韦伯分布.docx
《韦伯分布.docx》由会员分享,可在线阅读,更多相关《韦伯分布.docx(9页珍藏版)》请在冰豆网上搜索。
韦伯分布
韋伯分佈
韋伯分佈(Weibulldistribution)以指數分佈為一特例。
其p.d.f.為
其中α,β>0。
以表此分佈,有二參數α,β,α為尺度參數,β為形狀參數。
若取β=1,則得分佈,以表之。
底下給出一些韋伯分佈p.d.f.之圖形。
韋伯分佈是瑞典物理學家WaloddiWeibull,為發展強化材料的理論,於西元1939年所引進,是一較新的分佈。
在可靠度理論及有關壽命檢定問題裡,常少不了韋伯分佈的影子。
分佈的分佈函數為
期望值與變異數分別為
CharacteristicEffectsoftheShapeParameter,β,fortheWeibullDistribution
TheWeibullshapeparameter,β,isalsoknownastheslope.Thisisbecausethevalueofβisequaltotheslopeoftheregressedlineinaprobabilityplot.Differentvaluesoftheshapeparametercanhavemarkedeffectsonthebehaviorofthedistribution.Infact,somevaluesoftheshapeparameterwillcausethedistributionequationstoreducetothoseofotherdistributions.Forexample,whenβ=1,thepdfofthethree-parameterWeibullreducestothatofthetwo-parameterexponentialdistributionor:
wherefailurerate.
Theparameterβisapurenumber,i.e.itisdimensionless.
TheEffectofβonthepdf
Figure6-1showstheeffectofdifferentvaluesoftheshapeparameter,β,ontheshapeofthepdf.Onecanseethattheshapeofthepdfcantakeonavarietyofformsbasedonthevalueofβ.
Figure6-1:
TheeffectoftheWeibullshapeparameteronthepdf.
For0<β1:
∙As(orγ),
∙As,.
∙f(T)decreasesmonotonicallyandisconvexasTincreasesbeyondthevalueofγ.
∙Themodeisnon-existent.
Forβ>1:
∙f(T)=0atT=0(orγ).
∙f(T)increasesas(themode)anddecreasesthereafter.
∙Forβ<2.6theWeibullpdfispositivelyskewed(hasarighttail),for2.6<β<3.7itscoefficientofskewnessapproacheszero(notail).Consequently,itmayapproximatethenormalpdf,andforβ>3.7itisnegativelyskewed(lefttail).
Thewaythevalueofβrelatestothephysicalbehavioroftheitemsbeingmodeledbecomesmoreapparentwhenweobservehowitsdifferentvaluesaffectthereliabilityandfailureratefunctions.Notethatforβ=0.999,f(0)=,butforβ=1.001,f(0)=0.ThisabruptshiftiswhatcomplicatesMLEestimationwhenβisclosetoone.
TheEffectofβonthecdfandReliabilityFunction
Figure6-2:
EffectofβonthecdfonaWeibullprobabilityplotwithafixedvalueofη.
Figure6-2showstheeffectofthevalueofβonthecdf,asmanifestedintheWeibullprobabilityplot.Itiseasytoseewhythisparameterissometimesreferredtoastheslope.Notethatthemodelsrepresentedbythethreelinesallhavethesamevalueofη.Figure6-3showstheeffectsofthesevariedvaluesofβonthereliabilityplot,whichisalinearanalogoftheprobabilityplot.
Figure6-3:
TheeffectofvaluesofβontheWeibullreliabilityplot.
∙R(T)decreasessharplyandmonotonicallyfor0<β<1andisconvex.
∙Forβ=1,R(T)decreasesmonotonicallybutlesssharplythanfor0<β<1andisconvex.
∙Forβ>1,R(T)decreasesasTincreases.Aswear-outsetsin,thecurvegoesthroughaninflectionpointanddecreasessharply.
TheEffectofβontheWeibullFailureRateFunction
ThevalueofβhasamarkedeffectonthefailurerateoftheWeibulldistributionandinferencescanbedrawnaboutapopulation'sfailurecharacteristicsjustbyconsideringwhetherthevalueofβislessthan,equalto,orgreaterthanone.
Figure6-4:
TheeffectofβontheWeibullfailureratefunction.
AsindicatedbyFigure6-4,populationswithβ<1exhibitafailureratethatdecreaseswithtime,populationswithβ=1haveaconstantfailurerate(consistentwiththeexponentialdistribution)andpopulationswithβ>1haveafailureratethatincreaseswithtime.AllthreelifestagesofthebathtubcurvecanbemodeledwiththeWeibulldistributionandvaryingvaluesofβ.
TheWeibullfailureratefor0<β<1isunboundedatT=0(orγ).Thefailurerate,λ(T),decreasesthereaftermonotonicallyandisconvex,approachingthevalueofzeroasorλ()=0.Thisbehaviormakesitsuitableforrepresentingthefailurerateofunitsexhibitingearly-typefailures,forwhichthefailureratedecreaseswithage.Whenencounteringsuchbehaviorinamanufacturedproduct,itmaybeindicativeofproblemsintheproductionprocess,inadequateburn-in,substandardpartsandcomponents,orproblemswithpackagingandshipping.
Forβ=1,λ(T)yieldsaconstantvalueofor:
Thismakesitsuitableforrepresentingthefailurerateofchance-typefailuresandtheusefullifeperiodfailurerateofunits.
Forβ>1,λ(T)increasesasTincreasesandbecomessuitableforrepresentingthefailurerateofunitsexhibitingwear-outtypefailures.For1<β<2,theλ(T)curveisconcave,consequentlythefailurerateincreasesatadecreasingrateasTincreases.
Forβ=2thereemergesastraightlinerelationshipbetweenλ(T)andT,startingatavalueofλ(T)=0atT=γ,andincreasingthereafterwithaslopeof.Consequently,thefailurerateincreasesataconstantrateasTincreases.Furthermore,ifη=1theslo