I-* •"5? '^ = 
 
 • -4^, xTV 4. J 
 
 
 
 
 
 
 
 
 
 ' t. 
 
 '^0^ 
 
 Jp-n,. 
 
 v^^°^ '. 
 
 
 ■-^ju^;^ oV'^^a"- '»j3^ f^^/^< "^^d^ ' oV'^^R'- V„<V 
 
 
 
 
 
 
 
 
 L» >„ c'^ 
 
 
 Ay ^^ * o a 
 
 
 *bv^ 
 
 
 
 
 
 bV 
 
 
 
 ^ •» 
 
 ' J "-. '^"^0^ ^O'^'S^ "^P"^° / % ^^^»>* .0^°^t. l^P".' / "-0 
 
 
 
 <5^ A .. 
 i/"^ 
 
 ■i 
 
 
 
 % 0" 
 
 0" 
 
 *bv^ 
 
 ^°"%., 
 
 y % *''^ .♦^ 
 
 
 
 
 
 
 
 t^ A^ **^Va^ \^/* /^fe\ ^^'^^ ^'^^Va^ -^^^ c^^* -^'- 
 
 ^^.^i' 
 
 v*^* 
 
 'o, *.t;t* a ^. 'o,»* .G^ ''V *^T*^ A <* '"•** ^C» '^ *;^T« A 
 
 IP's 
 
 
 # * 
 
 
 
 
 '^ .-^'•. "^ 
 
 
 
 ■J? . 
 
 S *5 
 
 
 rO' 
 
 
 
 
 
 
 
 
 5> .»»'^ 
 
 <>» ^OMO" <5,> 
 
 Ci. * 
 
 
 
 50 
 
 IN/ 
 
 V^ ^o 
 
 
 « ■ te. A^ ' »>^^A" '"^^^ .c#^ *'«K'. ■ -e.. A^ ' /. 
 
 ■^ A>' ♦ 
 
 
 
 • " ^. 4 ^ 
 
 to 
 
 • •a.i 
 
 r 
 
 _^^°^ 
 
 o ■*■ 
 
 "^bt^ y^^^'". '-^U.a'? 
 
 ^^-n^. 
 
 
 ';^o^ 
 
 
 ^oV 
 
 v^o^ 
 
 
 Vj. * o « ' •o-^ 
 
 
 "q,/*lT;'' v^o 
 
 ■^ A 
 v-^^ 
 
 
 
 
 J^ * 
 
 
 - -t. 
 
 '-^^0^ o^ 
 
 
 
 • % A^ ' 
 
 \y "^ •"'• ^A" . . . 
 
 A^ 
 
 
 , t t- (» » a V ^-^ - " « e 3 ' "t' *■ "^-^ 
 
 '^^^o^ 
 
 *Pvs ; 
 
 *■"-%. 
 
 '. -^ 
 
 
 
 ^> 4 
 
 '^^..♦^ .'I(«^^o Vc/ /^*-. '^^^..♦^ /i^i^'".. V# 
 
 
 ^^"i^ 
 
 .\. /,.^>.>o ./\c:^%^^.. /..^>>o ..^^\c:^/V /.»^>A 
 
 ^'. '-^^.V oV^ma'- "^-oV^ ;^^'-' ■^-^^O*' f^'^a"". ''^oV^^ 
 
 
 c-^^o 
 
 
 o V 
 
 
 
 
 
 
 
 "-^^^' 
 .^^•v. 
 
A-* o 
 
 \ 
 
 
 
 v^' 
 
 •^ " " * A 
 
 " a « *o 
 
 
 4*" ..^'•. •* 
 
 
 
 
 
 M" ^ov^ :mm^- "^-"0^ ° 
 
 --■\. 
 
 
 
 ^ '*^t;t*' a 
 
 
 •5."' "-^ %' 
 
 > .0' 
 
 ^ ♦^T.T* A 
 
 
 ^" '^ov^" ;:^^'-. ^^^0^ f^^K". ^ov* 
 
 
 
 
 X3 V 
 
 ^0^ o , 
 
 >°-^^. 
 
 
 ° ^ 
 
 / 
 
 \ 
 
 
 
 ^» # "^ -: 
 
 0" o'*^''*^ o 
 
 
 
 
 <.^' '^«^. 
 
 ^ -.;:w^. 
 
 jP'j!' 
 
 
 <V * rCV\ SSA "^n V * OIKS * rJ» •vC*^ * J 
 
 
 
 
 1- ^^0^ 
 
 <V *'"■ 
 
 
 •^ 
 
 ^^' ^'A 
 
 
 ^'* V 
 
 
 V'^'' 
 
 
 <^^> 
 
 ^ ...... '^^ 
 
 .^ .*'. 
 
 
 ' -^^^ 
 
 
 
 r» -e-d^ :- 
 
 .^^°- 
 
 
 
 ^""^4, 
 
 
 • -ay Cl. *, 
 
 o^ * 
 
 ^^ o. 
 
 
 
 
 ^^ . 'WW/ ..^^\. '•' 
 
 
 ^' /% 
 
 %, . » . * - , 
 
 ..o« 0^ **© 
 
 <> '° 
 
 
 V » ' * "^ C5. 
 
 
 o, *^TeT» A 
 
 
 
 
 ^bt?' 
 
 n4q. 
 
 A'- ,l/fl^ <*. fsV ^OKO. -^^ 
 
 
 *bv" 
 
 
 t^o^ 
 
 
 
 "-^..^^ 
 
 ■^^^^^ 
 
 
 O, *J7Vi* A 
 
 'V ^ ", 
 
 1 » 'p- 
 
 A> ,>.-<i^ -^^ 
 
 o > 
 
 
 o^ *^t;T* a 
 
 
 v.tff 
 
 
 ^oV" 
 
 ',* qO 
 
 jP-^h 
 
 
 
 r.-' ^0^ 
 
 
 
 
 •y ^ ".^ 
 
 
 
 ^ ^ 
 
 
 <* 'O.I.* A 
 
 1^ f "^ 
 
 
 
 .0'* ^^ \.**^^*' aO 
 
 5- ^VvVa'- "^^^ cV" 
 
 
 <r <;. 'o,»* A^ 
 
 ' ^. 4 
 
 
 
 "*bV^ 
 
 •^ i'^' 
 ^^\^ 
 
 V 
 
 
 
 IliB^ o A^O. o 
 
 
 "-^-^ 
 
 ■^, 4 =" 
 
 
 ' tc. A^ ■ ^^tK^^A'-. %. ^^ - 
 
 
 C.-' "/r. -^ ^^illl^' o A.^r^. 
 
 
 4 o 
 
IC 
 
 8973 
 
 Bureau of Mines Information Circular/1984 
 
 Mine Ground Control 
 
 Proceedings: Bureau of Mines Technology Transfer 
 Seminars, Pittsburgh, PA, December 6-7, 1983, 
 and Denver, CO, December 8-9, 1983 
 
 Compiled by Staff, Bureau of Mines 
 
 UNITED STATES DEPARTMENT OF THE INTERIOR 
 
 
Information Circular 8973 
 
 V 
 
 Mine Ground Control 
 
 Proceedings: Bureau of Mines Technology Transfer 
 Seminars, Pittsburgh, PA, December 6-7, 1983, 
 and Denver, CO, December 8-9, 1983 
 
 Compiled by Staff, Bureau of Mines 
 
 UNITED STATES DEPARTMENT OF THE INTERIOR 
 William P. Clark, Secretary 
 
 BUREAU OF MINES 
 Robert C. Norton, Director 
 

 Library of Corigress Cataloging In Publication Data: 
 
 Bureau of Mines Technology Transfer Seminars (1983 : 
 Pittsburgh, PA, and Denver, CO) 
 
 Mine ground control. 
 
 (Information circular / Bureau of Mines ; 8973) 
 
 Includes bibliographies. 
 
 Supt. of Docs, no.: I 28.27:8973. 
 
 1. Ground control (Mining)— Congresses. I. United States. Bureau 
 of Mines. II. Title. III. Series: Information circular (United States. 
 Bureau of Mines) ; 8973. 
 
 TN295.U4 [TN288] 622s [622'.28] 83-600351 
 
^ PREFACE 
 
 N This Information Circular summarizes recent Bureau of Mines research 
 fv^ results concerning ground control in mining environments. The papers 
 
 represent only a sample of the Bureau's total research effort in this 
 
 area, but they outline major portions of the program. 
 
 :^^ The papers reproduced here were presented at Technology Transfer Sem- 
 
 '^ inars on Mine Ground Control given in December 1983 in Pittsburgh, PA, 
 
 xA| and Denver, CO. Those desiring more information on the Bureau's ground 
 
 X control program or health and safety technology programs in general, or 
 
 V information on specific situations, should contact the Bureau of Mines, 
 
 ^ Division of Health and Safety Technology, 2401 E Street, NW. , Washing- 
 ton, DC 20241, or the appropriate author listed in these proceedings. 
 
 1^ 
 
 
iii 
 
 CONTENTS 
 
 Page 
 
 Preface 1 
 
 Abstract 1 
 
 Introduction, by Chi-shing Wang and John M. Karhnak 2 
 
 Mine Design 
 
 Geologic Studies in Ground Control, by Noel N. Moebs 4 
 
 Coal and Rock Properties for Premine Planning and Mine Design, by 
 
 Richard E. Thill 15 
 
 Pillar Design Equations for Coal Extraction, by Clarence 0. Babcock 36 
 
 Underhand Cut-and-Fill Stoping for Rock Burst Control, by F. Michael Jenkins 
 
 and K. Robert Dorman 49 
 
 Hazard Detection 
 
 Acoustic Cross-Borehole System for Hazard Detection, by Karen S. Radcliffe and 
 
 Richard E. Thill 64 
 
 In-Seam Geophysical Techniques for Coal Mine Hazard Detection, by 
 
 Richard J. Leckenby and James J. Snodgrass... 75 
 
 Satellite Imagery as an Aid to Mine Hazard Detection, by Robert A, Speirer and 
 
 Stanton H. Moll 89 
 
 Microseismic Techniques Applied to Failure Warning in Mines , by 
 
 Fred W. Leighton 96 
 
 Mechanical and Ultrasonic Closure Rate Measurements, by Roger McVey 108 
 
 Ground Installation Equipment 
 
 Remote Manual Roof Bolters, by John E. Bevan 117 
 
 Field Test of an Automated Temporary Roof Support (ATRS) Used on a Low-Coal, 
 Single, Fixed-Head Roof Bolting Machine (Squirmer), by Charles T. Chislaghi 
 and Thomas E. Marshall 122 
 
 Roof Support Systems 
 
 Development of Epoxy Grouts and Pumpable Bolts, by Robert R. Thompson 126 
 
 Development of Lightweight Hydraulic Supports , by John P. Dunf ord 129 
 
 Mobile Roof Support and Applications in Retreat Mining, by Robert R. Thompson.. 133 
 
 Inorganic Grouts for Roof Bolting, by Jack E. Fraley 138 
 
 Research, Development, and Use of Steel-Fiber-Reinforced Concrete Cribbing for 
 
 Mine Roof Support, by Thomas W. Smelser and Dale A. Didcoct 146 
 

 UNIT OF MEASURE ABBREVIATIONS 
 
 USED IN 
 
 THESE PAPERS 
 
 A 
 
 ampere 
 
 lb 
 
 pound 
 
 cfm 
 
 cubic foot per minute 
 
 L/min 
 
 liter per minute 
 
 cm 
 
 centimeter 
 
 L/s 
 
 liter per second 
 
 deg 
 
 degree 
 
 m 
 
 meter 
 
 ft 
 
 foot 
 
 MHz 
 
 megahertz 
 
 ft/min 
 
 foot per minute 
 
 mi 
 
 mile 
 
 ft/s 
 
 foot per second 
 
 min 
 
 minute 
 
 gal 
 
 gallon 
 
 pin/in 
 
 microinch per inch 
 
 g/cm^ 
 
 gram per cubic centimeter 
 
 mm 
 
 millimeter 
 
 GPa 
 
 gigapascal 
 
 pm 
 
 micrometer 
 
 gpm 
 
 gallon per minute 
 
 MPa 
 
 megapascal 
 
 h 
 
 hour 
 
 ms 
 
 millisecond 
 
 hp 
 
 horsepower 
 
 US 
 
 microsecond 
 
 Hz 
 
 hertz 
 
 mV 
 
 millivolt 
 
 in 
 
 inch 
 
 ns 
 
 nanosecond 
 
 in/min 
 
 inch per minute 
 
 pet 
 
 percent 
 
 kHz 
 
 kilohertz 
 
 psi 
 
 pound per square inch 
 
 km 
 
 kilometer 
 
 s 
 
 second 
 
 km/s 
 
 kilometer per second 
 
 V 
 
 volt 
 
 kV 
 
 kilovolt 
 
 wt pet 
 
 weight percent 
 
 L 
 
 liter 
 
 yr 
 
 year 
 
MINE GROUND CONTROL 
 
 Proceedings: Bureau of Mines Technology Transfer Seminars, Pittsburgfi, PA, 
 December 6-7, 1983, and Denver, CO, December 8-9, 1983 
 
 Compiled by Staff, Bureau of Mines 
 
 ABSTRACT 
 
 These proceedings consist of papers presented at Bureau of Mines Tech- 
 nology Transfer Seminars in December 1983 for the purpose of disseminat- 
 ing recent advances in mining technology in the area of mine ground con- 
 trol. The Bureau of Mines conducts several of these seminars each year 
 in order to bring the latest results of Bureau research to the attention 
 of the mining industry as quickly as possible. 
 
INTRODUCTION 
 By Chi-shing Wang "I and John M. Karhnak'' 
 
 The basic goal of the Bureau of Mines 
 Ground Control research program is to 
 provide the mining industry with technol- 
 ogy that will lead to the reduction of 
 accidents due to falls of ground. The 
 problems of ground control are the ina- 
 bility to "see" geologic anomalies ahead 
 of the mine workings, the difficulty 
 in predicting ground movements induced 
 by excavation, and the need to provide 
 efficient ground control over the wide- 
 ly varying, and frequently unexpected, 
 conditions encountered from one place 
 to another. Current program objectives 
 include — 
 
 Exploring new methodologies and vali- 
 dating available tools for predicting 
 hazardous geologic features in ad- 
 vance of mining and for selecting and 
 designing safe and efficient mining 
 systems and layout plans. 
 
 Developing Instrumentation methods 
 and equipment able to detect and warn 
 of imminent failure of ground during 
 mining. 
 
 Ascertaining the most effective per- 
 manent support methods for specific 
 ground conditions. 
 
 Investigating roof support installa- 
 tion methods that improve working 
 conditions at the face and remove ma- 
 chine operators from dangerous unsup- 
 ported areas. 
 
 Advance knowledge of potentially haz- 
 ardous geologic features is a prerequi- 
 site to success in designing out the as- 
 sociated safety risks from the mining 
 plan during the early stage of mine de- 
 sign and development. To enable mine op- 
 erators to satisfy this requirement, the 
 Bureau has continued its efforts in 
 
 ^ Staff engineer, Division of Health and 
 Safety Technology, Bureau of Mines, Wash- 
 ington, DC. 
 
 assessing the applicability of various 
 geologic mapping methods and geophysical 
 techniques to detection and delineation 
 of geologic anomalies and mine voids from 
 the surface, boreholes, or mine workings. 
 
 The ability of miners to predict poten- 
 tial hazards ahead of working faces and 
 to detect impending ground failures at 
 the working place during mining is cru- 
 cial to timely evacuation and safety pre- 
 cautions to avoid injuries. Therefore, 
 Bureau research continues to assess in- 
 seam geophysical techniques for detecting 
 geologic anomalies and mine voids ahead 
 of coal faces in underground coal mines , 
 automatic microseismic roof fall warning 
 systems, and mechanical and ultrasonic 
 devices for continuous measurement of 
 roof-to-floor closure rates. 
 
 Better understanding and prediction 
 capability of mine geology and ground 
 movements serve as preventive measures 
 permitting mine operators to design out 
 potential ground hazards, and enabling 
 miners to avoid or escape from dangerous 
 areas in time. However, protective mea- 
 sures for assurance of safe mine workings 
 with an adequate ground support are es- 
 sential to maintaining active mining op- 
 erations. Hence, the Bureau has invested 
 considerably in research of more cost- 
 effective and more efficient temporary 
 and permanent roof support systems, al- 
 ternative ground support materials, and 
 remotely operated roof bolt installation 
 equipment and procedures. 
 
 This technology transfer seminar intro- 
 duces to the mining industry the signifi- 
 cant technological advances made by Bu- 
 reau research in recent years in various 
 aspects of ground control in mines. Em- 
 phasis is placed on those innovations 
 that have been validated by industrial 
 applications or by extensive field demon- 
 strations. Although many of those stud- 
 ies and innovations were conducted in or 
 designed for use in coal mine ground 
 
control, most of the new technologies 
 can be adapted for use in various mining 
 environments including metal and nonmetal 
 mines. As the current research policy 
 of the U.S. Government is to shift empha- 
 sis toward long-term, high-risk, and 
 high-payoff basic research, the mining 
 
 industry and mining equipment manufac- 
 turers are encouraged to assess the po- 
 tential of immediate uses or private 
 investments in commercial development of 
 Bureau research products that are offered 
 at this seminar. 
 
MINE DESIGN 
 
 GEOLOGIC STUDIES IN GROUND CONTROL 
 By Noel N. Moebs' 
 
 ABSTRACT 
 
 Several geologic techniques are re- 
 viewed that are useful in planning mines 
 and in solving ground control problems. 
 A summary is presented of a study of coal 
 
 mine roof structures and the use of 
 this information in improving supple- 
 mentary support and anticipating support 
 problems. 
 
 INTRODUCTION 
 
 The Bureau of Mines has been conducting 
 studies directed toward increasing the 
 utilization of geologic methods in coal 
 mine ground control, and more recently in 
 quarry operations, for the purpose of re- 
 ducing hazards from roof and rock falls. 
 It is commonly recognized that geology is 
 the key to effective ground control; that 
 is , a sound knowledge of the character 
 and structure of rock provides a sound 
 basis for mine planning and selection of 
 appropriate roof support methods. 
 
 For example, studies by the Bureau 
 of Mines have identified geologic struc- 
 tures in mine roof rock that contrib- 
 ute to many roof falls in Appalachian 
 coal mines. These structures, includ- 
 ing paleochannels , kettlebottoms , scours, 
 pinchouts, slickensides , clay veins, cre- 
 vasse splays, and joints, can often be 
 identified during, and sometimes before, 
 mine development. Mine projections can 
 be revised to reduce the adverse effects 
 of discontinuities in roof structure, 
 large roof areas of laminated sandstone 
 or incompetent strata generally can be 
 delineated or inferred from exploratory 
 drill-hole data, and the need for sup- 
 plementary support can be anticipated. 
 Accurate descriptions of roof geology 
 also provide some indication of optimum 
 length and type of roof bolts that should 
 be installed. 
 
 ^Geologist, Pittsburgh Research Center, 
 Bureau of Mines, Pittsburgh, PA. 
 
 In addition to studying geologic struc- 
 tures in coal mine roof , which are de- 
 scribed in another section, the Bureau 
 investigated the concept of a hazard map 
 that would delineate potentially hazard- 
 ous zones prior to mining. A method of 
 preparing a hazard map through the compu- 
 ter processing of drill-log data was de- 
 veloped, and the method was demonstrated 
 by preparing a hazard map of a small por- 
 tion of the Pittsburgh Coal Basin. This 
 methodology also should be applicable to 
 individual mines where an adequate den- 
 sity of exploratory drill holes is avail- 
 able. A computer program is available 
 for drill-hole data processing, which re- 
 sults in a printout map of anticipated 
 mining conditions. The identification 
 and weighting of hazardous geologic vari- 
 ables are provided for by modification of 
 the program. 
 
 Another geologic tool applied to ground 
 control studies by the Bureau is linear 
 analysis, 2 whereby an attempt is made to 
 correlate linear features from aerial 
 photographs or satellite images with sub- 
 surface rock stability conditions. The 
 results of this research in the Appa- 
 lachian coal region have, to date, been 
 inconclusive in that at some locations a 
 
 ^Linear, or photolineament, is a line 
 on an aerial photograph (or transposed 
 from satellite imagery) that is con- 
 trolled by an alignment of surface or 
 near-surface structures or conditions. 
 
correlation was discovered but at many it 
 was not. The lack of precise data on the 
 location of roof failure, the failure to 
 identify the "ground truth" or true geo- 
 logic character of a linear, and the 
 inability to selectively choose valid 
 lineaments all have severely limited the 
 practical application of linear analysis 
 to ground control studies. Continued re- 
 search is urgently needed to establish 
 the limitations and the potential of 
 linear analysis as a valid geologic meth- 
 od in ground control studies. 
 
 Virtually the only direct method of ob- 
 taining data on subsurface strata prior 
 to mining is through core drilling. Un- 
 fortunately, most drill core in the coal 
 regions has been logged by the driller 
 himself , and while this might have been 
 adequate to determine coal thickness and 
 general rock units, it is entirely inade- 
 quate for use in the assessment of an- 
 ticipated roof conditions, cavability, or 
 paleoenvironmental reconstructions to 
 determine coal continuity and quality. 
 Even with the increasing use of trained 
 professional geologists to log drill 
 core, there is still much to be desired 
 in the accurate and uniform identifica- 
 tion of rock types. Recognizing that 
 ground control investigations prior to 
 mining require the best possible drill- 
 core records , especially where no provi- 
 sion is made to preserve the core, the 
 Bureau has developed field guides for the 
 identification of cored rock in the 
 Pittsburgh Basin of Pennsylvania, Ohio, 
 and West Virginia, and in the Pocahontas 
 Basin of southeastern West Virginia. 
 These guides are pocket size, on weather- 
 proof paper, with a printed key to rock 
 types and color photographs of the actual 
 core. The Bureau produced these guide- 
 books on an experimental basis. They 
 have proven to be immensely popular and 
 useful, and as they become commercially 
 available a great deal more information 
 should be extracted from drill core than 
 before. 
 
 While most geologic studies have been 
 directed toward improving coal mine 
 ground control, a project recently initi- 
 ated in slate quarry operations utilizes 
 some of the same research methodologies 
 and thus is included in this overview. 
 The ground control problems of slate 
 quarry operations result partly from 
 geologic discontinuities in the vertical 
 quarry walls , which range from 80 to 560 
 ft in height (fig. 1). These geologic 
 discontinuities account for an occasional 
 rockfall or sudden collapse of an entire 
 highwall. These failures occur despite 
 scaling and pinning. Bolts are not em- 
 ployed for highwall support. Some rock 
 pressure problems are evident on the 
 quarry floors; the integrity of rock 
 barriers separating quarries sometimes is 
 in doubt ; and during the winter months 
 ice accumulations at the quarry brinks 
 are severe, and large masses or ice 
 stalactites that become dislodged consti- 
 tute a real hazard. Research to abate 
 these ground control problems includes 
 the use of strain gauges to detect immi- 
 nent falls of rock, the identification 
 and mapping of geologic features that 
 have contributed to highwall collapse, 
 the measurement of developed rock stress- 
 es, and the testing of both mechanical 
 and resin bolts to secure wall rock. 
 
 Following this brief review of some 
 geologic studies for ground control, the 
 remainder of this paper will summarize 
 the results of research on hazardous geo- 
 logic structures in coal mine roof. This 
 research provides an example of how the 
 recognition of coal mine roof structures 
 can lead to improved mine planning and 
 supplementary support. A bibliography of 
 selected references on this topic is pro- 
 vided at the end of this paper. Because 
 of the wide variance and diversity of 
 geologic structures in mine overburden, 
 only the most significant ones are dis- 
 cussed in the following sections. 
 
FIGURE 1, - Highwall of slate quarry operation. 
 
COAL MINE ROOF STRUCTURES 
 
 Among the causes of coal mine roof 
 failure are the hazardous roof conditions 
 produced by certain geologic structures. 
 The role of these structures in mine roof 
 falls is not always apparent because the 
 formations involved are sometimes poorly 
 exposed or difficult to recognize without 
 special training and underground experi- 
 ence. Most of the structures are small 
 scale, less than a few tens of feet in 
 width. In the underground coal produc- 
 tion environment, these structures com- 
 monly are unrecognized or are ignored by 
 all except those responsible for the 
 proper roof support installation. 
 
 PALEOCHANNELS 
 
 A paleochannel, or miners' "roll," is 
 the trough-shaped remnant of an ancient 
 stream channel that has been cut into 
 older rock, such as the roof shale and 
 coalbed. The channel subsequently has 
 been filled in with younger sediments, 
 usually sandstone. The paleochannels 
 most troublesome in mine roof support 
 seldom exceed about 30 ft (9.3 m) in 
 width and range from a few tens to a few 
 hundreds of feet in length. 
 
 Channels that are more than 30 to 50 ft 
 (9.3 to 15.5 m) wide and have cut down 
 through most of the coalbed are termed 
 "washouts." Although washouts can be a 
 severe problem in mine development or 
 longwall operations , they generally do 
 not cause serious roof fall problems for 
 more than about 50 ft (15.5 m) outside 
 the margins of the channel. 
 
 The most typical paleochannel is a 
 linear feature, often extending across 
 several adjacent entries. The slicken- 
 sides on the undersurf aces of the U- 
 shaped channel constitute planes of weak- 
 ness and a lack of bonding. Thus the 
 weak shales and coal adjacent to and 
 directly beneath the sandstone-filled 
 channel separate readily from it as soon 
 as the underlying coalbed is mined. The 
 most adverse roof condition occurs when 
 the channel trend is parallel to an en- 
 try, resulting in virtually continuous 
 
 roof support problems. Some shale-filled 
 channels also exist and constitute a sim- 
 ilar problem. 
 
 Paleochannels are best recognized by 
 their troughlike shape, slickensided mar- 
 gins, "horse-belly" appearance, and gen- 
 erally hard, crossbedded, poorly sorted 
 sandstone filling. Methods that have 
 been most often successful in supporting 
 shale channel margins are strapping and 
 the use of wood headers, and, for severe 
 cases, posts and crossbars. The use of 
 angle bolts that anchor into the hard 
 channel filling (fig. 2) is suggested as 
 another possible means of support. Un- 
 derground mapping is the best method of 
 determining the trends of channels. 
 Sometimes mine intersections can be re- 
 located to avoid an individual channel, 
 but, more often, several parallel chan- 
 nels occur, and the only option is to re- 
 vise the entry projections to intersect 
 channels at an obtuse angle. 
 
 Paleochannels in roof strata can rarely 
 be detected in advance of mining by nor- 
 mally spaced exploratory core drilling 
 (approximately 2,500-ft centers) because 
 of their small dimensions. However, 
 their presence can be inferred in areas 
 where drill core data indicate that 
 
 - UnderclQy 
 
 FIGURE 2. - Useof angleboltstosupportshale 
 strata at channel margins. Dashed line outlines 
 rock not fully supported by vertical bolting. 
 
thick, lenticular, crossbedded sandstone 
 occurs close to the top of the coalbed. 
 
 KETTLEBOTTOMS 
 
 A kettlebottom is a columnar mass of 
 rock embedded in and comprising a part of 
 the coal mine roof strata. Kettlebottoms 
 are the preserved casts of ancient tree 
 stumps that grew in the peat swamp, now 
 the coalbed. Once undermined, unsupport- 
 ed kettlebottoms can detach from the roof 
 at any time without warning, presenting a 
 hazard to miners. The size and frequency 
 of kettlebottoms in mine roof is depen- 
 dent upon geologic events and biological 
 processes active during deposition of 
 roof sediments. The preserved casts tend 
 to be small local features of erratic oc- 
 currence, which cannot be detected by 
 core drilling. They occur in the coal 
 measures throughout the Appalachians , but 
 are most abundant in the Pottsville age 
 deposits of southern West Virginia and 
 eastern Kentucky. All kettlebottoms not 
 dislodged after initial mining should be 
 properly supported. 
 
 SCOURS 
 
 The erosional action of an ancient 
 stream produces an almost endless variety 
 of cut-and-fill structures in rock. 
 Scours, high curved, oblong, or saucer- 
 shaped channel structures, must be ap- 
 proached with great caution, and the same 
 roof support methods should be utilized 
 as with the linear paleochannels. 
 
 Scours cannot be projected from entry 
 to entry as can the persistent and linear 
 channels, although, as with channels, 
 their presence in mine roof is more prob- 
 able where drilling indicates thick, len- 
 ticular, crossbedded sandstone close to 
 the top of the coalbed. 
 
 PINCHOUTS 
 
 The term "pinchout" is used here to 
 designate the abrupt termination of 
 a roof stratum (fig. 3). Some pinch- 
 outs have been observed near the flanks 
 of channels and washouts, and presuma- 
 bly were formed by the same cutting 
 
 action. Others appear to be due to very 
 rapid thinning during normal sediment 
 deposition. 
 
 If a pinchout occurs in a stratum that 
 is fairly thick (more than 1.5 ft or 0.5 
 m) , competent, and located in the immedi- 
 ate roof, the beam strength of the roof 
 will be weakened seriously by the discon- 
 tinuity. Pinchouts are not easily de- 
 tected until exposed by a roof fall. 
 They can sometimes be discovered or in- 
 ferred during drilling of roof bolt holes 
 when the roof bolter finds a pronounced 
 contrast in penetration rates between op- 
 posite sides of the entry. A pinchout 
 requires effective roof bolting to re- 
 inforce the weakened beam structure in 
 the roof through the normal roof support 
 plan or supplementary longer bolts that 
 anchor into competent overlying strata. 
 
 The pinchout most troublesome in mine 
 roof is a relatively small local feature 
 and cannot be detected by normal explora- 
 tory core drilling. However, its pres- 
 ence can be inferred in areas where the 
 same individual sandstone strata close to 
 the coalbed do not occur in adjacent 
 holes. 
 
 SLICKENSIDES 
 
 Probably the most common hazardous roof 
 structure occurring throughout the Appa- 
 lachian coal region is the slickenside, 
 or "slip." It is best developed and most 
 common in highly argillaceous rocks such 
 as shales, claystones, or miners' "clod." 
 A slickenside is a smooth, polished, and 
 sometimes striated or grooved surface re- 
 sulting from movement of rock on either 
 side of the surface. A slickenside usu- 
 ally is curved and in coal mine roof rock 
 generally is convex toward the coalbed. 
 
 The slickenside constitutes a disconti- 
 nuity in the beam roof structure and 
 therefore weakens the roof, creating a 
 potential hazard unless detected and 
 properly supported. It is difficult to 
 estimate the extent of the slickenside 
 simply by observing the trace in the mine 
 roof, but it can probably be assumed that 
 the longer the trace, the greater the 
 extent. 
 
Coalbed 
 
 Underclay- 
 
 _4Ft 
 
 Scale 
 
 FIGURE 3. - Abrupt termination of sandstone roof stratum. Dashed lines show roof structure when 
 first exposed; original roof was at top of the coalbed. 
 
 Small-sized slickensldes (less than 3 
 ft or 0.9 m) , nearly always found in mine 
 roof , are adequately supported by the 
 normal roof support plan at each mine. 
 However, large slickensldes should be re- 
 garded with extreme caution, and some 
 form of supplementary support used. In 
 practice, this commonly consists of ex- 
 tending the support area of the roof 
 bolts by using straps or wood headers , 
 installing additional bolts, or angle 
 bolting to more likely penetrate the 
 slickenside at a right angle and avoid 
 cantilevered segments of roof rock (fig. 
 4). In situations where heavy concentra- 
 tions of slickensldes are encountered, it 
 is advisable to consider using posts and 
 crossbars, full column resin bolts, or, 
 in severe cases, roof trusses. 
 
 There is currently no practical method 
 of detecting slickensldes in advance of 
 mining, although, if core drilling in- 
 dicates that the immediate roof con- 
 sists of a nonlaminated claystone, then 
 slickensldes almost certainly will be 
 present. 
 
 CLAY VEINS 
 
 Clay veins, or more properly, claystone 
 dikes, are wedge-shaped masses of slick- 
 ensided claystone or mudstone filling a 
 crevice in a coalbed (fig. 5). These 
 generally range up to 6 ft (2 m) in width 
 and persist for sometimes hundreds of 
 feet in length. 
 
10 
 
 :.Underclay 
 
 FIGURE 4. - Angle bolting of slickensides. 
 
 Clay veins are particularly prevalent 
 and troublesome in the Pittsburgh Coalbed 
 near Wheeling, WV, but occur sporadically 
 throughout the Appalachian region and in 
 many other coalbeds. Sometimes the term 
 "spar" is applied to a very narrow clay 
 vein or one that extends downward only a 
 few inches into the coalbed. Other types 
 of clay veins, such as steeply dipping, 
 narrow, claystone-f illed joints, occur in 
 the Appalachian region but are rare and 
 seldom cause roof problems. 
 
 The weakening effect of clay veins on 
 mine roof extends from 3 to 12 ft (0.9 to 
 3.7 m) above the top of the coalbed. In 
 addition to forming a discontinuity in 
 the coalbed and immediate roof , the clay 
 vein and surrounding rock are heavily 
 slickensided and therefore likely to 
 
 FIGURE 5. - Clay vein in entry rib and roof. 
 
11 
 
 break away from the roof in thick blocks 
 as the supporting coalbed is mined. The 
 slickensides along the margins and in- 
 terior of a clay vein tend to be aligned 
 parallel to the trend of the vein and to 
 dip toward the center of the vein forming 
 a troughlike effect. While prompt bolt- 
 ing, blocking, and strapping prevent im- 
 mediate falls of roof, the clay-rich rock 
 is susceptible to slaking and gradually 
 disintegrates, falling little by little. 
 
 Clay veins have rarely, if ever, been 
 detected by exploratory core drilling be- 
 cause of their narrow width. They are 
 particularly abundant and should be ex- 
 pected where the immediate roof consists 
 of a thick, clay-rich rock, although they 
 also occur immediately beneath limestone, 
 sandstone, or shale strata where perhaps 
 the clay-rich layer has been removed by 
 erosion. Clay veins can be predicted in 
 
 advance of mining only by underground 
 mapping and judicial projection along 
 their trends. 
 
 CREVASSE SPLAYS 
 
 The term "crevasse splay," used here in 
 a nongenetic, descriptive sense, desig- 
 nates a lithologic unit consisting of 
 sandstone thinly interbedded with shale, 
 or thin-bedded, laminated, micaceous 
 sandstone (miners' "stackrock") . The 
 unit ranges from 6 to 30 ft (2 to 9.3 m) 
 in thickness and persists laterally as a 
 sheetlike or lenticular body (fig. 6). 
 
 The splay type deposit that is sig- 
 nificant in mine roof stability is the 
 so-called low splay, where a predominant- 
 ly flat-bedded, laminated sandstone unit 
 lies within 6 to 10 ft (2 to 3 m) of the 
 top of the coalbed and, therefore, 
 
 FIGURE 6. ■ Example of splay type deposit in mine roof. 
 
12 
 
 constitutes part of the inunediate roof. 
 Splays that occur much higher than 10 ft 
 (3m) above the coalbed do not directly 
 affect roof conditions. 
 
 Roof falls in splay sequences generally 
 occur when a separation along a bedding 
 plane occurs at or above the horizon of 
 the roof bolt anchors, and the bolts are 
 unable to maintain the integrity of the 
 rock spanning the immediate roof. Some- 
 times failure can be prevented by stag- 
 gering the lengths of roof bolts so 
 that tensile stresses are not concen- 
 trated along one horizon. Full-column 
 resin bolts resist slippage along bedding 
 planes and thus reduce roof sag, which 
 might reduce the possibility of failure. 
 
 A low splay deposit can be inferred 
 from exploratory drill core information 
 that indicates a laterally widespread, 
 thin-bedded, laminated sandstone unit 
 lying within 6 to 10 ft (2 to 3 m) of 
 the coalbed, generally with some shale 
 above and below and situated adjacent to 
 channel-fill formations. 
 
 JOINTS 
 
 The term "joint" is used here to desig- 
 nate a nearly vertical, planar fracture 
 in rock, as distinguished from a slicken- 
 side, which is a curved, grooved surface 
 along which some movement has occurred. 
 In the central Appalachian coal region, 
 where coal measure strata are nearly 
 horizontal, joints in mine roof rock 
 usually are intraf ormational and of minor 
 
 significance to roof stability. Roof 
 falls may be terminated by a joint plane, 
 but there is little evidence to indicate 
 that the presence of a joint directly 
 contributes to a fall. 
 
 The situation is different along the 
 eastern margin of the Appalachian coal 
 region where the strata have been de- 
 formed by folding and faulting. Here, 
 jointing is more highly developed and 
 significantly contributes to roof prob- 
 lems, especially where the joints occur 
 as closely spaced, parallel sets or 
 groups. Highly jointed roof sometimes 
 can be supported satisfactorily with 
 bolted channel, strap, or mat, but severe 
 conditions may require trusses or posts 
 and crossbars. 
 
 Highly jointed roof is very difficult 
 to predict with certainty, but some de- 
 gree of success has been claimed by peo- 
 ple in the field, using interpretation 
 of high-altitude aerial photography and 
 satellite imagery, with the assumption 
 that at least some photolineaments repre- 
 sent rock joint sets. It is essential to 
 conduct a critical field check to deter- 
 mine the so-called ground truth of a lin- 
 ear lest cultural features be mistaken 
 for geologic structures. In some areas, 
 linear images are quite abundant , and the 
 problem is to distinguish between those 
 that are significantly related to roof 
 conditions or joints and those that are 
 not — a task for the most skilled and con- 
 scientious technologist. 
 
 DISCUSSION 
 
 There are a number of reasons for mine 
 roof failure. The cause of some roof 
 falls is obscure and attributed to poorly 
 understood stress concentrations sur- 
 rounding the mine openings; these falls 
 occur most commonly when mine entries are 
 located beneath stream valleys, and topo- 
 graphic relief is 200 ft (62 m) or more. 
 Some falls result from poor roof-bolting 
 techniques or improper installation of 
 bolts. However, most roof falls associ- 
 ated with the geologic hazards referred 
 to in this paper occur because unusual 
 
 geologic structures are not recognized or 
 anticipated and adequate support is not 
 provided. 
 
 The identification of the structures 
 that have been described here will re- 
 quire either the services of a geologist 
 with underground experience or the train- 
 ing of miners (especially mining machine 
 and roof -bolting machine operators) and 
 their supervisors. Through regular ex- 
 amination of mine roof structures and 
 conditions and the recording of this 
 
13 
 
 information systematically on the operat- 
 ing maps , and through some trial and er- 
 ror, it should become possible to deter- 
 mine the trends of hazardous structures 
 so that mine entry projections can be re- 
 vised or potentially troublesome zones 
 anticipated, roof support practices up- 
 graded, and roof falls reduced. 
 
 The task will not be easy because of 
 the complex character of many geologic 
 features, but, in consideration of both 
 
 the high priority of accident prevention 
 and the immense financial investment at 
 stake in developing a mine, the attempt 
 to identify and properly support roof 
 structures should become an integral part 
 of every underground operation. 
 
 The application of geologic methods to 
 other mines and quarry operations should 
 provide similar opportunities to improve 
 ground control practices and thus reduce 
 the inherent safety hazards. 
 
 BIBLIOGRAPHY 
 
 Alison, D. R. , E. T. Ohlsson, and K. V. 
 Whitney. Geologic and Engineering Data 
 Acquisition for Underground Coal Mine 
 Ground Control (contract J0395010, Arthur 
 D. Little, Inc.). BuMines OFR 89-80, 
 1980, 98 pp.; NTIS PB 80-219272. 
 
 Chase, F. E., and G. P. Sames. Kettle- 
 bottoms: Their Relation to Mine Roof and 
 Support. BuMines RI 8785, 1983, 12 pp. 
 
 Cummings , R. A., M. M. Singh, and N. N. 
 Moebs. Effect of Atmospheric Moisture on 
 the Deterioration of Coal Mine Roof 
 Shales. Min. Eng. (N.Y.), v. 35, No. 3, 
 Mar. 1983, pp. 243-245. 
 
 Cummings, R. A., M. M. Singh, S. E. 
 Sharp, and A. W. Laurito. Control of 
 Shale Roof Deterioration With Air Temper- 
 ing (contract J0188028, Engineers Inter- 
 national, Inc.). Volume 1 — Field and 
 Laboratory Investigations. BuMines OFR 
 41(l)-82, 1981, 162 pp. 
 
 Ellenberger, J. L. Slickenside Occur- 
 rence in Coal Mine Roof of the Valley 
 Camp No. 3 Mine Near Wheeling, W. Va. 
 BuMines RI 8365, 1979, 17 pp. 
 
 Ferm, J. C. , R. A. Melton, G. D. Cum- 
 mins, D. Mathew, L. L. McKenna, C. Muir, 
 and G. E. Norris. A Study of Roof Falls 
 in Underground Mines on the Pocahontas 
 No. 3 Seam, Southern West Virginia and 
 Southwestern Virginia (contract H0230028, 
 Univ. SC). BuMines OFR 36-80, 1978, 83 
 pp.; NTIS PB 80-158983. 
 
 Hylbert, D. K. Delineation of Geologic 
 Roof Hazards in Selected Coal Beds in 
 Eastern Kentucky, With Landsat Imagery 
 Studies in Eastern Kentucky and the 
 Dunkard Basin (contract J0188002, More- 
 head State Univ.). BuMines OFR 166-81, 
 1980, 97 pp.; NTIS PB 82-140336. 
 
 . Developing Geological Struc- 
 tural Criteria for Predicting Unstable 
 Mine Roof Rocks (contract H0133018, More- 
 head State Univ.). BuMines OFR 9-78, 
 1977, 246 pp.; NTIS PB 276 735. 
 
 Jansky, J. H. , and R. F. Valane. Cor- 
 relation of LANDSAT and Air Photo Linears 
 With Roof Control Problems and Geologic 
 Features. BuMines RI 8777, 1983, 22 pp. 
 
 McCulloch, C. M. , P. W. Jeran, and 
 C. D. Sullivan. Geologic Investigations 
 of Underground Coal Mining Problems. Bu- 
 Mines RI 8022, 1975, 30 pp. 
 
 Moebs, N. N. The Geological Character 
 of Some Coal Wants at the Westland Mine 
 in Southwestern Pennsylvania. BuMines 
 RI 8555, 1980, 25 pp. 
 
 . Geologic Guidelines in Coal 
 
 Mine Design. Paper in Ground-Control As- 
 pects of Coal Mine Design. Proceedings: 
 Bureau of Mines Technology Transfer Sem- 
 inar, Lexington, KY, March 6, 1973. Bu- 
 Mines IC 8630, 1974, pp. 63-69. 
 
14 
 
 Moebs, N. N. Roof Rock Structures and 
 Related Roof Support Problems in the 
 Pittsburgh Coalbed of Southwestern Penn- 
 sylvania. BuMines RI 8230, 1977, 30 pp. 
 
 . Subsidence Over Four Room-and- 
 
 Pillar Sections in Southwestern Pennsyl- 
 vania. BuMines RI 8645, 1982, 23 pp. 
 
 Moebs, N. N. , and E. A. Curth. Geolog- 
 ic and Ground Control Aspects of an Ex- 
 perimental Shortwall Operation in the Up- 
 per Ohio Valley. BuMines RI 8112, 1976, 
 30 pp. 
 
 Moebs, N. N. , and J. L. Ellenberger. 
 Geologic Structures in Coal Mine Roof. 
 BuMines RI 8620, 1982, 16 pp. 
 
 . Hazardous Roof Structures in 
 
 Appalachian Coal Mines. Ch. in Ground 
 
 Control in Room-and-Pillar Mining. AIME, 
 New York, 1982, pp. 9-16. 
 
 Moebs, N. N. , and J. C. Perm. The Re- 
 lation of Geology to Mine Roof Conditions 
 in the Pocahontas No. 3 Coalbed. BuMines 
 IC 8864, 1982, 8 pp. 
 
 Overbey, W. K. , Jr., C. A. Komar, and 
 J. Pasini III. Predicting Probable Roof 
 Fall Areas in Advance of Mining by Geo- 
 logical Analysis. BuMines TPR 70, 1973, 
 17 pp. 
 
 Stingelin, R. W. , J. R. Kern, and 
 S. L. Morgan. Pre-Mining Identification 
 of Hazards Associated With Coal Mine Roof 
 Measures (contract J0177038, HRB-Singer, 
 Inc.). BuMines OFR 167-81, 1979, 206 
 pp.; NTIS PB 82-140344. 
 
15 
 
 COAL AND ROCK PROPERTIES FOR PREMINE PLANNING AND MINE DESIGN 
 By Richard E. Thill'' 
 
 ABSTRACT 
 
 Nearly all phases of the mining opera- 
 tion require input on the engineering 
 properties of rock. Engineering proper- 
 ties are particularly useful in the pre- 
 mine investigation and mine design phases 
 of the mining operation. 
 
 Because of the general scarcity of en- 
 gineering property data for coal measure 
 strata in major U.S. coal districts and 
 the lack of organized data bases to house 
 and disseminate such information, the 
 
 Bureau of Mines undertook a wide-ranging 
 testing program to determine engineering 
 properties of coal measure rocks in sev- 
 eral U.S. coal basins. The Bureau also 
 initiated the development of a data base 
 for the housing, sorting, manipulation, 
 and retrieval of property data by com- 
 puter. Typical results are given for 
 both field and laboratory determinations 
 of geotechnical, mechanical, and geophys- 
 ical properties. 
 
 INTRODUCTION 
 
 Engineering properties are important in 
 nearly every phase of mining, but partic- 
 ularly in mine design, layout, and ground 
 control. Reviews of current literature 
 revealed that (1) not many data exist for 
 the engineering properties of strata as- 
 sociated with U.S. coal deposits, (2) 
 data available are scattered in the lit- 
 erature and difficult to retrieve, and 
 (3) very few data exist in data sets that 
 permit correlations to be made between 
 properties for predictive purposes. 
 
 Because of the general lack of data 
 in major U.S. coal-producing districts, 
 the Bureau of Mines undertook a testing 
 program to determine a variety of engi- 
 neering properties, including sets of 
 
 geological, geophysical, physical, and 
 mechanical properties, in several major 
 U.S. coalfields. In addition, data from 
 earlier Bureau studies on all types of 
 crystalline and sedimentary rock were 
 summarized in data tables and incorpo- 
 rated into a rock property data base. 
 The testing program and typical examples 
 of data and analyses for the coal measure 
 rocks, and a brief description of the 
 mechanical properties data base and data 
 tables are presented in this paper. The 
 property test results are being published 
 in Bureau and outside publications to 
 assist mine designers and planners in the 
 design of openings and support struc- 
 tures, and in the selection of appropri- 
 ate excavation and support equipment. 
 
 COAL MEASURE ROCK TEST PROGRAM 
 
 The testing program consisted of both 
 laboratory and field investigations on 
 rock core taken at several field sites 
 
 'Supervisory geophysicist, Twin Cities 
 Research Center, Bureau of Mines, Minne- 
 apolis, MN. 
 
 (fig. 1). In the field, lithologic and 
 geotechnical descriptions of core were 
 made and downhole geophysical logging 
 was performed. Core then was boxed 
 and shipped to the Bureau for property 
 testing. Portions of the tests in Illi- 
 nois rock were conducted under contract. 
 
16 
 
 PROGRAM ORGANIZATION 
 
 In situ investigations 
 
 Core 
 extraction 
 
 Lithologic 
 logging 
 
 Geotechnical 
 logging 
 
 Geophysical 
 logging 
 
 Specimen selection 
 and preparation 
 
 Laboratory investigations 
 
 Geological 
 properties 
 
 Acoustic 
 properties 
 
 Electrical 
 properties 
 
 Mectionical 
 properties 
 
 Physical and index 
 properties 
 
 Data reduction 
 and analyses 
 
 FIGURE 1. - Cool measure rock testing program. 
 
 In situ properties determined included 
 compressional (P) and shear (S) wave 
 velocities, formation density, derived 
 elastic moduli, and other geotechnical 
 properties. Laboratory determinations 
 included mechanical properties such as 
 uniaxial, triaxial, and tensile strength, 
 and static elastic properties; acous- 
 tic properties such as P- and S-wave 
 velocities, dynamic elastic moduli, and 
 acoustic impedance and attenuation; 
 petrographic properties; electrical prop- 
 erties such as dielectric constant, dis- 
 sipation factor, conductivity, and skin 
 depth; and physical or index properties 
 
 such as bulk density, porosity, permea- 
 bility, shore hardness, slake durability, 
 and point-load strength index. 
 
 MINESITES 
 
 The several minesites investigated in- 
 cluded the Gateway Mine in Pennsylvania, 
 the York Canyon Mine in New Mexico, and 
 three sites in the Illinois Basin. Ta- 
 ble 1 indicates the types of studies un- 
 dertaken at each site. The Gateway and 
 York Canyon studies were more extensive 
 and comprehensive than those in the 
 Illinois Basin. 
 
17 
 
 TABLE 1. - Properties determined at different sites 
 
 Property test or 
 
 Gateway Mine, 
 
 southwestern 
 
 Pennsylvania 
 
 York Canyon 
 
 Mine, northern 
 
 New Mexico 
 
 Illinois Basin 
 
 characteristic 
 
 East-central 
 Illinois 
 
 Southern Illinois 
 and Indiana 
 
 In situ: 
 
 Lithologic 
 
 Geotechnical 
 
 Geophysical 
 
 Laboratory: 
 
 Petrographic 
 
 Mechanical. ........ 
 
 X 
 X 
 X 
 
 X 
 X 
 X 
 X 
 X 
 
 X 
 
 X 
 
 X 
 X 
 X 
 
 X 
 X 
 X 
 
 X 
 
 X 
 
 X 
 
 X 
 X 
 X 
 
 X 
 X 
 X 
 
 X 
 
 X 
 
 X 
 
 X 
 X 
 
 Acoustical 
 
 Electrical 
 
 X 
 
 Physical and index. 
 
