ACARP PROJECT C14029

PUBLISHED February 2009

ANALYTICAL TOOLS FOR MANAGING ROCK FALL HAZARDS IN AUSTRALIAN COAL MINE ROADWAYS

R Seedsman, N Gordon & N Aziz CIVIL, MINING AND ENVIRONMENTAL ENGINEERING, UNIVERSITY OF WOLLONGONG, SEEDSMAN GEOTECHNICS PTY LTD & GORDON GEOTECHNIQUES PTY LTD

DISCLAIMER
No person, corporation or other organisation (person) should rely on the contents of this report and each should obtain independent advice from a qualified person with respect to the information contained in this report. Australian Coal Research Limited, its directors, servants and agents (collectively ACR) is not responsible for the consequences of any action taken by any person in reliance upon the information set out in this report, for the accuracy or veracity of any information contained in this report or for any error or omission in this report. ACR expressly disclaims any and all liability and responsibility to any person in respect of anything done or omitted to be done in respect of the information set out in this report, any inaccuracy in this report or the consequences of any action by any person in reliance, whether wholly or partly, upon the whole or any part of the contents of this report.

ANALYTICAL TOOLS FOR MANAGING ROCK FALL HAZARDS IN AUSTRALIAN COALMINEROADWAYS

ROSSSEEDSMAN1,2 NICKGORDON3 NAJAZIZ1 1Civil,MiningandEnvironmentalEngineering,UniversityofWollongong 2SeedsmanGeotechnicsPtyLtd 3GordonGeotechniquesPtyLtd

UNIVERSITYOFWOLLONGONGDISCLAIMER The information contained in this report is for general information purposes only. The University of Wollongong, its officers, employees and agents make no representations or warranties, express or impliedastotheaccuracy,reliabilityorcompletenessoftheinformationcontainedwithinthisreport. All liability for loss or damage of any kind at all, whether indirect or consequential, against the University of Wollongong, its officers, employees and agents arising from or through the use of the information in this text is excluded. Your use of this text is acknowledgement that the information providedhereinistoassistyouwithundertakingyourownenquiriesandanalysisandthatyoushould alwaysseekindependentprofessionaladvicebeforeactingontheinformationcontainedtherein. SEEDSMANGEOTECHNICSPTYLTDDISCLAIMER The information contained in this report is for general information purposes only. Seedsman Geotechnics, its directors and servants make no representations or warranties express or implied as to the accuracy, reliability or completeness of the information contained therein. Seedsman GeotechnicsPtyLtd,itsdirectorsanditsservantsexcludeallliabilityforlossordamageofanykindat all(includingindirectorconsequentiallossordamage)arisingfromtheinformationinthistextoruse of such information. You acknowledge that the information provided in this text is to assist you with undertaking your own enquiries and analysis and that you should seek independent professional advicebeforeactinginrelianceontheinformationcontainedtherein.

GORDONGEOTECHNIQUESPTYLTDDISCLAIMER The information contained in this report is for general information purposes only. Gordon Geotechniques, its directors and servants make no representations or warranties express or implied as to the accuracy, reliability or completeness of the information contained therein. Seedsman GeotechnicsPtyLtd,itsdirectorsanditsservantsexcludeallliabilityforlossordamageofanykindat all(includingindirectorconsequentiallossordamage)arisingfromtheinformationinthistextoruse of such information. You acknowledge that the information provided in this text is to assist you with undertaking your own enquiries and analysis and that you should seek independent professional advicebeforeactinginrelianceontheinformationcontainedtherein.

 ABSTRACT This report provides a reference source for the design of ground control measures in coal mine roadways using analytical methods. Analytical methods are based on identifying potential failure modes from knowledge of the geology, followed by simple analyses based on hand calculations, spreadsheets and design charts. It is intended that this report will complement other publications thatprovideempiricalandmoreintensivenumericalapproaches. Coal mining is conducted in layered materials different rock types, different strengths, different layer thickness, different joint spacing. Underground coal mines openings are rectangular, and at least in Australia, the roadway axes are parallel to the dominant discontinuity sets bedding and joints.Thestressfieldondevelopmenthastheprincipalaxesparalleltotheroadwayaxes,andthere canbemajorchangesinandaroundextractionpanels. These geometrical factors require a different approach to the empirical rock mass rating approaches used in metal mines. Fortunately, the comparatively simple geometry allows analytical methods based on identifying collapse modes associated with jointed and bedded material. This ability to analyse the relatively simple geometries allows a more anticipatory approach to ground control management including anticipation of geological conditions, prediction of collapse modes, the design of support or reinforcement system, and the monitoring of ground conditions for exceptions. The approachprovidesmorerobustoutcomesandlimitsthedemandsontheobservationalmethod. Collapse models are provided for roof and rib. The roof models recognise that different collapse modescanapplyindifferentstressfieldshigh,intermediate,andzerocompressivestresses.Therib modelsdrawanalogiestorockslopestabilityandalsotheimpactofhighverticalstresses Methods for determining support or reinforcement requirements are provided. Suspension of collapsed masses is identified as the basis for roof support in both very high and zero compressive stress regimes. Reinforcement of bedding discontinuities is advocated for intermediate compressive stresses. For the ribs, restraint of coal blocks defined by preexisting joints or by mining induced fracturesisrequired.

EXECUTIVESUMMARY The mining hazard that is addressed in this report is the gravitydriven fall of a block of ground into a roadway. The block of ground may be defined by preexisting discontinuities and/or by mining inducedfractures. The purpose of this report is to provide information to assist the geotechnical officer in identifying potentially unstable roof and rib, and to document design methods that may be considered in the specificationofgroundsupport.Thereportisfocusedonthestabilityofthegroundintheimmediate vicinityoftheroadwaythatistheimmediateroofandrib.
The report has been written for geotechnical engineers, geologists, and mining engineers charged with specifyinggroundsupport.Personsoccupyingthepositionsofminemanagerortechnicalservicesmanager should find the report of value. The report addresses some of the steps (site characterisation, model formulation, design, implementation, monitoring, and review) that can be found in Clause 48 of NSW regulationsforcoal,ortherockfallriskmanagementoftheMineralsCouncilofAustralia(PotvinandNedin, 2003). It is stressed that a full understanding of coal mine ground control is not yet available hopefully this report represents a step forward. One use of this report can be to challenge the standard interpretations of how a rock mass behaves around openings so that there can be a gradual improvement in the ability to design safe and productive mine roadways. The concepts introduced can be applied at early stages prefeasibility and feasibility using presumed values based on a review of the geological conditions. As the project progresses to operational phase, the greater the importance to move from presumed to demonstratedvaluesforthevariousinputparameters. Thestructureofreportisasfollows: Section1servesasanintroduction Section 2 of the report discusses design in rock masses, highlighting the uncertainties that are intrinsic to design in rock masses which are very difficult to fully characterise. The key message is that any design in rock mechanics will need to be implemented with a commitment to monitor andrespondasthegeologicalconditionsarerevealed. Section 3 defines rock falls and makes a distinction between the nature of the hazards in metal minesandcoalmines Section 4 summarises underground coal mining practices in Australia so that the practical constraintsonroadwaygeometriesandexcavationmethodscanbeappreciated. Section 5 provides a reference guide to the engineering geology of coal measures and includes discussionsondiscontinuities,rockstrengthanddeformationproperties,andtheinsitustresses. Section 6 examines the redistribution of stresses around a retreat longwall panel and particularly in the immediate vicinity of a roadway. The key point is that our knowledge of stress above chain pillars needs to be interrogated in more detail to understand the stresses that are present at the roof line. Simple numerical analyses for stress about openings in a homogeneous material are presented Section 7 presents the key strength properties of the popular support and reinforcement elementsinuseinAustraliaandpresentsmodelsfortheirreinforcingaction. Section 8 provides a flow chart for the specification of roof support and a number of analytical toolsthatcanbeusedtodetermineboltingdensitiesandboltlengths.
IV

Section9providesaflowchartforribsupportdesign.

Geotechnical engineering requires consideration of the inevitable uncertainties of the properties of rock masses. It is not possible to know all of the properties of a rock mass: conversely it is evident from the range of successful rock engineering ventures that this is also not necessary. It follows that there is a need to identify the key parameters in the rock mass and the ways of dealing with them. Thusgeotechnicalengineeringrequiresbothgeologicalandengineeringinput. Inananalyticalengineeringapproach,akeystepisidentifyingmodeofcollapse.Forthisgeotechnical models of behaviour are created. Characteristics of such models are that they capture the key features (efficient), they are amenable to mathematics (practical), and they anticipate conditions to be encountered (predictive). An important aspect of the approach is to identify failure and collapse modes based on observation and only then to seek to analyse the mechanics involved. The identificationof failure andcollapse modes allows the application of limit equilibrium techniques and theuseoffactorsofsafetytoassessthelevelofstability. The starting model is jointed layered roof with layers parallel to the roadway. This is much simpler than the metaliferous mining problem which involves arched roofs and rock wedges. Depending on thestresscondition,thissystemmayfailundertension,compression,orselfweight.Duringroadway development, the immediate stone roof is exposed to high levels of deviatoric stress, which quickly dissipatesastheroofdeflects.Forcoalroof,theimmediateroofstressesaretensile.Atthemaingate corner, the deflectedsoftened roof means that the increased stresses around the longwall goaf are deflected higher into the roof. At the tailgate, the vertical stresses become dominant and there is reliefofhorizontalstressestowardsthegoaf. A logical framework to assess stability and specify support is provided by way of flow charts for each of the roof and the ribs. The flow charts need to be applied at each stage of mining. Commentaries onpragmaticaspectsofthesupportdesignareofferedforguidance.

TABLEOFCONTENTS ABSTRACT...........................................................................................................................................................iii EXECUTIVESUMMARY .......................................................................................................................................iv TABLEOFCONTENTS.........................................................................................................................................vi 1 INTRODUCTION........................................................................................................................................... 1 1.1 1.2 1.3 1.4 2 2.1 2.2 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.4 2.4.1 2.4.2 3 3.1 3.2 3.3 3.4 4 4.1 4.2 4.2.1 Thechallenge................................................................................................. 2 Targetreaders............................................................................................... 2 Structureofthereport.................................................................................. 3 LimitatonsANDOPPORTUNITIESFORIMPROVEMENTS...............................3 DesignProcess............................................................................................... 6 Rockasanengineeringmaterial....................................................................9 Analysistools................................................................................................. 9 Precedentandpractice...........................................................................10 Empirical.................................................................................................. 10 Analytical................................................................................................. 11 Numerical................................................................................................ 13 Uncertainty,risk,andtheobservationalmethod........................................14 ObservationalMethod............................................................................15 Otherrisktools........................................................................................ 16 Differencebetweencoalandmetalmines..................................................17 Roofcollapse............................................................................................... 18 Ribcollapse.................................................................................................. 22 Approachestopreventrockfalls.................................................................23 Coalseams................................................................................................... 25 Miningmethods ........................................................................................... 25 Development........................................................................................... 25 4.2.1.1 Methods.......................................................................................... 25 4.2.1.2 SupportRuleDatabase...................................................................27 4.2.2 Extraction................................................................................................ 28 4.2.2.1 Longwalls........................................................................................ 28 4.2.2.2 Secondarysupport..........................................................................28 5 ENGINEERINGGEOLOGY ...........................................................................................................................31 5.1 5.1.1 5.1.2 5.1.3 5.1.4 5.1.5 Discontinuities............................................................................................. 31 Descriptionofdiscontinuities ..................................................................32 Beddingpartings..................................................................................... 34 Jointsinstone .......................................................................................... 35 Jointsincoalcleats................................................................................ 37 Faults....................................................................................................... 40
VI

DESIGNINROCKMECHANICS ..................................................................................................................... 5

ROCKFALLS ................................................................................................................................................17

UNDERGROUNDCOALMININGINAUSTRALIA.......................................................................................25

5.2 5.3 5.3.1

Lithologies .................................................................................................... 41 Strength....................................................................................................... 44 Uniaxialcompressivestrength................................................................44 5.3.1.1 Stone............................................................................................... 44 5.3.1.2 Coal................................................................................................. 46 5.3.2 5.3.3 5.3.4 5.3.5 5.3.6 Tensilestrength....................................................................................... 47 Triaxialparameters .................................................................................. 48 Brittlestrength........................................................................................ 49 Shearstrengthofdiscontinuities............................................................50 Massstrength.......................................................................................... 51 5.3.6.1 Stone............................................................................................... 51 5.3.6.2 Coal................................................................................................. 52

5.4 5.4.1 5.4.2 5.5 5.5.1 5.5.2 5.5.3 5.5.4 5.6 6 6.1 6.1.1 6.1.2 6.1.3 6.2 6.2.1 6.2.2 6.2.3 6.2.4 6.2.5

Deformationproperties............................................................................... 52 Modulus.................................................................................................. 52 Poissonsratio......................................................................................... 54 InsituStresses............................................................................................. 54 Stressinstone......................................................................................... 54 Stressincoal............................................................................................ 55 Faultedground........................................................................................ 57 Topography............................................................................................. 58 Localterminology........................................................................................ 59 Redistibutionaboutalongwall....................................................................60 Maingatecorner...................................................................................... 60 Bleeder/Tailgate...................................................................................... 62 Tailgatecorner........................................................................................ 62 Redistributionaboutaroadway..................................................................63 Basicconcepts......................................................................................... 63 Elasticstressredistributionaroundarectangularroadway ....................65 ThreeDimensionalstresschanges..........................................................70 Nonlinearstressredistributions.............................................................73 Stressesinducedwithinablockyroof.....................................................75 6.2.5.1 Voussoirbeams............................................................................... 75 6.2.5.2 Cantilevers...................................................................................... 76

MININGINDUCEDSTRESSCHANGES.......................................................................................................60

6.3 6.3.1 6.3.2 6.4 6.4.1 6.4.2 6.4.3 7 7.1

OTHERSTRESSREDISTRIBUTIONS...............................................................77 Tailgatepillars......................................................................................... 77 Veryhighstressesunderorabovepillars...............................................77 Compiledmodelsforstresspathsintheimmediateroof...........................78 Stoneroofsingleseam.........................................................................78 Coalroofsingleseam...........................................................................80 Stresspathsintheribs............................................................................81 Reinforcementaction.................................................................................. 82

VII

SUPPORTANDREINFORCEMENTTECHNOLOGIES..................................................................................82

7.1.1 7.1.2 7.2 7.3 7.4 8 8.1 8.1.1 8.1.2 8.1.3 8.2 8.2.1 8.2.2 8.2.3 8.3 8.3.1 8.3.2 8.3.3 8.4 8.4.1 8.4.2

Doweleffect............................................................................................ 83 Frictioneffect.......................................................................................... 83 Tendons....................................................................................................... 85 Anchorages.................................................................................................. 86 Strapsandpanelsskinrestraint................................................................87 Nonverticaljoints....................................................................................... 91 Collapsemodel........................................................................................ 92 Supportdesign........................................................................................ 93 Commentaryondesigntopreventcollapsewithnonverticaljoints.....93 Compressivefailure..................................................................................... 93 Collapsemodel........................................................................................ 94 Supportdesign........................................................................................ 98 Commentaryondesigntopreventcompressivecollapse....................101 Tensilefailure............................................................................................ 103 Collapsemode....................................................................................... 104 Support.................................................................................................. 105 Commentaryondesigntopreventtensilecollapse..............................105 Delaminationfailure.................................................................................. 106 Collapsemodel...................................................................................... 107 Reinforcementdesign...........................................................................108 8.4.2.1 Drivingforces................................................................................ 109 8.4.2.2 Resistingforces................................................................................. 1

PREVENTIONOFROOFCOLLAPSE............................................................................................................89

8.4.3 8.4.4 9 9.1 9.1.1 9.1.2 9.1.3 9.2 9.2.1 9.2.2 9.3

Densitiesandpatterns..........................................................................111 Commentaryondesigntopreventdelamination.................................113 Structurecontrol....................................................................................... 116 Slides..................................................................................................... 117 Wedges.................................................................................................. 118 Topples.................................................................................................. 118 Stressinducedribcollapse........................................................................119 Mininginducedfractures......................................................................119 Buckling................................................................................................. 120 Commentaryonribsupportdesign...........................................................121

PREVENTATIONOFRIBCOLLAPSE.........................................................................................................115

10 REFERENCES...........................................................................................................................................123 ListofTables Table1Classificationofprediction(afterLambe,1973)..........................................................................5 Table2Predictionclasses(afterMorganstern,2000).............................................................................. 5 Table3CommentaryonrecentimplementationofSUconstitutivemodelsinFLAC............................14 Table4Observationsanddatarequirementsforroofcollapsemodes.................................................21

VIII

Table5AustralianLongwallstatistics(afterCram,2008)......................................................................28 Table6Classificationofdiscontinuityspacing....................................................................................... 32 Table7Classificationofdiscontinuitypersistence................................................................................. 32 Table8Roughnessofbeddingpartings .................................................................................................. 35 Table9Jointroughnesscoefficient(JRC)ofcoaljoints..........................................................................40 Table10Visualandtactiletests(ISRM,1978)........................................................................................ 46 Table11Rangeofcoalunconfinedcompressivestrengths(ACIRL,1986).............................................47 Table12Presumedfrictionanglesandmvalues................................................................................... 49 Table13HoekBrownparametersforvariouscoals.............................................................................. 52 Table14Guidelinesfortheselectionofmodulusratio(afterHoekandDiederichs,2006)..................52 Table15Compilationofstressmeasurementstakenfromunderexcavations.....................................56 Table16Localgeologicalterminology................................................................................................... 59 Table17Roofandsidestressesforcirclesandellipses.........................................................................64 Table18Summaryofstressesforrectangularroadway(1=10MPa).................................................65 Table19SummaryofAustralianbolttypes........................................................................................... 85 Table20SummaryoftensilestrengthofUSAbolttypes.......................................................................85 Table21Longtendons........................................................................................................................... 85 Table22Databaseforbrittlefailureanalyses....................................................................................... 94 Table23BoltlocationstoresistBPXS.................................................................................................. 112 ListofFigures Figure1Stepsingeotechnicaldesignprocesscoveredinthisreport ......................................................8 Figure2Accuracyofapredictionisafunctionoftheamountofdataavailableandthemethodused (afterLambe,1973)........................................................................................................................ 10 Figure3Logicalframeworkformineexcavationdesigninmassiverock(BradyandBrown,1985)......12 Figure4Geometryofacoalmineroadwayanditscoaxialrelationshipwithdiscontinuitiesand stresses...........................................................................................................................................18 Figure5Typicalgeometryanddiscontinuityfieldforametalmineroadway(BradyandBrown,1985) ........................................................................................................................................................18 Figure6Generalcollapsemodesforabeddedroofwithsubverticaljoints .........................................19 Figure7Dippingjointscanallowblockmovementwithorwithouthorizontalroofstresses...............20 Figure8Collapsefollowingdelamination.............................................................................................. 20 Figure9Collapserelatedtoreleasealongverticaljoints.......................................................................20 Figure10Thinslabsbetweenboltsretainedwithmeshpanels............................................................21 Figure11Ribcollapseanalogiestorockslopes(topples,planarslides,wedges)..................................22 Figure12Bucklingofacoalrib............................................................................................................... 22 Figure13Backsurfacetoaslabformedbyamininginducedfractureinacoalrib..............................23 Figure14Flowchartfortheanalysisphasesforthepreventionofcollapseofground(Potvinand Nedin,2003)...................................................................................................................................24 Figure15Typicalbolterminer(SourceJoy)........................................................................................ 26 Figure16Typicalplacechangeminerandmobilebolter(SourceJoy)...............................................26 Figure17SummaryofInstalledCapacityofPrimarySupportforGateroads.........................................27

IX

Figure18SummaryofInstalledCapacityofPrimarySupportforIntersections....................................27 Figure19Summaryofcombinedminimummaingateprimaryandsecondarysupport ........................29 Figure20Summaryofcombinedmaximummaingateprimaryandsecondarysupport.......................29 Figure21Summaryofminimummaingatesecondarysupport.............................................................30 Figure22Summaryofmaximummaingatesecondarysupport............................................................30 Figure23Differencebetweentruecohesionofanintactsampleandapparentcohesionofarough discontinuity...................................................................................................................................32 Figure24JointRoughnessProfilesandCorrespondingJRCValues(BartonandChoubey,1977).........33 Figure25AlternativemethodforestimatingJRCfrommeasurementsofsurfaceroughnessamplitude fromastraightedge(BartonandChoubey,1977)........................................................................34 Figure26Exampleofjointpatternsinsedimentarystrata....................................................................35 Figure27Jointsinundeformedsediments(Fookesetal,2000)............................................................36 Figure28Relationshipbetweenjointspacingandbeddingthickness(Jietal,1998)...........................37 Figure29Associationbetweenjointsandfaults/folds(Fookesetal2000)...........................................38 Figure30Jointedandcleatedcoalnoteverylowpersistenceofcleat(MedhurstandBrown,1998) 39 Figure31Anexampleofrotationofcoaljointsaboutthrustfaults......................................................39 Figure32CleatandjointorientationsinBowenBasin(Pattison,1995)................................................40 Figure33Distributionoflithologiesinfluviatilesystem(Fookesetal,2000) ........................................42 Figure34Distributionoflithologiesinadeltaicsystem(Fookesetal,2000)........................................43 Figure35Plotshowingtherangeinstrengthanddeformationmodulusfordifferentlithologies (ShepherdandGale,1982)............................................................................................................. 44 Figure36AvarietyofUCSsonicrelationshipsforstone........................................................................45 Figure37UCSsonicrelationshipsforstoneatlowerstrength..............................................................45 Figure38UCSsonicvelocityrelationshipforcoal.................................................................................. 47 Figure39Strengthenvelopes(dashMohrCoulomb,solidHoekandBrown)..................................48 Figure40Brittlerockparameters.......................................................................................................... 49 Figure41Strengthenvelopeforadiscontinuity.................................................................................... 50 Figure42Geologicalstrengthindexformolasses.................................................................................. 51 Figure43Modulusreductionforcaseofdamageindex=0..................................................................53 Figure44Simplemodelfortheimpactofdiscontinuityspacingonthedeformationmodulusand Poissonsratio................................................................................................................................53 Figure45SummaryofstressmeasurementdatafromNewSouthWalesandQueenslandcoalfields (Nemciketal,2006)....................................................................................................................... 55 Figure46Measuredminimumstressincoalasafunctionofdepth(Eneveretal,2000).....................56 Figure47Reducedlateral(horizontal)stressassociatedwiththepresenceofsurfaceswithlower frictionalresistancebasedon(a)passiveearthpressures,(b)UDECmodeling(Nemciketal2006). ........................................................................................................................................................57 Figure48Effectoftopographyonstressdistribution............................................................................58 Figure49Horizontalstressesinhilltoporridgemining.........................................................................59 Figure50Pillardesignverticalstressesdevelopedabovechainpillarsassuminga200mwidepanel..61 Figure51Generalpatternofverticalandhorizontalstressredistribution(Gale2008).......................61 Figure52Concentrationofhorizontalstressmagnitudeatthemaingatecornerasafunctionofthe anglebetweenprincipalhorizontalstressaxisandtheroadwaydirection(afterGale2008).......62 Figure53Stressesaboveapillarinatailgate(Shenetal2006)............................................................63 Figure54Effectofplanesofweaknessondistributionofroofstresses................................................65
X

Figure55Distributionsofdeviatoricstressesabouta1.86:1roadway(1=10MPa)..........................66 Figure56Distributionofhorizontalstressesabouta1.86:1roadway(1=10MPa)............................67 Figure57Distributionofverticalstressesabouta1.86:1roadway(1=10MPa)................................68 Figure58Negativeminimumprincipalstressesinducedarounda1.86:1roadway(1=10MPa).......69 Figure59Exampleofshearandnormalstressesandbeddingparallelexcessshearstressassuminga .......................................................................................................... 70 35ofrictionangleandK=2.0 Figure60ContoursofBPXSasafunctionofheightintodistanceacrosstheroof(expressedas MPa/MPa)......................................................................................................................................70 Figure61Developmentofdeviatoricstressesat0.2mabovetherooflineastheheadingisadvanced (K=2)...............................................................................................................................................71 Figure62DistributionofnegativemeanstressintheimmediateroofofaroadwayforK=0.15..........71 Figure63ThreedimensionalviewofthedistributionofBPXS(ignorenegativesign) ...........................72 Figure64VerticalslicethroughFigure63showinghowBPXSdevelops(ignorenegativesign)............72 Figure65BPXSonasurface0.2mabovetheroofline...........................................................................73 Figure66StressmeasurementsatEmeraldMine.................................................................................. 74 Figure67Simulationofstressredistributionabovearoadwayusinganelasticmodel........................74 Figure68Conceptofastressrelievingroadway.................................................................................... 75 Figure69Voussoirbeamdeformationsinducecompressivestressesattheroofcornersandtensile stressesattheroadwaycentreline................................................................................................ 76 Figure70Exampleofashearsurfacegeneratedbycantileveringaction..............................................76 Figure71Relaxationofarooflineasaresultofverticaldeformationinoneofthesides....................77 Figure72Stressesunderarigidfooting................................................................................................. 78 Figure73Evolutionofdeviatoricandnegativeminorstressesduringlongwallretreat........................79 Figure74RedistributedinsituandinducedbodystressesaboutaroadwaywithK>0.8oncetheroof andfloordeflects............................................................................................................................80 Figure75RedistributedinsituandinducedbodystressesaboutaroadwaywithK<<0.8oncetheroof andfloordeflects ............................................................................................................................ 81 Figure76Tendonsinshear.................................................................................................................... 83 Figure77Dowelresistancefor21mmdiametertendonsasafunctionofUCS.....................................83 Figure78Importanceofboltangleinmaintainingclosedbedding.......................................................84 Figure79Frictionalshearresistanceprovidedbybolts.........................................................................84 Figure80Recommendedminimumanchoragelengthsincoalmeasurerockswithresinanchorages .86 Figure81Loadingofastraporpanelifloadedasacatenary................................................................88 Figure82Resultsofloading1.5mand2.0msquaremeshpanels(Thompson,2004)...........................88 Figure83Logicalframeworkappliedtocoalmineroofsupport...........................................................90 Figure84Thehazardofparallelnonverticaljoints............................................................................... 91 Figure85Nonparalleljointsdefiningtriangularprisms........................................................................92 Figure86Relationshipbetweenjointdip,jointfrictionandhorizontalroofstressforastable symmetricprism............................................................................................................................. 92 Figure87TwoanalysisofbrittlestrengthfactorusingExamine2Dshowingzonesofbrittlefailureand possibleboltingandcablepatterns................................................................................................ 94 Figure88RelationshipbetweenheightofstrengthfactorandtheKvaluefora5.2mby2.8mroadway andamajorprincipalstressof10MPa.......................................................................................... 96 Figure89HeightoffailureasafunctionofroofstrengthindexforvariousKvalues...........................97 Figure90Comparisonbetweentheresultsofmultiplelinearregressionandthenumericaldata.......97
XI

Figure91Slightchangesinfailurezonesneartheexcavationwithnodifferencebeyond2m.............97 Figure92Differentisotropyandfailureconditions.............................................................................102 Figure93Negativehorizontalstress(K=0.2)fora5mby3mroadway ................................................103 Figure94Heightofnegativehorizontalstress(5.5mby3.3mroadway) .............................................104 Figure95Collapsemodefortensilestressregime ............................................................................... 104 Figure96Roadwaysobliquetojointsetswillproducebetterconditionsinbothheadingsand cutthroughs..................................................................................................................................105 Figure97Loadingonmeshpanels....................................................................................................... 105 Figure98Generationofcompressivestressesandfailureinaregimeofnoimposedhorizontalstress ......................................................................................................................................................106 Figure99Slipandseparationinalayeredroofrock(BradyandBrown,1985)...................................107 Figure100Criticalthicknessanddeflectionofvoussoirbeamsasafunctionofspanandrockstrength (E/UCS=250).................................................................................................................................107 Figure101Criticalthicknessanddeflectionofa5.5mspanvoussoirbeam(1msurcharge) ..............108 Figure102Stabilityofavoussoirbeamincreaseswithincreasingthickness(1mequivalentsurcharge) ......................................................................................................................................................108 Figure103Examine3Dgeometry........................................................................................................ 109 Figure104BPXSat2.31mfromtheface............................................................................................. 110 Figure105CumulativeincreaseinBPXSwithdistancefromtheribline.............................................110 Figure106AverageBPXSacrosstherooflinefordifferentheightsintotheroofanddifferentlocations ofbolts(areaundercurvesinFigure107expressedasuniformlydistributedload....................111 Figure107BPXSasafunctionofthelocationandheight....................................................................111 Figure108XSfactorfor0.2mintotheroof .............................................................................................. 1 Figure109XSfactorfor0.4mintotheroof .............................................................................................. 1 Figure110Gparametertoaccountfordifferentfrictionangles.............................................................1 Figure111Thebolttimingfactor............................................................................................................. 1 Figure112Fullygroutedboltsareverystiffduringinitialshearloading(Stjern,1995)......................111 Figure113ComparisonbetweenidealBPXSsupportcapacityat0.4mintotheroof(Kn=2.0,Kp=1.5, G=1.3)andtheAustraliandatabase............................................................................................. 111 Figure114Optimumboltpatternsfromphysicalmodels(Stimpson,1987).......................................112 Figure115Logicalframeworkforsupportofribs................................................................................ 115 Figure116Failuremodesforrockslopesthatcanbeobservedincoalribs(HoekandBray,1981)...116 Figure117Allintersectionsarenoses.................................................................................................. 117 Figure118Planargeometryandrequiredfacesupportfora3mhighrib...........................................117 Figure119Wedgegeometryandrequiredfacesupport.....................................................................118 Figure120Anchorforcetopreventtoppling....................................................................................... 118 Figure121Distributionofthecohesivelosscomponentaboutroadways..........................................119 Figure122ValueofspallingratioalongthespringlineoftheribfordifferentKratiosandroadway aspectratios.................................................................................................................................120 Figure123Estimationofmaximumdepthofbrittlefailure.................................................................120 Figure124Thicknessofslabsthatmayundergobuckling...................................................................121

