Empirical approaches for opening design in weak rock masses

A major focus of ground control research presentlybeing conducted by the Spokane Research Laboratoryof the National Institute for Occupational Safety andHealth (NIOSH) is to incorporate weak rock masses(such as are associated with operations in the CarlinTrend in Nevada) into existing design relationshipsThe original database that led to most of the empiricaldesign relationships presently employed in hard-rockmining was derived from fair-to-good-quality rock Inthis study the relationship between weak rock qualityand opening design (non-entryentry methods) is beinginvestigated The common factor in all mines is a weakback or wall This work attempts to provide tools thatwill enable a mine operator to make economic decisionsthat will also ensure a safe working environment

Tom Brady (for correspondence ndash E-mail thb6cdcgov)andLewis Martin are at the Spokane Research Laboratory NationalInstitute for Occupational Safety and Health 315 EastMontgomery Avenue Spokane WA 99207-2291 USA (E-mailljm8cdcgov) Rimas Pakalnis is at the Mining DepartmentUniversity of British Columbia Vancouver British ColumbiaCanada (E-mail rcpminingubcca)

copy 2005 Institute of Materials Minerals and Mining andAustralasian Institute of Mining and Metallurgy Publishedby Maney on behalf of the Institutes Manuscript accepted infinal form 24 February 2005

Keywords Weak rock masses empirical design openingdesign

INTRODUCTIONResearchers at the Spokane Research Laboratory ofthe National Institute for Occupational Safety andHealth Spokane Washington USA and theUniversity of British Columbia Vancouver Canadahave assembled a team to develop underground designguidelines for safe and cost-effective mining within aweak rock mass Such work also includes developingnovel support methods such as the use of syntheticfibre reinforced shotcrete ways to undermine under-hand-and-fill backfilled stopes and assess supportspresently in place in weak rock masses In the presentstudy rock mass interaction with grouted bolt supportswas investigated in three mines in Nevada and backfillpillar and bolt support were studied in one

Many Nevada gold deposits are found in intenselyfractured faulted and argillised host rock As a result

underground mining is often difficult and hazardousas indicated by the number of injuries and fatalitiesfrom uncontrolled falls of ground (Table 1)

A comparative analysis by the Mine Safety andHealth Administration (MSHA)9 for the years1990ndash1999 indicated that the number of roof-fallinjuries in Nevada has varied from a low of 8 in 1991to a high of 28 in 1995 and 1997 As late as 1999 thenumber of injuries was still in double figures (Fig 1)Analysis of the MSHA data shows 76middot7 of the rooffalls were from the back 18middot6 were rock falls fromface or ribs with the remaining 4middot7 of an unknownnature The mines are required to report all injuries toMSHA by law

Mining is a dynamic process and ground conditionscan change over a short distance A mine opening mustperform in a predictable manner over its expected life

Mining Technology (Trans Inst Min Metall A) March 2005 Vol 114 A13DOI 101179037178405X44494

Empirical approaches for opening design in weak rock masses

Tom Brady Lewis Martin and Rimas Pakalnis

1 Injuries in Nevada 1990ndash2004

Table 1 Ground control injuries and fatalities in undergroundground gold mines in Nevada 1985ndash2000

Fatalities 7Permanent disabilities 4Lost-time injuries 49Restricted-activity injuries 46Other injuries 110

Total 216

Reported rock falls with no injuries 69

Includes MSHA data for non-injury incidents where areportable fall of ground occurred but did not cause injuriesbecause the mining area was unoccupied

Empirical design methods have been used successfullyover the past 30 years largely because they permit theoverall behaviour of a rock mass to be predicted easilyand accurately The basis for the success of theempirical method is a strong foundation of field datacoupled with on-going field observations that allowchanging rock conditions to be evaluated as miningprogresses Two systems were initially developed fromcase studies and databases originally derived fromcivil engineering applications and augmented by minestudies ndash the rock mass rating (RMR) system and theQ system As stated by the author lsquothe Q system isspecifically the permanent lining estimation systemfor tunnels and caverns in rock and mainly for civilengineering projectsrsquo1

There are several advantages for utilising the RMRsystem over the Q system The RMR system has arelatively straight-forward scoring system for eachparameter on a 0ndash10 scale and consequently is easierto learn1011

