Reliability Test Introduction and Application

ReliabilityTestIntroduction&ApplicationWhoshouldattendthise-LearningTargetaudience•PE,ME,QE,QAPre-requisite•BasicProbabilityandstatisticscoursesCoursecontent•What’sthereliability•Bathtubcurve•MTBF(MeanTimeBetweenFailure)•HALT(HighlyAcceleratedLifeTesting)Andthentheybeabletoknow•What’sthereliability•MethodsofestimatingtheintrinsicreliabilityofasystembeforeitsactualmanufactureanduseOutlineofReliability„0.Overview‰0.1.What’sthereliabilityTest‰0.2.Reliabilitydefinitions‰0.3.WhyneedReliabilityTest‰0.4.Reliabilitytestitems„1.BasicConcept‰1.1.Bathtubcurve‰1.2.Weibulldistribution‰1.3.MTBF(MeanTimeBetweenFailure)‰1.4.HALT(HighlyAcceleratedLifeTest)HASA(HighlyAcceleratedStressAudit)HASS(HighlyAcceleratedStressScreen)„2.ApplicationOpportunity‰2.1.CurrentourReliabilitytestitems‰2.2.ReliabilitylinkwithCflow0.1.What’stheReliabilityTestReliabilityisdefinedastheprobabilitythatagivenproductwillsuccessfullyperformarequiredfunctionwithoutfailure,underspecifiedenvironmentalconditions,foraspecifiedperiodoftime.0.2.Reliabilitydefinitions•arequiredfunction:Inmustberealized,whiledefiningthereliabilityofaproduct,itisnecessarytodefineexactlywhatfunctiontheproductisexpectedtoperformandwhatwouldbeconsideredafailure,beforeitslifecanbemeasured.•aspecifiedperiodoftime:Thespecifiedperiodoftimeisthetimeduringwhichsatisfactoryoperationisdesired,commonlyreferredtoasservicelife.•Specifiedenvironmentalconditions:Thestatedconditionsarethetotalphysicalenvironmentincludingmechanical,thermalandelectricalconditions.•ReliabilityR(t):Apointestimateofreliabilityequaltotheprobabilityofsurvivalforaspecifiedoperationtime,t,giventhattheequipmentwasoperationalatthebeginningoftheperiod.DefinitionsDefinitions0.3.WhyneedReliabilityTest9Identifyproductweakness9Reducelikelihoodproblem9EnsurecustomersatisfactionPurposePurpose0.4.ReliabilityTestItemsManywaystotestManywaystotest•TemperatureCycling•RandomVibration•HighTemperature•ElectricityStress•ThermalShock•Sinewavevibrationwithfixedfrequency•LowTemperature•Sinewavevibrationwithvariationfrequency•Combinedenvironment•Shock•Humidity•Acceleration•AltitudeEffect?Effect?Effect?Time?Time?Time?Cost?Cost?Cost?1.1.BathtubCurveExplication:Explication:ThefailurerateofaproductdependsonthechangeofthelifeThefailurerateofaproductdependsonthechangeofthelifecycle,thechangetrendchartnormallyshownlikeabathtubcurvcycle,thechangetrendchartnormallyshownlikeabathtubcurve.e.FailureRateOperationTimeInfantMortalitySteady-stateWearout1.1.BathtubCurve•InfantMortality:Thefailurerateishigh,butdecreasesrapidly.TimetofailureismodeledasWeibulldistribution.•Steady-state:Failuresoccurataconstantrate,referredtoasthesteady-statefailurerate.Thatis,thetimetofailurefollowsanexponentialdistribution.OtherprocedureslikeMIL217Fassumethatthefailurerateisalwaysconstant.•Wearout:Thefailurerateincreasesrapidly.Usually,wearoutdoesnotoccurforelectronicdevicesduringtheservicelife.BathtubCurveBathtubCurve1.2.Weibulldistribution•ProbabilityDensityFunction•ReliabilityFunction•HazardFunction,)(),;()(1βθβθθββθtettf−−=0≥t,)()(βθtetR−=0≥t,)()(1−=βθθβλtt0≥t:β:θshapeparameterscaleparameterTheWeibulldistributionisoneofthemostwidelyusedlifedistribution,especiallyinmodelinginfantmortalityfailure.1.2.Weibulldistributionβ=4.0f(t)tβ=2.0β=1.0β=0.25ProbabilityDensityFunction1.2.Weibulldistribution0.00.51.01.52.01.00.750.50.25β=4.0β=0.25β=1.0β=2.0R(t)tReliabilityFunction1.2.Weibulldistributionβ=4.0tλ(t)β=2.0β=1.0β=0.25β1:infantmortalityβ=1:steady-stateβ1:wearoutHazardRate1.3.MTBF(MeanTimeBetweenFailure)•MTBFisameasureofhowreliableaproductis.MTBFisusuallygiveninunitsofhours;thehighertheMTBF,themorereliabletheproductis.(i.e.repairable)•MTTFisabasicmeasureofreliabilityfornon-repairablesystems.Itisthemeantimeexpecteduntilthefirstfailureofapieceofequipment.(i.e.notrepairable)•MTTRistheaveragetimetorepairamachinebacktoacceptableoperatingconditions.MTBFisanabbreviationforMeanTimeBetweenFailureMTTFisanabbreviationforMeanTimeToFailureMTTRisanabbreviationforMeanTimeToRepairWhatisMTBF,MTTR,MTTF?WhatisMTBF,MTTR,MTTF?1.3.MTBFcalculationMTBF’scalculation:a.TobasedonStatisticstheory“Chi-squaredistribution”b.Touseone-Sidedtypeforthepracticalreliabilitydemonstrationtest.2T*A(t)MTBF=――――――――,TR=χ2(α,2γ+2)/2χ2(α,2γ+2)WhereT:Accumulatetesttimeγ:Numberoffailuresχ2():Chi-squaredistributionα:Producer’srisk=1–ConfidencelevelA(t):AccelerationFactor1.3.AcceleratedLifeModelAcceleratedLifemodelTheHighlyreliableandcomplexsystemrequireextremelyreliablecomponents.Wesimplycannotwaittoaccumulatefailuredatafromfieldperformance.Ã≡tn/taWheretn:timetofailureundernormalconditions.ta:timetofailureunderaccelerated(highstress)conditions.Ã:theaccelerationfactor•Themarketwindowoftheparticularproductthananelectronicdevicesisuseinissmall.Forexample:oneortwoyears,thenthetimeallowedforreliabilityevaluationofthatdevicemaybedays(orweeksatbest),notmonths(oryears).•Onesuchmethodisacceleratedlifetesting.Thetemperature-relatedfailurerate,accordingtotheArrheniusrelation,is:A=AccelerationFactorEa=Activationenergy(ineV)GeneralEaofel