XII

GLOSSARY outputisdependentonaleastoneinputvariableappliedtoalaw ofphysics Brittle Completelossofstrengthafterfailure. CHILE Continuous,Homogeneous,Isotropic,LinearElastic Cohesion The intercept on the shear stress axis of a straight line fitted to a scatterplotofnormalandshearstresses Continuum Amaterialwithnogaps Deviatoricstress The difference between the magnitude of the major and minor principalstresses DIANNE Discontinuous,inhomogeneous,anisotropic,notelastic Discontinuity Afeatureisarockmasswithzeroornegligibletensilestrength Empirical Basedoractingonobservationandexperiment,notontheory Frictionangle Theslopeofthestraightlinefittedtoascatterplotofnormaland shearstresses Homogeneous,inhomogeneous Samematerial Isotropicanisotropic Samepropertiesinalldirections Joint An approximately planar natural fracture in a rock mass that is typicallypartofaparallelset Linearelastic An elastic body returns to its original form after a displacing stressisremoved. Modulus Theratiobetweenappliedstressandresultantstrain Numerical Useofcomputerstosolvecomplexstressredistributions Parting,beddingparting A natural fracture in a sedimentary rock that is parallel to the beddingtexture Planestrain Deformationsoutoftheplanebeingconsideredarezero. PoissonsRatio The ratio of the normal strain to the transverse strain of a body underuniaxialstrain Principal stress, major, The magnitude and direction of the stresses that are normal to intermediate,minor planes where the shear stresses are zero. The stress field is 3 dimensionalsothereare3principalstresses Reinforcement Theadditionofrestrainingforcestojointsorpartings Rib Sidesofanexcavation Rockfall Anunplannedfallofgroundofanysizethatcauses(orpotentially causes)injuryordamage Slickenside Polished or striated surfaces that result from friction along movementsurfaces. Support The addition of tendons to suspend blocks from higher rock units thatarestable Uniaxialcompressivestrength The strength obtained in a laboratory test where the rock is loadeduniaxially Voussoir Abockofrockinamasonaryorrockarch Analytical

XIII

ABBREVIATIONSANDSYMBOLS a Boltfriction BPXS By c C D d Dowel E Fbolttiming Ffriciton Fu G GSI H Hf JCS JRC Kn kT L,M L,l Lf m Mv PLSI PR Pv q q RSI s T t Tf U UCS,1 vel W XS 1,2,3 n AconstantintheHoekBrownfailurecriterion The angle between the roadway direction and the direction of the major principal horizontalstress Theadditionalshearresistancesuppliedbyatendoninstalledacrossaclosedparting Beddingparallelexcessshear Yieldstressofsteel Cohesion Constant Damage constant diameter Theadditionalshearresistancesuppliedbyatendoninstalledacrossanopenparting ModulusofElasticity,Youngsmodulus A factor in the determination of the BPXS to account for different friction angles on the beddingpartings A factor in the determination of the BPXS to account for different friction angles on the beddingpartings Tensilestrengthofsteelsheet Independentshearmodulusinatransverseanisotropicmaterial GeologicalStrengthIndex Roadwayheight Heightofbrittlefailure JointCompressiveStrength JointRoughnessCoefficient Jointstiffness Parameterintopplingequation Theratioofthemajorandminorprincipalstressestotheverticalstress Length Horizontalstressconcentration aroundthemaingatecorner AconstantintheHoekBrownfailurecriterion Vertical stressconcentration atvariousstagesofmining PointLoadStrengthIndex PoissonsRatio Verticalloadonmesh Sagofmesh constantwithvaluesof1forbothendspinjointed,and0.5forbothendsclamped. RoofStrengthindex AconstantintheHoekBrownfailurecriterion Anchoragelength Strapthickness Anchoragelength Tensileloadinbolt UniaxialCompressiveStrength Sonicvelocity Roadwaywidth A variable used to determine the BPXS from the vertical stress assuming a friction angle of35oandbolting2.31mfromtheface Density Major,intermediate,andminorprincipalstresses Normalstress Shearstress Frictionangle

XIV

INTRODUCTION

Rock fall An unplanned fall of ground of any size that causes (or potentially causes) injury or damage1. Rock falls have been and continue to be unacceptable events in underground coal mines. Over the last 30 years there have been major improvements in the way in which Australian coal mines have managed the rock fall hazard through the application of safe methods of work, technological improvements,alternativeminingsystems,andriskmanagementconcepts. All Australian underground coal mines now have strata hazard management plans. These plans nominate a person to manage the mines ground control regime depending on the size of the operation they may have tertiary level qualifications or they may hold the position based on their industryexperience.Inthisreport,thispersonwillbereferredtoasthegeotechnicalofficer. The geotechnical officer usually has mining/civil engineering or geology qualifications, and in many cases a postgraduate qualification with a component of rock mechanics. Historically, the subject of mining rock mechanics has mostly focused on metal mining and tunneling. Underground coal mining hasanumberofsignificantdifferencesfrommetalminingintermsofthenatureoftherockmass,the mining machines and the shape of the openings. This report has been written to specifically address thisvoidinthetechnicalliterature. The purpose of this report is to provide information to assist the geotechnical officer in identifying potentially unstable roof and rib, and to document design methods that may be considered in the specificationofgroundsupport.Thereportisfocusedonthestabilityofthegroundintheimmediate vicinity of the roadway that is the immediate roof and rib. Specifically it should assist in answering thefollowingquestionswhichthegeotechnicalofficerroutinelyfaces: Whatfailuremodesarepossiblefortheoperation? Howcantheoperationanticipatethem? Howcanthepotentialfailuremodesbeidentifiedattheface? HowdoIspecifythesupportpatterninthegroundcontrolplan? HowdoImodifysupport,ifnecessary?

The report is unapologetically about the use of analytical methods: methods by which collapse modes are identified/inferred at the start of the process, driving forces then estimated, and support components specified in response. Others have developed empirical methods for coal mine roof

SofarinAustraliancoalminesalltherockfallsaregravitydriven.Thisreportdoesnotaddressgasoutbursts.

support for Australian underground coal mines and these form an invaluable complement to the analyticaltoolspresentedinthisreport. Some parts of the report may be considered to be controversial and there are some concepts/approaches presented that are yet to be fully validated. It is emphasized that the concepts are consistent with the available data and that there is no reason why they should not be widely applied. It is through this application that validation will be obtained and the appropriate factors of safety obtained. In the meantime, appropriate roof and rib stability will be obtained through the applicationofstandardengineeringdesignpracticetheuseofacombinationofindependentdesign tools.

1.1

THE CHALLENGE

Coal mining needs safe and fast roadway development. Rock falls present an obvious hazard to the workforce and, when large, to business continuity. The objective must be no rock falls, both from workplacesafetyandbusinessperspective. Currently,thesafetyhazardsincoalmineroadwaysareprimarilyassociatedwiththefallofscatfrom between the bolts2. Based on a survey of conditions in mines in the USA, Molinda (2003) suggests that the hazards are related to the presence of weak roof and unexpected discontinuities in the roof. There is an obligation to reduce the unexpected and to have support regimes that intrinsically addressthesupportoftheunexpected. Theinstallationofroofandribsupportiscurrentlymanuallyintensive,soitfollowsthatthenecessary levelofsafetymaynotbesimplyachievedbyinstallingmorebolts.Thereisaseverebusinessimpact of oversupporting the roof beyond the level needed to provide safety to the work force. New minesmaynotbeopened,orexistingonesclosed,ifthesupportisneedlesslyintense.

1.2

TARGET READERS

The report has been written for geotechnical engineers, geologists, and mining engineers charged with specifying ground support. Persons occupying the positions of mine manager or technical services manager should find the report of value. The report addresses some of the steps (site characterisation, model formulation, design, implementation, monitoring, and review) that can be found in Clause 48 of NSW regulations for coal, or the rockfall risk management of the Minerals CouncilofAustralia(PotvinandNedin,2003). It is emphasised that a full understanding of coal mine ground control is not yet available hopefully this report represents a step forward. One use of this report can be to challenge the standard

Thisreportdoesnotaddressthedesignofpillarsandthesequencingofcoalextraction.

interpretations of how a rock mass behaves around openings so that there can be a gradual improvementintheabilitytodesignsafeandproductivemineroadways. The concepts introduced can be applied atearly stages prefeasibility and feasibilityusing presumed valuesbasedonareviewofthegeologicalconditions.Astheprojectprogressestooperationalphase, the greater the importance to move from presumed to demonstrated values for the various input parameters.

1.3

STRUCTURE OF THE REPORT

Section2ofthereportdiscussesdesigninrockmasses,highlightingtheuncertaintiesthatareintrinsic to design in rock masses which are very difficult to fully characterise. The key message is that any designinrockmechanicswillneedtobeimplementedwithacommitmenttomonitorandrespondas thegeologicalconditionsarerevealed. Section3definesrockfallsandmakesadistinctionbetweenthenatureofthehazardsinmetalmines andcoalmines Section 4 summarises underground coal mining practices in Australia so that the practical constraints onroadwaygeometriesandexcavationmethodscanbeappreciated. Section 5 provides a reference guide to the engineering geology of coal measures and includes discussionsondiscontinuities,rockstrengthanddeformationproperties,andtheinsitustresses. Section 6 examines the redistribution of stresses around a retreat longwall panel and particularly in theimmediatevicinityofaroadway.Thekeypointisthatourknowledgeofstressabovechainpillars needs to be interrogated in more detail to understand the stresses that are present at the roof line. Simplenumericalanalysesforstressaboutopeningsinahomogeneousmaterialarepresented Section 7 presents the key strength properties of the popular support and reinforcement elements in useinAustraliaandpresentsmodelsfortheirreinforcingaction. Section 8 provides a flow chart for the specification of roof support and a number of analytical tools thatcanbeusedtodetermineboltingdensitiesandboltlengths. Section9providesaflowchartforribsupportdesign.

1.4

LIMITATONS AND OPPORTUNITIES FOR IMPROVEMENTS

Thisreportrepresentsaninitialefforttoformalizeananalyticalapproachtothedesignofroofandrib support:anapproachthathastendedtobeignoredIncomparisontosophisticatednumericaldesign. The report presents a number of simplified analyses of very complex mechanisms. In particular, the documented stress analyses are for homogenous materials to take advantage of the availability of some software and to demonstrate some basic concepts. More sophisticated numerical codes are

available for layered materials, and these could be incorporated within the proposed framework to improvethepredictions.

DESIGN IN ROCK MECHANICS

Tunnel design is different from many other engineering design processes. However, it can be performed on a scientific basis using an intimate blend of engineering geology, precedent, structural analysisandtheobservationalmethodduringconstruction(Pells,2002). The objectiveof any engineering design is that the constructed entity fulfils its intended function and in a way that is safely, economically, and environmentally acceptable. Engineering design needs to address both the stability and the deformations of the structure and the adjacent area, and at an acceptablecost. Because we are constructing something now that will perform in the future, all engineering design requires a component of prediction. This is particularly the case in geotechnical engineering because thepropertiesofthegeologicalmaterialsmustbeinferredfromacomparativelysmalldataset. Good predictions and the associated support designs are necessary but not necessarily sufficient. Performance assurance (both safety and business outcomes) comes from risk management of the design. Having said that, this report is only focused on the design/prediction component of the process. Predictions, which are forecasts of events in the future, are the essence of geotechnical engineering. Lambe (1973) provides a classification of predictions in geotechnical engineering (Table 1). Morganstern (2000) provides a framework for assessing the quality of predictions (Table 2) note thatachievinganoutcomewithin15%ofapredictionisconsideredtobegoodthisisareflectionof theuncertaintiesthatareinherentingeotechnicalpractice.
Table1Classificationofprediction(afterLambe,1973)

Predictiontype A B B1 C C1

Whenpredictionmade Beforeevent Duringevent Duringevent Afterevent Afterevent

Resultsattimeofprediction Notknown Known Notknown Known

Table2Predictionclasses(afterMorganstern,2000)

Category Excellent Good Fair Poor Bad 95105%(within5%)

Accuracy

8595%or105115%(within+/15%) 7585%or115125%(within25%) 5075%or125150%(within+/50%) <50%or>150%

Ideally, all predictions should be Class A and excellent; this is usually unattainable in practice. The goalshouldbetoachieveClassAasmuchaspossibleastheydorepresenthelowestcostandhighest level of intrinsic safety. However, in practice we must be satisfied with Class B predictions and within the excellent category from the safety perspective and within the good category from a business perspective. Class C are of no value and are best considered as calibrations. In the soil mechanics branch of geotechnical engineering, design competitions on Class A predictions reinforce theinabilitytoaccuratelypredict.Nosimilardataexistsinundergroundrockmechanics. The recognition of the inability to consistently produce good or excellent Class A predictions is the basis of the design process used in geotechnical engineering. The design process implicitly incorporatesmanyofthemorerecentconceptsofriskmanagement.

2.1

DESIGN PROCESS

Do the best prediction as possible, accept that ability to get 100% right is limited, and manage risks throughobservationalmethod. Incivilconstruction,thedesignprocessincludesthefollowingsteps: 1. 2. 3. 4. 5. 6. Designbrief Conceptdesign Detaileddesign Finaldesign, Designanddrawingsforconstruction Specification,Workmethodstatements,JobSafetyAnalyses,InspectionandTestplans,

OftenthereisanIndependentVerifiertoreviewthedetailedandfinaldesigns. Mining differs from civil engineering only in that the economic value is in what is removed, not what is left behind. Until recently, the mine owner was the operator so there were no contractual issues surroundingtheconstruction. In mining, the concept and detailed design are done in the prefeasibility and feasibility stages. The geotechnicaldesignshouldevolvealongwiththeJORCstages(Haile,2004).Theapproachintroduced in this report can be applied at early stages prefeasibility and feasibility using presumed values based on a review of the geological conditions. As the project progresses to operational phase, the greatertheimportancetomovefrompresumedvaluestodemonstrated. Forsteps4to6,theflowchartofBieniawski(1993)providesthestepsthatshouldbefollowed(Figure 1).TheseareverysimilartotherequirementofClause49ofthe1999regulations.

Limits on solution There is always a possibility that there may be practical limits to the design solution that relate to the mine layout and the machinery that are to be used. It is to be hoped that thefeasibilitystudywasvalidandthattheconstraintsarenotsevere. Collect data This stage refers to collecting and documenting the geological information. It is important to document any assumptions and simplifications made in the engineering geological model.Ifrockmassclassificationsystemsarequoted,theinputparametersshouldbestated. Identify collapse modes Domains should be defined based on depth, variations in the immediate roof lithologies, and the presence of faults. In longwall mines, there may be changes in the collapse modes related to changes in the stress field developed during the extraction phases. It is necessary definethesecollapsemodesastheyformthebasisoftheanalysesandhazardrecognitionlaterinthe process. Analysis In the context of this report, this step relates to analyses in excess of empirical rock mass classification systems. Where possible, a number of alternatives should be analysed. Adequate details ofcalculations shouldbe provided to allow it to beduplicated.All inputdata shouldbe listed, togetherwithrelevantoutput. Evaluation Identification of the preferred solution, possibly involving an assessment of the risks of implementing the candidate design. A comparison with precedent/practice and empirical systems should be presented. There should be documentation of the limits of acceptable deformations and designofmonitoringprograms. Implement This includes the drawing of support rules, preparation of TARPS, training in hazard identification,andaworkplaceriskassessment. Monitor hazards Proactively check geology assumptions, collapse modes, and confirm acceptable deformations.Reportsonfallsofgroundshouldbeinstandardiseddocumentedformat. All of this should be in the form of a written report. A complete record of the context of the design exercise, the data used, and the procedures followed makes review and extension of the design to similar work in the future much easier. The report should be written as part of the activity and referredtointhesupportrulesandTARPS(TriggeredActionResponsePlans).


ACTIONS Whocommissionedreport Context(prefeasibility,feasibility,operations) 2CONSTRAINTS Roadwaygeometry,machines,consumables 3GEOLOGY Seam geometry, roof and floor conditions (strength,bedding,joints),faults,stresses 4GEOTECHNICALMODELS Collapse modes, selection of design values, domainsofdifferentrockclasses 5ANALYSIS Support of detached blocks, reinforcement of discontinuities 6,7,8 SYNTHESIS, EVALUATION, Identification of preferred solution, comparison OPTIMIASATION with precedent/practice and empirical tools, Documentation of acceptable deformations and designofmonitoring 9IMPLEMENTATION Support rule, TARPS, training in hazard awareness, workplaceriskassessment. 10.MONITORING Check geology assumptions, confirm acceptable deformations,reportfallsofground.
Figure1Stepsingeotechnicaldesignprocesscoveredinthisreport

STAGE 1DEFINITION

NSWREGULATIONS 1f, 1a,1i 1c 1c 1c

1b, 1d, 1e,1h, 1k, 1l, 2ae, 1g,1j

2.2

ROCK AS AN ENGINEERING MATERIAL

Discontinuitiesfeaturesinarockmasswithzeroornegligibletensilestrength.

The defining characteristic of rock as an engineering material is the presence of discontinuities. Discontinuities include weak bedding planes, joints, cleats, faults, and shears. The term parting can beusedtodistinguishbeddingplanesfrombeddingtexturesthatareoftenseeninsedimentaryrocks. For the rockfall problem in coal mining, the scale of the roadway compared to the spacing of discontinuities is such that we are typically dealing with either intact rock or a few blocks defined by threesetsofdiscontinuitiesbeddingpartingsandtwosetofjoints(orthogonal). Asaconsequenceofthepresenceofdiscontinuities,rockmassesshouldalwaysbepresumedtohave zero tensile strength. A practical implication of this presumption is that a rock mass should never be put into tension. Brady and Brown (1985) state The important point is that a rock mass in compressionmaybehaveasastablecontinuum.Inadestressedstate,smallimposedorgravitational loadscancauselargedisplacementsofcomponentrockunits. In addition to the presence of discontinuities, there are a number of other inherent complexities in rock mechanics. Included in these are the influence of groundwater and the deterioration of freshly exposedrocktoweathering.Inthecontextofrockfalls,thereissignificantuncertaintyastohowrock fails under compression (Brady and Brown (1985). This issue is highlighted in the report whereby research reported in the early 1990s is used as the basis for assessing compressive failure around roadways. Recognising the role of discontinuities and the variability with different lithologies, rock masses are discontinuous, inhomogeneous, anisotropic, and not elastic (DIANNE3). DIANNE materials have not been, and possibly still are not, amenable to routine engineering analysis and hence routine design. Numerical codes such as UDEC and 3DEC allow the analysis of DIANNE materials but these are best consideredasspecialisedconsultingandresearchtools. Toenableroutinedesign,rockmassesareoftenascribedmodifiedpropertiesbasedonanassumption of equivalent continuous, homogeneous, isotropic, and linear elastic materials (CHILE). CHILE assumptions are required particularly when continuum numerical codes such as FLAC and Phase2 are used.

2.3

ANALYSIS TOOLS

Previous sections have discussed the integration of a number of steps into the design methodology. At the analysis stage, it is good engineering practice to use more than 1 design tool. The decision on
discontinuous presence of discontinuities (not a continuum), inhomogeneous not all the same rock type, anisotropic different properties in different directions, not elastic same increment or decrement of stress may produce a different incrementordecrementofdeformation(thisbasicallymeansthatrocksfail)
3

which tools to use depends to a large degree on the amount of data that is available. Lambe (1973) proposed that for the same amount of data, the accuracy of a prediction may decrease if an overly sophisticated method is used or if a simple method is employed when more data is available (Figure 2). In this figure, it is proposed that the method spectrum runs from rock mass ratings, through limit equilibriumanalyticaltools,to2dimensionalandthen3dimensionalnumericalcodes.

Figure2Accuracyofapredictionisafunctionoftheamountofdataavailableandthemethodused(afterLambe,1973).

2.3.1 PRECEDENT AND PRACTICE

There is no doubt that reference to precedent and current practice is an essential component to ground support design in ongoing operations. If there is confidence that the geotechnical conditions will remain the same, then it is reasonable to argue that a demonstrated successful support in immediatelyadjacentareascancontinuetobeapplied.Giventheinherentvariabilityofallgeological materials including coal measures, the maximum geographic extrapolation is perhaps in the order of onehundredmeters,andwillbelessifthegeologyisobservedtochange.

2.3.2 EMPIRICAL
Empirical:basedoractingonobservationandexperiment,notontheory. In the face of the complexity presented by DIANNE rock, classification schemes for rock masses have been developed in the past. In these schemes, numerical values are assigned to parameters that are considered to influence the behaviour of a rock mass and then these values are combined to give a single numerical rating value. These rating values are then used to interrogate databases of case histories. Two of the early and popular rating systems are the RMR system of Bieniawski (1976) and

10

theQsystemofBartonetal(1974).TheCoalMineRoofRating(CMRR,MarkandMolinda,2005)isa derivative of the RMR system that the authors propose to be more applicable to sedimentary rock masses. Theclassificationapproachisvalidiftheratingvaluecanbereliablyandrepeatedlyobtainedfromthe geological information, the mechanics of the system is well understood, and the data base includes case studies from similar geological and mining system. Attaining this goal may not be as simple as therockmassclassificationsystemswouldsuggest. Brady and Brown (1985) state Although the use of this approach is superficially attractive, it has a numberofseriousshortcomingsandmustbeusedonlywithextremecare.Theclassificationscheme does not always fully evaluate important aspects of a problem, so that if blindly applied without supporting analysis of the mechanics of the problem it can lead to disastrous results. This report argues that if the mechanics of the problem are understood, the simplicity of the coal mining geometryandmodernstressanalysistoolsmakestheempiricalapproachunnecessary. There are recommendations for ground support as a function of both RMR and Q values, particularly for metal mines and tunnels. Workers in horizontally bedded rocks have reported failures in the use ofRMRandQmethodstoprovideappropriategroundsupportrecommendations.(PellsandBertuzzi, 2007). For underground coal mines, the empirical relationships based on CMRR should be used, but once again with extreme care. The geomechanics of extended cut mining is substantially different from closeface bolting and this means that the relationship between the CMRR and the support designs will be different. It is inevitable that extended cutmining will have lower support densities for the samerating. Palmstrom and Broch (2006), in discussing rock mass classification systems and particularly the Q system, concluded that classification systems are useful for planning and less useful for the prescriptionofrocksupportduringconstruction.PellsandBertuzzi(2007)supportratingsystemsfor communicating rock conditions but not for the specification of support. Extending this observation further, rating systems are useful to incorporate consideration of rock conditions in the design of anotheraspectanoverallminelayout,orthedimensionsofachainpillar(Colwell,1998).

2.3.3 ANALYTICAL
Analyticaloutputisdependentonaleastoneinputvariableappliedtoalawofphysics The most accessible analytical tools are based on the limit equilibrium method, whereby a failure mode is identified, driving forces are estimated at the limit of stability, and resisting forces determined to provide equilibrium of forces with the application of a safety factor to account for inherentvariabilityanderror.Thesamecasehistorydatausedfordevelopingtheempiricalmethods is used to check the validity of method. Limit equilibrium methods require a smaller data base than theempiricalmethods,recognisingthatthedatabasemustbebasedonthesamefailuremechanism. Limit equilibrium methods are considered to be very suitable for the rockfall design problem as the objective is ultimately the prevention of gravity driven collapse of blocks of rock. This is particularly the case incoal mines as thegeometry of the block is comparatively simple, being rectangular prisms inmanycases.
11

While very well developed in soil mechanics, there are comparatively few limit equilibrium methods available in rock mechanics. This report presents a number of such methods for the underground rockfallproblemincoalmines. Experience and training is needed to correctly identify and anticipate the failure modes and this may alsorequiremoredetailedknowledgeoftherockmass.Thearithmeticinquantifyingthedrivingand resisting forces may be intimidating. Analytical tools do not have to be applied in isolation of numerical methods in fact it is often useful to use computer codes to estimate the driving stresses forcomplexgeometriesandthentoseparatelyapplythosestressestotheidentifiedfailuremodes. This report seeks to present a simplification of the arithmetic, allowing more focus on examining the rock mass and the failure modes that can develop. By having simple arithmetic tools, easier back analysisofobservedcollapseswillallowabetterunderstandingtodevelop. In many respects, this report develops the proposition of Brady and Brown (1985) that a rock mass traversed by one or two persistent structural features can be considered massive, openingup the abilitytoapplyarangeofsimpledesigntoolsinastructuredflowchart(Figure3).Inalatereditionof the book, Brady and Brown noted that the logical framework can also be applied to moderately jointedrock.

Figure3Logicalframeworkformineexcavationdesigninmassiverock(BradyandBrown,1985)

All factors of safety need to be developed and applied with care. The choice of an appropriate value dependsonconsiderationsrelatingtotheconfidenceinthematerialproperties,thesimplificationsof the behaviour models, and the assumptions made regarding various inputs. The best way to determinethevaluestouseindesignistobackanalysefailureconditionsandotherexperiencebases. Since many of the users of this report will have been exposed to the concept of a factor of safety in empirical pillar design it is important to highlight that there is a subtle but nonetheless important

12

difference. In empirical design methods where there has been a statistical analysis of a data base, factors of safety can be selected based on multiples of standard deviations to give probability of safety. Such a statistical approach is not possible in the early stages of the development of an analyticaldesigntool.

2.3.4 NUMERICAL
Numericaluseofcomputerstosolvecomplexstressredistributions In a numerical analysis, the key parts of the problem are simplified and a variety of computer programs areused to solve the equations of elasticity based on the imposed stress and the geometry of the openings. It is possible to extend the analysis into the nonlinear domain by implementing yielding. The engineer inputs the elastic deformation and failure properties of the materials being modeled and, in more sophisticated programs, the deformation parameters and strength parameters afterfailure(elastoplastic). The programs basically calculate stresses. Numerical models have the superficial attraction that they donotrequireknowledgeaboutthefailuremodes.However,thisisfarfromthecase.Theirabilityto identifyfailuredependstotallyonthedecisiononwhatcodetouse,thenatureofthesimplifications, and particularly the failure criteria and plasticity parameters that are decided upon. Our knowledge ofthedeformationandfailureofrockmasses(e.g.theCHILEassumption)andparticularlyabouthow rock masses deform after failure is extremely limited. The danger with numerical methods is that they can downgrade the importance of observations underground, replacing them with reliance on unverified assumptions on how rock behaves and simplifications made by the analyst remote from theoperationalproblem. The Geological Strength Index (GSI, Hoek and Brown, 1998) has been developed as the basis to modify the engineering properties of laboratory samples to give equivalent continuum values for the rock masses this formalizes the DIANNE to CHILE transition. This report highlights the fact that around the excavation boundary itself this may not be the appropriate strategy and that considerationofbrittlefailure(Martinetal,1991)ismoreappropriate. Numerical methods are the subject of intensive ongoing research. However, it has yet to be demonstrated that they are suitable for the design of ground support. Major simplifications of the real conditions are still required (best made by the geotechnical officer rather than by a numerical analyst)andjudgmentisneededinselectionofthesoftware(continuumordiscreteelementscodes). Continuum codes are unrivalled for the analysis of the farfield where multiple discontinuities such that the rock behaves as a homogenous material with reduced properties (Hoek and Brown, 1997) or inhomogenousrockmasseswithnodiscontinuities.Discretecodeswouldappeartobebetterforthe rockfallproblem,butsuchcodesarenotaswelldevelopedandareslowertorun. Given the overwhelming dominance of continuum codes (FLAC, Phase2) in the market place, there are a number of points that can assist in conducting and/or commissioning numerical analyses for rockfallinvestigations: Finite difference and finite element codes are just different ways of solving the equations of elasticity. With more complex models, the shorter run times of finite difference codes can havecommercialadvantages.
13

Acontinuumwithequivalentpropertiesmusthavezerotensilestrength. Beddingpartingsandjointsmusthaveacohesionvalueofveryclosetozero. To address the uncertainties in numerical analysis, plots of the distribution of stresses, separate fromthe distribution of displacements and failure zones, should be examined. Any model with zones of negative minimum principal stress but no corresponding failure should bechallenged. If bolts are explicitly modelled, plots of failures zones with and without bolts should be examined. Models that use plasticity are of particular concern as the state of the art in the plasticity of rockmassesispoorlydeveloped.

With the failure criteria available in continuum models, there is an ability to include transverse anisotropy(theimpactofbeddingpartingsviatheselectionofalowvalueforG,theshearmodulus) andthereistheabilitytomodelpostfailurebehaviourofisotropicmaterials;butthereisnovalidated model that allows both. There is a trend to seek to model transverse anisotropy and yield together through the use of strain softening ubiquitous joint (SU constitutive models) models in FLAC. These models contain algorithms for material behaviour, the parameters of which are claimed to be validated by independent material property tests. Zipf (2005) presents the state of the art in SU models.Table3presentsasummaryofthematerialpropertyassumptionsandacommentaryonthe possible impact and highlights the fact that there are a number of key parameters in the SU constitutivemodelforwhichthereislimitedifanyjustificationinthepublishedtechnicalliterature.
Table3CommentaryonrecentimplementationofSUconstitutivemodelsinFLAC

Parameter Cohesion

Frictionangle Dilation Tensile strength

CommonFLACassumption MohrCoulombassumed. Lab UCS values reduced by 0.56 before calculatingcohesion. Cohesion decreases to 10% over 5 millistrain postfailure 2136o,constantwithdeformation

Commentary Otherfailurecriteriaarepossible Basis for strength reduction may need to have thefrictionangleassumedtobezero No experimental support for the assumed rate ofcohesionloss. Lower friction angles for the sheared condition wellestablishedintheliterature 10o decreasing to 0o over 5 millistrain of No experimental or independent literature postfailureshearstrain supportforinitialorrateofloss Tensile strength ranges from about 10% of Zero tensile strength of rock should be UCS,tozeroover1millistrainpostfailure assumed with bedding discontinuities and joints

2.4

UNCERTAINTY, RISK, AND THE OBSERVATIONAL METHOD

Uncertaintyisacharacteristicofallgeotechnicalventuresandthemanagementoftheseuncertainties has been and continues to be an essential aspect of geotechnical engineering practice. The elegance orsimplicityofanydesigntooldoesnotreducetheuncertainty. Withuncertaintycomesrisk.Whilstthisreportseekstoreducetheseuncertaintiestosomedegree,it is essential to remember at all times that we are providing tools to better understand the uncertainties. Any perceived sophistication of the tools is more a reflection of the current limited alternativestodesignandnotnecessarilyamaterialreductionintheuncertainties.