One of the original extensions of the rock mass ratingsystem from civil into mining was conducted under aUS Bureau of Mines Spokane Research Laboratoryresearch grant which developed empirical methods forblock caving operations for US copper mines7

It is important to remember that any method ofdesigning an opening must be easy to assessunderstand apply modify if necessary and reproducefor the next application if accepted as an on-goingoperational tool for design A critical factor is that thedesign incorporates the degree of stability required forany mine entry

SPAN DESIGN MAN-ENTRY METHODSThe lsquocritical span curversquo was developed in 1994 toevaluate back stability in cut-and-fill mines13 In 2000the span curve database of 172 observationsdeveloped by the University of British Columbia wasexpanded to include a total of 292 case histories frommines primarily in Canada21 The information fromthese case histories provides the basis for the spandesign curve shown in Figure 2

A14 Mining Technology (Trans Inst Min Metall A) March 2005 Vol 114

Brady Martin Pakalnis Empirical approaches for opening design in weak rock masses

2 Design span curve

Table 2 Nevada mines database back spans in weak rock

RMR Span () (m) Condition Other

Mine 145 5middot5 S Stable with support45 9middot0 S Stable with support40 6middot0 S Stable with support

Mine 240 4middot0 S Stable with support45 4middot3 U Caved with support30 3middot7 S Stable with support

Mine 340 7middot0 S Stable with support45 2middot1 S Stable with support26 2middot1 S Stable with support25 4middot6 U Caved with support55 7middot6 S Stable with support45 3middot0 S Stable with support

Mine 470 4middot6 S Stable with support40 4middot6 S Stable with support25 4middot6 S Stable with support55 5middot5 S Stable with support30 6middot1 S Caved upon longhole30 6middot1 S Caved upon longhole45 4middot6 S Stable with support50 6middot1 S Stable with support70 11middot3 S Stable with support25 7middot3 S Stable with support30 3middot0 U Prior to support placement30 1middot8 S Prior to support placement50 6middot1 S Stable with support55 7middot6 S Stable with support55 6middot1 S Stable with support

Mine 530 3middot0 S Stable with support30 4middot3 U Caved with support20 5middot8 U Caved with support15 3middot7 S Stable with support


Mine 740 4middot6 S Stable with support35 4middot6 S Stable with support

Mine 825 5middot0 U Caved had to spile

20 1middot2 S No support maximum round possible



55 3middot7 S No support typical round no problems

35 4middot6 S No support typical round no spileshotcrete

20 7middot6 U Caved had to spile45 6middot0 S Stable with split-sets only

A lsquocritical spanrsquo is defined as the diameter of thelargest circle that can be drawn within the boundariesof the exposed back The stability of this exposed spanis related to the type of rock in the immediate backThe lsquodesign spanrsquo refers to backs that have no supportandor spans that are supported with localised patternbolting (1middot8-m long mechanical bolts on 1middot2 times 1middot2-mspacings) Local support is deemed as support used toconfine blocks that may be loose or that might open orfall because of subsequent mining in surrounding areasExcavation stability is classified into three categories

(i) Stable excavation (a) no uncontrolled falls ofground (b) no movement of the back is observedand (c) no extraordinary support measures havebeen employed

(ii) Potentially unstable excavation (a) extra groundsupport has been installed to prevent potentialfalls of ground (b) movement has occurred inthe back and (c) increased frequency of groundmovement has been observed

(iii) Unstable excavation (a) area has collapsed (b)depth of failure of the back is 0middot5 times the span (inthe absence of major structures) and (c) supportwas not effective in maintaining stability

A ndash10 correction factor is applied to the final RMR76

value when evaluating rock with shallow-dipping or flatjoints However the applicability of this factor is beingre-assessed for weak ground because of its amorphousnature and because joint direction is expected to play a

minor role Where discrete ground wedges have beenobserved and identified they must be supported prior toemploying the critical span curve

Stability is generally defined in terms of short-termstability because the database is based largely onstoping methods that by their nature are of shortduration Movement of the back greater than 1 mmwithin a 24-h period has been defined as a criticalamount of movement for safe access18

