++++++++FMEA

FMEAFailureModeandEffectsAnalysisAdaptedfrompresentationbyDr.StamperOutline•FailureModeandEffectsAnalysis–Whatisit?–Motivation–FMEAMethods–Example•FMEAin-classexerciseWhatisanFMEA?Description:•Aprocedurethatexamineseachiteminasystem,considershowthatitemcanfailandthendetermineshowthatfailurewillaffect(orcascadethrough)thesystemAcronyms•FMEA:FailureModesandEffectsAnalysis•FMECA:FailureModesandEffectsandCriticalityAnalysisMotivationforConductingaFMEA•Improvesdesignbydiscoveringunanticipatedfailures•Highlightstheimpactofthefailures•Potentiallyhelpfulduringlegalactions•Providesamethodtocharacterizeproductsafety•Oftenrequired(e.g.FDAandDODprocurement)MethodtoConductanFMEA(takenfromASMHandbookVol.11)–Identifyallcomponentsorsystemsatgivenlevelofthedesignhierarchy.–Listthefunctionofeachidentifiedcomponentorsystem.–Identifyfailuremodesforeachcomponent/system.Typicallytherewillbeseveralwaysinwhichacomponentcanfail.–Determinetheeffect(bothlocallyandglobally)onthesystem.MethodtoConductanFMEA(takenfromASMHandbookVol.11)–Classifythefailurebyitseffectsonthesystemoperation.–Determinethefailure’sprobabilityofoccurrence.–Identifyhowthefailuremodecanbedetected(maypointoutwhatneedstobeinspectedonaregularbasis).–Identifyanycompensatingprovisionsordesignchangestomitigatethefailureeffects.OtherVariationsoftheMethod•Mil-STD-1629A,Fig101.3•SAEJ1739(DFMEA)•TheDesignFMEAisusedtoanalyzeproductsbeforetheyarereleasedtoproduction.•Itfocusesonpotentialfailuremodesofproductscausedbydesigndeficiencies.•DesignFMEAsarenormallydoneatthreelevels–system,subsystem,andcomponentlevels•ThistypeofFMEAisusedtoanalyzehardware,functionsoracombinationPROCESSFMEA(PFMEA)•TheProcessFMEAisnormallyusedtoanalyzemanufacturingandassemblyprocessesatthesystem,subsystemorcomponentlevels.•ThistypeofFMEAfocusesonpotentialfailuremodesoftheprocessthatarecausedbymanufacturingorassemblyprocessdeficiencies.WhattoDo•FunctioncomesfromFunctionalAnalysis,FunctionalDecomposition•PotentialFailureModecomesfromthingsthathavegonewronginthepast,concernsofdesigners,andbrainstorming.Possibleconsiderationsarepartialfunction,intermittentfunction,excessfunction.•PotentialEffectsareconsequencestothedesign,theuser,andtheenvironment.Safetyandregulationnoncompliancearecriticalissues.IN-CLASSPROBLEM1Mostcarshaveafuelfilterbetweenthefueltankandtheengine.Considertheconsequencesofa)Anti-function(oppositeoffiltering)b)Partialfunctionc)Intermittentfunctiond)ExcessfunctionWhattoDo•PotentialCausesoffailureshouldbeengineeringrelatedsuchasincorrectmaterial,corrosion,wearandhumanrelatedsuchasinexperience,misuse,etc.•CurrentDesignControlsarethingslikeinspections,testing,pokeyoke,andotherdesignchecksthatareintendedtopreventtheproblem.WhattoDo•AssignvaluestoSeverity,Occurrence,andDetectionusingthetablesonthenextthreepages.•DeterminetheRiskPriorityNumber(Severity*Occurrence*Detection)•Developanactionplan•ImplementanactionplanRiskGuidelinesEffectRankCriteriaNone1NoeffectVerySlight2NegligibleeffectonPerformance.Someusersmaynotice.Slight3Slighteffectonperformance.NonvitalfaultswillbenoticedbymanyusersMinor4Minoreffectonperformance.Userisslightlydissatisfied.Moderate5Reducedperformancewithgradualperformancedegradation.Userdissatisfied.Severe6Degradedperformance,butsafeandusable.Userdissatisfied.HighSeverity7Verypoorperformance.Verydissatisfieduser.VeryHighSeverity8Inoperablebutsafe.ExtremeSeverity9Probablefailurewithhazardouseffects.Compliancewithregulationisunlikely.MaximumSeverity10Unpredictablefailurewithhazardouseffectsalmostcertain.Non-compliantwithregulations.OccurrenceRankingOccurrenceRankCriteriaExtremelyUnlikely1Lessthan0.01perthousandRemoteLikelihood20.1perthousandrateofoccurrenceVeryLowLikelihood30.5perthousandrateofoccurrenceLowLikelihood41perthousandrateofoccurrenceModeratelyLowLikelihood52perthousandrateofoccurrenceMediumLikelihood65perthousandrateofoccurrenceModeratelyHighLikelihood710perthousandrateofoccurrenceVeryHighSeverity820perthousandrateofoccurrenceExtremeSeverity950perthousandrateofoccurrenceMaximumSeverity10100perthousandrateofoccurrenceDetectionRankingDetectionRankCriteriaExtremelyLikely1Canbecorrectedpriortoprototype/ControlswillalmostcertainlydetectVeryHighLikelihood2Canbecorrectedpriortodesignrelease/VeryHighprobabilityofdetectionHighLikelihood3Likelytobecorrected/HighprobabilityofdetectionModeratelyHighLikelihood4DesigncontrolsaremoderatelyeffectiveMediumLikelihood5DesigncontrolshaveanevenchanceofworkingModeratelyLowLikelihood6DesigncontrolsmaymisstheproblemLowLikelihood7DesigncontrolsarelikelytomisstheproblemVeryLowLikelihood8DesigncontrolshaveapoorchanceofdetectionRemoteLikelihood9Unproven,unreliabledesign/poorchancefordetectionExtremelyUnlikely10Nodesigntechniqueavailable/ControlswillnotdetectIN-CLASSPROBLEM2Supposeyouwerethedesignerofcarsbeforeairbagsandseatbelts.Supposeyourfunctionis“Protectoccupantsduringhead-oncollisionatlessthan40mph.”EstimatenumbersforSeverity,Occurrence