14

Morgenstern(2000)discussesthesourcesofuncertaintyingeotechnicalengineering: model uncertainty this results from gaps in scientific knowledge such that simplified models need to be created to allow for subsequent engineering analysis. This is explicit in the analytical approach that requires the formulation of models of ground behaviour; importantlyitisalsoimplicitinempiricalandnumericalapproachesaswell. parameter uncertainty this results from the difficulties in ascribing values to the necessary input variables. Parameter uncertainty tends to increase with increasingly more complex models,forexamplewithFLACmodelsthatinvokeplastic(postfailure)propertiesofrock human uncertainty The quality of the workmanship applied to the analysis and also to the implementationmustalwaysbecontrolled.

In analytical design, the first and second uncertainties can be addressed to some extent through the selection of factors of safety. Judgment and experience is used to increase driving stress or reduce restrainingstressesinanalyticalmethods.Inempiricalmethods,thefactorsofsafetymaybederived fromregressionstatistics. In some respects, the observational method (Peck, 1969) evolved to address these uncertainties in soilengineeringandhassincebeenappliedtorockengineeringaswell.Itisstressedthatthismethod was developed primarily as a way of managing contractual risk and is not necessarily sufficient to managesafetyandotherbusinessrisks.

2.4.1 OBSERVATIONAL METHOD

Peck (1969) defines the observational method in the context of managing construction risk on major projects. The steps in the observational method as applied to the ground support problem in an undergroundcoalminewouldbe: Explorationtoestablishthegeneralnatureofmine. Assessment of most probable conditions and most unfavourable conceivable based on the knowngeology. Establishment of a design based on a working hypothesis of behaviour under the most probableconditions. Selection of variables to be observed and calculation of their anticipated values on the basis ofworkinghypothesis. Calculationofsameformostunfavourablecondition. Selection in advance of a course of action or modification of design for every foreseeable deviation,withparticularreferencetowhethertimeisavailabletoreact. Measurementofquantitiestobeobserved. Modificationofdesigntosuitactualconditions.

Someofthestepsaremadebythedesignengineersandsomearemadebyfaceworkers. It should be apparent that the observational method is not an alternative to basic engineering analysis and design. There is a need to have assessed the range of hazards and the implications if anythingtheobservationalmethodrequiresagreaterdesigneffort. It is important to note that monitoring (for example through the use of roof extensometers) is only one component of the observational method. In the complete application of the observational method, a full range of hazards will have been anticipated and candidate solutions already reviewed suchthatthereisconfidencethattheycanbeimplementedintimeandwithavailableequipment.
15

In the Australian coal industry, TARPS are used to give face workers direction on how to react. In developing TARPS it must be recognised that not all conditions may have been anticipated and that crews need to be given tools to allow them to determine if unanticipated geology has been encountered. There is also a danger that, in the absence of detailed design, the TARPS will require unnecessary additional support to be installed to the detriment of the business and without increasingsafety. Perversely, mining is often considered an ideal application of the observational method because of thelackofcontractualrestraints,althoughtheserestraintswerethefundamentalreasonforitsinitial development. Implicit in the application of the observational method is the availability of time and flexibility to change the support mining may have less flexibility in this regard and particularly mass mining techniques such as longwall. The method may not be successful if there are brittle failure modes where indications of collapse are very short, or where the result is an unacceptable delay (for examplecoalflowoffalongwall).

2.4.2 OTHER RISK TOOLS

Monitoring of the ground, involving inspections complemented with deformation monitoring as appropriate, is essential to the maintenance of a safe workplace. The greater the vigilance the more likelyitisthattheworkforcecanbeevacuatedfromanunsafeposition. By itself, monitoring is not sufficient to manage safety or other business risks associated with geologicalandgeotechnicaluncertainty. Workplace Risk And Consequence forums (WRAC) are not suitable for assessing design and business risks.Theyshouldbeusedtoassessthehazardsandriskofimplementingaspecificsupportrule. Of the risk tools that are currently available to the coal industry, those based on identification of failure modes are probablymore appropriate for business risks. Tools suchas FEMCA should provide theopportunitytoidentifytheweaknessesinthegeologicalandgeotechnicalmodels.

16

ROCK FALLS

This report is concerned with the rock fall hazard and specifically the way to prevent the gravity drivenfallofblocksofrock.Insomecasestheblocksofrockmaybedefinedbypreexistingsurfaces and the prevention of rock fall is by way of support. In other cases, the rock mass may be reinforced sothatsomeorallofthenecessarysurfacescanbepreventedfromforming(reinforcement). As will be seen later, rock falls do not necessarily mean elevated horizontal stresses are present in facttheycanoftenindicatetheoppositethelackofconfinement. Rock fall hazards can relate to local collapse of the strata (skin effects, scat) or more general collapse oftheroofspan.Bothhavetobemanaged.

3.1

DIFFERENCE BETWEEN COAL AND METAL MINES

The geometry of coal mine roadways is substantially different from that typically assumed for metal minesandtunnels.IntypicalAustraliancoalmines,thehorizontalaxisoftheroadwayiseffectivelyin the same plane as the dominant bedding discontinuity (Figure 4). Roadways are typically confined to the seam and the dips are typically less than 5o. One of the principal stresses is vertical and this meansthatondevelopment,theroadwayaxisisparalleltotheplaneoftwooftheprincipalstresses. On retreat, different stress directions may develop. This geometry should be compared to that presented as the basis for metal mining and tunnelling (Figure 5) whereby the excavation is equi dimensional, with an arched roof, and with joint sets and stress axes that are neither horizontal or vertical. In coal mining, the major rock fall hazards in development roadways relate to delamination along bedding in the roof, the slide of thin slabs of coal along steeply dipping joints in the sides, and the onset of failure of the coal in the sides. There are some important additional hazards that will be discussed later in the report related to the onset of compressive failure. In metal mines, the major rock fall hazard in the advancing roadways is the fall of wedges and in some circumstances (deep mines) the onset of compressive failure. All mines and tunnels are exposed to the problem of highly brokenrockwithinafaultzone. For longwall coal mines, the roadways at the maingate corner undergo increases in both horizontal and vertical stresses in the former and possibly vertical stress increases and horizontal stress decreasesinthecaseoftailgates.

17

Stressfield

Figure4Geometryofacoalmineroadwayanditscoaxialrelationshipwithdiscontinuitiesandstresses.

Figure5Typicalgeometryanddiscontinuityfieldforametalmineroadway(BradyandBrown,1985)

3.2

ROOF COLLAPSE

As an initial simplifying assumption, a coal mine roof can be considered as a jointed bedded beam made up of material with a finite compressive strength. Depending on the horizontal stress and vertical surcharge loading applied to the beam, the beam will collapse by either vertical shear along thejoints,beddingparallelshearingleadingtodelaminationandthesubsequentfailureofthinlayers, or the onset of compressive failure through the material causing fractures to develop from both ribsides until a block is defined above the installed support (Figure 6). In addition, there can be the local collapse of thin veneers of rock between zones where there is some tensile strength on the discontinuitiesorbetweenboltsoncetheyareinstalled(oftenreferredtoasscat). In the face of this simplification, there is a particularly serious collapse mode based on the presence ofjointsthatdipatlessthanabout65o70o(Figure7).Atthisorientation,anycompressivehorizontal
18

stresses can induce shearing along the dipping joints leading to collapse. Conversely, the lack of any compressivestressallowsshearingalongtheverticaljointdrivenbythedeadweightoftheblocksand anyverticalsurcharge. Examples of the types of roof collapses are shown in the following figures. Figure 8 shows a general roof collapse that resulted from delamination along bedding partings. Note the evidence of bedding parallel shear on the roof bolt still anchored in the roof. Figure 9 shows a local collapse by shear along vertical joints. In thiscase the roof material is thickly beddedcoal. Figure10 shows the role of meshpanelsincontrollingthehazardofrockfallsassociatedwithscat.

Low stress levels

High stress levels

Massive rock Bedded rock

Linear elastic response with little or Spalling and crushing initiates at points of high stress concentration at no rock failure. the roof /rib corners.

Bedded rock subjected to low horizontal stress in the roof line. Voussoir beam action develops and if the beams are too thin, compressive failure develops at the roof corners and joints open near the centreline.

Bedding results in higher induced stresses in the roof. Failure can develop higher into the roof than for massive rock.

Jointed rock

If roof stresses are very low, jointedbounded blocks may fall. Cantilever action in any blocks that remain insitu may result in crushing near the roof/rib corner.

Spalling and crushing initiates at points of high stress concentration at the roof /rib corners (similar to massive rock.

Figure6Generalcollapsemodesforabeddedroofwithsubverticaljoints

19


Figure7Dippingjointscanallowblockmovementwithorwithouthorizontalroofstresses

Figure8Collapsefollowingdelamination

Figure9Collapserelatedtoreleasealongverticaljoints

20

Figure10Thinslabsbetweenboltsretainedwithmeshpanels

The various collapse modes can be anticipated in the analysis phase and identified during mining operations (Table 4). An important point to note is that observations may not conclusively identify the collapse mode. In particular, it is noted that rib line guttering maydevelop in all general collapse modes,asitreflectsonlythelocalisedconcentrationofhighcompressivestressesandnotnecessarily elevatedinsitustresses.
Table4Observationsanddatarequirementsforroofcollapsemodes

Collapsemode Delamination

Compressivefailure

Observations Geologicalcontrols Rib line guttering, Spacing of bedding Centrelinecracking partings Shearstrengthofbedding partings Riblineguttering Unconfined compressive strength

Informationsources Lithologies Geotechnicallogging

Jointshear

Dippingjoints

Scat

Laboratorytests Geotechnicallogging Sonicvelocitylogs Open joints, joint Kinematically acceptable Joint orientation and boundedcollapse jointblocks spacing Riblineguttering Open joints, joint Joints dipping less than Joint orientation and spacing boundedcollapse 45plusfrictionangle/2 o Riblineguttering Joints striking within 20 ofroadwaytrend Loadingonmesh Closely spaced bedding Lithologies Thin rock slabs between partings straps

21

3.3

RIB COLLAPSE

Fromoneperspective,ribsaresimplyverticalrockslopesandhenceareexposedtothesamecollapse modesplanarslides,wedgeslides,andtopplingasseeninrockslopeengineering(Figure11).
Planar Wedge Topple

Figure11Ribcollapseanalogiestorockslopes(topples,planarslides,wedges)

In addition, there are additional collapse modes related to the vertical stresses that are applied to a coal rib and which are not present in a rock slope. Theses elevated stresses can induce buckling of jointboundedslabs(Figure12)and,inmassivecoal,theonsetofmininginducedfracturesthatdefine blocksthatsimplytoppleintotheroadway(Figure13).

Figure12Bucklingofacoalrib

22


Figure13Backsurfacetoaslabformedbyamininginducedfractureinacoalrib

3.4

APPROACHES TO PREVENT ROCK FALLS

Rock falls are gravitydriven collapse of blocks. There are therefore two approaches to prevent rock falls. Firstly, there is a support strategy whereby it is accepted that the block will form or is already present so that the approach is to support the block so that its gravity fall into the roadway is prevented. Secondly, there is the reinforcement approach based on preventing the block from forming in the first place, based on the recognition that blocks require a minimum of five surfaces before the kinematics allow the fall and that some of these surfaces may already exist in the form of bedding partings or joints. Referring back to the logical framework (Figure 3), support is required if compressive or tensile failure is induced, and reinforcement is required if there is a hazard of slip on discontinuities. Distinguishing between support and reinforcement can sometimes be an academic exercise, for example the case of shear translation of preexisting blocks along nonvertical surfaces. From a pragmatic designperspective, it may be preferable to take a step back from thecollapse concept and instead distinguish between structurallycontrolled failures and stress inducedfailures. Figure 14 is the flow chart proposed by Potvin and Nedin (2003) for metaliferous mines. It is apparent that Figures14and3presentverysimilarideas.

23


Figure14Flowchartfortheanalysisphasesforthepreventionofcollapseofground(PotvinandNedin,2003)

24

UNDERGROUND COAL MINING IN AUSTRALIA

Ifyoucantmineitwithouthurtingpeopleandyoucantmineitwithoutmakingaprofititisjust anotherblackrock

4.1

COAL SEAMS

In 2008, underground coal mining in Australia is being conducted in New South Wales and Queensland with one operation in Tasmania. Both coking and thermal coal are mined by undergroundmethods. For underground extraction, the coal seams are in excess of 1.8m thick, although thinner seams have beenconsideredwithalowerboundofabout1.2m1.3m.Themaximumseamthicknessisinexcess of 7m (depending on the definition of economic coal) and the maximum roadway height is 3.6m to 3.8m. In the thick seams, the better coal is most often found in the lower portions of the seam so a coalroofiscommon.Wheretheseamallowsit,coalfloorisoftenlefttoimprovetrafficabilityonlow strengthclayeyfloors. The dips are relatively flat (say less than 5o) and the depths of cover range from as low as 50m to in excessof550m. Thereareanumberofaccessestotheworkingseameithershafts,declines,orfinalhighwalls.

4.2

MINING METHODS

4.2.1 DEVELOPMENT 4.2.1.1 METHODS

All development mininguses continuous miners andhence the openings are rectangular with heights ranging from 1.5m to 3.8m, and widths ranging typically from 4.8m to 5.5m (an exemption by the NSWminingregulatorisrequiredforwidthsinexcessof5.5m). Presently, the most popular system for heading development in Australia is integrated cutting and bolting(bolterminers):therearesomeplacechangingsystem(cutandflit,extendedcutmining) Integrated cut and bolt systems are more popular in Australian mines and in part this may be a reflection of the higher frequency of faulting in our coal measures compared to seams in the USA where place changing is more common. The integrated system is efficiently implemented with two heading gate roads, and rectangular pillars with headings and cutthroughs at right angles. In the integrated system, roof bolting is carried out using minermounted hydraulicallyoperated units (Figure 15). Generally, the integrated system is used in conjunction with widehead continuous minersandthisenablestheinstallationofsupportintheroofatadistanceofabout2.5m3mbehind theface.Mostminesuseshuttlecarsforcoalclearancebehindthecontinuousminer.

25

The Place Changing methodinvolves using high capacity continuous miners alternating with purpose built high capacity multibolter roadway support machines (Figure 16). Narrow head continuous minersmostlysuitthesystem,forfrequentflittingoperations.Typicalmachineheadsare3.5mto4m wide.Thelengthofthecutistypicallyintheorderof8m10m. Thefallofeventhinslabsofroofcaninterruptthesystem,sotheplacechangingmethodrequiresthe ability for unbolted roof to stand for about up to four hours without either local or general collapse. Lithologies with closespaced bedding partings and jointed roof areas near faults may not have the requiredlevelofstanduptime.

Figure15Typicalbolterminer(SourceJoy)

Figure16Typicalplacechangeminerandmobilebolter(SourceJoy)

26

4.2.1.2 SUPPORT RULE DATABASE

A database of primary and secondary support installed in Australian longwall coal mines as of 2007 is discussed below. The database includes both the primary and secondary support installed in the gateroads. Data have been sourced primarily from site visits. An additional source of data has been obtained from ACARP project C15005 which tabulates the primary and secondary support rules for each of the mines in the development project database. In this database, the installed capacity is calculated as the number of bolts per square metre of roadway times the ultimate strength of the bolts(40tonnes/m2isequivalenttoabout6Xgradeboltspermetreofroadwayadvance). The installed capacity of the primary support for both the gateroads and intersections has been plotted as a function of depth in Figures 17 and 18 respectively. Stone roofs have been separated from coal roofs. The figures highlight thatdepth (as a proxy for stress) is not the sole determinant of roof support, with other issues dominating: these issues may include the strength of the rock, operationaldecisions to install support required for longwall retreat on advance, anddecisions based on precedent and practice. In Figure 17, the high density support (40 t/m2) at depths of 100m200m isfrommineswithverylowstrengthroofstrata.Alinearregressionlineisincludedfordiagrammatic purposes only. In general, there is no additional primary support routinely installed in the intersectionspriortolongwallretreat.
100 90

Primary Support Capcity (tonnes/m2)

80 70 60 50 40 30 20 10 0 0 100 200 300 Depth (m) 400 500 600

Stone roof Coal Roof Linear (Stone roof)

Figure17SummaryofInstalledCapacityofPrimarySupportforGateroads
100 90 Primary Support Capcaity (tonnes/m2) 80 70 60 50 40 30 20 10 0 0 100 200 300 Depth (m) 400 500 600 Stone roof Coal Roof

Figure18SummaryofInstalledCapacityofPrimarySupportforIntersections

27

4.2.2 EXTRACTION 4.2.2.1 LONGWALLS

The typical Australian longwall layout utilizes twoheading gateroads, mainly as a result of the choice to use bolter miners. Where the coal seam is gassy, the mine usually implements methane gas drainage programs, involving pre and post drainage systems of inseam and interseam drilling as well asdrillingfromthesurface.Multipleheadingshavebeenrequiredinsomeminestocontrolexposure oftheworkforcetoheat. Currently (2008) there are a total of 29 operation faces in Australia, producing around 79 million tonnes(Table5).Theaveragelongwallfacewidthis233mwithfacelengthsvarybetween150mand 400 m. The panel length runs between 520m and 4800m. The range of the extracted seam thickness variesbetween2.0mand4.5m(average3.44m).Thedepthofcovervariesbetween50mand500m.
Table5AustralianLongwallstatistics(afterCram,2008)

Mine

AngusPlace AppinWest Ashton Austar BaalBone Beltana Broadmeadow Crinium Bundoora Dendrobium Integra Grasstree Kestrel Mandalong Metropolitan MoranbahNorth NewlandsNorth Ravensworth Newstan NorthGoonyella OakeyCreekNo.1 OakyNorth Springvale Tahmoor Ulan United WamboNorth WestCliff WestWallsend

Longwall production (t) 3248500 1947600 2904800 1505600 1922300 7849500 3559600 4145400 1218000 3619000 2955900 3843000 4449000 4767300 1484000 4548000 4894900 1096800 2708100 2426300 6256500 5340800 3002400 1924900 3365100 3403100 1478100 3372700 830000

Seam

Lithgow Bulli Pikegully Greta Lithgow LowerWhybrow GoonyellaMiddle Lilyvale GermanCreek Wongawilli Liddell GermanCreek GermanCreek WestWallarah Bulli GoonyallaMiddle UpperNewlands PikesGully WestBorehole GoonyellaMiddle GermanCr GermanCr Lithgow Bulli Ulan WoodlandHill Wambo Bulli WestBorehole

Longwall facewidth (m) 250 250 205 150,220 240 264 200 270 225 236 246 300 250 150 154 300 300 250 220 300 300,200 250 305 275 400 200 250 300 165

Extracted thickness(m) 2.9 3.05 2.35 5.75 2.25 3 4.25 3.4 2.8 3.5 2.55 2.65 3.15 4.2 3.2 4.15 4.5 2.35 3.4 4.2 2.5 3.05 3 2.0 3.05 3.1 2.1 2.65 4.8

Panel1length (m) 2997 1400,520 1700,1830 1500,1300 1500,1000 3x3300 2x2100 2494,2400 2000 1955,1615 2213,2331 2400,2400 3186,3579 3020,2766 1167,1208 1900,2500 4800 1500,1800 3090 3000 3338,3489 3500 3443,3443 2180,988 3018,1200 3300,3120 3650 3227,3500 1667,1208

4.2.2.2 SECONDARY SUPPORT

Unliketheprimarysupport,whichisrelativelyconsistentalongthelengthofasinglelongwallblock,a wide range of maingate secondary support patterns are used. For example different patterns are

28

used for intersections, gateroads, structured zones etc. The primary support capacity has been included in these figures to capture the variation between gateroads and intersections e.g. many minesinstalla4or6boltprimarysupportwithnosecondarysupportmidpillar.Inordertocompare support capacities both minimum and maximum values have been plotted in Figures 19 and 20 the calculations assume a 50 tonne capacity cable. Similar to primary support, there is not a strong relationshipwithdepth.

80 Combined Primary and Minimum Secondary Support 2 Capacity (tonnes/m ) 70 60 50 40 30 20 10 0 0 50 100 150 200 250 300 Depth (m) 350 400 450 500 550 600

Stone roof Coal Roof

Figure19Summaryofcombinedminimummaingateprimaryandsecondarysupport

80 Combined Primary and Maximum Secondary Support 2 Capacity (tonnes/m ) 70 60 50 40 30 20 10 0 0 50 100 150 200

Stone roof Coal Roof

250

300 Depth (m)

350

400

450

500

550

600

Figure20Summaryofcombinedmaximummaingateprimaryandsecondarysupport

29

The minimum and maximumsecondary support only is plotted inFigures 21 and22.Overall, there is no relationship betweendepth of coverand secondary support intensity. There aremany mines that donotinstallanyroutinesecondarysupport(Figure21).Mostlongwallminesinstallbetween8to16 tonnes/m2(Figure22).Cablelengthrangesfrom4.1mto8.1m.
40 Minimum Secondary Support Capacity (tonnes/m2) 35 30 25 20 15 10 5 0 0 100 200 300 Depth (metres) 400 500 600
Stone roof Coal Roof

Figure21Summaryofminimummaingatesecondarysupport

40 36 Maximum Secondary Support Capacity 2 (tonnes/m ) 32 Stone roof Coal Roof

4 Number of 50 tonne cables /metre of roadway

3 28 24 20 16 12 1 8 4 0 0 100 200 300 Depth (metres) 400 500 0 600 2

Figure22Summaryofmaximummaingatesecondarysupport

30

ENGINEERING GEOLOGY

Thegeotechnicalengineershouldapplytheoryandexperimentationbuttemperthembyputtingthem into the context of the uncertainty of nature. Judgment enters through engineering geology. (Terzaghi,quotedinPalmstromandBroch,2006) The report has already highlighted the defining characteristic of rocks the discontinuities. In coal measure rocks, the main discontinuities are joints and bedding partings. Bedding textures such as cross bedding are not necessarily discontinuities. This section starts with a discussion of discontinuities,thenmovesontostrengthanddeformationpropertiesandtheninsitustresses. It is emphasised that the report is concerned with roof and rib support, whereby the scale is set by the roadway width say less than 6m. Discontinuities with spacings less than this may impact rock performance.

5.1

DISCONTINUITIES

Rock mass performance is controlledby discontinuities defined as features in a rockmass with zero or negligible tensile strength. Joints and bedding partings are the major discontinuities in coal measurerocks.Thereisnothingfundamentallydifferentabouttheengineeringpropertiesofbedding partings compared to joints they are both discontinuities albeit with different roughness and persistenceandtheonlydefiningcharacteristicisthedipofthesurfaces. The reader should be aware that this definition of a bedding parting as a discontinuity is consistent with the RMR system of Bieniawski (1976), the Q system of Barton et al (1974) and the ISRM Commission (1978) but conflicts with the use of the same term in the Coal Mine Roof Rating (CMRR) system (Molinda and Mark, 1994) where discontinuities are allocated a cohesive strength rating. A discontinuity with negligible tensile strength will also have negligible true cohesion when tested in shear. A discontinuity may still have an apparent cohesion intercept related to the roughness of the discontinuity surface (Figure 23), however this apparent cohesion is not the same as true cohesion related to the cementation of particles. The difference in definition is particularly stark in the more recentrecommendationsontheCMRRthatpromotetheuseofanindextodiametricpointloadtests insteadoftheindextodiscontinuityspacinginsomecircumstances(MarkandMolinda,2005).

31

Shearstress

N Truecohesion Apparent cohesion roughnessanglei NormalstressN

Figure23Differencebetweentruecohesionofanintactsampleandapparentcohesionofaroughdiscontinuity

5.1.1 DESCRIPTION OF DISCONTINUITIES

The description of discontinuities is discussed by the ISRM Commission (1978), and also included in thefieldgeologisthandbook(AusIMM,1995).Aspectscoveredincludeorientation(dip,dipdirection, strike),spacing,RQD,persistence,roughness,aperture,andfilling.
Table6Classificationofdiscontinuityspacing

Term Extremelyclose Veryclose Close Moderate Wide Verywide Extremelywide

Spacing(mm) <20 2060 60222 200600 6002000 20006000 >6000

Table7Classificationofdiscontinuitypersistence

Term Verylow Low Medium High Veryhigh <1 13

Length(m)

310 1020 >20

Thespacingofbeddingpartingscanbeexpectedtovaryovershortdistancesreflectingthechangesin lithology. Shales may have extremely close partings, and sandstones and conglomerates may have extremelywide spaced partings. In general, bedding partings will have high to very high persistence, andjointpersistencemayrangefromverylowtoveryhigh

32

Joints and bedding are not perfectly flat planes. On the relevant scale for a coal mine roadway, the surfaces may be planar, undulating or stepped. The Joint Roughness Coefficient (Barton et al, 1974) can be used to quantify this shape. The Joint Roughness Coefficient (JRC) can be estimated as per Figure 24. In determining the JRC value, not only is the profile of the discontinuity surface important butalsolengthoftheprofilemeasured,asshowninFigure25thisfigureprovidesaconvenientway ofdeterminingtheJRCinamineroof.

Figure24JointRoughnessProfilesandCorrespondingJRCValues(BartonandChoubey,1977)

Onthesmallerscalethesurfacemaybeclassifiedasrough,smooth,orslickensided. roughsandpaper smoothtabletop slickensidedpolished

33


Figure25AlternativemethodforestimatingJRCfrommeasurementsofsurfaceroughnessamplitudefromastraightedge (BartonandChoubey,1977)

5.1.2 BEDDING PARTINGS

Bedding discontinuities are often referred to as partings, this terminology has the advantage of emphasisingthedifferencefrombeddingtextures.Persistenceofbeddingpartingsistypicallyhighor greater. Spacing will vary within the same lithology and in response to changes in lithology. There is no easy way of assessing the spacing of partings, and certainly not from the current generation of geophysical logs. Core drilling is required. Some inference can be made from knowledge of the lithology,butthelackofprecisionmakesthisapproachunsuitablefordesign. The results of JRC measurements on bedding partings during this project are presented in Table 8. TheaverageJRCforpartingswas9,withaminimumof3.
34

Table8Roughnessofbeddingpartings

SampleType BeddingPlane BeddingPlane BeddingPlane BeddingPlane BeddingPlane BeddingPlane BeddingPlane BeddingPlane BeddingPlane Joint BeddingPlane BeddingPlane BeddingPlane BeddingPlane Siltstone,coalyfossils

Lithology 12 4 6 3 5 7 9 11 8 5 20 5 20 12

JRC (10cminterval)

Mudstone,coalylaminations Siltysandstone,minorcoalywisps Siltstone, Finegrainedsandstone/siltstone,micaceous Finegrainedsandstone,coalywisps Finegrainedsandstone,coalyfossils Mudstone Siltstone Finegrainedsandstone Siltstone Finetomediumgrainedsandstone,carbonaceouslaminae Siltstone,coalyfossils Finegrainedsandstone/siltstone,coalyfossils

5.1.3 JOINTS IN STONE

There is a very strong trend in sedimentary sequences, and particularly in coal measures, for the developmentof2jointsetsatrightanglestoeachotherandorthogonaltobedding(Figure26).

Figure26Exampleofjointpatternsinsedimentarystrata.

Hobbs (1967) was one of the first to discuss the possible mechanics behind the observed linear relationship between bed thickness and the spacing of joints (Figure 27). In relatively undeformed strata,thespacingofjointstendstofollowalognormaldistributionwiththeratioofthemedianjoint spacing to bedding spacing being approximately 1.0 (Figure 28). The implication of Figures 27 and 28 isthatasafirstapproximation,thestoneroofofcoalmineroadwaysiscomposedofcubesofrock.

35

 Figure27Jointsinundeformedsediments(Fookesetal,2000)

36


Figure28Relationshipbetweenjointspacingandbeddingthickness(Jietal,1998)

Thedataintheliteraturearelimitedtobedswiththicknessesoflessthanabout3m.Observationsby the authors of highwalls and cliff faces suggest that the general trend may be present to at least about20mthickness. Determining the orientation of joints during exploration programs is now a relatively easy task through the use of the acoustic scanner, and in many cases in the Bowen Basin from immediately adjacent highwalls. General trends can also be inferred from a knowledge of the regional geology, in particular the orientation of fold axes (Figure 29). Different joint orientations will be present in the vicinityoffaultsordykes.

5.1.4 JOINTS IN COAL CLEATS

Inthisreport,thetermscleatandcoaljointsareusedinterchangeably.Inapuresense,cleatrefersto the small scale (very low persistence, extremely close spacing) fractures, particularly within bright coal (nonpersistent cleat, Figure 30). From a geotechnical perspective, there is more interest in the features with greater persistence. The latter are more correctly called joints, but common usage in theminesoftenreferstothemascleatsaswell.

