AnnexI.2FailureMode

EffectsAnalysis失效模式效應(yīng)分析

(FMEA)I.2:FailureModeEffectsAnalysis失效模式效應(yīng)分析(FMEA)(seeIEC60812

見IEC60812)EvaluationofpotentialfailuremodesforprocessesThelikelyeffectonoutcomesand/orproductperformanceOncefailuremodesareestablished,

riskreductioncanbeusedto

eliminate,reduceorcontrolthepotentialfailuresFMEAreliesonprocessunderstandingSummarizetheimportantmodesoffailure,factorscausingthesefailuresandthelikelyeffectsofthesefailures

Howtoperform?

BreakdownlargecomplexprocessesintomanageablestepsICHQ9I.2:FailureModeEffectsAnalysis失效模式效應(yīng)分析(FMEA)(seeIEC60812

見IEC60812)工藝的潛在失效模式的評(píng)估對(duì)結(jié)果和產(chǎn)品性能的可能的影響一旦失效模式建立，風(fēng)險(xiǎn)降低行動(dòng)用于來消除、降低或控制潛在的失效依賴于對(duì)工藝的理解概述重要的失效模式，引起失效的因數(shù)和失效的可能的影響

如何實(shí)施?

將大的復(fù)雜的工藝分解成易處理的步驟ICHQ9I.2:FailureModeEffectsAnalysis(FMEA)

PotentialAreasofUse(s)PrioritizerisksMonitortheeffectivenessofriskcontrolactivitiesEquipmentandfacilitiesAnalyzeamanufacturingprocess

toidentifyhigh-riskstepsorcriticalparametersRNP:RiskPriorityNumberC.Kingery,TheSixSigmaMemoryJoggerIIICHQ9I.2:失效模式效應(yīng)分析(FMEA)

可使用的潛在領(lǐng)域：Prioritizerisks優(yōu)先考慮的風(fēng)險(xiǎn)Monitortheeffectivenessofriskcontrolactivities監(jiān)控風(fēng)險(xiǎn)控制活動(dòng)的效力Equipmentandfacilities

設(shè)備和設(shè)施Analyzeamanufacturingprocesstoidentifyhigh-riskstepsorcriticalparameters為了識(shí)別高風(fēng)險(xiǎn)步驟和關(guān)鍵參數(shù)分析一個(gè)生產(chǎn)工藝RNP:RiskPriorityNumberC.Kingery,TheSixSigmaMemoryJoggerIIICHQ9I.2:FailureModeEffectsAnalysis(FMEA)Howtoperform?1.Establishateam2.Identifytheknownandpotentialfailuremodes:Developlistsofknownproblemsandbrainstormotherpotentials…

e.g.ProductnotmeetingspecificationProcessnotmeetingyieldrequirementsMalfunctioningequipmentSoftwareproblemsNewlyidentifiedfailuremodesshouldbeaddedatanytimeEXAMPLEI.2:失效模式效應(yīng)分析(FMEA)如何實(shí)施?1.成立一個(gè)小組2.識(shí)別已知和潛在的失效模式生成已知問題和集體討論的其他潛在問題的列表例如.產(chǎn)品不符合標(biāo)準(zhǔn)工藝不滿足收率需要設(shè)備故障軟件問題新發(fā)現(xiàn)的失效問題應(yīng)該隨時(shí)加進(jìn)來。EXAMPLEI.2:FailureModeEffectsAnalysis(FMEA)Howtoperform?3.Characterisetheseverity,probabilityanddetectability

AnequalnumberoflevelsissometimeshelpfulSomepreferenceto3,4,5,6or10levelsBut:anevennumberoflevelsavoidsthemidpointUsedifferentscalesLinear: 1,2,3,4Exponential: 1,2,4,8Logarithmic: 1,10,100,1000Selfmade: 1,3,7,10MultiplyingdifferentscaleswilldifferentiatetheoutcomeTheaimistocomeupwithamethodofprioritisingEXAMPLEI.2:失效模式效應(yīng)分析(FMEA)如何實(shí)施?3.識(shí)別嚴(yán)重性、可能性和可檢測性一個(gè)相當(dāng)?shù)臄?shù)字等級(jí)常常是很有用的一般優(yōu)先選擇等級(jí)3，4，5，6或10但是:一個(gè)偶數(shù)等級(jí)會(huì)避免中位數(shù)用不同的等級(jí)線性的: 1,2,3,4直數(shù)的: 1,2,4,8對(duì)數(shù): 1,10,100,1000自定義: 1,3,7,10不同的等級(jí)的乘積將區(qū)分結(jié)果目標(biāo)是提出一個(gè)定義優(yōu)先順序的方法EXAMPLEI.2:FailureModeEffectsAnalysis(FMEA)Howtoperform?4.Defineactions5.Revisittheranking6.Defineresidualrisk7.PerformashortsummaryScopeDatafromtheassessment&control

