and Table VIII, the probability evaluation indicates a requirement for enhanced support after mining step 6 in the roof. However, the north wall (NW) is stable and does not require any secondary support throughout the whole mining sequence. The PDF for the roof at the intersection 2981 after mining step 5 and step 18 is shown in Figure 7. It can be seen from Figure 7 that the probability of occurrence having a value below the threshold (e.g. < 1.4) increases as mining progresses. The probability of unsatisfactory performance (Pf) is shown in Figure 8. It can be seen that the probability of instability for the unsupported roof of intersection 2981 is certain (Pf >85%) after mining step 11. Due to the likelihood of failure, enhanced support is recommended on the roof. For the north wall (NW), the probability of unsatisfactory performance is rare to possible. The cost of consequences associated with intersection instability should be estimated, and is discussed in the next section.

consequence that results ffrom ffailure or blockage off the mine development intersections. If the mine development intersection fails or becomes blocked, then action is necessary to regain access to the stopes. The following three scenarios have been used to choose the most economical technical solution. 1. Mining blocks associated with the failed intersections will be left, as shown in Figure 9 2. The failed intersection will be rehabilitated using secondary support or shotcrete 3. A new bypass will be developed to mine out the stopes associated with failed intersection, as shown in Figure 10. The third option (developing a new bypass) is found to be the most economical solution; therefore it will be presented in details in the following section.

Methodology The following key parameters are taken into consideration when the cost of developing a new bypass is calculated. All the cost values are hypothetical and do not represent actual Garson Mine numbers. 1. Length of developed bypass 2. Cost of development ($1000 per ft.)

Cost of consequences The objective of this section is to estimate the cost of

Figure 8—Probability of unsatisfactory performance (Pf) for roof and north wall (NW) of intersection 2981-5000 level

Figure 7—Probability density function (PDF) for the roof of intersection 2981-5000 level

▲

440

JUNE 2014

VOLUME 114

Figure 9—Plan view showing the unmined blocks associated with failed intersection The Journal of The Southern African Institute of Mining and Metallurgy

Risk indexing tool for mine planning Therefore, f the cost off new bypass development is the sum of lost interest value of money (only unmined stopes related to failed intersection) due to the delay of production to develop the new bypass and the cost of the development itselff (length of the bypass). The cost of development of the new bypass for the studied intersection (no. 2981) is given in Table IX. Table IX shows the total cost of developing the new bypass is about $1.097 million. The rating and ranking for the cost of consequence to the intersection under study are given in Table X, according to the assumed ratings and rankings of the cost of consequence given in Table VI.

Risk-index estimation

Figure 10â&#x20AC;&#x201D;Plan view showing the new bypass developed to access the unmined blocks

3. Mineral value of unmined blocks due to intersection failure (the mineral value is $200 per ton, the operating cost is $90 per ton) 4. Time value due to delay of production (developing time/bypass = 2 months) 5. Operating cost (three men plus truck haulage = $900 per shift) 6. Interest rate = 10% per annum 7. Shift length = 12 hours (actual shift length =10 hours) 8. Advance rate = 20 ft. per day 9. Average ore density = 4.53 t/m3. First, the cost of consequence of unmined stopes (CC1) due to intersection failure or blockage is calculated as follows: [2] Then the (lost interest) time value (CC2) due to delay of production can be calculated as:

The risk-index calculation is given in Table XI. It can be seen that the risk-index on the roof of the intersection 2981 on 5000 level is high to extreme. Thus, secondary support is highly recommended, especially prior to mining step 7. For the north wall (NW) of the intersection, the risk is moderate to high. Therefore, secondary support may be recommended especially at later mining steps. The risk-index with respect to mining steps is shown in Figure 11. The risk matrices for the roof and north wall (NW) of intersection 2981 on 5000 level are shown in Figures 12 and 13 respectively.

Conclusion This paper presents a simple methodology to estimate the risk-index of an intersection. A geotechnical risk assessment scheme is used for deciding when and where secondary support is required with respect to planned mining sequences. This index is the product of probability of unsatisfactory performance and the cost of consequence due to failure of the intersection. A case study is presented where risk-index is estimated for intersection 2981 on 5000 level with respect to the mining activity. The results show that intersection 2981 is a crucial intersection for 5000 level (e.g. the risk-index is extreme in the roof after mining step 7 and high in the north wall after mining step 3). The model is

[3] Also, the development cost for the new bypass is calculated from:

Table X

[4] Finally the total cost of developing the bypass is given as: [5]

Rating and ranking of the cost of consequence of bypass (5000 level) Intersection ID

Total cost, CC4, $ million

Rating

Ranking

1.097

Major

2981

Table IX

Total cost of development of new bypass due to intersection failure Intersection

Development length, ft.

Cost of Development, CC3 ($ million)

Mineral value, CC1 ($ million)

Time value due to delay of Production, CC2 ($ million)

Total cost, CC4 ($ million)

60.9754

1.016

1.097

2981

80.65

0.08065

The Journal of The Southern African Institute of Mining and Metallurgy

VOLUME 114

JUNE 2014

441

â&#x2013;˛

($1000/ft.)

Risk indexing tool for mine planning Table XI

Risk-index calculations for the back and NW of the intersection 2981 at 5000 level Mining step

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Roof

Risk-index

(Pf %)

Rating

13.79 16.35 24.51 36.69 45.22 58.32 62.93 66.64 74.86 78.81 88.88 94.29 93.06 96.41 97.61 99.55 100 100

2 2 3 3 3 3 4 4 4 4 5 5 5 5 5 5 5 5

4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

8 8 12 12 12 12 16 16 16 16 20 20 20 20 20 20 20 20

High High High High High High Extreme Extreme Extreme Extreme Extreme Extreme Extreme Extreme Extreme Extreme Extreme Extreme

Risk-index

(Pf%)

Rating

1.83 4.46 7.93 13.79 18.41 23.89 26.76 30.5 33 34.46 37.45 40.13 43.25 45.22 47.21 48.8 51.6 55.17

1 1 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3

4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

4 4 8 8 8 12 12 12 12 12 12 12 12 12 12 12 12 12

Moderate Moderate High High High High High High High High High High High High High High High High

Figure 11—Risk-index versus mining steps for the roof and north wall (NW) of the intersection 2981 (5000 level)

calibrated with the in situ stress measurements that were undertaken by MIRARCO, and validated with results from deformation monitoring extensometers (MPBX) installed in the study area. The Mohr-Coulomb elastoplastic strength-tostress ratio is adopted as a failure evaluation criterion. The threshold value of this criterion for temporary openings (e.g. mine developments) is taken as 1.4. Thus the performance of the mine development intersection is considered unsatisfactory when the strength-to-stress ratio falls below 1.40. The probabilistic analysis, in combination with numerical modelling, is necessary to account for the inherent uncertainty associated with the rock mass properties. The cost of consequence scenarios provide comparative information for evaluating alternative solutions if the intersection fails.

Recommendation for future work Figure 12—Risk-index matrix for the roof of the intersection 2981 (5000 level)

▲

442

JUNE 2014

VOLUME 114

The stochastic analysis techniques developed can be used in the future for other mining applications such as pillar stability (diminishing ore pillar) or the probability of The Journal of The Southern African Institute of Mining and Metallurgy

Risk indexing tool for mine planning BEWICK, R.P., VALLEY, B., RUNNALLS, S., WHITNEY, J., and KRYNICKI, Y. 2009. Global approach to managing deep mining hazards. ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, 9–14 May 2009. Diederichs, M.S. and Grasselli, G. (eds.). d EINSTEIN, H.H. 1996. Risk and risk analysis in rock engineering. Tunnelling and Underground Space Technology, vol. 11, no. 2. pp. 141–55. ELLEFMO, S.L. and EIDSVIK, JO. 2009. Local and spatial joint frequency uncertainty and its application to rock mass characterisation. Rock Mechanics and Rock Engineering, g vol. 42. pp. 667–88. HAMMAH, R.E., YACOUB, T.E., and CURRAN, J.H. 2008. Probabilistic slope analysis with the finite element method. 43rd US Rock Mechanics Symposium and 4th US-Canada Rock Mechanics Symposium, Asheville, NC, 28 June–1 July 2009. ITASCA. 2009. Fast Lagrangian analysis of continua in 3 dimensions (Flac3d),

Figure 13—Risk-index matrix for the NW of the intersection 2981 (5000 level)

User's Manual Ver. 4.0. 4th edn. Itasca Consulting Group Inc., Minneapolis, Minnesota. MALONEY, S. AND CAI, M. 2006. In situ stress determination, Garson Mine. Project Report: 06-015. Geomechanics Research Center, r MIRARCO, Sudbury, Ontario. MCKINNON, S.D. 2001. Analysis of stress measurements using a numerical

rockburst or fault slip occurrence. The risk methodology can also be used for future feasibility studies to determine the ideal location of the haulage drifts with respect to mining methods and sequence.

model methodology. International Journal of Rock Mechanics and Mining Sciences, vol. 38. pp 699–709. MIRARCO AND GOLDER ASSOCIATES. 2008. Garson Mine geomechanical study. Draft report submitted to Vale Inco Limited, Garson Mine.

Nomenclature

PERMAN, F., SJÖBERG, J., and DAHNÉR, C. 2011. Detailed three-dimensional stress

C: Cohesion of the rock mass

analysis of complex orebody geometry – model setup and results for the

γ: Unit weight of the rock mass υ: Poisson’s ratio E: Young Modulus of rock mass Φ: Friction angle of rock mass ψ: Dilation angle of rock mass ∂ t: Tensile strength of rock mass σ COV: Coefficient of variation (COV = μ ) σ: Standard deviation (SD) μ: Mean value of rock mass n: Number of variables Pf: Probability of unsatisfactory performance PDF: Probability density function

Malmberget Mine. Continuum and Distinct Element Numerical Modeling in Geo-Engineering, g Melbourne, Australia, 14–16 February 2011. Itasca International Inc., Minneapolis. Paper 02–04. PESCHL, G.M. and SCHWEIGER, H.F. 2003. Reliability analysis in geotechnics with finite elements- comparison of probabilistic, stochastic and fuzzy set methods. ISIPTA ’03. 3rd International Symposium on Imprecise Probabilities and Their Applications, University of Lugano, Lugano, Switzerland, 14–17 July 2003. vol. 3. pp. 437–451. SHNORHOKIAN, S., MITRI, H.S., and THIBODEAU, D. 2013. Numerical simulation of pre-mining stress field in a heterogeneous rockmass. International Journall of Rock Mechanics and Mining Sciences. In press. STURK, R., OLSSON, L., and JOHANSSON, J. 1996. Risk and decision analysis for

Acknowledgment

large undeground projects as applied to the Stockholm Ring Road tunnels.

This work was financially supported by a research grant from the Natural Sciences and Engineering Research Council of Canada in partnership with Vale ltda. The authors are grateful for their support.

Tunnelling and Underground Space Technology, vol. 11, no. 2. pp. 157–64. VALE INCO LIMITED. 2009. Garson Deep Pre-Feasibility Study/Fel 2. Geotechnical Report.

References

ZHANG, Y. and MITRI, H.S. 2008. Elastoplastic stability analysis of mine haulage drift in the vicinity of mined stopes. International Journal of Rock

ABDELLAH, W., MITRI, H.S., THIBODEAU, D., and MOREAU-V VERLAAN, L. 2012.

Mechanics and Mining Sciences, vol. 45, no. 4. pp. 574–593.

Stochastic evaluation of haulage drift unsatisfactory performance using Mineral Engineering, g vol. 4, no. 1. pp. 63–87. The Journal of The Southern African Institute of Mining and Metallurgy

ZHOU, J. and NOWAK, A.S. 1988. Integration formulas to evaluate functions of random variables. Structural Safety, vol. 5, no. 4. pp. 267–284. VOLUME 114

JUNE 2014

◆ 443

▲

random Monte-Carlo simulation. International Journal of Mining and

tough in the extreme

There are few materials that can stand up to the HQRUPRXV ULJRXUV RI LQGXVWULHV OLNH PLQLQJ UHÍ¤ QLQJ and power production. CeramiteÂŽ can. CeramiteÂŽ not only withstands temperatures up to 1500Âş Celsius, it provides unmatched wear resistance, itâ&#x20AC;&#x2122;s the perfect protection for your hardworking machinery and equipment, and a superior safeguard for your workforce. The CeramiteÂŽ wear-protection system takes care of product integrity and productivity levels too. It not only prevents breakdowns and accidents, its quick and easy application and repair processes minimise down time with far fewer, far shorter maintenance stops than those demanded by other wear protection methods, at drastically reduced costs. In harsh operating conditions, only CeramiteÂŽ is tough enough to guard your assets, workforce, output levels and product quality. Let CeramiteÂŽ protect your business.

NO NEED TO REPLACE - EVERLASTING MAINTENACE FOR GAS/SOLID CYCLONES, VESSELS, BINS, PIPES, FAN CASINGS and more. Minimum 15mm thick CeramiteÂŽ protection ction Spot weld mesh Original Surface YEAR 1: INSTALLATION

Manufacturers of CeramiteÂŽ

YEAR 3: ARBITRARY WEAR POINT

Installation Contractor

Visit www.processconsultants.co.za for data sheets and case studies or contact us on 011 803 2013, 082 657 2610 or processconsultants@mweb.co.za

YEAR 4: PATCH TROWELLED TO EXISTING THICKNESS

PROCESS CONSULTANTS CONSULTANTS TO THE METALLURGY INDUSTRY

Manufacturing and Distribution Licensee

Dealing with open fire in an underground coal mine by ventilation control techniques by N. Sahay*, A. Sinha*, B. Haribabu†, and P.K. Roychoudhary†

Open fire in coal mines is one of the most serious threats to miners, as well as to the mine. Open fire can often be effectively dealt with by prompt local action, otherwise it very quickly becomes uncontrollable. In one incident, none of the available open fire control technologies, viz., water deluge and sprinkler systems, high-expansion foam, high-pressure high-stability nitrogen foam, water misting, and ventilation and pressure control techniques, were effective for saving the mine without sealing from surface, since the fuel-rich environment prohibited underground access due to the methane explosion hazard. The authors have developed a methodology for dealing with advancedstage open fires underground by the application of a modified ventilation control technique. It is based primarily on a better understanding of the behaviour of open fires, proper diagnosis of the problem, application of judicious ventilation control techniques, and selection of suitable fire indices for assessing the status of an open fire. This methodology was used to successfully control an open fire in Surakachhar 3 and 4 incline mine Surakachhar, central India. The fire area was sealed underground and production subsequently resumed in record time. The paper discusses the behaviour of open fires, particulars of the mine, diagnosis of the problem, experimentation methods, and the results obtained. Keywords spontaneous combustion, high-pressure hig-stability nitrogen foam, water misting technology, ventilation and pressure control techniques, oxygenrich and fuel-rich fire.

Introduction Occurrence of open fire in underground coal mines is one of the most feared hazards throughout the world. During the 18th and 19th centuries, enormous losses of life and property occurred due to mine fires and explosions. Mine fires are mainly caused by sluggish ventilation, frictional heat, electrical sparks, or spontaneous combustion in coal heaps lying in airways that form a part of the active ventilation system. A mine fire can pollute the entire underground atmosphere in a very short time, resulting in loss of life and sometimes closure of the operation. Recent occurrences of open fire in Indian coal mines are Noonidih-Jitpur mine, M/s Steel Authority of India Limited, Burnpur, (2007), Kunustoria colliery, M/s Eastern Coalfields Limited, Sunctoria (2009), Majiri mine No. 3, M/s Western Coalfields Limited, (WCL), Nagpur (2010), Anjan Hill mine, (2010), Bartunga The Journal of The Southern African Institute of Mining and Metallurgy

Methods for controlling open fires The techniques and methodologies for controlling open fire may be broadly divided in two groups, viz., direct and indirect methods.

Direct methods The methods for direct fighting of open fire, such as water deluge and sprinkler systems (McPherson, 1993a, ch. 21), can be very effective in areas close to fixed equipment and activated by thermal sensors. A major difficulty in subsurface firefighting is the limited reach of water jets imposed by the height of the airway. It has been projected that a line pressure of the order of 800–1400 kPa would be required to increase the effective range of a typical water spray to 30 m. High-expansion foam (McPherson, 1993b) containing large volumes of water-based foam can provide a valuable tool for fighting fires in enclosed spaces such as a single blind gallery. The bubbles are generated by a fan that blows air through a fabric net stretched across the

* Mine Ventilation Discipline CIMFR, Dhanbad (Jharkhand), India. † SECL, Bilaspur, India. © The Southern African Institute of Mining and Metallurgy, 2014. ISSN 2225-6253. Paper received r Aug. 2012; revised paper received Feb. 2014. VOLUME 114

JUNE 2014

445

▲

Synopsis

mine (2010), and 3 & 4 Incline mine Surakachhar (2011), M/s South Eastern Coalfields Limited, Bilaspur, 3 & 4 Pits Ballarpur colliery, WCL(2012), Nagpur. The efficacy of the measures and technologies for controlling open fire are mostly dependent on the location, status, and duration of the fire. intensity and duration are also dominating factors for the better efficacy of the measures. Effective techniques require a better understating of the strengths and limitations of each measure and the behaviour of the fire, and a realistic work programme based on the extent and rate of progress of the fire.

Dealing with open fire in an underground coal mine by ventilation control techniques diffuser. ff The net is sprayed continuously with a mixture off water and foaming agent. Ammonium lauryl sulphate foaming agent mixed with carboxyle-methyl-cellulose improves the stability of the bubbles. A major limitation is obstructions caused by roof falls, which are quite liable to occur during fighting of a large underground fire. A foam plug may not be able to climb over such obstructions with the available ventilating pressure. High-pressure high-stability nitrogen foam (HPHS) (Voracek, 1994) with cooling, inertizing, and inhibiting properties is effective when applied upstream of the fire location under a controlled air flow rate. HPHS nitrogen foam in conjunction with the chamber method of ventilation was applied for successful control of an open fire in the goaf of a longwall panel at 1&2 Incline mine Jhanjra, ECL (Sahay et al., 2001) under the aegis Coal India Limited, Kolkata, India. Another underground open fire close to a downcast shaft in one of the intake airway at Noonidih-Jitpur colliery (Sahay et al., 2008) was also controlled by application of HPHS nitrogen foam in conjunction with regulation of the ventilation control to avoid methane accumulation. Water misting technology (Liu and Kim, 2000), based on wetting of air flowing to the seat of fire through adding water droplets of the order of 400 μm or less in size to extract heat directly from the seat of the fire by evaporation with the creation of an oxygen-deficient environment near the seat of the fire, is an added advantage. Water mist does not behave like a true gaseous agent in fire suppression. The effectiveness of a water misting system is dependent on spray characteristics like the size distribution of the droplets, flux density, and spray dynamics with respect to the fire scenario. Due to the complex extinguishing process the relationship between a fire scenario and the characteristic of a water misting system is not understood well enough to design a compatible water misting system. The efficacy of these technologies is location- as well as capacity-dependent. The water misting technology has yet to be applied in Indian coal mines, although its efficacy was found to be better than that of HPHS nitrogen foam when it was applied from the close vicinity of a fire during an experiment in a mine fire model gallery (MFMG) (Singh et al., 2007).

Indirect methods Ventilation control techniques (McPherson, 2003a, ch. 21) have three types of effect on fire: (i) the combustion process and distribution of products of combustion, (ii) direction and rate of propagation of the fire, and (iii) air flow distributions in other parts of the mine. In controlling of air flow rate in an open ventilation circuit, the probability of shifting the environment of the fire area from oxygen-rich to fuel-rich is increased due to less dissipation of convective heat. Moreover, the status of an open fire is very sensitive to air velocity. Ventilation control techniques therefore required to be applied judiciously in the case of a complex ventilation network. In the pressure control technique (Kissell and Timko1991), airways parallel and adjacent to the fire path are kept free from fire gases by maintaining higher atmospheric pressure with the help of brattice stopping erected in crosscuts to facilitate the application of water sprays into the fire path. Devices such as the ‘parachute

▲

446

JUNE 2014

VOLUME 114

stopping’ or ‘inflatable f seal’ have been developed to replace brattice cloths in such circumstances. These can be erected quickly and give improved seals around the perimeter of the airway. A consequence of this technique is that the air flow over the fire is increased to an extent that depends upon the configuration and resistances of the local airways. Pressure differentials between airways can also be modified by the use of a temporary fan instead of a restriction in the adjacent airway. In this case, air flow over the fire will be reduced. The location of and pressure developed by the fan is required to be selected with care in order to avoid aggravation of the fire. In many cases fire is located at a remote place in the mine and is detected only at the advanced stage, while main return airways become badly polluted with hazardous gases. This necessitates suspending the underground deployment of personnel due to the methane explosion hazard. In this situation two options are adopted. One option is sealing of the mine entries from surface followed by application of measures for controlling the fire followed by recovery of the mine, Moreover, the danger of methane explosion is always associated with the simultaneous sealing upstream and downstream or only downstream of an open fire, particularly in a fuel-rich environment. Methane explosions have occurred (Urosek et al., n.d.) 30 hours and 77 hours after sealing of open fires in longwall panels in coal mines at Delta Colorado and Morgantown, West Virginia respectively. At the same time, sealing downstream of an open fire is difficult and risky due to the hot and hazardous environment that prevails after sealing upstream of the fire location. Hence sealing of the area affected by open underground fire by conventional approaches is a risky proposition. Conversely, sealing of the mine from the surface is time-consuming and costly, causing suspension of production for a long period. Sometimes the mine, once sealed off due to fire, has to face problems of waterlogging and deterioration of roof conditions. In the light of the above, to overcome the constraints and limitations of various open fire control techniques, the authors have developed a methodology for dealing with open fire in critical situations and the safe isolation of the fireaffected area by a modified ventilation-control technique. The sequence of measures is sealing at a strategic location upstream of the fire to short-circuit the air flow adjacent to the sealing, stagewise reduction of fan pressure and barometric pressure over the fire while closely monitoring the status of the fire, followed by sealing of the fire area underground. This technique was applied for controlling an open fire at 3 and 4 Inclines Mine Surakachhar, (SECL), Bilaspur, situated in central India. The behaviour of the fire, particulars of the mine, and diagnosis of the problem, experimentation, and results are discussed in subsequent sections.

Behaviour of underground fires Fires in underground coal mines are classified broadly into two groups, viz. open and concealed fires (McPherson, 1993b). Open fires are accompanied by flaming combustion because of the availability of sufficient oxygen in the ventilation circuit. Conversely, concealed fires occur in areas that are inaccessible such as caved goaf or abandoned zones or sealed-off area. A closed fire can also become an open fire after crossing over the seals/stoppings. The behaviour of open fire depends on factors such as the nature and amount The Journal of The Southern African Institute of Mining and Metallurgy

Dealing with open fire in an underground coal mine by ventilation control techniques off fflammable material, ventilation system arrangement, the duration of the fire, the extent of the spread of combustion products, and ignition location. During initiation of an open fire the available oxygen is utilized in the formation of principally carbon dioxide and carbon monoxide. The chemical reactions that take place in the oxidation zone are (Pandey et al., 2003): [1]

C+O2 = CO2 + 24.03 Kcal/g mole C+ ½ O2 = CO + 26.42 Kcal/g mole CO+H H2O = CO2+H H2 + 9.8 Kcal/g mole

[2] [3]

All these reactions are exothermic, and raise the temperature of the coal to a high value. In this situation, the air temperature downstream of the fire increases due to the convective and radiative heat carried by the air current. Depleted of free oxygen, the fire situation moves into the reduction zone, where CO2 is reduced to CO in contact with hot coal. Where water is present in the fire zone, steam is produced, which reacts with the hot coal and produces additional combustibles gases and frequently water gas. The chemical reactions in this zone are believed to be: C+ CO2 = 2 CO + (-) 41.2 Kcal/g mole [4] C+ H2O = CO+H H2 + (-)31.4 Kcal/g mole [5] C+2H H2 = CH H4 +17.9 Kcal/g mole [6] The heat generated during reactions [1]–[3] in the oxidation zone is neutralized by endothermic reactions [4] and [5] in the reduction zone. Hence the area downstream of the fire zone initially contain hot combustibles and is deficient in oxygen. The progress of the reaction from oxidation to reduction may be identified by the increasing trend of CO, H2, and CH4 in air samples, followed by a decreasing temperature trend.

Effect of ventilation on open fire The effect of ventilation on open fire is a complex phenomenon. Initially, ventilation is influenced by an open fire. An open fire causes a sharp increase in the temperature, resulting in expansion of the air and the addition of combustion products and evaporated water (Gillies, Wu, and Humphreys (2004). The expanded air moves in both directions along the airway, which reduces the air flow rate in an upstream direction and increases the air velocity downstream of fire location, causing an additional pressure loss. This is known as the choking or throttling effect. The choking effect is analogous to increasing the resistance of the airway, and is increased in the case of a single airway. The immediate effect of heat on the ventilating air stream is very

(a) Oxygen-rich open fire in MFMG

localized (Litton, Derosa, and Li, 1987). The reduced density causes the mixture of hot air and combustion products to rise and flow preferentially along the roof of the airway, and under low air velocity in a level or descending airway they back up against the direction of air flow. In this way the air flow to the fire and discharge of fire gases become bidirectional. This situation may also prevail in case of choking upstream of the fire location. Similar results have also been observed during experiments (CIMFR, 2004; Singh et al., 2004) carried out in a 65.5 m mine fire model gallery (MFMG) at the Central Institute of Mining and Fuel Research, Dhanbad, India. This facility, which is equipped with stateof-the-art instrumentation to study the behaviour of open fires at 30-second intervals, is used to assess the efficacy of fire control measures. The experiment was conducted with 10 cm thick coal slab layered on the inner wall of an archedshaped gallery of cross section 5.78 m2 up to 22.0 m by establishing an air velocity of 1 m/s with the help of axial flow fan (AF-50) installed at the other end of the gallery. Photographs of the experiment are shown in Figure 1. Initially, the air flow was unidirectional (Figure. 1a). During the experiment it was observed that the lower portion of the gallery (from the floor to a height of about 1.4 m) from entry to the fire location was acting as intake air, while upper portion of the gallery was acting as a return (Figure 1b), despite the fan in operation, after about two hours from the initiation of the fire. It was also observed that hot gases containing coal dust started burning at the entry (Figure 1c). Temperatures in the fire zone between 800°C and 1058°C were recorded during the experiment. A similar phenomenon may occur in a mine. This phenomenon is more probable in the case of a single or blind gallery. The availability of oxygen at the fire site regulates the development of the fire and the shifting of the environment from oxygen-rich to fuelrich, as indicated by a sharp decrease in oxygen percentage and rise in concentration of CO, CH4, and H2 in air samples. In case of fire in a fuel-rich environment the value of the Jones and Tricket ratio (JTR) lies between 1 and 1.5 (Banerjee, 2000). This is a serious progression and makes it difficult for firefighters to take proper decisions.

Diagnosis of problem and experimentation Surakachhar 3 & 4 Incline Mine belongs to M/s SECL, and is situated in Bilaspur in the Korba district of Chhattisgarh State in central India. The mine produces approximately 500 t of coal per day and deploys 1200 persons in the largest shift.