37


Figure29Associationbetweenjointsandfaults/folds(Fookesetal2000)

38

Figure30Jointedandcleatedcoalnoteverylowpersistenceofcleat(MedhurstandBrown,1998)

Jemeric (1985) presents aclassification system for cleats and highlights that they mayhave a number ofdifferentorigins. EndogenousCleavagemainlyorientedperpendiculartothebeddingplanes.Relatedtothe drying and shrinkage of organic material associated with its compaction and release of volatilematter.Endogenouscleattypicallyhaslowpersistence. Exogenic Cleavage formed by external forces related to tectonic events. This cleat would tend to be aligned in the direction of the major principal stress active at the time of formation.Thisistheprobablytheoriginofthepersistentcleatorcoaljoints.Thisdefinition may explain how the orientation of coal joints may vary either side of a thrust fault (Figure 31).

It could be expected that cleat sets in coal and joints in stone may be parallel. Information from the Bowen Basin suggests that this may be generally true (Figure 32), but that there can be important deviations. Since bedding surfaces can be the locus for major tectonic horizontal movements, it is reasonable to anticipate rotations as well as translations along such surfaces. The result would be differentdiscontinuitysetsinthecoalcomparedtothestone.

Figure31Anexampleofrotationofcoaljointsaboutthrustfaults

39


Figure32CleatandjointorientationsinBowenBasin(Pattison,1995)

The results of initial testing into the roughness of coal joints indicates that the JRC is greater than for beddingpartings(Table9).
Table9Jointroughnesscoefficient(JRC)ofcoaljoints

SampleType Cleat Cleat Cleat Cleat Cleat Cleat Cleat Cleat Cleat Cleat Cleat Cleat Cleat

DirectionofTestingwith respecttobedding Perpendicular Parallel Parallel Parallel Parallel Parallel Parallel Parallel Perpendicular Parallel Parallel Parallel Parallel 9 3 12 4 4 4 12 4 14 4 12 8 8

JRC (10cminterval)

5.1.5 FAULTS
The eastern Australian coal fields have had a complex tectonic history, with the result being the overprintingofseveralfaultregimes.Thestateoftheartissuchthatonaminescaleitisnotpossible to predict in advance the full suite of fault styles and directions that may be encountered. Close spaced drilling and seismic surveys can provide information on the major faults in a block of coal but

40

are limited in their ability to resolve small scale faults, fault zones with little net throw, or bedding parallel faults. The application of drilling and reflection seismics is to identify ground that should not be mined, but sesimics in particular has limited application in providing geotechnical details of the groundthatistobemined. Oncetheoverallfaultregimeisknown,itispossibletoanticipatethejointingpatterns(Figures27and 29)anditisthejointingpatternthatisofconcernfromagroundcontrolperspective.Thereislikely to be an increase in the density of bedding partings in proximity to faults it would appear that the tectonicstressestendtoconvertbeddingtexturestopartings. It is highlighted that faults are characterised by broken ground and this implies that the ground has a less ability to accommodate high deviatoric stresses. It is more likely that stress magnitudes will be locallyreducedinthevicinityoffaults.Infact,thepoorergroundconditionsinproximitytofaultsare morelikelytoberelatedtothehigherdensityofjointsandpartingsdefiningpreexistingsmallblocks. Based on the authors experience, the following should be anticipated for the 3 typical fault types encountered: Normalfaults 70o90odip. Gougelessthan20mmwide. Paralleljointsfor5meitherside. Increasedbeddingpartingintensity10meitherside. Increasedpermeability.

Thrustfaults 15o40odip. Gougezoneof0.5mto1.5minwidth. Verycloseandclosespacedshearsandslickensides. Undulatingsurfacesdefininglozengeshapedblocksrangingfrom10cmupwards. Increasedpermeability.

Beddingparallelfaults 10mm250mmthick. Veryclosespacedslickensides.

5.2

LITHOLOGIES

Knowledgeofthelithologiesintheroofisimportantasitgivesthekeypieceofinformationonwhich to anticipate the spacing of bedding partings. There is no relationship between lithology and rock strength. Coalmeasureshavebeendepositedinarangeoffluviatile(Figure33)ordeltaicenvironments(Figure 34).Thefluviatilesystemcanincludemeanderingriversdepositionalenvironmentssuchthattheroof of coal seams can range from thick mudstones (flood plain) through interbedded sandstones and mudstones (crevasse splays and levees) to crossbedded and planarbedded sandstones (channel

41

sands). Of these, the crevasse splay and levee environments tend to produce the closerspaced and laterallypersistentbeddingpartingse.glaminitesintheSouthernCoalfield.Thespacingofbedding partingsinthechannelsandscanbehighlyvariable.

Figure33Distributionoflithologiesinfluviatilesystem(Fookesetal,2000)

42


Figure34Distributionoflithologiesinadeltaicsystem(Fookesetal,2000)

The braided channel systems tend to have coarser grained sediments in thicker beds. The cross bedding tends not to develop as a parting, so verywide and extremelywide spaced bedding partings areoftenpresent.Therooftypesassociatedwithdeltaicsystemstendtohavecloserspacedpartings whencomparedtofluviatilesystems.

43

An important aspect of Figures 33 and 34 is the scale. Lateral variations in lithologies can be on the scaleoftensofmetres.Theimplicationofthisisthatevendrillholespacingsof100mcannotbeused topredictthelithologiesthatwilldevelopaboveeveryroadway.

5.3

STRENGTH

Depending on the scale and geometry of the excavation, the strength of the rock mass around the excavation may be controlled by either the discontinuities or by the properties of the rock substance betweenthem. Strength is an all encompassing term, and covers the unconfined strength, confined strength and tensilestrengthofintactsamplesandtheshearstrengthalongdiscontinuities.

5.3.1 UNIAXIAL COMPRESSIVE STRENGTH 5.3.1.1 STONE

For coal measure rocks, the unconfined compressive strength (UCS) of the roof and floor stone as measuredinthelaboratoryrangesfromaslowas5MPa10MPatoinexcessof100MPa(Figure35) thereisnorelationshipbetweenlithologyandrockstrength.Igneousmaterials,suchasdykesorsills, may be less than 5 MPa (for example weathered or chilled margins) or in excess of 300 MPa (fresh dolerite).

Figure35Plotshowingtherangeinstrengthanddeformationmodulusfordifferentlithologies(ShepherdandGale,1982)

44

Due to the time and cost in testing core, Australian coal mines often utilise a sitespecific sonic velocity to UCS correlation. Core is selected for laboratory testing from drillholes in which sonic velocity logs have been run. Figure 36 shows a range of relationships with examples from both QLD and NSW minesites and also the McNally (1987) formula for stone. The figure highlights the variability in the relationships and suggests that there are other factors besides rock strength that controlthesonicvelocity.Otherrelationshipsincludeconsiderationofdensitybuteventhesedonot coverallfactors. From a ground control perspective, it is the low strength rocks that are of particular interest. The difference in the various relationships at low strengths is highlighted inFigure 37 where 10MPa has a range of 1300m/s in its associated sonic velocity. It is generally accepted that the original McNally database included many samples that hadbeen allowed to dry out before testing withthe result that anomalously high strength values were obtained (particularly for the lower strength range). The continueduseoftheMcNallyequationisnotrecommended.
140 Mine 1 Mine 3 Mine 5 Mine 7 Mine 9 Mine 2 Mine 4 Mine 6 Mine 8 McNally (stone)

120

100

UCS (MPa)

80

60

40

20

0 2000

2500

3000

3500

4000

4500

Average Sonic Transit Time (metres/second)

Figure36AvarietyofUCSsonicrelationshipsforstone
30

25

Mine 1 Mine 3 Mine 5 Mine 7 Mine 9

Mine 2 Mine 4 Mine 6 Mine 8 McNally (stone)

20 UCS (MPa)

15

10

0 2000

2100

2200

2300

2400

2500

2600

2700

2800

2900

3000

Average Sonic Transit Time (metres/second)

Figure37UCSsonicrelationshipsforstoneatlowerstrength

45

There are a number of practical constraints to the accuracy of any strength/sonic correlation. These include matching the depth of the samples with the sonic log, allocating a sonic velocity value, preserving the water content prior to testing (especially important for rocks less than about 20MPa), sampling and testing the lower strength rocks, the reproducibility of laboratory testing, and the limitationsoftheregressionanalyses(typicallylimitedtostandardspreadsheetfunctions). Theconclusiontobedrawnfromthisdiscussionisthatitisgoodpracticetouseexistingrelationships onlywithgreatcare,andtodeveloparelationshipforeachmineusingthehighestlevelofcontrolon the sampling and testing procedures. Given the greater significance of low strength materials, it is recommended that the relationship is created from a data base biased to this end of the spectrum. Having said that, a relationship that the authors have found to be reasonably valid across a range of coalfieldsis: UCS(MPa)=5785e(17374/vel) wherevel=sonicvelocityinm/s. The use of Point Load Strength Index testing (PLSI) to estimate UCS needs to be considered carefully. ThePLSIcanberelatedtotherockstrengthusingsimilarstatisticalmethodsasforsonicvelocity.The test suffers from a high degree of variabilityand a largenumber of data points are required to obtain a reliable estimate of the UCS. It is known from civil engineering that the reliability of the PLSI estimateforvalueslessthanabout20MPaisverylow.ThesonicvelocityapproachtoestimatingUCS ispreferred. For very low strength range, and for situations where the thickness of the band does not allow laboratory testing, good estimates of strength can be obtained from standard visual and tactile tests (Table10).
Table10Visualandtactiletests(ISRM,1978)

Equation1

Grade RO R1 R2 R3 R4 R5 R6

Term Extremelyweak Veryweak Weak Mediumstrong Strong Verystrong Extremelystrong

UCS (MPa) 0.251 15 525 2550 50100 100250 >250

Fieldobservations Indentwithfingernail Peel with pocket knife, crumbles under firm blows with point of geologicalhammer Peeled with difficulty with pocket knife, shallow indentation withfirmblowswithpointofgeologicalhammer Cannot scrap or peel with knife, specimen fractured with single blowfromthebluntendofageologicalhammer Morethan1blowfromthebluntendofageologicalhammer Manyblowsfromthebluntendofageologicalhammer Canonlychipthesamplewithageologicalhammer

5.3.1.2 COAL
Becauseofits cleatandjointing,theUCSofcoalisarelativelydifficultparametertodetermineinthe laboratory. A compilation of the results of testing coal samples, mainly by ACIRL (1986), is presented in Table 11 where it can be seen to range between 10MPa for Middle Goonyella Seam (coking) and 35MPa Great Northern Seam (thermal). Medhurst and Brown (1998) suggest that the UCS of coal is related to the brightness of the coal and will range from 35 MPa for dull coal to 9.7MPa for high brightnesscoalsthisisconsistentwiththedatainTable11.

46


Table11Rangeofcoalunconfinedcompressivestrengths(ACIRL,1986)

Seam HarrowCreek Bulli Tongarra Katoomba MtArthur Bulli Wallarah MouraDSeam Lithgow Wongawilli VictoriaTunnel GreatNorthern MiddleGoonyella Newlandsbrightbottoms GermanCreek(Southern)

UCS(MPa) 8.7 28.9 17.9 33.1 21.9 20.2 27.4 11.4 26.4 16.0 28.0 35.6 10 14 10

Modulus(GPa) 1525 4675 3403 4013 3146 3191 3838 2488 4096 3080 4458 4560

Adifferentsonicvelocity/strengthrelationshipexistsforcoal(Figure38).
60 50 40 30 20 10 0 2000 -10 Sonic velocity (m/sec) y = 0.07x - 148.57 R2 = 0.5178

Laboratory UCS (MPa)

2100

2200

2300

2400

2500

2600

2700

2800

Figure38UCSsonicvelocityrelationshipforcoal

5.3.2 TENSILE STRENGTH

It cannot be stressed enough that the presence of discontinuities in a rock or coal mass means that thetensilestrengthofthemassiszero. Thetensilestrengthofanintactrockcanbeassumedtobeintheorderof10%oftheUCS.

47

5.3.3 TRIAXIAL PARAMETERS

Increased confinement of samples results in an increase in the stress that will cause failure. Conventionally,therelationshipbetweenfailurestressandconfiningstress(Figure39)issimplifiedto astraightline(MohrCoulombfailurecriterion)orasacurve(HoekBrownfailurecriterion).
Figure39Strengthenvelopes(dashMohrCoulomb,solidHoekandBrown)

FortheMohrCoulombcriterion, 1=c+3tan Equation2

where 1 and3 are the maximum and minimum effective stresses4 at failure and c is the uniaxial compressive strength, and the relationship between uniaxial compressive strength, cohesion(c),andfrictionangle()is: c=2ccos/(1sin),or c=c(1sin)/(2*cos),and tan=(1sin)/(1sin),or sin=(tan1)/(tan+1). FortheHoekBrowncriterion, 1=3+c(mc3+s)a Equation3

effectivestress=totalstressporepressure.Coalseamsareaquifersandhavesignificantporepressurespriortomining.At the mining face the pore pressures are zero so it follows that there have been significant effective stress changes ahead of mining.

48

where m is a parameter related to lithology (somewhat analogous to friction angle), a = 0.5 for rock and 0.65 for coal, and s= 1.0 for intact rock. As will be discussed, m and s can be variedtoreflectrockmassstrength. The c parameter is calculated from the curve fitting and may not have the same value in the two criteriaandmaynothavethesamevalueastheUCSvalue. Asagoodapproximation,thefrictionangleandmvaluescanberelatedtolithology(Table12)
Table12Presumedfrictionanglesandmvalues

Lithology Siltstones,finesandstones Mudstone Mediumandcoarsesandstones Claystones Frictionangle 30 25 35 20 mi 10 7 15 4

5.3.4 BRITTLE STRENGTH

RecentresearchsuggeststhatneithertheMohrCoulombnortheHoekBrowncriterionissuitablefor the behaviour an excavation (Kaiser and Kim, 2008). It appears that cohesion and friction are not mobilised at the same time, and that initial rock failure relates only to a loss of cohesion. One of the earlypapersonbrittlefailure(Martinetal,1999)concentratesonmassiveigneousrocks,butacloser examination reveals that one of the early case studies involved the DonkinMorien tunnel in the coal measuresoftheSydneyBasinofNovaScotia.Theresearchprovidesthetechnicalexplanationforthe early observations that around openings the mass strength appears to be 0.5 the laboratory measuredstrength. This concept can be incorporated in a failure criterion either as13 =c/3 toc /2 ( =0.0), or m = 0.0, and s =0.11 to 0.25 (Figure 40). There is no tensile cutoff with the Hoek Brown criterion so, if tensileroofstressesarepossible,itmaybepreferabletousetheMohrColoumbcriterionwithazero tensilestrengthcutoff.

Figure40Brittlerockparameters

49

This brittle rock concept is applicable to estimating the zone of fracturing about excavations and henceisparticularlyrelevanttothedesignofgroundsupportsystemsthesubjectofthisreport.For coalmeasurerocks,s=0.11hasbeenfoundtobeappropriate(seelater). Full implementation of the brittle rock concept also requires the definition of the spalling limit. As a concept,thespallinglimitisnotwelldefinedanditneedstobedeterminedempirically(Diederichset al, 2004). The height of roof falls on development suggests that the spalling limit for coal measure rocksmaybeintheorderof35assumingisotropicdeformationparameters.Thisislowerthanthat inferred in the work on igneous and metamorphic rocks but may be an artifact of the assumption of isotropic deformation properties. The inferred spalling limit increases to a value of about 10 if the independent shear modulus (G) is reduced to simulate the impact of bedding anisotropy. For coal in theribs,thespallinglimitappearstobegreaterabout10(seelater).

5.3.5 SHEAR STRENGTH OF DISCONTINUITIES

TheshearstrengthofadiscontinuitycanbeexpressedintermsoftheMohrCoulombcriterion:

=c+ntan

Equation4

where = the shear strength, c= cohesion andn = the normal stress on the surface (Figure 41). For perfectly flat surfaces, the cohesion is equal to zero. As discussed above, real world discontinuities are rough and uneven and the effect of this can be considered either in terms of an effectivecohesionorintermsofanapparentlyhighfrictionangleatlownormalstresses. c +i

n
Figure41Strengthenvelopeforadiscontinuity

TheBartonBandisshearstrengthcriterionfordiscontinuitiesincorporatesthecontributionofsurface roughness(Barton,1973):

=ntan[JRClog10(JCS/n)+]
whereJCS=JointCompressiveStrength.

Equation5

50

ForbeddingpartingswithaJRCvalueof3inamudstoneof40MPa,thisrelationshipgivesanincrease of6otothefrictionangleforanormalstressof300kPa.

5.3.6 MASS STRENGTH

The strength of the rock mass when considered as a continuum is of limited value to the rockfall problem.Thissectionisincludedtogiveanintroductiontothetechnicalliterature. Bydefinition,thetensilestrengthofanyrockmassiszero.

5.3.6.1 STONE
The Geological Strength Index (GSI), Hoek and Brown (1997) provides the latest and most comprehensive approach to determining the equivalent continuum properties of a rock mass when the scale being considered results in more than 2 or 3 sets of discontinuities. The GSI is an extension of the Rock Mass Rating system of Bieniawski (1976) and can be estimated readily from a number of charts. Of the available charts, the one created for molasse (Hoek et al, 2004) is perhaps the most applicable to coal measures (Figure 42). The surface conditions for coal measures are best described as fair reflecting the smooth nature of the bedding partings. The suggestion by Hoek at al is that quotingtothenearest5pointsisappropriate.AGSIofabout50shouldbetypicalforcoalmeasures.

Figure42Geologicalstrengthindexformolasses

TheGSIisusedtomodifythemandsparametersintheHoekBrowncriterion:
51

mb=miexp((GSI1000/(2414D), sb=exp((GSI100)/(93D)),and a=0.5+1/6*(exp(GSI/15)exp(20/3)) where D= damage index. For coal mine roadways excavated with continuous miners, D can be taken as0.0. Brown(2008)suggeststhattheGSIreductionsshouldnotbeappliedforGSIvaluesgreaterthan70or lessthan30,notforrockswithaUCSlessthan15MPa.

5.3.6.2 COAL
MedhurstandBrown(1998)andMedhurst(1999)provideamethodforestimatingthemassstrength of coal based on the Hoek and Brown failure criterion (Table 13). The m and s values are related to the vitrinite reflectance values, seam brightness logging, and thec value is 35 MPa. For coal, the a valuewasfoundtobe0.65.
Table13HoekBrownparametersforvariouscoals

Vitrinite reflectance 0.91.1 1.11.30 1.31.5

mi 2016 1612 1210

mb 3.02.4 2.41.8 1.81.5 C5 0.085 0.08 0.075 C4/C3 0.075 0.07 0.675

S C3/C2/C4 0.07 0.065 0.0625

C2 0.065 0.0625 0.06

C1 0.06 0.0575 0.055

Care is needed when using these relationships, particularly for high brightness, highly cleated, low strength coals as there is a possibility to underestimate the strength. Medhurst (pers comm.) also advises that determining the laboratory strength of coal is difficult and is perhaps best estimated by testingatverylowconfiningpressuresandthenextrapolating.

5.4

DEFORMATION PROPERTIES

5.4.1 MODULUS
Modulus values for the range of coal measure rocks are given in Figure 35. Ratios of the modulus to theUCShavebeenfoundtobereasonablyconsistentforeachrocktype(Table14).
Table14Guidelinesfortheselectionofmodulusratio(afterHoekandDiederichs,2006)

Lithology Conglomerates Sandstones Siltstones Claystones Shales

Modulus/UCS 300400 200350 350400 200300 150250

Reductions in the laboratory values (Ei) to rock mass values (Erm) can be made using the GSI value (HoekandDiederichs,2006)whereby:

52

Erm=Ei*[0.02+(1D/2)/(1+exp((60+15DGSI)/11))]. The reduction factor is shown in Figure 43 for the case where D = 0.0: it can be seen that intact or massiverockshaveavalueclosetounity,andfortypicalcoalmeasuresrocks(GSI=50)thereduction factorisintheorderof0.3.

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 10 20 30 40 50 GSI 60 70 80 90 100

Modulus reduction factor

Figure43Modulusreductionforcaseofdamageindex=0

Brady and Brown (1985) analyse the simple case of parallel discontinuities to model transverse anisotropy in a continuum model. They provide equations to estimate the equivalent deformation modulus and Poissons ratio values as a function of the spacing and properties of the discontinuities (Figure 44). The trend is for a reduction in the equivalent modulus and Poissons ratio normal to the discontinuityasthespacingreduces.
2 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1 0 0.25 0.226 0.202 0.178 E=2 GPa, Kn = 10 GPa/m PR=0.25 1 2 3 Modulus Poisson's ratio 4 5 0.154 0.13 Spacing of discontinuites (m) Equivalent Poissons ratio

Equivalent modulus (GPa)

Figure44SimplemodelfortheimpactofdiscontinuityspacingonthedeformationmodulusandPoissonsratio

The previous discussion has assumed the material is an isotropic continuum. Coal measures are characterisedbythepresenceofbeddingnotonlyasatexturebutalsoasthedominantdiscontinuity. It follows that a better assumption would be a transverse anisotropic continuum with the major additional parameter being the independent shear modulus (G). There is little guidance for the selection of the independent shear modulus for coal measures. Through back analyses using
53

continuum codes, the principal author has found values of between 30MPa and 250MPa give reasonablesimulationoffallcavities.

5.4.2 POISSONS RATIO

There is little information about the variations in Poissons ratio for different rock types, mainly becauseitsvaluehaslittleimpacton2dimensionalstressmodels.Aswillbeseen,thisisnotthecase wheninthecaseof3dimensionalstressanalyses. Figure 44 shows the decreasing trend in Poissons ratio as the spacing of discontinuities decreases. For rock masses, it can be anticipated that the Poissons ratio is less than the laboratory values, and thatitincreasesasthemeanstresslevelincreases.

5.5

IN SITU STRESSES

It is not possible to predict the state of stress in the ground from a knowledge of the depth of cover. The state of stress at any point will vary depending on the depth of cover, the presence of faults, the natureofanydiscontinuities,andthestiffnessofthelocalrocktypes. Thefollowinggeneralrulesneedtobeappliedwithcare.Ideally,thestressstatewillbemeasuredat each mine site. However, such a measurement is unlikely to be representative of all of the mine, especially when faults are traversed. Furthermore, stress measurements are difficult and expensive. The approach should be to apply regional knowledge of the general stress field to point measurements at the mine site, and particularly to observations of how excavations behave underground. In the latter case, it is essential not to jump to the paradigm that all roof falls are due to elevated horizontal stresses; in fact this is unlikely to be the case once the roof is supported at the densitiestypicalofcurrentAustraliancoaloperations.

5.5.1 STRESS IN STONE

Data from New South Wales and Queensland (Nemcik et al, 2006) show that the ratio of the horizontal stress to the vertical stress is typically between 1.0 and 2.5 times at typical mining depths, with even higher values at shallow depths (Figure 45). Hillis et al (1999) report that 80% of the data from the Bowen Basin has the vertical stress being the lowest of the 3 principal stresses, and 17% with the vertical being the intermediate stress; in the Sydney Basin 90% have the vertical being the minimumstress. Within the limitation of any model for stress magnitudes and directions, the following considerations haveprovedtobeuseful: The high horizontal stresses are most likely related to Tertiary and Recent Age migration of theAustralianplatetothenorthnortheast. At depths in excess of about 200250m the major principal horizontal stress is likely to be orientednorthnortheast. At depths less than about 200m250m, stress relief to the surface along bedding will mean that the minor principal horizontal stress will be parallel to the dip direction and hence the majorprincipalhorizontalstresswillbeparalleltothestrikeofthecoalmeasures.
54

Thepresenceofbeddingasthedominantdiscontinuitymeansthattheverticalstressislikely tobeaprincipalstress. Theverticalstress(inMPa)canbeestimatedasthedepthofcover(m)times0.025 Large deviations in the general stress pattern will be encountered around large faults and foldstructures. Areductioninstressmagnitudesshouldbeanticipatednearfaults. As a starting position, the magnitude of the major principal horizontal stress should be assumed to be twice the vertical stress, and the minimum principal horizontal stress should beassumedtobe1.5timesthemaximumprincipalhorizontalstress.

Figure45SummaryofstressmeasurementdatafromNewSouthWalesandQueenslandcoalfields(Nemciketal,2006)

5.5.2 STRESS IN COAL

Thestressfieldincoalissubstantiallymorecomplexthaninstone.Eneveretal(2000)havereported consistently low values of the minimum horizontal stress based on step tests during coal seam methane exploration (Figure 46). All the data indicate that the vertical stress is either the intermediateorthemajorstress.Nodefiniteexplanationhasbeenprovided. The situation is further complicated as overcore measurements conducted from mine openings into thick coal seams have revealed a different stress field again (Table 15). There are not many test results available and the fewthat are do show a consistent pattern. The field datagives a stress field thatisnotperfectlyalignedtotheverticalandhorizontal.Itisnotknownifthisismeasurementerror or a reflection of the true stress orientation. The key observations are that the major stress is approximately vertical and significantly less than the vertical stress that would be expected from the depthofcover.Thehorizontalstressesarelowerthanthemeasuredverticalstressesandaslowas 25%inmanycases.

55


Figure46Measuredminimumstressincoalasafunctionofdepth(Eneveretal,2000) Table15Compilationofstressmeasurementstakenfromunderexcavations

SeamA SeamB

SeamC SeamD

Nominalvertical Verticalstress overburdenstress (MPa) (MPa) 170 4.3 1.3 165 4.1 1.9 200 5.0 2.3 200 5.0 2.6 220 5.5 1.3 220 5.5 3.3 Testsunsuccessful(EneverandDoyle1996)

Depth (m)

Greaterhorizontal (MPa) 0.6 1.2 1.0 1.1 0.46 4.1

Lesserhorizontal (MPa) 0.3 0.5 0.4 0.6 0.27 1.3

Seedsman (2004) has speculated on the mechanisms that may be acting to produce this measured stress field in coal. He noted that the ratios of the vertical to horizontal stresses are similar to those expected from the vertical loading with lateral restraint (Poissons ratio effect) and that the vertical stresses are lower than those anticipated from depth of cover. He proposed a model based on shrinkage of coal ahead of the mining face as the groundwater (and gas?) pressures are reduced by the mining face. The overlying stone does not shrink and as a result there is a decoupling between the stone and the coal. The coal may undergo some sort of passive failure at this stage (as yet not defined). Closer to the mining face and as a result of a broadening of the dewatering shrinkage zone, the overlying stone sags and loads the coal vertically. These reimposed vertical stresses generatehorizontalstresses. The timing of the reloading needs to be considered. The discrepancy in the vertical stress suggests that separation has developed and this must disappear over time. It is possible that there are lower vertical stresses at the face compared to outbye. This matches observations underground where thereisoutbyeand/ortimedependentdeteriorationoftheribs.

56

If the reloading model is accepted, it is possible to extend the analysis to consider the impact of intense cleating in the coal. For the simple case of a coal mass containing a single set of parallel discontinuities, it can be shown that the closer the spacing of the discontinuities, the lower is the Poissons ratio of an equivalent transversely anisotropic continuum (Figure 44). This lower Poissons ratiowouldresultinalowerinducedhorizontalstressinmorejointedcoalasthecoalisreloading.

5.5.3 FAULTED GROUND

Ingeneral,stressmagnitudesshouldbelowerinfaultedground. Elevateddeviatoricstressesgeneratefailureinitiallyandthisfailuremaybeseenasfaults,moreoften than not with associated subparallel joints. Once the rock is broken, the maximum deviatoric stresses within the broken rock are controlled predominantly by the frictional resistance of the surfaces generated by the faulting. Indications of the scale of this effect can be obtained by considering passive earth pressures (Herget, 1988) assuming a cohesion of zero (Figure 47). Nemcik etal(2006)producedsimilarresultsthroughnumericalmethods.BradyandBrown(1985)discussthe application of specific tests for the mechanical acceptability of a stress state with respect to equilibriumonpervasiveplanesofweakness.
Magnitude of horizontal stress at passive failure (MPa) 0 0 Vertical stress 100 5 10 200 15 20 Depth (m) 300 25 30 35 500 5 10 15 20 25 30 35 40 45

400

600

700

Figure47Reducedlateral(horizontal)stressassociatedwiththepresenceofsurfaceswithlowerfrictionalresistancebased on(a)passiveearthpressures,(b)UDECmodeling(Nemciketal2006).

