六西格瑪管理簡(jiǎn)介(英文版)課件

GlossaryGlossary4-Block……………aerror……………arisk………………

Accuracy…………

Active(opportunityordefect)…

AdvocacyTeam…AlternateHypothesis…………

ANOVA……………

ANOVAmethod(GaugeR&R)…

Assignablecausevariation……

AttributeChart…

Attributedata……

Average………Graphicaltooltoshowtherelationshipbetweenprocesscapability,control&technology.

Theerrormadeifdifferenceisclaimed,whentherealityissameness(e.g.rejectinggoodparts;Producer’sRisk).

Therisk(probability)ofmakinganaerror(frequentlysetat5%).

Howclosemeasurementsare,onaverage,totheirtarget.

Anopportunityordefectthatisbeingmeasured(adefectwearelookingfor).

ThegroupofpeoplewhohaveastakeintheSixSigmaproject,includingthosewhomustkeepitincontrol.

SeeHa

AnalysisofVariance.Astatisticalmethodofquantifyingcontributionsofdiscretelevelsof“X”stothevariationina“Y”response.AMinitabselectionforGaugeR&Rthatincludesoperator-partinteractioninthecalculationofvariationcontributions.ThemostaccuratemethodforGaugeR&R.Removablevariationinaprocess;variationduetooutsideinfluences.See‘BlackNoise’.StatisticalProcessControl(SPC)chartfordiscretedata.Includesp,np,canducharts.

Datathatcanbedescribedbylevels,integervaluesorcategoriesonly.SeeDiscretedata.Thesumofalldatainasampledividedbythenumberofdatapointsinthesample.SeeMean.4-Block……………Graphicaltberror……………brisk……………

Baselining………

Benchmarking……

BlackBelt………BlackNoise………

Boxplot…………

Brainstorming……Centring…………CentringofXvariables…………CentralLimitTheorem…………Theerrormadeifsamenessisclaimed,whentherealityisdifference(e.g.acceptingbadparts-Consumer’sRisk).

Therisk(probability)ofmakingabetaerror(frequentlysetat10%).Evaluatingthecapabilityofaprocessasitstandstoday,without“tweaking”-i.e.passiveobservation.

Evaluatingthecapabilityofsimilarprocessestoquantifywhatconstitutes‘theBest’.ApersonwhosefulltimejobconsistsofapplicationofSixSigmatools/methodsonprojects.

Processvariationdueto‘outsideinfluences’.SeeAssignableCauseVariation.Graphshowingtheportionofadistributionbetweenthefirstandthirdpercentileswithina‘box’.Theboxplotalsoshowsthemedianofthedistributionandtheextremevalues.Oftenusedtocomparepopulation.

AtechniqueusedbyanAdvocacyTeamto,fore.g.,developalistofpotentialX’satthebeginningofproject.

Aprocesscharacteristicdescribinghowwellthemeanofthesamplecorrespondstothetargetvalue.

AmethodusedtotransformXvariablesinDoE’sthatdevelophigherorder(quadratic)models;reducescorrelationbetweenX’s.

Afundamentalstatisticaltheoremstatingthatthedistributionofaveragesofacharacteristictendstobenormal,evenwhentheparentpopulationishighlynon-normal.

berror……………TheerrorCentralCompositeDesign……

Champion…………

ChampionReview………………

Chi-Squaredtest…

ClassicalYield……CommonCauseVariation………

ComponentsSearch……………

Confidence………

ConfidenceInterval………………

Consumer…………

ContinuousData…ADesignofExperiments(DoE)methodwhereeachXistestedat5levels(see‘StarPoints’).

ACCDprovidesthecapabilitytomodela

processwithaquadraticequationORalinear

equation.

Typicallyadirector-someonewhocansupporttheSixSigmaprojectandhastheauthoritytoremovebarriersandprovideresources.TakesanactivepartinProjectReview.

AregularmeetingtopresentSixSigmaprojects,shareexperiencesandremoveroadblocks.