(b) Stage 1 Fuel-rich open fire in MFMG

(c): Stage 2 fuel-rich open fire in MFMG

The Journal of The Southern African Institute of Mining and Metallurgy

VOLUME 114

JUNE 2014

447

▲

Figure 1—Combustion experiment in the mine fire model gallery (MFMG) at the Central Institute of Mining and Fuel Research, Dhanbad

Dealing with open fire in an underground coal mine by ventilation control techniques Particulars of the mine In the mine there are two seams viz., G-1 and G-3. The lowermost seam is G-1, which is 1.5–3.0 m thick and has a gradient of 1:9.5, dipping S-15°W, with reserves of 2.554 Mt. The mine has been developed by the bord and pillar method along the G-1 seam, at depths from 46 m to 240 m. The extent of development in N-S direction along dip is up to 72L and in an E-W direction along strike is from 3D up to the mine boundary. Gassiness of the seam is Degree-II as per Coal Mines Regulations-1957, India. The G-1 seam (Figure 4) is accessed from surface through two inclines. viz., Inclines 3 and 4, and an air shaft. The No. 3 incline connected with 0D and No.4 connected with 1DW are equipped with conveyor belt and haulage respectively. There are two major faults on the south side of the mining area. A downthrow fault F3-F3 with 8.5 m displacement intersects the seam between 29L and 32L and an upthrow fault F4-F4 with 30 m displacement intersects the seam between 38L and 39L. These faults thus create a natural basin formation of the seam between 29L and 39L. The airways, viz., 0D, 1DW, and 1DE, on both sides of faulted zone are connected through drifts. The roof of the seam consists of alternating shale and sandstone whereas the floor consists of shale and coal. The immediate roof of the airway consists of a coal layer about 0.5 m–0.8 m in thickness. There is one depillaring district (W-5B) in the west side of the mine and a pumping station on 62L in 0D. On the downdip side there are two boreholes (depth 116 m, diameter 100 mm) at 40L/1DE drilled from surface. One borehole is for laying electrical cables and another for discharging of underground water. The mine ventilation is via an exhaust system, achieved by an axial flow fan (SIWAX-APG-2159) of capacity 67 m3/s at a pressure of 650 Pa installed on the surface and connected with the air shaft through a fandrift. There is another fandrift connected with the same airshaft with proper locking arrangements. Prior to the fire, intake air entered the mine through Inclines 3 and 4 and flowed along 0D and 1DW up to 25L, where it was split in two streams. The majority of the intake air (about 45.0 m3/s) was directed along 25L and 26L to ventilate the depillaring district (W5B), while a more restricted part of the intake air (about 11 m3/s) flowed through 0D and 1DW to ventilate the pumping station. To control air flow rate on the dip side below 25L a regulator was installed on 62L beween 0D and 1DE. The return air from the pumping station flowed through a single airway 1DE. In the ventilation circuit doors at 12L and 27L between 0D and 1DE are provided. Other details are shown in schematic diagram of the ventilation network of the mine (Figure 2) A resistance of the mine in SI unit using Atkinson’s equation was calculated of the order of 146.25 Ns2/m8. The fan and mine characteristics are depicted in Figure 3.

proceed below 28L along 0D and 1DE due to the hot and hazardous environment and poor visibility. In this situation attempts were made to isolate the fire from the main ventilation circuit below 27L by erecting brick stoppings at 0D, 1DW, 1DE, and 2DE between 27L and 28L. Two stoppings, at 0D and 1DW between 27L and 28L upstream of the fire location (intake), were completed. Downstream of the fire location (return airways) partial CGI sheet and sandbag stoppings at 1DE between 27L and 28L could be not erected due to the unbearably hot and hazardous atmosphere at the proposed location. Similarly, the proposed location for erection of stopping at 2DE could not be approached. On 10 September 2011, doors at 27L between 0D and 1DE were partially opened and personnel evacuated from underground. These measures did not yield the desired results, as indicated by continued discharge of CO at a concentration of about 110 ppm through the fan drift. Sealing of the mine from surface at the main entries was inevitable due to apprehension of the damage potential of such an unsealed fire. At this juncture, experimentation was carried out by the authors for controlling the fire and safely isolating the fire

Figure 2—Schematic diagram of the ventilation network of the mine

History of the fire The open fire occurred at about 1.5 km from main mine entries, in 0D of G-1 seam (Intake airways) between 41L and 42L on 8 September 2011 at 1:00 a.m. The fire progressed very rapidly The intensity of the fire can be judged from the fact that the values of Graham’s ratio and oxides of carbon ratio had reached levels of 28.2% and 17.7% respectively. During efforts to locate the fire the rescue team could not

▲

448

JUNE 2014

VOLUME 114

Figure 3—Mine and fan characteristics The Journal of The Southern African Institute of Mining and Metallurgy

Dealing with open fire in an underground coal mine by ventilation control techniques area underground so that early resumption off coal production would be possible. Initially, the behaviour of the fire was studied as discussed in subsequent paragraphs.

Influence of barometric pressure on fire behaviour The behaviour of the fire was studied by measuring the CO gas content in air discharged through the fan drift on 14 and 15 September 2011. The results are represented in Figure 5. The results indicate minimum CO levels between 11:00 pm and 12:15 am when the barometric pressure was maximum, and maximum CO levels during minimum barometric pressure. Hence the fire was responding to diurnal changes in barometric pressure. It was inferred that the life of the fire could be prolonged by the feeding of air due to the diurnal change in barometric pressure.

Identification of chemical reaction zone Air samples were collected through the borehole at 1DE/40L and analysed using a gas chromatograph. The results are presented in Table I. The trends of combustibles generated during the period 10–15 September 2011 were examined. The CO, H2, and CH4 trends during that period are shown graphically in Figure 6. The CO, H2, and CH4 levels rose abruptly after partial isolation of the fire area during the period between 10 and 12 September, and then levelled off. From this it may be inferred that chemical reaction was taking place in the reduction zone (see Equations [1] and [4–6]. The fire was thus in an oxygen-rich environment as the minimum O2 level was 13.3%. It was apprehended that the presence of water, heat, and coal could prolong the fire.

Summary of investigation and observations The results of the investigation and observation revealed that: 1. Mine water gauge was of the order of 650 Pa 2. Air circulation into the mine was of the order of 56 m3/s 3. Double doors at 27L between 3D and 1DE were partially opened providing a parallel airway path to fire area 4. Suction pressure at the seat of fire measured through the borehole against the atmosphere was of the order of 300 Pa. 5. Gas analysis results from fan drift indicated that the fire was responding to diurnal variation in barometric pressure 6. A concentration of CO of the order of 110 ppm was measured at the fan discharge point 7. After erection of stoppings in intake airways (0D and 1DW), the direction of air flow in 1DE and 2DE was found to be bi-directional between 27L and the fire location. The return airway (1DE) alone was acting both for feeding and discharging of air from the fire area 8. The chemical reaction may be considered as being partially in the reduction zone.

Assessment of status of fire The values of the fire indices (Singh et al., 2004) viz., Graham’s ratio, oxides of carbon ratio, and Jones and Tricket ratio calculated from the air sample analyses on 15 September were of the order of 6.67%, 10.88%, and 0.66 respectively. As per the index values of Graham’s ratio and oxides of carbon ratio, the status of fire was considered active.

Figure 5—Gradient of CO concentration with time

The Journal of The Southern African Institute of Mining and Metallurgy

VOLUME 114

JUNE 2014

449

▲

Figure 4—Partial plan of G-1 seam showing location of fire

Dealing with open fire in an underground coal mine by ventilation control techniques Table I

Analysis of air samples collected through boreholes at 1DE/40L Date 10-9-11 11-9-11 12-9-11 13-9-11 15-9-11

CO%

CO2 %

CH4 %

C2H6 %

C2H, %

O2 %

H2%

N2 %

Fan pressure, Pa

0.0347 0.17 0.512 0.5284 0.5259

0.226 1.730 3.710 4.000 4.860

0.0122 0.0674 0.181 0.1762 0.174

0.001 0.0049 0.0092 0.0096 0.0073

0.0004 0.0016 0.0015 0.0003 0.0023

20.6 18.6 13.49 13 13.3

0.0196 0.146 0.33 0.2728 0.259

78.2 78.96 80.84 81 79.94

640 640 640 640 640

0.2253 0.2389 0.127 0.1412 0.0791 0.0644 0.0459 0.042 0.0352

81.56 81.877 82.7 82.93 83.3706 83.1166 83.1464 83.7699 82.25

400 380 360 340 320 300 280 280 280

Reduction of fan pressure on 15 September, 2011 16-9-11 17-9-11 18-9-11 19-9-11 20-9-11 21-9-11 22-9-11 23-9-11 28-9-11

0.4336 0.4598 0.3447 0.3771 0.3106 0.2844 0.2458 0.229 0.0786

3.766 4.145 4.016 4.630 4.416 4.243 4.251 4.480 5.207

0.1162 0.126 0.1277 0.14 0.116 0.1037 0.0942 0.0921 0

0.006 0.0064 0.0066 0.0066 0.0056 0.0053 0.0054 0.0051 0

0 0.0007 0.0003 0.0004 0.0003 0 0 0.0001 0

12.94 12.2 11.71 10.75 10.7392 11.2237 11.2515 10.4145 12.432

Measures taken

Figure 6—Trend of CO, H2, and CH4

Hence it was concluded that after partial sealing, the fire was active in the reduction zone under the influence of barometric pressure and fan pressure. Feeding of air to the fire was possible either through bi-directional flow of air and discharge of gases from the fire in 1DE or leakage of air through the stoppings in 3D and 4D between 27L and 28L.

Design of ventilation parameters To control feeding of air to the fire, the optimum fan pressure was found to be in the order of 280 Pa to establish an air velocity below 0.1 m/s at 27L near fire area and 0.5 m/s in other part of the mine so as to avoid accumulation of smoke and foul gases in the main intake airways. For this purpose, the ventilation circuits of the mine were simulated using ‘Vent’ computer software, and taking the main branches viz. ventilation circuit from surface (INC-3)-0D and 1DW/25L-W5 panel-27L-across door -1DE(27L-10L)- main return to fan drift) and fan system, including the surface leakage path. The results of the simulation studies further indicated that by increasing surface leakage to the order of 65.0 m3/s at the fan, the negative pressure over the fire created by the fan would be reduced to 10 Pa. In this case the fan pressure and air circulation in the mine would be of the order of 300 Pa and 11.0 m3/s respectively Furthermore, erection of a brattice curtain absorbing a pressure of 20 Pa at mouth of INC-3 during the night would reduce the effect of diurnal variation in barometric pressure over the fire by 17 Pa.

▲

450

JUNE 2014

VOLUME 114

Accordingly, fan pressure was reduced from 650 Pa to 400 Pa in the evening of 15 September, then gradually reduced to 280 Pa in steps of 20 Pa a day from 17–22 September by opening airlock in the fan drift of the standby fan. Simultaneously, the effect of the diurnal variation in barometric pressure was neutralized by increasing the mine water gauge by 20 Pa during the night. During this process, generation and burning of methane gas in air samples collected through the borehole were monitored to maintain an oxygen-rich environment in the fire area before reducing the fan pressure. Result of analyses, using gas chromatography, of air samples collected through the borehole during the period 16–28 September are also given in Table I.

Monitoring of status of fire In the case of an open fire, indices like Graham's ratio [CO/ΔO2] and oxides of carbon ratio [CO/CO2] are important, as these are independent of dilution by fresh air subject to variations in carbon monoxide and carbon dioxide that are not caused by the fire. The Jones-Tricket ratio is also considered as a measure of reliability of sample analysis as well an indicator of the type of fuel involved. The trend of saturated hydrocarbons, particularly ethylene, is a useful indicator to assess the heating of coal at low temperature or the combustion of fresh coal. Trending of CO, CH4, H2, and O2 may provide information about whether the environment is oxygen-rich or fuel-rich. The temperature at the seat of the fire could not be measured due to electrical cables laid in the borehole. The status of the fire and environment, including the efficacy of the measures taken, were assessed on the basis of analyses of air samples collected through the borehole at 1DE/40L during the period 10–28 September 2011 (Table I).

Trending of combustibles A close vigil was kept on the status of the fire to optimize the extent and duration of the measures taken. For this purpose, combustibles gases viz. CO, CH4, and H2, from the results of The Journal of The Southern African Institute of Mining and Metallurgy

Dealing with open fire in an underground coal mine by ventilation control techniques analysis off air samples (Table I) were trended graphically as shown in Figure 7. The results reveal that during the period 10–15 September the values of CO, CH4, and H2 increased sharply and then leveled-off at 0.5259 %, 0.1762%, and 0.259% respectively. The results also reveal that after starting the reduction in fan pressure on 15 September the values of CO, CH4, and H2 decreased sharply. At about 280 Pa of fan pressure the decreasing trend of combustibles in the air samples continued even without further reduction in fan pressure. The values of CO, CH4, and H2 in air samples collected on 28 September were of the order of 0.0786%, zero, and 0.0352% respectively.

fire. This ratio, together with Graham’s ratio, can be a better indicator in assessment of the status of an open fire. The IOC value of air samples (Table I) was calculated using

Assessment of status of fire using Graham's ratio

Assessment of status of fire using Jones-Tricket ratio

Graham’s ratio (Ray et al., 2004) [CO/ΔO2] is an important tool for assessing the status of an open fire. It is favored as it is independent of dilution by fresh air. It defines different stages of fire from initiation of spontaneous combustion to the blazing stage. An index value of less than 1.0 indicates a fire in the smouldering state and an index of more than 2.0 a blazing state. Trending of this ratio is essential when dealing with an open fire, and may give information regarding the progress of the fire. The values of the index were calculated from air samples taken during the period 10–28 September using the formula:

The Jones-Tricket ratio (Banerjee, 2000) is used as a measure of the reliability of sample analysis, and also as an indicator of the type of fuel involved in the fire. This ratio (JTR) is also unaffected by inflows of air, methane, or injected nitrogen. It can be used for the assessment of gaseous products in case of fires and explosions. This is a useful pointer to the progression of the fire, rising during the early stages and tending to remain constant during flaming combustion. However, the JTR rises rapidly again to between 1.0 and 1.5 as the fire becomes fuel-rich. In certain types of combustion with 50% conversion from CO to CO2, the JTR may increase to 7 or more. The value of this ratio in the case of <17.0% oxygen in air samples collected from the fire area through boreholes was 0.65 to 0.85. The JTR was calculated from the air sample analysis results (Table I) using the formula

[7] Trends of GR and fan pressure are presented in Figure 8. The results reveal a sharp decrease in index value from 28.21% to 6.45% during the period 10–12 September, followed by a levelling-off to 6.45–6.67% during the period 12–15 September. This may be due to sealing upstream of the fire. Furthermore, GR value followed a decreasing trend during the period 16–28 September and reached to a level of 0.8%, indicating a smouldering state. This may be due to the measures taken.

[8] The trends of IOC and fan pressure are presented in Figure 9. The results reveal a consistent decreasing trend during the period 16–28 September, when the IOC ratio reached a level of 2.63%, indicating a superficial state of the fire.

[9]

Assessment of status of fire using oxides of carbon ratio

Figure 7—Trend of CO, H2, CH4, and fan pressure The Journal of The Southern African Institute of Mining and Metallurgy

Figure 8—Trend of GR and fan pressure

Figure 9—Trend of oxides of carbon ratio and fan pressure VOLUME 114

JUNE 2014

451

▲

The oxides of carbon ratio (IOC) [CO/CO2] (Banerjee, 2000) is also important and favoured for assessment of the status of a fire. The IOC value is also independent of dilution by fresh air or even by nitrogen. The status of the fire on the basis of this ratio is interpreted as: less than 2.0, no fire; between 2.0 and 13.0, superficial state of fire; and more than 13.0, blazing

Dealing with open fire in an underground coal mine by ventilation control techniques The results and ffan pressures are presented in Figure.10. The results reveal that the JTR was of the order of 2.0 on 10 September, indicating a fuel-rich environment. This may indicate fuel-rich combustion conversion of CO to CO2 in the very active stage of the fire. A consistently decreasing trend can be seen during the period from 16–28 September, with the JTR reaching a level of 0.57 indicating an oxygen-rich state of the fire.

Assessment of status of fire by trending of ethylene Previous work carried out by the Bureau (Xie et al., 2011) has shown that the desorption of low molecular weight hydrocarbon gases from coal is strongly temperaturedependent. Changes in the concentration of desorbed hydrocarbons can be detected at temperatures even less than 100°C. This may also indicate the involvement of fresh coal in the fire or the fuelling of the fire by coal detached from the sidewalls and roof. The ethylene levels from air samples (Table I) during the period 10–28 September are presented in Figure 11, together with the fan pressure. An increasing trend of ethylene is evident during the period 10–15 September, followed by a decreasing trend from 16–23 September, and finally zero ethylene concentration from 23 September onwards. The results corroborate the similar finding by reduction in values of ethylene with reduction in fan pressure.

Trending of CO, H2, CH H4, and O2 The trends of CO, H2, CH H4, and O2 are presented in Figure 13. The results show a decrease in the values of CO, H2, CH H4, and O2 from 0.5259%, 0.1762%, 0.259%, and 13.3% to 0.0786%, zero, 0.0352, and 12.43% respectively. Hence the environment of the fire may be considered as being oxygen-rich. From the results of analyses of air samples and the assessment of the status of the fire, it was concluded that the intensity of fire had reduced to a smouldering/superficial state but combustion was continuing. In this situation, sealing of the fire area was called for. On the basis of the results of air samples, the underground environment was assessed as being non-explosive as confirmed by the Explo software based on Ellicott’s extension of Coward’s diagram (Vutukuri & Lama, 1986) for prediction of the explosibility of mine air mixture.

Hydrogen/methane [H H2/CH H4] ratio The hydrogen/methane ratio is an indicator of flaming combustion and increases with temperature. Since temperature could not be measured directly as electrical cables were laid in the borehole, the value of the index (T) was calculated from air sample analyses using Equation [10] The results and fan pressures are presented in Figure 12.

Figure 11—Trend of C2, H4, and fan pressure

[10] Figure 12 reveals that value of T, after reduction of fan pressure, increased to 1.9%, almost equal to the value on 10 September, and then decreased with reduction of fan pressure to a level of 0.45%. This may reflect the reduction in exchange of air after the implementation of control measures and a subsequent increase in the endothermic reaction in the fire-affected area. The trend also indicates a reduction in temperature at the seat of the fire. Figure 12—Trend of ratio (T) and fan pressure

Figure 13—Trend of CO, H2, CH4, O2 and fan pressure

Figure 10—Trend of JTR and fan pressure

▲

452

JUNE 2014

VOLUME 114

The Journal of The Southern African Institute of Mining and Metallurgy

Dealing with open fire in an underground coal mine by ventilation control techniques Finally, personnel under rescue cover were deployed underground on 10 October 2011 to isolate the fire area by erecting stopping at 1DE and 2DE between 27 L and 28L. The task was completed the same day. Subsequently, production in the mine was resumed on 12 October 2011.

Conclusions A modified ventilation control technique was applied to deal with an open underground fire located in a critical situation in Surakachhar 3 & 4 Inclines mine, (SECL), Bilaspur, India. Sealing upstream of the fire, short-circuiting of air flow rate at the nearest location to the seat of the fire, and reduction in the effect of fan pressure and barometric pressure over the fire, based on results from monitoring the status of the fire, were effective in controlling the fire. This methodology is general enough to apply elsewhere under similar situations.

Acknowledgements The authors are thankful to the CSIR-Central institute of Mining & Fuel Research institute (CIMFR), Dhanbad, for permission to publish this paper. The authors also acknowledge the members of Mine ventilation Discipline, particularly Sri A. Ansari, Technician, and Sri R.P. Dasaundhi, Laboratory Assistant, for support in field investigation. We thank the mine management, Sri S.N. Mishra (Manager), Sri Ajoy Kumar (Survey Officer), and Sri Z.H. Khan (General Manger) for their cooperation and support in during the investigation. The guidance and suggestions of Sri B.C. Bhowmick, Ex. Scientist and r Head, Mine Ventilation Discipline, CSIR-CIMFR, Dhanbad, is gratefully acknowledged. The opinions expressed in the paper are those of the authors and not necessarily those of CSIR-CIMFR, Dhanbad and M/s South Eastern Coalfield Limited, Bilaspur, India.

References BANERJEE, S.C. 2000. Early detection of heating and assessment of status of fire in mines, Prevention and Combating Mine Fires. Oxford & IBH, New Delhi, India Chapter-6. p. 165.

CIMFR. 2004. Studies on simulation of open fire in mine fire model gallery under varied airflow rate for suppression of fire and explosion in coalmines. Central Institute of Mining and Fuel Research, Dhanbad,

LIU, L and KIM, A.K. 2000. A review off water mist ffire suppression systems: fundamental studies. Journal of Fire Protection Engineering, g vol. 10, no. 3. pp. 32–50.

MCPHERSON, M.J. 1993a. Subsurface Ventilation Engineering. Chapman and Hall, London.

MCPHERSON, M.J. 1993b. Development and control of open fires in coal mine entries. Proceeding of the 6th US Mine Ventilation Symposium, University of Utah, Salt Lake City, 21-23 June. pp. 197–202.

PANDEY, B.P., CHOUDHARY, P., SINGH, A.K., and MENDHE, V.A. 2003. Laboratory study of channel gasification with steam-air blast in sub-bituminous coal from Pindra Raniganj coalfield. Minetech, vol. 24, no. 6. pp. 37–49.

RAY, S.K, SINGH, R.P., SAHAY, N., and VARMA, N.K. 2004. Assessing the sealed fire in underground coal mines. Journal of Scientific and Industrial Research, vol. 63. pp. 579–591.

SAHAY, N., BHOWMICK, B.C., VARMA, N.K., RAY, S.K., VERMA, S.M., and MONDAL, P.K. 2001. Control of fire in a longwall panel under shallow cover with chamber method of ventilation and high pressure high stability nitrogen foam - a case study. Proceedings of the 7th International Mine Ventilation Congress, Krakow, Poland. pp. 971–978.

SINGH, R.P., RAY, S.K,, SAHAY, N., and BHOWMICK, B.C. 2004. Study on application of fire suppression techniques under dynamic fire condition. Journal of the South African Institute of Mining and Metallurgy, vol. 104, no. 11. pp. 607–616.

SAHAY, N., SINHA, A., CHAKRAVORTY, R.B., PRASAD, P., AHMAD I., AND VARMA, N.K. 2008. Control of open fire in underground at Noonidih-Jitpur Colliery, SAIL [Steel Plant]. Journal of Mines Metals and Fuels, vol. 56, no. 12. pp. 241–248

UROSEK, J.E., BEITER, D.A., STOLTZ, T.T., and FRANCART, W.J. Not dated. Recent United State coal mine fires lessons learned. http://www.msha.gov/ s&hinfo/techrpt/MEO/Recent United States Coal Mine Fires Lessons Learned.pdf.

Jharkhand, India. VORACEK, V. 1994. Use of nitrogen foam for both prevention and suppression of GILLIES, S., WU, H.W., and HUMPHREYS, D. 2004. Case studies from simulating mine fires coal mine and their effect on mine ventilation system. Australasian Institute of Mining and Metallurgy Branch, Coal Operators

spontaneous combustion of coal in Ostrava Karvina Coalfields. Proceeding of the Workshop on Occupational Safety and Environment Protection in Underground Coal Mining Industry, SCZYRK, Poland.

Conference, University of Wollongong, New South Wales. pp. 111–125. VUTUKURI, V.S. and LAMA, R.D. 1986. Explosibility of mine atmospheres and fire KISSELL, F.N. and TIMKO, R.J. 1991. Pressurization of intake escape ways with parachute stoppings to reduce infiltration of smoke. Proceedings of the 5th

gases, Environmental Engineering in mines, Cambridge University press, Melbourne 3166, Australia

US Mine Ventilation Symposium, West Virginia. pp. 28–34. XIE, J., XUE, S., CHENG, W., and WAN, G. 2011. Early detection of spontaneous

ventilation airflow. Report of Investigation RI 9076. US Bureau of Mines. The Journal of The Southern African Institute of Mining and Metallurgy

combustion with development of an ethylene enriching system. International Journal of Coal Geology, vol. 85, no. 1. pp. 123–127. VOLUME 114

JUNE 2014

453

◆ ▲

LITTON, C.D., DEROSA, M.I., and LI, J.S. 1987 Calculating fire-throttling of mine

Empirical observations of dilution in panel caving by R.L. Castro* and P.S. Paredes*

Synopsis The extraction of ore in panel caving operations also involves the coextraction of non-economic material, known as dilution. Forecasting, avoiding, and understanding the mechanics of dilution is critical to longterm planning. The models of dilution that are currently in use in panel caving operations are those by Laubscher (1994), who postulated that dilution is due to gravitational flow and depends on the height of interaction, the column height, and the draw uniformity. This paper presents a back-analysis of extraction and dilution behaviour at Codelco’s El Salvador and Andina Divisions, which use panel caving as a mining method. This analysis was conducted in order to review the existent dilution entry mechanisms and calculation methods. Three sources of dilution were identified. The first source is ‘gravitational dilution’, which is due to gravitational flow and mixing as dilution migrates by gravity from a sector located above the panel under analysis. The second source of dilution is ‘caving dilution’, which occurs when an air gap is formed in a large area and a sudden propagation of a hang-up occurs, inducing early ingress of waste into the production level from a level above. The third mode observed is ‘lateral dilution’, which occurs when the panel is located next to an exhausted sector and the broken material enters the draw column early during draw. In the last two cases, dilution could migrate large distances horizontally. The mine data indicated that draw and caving strategy are key for long-term planning of large panel caves. Keywords panel caving, dilution entry, gravity flow mechanics, draw control, uniform draw.

These parameters include the spacing of drawpoints, the geometry of the orebody, the way in which drawpoints are extracted, and the flow characteristics of the ore and waste. A review of the parameters affecting dilution is shown in Table I. In quantitative terms, there have been attempts to derive formulae to evaluate dilution entry at caving operations and from them the amount of dilution. Laubscher (1994) defines the dilution entry as the percentage of column drawn when dilution is observed at a drawpoint. Based on empirical observations he proposed the following relationship to calculate the dilution entry point (DEP) [1] where HIZ is the height of the interaction zone, Hc is the draw column height, S is the swell factor of the column, and DCF is the draw control factor, which is determined from the standard deviation of the tonnage extracted from the working drawpoints normalized by 100 at a monthly scale (SD ) as: [2]

Dilution is an integral part of a cave mining operation and thus the control and understanding of its main sources are crucial in managing a caving operation. Dilution may have diverse sources, including the position of the ore/waste, the percolation of fines, and the draw control at the mine. There are ways to control the amount of dilution in panel caving through mine planning. For example, DeWolfe (1981) states that, in order to control dilution in a panel caving operation, the gap between the extraction pile and the cave back must be kept to a minimum so that the waste does not rill to the new open drawpoints. This is achieved by adjusting the angle of draw, which should be set below 45° to avoid early dilution from previously exhausted levels. Laubscher (1994, 2000) postulated that dilution content is controlled by a set of parameters ranging from the design to the mine operation itself. The Journal of The Southern African Institute of Mining and Metallurgy

Susaeta (2004) based on an analysis of Codelco´s extraction data postulated that dilution behaviour behaviour would depend on the way the draw is conducted at a mine, thus expanding on Laubscher’s model and defining three dilution mechanisms: interactive, isolated-interactive, and isolated. Therefore, when a drawpoint is drawn in isolation, dilution entry is low and the amount of dilution increases rapidly during extraction. In isolated-interactive draw, dilution entry is moderate and dilution content will also increase moderately. In interactive draw, which

* University of Chile, Santiago, Chile. © The Southern African Institute of Mining and Metallurgy, 2014. ISSN 2225-6253. Paper received Oct. 2012; revised paper received Jul. 2013. VOLUME 114

JUNE 2014

455

▲

Introduction

Empirical observations of dilution in panel caving Table I

Parameters affecting dilution Parameter

Effect on dilution

Author

Ore/waste contact surface inclination

To minimize dilution content extraction, the ore/waste contact interface must be kept at 45° to 50° moving away from the uncaved ore through draw control.

Julin (1992)

Ore volume to ore/waste interface area

The higher the ratio of ore volume to the surface area of the ore/waste interface, the lower the overall percentage of dilution.

Laubscher (2000)

Finely fragmented waste and coarse ore means early and extensive dilution, while coarse waste and fine ore means a low overall extracted dilution percentage.

Laubscher (2000)

Good drawzone interaction and parallel flow will represent the optimum conditions. Poor drawpoint interaction and drawzones angled according to local variations will lead to high dilution.

Laubscher (2000)

Fragmentation range from ore and waste Height of the interaction zone

Differences in density of ore and waste Block or panel caving strategy Uniformity of draw

High-density ore and low-density waste lead to low dilution and vice versa.

Laubscher (2000)

A block cave strategy will lead to more lateral dilution mixing than panel caving.

Laubscher (2000)

High uniformity of the tonnage drawn from the neighbour drawpoints will lead to high interaction and late dilution entry.

Julin (1992), Susaeta (2004)

occurs under uniform f draw, dilution content would be low. The definition of the flow mechanism is postulated to depend on the value of the uniformity index (IU), an index that accounts for the uniformity of the draw control at the mine. The latest research conducted in the field using markers has helped understanding of near-field flow, and demonstrated that flow is quite chaotic in block caving environments (Brunton et al., 2012). New experirments are planned to confirm these results (Castro and Armijo, 2012). The use of markers and the results of the data analysis could be used to improve understanding of the ore flow and to improve mine planning, especially as more markers are being recovered under this research programme. Despite the large amount of research and practical observations of panel caving operations, no guidelines have emerged for long-term planning to control dilution entry. This paper presents a thorough analysis of dilution at six productive sectors of Codelco’s caving operations as part of a broad and exhaustive review of dilution observed at the mines. In addition, the framework for building and calibrating a model for predicting the mean dilution entry point, based on mine data, is presented.