In Figure 47, it can be seen that the maximum horizontal stresses decrease as the friction angle decreases for example as would happen in the presence of slickensided surfaces. This observation is important for 2 reasons: (1) it would suggest that horizontal stresses are lower in proximity to thrustfaultsinclayeyrocksthaninsandstonerocks,and(2)thepoorroofconditionsthataretypically encountered in the vicinity of thrust faults are not the result of elevated horizontal stresses but are a consequenceofthepresenceofbrokenrockaboutthefaults.
57

5.5.4 TOPOGRAPHY
For most practical purposes, the influence of topography on the stress field is relatively small. Some simple continuum models of a 100m deep valley with valley walls at 62.5o and a 2:1 horizontal to vertical stress field have been conducted to demonstrate this (Figure 48). The models show that there are concentrations of horizontal stresses immediately beneath valleys along with the obvious reductions in vertical stresses leading to increases in deviatoric stresses. However, the impact of the valleyisnegligiblewithinabout1.5timesthedepthofthevalley. Hilltop or ridgetop mining (Figure 49) will be different with potential low horizontal stresses normal totheridgelineandmuchhighervaluesparalleltotheridgeline.
G ro und E levation : 1000.00

2.00 6.00

Sigm a1 Field stress: gravity Ground surface elevation: 1000 m Unit weight of overburden: 0.025 MN/m 3 Total stress ratio (horizontal/vertical in-plane): 2 Total stress ratio (horizontal/vertical out-of-plane): 1.5 Locked-in horizontal stress (in-plane): 1 Locked-in horizontal stress (out-of-plane): 1Stress Analysis

10.00

14.00

G ro und E levation : 1000.00

2.00

4.00 Deviatoric stress Field stress: gravity Ground surface elevation: 1000 m Unit weight of overburden: 0.025 MN/m 3 Total stress ratio (horizontal/vertical in-plane): 2 Total stress ratio (horizontal/vertical out-of-plane): 1.5 Locked-in horizontal stress (in-plane): 1 Locked-in horizontal stress (out-of-plane): 1Stress Analysis

6.00

8.00

G ro und E levation : 1000.00

5.00

10.00 Horizontal stress Field stress: gravity Ground surface elevation: 1000 m Unit weight of overburden: 0.025 MN/m 3 Total stress ratio (horizontal/vertical in-plane): 2 Total stress ratio (horizontal/vertical out-of-plane): 1.5 Locked-in horizontal stress (in-plane): 1 Locked-in horizontal stress (out-of-plane): 1Stress Analysis

15.00

Figure48Effectoftopographyonstressdistribution

58

Reductioninhorizontalstresses, relatedtoproximitytofreefaceofvalley
Figure49Horizontalstressesinhilltoporridgemining

5.6

LOCAL TERMINOLOGY

Over time, different coalfields have evolved different expressions for various geological features (Table 16). Some of the terms are colourful. There are some significant differences between the usage of the term cutter the USA uses the term for stress induced failures at the rib line while in Australiathetermisusedtodescribethetraceofbasicallyverticaljoints. ThereisatendencyinAustraliatousethetermgutterasageneraltermforallroofcracking.
Table16Localgeologicalterminology

Drawrock Stackrock,cataloguerock Headcoal Rash Kettlebottoms Horsebacks Hillseam Greasybacks Cutterroof(USA) Watercracks Gutter Trench Channelroof Cutters(Australia)

low strength shale/coal that often drops before bolts installed laminite Topcoal,coaltops Carbonaceousshalesandcoal Potarses fossilizedtreetrunks Rooffallsthatdevelopbetweenbolts Dilatedverticaljointsparalleltotheslope Slickensidedsurface Guttering Openjoints Shearing of the roof at the roof/rib corner that tends to hadeatasteepangleovertheroadway. Joint bounded rectangular cavities in the roof either betweenboltsoroverfullwidthofroadway Archshapedcavitiesthatdeveloppriortobolting. Smallscalenormalfaultsorjoints

59

MINING INDUCED STRESS CHANGES

Inthissection,boththeverticalandhorizontalstresschangesinducedaboutcoalmineroadwaysand also the immediately adjacent extraction areas are discussed. The focus is on the stresses that develop in the immediate vicinity of the roadway, and not necessarily on the stresses developed elsewhereinthesystem. There can be no doubting that considerations of the stress redistribution is complex, involving interactions between the insitu stresses and the large longwall goaf, stress redistribution around the roadway itself, and any body stresses induced by the deformation and movement of the immediate rockandcoalmass.

6.1

REDISTIBUTION ABOUT A LONGWALL

Boththeverticalandhorizontalstressesareconcentratedatthemaingatecorner(Figure50).Vertical stresses in coal have been measured in the context of pillar design (Mark, 1990, Colwell, 1998). Our knowledge of the redistribution of horizontal stresses about a longwall is based on extensive stress monitoring work conducted by Dr Gale and his associates over the last 20 years. In these latter studies, stresses and stress changes have been measured using hollow inclusion cells installed in stone above the chain pillars and about 5m 10m into the roof. As will be discussed, these are not thestressesattheroofline.

6.1.1 MAINGATE CORNER

Standard chain pillar design methods provide an estimate of the vertical stress expressed as a uniformlydistributedload.Theactualstressmagnitudesdependonthedepthofcoverandthewidth of the pillars. Typically the vertical stress is doubled at the maingate corner (Figure 50). Similar resultscanbeobtainedfromasimple2Delasticstressanalysis.Attheface/tailgatecornertheverticla stresshasincreasedfourfold. Measurements of horizontal stress concentrations and reductions about the maingate corner have beenconductedbySCT(Gale,2008)andothers.Thereisaconcentrationofhorizontalstressesabove the chain pillar at the maingate corner and a reduction in the horizontal stresses above the pillar behind the face adjacent to the goaf (Figure 51). There is also an increase in the vertical stress observedabovethepillarbehindtheface.

60

40m pillar, 500m depth (MPa)

50 Vertical stress (MPa) 40 30 20 10 0 Virgin

25m pillar, 250m depth (MPa) 40m pillar, 500m depth 25m pillar, 250m depth

5 4 3 2 1 0

Maingate

Bleeder

Tailgate

Double goaf

Figure50Pillardesignverticalstressesdevelopedabovechainpillarsassuminga200mwidepanel.

Figure51Generalpatternofverticalandhorizontalstressredistribution(Gale2008)

61

Vertical stress as a proportion of the virgin stress

60

The magnitude of the concentration of the major principal horizontal stress depends on the angle betweenthestressaxisandgateroaddirection(Figure52),withthepossibilitythatthereisadoubling ofthemagnitudeata45oangle.Asimple2Delasticanalysisproducesthesametrendwithrespectto the angle between the stress direction and the roadway but the magnitudes are higher this indicates that some failure may be developing in rock mass in the horizontal plane. This failure is likelytobebeddingplaneshear. As will be discussed, the simultaneous increases in both the horizontal and vertical stress are significantintermsofthestressesinducedintheimmediateroofofanexcavation.Dependingonthe concentrationfactorthatappliestothehorizontalstress,itispossiblethattheverticalstressactingin themaingatemaybecomethemajorprincipalstress.

Concentration of the magnitude of the princiapl stress tangential to the longwall excavation

2.25

Rangeofvalues
1.75

1.5

1.25

1 0 20 40 60 80

Angle between direction of major principal horizontal stress and gateroad (degrees)

Figure52Concentrationofhorizontalstressmagnitudeatthemaingatecornerasafunctionoftheanglebetweenprincipal horizontalstressaxisandtheroadwaydirection(afterGale2008)

6.1.2 BLEEDER/TAILGATE
InspectionofFigures50and51revealsthatthereareadditionalvertical stressincreasesonthechain pillar behind the longwall face (this will be referred to as a bleeder roadway) and reductions in the principal horizontal stress. The horizontal stress reduction is greater than that indicated from a 2 D elastic analysis. Tarrant (2006)possibly provides the mechanism for this greater stressrelief and also for the lesser stress concentration at the maingate corner shear along bedding in the direction of thegoaf. The combination of an increase in vertical stress with a decrease in horizontal stress would appear to be inconsistent with the socalled Poissons ratio effect. This effect cannot be used to explain the generation of stresses near excavations. Jaeger and Cook (1979) stress that the horizontal to vertical stress ratio is related to the Poissons ratio only in the case of complete lateral restraint (uniaxial strain).Suchcompletelateralrestraintisnotavailableinproximitytoanexcavation.

6.1.3 TAILGATE CORNER

Monitoring of horizontal stresses above pillars in tailgates has rarely been reported. The vertical stresses within the pillar have been comparatively better studied. Shen et al (2006) instrumented a
62

tailgateatUlanatadepthofabout200mandshowedthatoveralltheverticalstressincreasedby34 MPa while at the same time the horizontal stress decreased by up to about 1 MPa (Figure 53). The overallpatterninthetailgateisconsistentwithincreaseinverticalstressesimplicitinthepillardesign models(Figure50)andfurtherreductioninhorizontalstressesresultingfrommorelateraltranslation oftheroofintothenowtwolongwallvoids.

Increasesinverticalstress

Decreasesinhorizontalstress
Figure53Stressesaboveapillarinatailgate(Shenetal2006).

6.2

REDISTRIBUTION ABOUT A ROADWAY

6.2.1 BASIC CONCEPTS

Pragmatically, stresses and failure state are better understood in 2 dimensions a vertical stress and onehorizontalstressnormaltotheroadwayaxis.Inthefollowingdiscussion,boththesestressesare conveniently taken to be principal stresses. Whilst this is a reasonable assumption for the vertical stressoninitialroadwaydevelopment(becausetheseamdipsaresoflat),itisnotstrictlyvalidforthe horizontal stress if the roadway is not aligned parallel to a principal stress. Also note that the assumption of a vertical principal stress may not be valid once the stresses are redistributed about a longwall goaf. A simplifying assumption needs to be made such that the magnitude of the horizontal principalstressinthefollowing2Dmodelscanbecalculatedfromthefollowingrelationship:
K=0.5(L+M)0.5(LM)cos2
63

Equation6

where K=ratioofhorizontaltoverticalstressinplanegeometry L=ratioofmajorprincipalhorizontalstresstoverticalstress M=ratioofminorprincipalhorizontalstresstoverticalstress =anglebetweenroadwayaxisanddirectionofthemajorprincipalhorizontalstress.

ThereaderisencouragedtodownloadExamine2DfromtheRocsciencewebsite.Thisallowsanalysis of elastic stress redistribution about any shaped roadway to complement the readily available solutionsforacircularexcavation.Whilstthefollowingelasticanalysesgiveinvaluableinsight,itmust always be remembered that rock and coal masses are not elastic and that even the smallest deformationscanresultinmajorchangesinthewaythestressesaresubsequentlyredistributed. Elastic solutions are available for the stresses induced about circular and elliptical holes (Table 17). Notethatthestressesareindependentofsizeofthehole.Foracircularhole,thestressesattheroof centerline are tensile for K values of less than 0.3; for ellipses with a width/height ratio of 2:1, the criticalKvalueis0.5.
Table17Roofandsidestressesforcirclesandellipses
p

Kp W

Centreline Rooforfloor Sides

Circle P(3K1) P(3K)

Ellipse P(K1+2KH/W) P(1K+2W/H)

These simple equations provide some insight into how coal mine roofs may behave. For the general stone stress model, where the horizontal stresses are greater than the vertical andhence K is greater than 1, the implication is that roadway roofs are under compression. Importantly the coal stress model, where the horizontal stresses are less than the vertical (Table 15), implies that the roof stressesmayapproachtheonsetoftension.Inthebleederroadwayandparticularlythetailgate,the higherverticalstressestogetherwithreducedhorizontalstresseswillmeanareductionintheKvalue andthetrendtoreducedroofstressesifnottheonsetoftensilestress5. Brady and Brown (1985) extend this elastic analysis of stresses to the case of a flat lying feature (beddingparting?)locatedtowardsthetopoftheexcavation(Figure54).Shouldtherebesliponsuch a surface, the result is a reduction in stresses acting in the crown of the excavation. If such a surface islocatedabovetheexcavation,theresultisanincreaseinthestressesactinginthecrown.

Ifitisassumedthattherockorcoalmassintheimmediateroofhaszerotensilestrengthbecauseofthepresenceofjointing, theactualstressescannotbecometensilebecauseoftheonsetoffailure.
5

64

Stressincreases Stressreductions

Figure54Effectofplanesofweaknessondistributionofroofstresses

6.2.2 ELASTIC STRESS REDISTRIBUTION AROUND A RECTANGULAR ROADWAY

TheelasticstressredistributionaboutrectangularroadwayscanbereadilyassessedusingExamine2D. It is emphasized that this simple model assumes no subsequent yielding of the rock. This should not beusedasareasontodismissthesimpleelasticcodesinfavourofmoresophisticatedonesbecause, whilst the latter may have the numerical ability to handle yielding, the lack of validated constitutive equationsandinputparametersmeansthattheyarestillintheresearchdomain. Figures55to57presenttheresultsoftheanalysesofatypicalrectangularroadwaywithanumberof stress fields. In these analyses the roadway is 2.8m high and 5.2m wide. Six stress ratios have been considered K= 0.2, 0.5, 1.0, 1.4, 1.7, and 2.0, with the major principal stress being 10 MPa. Three stresscomponentsarepresented:deviatoricstress(13)whichisthedriverforcompressive/shear failure (Figure 55),h the horizontal stress which, if tensile, would allow the onset of shear along vertical joints (Figure 56), and v the vertical stress, which is significant in terms of rib behaviour (Figure57).AsummaryoftheresultsispresentedinTable18.
Table18Summaryofstressesforrectangularroadway(1=10MPa)

Horizontalstressat0.2m intoroofatcenterline (MPa) 5 1 6 8 9 10

0.2 0.5 1.0 1.4 1.7 2.0

Horizontalstress 0.1mfromriband 0.1mintoroof (MPa) 8.3 13.5 24 22 22 21

Maximumdeviatoric stress0.1mfromrib and0.1mintoroof (MPa) 19.5 21 27 24 21 21

Verticalstressattop ofribat0.2mintorib (MPa) 19 22.5 20.5 14 10.5 9

When the K value is greater than 1, the contours of deviatoric stress tend to form an arch over the roof line but this does not develop when the K values are less than 1 (Figure 55). The highest values areattheroof/ribcornerwiththemagnituderangingfromabout30MPaforaKvalueof1.0toabout 21 MPa for higher or lower K values. It is this concentration at the roof corners that is one of the sourcesofstressguttering.

65

From Figure 56, the centreline of the roof has tensile horizontal stress for K values less than 0.5, and the roof stress becomes increasingly compressive as the K value increases. Near the rib line the horizontalstressesarecompressive,evenforlowKvalues. Theverticalstressesimposedontheribvaryfrom22.5MPato9MPa(Figure57).Whenthesevalues are considered in the context of percentage increases, the variation is relatively small 225% to 180%. It is of value to highlight that elastic theory proposes that the stresses in the roof of an excavation may reduce and become tensile in the dominantly vertical stress field. This result is the direct opposite to that proposed by Colwell and Frith (2006) whereby the Poisson ratio effect is invoked in order to induce elevated horizontal stresses that are required for their failure mechanism. As mentioned above, the Poisson ratio effect is not applicable adjacent to excavation voids because the requirementforabsolutelateralrestraintcannotbemet.

K=2.0

K=1.7

K=1.4

K=1.0

K=0.5

K=0.2

Figure55Distributionsofdeviatoricstressesabouta1.86:1roadway(1=10MPa)

66

K=2.0

K=1.7

K=1.4

K=1.0

K=0.5

K=0.2

Figure56Distributionofhorizontalstressesabouta1.86:1roadway(1=10MPa)

67

K=2.0

K=1.7

K=1.4

K=1.0

K=0.5

K=0.2

Figure57Distributionofverticalstressesabouta1.86:1roadway(1=10MPa)

Figure58presentsanumberofanalysesforlowKvalues,suchthatthehorizontalstressmagnitudeis significantly less than the vertical stress magnitude (10 MPa). In this case the contours are the minimumprincipalstressandonlythenegative(tensile)valuesareshown.Thetensilezoneisweakly developed at a K value of 0.5 and it increases in height with decreasing K. Whereas the height of the tensilezonesdoesnotchangewithincreasingstressmagnitudes,themagnitudeofthestresseswithin theenvelopedoes.

68

K=0.2

K=0.4

K=0.3

K=0.5

Figure58Negativeminimumprincipalstressesinducedarounda1.86:1roadway(1=10MPa)

Shear stressesare developedabout the roadway, and of particular interest are those that are aligned parallel to bedding surfaces and normal to the roadway centreline. Shear displacements may be induced along bedding if the shear stresses are in excess of the frictional restraint that can be developed in response to the normal load across the bedding. The bedding parallel excess shear stress(BPXS)issimplycalculatedfromelasticmodelsas:
BPXS= n tan() Equation7

where

=beddingparallelshearstress n=stressnormaltobedding, =frictionangle.

Figure 59 provides an example of how the shear, normal, and bedding parallel excess shear stresses developalongasurface0.2mintotheroofofa5.2mwideroadway.Thevaluesarenormalizedtothe magnitude of the farfield vertical stress. It can be seen there are no shear stresses at the centre of span and both the normal and shear stresses increase towards the ribline. The bedding parallel excess shear stress peaks about 0.4m from the rib for a surface 0.2m into the roof and this position trends towards the centreline higher into the roof (Figure 60). Integrated across the roadway, the maximumBPXSisdevelopedabout0.5mintotheroof.

69

2 1.8 1.6 1.4

Stress(MPa/MPa)

400 350
Excessbeddingparallelshearstress(kN/MPa)

Normalstress Shearstress Excessshearstress 250 200 150 100 300

1.2 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 Distanceacrossroadway(m)

50 0

Figure59Exampleofshearandnormalstressesandbeddingparallelexcessshearstressassuminga35 frictionangleandK= 2.0

o

1 Height into roof (m)

0.5

0.5

1.5

2.5

3.5

4.5

5.5

Distance across roadway (m)

Figure60ContoursofBPXSasafunctionofheightintodistanceacrosstheroof(expressedasMPa/MPa)

6.2.3 THREE DIMENSIONAL STRESS CHANGES

Theanalysestodatehavebeenin2dimensions.Inthemine,thecoalfaceitselfprovidesrestraintto the immediately adjacent excavation and this results in changes in the way the various stress components develop. The following plots are from a 3D elastic boundary element code (Examine3D) andshowthestressesdevelopedonasurface0.2mabovetheroofline. Forthecaseofthedeviatoricstress,themaximumvaluesarenearthefacelineandalongtheribline (Figure 61). In this analysis, the roadway is aligned at 30o to the direction of the major principal horizontalstressandthebiasinthestressmagnitudestowardsonesideoftheroadnearthefaceline isapparent.

70


Figure61Developmentofdeviatoricstressesat0.2mabovetherooflineastheheadingisadvanced(K=2)

For a stress field where the vertical stress is the major principal stress, there is no bias in the development of tensile roof stresses and the peak of the tensile stresses is distant from the face (Figure62).

Figure62DistributionofnegativemeanstressintheimmediateroofofaroadwayforK=0.15

Forbeddingparallelexcessshear,mostoftheshearstressesdevelopclosetotheface(Figures63,64, 65). Note that in this 3D analysis, the shear stress that is considered is the bedding parallel shear stress normal to the roadway centreline. A key observation that must be emphasized is the rate at which BPXS develops away from the face at a distance of 2m, some 7580% of the final BPXS has
71

developed.Thisobservationiscriticaltounderstandinghowroofboltsareloadedinshearthebolts areonlyexposedtotheshearstressesthatareinducedaftertheboltsareinstalled.

Figure63ThreedimensionalviewofthedistributionofBPXS(ignorenegativesign)

Figure64VerticalslicethroughFigure63showinghowBPXSdevelops(ignorenegativesign)

72

Distance from centreline (m)

-1

-2 0 1 2 3 4 5 6 7 8 9 10

Distance from the face (m)

Figure65BPXSonasurface0.2mabovetheroofline

6.2.4 NON LINEAR STRESS REDISTRIBUTIONS

It is well established that coal mine roofs deform into the excavation as the roadway is advanced. This deformation zone is routinely identified with roof extensometry and is loosely referred as the softened zone or the height of softening. Whilst the term is somewhat misleading as the observations are simply of vertical movement and the origin of the movement maybedilation across beddingortheonsetofcompressivefailure,theuseofthetermsofteningdoeshavetheadvantage of focusing attention on what the impact may be in terms of the immediate roof stresses. Softening implies a lower modulus of deformation, which should mean that there is less of an ability to bear stressescomparedtostifferunitsnearby. The scale of roof movement and the associated stress redistribution has been demonstrated recently by Mark et al (2007). These authors were able to show that even at less than 20mm of roof movement; the horizontal stresses were already redistributed into an arch over the roadway (Figure 66). To quote from the paper The orientations of the principal stresses increases imply that the immediate roof of the crosscut yielded or softened and was not capable of transmitting additional horizontal stress.The measurements made during the study showed that the additional stresses were redirected above the immediate roof of the crosscut even before significant roof deformation hadoccurred.Thisredirectioncontinuedasthelongwallretreatedandtheroadwaywasexposedto maingatestressconcentrations. The samegeneral mechanism can be demonstrated ina2Dcontinuumnumerical model whereby the softening is simulated by the onset of compressive failure. Figure 67a show onset of brittle failure above a rectangular roadway. If the excavation shape is changed to reflect the failure zone, it can be seennotonlythestresstrajectoriesareredirectedaroundthefailurezonebutalsohowtheremaybe additionalfailureatthecrownofthesimulatedexcavation(Figure67b).

73

Afterdevelopment,priortolongwall

Afterlongwall

Figure66StressmeasurementsatEmeraldMine

(a)Failurezoneandstresstrajectoriesdevelopedabovearectangularroadway

(b) Failure zone and stress trajectories developed if the failure zone acts as a fully softenedzone.
Figure67Simulationofstressredistributionabovearoadwayusinganelasticmodel

74

Further evidence in support of the nonlinear redistribution of stresses can be found in the data on stress relieving roadways. Gale and Matthews (1992) reported that substantial relief from lateral stress can be obtained if roadways are driven close to an earlier driven roadway (Figure 68). The quantification of roof softening was not available at the time, but experience in the Southern Coalfield is that a bolted roof not an overall collapse is adequate to generate the stress relief. However, thekeypoint is that Newtons third law requires that the stress relief is also present within theexistingroadway.

Figure68Conceptofastressrelievingroadway

This nonlinear stress redistribution is particularly problematical when considering the stresses in the roof during the formation of an intersection. The second roadway is driven in a completely different stress field to that of the first roadway such that the stresses in the intersection roof at the point of breakthroughwillbelessthanthoseencounteredduringthestraightdriveage. There is no doubt that this behaviour is a key feature of underground roadway. Unfortunately, with the current state of the art, it is difficult to incorporate it in numerical analyses, and certainly not in routine design using numerical approaches. This limits the applicability of the numerical tools and requirestheapplicationofmorejudgmentbythedesignengineer.

6.2.5 STRESSES INDUCED WITHIN A BLOCKY ROOF

This discussion on mininginduced stress so far has been based primarily on continuum concepts (linear and nonlinear). As discussed earlier, a rock or coal mass is not a continuum and it is possible that its behaviour as a discontinuous medium can significantly modify the stress around an opening. There is value in considering a block model for the immediate roof and how stresses may be induced by the reorientation of the blocks in response to the formation of the opening. The simplification to rectangularblocksthatispossiblewithcoalmeasuresallowsconsiderationoftwosimpleanalogues.

6.2.5.1 VOUSSOIR BEAMS

The bedded nature of coal measures allows the ready application of the voussoir beam model (Brady and Brown, 1985). The concept here is that under situations of no applied lateral force, the incipient

75

rotation of the voussoirs induces a lateral thrust in the beam (Figure 69). The magnitude of this inducedlateralthrustdependsonthespan,densityandthicknessofthebeam.Atthepointoffailure ofvoussoirbeam,thecompressivestressesattheroof/ribcornerapproachthemagnitudeoftheUCS of the rock. An important point to note is that the result of voussoir action is the possible developmentofcompressivestressesattheroof/ribcornerandtensilestressesattheroofcentreline. An underground observer may observe the development of a stress gutter. The voussoir beam modelwillbediscussedinmoredetailinlaterchapters.
Figure69Voussoirbeamdeformationsinducecompressivestressesattheroofcornersandtensilestressesattheroadway centreline

6.2.5.2 CANTILEVERS
If the roof lineis exposed to the onset of tensile stress and there is sufficient relaxation such that the joints dilate, it is possible that a cantilever will develop (Figure 70). The failure of a cantilever in this situation will be through elevated shear/compressive stresses at the roof/rib corner. Once again thereisthepossibilityofgeneratingcompressivefailureinasituationofnoimposedhorizontalstress attheroofline.

Sigma 1 [MPa] -0.50 0.00

0.053 0.053 0.053 0.053 0.053 0.053 0.053

0.50 1.00 1.50

1

2.00 2.50 3.00 3.50

0

4.00 4.50

-1

10

11

12

13

14

Figure70Exampleofashearsurfacegeneratedbycantileveringaction

76

6.3

OTHER STRESS REDISTRIBUTIONS

6.3.1 TAILGATE PILLARS

Ifthereisdifferentialmovementbetweenthe2sidesofaroadway,theresultisanincreaseinthebay length of the roof line (Figure 71, after Diederichs and Kaiser, 1999) and a consequence reduction in the stress acting across the roof. The scale of this effect is in the order of 1MPa to 2MPa for a differential compression of 100mm, with a greater reduction for roofs with higher modulus values. Thescaleofstressreductionissignificantwhenitisrecalledthatthestressesattherooflineafterthe formation of the roadway may be already low as a result of the stress relief into the goaf and non linear effects discussed above. A potential location for such differential movement is when the roadway is bounded by a yielding pillar or coal fender and this is considered to be the basis of the relationshipbetweenchainpillardesignandtailgateroofsupportdiscussedbyColwell(1998).
Compressionofribside(mm) 0 0 10 20 30 40 50 60 70 80 90 100

0.5

Stresschangeinroof(MPa)

1.5 Modulus 5000MPa 2 10000MPa 15000MPa 2.5

Figure71Relaxationofarooflineasaresultofverticaldeformationinoneofthesides

6.3.2 VERY HIGH STRESSES UNDER OR ABOVE PILLARS.

For multiple seam operations, and particularly when the seams are close together, there may be a significant stress footprint about the existing pillars or goaf edge. The shape of the footprint has not been measured, but there are indications from mining operations that the significant stress changes extendfurtherthanasimpleanalogytofoundationsengineeringwouldsuggest. Insight into the stress footprint can be gained by considering the stress under a rigid footing for isotropic and transversely isotropic materials. When a high level of transverse anisotropy induced by joints or bedding partings is considered (modelled by using a low value for the shear modulus), the distributionofverticalstresschangessignificantlyfromthatforanisotropicmaterial(Figure72).

77

User Data Sigma YY 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00

User Data Sigma YY 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00

Shearmodulus=8333MPa

Shearmodulus=250MPa

Figure72Stressesunderarigidfooting

The pragmatic response should be to anticipate elevated vertical stresses when mining within say 30m of overlying or underlying pillars. When combined with what is known about the nonlinear distribution of horizontal stresses about voids, it would be prudent to consider the implications of a very low K value for the stress field and the possibility of induced tensile stresses in the roof (Seedsman,2003).

6.4

COMPILED MODELS FOR STRESS PATHS IN THE IMMEDIATE ROOF

Even allowing for depth variations, there is not a unique stress field acting in a coal mine roof. The stress conditions vary depending on the lithology that is present, the stage of mining,the dimensions ofthepillars,andthepresenceofworkingsinotherseams.Furthermore,thestressfieldmaychange in proximity to faults and other major geological features. Care needs to be taken when interpreting point measurements at one mine site as they may be more related to smallscale body stresses inducedbyblockrotationsandnotindicativeoftheoverallstressfield. Thesituationisnothopelessitispossibletoformulatesomegeneralmodels.Thesemodesprovide thebasisfromwhichdeviationscanbeidentifiedanduncertaintiesidentified.

6.4.1 STONE ROOF SINGLE SEAM

A model for the stress path for a stone roof exposed to an initial stress field of 6 MPa vertical and 9 MPa horizontal is discussed below for initial development, the maingate corner and the tailgate corner(Figure73).Deviatoricstressesandminorprincipalstressesareconsidered. Atthepointofexcavation,thereactiontotheoverburdenloadisremovedandtheverticalstressesat the roof line vanish. Until the roof deforms, the horizontal stresses are not yet redistributed. High deviatoric stresses and bedding parallel shear stresses develop immediately. These stresses may induce stress guttering at the roof corners, either by compressive failure of low strength rock or by theincipientdeflectionoftheroofbeams. Astheminingfaceadvances,saytoinexcessofroadwaywidth,theroofwillhavedeformedtoafinal state, or in some situations failed if overall compressive failure develops. In either case, the
78

horizontal stress in the immediate roof will have decreased to very low values and a stress arch developed higher in the roof. Deflections of the bolted roof will generate body stresses associated withtheformationofavoussoirbeamintheimmediateroof(Figure74). At the maingate, increases in both the vertical and horizontal stresses around the retreating goaf will alter the stress arch above the roadway and this may lead to an increase in the height of softening. There is no direct increase in the horizontal stress acting in the immediate roof because that roof has already deformed and softened. The extra softened material will be an additional surcharge loadingontheboltedbeamwhichwillthencauseanincreaseinthebodystresseswithinthevoussoir arch and an indirect increase in the horizontal stresses at the roof line. This may lead to the onset of stressguttering. DEVELOPMENT Verticalstress:Horizontal stress=6:9 MAINGATE Verticalstress:Horizontal stress=12:10 TAILGATE Verticalstress:Horizontal stress=25:7

guttering?

Increased size of softenedzone?

Loosening of rock in crown of softened zone?

Figure73Evolutionofdeviatoricandnegativeminorstressesduringlongwallretreat

Atthetailgateendoftheface,theimposedstressfieldisnowdominantlyverticalwithareductionin the horizontal stress due to the presence of the goaf on one side and also behind the faceline. The stresses within the stress arch decrease and there are some very high verticallyoriented deviatoric

79

stresses at the roof/rib corner. The horizontal stresses in the immediate roof remain relatively constant at low magnitudes, unless the chain pillar yields. In the latter case, reaction to the body stresses is lost and the horizontal stresses vanish. Localised compressive stresses may develop if the roofstructureallowstheformationofcantilevers.

Softenedzone

Vertical dead weight acting on boltedrockbeam

Stress arch induced in rockbeam

Insitu stresses redirected aroundsoftenedzone

Figure74RedistributedinsituandinducedbodystressesaboutaroadwaywithK>0.8oncetheroofandfloordeflects.