Hypothesistestfordiscretedata.Evaluatestheprobabilitythatcountsindifferentcellsaredependentononeanother,ortestsGoodnessofFittosomeaprioriprobabilitydistribution.

See“FirstPassYield”.GoodunitsproduceddividedbyTotalUnitsProduced.

See“WhiteNoise”.Theinherentvariationofaprocess,freefromexternalinfluences.Usuallymeasuredoverashorttimeperiod.AmethodofscreeningforVitalFewX’sinmanufacturedassemblies.Alsoknownas‘PartSwapping’.Thecomplementofalpharisk.Confidence=1-a.

Arangeofplausiblevaluesforapopulationparameter,suchasmeanorstandarddeviation.

Theenduserofaproduct(thehomeowner,fore.g.).Theconsumerisexternaltothebusiness.

Datathatcanbemeaningfullybrokendownintosmallerandsmallerincrements-e.g.length,temperatureetc.)CentralCompositeDesign……ADContourPlot…ControlLimits………………CostofQuality………………Cp……………Cpk……………

CQ……………

CTQ…………CubePlot……

Customer……

DataWindow…Defect…………DependentVariable…………AgraphusedtoanalyzeexperimentsofaCentralCompositeDesign.TwoX’scomprisetheaxes,andlevelsofconstantYareshowninthebodyofgraph.Resemblesatopographicalmap.

LinesonaStatisticalProcessControl(SPC)chartthatrepresentdecisioncriteriafortakingactionontheprocess.Linesaredrawn+/-3standarddeviations(s)fromthemean.

Afinancialreconciliationofallthecostsassociatedwithdefects(scrap,rework,concessionsetc.)

StatisticusedtomeasureProcessCapability.Assumesdataiscentredontarget.SimilarinconcepttoZ.st

StatisticusedtomeasureProcessPerformance.Doesnotassumecentreddata.SimilarinconcepttoZ.lt

CommercialQuality.Usedtocategorizenon-manufacturingprojectsthatimpacttheconsumerand/orcustomer.

Critical-to-Qualitycharacteristic.Anaspectoftheproductorservicethatisimportanttothecustomer/consumer.

Agraphusedforanalysisoftheresultsofafactorialdesignedexperiment(DoE).Showstestconditionsthatoptimizetheresponse.

Therecipientoftheoutputofaprocess.Maybeinternal(e.g.Assemblyisacustomeroffinishingshops),orexternal(e.g.Currys,Bellingetc.)whothensellourproductstoconsumers.

ThespreadsheetwindowinMinitabwheredataisenteredforanalysis.

Anyaspectofapartorprocessthatdoesnotconformtorequirements.

Theoutputofaprocess.The“Y”response.ContourPlot…AgraphusDescriptiveStatistics………

DesignofExperiments(DoE)DiscreteData………………Dotplot………DPMO…………

DPO…………

DPU…………e(ExponentialFunction)……Entitlement…

ExecutiveSummary………F-test…………Mean,StandardDeviation,Varianceandothervaluescalculatedfromsamplecharacteristics.Alsoincludesassortedgraphs.

Astatisticalfieldofstudywhereindependentvariables(X’s)aresystematicallymanipulatedandtheresponseobserved.UsedtodemonstratewhichX’saretheVitalFew,andtooptimizetheresponse.

Datathatcanonlybedescribedbylevels,i.e.pass/fail,operatora/b/c,integervalues(e.g.numberofdefects).Datathatcannotbebrokendownintofinerincrements.

Frequencydiagramrepresentingdataby‘dots’alongahorizontalaxis.Generallyusedasanalternativetoahistogramforsmallsamplesizes.

DefectsPerMillionOpportunities-1,000,000multipliedbytotalnumberofdefects,dividedbythetotalnumberofopportunities.Ametricfordefectsequivalenttoppmusedfordefectives.

DefectsPerOpportunity-totalnumberofdefectsdividedbytotalnumberofopportunities.UsedtoentertheNormalTabletoobtainZvalues.

