六西格瑪管理的概述(英文版)

六西格瑪管理的概述(英文版)Passion+Execution=FastandLastingResultsSixSigmaFocusDelightingthecustomerthroughflawlessexecutionRapidbreakthroughimprovementAdvancedbreakthroughtoolsthatworkPositiveanddeepculturechangeRealfinancialresultsthatimpactthebottomlineWhatisSixSigma?VisionPhilosophyAggressivegoalMetric(standardofmeasurement)BenchmarkMethodVehiclefor:CustomerfocusBreakthroughimprovementContinuousimprovementPeopleInvolvementWhatisSixSigma?Bottomline:SixSigmaDefinesthegoalsofthebusinessDefinesperformancemetricsthattietothebusinessgoalsIdentifies,usingperformancemetrics,projectsthatwillyieldclearbusinessresultsAppliesadvancedqualityandstatisticaltoolstoachievebreakthroughfinancialperformanceSigmaisastatisticalunitofmeasurethatreflectsprocesscapability.Thesigmascaleofmeasureisperfectlycorrelatedtosuchcharacteristicsasdefects-per-unit,parts-permilliondefective,andtheprobabilityofafailure/error(DistributionShifted±1.5s)2308,537366,80746,210523363.4sPPMProcessCapabilityDefectsperMillionOpp.Copyright1995SixSigmaAcademy,Inc.SixSigmacorrespondstopartsperbillionifprocessiscenteredSixSigma-Goal3Sigma6Sigma5Sigma4Sigma93.32%99.379%99.9767%99.99966%HistoricalCurrentIntermediateLong-termSigmaLong-TermYieldStandardSixSigma-PerformanceTarget99.99966%Good(6Sigma)20,000lostarticlesofmailperhourUnsafedrinkingwaterforalmost15minuteseachday5,000incorrectsurgicaloperationsperweekTwoshortorlonglandingsatmostmajorairportseachday200,000wrongdrugprescriptionseachyearNoelectricityforalmostsevenhourseachmonthSevenarticleslostperhourOneunsafeminuteeverysevenmonths1.7incorrectoperationsperweekOneshortorlonglandingeveryfiveyears68wrongprescriptionsperyearOnehourwithoutelectricityevery34years99%Good(3.8Sigma)SixSigma--PracticalMeaningIssueClassicalSixSigmaAnalyticalPerspective PointEstimate VariabilityManagement Cost&Time Quality&TimeManufacturability Trial&Error RobustDesignTolerancing WorstCase RootSumofSquaresVariableSearch OneFactor@Time DOEProcessAdjustment Tweaking SPCChartsProblems Fixing PreventingProblemSolving ExpertBased SystemsBasedAnalysis Experience DataFocus Product ProcessBehavior Reactive ProactiveSuppliers Cost RelativeCapabilityOutlook ShortTerm LongTerm?1994Dr.MikelJ.Harry-V4.0SixSigma--AsAValueIssueClassicalSixSigmaOutlook ShortTerm LongTermDecisionMaking Intuition ProbabilityDesign Performance ProducibilityAim Company CustomerOrganization Authority LearningTraining Luxury NecessityChainofCommand Hierarchy EmpoweredTeamsDirection SeatofPants BenchMark/MetricsGoalSetting RealisticPerception Reach-OutStretchPeople Cost CompetitiveAdvantageControl Centralized LocalizedImprovement Automation Optimization?1994SixSigmaAcademySixSigma--AsAValue....ToProduceImprovedRolledThroughputYield,DefectsPerUnit(DPU)DefectsPerMillionOpportunities(DPMO)ReducedCostofPoorQuality(COPQ)ImprovedCapacityandProductivityReducedVariationInOurProcesses/Products$$’sAProblemSolvingMethodologyCharacterizeOptimizeBreakthroughUSLTLSLUSLTLSLTUSLLSLUSL’LSL’TheStrategyCustomerFocused-BothInternally&ExternallyWhichBusinessFunctionNeedsIt?Aslongasthereisaprocessthatproducesanoutput,whetheritisamanufacturedproduct,data,aninvoice,etc…,wecanapplytheSixSigmaBreakthroughStrategy.Fortheseprocessestoperformtoacustomerstandardtheyrequirecorrectinputs!!!6SigmaMethodsMFG.DESIGNSERVICEPURCH.MAINT.ADMIN.QAMarketingIfwearesogoodatX,whydoweconstantlytestandinspectY?YDependentOutputEffectSymptomMonitorX1...XNIndependentInput-ProcessCauseProblemControlTogetresults,shouldwefocusourbehaviorontheYorX?f(X)Y=FocusonXratherthanY,asdonehistorically?1994SixSigmaAcademyTheFocusofSixSigmaIsYourLeadershipOnBoard?“Everyonehasthewilltowin,fewhavethewilltoworktowin.”--BobbyKnightSixSigmaonlyworkswhenLeadershipispassionateaboutexcellenceandwillingtochange.