Some 44 case histories from five different mineswith RMR76 values varying between 20 and 85 wereadded to the information base for the critical spancurve (Table 2) Several values were less than 55RMR76 the lowest RMR76 value calculated for anylocation was 25 This information was used toaugment the original lsquospan design curve for man-entrymining13 as shown in Figure 3A The span curveenables an operator to assess back stability withrespect to the rock mass The information has beenused successfully to predict the stability of weak rockmasses and has provided operators with an additionaldesign tool for making decisions concerning thestability of mine openings The data are being coupledwith depth of failure to define the amount of supportrequired to arrive at a safe cost-effective man-entrydesign A brief description of the use of the criticalspan curve is presented however the reader is referredto the detailed reference as outlined by Pakalnis18

STABILITY GRAPH METHOD NON-ENTRY MININGThe original stability method for open stope design wasbased largely on Canadian operations from 55 casehistories which included 48 back studies and 7 wall datapoints15 The original database of case historiesexhibited a rock mass rating (RMR76) in excess of 50or Q values of 2middot0 or greater4 The method wasextended and modified by Potvin19 based upon datafrom 34 mines with 175 open stope case histories and67 cases of supported stopes Nickson17 expanded theexisting database of supported stopes by Potvin bycollecting 46 cases histories while visiting 13 mines In

Mining Technology (Trans Inst Min Metall A) March 2005 Vol 114 A15


3 Distribution of RMR (A) Span database13 (B)stability graph database (ELOS)6

4 Augmented span curve Numbers in key correspondto mine numbers in Table 2 Letters indicatelocation on the span curve

all instances stability was qualitatively assessed aseither being stable potentially unstable or cavedResearch by Mah14 and Clark and Pakalnis6 at theUniversity of British Columbia augmented thestability graph by using stope surveys in which cavitymonitoring systems were employed Mahrsquos workadded 96 data points onto Matthewsrsquos stability graph

under mining conditions whereas Clark5 added anadditional 88 data points This research has enabledquantification of the amount of wall slough Aparameter defined by Clark5 as the equivalent linearoverbreakslough (ELOS Fig 5) was used to expressvolumetric measurements of overbreak as an averagedepth over the entire stope surface ELOS is defined as



Table 3 Nevada mines database stope spans and walls in weak rock (all values of A = 1)

RMR Dimensions () height times length (m) Dip B C N HR (m) ELOS (m) Comments

Mine 145 20 times 17 90 0middot3 8 2middot7 4middot6 lt1middot0 lt 1 m of ELOS40 20 times 16 90 0middot3 8 1middot5 4middot4 2middot055 49 times 18 90 0middot3 8 8middot1 6middot6 1middot039 34 times 34 90 0middot3 8 1middot4 8middot5 4middot625 90 0middot3 8 0middot3 1middot8 lt 1middot0 lt 1 m of ELOS stable (estimated)34 90 0middot3 8 0middot3 3middot4 lt 1middot0 lt 1 m of ELOS stable (estimated)42 90 0middot3 8 0middot3 2middot8 lt 1middot0 lt 1 m of ELOS stable (estimated)

Mine 240 11 times 21 90 0middot3 8 1middot5 3middot4 lt 1middot0 lt 1 m of ELOS50 11 times 21 90 0middot3 8 4middot7 3middot4 lt 0middot5 lt 0middot5 m of ELOS

Mine 355 18 times 18 70 0middot2 5middot9 4middot0 4middot6 0middot626 12 times 18 70 0middot3 5middot9 0middot2 3middot7 gt 2middot0 gt 2 m of ELOS