Defectsperunit-totalnumberofdefectsdividedbytotalnumberofunits.UsedprimarilytocalculateRolledThroughputYield(Y.rt)throughthePoissonformulaY.rt=e-DPU.Amathematicconstantroughlyequalto2.718

Mathematicalidentity:ln(e)=1Z.stThebesttheprocesscanbe.WhattheprocesswouldlooklikeifallAssignableCauseVariationwascontrolled.ThefirstpageofoutputfromtheMinitabProcessCapabilityselection.Atesttocomparevariancesof2ormoresamples,andtocomparetheequalityoftwoormoremeans(inANOVA).DescriptiveStatistics………MeanFactorialExperiment………FractionalFactorialExperiment.

FirstPassYield………………

FMEA……………

FunctionalOwner……………GaugeXBRmethod…………

GanttChart……

GaugeR&R……

GreenBelt………

Ha………………

Ho………………Adesignedexperiment(DoE)whichinvolvestestingofallpossiblecombinationsofindependent(X)variables.

Adesignedexperiment(DoE)whichinvolvestestingafractionofallpossiblecombinationsofindependent(X)variablesinafullFactorialexperiment.Resultsinfewertestruns.See‘ClassicalYield’.Equaltothenumberofgoodunitsproduceddividedbythetotalnumberofunitsproduced.FailureModeandEffectsAnalysis-ateam-basedprocedurethatidentifiesanddocumentsallpossiblefailuremodes,effects,causesandassociatedcorrectiveactions.

Thepersonwithfinancialresponsibilityfortheprocessunderconsideration.

GaugeR&Rmethod-anoptioninMinitab.Aprojectmanagementtoolthatgraphsmilestonesvs.thecalendar.Barsareusedtoindicatebothplannedandactualdurationoftasks.

Ameansofdeterminingtheacceptabilityofthevariabilityinthegaugingsystemforuseintheprocess.

ApersonwhousesSixSigmatoolsandmethodologyinthecourseoftheirwork,andwhoalwayshasaSixSigmaprojectactiveintheirplaceofwork.

AlternateHypothesis(hypothesisofdifference).Thehypothesisbeingproveninastatisticalhypothesistest.

Nullhypothesis(hypothesisofsameness).Thestartingassumptioninastatisticalhypothesistest.NB.Thenullhypothesiscannotbeproved!FactorialExperiment………AdesiHistogram……

HomogeneityofVariance……Hypothesistest………………

I/MRChart………

IndependentVariable…………Inferentialstatistics……………

InherentProcessCapability…

Interactionplot………………Afrequencydiagramcomposedofrectangularbarswhoserelativeheightsindicatethenumberofcounts(orrelativefrequency)ataparticularlevel.AmenuselectioninMinitabunderwhichtheF-test(comparisonofvariances)isperformedAnyofseveralstatisticaltestsof2ormoresamplesfrompopulations.Usedtodetermineiftheobserveddifferencescanbeattributabletochancealone.Theresultofthetestistoeitheracceptorrejectthealternatehypothesis(Ha).(t-test,F-testandChi-Squaredtestareexamples.)

Individual/MovingRangechart-aStatisticalProcessControl(SPC)chartinwhichtheuppergraphisusedtoplotindividualdatapointscomparedtocalculatedcontrollimits;thelowergraph(MovingRange)plotsthedifferencebetweensequentialdataaspointsonthechart.Controllimitsarealsocalculatedforthischart.Variables(X’s)thatinfluencetheresponseofadependentvariable(Y)Statisticalanalysesthatquantifytheriskofstatementsaboutpopulations,basedonsampledata.Inferentialstatisticsareusuallyhypothesistestsorconfidenceintervals.