 Environmental 
 effects: 
 Stress ............. 
 
 X 
 X 
 
 Moisture 
 
 X 
 
 The Gateway Mine is located in the 
 northern part of the Eastern Coal Prov- 
 ince (fig. 2). Core was taken with a 
 
 conventional drill rig equipped with a 
 20-ft-long, wire line retrievable, inner 
 core barrel (fig. 3). Roughly 500 ft of 
 
 100 
 
 Scale, mi 
 
 200 
 
 wv 
 
 LEGEND Scale, mi 
 
 Underlain by Pittsburgh Coalbed 
 
 10 
 
 h! '. ,' H Gateway mines 
 <-\-^ Anticline 
 -H^ Syncline 
 
 FIGURE 2, - Location of Gateway minesite. 
 
18 
 
 FIGURE 3. - Drill rig at the Gateway site. 
 
19 
 
 FIGURE 4. - Core extraction at the Gateway site. 
 
 core was taken at the site (fig. 4) 
 and logged for lithology, percent re- 
 covery and rock quality (RQD). Three- 
 dimensional acoustic and gamma-gamma den- 
 sity surveys also were conducted in the 
 borehole. 
 
 The York Canyon site is in the north- 
 eastern portion of the Southwest Coal 
 Province (fig. 5). Figure 6 shows the 
 drill rig used at the site. As core was 
 taken from the hole, natural moisture 
 content was preserved by wrapping the 
 core in tinfoil and sealing it with wax 
 (fig. 7). 
 
 The localities of the Illinois Coal 
 Basin sites are shown in figure 8. 
 The Danville, IL, site is located in 
 the east-central portion and the other 
 two sites are located in the south- 
 east portion of the Interior Coal 
 Province. 
 
 FIGURE 5. - Location of York Canyon Mine. 
 
20 
 
 FIGURE 7. = Sealing moisture content in the 
 York Canyon core. 
 
 INDIANA 
 
 FIGURE 6, - Drill rig at the York Canyon site. 
 
 LABORATORY TESTING 
 
 Test specimens were prepared from the 
 core arriving from the field. Laboratory 
 property determinations included petro- 
 graphic, acoustic, electrical, mechani- 
 cal, physical, and index properties. 
 Specimen preparation and testing was con- 
 ducted in accordance with standardized 
 procedures described in the Bureau of 
 Mines Test Procedures for Rocks {l)^ and/ 
 or as recommended by the International 
 
 ^Underlined numbers in parentheses re- " 
 fer to items in the list of references at 
 the end of this paper. 
 
 FIGURE 8. - Location of Illinois sites. 
 
21 
 
 FIGURE 9. - Mechanical property testing apparatus. 
 
 Society of Rock Mechanics (2^) . The me- 
 chanical property testing facilities in- 
 clude two hydraulic, servo-controlled 
 testing machines, one of 200,000-lb and 
 the other of 500,000-lb load capacity 
 
 (fig. 9). For triaxial confinement test- 
 ing, separate servo controls are used 
 for applying the uniaxial and confining 
 pressure. 
 
 RESULTS 
 
 FIELD GEOTECHNICAL AND 
 GEOPHYSICAL PROPERTIES 
 
 As core was taken in the field, litho- 
 logic descriptions were made to construct 
 stratigraphic columns for the vicinity of 
 the working coal seam and the overburden. 
 As indicators of rock quality, percent 
 recovery, RQD, and fracture frequency 
 were measured. Figure 10 represents re- 
 sults obtained at the Gateway minesite. 
 Poor quality rock is designated by zones 
 having low RQD, low core recovery, or 
 high fracture frequency. 
 
 Borehole geophysical logs were run at 
 the Gateway, York Canyon, and Danville 
 sites. The logs obtained at the Gateway 
 site (fig. 11), for instance, permit 
 comparison between elastic wave velocity 
 and density response. Coal seams and 
 fracture zones are readily detected by 
 anomalously low-velocity, low-density re- 
 sponse. From the full waveform velocity 
 log, both P- and S-wave velocities can be 
 determined, and when combined with den- 
 sity, these permit calculation of the dy- 
 namic elastic moduli, indicating stress- 
 strain response and, by correlation, the 
 
22 
 
 LITHOLOGIC 
 LOG 
 EL -7 
 
 DRILLER'S 
 LOG 
 EL-7 
 
 Or 
 
 SO- 
 
 SO- 100- 
 
 150 - 
 
 60 - 200 - 
 
 90 
 
 120 
 
 1501- 
 
 250 
 
 300 
 
 350 - 
 
 400 
 
 450 
 
 500^ 
 
 Coal 
 
 Coal 
 
 Coal 
 
 Waynesburg 
 cool 
 
 Coal 
 
 LA. .li. .E- -UT^ 
 
 Sewickley 
 coal 
 
 Pittsburgh 
 cool 
 
 
 ROD, 
 
 pet 
 
 40 60 80 100 
 
 -\ r 
 
 CORE 
 
 RECOVERY, 
 
 pet 
 
 94 96 98 100 
 
 Coal 
 
 Coal 
 
 Waynesburg 
 coal 
 
 Sewickley 
 deool 
 
 Pittsburgh 
 eoal 
 
 FRACTURE 
 
 FREQUENCY, 
 
 per meter 
 
 2 4 6 
 
 FIGURE 10. - Geotechnical logs for the Gateway site. 
 
 strength of various strata in the over- 
 burden. The geophysical data, in addi- 
 tion, are useful for interpreting strati- 
 graphic boundaries and changes O) . 
 
 LABORATORY GEOPHYSICAL PROPERTIES 
 
 Geophysical properties of coal measure 
 strata, such as elastic wave velocities 
 and electrical properties, are not well 
 documented in the literature but are re- 
 quired for the design of geophysical 
 probes and interpretation of field geo- 
 physical data. Hence, laboratory tests 
 were conducted for these properties. 
 Acoustic core logging was conducted on 
 the core as it arrived from the field, 
 before specimens were prepared for other 
 laboratory tests. Measurements of P-wave 
 traveltimes were made in different di- 
 rections parallel to the bedding at 0.5- 
 to 1-ft depth increments along the core. 
 
 Velocity averages for the plane and ve- 
 locity difference expressed in terms of 
 anisotropy were then plotted as functions 
 of depth and compared with the density 
 and lithologic logs (fig. 12). As in 
 field results, combinations of low veloc- 
 ity and density normally signify coal 
 seams and fractured zones . High values 
 of velocity anisotropy usually suggest 
 fracturing and poor quality of rock (4^) . 
 
 Although electrical properties were de- 
 termined only in the Gateway strata for 
 the dry condition, hundreds of tests were 
 conducted in various categories of coal 
 measure rocks over frequency ranges from 
 1 kHz to 100 MHz. Predicted propagation 
 distance (skin depth) for electromagnetic 
 energy over the frequency range from 1 
 kHz to 20 MHz is indicated for the com- 
 posite rock types in figure 13. 
 
23 
 
 03 
 
 ■>. 
 
 a 
 
 a> 
 a 
 O 
 
 I/) 
 en 
 O 
 
 I/) 
 
 Q. 
 O 
 0) 
 
 O 
 
 LU 
 
 o 
 
24 
 
 DENSITY, g*m' 
 
 DIAMETRAL P-WAVE VELOCITY , km/s 
 
 VELOCITY ANISOTROPY, pet 
 
 20 30 00 50 
 
 FIGURE 12. - Laboratory wave velocity and density profiles. 
 
 5 I02- 
 
 10' 102 10= 
 
 FREQUENCY, kHz 
 
 FIGURE 13, - Skin depth as a function of frequency, 
 
 MECHANICAL PROPERTIES 
 
 Mechanical properties were determined 
 in uniaxial and triaxial compression 
 tests using the closed-loop, servo- 
 controlled testing machines and confin- 
 ing pressure cell. Resultant data were 
 plotted as axial, circumferential, and 
 
 volumetric stress-strain curves (fig. 
 14). Apparent Poisson's ratio, that is, 
 the ratio of lateral to axial strain, was 
 also determined throughout the range of 
 stress. Concurrent with uniaxial com- 
 pression testing, the change in P-wave 
 velocity as a function of stress was 
 determined (fig. 15). The effect of 
 confinement on shear strength was deter- 
 mined using conventional Mohr-Coulomb 
 plots and analysis (fig. 16). Although 
 stress-strain behavior in most of the 
 coal measure rocks exhibited nonlinear 
 behavior, the linear Coulomb failure cri- 
 teria appeared to provide a good first 
 approximation to shear strength for most 
 of the rock types tested. Confining 
 pressures ranged up to 2,500 psi (17.25 
 MPa) , simulating expected lithostatic 
 pressures to depths of roughly 2,750 ft. 
 Typical triaxial test results are given 
 in table 2. Enhancement of compressive 
 strength with confinement is indicated in 
 figure 17. More detailed analyses, pre- 
 dicting failure by more complicated fail- 
 ure criteria, are still underway. In- 
 direct tensile strength also was deter- 
 mined using the Brazilian test method. 
 

 cu ' 
 
 
 
 
 
 
 
 
 
 
 
 
 
 0) (J 4-1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 bO 3 C 
 
 00 
 
 in 
 
 O 
 
 f—^ 
 
 CO 
 
 ON 
 
 CM 
 
 o 
 
 CO 
 
 -* 
 
 
 
 CO 4-1 0) 4-1 
 
 o 
 
 f-H 
 
 vo 
 
 t — 1 
 
 -* 
 
 ~d- 
 
 -a- 
 
 r-- 
 
 CO 
 
 
 
 
 l-l CO 4-1 O 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 
 
 
 • 
 
 
 
 Q> -H C CX 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 > O O 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 
 <: S O 
 
 
 
 
 
 
 
 
 
 
 
 CO 
 4J 
 O 
 
 
 (U 
 
 
 
 
 
 
 
 
 
 
 
 
 1-1 ' 
 
 
 
 
 
 
 
 
 
 
 
 iH 
 
 
 C 3 (U 
 
 
 
 
 
 
 
 
 
 
 
 Q, 
 
 
 cO 4-1 fH bO 
 
 o 
 
 . — 1 
 
 csl 
 
 00 
 
 00 
 
 r-~ 
 
 o 
 
 00 
 
 -* 
 
 r^ 
 
 
 
 0) O bO (U 
 
 CO 
 
 CO 
 
 CO 
 
 CM 
 
 OvI 
 
 CM 
 
 CO 
 
 CM 
 
 og 
 
 CNl 
 
 0) 
 
 
 S CO c -a 
 
 
 
 
 
 
 
 
 
 
 
 r-l 
 
 
 l-l CO 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 
 14-1 
 
 
 
 
 
 
 
 
 
 
 
 u 
 
 
 
 
 
 
 
 
 
 
 
 
 
 •H 
 
 
 3 4-1 
 
 
 
 
 
 
 
 
 
 
 
 u 
 
 
 O C 
 
 
 
 
 
 
 
 
 
 
 
 
 
 T^ (U 
 
 
 
 
 
 
 
 
 
 
 
 u 
 
 
 4-1 -H 
 
 vD 
 
 <3^ 
 
 VD 
 
 ON 
 
 t-H 
 
 r-. 
 
 CO 
 
 <JN 
 
 p~ 
 
 eg 
 
 J= 
 
 
 CO O 
 
 <T\ 
 
 a\ 
 
 r-~ 
 
 ON 
 
 ON 
 
 CO 
 
 in 
 
 CO 
 
 r-> 
 
 ON 
 
 o 
 
 
 iH -H 
 
 o\ 
 
 ON 
 
 ON 
 
 <JN 
 
 CJN 
 
 CJN 
 
 ON 
 
 <7N 
 
 CJN 
 
 CJN 
 
 s 
 
 
 (U 4-1 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 
 
 
 l-l 4-1 
 
 o 
 
 
 
 
 
 
 
 
 
 
 0) 
 
 
 l-l 0) 
 
 
 
 
 
 
 
 
 
 
 
 > 
 
 
 O O 
 
 
 
 
 
 
 
 
 
 
 
 iH 
 
 
 CJ> O 
 
 
 
 
 
 
 
 
 
 
 
 u 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 A 
 
 
 
 
 
 
 
 
 
 
 
 13 
 
 
 <4-i .H n 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O CO o 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 3 -H bO 
 
 
 
 
 
 
 
 
 
 
 
 4-1 
 
 
 0) l-l 4J (1) 
 
 00 
 
 o 
 
 r^ 
 
 CO 
 
 t-H 
 
 <t 
 
 o 
 
 00 
 
 CO 
 
 00 
 
 
 
 iH 0) a T3 
 
 CNJ 
 
 CO 
 
 CM 
 
 CO 
 
 CO 
 
 CM 
 
 -* 
 
 CO 
 
 -* 
 
 CO 
 
 •a 
 
 
 bO 4-1 T^ 
 
 
 
 
 
 
 
 
 
 
 
 <u 
 
 CO 
 
 3 3 l-l 
 < 1-1 IM 
 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 4J 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 rH 
 
 3 
 
 
 
 
 
 
 
 
 
 
 
 
 
 4-1 M 
 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 03 
 
 3 CO 
 
 
 
 
 
 
 
 
 
 
 
 4J 
 
 0) 
 
 <U 3 3 
 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 M 
 
 •H iJ O 
 
 
 
 
 
 
 
 
 
 
 
 -3 
 
 
 a (1) -H 
 
 CN 
 
 00 
 
 o 
 
 vi- 
 
 o 
 
 -<3- 
 
 CO 
 
 00 
 
 CM 
 
 t~~ 
 
 
 J-l 
 
 •H 4J 4J 
 
 -* 
 
 m 
 
 in 
 
 vo 
 
 vO 
 
 o- 
 
 00 
 
 r^ 
 
 CJN 
 
 1^ 
 
 4J 
 
 CO 
 
 iM 3 O 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 B 
 
 • 
 
 • 
 
 9 
 
 • 
 
 CO 
 
 Q) 
 
 14-1 -H -H 
 
 o 
 
 
 
 
 
 
 
 
 
 
 (U 
 
 4-1 
 
 (U l-l 
 O <4-l 4-1 
 
 
 
 
 
 
 
 
 
 
 
 4-1 
 
 iH 
 
 o o 
 
 
 
 
 
 
 
 
 
 
 
 (U 
 
 CO 
 •H 
 
 
 
 
 
 
 
 
 
 
 
 
 1-) 
 
 > 
 
 
 
 
 
 
 
 
 
 
 
 X 
 
 4-> 
 
 
 
 
 
 
 
 
 
 
 
 09 
 
 CO 
 
 3 3 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 •H 
 
 O (U 
 
 
 
 
 
 
 
 
 
 
 
 0) 
 
 U 
 
 •H -H 
 
 u-1 
 
 in 
 
 o 
 
 o 
 
 o 
 
 o 
 
 in 
 
 o 
 
 o 
 
 m 
 
 4-1 
 
 4-1 
 
 CO O CO 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 9 
 
 • 
 
 
 
 a) -H CM 
 
 CM 
 
 00 
 
 vD 
 
 r-^ 
 
 r-~ 
 
 <f 
 
 00 
 
 r-- 
 
 r^ 
 
 i-H 
 
 3 
 
 M-l 
 
 x: 4-1 S 
 
 f-H 
 
 
 ^H 
 
 .— ) 
 
 t — 1 
 
 t-H 
 
 t-H 
 
 t-H 
 
 t-H 
 
 t-H 
 
 CO 
 
 O 
 
 O 4-1 
 O 0) 
 
 
 
 
 
 
 
 
 
 
 
 •H 
 iH 
 
 >> 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 1-1 
 
 u 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 N 
 
 
 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 4-1 
 
 
 
 
 
 
 
 
 
 
 
 U 
 
 s 
 
 O CO- 
 
 
 
 
 
 
 
 
 
 
 
 P3 
 
 3 
 
 (U "O 
 
 
 
 
 
 
 
 
 
 
 
 
 w 
 
 l-l ^ 1) 
 
 
 
 
 
 
 
 
 
 
 
 T3 
 
 
 (U a, 4-1 
 
 (T. 
 
 <-H 
 
 CO 
 
 1— 1 
 
 <f 
 
 r~ 
 
 CO 
 
 r^ 
 
 00 
 
 O 
 
 3 
 
 1 
 
 ^ d CO 
 
 in 
 
 f-H 
 
 i-H 
 
 CO 
 
 CM 
 
 -* 
 
 t-H 
 
 t-H 
 
 in 
 
 CO 
 
 CO 
 
 
 B CO 0) 
 
 
 
 
 
 
 
 
 
 
 
 
 • 
 
 3 CO 4-1 
 
 
 
 
 
 
 
 
 
 
 
 M 
 
 (M 
 
 IS 
 
 
 
 
 
 
 
 
 
 
 
 3 
 O 
 
 w 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 ij 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 to 
 
 ^ 
 
 CO 
 
 r-. 
 
 eo 
 
 Cvl 
 
 r^ 
 
 CO 
 
 -* 
 
 t-H 
 
 o 
 
 CO 
 
 vO 
 
 CO 
 
 H 
 
 •O 3 4J 
 
 9 
 
 • 
 
 • 
 
 • 
 
 • 
 
 
 • 
 
 
 • 
 
 • 
 
 p 
 
 
 <U ^ 4-1 
 
 ■vT 
 
 vO 
 
 CN 
 
 ^ 
 
 -0- 
 
 r^ 
 
 o 
 
 CM 
 
 vO 
 
 t-H 
 
 <x 
 
 
 P9 CJ 
 
 CNJ 
 
 
 t-H 
 
 CSl 
 
 rH 
 
 CM 
 
 t-H 
 
 t-H 
 
 CM 
 
 I-H 
 
 
 
 
 •H 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 X: 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 4-1 
 
 
 
 
 
 
 
 
 
 
 
 l-l 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 in 
 
 <T\ 
 
 in 
 
 CM 
 
 <3- 
 
 00 
 
 m 
 
 r~~ 
 
 v£> 
 
 CO 
 
 
 A 
 
 • 
 
 • 
 
 ■ 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 •H 
 
 
 i-l 
 
 o 
 
 CJN 
 
 CO 
 
 CO 
 
 CM 
 
 vO 
 
 00 
 
 00 
 
 vO 
 
 in 
 
 >4 
 
 
 ^ CO 
 
 O) 
 
 m 
 
 m 
 
 t^ 
 
 <»■ 
 
 t-H 
 
 ^ 
 
 r^ 
 
 CO 
 
 o 
 
 CO 
 
 
 4-1 > 
 
 f-H 
 
 t-H 
 
 CNj 
 
 -* 
 
 
 CM 
 
 -* 
 
 CO 
 
 CO 
 
 <r 
 
 1-1 
 
 
 O. H 4-1 
 
 1 
 
 1 
 
 1 
 
 1 
 
 1 
 
 1 
 
 1 
 
 1 
 
 1 
 
 1 
 
 u 
 
 
 0) 0) 4-1 
 
 ro 
 
 CO 
 
 t-~ 
 
 00 
 
 ON 
 
 O 
 
 r~. 
 
 in 
 
 v£> 
 
 o 
 
 4-1 
 
 
 n 4-1 
 
 • 
 
 • 
 
 • 
 
 • 
 
 « 
 
 • 
 
 • 
 
 • 
 
 • 
 
 • 
 
 
 
 3 
 
 in 
 
 CO 
 
 t-H 
 
 00 
 
 r-- 
 
 ON 
 
 CO 
 
 NO 
 
 o 
 
 -* 
 
 t3 
 
 
 •H 
 
 (J^ 
 
 in 
 
 o- 
 
 <!• 
 
 Csl 
 
 00 
 
 CO 
 
 VO 
 
 f—i 
 
 ON 
 
 3 
 
 
 
 
 '~' 
 
 CM 
 
 -* 
 
 t-H 
 
 t-H 
 
 -^ 
 
 CO 
 
 CO 
 
 CO 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 
 
 i-l 
 
 
 
 
 
 
 
 
 
 • 
 
 
 • 
 
 
 CO 
 
 
 
 
 
 
 
 
 1 • 
 
 • 
 
 
 • 
 
 
 iH 
 
 
 
 
 
 
 
 
 X) dJ 
 
 • 
 
 
 <u 
 
 
 X 
 
 
 
 
 
 
 
 
 3 >-i 
 
 0) 
 
 
 3 
 
 
 CO 
 
 
 O 3 
 
 
 
 
 
 
 CO to 
 
 3 
 
 
 O 
 
 
 1-1 
 
 
 •H O 
 
 
 
 
 
 
 CO ^ 
 
 o 
 
 
 4J 
 
 
 3 
 
 
 bO -H 
 
 
 
 
 
 
 CO 
 
 4J 
 
 
 CO 
 
 
 3 
 
 
 O 4-1 
 
 
 
 
 
 <u 
 
 TJ 
 
 CO 
 
 
 2 
 
 
 
 
 .-( a 
 
 
 
 
 
 iH 
 
 (U 'V 
 
 4J 
 
 
 s 
 
 
 CO 
 
 
 O -H 
 
 01 
 
 
 
 
 CO 
 
 TS 3 
 
 tH 
 
 
 1-1 
 
 0) 
 
 (U 
 
 
 J3 M 
 
 3 
 
 
 
 
 ,3 
 
 T3 CO 
 
 1-1 
 
 
 tH 
 
 3 
 
 CO 
 
 
 4J O 
 
 o 
 
 
 
 
 CO 
 
 0) 
 
 CO 
 
 
 
 o 
 
 •H 
 
 
 •H CO 
 
 4J 
 
 
 
 
 
 ^ 4) 
 
 
 
 >N 
 
 4J 
 
 M 
 
 
 I-] (U 
 
 CO 
 
 
 
 
 >. 
 
 U 3 
 
 >% 
 
 <u 
 
 0) 
 
 CO 
 
 & 
 
 
 •a 
 
 T3 
 
 o 
 
 O 
 
 o 
 
 T3 
 
 0) o 
 
 T) 
 
 r-t 
 
 rH 
 
 01 
 
 
 
 
 
 3 
 
 Q 
 
 Q 
 
 a 
 
 3 
 
 4J 4J 
 
 3 
 
 CO 
 
 CO 
 
 e 
 
 O 
 
 
 
 CO 
 
 
 
 
 CO 
 
 3 CO 
 
 CO 
 
 J3 
 
 Si 
 
 1-1 
 
 O 
 
 
 
 en 
 
 
 
 
 en 
 
 M 
 
 en 
 
 CO 
 
 en 
 
 kJ 
 
 ~ 
 
 25 
 
26 
 
 □ Lateral strain 
 A Volumetric strain 
 V Axial strain 
 O Apparent Poisson-s 
 ratio 
 
 4 6 
 
 
 
 MILLISTRAIN 
 
 FIGURE 14. - Volumetric stress-strain curve. 
 
 0.2 
 
 0.4 
 
 0,6 
 
 0.8 
 
 PHYSICAL AND INDEX PROPERTIES 
 
 Physical and index properties deter- 
 mined in the laboratory included shore 
 hardness, porosity, permeability, bulk 
 density, point-load strength index, and 
 slake durability. Moisture content was 
 determined in the specimens tested 
 for uniaxial and triaxial compression 
 strength. In certain cases, studies also 
 were made of the effect of moisture on 
 elastic and strength properties and on 
 the elastic wave velocities. 
 
 STATISTICAL ANALYSES 
 
 All rock property data were input into 
 a computer for further analyses. Statis- 
 tical analyses were conducted to obtain 
 the mean, standard deviation, and coeffi- 
 cient of variability for all measurements 
 in a particular rock type at each site. 
 
 Bar charts were constructed providing the 
 mean and range (one standard deviation) 
 of property values for each property 
 (figs. 18-19). Because of lithologic 
 variability within each rock category, 
 standard deviations frequently were quite 
 large. 
 
 Crossplots were constructed to deter- 
 mine correlations between rock properties 
 by regression analyses for each rock type 
 (fig. 20). Correlation coefficient ma- 
 trices were obtained to indicate the 
 degree of correlation between any two 
 properties. Those properties that ex- 
 hibit significant correlation with elas- 
 tic moduli and strength properties and 
 that could be easily and inexpensively 
 obtained during exploration investiga- 
 tions might, in some cases, be used as 
 index properties to predict modulus or 
 strength (fig. 21). 
 
27 
 
 VELOCITY vs STRESS 
 
 STRESS-STRAIN 
 
 T 1 r 
 
 Mudstone 
 
 EL-7 No 303 
 
 10 15 20 25 30 35 40 
 
 40 
 35 
 30 
 25 
 20 
 15 
 10 
 5 
 
 
 
 I 
 
 1 1 1 1 1 1 1 
 
 p 
 
 ~ 
 
 Mudstone ^ 
 
 - 
 
 - 
 
 y Z\.-l No. 303 
 
 - 
 
 ^-rr"- 
 
 "1 1 1 1 1 1 1 
 
 
 I 23456789 10 
 
 "T — 1 — r 
 
 1 — I — I — I — I — r 
 
 Shale 
 
 EL-7 No 125 B 
 
 I I I I ] I J I J J 
 
 10 20 30 40 50 60 70 80 90 100 110 
 
 120 
 100 
 80 
 60 
 40 
 20 
 
 
 
 1 
 
 1 1 1 
 
 1 1 1 1 1 1 
 
 r^- 
 
 - 
 
 Shale 
 
 ^^ 
 
 
 ^ 
 
 "i 1 1 
 
 EL-7 No. 125 B 
 1 1 1 1 1 1 
 
 - 
 
 I 2 3 4 5 6 7 8 9 10 II 12 
 
 4 - 
 
 3 - 
 
 1 
 Coarse 
 
 1 1 
 -grain 
 
 1 1 
 sandstone 
 
 — 
 
 1 
 
 ' 1 
 
 1 1 
 
 EL-7 
 
 1 1 
 
 No. 
 
 IIOA 
 
 1 
 
 80 
 70 
 60 
 50 
 40 
 30 
 20 
 10 
 
 1 1 1 1 1 1 1 1 1 
 
 ~~o- 
 
 Coarse-groin sondstone X 
 
 - 
 
 j^ EL-7 No. IIOA 
 
 _ 
 
 ^-°r^i 1 
 
 " 
 
 10 20 30 40 50 60 70 80 
 
 23456789 10 
 
 Shaley linnestone 
 
 EL-7 No 296 
 
 J I 1 I I I : L_ 
 
 I8U 
 
 1 
 
 I 
 
 
 1 1 I 1 
 
 p 1 
 
 160 
 
 - 
 
 
 
 y 
 
 / 
 
 140 
 120 
 
 - 
 
 Sh 
 
 oley 
 
 limestone /^ 
 
 - 
 
 100 
 
 - 
 
 
 
 y^ 
 
 - 
 
 80 
 
 - 
 
 
 
 y 
 
 - 
 
 60 
 40 
 
 - 
 
 
 
 y EL-7 No, 296 
 
 - 
 
 20 
 
 - 
 
 -°^i 
 
 ^ 
 
 1 1 1 1 
 
 - 
 
 
 
 
 1 
 
 
 2 3 4 5 
 
 MILLISTRAIN 
 
 e 
 
 20 40 60 80 100 120 140 160 180 
 UNIAXIAL STRESS, MPa 
 
 FIGURE 15. - P-wave velocity as a function of uniaxial stress. 
 
28 
 
 -20 
 
 20 
 
 Micaceous 
 sandstone 
 
 40 60 80 100 120 140 
 
 NORMAL STRESS, MPa 
 FIGURE 16. - Mohr-Coulomb plot for the Pittsburgh Sandstone. 
 
 180 200 220 
 
 250 
 
 o 
 
 CL 
 
 200 
 
 2 
 
 
 m 
 
 
 CO 
 
 
 UJ 
 
 
 oc 
 
 
 \- 
 
 150 
 
 U) 
 
 
 _) 
 
 
 < 
 
 
 Q. 
 
 
 O 
 
 
 tr 
 
 100 
 
 Q. 
 
 
 S 
 
 
 Z3 
 
 
 S 
 
 
 V 
 
 
 < 
 
 bO 
 
 -| I I I r 
 
 Micaceous sandstone 
 
 T I I F 
 
 o 
 
 ^0* 
 
 Principal stress equation, oj = III + 5.23 o-. 
 Correlation coefficient, r = 0.96 
 
 5 10 15 
 
 CONFINING PRESSURE, MPa 
 FIGURE 17. - Increase in compressive strength with confinement-Pittsburgh Sandstone. 
 
 20 
 
29 
 
 COMPRESSIVE STRENGTH. MPa 
 
 BULK DENSITY, g/cm' 
 
 Rock 
 
 40 80 120 160 200 240 
 
 Number 
 tested 
 
 Sandstone 
 
 t 1 1 1 I 1 
 
 21 
 
 W////.k'<'////A 
 
 Shole 
 
 
 14 
 
 1//////l^//////\ 
 
 Siltstone 
 
 V/Al/M 
 
 5 
 
 Mudstone 
 
 ■' V/I^M 
 
 5 
 
 Shaley limestone 
 
 
 26 
 
 [■////■//////♦■/////■/////J 
 
 Silty shale 
 
 ry^ 
 
 4 
 
 Limestone 
 
 
 34 
 
 l/////////^^/////////A 
 
 Rock 
 
 20 25 30 
 
 Number 
 tested 
 
 Sandstone 
 
 V/^/A 
 
 105 
 
 Sliole 
 
 V/////l^////A 
 
 64 
 
 Siltstone 
 
 * 
 
 17 
 
 Mudstone 
 
 V/%'A 
 
 IS 
 
 Stioley limestone 
 
 V/y^/A 
 
 68 
 
 Silty sltole 
 
 Vj^ 
 
 56 
 
 Limestone 
 
 V/^/A 
 
 90 
 
 INDIRECT I8RAZILIAN) TENSILE STRENGTH, MPa 
 
 Rock 
 
 1 3 5 7 9 II 
 
 Number 
 tested 
 
 Sandstone 
 
 ■ 1 I 1 1 1 
 
 65 
 
 l'////////A/////////A 
 
 Shale 
 
 V//////^///M///////Af///////////////////M 
 
 119 
 
 
 Siltstone 
 
 V/////%^////A 
 
 3 
 
 Mudstone 
 
 V/W/M/IV////U/M 
 
 33 
 
 Shaley limestone 
 
 w/////////l^////////A 
 
 52 
 
 Silty shole 
 
 V////////J^///////M 
 
 72 
 
 Limestone 
 
 U//////////^///////////A 
 
 105 
 
 POROSITY, pel 
 
 Rock 
 
 2 4 6 8 10 
 
 Number 
 tested 
 
 Sondstone 
 
 ■ ■■III 
 
 33 
 
 v//////^//////x 
 
 Shole 
 
 f7///f////J 
 
 39 
 
 Siltstone 
 
 V//^///A 
 
 8 
 
 Mudstone 
 
 y///ifV/A 
 
 8 
 
 Shaley limestone 
 
 V/////////^/////////M 
 
 29 
 
 Silty shale 
 
 V///^///A 
 
 38 
 
 Limestone 
 
 V/////////Af/////////A 
 
 38 
 
 STRAIN AT FAILURE, pet 
 
 Rock 
 
 Q2 04 06 08 10 1.2 
 
 Number 
 tested 
 
 Sandstone 
 
 1 1 I 1 1 1 
 
 21 
 
 W///W/W//Wk'///////W//////A 
 
 Shale 
 
 
 14 
 
 V///////l^///////A 
 
 Siltstone 
 
 
 5 
 
 y//////////^//////////i 
 
 Mudstone 
 
 
 5 
 
 V////J^/////i 
 
 Shaley limestone 
 
 
 29 
 
 v/////A/.'^///A 
 
 Silty shale 
 
 
 4 
 
 VW///^/////M 
 
 Limestone 
 
 
 34 
 
 V/////%/////,\ 
 
 PERMEABILITY, dorcy 
 
 Rock 
 
 IO-« 10"' 10'* 
 
 Number 
 tested 
 
 Sandstone 
 
 
 27 
 
 
 
 Shale 
 
 
 19 
 
 
 
 Siltstone 
 
 
 4 
 
 V//////////////////////^V///A 
 
 
 Mudstone 
 
 ♦ 
 
 1 
 
 Shaley limestone 
 
 
 13 
 
 
 
 Silty shale 
 
 
 14 
 
 
 
 Limestone 
 
 
 21 
 
 
 
 STATIC YOUNG'S MODULUS, GPa 
 
 Rock 
 
 10 20 30 40 50 
 
 Number 
 tested 
 
 Sandstone 
 
 
 21 
 
 V////////i%/W//////i 
 
 Shale 
 
 
 14 
 
 V//i%///M 
 
 Siltstone 
 
 yJ^^ 
 
 5 
 
 Mudstone 
 
 V/I^M 
 
 5 
 
 Shaley limestone 
 
 
 29 
 
 r///////l^'//WM 
 
 Silty shole 
 
 rz^T?, 
 
 4 
 
 Limestone 
 
 
 34 
 
 y//////4v/////A 
 
 SHORE HARDNESS 
 
 Rock 
 
 10 20 30 40 50 60 
 
 Number 
 tested 
 
 Sondstone 
 
 
 64 
 
 V//////AIM///A 
 
 Shole 
 
 V///////M'//////A 
 
 115 
 
 Siltstone 
 
 \Zdk!A 
 
 3 
 
 Mudstone 
 
 V////////J^>///////M 
 
 32 
 
 Shaley limestone 
 
 Y////A////M 
 
 53 
 
 Silty shale 
 
 l'////////A'////////A 
 
 78 
 
 Limestone 
 
 V/////%////M 
 
 108 
 
 POISSON'S RATIO, stotic 
 
 Rock 
 
 05 10 015 20 25 30 
 
 Number 
 tested 
 
 Sondstone 
 
 r I 1 1 I 1 
 
 17 
 
 1//////<'^A(///<'WM 
 
 Shale 
 
 
 1 1 
 
 V//////^//////M 
 
 Siltstone 
 
 
 4 
 
 1////i%'////A 
 
 Mudstone 
 
 
 5 
 
 'iy','///y/y//y//////y///y%//////////////^////^//» 
 
 
 Shotey limestone 
 
 
 27 
 
 y////////^///////x 
 
 Silty shole 
 
 
 3 
 
 y////^y///A 
 
 Limestone 
 
 
 32 
 
 V/////^'/////i 
 
 POINT LOAD INDEX, MPo 
 
 Rock 
 
 5 10 15 2 2.5 30 3 5 
 
 Number 
 tested 
 
 Sondstone 
 
 
 18 
 
 W//////J^///////////X 
 
 
 Shale 
 
 * 
 
 50 
 
 Siltstone 
 
 
 51 
 
 
 
 Mudstone 
 
 
 21 
 
 
 
 Shaley limestone 
 
 
 26 
 
 
 
 Cool 
 
 ♦ 
 
 Z 
 
 Limestone 
 
 
 19 
 
 ■///////////^///^//AAl/A////^A/,'//////y^////yM 
 
 
 FIGURE 18. - Physical-mechanical property mean and standard deviation by rock type. 
 
30 
 
 AXIAL P-.WAVE VELOCITY, km/s 
 
 Rock 
 
 1 2 3 4 5 6 
 
 Number 
 tested 
 
 Sandstone 
 
 w^y/A 
 
 105 
 
 Shale 
 
 V///^///A 
 
 64 
 
 Silretone 
 
 V/AfVA 
 
 17 
 
 Mudstone 
 
 W///\////A 
 
 18 
 
 Shaley limestone 
 
 Y/////!lt/////A 
 
 65 
 
 Silty sliole 
 
 V///A,V//A 
 
 56 
 
 Limestone 
 
 v//////Ay/////A 
 
 83 
 
 DYNAMIC YOUNG MODULUS, GPo 
 
 Rock 
 
 10 20 30 40 50 60 
 
 Number 
 tested 
 
 Sondstone 
 
 
 34 
 
 f/'/V/V^y/V/y/i 
 
 Shole 
 
 
 24 
 
 V///,\///M 
 
 Stltstone 
 
 
 6 
 
 V///^'///A 
 
 Mudstone 
 
 
 7 
 
 vm/z.^v/////! 
 
 Sttdtey limestone 
 
 
 27 
 
 V//////////^//////////A 
 
 Sllty sttole 
 
 
 8 
 
 V////^////A 
 
 Limestone 
 
 
 34 
 
 
 
 AXIAL S-WAVE VELOCITY, km/s 
 
 Rock 
 
 0,5 1.0 15 20 25 3,0 
 
 Numlier 
 tested 
 
 Sandstone 
 
 1 1 1 ' ■ ■ 
 
 39 
 
 K////4y////i 
 
 Shale 
 
 W///A^/////X 
 
 31 
 
 Siltstone 
 
 yy//////i(///////A 
 
 5 
 
 Mudstone 
 
 y//////J^/////A 
 
 10 
 
 Stioley limestone 
 
 V////>y///A 
 
 22 
 
 Sitty shole 
 
 V/////Jlf'////A 
 
 7 
 
 Limestone 
 
 W///My////A 
 
 34 
 
 DYNAMIC SHEAR MODULUS,GPo 
 
 Rock 
 
 4 8 12 16 20 24 
 
 Number 
 tested 
 
 Sandstone 
 
 
 34 
 
 V/WZ.\/////A 
 
 Shale 
 
 
 24 
 
 V////.^////A 
 
 Siltstone 
 
 
 6 
 
 W///^////i 
 
 Mudstone 
 
 
 7 
 
 V/////^'/////A 
 
 Snoiey limestone 
 
 
 27 
 
 V/////////y///////^^A 
 
 Silty shole 
 
 
 8 
 
 K/'///X^.f//////J 
 
 Limestone 
 
 
 34 
 
 fy//'//'//'/V/fW/'/y/'/'///'l 
 
 POISSON'S RATIO, dynamic 
 
 Rock 
 
 010 015 020 025 30 35 
 
 Number 
 tested 
 
 Sandstone 
 
 
 34 
 
 y///////////^//MM///M 
 
 Shale 
 
 
 24 
 
 'f////////ll////////A 
 
 Siltstone 
 
 
 6 
 
 l'//////'////y///i^V/////////////A 
 
 Mudstone 
 
 
 7 
 
 vw//////A^y//////////i 
 
 Stioley limestone 
 
 
 26 
 
 ty/^///^'^/y'//i 
 
 Silty stiole 
 
 
 8 
 
 Y/////////i^/////////A 
 
 Limestone 
 
 
 34 
 
 V/////,^//////i 
 
 FIGURE 19. - Acoustical property mean and standard deviation by rock type. 
 ROCK CLASSIFICATION FOR ENGINEERING PURPOSES 
 
 INTACT ROCK 
 
 The strength and the modulus ratio 
 were plotted according to the Deere (_5) 
 intact-rock classification scheme that is 
 widely used in civil engineering and 
 
 mining (fig. 22). The scheme describes 
 the rock in terms of five categories of 
 compressive strength, and three catego- 
 ries, high, medium, and low, of the ratio 
 of modulus to strength. 
 
31 
 
 •: 5 - 
 
 en 
 
 ^"4 
 
 o 
 
 o 
 
 _I 
 
 UJ 
 
 > 
 
 > 3 
 
 < 
 
 5 
 
 
 1 
 
 1 1 
 KEY 
 
 1 
 
 1 
 
 1 
 V 
 
 
 V 
 
 Limestone 
 
 -sholey limestone 
 
 
 
 
 
 ▲ 
 
 Mudstone- 
 
 ■siltstone 
 
 
 V 
 
 
 
 o 
 
 Sandstone 
 
 
 a 
 
 2^^^.^.— - 
 
 ,^ — 
 
 
 D 
 
 Shale 
 
 
 
 
 
 • 
 
 Coal 
 
 OV o ^^-^^ 
 
 V V 
 
 D 
 V 
 
 
 
 " 
 
 D 
 D C 
 
 o v- — 
 
 O 
 
 
 
 ~ 
 
 ▲ 
 
 X 
 
 
 
 
 
 
 a 
 
 x^ 
 
 ▲ 
 
 
 
 
 
 D 
 
 / 
 
 
 
 y = x/ (0.3559+ O.I46x) 
 
 - 
 
 
 a 
 
 
 
 r = 0.92 
 
 
 
 /A 
 
 D 
 
 1 
 
 1 1 
 
 1 
 
 1 
 
 1 
 
 
 2 3 4 5 6 
 
 P-WAVE VELOCITY, INTACT ROCK,km/s 
 FIGURE 20. - Crossplot of in situ versus laboratory P-wave velocity. 
 
 AL 
 
 RENGTH, MPa 
 
 OJ X 
 
 o c 
 o c 
 
 - 
 
 1 
 
 \ 1 1 
 
 1 1 
 
 D 
 
 D 
 
 D 
 D 
 
 ^ 
 
 UNIAX 
 SSIVE ST 
 
 O 
 O 
 
 
 
 a 
 
 
 
 ^ 100 
 
 - 
 
 
 1 1 1 
 
 
 - 
 
 Q. 
 
 2 
 
 
 a 
 
 y= -42.0+45.2X 
 
 
 O 
 
 
 ^^ 
 
 r = 0.793 
 1 i 
 
 
 
 
 I 
 
 6 
 
 2 3 4 5 
 
 P-WAVE VELOCITY, km/s 
 FIGURE 21. - Crossplot of uniaxial compressive strength versus P-wave velocity. 
 
32 
 
 Very low 
 strength 
 
 100 
 
 o 
 
 CL 
 (D 
 
 (n 
 
 _l 
 O 
 
 o 
 
 CO 
 
 "o 
 
 z 
 
 o 
 
 iLl 
 
 o 
 
 z 
 < 
 
 ./*'* 
 
 
 - / 
 
 / 
 / 
 
 D 
 
 Low 
 strength 
 
 .^^ 
 
 / <^^ 
 
 
 /. 
 
 / 
 / 
 
 >.° 
 
 V 
 
 Medium 
 strength 
 
 T — I I I 
 
 
 / 
 
 
 . 8 
 
 a o 
 
 D 
 
 .^o" 
 
 J I I L 
 
 B 
 
 High 
 strength 
 
 Tf7 
 
 
 "a^ 
 
 
 Very high 
 strength 
 
 "I — TT 
 / 
 
 "I — I I I 
 
 
 v^ 
 
 KEY 
 
 o Sandstone 
 
 D Shale 
 
 V Limestone 
 
 ^ Shaley limestone 
 
 A Mudstone 
 
 ▼ Silt stone 
 
 ■ Silty shale 
 
 J I L 
 
 10 
 
 100 
 COMPRESSIVE STRENGTH, MPa 
 
 FIGURE 22. - Intact-rock engineering classification. 
 
 1,000 
 
33 
 
 Figure 22 shows that for the coal mea- 
 sures at the Gateway Mine, the limestones 
 and shaley limestones are classified 
 in the high- to very-high-strength, and 
 medium- to low-modulus-ratio categories. 
 Siltstone and silty shales are predomi- 
 nantly medium strength with low modulus 
 ratio, and mudstone is low strength with 
 medium to low modulus ratio. Such clas- 
 sification schemes for intact rock are 
 relevant to drilling, blasting, and 
 fragmentation on a smaller scale, and 
 for massive rock without joints. These 
 schemes also aid in selection of appro- 
 priate mining excavation and fragmenta- 
 tion equipment. 
 
 ROCK MASS 
 
 Rock mass classification schemes take 
 into account the influence of discontinu- 
 ities and often, directly or indirectly, 
 in situ environmental factors such as 
 stress and moisture for estimating 
 strength and def ormational behavior of 
 rock masses. The geomechanics classifi- 
 cation proposed by Bieniawski (6) appears 
 
 adaptable to coal mining applications, 
 and has been used to a limited extent in 
 classifying roof conditions. The classi- 
 fication is based on uniaxial compressive 
 strength, RQD, the spacing, orientation 
 and condition of joints, ground water 
 conditions, and sometimes other modifying 
 parameters. Significant parameters in 
 the geomechanics classification for de- 
 termining roof conditions were estimated 
 for 50 ft of strata overlying the Pitts- 
 burgh Coalbed. Roof strata at the Gate- 
 way Mine were divided into three distinct 
 lithologic units, and the rock mass rat- 
 ing was applied to each lithologic member 
 (table 3). The lowermost member was 
 classified as poor rock, the middle mem- 
 ber as good rock, and the uppermost mem- 
 ber as fair rock. Such determinations 
 permit speculation on maximum unsupported 
 roof span and standup time and assist in 
 selection of appropriate support. For 
 application in coal measure roof rocks, 
 however, modification and improvement of 
 the classification scheme is needed to 
 obtain a higher degree of predictability 
 of standup time and support requirements. 
 
 TABLE 3. - Rock mass classification of Gateway roof rock 
 
 Lithologic member and thickness. . .ft. 
 
 I, 10.8 
 
 II, 24.7 
 
 III, 19. 
 
 Strength of intact rock: 
 
 Uniaxial compressive strength. .MPa. . 
 
 Rating 
 
 Point-load index ' MPa . . 
 
 73... 
 (7).. 
 0.08. 
 
 RQD 
 
 Rating. 
 
 ,pct, 
 
 49., 
 (8), 
 
 106 , 
 
 (12) , 
 
 0.37-0.66.., 
 
 98.. 
 (20), 
 
 Spacing of discontinuities m. 
 
 Rating 
 
 0.1-0.5. 
 (10).... 
 
 0.5. 
 (20), 
 
 Condition of discontinuities, 
 
 Slickensided 
 
 Rating. 
 
 (6), 
 
 Slightly rough, 
 
 separation 
 
 < 1 mm. 
 (12) 
 
 Ground water general conditions. 
 Rating 
 
 Moist. 
 (7)... 
 
 Moist. 
 (7).., 
 
 Total rating 
 
 Class No 
 
 Description 
 
 Potential modifier; 
 durability. 
 
 slake 
 
 pet. 
 
 38 
 
 IV 
 
 Poor rock... 
 
 75.5 (silt- 
 stone) . 
 91.1 (shale) 
 
 71 
 
 II 
 
 Good rock. 
 
 99.3. 
 
 81. 
 
 (7). 
 
 0.09. 
 
 97. 
 (20). 
 
 0.45-0.48. 
 (20). 
 
 Slightly rough, 
 
 separation 
 
 1-5 mm. 
 (6). 
 
 Moist. 
 (7). 
 
 60. 
 
 III. 
 
 Fair rock. 
 
 87. 
 
 •No rating is given because uniaxial compressive 
 range. 
 
 stength is preferred in the low 
 
34 
 
 ROCK PROPERTIES DATA BASE 
 
 Although the rock properties deter- 
 mined in these investigations will be 
 published, there is need for the es- 
 tablishment of a computerized data base 
 where rock property data can be obtained 
 through search and retrieval opera- 
 tions from remote terminal. The Bureau 
 is working toward this goal by input- 
 ting the property data into computerized 
 files. The numerical data base manage- 
 ment system, in addition, needs to be 
 able to sort and retrieve coded numerical 
 
 information so that the data can be 
 manipulated and various mathematical and 
 statistical analyses performed. As a 
 start in this direction, property data 
 tables from Bureau task files were orga- 
 nized into a data base with standardized 
 format for a wide variety of rock types 
 (table 4). These property tables and a 
 detailed description of the data base 
 structure are being compiled into a Bu- 
 reau Information Circular. 
 
 TABLE 4. - Mechanical property tables for mine rock 
 
 ID 
 
 Rock type and 
 modifier 
 
 Location 
 
 and 
 
 description 
 
 Source 
 
 Young's 
 
 modulus , 
 
 GPa 
 
 Pois- 
 son's 
 ratio 
 
 Compressive 
 
 strength, 
 
 MPa 
 
 Tensile 
 
 strength, 
 
 MPa 
 
 Brazilian 
 
 line-load 
 
 strength, 
 
 MPa 
 
 1103 
 
 Shale , 
 calcareous 
 kerogen. 
 
 CO 
 
 TCRC 
 
 5.9 
 (N=20) 
 
 
 85.0 
 (N=20) 
 
 7.6 
 (N=32) 
 
 
 
 
 
 1104 
 
 . . .do 
 
 Garfield 
 Co., CO. 
 
 TCRC 
 
 7.0 
 (N=20) 
 
 0.358 
 (N=20) 
 
 79.6 
 (N=20) 
 
 
 
 1105 
 
 . . .do 
 
 ...do 
 
 TCRC 
 
 3.4 
 (N=2) 
 
 0.370 
 (N=2) 
 
 62.0 
 (N=2) 
 
 
 
 1106 
 
 . . .do 
 
 . . .do 
 
 TCRC 
 
 8.0 
 (N=56) 
 
 0.183 
 (N=55) 
 
 90.1 
 (N=57) 
 
 13.2 
 (N=54) 
 
 
 1107 
 
 Shale. 
 
 PA 
 
 TCRC 
 
 16.1 
 (N=22) 
 
 
 74.4 
 (N=22) 
 
 
 6.4 
 
 
 
 (N=17) 
 
 1108 
 
 . . .do 
 
 PA 
 
 TCRC 
 
 13.7 
 (N=24) 
 
 
 75.0 
 (N=24) 
 
 
 6.1 
 
 
 
 (N=24) 
 
 1109 
 
 . . .do 
 
 Rice Co. , 
 KS. 
 
 TCRC 
 
 15.2 
 (N=7) 
 
 
 72.5 
 (N=7) 
 
 
 
 1110 
 
 . . .do 
 
 . . .do 
 
 TCRC 
 
 21.0 
 (N=5) 
 
 
 80.5 
 (N=5) 
 
 
 
 N 
 TCRC 
 
 Number of samples tested. 
 Twin Cities Research Center. 
 