Mine 425 6 times 29 55 0middot2 4middot5 0middot2 2middot5 0middot3 Cluster average heightwidth25 8 times 36 90 0middot3 8 0middot5 2middot4 0middot1 Cluster average rib25 17 times 12 90 0middot3 8 0middot5 3middot5 0middot1 Cluster average rib25 21 times 15 90 0middot3 8 0middot5 4middot5 0middot1 Cluster average rib55 30 times 26 90 0middot3 8 7middot2 7middot0 0middot1 Cluster average rib45 6 times 25 90 0middot3 8 2middot4 2middot5 0middot1 Cluster average heightwidth45 18 times 12 90 0middot3 8 2middot4 3middot5 0middot1 Cluster average rib45 19 times 16 90 0middot3 8 2middot4 4middot4 0middot1 Cluster average rib45 6 times 22 90 0middot3 8 2middot4 2middot4 0middot5 Moderate25 16 times 5 0middot5 3middot2 0middot5 Moderate25 6 times 26 0middot5 2middot5 0middot5 Moderate45 6 times 22 0middot3 8 2middot4 2middot4 0middot5 Moderate25 16 times 27 90 0middot3 8 0middot5 2middot5 0middot5 Moderate25 6 times 20 90 0middot3 8 0middot5 2middot3 0middot6 Moderate45 6 times 20 90 0middot3 8 2middot4 2middot3 0middot6 Moderate25 6 times 20 90 0middot3 8 0middot5 2middot3 0middot6 Moderate25 6 times 24 90 0middot3 8 0middot5 2middot4 0middot9 Moderate35 6 times 25 90 0middot2 4middot8 2middot4 2middot4 1middot0 Moderate25 15 times 13 90 0middot3 8 0middot5 3middot5 gt 2middot0 Caved visually estimated lt 2 m25 20 times 15 90 0middot3 8 0middot5 44 gt 2middot0 Caved visually estimated lt 2 m25 21 times 15 90 0middot3 8 1middot0 4middot4 gt 2middot0 Caved visually estimated lt 2 m25 15 times 13 90 0middot3 8 0middot5 3middot5 gt 2middot0 Caved visually estimated lt 2 m25 20 times 15 90 0middot3 8 0middot5 4middot4 gt 2middot0 Caved visually estimated lt 2 m25 21 times 15 90 0middot3 8 0middot5 3middot8 gt 2middot0 Caved visually estimated lt 2 m25 21 times 10 90 0middot3 8 0middot5 3middot5 1middot5 Caved visually estimated lt 2 m25 22 times 14 90 0middot3 8 0middot5 4middot4 1middot5 Caved visually estimated lt 2 m25 23 times 12 90 0middot3 8 0middot5 4middot4 1middot5 Caved visually estimated lt 2 m25 21 times 10 90 0middot3 8 0middot5 3middot5 1middot5 Failed visually estimated 1ndash2 m25 22 times 14 90 0middot3 8 0middot5 4middot3 1middot5 Failed visually estimated 1ndash2 m25 23 times 12 90 0middot3 8 0middot5 3middot8 1middot5 Failed visually estimated 1ndash2 m25 19 times 13 90 0middot3 8 0middot5 3middot 1middot5 Failed visually estimated 1ndash2 m25 19 times 13 0middot3 8 0middot5 3middot8 1middot5 Failed visually estimated 1ndash2 m25 22 times 15 0middot3 8 0middot5 4middot4 1middot5 Failed visually estimated 1ndash2 m25 19 times 13 0middot3 8 0middot5 3middot9 1middot5 Failed visually estimated 1ndash2 m25 19 times 13 0middot3 8 0middot5 3middot8 1middot5 Failed visually estimated 1ndash2 m25 22 times 15 0middot3 8 0middot5 4middot4 1middot5 Failed visually estimated 1ndash2 m

Mine 645 2middot6 4middot4 lt 0middot5 Typical stope45 2middot6 6middot2 1middot8 Caved stope

the volume of slough from the stope surface dividedby the product of stope height times wall strike lengthknown as the hydraulic radius (HR)

ELOSHR

Volume of Slough= (1)

The stability graph relates hydraulic radius (HR) ofthe stope wall to empirical estimates of overbreakslough Additionally the database for the Matthewsrsquomethod has been augmented by the addition of 400case histories which includes open stoping experiencesfor a broad range of rock mass conditions in Australiaspecifically at the Mount Charlotte mine Theadditional case histories include much larger stopesand extend the modified stability graph to a hydraulicradius of 55 m as compared to the previous maximumvalue of 20 m16 The drawback of this study was that itwas largely qualitative because surveys were notavailable to measure the wall slough

A limited number of observations existed forRMR76 values under 45 (Fig 3B) An additional 45data points were added on the stability graphndashnon-entry from Nevada operations having an RMR76

under 45 (Fig 3B and Table 3) In addition mine 4reflects over 338 observations that have been averagedto reflect the design points shown in Table 2 Thestability graph relates hydraulic radius of the stopewall to empirical estimates of overbreak sloughHydraulic radius (HR) is defined as the surface areaof an opening divided by the perimeter of the exposedwall analysed as shown in Figure 5