(e.g.No.ofidentifiedfailuremodes)Levelofacceptedriskwithoutactionsi.e.residualrisk

(e.g.RiskpriorityNumber<50)Recommendedactions,responsibilitiesandduedates

(includingapproval,ifappropriate)Personinchargeforfollow-upofFMEAEXAMPLEI.2:失效模式效應(yīng)分析(FMEA)如何實(shí)施?4.詳細(xì)說明行為5.重新評(píng)級(jí)6.詳細(xì)說明剩余的風(fēng)險(xiǎn)7.完成一個(gè)簡短的摘要范圍從評(píng)估和控制中得到的數(shù)據(jù)

(例如已識(shí)別失效模式的編號(hào))不需要采取措施的可接受的風(fēng)險(xiǎn)的等級(jí)即剩余風(fēng)險(xiǎn)

(比如風(fēng)險(xiǎn)優(yōu)先號(hào)<50)建議采取的措施,相關(guān)職責(zé)和完成日期

(如果需要也包括批準(zhǔn))負(fù)責(zé)下一步FMEA的責(zé)任人EXAMPLESeverity(Consequencesoffailure)10ExtremePredictedtocausesevereimpacttoquality(Productoutofspecifications,noExpertStatementpossible)

7HighPredictedtocausesignificantimpactonquality(Failuretomeetspecifications,noStabilitydata,ExpertStatementpossible)

3ModeratePredictedtocauseminorimpactonquality(Failuretomeetspecifications,Stabilitydataavailable)

1LowPredictedtohaveno/minorimpactonqualityoftheproduct(Qualitywithinspecifications)EXAMPLE嚴(yán)重性(失效的結(jié)果)10極高可能會(huì)對(duì)產(chǎn)品產(chǎn)品產(chǎn)生嚴(yán)重的影響(產(chǎn)品OOS,可能無專家聲明)

7高可能對(duì)產(chǎn)品質(zhì)量產(chǎn)生重大的影響(不能滿足指標(biāo),沒有穩(wěn)定性數(shù)據(jù),可能需要專家聲明)

3中可能對(duì)產(chǎn)品質(zhì)量產(chǎn)生小的影響(不能滿足指標(biāo)要求,可能沒有穩(wěn)定性數(shù)據(jù))

1低可能對(duì)產(chǎn)品質(zhì)量無影響或很微小的影響(質(zhì)量在指標(biāo)之內(nèi))EXAMPLEProbability(Likelihoodfailurewillhappen)8RegularfailuresExpectedtohappenregularly

4RepeatedfailuresExpectedtohappeninalowfrequency

2OccasionalfailuresExpectedtohappeninfrequently

1UnlikelyfailuresUnlikelytohappenEXAMPLE可能性(失效發(fā)生的概率)8經(jīng)常失效期望有規(guī)律的發(fā)生

4重復(fù)的失效期望以較低頻率發(fā)生

2偶爾失效期望很少發(fā)生

1幾乎不會(huì)失效不太可能發(fā)生EXAMPLEDetectability(Abilitytofindthefailure)4NormallynotdetectedFailureverylikelytobeoverlooked,hencenotdetected

(notechnicalsolution,nomanualcontrol)3LikelynotdetectedFailuremaybeoverseen

(manualcontrol,spotchecks)2RegularlydetectedFailurewillnormallybedetected

(manualcontrol,routineworkwithstatisticalcontrol)1AlwaysdetectedFailurecanandwillbedetectedinallcases

(monitoring,technicalsolutionavailable)EXAMPLE可檢測性(發(fā)現(xiàn)失效的能力)4通常不能被發(fā)現(xiàn)失效非常容易被忽視,因此不能發(fā)現(xiàn)

