《σ标准差教材》PPT课件.ppt

AnOverview NotalotofDetails 6 Overview SixSigma Ifwecan texpresswhatweknowintheformofnumbers wereallydon tknowmuchaboutit Ifwedon tknowmuchaboutit wecan tcontrolit Ifwecan tcontrolit weareatthemercyofchance MikelJ HarryPresident CEOSixSigmaAcademy Inc ARigorousMethodforMeasuring ControllingOurQuality willbringGEtoawholenewlevelofqualityinafractionofthetimeitwouldhavetakentoclimbthelearningcurveonourown JohnF Welch Jr 1995GEAnnualReport 6 Overview WhatDoes Sigma Mean SigmaisaMeasureoftheConsistencyofaProcess It isAlsothe18thLetterintheGreekAlphabet WhyDoesGENeedAQualityInitiative GERaisingTheBarNewGoaltobe BestintheWorld vs 1or 2CustomersareExpectingMore weMustDeliver Ship and fix ApproachnoLongerToleratedintheMarketAimtoSpeedPastTraditionalCompetitorsin5YearsGoalConsistentwithReducedTotalCostsWeMustAcknowledgeOurVulnerabilitiesPoorQualityThatImpactsCustomersProblemswithNPITooHighInternalCosts 6 Overview WeNeedaMajorInitiativetoMoveFromWhereweAretoWhereweWanttobe 6 Overview WhyDoesGENeedAQualityInitiative 40 35 30 25 20 10 15 5 CostofFailure ofSales DefectsperMillion 3 4 233 6210 66 807 308 537 500 000 Sigma 6 5 4 3 2 1 EstimatedCostofFailureinUSIndustryis15 ofSales TakingGEFroma3 toa6 CompanyWillSave 10 5BillionperYear Why SixSigma ProvenSuccessfulin Quality Demanding Industriese g Motorola TexasInstruments manyprocessstepsinseries ProvenMethodtoReduceCostsHighlyQuantitativeMethod ScienceandLogicInsteadofGutFeelIncludesManufacturing Service closetocustomer andProvidesBridgetoDesignforQualityConceptsHasSupportandCommitmentofTopManagement ItWorks 6 Overview 6 isSeveralOrdersofMagnitudeBetterThan3 Sigma AMeasureofQuality 6 Overview WhereDoes SixSigma ComeFrom MikelJ HarryoneoftheOriginalArchitectsPreviouslyHeadedQualityFunctionatABBandMotorolaNowPresident CEOofSixSigmaAcademyinPhoenix ArizonaHasConsultedforTexasInstruments AlliedSignal andothers CurrentlyRetainedbyGEtoTeachtheImplementation DeploymentandApplicationofSixSigmaConcepts Tools LearningfromThoseWhoHavehadSuccessWith6 WillAccelerateitsImplementationatGE 6 Overview So WhatisSixSigma THESIXSIGMABREAKTHROUGHSTRATEGY 6 Overview HowDoWeArriveatSigma Measuring EliminatingDefectsisthe Core ofSixSigma MeasurementSystem IdentifytheCTQs LookforDefectsinProductsorServices CriticaltoQuality CharacteristicsortheCustomerRequirementsforaProductorService CountDefectsorfailurestomeetCTQrequirementsinallprocesssteps DefineDefectOpportunities AnystepintheprocesswhereaDefectcouldoccurinaCTQ ArriveatDPMO UsetheSIGMATABLE ConvertDPMOtoSigma DefectsPerMillionOpportunities 23456 308 53766 8076 2102333 4 PPM DefectsperMillionofOpportunity SigmaLevel 6 Overview MeasurementSystem 23456 308 53766 8076 2102333 4 PPM SIGMALEVEL DEFECTSperMILLIONOPPORTUNITY IRSTaxAdvice BestCompanies AirlineSafety AverageCompany GE AverageCompanyin3 to4 Range SomeSigma Benchmarks 6 Overview MeasurementSystem AGraphic QuantitativePerspectiveonVariation AverageValue ManyDataSetsHaveaNormalorBellShape NumberofPeopleArrivingatCRD Time 7 00 7 15 7 30 7 45 8 00 8 15 8 30 8 45 9 00 9 15 6 Overview ProblemSolvingApproach 6 HelpsusIdentifyandReduceVARIATIONdueto InsufficientProcessCapability UnstableParts Materials InadequateDesignMargin Target USL LSL Target USL LSL Target USL LSL CenterProcess ReduceSpread Off Target Unpredictable On Target 6 Overview ProblemSolvingApproach LowerSpecificationLimit UpperSpecificationLimit LessVariationMeansFewerDefects HigherProcessYields 6 Overview ProblemSolvingApproach KeyComponentsof BREAKTHROUGHSTRATEGY IdentifyCTQ CTP CriticaltoProcess VariablesDoProcessMappingDevelopandValidateMeasurementSystems