Nprime = Qprime times A times B times C (2)

where Nprime is the modified stability number Qprime is theNorwegian Geological Institute (NGI) rock qualityindex (after Barton et al2) with stress reductionfactor (SRF) and Jw = 1middot0) A is high stress reductionB is orientation of discontinuities and C is theorientation of surface

The stress reduction factor and joint waterreduction factor are equal to 1 as they are accountedfor separately within the analysis For the ELOS graphpoints the database for the stability graph was derivedfrom mining operations that are generally dry

The following relationship was used to convertRMR to Qprime (from Bieniawski3)

RMR = 9LnQprime + 44 (3)

The A factor accounts for the influence of highstresses that reduce rock mass stability and isdetermined by the ratio of unconfined compressivestrength of intact rock to maximum induced stressparallel to the opening surface It is set to 1middot0 if intactrock strength is 10 or more times induced stress whichindicates that high stress is not a problem It is set to0middot1 if rock strength is two times induced stress or lesswhich indicates that high stress significantly reducesopening stability In the mines visited for this studythe value of A was equal to 1middot0 because the hangingwall was largely in a relaxed state

The B factor looks at the influence of the orientationof discontinuities with respect to the surface analysed andstates that joints oriented 90deg to a surface do not createstability problems and the B factor is assessed a value of1middot0 Discontinuities dipping up to 20deg to the surface arethe least stable and represent geological structures thatcan topple In this case the B factor is equal to 0middot2 whichwas the value used for the Nevada database In extremelyweak rock masses (RMR76 = 25) the material largelyresembles an amorphous mass with geological structuresthroughout therefore reduction due to jointing issuspect The authors are presently analysing the data toaugment this factor

The C factor considers orientation of the surfacebeing analysed and is assigned a value of 8 for the designof vertical walls and a value of 2 for horizontal backsThe C factor reflects the inherently more stable nature ofvertical walls compared to a horizontal back In thispaper the ELOS curves employ a value for C of 8middot0 forthe footwalls For a more complete explanation offactors A B and C refer to papers by Potvin19 and Clarkand Pakalnis6

SUPPORT GUIDELINESThe development of support capacity guidelines iscritical to the overall success of the mining methodselected in terms of ensuring a safe work place (see Fig 2)Ground support in weak rock presents specialchallenges Under-design can lead to costly failureswhereas over-design can lead to high costs forunnecessary ground support Figure 6 depicts a classicwedge failure controlled by structure if the groundsupport has been under-designed It is critical to designfor the dead weight of the wedge in terms of the breakingload of the support as well as the bond strengthassociated with embedment length8 Split-Set andSwellex have been defined as continuous friction coupled(CFC) non-grouted bolts by CSIRO24 The load transfermechanism between the rock and borehole depends onthe borehole characteristics (hole diameter and



5 Stability graph

roughness) compared to the outside diameter of the boltelement Independent tests conducted by NIOSH12

Tomory et al20 and Villaescusa et al25 confirm thatrelationship for all diameter of bolts ndash 33-mm 39-mmand 46-mm Split-Sets

Over 400 000 Split-Set8 friction bolts are used inNevada mines for primary support Friction bolts areparticularly useful in fissile buckling or shearedground where it is difficult to secure a point anchorCaution must be used when using this method ofprimary support because of the low bond strengthbetween the weak rock mass and the bolt and becauseof the susceptibility of the bolt to corrosion In mine4 Split-Set bolts had a life of 6 months because ofcorrosion resulting from acidic ground conditions Ananalysis of the performance of friction bolts in mineswith weak rock (as determined by RMR) needed to beaddressed With one exception Nevada mines use 39-mm Split-Set bolts (the exception uses 46-mm Split-Set bolts) however mines in Canada generally use33-mm Split-Set bolts Canadian mines normally usethese bolts only in the walls and not in the back Table4 shows an updated support capacity chart asaugmented by this study

Data points gathered from several pull tests inweak rock were plotted as shown in Figure 7 A neural

net23 was superimposed on the mine data to determineif trends or predictions could be made The neural netmethodology has been used in establishing designcurves for span and stope design by Wang et al22 Thegraph shows a strong trend between RMR and bondstrength this relationship is being assessed as part ofon-going research Preliminary results are shown inFigures 7 and 8

