UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Cable support guidelines for underground hard rock mine operations Nickson, Simon D. 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1993_spring_nickson_simon.pdf [ 11.53MB ]
Metadata
JSON: 831-1.0081080.json
JSON-LD: 831-1.0081080-ld.json
RDF/XML (Pretty): 831-1.0081080-rdf.xml
RDF/JSON: 831-1.0081080-rdf.json
Turtle: 831-1.0081080-turtle.txt
N-Triples: 831-1.0081080-rdf-ntriples.txt
Original Record: 831-1.0081080-source.json
Full Text
831-1.0081080-fulltext.txt
Citation
831-1.0081080.ris

Full Text

CABLE SUPPORT GUIDELINESFORUNDERGROUND HARD ROCK MINE OPERATIONSbySIMON DAVID NICKSONB.Eng., McGill University, 1983A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE STUDIESMining and Mineral Process EngineeringWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIADECEMBER 1992©. Simon David Nickson, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of  Mining and Mineral Process EngineeringThe University of British ColumbiaVancouver, CanadaDate December 11, 1992DE-6 (2/88)11ABSTRACTThe objective of this thesis is to expand upon the existing database of cable supportpractice, and develop revised design criteria for the support of underground openings. A newempirical database of 46 supported case histories was assembled from a six month field study,involving visits to 13 mines in Western Canada, the United States, and Ireland. A comprehensivereview of current cable support theory, practice and design is presented. A statisticalmethodology is introduced to define zones of stability from an empirically collected database.Existing guidelines for the design of underground support are reviewed and calibrated. Currentdesign criteria for cable support are often based on an even distribution of bolts over thesupported surface. The point anchor approach involves the use of a high concentration of cables,installed into large open stope hangingwalls from sublevel access drifts. This thesis proposesrevised cable design guidelines for even distributions in open stope backs, and develops anapproach for the point anchor design of hangingwall surfaces.iiiTABLE OF CONTENTSABSTRACT^ iiTABLE OF CONTENTS^ iiiLIST OF TABLES viiiLIST OF FIGURES^ xACKNOWLEDGEMENTS xivDEDICATION^ xvCHAPTER 1: INTRODUCTION^ 11.1 GENERAL^ 11.2 OBJECTIVE 11.3 INTRODUCTION^ 21.4 EMPIRICAL DESIGN 3CHAPTER 2: CABLE SUPPORT PRINCIPLES^ 52.1 INTRODUCTION^ 52.2 INSTALLATION 62.3 FAILURE MODES^ 82.4 LABORATORY TESTING 82.4.1 Commonwealth Scientific & Industrial ResearchOrganization Testing^ 82.4.2 United States Bureau of Mines Testing^ 122.5 CRITICAL BOND LENGTH^ 232.6 GROUT^ 242.7 CABLE GEOMETRY^ 242.8 CONFINEMENT 27iv2.9 STRESS^ 282.10 CABLE ORIENTATION^ 282.11 SUPPORT STIFFNESS 302.12 DISCUSSION^ 31CHAPTER 3: CABLE SUPPORT PRACTICE^ 333.1 INTRODUCTION^ 333.2 GROUT^ 333.3 GROUT PUMPS^ 373.4 PLATES AND STRAPPING^ 393.5 CABLE GEOMETRY 403.6 CABLE ANCHORS^ 413.7 INSTALLATION PROCEDURE^ 423.8 DESIGN LAYOUTS^ 453.9 CABLE PATTERNS 483.9.1 Back Support^ 483.9.2 Hangingwall Support 503.9.3 Other Support Patterns^ 523.10 HEALTH AND SAFETY ASPECTS OF CABLE BOLTING^523.10.1 Cable Pushing^ 523.10.2 Cable Anchoring 553.10.3 Effect of Working with Cement^ 553.10.4 Handling Cable Bolts^ 563.10.5 Grout Pressure 57V3.10.6 Manpower^ 573.11 COSTS^ 583.12 DISCUSSION 58CHAPTER 4: CABLE BOLT DESIGN METHODS^ 614.1 INTRODUCTION^ 614.2 DISCRETE ANALYSIS 634.3 COLLECTIVE ANALYSIS^ 654.3.1 Dead Weight Design 664.3.2 Rock Classification^ 684.3.3 Beam Theory 744.3.4 Past Experience^ 764.3.5 The Mathews Method and Bolt Factor^ 784.3.6 Design Based on Rock Mass Stiffness 824.3.7 The Potvin Method^ 844.4 WESTERN CANADIAN PRACTICE^ 86CHAPTER 5: THE POTVIN METHOD 875.1 INTRODUCTION^ 875.2 THE MODIFIED STABILITY GRAPH^ 875.2.1 The Block Size Factor 925.2.2 The Stress Factor^ 935.2.3 The Joint Orientation Factor^ 955.2.4 The Gravity Factor^ 95vi5.2.5 The Unsupported Modified Stability Graph^ 975.2.6 The Supported Modified Stability Graph 985.3 DESIGN CHART FOR CABLE BOLT DENSITY^ 1025.4 DESIGN CHART FOR CABLE BOLT LENGTH 1065.5 DISCUSSION^ 106CHAPTER 6: DATABASE 1096.1 INTRODUCTION^ 1096.2 DATA COLLECTION 1106.3 DATABASE^ 1106.3.1 Comparison with the Modified Stability Graph Design Ranges^1166.3.2 Comparison with the Design Chart for Cable Bolt DensityDesign Ranges^ 1206.4 STATISTICAL METHODOLOGY^ 1206.4.1 Linear Regression Analysis 1206.4.2 Discriminant Analysis^ 1276.5 STATISTICAL ANALYSIS 1326.5.1 Unsupported Database (Case #1)^ 1396.5.2 Supported Database (Case #2) 1416.5.3 Back Cable Support Database (Case #3)^ 1446.6 CONCLUSIONS^ 155CHAPTER 7: CABLE BOLT SUPPORT GUIDELINES^ 1627.1 INTRODUCTION^ 1627.2 CABLE BOLT DESIGN METHODOLOGY^ 162vii7.3 PATTERN APPROACH TO CABLE DESIGN^ 1667.3.1 Cable Bolt Density for Back Support 1667.3.2 Cable Bolt Length for Back Support^ 1717.3.3 Other Support Patterns^ 1737.4 POINT ANCHOR APPROACH TO CABLE DESIGN^1757.4.1 Hangingwall Cable Support^ 1767.4.2 Design Chart for Point Anchor Hangingwall Cable Support^1807.4.3 Cable Bolt Length for Point Anchor Hangingwall Support^1867.5 DESIGN CASE HISTORIES^ 1867.5.1 Detour Lake Mine 1877.5.2 Wilroy Mine^ 1877.6 DISCUSSION 189CHAPTER 8: CONCLUSIONS^ 1928.1 INTRODUCTION 1928.2 CONCLUSIONS^ 1928.3 FUTURE WORK 1958.4 FINAL REMARKS^ 197BIBLIOGRAPHY^ 198APPENDIX A - DATABASE SUMMARY^ 204APPENDIX B - BOLT DENSITY CONVERSION CHART^ 215APPENDIX C - SUMMARY OF INSTALLATION PROCEDURES^218TableLIST OF TABLESvii'Page2.1 Summary of 28-day USBM Test Results 132.2 Comparison of Laboratory Test Results 323.1 Case Study Support Practice 343.2 Cable Bolt Installation Procedure Summary 353.3 Grout Characteristics (after Hyett, Bawden, and Coulson 1992) 363.4 Cable Bolt Grouting Procedure (after Cluett 1991) 443.5 Summary of Cable Patterns in Western Canadian Practice 483.6 Summary of Cable Component Costs (1992 Canadian Dollars) 594.1 Description of Q-system Parameters (Barton, Lien, and Lunde 1974) 694.2 Design Methods in Western Canadian Practice 865.1 Summary of Potvin Unsupported Main Database 885.2 Summary of Potvin Unsupported Complementary Database 905.3 Summary of Potvin Supported Database 1006.1 Summary of Unsupported Database 1126.2 Summary of Supported Database 1136.3 Summary of Average Cable Density and Length 1156.4 Hydraulic Radius Comparison 1196.5 Values of r for Different Levels of Significance 1266.6 Cases of Back Support in the Supportable Region of theRevised Modified Stability Graph - Combined Database 1496.7 Possible Linear Relationships for Cable Bolt Density 1517.1 Percentage of Supported Cases in the Revised ModifiedStability Graph Design Zones 165ix7.2 Relationship Between Cable Length and Span for Back Support^1717.3 Point Anchor Hangingwall Supported Database^ 181A.1 Case Study Summary (Case 1 to 6)^ 205A.2 Case Study Summary (Case 7 to 12) 206A.3 Case Study Summary (Case 13 to 18)^ 207A.4 Case Study Summary (Case 19 to 24) 208A.5 Case Study Summary (Case 25 to 30)^ 209A.6 Case Study Summary (Case 31 to 36) 210A.7 Case Study Summary (Case 37 to 42)^ 211A.8 Case Study Summary (Case 43 to 48) 212A.9 Case Study Summary (Case 49 to 54)^ 213A.10 Case Study Summary (Case 55 to 59) 214B.1 Bolt Density Conversion Chart^ 216LIST OF FIGURESFigurexPage2.1 Cable bolt installation methods 72.2 Cable bolt failure modes 92.3 Typical single wire load-displacement curve (After Fuller and Cox 1975) 112.4 Typical strand load-displacement curve (After Fuller and Cox 1975) 112.5 USBM pull test apparatus (After Goris 1990) 142.6 Averaged 7, 14, and 28 day load-displacement curves for test series 1(After Goris 1990) 152.7 The effect of pulling on a grouted cable (After Noranda TechnologyCenter 1990) 172.8 Maximum load versus embedment length (After Goris 1990) 182.9 Averaged 28-day load-displacement curves (After Goris 1990) 182.10 The effect of embedment length 192.11 Load-averaged displacement curves for double cables (After Goris 1990) 222.12 Grout unconfined compressive strength versus water:cement ratio(After Reichert, Bawden, and Hyett 1992) 252.13 Comparison of U.C.S. versus water:cement ratio for different mixingmethods (After Gendron et al. 1992) 262.14 Cable bolt capacity vs embedment length at 0.3 w:c(After Reichert, Bawden, and Hyett 1992) 293.1 Sample cable bolt layout sheet 473.2 Typical back support for open stope and cut and fill mining 493.3 Hangingwall cable support patterns (After Fuller 1983b) 513.4 Sling approach to open stope support 534.1 Cable design flowchart 62xi4.2 Sliding block discrete analysis (after Hoek and Brown 1980) 644.3 Tunnel Support Chart (After Barton, Lien, and Lunde 1974) 734.4 Beam approach to hangingwall design (After Fuller 1983b) 754.5 Stability graph proposed by Mathews et al. (1981) 804.6 Factors A, B, C proposed by Mathews et al. (1981) 814.7 Bolt factor design chart (After Bawden et al. 1989) 834.8 Relationship between radial stiffness, embedment length, and water:cementratio required to mobilize 24 tonnes of cable load carrying capacity(After Reichert, Bawden, and Hyett 1992) 855.1 Factors A, B, C proposed by Potvin (1988) 945.2 Critical joint concept (After Potvin 1988) 965.3 Unsupported Modified Stability Graph (After Potvin 1988) 995.4 Supported Modified Stability Graph (After Potvin 1988) 1035.5 Design Chart for Cable Bolt Density (After Potvin 1988) 1045.6 Revised Design Chart for Cable Bolt Density (After Potvin and Milne 1992) 1055.7 Design Chart for Cable Bolt Length (After Potvin, Hudyma, and Miller 1989) 1076.1 Unsupported case histories compared to the design ranges proposed byPotvin (1988) 1176.2 Supported case histories compared to the design ranges proposed byPotvin (1988) 1186.3 Cable bolted backs compared to the design ranges proposed by Potvin (1988) 1216.4 Concepts of linear regression analysis 1226.5 Methods of separating two classes of data 1306.6 Dot plots for the combined unsupported database (Case #1) 1346.7 Dot plots for the combined unsupported database (Case #1 - transformed data) 136xii6.8 Probability plots for stable cases from the combined unsupporteddatabase (Case #1) 1376.9 Probability plots for caved cases from the combined unsupporteddatabase (Case #1) 1386.10 Statistical analysis of stable and caved combined unsupporteddatabase (Case #1) 1406.11 Statistical analysis of stable and caved combined supporteddatabase (Case #2) 1426.12 Proposed modifications to the Modified Stability Graph 1436.13 Statistical analysis of combined cable bolted back database (Case #3) 1456.14 Possible transition zone for the Design Chart for Cable Bolt Density 1476.15 Minimum cable density for design regions of the revised ModifiedStability Graph 1486.16 Regression line for cable bolt density versus RQD/Jn/HR 1506.17 Regression analysis with the relative block size factor and thestress factor 1536.18 Regression analysis with the relative block size factor and thestress factor 1546.19 Regression analysis with Q' 1566.20 Regression analysis with Q' 1576.21 Regression analysis with Q' 1586.22 Regression analysis with N' 1596.23 Cable support design ranges for the revised Modified Stability Graph 1607.1 Proposed design methodology for cable bolt support 1637.2 Proposed revisions to the Modified Stability Graph design regions 1647.3 Design Chart for Back Cable Support 167xiii7.4 Proposed back cable support design ranges on the revised ModifiedStability Graph 1697.5 Proposed minimum bolt density design ranges for back cable support 1707.6 Cable bolt length for back support 1727.7 Relation between potential failure zone and cable length 1747.8 Wilroy case history of the point anchor approach to hangingwall support 1777.9 N663 stope case history of point anchor hangingwall cable support 1797.10 Description of geometry for point anchor approach to cable support design 1827.11 Design Chart for Point Anchor Hangingwall Cable Support 1837.12 Point anchor hangingwall database compared to the revised ModifiedStability Graph 1857.13 Detour Lake Mine case history of point anchor hangingwall cable support 188xivACKNOWLEDGEMENTSThe author greatly appreciates the assistance of all who have contributed to this research.In particular, the support received from Noranda Inc., Inco Ltd. (Manitoba Division), and thethirteen mines from which the database was collected, is gratefully acknowledged. Many usefulideas and observations came directly from operating and technical personnel from each minesite.Numerous friends and colleagues assisted during the data collection phase of this project. Specialthanks is extended to Nan Lee, Dave Neuburger, Lester and Jacquie Jordan, Matt Sutcliffe,Philip Ng, Al Meston, Rick and Val Sawyer, Tony Reshke, Jan Romanowski, John and CannelKelly, Brian Keady, Charlie and Yvette Brown, Mathew and Rita Chand, John Goris, JimWilson, Jim Spencer, Mark Brown, Bill Steenson, Brent Lamore, Chris Kushnir, JozefSzymanski, Mike Cullen, Dave Nicholls, Chantale Doucet, Steve Davies, Bob Peddigrew, BillGreeley, Paul Pawliw, Nick Maniel, Mark Laueranti, George Greer, Hadyn Figueroa, GeorgeNagy, Marty Klotz, Hiram Rogers, Keith Patterson, Jim Ashcroft, Gary Allen, Bruce Tokle,Ken Fedak, Jim Bell, Mel Sasek and Carl Sauer. Appreciation is extended to the staff of theMining and Mineral Processing Department at the University of British Columbia for theirassistance during the term of this research. Dr. Rimas Pakalnis, Doug Milne, and Dr. YvesPotvin offered valuable advice and consultation throughout the project. The author is alsograteful for the assistance provided by the review committee, Dr. Rimas Pakalnis, Andy Mular,Chuck Brawner, and John Goris.To Michael Anthony NicksonXV1CHAPTER 1INTRODUCTION1.1 GENERALThe use of cable bolts in underground mines has evolved from the need for a long,flexible, high capacity support system, that could be installed in advance of mining The designof a cable bolt system is often a subjective process. The mining engineer can design a cablepattern based on dead weight, but lacks guidelines for cases where dead weight considerationsmay not be cost efficient. Potvin (1988) proposed a method for cable bolt design that was basedon the empirical analysis of back support derived from sixty-six case histories of Canadiansupport practice. This thesis reviews the method proposed by Potvin (1988) and develops arevised design methodology for cable bolt support.1.2 OBJECTIVEThe objective of this thesis is to expand upon the existing database of cable supportpractice and develop revised design criteria for the support of underground openings. Six monthswere spent in the field during the data gathering stage of this study, and visits were made tothirteen mines to review cable support practice. The majority of the mines were located inwestern Canada, but the application of cable support was also observed in the western UnitedStates and Ireland. A total of 13 unsupported and 46 supported case histories have been collectedduring this study. The terms "supported" and "unsupported" refer to the use of cable support.A supported case history incorporates cable bolts, while an unsupported case history does not.2An unsupported case history may however incorporate short pattern bolting that is installed toprovide primary support during the development phase. Stope backs usually incorporate shortpattern bolting based on operational or legal standards. Hangingwalls are occasionally not boltedduring the development phase due to limited surface exposure. The design methodology proposedby Potvin (1988) was directed at the cable support of open stope backs. Cable bolts are alsofrequently used to support open stope hangingwalls, although limited design techniques arecurrently available. In this study, particular emphasis was placed on collecting a database ofhangingwall cable support that would aid in the interpretation of related design criteria.1.3 INTRODUCTIONThe first half of this thesis reviews the current state of cable support theory, practice anddesign. The discussions are based on a literature review that has been complemented byobservations from the field study phase of this project. A review of cable support practice ispresented in Chapter 3, and has been designed as a guide for the development of operationalprocedures. Chapter 4 presents a review of current cable design methods, based on observedpractice and techniques described in literature. The design methodology proposed by Potvin(1988) is reviewed in Chapter 5, and introduces the development of revised cable design criteriain the second section of this thesis. The database assembled in this study is presented andcompared to the Potvin (1988) design proposals in Chapter 6. A statistical approach is used todevelop relevant relationships for use in the design of cable support. Chapter 7 introduces reviseddesign guidelines for back support and proposes a new technique for use in hangingwall cabledesign.31.4 EMPIRICAL DESIGNEmpirical analysis provides a tool for the mining engineer to use in the design ofunderground openings and support. Empirical is defined as "Relying or based on practicalexperience without reference to scientific principles" (Webster's New World Dictionary ofAmerican English 1991). Scientific principles cannot be ignored in proper empirical analysis andideally are used to complement results suggested through observation and experience. Empiricalcharts or graphs developed from observational data are frequently used for the design ofunderground openings and associated rock support. Several examples of such charts (Barton,Lien, and Lunde 1974; Mathews et al. 1981; Potvin 1988) will be reviewed later in this thesis,and can be a valuable asset to the mining engineer if used with proper care. It is recommendedthat the original published document be reviewed prior to the utilization of a particular designchart or graph. This permits a full understanding of the original objectives associated with thechart, and provides a means of assessing its applicability. For example, the Design Chart forCable Bolt Density produced by Potvin (1988), relates cable bolt density to a relative block sizefactor. This chart was developed from a database that is made up of cable bolted backs, and wasnot intended for use in hangingwall design.Charts can be custom built or modified to suit particular characteristics of one operation.Greer (1989) reviewed both the Mathews (1981) and Potvin (1988) empirical design approaches,to determine which displayed the best correlation with a database assembled from the ManitobaDivision of Inco Ltd. Based on this data, the Mathews (1981) approach was suggested as the bestguideline for hangingwall design, but back design was better suited to the Potvin (1988)methodology. Greer (1989) recommended that the database be expanded as mining progressedto allow for further calibration of design criteria. This is a good illustration of the modification4of empirical design methods to reflect on site operating experience. Cullen (1991) discusses asite specific rock classification system developed for use at the H-W mine of Westmin ResourcesLtd. The system involved a quantitative assessment of drill core quality, degree of schistosity,rock hardness and total gouge to produce a qualitative description of the rock mass that variedfrom very poor to very good ground. The classification was used as an indicator of groundconditions on geological sections, and related to planned excavations. Hoek and Brown (1980,131) discuss the transition from intact rock material to a heavily jointed rock mass. Laboratorytesting can be applied to determine the properties of intact rock but in terms of stope design, itis necessary to consider the rock mass. The empirical design approaches that are discussed in thisthesis attempt to quantify rock mass conditions for practical use in the design of undergroundopenings.5CHAPTER 2CABLE SUPPORT PRINCIPLES2.1 INTRODUCTIONCable bolts as a means of rock support were first introduced to the mining industry in the1960's. Gramoli (1975) describes the installation of discarded locked coil hoisting rope as asupport system adopted by the Geco Mine in 1963. This is one of the earliest references to cablebolting on a large scale, but Garcia (1929) suggests that the installation of cable slings to supportthe back of a coal mine was a convenient way of utilizing scrap hoist rope. The initial applicationof cable support often involved the use of discarded hoist rope that was installed in a borehole,tensioned and subsequently grouted. Due to its flexibility, a long cable could be installed intoa drill hole from a drift that had limited height. This was an advantage over traditional methodsof support and quickly lead to widespread use of the cable bolt in the 1970's. Windsor (1992)describes the progression of cable bolting from the use of discarded hoist rope to pre-stressingwire and eventually to a pre-stressing strand. Hoist rope required extensive cleaning to removegrease and was abandoned in favour of a seven wire steel strand. The cable strand is made upof one central wire surrounded by six slightly smaller outside wires, and typically has a diameterof 15.2 to 15.9 mm (0.6" to .625"). The cable wires are made from high-strength steel with amodulus of elasticity of approximately 203.4 GPa (29.5 x 106 psi) and an ultimate strength ofapproximately 26.3 tonnes or 58,000 lbs (Goris 1990). There is a small variation in reportedcable breaking strengths noted in literature due to different steel specifications and frequentconversion between metric and imperial units. In this thesis, the ultimate strength of a cable boltis considered to be approximately 26.3 tonnes, but will be subsequently noted as 26 tonnes.6Cable breaking strength is not commonly attained in the laboratory and reference is often madeto the pull out strength or load carrying capacity of the cable. These terms refer to loads at whichthe cable fails or pulls out of a grout column, and can be well below the steel breaking strength.Fuller (1983a) notes that steel failure was not common in a review of Australian supported openstopes. Gendron et al. (1992) indicate that cable bolts are typically left dangling from the backin failed cases observed at several Canadian mining operations. These comments suggest that thefull use of steel breaking strength is not reflected in current cable bolt practice. The objectiveof cable bolt design should be to maximize the load carrying capacity of the cable so that itapproaches the breaking strength. It is important however to recognize the distinction betweenbreaking strength and pull out load in the design of cable support. This chapter discusses thefactors involved in the determination of cable load carrying capacity.2.2 INSTALLATIONA discussion of the common installation practices for cable bolts may be helpful as apreamble to the review of support principles. Three methods of installation are currentlypracticed as illustrated in Figure 2.1. Methods A and B refer to cable bolts installed in upholeswhile method C is applicable to downholes. For all methods, the cable is inserted into the holeprior to grouting. The breather tube method involves the use of a 9.5 mm (3/8") to 12.7 mm(1/2") diameter polyethylene tube that is installed to the toe of the hole. A 19.1 mm (3/4")diameter grout tube is inserted just beyond the hole collar, and grout is pumped from the collarto the toe. Trapped air is exhausted through the breather tube as grout is pumped into the hole.Grout coming out of the breather tube is an indication that the process is complete. The grouttube method does not require a breather tube to exhaust air, as grout is pumped from the toe to7Figure 2.1: Cable bolt installation methods8the collar. Air is naturally exhausted through the drillhole column. This method can be used inupholes providing that the grout is thick enough to remain in the hole. The breather tube methodrequires a collar plug while the grout tube method does not. A more detailed discussion oninstallation methods along with some of the problems encountered will be covered in Chapter 3.2.3 FAILURE MODESJeremic and Delaire (1983) identified four possible failure modes (Figure 2.2) that applyto cable bolt support. Mode A describes the failure of the grout-rock bond, which is notencountered often in practice due to the roughness of the rock surface and the large contactsurface area. The failure of the grout-cable bond is shown in mode B, while modes C and Dinvolve failure of the grout and rock respectively by internal friction. Rock failure is generallynot a concern in hard rock mining practice but could be encountered in soft rock situations.Jeremic and Delaire (1983) suggest that the most frequent failure mode in practice involvesfailure of the grout, but in fact the most frequent mechanism is thought to be a combination ofmode B and C. Failure of the steel cable is a fifth possible failure mode but is not commonlyencountered in practice.2.4 LABORATORY TESTING2.4.1 Commonwealth Scientific & Industrial Research Organization TestingFuller and Cox (1975) completed pull tests on 7 mm stress relieved round wire groutedin a portland cement paste at a 0.45 water:cement ratio by weight. A typical load-displacementcurve is shown in Figure 2.3. The peak load was reached with little displacement and was10followed with a gradually reducing residual load. The fluctuations in residual load were thoughtto be a result of small variations in wire diameter. The embedment length was varied from 100to 700 mm during these tests and the peak load was found to increase linearly with increasingembedment length. Fuller and Cox (1975) proposed that the grout-steel bond initially failedprogressively along the embedment length of the wire. This process occurs prior to the peak loadbeing attained and results in limited wire displacement. The peak load occurs when the completebond is broken along the embedment length, and frictional resistance continues to generateresidual load as the wire moves out of the grout column. The frictional resistance is graduallyovercome as the wire moves relative to the grout, and the residual load would be expected todecrease. A second series of tests were completed with rusted wires at an embedment length of700 mm The peak load at this embedment was found to be three times higher than smooth wire.It was proposed that increased frictional resistance due to the presence of rust was responsiblefor this increase in peak load. A third series of tests used indented wire at a 700 mm embedmentlength. The load-displacement curve for these tests was very similar to Figure 2.3, except thatthe peak load was approximately twice as high. The residual load fluctuations were still presentand occurred at the same interval as the indentations on the wire. As displacement occurred, itwas proposed that the indentations changed the failure mechanism and forced the wire to crushthe grout. Rust on the indented wire surface was found to improve the peak strengthapproximately 30% . Pull tests were also conducted on 12 mm diameter 7 wire stress relievedstrand, with an embedment length of 450 mm. A typical load-displacement curve for the strandpull testing is shown in Figure 2.4. The peak load attained was just over 11 times the load fora single smooth wire at a similar embedment. One marked difference in the strand load-displacement curve is the increase in post peak residual load. The grout-steel bond was expectedto be higher for strand since the steel surface area was 2.3 times greater than a single 7 mm11Figure 2.3: Typical single wire load-displacement curve (After Fuller and Cox 1975)Figure 2.4: Typical strand load-displacement curve (After Fuller and Cox 1975)1 2wire. The additional increase in strength was thought to be caused by the individual wires havingto fracture the grout in order to pull out. This failure mechanism takes advantage of the fullgrout compressive strength, which is higher than the grout-steel bond strength. Fuller and Cox(1975) proposed that grout fracturing caused the additional increase in peak load, and increasedfrictional resistance resulted in higher post peak residual loads. This testing illustrated that loadtransfer between steel and grout is dependent on the geometry and condition of the steel.2.4.2 United States Bureau of Mines TestingThe United States Bureau of Mines conducted a series of laboratory studies on the supportproperties of cable bolts (Goris 1990, 1991). Thirteen series of pull tests were completed and theresults are itemized in Table 2.1. The cable to be tested was grouted into two segments of 66.5mm diameter steel pipe that were separated by a rubber washer, as illustrated in Figure 2.5. Theportion of the cable installed within the 305 mm pipe was the length that was actually beingtested. The "fixed" end of the cable was anchored in a 508 mm length of pipe with a barrel andwedge anchor. The sample was pulled apart by a hydraulic test machine and the cable would failin the shorter section of pipe. The rate of displacement was set at 0.6 inches/minute untilapproximately 6" of total displacement was reached. These tests simulate a rock slipping off theend of a cable, as shown in Figure 2.5. Series 1 tests were set as the standard, and consisted ofa single 15.9 mm (5/8") cable strand grouted in a 0.45 water:cement ratio by weight, with nobreather tube or additives, and an average embedment length of 287 mm (11.3"). As shown inTable 2.1, the maximum load for the Series 1 test was 9.0 tonnes, and all other tests arecompared as a percentage of this result. The average grout strength was also recorded for eachtest and is shown to be quite consistent, except where the water:cement ratio was varied.A typical load displacement curve for the standard test is shown in Figure 2.6. It is veryTable 2.1: Summary of 28-day USBM Test ResultsTESTSERIESVARIABLE MAX. LOAD(lb)MAX. LOAD(tonnes)PERCENT OFSTANDARDGROUTSTRENGTH(psi)GROUTSTRENGTH(MPA)1 Standard (0.45 w:c/11.3" Embedment) 19820 9.0 100 6940 47.82A 8" Embedment 14700 6.7 74 7291 50.32A 10" Embedment 18960 8.6 96 7291 50.32A 12" Embedment 19300 8.8 97 7291 50.32A 14" Embedment 21600 9.8 109 7291 50.32A 16" Embedment 23120 10.5 117 7291 50.32A 18" Embedment 25360 11.5 128 7291 50.32A 20" Embedment 28840 13.1 146 7291 50.32B 22" Embedment 31020 14.1 157 7175 49.52B 24" Embedment 36320 16.5 183 7175 49.52B 26" Embedment 37920 17.2 191 7175 49.52B 28" Embedment 41360 18.8 209 7175 49.52B 30" Embedment 43040 19.5 217 7175 49.53A 1/4" Breather (full) 19650 8.9 99 7109 49.03A 3/8" Breather (full) 18982 8.6 96 7109 49.03A 1/2" Breather (full) 19200 8.7 97 7109 49.03B 1/4" Breather (full) 19888 9.0 100 7265 50.13B 3/8" Breather (full) 20050 9.1 101 7265 50.13B 1/2" Breather (full) 19934 9.0 101 7265 50.13C 1/2" Breather (MT) 17660 8.0 89 7258 50.04 0.30 W:C Ratio 36820 16.7 186 9844 67.94 0.35 W:C Ratio 32080 14.6 162 8175 56.44 0.40 W:C Ratio 26100 11.8 132 7580 52.34 0.45 W:C Ratio 19820 9.0 100 6940 47.85 Cured 127 degrees F. 22900 10.4 116 N/A N/A6A 1.4:1 Sand/Cement Grout 27920 12.7 141 7600 52.46B 6A & 0.25Ib additiye/100Ib cement 27592 12.5 139 7490 51.66C 6A & 0.451b additive/100lb cement 27195 12.3 137 7605 52.47 Two Cables - 9.36" Emb. 41080 18.6 207 6748 46.58 Two Cables & 1/2" tube - 9.33" Emb. 43058 19.5 217 7103 49.09A Steel Button @ 2" 26500 12.0 134 7375 50.89B Steel Button @ 4' 55840 25.3 282 7337 50.69C Steel Button @ 6" 53950 24.5 272 7470 51.510A Birdcage(A-Node @ Pipe) - 10" Emb. 33960 15.4 171 7600 52.410B Birdcage(Node @ Pipe) - 10" Emb. 25760 11.7 130 7357 50.711A Two Birdcage(AN @ Pipe) - 10" Emb. 77300 35.1 390 7775 53.611B Two Birdcage(N @ Pipe) - 10" Emb. 79750 36.2 402 7250 50.012 Epoxy Coated Cable 27875 12.6 141 7094 48.913 Two Epoxy Coated Cables - 10" Emb. 57550 26.1 290 7061 48.7Reference: (Goris 1990, 1991)Standard: Single 5/8" diameter cable with 0.45 w:c ratio and 11.3" embedment length.Bleeding: W:C Ratio Bleed(in/ft)0.45 1.150.40 0.760.35 0.390.30 0.1913USBM Laboratory TestingPlug riI^ " . : r : : PNWN:**4 .brON MIA ' , , ►. . • . . S 4* . 40=AI r 1111111riiA NIM VW**. V 1P' 'yew : :   ),40,4444,_ _414,014W4X40+#44 • ; ; ; *NM Z;_2,4.4.4 . t•ititty;1 4 %V : : : : INWASSWBarrel and Wedge We .14-4 te4.4 0   : : otto. ,4 004.44 vow-.0.1.44,404,40,40 : : : : Otto, 4, 444040 40.41,Anchor-.- - - - -4 74 7" 4 ' ' : ' A Wet 0.4"4 0, 0I,I Me 6,686. 4 04.4 ; ;. • , „,4).A"."   . W. 40 ' : : : Vegi NW pa,..li^44 14.4 4 • 4" e...41 :  ' ' Wit" . • 44.,"16^1 4 t.4 W Vesit. . • 4 0  0 : : : o , . . 4,4 ....* Ai • 4, ,0 - .. .41) • v. 4 . . . . 4,0"1 tat A 0 04 iss* . . . , : : 04....404.......4,8 C A^1^a 15.9mm Cable^0.- 0:4 vo, iti, 0, 00, A 0 . . , ; 004.44) ," . 4, . 4 ,L r )1PZ, ■ ASV 0. 00"*".44 : : : , I) wit tit AA tsrPP-  4 0 - . -. 4 4 • 4 044 044 : : i ►Ov+e."".4 0.4.4,t."4".".....4 ►144.......44. ...******** :: 44.v . 40 ..i ntitto,,SX4AA& 0 : :  ■ %ow 4, svGrout astinta. it4  Via 14 orti-tilt. 24 V40,  44 ■44'  s.  . ;i '   ^L+„4  AA 0. A.....„wAut...."....   ^lwiriviv.V ^M mi 4 - - Rubber WasherA- - - - - - -- I^1 eirtgaiN NS ,: Feslasioss' S...4 "4 4.4.44 ' : 04.4 AU** 0tiattitiitta 0$  i i :'  littirtaft-, i i Sch 80 Steel P i p e^iitatal i i towatilN^' 1 '. Vs*:440:40-I^I^1IE^ 1111101111111111111$0 : i t.44000000,010141wiAmili^NIAIRI #4.40444W4F 40A : : : 4 4, 4 . 04, 06 , 4 A duiV. 02 tid oit% Oa% . A r.40",4 I E ' ' ' 14 0, A 4 0, 6 A . , .li^' PA (After Goris 1990)A A0—, 011v:::) 1--•<:, Cmo......, ,-,coI.)O.CCl:..1AD0',::CD=6,■—,Cl.ts.)00CI.A:t...5—'POCi.,..,.'7:3tr:C:CDCD=n=CDGI5oy,0*1t—rCDCI1e-r-GO,cococA0—>g,co1-1C.10:-.1 .c.)16similar to the curve produced by Fuller and Cox (1975) for a single strand cable (Figure 2.4).The points on the seven day test curve represent the actions of the cable as it is pulled out of thegrout. Between points A and B, the bond is broken between the steel and the grout, starting atthe pipe joint and propagating to the other end. Between B and C the cable is free to move, butmust break the ridges of grout that are between the individual wires of the strand. Brokenparticles from these ridges provide increased resistance to movement and result in a continualincrease in load until a maximum is reached at point C. Displacement continues between C andD, but dilation, or expansion of grout particles, maintains a high residual strength. The effectof pulling on a cable, as illustrated in Figure 2.7, is to generate a normal force into the groutcolumn as a result of grout particle dilation. Assuming that the rock stiffness is sufficient tocounteract this force, the frictional resistance will remain high, until the strand wires are ableto shear through the grout ridges.Series 2 tests varied the embedment length from 203 mm (8") to 762 mm (30"). Goris(1991) found that the load was linearly related to the embedment length as shown in Figure 2.8.An extrapolation of these results indicates that 1.06 m (41.9") of embedment is required tomobilize the full 26 tonne breaking strength of the cable. The load-displacement curve for eachembedment length is illustrated in Figure 2.9 and clearly shows the limited displacement as thecable-grout bond is progressively broken. The importance of embedment length to cable designis illustrated in Figure 2.10. A 20 tonne block of rock is supported by a single cable that isgrouted with a water:cement ratio of 0.45. Based on the USBM test results, the cable willsupport the block with an embedment length of 1 m. However, if the block is rotated 90°, theembedment length changes to 0.5m and the block can potentially slide off the cable. The jointingwithin a rock mass or the geometry of a particular block are thus important considerations in thedesign of cable support.'71.—.co.=.-1coN:-.)0-3coco000r--n*c)=..=Cr40ZpoCrA0Zi" ITCs.0cocrFr>0■-iZ01-:a:O.,co.-3a0g00Mr'--Cn0=0-I‘.