(沒有技術(shù)解決方法,沒有手動(dòng)控制)3可能發(fā)現(xiàn)不了失效可能被檢查到

(手動(dòng)控制,現(xiàn)場檢查)2定期檢查可以發(fā)現(xiàn)失效通常都能發(fā)現(xiàn)

(手動(dòng)控制,

統(tǒng)計(jì)控制的例行工作)1隨時(shí)可以發(fā)現(xiàn)失效隨時(shí)會(huì)被發(fā)現(xiàn)

(檢測器,技術(shù)措施有效)EXAMPLEFMEA:QuantitationofRisk:Severity10DangerouslyHighFailurecouldleadtodeathorpermanentinjurytothecustomer.Financial:>$1,000,0009ExtremelyhighFailurecouldleadtoinjurytothecustomer.Failurewouldcreatenon-compliancewithregisteredspecifications.Failurelikelytoleadtorecall.Financial:$1,000,0008VeryHighFailurecouldleadtoadversereactionforcustomer.FailurewouldcreatenoncompliancewithGMPregulationsorproductregistrations.Failurepossibletoleadtorecall.Financial:$500,0007HighFailureleadstocustomerperceptionofsafetyissue.Failurerendersindividualunit(s)unusable.Failurecausesahighdegreeofcustomerdissatisfaction.RecallforbusinessreasonspossiblebutAuthorityrequiredrecallunlikely.Financial:$100,0006ModerateFailurecausesahighdegreeofcustomerdissatisfactionandnumerouscomplaints.Failureunlikelytoleadtorecall.Financial:$50,0005LowFailurelikelytocauseisolatedcustomercomplaints.Financial:$10,0004VeryLowFailurerelatestonon-dosageformissues(likeminorpackagingproblems)andcanbeeasilyovercomebythecustomer.Financial:$5,0003MinorFailurecouldbenoticedbythecustomerbutisunlikelytobeperceivedassignificantenoughtowarrantacomplaint.2VeryMinorFailurenotreadilyapparenttothecustomer.Financial:<$1,0001NoneFailurewouldnotbenoticeabletothecustomer.Financial:noneDr.GaryHarbour,PfizerEXAMPLEFMEA:風(fēng)險(xiǎn)的數(shù)值:嚴(yán)重性10高危險(xiǎn)失效可能導(dǎo)致消費(fèi)者死亡或永久的傷害，財(cái)產(chǎn)損失：>$1,000,0009極端高失效可能導(dǎo)致消費(fèi)者的傷害。失效將產(chǎn)生與已注冊(cè)的標(biāo)準(zhǔn)不符合，失效會(huì)導(dǎo)致產(chǎn)品召回財(cái)產(chǎn)損失：$1,000,0008很高失效會(huì)導(dǎo)致消費(fèi)者不良反應(yīng)?？赡墚a(chǎn)生不符合GMP法規(guī)或產(chǎn)品注冊(cè)?？赡軙?huì)導(dǎo)致召回。財(cái)產(chǎn)損失：$500,0007高失效會(huì)引起客戶對(duì)安全問題的擔(dān)心，失效必須給單個(gè)單位給與補(bǔ)償，失效會(huì)導(dǎo)致客戶高度的不滿意，產(chǎn)品召回只是因?yàn)樯虡I(yè)原因而不會(huì)引起強(qiáng)制召回，財(cái)產(chǎn)損失:$100,0006中失效會(huì)引起客戶高度的不滿意和大量的投訴。失效未必會(huì)引起產(chǎn)品召回，財(cái)產(chǎn)損失:$50,0005低失效會(huì)引起單個(gè)客戶投訴，財(cái)產(chǎn)損失:$10,0004非常低失效不涉及到劑型等問題（象包裝問題），可以很容易被客戶接受，財(cái)產(chǎn)損失:$5,0003較小失效會(huì)被客戶注意但是未必會(huì)引起客戶投訴2非常小失效不容易被客戶接受.財(cái)產(chǎn)損失:<$1,0001無失效不會(huì)引起客戶注意。財(cái)產(chǎn)損失：無Dr.GaryHarbour,PfizerEXAMPLEFMEA:QuantitationofRisk:Probability10VeryHigh:FailureisalmostinevitableMorethanoneoccurrenceperdayoraprobabilityofmorethanthreeoccurrencesin10units(Cpk<0.33or<1σ).9