Mine data collection and analysis The information collected at Codelco’s underground operations is presented in Table II. It is worth noting that the geological marker used to describe dilution is different for each mine. The data is also recorded differently. At El Salvador, only the limonite content (percentage observed at drawpoint) is declared and is usually recorded by the mine production control (MPC) personnel. The case of Andina is different: a trained geologist records the rhyolite (percentage observed at drawpoint) as well as the caved rock material (percentage observed at drawpoint) coming from the Panel II exhausted sector. For El Salvador and Andina, the geological marker information is collected on a drawpoint basis every 2000 t, along with grade control. This investigation is based on one of the largest data-sets used to date, where 2065 drawpoints were analysed corresponding to the extraction of 166 Mt. This long draw

▲

456

JUNE 2014

VOLUME 114

history corresponds to a period off 737 months in total (as per Table III).

Geological marker entry definition The first stage in studying the mechanisms by which dilution enters the drawpoints is to establish a generalized criterion for declaring dilution entry at a drawpoint. To do this, the dilution curves of the 1674 drawpoints that reported geological markers were analysed. After analysing the observed dilution content as a function of the percentage extracted from the in situ column, it is possible to identify two types of behaviour. The first is the pulse-shaped behaviour, where dilution reverts to zero after its first appearance (Figure 1a). The second is the continuous-entry behaviour, where dilution percentage does not revert to zero after the first appearance and may or may not show a continuous increase (Figure 1b). As approximately 51% of the dilution curves in the database have a pulse-shaped behaviour, we developed a criterion to define a range when a pulse of dilution is significant. This criterion can then be used towards determining that dilution entered the drawpoint. The cumulative dilution values for a period t are defined as: [3]

where observed dilution%i and tonnage i are the dilution percentages observed at the drawpoint and the tonnage drawn during period i, respectively. The cumulative curves allow for regularizing the curve behaviour and analyses when the overall dilution content exceeds a certain threshold. Once the use of the cumulative curves is selected, the criterion for declaring dilution entry at a drawpoint is reduced to selecting the threshold to be exceeded by cumulative dilution content. Figure 2 illustrates an example of the dilution content for the drawpoint A69W51 at El Salvador Mine. It also indicates the change in the tonnage at 3% [DEP(3%)] and 5% [DEP(5%)] The Journal of The Southern African Institute of Mining and Metallurgy

Empirical observations of dilution in panel caving Table II

Mine data collected at Codelco's operations Codelco Division El Salvador

Andina

Mine sector

Inca Central East Inca Central West Inca North Inca West Panel III

General information

ICE • Block models of the date extraction began ICW • Topography IN • Drawbell design IW • Extraction and undercut levels

Extracted tonnages and copper grades

Geological markers

Tonnages and CuT grades at a drawpoint scale on a shift-to-shift basis, consolidated with treatment plant measurements

Dilution (limonite) content estimated by means of monthly observations at a drawpoint scale, available throughout mine life Dilution (rhyolite and broken material) content estimated by means of monthly observations at a drawpoint scale, available from Jan-06

Table III

Summary of database used in this study Sector

ICE ICW IN IW LHD cluster Grizzly cluster

Number of drawpoints

Sector extracted tonnage

Sector extracted Cu

Number of periods

Total extracted

with extraction

(kt, monthly)

grade (%, monthly)

(months)

tonnage (t)

108 471 363 578 397 148

Mean

Std. dev.

Mean

Std. dev.

57.7 259.5 297.6 164.5 371 149

51.8 98.5 205.2 87.3 353 90

0.56 0.64 0.57 0.54 1.1 1.0

0.06 0.05 0.11 0.05 0.1 0.13

88 133 203 189 74 50

5.1 34.5 60.4 31.1 27.1 7.5

Figure 1—Observed dilution percentage as a function of the percentage extracted from the in situ column for: (a) a drawpoint with pulse-shaped behaviour and (b) a drawpoint with continuous-entry behaviour

The Journal of The Southern African Institute of Mining and Metallurgy

VOLUME 114

JUNE 2014

457

▲

Figure 2—Illustration of the DEP declared at 3% and 5% of the cumulative content for a drawpoint, where (a) is the cumulative dilution curve and (b) is the observed dilution curve

Empirical observations of dilution in panel caving off the overall (cumulative) dilution content. In this case, it is shown that a 3% cumulative is equivalent to 40% observed dilution at a drawpoint. Four cumulative dilution values were investigated to define a significant value for dilution and to carry out a statistical analysis. It is worth noting that the dilution entry concept of Laubscher (1994) corresponds to zero cumulative dilution and has a large error or deviation for the purpose of carrying out statistical data analysis. The data analysis shows (Table IV) that the 3% threshold criterion (DEP(3%)) represents a robust volume of data and it is sufficient to reach steady state with respect to the numbers of drawpoints with declared dilution in a given sector. Therefore, in this paper we used the 3% threshold to define when a drawpoint has dilution. At Andina, there are two different geological markers considered as dilution: (1) the broken rock from the previous panel and (2) the rhyolite. In this case, the dilution entry point is calculated separately for each of the markers, which are called broken entry (BEP) and rhyolite entry (REP), respectively.

the sector. This has an important influence f on the dilution entry profile for the different sectors, except for the IW sector, where a hang-up of the cave back that occurred during the first stages of extraction caused rapid dilution entry after an air blast (De Nicola and Fishwick, 2000). Table V summarizes the results of the analysis at the El Salvador sectors. It shows that two sectors with the same uniformity, but with different in situ column height profiles (ICE and ICW), have significantly different mean dilution entry points. In other words, dilution entry does not depend on the uniformity percentage when the dilution source is located next to the sector instead of above it. Table V summarizes the dilution entry hypothesis for the El Salvador sectors based on the analysis. The mechanisms observed for dilution entry depend on the initial boundary conditions of the different sectors. For Inca Central East, the extraction sequence proceeds from the lowest in situ column (under 50 m) to the higher ones (over 400 m). This means that the extraction sequence

Dilution entry analysis The dilution entry analysis consisted of evaluating the dilution entry behaviour at a drawpoint scale and considering the diverse initial conditions of the different sectors, that is, in situ column heights, extraction sequence, and uniformity of extraction. This empirical analysis was carried out to develop a hypothesis for the mechanisms by which dilution entered the drawpoints. The uniformity index introduced by Susaeta (2004) was used to calculate the uniformity with which extraction was performed (IU).

El Salvador The four sectors analysed at El Salvador have different column height profiles and their extraction sequences have different directions relative to the location of the corresponding dilution source. Thus, dilution entry profiles at a drawpoint scale vary from one sector to another. Figure 3 shows isometric views of the broken and in situ material contact surfaces for the El Salvador sectors. For Inca Central East (ICE), Inca North (IN), and Inca West (IW), the dilution sources are located mainly to the side of the sector, while for Inca Central West (ICW), the source is located above

Figure 3â&#x20AC;&#x201D;Isometric views of broken and in situ material contact surfaces for El Salvador sectors

Table IV

Number of drawpoints with dilution entry declared for different thresholds Division

El Salvador

Andina

â&#x2013;˛

458

Sector

Geological marker

Number of drawpoints with significant dilution DEP(0%)

DEP(3%)

DEP(5%)

DEP(10%)

ICE ICW IN IW

Limonite Limonite Limonite Limonite

97 289 537 170

58 167 361 136

55 143 329 117

49 68 262 66

LHD cluster LHD cluster Grizzly cluster Grizzly cluster

Broken Rhyolite Broken Rhyolite

137 150 147 147

53 87 137 120

37 49 127 92

17 5 98 52

JUNE 2014

VOLUME 114

The Journal of The Southern African Institute of Mining and Metallurgy

Empirical observations of dilution in panel caving Table V

Dilution analysis indexes summary for El Salvador sectors El Salvador sector

Inca Central East Inca Central West Inca North Inca West

ICE ICW IN IW

In situ column height (m)

Percentage of uniformity (uniform draw) of the interior drawpoints (%)

DEP(3%) (%)

226 ± [173] 159 ± [53] 181 ± [85] 247 ± [122]

88 ± [10] 88 ± [10] 79 ± [18] 84 ± [10]

57 ± [54] 103 ± [40] 78 ± [70] 67±[48]

starts ffrom the dilution source in the north. Following the sequence, the dilution entry point decreases until a broken rock pillar created by the inactive drawpoints (orepasses) acts as a dilution barrier. In this case, dilution travels from the source at the northern zone to the central zone (Figure 4d). For the Inca Central West (ICW), the extraction sequence starts with 130 m in situ column heights (zone 1), passing through column heights of over 200 m (zone 2) and then back to 130 m in situ column heights (zone 3). In this case, the dilution source is located above the sector. The DEP(3%) goes from 40%to 100% for zone 1, then increases to over 100% for zone 2, and goes from 80% to over 100% in zone 3. Data shows that dilution enters due to the vertical displacement of the broken material from the sector previously extracted and located above ICW. For the Inca North, the sequence started with in situ column heights ranging from under100 m to 150 m (most of the drawpoints), and finished extracting columns of over 200 m at the end of the sequence. Following the sequence, the dilution entry point decreases to the central zone of the sector (DEP(3%) under 40%) and then increases to the northern zone. The data indicates that dilution in the central zone comes from the southern zone with a large horizontal displacement, despite the uniform draw. Finally, for the Inca West sector and as indicated in Figure 7, the dilution entered the sector after a large hang-up of the cave back collapsed at the beginning of the extraction sequence. This indicates that dilution entry is controlled by the sudden propagation of the cave back (De Nicola and Fishwick, 2000).

At the Grizzly and LHD clusters, different ff conditions prevailed regarding the original position of the broken material before extraction began. On the western side of the

Figure 4—Plan views of ICE's drawpoints, showing (a) in situ column heights, (b) extraction sequence, (c) percentage of uniformity in uniform draw, and (d) DEP (3%)

Andina

The Journal of The Southern African Institute of Mining and Metallurgy

Figure 5—Plan views of ICW's drawpoints, showing (a) in situ column heights, (b) extraction sequence, (c) percentage of uniformity in uniform draw, and (d) DEP(3%) VOLUME 114

JUNE 2014

459

▲

Andina’s Panel III mine contains two rock types (primary and secondary mineralization) in the in situ columns. Thus, there are two main areas with different extraction layouts: the LHD area (mechanized extraction) and the Grizzly area. There are also two main geological markers (considered dilution): a rhyolite chimney, located in the northeastern side of Panel III, and broken material coming from the previously extracted sector (Panel II), located above the central area of the Panel III sector (Figure 8b). These two markers originated from different locations and therefore were analysed separately. Information on the geological markers at the Andina mine’s drawpoints is available from January 2006. Thus, the geological markers’ entry behaviour was analysed over two clusters where extraction started in January 2006. These clusters are located in the LHD and in the Grizzly area (Figure 8a).

Empirical observations of dilution in panel caving Grizzly cluster, there are two exhausted panels that fform a wall of broken material; thus, the dilution source for this panel was located mainly next to the cluster and not above it (Figure 9b). On the other hand, the panel located at the southern side of the LHD cluster had extracted less than 10 m of the in situ columns when extraction began; thus, the dilution source was the broken material left from the

extraction off the previous panel. Well-controlled extraction in terms of the uniformity at the LHD dilution cluster, together with the fact that dilution was located above the panel, led to high dilution (broken material) entry points for the cluster’s drawpoints (Figure 10d). The broken material entry analysis for the Grizzly cluster shows that, despite the broken material source being located on the western side of the panel, the whole cluster shows evidence of early entry of this kind of dilution (Figure 9d). This could be due to the extraction of the old panel, which caused subsidence and flow of the broken ore to the new panel.

Dilution entry observations Considering the dilution analysis performed for El Salvador and Andina mines, it is possible to hypothesize the following dilution mechanisms.

Vertical dilution Dilution enters the drawpoints by descending gravitationally from the source, which in this case, is located above the panel under analysis. However, other phenomena, such as percolation, could occur for the finer rock. This phenomenon was observed at the Inca Central West (ICW) sector and the LHD cluster (Figure 11a). The dilution entry and ore recovery are high in relation to the other types of dilution. Figure 6—Plan views of IN's drawpoints, showing (a) in situ column heights, (b) extraction sequence, (c) percentage of uniformity in uniform draw, and (d) DEP(3%)

Lateral dilution The cave back could preferentially propagate, in terms of rate, to the lateral interface of the in situ and broken material, when the dilution source is located next to the panel under analysis. This interface may not necessarily be vertical but have a smaller angle depending on geomechanical properties of the rock mass and the amount of ore removed at boundary points. If the air gap between the cave back and the muck pile is sufficiently large, caved waste rock will enter the draw columns and may travel horizontally depending on the pile slope. This phenomenon was observed at the Inca Central East (ICE), Inca North (IN), and the Grizzly cluster (Figure 11b). In this case the dilution entry is low, and if not controlled, such dilution could migrate large horizontal distances.

Dilution entry after an air blast event (caving dilution) Figure 7—IW extracted dilution tonnage and content; dilution enters the sector after the air blast occurs

If the hydraulic radius of the initial opened area is insufficient for continuous propagationof the cave, a hang-up

Figure 8—(a) Plan view of Panel III’s drawpoints showing cluster locations and primary-secondary mineralization contact, (b) isometric view of Panel III’s drawpoints showing the rhyolite chimney and broken material contact surfaces with the in situ material

▲

460

JUNE 2014

VOLUME 114

The Journal of The Southern African Institute of Mining and Metallurgy

Empirical observations of dilution in panel caving Table VI

Dilution analysis index summary for Andina clusters Andina cluster

LHD Grizzly

Area

North South North South

In situ column heights

Percentage of uniformity (uniform draw index IU)

BEP(3%)

REP(3%)

(m)

(%)

422 ± [127] 331 ± [121] 290 ± [62] 433 ± [22]

82 ± [12] 89 ± [6] 62 ± [10] 68 ± [11]

73 ± [40] 33 ± [21] 32 ± [19]

54 ± [18] 90 ± [28] 41 ± [23] 43 ± [22]

Figure 9—Plan views of the Grizzly cluster’s drawpoints, showing (a) in situ column heights, (b) extraction sequence, (c) percentage of uniformity in uniform draw, and (d) BEP(3%) with broken material contour lines

Figure 10—Plan views of the LHD cluster’s drawpoints, showing (a) in situ column heights, (b) extraction sequence, (c) percentage of uniformity in uniform draw, and (d) BEP(3%) with broken material contour lines

Figure 11—Schematic sequential drawbell sections, showing (a) gravity dilution and (b) lateral dilution

Conclusions In this paper we presented an analysis of draw control data The Journal of The Southern African Institute of Mining and Metallurgy

towards understanding dilution entry for f long-term purposes. The data showed that boundary conditions (where dilution is located) and caving propagation are the most important factors to be considered for long-term planning purposes. Other variables, such as uniform draw at a short term scale, are not so important for determining the entry of dilution for long-term planning. Rather, the draw strategy and cave back management seem to be the key for draw control purposes. In terms of the stage of mining, dilution could occur during caving and after the cave has broken through to the surface. During cave propagation, if the extraction pile height VOLUME 114

JUNE 2014

461

▲

off the cave back will occur. Iff extraction continues in the opened area without incorporating new area, a large air gap will be generated. Once new area is incorporated, the cave back will suffer a sudden propagation and dilution located above the panel will enter the first opened drawpoints following an air blast. This is the case observed at the Inca West (IW) sector (Figure 12). In this case, the dilution entry and the recovery are low.

Empirical observations of dilution in panel caving Acknowledgments The authors would like to thank the colleagues who collaborated with the mine data for this paper, especially M. Fidel Baez (Codelco), M. Mauricio Melendez (El Teniente), M. Ernesto Arancibia (El Salvador). and M. Felipe Marinkovic (Andina). The assistance of the Laboratorio de Block Caving is also acknowledged. Finally, the authors would like to acknowledge the financial support of the Chilean Government through the project Conycit FB0809.

References BRUNTON, I., SHARROCK, G., and LETT, J. 2012. Full scale near field flow behaviour at the Ridgeway Deeps Block Cave Mine. Proceedings of MassMin 2012: 6th International Conference and Exhibition on Mass Mining, g Sudbury, Canada, 10–14 June 2012. Canadian Institute of Mining, Metallurgy and Petroleum, Westmount, Quebec. CASTRO, R. and PAREDES, P. 2012. Comparison of REBOP to large panel caving at

Figure 12—Caving dilution

Codelco’s Operations. Final Internal Technical Report. t Mass Mining Technology Project II, Toronto, Canada, 5–7 June 2012. CASTRO, R. and ARMIJO, F. 2012. Experimental design for the full scale flow test

is lower than the cave back height for a given area, an air gap could develop, allowing dilution to enter the draw column. Thus, if the draw rate (V VEXX) is larger than the cave propagation, a gap will form if:

marker at El Teniente. Internal report to Codelco Chile. Laboratorio de Block Caving, January 2012. COULOMB, C. 1776. Essaisurune application des regles des maximisetminimis a quelquesproblems de statiquerelatifs a la architecture. Mem. Acad. Roy.

[4]

Div. Sav., vol. 7. (in French). DE NICOLA, R. and FISHWICK, M. 2000. An underground air blast – Codelco-Chile

where e is the void ratio in the pile, (ρS) is the solid density of the ore (t/m3), and VCP is the rate of caving. On the other hand, if the local slope (β) of the extraction pile heights of two contiguous drawbells in the sequence direction is greater than the friction angle of the diluting material, dilution will be able to rill and move laterally. The pile’s slope angle is defined as: [5]

– Division Salvador. Proceedings of MassMin 2000, Brisbane, Australia. Australasian Institute of Mining and Metallurgy, Melbourne. pp. 173–178. DEWOLFE, V. 1981. Draw control in principle and practice at Henderson Mine. Design and Operation of Caving and Sublevel Stoping Mines. Stewart, D.R. (ed.). Society of Mining Engineers of AIME, New York. pp. 729–735. JANSSEN, H. 2004. Experiments regarding grain pressure in silos written in1895. Proceedings of MassMin 2004, Santiago, Chile, 22–25 August 2004. Instituto de Ingenieros de Chile, Chile. pp. 293-300. JULIN, D. Block Caving. SME Mining Engineering Handbook. Vol. 2, ch. 20.3.

where(W ) is the front width (m), and (V VD) is the development rate (m2/d). There will be lateral displacement of dilution if β is greater than the friction angle of the diluting material θ or when tan (β) > tan(∅). In addition, the ratio of ore volume to dilution contact surface area (R ( OD) and the difference between column heights influence the dilution entry. As the ratio of the ore volume to dilution contact surface area increases, the ore subjected to dilution decreases. Thus, dilution entry will be high, as ROD (in metres) is higher. On the other hand, the bigger the difference between the sector’s column heights, the higher the potential for lateral dilution mixing; therefore, dilution entry will be small. There are other factors that influence dilution once the cave back has propagated, including draw control and percolation rate of the finer diluting rock such as cave propagation and extraction velocity, volume of ore to waste, and draw angle. These complexities would have to be taken into account in a new methodology for dilution strategy control at the mines.

▲

462

JUNE 2014

VOLUME 114

Society for Mining, Metallurgy, and Exploration, Inc., Littleton, CO. 1992. pp. 1815–1821. LAUBSCHER, D. 2000. Block Cave Manual prepared for the International Caving Study. Julius Kruttschnitt Mineral Research Centre, University of Queensland. pp. 111-118. Laubscher, D. 1994. Cave mining – the state of the art. Journal of the South African Institute of Mining and Metallurgy, vol. 94, no. 10. pp. 279–293. PAREDES, P. 2012. Dilution entry mechanisms in block and panel caving operations. Master’s thesis, University of Chile (in Spanish). SUSAETA, A. 2004. Theory of gravity flow (Part 1). Proceedingsof MassMin 2004, Santiago, Chile, 22–25 August 2004. Instituto de Ingenieros de Chile, Chile. pp. 167-172. SUSAETA, A., RUBIO, E., PAIS, G., and ENRIQUEZ, J. 2008. Dilution behaviour at Codelco panel cave Mines. Proceedings of MassMin 2008, Luleå, Sweden. Lulea University of Technology. pp. 167–178.

◆

The Journal of The Southern African Institute of Mining and Metallurgy

Design principles for optimizing an established survey slope monitoring system by N. Mphathiwa* and F.T. Cawood*

When slope angles are designed during open pit optimization, there is a risk factor applied in steepening the slopes. The steepening of slope angles has implications for the safety and economics of the mining operation. The steeper the slope angles, the greater the probability of slope failure. Although a slope failure will result in added costs, the challenge is to compile an accurate cost-benefit exercise optimizing the economic benefits of the project without exposing mine workers and equipment to unacceptable risk of rockfalls. A balance between the safety of the operation and the economics of the investment is therefore required. The ideal situation is to have a slope monitoring system that will predict slope failure by detecting any ground movement before the actual failure occurs. This early warning will allow the risk factor to be applied with a high degree of confidence, knowing that the risk will be adequately mitigated. The objective of this paper is to provide guidelines on how to design an optimal survey slope monitoring system. It is the authors’ view that for a survey monitoring system to yield desirable results, it should adhere to survey principles such as working from the whole to part and consistently cross-checking. The case study used is Jwaneng Mine, and the design strategy outlined can be used as a guideline for developing a new slope monitoring system or to optimize an existing one. Keywords open pit optimization, slope angle, slope failure, slope monitoring system design.

Introduction Monitoring is defined as the regular observation of activities taking place in a project or programme, and is a process of routinely gathering information on all aspects of the project (Bartley, 2007). There are different types of monitoring surveys, but this paper focuses on slope stability monitoring. Slope stability monitoring can be defined as the science of measuring ground movements and detecting instability before failure occurs. Monitoring is an invaluable tool for assessing design performance and failure risk and for aiding risk minimization (Read and Stacey, 2009). The objective of slope stability monitoring is to balance mine safety with the economics of the project. The safety of workers in any mining operation is the number one priority of every mining manager. This is both a moral and legal obligation. It is, therefore, critical to have a reliable slope monitoring system so that The Journal of The Southern African Institute of Mining and Metallurgy

* School of Mining Engineering, University of the Witwatersrand. © The Southern African Institute of Mining and Metallurgy, 2014. ISSN 2225-6253. Paper received r Nov. 2012; revised paper received Jan. 2014. VOLUME 114

JUNE 2014

463

▲

Synopsis

any potential ffailure can be detected well in advance, allowing for workers and equipment to be evacuated promptly from the hazard areas. The steepening of slopes results in less waste rock stripping, and hence reduces the costs of mining significantly. However, by steepening the slopes, the probability of slope failure is increased. This risk associated with the steepening of slopes is mitigated by slope stability monitoring. It follows, then, that the more reliable the slope monitoring system, the more risk can be taken when designing the slopes, hence reducing the cost per ton mined further. It is expected that, in the near future, data from slope monitoring equipment will add a much-needed dimension to slope engineering, when used to improve slope designs and to optimize slope angles (Cawood and Stacey, 2006). Slope failures can also result in ore dilution when sliding waste rock mixes with the ore. This will inevitably reduce the grade and increase mining and treatment costs. A rockslide at Kumtor gold mine in Kyrgyzstan resulted in 100 000 ounces being cut from the 2006 production forecast (Mining News, 2006). The slope monitoring system allowed the area to be safely evacuated in advance, and there were no injuries, although a diamond drill was covered by rock. It is the authors’ opinion that the design of the slope monitoring system is the determining factor in setting up a reliable early warning. In this paper, the authors will attempt to answer the fundamental question of how to design a slope monitoring system? The focus will be on geo-referenced systems, otherwise known as survey slope monitoring systems. These systems include, among others, the Geodetic Monitoring System (GeoMos), slope

Design principles for optimizing an established survey slope monitoring system monitoring radar (SSR), and the global positioning system (GPS) technology. The introduction of automated slope monitoring systems was a major step in optimizing the whole concept of monitoring. However, in the authors’ opinion, no matter how sophisticated the instrumentation or the software is, if the foundation or design is not optimal, the level of confidence in the monitoring results will be low. This paper will be of interest to professionals involved in open-pit mining, including mine surveyors, mine planners, geotechnical engineers, mine safety officers, and all employees working in open-pit operations.

The case study Jwaneng Mine, which is owned by Debswana Diamond Company, is used as a case study. Jwaneng Mine is currently extending its open pit mining through its Cut 8 project (Debswana, 2010), which will deepen the pit from 330 m to 624 m, with a length of 2.7 km and width of 1.7 km. A prefeasibility study is being undertaken for a Cut 9 project, which will deepen the mine further to 850 m, with a possibility of extending the dimensions of the pit further with a Cut 10 project (Mining Weekly, 2010). The deepening of the pit and the general increase in the footprint increases the risk associated with slope failures. The Cut 8 mining limit will be approximately 100 m from the main treatment plant infrastructure. Movement of the ground in the vicinity of the plant infrastructure can result in production losses for the company and significant unplanned replacement or repair costs. The abovementioned scenarios call for a robust slope monitoring system design to successfully mitigate the risk of slope failure. In this paper, we assess the existing slope monitoring design at Jwaneng Mine and develop recommendations in order to make it optimal. Jwaneng Mine has been running a slope stability monitoring programme since 1989. In 1995, there was a proposal to upgrade the monitoring programmes at the Letlhakane, Jwaneng, and Orapa diamond mines (Watt, 1996). The focus was on the actual monitoring using conventional survey instruments such as the Wild DI 2202, precise levelling, and the calculation of the survey observations to reduce them to useable information. Most of the recommendations were implemented by all the three Debswana mines, and benefits were realized at that time. However, with the passage of time, developments have increased the need for a different approach to monitoring. The mines have gone deeper and wider with mining of additional cuts. For example, the Cut 8 limit at Jwaneng Mine is less than 100 m from the plant infrastructure. These developments, especially the deepening of the pits, have increased the risk associated with slope failure. To mitigate this heightened risk, Debswana responded by intensifying the monitoring and by increasing the number of targets and the frequency of the monitoring. Debswana introduced automated monitoring systems in 2001, and has been gradually updating them at all of its mines. The GeoMos system was introduced to the company and implemented at the Letlhakane Mine in 2002, followed by Orapa and Jwaneng Mines. Similarly, SSR was first implemented at Jwaneng Mine in 2005, followed by Letlhakane and Orapa. Jwaneng has recently started installing GPS receivers in and around the pit to enhance the

▲

464

JUNE 2014

VOLUME 114

existing monitoring systems to mitigate the heightened risk of mining Cut 8.

Survey monitoring design considerations The following design parameters of survey monitoring systems were considered:

Survey control network This is the basis of the design. The integrity of any survey measurements depends on the accuracy of the survey control network. In the case of slope stability monitoring, all movements are referenced to the survey control network. The first set of survey stations to be installed consists of primary beacons. The optimal distance of the primary beacons from the pit rim must be determined systematically. The positioning of the secondary beacons with respect to the monitoring site should also be established.

Construction of the survey beacons ➤ Primary beacons—the s stability of primary beacons is critical because they are used as reference stations for orientation and to determine the position of the monitoring station when using the GeoMos ➤ Secondary beacons—the s secondary beacons are constructed close to the rim of the pit so that there is a clear line of sight to the monitoring targets. Although the secondary beacons’ stability is inevitably affected by blast vibrations, because of their close proximity to the pit, there is need for a structural design that can withstand blast vibrations as much as possible ➤ Instrument shelter—the r construction of the shelter for housing the monitoring equipment should be investigated. Instruments must be well protected against corrosion, moisture, other aggressive agents, and vandalism (Abramson et al., 2001). The shelter protects the instrument from mining conditions such as dust and flying rocks. However, the materials used must not compromise the accuracy of the monitoring measurements.

Equipment election The choice of equipment depends primarily on the required accuracy, and also the type of movement to be detected. Some instruments such as levels are good for vertical movements, while others, such as global positioning systems, are suitable for horizontal movements. The authors investigated ways to utilize different monitoring equipment to complement each another.

Software selection The focus is on how to analyse and present data from various monitoring systems. Most monitoring systems are equipped with software for interpreting and presenting results. The aim is to investigate ways of integrating data from these different systems to ease the flow of information.

Skills and competencies For the design to produce desired results, it is imperative to have people with right skills and competencies to implement The Journal of The Southern African Institute of Mining and Metallurgy

Design principles for optimizing an established survey slope monitoring system and maintain it. It is recommended that a qualified f and competent person should oversee the slope monitoring programme and conduct the data analysis (Jooste and Cawood, 2006).

Optimization of a typical monitoring system The optimization strategy will consider the following parameters.