6.4.2 COAL ROOF SINGLE SEAM

The stress field in coal just ahead of mining should be considered to have all three components compressive but with the vertical stress dominant K values of about 0.2 should be assumed, with evenlowervaluesinthepresenceofcloselyspacedjoints. This stress field results in very low horizontal stresses in the roof of the excavations and possibly the onsetoftensilestress.Thesizeofthetensilezoneincreasesawayfromthefacewhichintroducesthe risk of falls developing outbye. If the model of the coal stress field being related to depressurization ofthecoalandreloadingbyagraduallydeformingoverburdenapplies,thereisalsothepossibilitythe verticalstress magnitudewillincreaseoutbye,butthatshouldnotaltertheheightofthetensilezone unlessthereisareductioninPoissonratiowiththehigherstressthatdevelopoutbye. Destressedroofmayshowsignsofelevatedbodystressesduetoblockrotationswhichmaymaskthe overall behaviour (Figure 75). This is more likely if there are only a few large blocks multiple small blocks may not be able to adequately interact. It should also be noted that a destressed coal roof mayhavedilatedjointsandcleatssuchthatresinlossispossible. Atthemaingatecornertheverticalandhorizontalstressesincrease,butifKvalue staysthesamethe resultwouldbenosubstantialchangeintheroofstressconditions. Withthelackofhorizontalroofstresses,coalroofsareparticularlyexposedtoanystressrelaxationin thetailgatethatcouldbeintroducedwithayieldingtailgatepillarsdesign.

80

Destressedroofduetoonset oftensileconditions

Stress arch induced in coalbeam

Figure75RedistributedinsituandinducedbodystressesaboutaroadwaywithK<<0.8oncetheroofandfloordeflects

6.4.3 STRESS PATHS IN THE RIBS

The stresses in coal are dominantly vertical throughout the longwall mining cycle. This means that the ribs are exposed to progressively increasing deviatoric stresses dominated by very high vertical stressesandnegligiblehorizontalstresses.

81

SUPPORT AND REINFORCEMENT TECHNOLOGIES

The terminology for the various support and reinforcement elements is rather loose, and is not consistent between mining and civil. In the following discussion the generic term tendon will be usedtohighlightthefactthattheseelementsworkintension. Thereare2basicapproachestothepreventionofroofandribcollapsesupportorreinforcement. Support implies that the ground did not have, or has lost most if not all, any selfsupporting capacity in the stress regime that is active at the time. The term support may have it origin in the use of propsandbeamsthatprovidedsupportfromunderneaththeroof.Nowadays,thefullcrosssectional area of the roadways is required and the support is most often in the form of suspension from a zone in the roof that isconsidered tobe stable. Recognising that the objective is suspension, the key aspectoftheinstalledtendonsistensileloading. Reinforcement involves conserving or improving the overall rock mass properties and, in the context of coal measures, this becomes the prevention of shear along bedding. During the reinforcement action,thetendonsshouldbeloadedintension. The various technologies can be used in either a support or reinforcement mode. Essentially the key aspectisthetensilecapacityofthetendon. In mining, some deformation is acceptable and they are implicitly if not explicitly included in the design of openings. These deformations can come from the dilation of a yielding rock/coal material, from shear along bedding, or from rotation of discrete blocks. The implication of this is that the tendons must have a relatively large degree of tolerance of deformation. The analytical tools presented in this report are based on a balance of driving and restraining stresses; deformations are not considered. Validation of the tools has been through back analysis of existing technologies. The implication of any differences in the postyield capacity of tendons, as compared to H and X grade steel, needs to be carefully considered. Higher strength, but more brittle steel may leads to major reductionsinroofconditions.

7.1

REINFORCEMENT ACTION

Inasupportfunction,thetendonscanbeassumedtobeloadedsimplybydeadweight.Theimposed loading is tensile controlled by the mass of the block that is being suspended and the capacity of the anchorage. As defined above, the reinforcement mode relates to shear along bedding discontinuities. The reinforcement action can be direct shear of the tendon (guillotine), the dowel effect where the shear resistanceiscontrolledbyabearingfailureintherock,orthefrictioneffect(bolt)wherethefrictional restraint provided by a closed bedding parting can be exploited (Figure 76). Coal measure rocks are notstrongenoughtogenerateaguillotineeffect.

82

 Guillotine
Figure76Tendonsinshear

Dowel

Bolt

7.1.1 DOWEL EFFECT

The dowel effect applies when the bedding surface is open and results in the controlling factor being the compressive strength of the rock. There are several different equations that describe this effect andallaresimilar.Bjurstom(1974)proposesthatthedowelresistance(Figure78)isgivenby:
Dowel=0.67d (UCS.By)
2

Equation8

whereBy=yieldstressofthesteelandd=diameteroftendon.

8 7 Dowel resistance (tonnes) 6 5 4 3 2 1 0 0 10 20 30 40 50 60 70 80 90 100 Rock UCS (MPa)

Figure77Dowelresistancefor21mmdiametertendonsasafunctionofUCS

X grade 650 MPa

S grade 400 MPa

Mild steel - 300 MPa

7.1.2 FRICTION EFFECT

The friction effect requires closed bedding partings. Noting that the BPXS has developed to a large degree by the time any tendons are installed, there is a high likelihood that bedding partings will be open during the bolting. Installing fully grouted dowels can lock in the bedding parting open. The pragmatic ways of closing any open bedding parting is either to pretension the bolt against a point anchor,ortojackuptheroofbyreactingagainstthefloor.

83

Thereisalsotherequirementthatthebeddingstaysclosedduringanydeformationthetrendofthe plunge of the bolt needs to be vertical or in the opposite direction to the sense of subsequent shear (Figure78).Forcoalmineroadwaysthismeanseitherverticalorfannedoutwards.

Increaseinshearresistanceduetogreaterfrictionalrestraintandtensilestressintendon

Decreaseinshearresistanceasaresultoflossoffrictionalrestraint
Figure78Importanceofboltangleinmaintainingclosedbedding

Forrationalboltorientationsandsheardisplacements,thefrictionalresistanceofabolt(Figure78)is givenby: BoltFriction=U*tan() Equation9

whereUisthetensileloadinthebolt.

25

20

20 25 30 35

SHEAR RESISTANCE (TONNES)

15

10

0 0 50 100 150 200 250 300 350

TENSILE LOAD IN BOLT (kN)

Figure79Frictionalshearresistanceprovidedbybolts

Comparing Figures 79 and 77, it can be seen that the same X grade tendon can provide either say 5 tonnes (as a dowel in 45 MPa rock) or 24 tonnes (as a bolt across a closed parting with a 35o friction angle).

84

7.2

TENDONS

Forcompleteness,alistingofthemostcommonboltsusedinAustraliaisprovidedbelowalongwitha tableoftypicalboltsusedinUSA.
Table19SummaryofAustralianbolttypes

Grade BoltDiameter(mm) CoreDiameter(mm) YieldStrength(MPa) Yieldstrength(kN) UltimateTensileStrength(MPa) UltimateTensileStrength(kN) Standardelongation(%) Uniformelongation(%)

DSIX 23.2 21.7 650 240 920 340 15 8

DSIH 23.2 21.70 400 150 680 250 22 10

DSIS 23.2 21.7 300 110 475 175 35 12

DSIAVH 22.5 20.7 775 260 935 315 14 3

JennmarJX 24 21.7 650 240 890 340 19

Table20SummaryoftensilestrengthofUSAbolttypes

Yieldload(kN) Reference Diameter 5 18.1mm 6 21.8mm 7 25.4mm 8 29.0mm 9 32.6mm Ultimateload(kN) 5 18.1mm 6 21.8mm 7 25.4mm 8 29.0mm 9 32.6mm

40 340 88 126 172 225 284 550 142 204 278 363 460
Table21Longtendons

60 480 124 178 243 317 401 720 186 268 364 476 602

75 585 151 217 296 386 489 790 204 294 400 522 660

90 690 178 256 349 456 577 930 240 346 470 614 777

Type Features

Diameter(mm) Holediameter(mm) Yieldstrength(kN) Ultimatestrength(kN)

HiTen Flexibolt Plainstrand cableresin capsules 23.4

TGCable Internalbreath tube

BowenCable Bulbedstrand alongfulllength

SSB

Megabolt Straight indentedwires Internaltube 2739 3545 480870

Strandwith bulbed anchorage 28 21.8 21.8 Bulbs:3238 Bulbs:3218 28 3842 4245 3555 500 560 525 525 580 630 590 590 N.B.Capacitymaybelimitedbythebarrelandwedgestrength

85

7.3

ANCHORAGES

Overwhelmingly in the Australian coal industry, the anchorages are resinbased to allow rapid installation.Postgroutingistypicallydonewithcementitousgrouts. The design of the anchors can be based on the ground anchor approach in civil engineering (Littlejohn, 1993). Recognising that the strength of the resins is in the order of 70 MPa, which is stronger than many of the coal measure rocks (Figure 35), failure will take place at the resin/rock interface, assuming that the steel/resin interface is adequately rough. The pull out capacity of an anchor(Tf)isgivenby: Tf=c*3.14159*d*L Equation10

wherec=UCS*(1sin)/(2*cos), L=length,andd=diameterofhole. It is recommended to apply a relatively high factor of safety for cartridge based systems to account for uncertainties with mixing andgloving.Civil engineering would have a minimum of 2.0 forcement grout systems, and for mining this value would be appear to be appropriate. Figure 80 provides a simpledesignchartforthelengthofgroutedanchorageforeachtonneofboltload.
0.08 Length of anchorage @ factor of safety of 2.0 (m/tonne)

0.07 Hole diameter 0.06 28 30 32 35 45 55

0.05

0.04

0.03

0.02

0.01
friction angle = 30
o

0.00 10 15 20 25 30 35 40 Unconfined compressive strength (MPa)

Figure80Recommendedminimumanchoragelengthsincoalmeasurerockswithresinanchorages

It is noted that this approach indicates that large holes provide shorter anchorages. It is essential to notethatthisassumesideal resinmixing whichmaynotbepossiblewithalargeannulusthatcould resultifstandarddiameterboltsareusedinlargeholes. Equation 10 can also be used in reverse, to estimate the UCS from the results of short encapsulation lengthpulltests,onceanassumptionismaderegardingthefrictionangle.

86

7.4

STRAPS AND PANELS SKIN RESTRAINT

Australian coal mines use straps or mesh panels, not shotcrete. The task of straps or panels is to protect the operators from the fall of scat and to provide restraint to the roof if there is a tendency fortherooftocollapsebetweenthebolts.Fromapragmaticoperationperspective,strapsandmesh also assist in the correct location of the bots. The straps or panels transfer vertical loads to the bolt collars. The mechanics of this transfer are complex as the loading is orthogonal to the orientation of thestraporpanel. Iftheroofisundercompressionandhasnotsufferedoverallfailure,theloadingonthestrapormesh is relatively low and is related to the collapse of thin slabs of the immediate roof (say in the order of 0.5 tonnes for bolts spaced 2m apart). If the roof has failed under either compressive or particularly tensile loading conditions, the loads on the skin may be significantly higher (readily in excess of 5 tonnes). Failure of the strap or mesh will be at the bolt locations and may be due to vertical shear through a rigidmembraneortensilefailureifdeformationsareallowedtodevelopthelatterisconsideredthe likelymechanismforsteelstrapsorpanels. Coates(1970)providesananalysisoftheloadingofstrapsorpanelsontheassumptionthatitformsa catenary.Thetension(T)inducedinthemeshcanbeestimatedby:

T=Pvs2/(8q)

Equation11

wherePvisverticalpressureonmesh,sisboltspacing,andqisthesagofthemesh. A design chart based on this equation is given in Figure 81. Inspection of this chart shows that very high loads can be readily induced in a strap or mesh panel especially if large deflections do not develop.Themorethemembranedeflects,theloweraretheinducedloads. Therearenodataontheloadingandfailureofcoalminingpanels,buttherearedataonsimplemesh panels used in the metal mines (Figure 82). For the mesh of 5mm diameter wires, the ultimate capacity is about 2 3 tonnes after about 100mm to 300mm deflection. Note that each weld is typically certified at 0.85 tonnes, so the ultimate load represents only about 3 welds this is not surprising once it is recognised thatultimately the load on the mesh is transferredto2wires reacting againstaboltorcablecollar. For W straps, reference can be made to the bearing capacity of bolts in sheet steel, whereby the bearingresistanceperbolt,Br,canbecomputedinaccordancewiththefollowingequation;

Br=CdtFu

Equation12

where C = 3 for this geometry, d = nominal bolt diameter, t = sheet thickness, and Fu = tensilestrengthofsteelsheet. For 350 grade 1.9mm thick steel sheet, this represents 4.2 tonnes. This suggests that W straps could betwiceaseffectiveameshpanelsinthisloadingcondition.
87

100

Induced load in membrane (tonnes)

10

50mm deflection, 0.5m bolt spacing 50mm deflection, 1m bolt spacing 50mm deflection, 2m bolt spacing 100mm deflection, 0.5m bolt spacing 100mm deflection, 1m bolt spacing 100mm deflection, 2m bolt spacing 200mm deflection, 0.5m bolt spacing 200mm deflection, 1m bolt spacing

0.1 0 1 2 3 4 5 6 7 8 9 10 11 Applied load (tonnes)

200mm deflection, 2m bolt spacing

12

13

14

15

Figure81Loadingofastraporpanelifloadedasacatenary

Figure82Resultsofloading1.5mand2.0msquaremeshpanels(Thompson,2004)

88

PREVENTION OF ROOF COLLAPSE

The previous sections have outlined the range of geological and strength conditions that may be presentintheroofofacoalmineandhavediscussedtherangeofstressesthatcanbeappliedorare induced about a coal mine roadway. This section provides specific steps in a framework for the specificationofsupportorreinforcementtopreventroofcollapse. The basis for the framework is the logical methodology for mine excavation design presented in BradyandBrown(1985)andsubsequenteditions.Thelogicalmethodologywasinitiallyproposedfor massive rock cut by one or two discontinuities, but in later editions its application was extended to moderatelyjointedrock.BradyandBrownalsoemphasizethatthepresenceoffailedzonesarounda mining opening is common; the mining problem is not necessarily to prevent failure but certainly to preventuncontrolleddisplacementofrockfromtheboundaryoftheexcavation. There are 4 fundamental steps in the framework that lead to 5 different roof support designs (Figure 83).Becauseofthecomplexstresspaththatacoalmineroofcanundergoduringlongwallextraction, the steps in the framework may need to be applied at least 3 times initial roadway formation, maingatecorner,andtailgatecorner.Thestepsare: 1. Checkforthepresenceofnonverticaljointstheoperationalconcernisthatsuchjointscan define key blocks that are unstable in any stress regime and may be difficult to support with the support hardware and drill rigs on standard continuous miners. Such joints are morelikelytobepresentaroundfaults,butshouldbeanticipatedtobepresentanywhere. Checkforthepossibilityofoverallcompressivefailure.Thisconditionisthemajorhazardat the maingate. It may also develop on initial roadway formation, particularly at low depths of cover where weathering/alteration effects have reduced rock strength. It would appear that current mining depths in Australia (less than 550m) do not induce such failures in standardroofrocks. Check for the possibility of tensile horizontal stresses in the roof. This condition should be anticipated for coal roofs, tailgates where the chain pillar is designed to yield, under pillar and goaf edges in multiple seam mining, and when mining close and parallel to valleys and highwalls. Assess whether large unsupported spans can be formed. It will be shown that a spanning unit needs to be only 0.25m 0.3m thick. While intrinsic geological variability does not allowtheassumptionthatsuchroofisalwayspresent,itmaybepossibletotestitspresence bytakinganextendedcut(unsupported)longerthantheroadwayspan. Ifsuchatestcannotbemadeoneachcut,theroofdesignmustbebasedontheassumption that thinly bedded units are present that need to be reinforced to stop bedding parallel shear. Because the behaviour of bedded roof is nonlinear and there are major stress redistributions in the roof after only very small deflections, this reinforcement needs to be consideredonlyatthetimeofinitialroadwayformation.

2.

3.

4.

5.

89

NON-VERTICAL Y JOINTS ? N

CHANGE LAYOUT SUSPENSION

COMPRESSIVE Y FAILURE ? N

SUSPENSION LONG TENDONS CROSS MEMBER

TENSILE FAILURE ? N

SUSPENSION CROSS MEMBER

STABLE OVER EXTENDED CUT ? N

SKIN RESTRAINT

REINFORCEMENT SKIN RESTRAINT

Figure83Logicalframeworkappliedtocoalmineroofsupport

The analyses contained within the 4 steps utilise relatively simple elastic stress models applied to a number of preidentified failure modes. A limit equilibrium approach, balancing driving and restraining stresses, is used to quantify thesupport or reinforcement requirements. In common with good engineering design practice and where possible, calibration/verification is sought by reference to actual mine experiences and this is used to provide guidance on the initial selection of factors of safety.

90

8.1

NON VERTICAL JOINTS

For the mining geometries and roof bolters that are used in Australia, the presence of a joint or a series of parallel joints dipping at less than a certain threshold and normal to the roadway centerline is a major hazard in any roof stress regime (Figure 84). For a compressive stress regime, shear can develop along the joints; conversely for a tensile regime the blocks can fall under gravity. Furthermore, a standard roof bolt pattern is unlikely to intersect the joints and provide adequate reinforcement. Note that loading on the skin restraint (straps or mesh) could readily exceed the availablecapacity.

Figure84Thehazardofparallelnonverticaljoints.

The model in Figure 84 is only in 2 dimensions. There is a need to consider the trend or strike of the joint with respect to the roadway axis. Direct analogy with rock slope design practice would suggest thatthehazardisreducedifthestrikeofthejointsisgreaterthan20ofromthetrendoftheroadway. The authors are not aware of any specific work on the appropriate angle for roof geometries so recommend this analogy should be validated by mine site observations and not relied on without challenge. Nonparalleljointscanformwedgesortriangularroofprismsthatwillsimplyfallfromtheroof(Figure 85). Whilst this hazard cannot be ignored, in some ways these represent a lesser hazard in the coal mining sector as they will tend to collapse prior to the installation of the support. Furthermore this conditionwillmostlikelyberelatedtothepresenceoffaultsandthesupportregimewillhavealready beenmodified.

91


Figure85Nonparalleljointsdefiningtriangularprisms

8.1.1 COLLAPSE MODEL

ThesystemshowninFigure85collapseswhensheardevelopsalongthejoints. There is a need toconsider the stress conditions that areactive in the roof. Close to the mining face, beforetheroofhasdeflected,thestressfieldneartherooflinewithastoneroofwillbecharacterized by a high horizontal compressive stress. For coal roof, the horizontal roof stresses may already be very low. Back from the face, and also in the tailgate, the roof may relax and the horizontal stresses vanish.Inthatcase,themajorstresswillbeverticalrelatedtotheselfweightoftheblock. For the compressive stress regime, the critical dip angle in the immediate roof is 90 , where = friction angle on the joint. For typical roof lithologies and joint roughness, an appropriate critical minimumacceptabledipis65o.Thismeansthatthehazardcoulddevelopifthejointsdipatlessthan 65o. Figure86isbasedonthemethodofBradyandBrown(1985)forsymmetrictriangularprismsandcan beusedtoestimatethelikelyroofstressesthatwerepresentpriortoajointboundedfall.Notethat for these triangular prisms a relatively low horizontal stress stabilizes the roof. The reference providesthebasisformoredetailedanalysesofdifferentprisms.
90 Stable dip of joints forming a symetrical triangular roof prism 85 80 75 70 65 60 55 50 45 0.01 25 degrees 35 degrees 45 degrees 1m high prism, density = 2.5 t/m3

0.10

1.00

10.00

Horizontal stress in the roof (MPa)

Figure86Relationshipbetweenjointdip,jointfrictionandhorizontalroofstressforastablesymmetricprism

92

8.1.2 SUPPORT DESIGN

Itisnotpossibletogivespecificrecommendationsonhowthissortofgroundcanbesupported. The orientation of the joints and the available bolt angles makes it unlikely that the joints can be reliablyintersectedandthenreinforcedwithbolts. One approach could be to invoke suspension from more stable ground. The mass of potential joint bounded blocks can be readily determined from knowledge of the joint geometries and bolt/cable densityandlengthassessed. An alternative is to strap or mesh at the roof line across the joints. In this approach would be necessary to consider the loads that may be induced on the membrane. A more closelyspaced bolt patternwillprobablybeneededtoachievethisoutcome.

8.1.3 COMMENTARY ON DESIGN TO PREVENT COLLAPSE WITH NON VERTICAL JOINTS

Nonvertical joints in the roof are a particular hazard that should not be ignored. The actual conditions that a mining crew may encounter may be difficult to characterize in advance. Furthermore, the signs of the presence of such joints may be subtle and difficult to identify. It is essentialthatthecrewsaremadeawareofthehazardandtheneedforcloseobservations. Many zones of nonvertical joints will be found to be related to fault structures. If they can be anticipated in advance, it would be good practice to seriously consider the alignment of roadways so thattheyareatleast20oawayfromthefaulttrends. Recognising the severity of the hazard, trigger action response plans (TARPS) for the face crews shouldbebasedoneither: Stop mining and seek advice if dipping joints are trending within 20o of the roadway trend, or Installarelativelyintensepredeterminedsupportpattern.

The presence of nonvertical joints is the defining feature of metaliferous mines (Figure 5). By recognizing the difficulties in securing wedges, it is perhaps understandable why that sector has tendedtouserockmassratingsystemsasthebasisoftheirroofsupportdesign.Equallyimportantis therecognitionoftheadvantagesofthedrillandblastminingsysteminsuchrockmassesextended cuts,highblastvibrationtodislodgejointblocks,andagreateroperationaltoleranceofoverbreak.

8.2

COMPRESSIVE FAILURE

In this failure mode, there is an overall compressive failure of the rock mass induced due to high deviatoric stresses compared to the compressive strength of the rock. The failed volume of rock is then assumed to undergo a gravitydriven collapse. Roof collapse is prevented by the suspension of therockmassfromunfailedmaterialhigherup.

93

To maximize the ease of use of thiscompressive failure mode, the rock mass will beconsidered asan elastic isotropic continuum with brittle failure properties. The validity of these simplifications are tested against two published case studies. The simplifications allow this failure mode to be readily examined using freely available software (Examine2D). This report presents some simple equations that can be used to rapidly assess the likelihood of compressive failure. A more accurate analysis could be obtained using more sophisticated codes such as Phase2 or FLAC that allow layering of differentmaterials.Sucheffortwillproduceafullerunderstandingofthemechanicsofroofcondition but may produce only marginal improvements in the support design due to the limitation of being abletocharacterisethematerialstoenablemoresophisticatedplasticityapproaches.

8.2.1 COLLAPSE MODEL

Figure 87 presents the results of two Examine2D models for a brittle material with a UCS of 10 MPa. Contours of the strength factor are presented. The classic representation of stress guttering can be seen whereby there are zones of more failure (low strength factors) at the roof corners. In the first case, it is possible that roof bolts could be used to suspend the fall mass, in the second case longer tendonswouldberequired.

Figure87TwoanalysisofbrittlestrengthfactorusingExamine2Dshowingzonesofbrittlefailureandpossibleboltingand cablepatterns.

Aseriesofanalysesofa5.2mwide,2.8mhighroadwayin10MParockhavebeenconductedwiththe major principal stress of 10 MPa, and the two other stress components equal (Table 22). In general, low values of the strength factor define zones at the corners and higher values extend across the roadway. A series of analyses with different K values allowed the production of Figure 88 which shows the maximum height of the various contours of the strength factor for different K values. The surfacedefinedbythese3variablesisdistinctivelynonplanar.
Table22Databaseforbrittlefailureanalyses Horizontal stress (MPa) 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 Vertical stress (MPa) 5.00 5.00 5.00 5.00 5.00 5.00 6.67 6.67 6.67 6.67 6.67 Strength factor 0.525 0.5 0.45 0.4 0.35 0.3 0.65 0.6 0.55 0.5 0.4 Maximum height (m) 6.37 5.57 4.39 3.42 2.54 1.33 6.44 5.65 4.90 4.21 2.92 K 2.00 2.00 2.00 2.00 2.00 2.00 1.50 1.50 1.50 1.50 1.50 RSI @UCS=10MPa 3.81 4.00 4.44 5.00 5.71 6.67 2.31 2.50 2.73 3.00 3.75

94

Horizontal stress (MPa) 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 8.33 8.33 8.33 8.33 8.33 8.33 8.33

Vertical stress (MPa) 6.67 6.67 7.14 7.14 7.14 7.14 7.14 7.14 7.14 6.25 6.25 6.25 6.25 6.25 8.33 8.33 8.33 8.33 8.33 8.33 8.33 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 5.55 5.55 5.55 5.55 5.55 5.55 7.52 7.52 7.52 7.52 7.52 7.52 7.69 7.69 7.69 7.69 7.69 7.69 7.69 10.00 10.00 10.00 10.00 10.00 10.00 10.00

Strength factor 0.35 0.32 0.7 0.6 0.5 0.4 0.35 0.33 0.32 0.6 0.5 0.4 0.35 0.31 0.8 0.7 0.6 0.5 0.4 0.35 0.34 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.35 0.55 0.5 0.45 0.4 0.35 0.31 0.7 0.6 0.5 0.4 0.35 0.325 0.75 0.7 0.6 0.5 0.4 0.35 0.325 1.1 0.9 0.8 0.7 0.55 0.5 0.45

Maximum height (m) 2.24 1.44 6.50 5.21 4.03 2.86 2.16 1.63 1.46 6.19 4.47 3.05 2.32 1.36 6.05 5.28 4.47 3.61 2.62 1.98 1.77 6.44 6.08 5.72 5.30 4.86 4.39 3.88 3.23 2.39 1.73 6.10 4.97 4.05 3.26 2.43 1.36 6.02 4.92 3.88 2.80 2.11 1.55 6.36 5.83 4.80 3.84 2.76 2.07 1.52 4.87 4.11 3.75 3.38 2.62 2.42 2.11

K 1.50 1.50 1.40 1.40 1.40 1.40 1.40 1.40 1.40 1.60 1.60 1.60 1.60 1.60 1.20 1.20 1.20 1.20 1.20 1.20 1.20 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.80 1.80 1.80 1.80 1.80 1.80 1.33 1.33 1.33 1.33 1.33 1.33 1.30 1.30 1.30 1.30 1.30 1.30 1.30 0.83 0.83 0.83 0.83 0.83 0.83 0.83

RSI @UCS=10MPa 4.28 4.69 2.00 2.33 2.80 3.50 4.00 4.24 4.38 2.67 3.20 4.00 4.57 5.16 1.50 1.71 2.00 2.40 3.00 3.43 3.53 0.83 0.91 1.00 1.11 1.25 1.43 1.67 2.00 2.50 2.86 3.28 3.60 4.00 4.50 5.15 5.81 1.90 2.22 2.66 3.32 3.80 4.09 1.73 1.86 2.17 2.60 3.25 3.71 4.00 0.91 1.11 1.25 1.43 1.82 2.00 2.22

95

Maximum height of above roof line (m)

5
Horizontal/vertica l stress ratio (K)

2.00 1.80 1.60 1.50 1.40 1.30 1.20 1.00 0.83

0 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 Strength factor - brittle failure

Figure88RelationshipbetweenheightofstrengthfactorandtheKvaluefora5.2mby2.8mroadwayandamajorprincipal stressof10MPa

The simplicity of the brittle rock parameters (only a single compressive strength parameter, as the frictionangleiszero)combinedwithanelasticanalysisallowsanumberofeasyparametricstudiesas itcanbeshownthatthecontoursinFigure89areconstantforthesamestrength/stressratios. The roof strength index (RSI) is defined as the ratio of the unconfined compressive strength of the rock to the vertical stress. Plots of RSI against height of failure for different K values (Figure 89) indicate that they may define a surface that is close to planar, at least for heights of greater than 1.5m.Amultiplelinearregressiongivesthefollowingrelationshipfortheheightoffailuredefinedby astrengthfactor=1.0(Hf):
Hf(m)=1.25+6.12K1.94RSI(r =0.97),
2

Equation13

onlyforheightsgreaterthan1.5m. The equation is extrapolating for K values of less than 0.83, and should not be used for K values less than 0.6. A comparison of the actual and fitted data shows that there is a systematic error in the fit, with a potential error of 0.5m, which is judged to be acceptable. A brief study of the impact of different roadway heights suggests that there are small differences in the strength factor contours near the roof/rib corners but no significant differences in the contours for heights above 2m (Figure 91). Theroofstrengthindexisdefinedinthiswayasitisreadilycalculatedfromstandardgeophysicallogs theUCScanbecalculatedfromthesonicvelocitylogsandtheverticalstresscanbecalculatedfrom thedensitylog.TheRSIhasthesameformulationastheCompetenceFactor(Muirwood,1972).