■-8■0010^15^20^25^30Embedment Length (inches)18Figure 2.8: Maximum load versus embedment length (After Goris 1990)Figure 2.9: Averaged 28-day load-displacement curves (After (ions 199U)20Since a breather tube is commonly left in the hole after grouting, the effect of differentsized tubes were analyzed in the third test series. The tube size was varied from 6 4 mm (1/4")to 12 7 mm (1/2") in diameter. The results revealed that the presence of a breather tube in thegrout column had no effect on the cable strength, as long as the tube was filled with grout. Whenthe breather tube was empty, the maximum load dropped slightly but the residual load carryingcapacity was significantly reduced. As the cable was displaced, the grout was able to move intothe void created by the empty breather tube, and frictional resistance was reduced.The water:cement ratio was varied from 0.30 to 0.45 in the next series of tests. Themaximum load increased as the water:cement ratio decreased. This indicates that an increase ingrout strength as a result of reducing water:cement ratio, will increase the load carrying capacityof a cable bolt. A decrease in water:cement ratio from 0.45 to 0.3 approximately doubles the pullout strength of the cable. Cement particles in a grout column will settle immediately afterplacement in a process that is referred to as bleeding. Water rises to the surface as the cementparticles settle. It was found that a grout column with a water:cement ratio of 0.45 would resultin 96 mm/m (1.15"/ft) of grout bleed. The amount of bleed decreased as the water:cement ratiowas reduced, as shown at the bottom of Table 2.1. This is important in terms of cable designsince a vertical uphole that is 18 m in length, would have 1.7 m of water at the toe of the hole,if the water:cement ratio was 0.45.Series 5 samples were cured at 127° F to determine the effect of high rock temperatureson a cable bolt installation. The higher temperature resulted in faster curing of the cement andhigh early strength. Adequate water was thought to be available to complete the hydrationprocess since a grouted cable bolt is well contained, and there is little chance of evaporation. Inseries six, a 1.4:1 mixture of sand and cement was found to increase the maximum load by 41% .The sand particles are thought to provide an interlocking force and therefore greater frictional21resistance as the cable displaces. Goris (1990) noted handling and flow problems with thesand:cement grout that would limit its underground application at this stage.The remaining tests varied the geometry of the cable and resulted in significant increasesin cable loads. The embedment lengths in some of these tests were reduced, since there was someconcern about the capacity of the testing equipment. The load-displacement curves for doublecables (Figure 2.11) reveal nearly twice the single cable maximum load. Higher maximum loadsare reached with very little displacement, indicating that double cables offer a stiffer supportsystem than single cables. Steel buttons were found to greatly increase the load carrying capacityof a cable, providing that the button is located greater than 102 mm (4") from the pipe joint. Thefailure mechanism changes from Figure 2.7, since the grout is forced to fail in compression inorder for the cable to displace. When the button was located 51 mm (2") from the pipe joint, theresulting short grout column was easily fractured and the maximum load was significantlyreduced. Laboratory testing indicates that button location with respect to rock mass jointing willcontrol the load carrying capacity of the cable. Due to the difficulty of predicting thisrelationship, buttons are not commonly encountered in practice. They do however have thepotential of significantly increasing the cable load carrying capacity, and allow largedisplacements at high loads.Birdcage cable refers to an ordinary 7 wire strand that has been untwisted. The result isan open wire cable with a series of nodes and antinodes at 178 mm (7") intervals. Singlebirdcage cable was tested in Series 10 and indicated a 30% to 70% increase in load over ordinarystrand. The open wires of a birdcage cable increase the steel surface area exposed to grout andresult in a higher bond strength. In addition, grout is able to penetrate inside the birdcageconfiguration and each node acts as an anchor. As the cable displaces, the antinode puts the groutin compression and results in high loads but offers a much stiffer system than conventional22Figure 2.11: Load-averaged displacement curves for double cables (After Goris 1990)23strand.The final test series looked at single and double epoxy coated cables with silica gritembedded in the epoxy. The silica grit provided increased frictional resistance and mobilized a40% increase in pull out strength for a single cable. Epoxy coated cable was also found tosignificantly reduce the amount of grout bleed. The configuration of a regular steel strand allowswater to enter into the wire arrangement and subsequently move to the top of the hole throughinternal voids. An epoxy covering does not allow water transmission as the strand arrangementis isolated from the grout mixture.2.5 CRITICAL BOND LENGTHThe critical bond length refers to the length of grouted cable that is required to mobilizethe full breaking strength of steel. An extrapolation of the USBM testing indicates that the criticalbond length for a single 0.625 mm cable grouted in a water:cement ratio of 0.45, is just over1 m. Observation of underground cable bolt systems frequently reveal that many failures leavethe cable intact, signifying that the full strength of the bolt is not fully mobilized prior to failure.Much of the research into cable support has been directed at reducing the critical bond length toincrease the potential of utilizing the full cable breaking strength. The research described inSection 2.4 illustrates several ways of decreasing the critical bond length. These involve:1) increasing grout strength and stiffness2) improving the cable geometry3) increasing frictional resistance by altering the cable bolt surface.Other factors that will increase the critical bond length include reduced grout stiffness due toempty breather tubes or voids and reduced confinement offered by low rock stiffness.242.6 GROUTReichert, Bawden, and Hyett (1992) found that grout unconfined compressive strengthincreased with decreasing water:cement ratio. Grout strengths were tested between awater:cement ratio of 0.28 to 0.60 and the results are shown in Figure 2.12. It was noted thatthe variability in strengths is much larger at lower water:cement ratios. Reichert et al. (1992)suggest that a water:cement ratio of 0.3 does not have adequate flow characteristics for use incable bolt systems, and propose the use of 0.35 to 0.40 as a practical compromise Goris (1991)also noted that a 0.3 water:cement ratio grout would not flow without the addition of a waterreducing agent. Gendron et al. (1992) found that laboratory grout strengths also varied as afunction of the mixing method, as shown in Figure 2.13. Hand mixing of grout produced thelowest strengths, while a drum mixer and Mix and Inject (MAI) system produced progressivelyhigher strengths. The Mix and Inject system (Reichert, Bawden, and Hyett 1992) utilizes a rotor-stator pumping assembly with a continuous mixing process. Underground grout samples werefound to plot towards the lower range of the laboratory curves.2.7 CABLE GEOMETRYThe geometry of a cable support system can be improved by increasing the exposedsurface area and improving the grout-steel bond. Birdcage and double cables both significantlyincrease the exposed surface area. The surface condition of a cable can be altered by the use ofan epoxy coating with embedded silica grit to improve the frictional resistance. The use ofbuttons and birdcage change the failure mechanism to take advantage of the full groutcompressive strength. Noranda has conducted research on a cable grip (Gendron et al. 1992)V(1)^,..,-t c•D0 Nn •=01., t..e.)04 • •0= nCD C,-,- 5AD ,z1— TA•q:, o'CN■-■ 0.4,ncl)..cp,-,Clz,0-1.6CD5nCco0,-.0...i.,0„C]....4;"..,cD"(ID5,-574 •C0'45CDE.0Cuc.,27made from the wedge portion of standard barrel and wedge anchors. The wedge is placed on thecable and covered with a plastic sleeve prior to grouting. Grout surrounding the sleeve functionsas a cement version of the barrel, and test results have shown similar strengths to birdcage cablewith reduced stiffness. The nutcase cable bolt is a recent cable configuration (Bawden, Hyett,and Cortolezzis 1992) that has been developed, and is currently being evaluated. The nutcasegeometry involves placing a hexagonal nut over the central wire of a cable strand and rewindingthe six outside wires so that each wire rests on one edge of the nut. The result is a more tightlywound version of the birdcage cable bolt. Preliminary testing (Bawden, Hyett, and Cortolezzis1992) has indicated slightly higher strengths than the birdcage geometry and further study isplanned to evaluate the ability of low water:cement ratios to penetrate the tighter geometry.2.8 CONFINEMENTThe failure mechanism proposed by Goris (1990) generates a force normal to the cabledue to grout dilation as a result of strand displacement. Reichert, Bawden, and Hyett (1992) havedemonstrated that the stiffness of the material surrounding the grout will affect the cable strength.A series of laboratory and field pull tests were conducted to relate different levels of confinementto the load carrying capacity of a cable bolt. The concept of radial stiffness is introduced tocompare the confinement offered by different pipe materials and rock types. Radial stiffness isdescribed (Reichert, Bawden and Hyett 1992) as the amount of pressure that is required to inducea specified internal radial deformation, expressed in units of MPa/mm The radial stiffness ofpipes was calculated based on thick wall cylinder theory and a borehole dilatometer was used toevaluate radial stiffness for different rock types. The borehole dilatometer is a high pressureinflated packer that records the deformability of 76 mm diameter holes. The relationship between28cable load carrying capacity and embedment length for steel, granite and shale is shown in Figure2.14. Steel pipe exhibited the highest radial stiffness, followed by granite and then shale. Theresults illustrated in Figure 2.14 suggest that a stiffer rock mass improves the load carryingcapacity of a cable bolt by providing increased confinement to dilation of the grout column. Thisin turn results in an increase in the level of frictional resistance to cable movement.2.9 STRESSStress change is to be expected within an active mining area and has been considered incable design by Kaiser, Maloney, and Yazici (1991). The previous discussion on confinementhas indicated that the resistance of a rock mass to grout dilation can have an effect on the cablebolt strength. Kaiser et al. (1991) have proposed that a change in the field stress will result ina corresponding change in the pressure at the grout-steel interface. An increase in stressperpendicular to a cable bolt hole, as might occur in the back of a cut and fill stope, wouldincrease the cable bolt strength by increasing the confinement around the steel. A stress decreasewould similarly decrease the cable bolt strength. Research in this area is currently limited butfuture work is planned to study the implications of this concept.2.10 CABLE ORIENTATIONThe preceding discussion has concentrated on axial loading of cable bolts, as limitedresearch has been conducted on the effects of shearing on cable bolt behaviour. Miller (1984)has suggested that cables are most efficient when oriented parallel to the shear direction andinclined between 17° and 27° from the joint. Fuller (1983a) suggested that this angle should30range between 15° and 30° for maximum efficiency. In situations where the restriction of jointopening is required, the most efficient orientation is at 90° to the joint.2.11 SUPPORT STIFFNESSPotvin, Hudyma, and Miller (1989) suggest that it is desirable in most situations to matchthe rock mass and support stiffness. The support stiffness refers to the amount of deformationthat is permitted prior to failure. The response of an ungrouted cable placed in tension can beevaluated by considering the induced strain and the modulus of elasticity of the steel. In thissituation the entire length of the cable is used to determine the induced strain. Crack dilationbetween the ends of a grouted cable bolt causes gradual debonding of the grout-steel interfaceaway from the crack in both directions. The induced strain is related to the debonded length andcan result in much higher loads than in the free cable situation. The relationship between theamount of debonding that occurs for a particular load has not been well defined and is still underinvestigation. Matthews, Tillmann and Worotnicki (1983) describe the application of an artificialdebonding procedure designed to reduce the stiffness of a cable support system in a highlystressed crown pillar. The technique involved installing plastic tubing over the cable betweenbutton type anchors placed along the length of the strand. The plastic tubing prevents theformation of the grout-steel bond and maximizes the strain potential of the cable. The supportstiffness of a cable system can be increased by tensioning the individual bolts. Aside from theapplication of small loads to seat plates at the hole collar, the tensioning of cables is rarelyencountered in practice.312.12 DISCUSSIONBased on laboratory observations it is evident that the full breaking strength of a cablebolt is not always mobilized, but is in fact dependent on many variables. Some of these variablesare related to different laboratory test results in Table 2.2. Very few cases of steel failure arenoted and most results are related to pull out loads. Villaescusa, Sandy, and Bywater (1992)report on recent laboratory testing that involve mobilization of the full breaking strength ofbirdcage, double and combination cables. Combination cables refer to the use of one standardstrand and one birdcage cable. These results are included in Table 2.2 and suggest that doublecables compare very closely to a single birdcage at a 0.55 water:cement ratio. In terms of cabledesign, laboratory results can be useful in an assessment of the inherent strength of a cable bolt.An estimate of grout quality combined with a review of laboratory test results can provide anindication of expected cable bolt strength. For example, if a 0.5 m thick block is to be supportedand the grout water:cement ratio is estimated at 0.45, Table 2.1 suggests that one cable bolt willsupport approximately 12 tonnes. Additional cables are required if the block is greater than 12tonnes. This type of design application is difficult at this stage since there are many variables toconsider. This thesis advocates an empirical design tool that relates support to the rock massbased on operational experience.Table 2.2: Comparison of Laboratory Test ResultsReference Support TypeEmbedmentLength(mm)Water:CementRatioMaximumLoad(tonnes)SteelFailureDisplacementat Maximum Load(mm)TestingMediumGroutedDiameterCuringTimeGoris 1990 Single 288 0.45 9.0 No 45 Steel Pipe 51 mm 28 daysVillaescusa et al. 1992 Single 500 0.55 10.0* No 8* Steel Pipe 68 mm 7 daysGoris 1990 Single 289 0.40 11.8 No 60* Steel Pipe 51 mm 28 daysGoris 1990 Single 300 0.35 14.6 No 41* Steel Pipe 51 mm 28 daysGoris 1990 Single 299 0.30 16.7 No 68* Steel Pipe 51 mm 28 daysVillaescusa et al. 1992 Sin gle 1000 0.55 15.3* No 25* Steel PipeSteel Pipe68 mm68 mm7 days7 daysVillaescusa et al. 1992 Single 500 0.30 17.1* No 43*Cods 1990 Single 763 0.45 19.5 No 61 Steel PipeSteel Pipe51 mm68 mm28 days7 daysVillaescusa et al. 1992 Single 1000 0.30 26.1* No 42*Goris 1990 Single 1064 0.45 26.3 Projected Steel Pipe 51 mm 28 daysVillaescusa et al. 1992 Double 500 0.55 17.8* No 20* Steel Pipe 68 mm 7 daysGoris 1990 Double 238 0.45 18.6 No 3 Steel Pipe 51 mm 28 daysOliver 1992 Double 305 0.30 (additive) 52.6 Yes Rock 24 hoursVillaescusa et al. 1992 Double 1000 0.55 30.0* No 31* Steel Pipe 68 mm 7 daysVillaescusa et al. 1992 Double 500 0.30 40.8* No 40* Steel Pipe 68 mm 7 daysVillaescusa et al. 1992 Double 1000 0.30 51.0 No 50* Steel Pipe 68 mm 7 daysGoris 1991 Button g 2" 305* 0.45 12.0 No 18* Steel Pipe 51 mm 28 daysCods 1991 Button @ 6" 305* 0.45 24.5 No 131* Steel Pipe 51 mm 28 daysGoris 1991 Button @ 4" 305* 0.45 25.3 No 120* Steel Pipe 51 mm 28 daysGoris 1991 Birdcage-Node 254 0.45 11.7 No 8 Steel Pipe 51 mm 28 daysGoris 1991 Birdcage-ANode 254 0.45 15.4 No 7 Steel Pipe 51 mm 28 daysVillaescusa et al. 1992 Birdcage 500 0.55 18.4* No 17* Steel Pipe 68 mm 7 daysVillaescusa et al. 1992 Birdcage 1000 0.30 24.3* Yes 14* Steel Pipe 68 mm 7 da sVillaescusa et al. 1992 Birdcage 500 0.30 24.4* Yes 18* Steel PipeSteel Pipe68 mm68 mm7 days 7 daysVillaescusa et al. 1992 Birdcage 1000 0.55 26.9* Yes 23*Goris 1991 Double Birdcage -ANode 254 0.45 35.1 No 120 Steel Pipe 51 mm 28 daysGoris 1991 Double Birdcage-Node 254 0.45 36.2 No 15 Steel Pipe 51 mm 28 daysVillaescusa et al. 1992 Combination 500 0.55 32.5* No 16* Steel Pipe 68 mm 7 daysVillaescusa et al. 1992 Combination 1000 0.55 45.6* Yes 29* Steel Pipe 68 mm 7 daysVillaescusa et al. 1992 Combination 500 0.30 44.1* Yes 24* Steel Pipe 68 mm 7 daysVillaescusa et al. 1992 Combination 1000 0.30 50.1 Yes 25* Steel Pipe 68 mm 7 days* Approximated33CHAPTER 3CABLE SUPPORT PRACTICE3.1 INTRODUCTIONCable bolt practice was noted in detail at all of the operations visited, and when theopportunity arose time was spent with the cable bolt crews. The database will be discussed indetail in Chapter 6, but the particulars of cable bolt practice for each case study are summarizedin Table 3.1. Table 3.2 summarizes the installation procedures for each minesite, and a moredetailed description is available in Appendix C. This chapter will present a review of currentcable bolt practice based on field observations and a literature review.3.2 GROUTGrout quality was found to vary quite extensively at each mine. An attempt was made tolog an estimate of the grout quality for each case study based on visual observation of thegrouting process. On average, the grout water:cement ratio was estimated to be in the 0.4 to 0.45range but varied from 0.35 to 0.55. Laboratory testing completed by Goris (1990) relates thegrout water:cement ratio to the maximum load. The results show a maximum pull out load of9 tonnes for a 0.45 water:cement ratio and 16.7 tonnes for a 0.30 water:cement ratio, based ona 287 mm (11.3") embedment length with a single 15 9 mm (5/8") cable and no breather tube.In similar tests, other authors (Reichert, Bawden, and Hyett 1992) have shown an increase in pullout load with decreasing water:cement ratio. Goris (1990) also found that grout bleed decreasedfrom 96 mm/m (1.15"/foot) of cable at a 0.45 water:cement ratio to 16 mm/m (0.19"/foot) atTable 3.1: Case Study Support PracticeCASE Stability Surface StrapsOtherSupport Grout Pump1 Caved HW No No S^del 60002 Stable Back No No S^del 60003 Stable HW No No S^del 60004 Caved HW No No S^del 60005 Stable Back No No S^del 60006 Caved HW No No S^del 60007 Unstable HW No No S^del 60008 Unstable HW No No S^del 60009 Stable HW Point Anchor Blasthole 51 • No No S^del 600010 Stable • No 2.4m Swellex S^del 600011 Stable HW S^del 600012 Stable Back • Steel Sets .7m Sw/1.5m RB S^del 600013 Stable Back • No Rockbolts S^del 600014 Stable Back No Rockbolts S^del 600015 Stable Back No S^del 600016 Caved Back No S^del 600017 Caved Back . No S^del 600018 Stable Back • No Rockbolts S^del 600019 Unstable Back • No Rockbolts S^del 600020 Caved HW No No S^del 600021 Stable HW No No S^del 600023 Stable No Rockbolts S^del 600025 Caved Back HW/FW & S•uare Blasthole 51 No Rockbolts S•-del 600026 Stable HW No S• del 600027 Stable HW No S^del 600029 Stable Back • No Rockbolts Mo no 3L330 Stable Back .. Yes 2.4m RB/Scr/Ex S^del 600032 Stable HW w • Yes No S^del 600033 Stable Back Fart VCR 51 • _ No 2.4m RB/Screen S^del 600034 Unstable Back • No 2.4m Rebar/Ex S^del 600036 Stable Back • • It No 2.4m RB/Ex S^del 600037 Stable Back • Yes 2.4m RB/Scr/Ex S^del 600043 Stable Back Yes 2.4m RB/Scr/Ex N/A45 Stable Back • _ No 2.4m RB/Swellex Min •ro 346 Stable Back • • No Rockbolts Min •ro 347 Stable Back • No Rockbolts Min •ro 348 Stable Back • No Rockbolts Min•ro 349 Stable HW No Min •ro 350 Caved HW No Min •ro 352 Stable Back • • No RB/Resin Min •ro 353 Stable Back • No Rockbolts Min •ro 354 Stable HW No No Min •ro 355 Stable Back No Rockbolts Min • ro 356 Stable HW No No Min •ro 357 Stable Back No Rockbolts Min ro 359 Stable Back • No 2.1m Swellex Min • ro 3Table 3.2: Cable Bolt Installation Procedure SummaryMine Pump InstallationMethodAnchor W:C Ratio #CablesLength(m)Mine #1 ' Spedel 6000 Breather  Bent Wire 0.45 2 6.1-22.0Mine #2 Minepro 3 Breather > 12.2mGrout < 12.2mSpring Steel 0.3750.322 8.0-15.0Mine #3 Spedel 6000 Breather Spring Steel 0.45 1 & 2 9.8-14.6Mine #4 Moyno 3L3 Breather Bent Toe Wire(1 or 2)0.50 2 18.3Mine #5 Spedel 6000 Grout Spring Steel 0.40 1 6.1-12.0Mine #6 Spedel 6000 Grout Spring Steel 0.4-0.45 1 12.2Mine #7 Minepro 3 Grout Bent Toe Wire 0.35-0.40 1 15.8Mine #9 Minepro 3 Breather > 13.7mGrout < 13.7mMessenger Wire& Wedge0.33-0.400.301 5.0-18.3Mine #10 Minepro 3 Grout(Retrieved)Bent Toe Wire(2)0.35-0.40 2 4.9Mine #11 Cabolt Grout(Retrieved)Cable Kink 0.30-0.35 2 8-1036a water:cement ratio of 0.30. Bleed has been found to be an important consideration in cut andfill mining where cables are typically installed at lengths of up to 18.3 meters. The last lift maynot contain fully grouted cable if the water:cement ratio is too high (Cluett 1991).Limited monitoring of the grout quality was found to exist and is recommended as animportant component of any established cable bolt procedure. A test kit (Gendron et al. 1992)from the Noranda Technology Center was made available for this study in order to collect groutsamples to be submitted for strength testing. Unfortunately, samples were collected on only oneoccasion due to the uncertainty of arriving on site while cables were being grouted. The kitcomes in a metal box and contains tubes that can be filled with grout and sent to the laboratoryfor testing. It is also possible to visually estimate the grout quality, but this requires someexposure to the flow characteristics of grout at different water:cement ratios. This is best doneTable 3.3: Grout Characteristics (After Hyett, Bawden, and Coulson1992)W:CRATIOGROUT CHARACTERISTICSAT END OF GROUT HOSEHANDLINGCHARACTERISTICS<0.30 Dry, stiff sausage structure Sausage fractures when bentGrout too dry to stick to handCan be rolled into balls0.30 Moist sausage structure'Melts' slightly with timeSausage is fully flexibleGrout will stick to handEasily rolled into wet, soft balls0.35 Wet sausage structureStructure 'melts' away with timeGrout sticks readily to handHangs from hand whenupturned0.40 Sausage structure lost immediatelyFlows viscously under its ownweight to form pancakeGrout readily sticks to hand butcan be shaken free0.50 Grout flows readily andsplashes on impact with groundGrout will drip from hand - noshaking required37by preparing samples and observing how each sample flows into the mixing container and fromthe grout hose. Hyett, Bawden and Coulson (1992) present a description of differentwater:cement ratios (Table 3.3) that describe the observation of grout quality. Normal portlandcement was used at all the mines visited in this study, but the use of high early strength cementwas noted on a few occasions. Admixtures were found to be limited to anti-bleed and water-reducing chemical agents that were used on an inconsistent basis. The admixtures encounteredrequired typical dosages of approximately 1% by weight of cement, and were added during thegrout mixing phase. Gendron et al. (1992) suggest that grout admixtures create quality controlproblems that overshadow any possible benefits. Premixed additives are available but result inincreased material costs.3.3 GROUT PUMPSTwo main types of grout pumps were found in use in western Canadian mines The mostcommon pump was the air powered Spedel 6000 reciprocating piston pump that was used inconjunction with a Spedel B3100 mixer (Oliver 1992). It was found that the use of a Spedelpump was generally associated with the use of a grout water:cement ratio above 0.4. This doesnot appear to be a limitation of the pump as 0.35 water:cement ratio grout was observed to besuccessfully pumped, and Oliver (1992) discusses successful pumping of 0.3 water:cement ratiogrout in a 6.1 meter length of plastic tubing. Industrial experience however indicates that theSpedel 6000 is limited to pumping 12 meters for a vertical hole at 0.35-0.40 water:cement ratio.Oliver (1992) describes the reduction in Spedel pump performance as a result of inadequatepump cleaning. The Minepro 3 pump is also used at a number of operations, but is still relativelynew on the market. The Minepro is a positive displacement pump with a rotor-stator assembly38that is designed to pump grout down to a 0.3 water:cement ratio. Practical applications observedin this study indicated that pumping problems occurred below a water:cement ratio of 0.35 invertical holes longer than 14 meters. Both a pneumatic and electric version of the Minepro 3pump are available and it can also be operated from the hydraulics of a host vehicle. The internalparts of the rotor-stator assembly are easy to remove and clean. The mixing tank is mountedwithin the pump hardware and offers a more effective shear mixing mechanism than the Spedelpump. The Minepro pump however is larger and more difficult to mobilize, whereas the Spedelpump is preferred for areas of restricted access. Most of the mining operations are movingtowards the Minepro 3 pump in an attempt to improve the quality of grout obtained. TheMinepro seems to handle the mine environment well but has yet to establish a lengthy operatinghistory.Both the Minepro and Spedel pumps involve a batch mixing process where a fixedquantity of grout is prepared and pumped prior to returning to the mixing phase. Continuousmixing presents additional quality control problems and has not yet been efficiently adapted tocable bolt applications. A recirculation system (Oliver 1992) is required during the groutingphase to permit continual flow during the pumping of a prepared batch. It was found in thisstudy that shutting down the pump with grout still in the system frequently lead to line blockage,especially utilizing a water:cement ratio below 0.40. When establishing a cable bolting procedureit is recommended that consideration be given to the hardware involved. A special vise forremoving the rotor from the stator for cleaning was found useful at one operation. Minimizingrestrictions in the line from the pump to the grout tube can help in reducing line blockages, andadapting quick coupling mechanisms to the hose ends can improve efficiency. Pagel (1987)describes the use of reusable rubber plugs to block hole collars while grouting with the breathertube method. These plugs are held in place during pumping and incorporate a slot for the39breather tube. In addition, the grout plug incorporates a 25 mm nipple that permits the couplingof a 25 mm hose directly from the pump. Typical practice in Western Canada involves couplingthe pump discharge directly to the 19.1 mm grout tube and generates frictional resistance thatseems to limit pump performance. Bourchier, Dib and O'Flaherty (1992) report on the use ofa grouting assembly that also permitted coupling of a grout hose directly from the pump. Thismethod was devised to pump 0.35 water:cement ratio grout with a Spedel 6000 pump using thebreather tube method and 15.2 meter vertical holes. The grouting assembly was found to reducethe frictional head losses and eliminate bursting hoses but was only reported successful when thegrout tube was recessed 8 meters from the hole toe.3.4 PLATES AND STRAPPINGPlates were used at four of the operations visited but only on a consistent basis at onelocation. The plates varied from 102 mm x 102 mm (4" x 4") to 305 mm x 305 mm (12" x 12")in size and 6.4 mm (0.25") to 9.7 mm (0.38") in thickness. The majority of the case historiescollected did not incorporate plates or straps within the design. The use of plates is importantwhere blocks have the ability to unravel around the cables or where tensioning may be desirable.Plates also mobilize the full strength of a cable bolt by preventing blocks from sliding off thestrand. Additional restraint can be provided by tying cables together with strapping or screen.In one case, cables were designed to be three meters longer than the holes so that they could belaced together using wire rope clips. This idea may have some merit but in practice the lacingwas not completed due to production requirements. The use of plates in conjunction with strapswas successfully observed at several operations. A strap thickness of 4.8 mm (3/16") wasselected in one case to allow for some ground movement to occur. In another case, 102 mm x40102 mm (4" x 4") square mesh screen was used in conjunction with cables and straps. Bothplates and straps were attached to cables with the use of barrel and wedge anchors. The barreland wedge anchor consists of a segmented tapered steel wedge that fits inside a tapered steelbarrel (Thompson 1992). The whole assembly slides over the cable with the tapered end of thebarrel against the plate or strap. Rock movement pushes the barrel against the wedge, which inturn provides a positive interlock against the cable. The barrel and wedge anchors encounteredin practice were flat on both ends, but cast domed anchors are also available to improve the loaddistribution at the plate-anchor junction. The installation of barrel and wedge anchors wasfrequently done by hand and rarely resulted in plates being tight to the rock. A hollow drill rodwas used on one occasion in conjunction with the deck of a scissor lift to tighten plates. Twomines used a hydraulic jack to apply approximately 2 to 4 tonnes to plated cable ends. This wasfound to be the most effective method of installing plates and straps, but results in reducedproductivity. The installation of plates and straps was generally left until two or three days aftercable grouting. Hyett, Bawden, and Coulson (1992) suggest that plates can be installed 16 to 24hours after grouting without damaging the cable-grout interface.3.5 CABLE GEOMETRYNinety percent of the mines visited used 15 9 mm (5/8") cables that were supplied in pre-cut lengths. One operation ordered their cable in coils that were shipped underground and cutby the cable bolt crew as required. This permits the installation of variable cable lengths, butcontinuous cable is susceptible to tangling and requires additional capital investment to providemobility. Double cables were used most of the time but single cables were used in 46% of thecases. Although birdcage cables have recently become available on the Canadian market, none41were found in use at present. One operation had installed birdcage cable bolts in the past butwere not using them on a regular basis.3.6 CABLE ANCHORSSupporting cable bolts in the hole prior to grouting is an important consideration sincethe integrity of the anchor must hold the cable until the grouting phase is completed. Cases ofanchor failure have been noted at different operations and can have serious consequences. Mostoperations simply bend back one or two wires on one of the cables, either 135° at the toe of thehole or 45° at the collar. The frictional resistance between the bent wire and the hole surface issufficient to support the weight of the cable. Pushing cables with bent wires at the toe of the holeis a difficult process due to the resistance of the hole walls, but has the advantage of immediatelysupporting the cable. As noted in Table 3.1, the bent wire anchor is occasionally placed at thecollar of the hole, resulting in an easier installation process, but does not fully support the cableuntil installation is complete. Pull test results from one operation using 18.3 meter single cablesfound that a single wire anchor at the toe of the hole required five times the weight of the cableto cause slippage. In the case of a double wire anchor, twelve times the cable weight wasrequired to induce slippage. Four of the mines visited now utilize a ferrule, or end holdingdevice, that is factory mounted at the end of the cable and supplied with spring steel strips thatare screwed on underground. The ferrule and spring clips can usually be pushed into the holeeasier than a bent wire anchor, but the required strip size varies with hole diameter. Two springsteel strips are usually screwed onto the end of the ferrule perpendicular to each other. Someslippage problems were noted by mine operators and were countered by increasing the numberof steel strips. Two other methods of anchoring cables in western Canada included utilizing two42lengths of messenger wire and placing a full kink in the cable. The messenger wire was clampedto the cable with packaging strapping and was found to provide successful anchorage at oneoperation Kinking of the full cable was used at two operations that employed the use of a fullymechanized cable bolting jumbo. Wedges were used on some occasions to provide secondarysupport at the hole collar but are not suitable for primary cable anchorage. In terms ofinstallation, it was common practice to anchor a large number of cables in one pass and completethe grouting phase at a later stage.Mechanical anchors are available for cable bolts but were only used on one occasion(Fraser 1976) in western Canada to tension 25 mm locked coil ropes prior to grouting. Themechanical anchor is similar to the standard mechanical wedge and bail assembly associated withstandard rockbolts. Ungrouted tensioned cables provide a soft support system that is dependenton no anchor slippage and good plate contact at the hole collar. Fuller, Dight, and West (1990)suggest that tensioning of cable bolts is not necessary as high loads are built up as a result ofrock mass dilation. Tensioning of cable bolts may be useful where support is installed aftersignificant rock mass movement has already occurred. In this situation, the inherent strength ofthe rock mass is reduced and tensioning is useful in limiting further dilation. Fuller (1981) alsosuggests that tensioning may be necessary in areas of low horizontal stress where minimalclamping forces are present.3.7 INSTALLATION PROCEDUREFifty percent of the operations visited were found to use the traditional breather tubetechnique of grouting cables, referred to as the breather tube method in this study. This methodinvolves attaching a breather tube to the top of the cable prior to installation and placing this end43at the toe of the hole (Figure 2.1). A grout tube is taped to the cable just inside the hole collarand the hole is plugged prior to grouting. Materials used for plugging holes include an expandingpolyurethane foam, burlap, rags, shredded cloth and cement grout plugs. The pump dischargehose is then hooked up to the grout tube and the hole is filled from the collar to the toe. Air isforced through the breather tube and can be monitored by bubbling the discharge through water.The grouting phase is complete when grout is observed returning through the breather tube. Thegrout tube is then disconnected and the tubes are tied off to prevent grout leakage. Forcing thegrout to return through the breather tube requires high pump pressures and often results inbursting of hoses (Bourchier, Dib, and O'Flaherty 1992). In addition, the breather tube methodis susceptible to leaks, especially near the hole collar where ground may be fractured. Minorleaks can be plugged as they occur but occasionally become too severe to permit successfulgrouting. In this situation, recommended procedure at most operations suggests leaving the holefor one hour to allow grout to gel within the cracks. The typical breather tube diameter observedwas 9.5 mm but was found to range up to 12.7 mm Cluett (1991) describes the closing ofbreather tubes near the hole collar when grouting 19.8 meter holes with less than 0.45water:cement ratios. This was believed to be the result of high hydrostatic pressures due to thelong grout columns and was corrected by utilizing high pressure breather tubes.The second method encountered for grouting upholes