MINING APPLICATIONS 
 
 35 
 
 The physical and mechanical properties 
 of coal measure and other mine rocks have 
 application in most aspects of premine 
 planning and mine design. Elastic and 
 strength properties are especially needed 
 in evaluating and modeling rock mass be- 
 havior and structural stability in mines. 
 Geologic data from both surface and un- 
 derground surveys are required to deter- 
 mine the continuity of coal seams and ore 
 reserves, identify lithologic changes and 
 trends, and delineate geological hazards 
 in the proximity of mine workings. Geo- 
 physical properties are necessary for 
 interpreting rock mass and coal or ore 
 characteristics in advance of mining and 
 in inaccessible zones between exploration 
 boreholes or adjacent to mine workings. 
 Acoustic and electrical property data are 
 particularly beneficial in the design of 
 geophysical probes that identify and 
 delineate rock mass conditions during the 
 exploration phase of mining or in advance 
 of the working face during mine develop- 
 ment. Index properties are needed to 
 infer other useful engineering proper- 
 ties, when formalized testing procedures 
 
 and facilities are unavailable or too 
 complicated and costly for a mining op- 
 eration to utilize on a cost-effective 
 basis. Index properties can be particu- 
 larly advantageous in coping with the 
 frequent and common lithologic changes 
 prevalent in coal measure rocks. More- 
 over, the index properties determined in 
 exploration or in-mine geotechnical pro- 
 grams can be utilized in engineering 
 classification schemes that infer rock 
 mass response, indicate stability and 
 support requirements , and permit experi- 
 ence gained at one site to be transferred 
 to other sites where similar conditions 
 exist. 
 
 Eventually, an interactive, computer- 
 ized data base of engineering properties 
 of rock can be established as an informa- 
 tion retrieval system for use by mine 
 operators, planners, and research or reg- 
 ulatory agencies. Such a comprehensive 
 engineering properties data base will 
 require input from numerous sources and 
 full cooperation of the industry. 
 
 REFERENCES 
 
 1. Lewis, W. E., and S. Tandanand 
 (eds.). Bureau of Mines Test Procedures 
 for Rocks. BuMines IC 8628, 1974, 
 223 pp. 
 
 2. Brown, E. T. (ed.). Rock Char- 
 acterization Testing and Monitoring — 
 ISRM Suggested Methods. Pergamon, 1981, 
 211 pp. 
 
 College Park, PA, June 11-14, 1972. Am. 
 Soc. Civil Eng., 1972, pp. 649-687. 
 
 5. Deere, D. V., and R. P. Miller. 
 Engineering Classification and Index 
 Properties for Intact Rock. U.S. Air 
 Force Systems Command, Air Force Weapons 
 Lab., Kirkland AFB, NM, Tech. Rep. AFWL- 
 TR-65-116, 1966, 308 pp. 
 
 3. Dresser Industries, Inc. Well Log 
 Interpretation Techniques. 1982, 481 pp. 
 
 4. Thill, R. E. Acoustic Methods for 
 Monitoring Failure in Rock. Proc. 14th 
 Symp. on Rock Mechanics, PA State Univ., 
 
 6. Bieniawski, A. T. , F. Rafia, and 
 D. A. Newman. Ground Control Investiga- 
 tions for Assessment of Roof Conditions 
 in Coal Mines. Proc. 21st Symp. on Rock 
 Mechanics, Rolla, MO, May 28-30, 1980. 
 Univ. MO— Rolla, 1980, pp. 691-700. 
 
36 
 
 PILLAR DESIGN EQUATIONS FOR COAL EXTRACTION 
 By Clarence 0. Babcock^ 
 
 ABSTRACT 
 
 Coal mine pillar design equations de- 
 veloped during the period 1833 to 1980 
 are reviewed. These equations suggest 
 that mine pillars of different sizes are 
 required for safety. Two widely used de- 
 sign equations, those of Wilson and 
 Wardell, give pillar areas that vary by 
 
 an average factor of 2.08. The pillar 
 width and height alone are not the pri- 
 mary parameters in the design problem. 
 The Mohn-Coulomb stress-failure criteria 
 can be used to explain the difference in 
 the estimated pillar sizes. 
 
 INTRODUCTION 
 
 Coulomb (J_)2 was the first person to 
 publish a rational theory of earth pres- 
 sures (1773), and this theory is in wide- 
 spread use today for both soil and rock 
 mechanics applications. His theory was 
 also the first to show that the strength 
 of a solid is related in part to the ma- 
 terial properties and in part to the 
 amount of constraint provided during the 
 testing. Since that time, nearly every 
 
 theory of failure of solids has been in 
 terms of the combined stress, strain, or 
 strain energy state. The role of con- 
 straint in strength has only recently 
 been emphasized in the design of mine 
 pillars. Too often, the investigator has 
 failed to realize that the constraint and 
 not the material strength is responsible 
 for pillar behavior when combined states 
 of stress or strain are involved. 
 
 STATE OF THE ART IN PILLAR DESIGN 
 
 Wilson (2^) gave the equation for pil- 
 lar size, W, in feet, for a safety factor 
 (S.F.) of 1.0 as 
 
 W = (R/3 + 2HD X 10"^) 
 
 + [R/3 + 2HD X 10-5)2 
 
 + (r2/3 - 4h2d2 X 10-6)]l/2^ (1) 
 
 where R, H, and D are the entry width, 
 entry height, and the depth of the coal 
 seam below the surface, respectively, all 
 
 in feet. 
 
 the width and height of the pillar in 
 feet, respectively. Wardell gives tables 
 of minimum pillar widths for coal seams 
 (pillar heights) of 4, 5, 6, 7, 8, 9, 10, 
 and 12 ft, for a safety factor of 1.5. 
 The pillar sizes are determined from a 
 tributary area relationship that he does 
 not define mathematically. A relation- 
 ship that generates those tabulated val- 
 ues was derived by Babcock and Hooker (4^) 
 as 
 
 (W + R)2 X 1.5 D 
 W2 
 
 1,000 _j_ 20 (W2) 
 
 /H 
 
 h2 
 
 (3) 
 
 Wardell (3^) proposed an equation of the 
 form 
 
 S = a/ »^ + b (W/H)2 
 
 = 1,000 / /H + 20 (W/H)2 
 
 (2) 
 
 for the strength of actual mine pillars. 
 Here the variable S is the strength of 
 the pillar in pounds per square inch; a 
 and b are coefficients; and W and H are 
 
 where D is the depth below surface, R is 
 the room width, and W and H are the width 
 and height of the coal pillar, all in 
 feet. 
 
 ^Mining engineer, Denver Research Cen- 
 ter, Bureau of Mines, Denver, CO. 
 
 ^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
37 
 
 Panek O) proposed an equation based on 
 a theory of similitude for the design of 
 mine pillars in which the physical prop- 
 erties of the roof, floor, and coal are 
 
 obtained from laboratory compressive 
 testing of model pillars of the same 
 geometry as the mine pillars with steel 
 platens. His equation is 
 
 Spred I cted _ ( ^r_ 
 ^ known \ ^s 
 
 C4 
 
 it 
 
 Es 
 
 C5 
 
 Ucr 
 Ucs 
 
 where Spredlcted is the expected pillar 
 strength, S^no^n is the measured model 
 strength, and E is the Young's modulus, 
 all in pounds per square inch; u is the 
 coefficient of friction; and v is the 
 Poisson's ratio. The subscripts c, f, r, 
 and s denote the coal, floor, roof, and 
 steel, respectively. The double sub- 
 scripts cf , cr, and cs denote the co- 
 efficient of friction between the coal 
 and floor, coal and roof, and coal 
 
 C6 
 
 Ucf 
 Ucs 
 
 C7 
 
 C8 
 
 _1 
 
 eg 
 
 (4) 
 
 and steel, respectively. The coeffi- 
 cients C4 , C5 , ... C9 are constants to be 
 determined by best fit statistical meth- 
 ods. This relationship assumes no bond- 
 ing between the coal and roof and floor 
 rock. A more complicated relationship 
 with three additional terms associated 
 with C 1 , C2 , and C3 is given by Panek (_5) 
 for the case when the model has a dif- 
 ferent geometric shape than the mine pil- 
 lar has . 
 
 COMPARISON OF LABORATORY AND IN SITU SPECIMEN TESTING 
 WITH OBSERVED MINE PILLAR BEHAVIOR 
 
 In general, the pillar failure predic- 
 tion equations are of the form 
 
 ^=A + B^ 
 
 (5) 
 
 where Op and Oq are the stresses at fail- 
 ure in the pillar and the specimen; A, B, 
 a, and 3 are constants; and W and H are 
 width and height of the pillar, respec- 
 tively. Numerous equations of this form 
 and the results given in the literature 
 were summarized by Babcock, Morgan, and 
 Haramy (6^) and are shown in table 1 and 
 figure 1. In figure 1, the zone between 
 the failed and unfailed pillars was given 
 by Warden O) for S.F. = 1.0. An empir- 
 ical curve for S.F. = 1.5 is also shown. 
 The theoretical result from reference 2 
 (table 1) is so marked. 
 
 It is apparent that the equations in 
 the literature predicted as safe many 
 pillars that were unsafe. While these 
 equations may be correct for a given 
 coal, none of them, other than Warden's, 
 predicted correctly the large-scale or 
 full-sized mine pillar behavior. 
 
 123456789 10 
 W/H RATIO 
 
 FIGURE 1.- Comparison of pillar strengths pre- 
 dicted by published equations and experimental 
 strengths of mine pi liars and large test specimens. 
 Source: Reference 6. (Underlined numbers in pa- 
 rentheses refer to items in the list of references 
 at the end of this paper. ) 
 
38 
 
 33 
 
 3 
 
 + 
 < 
 
 D 
 
 B 
 u 
 o 
 
 0) 
 
 0) 
 
 c 
 o 
 •w 
 
 4-1 
 
 nj 
 
 3 
 
 cr 
 
 0) 
 
 c 
 
 M 
 •H 
 CO 
 0) 
 
 l-i 
 
 I 
 
 
 1 
 
 )-l 
 
 <u 
 
 (1) 
 
 o 
 
 <w 
 
 c 
 
 (1) 
 
 0) 
 
 Pi 
 
 
 r-» 00 ON o -^ — I 
 
 M 
 4J 
 
 C 
 3 
 
 o 
 
 Q) 
 
 03 
 CU 
 U 
 
 e 
 
 0) 
 
 <U 4J 
 
 M 
 O 
 U 
 CO 
 bO 
 1-1 
 4J 
 CO 
 (U 
 > 
 
 u 
 to 
 (1) 
 
 
 T3 
 
 c 
 
 C tS 
 
 O M 
 
 N <D 
 
 Z CO 
 
 0) <u 
 
 c c 
 o o 
 
 CO CO O l-l 
 0) 13 T3 X 
 
 •H tO 
 
 :^ 
 
 bO 
 3 
 CO o 
 
 u x) 
 
 CO 
 CO <U 
 
 a w 
 1-1 n) 
 1-1 iJ 
 
 :^^ 
 
 U J-l 
 
 3 1-1 
 O C 
 
 M 
 l-i 
 3 
 Xl 
 
 CO 
 4J 
 4J 
 •H 
 
 P- 
 
 ■V 
 
 c 
 
 (0 
 ^ • 
 
 o 
 
 (U 
 PQ 
 
 s 
 o 
 
 t3 
 bO 
 C 
 
 CO 
 O 
 
 i-( 
 
 x: o 
 
 4-1 T3 
 
 3 
 O 
 
 cn 
 
 CO 
 
 c u 
 
 >. CO 
 H ffi 
 
 14-1 
 
 >4-l >n 
 
 3 0) 
 
 O rH 
 
 CO 
 
 Q. C 
 CD l-i 
 (U (0 iH 
 
 Q m pa 
 
 CO 
 
 3 
 o 
 
 ■H C 
 
 e CO o 
 
 3 ^ T3 
 
 B 
 o 
 
 T3 CO 
 bO O 
 
 o 
 bO 
 u 
 
 CO 
 ►J 
 
 < 0) 
 
 CO TJ 
 
 B 
 o 
 
 bO 
 
 CO 
 CO 
 
 o 
 
 z 
 
 COC/)W3C/3cncrt|.Ji-JiJC/3C/lC/lCO(/lHJiJl-JHh4i-Ji-)CO 
 
 IT) 
 
 in u-1 CO 
 
 vO u-1 
 
 CM 
 
 m lA CO 
 
 -H sa- 
 
 cs) ^^ ,-H O 
 
 I I 
 
 CM Ui 
 
 o 
 
 o o o u 
 
 CJ 
 
 o o 
 in o 
 
 l-s. CO 
 
 00 -sT O O 
 
 e» 00 o 
 -* in vo 
 
 o 
 
 CSI 
 
 PO CO 00 
 
 o o o o 
 
 u-1 O 
 
 oooooooooo 
 
 o o o o o 
 
 ^ o 
 o 
 
 ^ 
 
 • u 
 
 • 0) 
 
 • bO 
 
 • 3 
 
 • -l-l 
 
 • XI 
 4-1 O 
 CO CO 
 
 o 
 
 0) 
 C 4-> 
 
 o c 
 
 CO 0) 
 
 O. 4-1 
 
 M c 
 
 CO 3 
 > M ►-J U M 
 
 3 XI 
 CO O 
 
 X rH 
 
 4J (0 
 
 •^ g 
 
 <4-l 0) 
 
 •H (I) 
 
 • T) 
 
 • 3 
 >s CO 
 
 -3 rH 
 
 CO O 
 
 crt o pd 
 
 c 
 o 
 B 
 to 
 
 rH 
 
 CO 
 CO 
 
 CO ^ r^ r-l ^ 
 
 en r-s On O -H 
 00 00 00 ^ On 
 
 CM <Ti 
 
 rH fO 
 ON 0\ 
 
 rH ^ VO <^ VO 
 
 -* m in vo vo 
 
 On On On On On 
 
 • 3 
 
 • 0) 
 
 • T3 
 
 • U 
 
 • 0) 
 3 0) 
 
 O ts 
 
 CO 
 
 (U 
 3 
 3 bO 
 iH CO CO 
 S > S 
 
 (M en -* 
 r-- p^ r^ 
 
 On ON ON 
 
 T3 
 
 iH 
 
 <-{ 
 
 3 
 M 
 4-) 
 CO 
 
 a 
 
 in 
 
 ON 
 
 0) 
 •3 
 U 
 
 ON 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 O 
 
 
 
 
 
 
 
 T3 
 
 
 
 
 
 
 
 (U 
 
 
 
 
 
 
 
 U 
 
 
 
 
 
 
 
 a 
 
 
 
 
 
 
 
 CO 
 
 
 • 
 
 
 
 
 
 M 
 
 
 B 
 
 
 
 
 
 cu 
 
 
 CO 
 
 
 
 
 
 X 
 
 
 a> 
 
 
 
 
 
 4J 
 
 
 CO 
 
 
 
 
 
 o 
 
 
 X 
 
 
 
 
 
 rH 
 
 
 o 
 
 
 
 
 
 ^ 
 
 
 CO 
 
 
 
 
 
 CO 
 
 
 0) 
 
 
 
 
 
 (U 
 
 
 u 
 
 
 
 
 
 rH 
 
 
 o 
 
 
 
 
 
 iH 
 
 
 MH 
 
 
 
 
 
 ^ 
 
 
 4J 
 
 
 
 
 
 
 
 3 
 
 
 
 
 
 «S 
 
 
 (U 
 
 
 
 
 
 X 
 
 
 M 
 
 • 
 
 
 • 
 
 
 4J 
 
 
 0) 
 
 0) 
 
 
 Q) 
 
 
 bO 
 
 
 >4H 
 
 ,-\ 
 
 
 rH 
 
 
 3 
 
 • 
 
 IIH 
 
 o. 
 
 • 
 
 a. 
 
 
 a; 
 
 vO 
 
 •rA 
 
 B 
 
 0) 
 
 B 
 
 
 u 
 
 
 "O 
 
 CO 
 
 rH 
 
 CO 
 
 
 u 
 
 OJ 
 
 
 en 
 
 X 
 
 en 
 
 
 m 
 
 CJ 
 
 CO 
 
 
 CO 
 
 
 
 
 3 
 
 iH 
 
 u 
 
 CJ 
 
 4-1 
 
 
 r^ 
 
 <u 
 
 
 CO 
 
 •rl 
 
 CO 
 
 
 to 
 
 u 
 
 4-1 
 
 0) 
 
 rH 
 
 (U 
 
 
 CJ 
 
 CD 
 
 3 
 
 4-1 
 
 O. 4J 
 
 • 
 
 iH 
 
 MH 
 
 CO 
 
 
 a 
 
 
 >%x 
 
 CU 
 
 4-1 
 
 0) 
 
 CO 
 
 rH 
 
 u 
 
 U 
 
 Pi 
 
 CO 
 
 bO 
 
 
 rH 
 
 o 
 
 3 
 
 
 3 
 
 V4 
 
 4J 
 
 CO 
 
 a> 
 
 CJ 
 
 
 o 
 
 CO 
 
 o 
 
 B 
 
 X 
 
 
 • • 
 
 o 
 
 hJ 
 
 z 
 
 CO 
 
 H 
 
 CO 
 CO 
 
 o 
 
 o 
 
 0) 
 CJ 
 
 u 
 
 3 
 
 o 
 
 u 
 
 t-I 
 
 z 
 
 CO 
 
 H 
 
 — 
 
 CO 
 
39 
 
 COMPARISON OF PILLAR SIZES DETERMINED WITH THE EQUATIONS 
 OF WARDELL AND WILSON 
 
 The equations of Wardell (2) and Wilson 
 (1) were used to calculate the minimum 
 pillar sizes required, in feet, for seam 
 thicknesses of 6 and 10 ft, entry widths 
 of 16, 20, and 24 ft, and depths of 200 
 to 3,000 ft (table 2). The average pil- 
 lar stresses corresponding to the pillar 
 sizes in table 2 are given in table 3. 
 Note that the stresses increase with 
 depth. These results can all be repre- 
 sented by straight line relationships as 
 indicated by the parameters given in ta- 
 ble 4. In this table, a, m, and r are 
 the unconfined or uniaxial strength at 
 
 the surface, the slope of the stress 
 (ordinate) versus depth (abscissa) , and 
 the correlation coefficient. If these 
 results were a perfect fit to a straight 
 line, r would have a value of 1.0. The 
 values of 0.992 or better indicate that a 
 straight line assumption is a reasonable 
 one. Notice that by changing the in- 
 tercept a and the slope, entry widths 
 from 16 to 24 ft can be fitted. All 
 these results could therefore be de- 
 scribed by a Mohr-Coulomb model with rea- 
 sonable success. 
 
 TABLE 2. - Comparison of pillar widths, in feet, calculated 
 with the equations of Wardell and Wilson 
 
 Depth, 
 ft 
 
 16-ft room width 
 
 Wardell Wilson 
 
 20-ft room width 
 
 Wardell | Wilson 
 
 24-f t room width 
 Wardell | Wilson 
 
 6-ft-THICK COAL SEAM 
 
 200 
 
 21 
 
 20 
 
 23 
 
 24 
 
 26 
 
 28 
 
 400 
 
 31 
 
 23 
 
 33 
 
 27 
 
 36 
 
 31 
 
 600 
 
 39 
 
 27 
 
 41 
 
 31 
 
 44 
 
 35 
 
 800 
 
 45 
 
 30 
 
 47 
 
 34 
 
 50 
 
 38 
 
 1,000 
 
 50 
 
 33 
 
 53 
 
 37 
 
 55 
 
 42 
 
 1,200 
 
 54 
 
 36 
 
 57 
 
 41 
 
 60 
 
 45 
 
 1,400 
 
 59 
 
 40 
 
 61 
 
 44 
 
 64 
 
 48 
 
 1,600 
 
 62 
 
 43 
 
 65 
 
 47 
 
 68 
 
 51 
 
 1,800 
 
 66 
 
 46 
 
 69 
 
 50 
 
 72 
 
 55 
 
 2,000 
 
 69 
 
 49 
 
 72 
 
 53 
 
 75 
 
 58 
 
 2,500 
 
 75 
 
 57 
 
 80 
 
 61 
 
 83 
 
 66 
 
 3,000 
 
 84 
 
 64 
 
 87 
 
 69 
 
 90 
 
 73 
 
 
 
 10-ft- 
 
 THICK COAL SEAM 
 
 
 
 200 
 
 29 
 
 22 
 
 33 
 
 26 
 
 36 
 
 30 
 
 400 
 
 46 
 
 28 
 
 49 
 
 32 
 
 52 
 
 36 
 
 600 
 
 58 
 
 33 
 
 61 
 
 37 
 
 64 
 
 42 
 
 800 
 
 68 
 
 39 
 
 71 
 
 43 
 
 74 
 
 47 
 
 1,000 
 
 76 
 
 44 
 
 80 
 
 48 
 
 83 
 
 52 
 
 1,200 
 
 84 
 
 49 
 
 87 
 
 53 
 
 90 
 
 58 
 
 1,400 
 
 91 
 
 54 
 
 94 
 
 59 
 
 97 
 
 63 
 
 1,600 
 
 97 
 
 59 
 
 100 
 
 64 
 
 103 
 
 68 
 
 1,800 
 
 103 
 
 64 
 
 106 
 
 69 
 
 109 
 
 73 
 
 2,000 
 
 108 
 
 69 
 
 112 
 
 74 
 
 115 
 
 78 
 
 2,500 
 
 121 
 
 81 
 
 124 
 
 86 
 
 127 
 
 91 
 
 3,000 
 
 132 
 
 93 
 
 135 
 
 98 
 
 139 
 
 103 
 
40 
 
 TABLE 3. - Comparison of average pillar stress, in pounds 
 per square inch, calculated for the equations of Wardell 
 and Wilson 
 
 Depth, 
 ft 
 
 16-ft room width 
 Wardell | Wilson 
 
 20-ft room width 
 
 Wardell | Wilson 
 
 24-f t room width 
 
 Wardell Wilson 
 
 6-ft-THICK COAL SEAM 
 
 200 
 
 745 
 
 778 
 
 839 
 
 807 
 
 888 
 
 705 
 
 400 
 
 1,103 
 
 1,380 
 
 1,238 
 
 1,454 
 
 1,333 
 
 1,299 
 
 600 
 
 1,432 
 
 1,826 
 
 1,594 
 
 1,949 
 
 1,720 
 
 1,778 
 
 800 
 
 1,764 
 
 2,257 
 
 1,951 
 
 2,422 
 
 2,103 
 
 2,236 
 
 1,000 
 
 2,091 
 
 2,646 
 
 2,277 
 
 2,848 
 
 2,476 
 
 2,615 
 
 1,200 
 
 2,420 
 
 3,004 
 
 2,628 
 
 3,188 
 
 2,822 
 
 3,004 
 
 1,400 
 
 2,715 
 
 3,293 
 
 2,962 
 
 3,554 
 
 3,176 
 
 3,372 
 
 1,600 
 
 3,039 
 
 3,615 
 
 3,283 
 
 3,902 
 
 3,514 
 
 3,721 
 
 1,800 
 
 3,334 
 
 3,924 
 
 3,594 
 
 4,234 
 
 3,840 
 
 4,016 
 
 2,000 
 
 3,642 
 
 4,223 
 
 3,919 
 
 4,553 
 
 4,182 
 
 4,341 
 
 2,500 
 
 4,417 
 
 4,921 
 
 4,688 
 
 5,290 
 
 4,986 
 
 5,094 
 
 3,000 
 
 5,102 
 
 5,625 
 
 5,445 
 
 5,989 
 
 5,776 
 
 5,843 
 
 10-ft-THICK COAL SEAM 
 
 200 
 
 578 
 
 716 
 
 619 
 
 751 
 
 667 
 
 778 
 
 400 
 
 872 
 
 1,185 
 
 952 
 
 1,268 
 
 1,025 
 
 1,333 
 
 600 
 
 1,172 
 
 1,587 
 
 1,270 
 
 1,709 
 
 1,361 
 
 1,778 
 
 800 
 
 1,465 
 
 1,909 
 
 1,577 
 
 2,061 
 
 1,684 
 
 2,191 
 
 1,000 
 
 1,758 
 
 2,231 
 
 1,875 
 
 2,408 
 
 1,994 
 
 2,563 
 
 1,200 
 
 2,041 
 
 2,534 
 
 2,178 
 
 2,732 
 
 2,310 
 
 2,878 
 
 1,400 
 
 2,323 
 
 2,823 
 
 2,471 
 
 3,012 
 
 2,614 
 
 3,204 
 
 1,600 
 
 2,606 
 
 3,103 
 
 2,765 
 
 3,308 
 
 2,919 
 
 3,514 
 
 1,800 
 
 2,883 
 
 3,375 
 
 3,052 
 
 3,594 
 
 3,216 
 
 3,814 
 
 2,000 
 
 3,164 
 
 3,642 
 
 3,334 
 
 3,873 
 
 3,506 
 
 4,104 
 
 2,500 
 
 3,846 
 
 4,302 
 
 4,046 
 
 4,558 
 
 4,241 
 
 4,791 
 
 3,000 
 
 4,526 
 
 4,945 
 
 4,746 
 
 5,219 
 
 4,950 
 
 5,473 
 
 TABLE 
 
 4. - Parameters for 
 
 straight line fit to table 3 data 
 
 Parameter 
 
 16-ft room width 
 
 20-ft room width 
 
 24-ft room width 
 
 
 Wardell | Wilson 
 
 Wardell | Wilson 
 
 Wardell | Wilson 
 
 6-ft-THICK COAL SEAM 
 
 
 
 505 
 
 812 
 
 615 
 
 862 
 
 684 
 
 699 
 
 m 
 
 1.560 
 
 1.682 
 
 1.639 
 
 1.809 
 
 1.734 
 
 1.796 
 
 r 
 
 .9995 
 
 .9931 
 
 .9993 
 
 .9920 
 
 .9989 
 
 .9940 
 
 10-ft-THICK COAL SEAM 
 
 a 
 
 329 
 
 667 
 
 387 
 
 739 
 
 445 
 
 786 
 
 m 
 
 1.411 
 
 1.476 
 
 1.469 
 
 1.553 
 
 1.524 
 
 1.636 
 
 r 
 
 .9999 
 
 .9963 
 
 .9997 
 
 .9950 
 
 .9995 
 
 .9939 
 
 The results given in tables 5 and 6 
 show the need for horizontal confine- 
 ment for pillar strength. The value d* 
 is defined as d* = d/(l-re), where d is 
 the depth below surface and r^ is the 
 fractional recovery by mining (i.e., 
 50 pet = 0.50). In table 5, the cohesive 
 strength in unconfined shear is 100 psi. 
 The small white area at the top of the 
 table indicates that only small depths 
 can be mined without some horizontal pil- 
 lar confinement. For angles of internal 
 
 friction of 30° or larger, a horizontal 
 stress of less than 30 pet of the verti- 
 cal is required for pillar stability away 
 from the pillar edges. For large angles 
 of friction, for instance, 50° or more, 
 the horizontal stress required even for 
 low 'a' values is only 11 pet or less of 
 the vertical stress. This means that 
 practically no horizontal confinement is 
 required, and most of the confinement 
 comes from the vertical stress loading 
 across the failure surface. 
 
TABLE 5. - Blatio of horizontal pillar 
 stress to vertical pillar stress needed 
 for pillar stability for a small value 
 of cohesive strength; 'a' = 100 psi 
 
 d*, 
 
 
 Friction 
 
 angle (()>), 
 
 deg 
 
 ft 
 
 
 
 10 
 
 20 
 
 30 
 
 40 
 
 50 
 
 60 
 
 200 
 
 0.17 
 
 
 
 
 
 
 ^-^ = 
 av 
 
 400 
 
 .58 
 
 0.36 
 
 0.20 
 
 0.09 
 
 0.03 
 
 
 600 
 
 .72 
 
 .47 
 
 .30 
 
 .17 
 
 .09 
 
 0.03 
 
 
 800 
 
 .79 
 
 .53 
 
 .34 
 
 .21 
 
 .12 
 
 .06 
 
 0.02 
 
 1,000 
 
 .83 
 
 .56 
 
 .37 
 
 .24 
 
 .14 
 
 .07 
 
 .03 
 
 1,200 
 
 .86 
 
 .59 
 
 .39 
 
 .25 
 
 .15 
 
 .08 
 
 .03 
 
 1,400 
 
 .88 
 
 .60 
 
 .41 
 
 .26 
 
 .16 
 
 .09 
 
 .04 
 
 1,600 
 
 .90 
 
 .62 
 
 .42 
 
 .27 
 
 .17 
 
 .10 
 
 .04 
 
 1,800 
 
 .91 
 
 .63 
 
 .43 
 
 .28 
 
 .18 
 
 .10 
 
 .05 
 
 2,000 
 
 .92 
 
 .64 
 
 .43 
 
 .29 
 
 .18 
 
 .10 
 
 .05 
 
 2,500 
 
 .93 
 
 .65 
 
 .44 
 
 .30 
 
 .19 
 
 .11 
 
 .05 
 
 3,000 
 
 .94 
 
 .66 
 
 .45 
 
 .30 
 
 .19 
 
 .11 
 
 .06 
 
 00 
 
 1.0 
 
 .70 
 
 .49 
 
 .33 
 
 .22 
 
 .13 
 
 .07 
 
 Next consider table 6 for a cohesive 
 strength in shear of 800 psi. While this 
 appears large, cohesion values up to 
 1,240 psi for coal have been observed in 
 the laboratory. Notice the large blank 
 areas indicating that no horizontal con- 
 finement is necessary. For angles of 
 internal friction as small as 30°, a 
 horizontal stress that is only 8 pet of 
 the vertical is required for pillar 
 
 41 
 
 TABLE 6. - Ratio of horizontal pillar 
 stress to vertical pillar stress needed 
 for pillar stability for a large value 
 of cohesive strength; 'a' = 800 psi 
 
 d*. 
 
 Friction angle ((Ji), deg 
 
 ft 
 
 
 
 10 
 
 20 
 
 30 
 
 40 
 
 50 
 
 60 
 
 200 
 
 
 
 
 
 
 
 
 400 
 600 
 
 
 
 ^=0 
 ov 
 
 
 
 
 
 800 
 
 
 
 
 
 
 
 
 1,000 
 
 
 
 
 
 
 
 
 1,200 
 
 
 
 
 
 
 
 
 1,400 
 
 0.05 
 
 
 
 
 
 
 
 1,600 
 
 .17 
 
 0.01 
 
 
 
 
 
 
 1,800 
 
 .26 
 
 .08 
 
 
 
 
 
 
 2,000 
 
 .33 
 
 .15 
 
 0.02 
 
 
 
 
 
 2,500 
 
 .47 
 
 .26 
 
 .12 
 
 0.3 
 
 
 
 
 3,000 
 
 .56 
 
 .33 
 
 .18 
 
 .08 
 
 0.01 
 
 
 
 00 
 
 .00 
 
 .70 
 
 .49 
 
 .33 
 
 .22 
 
 0.13 
 
 0.07 
 
 stability, away from the pillar edge 
 where the opening stress concentration 
 exists. For example, if a recovery of 50 
 pet is used, the d* value would be 3,000 
 for a depth of 1,500 ft. The <=° symbol at 
 the bottom of tables 5 and 6 indicates 
 the limits of horizontal constraint re- 
 quired at great depth assuming elastic 
 behavior still applies. 
 
 THE ROLE OF CONSTRAINT IN PILLAR DESIGN 
 
 WHAT IS FAILURE? 
 
 What passes for coal or rock strength 
 is often not material strength at all but 
 constraint behavior. Consider the be- 
 havior of the coalbed in figure 2 at some 
 depth D below the surface. Regardless of 
 the coal strength, the coalbed will re- 
 main intact as long as it is confined by 
 the rock strata above and below it. This 
 will be true for any depth and any mate- 
 rial. In other words, when completely 
 confined, any coal or rock appears to 
 have an 'infinite' strength. If a piece 
 of this coal or rock could be tested in 
 the laboratory, it would be readily ap- 
 parent that the constraint and not the 
 coal or rock strength was responsible for 
 this behavior. 
 
 Surface-N 
 
 ^^?<^^^<5^ 
 
 ^^^m 
 
 Rock 
 
 Vertical confining stress 
 
 i i i J, i i J, i 1 i t i- 
 
 FIGURE 2. - Confined coalbed. Any coal seam 
 of any strength at any depth constrained as shown 
 is "infinitely" strongfor a uniform vertical stress 
 
42 
 
 Another way of showing the effect of 
 constraint is to consider a single entry 
 in the coal layer as shown in figure 3. 
 If D is large enough and the coal in the 
 rib is unconfined, it will break, adja- 
 cent to the opening, for a distance that 
 is some function of the coal seam height 
 to point B. This broken zone will pro- 
 gressively provide more constraint away 
 from the rib until the combination of 
 coal strength and constraint halts the 
 breaking process. The constrained coal 
 at A is unbroken. 
 
 Next, consider several entries as shown 
 in figure 4. Each entry will have two 
 unconfined ribs in coal and an unconfined 
 roof and floor. Because of less con- 
 straint, a relatively greater volume of 
 
 Surface^N, 
 
 P^<^5^?^5^2s^J^?J^:?f^5^?^5^?^?;?:^J^?«?*5|^^ 
 
 Rock 
 
 Unconstrained at free 
 surface of opening 
 
 Constrained; 
 
 rrrr 
 
 Room 
 
 JUJ, 
 
 ^Constrained 
 
 A B C C B A 
 
 ^ 
 
 t'T't^r t 
 
 FIGURE 3. - Single entry in coal layer. Con- 
 strained coal seam is "infinitely" strong except 
 at the free surfaces and adjacent to them. 
 
 Surface^ 
 
 Rock 
 Unconstrained at free surfaces 
 
 
 Room kPiIIQ''^ Room SPillar? Room 
 
 rfT 
 
 A --Broken-' ^ 
 ^--Constrained-- 
 
 FIGURE 4. - Several entries in coal layer. As 
 the number of free surfaces increases, the overall 
 constraint decreases, and the overall coal strength 
 decreases as well. 
 
 coal will break until the combined ef- 
 fects of constraint and strength halt the 
 breaking process. As the pillars become 
 smaller, constraint plays a decreasing 
 role in pillar stability and the coal 
 strength without constraint a more impor- 
 tant role. 
 
 The many attempts to define pillar 
 strength in terms of the width-to-height 
 ratio (W/H) for the pillar imply that 
 constraint is taken as necessary to en- 
 sure pillar survival. If the coal alone 
 is strong enough to support the applied 
 load, constraint will be unnecessary and 
 the value of W/H does not enter the prob- 
 lem. What the W/H ratio represents is 
 the amount and importance 
 straint to pillar survival, 
 en a physical meaning if 
 Mohr-Coulomb behavior. 
 
 of the con- 
 
 This is giv- 
 
 expressed in 
 
 A common model for failure prediction 
 is the Coulomb (J_) behavior of soil 
 mechanics or the Mohr-Coulomb behavior of 
 the theory of elasticity. Figure 5 shows 
 the case of Coulomb failure in soil 
 mechanics. The change in shearing 
 strength with normal stress across the 
 shear surface results from frictional 
 effects, defined by the angle of internal 
 friction ()>. That is, the shearing 
 strength is the cohesive strength C in 
 shear plus the frictional strength, a^ 
 tan <t), that results from confinement of 
 the shear surface by the normal stress 
 On. If the normal stress is removed at 
 any time, the cohesive strength in shear 
 
 T5= C + CTp ton (^ 
 
 Coulomb 
 failure' 
 
 1773 
 
 Apparent 
 strength 
 
 Constraint 
 
 Material 
 I strengt h 
 
 FIGURE 5. - Coulomb failure. Apparent strength 
 as defined by the Coulomb a„ failure condition is 
 partly material strength and partly the effects of 
 constraint across the failure surface in shear. 
 
43 
 
 is still C. In other words, the apparent 
 increase in strength is only the effect 
 of one stress, the shearing stress, act- 
 ing against the confining stress. 
 
 Another way of showing that the in- 
 creased strength is not strength at all 
 is shown in figure 6, which shows the 
 frictional effect alone of one rigid sol- 
 id surface being pressed against a second 
 rigid solid surface. The resulting curve 
 is the same as that for a Coulomb failure 
 when the cohesive strength is zero. That 
 is, the total apparent strength is only 
 the result of constraint. Since the bod- 
 ies are rigid they do not deform, and ma- 
 terial strength is not considered. 
 
 Next, consider the Mohr-Coulomb stress- 
 failure relationship shown in figure 7. 
 It is assumed in this theory that the 
 
 C+cr, 
 
 Coulomb failure 
 
 Tg = CTn tan<^ , tan <^ = coefficient of friction 
 
 A Shearing stresses produced when two rigid 
 bodies are displaced parallel to the contact 
 area. 
 
 Constraint only, no 
 material strength 
 
 B The Coulomb relationship between the normal 
 and shearing stresses oi A. 
 
 FIGURE 6.. Shearing stresses and Coulomb fail- 
 ure. The Coulomb failure relationship describes 
 the effect of a shearing stress acting against a nor- 
 mal stress across the shear zone. 
 
 ,-Mohr stress 
 circle 
 
 Apparent 
 strength 
 
 FIGURE 7. - Mohr-Coulomb stress-failure rela- 
 tionship. This isof the same form as the relation- 
 ship shown in figure 6 but with a cohesive shear 
 strength added. 
 
 Mohr's stress circle will produce failure 
 if it touches the Coulomb failure line in 
 any way. The Coulomb condition is a 
 straight line as shown or a concave down- 
 ward line when defined by the Mohr's 
 stress envelope, with increasing dis- 
 placement of the circle to the right on 
 the normal stress axis. The strength of 
 a solid is generally taken to be the val- 
 ue of the shear stress at the point the 
 circle touches the Coulomb surface. How- 
 ever, in view of the fact that this is 
 part shear strength and part confinement, 
 this is an incorrect interpretation of 
 pillar behavior. Since the solid can be 
 broken with a shear stress C at any time 
 when unconstrained, this much of the 
 shear value is material strength and the 
 remainder is constraint strength. The 
 notion that the combined state of stress 
 somehow measures strength has lead to 
 confusion in the evaluation of pillar be- 
 havior, particularly with regard to 
 laboratory testing of rock and coal sam- 
 ples, where unusually high constraint 
 strength is provided by very stiff steel 
 platens or triaxial confinement. 
 
 The stress states at points A, B, and C 
 in a coal seam as shown in figure 3 are 
 defined in figure 8. If the coal has a 
 shear strength C2, the Mohr's stress cir- 
 cles A, B, and C will not reach the fail- 
 ure condition, the rib will stand intact, 
 and no constraint will be necessary for 
 pillar stability. If the strength is C], 
 the coal will break at the rib until the 
 broken coal provides enough constraint or 
 
44 
 
 FIGURE 8. - The Mohr's stress circles for 
 points A, B, and C shown in figure 3. The coal 
 with shear strength C, has failed at the rib, and 
 constraint is necessary to pillar survival. The 
 coal v/ith shear strength C2 does not fail at the 
 rib, and constraint is unnecessary to pillar survival. 
 
 the stress concentration disappears so 
 that the Mohr's stress circle B will just 
 touch the Coulomb failure condition. 
 Constraint will be necessary for pillar 
 survival. 
 
 DOES A LARGE W/H RATIO ALWAYS ENSURE 
 PILLAR STRENGTH? 
 
 At least one implicit assumption inher- 
 ent in the use of the W/H ratio as a mea- 
 sure of pillar stability is that this ra- 
 tio provides constraint to most of the 
 pillar volume away from the coal rib. In 
 part, this belief is the result of test- 
 ing flat rock or coal samples in the 
 laboratory between steel platens that 
 provide frictional constraint to the sam- 
 ple, making it almost incompressible. 
 
 Consider the relationships shown in 
 figure 9 where steel cubes load a rock 
 or coal cube. For 30 psi of vertical 
 stress in the steel, a vertical strain 
 of 1 uln/in results. If the Poisson's 
 ratio is 0.25, the horizontal strain is 
 0.25 pin/in. This is an upper bound for 
 the horizontal strain because the steel 
 block is not constrained horizontally as 
 would be the case for a steel platen that 
 is much larger than the specimen. The 
 cube of coal with a Young's modulus of 
 
 Strong 
 
 = Weak 
 
 I^t MCTy 
 
 Coal is 
 constrained 
 by shear 
 stresses 
 
 Strong 
 
 FIGURE 9.- Steel cubes loading a coal cube. 
 If the coal is confined between two more rigid ma- 
 terials such as steel platens or strong roof and 
 floor rock, the apparent strength of the coal in- 
 creases from constraint. 
 
 1 X 10^ psi or less would have a vertical 
 strain of 30 yin/in and a horizontal 
 strain of 7.5 yin/in for a Poisson's ra- 
 tio of 0.25. The strains in the coal 
 would be 30 times those in the steel. 
 The net effect is that the steel con- 
 strains the horizontal expansion of the 
 coal and gives it an apparent strength 
 that is much greater than the strength of 
 the coal when unconfined. 
 
 Next consider the relationships shown 
 in figure 10 where cubes of rock with low 
 Young's moduli, for example, 0.1 x 10^ 
 psi, are used to load the cube of coal 
 with Young's modulus of 1 x 10^ psi or 
 less. A vertical stress of 30 psi will 
 produce a vertical strain of 300 uin/in 
 in the rock cubes. If the soft roof and 
 floor are a claylike material, the Pois- 
 son's ratio may be very high, approaching 
 0.5. To be conservative, let the Pois- 
 son's ratio be 0.25. This would result 
 in a horizontal strain in the roof and 
 floor of 75 pin/in. This is an upper 
 bound because the roof and floor are con- 
 strained laterally in the mine. For this 
 
45 
 
 Weak- 
 
 Flow 
 
 U t t { ^y 
 
 Strong 
 
 Weak- 
 
 Coal is not 
 constrained 
 
 Flow 
 
 FIGURE 10. - Strong coal between weaker ma- 
 terials. If the coal is stronger than the roof and 
 floor rock or the material used to load the coal in 
 the laboratory, the coal is not constrained. In the 
 mine the roof and floor must flow for the constraint 
 to disappear. 
 
 horizontal strain to occur, the floor 
 must heave and the roof must sag. These 
 conditions are not uncommon in actual 
 mining operations. The vertical strain 
 in the coal would be 30 uin/in. The hor- 
 izontal strain in the coal for a Pois- 
 son's ratio of 0.25 would be 7.5 pin/in. 
 The horizontal strains in the roof and 
 floor would be 10 times that for the 
 coal. The net effect is that, instead of 
 confining the coal seam, the roof and 
 floor would actually pull the coal seam 
 apart. In this case the W/H ratio does 
 not indicate increased strength. 
 
 Future work on the effects of W/H ratio 
 on pillar strength should evaluate these 
 two important factors in terms of Mohr- 
 Coulomb behavior. First, the strength of 
 the coal itself, when unconfined, must be 
 determined; and second, the role of con- 
 straint or lack of it must be evaluated. 
 One way of doing this, in situ, would be 
 to use the drilling yield method devel- 
 oped in Europe. 
 
 POTENTIAL FOR USING THE DRILLING YIELD METHOD FOR DESIGNING MINE PILLARS 
 
 Kidybinski and Stranz (26) reported on 
 the drilling yield method to the U.S. De- 
 partment of the Interior. In this method 
 a hole drilled into the coal seam is used 
 to study the conditions of stress in the 
 seam. The volume of drill cuttings is 
 expected to be 2 to 3 L/m of drill-hole 
 length when the hole diameter is 42 mm. 
 If the hole squeezes shut during drill- 
 ing, indicating that the coal is not 
 strong enough to handle the stress, the 
 volume of cuttings will increase. If the 
 volume is 6 L or more, a serious coal 
 bump condition exists. The method has 
 been in field use since the 1960's ( 27 ) 
 in the Federal Republic of Germany, since 
 the 1950's in the U.S.S.R. , since 1965 in 
 Poland, and since 1968 in Yugoslavia. A 
 2.5-hp motor was used, and the stalling 
 behavior of the motor was also a part of 
 the analysis. 
 
 The use of the drilling yield method 
 could be helpful in the design of 
 mine pillars by establishing the broken 
 or squeezing zone locations or depths. 
 The concept is shown in figure 11. In 
 
 part A, a point in the coal seam de- 
 noted by P is confined and stable. If a 
 hole is drilled through this location 
 (part B ) , the constraint is removed on 
 the hole surface. The tangential stress 
 
 Roof 
 
 Floor 
 A Confined coal seam is stable at point P. 
 
 Roof 
 
 Drill 
 hole- 
 
 Floor 
 B Unconfined coal seam may be unstable 
 at point P. 
 
 FIGURE 11. - Use of drilling yield method to 
 establish pillar dependence on constraint for 
 survival. 
 
46 
 
 concentration on the hole boundary would 
 be in the range of 2 to 3 times the ver- 
 tical stress for an elastic condition. 
 If the coal breaks around the hole, it 
 does not mean that the pillar is unstable 
 but that the coal strength is less than 2 
 to 3 times the most compressive prin- 
 cipal stress. With breaking, the stress 
 concentration will be reduced and in the 
 
 limit approach the coal seam stress. 
 The breaking should then stop. If it 
 does not , this indicates that the coal 
 strength is inadequate to support the 
 stresses when unconfined. The volume of 
 cuttings is therefore an indicator of the 
 pillar stability versus depth from the 
 rib. 
 
 SUMMARY AND CONCLUSIONS 
 
 The number of studies of pillar behav- 
 ior based on testing of samples has in- 
 creased rapidly during the last 25 yr. 
 The concepts on which many, if not most, 
 of these tests were based often date 
 back to 1912 or before. That is, the 
 number of testing procedures or methods 
 of analysis used has not kept pace with 
 the number of studies. Recent experimen- 
 tal testing has emphasized the effects of 
 sample size and conditions of testing, 
 including constraint. One reason for the 
 slow progress in pillar design is that 
 the separate effects of coal strength 
 and the apparent increase in strength re- 
 sulting from confinement have not been 
 appreciated. 
 
 In pillar design, if the pillar must 
 have constraint from the roof and floor 
 to survive, the pillar becomes unstable 
 if such constraint disappears with time 
 or change in mine geometry. If the coal 
 can survive in pillar form without con- 
 straint, it is more likely to remain sta- 
 ble when constraint conditions are 
 changed. 
 
 The practice of assuming that a flat 
 coal pillar will be "infinitely" strong 
 because this is the case for laboratory 
 testing of coal between steel platens 
 should be examined carefully. In such 
 tests, the strength of the platens and 
 not of the coal is determined, and there 
 is no correlation to mining conditions 
 with roof and floor of rock, sometimes 
 rock that is not very strong or struc- 
 turally stable. 
 
 It is commonly considered that the re- 
 cent emphasis on constraint is new, but 
 in fact, any equation using W/H relation- 
 ships implies this condition. 
 
 Most of the pillar strength results 
 from constraint across the failure sur- 
 face by normal compressive stresses 
 according to the Mohr-Coulomb stress- 
 failure criteria. This constraint is 
 supplied for the most part by the ver- 
 tical component of stress from the over- 
 lying rock. This is particularly true if 
 the angle of internal friction is 30° or 
 more. For angles as large as 50° or 60° 
 only a very small horizontal stress com- 
 ponent is needed for pillar stability. 
 In addition, the small value of the un- 
 confined shear strength often assumed for 
 in situ behavior is largely responsible 
 for the very different results obtained 
 in the laboratory and in the mine. 
 
 Nearly all the theories with width- 
 to-height relationships can be represent- 
 ed equally well using the Mohr-Coulomb 
 stress-failure models. There is no magic 
 in the width-to-height ratios for pillar 
 design. These occur naturally when the 
 strength increases with depth and a given 
 entry size is used. 
 
 The cores of pillars are confined for 
 the most part not by the horizontal pil- 
 lar stresses, as in the constrained core 
 concept, but through the action of the 
 vertical stress away from the pillar 
 edges. The effect that exists with 
 respect to horizontal stress is one of a 
 confined edge rather than a core. 
 
47 
 
 REFERENCES 
 
 1. Coulomb, C. A. Essai sur une Ap- 
 plication des Regies des Maximis et 
 Minimis a Quelques Problems de Statlque 
 Relatifs a 1 'Architecture (Tests for the 
 Application of Maxima and Minima Rules 
 and Statistical Problems Relative to Ar- 
 chitecture). Mem. Acad. R. Pres. Divers 
 Savants (Paris), v. 7, 1773. 
 
 2. Wilson, A. H. , and D. P. Ashwin. 
 Research Into the Determination of Pillar 
 Size. Part I. An Hypothesis Concerning 
 Pillar Stability. Min. Eng. (N.Y.), v. 
 131, June 1972, p. 409-417. 
 
 3. Warden, K. , and Partners (Newcas- 
 tle, United Kingdom). Guidelines for 
 Mining Near Surface Waters (contract 
 H0252021). BuMines OFR 30-77, 1977, 59 
 pp.; NTIS PB 264 729. 
 
 4. Babcock, C. C, and V. E. Hooker. 
 Results of Research To Develop Guidelines 
 for Mining Near Surface and Underground 
 Bodies of Water. BuMines IC 8741, 1977, 
 17 pp. 
 
 5. Panek, L. Estimating Mine Pillar 
 Strength From Compression Tests. Trans. 
 Soc. Min. Eng. AIME, v. 268, 1980, 
 pp. 1749-1761. 
 
 6. Babcock, C. 0., T. Morgan, and K. 
 Haramy. Review of Pillar Design Equa- 
 tions Including the Effects of Con- 
 straint. Paper in 1st Annu. Conf. on 
 Ground Control in Mining, WV Univ. , Mor- 
 gantown, WV, July 1981. Dep. Min. Eng., 
 WV Univ., 1981, pp. 23-34. 
 
