ContentslistsavailableatScienceDirectJournalofCleanerProductionjournalhomepage:www.elsevier.com/locate/jcleproModelingandevaluationofthesafetycontrolcapabilityofcoalminebasedonsystemsafetyQuan-longLiua,Xin-chunLib,*abManagementScienceandEngineering,SchoolofManagement,ChinaUniversityofMiningandTechnology,Xuzhou221116,ChinaComplexityResearchInstituteofEconomyandManagement,SchoolofManagement,ChinaUniversityofMiningandTechnology,Xuzhou221116,ChinaarticleinfoArticlehistory:Received30May2013Receivedinrevisedform10November2013Accepted16November2013AvailableonlinexxxKeywords:SafetycontrolcapabilitySystemsafetyBack-propagationneuralnetworkOffsetriskfactorsHazardfactorsabstractTheintrinsiclevelofcoalminesafetyisdirectlyrelatedtoitsconnaturalrisks,namelythehazardfactors.However,theoffsetriskfactors,namelythesafetycontrolcapability,ultimatelydeterminethedesirableoperatinglevel.Inthispaper,theoffsetriskfactorswerestudiedfromtheaspectsofpersonnel,material(machine),environmentandmanagementtocontrolcoalminesafetyincontrasttodescribingsystemrisk.Themodelwasconstructedanddescribedusingsafetyentropyandcasktheorybyfullyunder-standingthehazardfactorsandoffsetriskfactors.Thenasetofmodelevaluationindexeswerederivedbasedontheaboveanalysis.Aback-propagationneuralnetwork(BPNN)wasdevelopedtoevaluatethesafetycontrolcapability.Itwastrainedandtestedwithdatacollectedfromforty-onestate-ownedcoalminesinChina;thirty-sixwereusedtotrainthenetworkandtherestwereusedtotestit.Theresultsshowedthatsimulationperformancewasacceptableandthegoodnessoffitwashigh.Theweightsofpersonnel,material(machine),environmentandmanagementonthesafetycontrolcapabilityare0.26,0.29,0.22and0.23respectively.Finally,thetrainednetworkwasappliedtoassesstheWulanmulunmine’ssafetycontrolcapability.Theresultsindicatedahighlevelofsafetycontrolcapability(0.93).Ó2013ElsevierLtd.Allrightsreserved.1.IntroductionCoalaccountsformorethan70%ofChina’senergy.Thecoal-dominatedenergystructureisunlikelytochangeforalongperiodoftimebecauseChinaisrichincoalbutpoorinoilandgas(Li,2012).Inrecentyears,frequentcoalmineaccidentsnotonlycausedsignificantlossoflifeandproperty,butalsohadanegativeinfluenceonsociety.Althoughthetotalaccidents,deathtollanddeathratepermilliontonneshavedecreasedyearafteryear,thecoalminesafetysituationisstillverygrimcomparedwithsomedevelopedcountries.Therefore,coalminesafetyremainsoneofthemostsignificantoutstandingproblemsinChina.Coalminesafetylevelisdirectlyrelatedtoitsconnaturalrisks,namelythehazardfactors.However,theoffsetriskfactors,namelythesafetycontrolcapability,ultimatelydeterminethedesiredoperatinglevel.Fromtheviewofsystemsafety,determininghowtoimprovethesafetycontrolcapabilityfromtheaspectsofpersonnel,material*Correspondingauthor.Tel.:þ8613815319956(mobile).E-mailaddresses:715664665@qq.com,lxc_ljx@263.net(X.-c.Li).0959-6526/$eseefrontmatterÓ2013ElsevierLtd.Allrightsreserved.http://dx.doi.org/10.1016/j.jclepro.2013.11.048(machine),environmentandmanagementwillultimatelyassistimprovingminesafety.Atpresent,coalminesafetyevaluationisoneofthemostimportantissuesinthefieldofcoalminesafetyresearch.Mostresearchtodatehasfocusedonassessingtherisksposedbyhazardfactorsfromtheperspectiveofsystemrisk.Thisissometimesreferredtoasintrinsic.Table1summarizesthisliterature.Inthispaper,westudiedtheoffsetriskfactorstocontrolcoalminesafetyincontrasttodescribingsystemrisk.Weconstructedamodelofsafetycontrolcapabilityusingsafetyentropyandcasktheorybyfullyunderstandingthehazardfactorsandoffsetriskfactors.Asetofmodelevaluationindexeswerethenderived.Aback-propagationneuralnetwork(BPNN)wasdevelopedtoeval-uatethesafetycontrolcapability.Itwastrainedandtestedwithdatacollectedfromforty-onestate-ownedcoalminesinChina,amongwhichthirty-sixwereusedtotrainthenetworkandtherestwereusedtotestit.Therelativeimportanceofeachoftheoffsetriskfactorswasamended.Thisprovidedamechanismforgraduallyimprovingtheriskcontrollevel.Finally,thenetworkwasappliedtoassesstheWulanmulunmine’ssafetycontrolcapability.Pleasecitethisarticleinpressas:Liu,Q.-l.,Li,X.-c.,Modelingandevaluationofthesafetycontrolcapabilityofcoalminebasedonsystemsafety,JournalofCleanerProduction(2013),http://dx.doi.org/10.1016/j.jclepro.2013.11.0482Q.Liu,X.Li/JournalofCleanerProductionxxx(2013)1e6
Table1Literaturereviewonassessingtherisksposedbyhazardfactors.ResearchersResearchcontentsYear19721983198619881992RiskJerryS.RosenbloomAcasestudyinriskmanagement(Riskandinsuranceseries).DavidM.SiegelFormaldehyderiskassessmentforoccupationallyetal.exposedworkers.RobertA.BariDecisionmakingandprobabilisticriskassessment.NickF.PidgeonRiskassessmentandaccidentanalysis.DenbyBandKizilApplicationofexpertsystemsingeotechnicalriskM.Sassessment.W.Hatton.M.K.G.Whateleyassessmentapplied1995tocoaltonnageestimationintheUK.DuzgunH.S.BRiskassessmentforZonguldakcoalbasinundergroundmines.ChengLeietal.Studyofmineventilationsystemassessmentbasedonartificialneuralnetwork.ZhangXiao-yuandAssessmentofventilationsystemforundergroundDouShi-qingmines.HiromitsuSatisfyingsafetygoalsbyprobabilisticriskKumamotoassessment.WangWen-jieandFuzzyarrangementcomprehensivesafetyLinJian-guangevaluationofmetalminingventilationsystem.YuTie-junetal.Theapplicationoffuzzycomprehensivejudgmentinsafetyassessmentofmineventilationsystem.KwanSeongJeongAqualitativeidentificationandanalysisofhazards,etal.risksandoperatingproceduresforadecommissioningsafetyassessmentofanuclearresearchreactor.LiRun-qiuetal.Summarizationofsafetyassessmentmethodsformineventilationsystem.HuXiao-xueetal.Setpairanalysismodelforriverhealthsystemassessment.YiShan-yongandStudyonscheduleriskevaluationoflargeR&DQiuZhi-mingprojects.LiuQuan-longetal.Studyontheriskmeasurementandthecouplinganalysisofmultihazardsourceincoalgasaccident.2005200520052007200720072008includeunsafeactsandfaultmanagement,suchasinsufficientskillstomeetrequirements,supervisionerrorsanddecisionerrors.Thematerial(machine)hazardfactorsmainlyrefertotheunsafestatesofmachines,namelyequipmentfaults,e.g.,fanfailure,un-successfulexplosionsandswitchshort-circuits.Theenvironmentalhazardfactorsmainlyrefertotheenvironment’sunsafecharac-teristics,suchasgasconcentrationexceedingthestandard.Themanagementhazardfactors,whichmainlyincludetheunsafefac-torsinmanagementregulations,programsororganizationalcul-ture.Fig.2representstherelationshipbetweenthesystemrisk(basedonhazardfactors)andthesafetycontrolcapability(sup-portedbyoffsetriskfactors).Fig.2representsthesystemsafetylevelastheresultsofjointeffectsfromthesafetycontrolcapabilitywithpositiveeffectsandthesystemriskwithnegativeeffects.Thesystemwillbeinastatusofgraduallyrisingsafetyleveliftheeffectsofthesafetycontrolcapabilityisstrongerthanthesystemrisk.Conversely,accidentsoccurmoreeasilyinthesystem.Theoffsetriskfactorsandtheirfunctionswereidentifiedandanalyzedaccordingtosystematicunderstandingofthefourkindsofhazardfactors,whichthenlaysafoundationforselectingevaluationindexesofthesafetycontrolcapability.