is referred to as the grout tubemethod, and involves the use of a thick grout and a single grout tube taped to the cable bolt endplaced at the toe of the hole. The grout is pumped through the grout tube and fills the hole fromthe toe to the collar. It is important that the grout used in this method has sufficient viscosity toremain in the hole and advance as a continuous front. Several authors (Oliver 1992; Reichert,Bawden, and Hyett 1992) have indicated that a 0.35 water:cement ratio grout will remain in anuphole. Some operators plug the hole upon completion, and continue pumping in an attempt to44pressurize the grout column and fill any remaining voids. The grout tube is normally left in thehole, but one operation retrieved the grout tube during grouting so that the tube could be reusedand installation time reduced. Oliver (1992) observed that it was almost impossible to completelyfill a clear plastic tube by retracting the grout tube, and it is believed that this method will onlyenhance the presence of voids within the grout column. Some mine operators (Cluett 1991;Bourchier, Dib, and O'Flaherty 1992) have observed separation of the grout column while usingthe grout tube method with low water:cement ratios in holes greater than 15 meters. Grout eitherappeared at the collar, falsely indicating that the hole was full, or left voids within the groutcolumn. The analysis of grout flow within a cable bolt hole is difficult to observe and would bean interesting area of future research.When observing the grouting operation during the collection of data for this study, thegrout tube method was found to be preferred over the breather tube method. The failure of groutTable 3.4: Cable Bolt Grouting Procedure (after Cluett 1991)HOLE LENGTH(m)HOLE ANGLE(FROM HORIZONTAL) PUMPBREATHERTUBEGROUTTUBECOLLARPLUG GROUTUphole > 13.7 m > 450 Minepro 3 3.1 MPa (450 psi) 1.7 MPa (250 psi) Cement 0.33-0.40Uphole > 13.7 m < 450 Spedel 6000 3.1 MPa (450 psi) 0.7 or 1.7 MPa(100 or 250 psi)Cement 0.40-0.45Uphole 9.1 - 13.7 m > 450 Minepro 3 None 0.7 or 1.7 MPa(100 or 250 psi)None 0.30-0.33Uphole < 9.1 m all Minepro 3 None 0.5 MPa (75 psi) None 0.30-0.33Uphole < 9.1 m all Spedel 6000 0.5 MPa (75 psi) 0.5 MPa (75 psi) Cement 0.40Downhole > 9.1 m all Minepro 3 None 0.7 or 1.72 MPa(100 or 250 psi)None 0.30-0.33Downhole > 9.1 m all Spedel 6000 None 0.7 or 1.72 MPa(100 or 250 psi)None 0.40Downhole < 9.1 m all Minepro 3 None 0.5 MPa (75 psi) None 0.30-0.33Downhole < 9.1 m all Spedel 6000 None 0.5 MPa (75 psi) None 0.4045to flow out of the breather tube was commonly observed, especially at lower water:cement ratios.This frequently lead to the addition of water to the grout mixture and a drastic increase in thewater:cement ratio. Cluett (1991) describes a grouting procedure (Table 3.4) adapted to a miningoperation in Manitoba that tackles some of the problems discussed in this section. This procedurereflects operational experience and limits the use of the Spedel pump to upholes either less than13.7 meters or less than 45° from horizontal. For holes that are greater than 13.7 m and 45°from horizontal, the Minepro 3 pump is used with the breather tube method and a water:cementratio of 0.33 to 0.40. The use of the grout tube method is preferred for shorter holes along witha lower water:cement ratio. Recent operational practice has favoured the use of the Minepro 3pump for all conditions. Table 3.4 is representative of the ideal practice observed in this studyand is recommended for consideration when establishing operational procedure. Observed cablebolt installation procedures usually required the flushing of the breather tube with water prior togrouting, to ensure that the tube is not internally blocked. It was also found beneficial to cutbreather and grout tube at 45° in order to minimize the size of the cable/tube configurationinserted to the hole toe. Downholes are grouted using the grout tube method but are not restrictedto the use of grout below a water:cement ratio of 0.35. Breakthrough holes are to be avoidedwith the use of downholes since they entail some method of blocking the hole toe. Uponcompletion of drilling, downholes should be blown clean prior to retrieving drill rods andblasthole plugs should be placed in the hole collars.3.8 DESIGN LAYOUTSIt was quite common for design layouts of cable bolt holes to be issued as a standarddrawing to be adapted by the driller to various stope widths. This method leads to some46organizational problems especially if the drill and cable bolt crews are different. If drill holequality is a problem, the use of issued engineering layouts would be recommended. A samplecable bolting layout sheet is shown in Figure 3.1.Hole diameter varied from 51 to 64 mm for both single and double cables. No installationproblems were encountered with this range of hole size, although procedures often required thatlarger diameter holes be drilled in poor ground. Schmuck (1979) suggests that drillhole diameterbe designed to provide 6 to 13 mm between the cable and the hole.Cable bolt holes are typically drilled with pneumatic percussive drills that are mountedon a mobile rig. Blasthole drilling experience with these types of drill rigs limit uphole lengthsto approximately 18.3 meters, and downholes to 25 meters. Fuller (1981) notes that holedeviation can alter the cable pattern at depths in the range of 20 meters and Hunt and Askew(1977) suggest a maximum length of 19.5 meters for 65 mm holes. Drilling and cable installationare frequently completed by different crews and it is important for both to understand thepurpose and importance of cable bolts. The layout sheet in Figure 3.1 can be used by both crewsand provides the installation crew with an idea of problems encountered during the drilling phase.This information is important where breakthrough holes or cracked ground can create groutingproblems. Cable lengths are usually designed to cover the full length of the drillhole butcountersinking is used where bolts are not required for the full hole length. Countersinkinginvolves recessing the cable beyond the hole collar, utilizing aluminium drill rods or segmentedloading sticks. The practice of countersinking produces lower insertion productivity butminimizes the unnecessary use of cable. The layout sheet in Figure 1 can be used to indicaterecess depths to the cable crew.47Figure 3.1: Sample cable bolt layout sheet483.9 CABLE PATTERNSSome typical examples of open stope patterns encountered are presented in Figures 3.2to 3.4. It should be noted that only one case study involved development that was specificallydesigned for the installation of cable bolts. In all other cases, the support pattern was designedaround existing development. Table 3.1 describes each case in terms of the support pattern anda summary of the most common patterns is presented in Table 3.5.Table 3.5: Summary of Cable Patterns in Western Canadian PracticeSUPPORT PATTERN NUMBER OF CASES PERCENTAGEFan Back 4 9%Square Back 20 45%Point Anchor Back 4 9%Even HW 2 4%HW Drift Fan 1 2%Point Anchor HW 13 29%Quasi-Mandolin HW 1 2%3.9.1 Back SupportCase studies of back support were made up largely of blasthole and vertical crater retreatopen stoping situations, but some cut-and-fill, room and pillar, and drift cases are also includedin the database. Figure 3.2a shows typical cable installations of square back support patternsencountered in cut and fill mining Long cables up to 18.3 m are normally installed in upholesto cover three or more mining lifts. Extra cables are often installed into the hangingwall.Footwall rolls within the ore zone require the placement of additional bolts on certain lifts.49Figure 3.2: Typical back support for open stope and cut and fill mining50Cables can also be installed from an overcut but are restricted by the length of holes that can beaccurately drilled. In narrow open stopes, cable bolts are installed in a fan pattern (Fan Back)from footwall to hangingwall (Figure 3.2b). Where the development is large enough, cable boltsare installed on a regular square pattern (Square Back) and sometimes angled into thehangingwall and footwall. Excessive dilution from the hangingwall or footwall can undercut thistype of back support and induce failure. Poor distribution of cables into a stope back often occursas a result of limited access, or where the drill drifts are not slashed to the full width of theorebody. In this case a regular square pattern is not possible and the point anchor approach(Point Anchor Back) is often adopted, as illustrated in Figure 3.4h.3.9.2 Hangingwall SupportHangingwall cable bolting was largely installed from a sublevel drill drift to act as a pointanchor (Point Anchor HW) as illustrated in Figure 3.3c. The bolt densities encountered on thesublevels varied from two to seven bolts installed on rings spaced 2.4 m along strike. The designstrategy in some cases was not to stabilize the whole hangingwall but to limit the effect ofundercutting as mining advanced to the next lift. A number of case histories of the point anchorapproach to cable support have been assembled in order to develop design strategies for this typeof bolt pattern. Further discussion on point anchor design will be pursued in Chapter 7. It isbelieved that the localized bolt density is not as important as the distance between each pointanchor. Hangingwall cables can be evenly distributed over the supported surface by drilling holesfrom a hangingwall drill drift (HW Drift Fan) or countersinking the bolts through the back ofa sublevel drift (Even HW) as illustrated in Figures 3.3a and 3.3b. Cables installed from aseparate hangingwall drift were encountered on only one occasion due to the high cost associatedwith the additional development required.523.9.3 Other Support PatternsCable slings were encountered in isolated cases of crown pillar recovery, bulkheadsupport and pillar reinforcement. In the case of crown pillar recovery, cable slings were used tosupport a log mat below slag or tailings fill as mining advanced forward (Figure 3.4a). Slingswere also used to reinforce the back and walls of conventional sublevel development for an openstope slot. Although cable slings were not the main focus of this study, a similar supportmechanism was encountered in two design applications. Figure 3.4b illustrates an application ofthe sling type approach applied to back support of an open stope. This type of support may beuseful where there is a poor distribution of cables over the surface. The drill crosscuts were closeenough to allow cable bolt holes to be drilled from one to the other. Cables could be installedand plated on each end in an attempt to sling the back between each drill crosscut. No patternsof this nature were encountered in practice but they were discussed at a design level. Mandolinbolting is another method of cable support that was encountered at the design stage. Cable boltsare installed parallel to the stope hangingwall and attached to a second set of angled cablesinstalled above the sublevel drill drift (Figure 3.4c). The cables parallel to the hangingwall areangled less than the dip of the surface in order to place the end of the cable into good qualityrock. The sublevel drill drift may be shotcreted to protect the exposed portion of the cables.3.10 HEALTH AND SAFETY ASPECTS OF CABLE BOLTING3.10.1 Cable PushingA single strand 15.9 mm cable typically has a nominal weight of approximately 1.1 kg/m.In a 9 meter vertical hole, a double strand bolt would generate an effective weight of 20 kg.Manually pushing cables into a drillhole is the most common method of installation. In addition53Figure 3.4: Sling approach to open stope support54to the cable weight, the operator has to contend with grout tube, breather tube and an additionalforce to overcome a cable holding device. The use of bent wire and spring steel anchors havebeen discussed earlier in this chapter. The pushing of cables into a drillhole is an awkwardprocess and can result in back injury, especially in long upholes. The cable pushing crew shouldbe provided with some means of reaching the hole collar in order to assume a effective stanceto manually push cables. A scissorlift is commonly used at many operations and should bededicated to the cable bolt crew. Cables and other equipment can easily be moved from area toarea, and a suitable installation platform is readily available. Specialized mobile cable boltingequipment incorporating a grout pump, storage area and an installation platform is available onthe market, or can be custom designed.Several mechanized methods of pushing cables have been developed in an attempt toreduce the manual effort required and increase productivity. An air powered inserter wascommonly encountered among western Canadian mines, but was reluctantly used by cable boltcrews due to cable slippage and lower installation productivity. The opposite is suggested byPagel (1983), who reports on the potential of doubling cable insertion productivity with the useof a mechanized inserter. The inserter operates similar to an old fashioned roller dryer by feedingcables through two rubber rollers into the drillhole. The deck of a scissor lift was used at oneoperation to push the final segment of cable incorporating a bent wire anchor at the collar. Theresistance provided by cable weight and the frictional component of the anchor was too high toallow for manual insertion. The exposed cable end was inserted into a hollow drill rod and thescissor deck was slowly raised to force the cable into the hole. This was found to be an effectivemethod but the drill rod must be kept aligned with the hole and secure footing ensured. Continualraising and lowering of the scissor deck during this process is bound to increase maintenancecosts. Hunt and Askew (1977) describe a procedure of mechanical cable insertion utilizing the55feed motor of a longhole drill rig and a cable grab attached to the shank adaptor. A similarmethod was encountered at one operation in western Canada.3.10.2 Cable AnchoringTypical installation practice in western Canada involves the insertion of a large numberof cables in one pass with grouting completed at a later stage. If a cable slips from an upholeprior to grouting, it presents a danger to the bolting crew and to other personnel in the area. Thecable weight is sufficient in most cases to cause serious injury and the possible whip can covera large area. It is good practice to push and grout cables with as little delay as possible, and torestrict access during this phase of the operation. Slippage has been reported at several operationsin western Canada utilizing both the bent wire and spring steel methods of anchoring. In an effortto reduce cable slippage, procedural changes have been incorporated to increase the number ofspring steel strips, or utilize a double instead of a single wire anchor. A double wire anchor atthe hole toe is difficult to manually install, and prompted one operation to return to the singlewire anchor with the adoption of a protective covering for the cable end. This covering is aplastic mushroom shaped fitting that is pushed onto the exposed cable. In the event that a cabledoes release from a hole, the protective cover provides a blunt face that reduces the chance ofa penetrating injury. Hunt and Askew (1977) suggest that the radius of the bent wire must bekept small, 75 mm for 65 mm hole diameters, in order to maximize the frictional resistance ofthe anchor. A longer length provides less resistance to slippage since the wire tends to bendfarther as the cable is pushed up the hole.3.10.3 Effect of Working with CementBatch mixing of grout is typically done with a mechanical mixer, but the water and56cement are added manually. Pouring cement into the mixer produces an excessive amount of dustin the vicinity and the use of masks are recommended. Cable bolting is frequently located insublevel stope development that requires adequate auxiliary ventilation. Hunt and Askew (1977)note that 80% of the injuries during cable installation at one operation were the result of cementburns from the grout. Schmuck (1979) also indicates that most injuries are a result of cementburns and recommends the use of long gloves, eye goggles and respirators. Skin/grout contactusually arises during the pumping phase of the operation as grout leaks from the hole or fromthe end of the grout tube. The use of protective overalls and waterproof suits were commonlyused by operators in western Canada. Water is normally on hand during the grouting phase andis recommended for use in the immediate washing of any areas of skin/grout contact.3.10 4 Handling Cable BoltsPre-cut bolts are cut to specific lengths in the factory and shipped in 1.2 meter diametercoils. Each wrap of the cable is secured by strapping and must be cut by the bolting crew priorto cable installation. This is a dangerous process as the cable tends to whip as each strap is cut.The recommended procedure is to cut the straps in sequence while standing in the middle of thecoil in order to avoid the cable whip. The cables are strapped in sequence in the factory and mustbe cut in the reverse sequence to limit the amount of cable released with each strap. It isimportant to use a cutting device that is quick and effective. Cutting with an air operated diskcutter is preferred over the use of a hacksaw. The use of leather gloves, safety glasses and acutter protective guard are an important part of this operation. Personnel should be cleared fromthe area where pre-cut cable coils are being opened. The cutting of cable underground is usuallynot as clean as in the factory and can result in sharp ends that require special care whenhandling. The use of the protective cover described in Section 3.10.2 is recommended where57cable must be cut underground. Utility knives are frequently used to cut grout or breather tubeand open cement bags. High pressure tubing is not easily pierced with a utility knife and the useof a hacksaw is recommended as a much more effective and safer cutting method.3.10.5 Grout PressureBourchier, Dib, and O'Flaherty (1992) describe bursting of the hoses due to back pressureat the pump hose-grout tube connection. Bursting of hoses and connections have been observedand frequently reported in western Canadian mines The direct coupling of the pump outlet hoseto a grouting assembly has been suggested (Bourchier, Dib, and O'Flaherty 1992) to eliminatethe safety hazard of bursting connections. The grout tube method can be used to reduce groutpressure, and some operators have adopted the use of high pressure (1.7 MPa) tubing to reducethe occurrence of hose bursting. High pressures at the hole collar generated with the breathertube method can induce forces in fractured ground that encourage the release of loose rock.Proper scaling practice at the start of each shift will minimize the occurrence of ground falls dueto grouting. Plate tensioning and the insertion of bent wire anchors can also induce ground falls.After grouting is completed, the line pressure will remain high and care must be taken whendisconnecting the pump hose from the grout tube.3.10.6 ManpowerCable bolting is typically a high turnover job since it is not often viewed as desirablework, and is usually associated with a nominal wage and bonus. Personnel are generally notveteran mine employees and have accumulated minimal training. An installation manual shouldbe developed and combined with a practical training program. Cable bolting requires a highdegree of quality control, and it is beneficial to retain experienced cable bolting personnel as long58as possible. Management could consider incorporating the drilling function into the jobdescription of the cable bolt crew as a way of providing additional drilling responsibility, andperhaps, a source of improved bonus potential. Bonus payments for cable bolting must carefullyweigh productivity against performance. Cable bolt quality control is difficult to check and highproductivity can often lead to a reduction in installation quality.3.11 COSTSA detailed review of cable bolt costs will not be pursued in this thesis but some ideas ofthe range in component prices is given in Table 3.6. It is recommended that a cable supplier becontacted to provide cost estimates for planning purposes. Cable bolt costs are usually includedin a general ground support account. Consideration should be given to establishing a charge codeto which cable bolt labour, materials, and maintenance may be assigned. A detailed financialrecord is then available for future cost analysis. Cable bolt costs in 1991 Canadian dollars rangedfrom $19.00/m to $36.00/m for the mines visited in this study. The average cost per meter ofinstalled cable was approximately $27, but different accounting structures and cable componentsmake it difficult to compare costs directly.3.12 DISCUSSIONThis chapter has been designed to briefly review the major components that should beconsidered in the development of a cable support program. Observation of cable performance canbe used as a guide to modify current practice. MacSporran, Bawden, and Hyett (1992) relatevisual cable observations after failure to an estimate of the maximum load. Undisturbed cables59Table 3.6: Summary of Cable Component Costs (1992 Canadian Dollars)COMPONENT COSTStandard 15 9 mm Cable $1.61 to $1.84 per meterBirdcage Cable $2.95 to 3.05 per meterBarrel & Wedge Anchor $4.25 to $7.50 eachPlates (152 x 152 x 9 5 mm) $1.65 to $2.45 eachEnd Holding Device (ferrule installed) $2.00 to $3.30 eachExpansion Shell Anchor $20.00 to $40.00 eachCable Button (installed) $1.50 eachGrout Tube $0.49 to $1.34 per meterBreather Tube $0.30 to $0.52 per meterSpedel 6000 Pump and B3100 Mixer $9000 eachMinepro 3 Grout Pump:Air/Hydraulic skid mountedElectric/Hydraulic skid mountedHost Hydraulic$29000 each$31000 each$25000 eachleft hanging after a failure are indications of load carrying capacities in the range of 0 to 5tonnes. Where unravelling of the lay occurs, the load is estimated at 5 to 15 tonnes. Loadsbetween 15 and 25 tonnes are suggested where cable ends are pigtailed. Pigtailing has beencoined as a descriptive term applied to failed cables that show evidence of having attainedsignificant load carrying capacity prior to failure. The term is analogous to the effect of curlinga strip of paper by pulling it through ones fingers. Pigtailed cables are typically unravelled andthe individual wires are curled. This can be a useful measure of cable load, but observations inthis study have occasionally indicated the presence of pigtailing with no apparent rock failure.This is typically encountered in hangingwalls where the action of blasted ore moving towards thedrawpoint can result in pigtailed cables. The only true measure of attaining the maximum load60carrying capacity is observation of steel failure. Frequent observation of undisturbed cables aftera failure situation is evidence of the need to revise current practice. Design applications will beintroduced in Chapter 4 and should be complemented with observations of cable boltperformance.61CHAPTER 4CABLE BOLT DESIGN METHODS4.1 INTRODUCTIONThis chapter will review cable bolt design methods with particular reference to westernCanadian practice. Discussions with different mine personnel have indicated that there is nostandard cable bolt design procedure used in western Canadian mines, but there is a strong desirefor some consistent criteria. The flowchart in Figure 4.1 proposes a methodology of cable designthat incorporates current practice and will be used later in this thesis to propose revisions. Theflowchart is split between discrete and collective analysis, reflecting a distinction betweenassociated design techniques. Discrete analysis is applied to cases of isolated blocks or structurethat require support. The discrete design method has been well defined by several authors (Hoekand Brown 1980, 246-248; Stillborg 1986, 58-62) and will be briefly reviewed in this chapter.The collective analysis segment of the flowchart deals with the in-situ rock mass and incorporatesdesign methods that reflect this approach. One of the most important considerations on both sidesof the flowchart is access. In this study, only 2 % of the supported case studies involveddevelopment that was specifically designed for the implementation of support. With this in mind,it is important to consider cable design in relation to the available access and the associatedpatterns that can result. The collective analysis approach incorporates a number of designtechniques reflecting the rock mass as a continuum. These techniques are frequently applied inclose collaboration with each other to obtain a final design. An economic analysis is normallyperformed and may indicate a review of the entire process prior to implementation. A successfuldesign and economic analysis leads to implementation of the support system. The support system62Figure 4.1: Cable design flowchart63performance is then observed and serves as input into subsequent design.4.2 DISCRETE ANALYSISDiscrete analysis is reviewed by Hoek and Brown (1980, 246-248) in relation to a specificblock or wedge that is either free to fall or slide. The bolt load required to support a block thatis free to fall is given by(4.1) N_ W x FTwhere^N = number of boltsW = weight of wedgeT = bolt loadand^F = factor of safety.For cases where the block is free to slide, as shown in Figure 4.2, the frictional resistance of thesliding surface must be considered using the following relation,(4.2)^ cA + (Wcosijr + TcosO)tan4Wsintly - Tsin9whereandc = sliding surface cohesive strengthA = surface area of the sliding surface* = dip of the sliding surface0 = angle between bolt and normal to the sliding surface(t) = friction angle of the sliding surface.Discrete Analysis -Sliding Block65Equation 4.2 can be rearranged to determine the required bolt load as shown in Equation 4.3.(4.3)^T = W(Fsimp - cosirtan4) - cAcos13tan4 + FsinOHoek and Brown (1980, 247) recommend that a factor of safety of 1.5 be used where groutedbolts or cables are used, and bolt length be based on suitable anchorage beyond the blockboundaries. Where sliding occurs along two planes, Hoek and Brown (1980, 248) suggest thatthe dip of the line of intersection be used in the support analysis. Discrete analysis is dependenton good structural definition of the block boundaries in order to estimate the block weight. Thebolt load used in equation 4.2 to determine a factor of safety may not relate to the cable breakingstrength unless the critical embedment length is realized, or plates are incorporated at the holecollar. Individual cable capacity used in equation 4.2 should reflect the embedment length andgrout water:cement ratio. Miller (1984) has shown that the most efficient shear resistance isoffered by cables oriented parallel to the shear direction and inclined from 17° to 27°.4.3 COLLECTIVE ANALYSISHelping the rock mass to support itself is described by Hoek and Brown (1980, 244) asthe principal objective of underground support. Lang (1961) referred to rockbolting as "thedesigned use of rock bolts to reinforce and develop the rock around an excavation into astructural entity". This concept takes advantage of the inherent strength of a rock mass and wasfirst applied to cable bolts in cut and fill mining Fuller (1981) describes the concept of pre-reinforcement that was successfully adapted to initial cable bolting applications. Experience withrockbolting had demonstrated that pattern support was most successful where installation occurred66immediately after mining With cable bolts, it was possible to install support prior to mining andtake advantage of the inherent rock mass strength by limiting joint dilation. Most cable designmethods have been developed on the basis of an even distribution of bolts on a regular squarepattern. Where access is restricted, an even pattern may not be possible, and bolts are frequentlyinstalled in a high density fan to act as a point anchor. The point anchor approach differs frompattern bolting as large areas of the supported surface are left unsupported. Discrete analysis hasbeen described in terms of isolated structure, and is specific to a particular situation. This thesisis concerned with a collective design approach that has the ability to reflect characteristics of theentire rock mass. This section will review pattern and point anchor collective design methods thatwere encountered in practice4.3.1 Dead Weight DesignDead weight design was encountered when cables were required to support the deadweight of a rock mass. This method involves the estimation of the tonnage of rock to besupported, and the use of a factor of safety to determine the number of bolts required, asreviewed in Section 4.2. The number of bolts are typically distributed evenly over the areainvolved, and extend 1 to 4 meters beyond the projected failure plane. Where the thickness ofthe supported area is variable, the bolt pattern should be adjusted to reflect higher loads inthicker areas. A typical factor of safety of 1.2 was incorporated in the final design andoccasionally the bolt tensile strength was reduced if the cable orientation was not vertical. Allof the designs utilizing a dead weight analysis in this study, assumed that the load carryingcapacity of the cable was equivalent to the steel tensile strength.The determination of the weight of rock to be supported is based on defining a zone ofexpected instability. This is not a well defined procedure unless a fault or shear zone isolates a67area of the rock mass. Stillborg (1986, 66-68) describes the formation of a natural arch abovean opening due to stress redistribution as the opening is created. A zone of 'loose' rock belowthe arch is subject to instability, and can be used to define the dead weight load that cables mustsupport. If the arch area is assumed to approximate a triangular shape, the bolt pattern can thenbe calculated using the following relationship,(4.4) s2 _ ^2T FxHxywhere^T = bolt load in tonnesS 2 = bolt square pattern (S x S) in metersH = height of relaxed zone below archand^y = unit weight of the rock mass in tonnes/meter 3 .This method assumes that the full bolt tensile strength is utilized, and there is competent groundabove the relaxed zone. The height of the natural arch formed, H, can be determined throughnumerical modelling or through the use of empirical formulations that relate the bolt length tothe opening span. Bolt length is typically determined by designing for at least 4 meters ofanchorage into the natural arch. Stheeman (1982) describes a procedure used at the Tsumeb Mineto mathematically approximate the curvature of a natural arch above a cut and fill stope. Thecurved shape of fracture planes forming the walls of a cavity after a rock fall, indicated the limitsof the self-supporting arch. Fracture plane angles were noted at different positions and used toderive an elliptical relationship that supported the observations. The relationship was used toestimate the volume of rock below the natural arch that required cable support. The number ofcables required was determined by dividing the weight of rock below the natural arch, by theestimated breaking strength of the cable. The calculated cable density was subsequently increased68to reflect a safety factor of 1.2. Bolt length was based on the coverage of five mining lifts, theheight of the natural arch, the bond length required to support the weight below the arch, andan allowance for grout bleed.4.3.2 Rock ClassificationSeveral methods of rock mass classification have been developed to assess the quality ofa rock mass and estimate stand-up time or support requirements. The Q-system of rock massclassification (Barton, Lien, and Lunde 1974) describes the rock mass in terms of six parametersas follows,RQD Jr Jw(4.5)^Q^x ^xJn^Ja SRFwhere^Q = rock mass qualityRQD = rock quality designationJn = joint set numberJr = joint roughness numberJa = joint alteration numberJw = joint water reduction factorand^SRF = stress reduction factor.A full description of each parameter is given in Table 4.1, but a review of the original text(Barton, Lien,and Lunde 1974) is recommended. The rock quality designation, or RQD, isdefined as the percentage of core recovered in intact lengths greater than 100 mm (Hoek andBrown 1980, 18). The rock mass quality, Q, increases on a logarithmic scale from 0.001 to 1000with increasing rock quality. In the initial Q-system proposal, Barton, Lien, and Lunde (1974)Table 4.1: Description of Q-system Parameters (Barton, Lien, and Lunde1974)1. ROCK QUALITY DESIGNATION (RQD)A. Very Poor 0-25 Note:(I) Where RQD isreputed or measured as5 10 (including 0) anominal value of 10 isused to evaluate Q.(ii) RQD intervals of 5,i.e. 100, 95, 90, etc. aresufficiently accurateB. Poor 25-50C. Fair 50-75D. Good 75-90E. Excellent 90-1002. JOINT SET NUMBER (Jn)A. Massive, no or few joints 0.5-1.0 Note:(i) For intersections use(3.0 x Jn)(11) For portals use(2.0 x in)B. One joint set 2C. One joint set plus random 3D. Two joint sets 4E. Two joint sets plus random 6F. Three joint sets 9G. Three joint sets plus random 12H. Four or more joint sets, random, heavily jointed, "sugar cube like",etc. 15J. Crushed rock, earthlike 203. JOINT ROUGHNESS NUMBER (Jr)(a) Rock wall contact and(b) Rock wall contact before10 ems shearA. Discontinuous joints 4 Note:(i) Add 1.0 if the meanspacing of the relevantJoint set is greater than 3m(u) Jr=0.5 can be usedfor planar slickensidedJoints having lineations,provided the lineationsare favourably orientatedB. Rough or irregular, undulating 3C. Smooth, undulating 2D. Slickensided, undulating 1.5E. Rough or irregular, planar 1.5F. Smooth, planar 1.0G. Slickensided, planar 0.5(c) No rock wall contact when shearedH. Zone containing clay minerals thick enough to prevent rock wall contact 1.0 (nominal)J. Sandy, gravelly or crushed zone thick enough to prevent rock wall contact 1.0 (nominal)69Table 4.1 (con't): Description of Q-system Parameters (Barton, Lien, andLunde 1974)4. JOINT ALTERATION NUMBER (Ja) tor (approx.)(a) Rock wall contactA. Tightly healed, hard, non-softening, impermeable filling i.e. quarts or epidote 0.75 (-) Note:(1) Values of (m),, theresidual friction angle,are intended as anapproximate guide to themineralogical propertiesof the alterationproducts, if presentNote:(i) For intersections use(3.0 x in)(ii) For portals use(2.0 x Jn)B. Unaltered joint walls, surface staining only 1.0 (25°-35°)C. Slightly altered joint walls. Non-softening mineral coatings, sandy particles,clay-free disintegrated rock etc.2.0 (25°-30°)D. Silty-, or sandy-clay coatings, small clay-fraction (non-softening) 3.0 (20°-25°)E. Softening or low friction clay mineral coatings, i.e. kaolinite, mica. Alsochlorite, talc, gypsum and graphite etc., and small quantities of swelling clays.(Discontinuous coatings, 1-2 mm or less in thickness)4.0 (8°-16°)(b) Rock wall contact before 10 cms shearF. Sandy particles, clay-free disintegrated rock etc. 4.0 (25°-30°)G. Strongly over-consolidated, non-softening, clay mineral fillings. (Continuous, <5 mm in thickness)6.0 (160-240)H. Medium, or low over-consolidation, softening, clay mineral fillings.(Continuous, < 5 mm in thickness)8.0 (12°46°)J. Swelling clay fillings, i.e. montmorillonite (Continuous, <5 mm in thickness).Value of Ja depends on percent of swelling clay-size particles, and arcros towater etc.8.0-12.0 (6°-12°)(c) No rock wall contact when shearedK,L,M.Zones or bands of disintegrated or crushed rock and clay (see G, H, J fordescription of clay condition)6.0, 8.0 or8.0-12.0(6°-24°)N. Zones or bands of silty- or sandy-clay, small clay fraction (non-softening) 5.00,P,R.Thick, continuous zones or bands of clay (see G,H,J for description of claycondition)10.0, 13.0 or13.0-20.0(6°-24°)5. JOINT WATER REDUCTION FACTOR (Jw) Approx. waterpressure(kg/cm2 )A. Dry excavations or minor inflow, i.e. , 5 I/min. locally 1.0 <1Note:(i) Factors C to F arecrude estimates. IncreaseJw if drainage measuresare installed.