计量型数据MSA

019MSA-1EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.计量性数据MSA019MSA-2EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.简单测量分析－续SimpleMeasurementStudy-continued•使用描述性统计，计算Delta的标准差UsingDescriptiveStatistics,calculatethestandarddeviationofDelta•这个标准差被称为测量标准误差ThisstandarddeviationiscalledtheStandardErrorforMeasurement•其主要代表了什么，重复性还是再现性？Whichdoesthisprimarilyrepresent,RepeatabilityorReproducibility?VariableNMeanMedianStDevDelta10-0.410-0.4000.387019MSA-3EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.简单测量分析－续SimpleMeasurementStudy-continued•使用上面的标准差，什么是测量分布的良好估计值Usingthestandarddeviationabove,whatwouldbeagoodestimateofmeasurementspread?VariableNMeanMedianStDevDelta10-0.410-0.4000.3872.32=Precision0.387*6=Precision*6=Precision•假设公差是5（上限－下限）Supposetolerance(upperspec-lowerspec)is5.•精确度和公差的比率如何？Whatwouldbetheratiobetweentheprecisionandthetolerance?019MSA-4EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.简单测量分析－续SimpleMeasurementStudy-continued这个比率被称为精确度和公差比率(P/T).这个例子中P/T比率是：ThisratioisknownasthePrecisiontoTolerance(P/T)Ratio.ForthisexampletheP/TRatiois:这个可以被解读为：“46％的规格公差被测量系统误差所占用了”Thisisinterpretedas:“46%ofthespecificationtoleranceistakenupbymeasurementerror”VariableNMeanMedianStDevDelta10-0.410-0.4000.387PT/.PrecisionTolerance46%2325019MSA-5EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.测量能力指数-P/TMeasurementCapabilityIndex-P/T•精确性和公差比率PrecisiontoToleranceRatio•描述有多少百分比的公差是被测量系统误差占用了Addresseswhatpercentofthetoleranceistakenupbymeasurementerror•同时包含了重复性和再现性Includesbothrepeatabilityandreproducibility操作者x单位x试验次数OperatorxUnitxTrialexperiment•最佳情况Bestcase:10%；可以接受Acceptable:30%通常以百分比表示Usuallyexpressedaspercent注意：6.0标准差代表了99.73%的测量系统变差。记住在做测量系统和以前分析结果的计算时，要经常检查这个值。这里用的6.0是AIAG的工业标准。Note:6.0standarddeviationsaccountsfor99.73%ofMSvariation.Alwayscheckthevalueusedwhenmakingcomparisonsbetweenmeasurementsystemsorpreviousanalysiscalculations.Theuseof6.0istheAIAGindustrystandard.PTToleranceMS/6.0*019MSA-6EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.精确度和公差比率PrecisiontoToleranceRatioProductToleranceLSLUSL测量系统变差measurementsystemvariationP/T=20%P/T=100%P/T=200%019MSA-7EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.测量能力指数－%GR&RMeasurementCapabilityIndex-%GR&R•描述有多少总变差比率是由测量系统误差造成的Addresseswhatpercentofthetotalvariationistakenupbymeasurementerror•同时包含了重复性和再现性Includesbothrepeatabilityandreproducibility操作者x单位x试验次数OperatorxUnitxTrialexperiment•作为一个目标，我们期望%R&R30%Asatarget,lookfor%R&R30%通常以百分比表示Usuallyexpressedaspercent%&RRMSTotal100019MSA-8EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.％重复性和再现性%Repeatability&ReproducibilityObservedProcessVariation测量系统变差measurementsystemvariation%R&R=20%%R&R=75%%R&R=100%019MSA-9EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.%R&R=20%%R&R=50%观察到的流程变差ObservedProcessVariation%R&R=100%产品公差ProductToleranceLSLUSL测量系统变差measurementsystemvariationP/T=20%P/T=50%P/T=100%019MSA-10EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.%R&R=25%%R&R=50%观察到的流程变差ObservedProcessVariation%R&R=100%产品公差ProductToleranceLSLUSL测量系统变差measurementsystemvariationP/T=50%P/T=100%P/T=200%019MSA-11EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.%R&R=20%%R&R=40%%R&R=100%产品公差ProductToleranceLSLUSL测量系统变差measurementsystemvariationP/T=10%P/T=20%P/T=50%观察到的流程变差ObservedProcessVariation019MSA-12EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.10%30%30%P/TRatio%R&RRedYellowGreen15%主要问题–我的测量系统是否适用于我的项目MainQuestion-Ismymeasurementsystemokaytouseonmyproject?GR&R矩阵GR&RMetrics019MSA-13EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.GR&R练习GR&RExercise你将如何处理一下情形Whatwouldyoudointhisscenario?A)%P/T=9%%RR=8%_______________B)%P/T=43%%RR=72%_______________C)%P/T=43%%RR=16%_______________D)%P/T=13%%RR=43%_______________019MSA-14EffectiveDate:2/2004Copyright2000-2002SigmaBreakthroughTechnologies,Inc.Usedwithpermission.Copyright2004Cummins,Inc.AllRightsReserved.测量误差对能力指数的影响MeasurementErrorEffectonCapabilityIndices测量误差越高，对你了解真正流程产能的影响越剧烈！Thehigherthe