Control network design The survey control network design process is as outlined as follows: ➤ A desktop exercise to determine the provisional positions of the survey beacons ➤ Determination of lines of sight to be used during geodetic surveys ➤ A reconnaissance to adjust the provisional positions to more practical positions ➤ Computation of observations from coordinates using survey applications such as resection ➤ Tests of the network accuracy by computing standard deviations of coordinates calculated from redundant observations (Kealey, 2004). The provisional positions of the primary beacons are established using the principle of locating the primary control points anywhere from 100 m to 3 km away from the pit rim (Cawood and Stacey, 2006). Figure 1 shows conceptual positions of the primary beacons established from a desktop study. The design entails two sets of primary beacons. The first set of primary beacons will be positioned 100 m away from the pit. The build-up of dumps and infrastructure around the pit is a constraint in placing the primary beacons further away because this will affect the line of sight. This set of primary beacons (100 m radius) is used for orientation during geodetic monitoring. It is also used to check and update the position of the monitoring beacon using the resection method, known as thee free station in GeoMos. The first set of primary beacons, 100 m away from the pit rim, is not be very stable as it is affected by blast vibrations. It is, therefore, critical to regularly update the beacon positions using the GPS post-processing method. The second set of primary beacons (3 km radius) is logged as known points when applying the post-processing method, as these beacons are more stable. The secondary beacons will be constructed on the rim of the pit. The guiding principle is to maximize the view onto the pit (Bannister Raymond, Baker, 1998). The current GeoMos design at Jwaneng mine requires only two secondary beacons to be used as monitoring beacons. However, additional secondary beacons should be built in case there is loss of line of sight to one of the monitoring beacons or the stability of the ground they are built on is compromised (Cawood and Stacey, 2006).

➤ Is the beacon design compatible with geotechnical properties of the ground on which the beacon will be constructed? ➤ Is the design easy to implement? ➤ How will the designer ensure that the structure is implemented as designed? ➤ Does the contractor have the right competencies to implement the design specification adequately? The structural design of the survey beacons is appropriate for the Jwaneng Mine stratigraphy. The 17-20 m top layer of sand has been designed for by incorporating piling to ensure the foundation of the beacon is built on solid rock. Piling is highly recommended when the bedrock is covered by less competent material, such as sand (Leica Geosystems. 2004). The construction notes explaining how the design should be implemented are clear and easy to understand, making the design easy to implement. It is advised that the construction specifications be made simple to interpret. To ensure that the beacon design is constructed to the correct specification, the company needs to consider the following: ➤ When evaluating tenders for the construction of the beacons, more weight should be given to the technical competencies of the company rather than general practice of giving the lowest bidder more points ➤ There is need for a construction schedule to accompany the structural design. The construction schedule should have gate release clauses stating stages of construction where progress cannot be made to the next stage until the built structure has been inspected and signed off by the relevant personnel.

Instrument shelter The next design aspect to consider is the instrument shelter that houses the Total Station when using the GeoMos for monitoring. The purpose of the shelter is to protect the instrument from theft, dust, rainfall, and flying rocks from blasting activities. When designing the instrument shelter, construction material that will not affect the accuracy of the measurements must be used. Figure 2 shows a typical design of an instrument shelter. The glass allows the Total Station to sight to any beacon or targets within its line of sight without hindrance from the shelter. Jwaneng Mine has experienced problems with

Beacon design and construction

The Journal of The Southern African Institute of Mining and Metallurgy

Figure 1—Provisional positions of the primary beacons VOLUME 114

JUNE 2014

465

▲

After completion of the design of the survey control network, the focus now shifts to the beacons’ structural design and construction. There are four fundamental questions to consider when designing and constructing survey beacons:

Design principles for optimizing an established survey slope monitoring system measuring through glass as it was affecting ff the accuracy off monitoring results. Glass with a thickness of 3 mm or less has a minimal impact on the accuracy of the measurements, and the errors can be adjusted using a tested formula. Tint and shape also matter, since clear flat glass has the least impact (Afeni and Cawood, 2010).To protect the glass from flying rocks during blasting, the shelter can be equipped with pull-down rubber curtains that can be pulled down during blasting.

Selection of monitoring instrumentation After constructing the infrastructure, such as control survey beacons and the housing of the instrument, the next design process involves the selection of suitable monitoring equipment. The selection process takes the following factors into consideration: ➤ The expected magnitude of the ground movement ➤ Most likely movement direction (horizontal or vertical)

Figure 2—Proposed instrument shelter (source: Read and Stacey, 2009)

➤ ➤ ➤ ➤ ➤ ➤

Accuracy and precision off the instrument Number and frequency of measurements Size of area to be monitored Level of automation Ease of interface with other monitoring instruments GIS adaptability (Cawood and Stacey, 2006).

The monitoring process should be started by the identification of risk areas by the geotechnical engineers (Jooste, 2005). The areas are then classified according to the severity of the risk (high, medium, and low) as shown in Figure 3. The severity of the risk is one of the determining factors in equipment selection. Jwaneng mine has two Total Stations connected to the GeoMos, two SSRs, six GPS receivers, one digital level, and one GPS/GNSS surveying system as part of the slope stability monitoring equipment. This combination of equipment can provide an optimal monitoring solution if it is appropriately utilized. Two Total Stations, which are components of the GeoMos, should continue to monitor either side of the pit. The GeoMos primarily tracks the direction of movement, while the SSR is used to measure the magnitude. There needs to be a systematic link between the SSR and the GeoMos. For example, when specific movement limits are reached when monitoring with the GeoMos, monitoring can be intensified by incorporating the SSR. It has been suggested that before taking any actions when movement limits are reached, the responsible personnel should confirm that the cause is actual ground movement (Jooste, 2005). The systematic deployment of monitoring equipment is illustrated in Figure 4. The area close to the Cut 8 mining limit has been identified as a high-risk area and its monitoring should be intensified by dedicating a SSR unit to monitor the highwalls continuously in the area, as shown in Figure 5. GeoMos targets should also be installed in the area to assist with establishing the direction of movement if detected. To enhance the monitoring further, GPS receivers should be installed on the highwall in that area to provide a cross-check to the GeoMos and the SSR. Cross-checking is a necessity in slope stability mornitoring (Abramson et al., 2001).

Figure 3—Risk areas (Source: Jooste, 2005)

▲

466

JUNE 2014

VOLUME 114

The Journal of The Southern African Institute of Mining and Metallurgy

Design principles for optimizing an established survey slope monitoring system levelling method (Jooste, 2005). The challenge inherent in the precise levelling method is that it is a point measuring method and will not adequately cover large areas. To enhance the precise levelling method, the mine should consider other monitoring methods suitable for subsidence monitoring and which can cover large areas, such as InSAR technology. Portable ground technology that produces high-resolution SAR images is the most suitable equipment (Canuti et al., 2002). Figure 6 illustrates the proposed deployment of the monitoring equipment at Jwaneng mine. Satellite images from the InSAR technology should be used to reconcile the monitoring systems at Jwaneng Mine. The InSAR technology tracks the impact of ground movement on infrastructure around the pit, dumps, and slimes dams (Altamira Information, n.d.). Figure 7 shows an example of a satellite image produced from InSAR. The magnitude of movement is presented in colour fringes in the comparison of satellite images from different dates. The images should be purchased on a quarterly basis, but more frequently should there be a need. At the start of monitoring, archived images should be used to identify hazard areas based on historical movements.

Figure 4—Systematic utilization of monitoring equipment

Figure 6—Monitoring equipment positioning

Figure 5—High-risk area associated with Cut 8 mining

The Journal of The Southern African Institute of Mining and Metallurgy

Figure 7—A satellite image from Altamira InSAR (Source: Altamira Information, n.d.) VOLUME 114

JUNE 2014

467

▲

Since the main treatment plant infrastructure is within 100 m of the mining activities in Cut 8, the built-up area should be monitored for movement. The mine should consider installing GPS receivers in this area. The GPS receivers should be strategically positioned to avoid measurement errors brought about by multi-pathing and dilution of geometric intensity of satellites because of the plant infrastructure. Multi-pathing and satellite availability can be a problem when monitoring around tall structures using GPS receivers. To compensate for the inaccuracies of GPS height measurements, the mine should use the precise

Design principles for optimizing an established survey slope monitoring system Data collection and processing

Monitoring procedures

This section discusses the data collection strategy suitable for Jwaneng Mine, focusing on the frequency of measurements and processing of the data for errors. The frequency of the slope monitoring measurements should be systematic and guided by rock behaviour. The movement rate of the rock should determine the frequency of the measurements. The frequency of the measurement can be determined as follows;

The next design criterion to consider is the monitoring procedures guiding the slope stability monitoring process. Jwaneng Mine procedures are categorized as follows:

➤ Movements of 0 to 2 mm per day are monitored once a month ➤ Movements of 0 to 5 mm per day are to be monitored once a week ➤ Movements of 5 to 10 mm per day to be monitored once every 2 days ➤ Movements of 10 to 50 mm per day will be monitored once per day ➤ Movements greater than 50 mm will require constant observation (Jooste, 2005). The measurement guidelines stated above are as applied at Venetia mine. It is recommended that each mine develops its own guidelines suitable for local prevailing conditions. The mine needs to be consistent with the checking of the positions of control points using the GPS post-processing method and precise levelling. These processes should be carried every six months as per survey procedures, and be repeated more frequently when movement limits are exceeded.

Code of practice (COP) All mines should develop a code of practice guiding slope stability monitoring. Although there are Acts guiding slope stability monitoring in Botswana, these are not very comprehensive. Such mines should look at Acts guiding slope stability monitoring in other countries for guidance, as the principles are the same. The South African Department of Mineral Resources (DMR) has prepared a guideline for the preparation of a COP to combat rockfall and slope instabilityrelated incidents in open pit mines (Cawood and Stacey, 2006). The guideline is available on the website of the Department of Mineral Resources (DMR, 2005). In developing g a COP, the mine could be guided by the following principles: ➤ Identification and documentation of rock-related incidents ➤ Development of appropriate strategies to eliminate or reduce risk caused by these hazards ➤ Allocation of duties for the execution of these strategies ➤ Training of personnel to enable them to carry out their duties (Gudmanz, 1998). The COP should be reviewed regularly to keep up with international standards guiding slope stability monitoring.

Process flows Analysis and reporting of monitoring results The mine should consider the following aspects when selecting the appropriate software to be used to analyse and report slope stability monitoring results. Since various instruments are used to collect slope stability monitoring data, there is need to integrate this data and analyse it from one point so that it can be subjected to the same level and standard of interpretation. If the data is analysed using the same software, it becomes easy to establish trends in data from different sources. Integration also allows for cross-checking between data sources (Abramson et al., 2001). Figure 8 illustrates how data from different sources can be integrated and the benefits derived. GIS is the most common software used to integrate data from various sources for analysis and presentation. Most GIS packages have least-square adjustment functionality for error analysis, graphic display functionality, and can produce movement graphs. GIS evolves with data collecting instruments, which makes it suitable for the ever-developing slope monitoring technology (Wolf and Ghilani, 1997). The other advantage of GIS is that because of its ability to handle large quantities of data, it can be used to manage other mine data such as rainfall figures, blasting data, pit dewatering information, and other hydrological data that has influence on the stability of pit slopes (Wolf and Ghilani, 1997). Integration allows for data from the various monitoring systems to be interpreted, analysed, and movement trend comparisons done within a short interval after collection. If data is allowed to accumulate without analysis, the integrity of the monitoring process will be compromised.

▲

468

JUNE 2014

VOLUME 114

These procedures list the step-by-step processes of slope stability monitoring activities. Examples of these procedures include the GeoMos operating procedure, SSR operating procedure, precise levelling procedure, and the GPS postprocessing procedure.

Warning systems and response This focuses on the action that will be taken when ground movements have been detected. The mine will develop guidelines on how to respond to movements of different magnitudes. The procedures listed above should be tested for practicability by running mock-ups regularly. The procedures must be stored in one place and made easily accessible.

Figure 8—Using GIS for data integration The Journal of The Southern African Institute of Mining and Metallurgy

Design principles for optimizing an established survey slope monitoring system Personnel responsibilities After the slope monitoring system has been implemented and procedures developed, there is a need to consider the personnel who operate the system. Geotechnical engineers are responsible for identification of hazardous areas and the classification of levels of risk. The level of risk determines the precision and the frequency of the measurements. The analysis and reporting of the monitoring results are also the responsibility of the geotechnical engineers. Mine surveyors are responsible for managing and maintaining the slope monitoring equipment in terms of availability and utilization. Furthermore, the surveyors are responsible for managing the data acquired by the monitoring equipment. They ensure that the data is processed for errors before being plotted for analysis. The mine surveyors are also responsible for the maintenance of the survey network. This maintenance is done by regularly carrying out activities such as GPS post-processing and precise levelling. The management of slope stability monitoring procedures is a joint responsibility of the mine surveyors and the geotechnical engineers. The information technology personnel are responsible for the security and backups of database storing the slope stability monitoring information. They ensure that the software used for analysis and the communication system used to relay slope stability information is always available. After these responsibilities have been allocated, a competency matrix is developed for each individual involved in the slope stability monitoring process. The competency matrix is then used to assess the level of competency, which informs the development programme for the individual.

Budget The mine needs to carry out a proper cost analysis to determine the cost implications. To justify the extra expenditure aimed at optimizing the existing design, the value-add of the new components should be clearly stated (Cawood and Stacey, 2006).

adjustment are still required ffor reliable results to be achieved. The amount of data that needs to be collected and analysed require a dedicated mine surveyor and a geotechnical engineer on a full-time basis.

Recommendation In addition to the strategy outlined in this paper, it is recommended that further research be conducted in the following areas. ➤ The correction for varying atmospheric conditions brought about by depth changes in the pit remains a challenge when using GeoMos and need to be investigated. It is critical to understand what actually happens to the signal that travels from the Total Station to the monitoring point. Varying temperatures and atmospheric pressure, coupled with dust and fumes in the pit, affect the accuracy of distance measurements and need to be investigated ➤ It is essential to develop a systematic approach to managing the large amounts of data collected by the different monitoring systems so that a single version of the truth can be detected from them. This approach should encompass data validation, processing, and interpretation ➤ Beacon design and construction standards should to be developed. These standards will ensure that the reference points for monitoring are robust and not easily affected by blasting activities. Challenges in the area of slope stability monitoring will always exist. The onus rests with mine surveyors and geotechnical engineers to turn these challenges into opportunities for continuous improvement. The current literature should be reviewed by the relevant parties, and they should participate in technical conference events.

References ABRAMSON, L.W., LEE, T.S., SHARMA, S., and BOYCE, G.M. 2001. Design, Construction and Maintenance. Slope Stability and Stabilization Methods. Wiley-Interscience, Hoboken, NJ. Chapter 8, pp. 604,607, 608–609, 626–627, 629.

A strategy for optimizing slope monitoring process has been developed. The efficiency of the monitoring system should be gauged by its ability to predict failures and its economic value-add during slope angle design. The strategy focused on large open pit mines, with the Debswana Jwaneng mine serving as a case study, It is concluded that slope monitoring requires a multifaceted approach focusing on the survey control network, beacon design and construction, the equipment shelter, equipment selection, data collection and processing, procedures, and personnel responsibilities. All of these factors are equally critical for an optimal monitoring process. Negligence in one area can negate all the good work done in other strategic areas, leading to unreliable monitoring results. It is evident that although slope monitoring has evolved over the years, with the process becoming more automated, basic survey principles such as working from whole to part, cross-checking, documentation of procedures, and error The Journal of The Southern African Institute of Mining and Metallurgy

AFENI, T.B. and CAWOOD, F. 2010. Do the properties of glass matter when taking Total Station distance measurements through an observation window? International Society for Mine Surveying XIV International Congress, Sun City, South Africa, 20–24 September 2010. p. 167. ALTAMIRA INFORMATION. Not dated. Ground motion InSAR. www.altamirainformation.com/html/1-18161-Techniques.php [Aaccessed 01/08/ August 2011]. BANNISTER, A., RAYMOND, S., and BAKER, R. 1998. Surveying. Pearson Education Limited. pp. 190, 207, 337-338, 431. BARTLEY, P. 2007. The nature of monitoring and evaluation; definition and purpose. www.scn.org/cmp/modules/mon-wht.htm, [Accessed 27 July /7/2010]. CANUTI, P., CASAGLI, N., MORETTI, S., LEVA, D., SIEBER, A. J., and TARCHI, D. 2002. Landslide monitoring by using ground-based radar differential interferometry. Proceedings of the First European Conference on Landslides, Prague, Chech Republic, June 2002, pp. 523–527. VOLUME 114

JUNE 2014

469

▲

Conclusion

Design principles for optimizing an established survey slope monitoring system CAWOOD, F.T. and STACEY, T.R. 2006. Survey and geotechnical slope monitoring considerations. Journal of the South African Institute of Mining and Metallurgy, vol. 106, no. 7. pp. 495–497, 500–501.

DMR. 2005. Guideline for the compilation of a mandatory code of practice to combat rock fall and slope instability related accidents in surface mines. www.dmr.gov.za/guidance-notes-for-medical-practitioners/finish/20mine-health-and-safety/350-combat-rockfall-and-slope-instabilityrelated-accidents-in-surface-mines/0.html [Accessed 8 May 2014]. GUDMANZ, K.M. 1998. The implementation of codes of practice, Symposium on Rock Mechanics and Productivity and the Implementation of Codes of Practice. West Rand, South Africa, 28 October 1998. Handley, M.F. (ed.), South African National Group of the International Society of Rock Mechanics. pp. 3–6.

Department of Geomatics, University of Melbourne, Victoria. [Accessed 13 December 2010]. LEICA GEOSYSTEMS. 2004. Reporter, r no. 50, Leica Geosystems, April 2004, p.12. MINING NEWS. 2006. Rockslide at Kumtor. www.miningnews.net/ /storyview.asp?storyid=62170&sectionsource=s0. 19 July 2006, [Accessed 29 July 2009]. MINING WEEKLY. 2010. Botswana, De Beers investing R25bn in ultra-rich Jwaneng Diamond Mine. www.miningweekly.com, 19 October 2010 [Accessed 25 November 2010]. READ, J. and STACEY, P. 2009. Guidelines for Open Pit Slope Design. CSIRO

JOOSTE, M.A. and CAWOOD, F.T. 2006. Survey slope stability monitoring: Lessons from Venetia Diamond Mine. International Symposium on Stability of Rock Slopes in Open Pit Mining and Civil Engineering, g Cape Town, 3–6 April 2006. pp. 361–363.

Publishing, Collingwood, Victoria, Australia. pp. 342, 346–353. WATT, I.B. 1996. Monitoring Surveys at Letlhakane and Orapa open pit mines. Consulting Report. March 1996. WOLF, R. and GHILANI, D. 1997. Adjustment Computations: Statistics and Least

JOOSTE, M.A. 2005. Slope Stability Monitoring In Open Pit Operations. Investigational Project for Masters Degree in Engineering, University of

▲

JUNE 2014

KEALY, A. 2004. 451-200 Survey Networks Theory, Design and Testing. www.geom.unimelb.edu.au/kealyal/200/Teaching/net_design_test.html

DEBSWANA., 2010. Jwaneng Long Term Plans. 2010.

470

the Witwatersrand, Johannesburg, 2005, pp. 16,27,36,37.

VOLUME 114

Squares in Surveying and GIS. John Wiley and Sons, Hoboken, NJ. pp. 1–11.

◆

The Journal of The Southern African Institute of Mining and Metallurgy

Estimation of rock strength from quantitative assessment of rock texture by C.A. Ozturk*, E. Nasuf*, and S. Kahraman†

Synopsis The compressive strength of rock (σ σc) has an important effect on design of structures in rock engineering. Compressive strength can be determined in the laboratory using the uniaxial compressive strength (UCS) test. Some other index tests, such as the point load test, are also used, particularly when suitable samples for UCS are not available. The quantification of rock texture has introduced a new method in rock engineering for estimating the mechanical and physical properties of rock materials from microscopic investigations. The aim of this study is to quantify rock texture to estimate rock strength from the texture coefficient (TC), which is determined from a statistical assessment of thin section images. Rock texture is quantified by twelve different images from a single thin section to increase the reliability of texture analysis. A data-set is prepared to investigate correlations between TC and σc. The statistical correlations are computed after classifying of the rock samples based on their lithology as well as grain features. Equations derived based on the results of this study are used to predict the approximate value of compressive strength from the texture coefficient. This method is particularly useful for preliminary studies in rock engineering projects prior to detailed site investigation. Keywords rock engineering, compressive strength, statistical correlation, grain features, texture coefficient.

beyond the scope off this paper, although they are useful for understanding the applications of the various test methods. Texture, which is defined as the degree of crystallinity (Williams et al., 1982), is the combination of mineral grains and matrix that includes the smallest particles of rock material. Nearly all mechanical and physical properties of interest to rock engineers depend on how the grains and matrix relate to the texture. A method for quantification of rock texture by combining geometrical parameters regarding grains and matrix was proposed by Howarth and Rowlands (1986, 1987). The proposed correlations of textural and mechanical properties from texture coefficient (TC) measurements were based mainly on statistical assessment. The purpose of the current study is to investigate the correlations between strength of rock material and rock texture properties in order to predict the UCS from studies of rock thin sections and to develop a statistical technique for thin section analysis.

Quantification of rock texture

The strength of rock masses depends mainly on the properties of the intact rock material and discontinuities. The strength of the intact rock material depends on the mineral composition and texture. The strength of rock masses can therefore be obtained by analysing the features of intact rock material and discontinuities (Singh and Goel, 2011). The compressive strength (σc) of rock is measured mostly from laboratory uniaxial compressive strength (UCS) tests. Some other index tests, such as point load index (Is) and Schmidt hammer rebound number (R), are also useful for determining the strength of rock materials. The methodology for UCS is standardized by International Society of Rock Mechanics (ISRM, 1981) and American Society for Testing and Materials (ASTM, 1984). The estimation of σc from index tests has been investigated by several researchers, based mostly on statistical investigations of datasets. The details of these investigations are The Journal of The Southern African Institute of Mining and Metallurgy

Rock texture is revealed by the geometrical relationships of the grains and matrix. Geometrical features of grains can be obtained from a photograph of a rock thin section. The following features have been used quantify rock texture.

Grain shape and size Grain shape can be quantified by measuring grain length (major axis) and width (minor axis), as well as perimeter and area. Grain shape is classified based on roundness and sphericity (Cox and Budhu, 2008). Roundness

* Istanbul Technical University, Mining Engineering Department, Istanbul, Turkey. † Hacettepe University, Mining Engineering Department, Ankara, Turkey. © The Southern African Institute of Mining and Metallurgy, 2014. ISSN 2225-6253. Paper received r Oct. 2012; revised paper received Apr. 2014. VOLUME 114

JUNE 2014

471

▲

Introduction

Estimation of rock strength from quantitative assessment of rock texture is the ratio off curvature off a grain’s edges to overall grain shape (Wadell, 1932), and sphericity is the ratio between grain volume and the smallest circumscribing sphere (Krumbein, 1941). These two parameters are formulized as (Cox and Budhu, 2008):

Dreyer (1973) also proposed an equation to quantify f the homogeneity of grains by means of an index of grain size homogeneity (t) (Equation [5]). [5]

[1]

[2] Brace (1961) reported a linear relationship between grain shape and strength, while increasing size of grains leads to a decrease in strength. Ehrlich and Weinberg (1970) proposed a grain roughness coefficient that was also used by Onodera and Kumara (1980). The roughness coefficient showed correlation between rock strength, fracture propagation, and grain shape. Beddow and Vetter (1977) described grain shape from Fourier series, while Kaye (1982) defined grain shape using fractal dimension. Prikryl (2001) investigated the effect of grain size on strength and proposed a logarithmically inverse relation between grain size and rock strength, as previously suggested by Eberhardt et al. (1999). Prikryl (2006) also proposed that grain shape and size can be characterized by using the following parameters: A Area Lp Perimeter Dmax, Dmin Major and minor axis length (Herdan and Smith, 1953) Dequiv Equivalent diameter (Petruk, 2000) C Compactness FF Form factor (Howarth and Rowlands, 1986, 1987) AR Aspect ratio GBS Grain boundary smoothness. Schematic views of these properties and calculations are given in Figure 1 and the following relationship.

[3]

where Aavg = Average cross-sectional area of grains Ai = Individual grain area. The high values of these parameters indicate the complexity of intergranular relationships in texture. As reported by Howarth and Rowlands (1987) from the study of Hoek (1965), the degree of grain interlocking plays an important role in the resistance of rock against applied stress by increasing the strength of the material. Higher values of stress are required to initiate cracks at grain boundaries for tightly packed and well cemented grains such as in igneous and metamorphic rocks.

Grain and matrix relationship The matrix also has an important effect on the strength of rock. Rock strength is controlled by the strength of matrix material, which is generally less than the grain strength. Generally, cracks start inside the matrix under stress conditions and propagate through the matrix material. Packing density is also correlated with strength properties. Increasing the packing density increases the strength (Bell, 1978). Voids between the grains (porosity) that are not cemented with matrix also have an adverse effect on the strength of material (Price, 1960; Smordinov et al., 1970; Dube and Singh, 1972; Tugrul and Zarif, 1999). Howarth and Rowlands (1987) proposed a relationship (Equation [6]) W) as to quantify packing density, which is area weighting (AW relative proportions of grains and matrix. [6]

Effect of mineral content Numerous studies have been conducted regarding the relationships between geomechanical and petrographical properties, including mineralogical content of intact rock material, mostly focusing on granitic rocks and indicating the importance of petrographic investigation (Mendes et al., 1966; Willard and McWilliams, 1969; Hallbauer et al., 1978; Irfan and Dearman, 1978; Fahy and Guccione, 1979; Verhoef and Van DeWall, 1998; El Bied et al., 2002). There are also

Consequently, the results show that there is an inverse relation between grain size and rock strength. With decreasing grain size, rock strength tends to increase and vice versa (Olsson, 1974; Hugman and Friedman, 1974; Onodera and Kumara, 1980; Tugrul and Zarif, 1999).

Grain relationships The complex relationships between adjacent grains have been characterized by Dreyer (1973). He proposed an index of interlocking (g), given by: [4] where n = Number of grains Lpi = Fraction of the perimeter of grain contacting the adjacent grain.

▲

472

JUNE 2014

VOLUME 114

Figure 1—Demonstration of grain shape parameters (Howarth and Rowlands, 1987) The Journal of The Southern African Institute of Mining and Metallurgy

Estimation of rock strength from quantitative assessment of rock texture some studies on estimating rock strength ffrom petrographical properties based on statistics (Shakoor and Bonelli, 1991; Richards and Bell, 1995; Bell and Culshaw, 1998; Prikryl, 2001). Previous research has focused mainly on the relation between quartz content and rock strength. Merriam et al. (1970) proposed a relationship between quartz content and tensile strength, while Irfan and Dearman (1978) developed a micropetrographic index that shows good correlation with geomechanical parameters for granite. This was also previously investigated and proposed by Mendes et al. (1966). However, most of the studies have shown conflicting results due to the nature of the material. For instance, Gunsallus and Kulhawy (1984) also found a definite relation between quartz content and rock strength. However, Bell (1978), Fahy and Guccione (1979), Shakoor and Bonelli (1991), and Tugrul and Zarif (1999) found no correlation. Similarly, Prikryl (2006) found no significant correlation between mineral content and mechanical properties, except for mica content. Phillipson (2008) also reported that there is no correlation between quartz content and rock strength in shale; on the other hand, he also proposed that sutured mica grains do not affect rock strength. Consequently, although nearly 50 years of investigations have shown that minerals such as quartz, feldspar, mica, etc. affect rock strength, owing to the complex textures as well as random behaviour, the relation between mineral content and strength has not been quantified to date.

Texture coefficient Rock texture, which is characterized by the relative amounts of minerals, grain sizes and shapes, and the manner in which the grains interlock (Merriam et al., 1970) has been described by several researchers. The pioneering work on quantification of rock texture was proposed by Howarth and Rowlands (1986; 1987). The main differences between the texture coefficient and the other texture models are based on the parameters used for the characterization. The quantitative assessment of rock texture depends on the parameters given in Equation 7. [7]

where

N0, N1 = The number of grains not elongated and elongated FF F0 = AR1 = AF1 = AW =

(see below) Form factor for grain circularity Aspect ratio Angle factor for orientation Area weighting for degree of grain packing.