(1)Thepersonnelsubsystemreferstothemeasurestakenbyindividualssuchasadjustmentandmaintenancetoensurethematerialproductionandtheenvironmentalcontrolsubsystemruninanorderlyfashion.Thesafetycontrolcapabilityismainlydeterminedbytheknowledgeandskillsofindividuals.(2)Thesafetycontrolcapabilityofthematerialproductionsubsystemisreflectedinthecoalminetechnicalconditionsandequipment,whichcanrestrictthecharacteristicsandstatusofthehazards.Theequipmentlevelandtechnologyofthematerialproductionsubsystemshouldbeimprovedtoenhanceitssafetycontrolcapability.(3)Theenvironmentcontrolsubsystemreferstoprovidingsafetyreliabilityoftheoperatingenvironmentincludingcomfortableworkingconditionsforminers.Specifically,thespacemustmeettheproductionrequirementandthetem-perature,humidityandlightingmustmeettheneedsoftheminers.(4)Themanagementsubsystemreferstoaseriesofactivitieswhichincludeplanning,organizing,conducting,coordi-natingandcontrolling.Theaimistoensuretheminers’safetyandhealth,protectthenationalorcollectivepropertyfromlossandpromoteenterprisestoimprovetheirman-agementefficiency.Moreover,itsfunctionismainlyreflectedinrestrainingandregulatingtheotherthreesubsystems’safetystatus.20082008200920112.Modelingofthesafetycontrolcapability2.1.ComponentanalysisonthesafetycontrolcapabilityThehazardfactorscancausephysicalinjury,equipmentdamageorpropertyloss.Theyarenotstaticbutdynamic.Consequently,theyshouldberestrictedorcounteractedtodecreasetheproba-bilityofaccidents.Suchcounteractivefactorsaretermedoffsetriskfactors.Thesafetycontrolcapabilitywasdefinedastheabilitytocontrolthesystemrisklevelbyreducingtheprobabilityofacci-dentsandlossesthatarecausedbythehazardfactors.Fig.1rep-resentsouroverallframework.SongandLi(2010)dividedthehazardfactorsintofourcate-gories,whicharethepersonal,material(machine),environmentalandmanagement.Inthispaper,thefourkindsofhazardfactorsweredealtwithinsummaryform.Thepersonalhazardfactorsmainlyrefertounsafebehaviorattheleveloftheindividualand2.2.ModelingofthesafetycontrolcapabilitySafetyentropyisameasuredescribingthesystem’sdegreeofdisordercausedbyhazardfactors.Thelargerthesafetyentropy,themorelikelytheaccidents.Consequently,thesafetycontrolcapa-bilityandthesafetyentropyhavetherelationshipthatoneisupandtheotherisdownaccordingtothedialecticalrelationshipofsafetyandrisk.Thecasktheoryisdescribedasacaskmadeofdifferentlengthsofboards(Shi,2004).Thecaskcapacitydependsontheheightoftheshortestboard.Toimprovethecaskcapacity,thelengthoftheshortestboardmustbeincreased.Thesafetycontrolcapabilityisusedtoachievethisinthisstudy.Fig.2alsorepresentstherelationshipbetweenthesafetyentropyandthecaskcapacity.ThesafetycaskmodelofthesafetycontrolcapacityThesafetycontrolcapability(supportedbyoffsetriskfactors)Therisks(basedonhazardfactors)ThesafetylevelFig.1.Thestructureofsafetylevel.
Pleasecitethisarticleinpressas:Liu,Q.-l.,Li,X.-c.,Modelingandevaluationofthesafetycontrolcapabilityofcoalminebasedonsystemsafety,JournalofCleanerProduction(2013),http://dx.doi.org/10.1016/j.jclepro.2013.11.048Q.Liu,X.Li/JournalofCleanerProductionxxx(2013)1e63
Fig.2.Therelationshipbetweenthesystemriskandthesafetycontrolcapability.
