计数型MSA

Copyright©2001-2005SixSigmaAcademyInternational,LLCAllrightsreserved;foruseonlyincompliancewithSSAlicense.MeasurementSystemsAnalysis(MSA)Attribute特性数据MSACopyright©2001-2005SixSigmaAcademyInternational,LLCAllrightsreserved;foruseonlyincompliancewithSSAlicense.MeasurementSystemsAnalysis(MSA)VariablesPg1TheBreakthroughStrategy®AndMeasurementSystemEvaluation•MeasurementSystemsAnalysiswillhelpusdecideifwecanadequatelymeasureouroutputcharacteristic测量系统分析有助于确认是否能够胜任测量输出特性的要求•Thenafterwe’veimprovedtheprocess,wewillneedtoassessitagainandpotentiallyneedtoimproveit/在改善过程以后，我们需要再次评价测量系统，以及潜在的测量系统的改进1.SelectOutputCharacteristic2.DefinePerformanceStandards3.ValidateMeasurementSystem确认测量系统4.EstablishBaselineProcessCapability5.DefinePerformanceObjectives6.IdentifyVariationSources7.ScreenPotentialCauses8.DiscoverVariableRelationships9.EstablishOperatingTolerances–ImplementImprovements10.ValidateMeasurementSystem确认测量系统11.DetermineFinalProcessCapability12.ImplementProcessControlsCopyright©2001-2005SixSigmaAcademyInternational,LLCAllrightsreserved;foruseonlyincompliancewithSSAlicense.MeasurementSystemsAnalysis(MSA)AttributePg2ModuleObjectivesBytheendofthismodule,theparticipantshouldbeableto:•DiscusstheneedforanAttributeMeasurementSystemAnalysis探讨属性数据测量系统分析所需条件•DescribethetypesofAttributeMeasurementSystemAnalysesandwhentousethem叙述属性数据测量系统分析以及何时使用•PerformanAttributeMeasurementSystemAnalysis实施属性数据测量系统分析Copyright©2001-2005SixSigmaAcademyInternational,LLCAllrightsreserved;foruseonlyincompliancewithSSAlicense.MeasurementSystemsAnalysis(MSA)AttributePg3TypicalMeasurementSystemsAttributeMSA•Examples-Datacollectionform-Survey-Customersatisfaction-Academicgrades-Pluggages(go/no-go)-Threadgages(go/no-go)-VisualdefectcheckNotethatoftenAttributeMeasurementSystemsinvolvehumanjudgment–Isitreliableforuseindecisionmaking?/注意属性数据MSA经常包含人员的判定–判定可靠吗？Copyright©2001-2005SixSigmaAcademyInternational,LLCAllrightsreserved;foruseonlyincompliancewithSSAlicense.MeasurementSystemsAnalysis(MSA)AttributePg4PurposeOfAttributeMSA•Accuracychecks准确性检查-Assessstandardsagainstcustomers’requirements评价标准和顾客要求对比-IdentifyhowwellMeasurementSystemconformstoa“knownmaster”/识别测量系统符合”已知标准”的程度•Precisionchecks精密性检查-Todetermineifinspectors(Appraisers)acrossallshifts,machines,lines,etc…usethesamecriteriatoevaluateitems–Reproducibility/确认检查员(评价者)所有班次、设备、生产线是否使用相同标准评价–再现性-Toquantifytheabilityofinspectors(Appraisers)orgagestoaccuratelyrepeattheirinspectiondecisions–Repeatability/量化检查员(评价者)或仪器准确重复检查结果的能力–重复性•Toidentifyhowwellinspectors/gagesmeasureaknownmaster(possiblydefinedbythecustomer)toensurenomisclassificationoccurs/识别检查员/仪器测量”已知标准”(可能是顾客定义)确保没有错误分类的能力-Howoftenoperatorsdecidetoshiptrulydefectiveproduct多久操作员会将不良品判为合格品？-Howoftenoperatorsdonotshiptrulyacceptableproduct多久操作员会将合格品判为不良品？•Todetermineareaswhere确定以下范围-Trainingisneeded是否需要培训-ProceduresorControlPlansarelacking是否缺乏程序或控制计划-Standardsarenotclearlydefined标准是否未定义-Gageadjustmentorcorrelationisnecessary仪器调整或相互性是必须的Copyright©2001-2005SixSigmaAcademyInternational,LLCAllrightsreserved;foruseonlyincompliancewithSSAlicense.MeasurementSystemsAnalysis(MSA)AttributePg5AttributeDataWithUnorderedCategories非顺序型的属性数据–Kappa•SupposethatinvoicequalityisakeytotheprocessThroughput.Inotherwords,ifaninvoiceisincomplete,thereworkrequiredimpactsthequantitythatcanbeprocessedinaday.TwoAppraisersareaskedtoindependentlyevaluateteninvoicesrandomlyselectedfromdifferentdays.Theresultsofthestudyareshownbelow:/假定发票质量是过程输出的关键，换句话说，如果一张发票不完整，再作业就会影响当天的产量。2个评价者独立评价随机在不同日期抽取的10张发票。结果如下所示：Invoice#Appraiser1Appraiser2Agreement?1BadBadY2GoodBadN3GoodGoodY4BadBadY5GoodGoodY6GoodGoodY7GoodBadN8GoodBadN9GoodGoodY10GoodGoodY•WecouldsimplylookatthepercentofthetimetheyagreeasmetricforbetweenAppraiseragreement:我们可以只考虑评价者判定一致的百分率作为衡量指标•Butwhatwouldthatnottakeintoaccount?但是还有什么没有考虑？_____%agree_____%disagreeCopyright©2001-2005SixSigmaAcademyInternational,LLCAllrightsreserved;foruseonlyincompliancewithSSAlicense.MeasurementSystemsAnalysis(MSA)AttributePg6KappacomparestheproportionofagreementbetweenAppraisersafterremovingagreementbychance/Kappa比较了评价者之间去除了的偶然性的一致性比率TheproportionthatthejudgesareinagreementisPobserved/判断一致的比率是PobservedTheproportionexpectedtooccurbychanceis:由于偶然性一致的比率Pchance=(PInsp1Good)(PInsp2Good)+(PInsp1Bad)(PInsp2Bad)1.WhatwouldtheKappabeifthejudgesagreedoneverypart?/如果每个零件的判定都一致，Kappa是多少？2.HowwouldyouinterpretaKappaofzero?怎么解释Kappa为0？3.WhatwouldbetheKappaifthejudgesneveragreed?如果判定都不一致Kappa又是多少？DefinitionOfKappachancechanceobservedP1PPKappaLet’swalkthroughthecalculationbyhand.Copyright©2001-2005SixSigmaAcademyInternational,LLCAllrightsreserved;foruseonlyincompliancewithSSAlicense.MeasurementSystemsAnalysis(MSA)AttributePg7HowToCalculateKappa–ByHandBuildcontingencytableSumcolumnsandrowsCalculatePobsbyaddingdiagonalCalculatePchanceCalculateKappaPchance=(PR1bad)(PR2bad)+(PR1good)(PR2good)=(.2)(.5)+(.8)(.5)=.5Pobs-PchanceKappa=1-Pchance=(.7-.5)/(1-.5)=0.4.2.8.5.5BadGoodAppraiser1GoodBadAppraiser2Pobs=.5+.2=.73/10=.32/10=.20/10=05/10=.5AddAddAddAddHowisthisint