Hence, this concept takes into account more geometrical properties of grains as well as grain and matrix relations, except mineral contents. In the formula, N0 and N1 are the number of grains whose aspect ratios are below and above a pre-set discrimination level respectively, which is defined as 2.0 by Howarth and Rowlands (1986, 1987). Grains with AR greater than 2.0 are described as elongated. FF F0 is the arithmetic mean of FF of grains that are not elongated as calculated by Equation [3]. AR1 is the arithmetic mean of the AR of elongated grains as per Equation [3]. AF1 is used to quantify orientation for elongated grains, while AW is calculated from Equation [6]. The application details of the texture coefficient are comprehensively described by Howarth and Rowlands (1987). The proposing of texture coefficient can also make it possible to investigate the relation between texture and mechanical properties of rocks, summarized in Table I. Although previous investigations (e.g. Tiryaki and Dikmen, 2006) included the relationship between TC and several mechanical, physical, drillability, and cuttability properties of rock, only the relationship between the TC and rock strength is given in Table I. Although TC is useful for predicting mechanical properties of rocks, especially for similar lithologies (Ersoy and Waller, 1995; Azzoni et al., 1996; Prikryl, 2006), there are also contradictory correlations between TC and strength (Ozturk et al., 2004; Alber and Kahraman, 2009). This is because the complexity of rock texture limits the correlation between texture and mechanical properties. However, texture and strength can be correlated at least to the extent that the strength class of rock material can be established, and this can also encourage researchers to extend experimental studies for the application of texture coefficient in rock engineering. On the other hand, the TC values in these previous studies were derived from only a single photograph.

Table I

Summary of TC application for uniaxial compressive strength estimation Sr# 1

Relation

Reference

σc = 104.80 x TC-55.14, r2 = 0.92

Howarth and Rowlands (1987)

Remarks Dry rock material.

σc = 96.4 x TC-56.48, r2 = 0.91

r2 = 0.62

Ersoy and Waller (1995)

Lithology: sandstone, limestone, siltstone, granite, diorite.

n/a

Azzoni et.al. (1996)

The authors confirmed the results of Howarth and Rowlands (1987) for granite, marble, and sandstone. The same results could not be obtained for other rock types such as gneiss, rhyolite, etc.

r2 = 0.11

Ozturk et.al. (2004)

No valid correlation was found for rocks that are mainly in micritic form, such as andesite.

r2 = 0.10 (for mean σc)

Prikryl (2006)

No valid correlation was found for granite and orthogneiss samples except for mica content and strength relation.

σc = -131.86xTC+86.20, r2 = 0.90

Alber and Kahraman (2009)

A well-correlated relation was found for fault breccia, especially for TC values between 0.30 and 0.60.

The Journal of The Southern African Institute of Mining and Metallurgy

VOLUME 114

JUNE 2014

473

▲

Saturated rock material

Estimation of rock strength from quantitative assessment of rock texture The current investigation uses a statistical assessment off quantified data from 12 photographs of a thin section to obtain a unique TC, thus increasing the reliability of estimation models for strength compared with the studies summarized in Table I.

399 m. Bilginet et al. (1988) researched the cuttability off a road header based on rock material and rock mass properties in a sewerage tunnel driven in Eyup, Istanbul in western Turkey. Kahraman (2001) and Kahraman et al. (2005) investigated the relationships between mechanical parameters of rock samples from different parts of Turkey. Ozturk and Nasuf (2002) studied the effect of loading conditions on texture in granites from the Balikesir region, also in western Turkey. All of these studies are used to obtain a wellestablished data-set to investigate the texture and strength relationship from experimental studies of thin sections.

Experimental studies The relationship between texture and strength of intact rock material was investigated by image analysis of thin sections of rock samples studied previously by Bilgin and Shahriar (1987), Bilginet et al., (1988), Kahraman (2001), Kahraman et al. (2005), and Ozturk and Nasuf (2002) that had been retained for further analysis. These studies and materials are summarized briefly below. Bilgin and Shahriar (1987) carried out an extensive rock cuttability test on samples taken from the Zonguldak Amasra bituminous coal basin in northern Turkey. The samples were obtained from a borehole drilled to a depth of nearly 1000 m, and were taken from depths ranging between 40 m and

Strength of rock materials Although the aforementioned studies investigated a series of mechanical and physical properties of rock material, only uniaxial compressive strength is of interest in the current investigation. Table II and Figure 2 present the location, rock type, and the rock strength (UCS) of the samples used in this study.

Table II

Strength of rock samples used in this study Sample code

Sample depth or location

A.1 A.2 A.3 A.4 A.5 A.6 A.7 A.8 A.9 A.10 A.11

Zonguldak Amasra Coal Basin (depth of samples m

B.1 B.2 B.3 B.4 B.5 B.6 B.7 B.8

Istanbul Eyup sewerage tunnel (tunnel m)

C.1 C.2 C.3 C.4 C.5 C.6 C.7 C.8 C.9 C.10 C.11 C.12 C.13 C.14 C.15 C.16 C.17 C.18 C.19 C.20 C.21 C.22 C.23 C.24 C.25 C.26 C.27

Gaziantep/Erikli Pozanti Osmaniye/Bahce Gaziantep/Erikli Gaziantep/Erikli Osmanbey/Bahce Gaziantep/Erikli Konya Kayseri/Yahyali Adana Misis Kutahya/Tuncbilek Kutahya/Emet Kutahya/Emet Konya/Karaman Icel/Mut Konya/Godene Antalya/Demre Antalya/Korkuteli Antalya/Finike Burdur/Bucak Antalya/Demre Sivas/Yildizeli Kayseri/Yahyali Kayseri/Bunyan Aksaray/Ortakoy Balikesir

â&#x2013;˛

474

JUNE 2014

VOLUME 114

Lithology 40 71 221 259 278 336 315 355 367 397 399 476 899 1334 1337 1452 1695 1752 1802

Limestone Andesite Tuff Tuff Limestone

UCS (MPa) 53.71 52.29 52.26 53.03 44.69 35.00 52.62 38.35 27.91 28.84 43.25

Siltstone Shale Shale Dyke Mudstone Sandstone Sandstone Siltstone

104.10 82.90 126.20 127.90 146.50 154.10 132.80 55.10

Serpentine Limestone Sandstone Diabase Marl Altered sandstone Limestone Serpentine Haematite Limestone Limestone Marl Sandstone Limestone Travertine Travertine Travertine Travertine Marble Travertine Travertine Travertine Travertine Limestone - dolomitic Limestone Granite Granite

69.11 123.80 45.20 110.90 39.50 20.10 51.30 54.30 61.80 15.70 85.20 21.40 70.50 42.10 50.30 60.00 45.40 57.60 134.20 80.00 50.30 112.30 83.30 136.70 175.00 114.50 106.30

The Journal of The Southern African Institute of Mining and Metallurgy

Estimation of rock strength from quantitative assessment of rock texture

Figure 2—Location map of rock samples

Texture analysis The TC for each rock sample was determined by analyses of textural properties. The main steps in TC determination from a thin section are summarized in Figure 3, and are divided into two groups. The first steps involve obtaining geometrical features of grains as independent parameters from image processing, and the second determining the dependent parameters for TC value. Thin sections were split into frames and images captured from a camera mounted on the microscope were taken from the centre of each frame (Figure 4). Using this method, multiple images from a single thin section can be used for a statistical assessment after determining the TC values for each image. The statistical assessment of all of these images can be used to calculate the average TC values, which decreases the possible bias based on microscopic analysis. The main difficulty in using texture analysis to obtain a reliable result from image processing is defining grain and matrix boundaries properly. As can be seen in Figure 1, drawing the perimeter of each grain is sometimes very difficult due to the subjective approach and complex structure of grain sutures. This is mainly because of picture frame limitations and the non-homogenous structure of texture. To eliminate these drawbacks, it is proposed to take at least twelve photographs by splitting the thin section. Statistical analysis of the data from image processing of each photograph is used to obtain a number that corresponds to a dimensionless value for rock texture. The study also proposes to correlate the number of photographs with the variance of the TC. If the variances of the overall data analysis for the TC are high, it is recommended that more photographs be taken to decrease the variance and bias as well as for re-analysis. The results of the texture analysis experiments based on the proposed technique are presented in Table III. The data-set used in this study represents the first application of this statistical evaluation technique of texture analysis, as summarized in Figure 3.

These basic statistical assessments are useful f in understanding the validity of strength prediction from texture as proposed in this study. Histograms of strength and TC values as well as statistical parameters show that the strength values are clustered around 75 MPa, while the minimum and maximum strength values are 15.70 MPa and 175.00 MPa. On the other hand, nearly half the samples, which have compressive strength value more than 100 MPa, have a variance greater than 0.15. Texture analysis of rocks stronger than 100 MPa should be conducted in detail by taking more photographs, and the image analysis results from each photograph should be checked statistically to decrease the variation of texture analysis. The number of grains in each texture photo should be between 20 and 50, as proposed by Howarth and Rowlands (1987). On the other hand, although it was proposed to take the AW as 1.0 for igneous rocks, AW values ranged from 0.95 to 1.0 for the samples coded as A[3;4], B4, C[1;4;9;19;26]. The possible causes of these small deviations from 1.0 are as follows.

Figure 3—Procedure for quantifying rock texture

Statistical assessments of the results

The Journal of The Southern African Institute of Mining and Metallurgy

Figure 4—Splitting of thin section to capture photographs VOLUME 114

JUNE 2014

475

▲

The statistical assessments of the results obtained from the strength and texture properties of the samples are given in Table IV and Figure 5.

Estimation of rock strength from quantitative assessment of rock texture Table III

Experimental studies – results of texture analysis (Nasuf and Ozturk, 2005; Ozturk, 2006) Sample code

Av. AW

Av. N0

Av. N1

Av. FF0

Av. AR1

Av. AF1

Av. TC

Var.* TC

A.1 A.2 A.3 A.4 A.5 A.6 A.7 A.8 A.9 A.10 A.11

0.25 0.18 0.98 0.96 0.45 0.32 0.11 0.26 0.15 0.12 0.18

17.50 19.83 21.72 39.97 21.50 20.17 33.50 18.83 22.00 27.80 25.00

4.00 4.83 2.89 1.39 4.25 7.83 7.17 6.00 3.17 3.60 3.80

0.58 0.59 0.89 0.91 0.58 0.59 0.63 0.57 0.65 0.65 0.63

2.41 2.48 2.02 2.03 2.50 2.63 2.36 2.38 2.04 2.30 2.87

1.28 1.59 1.56 1.07 1.52 2.70 2.82 2.19 0.43 1.26 1.14

0.53 0.41 1.34 1.06 0.91 1.02 0.26 0.67 0.24 0.21 0.38

0.03 0.03 0.04 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

B.1 B.2 B.3 B.4 B.5 B.6 B.7 B.8

0.52 0.22 0.32 0.96 0.49 0.54 0.50 0.32

19.80 21.83 17.83 45.36 20.17 21.83 31.50 18.00

2.20 5.50 6.83 2.18 7.33 6.00 7.50 2.39

0.59 0.62 0.56 0.91 0.62 0.58 0.64 0.64

2.64 2.64 2.44 2.03 2.68 2.40 2.45 2.24

0.49 1.41 2.43 1.02 2.27 2.23 2.10 0.60

0.86 0.50 0.94 1.07 1.37 1.41 1.20 0.53

0.07 0.03 0.03 0.02 0.04 0.23 0.24 0.01

C.1 C.2 C.3 C.4 C.5 C.6 C.7 C.8 C.9 C.10 C.11 C.12 C.13 C.14 C.15 C.16 C.17 C.18 C.19 C.20 C.21 C.22 C.23 C.24 C.25 C.26 C.27

0.97 0.48 0.43 0.97 0.16 0.34 0.33 0.22 0.96 0.27 0.47 0.11 0.30 0.21 0.41 0.61 0.60 0.40 0.83 0.50 0.50 0.45 0.60 0.70 0.70 0.95 1.00

43.29 20.67 49.57 32.75 22.50 22.00 12.10 16.83 46.79 17.50 23.00 17.00 20.50 14.40 20.33 22.00 25.75 22.20 34.54 31.20 24.67 23.29 23.43 21.88 12.58 13.58 13.63

3.18 7.67 9.71 6.21 6.17 4.67 7.9 6.83 1.95 3.33 7.50 4.00 6.17 6.6 3.33 1.80 2.00 1.40 3.89 5.00 2.33 5.43 1.14 7.38 9.21 6.97 5.00

0.91 0.64 0.66 0.58 0.67 0.68 0.7 0.72 0.91 0.81 0.66 0.71 0.7 0.71 0.71 0.74 0.80 0.71 0.88 0.68 0.80 0.69 0.76 0.69 0.65 0.59 0.58

2.09 2.34 2.99 2.23 2.53 2.51 2.77 2.51 2.05 2.65 2.16 2.50 2.44 2.14 2.86 2.47 2.41 2.38 2.05 2.47 2.37 2.39 2.27 2.61 2.85 2.78 2.60

1.23 1.92 2.64 1.59 3.08 1.64 1.64 1.14 0.62 0.70 1.90 0.55 1.59 0.97 1.24 1.05 0.53 0.39 2.12 1.85 0.64 2.25 0.83 2.51 1.45 0.91 1.30

1.13 1.13 1.13 1.94 0.46 0.66 0.88 0.40 1.04 0.37 1.00 0.16 0.63 0.34 0.85 0.90 0.75 0.56 1.22 1.01 0.62 1.03 0.80 1.98 1.84 1.89 2.22

0.01 0.04 0.07 0.13 0.04 0.03 0.11 0.01 0.01 0.03 0.03 0.01 0.07 0.01 0.08 0.01 0.01 0.01 0.01 0.07 0.01 0.64 0.01 0.53 0.66 0.44 0.04

*Variance

Table IV

Statistical assessment of strength and TC Parameter

Compressive strength

74.53 40.74 1,660.03 15.70 175.00 0.68 -0.63

0.91 0.50 0.25 0.16 2.22 0.76 0.15

Mean Standard deviation Sample variance Minimum value Maximum value Skewness Kurtosis

The volcanic glass in andesite (A.3, A.4) is regarded as non-granular material ii) The dyke consists of 85% crystalline feldspar and some of the remaining grains are altered for B.4, based on mineralogical investigation

▲

476

JUNE 2014

VOLUME 114

Figure 5—Histograms of compressive strength (a) and TC values (b)

iii) Due to the grain contacts and in situ stress conditions, voids are observed among grains in serpentine and diabase C.1 and C.4) iv) The alteration of haematite is the reason for the AW value of 0.96 for C.9 The Journal of The Southern African Institute of Mining and Metallurgy

Estimation of rock strength from quantitative assessment of rock texture v)

The AW off 0.83 ffor C.19 is due to a matrix off micritic limestone, as shown by mineralogical analysis vi) Due to the high values of elongation caused by in situ stress for granite (C.26), heavily milled quartz material is taken as non-granular, resulting in an AW of 0.95.

Strength and texture relationships The relationships between rock strength and quantified rock texture are obtained by regression based on constructing a scattergram and fitting a trend line. The value of the correlation coefficient (r2) was used to ascertain the validity of the trend line. Investigations for the prediction models were carried out based on the lithology of the materials and geometrical features of the texture constituents.

Investigation based on lithology Firstly, estimation was performed and a trend line was obtained as shown in Figure 6, without classifying the data. However, the results do not indicate any significant correlation due to the low value of the correlation coefficient (r2=0.52). It can be seen that an increasing value of TC indicates an increasing compressive strength. In the next step, the data for sandstone, siltstone, marl, and shale was separated (Figure 7). The correlation coefficient of the dotted line (r2=0.93) is quite acceptable when the two data points C.3 and C.6, which represent sandstone and altered sandstone, are omitted from the graph due to their magnitudes of scattering. The model for limestone can be seen in Figure 8. According to the result of the analysis, compressive strength can be reliably estimated from the rock texture coefficient (r2=0.87). The results show that classifying the data based on lithology can increase the reliability of the models for estimating rock strength from the TC. The values of r2 that can be used to ascertain the reliability of the models are different in the regression models. Increasing or decreasing values of r2 can be explained only by differences based on lithology. Consequently, materials consisting of small, wellrounded grains, such as sandstone or siltstone, display a more reliable relationship between texture and strength. This is probably the most important reason for the high values of r2 obtained for sandstone, siltstone, marl, and shale. The reliability of the regression model proposed for limestone is

lower due to the larger and more elongate grains compared with the sandstone samples.

Investigation based on grain features Correlation of rock strength versus TC was also investigated based on grain features. Each data sample was classified F0) and aspect according to the value of the form factors (FF ratios (AR1), which were used to characterize the circularity as well as elongation of grains. A regression model was F0 values applied for the TC and σc of the data with FF between 0.65 and 0.88, and AR1 between 2.04 and 2.61, and a trend line between quantified rock texture and strength was constructed (Figure 9). The intervals are selected to propose a model for rocks that consist of grains with intermediate elongation and circularity. The results show that rock strength tends to increase with increasing TC and can be estimated from the equation given in Figure 9, with r2=0.76.

Figure 7—Correlation between σc and TC for sandstone, siltstone, marl, and shale

Figure 8—orrelation between σc and TC for limestone

The Journal of The Southern African Institute of Mining and Metallurgy

Figure 9—Correlations between σc and TC based on grain features VOLUME 114

JUNE 2014

477

▲

Figure 6—Correlation between σc and TC

Estimation of rock strength from quantitative assessment of rock texture A summary off the output off this study ffor the prediction is given in Table V.

Discussion Apart from the current studies to quantify rock texture by using the TC, a procedure is given based on statistical evaluation of a thin section of a rock sample. This procedure is based on dividing the thin section by means of a grid of squares and then capturing images from the centre of each square. Hence, at least 12 photos of a single thin section are processed by image analysis and statistical analysis can be applied to obtain a dimensionless value of rock texture. Since texture analysis sometimes delivers contradictory results, as summarized in Table I, a greater number of photographs from a single thin section are taken for analysis in order to eliminate this difficulty, and statistical analysis is used to obtain the optimum value of the TC. The number of images should be based on the variance of the TC. The results of the experimental analyses show that if the rock strength is greater than 100 MPa, it is preferable to take more images for processing due to the increasing variance for the TC. The increasing strength values may be the consequence of the increases in TC due to the greater number of elongated and non-circular grains. This probably explains the high variance obtained with higher strength values. The data-set obtained from the studies was used to set up regression models to estimate rock strength from the TC. Four different regression models are proposed to predict rock strength based on lithology and grain features. Firstly, the plot of σc versus TC for all samples shows a high degree of scatter. The result shows that although there is a broad trend between σcc and TC, strength cannot be estimated due to the poor value of the correlation coefficient (r2=0.52). However, this trend can be used to classify the rock samples based on relative strength. Hence, the data is classified based on lithology. A regression model, which is σc=106.51xTC+7.46, is proposed for the materials classified as sandstone, siltstone, marl, and shale. The high correlation coefficient of 0.93 shows that the strength can be reliably predicted from the TC. The regression model proposed for limestone, that is σc=72.37xTC+10.38, also has a high r2 (0.87). The strength and the TC values for limestone display a linearly increasing trend, which is also valid for the materials classified as sandstone, siltstone, marl, and shale. These results show that classifying the rock materials based on lithology increases the reliability of the prediction models. Apart from the lithological classification, grain features are used to classify the rock F0 and AR1 parameters. FF F0 is taken materials, using FF between 0.65 and 0.88, while AR1 is taken between 2.04 and

2.61 ffor this classification f to obtain materials with grains that are neither over-elongated nor over-circular. A regression model obtained from this classification, that is σc = 70.83xTC+12.83, with r2=0.76, is usable for estimation, but is not as reliable as the regression models determined from lithological classifications. This equation is proposed to predict strength values for materials that are not classified as either sandstone, siltstone, marl, shale, or limestone. These proposed regression models can be particularly useful for projects that involve the mechanical properties of rock environments. The uniaxial compressive strength value for rock materials is one of the important and useful parameters employed in design studies carried out in rock environments. The proposed equations are also useful for predicting the value of compressive strength from a rock thin section obtained from a simple rock sample. These investigations will be particularly applicable for projects at the prefeasibility stage. Ozturk and Nasuf (2013) demonstrate a simple application of strength classification of rock materials based on rock texture that also used some the data from this study. In their example, the value of the intact rock rating for rock mass rating classification system (RMR) is predicted from TC, which is a good example of the application of texture analyses in rock engineering.

Conclusions The texture coefficient (TC) was used to interpolate the correlation between rock strength and texture so as to estimate the uniaxial compressive strength (σc) of rock material. In this study, a procedure is proposed to determine the TC from 12 images taken from a single rock thin section. This procedure can be used to increase the reliability of the texture analysis, which sometimes suffers from subjective interpretation in defining the grain boundaries. A higher variance for the TC indicates that more images should be used for image processing to decrease the variance and increase the reliability of the TC value for quantifying rock texture. A data-set including 46 data samples from 15 rock samples of different lithologies, with strength values ranging from 15.70 MPa to 175.00 MPa, was used in this study based on statistical assessment of texture analyses. Four different regression models are also proposed. The first may be of use only to establish the strength class of intact rock, due to its low correlation coefficient (r2) value. The classification of the material based on lithology is used to propose two more regression models. The one can be used to predict σc for sandstone, siltstone, marl, and shale, while the second is used for limestone. The final regression model is proposed to

Table V

Summary of regression models Model no. 1 2 3 4

▲

478

Equation

Correlation coefficient

Remarks

σc = 58.86 x TC + 21.16 σc = 106.51 x TC + 7.46 σc = 72.37 x TC + 10.38 σc = 70.83 x TC + 12.83

0.52 0.93 0.87 0.76

Non-classified Sandstone, siltstone, marl, and shale Limestone 0.65 < FF0 < 0.88 and 2.04 < AR1< 2.61

JUNE 2014

VOLUME 114

The Journal of The Southern African Institute of Mining and Metallurgy

Estimation of rock strength from quantitative assessment of rock texture predict σc from f the TC based on grain ffeatures. Form ffactor F0) and aspect ratio (AR10) are used as the parameters to (FF represent grain features of intact rock. These results show that the classification of the intact rock based on lithology increases the reliability of the prediction models derived from regression analysis. Further investigations based on different lithologies are highly recommended to investigate the feasibility of estimating at least the strength class of rock material by microscopic investigations.

References ALBER, M. and KAHRAMAN, S. 2009. Predicting the uniaxial compressive strength and elastic modulus of a fault breccia from texture coefficient. Rock Mechanics and Rock Engineering, g vol. 42. pp. 117–127. ASTM. 1984. Standard test method for unconfined compressive strength of intact rock core specimens. Soil and Rock, Building Stones. Annual Book of ASTM Standards 4.08. American Society for Testing and Materials, Philadelphia, PA. AZZONI, A., BAILO, F., RODENA, E., and ZANINETTI, A. 1996. Assessment of texture coefficient for different rock types and correlation with uniaxial compressive strength and rock weathering. Rock Mechanics and Rock Engineering, g vol. 29, no.1. pp. 39–46. BEDDOW, J.K. and VETTER, A.F. 1977. The use of classifiers in morphological analysis of articulates. Journal of Powder and Bulk Solids Technology, vol. 1. pp. 42–45. BELL, F.G. 1978. The physical and mechanical properties of the Fell sandstones, Northumberland, England. Engineering Geology, vol. 12. pp. 1–29.

GUNSALLUS, K.L. and KULHAWY, F.N. 1984. A comparative evaluation off rock strength measures. International Journal of Rock Mechanics and Mining Science and Geomechanical Abstracts, vol. 21, no. 5. pp. 233–248. HALLBAUER, D.K., NIEBLE, C., BERARD, J., RUMMEL, F., HOUGHTON, A., BROCH, E., and SZLAVIN, J. 1978. Suggested methods for petrographic description. ISRM commission on standardization of laboratory and field tests. International Journal of Rock Mechanics and Mining Science, vol. 15. pp. 41–45. HERDAN, G. and SMITH, M.L. 1953. Small Particle Statistics. Elsevier, Houston, TX. HOEK, E. 1965. Rock fracture under static stress conditions. Report no. MEG 383. CSIR, Pretoria, South Africa. HOWARTH, D.F. and ROWLANDS, J.C. 1986. Development of an index to quantify rock texture for qualitative assessment of intact rock properties. Geotechnical Testing Journal, vol. 9. pp. 169–179. HOWARTH, D.F. and ROWLANDS, J.C. 1987. Quantitative assessment of rock texture and correlation with drillability and strength properties. Rock Mechanics and Rock Engineering, g vol. 20. pp. 57–85. HUGMAN, R.H. and FRIEDMAN, M. 1979. Effects of texture and composition on mechanical behaviorof experimentally deformed carbonate rocks. Association of Petroleum Geological Bulletin, vol. 63, no. 9. pp. 1478–1489. IRFAN, T.Y. and DEARMAN, W.R. 1978. The engineering petrography of a weathered granite in Cornwall, England. Quarterly Journal of Engineering Geology, vol. 11. pp. 233–244. INTERNATIONAL SOCIETY OF ROCK MECHANICS (ISRM). 1981. Rock characterization testing and monitoring. ISRM Suggested methods. Brown E.T. (ed.). Pergamon Press, Oxford.

BELL, F.G. and CULSHAW, M.G. 1998. Petrographic and engineering properties of sandstones from the Sneinton Formation, Nottinghamshire, England. Quarterly Journal of Engineering Geology, vol. 31. pp. 5–19.

KAHRAMAN, S., GÜNAYDıN, O., and FENER, B. 2005. The effect of porosity on the relation between uniaxial compressive strength and point load index. International Journal of Rock Mechanics and Mining Science, vol. 42. pp. 584–589.

BILGIN, N., SEYREK, T., and SHAHRIAR, K. 1988. Roadheaders in Istanbul Golden Horn clean-up contributes valuable data. Tunnels and Tunnelling, g June. pp. 41–44.

KAHRAMAN, S. 2001. Evaluation of simple methods for assessing the uniaxial compressive strength of rock. Engineering Geology, vol. 38. pp. 981–994.

BILGIN, N. and SHAHRIAR, K. 1987. Developing a measurement system for mechanized excavation in mines and application to Amasra bituminous coal basin of Turkish Hard Coal Enterprise (in Turkish). Report no. 674. Science and Technology Research Council of Turkey (TUBITAK), Istanbul. BRACE, W.F. 1961. Dependence of fracture strength of rocks on grain size. Proceedings of the 4th Symposium on Rock Mechanics, University Park, Penn., PA. pp. 99–103. BROCH, E. and FRANKLIN, J.A. 1972. Point-load strength test. International Journal of Rock Mechanics and Mining Science, vol. 9, no. 6. pp. 669–697. COX, M.R. and BUDHU, M.A. 2008. Practical approach to grain shape quantification. Engineering Geology, vol. 96. pp. 1–16. DREYER, W. 1973. The Science of Rock Mechanics. Part I. The Strength Properties of Rocks. 2nd edn. Series on Rock and Soil Mechanics, 1 (1971/73). Trans. Tech. Publications, Clausthal. DUBE, A.K. and SINGH, B. 1972. Effect of humidity on tensile strength of sandstone. Journal of Mining Metallurgy and Fuel, vol. 20, no. 1. pp. 8–10. Eberhardt, E., Stimpson, B., and Stead, D. 1999. Effects of grain size on the initiation and propagation thresholds of stress-induced brittle fractures. Rock Mechanics and Rock Engineering, g vol. 32, no. 2. pp. 81–99. EHRLICH, R. and WEINBERG, B. 1970. An exact method for characterization of grain shape. Journal of Sedimentary Petrology, vol. 40, no. 1. pp. 20–-212. EL BIED, A., SULEM, J., and MARTINEAU, F. 2002. Microstructure of shear zones in Fontainebleau sandstone. International Journal of Rock Mechanics and Mining Science, vol. 39. pp. 917–932.

KAYE, B.H.1982. Review of new methods for characterizing the shape and texture of fine particles. Journal of Powder and Bulk Solids Technology, vol. 6, no. 2. pp. 1–4. KRUMBEIN, W.C. 1941. Measurement and geological significance of shape and roundness of sedimentary particles. Journal of Sedimentary Petrology, vol. 11. pp. 64–72. MENDES, F.M., AIRES-BARROS, L., and RODRIGUES, F.P. 1966. The use of modal analysis in the mechanical characterization of rock masses. Proceedings off the 1st International Congress On Rock Mechanics, Lisbon, vol. 1. pp. 217–223. MERRIAM, R., RIEKE, H.H., and KIM, Y.C. 1970. Tensile strength related to mineralogy and texture of some granitic rocks. Engineering Geology, vol. 4. pp. 155–160. NASUF, E. and OZTURK, C.A. 2005. Investigation into the texture properties of rock by image processing and correlations between texture and mechanical properties (in Turkish). Report no. 199Y046. Science and Technology Research Council of Turkey (TUBITAK), Istanbul, Turkey. OLSSON, W.A. 1974. Grain size dependence on yield stress in marble. Journal of Geophysical.Research, vol. 79, no. 32. pp. 4859–4861. ONEDERA, T.F. and ASOKA KUMARA, K.H.M. 1980. Relation between texture and mechanical properties of crystalline rocks. Bulletin of the International Association of Engineering Geology, vol. 22. pp. 173–177. OZTURK, C.A. 2006. Classification of rock texture properties and rock engineering applications (in Turkish). PhD thesis, Istanbul Technical University, Istanbul, Turkey. OZTURK, C.A. and NASUF, E. 2002. Investigation into the textural properties of rocks under different load conditions (in Turkish). Proceedings of the 6th National Rock Mechanics Symposium, Turkey. pp. 147–153.

