10 Steps to Conduct a PFMEA(創(chuàng)建 FMEA 權(quán)威 步驟 )

1、10 Steps to Conduct a PFMEA HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#Step_1:_Review_the_Process l Step_1:_Review_the_Process Step 1Review the processUse a process flowchart to identify each process component. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#Step_2:_Brainstorm_Potential_Failure

2、_Modes l Step_2:_Brainstorm_Potential_Failure_Modes Step 2Brainsttorm ppotenttial ffailurre moddesReeview existting ddocumeentatiion annd datta forr cluees. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#Step_3_List_Potential_Effects_of_Failure l Step_3_List_Potential_Effects_of_Failure Step 3List

3、pootentiial efffectss of ffailurreTheere maay be more than one ffor eaach faailuree. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#Step_4_Assign_Severity_Rankings l Step_4_Assign_Severity_Rankings Step 4Assign Severrity rrankinngsBaased oon thee seveerity of thhe connsequeences of faailuree. HYPER

4、LINK /resources/fmea/fmea_10step_pfmea.htm#Step_5_Assign_Occurrence_Rankings l Step_5_Assign_Occurrence_Rankings Step 5Assign Occurrrencee rankkingsBasedd on hhow frrequenntly tthe caause oof thee faillure iis likkely tto occcur. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#Step_6_Assign_Detectio

5、n_Rankings l Step_6_Assign_Detection_Rankings Step 6Assign Detecction rankiingsBBased on thhe chaances the ffailurre willl be deteccted pprior to thhe cusstomerr findding iit. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#Step_7_Calculate_the_RPN l Step_7_Calculate_the_RPN Step 7Calculaate thhe RP

6、NNSeveerity X Occcurrennce X Detecction. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#Step_8_Develop_the_Action_Plan l Step_8_Develop_the_Action_Plan Step 8Developp the actioon plaanDeffine wwho wiill doo whatt by wwhen. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#Step_9_Take_Action l Step_9_

7、Take_Action Step 9Take acctionImpleement the iimprovvementts ideentifiied byy yourr PFMEEA teaam. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#Step_10_Recalculate_the_Resulting_RPN_ l Step_10_Recalculate_the_Resulting_RPN_ Step 100Calculaate thhe ressultinng RPNNRe-eevaluaate eaach off the potenn

8、tial failuures oonce iimprovvementts havve beeen madde andd deteerminee the impacct of the iimprovvementts. Step 1: Revview tthe Prrocesss Review the pprocesss commponennts annd thee inteended functtion oor funnctionns of thosee compponentts. Use of a dettailedd flowwchartt of tthe prrocesss or aa t

9、ravveler (or rrouterr) is a goood staartingg poinnt forr reviiewingg the proceess. There aare seeverall reassons ffor reeviewiing thhe proocessFirst, the rrevieww helpps asssure tthat aall teeam meemberss are familliar wwith tthe prrocesss. Thiis is especciallyy impoortantt if yyou haave teeam meemb

10、erss who do noot worrk on the pprocesss on a daiily baasis. The seccond rreasonn for revieewing the pprocesss is to iddentiffy eacch of the mmain ccomponnents of thhe proocess and ddetermmine tthe fuunctioon or functtions of eaach off thosse commponennts. Finallyy, thiis revview sstep wwill hhelp aa

11、ssuree thatt you are sstudyiing alll commponennts off the proceess wiith thhe PFMMEA. Using tthe prrocesss flowwchartt, labbel eaach coomponeent wiith a sequeentiall refeerencee numbber. These rrefereence nnumberrs willl be used throuughoutt the FMEA proceess. The marrked-uup floowcharrt willl givve

12、 youu a poowerfuul vissual tto reffer too throoughouut thee PFMEEA. With thhe proocess flowcchart in haand, tthe PFFMEA tteam mmemberrs shoould ffamiliiarizee themmselvees witth thee proccess bby phyysicallly waalkingg throough tthe prrocesss. Thiis is the ttime tto asssure eeveryoone onn the team u

13、nderrstandds thee basiic proocess flow and tthe woorkinggs of the pprocesss commponennts. For eacch commponennt, liist itts inttendedd funcction or fuunctioons. The funnctionn of tthe coomponeent iss the valuee-addiing roole thhat coomponeent peerformms or proviides. Many coomponeents hhave mmore tt