 9. Johnson. Materials of Construc- 
 tion. 1897. (Cited in Gonnerman, H. F. 
 Effect of Size and Shape of Test Speci- 
 mens on Compressive Strength of Concrete. 
 Structural Materials Res. Lab., Bull. 16, 
 Oct. 1925, 18 pp.; ASTM, v. 25, pt. 2, 
 1925.) 
 
 10. Carpenter, R. C. Article in Sib- 
 ley J. Eng., V. 16, No. 3, Dec. 1901, 
 p. 105. (Cited in reference 11.) 
 
 11. Bunting, D. Chamber-Pillars in 
 Deep Anthracite-Mines. Trans. AIME, v. 
 42, 1912, pp. 236-245. 
 
 12. Griffith, W. , and E. T. Conner. 
 Mining Conditions Under the City of 
 Scranton, Pa. BuMines B 25, 1912, 89 pp. 
 
 13. Greenwald, H. P., H. C. Howarth, 
 and I. Hartmann. Experiments on Strength 
 of Small Pillars of Coal in the Pitts- 
 burgh Bed. BuMines TP 605, 1939, 22 pp. 
 
 14. . Experiments on Strength of 
 
 Small Pillars of Coal in the Pittsburgh 
 Bed. BuMines RI 3575, 1941, 7 pp. 
 
 15. Steart, F. A. Strength and Sta- 
 bility of Pillars in Coal Mines. Chem. , 
 Metall. , and Min. Soc. S. Afr. , v. 54, 
 1954, pp. 307-325. 
 
 16. Gaddy, F. L. A Study of the Ulti- 
 mate Strength of Coal as Related to the 
 Absolute Size of Cubical Specimens Test- 
 ed. Bull. VA Polytech. Inst, and State 
 Univ., No. 49, 1954, pp. 1-27. 
 
 7. Vicat, L. J. Researches on Physi- 
 cal Phenomena Which Precede and Accompany 
 Rupture or Deformation of a Certain Class 
 of Solids. Ann. Fonts et Chaussees, pt. 
 2, 1833, p. 201. 
 
 17. Holland, C. T. The Strength of 
 Coal in Mine Pillars. Paper in Proc. 6th 
 Symp. on Rock Mechanics, Univ. MO, Rolla, 
 MO, Apr. 1964. Univ. MO~Rolla, 1964, 
 pp. 450-456. 
 
 8. Bauschinger, J. Mitteilungen aus 
 dem Mechanisch-Technischen Laboratorium 
 der K. Technischen Hochschule in Munchen 
 (Reports From the Mechanical-Technical 
 Laboratories of the K. Technical College 
 in Munich). V. 6, 1876. 
 
 18. Evans, I., and C. D. Pomeroy. The 
 Strength, Fracture and Workability of 
 Coal. Pergamon, 1966, 277 pp. 
 
48 
 
 19. Salamon, M. D. G. , and A. H. 
 Munro. A Study of the Strength of Coal 
 Pillars. J. S. Afr. Inst. Min. and 
 Metall,, V. 68-2, Sept. 1967, pp. 55-67. 
 
 20. Bieniawski, Z. T. In Situ 
 Strength and Deformation Characteristics 
 of Coal. Eng. Geol. (Amsterdam), v. 2, 
 1968, pp. 325-340. 
 
 21. 
 
 In Situ Large Scale Test- 
 Paper in Proc. Conf. on In 
 and Rock. 
 
 ing of Coal. 
 
 Situ Investigations on Solids 
 
 Brit. Geotech. Soc. , 1969, pp. 67-74. 
 
 22. Van Heerden. In Situ Determina- 
 tion of Complete Stress-Strain Character- 
 istics for 1.4 M Square Coal Specimens 
 With Height to Width Ratios of Up to 3.4. 
 Rep. CSIR (S. Afr.), No. ME 1265, 1974, 
 p. 30. 
 
 23. Wagner, H. Determination of Com- 
 plete Load Deformation Characteristics 
 of Coal Pillars. Paper in Proc. 3d 
 Int. Conf. on Rock Mechanics, Denver, 
 CO. Natl. Acad. Sci., v. 11-B, 1974, 
 pp. 1076-1082. 
 
 24. Hustrulid, W. A. A Review of Coal 
 Pillar Strength Formulas. Rock Mech. , 
 V. 8, 1976, pp. 115-145. 
 
 25. Skelly, W. A., J. Wolgamott, and 
 F. Wang. Coal Mine Pillar Strength and 
 Deformation Prediction Through Laboratory 
 Sample Testing. Paper in Proc. 18th 
 Sjnmp. on Rock Mechanics, Keystone, CO, 
 June 1977. CO School Mines Press, 1977, 
 pp. 2B5-1 to 2B5-5. 
 
 26. Kidybinski , A., and B. Stranz. 
 Coal Mine Safety Hazards Related to Rock 
 Stresses. ' Res. Prog. Rep. — P. L. 480 to 
 U.S. Dep, Interior by Polish Central 
 Inst, of Mining (Katowice, Poland), Dec. 
 1972, 56 pp. 
 
 27. Jahn, H. (Identification and Dis- 
 posal of Dangerous Stresses in the Coal- 
 side of a Gateroad in the Coal Bump 
 Prone Seam Sonnenschein. ) Gluckauf, 
 1965, p. 101. 
 
49 
 
 UNDERHAND CUT-AND-FILL STOPING FOR ROCK BURST CONTROL 
 By F. Michael Jenkins^ and K. Robert Dormant 
 
 ABSTRACT 
 
 The occurrence of rock, bursts in deep 
 metal and nonmetal mines presents a major 
 hazard to their safe and economical oper- 
 ation. As part of a Bureau of Mines pro- 
 gram to reduce rock bursts in deep vein 
 mining, a project was conducted in the 
 Coeur d'Alene mining district of Idaho to 
 evaluate and demonstrate the mining of a 
 destressed sill pillar using underhand 
 cut-and-fill methods. A 50-ft sill pil- 
 lar in a burst-prone area was precondi- 
 tioned by drilling and blasting vertical 
 
 holes in the ore. After raises were 
 driven through the destressed pillar to 
 the level above, it was mined by the un- 
 derhand cut-and-fill method. The con- 
 clusion from this demonstration is that 
 the combination of ore preconditioning 
 and underhand mining resulted in great- 
 ly improved rock burst and ground con- 
 trol, allowing safe and efficient mining 
 of a potentially hazardous, burst-prone 
 pillar. 
 
 INTRODUCTION 
 
 The occurrence of rock bursts in deep 
 metal and nonmetal mines presents a major 
 hazard to their safe and economical oper- 
 ation. A rock burst is the sudden vio- 
 lent release of strain energy stored in 
 the rock as a result of mining. The 
 causes of rock bursting can be traced to 
 the geometry of mine openings and rock 
 characteristics. Strong, brittle rock in 
 high stressed areas of the mine has a 
 high potential for bursting. Bursting 
 associated with stoping usually occurs in 
 sill pillars, converging stopes, initial 
 or subdrifts, and raises. 
 
 As part of a research program to re- 
 duce rock bursts in deep vein mining, a 
 
 project was conducted under a Bureau con- 
 tract to evaluate and demonstrate the 
 mining of a destressed sill pillar using 
 underhand cut-and-fill methods. 3 The 
 project was carried out in two phases: 
 first, to study the feasibility and cost 
 effectiveness of the method and present 
 design recommendations, and second, to 
 enlist the cooperation of a mine and dem- 
 onstrate that the method could be used to 
 reduce rock bursts. This report presents 
 some background information from phase I 
 and the findings and conclusions of phase 
 II as demonstrated by the test stopes, 
 including cost, production, and safety 
 studies. 
 
 BACKGROUND 
 
 In the Coeur d'Alene mining district of 
 northern Idaho, where lead-zinc silver 
 veins are being mined to 8,000 ft below 
 the surface, rock bursting continues to 
 be a severe operational problem. Finding 
 a means of controlling rock bursting has 
 been the goal of research for almost 80 
 yr. However, the problem still persists 
 
 ^Mining engineer. 
 ^Supervisory mining engineer. 
 Spokane Research Center, Bureau of 
 Mines, Spokane, WA. 
 
 because of its complexity and the diffi- 
 culty of finding practical solutions that 
 can be applied economically. 
 
 The predominant mining method in the 
 Coeur d'Alene mining district is overhand 
 
 ^Bush, D. D., W. Blake, and M. P. 
 Board. Evaluation and Demonstration of 
 Underhand Stoping To Control Rock Bursts. 
 BuMines contract H0292013; for informa- 
 tion, contact F. M. Jenkins, TPO, Spokane 
 Research Center. 
 
50 
 
 cut and fill. The general plan is to de- 
 velop horizontal levels from vertical 
 shafts, produce ore from stopes that are 
 mined upwards toward the mined-out level 
 above, and backfill each stope cut with 
 mill tailings. The result is an ever- 
 decreasing pillar of ore between the lev- 
 el above and the miners working in the 
 stope, with lateral stress concentrated 
 in the pillar. Structural failure of the 
 sill pillar often produces rock bursts. 
 
 Experience indicates that sill-pillar 
 bursting usually begins when the pillar 
 height is reduced to about 80 ft, with 
 the greatest frequency and severity oc- 
 curring at pillar heights of 40 to 30 ft. 
 These bursts often inflict great damage 
 to the stope as well as to haulage drifts 
 and crosscuts above and below. A large 
 burst can displace more than 1,000 tons 
 of rock, heave the drift floor, and break 
 numerous caps and posts. The results 
 are extensive repair costs and loss 
 
 of production. Statistics indicate that 
 failure associated with sill pillars ac- 
 counts for more than 60 pet of all rock 
 bursts.^ 
 
 Ore preconditioning, the destress 
 blasting of a pillar prior to mining, has 
 been shown to effectively reduce rock 
 bursts. 5 Blasting changes the character- 
 istics of the ore from a strong, brittle 
 material to a yielding material incapable 
 of storing strain energy. To be effec- 
 tive, however, the preconditioning must 
 thoroughly fragment the ore. This cre- 
 ates a hazard to those working in over- 
 hand stopes because they are working be- 
 neath fragmented rock. The underhand 
 cut-and-fill mining method was developed 
 in Canada to mine pillars of ore similar 
 to those created by preconditioning. ^ A 
 combination of the two techniques (pre- 
 conditioning and underhand cut and fill) 
 seemed an attractive solution to recover- 
 ing burst-prone sill pillars. 
 
 PHASE I— EXAMINATION AND DESIGN RECOMMENDATIONS 
 
 During phase I of this project, the un- 
 derhand cut-and-fill practice was ex- 
 amined at two mining operations in Canada 
 and two in the United States where the 
 method was being used for primary pro- 
 duction and for pillar recovery under 
 difficult ground conditions. An artist's 
 conception of a typical underhand cut- 
 and-fill stope is shown in figure 1. 
 
 Stress analyses were made of mining 
 by overhand cut and fill and by under- 
 hand cut and fill to compare their ef- 
 fects on ground control in the Coeur 
 d'Alene District. 
 
 To examine the bursting potential of 
 underhand cut-and-fill stoping, the ini- 
 tial simulation assumes five end-to-end 
 stopes mined in a flat-back arrangement 
 toward an unmined level below. This sim- 
 ulation ignores active mining on multiple 
 levels; thus, no sill pillar is created. 
 The maximum stress (30,000 to 50,000 psi) 
 occurs at the stoping horizon. With no 
 sill pillar being created, the stress 
 does not change as mining progresses. 
 Because of the constant rate of energy 
 
 release, little bursting should be en- 
 countered with a flat-back underhand 
 method if no pillars are created. 
 
 Because mining is not normally confined 
 to one level, more realistic, multilevel 
 mining sequences were examined. A series 
 of simulations with simultaneous mining 
 on three levels was made. The potential 
 for bursting approaches that of the over- 
 hand method soon after multilevel mining 
 occurs. In this case, the pillar stress 
 is between 55,000 and 60,000 psi or near- 
 ly that amount induced in the sill pillar 
 
 ^McLaughlin, W. C, G. C. Waddell, 
 and J. C. McCaslin. Seismic Equipment 
 Used in Rock Burst Control in the Coeur 
 d'Alene Mining District, Idaho. BioMines 
 RI 8138, 1976, 27 pp. 
 
 ^Karwoski, W. J., W. C. McLaughlin, 
 and W. Blake. Rock Preconditioning To 
 Prevent Rock Bursts — Report on a Field 
 Demonstration. BuMines RI 8381, 1979, 
 45 pp. 
 
 ^Society of Mining Engineers of AIME. 
 Underground Mining Methods Handbook. New 
 York, 1982, p. 631 . 
 
51 
 
 -L 
 
 ^9^ 
 
 <*'>' iw'TSi'- 
 
 FIGURE 1. - Artist's conception of an underhand cut=and-fill stope. 
 
 for the overhand case. Though energy re- 
 lease rates are lower, bursting will not 
 be eliminated by underhand mining alone. 
 However, underhand cut and fill offers 
 distinct advantages where destressing is 
 used in rock burst control: (1) Precon- 
 ditioning techniques are easily applied 
 since destress holes can be drilled ver- 
 tically in the ore body, (2) damage dur- 
 ing preconditioning will generally be 
 less severe and will not produce the mas- 
 sive caving that often occurs during con- 
 ventional destress blasting, and (3) 
 blast-induced caving and timber damage 
 during mining are eliminated. 
 
 Following the mine examinations and 
 stress analysis studies, three alternate 
 
 methods of underhand cut-and-fill stoping 
 were suggested for the Coeur d'Alene. A 
 cost-and-production estimate was made for 
 the three methods as well as for typical 
 timbered cut and fill. They indicated 
 that efficiencies (tons per worker-shift) 
 for the underhand cut-and-fill method 
 were comparable to practices in use, and 
 that considerable cost savings could be 
 obtained over conventional timbered cut 
 and fill, owing to the reduction of tim- 
 ber required for support. The cost esti- 
 mate found that underhand cut and fill 
 would not, however, be an economical al- 
 ternative for primary production at all 
 mines. 
 
52 
 
 PHASE II~FIELD TEST 
 
 The phase I study indicated underhand 
 cut-and-fill stoping could be a cost- 
 effective method of rock burst control 
 when used in conjunction with ore pre- 
 conditioning. During phase 11, the meth- 
 od was tested in an operating mine, and 
 productivity, relative cost, and ground 
 control were evaluated. This section 
 discusses the field test, from site se- 
 lection through mining of the final cut. 
 
 SITE SELECTION AND MINING PLAN 
 
 The results of the phase I study were 
 presented to the major mining companies 
 in the Coeur d'Alene District. The re- 
 sponse was generally positive, although 
 most of the companies questioned the eco- 
 nomics of the system. One company was 
 interested enough to cooperate with the 
 Bureau and Terra Tek, Inc., in a field 
 test. An ideal test area was selected in 
 an unusual manner. Miners were preparing 
 a relatively stable area for the test 
 when a major burst occurred in a 50-ft 
 sill pillar in another part of the mine. 
 Based on past experience, management pre- 
 dicted that additional pillar bursts of 
 possibly greater magnitude were likely to 
 occur during mining of this sill pillar. 
 They chose this site for testing the un- 
 derhand cut and fill combined with ore 
 preconditioning. 
 
 A mining plan, illustrated in figure 2, 
 was prepared by Terra Tek and submitted 
 to the Bureau and the mine for approval. 
 The plan called for the following se- 
 quence of events: 
 
 1. Complete repair of the rock burst 
 and complete the present stope cut. 
 Three of the stopes in the pillar were to 
 be brought to the same elevation, thus 
 creating a flat back. One stope had been 
 dropped because of poor-grade ore. 
 
 2. Clean out and repair the haulage 
 lateral above the pillar. Timbersets 
 were to be repaired or replaced, slabs 
 were to be removed, and track was to be 
 repaired where possible. 
 
 3. Prepare the haulage drift for ce- 
 mented fill after cleanup and repair. A 
 floor mat, consisting of 12- by 12-in 
 stringers, lagging, 4- by 4-in No. 8 wire 
 mesh, and a layer of woven polyethylene 
 cloth (Fabrene),^ would be laid over the 
 entire 600-ft length of drift. At 100-ft 
 intervals, fill fences would be con- 
 structed from light timber, wire mesh, 
 and Fabrene to limit the extent of any 
 individual pour. 
 
 4. Drill the pillar with vertical de- 
 stress holes and blast using ammonium 
 nitrate-fuel oil (ANFO) or water gel. 
 
 5. Perform pre- and post-precondi- 
 tioning seismic velocity surveys through 
 the pillar to evaluate effectiveness of 
 preconditioning. 
 
 6. Fill the haulage drift with ce- 
 mented sandfill. An 8:1 sand-to-cement 
 dry weight ratio was chosen, based on ex- 
 perience in other mining districts. 
 
 7. Drive three-cap raises (manway and 
 timber slide, plus two joker chutes) 
 through to the haulage drift. Raise 
 through cemented fill to provide second- 
 ary access to the stopes. Provide air- 
 tight raise covers to avoid upsetting the 
 ventilation flow. 
 
 8. Leave the present overhand mining 
 floor open for access way between stopes. 
 
 9. Begin underhand cut-and-fill min- 
 ing by breasting beneath the haulage- 
 level cemented fill. Post or timber be- 
 neath the haulage floor mat if necessary, 
 making the stope as narrow as possible. 
 Leave the raises open for secondary es- 
 cape as mining progresses downward. 
 
 10. Prepare the floor of each cut with 
 caps on 6-ft centers, lagging, wire mesh, 
 and Fabrene. Pour cemented fill with 8:1 
 
 ^Reference to specific products does 
 not imply endorsement by the Bureau of 
 Mines. 
 
53 
 
 Mined and filled 
 2 cap raise 2 cap raise 
 
 3 cap raise 
 
 Level 
 
 Sill pillar 
 
 I 
 
 / 
 
 I Ifcut 11 II II 
 
 nn 
 
 IZL 
 
 3ui 
 
 &^CutJ4_Lr 
 
 T 
 
 -I II 
 
 TTTT 
 
 I 
 
 Undercut and fill stopes 
 
 Mined and filled 
 
 TTTT 
 
 ^JaV 
 
 :tft 
 
 Level 
 
 Longitudinal section or undercut and fill stopes 
 
 FIGURE 2. " General layout of pillar recovery plan. 
 
 m 
 
 sand-to-cement ratio (by weight) to a 
 depth of 3 to 5 ft, followed by uncement- 
 ed tailings for the remainder of each 10- 
 ft cut. 
 
 period prior to blasting the destress 
 round. A second group of instruments was 
 installed during preparation of the de- 
 stress round. 
 
 11. Instrument each cut with closure 
 extensometers and fill pressure cells to 
 monitor load-displacement behavior of the 
 fill. 
 
 12. Take the last cut by end slicing. 
 
 Preparations for the underhand cut- 
 and-fill experiment began in November 
 1980 based on the mining plan. 
 
 INSTRUMENTATION 
 
 The instrumentation program was aimed 
 at quantifying the stress and displace- 
 ment behavior of the wall rock, ore body, 
 and fill before and after the destress 
 blasting as well as during subsequent 
 mining. 
 
 Initial closure and stress change in- 
 struments were installed on the upper 
 level during cleanup and repair of the 
 September 1980 rock burst damage. These 
 were to supply baseline data during 
 leveling of the stopes and during the 
 
 Instrumentation performance was mixed. 
 The most consistent and useful data were 
 obtained from the closure points and 
 stope closure extensometers. This clo- 
 sure data nicely illustrated the effects 
 of preconditioning and pillar behavior 
 before and during mining. The stress 
 meters provided qualitative data on 
 stress conditions in the walls surround- 
 ing the test stopes, which could be of 
 possible use in a modeling study. How- 
 ever, there were too few gauges for a 
 quantitative evaluation of the stress 
 conditions in the pillar. The soil pres- 
 sure cells placed in the fill proved 
 disappointing. The gauges are thought 
 to have suffered from electronic, envi- 
 ronmental, and, possibly, instrument- 
 construction problems. Little quantita- 
 tive data were produced by the pressure 
 cells prior to failure. One point is 
 evident from this test: Hand-measurable 
 instruments, such as the tape extensom- 
 eter, provide the most reliable data in a 
 harsh mining environment. 
 
54 
 
 SITE PREPARATIONS 
 
 Because the mine had no facilities for 
 adding cement to the sandfill (mill tail- 
 ings), a small cementing system was de- 
 signed and constructed on-site. The 
 mine's existing sandfill system consisted 
 of two major parts (fig. 3): the surface 
 sand-pumping facilities located within 
 the mill and underground sand storage and 
 distribution system. At the mill, a du- 
 plex piston pump was used to pump the 
 sandfill slurry (40 to 50 wt pet solids) 
 through 11,000 ft of 3-in line to the 
 2000 level of the mine. At the 2000 
 level, the slurry was discharged to a 
 cyclone where the underflow was approx- 
 imately 53 to 60 pet pulp density and 
 contained a high percentage of fines. 
 The high percentage of fines bears di- 
 rectly on the ability to achieve a good 
 sand-cement set. 
 
 After passing through the cyclone, the 
 underflow was dumped to a storage tank 
 (the storage tank was blasted out of 
 country rock and lined with shotcrete) 
 and air-agitated. The sandfill was 
 gravity-drained from the tank and dis- 
 tributed to the required stope through 
 3-in black pipe with victualic-style 
 couplings. 
 
 Ill Teea 
 
 Gankier-Denver 
 duplex piston 
 point 
 
 11,000 ft 
 
 3-ln sch 40 pipe 
 
 at 460 pal 
 
 To stopes above 
 2000 level 
 
 Gardner-Denver 
 -^ duplex piston pump 
 at 500 psi 
 
 Overflow — 1 — , 
 to-*— I 
 tailings I I 
 
 ^T^ Cyclone underflow 
 1 at 55-60 pet 
 
 Sand storage 
 
 Cement line 
 from surface 
 plant 
 
 To stopes 
 at 90 tons/h 
 solids 
 
 2000 level 
 
 FIGURE 3, - Schematic of existing sandfill system. 
 
 A study was conducted to determine the 
 most cost-effective method of introducing 
 cement into the sandfill for distribution 
 to the stopes below the 2000 level. Be- 
 cause of operational problems (tramming 
 and inability to mix cement in the under- 
 ground storage tank) , as well as the cost 
 of handling and preparing large tonnages 
 of cement underground, the decision was 
 made to place the cement slurrying and 
 pumping facilities outside the mine. 
 
 The cementing system can be divided in- 
 to two component parts: cement handling 
 and cement pumping. The components are 
 diagramed in figure 4. 
 
 The cement-handling system is that por- 
 tion of the cementing system that stores 
 bulk cement, delivers it at a desired 
 rate, and slurries it to the desired bulk 
 density. The functions of the handling 
 system are — 
 
 1. Bulk storage of up to 60 dry tons 
 of cement . 
 
 2. Delivery of from to 5.5 cfm bulk 
 cement for slurrying, with adjustment for 
 any range in between. 
 
 3. Capacity for slurrying up to 3,000 
 gal of a 50-pct-solids cement slurry. 
 
 Variable speed 
 rotary valve feeder 
 
 3,500-gBl 
 slurry tank 
 
 Butterfly valve 
 
 Water flush , , ^ 
 
 Accumulator 
 Strainer box 
 
 Suction r 
 
 -Dump to tank 
 
 Plug valves 
 
 stabilizer' 
 Ash 2-in 
 charging pump 
 
 Emergency dump-^ 
 to tailings 
 
 2-in mi?\ 
 11.000 ft O 
 
 Wilson-Snyder 
 43-85T triplex 
 plunger pump 
 
 to 2000 
 
 level sandllne 
 
 FIGURE 4, - Schematic of cement-handling system. 
 
55 
 
 The pumping system referred to here in- 
 cludes all pumps, valves, pipelines, and 
 accessory equipment for delivering the 
 cement slurry to the 2000 level. The 
 governing functions of the system are — 
 
 1. A system capable of pumping at a 
 rate of 30 to 60 gpm at pressures to 600 
 psi during normal operation. 
 
 2. In the event of a need to stop 
 pumping into the sandline for short peri- 
 ods of time (for example, sandline breaks 
 in shaft), the pump should be capable of 
 displacing less than 5 gpm of slurry. 
 
 3. The underground line must be capa- 
 ble of being flushed both from outside 
 with fresh water, or from inside the mine 
 in the event of pump failure. 
 
 The design of the cementing system was 
 completed in late July of 1980, and or- 
 ders were initiated for components in 
 August. The initial testing of the sys- 
 tem was completed in January 1981. After 
 extensive use of the system, the normal 
 operating conditions for a slurry pulp 
 density of 50 pet were a flow rate of 50 
 gpm at a discharge pressure of 450 psi. 
 
 During construction of the cementing 
 system, preparation of the floormat in 
 the sill drift was completed. When this 
 drift was originally driven, 12- by 12-in 
 stringers were placed at each rib-floor 
 intersection and run parallel to the 
 drift axis for its entire length. The 
 posts for the drift caps were footed on 
 these stringers. The floormat was con- 
 structed by placing lagging across the 
 stringers and covering the lagging with a 
 single layer of 4- by 4-in wire mesh 
 (fig. 5) followed by a single layer of 
 Fabrene. As no rail track existed, all 
 timber was hand-trammed into the drift. 
 The rock burst had reduced the drift sec- 
 tion to less than 6 ft in height and 
 width in certain areas and eliminated 
 air and water services. Slabs pulled 
 loose from the walls were broken and re- 
 moved by hand. Consequently, repairs and 
 mat preparation required approximately 
 1 month. 
 
 4- by 4-ln No. 8 or doublg 
 layer of 6- by 6-ln No. 10 wire 
 
 10- or 12-in lagging 
 -12- by 12-ln stringer 
 
 FIGURE 5. - Floormat preparation in over- 
 lying drift. 
 
 Shortly after completion of the ce- 
 menting system, the drift was filled 
 in a series of five individual pours, 
 each 100 to 200 ft in length. Owing to 
 poor availability of sandfill, this pro- 
 cess required approximately 1 month for 
 completion. 
 
 By end of January 1981, stope prepara- 
 tions were completed. The three stopes 
 had been brought to the same elevation 
 and were sandfilled (with the exception 
 of each raise area) to within 4 ft of the 
 back. It was decided to leave this 4-ft 
 access way between stopes to provide ease 
 of movement during mining and simplify 
 preconditioning of the pillar. Each 
 raise area was timbered in preparation 
 for raise driving from the stope below. 
 
 The original preconditioning design 
 called for a series of 35-ft-long, 2- 
 1/4-in-diam vertical blastholes drilled 
 upward on 6-ft centers the entire length 
 of the stope block. Based upon the un- 
 derground crew's recommendation, 10-ft 
 spacing was adopted, and the holes were 
 drilled with jacklegs and stopers using 
 1-in rope-thread steel with 2-in and 2- 
 1/4-in cross bits. Two to four drillers, 
 working on a single-shift basis, required 
 less than 2 weeks to drill the 29 holes. 
 No holes were drilled at each future 
 raise location for fear that blasting 
 would create a slabby back. 
 
 Prior to shooting the destress round, a 
 seismic velocity survey was performed. 
 
56 
 
 The data showed P-wave velocities averag- 
 ing about 14,000 ft/s. Higher veloci- 
 ties, indicating either more competent 
 rock, stress concentrations, or both, 
 were recorded at the ends of the stope. 
 The waves passed through the west area at 
 significantly lower velocities. This 
 agrees well with the highly broken nature 
 of the ore body observed in the stope in 
 the vicinity of the September 1980 rock 
 burst. In general, the survey velocities 
 indicated that both ends of the sill pil- 
 lar (not affected by the prior burst) had 
 the highest probability of future burst- 
 ing. The destress round was loaded and 
 shot on March 3, 1981. 
 
 Following preconditioning, the mining 
 crews were returned to the three stopes 
 and began raising through the sill pil- 
 lar. The mining plan called for the 
 driving of three-cap raises consisting of 
 manway, timber slide, and two joker 
 chutes above each previous raise. The 
 underground crew, based upon past experi- 
 ence, felt that driving three-cap raises 
 might open too much ground, increase the 
 risk of rock bursting, and create a slab- 
 by back. A compromise was reached to 
 drive two-cap raises above the two west 
 stopes and a three-cap raise in the east 
 stope, where conditions were generally 
 better. The west stopes would be ser- 
 viced from above, and the east stope from 
 below. The geometry of the raises and 
 the stopes and the preconditioning drill 
 pattern are shown in figure 6. 
 
 LavBl 
 
 ; I Destress holes jy' Plller jf 
 
 Ik^iilllllllllllll! Ml> 3s>. 
 
 FIGURE 6. • Drilling pattern for pillar preconditioning. 
 
 The driving of the raises proved to be 
 fairly difficult in B and C stopes, but 
 was accomplished with little difficulty 
 in A. The rock burst in the B-raise area 
 had left a badly broken ore body and a 
 slabby back. C-raise was located in 
 fairly competent ground, but an inexperi- 
 enced crew overloaded the raise rounds, 
 shattering the back and creating a prob- 
 lem. Raising required approximately 1 
 month in A stope and 1-1/2 months in B 
 and C. 
 
 For procedure evaluation, it is noted 
 that the miners had some work reserva- 
 tions, and absenteeism was a problem. 
 Also, the mining method conventionally 
 employed at the mine and the particular 
 site geometry necessitated some of the 
 above-mentioned special preparations. 
 The haulage drift above the sill pillar 
 had been driven in the vein and, there- 
 fore, had to be filled before underhand 
 stoping could begin. In addition, raises 
 had to be driven through the sill pillar 
 because the mine used blind stoping. 
 This would not be required where raises 
 are developed from level to level before 
 stoping commences. 
 
 MINING 
 
 Stoping beneath the sandfill of the 
 sill drift began in earnest in May 1981. 
 Drifting was accomplished by breasting 
 from the raise in each direction using a 
 modified V-cut (fig. 7). During the ini- 
 tial cut, timber sets were stood beneath 
 the floormat because of slimes, old 
 track, wood, and loose rock beneath the 
 mat. Raise timber required excessive re- 
 pairs because of high closure rates. 
 This proved to be a continuous problem 
 throughout the project and resulted in 
 much lost time. Mining the first cut re- 
 quired approximately 1-1/2 months in A 
 and 2 months in B and C stopes. During 
 mining of the first cut, a moderate rock 
 burst occurred at some distance out in 
 the wall below A and B stopes. Only 
 minor damage was seen in the stopes, pri- 
 marily loose slabs shaken down in the 
 access way at the bottom of the pillar. 
 
57 
 
 o 
 
 o 
 o 
 
 c 
 o 
 
 3 
 
 o- 
 o 
 (0 
 
 -10' 
 
 'I , , ' , t_ 
 
 CROSS SECTION 
 
 LONGITUDINAL SECTION 
 
 FIGURE 7. - Modified V-cut. 
 
 The stopes were prepared for sandfill- 
 ing of the first cut. A nominal 1-ft 
 layer of broken muck was leveled on the 
 floor as a blast cushion, followed by 
 heading in of caps on 6-ft centers on top 
 of the muck floor. Two rows of lagging 
 were nailed down from cap to cap, fol- 
 lowed by a layer of 4- by 4-in No. 8 wire 
 mesh and Fabrene cloth. The floor caps 
 were cabled to the roof caps to prevent 
 their slippage and to eliminate the need 
 for posting on the cut below. A conven- 
 tional sand wall, seen in figure 8, was 
 constructed. An initial 8:1 sand-to- 
 cement mix was poured to a thickness of 
 3 to 4 ft, followed by uncemented sand- 
 fill in the upper portion of the cut. 
 The high fines content and low pulp den- 
 sity of the fill, as well as poor drain- 
 age qualities, resulted in a large loss 
 of cement down the chutes to the level 
 
 below. Most of the cement and fines re- 
 maining in the stope were concentrated 
 at the front of the stope near the sand 
 wall, leaving an inconsistently cemented 
 sandfill over the length of the stope. 
 
 Mining in stope A progressed well dur- 
 ing the second cut despite a poorly ce- 
 mented sandfill. As shown in figure 9, 
 there was no need for additional timber 
 in this stope. 
 
 Mining in B and C stopes was combined 
 on a double-shift basis, with all broken 
 ore slushed to raise B. Inexperienced 
 mining crews occasionally blasted down 
 some of the overlying caps and fill. 
 Raise closure continued at a nearly con- 
 stant rate of 1.5 in per month. Raise A 
 experienced nearly 18 in of total closure 
 during 12 months of mining. The stope 
 downtime for replacement of broken raise 
 caps was significant, requiring nearly 3 
 weeks between cuts. The second cut re- 
 quired approximately 1 month to complete 
 in stope A and nearly 3 months in B. 
 
 At this time, a complete report of the 
 mining and the problem with the cemented 
 sandfill was made to the mine management. 
 It was agreed that the sandfill crew 
 would make a concerted attempt to in- 
 crease the pulp density of the fill and 
 eliminate the high slimes content. 
 
 Preparation of the floormat was the 
 same as in the previous cut, with two ex- 
 ceptions: an 18-inch layer of broken 
 muck was left on the stope floor as a 
 blast cushion, and the caps were tied 
 with wire rope to split-set rock bolts 
 placed in the wall, as shown in fig- 
 ure 10, rather than to the overlying 
 caps. Upon filling, however, problems 
 with low pulp density and high slimes 
 content were again encountered, and poor 
 sandwall and stope drainage technique 
 caused a high cement loss down the 
 chutes. 
 
 The mining of the third cut in stope A 
 went quite well. Experience, gained in 
 the previous cuts, eliminated basic oper- 
 ational problems. Once mining had pro- 
 gressed two rounds on either side from 
 
58 
 
 FIGURE 8. - Sand wall construction. 
 
 the raise, it was possible to cycle each 
 stope face almost nightly. Tying the 
 caps to split-set bolts with wire rope 
 eliminated the need for additional tim- 
 ber. The marginal cemented fill in this 
 cut presented no significant problems 
 due to the narrow (6 to 7 ft) width of 
 the stope. Mining of this third cut 
 required approximately 1 month, includ- 
 ing 1 week of raise repair prior to min- 
 ing. Productivity and cost were greatly 
 improved. 
 
 Closure of the stope continued at ap- 
 proximately the same rate as before. 
 Low mlcroseismic activity was likely re- 
 lated to frictional sliding along frac- 
 tures in the crushing pillar. The data 
 indicated the pillar was failing in a 
 
 nonviolent manner with 
 of blasting. 
 
 little likelihood 
 
 During this time, the second cut in 
 stope B was completed and prepared for 
 fill. During filling, greater attention 
 was directed to proper drainage, result- 
 ing in less spillage of cement to the 
 level below. This produced the best fill 
 to date, as was observed when mining be- 
 gan in the next cut ,(fig. 11). Though 
 still not of the quality desired, the 
 fill was strong enough to support itself. 
 Both stopes were experiencing a continual 
 problem with much removal. Prior cement 
 spills , which had reached the pockets on 
 the lower level, cemented the broken ore. 
 The result was inadequate storage capa- 
 city and muckbound stopes. 
 
59 
 
 \ ■■ V • 
 
 V 1 '. 
 
 FIGURE 9. - Completion of mining in second cut. 
 
60 
 
 FIGURE 10. - Stope prepared for filling. 
 
 Preparation of the third-floor mat in 
 stope A required approximately 1 week. 
 The mining crew handled the sandfilling 
 duties themselves. They made an initial 
 35-min pour, then held the water behind 
 the sand wall, allowing the cement to 
 settle out. The water was then decanted 
 off, and the remainder of the stope was 
 filled. A fairly well-cemented fill was 
 obtained. By the end of December 1981 
 mining was completed on the third cut of 
 stope B without major mining problems. 
 The hung ore pocket continued to cause a 
 muck removal problem. 
 
 The fourth and final cut, which was 14 
 ft in height, was mined by end slicing, 
 
 but with conventional br easting-down 
 rounds rather than the horizontal V-cut 
 used thus far. The mining in stope A 
 progressed rapidly and was completed by 
 mid-March. This crew had become quite 
 proficient in underhand mining and had 
 few problems over the last two cuts. 
 
 In this final cut, a decrease in clo- 
 sure rate was reflected by fill pressure 
 of approximately 100 psi (before erratic 
 readings developed) . Slow progress was 
 made in stope B because of many boulders, 
 which restricted the access way at the 
 bottom of the pillar. Poor ore prices 
 caused layoffs at the mine, and the 
 crews were changed in B, further slowing 
 
61 
 
 ^^^ ' 
 
 m;r 
 
 FIGURE 11. . Cement fill above third cut. 
 
 progress. A general mine closure even- 
 tually halted mining at the end of May 
 1982, with an approximate 50-ft length 
 of pillar remaining. Mining of the pil- 
 lar following initial raising had re- 
 quired 12 months in stope A and some 
 14 months in B. 
 
 COST COMPARISON 
 
 For cost analysis during the demonstra- 
 tion phase, the actual production costs 
 are given in table 1. Data comparison is 
 made to mining of a 35-ft-thick sill 
 
 pillar in a nearby area of the mine by 
 the overhand method. 
 
 It is estimated that underhand cut- 
 and-fill costs would be $48.81 per ton If 
 the method was adopted as a standard pro- 
 cedure for mining sill pillars. This 
 estimate argues that experienced crews 
 in well-prepared stopes would reduce 
 costs in the areas where unscheduled 
 maintenance proved costly in terms of 
 lost production, as well as extra labor 
 and materials. 
 
62 
 
 TABLE 1. - Comparison of sill pillar recovery costs, per ton 
 
 Item 
 
 Labor 
 
 Explosives 
 
 Rock bolts , 
 
 Sandf ill 
 
 Drill repair and drills , 
 
 Bits and rods 
 
 Timber 
 
 General-conventional-miscellaneous, 
 
 under hand-Fabrene , cement 
 
 Total direct costs per ton , 
 
 Total indirect costs per ton , 
 
 Total costs (excluding milling and 
 general and administrative) 
 
 Productivity. .. .tons per worker shift., 
 
 Overhand 
 
 cut and 
 
 fill 
 
 $19.94 
 1.54 
 1.81 
 1.25 
 1.01 
 .77 
 1.97 
 
 1.99 
 
 30.28 
 27.08 
 
 57.36 
 
 12.04 
 
 Underhand 
 
 cut and 
 
 fill 
 
 $22.56 
 
 1.15 
 
 .25 
 
 1.42 
 
 .95 
 
 .91 
 
 2.09 
 
 6.68 
 
 36.01 
 27.41 
 
 63.42 
 
 7.96 
 
 CONCLUSIONS 
 
 Little difficulty was experienced in 
 applying underhand cut-and-fill mining 
 procedures to the Coeur d'Alene mines. 
 After initial problems were resolved, 
 rounds were cycled daily in the later 
 cuts. Two noteworthy problems were en- 
 countered, however. Excessive closure in 
 the raises resulted in a great deal of 
 raise repair. Carefully designed and 
 constructed raises would save considera- 
 ble time and expense in light of the 
 amount of closure that can be expected in 
 a preconditioned pillar. The second 
 problem was getting a well-consolidated 
 cemented fill. Particle size of the mill 
 tailings was very fine and detrimental to 
 setting of the sand-cement mix. Also, 
 the weight percent solids of the mix was 
 too low. Excess water prevented cement 
 from adhering to the sand particles, al- 
 lowing it to be carried off with the 
 slimes. Control of the pulp density is 
 critical and should be kept at 65 pet 
 minimum. 
 
 Regardless of operational problems, the 
 success of the project must be measured 
 in terms of rock burst reduction. The 
 only rock burst occurred some 100 ft out 
 in the wall below stopes A and B (during 
 mining of the first cut), causing minimal 
 damage in the undercut stopes. Five caps 
 
 in stope B were broken, but it was not 
 apparent whether this was caused by the 
 burst or by the 4 in of closure that re- 
 sulted from mining. A minor amount of 
 rock slid off the walls. Only 0.3 In of 
 closure was due to the burst. In the old 
 stopes below, considerable rock was 
 shaken from the back. There was no 
 seismic buildup prior to the burst, and 
 the popping and cracking accompanying 
 mining was of a destressing nature, too 
 small to register on the mine micro- 
 seismic monitoring system. 
 
 The last three cuts were mined without 
 the occurrence of bumps or rock bursts. 
 The minor popping and cracking that 
 accompanied mining of the first three 
 cuts disappeared by the fourth (final) 
 cut as the stopes continued to squeeze 
 and further destress. The high closure 
 accompanying the mining of all cuts , 2 
 to 4 in per cut, verified the effective- 
 ness of destress blasting in softening 
 the pillar. 
 
 Underhand mining beneath a cemented 
 backfill appears to offer greatly im- 
 proved ground control during sill pillar 
 mining. Provided the cemented fill is of 
 good quality, there is no potential roof 
 fall problem. 
 
63 
 
 A comparison shows the cost of under- 
 hand cut and fill was 11 pet higher and 
 productivity was 34 pet lower than that 
 associated with mining a similar pillar 
 using conventional overhand cut and fill. 
 This is largely due to the mine's inex- 
 perience with the mining system (includ- 
 ing preparation) and cemented backfill. 
 If the costs of the underhand cut and 
 fill are adjusted to reflect an accepta- 
 ble productivity level, the underhand 
 
 cut-and-fill mining will reduce sill pil- 
 lar mining costs by 15 pet. 
 
 The conclusion from this demonstration 
 is that the combination of ore precondi- 
 tioning and underhand mining resulted in 
 greatly improved rock burst and ground 
 control, allowing safe and efficient min- 
 ing of a potentially hazardous, burst- 
 prone pillar. 
 
64 
 
 HAZARD DETECTION 
 
 ACOUSTIC CROSS-BOREHOLE SYSTEM FOR HAZARD DETECTION 
 
 By Karen S. Radcliffe^ and Richard E. Thill2 
 
 ABSTRACT 
 
 A high-frequency (20 kHz) acoustic 
 cross-borehole system has been devised 
 and field tested to remotely investigate 
 the structural conditions of a rock mass 
 in advance of mining. Elastic properties 
 of the rock can be determined by monitor- 
 ing acoustic waves generated between 
 boreholes, and the structural integrity 
 of the rock mass can be interpreted by 
 
 evaluation of the acoustic signal charac- 
 teristics. Identification of hazardous 
 ground conditions prior to mining can 
 help reduce, and ultimately prevent, in- 
 juries and fatalities to miners as well 
 as interruptions in the mining operation 
 due to encounters with unstable ground 
 and geologic hazards. 
 
 INTRODUCTION 
 
 Geologic anomalies can have serious 
 effects on the safety of miners and on 
 production when encountered unexpectedly 
 during a mining operation. Unidentified 
 fracture zones, voids (such as abandoned 
 mine workings or solution cavities) , 
 lithologic f acies changes , and inclu- 
 sions, both within and surrounding the 
 ore body, create potentially hazardous 
 conditions for the miner. Fractures, 
 joints, faults, and bedding planes are 
 also critical in determining environmen- 
 tal effects from mining, controlling 
 ground movements, and supporting excava- 
 tions and mine structures. 
 
 These structural features create zones 
 of increased permeability, often con- 
 taining large quantities of ground water 
 that can quickly inundate the mine when 
 encountered during mining. These same 
 structural features can also produce 
 weakened and therefore unstable roof 
 conditions, resulting in roof falls or 
 collapse at the working face. Also haz- 
 ardous in coal mining is the methane 
 that may be contained in fractures or 
 joints. Abandoned and unmapped oil and 
 gas wells, and especially abandoned mine 
 workings, create vulnerable conditions 
 
 Geologist. 
 
 ^Supervisory geophysicist. 
 Twin Cities Research Center, Bureau of 
 Mines, Minneapolis, MN. 
 
 for inundation and mine 
 cially in coal mines. 
 
 flooding, espe- 
 
 In addition to the structural weakness 
 inherent in a rock mass from natural dis- 
 continuities and geologic features, dam- 
 age can also be introduced into the 
 mine's supporting structures from mine 
 excavation and blasting operations. Even 
 controlled blasting can result in over- 
 break into mine structures or weaken al- 
 ready unstable areas. 
 
 To help reduce, and ultimately prevent, 
 injuries and fatalities to miners and 
 interruptions in the mining operation 
 from unstable ground and other geologic 
 hazards, it is necessary to remotely in- 
 vestigate the structural conditions in 
 advance of mining. This can be accom- 
 plished on a large scale during mine ex- 
 ploration and development, or on a small 
 scale as mining progresses by probing the 
 rock mass structure ahead of the working 
 face. It is necessary to characterize 
 the nature and structural integrity of 
 the rock mass surrounding the material to 
 be mined as well as the integrity of the 
 ore body itself. By characterizing these 
 materials and identifying any hazardous 
 anomalies prior to rock excavation, ap- 
 propriate safeguards can be taken to re- 
 duce or prevent the hazards posed by un- 
 stable conditions. 
 
65 
 
 Various techniques exist for assessing 
 the condition of a rock mass , ranging 
 from mechanical property tests on small- 
 scale samples in the laboratory to large- 
 scale testing under field conditions. 
 Although laboratory testing provides val- 
 uable information concerning the intact 
 elements of the rock mass, the effects of 
 large-scale discontinuities in that rock 
 mass are best determined in situ. Me- 
 chanical property data that are truly in- 
 dicative of the behavior of the rock mass 
 under in situ conditions of stress, mois- 
 ture, and other environmental factors are 
 also best determined in the field. In 
 situ techniques also provide the only op- 
 portunity to identify and specifically 
 locate the presence of geologic anomalies 
 and hazardous conditions. 
 
 Depending upon the type and size of 
 feature to be located, several in situ 
 geophysical methods for evaluating the 
 structure of a rock mass are available. 
 Nonseismic geophysical surveying includes 
 gravimetric, magnetometric and geoelec- 
 tric, thermal, radioactive, and geochemi- 
 cal methods (J_).2 Application of these 
 techniques, as well as seismic methods, 
 is based on the presence of a measurable 
 difference in a physical property within 
 the earth materials under evaluation, 
 such as acoustic velocity, density, mag- 
 netic susceptibility, or resistivity. 
 Use of nonseismic geophysics in detecting 
 mining hazards is limited by the range 
 over which the methods can be applied and 
 by the degree of variation in physical 
 properties required within the structures 
 of interest. Interpretation of field 
 data is also limited because of its de- 
 pendence upon the model chosen in the 
 investigation. 
 
 ACOUSTIC CROSS 
 
 The acoustic cross-borehole system op- 
 erates at a frequency of 20 kHz. It cou- 
 ples under pneumatic or hydraulic pres- 
 sure to the borehole wall, and can there- 
 fore function in water-saturated or dry 
 holes. Acoustic measurements can be made 
 in vertical holes from the surface, or in 
 
 ■^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
 Seismic techniques have been used ex- 
 tensively by the petroleum industry for 
 locating geologic structures favorable 
 for economic accumulations of oil and 
 gas. Surface seismic exploration methods 
 have also been used for determining sub- 
 surface geology and mapping. Of all the 
 physical methods used in geological ex- 
 ploration, the seismic methods are con- 
 sidered to be the most direct, and when 
 applicable, give the least ambiguous 
 results. Most of the techniques used are 
 surface surveys that operate over a large 
 horizontal distance. The vertical depth 
 to which seismic surveys are effective 
 depends upon the stratigraphy and struc- 
 ture as well as the strength and fre- 
 quency of the seismic signal utilized in 
 the survey. The applicability of seismic 
 methods is based on the relationship 
 between the acoustic properties of rocks 
 and their physical properties , mineral 
 composition, and structural integrity 
 (2). 
 
 To overcome the deficiencies of many 
 surface geophysical methods, an acoustic 
 cross-borehole device was developed and 
 field tested by the Bureau of Mines (3^) . 
 The apparatus is designed to operate at 
 high frequency (20 kHz) over moderate 
 distances between boreholes. These dis- 
 tances range from a few meters to tens of 
 meters, depending upon the acoustic 
 transmission characteristics of the rock. 
 Applications include evaluating the elas- 
 tic properties and integrity of under- 
 ground structures, monitoring stress 
 changes in and around underground work- 
 ings, evaluating fragmentation and over- 
 break from blasting, and detecting dis- 
 continuities in advance of mining. 
 
 -BOREHOLE SYSTEM 
 
 horizontal or inclined holes in the roof, 
 rib, or floor of an underground mine. 
 
 SYSTEM COMPONENTS 
 
 The basic acoustic cross-borehole sys- 
 tem consists of a high-frequency pulse 
 generator, transmitting and receiving 
 borehole transducers, and monitoring and 
 
 timing electronics (fig. 1). Auxililary 
 components include a downhole amplifier 
 in the receiver transducer to improve the 
 
66 
 
 FIGURE 1. - Basic components of acoustic cross-borehole system: borehole transducers, pulse gen- 
 erator, and signal monitoring and timing electronics. 
 
 received acoustic signal, and high- 
 voltage coaxial cable for driving the 
 source. 
 
 Pulse Generator 
 
 A 5-kV electronic pulse generator 
 drives a piezoelectric source in the 
 downhole transmitter. A fast-rising out- 
 put trigger pulse is provided for syn- 
 chronizing the timing circuitry. The 
 unit is operable from a conventional 60- 
 Hz, 120-V line source. 
 
 Borehole Transducers 
 
 The transmitter and receiver transduc- 
 ers are pressure-sensitive hydrophones. 
 Uniform lateral radiation and reception 
 of acoustic energy is achieved through 
 use of cylindrical lead-zirconate piezo- 
 electric transducers operating in the ra- 
 dial expansion mode. A cross-sectional 
 view of the transducer assembly is shown 
 in figure 2. An internal inflatable 
 bladder causes displacement of fluid in 
 the transducer cavity and outward expan- 
 sion of the outer neoprene boot to cou- 
 ple to the borehole wall. The transducer 
 
 elements are moderately damped, giving 
 high sensitivity but producing some re- 
 verberation ringing. The reverberation 
 is not detrimental to picking the onset 
 of the first arrival normally associated 
 with the compressional wave (P-wave) , but 
 it can mask or interfere with subsequent 
 wave arrivals. The emitter and receiver 
 transducers are identical in construction 
 and are 6.67 cm diam by 20 cm long. The 
 central operating frequency of the emit- 
 ter is 20 kHz. Frequency response of 
 the receiver is essentially flat between 
 19 and 31 kHz. Operating voltage for 
 the emitter is limited to 2 kV to pre- 
 vent deterioration of the piezoelectric 
 elements. 
 