FAHY, M.P. and GUCCIONE, M.J. 1979. Estimating strength of sandstone using petrographic thin-section data. Bulletin of the International Association of Engineering Geology, vol. 16, no. 4. pp. 467–485.

OZTURK, C.A., and NASUF, E. 2013. Strength classification of rock material based on textural properties. Tunnelling and Underground Space Technology, vol. 37. pp. 45–54.

The Journal of The Southern African Institute of Mining and Metallurgy

VOLUME 114

JUNE 2014

479

▲

ERSOY, A. and WALLER, M.D. 1995. Textural characterization of rocks. Engineering Geology, vol. 39. pp. 123–136.

Estimation of rock strength from quantitative assessment of rock texture OZTURK, C.A., NASUF, E., and BILGIN, N. 2004. The assessment off rock cutability, and physical and mechanical rock properties from a texture coefficient. Journal of the South African Institute of Mining and Metallurgy, vol. 7. pp. 397–403. PETRUK, W. 2000. Applied Mineralogy in the Mining Industry. Elsevier, Amsterdam. PHILLIPSON, S.E. 2008. Texture, mineralogy, and rock strength in horizontal stress-related coal mine roof falls. International Journal of Coal Geology, vol. 75. pp. 175-–84. PRICE, N.J. 1960. The compressive strength of coal measure rocks. Colliery Engineering, g vol. 37. pp. 283–292. PR˘IKRYL, R. 2006. Assessment of rock geomechanical quality by quantitative rock fabric coefficients: limitations and possible source of misinterpretations. Engineering Geology, vol. 87. pp. 149–162. PR˘IKRYL, R. 2001. Some microstructural aspects of strength variation in rocks. International Journal of Rock Mechanics and Mining Science, vol. 38. pp. 671–682. RICHARDS, N.P. and BELL, F.G. 1995. A survey of some petrographic and engineering properties of sandstones from the Natal Group, South Africa. Proceedings of the 8th International Congress on Rock Mechanics, vol. 3. pp. 1131–1136. SHAKOOR, A. and BONELLI, R.E. 1991. Relationship between petrographic characteristics, engineering index properties, and mechanical properties of

▲

480

JUNE 2014

VOLUME 114

selected sandstones. Bulletin of the International Association of Engineering Geology, vol. 28. pp. 55–71. SINGH, B. and GOEL, R.K. 2011. Engineering Rock Mass Classification, Tunnelling, Foundations, and Landslides. Elsevier, Amsterdam. SMORDINOV, M.I., MOTOVILOV, E.A., and VOLKOV, V.A. 1970. Determination of correlation relationships between strength and some physical characteristics of rocks. Proceeding of the 2nd International Congress of the Society on Rock Mechanics, vol. 2. pp. 35–37. TIRYAKI, B. and DIKMEN, A.C. 2006. Effects of rock properties on specific cutting energy in linear cutting of sandstones by picks. Rock Mechanics and Rock Engineering, g vol. 39, no. 2. pp. 89–120. TUGRUL, A. and ZARIF, I.H. 1999. Correlation of mineralogical and textural characteristics with engineering properties of selected granitic rocks from Turkey. Engineering Geology, vol. 51. pp. 303–317. VERHOEF, P.N.W. AND VAN DE WALL, A.R.G.1998. Application of petrography in durability assessment of rock construction materials. Aggregate Resources: a Global Perspective. Bobrowsky, P.T. (ed.). Balkema. Rotterdam. pp. 307–330. WADELL, H. 1932. Volume, shape and roundness of rock particles. Journal of Geology, vol. 40. pp. 443–451. WILLARD, R.J. and MCWILLIAMS, J.R. 1969. Microstructural techniques in the study of physical properties of rocks. International Journal of Rock Mechanics and Mining Science, vol. 6. pp. 1–2. WILLIAMS, H. 1982. Petrography. WH Freeman, San Francisco, CA.

◆

The Journal of The Southern African Institute of Mining and Metallurgy

A new process for the recovery of iron, vanadium, and titanium from vanadium titanomagnetite by S.Y. Chen* and M.S. Chu*

A new process comprising metallizing reduction, magnetic separation, and electroheat melting separation is proposed with the aim of achieving high recoveries of iron, vanadium, and titanium from vanadium titanomagnetite. The effects of magnetic intensity, reduction temperature, reduction time, carbon ratio, and coal particle size on the efficiency of reductionmagnetic separation were investigated, the reaction mechanisms elucidated by SEM and EDS analysis, and the optimal process parameters established. Recoveries of up to 80.08% for titanium, 95.07% for iron, and 71.60% for vanadium were achieved. The reduction sequence is shown to be: Fe2O3 → Fe3O4 → FeO → Fe; Fe2TiO5 → Fe2TiO4 + TiO2 → Fe + FeTiO3 → Fe + FeTi2O5 → Fe + TiO2. Keywords VTM, metallization reduction, magnetic separation, electrothermal smelting separation, phase transition.

Introduction Vanadium titanomagnetite (VTM), which contains valuable elements such as iron, vanadium, and titanium, has an extremely high potential value. VTM resources in the PanXi regions of China are estimated at up to 10 Gt (billion tons), and account for 93% and 63% of the country’s titanium and vanadium resources respectively. The exploitation of VTMs has thus received much attention (Barksdale, 1966; Chen et al., 2011). Owing to their poor grades, fine grain size, and complex mineralogy, VTMs are difficult to treat (Katsura and Kushiro, 1961; Akimoto and Katsura, 1959). In China, the main method of VTM utilization is in ironmaking via the blast furnace route and the basic oxygen furnace process for vanadium recovery. However, the recovery ratios of Fe (into pig iron), V (into vanadium slag), and Ti (into ilmenite) are only 70%, 42%, and 25% respectively, which are relatively low. Meanwhile, 70 Mt of high-titanium slag are discarded on the riversides of the Jinsha River, resulting in waste of resources and environmental pollution (Diao, 1999, Ma et al., 2000). Many new processes have been recently proposed and studied both in China and abroad (Roshchin et al., 2011; Wang et al., 2008; The Journal of The Southern African Institute of Mining and Metallurgy

Experimental methods Materials The VTM concentrate used in this investigation was obtained from the PanXI region in China. The chemical composition and size distribution of the VTM concentrate are shown in Table I. The VTM concentrate contained 53.91% total iron, 13.03% titanium oxide, and 0.52%. vanadium oxide. The X-ray diffraction (XRD) patterns of VTM concentrate are shown in Figure 2. The main phases are titanomagnetite (FexTi3-xO4), magnetite (Fe3O4), ilmenite (FeTiO3), vanadium spinel (FeO•V2O3), and augite [(Mg,Fe,Al,Ti)(Ca,Mg,Fe)(Si,Al)2O4].

* School of Materials and Metallurgy, Northeastern University, China. © The Southern African Institute of Mining and Metallurgy, 2014. ISSN 2225-6253. Paper received r Sep. 2013; revised paper received Nov. 2013. VOLUME 114

JUNE 2014

481

▲

Synopsis

Yuan et al., 2006; Lei et al., 2011), such as chemical extraction, direct reduction, and EAF smelting. However, none of the proposed processes have been shown to be commercially viable, owing to poor selectivity and recoveries, and economic efficiency. In this paper, a new process for VTM beneficiation is proposed, comprising metallizing reduction, magnetic separation, and electroheat melting separation. The proposed flow sheet is shown in Figure 1. The VTM is firstly reduced by carbothermal reduction, and the titanium (non-magnetic fraction) and iron and vanadium (magnetic fraction) are separated by magnetic separation. The magnetic fraction bearing iron and vanadium is melted and separated by electrothermal smelting separation. The hightitanium slag in the non-magnetic fraction and vanadium slag could be further treated by a hydrometallurgical route.

A new process for the recovery of iron, vanadium, and titanium

Figure 1—Flow chart of the new process

Table I

Chemical analysis of materials Chemical composition of VTM Fetot FeO V2O5 TiO2 CaO SiO2 MgO Al2O3 -

53.91% 31.13% 0.52% 13.03% 0.68% 3.20% 2.71% 3.82% -

Size distribution of VTM < 40 27.93% 40 ~ 74 18.95% 74 ~ 100 12.07% 100 ~ 200 28.37% Coal chemical properties Aad 4.29% Vdaf 33.64% Std 0.16% FC 62.12%

The magnetic ffraction was then treated by electrothermal smelting. A 135 g sample of the magnetic concentrate was dried for 3 hours at 105°C, then mixed with 0.9 g of coal, placed in a high-purity graphite crucible, and heated in a high-temperature resistance furnace at 1550°C for 120 minutes to separate a metallic iron phase and a vanadium slag. The chemical compositions of the metallization reduction product, the magnetic and non-magnetic fractions, and the iron phase and vanadium slag were analysed by X-ray diffraction XRD and scanning electron microscopy (SEM) to investigate the effects of process conditions and the reaction mechanisms.

Results and discussion Effect of magnetic intensity

Figure 2—XRD pattern of VTM concentrate

The finely ground metallization-reduction product was separated in the DTCXG-ZN50 magnetic tube at magnetic intensities from 25 mT to 125 mT. The process parameters for the metallization-reduction step were: reduction temperature 1350°C, reduction time 30 minutes, carbon ratio 1.0, coal particle size -74 μm. The chemical composition and metallization rate of the reduction product are listed in Table II, and the effects of magnetic intensity on the metallizing reduction and magnetic separation indexes are shown in Figure 4. The result shows that titanium partitions into the high-titanium slag phase of

Coal is used as reductant. The chemical analysis of the coal is listed in Table I, which shows that the coal is high in fixed carbon (FC), and low in ash (Aad) and total sulphur (Std).

Apparatus and procedure Figure 3 shows the schematic diagram of the experiment procedure, which consists of three main steps. The first step consists of metallization reduction. VTM concentrate (100 g) and coal were combined in mole ratios of fixed carbon to reducible oxygen (carbon ratio) of 0.8, 1.0, 1.2, 1.4, 1.6, and 1.8. The well-mixed sample was placed in a graphite crucible (Wang et al., 2008) and reduced in a hightemperature laboratory furnace at various temperatures and times. The crucible was removed from the furnace and allowed to cool. The reaction products were ground and a 10 g sample was placed in a DTCXG-ZN50 magnetic tube to separate the magnetic and non-magnetic fractions.

▲

482

JUNE 2014

VOLUME 114

Figure 3—Schematic diagram of experiment procedure and equipment

Table II

Chemical composition and metallization rate of reduction product Fetot MFe V2O5 TiO2 Metallization

65.59 % 61.84 % 0.728 % 16.5 % 94.28 %

The Journal of The Southern African Institute of Mining and Metallurgy

A new process for the recovery of iron, vanadium, and titanium equipment and the high recoveries off iron, vanadium, and titanium, an optimum reduction temperature of 1350°C was selected for the subsequent experiments. Figure 8 shows that the reducibility of the iron-bearing phase in the reduction product is significantly improved by increasing the reduction temperature. Aggregation and growth of the iron particles is also facilitated and much larger iron particles are formed. As shown in Figure 8(e)-(f), the iron phase (A) can be separated from the high-titanium slag phase (B) by magnetic separation.

Effects of reduction time

magnetic separation indexes

the non-magnetic fraction while iron and vanadium partition into the iron phase of the magnetic fraction. As shown in Figure 4, all of the process indexes, including titanium content (ω(TiO2)), recovery ratio of titanium (η(TiO2)) in the non-magnetic fraction, recovery ratio of vanadium (η(V2O5)), iron content (ω(TFe)), and recovery ratio of iron (η(Fe)) in the magnetic fraction, increase with increasing magnetic intensity from 25 mT to 50 mT, but show a downtrend trend as the magnetic intensity is increased beyond 50 mT. An optimal magnetic intensity of 50 mT was therefore selected for the subsequent experiments. The reaction mechanisms were investigated based on SEM and energy dispersive spectrometry (EDS) of the reduction product (Figure 5). Figure 5(a) shows that the dispersed iron particles agglomerate due to metallizing reduction and carburization. Figures 5(a) and 5(b) show that the high-titanium slag phase (gray phase) becomes entrained in the iron phase (white phase) on account of excessive magnetic force, leading to the decrease of η(Fe), η(V2O5), and η(TiO2).

To determine the optimum reduction time, a series of metallizing reduction and magnetic separation tests were carried out at reduction times of 10, 20, 30, 40, 50, and 60 minutes. The other process parameters were kept constant as follows: reduction temperature 1350°C, carbon ratio 1.0, coal particle size -74 μm, magnetic intensity 50 mT. The effects of reduction time on metallization and the iron content are

Figure 5—SEM and EDS images of the reduction product

Effect of reduction temperature

The Journal of The Southern African Institute of Mining and Metallurgy

Figure 6—Effects of temperature on the reduction product

Figure 7—Effects of reduction temperature on the new process VOLUME 114

JUNE 2014

483

▲

To determine the optimum reduction temperature, a series of metallizing reduction and magnetic separation tests were carried out at reduction temperature of 1250°C,1275°C, 1300°C, 1325°C, and 1350°C. The other process parameters were kept constant as follows: reduction time 30 minutes, carbon ratio 1.0, coal particle size -74 μm, magnetic intensity 50 mT. The effects of reduction temperature on metallization rate and the iron content are shown in Figure 6. The ω(TFe)) and metallization rate of the reduction product increase gradually with increasing reduction temperature. At a reduction temperature of 1350°C, the ω(TFe) and metallization rate are 65.59% and 94.28%, respectively. The effects of reduction temperature on the metallizing reduction and magnetic separation indexes are shown in Figure 7. The ω(TiO2) and η(TiO2) of the non-magnetic fraction increase from 26.43% to 54.73% and from 62.58% to 80.77% respectively. The ω(TFe), η(Fe), and η(V2O5) of the magnetic fraction also increase, from 78.94% to 84.55%, from 75.58% to 96.23%, and from 44.54% to 66.97% respectively with the reduction temperature increasing from 1250°C to 1350°C. Considering the heating capacity of the reduction

A new process for the recovery of iron, vanadium, and titanium

Figure 8—SEM image of reduction product at different reduction temperature

shown in Figure 9. The ω(TFe)) and metallization off the reduction product show a upward trend. At the maximum reduction time of 60 minutes, the ω(TFe) and metallization are 68.60% and 95.93%, respectively. The effects of reduction time on metallizing reduction and magnetic separation indexes are shown in Figure 10. With increasing reduction time from 10 minutes to 60 minutes, the (TiO2) and η(TiO2) of the non-magnetic fraction increase from 18.33% to 55.39% and from 79.81% to 80.08% respectively. The ω(TFe), η(Fe), and η(V2O5) of the magnetic fraction also increase, from 74.82% to 86.56%, from 39.74% to 97.48%, and from 29.22% to 75.68%. An optimum reduction time of 60 minutes was therefore selected for the subsequent experiments. The SEM micrographs of the six reduction products are shown in Figure 11. With increasing reduction time, the bearing-iron (white) phase aggregates and grows owing to the ongoing reduction reaction between VTM and coal. At the same time, iron-iron bonds form instead of iron-titanium bonds, accelerating the separation of the iron phase from the other phases.

Figure 9—Effects of reduction time on the reduction product

Effects of carbon ratio In test work to optimize the carbon ratio, the VTM and coal mixtures were prepared with carbon ratios of 0.8, 1.0, 1.2, 1.4, 1.6, and 1.8. The other process parameters were kept constant as follows: reduction temperature 1350°C, reduction time 60 minutes, coal particle size -74 μm, and magnetic

Figure 10—Effects of reduction time

Figure 11—SEM images of reduction product with different reduction time

▲

484

JUNE 2014

VOLUME 114

The Journal of The Southern African Institute of Mining and Metallurgy

A new process for the recovery of iron, vanadium, and titanium intensity 50 mT. The effects ff off carbon ratio on the metallization and iron content are shown in Figure 12. The metallization of the reduction product increases as the carbon ratio increases from 0.8, to 1.0, and then decreases as carbon ratios above 1.0. The effects of carbon ratio on the metallizing reduction and magnetic separation indexes are shown in Figure 13. With the carbon ratio increasing from 0.8 to 1.0, the ω(TiO2) and η(TiO2) of the non-magnetic fraction increase from 37.84% to 55.39% and from 76.07% to 80.08% respectively, and ω(TFe), η(Fe), and η(V2O5) of the magnetic fraction increase from 83.81 % to 86.56 %, from 80.68 % to 97.48 %, and from 64.00 % to 75.68 %, respectively. With further increases in the carbon ratio from 1.0 to 1.8, the ω(TiO2), ω(TFe), η(Fe), and η(V2O5) show a downward trend. Taking into account the economics of the new process and effective utilization of resources and energy, the an optimum carbon ratio of around 1.0 was selected for metallizing reduction and magnetic separation. The SEM micrographs of the reduction products obtained with different carbon ratios are shown in Figure 14. It can be seen that with an increase in the carbon ratio from 0.8 to 1.0, more iron particles (white) formed, aggregated, and grew, and a large number of bigger iron particles appeared to be separating the iron phase from the other phases. With further increases in the carbon ratio from 1.0, 1.4 to 1.8, the carbon is in excess, and the aggregation of reduced iron particles is hindered by superfluous graphitized carbon.

reduction and magnetic separation indexes are shown in Figure 16. Increasing in coal size from -74 μm to -2 mm had no significant effect on the η(TiO2) of the non-magnetic fraction. The ω(TFe), η(Fe), and η(V2O5) of the magnetic fraction gradually decreased from 85.56% to 78.13%, from 97.48% to 75.57% and from 75.68% to 56.82 %, respectively. Thus, the optimum coal particle size for metallizing reduction and magnetic separation should be -74 μm.

Figure 13—Effects of carbon ratio

Effects of coal particle size

Figure 12—Effects of carbon ratio on the reduction product The Journal of The Southern African Institute of Mining and Metallurgy

Figure 14—SEM and EDS photos of reduction product with different carbon ratios

Figure 15—Effects of coal size on the reduction product VOLUME 114

JUNE 2014

485

▲

To determine the effects of coal particle size on the metallizing reduction and magnetic separation indexes, a series of reduction and magnetic separation tests were performed at coal particle sizes of -74 μm, -0.5 mm, -1 mm, -1.5 mm, and -2 mm. The other parameters were kept constant as follows: reduction temperature 1350°C, reduction time 60 minutes, carbon ratio 1.0, and magnetic intensity 50 mT. The effects of coal particle size on metallization and the iron content are shown in Figure 15. The ω(TFe)) and metallization of the reduction product show a downward trend with increasing coal size. When the coal particle size increased from -74 μm to -2 mm, the ω(TFe) and metallization decreased significantly from 68.60% to 60.07%, and from 95.93% to 85.73%, respectively. The effects of coal particle size on the metallizing

A new process for the recovery of iron, vanadium, and titanium The main phases in the VTM concentrate are FexTi3-xO4, Fe3O4, FeTiO3, and augite. As reduction proceeds, Fe, FeO, and Fe5TiO8 appear (5 minutes); Fe2TiO5 appears and FeTiO3 disappears (10 minutes); Fe2TiO4 and FeTi2O5 appear but FeO disappears (25 minutes); Fe3O4 and Fe2TiO4 disappear and FeTiO3 appears (30 minutes); Fe2TiO5 disappears after 40 minutes, followed by FeTiO3 after 50 minutes. Further increase in the reduction time to 60 minutes does not result in any further change in the phase composition, and the VTM is considered to be reduced completely. Considering the analyses of the thermodynamics and phase transition process, the reduction processes of iron oxide and ilmenite in VTM proceed as follows: Figure 16—Effects of coal particle size index

Fe2O3 → Fe3O4 → FeO → Fe Fe2TiO5 → Fe2TiO4 + TiO2 → Fe + FeTiO3 → Fe + FeTi2O5 → Fe + TiO2

Results and analysis of electrothermal smelting experiment The chemical analyses of the magnetic (7.71 g) and nonmagnetic substance (2.21 g) fractions obtained by magnetic separation are shown in Table III. The chemical compositions of the iron and vanadium-bearing slag obtained by electrothermal smelting are analysed and listed in Table IV. The results show that 97.53% of the iron partitions to the iron phase and 94.61% of the vanadium to the vanadiumbearing slag. Considering together the results of the metallizing reduction, magnetic separation, and electrothermal smelting separation experiments, the η(TiO2) of the non-magnetic fraction is 80.08%, and the η(Fe) in the iron phase and η(V2O5) of the vanadium-bearing slag are 95.07% and 71.60%, respectively.

Thermodynamic analysis of the metallization reduction All possible chemical reactions between VTM and coal are depicted in Figure 17. The standard Gibbs free energies (ΔG°) of solid reactions and the gas-solid equilibrium of VTM reduced by CO are also shown. As depicted in Figure 17, the solid reactions take place at first, and are then restricted by solid diffusion and interface reactions. At the same time, the gas–solid reactions taking place are greatly accelerated by the increasing CO partial pressure. For the thermodynamic analysis, it can be seen that the ΔG° values for all solid-solid and gas-solid reactions are negative at 1350°C, therefore the reduction reaction of VTM can proceed and the reduction products of iron oxide and ilmenite are iron and titanium oxides of low valences (Sadykhov et al., 2010; Paunova, 2002).

Conclusion Based on the results obtained in this study, the following conclusions can be drawn.

Table III

Chemical composition of magnetic and nonmagnetic fraction % Constituent Fetot FeO TiO2 V2O5 CaO SiO2 MgO Al2O3

Magnetic 86.56 4.26 3.95 0.649 0.22 1.18 1.72 1.16

Non-magnetic 6.19 55.39 0.667 1.90 13.01 6.77 11.50

Table IV

Chemical composition of iron and slag phase Element

Iron phase

Slag phase

Mass Fetot V2O5 TiO2

115.45 g 98.72% 0.037% -

16.29 g 8.84% 5.39% 28.24%

Element migration process during metallization reduction In order to illustrate the element migration process based on the optimum conditions, including reduction temperature of 1350°C, carbon ratio of 1.0, and coal particle size of less than 74 μm, the XRD patterns and phase compositions of the reduction product at different stages of reduction are given in Figure 18 and Table V (Pogudin et al., 2010; Welham, 1996).

▲

486

JUNE 2014

VOLUME 114

Figure 17—The results of thermodynamic analysis (after (Pesl and Eric, 2002; Huang, 2007) The Journal of The Southern African Institute of Mining and Metallurgy

A new process for the recovery of iron, vanadium, and titanium

Figure 18—X-ray diffraction analysis of phase containing iron and titanium

Phase composition of reduction product with different reduction times Time, min

Phase composition

Fe3O4, FexTi3-xO4, FeV2O4, FeTiO3, augite

Fe, FeO, Fe3O4, Fe5TiO8, FeTiO3, MgAl2O4

Fe, FeO, Fe3O4, Fe2TiO5, Fe5TiO8, MgAl2O4, MgSi2O4

Fe, FeO, Fe3O4, Fe2TiO5, MgAl2O4, MgSi2O4

Fe, Fe3O4, Fe2TiO5, Fe2TiO4, MgAl2O4, MgSi2O4

Fe, Fe2TiO5, FeTiO3, (Fe,Mg)Ti2O5, MgAl2O4, MgSi2O4

Fe, FeTiO3, MgAl2O4, (Fe,Mg)Ti2O5, MgSi2O4

Fe, (Fe,Mg)Ti2O5, MgAl2O4, MgSi2O4

The optimum process conditions for metallizing reduction and magnetic separation are a carbon ratio of 1.0, reduction temperature 1350°C, reduction time 60 minutes, magnetic intensity 50 mT, and coal particle size less than 74 μm. Under the above conditions, after electroheat melting separation of the magnetic fraction, the recovery ratio of titanium in the non-magnetic fraction is 80.08%, and recovery ratios of iron in molten iron and vanadium in molten slag are 95.07% and 71.60%, respectively. Through the new process of metallizing reduction, magnetic separation, and electroheat melting separation, the effective separating of iron, vanadium and titanium in VTM could be achieved and iron in iron phase, vanadium in vanadium slag, and titanium in high titanium slag were successfully obtained. Based on the analyses of the thermodynamics and phase transitions, the reduction of iron oxide and ilmenite in VTM proceeds as follows:

Fe2O3 → Fe3O4 → FeO → Fe Fe2TiO5 → Fe2TiO4 + TiO2 → Fe+FeTiO3 → Fe + FeTi2O5 → Fe + TiO2 Acknowledgements The research presented this paper was supported by National High-tech Research and Development Projects (Grant No. 2012AA062302) and Major Program of the National Natural Science Foundation of China (Grant No. 51090384). The The Journal of The Southern African Institute of Mining and Metallurgy

work was also supported by the Fundamental Research Funds for the Central Universities (Grant No. N110202001).

References AKIMOTO, S. and KATSURA, T. 1959. Magnetochemical study of the generalized titanomagnetic in volcanic rocks. Journal of Geomagnetism and Geoelectricity, vol. 10. pp. 69–90. BARKSDALE, J. 1966. Titanium, its Chemistry and Technology. 2nd edn, Roland Press, New York. CHEN, D.S., SONG, B., and WANG, L.N. 2011. Solid state reduction of Panzhihua titanomagnetite concentrates with pulverized coal. Minerals Engineering, g vol. 24, no. 8. pp. 864–869. DIAO, R.S. 1999. New understanding about special problems of smelting vanadium-bearing titanomagnetite with BF. Iron and Steel, vol. 34, no. 6. pp. 12–14. HUANG, X.H. 2007. Ferrous Metallurgy Theory. Metallurgical Industry Press, Beijing. pp. 268–277. KATSURA, T. and KUSHIRO, I. 1961. Titanomagnetic in igneous rocks. American Mineralogist, t vol. 46. pp. 134–145. LEI, Y., LI, Y., and PENG, J.H. 2011. Carbothermic reduction of Panzhihua oxidized ilmenite in a microwave field. ISIJ International, vol. 51, no. 3. pp. 337−343. MA, J.Y., SUN, X.W., and SHENG, S.X. 2000. Intensified smelting of vanadium and titanium magnetite in blast furnace. Iron and Steel, vol. 35, no. 1. pp. 4–8. PAUNOVA, R. 2002. Thermodynamic study of the reduction of titanium magnetite concentrate with solid carbon. Metallurgical and Materials Transactions B, vol. 33, no. 8. pp. 633−638. PESL, J. and ERIC, R.H. 2002. High temperature carbothermic reduction of Fe2O3-TiO2-MxOy oxide mixtures. Minerals Engineering, g vol. 15. pp. 971−984. POGUDIN, D.S., MOROZOV, A.A., and SADYKHOV, G.B. 2010. Phase formation during the reduction of titanomagnetite from the Gremyakha-Vyrmes deposit. Russian Metallurgy, vol. 9. pp. 759–762. ROSHCHIN, V.E., ASANOV, A.V., and ROSHCHIN A.V. 2011. Possibilities of twostage processing of titaniferous magnetite ore concentrates. Russian Metallurgy, vol. 6. pp. 15–25. SADYKHOV, G.B., GONCHAROV, K.V., and GONCHARENKO, T.V. 2010. Phase composition of the vanadium-containing titanium slags forming upon the reduction smelting of the titanomagnetite concentrate from the Kuranakhsk deposit. Russian Metallurgy, vol. 7. pp. 581−587. WANG, Y.M., YUAN, Z F., and GUO, Z.C. 2008. Reduction mechanism of natural ilmenite with graphite. Transactions of the Nonferrous Metals Society, vol. 18, no. 4. pp. 962−968. WELHAM, N.J. 1996. A parametric study of the mechanically activated carbothermic reduction of ilmentite. Minerals Engineering, g vol. 9. pp. 1189−1200. YUAN, Z.F., WANG, X.Q., and XU, C. 2006. A new process for comprehensive utilization of complex titania ore. Minerals Engineering, g vol. 19, no. 11. pp. 975−978. ◆ VOLUME 114

JUNE 2014

487

▲

Table V

1THREAD_4139_SAI

Geared Motors \ Drive Electronics \ Drive Automation \ Industrial Gears \ Services

We drive the mining industry

SEW-EURODRIVE offers innovative drive solutions for all applications in the mining industries. All SEW-EURODRIVE products and systems make the best use of the space available around the machine and ensure great ďŹ&#x201A;exibility and reliability. Minimum maintenance and simple operation ensure that you will operate machines and equipment efďŹ ciently from the very beginning. Thanks to the modular design and countless combination options, all drive engineering components can be replaced quickly, if the need arises. From gravel mining to the excavation of gold, platinum, coal and diamonds - we put the drive into all facets of the mining industry. SEW-EURODRIVE - Driving the world.