wasdescribedusingsafetyentropyandcasktheoryasfollows.ThesafetycaskmodelofthecontrolcapacityisshowninFig.3.Therefore,itisnecessarytoevaluatethesafetycontrolcapacityofpersonnelsubsystems,machinerysubsystems,environmentsubsystemsandmanagementsubsystemstofindtheshortboard.Thentakesomerelatedandtargetedmeasurestoimprovethesafetycontrolcapacity.3.EvaluationofthesafetycontrolcapabilityTheshortboardThesafetyentropyThepersonnelsubsystemThedeterminationofevaluationindexesistheprimarystepbeforeevaluatingthesafetycontrolcapability.Theevaluationin-dexesofthecontrolcapacitywereidentifiedbyinvestigatingandanalyzingitscomponentsandtheprincipleofcoveringallthemainproductionprocesses(Liuetal.,2011).TheevaluationindexesincludefourheadindexesandeighteensecondaryindexesasshowninTable2.3.1.EvaluationofthesafetycontrolcapabilitybasedonBPNNBack-propagationneuralnetworks(BPNN)areeffectiveinsolvingcomplicatednonlinearproblems(Hu,2006).Theyhavefeaturesofself-organizingandself-learningthatmakethemeffectiveatnonlinearapproximation.Further,theinformationstoredinthenetworkcanbemademoreaccuratewithsuccessiveretrainingwithadditionaldata.Therefore,theinformationcontentincreasestherebyimprovingthegeneralizationability.Whenconditionschange,newsampledatacanbeputintothenetworktorelearnimprovinginformationcontent,sothewelltrainednetworkcanbeappliedtoevaluatedifferentconditions.ThetopologicalstructureofaBPNNiscomposedofaninputlayer,ahiddenlayerandanoutputlayer(Fig.4).Eachneuronmodelinthehiddenlayerincludesonenonlinearactivationfunction;thesigmoidfunctionisthemostcommonlyused.Wecanapproximateanynonlinearfunctiontoanyaccuracybyadjustingtheconnectionweights,thenumberofinputvariables,outputvariables,networklayersandneurons(Haykin,2009).Thedesignofthenetworkstructuredirectlydeterminesthefinalsimulationresultsandthereliabilityofthenetwork.Toobtainthewelltrainednetworkinwhichsimulationper-formanceisacceptableandthegoodnessoffitishigh,ourBNPPprocedureisundertakenin5steps.(1)Designthenetworkstructureanddeterminethetrainingsamples.TheenvironmentsubsystemThematerialsubsystemThemanagementsubsystemFig.3.Thesafetycaskmodelofthecontrolcapacity.
(1)Thesafetycaskcapacityrepresentsthesafetycontrolcapability.Thegoalofsafetyworkistomakethesafetycaskcapacitysufficientlylargetocontainallthesafetyentropy,namelythereisnoshortboardinthefoursubsystems,thusthepotentialaccidentisundercontrol.(2)Thesafetyentropylevelinsafetycaskrepresentsthesystemrisklevel.Thesystemwillbeinahigherriskstatusandhashigherlikelihoodofaccidentswhenthesafetyentropylevelishigherthanthesafetycaskcapacity.(3)Theheightofthecaskrepresentstheabilitytoresistrisk,namelythesafetycontrolcapacity,anditwillbeimprovedwiththeraiseofcaskheight.(4)Theheightoftheboardsrepresentsthesafetycontrolcapacityofpersonnelsubsystems,machinerysubsystems,environmentsubsystemsandman-agementsubsystems.Theirwidthcanbeseenastheinfluencedegreetothewholesystem,namelyweight.(5)Thehoopofthesafetycaskrepresentsthemanagementsubsystemthataffectsotherthreesubsystemsthroughitssafetycontrolcapability.Ithooksthemupandformsacompletesafetycaskwhichestablishesacollaborativerelationshipwitheachother,jointlypreventtherisks.(6)Theshortboardofthesafetycaskisthesubsystemwhichhaslowsafetycontrolcapacityandhighprobabilityofriskoccurrence.Ifthesafetycaskhasashortboard,nomatterhowreliablethesafetycontrolcapacityofotherthreesubsystemshave,thesafetycontrolcapacityisdeterminedbythisshortboard.Pleasecitethisarticleinpressas:Liu,Q.-l.,Li,X.-c.,Modelingandevaluationofthesafetycontrolcapabilityofcoalminebasedonsystemsafety,JournalofCleanerProduction(2013),http://dx.doi.org/10.1016/j.jclepro.2013.11.0484Q.Liu,X.Li/JournalofCleanerProductionxxx(2013)1e6