 Downhole Amplifier 
 
 A solid-state amplifier was designed to 
 fit into a small (7-cm-diam) compartment 
 located near the receiver transducer. 
 The amplifier provides substantial sig- 
 nal gain and drives a long length of 
 cable (up to 25 m) with minimum signal 
 loss or distortion. Frequency response 
 of the amplifier is flat over the range 
 5 to 50 kHz. 
 
67 
 
 Pressurizin 
 port 
 
 Inflatable 
 bladder 
 
 Stainless 
 
 steel 
 
 housing 
 
 Couplant 
 
 Neoprene — ^ 
 
 expandable 
 
 boot 
 
 Electrical 
 insulator 
 
 Coaxial 
 connector 
 
 Air or fluid 
 
 pressurizing 
 
 media 
 
 Electrode 
 feedthrough 
 
 Cylindrical 
 piezoelectric 
 transducer 
 element 
 
 Transducer 
 backing 
 
 Boot clamp 
 
 FIGURE 2. - Cross-sectional diagram of bore- 
 hole transducer assembly. 
 
 Monitoring and Timing Circuitry 
 
 Conventional oscilloscope waveform dis- 
 play and timing circuitry are used to 
 monitor the received signal and to obtain 
 elastic wave transmit times. Acoustic 
 signals are captured and digitized on a 
 floppy disk to permit detailed waveform 
 analysis. 
 
 PRINCIPLES OF OPERATION 
 
 The principle of the acoustic cross- 
 borehole system is to propagate a seismic 
 wave from an emitter in one borehole to a 
 receiver in a second borehole (fig. 3). 
 By measuring the wave's traveltime and 
 the distance between boreholes, the 
 acoustic velocity of the seismic wave 
 through the rock mass can be determined. 
 By moving the emitter and receiver along 
 the lengths of their boreholes, profiles 
 of wave velocity or amplitude in the 
 
 plane including the boreholes can be 
 made. Proper geometric arrangement of 
 the boreholes then enables delineation of 
 the subsurface structures. Calculation 
 of wave velocity assumes a straight line 
 of propagation between the emitter and 
 receiver transducers, and for short prop- 
 agation paths, the direct compressional 
 wave is normally the first energy to ar- 
 rive at the receiver. At greater dis- 
 tances, however, and in the presence of 
 higher velocity zones adjacent to the 
 path between the emitter and receiver, 
 refracted waves may begin to reach the 
 receiver before the direct P-wave. Typi- 
 cal wave traces are shown in figure 4. 
 
 Deviations from the expected character- 
 istics of the propagated wave can be used 
 to identify structurally hazardous fea- 
 tures in mine strata. Characteristics of 
 the seismic wave arriving at the receiver 
 (such as the wave velocity, amplitude, 
 and frequency content) are related to the 
 physical and mechanical properties of the 
 geologic media through which they travel. 
 In rock masses approximating homogeneous, 
 elastic, isotropic media of semi-infinite 
 extent, a seismic disturbance will gen- 
 erate compressional (P-wave) and shear 
 (S-wave) waves that travel with veloci- 
 ties related to the density and elastic 
 constants of the rock medium according to 
 the equations 
 
 and 
 
 "P [p (1 + a)(l-2a)J 
 
 /2 
 
 (1) 
 (2) 
 
 where Vp is the compressional wave veloc- 
 ity, Vs is the shear wave velocity, p is 
 the density, a is Poisson's ratio, and E 
 and G are the Young's and shear moduli, 
 respectively. Surface waves may also be 
 generated at free surfaces of the rock 
 mass by the seismic disturbance (_3 ) . 
 
 Unfortunately, rock masses are seldom 
 homogeneous, perfectly elastic, or iso- 
 tropic, but are more often complex struc- 
 tures containing joints, bedding, frac- 
 tures, and weathered zones. Discon- 
 tinuities such as these, in addition to 
 
68 
 
 — il^Surface support system 
 
 ^^T^^^^^J^^^JT-f- 
 
 \t='y^^^ 
 
 Emitter 
 
 Zone of Discontinuities 
 
 FIGURE 3. - Acoustic system for detecting discontinuities between boreholes. 
 
 cavities and voids within the rock mass, 
 all influence the transmission of seismic 
 waves through the rock. The presence of 
 structural features in a rock mass can be 
 identified by monitoring deviations in 
 the compressional- and shear-wave veloci- 
 ties as well as the associated mechanical 
 properties that can be calculated. The 
 structural integrity of a rock mass is 
 assessed by comparing the wave velocity, 
 wave amplitude, other signal characteris- 
 tics, and the physical and mechanical 
 properties of the rock mass with known 
 characteristics of the intact material. 
 
 Reliable detection of geologic hazards 
 requires that there be sufficient acous- 
 tic contrast between the target hazard 
 and the surrounding material so that the 
 probing energy will be reflected, re- 
 fracted, attenuated, or otherwise identi- 
 fiably modified. Also important with 
 respect to the frequency of the seismic 
 signal generated are the target size and 
 shape compared with the wavelength of the 
 
 signal. The ability to accurately inter- 
 pret information about geologic condi- 
 tions from the detection signals requires 
 prior identification of various types of 
 geologic anomalies and what their effects 
 on propagated seismic signals might be. 
 
 Propagating and detecting seismic waves 
 in a discontinuous rock mass requires a 
 tradeoff between range and resolution. 
 Use of high-frequency seismic energy re- 
 sults in better resolution of small geo- 
 logic features but limits the penetration 
 distance, because of the selective atten- 
 uation and absorption of the higher fre- 
 quency components of the propagating 
 wave. Seismic resolution of geologic 
 anomalies that may cause mining hazards 
 requires use of a wavelength comparable 
 to, and preferably smaller than, the 
 dimensions of the anomaly to be observed. 
 Table 1 describes typical applications 
 of the cross-borehole acoustic technique 
 relative to the seismic frequencies 
 required. 
 
69 
 
 157 cm 
 
 319 cm 
 
 473 cm 
 
 603 cm 
 
 Wiiiil 
 
 I 
 
 .5 V 
 
 h 
 
 
 0.5 ms 
 
 FIGURE 4. - Typical wave traces obtained in situ. 
 
 TABLE 1. - Typical applications of acoustic cross-borehole methods 
 
 Operational frequency 
 
 20 kHz 
 
 2 kHz 
 
 Typical wavelength in rock, m: ' 
 
 
 
 Sandstone. ................... 
 
 0.10-0.18 
 
 .16- .28 
 
 .11 
 
 .06- .14 
 
 Fractures, joints. 
 
 1.0-1.8 
 
 Limestone. ................... 
 
 1.6-2.8 
 
 Shale 
 
 1.1 
 
 Coal 
 
 .6-1.4 
 
 Geologic hazards or structural 
 
 Faults , channel 
 
 features detected. ^ 
 
 kettlebottoms, 
 
 sands, voids. 
 
 
 blast damage. 
 
 abandoned mines , 
 
 
 small-scale 
 
 large-scale 
 
 
 geologic dis- 
 
 geologic dis- 
 
 
 continuities. 
 
 continuities. 
 
 'Reference 2. 
 
 ^Rock mass properties (Young's modulus, shear modulus, bulk modulus, 
 Poisson's ratio) can also be determined at each frequency for various 
 rock types. 
 
70 
 
 PROOF TESTING AND LABORATORY CALIBRATION 
 
 The performance capability of the 
 acoustic cross-borehole apparatus was 
 first tested in a block of granite in the 
 laboratory. Tests were conducted to de- 
 termine waveform characteristics, P-wave 
 velocity, the radiation pattern from the 
 source, the traveltime correction factor 
 for electronic delays and pulse buildup 
 in the transducer , and the optimum cou- 
 pling pressure. 
 
 Testing determined that a minimum pres- 
 sure of 10 psi was required to couple the 
 transducer to the borehole wall and that 
 amplitude of the first arrival gradually 
 increased until 50 psi, after which there 
 was little change. All subsequent tests 
 were therefore conducted at coupling 
 pressures between 50 and 70 psi. A near- 
 ly uniform radiation pattern was observed 
 by monitoring the output from an accel- 
 erometer as the receiver was rotated 
 around the source. Small irregularities 
 were attributed to the coupling and sur- 
 face conditions. Wave transmission tests 
 produced excellent signal reception with 
 well-developed first arrivals. Rise time 
 of the first arrival was about 5 us, 
 indicating that high frequencies were 
 
 retained over short transmission dis- 
 tances in the granite. This permitted 
 wave velocity to be determined within 1 
 pet precision (_3) . 
 
 Mechanical properties of the block were 
 established in advance by conventional 
 ultrasonic and bar resonance tests (4_) on 
 core extracted from boreholes in the 
 granite block. These properties provided 
 a basis for comparison of the laboratory 
 data with in situ measurements. 
 
 Laboratory testing was also performed 
 in a concrete slab to determine the ef- 
 fects of propagation distance on the am- 
 plitude of the first wave arrival. The 
 wave traces verified that the first arri- 
 val diminishes rapidly with increasing 
 propagation distance. As demonstrated 
 in the granite block, the high frequen- 
 cies were filtered through the concrete. 
 Although the time difference between 
 first and second arrivals increases with 
 propagation distance, the change is not 
 large and again introduces an error of 
 less than 1 pet in determining transit 
 time at these distances. 
 
 FIELD VERIFICATION 
 
 EMERALD ISLE MINESITE 
 
 The first field tests of the high- 
 frequency cross-borehole system were con- 
 ducted at the Emerald Isle open pit mine 
 near Kingman, AZ, where the Bureau of 
 Mines was experimenting with blasting to 
 fracture the rock mass for in situ leach- 
 ing {5). Acoustic velocities of a cop- 
 per conglomerate were determined in deep 
 (76 to 85 m) , vertical, water-saturated 
 boreholes following blasting. The tests 
 were designed to determine whether the 
 cross-borehole system could be used to 
 detect increased fracturing following 
 blasting. Results showed a decrease in 
 
 acoustic velocity with depth, correspond- 
 ing to a deterioration of the rock due to 
 blasting, and delineation of a weathered 
 zone at the granite-gneiss interface. 
 Wave velocities calculated from field 
 data are shown in relation to velocities 
 determined from laboratory tests on pre- 
 shot core recovered from the boreholes 
 (fig. 5). 
 
 SENECA MINESITE 
 
 Performance tests of the cross-borehole 
 system were also conducted in inclined, 
 water-filled boreholes underground at 
 the Seneca Mine near Mohawk, MI. In 
 
71 
 
 175 
 
 205 
 
 t235 
 
 265- 
 
 295 
 
 KEY 
 ^ Cross-borehole velocity 
 
 determinations 
 — Preshot P-wave velocity 
 
 Overburden 
 contact I 
 
 \ Gila conglom- 
 erate 
 
 A Twoaiiioiou &UIIO 
 
 RubbI 
 
 ^^ 
 
 Weatheredzone 
 
 Granite gneiss 
 contract 
 
 O A §1 
 
 MEAN VELOCITY, km/s 
 
 8 
 
 60 
 
 70 E 
 
 I 
 I- 
 
 Q. 
 LU 
 
 Q 
 80 
 
 90 
 
 FIGURE 5. - Comparison of laboratory and field 
 acoustic velocities at the Emerald Isle minesite. 
 
 Studies similar to those at the Emerald 
 Isle site, the Bureau of Mines was con- 
 ducting research to develop methods of 
 confined blasting to create fracture 
 permeability for in situ leaching of na- 
 tive copper ores (6^). Acoustic veloci- 
 ties were determined before and after 
 blasting at a separation distance of 5m 
 between boreholes. The velocity deter- 
 minations indicated little, if any, ef- 
 fect from blasting in the zone of the 
 explosives column. Results from permea- 
 bility and Rock. Quality Designation (RQD) 
 measurements conducted at the same time 
 also indicated that the rock was not 
 highly fractured by the blast and that 
 fractures which were present were tightly 
 closed by overburden pressure O) . The 
 excellent quality of the signals received 
 indicated that considerably larger dis- 
 tances (>15 m) could have been traversed 
 in the amygdaloidal basalt host rock. 
 
 LOGAN WASH MINESITE 
 
 The most recent series of field tests 
 was conducted in cooperation with 
 Occidental Oil Shale, Inc., at the Logan 
 Wash experimental mine near DeBeque, CO. 
 Acoustic velocities were determined under 
 preshot and postshot conditions as in 
 situ oil shale retorts were rubblized us- 
 ing explosives. The field tests were 
 designed to evaluate the cross-borehole 
 system's capability to detect blast dam- 
 age to the pillars separating the re- 
 torts. One set of measurements was taken 
 in vertical holes, and two other series 
 were made at inclinations of 30° and 45° 
 in water-saturated boreholes. 
 
 Slight decreases in acoustic velocities 
 were observed following retort rubbliza- 
 tion, suggesting increased fracturing in 
 the pillars (fig. 6). The greatest over- 
 all decrease in velocity was on the order 
 of 4.5 pet in a zone between 12 to 15 m 
 from the borehole collar and 14 m in from 
 the retort wall. It was expected that 
 there would be a minimal change in veloc- 
 ity before and after blasting, since the 
 measurements were made in the pillar and 
 the blasting was designed to contain 
 fragmentation within the rubble bed in 
 the retort. 
 
 Cores extracted from the pillar be- 
 fore and after retort rubblization were 
 subjected to ultrasonic velocity deter- 
 mination in the laboratory to provide 
 another set of data for detecting veloc- 
 ity changes. The observed decrease in 
 field velocity following blasting cor- 
 responds favorably with the laboratory 
 acoustic data, as well as with RQD and 
 fracture frequency determined from the 
 preshot and postshot core. This posi- 
 tive correlation suggests that the cross- 
 borehole system is a reliable tool for 
 detecting small variations in geologic 
 structure in situ. 
 
72 
 
 E 
 
 u 
 o 
 
 ui 
 
 > 
 
 lU 
 
 (9 
 
 < 
 oc 
 IIJ. 
 
 5.50 
 
 5.25 
 
 5.00 
 
 4.75 
 
 4.50 
 
 4.25 
 
 4.00 
 
 3.75 
 
 3.50 
 
 KEY 
 
 Preshot R-8 
 Postshot R-8 
 
 10 
 
 15 
 
 20 
 
 25 
 
 45 
 
 50 
 
 55 
 
 60 
 
 65 
 
 30 35 40 
 
 DEPTH, ft 
 
 FIGURE 6. - Depth versus acoustic velocity, preshot versus postshot Retort 8, Logan Wash site. 
 DEVELOPMENT OF LOW-FREQUENCY CROSS-BOREHOLE SYSTEM 
 
 70 
 
 A second cross-borehole system is cur- 
 rently in development by the Bureau of 
 Mines that operates at lower frequency (1 
 to 2 kHz) over wider separation distances 
 between boreholes. The low-frequency 
 seismic signal generated by this system 
 will provide complementary hazard detec- 
 tion capabilities to the high-frequency 
 system by detecting much larger anomalous 
 features prior to mining, and at dis- 
 tances of up to 150 m. This system is 
 especially designed to detect uncharted 
 underground voids in coal measure rocks. 
 It is unique in that it will operate 
 simultaneously in combined modes of re- 
 flection and through-transmission to pro- 
 vide the best opportunity to detect haz- 
 ards in advance of mining (fig. 7). 
 
 The acoustic principles utilized by 
 this low-frequency system are the same as 
 those governing the operation of the 
 
 high-frequency cross-borehole system cur- 
 rently in use. The combined reflection, 
 through-transmission capability, however, 
 provides for a modified method of struc- 
 ture interpretation. In this case, the 
 emitter probe located in one borehole 
 sends out an acoustic pulse toward the 
 receiver probe in a second borehole. A 
 void space, or other geologic discon- 
 tinuity, will be delineated by the record 
 of transmitted and reflected energy in 
 the acoustic waveform generated by the 
 emitter. A sharp discontinuity between 
 boreholes will reflect back nearly all 
 energy to the receiver-emitter probe and 
 transmit little energy to the distant 
 receiver. Distance to the discontinu- 
 ous surface can be calculated from the 
 traveltime data. Conversely, in a solid, 
 homogeneous rock mass, nearly all of the 
 energy will be transmitted and little re- 
 flected. By profiling along the lengths 
 
73 
 
 Surface support 
 system and vehicle 
 
 o^ 
 
 \x 
 
 .,-m^^-'- -'-' 
 
 <??? 
 
 '^r^^'??^^;?^!^^^::^^^. 
 
 '"mw^m%' 
 
 Transmitter- 
 receiver probe 
 
 Acoustic 
 
 source 
 
 Acoustic 
 receiver 
 
 Abandoned mine 
 
 FIGURE 7. - Combined reflection and through-transmission cross-borehole system. 
 
 of the boreholes , and by arranging geo- 
 metric patterns of boreholes, voids and 
 other discontinuities could be delineated 
 in size, shape, and location. Comparison 
 of the transmitted and reflected energy 
 
 will be useful in interpreting hazards 
 ahead of mining. Sophisticated data pro- 
 cessing techniques will also permit more 
 detailed reduction of velocity and fre- 
 quency data. 
 
 CONCLUSIONS 
 
 A high-frequency acoustic cross- 
 borehole system has been developed and 
 field tested by the Bureau of Mines for 
 in situ investigation of rock mass prop- 
 erties and detection of geologic hazards 
 in advance of mining. The system oper- 
 ates at a frequency of 20 kHz between 
 boreholes separated by distances from a 
 few meters to tens of meters. Changes in 
 the characteristics of a seismic wave as 
 it propagates through geologic structures 
 can be used to detect potential hazards 
 in the mining environment. Applications 
 of the cross-borehole system include de- 
 tection of fracturing and overbreak from 
 mine blasting, locating underground voids 
 or abandoned mine workings , inferring 
 stress distributions in mine structures. 
 
 and detecting geologic hazards in advance 
 of mining. 
 
 The second cross-borehole system cur- 
 rently under development will propagate 
 much lower frequency (1 to 2 kHz) seismic 
 waves between boreholes separated by up 
 to 150 m. This system will operate in a 
 combined reflection, through-transmission 
 mode to detect and define the size, 
 shape, and location of larger geologic 
 hazards in advance of mining. The system 
 is specially designed to detect uncharted 
 voids in coal measure rocks. Field test- 
 ing of the low-frequency system will fol- 
 low assembly of the equipment components, 
 currently in progress. 
 
74 
 
 REFERENCES 
 
 1. Dohr, G. Applied Geophysics. Ge- 
 ology of Petroleum, v. 1. Halsted Press, 
 New York, 1981, 231 pp. 
 
 2. Rzhevsky, V., and G. Novik. The 
 Physics of Rocks. MIR Publ. , Moscow, 
 1971, 320 pp. 
 
 3. Thill, R. E. Acoustic Cross-Bore- 
 hole Apparatus for Determining In Situ 
 Elastic Properties and Structural Integ- 
 rity of Rock Masses. Paper in Proc. 19th 
 Symp. on Rock Mechanics, Stateline, NV, 
 May 1-3, 1978. Univ. NV, Reno Press, 
 pp. 121-129. 
 
 4. Thill, R. E., and S. S. Peng. Sta- 
 tistical Comparison of the Pulse and 
 
 Resonance Methods for Determining Elastic 
 Moduli. BuMines RI 7831, 1974, 24 pp. 
 
 5. D'Andrea, D. V., W. C. Larson, 
 L. R. Fletcher, P. G. Chamberlain, and 
 W. H. Engelmann. In Situ Leaching Re- 
 search in a Copper Deposit at the Emer- 
 ald Isle Mine. BuMines RI 8236, 1977, 
 43 pp. 
 
 6. Chamberlain, P. G. In-Place Leach- 
 ing at the Seneca Mine, Mohawk, Michigan. 
 Paper in Upper Peninsula AIME Spring 
 Technical Meeting Review. Skillings' 
 Min. Rev., v. 66, No. 21, May 21, 1977, 
 pp. 19-20. 
 
75 
 
 IN-SEAM GEOPHYSICAL TECHNIQUES FOR COAL MINE HAZARD DETECTION 
 By Richard J. Leckenby^ and James J. Snodgrass^ 
 
 ABSTRACT 
 
 The Bureau of Mines has helped to make 
 a number of advances in improving and de- 
 veloping geophysical methods for detect- 
 ing mining hazards from the working face 
 in underground coal mines. This paper 
 presents the results of some of the Bu- 
 reau's recent work in four, in-seam geo- 
 physical methods: 
 
 1. A synthetic pulse radar system with 
 initial field test results showing its 
 potential. 
 
 2. A high-frequency prototype seismic 
 system used to ascertain development cri- 
 teria for a hand-held scanner. 
 
 3. Controlled-source piezoelectric 
 transducers used to generate predominant- 
 ly compressional or shear wave energy for 
 better control over propagation of guided 
 waves . 
 
 4. A dry-hole borehole sonic probe 
 adapted to determine compressional and 
 shear wave velocities in a coal seam from 
 a horizontal borehole. 
 
 INTRODUCTION 
 
 Coal mines are often plagued by adverse 
 ground conditions because of geology or 
 previous mining or drilling in the area. 
 Common geologic problems include faults , 
 channel sands, split seams, and clay 
 veins , while manmade problems include 
 abandoned mines and wells. These prob- 
 lems can cause roof falls , blowouts , 
 flooding, or inundation. It is important 
 to determine the existence and locations 
 of these problem areas so that precau- 
 tions can be taken to correct or avoid 
 potential hazards. In addition, timely 
 detection can lead to increased produc- 
 tivity and resource recovery. 
 
 One of the most common and costly meth- 
 ods of hazard detection is drilling, but 
 even extensive drilling programs can miss 
 
 ^Physicist. 
 ^Geophysicist. 
 Denver Research Center, Bureau of 
 Mines, Denver, CO. 
 
 significant geologic and manmade con- 
 ditions. The Bureau has been investi- 
 gating, devising, and improving in-seam 
 geophysical methods for accurate, relia- 
 ble, economic, and practical detection 
 and mapping of potential hazards from the 
 working face. 
 
 There is no one proven method that can 
 detect all hazards in all conditions and 
 locations. It is necessary to undertake 
 research to devise a variety of methods 
 that can be used either singly or in 
 combination to provide for the most 
 efficient, reliable, and accurate hazard 
 detection. This paper considers some of 
 the most recent technological advances 
 the Bureau has made in in-seam geophysi- 
 cal methods. It includes a new ground 
 probing radar system called "synthetic 
 pulse radar" and seismic methods which 
 include developments in high-frequency, 
 seismic, guided waves and borehole 
 techniques. 
 
76 
 
 SYNTHETIC PULSE RADAR 
 
 Ground probing radar is a method using 
 electromagnetic waves for the detection 
 and mapping of anomalous conditions in 
 the ground. A number of ground probing 
 radar methods are either being used or 
 being investigated for use.^ Short pulse 
 radar is probably the most commonly used 
 method, but electronically it is diffi- 
 cult to generate a high-power, high- 
 frequency, broadband short pulse that can 
 be used to detect in-seam coal mine haz- 
 ards up to 200 ft from the working face. 
 Synthetic pulse radar is a method that 
 overcomes some of the difficulties found 
 in short pulse systems. 
 
 Synthetic pulse radar is based on a 
 proprietary concept developed by ENSCO, 
 Inc. A feasibility study was conducted 
 first under a Bureau contract to deter- 
 mine the potential of the concept for im- 
 proving radar technology.^ Based on that 
 first study a cost-sharing contract was 
 awarded to XADAR Corp. (a subsidiary of 
 ENSCO, Inc.) to design, build, and test a 
 prototype system for use in underground 
 coal mines. 
 
 short pulse radar, it is useful to review 
 some of the basic principles behind 
 Fourier analysis. According to Fourier 
 theorem, any periodic function (wave) can 
 be represented as a sum of a number (pos- 
 sibly infinite) of sine and cosine func- 
 tions. That is, a periodic wave can be 
 given by the equation: 
 
 Y = ao + ai sin tot + a2 sin 2ajt 
 
 + a3 sin 3tot + . . . 
 
 ai'cos cot + a2'cos 2(jot 
 
 + a3'cos 3a)t + ... (1) 
 
 This equation is known as a Fourier 
 series, which is composed of a constant 
 ao , amplitudes a], a2 , a.-^,,,. a^", a2'', 
 a3' ... and angular frequencies o), 2a), 
 3oj, ... The resultant wave is regarded 
 as being built up by a number of waves 
 whose wavelength ratios are 1, 1/2, 1/3, 
 1/4, ... From the study of sound this 
 represents the fundamental and its vari- 
 ous harmonics. 
 
 The construction and initial testing of 
 the synthetic system has been completed 
 to the point where it is now feasible to 
 use the system and to design new measur- 
 ing and processing techniques to take ad- 
 vantage of this new radar system. 
 
 SYNTHESIZED PULSE 
 
 To understand synthetic pulse radar and 
 how it differs from the more conventional 
 
 ^Leckenby, R. J. Electromagnetic 
 Ground Radar Methods. Paper in Premining 
 Investigations for Hardrock Mines. Pro- 
 ceedings: Bureau of Mines Technology 
 Transfer Seminar, Denver, CO, Sept. 25, 
 1981. BuMines IC 8891, 1982, pp. 36-45. 
 
 ^Fowler, J. C, S. D. Hale, and R. T. 
 Houck. Coal Mine Hazard Detection Using 
 Synthetic Pulse Radar (contract H0292025, 
 ENSCO, Inc.). BuMines OFR 79-81, 1981, 
 84 pp. 
 
 For waves that are not periodic, but 
 have zero displacement beyond a certain 
 finite range, a Fourier series cannot be 
 used. Instead, a Fourier integral is em- 
 ployed. The nonperiodic wave is composed 
 of (or can be synthesized by adding) an 
 infinite number of wave trains, each hav- 
 ing a frequency differing only infinites- 
 imally from the next. The process of de- 
 termining the components at each frequen- 
 cy is done by what is known as the 
 Fourier transform, and likewise the wave 
 synthesis is done by the inverse Fourier 
 transform. 
 
 The above discussion is based on 
 continuous waves, but in actual prac- 
 tice, waves are measured for finite 
 periods. This leads to a whole new 
 discussion on discrete Fourier trans- 
 forms and a computational method known 
 as the Fast Fourier Transform (FFT) 
 and the Inverse Fast Fourier Transform 
 
77 
 
 (IFFT),^ which will be described by sim- 
 ply stating that a FFT of a discrete time 
 series using prescribed cautions, meth- 
 ods, and rules will yield a Fourier 
 series of finite components which will be 
 a good estimate of the actual Fourier 
 transform. 
 
 With this general background of the 
 Fourier transform it is possible to see 
 the difference, or probably better 
 stated, the similarity between a short 
 pulse and the synthetic pulse. Figure 1 
 is composed of three graphs. The top 
 
 ^Bracewell, R. N. The Fourier Trans- 
 form and Its Applications. McGraw-Hill, 
 1978, pp. 356-381 . 
 
 Brigham, E. O. The Fast Fourier 
 Transform. Prentice-Hall, 1974, 252 pp. 
 
 graph is a wave contour of a digitized 
 transmitted pulse after traveling through 
 about 12 m of coal. A FFT was applied to 
 this time domain data to yield the fre- 
 quency domain data plotted in the lower 
 two graphs. The center graph is the am- 
 plitude spectrum, which is the magnitude 
 of the sine and cosine components at each 
 discrete frequency. The bottom graph is 
 the phase spectrum which is the tangen- 
 tial angle between the two components. 
 The wave was digitized with 1,024 sam- 
 ples of equal time intervals, between 
 to 500 ns. This yielded a corresponding 
 spectrum with a fundamental frequency of 
 2 MHz and 512 harmonic frequencies, of 
 which only the first 65 are plotted be- 
 cause the remaining frequencies are so 
 low in amplitude that they appear almost 
 zero on the graph. 
 
 50 100 150 200 250 300 350 400 450 500 
 TIME, ns 
 
 -180, 
 
 12 25 37 50 62 75 87 100 112 125 
 FREQUENCY, MHz 
 
 ■'■I|' I'l I'l'''''^'' P'l' l' I'll 
 
 25 
 
 37 50 62 75 87 
 
 FREQUENCY, MHz 
 
 100 112 
 
 FIGURE 1. - Three graphs showing the short 
 puise signal in the time domain (top graph) and 
 the frequency domain (bottom two graphs). From 
 top to bottom the graphs are (1) wove contour, 
 (2) amplitude spectrum, and (3) phase spectrum. 
 
 In operation each discrete frequency 
 needed to synthesize a pulse is transmit- 
 ted one at a time, and the resultant am- 
 plitude and phase at each frequency is 
 recorded. That is, the acquired data are 
 from the frequency domain and can be pre- 
 sented in a similar fashion as the lower 
 two graphs in figure 1. To ascertain the 
 time domain or corresponding pulse data 
 an IFFT or equation 1 is used to synthe- 
 size a wave, which can be presented in a 
 fashion similar to the wave contour of 
 the upper graph in figure 1. 
 
 SHORT PULSE VERSUS SYNTHETIC PULSE 
 
 The discussion thus far has served to 
 explain how a short pulse and synthetic 
 pulse will basically produce the same re- 
 sults. The major difference is in the 
 hardware and how it controls the energy 
 of the transmitted and received signals. 
 On a short pulse system the pulse is usu- 
 ally generated by a fast discharge of a 
 high-voltage potential into an antenna 
 and a complex impedance. The potential, 
 rate of discharge, antenna, and impedance 
 control which frequencies and their am- 
 plitudes will be transmitted into the 
 ground. This in turn will determine the 
 shape of the pulse, the resolution and 
 the effective pulse power, the sum of the 
 power at each frequency, which in turn 
 determines the range of the system. 
 
78 
 
 Figure 1 is a good example of an erratic 
 pulse spectrum that can be ascertained 
 from a short pulse system. On the other 
 hand the synthetic pulse transmits each 
 known frequency at a given power, which 
 assures better control over the shape, 
 resolution, and power of the transmitted 
 signals. It is also easier to obtain 
 higher effective pulse power than by us- 
 ing a broadband short pulse. That is, a 
 small increase in power for each frequen- 
 cy, when summed, gives a large increase 
 in pulse power. Likewise, pulse power 
 may be increased by increasing the number 
 of frequencies transmitted. 
 
 THE SYNTHETIC PULSE SYSTEM 
 
 The major components of the synthetic 
 pulse system are shown in figure 2. The 
 system employs a hetrodyne receiver, 
 which requires two synthesizers that are 
 phase locked together with a 1-mHz off- 
 set. By mixing the transmitted signal 
 and the intermediate frequency (IF) of 1 
 MHz with the received signal, the detec- 
 tion circuits only have to operate at the 
 
 IF frequency and not over the entire band 
 of transmitted frequencies. This mixing 
 allows for easier shielding of the input 
 from direct feedthrough of the transmit- 
 ted signal. A fiber optic link is also 
 used between the control unit and the 
 transmitter to eliminate direct feed- 
 through signals. 
 
 The transmission frequencies are in the 
 range of 20 to 160 MHz, with a mini- 
 mum frequency spacing of 100 kHz, thus 
 providing a time window of 10,000 ns. 
 The control unit, shown in figure 3, is 
 housed in an aluminum case with its own 
 batteries. The receiver electronics are 
 contained in the receiving antenna. The 
 transmitter, figure 4, consists of a 
 small box containing batteries and elec- 
 tronics with the power amplifier being 
 mounted in the transmitting antenna. The 
 antennas are loaded wideband dipoles sim- 
 ilar to those used in short pulse sys- 
 tems. A host computer, which is required 
 to perform the IFFT, can access the digi- 
 tal tape by either a serial or parallel 
 interface. 
 
 Tape 
 drive 
 
 Control 
 
 micro- 
 
 processor 
 
 A/D 
 conversion 
 
 CONTROL UNIT 
 
 Synthesizer 
 
 Fiber optic 
 transmitter 
 
 Offset 
 synthesizer 
 
 1 MHz IF 
 reference 
 
 nrRANSMITTER 
 
 Fiber optic 
 
 cable 
 
 Fiber optic [J Power 
 amplifier 
 
 Sample 
 and hold 
 
 Quadrature 
 hybrid 
 
 Sample 
 and hold 
 
 0" 
 
 90' 
 
 Q mixer 
 
 XMIT+IF 
 reference 
 
 1st mixer 
 
 Antenna 
 
 Amplifier 
 
 RECEIVER 
 
 Antenna 
 
 ^ 
 
 FIGURE 2. - Block diagram of the synthetic pulse radar system. 
 
79 
 
 -Q 
 
 -a 
 
 c 
 a 
 
 -a 
 
 c 
 
 3 
 
 o 
 
 E 
 
 0) 
 
 >. 
 
 
 3 
 O 
 
 a> 
 
 U 
 
 a 
 
 J3 
 
80 
 
81 
 
 Receiver station 
 
 170' 
 
 170' 
 
 Transmitter 
 
 FIGURE 5. - Main travel paths for radar sig- 
 nals through a 170-ft coal pillar. 
 
 SYNTHETIC PULSE FIELD TESTS 
 
 The synthetic pulse system was tested 
 at four separate sites to determine its 
 capabilities of detecting hazards in a 
 coal seam. Two tests were in Eastern 
 coal mines and two were in Western coal 
 mines. Figure 5 shows possible travel 
 paths from a transmission test through 
 170 ft of coal at Consolidation Coal 
 Co.'s Humphrey #7 Mine near to Morgan- 
 town, WV. These paths include direct 
 transmission, reflection off the side 
 
 CO 
 UJ 
 
 > 
 
 LiJ 
 
 a 
 us 
 tr 
 
 30 
 
 20 
 10 
 
 
 ^ — AA/n/vk/ 
 
 100 200 300 400 
 TIME, ns 
 
 500 600 
 
 FIGURE 6. - Synthetic pulse radar signals 
 through 170 ft of coal. 
 
 rib, and refraction along the back rib. 
 Figure 6 shows the synthesized pulses 
 as the receiver was moved along the back 
 rib. Additional "ghost" signals are 
 present in this data, because at the time 
 of this test the power amplifier was in 
 a separate box from the antenna and ring- 
 ing occurred in the connecting coaxial 
 cable. This problem was later solved by 
 mounting the amplifier directly on the 
 antenna. 
 
 The field tests were very successful. 
 The synthetic pulse system more than 
 doubled the probing distance of previous- 
 ly tested short pulse systems. The sys- 
 tem was also successfully used in deter- 
 mining the in situ electrical properties 
 of coal. However, the ultimate range and 
 limitations of the system have yet to be 
 determined. 
 
 SEISMIC METHODS 
 
 Seismic methods in underground coal 
 mines are divided into three categories. 
 The first uses relatively high frequen- 
 cies for near-field, high-resolution of 
 smaller reflection targets. The second 
 uses guided elastic waves for delineation 
 of major anomalies in the far-field. The 
 third, a borehole technique, provides 
 calibration velocities for the first two 
 techniques with the potential for mapping 
 coal thickness from a horizontal hole. 
 
 HIGH-FREQUENCY TECHNIQUE 
 
 Operating frequencies in the range of 
 5,000 Hz have potential for resolving 
 targets as small as 6-in diam within 20 
 to 30 ft of the rib or face. Such reso- 
 lution is necessary to locate small voids 
 (e.g., well bores) and small faults ahead 
 of the face. An advantage of the high- 
 frequency technique is that the source 
 and receiver can be configured to 
 
82 
 
 provide a directional beam of energy for 
 a relatively narrow field of view. Thus 
 the boundary effects of reflections from 
 the roof and floor are minimized. 
 
 Figures 7 and 8 illustrate a prototype 
 system used to demonstrate the high- 
 frequency technique. The system was de- 
 veloped under a Bureau of Mines contract 
 by the Energy and Minerals Research Co. 
 In figure 7, the source and receiver 
 transducers are shown mounted side- 
 by-side on the rib of a coal pillar. The 
 source and receiver are piezoelectric 
 transducers tuned to the same (5,000 Hz) 
 frequency with a relatively low-Q (broad- 
 band) response found appropriate to cou- 
 ple signal energy into coal media. Elec- 
 tronic circuitry (fig. 8) includes a var- 
 iable pulse-width drive circuit for the 
 source transducer, and signal condition- 
 ing circuits for the receiver output. 
 
 The system was calibrated by transmit- 
 ting a pulse through the 18-ft-wide pil- 
 lar and measuring the traveltime with a 
 receiver mounted on the opposite rib 
 
 (fig. 9). Velocity of sound through the 
 coal was determined to be 6,922 ft/s, 
 consistent with laboratory measurements 
 of coal samples from this seam. With 
 this information, and by selectively fil- 
 tering the received signal on the same 
 rib (fig. 7), an easily discernible echo 
 from the opposite rib was observed. 
 These tests show that it is possible to 
 obtain a reflection from a planar inter- 
 face at distances of at least 18 ft. The 
 detection circuitry, with the velocity 
 calibration, lends itself to numerical 
 readout of distance to the reflector. 
 This will provide an on-site interpreta- 
 tion of the results and allow an un- 
 skilled operator to scan a volume of coal 
 ahead of the face for potentially hazard- 
 ous conditions. 
 
 In the preliminary tests of the high- 
 frequency system, transducers were bonded 
 to the coal rib with resin grout to 
 achieve adequate coupling of energy into 
 the coal. Subsequent research has empha- 
 sized development of a force-insensitive 
 mounting configuration, which will be a 
 
 FIGURE 7. - High-frequency piezoelectric transducers mounted on the rib of a coal pillar. 
 
83 
 
 FIGURE 8. - Breadboard electronic circuitry for pulse shaping, signal conditioning, and detection 
 of reflections. 
 
 FIGURE 9. • Small piezoelectric transducer mounted on the opposite rib for velocity calibration. 
 
84 
 
 hand-held scanning unit employed at vari- 
 ous locations to map anomalous conditions 
 ahead of the face. It was found that an 
 equivalent amount of transmitted energy 
 can be achieved with reduced drive volt- 
 age on the transmitter by employing gated 
 bursts of single frequency, rather than 
 single pulses. With lower power levels 
 for energy transmission, the system can 
 readily be made intrinsically safe for 
 operation in coal mines. 
 
 GUIDED WAVES 
 
 Under certain conditions, a coal seam 
 may be a waveguide for long-range propa- 
 gation of seismic energy. This requires 
 a seam that is bounded by roof and floor 
 rock that have a greater density and a 
 higher sound velocity than the coal. 
 When these criteria are met, as they gen- 
 erally are in nature (the coal seam being 
 bounded by shale or sandstone), then the 
 multiply reflected waves from the roof 
 and floor will constructively interfere 
 to produce resonant modes of propagation 
 that undergo less attenuation than the 
 direct body waves. These normal resonant 
 modes are dispersive (different frequen- 
 cies travel at different speeds) and are 
 referred to as Rayleigh- or Love-type 
 waves because of their similarity to 
 earthquake-generated waves, which travel 
 large distances over the earth's surface. 
 
 The dispersive nature of the normal 
 modes and the restrictive nature of the 
 underground environment complicate the 
 application of the technique, but the po- 
 tential benefits of far-range detection 
 of faults and abandoned workings justify 
 research to develop the concept. 
 
 Two procedures are used in underground 
 mines: through-transmission and reflec- 
 tion surveys. Seismic sources used are 
 generally small explosive charges placed 
 in drill holes or hammer blows on the 
 rib. Because explosives present an obvi- 
 ous safety hazard, and because hammer 
 blows are not reliably repeatable, the 
 Bureau developed, under a contract, 
 controlled-source piezoelectric trans- 
 ducers (fig. 10) to generate predominant- 
 ly compressional or shear wave energy. 
 
 The most desirable type of wave from the 
 standpoint of simplicity in interpreta- 
 tion is a horizontally polarized shear- 
 wave, which will be totally internally 
 reflected within the seam for the normal 
 modes of the Love type. The shear wave 
 source and receiver can be mounted on the 
 rib of a coal pillar (fig. 11) to prefer- 
 entially excite horizontal particle mo- 
 tion and generate the Love-type modes. 
 
 The shear wave source and matching re- 
 ceiver were used to demonstrate the 
 through-transmission and reflection meth- 
 ods in a coal pillar at the Bureau's 
 Safety Research Coal Mine, Bruceton, PA 
 (fig. 12). Waveforms recorded at the re- 
 ceiver directly opposite from the trans- 
 mitter having the shortest travel path 
 are reproduced in figure 13. The top 
 
 FIGURE lOo - Controlled-source shear wave 
 transducers. 
 
 FIGURE IL • Shear wave source mounted on 
 the rib of a cool pillar to generate horizontal 
 vibration. 
 
85 
 
 
 « 37.0' i 
 
 11.0' 
 
 ■ 50.0' i 
 
 
 
 
 
 
 ^Detector 
 
 u^n 1 
 
 82 
 
 
 
 Detector 
 \ 
 
 N 
 \ 
 N 
 \ 
 
 s 
 
 
 «s 
 
 
 "^ 
 
 
 Source 
 
 20 
 
 Scale, ft 
 
 FIGURE 12. - Plan view of the test pillar at the 
 Bruceton Mine with transducer locations and trav- 
 el paths for seismic signals. 
 
 FIGURE 13. • Waveforms recorded at the nearest 
 receiver location. The shear wave source has been 
 rotated 180° to obtain the bottom trace, reversing 
 the polarity of the arrival indicated by the arrow. 
 
 trace shows a low-amplitude first arrival 
 followed by a large secondary signal, 
 which Is Interpreted as the shear wave 
 traveling In the coal seam at a speed of 
 
 approximately 2,300 ft/s. Support for 
 this interpretation is provided by the 
 lower trace, which was obtained by re- 
 versing the pulse direction of the source 
 transmitter. The shear wave arrival 
 clearly shows a 180° phase shift (arrows) 
 as expected. 
 
 In figure 14, the waveforms recorded at 
 the two receiver locations are compared. 
 Here the top trace is an amplified ver- 
 sion of the previous data for the nearest 
 receiver. The directly transmitted shear 
 wave arrival is again indicated, and a 
 later very similar arrival with lower am- 
 plitude and reversed polarity is clearly 
 evident at about 100 ms. Arrival time 
 for this event corresponds to reflection 
 of the shear wave from the end of the 
 coal pillar; a polarity reversal would be 
 expected from the negative reflection co- 
 efficient at the coal-air interface. The 
 bottom trace shows the waveform at the 
 far receiver position of figure 12. The 
 direct shear wave is obscured in the 
 early portion of the trace where some ap- 
 parent resonance or ringing appears; how- 
 ever, after the amplitudes die off, an 
 arrival at about 150 ms corresponds to 
 
 FIGURE 14. - Waveforms recorded at the near 
 receiver (top) and far receiver (bottom). 
 
86 
 
 reflection from the end of the pillar 
 over the longer travel path for the sec- 
 ond receiver position. It should be not- 
 ed in figure 12 that the various travel 
 paths are at different angles to the 
 direction of particle motion excitation; 
 for the far receiver position, propor- 
 tionately more compressional wave com- 
 ponent of energy would be detected, 
 possibly contributing to the higher am- 
 plitude early arrival and obscuring the 
 direct shear arrival. 
 
 BOREHOLE TESTS 
 
 When explosives or hammer blows are 
 used in underground seismic surveys, the 
 more complex waveforms generated require 
 accurate determination of both the com- 
 pressional and shear wave velocities for 
 interpretation. A borehole sonic log- 
 ging probe was adapted for use in a hori- 
 zontal drill hole to investigate veloci- 
 ties near the edges of a coal panel. The 
 probe is a dual-receiver, two-component 
 tool designed to selectively detect par- 
 ticle motion parallel or radial to the 
 borehole axis (compressional or shear 
 waves, respectively). Hydraulic pistons 
 clamp the transducers in rigid contact 
 with the borehole wall, and the probe can 
 be used in either fluid-filled or dry 
 holes. Figure 15 illustrates the probe 
 
 Transmitter 
 
 Waveforms recorded 
 at near detector 
 
 S'Wave P-wave 
 channel channel 
 
 Waveforms recorded 
 at far detector 
 
 FIGURE 15. - Diagram of dry hole sonic probe 
 and sample waveforms. 
 
 configuration, with typical waveforms 
 observed on the compressional and shear 
 wave channels at the two receivers. 
 
 The logging probe was positioned at a 
 starting depth of 17 ft in a drill hole 
 in a coal pillar (fig. 16). Transducers 
 were clamped to the borehole wall, 
 and full waveform recordings were taken 
 on the compressional and shear wave 
 channels for both receivers , Rl spaced 
 at 4 ft and R2 6 ft from the transmit- 
 ter. Pistons were then retracted and 
 the process was repeated at 1-ft inter- 
 vals toward the rib. Velocity deter- 
 minations from these data are plotted 
 in figures 17 and 18. Both the compres- 
 sional and shear wave velocities exhibit 
 low values near the rib, reach a shoulder 
 a few feet into the rib, and tend to 
 level off toward a constant value with 
 depth. 
 
 2 3 15 6 Z a 9 10 I! 12 13 14 
 
 4-in-diam 
 
 Start position 
 Tx depth (17ft) 
 
 . End position 
 Tx depth (4ft) 
 
 in-- 10' 
 
 FIGURE 16. - Cross section view of horizontal 
 drill hole in a coal pillar and locations of veloc- 
 ity measurements. 
 
 2,000 
 
 2 
 
 4 6 8 10 12 14 16 
 
 SOURCE-RECEIVER MIDPOINT DEPTH, ft 
 
 FIGURE 17. - Compressional wave velocities 
 versus depth of measurement in a coal pillar. 
 
87 
 
 5,000 
 
 o 3,000 
 
 2 4 6 8 10 12 14 16 
 
 SOURCE-RECEIVER MIDPOINT DEPTH, ft 
 
 FIGURE 18. - Shear wave velocities versus 
 depth of measurement in a coal pillar. 
 
 The two lower plots in figures 17 and 
 18 represent average velocity deter- 
 minations over the 4-ft and 6-ft travel 
 paths from transmitter to receiver, and 
 include the effects of variation in 
 borehole diameter and electronic delays 
 in the probe circuitry, producing some- 
 what low values. The upper curves rep- 
 resent differential determination of 
 velocities over the 2-ft travel path 
 between the two receivers. They should 
 represent more realistic values because 
 any constant electronic delays are ac- 
 counted for; however, the values are 
 more variable because of greater sensi- 
 tivity to local anomalies in the borehole 
 wall. 
 
 4,000 
 
 3,000 
 
 UJ 
 
 2,000 
 
 1,000- 
 
 4 6 8 10 
 DISTANCE, ft 
 
 FIGURE 19. • Traveltime curves for determin- 
 ing average velocities within the coal pillar. 
 
 Average velocities for calibration of 
 the seismic surveys were established by 
 plotting a summation of the differential 
 traveltimes versus distance along the 
 borehole (fig. 19), neglecting the low- 
 er velocity values shallower than the 
 shoulder at about 4-ft depth in fig- 
 ures 17 and 18. A good correlation is 
 achieved with this method using linear 
 regression to provide values of shear and 
 compressional wave velocities which are 
 consistent with previous well-log sonic 
 data in vertical exploration holes. 
 
 CONCLUSION 
 
 Four geophysical methods — synthetic 
 pulse radar, high-frequency seismic, 
 guided waves, and borehole velocity log- 
 ging — were investigated by the Bureau to 
 devise better in-seam hazard detection 
 techniques. It is anticipated that these 
 techniques will complement each other and 
 will provide a valuable tool to the min- 
 ing industry. The synthetic pulse radar 
 provides high-resolution reflection capa- 
 bilities in the range between 10 ft and 
 200 ft. The high frequency seismic meth- 
 od, when fully developed, should provide 
 near-range, up to 30 ft, high-resolution 
 reflection detection capabilities, and 
 
 the lower resolution, long-range guided- 
 wave method using the control sources 
 will provide detection in the hundreds of 
 feet. The borehole sonic probe will be 
 useful for determining the seismic veloc- 
 ities needed for both the high-frequency 
 seismic and guided-wave methods. 
 
 Research is continuing on all four 
 methods. This paper demonstrates the po- 
 tential of the methods. Further re- 
 search, development, testing, and actual 
 use are required to reach the expectation 
 that is expected for each method. 
 
88 
 
 BIBLIOGRAPHY 
 
 Snodgrass, J. J. A New Sonic Velocity- Suhler, S. A., T. E. Owen, B. M. Duff, 
 Logging Technique and Results in Near- and R. J. Spiegel. Geophysical Hazard 
 Surface Sediments of Northeastern New Detection From the Working Face (contract 
 Mexico. BuMines TPR 117, 1982, 24 pp. H0272027, Southwest Res. Inst.). BuMines 
 
 OFR 69-83, 1981, 176 pp. 
 
SATELLITE IMAGERY AS AN AID TO MINE HAZARD DETECTION 
 By Robert A. Speirer'' and Stanton H. Moll^ 
 
 89 
 
 ABSTRACT 
 
 The Bureau of Mines is involved in 
 ongoing research to develop potential 
 hazard evaluation maps for mine areas . 
 These maps will be generated using 
 computer-aided methods to analyze Land- 
 sat imagery and multivariate data sets. 
 A means of image processing whereby 
 lineament information is enhanced and 
 
 extraneous information suppressed has 
 been devised. Digital processing is 
 particularly appropriate for picking 
 lineaments from Landsat data because it 
 is faster, less biased, more repeatable 
 and, ultimately, less costly than manual 
 interpretation. 
 