Oneoccurrenceeverythreetofourdaysoraprobabilityofthreeoccurrencesin10units(Cpk~0.33or~1σ).8High:RepeatedfailuresOneoccurrenceperweekoraprobabilityof5occurrencesin100units(Cpk~0.67or~2σ).7

Oneoccurrenceeverymonthoroneoccurrencein100units(Cpk~0.83~2.5σ).6Moderate:OccasionalFailuresOneoccurrenceeverythreemonthsorthreeoccurrencesin1,000units(Cpk~1.00or~3σ).5

Oneoccurrenceeverysixmonthstooneyearoroneoccurrencein10,000units(Cpk~1.17or~3.5σ).4

Oneoccurrenceperyearorsixoccurrencesin100,000units(Cpk~1.33or~4σ).3Low:RelativelyfewFailuresOneoccurrenceeveryonetothreeyearsorsixoccurrencesin10,000,000units(Cpk~1.67or~5σ).2

Oneoccurrenceeverythreetofiveyearsor2occurrencesin1,000,000,000units(Cpk~2.00OR~6σ).1Remote:FailureisunlikelyOneoccurrenceingreaterthanfiveyearsorlessthantwooccurrencesin1,000,000,000units(Cpk>2.00OR>6σ).Forbatchfailuresusethetimescaleforunitfailuresusetheunitscale.Dr.GaryHarbour,PfizerEXAMPLEFMEA:風(fēng)險(xiǎn)數(shù)值:可能性10非常高:失效幾乎不可避免每天發(fā)生超過一次或可能性大于3/10(Cpk<0.33or<1σ).9

每三天或四天發(fā)生一次或可能性是3/10(Cpk~0.33or~1σ).8高:重復(fù)失效每周一次或可能性為5/100(Cpk~0.67or~2σ).7

每月發(fā)生一次或可能性為1/100(Cpk~0.83~2.5σ).6中:偶爾失效每三個(gè)月發(fā)生一次或可能性為3/1000(Cpk~1.00or~3σ).5

每六個(gè)月到一年發(fā)生一次或可能性為1/10，000(Cpk~1.17or~3.5σ).4

一年發(fā)生一次或可能性為6/100，000(Cpk~1.33or~4σ).3低:相對(duì)很少失效每一年至三年發(fā)生一次或可能性為6/10，000，000(Cpk~1.67or~5σ).2

每三至五年發(fā)生一次或可能性為1/1，000，000，000(Cpk~2.00OR~6σ).1微笑:失效幾乎不太可能超過五年發(fā)生一次或小于2/1，000，000，000(Cpk>2.00OR>6σ).Forbatchfailuresusethetimescaleforunitfailuresusetheunitscale.對(duì)于批失效用時(shí)間尺度，對(duì)于單元失效使用單位尺度Dr.GaryHarbour,PfizerEXAMPLEFMEA:QuantitationofRisk:Detection10AbsoluteUncertaintyTheproductisnotinspectedorthedefectcausedbythefailureisnotdetectable.9VeryRemoteProductissampled,inspected,andreleasedbasedonAcceptableQualityLevel(AQL)samplingplans.8RemoteProductisacceptedbasedonnodefectsinasample.7VeryLowProductis100%manuallyinspectedintheprocess.6LowProductis100%manuallyinspectedusinggo/no-goorothermistake-proofinggauges.5ModerateSomeStatisticalProcessControl(SPC)isusedintheprocessandproductisfinalinspectedoff-line.4ModeratelyHighSPCisusedandthereisimmediatereactiontoout-of-controlconditions.3HighAneffectiveSPCprogramisinplacewithprocesscapabilities(Cpk)greaterthan1.33.2VeryHighAllproductis100%automaticallyinspected.1AlmostCertainThedefectisobviousandthereis100%automaticinspectionwithregularcalibrationandpreventivemaintenanceoftheinspectionequipment.Dr.GaryHarbour,PfizerEXAMPLEFMEA:風(fēng)險(xiǎn)數(shù)值:可檢測性10絕對(duì)不確定產(chǎn)品不檢查或失效引起的缺陷檢測不到9非常微弱.產(chǎn)品基于可接受的質(zhì)量平行取樣計(jì)劃進(jìn)行取樣，檢查和放行8微弱樣品沒有缺陷產(chǎn)品就可以被接受。7很低在工藝中產(chǎn)品被100%的人為檢查。6低產(chǎn)品使用行/不行或其他防錯(cuò)測量來進(jìn)行100%的人為檢查5中一些統(tǒng)計(jì)過程控制（SPC）被用于工藝，產(chǎn)品生產(chǎn)完后做最終檢測。4中高使用SPC，對(duì)不受控制的條件有中間反應(yīng)3高用工藝能力大于1.33的一個(gè)有效的SPC程序2很高所有產(chǎn)品100%自動(dòng)檢查。1幾乎確定缺陷很明顯，定期校驗(yàn)的100%的自動(dòng)檢查和檢查設(shè)備的預(yù)防性維護(hù)。Dr.GaryHarbour,PfizerEXAMPLESeverity/Probability/Detection(SPD)PhDR.C.Mendson,Menson&Associations,Inc