BenchmarkandBaselineProcessesCalculateYieldandSigmaTargetOpportunitiesandEstablishImprovementGoalsUseofParetoChart FishboneDiagrams UseDesignofExperimentsIsolatethe VitalFew fromthe TrivialMany SourcesofVariationTestforImprovementinCenteringUseofBrainstormingandActionWorkouts SetupControlMechanismsMonitorProcessVariationMaintain InControl ProcessesUseofControlChartsandProcedures AMixofConceptsandTools WillAlsoIntegratewithNPIProcess 6 Overview DisciplinedChangeProcess ANewSetofQUALITYMEASURES CustomerSatisfactionCostofPoorQualitySupplierQualityInternalPerformanceDesignforManufacturability WillApplytoManufacturing Non ManufacturingProcessesandbeTracked ReportedbyEachBusiness 6 Overview Structure 6 ProjectswiththeGEBusinesses TabulationofGESixSigmaResults BenefitTarget Update Currentbenefitslevel 10 865MM QPIDloading Carryoverfrom1999 4 059CompletedProjects2000 3 313ActiveProjects2000 3 285Total 10 865MM KeyConcepts Tools 6 Overview 6 Overview ChangingFocusFromOutputtoProcess IdentifyingandFixingRootCausesWillHelpusObtaintheDesiredOutput f X Y ProcessCapability 6 Overview SustainedCapabilityoftheProcess longterm USL T InherentCapabilityoftheProcess shortterm LSL Target OverTime a Typical ProcessWillShiftandDriftbyApproximately1 5 6 Overview ShortTermCentered versus LongTermShifted SixSigmaCentered LSL T ProcessCapability 001ppm 001ppm 6 LSL USL T 3 4ppm SixSigmaShifted1 5 ProcessCapability HigherDefectYieldinLongTermProcessCapabilitythanShortTermProcessCapability 6 6 Overview TyingitAllTogether shift C D A B 0 51 01 52 02 5 123456 CONTROL POOR GOOD TECHNOLOGY POOR GOOD shortterm ProblemCouldbeControl TechnologyorBoth 6 Overview ShortTermCapability ShortTermCapabilityRatio Cp Cp LSL 6 USL Example 6 3 0 3 0 Cp Cp 1 LSL USL 2 5 0 5 3 0 ProcessMean T Target A3 Process ThePotentialPerformanceofaProcess ifitWereonTarget 6 Overview LongTermCapability Cpk LongTermCapabilityRatio Example Cp 1 previouschart Target 0 5 0 Cpk 1 0 5 0 3 Cpk 0 83 Off TargetPenalty ThePotentialPerformanceofaProcess CorrectedforanOff TargetMean LSL USL 2 5 0 5 3 0 ProcessMean T Target A3 Process 6 Overview Z ScaleofMeasure Z AUnitofMeasureEquivalenttotheNumberofStandardDeviationsthataValueisAwayfromtheTargetValue 3 0 0 5 3 0 Z Values USL LSL 2 5 0 5 3 0 ProcessMean Z T Target 0 A3 Process TheDefinitionsofYield FirstTimeYield Yft UnitsPassed UnitsTested 65 70 0 93 RolledThruputYield Yrt Yield1 Yield2 Yield3 91 82 65 70 0 65 100 91 70 82 NormalizedYield Ynm 1 n Yrt 0 65 1 4 0 89 n TotalNumberofProcesses 6 Overview YieldExclusiveofRework ProbabilityofZeroDefects AverageYieldofAllProcesses 6 Overview AstheNumberofOperationsIncreases aHighRolledYieldRequiresaHigh forEachOperation 5 4 3 6 ProcessMeanShifted1 5 atEachOperation 6 Overview Baselining BenchmarkinganExistingProcess p x Defects Benchmark Baseline Entitlement Baselining CurrentProcess Benchmarking UltimateGoal SomeBasic6 RelatedTools 6 Overview ScatterDiagram OverSlept CarWouldNotStart Weather FamilyProblems Other ParetoDiagram FrequencyofOccurence ReasonsforBeingLateforWork ArrivalTimeatWork TimeAlarmWentOff Materials People TheHistogram 6 Overview SomeBasic6 RelatedTools PlotofDailyArrivalTime 9 15 7 00 7 15 7 30 7 45 8 00 8 15 8 30 8 45 9 00 AverageValue NumberofPeopleArrivingatCRD Time 6 Overview LCL UCL RangeChart SomeBasic6 RelatedTools MonitorsChangesinAverageorVariationOverTime DesignofExperiments 6 Overview SCREENING OPTIMIZATION CHARACTERIZATION ForExperimentsInvolvingaLargeNumberofFactorsUsefulinIsolatingthe VitalFew fromthe TrivialMany ForExperimentsInvolvingaRelativelySmallNumberofFactorsUsefulWhenStudyingRelativelyUncomplicatedEffects