 INTRODUCTION 
 
 The geologic environment in mining ar- 
 eas directly influences the safety of 
 mine workers. Where possible hazards 
 exist, it is essential that their nature 
 and location be determined before mining 
 into them. The basic mine plan can then 
 be modified at an early date for safety 
 and economy. 
 
 Geologic features in coal mines , such 
 as faults or sand channels, cause zones 
 of weakness in the roof due to fracturing 
 or differential compaction. These fea- 
 tures usually require that roof bolt 
 plans in their vicinity be modified. 
 Roof or rib falls in underground coal 
 mines and slides in open pits are related 
 to these geologic features and still 
 account for a large number of fatali- 
 ties. Although fatalities were reduced 
 since the enactment of the Federal Coal 
 Mine Health and Safety Act of 1969, roof 
 falls still cause some 40 pet of all 
 
 underground coal mine fatalities and dis- 
 abling injuries. 
 
 Detection of possible zones of weakness 
 using data derived from Landsat satel- 
 lites has been demonstrated by many re- 
 searchers. Since the zones of weakness 
 that may cause mine hazards may be re- 
 flections of discontinuities in the 
 earth's crust, surface expressions of 
 such discontinuities should be, and of- 
 ten are, visible on Landsat images to 
 the trained interpreter. Much effort 
 has been devoted to mapping lineaments 
 (linear features often representing dis- 
 continuities) and demonstrating corres- 
 pondences between lineaments and natu- 
 ral features (i.~2^)»^ Additional effort 
 has been applied to the specific task of 
 delineating lineaments in mine areas (2) , 
 The Mine Safety and Health Administration 
 (MSHA) also conducts a program to provide 
 technical support for lineament analysis 
 to mine operators. 
 
 MANUAL LINEAMENT ANALYSIS 
 
 Current technology consists of a 
 trained operator using visual techniques 
 to plot linear features onto a base map 
 from Landsat images. Rinkenberger (j4) 
 used this method to demonstrate the cor- 
 relation of lineaments with known faults 
 and fractures. Generally, such methods, 
 including those used by MSHA, involve the 
 
 'Geologist, Denver Research Center, Bu- 
 reau of Mines, Denver, CO. 
 
 use of analog equipment for edge enhance- 
 ment and density slicing (mapping ranges 
 of brightness to a single color or 
 brightness) of subsets of standard Land- 
 sat scenes . The visually enhanced images 
 are manually interpreted to obtain the 
 
 ^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
90 
 
 lineament map, which is then registered 
 and overlaid onto a base map. 
 
 An example of a lineament map over a 
 mine area is shown in figure 1. The 
 lineaments have been manually interpreted 
 from a Landsat scene. When overlaid on a 
 mine outline, a high correlation is seen 
 between the lineaments and known roof 
 fall occurrences (small dots). With the 
 aid of a lineament map , the mine operator 
 has some warning of zones of potential 
 hazards and can use the map as a guide in 
 anticipating hazardous ground conditions 
 (large dots). Verification of these haz- 
 ards may lead to installing additional 
 roof support or to modifying the mine 
 plan to avoid the hazards. 
 
 Although this method has proven via- 
 ble, and is generally accepted by mine 
 planners, it has a number of substantial, 
 interrelated drawbacks. These include: 
 
 1. Repeatability . Repeated interpre- 
 tations of the same image are rarely, if 
 ever, identical. Two interpreters will 
 seldom agree about the length, azimuth, 
 or number of lineaments . Even the same 
 trained interpreter rarely obtains the 
 same results on subsequent trials. 
 
 2. Bias . Each interpreter brings his 
 or her own biases to the task. Fur- 
 thermore, since these are an unconscious, 
 and hence, of unknown quantity, they are 
 extremely difficult to control and 
 
 MINE 'a' 
 
 FIGURE 1. - Lineament map registered to a mine map, generated by the manual method. The areas 
 where greater roof instability may be expected are shown in large dots. 
 
91 
 
 compensate for. For example, an inter- 
 preter with some previous knowledge of 
 the structure of an area will probably 
 color the interpretation with those pre- 
 conceptions. One interpreter may prove 
 adept at locating "local" or short fea- 
 tures, whereas another may, for example, 
 preferentially locate northeast-trending 
 features over north-trending lineaments. 
 Again, the individual biases may change 
 from trial to trial. 
 
 3. Experience . The interpreter must 
 be trained in the "art" of lineament 
 analysis. The more experience an in- 
 terpreter has , the more he or she 
 will be able to control the problems of 
 
 repeatability and bias. At the same 
 
 time, however, the cost of analysis 
 
 will increase with the increase in 
 experience. 
 
 4. Speed and cost . Although a Lands at 
 scene can be analyzed by a trained inter- 
 preter using just an image and a pencil, 
 the process is slow and tedious. The 
 cost is low because it consists of only 
 an interpreter's wages and the cost of 
 the image. Enhancement devices, at added 
 cost, can be used to speed up the analy- 
 sis. Such devices do not, however, re- 
 lieve the tedium of the manual interpre- 
 tation process nor eliminate the other 
 drawbacks . 
 
 COMPUTER-AIDED LINEAMENT ANALYSIS 
 
 Because of the success of the manual 
 lineament analysis technique, the Bureau 
 is devising a program to improve the 
 technique with computer processing. Com- 
 puter processing and advanced analytical 
 techniques can reduce or eliminate the 
 drawbacks of the manual method while im- 
 proving throughput and making the tech- 
 nology more widely accessible. 
 
 The advantages of using digital meth- 
 ods for lineament analysis are several. 
 These include: 
 
 1. Suitability . Landsat data is dis- 
 crete digital data, and hence lends it- 
 self readily to digital processing. 
 Images are available from EROS Data 
 Center, Sioux Falls, SD, that have al- 
 ready been enhanced by digital means to 
 Improve image quality. Scenes can also 
 be acquired in digital form as Computer 
 Compatible Tapes (CCT's) and further 
 processed to enhance desirable features 
 such as lineaments. 
 
 2. Repeatability . If identical param- 
 eters are entered into a program, the 
 computer will repeatedly find an identi- 
 cal crop of lineaments from the same 
 s cene . 
 
 3. Controlled bias . The computer can 
 be directed to find features of a given 
 size or shape and can be expected to find 
 
 all features for which it has been di- 
 rected to search. Different processing 
 methods can enhance or suppress certain 
 features. These capabilities cannot be 
 expected from a human interpreter. Most 
 importantly, the biases of the computer 
 can be known, whereas those of a human 
 interpreter cannot. 
 
 4. Flexibility . Because computers are 
 more easily "retrained" than human inter- 
 preters, they can be directed to perform 
 many different tasks. 
 
 5. Speed and cost . Although initial 
 costs for the computer method are like- 
 ly to be greater than for the manual 
 technique, this disadvantage is soon 
 overcome by greater throughput. Increased 
 reliability and repeatability, and less 
 need for a trained Interpreter. Soft- 
 ware development costs may be high for a 
 single application, but they can usually 
 be amortized and shared among several 
 uses. Less time is needed to train an 
 operator for the image processing system 
 than to train an interpreter for the man- 
 ual technique. 
 
 An earlier study, now complete, was 
 made to try to devise a computer tech- 
 nique to detect lineaments in Landsat 
 scenes. After transferring the linea- 
 ments to a base map , they were field 
 checked to verify the correspondence 
 
92 
 
 between the plotted lineaments and sur- 
 face geology. The technique proved suc- 
 cessful in discovering some features in 
 the mining areas where it was tested. It 
 was apparent, however, that further re- 
 finement was needed to learn how to dis- 
 criminate between manmade features such 
 as roads, canals, fields, vapor trail 
 
 shadows, etc., and natural phenomena such 
 as faults, fractures, joints, and paleo- 
 channels. Furthermore, the technique 
 proved to be "blind" to lineaments in 
 certain orientations, and also to other 
 nonlinear features that may be important, 
 such as circular or serpentine patterns 
 (e.g. calderas and thrust faults). 
 
 AUTOLINER PROJECT 
 
 The Bureau has recently concluded an 
 interagency project with the U.S. Geo- 
 logical Survey to develop a method of 
 automatic lineament mapping O ) . The 
 Autoliner project, although not capable 
 of generating a lineament map per se, did 
 result in a method of processing the 
 Landsat scenes to enhance those natural 
 features of interest to the image analyst 
 while suppressing extraneous information. 
 The resultant image is far superior to a 
 standard image for lineament analysis. 
 
 AUTOLINER METHOD 
 
 Basically, the autoliner works as 
 follows: 
 
 A Landsat satellite records the reflec- 
 tance of an area on the ground in each of 
 four spectral bands. The smallest area 
 that can be resolved, called a pixel 
 (picture element), measures about 75 m 
 by 75 m. Each pixel, in each band, is 
 sent back to earth as a value between 0^ 
 (for no reflectance) to M^ (for total 
 reflectance) . 
 
 On earth the data is cleaned up (pre- 
 processed) to correct for such aberra- 
 tions as differences in detector calibra- 
 tion, atmospheric haze, data transmission 
 errors, and geometric distortion. When 
 this procedure is complete, the data 
 still comprises four bands ranging in 
 value from 0^ to W^ each, but now is geo- 
 graphically correct. An image made at 
 this point would be brighter and easier 
 to interpret than an image made before 
 preprocessing. These images are marketed 
 by EROS Data Center, and are the images 
 usually used in manual interpretations. 
 
 A great deal more processing can be 
 done, however, to enhance specific fea- 
 tures. Since the pixel values (called 
 density numbers, or DN) are directly pro- 
 portional to reflectivity, the difference 
 between pixel values is their contrast. 
 Linear features will usually have a mod- 
 erate contrast, which will be due to veg- 
 etation or elevation differences across 
 the feature. Other features, such as 
 snow-field boundaries, will have extreme 
 contrast, whereas very low contrast is 
 usually minor in content. Consequently, 
 a window can be specified. Outside this 
 window data can be ignored (or set to no 
 contrast), whereas contrasts inside the 
 window can be magnified to further in- 
 crease the contrast. 
 
 There are many ways of establishing the 
 contrast of a pixel with its surround- 
 ings . A pixel may be compared with a 
 single immediate neighbor or with any 
 number of surrounding pixels. A small 
 area will more closely reflect the value 
 of its central pixel than will a large 
 area, hence the more distant pixels 
 should be more suppressed than the proxi- 
 mal pixels. These methods of establish- 
 ing and enhancing contrast were studied 
 in the Autoliner project. 
 
 The results of the optimal method for 
 enhancing lineament information is shown 
 in figure 2. Called a "thematic linear 
 feature map," the values within the se- 
 lected window are shown in black, and all 
 contrast values outside the window are in 
 white. Selected lineaments are shown by 
 pairs of arrows. 
 
93 
 
 FIGURE 2. - Thematic (binary) lineament feature map of a Denver Landsat 3 image, 
 produced using the modified-modified gradient method. Arrows show lineaments. 
 
 AUTOLINER CONCLUSIONS 
 
 The Autoliner project resulted in prom- 
 ising techniques of image processing and 
 enhancement for lineament analysis. 
 These techniques permit the highlighting 
 of data that contain lineament informa- 
 tion, while suppressing information ex- 
 traneous to the interpretation. However, 
 the procedures outlined so far must still 
 be interpreted by a trained lineament 
 analyst. That is, they are still subject 
 to visual analysis for lineament picking, 
 albeit with a much improved product. 
 
 The image processing software used in 
 the Autoliner project will shortly become 
 
 available to the public. MIPS (Minicom- 
 puter Image Processing System) was de- 
 signed to operate on a DEC PDP-11 mini- 
 computer, model 23 or higher, running the 
 RSX-llM operating system. Color graphics 
 are provided by a Grinnel model GMR 27 or 
 270 image processing display system, with 
 hardcopy output via color Optronics or 
 high-resolution Dunn camera.-^ 
 
 ^Reference to a specific brand, equip- 
 ment, or trade name in this report is 
 made to facilitate understanding and does 
 not imply endorsement by the Bureau of 
 Mines. 
 
94 
 
 Most of the processing routines in MIPS 
 are written in DEC FORTRAN IV PLUS, and 
 hence should be readily transportable to 
 other computers. However, since image 
 processing deals with such large data 
 sets, the decision was made to optimize 
 I/O (data input and output operations) 
 and data transformation routines by 
 
 writing them in machine-specific assembly 
 language, and using low-level operating 
 system routines. Consequently, translat- 
 ing these routines to another computer 
 will involve some investment of time and 
 probably greater program execution time 
 as well. 
 
 FUTURE RESEARCH 
 
 The primary objective of Bureau re- 
 search with the MIPS system is to develop 
 a potential hazard map for use by mine 
 planners. A similar product acquired by 
 the manual method was shown in figure 1. 
 Using the MIPS system, it is hoped that 
 this product can be improved in two ways : 
 (1) by incorporating the "Autoliner" 
 methodology and its derivatives, and (2) 
 by integrating other data sets into the 
 hazard analysis. 
 
 Figure 3 shows a prototype of a hazard 
 map generated using digital means. Al- 
 though numbers have not , at this point , 
 been assigned to the contours, they could 
 represent a variety of quantities, such 
 as degree of hazard or length of roof 
 bolts to be installed. Assessing the 
 parameters to incorporate into the plan, 
 and their respective weights, is one of 
 the objectives of the Bureau's research. 
 
 
 
 Potential 
 hazard 
 
 □ High 
 
 □ Med 
 
 □ low 
 
 FIGURE 3. - Preliminary potential hazard map of the some area as shown in figure 1, to be gen- 
 erated by integration of Autoliner results and geologic, geophysical, geochemical, and miscellane- 
 ous data sets. Contours shown represent arbitrary units and values. 
 
95 
 
 For example, lineament analysis alone 
 cannot provide data about the degree of 
 weakness each lineament represents. For 
 this information, ancillary and comple- 
 mentary data sets must be used. Part of 
 the Bureau's research will be to assess 
 the applicability of these data sets. 
 
 Data sets which are anticipated as being 
 useful include radar, magnetics, gravity, 
 digital elevation data, geologic mine 
 maps, previous roof -fall data, drilling 
 data, methane and helium concentrations, 
 and other geophysical, geochemical, geo- 
 logical, and miscellaneous data. 
 
 CONCLUSIONS 
 
 Computer applications are becoming com- 
 monplace in mine planning, and we expect 
 that they will soon be common in daily 
 operations as well. Furthermore, given 
 the massive amounts of mine-related data 
 available, the only feasible means of 
 assimilating it is by computer. Our re- 
 search was undertaken to assist mine 
 planners and operators in promoting safe 
 and economic operating conditions. 
 
 Since much of the current work in pro- 
 viding lineament information to mine 
 planners is done by MSHA, the Bureau in- 
 tends to work closely with MSHA in devel- 
 oping the methodology. It is hoped that 
 eventually MSHA will have image process- 
 ing systems in their field offices where 
 a semi-automatic analysis can be generat- 
 ed locally for mine operators. 
 
 REFERENCES 
 
 1. Lillesand, T. M. , and R. W. Kiefer. 
 Remote Sensing and Image Interpretation. 
 Wiley, 1979, 597 pp. 
 
 2. Short, N. M. The Landsat Tutorial 
 Handbook. NASA Ref. Publ. 1078, 1982, 
 553 pp. 
 
 3. Burdick, R. G. , and R. A. Speirer. 
 Development of a Method To Detect Geo- 
 logic Faults and Other Linear Features 
 From LANDSAT Images. BuMines RI 8413, 
 1980, 74 pp. 
 
 4. Rinkenberger , R. K. Implementing 
 Remote Sensing Techniques for Evaluating 
 Mine Ground Stability. Mining Enforce- 
 ment and Safety Administration (now Mine 
 Safety and Health Administration), Inf. 
 Rep. 1057, 1977, 34 pp. 
 
 5. Chavez, P. S., Jr. Autoliner Pro- 
 ject. BuMines Interagency Agreement 
 J0215036; for inf., contact R. A. 
 Speirer, TPO, Denver Research Center. 
 
96 
 
 MICROSEISMIC TECHNIQUES APPLIED TO FAILURE WARNING IN MINES 
 By Fred W. Leighton^ 
 
 ABSTRACT 
 
 Miners have long known that rock noise, 
 or the popping and cracking of the rock 
 commonly heard during mining, can be in- 
 dicative of the stability of the mine 
 structure. For many years, miners have 
 "listened" to the rock talk and many 
 times have interpreted changes in rock 
 noise activity to be a warning of fail- 
 ure and have retreated from the failure 
 area. Microseismics , or the study of 
 rock noises, was begun in the early 
 I940's, partly because of this historical 
 fact. 
 
 Microseismics uses geophysical equip- 
 ment to detect and analyze rock noises on 
 both the audible and subaudible level. 
 Thus, these systems are much more sensi- 
 tive than the human ear and "hear" even 
 more of the rock "talk" than do miners. 
 
 Research has shown that microseismics 
 can be used to precisely locate those 
 portions of a working area that are gen- 
 erating rock noise, and that the rock 
 noise release rate information from each 
 area can be used to analyze its stabil- 
 ity. In some instances, rock noise data 
 have been analyzed to provide warning of 
 imminent structural failure, and person- 
 nel have been removed from or prohibited 
 from entering a failure area. This paper 
 presents a brief history of how micro- 
 seismics evolved, explains why the tech- 
 nique works, and describes the basic 
 equipment used. Past results in both 
 coal and metal and nonmetal mining sys- 
 tems are shown, and recent results con- 
 cerning the occurrence of a failure in a 
 coal mine advancing longwall section are 
 presented. 
 
 INTRODUCTION 
 
 When a rock mass is subjected to chang- 
 ing stress conditions, such as those 
 caused by mining, small-scale adjustments 
 occur within the rock that release seis- 
 mic energy. This energy, when in the 
 audible range, is called rock noise. 
 Those areas in which stress changes occur 
 are also the areas of the structure most 
 likely to fail. Individual rock noises 
 can be detected and analyzed to determine 
 their precise location relative to the 
 mine structure. Over a period of time, 
 plots of these data provide a pictorial 
 representation of where stress changes 
 are occurring, because rock noise activ- 
 ity tends to concentrate in those areas 
 of the structure most actively adjusting 
 to the changing stresses. Since these 
 areas also represent those areas most 
 likely to fail, potential failure areas 
 can be pinpointed and mapped relative to 
 
 ^Supervisory mining engineer, Denver 
 Research Center, Bureau of Mines, Denver, 
 CO. 
 
 the structure. Rock noise rates, or the 
 number of rock noises occurring per unit 
 of time, also tend to vary dramatically 
 prior to failure. Thus, the ability to 
 locate the source of individual rock 
 noises provides a means of determining 
 where failure may occur, and rate count- 
 ing within each area offers a means of 
 assessing when failure may occur. This 
 information, properly treated, can be 
 used as an aid to help avoid, control, or 
 provide warning of impending failure. 
 Successful applications of this technique 
 have provided recent impetus to the ef- 
 fort of developing microseismics into a 
 practical, reliable, and economically 
 feasible tool. 
 
 The phenomenon of naturally occurring 
 rock noise was discovered in 1938 by 
 Obert (J_),2 who was measuring seismic 
 
 ^Underlined numbers in parentheses re- 
 fer to items in the list of references at 
 the end of this paper. 
 
97 
 
 velocities in mine pillars. Seismic 
 energy other than that which he was gen- 
 erating continually registered on the 
 recording equipment. Further study by 
 Obert and Duvall showed that the extrane- 
 ous seismic energy being recorded was 
 from rock noises that were being gener- 
 ated naturally within the rock in highly 
 stressed areas. Pursuing this interest- 
 ing phenomenon, both in the laboratory 
 and in the field, they documented the 
 dramatic change in rock noise rate prior 
 to failure (^~3^) • Carrying this work 
 further, they established the fact that 
 in many instances, rock burst failures 
 could be predicted by monitoring and 
 "listening" to the rock noise activity in 
 rock-burst-prone areas (2^). 
 
 These early efforts clearly showed that 
 microseismics had great potential as a 
 tool for measuring or estimating the sta- 
 bility of mine structures. Extended 
 testing, however, showed that sometimes 
 rock bursts occurred with no apparent 
 microseismic warning, and at other times, 
 sharp increases in microseismic data were 
 not accompanied by failure. Also, be- 
 cause precise location of individual rock 
 noises was not at that time possible, one 
 never knew where the failure was going 
 to occur, only that failure nearby the 
 
 observation point was likely. Thus, 
 
 while the technique clearly offered 
 
 promise, it was not at that time consid- 
 ered practical. 
 
 In the mid-1960' s, the Bureau of Mines 
 began a new research effort to improve 
 the microseismic technique. Major im- 
 provements were judged possible, in great 
 part due to the availability of new and 
 vastly improved electronic and system 
 components which had resulted from in- 
 strumentation developed during the space 
 program. Thus, in 1967, development of 
 a multichannel, broadband, microseismic 
 system began (^) • Application of this 
 system in rock-burst-prone mines showed 
 it to work well and provided the incen- 
 tive to develop methods whereby the 
 source location of individual rock noises 
 could be easily and directly calculated 
 0-_7) . The improved system and the abil- 
 ity to locate rock noises showed through 
 application that the microseismic tech- 
 nique offered new and increased potential 
 as a useful engineering tool for the min- 
 ing industry (8^) . The following sections 
 briefly describe the basis for microseis- 
 mics and current microseismic systems, 
 and offer examples of how the technique 
 has been and is being applied to mining 
 problems and failure warning research. 
 
 HOW MICROSEISMICS WORK 
 
 As has been stated, mining results in 
 ever-changing load and stress conditions 
 in the ore body and in the mine structure 
 support system. These changes are accom- 
 panied by rock noise, or the release of 
 low-level seismic energy, which can be 
 detected by geophones placed throughout 
 the mine structure. Each individual rock 
 noise can be accurately located relative 
 to the mine structure, thus delineating 
 those areas most actively adjusting to 
 mining. This process is simple and 
 straightforward. 
 
 An example of this phenomenon is shown 
 in figure 1 , which depicts a plan view of 
 a rock noise occurring in the barrier 
 pillar of a room-and-pillar retreat sec- 
 tion in a coal mine. The seismic energy 
 released by the rock noise travels out- 
 ward from the source of the rock noise 
 in all directions at some measurable 
 
 velocity (not necessarily the same in all 
 directions), thus arriving at different 
 geophone locations at different times, as 
 shown in the example seismic record in 
 the lower portion of the figure. Knowing 
 the coordinates of each geophone , the 
 velocity at which the seismic energy 
 travels, and the time at which each geo- 
 phone in the array sensed the arrival of 
 the seismic energy allows one to calcu- 
 late the coordinates of the rock noise 
 source. Each source is then plotted on 
 the mine structure map , and those areas 
 most actively adjusting to the new stress 
 regime are delineated by the areas in 
 which rock noise activity is most dense. 
 In this manner, potential failure areas 
 are delineated and pinpointed relative to 
 the mine structure. 
 
 Figure 2 shows such a rock noise 
 concentration in the two orthogonal 
 
98 
 
 Rock noise source 
 
 Seismic wave 
 
 Geophone 
 
 A Example geopiione array 
 
 2 
 
 3 
 
 4 
 5 
 
 6 
 
 /A/^/VA^.A ^. 
 
 -A^A/^AAr- 
 
 t. 
 
 -V\A/V- — -' — 
 
 t. 
 
 -V\A'*^ 
 
 - — A/\/v^ 
 
 TIME — ► 
 
 B Example record of rock noise 
 
 FIGURE 1. - Rock noise. A, Plan view of a 
 geophone array and rock noise in a coal mine 
 room-and-pillar section. B, Example of arrival 
 time information used to calculate rock noise 
 source location. 
 
 . (1 ' ■• •■ ■• 
 
 p 
 
 '• \ ' \ ''wi •' 
 
 t 
 
 
 1 
 
 1 
 
 • • 
 
 
 * 
 
 FIGURE 2. - Elevational views of rock noise 
 concentration in a rock-burst-prone pillar. 
 
 elevation views of a rock -burst-prone 
 pillar (a rock burst is a sudden massive 
 and sometimes catastrophic failure) . 
 This concentration happened over a 
 several-hour period and precisely located 
 the area of a rock burst that occurred at 
 the beginning of the day shift in the 
 area. This example shows the importance 
 of being able to locate where rock noises 
 are being generated and demonstrates the 
 ability of the technique to accurately 
 delineate problem areas. 
 
 Another feature of rock noise activity 
 is that the rate at which it occurs tends 
 to fluctuate dramatically in the area 
 prior to failure, which in many instances 
 provides a warning of the failure. Fig- 
 ure 3 shows the rock noise rate, or num- 
 ber of rock noises recorded per unit of 
 time, for the location data shown in 
 figure 2. Note the dramatic increase in 
 the rock noise rate prior to the rock 
 burst. 
 
 120 
 
 90- 
 
 z 
 
 i 60 
 
 HI 
 
 (A 
 O 
 
 Z 
 
 o 
 o 
 
 cc 
 
 30 
 
 T 
 
 Small burst 
 
 J L 
 
 _L 
 
 2 4 6 8 10 12 14 16 
 
 TIME, days » 
 
 FIGURE 3. - Rock noise rate prior to a rock burst. 
 
99 
 
 These two analyses procedures — i.e., 
 event location and rock noise rate 
 counting — in combination represent the 
 way in which microseismic techniques are 
 used in current practice, when using mul- 
 tichannel microseismic systems. As will 
 
 be shown, worthwhile applications of 
 single-channel systems to monitor local- 
 ized areas of interest are also possible, 
 using the rate counting method of analy- 
 sis without regard to the location of in- 
 dividual events. 
 
 MICROSEISMIC SYSTEMS 
 
 Microseismic systems may vary somewhat 
 according to the dictates of their appli- 
 cation, but essentially all systems in- 
 clude the same fundamental components. A 
 detailed examination of microseismic sys- 
 tems may be found elsewhere (9-11) , and 
 only an overview is presented here. 
 
 Figure 4 shows a block diagram of a 
 basic multichannel microseismic system. 
 Each channel consists of a sensor, a pre- 
 amplifier, the data transmission cable, 
 and sometimes a postamplif ier to condi- 
 tion the signal for final recording. The 
 sensor may be either an accelerometer or 
 a velocity gauge, depending on the appli- 
 cation. Each channel is connected to a 
 multichannel, high-speed magnetic tape 
 recorder and/or an automatic monitoring 
 system that electronically measures the 
 necessary information from the seismic 
 signals. The data measured by this sys- 
 tem are fed into a minicomputer or micro- 
 processor , where the data are analyzed to 
 
 Sensor 
 
 Preomp 
 
 High 
 
 gain 
 
 amplifier 
 
 
 FM to 
 
 pe 
 
 
 
 
 
 
 
 
 
 recorder 
 
 
 
 
 
 
 
 
 
 
 
 
 
 L- 
 
 3sciMograph 
 
 
 
 
 
 
 
 
 
 
 
 RBM 
 
 
 Computer 
 
 
 Plotter 
 
 
 
 
 
 
 
 1 
 
 
 
 
 Printer 
 
 
 FIGURE 4. - Block diagram of a multichannel 
 microseismic system. 
 
 compute the coordinates of the source of 
 each rock noise, which is then plotted on 
 a map of the mine. 
 
 Simpler versions of microseismic sys- 
 tems are available for application where 
 "listening" is carried out using only one 
 channel of equipment. Systems such as 
 these are used in certain specific in- 
 stances such as in roof fall warning mon- 
 itoring, where the area of interest is 
 well defined and limited in size. Exam- 
 ples of how these systems can be used are 
 discussed in "Field Applications." 
 
 Microseismic systems are now available 
 commercially, either as a total system or 
 by purchasing and assembling individual 
 components . While these systems are not 
 difficult to use, they do require full- 
 time attention and maintenance and are 
 not to be considered a "turnkey" opera- 
 tion requiring little or no ongoing 
 commitment of personnel and capital ex- 
 penditure. Sensors are sometimes lost 
 during mining and must be replaced, and 
 data transmission lines are often severed 
 or damaged, requiring attention. The 
 data recorded require daily, preferably 
 continuous, analysis and interpretation 
 to be of maximum value. Thus, a micro- 
 seismic system should be considered a 
 tool to be applied to stability problems 
 and requires the same dedication in terms 
 of attention and maintenance as do other 
 tools used in the mining routine. Sev- 
 eral applications of these systems have 
 shown this effort to be worthwhile and 
 contributory to a safer and more pro- 
 ductive mine. 
 
100 
 
 FIELD APPLICATIONS 
 
 Microseismic systems have been in- 
 stalled and are in use at a number of 
 locations throughout the world. Applica- 
 tions vary from those using single- 
 channel "listening" systems and rate 
 counting methods to those incorporating 
 24 or more channel systems and the com- 
 bination of source location and rate 
 counting methods. The following is 
 therefore broken into two discussions; 
 i.e., those applications using single 
 channel systems, and those using multi- 
 channel systems. 
 
 SINGLE-CHANNEL APPLICATIONS 
 
 In terms of failure warning, highly 
 promising research has recently been done 
 using equipment sensitive to the fre- 
 quency range of from 40 to 100 Hz (12- 
 13). A major advantage of this system 
 over the systems sensitive to the lower 
 frequency ranges is that manmade noise, 
 such as that due to mining, is comprised 
 mostly of frequencies lower than 400 kHz, 
 hence the data recorded are mostly, if 
 not entirely, made up of noises naturally 
 occurring in the mine structure. The 
 system can thus monitor right in the 
 working area, where there is the highest 
 risk for personnel. 
 
 Single-channel microseismic systems are 
 designed for monitoring well-defined 
 problem areas of limited extent. These 
 systems detect rock noise activity in 
 several frequency ranges, the most common 
 being in the 100- to 5,000-Hz range or 
 the 40- to 100-Hz range. An advantage of 
 single-channel units is that they are 
 portable and may be used for periodic 
 sampling of many areas within the mine, 
 or for continuous monitoring at the work- 
 ing area. The disadvantage is that the 
 source of the rock noise is never known, 
 hence the precise location of a potential 
 failure cannot be delineated. As will be 
 shown, this disadvantage does not pre- 
 clude beneficial use of single -channel 
 technology in many instances. 
 
 The system as discussed below was con- 
 structed specifically to provide warning 
 of impending roof falls. It was designed 
 to provide coverage of about a 50- or 
 100-ft radius from the sensor, so that it 
 would monitor only the working area mini- 
 mizing the importance of individual rock 
 noise locations. When a warning was 
 sounded, one would simply evacuate the 
 area immediately surrounding the sensor. 
 In practice, this system has been found 
 to perform well with large-scale roof 
 falls . Figure 5 shows the commercial 
 prototype of this device. The system is 
 a stand-alone design, providing continu- 
 ous monitoring, automatic data analysis, 
 and warning of failure independent of 
 human input . 
 
 Single-channel data may be recorded and 
 used in a variety of ways, ranging from 
 counting noises heard through a set of 
 earphones to permanently recording and 
 analyzing data using sophisticated elec- 
 tronics. In the former instance, the 
 success of the application relies heavily 
 on the dedication and skill of the ob- 
 server. In the latter case, the system 
 can perform its function essentially 
 unattended. Simple systems using head- 
 phones and an observer may cost in the 
 $1,000 range, while more sophisticated 
 systems may cost $10,000. The choice of 
 which system to use and consequently how 
 much money to invest is dictated by the 
 application. 
 
 This system differs from previous 
 single-channel monitoring systems in that 
 it measures not only the number of events 
 that occur within its range, as do stan- 
 dard systems, but also estimates the 
 total amount of energy released by those 
 events . The event count and energy value 
 are accumulated for 60-s intervals, and 
 then the energy value is divided by the 
 total event count to obtain the energy- 
 event ratio. This ratio calculation is 
 unique to this instrument. In applica- 
 tion, the energy-event ratio behaves 
 anomalously before failure, and the anom- 
 aly is sufficiently large to be easily 
 detectable and to be used as a failure 
 warning. 
 
101 
 
 FIGURE 5. - Components of a high-frequency, single sensor microseismic system. 
 
 Figure 6 shows the behavior of the 
 energy-event ratio before a large scale 
 roof fall (fig. 6A) and before a coal and 
 gas outburst, a coal mine failure similar 
 to a rock burst (fig. 6B ) . In both in- 
 stances, the anomaly is large and occurs 
 sufficiently prior to the actual failure 
 to ensure removal of personnel from the 
 failure area. 
 
 This system has shown much success in 
 providing warning of large-scale roof 
 falls. Its efficacy as a warning device 
 for small-scale roof falls, coal and gas 
 outbursts, and other types of failures 
 has not yet been determined, but it has 
 been shown to work on some occasions. 
 
 These areas of application are the 
 focus for current research efforts by 
 the Bureau of Mines with the goal of 
 establishing the procedures to ensure 
 proper use of the device in the field 
 and its reliability in different 
 applications. 
 
 Single-channel applications of this 
 device and others have been historically 
 proven to be of value and to be capable 
 of providing meaningful information about 
 the mine structure in specific areas of 
 limited size. To carry out similar stud- 
 ies over a larger area, even up to the 
 size of the entire mine, multichannel 
 systems are necessary. 
 
102 
 
 4:00 6:00 
 
 z 
 
 LU 
 
 > 
 Ui 
 
 >^ 
 
 oc 
 
 UJ 
 
 z 
 
 UI 
 
 8:00 10:00 
 TIME, h 
 
 12:00 
 
 8,834 
 7,848 
 6,867 
 5,886 
 4,905 
 3,924 
 2,943 
 1,962 
 981 
 
 
 1 
 
 Event 
 
 8:30 
 
 9:00 
 
 9:30 
 TIME, h 
 
 10:00 10:30 
 
 FIGURE 6. - Energy-event ratio prior to a 
 roof fall {A) and an outburst (6). 
 
 MULTICHANNEL APPLICATIONS 
 
 Multichannel microseismic systems were 
 developed by the Bureau in the late 
 I960' 8, and research is continuing to 
 demonstrate their usefulness in solving 
 specific mine failure problems and in 
 improving total systems reliability. 
 Meanwhile, several applications of the 
 technology have been made at its present 
 level of development. 
 
 Early examples of systems application 
 can be seen in figures 7 and 8, which are 
 Bureau of Mines research results (12, 
 14) . The important difference between 
 single-channel applications and multi- 
 channel monitoring is that the location 
 of each individual rock noise can be cal- 
 culated and plotted on mine maps with the 
 latter system. Thus, a larger area of 
 the mine, even the entire mine if neces- 
 sary, can be monitored on a continuous 
 basis, and potential failure areas can be 
 
 accurately pinpointed within the mine 
 structure. This capability sometimes 
 allows for mine support to be modified to 
 avoid failure, for the application of 
 destressing techniques to control fail- 
 ure, or for the removal of personnel from 
 the area prior to failure. 
 
 Figure 7 shows the rock noise activity 
 plots for a 5-day period before a rock 
 burst in a stope pillar in a metal mine. 
 As can be seen, this procedure precisely 
 defined the location of the eventual 
 failure. Figure 7F^ shows the rock noise 
 rate which describes how rock noises in 
 the potential failure area occurred as a 
 function of time. As in single-channel 
 applications, note the dramatic increase 
 in rock noise rate prior to the rock 
 burst. 
 
 A similar example, this time comprised 
 of data from a Bureau study in a coal 
 mine, shows the rock noise data recorded 
 in conjunction with a coal mine bounce (a 
 failure very similar to a rock burst) 
 (fig. 8). The rock noise activity has 
 been contoured in terms of density, so 
 that the inner contours represent the 
 area in which the most rock noise is 
 occurring. Again, over a period of days, 
 notice the dense pattern of rock noises 
 that occurred in the eventual failure 
 area, precisely defining its location. 
 Also, note the similar reaction of the 
 rock noise rate from the failure area be- 
 fore its failure, as shown in figure 8F^. 
 
 Another recent example of results, from 
 the current Bureau study in a longwall 
 coal mining system, is shown in figure 9. 
 The cumulative rock noise location data 
 shown in figure 9A, were recorded during 
 the period April 7-14, 1983. A major 
 bounce occurred in this section on April 
 20, 1983, and caused damage in the tail- 
 gate and along the face exactly in the 
 area delineated by the rock noise data. 
 The rock noise rate also changed during 
 this time period, indicating unusual be- 
 havior of that face area. 
 
103 
 
 279 
 
 Ji 
 
 • • 
 
 284 279 
 
 » ^^ • • 
 
 • ■ 
 
 284 
 
 May 20 
 
 279 
 
 ' , y v 
 
 279 
 
 May 22 
 
 B 
 
 284 279 
 
 May 21 
 
 
 /// • 
 
 • 
 
 
 * 1 i 
 
 • • • 
 
 * • * 
 
 
 . • .» ^... . .. 
 
 
 • 
 • 
 
 • 
 
 284 
 
 May 23 
 
 
 284 
 
 o 
 
 lUU 
 
 80 
 
 - 
 
 1 
 
 1 
 
 I 
 
 1 
 Burst; 
 
 
 60 
 
 - 
 
 
 
 
 /- 
 
 Si 
 
 40 
 
 - 
 
 
 
 
 / - 
 
 0^4 
 
 
 
 
 
 
 
 Oc5 
 
 20 
 
 _ 
 
 
 
 
 / 
 
 z 
 
 n 
 
 
 
 9- 
 
 J — - 
 
 1 
 
 
 ^9 
 
 20 
 
 21 
 
 22 
 
 23 2 
 
 F 
 
 
 
 
 MAY 1979 
 
 
 F May 24 
 
 FIGURE 7. - Elevational views of rock noise concentrations and rate change prior to a rock burst. 
 
 This data set is incomplete, owing to 
 Irregularity in the monitoring procedure 
 during this time period, and the pre- 
 history and posthistory of this isolated 
 data set are unknown. The data cannot 
 therefore be said to have offered con- 
 clusive evidence regarding failure warn- 
 ing. The data are important however, in 
 that they (1) provide additional support 
 to the hypothesis that bounce areas can 
 be delineated well in advance of their 
 failure, and (2) indicate the viability 
 of these techniques In the mechanized 
 
 longwall mining environment, an important 
 contribution in light of the growing num- 
 ber of longwalls in the coal mining 
 industry. 
 
 The above examples were obtained using 
 small geophone arrays comprised of many 
 sensors to obtain precise locations of 
 rock noises and to provide precise fail- 
 ure data of specific areas of interest. 
 Similar applications are made in indus- 
 try, but the geophones are more widely 
 spaced to obtain widespread coverage, and 
 
104 
 
 April 3-30, 1973 
 
 15 20 
 APRIL 
 
 LEGEND 
 Event ^ Active mining 
 
 Number ^caved 
 3 ' of events 
 
 FIGURE 8. - Plan views of rock noise concentrations and rate change prior to a coal mine bounce. 
 
105 
 
 B 
 
 11 14 
 APRIL 
 
 Headgate 
 
 FIGURE 9„ = Plan view of rock noise concentration and rate change before a coal mine bounce in a 
 longwall section. 
 
 accuracy of rock noise locations is only 
 valid to about a 40-ft limit. This 
 proves sufficient for monitoring several 
 stoping areas and provides information 
 relavent to the whole of the stope pil- 
 lar. This practice, called minewide mon- 
 itoring, has been and is presently being 
 used with moderate success around the 
 world. This type of monitoring provides 
 the potential to evacuate a specific 
 working area before a burst, as can be 
 
 seen from data such as shown in figure 
 10, a case history from the Star Mine, 
 Burke, ID (15). It also provides an 
 insight into general mine stability and 
 an opportunity to watch the development 
 of problem areas and to take remedial 
 action to control or avoid failure. In 
 the Couer d'Alene mining district of 
 northern Idaho this type of analysis 
 has successfully been used to deter- 
 mine when destressing techniques should 
 
106 
 
 be Initiated to control rock -burst-prone 
 pillars (16). CO 
 
 Similar applications have been made in 
 other parts of the world for both re- 
 search and production. Research efforts 
 have increased dramatically around the 
 world within the last 5 yr, in both hard 
 rock and coal mining. Mining companies 
 have installed available systems to moni- 
 tor and study specific problems, and in 
 some instances are carrying out their own 
 research efforts to study new problems. 
 These new efforts, on many fronts, will 
 have a major impact on the future of 
 microseismic techniques in the mining 
 industry. 
 
 UJ 
 
 > 
 
 UJ 
 
 00 
 
 3 
 
 18 
 
 FIGURE 10. 
 a rock burst. 
 
 12 14 
 
 TIME, h 
 
 Rock noise rate change prior to 
 
 SOME PROBLEMS 
 
 The microseismic technique has devel- 
 oped into a potential tool that can pro- 
 vide for vastly increased safety to mine 
 personnel and can be used as an aid in 
 production planning and problems. To 
 demonstrate this, clear-cut and highly 
 successful applications have been shown; 
 however, problems and deficiencies of the 
 present technique do exist. 
 
 Reliability is the foremost of the 
 problems with microseismic techniques at 
 their present level of development. 
 Reliability does not mean equipment reli- 
 ability, although microseismic systems do 
 require constant maintenance and atten- 
 tion, but rather tlie fact than not all 
 failures are predicted as easily or 
 clearly as those presented, and sometimes 
 indicated failures may not occur. This 
 presents the user with the problem of 
 determining how many times he is willing 
 
 to be wrong. Experience has shown that 
 the technique has undeniable potential 
 and is beneficial in its present form 
 when properly used. That same experi- 
 ence, however, shows that at times, false 
 alarms will be sounded, and even worse, 
 sometimes no alarm will be sounded when 
 one was necessary. 
 
 The reliability problem cannot be 
 attributed solely to improper use of 
 microseismic systems or inadequate anal- 
 ysis of their data. The solution to the 
 problem lies in further research to de- 
 velop better and more accurate methods of 
 data analysis, and possibly the incorpor- 
 ation of supplemental methods to be used 
 in conjunction with the application of 
 microseismic techniques. While not per- 
 fect, the positive aspects of the method 
 generally outweigh the potential prob- 
 lems to the user. 
 
 CONCLUSIONS 
 
 The microseismic technique has been 
 developed a great deal in recent years 
 and has been used with moderate success 
 to provide warning of failure in several 
 different applications in the mining in- 
 dustry. Its use is growing and more 
 potential users of the technique are 
 becoming aware of its capabilities and 
 are making efforts to use the technology. 
 As mining continues to greater depths and 
 
 stability problems become more intense, 
 this use can only be expected to expand, 
 and with expanded use, we can expect 
 expanded technology and improved methods 
 of utilization. 
 
 Current Bureau research, in both hard 
 rock and coal mine environments, is aimed 
 at establishing better utilization of the 
 microseismic technique. Results continue 
 
107 
 
 to indicate the viability of the tech- warning of failure. The continued ef- 
 
 nique and that its use can contribute 
 substantially to both safety and mine 
 design. The ability to delineate problem 
 areas before failure is becoming better 
 established, but the overall reliability 
 of current techniques remains undefined, 
 particularly with regard to providing 
 
 forts of the Bureau are aimed at estab- 
 lishing the data base and experience 
 necessary to evaluate present reliability 
 problems and to improve the overall 
 effectiveness of microseismic techniques 
 applied to the problems of the mining 
 industry. 
 
 REFERENCES 
 
 1. Obert, L. Measurement of Pressures 
 on Rock Pillars in Underground Mines. 
 Pt. I. BuMines RI 3444, 1939, 15 pp. 
 
 2. . Use of Subaudible Noises 
 
 for Prediction of Rock Bursts. BuMines 
 RI 3555, 1941, 4 pp. 
 
 3. Obert, L. , and W. I. Duvall. The 
 Microseismic Method of Predicting Rock 
 Failure in Underground Mining. Part II. 
 Laboratory Experiments. BuMines RI 3803, 
 1945, 14 pp. 
 
 4. Blake, W. , and F. Leighton. Re- 
 cent Developments and Applications of the 
 Microseismic Method in Deep Mines. Ch. 
 23 in Rock Mechanics — Theory and Prac- 
 tice. AIME, 1970, pp. 29-443. 
 
 5. Leighton, F. , and W. Blake. Rock 
 Noise Source Location Techniques. Bu- 
 Mines RI 7432, 1970, 14 pp. 
 
 6. Leighton, F., and W. Duvall. A 
 Least Squares Method for Improving the 
 Source Location of Rock Noise. BuMines 
 RI 7626, 1972, 19 pp. 
 
 7. Redfern, F. R. , and R. D. Munson. 
 Acoustic Emission Source Location — A 
 Mathematical Analysis. BuMines RI 8692, 
 1982, 27 pp. 
 
 8. Blake, W. Microseismic Applica- 
 tions for Mining — A Practical Guide (con- 
 tract J0215002). BuMines OFR 52-83, 
 1982, 208 pp.; NTIS PB 83-180877. 
 
 9. . An Automatic Rock Burst 
 
 Monitor for Mine Use. Paper in Proc. 
 
 Conf, on the Underground Mining Environ- 
 ment, Univ. MO, Rolla, MO, 1971. Univ. 
 MO— Rolla, 1971, pp. 69-82. 
 
 10. Blake, W. , F. Leighton, and W. 
 Duvall. Microseismic Techniques for 
 Monitoring the Behavior of Rock Struc- 
 tures. BuMines B 665, 1974, 65 pp. 
 
 11. Coughlin, J. P. Software Tech- 
 niques in Microseismic Data Acquisition. 
 BuMines RI 8961, 1982, 51 pp. 
 
 12. Leighton, F., and B. Steblay. 
 Applications of Microseismics in Coal 
 Mines. Paper in Proc. 1st Conf. AE/MS 
 Activity in Geologic Structures and Mate- 
 rials (PA State Univ., June 1975). 
 Trans. Tech. Publ., 1977, pp. 205-229. 
 
 13. Steblay, B, Progress in the 
 Development of a Microseismic Roof Fall 
 Warning System. Paper in Proc. 10th An- 
 nual Institute on Coal Mining Health, 
 Safety & Research. VA. Polytech. Inst, 
 and State Univ., Blacksburg, VA, 1979, 
 pp. 177-195. 
 
 14. Leighton, F. A Case History of a 
 Major Rock Burst. BuMines RI 8701, 1982, 
 14 pp. 
 
 15. Langstaff, J. J. Seismic Detec- 
 tion System at the Lucky Friday Mine. 
 World Min., Oct. 1974, pp. 58-61. 
 
 16. Blake, W. Rock Burst Research at 
 the Galena Mine, Wallace, Idaho. BuMines 
 TPR 39, Aug. 1971, 22 pp. 
 
108 
 
 MECHANICAL AND ULTRASONIC CLOSURE RATE MEASUREMENTS 
 By Roger McVey^ 
 
 ABSTRACT 
 
 The Bureau of Mines has constructed two 
 intrinsically safe closure rate instru- 
 ments that provide the mine operator a 
 means for predicting an imminent roof 
 fall during pillar robbing. This im- 
 proves operator and machine safety and 
 prevents delays in digging out equipment. 
 One instrument system consists of two 
 rugged retrievable extensometers con- 
 nected by long electrical cables to a 
 digital readout unit for reading closure 
 and closure rate. Once a predesignated 
 closure rate is reached, the extensometer 
 is retrieved by pulling it from the im- 
 minent roof fall area by its electrical 
 cable. The equipment and mine personnel 
 are also pulled back to await the fall, 
 which usually occurs within minutes after 
 
 the designated rate is reached. Although 
 the unit is primarily designed for re- 
 treat mining operations , it can be used 
 for any activity requiring measurement of 
 displacement or rate of displacement. 
 Measurement range is to 6 in with 0.1 
 pet accuracy for openings of 4-1/2 to 
 12 ft. 
 
 The Bureau is also evaluating a small 
 ultrasonic unit to make these measure- 
 ments. The new instrument provides unob- 
 structing measurements up to 35 ft. The 
 ultrasonic transducer can be attached to 
 a roof bolt, tossed into an unsupported 
 area, or handheld. Total distance and 
 rate of change are displayed digitally to 
 0.001 ft. 
 
 INTRODUCTION 
 
 Roof control is a major problem in all 
 aspects of underground mining, especially 
 in room-and-pillar retreat operations. 
 Room-and-pillar retreat mining is begun 
 by developing a state of multiple en- 
 tries. The coal pillars between the 
 entries and crosscuts are then extracted 
 in a retreat sequence and the roof 
 allowed to cave in. The key is to mine 
 as much of the pillar as possible, then 
 remove both personnel and equipment be- 
 fore the final portion of the roof 
 collapses. 
 
 An extensive study had been made ear- 
 lier at the Southern Utah Fuel Company 
 (SUFCO) No. 1 Mine by Hamid Maleki, Colo- 
 rado School of Mines, and Doug Johnson, 
 
 ^Supervisory electronic technician, 
 Spokane Research Center, Bureau of Mines, 
 Spokane , WA . 
 
 of SUFCO, in determining the critical 
 roof-to-floor closure rate for predicting 
 a roof -caving during retreat mining. 
 They timed the roof-to-floor clo- 
 sure change to determine the critical 
 rate of closure. The critical rate was 
 determined to be 0.2 in/min. Roof -caving 
 prediction based on this value was so 
 successful that the mine suggested the 
 Bureau develop an automatic closure-rate 
 instrument. 
 
 Subsequently, the Bureau designed and 
 built two types, mechanical and ultra- 
 sonic, of automatic closure-rate instru- 
 ments that would digitally record both 
 the rate of closure and the accumulative 
 closure, and also provide audiovisual 
 warnings when the closure rate reached a 
 preset critical value. 
 
109 
 
 ACKNOWLEDGMENTS 
 
 The author wishes to thank Bob 
 Ochsner and Tom Heaps of SUFCO for 
 
 their help in testing the closure-rate 
 instruments. 
 