ICHEWGLondon,March2004EXAMPLEI.2:FailureModeEffectsAnalysis(FMEA)Severity(S)LinktoendproductfunctionalfailureMedicalDepartmentinvolvementProbability(P)UsehistoricaldataSimilarprocessesproductsDetectionMethodvalidationstudiesHistoricaldataEXAMPLETakayoshiMatsumura,EisaiCo.DryingProcessI.2:失效模式效應(yīng)分析

(FMEA)嚴(yán)重性

(S)聯(lián)系最終產(chǎn)品功能性失效引入醫(yī)學(xué)部門可能性

(P)使用歷史數(shù)據(jù)類似工藝產(chǎn)品可檢測性

方法驗(yàn)證研究歷史數(shù)據(jù)EXAMPLETakayoshiMatsumura,EisaiCo.干燥過程/工藝RankingSeverity(S)Probability(P)Detection(D)10DeathMorethanonceadayImpossibletodetect9↓3–4timesadayRemote8PermanentinjuryOnceaweekVeryslight7↓OnceamonthSlight6TemporaryinjuryOnceinthreemonthLow5↓Onceinhalf–oneyearMedium4Reported/dissatisfiedOnceayearModeratelyhigh3↓Oncein1–3yearsHigh2Notice/noreportOncein3–5yearsVeryHigh1↓Lessthanoncein5yearsVirtuallycertainI.2:FailureModeEffectsAnalysis(FMEA)TakayoshiMatsumura,EsaiCoEXAMPLEDryingProcess級(jí)別嚴(yán)重性

(S)可能性

(P)可檢測性

(D)10死亡超過一天一次不可能9↓3–4次/每天微弱的8永久傷害1次/周非常微小7↓1次/月微小6暫時(shí)傷害1次/3月低5↓1次/6月中4報(bào)告的/不高興1次/年適度高3↓1次/1~3年高2通知/無報(bào)告1次/3~5年很高1↓<1次/5年近乎必然I.2:失效模式效應(yīng)分析

(FMEA)TakayoshiMatsumura,EsaiCoEXAMPLE干燥過程/工藝ProcessPotentialFailureModePotentialCauseSPDRPN1.Setupcontaminationdisheveledgownofoperatorinsufficientcleaningofequipment2.Startdryingcontaminationdamageofinlet-airfilterdegradationofproductdamageofthermometer3.Maintaintemperaturelongdryingtimeunstablesupply-airvolumehighLossOnDrying(LOD)damageoftimerlowLODhighdew-pointnon-uniformityofLODuneventemperaturedistributionDryingProcessI.2:FailureModeEffectsAnalysis(FMEA)TakayoshiMatsumura,EsaiCoEXAMPLERPN:RiskPriorityNumber=S*P*D工藝潛在失效模式潛在原因SPDRPN1.進(jìn)料污染操作者不潔凈的服裝設(shè)備清洗不充分2.開始干燥污染進(jìn)口過濾器損壞產(chǎn)品降解溫度計(jì)損壞3.保持溫度長的烘干時(shí)間不穩(wěn)定的供應(yīng)空氣體積高的干燥失重計(jì)時(shí)器損壞低的干燥失重高露點(diǎn)不均勻的干燥失重不均勻的溫度分布干燥過程/工藝I.2:失效模式效應(yīng)分析

