IBM’sQualityEarlyWarningSystem

EarlierandMoreDefinitiveProblemDetection

IntegratedSupplyChainEngineering22ISCEngineeringHistoryofsixsigma(1)Sixsigmaismostlyfoundedonmethodsthathavebeenaroundfordecades:Pre-industrialrevolutionSkilledcraftsmencontrolledthequalityandthedesignoftheirproductsfrombeginningtoend1798EliWhitneyMassproductionandinterchangeableparts1850'sIntroductionofgages1880'sFrederickTaylorIntroductionofassemblylinesandpartitioningofwork1920'sW.A.ShewhartIntroductionofStatisticalProcessControlAssignablecausevs.changecauseUseofstatisticforimprovement1930'sDodge&RomigConceptofacceptabilitybasedonsamplingresult(AQL)1950'sDemingPromotionofPlan-Do-Study-ActcycleTopmanagementinvolvementConcentrationonsystemimprovement1950'sJuranQualityplanningisusedtocreatetheprocessthatwillenableone

tomeetthedesiredgoals33ISCEngineeringHistoryofSixSigma(2)Sixsigmaismostlyfoundedonmethodsthathavebeenaroundfordecades:1960'sKaoruIshikawaQualitycomesfirstCustomercomesfirstDecisionsarebasedonfactsanddataEngagementofManagementCross-functionalcommittees1980'sPhilipCrosby14Stepapproachtoachievecompanywidequalityimprovement1987ISOSeriesofqualitystandardsthatdetailedthekeyelements

ofsoundqualitypractices1987MalcolmBaldrigeNationalQualityAwardPromotionofbestpracticesharingandtheestablishmentofabenchmarkforqualitysystems1987MotorolaStructuredmethodologyFocusoncustomerneeds1995GeneralElectricCompanywideimplementation,demonstratedleadership44ISCEngineeringWhatisSixSigma?SixSigmaemergedin1987whenMotorolapublishedtheirSixSigmaqualityprogram

SixSigmaisametricthatdemonstratesqualitylevelsat99.9997%performance

forproducts,processesandservices

SixSigmaisavisionandanapproachtoachievingthehighestcustomer

satisfactionthroughofferingproducts,processesandservicesatthehighest

qualityandlowestcosts

SixSigmaisanintegratedapproachtoprocessexcellence

SixSigmaisabusinessconceptinresponsetocustomers’demandforhighquality

SixSigmademandscompetenceinstatisticstoensuredecisionsbasedonfacts-7-6-5-4-3-2-10+1+2+3+4+5+6+755ISCEngineeringLeanSigmaisbasedonthepracticallearningoforganizationsimprovingtheirprocessesforover50yearsHistoryofLeanSigma66ISCEngineeringWhatisLeanSigmaLeanSigmaisanevolutionaryimprovementtoSixSigmafortransactionalprocessescommone.g.inservicesbutalsomanufacturingItdevelopsadeepprocessunderstandingintermsofvalueflowdependenciesIttriestoimproveefficiencyinprocessesItsrootsareindependentfromSixSigma;acommonreferenceforleanmanufacturingistheToyotaProductionSystem(TPS)LeanMethodologyIncreaseefficiencySimplifyworkflowsFocusonhigh-valuestepsEliminatewasteALeanenterpriseisonethatdeliversvaluetoitsstakeholderswithlittleornowastefulconsumptionofresources.ProductorServiceOutputsSixSigmaTMMethodologiesIncreaseconsistencyReducevariationEliminatedefectsInaSixSigmaenterprise,everyoneisfocusedonidentifyingandeliminatingdefects.Customer-drivenCustomer-driven77ISCEngineeringWhyisitimportanttocompanies?QualityCostofFailureCostofavoidingfailuresHiddenCostsWarrantyCostsTrainingReducedproductivityduetolowerutilizationofexistingresourcesCustomerComplaintVisitsCapabilityStudiesFieldserviceAdvancedQualityPlanningReturns&RecallsSalesthatwouldhavebeenoccurredduetomeetingcustomerneedsLiabilitySuitsVendorSurveysInspection&TestCosttoCustomerMaintenanceAuditsScrap&ReworkChangestoDesignChangestoProcessQualitycostisanycostthatacompanywouldnothaveincurrediftheirproductorprocesswereperfect.8MostCompaniesuseSPCinQualitymonitoringThenatureofSPC/RateBasedmanagementisreactiveBasedonpastperformanceandstatisticalrelevanceUnabletopredictwhat“MAY”occurinthefutureDoesnotrankwarningstofocusonpotentialemergingissuesNeedtopredictdefecttrendsbeforecumulativeevidenceisavailable9BladesIBMIntegratedSupplyChainEngineeringScopeRawParts/