Variability in test results shows the difficulty inassessing overall support for a given heading Thus itis important that mines develop a database withrespect to the support used so as to design for variableground conditions Factors critical to design such asbond strength hole size support type bond lengthand RMR should be recorded so as to determinetheir influence on the design curve to be determined




Rock properties (t)

Bolt strength Yield strength Breaking strength

58-inch mechanical 6middot1 10middot2Split-Set (SS 33) 8middot5 10middot6Split Set (SS 39) 12middot7 14middot0Standard Swellex NA 11middot0Yielding Swellex NA 9middot5Super Swellex NA 22middot020-mm rebar No 6 12middot4 18middot522-mm rebar No 7 16middot0 2325-mm rebar No 8 20middot5 30middot8No 6 Dywidag 11middot9 18middot0No 7 Dywidag 16middot3 24middot5No 8 Dywidag 21middot5 32middot3No 9 Dywidag 27middot2 40middot9No 10 Dywidag 34middot6 52middot012-inch cable bolt 15middot9 18middot858-inch cable bolt 21middot6 25middot514 times 4-inch strap 25middot0 39middot0

Note No 6 gauge = 68-inch diameter No 7 gauge = 78-inchdiameter No 8 gauge = 1-inch diameter NA = Not applicable

Screen Bag strength (t)

4 times 4-inch welded mesh 4 gauge 3middot64 times 4-inch welded mesh 6 gauge 3middot34 times 4-inch welded mesh 9 gauge 1middot94 times 2-inch welded mesh 12 gauge 1middot42-inch chain link 11 gauge bare metal 2middot92-inch chain link 11 gauge galvanised 1middot72-inch chain link 9 gauge bare metal 3middot72-inch chain link 9 gauge galvanised 3middot2

Note 4 gauge = 0middot23-inch diameter 6 gauge = 0middot20-inchdiameter 9 gauge = 0middot16-inch diameter 11 gauge = 0middot125-inchdiameter 12 gauge = 0middot11-inch diameter Shotcrete shear strength = 2 MPa (200 t mndash2)

Bond strength

Split-Set hard rock 0middot75ndash1middot5 Mt per 0middot3 mSplit-Set weak ground 0middot25ndash1middot2 Mt per 0middot3 mSwellex hard rock 2middot70ndash4middot6 Mt per 0middot3 mSwellex weak rock 3-3middot5 Mt per 0middot3 mSuper Swellex weak rock gt 4 Mt per 0middot3 m58-inch cable bolt hard rock 26 Mt per 1 mNo 6 rebar hard rock 18 Mt per 0middot3 m ~12-inch

granite

7 Rock mass rating versus pull-out strength A neuraltrend line is superimposed

6 Classic wedge failures

CONCLUSIONSThe Spokane Research Laboratory and the Universityof British Columbia Geomechanics group arefocusing on the development of safe and cost-effectiveunderground design guidelines for weak rock masseshaving an RMR in the range 15ndash45 Weak groundconditions ground support and mining methods usedin several Nevada underground mines were observedThe RMR values were calculated to update both spandesign calculations and stability graphs and adatabase on underhand mining methods was developedto reflect existing Nevada mining conditions Theimmediate rock mass was also characterised andanalysed in terms of prevailing type of ground supportpotential failure mechanisms and rock behaviour

Variability in field conditions shows the difficulty inassessing overall support for a given heading It isimperative that mines develop their own databasesbased on the type of support used in their mines sounexpected ground conditions can be analysed Theresults from augmented design curves and pull-out testsare presented in the hope that they will aid mineprofessionals in the task of designing a safe workplaceA systematic approach allows an operator to under-stand overall failure mechanisms and resultant loadsthat could affect the system This approach would allowan engineer to develop an optimal support strategy forthe mining method employed

The work would not have been possible without thepartnership between NIOSH the University of BritishColumbia Geomechanics Group and the Nevada goldmining company personnel This continued partnershipis critical to the development of safe and cost effectivemine strategies Figure 2 shows that since the inceptionof the lsquoteamrsquo approach and resultant collaboration thatinjury statistics have declined dramatically The causeand effect may be a result of many factors however it isclear that this approach is important and relevant tomine operations

REFERENCES1 N BARTON lsquoSome new Q-value correlations to assist in

site characterization and tunnel designrsquo Int J RockMech Min Sci 2002 39 185ndash216