Table3Theactualandmodeledvaluesofthetestsamples.Testsamples120.70.70À0.000.70.79À0.090.40.40À0.000.30.30À0.000.30.290.01345Table2Theevaluationindexesofthesafetycontrolcapabilityofcoalmine.HeadindexesTheevaluationindexesofthesafetycontrolcapabilityofcoalmineThepersonnelsubsystem(a1)EvaluationindexesMemberswithcollegedegreeoraboveproportion(a11)Engineersandtechniciansproportion(a12)Memberswithjoblicenseproportion(a13)Specialworkerswithjoblicenseproportion(a14)Mechanicalminingrate(a21)Mechanicaldrivingrate(a22)Assurancecoefficientofliftcapability(a23)Assurancecoefficientoftransportcapability(a24)Gasdrainagesystemcompleteness(a31)Ventilationsystemcompleteness(a32)Dustcontrolsystemcompleteness(a33)Firecontrolsystemcompleteness(a34)Drainagesystemcompleteness(a35)Safetymonitoringsystemcompleteness(a36)Safetycultureconstructionlevels(a41)Safetyproductionregulations(a42)Managementinformationlevel(a43)Safetymanagementorganizationstructure(a44)Thematerialproductionsubsystem(a2)Theenvironmentalsubsystem(a3)ThepersonnelsubsystemActualvalue0.9Modeledvalue0.91ErrorÀ0.01ThematerialproductionActualvalue0.3subsystemModeledvalue0.30ErrorÀ0.00TheenvironmentalActualvalue0.3subsystemModeledvalue0.31ErrorÀ0.01ThemanagementActualvalue0.9subsystemModeledvalue0.88Error0.02ThewholesystemActualvalue0.3Modeledvalue0.29Error0.010.50.90.50.510.890.50À0.010.01À0.000.90.50.90.790.510.900.11À0.010.000.40.30.70.400.280.83À0.000.02À0.130.10.10.30.110.130.26À0.01À0.030.040.50.70.70.520.710.71À0.02À0.01À0.01Themanagementsubsystem(a4)TheBPNNmodelwasimplementedinMATLAB.Thetrainingparametersweresetasfollows.Expectederroris0.001andthemaximumvalueis2000.Thetransferfunctionofthehiddenlayerandoutputlayerarebothlogsigfunctions.Thetrainingfunctionofthenetworkisthetrainlmfunction.Thelearningfunctionofthenetworkisthelearngdmfunction.Theperformancefunctionofthenetworkisthemsefunction.Thenumberofneuronsinthehiddenlayerwaschosenbyselectingthebesttrainingresultsoutofthetestusingbetween5and15layers.(4)Trainthenetwork.Thenetworkwastrainedusing36datauntilthetrainingprocessshowedthatitssimulationperformancemettheoriginalsettingaccuracy.(5)Testtheperformanceofthenetwork.Theremainingfivedatawereusedtotesttheperformanceofthetrainednetwork.TheactualandmodeledvaluesofthetestsamplesineachsubsystemsandthewholesystemaresummarizedinTable3.Table3showsthatthegoodnessoffitofthetrainednetworkishigh.Specifically,withoneortwoexceptions,morethan90%oftheerrorsarelessthan0.1andmorethan80%oftheerrorsarelessthan0.05,sothefittingeffectofthetrainednetworkwasexcellent,namelywehavegotthenetworkwhichsimulationperformancewasacceptableandthegoodnessoffitwashigh.Basedontheabovenetwork,theweightofeachsubsystemonthesafetycontrolcapabilitywasobtainedbyusingthecommandofNET.IWandNET.LWinMATLAB,namelybyextractingandweightingtheconnectionweightofthenetwork’sneurons,andtheyareshowninTable4.AsshowninTable4,thematerialproductionsubsystemhasthehighestweight(0.29)inthefoursubsystems,soitssafetycontrolcapabilityhasthemostsignificantinfluenceonthecoalminesafetylevel.Thenetworkstructurewasdesignedasmultiple-inputs,single-hiddenlayerandsingle-outputaccordingtothecharacteristicsoftheevaluationindexes.Itwastrainedandtestedwithdatacollectedfromforty-onestate-ownedcoalminesinChina;thirty-sixwereusedtotrainthenetworkandtherestwereusedtotestit.(2)Initializethetrainingsamplesdata.Thetrainingdatamustbeinitializedandnormalizedin[0,1]forthesigmoidtrans-activationfunction.Theinitializedmagnitudewasdeterminedbythemaximumvalueofalltheinputdata.The41samplesareshowninAppendices1e4foreachsubsystem(Fig.2).(3)Determinetheparametersofnetworktraining.Table4Theweightofeachsubsystem.ThepersonnelThematerialTheenvironmentalThemanagementcontrolsubsystemsubsystemsubsystemproductionsubsystemFig.4.Theneuralnetworkstructure.
