|
FYR
FieldFailureRateEstimatefromHALTResults
OverviewofEstimator
TheAFREstimatorisapatentpendingmathematicalmodelthat,whenprovidedwiththeappropriateHALTand
productinformation,willaccuratelyestimatetheproduct’sfieldAFRorAnnualFailureRate.Thismethodology
hasbeenusedonanumberofproductswithsignificantpositivefinancialresults.Thismodelwillalsoprovide
HASSorHASAdataforthedetectionofanoutgoingqualityprocessshiftintimefromastatedAFR.Herearea
fewreasonsforwhyandwhenthismodelshouldbeused:
•AsHALTtakesonlyafewdaystorunandtoimplementitscorrectiveaction(s),andevenifittookabit
longer,thistimewouldbefarlessthanwaitingforalifeoracceleratedlifetesttoberunandto
implementitscorrectiveaction(s).Theapplicationofthismodelcanbeahugetimeandcostsaver.
•AshigherHALTlimitsequatetolowerAFR,younowhaveatoolthatcanaccuratelyestimatethefieldAFR
beforelaunchingtheproduct.StresslevelsthataredepictedinthetableinSectionEarehighly
recommendedforHALT.Theselevelscanassuretheproducerthattheproductwillexceedcustomer
expectationsandallowtheproducertoaccuratelyforecastwarrantyexpenditures.
•BynotperforminglifetestsandsimplydoingHALT,timeandmoneywillbesaved.Thisisnottosaythat
lifetestingisn’timportant.Itshouldbeconsideredfornewtechnologiesandforanexistingpart/design
withadifferentapplicationbutnotasaprocesstoaccuratelyestimateAFR.
•WithseventotensimpledataentrypointsandmostofthemcomingfromtheHALTeffort,theAFR
EstimatorwillprovideanaccuratefieldAFRinstantaneouslywithitsassociated90%statisticalconfidence
limits.TheinputsforHASSandHASAare:willyouperformHASSorHASA,thedailysamplesize,andthe
detectableshiftintheAFRyouwishtodetect.
•TheAFREstimatorhasbeenvalidatedonovertwentyproductsfromdiversemanufacturersanddesign
environments.
•ThemodelcanaccommodateHALTsamplessizesfromonetosixwiththeoptimumsizebeingfour.
Samplesizesofgreaterthanfourwilldefaulttofour.
•90%upperandlowerconfidencelimitsarecalculatedbasedontheHALTAFRandtheHALTSampleSize. |
本帖子中包含更多资源
您需要 登录 才可以下载或查看,没有账号?-注 册-
×
|