FundamentalsofLeadershipChallengetheprocessInspireasharedvisionEnableotherstoactModelthewayEncouragetheheartSixSigmaisacatalystforleadersBreakthroughStrategyCharacterizationPhase1:MeasurementPhase2:AnalysisOptimizationPhase3:ImprovementPhase4:Control?1994SixSigmaAcademyTheBreakthroughPhasesTheBreakthroughStrategyABCDEFG1SelectOutputCharacteristic2DefinePerformanceStandards3ValidateMeasurementSystem4EstablishProductCapability5DefinePerformanceObjectives6IdentifyVariationSources7ScreenPotentialCauses8DiscoverVariableRelationships9EstablishOperatingTolerances10ValidateMeasurementSystem11DetermineProcessCapability12ImplementProcessControlsApplicationProjectsBreakthroughCookbook(ProductBenchmarking)(ProcessBaselineAnalysis)ManagementLeadersManagement&TechnicalLeaders

Yi=fX1,...,XNGoal:MeasureAnalyzeImproveControl?1994SixSigmaAcademyProductCapabilityAnalysisProcessCapabilityAnalysisFocusontheY’sFocusontheX’sDefineRecognizeTimePerformanceGOODBAD3Sigma6SigmaThisDrivesBreakthroughImprovementSixSigmaBreakthroughDataisderivedfromobjects,situations,orphenomenonintheformofmeasurements.Dataisusedtoclassify,describe,improve,orcontrolobjects,situations,orphenomenon.1.Weonlyuseexperience,notdata.2.Wecollectdata,butjustlookatthenumbers.3.Wegroupthedatasoastoformchartsandgraphs.4.Weusecensusdatawithdescriptivestatistics.5.Weusesledatawithdescriptivestatistics.6.Weusesledatawithinferentialstatistics.LevelsofAnalysis:Whatlevelarewe??TheFoundationoftheSixSigmaToolsWhySixSigma?ExecutivePerspective“2,400processimprovementsweremade,reducingdefectsby61percent-yielding$350millioninannualizedcostsavings.” --LarryBossidy,ChairmanandChiefExecutiveOfficer,AlliedSignal“SixSigma—GEQuality2000—willbethebiggest,themostpersonallyrewardingand,intheend,themostprofitableundertakinginourhistory.” --JackWelch,ChairmanandChiefExecutiveOfficer,GE“Theresimplyaren’tmanyskepticsofJackWelchorLarryBossidy.WhilebothcompanieshavementionedSixSigma,webelieveinvestorshavenotfullyappreciatedthepowerorthepotentialofthisprocesstogeneratelow-riskearningsgrowthandsharegains.” --JenniferPokrzywinski,FirstCallReport,MorganStanleyAnalystWhySixSigma?

“Quality...isthenextopportunityforourCompanytosetitselfapartfromitscompetitors... ...Dramaticallyimprovedqualitywillincreaseemployeeandcustomersatisfaction,willimprovesharepriceandprofitability,andwillenhanceourreputation. ...Thisqualitydrivewillrequirethepassionatecommitmentofallofyoutomakeithappen...ASixSigma2000targetwillrequireasignificantcommitmentofresourcesandyourstrong,highlypersonalcommitmenttotheobjective.”