Weight0.260.290.220.23Pleasecitethisarticleinpressas:Liu,Q.-l.,Li,X.-c.,Modelingandevaluationofthesafetycontrolcapabilityofcoalminebasedonsystemsafety,JournalofCleanerProduction(2013),http://dx.doi.org/10.1016/j.jclepro.2013.11.048Q.Liu,X.Li/JournalofCleanerProductionxxx(2013)1e6
Table5TheevalutionresultsofWulanmulunmine’ssafetycontrolcapability.ThepersonnelsubsystemEvaluationvalue0.82Thematerialproductionsubsystem0.99Theenvironmentalcontrolsubsystem0.76Themanagementsubsystem0.65Thewholesystem0.935
3.2.ApplicationofthenetworkIthasbeenshownthatthesimulationperformanceoftheabovetrainednetworkmettheoriginalsettingaccuracyandthegoodnessoffitwashigh.Itwasappliedtoevaluatethesafetycontrolcapabilityofaminenotincludedinthetrainingandtestingsamples.TheWulanmulunminebelongstoShenhuagroupandislocatedatthesouthernpartofWulanmuluntowninShendongminingareainChina.Itisamodernandfullyautomaticminewhichproduces6.2milliontonsperannumandhas771staffs.Itssafetycontrolcapabilitywasassessed,andtheinitializeddataareshowninAppendices1e4.TheresultswereshowninTable5.TheresultsshowedthattheWulanmulunminehasahighlevelofsafetycontrolcapability(0.93).Specifically,thematerialpro-ductionsubsystemhasthehighestsafetycontrolcapability(0.99),whichreflectsitshighdegreeofmechanizationandadvancedtechnology.Thesecondhighestisthepersonnelsubsystem(0.82),whichreflectsitshigh-qualitypersonnel.Thelowestistheman-agementsubsystem(0.65).Therefore,themanagementsubsystemistheshortestboardofthecontrolcapacity,andsomerelatedandtargetedmeasuresshouldbetakentoimproveitssafetycaskca-pacity,namelythecontrolcapacity.4.ConclusionsThefollowingconclusionscanbedrawnfromthisstudy:(1)Coalminesafetyisdirectlyrelatedtoitsconnaturalrisks,butmorerelevanttotheoffsetriskfactors,namelythesafetycontrolcapability.Sowecanimprovecoalminesafetylevelbyenhancingthesafetycontrolcapabilityfromtheaspectsofpersonnel,material(machine),envi-ronmentandmanagementfromtheperspectiveofsystemsafety.(2)Themodelofsafetycontrolcapabilitywasconstructedanddescribedusingsafetyentropyandcasktheoryandthenasetofmodelevaluationindexeswerederived.Aback-propagationneuralnetwork(BPNN)wasdevelopedtoeval-uatethesafetycontrolcapabilityofcoalmine.Itwastrainedandtestedwithdatacollectedfromforty-onefirms.There-sultsshowedthatthesimulationperformanceofthetrainednetworkmettheoriginalsettingaccuracyandthegoodnessoffitwashigh.Soweconcludethatitssimulationperfor-mancewasacceptableandfittingeffectwasexcellent,anditcanbeappliedtoevaluateothermines’safetycontrolcapability.AcknowledgmentsTheresearchdescribedinthispaperwassubstantiallysup-portedbytheNationalNaturalScienceFoundationProjectsofChina(71271206)andProgramforNewCenturyExcellentTal-entsinUniversityofMinistryofEducationofChina(NCET-10-0766).AppendixesAppendix1Thesampledataofpersonnelsubsystem.SamplesSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSample123456789101112131415161718192021a110.10.30.510.50.910.90.50.30.90.30.30.30.10.70.50.30.10.10.1a120.50