DiscreteComponentsSub-SystemsIntegratedSystems/SolutionsCables/ConnectorsLogic/Active/Optic/PassiveMemory/DIMMs/SRAM/DRAMDrivesCardAssembliesPowerSubsystemsThermalSubsystemsMechanicalsNodeAssembliesPrintedCircuitBoards/FlexCableAssemblyMainframesServersandHPCStorage1010ISCEngineeringQEWSisanenterprise-levelsystemwhichusesuniqueIBMtechnologytodetectandprioritizequalityproblemsandparametricshiftsearlierandmoredefinitivelythancanbedoneusingtraditionaltechniquesofstatisticalprocesscontrol.QEWSanswersthequestion:WhatistheIBMQualityEarlyWarningSystem?Hasanythingchangedenoughtorequireaction?1111ISCEngineeringQualityproblemsareidentifiedmoreeffectively:earliermoredefinitivelyvisiblyEngineeringproductivityishigher:muchlesstimeisspentdeterminingwheretheproblemsaremoretimeisspentworkingproactivelyonissueswhichcouldeproblemsallengineersareempoweredwithexpert-levelanalyticaljudgmentBrandimageandbrandvaluehavebeenwellprotected,despitecontinuingcostreductions,manpowerreductions,andincreasedrelianceonsuppliersIBM’sexperiencewithQEWS1212ISCEngineeringThevalueofQEWSLowerCostsinManufacturing

reducedreworkreducedscrapinDistribution

fewerrecallsinWarranty

fewerclaimsImprovedProductivityinManufacturing/PurchasedProduct

reducedreworkreducedscrapincreasedcapacityutilizationmoreon-timeshipmentshigherassuranceofdeliveryofqualityproductsinEngineering

moreproductcoverage,andmoreeffectivecoverage

withexistingengineeringresourcesprioritizationofthemostpressingissuessingleeffectiveprocessfortheentireenterpriseImprovedBrandValue

improvedbrandimageprotecthigh-stakes,highvolumeproductlauncheshighertop-linegrowthandcustomerretention1313ISCEngineeringAtpointswhereatest,measurement,orinspection

ismade:Inthesupplychain:suppliers’finaltestinginspectionofrawmaterialsinginspectionofprocuredcomponentsInmanufacturing:atindividualproductionoperationsatfinalproducttestInproductfieldperformancewarrantyclaimsWherecanQEWSbeapplied?1414ISCEngineeringQEWSAnalysisCapabilitiesModuleExamplesTypicalApplicationsAttributedatafailureratesyieldssortcategoriesMonitoringqualityofcomponentsprocuredfromsuppliersMonitoringqualityofmanufacturingoperationsviainlinemeasurementsMonitoringqualityofmanufacturingoperationsviafinalproducttestReliabilitydatawarrantyclaimsstresstestsMonitoringqualityofproductviaperiodicreliabilitymonitoringtestsMonitoringqualityofproductincustomeruseenvironmentsDetectingproductwearout1515ISCEngineeringSPCcharts:Whichonesrequireattention?1616ISCEngineeringQEWSmakeschangesvisible1717ISCEngineeringQEWSDemonstrationResults:EarlierDetectionThischartshowsQEWSanalysisresultsforthesamesetofdataasabove.Thex-axisisalignedintimetothechartabove.QEWSalertswhenthecumulativeevidencecrossesabovethehorizontalthresholdline(inblack.)Inthiscase,QEWSalerted8weeksearlierthanSPC.ThischartshowsSPCanalysisresultsforasetofyielddata.SPCalertswhenapointfallsoutsidethecontrollimits(attheextremeright-handsideofthechart.)1stQEWSalert1stSPCalert8weeks1818ISCEngineeringQEWSDemonstrationResults:DefinitiveDetection1stQEWSalertcontinuingalertsmountingevidence1stSPCalertnoSPCalerts2ndSPCalertForthesamedataasabove,QEWSalerts,thenstaysinalertmode(abovethehorizontalblackthresholdline.)Thepositiveslopeofthecumulativeevidencelineindicatesthequalityproblemisgettingworse.Fromthefirstalertonward,QEWSpresentsaclearmessagethatactionshouldbetaken.Forthissetofdata,SPCalertsonce,thendoesnotalertagainuntilmanypointslater.Manyengineerswoulddismissthefirstalertasananomaly,andnottakeactionuntilthesecondalert.1919ISCEngineeringABStrategicdatasources:

linksanddataformatsTargetoptimizationalgorithmInnovative,verylarge-scale

datacubedeploymentDatastoreanddashboardUserinterface,dashboardandanalytics(supplierandOEM)PrioritizationalgorithmCPatentpendingSupplier,Manufacturing,FieldQEWSEngineThresholdsettingalgorithmQEWScombinesadvancedanalytics,visualizationandworkflowtocreateasystemthatiseffective,easytouse,andeasytodeploy

2020ISCEngineeringQEWSDashboard:VisualQualityManagement2121ISCEngineeringQEWSDashboard:drill-downengineeringtoolsViewSPCdataViewAVTdataViewQEWSdataViewHistoryRawdatadrill-down2222ISCEngineeringQEWScasestudy:fieldreliabilitymonitoringIndustryAutomotiveClientGlobalautomobilemanufacturerQEWSmoduleReliabilitydataanalysisClient’sObjectivesImproveproductreliabilityImprovebrandimageReducewarrantyclaimscostsClient’sChallenges/IssuesThenumberofvehiclesrepairedunderwarrantyissmallcomparedtotheentirefieldpopulation,makingearlydetectionoftrendsdifficultWearoutmechanisms,ifnotdetectedearly,willcauseepidemicratesofproductfailureAnalyticmethodsforwarrantydataarenotautomatedIBM’sApproachUsedclientexperiencetotargetareasforanalyticimprovementEstablishedmethodologyforQEWSmonitoringofvehiclewarrantyclaimsdataTailoredQEWSanalyticstoclient-specificrequirementsResults