(FMEA)TakayoshiMatsumura,EsaiCoEXAMPLERPN:風(fēng)險(xiǎn)優(yōu)先因子

=S*P*DExistingcontrols:IPCofLODanddegradationproductafterdryingprocessDryingProcessProcessPotentialFailureModePotentialCauseSPDRPN1.Setupcontaminationdisheveledgownofoperator358120insufficientcleaningofequipment7281122.Startdryingcontaminationdamageofinlet-airfilter736126degradationofproductdamageofthermometer733633.Maintaintemperaturelongdryingtimeunstablesupply-airvolume24540highLODmalfunctionoftimer2228lowLODhighdue-point33327non-uniformityofLODuneventemperaturedistribution35345I.2:FailureModeEffectsAnalysis(FMEA)TakayoshiMatsumura,EisaiCoEXAMPLERPN:RiskPriorityNumber=S*P*D現(xiàn)有的控制:LOD的IPC和

烘干過程后的降解產(chǎn)品干燥過程/工藝工藝潛在失效模式潛在原因SPDRPN1.進(jìn)料污染操作者不潔凈的服裝358120設(shè)備清洗不充分7281122.開始干燥污染進(jìn)口過濾器損壞736126產(chǎn)品降解溫度計(jì)損壞733633.保持溫度長的烘干時(shí)間不穩(wěn)定的供應(yīng)空氣體積24540高的干燥失重計(jì)時(shí)器損壞2228低的干燥失重高露點(diǎn)33327不均勻的干燥失重不均勻的溫度分布35345I.2:失效模式效應(yīng)分析

(FMEA)TakayoshiMatsumura,EisaiCoEXAMPLERPN:風(fēng)險(xiǎn)優(yōu)先因子=S*P*DTakeactionwhenRPNisover100Takeactionwhenseverityisover5Remainingcriticalparametersaftertakingaction;furthercontrolsrequiredDryingProcessProcessPotentialCauseRPNRecommendedActionSPDRPN1.Setupdisheveledgownofoperator120uselongglovesandgoggles32848insufficientcleaningofequipment112changecleaningprocedure724562.Startdryingdamageofinlet-airfilter126changemaintenanceperiod72684damageofthermometer63changecalibrationperiod723423.Maintaintemperatureunstablesupply-airvolume40―24540malfunctionoftimer8―2228highdew-point27―33327uneventemperaturedistribution45―35345I.2:FailureModeEffectsAnalysis(FMEA)BasedonTakayoshiMatsumura,EsaiCo.EXAMPLERPN:RiskPriorityNumber=S*P*D當(dāng)RPN大于40采取措施當(dāng)嚴(yán)重性超過5時(shí)采取措施采取措施后保留關(guān)鍵參數(shù);需要進(jìn)一步控制干燥過程/工藝工藝潛在原因RPN建議措施SPDRPN1.進(jìn)料操作者不潔凈的服裝120使用長手套和眼睛32848設(shè)備清洗不充分112變更清洗程序724562.開始干燥進(jìn)口過濾器損壞126變更維護(hù)周期72684溫度計(jì)損壞63變更檢驗(yàn)周期723423.保持溫度不穩(wěn)定的供應(yīng)空氣體積40―24540計(jì)時(shí)器損壞8―2228高露點(diǎn)27―33327不均勻的溫度分布45―35345I.2:失效模式效應(yīng)分析

(FMEA)BasedonTakayoshiMatsumura,EsaiCo.EXAMPLERPN:風(fēng)險(xiǎn)優(yōu)先因子=S*P*DI.2:FailureModeEffectsAnalysis(FMEA)Analyseagranulationprocessstep

becauseonlyafewparametersareadjustableandmanyproblemscanoccurbymanualoperationsEXAMPLES.R?nninger,RocheI.2:失效模式效應(yīng)分析(FMEA)分析制粒工藝步驟因?yàn)閮H僅很少的參數(shù)可以調(diào)整，手工操作許多問題都可能發(fā)生。EXAMPLES.R?nninger,RocheFailureModeEffectsAnalysis(FMEA)