TheBesttheprocesscanbe,withonlyvariationduetowhitenoisepresent.SeeEntitlement,Z.stAgraphusedtoanalysefactorialandfractionalfactorialdesignsofexperiments.IndicatestheeffectonYwhentwoX’sarechangedsimultaneously.ThegreaterthedifferenceinslopesbetweentheX’s,thegreatertheinteraction.Histogram……AfrequencyKurtosis…………L1Spreadsheet………………L2Spreadsheet……………

LCL(LowerControlLimit)…

LeverageVariable……………Linearity(gauge)………………

Longtermdata…LSL………………m…………………Macro……………

MainEffectsPlot………………MasterBlackBelt……………Comparisonoftheheightofthepeakofadistributiontothespreadofthe‘tails’.Thekurtosisvalueis3foraperfectnormaldistribution.ExcelspreadsheetfordiscretedatathatcalculatessubsystemZvaluesand‘rolls’themintoasystem-levelZvalue.ReplacedbyProductReportinMinitabrelease11.2

ExcelspreadsheetforcontinuousdatathatcalculatesZ.standZ.lt

ReplacedbyProcessReportsinMinitabrelease11.2

ThelowercontrolboundaryonaStatisticalProcessControl(SPC)chart.Alimitcalculatedasthemeanminus3standarddeviations.Note:SEM(StandardErroroftheMean)isusedfors;stdev=s/sqrt(n).

AnXvariablewithastronginfluenceontheYresponse.OneoftheVitalFew.

Thedifferenceintheaccuracyofthegaugefromthelowendtothehighendofthetestrange.

Dataobtainedinsuchawaythatitcontainsassignablecausevariation(‘blacknoise’).

LowerSpecificationLimit

ThemeanoraverageofapopulationAminiprogramwithinasoftwarepackagedesignedtoprovideaparticularoutput(e.g.GaugeR&R)Agraphusedtoanalyzefactorialandfractionalfactorialdesignsofexperiments.ComparestheeffectonYofanXatthe‘high’levelvs.itseffectatthe‘low’level.Slopeofthelineonthegraphindicatessignificance.

Acoach,mentorandtraineroftheSixSigmamethodologiesandtools.Kurtosis…………ComparisonMean……………MeasurementsSystems

Analysis………Median…………Minitab…………

NormalCurve…

NormalProbabilityPlot………

Normalize………NormalizedAverageYield……NullHypothesis………………Orthogonal……

p-value…………ParetoAnalysis………………Theaverage.Maybetheaverageofasample(x-bar),ortheaverageofapopulation(m).See‘GaugeR&R’.Themiddlevalueofasetofdata(the50thpercentile).AstatisticalsoftwarepackagecontainingthemajorityofSixSigmatools.Awidely-used,commonly-seendistributionwheredataissymmetricallydistributedaroundthemean(‘bellcurve’).Agraphicalhypothesistestinwhichsampledataiscomparedtoa‘perfectnormal’distribution.Ho:thesampledataisthesameasthe‘perfectnormal’distribution.Ha:thesampledataisdifferent(i.e.non-normal).Theprocessofconvertingnon-normaldatathroughtheuseofatransformationfunction.Theaverageyieldofaprocesswithmultiplestepsoroperations.Y.na=(Y.rt)1/nSee‘Ho’.Literally,“rightangles”.Afeatureofawell-definedexperimentthatallowsmaineffectstobeseparatedfrom2-wayandhigherorderinteractions,aswellasquadratic(squared)terms.Theprobabilityofmakinganalpha(a)error.Avalueusedextensivelyinhypothesistesting.Alsoreferredtoasthe‘observedlevelofsignificance’.p-valuesarecomparedtothe‘a(chǎn)cceptable’levelofalphariskinordertomakedecisionsinhypothesistests.Aproblemsolvingtoolthatallowscharacteristicstoberankedindescendingorderofimportance.Mean……………Theaverage.ParetoPrinciple………………Passive(opportunity/defect)…PointofInflexion………………PoissonApproximation………Population……PoweroftheTest……………ppm……………