J.F.Welch,10/15/95JackWelch’s“Vision2000”EachBlackBeltProjectaverages$250,000savingstothebottomline.CostofPoorQuality40%35%30%25%20%15%10%5%3.4233621066,807308,537691,462CostofFailure(%Sales)DPMO

654321SigmaAonesigmashiftwillresultin20%marginimprovement,12-18%capacityincrease,12%reductioninorredeploymentofemployees,and10-30%capitalreduction.WhatisCostofPoorQuality?Inadditiontothedirectcostsassociatedwithfindingandfixingdefects,“CostofPoorQuality”alsoincludes:

ThehiddencostoffailingtomeetcustomerexpectationsthefirsttimeThehiddenopportunityforincreasedefficiencyThehiddenpotentialforhigherprofitsThehiddenlossinmarketshareThehiddenincreaseinproductioncycletimeThehiddenlaborassociatedwithorderingreplacementmaterialThehiddencostsassociatedwithdisposingofdefectsFormostcompaniestoday,thecostofpoorqualityislikelytobe25%ofsales.InalmosteverycompanywheretheCOPQisunknown,theCOPQexceedstheprofitmargin.Howfarcan

inspection

getusto6Sigma?(DistributionShifted±1.5s)s23456PPM308,53766,8076,2102333.4GettingThereThroughInspectionUnlessyouhaveabsolutely100%perfectinspection,themoredefectsyoucreate,themoredefectsescapetothecustomer.TotalDefects/UnitEscapingDefects/UnitTheoryofDefectsTask:

Countthenumberoftimesthe6thletterofthealphabetappearsinthefollowingtext.Thenecessityoftrainingfarmhandsforfirstclassfarmsinthefatherlyhandlingoffarmlivestockisforemostinthemindsoffarmowners.Sincetheforefathersofthefarmownerstrainedthefarmhandsforfirstclassfarmsinthefatherlyhandlingoffarmlivestock,thefarmownersfeeltheyshouldcarryonwiththefamilytraditionoftrainingfarmhandsoffirstclassfarmsinthefatherlyhandlingoffarmlivestockbecausetheybelieveitisthebasisofgoodfundamentalfarmmanagement.GettingThereThroughInspectionMotorola TexasInstruments ABB(AseaBrownBoveriLTD.)AlliedSignal GEBombardierNokiaToshibaDuPontAmericanExpressBBAFordDowChemicalJohnsonControlsNorandaCompaniesImplementingSixSigmaSixSigmaOverviewOverallPerspectiveManufacturingProcessPictureDefectsandVariationVariationandProcessCapabilityProcessCapabilityandImprovementTheDefectEliminationSystemCustomerRequirementsProcessCapabilitiesTechnology(Design)CapabilitiesMaterialsCapabilitiesPeopleCapabilitiesGreatDesignsPredictableFlexibleFactoriesCustomerDelightSafetyRolledThroughputYieldProductCostCapacity/CycleTimeMaterialsManagement&InventoryOverallPerspectiveSixSigmaisaboutidentifying/quantifyingandeliminatingtheHiddenFactorythroughDefectReductionandultimatelyDesignfor6Sigma90%YieldAfter

InspectionorTestEachdefectmustbedetected,repairedandplacedbackintheprocess.ThisresultsinWasted:TimeMoneyResourcesFloorSpaceScrapReworkHiddenFactoryNOT

OKOperationInputsInspectFinalTestYieldOK90%

CustomerQualityDefectsandtheHiddenFactoryThisclassicapproachtocalculatingyielddoesnotaccountforthehiddenfactory.DescribedbelowisanyprocessthatmeetsacustomerneedRolledThroughputYieldVersus

FirstTimeYield66%190%...whynot?Scrap90%

CustomerQualityReworkHiddenFactoryNOT

OKYieldAfter

InspectionorTestOperationInputsInspectFinalTest

Yield=OKProcessABC90%

Yield90%

Yield81%73%FinalTestD

Using“finaltestyield”ignoresthehiddenfactory.Finaltestperformanceisafunctionofinspection&testnotactualdefectdata.Rolled-ThroughputYieldFinalTestYield90%