96

6 5.5 5 Maximum height of failure (m) 4.5 4 3.5 3 2.5 2 1.5 0 1 2 3 4 5 6 7 Roof strength index
Horizontal/vertica l stress ratio (K) 2.00 1.80 1.60 1.50 1.40 1.30 1.20 1.00 0.83

Figure89HeightoffailureasafunctionofroofstrengthindexforvariousKvalues

Numerical analysis value

0 0 1 2 3 4 5 6 7 Multiple linear regression value

Figure90Comparisonbetweentheresultsofmultiplelinearregressionandthenumericaldata

2.2mhigh

2.8mhigh

3.3mhigh

Figure91Slightchangesinfailurezonesneartheexcavationwithnodifferencebeyond2m

97

8.2.2 SUPPORT DESIGN

Rock that has undergone compressive failure should not be assumed to have any self supporting capacity. The appropriate strategy is to suspend the failure mass from above. The design requires a predictionoftheheightandweightofthefailedzone. Anequationfortheheightoffailurebasedonespecificroadwaydimensionhasbeendeterminedand it would appear to be applicable for different roadway heights. It can used to estimate the length of thetendonsandtherequiredcapacity. Theheightofcompressivefailureiseither, Hf(development)=1.25+6.12K1.94RSI Equation14 Equation15

Hf(longwall)=1.25+6.12(K*(Lf/Mv)1.94RSI/Mv

where Lf = the horizontal stress concentration at the maingate corner (Figure 52), and Mv = maingate verticalconcentration(Figure50,assume=2?). The shape of the strength factor contours is not readily amenable to curve fitting to allow the containedareatobedeterminedbyintegration.Itisassessedthatareasonableapproximationtothe areaunderthecurveis0.7timesHf*roadwaywidth. AnalternativeapproachistodeterminetheRSIvaluewhichdefinesaheightoffailurethatpotentially overrides the primary bolts. For example, allowing for the bolt tail and the anchorage, the active length of a 2.1m bolt may be 1.8m. Setting Hf to this value on development, the minimum RSI that wouldpermitprimaryboltonlyis: MinimumRSI(2.1mbolt)=3.15K0.28. Equation16

Thefollowingisanexampleofhowthesupportdesigncouldprogress: Input:Depthof250m,sonicvelocity=3240m/sec,indicatedUCS=27,averagedensity=2.4t/m3.L= major horizontal/vertical stress ratio = 1.6, M = minor horizontal/vertical stress = 1.4, roadway driven at20otothemajorprincipalhorizontalstress.Secondarysupportistouse45tonnecablesinstalledin 28mmholes. Designvalues:Verticalstress=250*2.4*0.0098=5.88MPa,RSI=27/5.88=4.59,Mv=2.0,Lf=1.75,K =1.42.Assumeafrictionangleof30o. Output: Height of compressive failure development = 1.1m, mass of failed volume development = 9 tonnes, height of compressive failure maingate = 4.4m, mass of failure volume maingate = 39 tonnes.Anchoragelength=1.31m. Interpretation: The height of compressive failure on development is 1.1m which is less than the length of the primary roof bolting that is to be used. The dead weight of 9 tonnes is very much less than the installed capacity of the primary bolting. No additional support on development is required. The height of compressive failure at the maingate is 4.4m which is higher than the length of the roof bolts. Long tendonsupportwillberequiredpriortolongwallretreat.Apatternof2by50tonnecablesevery2.4m would be adequate to suspend the potential fall mass, and a pattern of 2 every 2m would give some

98

contingency. Long tendons would need to be 6m long if installed along the centreline or less if installedinpairsacrosstheroadway.

CaseStudy#1KestrelMine The roof strength index has been used in a study of maingate failure in Kestrel Mine in Central Queensland (Gordon and Tembo, 2005). The trend at this mine has been for long tendon support to be required at shallow depths and not at greater depths where the overall stresses would tend to be higher. It was found that an RSI value of 3.5 was a good predictor of the need to install additionalroofboltsondevelopmentandtoinstallcablespriortolongwallretreat.Avalueofless than 3.2 was a good predictor for the need to install cables on development. At Kestrel, the UCS waspredictedfromsonicvelocitylogs. The minimum RSI value at the adjacent Gordonstone operation was about 2.5 and this operation was characterized with very difficult development roof conditions requiring long tendon support on development. Similar low strength roof units are also present in the adjacent Crinum Mine (payne,2008) In these operations the roof bolts were 2.1m long. The horizontal to vertical stress ratios was 1.2 and1.6andthegateroadswerealignedsubparalleltothemajorprincipalhorizontalstress. The figure shows the relationship between the RSI and the height of brittle failure on the assumption that the longwall stress concentration was 1.2. Recognising the variations in roof strengths and the uncertainties in the stress information, it is considered that there is reasonable agreement between the mining conditions and a mode of when the bolts could be overridden by afailurezone.
3.0

2.5 Height of compressive failure (m)

2.0

1.5

1.0

Development Maingate

0.5

0.0 3.0 3.1 3.2 3.3 3.4 3.5 RSI 3.6 3.7 3.8 3.9 4.0

FigureCS11HeightofcompressivefailureasafunctionoftheRSIforthestressassumptionsat Kestrel

99

Casestudy#2EmeraldMine Emerald Mine has been referred to in the body of this report in the context of measured stress directions. The paper also includes information on the stress magnitudes and the height of softening measured with extensometers. The UCS ranges between 24.4 MPa 36.4 MPa, the immediate roof is coalandthenshalesandclays.Theroadwaydimensionsare4.9mby2.4m. This stress data indicates an initial K value of 1.78 and then both the vertical and the horizontal stress increaseequally.

The predictedheight of brittle failure canbe calculated as a function of the horizontal stress magnitude for different UCS values. It can be shown that the roof displacement data is consistent with the brittle failuremodeliftheUCSisintherangeof30to35MPa(whichisthemeasuredrange).
8 7.5 7 Height of failure using RSI model (m) 6.5 6 5.5 5 4.5 4 3.5 3 2.5 2 1.5 10 11 12 13 14 15 16 17 18 19 20 Horizontal stress (MPa) UCS = 20 MPa UCS = 25 MPa UCS = 30 MPa UCS = 35 MPa Mark et al (2006), Figure 15

100

Casestudy#3CrinumMine At the adjacent Crinum Mine a move from a 6 bolt to 6:2 pattern was found to be necessary when the RSI approached 3.5 (Payne 2007). Once again, at Crinum the UCS was predicted from sonic velocity logs. Figure CS2 1 shows the shape of the failure zone and the location of the 6 bolt pattern (thin lines) used at Crinum. The inner 4 bolts possibly do not have adequate anchorage. Thethicklinesshowthelocationoftheadditional2bolts.Itcanbeseenthattheseboltshavemore oftheanchorageoutsidethefailurezone.

FigureCS31StrengthfactorforRSI=3.5underdevelopmentconditions(K=1.2)

8.2.3 COMMENTARY ON DESIGN TO PREVENT COMPRESSIVE COLLAPSE

The design assumption is that the collapse zone can be suspended. This requires that the immediate roof line has some loadcarrying capacity so that the vertical loads can be transferred to the cables. The primary bolting and mesh is probably adequate for this role so long as the cables are located to minimisespans. The three case studies from the literature suggest that the prediction of compressive failure rising above the bolt anchorage is in reasonable agreement with some mining conditions. It is noted that there is some uncertainty in the estimation of the UCS from the sonic velocity logs. This would suggestthatanyfactorsofsafetyshouldbeclosetounity. Advocating compressive failure as the source of maingate instability, which is the implication of the RSIapproachprovidesexplanationstoanumberofkeyobservationswithrespecttodepthandstress: At the same mine, with the same roof types, there is no strong relationship between the needforsecondarysupportandincreasingdepth. Intheindustrydatabase,thereisawidedivergenceintheuseoflongtendonsupport. Compressive failure higher in the roof is consistent with stress reductions within the softenedzone.

101

Compressive failure higher in the roof can induce stress guttering at the roof line as a result ofverticalloadingoftheboltedroof. The magnitude of the bedding parallel shear stress that would be induced at the maingate if there was no relaxation in the softened zone exceeds the capacity of the bolting patterns used.

Coal measures are transversely anisotropic, especially for the finer grained units. Transverse anisotropycanincreasetheroofstressessuchthatthefailureheightsincrease(Figure92),soignoring them is nonconservative to the roof support design. Because of this, there may be a need to apply higher factors of safety in highly bedded material if using the simple equation and prior to field verification.Alternatively,aseparateanalysisusingExamine2Dormoresophisticatedcodescouldbe considered. As discussed earlier in this report, there is little guidance on the choice of the independentshearmodulus.

(a) Isotropicelastic

(b) Transverseanistropicelastic(G=200MPa)
Figure92Differentisotropyandfailureconditions

102

8.3

TENSILE FAILURE

The importantto point isthat a rockmass in compressionmaybehave as a stable continuum. In a destressed state, small imposed or gravitational loads can cause large displacements of componentrockunits.(BradyandBrown,1985) A rock mass must be assumed to have zero tensile strength (by virtue of its discontinuities) so the developmentofazoneofelastictensilestressmustbeassumedtobeazoneoffailureandassociated redistribution of stresses. Whether a roof collapse develops depends on the nature of the joints sets closelyspacedjoints,particularlyiforientedparalleltotheroadwayaxeswillpromoteacollapse. Since the collapse mode requires shear along vertical joints, it is appropriate to examine the elastic horizontal stresses that are developed in the immediate roof (Figure 93). As discussed in an earlier chapter, the stresses will become negative (tensile) when the K ratio is less than 0.65. Note how the stresses are still compressive at the roof/rib corner and that the maximum height is at the centre of the roadway. The height of the tensile zone (Ht) increases as the K value decreases (Figure 94) but does not change with increasing magnitudes of the stresses. Slightly different relationships exist for differentroadwayaspectratios. AnapproximaterelationshipbetweentheKvalueandheightofthetensilezoneis: Ht=2.3K24.74K+2.12(forKbetween0.1and0.62). Equation17

Figure93Negativehorizontalstress(K=0.2)fora5mby3mroadway

103

Maximumheightofzoneofnegative horizontalstress(m)

1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Kvalue

Figure94Heightofnegativehorizontalstress(5.5mby3.3mroadway)

5mby3.5m 5mby3m 5mby2.5m 5mby2m

K values of less than 0.65 may be more common than expected in coal mines. Certainly, the stress fieldsmeasuredwithincoalseamsinadvanceoftheminingfacehavesuchKvalues(Table15).Other locations where the mining geometry results in an increase in vertical stress and a reduction in horizontal stresses include close proximity to large excavation voids where nonlinear behaviour of theoverburdenresultsinthedevelopmentofstressarcheshigherintheroofandlowerintothefloor. Possible locations include tailgates where there is an adjacent goaf and under pillar goaf edges in multipleseamoperations.Thereisaparticularhazardifpillarsaredesignedtoyield.

8.3.1 COLLAPSE MODE

The collapse mode is simply the gravitydriven collapse of joint bounded blocks (Figure 95) or, in the case of a tailgate, may be the gravity collapse of blocks defined by the onset of brittle failure at the maingate. In the case of joints, whether the collapse occurs may be related to the kinematic acceptabilityoftheblocks(Figure96)wherebyblockinteractionsmayresultinametastableroofthat has adequate serviceability. It is apparent that roadway orientation has the potential to be a major controloncollapse.

Figure95Collapsemodefortensilestressregime

104


Figure96Roadwaysobliquetojointsetswillproducebetterconditionsinbothheadingsandcutthroughs

8.3.2 SUPPORT
Roofs exposed to tensile stresses are particularly difficult to stabilize. The quote at the beginning of this section highlights the fact that such roofs can appear to be stable and then collapse with little if anywarning.VisualinspectionsandTARPScannotbereliedonastheremaybenoslowdeterioration andnonoise. The support strategy should be to provide a tensile member in the roof (strap or mesh) and then to suspend from the zone of compressive stresses above the roadway which also has an adequate spanningcapability.Intermsofdeadweightloading,theboltingdensitywouldbeverylow,butthere is a need to consider bolt spacings and the loading of the straps or mesh. With the roadway geometries under considerations, it is not practical to attempt to reinforce the vertical joints themselves.Theloadingontheboltswillnotbehigh,themajorpracticalconsiderationistheloading onthestraporpanel.

Figure97Loadingonmeshpanels

8.3.3 COMMENTARY ON DESIGN TO PREVENT TENSILE COLLAPSE

Supporting roofs exposed to tensile stresses is relatively difficult as it is easy to overload the mesh or strap. The use of standing support would remove a number of uncertainties but is it compatible withtheroadwayuse?

105

Ifexaminingthiscollapsemodeinnumericalmodels,youshouldexamineplotsofhorizontalstressto assesspossibilityofmovementonjoints,plotsofnegative1andmeanstress,andthedistributionof failurezonesprovidedthatthetensilestrengthissettozero(intheMohrCoulombcriterion). In a zone of tensile roof stresses, it is possible for cantilevers to form which can then fail (Figure 98). Thisfailuremodecanbedeceptiveastheonsetofcollapse,asviewedfromtheroadway,mayinvolve compression. Given the hazards intrinsic in tensile failures, it is important that the full geometry of a rooffallisexaminedbeforeamechanismisdecidedupon.

Figure98Generationofcompressivestressesandfailureinaregimeofnoimposedhorizontalstress

8.4

DELAMINATION FAILURE

In this section we analyse the impact of the presence of bedding partings. Partings define jointed rock beams. These beams may be able to span across the coal mine roadway or, if they too thin, there is a need to use roof bolts to create acomposite beam. This composite beam can be created if shearalongthebeddingpartingsisprevented. Unless it can be demostrated otherwise, geological variability is such that it should always be assumed that close to extremelyclose spaced bedding partings will be present just ahead of mining. Extendedcutminingcanprovide such ademonstration.For immediate facebolting,where boltingis conducted within about 2.7m of the face, this cannot be demonstrated and, furthermore, there are stillchangesinthestressesactingintheroofatthepositionatwhichtheboltingisconducted. As the roadway is formed, shear can develop along bedding partings, the bedding may open, and individualbeamsofrockdevelop(Figure99).Ifthebeamsaretoothin,theymaycollapse. So there are two parts to the design exercise identifying the characteristics of a spanning unit and determininghowtocreatesuchaunitifthereisinsufficientconfidencethatoneispresent.

106


Figure99Slipandseparationinalayeredroofrock(BradyandBrown,1985)

8.4.1 COLLAPSE MODEL

The mechanics of the jointed rock beam is shown in Figure 99 failure can be by shear along joints, compressive failure through the rock substance, or by snapthrough (buckling). Voussoir beam or linear arch analysis can used to analyse the mechanics: a number of methods are available Brady and Brown (1985), Cpillar (www.rocscience.com), or Sofianos and Kapensis (1998). The latter is amenabletospreadsheetproceduresandhasbeenusedinthisreport.Alistingofthespreadsheetis includedinanappendix. Figures 100 and 101 present some typical relationships. As expected, the critical thickness (factor of safety = 1.0) increases with span and decreases with rock strength. For a 5.5m span, and a typical strength of 40 MPa, the necessary thickness without a surcharge is about 0.1m, and this increases to 0.3m for a 2m surcharge or a 10m span and a 1m surcharge. The associated deflections are about 20mm. The authors experience is that these numbers are reasonably consistent with interpreted fieldbehaviour.
0.5 0.45 0.4 0.35 Critical thickness (m) 0.3 0.25 0.2 0.15 0.1 0.05 0 1 2 3 4 5 Span (m) 6 7 8 9 10

UCS = 10 MPa UCS = 20 MPa UCS = 40 MPa UCS = 60 MPa

Figure100Criticalthicknessanddeflectionofvoussoirbeamsasafunctionofspanandrockstrength(E/UCS=250)

107

0.7

0.6

0.5 Critical thicknesss (m)

0.4

0.3

0.2 UCS = 10 MPa UCS = 20 MPa UCS = 40 MPa UCS = 60 MPa

0.1

0 0 0.5 1 1.5 Surcharge (m) 2 2.5 3 3.5

Figure101Criticalthicknessanddeflectionofa5.5mspanvoussoirbeam(1msurcharge)

Based onFigures 100 and 101, some simple rules ofthumb canbederived the required thickness is about2.5%to4%ofthespan,andtheresultingdeflectionisabout0.5%ofthespan. The critical thickness quoted above relates to a factor of safety of 1.0, and in most cases against compressivefailure.Greaterthicknessesresultinarapidincreaseinthefactorofsafety(Figure102). A30%40%increaseinthicknessresultsinchangefrom1.0to2.0inthefactorofsafety.
5

4 Factor of safety (compression)

2
UCS = 10 MPa UCS = 20 MPa UCS = 40 MPa UCS = 60 MPa

1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Thickness (m)

Figure102Stabilityofavoussoirbeamincreaseswithincreasingthickness(1mequivalentsurcharge)

8.4.2 REINFORCEMENT DESIGN

If a spanning unit has not or cannot be determined to be present, there is a need to install reinforcement across the bedding surfaces so that an adequately spanning beam can be formed. Determining the stresses to be applied to the reinforcement element is complex involving considerations of the full three dimensional stress field in a discontinuous material which is known to
108

behavenonlinearly.Inthefollowinganalysis,a3dimensionalelasticcontinuumcodewillbeutilized. This approach is assessed to be conservative (i.e. will over estimate the bedding parallel shear stress) as it does not account for the established stress redistributions to higher in the roof and relaxation near the roof line once the roof deflects. It will be assumed that at thedevelopmentmining face the roofbeamisexposedtothefullelastichorizontalstressregime. Section 6.2.2 discussed beddingparallel excess shear stresses (BPXS) and how they develop away from the face. Figure 64 highlights the fact that about 6070% of the BPXS have developed by the time the roadway has advanced 2.7m. If there are bedding partings present in the zone of positive BPXS, shear movements will have occurred. It follows that bolts are installed in a rock mass where movementshavedeveloped.Theboltswillonlybeexposedtosubsequentpotentialmovementsand theassociatedstresses.

8.4.2.1 DRIVING FORCES

A series of analyses have been conducted using Examine3D (www.rocscience.com.au) with a grid of predictionpointsnearthefaceandaplanlocated37.5mdistant(Figure103).Theroadwaywas5.0m wideand3mhigh,andaPoissonsratioof0.25wasused(moredetailsinAppendixB).TheSuuandSue stress tensors were imported into an Excel spreadsheet to be manipulated to generate data files of positive BPXS for a range of friction angles. Simple incremental linear assumptions were used to integrate the area under the curves. Because this is a linear elastic code, the various stress componentscanbenormalizedinthiscasetothemagnitudeoftheverticalstress.

Planestrain

Miningface

Figure103Examine3Dgeometry

The approach can result in a variety of plots. Figure 104 shows the distribution of BPXS that will develop after the face has advanced 2.31m for 3 different heights into the roof. The peak values are about 0.5m from the rib line, which is closer to the rib line when compared to the simple beam models (Stimpson, 1987). The data can be further manipulated to show the progressive increase in BPXSwithdistancefromtheribline(Figure105)

109

2.31M FROM FACE LINE

0.6 0.2M INTO ROOF BPXS @ 35 degree friction angle (MN) 0.5 0.4M INTO ROOF 0.6M INTO ROOF

0.4

0.3

0.2

0.1

0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Distance across roadway (m)

Figure104BPXSat2.31mfromtheface
1 0.9 0.8 INCREMENTAL BPXS (%) 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1 1.5 2 2.5 DISTANCE FROM RIB LINE (m)

Figure105CumulativeincreaseinBPXSwithdistancefromtheribline

The average BPXS decreases as an inverse function of the distance from the face raised to the fourth power (Figure 106). For typical bolting locations (about 2.3m to 2.7m from the face), the BPXS represent about 2025% of the plane strain value. Figure 107 shows how the BPXS varies with both height into the roof and timing of the bolting. It can be seen that there is a ridge of high values about0.5m1.0mintotheroofforaboutthefirst2mofthedriveage.Attypicalboltinglocations,the ridgeisverymuchlessprominent.

110

1.6 Average BPXS at 35 degree friction (MNm) 1.4 1.2

Height into roof (m)

1 0.8 0.6 0.4 0.2 0 0 2 4 6

0.2 0.4 0.6 0.8 1 1.2

10

12

Distance from the face line (m)

Figure106AverageBPXSacrosstherooflinefordifferentheightsintotheroofanddifferentlocationsofbolts(areaunder curvesinFigure107expressedasuniformlydistributedload
1.5

Height into roof (m)

0.5

0.5

1.5

2.5

3.5

4.5

5.5

Bolting location with respect to face (m)

Figure107BPXSasafunctionofthelocationandheight

Inthefollowinganalysis,theBPXSdevelopedontwobeddinghorizonsarediscussed0.2mand0.4m above the roof line (dataset in an appendix). The data was manipulated in the context of the followingrelationship: BPXS(kN/m)=XS*verticalstress(MPa)*Ffriction*Fbolttiming, Equation18

where XS is a function of theheight into roof and the K values acting normal and parallel to the roadway, Ffriction is a correction for different friction angles, and Fbolt timing is a correction forboltlocation. The XS values are presented in Figure 108 (0.2m into roof) and Figure 109 (0.4m into roof) for the caseofafrictionangleof35oand2.3mfromtheface.

111

2.5

2.5

2 Parallel/vertical
Parallel/vertical

1.5

1.5

0.5

0.5

0.5

1.5 Normal/vertical

2.5

0.5

1.5 Normal/vertical

2.5

Figure108XSfactorfor0.2mintotheroof

Figure109XSfactorfor0.4mintotheroof

The Ffriction parameter is itself a function of the K values suchthat Ffriction= 1+(G *(35x)/25) where G is given in Figure 110 and x is the design friction angle. The Fbolt timing parameter (Figure 111) does vary withtheKratios,butatthisstageithasnotbeenfurtheranalysed.
3

10

2.5

Normal-parallel-vertical

Bolting location factor

2 Parallel/vertical

2.5-1.5-1 1.5-1-1 2.0-1.33-1

1.5

0.5

0.1
0 0 0.5 1 1.5 Normal/vertical 2 2.5 3

Distance from the face line (m)

Figure110Gparametertoaccountfordifferent frictionangles

Figure111Thebolttimingfactor

8.4.2.2 RESISTING FORCES

The deformations that develop close to the roof mean that it should be assumed that the bedding surfaceswillbeopen.Pragmaticallythereisaneedtoclosetheopenbeddingsurface,otherwiseonly therelativelyinefficientdowelresistancewillbeavailable. Based on observations of the roof lifting and adjacent bolts losing tension made during the introduction of pretensioned roof bolts in the Australian industry, it is assessed that correct installation of tensioned bolts can close bedding surfaces. It is not necessary to apply a preload against,notnecessarilytoapplysignificantpreloadacrosstheclosedsurfaceandtheshearstiffnessis such that full capacity is mobilized within a few milimetres of movement (Figure 112). Shear resistanceisgiveninFigure79.


Figure112Fullygroutedboltsareverystiffduringinitialshearloading(Stjern,1995)

8.4.3 DENSITIES AND PATTERNS

ThenumberofideallyplacedboltsiscalculatedbydividingtheBPXSvaluebythefrictionalresistance provided by the bolts. The number of bolts will depend on the strength of the bolts, the depth of cover, the K ratios, and the friction angle of the bedding. Expressed in terms of installed capacity (Section 4.2.1.2), there is good agreement between this BPXS model and current mining practice (Figure 113) especially if the mines with low RSI values are excluded. Friction angles of between 30o and35oareindicatedandthisisconsistentwiththerocktypesandjointroughness.
80 70 Primary Support Capcity (tonnes/m 2) 60 50 40 30 20 10 0 0 100 200 300 400 500 600 700 Depth (m) Stone roof 25 30 35 Linear (Stone roof) Low RSI values

Figure113ComparisonbetweenidealBPXSsupportcapacityat0.4mintotheroof(Kn=2.0,Kp=1.5,G=1.3)andtheAustralian database

Ideally the bolts should be placed so that they are equally loaded and Figure 105 can be used for guidance. According to this analysis, there is no advantage with centreline bolts as the locations are strongly biased towards the rib sides (Table 23) and this can result in a large unsupported span in the centre of the roadway that is bigger than the cut out distance. If the roof is thinly bedded this
111

span may not be adequately stable (Figure 100) indicating the need for mesh panels. The possibility ofadverseloadingonthepanelsmayrequireadditionalboltsortherelocationoftheidealtowards theroadwaycenterline.Theribsidebiashasbeenexaminedwithphysicalmodelsandthesecertainly showtheadvantageofabiastowardstotheribs(Figure114).
Table23BoltlocationstoresistBPXS

Pattern 2 4 6 8

Distancefromribline(m) 0.55 0.3,0.8 0.2,0.55,1.05 0.15,0.4,0.7,1.15

Centrelinespan(m) 4.1 3.6 3.1 2.9

Compromiselocations(m) 1.35 0.6,1.6 0.4,0.9,1.6 0.2,0.5,1.0,1.7

Figure114Optimumboltpatternsfromphysicalmodels(Stimpson,1987)

Thefollowingisanexampleofhowthesupportdesigncouldprogress: Input: Depth of 250m, average density = 2.4 t/m3. L=1.6, M=1.4, roadway driven at 20o to the major principal horizontal stress. Thinly bedded mudstone roof, JRC = 4. Bolting conducted at 2.3m from face.Pretensioned22mmdiameterXgradebolts. Designvalues:Verticalstress=250*2.4*0.0098=5.88MPa,Kn=1.42,Kp=1.52.Assumeabasefriction angle of 25o, and a JRC contribution of 8o. Required beam thickness is at least 0.3m to carry 2m of surcharge. Output:XS=122(use0.4mheightchart),G=1.48,Ffriction=1.11,Fbolttiming=1.0,BPXS=796kN,Boltfriction= 340*tan32=212kN,Anchorlength=0.9m. Interpretation: The ideal bolt density is 3.75 bolts per metre of roadway advance. A 4 bolt pattern is recommended. The thinly bedded mudstone could mean large loads on the mesh panel, so less than ideal bolt locations (in terms of BPXS) is recommended. Bolt locations of 0.6m and 1.6m from each ribs is recommended. The centre span is 2.0m and some loading on the mesh is anticipated. An outbye centre bolt could be considered if there is concern about mesh loading. A minimum bolt length of 1.4mispossible,butprecedentpracticeistheuseof1.8mbolts.1.8mboltscanbeinstalledinasingle passoperation.

112

8.4.4 COMMENTARY ON DESIGN TO PREVENT DELAMINATION

Based on the authors experiences, the voussoir beam tool is appropriate for estimating spanning thickness. The tool can also be used to identify the requirements for a spanning unit higher in the roof if a suspension strategy is to be adopted (e.g. if destressed coal is to be suspended from a stone roof). A key aspect is the estimation of the surcharge, which should include the anchorage length of theboltandanysuspendedload.Atthisstage,itisconsideredthatafactorofsafetyof1.52.0should beappliedtotherequiredthicknessthismeansathicknessof0.4mmaybeappropriate. Itisnotedthatyoucouldincludethemagnitudeofthehorizontalstressinthebeamcalculationifyou wishtheinputUCSvalueistheavailablecompressivecapacity(UCSh). One of the implications ofthe BPXSanalysis is that bolting densities reduce ifbolting can bedeferred until the roadway is advanced. The voussoir beam analysis can be used to assess the risk of loss of roof skin with span in this case being from face to last line of bolts. If risk is acceptable, the bolting densitiesaresubstantiallyreduced.ThisisprobablywhytheUSAindustriesusesgenerallylowerbolt densitiesasthatindustryusesextendedcutminingtoamuchgreaterlevelthanAustralia. The design tool for bedding plane shear assumes that the roof beam is exposed to the elastic stress field in a continuum with no relaxation due to roof deflection. While this appears to be a major simplification,thevalidationtocurrentminingpracticeindicatesthatitisreasonablyvalidatleastfor aheightof0.4m.Itappearsthatfactorsofsafetyclosetounitymaybeappropriate. The XS value varies by no more than 20 % within the range of K values encountered in the Australian industry. This suggests that there is no real advantage in having different support densities in headingscomparedtocutthroughs. The use of ultimate strength of the bolts is acceptable, if you can rely on their post yield and post failure capacity. The greater the tolerable deformations in the system, the more likely that the immediateroofstresseswillrelax. TheBPXStoolshouldonlybeappliedforinitialroadwaydevelopmentandnotatthemaingatecorner because there is little doubt that the relaxation of roof stresses has developed. If there was no such relaxation, the maingate BPXS magnitudes would be similar to those shown in Figure 110 with the bolt timing being set at 0m say 5 times greater than those used in the design for the development face.Ifthiswasthecase,oneimplicationwouldbethatmaingateroofswithoutcablesupportwould fail and this is demonstrably not the case in Australia or other countries. It is noted that Seedsman (2000)hadnotidentifiedthisrelaxationmechanism. Deterioration of the maingate conditions can be explained increase in height of stress arch and the resulting increase in the vertical surcharge to the beam when the RSI is low. This vertical surcharge induces compressive stresses in the roof beam that has been created in the softened zone by the roofboltssothattheobservermayseeadditionalguttering. It is possible that the reports of stress notching in intersections are a reflection of changes in the heightofthestressarchandnotincreasedhorizontalstressesattheroofline. ItisnotedthattheroadwaywidthusedintheBPXSanalysiswas5.0m.Theresultscanbereasonably applied to roadways with widths between 4.5m and 5.5m. They should not be directly extrapolated

113

to wider roadways because the impact both in terms of the wider roadway and the bolt location factorhasnotbeenexamined.

114

PREVENTATION OF RIB COLLAPSE

Asimilarlogicalframeworkhasbeendevelopedforribs(Figure115).Forribs,thereare3stepsinthe framework. Different from roof, ribs undergo a simple stepwise increase in the applied stress field from development to the maingate and then the tailgate. At one level, it is appropriate to consider the hazards of coal ribs in the context of a vertical pitwall. This analogy is not sufficient so it is then necessarytoconsiderthepossibilityoftheonsetofadditionalfailuremodesrelatedtotheadditional verticalstressesthatribscarrycomparedtorockfaces. Thestepsare: 1. Check the height of the roadway. Rib support is concerned with falls from height, with the greatest concern with falls from above shoulder level. Assuming that roadway width is such thatreasonableaccessisavailablearoundthevariouspiecesofminingmachineryandhence a relativelyclear workspace,a roadwayheight of1.8m isconsidered tobe a reasonable limit forpatternribsupport. Checkforalignmentwithrespecttocoaljointsandcleat. Checkforthepossibleonsetofmininginducedfailure.Notethatthismaynotdevelopifthe alignment is such that joints and cleats dominate behaviour. Also note that this failure may betimeandminingsystemdependent.

2. 3.

Extraction height greater N than 1.8m Y

Pattern support not required

More than 20 degrees to N cleat Y Secure intersections

Consider realignment of roadway Possibiity of large wedge and planar slides Bolt length based on slide geometry

Strength/vertical N greater than 5.0 Y

Mining-induced fracturing will develop immeidately or over time

Pattern support not required

Figure115Logicalframeworkforsupportofribs

115

9.1

STRUCTURE CONTROL

Inoneaspect,ribscanbeconsideredasverticalpitwalls(Figure116).Theyarethereforeexposedto the same hazards of jointcontrolled failure with planar slides, wedges and topples. Highly broken coal, for example in a fault zone, could also generate a circular slip mode. Numerous papers and books discuss planar slides and wedges and there are software programs available (such as Rocplane andSwedge,www.rocscience.com).