PracticalProblem……………

PracticalSolution……………

Precision………Pre-Control……

PrincipleofReverseLoading..Probabilityofadefectp(d)…The“80-20”rule.Theprinciplethat20%ofthevariablescause80%ofthevariation.Adefectoropportunitythatiscounteduponoccurrence,butthatisnotpartoftheactivemonitoringprocess.Pointonthenormalcurvewhereitchangesfromconvextoconcave.Mathematicallydefinedbysettingthethirdderivativetozero.AmathematicalapproximationforRolledThroughputYield,givenDPU:Y.rt=e-DPU.Alldataofinterestforaparticularprocess,recordedornot.Usuallymodelledwithsamples.Thelikelihoodofdetectingbeneficialchange.Representedas1-b.Theprobabilityofrejectingthenullhypothesis.Partspermilliondefective.AdiscretemeasurementofdefectivesforlongtermdataTheoutputoftheMeasurephase.AcharacterizationoftheZvalue,centringandspreadforY.TheoutputoftheControlPhase.TheoptimisedXlevelsandcontrolplantomaintaintheprocessatitshighestZvalue.Howcloselythedataisclusteredaroundtheirmean.Describesthespreadofthedata.AStatisticalProcessControl(SPC)methodthatallowsanoperatortotakeactiononaprocessbasedonwherethepartmeasurementsfallinanormaldistribution.Partsarecodedred,yelloworgreen.Planningahead–Needtodefinewhatdoyouwanttoknow,sowhattool/testshouldbeused,sowhatdatadoyouneed?The‘tail’areaofthenormalcurve,beyondthespecificationlimit(s).ParetoPrinciple………………The“80ProblemStatement……………ProcessCapability……………ProcessCharacterization……ProcessMap…ProcessOptimisation…………

ProjectHopper………………QFD……………

Quartiles………R-bar/d………

RandomCauseVariation……Range…………RationalSubgrouping………Abriefbutsuccinctdescriptionoftheissueunderinvestigation.Includesthepracticalandbusinessreasonsfortheproject.Astatisticthatnumericallydescribeshowwelltheprocesscouldperformintheabsenceof‘blacknoise’.Examples:Z.st,CpUnderstandingtheY’sandX’sinaprocess.DevelopedthroughthetoolsoftheDefine,MeasureandAnalysephases.Aproblemsolvingtoolthatgraphicallydescribeseachsteporphaseinaprocess.DefiningthebestoperatingpointforX’sinaprocess.DevelopedthroughtoolsoftheImprove/Controlphases.AstackofpotentialSixSigmaprojects,tobepickedupbyBlackBeltsorGreenBeltswhenresourcesallow.QualityFunctionDeployment.ArigorousmethodofdeterminingtechnicalrequirementsandCTQ’sfromthedefinitionofConsumerCues.‘Quarters’ofapopulation.1/4ofthedatafallbelowthefirstquartile,1/4ofthedatafallabovethe3rdquartile.Anestimateofstandarddeviationusingtherangeofthedataandtabledadjustmentfactors.UsedincalculationofcontrollimitsinMinitabGaugeR&RXbargraphicaloutput.See‘WhiteNoise’.Theinherentvariationoftheprocess,freefromexternalinfluences.Thelargestvalueinadatasetminusthesmallestvalueinthedataset.Adatacollectiontechniquethatallowstheseparationofshorttermvariationfromlongtermvariation.ProblemStatement……………AbrieRegression……Repeatability(Gauge)………Repetition………Reproducibility(Gauge)………ResponseSurfaceExperiment