14、han oone fuunctioon. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#10_Steps_to_Conduct_a_PFMEA l 10_Steps_to_Conduct_a_PFMEA Step 2: Braainstoorm Pootentiial Faailuree Modees In Stepp 2, cconsidder thhe pottentiaal faiilure modess for each compoonent and iits coorresppondinng funnctionn. A potennt

15、ial failuure moode reepreseents aany maanner in whhich tthe coomponeent orr proccess sstep ccould fail to peerformm its intennded ffunctiion orr funcctionss. Using tthe liist off compponentts andd relaated ffunctiions ggeneraated iin Steep 1, as a team, braiinstorrm thee poteentiall faillure mmodes

16、for eeach ffunctiion. Dont ttake sshortccuts hhere; this is thhe timme to be thhorouggh. Preparee for the bbrainsstormiing seessionn. Before you bbegin the bbrainsstormiing seessionn, revview ddocumeentatiion foor cluues abbout ppotenttial ffailurre moddes. HYPERLINK /resources/fmea/fmea_10step_pfme

17、a.htm#10_Steps_to_Conduct_a_PFMEA l 10_Steps_to_Conduct_a_PFMEA Step 3: Lisst Pottentiaal Efffects of FaailureeDetermiine thhe efffects assocciatedd withh eachh faillure mmode. The eeffectt is rrelateed dirrectlyy to tthe abbilityy of tthat sspeciffic coomponeent too perfform iits inntendeed funncti

18、onn. An effeect iss the impacct a ffailurre couuld maake iff it ooccurrred. Some faailurees willl havve an effecct on the ccustommers aand otthers on thhe envvironmment, the ffaciliity, aand evven thhe proocess itsellf. As withh faillure mmodes, use descrriptivve andd detaailed termss to ddefinee ef

19、feects. The efffect sshouldd be sstatedd in tterms meaniingfull to pproducct or systeem perrformaance. If the effeccts arre deffined in geenerall termms, itt willl be ddifficcult tto ideentifyy (andd reduuce) ttrue ppotenttial rrisks. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#10_Steps_to_Condu

20、ct_a_PFMEA l 10_Steps_to_Conduct_a_PFMEA Step 4: Asssign SSeveriity RaankinggsAssign a sevverityy rankking tto eacch efffect tthat hhas beeen iddentiffied. The sevverityy rankking iis an estimmate oof howw seriious aan efffect wwould be shhould it occcur. To deteerminee the severrity, consiider tthe

21、 immpact the eeffectt woulld havve on the ccustommer, oon dowwnstreeam opperatiions, or onn the emplooyees operaating the pprocesss. The sevverityy rankking iis bassed onn a reelativve scaale raangingg fromm 1 too 10. A “10” meanss the effecct hass a daangeroously high severrity lleadinng to a hazza

22、rd wwithouut warrning. Converssely, a sevverityy rankking oof “1” meaans thhe sevverityy is eextremmely llow. The rannking scalees (foor sevverityy, occcurrennce, aand deetectiion) aare miissionn crittical for tthe suuccesss of aa PFMEEA beccause they estabblish the bbasis for ddetermminingg riskk o

23、f oone faailuree modee and effecct rellativee to aanotheer. The samme rannking scalees forr PFMEEAs shhould be ussed coonsisttentlyy throoughouut youur orgganizaation. Thiss willl makee it ppossibble too comppare tthe RPPNs frrom diiffereent FMMEAs tto onee anotther. See HYPERLINK /resources/fmea/in

24、dex.htm FMEEA Cheecklissts annd Forrms foor an exampple PFFMEA SSeveriity Raankingg Scalle. The besst wayy to ccustommize aa rankking sscale is too starrt witth a sstandaard, ggeneriic scaale annd theen moddify iit to be moore meeaninggful tto youur orgganizaation. As you add eexamplles sppecifiic t

25、o your organnizatiion, cconsidder addding severral coolumnss withh eachh coluumn foocusedd on aa topiic. One toppic coould pprovidde desscripttions of seeveritty levvels ffor opperatiional failuures, anothher coolumn for ccustommer saatisfaactionn faillures, and a thiird foor envvironmmentall, heaal

26、th, and ssafetyy issuues. See HYPERLINK /resources/fmea/index.htm FMEEA Cheecklissts annd Forrms foor an exampples oof Cusstom PPFMEA Rankiing Sccales. (Exaampless of ccustomm scalles foor sevverityy, occcurrennce, aand deetectiion raankinggs aree inclluded in thhis Apppendiix.) HYPERLINK /resources