.510.90.70.50.70.90.50.70.90.70.60.50.10.50.70.30.50.10.3a130.50.110.10.50.50.50.90.310.50.90.310.50.50.10.10.50.11a1410.30.10.30.30.611110.90.90.11110.10.6110.1SamplesSample22Sample23Sample24Sample25Sample26Sample27Sample28Sample29Sample30Sample31Sample32Sample33Sample34Sample35Sample36Sample37Sample38Sample39Sample40Sample41Wulanmulunminea110.50.30.30.30.50.10.70.50.70.90.70.90.610.90.60.50.50.50.70.4a120.30.90.10.10.50.50.50.30.50.70.30.60.90.70.910.110.50.90.3a130.110.10.110.311110.30.50.90.30.810.80.50.90.10.8a140.610.90.31111111110.60.90.90.90.90.90.90.9Appendix2Thesampledataofmaterialproductionsubsystem.SamplesSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSample123456789101112131415161718192021a210.20.321110.210.60.310.50.20.30.90.90.810.70.50.21a220.20.20.050.30.80.210.60.20.70.20.20.30.60.90.30.90.50.20.20.2a230.210.80.10.20.20.20.10.310.20.20.90.40.50.310.90.30.70.9a240.111110.71111110.911110.9111SamplesSample22Sample23Sample24Sample25Sample26Sample27Sample28Sample29Sample30Sample31Sample32Sample33Sample34Sample35Sample36Sample37Sample38Sample39Sample40Sample41Wulanmulunminea210.50.90.910.210.50.80.20.20.20.1110.20.40.70.70.710.9a220.20.50.910.210.50.10.20.10.20.10.50.30.20.20.20.20.70.70.9a230.8110.810.40.310.30.20.10.710.60.20.90.20.20.90.40.6a2410.110.20.21111110.20.210.20.2110.210.6Pleasecitethisarticleinpressas:Liu,Q.-l.,Li,X.-c.,Modelingandevaluationofthesafetycontrolcapabilityofcoalminebasedonsystemsafety,JournalofCleanerProduction(2013),http://dx.doi.org/10.1016/j.jclepro.2013.11.0486
Appendix3Thesampledataofenvironmentalcontrolsubsystem.SamplesSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSample123456789101112131415161718192021a311110.90.9110.910.911110.1110.1111a320.80.90.80.810.90.20.80.90.90.60.30.09110.70.40.30.30.40.7a330.2110.910.9110.10.1111110.910.9111a341110.81110.811110.81110.91110.9Q.Liu,X.Li/JournalofCleanerProductionxxx(2013)1e6
a3510.51110.910.20.9110.910.90.90.910.9110.9a360.810.70.90.90.90.910.90.90.810.60.90.50.60.70.90.90.50.3SamplesSample22Sample23Sample24Sample25Sample26Sample27Sample28Sample29Sample30Sample31Sample32Sample33Sample34Sample35Sample36Sample37Sample38Sample39Sample40Sample41Wulanmulunminea3110.21111110.910.9110.21110.2110.8a321110.90.70.40.30.90.710.90.910.40.50.60.40.50.50.90.9a330.910.310.810.9110.711110.810.10.8110.7a34110.81110.910.910.910.811111110.8a351111110.910.910.911110.211110.7a360.30.60.60.80.70.80.8110.90.60.70.40.40.40.80.90.50.90.70.9Appendix4Thesampledataofmanagementsubsystem.SamplesSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSampleSample123456789101112131415161718192021a410.10.170.590.920.950.830.130.820.160.8510.980.770.530.3110.80.690.950.691a42a430.40.60.20.60.410.40.60.8110.610.60.80.80.810.60.810.40.60.20.60.40.80.40.60.60.610.40.60.40.60.60.610.40.60.15a440.60.40.850.70.90.910.950.80.850.870.750.9510.9510.90.80.90.70.95SamplesSample22Sample23Sample24Sample25Sample26Sample27Sample28Sample29Sample30Sample