Resolution(Gauge)…………Resolution(Fractional

Factorial)…RolledThroughputYield……Astatisticalmodellingtoolthatallowsdatatoberepresentedbyanequation.UsedforcontinuousYresponses,usuallywithcontinuousXinputs.(ThereisspecialtechniquewithinMinitabcalledLogisticRegressionwhichhandlesspecialformsofdiscreteX’s.)Abilityofagaugetoconsistentlymeasurethesamepartwiththesameresults.PartoftheoutputofaGaugeR&Rstudy.Collectingmultipledatapointssequentiallyfromaprocess,withoutre-settingtheprocessAbilityofoperatorsofagaugetogenerateconsistentmeasurements.PartoftheoutputofaGaugeR&Rstudy.Adesignedexperiment(DoE)thatallowstheYresponsetobemodelledasafunctionofcontinuousXvariables.SeeRegressionalso.Theabilityofagaugetodiscriminateincrementsofacontinuousmeasurement.Gaugeresolutionisusuallyrequiredtobetentimesgreaterthanthemeasurementofinterest;i.e.,afeaturespecifiedwithaspecificationtoonedecimalplacewouldrequireagaugewitharesolutionoftwodecimalplacesetc.Aromannumeralthatindicatesthedegreeofconfoundinginafractionalfactorialdesign.Higherresolutionindicateslessconfounding-i.e.lessambiguityinthesourceofeffects.Y.rtTheproductofyieldsateachstepofaprocess.CanbeestimatedusingthePoissonApproximation.Regression……Astatistics…………………Sample………SessionWindow……………Shift……………Shorttermdata……………Sigma(s)………SixSigmaTeamMember……Skewness……Specification…Spread…………Stability(Gauge)………………StandardDeviation…………Thestandarddeviationofasample.Ameasureofspread(orvariability)ofthedata.s=sqrt[S(x-xbar)/(n-1)]Acollection(subset)ofdataintendedtorepresentthecharacteristicsoftheparentpopulation.Oneofthe4Minitabwindows.Usedforcommandentryanddataoutput.Thedifferencebetweenshort-termandlong-termprocessvariation.Z.shift=Z.st-Z.ltDataobtainedinsuchawaythatitcontainsNOassignablecausevariation(‘blacknoise’).Onlytheinherentprocessvariationisrepresented,whichallowscalculationofZ.stThestandarddeviationofapopulation.AstakeholderintheSixSigmaprocess.Apersonwhoneedstohaveanunderstandingofthemethodology,butdoesnotformallyusethetools.Evaluationofthesymmetryofadistribution.Skewness=0forperfectsymmetry;skewnessisnegativeifthedistributionisshiftedtotheright,positiveifshiftedtotheleft.Therequirementsofadesign,usuallyexpressedasatarget(ornominal)valuewithanassociatedallowabletoleranceforvariation(e.g.5.00cm+/-0.05cm)Howfarthedataisdistributedawayfromtheirmean.Consistencyofmeasurementvaluesobtainedwiththesamegaugeonthesamesetofparts,withmeasurementstakenatdifferenttimes.Gaugeinstabilitycanleadtocalibrationissues.Astatisticalmeasureofspreadordispersionfromameanvalue.s…………………ThestandarddStandardErroroftheMean…StandardNormalDeviate……StandardOrder………………StarPoint(s)……StatisticalProblem……………StatisticalProcessControl…StatisticalSolution……………Statistics………StepwiseRegression…………StructureTree………………Thestandarddeviationofxbar,basedonasamplesizeofn.(Alsoacorrectionfactorforstandarddeviationofrelativelysmallsamplesizes(<30).)Reducesthestandarddeviationofthesamplebysqrt(n).SEM=s/sqrt(n).See“Ztransform”.AfeatureoffactorialDesignofExperiments(DoE)thatdeterminestheorderofthehigh/lowsettingsoftheX’sforeachrunofanexperimentbyusingapre-determinedpatternof+1’sand-1’sforeachX.ExtremetestpointsinaCentralCompositeDesignofExperiments.Foundbytakingthefourthrootofthenumberof‘Cubepoints’(factorialpoints)inthedesignandadding/subtractingthisvaluefromtheCentrePoint.TheoutcomeoftheAnalyzephase.Istheproblemcentring,spreadorboth?SPC.Agraphicalmethodofmonitoringaprocessanddeterminingstatisticallywhentheprocessrequiresattentionbycomparingittoahistoricalmeanandcalculatedcontrollimitsat+/-3sigma.OutputoftheImprovephase.WheredotheX’sneedtobesettocontroltheY?Thestudyofvariation,includingmethodsofdescribing,quantifyingandreducingvariation,aswellasestimatingrisks.Aregressiontechniquewherethemodelisdevelopedonestepatatime,addingXvariablesoneatatimetothemodelinorderoftheircontributiontochangesinY.Aproblemsolvingtoollistingthecharacteristicsofinterestononesideofthepage,andshowingcontributingfactorstothecharacteristicsasbranches.StandardErroroftheMean…ThSubgroup………SustainedProcessCapabilityt-test……………