Yield66%90%

YieldRolledYieldWhatCausesDefects?ExcessVariationdueto:ManufacturingprocessesSupplier(incoming)materialvariationUnreasonablytightspecifications(tighterthanthecustomerrequires)InadequateDesignMarginUnstablePartsandMaterialsInsufficientProcessCapabilityDissectingProcessCapabilityDefectsProcessCapabilityInadequateDesignMarginInadequateProcessCapabilityUnstableParts&MaterialsDefectsAcceptablePremiseof6Sourcesofvariationcanbe:IdentifiedQuantifiedEliminatedorControlledLSLUSLHowDoWeImproveCapabilityOurOutputs(Y’s)aredeterminedbyourInputs(X’s).IfweknowenoughaboutourX’swecanaccuratelypredictYwithouthavingtomeasureit.Ifwedon’tknowmuchaboutourX’s,thenwehavetoresorttoinspectionandtest(nonvalueaddedoperations).ByknowingandcontrollingtheX’s,wereducethevariabilityinY,whichdecreasesthenumberofdefects,improvesRTY,cycletime,etc.Wecanalsoeliminateorreduceinspection,test,andrework.SixSigma,Metricsand

ContinuousImprovementSixSigmaischaracterizedby:definingcriticalbusinessmetricstrackingthemimprovingthemusingproactiveprocessimprovementSixSigma’sprimarymetricisDefectsperUnit(DPU),whichisdirectlyrelatedtoRolledThroughputYield(Yrt) Yrt=e-dpuCostofPoorQualityandCycleTime(Throughput)aretwoothersSUPPLIERProcess

Step1Process

step2CUSTOMERProcess

step3RECORD

DEFECTSRECORD

DEFECTSRECORD

DEFECTSRECORD

DEFECTSRECORD

DEFECTSMEASURECONTROLIMPROVEANALYZECONTINUOUSIMPROVEMENTMODELPROCEDURETIMEGOALTRENDCHARTACTIONPLANACTIONNAMEDATEVitalFewIdentificationContinuousImprovementThefollowingvisualillustrateshowthesystemcanworkY1Y2Y3YrtMetricsDefectsperunit(DPU)drivesplant-wideimprovementDefectspermillionopportunities(DPMO)allowsforcomparisonofdissimilarproductsSigmalevelallowsforbenchmarkingwithinandacrosscompaniesCalculatingtheProductSigma-LevelCalculateDefectsperUnit(DPU):DPU=Total#ofDefectsTotal#ofUnitsProducedCalculateDefectsperMillionOpportunities(DPMO):GotoaSigmaChartandEstimatetheSigmaLevelDPMO=DPOx1,000,000TotalOpportunitiesforErrorinOneUnitCalculateDefectsperOpportunity(DPO):DPO=DPU

(#ofOpportunitiesforErrorinOneUnit)CalculatingYieldCalculateYieldYield=e-DPU(UsingPoissonApproximation)OrYield=(1-DPO)#ofOpportunitiesRolledThroughputYield=Y1*Y2*Y3*Y…Thefollowingexletracksactualproductiondataforonemonthinatransformerfactory.Listedaretheprocesssteps.Noticeengineeringandpurchasingarelistedaswell.ExlefromDiscreteManufacturingProcessSteps,Defects,UnitsandDPU(1)Whichthreeprocessareastoattackfirst?ProcessSteps,Defects,UnitsandDPU(2)TanksPaintMaterialsDefectsperunitdrivesplantwideimprovementBenchmarkingprocessstepsandproductProcessSteps,Complexity,

andSigmaLevelRolledThroughoutYield=18.4AlthoughtheFirstTimeYieldattheFinalTestis99.8%,only18.4%ofthetransformersmakeitthroughthefactorywithoutrework.ProcessSteps,FTYandRTYDynamicsoftheSystemEnsureyouReviewtheCostsandCustomerImpactofeachprocessDrivesplantwideimprovementfocusDrivesSBUwidebenchmarkingDefectdatadrive’sprojectteamimprovementRolledThroughoutYield=18.450%ReductioninDefectsRolledThroughputYieldimprovedfrom18.4%to43%(134%improvement).Sigmawentfrom3.99to4.23.TrackingTrendsinMetricsOurobjectiveistotrackthetrendsintheMetricstoestablish,basedonfact,ourimprovements.ThesemetricscanbeDefectsPerUnit,PartsPerMillionDefective,Time,Defects,RolledThroughputYield,etc.---------------------------------------SweetFruit

DesignforManufacturability,DesignforSixSigmaBulkofFruitProcessCharacterizationandOptimizationLowHangingFruitSevenBasicToolsGroundFruitLogicandIntuiti