2 N BARTON R LIEN and J LUNDE lsquoEngineeringclassifications of rock masses for the design of tunnelsupportrsquo Rock Mech 1974 6 189ndash236

3 Z T BIENIAWSKI lsquoRock nass classifications in rockengineeringrsquo Proc Symposium on Exploration for RockEngineering Johannesburg 1976 97ndash106

4 Z T BIENIAWSKI lsquoThe aeromechanics classification inrock engineering applicationsrsquo Proc 4th InternationalCongress on Rock Mechanics Vol 2 Rotterdam A ABalkema 1979 41ndash48

5 L CLARK lsquoMinimizing dilution in open stope miningwith focus on stope design and narrow vein longholeblastingrsquo Unpublished MSc thesis University of BritishColumbia Vancouver 1998

6 L CLARK and R PAKALNIS lsquoAn empirical designapproach for estimating unplanned dilution from openstope hangingwalls and footwallsrsquo 99th AnnualAGMndashCIM conference Vancouver 1997

7 R A CUMMINGS F S KENDORSKI and Z TBIENIAWSKIlsquoCaving mine rock mass classification and supportestimationrsquo US Bureau of Mines Contract ReportJ0100103 Washington DC

8 J M GORIS Personal communication Thiessen TeamUSA Inc February 20 2003

9 T HOCH Ground control seminar for underground goldmines Elko NV September 20 Sponsored by MSHAGround Control Division Pittsburgh PA 2000

10 E HOEK P KAISER and W F BAWDEN lsquoSupport ofunderground excavations in hard rockrsquo BrookfieldMA AA Balkema 1995

11 J A HUDSON and J P HARRISON lsquoEngineering rockmechanics an introduction to the principlesrsquo NewYork Elsevier Science 1997

12 J C JOHNSON T WILLIAMS C SUNDERMAN S SIGNER andD BAYER lsquoField test with strain gauged friction bolts atthe gold Hunter Mine Mullan Idaho USArsquo 22ndInternational Conference on Ground ControlMorgantown WV 2003 233ndash239

13 B LANG lsquoSpan design for entry type excavationsrsquoUnpublished MSc thesis University of BritishColumbia Vancouver 1994

14 S MAH lsquoQuantification and prediction of wall slough inopen stope mining methodsrsquo Unpublished MSc thesisUniversity of British Columbia Vancouver 1997

15 K E MATTHEWS E HOEK D WYLIE and M STEWARTlsquoPrediction of stable excavation spans for mining atdepths below 1000 m in hard rockrsquo CANMET DSSSerial No 0sQ80-00081 Ottawa 1981

16 C MAWDESLEY R TRUEMAN and W J WHITENlsquoExtending the Matthews stability graph for open-stopedesignrsquo Min Technol (Trans Inst Min Metall A)2001 110 27ndash39

17 S D NICKSON lsquoCable support guidelines forunderground hard rock mine operationsrsquo UnpublishedMSc thesis University of British Columbia Vancouver1992

18 R PAKALNIS lsquoEmpirical design methods ndash UBCgeomechanics updatersquo NARMS-TAC 2002

19 Y POTVIN lsquoEmpirical open stope design in CanadarsquoUnpublished PhD thesis University of BritishColumbia Vancouver 1988

20 P B TOMORY M W GRABINSKY J H CURRAN and J

CARVALHO lsquoFactors influencing the effectiveness ofSplit Set friction stabilizer boltsrsquo Can Inst Min 199891 205ndash214

21 J WANG R PAKALNIS D MILNE and B LANG lsquoEmpiricalunderground entry type excavation span design



8 Pull-out load versus RMR at mine 4

modificationrsquo Proc 53rd Annual Conference CanadianGeotechnical Society 2000

22 J WANG D MILNE and R PAKALNIS lsquoApplication of aneural network in the empirical design of undergroundexcavation spansrsquo Min Technol (Trans Inst MinMetall A) 2002 111 A73ndashA81

23 WARD SYSTEMS GROUP (Frederick MD) NeuroShellVersion 201 ltwwwwardsystemcomgt 2003

24 C R WINDSOR and A G THOMPSON lsquoRock reinforcementtechnology testing design and evaluationrsquo Vol 4 (edJ A Hudson) Comprehensive rock engineeringLondon Pergamon Press 1993 451ndash484