(ii) Special problemscaused by ice formationare not consideredB. Medium inflow or pressure, occasional outwash of joint fillings 0.66 1.0-2.5C. Large inflow or high pressure in competent rock with unfilled joints 0.5 2.5-10.0D. Large inflow or high pressure, considerable outwash of joint fillings 0.33 2.5-10.0E. Exceptionally high inflow or water pressure at blasting, decaying with time 0.2-0.1 >10.0F. Exceptionally high inflow or water pressure continuing without noticeable delay 0.1-0.05 >10.070Table 4.1 (con't): Description of Q-system Parameters (Barton, Lien, andLunde 1974)6. STRESS REDUCTION FACTOR (SRF)(a) Weakness zones intersecting excavation, which may cause loosening of rock mass whentunnel is excavatedNote:(i) Reduce these values of SEEby 25-50% if the relevant shearzones only influence but do notintersect the excavation(ii) For strongly anisotropic stressfield (if measured):when 5 50 1103 510, reduce acand at to 0.8ac and 0.8at ;when 0 1/03 >10, reduce ac andat to 0.6ac and 0.6at where:ac = unconfined compressivestrength, at = tensile strength(point load), o f and 03 = majorand minor principal stresses(m) Few case records availablewhere depth of crown belowsurface is less than span width.Suggest SRF increase from 2.5 to5 for such cases (see H)A. Multiple occurrences of weakness zones containing clay or chemically disintegrated rock, veryloose surrounding rock (any depth)10.0B. Single weakness zones containing clay, or chemically disintegrated rock (depth of excavation 550 m)5.0C. Single weakness zones containing clay, or chemically disintegrated rock (depth of excavation >50 m)2.5D. Multiple shear zones in competent rock (clay free), loose surrounding rock (any depth) 7.5E. Single shear zones in competent rock (clay free) (depth of excavation 5 50 m) 5.0F. Single shear zones in competent rock (clay free) (depth of excavation > 50 in) 2.5G. Loose open joints, heavily jointed or "sugar cube" etc. (any depth) 5.0(b) Competent rock, rock stress problems gc/01 crti°1H. Low stress near surface >200 >13 2.5I. Medium stress 200-10 13-0.66 1.0K. High stress, very tight structure (Usuallyfavourable to stability, may beunfavourable to wall stability)10-5 0.66-0.33 0.5-2.0L. Mild rock burst (massive rock) 5-2.5 0.33-0.16 510M. Heavy rock burst (massive rock) <2.5 <0.16 10-20(c) Squeezing rock; plastic flow of incompetent rock under the influence of high rock pressuresN. Mild squeezing rock pressure 5-100. Heavy squeezing rock pressure 10-20(d) Swelling rock; chemical swelling activity depending on presence of waterP. Mild swelling rock pressure 5-10R. Heavy swelling rock pressure 10-157172recommend different support ranges as illustrated in Figure 4.3. In this chart, the equivalentdimension, De , is related to the rock mass quality, Q. The equivalent dimension relates the sizeof an excavation to its purpose, and is defined as the span, diameter, or height of the openingdivided by the excavation support ratio, ESR. The span or diameter is used in relation to backsupport, and height is related to wall support. The excavation support ratio is higher for mineopenings than for underground civil chambers, and reflects the increased tolerance for instabilityin underground mines. Barton, Lien, and Lunde (1974) suggest an ESR of 3 to 5 for temporarymine openings, 1.6 for permanent mine openings, and 1.0 for major road or railway tunnels.Each numbered region in Figure 4.3 represents a different level of support, and generally reflectincreasing support requirements with higher numbers. Several regions in Figure 4.3 recommenduntensioned, grouted pattern support, which has been related to a cable bolt pattern design. Thedesign length is typically based on extending past some critical feature with suitable anchorage.It is important to note that the support recommendations from this method are based on a largedatabase, but only two cases are noted for temporary mine openings. The majority of the casestudies were collected from permanent mine or civil excavations, and this method can only berecommended for an initial pass at cable bolt design for mining blocks.The Geomechanics Classification of Jointed Rock Masses (Bieniawski 1976) is a methodof classifying rock for engineering purposes through the estimation of a rock mass rating, orRMR. The RMR is a rating between 0 and 100 that is based on the intact rock strength, drillcore RQD, joint spacing, joint condition, and the presence of groundwater. The classificationsystem is described in detail in Hoek and Brown (1980, 22-27) and has been related to rocksupport by Bieniawski (1976). It has not been directly applied in this thesis since existingmethods of cable design are associated with the Q-system of rock classification. Bieniawski(1976) proposed the following relationship between Q and RMR based on the regression analysis100Ex;112z10EXCEPTPO10NARLLY0GOODVERYGOOD1,0001001 100.10.01Tunnel Support Chart(After Barton, Lien, and Lunde 1974)somas momsMON^momIIIMO1191111111111 7^111^101 MLIIIIN•1111W16oF1111111 MI' Eli ISM3214381=11111=11111EXTREMELYPOOR35VERYPOOREXT.GOODEXCGOOPOOR FAIRPRP NON^• ORMiiult Imp /a 111111■MN_ —.1011 111 EINM 111M IMINE15D...MMI1111•111•11111MrilINIMILM•110_•.!: iioN111111111/1111111 MIAMI!EWEN7Mill P!111111111" 111117.11 0'1zz1Hz^11■• mill^1111111137362933NO UP 01111■111^1■7 21^literAMIM 111111^1111PC' -^III^111111 1111^MIIII1IRED0.10.001ROCK MASS QUALITY Q = RQD x Jr X JwJn^Ja SRFmum^ New% iiimemmum Pnigsum msimP:idermso^mineMINI^31 PoMifila NIMM•Po.r.r■ OMNI •111/./1•11111=11111•11 MEINMINIM PP■11111■1^MININKOMOMMOMMINI74of a series of case histories,(4.6)^RMR = 91nQ + 44.Equation 4.6 can be used to convert between Q and RMR but the large data scatter (Bieniawski1976) around this line should be noted.4.3.3 Beam TheoryBeam theory has been applied to cable bolt design where structure is parallel to thesurface to be supported. This is most commonly applied in a layered rock mass, where cablesare used to tie a number of layers together into a stable beam. The design method is similar tothe arch concept discussed in Section 4.3.1, and is based on using the dead weight of the beamto determine cable spacing. The volume of rock is approximated as a rectangular area andequation 4.4 becomes,(4.7) S2 - ^FxLxywhere L = the beam thickness.Fuller (1983b) describes a method of cable design for a point anchor cable pattern thatis based on beam theory. A hangingwall in layered rock was assumed to behave like a beam, asillustrated in Figure 4.4. Cable fans installed into the hangingwall at two intermediate sublevelsare simulated by reactions, R. The stope abutments were fixed, but the analysis allowed forsome beam deflection at the intermediate points to represent the true action of cables taking load.In this analysis, Fuller (1983b) assumed that the beam experienced uniformly distributed loading80^0.5^1^1.5^2^2.5^3Beam Thickness (m)3.5 4Beam Approach to Hangingwall DesignY6D)0Itsm4045Cl)a)15"(U20S.G. = 4.0E = 80 GPaBeam Length = 90 mPc based on pull testsDip = 45°t = beam thickness(After Fuller 1983b)ked cables with lowerMc:bistiifiness76based on a rock specific gravity of 4.0, and related the cable reactions to laboratory pull testcurves for different bond lengths of 15.2 mm diameter double cables. The results of this analysisrelate the number of cables required along strike to the beam thickness, as illustrated in Figure4.4. A minimum bolt density of 4 cables for each meter of strike length was determined for thisparticular stope. Beams less than one meter in thickness were found to fail in tension. Cases ofhangingwall support collected from western Canadian experience typically reflect major structureparallel to the surface. In most cases however, there is usually a second minor joint set thatencourages failure between sublevels, and limits the success of point anchor bolting.4.3.4 Past ExperiencePast experience was the most common design method encountered in Western Canada.This method relies on an initial trial and error process with subsequent adjustments based onperformance. In many cases, the initial design is based on typical patterns that are used at otheroperations. Observation of support performance establishes a database of internal case histories,that can then be used to modify cable design. One operation used past failure situations todetermine the opening span at which cable support was required. Larger spans dictated the useof a tighter cable pattern. This method of design was most noticeable where cut and fill miningwas in progress. The cables would be installed at lengths of up to 18 meters to cover severalmining lifts. In these cases, the performance of a cable pattern can be evaluated fairly easily,since cut and fill lifts are mined rapidly relative to open stoping situations. This approach wasalso encountered to design open stope back support, although the initial pattern selection wasmuch more critical. Modifications to the installation after mining had begun were difficult if notimpossible. Cables were required in the back if the span exceeded an amount determined byexperience. This critical span was reduced if the quality of the rock mass deteriorated. In some77cases, the cable bolt spacing was determined by the blasthole ring burden, so that bolt holescould be drilled from existing set up points.Instrumentation can be used to record support performance and develop design criteria.Fuller (1981) describes the monitoring of cables in the immediate back of a cut and fill stope.The strain distribution immediately after a blast, indicated a tensile load in the first 5 meters ofcable above the back. Approximately one hour after the blast, the tensile load range decreasedto 2 meters from the stope back, and conformed the opinion that cable support helps the rockmass to support itself. Similar observations resulted from an instrumented 2 m x 2 m cable grid,but the ground became self-supporting within minutes of the blast. Based on this experience,Fuller (1983) recommended a 2 m x 2 m pattern for design with a minimum cable overlap of2 meters. Choquet and Miller (1988) describe the development of a tension measuring devicespecifically designed for cable bolts. The technique involves the measurement of resistancevariation in a thin wire that is wound along a cable strand, and isolated from the grout.Measurements using this device can provide an estimate of cable strain and actual load carryingcapacity. It is however highly dependent on location with respect to dilation within the rockmass. Since the wire in the tension measuring device is free along its entire length, cable strainmay be significantly higher than the measuring device might indicate.Hunt and Askew (1977) describe the use of empirical design guidelines developed by theSnowy Mountains Authority for use in the design of cable patterns for wide permanent openings.In this situation the bolt length, L, was determined by the minimum of Equation 4.8,(4.8)^L = 1.83 + 0.013D 2or three times the spacing of the smallest joint set, where D represents the opening span in78meters. Stillborg (1986, 69) describes empirical design recommendations formulated by the U.SCorps of Engineers for use in rockbolt design. These are occasionally adapted for use in cabledesign and suggest that the minimum bolt length should the greatest of:a) 2 x the bolt spacingb) 3 x the thickness of critical and potentially unstable blocksc) For spans less then 6 m: 0.5 x spanFor spans between 18 and 30 m: 0.25 x spanFor spans between 6 and 18 m: interpolate between 3 and 4.5 meter lengths.The spacing is based on half the bolt length or 1.5 times the width of critical and potentiallyunstable blocks, up to a maximum of 2 meters. There are many similar empirical designrecommendations, but most are specifically based on rockbolting experience in tunnellingapplications. The applicability of these design criteria to cable bolting is limited, but they areuseful as initial guidelines.4.3.5 The Mathews Method and Bolt FactorMathews et al. (1981) developed an empirical relationship between the stability number,N, and the shape factor, S, of a stope surface. The stability number can be evaluated by(4.9)^N=Q'xAxBxCwhere^Q' is the Q-system rock mass rating with the stress reduction factor set to oneA is the stress factor,B is the rock defect orientation factor,and^C is the design surface orientation factor.The shape factor is also called the hydraulic radius, and is determined by dividing the surface79area by the perimeter. It was originally used in mining applications by Laubscher (1976) to relatethe undercut area of a stoping block to cavability. The terms span and length are often appliedin open stope terminology, where the span of a surface can be defined as the minimumdimension, and length refers to the maximum dimension. A traditional longitudinal stopingsequence relates length to the distance along strike, and span to the orebody width. Hydraulicradius considers both the span and length of a particular surface. For square openings, hydraulicradius is one fourth of the span, but as the ratio of span to length decreases, the hydraulic radiusconverges to half the span. Mathews et al. (1981) proposed a design chart (Figure 4.5) thatrelates the stability number to the shape factor, and defines zones of stability, potentialinstability, and potential cave. These zones are further described as follows,(i) Stable - the excavation will stand unsupported with occasional localized ground supportto control slabbing.(ii) Unstable - the excavation will experience localized caving but tend to form a stablearch. Cable bolts and modification of the extraction sequence are suggested as ways tomake open stoping feasible in this region.(iii) Caved - the excavation will not stabilize until the void is full.The factors A, B, and C are presented graphically on a series of charts in Figure 4.6. The rockstress factor, A, replaces the SRF in the Q-system and is related to a s/v„ the ratio of the uniaxialcompressive strength of intact rock, to the induced compressive stress parallel to the surfaceunder consideration. Cases of high induced stress will reflect a lower a chy, ratio and an overallreduction in the stability number through a reduction in the factor A. The rock defect orientationfactor, B, is based on the orientation of the most persistent joint set with respect to the stopesurface. Structure perpendicular to the surface reflects the most favourable orientation, and isgiven the highest rating. The design surface orientation factor, C, is based on the assumption thatShape Factor (S) = Area/Perimeter (m)80Figure 4.5: Stability graph proposed by Mathews et al. (1981)0.8200^806040Mathews Factors A, B, and C(After Mathews et al. 1981)s.0C0.80.60.40.2oo108642Zone of pote atial instability ( — < 2)Qi5 10^15^20ac /a1ORIENTATION^FACTOR B^ORIENTATIONOF ROOF OF WALL110.31.0Angle of Dip from Horizontal (degrees) 0.5O82a vertical wall is eight times as stable as a horizontal surface under the effects of gravity.Bawden et al. (1989) present an attempt to modify rock classification results to reflect theinstallation of cable support. The bolt factor, BF, is determined by dividing the total length ofinstalled cables by the area of the supported surface. The bolt factor is then related tounsupported and supported Q' values as shown in Figure 4.7. It is important to note that thechart in Figure 4.7 represents a proposed design method that has received little practicalcalibration, and therefore is not recommended for design. However, the idea of increasing rockmass quality to account for support is a valuable concept. The initial phase of the bolt factorapproach, involves completing a stability analysis for the stope surface using the techniquedeveloped by Mathews et al. (1981). If the surface plots within the stable zone, cable supportis not required. If potential instability or caving are indicated by the stability analysis, a boltfactor is selected to estimate a supported Q' value using Figure 4.7. The selection of the boltfactor is based on establishing a supported Q' that can be related to the prediction of a stablesurface in the original stability analysis.4.3.6 Design Based on Rock Mass StiffnessFuller, Dight, and West (1990) describe a design technique that was developed under thecoordination of the Australian Mineral Industries Research Association (AMIRA). The methodrelates the rock mass deformation modulus to cable load-displacement curves obtained fromlaboratory testing. Section 2.8 discussed an increase in cable load carrying capacity withincreasing rock mass stiffness. The Mathews method is used to evaluate initial stability and theneed for additional support. The in situ rock mass modulus is estimated from a relationship withthe Q classification system described by Hoek and Brown (1980). The size and displacement ofa potential failure zone is estimated from numerical modelling or past experience. The specific10 100101BOLT FACTOR METHOD(After Bawden et al. 1989)wiwmor....wp•-•11^...Pageorica■• 11/...Mrjerti/IMMINSIN -e'riwir . mippl,_......4.4"gri_...  ■Illiral____=A1-..amaamiplamraill...-_,...Jal.---_—_.--^.....--_—.^....mii.1,—.... i. ..m.m.......-- ....PP%1PM111111/11121 NIWW I"BF = Bolt Fa for (me ers c ble/BF = 8BF = 7BF = 6BF = 5BF = 4BF = 3BF = 2BF = 1Q' (UNSUPPORTED)100084load-displacement curve for a particular rock modulus provides an estimate of the cable loadcarrying capacity that can be used for design. The load carrying capacity is then related to thesize of the failure zone to determine the number of cables required. A factor of safety of between1.2 and 2.0 is suggested and cable length is based on extending 3 meters beyond the deepestpoint of any potential failure. This method has not been widely used in Canadian practice dueto the difficulty of estimating the size and allowable displacement of potential failure zones.Reichert, Bawden and Hyett (1992) suggest a design method that is based on therelationship between embedment length and radial stiffness illustrated in Figure 4.8. The solidline in Figure 4.8 is based on a linear extrapolation of the relationships between load andembedment length in Figure 2.14. The extrapolated values of embedment length are taken at 24tonnes to reflect an optimum load carrying capacity. This relationship is based on laboratorytesting at a 0.30 water:cement ratio and additional curves were estimated for higher ratios. Interms of design, an estimate of radial stiffness and embedment length will provide an indicationof expected cable performance.4.3.7 The Potvin MethodThe Modified Stability Graph (Potvin 1988) was developed empirically from the analysisof 175 case histories of open stoping situations in Canada. Based on these case studies and theconcepts behind the Mathews method, a stable and a caving zone were identified by relating amodified stability number, N', to the hydraulic radius, HR, of the surface. The value of thestability number is determined in a similar procedure to the Mathews method (Mathews et al.1981), but the term N' is used to signify the use of slightly different A, B, and C factors. Potvin(1988) suggested that the first stage of the cable bolt design process was to complete a stabilityanalysis to determine the expected stability condition of the surface. A zone within the cavingZ *4 171*-< olZ 6.4 .CD ==^I-Ib.., CD CDv.:) rz'' 4=..vD ,•-t- •ts..) 0 00• 5 vO cpIT ADFr] oIN..) =4., l•-•"c)an a'5^-(1)rn• nco0 =i-sa 'illpp cl,cr .11'O" a. ,..„Q., 5.,a eD/..... 0ZC nl cia aco cm.17) 5coa cp..•L7 r.",--, c7> 3P>pvCDa co• ,-t-. aNItXJ iioo 0• 5a cpa ze-t-* ilw rl:= Z-.a086region was identified on the Modified Stability Graph to be stable with the addition of cablesupport. Assuming that cable support was required, the design density could be determined fromthe Design Chart for Cable Bolt Density (Potvin 1988). This chart relates the cable density toa relative block size factor, RQD/Jn/HR, and was developed based on the analysis of 66 casehistories of stope backs that were supported with cable bolts. The Modified Stability Graph wasfound to be used at about half of the operations visited. The Design Chart for Cable Bolt Densitywas rarely used for fmal design, but was consulted in some cases prior to determining the desiredpattern. Chapter 5 will discuss the Potvin Method in more detail.Table 4.2: Design Methods in Western Canadian PracticeDead Weight 9%Rock Classification 4%Beam Theory 4%Past Experience 71 %Bolt Factor 5%Potvin Method 7%4.4 WESTERN CANADIAN PRACTICEFifty-six design applications were reviewed during the mine visits conducted in this study.A summary of the different design methods encountered is given in Table 4.2. Experience wasfound to be the most frequent design method and reflects the lack of acceptable guidelines forcable support. A combination of several design methods is often used to develop a cable supportproposal.87CHAPTER 5THE POTVIN METHOD5.1 INTRODUCTIONThis chapter will review the method of stope design that was proposed by Yves Potvinin his PhD thesis at the University of British Columbia (Potvin 1988). Forty mines were visitedin 1986 and 1987, with the objective of back analyzing the stability condition of open stopes tocreate a database, and subsequently develop a model for stope design. The model database,assembled from 34 mines, consisted of 176 case histories of unsupported stope surfaces and 66supported surfaces. Five stages of development were involved in creating the model. First,geotechnical parameters were assembled from mine visits. These parameters were subsequentlyexpressed in terms of factors to form a geotechnical model. The Mathews method (Mathews etal. 1981) was chosen as the best method of design analysis since it is based on case histories ofopen stoping situations. Rock mass performance was related to the stability condition of thesurface, and the final stage was to calibrate the effect of each factor by back analysis.5.2 THE MODIFIED STABILITY GRAPHThe Potvin (1988) case histories were divided into a main and complementary database.The main database contained 84 case histories collected during visits to different miningoperations. The complementary database contained 92 case histories that were collected fromliterature or involved some uncertainty in one or more parameters. A summary of the main andcomplementary Potvin database can be found in Tables 5.1 and 5.2 respectively.Table 5.1: Summary of Potvin Unsupported Main DatabaseCase#SurfaceHydraulicRadius(m)Q' A B C N' Stability1 HW 5.0 54.0 1.0 0.65 6.5 228 Stable2 Wall 8.9 6.0 0.2 0.25 2.5 0.7 Unstable3 Wall 7.7 6.0 0.1 0.2 2.5 0.3 Caved4 HW 7.1 10.5 1.0 0.2 3.7 7.8 Unstable5 HW 14.0 40.0 1.0 1.0 8.0 320 Stable6 HW 11.0 40.0 1.0 1.0 8.0 320 Stable7 HW 5.2 40.0 1.0 1.0 6.5 260 Stable8 HW 8.5 9.0 1.0 0.4 5.0 18 Stable10 End 4.7 3.2 0.3 0.2 3.5 0.7 Unstable12 HW 9.1 4.2 1.0 0.2 6.5 5.5 Unstable13 HW 8.3 30.0 1.0 0.2 7.0 42 Stable16 Back 5.8 6.25 0.1 0.85 2.0 1.1 Caved17 Back 4.2 6.25 0.1 0.85 2.0 1.1 Stable18 HW 8.8 30.0 1.0 0.6 8.0 144 Stable19 Back 3.5 30.0 0.1 0.4 2.0 2.4 Unstable20 Back 1.8 16.5 1.0 0.2 2.0 6.6 Stable21 HW 4.7 16.5 1.0 0.2 4.5 15 Stable22 HW 8.8 16.5 1.0 0.2 4.5 15 Stable23 Back 2.1 16.5 1.0 0.2 2.0 6.6 Stable24 Back 10.5 34.0 1.0 0.2 2.0 14 Caved25 Back 11.3 34.0 1.0 0.2 2.0 14 Caved26 Back 12.2 34.0 1.0 0.2 2.0 14 Caved27 Back 4.1 34.0 1.0 0.2 2.0 14 Stable28 Wall 7.6 12.0 1.0 0.3 2.0 6.9 Stable29 Wall 7.6 34.0 1.0 0.2 3.0 20 Stable30 HW 9.0 34.0 1.0 0.2 5.0 34 Stable31 HW 16.6 90.0 1.0 1.0 8.0 720 Stable32 Back 4.0 90.0 0.1 1.0 2.0 18 Stable33 HW 23.0 90.0 1.0 1.0 8.0 720 Stable34 Back 10.7 90.0 0.4 1.0 2.0 72 Stable35 Back 10.5 9.0 0.6 0.3 2.3 3.9 Caved36 HW 9.0 9.0 0.9 0.3 5.0 13 Stable53 Back 2.4 43.5 0.5 0.2 2.0 8.8 Stable54 Back 6.8 43.5 0.5 0.2 2.0 8.8 Caved55 Back 8.0 43.5 0.5 0.2 2.0 8.8 Caved56 Wall 19.0 2.0 1.0 0.3 8.0 5.2 Caved57 Back 3.7 43.5 0.2 0.2 2.0 3.5 Stable58 Wall 8.4 43.5 1.0 1.0 8.0 352 Stable59 Wall 4.5 2.0 1.0 0.3 8.0 5.2 Stable61 HW 7.5 25.5 1.0 0.3 6.0 45 Stable62 FW 7.5 25.5 1.0 0.3 4.0 30 Stable132 HW 5.6 6.0 1.0 0.2 8.0 10 Stable133^_ HW 6.7 6.0 1.0 0.2 8.0 9.4 Stable88Table 5.1: Summary of Potvin Unsupported Main Database (con't) Case#SurfaceHydraulicRadiusOn)Q' A B C N' Stability134 Back 1.9 5.0 0.1 0.2 2.0 0.2 Stable135 Back 2.1 26.0 0.6 0.6 2.0 19 Stable136 Back 2.4 26.0 0.5 0.6 2.0 16 Stable137 Back 2.9 26.0 0.4 0.6 2.0 13 Stable138 Back 3.1 26.0 0.4 0.6 2.0 13 Stable139 Back 3.0 26.0 0.3 0.6 2.0 10 Stable140 HW 7.5 8.0 1.0 0.3 6.0 15 Stable141 HW 8.1 8.0 1.0 0.3 6.0 15 Unstable142 HW 5.3 8.0 1.0 0.2 5.5 9.2 Stable143 HW 5.7 8.0 1.0 0.2 5.5 9.2 Stable144 Back 1.9 8.0 0.1 0.2 2.0 0.3 StableBack 0.3 0.2 2.0 1.0 StableBack • 0.1 0.2 2.0 0.3 UnstableBack 0.1 0.2 2.0 0.3 UnstableBack 0.7 0.2 2.0 5.9 StableHW 1.0 0.2 6.0 0.8 Caved0.2 6.0 0.8 Caved0.2 2.0 4.8 Caved0.2 2.0 12 Stable0.5 8.0 120 Stable0.2 3.0 19 Stable1.0 3.0 10 Stable• 0.2 8.0 26 Stable0.2 2.0 1.0 Stable0.2 2.0 0.6 Caved0.2 8.0 4.8 Caved0.8 2.0 3.3 Caved9.9 0.2 8.0 33 Stable9.9 • 0.2 3.0 8.3 Unstable• 12.5 0.8 2.8 60 Caved15.0 0.8 2.8 60 Caved15.9 0.8 2.8 60 Caved7.7 0.8 2.8 60 Stable5.4 0.8 2.8 60 Stable11.6 0.3 8.0 32 Unstable7.3 0.85 2.5 29 Stable177 9.9 27.0 0.5 0.85 2.5 29 Stable178 Back 11.1 27.0 0.5 0.85 2.5 29 Unstable180 HW 6.9 6.0 1.0 0.3 6.0 10 Unstable183 Wall 4.9 24.0 0.1 0.3 8.0 5.8 Stable184 HW 6.7 6.0 1.0 0.3 7.0 12 Stable89Table 5.2: Summary of Potvin Unsupported Complementary DatabaseCase#SurfaceHydraulicRadiusOn)Q' A B C N' Stability64 HW 6 6.0 1.0 0.3 5.5 10 Stable65 HW 12 6.0 1.0 0.3 5.5 10 Caved66 HW 3 2.4 1.0 0.3 7.0 5.0 Stable67 HW 9 2.4 1.0 0.3 7.0 5.0 Unstable68 HW 12 2.4 1.0 0.3 7.0 5.0 Caved69 HW 16 54.0 1.0 0.3 4.5 73 Unstable70 HW 5 4.8 1.0 0.3 8.0 12 Unstable71 HW 8 4.8 1.0 0.3 8.0 12 Caved72 HW 16 0.25 1.0 0.3 6.5 0.5 Caved73 HW 7 48.0 1.0 0.3 8.0 115 Stable74 HW 2 12.0 1.0 0.3 4.5 16 Stable75 HW 11 12.0 1.0 0.3 4.5 16 Stable76 HW 5 54.0 1.0 0.3 5.0 81 Stable77 HW 14 0.75 1.0 0.3 8.0 1.8 Caved78 HW 6 0.75 1.0 0.3 7.0 1.6 Caved79 HW 10 0.75 1.0 0.3 7.0 1.6 Caved80 HW 11 0.25 1.0 0.3 6.5 0.5 Caved81 HW 9 54.0 1.0 0.3 5.0 81 Stable82 HW 6 2.4 1.0 0.3 5.5 4.0 Unstable83 HW 13 2.25 1.0 0.3 5.0 3.4 Caved84 HW 10 54.0 1.0 0.3 4.5 73 Stable85 HW 4 21.0 1.0 0.3 5.5 35 Stable86 HW 1 60.0 1.0 0.3 4.5 81 Stable87 FW 12 60.0 1.0 0.3 4.5 81 Unstable88 HW 4 16.0 1.0 0.3 8.0 38 Stable89 HW 11 16.0 1.0 0.3 8.0 38 Stable90 HW 3 0.75 1.0 0.3 4.0 0.9 Stable91 HW 11 0.75 1.0 0.3 4.0 0.9 Caved92 HW 2 0.75 1.0 0.3 5.5 1.2 Stable93 HW 7 0.75 1.0 0.3 5.5 1.2 Caved94 HW 9 0.75 1.0 0.3 5.5 1.2 Caved95 HW 16 0.75 1.0 0.3 5.5 1.2 Caved96 HW 8 0.25 1.0 0.3 7.0 0.5 Caved97 HW 3 0.25 1.0 0.3 8.0 0.6 Unstable98 HW 5 0.25 1.0 0.3 8.0 0.6 Caved99 HW 3 16.0 1.0 0.3 5.5 26 Stable100 HW 3 3.0 1.0 0.3 5.0 4.5 Stable101 HW 6 3.0 1.0 0.3 5.0 4.5 Unstable102 HW 14 3.0 1.0 0.3 5.0 4.5 Caved103 HW 3 1.5 1.0 0.3 5.5 2.5 Stable104 HW 8 1.5 1.0 0.3 5.5 2.5 Unstable105 HW 13 1.5 1.0 0.3 5.5 2.5 Caved106 HW 10 30.0 1.0 0.3 6.0 54 Stable107 HW 4 1.6 1.0 0.3 7.0 3.4 Caved108 HW 10 1.6 1.0 0.3 7.0 3.4 Caved109 HW 6 3.0 1.0 0.3 5.0 4.5 Unstable110 HW 12 3.0 1.0 0.3 5.0 4.5 Caved90Table 5.2: Summary of Potvin Unsupported Complementary Database (con't)Case#SurfaceHydraulicRadius (m)Q' A B C N' Stability111 HW 3 1.0 1.0 0.3 6.0 1.8 Stable112 HW 8 1.0 1.0 0.3 6.0 1.8 Unstable113 HW 14 1.0 1.0 0.3 6.0 1.8 Caved114 HW 2 2.4 1.0 0.3 5.5 4.0 Stable115 HW 8 2.4 1.0 0.3 5.5 4.0 Unstable116 HW 10 2.4 1.0 0.3 5.5 4.0 Unstable117 HW 10 6.0 1.0 0.3 5.5 10 Stable118 HW 6 0.25 1.0 0.3 6.5 0.5 Unstable119 HW 9 0.25 1.0 0.3 6.5 0.5 Caved120 HW 1 0.25 1.0 0.3 5.0 0.4 Stable121 HW 2 0.25 1.0 0.3 5.0 0.4 Unstable122 HW 13 0.25 1.0 0.3 5.0 0.4 Caved123 HW 6 0.25 1.0 0.3 5.5 0.4 Unstable124 HW 10 0.25 1.0 0.3 5.5 0.4 Caved125 HW 1 0.25 1.0 0.3 6.0 0.5 Stable126 HW 2 0.25 1.0 0.3 6.0 0.5 Unstable127 HW 13 0.25 1.0 0.3 6.0 0.5 Caved128 HW 7 9.6 1.0 0.3 4.5 13 Stable129 HW 12 1.5 1.0 0.3 5.5 2.5 Unstable130 HW 4 0.25 1.0 0.3 5.0 0.4 Unstable131 HW 3 0.25 1.0 0.3 5.5 0.4 Unstable9 Wall 4.7 24.0 0.3 0.2 8.0 12 Stable11 HW 7.9 3.0 1.0 0.2 7.0 4.2 Stable14 HW 8.8 4.5 1.0 0.2 6.0 5.4 Caved15 HW 8.8 4.5 1.0 0.3 7.0 9.5 Caved154 Back 5.2 32.0 0.1 0.3 2.0 1.9 Unstable167 HW 7.8 13.5 1.0 0.2 8.0 22 Stable168 HW 6 22.5 1.0 0.2 8.0 36 Stable169 Back 5 22.5 0.3 0.3 2.0 4.1 Stable179 Back 4.1 22.5 0.1 0.85 2.0 3.8 Stable181 Back 4 22.5 0.1 0.85 2.0 3.8 Stable182 HW 4.9 22.5 1.0 0.3 8.0 54 Stable37 Back 2.7 121.5 0.4 1.0 2.0 97 Stable38 Back 6.1 121.5 0.4 1.0 2.0 97 Unstable39 Back 7.6 121.5 0.6 1.0 2.0 146 Unstable40 Back 8.8 39.0 0.6 1.0 2.0 47 Unstable41 Back 13.4 39.0 0.6 1.0 2.0 47 Unstable42 Back 6.1 18.2 0.5 0.3 2.0 5.5 Unstable43 Back 15.2 18.2 0.5 0.3 2.0 5.5 Caved44 Back 6.4 18.2 0.3 0.3 2.0 3.3 Unstable46 HW 13.1 39.0 1.0 0.3 8.0 94 Stable47 Back 7.3 18.0 0.3 1.0 2.0 11 Caved48 Back 5 18.0 0.1 1.0 2.0 3.6 Unstable49 Back 9.9 18.0 1.0 1.0 2.0 36 Caved50 Back 6.8 18.0 0.4 1.0 2.0 14 Caved9192The Mathews method, as briefly reviewed in Chapter 4, was selected as the starting pointfor the model. This method involves the determination of a stability number, N, from a modifiedQ value and three factors as follows:(5.1) N=QIxAxBxC.Potvin (1988) starts with the same idea, but expresses the factors A, B and C slightly differently.The Potvin approach described the stability number in terms of a block size factor, a stressfactor, a joint orientation factor and a gravity factor. The Potvin stability number is documentedas N' since modifications have been made to the original Mathews stability number.5.2.1 The Block Size FactorThe block size factor is expressed as RQD/Jn and remains unchanged from the Q systemof rock classification. Barton, Lien, and Lunde (1974) describe the quotient RQD/Jn as a crudemeasure of the relative block size within a rock mass. The ratio can differ by a factor of 400since it ranges from a minimum value of 0.5 to a maximum of 200. RQD can be determinedfrom an examination of drill core, but is subject to the effects of drill hole orientation. Potvin(1988) recommended that RQD be determined from direct underground mapping and theutilization of methods proposed by Palmstrom (1985) and Priest and Hudson (1976). Priest andHudson (1976) related RQD to the average number of discontinuities per meter using thefollowing relationship:(5.2)^RQD = 100e -0.11(0.11+1)where A represents the average number of discontinuities per meter obtained from a scanline93mapping survey. Palmstrom (1985) describes the volumetric joint count, J„ as the number ofjoints intersecting one unit volume of rock mass. The value of J v in joints/m3 can be evaluatedby adding the average number of joints per meter for each joint set in a rock mass. Palmstrom(1976) related RQD to the volumetric joint count with the following relationship:(5.3)^RQD = 115 - 3.3./vwhere RQD is set to 100% when J, is less than 4.5 joints/in'. Potvin (1988) suggests that theratio of RQD/Jn ranges from 1 to 90 for Canadian open stope situations.5.2.2 The Stress FactorThe stress factor is related to the uniaxial compressive strength to induced stress ratio,as shown in Figure 5.1a. Mathews et al. (1981) suggested that rock would fail if the ratio ofuniaxial compressive strength to induced stress was 2.0 or lower. Potvin (1988) proposed thatthe lower bound for the stress factor be kept at 0.1 based on the observation of several highlystressed backs that remained stable due to their small size. The uniaxial compressive strength canbe taken from the results of laboratory testing and an estimate of induced stress determined fromnumerical modelling. Since access to numerical modelling software was limited at manyminesites, Potvin (1988) developed a series of induced stress curves to aid in the evaluation ofthe stress factor. The use of two and three dimensional numerical modelling software at minesis increasingly evident and provides a better approach to stress estimation for complexgeometries. The stress factor can vary from 0.1 for highly stressed backs to 1.0 for surfaces thatare in a state of relaxation.955.2.3 The Joint Orientation FactorPotvin found that most structurally controlled failures occurred where joints formed ashallow angle with the stope surface. This introduced the concept of the critical joint, asillustrated in Figure 5.2. The critical joint represents the joint set that is most likely to detractfrom the stability of a particular surface. The least favourable case is where the critical joint dip,0, is found to be ten to thirty degrees from the dip of the stope surface. The distance, d, is smallenough in this case to enhance the chance of failure. The orientation of the critical joint withrespect to the stope surface is related to a joint orientation adjustment factor, B, as illustrated inFigure 5.1b. The joint orientation adjustment factor was set to 0.2 for the worst case, andimproves to a maximum of 1.0 when the dip difference approaches 90 degrees. A slightimprovement in stability is found when the critical joint parallels the stope surface, and theadjustment factor increases marginally to 0.3. A difference in strike increases the difference intrue dip and is reflected in the dashed lines in Figure 5. lb. The second parameter that is includedin Potvin's Joint Orientation Factor is the shear strength of the critical joint. This value isrepresented by the Jr/Ja term of the Q system, and is estimated by underground geotechnicalmapping. Barton, Lien, and Lunde (1974) suggested that Jr/Ja is representative of the shearstrength that might result from various combinations of joint roughness and alteration. Potvin(1988) suggests that Jr/Ja can range from 0.05 to 3 for Canadian open stope practice.5.2.4 The Gravity FactorGravity is an important consideration in the kinematic analysis of rock structures. Potvinidentified five modes of failure encountered in open stoping situations, gravity fall, slabbing,buckling, sliding and shearing. Buckling and shearing are similar mechanisms to slabbing andsliding modes of failure, except that the principal driving force is stress rather than gravity. The96Figure 5.2: Critical joint concept (After Potvin 1988)97effects of stress have already been incorporated in the stress factor. The remaining gravityinduced failure modes can be summarized into three mechanisms, namely gravity fall, slabbingand sliding. Gravity fall and slabbing modes of failure are dependent on gravity, and have beenrelated to the inclination of the stope surface as shown in Figure 5.1c. This factor is based onthe idea that a vertical stope wall is four times as stable as a stope back. The sliding mode offailure is dependent on the inclination of the critical joint and is derived from Figure 5.1d. Theadjustment factor is a maximum where the inclination of the critical joint is less than thirtydegrees, since most rocks have an angle of friction in this range. As the inclination of the criticaljoint increases, the probability of a sliding failure rises, and the adjustment factor decreases. Thegravity factor for all failure modes ranges from a value of 2 to 8.5.2.5 The Unsupported Modified Stability GraphPotvin proposed that all the above factors be combined to determine a modified stabilitynumber where:JointRQD Compressive Orientation jr^Gravityx AdjustmentFactor(5.4)^N -Jn x^Stress xAdjustment JaFactor^FactorTo avoid confusion, Equation 5.4 can be expressed in terms of the original Mathews terminologywhere:(5.5)^N/ = (2 1 x A xBxCand the factors A, B and C represent the stress factor, joint orientation adjustment factor and98gravity adjustment factor proposed by Potvin (1988). The modified stability number is relatedto the hydraulic radius on the Modified Stability Graph (Figure 5.3). The hydraulic radius isadapted directly from the Shape Factor used by Mathews et al. (1981), and is calculated bydividing the area of the surface by its perimeter. Potvin proposed that a stable and a caved zonecould be identified and used for stope design. The main database was plotted on the ModifiedStability Graph and showed good correlation with the proposed design regions. A transition zonewas indicated within the shaded region, where many of the unstable cases were found toconcentrate. The complementary database was