Target…………TechnicalRequirement………TestSensitivity(d/s)…………Tolerance………TOP(TotalOpportunities)……Transfer………Transform……TrivialManyX’s………………

UCL(UpperControlLimit)……Unit……………Asampleoflikepartsorrelateddatatakenconsecutivelythatcontainsonlyinherentprocessvariation(‘whitenoise’)CapabilityofaprocessinthelongtermZ.ltAstatisticaltestusedtocomparetwomeans,ortocompareameantoastandardvalue.ThespecifiedordesiredaverageofaprocessPhysicalorprocesscharacteristicthatmustbecontrolledtoaddressaConsumerCue-alsoknownas“TheGap”.Astatisticusedtodeterminesamplesizeforhypothesistesting.Comparesthedifferenceinmeanstothespreadofthedata.Theamountofvariationallowablebydesigninaprocess.Tolerance=USL-LSL.Numberofopportunitiesperunittimesthenumberofunits.ThelastphaseofaSixSigmaproject,whereknowledgegainedistransferredtoallothersimilarprocesses-iesynergy.Anymathematicalrelationshipusedtotranslatedataofonespaceintodataofanotherspace(e.g.transformstoconvertnon-normaldatatonormaldata;log,reciprocal,powerfunctionsetc.)The80%oftheindependentvariables(X’s)thatgenerateonly20%ofthetotalprocessvariation.Variablesthatinfluencetheprocess,butatamuchlesssignificantlevelthanthe‘VitalFew’.TheuppercontrolboundaryonaStatisticalProcessControl(SPC)chart.Alimitcalculatedasthemeanplus3standarddeviations.NOTE:SEM(StandardErroroftheMean)isusedfors:stdev=s/sqrt(n)Auser-definedquantityrepresentingtheoutputofaprocess.Maybeapart,systemSubgroup………AsampleofUnit……………

USL……………Variance………VitalFewX’s…WhiteNoise……X………………X-bar…………X-bar/Rchart…Yresponse……Y.ft……………Y.na……………Y.rt………………Z.bench………Auser-definedquantityrepresentingtheoutputofaprocess.Maybeapart,system,componentofapartorasub-system.UpperSpecificationLimit(StandardDeviation)2The20%oftheindependentvariablesthatgenerate80%ofthetotalprocessvariation.TheseareX’swhichmustbecontrolledtobringaprocesstoSixSigmalevelsofperformance.See‘CommonCauseVariation’.Thenaturalvariationwithintheprocess,freeofexternalinfluences.Theindependentvariable(s),orinput(s),ofaprocess.Themeanoraverageofasample.Thesumofalldatainthesampledividedbythenumberofsamples.AStatisticalProcessControl(SPC)chartinwhichtheuppergraphisusedtoplotsubgroupaveragescomparedtocalculatedcontrollimits;thelowergraph(Range)plotsthedifferencebetweenthehighandlowvalueofthesubgroup.ControllimitsarealsousedontheRangechart.Thedependentvariable,oroutput,ofaprocess.‘FirstTime’or‘FirstPass’Yield.ClassicalYield.Numberofgoodunits/totalproduced.‘NormalizedAverageYield’.

(RolledThroughputYield)1/n.Averageyieldateachstepoftheprocess.‘RolledThroughputYield’.Yieldsofallstepsoftheprocessmultipliedtogether.Thereportedprocesscapability.Avaluederivedbycombiningalldefectsintoonetailofthedistribution,thenreadingtheZvalueUnit……………Auser-definedZ.bench………Ztransform……Z.lt………………

Z.st……………Thereportedprocesscapability.Avaluederivedbycombiningalldefectsintoonetailofthedistribution,thenreadingtheZvaluefromaNormaltable.Maybeshorttermorlongterm(mustquotewhich).