31Sample32Sample33Sample34Sample35Sample36Sample37Sample38Sample39Sample40Sample41Wulanmulunminea41110.380.3210.550.670.980.890.861110.9510.930.090.0310.80.7a420.60.40.60.410.2510.6110.20.120.20.4110.40.20.60.150.7a430.60.60.20.40.80.20.610.810.20.60.20.8110.40.10.60.150.6a441110.850.650.40.70.50.80.90.91111110.80.70.80.8ReferencesCheng,L.,Yang,Y.L.,Xiong,Y.X.,2005.Studyofmineventilationsystemassessmentbasedonartificialneuralnetwork.ChinaSaf.Sci.J.15(5),88e91.David,M.S.,Vasilios,H.F.,Marvin,A.S.,1983.Formaldehyderiskassessmentforoccupationallyexposedworkers.RegulatoryTexicol.Pharmacol.3(4),355e371.Denby,B.,Kizil,M.S.,1992.Applicationofexpertsystemsingeotechnicalriskassessmentforsurfacecoalminedesign.Int.J.RockMech.MiningSci.Geo-mech.Abst.2(2),110.Duzgun,H.S.B.,2005.AnalysisofrooffallhazardsandriskassessmentforZon-guldakcoalbasinundergroundmines.Int.J.CoalGeol.64(1e2),104e115.Hatton,W.,W,M.K.G.,1995.RiskassessmentappliedtocoaltonnageestimationintheUnitedKingdom.Int.J.RockMech.MiningSci.Geomecha.Abst.32(6),276.Haykin,S.,2009.NeuralNetworksandLearningMachines,pp.120e135.Hiromitsu,K.,2007.SatisfyingSafetyGoalsbyProbabilisticRiskAssessment,pp.95e111.Hu,W.S.,2006.TheTheoryofNeuralNetworkanditsApplicationsinEngineering,pp.100e109.Hu,X.X.,Yang,X.H.,Li,J.Q.,Geng,L.H.,2008.Setpairanalysismodelforriverhealthsystemassessment.Syst.Eng.-TheoryPract.28(5),164e170.Jerry,S.R.,1972.ACaseStudyinRiskManagement.In:RiskandInsuranceSeries,p.2.Kwan,S.J.,Dong,G.L.,Kune,W.L.,HyeonKyo,L.,2008.Aqualitativeidentificationandanalysisofhazards,risksandoperatingproceduresforadecommissioningsafetyassessmentofanuclearresearchreactor.Ann.Nucl.Energy35(10),1954e1962.Li,J.K.,2012.ChinaStrategicCoalStockpileanditsRegulationMechanism,pp.35e37.LI,R.Q.,Shi,S.L.,Peng,X.,2008.Summarizationofsafetyassessmentmethodsformineventilationsystem.ChinaSaf.Sci.J.18(1),112e118.Liu,Q.L.,Li,X.C.,Zhang,Q.C.,2011.Studyontheriskmeasurementandthecouplinganalysisofmultihazardsourceincoalgasaccident.J.Saf.CoalMines42(7),189e192.Nick,F.P.,1988.Riskassessmentandaccidentanalysis.ActsPsychologica68(1e3),355e368.Robert,A.B.,1986.Decisionmakingandprobabilisticriskassessment.Nucl.Eng.Des.93(2e3),341e348.Shi,L.,2004.CaskEffect,pp.1e10.Wang,W.J.,Lin,J.G.,2007.Fuzzyarrangementcomprehensivesafetyevaluationofmetalminingventilationsystem.MiningMetallurgy16(1),4e7.Yi,S.Y.,Qiu,Z.M.,2009.StudyonscheduleriskevaluationoflargeR&Dprojectsbasedonthetheoryofsetpairanalysis.J.Stat.Dec.(1),44e47.Yu,T.J.,Xu,M.G.,Yao,G.F.,2007.Theapplicationoffuzzycomprehensivejudgmentinsafetyassessmentofmineventilationsystem.Gold28(5),18e21.Zhang,X.Y.,Dou,S.Q.,2005.Theassessmentofventilationsystemforundergroundminesbasedonneuralnetwork.Non-FerrousMiningTechnol.21(4),11e13.Pleasecitethisarticleinpressas:Liu,Q.-l.,Li,X.-c.,Modelingandevaluationofthesafetycontrolcapabilityofcoalminebasedonsystemsafety,JournalofCleanerProduction(2013),http://dx.doi.org/10.1016/j.jclepro.2013.11.048
因篇幅问题不能全部显示，请点此查看更多更全内容