27、/fmea/fmea_10step_pfmea.htm#10_Steps_to_Conduct_a_PFMEA l 10_Steps_to_Conduct_a_PFMEA Step 5: Asssign OOccurrrence RankiingsNext, cconsidder thhe pottentiaal cauuse orr faillure mmechannism ffor eaach faailuree modee; theen asssign aan occcurrennce raankingg to eeach oof thoose caauses or faailuree

28、mechhanismms. We needd to kknow tthe pootentiial caause tto detterminne thee occuurrencce rannking becauuse, jjust llike tthe seeveritty rannking is drriven by thhe efffect, the ooccurrrence rankiing iss a fuunctioon of the ccause. The occurrrencee rankking iis bassed onn the likellihoodd, or frequu

29、ency, thatt the causee (or mechaanism of faailuree) willl occcur. If we kknow tthe caause, we caan bettter iidentiify hoow freequenttly a speciific mmode oof faiilure will occurr. Howw do yyou fiind thhe rooot cauuse? There aare maany prroblemm-findding aand prroblemm-solvving mmethoddologiies. One

30、of the eeasiesst to use iis thee 5-Whhys teechniqque. Once thhe cauuse iss knowwn, caapturee dataa on tthe frrequenncy off causses. SSourcees of data may bbe scrrap annd rewwork rreportts, cuustomeer commplainnts, aand eqquipmeent maaintennance recorrds. The occcurrennce raankingg scalle, liike thhe

31、 sevverityy rankking, is onn a reelativve scaale frrom 1 to 100. An occuurrencce rannking of “110” meeans tthe faailuree modee occuurrencce is very high, and happeens alll of the ttime. Conveerselyy, a “1” means the probability of occurrence is remote. See HYPERLINK /resources/fmea/index.htm FMEEA C

32、heecklissts annd Forrms foor an exampple PFFMEA OOccurrrence Rankiing Sccale. Your orrganizzationn may need an occcurreence rrankinng scaale cuustomiized ffor a low-vvolumee, commplex assemmbly pprocesss or a mixxture of hiigh-voolume, simpple prrocessses annd loww-voluume, ccompleex proocessees. Co

33、nsideer cusstomizzed occcurreence rrankinng scaales bbased on tiime-baased, eventt-baseed, orr piecce-bassed frrequenncies. See HYPERLINK /resources/fmea/index.htm FMEEA Cheecklissts annd Forrms foor exaampless of CCustomm PFMEEA Rannking Scalees. (EExamplles off custtom sccales for sseveriity, oocc

34、urrrence, and detecction rankiings aare inncludeed in this Appenndix.) HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#10_Steps_to_Conduct_a_PFMEA l 10_Steps_to_Conduct_a_PFMEA Step 6: Asssign DDetecttion RRankinngsTo assiign deetectiion raankinggs, iddentiffy thee proccess oor prooduct relatted coo

35、ntrolls in placee for each failuure moode annd theen asssign aa deteectionn rankking tto eacch conntrol. Deteectionn rankkings evaluuate tthe cuurrentt proccess ccontrools inn placce. A contrrol caan rellate tto thee faillure mmode iitselff, thee causse (orr mechhanismm) of failuure, oor thee effeec

36、ts oof a ffailurre modde. To makee evalluatinng conntrolss evenn moree compplex, contrrols ccan eiither preveent a failuure moode orr causse froom occcurrinng or detecct a ffailurre modde, caause oof faiilure, or eeffectt of ffailurre aftter itt has occurrred. Note thhat prreventtion ccontrools caan

37、not relatte to an efffect. If ffailurres arre preeventeed, ann effeect (oof faiilure) cannnot exxist! The Dettectioon rannking scalee, likke thee Seveerity and OOccurrrence scalees, iss on aa relaative scalee fromm 1 too 10. A Detecction rankiing off “1” meaans thhe chaance oof dettectinng a ffailur

38、re is certaain. Converssely, a “100” meaans thhere iis abssolutee certtaintyy of nnon-deetectiion. TThis bbasicaally mmeans that theree are no coontrolls in placee to pprevennt or detecct. See HYPERLINK /resources/fmea/index.htm FMEEA Cheecklissts annd Forrms foor an exampple PFFMEA DDetecttion RRan

39、kinng Scaale. Taking a leaad froom AIAAG, coonsideer thrree diiffereent foorms oof Cusstom DDetecttion RRankinng opttions. Custtom exxamplees forr Misttake-PProofiing, GGauginng, annd Mannual IInspecction contrrols ccan bee helppful tto PFMMEA teeams. See HYPERLINK /resources/fmea/index.htm FMEEA Ch