Figure116Failuremodesforrockslopesthatcanbeobservedincoalribs(HoekandBray,1981).

In the rock slope engineering, there is a preference to avoid noses in pit walls as these protuberances are unstable due to the lack of lateral restraint. In coal mine roadways, all intersectionsarenoses(Figure117)andmayneedspecialattention.

116

 Figure117Allintersectionsarenoses

9.1.1 SLIDES
The key variables in a planar slide analysis are the dip and friction angle of the discontinuity. Recognizing the low heights, the stresses are low and JRC can become a significant contributor to friction angle. The JRC for coal surfaces ranges between 4 and 14, averaging 8 (Table 9). Thus, for a typicalribgeometry,thefrictionangleincrementrelatedtoJRCmaybeupto20o. Figure 118 shows the installed bolt capacity to achieve a factor of safety of 1.5 using with passive bolts. The installed capacities are only in the order of 1 tonne/m advance to 2.5 tonnes/m advance for a 3m high rib. For different rib heights, the loads can be reduced according to the square of the height.
Bolt force at 3.0m high Factor of safety without passive bolts Mass at 3m 25

30

10 Factor of safety without bolts or fall mass (tonnes/m)

9 8 7 6

Bolt force (kN/m)

20

15

5 4

10

3 2 1

0 40 45 50 55 60 65 70 75 80 85 Dip of plane with friction angle of 45o

Figure118Planargeometryandrequiredfacesupportfora3mhighrib

117

9.1.2 WEDGES
At more than 20o to the strike of a discontinuity, there is still the possibility of a wedge falling from theribline.Themaximumdimensionsofthewedgeareintheorderof5mlongandabout1.5minto the face (Figure 119). The required bolt capacity is in the order of 0.5 tonnes/metre of roadway advancefora3mhighrib.
Length (m) 10 Depth (m) Mass (tonnes) 9 8 kN/metre 7 6 5 4 3 2 1 0 40 45

10 9 8 7 6 5 4 3 2 1 0 10 15 20 25 30 35

Angle between face and strike of one of the conjugate joint sets
Figure119Wedgegeometryandrequiredfacesupport

9.1.3 TOPPLES
Sagaseta et al (2001) provides a general solution for the anchor force to limit toppling failures. They arguethattheanchorforce(Ft)isgivenbythefollowingrelationship: Ft=0.5*kTH2

Equation19

Where = density, H is face height, and kT is a dimensionless factor which is provided in a series of designcharts.Forthegeometriesofinterest,therangeofkTvalueis0.2to0.3. The anchor force to prevent toppling is therefore in the order of 1 tonne/m to 2 tonnes/m (Figure 120).
30

25

Installed capacity (kN/m)

20

15

10
0.2 0.3

0 1.8 2 2.2 2.4 2.6 Rib height (m) 2.8 3 3.2 3.4

Figure120Anchorforcetopreventtoppling

118

Installed passive support to give FoS = 1.5 (kN/m)

Length and depth of rib impacted (m) or weight of wedge (tonnes)

9.2

STRESS INDUCED RIB COLLAPSE

9.2.1 MINING INDUCED FRACTURES

OBierne et al (1986) discussed mining induced fractures in coal ribs. Brittle rock behaviour provides thetheoreticalbackgroundtotheseobservations. As discussed in Section 5.3.4, there are 3 components to the brittle rock failure criterion the onset of failure related to the loss of cohesive strength, a spalling limit, and then the standard Hoek Brown criterion in the far field. The far field in this context may be only a few metres away from the excavation. Readily available software cannot yet handle this composite criterion, so there is a need toapplyitstepwise. Figure 121 shows the distribution of the strength factor with respect to the brittle parameter for the case of 20 MPa coal with vertical stresses of 5 MPa and 10 MPa. This simple analysis would suggest that rib failure would progressively extend well into the rib with increasing depth. It is known that thisisnotthecaseasobservationsofribspallatdepthsuggestthatitisstillonlyaskineffect.Colwell (2004) examined rib data over a depth range of 125m to 525m and reported that average rib bolt lengthwas1.2mandthemaximumlengthwas1.8m.

(a)5MPaverticalstress (b)10MPaverticalstress

Figure121Distributionofthecohesivelosscomponentaboutroadways

Theexplanationforthisdiscrepancyistheneedtoconsiderthespallinglimit.Thespallinglimitisthe ratio of the magnitude of the major and minor principal stress and hence is independent of any strengthparameter.Thedistributionofthespallinglimitwithdepthatmidheightisafunctionofthe K ratio and the aspect ratio of the opening, and in general is reaches a value of 10 somewhere between0.6mand1.1mintotherib(Figure122). SimilartotheRSIconcept,itispossibleto manipulatethedatainthecontextofacoalstrengthindex (CSI) to produce a design chart for the depth of brittle failure (Figure 123). The shape of the failure zones can be obtained by inspection of Figure 121a. Pattern rib support may only be required for failuredepthsinexcessofabout0.1m,oraCSIofabout7.5.

119

20 18 16 14 Spallingratio 12 10 8 6 4 2 0 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 Depthintorib(m)
K=0.1,5mby3m K=0.2,5mby3m K=0.4,5mby3m K=0.1,5mby2m K=0.2,5mby2m K=0.4,5mby2m

Figure122ValueofspallingratioalongthespringlineoftheribfordifferentKratiosandroadwayaspectratios

1.2

1 Depth of brittle failure (m)

maximum depth set by spalling limit = 10

0.8

0.6

0.4

0.1, 5m 0.2, 5m 0.3, 5m 0.4, 5m 0.1, 5m 0.2, 5m 0.3, 5m 0.4, 5m

by 3m by 3m by 3m by 3m by 2m by 2m by 2m by 2m

0.2

0 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 Coal strength index

Figure123Estimationofmaximumdepthofbrittlefailure

9.2.2 BUCKLING
Forroadwayorientationsthatarewithinsay20oofthetrendofacoaljointset,thereisapossibilityof bucklingofcoalbeamsifthejointsarecloselyspaced.Eulerbucklingconceptsoffersomeinsightinto themechanismsinvolved,withthelimitationthatthereisalargedegreeofuncertaintyregardingthe imposedloading. The axial loading of slabs was discussed by Hoek and Brown (1980) who provided the following relationshipfortheaxialstress()atwhichtheplatewillbuckle:

120

=2E/(12q2(l/t)2

Equation20

where E=modulusofelasticity, l=heightoftheslab, t=thicknessofslab,and q = constant with values of 1 for both ends pinjointed, and 0.5 for both ends clamped. The axial stress applied to a joint bound slab at the side of a roadway is difficult to accurately determine due to the development of yielding in the coal and any low strength bands. A simple 2 dimensionalelasticanalysissuggeststhattheverticalstressattheriblinecanincreasebyabout220% (Table18,Figure57).Usingthisresult,Figure124quantifiestherelationshipbetweenslabthickness, extracted coal thickness and depth. At shallow depths, slabs less than about 200mm thick may buckle,andthisincreaseswithdepthto400mmatdepthsinexcessof400m.

0.6
Simple beam, pin jointed E= 1500 MPa Vertical stress related to depth of cover and density of 2.5 t/m3 , and an abutment peak at 2.16 the average stress

0.5

Buckling thickness (m)

0.4

0.3 DEPTH 200 300 0.1 400 500 0 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4

0.2

Roadway Height (m)

Figure124Thicknessofslabsthatmayundergobuckling.

9.3

COMMENTARY ON RIB SUPPORT DESIGN

Hoek and Bray (1981) suggests that the kinematic acceptability of planar slides is reduced if the rock face in an open excavation is aligned at more that 20o from the strike of the joint. This recommendation should be considered to be a minimum in the underground coal environment as there may still be concerns with the practicalities of the installation of rib support at greater angles. Theauthorshavefoundthatanangleofabout35owouldbepreferred. Thediscussionaboutcoalstresses(Section5.5.2)speculatedthattheremaybeareductioninvertical stresstobelowthatrelatedtodepthofcoverattheminingfaceandthatthisincreasesoutbye.Ifthis

121

is the case, there may be an apparent time dependent deterioration of the ribs related to brittle failureastheCSIreducesastheverticalstressincreases. The Euler analogy needs to be applied with care as it requires that the loading system is of low stiffnesssuchthatthereisnostressrelaxationiftheribcompressestoanydegree.Iftheribslabsare formedbyabrittlefailuremechanism,thedevelopmentofmoreintensefailurezonesattheroofand ribcornerwouldnotbeexpectedtotransferthesamelevelsofverticalloading.Themechanicsbeing invokedarethesameasthosethatdrivetherelaxationofthesoftenedzoneintheroof.

122

10 REFERENCES
ACIRL.(1986).Evaluationofcoalpillarstabilityandextractionmethods.InACAEndofGrantReport. AusIMM,(1995).FieldGeologistManual(3rde.d.).(p.390).TheAustralasianInstituteofMiningand Metallurgy.Carlton,Australia. Barton,N.R.(1973).Reviewofanewshearstrengthcriterionforrockjoints.EngineeringGeology8, 287322. Barton,N.R.(1980).Evaluationofshearstrengthinrockfillandbetweenrockfillsandrock foundations.NGIreport. Barton,N.R.,&Choubey,V.(1977).Theshearstrengthofrockjointsintheoryandpractice.Rock Mechanics1/2:154. Barton,N.,LienR.,&Lunde,J.(1974).Engineeringclassificationofrockmassesforthedesignof tunnelsupport.RockMechanics6(4),163177. Bieniawski,Z.T.(1976).(Eds.Bieniawski,Z.T.)Rockmassclassificationsinengineeringpractice.In Explorationforrockengineering.1,97106.CapeTown:AABalkema. Bieniawski,Z.T.(1993).(Eds.Hudson,J.,Brown,E.T.,Fairhurst,C.,&Hoek,E.).Designmethodology forrockengineering:Principlesandpractice.ComprehensiveRockEngineering,2:77993. Oxford:Pergamon. Bjurstrom,S.(1974).Shearstrengthofhardrockjointsreinforcedbygrouteduntensionedbolts. Proceedings3rdISRMCongress,Denver,11941999. Brady,B.H.G.&Brown,E.T.(1985).Rockmechanicsforundergroundmining.London,UK:George Allen&Unwin. Brown,E.T.(2008).EstimatingtheMechanicalPropertiesofRockMasses.SHIRMS2008(EdPotvin,Y. Carter,J.,Dyskin,D.andJeffrey,R),Perth,317. Canmet.(1977).PitSlopeManual,Chapter6,MechanicalSupport.Ottawa:Canmet. Coates,D.F.(1970).Rockmechanicsprinciples.MinesBranch,DepartmentofEnergy,Minesand Resources.Monograph874. Colwell,M.G.(1998).Chainpillardesign:CalibrationofALPS.(p.67).FinalReport:ACARPProject C6036. Colwell,M.G.(2004).Analysisanddesignofribsupport(ADRS):Aribsupportdesignmethodologyfor Australiancollieries.(p.325).Finalreport:ACARPProjectC11027. Colwell,M.G.,&Frith,R.(2006).WhyuniaxialcompressivestrengthandYoungsmodulusare commonlypoorindicatorsofroadwayroofstabilityexceptinthetailgate.Paperpresentedin CoalOperatorsConferenceatUniversityofWollongong,Illawarrabranch,AusIMM. Cram,K.(September,2008).ProductionfromAustralianlongwallmines.AustralianLongwall Magazine. Diederichs,M.S.,&Kaiser,P.K.(1999).Tensilestrengthandabutmentrelaxationasfailurecontrol mechanismsinundergroundexcavations.InInternationalJournalofRockMechanicsand MiningSciences:36,6996.

123

Diederichs,M.S.,Kaiser,P.K.,&Eberhardt,E.(2004).Damageinitiationandpropagationinhardrock duringtunnelingandtheinfluenceofnearfacestressrotations.InInternationalJournalof RockMechanicsandMiningSciences:41,785812. Donaldson,A.C.(1979).(Eds.Donaldson,A.C.,Presley,M.K.,&Renton,J.J.)Originofcoalseam discontinuities.(p.102132).InCarboniferousCoalGuidebook.WVGeologicalSurvey,VolB37 1. Enever,J.R.&Doyle,R.(November,1996).(Eds.NcNally,G.H.,&Ward,C.R.).Theroleofhorizontal stressinlongwalldesign:AcasehistoryatDartbrookColliery,NSW.(p.143151).Symposium onGeologyinLongwallMining,1213. Enever,J.R.,Jeffrey,R.G.&Casey,D.A.(2000).(Eds.Girard,J.,Liebbman,M.,Breeds,C.&Doe,T.). Therelationshipbetweenstressincoalandrock.(p.409414).PacificRocks2000.Rotterdam: Balkema. Fookes,P.G.,Baynes,F.J.,&Hutchinson,J.N.(2000).Totalgeologicalhistory:Amodelapproachto theanticipation,observation,andunderstandingofsiteconditions.PaperpresentedinGeoEng 2000,AninternationalconferenceongeotechnicalandgeologicalengineeringinMelbourne. Lancaster,USA:TechnomicPublishingCompany. Gale,W.J.(2008).Stressissuesinundergroundcoalminesanddesignapproach.Paperpresentedin StressMeasurements,MonitoringandModellingTechniquesandtheirDesignApplications, Wollongong:EasternAustraliaGroundControlGroup. Gale,W.J.&Matthews,S.M.(1992)Stresscontrolmethodsforoptimiseddevelopmentand extractionoperations.ReporttoNationalEnergyResearchDevelopment,andDemonstration (NERD&D)Program,Project1301. Gordon,N.&Tembo,E.(2005).Theroofstrengthindex:Asimpleindextoonepossiblemodeofroof collapse.PaperpresentedinBowenBasinSymposium.(Eds.Beeston,J.W.)Yeppoon, Queensland. Haile,A.(2004).Areportingframeworkforgeotechnicalclassificationofminingprojects.Bulletinof theAusIMM.September,p.3037. Hanna,T.H.(1982).Foundationsintension:groundanchors.TransTechPublicationSeriesonRock andSoilMechanics,Volume6. Hasenfus,G.J.&Su,D.W.H.(2006).Horizontalstressandcoalmines:Twentyfiveyearsexperienceand perspective.(p.256267).Paperpresentedin25thInternationalConferenceonGroundControl inMining. HergetG.(1988).Stressinrock.Balkema. Hillis,R.R.,Enever,J.R.&Reynolds,S.D.(1999).InsitustressfieldofeasternAustralia.Australian JournalofEarthSciences.46,813825. Hobbs,D.W.(1967).Theformationoftensionjointsinsedimentaryrocks:Anexplanation.Geology Magazine,104(6),550556. Hobst,L.&Zajic,J.(1983).Anchoringinrockandsoil.(p.570).DevelopmentsinGeotechnical Engineering,33.Elsevier,Amsterdam. Hoek,E.&Bray,J.W.(1981).Rockslopeengineering.3rdEdition.London:InstnMin.Metall.:London. Hoek,E&BrownE.T.(198).UndergroundExcavationsinRock.InstnMinMetall.:London. Hoek,E&BrownE.T.(1997).Practicalestimatesofrockmassstrength.InternationalJournalofRock MechanicsandMiningSciences,34(8),11651186

124

Hoek,E.,Marinos,P.G.&Marinos,V.P.(2005).Characterisationandengineeringpropertiesof tectonicallyundisturbedbutlithologicallyvariedsedimentaryrockmasses.International JournalofRockMechanicsMiningandSciences,42(2),277286. Hoek,E&Diederichs,M.(2006).Empiricalestimatesofrockmassmodulus.InternationalJournalof RockMechanicsandMiningSciences,43,203215. ISRM(InternationalSocietyforRockMechanics)CommissiononStandardizationofLaboratoryand FieldTests.(1978).SuggestedMethodsforthequantitativedescriptionofdiscontinuitesin rockmasses.InternationalJournalofRockMechanicsandMiningSciences,5,316368. Jaeger,J.C.,&Cook,N.G.(1979).FundamentalsofRockMechanics.London:ChapmanandHall. Jemeric,M.L.(1985).Stratamechanicsincoalmining.Rotterdam:A.A.Balkema. Ji,S.,Zhu,Z.,&Wang,Z.(1998).Relationshipbetweenjointspacingandbedthicknessinsedimentary rocks:Effectsofinterbedslip.GeologyMagazine,135(5),637655. Kaiser,P.K&Kim,BH.(2008).RockMechanicsChallengesinUndergroundConstructionandMining. SHIRMS2008(EdPotvin,Y.Carter,J.,Dyskin,D.andJeffrey,R),Perth,2338. Lambe,T.W.(1973).Predictionsinsoilengineering.Geotechnique,23(2)149202. Littlejohn,S.(1993).Overviewofrockanchorages.ComprehensiveRockEngineering,4:413450. Oxford:Pergamon. Mark,C.(1990).PillarDesignMethodsforLongwallMining.UnitedStatesBureauofMines InformationCircular9247,53p. Mark,C.,&Molinda,G.M.(2005).Thecoalmineroofrating(CMRR):Adecadeofexperience.Coal Geology,64,85103. Mark,C.,Gale,W.,Oyler,D.,&Chen,J.(2007).Casehistoryoftheresponseofalongwallentry subjectedtoconcentratedhorizontalstress.InternationalJournalofRockMechanicsand MiningSciences,44,210221. Martin,C.D.,Kaiser,P.K.,&McCreath,D.R.(1991).HoekBrownparametersforpredictingthedepthof brittlefailurearoundtunnels.CanadianGeotechnicalJournal,36:136151. McCabe,K.W.,&Pascoe,W.(1978).Sandstonechannels:Theirinfluenceonroofcontrolincoalmines. (p.24).MineSafetyandHealthAdministration,IR1096. Medhurst,T.P.(1999).Highwallmining:Practicalestimatesofcoalseamstrengthandthedesignof slenderpillars.TransInstnMinMetall(SectAMinIndustry),108,A161A171. Medhurst,T.P.,&Brown,E.T.(1998).Astudyofthemechanicalbehaviourofcoalforpillardesign. InternationalJournalofRockMechanicsandMiningSciences,35,10871105 Molinda,G.M.(2003).Geologicalhazardsandroofstabilityincoalmines.USDepartmentofHealth andHumanServices,NIOSH,PittsburghResearchLaboratory,InformationCircularIC9466, Publicationnumber2003152. Molinda,G.M.,&Mark,C.(1994).Thecoalmineroofrating(CMRR):Apracticalrockmass classificationforcoalmines.UnitedStatesBureauofMinesIC9387. Morgenstern,N.R.(2000).Performanceingeotechnicalpractice.PaperpresentedattheInaugural LumbLecture,10May,HongKong. Muirwood,A.M.(1972).Tunnelsforroadsandmotorways.QuartJEngGeol,5,111126. Nemcik,J.A.,Gale,W.J,&Fabjanczyk,M.W.(2006).Methodsofinterpretinggroundstressbasedon undergroundstressmeasurementsandnumericalmodelling.PaperpresentedinCoal2006,7th UndergroundCoalOperatorsconference.104112.
125

OBeirne, T, Shepherd, J, Rixon, K and Napper, A, 1986. Coal rib stabilisation a new perspective. The CoalJournal,pp711. Palmstrom,A.,&Broch,E.(2006).Useandmisuseofrockmassclassificationsystemswithparticular referencetotheQsystem.TunnellingandUndergroundSpaceTechnology,21,575593. Pattison,C.I,Fielding,C.R,McWatters,R.H.&Hamilton,L.H.(1995).NatureandOriginofFracturesin PermianCoalsfromtheBowenBasinandtheRelationshipwithinsitureservoirpermeability. BowenBasinSymposium,Mackay.217234. Payne,D.(2007).CrinumMine15longwallsinweakroof.Whatdidwelearn?.Paperpresentedin Coal2006,7thUndergroundCoalOperatorsconference.2241. Peck,R.B.(1969).Advantagesanddisadvantagesoftheobservationalmethodinappliedsoil mechanics.Geotechnique,19,171187. Pells,P.J.N.(2002).DevelopmentsinthedesignoftunnelsandcavernsintheTriassicrocksofthe Sydneyregion.InternationalJournalofRockMechanicsandMiningSciences,39,5,569588 Pells,P.J.N.,&Bertuzzi,R.(2007).Limitationsofrockmassclassificationsystems.Tunnelsand TunnellingInternational,April,3337. Peng,S.S.(1986).CoalMineGroundControl.(p.491).JohnWileyandSonsInc. Potvin,Y.&Nedin,P.(2003).Managementofrockfallrisksinundergroundmetaliferousmines:A referencemanual.MineralsCouncilofAustralia Sagaseta,C.,Sanchez,J.J.,&CanizalJ.(2001).Ageneralanalyticalsolutionfortherequiredanchor forceinrockslopeswithtopplingfailure.InternationalJournalofRockMechanicsandMining Sciences,38(3),421435. Schach,R.,Garshol,K.,&Heltzen,A.M.(1979).Rockbolting:Apracticalhandbook.(p.84).Oxford: PergamonPress. Seedsman,R.W.(2000).Progressinthedevelopmentofaroofboltdesignmethodologybasedon resistingshear.Paperpresentedin19thInternationalConferenceonGroundControlinMining. 272277. Seedsman,R.W.(2003).Roofsupportingateroadsinmultipleseamlongwalls:Lessonsfromultra closeminingatGibsonsColliery.PaperpresentedinCoal20034thAustralasianCoalOperators Conference,Wollongong. Seedsman,R.W.(2004).(Eds.Villaescua,E.,&Potvin,Y.).Failuremodesandsupportofcoalroof. PaperpresentedinGroundSupportinMiningandUndergroundConstruction.Proceedingsof the5thInternationalSymposiumonGroundSupport,Perth.AABalkema. Shen,B.,Guo,H.,King,A.&Wood,M.(2006).AnIntegratedRealtimeRoofMonitoringSystemfor UndergroundCoalMines.PaperpresentedinCoal20067thAustralasianCoalOperators Conference,Wollongong. Shepherd,J.&Gale,W.(1982).Collieryroofstabilityandtheroleofgeology.TheAustralianJournal ofCoalMiningTechnologyandResearch,No1,4767. Sofianos,A.&Kapenis,A.P.(1998).Numericalevaluationoftheresponseinbendingofan undergroundhardrockvoussoirbeamroof..InternationalJournalofRockMechanicsand MiningSciences,35(8),10711086. Stimpson,B.(1987).Optimisingthereinforcementeffectoffullcolumn,untensioned,groutedboltsin beddedmineroof.InternationalJournalofMiningandGeologicalEngineering,5,285298.

126

Stjern,G.(1995).PracticalPerformanceofRockBolts.DoktorIngeniorThesis,InstitutttforGeologi Bergteknikk,norgesTekniskeHogskole,UniverstitetetiTrandheim. Tarrant,G.(2006).Skewroofdeformationsmechanisminlongwallgateroads:Conceptsand consequences.PaperpresentedinCOAL2005,6thAustralianCoalOperatorsConference. Brisbane,AusIMM. Thompson,A.G.(2004).Rocksupportactionofmeshquantifiedbytestingandanalysis.Paper presentedinSurfacesupportinMining.(Eds.Potvin,Y.,Stacey,D.,&Hadjigeorgiou,J.) AustralianCentreforGeomechanics,pp391398. Zipf,R.K.(2005).Failuremechanismsofmultipleseammininginteractions.Paperpresentedin24th InternationalConferenceonGroundControlinMining.93106.

127

APPENDIXAVOUSSOIRBEAMSPREADSHEET By implementing the spreadsheet as detailed below, you acknowledge that you appreciate the limitations of the method as outlined in Sofianos, A. & Kapenis, A.P. (1998). Numerical evaluation of the response in bending of an underground hard rock voussoir beam roof. International Journal of RockMechanicsandMiningSciences,35(8),10711086.

UCS,modulus,density(C6,C4,C5)

Surchargethicknessanddensity(C7,C8)

thickness(C3)

span(C2)
Span Beam thickness Modulus Density UCS Surcharge thickness Surcharge density INPUT 70 10 7.5 25 30 5 24 Output =C18/C23 =C19/C23 mm =C30*C27*1000 calculations check 0.1-0.3 =0.3-0.14*C16*POWER(C23,0.333) =C2/C3 Equivalent surcharge =C7*C8/C5 =4*C14*C6/C4/1000/C26*SQRT((10.5*C14*C6/C4/1000*C24*C26*C26)) =3/C26/C26/C26 =(C3+C17)/C3*C5*C3 =(C17+C3)/C3*C5*C2/C4/1000000 =0.2+0.06*SQRT(C16)+0.0005/SQRT(C23) =1-2*C14/3 =C2/C27 =C25*C3 =C23*C26*C26*C26/16 =1/3*ATAN(SQRT(1/27-C28*C28)/C28) =1-COS(C29)/SQRT(3)-SIN(C29)

m m Gpa kn/m3 MPa m kN/m3 Crushing factor Buckling factor Deflection n Sn m equation 9 buckling Q Qn lamda Zon Sz Zo delzo omega delz

128

APPENDIXBEXAMINE3DMODELLING ThisappendixprovidesmoredetailsonsomeoftheExamine3dmodelling. FigureB1showsthedifferencebetween5m,5.5m,and8mwideroadways.Theplotssupporttheuse oftheBPXSplotsforroadwaysbetween4.5mand5.5minwidth,butnotforroadwaysthatarewider.

2

1.5

0.5 8M WIDE 0 0 2 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

1.5

0.5 5.5M WIDE 0 0 2 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

1.5

0.5 5M WIDE 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

FigureB1Resultsfordifferentroadwaywidths,assuming3mheight,1=12MPa,2=9MPa,and3 =v=6MPa,roadwayparallelto1,andPR=0.25

129

FigureB2 shows the effect of Poissons ratio. For Bottling at 2.31m anda 35 friction angle, the value of the BPXS at 0.4m into the roof is 0.906, 0.863, and 0.838 for Poissons Ratio values of 0.35, 0.25, 0.15respectively.
2

Height into roof (m)

1.5

0.5

0.15
0 0 0.5 1 1.5 2 2.5 Bolt location (m) 2 3 3.5 4 4.5 5

1.5

0.5

0.25
0 0 2 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

1.5

0.35
0.5

0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

FigureB2ResultsfordifferentPoissonsratios,assuming3mheightand5mwide,1=12MPa,2=9 MPa,and3=v=6MPa,roadwayparallelto1.

130

TableB1SummaryofBPXSdataasreferedtointhebodyofthereport.
Normal Parallel Vertical to to roadwa roadwa y y Height into roof (m0 Friction angle 1 1 1 1 3 3 3 3 3 3 6 6 6 6 6 6 6 6 6 6 6 8 9 9 9 9 9 9 9 9 12 12 12 12 12 12 12 12 12 12 15 15 15 18 18 18 1 3 6 12 1 3 6 9 12 24 1 3 6 9 9 12 12 12 15 18 24 12 3 6 9 12 12 15 18 24 1 3 6 8 9 12 12 15 18 24 6 9 12 1 3 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 12 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 12 6 6 6 6 6 6 6 6 6 BPXS (kN)

0.2 35 277 345 421 560 312 370 452 529 603 879 347 401 490 576 576 656 656 904 730 804 951 683 437 523 608 693 693 775 851 999 413 469 554 612 639 723 981 808 890 1045 588 671 756 476 534 601

0.4 35 387 483 599 820 410 490 607 724 833 1240 422 497 620 738 738 853 853 1214 962 1071 1278 862 511 628 746 863 863 976 1085 1304 440 519 637 715 753 870 1240 987 1099 1317 649 764 881 461 541 636

0.4 30 440 538 659 894 466 551 673 793 913 1356 490 571 700 820 820 940 940 1347 1060 1177 1402 956 597 722 842 962 962 1082 1201 1434 533 617 744 825 863 983 1399 1103 1222 1461 767 886 1006 574 660 765

0.4 25 497 588 705 932 526 606 724 842 955 1394 559 635 758 877 877 993 993 1449 1106 1219 1444 1017 671 789 908 1026 1026 1141 1254 1480 623 703 825 907 947 1069 1515 1191 1309 1539 859 980 1102 685 767 872

0.4 20 554 644 759 988 593 672 788 903 1019 1466 635 710 830 947 947 1063 1063 1575 1179 1294 1516 1090 754 870 986 1102 1102 1217 1333 1563 718 795 912 989 1027 1143 1660 1258 1374 1605 955 1070 1185 802 881 981

0.4 15 615 699 810 1029 658 734 846 956 1066 1499 711 784 900 1011 1011 1121 1121 1692 1230 1340 1558 1156 838 951 1062 1172 1172 1281 1442 1609 814 889 1001 1076 1113 1222 1799 1332 1441 1660 1053 1164 1274 915 991 1090

0.4 10 673 751 856 1066 720 791 897 1002 1108 1524 782 850 959 1065 1065 1171 1171 1794 1277 1382 1589 1211 913 1019 1125 1230 1230 1336 1442 1652 903 973 1079 1150 1184 1289 1918 1395 1501 1711 1139 1244 1350 1022 1093 1188

131

Normal Parallel Vertical to to roadwa roadwa y y Height into roof (m0 Friction angle 18 18 18 18 24 24 24 24 24 24 24 24 9 12 18 24 1 3 6 6 9 12 18 24 6 6 6 6 6 6 6 12 6 6 6 6

BPXS (kN)

0.2 35 703 788 957 1125 509 528 655 938 737 814 1021 1190

0.4 35 776 893 1127 1352 448 566 650 1037 765 873 1151 1385

0.4 30 908 1028 1267 1506 583 667 798 1234 921 1034 1312 1551

0.4 25 1013 1125 1380 1618 722 804 930 1404 1050 1166 1446 1690

0.4 20 1112 1228 1459 1690 864 941 1060 1590 1174 1284 1544 1775

0.4 15 1215 1325 1544 1763 998 1072 1186 1777 1297 1403 1647 1866

0.4 10 1304 1409 1620 1831 1126 1196 1303 1945 1407 1509 1740 1951

132