 MECHANICAL INSTRUMENT 
 
 DESCRIPTION 
 
 A mechanical closure-rate instrument 
 was built first. It consists of two 
 rugged telescoping potentiometric exten- 
 someters and a digital readout-control 
 box (fig. 1). The extensometer (fig. 2) 
 is designed to accommodate a height of 
 from 4-1/2 to 12 ft with a measurement 
 It is spring-loaded over 
 Long (100- to 125-ft) 
 cables connect the extensometers to the 
 readout box, permitting the operator to 
 remain in a safe, supported area. A 
 breakaway feature on each extensometer 
 allows it to be pulled from the fall area 
 by its electrical cable. 
 
 range of 6 in. 
 this 6-in range. 
 
 The operator watches the digital read- 
 out on the control box (fig. 3) during 
 mining of the pillar. When a predeter- 
 mined critical rate of closure is indi- 
 cated, he retrieves the extensometer 
 (fig. 4) and signals the miner operator 
 to pull back. The control box (fig. 5) 
 provides two digital visual readouts. 
 One display shows the closure rate, the 
 other total accumulative closure from 
 time zero. The operator can preset any 
 closure rate from to 1 in/min. When 
 the preset rate is reached, an alarm 
 light illuminates and an audible alarm is 
 sounded. The control box can monitor two 
 extensometers individually or alter- 
 nately. The system is battery-operated. 
 
 FIGURE 1. - Jvlechanical closure-rate system. 
 
 FIGURE 2. - Extensometer. 
 
110 
 
 FIGURE 3. - Control box, on-site. 
 
 conqjletely portable, and can be quickly 
 moved from one place to another. 
 
 The closure-rate instrument, though 
 primarily designed for retreat mining, 
 can be used for any activity where knowl- 
 edge of roof-to-floor closure rate or 
 total displacement is required. The sys- 
 tem provides a 0- to 6-in measurement 
 range with 0.1 pet accuracy for both rate 
 
 and total closure. Any extensometer in- 
 stallation displacement can be zeroed out 
 with zeroing potentiometers. This zero 
 value can be recorded and reset if 
 multiple extensometers (greater than two) 
 are used. The auto alarm can be set to 
 any closure-rate value from to 1 in/ 
 min, with a 0.01-in resolution and 0.1 
 pet accuracy. 
 
HI 
 
 FIGURE 4. - Installed extensometer. 
 
 
 CIRCUIT DESCRIPTION 
 
 Figure 6 is a timing diagram for the 
 operation of the electronic measurement 
 circuits. Because of the small changes 
 in actual rate, normally up to about 0.2 
 in/min, it was decided to use a 10-s sam- 
 ple period for greater accuracy. The ex- 
 tensometer is sampled at the beginning 
 (S2) and at the end (S,) of a 9-s period 
 by sample and hold circuits. The two 10- 
 ms samples are compared, and any differ- 
 ence is converted to a directly propor- 
 tional frequency, counted for 100 ms , and 
 displayed as rate in inches per minute. 
 A total read cycle consists of taking the 
 two samples, resetting the counter, read- 
 ing the value, and displaying it dig- 
 itally. The extensometers are auto- 
 matically measured alternately or can be 
 continuously monitored on a singular bas- 
 is. The extensometer circuit (if in 
 auto) will automatically cycle to extens- 
 ometer No. 2 with the extensometer toggle 
 switch in auto position. When No. 2 has 
 been sampled and displayed. No. 1 extens- 
 ometer is automatically toggled back into 
 the circuit. 
 
 FIGURE 5. - Control box, closeup. 
 
 Figure 7 is an electrical diagram 
 of the readout-control box. Each 
 
112 
 
 1 Hz dispiay 
 
 K 
 
 Time zero 
 
 1 
 
 J 
 
 1 Hz clock .JoUUJIUIUTLfBUeUTLJsLJ^^ 
 
 Sj sample 
 S, sample 
 Reset counter 
 Read 
 
 il 
 
 10 ms 
 
 il 
 
 10 ms 
 
 il 
 
 10 ms 
 
 Jl 
 
 10 ms 
 
 Toggle ext H ^^ "^^ 
 
 F/F 
 
 U 100 ms 
 [-[10 ms 
 
 Note: Timing as follows 
 
 1. Extensometer 1 is selected by "Toggle ext pulse." 
 
 2. Extensometer voltage is sampled "Sz" and is held to be compared 
 with Si. 
 
 3. Si is sampled and compared with Sz- 
 
 4. Counter is reset ready to count. 
 
 5. Counter reads V to F count. 
 
 6. Count displayed for 5 s. 
 
 7. Extensometer 2 is selected by "Toggle ext pulse." 
 
 8. Same sequence of reading for Extensometer 2. 
 
 9. Extensometer 1 is selected by "Toggle ext pulse." 
 
 FIGURE 6. - Timing diagram. 
 
 extensometer electrical output is fed Accuracy: 
 directly to a high-input impedance ampli- 
 fier. The outputs from these amplifiers Resolution: 
 are routed to a field-effect transistor 
 switch, which selects the extensometer to Alarm: 
 be measured. One route converts the ex- 
 tensometer voltage to frequency and is 
 displayed as inches of total accumulative Readout: 
 displacement. The same voltage is routed 
 to the sample-and-hold circuitry, where 
 it is sampled, summed, and converted to 
 rate in inches per minute. This differ- 
 ence, or summed voltage, is also sent to Power: 
 the comparator for the alarm function. 
 
 INSTRUMENT SPECIFICATIONS 
 
 Measured range: to 6 in displacement 
 
 Closure rate: to 6 in/min 
 
 +0.1 pet 
 
 0.001 in 
 
 to 1 in/min with 
 0.01-in resolution 
 
 Digital LED display: 
 two; one for total 
 displacement; one 
 for closure rate. 
 
 Battery power 12-h 
 capacity. Intrinsi- 
 cally safe (MSHA 
 approved) . 
 
113 
 
 Ext. No. 1 
 
 »- 6.9 V regulated 
 
 V/F, 
 
 SW1 
 
 Ext. No. 2 
 
 >- 6.9 V regulated 
 
 FET 
 SW 
 
 Toggle 
 
 Voltage 
 to f req 
 
 Counter 
 
 ' Display \ 
 
 jf Warning 
 < — 1^ ''9'^* ^ !► Alarm 
 
 Total closure 
 
 ^ 
 
 Audible 
 alarm 
 
 J-*fset 
 
 Cpl 
 
 INVn 
 
 osc 
 
 Auto- 
 Dlvider circuits 
 
 Sample 
 
 and 
 hold S, 
 
 VF, 
 
 Voltage 
 to freq 
 
 100 
 Hz 
 
 10 
 
 10 
 
 Hz 
 
 -MO 
 
 Hz 
 
 Closure rate 
 
 Timing circuit 
 
 100 ms 
 — ► Read 
 
 ►lO ms 
 
 Sample 
 
 FIGURE 7. - Block diagram of closure-rote instrument. 
 ULTRASONIC INSTRUMENT 
 
 A nonobstructing method of measuring 
 roof-to-floor closure has always been 
 desirable in underground ground-control 
 measurements. The mechanical extensom- 
 eter has several drawbacks, such as high 
 cost and being hard to use in high- 
 traffic areas. Thus the Bureau undertook 
 a project to determine the feasibility of 
 using ultrasonics for convergence mea- 
 surement underground. The new Polaroid 
 sonar camera transducer and ranging sys- 
 tem became a prime candidate. A small 
 ultrasonic transducer of this type could 
 provide an inexpensive, nonobstructing 
 means of measurements. 
 
 The Bureau has thus far developed a 
 small, yet inexpensive, handheld ultra- 
 sonic unit, shown in figure 8, that can 
 measure distances from 1 to 35 ft ±0.02 
 ft. The unit is excellent for general 
 survey work, measuring high roofs, etc. 
 
 A small, portable closure-rate measuring 
 device was also built with an ultrasonic 
 remote transducer that can be placed up 
 to 100 ft or more from the readout in- 
 stmment. This unit is shown in figure 
 9. This instrument reads both distance 
 as well as rate of closure. It has a 
 measurement range of 1 to 35 ft ±0.01 ft 
 with 0.001-ft resolution. The first 
 reading displayed is the distance. The 
 rate is displayed 6 s later. A thumb- 
 wheel switch allows for setting an alarm 
 limit for rate of closure. A visual and 
 audible alarm is provided. The alarm 
 limit can be set from to 9.9 ft/min. 
 
 Because the ultrasonic instrument is 
 still in the design and testing phase, 
 and since permissibility approval has not 
 been received from MSHA, its electrical 
 schematics and operational information 
 will be made available at a later time. 
 
114 
 
 FIGURE 8. • Portable handheld ultrasonic unit. 
 
115 
 
 FIGURE 9. - Portable ultrasonic closure-rate instrumento 
 
 FIELD TESTS 
 
 In June 1981, the mechanical closure- 
 rate instrument was placed for field 
 testing in SUFCO's No. 1 Mine (figs. 3-4) 
 near Salina, UT. The test results have 
 exceeded expectations, and roof-caving 
 predictions have proven very accurate. 
 In most cases, the instrument has warned 
 of an impending roof -caving within min- 
 utes of the event. It was reported that 
 use of the closure-rate instrument has, 
 in general, led to increased coal recov- 
 ery and productivity. 
 
 The readout instrument, except for a 
 fuse blown during battery replacement, 
 has been trouble-free. A design change 
 in the breakaway mechanism and substitu- 
 tion of heavier gauge extensometer rods 
 are the only modifications made to the 
 system thus far. It is noteworthy that, 
 to date, no extensometer s have been lost 
 during caving. 
 
116 
 
 CONCLUSIONS 
 
 With 22 months of field tests we con- 
 clude that the mechanical closure-rate 
 Instrument appears to be a viable tool 
 for roof hazard prediction In retreat 
 mining operations once the critical clo- 
 sure rate has been determined for a par- 
 ticular mining location. Most retreat 
 
 cycles. In general, have Increased ton- 
 nage per production shift. 
 
 Preliminary test results of the ultra- 
 sonic closure-rate unit appear excellent. 
 It Is hoped that this unit will also Im- 
 prove safety underground. 
 
117 
 
 GROUND INSTALLATION EQUIPMENT 
 
 REMOTE MANUAL ROOF BOLTERS 
 By John E. Bevani 
 
 ABSTRACT 
 
 Coal mine accident statistics show that 
 18 pet of all roof-fall fatalities in- 
 volve roof bolting. This is 30 pet 
 higher than for any other occupation 
 category. Industry analyses show that 
 roof bolters were involved in 15 pet of 
 
 lost-time accidents. Rapid placement of 
 permanent roof support appears essential 
 to safety as well as long-term roof sta- 
 bility. This paper investigates one 
 method for placement of permanent roof 
 support. 
 
 BACKGROUND 
 
 The MESA report, "Analysis of Fatal 
 Roof-Fall Accidents in Coal Mines, 1972- 
 1975," indicates that 18 pet of all roof- 
 fall fatalities involved roof-bolter 
 operators and helpers. This group of 
 workers experienced 30 pet more fatal- 
 ities than any other occupation category. 
 
 The report "Injuries Associated With 
 Roof or Rib Bolting and Bolting Machines 
 in Underground Coal Mines, 1978-1982" 
 analyzed 5,777 underground coalmine bolt- 
 ing or related bolting machine injuries 
 reported to HSAC from 1979 through 1982. 
 The results were — 
 
 1. Drilling of the roof or rib 
 accounted for 2,457 injuries (45 pet). 
 
 2. Installation of bolts in roof or 
 rib~l,634 (28.3 pet). 
 
 3. Tramming the machine — 727 (12.5 
 pet). 
 
 4. Unknown (due to insufficient data 
 to classify) — 959 (16.6 pet). 
 
 Industry analyses of lost-time acci- 
 dents are not complete, but the following 
 figures are typical. One company's 
 Safety Department reports that roof-bolt 
 operators were involved in 15 pet of 
 their lost-time accidents. Of the roof- 
 bolting accidents, 19 pet were caused by 
 roof falls , and 48 pet were caused by 
 being struck or caught by rotating tools. 
 
 The above information indicates the 
 danger involved with roof bolting and 
 roof control. Many other non-roof-bolter 
 injuries and fatalities are also a result 
 of insufficient roof support. Rapid 
 placement of permanent roof support also 
 appears to be beneficial to mid- and 
 long-term roof stability as well as imme- 
 diate safety. 
 
 RATIONALE 
 
 The Bureau of Mines has addressed 
 these problems through two areas of re- 
 search: (1) by removing the bolter oper- 
 ator from the immediate dangers of bolt 
 installation; and (2) by providing means 
 for a safer and more timely bolting 
 system. 
 
 'Mechanical engineer, Spokane Research 
 Center, Bureau of Mines, Spokane, WA. 
 
 A roof-bolt inserter (RBI) developed 
 allowed a longer-than-seam-height bolt to 
 be installed by bending the horizontally 
 carried bolt into the vertical orienta- 
 tion of the mine roof-bolt hole. A 
 longer-than-seam height (LTSH) drill or 
 flexible roof drill developed by con- 
 tract allowed long holes to be drilled 
 in low coal. The RBI and LTSH drill add 
 flexibility of package design, which 
 
118 
 
 allows the operator to be removed from 
 the immediate bolting area, under sup- 
 ported roof and away from the dangers of 
 rotating and moving equipment. The oper- 
 ator can be placed outby the last perman- 
 ent row of roof support and/or under can- 
 opy protection. Here the operator can be 
 protected while still using his or her 
 abilities. 
 
 An automated miner-bolter to use the 
 concept of mounting the two systems on a 
 continuous miner would allow one-pass or 
 truly continuous mining. 
 
 However, many problems of an automated 
 system may be caused by the substitu- 
 tion of the operator's function with 
 mechanical-electrical-hydraulic systems 
 and artificial intelligence. Manual dex- 
 terity, memory, logic, audio, visual, and 
 sensory capabilities of the operator have 
 been replaced in an attempt to remove him 
 or her from the dangers of the immediate 
 bolting station. It appears that any 
 viable system must be greatly simplified; 
 hence, replacing some of the capabilities 
 of the operator is unjustified and unwar- 
 ranted if the operator can supply these 
 functions while being protected. 
 
 Six concepts featuring hands-off drill- 
 ing, remote control and/or automated 
 
 sequencing, and improved productivity 
 features were developed. From these six 
 concepts and initial designs, the artic- 
 ulated remote manual roof -bolter (ARM 
 bolter) and the remote-manual bolter 
 (REM bolter) were chosen for further 
 development. 
 
 The ARM bolter (fig. 1) is a roof 
 bolter for low seams , which enables an 
 operator to perform bolting functions 
 while under a canopy protection and per- 
 manent roof support. The operator sits 
 in a reclined position in a cab with his 
 head approximately 8 ft from the bolt- 
 hole location. From this location, the 
 drilling, bolting, and torquing of bolts 
 are controlled. Bolts ranging in length 
 from 4 to 8 ft are fed into the bolter 
 component assembly. The machine's over- 
 all tram height is approximately 33 in 
 and it is capable of installing bolts in 
 seams ranging from 37 to 60 in. The lim- 
 itation in seam height is due to bolter 
 design and not necessarily limited by 
 conceptual considerations . 
 
 The heart of the ARM bolter is the 
 bolter component assembly. This assembly 
 houses a flexible roof drill, roof -bolt 
 inserter (RBI), torque thrust assembly, 
 plate magazine, feed and receive mechan- 
 isms, and a component carriage to house 
 
 I 
 
 FIGURE lo - Articulated remote manual (ARM) roof bolter. 
 
119 
 
 the above elements. The bolter component 
 assembly interfaces with the front end of 
 the articulated vehicle and is raised and 
 lowered by means of two elevation assem- 
 blies. The operator manually assembles 
 the mechanical bolt and anchor, then 
 loads the assembled bolt (less anchor 
 plate) into the feed and receive mechan- 
 ism prior to drilling the hole. The 
 anchor plates are loaded into the plate 
 magazine at the same time. The operator 
 enters the operator station (under canopy 
 support) and trams the bolter to the 
 proper bolting position as dictated by 
 the mine plan. The bolter component as- 
 sembly is raised to the roof, and the 
 bolt hole is drilled by the longer- 
 than-seam-height drill. After the hole 
 is drilled to the desired length, the 
 drill string is retracted and the drill 
 is indexed away allowing the RBI-torquer 
 assembly to be indexed to align with the 
 drilled hole. The plate feed assembly 
 installs the plate on the roof bolt, then 
 the RBI installs the bolt into the hole. 
 
 The RBI expands allowing the torquer to 
 engage, insert, and torque the bolt as- 
 sembly. The RBI and drill are indexed to 
 their stow position, the bolter component 
 assembly and roof jacks are lowered, and 
 the ARM bolter is ready to tram to the 
 next bolt installation. The control is a 
 combination of air logic control and re- 
 mote operator control. Remote operator 
 control is available in the event of an 
 air logic system failure. These opera- 
 tions, done during bolt installation, are 
 achieved while the operator is under can- 
 opy and permanently supported roof. 
 
 The second concept being developed 
 (fig. 2) by the Bureau is the remote man- 
 ual bolter (REM bolter). This concept 
 uses items from the longer-than-seam- 
 height drill program, but previous prob- 
 lems with automated modules directed 
 attention toward enchancing, rather than 
 replacing the operator. Many of the same 
 design criteria used for the ARM bolter 
 were used in the development of the REM 
 
 FIGURE 2. . Remote manual (REM) roof bolter. 
 
120 
 
 bolter. The operator is used more in the 
 REM bolter concept than the ARM bolter 
 (attempting to make a simpler system). 
 The same steps are required to install 
 the bolt, but here the operator is re- 
 sponsible for assembling the complete 
 bolt. Any anomaly must be compensated 
 for by the operator. 
 
 The REM bolter is trammed into posi- 
 tion, and the drilling is initiated. The 
 RBI is mounted on a track which allows it 
 to slide back near the operator where the 
 assembled bolt is manually placed into 
 the RBI. When the drilling cycle is com- 
 pleted, the drill is indexed to its 
 stowed position, and the RBI is run for- 
 ward on the track and indexed under the 
 drilled hole. Final positioning of the 
 RBI, if required, can be done by moving 
 the bolter or moving the arm that carries 
 the RBI. The RBI then inserts the bolt 
 into the hole and is indexed to its stow 
 position. The torquer is indexed into 
 position and coupled with the bolt. The 
 
 bolt is torqued, the torquer is stowed, 
 the floor jack is released, and the 
 bolter is ready to be trammed to the next 
 bolting position. The operator has not 
 moved from his protected position. 
 
 All control functions of the REM bolter 
 are operator-controlled. No sensory 
 feedback or logic circuit, lockout, etc., 
 is employed. This is done for simplicity 
 and its associated reduction in cost, 
 weight, and size. 
 
 The future for remote bolters should 
 hold great promise. The remote-manual 
 concept can easily be adapted to resin 
 bolts or a combination of mechanical 
 and resin. Water-jet-assisted flexible 
 drills could increase the number of roof 
 conditions where the bolter can be used. 
 (Water-jet-assisted flexible drills may 
 replace rotary-impact drilling now used 
 on severe roof conditions). New concepts 
 like the inorganic grout injection de- 
 vices could be easily adapted. 
 
 RESULTS 
 
 A 4-month underground test of the REM 
 bolter began February 25, 1983, at a mine 
 near Daisytown, PA. A soft band of shale 
 in the mine roof resulted in flex drill 
 problems. The dust system appeared to be 
 plugging, which resulted in impacting the 
 drill string in the hole. Another prob- 
 lem area was the operator's difficulty in 
 inserting the bolt and coupling the 
 torquer to the bolt head after it was 
 inserted. 
 
 Another working section was made avail- 
 able by mine personnel and initial drill 
 
 problems did not reoccur. Bolt installa- 
 tion of 50 holes per shift has been 
 achieved. 
 
 The ARM bolter has been tested for 
 a total of 5 weeks in a West Virginia 
 mine. A total of 163 bolts have been 
 installed. No major problems were en- 
 countered; however, some correctable 
 problems have been encountered with the 
 drill, cycle time, and dust collection 
 system. 
 
 CONCLUSIONS 
 
 Both the ARM bolter and the REM bolter 
 address the same problems, but the path 
 taken by each is different. Some of the 
 questions to be answered are: 
 
 3. Can the operator find the hole to 
 insert the bolt, and can it be done in 
 varying seam heights with undulating top 
 and /or bottom? 
 
 1. What degree of automation or semi- 
 automation is required? 
 
 2. Can the operator develop the 
 needed skills to couple and uncouple the 
 torquer to the bolt or can a control sys- 
 tem do it better? 
 
 4. Will the operator be fatigued or 
 will active participation result in a 
 safer, more alert operator? 
 
 The program goal for the REM and ARM 
 bolters was to develop a concept that 
 
121 
 
 will protect the operator , be econom- 
 ically feasible, and be implemented by 
 industry into a production machine. The 
 program development and subsequent test- 
 ing has shown the concept as viable. 
 Automated bolters were costly and com- 
 plex. The remote bolters are less expen- 
 sive and much simpler while maintaining 
 operator safety and high production 
 levels. 
 
 Future generation machines will be more 
 efficient and probably less costly. The 
 degree of automation that can be effec- 
 tively used in the commercial underground 
 mining environment will increase as the 
 state-of-the-art in robotics, controls, 
 software, etc., increases, but any suc- 
 cessful machine must be designed around 
 the human operator, still the most impor- 
 tant element of any system. 
 
122 
 
 FIELD TEST OF AN AUTOMATED TEMPORARY ROOF SUPPORT (ATRS) USED ON A LOW-COAL, 
 SINGLE, FIXED-HEAD ROOF BOLTING MACHINE (SQUIRMER) 
 
 By Charles T. Chislaghi^ and Thomas E. Marshall^ 
 
 ABSTRACT 
 
 An economical, remotely operated (auto- 
 mated), temporary roof support (ATRS) has 
 been developed by the Bureau of Mines for 
 use on a single, fixed-head roof bolting 
 machine (squirmer) that operates in low- 
 coal seams (<42 in thick). This ATRS 
 eliminates the need for workers to go 
 under unsupported roof to set or remove 
 temporary support prior to or during the 
 roof bolting cycle — a task that annually 
 accounts for approximately 20 pet of all 
 roof fall fatalities. It can be adapted 
 
 on any squirmer used in the U.S. low-coal 
 fields. A prototype ATRS, designed and 
 built at the Bureau's Pittsburgh Research 
 Center, was field-tested at Imperial Col- 
 liery Co.'s Mine No. 20 in Eskdale , WV. 
 The Mine No. 20 amended roof control 
 plan, which requires the use of the Bu- 
 reau's ATRS as temporary support during 
 face bolting, has been approved by the 
 Mine Safety and Health Administration 
 (MSHA) . 
 
 INTRODUCTION 
 
 A statutory provision of the Federal 
 Coal Mine Health and Safety Act of 1969 
 states that "No person shall proceed be- 
 yond the last permanent support unless 
 adequate temporary support is provided."-^ 
 However , since the time the law was 
 written, there have been no means avail- 
 able for squirmer operators and helpers 
 to set temporary supports from under per- 
 manently supported roof. This provision 
 was interpreted to mean "In areas where 
 permanent artifical support is required, 
 temporary support should be used until 
 such permanent support is installed,"'* 
 and "Only those persons engaged in in- 
 stalling temporary support should be 
 allowed to proceed beyond the last per- 
 manent support until such temporary sup- 
 ports are installed."^ Annually, ap- 
 proximately 20 pet of all roof-fall 
 fatalities involve miners who have gone 
 beyond the last permanent support to set 
 or remove temporary roof support prior to 
 or during the roof bolting cycle. 
 
 'Mining engineer. 
 ^Engineering technician. 
 Pittsburgh Research Center, 
 Mines, Pittsburgh, PA. 
 ^30 CFR 75.200. 
 "^SO CFR 75.200-13(a){1 ) . 
 ^30 CFR 75.200-13(a)(2) 
 
 Bureau of 
 
 Because of space limitations in low 
 coal, not many ATRS's have been commer- 
 cially developed for squirmers, although 
 many different ATRS systems have been 
 commercially developed for roof bolting 
 machines used in high coal. Most ATRS's 
 designed for squirmers create a situation 
 that reduces or compromises the existing 
 safety level, with a greater safety haz- 
 ard to squirmer operators and helpers 
 working inby the last row of permanent 
 support. 
 
 Over 3,500 squirmers are in use today 
 in southern West Virginia, eastern Ken- 
 tucky, and southwestern Virginia, and 
 approximately 60 pet of these have no 
 ATRS, cab, or canopy. Moreover, low-coal 
 mine operators and owners in West Vir- 
 ginia have a need for ATRS because West 
 Virginia mine law requires that roof 
 bolting machines used in working places 
 of West Virginia coal mines be equipped 
 with ATRS, regardless of coal seam 
 height. 
 
 All design work and prototype fabrica- 
 tion was done by the Roof Support Group 
 at the Pittsburgh Research Center. All 
 fieldwork was done in the No. 2 Gas seam 
 (36 to 42 in thick) at Mine No. 20 of the 
 Imperial Colliery Co. in Eskdale, WV. 
 
123 
 
 ACKNOWLEDGMENTS 
 
 The Bureau acknowledges the coopera- 
 tion it received from personnel of the 
 Imperial Colliery Co., MSHA Mt. Hope 
 Subdistrict, and MSHA Bruceton Safety 
 
 Technology Center. Without their techni- 
 cal suggestions and assistance, this 
 project could not have been completed. 
 
 DESCRIPTION OF ATRS 
 
 The Bureau of Mines ATRS is based on a 
 modified and improved Lee Engineering 
 design for a squirmer ATRS. It consists 
 of a 10-ft-long, steel, wide-flange beam 
 supported by two double-acting, telescop- 
 ing hydraulic cylinders (fig. 1). A 
 steel sleeve, mounted on the bottom cen- 
 ter of the beam, is designed to fit over 
 the top of the squirmer drill head. The 
 ATRS is carried from place to place and 
 row to row on the squirmer drill head, 
 but is not an integral part of the 
 squirmer during bolting. During bolting 
 it is connected to the squirmer only by 
 two hydraulic lines. Because the ATRS 
 only weighs about 400 lb, the squirmer 
 
 drill head and boom do not have to be re- 
 built to carry it. Total cost of the 
 beam and cylinders is approximately 
 $1,800. In-house fabrication of the ATRS 
 took 8 worker-hours. 
 
 The Bureau's ATRS design meets MSHA's 
 general design requirements and West Vir- 
 ginia's design and operating requirements 
 for such support. Both hydraulic cylin- 
 ders supporting the ATRS have check 
 valves to prevent sudden collapse of the 
 ATRS in the event of a ruptured hydraulic 
 line or broken hydraulic fitting. In 
 addition, the ATRS hydraulic circuit con- 
 tains an accumulator , charged by squirmer 
 
 FIGURE 1. - Bureau of Mines ATRS. 
 
124 
 
 line hydraulic pressure, which keeps the 
 ATRS firmly set against the mine roof 
 even if the roof rock is pulled up dur- 
 ing the bolting cycle. The ATRS can 
 
 elastically support the minimum required 
 deadweight load of 33,750 lb; this 
 capacity is certified by a professional 
 engineer. 
 
 SQUIRMER STREAMLINING 
 
 Imperial Colliery personnel streamlined 
 a 15-yr-old FMC model 300 squirmer for 
 the field test. West Virginia State Mine 
 Law requires the streamlining of any roof 
 bolting machine before it can be retro- 
 fitted with ATRS. The ATRS controls were 
 located 5 ft back from the drill head so 
 that they can only be operated from be- 
 neath permanently supported roof. Inch 
 tram controls were located at the drill 
 station, and inch tram speed was reduced 
 to 65 ft/min. Full tram controls were 
 located with the ATRS controls, and the 
 
 full tram speed was left at 150 ft/min. 
 No ATRS controls were located at the 
 drill station. Other streamlining work 
 included removal of the bolt tray and 
 tram deck, installation of low-value, 
 high-torque tram motors, and moving the 
 squirmer front wheels 8 in forward to 
 provide space for the ATRS and full tram 
 controls. Total cost of this work, which 
 required 96 worker-hours, was $5,500. 
 The Bureau piped the drill circuit to the 
 ATRS and piped the ATRS circuit on beam 
 at a cost of $150 and 32 worker-hours. 
 
 FIELD TEST AND RESULTS 
 
 With MSHA and West Virginia approval. 
 Imperial Colliery placed the ATRS in the 
 production cycle at Mine No. 20 for 5 
 months. Bolting was on 4-ft center, with 
 20-ft-wide entries and crosscuts. The 
 cycle was the following: 
 
 Step Description 
 
 1 The squirmer operator, at the full 
 tram controls, trams into the cen- 
 ter of a place and stops when the 
 ATRS is under the last row of per- 
 manent support. 
 
 2 The operator lowers the drill head 
 (and ATRS) to the mine floor using 
 a boom control located beside 
 the full tram and ATRS controls; 
 moves from the full tram position 
 to the beam (ATRS); and unhooks 
 the hydraulic cylinder leg on the 
 operator side that is chained to 
 the beam, while the helper does 
 the same to the leg on the right 
 side. 
 
 3 The operator raises the drill head 
 (and ATRS) , using a boom control at 
 the drill station, just high enough 
 to let the legs hang down without 
 scraping the mine floor; locks the 
 legs perpendicular to the mine 
 floor; moves back to the full tram 
 and ATRS controls; and trams the 
 
 squirmer inby without the legs 
 scraping mine floor or the beam 
 scraping mine roof. 
 
 The operator stops when under the 
 last row of permanent support. The 
 ATRS is now 5 ft inby the last row 
 of permanent support and 5 ft from 
 each rib. 
 
 The operator places the ATRS 
 against the roof using the boom 
 control located beside the full 
 tram and ATRS controls, and then 
 lowers the legs to the mine floor 
 using the ATRS control until the 
 beam is firmly set against the roof 
 and the legs are firmly set against 
 the mine floor. 
 
 The operator lowers the drill head 
 away from the beam using the boom 
 control located beside the full 
 tram and ATRS controls. 
 
 At this point, the operator moves 
 to the drill station, pushes in the 
 diversion valve which diverts all 
 hydraulic fluid from the full tram 
 circuit to the inch tram circuit, 
 and "inches" the squirmer to the 
 left rib to begin bolting. During 
 bolting the squirmer is connected 
 to the ATRS by only two hydraulic 
 lines. 
 
125 
 
 8 After a row of permanent support is 
 installed, the operator raises the 
 drill head into the beam using the 
 boom control at the drill station; 
 pulls out the diversion valve which 
 diverts all the hydraulic fluid 
 back to the full tram circuit from 
 the inch tram circuit; moves to the 
 full tram and ATRS controls; and 
 raises the legs using the ATRS 
 control. 
 
 9 The operator lowers the drill head 
 (and ATRS), using the boom control 
 located beside the full tram and 
 ATRS controls , just enough to tram 
 the squirmer inby without the legs 
 scraping mine floor or the beam 
 scraping mine roof. 
 
 10 When under the row of permanent 
 roof support that has just been in- 
 stalled, the operator stops and re- 
 peats steps 5 through 9. This 
 cycle is repeated until the last 
 bolt is in place. 
 
 11 Then the operator raises the drill 
 head into the beam using the boom 
 control at the drill station; un- 
 locks the legs ; pulls out the di- 
 version valve; moves to the full 
 
 tram and ATRS controls; raises the 
 legs using the ATRS control; moves 
 back to the drill station; lowers 
 the drill head (and ATRS) to the 
 mine floor using the boom control 
 at the drill station; chains the 
 leg, on the operator side, to the 
 beam while the helper does the same 
 to the leg on the right side; moves 
 back to the full tram and ATRS con- 
 trols; turns the squirmer 180°; and 
 trams to the next place where steps 
 1 to 11 are repeated. 
 
 No operating or maintenance problems 
 were encountered during the 5 months of 
 testing. With the addition of the ATRS, 
 the squirmer could still turn 180° within 
 the 20-ft-wide entries and crosscuts and 
 could tram through check curtains and 
 line brattice without pulling them down. 
 Comparative time studies of the same 
 bolting crews showed that it took an av- 
 erage of 5 min less to bolt a place with 
 the ATRS than with mechanical jacks. The 
 bolting crews preferred the ATRS. After 
 testing, an amended roof control plan re- 
 quiring the use of the Bureau's ATRS dur- 
 ing face bolting at Mine No. 20 was sub- 
 mitted by the Imperial Colliery Co. and 
 approved by MSHA — District 4, Mount Hope 
 Subdistrict, Montgomery Field Office. 
 
 CONCLUSIONS 
 
 The Bureau's ATRS eliminates the need 
 for squirmer operators and helpers to go 
 under unsupported roof to set or remove 
 temporary support prior to or during the 
 roof bolting cycle. The squirmer oper- 
 ator will always be under permanently 
 supported roof while setting or removing 
 the ATRS and will not be able to bolt 
 inby the ATRS because of its control 
 location. The Inexpensive and light- 
 weight ATRS does not reduce the squirmer 
 operator's work space. It has the poten- 
 tial to be immediately used in some 70 
 pet of U.S. low-coal mines. Although the 
 Bureau's ATRS was field tested only with 
 
 the FMC model 300 squirmer , it can be 
 adapted to the drill head of any squirmer 
 operating in low coal and it can be 
 fabricated in any mine shop. If a 
 streamlined squirmer is available, the 
 ATRS can be retrofitted to the squirmer 
 during maintenance shifts , and if prob- 
 lems occur with the ATRS during the bolt- 
 ing cycle, it can be disconnected from 
 the squirmer to allow bolting to continue 
 with mechanical jacks. It has the poten- 
 tial to eliminate roof fall fatalities 
 and injuries and may lead to increased 
 productivity. 
 
126 
 
 ROOF SUPPORT SYSTEMS 
 
 DEVELOPMENT OF EPOXY GROUTS AND PUMPABLE BOLTS 
 By Robert R. Thompson'' 
 
 ABSTRACT 
 
 Good roof control is critical to the 
 coal mining industry. The Bureau of 
 Mines recognized the need for a remotely 
 placed roof bolt of noncorrosive materi- 
 als, which could be remotely installed 
 in longer -than-seam-height lengths. The 
 system developed has a fiberglass core. 
 The core is made up of four 1/4-round 
 sections, which can be coiled on the 
 machine for storage and then formed into 
 a hollow core during insertion into the 
 drilled hole. The adhesive used is a 
 
 fast-setting epoxy resin. The resin was 
 tested underground in cartridge form. 
 The test results showed bonding strengths 
 greater than those available with the 
 polyester resin now being used. Core 
 handling and pumping equipment was de- 
 signed, built, and placed on an existing 
 roof bolter. The equipment is now being 
 laboratory-tested and will be tested 
 underground in a coal mine using a stan- 
 dard MSHA-def ined, two-intersection test. 
 
 INTRODUCTION 
 
 Personnel safety is the first benefit 
 of good roof control, but productivity 
 can also be affected significantly by im- 
 proved roof control procedures. Most of 
 the easily mined coal has been extracted, 
 and future mining will take place in 
 
 areas with difficult roof-control prob- 
 lems . The Bureau of Mines recognized the 
 need for a remotely placed roof bolt 
 of noncorrosive materials that could be 
 remotely installed in longer-than-seam- 
 height lengths . 
 
 BACKGROUND 
 
 The system envisioned would consist of 
 a hollow fiberglass-reinforced bolt and 
 an adhesive that would be pumped into the 
 
 annulus between the bolt and rock. The 
 Bureau initiated work to develop such a 
 system in 1978. 
 
 EARLY DEVELOPMENT 
 
 During the early part of the work, pol- 
 yesters, urethanes, acrylics, epoxies, 
 and inorganics (cement and gypsum) were 
 examined for use as an adhesive. Poly- 
 ester resins widely used today in roof 
 bolting are supplied in cartridges that 
 are used with steel bolts. The poly- 
 esters did not lend themselves to pumping 
 because of resin instability and low 
 flashpoints. Urethanes and acrylics were 
 easily pumped and had fast set times, but 
 were either too expensive or could not 
 
 'Research structural engineer, Spokane 
 Research Center, Bureau of Mines, Spo- 
 kane , WA . 
 
 meet adhesive strength requirements. A 
 coal-tar-based epoxy, one part resin to 
 one part hardener , was developed which 
 could be mixed in a static mixer and 
 pumped by conventional liquid-handling 
 equipment. Gel time requirements of the 
 epoxy were met by preheating of the com- 
 ponents. Cements had fast set times but 
 unacceptable mixing and pumping charac- 
 teristics. It was determined that the 
 epoxy and cements were the most promising 
 and would require additional laboratory 
 testing during the second phase of the 
 program. 
 
127 
 
 The core shape 
 that the core 
 
 Also, during the initial phase of the 
 program, several shapes of fiberglass 
 bolt cores were examined, 
 requirements were such 
 could be coiled on reels for storage on 
 the machine, then formed into a hollow 
 core bolt during installation. Surface 
 modifications on the smooth fiberglass- 
 reinforced polyester (FEIP) core to in- 
 crease bonding strengths of the adhesives 
 appeared necessary. A variety of special 
 wraps and cloth meshes, fabricated on the 
 outer surface of the smooth cores, delam- 
 inated under the severe pull strengths 
 used to evaluate the system. The best 
 method found was to cut a diagonal groove 
 into the outer surface. This provided 
 the mechanical interlock necessary to 
 achieve the required strengths. 
 
 Initial designs led to two FRP bolt 
 core shapes for laboratory testing. 
 An axially pleated, flat-sheet material 
 was tried. It could be stored in a 
 flat coil and pulled as a "tape" to 
 the placement head, which would fold 
 it to form a hollow, cylindrical core. 
 The second core configuration was 
 based on discreet lengths of FRP, each 
 a 1/4-round section which could easily 
 be bent at the head, and the four pieces 
 formed into a hollow core prior to in- 
 sertion into the drilled hole. Again, it 
 was determined that both bolt core con- 
 cepts required additional laboratory 
 testing during the second phase of the 
 program. 
 
 PHASE II DEVELOPMENTS 
 
 The second phase of the program per- 
 mitted more extensive laboratory testing 
 of the core and grout candidates. Con- 
 crete blocks and test equipment were con- 
 structed to allow grouting and pull test- 
 ing of both FRP bolt cores with epoxy and 
 cement adhesives. 
 
 With the aid of several chemical com- 
 panies, providing their own funding, 
 inexpensive hardeners were developed 
 which produced fast gel times without 
 need for preheating. Additional labora- 
 tory work with the cements still produced 
 unacceptable mixing and pumping. It was 
 determined that the epoxy showed the most 
 promise and that work with the cements 
 would be terminated. 
 
 A series of four epoxy formulations was 
 developed that met the 1-min gel times 
 and early strengths needed over the re- 
 quired temperature range of 40° to 
 
 100° F. They had a 1:1 ratio of resin to 
 hardener, were highly filled, and easily 
 mixed within a static mixer. A twin cyl- 
 inder, positive-displacement piston pump 
 was used to meter, mix, and dispense the 
 epoxy resin. 
 
 Both the flat-sheet and the segment 
 core designs were adaptable for continu- 
 ous insertion of longer-than-seam-height 
 lengths. The folded sheet core tended to 
 collapse on itself under cross-shear, 
 while the segmented core design remained 
 stable and was stronger in cross-shear. 
 It was therefore decided to use the seg- 
 mented core design. 
 
 Laboratory tests are continuing to ver- 
 ify that the segmented core design and 
 the new epoxy systems meet or exceed all 
 the strength requirements recommended by 
 the Bureau. 
 
 EPOXY CARTRIDGE DEVELOPMENT AND FIELD TEST 
 
 Evaluation of epoxy grouts in mine con- 
 ditions bypassed the pumping system de- 
 velopment by packaging in cartridge form. 
 Steel bolts with epoxy cartridges were 
 installed in several mines under condi- 
 tions similar to those used for polyester 
 cartridges. In this way, fully grouted 
 
 roof bolts using the epoxy formulation 
 could be easily compared to existing sup- 
 port systems. 
 
 Pull tests on roof bolts with 1- 
 ft point anchors were conducted off- 
 section in the mine. Underground results 
 
128 
 
 verified laboratory testing, in that the 
 epoxy bolts could be installed as easily 
 as the polyester and with greater bonding 
 strengths. 
 
 A double intersection roof-bolt test 
 was carried out in a freshly mined area 
 (pretimbered) of a working Eastern coal 
 mine. Over 200 epoxy-resin-grouted steel 
 bolts were installed in the test without 
 any installation problems. After a 
 
 suitable return passage had been mined, 
 the timbers were removed and the roof was 
 evaluated for stability. The results 
 indicated that the bolted strata were 
 successfully supported and that the epoxy 
 cartridges could be used as a viable 
 alternative to the polyester system. 
 Several private con^janies are pursu- 
 ing the possible marketing of epoxy 
 cartridges. 
 
 NEW PROGRAM 
 
 As a result of the laboratory and mine 
 evaluations of the epoxy bolting system, 
 the Bureau entered into a cost-sharing 
 research program to develop the necessary 
 mineworthy installation equipment to 
 evaluate the concept of a remote- 
 ly placed, pumpable, longer-than-seam- 
 height, epoxy FRP, noncorrosive, roof- 
 bolt system. 
 
 The contractor is furnishing a bolter 
 chassis for the test period. Four major 
 subsystems must be added to the chassis 
 for placement of the pumpable roof bolts: 
 the pumping equipment, purge system, 
 fiberglass core-forming equipment, and a 
 placement head for interfacing each of 
 the subsystems with the drilled roof-bolt 
 holes. 
 
 The contractor will incorporate, on the 
 bolter, all the new design features shown 
 to be needed. The compact pumping system 
 design enables easy material loading and 
 repair from the rear of the bolter. The 
 resin will be pumped by two piston pumps 
 
 which are driven by a cylinder mounted 
 between the pumps. The system also in- 
 cludes a static mixer, valves, and 
 hoses. 
 
 Laboratory testing indicates that a 
 high-pressure inexpensive water purge 
 completely cleans the static mixers. 
 
 The bolter system design includes a 
 three-segment 3/4-in-diam core, three- 
 reel storage, and handling system. The 
 three segments of core will be pulled 
 from the storage reels and driven through 
 rollers designed to form a 3/4-in hollow 
 core bolt when the three segments arrive 
 at the drilled hole. 
 
 The placement device is designed to 
 bring together the epoxy dispenser, purg- 
 ing system, and core former into a conmion 
 head where the components can be prepared 
 for roof bolting. This head will inter- 
 face directly with the mine roof to pro- 
 vide multiple composite bolt placement. 
 
 FUTURE PLANS 
 
 During the next 4 months , the equipment 
 will be laboratory tested. The results 
 will be reviewed and systems redesigned 
 as needed to ensure a meaningful field 
 evaluation. 
 
 involved are satisfied with the system, a 
 standard MSHA-def ined, two-intersection 
 test will be conducted in a cooperating 
 coal mine. 
 
 Field evaluation will be conducted off- 
 section in a coal mine. When all 
 
129 
 
 DEVELOPMENT OF LIGHTWEIGHT HYDRAULIC SUPPORTS 
 By John P. Dunfordi 
 
 ABSTRACT 
 
 The Installing of temporary roof sup- 
 port is an integral part of underground 
 coal mining. The most common forms of 
 temporary supports are wooden posts and 
 metal jacks. Both wooden posts and non- 
 yielding steel jacks are heavy and cum- 
 bersome to install. Yielding hydraulic 
 jacks, while easier to install and more 
 functional, are extremely heavy. With 
 this in mind, the Bureau of Mines studied 
 ways to reduce the weight of contemporary 
 hydraulic supports without sacrificing 
 performance. 
 
 In 1979 a contract was awarded to de- 
 sign and test a lightweight hydraulic 
 mine support. Laboratory testing indi- 
 cated some changes were needed in the 
 basic design selected. After these modi- 
 fications were made, 33 units were sent 
 to the field at three different loca- 
 tions. These units were designed for use 
 in a 6- to 8-ft seam, have a 22-ton ca- 
 pacity, are fully self-contained, provide 
 a 5- to 7-ton roof preload, and yield to 
 
 overload. The total weight of each unit 
 is 55 lb, compared with 110 lb for a com- 
 mercial steel unit. All three of the 
 Western coal mines were extremely pleased 
 with the units and requested to keep 
 using them for as long as possible. Dur- 
 ing the field testing some problems did 
 occur, such as corrosion on the piston 
 surfaces , weak pump handles , and short 
 life span of some internal seals. All of 
 these problems were corrected. 
 
 During the field tests, the need became 
 apparent for units that would function in 
 seam heights other than 6 to 8 ft. 
 Units of the same configuration and ca- 
 pacity were built and tested for seam 
 heights in the 4.5- to 6-ft, 8- to 10-ft, 
 and 10- to 12-ft ranges. After the test- 
 ing proved successful, arrangements were 
 made to install 10 of the 10- to 12-ft 
 units in a mine for field testing. The 
 results have not been completed at this 
 time. 
 
 INTRODUCTION 
 
 The installing of ten^)orary roof sup- 
 port Is an integral part of underground 
 coal mining. Historically, wooden posts, 
 metal screw jacks, and, more recently, 
 telescoping hydraulic supports are used 
 for this purpose. Wooden posts are time- 
 consuming and cumbersome to install, not 
 consistent in size and strength, and also 
 represent a sizable constant cost. Screw 
 jacks, while easier to install, have no 
 yield capability and are prohibitively 
 heavy where high strength supports are 
 required. The all-hydraulic, telescoping 
 cylinder concept remains the best choice 
 for reusable temporary support. 
 
 There are now on the market various 
 hydraulic supports made of steel tubing 
 and steel components that perform very 
 
 ^Mining engineer, Spokane Research Cen- 
 ter, Bureau of Mines, Spokane, WA. 
 
 well as a temporary support. Unfortu- 
 nately, due to increasing support-load 
 requirements and the thicker coal seams 
 found in the West, these steel supports 
 are becoming extremely heavy and cumber- 
 some to use. A support rated at 22 tons 
 and used in a 6- to 8-ft coal seam weighs 
 in excess of 110 lb. In addition, the 
 same type support designed for use in a 
 12-ft seam weighs approximately 300 lb. 
 
 With this in mind, the Bureau of Mines 
 funded a contract to look at state- 
 of-the-art technology in lightweight hy- 
 draulic cylinder design in an effort to 
 construct a high strength-to-weight ratio 
 yielding temporary roof support. 
 
 This work was started in 1978 and com- 
 pleted in February 1982. This paper 
 deals with the evolution of the project 
 through the contract and in-house phases. 
 
130 
 
 CONCEPT EVALUATION AND DETAILED DESIGN 
 
 The contract initially called for exam- 
 ination of previous work in the area of 
 lightweight supports, as well as existing 
 commercially available materials and 
 products. The review identified the fol- 
 lowing major desirable designs features: 
 
 1. A 22-ton capacity. 
 
 2. A goal weight of 50 lb. 
 
 3. A prop with controlled yielding at 
 overload. 
 
 4. A fully self-contained unit (i.e., 
 integral reservoir and pressurization 
 unit). 
 
 5. A remotely recoverable prop. 
 
 6. A prop easily and quickly 
 installed. 
 
 Two concepts selected for detailed design 
 incorporated hydraulic mechanisms that 
 would yield at overload and be remotely 
 recoverable. Of the two designs (totally 
 hydraulic and hydromechanical) , the all- 
 hydraulic version was selected for proto- 
 type production and testing (fig. 1). 
 Although the hydromechanical support was 
 lighter in weight, it was rejected be- 
 cause of its mechanical complexity, cost, 
 and complexity of installation in a min- 
 ing situation. 
 
 FABRICATION AND TESTING OF LIGHTWEIGHT SUPPORTS 
 
 FABRICATION 
 
 Four prototype models of the all- 
 hydraulic lightweight support were fabri- 
 cated to verify the design. Based on 
 value and manufacturing engineering con- 
 siderations, some design changes were 
 made prior to release for fabrication. 
 Additional changes were found to be 
 
 FIGURE 1. - Six- to eight-hydraulic support. 
 
 necessary during assembly and functional 
 testing. Chief among these were modify- 
 ing the pump block assembly and redesign- 
 ing the main cylinder. 
 
 For structural testing, all four proto- 
 type units were modified. The four were 
 laboratory tested to ensure functional 
 stability and correctness of design. Two 
 props were tested to ultimate load fail- 
 ure to verify column ultimate strength 
 calculations. 
 
 Following the completion of successful 
 functional and structural testing of the 
 models, a production lot of 40 additional 
 units was constructed (fig. 2). It was 
 intended that 10 of these supports would 
 be provided to each of four different 
 mines for a 6-month demonstration and 
 evaluation period. 
 
 Following the final assembly, each of 
 the production units was subjected to 
 operational testing. This consisted of 
 pressure testing the internal preload 
 limit valve, the 22-ton yield bypass 
 valve, and preloading the unit in a test 
 frame overnight to detect valve or seal 
 leaks. 
 
 An independent testing program was per- 
 formed by the Bureau to corroborate the 
 
131 
 
 Manual pump 
 handle 
 
 ^k 
 
 Remote retrieval 
 lanyard ring 
 
 [J 
 
 -Hydraulic fluid 
 reservoir 
 
 Pressure release 
 valve 
 
 r-1 
 
 -Carrying handle 
 bracket 
 
 1^ 
 
 Telescoping hydraulic 
 piston 
 
 FIGURE 2. - Components of the hydraulic support. 
 
 contractors' findings. During these 
 tests, one of the reservoir tube welds 
 cracked and failed. Subsequent analysis 
 by both the Bureau and FMC Corp. 's Mate- 
 rials Laboratory revealed that the weld 
 was faulty. Other faulty welds were 
 detected through radiographic analysis, 
 so it was decided to reweld and re- 
 heat-treat all of the reservoir tube 
 assemblies. 
 