added to the chart and used to confirm the originalhypothesis. Potvin (1988) recommended that the Modified Stability Graph be used in conjunctionwith economic, scheduling and mining constraints for the design of underground openings. It wassuggested that critical openings, where absolute stability is required, be designed to plot withinthe stable zone. Designs that involve multiple openings in the same area should also plot withinthe stable zone. It was noted that non-entry methods of mining may still be viable when plottingbelow the transition zone, since there is greater tolerance for instability. Potvin found thatopenings plotting within the transition zone were sensitive to the effects of blasting and time.Dilution is expected to increase as a point moves farther into the caved zone, but has not yetbeen calibrated with the Modified Stability Graph.5.2.6 The Supported Modified Stability GraphPotvin (1988) noted that surfaces plotting below the transition zone required support inmany cases to remain stable. To investigate the effects of support, 66 case studies of supportedsurfaces were assembled and plotted on the Modified Stability Graph. The parameters for thesupported case histories are summarized in Table 5.3. A region below the transition zone was1• APHMO IED STA ILITY G0 5 10 15 20 250.1HYDRAULIC RADIUS (m)1000C:4•T4100z:=4▪ -4▪ 0E-1cir).10STABLECAVED99Figure 5.3: Unsupported Modified Stability Graph (After Potvin 1988)Table 5.3: Summary of Potvin Supported Database *^**Case#SurfaceHydraulicRadius(m)0' A B C N'Cable BoltLength(m)251 *Back 8.4 18.75 0.25 0.2 2.0 1.9 • 21.0252 *Back 8.4 18.75 0.5 0.2 2.0 3.8 • 21.0253 *Back 5.3 18.75 0.25 0.2 2.0 1.9 • 21.0254 *Back 6.4 54.0 0.1 0.2 2.0 2.2 • 21.0255 HW 5.0 18.75 1.0 0.2 8.0 30256 *Back 5.9 54.0 0.1 0.2 2.0 2.2 9.0257 *Back 6.7 54.0 0.1 0.2 2.0 2.2 Stable 18 • 9.0258 *Back 7.1 54.0 0.1 0.2 2.0 2.2 • 9.0259 *Back 4.6 54.0 0.1 0.2 2.0 2.2 9.0260 *Back 5.0 54.0 0.1 0.2 2.0 2.2 9.0261 *Back 13.9 7.0 0.1 0.4 2.6 0.7 6.0• 24.0262 *Back 16.0 7.0 0.1 0.4 2.6 0.7 6.0• 24.0263 Back 7.3 4.2 0.36 0.2 2.4 0.7 3.0• 18.0264 Back 6.0 0.8 0.1 0.4 2.0 0.1 3.0• 18.0265 Back 8.0 4.2 0.35 0.2 2.4 0.7 3.0• 18.030.0266 Back 14.8 4.2 0.77 0.2 2.4 1.6 3.0• 18.0267 Back 7.8 0.8 0.1 0.85 2.0 0.1 18.0268 Wall 8.9 6.0 1.0 0.2 3.5 4.2 10.0269 Back 4.4 6.0 0.2 0.2 2.0 0.5 3.0270 Back 5.3 6.0 0.1 0.2 2.0 0.2 6.0271 Back 5.3 10.5 0.1 0.5 2.0 1.1 1.32 0.30 9.0272 Back 6.2 40.0 0.1 1.0 2.0 8.0 3.0273 *Back 2.6 9.0 0.1 0.2 2.0 0.4 6.0274 *Back 4.2 9.0 0.1 0.2 2.0 0.4 • 6.0275 End 4.7 9.0 0.1 0.4 5.0 1.8 9.0276 End 6.1 9.0 0.23 0.25 4.6 2.4 9.0277 Back 5.2 9.0 0.1 0.2 2.0 0.4 • 9.0278 **Back 2.5 3.2 0.1 0.2 2.0 0.1279 **Back 7.5 3.2 0.2 0.2 2.0 0.3280 **Back 2.7 30.0 0.6 0.2 2.0 7.2OTable 5.3: Summary of Potvin Supported Database (con'n * Cables with rebar. ** Re bar onlCase#Su rfaceHydraulicRadius(m)Q' A B C N' Stability RQD/Jn RQD/Jn/HRCable BoltDensity(Bolts/Sq. m)Cable BoltLength(m)281 **Back 3.6 30.0 0.8 0.2 2.0 9.6 Stable 15 4.17282 Back 4.1 6.25 0.1 0.9 2.0 1.1 Stable 25 6.10 0.10 15.0284 HW 7.5 4.5 0.45 0.3 6.0 3.6 Stable 9 1.20 0.03 15.0285 **Back 6.3 6.0 0.5 0.2 2.0 1.2 Unstable 8 1.27286 **HW 10.0 6.0 1.0 0.3 6.0 11 Stable 8 0.80287 HW 19.7 30.0 1.0 0.3 8.0 72 Stable 30 1.52 0.07 21.0289 Back 6.2 34.0 1.0 0.2 2.0 14 Stable 17 2.74 0.07 11.0290 Back 11.4 9.0 0.6 0.3 2.4 3.9 Caved 6 0.53 0.10 20.0291 Back 8.0 9.0 1.0 0.3 2.4 6.5 Stable 6 0.75 0.10 20.0292 Back 20.8 9.0 1.0 0.3 2.4 6.5 Caved 6 0.29 0.10 20.0293 Back 9.2 9.0 1.0 0.3 2.4 6.5 Stable 6 0.65 0.10 20.0294 Wall 19.0 2.0 1.0 0.3 8.0 4.8 Caved 4 0.21 0.02 6.0295 Back 3.7 43.5 0.2 0.3 2.0 5.2 Stable 29 7.84 0.15 6.0296 Back 5.3 25.5 0.1 0.2 2.0 1.0 Stable 17 3.21 0.23 9.0297 Back 9.0 50.0 0.3 0.2 2.0 6.0 Stable 25 2.78 0.27 10.0298 Back 3.9 50.0 0.1 0.2 2.0 2.0 Stable 25 6.41 0.20 5.0299 Back 8.0 25.5 0.3 0.9 2.0 11 Stable 17 2.13 0.21 10.0300 Back 4.7 16.2 0.1 0.2 2.0 0.6 Stable 9 1.91 0.22 12.0301 Back 7.7 16.2 0.1 0.2 2.0 0.6 Caved 9 1.17 0.22 12.0302 Back 5.6 2.0 1.0 0.2 2.0 0.8 Stable 2 0.36 0.33 10.0303 Back 4.3 10.0 1.0 0.2 2.0 4.0 Stable 10 2.33 0.33 7.5304 Back 2.7 5.0 1.0 0.9 2.0 9.0 Stable 5 1.85 0.37 7.5305 Back 14.0 2.0 1.0 0.2 2.0 0.8 Caved 2 0.14 0.22 10.0306 Back 9.3 5.0 1.0 0.2 2.0 2.0 Unstable 5 0.54 0.19 10.0307 Back 12.7 1.0 1.0 0.2 2.0 0.4 Caved 1 0.08 0.28 10.0308 Back 14.6 8.0 1.0 0.2 2.0 3.2 Stable 8 0.55 0.20 10.0309 Back 7.1 15.0 0.7 0.2 2.0 4.2 Stable 15 2.11 0.16 15.0310 Back 8.0 25.0 0.7 0.2 2.0 7.0 Stable 25 3.13 0.16 10.0311 Back 7.4 20.0 0.7 0.2 2.0 5.6 Stable 20 2.70 0.16 25.0312 Back 13.7 5.0 0.7 0.2 2.0 1.4 Caved 5 0.36 0.25 10.0313 Back 10.0 10.0 0.7 0.2 2.0 2.8 Stable 10 1.00 0.16 18.0314 Back 5.3 20.0 0.5 0.2 2.0 4.0 Stable 20 3.77 0.16 15.0315 Back 6.9 20.0 0.5 0.2 2.0 4.0 Stable 20 2.90 0.13 25.0316 HW 8.4 16.2 1.0 0.2 5.5 18 Stable 9 1.07 0.07 12.0317 Back 8.6 21.0 0.1 0.8 2.0 3.4 Unstable 14 1.63 0.11 15.0318 Back 5.9 23.4 0.1 0.9 2.0 4.2 Stable 13 2.20 0.31 6.0102identified where supported surfaces seemed to remain stable. Cable bolts were the most dominanttype of support within the database of supported case histories, and Potvin proposed that the areabelow the transition zone illustrated in Figure 5.4 be used for the design of cable bolt support.If a surface plots within this supportable region, the use of cable support is recommended.5.3 DESIGN CHART FOR CABLE BOLT DENSITYIf cable bolt support is required, Potvin (1988) proposed that the Design Chart For CableBolt Density (Figure 5.5) be used to estimate the bolt density required for stope backs. The cabledensity was related to the ratio of the block size factor to the surface hydraulic radius,RQD/Jn/HR. This term provides a relative estimate of the block size with respect to the surfacehydraulic radius, and is referred to as the relative block size factor. A smaller block size, or alarger hydraulic radius, decreases the relative block size and suggests a higher bolt density. Theband of three diagonal lines represents the recommended region for cable bolt design. The centerline was suggested as a conservative design guideline for cable density. Only one case of cablesupport was encountered below a density of 0.10 bolts per square meter, and the region belowthis value is not recommended for design. A success rate of 20% was indicated to the left of thevertical dashed line on the design chart. Potvin (1988) suggested that cable bolts will havelimited effectiveness when the relative block size factor is below 0.75. A revised interpretation(Potvin and Milne 1992) of the Design Chart for Cable Bolt Density is illustrated in Figure 5.6.The hatched region in Figure 5.6 represents a caving zone where the cable density or relativeblock size factor is too low. The area between line #1 and #2 is suggested for the design of non-entry open stope backs. A conservative design density for non-entry mining and a realistic design103Figure 5.4: Supported Modified Stability Graph (After Potvin 1988)106for entry mining methods is bounded by line #2 and #3. The region above line #3 is associatedwith very conservative design.5.4 DESIGN CHART FOR CABLE BOLT LENGTHThe line, L, on Figure 5.7 was suggested as an approximate and conservative guidelinefor the determination of cable bolt length (Potvin, Hudyma, and Miller 1989). This line indicatesthe point at which the cable bolt length is approximately equivalent to stope span for a range ofhydraulic radii. This is based on the concept that hydraulic radius converges to half the span atlow span to length ratios. The hydraulic radius encountered by Potvin (1988) varied from 3 mto 21 m, but limited success was found with values exceeding 10m.5.5 DISCUSSIONThe design techniques developed by Potvin (1988) are based on Canadian open stopingexperience. The application of the Modified Stability Graph for stope design was widelyencountered in western Canadian practice, although it was not used extensively for the designof cable support. The Design Chart for Cable Bolt Density was rarely consulted due to theuncertainty in locating particular design ranges. The data scatter was noted by Potvin (1988) andrevised design bands have been suggested (Potvin and Milne 1992) to improve the cable densityselection process. The revised guidelines proposed by Potvin and Milne (1992) offer moredescriptive design criteria, but still require further calibration. The applicability of the DesignChart for Cable Bolt Density is restricted to an even distribution of cables over the design108surface, and is only intended for use with stope backs that plot in the supportable zone of theModified Stability Graph. The chart was not intended for use with cases of hangingwall supportor point anchor back support. Potvin and Milne (1992) also suggest that inclusions of weakmaterial, slot raise or brow development, and poor quality grout will limit the application of thedensity chart.The ideas proposed by Potvin (1988) are valuable resources to the mining engineer sincethey relate the characteristics of a rock mass to the design of underground openings. Theadditional database collected in this study will be used to evaluate and improve upon existingdesign guidelines. Limited statistical analysis of the database was completed by Potvin (1988),but Chapter 6 will explore this topic in more detail.109CHAPTER 6DATABASE6.1 INTRODUCTIONThe database assembled for this study was obtained during visits to twelve minesin Western Canada and the northwestern United States in the spring, summer and fall of1991. One overseas mining operation was visited to review cable bolting procedures butdid not contribute to the case history database. Eleven of the twelve mines in NorthAmerica used cable bolts for the support of underground openings.Design regions for the Modified Stability Graph and the Design Chart for Cable BoltDensity (Potvin 1988) were visually calibrated through the empirical analysis of 242 casehistories of supported and unsupported open stope surfaces. In this chapter, a newdatabase will be evaluated in relation to the design methods proposed by Potvin (1988).Section 6.4 will review statistical methods that were applied to aid in the identificationof design trends. Although statistical methods have traditionally not been widely appliedin the field of empirical rock mechanics, any collection of data warrants a numericalreview to aid in the identification of an implied trend. The following quote provides someinsight into the goals of this statistical analysis.There is no magic about numerical methods, and many ways in which they can breakdown. They are a valuable aid to the interpretation of data, not sausage machinesautomatically transforming bodies of numbers into packets of scientific fact (Marriott1974).1106.2 DATA COLLECTIONThe collection of data was done during visits arranged to different miningoperations. There is some variability in the methods of data collection due to the variationin time allocated to each particular mine Some operations had the flexibility to allow themine visit to extend beyond a week, but most were limited in the amount of time thatcould be allocated to the purposes of this study. Where time permitted, detailed mappingand underground investigation were used to evaluate the parameters involved. On someoccasions this was not possible, and the data was collected through discussions with minepersonnel and a review of underground plans. Given the limited time constraints, this typeof study requires the ability to quickly gain an understanding of the general mineenvironment and subsequently extract relevant information. The geological departmentat most operations is a tremendous resource in this regard and often have mapping andclassification data logged on site.6.3 DATABASEThe new database is made up of 13 unsupported and 46 supported case histories.A supported case history refers to a stope surface that contains cable bolts as a means ofreinforcing the rock mass. One isolated supported case applies to a surface reinforcedwith long extension bolts. An unsupported case history is defined as a surface that doesnot contain cable bolts. Supported and unsupported surfaces may incorporate short patternbolting with point anchor or fully bonded bolts. In an effort to distinguish between theeffect of long bolts and short pattern bolting, surfaces with cable bolt lengths less than1113.7 m were not considered in this database.The parameters collected for this study are identical to those assembled in thePotvin (1988) database. A stability number, N', and hydraulic radius, HR, weredetermined for each case history to compare the collected database to the design rangeson the Modified Stability Graph proposed by Potvin (1988). The cable bolt density andrelative block size factor, RQD/Jn/HR, were also evaluated for the supported casehistories to compare results to the Design Chart for Cable Bolt Density (Potvin 1988). Theunsupported database is summarized in Table 6.1 and the supported database in Table 6.2.A more detailed summary of each case history is presented in Appendix A.The Potvin (1988) unsupported database is made up of stable, unstable and cavedcase histories. A stable stope surface contributes low amounts of dilution, while anunstable surface is characterized by operational problems due to dilution or ground falls.A caved surface represents case histories with severe ground control problems. In thisstudy, unsupported surfaces are described in a similar fashion. The caved classificationwas applied to those surfaces that were observed or reported to be well beyond thedesigned excavation limits, due to ground fall or excessive dilution. Several cases ofground instability that lead to stope closure were classified as caved. The stableclassification was applied to surfaces that were observed or reported to be within thedesigned excavation limits. Cases that did not fit clearly into either the stable or cavedcategories were classified as unstable.The supported case histories assembled by Potvin were classified as stable,unstable or caved in relation to the support system. The caved classification was appliedto cases where the support system failed. Cases involving ravelling of the rock massbetween the cables were classified as unstable. The stable classification was applied toTable 6.1: Summary of Unsupported Database Case#SurfaceHydraulicRadius(m)Q' A B C N' Stability22 Back 6.2 13.3 0.1 0.2 2.0 0.53 Caved24 Back 5.2 13.3 0.1 0.2 2.0 0.53 Caved28 HW 10.3 10.0 1.0 0.3 5.0 15.0 Caved31 HW 16.4 5.9 1.0 0.2 5.5 6.5 Caved35 HW 7.0 13.1 1.0 0.2 8.0 21.0 Stable38 HW 5.2 7.2 1.0 0.2 5.0 7.2 Stable39 Back 1.3 15.8 0.1 0.2 2.0 0.63 Stable40 HW 6.1 21.5 1.0 0.3 6.0 38.7 Stable41 Back 1.8 15.8 0.1 0.2 2.0 0.63 Stable42 HW 6.1 21.5 1.0 0.2 5.0 21.5 Unstable44 HW 5.9 7.2 1.0 0.2 5.0 7.2 Caved51 HW 10.4 8.3 1.0 0.3 5.0 12.5 Stable58 HW 4.9 3.1 1.0 0.3 6.0 5.6 UnstableTable 6.2: Summary of Supported Database Case#SurfaceHydraulicRadius(m)Q' A B C N' Stability RQD/Jn RQD/Jn/HRCable BoltDensity(Bolts/Sq. m)Cable BoltLength(m)1 HW 10.0 11.7 1.0 0.2 6.0 14.0 Caved 11.7 1.17 0.018 9.1 -13.32 Back 2.3 11.7 1.0 0.2 2.0 4.7 Stable 11.7 5.09 0.130 6.13 HW 11.7 2.5 1.0 0.3 7.5 5.6 Stable 6.7 0.57 0.021 6.14 HW 19.1 2.5 1.0 0.3 7.5 5.6 Caved 6.7 0.35 0.018 6.15 Back 2.6 11.7 1.0 0.2 2.0 4.7 Stable 11.7 4.50 0.160 9.16 HW 17.1 27.3 1.0 0.3 7.5 61.4 Caved 13.7 0.80 0.022 6.17 HW 10.9 2.5 1.0 0.2 7.5 3.8 Unstable 6.7 0.61 0.011 6.18 HW 12.7 2.5 1.0 0.3 7.5 5.6 Unstable 6.7 0.53 0.020 6.19 HW 13.2 23.7 1.0 0.3 8.0 56.9 Stable 11.8 0.89 0.018 6.110 Back 5.0 18.8 0.2 0.3 2.0 2.3 Stable 12.5 2.50 0.130 22.011 HW 10.8 30.0 1.0 0.3 7.5 67.5 Stable 15 1.39 0.023 6.112 Back 1.6 18.8 0.1 0.2 2.0 0.75 Stable 12.5 7.81 0.580 7.613 Back 3.6 11.7 1.0 0.2 2.0 4.7 Stable 11.7 3.25 0.116 6.114 Back 4.3 0.6 1.0 0.4 2.0 0.48 Stable 1.67 0.39 0.180 14.015 Back 7.6 0.5 1.0 0.3 3.0 0.45 Stable 2.5 0.33 0.280 8 &1516 Back 11.2 0.5 1.0 0.3 3.0 0.45 Caved 2.5 0.22 0.180 8 &1217 Back 8.6 0.6 1.0 0.2 2.0 0.24 Caved 1.67 0.19 0.140 12 &1518 Back 4.2 13.3 0.1 0.3 2.0 0.80 Stable 13.3 3.17 0.290 9.819 Back 5.2 13.3 0.1 0.3 2.0 0.80 Unstable 13.3 2.56 0.270 9.820 HW 12.4 15.8 1.0 0.2 5.0 15.8 Caved 15.8 1.27 0.035 14.621 HW 10.8 15.8 1.0 0.3 6.0 28.4 Stable 15.8 1.46 0.031 14.623 Back 5.2 13.3 0.1 0.2 2.0 0.53 Stable 13.3 2.56 0.330 9.825 Back 6.4 13.3 0.1 0.2 2.0 0.53 Caved 13.3 2.08 0.170 9.826 HW 11.5 15.8 1.0 0.3 7.0 33.2 Stable 15.8 1.37 0.025 14.627 HW 10.7 15.8 1.0 0.3 6.5 30.8 Stable 15.8 1.48 0.041 14.629 Back 2.1 0.9 0.1 0.8 2.0 0.14 Stable 2.5 1.19 0.167 18.330 Back 2.0 11.6 0.1 0.2 2.0 0.46 Stable 6.6 3.30 0.410 6.132 HW 4.9 10.4 1.0 0.2 5.0 10.4 Stable 6.9 1.41 0.070 12.033 Back 1.7 8.9 0.1 0.4 2.0 0.71 Stable 8.9 5.24 0.540 6.134 Back 5.1 8.3 0.1 0.2 2.0 0.33 Unstable 4.7 0.92 0.300 9.136 Back 1.8 29.2 1.0 0.5 2.0 29.2 Stable 11.1 6.17 0.550 6.137 Back 2.3 12.3 0.1 0.2 2.0 0.49 Stable 5.8 2.52 0.410 12.243 Back 2.4 11.1 0.1 0.2 2.0 0.44 Stable 6.25 2.60 *Roc kbolts 4.945 Back 3.6 26.1 0.1 0.3 2.0 1.6 Stable 9.8 2.72 0.210 15.846 Back 5.0 5.4 0.1 0.2 2.0 0.22 Stable 10.8 2.16 0.304 18.347 Back 5.1 5.4 0.1 0.2 2.0 0.22 Stable 10.8 2.12 0.308 18.348 Back 5.0 25.0 0.4 0.2 2.0 4.0 Stable 16.7 3.34 0.245 18.349 HW 15.5 9.9 1.0 0.3 4.7 14.0 Stable 13.1 0.85 0.160 9.150 HW 17.0 3.1 1.0 0.3 4.7 4.4 Caved 8.3 0.49 0.180 9.152 Back 5.2 15.5 0.1 0.2 2.0 0.62 Stable 15.5 2.98 0.110 10.253 Back 3.8 5.4 0.1 0.2 2.0 0.22 Stable 10.8 2.84 0.330 10.754 HW 7.9 3.1 1.0 0.3 5.0 4.7 Stable 8.3 1.05 0.130 9.155 Back 6.2 25.0 0.2 0.2 2.0 2.0 Stable 16.7 2.69 0.350 5.056 HW 10.9 3.1 1.0 0.3 5.0 4.7 Stable 8.3 0.76 0.120 7.6 -9.157 Back 5.4 5.4 0.1 0.2 2.0 0.22 Stable 10.8 2.00 0.340 6.459 Back 5.6 25.0 0.4 0.3 2.2 6.6 Stable 12.9 2.30 0.260 4.9114cases where the support system was effective in maintaining the integrity of the surface,and therefore minimizing operational problems and dilution. A similar classificationsystem was used for the supported database collected in this study. Dilution was used insome cases to gauge the effectiveness of a support system. Pakalnis et al. (1987) describethe empirical prediction of dilution at one operation, and relate the results to the designranges proposed on the Mathews (1981) stability graph. Dilution values for stopesplotting in the potentially caving zone ranged from 9% to 25%, but seem to concentrate onthe 20% to 30% range. The relationship between dilution and the stability condition ofa surface has not yet been well defined and would be an interesting area of futureresearch. In this study, the caved classification was generally applied to stopes thatexperienced greater than 30% dilution, but it must be noted that the effect of dilution willvary depending on mineralized content and individual mine financial structure. Theoperational philosophy at a particular minesite also has to be considered when discussingthe effect of dilution. Smaller stopes tend to have longer stand-up times, and as a result,may be mucked longer and generate high dilution values. Larger stopes may be moresusceptible to caving and a reduction in the planned production period, which in turn canproduce low dilution.A summary of average cable densities and lengths is presented in Table 6.3 for thedatabase collected in this study. The results vary with the type of cable pattern and arepresented in relation to the pattern descriptions discussed in Chapter 3. Mandolin andhangingwall drift fan patterns are not included in Table 6.3 since only one case of eachwas collected in this database. The average density for stable square back patterns is 0.25cable bolts/m2 , or 2 x 2 meters. Single cables were found to reflect a higher installeddensity than double cables. The average density for fan back patterns is in the region ofTable 6.3: Summary of Average Cable Density and Lengtha) Square Back Pattern AllCasesStableCasesStable Cases(single cables)Stable Cases(double cables)Non-entryCasesNon-entryStable CasesNumber of casesAverage cable density (bolts/sq. meter)Average square pattern (m x m)Average length (m)200.242.0411.5170.252.0011.780.301.8313.090.212.1810.5110.232.099.380.242.048.8b) Fan Back Pattern AllCasesStableCasesStable Cases(single cables)Stable Cases(double cables)Non-entryCasesNon-entryStable CasesNumber of casesAverage cable density (bolts/sq. meter)Average square pattern (m x m)Average length (m)40.491.438.040.491.438.030.581.318.110.451.497.630.451.498.130.451.498.1c) Even Hangingwall Pattern AllCasesStableCasesStable Cases(single cables)Stable Cases(double cables)Non-entryCasesNon-entryStable CasesNumber of casesAverage cable density (bolts/sq. meter)Average square pattern (m x m)Average length (m)20.132.777.120.132.777.120.132.777.10N/AN/AN/A20.132.777.120.132.777.1d) Point Anchor Hangingwall Pattern AllCasesStableCasesNumber of casesAverage cable density (bolts/sq. meter)Average ring spacing (m)Average number of cables/ringAverage length (m)  130.0472.435.49.570.0462.346.110.7e) Point Anchor Back Pattern AllCasesStableCasesNumber of casesAverage cable density (bolts/sq. meter)Average ring spacing (m)Average number of cables/ringAverage length (m)40.201.985.39.020.161.755.59.51160.49 bolts/m2 , reflecting a high concentration of bolts within a small area. Point anchorhangingwall patterns exhibit an average 2.4 meter ring spacing along strike with 5 boltsinstalled on each ring. The average point anchor density for hangingwalls is 0.047bolts/m2 , approximately one fifth of the average density found for square back patterns.The low densities are a direct result of limited access and larger hydraulic radii associatedwith hangingwall support. The point anchor approach applied to backs is associated withsmaller surface areas than hangingwalls, and exhibits densities above 0.10 bolts/m 2 . Theaverage cable lengths for square back support is expressed in terms of non-entry cases toreflect the open stope situation. Some cut and fill cases are included in this database andtypically involve cables up to 18.3 meters that are designed to cover a number of mininglifts. The average length for open stope square back patterns is 9 meters. Average lengthsfor other patterns range from 7 to 11 meters.6.3.1 Comparison with the Modified Stability Graph Design RangesThe unsupported and supported data points are plotted on the Modified StabilityGraph in Figures 6.1 and 6.2, respectively. Eighty-five percent of the collected casehistories were found to agree with the design ranges proposed by Potvin. The unsupporteddatabase is limited in size since the main interest of this study was to examine supportedcase histories. In terms of the unsupported database, eighty-three percent of the stablesurfaces plot above the transition zone while eighty percent of the caved surfaces plotbelow the transition zone. Figure 6.2 illustrates that twenty percent of the supported casesplot above the transition zone. This region is classified as stable without support andindicates a conservative approach to cable support design by some operators. Within the.^.......^.^.0^if117MO IFIE STABILITY GRAPHUNSUPPORTED DATABASE10000.10^5^1 0^15^20^25HYDRAULIC RADIUS (m)Without Cables:0 Stable111 UnstableV CavedFigure 6.1: Unsupported case histories compared to the design ranges proposed by Potvin (1988)0. 11000...•^•^• ^•^•VV •I ^118MODIFIK IJ STA is ILITY GRAPHSUPPORTED DATABASE0^5^10^15^20^25HYDRAULIC RADIUS (m)With Cables:• Stable■ UnstableCavedFigure 6.2: Supported case histories compared to the design ranges proposed by Potvin (1988)119Table 6.4: Hydraulic Radius ComparisonDATABASE NICKSON POTVIN (1988)Back Support Cases 29 (63%) 57 (86%)Hangingwall SupportCases17 (37%) 5 (8%)Average HR (m)Backs4.5Range 1.6 - 11.27.5Range 2.5 - 20.8Average HR (m)Hangingwalls12.2Range 4.9 - 19.110.1Range 5.0 - 19.7Average HR (m)Stable Backs3.9Range 1.6 - 7.66.1Range 2.5 - 14.6Average HR (m)Stable Hangingwalls10.8Range 4.9 - 15.510.1Range 5.0 - 19.7supportable region below the transition zone, seventy-one percent of the cases were foundto be stable. The transition zone was identified by Potvin (1988) as a region sensitive tothe effects of blasting and time. Twenty-four percent of the supported cases were foundto plot within the transition zone. A high degree of confidence is indicated by designingin or above the transition zone with support, since no cases of instability were found inthese regions. The Potvin supported database has few successful cases with a hydraulicradius greater than 10 meters, and reflects the large percentage (86%) of back supportcases. Table 6.4 shows that the average hydraulic radius of stope backs is significantlysmaller than hangingwalls. In terms of design, Figure 6.2 shows that the supportable zoneabove a hydraulic radius of 10m is not as reliable as the region below 10m. This suggeststhat there is a limit to the hydraulic radius that can be effectively supported and thesupportable region defined by Potvin (1988) on the Modified Stability Graph may not120necessarily parallel the transition zone.6.3.2 Comparison with the Design Chart for Cable Bolt Density Design RangesThe collected case histories of back support are plotted on the Design Chart forCable Bolt Density in Figure 6.3. Seventy-nine percent of the cases were found to be inagreement with the design ranges proposed by Potvin (1988). Only one case of instabilitywas found when plotting in the Conservative and Very Conservative design zones. Thedatabase supports Potvin's observation that bolt densities of less than 0.10 bolts/m 2 arenot used by mine operators. Unfortunately, the availability of caved cases was limited andthe majority of the case studies obtained for stope backs were found to be stable.6.4 STATISTICAL METHODOLOGYTwo statistical methods will be used in this chapter to analyze the databaseassembled for this study. The methods involve linear regression analysis and a form ofdiscriminant analysis that will be used as a tool to develop design proposals for cablesupport. A brief review of both of these techniques is presented in this section.6.4.1 Linear Regression AnalysisThe determination of a relationship between two variables is often of interest inthe statistical evaluation of a database. Regression defines the process of estimating adependent variable y, from an independent variable x. The scatter diagram in Figure 6.4arepresents a plot of two variables x and y, where (x1,y1), (x2 ,y2), (x„,yn) represent n121Figure 6.3: Cable bolted backs compared to the design ranges proposed by Potvin (1988)122Figure 6.4: Concepts of linear regression analysis123individual sample points. A visual relationship may be apparent from the scatter diagram,but the regression technique provides a method of evaluating a mathematical relationship.Figure 6.4b illustrates the application of the method of least squares to estimate the "best-fitting curve" through a series of data points. The "best-fitting curve", or least squaresregression curve, is defined by minimizing the sum of the residual values d 1 2 ,Where a linear relationship is desired, the "line of best fit" can be derived (Spiegel 1975,258) using the method of least squares, and is defined by(6.1)^ y = ax + bwhere a and b are constants determined by simultaneously solving the equations(6.2) Eyi = an + bExii=i^i=i(6.3)^ xiyi = aE x + bE x2;i=i^i=1^i=1The standard error of estimate (Spiegel 1975, 262) of y on x is a measure of the scatteraround the regression line and can be calculated by(6.4) E^)2i =1 estS =Y n-2where yest is the estimated value of y for a given value of x, as obtained from theregression line. The value s y x has similar properties to standard deviation and can be used124to construct lines parallel to the regression line. Lines constructed at vertical intervals ofsY . x , 25„ and 3s,, „, would respectively bound 68%, 95% and 99.7% of the sample points,providing the sample size is large (n > 30) and approximates a normal distribution.Correlation analysis can be used to measure the strength of a linear relationshipbetween two variables. The correlation coefficient, r, is related (Johnson 1976, 100-107)to the sample variances and covariance of x and y wheren(6.5a)ESample variance of x = s2 - ^ n-1nE (yi -i3)2(6.5b) Sample variance of y = sy2 - i=1n -1(6.5c)nE (xi -x)(y,-;)Sample covariance = sue, - 1 = 1^n-1and x, y represent the sample mean of x and y.The sample correlation coefficient, r, is defined by(6.6) xy r-s sx yand has a value between -1 and +1. A positive correlation reflects a simultaneousincrease in both variables, while a negative correlation indicates that one variable125decreases as the other increases. When the absolute value of the correlation coefficientis high, a close linear relationship is said to exist between two variables. In particular,perfect linear correlation results in a correlation coefficient of +1, and arises when allpoints fall directly on a straight line. A value of zero is assigned to r if there is nocorrelation between the variables. Between r 0 and r = 1+11, there is a statisticaldecision point related to the sample size, that separates a zone of correlation from nocorrelation. The values of x and y discussed in this section represent a random sample thatis taken from a population. Since every member of the population cannot be collected ina typical database, a sample is used to represent that population. The final analysis is todetermine if the sample correlation coefficient, r, indicates that there is a dependencybetween the variables for the population from which the sample was taken. If p representsthe population linear correlation coefficient, it is possible to propose that p =0, whichwould indicate that the two variables are linearly unrelated. The proposal that p =0 isreferred to as the null hypothesis. Rejection of the null hypothesis implies that there islinear dependency between the variables, and the sample correlation coefficient isstatistically significant (Johnson 1976, 480-483). Table 6.5 contains critical values forthe sample correlation coefficient that can be used to consider the null hypothesis atdifferent levels of significance, a (Johnson 1976, A47). The sample correlationcoefficient is used as the test statistic and related to the number of degrees of freedom,df, where df = n - 2. The null hypothesis is rejected for a certain level of significanceif the test statistic, r, is greater than the critical value found in Table 6.5 for a particulardf and a. Spiegel (1975, 211-223) suggests that results found to be significant at a =0.01are highly significant, results significant at a =0.05 but not at a =0.01 are probably126significant, and results found to be significant at levels larger than 0.05 are notsignificant. If the sample correlation coefficient is significant at a=0.05, then there isa 5% chance that the null hypothesis should have been accepted. The 5% level ofsignificance is generally standard in hypothesis testing.Table 6.5: Values of r for Different Levels of Significancedf a df a0.10 0.05 0.01 0.10 0.050 0.011 0.988 0.997 1.00 16 0.400 0.468 0.5902 0.900 0.950 0.990 17 0.389 0.455 0.5753 0.805 0.878 0.959 18 0.378 0.444 0.5614 0.729 0.811 0.917 20 0.360 0.423 0.5375 0.669 0.754 0.874 25 0.323 0.381 0.4876 0.621 0.707 0.834 30 0.296 0.349 0.4497 0.582 0.666 0.798 35 0.275 0.325 0.4188 0.549 0.632 0.765 40 0.257 0.304 0.3939 0.521 0.602 0.735 45 0.243 0.287 0.37210 0.497 0.576 0.708 50 0.231 0.273 0.35411 0.476 0.553 0.683 60 0.211 0.250 0.32512 0.457 0.532 0.661 70 0.195 0.232 0.30213 0.441 0.514 0.641 80 0.183 0.217 0.28314 0.426 0.497 0.623 90 0.173 0.205 0.26715 0.412 0.482 0.605 100 0.164 0.195 0.254df = the number of degrees of freedom1276.4.2 Discriminant AnalysisIn this study a series of case histories have been collected and described in termsof three variables. In statistical terms the case histories form a three dimensionalmultivariate database. In all cases, two variables, for example (HR, N') or (RQD/Jn/HR,Cable Density), are quantified and a third is determined from the stability condition ofthe surface. The stability condition is a qualitative variable, and is expressed in terms ofbeing either stable, unstable or caved. For design purposes, it is necessary to separate thedatabase into zones of stability. Discussions with the Department of Statistics at theUniversity of British Columbia were initiated with the goal of developing a method ofstatistical analysis that could be used to separate an empirical database based on stability.Discriminant analysis is a multivariate technique that is concerned with either separatingobservations into different groups, or assigning a new observation to existing groups(Johnson and Wichern 1988, 470). The method that was selected for use in this study,was a form of discriminant analysis that has the ability of separating three dimensionalmultivariate data into different groups, using a statistical measure of distance called theMahalanobis distance (Seber 1984, 10). This form of discriminant analysis is designedto derive a linear function that will separate a database into two or more select groups.If a random sample of two dimensional stable, unstable and caved points arerepresented by the following variables:[1-11- class of stable points where x1 =^etc.1y1 ^ yhi - class of unstable points128z1 ^ zn3 - class of caved points,then the sample mean vector for each class is given by Equation 6.7a, 6.7b and 6.7c(Johnson and Wichern 1988, 221).(6.7a)n11—^1 It---■X = —2_, Xin i i=in2(6.7b)^-^1 x--,Y = —2., Yin2 fri(6.7c)n3_^1 v.Z = — Li li •n3 i=iThe sample covariance matrices can be evaluated for each class using the followingrelationships:(6.8a)n iStable Class:^S1 = 1n -1E (xci)(xi-x)T^1 ^i=i(6.8b)n2%--,Unstable Class: S2 =^1 1 ^(Yi -.YX,Y i -17) Tn2 -1 i=1713Caved Class:^S3^1  E (z;-i)(zi -z-Y.n3 -1 i = i(6.8c)129If each class can be considered as multivariate normal distributions with the samevariance, then a pooled sample covariance matrix, Sp, (Johnson and Wichern 1988, 222)can be formed where:(6.9)^Sp -(n -1)S1 +(n2 -1)S2 +(n3 -1)S3nl +n2 +n3 -3Figure 6.5a contains a representation of two classes of data, say stable and unstable cases,plotted on some fictional axis. If a new point, 1 = (4,4) is added to the database, intowhich group should it be placed? One technique that can be used to answer this questionis the Mahalanobis squared distance technique, where 13, 2 is the Mahalanobis squareddistance from 1 to each class, determined using the following relationships (Seber 1984,10):(6.10a) Di = (1-WV (1-i) for the stable class(6.10b) D22 = (14)%1(14) for the unstable class.The Mahalanobis distance differs from the straight Euclidean distance (Manly 1976, 47-52) in that it considers correlations between variables. The minimum Mahalanobis squareddistance will define the class to which the point, 1, should be assigned. This techniquewould have to be repeated for every possible point within a design population and hasbeen simplified to deal more efficiently with the data available in this study. Instead ofassigning a point to a particular class, as illustrated in Figure 6.5a, the procedure hasbeen revised to develop a linear boundary that separates two classes. A line exists130Figure 6.5: Methods of separating two classes of data131between two classes where the Mahalanobis squared distance to both classes is equivalentfor any point on that line (Figure 6.5b). The case where D i 2 =D22 could represent astatistically derived division between a stable and unstable data class. Seber (1984, 10)notes that the Mahalanobis distance can be interpreted as a probabilistic distance whereequal distances imply equal probability. Points that plot on the line defined by D i 2 =D2can therefore be interpreted as having an equal likelihood of being in either class. Alinear relationship can be determined for this case by first equating the squared distancesfor each class as in Equation 6.11.