Tel: +27 11 248-7000 Web: www.sew.co.za

A methodology for determining the erosion profile of the freeze lining in submerged arc furnace by H. Dong*, H.-J. Wang*, and S.-J. Chu*

This paper presents a heat conduction model to determine the erosion profile of freeze linings in a ferronickel submerged arc furnace. Kirchhoff transformation was adopted to simplify the nonlinear equation governing heat conduction. The boundary element method (BEM) and Nelder-Mead simplex optimization method were employed to solve an inverse heat conduction problem. Erosion profiles could be obtained when calculated temperatures at sensor locations agreed well with measured temperatures. The model was validated by comparing the heat loss through the linings, which was calculated by BEM and integral operation after erosion profiles were obtained, with the value calculated from industrial data such as actual power consumption, tap-to-tap time, and ferronickel output. The optimal solution for the erosion profile could be acquired when the difference in heat losses calculated by these two methods is at a minimum. Keywords submerged arc furnace, freeze lining, erosion profile, boundary element method.

Introduction The submerged arc furnace (SAF) is still the predominant route for ferroalloy production, accounting for over 70% of production. SAF campaigns are strongly influenced by the lifetime of the lining, which is gradually eroded or worn due to physical and chemical wear by molten alloy and slag. The design of a SAF is a lengthy process, and construction is expensive. In order to prevent breakouts, it is important to monitor the residual thickness of linings. To estimate the erosion profile of SAF linings, a few online monitoring models (Kievit et al., 2004; Rodd, Voermann, and Stober, 2010; Karstein and Skaar, 1999) have been developed since 1999, which are similar to that applied in monitoring the erosion of the hearth and bottom of a blast furnace (BF). The methodologies by which information on a BF hearth state is obtained include core sampling (Akihiko, 2003), non-destructive testing (NDT) (Afshin and Pawel, 2009), and theoretical prediction (Zhao et al., 2007; Swartling et al., 2010; Surendra, 2005; Brännbacka and Saxén, 2008; Zagaria, Dimastromatteo, and Colla, 2010; Torrkulla The Journal of The Southern African Institute of Mining and Metallurgy

* School of Metallurgical and Ecological Engineering, University of Science and Technology Beijing, Beijing, China. © The Southern African Institute of Mining and Metallurgy, 2014. ISSN 2225-6253. Paper received Mar. 2013; revised paper received Feb. 2014. VOLUME 114

JUNE 2014

489

▲

Synopsis

and Saxén, 2000; Lijun and Huier, 2003; Yu and Robit, 2008; Kuncan, Zhi, and Xunliang, 2009; Kouji, Takanobu, and Kouzo, 2001). Thicknesses of linings, provided by core sampling and NDT methods are precise, but these methods have obvious drawbacks. Core sampling requires a furnace shutdown, with consequent loss of production. The thickness of lining acquired by NDT has a local character and cannot represent the status of the entire BF hearth. Theoretical prediction, on the other hand, calculates the location of 1150°C isotherm based on measured temperatures from thermocouples embedded in hearth linings. This isotherm was considered as the wear-line in the BF hearth model. Theoretical prediction models involve a heat conduction model and a composite model that combines heat conduction theory with CFD (Kuncan, Zhi, and Xunliang, 2009). With respect to the heat conduction model, there are two ways to approach a heat transfer problem: direct and inverse heat conduction problems, known as DHCP and IHCP (Swartling et al., 2010). Boundary conditions and thermo-physical properties are known in the DHCP models (Swartling et al., 2010; Surendra, 2005; Lijun and Huier, 2003). A significant limitation of DHCP models is that the computational domain is not immutable. Hence it is necessary to continually change the computing domain, which varies with the location of the erosion boundary. However, it is difficult to accomplish this by the DHCP approach. For instance, Lijun and Huier (2003) attempted to employ BEM to solve a DHCP and the 1150°C isotherm was investigated by the orthogonal design. Either thermo-physical properties or boundary conditions could be known in the IHCP model; instead, interior temperatures had

A methodology for determining the erosion profile of the freeze lining to be known ffor some points off the domain (Swartling et al., 2010). Monitoring models to solve IHCP and estimate the shape of erosion profile were developed, the system geometry of which is not fully known but is being investigated in the solution. Some investigators (Brännbacka and Saxén, 2008; Zagaria, Dimastromatteo, and Colla, 2010; Torrkulla and Saxén, 2000) combined the optimization technique with a numerical method to solve IHCP, obtaining several acceptable solutions for the erosion profile. It is important to note that the approximate solutions acquired from IHCP models are not unique. Therefore, validation is very important in numerical simulation because the availability of a computing model should be carefully considered. The most common validating method is the comparison between calculated temperatures and measured temperatures. If calculated temperatures at thermocouple locations approximate measured ones, the solution for the erosion profile can be accepted. Due to the non-uniqueness of solutions in IHCP, some other measurements were employed to validate the computation models in addition to comparison of temperatures. Dig-out investigations (Zagaria, Dimastromatteo, and Colla, 2010; Kouji, Takanobu, and Kouzo, 2001) are also a good validation method, but this process is available only during the blow-out of a BF. Torrkulla and Saxén (2000) validated the model by analysing various hearth phenomena such as pressure drop, coke voidage, production rate, and slag delay. In addition, Yu and Robit (2008) combined the CFD model with a one-dimensional heat conduction to perform a numerical analysis on the inner profile of a BF hearth. However, this calculation procedure is overly complex. Many researchers have contributed to modelling the complex process of lining erosion in a BF hearth, while only few have paid attention to the erosion profile of SAFs, and monitoring of SAF linings is still in its infancy. Kievit et al. (2004) utilized a one-dimensional steady heat conduction method to monitor a ferromanganese SAF. Rodd, Voermann, and Stober (2010) developed a two-dimensional DHCP model that was used for a SAF for ferronickel production. Karstein and Skaar (1999) exploited two different numerical algorithms and employed the finite element method (FEM) to solve IHCP for an ilmenite melting furnace. The simulated results obtained from this work were validated by the method of temperature comparison. In view of the lack of methods for determining the erosion profile of the freeze lining in SAFs and the shortcomings of the method of temperature comparison, we have developed a model for monitoring the erosion profile in a ferronickel SAF at the Liang-Da cooperation in Shandong Province, China. Measured temperatures from the thermocouples were utilized to solve an inverse heat conduction problem. The nonlinear differential equation could be converted into a Laplace equation after introducing Kirchhoff transformation. BEM and the Nelder-Mead simplex optimization method (Nelder and Mead. 1965) were adopted simultaneously in order to improve the speed of solution. Furthermore, the computing model was validated with industrial data such as actual power consumption, tap-to-tap time, and output of ferronickel. Compared with previous monitoring models for SAFs, there are two main advantages in the methodology presented in this paper. Improved accuracy and efficiency are obtained

▲

490

JUNE 2014

VOLUME 114

by employing BEM and the Nelder-Mead simplex algorithm, and a new validation method is presented, in which the model is validated utilizing industrial data.

Mathematical model Governing equations and boundary conditions The SAF lining model studied in this paper is built with reference to the configuration illustrated in Figure 1. According to this concept, the ceramic protection layer is a sacrificial layer, and a ’skull’ will be formed at the interface between carbon brick and molten iron after the ceramic layer has been eroded away. The erosion profile of the freeze lining could therefore be determined if the inner contour of the carbon brick region could be calculated. Figure 2 illustrates a calculation domain Ω to represent the carbon brick region, the inner boundary of which is expressed as B5, and the determination of its shape is the aim of the present work. Sensor locations in the system geometry Ω are indicated by the open circles in Figure 2. Six thermocouples numbered from 1–6 are utilized to measure internal temperatures. A further six thermocouples, numbered 7–12, are embedded at the interface between carbon brick and graphite to measure the boundary temperature. It should be pointed out that the actual number of sensors in a SAF is larger. In region Ω, heat conduction in the steady state can be described by the 2D rotational symmetry equation:

Shell Graphite Hot-pressed Carbon brick Ceramic protection layer

Figure 1—Freeze lining configuration in a submerged arc furnace (after Duncanson and Toth, 2004)

Figure 2—2D physical model for computing erosion profile The Journal of The Southern African Institute of Mining and Metallurgy

A methodology for determining the erosion profile of the freeze lining [1] where T is the temperature at a point (r, r z), r and z are the radial and axial coordinates, respectively, and k(T ) is the thermal conductivity of the carbon brick, which is a function of temperature as shown in Equation [2] and can be detected by micro-laser flash apparatus (LFA427, NETZSH group, Germany) . [2] Equation [1] is a nonlinear partial differential equation because the function of thermal conductivity is a logarithmic equation. Kirchhoff transformation (Carslaw and Jaeger, 1959) is applied to simplify the differential equation of heat conduction. A new temperature variation U is used to replace T (Equation [3]), where k0 is thermal conductivity of carbon brick at a known temperature T0. The Laplace equation can be obtained after substituting U into Equation [1] as follows. [3]

[4] In order to solve Equation [4], boundary conditions should be specified and transferred by Kirchhoff transformation for the domain Ω border. Heat fluxes of B1 and B4 are assumed to be zero since the model is rotational symmetric.

Numerical and optimization method for inverse heat conduction problem Normally, the solution for IHCP consisted of two main parts. First, a numerical method was utilized to calculate the temperatures at the sensor locations, and an optimization method was then adopted to minimize the difference between the calculated temperature (T TC) and measured temperature (T TM) (Equation [11]). The unknown erosion profile will be estimated by this computation scheme, illustrated as Equation [11]: [11] where Tc is the temperature calculated by the numerical methods, TM is the measured temperature from the thermocouples, and superscript i is the number of thermocouples in domain Ω. The Laplace equation describing the domain Ω in Figure 2 could be solved by BEM with high efficiency. Temperatures at the thermocouple locations could be calculated from Equation [12]. The integrating range C consists of the entire boundary T* represents the fundamental (B1∪B2∪B3∪B4∪B4∪). T solution of the Laplace equation and q* is its derivative (Kurpisz and Nowak, 1995). [12]

[5]

[6] Temperatures of the boundary between B2 and B3 were measured by thermocouples 7-9 and 10-12 respectively. A function f(r, r zz) was selected to describe the Dirichlet boundary condition of B2 and B3. A new function g(r, r zz) of boundary conditions could also be established from the Kirchhoff transformation. [7] [8]

The inside boundary B5, as the optimizing object, should be determined by some parameters at first. The parameterization schedule to B5 is shown in Figure 3 and described as follows. Two straight lines were extended from the B1 and B4 boundary, which intersect at structure point A. Structure lines (L1, L2, L3, ... Ln) were built after connecting point A with some points on B2 and B3. B5 would be formed by cubic spline interpolation to boundary points (P1, P2, P3, ... Pn) that lie on the structure lines. Because the relationship between r and z is linear on the structure lines, a set of radial coordinates (r1, r2, r3, ... rn) of boundary points is defined as optimizing parameters. The ill-posed property is a major obstacle in the solution process. Small changes in measured data would lead to significant fluctuations. In order to prevent the emergence of unreliable solutions, a regularization method was adopted to modify the equation (Brännbacka and Saxén, 2008; Zagaria, Dimastromatteo, and

The 1150°C isotherm is defined as the inner boundary when the erosion profile of the BF hearth is monitored. This temperature corresponds to the solidification of the eutectic Fe-C, and the hot metal is in solid state at lower temperatures than this (Kuncan, Zhi, and Xunliang, 2009). In industrial ferronickel production, the FeNi liquidus temperature is around 1300°C and the slag liquidus temperature is 1550°C. Since the amount of slag is very large in ferronickel production, the slag liquidus temperature 1550°C is defined as the hot face boundary condition (Rodd, Voermann, and Stober, 2010. [9]

The Journal of The Southern African Institute of Mining and Metallurgy

Figure 3—Construction scheme of inner boundary in computational model VOLUME 114

JUNE 2014

491

▲

[10]

A methodology for determining the erosion profile of the freeze lining Colla, 2010; Torrkulla and Saxén, 2000). Therefore, f instead of the minimizing formula in Equation [11], the minimization process was modified with a regularizer, as shown in Equation [13]. [13]

are shown in Figure 5. It can be seen that the calculated shapes of the inner profile vary widely, although the difference between two calculated temperatures is not obvious. In this case, it is preferable to introduce some other characteristics or industrial data for SAFs to validate the computing model and determine a reasonable solution.

Validation where Ψ is the objective function, j is the number of boundary points, αj is the angle between structure line and B5, and γ is a regularization term. The Nelder-Mead simplex optimization method (Nelder and Mead. 1965) belonging to a direct searching procedure is introduced. The criterion for stopping iteration of the optimization process is shown in Equation [14]:

In previous work, Chu, Liu, and Wang (2009) assumed that an SAF is an adiabatic system and calculated its theoretical power consumption by material and energy balance theory. The results showed that the theoretical power consumption is 2713.46 kWh for producing one ton of FeNi (12%Ni) when the charge has been heated to 700°C. However, a real SAF is not an adiabatic system, hence it has a higher power

[14] where ε is a convergence criterion, R l is the element from the collection of ‘simplex’, and xc is the centroid (Nelder and Mead, 1965). The flow diagram of the computational model is shown in Figure 4. Since Equation [12] does not contain any domain integrals, only the boundary needs to be discretized. In addition, the solution of the Nelder-Mead simplex method (modified simplex method) does not need derivation to optimizing parameters, and this program would converge in a few minutes. Detailed information about BEM and the modified simplex method is given by Nelder and Mead (1965) and Kurpisz and Nowak (1995). The setting up of the mathematical model is described by Dong and Shaojun (2013).

Results The average values of measured data from thermocouples 1–6 in Figure 2 were selected to calculate the erosion profile. Measured temperatures in domain Ω also needed to be transferred by the Kirchhoff method (Table I). The computing methodology in Figure 4 was programmed in MATLAB (R2010a). The convergence criterion was chosen as 25 and the regularization term was defined as 0.015. More than one result could be acquired after the iterative convergence because of the diversity of the initial simplex and illposedness of the inverse problem. Two sets of calculated temperatures are illustrated in Table I. It can be seen that the difference between measured temperatures and calculated temperatures is less than 5°C, which indicates the effectiveness of the computing methodology developed. The corresponding erosion profiles from calculated temperature

Figure 4—Flow chart of the computational model

Figure 5—Computed lining erosion profiles obtained by (a) calculated temperature 1 and (b) calculated temperature 2 (Table I)

Table I

Measured and calculated temperatures at sensor locations Temperature (°C)

Measured Measured (Kirchhoff transformation) Calculated 1 Calculated 2

▲

492

JUNE 2014

VOLUME 114

Point 1

Point 2

Point 3

Point 4

Point 5

Point 6

733.58 676.83 677.39 679.41

677.39 623.56 625.90 621.67

333.46 295.44 295.46 298.78

638.87 586.99 587.08 589.19

755.11 697.23 696.27 694.05

773.29 714.44 715.74 713.93

The Journal of The Southern African Institute of Mining and Metallurgy

A methodology for determining the erosion profile of the freeze lining consumption. According to investigations in an industrial plant, the actual power consumption is 3650 kW/t. A batch of FeNi is about 35 t and tap-to-tap time is around 4 hours. The modes of heat loss from a SAF have also been studied, and the heat losses through different paths are shown in Table II. It is assumed that all these energies, which are generated by extra power consumption, are changed into heat. If the heat emissions from a SAF as indicated in Table II are correct, heat loss from the current SAF linings per unit time can be calculated as Equation [15]: [15]

[18] Six vertical sections were selected to arrange the thermocouples in the SAF, as shown in Table III. The positions of the sensors in each vertical sections were similar. Table III also shows three different calculated temperatures and the corresponding heat fluxes of 3D models per unit time, assuming three possible results for the erosion profile could be obtained in each section. Total heat loss of the lining per unit time could be calculated as follows: [19]

1 QR

is the heat loss through the linings as calculated where from the production data. Since the heat flux density of all boundaries in domain Ω can be calculated by virtue of BEM after erosion profiles have been obtained, the heat loss from the linings can also be calculated by integration. The computing process can be expressed as follows: since more than one vertical section, similar to domain Ω in Figure 2, was fitted with thermocouples to measure temperatures in a production SAF, several 3D solid lining models can be built by integrating to vertical sections; then the total heat loss of the linings can be calculated by summing the heat losses of all 3D models. The computational scheme, which takes one vertical section of domain Ω as an example, is illustrated in Figure 6. B4 in Figure 6 is assumed to be an axis located in the center of the SAF. One 3D model s generated by rotating the section around B4 through an angle φ. Surfaces S1 and S2 are formed by the sweep of B2 and B3, with lengths of H and G, respectively. In addition, B2 and B3 in Figure 6 are discretized into some boundary elements, with the amounts of M1 and M2. The element on boundary B2, numbered as u, can be u indicated by EB2 . Similarly, the v vth element on boundary B3 v is expressed as EB3 . Two surfaces can be formed by the u u v sweep of elements EB2 . and EB3 , with heat fluxes of QB2 and v QB3 respectively, calculated by Equations [16] and [17].

Where superscript a is the number of the vertical section, and subscript b is the possible number of calculated temperatures in vertical section a. Since three possible results for the erosion profile are obtained in each vertical section, the 2 1 amount of possible QR is equal to (C (C3 )6. All computed results 2 1 for QR need to be compared with QR. The optimal solution of 2 1 QR is the value closest to QR. 2 An optimal calculated result for QR is acquired after 729 comparisons, as illustrated in Equation [20]. The difference 2 1 between optimal QR and QR is 168.2 KJ/s. [20] 2

In the process of calculating QR , the SAF lining is supposed to be an ideal 3D solid that is similar to a cylinder, with tap-holes ignored. It should be noted that the temperature at the region near the tap-holes is higher. In the current model, the computed results for heat loss, ignoring the tap-holes, are lower than for a realistic situation. Erosion profiles obtained after validation could meet the requirements of temperature comparison method; furthermore, the calculated heat loss obtained by the current model approximates that computed with industrial data.

Conclusions [16]

In order to monitor the erosion profile of the freeze lining in a SAF, a new two-step mathematical model was developed.

[17] where h is the length of an element and q the heat flux o density. It should be noted that the special parameter hB3 in Equation [17] is equal to zero. The heat loss of the 3D model, denoted by Γ, is the sum of heat flux through S1 and S2, which can be calculated by Equation [18].

Table II

Modes of heat losses from a SAF

Off-gases Resistance heat generated in circuit Linings (bottom and sidewall) Other

% of total heat loss 31.74 12.80 25.17 30.29

The Journal of The Southern African Institute of Mining and Metallurgy

Figure 6—Computing model for heat loss with BEM and integral VOLUME 114

JUNE 2014

493

▲

Mode

A methodology for determining the erosion profile of the freeze lining Table III

Measured and calculated temperatures at different vertical sections Heat release Γ

Temperature (°C) Point 1

Point 2

Point 3

Point 4

Point 5

Point 6

(KJ/s)

Section 1 Calculated 1 Calculated 2 Calculated 3

Measured 735.42 736.44 736.81

733.58 679.82 676.07 683.25

677.39 331.76 329.63 332.83

333.46 641.79 640.27 647.27

638.87 757.94 754.71 759.71

755.11 775.18 771.50 777.43

773.29 329.21 322.98 333.52

unknown

Section 2 Calculated 1 Calculated 2 Calculated 3

Measured 759.45 758.28 753.14

756.35 692.33 694.49 688.21

690.25 354.01 357.27 353.72

356.78 659.41 658.85 652.20

655.65 772.34 777.52 769.45

774.97 792.22 794.44 787.50

791.74 362.14 359.67 355.87

unknown

Section 3 Calculated 1 Calculated 2 Calculated 3

Measured 721.54 726.85 728.26

725.41 656.85 660.15 662.41

658.17 323.02 326.78 327.11

325.97 624.58 630.96 628.30

626.22 739.95 745.33 744.02

742.01 766.01 765.11 763.27

761.19 308.35 317.57 315.52

unknown

Section 4 Calculated 1 Calculated 2 Calculated 3

Measured 695.97 701.21 693.33

697.12 638.55 644.98 637.91

641.95 304.95 306.60 303.10

306.20 605.92 610.53 602.01

606.08 728.95 730.97 724.08

727.72 742.08 747.52 739.58

744.74 287.08 290.93 286.65

unknown

Section 5 Calculated 1 Calculated 2 Calculated 3

Measured 716.29 713.32 708.11

712.87 660.31 657.03 649.73

654.01 313.43 315.08 310.14

311.36 615.47 614.01 607.61

613.24 740.15 739.92 735.39

737.47 759.54 758.66 751.53

756.09 296.96 298.05 292.49

unknown

Section 6 Calculated 1 Calculated 2 Calculated 3

Measured 703.21 706.42 708.98

704.52 648.57 652.65 653.02

650.25 304.10 306.85 305.91

304.25 605.44 605.70 606.63

604.23 725.14 728.88 726.20

724.51 741.14 743.99 745.28

742.24 289.87 292.20 290.74

unknown

Kirchhoff transformation was applied in the first step to deal with the nonlinear partial differential equation, boundary conditions, and measured temperatures. The boundary element method was combined with the Nelder-Mead simplex optimization method in the second step to solve an inverse heat conduction problem using the measured data from thermocouples. The difference between calculated temperature and measured temperature is less than 5°C after the erosion profile is obtained. The solutions are not unique owing to the differences of initial simplex and ill-posedness of inverse problems. The computational model was validated by comparison of heat losses from the lining. The heat release through the linings, which was calculated by BEM and integral calculus after the erosion profiles were obtained, could be computed using industrial data such as actual power consumption, tap-to-tap time, and output of ferronickel. The results showed that the difference in lining heat losses per unit time obtained by these two methods is around 168.2 KJ/s. The possible reason for this is that the tap-hole region is ignored in the model. A transient model and 3D lining model including the tap-hole will be developed in future work.

The authors gratefully acknowledge the financial support from National Natural Science Foundation of China (No.51274030) for this project. Our thanks to Mr Jichao Li from Ohio State University for useful discussions, and to Mr. Peixiao Liu from Sinosteel Jilin Electro-mechnical Equipment Co. Ltd providing valuable industrial data for the SAF.

▲

JUNE 2014

v EB3

Boundary element that is numbered as v on boundary B2, {v: v∈ (1,2...,M2)}

TMi

Measured temperature of no. i thermocouple, {i: i∈ (1,2,3,4,5,6)}

u hB2

Length of element EB2

TCi

Calculated temperature at the location of no. i thermocouple, {i: i∈ (1,2,3,4,5,6)}

v hB3

Length of element EB3

Temperature transferred by Kirchhoff transformation

u qB2

Heat flux density of element EB2

Radial coordinates

v qB3

Heat flux density of element EB3

Axial coordinates

u QB2

Heat flux of the surface that is formed by the sweep u of EB2 at an angle of φ

thermal conductivity

Acknowledgements

494

Nomenclature

VOLUME 114

QB3

Heat flux of the surface that is formed by the sweep v of EB3 at an angle of φ

Boundary of domain Ω, {o: oa∈ (1,2,3,4,5)}

Number of lining vertical section, {a: a∈ (1,2,3,4,5,6)}

Temperatures function of boundary B2 and B3 The Journal of The Southern African Institute of Mining and Metallurgy

A methodology for determining the erosion profile of the freeze lining b g

Number off calculated temperature in vertical section, {b: b∈ (1,2,3)} Temperatures function transferred from f by Kirchhoff transformation

Amount of structure lines

Structure point

Amount of element on boundary B2

Structure line, {j: j ∈ (1,2,...n)}

Amount of element on boundary B3

Boundary point of B5, {j: j ∈ (1,2,...n)}

CHU, S.J., LIU, P.X, and WANG, H.L. 2009. Calculation off theoretical power consumption and production cost of nickel ferroalloy. Ferro-Alloys, no. 4. pp. 1–7. DONG, H. and SHAOJUN, C. 2013. Mathematical model for fast comptation of erosion profile in submerged arc furnace with freeze lining. 13th International Ferroalloys Congress Proceedings, Almaty, Kazakhstan. pp. 799–809. DUNCANSON, P.L. and TOTH, J.D. 2004. The truths and myths of freeze lining technology for submerged arc furnace. 10th International Ferroalloys Congress, Cape Town. pp. 488–499. KARSTEIN, S.I. and SKAAR, M. 1999. Monitoring the wear-line of a melting furnace. 3rd International Conference on Inverse Problems in Engineering, g

Greek symbols

Port Ludlow. pp. 1-11.

Element from the collection of regular ‘simplex’ in modified simplex method, {l: l∈ (1,2,...n+1)}

αj

Angles between B5 and structure line, {j: j ∈ (1,2,...n)}

Centriod in modified simplex method

Convergence criteria

erosion process of blast furnace hearth. ISIJ International, vol. 41, no. 10.

Fundamental solution in BEM

pp. 1139–1145.

Regularization term * TAS

control of furnace 1 freeze lining at Tasmanian electro-metallurgical company. 10th International Ferroalloys Congress, Cape Town. pp. 477–483. KOUJI, T., TAKANOBU, I., and KOUZO, T. 2001. Mathematical model for transient

KUNCAN, Z., ZHI, W., and XUNLIANG L. 2009. Research status and development trend of numerical simulation on blast furnace lining erosion. ISIJ

Derivative of

Computing region

Integrating range

Objective function

Rotating angle of vertical section

Length of B2 boundary

element method. Applied Thermal Engineering, g vol. 23, no. 16.

Length of B3 boundary

pp. 2079–2087.

1 QR a Γb

in BEM

KIEVIT, A.D., GANGULY, S., DENNIS, P., and PEIETRS, T. 2004. Monitoring and

Heat release calculated with industrial data Heat release of 3D model formed by rotation of vertical section at an angle φ

QR2

Heat release calculated by presented model

EBu2

Boundary element that is numbered as u on boundary B2, {u: u∈ (1,2,...,M1)}

International, vol. 49, no. 9. pp. 1277–1282. KURPISZ, K. and NOWAK, A.J. 1995. Inverse thermal problems. Computational Mechanics Publications. pp. 53–65. LIJUN, W. and HUIER, C. 2003. Mathematical model for on-line prediction of bottom and hearth of blast furnace by particular solution boundary

NELDER, J.A. and MEAD, R. 1965. A simplex method for function minimization. The Computer Journal, vol. 7, no. 4. pp. 308–313. RODD, L., VOERMANN, N., and STOBER, F. 2010. SNNC: a new ferro-nickel smelter in Korea. 12th International Ferroalloys Congress, Helsinki. pp. 698–700. SURENDRA, K. 2005. Heat transfer analysis and estimation of refractory wear in an iron blast furnace hearth using finite element method. ISIJ International, vol. 45, no. 8. pp. 1123–1128.

References AFSHIN, S. and PAWEL, G. 2009. Non-destructive testing (NDT) and inspection of the blast furnace refractory lining by stress wave propagation technique. 5th International Congress on the Science and Technology of Ironmaking, g Shanghai. pp. 951–955. AKIHIKO, S., HITOSHI, N., NARIYUKI, Y., YOSHIFUMI, M., and MASARU, M. 2003.

SWARTLING, M., SUNDELIN, B., AND TILLIANDER, A, and JÖNSSON, P.G. 2010. Heat transfer modelling of a blast furnace hearth. Steel Research International, vol. 81, no. 3. pp. 186–196. TORRKULLA, J. and SAXÉN, H. 2000. Model of the state of the blast furnace hearth. ISIJ International, vol. 40, no. 5. pp. 438–447. YU, Z. and ROBIT, D. 2008. Numerical analysis of blast furnace hearth inner

Investigation of blast-furnace hearth sidewall erosion by core sample

profile by using CFD and heat transfer model for different time periods.

analysis and consideration of campaign operation. ISIJ International,

International Journal of Heat and Mass Transfer, r vol. 51, no. 1.

vol. 43, no. 3. pp. 321–330. BRÄNNBACKA, J. AND SAXÉN, H. 2008. Model for fast computation of blast furnace hearth erosion and buildup profiles. Industrial and Engineering Chemistry Research, vol. 47, no. 20. pp. 7793–7801.

pp. 186–197. ZAGARIA, M., DIMASTROMATTEO, V., and COLLA, V. 2010. Monitoring erosion and skull profile in blast furnace hearth. Ironmaking and Steelmaking, g vol. 37, no. 3. pp. 229–234. ZHAO H.B., CHENG S.S., and ZHAO, M.G. 2007. Analysis of all-carbon brick

Oxford University Press, Southampton. pp. 8–13. The Journal of The Southern African Institute of Mining and Metallurgy

bottom and ceramic cup synthetic hearth bottom. Journal of Iron and Steel Research International, vol. 14, no. 2. pp. 6–12. VOLUME 114

◆

JUNE 2014

495

▲

CARSLAW, H. and JAEGER, J.C. 1959. Conduction of heat in solids. 2nd edn.