 UNDERGROUND TESTING 
 
 One of the test sites in eastern Utah 
 incorporated the supports into a continu- 
 ous mining development section. The min- 
 ers used the supports continuously, ex- 
 cept for the 3-month-long United Mine 
 Workers strike (six of the supports were 
 left, fully loaded, across an entry for 3 
 months with no failures) , and felt the 
 lightweight supports were superior to 
 what they had been using as far as hand- 
 ling and setting were concerned. 
 
 The supports were removed from the mine 
 in November 1981 for repair. All of the 
 handles were replaced with steel ones. 
 Eight of the ten supports were rebuilt 
 with cannibalized parts from two supports 
 that had been hit by mining equipment and 
 were beyond repair. All of the work was 
 done locally and sent back to the mine 
 for further use. The units were rebuilt 
 again in June 1982. The six remaining 
 units were then used in a continuously 
 advancing development section as part of 
 a prop and beam temporary support system. 
 These were used until August 1982 when 
 they were pulled from the mine for over- 
 haul. Due to lack of production, it was 
 decided to send the units back to the 
 Bureau for rebuilding and inspection. 
 
 Another site, located in western Colo- 
 rado, did not install the supports until 
 May 1981. Eight units were being used as 
 part of their longwall tailgate support 
 plan. As of April 1982, the only prob- 
 lems encountered were one broken pump 
 handle and one sticking release valve. 
 After that date the longwall was shut 
 down and the supports were used in vari- 
 ous applications such as setting brattice 
 curtain lines, setting chain-link fence 
 at pillar ribs , and in longwall panel 
 development work. 
 
 During the latter part of the fabrica- 
 tion phase, underground coal mines were 
 contacted to solicit interest in partici- 
 pation in the demonstration of the light- 
 weight hydraulic supports. Three mines, 
 two in Colorado and one in Utah, were 
 selected to receive the props for under- 
 ground testing. Each mine was given 10 
 supports to use under actual mining 
 conditions. 
 
 The supports were removed from the mine 
 in November 1982 for overhaul. In March 
 1983, the units were again sent under- 
 ground for use in a development section 
 and also to be used as additional support 
 around a drilling operation for a methane 
 drainage project. 
 
 At the third test site, located outside 
 Grand Junction, CO, the supports were 
 
132 
 
 used in development work. In order to 
 work in the 8-1/2-ft coal seam, a 2-ft 
 steel extension was added to the support. 
 
 The supports were used in slow develop- 
 ment work from April until November 1981, 
 
 at which time most of the units needed 
 some repair. Problem areas were failure 
 of the main seal , sticking pressure re- 
 lease valve, and pressure pump piston. 
 These supports were sent back to the Bu- 
 reau for inspection and rebuilding. 
 
 OVERALL TEST CONCLUSIONS 
 
 In the three test mines, all comments 
 about the test were positive. In all 
 cases , both miners and management liked 
 the supports and wanted to continue to 
 use them. Although several areas of 
 weakness became evident, they were not 
 
 major problems. It should be noted that 
 although the duty-cycle was only about 6 
 months , the problems exhibited were con- 
 sistent with those props currently on the 
 market. 
 
 CURRENT STATUS 
 
 All three test mines have expressed a 
 desire to continue using the supports in 
 their mining sequence. These mines and 
 various other mines have requested infor- 
 mation about commercial availability of 
 the props. 
 
 A project to test the various lengths 
 of lightweight supports is being per- 
 formed by the Bureau, since the various 
 lengths require slightly different de- 
 signs. The first step in this project 
 was to have prototype supports built 
 using the updated shop drawing package. 
 A contract was let to fabricate light- 
 weight supports in the 4.5- to 6-ft, 8- 
 to 10-ft, and 10- to 12-ft range. These 
 supports were received in February 1983. 
 A structural testing program to confirm 
 design calculations was performed in late 
 June 1983. At the same time, modifica- 
 tions are being made to the support, 
 based on field test results, that will 
 increase the duty-cycle and operating 
 ability. Some of these changes included: 
 redesign of the top end of the cap for 
 
 more bearing surface and to facilitate 
 nailing a cap piece on to the longer sup- 
 ports prior to setting the unit. Also, a 
 redesign of the lanyard ring assembly 
 will be performed along with plating the 
 pump pressure piston and pressure release 
 piston. Adding 3 in to the pump handle 
 will help in applying the preload pres- 
 sure. The use of various seals in the 
 longer units will be investigated. An 
 aluminum extension has been fabricated 
 and is due for field testing in June 
 1983. Three units were fitted with sight 
 pressure gauges and sent to Alaska to be 
 used during a rock -mechanics project in a 
 permafrost gold mine. Six other 6- to 
 8-ft units were fitted with pressure 
 transducers and will be used in a retreat 
 coal mining sequence to monitor loading 
 in a breaker prop row. 
 
 Several of the longer supports will be 
 fabricated and tested. Evaluations will 
 be completed and results published in the 
 near future. 
 
133 
 
 MOBILE ROOF SUPPORT AND APPLICATIONS IN RETREAT MINING 
 By Robert R. Thompsoni 
 
 ABSTRACT 
 
 Retreat pillar mining is highly pro- 
 ductive, but dangerous. The primary dan- 
 ger during pillar removal is premature 
 caving of the roof. The Bureau of Mines 
 has developed a remotely operated machine 
 that will place and retrieve temporary 
 roof support. The prototype machine 
 worked well but had several problems, the 
 
 primary one being tramming. Two second- 
 generation machines were built under a 
 cost-sharing program with a Utah coal 
 mine and a mining equipment con^jany. The 
 machine carries four 50-ton jacks and is 
 remotely controlled by radio. The ma- 
 chines are presently being tested under- 
 ground in a Utah coal mine. 
 
 INTRODUCTION 
 
 Retreat pillar mining is highly pro- 
 ductive because supply, haulage, ventila- 
 tion, and power systems are established, 
 and there is also the advantage of the 
 knowledge gained during development such 
 as roof and ground behavior and hydro- 
 logic factors. The primary danger during 
 pillar removal is premature caving of the 
 roof. 
 
 The roof must cave in a predictable and 
 dependable manner to prevent inducing 
 excessive abutment loads in adjacent pil- 
 lars, which can result in rib bursts. 
 
 floor heave, or crushed pillars. The 
 safety of the miners is dependent on 
 successfully controlling the roof. Roof 
 support during retreat is usually ob- 
 tained by setting posts, cribs, hydraulic 
 props, roof bolts, or a combination of 
 these devices. They are set manually by 
 a miner working in a hazardous area. 
 Many of these devices can never be re- 
 covered and thus become part of the cost 
 of extracting coal. The problem is how 
 to set and retrieve these roof supports 
 safely. 
 
 MOBILE ROOF SUPPORT SYSTEM 
 
 The Bureau embarked on a project to 
 develop a system that would place and 
 retrieve temporary roof supports without 
 danger to the operator. In conjunction 
 with a contractor, a mobile roof support 
 (MRS) machine (fig. 1) was designed, 
 built, and field-tested. THE MRS was re- 
 motely operated, battery-powered, and 
 rubber-tired. It carried four jacks, two 
 on the body of the machine, and two at 
 the end of hinged arms. The jacks extend 
 
 ^Research structural engineer, Spokane 
 Research Center, Bureau of Mines, Spo- 
 kane, WA. 
 
 to form columns between the floor and 
 roof, each with 30 tons of potential sup- 
 port. The jacks were hydraulically 
 locked, and the load distributed to three 
 points on each jack, without loading the 
 machine chassis. 
 
 The MRS was tested underground in an 
 Illinois coal mine. The prototype ma- 
 chine worked well but had several prob- 
 lems, the primary one being tramming over 
 the soft floor. Tramming was slow and, 
 at times, the tires became buried and 
 machine had to be pulled out. 
 
134 
 
 FIGURE 1. - First-generation mobile roof support. 
 
 SECOND-GENERATION MOBILE ROOF SUPPORT MACHINE 
 
 Results of the field test were encour- 
 aging. The concept had acceptance by 
 both mine management and the miners. 
 Several industry personnel witnessed the 
 field trials and felt the MRS would im- 
 prove the safety and productivity of 
 their retreat mining sections. With this 
 encouragement , the Bureau decided that a 
 second-generation machine should be 
 built, correcting the problems of the 
 prototype. 
 
 A Utah coal mining company offered its 
 mine as a test site. It also offered to 
 work with the Bureau during the design 
 
 and fabrication of the second-generation 
 machine and to share some of its costs. 
 This operation removed 12 ft of a 15-ft 
 coal seam and had a four-member support 
 crew setting posts. This time-consuming 
 and hazardous task, at times, occurred 
 under unsupported roof. 
 
 A mining equipment company became in- 
 terested in producing the MRS. It 
 offered to help cost-share by supplying 
 some of the parts for the new machines. 
 It also offered to assist during the 
 design and fabrication. 
 
135 
 
 FIGURE 2. - Second-generation mobile roof support. 
 
 BASIC MACHINE REQUIREMENTS 
 
 The Bureau embarked on a research pro- 
 gram to design, fabricate, and field-test 
 two second-generation MRS's (fig. 2). 
 During the first 3 months, participants 
 met monthly. During this preliminary 
 design phase, all participants agreed 
 that the machine in the tram mode should 
 be 8 ft wide and 10 ft long, or less, 
 with at least a 12-in ground clearance. 
 Also, it should be of rugged construction 
 with towing hooks on both ends and 
 mounted on independently controlled 
 crawlers that exerted 20 psi, or less, 
 ground pressure. The machine should be 
 
 powered by a 40-hp, 460-V ac permissible 
 motor from a 260-ft reeled trailing cable 
 and transmitter remote controls. Re- 
 versible variable tram speed of at least 
 80 ft/min on 20 pet grade, and free- 
 wheeling for emergency towing, were to be 
 included. Front and back dozer blades of 
 12-in range are required. 
 
 The machine was to carry two chassis 
 mounted and two swing-arm jacks of 7- to 
 15-ft working height. The jacks were to 
 have a 3- to 8-ton installation and 50- 
 ton maximum loading capabilities. The 
 
136 
 
 swing jacks are to form a breaker row conditions, the jacks will be capable of 
 
 with 6-ft separation between chassis 
 jacks and have a visual load indicator. 
 In case of heavy ground, the ability to 
 remote-jettison either or both swing 
 jacks is incorporated. In order to 
 attain rapid egress under bad roof 
 
 retraction, so as to provide 1 ft of 
 ground clearance and 1 ft of roof clear- 
 ance in 30 s. The mine requested that 
 the machines be remotely controlled by 
 radio. 
 
 UNDERGROUND TESTING 
 
 After the machines are fabricated, they 
 are being shipped to the mine to be 
 tested for a period of 6 months. Figure 
 3 shows the test mine's ground control 
 plan during full-pillar extraction. It 
 required setting 24 posts for pulling 
 each fender. The use of the two machines 
 will eliminate the requirement for set- 
 ting these posts. Figure 4 shows the 
 planned sequential operation with the 
 MRS. Again, the machine will be moved 
 and set remotely, thus eliminating expo- 
 sure of the miners setting the posts by 
 hand. Yet to be determined is the 
 
 possible use of some posts to act as 
 "squealers" or warning devices. These 
 posts, if required, would be set after 
 the MRS is in place and supporting the 
 roof. 
 
 The mine is expected to use the ma- 
 chines for a period of up to 10 jrr , thus 
 providing long-term testing. The pro- 
 jected mass-produced costs of the ma- 
 chines, with tethered remote-control 
 rather than radio, are estimated at 
 $125,000. The radio remote-control is 
 expensive and is considered as an extra. 
 
 • • • • 
 
 • • • • 
 
 
 FIGURE 3. - Ground control plan during full 
 pillar extraction. 
 
 FIGURE 4. - Sequential operation with machines. 
 
137 
 
 ROOM-AND-PILLAR RETREAT MINING 
 
 The Bureau published a manual for the 
 coal industry, 2 which is to provide mine 
 managers and engineers with: 
 
 1. Assistance in making decisions to 
 retreat mine and in selecting the best 
 mining technique for their specific 
 condition. 
 
 3. Information to develop a section 
 foreman's handbook on retreat mining 
 safety and operation. 
 
 Copies may be obtained from the Su- 
 perintendent of Documents, U.S. Govern- 
 ment Printing Office, Washington, DC 
 20402. 
 
 2. Information on efficient retreat 
 mining design. 
 
 SUMMARY 
 
 The Bureau has des^-gned and built a 
 support system for retreat mining that 
 can be set and retrieved remotely. The 
 
 ^Kauffman, P. W., S. A. Hawkins, and 
 
 R. R. Thompson. Room and Pillar Retreat 
 
 Mining. A Manual for the Coal Industry. 
 
 BuMines IC 8849, 1981, 228 pp. 
 
 system is now being tested in a Utah coal 
 mine. This system provides added safety 
 for the miner, by eliminating the need to 
 work in a hazardous area setting posts, 
 cribs, or hydraulic props. The MRS will 
 also increase productivity, since the 
 number of manual support setting opera- 
 tions has been decreased. 
 
138 
 
 INORGANIC GROUTS FOR ROOF BOLTING 
 By Jack E. Fraley^ 
 
 ABSTRACT 
 
 The Bureau of Mines investigated rapid- 
 hardening material substitutes for the 
 resin used in mine roof bolts. Gypsum 
 plasters (CaS04 • I/2H2O) were selected be- 
 cause they have high early strength while 
 being readily available and inexpensive. 
 
 Gypsum plaster-water capsule cartridges 
 provide a substitute for resin car- 
 tridges. Gypsum plaster holds promise 
 for injection in roof bolt holes as a 
 premixed slurry because of improved oper- 
 ator safety and greater economy. 
 
 INTRODUCTION 
 
 Fully grouted resin bolts are a rela- 
 tively new phenomenon to the mining in- 
 dustry. In 1972, the advantages of resin 
 bolts became apparent, and by 1980, an 
 estimated 20 million of these bolts were 
 installed. Since resin bolts are more 
 costly than mechanical bolts, their su- 
 perior performance is illustrated by the 
 large increase in their usage. 
 
 The price of resin doubled between 1973 
 and 1975. Since resin is petroleum- 
 derived, its cost and future supply are 
 uncertain. Because resin cartridges are 
 flammable, they are a potential under- 
 ground fire hazard. 
 
 To overcome these resin disadvantages, 
 the Bureau of Mines started to investi- 
 gate rapid-hardening material substi- 
 tutes. Gypsum plasters (CaS04 •I/2H2O) 
 were selected because they have high 
 early strength while being readily avail- 
 able and inexpensive. To achieve desired 
 rapid hardening, the plaster is acceler- 
 ated by adding 1 pet K2SO4 (by weight of 
 dry plaster). 
 
 ^Chemical engineer, Spokane Research 
 Center, Bureau of Mines, Spokane, WA. 
 
 JV^ 
 
 -Cartridge 
 
 -Voids (air) 
 
 -Cement 
 
 Microcapsules 
 of water 
 
 -Grout 
 
 -Rebar 
 
 ABC 
 
 FIGURE 1. - Gypsum-plaster, water-capsule bolt. 
 
139 
 
 WATER-CAPSULE CARTRIDGES 
 
 A gypsum-plaster, water-capsule car- 
 tridge was made, as shown in figure 1. A 
 cartridge (packaged similarly to resin 
 cartridges) (fig. lA) is inserted into a 
 drilled hole. The cartridge wrapper is 
 filled with accelerated gypsum plaster 
 and water capsules, but also contains air 
 as void spaces between the fine gypsum 
 particles, as shown in figure IB^. During 
 rebar insertion (fig. 1£) , the water cap- 
 sules rupture, releasing the water, which 
 mixes with the plaster to form hardened 
 gypsum. 
 
 Figure 2 shows each component in the 
 system. The plaster is on the upper 
 
 left, and the water capsules on the upper 
 right. A cartridge is in the center, 
 while a short length of rebar is at the 
 bottom. 
 
 The water capsules appear in figure 3 
 alongside a penny, so their size can be 
 noted. Typically, the capsule diameters 
 are 1,800 pm (0.071 in). The water cap- 
 sules are a modified wax shell surround- 
 ing water (encapsulated water). They 
 contain over 60 wt pet water; and to be 
 of adequate quality, they must retain the 
 water and be durable enough to withstand 
 normal handling during cartridge produc- 
 tion and installation. 
 
 FIGURE 2. - Components of the gypsum-plaster, water-capsule bolt. 
 
140 
 
 FIGURE 3. - Water capsules. 
 
 During installation, the cartridges are 
 manually inserted, as shown in figure 4. 
 The plaster-water mixing is shown in fig- 
 ure 5. The water capsules contain blue 
 dye, which is visible as the rebar is 
 inserted through a cartridge inside a 
 clear plastic tube. 
 
 The installation procedure for the 
 inorganic cartridges is similar to that 
 for resin cartridges, except that less 
 rotation for mixing is required. After 
 the cartridges are inserted in a drilled 
 hole, the bolt is inserted part way while 
 being rotated. A wrench is placed on the 
 bolting machine rotation chuck, which 
 allows complete insertion with the avail- 
 able vertical movement. After the wrench 
 
 is in place, the bolt is rotated during 
 the remainder of the insertion. Extended 
 spinning, with the bolt fully inserted, 
 is not required. As the bolt is in- 
 serted, pressure builds within the hole 
 and ruptures the water capsules. The wet 
 mix that is formed is stiff enough so the 
 bolt remains in place, allowing immediate 
 lowering of the bolter head. 
 
 Ninety percent of the gypsum strength 
 is developed in less than 10 min. Pull 
 strengths of the bolts (4-ft, Grade 50, 
 No. 6 rebar) done in concrete blocks 
 exceeded the bolt yield point which 
 is over 22,000 lb. Table 1 shows the 
 pull strengths 10 to 13 min after 
 installation. 
 
141 
 
 FIGURE 4. - Manual insertion of cartridges. 
 
 TABLE 1. - Pull strength of gypsum-water 
 capsule bolts 
 
 Bolt 
 
 Pull, lb 
 
 Time , mln 
 
 2 
 
 26,150 
 22,400 
 21,400 
 25,200 
 22,400 
 21,400 
 21,400 
 24,300 
 22,400 
 
 10 
 
 3 
 
 13 
 
 4 
 
 10 
 
 5 
 
 10 
 
 6 
 
 10 
 
 7 
 
 10 
 
 9 
 
 11 
 
 11 
 
 12 
 
 17 
 
 10 
 
 When the bolt is thrust into a confined 
 cartridge, the pressure has an effect on 
 subsequent pull strength. Pressure as 
 high as 5,500 psi has been measured dur- 
 ing bolt insertion in 4-ft holes. Four 
 pull test samples were made by pouring 
 concrete in 4-in pipes that were 4 ft 
 long. After installation of 4-ft bolts 
 in the samples , the samples were cut into 
 eight 6-in sections. Each section had an 
 extension rod attached to the bolt for 
 pull testing. The average pull strength 
 for the four 6-in sections closest to the 
 
142 
 
 bolt head (sections A-D) , as well as that 
 for the four 6-in sections farthest from 
 the bolt head (sections E-H) , is given in 
 table 2. The pull strengths were higher 
 in the half of the bolt furthest up the 
 hole, where pressures were higher, owing 
 to increased cartridge confinement. 
 
 TABLE 2. - Average pull strength of 6-in 
 bolt section, pounds 
 
 Bolt 
 
 Section A-D 
 average 
 
 Section E-H 
 average 
 
 1. 
 8., 
 
 12, 
 16, 
 
 8,119 
 
 6,450 
 
 11,156 
 
 14,725 
 
 8,756 
 
 6,963 
 
 17,538 
 
 14.794 
 
 Bolts were installed in foamed concrete 
 to measure the effect of the weaker rock 
 on pull strength. One bolt gave 9,000 lb 
 and another 15,700 lb pull. More spin- 
 ning during bolt insertion appears to en- 
 large the hole in weaker materials like 
 foamed concrete. The smaller amount of 
 required spinning in installing the 
 gypsum-water capsule bolt may become an 
 advantage in weaker rock. 
 
 The cartridges are made to have a 
 water-to-plaster weight ratio of 0.30 to 
 0.35. Less water gives Increased 
 strength; however, rebar insertion is 
 more difficult, so the 0.30 weight ratio 
 is the approximate lower limit for re- 
 peatable rebar installation. 
 
 An additive that increases the fluidity 
 of freshly mixed grout at any water- 
 to-plaster ratio makes the mixing and re- 
 bar insertion easier and reduces bolt in- 
 sertion time. The additive allows rebar 
 insertion at water-to-plaster ratios as 
 low as 0.21. This provides a greater 
 margin for successful underground instal- 
 lation with variations in the cartridges 
 and operator-equipment techniques. By 
 lowering the quantity of water as water 
 capsules, cost savings of a couple cents 
 per cartridge are possible. 
 
 FIGURE 5. - Mixing of gypsum plaster and 
 water capsules. 
 
 A research contract was issued to de- 
 velop the packaging technology for com- 
 mercial production of the cartridges. 
 An encapsulation system to provide the 
 
143 
 
 water capsules was developed along with 
 the equipment to produce 20 cartridges 
 per minute. The contractor needs private 
 
 venture capital to develop a complete 
 production plant and marketing system for 
 the cartridges. 
 
 WATER TUBE CARTRIDGES 
 
 Cartridges that replace the water cap- 
 sules with a tube of water are being 
 studied. Figure 6 shows two methods of 
 storing the water within the cartridge. 
 Panel A shows a continuous-length water 
 tube inside the cartridge. As the rebar 
 begins to push on the cartridge, the com- 
 pression shortens the cartridge length. 
 Flexible water tubes like lay-flat tubing 
 are thought to bend as the cartridge com- 
 presses, rather than rupture and release 
 the water. The result is hard rebar 
 insertion due to poorly mixed plaster. 
 
 A semirigid tube that retains its shape 
 releases its water sooner to enhance mix- 
 ing and ease rebar insertion. Several 
 tube materials have been investigated to 
 find one capable of retaining water, 
 withstanding normal handling during 
 
 V ////////////y////// 
 
 Gypsum 
 
 Water tube 
 
 - f^Z^QOO 
 
 B 
 
 / A Water package 
 
 C J Gypsum package 
 
 FIGURE 6. - Water tube cartridges. 
 
 cartridge preparation and installation, 
 and breaking as soon as the cartridge is 
 compressed. 
 
 Figure 6B^ illustrates alternate pack- 
 ages of water and plaster within the car- 
 tridge wrapper. As with the continuous- 
 length water tube, best rebar insertion 
 is obtained if the water is released as 
 the cartridge is first compressed. 
 
 To reduce the loss of fluid material 
 from the hole during rebar insertion, a 
 plastic cap can be inserted over the 
 rebar , which plugs the hole as the rebar 
 installation progresses (fig. 7). 
 
 r~ji 
 
 Cap 
 
 FIGURE 7, - Cap to reduce loss of fluid from 
 hole during rebar insertion. 
 
144 
 
 Water tube cartridges require more mix- 
 ing than water capsule cartridges because 
 the water is not dispersed throughout the 
 cartridge. To ease water-plaster mixing, 
 an additive can be mixed with the plaster 
 
 to increase its fluidity at a given water 
 content. Also, the surface tension of 
 the water can be reduced to increase its 
 ability to wet the plaster. 
 
 SLURRY INJECTOR 
 
 The slurry injector is a bulk injection 
 method that lends itself well to remote- 
 control operations. The operator can 
 work under supported roof during full- 
 column bolting. Dry gypsum plaster is 
 automatically mixed with water to form a 
 slurry, pumped into a delivery hose, and 
 injected up the roof -bolt hole, without 
 placing any mechanical device in the 
 hole. The system uses a twin-screw ex- 
 truder (fig. 8) normally used for pro- 
 cessing plastics. The geometry of the 
 screws makes them self-cleaning. The 
 extruder mixes and pumps the grout into a 
 20-ft delivery hose attached to a trans- 
 fer device. After the grout is in the 
 hose, the transfer device inserts a 
 plastic "rabbit" or plug behind the 
 grout. High-pressure air then drives the 
 
 rabbit and grout through the hose and 
 nozzle into the roof -bolt hole. The 
 delivery hose is cleaned by the rabbit. 
 The nozzle is positioned beneath the hole 
 by a linkage mounted on the bolter drill 
 head. 
 
 To operate the system, the hopper is 
 filled with gypsum plaster and a tank is 
 filled with water. A control knob se- 
 lects the proper grout volume and water- 
 to-plaster weight ratio for the size hole 
 being drilled. After the hole is 
 drilled, the nozzle is positioned under 
 the hole and the rabbit inserted into the 
 transfer device. The machine is switched 
 on and all operations through complete 
 bolt insertion are then automatic. 
 
 Transfer device 
 
 Transmission 
 
 Delivery hose 
 
 screws 
 
 Knife 
 valve 
 
 Hydraulic 
 motor 
 
 FIGURE 8. • Slurry injector components. 
 
CONCLUSIONS 
 
 145 
 
 1. Gypsum plaster-water capsule car- 
 tridges provide adequate bonding of No. 6 
 rebar in 4-ft holes. 
 
 4. Capital is necessary for commercial 
 production of the gypsum plaster-water 
 capsule cartridges. 
 
 2. The gypsum plaster-water capsule 
 cartridges provide a substitute for resin 
 cartridges. 
 
 5. Water tube cartridges should be 
 further investigated, as they may be eas- 
 ier to manufacture and more economical. 
 
 3. The gypsum plaster-water capsule 
 cartridges may be an advantage in softer 
 rock where more spinning disturbs the in- 
 tegrity of the hole. 
 
 6. The slurry injector is a promising 
 concept because of possible improvements 
 in operator safety and econoiny of bolt 
 installations. 
 
146 
 
 RESEARCH, DEVELOPMENT, AND USE OF STEEL-FIBER-REINFORCED 
 CONCRETE CRIBBING FOR MINE ROOF SUPPORT 
 
 By Thomas W. Smelser^ and Dale A. Didcoct^ 
 
 ABSTRACT 
 
 Through the combined efforts of the health and safety conditions with a 
 
 Bureau of Mines and private industry, 
 steel-fiber-reinforced concrete crib 
 block to improve coal mine safety in the 
 area of roof control has been developed 
 and is currently in use commercially. 
 The development objective was to improve 
 
 stiffer, stronger, nonflammable roof sup- 
 port at a competitive cost. During the 
 time this method of roof control has been 
 in use in the mining industry, it has 
 proven to have definite functional and 
 economic advantages. 
 
 INTRODUCTION 
 
 In 1975, the Bureau initiated develop- 
 mental research in the area of a steel- 
 fiber-reinforced concrete crib block as 
 an alternative to the use of wood cribs. 
 Aside from the obvious disadvantages of 
 low stiffness and strength, deterioration 
 from chemical and bacteriological attack, 
 methane liberation, and f lammability , 
 wood cribs are subject to variables in 
 cost and availability of suitable vari- 
 eties in many areas of the country. In 
 the Eastern States, a reliable source of 
 pressure-treated wood is often difficult 
 to find, and the scarcity of any type of 
 wood in some Western States makes it ex- 
 tremely expensive. Concrete offered the 
 lowest cost support and appeared to be 
 the best choice for replacing wood sup- 
 ports if a solution could be found for 
 its poor failure characteristics. The 
 addition of steel fibers to the concrete 
 would greatly improve the safety factor 
 by eliminating the possibility of sudden 
 brittle failure. 
 
 Starting in 1976, a project was con- 
 ducted at the Bureau's Spokane (WA) Re- 
 search Center, under the direction of 
 G. L. Anderson, Research Structural En- 
 gineer, and Thomas W. Smelser , Super- 
 visory Mechanical Engineer. The many 
 
 ^Supervisory mechanical engineer, Spo- 
 kane Research Center, Bureau of Mines, 
 Spokane , WA . 
 
 ^vice president. Underground Technology 
 Div., Burrell Group, Morristown, TN. 
 
 types of fibers that were considered in 
 the initial testing included glass, as- 
 bestos, aramid, nylon, carbon, polypropo- 
 lene, and steel. Except for steel, the 
 aforementioned fibers were each elimi- 
 nated owing to various problems with mix- 
 ing characteristics, economy, health haz- 
 ards, strength, etc. Steel fibers also 
 appeared to be the lowest cost solution, 
 as compared with the use of steel- 
 reinforcing bars or steel wire mesh. The 
 steel fiber types considered included 
 straight-round, straight-flat, crimped- 
 full-length, melt-extracted, deformed- 
 full-length, and bent-end. The bent-end 
 type fiber was selected, based on per- 
 formance and cost considerations. 
 
 Upon completion of the selection of the 
 fibers to be used, the next step was to 
 select a concrete mix design and deter- 
 mine the quantity of fibers for the mix. 
 The steel-fiber-reinforced concrete (SFC) 
 support members that were developed 
 offered a significant improvement in 
 stiffness and strength in compression 
 compared with wooden cribs, yet they 
 avoided the brittle or catastrophic com- 
 pressive failure mode of plain concrete. 
 Full-scale compression testing showed the 
 ultimate compressive strength of SFC 
 blocks to be 4,000 psi, compared with 500 
 
 psi for wood. Because eight times more 
 wood is necessary to equal the strength 
 of the steel-fiber-reinforced concrete 
 crib block, the use of SFC greatly in- 
 creases the area for movement of 
 
147 
 
 personnel and equipment and for ventila- 
 tion airflow. The 4,000-psi concrete was 
 selected as optimum strength by most 
 closely approximating the strength of a 
 typical roof and floor structure of a 
 coal mine. 
 
 Bureau Report of Investigations 8412, 
 published in 1980, contains support sys- 
 tem design and results of laboratory 
 investigation and full scale testing. 3 
 
 A successful installation of steel- 
 fiber-reinforced concrete cribs at 
 Kaiser's Sunnyside Mine in Utah was made 
 in 1976 as part of the single-entry 
 
 longwall demonstration at that mine. The 
 wood cribs being used to support the 
 single entry were not stiff enough to 
 hold the tailgate section of the entry 
 open after passage of the first longwall 
 face. Although limited, the demonstra- 
 tion showed the promise of concrete mine 
 support systems and resulted in the 
 project, covered by this report, to more 
 thoroughly characterize and evaluate 
 materials in the laboratory for improved 
 support characteristics, safety, and 
 economy. A follow-on program field- 
 tested the support systems to verify in- 
 stalled cost, structural behavior, and 
 industry acceptance. 
 
 DEVELOPMENT OF MANUFACTURING PROCESS 
 
 The prototype blocks made by the Bureau 
 in the research and testing phase were 
 manufactured on a small production basis 
 and cost approximately $7 per block. 
 This cost was prohibitive, and a meth- 
 od of mass-producing the steel-fiber- 
 reinforced blocks had yet to be devel- 
 oped. A major manufacturer of concrete 
 block and steel fiber gunite mixes, Bur- 
 rell Construction and Supply Co., New 
 Kensington, PA, was contacted by the re- 
 search team. 
 
 With the Bureau contributing their 
 technical specifications, and Burrell 
 Construction and Supply Co. providing the 
 formulation, production technique, and 
 
 ■^Anderson, G. L., and T. W. Smelser. 
 Development Testing and Analysis of 
 Steel-Fiber-Reinforced Concrete Mine Sup- 
 port Members. BuMines RI 8412, 1980, 
 38 pp. 
 
 mix methods, an economical SFC block was 
 produced in the summer of 1980. 
 
 By summer 1981, Burrell had developed a 
 process and method (patent applied for) 
 for mixing the concrete and steel fibers 
 to produce crib blocks that fulfilled the 
 standards set by the Bureau. Many full- 
 scale crib tests were performed at Lehigh 
 University, Pittsburgh Testing Lab, and 
 on the Mine Roof Simulator at the U.S. 
 Department of Energy (DOE) in Bruceton, 
 PA (figs. 1-4). The results of the DOE 
 testing are contained in the DOE Report 
 MRS-DR-81-05.4 
 
 ^yrd, R. J., and J. L. Thompson. Mine 
 Roof Simulator Data Report Steel- 
 Fiber-Reinforced Concrete Roof Cribs 
 (Three Configurations). U.S. Dep. 
 Energy, Min. Equip. Test Facility, MRS- 
 DR-81-05, Aug. 1981, 40 pp. 
 
148 
 
 FIGURE 1. - Solid crib failure mode (DOE tests). 
 
 2,100 
 
 
 ANALOG DATA 
 
 Text No. 30901 
 
 1 
 
 1 1 1 1 
 
 1 1 
 
 1,800 
 
 - 
 
 
 - 
 
 1,500 
 
 A 
 
 
 - 
 
 1,200 
 
 - / 
 
 \ 
 
 - 
 
 900 
 
 - / 
 
 \ 
 
 - 
 
 600 
 
 - / 
 
 \ 
 
 - 
 
 300 
 
 y 
 
 \...^ 
 
 - 
 
 
 1 1 1 1 
 
 1 1 
 
 0.25 0.50 0.75 1.00 1.25 ISO 1.75 2.00 
 
 Text No. 30902 
 
 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 
 
 VERTICAL DISPLACEMENT, in 
 
 FIGURE 2. - Solid crib-vertical load versus 
 displacement (DOE tests). 
 
 FIGURE 3, • Open crib failure mode (DOE tests). 
 
 ANALOG DATA 
 
 Test No. 30903 
 
 0.50 0.75 1.00 1.25 1.50 
 
 VERTICAL DISPLACEMENT, in 
 
 FIGURE 4. - Open crib-vertical load versus 
 displacement (DOE tests). 
 
 The resulting mass production technique 
 reduced the cost per unit to $2.29 FOB 
 plant. The block is presently being man- 
 ufactured under licensed agreement with 
 Burrell at locations in Pennsylvania, 
 Ohio, Virginia, Utah, and Alabama. Other 
 block producers are beginning to enter 
 the market with similar products. 
 
149 
 
 UNDERGROUND EXPERIENCE 
 
 Since the original installations in 
 1976, at Kaiser's Sunnyside Mine, the 
 Bureau has been involved in cooperative 
 underground evaluations at several other 
 locations: 
 
 Kaiser Steel, York Canyon, NM 
 
 Price River Coal, Helper, UT 
 
 Snowmass Coal, Carbondale, CO 
 
 Bethlehem #131, Van, WV 
 
 U.S. Steel #9, Gary WV 
 
 In all these cases, the cribs were used 
 in the tailgate section of the entries on 
 longwall panels (fig. 5). Because of the 
 successes of the initial installations, 
 some of the above are expanding the use 
 of SFC to areas where permanent support 
 
 systems are required, such as ventilation 
 entries, and to support the base of ven- 
 tilation shafts. 
 
 Several other coal companies have 
 elected to institute the use of this 
 product in longwall entries, long- 
 life main entries, bleeder entries, and 
 stoppings: 
 
 Eastern Associated Coal Company — WV 
 
 Emery Mining Company — UT 
 
 Trail Mountain Mining Company — UT 
 
 Barnes & Tucker #25 — PA 
 
 Penn Allegheny Coal — PA 
 
 Canterbury Coal — PA 
 
 FIGURE 5. - Concrete cribs supporting tailgate after passage of longwall face. Kaiser's York 
 Canyon, NM, mine. 
 
150 
 
 Carpentertown Coal & Coke — PA 
 Scott's Branch Mine — KY 
 Martin County Coal — KY 
 Texas Gulf, Inc. — WY 
 Westmoreland Coal Company — VA 
 Helvetia Coal Company — PA 
 ARMCO Inc.~WV 
 Consolidation Coal Co. — WV & PA 
 
 Island Creek Coal Co. — VA 
 
 Jim Walters Resources Inc. — AL 
 
 Kit Energy — WV 
 
 In almost all applications of the tail- 
 gate entries for longwall, the SFC has 
 not only provided added safety, better 
 airflow, and more area for movement, but 
 has also proven cost-effective. For 
 example, in a mine using a double row of 
 wood cribs set on 5-f t centers , they are 
 now installing SFC cribs in a single row 
 on 7-ft centers (figs. 6-7). 
 
 FIGURE 6. - Longwall tailgate entry with double row of wood cribbing. West Virginia mine. 
 
151 
 
 FIGURE 7c, - Longwall tailgate entry with single row of concrete cribbingo West Virginia mine,, 
 
 COST ANALYSIS 
 
 Following is an example of a cost com- 
 parison for this West Virginia mine: 
 
 FIBERCRIB versus Wooden Cribs 
 
 (Based on 84-in height — 1,000-ft 
 advancement 
 
 Wooden Cribs - (6 by 8 by 30 in) — Double 
 Row on 5-ft centers — 
 28 blocks per crib 
 @ $2.39 FOB Mine = $66.92 
 per crib 
 
 400 Cribs (g $66.92 = $26.768.00 
 
 FIBERCRIB Cribs 
 
 (3-5/8 by 7-5/8 by 
 23 in) —Single Row 
 on 7-ft centers — 
 46 blocks per crib 
 (a $2.88 FOB Mine 
 = $132.48 per crib 
 
 142 Cribs @ $132.48 = $18,812.00 
 
 Labor Costs - The labor cost to build 
 each crib is approximately equal. Note 
 that 400 wood cribs were required in this 
 example, as opposed to 142 FIBERCRIB 
 cribs, resulting in a labor cost saving 
 of over 60 pet. The total resulting cost 
 saving in this example is over 40 pet for 
 the FIBERCRIBS. 
 
 In another instance, where the mine was 
 setting a single row of wood cribs skin- 
 to-skin, they have found the use of SFC 
 cribs on 6-ft centers to be cost- 
 effective. 
 
 In an Alabama mine, a longwall tail- 
 gate, originally supported with a single 
 row of wooden cribbing 9 ft on center, is 
 being successfully supported with a 
 single row of SFC cribs 15 ft on center. 
 This will result in a total savings of 
 over 55 pet for labor and materials ex- 
 pended for support of this tailgate 
 (figs. 8-9). 
 
152 
 
 FIGURE 8. - Tailgate supported with concrete cribbing ahead of longwall face. Alabama mine. 
 
 Many of these mines have also found the 
 SFC cribs to be efficient in areas where 
 longevity is a factor and, although the 
 cost may be initially higher in some 
 cases, the longer life of the SFC crib- 
 bing will make them more economical in 
 the long run. The industry has found 
 that in the case of long-life entries, 
 the SFC cribbing is an advantage because 
 of the inherent qualities of stiffness, 
 strength, nonshrinking, nonrotting, and 
 nonflammability. Replacement costs on 
 wood cribbing in these applications have 
 proven to be very high. In any area 
 where a wood crib would need replacing 
 because of rot and/or failure, the SFC 
 block would eventually prove to be more 
 economical. 
 
 In most applications it is possible to 
 use fewer SFC cribs than wood cribs in 
 order to achieve the same support. A 
 Pennsylvania mine, with their belt and 
 track lines running parallel, was using 
 wood posts skin-to-skin and, in some 
 cases, two to three deep to support the 
 roof. They are now building solid SFC 
 cribs on 6-ft centers with an "I" beam 
 from crib to crib. This has given them 
 much greater support and they now have 
 access between beltline and track. This 
 particular crib installation was in- 
 spected by representatives of the Office 
 of Deep Mine Safety, Pennsylvania Depart- 
 ment of Environmental Resources. The two 
 inspectors recommended, "An approval be 
 granted to Burrell Construction and 
 
153 
 
 fT^ 
 
 ^»*i.^*»" 
 
 FIGURE 9, = Tailgate supported with concrete cribbing after passage of longwall face.. Alabama mine. 
 
 Supply Company for use of FIBERCRIB in 
 bituminous underground coal mines in 
 Pennsylvania, providing FIBERCRIB blocks 
 strictly comply with the manufacturer's 
 specifications" (fig. 10). 
 
 To date, a combined total of about 20 
 miles of longwall tailgate entry, main 
 entry, and bleeder entry is being 
 
 supported with SFC cribbing in the major 
 U.S. coal mining regions. SFC cribbing 
 has also been used to build stoppings and 
 overcasts where normal cinder block had 
 proved ineffective due to crushing. The 
 SFC block is approximately 2-1/2 times 
 stronger than a regular cinder block and 
 is not subject to brittle failure. 
 
154 
 
 FIGURE lOo " Solid concrete cribs placed 6 ft on center supporting a main haulage entryo 
 Pennsylvania mine. 
 
 INSTALLATION PROCEDURES 
 
 The procedure for installation of SFC 
 crib blocks, suggested by the Bureau, is 
 as follows: 
 
 CONCRETE CRIBS (FIBERCRIB) 
 
 1. Prepare floor area level and flat, 
 large enough for 23- by 23-in, 15- by 23- 
 in, or 15- by 15-in crib, as required. 
 
 2. Place first layer of crib blocks 
 and check for level with hand level. 
 
 3. Stack block in open or solid con- 
 figuration according to plan. 
 
 4. While stacking, keep blocks 
 straight, square, and plumb and remove 
 dirt, etc., from each layer before plac- 
 ing next layer . 
 
 5. Top of crib must be finished with 
 wood plank or beam and wedged tightly 
 
 with wood wedges. The total thickness of 
 wood including wedges must be a minimum 
 of 1 in for each foot of crib height (4 
 in for 4-ft crib, 6 in for 6-ft crib, 
 etc.) (fig. 11). 
 
 CONCRETE STOPPINGS USING 
 FIBERCRIB BLOCKS 
 
 1. Follow above procedure except ex- 
 cavate floor for stopping dimensions. 
 
 2. Stack block flat and in overlapping 
 fashion as brick is typically laid. 
 
 3. Finish top of stopping with the 
 least amount of wood wedging possible or 
 fill with mortar mix. 
 
 4. Seal one side of stopping with 
 suitable mortar or sealant. 
 
155 
 
 FIGURE 11. - Example of insufficient wood at top of crib. Point loading. 
 
 SUMMARY AND CONCLUSIONS 
 
 There is no question of the strength 
 and longevity of the steel-fiber- 
 reinforced concrete cribbing and its suc- 
 cess in present installations. Although 
 wood has been the traditional material in 
 use for roof support in the mining in- 
 dustry, the problems of supply and cost 
 are increasing rapidly to the point where 
 a viable alternative must be found. Con- 
 crete products are readily available in 
 all parts of the country, and, indeed, 
 the world. The steel-fiber-reinforced 
 concrete cribs are proving to be the 
 answer to many roof control problems 
 
 experienced in the past and are 
 economical alternative to wood. 
 
 the most 
 
 One industry source in the field of 
 mine safety asserted: "The new cribs are 
 much safer due to the inherent character- 
 istics of steel reinforced concrete. We 
 have had no failures, and we have long- 
 term installations in which we have con- 
 fidence that these cribs are a reliable 
 source of roof support." Some mines have 
 already decided that, in their opera- 
 tions, wood cribbing is a thing of the 
 past. 
 
 INT.-BU.OF MINES, PGH., PA. 27440 
 
H 23b 84 
 
-C^ -00*" ^ 
 
 ■r •^ 
 
 
 * tCvn^v-t**^.. '^ 
 
 5>.' •'^^. -«^ ^v-^a-. "<»^^^^^ ^s^^. '^^McS' ^^-K^ar- —^ov^ <.'!^&'- ''^'^0^ f^^K" ^ov^ 
 
 'oK 
 
 .o^..»-%>^ \*^\!;;z^',\. .c^/^;!^"'" 
 
 ■^^0^ 
 
 
 \^^^' "v*^"-^'/ V**^\/ V*"^"'!/ ,.\^^^'*"/' '°* *"-" ♦^ 
 
 ^^--.^ 
 
 
 ^ ^^"X ^. 
 
 ./"%. 
 
 
 
 
 A^-^ 
 # %. o- 
 
 
 
 *bv^ 
 
 
 
 -^^d^ 
 
 • ,sS^^v<'.. o 
 
 .■i ymi/;^. --^ ^- .^^N^-. -^^^ -^m^^^ ^^>.^ 
 
 
 
 *bV" 
 
 / -o^^^5*/ <V^*^\^'' "-^^^'^^"'V 
 
 %^0^ 
 
 o o > 
 
 
 5" »!f(\%^/)u" "^^ .^'V' 
 
 
 1.° •<^r. .& »l 
 
 0* »i,!oL'* <?■ 
 
 ^ *»«(»» ^ 
 
 
 
 
 "-.s"^' 
 
 .^ ^^^'^^ "me:*^ .,<5.^^"^^, ». 
 
 
 ,** 
 
 
 
 
 
 
 
 »f* ^ <^^ *' 
 
 
 
 
 -'- "-".o^-.. V'"- 
 
 <j>, »e«o' .1,' 
 
 
 
 V.<^^ 
 
 
 
 
 
 
 
 
 
 ^^'% ^; 
 
 > .0 
 
 ^°'%.. 
 
 4 V c B « 
 
 <?^. o « < 
 
 
 
 
 •^ 
 
 
 ^°^^^ ' - 
 
 "^J**^'* ^v^' 
 
 4. »^ 
 
 a; 4^ 
 
 
 .0 
 
 "u-";.^ 
 .i"^.. 
 
 
 
 ^' ^^ -^^.WC^^/ \/°^!^,v.^^' '^^.-«l^_r.^ rv >. 
 
 
 
 ;\ 
 
 J' \. ^-^ 
 
 5^' ^> <« o" 
 
 
 ^^* 
 
 
 
 
 
 
 •^^o^ 
 
 
 IWII 
 
 
 *bV 
 
 «V *-'^- 
 
 
 
 o A>^ 
 
 
 
 
 -1 r\^ 
 
 
 
 "^ t? » K^ <, ■ay >t» • ™,„-^ , 
 
 
 
 
 ,* 4 o 
 
 ^r?^ 
 
 >■ 
 
 
 \/ « 
 
 
 
 *^».^*' /JIK- ■ V*^ 
 
 V^^ 
 
 ,\, /.^v'^°o .^^\^j^:^%\ /.^i^'^^o /\^-;^%\. /.^j^^>A 
 
 :. '^^.^ ov^^gi'"^ -^-ov^" /^^.'- '-^^o^' »<.^»'". '^bv*' :^^'' "^--0^ i' 
 
 
 >7 J-' -*• 
 
 
 
 ^^-^K. 
 
 "^^^ »o 
 
 
 rr^*.^-^ "°w. *-^'\^o 
 
 ^ 
 
 ^ /.aWa^o -n*,^^,-^ ;,fl^-. \^^^ ^'Mik" "-/ ='am'^ ^-"^^ -' 
 
 o°^^a;;.'>- j'^\^^:^^\. .co^.^;:>o ^,,-^\.^, 
 
 'bt 
 
 
 ./%. 
 
 a5>"^ 
 
 ^ .>•' 
 
 
 
 -^oi^ 
 
 '■^^0^ 
 
 'bv 
 
 
 'bV 
 
 ^^c 
 
'"% ^-^ 
 
 <^ '■"' ^^ 
 
 '^'^o-^'""'..-^*^.^!^;^'. "^^ <P' .°-^% '°o .,-2^"" ..c:.^% ^-t 
 
 
 ^a^^ :^^- '^"^^^^ oV-^a--" V^C^" Vi5^.'. ''-t^X^ 
 
 o " " . *0 
 
 ':^o^ 
 
 #y VW.o'^ VW./ 'o./W-.o' \;*-\/ v^y„ 
 
 v.^ 
 
 ^^*^-. 
 
 • y ^f "J^ 
 
 
 
 ■>""* *o 
 
 
 .o .^''^ 
 
 iP-r.. 
 
 
 o . * "■ - G^ v5 
 
 s> » 
 
 ,*' .>«^-. >^ .^•.>,v^-..X>*-'/^-,%.,*^\.;&aX,/'/,^-,--.,,*v.-^ 
 
 
 
 
 
 
 
 .t* ,& 
 
 
 ^^ ^ '*7rr*' A 
 
 
 
 ^" ^^'% ..' 
 
 v^' 
 
 "•\'>o '">\.1^,'\ 
 
 
 
 
 
 
 4^ .*^ 
 
 "''e^o* 
 
 „'' o 
 
 
 
 
 
 .0 if. V 
 
 
 h4 
 
 ,0^ %, *75^T^^ A 
 
 ^. 
 
 
 
 
 te. .^^ ■ »>^^A-^ "%. ^^ 
 
 
 
 '^9 
 
 
 <3. *^T« A 
 
 V ^* ". 
 
 
 
 '^ ■.'<' /j^l 
 
 
 
 <. "o 5 d « .G^ 
 
 ^ *-'V^5* A 
 
 
 
 •0" 
 
 jP-7j ..i 
 
 5^ V>. -J 
 
 
 — r. o 
 
 ./ ^o.^^f^*/ v^^!^\/ \.^^'^'/ v.'*^\/ °^ ^°' / 
 
 V 
 
 
 
 ^oV 
 
 *by 
 
 
 
 
 
 
 ♦^"V - 
 
 »> ^ ". 
 
 
 V.^'^" 
 
 
 
 
 
 <^, '"■"" a"- „„ -^ 
 
 
 
 iP -n*. t 
 
 
 
 ««»-^\<!>^" "q.. *^'^'\o'^ 
 
 
 "^O. 
 
 ip .*iJ 
 
 HECKMAN 
 
 BINDERY INC. 
 
 ^^ NOV 84 
 
 ^fey N. MANCHESTER, 
 ^=^ INDIANA 46962 
 
 :j^%^ 
 
 -• \,** •• 
 
 i» or ^' » 
 <5' % 
 
 
 
 x^vl,. 
 
 ^0 %^ 
 
 
 <\ 'o . » * , G^ '^ *;^T' A rf . 'o 
 
 
 
 
 
 
 
 ,G^ '^^^ '*.^T-' A .^ 
 
 6^ ^."-^.'^o, ^^-^^ ..-J., ^^ 
 
 ■ * 'V' '<^ •^.^ 
 
 <7i * V- 0* 
 
 ^oV 
 
 v^o^