^(6.11)^ = (14) 7'41(1--it)Upon expanding this expression:^(6.12)^/ Ts^T^s-1-p X+ Tp T^a^Tp x = I^t-y T 31, i T Sp Y -1 T Sp yand cancelling like terms, the following relation is obtained:(6.13)^-T T 1 7'^ T -1-(y -x )Sp 1+1 Sp -x)+x ,Sp x-y Sp y = 0Since yTx = xTy then Equation 6.13 can be simplified to:(6.14)^Z T[2Spi(i -X)] +C = 0Equation 6.14 represents the equation of a straight line where,132(6.15)^a111 + a2/2 + C = 0 ,(6.16)and(6.17)a1^-1 —= a = 2Sp (y ^,{ 21C = x-TS;1X- - Y—TsCp-1;If (1i ,12) = (HR,N'), then an evaluation of a 1 , a2 , and C can generate a linear divisionbetween two different classes, where each point is a three dimensional variable definedby the hydraulic radius, stability number, and stability condition. The method presentedin this section can be used to separate more than two classes of points if desired.6.5 STATISTICAL ANALYSISA series of three statistical tests were completed using the format outlined inSection 6.4.2 and a spreadsheet software package. The result of each test was a linearseparation between stable and caved points that will be used to review the design rangesproposed by Potvin (1988). The three tests are summarized as follows:Case #1: Derive a linear separation between stable and caved unsupported casehistories and compare the result to the transition zone proposed by Potvin (1988)on the Modified Stability Graph.Case #2: Derive a linear separation between stable and caved supported casehistories and compare the result to the supportable region proposed by Potvin133(1988).Case #3: Derive a linear separation between stable and caved cases of back supportand compare the result to the stable and caved regions on the Design Chart for CableBolt Density.The method of multivariate analysis described in the previous section requires that thesample be randomly collected, satisfy the condition of multivariate normality, and havesimilar variances. The database obtained in this study was combined with the Potvin(1988) database to improve upon the quality of analysis. This was also required to meetthe conditions of multivariate normality and similar variance. Unstable cases wereremoved from the database, since their defmition makes it difficult to consider a separateunstable class. In addition, the unstable case histories are limited, and an excessivevariance was noted for Case #2 and #3 data, when compared to stable and caved classes.The results of this analysis therefore predicts a separation between stable and caved casehistories that will enable the definition of possible design regions. The case histories wererandomly selected in the sense that they were collected on the basis of availability, withno bias to the addition of one case or another. Unfortunately, the availability of casehistories is limited due to access and time constraints, so there is not a large databasefrom which to select a true random sample. The conditions of multivariate normality andsimilar variance were tested using dot plots and probability plots obtained using Systat(Wilkinson 1990b), a statistical software package. A dot plot is similar to a histogram(Wilkinson 1990a, 182) and graphically compares the variance of different classes ofpoints. In order to have similar variances, which is a condition of the pooled samplecovariance matrix, the lines for each class on a dot plot should have a similar measure ofdispersion. Figure 6.6 illustrates the dot plots obtained for the combined unsupported252050(a)StabilityCaved^Stable^UnstableStabilityCaved^Stable^Unstable800700600500(b) z 400300200100I.134Figure 6.6: Dot plots for the combined unsupported database (Case #1)135database (Case #1). The stability number dot plot was interpreted as having excessivevariance, and resulted in the application of a logarithmic transformation. Although notideal, the dot plots obtained for the transformed data (Figure 6.7) were thought to havesimilar variances. The assumption of multivariate normality was checked through the useof normal probability plots. A normal probability plot compares the sorted database to thecorresponding values of a mathematical normal distribution (Wilkinson 1990a, 345). Thedata should plot along a straight line if it is normally distributed. If the individualvariables are normally distributed, then it can be assumed that the joint distribution ismultivariate normal (Manly 1986, 15). The probability plots obtained from theunsupported database are shown in Figures 6.8 and 6.9. Both stable and caved classescannot be considered normally distributed unless a logarithmic transformation is applied.Similar results were found for Case #2 and Case #3 data. The outcome of the analysisindicated that a logarithmic transformation had to be applied to meet the condition ofnormality. The transformation occasionally detracted from the condition of similarvariance, but significantly improved the normality characteristics.Prior to proceeding with the statistical analysis, a final review of the database wasundertaken to determine if any case studies should not be considered. No cases from thecombined unsupported database were eliminated and it is fully represented by the data inTables 5.1, 5.2 and 6.1. Several cases of rebar support were removed from the Potvinsupported database presented in Table 5.3, namely case number 278, 279, 280, 281, 285,and 286. It was felt that the statistical analysis should be restricted to cable support. Casenumber 43 in Table 6.2 was supported with grouted extension bolts and was removed forthe same reason. Two other cases, number 1 and 25 were removed from the data in Table6.2 prior to the statistical analysis. Case number 1 involves the use of cables installed136Figure 6.7: Dot plots for the combined unsupported database (Case #1 - transformed data)137Figure 6.8: Probability plots for stable cases from the combined unsupported database (Case #1)138Figure 6.9: Probability plots for caved cases from the combined unsupported database (Case #1)139from a hangingwall drill drift, but the cable pattern was not designed to cover thecomplete hangingwall. In addition, a limited free face was available at the time of thefinal blast in this stope, and was thought to significantly contribute to the ensuing failure.Case number 25 involved two cable bolted stopes that were originally separated by asmall rib pillar. The pillar between these stopes partially collapsed, and the case studywas based on the new exposed stope back. Blasting practice was again thought to affectthe stability of the surface and was combined with an uneven distribution of cables in theback. The remaining 103 cases in Tables 5.3 and 6.2 comprise the combined supporteddatabase used in the statistical analysis.6.5.1 Unsupported Database (Case #1)The results of the statistical analysis performed on the combined unsupporteddatabase is shown in Figure 6.10, and is mathematically represented in Equation 6.18.(6.18)^HR (meters) = 100.573 + 0.338logNI)This line was found to compare well with the transition zone proposed by Potvin (1988),except for a possible increase in the size of the stable region at a higher hydraulic radius.The extremities of the statistical line however are prone to the most error, since the datapoints are sparsely distributed in these regions. The location of many unstable surfaceswithin the transition zone proposed by Potvin (1988), suggests that this zone has alreadybeen well defined, and no further adjustment is required at this stage. The Potvin (1988)transition zone is recommended for the design of unsupported stope surfaces, and will besubsequently referred to as the unsupported transition zone.140Figure 6.10: Statistical analysis of stable and caved combined unsupported database (Case #1)1416.5.2 Supported Database (Case #2)The results of the statistical analysis for the combined supported database is shownin Figure 6.11, and is mathematically represented in Equation 6.19.(6.19)^HR (meters) = 10(0.872 + 0.1711°gN)The line separating stable and caved supported surfaces intersects the unsupported transitionzone at a hydraulic radius of approximately 15 m. This suggests that the supportableregion may be smaller than originally proposed by Potvin (1988) and does not necessarilyparallel the unsupported transition zone.In terms of design confidence, 100% of the supported case histories that plot in orabove the unsupported transition zone are stable. Eighty-five percent of the cases plottingbetween the unsupported transition zone and the derived statistical line (Zone A on Figure6.11) are stable. The stability of the points plotting between the derived statistical lineand the Potvin (1988) lower supportable line (Zone B) is much more variable with only58% of the cases being stable. This analysis suggests there is a reduction in designconfidence as points plot farther into the caved zone. A modification to the supportedModified Stability Graph (Potvin 1988) is proposed in Figure 6.12. The statistical analysisrevealed that the supportable region does not parallel the unsupported transition zone andis therefore reflected in the proposed modification. This concept is an indication of thelimitations of cable bolt support in underground mining applications. It suggests that theaddition of cable support beyond a stability number of approximately 80 and a hydraulicradius of 16 m will not change the stability condition of the surface. The block size at this142Figure 6.11: Statistical analysis of stable and caved combined supported database (Case #2)143Figure 6.12: Proposed modifications to the Modified Stability Graph144cutoff point is likely too large to be effectively supported given economic constraintspertaining to cable length. In terms of scale, it has been shown in Table 6.4 that there islimited success with a hydraulic radius greater than 10 m. A supported transition zone hasbeen incorporated in Figure 6.12 to reflect the reduction in design confidence farther intothe caved zone of the graph. The location of the supported transition zone is based on thestatistical line that was derived in this analysis and the edge of the supportable regionproposed by Potvin (1988). Since a transition zone was found to exist for the unsupporteddatabase, it is unlikely that a single line can adequately separate stable and cavedsupported cases. The region between the two proposed transition zones is called the stablewith support zone and reflects the high degree of design confidence suggested by thedatabase.6.5.3 Back Cable Support Database (Case #3)The method of discriminant analysis discussed in Section 6.4.2 was used to dividethe stable and caved cases of back support on the Design Chart for Cable Bolt Density(Potvin 1988). The statistical line is illustrated on Figure 6.13 and is mathematicallyreflected in Equation 6.20.(6.20) Cable Density (bolts1m2 ) 10 (-0.697 - 4.2071og(RQD/Jn/HR))The line proposed by Potvin (1988) traces a line that joins points A, B, C, and D onFigure 6.13. The statistical interpretation follows the same trend as Potvins' line, butsuggests that the limiting relative block size factor may be closer to 1.0. A possible145Figure 6.13: Statistical analysis of combined cable bolted back database (Case #3)146transition zone, as illustrated on Figure 6.14, could be considered between a relativeblock size factor of 0.75 and 1.0. The design band proposed by Potvin (1988) suggeststhat cable bolt density should increase with a decreasing relative block size factor. Thecombined database illustrated in Figure 6.13 does not seem to correlate with thisparticular trend. However, it should be noted that the design ranges proposed by Potvin(1988) and the statistical analysis illustrated on Figure 6.13, involved the use of the entireback support database. Figure 6.15 isolates cases of back support that plot within theunsupported transition zone, the stable with support zone, and the supported transition zoneproposed in Figure 6.12. A design line that reflects a minimum cable bolt density isincorporated within each chart, based on maintaining 100% stability above the line. Thismethod of analysis produces increasing cable bolt density farther into the caving zone ofthe Modified Stability Graph, but shows no correlation with the relative block size factor.The relative block size factor may not represent the ideal relationship with cable boltdensity for back support. This thesis will examine other possible relationships to improveupon the definition of design zones for back cable support.The techniques of linear regression analysis discussed in Section 6.4.1 were usedto review statistically relevant design trends related to stable cases of back support. Table6.6 summarizes the cases of back support that plot within the stable with support zone andthe supportable transition zone. With cable bolt density as the dependent variable, a seriesof linear regression relationships were evaluated for different independent variables,utilizing the stable cases from Table 6.6 and the Systat (Wilkinson 1990b) softwarepackage. These relationships are illustrated in Figures 6.16 to 6.22, and are summarizedin Table 6.7. Figure 6.16 shows the regression line produced for the relationship betweencable bolt density and the relative block size factor, RQD/Jn/HR. This is the basis for the147Figure 6.14: Possible transition zone for the Design Chart for Cable Bolt DensityMinimum Density = 0.10-7 8(b)E6_ 0.3500.300.450.40as 0.250as5:2-0.202 0.15a)0.1000.051 7 82^3^4^5^6(RQD/Jn)/Hydraulic Radius(c)2^3^4^5^6(RQD/Jn)/Hydraulic Radius0.000 10.450.40Eci- 0.35cnLI 0.300w 0.25asas2- 0.200.15a)00.100.05(a)(ROD/Jn)/Hydraulic Radius148Figure 6.15: Minimum cable density for design regions of the revised Modified Stability GraphTable 6.6: Cases of Back Support in the Supportable Region of the Revised Modified Stability Graph - Combined DatabaseCase#SurfaceHydraulicRadius(ni)Q' A B C N' Stability RQD/Jn RQD/Jn/HRCable BoltDensity(Bolts/Sq. n1)Cable BoltLength(ro)14 Back 4.3 0.6 1.0 0.4 2.0 0.48 Stable 1.67 0.39 0.180 14.015 Back 7.6 0.5 1.0 0.3 3.0 0.45 Stable 2.5 0.33 0.280 8 &1518 Back 42 13.3 0.1 0.3 2.0 0.80 Stable 13.3 3.17 0290 9.819 Back 5.2 13.3 0.1 0.3 2.0 0.80 Unstable 13.3 2.56 0.270 9.823 Back 5.2 13.3 0.1 0.2 2.0 0.53 Stable 13.3 2.56 0.330 9.825 Back 6.4 13.3 01 0.2 2.0 0.53 Caved 13.3 2.08 0.170 9.834 Back 5.1 8,3 0.1 0.2 2.0 0.33 Unstable 4.7 0.92 0.300 9.146 Back 5.0 5.4 0.1 0.2 2.0 0.22 Stable 10.8 2.16 0.304 18.347 Back 5.1 5.4 0.1 0.2 2.0 0.22 Stable 10.8 2.12 0.308 18.352 Back 5.2 15.5 0.1 0.2 2.0 0.62 Stable 15.5 2.98 0.110 10.253 Back 3.8 5.4 0.1 0.2 2.0 022 Stable 10.8 2.84 0.330 10.755 Back 6.2 25.0 0.2 0.2 2.0 2.0 Stable 16.7 2.69 0.350 5.057 Back 5.4 5.4 0.1 0.2 2.0 0.22 Stable 10.8 2.00 0.340 6.4251 +Back 8.4 18.75 0.25 0.2 2.0 1.9 Caved 25 2.98 0.17 21.0252 +Back 8.4 18.75 0.5 0.2 2.0 3.8 Caved 25 2.98 0.17 21.0253 +Back 5.3 18.75 0.25 0.2 2.0 1.9 Stable 25 4.72 0.17 21.0254 +Back 6.4 54.0 0.1 0.2 2.0 2.2 Stable 18 2.81 0.17 21.0256 +Back 5.9 54.0 0.1 0.2 2.0 2.2 Stable 18 3.05 0.16 9.0257 +Back 6.7 54.0 0.1 0.2 2.0 2.2 Stable 18 2.69 0.16 9.0258 +Back 7.1 54.0 0.1 0.2 2.0 2.2 Stable 18 2.54 0.16 9.0263 Back 7.3 4.2 0.36 0.2 2.4 0.7 Stable 6 0.82 0.17 3.00.17 18.0264 Back 6.0 0.8 0.1 0.4 2.0 0.1 Stable 4 0.67 0.17 3.00.17 18.0265 Back 8.0 4.2 0.35 0.2 2.4 0.7 Stable 6 0.75 0.17 3.00.17 18.00.05 30.0269 Back 4.4 6.0 0.2 0.2 2.0 0.5 Stable 6 1.36 0.20 3.0270 Back 5.3 6.0 0.1 0.2 2.0 0.2 Stable 6 1.13 0.20 6.0271 Back 5.3 10.5 0.1 0.5 2.0 1.1 Stable 7 1.32 0.30 9.0274 +Back 4.2 9.0 0.1 02 2.0 0.4 Stable 6 1.43 0.24 6.0277 Back 5.2 9.0 0.1 0.2 2.0 0.4 Stable 6 1.15 023 9.0291 Back 8.0 9.0 1.0 0.3 2.4 6.5 Stable 6 0.75 0.10 20.0293 Back 9.2 9.0 1.0 0.3 2.4 6.5 Stable 6 0.65 0.10 20.0296 Back 5.3 25.5 0,1 0.2 2.0 1.0 Stable 17 3.21 0.23 9.0297 Back 9.0 50.0 0.3 0.2 2.0 6.0 Stable 25 2.78 0.27 10.0300 Back 4.7 16.2 0.1 0.2 2.0 0.6 Stable 9 1.91 0.22 12.0301 Back 7.7 16.2 0.1 0.2 2.0 0.6 Caved 9 1.17 0.22 12.0302 Back 5.6 2.0 1.0 0.2 2.0 0.8 Stable 2 0.36 0.33 10.0306 Back 9.3 5.0 1.0 0.2 2.0 2.0 Unstable 5 0.54 0.19 10.0309 Back 7.1 15.0 0.7 0.2 2.0 4.2 Stable 15 2.11 0.16 15.0310 Back 8.0 25.0 0.7 0.2 2.0 7.0 Stable 25 3.13 0.16 10.0311 Back 7.4 20.0 0.7 02 2.0 5.6 Stable 20 2.70 0.16 25.0313 Back 10.0 10.0 0.7 0.2 2.0 2.8 Stable 10 1.00 0.16 18.0315 Back 6.9 20.0 0.5 0.2 2.0 4.0 Stable 20 2.90 0.13 25.0317 Back 8.6 21.0 0.1 0.8 2.0 3.4 Unstable 14 1.63 0.11 15.0Cables with Rebar)1490 1 2^3RQD/Jn/HR4 5Cable Bolt Density vs RQD/Jn/HR0.400.35ESir 0.30co0.25_c)a)-aw 0.20C.)cnCa)°0.150.10 0co0.050.00•• • •••• •■ ••■^•••■ •■•11^•• • •Regression Liner=0.029•• SSD • v^• v• •• • ••• Stable■ Unstablev Caved151Table 6.7: Possible Linear Relationships for Cable Bolt DensityTESTSERIESINDEPENDENTVARIABLE(x)EQUATION r SIGNIFICANCEI 1 1^RQD/Jn/HR y = 0.212 + 0.002x +0.029 I^NOT SIGNIFICANT I2A x RQD/ln/IIR y = 0.243 - 0.053x -0.369 PROBABLY SIG.A l x RQD/Ja/HR y = 0.228 - 0.013x -0.203 NOT SIGNIFICANTA2 x RQD/Ja/HR y = 0.235 - 0.027x -0.314 NOT SIGNIFICANTLog(A x RQD/ln/HR) y = 0.184 - 0.072x -0.344 PROBABLY SIG.Log(A1 x RQD/Ja/HR) y = 0.203 - 0.041x -0.268 NOT SIGNIFICANTLog(A2 x RQD/Ja/HR) y = 0.193 - 0.061x -0.351 PROBABLY SIG.3Q' y --- 0.234 - 0.001x -0.231 NOT SIGNIFICANTQ'/HR y = 0.231 - 0.006x -0.183 NOT SIGNIFICANTA x Q'/HR y = 0.250 - 0.057x -0.432 PROBABLY SIG.Al x Q'/HR y = 0.223 - 0.007x -0.126 NOT SIGNIFICANTA2 x Q'/HR y = 0.230 - 0.016x -0.248 NOT SIGNIFICANTLog(Q'/HR) y = 0.220 - 0.017x -0.113 NOT SIGNIFICANTLog(A x Q'/HR) y = 0.189 - 0.061x -0.397 PROBABLY SIG.Log(A 1 x Q'/HR) y = 0.202 - 0.043x -0.341 PROBABLY SIG.Log(A2 x Q'/HR) y = 0.195 - 0.056x -0.409 PROBABLY SIG.4N'/HR y = 0.256 - 0.148x -0.495 HIGHLY SIGNIFICANTLog(N'/HR) y = 0.160 - 0.072x -0.443 HIGHLY SIGNIFICANTDesign Chart for Cable Bolt Density (Potvin 1988), but no significant relationship isrevealed. It is interesting to note that the regression line is roughly horizontal at a densityof between 0.21 and 0.22 bolts/m 2 . This corresponds closely with the average density of0.25 bolts/m2 presented in Table 6.3 for stable cases of square back support.Stope backs that are subject to high stress may exhibit stress induced rock failure.The relative block size factor considers the effect of gravity but does not consider theinfluence of stress. There are two failure modes that must be considered when discussingthe effects of stress in relation to cable bolt support. A low uniaxial compressive strength152(a) to induced stress (a) ratio, can induce failure within a rock mass and result in apotential reduction in the inherent block size. Mathews et al. (1981) proposed that failureoccurs at a °jai ratio less than two and this is incorporated in the Potvin (1988) stressfactor, A, at a value of 0.1. Jaeger and Cook (1979) suggest that rock can be regardedas being strong, massive and competent if its uniaxial strength is three or more times thefield stress around an excavation. This suggests that stress will affect a rock mass belowa o-ja, ratio of 3, which is equivalent to an A factor less than 0.3. Stress in stope backscan also enhance stability by clamping blocks together, but depending on the orientationof the critical joint, may induce a sliding failure. To consider the two stress inducedfailure mechanisms discussed, the stress factor, A, was considered in combination withthe relative block size factor, as illustrated in Figure 6.17a. A modification of the Afactor was expressed in terms of A, and A2, to consider a reduction in block size as aresult of high stress. The A, factor had a value of 0.1 or 1.0, and was created to accountfor an increase in block size due to stress induced fractures. The value of A, was set to1.0 when the modified stability number stress factor was greater than 0.1. All other caseswere assigned an A, value of 1.0. The A2 factor considered the comments of Jaeger andCook (1979), and was assigned a value of 0.1 when the modified stability number stressfactor was less than 0.3. The A 2 factor for all other cases again went to 1.0. The relativeblock size factor is combined with A, and A 2 in Figures 6.17b and 6.17c, and theregression results are shown in Table 6.7. The best correlation resulted with the use ofthe modified stability number stress factor (A), and suggests that all ranges of stressshould be considered in an empirical relationship with cable bolt density. A logarithmictransformation was applied, as illustrated in Figure 6.18, and displayed slightly highercorrelations.(a) -2 0.200.15U)a)cl 0.10oco0.054 5RegressionLine•42^3A2 x RQD/Jn/HR• Stable• Unstablev Caved2^3A x RQD/Jn/HR0.000 10.400.35E55;0.30▪ 0.25-2 0.20A"'0.15U)In 0.10O0.05(b)• • Stable■ Unstablev Caved2^3^4^5A, x RQD/Jn/HR0.00 ^0 15• Stable■ Unstablev Caved-2 0.20rnCa 0.10oco0.05-0 • •• 11191^••• 001• DrLino''03140.000•0.400.35E5 0.300.25_o(c)0.400.35E,5; 0.307:) 0.25• • ••- ••• ••011111^40153Figure 6.17: Regression analysis with the relative block size factor and the stress factor• •0.204'0.15000.1000.05•• Stable■ Unstablev Caved0.400.35EFoi" 0.30o 0.25a)fa 0.200.150 0.10oco0.05•• •• V1•^00•^el•• Stable■ Unstablev Caved■000.000.01^0.1^ 1A l x RQD/Jn/HR100.400.35E5r, 0.300.25••fib035j^■• s•0.000.01^0.1^ 1 10Ra0.35E0.30•• •• •■ •• Stable■ Unstablev Caved•11^•7) 0.25xj 0.200 •0.15en0 0.1000.050.40■ • •••0.000.01^0.1^ 1A x RQD/Jn/HR10A 2 x RQD/Jn/HR(a)(b)(c)154Figure 6.18: Regression analysis with the relative block size factor and the stress factor155Figures 6.19 to 6.21 incorporate the use of Q' within the independent variable.Table 6.7 shows that correlations are generally highest when Q' is combined with thestress factor, A. Based on this result, it was decided to incorporate the entire modifiedstability number within the independent variable. The results, as illustrated in Figure6.22, revealed highly significant correlations for both N'/HR and Log (N'/HR), withrespect to cable bolt density. The best correlation was obtained for the plot of bolt densityversus N'/HR. The regression line shown in Figure 6.22a represents the line of best fitfor the stable cases of back support, and is proposed as a guideline for the determinationof cable bolt density for stope backs. The 68% confidence interval has been plotted onFigure 6.22 for both cases of regression analysis with N'/HR. Relating cable density tothe modified stability number allows for the consideration of the state of stress in asupported back. Although stress is often thought of as a clamping force in relation tostope backs, it can also act as a driving force in a sliding failure defined by joint structureintersecting the design surface. The relative block size factor does not consider the effectof stress or joint orientation. The ratio N'/HR is directly related to the revised ModifiedStability Graph and a range of selected ratios are plotted in Figure 6.23. Design cable boltdensities can be directly related from Figure 6.22a for each region on the revised ModifiedStability Graph. Chapter 7 will pursue this concept further in the context of design.6.6 CONCLUSIONSThe unsupported Modified Stability Graph (Potvin 1988) has been statisticallyvalidated and is recommended for the design of stope surfaces. The supportable region156Figure 6.19: Regression analysis with Q'157Figure 6.20: Regression analysis with Q'1 100.10.011 100.10.011 100.10.01A x Q'/HR&i x Q'/HR• Stable▪ Unstablev Caved••0.400.35E53,- 0.300.25a.)0.200.15(7)c=1 0.10oco0.050.00(c)O0• 0A2 x Q ./HR0.400.35Eg 0.30_ao 0.25-2 0.2000.15(,)a 0.10oco0.050.00(a)• Stable▪ Unstablev Caved•■ •••deo Jo.•0.400.35E5;0.30_ao 0.25fj 0.200U) 0.15• 0.10O0.050.00(b)• • • •■• •■• •• Stable■ Unstable• Caved• •gr.Linr'0.347•158Figure 6.21: Regression analysis with Q'159Figure 6.22: Regression analysis with N'160Figure 6.23: Cable support design ranges for the revised Modified Stability Graph161proposed by Potvin (1988) has been revised to incorporate a supported transition zone anda stable with support zone for use in the design of cable support. The supported transitionzone was found to intersect the unsupported transition zone, and suggests that there is a limitto the effect of cable support on the stability of large competent stope surfaces. Thecombined database suggests that the supported transition zone reflects a lower designconfidence than the stable with support zone. The failure regions identified by Potvin forthe Design Chart for Cable Bolt Density are in close agreement with the statistical divisionbetween stable and caved cases. No significant correlation however, was found to existbetween cable bolt density and the relative block size factor, RQD/Jn/HR. A statisticallysignificant relationship between cable bolt density and N'/HR is proposed for futuredesign purposes. The proposed relationship can be directly related to design ranges on theModified Stability Graph.162CHAPTER 7CABLE BOLT SUPPORT GUIDELINES7.1 INTRODUCTIONThe concepts of current cable bolt practice and design have been reviewed in the firstsection of this thesis. A statistical analysis was introduced in Chapter 6 to develop revised designguidelines for cable bolt support. This chapter will propose a methodology for cable design anddiscuss recommended design procedures.7.2 CABLE BOLT DESIGN METHODOLOGYA proposed design methodology is presented in Figure 7.1 based on the flowchartdiscussed in Section 4.1. The first priority in cable design is to assess the need for cable support.In the case of discrete design, this is related to the feature to be supported. The advantages oflength and bolt capacity are often considered as justification for the use of cables, but dueconsideration should be given to other support mechanisms. Prior exposure to cable bolt practiceis helpful, but not required to successfully implement a cable bolting proposal. Groutingequipment can be obtained on the rental market and cable bolt materials are readily availablefrom several suppliers. The concepts of discrete analysis have been discussed in Section 4.2, andshould be applied to isolated blocks or structure that require support. In the case of collectiveanalysis, a revised version of the Modified Stability Graph (Potvin 1988) is proposed in Figure7.2 for use in assessing the need of cable support. The supportable region originally proposedby Potvin (1988) has been divided into a stable with support zone and a supported transition164Figure 7.2: Proposed revisions to the Modified Stability Graph design regions165Table 7.1: Percentage of Supported Stable Cases in the Revised ModifiedStability Graph Design ZonesDESIGN ZONE % STABLEStable Zone 100%Unsupported Transition Zone 100%Stable with Support Zone 85%Supported Transition Zone 58%Caved Zone 25%zone, based on the statistical analysis discussed in Chapter 6. The supportable region includesboth of the above zones, but based on the combined database, increased design confidence issuggested in the stable with support zone, as summarized in Table 7.1. There is a notabledecrease in the percentage of stable cases plotting in the supported transition zone. All of thecases plotting in the stable zone or the unsupported transition zone are stable, and indicate a highdegree of design confidence. Cable support should be considered when a design surface plotswithin the supportable region of the Modified Stability Graph in Figure 7.2. The revisedsupportable region proposed for cable design is similar to the original Potvin (1988) proposal.The trend of the supported transition zone however, reflects a slight difference as it does notparallel the unsupported transition zone. It has been suggested in Section 6.5.2, that this mayreflect economic and technological constraints of supporting large competent surfaces. Futureimprovements in the application of cable support should move the supported transition zonefarther into the caved zone. If the revised Modified Stability Graph suggests the use of cables,collective analysis then considers the available access and potential design geometries. Cablesupport design has been classified into two approaches related to the potential pattern. Section1667.3 will discuss a design method to determine cable density for back support, where cables areevenly distributed over the surface. A proposed method for the design of point anchorhangingwall support is reviewed in Section 7.4. Prior to implementation, an economic analysisis a necessary constituent of both discrete and collective design proposals. Observation andmonitoring comprise the final and perhaps most important stage of the design process. Aninternal database and foundation for future design modifications can be developed through anassessment of support performance.7.3 PATTERN APPROACH TO CABLE DESIGNThe pattern approach to cable design applies to situations where cables are distributedevenly over the supported surface. Square and fan back patterns are the best application of thistype of support and comprise most of the cases in the combined database where an evendistribution of cables exist. Bolt density has been discussed in terms of the number of bolts persquare meter, and will be included in the pattern approach to cable design. A bolt densityconversion chart has been included in Appendix B, for use in converting bolt density frombolts/m2 to a square pattern equivalent in metric or imperial units.7.3.1 Cable Bolt Density for Back SupportA relationship between cable bolt density and the ratio N'/HR is proposed for the designof back support, where cables are evenly distributed over the supported surface. This relationshipis reflected in the Design Chart for Back Cable Support, illustrated in Figure 7.3. This chart isrecommended for use with square and fan back cable patterns. The design line indicated inFigure 7.3 represents the regression line obtained in Section 6.5.3. The regression analysis was168based on stable cases of back support within the combined database that fall in the supportableregion of the revised Modified Stability Graph. This design line reflects a minimum cable densitythat is higher than 59% of the stable cases in the combined database, and above the region thatcontains caved case histories. It is recommended that cable densities be located above the designline for cases that plot in the supportable region of the revised Modified Stability Graph. Areduction in design confidence within the supported transition zone suggests that a higher boltdensity should be considered for cases plotting within this zone. The upper limit of the 68%confidence band, or the 84% confidence line in Figure 7.3, is suggested as a minimum designdensity for the supported transition zone. This line represents a design level that is greater than84% of the cases within the combined stable database. No unstable or caved cases are locatedin the region above the 84% confidence line. Further research is required to calibrate regions ofhigher design confidence on the Design Chart for Back Cable Support. A proposal to relateN'/HR directly to the revised Modified Stability Graph is illustrated in Figure 7.4. Thesupportable region has been divided into design ranges determined by different ratios of N'/HR.Minimum design densities have been determined from the Design Chart for Back Cable Supportfor each range, as illustrated in Figure 7.5. The 84% confidence line has been used for theranges that plot within the supported transition zone. The Design Chart for Back Cable Supportis based on average conditions within a database of stable case histories. The majority of the casehistories within this database incorporate single cables and limited use of plates. The averagewater:cement ratio is approximately 0.40. Improvements in design confidence can be expectedwith the use of alternate cable geometries, water:cement ratios less than 0.40, and plates installedat the hole collar. Plates are recommended if ravelling of the rock mass is a possibility due tosmall block size or high critical bond lengths.169Figure 7.4: Proposed back cable support design ranges on the revised Modified Stability Graph170Figure 7.5: Proposed minimum bolt density design ranges for back cable support1717.3.2 Cable Bolt Length for Back SupportPotvin, Hudyma, and Miller (1989) related cable bolt length to the hydraulic radius ofthe supported surface for stope backs. Figure 7.6 illustrates this approach for cases of open stopeback support in the combined database that plot within the supportable zone of the revisedModified Stability Graph. Regression analysis was used to define a linear relationship betweencable length and hydraulic radius based on the stable cases within this database. A highlysignificant (r=0.495) relationship was revealed and is defined by(7.1)^Cable Length = 1.30 + 1.84HR.Figure 7.6 shows that the regression line is very similar to the design line proposed by Potvin,Hudyma, and Miller (1989), and it is suggested as a guideline for the determination of cablelength for back support. The suggested ratio of cable length to hydraulic radius ranges from 3.1to 2.0, and can be related to span as shown in Table 7.2.Table 7.2: Relationship Between Cable Length and Span for Back SupportSTOPE DIMENSIONSCABLE LENGTHLENGTH/SPAN HR/SPAN1:1 0.25 (0.5 to 0.8) x SPAN2:1 0.33 (0.7 to 1.0) x SPAN4:1 0.40 (0.8 to 1.2) x SPAN9:1 0.45 (0.9 to 1.4) x SPANDesign methods related to arch theory and discrete analysis presented in Chapter 4, oftensuggest lengths that permit cable anchorage in competent rock beyond a potential failure zone,173as illustrated in Figure 7.7a. It is important in cases of discrete analysis to anchor into competentmaterial. Economic conditions or large stope surfaces may produce a situation where cablescannot be anchored beyond a potential failure zone, as illustrated in Figure 7.7b. In terms ofcollective analysis, cable bolts act to reinforce the rock mass and develop the inherent strengthby limiting relative block movement. A reduction in stability however is expected where cablescannot penetrate beyond a potential failure zone, and the use of plates and strapping isrecommended to prevent ravelling of the rock mass.7.3.3 Other Support PatternsThe design guidelines proposed in this section have been discussed in terms of an evendistribution of cables over the supported surface, and include square back and fan back cablepatterns. Other patterns that were encountered in practice include point anchor backs, pointanchor hangingwalls, even hangingwalls, and hangingwall drift fans. It is suggested that theproposed design guidelines for cable length and density discussed in this section are appropriatefor use with all of the above, except point anchor hangingwalls. Even hangingwall patterns andhangingwall drift fan support reflect an even distribution of bolts over the supported surface.Point anchor hangingwalls and backs reflect a concentration of cables at particular points alongthe surface, that depend upon the available development. The average density of point anchorback support from Table 6.3 is 0.16 bolts/m2 for stable cases compared to 0.046 bolts/m 2 forstable point anchor hangingwalls. The higher densities are a result of smaller hydraulic radiicombined with improved accessability, and suggest that point anchor back support is similar toan even distribution of cables. The cases of point anchor back support in this database fit bestwith the design guidelines proposed in this section, but require further calibration. The boltPotential failure zone174Figure 7.7: Relation between potential failure zone and cable length175density of point anchor backs in this study was determined based on the number of installedbolts, divided by the surface area. Occasionally, cable back patterns are fanned into the footwalland hangingwall, making it difficult to calculate a true back cable density, especially in the caseof small openings. In this study, it was assumed that bolts fanned more than 45° from verticalwere not included in the density calculation for back support. Point anchor hangingwalls exhibitdensities that are well below patterns encountered for even cable distributions, and will bediscussed separately in Section 7.4.7.4 POINT ANCHOR APPROACH TO CABLE DESIGNFuller (1983b) made the distinction between a localized and uniform cable distributionin open stope hangingwalls. The purpose of the localized, or point anchor approach tohangingwall support, is described as dividing the hangingwall into smaller unsupported stablespans. The location of cable support is usually determined by sublevel development, and spanis