25 E VILLAESCUSA and J WRIGHT lsquoExcavation reinforce-ment using cemented grouted Split Set boltsrsquo AusIMMProc 1997 1 65ndash69

Authors

Tom Brady is a Senior Scientist with the Spokane ResearchLaboratory formerly of the US Bureau of Mines and nowpart of National Institute for Occupational Safety andHealth (NIOSH) He works on health and safety issues forunderground mining with his main focus or empirical minedesign for metal and nonmetal mines Recently Tom Bradywas selected as a Henry Crumb lecturer by SME

Lewis Martin is a Mechanical Engineer with Spokane ResearchLaboratory He works on safety issues for under-groundmining with emphasis on development of equipment Lewisholds several patents for instruments which monitorunderground loads and movements

Rimas Pakalnis is the Rock Mechanics Professor at theDepartment of Mining Engineering at the University ofBritish Columbia



Empirical design methods have been used successfullyover the past 30 years largely because they permit theoverall behaviour of a rock mass to be predicted easilyand accurately The basis for the success of theempirical method is a strong foundation of field datacoupled with on-going field observations that allowchanging rock conditions to be evaluated as miningprogresses Two systems were initially developed fromcase studies and databases originally derived fromcivil engineering applications and augmented by minestudies ndash the rock mass rating (RMR) system and theQ system As stated by the author lsquothe Q system isspecifically the permanent lining estimation systemfor tunnels and caverns in rock and mainly for civilengineering projectsrsquo1

There are several advantages for utilising the RMRsystem over the Q system The RMR system has arelatively straight-forward scoring system for eachparameter on a 0ndash10 scale and consequently is easierto learn1011

One of the original extensions of the rock mass ratingsystem from civil into mining was conducted under aUS Bureau of Mines Spokane Research Laboratoryresearch grant which developed empirical methods forblock caving operations for US copper mines7

It is important to remember that any method ofdesigning an opening must be easy to assessunderstand apply modify if necessary and reproducefor the next application if accepted as an on-goingoperational tool for design A critical factor is that thedesign incorporates the degree of stability required forany mine entry

SPAN DESIGN MAN-ENTRY METHODSThe lsquocritical span curversquo was developed in 1994 toevaluate back stability in cut-and-fill mines13 In 2000the span curve database of 172 observationsdeveloped by the University of British Columbia wasexpanded to include a total of 292 case histories frommines primarily in Canada21 The information fromthese case histories provides the basis for the spandesign curve shown in Figure 2



2 Design span curve


RMR Span () (m) Condition Other

Mine 145 5middot5 S Stable with support45 9middot0 S Stable with support40 6middot0 S Stable with support


Mine 340 7middot0 S Stable with support45 2middot1 S Stable with support26 2middot1 S Stable with support25 4middot6 U Caved with support55 7middot6 S Stable with support45 3middot0 S Stable with support

Mine 470 4middot6 S Stable with support40 4middot6 S Stable with support25 4middot6 S Stable with support55 5middot5 S Stable with support30 6middot1 S Caved upon longhole30 6middot1 S Caved upon longhole45 4middot6 S Stable with support50 6middot1 S Stable with support70 11middot3 S Stable with support25 7middot3 S Stable with support30 3middot0 U Prior to support placement30 1middot8 S Prior to support placement50 6middot1 S Stable with support55 7middot6 S Stable with support55 6middot1 S Stable with support

Mine 530 3middot0 S Stable with support30 4middot3 U Caved with support20 5middot8 U Caved with support15 3middot7 S Stable with support


Mine 740 4middot6 S Stable with support35 4middot6 S Stable with support

Mine 825 5middot0 U Caved had to spile




55 3middot7 S No support typical round no problems

35 4middot6 S No support typical round no spileshotcrete

20 7middot6 U Caved had to spile45 6middot0 S Stable with split-sets only



























ELOSHR
















5 Stability graph










Rock properties (t)







Bond strength


granite





































Authors
































ELOSHR
















5 Stability graph










Rock properties (t)







Bond strength


granite





































Authors


















ELOSHR
















5 Stability graph










Rock properties (t)







Bond strength


granite





































Authors







ELOSHR
















5 Stability graph










Rock properties (t)







Bond strength


granite





































Authors















Rock properties (t)







Bond strength


granite





































Authors








































Authors











Authors






Documents

Empirical approaches for opening design in weak rock masses