EXAMPLES.R?nninger,RocheI.2FailureModeEffectsAnalysis(FMEA)EXAMPLES.R?nninger,RocheI.2:FailureModeEffectsAnalysis(FMEA)PrepareariskprofileSeverity/Consequencesi negligible

ii marginaliii criticaliv catastrophicProbabilityA frequent

B moderate

C occasionalD rareE unlikelyF veryunlikelyProbabilityConsequencesRisk

protection

levelEXAMPLES.R?nninger,RochePicture:?ZurichInsuranceLtd,SwitzerlandI.2:失效模式效應(yīng)分析(FMEA)準(zhǔn)備一個(gè)風(fēng)險(xiǎn)輪廓嚴(yán)重性/重要性i 可以忽略

ii 邊際iii 嚴(yán)重的iv 災(zāi)難性的可能性A 頻繁的B 中等的

C 偶爾的D 很少E 不太可能F 完全不可能可能性重要性風(fēng)險(xiǎn)保護(hù)水平EXAMPLES.R?nninger,RochePicture:?ZurichInsuranceLtd,SwitzerlandI.2:FailureModeEffectsAnalysis(FMEA)Prepareariskprofile:ProbabilityEXAMPLES.R?nninger,RocheI.2:失效模式效應(yīng)分析(FMEA)準(zhǔn)備一個(gè)風(fēng)險(xiǎn)輪廓:可能性EXAMPLES.R?nninger,RocheI.2:FailureModeEffectsAnalysis(FMEA)RiskEvaluationPrepareariskprofile:ConsequencesEXAMPLES.R?nninger,

RocheI.2:失效模式效應(yīng)分析(FMEA)風(fēng)險(xiǎn)評(píng)估準(zhǔn)備一個(gè)風(fēng)險(xiǎn)輪廓:重要性（后果）EXAMPLES.R?nninger,

RocheI.2:FailureModeEffectsAnalysis(FMEA)RiskEvaluationPrepareariskprofile:ConsequencesEXAMPLES.R?nninger,

RocheI.2:失效模式效應(yīng)分析(FMEA)風(fēng)險(xiǎn)評(píng)估準(zhǔn)備一個(gè)風(fēng)險(xiǎn)輪廓:重要性（后果）EXAMPLES.R?nninger,

Roche風(fēng)險(xiǎn)評(píng)估:

風(fēng)險(xiǎn)輪廓

Forhighrisks,whicharenotacceptable,riskreductionmeasureshavetobetakenasahighpriorityI.2:失效模式效應(yīng)分析(FMEA)EXAMPLES.R?nninger,RocheHighriskLowriskRiskEvaluation:RiskProfile

對(duì)于不可接受的高風(fēng)險(xiǎn)，因?yàn)橐粋€(gè)高的風(fēng)險(xiǎn)因子風(fēng)險(xiǎn)消減措施必須實(shí)施I.2:FailureModeEffectsAnalysis(FMEA)EXAMPLES.R?nninger,RocheHighriskLowriskI.2:FailureModeEffectsAnalysis(FMEA)Howtoperform?3.Summary(RiskEvaluation)Theeffectsareratedintermsoftheir

consequencesandthe

causesareassessedintermsoftheirprobabilities

a)qualitativeorb)quantitativeBasedontheseresultsariskprofileiscompleted.Inthisprofiletherisksarecomparedwiththe

riskprotectionlevel,whichdeterminestheacceptedprobabilityfordefinedconsequencesUseasanaidtoprioritiseactions!EXAMPLES.R?nninger,RocheI.2:失效模式效應(yīng)分析(FMEA)如何實(shí)施?3.總結(jié)(風(fēng)險(xiǎn)評(píng)估)效果在術(shù)語后果中評(píng)估，原因在術(shù)語可能性中評(píng)估a)定性或

b)定量在這些結(jié)果基礎(chǔ)上完成一個(gè)風(fēng)險(xiǎn)矩陣在這些矩陣中風(fēng)險(xiǎn)使用風(fēng)險(xiǎn)保護(hù)水平來比較。為描述的后果來確定可接受的可能性。將預(yù)先的措施當(dāng)作幫助EXAMPLES.R?nninger,RocheI.2:FailureModeEffectsAnalysis(FMEA)QRMforfacilities,equipmentandutilities

Assessanexistingcompressedairsys