therefore related to the distance between sublevels. Fabjanczyk (1982) suggests that this typeof support simulates a series of reinforced beams along the hangingwall. It has evolved as amethod of installing cable bolts where access is typically restricted. Beer, Meek, and Cowling(1983) have described the formation of individual plates parallel to the hangingwall, due todeformation into the open stope. The stability of the hangingwall is then related to beamthickness, frequency of cross jointing, surface dimensions, and ground support. Hangingwallstructure reviewed in this study was predominantly parallel to the surface, and stability seemedto be related to the unsupported span between sublevels and the size of blocks created byassociated cross jointing. This section will review applications of point anchor support andpropose a design method that relates factors controlling surface stability to the unsupported span.1767.4.1 Hangingwall Cable SupportAn interesting case of point anchor hangingwall support (Marshall 1963) was applied atthe Wilroy Mine in the early sixties. This particular installation was designed to reducehangingwall spalling and control dilution in the No. 3 orebody. The hangingwall and footwallrock was described as a fairly massive grey gneiss with occasional biotitic bedding planes andlittle alteration. The orebody was mined using sublevel blasthole benching with graduallyadvancing stoping blocks down dip. Hangingwall dilution was a concern during mining of thefirst three stopes. Efforts to control dilution included the separation of stoping blocks by sillpillars, leaving random pillars within the stope limits to reduce the unsupported wall dimensions,and leaving a skin of ore to absorb blast damage. Dilution however continued to remain high,and planning for the third stope included an increase in production rate to reduce exposure time,along with an attempt to pre-shear the hangingwall. In addition, 2.4 m rockbolts were installedwith strapping at intervals of 1.2 m along each sublevel. The support system was not successfulas rockbolt anchor slippage was noticed, and failed waste slabs that contained undisturbed boltsindicated that the support design length was insufficient. Breakage of some bolts also suggestedthat the support tensile strength was too low. The 1210 stope between the 10th and 12th levelwas the fourth stope to be mined, and included a point anchor approach to hangingwall supportusing 6.7 m (22') steel bars with a diameter of 28 6 mm (1 1/8"). The bolts were grouted alongthe hangingwall of each sublevel at a 2.4 m spacing, as illustrated in Figure 7.8. Bolts weregrouted with a mixture of water and High Early Strength cement in 76 2 mm (3") diameter holesthat were inclined downwards at 1 ° . A 63.5 mm (2.5") anchor nut and washer was installed atthe end of the bolt, and the bar was greased prior to installation. A bearing plate was placed atthe hole collar on a grout pad, and the bar was tensioned three days after grouting toapproximately 45 tonnes. The grouted bolt length was actually 6.1 m (20') since the collar end178protruded to allow for plate attachment and tensioning. The sublevels were slashed to full widthand eliminated problems of handling 6.7 m bolts. In cases of narrow ore width however, the useof coupled bars were required. At an estimated hydraulic radius of 19.3 m, this stope was minedsuccessfully using the point anchor approach to hangingwall support. Sublevel spacings variedfrom 19 to 21 meters along the hangingwall.Greenelsh (1985) describes the use of the point anchor approach to cable support in thedesign of the N663 stope experiment at Mount Isa Mine. The N663 stope was located at a depthof approximately 1000 meters, and was designed to test the feasibility of open stope mining inplace of mechanized cut and fill. The structure of the N663 stope hangingwall is characterizedby shale bedding planes that are described as being fissile, frequently smooth, continuous andgraphitic. Joints and faults frequently intersect the bedding planes. A typical section through theN663 stope from the 19C to 16E sublevel is shown in Figure 7.9. The stope was 170 meters inheight from the 19C to 16B sublevel, and was mined in two separate lifts. The lower section ofthe stope from the 19C sublevel to the limit of the upholes above 17L sublevel was mined first.The stope dimensions were 95 meters high, 15 meters wide, 20 meters along strike, and thehangingwall was dipping at 65° . These dimensions translate to a hangingwall hydraulic radiusof approximately 8.4 meters. Four double cable bolts were installed on rings spaced 2 metersapart on level 17 and 18B sublevel. Two meters of the cable bolts at the hole collar weredebonded to reduce the support stiffness. Steel straps were used at the hole collar in conjunctionwith barrel and wedge anchors, and the debonded section was tensioned to a 2 tonne load.Blastholes were 70 mm in diameter and rings were fired singly or in pairs. Instrumentation wasinstalled on the level 17 cables and rod extensometers were used to monitor hangingwallmovement. The N663 stope was mined successfully with an estimated hangingwall dilution ofjust over 3 % . Cable instrumentation indicated that the debonded section of cable at the holeF,--i, \--1 16E16B-- Top of lower sectionCablesMonitoringDriftN663 Stope(After Greenelsh 1985)179Figure 7.9: N663 stope case history of point anchor hangingwall cable support180collars exhibited loads close to the cable yield strength. Debonding is a useful designconsideration in hangingwall support, as large amounts of movement are expected near thesurface. The cable instrumentation revealed minimal loading at the toe of the holes and indicatedthat 9 to 12 meters was adequate for cable length.7.4.2 Design Chart for Point Anchor Hangingwall Cable SupportBeer, Meek, and Cowling (1983) proposed that hangingwall behaviour in bedded rockis based on joint frequency and spacing, joint frictional and cohesive properties, excavationgeometry, the virgin stress field, and ground support. Thirteen cases of point anchor hangingwallsupport were assembled from the database and are summarized in Table 7.3. Supported andunsupported spans have been defined as illustrated in Figure 7.10. Unsupported spans for aparticular case were variable, and the maximum unsupported span was used in this analysis. Itis proposed that the success of the point anchor approach to cable support is related to thedistance between sublevels and the rock mass block size. The Design Chart for Point AnchorHangingwall Cable Support is illustrated in Figure 7.11. The chart relates the maximumunsupported hydraulic radius to the relative block size factor expressed in terms of the supportedhydraulic radius. The point anchor database was used to derive a support line for design, basedon the method of discriminant analysis discussed in Section 6.4.2. The conditions of multivariatenormality and similar variance are not ideal due to the limited size of the database, but thesupport line does approximate a visual division between stable and caved points. The derivedstatistical line favoured the caved cases and was shifted vertically down in order to place all thecaved cases above the line. For design purposes, underground mapping and stope planning willgive an indication of the relative block size factor. An acceptable design is indicated byprojecting vertically up from the horizontal axis to the design line, and reading a recommendedTable 7 3: Point Anchor Hanainawall Supported DatabaseCase Stability SupportedHR (m)AverageDipStrike(m)SupportedSpan (m)# Spans Unsupported Soan (m) Unsupported HR (m) RQD/Jn RQD/JN/HR N' Length (L)(meters)L/Max. UnsupportedSpanMin. Max. Avg. Min. Max. Avp.3 Stable 11.7 80 66.4 35.9 2 14.4 21.3 17.9 5.9 8.1 7.0 6.7 0.57 5.6 6.1 0.294 Caved 19.1 80 59.7 106.1 4 21.3 32.2 26.6 7.7 10.7 9.2 6.7 0.35 5.6 8.1 0.198 Caved 17.1 82 67.2 69.6 3 19.9 28.8 23.3 7.8 9.9 8.7 13.7 0.80 61.4 6.1 0.217 Unstable 10.9 85 32.6 66.3 2 20.1 46.4 33.2 8.2 9.8 8.2 8.7 0.61 3.8 6.1 0.138 Unstable 12.7 84 40.2 67.3 2 19.8 47.4 33.6 8,8 10,9 9.2 6.7 0.53 5.6 6.1 0.139 Stable 13.2 87 68.6 44.2 2 18.5 25.6 22.1 7.2 9.4 8.4 11.8 0.89 56.9 6.1 0.2411 Stable 10.8 85 33.4 83.2 3 16.5 26.4 20.3 5.5 7.5 6.3 15 1.39 67.5 6.1 0.2320 Caved 12.4 62 33.3 103 2 29.0 74.0 51.5 7.8 11.5 10.1 15.8 1.27 15.8 14.6 0.2021 Stable 10.8 72 28.8 95.5 2 26.5 69.0 47.3 6.9 10.2 9.0 15.8 1.46 28.4 14.6 0.2126 Stable 11.5 79 27.7 139 2 66.0 76.0 71.0 10.0 9.8 10.0 15.8 1.37 33.2 14.8 0.1927 Stable 10.7 74 28.6 B4 2 21.8 66.0 43.9 6.2 10.0 8.7 15.8 1.48 30.8 18.3 0.2849 Stable 15.5 62 130 44 2 20.4 23.6 22.0 8.9 10.1 9.5 13.1 0.85 14.0 9.1 0.3950 Caved 17 82 111 82.1 2 23.6 38.5 31.1 10.1 14.3 10.3 8.3 0.49 4.4 9.1 0.2400N184unsupported hydraulic radius on the vertical axis. The unsupported hydraulic radius wasdetermined by considering the span and associated strike length for each sublevel interval. It isintended that this procedure be reversed to derive either an acceptable span or strike length tobe excavated. The state of stress and joint properties discussed by Beer, Meek, and Cowling(1983) are not directly included in this analysis. Due to their geometry, hangingwalls aretypically destressed and a zone of relaxation subject to the effects of gravity, is created. Pakalnis(1991) suggests that hangingwall dilution is generally due to slough within the relaxed zone. Allof the hangingwalls in this study were found to be in a state of relaxation, and therefore the stateof stress was not directly included in this analysis. The effect of joint properties and surfaceorientation were addressed by attempting to relate N'/HR, the modified stability number tosupported hydraulic radius ratio, directly to the unsupported hydraulic radius. No apparent designcriteria resulted from this analysis. Since the relaxed zone is subject to the effects of gravity, itis suspected that surface orientation should be included in hangingwall point anchor design. Thishas been accomplished to some degree by determining span and hydraulic radius based on thedip of the surface. The use of the gravity adjustment factor (C) from the modified stabilitynumber calculation, was related to the relative block size factor, but no design relationship wassuggested. It is recommended that further research into point anchor cable support consider theeffects of surface orientation and joint properties. The size of this database is limited, andadditional case histories are required to improve the reliability of this design method.The Design Chart for Point Anchor Hangingwall Cable Support is based on theassumption that the revised Modified Stability Graph can be used to determine if cable supportis required. Figure 7.12 shows that the case histories of point anchor cable support do notstrongly correlate with the supportable region of the revised Modified Stability Graph. Thecollection of additional case histories is required to improve upon this relationship. This design185Figure 7.12: Point anchor hangingwall database compared to the revised Modified Stability Graph186method also assumes that adequate support is installed at each sublevel, and it is recommendedthat the use of plates be incorporated. The average water:cement ratio for the point anchordatabase is in the 0.40-0.45 range. Due to the large surface areas involved, it is difficult to relatebolt density to the number of bolts per square meter of surface area, as proposed in the DesignChart for Back Cable Support. Preliminary guidelines for bolt density and length can be basedon current practice. Bolt density for point anchor support is related to the number of boltsinstalled on each ring, and the spacing between rings. The database in this study reflects anaverage of 4 bolts per ring and 2.4 meters between rings. If double cables are treated as twoseparate bolts, then an average of slightly over 5 cable strands per ring is reflected. It issuggested that the average values for ring spacing and the number of cable strands per ring, beused as preliminary design guidelines. Further calibration of these parameters is required.7.4.3 Cable Bolt Length For Point Anchor Hangingwall SupportIt is proposed that design bolt length for a point anchor approach to hangingwall cablesupport is related to distance between sublevels. The ratio of cable length to maximumunsupported span listed in Table 7.3 can be used as a preliminary guideline for the determinationof cable length. Table 7.3 indicates that cable length should exceed 25% of the maximumunsupported span.7.5 DESIGN CASE HISTORIESThe Design Chart for Point Anchor Hangingwall Cable Support is based on a minimalnumber of case histories, and further calibration is required. Two case histories of hangingwallsupport design will be briefly reviewed as an illustration of the proposed design method.1877.5.1 Detour Lake MineHangingwall cable design for the 560-660 Block at the Detour Lake Mine (Detour LakeMine 1992) is based on the point anchor approach described in this chapter. The structure of thehangingwall rock was assessed as indicated below:Q'= 68, A = 1.0, B = 0.2, C = 5.0, and N'= 68.RQD was estimated at 85 % and in was set to 3 based on one major joint set and additionalrandom jointing in the hangingwall rock. A typical section through the proposed stope is shownin Figure 7.13. The supported hydraulic radius is approximately 30 meters, based on a strikelength of 150 meters and a supported span of 100 meters. The relative block size factor is 0.9(RQD/Jn/Supported HR = 85/3/30) and the Design Chart for Point Anchor HW Cable Supportsuggests that the unsupported hydraulic radius should be kept to a maximum of 9.6 meters. Basedon the 150 meter strike length, this hydraulic radius can be back calculated to reflect a maximumunsupported span of 20.5m. Cable rings are planned every 3 meters along strike with threebolts/ring. Cable orientation is designed to establish a pattern approximately perpendicular to thehangingwall with cables at 0°, +20°, and +40° from horizontal.7.5.2 Wilroy MineIn an effort to compare the Wilroy case history discussed in Section 7.4.1 to the proposedDesign Chart for Point Anchor Hangingwall Cable Support, an estimate of the parametersinvolved was taken from information provided in the literature (Marshall 1963). This applicationrelates to the use of steel bar for support, but it is suspected that this could be correlated witha plated fan of cable bolts. The hangingwall is described as a massive rock and the RQD wasestimated at 100% . Gneiss is a metamorphic rock that is associated with a banded distributionand may grade to a schist (Kyrine and Judd 1957). Occasional biotitic planes at Wilroy were188Figure 7.13: Detour Lake Mine case history of point anchor hangingwall cable support189noted to affect the hangingwall stability. Based on a joint set number of 4, a joint roughnessnumber of 2, and a joint alteration number of 4, the Q' value is estimated at 12.5. Anapproximation of the stability number is outlined as follows:Q' = 12.5A = 1.0 since hangingwall in relaxationB = 0.3 as critical joint appears to be parallel to the hangingwallC = 6.5 for a 75° wallN' = 24.4.The hangingwall hydraulic radius was estimated at 19.2 m from Figure 7.8, and the surface plotswithin the caved zone of the revised Modified Stability Graph. This reflects some of theuncertainty with the relationship between design ranges on the revised Modified Stability Graphand point anchor hangingwall cable support. The maximum supported hydraulic radius wasestimated at 8.4 m for the lift between D Sublevel and 10 Level (Figure 7.8). With anunsupported relative block size factor (RQD/Jn/Supported HR) of approximately 1.3, the DesignChart for Point Anchor Hangingwall Support suggests that the 1210 stope should be stable. Theanalysis also suggests that the sublevel interval could be increased to reflect a maximumhydraulic radius of 10.5 meters. A back analysis of this hydraulic radius would suggest thatconsideration could be given to eliminating one sublevel.7.6 DISCUSSIONThe design proposals presented in this chapter have been derived from an empiricaldatabase assembled from Canadian hard rock mining experience. The collective design of cable190bolt support has been related to a revised version of the Modified Stability Graph, that wasoriginally proposed by Potvin (1988). The unsupported transition zone has been statisticallycalibrated with the addition of a new database, and is recommended for the design ofunsupported surfaces. The supportable region of the Modified Stability Graph (Potvin 1988) hasalso been calibrated, and a stable with support zone and a supported transition zone have beenproposed for the design of supported surfaces. Cable support is suggested for design surfaces thatplot within the supportable region of the revised Modified Stability Graph.No significant relationship was found between the relative block size factor and cable boltdensity. Block size will affect a cable bolt system, but the statistical analysis suggests that otherfactors are also involved. The modified stability number accounts for block size, surfaceorientation, joint properties, stress, and joint orientation. These factors have been successfullyrelated to stope design by Potvin (1988), and this thesis suggests that they are also related to thedetermination of cable bolt density. The Design Chart for Back Cable Support is recommendedfor use with fan and square back cable patterns. The application of even hangingwall and pointanchor back patterns to this chart is suggested, but requires further analysis. Minimum cable boltdensities have been related directly to the revised Modified Stability Graph.The Design Chart for Point Anchor Hangingwall Cable Support is proposed for use inthe determination of hangingwall cable patterns. Support design is related to rock mass blocksize, surface hydraulic radius, and the sublevel interval. It is proposed that the supportable regionof the revised Modified Stability Graph be used as an indicator of the requirement for cablesupport. The case histories discussed in Section 7.5 suggest that point anchor design may extendinto the caving zone of the revised Modified Stability Graph. The hangingwall database is limiteddue to restricted accessibility, and further case histories are required to calibrate the proposed191design criteria.The design proposals for point anchor and back cable support do not directly distinguishbetween cable geometry and the application of plates at the hole collar. The database of stablebacks collected in this study reflect the limited use of plates. Single cables were used in 60% ofthe cases. Increasing support stiffness by using double cables or installing plates at the holecollar, can be used as an additional safety factor in cable design. Plates prevent blocks fromsliding off the cable and maximize the available load carrying capacity. This is important inhangingwall design, since the surface is typically in a state of relaxation. The recommendednumber of cables for point anchor support has been based on the number of installed cablestrands. It is also recommended that cable distribution be maximized for point anchor support.There are several external factors that relate to successful cable design. Undercutting ofore contacts can frequently lead to a progressive failure of hangingwall surfaces, and detract fromthe performance of a cable installation. Drifts driven under surveyed or geological control, canstill break the ore contact due to the uncertainty in diamond drilling predictions. A skilleddevelopment crew can often follow a predetermined contact with only minor guidance forelevation. High density point anchor fans with plates were used on occasion to control, or stopprogressive hangingwall failures initiated from undercutting on previous mining horizons.Hangingwall failures can effect the stability of stope backs by increasing the exposed hydraulicradius. Blasting practice can also affect the performance of cable support. Bywater and Fuller(1983) have suggested that slot raises should be located on the footwall side of the stope, toreduce the effect of initial tight slot blasts on hangingwall stability. The use of decoupled chargesalong the hangingwall can also assist in this regard. Cable design should consider manpoweravailability and time constraints in relation to planned stope production. When time is a factor,production requirements often take priority, and support installations may be left uncompleted.192CHAPTER 8CONCLUSIONS8.1 INTRODUCTIONThe objective of this thesis was to expand upon the existing database of cable supportpractice and develop revised design criteria for the support of underground openings. Anempirical database was collected during an extensive field study that involved visits to operatingmines in Western Canada, the United States and Ireland. Guidelines proposed by Potvin (1988)for the design of open stope surfaces were applied to the data collection process. The Potvin(1988) stope design criteria were reviewed and revised in the context of cable support. Astatistical analysis was presented as a tool to aid in the interpretation of empirical data. Revisedguidelines are proposed in this thesis for use in the design of back and hangingwall cablesupport.8.2 CONCLUSIONSA new empirical database of 46 case histories has been assembled during this study toexpand upon the existing knowledge of cable support in underground mining applications. Thenew data was combined with the Potvin (1988) database to develop design guidelines for cablesupport. Improvements in blast technology, monitoring, quality control and production methodswill inevitably alter the picture described by this empirical database. The process of calibrationand revision suggested in this thesis should continue, in order to reflect technology improvementsand changes in operational procedure.193A methodology for cable design is proposed based on a distinction between discrete andcollective analysis. Discrete analysis should be applied to isolated blocks or structure that requiresupport. Collective analysis applies to the rock mass, rather than distinct features, and providesa method of relating structure and stress to the design of cable support. The Potvin (1988)approach has related characteristics of the rock mass to the design of open stope surfaces, byrelating a modified stability number to the surface hydraulic radius on the Modified StabilityGraph. The unsupported Modified Stability Graph proposed by Potvin (1988) has beenstatistically verified using a method of discriminant analysis, and is recommended for use in thedesign of open stope surfaces. The supportable region defined on the supported Modified StabilityGraph (Potvin 1988) has been revised for use in cable design. The revised supportable regionincorporates a stable with support zone and a supported transition zone that are based on differentlevels of design confidence. The revised supportable region is very similar to that proposed byPotvin (1988), but does not parallel the unsupported transition zone. This suggests that there isa limit to the effectiveness of cable support with respect to large competent stope surfaces.Consideration of cable support is recommended for surfaces that plot in the supportable regionof the revised Modified Stability Graph (Figure 7.2). This thesis advocates the use of on-sitecalibration as the ideal method of applying empirical design techniques. The mining engineer willfind that the best use of time is often spent in observing underground geotechnical activity inrelation to production. The revised Modified Stability Graph provides a tool for the miningengineer to use in the documentation process, and the development of site specific designguidelines. A rough relationship between the increase in rock mass quality due to the additionof support can also be derived. The stable zone for an unsupported surface can be related to thestable with support zone for a supported surface. This provides a method of quantifying the effect194of support in terms of either an increase in Q', or RMR. This is a valuable concept but requiresadditional study.The collective design process uses the revised Modified Stability Graph to determine ifcable support is warranted for a particular surface. Cable design has been classified into twocategories, a pattern approach and a point anchor approach. The pattern approach involves aneven distribution of cable support over the design surface, and is the most common methodencountered in practice. The point anchor approach is characterised by a large concentration ofbolts placed at particular points along a surface. Two design methods that reflect the distinctionbetween the point anchor and pattern approaches, are proposed in this thesis.The new database was compared with the Design Chart for Cable Bolt Density that wasproposed by Potvin (1988) for the design of cable support for stope backs. Close agreement wasfound with the stability conditions proposed by Potvin (1988), but no significant statisticalrelationship was found between cable density and the relative block size factor. A statisticalanalysis in this thesis suggests that joint properties, stress, and joint orientation should also beconsidered in the determination of cable density for stope backs. The modified stability numberproposed by Potvin (1988) and the surface hydraulic radius have been statistically related to cabledensity in the Design Chart for Back Cable Support. It is recommended for use in thedetermination of cable patterns for back support, where cables are evenly distributed over theentire surface. Cable length for back support has been related to the surface hydraulic radius.Point anchor hangingwall cable design is related to the determination of a maximum stableunsupported span between beams of high density cable fans. The block size and surfacedimensions have been related to unsupported span on the Design Chart for Point AnchorHangingwall Cable Support. Attempts to include the modified stability number and surface195orientation into the design proposals were unsuccessful. Cable support is based on averageconditions. It is recommended that cable rings be spaced 2.4 meters along strike and five platedcables be incorporated within each ring. The point anchor design proposals are based on a limiteddatabase and require additional calibration. Cable length for point anchor cable support is relatedto the distance between sublevels.8.3 FUTURE WORKPakalnis et al. (1987) indicate that dilution is used as a measure of stope design quality,but is not necessarily defined in the same manner by open stope operators. In this study, designdilution estimates were difficult to obtain, but actual dilution values were frequently recorded.This indicates that the relationship between design and actual dilution merits further investigation.Non-entry mining methods typically exhibit a tolerance for instability that is difficult to describe,but can be related to the degree of dilution. It is recommended that dilution be considered infuture expansion of the empirical database, and related to design ranges on the revised ModifiedStability Graph. The influence of economic and operational parameters should also beconsidered.A review of design practice has indicated that cable length is often related to a zone ofinstability. Pakalnis (1991) suggests that hangingwall dilution is usually a result of slough withinthe zone of relaxation. A parametric modelling study (Pakalnis 1991) showed that the zone ofrelaxation for hangingwalls, predicted by traditional two dimensional modelling, can be 300%higher than three dimensional modelling predictions. Further examination of the relationshipbetween the zone of relaxation obtained from three dimensional modelling, and cable length, issuggested.196The characteristics of grout flow in a cable bolt hole are difficult to observe and arefrequently related to deficiencies in cable bolt support. Chapter 2 notes that improvements ingrout quality can significantly improve cable load carrying capacity. Further research in this areais recommended to consider the variables involved in grout flow. The analysis should considerpumping equipment, installation methods, and properties of grout mixtures. Steel pipes have beenused to simulate underground cable installations (Cluett 1991). This type of laboratory testingwould be useful in the evaluation of grout flow under operating conditions.The Design Chart for Back Cable Support has been proposed for use with all forms ofcable support that feature an even distribution of cables over the design surface. Most of thedatabase used in the determination of this chart was based on case histories of square back andfan back cable patterns. Additional case histories of point anchor back and even hangingwallsare required to verify that they agree with the proposed design ranges.The Design Chart for Point Anchor Hangingwall Cable Support is based on a limiteddatabase. The chart requires the collection of additional case histories to calibrate the proposeddesign ranges. It is recommended that future research consider point anchor hangingwall cabledesign in terms of surface orientation and the modified stability number. Mandolin bolting wasalso encountered on a limited basis for hangingwall support. This methods simulates a cablesling, and may merit further investigation for small hangingwalls.Discrimination based on the Mahalanobis distance, has been introduced as a method ofseparating a multidimensional database based on stability. It produces a division between twoclasses of points, but since it is a linear technique, it is not sensitive to non-linear division. It istherefore difficult to use this method to extrapolate a division between classes beyond the rangeof collected data. With this limitation in mind, discrimination based on the Mahalanobis distanceis recommended for use with forms of empirical analysis described in this thesis. A logarithmic197transformation has been successfully applied to meet the conditions of multivariate normality andsimilar variance. There are other methods of statistical analysis that deal with data classificationor group assignment. Cluster analysis is one such technique that is concerned with theidentification of groupings within a database (Manly 1986, 13). The method of discriminantanalysis discussed in this thesis, assumes that the database is separated based on the stabilitycondition of the surface. Cluster analysis is a numerical process that determines the number ofclasses based on the database. It is inherently more complex and may not apply to the type ofdatabase discussed in this thesis, but it is recommended for further study. It is not restricted toa linear separation between classes, but instead identifies the nature of true groupings within adatabase.8.4 FINAL REMARKSThe design methods proposed in this thesis have been derived from an empirical databasebased on Canadian hard rock mining experience. They are not intended to suggest rigidguidelines, but instead to provide a wide degree of latitude in the design of cable supportsystems. The techniques described in this thesis can be applied at operating mines to develop sitespecific criteria. Expansion of the empirical database and calibration of the proposed designguidelines, is required to account for operational and technological improvements.198BIBLIOGRAPHYBarton, N., R. Lien, and J. Lunde. 1974. Engineering classification of rock masses for thedesign of tunnel support. Rock Mechanics. 6 (May 1974): 189-236.Bawden, W.F., A.J. Hyett, and D. Cortolezzis. 1992. Towards a methodology for performanceassessment in cable bolt design. In Rock support in mining and underground construction: Proceedings of the international symposium on rock support in Sudbury, Ontario, June16-19, 1992, edited by P.K. Kaiser and D.R. McCreath, 277-284. Rotterdam andBrookfield: A.A. Balkema.Bawden, W.F., G. Sauriol, D. Milne, and P. Germain. 1989. Practical rock engineering stopedesign case histories from Noranda Minerals Inc. CIM Bulletin 82 (July 1989): 37-45.Beer, G., J.L. Meek, and R. Cowling. Prediction of the behaviour of shale hanging walls indeep underground excavations. In • • lin  s f fifth on , s of the InternationalSociety for Rock Mechanics held in Melbourne, Australia in 1983, 2, D45-D51.Rotterdam: A.A. Balkema.Bieniawski, Z.T. 1976. Rock mass classifications in rock engineering. In Proceedings of thei m sn ex 1 r. on f rr k •n me•rin h- l• in oh nne r fr m ovember 1-III5, 1976, edited by Z.T. Bieniawski, 97-106. Rotterdam, A.A. Balkema.Bourchier, F., E. Dib, and M. O'Flaherty. 1992. Practical improvements to installation of cablebolts: Progress at Campbell Mine. Iniloct—c ,support in mining and undergroundconstruction: Proceedings of the international symposium on rock support in Sudbury, Ontario, June 16-19, 1992, edited by P.K. Kaiser and D.R. McCreath, 311-318.Rotterdam and Brookfield: A.A. Balkema.Bywater, S. and P.G. Fuller. 1983. Cable support of lead open stope hangingwalls at Mount IsaMines Limited. In Pr* -^of int mn sm.16^ n r k bol 'n in A i kAugust 28 - September 3, 1983, 539-555.Choquet, P. and F. Miller. 1988. Development and field testing of a tension measuring gaugefor cable bolts used as ground support. CIM Bulletin 81 (July 1983): 53-59.Cluett, J.L. 1991. HBM & S cable bolt practices at Trout Lake Mine - Our success and failures.Paper presented at the 93rd Annual General Meeting of the Canadian Institute of Mining,Metallurgy, and Petroleum, Vancouver, April 28 - May 2, 1991.Cullen, M. 1991. Empirical and analytical excavation design at the H-W Mine. Paper presentedat the 93rd Annual General Meeting of the Canadian Institute of Mining, Metallurgy, andPetroleum, Vancouver, April 28 - May 2, 1991.16199Curran, J.H. and B.T. Corkum. 1989. Examine 2D boundary element program manual, R 2.1.Detour Lake Mine. 1992. Internal report to Detour Lake Mine - DLM-11-92.Fabjanczyk, M. 1982. Review of ground support practice in Australian undergroundmetalliferous mines. In Proceedings of the annual conference of the Australian Instituteof Mining and Metallurgy held in Melbourne in August 1982, 511-522.Fraser, H. 1976. Mechanical anchors for tensioned cable bolts at International Nickel's BirchtreeMine, Manitoba Division, Thompson Manitoba. Paper presented at the UndergroundOperators Conference of the Metal Mining Division of the Canadian Institute of Miningand Metallurgy, Val d'Or, February 23-26, 1976.Fuller, P.G. 1981. Pre-reinforcement of cut and fill stopes. In Application of rock mechanicsto cut and fill mining: Proceedings of the conference on the application of rock mechanicsto cut and fill mining at the University of Lulea, Sweden, June 1-3, 1980, edited by OveStephansson and Michael Jones, 55-63. London: The Institution of Mining andMetallurgy.Fuller, P.G. 1983a. Cable support in mining: A keynote lecture. In Proceedings of theinternational symposium on rock bolting in Abisko, August 28 - September 2. 1983,  511-522. Rotterdam: A.A. Balkema.Fuller, P.G. 1983b. The potential for cable support of open stopes. In Proceedings of the 5thinternational congress on rock mechanics held in Melbourne, Australia, 1983,  2, D39-D44. Melbourne: A.A. Balkema.Fuller, P.G., and R.H.T. Cox. 1975. Mechanics of load transfer from steel tendons to cementbased grout. In Proceedings of the fifth Australasian conference on the mechanics ofstructures and materials in Melbourne, Australia, August 25-27, 1975,  189-203.Fuller, P.G., P.M. Dight, and D. West. 1990. Cable support design for underground mines.Paper presented at the 92nd Annual General Meeting of the Canadian Institute of Miningand Metallurgy, Ottawa, Ontario, May 6-10, 1990.[Garcia, John A.]. 1929. Operating ideas from production, electrical and mechanical men. CoalAge. 34 (June 1929): 380-385.Gendron, A., D. Jacob, Y. Potvin, and D. Milne. 1992. Grout evaluation for cable bolt support.In Rock support in mining and underground construction: Proceedings of the internationalsymposium on rock support in Sudbury, Ontario, June 16-19, 1992,  edited by P.K.Kaiser and D.R. McCreath, 335-339. Rotterdam and Brookfield: A.A. Balkema.Goris, J.M. 1990. Laboratory evaluation of cable bolt supports (in two parts) 1. Evaluation ofsupports using conventional cables.  