Interactions ForExperimentsInvolvingOnly2or3FactorsUsefulWhenStudyingHighlyComplicatedEffects Relationships DOEisMoreEffectiveThanTestingOneFactorataTime 6 Overview Usingthe OneFactorataTime Approach TimeofDeparture 321 7 15 7 30 7 45 8 00 8 15 Route CombinationSelected TheResult UseRoute2andLeaveat7 15toReachGoal 6 Overview Using DesignofExperiments DOE TimeofDeparture DOE i BetterAccountsforInteractiveVariablesMissedby OneFactorataTime and ii EfficientlySearchesfor SweetSpot inParameterSpace TheVariables TimeofDeparturefromHome RouteTakentoWork TheResult ABetterCombinationAllowing15MoreMinutesofSleep ActualCommutingTimeAverages minutes 321 7 15 7 30 7 45 8 00 8 15 Route 1720232119 1518201916 1215212018 OriginalConclusion BestCombination SweetSpot APracticalExample The Cookbook 6 Overview 6 andBakingBread Usinga12StepProcess 6 Overview WhatisImportanttotheCustomer RiseTextureSmellFreshnessTaste Y Taste 6 Overview Measure 6 Overview HowCouldWeMeasureTaste Y PanelofTastersRatingSystemof1to10Target AverageRatingat8Desired NoIndividualRatings defects Below7 Y 12345678910 Target Defects Worst Best But IsthistheRightSystem Measure 6 Overview HowCouldWeApproachThis BlindfoldedPanelRatesSeveralLoafSamplesPut Repeat PiecesfromSameLoafinDifferentSamplesConsistentRatings onPiecesfromSameLoaf Repeatability ConsistentRatings onSamplesAcrossthePanel Reproducibility Repeatability Reproducibility SuggestValidMeasurementApproach PanelMember Loaf1Loaf2Loaf3 A589B491C492D898E482F591G892 Within OneTasteUnit Measure 6 Overview Thisisa3 Process 7Defects ratingsbelow7 24Ratings fromourpanel 292 292 000Defectsper1 ooo oooLoaves OR Analyze HowDoWeApproachThis BakeSeveralLoavesUnder Normal ConditionsHaveTasterPanelAgainDotheRatingAverageRatingis7 4ButVariationistooGreatfora6 Process 6 Overview HowdoweDefineImprovement BenchmarktheCompetitionFocusonDefects i e tasterating 7 DetermineWhatisan AcceptableSigmaLevel SetImprovementObjectivesAccordingly Maybea5 ProcessWillSuffice 1 000 000 100 000 10 000 1 000 100 10 1 234567 BETTERBREAD BakingProcess BestCompetitor RangeforImprovement DefectsPerMillion SigmaScale Analyze 6 Overview HowdoweDeterminethePotentialSourcesofVariation Xs HavetheChefsBrainstormSomeLikelyOnesMightbe AmountofSaltUsed BrandofFlour BakingTime BakingTemperature BrandofYeast MultipleSources Chefs Suppliers Controls Analyze 6 Overview HowdoweScreenforCausesofVariation Xs DesignanExperimentUseDifferentSourcesofPotentialVariationHavePanelRatetheBreadUsedintheExperimentResultsLeadtothe VitalFew Causes Source Conclusion Negligible MajorCause Negligible MajorCause Negligible FocusonThe VitalFew Improve 6 Overview HowdoweFindtheRelationshipBetweenthe VitalFew Xs andTaste Y ConductaMoreDetailedExperimentFocus OvenTemperaturefrom325to375and3BrandsofFlourRUN TEMPBRAND1325A2325B3325C4350A5350B6350C7375A8375B9375C Improve Note TimeisaFactorOnlyifTemperatureChangesSignificantly 6 Overview But IsOurMeasurementSystemCorrect Improve 6 Overview HowCouldWeApproachThis NeedtoVerifytheAccuracyofOurTemperatureGaugesNeedfor Benchmark InstrumentationforComparisonRentSomeOther HighEnd GaugesComparetheResults VerifythatourInstrumentsareAccurate Control 6 Overview HowCouldWeApproachThis CheckANumberofOvensMonitorTemperaturesOverTimeFocusontheProcessCapabilityLookforDegreeofVariation VariationOKBut AverageisHigh andthealgorithmshouldbechecked Control 6 Overview WhatdowedoGoingForward CheckOvensDailyforTemperatureLevelsAuditUsageFrequencyofAlternativeFlourSupplier e g BrandC PeriodicallyReassemblethePaneltoTestTasteCharttheResults And PlottheDataOverTime 354353352351350349348 135791113151719212325 Control