JULY2014IntroductiontoLayerofProtectionAnalysis(LOPA)Isitsafeenough?Thiscanbeadifficultquestion.LevelofProtectionAnalysis(LOPA)isastructuredmethodthatyieldsadefendableanswertothatquestion.TheBasisTheaverage30yearoldhasabouta1/1000(10-3)probabilityofdyinginthisyear(muchofthatduetoautomobilerisk).Eventhoughmanypeoplearesurprisedwhentheyfirsthearthisnumber,itisalevelofriskthatweimplicitlyaccept.Supposethatyouweretolearnthatyouroddsofbeingkilledonthejobis1/100(10-2),tentimeshigher.Youwilllikelybeupsetandmayinsistonimprovements.Fortunately,justtheoppositeisthecaseinmostoilandgasoperations.Companiesthatexplicitlysetatargetseektobeatleastanorderofmagnitudesaferthantheworldatlarge(10-4).LOPAprovidesaconsistentbasisforjudgingwhethertherearesufficientindependentprotectionlayersagainsthazardouseventstoachievetheriskreductionrequiredtoachievesuchanexplicittarget.TheMethodLOPAusesconservative,orderofmagnitudevaluesforinitiatingeventfrequency,consequenceseverityandlikelihoodoffailureofprotectivelayerstoapproximatearisklevelforanygivenscenario.Inrigor,itfallsbetweenatypicalriskmatrixapproach(ascommonlyusedinHAZOPs)andaquantitativemethod(QRA).ALOPAisfrequentlyperformedafteraHAZOPtofurtherinvestigatesignificantfindings.TheStandardApproachis:1.Describetheaccidentscenariotobestudied.2.Identifytheinitiatingeventanddeterminethefrequencyorlikelihoodoftheinitiatingevent.SeeTable1and2fortypicalinitiatingeventfrequencies.3.Identifytheconsequencelevel.Theconsequenceseverityisjudgedbasedonspecifiedcriteria.Table3isasimplifiedexample.4.DeterminetheRiskReductionRequirementviaaCalibratedRiskMatrix.Figure1isanexamplematrix.5.IdentifytheIndependentProtectionLayers(IPLs),estimatetheprobabilityoffailureondemand(PFD)ofeachIPLandmathematicallycombinetheIPLs.SeeTable4fortypicalIPLPFDs.6.ComparethecombinedriskreductioneffectivenessofallidentifiedIPLswiththeRequiredRiskReductiontodetermineifadditionalriskreductionisrequired.RiskReductionMatrixFigure1isasectionofanexampleriskreductionmatrix.Thetypicalred,yellow,greencolor-codingisretainedonthisexample,buttheimportantfeatureisthenumbersinthecells.Theseareorderofmagnituderiskreductionrequirements.Thisisa“10-4matrix”.Theentryofa‘0’inrow4-columnAindicatesthatnofurtherriskreductionisrequiredforascenariofeaturingaMajorconsequence(singlefatality)estimatedtooccuratafrequencynogreaterthan1/10000years.(Recallfromthediscussionearlierthatthisis1orderofmagnitudesaferthantheworldatlarge.)Thismatrixhastheagreeablepropertythatmovingoneroworcolumninanydirectionchangestheseverityorfrequencyby1orderofmagnitude.Hence,havingestablishedthebasis‘0’inrow4–columnA,itisasimpletasktocompletetherestoftheentries.Moveonecolumntotheright–add1.Moveonerowup–add1.Table1:InitiatingEventFrequenciesExampleInitiatingCauseLikelihood,events/yrLikelihood,10-xControlLoopFailure1/1010-1SealFailure1/1010-1GasketFailure1/10010-2RotatingEquipTrip1/1100FixedEquipFailure1/10010-2LossofPower1/1010-1UtilityFailure1/1010-1HumanErrorsLikelihoodWelltrainedoperatorw/stress1/(10opportunities)Welltrainedoperator,nostress1/(100opportunities)Table2:HumanErrorFrequenciesExampleTable3:ConsequenceSeverityTableExampleSeveritySafetyCost($US)5:CatastrophicMultipleFatalities1Billion4:MajorSingleFatality100Million3:SevereSeriousInjury10Million2:MinorMinorInjury1Million1:SlightFirstAid100,000Table4:SomeTypicalPFDsIn...
优质VIP