United States. Bureau of Mines. RI 9308.200Goris, J.M. 1991. Laboratory evaluation of cable bolt supports (in two parts) 2. Evaluation of in on en on.1 ble ith s 1 b ons irdc e bl s nd - •ox -coatedI LScables.United States. Bureau of Mines. RI 9342.Gramoli, L. 1975. Tensioned^ f Noranda Mines tdPaper presented at the First Annual Underground Operators Conference of the CanadianInstitute of Milling and Metallurgy, Sudbury, February 25-28, 1975.Greenelsh, R.W. 1985. The N663 stope experiment at Mount Isa Mine. International Journal ofMining Engineering. 3: 183-194.Greer, G.J. 1989. Empirical modelling of open stope stability in a vertical crater retreatapplication at Inco's Thompson Mine. Paper presented at the 91st Annual GeneralMeeting of the Canadian Institute of Mining and Metallurgy, Quebec City, April 30 -May 3, 1989.Hoek, E. and E.T Brown. 1980. Underground Excavations in Rock. London: The Institution ofMining and Metallurgy.Hunt, R.E.B, and J.E. Askew. 1977. Installation and design guidelines for cable dowel groundsupport at ZC/NBHC. In Proceedings of the Australian Institute of Mining andMetallurgy, Broken Hill Branch. Underground Operators Conference, Australia, October1977. 113-122.Hyett, A.J., W.F. Bawden, and A.L. Coulson. 1992. Physical and mechanical properties ofnormal Portland cement pertaining to fully grouted cable bolts. In Iocksupwgin miningand underground construction: Proceedings of the international symposium on rocksuppo1992, edited by P.K. Kaiser and D.R.McCreath, 341-348. Rotterdam and Brookfield: A.A. Balkema.Jaeger, J.C. and N.G.W. Cook. 1979. Fundamentals of Rock Mechanics.  3rd ed. New York:Chapman and Hall Ltd.Jeremic, M.L., and G.J.P. Delaire. 1983. Failure mechanics of cable bolt systems.CIM Bulletin 76 (August 1983): 66-71.Johnson, R.A., and D.W. Wichem. 1988. Applied multivariate statistical analysis,  2d ed.Englewood Cliffs, N.J.: Prentice-Hall, Inc.Johnson, R.R. 1976. Elementary Statistics. 2nd ed. North Scituate, Massachusetts: DuxburyPress.Kaiser, P.K., S. Maloney, and S. Yazici. 1992. A new perspective on cable bolt design. CIMBulletin 85 (July/August 1992): 103-109.201Krynine, D.P. and W.R. Judd. 1957. Principles of engineering geology and geotechnics. NewYork: McGraw-Hill Book Company, Inc.Lang. 1961. Theory and practice of rock bolting. Transactions of the American Institute ofMining. Metallurgical, and Petroleum Engineers 220: 333-348.Laubscher, D.H. 1976. The importance of geomechanics classification of jointed rock massesin mining operations. In Proceedings of the symposium on exploration for rockengineering held in Johannesburg from November 1-5. 1976, edited by Z.T. Bieniawski,119-128. Rotterdam, A.A. Balkema.MacSporran, G.R., W.F. Bawden, A.J. Hyett, D.J. Hutchinson, and P.K. Kaiser. 1992. Anempirical method for the analysis of failed cable bolted ground.  Paper presented at therock mechanics and strata control session at the 94th Annual General Meeting of theCanadian Institute of Milling, Metallurgy and Petroleum, Montreal, April 26-30, 1992.Manly, B.F.J. 1986. Multivariate statistical methods: a primer.  London: Chapman & Hall.Marriott, F.H.C. 1974. The interpretation of multiple observations.  London: Academic Press.Quoted in R.A. Johnson and D.W. Wichern. Applied multivariate statistical analysis, 2ded, 3. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1988.Marshall, D. 1963. Hanging-wall Control at Willroy.CIM Bulletin 56 (April 1963): 327-331.Mathews, K.E., E. Hoek, D.C. Wyllie, and S.B.V. Stewart. 1981. Prediction of stablex vtion I.n fir^n^t • ' I s ►1.^11,11 m- r in h, • rick. Canada:CANMET, Department of Energy, Mines and Resources. DSS Serial No. OSQ80-00081,DSS File No. 17SQ.23440-0-9020.Matthews, S.M., V.H. Tillmann, and G. Worotnicki. 1983. A modified cable bolt system forthe support of underground openings. In Proceedings of the annual conference of theAustralasian Institute of Mining anclMetallurgyin&Dlecn Hill, New South Wales in July, 1983, 243-255. Parkville: The Australasian Institute of Mining and Metallurgy.Miller, D.R. 1984. Design of cable reinforcement pattern to resist shear failure in open stopewalls. In Stabili in  Proceedings of the second internationalconference on stability in underground mining held in Lexington, Kentucky August 6-8, 1984, edited by A.B Szwilski and C.O. Brawner, 346-362. New York: Society of MiningEngineers of the American Institute of Mining, Metallurgical and Petroleum Engineers,Inc.Milne, D. and Y. Potvin. 1992. Measurement of rock properties for mine design. In Proceedingsof the ISRM symposium on rock characterization: Eurock '92 in Chester, England. September 14-17, 1992, edited by J.A. Hudson, 245-250. London: Thomas TelfordServices Ltd.202Noranda Technology Center. 1990. Noranda Geomechanics Instrumentation and Cable BoltingShort Course - Ansil Mine. Notes.Oliver, P.H. 1992. The evolution of the thick grout cable bolting system at Inco's Sudbury areaoperations. In Rock support in mining and underground construction: Proceedings of theinternational symposium on rock support in Sudbury, Ontario, June 16-19. 1992,  editedby P.K. Kaiser and D.R. McCreath, 303-309. Rotterdam and Brookfield: A.A. Balkema.Pagel, C.W. 1987. Cable bolting for ground support at No. 12 Mine of Brunswick Mining. InUnderground Support Systems: CIM special volume 35,  edited by J.E. Udd, 59-62.Montreal: Canadian Institute of Mining and Metallurgy.Pakalnis, R.C.T. 1986. Empirical stope design at Ruttan.Ph.D. dissertation, University of British Columbia.Pakalnis, R.C.T. 1991. Three dimensional modelling - an applied approach. In Proceedings ofthe second Canadian conference on computer applications in the mineral industry held inVancouver, from September 15-18, 1991,  2, edited by R. Poulin, R.C.T. Pakalnis andA.L. Mular, 619-630. Vancouver: Department of Mining and Mineral ProcessEngineering, University of British Columbia.Pakalnis, R., H.D.S. Miller, S. Vongpaisal, and T. Madill. 1987. An empirical approach toopen stope design. In Proceedings of the sixth international congress on rock mechanicsheld in Montreal, Quebec, 1987. 1191-1196.Pakalnis, R., D. Tenney, and B. Lang. 1991. Numerical modelling as a tool in stope design.CIM Bulletin 84 (July 1991): 64-73.Palmstrom, A. 1985. Application of the volumetric joint count as a measure of rock massjointing. In Proceedingsf the imernational symposium on fundamentals of rock jointsheld in BjOrkliden from September 15-20, 1985,  103-11.Potvin, Yves. 1988. Empirical open stope design in Canada.Ph.D. dissertation, University of British Columbia.Potvin, Y., M. Hudyma, and H.D.S. Miller. 1989. Design guidelines for open stope support.CIM Bulletin 82 (June 1989): 53-62.Potvin, Y. and D. Milne. 1992. Empirical cable bolt support design. In Rock support in miningn un r' n ons i. ti S n . Pr ins i f e international symposium on rocksupport in Sudbury, Ontario, June 16-19,1992, edited by P.K. Kaiser and D.R.McCreath, 269-275. Rotterdam and Brookfield: A.A. Balkema.Priest, S.D. and J.A. Hudson. 1976. Discontinuity spacings in rock. Int. J. Rock Mech. MM. Sci. & Geomech. Abstr. 13: 135-148.203Reichert, R.D., W.F. Bawden, and A.J. Hyett. 1992. Evaluation of design bond strength forfully grouted cables. CIM Bulletin 85 (July/August 1992): 110-118.Schmuck, C.H. 1979. Cable bolting at the Homestake Gold Mine. Mining Engineerin' 31(December 1979): 1677-1681.Seber, G.A.F. 1984. Multivariate Observations. New York: John Wiley & Sons, Inc.Spiegel, M.R. 1975. Schaum's outline of theory and problems of probability and statistics.  NewYork: McGraw-Hill, Inc.Stheeman, W.H. 1982. A practical solution to cable bolting problems at the Tsumeb Mine.CIM Bulletin 75 (February 1982): 65-77.Stillborg, Bengt. 1986. Professional users handbook for rock bolting.  Federal Republic ofGermany: Trans Tech Publications.Thompson, Alan G. 1992. Tensioning reinforcing cables. In Rock support in mining andunderground construction: Proceedings of the international symposium on rock supportin Sudbury, Ontario, June 16-19, 1992, edited by P.K. Kaiser and D.R. McCreath, 285-291. Rotterdam and Brookfield: A.A. Balkema.Webster's New World Dictionary of American English, Third College Edition.  1991 ed. S.v."Empirical".Wilkinson, Leland. 1990a. Sygraph: The system for graphics.  Evanston, IL: Systat, Inc.Wilkinson, Leland. 1990b. Systat: The system for statistics. Evanston, IL: Systat, Inc.Windsor, C.R. 1992. Invited lecture: Cable bolting for underground and surface excavations.In Rock support in mining and underground construction: Proceedings of the internationalsymposium on rock support in Sudbury, Ontario, June 16-19, 1992, edited by P.K.Kaiser and D.R. McCreath, 349-366. Rotterdam and Brookfield: A.A. Balkema.APPENDIX A204DATABASE SUMMARYTABLE A.1: Case Study Summary (Case 1 to 6) CASE # 1 2 3 4 5 6Surface HW Back HW HW Back HWCables/Hole 2 2 2 2 2 2Hole Length (m) 9.1 - 18.3 6.1 6.1 6.1 9.1 6.1Pattern HW Drift Fan Point Anchor Square Point AnchorDensity (bolts/m2) 0.018 0.018 0.16 0.022Equiv. Pattern (ft x ft) 24.5 24.46 8.2 22.12Stability Caved Caved Stable CavedStrike Length (m) 32 • 59.7 26.5 67.2Average Width (m) 4.6 8.1 4.1 6.1Dip (degrees) 74 80 76 82Vertical Height (m) 69.5 104 33 69Area (m2) 2063 6345 167 4826Perimeter (m) 207 • 332 64 282Hydraulic Radius (m) 10.0 2.3 19.1 2.6 17.1Mining Method Blasthole Blasthole Blasthole BlastholeDilution (%) 24 • 42 25 66Cable Diameter (mm) 15.9 15.9 15.9 15.9 15.9Hole Diameter (mm) 51 51 51 51 51Cost ($/m) 29.46 29.46 29.46 29.46 29.46Bonus $40/Shift $40/Shift $40/Shift $40/ShiftSupplier Thiessen Thiessen Thiessen ThiessenTiming 7-14d 7-14d 14d 7-14dPlates No No 305 x 305 mm 305 x 305 mm No NoStraps No No No No No NoOther Support No No No No No NoGrout Pump Spedel 6000 Spedel 6000 Spedel Spedel 6000 Spedel 6000 Spedel 6000Water:Cement Ratio 0.45 0.45 0.45 0.45 0.45 0.45Grout Tube Dia. (mm) 19 19 19 19 19 19Breather Tube Dia. (mm) 9.5 9.5 9.5 9.5 9.5 9.5Rock Type AA & CA NS TS CA/AARMR' 56 59 40 40 59 70Q' 11.7 11.7 2.5 2.5 11.7 27.3A 1.0 1.0 1.0 1.0 1.0 1.0B 0.2 0.2 0.3 0.3 0.2 0.3C 6.0 2.0 7.5 7.5 2.0 7.5N' 14.0 4.7 5.6 5.6 4.7 61.4RQD 70 70 40 40 70 82Jn 6 6 6 6 6 6Jr 2 2 1.5 1.5 2 2Ja 2 2 4 4 2 1RQD/Jn/H.R. 1.17 5.09 0.57 0.35 4.50 0.80Blasthole Dia (mm) 51 51 51 51 51 51Backfill None None None None None C&F Below205TABLE A.2: Case Study Summary (Case 7 to 12) CASE # 7 8 9 10 11 12Surface HW HW HW Back HW BackCables/Hole 2 2 2 2 2 2Hole Length (m) 6.1 6.1 6.1 22 6.1 7.6Pattern Point Anchor Point Anchor Point Anchor Square C&F Point Anchor FanDensity (bolts/m2) 0.011 0.020 0.018 0.13 0.023 0.58Equiv. Pattern (ft x ft) 31.28 23.2 24.46 9.10 21.63 4.31Stability Unstable Unstable Stable Stable Stable StableStrike Length (m) 32.6 40.2 68.6 21.0 33.4 21.9Average Width (m) 9.1 7.0 4.9 15.5 8.3 4.0Dip (degrees) 85 84 87 85 85 N/AVertical Height (m) 66.3 68 44 4.3 63 3.7Area (m2) 2171 2897 3103 412 2069 52Perimeter (m) 199 10.922912.723513.281.7 5.019210.833 1.6Hydraulic Radius (m)Mining Method Blasthole Blasthole Blasthole Cut and Fill Blasthole DriftDilution (%) 43 30 30 10-15 30 N/ACable Diameter (mm) 15.9 15.9 15.9 15.9 15.9 15.9Hole Diameter (mm) 51 51 51 57 51 51Cost ($/m) 29.46 29.46 29.46 40.3 29.46 29.46+Bonus $40/Shift $40/Shift $40/Shift Contractor $40/Shift $40/Shift+Supplier Thiessen Thiessen Thiessen Thiessen Thiessen ThiessenTiming 14-28d 7-14d 14-28d 28d+ 7-14d 28d+Plates 305 x 305 mm 305 x 305 mm No No No 305 mm dia.Straps No No No No No Steel SetsOther Support No No No Swellex No Swellex/RBGrout Pump Spedel 6000 Spedel 6000 Spedel 6000 Spedel 6000 Spedel 6000 Spedel 6000Water:Cement Ratio 0.45 0.45 0.45 0.45 0.45 0.45Grout Tube Dia. (mm) 19 19 19 19 19 19Breather Tube Dia. (mm) 9.5 9.5 9.5 9.5 9.5 9.5Rock Type NS NS CA TS CA TSRMR' 40 40 70 75 75Q' 2.5 2.5 23.7 18.8 30 18.8A 1.0 1.0 1.0 0.2 1.0 0.1B 0.2 0.3 0.3 0.3 0.3 0.2C 7.5 7.5 8.0 2.0 7.5 2.0N' 3.8 5.6 56.9 2.3 67.5 0.75RQD 40 40 71 75 90 75Jn 6 6 6 6 6 6Jr 1.5 1.5 2 1.5 2 1.5Ja 4 4 1 1 1 1RQD/Jn/H.R. 0.61 0.53 0.89 2.50 1.39 7.81Blasthole Dia (mm) 51 51 & 76 51 38 51 38Backfill 3&F Below C&F Below C&F Below Sand None None206TABLE A.3: Case Study Summary (Case 13 to 18) CASE # 13 14 15 16 17 18Surface Back Back Back Back Back BackCables/Hole 2 2 2 2 2 2Hole Length (m) 6.1 14 8 & 15 8 & 12 12 & 15 9.8Pattern Square Point Anchor Square Point Anchor Square SquareDensity (bolts/m2) 0.116 0.18 0.28 0.18 0.14 0.29Equiv. Pattern (ft x ft) 9.63 7.80 6.20 7.73 8.77 6.09Stability Stable Stable Stable Caved Caved StableStrike Length (m) 24.4 15 20 34 62 28.4Average Width (m) 7.3 20 27 21Dip (degrees) 84 15 30-40 30-40 68Vertical Height (m) 35.4 22 28 95Area (m2) 231.9 300 1300 1930 1604 358Perimeter (m) 64.9 70 170 172 187 85Hydraulic Radius (m) 3.6 4.3 7.6 11.2 8.6 4.2Mining Method Blasthole/Sh Blasthole Blasthole Blasthole Blasthole BlastholeDilution (%) 25 18Cable Diameter (mm) 15.9 15.9 15.9 15.9 15.9 & 12.7 15.9Hole Diameter (mm) 51 51 & 57 51 & 57 51 & 57 51 & 57 51Cost ($/m) 29.46 16.40 19.71 19.71 19.71 28.84Bonus $40/Shift $45-55/MS $45-55/MS $45-55/MS $45-55/MSSupplier Thiessen ThiessenTiming 7-14d 28dPlates No No No No No 254 x 254 mmStraps No No No No No NoOther Support No RB RBGrout Pump Spedel 6000 Spedel 6000 Spedel 6000 Spedel 6000 Spedel 6000 Spedel 6000Water:Cement Ratio 0.45 0.55 0.55 0.45Grout Tube Dia. (mm) 19 19 19 19Breather Tube Dia. (mm) 9.5 9.5 9.5 9.5Rock Type CA/AA Fault Fault Fault Fault MSRMR' 56 28 80Q' 11.7 0.6 0.5 0.5 0.6 13.3A 1.0 1 1 1 1.0 0.1B 0.2 0.4 0.3 0.3 0.2 0.3C 2.0 2.0 3.0 3.0 2.0 2.0N' 4.7 0.48 0.45 0.45 0.24 0.80RQD 70 5 10 10 5 80Jn 6 3 4 4 3 6Jr 2 1.5 1.5 1Ja 2 4 4 1RQD/Jn/H.R. 3.25 0.39 0.33 0.22 0.19 3.17Blasthole Dia (mm) 51 51? 51 114Backfill None Yes Yes No Yes (P5 Tight) None207TABLE A.4: Case Study Summary (Case 19 to 241CASE # 19 20 21 22 23 24Surface Back HW HW Back Back BackCables/Hole 2 1 2 N/A 2 N/AHole Length (m) 9.8 14.6 14.6 N/A 9.8 14.6Pattern Sq & HW Fan Point Anchor Point Anchor N/A Sq. & HW Fan HW/FW FanDensity (bolts/m2) 0.27 0.035 0.031 N/A 0.33 N/AEquiv. Pattern (ft x ft) 6.31 17.54 18.63 N/A 5.19 N/AStability Unstable Caved Stable Caved Stable CavedStrike Length (m) 61 33.3 28.8 41.5 26 41.2Average Width (m) 11.9 10 62 13.2 15 14Dip (degrees) 70 62 72 84 69 70Vertical Height (m) 97 89 90 59 55 56.8-70Area (m2) 742 3187 2477 644 455 573Perimeter (m) 144.1 256.4 229.6 103.4 80 109.3Hydraulic Radius (m) 5.2 12.4 10.8 6.2 5.2 5.2Mining Method Blasthole Blasthole Blasthole Blasthole Blasthole BlastholeDilution (%) 19 30 26.9 58.4 28.4Cable Diameter (mm) 15.9 15.9 15.9 N/A 15.9 N/AHole Diameter (mm) 51 51 51 N/A 51 N/ACost ($/m) 28.84 28.84 28.84 N/A 28.84 N/ABonus N/A N/ASupplier Thiessen Thiessen Thiessen N/A Thiessen N/ATiming 28d 28d 28d N/A 28d N/APlates 254 x 254 mm 102 x 102 mm 102 x 102 mm N/A 102 x 102 mm N/AStraps No No No No No N/AOther Support RB No No RB RB RBGrout Pump Spedel 6000 Spedel 6000 Spedel 6000 N/A Spedel 6000 N/AWater:Cement Ratio 0.45 0.45 0.45 N/A 0.45 N/AGrout Tube Dia. (mm) 19 19 19 N/A 19 N/ABreather Tube Dia. (mm) 9.5 9.5 9.5 N/A 9.5 N/ARock Type MS Acid Sed. Acid Sed/Qte MS MS MSRMR' 80 69 64 70-80Q' 13.3 15.8 15.8 13.3 13.3 13.3A 0.1 1.0 1.0 0.1 0.1 0.1B 0.3 0.2 0.3 0.2 0.2 0.2C 2.0 5.0 6.0 2.0 2.0 2.0N' 0.80 15.8 28.4 0.53 .53 0.53RQD 80 95 95 80 80 80Jn 6 6 6 6 6 6Jr 1 1 1 1 1 1Ja 1 1 1 1 1 1RQD/Jn/H.R. 2.56 1.27 1.46 2.15 2.56 2.56Blasthole Dia (mm) 114 114 114 114 114 114Backfill None None None None None None208TABLE A.5: Case Study Summary (Case 25 to 30) CASE # 25 26 27 28 29 30Surface Back HW HW HW Back BackCables/Hole 2 1 1 N/A 2 1Hole Length (m) ^9.8 HW/FW & Sq.14.6Point Anchor14.6Point AnchorN/A N/A18.3 Square C&F6.1 FanPatternDensity (bolts/m2) 0.17 0.025 0.041 N/A 0.167 0.41Equiv. Pattern (ft x ft) 7.96 20.75 16.20 N/A 8.0 5.12Stability Caved Stable Stable Caved Stable StableStrike Length (m) 71 27.7 28.6 28.4 4.6 68.1Average Width (m) 14 9 8.7 11.5 42.6 4.3Dip (degrees) 70 79 74 59 70 60(30-80)Vertical Height (m) 55-70 139 85 72 4.0 54Area (m2) 1105 3854 2402 2297 195 289Perimeter (m) 172.1 335 225 222 95 144.7Hydraulic Radius (m) 6.4 11.5 10.7 10.3 2.1 2.0Mining Method Blasthole Blasthole Blasthole Blasthole Cut and Fill BlastholeDilution (%)Cable Diameter (mm) 15.9 15.9 15.9 N/A 15.9 15.9Hole Diameter (mm) 51 51 51 N/A 57 51Cost ($/m) 28.84 28.84 28.84 N/A 23.79Bonus N/A $0.42/mSupplier Thiessen Thiessen Thiessen N/A ThiessenTiming 28d 28d 28d N/A 28d 28dPlates 102 x 102 mm N/A No 152 x 152 mmStraps No No No No No YesOther Support RB No No No RB RB/Scr/ExGrout Pump Minepro 3? Spedel 6000 Spedel 6000 N/A Moyno Spedel 6000Water:Cement Ratio 0.27-0.33? 0.45 0.45 N/A 0.5 0.4-0.45Grout Tube Dia. (mm) 19 19 19 N/A 19 19Breather Tube Dia. (mm) 9.5 9.5 9.5 N/A 12.7 N/ARock Type MS MS/Chl Sch MS/Chl Sch MS/Sx/QteRMR' 60 68 30 56Q' 13.3 15.8 15.8 10 0.9 11.6A 0.1 1.0 1.0 1.0 0.1 0.1B 0.2 0.3 0.3 0.3 0.8 0.2C 2.0 7.0 6.5 5.0 2.0 2.0N' 0.53 33.2 30.8 15.0 0.14 0.46RQD 80 95 95 90 10 79Jn 6 6 6 9 4 12Jr 1 1 1 1 1.5 2.3Ja 1 1 1 1 4.0 1.3RQD/Jn/H.R. 2.08 1.37 1.48 0.97 1.19 3.30Blasthole Dia (mm) 114 114 114 114 38 114Backfill None None None None Tailings None209TABLE A.6: Case Study Summary (Case 31 to 36) CASE # 31 32 33 34 35 36Surface HW HW Back Back HW BackCables/Hole N/A 1 1 1 N/A 1Hole Length (m) N/A 12 6.1 9.1 N/A 6.1Pattern N/A Quasi - Mandolin Fan Point Anchor N/A SquareDensity (bolts/m2) N/A 0.07 0.54 0.30 N/A 0.55Equiv. Pattern (ft x ft) N/A 12.4 4.46 5.99 N/A 4.42Stability Caved Stable Stable Unstable Stable StableStrike Length (m) 68.1 23.8 34.1 22.6 22.6 27.1Average Width (m) 4.3 3.7Dip (degrees) 60(30-80) 83 90 90 N/AVertical Height (m) 54 40.4 42.7 35.1 35.1 3.7Area (m2) 4290 295 135.4 420 825 169.8Perimeter (m) 262.2 60 80.6 82 118 92.7Hydraulic Radius (m) 16.4 4.9 1.7 5.1 7.0 1.8Mining Method Blasthole Blasthole VCR Blasthole Blasthole DriftDilution (%)Cable Diameter (mm) N/A 15.9 15.9 15.9 N/A 15.9Hole Diameter (mm) N/A 64 51 51 N/A 51Cost ($/m) N/A N/ABonus N/A N/ASupplier N/A Thiessen Thiessen Thiessen N/A ThiessenTiming N/A 28d 28d 28d N/A 28dPlates N/A 152 x 152 mm 152x 152 mm 152 x152 mm N/A 152 x 152 mmStraps Sub1 Yes No No NoOther Support Sub1 No RB/Scr Rebar/Ex RB/ExGrout Pump N/A Spedel 6000 Spedel 6000 Spedel 6000 N/A Spedel 6000Water:Cement Ratio N/A 0.4 0.45 0.4 N/A 0.4-0.45Grout Tube Dia. (mm) N/A 19 19 19 N/A 19Breather Tube Dia. (mm) N/A None 9.5 9.5 N/A N/ARock Type Qte Prdt M.Sch MS/Sx/Prdt/Msh Msch/Sch QteRMR' 55 49 60 48 64Q' 5.9 10.4 8.9 8.3 13.1 29.2A 1.0 1.0 0.1 0.1 1.0 1.0B 0.2 0.2 0.4 0.2 0.2 0.5C 5.5 5.0 2.0 2.0 8.0 2.0N' 6.5 10.4 0.71 0.33 21.0 29.2RQD 55 83 80 70 82 100Jn 15 12 9 15 15 9Jr 2.1 2.4 2 2.5 2.4 2.1Ja 1.3 1.6 2 1.4 1.0 0.8RQD/Jn/H.R. 0.22 1.41 5.24 0.92 0.78 6.17Blasthole Dia (mm) 114 114 114 114Backfill None C&F Below Cem R/F Cem R/F None210TABLE A.7: Case Study Summary (Case 37 to 421 CASE # 37 38 39 40 41 42Surface Back HW Back HW Back HWCables/Hole 1 N/A N/A N/A N/A N/AHole Length (m) 12.2 N/A N/A N/A N/A N/APattern Fan N/A N/A N/A N/A N/ADensity (bolts/m2) 0.41 N/A N/A N/A N/A N/AEquiv. Pattern (ft x ft) 5.12 N/A N/A N/A N/A N/AStability Stable Stable Stable Stable Stable UnstableStrike Length (m) 15.2 15.2 20.1 20.1 19.8 19.8Average Width (m) 6.4 6.4 3.0 3.0 3.1 3.1Dip (degrees) 54 54 70 70 59 59Vertical Height (m) 29 29 30.5 30.5 29 29Area (m2) 97.3 500 60.1 613 90 628Perimeter (m) 43.2 96 46.2 101.2 51 103Hydraulic Radius (m) 2.3 5.2 1.3 6.1 1.8 6.1Mining Method VCR VCR VCR VCR VCR VCRDilution (%)Cable Diameter (mm) 15.9 N/A N/A N/A N/A N/AHole Diameter (mm) 51 N/A N/A N/A N/A N/ACost ($/m) N/A N/A N/A N/ABonus N/A N/A N/A N/ASupplier Thiessen N/A N/A N/A N/A N/ATiming 28d+ N/A N/A N/A N/A N/APlates 152 x 152 mm N/A N/A N/A N/A N/AStraps Yes(Laced) N/A No No No NoOther Support RB/Scr/Ex RB/Scr RB/Scr/Ex RB/Scr RB/Scr/Ex RB/Scr/ExGrout Pump Spedel 6000 N/A N/A N/A N/A N/AWater:Cement Ratio 0.4-0.45 N/A N/A N/A N/A N/AGrout Tube Dia. (mm) 19 N/A N/A N/A N/A N/ABreather Tube Dia. (mm) None N/A N/A N/A N/A N/ARock Type Sumx Prdt Sumx Bio Sch Sumx SchRMR' 46Q' 12.3 7.2 15.8 21.5 15.8 21.5A 0.1 1.0 0.1 1.0 0.1 1.0B 0.2 0.2 0.2 0.3 0.2 0.2C 2.0 5.0 2.0 6.0 2.0 5.0N' 0.49 7.2 0.63 38.7 0.63 21.5RQD 70 60 90 92 90 92Jn 12 12 12 9 12 9Jr 2.1 2.3 2.1 2.1 2.1 2.1Ja 1 1.6 1 1 1 1RQD/Jn/H.R. 2.52 0.96 5.77 1.68 4.17 1.68Blasthole Dia (mm) 114 114 114 114 114 114Backfill C&F Below C&F Below None None211TABLE A.8: Case Study Summary (Case 43 to 48)CASE # 43 44 45 46 47 48Surface Back HW Back Back Back BackCables/Hole N/A N/A 1 1 1 1Hole Length (m) N/A N/A 15.8 18.3 18.3 18.3Pattern Fan N/A Square C&F Square C&F Square C&F Square C&FDensity (bolts/m2) N/A N/A 0.21 0.304 0.308 0.245Equiv. Pattern (ft x ft) N/A N/A 7.16 5.99 5.89 6.63Stability Stable Caved Stable Stable Stable StableStrike Length (m) 15.8 15.8 31 76 112 184Average Width (m) 6.7 6.7 10.1 12 11.6 10.7Dip (degrees) 56 56 33 65-70 70-80 60-70Vertical Height (m) 40.5 40.5 5.2 6.1 6.1 6.1Area (m2) 105.9 746 316 875 1295 2020Perimeter (m) 45 126 87 175 253 406Hydraulic Radius (m) 2.4 5.9 3.6 5.0 5.1 5.0Mining Method VCR VCR Cut and Fill Cut and Fill Cut and Fill Cut and FillDilution (%)Cable Diameter (mm) N/A N/A 15.9 15.9 15.9 15.9Hole Diameter (mm) N/A N/A 51 51 51 51Cost ($/m) N/A N/A 31.8 19.69 19.69 19.69Bonus N/A N/A $0.33/m $0.35/ft - $4/hr $0.35/ft - $4/hr $0.35/ft - $4/hrSupplier N/A N/A Thiessen Thiessen Thiessen ThiessenTiming N/A N/A 28d+ 28d 28d 28dPlates 152 x 152 mm N/A No Yes(#2&3 cut) Yes(#2&3 cut) Yes(#2&3 cut)Straps Yes & 4.9m Ex Yes No No No NoOther Support RB/Scr/Ex RB/SS/Scr Swellex/RB RB RB RBGrout Pump N/A N/A Minepro 3 Minepro 3 Minepro 3 Minepro 3Water:Cement Ratio N/A N/A 0.35-0.4 0.35-0.40 0.35-0.40 0.35-0.40Grout Tube Dia. (mm) N/A N/A 19 19 19 19Breather Tube Dia. (mm) N/A N/A 9.5 9.5 Hi Press 9.5 Hi Press 9.5 Hi PressRock Type Sumx/Prdt Prdt SS DS/SS DS/SS SSRMR' 46 79 55 55 84Q' 11.1 7.2 26.1 5.4 5.4 25A 0.1 1.0 0.1 0.1 0.1 0.4B 0.2 0.2 0.3 0.2 0.2 0.2C 2.0 5.0 2.0 2.0 2.0 2.0N' 0.44 7.2 1.6 0.22 0.22 4.0RQD 75 60 88 65 65 100Jn 12 12 9 6 6 6Jr 2.3 2.3 2.0 2.0 2.0 1.5Ja 1.3 1.6 0.75 4.0 4.0 1.0RQD/Jn/H.R. 2.60 0.85 2.72 2.16 2.12 3.34Blasthole Dia (mm) 114 114 38 38 38 38Backfill C&F Below C&F Below Tailings Sand/Hyd Sand/Hyd Sand/Hyd212TABLE A.9: Case Study Summary (Case 49 to 54)CASE # 49 50 51 52 53 54Surface HW HW HW Back Back HWCables/Hole 1 1 N/A 1 1 1Hole Length (m) 9.1 9.1 N/A 10.2 10.7 9.1Pattern Point Anchor Point Anchor N/A Square Square EvenDensity (bolts/m2) 0.16 0.18 N/A 0.11 0.33 0.13Equiv. Pattern (ft x ft) 8.20 7.73 N/A 9.89 5.71 9.10Stability Stable Caved Stable Stable Stable StableStrike Length (m) 130 111 76 76 70 70Average Width (m) 11.5 11.0 7.8 12 17.2 17.2Dip (degrees) 62 62 60 60 60 60Vertical Height (m) 41 57.5 25 25 17.5 17.5Area (m2) 6111 7565 2236 936 654 1378Perimeter (m) 395 445 215 180 172 175Hydraulic Radius (m) 15.5 17.0 10.4 5.2 3.8 7.9Mining Method Blasthole Blasthole Blasthole Blasthole Blasthole BlastholeDilution (%)Cable Diameter (mm) 15.9 15.9 N/A 15.9 15.9 15.9Hole Diameter (mm) 51 51 N/A 51 51 51Cost ($/m) 19.69 19.69 N/A 19.69 19.69 19.69Bonus $0.34/ft-$4/hr $0.34/ft-$4/hr N/A $0.34/ft-$4/hr $0.34/ft-$4/hr $0.34/ft-$4/hrSupplier Thiessen Thiessen N/A Thiessen Thiessen ThiessenTiming 28d 28d N/A 28d 28d 28dPlates No No N/A No Yes NoStraps No No No No No NoOther Support RB/Resin RB/Resin RB NoGrout Pump Minepro 3 Minepro 3 N/A Minepro 3 Minepro 3 Minepro 3Water:Cement Ratio 0.35-0.40 0.35-0.40 N/A 0.35-0.40 0.35-0.40 0.35-0.40Grout Tube Dia. (mm) 19 19 N/A 19 19 19Breather Tube Dia. (mm) 9.5 (hi press) 9.5 (hi press) N/A 9.5 (hi press) 9.5 (hi press) 9.5 (hi press)Rock Type SCQP SCQP SCQP SS(80%)/DS DS SCQPRMR' 62 55 59Q' 9.9 3.1 8.3 15.5 5.4 3.1A 1.0 1.0 1.0 0.1 0.1 1.0B 0.3 0.3 0.3 0.2 0.2 0.3C 4.7 4.7 5.0 2.0 2.0 5.0N' 14.0 4.4 12.5 0.62 0.22 4.7RQD 92 75 75 93 65 75Jn 7 9 6 6 6 9Jr 1.5 1.5 2 1.6 2.0 1.5Ja 2 4 3 1.6 4 4RQD/Jn/H.R. 0.85 0.49 1.20 2.98 2.84 1.05Blasthole Dia (mm) 76 76 51 51 76 76Backfill No No Waste/Hyd Waste/Hyd No No213TABLE A.10: Case Study Summary (Case 55 to 59)CASE # 55 56 57 58 59Surface Back HW Back I-1W BackCables/Hole 1 1 1 N/A 2Hole Length (m) 5.0 7.6-9.1 6.4 N/A 4.9Pattern Point Anchor Even Square N/A Square R&PDensity (bolts/m2) 0.35 0.12 0.34 N/A 0.26Equiv. Pattern (ft x ft) 5.55 9.47 5.71 N/A 6.43Stability Stable Stable Stable Unstable StableStrike Length (m) 141 141 25 25 144Average Width (m) 13.7 13.7 18.8 18.8 12.2Dip (degrees) 60 60 70 70 20Vertical Height (m) 15.7 15.7 15 15 5.2Area (m2) 1862 2439 473 400 1757Perimeter (m) 302 317 88 82 312Hydraulic Radius (m) 6.2 10.9 5.4 4.9 5.6Mining Method Blasthole Blasthole Blasthole Blasthole Room & PillarDilution (%)Cable Diameter (mm) 15.9 15.9 15.9 N/A 15.9Hole Diameter (mm) 51 51 51 N/A 64Cost ($/m) 19.69 19.69 19.69 N/ABonus $0.34/ft-$4/hr $0.34/ft-$4/hr $0.34/ft-$4/hr N/A $133/shiftSupplier Thiessen Thiessen Thiessen N/A ThiessenTiming 28d 28d 28d N/A 28dPlates No No No No NoStraps No No No No NoOther Support RB No RB RB SwellexGrout Pump Minepro 3 Minepro 3 Minepro 3 N/A Minepro 3Water:Cement Ratio 0.35-0.40 0.35-0.40 0.35-0.40 N/A 0.35-0.40Grout Tube Dia. (mm) 19 19 19 N/A 19 recoveredBreather Tube Dia. (mm) 9.5 (hi press) 9.5 (hi press) 9.5 (hi press) N/A NoneRock Type SS SCQP DS SCQP SulphideRMR' 84 59 55 59 69Q' 25 3.1 5.4 3.1 25A 0.2 1.0 0.1 1.0 0,4B 0.2 0.3 0.2 0.3 0.3C 2.0 5.0 2.0 6.0 2.2N' 2.0 4.7 0.22 5.6 6.6RQD 100 75 65 75 90Jn 6 9 6 9 6-9Jr 1.5 1.5 2.0 1.5 1.5-2.1Ja 1.0 4.0 4.0 4.0 0.75-1RQD/Jn/H.R. 2.69 0.76 2.00 1.70 2.30 avg.Blasthole Dia (mm) 76 76 51 51 38Backfill No No No No None214APPENDIX B215BOLT DENSITY CONVERSION CHARTTABLE B.1: Bolt Density Conversion ChartBOLTS/SQ. METER BOLTS/SQ. FOOTSQUARE PATTERNEQUIVALENT(m a m)SQUARE PATTERNEQUIVALENT(ft a ft): OLTS/SQ. METER BOLTS/SQ. FOOTSQUARE PATTERNEQUIVALENT(m a m)SQUARE PATTERNEQUIVALENT(ft a ft)0.000 0.0000 0.00 0.00 0.275 0.0255 1.91 6.260.005 0.0005 14.14 46.40 0.280 0.0260 1.89 6.200.010 0.0009 10.00 32.81 0.285 0.0265 1.87 6.150.015 0.0014 8.16 26.79 0.290 0.0269 1.86 6.090.020 0.0019 7.07 23.20 0.295 0.0274 1.84 6.040.025 0.0023 6.32 20.75 0.300 0.0279 1.83 5.990.030 0.0028 5.77 18.94 0.305 0.0283 1.81 5.940.035 0.0033 5.35 17.54 0.310 0.0288 1.80 5.890.040 0.0037 5.00 16.41 0.315 0.0293 1.78 5.850.045 0.0042 4.71 15.47 0.320 0.0297 1.77 5.800.050 0.0046 4.47 14.67 0.325 0.0302 1.75 5.760.055 0.0051 4.26 13.99 0.330 0.0307 1.74 5.710.060 0.0056 4.08 13.39 0.335 0.0311 1.73 5.670.065 0.0060 3.92 12.87 0.340 0.0316 1.71 5.630.070 0.0065 3.78 12.40 0.345 0.0320 1.70 5.590.075 0.0070 3.65 11.98 0.350 1.69 5.550.080 0.0074 3.54 11.60 0.355 1.68 5.510.085 0.0079 3.43 11.25 0.360 1.67 5.470.090 0.0084 3.33 1.66 5.430.095 0.0088 3.24 1.64 5.390.100 0.0093 3.16 1.63 5.360.0098 3.09 1.62 5.320.0102 3.02 1.61 5.29• 0.0107 2.95 1.60 5.250.0111 2.89 1.59 5.22• 0.0116 2.83 1.58 5.190.0121 2.77 1.57 5.16• 0.0125 2.72 1.56 5.120.0130 2.67 1.55 5.09• 0.0135 2.63 1.54 5.062.58 0.425 1.53 5.032.54 0.430 1.52 5.002.50 • 0.435 1.52 4.972.46 0.440 1.51 4.952.43 0.445 1.50 4.922.39 0.450 1.49 4.892.36 0.455 1.48 4.862.32 0.460 1.47 4.842.29 0.465 1.47 4.812.26 0.470 1.46 4.792.24 0.475 1.45 4.762.21 0.480 1.44 4.742.18 0.485 1.44 4.712.16 0.490 1.43 4.692.13 0.495 1.42 4.662.11 0.500 1.41 4.642.09 0.505 1.41 4.622.06 0.510 1.40 4.592.04 • 0.515 1.39 4.570.245 2.02 0.520 1.39 4.550.250 0.0232 2.00 6.56 0.525 1.38 4.530.255 0.0237 1.98 6.50 0.530 1.37 4.510.260 0.0242 1.96 6.43 0.535 1.37 4.490.265 0.0246 1.94 6.37 0.540 1.36 4.460.270 0.0251 1.92 6.31 0.545 0.0506 1.35 4.44216Table B.1: Bolt Density Conversion Chart (con't) BOLTS/SQ. METER BOLTS/SQ. FOOTSQUARE PATTERNEQUIVALENT(m a m)SQUARE PATTERNEQUIVALENT(ft a ft)BOLTS/SQ. METER BOLTS/SQ. FOOTSQUARE PATTERNEQUIVALENT(m a m)SQUARE PATTERNEQUIVALENT(ft a ft)0.550 0.0511 0.00 0.00 0.825 0.0766 1.10 3.610.555 0.0516 1.34 4.40 0.830 0.0771 1.10 3.600.560 0.0520 1.34 4.38 0.835 0.0776 1.09 3.590.565 0.0525 1.33 4.36 0.840 0.0780 1.09 3.580.570 0.0529 1.32 4.35 0.845 0.0785 1.09 3.570.575 0.0534 1.32 4.33 0.850 0.0790 1.08 3.560.580 0.0539 1.31 4.31 0.855 0.0794 1.08 3.550.585 0.0543 1.31 4.29 0.860 0.0799 1.08 3.540.590 0.0548 1.30 4.27 0.865 0.0804 1.08 3.530.595 0.0553 1.30 4.25 0.870 0.0808 1.07 3.520.600 0.0557 1.29 4.24 0.875 0.0813 1.07 3.510.605 0.0562 1.29 4.22 0.880 0.0817 1.07 3.500.610 0.0567 1.28 4.20 0.885 0.0822 1.06 3.490.615 0.0571 1.28 4.18 0.890 0.0827 1.06 3.480.620 0.0576 1.27 4.17 0.895 0.0831 1.06 3.47• 0.0581 1.26 4.15 0.900 0.0836 1.05 3.460.0585 1.26 4.13 0.905 0.0841 1.05 3.45• 0.0590 1.25 4.12 0.910 0.0845 1.05 3.440.0595 1.25 4.10 0.915 0.0850 1.05 3.43• 0.0599 1.25 4.09 0.920 0.0855 1.04 3.420.0604 1.24 4.07 0.925 0.0859 1.04 3.41• 0.0608 1.24 4.05 0.930 0.0864 1.04 3.401.23 4.04 0.935 0.0869 1.03 3.391.23 4.02 0.940 0.0873 1.03 3.381.22 • 0.945 0.0878 1.03 3.381.22 0.950 0.0882 1.03 3.371.21 0.955 0.0887 1.02 3.361.21 0.960 0.0892 1.02 3.351.20 0.965 0.0896 1.02 3.341.20 0.970 0.0901 1.02 3.331.20 0.975 0.0906 1.01 3.321.19 0.980 0.0910 1.01 3.311.19 0.985 0.0915 1.01 3.311.18 0.990 0.0920 1.01 3.301.18 0.995 0.0924 1.00 3.291.17 1.000 0.0929 1.00 3.281.17 1.005 0.0934 1.00 3.271.17 1.010 0.0938 1.00 3.261.16 1.015 0.0943 0.99 3.261.16 1.020 0.0948 0.99 3.251.15 3.79 1.025 0.0952 0.99 3.241.15 3.78 1.030 0.0957 0.99 3.231.15 3.76 1.035 0.0961 0.98 3.231.14 3.75 1.040 0.0966 0.98 3.221.14 3.74 1.045 0.0971 0.98 3.211.14 3.73 1.050 0.0975 0.98 3.200.780 0.0725 1.13 3.72 1.055 0.0980 0.97 3.190.785 0.0729 1.13 3.70 1.060 0.0985 0.97 3.190.790 0.0734 1.13 3.69 1.065 0.0989 0.97 3.180.795 0.0739 1.12 3.68 1.070 0.0994 0.97 3.170.800 0.0743 1.12 3.67 1.075 0.0999 0.96 3.160.805 0.0748 1.11 3.66 1.080 0.1003 0.96 3.160.810 0.0752 1.11 3.65 1.085 0.1008 0.96 3.150.815 0.0757 1.11 3.63 1.090 0.1013 0.96 3.140.820 0.0762 1.10 3.62 1.095 0.1017 0.96 3.14217APPENDIX C218SUMMARY OF INSTALLATION PROCEDURES219CABLE BOLT INSTALLATION PROCEDURESMINE #1Spedel 6000 Grout PumpDouble cables2 man crewDownholes: Use 3/4" loading hose to clean all holes prior to inserting cable bolts.Tape 3/4" grout tube 6" from one end of the cable using enough tube to run the whole lengthof the hole and attach to the grout pump.Lower cables into holes to be grouted with taped end of grout tube at toe of hole.Prepare .35 w:c grout and hook up pump to grout tube.Pump grout until it begins to come out of the collar.Upholes: Tape 3/8" breather tube to end of cable and insert into hole with taped end of breather tube atthe toe of hole.Insert grout tube 2' into collar of hole and leave enough to reach to the pump.Plug hole collar with rags.Prepare 0.35 w:c grout and hook up pump to grout tube.Pump grout until it comes out of the breather tube.Bend and tie off breather and grout tubes to prevent grout from leaking.MINE #2Minepro 3 Pneumatic Grout Pump.213'/ms (65m/ms) inserted (Double cables, 30-49' bolts (9-15m) ,countersink 0-16' (0-5m))591 '/ms (180m/ms) grouted (w:c=0.32, 9-19m)2 man crewUpholes > 40' (12m): Install spring steel on end holding device.Tape 1/2" breather tube to end of cable with end of breather tube cut at 45 degrees.Insert cable in hole and ensure that breather tube is not pinched.Connect water hose to breather tube and flush hole.Place 3/4" grout tube lm into hole collar with end cut at 45 degrees.Plug hole with MONOFOAM and allow to cure 24 hours.Fill hole with 0.375 w:c groutEnsure breather tube is completely full of grout.220Upholes < 40' (12m): Install spring steel on end holding device.Tape 3/4" grout tube to end of cable.Insert cable into hole.Connect water hose to breather tube and flush hole.Fill hole with 0.32 w:c groutNo plug to be used at collarMINE #3Spedel 6000 Pump & B3100 MixerGrout & Install in 45 minutes.Double cables @ 33' (10m)2 man crewUpholes: Wash down cable to be installed.Install spring steel on end holding device.Attach 3/8" breather tube 6" (15cm) from end of cable with electrical tape.Wrap with tape at several locations along the cable.Tape 3/4" grout tube 6.5' (2m) from the hole collar.Insert cable into the hole.Seal hole collar with burlap and wooden wedges.Cut breather & grout tubes 10' (3m) from the collar.Attach grout tube to grout pump feeder hose and pump 0.4 w:c grout until it discharges fromthe breather tube.MINE #4Moyno 3L3 Pump with Chemgrout CG-550 mini grout plant.600'/ms installed and grouted (60' upholes)2 man crewUpholes: Bend back one wire if hole < 40' or 2 wires if > 40'Use cable pusher if greater than 40'Install breather and grout tube combinationPlug holes with shredded cloth.Mix grout like pudding so will squeeze through fingers.MINE #5Spedel 6000 grout pump.8 40' holes grouted in three hours - w:c 0.35-0.4Single cables2 man crewUpholes: Install spring steel on end holding device.Tape 3/4" grout tube to top end of cable with 2' taped segment.Insert cable into holeCut grout tube to allow end to reach grout pumpMix 1 bag (40kg) cement with 12 litres water. (w:c = 0.3)Pump grout until expelled at hole collar.Plug collar with steel wool (4 rolls)Pump grout until pump stalls.Crimp and tape grout tube at collar.MINE #6Spedel 6000 grout pumpMinpro 3 Host Hydraulic purchased June 91.Procedure same as Mine #5MINE #7Pneumatic skid mounted Minpro 3 grout pump.Single cables2 man crewUpholes: Bend wire at top of cable 135 degrees.Install cable in hole.Push in grout tube to toe of hole and pull back 6".Tape grout tube to cable sticking out of collar.Cut off grout tube with enough to reach the grout pump hose.Install wedges in hole collar.Pump grout and bend end of grout tube and tie off.221222MINE #9Minpro 3 Grout pump2 man crewUpholes > 45': Install spring steel on end holding device.Tape 3/8" (450 psi) breather tube to the cable in two or three places.Keep the breather tube end within 6" of the end holding device and cut the end at 45 degrees.Install the cable in the hole.Push a 3/4" (250 psi) grout tube 1' to 3' into the hole and leave a 3' tail out of the hole.Wedge the cable in the hole and plug the collar with cotton waste and thick grout.Let the collar seals set for eight hours.Mix grout between 0.33 and 0.40 w:c (90 litres water and pump until grout flows out of thebreather tube.Bend end of the grout and breather tube 180 degrees and tie in this position.Upholes < 45': Install spring steel on end holding device.Tape a 3/4" (100 psi) grout tube to the cable in two or three places.Keep the grout tube end within 6" of the end holding device.Insert the cable in the hole.Wedge the cable at the hole collar with a wooden wedge.Do not pinch the grout tube.Mix grout to a 0.30 w:c ratio and pump until it squeezes out of the hole collar.Kink and tie off the grout tube.Grout w:c ratios > 0.35 will require a grout plug so it is important to maintain a 0.3 w:c ratio.Downholes: Tape a 3/4" grout tube (100 psi) within 6" of the cable end.Insert the cable into the hole and push to the end of the hole.Mix a 0.3 w:c ratio grout and pump until the grout appears at the collar.Do not extract the grout tube while pumping as tests have shown that air gaps may result in thegrout column223MINE #10Pneumatic skid mounted Minpro 3 grout pump.Double cables, 16'(4.9m)2 man crewBend 9" length of two cable wires back 135 degrees.Install cables in hole with anchor end towards toe of hole.Push 3/4" grout tube to toe of hole.Pump 0.32 w:c ratio grout and retrieve grout tube as grout is pumped into hole.Grout tube should gently push back as grout is pumped and a slight hand pressure is required tohold the tube in place.When hole is full push a pilgrim hat plug into the collar.MINE #11Tamrock Cabolt Machine3/4" grout tube is fed into the hole from a reel on the Cabolt rig.Grout is mixed to a 0.3 w:c ratio and pumped into the hole.Cable is fed into the grouted hole from a reel underneath the bolting rig and cut with a hydrauliccutting device.The cable is kinked prior to fmal installation in order to aid in anchoring the cable.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0081080/manifest

Comment

Related Items