40、eecklissts annd Forrms foor exaampless of CCustomm PFMEEA Rannking Scalees. (EExamplles off custtom sccales for sseveriity, ooccurrrence, and detecction rankiings aare inncludeed in this Appenndix.) HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#10_Steps_to_Conduct_a_PFMEA l 10_Steps_to_Conduct_a_P

41、FMEA Step 7: Callculatte thee RPNThe RPNN is tthe Riisk Prrioritty Nummber. The RRPN giives uus a rrelatiive riisk raankingg. Thee highher thhe RPNN, thee highher thhe pottentiaal rissk. The RPNN is ccalcullated by muultipllying the tthree rankiings ttogethher. MMultipply thhe Sevverityy rankking tt

42、imes the OOccurrrence rankiing tiimes tthe Deetectiion raankingg. Callculatte thee RPN for eeach ffailurre modde andd effeect. Editoriial Noote: The ccurrennt FMEEA Mannual ffrom AAIAG ssuggessts onnly caalculaating the RRPN foor thee highhest eeffectt rankking ffor eaach faailuree modee. We do noot

43、 agrree wiith thhis suuggesttion; we beelievee thatt if tthis ssuggesstion is foolloweed, itt willl be ttoo eaasy too misss the need for ffurtheer impprovemment oon a sspeciffic faailuree modee. Since eeach oof thee threee rellativee rankking sscaless rangges frrom 1 to 100, thee RPN will alwayys be

44、 betweeen 1 and 11000. The hhigherr the RPN, the hhigherr the relattive rrisk. The RRPN giives uus an excelllent tool to prriorittize ffocuseed impprovemment eeffortts. HYPERLINK /resources/fmea/fmea_10step_pfmea.htm#10_Steps_to_Conduct_a_PFMEA l 10_Steps_to_Conduct_a_PFMEA Step 8: Devvelop the AAct

45、ionn PlannTaking actioon meaans reeducinng thee RPN. The RPN ccan bee reduuced bby lowweringg any of thhe thrree raankinggs (seeveritty, occcurreence, or deetectiion) iindiviiduallly or in coombinaation with one aanotheer. A reducction in thhe Sevverityy rankking ffor a PFMEAA is ooften the mmost dd

46、ifficcult. It ussuallyy requuires a phyysicall modiificattion tto thee proccess eequipmment oor layyout. Reductiion inn the Occurrrencee rankking iis acccompliished by reemovinng or contrrollinng thee poteentiall causses. Mistakee-prooofing toolss are oftenn usedd to rreducee the frequuency of occcu

47、rreence. A reducction in thhe Dettectioon rannking can bbe acccompliished by immproviing thhe proocess contrrols iin plaace. Adding proceess faail-saafe shhut-doowns, alarmm signnals (sensoors orr SPC), andd valiidatioon praacticees inccludinng worrk insstructtions, set-up prroceduures, calibbratioo

48、n proogramss, andd prevventattive mmainteenancee are all ddetecttion rrankinng impprovemment aapproaaches. What iss conssidereed an accepptablee RPN? The answeer to that questtion ddependds on the oorganiizatioon. For exaample, an oorganiizatioon mayy deciide anny RPNN abovve a mmaximuum tarrget oof

49、 2000 pressents an unnaccepptablee riskk and must be reeducedd. If so, tthen aan acttion pplan iidentiifyingg who will do whhat byy whenn is nneededd. There aare maany toools tto aidd the PFMEAA teamm in rreduciing thhe rellativee riskk of ffailurre moddes reequiriing acction. Amonng thee mostt powe

50、erful toolss are Mistaake-Prroofinng, Sttatisttical Proceess Coontroll, andd Desiign off Expeerimennts. Mistakee-Prooofing (Pokaa Yokee) Techniqques tthat ccan maake itt impoossiblle forr a miistakee to ooccur, reduucing the OOccurrrence rankiing too 1. Especiaally iimporttant wwhen tthe Seeveritty

51、rannking is 100. Statisttical Proceess Coontroll (SPCC) A statiisticaal toool thaat hellps deefine the ooutputt of aa proccess tto detterminne thee capaabilitty of the pprocesss agaainst the sspecifficatiion annd theen to mainttain ccontrool of the pprocesss in the ffuturee. Design of Exxperimments (DOE) A familly of powerrful sstatissticall