Invitation to the members of the SAIMM to form part of qualitative research to be conducted by members of the Department of Communication Science, Unisa Members from the Department of Communication Science, Unisa, under the leadership of Dr Elnerine Greeff extend an invitation to all SAIMM members to participate in the new research in the mining sector, which will focus on the history and contemporary perspectives of Fanakalo and Activism. The main objectives of the research will be to explore and describe the perceptions of those working, and otherwise engaged in this sector, as it pertains to these two phenomena, and as they manifest uniquely in the South African mining context. To this end, we invite members to participate in interviews with the research team, which will explore perceptions on: 1. Fanakalo (also Fanagalo) It is often seen as a mystical phenomenon – without proven origins and with as many negative connotations (such as racist subtexts) as positive ones (such as solidarity). A whole host of research has been done on the philological aspects of the language, but little attention has been given to contemporary opinions regarding its use, and place, in the industry today. We would like to explore opinions to this effect. 2. Activism In recent, postmodernistic views on organisational management, activism is seen as a positive phenomenon – as a symptom of a healthy organisation. The mining industry has a unique history concerning activism and in recent times, arguably, a uniquely negative application (or at least with negative consequences). We would like to explore where the line is drawn – how much activism is too much activism, and how do those situated in this industry view activism and its effects. Our guiding approach is that a whole host of research (especially in the Humanities) presumes to talk to the industry, and not with or from the industry. We would like to help fill this void by documenting and exploring opinions in an academic and responsible way. Should you wish to participate in this research excercise, please contact the project leader, Dr Elnerine Greeff in one of the following ways: Email: greefwj@unisa.ac.za, Tel: 012 429 3886, Cell: 082 446 0351.

Please note that the SAIMM is only involved in facilitating this process and, should you want to take part in the research, you act in your personal capacity (opinions do not necessarily reflect that of the SAIMM, and participation is completely voluntary). All research participants, should they wish it, will be treated with the utmost and strictest confidentiality – anonymity in the research process can be secured.

▲

496

JUNE 2014

VOLUME 114

The Journal of The Southern African Institute of Mining and Metallurgy

Life-of-mine ventilation and refrigeration planning for Resolution Copper Mine by S. Bluhm*, R. Moreby†, F. von Glehn*, and C. Pascoe‡

Synopsis Resolution Copper Mine is planned to be a 2000 m deep panel cave mine with virgin rock temperatures above 80°C in rock with a high crystalline silica content. The planned run-of-mine production rate is 120 kt/day. The project is in prefeasibility evaluation. This paper discusses features of the ventilation system design, which include multiple ventilation shafts with total flow of about 3000 m³ m³/s and both surface and underground refrigeration systems with more than 140 MW total capacity. This will be a very challenging mine to ventilate, but this work has demonstrated that it will be technically achievable with the application of existing technology. Keywords mine ventilation, mine refrigeration, dust management, heat load, modelling.

be an electrical rail haulage system reporting to the crushers, from which a belt conveyor will report to the hoisting shaft(s) loading facilities. There will be midshaft skip discharge, and from there another conveyor system will transport ore to the surface plant. The orebody has a high silica content that will create serious challenges for dust control. Furthermore, the mine will be deep, with high virgin rock temperatures. Thus dust and thermal issues dominate the design evaluations and, obviously, diesel particulate matter (DPM) management will also play an important role.

Design criteria Silica/quartzite levels The Resolution Copper project is located 110 km southeast of Phoenix, Arizona and is run by Resolution Copper Mining (RCM), which is owned by Rio Tinto and BHP Billiton. The orebody is a large, deep, high-grade porphyry copper deposit located close to the historic Magma Vein (Pascoe, Oddie, and Edgar, 2008). The mine is planned to be a large block cave operation and is currently at the prefeasibility stage. Assessment for the life-of-mine and ventilation studies has been done to establish ventilation distribution layouts, and ventilation and refrigeration needs and strategies. Although the mine development phases are hugely important to the project feasibility, this paper concentrates more on the ventilation designs for the fully established life-of-mine scenarios. Furthermore, while refrigeration will be important to the project, this paper discusses the ventilation issues in more detail than those of refrigeration. The run-of-mine production rate will be 120 kt/day and the mining method will be an advance undercut panel cave operation using single panels. The mine will employ three hoisting and service shafts and three upcast shafts. Electrical loaders will be used for production, but diesel equipment will be used for development and undercutting. There will The Journal of The Southern African Institute of Mining and Metallurgy

The main mining activity will be in rock with an average quartz content of 37 per cent and – depending on how crystal boundaries break – it is possible that the respirable dust fraction will have a free crystalline silica content up to 45 per cent. This high silica content will create serious challenges for dust control; however, with the highly mechanized operation most workers will not be in dusty areas for significant periods. Thus the maximum free crystalline silica limit was taken as 0.1 mg/m³ (40-hour time-weighted average) with a shortterm exposure limit of 0.3 mg/m³. The control of respirable dust will be a very significant issue and the best practice in terms of dust control measures will be applied.

* † ‡ ©

BBE Consulting. Morevent Mining. Mine Consulting, Rio Tinto. The Southern African Institute of Mining and Metallurgy, 2014. ISSN 2225-6253. This is an abridged version of the 2014 Best Paper Prize, originally presented at The Australian Mine Ventilation Conference 2013. A full version of the paper is available via the AusIMM website at www.ausimm.com.au/shop. VOLUME 114

JUNE 2014

497

▲

Introduction

Life-of-mine ventilation and refrigeration planning for Resolution Copper Mine Temperature issues The general geothermal gradient is 2.7°C per 100 m and the virgin rock temperatures on the crusher-to-conveyor level at 2180 m depth will be 83°C. The thermal conductivity will also be relatively high-particularly in quartz rich areas (4.8-6.6 W/m² °C). Figure 1 presents surface design ambient temperature data. After detailed statistical evaluation, the design conditions were taken as 21/37°Cwb/db (wet bulb/dry bulb) at 88 kPa barometric pressure. The refrigeration required depends on the design temperatures and balancing of the heat stress management needs with the cost of refrigeration. The design reject temperatures have been taken as 27.5/37.5°Cwb/db in development faces where personnel could be outside air-conditioned cabins, and 30/40°Cwb/db in production crosscuts where equipment is operated remotely. Heat stress management systems will be set up to determine different levels of operational control for rapid decision-making with various heat-related interventions.

Diesel particulate matter exposure management In addition to the dust and heat, DPM management will be an important design issue (NIOSH, 2011). Controlling employee DPM exposure to less than the current tolerance level values and potential future values will be addressed by management plans, including:

refrigeration f needs, and this is the basis off the detailed VUMA modelling. At this phase, the basic underground vehicle complement and equipment will be: ➤ ➤ ➤ ➤

Diesel auxiliary and service vehicles Diesel loaders and diesel rockbreakers Electrical loaders on the extraction level Drill rigs (blind borers, raise borers)

8.0 MW 4.2 MW 4.6 MW 4.1 MW

The following general ‘infrastructure’ zones were accounted for in the modelling: workshops, warehouses, batch plant, pump stations, refrigeration plant chamber, conveyor belt system to hoisting shaft facilities, crusher station, electrical substations, etc. In these zones the ventilation requirements will relate to the equipment/activities such as diesel vehicles, welding, pumps, compressors, vibratory feeders, belt conveyors, dust collection, fans, transformers, belts, fans, lights, etc.

Heat from broken rock The broken rock flow rate will be large and the rock will be hot, thus creating a significant heat load. Hypothetically, the ‘worst case’ would be if all rock enters the draw points at the virgin rock temperature and leaves the mine at the ambient underground temperature.

➤ Underground diesel vehicles with tier-3 or tier-4 engines using high-quality, low-sulphur fuel ➤ Exhaust conditioning over and above that provided normally on engines will be applied ➤ Operators to be in air-conditioned cabins with filtered intake where possible ➤ Series ventilation will be minimized or avoided altogether due to heat considerations ➤ Diesel vehicles will be operated according to a sitespecific, risk-assessed DPM control plan.

Features of the fully established mine The critical design year snapshot in the life-of-mine (Figure 2) was selected for sizing the ultimate ventilation and

Figure 2—Snapshot at critical design condition

Figure 1—Surface ambient temperature data

▲

498

JUNE 2014

VOLUME 114

The Journal of The Southern African Institute of Mining and Metallurgy

Life-of-mine ventilation and refrigeration planning for Resolution Copper Mine Assuming basic parameters, this ‘worst case’ hypothetical potential heat flow will be of the order of 60 MW. This is large but it presents potential mitigation opportunities by reduced residence times in intake airways or separation of haulage route and crusher ventilation circuits from other intake airways. The estimation of the heat load from the broken rock involved the fundamental assessment of residence time in various sections (Figure 3), rock sizes, and transient cooling in broken rock masses. While this required simplifying assumptions, it supplied order-of-magnitude guidance and indicated that about 50 per cent (30 MW) of the ‘worst case’ hypothetical potential will manifest itself in the underground mine. The remaining potential heat will be in rock that leaves the mine in a ‘still hot’ condition.

Primary ventilation system Primary ventilation flow rates The overall primary ventilation distribution is 2760 kg/s (or 2345 m³/s) to the underground workings. The total flow from surface will be 3120 kg/s, with the difference being that used on the midshaft skip discharge level. During full production, the needs will be dominated by the 70-odd extraction drives with 35 regulated exhaust raises serving each pair. Some of the specific considerations included: ➤ Development crew allocations based on two separate active faces per crew with exhaust vent raises installed to limit force duct lengths to <500 m ➤ Undercut and drawbell levels involve both development and mucking of swell ➤ Extraction drives will have remotely controlled electric loaders only ➤ Potential for high dry bulb temperatures will be prevalent on the production level and liberal amounts of water will be used for controlling temperature and dust

➤ Ventilation ffor ffuel/tyre bays will report directly to return airways, but other workshop ventilation will be re-used ➤ Dust management will be based on capture velocities of about 2.5 m/s applied to the cross-sectional area of the capture point ➤ Midshaft skip discharge dust control will use a return connection to No. 14 Shaft exhaust ➤ The underground conveyor ramp with crusher station and rock handling systems will be isolated as an independent ventilation district.

Comparison with other block cave mines Figure 4 compares the planned ventilation requirements for RCM and those of some other block cave mines. Considering the need at RCM to manage high heat loads and to allocate relatively large flow rates for controlling dust, this comparison indicates that the planned ventilation rates are consistent with those employed at other block cave mines (assuming flows reported for the other mines are at surface density).

Figure 4—Resolution ventilation rate benchmarking

The Journal of The Southern African Institute of Mining and Metallurgy

VOLUME 114

JUNE 2014

499

▲

Figure 3—Ore transport flow chart

Life-of-mine ventilation and refrigeration planning for Resolution Copper Mine Shafts and primary ventilation infrastructure Figure 5 shows the life-of-mine primary ventilation circuit. No. 11, No. 12, and No. 13 Shafts will downcast and No. 9, No, 10, and No. 14 Shafts will upcast together with exhaust via the conveyor drift. Provision is made for a return connection between the skip discharge zone and midshaft pump station to the exhaust shafts and conveyor tunnel to surface. The ventilation distribution in each airway, with the exception of the conveyor and regulated returns, is based on free splitting out of the shafts and through the workings. The return ventilation from the mining block will join the return air from the ore handling system and flow in the main return system to the upcast shafts. The underground return airways will operate at near-limiting air velocities for occasional vehicle or pedestrian access. The main fan stations will be installed on surface and will operate in exhaust mode. To allow for phased increase in ventilation capacity, available surface space during sinking, and long-term security of production, it is planned to connect all exhaust shafts together in a manifold configuration (all fans connected to all shafts).

High-speed intake airways with booster fans The primary intake infrastructure will include two large airways (each >60 m²) that will be used as high-speed dedicated intakes. These airways will be no-go zones and will be operated at air speeds >11 m/s, and they will carry more than 60 per cent of all underground ventilation (below skip discharge level). Booster fans will be installed in each of the airways with the objective of optimizing the carrying capacity of these large airways as well as controlling air velocity in other trafficable airways. Each of these fan stations will have four fan motor sets (each 250 kW) installed in parallel.

Undercut levels There will be two undercut levels (spaced at 20 m) in order to achieve the required wide inclined undercut for cave initiation. The undercut levels are generally challenging from the ventilation perspective, and they were examined in particular detail with stand-alone VUMA models. In the

models, provision was made ffor dead-end swell mucking, rim tunnel and crosscut development, and cubbies for swell muck dump every 200 m.

Production level Once steady production has been achieved, there will be up to 70 open, 150 m long production crosscuts. Of these, 35 production crosscuts will be active (35 electric loaders will be mucking) and 35 production crosscuts will be available for operation. The production crosscut allocation alone is 840 kg/s. The air will be distributed to open production crosscuts by regulators or secondary fans in the return raises, controlled from a central control station. This is a relatively low air allocation and the ventilation distribution control will be challenging. This specific VUMA model was scrutinized in detail (Figure 6).

Dust management The management of respirable dust will be very important because of the high levels of silica. The dust will have to be suppressed, captured, or report to return airways. The main strategy will be to direct contaminated air to return airways and for this, significant ventilation capacity has been allocated. With respect to control of respirable dust, provision will be made for industry best-practice dust controls such as: ➤ Dedicated exhaust ventilation systems with large flows. In total, there will be 755 kg/s (25 per cent of the entire primary ventilation) allocated to dust control direct to return ➤ Remote loading of extraction level with operator location on the intake side ➤ Extensive use of water sprays (which will also help control high dry bulb temperature) at draw points and loading/transfer points, thus wetting broken rock to at least 2 per cent moisture ➤ Road base maintenance and dust suppression with water suppression sprays in all roadways for routine and controlled wetting ➤ Limited access to return airways ➤ Overall layout design will minimize the need for filtration and re-use of contaminated air. However,

Figure 5—Distribution of ventilation

▲

500

JUNE 2014

VOLUME 114

The Journal of The Southern African Institute of Mining and Metallurgy

Life-of-mine ventilation and refrigeration planning for Resolution Copper Mine

Figure 6—Production level temperatures (blue = 20°Cwb; red = 30°Cwb)

Figure 7—Ultimate VUMA model (blue = 20°Cwb; red = 30°Cwb)

The following dust generation and ore transfer points have been accounted for: extraction level loader transport and tipping; rail loading level; rail tip to belt feeders (rail tip level, crusher level, transfer conveyor feeders); skip loading; skip discharge, shaft-to-surface conveyor drift.

VUMA heat load and energy balance modelling The design snapshot scenario (Figure 2) was selected as the critical design year for sizing the ultimate ventilation and refrigeration needs. VUMA software was used to model the full mine layout and to simulate the different ventilation The Journal of The Southern African Institute of Mining and Metallurgy

strategies. This includes the ffull interactive simulation off the heat flow, ventilation, and cooling systems to determine the air temperatures, flow rates, heat loads, and cooling requirements (Figure 7). The simulations take full account of the block cave mining details, and this software is unique in that it deals with the important effects of broken rock and advancing rock faces. This design snapshot relates to the period when the first panel is approaching completion and the second panel is starting production. Other scenarios towards the end of lifeof-mine as well as the development phases were modelled separately (but are not discussed here). The heat load due to all the mobile equipment and static equipment facilities was input at the relevant nodes and branches, as was the heat flow from the broken rock discussed above. The models indicated that the mine heat loads will be satisfied with the following ventilation resources: ➤ Chilled ventilation downcast ex from surface 3120 kg/s at 10.5°Cwb VOLUME 114

JUNE 2014

501

▲

some ffiltration may be needed in areas remote ffrom return airways such as main conveyor transfer points and secondary belts ➤ Extensive use to be made of PPE and personnel will be trained in the control of exposure to respirable dust ➤ Respirable dust sampling regime will be set-up to monitor employee exposure and samples will be analysed for composition, including respirable free silica in its various forms.

Life-of-mine ventilation and refrigeration planning for Resolution Copper Mine ➤ Surface f bulk air cooler duty ➤ Chilled service water ➤ Underground secondary air coolers

105 MW 50 L/s at 4°C 35 MW

The global energy balance can be satisfied by different combinations of higher air flow rates with cool air or lower air flows with colder air. The optimum is dictated by issues such as available downcast capacities, overall costs of the ventilation and cooling systems, standard equipment capacities, and phase-in needs. Following a number of iterations that included the shaft(s) sizing, sensitivity studies (more/less underground refrigeration, use of ice, etc.), the above mix of flow rate and refrigeration capacity is considered to be close to optimum. Trade-off studies were conducted with different splits between surface and underground refrigeration and the manner in which the cooling is distributed. In summary, a significant refrigeration capacity will be on surface. However, the underground refrigeration will be extremely important during the development phases and will provide the essential high positional efficiency air coolers directly in the workings during production phases.

Description of refrigeration system Surface refrigeration system The surface refrigeration system will comprise: ➤ Central surface refrigeration plant room and refrigeration machines (and thermal store) ➤ Surface bulk air coolers at each downcast shaft ➤ Service water refrigeration system to provide chilled surface water to underground. From the central plant room, chilled water will be served to surface bulk air coolers at each of the downcast shafts. In addition, there will be a supplementary surface refrigeration system that will provide general chilled service water to underground. The primary refrigeration system will comprise main base load machines prechilling water flow from the bulk air coolers. From these plants, chilled water will then flow to the thermal storage dam containing tube banks through which subzero glycol is circulated. Ice will be formed on the outside of the tubes during the colder part of the day and then melted by the circulating water during the warm part of the day. The chilled water will leave the thermal storage dam at temperatures close to 0°C. The thermal storage will allow peak load damping and optimal energy management. The combined plant and ice store system will provide 109 MW nominal refrigeration capacity. There will be a number of large refrigeration machine modules chilling water and glycol. All the machines will be similar, with interchangeable components. The refrigeration machine modules will be factory-assembled and packaged plants with R134a centrifugal compressors and shell-andtube evaporators/condensers. Each plant will have differing process conditions that will depend on final equipment selection. However, for example, the lead water chilling plants will have a refrigeration duty of 22 MW. Each of the downcast shafts will be served by bulk air coolers in the form of horizontal spray heat exchangers in which the air is forced through an intense spray of chilled

▲

502

JUNE 2014

VOLUME 114

water in a horizontal concrete tunnel (Bluhm et al., 2001). Within the sprays, heat exchange will occur directly across the large surface area of the spray drops. Where the cool air emerges from the chamber, mist eliminators will be installed to ensure that no water is carried out. For example, the No. 13 Shaft bulk air cooler will have a cooling duty of 40 MW. Chilled service water will be a very important part of the underground cooling. The chilled service water will be used for dust control, localized cooling sprays, and mine service needs and will provide effective localized cooling wherever it is applied. Thus, on surface, in addition to the main refrigeration system, there will be a separate independent surface refrigeration system that will provide general chilled service water to all the underground workings.

Underground refrigeration system The underground refrigeration system will comprise two main components: ➤ Centralized underground refrigeration plant chamber and refrigeration machines ➤ Suite of underground air coolers and cold water distribution system. There will be a single central underground refrigeration plant chamber located near the mining block off the exhaust vent level. The return ventilation system will be used for heat rejection from the refrigeration plant. The refrigeration machines will provide cold water, in an insulated closed circuit network, to air coolers situated strategically throughout the workings. The required secondary air cooler overall duty will be 35 MW and the underground refrigeration plant will provide 40 MW capacity to overcome losses. The refrigeration installation will ultimately have five 8 MW machines operating (plus one standby). The machines will operate in parallel and this arrangement will allow the system to adjust to changes in demand. The machines will use refrigerant R134a and will include high-speed, multistage, centrifugal compressors. The refrigeration machines will be identical packaged units with compressor motor sets and shell and tube evaporators/condensers. The plant room layout includes two machine rooms with the chilled water pumps and condenser water pumps grouped in common pump chambers adjacent to their respective dams. Rejected heat will be in cooling towers in the form of spray filled vertical excavations rejecting heat into the return ventilation. The refrigeration plant will be adjacent to two large main return airways and a large part of the return flow will be directed to the cooling towers. For the ultimate life-of-mine scenarios, there will be numerous underground secondary air coolers with duties ranging from 0.5-3 MW, with a total duty of 35 MW. These secondary air coolers will be in the form of closed-circuit cooling coils. The main chilled water piping will comprise a 500 mm insulated system near the plants, reducing as the network splits up to ultimately 150 mm insulated pipe sections.

Conclusion Resolution will be a deep, hot, block cave operation with 120 kt/day production. The rock will have a high silica The Journal of The Southern African Institute of Mining and Metallurgy

Life-of-mine ventilation and refrigeration planning for Resolution Copper Mine content and the mine will be deep in hot virgin rock temperature. Thus dust issues and thermal issues dominate the design evaluations. The mine will employ three hoisting and service shafts and three upcast shafts, and the total primary ventilation capacity will be about 3000 m³/s. The general approach to dust management will be to direct contaminated air to return, and some 25 per cent of the total primary air flow will be allocated in this manner. The primary intake system will include two large airways which will be used as high-speed dedicated intakes with booster fans. These airways will be no-go zones operating at high air speeds and will carry more than 60 per cent of all underground ventilation. VUMA software was used to model the full mine layout and simulated different ventilation and refrigeration strategies. The simulation uses an iterative process to determine heat loads and cooling needs as well as the sizing and positioning of refrigeration system components. The large heat load components were the broken rock flow, surrounding rock conduction and mobile and static equipment facilities. The refrigeration systems will include 105 MW surface bulk air cooler duty and 35 MW underground secondary air cooler duty. This will be a very challenging mine to ventilate and cool, but this work has demonstrated that it will be technically achievable with the application of existing technology.

ISO. 2004. Standard 7933, Ergonomics off the thermal environment – analytical determination and interpretation of heat stress using calculation of the predicted heat strain. 2nd edn. Geneva. NATIONAL INSTITUTE OF OCCUPATIONAL SAFETY AND HEALTH (NIOSH). 2002. NIOSH Hazard Review, Health effects of Occupational Exposure to Respirable Crystalline Silica. Atlanta, GA. NATIONAL INSTITUTE OF OCCUPATIONAL SAFETY AND HEALTH (NIOSH), 2011. Diesel Aerosols and Gases in Underground Mines: Guideto Exposure Assessment and Control. (eds.) Bugarski, A.D., Janisko, S.J., Cauda, E.G., Noll, J.D., and, Mischler, S.E. Report of Investigations 9687. Atlanta, GA. NATIONAL TOXICOLOGY PROGRAM (NTP). 2003. Silica Crystaline (Respirable Size), US National Toxicology Program. US Department of Health and Human Services, Research Triangle Park, NC. PASCOE, C., ODDIE, M., and EDGAR, I. 2008. Panel caving at the Resolutioncopper project. Proceedings of the Fifth International Conference and Exhibition on Mass Mining, g Lulea, Sweden. ◆

Acknowledgement Permission from Rio Tinto RCM to present this paper is gratefully acknowledged.

Refserences

The Journal of The Southern African Institute of Mining and Metallurgy

AusIMM President Geoff Sharrock FAusIMM(CP) presenting the 2014 AusIMM Best Paper prize to Steven Bluhm in Cape Town

VOLUME 114

JUNE 2014

503

▲

BLUHM, S., FUNNELL, R., and SMIT, H. 2001. Horizontal spray chambers for surface bulk air cooling. Proceedings of the Seventh International Mine Ventilation Congress, Cracow, Poland.

INTERNATIONAL ACTIVITIES 9–11 September 2014 — 3rd Mineral Project Valuation School Mine Design Lab, Chamber of Mines Building, The University of the Witwatersrand

11–12, June, 2014 —

26–30 June 2014 —

16–17 September 2014 —

E-mail: raymond@saimm.co.za 15–16 July 2014 — Mine Planning School Mine Design Lab, Chamber of Mines Building, The University of the Witwatersrand

E-mail: camielah@saimm.co.za 4–5 August 2014 — Emperors Palace Hotel Casino Convention Resort, Johannesburg

E-mail: camielah@saimm.co.za 23–24 September 2014 —

E-mail: raymond@saimm.co.za 20–24 October 2014 — 6th International Platinum Conference Sun City, South Africa

E-mail: raymond@saimm.co.za E-mail: camielah@saimm.co.za

12 November 2014 —

6–8 August 2014 — MinPROC 2014 Lord Charles Hotel, Somerset West, Cape Town 20–22 August 2014 — MineSafe Conference 2014 Technical Conference and Industry day 20–21 August 2014: Conference 22 August 2014: Industry day Emperors Palace, Hotel Casino Convention Resort, Johannesburg

E-mail: raymond@saimm.co.za 18–19 November 2014 —

E-mail: raymond@saimm.co.za 19–20 November 2014 —

1–2 September 2014 —

Emperors Palace, Hotel Casino Convention Resort, Johannesburr

E-mail: raymond@saimm.co.za

▲

JUNE 2014

The Journal of The Southern African Institute of Mining and Metallurgy

INTERNATIONAL ACTIVITIES 11–13 March 2015—

12–14 August 2015—

E-mail: camielah@saimm.co.za 8–10 April 2015—Sulphur and Sulphuric Acid 2015 Conference 8–9 April 2015—Conference 10 April 2015—Technical Visit Southern Sun Elangeni Maharani KwaZulu-Natal, South Africa

E-mail: raymond@saimm.co.za 28 September-2 October 2015— —

E-mail: camielah@saimm.co.za 9–10 June 2015— E-mail: raymond@saimm.co.za E-mail: raymond@saimm.co.za

12–14 October 2015—

14–17 June 2015 —

E-mail: raymond@saimm.co.za

16–20 June 2015 —

28–30 October 2015— —

6–8 July 2015—Copper Cobalt Africa Incorporating The 8th Southern African Base Metals Conference The Falls Resort, Victoria Falls, Livingstone, Zambia

E-mail: raymond@saimm.co.za

8–13 November 2015—

13–17 July 2015—School on Production of Clean Steel Misty Hills Conference Centre, Muldersdrift

The Journal of The Southern African Institute of Mining and Metallurgy

Raj Singhal, E-mail: singhal@shaw.ca or E-mail: raymond@saimm.co.za, Website: http://www.saimm.co.za

JUNE 2014

▲

E-mail: raymond@saimm.co.za

Company Affiliates The following organizations have been admitted to the Institute as Company Affiliates

â&#x2013;˛

xii

JUNE 2014

The Journal of The Southern African Institute of Mining and Metallurgy

Forthcoming SAIMM events...

IP PONSORSH EXHIBITS/S ng to sponsor ishi e Companies w t at any of thes bi hi ex or and/ t the ac nt co ld ou events sh -ordinator conference co ssible po as on so as

2014

◆ SEMINAR 26–30 June 2014, ◆ SCHOOL Mine Planning School 15–16 July 2014, Mine Design Lab, Chamber of Mines Building, The University of the Witwatersrand ◆ CONFERENCE 4–5 August 2014, Emperors Palace, Hotel Casino Convention Resort, Johannesbur ◆ CONFERENCE MineSafe Conference 2014 Technical Conference and Industry day 20–21 August 2014, Conference 22 August 2014, Industry day Emperors Palace, Hotel Casino Convention Resort, Johannesburg ◆ 1–2 September 2014, ◆ SCHOOL 3rd Mineral Project Valuation School 9–11 September 2014, Mine Design Lab, Chamber of Mines Building, The University of the Witwatersrand ◆ CONFERENCE 16–17 September 2014, ◆ 23–24 September 2014, ◆

6th International Platinum Conference 20–24 October 2014, Sun City, South Africa

◆

For further information contact: Conferencing, SAIMM P O Box 61127, Marshalltown 2107 Tel: (011) 834-1273/7 Fax: (011) 833-8156 or (011) 838-5923 E-mail: raymond@saimm.co.za

12 November 2014, ◆ 19–20 November 2014, Emperors Palace, Hotel Casino Convention Resort, Johannesbur

Website: http://www.saimm.co.za